From 2c87bbe6bcb5975e3921ae092c1773753bd846d5 Mon Sep 17 00:00:00 2001 From: MSVSphere Packaging Team Date: Tue, 26 Nov 2024 16:06:08 +0300 Subject: [PATCH] import glibc-2.39-22.el10 --- .gitignore | 1 + .glibc.metadata | 1 + SOURCES/ChangeLog.old | 11823 ++++++++++++++++ SOURCES/RHEL-18039-1.patch | 79 + SOURCES/RHEL-18039-2.patch | 204 + SOURCES/bench.mk | 77 + SOURCES/glibc-RHEL-22226.patch | 121 + SOURCES/glibc-bench-compare | 153 + SOURCES/glibc-cs-path.patch | 44 + SOURCES/glibc-fedora-linux-tcsetattr.patch | 61 + SOURCES/glibc-fedora-localedata-rh61908.patch | 50 + SOURCES/glibc-fedora-manual-dircategory.patch | 31 + SOURCES/glibc-python3.patch | 30 + SOURCES/glibc-rh2292195-1.patch | 72 + SOURCES/glibc-rh2292195-2.patch | 84 + SOURCES/glibc-rh2292195-3.patch | 31 + SOURCES/glibc-rh827510.patch | 37 + SOURCES/glibc-upstream-2.39-1.patch | 264 + SOURCES/glibc-upstream-2.39-10.patch | 16 + SOURCES/glibc-upstream-2.39-11.patch | 80 + SOURCES/glibc-upstream-2.39-12.patch | 1453 ++ SOURCES/glibc-upstream-2.39-13.patch | 496 + SOURCES/glibc-upstream-2.39-14.patch | 250 + SOURCES/glibc-upstream-2.39-15.patch | 38 + SOURCES/glibc-upstream-2.39-16.patch | 446 + SOURCES/glibc-upstream-2.39-17.patch | 221 + SOURCES/glibc-upstream-2.39-18.patch | 24 + SOURCES/glibc-upstream-2.39-19.patch | 146 + SOURCES/glibc-upstream-2.39-2.patch | 35 + SOURCES/glibc-upstream-2.39-20.patch | 30 + SOURCES/glibc-upstream-2.39-21.patch | 24 + SOURCES/glibc-upstream-2.39-22.patch | 32 + SOURCES/glibc-upstream-2.39-23.patch | 669 + SOURCES/glibc-upstream-2.39-24.patch | 54 + SOURCES/glibc-upstream-2.39-25.patch | 144 + SOURCES/glibc-upstream-2.39-26.patch | 108 + SOURCES/glibc-upstream-2.39-27.patch | 61 + SOURCES/glibc-upstream-2.39-28.patch | 50 + SOURCES/glibc-upstream-2.39-29.patch | 246 + SOURCES/glibc-upstream-2.39-3.patch | 78 + SOURCES/glibc-upstream-2.39-30.patch | 2143 +++ SOURCES/glibc-upstream-2.39-31.patch | 207 + SOURCES/glibc-upstream-2.39-32.patch | 49 + SOURCES/glibc-upstream-2.39-33.patch | 20 + SOURCES/glibc-upstream-2.39-34.patch | 89 + SOURCES/glibc-upstream-2.39-35.patch | 32 + SOURCES/glibc-upstream-2.39-36.patch | 53 + SOURCES/glibc-upstream-2.39-37.patch | 54 + SOURCES/glibc-upstream-2.39-38.patch | 384 + SOURCES/glibc-upstream-2.39-39.patch | 49 + SOURCES/glibc-upstream-2.39-4.patch | 42 + SOURCES/glibc-upstream-2.39-40.patch | 209 + SOURCES/glibc-upstream-2.39-41.patch | 368 + SOURCES/glibc-upstream-2.39-42.patch | 31 + SOURCES/glibc-upstream-2.39-43.patch | 73 + SOURCES/glibc-upstream-2.39-44.patch | 34 + SOURCES/glibc-upstream-2.39-45.patch | 57 + SOURCES/glibc-upstream-2.39-46.patch | 34 + SOURCES/glibc-upstream-2.39-47.patch | 32 + SOURCES/glibc-upstream-2.39-48.patch | 40 + SOURCES/glibc-upstream-2.39-49.patch | 27 + SOURCES/glibc-upstream-2.39-5.patch | 143 + SOURCES/glibc-upstream-2.39-50.patch | 34 + SOURCES/glibc-upstream-2.39-51.patch | 57 + SOURCES/glibc-upstream-2.39-52.patch | 34 + SOURCES/glibc-upstream-2.39-53.patch | 207 + SOURCES/glibc-upstream-2.39-54.patch | 45 + SOURCES/glibc-upstream-2.39-55.patch | 38 + SOURCES/glibc-upstream-2.39-56.patch | 71 + SOURCES/glibc-upstream-2.39-57.patch | 143 + SOURCES/glibc-upstream-2.39-58.patch | 23 + SOURCES/glibc-upstream-2.39-59.patch | 40 + SOURCES/glibc-upstream-2.39-6.patch | 1052 ++ SOURCES/glibc-upstream-2.39-60.patch | 195 + SOURCES/glibc-upstream-2.39-61.patch | 1824 +++ SOURCES/glibc-upstream-2.39-62.patch | 38 + SOURCES/glibc-upstream-2.39-63.patch | 62 + SOURCES/glibc-upstream-2.39-64.patch | 68 + SOURCES/glibc-upstream-2.39-65.patch | 44 + SOURCES/glibc-upstream-2.39-66.patch | 57 + SOURCES/glibc-upstream-2.39-67.patch | 232 + SOURCES/glibc-upstream-2.39-68.patch | 244 + SOURCES/glibc-upstream-2.39-69.patch | 107 + SOURCES/glibc-upstream-2.39-7.patch | 46 + SOURCES/glibc-upstream-2.39-70.patch | 21 + SOURCES/glibc-upstream-2.39-71.patch | 69 + SOURCES/glibc-upstream-2.39-72.patch | 82 + SOURCES/glibc-upstream-2.39-75.patch | 28 + SOURCES/glibc-upstream-2.39-76.patch | 155 + SOURCES/glibc-upstream-2.39-77.patch | 107 + SOURCES/glibc-upstream-2.39-78.patch | 46 + SOURCES/glibc-upstream-2.39-79.patch | 30 + SOURCES/glibc-upstream-2.39-8.patch | 284 + SOURCES/glibc-upstream-2.39-80.patch | 66 + SOURCES/glibc-upstream-2.39-81.patch | 32 + SOURCES/glibc-upstream-2.39-82.patch | 53 + SOURCES/glibc-upstream-2.39-83.patch | 153 + SOURCES/glibc-upstream-2.39-84.patch | 43 + SOURCES/glibc-upstream-2.39-85.patch | 137 + SOURCES/glibc-upstream-2.39-86.patch | 58 + SOURCES/glibc-upstream-2.39-87.patch | 81 + SOURCES/glibc-upstream-2.39-88.patch | 201 + SOURCES/glibc-upstream-2.39-89.patch | 209 + SOURCES/glibc-upstream-2.39-9.patch | 172 + SOURCES/glibc-upstream-2.39-90.patch | 61 + SOURCES/glibc-upstream-2.39-91.patch | 88 + SOURCES/glibc-upstream-2.39-92.patch | 247 + SOURCES/glibc-upstream-2.39-93.patch | 52 + SOURCES/glibc-upstream-2.39-94.patch | 31 + SOURCES/glibc.attr | 3 + SOURCES/glibc.req.in | 42 + SOURCES/parse-SUPPORTED.py | 40 + SOURCES/wrap-find-debuginfo.sh | 147 + SPECS/glibc.spec | 4765 +++++++ 114 files changed, 34198 insertions(+) create mode 100644 .gitignore create mode 100644 .glibc.metadata create mode 100644 SOURCES/ChangeLog.old create mode 100644 SOURCES/RHEL-18039-1.patch create mode 100644 SOURCES/RHEL-18039-2.patch create mode 100644 SOURCES/bench.mk create mode 100644 SOURCES/glibc-RHEL-22226.patch create mode 100755 SOURCES/glibc-bench-compare create mode 100644 SOURCES/glibc-cs-path.patch create mode 100644 SOURCES/glibc-fedora-linux-tcsetattr.patch create mode 100644 SOURCES/glibc-fedora-localedata-rh61908.patch create mode 100644 SOURCES/glibc-fedora-manual-dircategory.patch create mode 100644 SOURCES/glibc-python3.patch create mode 100644 SOURCES/glibc-rh2292195-1.patch create mode 100644 SOURCES/glibc-rh2292195-2.patch create mode 100644 SOURCES/glibc-rh2292195-3.patch create mode 100644 SOURCES/glibc-rh827510.patch create mode 100644 SOURCES/glibc-upstream-2.39-1.patch create mode 100644 SOURCES/glibc-upstream-2.39-10.patch create mode 100644 SOURCES/glibc-upstream-2.39-11.patch create mode 100644 SOURCES/glibc-upstream-2.39-12.patch create mode 100644 SOURCES/glibc-upstream-2.39-13.patch create mode 100644 SOURCES/glibc-upstream-2.39-14.patch create mode 100644 SOURCES/glibc-upstream-2.39-15.patch create mode 100644 SOURCES/glibc-upstream-2.39-16.patch create mode 100644 SOURCES/glibc-upstream-2.39-17.patch create mode 100644 SOURCES/glibc-upstream-2.39-18.patch create mode 100644 SOURCES/glibc-upstream-2.39-19.patch create mode 100644 SOURCES/glibc-upstream-2.39-2.patch create mode 100644 SOURCES/glibc-upstream-2.39-20.patch create mode 100644 SOURCES/glibc-upstream-2.39-21.patch create mode 100644 SOURCES/glibc-upstream-2.39-22.patch create mode 100644 SOURCES/glibc-upstream-2.39-23.patch create mode 100644 SOURCES/glibc-upstream-2.39-24.patch create mode 100644 SOURCES/glibc-upstream-2.39-25.patch create mode 100644 SOURCES/glibc-upstream-2.39-26.patch create mode 100644 SOURCES/glibc-upstream-2.39-27.patch create mode 100644 SOURCES/glibc-upstream-2.39-28.patch create mode 100644 SOURCES/glibc-upstream-2.39-29.patch create mode 100644 SOURCES/glibc-upstream-2.39-3.patch create mode 100644 SOURCES/glibc-upstream-2.39-30.patch create mode 100644 SOURCES/glibc-upstream-2.39-31.patch create mode 100644 SOURCES/glibc-upstream-2.39-32.patch create mode 100644 SOURCES/glibc-upstream-2.39-33.patch create mode 100644 SOURCES/glibc-upstream-2.39-34.patch create mode 100644 SOURCES/glibc-upstream-2.39-35.patch create mode 100644 SOURCES/glibc-upstream-2.39-36.patch create mode 100644 SOURCES/glibc-upstream-2.39-37.patch create mode 100644 SOURCES/glibc-upstream-2.39-38.patch create mode 100644 SOURCES/glibc-upstream-2.39-39.patch create mode 100644 SOURCES/glibc-upstream-2.39-4.patch create mode 100644 SOURCES/glibc-upstream-2.39-40.patch create mode 100644 SOURCES/glibc-upstream-2.39-41.patch create mode 100644 SOURCES/glibc-upstream-2.39-42.patch create mode 100644 SOURCES/glibc-upstream-2.39-43.patch create mode 100644 SOURCES/glibc-upstream-2.39-44.patch create mode 100644 SOURCES/glibc-upstream-2.39-45.patch create mode 100644 SOURCES/glibc-upstream-2.39-46.patch create mode 100644 SOURCES/glibc-upstream-2.39-47.patch create mode 100644 SOURCES/glibc-upstream-2.39-48.patch create mode 100644 SOURCES/glibc-upstream-2.39-49.patch create mode 100644 SOURCES/glibc-upstream-2.39-5.patch create mode 100644 SOURCES/glibc-upstream-2.39-50.patch create mode 100644 SOURCES/glibc-upstream-2.39-51.patch create mode 100644 SOURCES/glibc-upstream-2.39-52.patch create mode 100644 SOURCES/glibc-upstream-2.39-53.patch create mode 100644 SOURCES/glibc-upstream-2.39-54.patch create mode 100644 SOURCES/glibc-upstream-2.39-55.patch create mode 100644 SOURCES/glibc-upstream-2.39-56.patch create mode 100644 SOURCES/glibc-upstream-2.39-57.patch create mode 100644 SOURCES/glibc-upstream-2.39-58.patch create mode 100644 SOURCES/glibc-upstream-2.39-59.patch create mode 100644 SOURCES/glibc-upstream-2.39-6.patch create mode 100644 SOURCES/glibc-upstream-2.39-60.patch create mode 100644 SOURCES/glibc-upstream-2.39-61.patch create mode 100644 SOURCES/glibc-upstream-2.39-62.patch create mode 100644 SOURCES/glibc-upstream-2.39-63.patch create mode 100644 SOURCES/glibc-upstream-2.39-64.patch create mode 100644 SOURCES/glibc-upstream-2.39-65.patch create mode 100644 SOURCES/glibc-upstream-2.39-66.patch create mode 100644 SOURCES/glibc-upstream-2.39-67.patch create mode 100644 SOURCES/glibc-upstream-2.39-68.patch create mode 100644 SOURCES/glibc-upstream-2.39-69.patch create mode 100644 SOURCES/glibc-upstream-2.39-7.patch create mode 100644 SOURCES/glibc-upstream-2.39-70.patch create mode 100644 SOURCES/glibc-upstream-2.39-71.patch create mode 100644 SOURCES/glibc-upstream-2.39-72.patch create mode 100644 SOURCES/glibc-upstream-2.39-75.patch create mode 100644 SOURCES/glibc-upstream-2.39-76.patch create mode 100644 SOURCES/glibc-upstream-2.39-77.patch create mode 100644 SOURCES/glibc-upstream-2.39-78.patch create mode 100644 SOURCES/glibc-upstream-2.39-79.patch create mode 100644 SOURCES/glibc-upstream-2.39-8.patch create mode 100644 SOURCES/glibc-upstream-2.39-80.patch create mode 100644 SOURCES/glibc-upstream-2.39-81.patch create mode 100644 SOURCES/glibc-upstream-2.39-82.patch create mode 100644 SOURCES/glibc-upstream-2.39-83.patch create mode 100644 SOURCES/glibc-upstream-2.39-84.patch create mode 100644 SOURCES/glibc-upstream-2.39-85.patch create mode 100644 SOURCES/glibc-upstream-2.39-86.patch create mode 100644 SOURCES/glibc-upstream-2.39-87.patch create mode 100644 SOURCES/glibc-upstream-2.39-88.patch create mode 100644 SOURCES/glibc-upstream-2.39-89.patch create mode 100644 SOURCES/glibc-upstream-2.39-9.patch create mode 100644 SOURCES/glibc-upstream-2.39-90.patch create mode 100644 SOURCES/glibc-upstream-2.39-91.patch create mode 100644 SOURCES/glibc-upstream-2.39-92.patch create mode 100644 SOURCES/glibc-upstream-2.39-93.patch create mode 100644 SOURCES/glibc-upstream-2.39-94.patch create mode 100644 SOURCES/glibc.attr create mode 100644 SOURCES/glibc.req.in create mode 100644 SOURCES/parse-SUPPORTED.py create mode 100644 SOURCES/wrap-find-debuginfo.sh create mode 100644 SPECS/glibc.spec diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..b446b34 --- /dev/null +++ b/.gitignore @@ -0,0 +1 @@ +SOURCES/glibc-2.39.tar.xz diff --git a/.glibc.metadata b/.glibc.metadata new file mode 100644 index 0000000..2501aee --- /dev/null +++ b/.glibc.metadata @@ -0,0 +1 @@ +4b043eaba31efbdfc92c85d062e975141870295e SOURCES/glibc-2.39.tar.xz diff --git a/SOURCES/ChangeLog.old b/SOURCES/ChangeLog.old new file mode 100644 index 0000000..c927cc3 --- /dev/null +++ b/SOURCES/ChangeLog.old @@ -0,0 +1,11823 @@ +* Wed Jul 27 2022 Arjun Shankar - 2.35.9000-32 +- Auto-sync with upstream branch master, + commit eaad4f9e8f07fc43618f6c8635a7e82831a423dd: +- arc4random: simplify design for better safety +- LoongArch: Update NEWS and README for the LoongArch port. +- LoongArch: Update build-many-glibcs.py for the LoongArch Port. +- LoongArch: Hard Float Support +- LoongArch: Build Infrastructure +- LoongArch: Add ABI Lists +- LoongArch: Linux ABI +- LoongArch: Linux Syscall Interface +- LoongArch: Atomic and Locking Routines +- LoongArch: Generic and soft-fp Routines +- LoongArch: Thread-Local Storage Support +- LoongArch: ABI Implementation +- LoongArch: Add relocations and ELF flags to elf.h and scripts/glibcelf.py +- LoongArch: Add LoongArch entries to config.h.in +- struct stat is not posix conformant on microblaze with __USE_FILE_OFFSET64 +- Linux: dirent/tst-readdir64-compat needs to use TEST_COMPAT (bug 27654) +- manual: Add documentation for arc4random functions +- s390x: Add optimized chacha20 +- powerpc64: Add optimized chacha20 +- x86: Add AVX2 optimized chacha20 +- x86: Add SSE2 optimized chacha20 +- aarch64: Add optimized chacha20 +- benchtests: Add arc4random benchtest +- stdlib: Add arc4random tests +- stdlib: Add arc4random, arc4random_buf, and arc4random_uniform (BZ #4417) +- locale: Optimize tst-localedef-path-norm +- malloc: Simplify implementation of __malloc_assert +- Update scripts/config.* files from upstream GNU config version +- linux: return UNSUPPORTED from tst-mount if entering mount namespace fails + +* Thu Jul 21 2022 Fedora Release Engineering - 2.35.9000-31 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_37_Mass_Rebuild + +* Sun Jul 17 2022 Patsy Griffin - 2.35.9000-30 +- Auto-sync with upstream branch master, + commit 49889fb256a7f9b894b2d16fea23de1ac25b65e2. +- x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level +- x86: Add support to build wcscpy with explicit ISA level +- x86: Add support to build strcmp/strlen/strchr with explicit ISA level +- elf: Fix wrong fscanf usage on tst-pldd +- Apply asm redirections in stdio.h before first use [BZ #27087] +- S390: Define SINGLE_THREAD_BY_GLOBAL only on s390x +- x86: Add missing rtm tests for strcmp family +- x86: Remove unneeded rtld-wmemcmp +- x86: Move wcslen SSE2 implementation to multiarch/wcslen-sse2.S +- x86: Move wcschr SSE2 implementation to multiarch/wcschr-sse2.S +- x86: Move strcat SSE2 implementation to multiarch/strcat-sse2.S +- x86: Move strchr SSE2 implementation to multiarch/strchr-sse2.S +- x86: Move strrchr SSE2 implementation to multiarch/strrchr-sse2.S +- x86: Move memrchr SSE2 implementation to multiarch/memrchr-sse2.S +- x86: Move strcpy SSE2 implementation to multiarch/strcpy-sse2.S +- x86: Move strlen SSE2 implementation to multiarch/strlen-sse2.S +- x86: Move strcmp SSE42 implementation to multiarch/strcmp-sse4_2.S +- x86: Move wcscmp SSE2 implementation to multiarch/wcscmp-sse2.S +- x86: Move strcmp SSE2 implementation to multiarch/strcmp-sse2.S +- x86: Rename STRCASECMP_NONASCII macro to STRCASECMP_L_NONASCII +- nptl: Fix ___pthread_unregister_cancel_restore asynchronous restore +- x86: Remove __mmask intrinsics in strstr-avx512.c +- x86: Remove generic strncat, strncpy, and stpncpy implementations +- i386: Remove -Wa,-mtune=i686 +- x86-64: Remove redundant strcspn-generic/strpbrk-generic/strspn-generic +- elf: Rename tst-audit26 to tst-audit28 +- x86-64: Don't mark symbols as hidden in strcmp-XXX.S +- stdlib: Tests for mbrtoc8, c8rtomb, and the char8_t typedef. +- stdlib: Implement mbrtoc8, c8rtomb, and the char8_t typedef. +- gconv: Correct Big5-HKSCS conversion to preserve all state bits. [BZ #25744] +- aarch64: Optimize string functions with shrn instruction +- test-container: return UNSUPPORTED for ENOSPC on clone() +- x86: Add support for building {w}memcmp{eq} with explicit ISA level +- x86: Add support for building {w}memset{_chk} with explicit ISA level +- x86: Add support for building {w}memmove{_chk} with explicit ISA level +- x86: Add support for building str{c|p}{brk|spn} with explicit ISA level +- x86: Add comment explaining no Slow_SSE4_2 check in ifunc-sse4_2 +- Replace __libc_multiple_threads with __libc_single_threaded +- linux: Add mount_setattr +- linux: Add tst-mount to check for Linux new mount API +- linux: Add open_tree +- linux: Add fspick +- linux: Add fsconfig +- AArch64: Reset HWCAP2_AFP bits in FPCR for default fenv +- elf: Fix direction of NODELETE log messages during symbol lookup + +* Fri Jul 8 2022 Stephen Gallagher - 2.35.9000-29 +- Modify glibc autorequires to exclude %%dist + +* Tue Jul 5 2022 Florian Weimer - 2.35.9000-28 +- ppc64le: Increase Clang compatibility of float128 redirects (#2100546) + +* Tue Jul 05 2022 Florian Weimer - 2.35.9000-27 +- Auto-sync with upstream branch master, + commit 7519dee356a0ab21c8990e59ed05dd48a4e573a0: +- malloc: Simplify checked_request2size interface +- stdlib: Simplify buffer management in canonicalize +- localedef: Support building for older C standards +- de_DE: Convert to UTF-8 +- locale: localdef input files are now encoded in UTF-8 +- locale: Introduce translate_unicode_codepoint into linereader.c +- locale: Fix signed char bug in lr_getc +- locale: Turn ADDC and ADDS into functions in linereader.c +- libc-symbols.h: remove unused macros + +* Mon Jul 04 2022 Carlos O'Donell - 2.35.9000-26 +- Auto-sync with upstream branch master, + commit 8ee2c043cfb35c48b45c7c5aed4022a8a7352bdc. +- Fix hurd namespace issues for internal signal functions +- argp: Remove old includes in !_LIBC case +- Use GCC 12 branch in build-many-glibcs.py +- Refactor internal-signals.h +- riscv: Use memcpy to handle unaligned access when fixing R_RISCV_RELATIVE +- AArch64: Add asymmetric faulting mode for tag violations in mem.tagging tunable +- linux: Fix mq_timereceive check for 32 bit fallback code (BZ 29304) +- x86: Add missing IS_IN (libc) check to strncmp-sse4_2.S +- x86: Add missing IS_IN (libc) check to strcspn-sse4.c +- x86: Add missing IS_IN (libc) check to memmove-ssse3.S +- x86-64: Properly indent X86_IFUNC_IMPL_ADD_VN arguments +- x86-64: Small improvements to dl-trampoline.S +- x86: Move mem{p}{mov|cpy}_{chk_}erms to its own file +- x86: Move and slightly improve memset_erms +- x86: Add definition for __wmemset_chk AVX2 RTM in ifunc impl list +- linux: Remove unnecessary nice.c and signal.c +- nptl: Remove unused members from struct pthread +- Linux: Forward declaration of struct iovec for process_madvise +- x86: Add more feature definitions to isa-level.h + +* Wed Jun 29 2022 DJ Delorie - 2.35.9000-25 +- Rebuild for fixed CI test for bz699724 + +* Tue Jun 28 2022 DJ Delorie - 2.35.9000-24 +- Auto-sync with upstream branch master, + commit a3563f3f369878467dd74aeb360448119a7a4b41. +- elf: Fix -DNDEBUG warning in _dl_start_args_adjust +- elf: Fix compile error with -Werror and -DNDEBUG +- x86-64: Only define used SSE/AVX/AVX512 run-time resolvers +- x86: Move CPU_FEATURE{S}_{USABLE|ARCH}_P to isa-level.h +- x86: Fix backwards Prefer_No_VZEROUPPER check in ifunc-evex.h +- x86: Rename strstr_sse2 to strstr_generic as it uses string/strstr.c +- x86: Remove unused file wmemcmp-sse4 +- x86: Put wcs{n}len-sse4.1 in the sse4.1 text section +- x86: Align entry for memrchr to 64-bytes. +- Makerules: Remove no-op -Wl,-d when linking libc_pic.os +- m68k: optimize RTLD_START +- misc: Optimize internal usage of __libc_single_threaded +- linux: Add move_mount +- linux: Add fsmount +- linux: Add fsopen +- resolv/tst-resolv-noaaaa: Support building for older C standards +- resolv: Implement no-aaaa stub resolver option +- support: Change non-address output format of support_format_dns_packet +- riscv: Use elf_machine_rela_relative to handle R_RISCV_RELATIVE +- x86: Remove faulty sanity tests for RTLD build with no multiarch +- stdlib: Fixup mbstowcs NULL __dst handling. [BZ #29279] +- x86: Replace all sse instructions with vex equivilent in avx+ files +- x86: Add support for compiling {raw|w}memchr with high ISA level +- x86: Add defines / utilities for making ISA specific x86 builds +- stdlib: Remove attr_write from mbstows if dst is NULL [BZ: 29265] +- stdlib: Remove trailing whitespace from Makefile +- debug: make __read_chk a cancellation point (bug 29274) +- s390: use LC_ALL=C for readelf call +- s390: use $READELF + +* Mon Jun 20 2022 Arjun Shankar - 2.35.9000-23 +- Auto-sync with upstream branch master, + commit e5446dfea11e969212939197b606424a718d9b65: +- i386: Fix include paths for strspn, strcspn, and strpbrk +- elf: Silence GCC 11/12 false positive warning +- x86: Rename generic functions with unique postfix for clarity +- x86: Add BMI1/BMI2 checks for ISA_V3 check +- x86-64: Handle fewer relocation types for RTLD_BOOTSTRAP +- aarch64: Handle fewer relocations for RTLD_BOOTSTRAP +- riscv: Change the relocations handled for RTLD_BOOTSTRAP +- x86: Cleanup bounds checking in large memcpy case +- x86: Add bounds `x86_non_temporal_threshold` +- Remove remnant reference to ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA +- elf: Remove ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA +- x86: Add sse42 implementation to strcmp's ifunc +- x86: Fix misordered logic for setting `rep_movsb_stop_threshold` +- elf: Refine direct extern access diagnostics to protected symbol +- Avoid -Wstringop-overflow= warning in iconv module. +- Add bounds check to __libc_ifunc_impl_list +- libio: Avoid RMW of flags2 outside lock (BZ #27842) +- x86: Optimize svml_s_tanhf4_core_sse4.S +- x86: Optimize svml_s_tanhf8_core_avx2.S +- x86: Add data file that can be shared by tanhf-avx2 and tanhf-sse4 +- x86: Optimize svml_s_tanhf16_core_avx512.S +- x86: Improve svml_s_atanhf4_core_sse4.S +- x86: Improve svml_s_atanhf8_core_avx2.S +- x86: Improve svml_s_atanhf16_core_avx512.S +- x86: Align varshift table to 32-bytes +- x86: Add copyright to strpbrk-c.c + +* Thu Jun 09 2022 Florian Weimer - 2.35.9000-22 +- Auto-sync with upstream branch master, + commit ace9e3edbca62d978b1e8f392d8a5d78500272d9: +- nss: handle stat failure in check_reload_and_get (BZ #28752) +- nss: add assert to DB_LOOKUP_FCT (BZ #28752) +- x86: Fix page cross case in rawmemchr-avx2 [BZ #29234] +- nptl_db: disable DT_RELR on libthread_db.so +- elf: add missing newlines in lateglobal test +- nptl: Fix __libc_cleanup_pop_restore asynchronous restore (BZ#29214) +- x86: ZERO_UPPER_VEC_REGISTERS_RETURN_XTEST expect no transactions +- x86: Shrink code size of memchr-evex.S +- x86: Shrink code size of memchr-avx2.S +- x86: Optimize memrchr-avx2.S +- x86: Optimize memrchr-evex.S +- x86: Optimize memrchr-sse2.S +- Benchtests: Improve memrchr benchmarks +- x86: Add COND_VZEROUPPER that can replace vzeroupper if no `ret` +- x86: Create header for VEC classes in x86 strings library +- powerpc: Fix VSX register number on __strncpy_power9 [BZ #29197] +- AArch64: Sort makefile entries +- AArch64: Add SVE memcpy +- x86_64: Add strstr function with 512-bit EVEX +- scripts/glibcelf.py: Add PT_AARCH64_MEMTAG_MTE constant + +* Mon Jun 06 2022 Carlos O'Donell - 2.35.9000-21 +- Auto-sync with upstream branch master, + commit 999835533bc60fbd0b0b65d2412a6742e5a54b9d: +- socket: Fix mistyped define statement in socket/sys/socket.h (BZ #29225) +- Declare timegm for ISO C2X +- Add PT_AARCH64_MEMTAG_MTE from Linux 5.18 to elf.h +- grep: egrep -> grep -E, fgrep -> grep -F +- string.h: Fix boolean spelling in comments +- elf: Add #include for use of E* constants. +- elf: Add #include for MAX usage. +- linux: Add process_mrelease +- linux: Add process_madvise +- linux: Set tst-pidfd-consts unsupported for kernels headers older than 5.10 +- testrun.sh: Support passing strace and valgrind arguments +- Linux: Adjust struct rseq definition to current kernel version +- iconv: Use 64 bit stat for gconv_parseconfdir (BZ# 29213) +- catgets: Use 64 bit stat for __open_catalog (BZ# 29211) +- inet: Use 64 bit stat for ruserpass (BZ# 29210) +- socket: Use 64 bit stat for isfdtype (BZ# 29209) +- posix: Use 64 bit stat for fpathconf (_PC_ASYNC_IO) (BZ# 29208) +- posix: Use 64 bit stat for posix_fallocate fallback (BZ# 29207) +- misc: Use 64 bit stat for getusershell (BZ# 29203) +- misc: Use 64 bit stat for daemon (BZ# 29203) +- linux: use statx for fstat if neither newfstatat nor fstatat64 is present +- Add MADV_DONTNEED_LOCKED from Linux 5.18 to bits/mman-linux.h +- Add HWCAP2_MTE3 from Linux 5.18 to AArch64 bits/hwcap.h +- i686: Use generic sincosf implementation for SSE2 version +- benchtests: Add workload name for sincosf +- i686: Use generic sinf implementation for SSE2 version +- i686: Use generic cosf implementation for SSE2 version +- benchtests: Add workload name for cosf +- x86_64: Optimize sincos where sin/cos is optimized (bug 29193) +- manual: fix reference to source file +- Add SOL_SMC from Linux 5.18 to bits/socket.h +- elf: Remove _dl_skip_args +- x86_64: Remove _dl_skip_args usage +- sparc: Remove _dl_skip_args usage +- sh: Remove _dl_skip_args usage +- s390: Remove _dl_skip_args usage +- riscv: Remove _dl_skip_args usage +- nios2: Remove _dl_skip_args usage (BZ# 29187) +- mips: Remove _dl_skip_args usage +- microblaze: Remove _dl_skip_args usage +- m68k: Remove _dl_skip_args usage +- ia64: Remove _dl_skip_args usage +- i686: Remove _dl_skip_args usage +- hppa: Remove _dl_skip_args usage (BZ# 29165) +- csky: Remove _dl_skip_args usage +- arc: Remove _dl_skip_args usage +- arm: Remove _dl_skip_args usage +- alpha: Remove _dl_skip_args usage +- benchtests: Improve benchtests for strstr, memmem, and memchr +- dlsym: Make RTLD_NEXT prefer default version definition [BZ #14932] +- x86-64: Ignore r_addend for R_X86_64_GLOB_DAT/R_X86_64_JUMP_SLOT +- x86_64: Implement evex512 version of strlen, strnlen, wcslen and wcsnlen +- Update kernel version to 5.18 in header constant tests +- String: Improve overflow test coverage for strnlen + +* Thu May 26 2022 Arjun Shankar - 2.35.9000-20 +- Auto-sync with upstream branch master, + commit 3d9926663cba19f40d26d8a8ab3b2a7cc09ffb13: +- Update syscall-names.list for Linux 5.18 +- Fix deadlock when pthread_atfork handler calls pthread_atfork or dlclose +- Use Linux 5.18 in build-many-glibcs.py +- stdio-common: Simplify printf_unknown interface in vfprintf-internal.c +- stdio-common: Move union printf_arg int +- stdio-common: Add printf specifier registry to +- elf/dl-reloc.c: Copyright The GNU Toolchain Authors +- benchtests: Improve bench-strnlen.c +- math: Add math-use-builtins-fabs (BZ#29027) +- linux: Add CLONE_NEWTIME from Linux 5.6 to bits/sched.h +- Revert "[ARM][BZ #17711] Fix extern protected data handling" +- Revert "[AArch64][BZ #17711] Fix extern protected data handling" +- elf: Rewrite long RESOLVE_MAP macro to an always_inline static function + +* Mon May 23 2022 DJ Delorie - 2.35.9000-19 +- Auto-sync with upstream branch master, + commit 748df8126ac69e68e0b94e236ea3c2e11b1176cb. +- dlfcn: Move RTLD_DEFAULT/RTLD_NEXT outside __USE_GNU +- elf: Optimize _dl_new_hash in dl-new-hash.h +- nss: Optimize nss_hash in nss_hash.c +- benchtests: Add benchtests for dl_elf_hash, dl_new_hash and nss_hash +- nss: Add tests for the nss_hash in nss_hash.h +- elf: Add tests for the dl hash funcs (_dl_new_hash and _dl_elf_hash) +- elf: Refactor dl_new_hash so it can be tested / benchmarked +- locale: Add more cached data to LC_CTYPE +- locale: Remove private union from struct __locale_data +- locale: Remove cleanup function pointer from struct __localedata +- locale: Call _nl_unload_locale from _nl_archive_subfreeres +- stdio-common: Add tst-memstream-string for open_memstream overflow +- __printf_fphex always uses LC_NUMERIC +- vfprintf: Consolidate some multibyte/wide character processing +- vfprintf: Move argument processing into vfprintf-process-arg.c +- stdio-common: Add tst-vfprintf-width-i18n to cover numeric field width +- string.h: fix __fortified_attr_access macro call [BZ #29162] +- Enable DT_RELR in glibc shared libraries and PIEs automatically +- S390: Enable static PIE +- linux: Add tst-pidfd.c +- linux: Add P_PIDFD +- linux: Add pidfd_send_signal +- linux: Add pidfd_getfd +- linux: Add pidfd_open +- aarch64: Move ld.so _start to separate file and drop _dl_skip_args +- linux: Add a getauxval test [BZ #23293] +- rtld: Remove DL_ARGV_NOT_RELRO and make _dl_skip_args const +- rtld: Use generic argv adjustment in ld.so [BZ #23293] +- scripts/glibcelf.py: Add *T_RISCV_* constants +- Remove dl-librecon.h header. +- elf: Remove ldconfig kernel version check +- Remove kernel version check +- linux: Use /sys/devices/system/cpu on __get_nprocs_conf (BZ#28991) +- csu: Implement and use _dl_early_allocate during static startup +- Linux: Introduce __brk_call for invoking the brk system call +- sys/cdefs.h: Do not require C++ compilers to define __STDC__ +- fortify: Ensure that __glibc_fortify condition is a constant [BZ #29141] +- Update RISC-V specific ELF definitions + +* Mon May 16 2022 Arjun Shankar - 2.35.9000-18 +- Auto-sync with upstream branch master, + commit 9403b71ae97e3f1a91c796ddcbb4e6f044434734: +- x86_64: Remove bzero optimization +- RISC-V: Use an autoconf template to produce `preconfigure' +- MIPS: Use an autoconf template to produce `preconfigure' +- m68k: Use an autoconf template to produce `preconfigure' +- C-SKY: Use an autoconf template to produce `preconfigure' +- Remove configure fno_unit_at_a_time +- stdio: Remove the usage of $(fno-unit-at-a-time) for siglist.c +- stdio: Remove the usage of $(fno-unit-at-a-time) for errlist.c +- Add declare_object_symbol_alias for assembly codes (BZ #28128) +- wcrtomb: Make behavior POSIX compliant + +* Tue May 10 2022 Patsy Griffin - 2.35.9000-17 +- Auto-sync with upstream branch master, + commit 8162147872491bb5b48e91543b19c49a29ae6b6d. +- nptl: Add backoff mechanism to spinlock loop +- Linux: Implement a useful version of _startup_fatal +- ia64: Always define IA64_USE_NEW_STUB as a flag macro +- linux: Fix posix_spawn return code if clone fails (BZ#29109) +- benchtests: Add wcrtomb microbenchmark +- clock_settime/clock_gettime: Use __nonnull to avoid null pointer +- clock_adjtime: Use __nonnull to avoid null pointer +- ntp_xxxtimex: Use __nonnull to avoid null pointer +- adjtimex/adjtimex64: Use __nonnull to avoid null pointer +- hurd spawni: Fix reauthenticating closed fds +- Linux: Define MMAP_CALL_INTERNAL +- i386: Honor I386_USE_SYSENTER for 6-argument Linux system calls +- i386: Remove OPTIMIZE_FOR_GCC_5 from Linux libc-do-syscall.S +- manual: Clarify that abbreviations of long options are allowed + +* Tue May 03 2022 Florian Weimer - 2.35.9000-16 +- Auto-sync with upstream branch master, + commit 8e28aa3a51bf0ef3683f2aed4b5b448744897b66: +- elf: Remove fallback to the start of DT_STRTAB for dladdr +- powerpc32: Remove unused HAVE_PPC_SECURE_PLT +- dlfcn: Implement the RTLD_DI_PHDR request type for dlinfo +- manual: Document the dlinfo function +- Do not use --hash-style=both for building glibc shared objects +- benchtests: Better libmvec integration +- benchtests: Add UNSUPPORTED benchmark status +- linux: Fix fchmodat with AT_SYMLINK_NOFOLLOW for 64 bit time_t (BZ#29097) +- Use __ehdr_start rather than _begin in _dl_start_final +- sysdeps: Add 'get_fast_jitter' interace in fast-jitter.h +- posix/glob.c: update from gnulib +- benchtests: Add pthread-mutex-locks bench +- linux: Fix missing internal 64 bit time_t stat usage +- elf: Fix DFS sorting algorithm for LD_TRACE_LOADED_OBJECTS with missing libraries (BZ #28868) +- posix: Remove unused definition on _Fork +- NEWS: Mention DT_RELR support +- elf: Add more DT_RELR tests +- elf: Properly handle zero DT_RELA/DT_REL values +- elf: Support DT_RELR relative relocation format [BZ #27924] +- Add GLIBC_ABI_DT_RELR for DT_RELR support +- elf: Define DT_RELR related macros and types +- elf: Replace PI_STATIC_AND_HIDDEN with opposite HIDDEN_VAR_NEEDS_DYNAMIC_RELOC +- i386: Regenerate ulps +- dlfcn: Do not use rtld_active () to determine ld.so state (bug 29078) +- INSTALL: Rephrase -with-default-link documentation + +* Mon Apr 25 2022 Carlos O'Donell - 2.35.9000-15 +- Auto-sync with upstream branch master, + commit 1305edd42c44fee6f8660734d2dfa4911ec755d6: +- elf: Move post-relocation code of _dl_start into _dl_start_final +- misc: Fix rare fortify crash on wchar funcs. [BZ 29030] +- elf: Remove unused enum allowmask +- scripts/glibcelf.py: Mark as UNSUPPORTED on Python 3.5 and earlier +- x86: Optimize {str|wcs}rchr-evex +- x86: Optimize {str|wcs}rchr-avx2 +- x86: Optimize {str|wcs}rchr-sse2 +- benchtests: Improve bench-strrchr +- x86-64: Fix SSE2 memcmp and SSSE3 memmove for x32 +- Default to --with-default-link=no (bug 25812) +- scripts: Add glibcelf.py module +- Add locale for syr_SY +- elf: Move elf_dynamic_do_Rel RTLD_BOOTSTRAP branches outside +- m68k: Handle fewer relocations for RTLD_BOOTSTRAP (#BZ29071) +- nptl: Fix pthread_cancel cancelhandling atomic operations +- x86: Fix missing __wmemcmp def for disable-multiarch build +- elf: Remove __libc_init_secure + +* Tue Apr 19 2022 DJ Delorie - 2.35.9000-14 +- Auto-sync with upstream branch master, + commit 78fb88827362fbd2cc8aa32892ae5b015106e25c. +- mips: Fix mips64n32 64 bit time_t stat support (BZ#29069) +- x86: Cleanup page cross code in memcmp-avx2-movbe.S +- x86: Remove memcmp-sse4.S +- x86: Optimize memcmp SSE2 in memcmp.S +- misc: Use 64 bit time_t interfaces on syslog +- misc: syslog: Move SYSLOG_NAME to USE_MISC (BZ #16355) +- misc: syslog: Use fixed-sized buffer and remove memstream +- misc: syslog: Simplify implementation +- misc: syslog: Fix indentation and style +- misc: Add syslog test +- support: Add xmkfifo +- stdio: Split __get_errname definition from errlist.c +- x86: Reduce code size of mem{move|pcpy|cpy}-ssse3 +- x86: Remove mem{move|cpy}-ssse3-back +- x86: Remove str{p}{n}cpy-ssse3 +- x86: Remove str{n}cat-ssse3 +- x86: Remove str{n}{case}cmp-ssse3 +- x86: Remove {w}memcmp-ssse3 +- nptl: Handle spurious EINTR when thread cancellation is disabled (BZ#29029) +- S390: Add new s390 platform z16. +- Replace {u}int_fast{16|32} with {u}int32_t +- stdlib: Reflow and sort most variable assignments +- elf: Fix memory leak in _dl_find_object_update (bug 29062) +- hurd: Define ELIBEXEC +- hurd: Fix arbitrary error code +- NEWS: Move PLT tracking slowdown to glibc 2.35. +- Remove _dl_skip_args_internal declaration +- test-container: Fix "unused code" warnings on HURD +- Add .clang-format style file +- manual: Avoid name collision in libm ULP table [BZ #28956] + +* Tue Apr 12 2022 Arjun Shankar - 2.35.9000-13 +- Add entry for Tarifit language locale for Morocco, and +- Auto-sync with upstream branch master, + commit 1a85970f41ea1e5abe6da2298a5e8fedcea26b70: +- powerpc: Relocate stinfo->main +- powerpc64: Set up thread register for _dl_relocate_static_pie +- powerpc64: Use medium model toc accesses throughout +- linux: Constify rfv variable in dl_vdso_vsym +- string: Replace outdated comments in strlen(). +- S390: Fix elf/tst-audit25[ab] +- sparc64: Remove fcopysign{f} implementation +- alpha: Remove fcopysign{f} implementation +- math: Use builtin for ldbl-96 copysign +- ia64: Remove fcopysign{f} implementation +- x86: Remove fcopysign{f} implementation +- powerpc: Remove fcopysign{f} implementation +- Add rif_MA locale [BZ #27781] +- tests/string: Drop simple/stupid/builtin tests +- test-memcpy: Actually reverse source and destination +- benchtests: Only build libmvec benchmarks iff $(build-mathvec) is set + +* Tue Apr 05 2022 Florian Weimer - 2.35.9000-12 +- Auto-sync with upstream branch master, + commit 053fe273434056f551ed8f81daf750db9dab5931: +- linux: Fix __closefrom_fallback iterates until max int (BZ#28993) +- Remove -z combreloc and HAVE_Z_COMBRELOC +- sparc: Remove s_abs implementations +- ia64: Remove fabs implementations +- x86: Remove fabs{f} implementation +- alpha: Remove s_abs implementations +- Allow for unpriviledged nested containers +- Increase the test timeout of some string tests +- realpath: Bring back GNU extension on ENOENT and EACCES [BZ #28996] +- stdlib: Fix tst-getrandom memcmp call +- stdlib: Fix tst-rand48.c printf types +- elf: Remove unused functions from tst-audit25(a,b) +- nptl: Use libc-diag.h with tst-thread-setspecific +- crypt: Remove unused variable on cert test +- elf: Remove unused variables in tests +- elf: Fix wrong fscanf usage on tst-pldd +- posix: Remove unused variable on tst-_Fork.c +- resolv: Initialize loop variable on tst-resolv-trailing +- locale: Remove set but unused variable on ld-collate.c +- localedate: Fix printf type on tst_mbrtowc +- localedata: Remove unused variables in tests +- x86: Small improvements for wcslen +- x86: Small improvements for wcscpy-ssse3 +- debug: Improve fdelt_chk error message +- Add HWCAP2_AFP, HWCAP2_RPRES from Linux 5.17 to AArch64 bits/hwcap.h +- x86: Remove AVX str{n}casecmp +- x86: Add EVEX optimized str{n}casecmp +- x86: Add AVX2 optimized str{n}casecmp +- string: Expand page cross test cases in test-strncmp.c +- string: Expand page cross test cases in test-strcmp.c +- x86: Optimize str{n}casecmp TOLOWER logic in strcmp-sse42.S +- x86: Optimize str{n}casecmp TOLOWER logic in strcmp.S +- string: Expand page cross tests in test-strncasecmp.c +- string: Expand page cross tests in test-strcasecmp.c +- benchtests: Use json-lib in bench-strncasecmp.c +- benchtests: Use json-lib in bench-strcasecmp.c +- x86: Fix fallback for wcsncmp_avx2 in strcmp-avx2.S [BZ #28896] +- x86: Remove strspn-sse2.S and use the generic implementation +- x86: Remove strpbrk-sse2.S and use the generic implementation +- x86: Remove strcspn-sse2.S and use the generic implementation +- x86: Optimize strspn in strspn-c.c +- x86: Optimize strcspn and strpbrk in strcspn-c.c +- benchtests: Use json-lib in bench-strspn.c +- benchtests: Use json-lib in bench-strpbrk.c +- x86: Code cleanup in strchr-evex and comment justifying branch +- x86: Code cleanup in strchr-avx2 and comment justifying branch +- benchtests: Add random benchmark in bench-strchr.c +- benchtests: Use json-lib in bench-strchr.c +- Update kernel version to 5.17 in tst-mman-consts.py +- iconvdata: Fix enum type on UTF-7 +- nscd: Remove unused variable +- support: Fix support_process_state_wait path size calculation +- support: Remove unused extract_8 function +- locale: Remove ununsed wctype_table_get function +- gmon: Remove unused sprofil.c functions +- Update syscall lists for Linux 5.17 +- Fix ununsed fstatat64_time64_statx +- malloc: Fix duplicate inline for do_set_mxfast +- elf: Remove inline _dl_dprintf +- configure.ac: fix bashisms in configure.ac +- getaddrinfo: Refactor code for readability +- Use Linux 5.17 in build-many-glibcs.py +- resolv: Fix unaligned accesses to fields in HEADER struct +- gai_init: Avoid jumping from if condition to its else counterpart +- gaiconf_init: Refactor some bits for readability +- gethosts: Return EAI_MEMORY on allocation failure +- gaih_inet: Split result generation into its own function +- gaih_inet: split loopback lookup into its own function +- gaih_inet: make gethosts into a function +- gaih_inet: separate nss lookup loop into its own function +- gaih_inet: Split nscd lookup code into its own function. +- gaih_inet: Split simple gethostbyname into its own function +- gaih_inet: make numeric lookup a separate routine +- gaih_inet: Simplify service resolution +- getaddrinfo: Fix leak with AI_ALL [BZ #28852] +- gaih_inet: Simplify canon name resolution +- Simplify allocations and fix merge and continue actions [BZ #28931] +- iconv: Add UTF-7-IMAP variant in utf-7.c +- iconv: make utf-7.c able to use variants +- iconv: Better mapping to RFC for UTF-7 +- iconv: Always encode "optional direct" UTF-7 characters +- stdio-common: Add wide stream coverage to tst-vfprintf-user-type +- libio: Flush-only _IO_str_overflow must not return EOF (bug 28949) +- libio: Convert tst_swprintf to the test framework +- scripts/dso-ordering-test.py: Fix C&P error in * callrefs processing +- stdio-common: Generate ja_JP.EUC-JP locale +- stdio-common: Re-flow and sort Makefile variables +- nss: Sort tests and tests-container and put one test per line +- benchtests: Use "=" instead of ":=" [BZ #28970] +- hppa: Use END instead of PSEUDO_END in swapcontext.S + +* Tue Mar 15 2022 DJ Delorie - 2.35.9000-11 +- Auto-sync with upstream branch master, + commit d05e6dc8d1032e1732542a48e0fb895432008b6e. +- hppa: Implement swapcontext in assembler (bug 28960) +- associate a deallocator for iconv_open +- associate a deallocation for opendir +- Add access function attributes to epoll_wait +- Add access function attributes to grp and shadow headers +- Define ISO 639-3 "tok" [BZ #28950] +- nss: Protect against errno changes in function lookup (bug 28953) +- nss: Do not mention NSS test modules in +- malloc: Exit early on test failure in tst-realloc +- Add some missing access function attributes +- libio: Ensure output buffer for wchars (bug #28828) +- inet: Return EAI_MEMORY when nrl_domainname() fails to allocate memory +- inet: Remove strdupa from nrl_domainname() +- inet: Fix getnameinfo (NI_NOFQDN) race condition (BZ#28566) +- benchtests: make compare_strings.py accept string as attribute value + +* Wed Mar 09 2022 Arjun Shankar - 2.35.9000-10 +- Drop glibc-rh1070416.patch; nscd related, thus no longer relevant. + +* Tue Mar 08 2022 Arjun Shankar - 2.35.9000-9 +- Auto-sync with upstream branch master, + commit 6de743a4e31a94e3d022e64a90c9082290a5a573: +- x86_64: Fix code formatting of vectorized math functions +- pthread: Do not overwrite tests-time64 +- x86_64: Fix svml_s_acosf16_core_avx512.S code formatting +- i386: Remove libc-do-syscall from sysdep-dl-routines [BZ #28936] +- linux/i386: remove dead assignment of sysdep-dl-routines + +* Tue Mar 08 2022 Siddhesh Poyarekar - 2.35.9000-8 +- Fix version check to accommodate gettext snapshot builds in rawhide. + +* Tue Mar 01 2022 Carlos O'Donell - 2.35.9000-7 +- Auto-sync with upstream branch master, + commit 2bbc694df279020a6620096d31c1e05c93966f9b: +- nptl: Fix cleanups for stack grows up [BZ# 28899] +- manual: SA_ONSTACK is ignored without alternate stack +- io: Add fsync call in tst-stat +- Linux: Consolidate auxiliary vector parsing (redo) + +* Fri Feb 25 2022 Arjun Shankar - 2.35.9000-6 +- Auto-sync with upstream branch master, + commit 1fe00d3eb602a0754873b536dc92fb6226759ee4: +- build: Properly generate .d dependency files [BZ #28922] +- benchtests: Generate .d dependency files [BZ #28922] +- benchtests: Remove duplicated loop in bench-bzero-walk.c +- localedata: Do not generate output if warnings were present. +- localedef: Update LC_MONETARY handling (Bug 28845) +- localedef: Handle symbolic links when generating locale-archive +- benchtests: Add small sizes (<= 64) to bench-bzero-walk.c +- math: Add more input to atanh accuracy tests +- resolv: Fix tst-resolv tests for 2.35 ABIs and later +- x86_64: Disable libmvec tests if multiarch not enabled [BZ# 28869] +- benchtests: Add benches for memset with 0 value +- i686: Remove bzero optimizations +- s390: Remove bzero optimizations +- powerpc: Remove powerpc64 bzero optimizations +- powerpc: Remove powerpc32 bzero optimizations +- sparc: Remove bzero optimization +- ia64: Remove bzero optimization +- alpha: Remove bzero optimization +- x86_64: Remove bcopy optimizations +- i386: Remove bcopy optimizations +- powerpc: Remove bcopy optimizations +- ia64: Remove bcopy +- hppa: Fix warnings from _dl_lookup_address +- hppa: Revise gettext trampoline design + +* Wed Feb 23 2022 Carlos O'Donell - 2.35.9000-5 +- Fix locale-archive generation (#2057697) + +* Tue Feb 22 2022 Carlos O'Donell - 2.35.9000-4 +- Auto-sync with upstream branch master, + commit fdc1ae67fef27eea1445bab4bdfe2f0fb3bc7aa1: +- Add SOL_MPTCP, SOL_MCTP from Linux 5.16 to bits/socket.h +- elf: Check invalid hole in PT_LOAD segments [BZ #28838] +- realpath: Do not copy result on failure (BZ #28815) +- x86: Fix TEST_NAME to make it a string in tst-strncmp-rtm.c +- x86: Test wcscmp RTM in the wcsncmp overflow case [BZ #28896] +- hppa: Fix swapcontext +- x86: Fallback {str|wcs}cmp RTM in the ncmp overflow case [BZ #28896] +- string: Add a testcase for wcsncmp with SIZE_MAX [BZ #28755] +- microblaze: Use the correct select syscall (BZ #28883) +- Update kernel version to 5.16 in tst-mman-consts.py +- pthread: Use 64 bit time_t stat internally for sem_open (BZ #28880) +- x86: Fix bug in strncmp-evex and strncmp-avx2 [BZ #28895] +- String: Strength memset tests in test-memset.c +- x86-64: Define __memcmpeq in ld.so +- htl: Destroy thread-specific data before releasing joins +- htl: Fix initializing the key lock +- mach: Fix LLL_SHARED value +- htl: Make pthread_[gs]etspecific not check for key validity +- x86-64: Remove bzero weak alias in SS2 memset +- hppa: Fix typo +- linux: Use socket-constants-time64.h on tst-socket-timestamp-compat +- x86/configure.ac: Define PI_STATIC_AND_HIDDEN/SUPPORT_STATIC_PIE +- Fix elf/tst-audit2 on hppa +- x86: Use CHECK_FEATURE_PRESENT on PCONFIG +- x86: Don't check PTWRITE in tst-cpu-features-cpuinfo.c +- x86: Set .text section in memset-vec-unaligned-erms +- Linux: Include in dl-sysdep.c only for SHARED +- Revert "Linux: Consolidate auxiliary vector parsing" +- String: Ensure 'MIN_PAGE_SIZE' is multiple of 'getpagesize' +- Use binutils 2.38 branch in build-many-glibcs.py +- elf: Remove LD_USE_LOAD_BIAS +- malloc: Remove LD_TRACE_PRELINKING usage from mtrace +- elf: Remove prelink support +- Linux: Consolidate auxiliary vector parsing +- Linux: Assume that NEED_DL_SYSINFO_DSO is always defined +- Linux: Remove DL_FIND_ARG_COMPONENTS +- Linux: Remove HAVE_AUX_SECURE, HAVE_AUX_XID, HAVE_AUX_PAGESIZE +- elf: Merge dl-sysdep.c into the Linux version +- hppa: Fix bind-now audit (BZ #28857) + +* Tue Feb 15 2022 Arjun Shankar - 2.35.9000-3 +- Reduce installed size of some langpacks by de-duplicating LC_CTYPE + +* Thu Feb 10 2022 Arjun Shankar - 2.35.9000-2 +- Drop glibc-fedora-localedef.patch and adjust locale installation + accordingly so that installed content remains unchanged. + +* Wed Feb 09 2022 Florian Weimer - 2.35.9000-1 +- Auto-sync with upstream branch master, + commit 3d9f171bfb5325bd5f427e9fc386453358c6e840: +- x86-64: Optimize bzero +- benchtests: Add benches for bzero +- linux: fix accuracy of get_nprocs and get_nprocs_conf [BZ #28865] +- x86: Remove SSSE3 instruction for broadcast in memset.S (SSE2 Only) +- benchtests: Sort benches in Makefile +- Benchtests: Add length zero benchmark for memset in bench-memset.c +- x86: Improve vec generation in memset-vec-unaligned-erms.S +- x86-64: Add vector tan/tanf to libmvec microbenchmark +- x86-64: Add vector erfc/erfcf to libmvec microbenchmark +- x86-64: Add vector asinh/asinhf to libmvec microbenchmark +- x86-64: Add vector tanh/tanhf to libmvec microbenchmark +- x86-64: Add vector erf/erff to libmvec microbenchmark +- x86-64: Add vector acosh/acoshf to libmvec microbenchmark +- x86-64: Add vector atanh/atanhf to libmvec microbenchmark +- x86-64: Add vector log1p/log1pf to libmvec microbenchmark +- x86-64: Add vector log2/log2f to libmvec microbenchmark +- x86-64: Add vector log10/log10f to libmvec microbenchmark +- x86-64: Add vector atan2/atan2f to libmvec microbenchmark +- x86-64: Add vector cbrt/cbrtf to libmvec microbenchmark +- x86-64: Add vector sinh/sinhf to libmvec microbenchmark +- x86-64: Add vector expm1/expm1f to libmvec microbenchmark +- x86-64: Add vector cosh/coshf to libmvec microbenchmark +- x86-64: Add vector exp10/exp10f to libmvec microbenchmark +- x86-64: Add vector exp2/exp2f to libmvec microbenchmark +- x86-64: Add vector hypot/hypotf to libmvec microbenchmark +- x86-64: Add vector asin/asinf to libmvec microbenchmark +- x86-64: Add vector atan/atanf to libmvec microbenchmark +- elf: Replace tst-audit24bmod2.so with tst-audit24bmod2 +- x86_64/multiarch: Sort sysdep_routines and put one entry per line +- string: Sort headers, routines, tests and tests-translation +- x86: Improve L to support L(XXX_SYMBOL (YYY, ZZZ)) +- Benchtests: move 'alloc_bufs' from loop in bench-memset.c +- x86-64: Fix strcmp-evex.S +- x86-64: Fix strcmp-avx2.S +- x86-64: Add vector acos/acosf to libmvec microbenchmark +- benchtests: Add more coverage for strcmp and strncmp benchmarks +- x86: Optimize strcmp-evex.S +- x86: Optimize strcmp-avx2.S +- string: Improve coverage in test-strcmp.c and test-strncmp.c +- string/test-str*cmp: remove stupid_[strcmp, strncmp, wcscmp, wcsncmp]. +- Open master branch for glibc 2.36 development + +* Tue Feb 08 2022 Florian Weimer - 2.35-2 +- Auto-sync with upstream branch release/2.35/master, + commit 24962427071fa532c3c48c918e9d64d719cc8a6c: +- Add BZ#28860 reference on NEWS +- linux: Fix missing __convert_scm_timestamps (BZ #28860) + +* Thu Feb 03 2022 Florian Weimer - 2.35-1 +- glibc 2.35 upstream release +- Auto-sync with upstream branch release/2.35/master, + commit a2f1675634b3513c09c38e55e6766e8c05768b1f: +- linux: __get_nprocs_sched: do not feed CPU_COUNT_S with garbage [BZ #28850] +- posix: Fix tst-spawn6 terminal handling (BZ #28853) +- Regenerate configure +- Create ChangeLog.old/ChangeLog.24. +- Prepare for glibc 2.35 release. +- Regenerate configure. +- Update install.texi, and regenerate INSTALL. +- Update NEWS bug list. +- Update NEWS. +- Update translations. +- Linux: Use ptrdiff_t for __rseq_offset +- Fix elf/tst-audit25a with default bind now toolchains +- posix: Replace posix_spawnattr_tc{get,set}pgrp_np with posix_spawn_file_actions_addtcsetpgrp_np +- or1k: Define PI_STATIC_AND_HIDDEN +- SET_RELHOOK: merge i386 and x86_64, and move to sysdeps/mach/hurd/x86 +- elf: Fix runtime linker auditing on aarch64 (BZ #26643) +- elf: Issue la_symbind for bind-now (BZ #23734) +- elf: Fix initial-exec TLS access on audit modules (BZ #28096) +- elf: Add la_activity during application exit +- localedata: Adjust C.UTF-8 to align with C/POSIX. +- localedef: Fix handling of empty mon_decimal_point (Bug 28847) +- malloc: Fix tst-mallocalign1 macro spacing. + +* Tue Feb 01 2022 Florian Weimer - 2.34.9000-39 +- Drop glibc-temp-Wno-use-after-free.patch, fixed upstream. +- Auto-sync with upstream branch master, + commit 3fb18fd80c5900cc82748f3320b30516c57d24da: +- elf: Add +- Mention _FORTIFY_SOURCE=3 for gcc12 in NEWS +- malloc: Fix -Wuse-after-free warning in tst-mallocalign1 [BZ #26779] +- Update libc.pot for 2.35 release. +- tst-socket-timestamp-compat.c: Check __TIMESIZE [BZ #28837] +- Add prelink removal plan on NEWS +- Linux: Only generate 64 bit timestamps for 64 bit time_t recvmsg/recvmmsg +- linux: Fix ancillary 64-bit time timestamp conversion (BZ #28349, BZ#28350) +- support: Add support_socket_so_timestamp_time64 +- Fix elf/loadfail test build dependencies +- Fix glibc 2.34 ABI omission (missing GLIBC_2.34 in dynamic loader) +- x86: Use CHECK_FEATURE_PRESENT to check HLE [BZ #27398] +- Guard tst-valgrind-smoke.out with run-built-tests +- hurd: Add posix_spawnattr_tc{get,set}pgrp_np on libc.abilist +- Avoid -Wuse-after-free in tests [BZ #26779]. +- elf: Replace tst-p_alignmod1-editX with a python script +- stdlib: Avoid -Wuse-after-free in __add_to_environ [BZ #26779] +- io: Fix use-after-free in ftw [BZ #26779] +- intl: Avoid -Wuse-after-free [BZ #26779] +- elf: Fix use-after-free in ldconfig [BZ #26779] +- posix: Add terminal control setting support for posix_spawn + +* Mon Jan 24 2022 DJ Delorie - 2.34.9000-38 +- Auto-sync with upstream branch master, + commit 5b8e7980c5dabd9aaefeba4f0208baa8cf7653ee. +- Linux: Detect user namespace support in io/tst-getcwd-smallbuff +- Fix handling of unterminated bracket expressions in fnmatch (bug 28792) +- realpath: Avoid overwriting preexisting error (CVE-2021-3998) +- elf: Add a test for PT_LOAD segments with invalid p_align [BZ #28688] +- elf: Add a test for PT_LOAD segments with p_align == 1 [BZ #28688] +- elf: Add a test for PT_LOAD segments with mixed p_align [BZ #28676] +- Add and use link-test-modules-rpath-link [BZ #28455] +- tst-realpath-toolong: Fix hurd build +- getcwd: Set errno to ERANGE for size == 1 (CVE-2021-3999) +- Add valgrind smoke test +- htl: Fix cleaning the reply port +- elf: Properly align all PT_LOAD segments [BZ #28676] +- realpath: Set errno to ENAMETOOLONG for result larger than PATH_MAX [BZ #28770] +- support: Add helpers to create paths longer than PATH_MAX +- nptl: Effectively skip CAS in spinlock loop +- mips: Move DT_MIPS into +- x86_64: Document libmvec vector functions accuracy [BZ #28766] +- x86: Black list more Intel CPUs for TSX [BZ #27398] +- elf: Fix tst-align3 +- elf: Move _dl_setup_hash to its own file +- htl: Fix build error in annexc +- elf: Reinstate tst-audit17 +- x86: use default cache size if it cannot be determined [BZ #28784] +- rt/tst-mqueue*: Return UNSUPPORTED when mq_open fails with ENOSYS +- Linux: Add epoll_pwait2 (BZ #27359) +- Properly handle --disable-default-pie [BZ #28780] +- elf: Fix 64 time_t support for installed statically binaries +- Revert "elf: Fix 64 time_t support for installed statically binaries" +- CVE-2022-23218: Buffer overflow in sunrpc svcunix_create (bug 28768) +- sunrpc: Test case for clnt_create "unix" buffer overflow (bug 22542) +- CVE-2022-23219: Buffer overflow in sunrpc clnt_create for "unix" (bug 22542) +- socket: Add the __sockaddr_un_set function +- elf/tst-dl_find_object: Disable subtests for non-contiguous maps (bug 28732) +- elf: Set l_contiguous to 1 for the main map in more cases +- elf: Introduce rtld_setup_main_map +- hurd: Make RPC input array parameters const +- hurd: optimize exec cleanup +- hurd: Add __rtld_execve +- hurd: Fix exec() leak on proc_task2proc failure +- htl: Hide __pthread_attr's __schedparam type [BZ #23088] +- htl: Clear kernel_thread field before releasing the thread structure +- hurd: drop SA_SIGINFO availability xfail +- hurd: Fix timer/clock_getres crash on NULL res parameter +- hurd: Fix pthread_kill on exiting/ted thread +- [hurd] Drop spurious #ifdef SHARED +- [hurd] Call _dl_sort_maps_init in _dl_sysdep_start +- elf tst-dl_find_object: Fix typo +- s390x: Use in early HWCAP check +- x86: Add x86-64-vN check to early startup +- powerpc64le: Use in early HWCAP check +- Add --with-rtld-early-cflags configure option +- elf: Split dl-printf.c from dl-misc.c +- elf/Makefile: Reflow and sort most variable assignments +- Generate gcc-macros.h +- x86: HAVE_X86_LAHF_SAHF, HAVE_X86_MOVBE and -march=x86-64-vN (bug 28782) +- math: Add more inputs to atan2 accuracy tests [BZ #28765] +- Disable debuginfod in printer tests [BZ #28757] +- Update syscall lists for Linux 5.16 +- i386: Remove broken CAN_USE_REGISTER_ASM_EBP (bug 28771) +- stdlib: Fix formatting of tests list in Makefile +- stdlib: Sort tests in Makefile +- x86_64: Fix SSE4.2 libmvec atan2 function accuracy [BZ #28765] +- debug: Synchronize feature guards in fortified functions [BZ #28746] +- debug: Autogenerate _FORTIFY_SOURCE tests +- Do not build libresolv module with 64 bit time_t flags +- Revert "linux: Fix ancillary 64-bit time timestamp conversion (BZ #28349, BZ #28350)" +- Revert "support: Add support_socket_so_timestamp_time64" +- timezone: Fix tst-bz28707 Makefile rule +- linux: Fix ancillary 64-bit time timestamp conversion (BZ #28349, BZ #28350) +- support: Add support_socket_so_timestamp_time64 +- elf: Fix 64 time_t support for installed statically binaries +- Enable _FORTIFY_SOURCE=3 for gcc 12 and above +- manual: Drop obsolete @refill +- aarch64: Add HWCAP2_ECV from Linux 5.16 +- Use Linux 5.16 in build-many-glibcs.py +- x86: Fix __wcsncmp_evex in strcmp-evex.S [BZ# 28755] +- x86: Fix __wcsncmp_avx2 in strcmp-avx2.S [BZ# 28755] +- math: Fix float conversion regressions with gcc-12 [BZ #28713] +- elf: Simplify software TM implementation in _dl_find_object +- Restore ENTRY_POINT definition on hppa, ia64 (bug 28749) +- elf: Fix fences in _dl_find_object_update (bug 28745) +- ttydefaults.h: Fix CSTATUS to control-t +- AArch64: Check for SVE in ifuncs [BZ #28744] +- debug: Remove catchsegv and libSegfault (BZ #14913) +- Documentation for OpenRISC port +- build-many-glibcs.py: add OpenRISC support +- or1k: Build Infrastructure +- or1k: ABI lists +- or1k: Linux ABI +- or1k: Linux Syscall Interface +- or1k: math soft float support +- or1k: Atomics and Locking primitives +- or1k: Thread Local Storage support +- or1k: startup and dynamic linking code +- or1k: ABI Implementation +- linux/syscalls: Add or1k_atomic syscall for OpenRISC +- elf: Add reloc for OpenRISC +- elf: Add a comment after trailing backslashes +- elf: Also try DT_RUNPATH for LD_AUDIT dlopen [BZ #28455] +- elf: Fix tst-linkall-static link when pthread is not in libc + +* Thu Jan 20 2022 Fedora Release Engineering - 2.34.9000-37 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_36_Mass_Rebuild + +* Tue Jan 04 2022 Florian Weimer - 2.34.9000-36 +- Auto-sync with upstream branch master, + commit 28713c06129f8f64f88c423266e6ff2880216509: +- elf: Sort tests and modules-names +- hurd: nuke all unknown ports on exec +- hurd: Fix auth port leak +- Remove stale reference to libanl.a +- elf: Add +- Properly check linker option in LIBC_LINKER_FEATURE [BZ #28738] +- hurd: Implement _S_msg_get_dtable +- Update automatically-generated copyright dates +- Sync move-if-change from Gnulib, updating copyright +- Update copyright dates not handled by scripts/update-copyrights. +- Update copyright dates with scripts/update-copyrights +- hurd: Use __trivfs_server_name instead of trivfs_server_name +- hurd: Bump BRK_START to 0x20000000 +- hurd: Avoid overzealous shared objects constraints +- time: Refactor timesize.h for some ABIs +- hurd: Make getrandom a stub inside the random translator +- open64: Force O_LARGEFILE on all architectures +- x86-64: Add vector tan/tanf implementation to libmvec +- x86-64: Add vector erfc/erfcf implementation to libmvec +- resolv: Do not install libanl.so symbolic link +- resolv: Do not build libanl.so for ABIs starting at 2.35 +- timezone: test-case for BZ #28707 +- timezone: handle truncated timezones from tzcode-2021d and later (BZ #28707) +- x86-64: Add vector asinh/asinhf implementation to libmvec +- x86-64: Add vector tanh/tanhf implementation to libmvec +- x86-64: Add vector erf/erff implementation to libmvec +- x86-64: Add vector acosh/acoshf implementation to libmvec +- x86-64: Add vector atanh/atanhf implementation to libmvec +- x86-64: Add vector log1p/log1pf implementation to libmvec +- x86-64: Add vector log2/log2f implementation to libmvec +- x86-64: Add vector log10/log10f implementation to libmvec +- x86-64: Add vector atan2/atan2f implementation to libmvec +- x86-64: Add vector cbrt/cbrtf implementation to libmvec +- x86-64: Add vector sinh/sinhf implementation to libmvec +- x86-64: Add vector expm1/expm1f implementation to libmvec +- x86-64: Add vector cosh/coshf implementation to libmvec +- x86-64: Add vector exp10/exp10f implementation to libmvec +- x86-64: Add vector exp2/exp2f implementation to libmvec +- x86-64: Add vector hypot/hypotf implementation to libmvec +- x86-64: Add vector asin/asinf implementation to libmvec +- x86-64: Add vector atan/atanf implementation to libmvec + +* Wed Dec 29 2021 Florian Weimer - 2.34.9000-35 +- Auto-sync with upstream branch master, + commit 5d28a8962dcb6ec056b81d730e3c6fb57185a210: +- elf: Add _dl_find_object function +- malloc: Remove memusage.h +- malloc: Use hp-timing on libmemusage +- Remove atomic-machine.h atomic typedefs +- malloc: Remove atomic_* usage +- microblaze: Add missing implementation when !__ASSUME_TIME64_SYSCALLS +- elf: Do not fail for failed dlmopen on audit modules (BZ #28061) +- elf: Issue audit la_objopen for vDSO +- elf: Add audit tests for modules with TLSDESC +- elf: Avoid unnecessary slowdown from profiling with audit (BZ#15533) +- elf: Add _dl_audit_pltexit +- elf: Add _dl_audit_pltenter +- elf: Add _dl_audit_preinit +- elf: Add _dl_audit_symbind_alt and _dl_audit_symbind +- elf: Add _dl_audit_objclose +- elf: Add _dl_audit_objsearch +- elf: Add _dl_audit_activity_map and _dl_audit_activity_nsid +- elf: Add _dl_audit_objopen +- hurd: Fix static-PIE startup +- hurd: let csu initialize tls +- hurd: Fix XFAIL-ing mallocfork2 tests +- hurd: XFAIL more tests that require setpshared support +- malloc: Add missing shared thread library flags +- stdio-common: Fix %m sprintf test output for GNU/Hurd +- x86: Optimize L(less_vec) case in memcmpeq-evex.S +- x86: Optimize L(less_vec) case in memcmp-evex-movbe.S +- elf: Remove AArch64 from comment for AT_MINSIGSTKSZ +- math: Properly cast X_TLOSS to float [BZ #28713] +- Set default __TIMESIZE default to 64 +- stdio: Implement %#m for vfprintf and related functions +- elf: Remove unused NEED_DL_BASE_ADDR and _dl_base_addr +- x86-64: Add vector acos/acosf implementation to libmvec +- intl/plural.y: Avoid conflicting declarations of yyerror and yylex +- elf: Remove excessive p_align check on PT_LOAD segments [BZ #28688] +- s_sincosf.h: Change pio4 type to float [BZ #28713] +- Linux: Fix 32-bit vDSO for clock_gettime on powerpc32 +- Regenerate ulps on x86_64 with GCC 12 +- Add ARPHRD_CAN, ARPHRD_MCTP to net/if_arp.h +- Remove ununsed tcb-offset +- riscv: align stack before calling _dl_init [BZ #28703] +- riscv: align stack in clone [BZ #28702] +- elf: Fix tst-cpu-features-cpuinfo for KVM guests on some AMD systems [BZ #28704] +- powerpc64[le]: Allocate extra stack frame on syscall.S +- Update copyright header in recently merged ab_GE locale +- fortify: Fix spurious warning with realpath + +* Tue Dec 28 2021 Florian Weimer - 2.34.9000-34 +- armhfp, i686: Revert 64-bit time_t/off_t for internal use (#2034715) + +* Fri Dec 17 2021 Florian Weimer - 2.34.9000-33 +- Auto-sync with upstream branch master, + commit b99b0f93ee8762fe53ff65802deb6f00700b9924: +- nss: Use "files dns" as the default for the hosts database (#2033020) +- arm: Guard ucontext _rtld_global_ro access by SHARED, not PIC macro +- Fix The GNU ToolChain Authors copyright notice +- Remove upper limit on tunable MALLOC_MMAP_THRESHOLD +- localedata: add new locale ab_GE +- Fix __minimal_malloc segfaults in __mmap due to stack-protector +- __glibc_unsafe_len: Fix comment +- malloc: Enable huge page support on main arena +- malloc: Move MORECORE fallback mmap to sysmalloc_mmap_fallback +- malloc: Add Huge Page support to arenas +- malloc: Add Huge Page support for mmap +- malloc: Move mmap logic to its own function +- malloc: Add THP/madvise support for sbrk +- malloc: Add madvise support for Transparent Huge Pages +- powerpc: Use global register variable in +- Use LFS and 64 bit time for installed programs (swbz#15333) + +* Wed Dec 15 2021 Arjun Shankar - 2.34.9000-32 +- Do not use --enable-static-pie configure flag since it is now ignored + +* Wed Dec 15 2021 Florian Weimer - 2.34.9000-31 +- Drop glibc-fedora-revert-PT_LOAD-segment-align.patch; fix applied upstream +- Auto-sync with upstream branch master, + commit 4435c29892c43ae9908a42e591747be63102689b: +- Support target specific ALIGN for variable alignment test [BZ #28676] +- NEWS: Document LD_PREFER_MAP_32BIT_EXEC as x86-64 only +- elf: Align argument of __munmap to page size [BZ #28676] + +* Tue Dec 14 2021 Arjun Shankar - 2.34.9000-30 +- Add glibc-fedora-revert-PT_LOAD-segment-align.patch to revert upstream + commits 718fdd87b1b9 and fc2334ab32e0. +- Auto-sync with upstream branch master, + commit 0884724a95b60452ad483dbe086d237d02ba624d: +- intl: Emit no lines in bison generated files +- hurd: Do not set PIE_UNSUPPORTED +- NEWS: Move LD_PREFER_MAP_32BIT_EXEC +- mach: Fix spurious inclusion of stack_chk_fail_local in libmachuser.a +- Disable DT_RUNPATH on NSS tests [BZ #28455] +- sysdeps: Simplify sin Taylor Series calculation +- math: Remove the error handling wrapper from hypot and hypotf +- math: Use fmin/fmax on hypot +- aarch64: Add math-use-builtins-f{max,min}.h +- math: Add math-use-builtinds-fmin.h +- math: Add math-use-builtinds-fmax.h +- math: Remove powerpc e_hypot +- i386: Move hypot implementation to C +- math: Use an improved algorithm for hypotl (ldbl-128) +- math: Use an improved algorithm for hypotl (ldbl-96) +- math: Improve hypot performance with FMA +- math: Use an improved algorithm for hypot (dbl-64) +- math: Simplify hypotf implementation +- Cleanup encoding in comments +- Replace --enable-static-pie with --disable-default-pie +- hurd: Add rules for static PIE build +- hurd: Fix gmon-static +- x86-64: Remove LD_PREFER_MAP_32BIT_EXEC support [BZ #28656] +- elf: Use errcode instead of (unset) errno in rtld_chain_load +- Add a testcase to check alignment of PT_LOAD segment [BZ #28676] +- elf: Properly align PT_LOAD segments [BZ #28676] +- elf: Install a symbolic link to ld.so as /usr/bin/ld.so +- nptl: Add one more barrier to nptl/tst-create1 +- Remove TLS_TCB_ALIGN and TLS_INIT_TCB_ALIGN +- nptl: rseq failure after registration on main thread is fatal +- nptl: Add public rseq symbols and +- nptl: Add glibc.pthread.rseq tunable to control rseq registration +- Linux: Use rseq to accelerate sched_getcpu +- nptl: Add rseq registration +- nptl: Introduce THREAD_GETMEM_VOLATILE +- nptl: Introduce for THREAD_* accessors +- nptl: Add for defining __thread_pointer +- String: test-memcpy used unaligned types for buffers [BZ 28572] +- localedef: check magic value on archive load [BZ #28650] +- x86: Don't set Prefer_No_AVX512 for processors with AVX512 and AVX-VNNI +- linux: Add generic ioctl implementation +- linux: Add generic syscall implementation +- misc, nptl: Remove stray references to __condvar_load_64_relaxed +- csu: Always use __executable_start in gmon-start.c +- elf: execve statically linked programs instead of crashing [BZ #28648] +- Add --with-timeoutfactor=NUM to specify TIMEOUTFACTOR + +* Mon Dec 13 2021 Arjun Shankar - 2.34.9000-29 +- Remove nscd (#1905142) + https://fedoraproject.org/wiki/Changes/RemoveNSCD + +* Fri Dec 10 2021 Pavel Březina - 2.34.9000-28 +- /etc/nsswitch.conf is now owned by authselect (rhbz#2023741) + +* Thu Dec 09 2021 Siddhesh Poyarekar - 2.34.9000-27 +- Set BuildFlagsNonshared only if _annotated_build is set. + +* Sat Dec 04 2021 Florian Weimer - 2.34.9000-26 +- Drop glibc-rh2026399.patch, not needed anymore due to upstream fix. +- Auto-sync with upstream branch master, + commit 4df1fa6ddc8925a75f3da644d5da3bb16eb33f02: +- x86-64: Use notl in EVEX strcmp (#2026399) +- nptl: Increase default TCB alignment to 32 +- elf: add definition for ELF_NOTE_FDO and NT_FDO_PACKAGING_METADATA note +- AArch64: Improve A64FX memcpy +- AArch64: Optimize memcmp +- powerpc64[le]: Fix CFI and LR save address for asm syscalls [BZ #28532] +- linux: Implement pipe in terms of __NR_pipe2 +- linux: Implement mremap in C +- linux: Add prlimit64 C implementation +- elf: Include in tst-tls20.c +- elf: Include in tst-tls20.c +- hurd: Let report-wait use a weak reference to _hurd_itimer_thread + +* Sat Dec 4 2021 Florian Weimer - 2.34.9000-25 +- x86_64: Disable additional EVEX string functions (#2026399) + +* Fri Dec 3 2021 Florian Weimer - 2.34.9000-24 +- x86_64: Disable EVEX *cmp* string functions (#2026399) + +* Thu Dec 2 2021 Florian Weimer - 2.34.9000-23 +- Drop glibc-sdt-headers.patch; the official systemtap headers work again. + +* Thu Nov 25 2021 Carlos O'Donell - 2.34.9000-22 +- Auto-sync with upstream branch master, + commit 137ed5ac440a4d3cf4178ce97f349b349a9c2c66: +- linux: Use /proc/stat fallback for __get_nprocs_conf (BZ #28624) +- linux: Add fanotify_mark C implementation +- linux: Only build fstatat fallback if required +- regex: fix buffer read overrun in search [BZ#28470] +- x86-64: Add vector sin/sinf to libmvec microbenchmark +- x86-64: Add vector pow/powf to libmvec microbenchmark +- x86-64: Add vector log/logf to libmvec microbenchmark +- x86-64: Add vector exp/expf to libmvec microbenchmark +- x86-64: Add vector cos/cosf to libmvec microbenchmark +- io: Refactor close_range and closefrom +- nptl: Do not set signal mask on second setjmp return [BZ #28607] +- powerpc: Define USE_PPC64_NOTOC iff compiler supports it +- setjmp: Replace jmp_buf-macros.h with jmp_buf-macros.sym +- Update kernel version to 5.15 in tst-mman-consts.py +- socket: Do not use AF_NETLINK in __opensock +- elf: Move la_activity (LA_ACT_ADD) after _dl_add_to_namespace_list() (BZ #28062) +- Add PF_MCTP, AF_MCTP from Linux 5.15 to bits/socket.h +- malloc: Fix malloc debug for 2.35 onwards +- elf: Introduce GLRO (dl_libc_freeres), called from __libc_freeres +- nptl: Extract from pthread_cond_common.c + +* Wed Nov 17 2021 Florian Weimer - 2.34.9000-21 +- Temporary patch glibc-sdt-headers.patch, to fix s390x build failure. +- Temporary patch glibc-dso-sort-makefile-fail.patch, to fix x86_64 build. +- Auto-sync with upstream branch master, + commit a43c0b5483da4c5e3796af309864cb44256c02db: +- x86-64: Create microbenchmark infrastructure for libmvec +- elf: hidden visibility for __minimal_malloc functions +- elf: Use a temporary file to generate Makefile fragments [BZ #28550] +- dso-ordering-test.py: Put all sources in one directory [BZ #28550] +- elf: Move LAV_CURRENT to link_lavcurrent.h +- Move assignment out of the CAS condition +- Add a comment for --enable-initfini-array [BZ #27945] +- tst-tzset: output reason when creating 4GiB file fails +- Add LLL_MUTEX_READ_LOCK [BZ #28537] +- Avoid extra load with CAS in __pthread_mutex_clocklock_common [BZ #28537] +- Avoid extra load with CAS in __pthread_mutex_lock_full [BZ #28537] +- String: Split memcpy tests so that parallel build is faster +- x86: Shrink memcmp-sse4.S code size +- Support C2X printf %b, %B +- Update syscall lists for Linux 5.15 +- s390: Use long branches across object boundaries (jgh instead of jh) + +* Tue Nov 16 2021 Arjun Shankar - 2.34.9000-20 +- Create /{bin,lib,lib64,sbin} as symbolic links in test-container + +* Wed Nov 10 2021 Arjun Shankar - 2.34.9000-19 +- Auto-sync with upstream branch master, + commit 0bd356df1afb0591470499813d4ebae9bcedd6a6: +- Remove the unused +mkdep/+make-deps/s-proto.S/s-proto-cancel.S +- Fix build a chec failures after b05fae4d8e34 +- elf: Use the minimal malloc on tunables_strdup + +* Mon Nov 08 2021 Arjun Shankar - 2.34.9000-18 +- Auto-sync with upstream branch master, + commit db6c4935fae6005d46af413b32aa92f4f6059dce: +- Fix memmove call in vfprintf-internal.c:group_number +- locale: Fix localedata/sort-test undefined behavior +- test-memcpy.c: Double TIMEOUT to (8 * 60) +- hurd: Remove unused __libc_close_range +- hurd: Implement close_range and closefrom +- x86: Double size of ERMS rep_movsb_threshold in dl-cacheinfo.h +- x86: Optimize memmove-vec-unaligned-erms.S +- benchtests: Add partial overlap case in bench-memmove-walk.c +- benchtests: Add additional cases to bench-memcpy.c and bench-memmove.c +- string: Make tests birdirectional test-memcpy.c +- Remove the last trace of generate-md5 [BZ #28554] +- Revert "benchtests: Add acosf function to bench-math" +- Configure GCC with --enable-initfini-array [BZ #27945] +- elf: Earlier missing dynamic segment check in _dl_map_object_from_fd +- gconv: Do not emit spurious NUL character in ISO-2022-JP-3 (bug 28524) +- [powerpc] Tighten contraints for asm constant parameters + +* Wed Nov 03 2021 Patsy Griffin - 2.34.9000-17 +- Auto-sync with upstream branch master, + commit d3bf2f5927d51258a51ac7fde04f4805f8ee294a. +- elf: Do not run DSO sorting if tunables is not enabled +- riscv: Build with -mno-relax if linker does not support R_RISCV_ALIGN +- x86-64: Replace movzx with movzbl +- regex: Unnest nested functions in regcomp.c +- Use Linux 5.15 in build-many-glibcs.py +- elf: Assume disjointed .rela.dyn and .rela.plt for loader +- i386: Explain why __HAVE_64B_ATOMICS has to be 0 +- benchtests: Add hypotf +- benchtests: Make hypot input random +- arm: Use have-mtls-dialect-gnu2 to check for ARM TLS descriptors support +- arm: Use internal symbol for _dl_argv on _dl_start_user +- x86-64: Remove Prefer_AVX2_STRCMP +- x86-64: Improve EVEX strcmp with masked load + +* Fri Oct 29 2021 DJ Delorie - 2.34.9000-16 +- Auto-sync with upstream branch master, + commit 79d0fc65395716c1d95931064c7bf37852203c66. +- benchtests: Add acosf function to bench-math +- benchtests: Improve bench-memcpy-random +- Disable -Waggressive-loop-optimizations warnings in tst-dynarray.c +- Fix compiler issue with mmap_internal +- Check if linker also support -mtls-dialect=gnu2 +- Fix LIBC_PROG_BINUTILS for -fuse-ld=lld +- elf: Disable ifuncmain{1,5,5pic,5pie} when using LLD +- Handle NULL input to malloc_usable_size [BZ #28506] +- x86_64: Add memcmpeq.S to fix disable-multi-arch build +- login: Add back libutil as an empty library +- riscv: Fix incorrect jal with HIDDEN_JUMPTARGET +- x86_64: Add evex optimized __memcmpeq in memcmpeq-evex.S +- x86_64: Add avx2 optimized __memcmpeq in memcmpeq-avx2.S +- x86_64: Add sse2 optimized __memcmpeq in memcmp-sse2.S +- x86_64: Add support for __memcmpeq using sse2, avx2, and evex +- Benchtests: Add benchtests for __memcmpeq +- String: Add __memcmpeq as build target +- NEWS: Add item for __memcmpeq +- String: Add tests for __memcmpeq +- String: Add hidden defs for __memcmpeq() to enable internal usage +- String: Add support for __memcmpeq() ABI on all targets +- configure: Don't check LD -v --help for LIBC_LINKER_FEATURE +- elf: Make global.out depend on reldepmod4.so [BZ #28457] +- x86: Replace sse2 instructions with avx in memcmp-evex-movbe.S +- bench-math: Sort and put each bench per line +- x86_64: Add missing libmvec ABI tests +- elf: Fix e6fd79f379 build with --enable-tunables=no +- elf: Fix slow DSO sorting behavior in dynamic loader (BZ #17645) +- elf: Testing infrastructure for ld.so DSO sorting (BZ #17645) +- iconv: Use TIMEOUTFACTOR for iconv test timeout +- posix: Remove alloca usage for internal fnmatch implementation +- Add alloc_align attribute to memalign et al +- linux: Fix a possibly non-constant expression in _Static_assert +- x86-64: Add sysdeps/x86_64/fpu/Makeconfig + +* Wed Oct 20 2021 Carlos O'Donell - 2.34.9000-15 +- Remove glibc-ld-readonly-revert.patch. +- Auto-sync with upstream branch master, + commit e037274c8ec86ca9d491331984b34f30701b23cf: +- stdlib: Fix tst-canon-bz26341 when the glibc build current working + directory is itself using symlinks. +- powerpc: Remove backtrace implementation +- Correct access attribute on memfrob (bug 28475) +- debug: Add tests for _FORTIFY_SOURCE=3 +- Make sure that the fortified function conditionals are constant +- Don't add access size hints to fortifiable functions +- glibcextract.py: Place un-assemblable @@@ in a comment +- nss: Unnest nested function add_key +- ld.so: Initialize bootstrap_map.l_ld_readonly [BZ #28340] +- timex: Use 64-bit fields on 32-bit TIMESIZE=64 systems (BZ #28469) +- manual: Update _TIME_BITS to clarify it's user defined +- nptl: Fix tst-cancel7 and tst-cancelx7 pidfile race +- elf: Fix elf_get_dynamic_info() for bootstrap +- hurd if_index: Explicitly use AF_INET for if index discovery +- hurd: Fix intr-msg parameter/stack kludge +- x86-64: Add test-vector-abi.h/test-vector-abi-sincos.h +- elf: Fix dynamic-link.h usage on rtld.c + +* Thu Oct 14 2021 Arjun Shankar - 2.34.9000-14 +- Adjust glibc-ld-readonly-revert.patch. +- Auto-sync with upstream branch master, + commit e59ced238482fd71f3e493717f14f6507346741e: +- x86: Optimize memset-vec-unaligned-erms.S +- x86: Optimize memcmp-evex-movbe.S for frontend behavior and size +- libio: Update tst-wfile-sync to not depend on stdin +- elf: Update audit tests to not depend on stdout +- elf: Fix elf_get_dynamic_info definition +- Add TEST_COMPARE_STRING_WIDE to support/check.h +- Fix nios2 localplt failure +- elf: Remove Intel MPX support (lazy PLT, ld.so profile, and LD_AUDIT) +- resolv: Avoid GCC 12 false positive warning [BZ #28439]. +- benchtests: Add medium cases and increase iters in bench-memset.c +- x86: Modify ENTRY in sysdep.h so that p2align can be specified +- resolv: make res_randomid use random_bits() +- Linux: implement getloadavg(3) using sysinfo(2) +- Remove unreliable parts of rt/tst-cpuclock2 +- elf: Avoid nested functions in the loader [BZ #27220] +- Add run-time check for indirect external access +- Initial support for GNU_PROPERTY_1_NEEDED +- io: Fix ftw internal realloc buffer (BZ #28126) +- Fix subscript error with odd TZif file [BZ #28338] + +* Thu Oct 07 2021 Carlos O'Donell - 2.34.9000-13 +- Auto-sync with upstream branch master, + commit f2e06656d04a9fcb0603802a4f8ce7aa3a1f055e: +- S390: Add PCI_MIO and SIE HWCAPs +- support: Also return fd when it is 0 +- ld.so: Don't fill the DT_DEBUG entry in ld.so [BZ #28129] +- S390: update libm test ulps +- powerpc: update libm test ulps +- math: Also xfail the new j0f tests for ibm128-libgcc +- y2038: Use a common definition for stat for sparc32 +- Fix stdlib/tst-setcontext.c for GCC 12 -Warray-compare +- aarch64: update libm test ulps +- Fixed inaccuracy of j0f (BZ #28185) +- Fix stdio-common tests for GCC 12 -Waddress +- benchtests: Building benchmarks as static executables +- elf: Avoid deadlock between pthread_create and ctors [BZ #28357] +- time: Ignore interval nanoseconds on tst-itimer +- io: Do not skip timestamps tests for 32-bit time_t +- Update to Unicode 14.0.0 [BZ #28390] + +* Fri Oct 01 2021 Florian Weimer - 2.34.9000-12 +- Auto-sync with upstream branch master, + commit eae81d70574e923ce3c59078b8df857ae192efa6: +- nptl: pthread_kill must send signals to a specific thread [BZ #28407] +- support: Add check for TID zero in support_wait_for_thread_exit +- nptl: Add CLOCK_MONOTONIC support for PI mutexes +- support: Add support_mutex_pi_monotonic +- nptl: Use FUTEX_LOCK_PI2 when available +- Linux: Add FUTEX_LOCK_PI2 +- Add C2X _PRINTF_NAN_LEN_MAX +- Add exp10 macro to (bug 26108) +- elf: Replace nsid with args.nsid [BZ #27609] +- Add missing braces to bsearch inline implementation [BZ #28400] +- Update alpha libm-test-ulps +- Suppress -Wcast-qual warnings in bsearch +- elf: Copy l_addr/l_ld when adding ld.so to a new namespace +- powerpc: Fix unrecognized instruction errors with recent binutils + +* Wed Sep 29 2021 Florian Weimer - 2.34.9000-11 +- Drop glibc-rh1992702-*.patch, applied upstream. +- Auto-sync with upstream branch master, + commit 9bd9978639c2f75dbea5c25226264b1ac11fdf05: +- Do not declare fmax, fmin _FloatN, _FloatNx versions for C2X +- Do not define tgmath.h fmaxmag, fminmag macros for C2X (bug 28397) +- Add fmaximum, fminimum functions +- Linux: Simplify __opensock and fix race condition [BZ #28353] +- pthread/tst-cancel28: Fix barrier re-init race condition +- powerpc: Delete unneeded ELF_MACHINE_BEFORE_RTLD_RELOC +- posix: Remove spawni.c +- Disable symbol hack in libc_nonshared.a +- linux: Revert the use of sched_getaffinity on get_nproc (BZ #28310) +- linux: Simplify get_nprocs +- misc: Add __get_nprocs_sched +- htl: Fix sigset of main thread +- htl: make pthread_sigstate read/write set/oset outside sigstate section +- Avoid warning: overriding recipe for .../tst-ro-dynamic-mod.so +- benchtests: Improve reliability of memcmp benchmarks +- Define __STDC_IEC_60559_BFP__ and __STDC_IEC_60559_COMPLEX__ +- build-many-glibcs.py: add powerpc64le glibc variant without multiarch +- Fix sysdeps/x86/fpu/s_ffma.c for 32-bit FMA processor case +- Linux: Avoid closing -1 on failure in __closefrom_fallback +- i386: Port elf_machine_{load_address,dynamic} from x86-64 +- aarch64: Disable A64FX memcpy/memmove BTI unconditionally +- xsysconf: Only fail on error results and errno set +- powerpc64le: Avoid conflicting types for f64xfmaf128 when IFUNC is not used +- Fix ffma use of round-to-odd on x86 +- vfprintf: Unify argument handling in process_arg +- vfprintf: Handle floating-point cases outside of process_arg macro + +* Thu Sep 23 2021 Florian Weimer - 2.34.9000-10 +- Fix ppc64le build failure by reverting DL_RO_DYN_SECTION removal + +* Thu Sep 23 2021 Florian Weimer - 2.34.9000-9 +- Auto-sync with upstream branch master, + commit 2849e2f53311b66853cb5159b64cba2bddbfb854: +- nptl: Avoid setxid deadlock with blocked signals in thread exit [BZ #28361] +- Add narrowing fma functions +- ld.so: Replace DL_RO_DYN_SECTION with dl_relocate_ld [BZ #28340] +- Adjust new narrowing div/mul tests for IBM long double, update powerpc ULPs +- Mention today's regex merge in SHARED-FILES +- Fix f64xdivf128, f64xmulf128 spurious underflows (bug 28358) +- regex: copy back from Gnulib +- nptl: Fix type of pthread_mutexattr_getrobust_np, + pthread_mutexattr_setrobust_np (bug 28036) +- powerpc: Fix unrecognized instruction errors with recent GCC +- elf: Include in elf/dl-debug-symbols.S +- nptl: pthread_kill needs to return ESRCH for old programs (bug 19193) +- Extend struct r_debug to support multiple namespaces [BZ #15971] +- Use $(pie-default) with conformtest +- Run conform/ tests using newly built libc +- posix: Fix attribute access mode on getcwd [BZ #27476] +- Fix build-many-glibcs.py --strip for installed library renaming +- benchtests: Fix validate_benchout.py exceptions +- elf: Remove THREAD_GSCOPE_IN_TCB +- htl: Reimplement GSCOPE +- htl: Move thread table to ld.so +- Redirect fma calls to __fma in libm +- time: Fix compile error in itimer test affecting hurd + +* Wed Sep 15 2021 Florian Weimer - 2.34.9000-8 +- Use system CPU count for sysconf(_SC_NPROCESSORS_*) (#1992702) + +* Wed Sep 15 2021 Florian Weimer - 2.34.9000-7 +- Auto-sync with upstream branch master, + commit 2444ce5421c6036a503842d8dd8d93c27aad59ee: +- mach lll_lock/unlock: Explicitly request private locking +- elf: Replace most uses of THREAD_GSCOPE_IN_TCB +- Add MADV_POPULATE_READ and MADV_POPULATE_WRITE from Linux 5.14 to + bits/mman-linux.h +- Update kernel version to 5.14 in tst-mman-consts.py +- configure: Fix check for INSERT in linker script +- iconvconfig: Fix behaviour with --prefix [BZ #28199] +- nptl: Fix race between pthread_kill and thread exit (bug 12889) +- nptl: pthread_kill, pthread_cancel should not fail after exit (bug 19193) +- benchtests: Remove redundant assert.h +- benchtests: Enable scripts/plot_strings.py to read stdin +- Add narrowing square root functions +- _Static_assert needs two arguments for compatibility with GCC before 9 +- testrun.sh: Add support for --tool=rpctrace + +* Thu Sep 09 2021 Patsy Griffin - 2.34.9000-6 +- Auto-sync with upstream branch master, + commit 89dc0372a9055e7ef86fe19be6201fa0b16b2f0e. +- Update syscall lists for Linux 5.14 +- Fix failing nss/tst-nss-files-hosts-long with local resolver +- MIPS: Setup errno for {f,l,}xstat +- Use Linux 5.14 in build-many-glibcs.py +- locale: Add missing second argument to _Static_assert in C-collate-seq.c +- Update hppa libm-test-ulps +- Add generic C.UTF-8 locale (Bug 17318) +- Add 'codepoint_collation' support for LC_COLLATE. +- AArch64: Update A64FX memset not to degrade at 16KB +- Revert "AArch64: Update A64FX memset not to degrade at 16KB" +- Remove "Contributed by" lines +- Port shared code information from the wiki +- AArch64: Update A64FX memset not to degrade at 16KB +- posix: remove some iso-8859-encoded characters +- configure: Allow LD to be LLD 13.0.0 or above [BZ #26558] +- hurd msync: Drop bogus test +- hurd: Fix typo in msync + +* Tue Aug 31 2021 Florian Weimer - 2.34.9000-5 +- Auto-sync with upstream branch master, + commit 3c8b9879cab6d41787bc5b14c1748f62fd6d0e5f: +- x86-64: Use testl to check __x86_string_control +- i686: Don't include multiarch memove in libc.a +- support: Add support_wait_for_thread_exit +- Allow #pragma GCC in headers in conformtest +- nptl: Fix tst-cancel7 and tst-cancelx7 race condition (BZ #14232) +- Use support_open_dev_null_range io/tst-closefrom, + misc/tst-close_range, and posix/tst-spawn5 (BZ #28260) +- support: Add support_open_dev_null_range +- llio.texi: Wording fixes in description of closefrom() +- Fix error message in memmove test to display correct src pointer + +* Wed Aug 25 2021 Arjun Shankar - 2.34.9000-4 +- Auto-sync with upstream branch master, + commit 9926f6e2eeb374cf729d4bb3f092dd4b36a8f861: +- elf: Skip tst-auditlogmod-* if the linker doesn't support --depaudit [BZ #28151] +- powerpc: Use --no-tls-get-addr-optimize in test only if the linker supports it +- x86-64: Remove assembler AVX512DQ check +- x86-64: Remove compiler -mavx512f check +- Use __executable_start as the lowest address for profiling [BZ #28153] +- hurd: Fix errlist error mapping +- hurd: Remove old test-err_np.c file +- Fix iconv build with GCC mainline +- rtld: copy terminating null in tunables_strdup (bug 28256) +- mtrace: Fix output with PIE and ASLR [BZ #22716] +- x86-64: Optimize load of all bits set into ZMM register [BZ #28252] +- Update string/test-memmove.c to cover 16KB copy +- elf: Fix missing colon in LD_SHOW_AUXV output [BZ #28253] +- x86: fix Autoconf caching of instruction support checks [BZ #27991] +- arm: Simplify elf_machine_{load_address,dynamic} +- riscv: Drop reliance on _GLOBAL_OFFSET_TABLE_[0] +- Remove sysdeps/*/tls-macros.h + +* Tue Aug 24 2021 Siddhesh Poyarekar - 2.34.9000-3 +- Disable dependencies and linking for libc_malloc_debug.so (#1985048). + +* Tue Aug 17 2021 DJ Delorie - 2.34.9000-2 +- Auto-sync with upstream branch master, + commit b37b75d269883a2c553bb7019a813094eb4e2dd1. +- x86_64: Simplify elf_machine_{load_address,dynamic} +- elf: Drop elf/tls-macros.h in favor of __thread and tls_model attributes [BZ #28152] [BZ #28205] +- hurd: Drop fmh kludge +- time: Fix overflow itimer tests on 32-bit systems +- mips: increase stack alignment in clone to match the ABI +- mips: align stack in clone [BZ #28223] + + +* Thu Aug 12 2021 Arjun Shankar - 2.34.9000-1 +- Auto-sync with upstream branch master, + commit 4cc79c217744743077bf7a0ec5e0a4318f1e6641: +- librt: add test (bug 28213) +- mtrace: Use a static buffer for printing [BZ #25947] +- hurd mmap: Reduce the requested max vmprot +- hurd mmap: Factorize MAP_SHARED flag check +- aarch64: Make elf_machine_{load_address,dynamic} robust [BZ #28203] +- elf: Unconditionally use __ehdr_start +- hurd: Add support for AT_NO_AUTOMOUNT +- [5/5] AArch64: Improve A64FX memset medium loops +- [4/5] AArch64: Improve A64FX memset by removing unroll32 +- [3/5] AArch64: Improve A64FX memset for remaining bytes +- [2/5] AArch64: Improve A64FX memset for large sizes +- [1/5] AArch64: Improve A64FX memset for small sizes +- Use binutils 2.37 branch in build-many-glibcs.py +- Add PTRACE_GET_RSEQ_CONFIGURATION from Linux 5.13 to sys/ptrace.h +- librt: fix NULL pointer dereference (bug 28213) +- powerpc64: Add checks for Altivec and VSX in ifunc selection +- powerpc64: Check cacheline size before using optimised memset routines +- powerpc64: Replace some PPC_FEATURE_HAS_VSX with PPC_FEATURE_ARCH_2_06 +- Linux: Fix fcntl, ioctl, prctl redirects for _TIME_BITS=64 (bug 28182) +- Add INADDR_DUMMY from Linux 5.13 to netinet/in.h +- tst-mxfast: Don't run with mcheck +- rt: Set the correct message queue for tst-mqueue10 +- Update sparc libm-test-ulps +- linux: Add sparck brk implementation +- test-dlclose-exit-race: avoid hang on pthread_create error +- gethosts: Remove unused argument _type +- hurd: Avoid spurious warning +- gaiconf_init: Avoid double-free in label and precedence lists +- copy_and_spawn_sgid: Avoid double calls to close() +- iconv_charmap: Close output file when done +- gconv_parseconfdir: Fix memory leak +- ldconfig: avoid leak on empty paths in config file +- Fix build of nptl/tst-thread_local1.cc with GCC 12 +- nis: Fix leak on realloc failure in nis_getnames [BZ #28150] +- Remove obsolete comments/name from several benchtest input files. +- Remove obsolete comments/name from acos-inputs, since slow path was removed. +- Open master branch for glibc 2.35 development + +* Mon Aug 2 2021 Florian Weimer - 2.34-1 +- Switch to glibc 2.34 release tarball: +- Update ChangeLog.old/ChangeLog.23. +- Prepare for glibc 2.34 release. +- po/nl.po: Update Dutch translation. +- Update install.texi, and regenerate INSTALL. +- Update translations. +- Update NEWS. +- NEWS: Fix typos, grammar, and missing words +- elf: Fix audit regression + +* Fri Jul 30 2021 Siddhesh Poyarekar - 2.33.9000-56 +- Port post scripts for gconv-extra to lua and drop dependency across + architectures for x86 multilib (#1988344). + +* Thu Jul 29 2021 Florian Weimer - 2.33.9000-55 +- Auto-sync with upstream branch master, + commit c37fc3ebf0607ce1953c565ffe56d56555eeb25e: +- Update libc.pot for 2.34 release. +- x86-64: Add Avoid_Short_Distance_REP_MOVSB +- Typo: Rename HAVE_CLONE3_WAPPER to HAVE_CLONE3_WRAPPER +- build-many-glibcs.py: Add x86_64-linux-gnu-minimal configuration +- tests: use xmalloc to allocate implementation array +- xmalloc: Fix warnings with gcc analyzer +- __cxa_thread_atexit_impl: Abort on allocation failure [BZ #18524] +- manual: Drop the .so suffix in libc_malloc_debug description +- hurd: _Fork: unlock malloc before calling fork child hooks + +* Tue Jul 27 2021 Florian Weimer - 2.33.9000-54 +- Revert to old C.UTF-8 locale + +* Mon Jul 26 2021 Siddhesh Poyarekar - 2.33.9000-53 +- Loosen dependency on glibc-gconv-extra (#1812191). + +* Mon Jul 26 2021 Florian Weimer - 2.33.9000-52 +- Switch to new version of C.UTF-8 locale + +* Mon Jul 26 2021 Florian Weimer - 2.33.9000-51 +- Auto-sync with upstream branch master, + commit ddcc612ce923038b867083a0c55d6e034951155a: +- Exclude static tests for mcheck and malloc-check +- i386: Regenerate ulps + +* Sat Jul 24 2021 Florian Weimer - 2.33.9000-50 +- Auto-sync with upstream branch master, + commit d34ed66f96fa9316654d7adb2afcce4be1d1c4f5: +- manual: Document unsupported cases for interposition +- x86: Install [BZ #27958] +- Fix build and tests with --disable-tunables + +* Sat Jul 24 2021 Florian Weimer - 2.33.9000-49 +- Remove both old and new library names in glibc-hwcaps removal (#1983677) + +* Fri Jul 23 2021 Florian Weimer - 2.33.9000-48 +- Auto-sync with upstream branch master, + commit 9a7ab0769b295cbf5232140401742a8f34bda3de: +- hurd: Fix glob lstat compatibility +- socket: Add time64 alias for setsockopt +- socket: Add time64 alias for getsockopt +- mcheck Fix malloc_usable_size [BZ #22057] +- Remove malloc hooks [BZ #23328] +- Move malloc_{g,s}et_state to libc_malloc_debug +- glibc.malloc.check: Wean away from malloc hooks +- mtrace: Wean away from malloc hooks +- Simplify __malloc_initialized +- mcheck: Wean away from malloc hooks [BZ #23489] +- Move malloc hooks into a compat DSO +- Remove __morecore and __default_morecore +- Remove __after_morecore_hook +- Make mcheck tests conditional on GLIBC_2.23 or earlier +- posix: Add sysconf(_SC_{MIN,}SIGSTKSZ) support +- malloc: Fix tst-mallocfork3-malloc-check link +- ARC: elf: make type safe +- ARC: fp: (micro)optimize FPU_STATUS read by eliding FWE bit clearing + +* Thu Jul 22 2021 Fedora Release Engineering - 2.33.9000-47 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_35_Mass_Rebuild + +* Wed Jul 21 2021 Florian Weimer - 2.33.9000-46 +- Auto-sync with upstream branch master, + commit 77ede5f010f1b144e067ec035e422a13bb57c55d: +- socket: Add time64 alias for sendmsg +- socket: Add time64 alias for recvmsg +- socket: Add time64 alias for sendmmsg +- Linux: Add time64 alias for prctl +- io: Add time64 alias for fcntl +- misc: Add time64 alias for ioctl +- RISC-V: Update rv64 ULPs +- resolv: Do not install libnss_dns.a, libnss_dns.so +- hurd: Add support for spawn_do_closefrom +- elf: Fix tst-cpu-features-cpuinfo on some AMD systems (BZ #28090) +- i386: Add the clone3 wrapper + +* Mon Jul 19 2021 Florian Weimer - 2.33.9000-45 +- Remove glibc-hwcaps multilibs on upgrade (#1983677) + +* Mon Jul 19 2021 Florian Weimer - 2.33.9000-44 +- Auto-sync with upstream branch master, + commit ee5ed99922ca90bcea4a2f9a48a0c9ae4b534ece: +- nss: Directly load nss_dns, without going through dlsym/dlopen +- resolv: Move nss_dns into libc +- resolv: Move res_query functions into libc +- resolv: Move res_mkquery, res_nmkquery into libc +- resolv: Move res_send, res_nsend into libc +- resolv: Move res_hostalias into its own file, along with hostalias +- resolv: Move __res_context_hostalias into its own file and into libc +- resolv: Move res_queriesmatch to its own file and into libc +- resolv: Move res_nameinquery to its own file and into libc +- resolv: Move ns_samename into its own file, and into libc +- resolv: Move ns_makecanon into its own file, and into libc +- resolv: Move res_isourserver to its own file and reformat to GNU style +- resolv: Move __res_get_nsaddr to its own file and into libc +- resolv: Rename res_comp.c to res-name-checking.c and move into libc +- resolv: Move dn_skipname to its own file and into libc +- resolv: Move dn_comp to its own file and into libc +- resolv: Move _getlong, _getshort, __putlong, __putshort to res-putget +- resolv: Move dn_expand to its own file and into libc +- resolv: Move ns_name_compress into its own file and into libc +- resolv: Move ns_name_pack into its own file and into libc +- resolv: Move ns_name_pton into its own file and into libc +- resolv: Move ns_name_uncompress into its own file and into libc +- resolv: Move ns_name_skip to its own file and into libc (bug 28091) +- resolv: Deprecate legacy interfaces in libresolv +- tst-safe-linking: make false positives even more improbable +- htl: Do not expose pthread hidden proto outside libpthread +- elf: Fix a wrong array access on tst-tls20 +- elf: Add -Wl,--no-as-needed for tst-tls-manydynamic*mod-dep-bad.so (BZ #28089) +- resolv: Move ns_name_unpack to its own file and into libc +- resolv: Remove unnecessary res_isourserver_p call from send_dg +- resolv: Move ns_name_ntop to its own file and into libc +- nss_dns: Do not use deprecated packet parsing functions +- resolv: Sort Makefile routines and Versions lexicographically +- socket: Add hidden prototype for setsockopt +- elf: Fix DTV gap reuse logic (BZ #27135) +- Fix linknamespace errors and local-plt-usages in nss_files. +- Add static tests for __clone_internal +- x86-64: Add the clone3 wrapper +- Add an internal wrapper for clone, clone2 and clone3 +- nss: Fix build error with --disable-nscd +- htl: Fix linking static examples against libpthread +- htl: Let libc call __pthread_mutex_{,try,un}lock +- posix: Ignore non opened files on tst-spawn5 +- mcheck: Align struct hdr to MALLOC_ALIGNMENT bytes [BZ #28068] + +* Thu Jul 15 2021 Siddhesh Poyarekar - 2.33.9000-43 +- Run postun only if the main gconv-modules configuration file exists. + (#1981013) +- Own %{libdir}/gconv/gconv-modules.d +- Rearrange file list command so that gconv-modules.cache is no longer + marked as %config + +* Tue Jul 13 2021 Florian Weimer - 2.33.9000-42 +- Perform systemd re-exec even if glibc.i686 is installed + +* Tue Jul 13 2021 Florian Weimer - 2.33.9000-41 +- Re-exec systemd on upgrades + +* Mon Jul 12 2021 Florian Weimer - 2.33.9000-40 +- Merge files NSS service module into libc. +- Restore vDSO acceleration for time functions on older kernels for i686. +- Auto-sync with upstream branch master, + commit 72e84d1db22203e01a43268de71ea8669eca2863: +- Linux: Use 32-bit vDSO for clock_gettime, gettimeofday, time (BZ# 28071) +- Reduce pollution due to dynamic PTHREAD_STACK_MIN +- Fix failing nss/tst-nss-files-hosts-long. +- nis: nis_local_group may read from __nisgroup[-1] (bug 28075) +- hurd _Fork: Drop duplicate malloc_fork_lock calls +- support: Replace _SC_MINSIGSTKSZ with _SC_SIGSTKSZ +- support: Replace MINSIGSTKSZ with sysconf (_SC_MINSIGSTKSZ) +- Define PTHREAD_STACK_MIN to sysconf(_SC_THREAD_STACK_MIN) +- Force building with -fno-common +- Add a generic malloc test for MALLOC_ALIGNMENT +- Properly run tst-spawn5 directly [BZ #28067] +- build-many-glibcs.py: Add glibcs-arm-linux-gnueabihf-thumb +- nptl: Use out-of-line wake function in __libc_lock_unlock slow path +- powerpc64le: Fix typo in configure +- powerpc64: Remove strcspn ifunc from the loader +- x86: Remove wcsnlen-sse4_1 from wcslen ifunc-impl-list [BZ #28064] +- x86-64: Test strlen and wcslen with 0 in the RSI register [BZ #28064] +- x86_64: Remove unneeded static PIE check for undefined weak diagnostic +- Add NT_ARM_PAC_ENABLED_KEYS to elf.h +- posix: Add posix_spawn_file_actions_addclosefrom_np +- io: Add closefrom [BZ #10353] +- linux: Add close_range +- support: Add support_stack_alloc +- _int_realloc is static +- Move mcheck symbol from stdlib to malloc +- nss: Do not install static linker input files for libnss_files +- elf/tests: Make thrlock and noload depend on libm +- Harden tcache double-free check +- nss: Access nss_files through direct references +- nss_files: Move into libc +- nss_files: Add generic code for set*ent, end*ent and file open +- nss_files: Allocate nscd file registration data on the heap +- libio: Replace internal _IO_getdelim symbol with __getdelim +- libio: Add hidden prototype for ungetc +- inet: Add hidden prototype for __inet_network +- Update MIPS libm-test-ulps +- Update powerpc-nofpu libm-test-ulps +- soft-fp: Add __extendhfsf2/__extendhfdf2, __truncsfhf2/__truncdfhf2, + __eqhf2/__nehf2 +- Update kernel version to 5.13 in tst-mman-consts.py +- tests-exclude-mcheck: Fix typo + +* Wed Jul 07 2021 Florian Weimer - 2.33.9000-39 +- Auto-sync with upstream branch master, + commit 7a5db2e82fbb6c3a6e3fdae02b7166c5d0e8c7a8: +- elf: Clean up GLIBC_PRIVATE exports of internal libdl symbols +- nptl: Reduce the GLIBC_PRIVATE ABI +- nptl: Remove GLIBC_2.34 versions of __pthread_mutex_lock, + __pthread_mutex_unlock +- nptl: Use internal low-level lock type for !IS_IN (libc) +- glibc.malloc.check: Fix nit in documentation +- Exclude tst-realloc from tests-mcheck +- linux: Fix setsockopt fallback +- linux: Use the expected size for SO_TIMESTAMP{NS} convertion +- linux: Consolidate Linux setsockopt implementation +- linux: Consolidate Linux getsockopt implementation +- manual: fix description for preadv() + +* Tue Jul 06 2021 Florian Weimer - 2.33.9000-38 +- Auto-sync with upstream branch master, + commit 832f50be6c9c010e46180d14126bbb81f35e808c: +- linux: Check for null value msghdr struct before use +- elf: Call free from base namespace on error in dl-libc.c [BZ #27646] + +* Mon Jul 05 2021 Florian Weimer - 2.33.9000-37 +- Move libanl.a into glibc-devel. +- Auto-sync with upstream branch master, + commit 91fb0f17a5779da6e7877eb74119a83dbe8bf167: +- hooks.c: Remove incorrect comment +- mtrace: Add attribute nocommon to mallwatch +- Move glibc.malloc.check implementation into its own file +- mtrace: Deprecate mallwatch and tr_break +- Drop source dependencies on hooks.c and arena.c +- malloc: Initiate tcache shutdown even without allocations [BZ #28028] +- Add mcheck tests to malloc +- iconvconfig: Use the public feof_unlocked +- resolv: Move libanl into libc (if libpthread is in libc) +- powerpc: optimize strcpy/stpcpy for POWER9/10 +- soft-fp: Add __extendhfxf2 and __truncxfhf2 +- x86: Check RTM_ALWAYS_ABORT for RTM [BZ #28033] +- Update syscall lists for Linux 5.13 +- s390: Fix MEMCHR_Z900_G5 ifunc-variant if n>=0x80000000 [BZ #28024] +- Fix extra PLT reference in libc.so due to __glob64_time64 if build with gcc 7.5 on 32bit. +- AArch64: Add hp-timing.h +- AArch64: Improve strnlen performance + +* Wed Jun 30 2021 Florian Weimer - 2.33.9000-36 +- Auto-sync with upstream branch master, + commit eb68d7d23cc411acdf68a60f194343a6774d6194: +- Linux: Avoid calling malloc indirectly from __get_nprocs (#1975693) +- Use Linux 5.13 in build-many-glibcs.py + +* Wed Jun 30 2021 Florian Weimer - 2.33.9000-35 +- Drop glibc-s390x-roundeven.patch, applied upstream. +- Move libutil.a into glibc-devel. +- Auto-sync with upstream branch master, + commit 734c60ebb607086ad6d67b2544d6b7baba72a652: +- login: Move libutil into libc +- login: Rework hidden prototypes for __setutent, __utmpname, __endutent +- login: Hidden prototypes for _getpt, __ptsname_r, grantpt, unlockpt +- nptl_db: Re-use the ELF-to-abilist converter for ABI checking +- Add RFC 8335 Definitions from Linux 5.13 +- nss: Fix NSS_DECLARE_MODULE_FUNCTIONS handling of _nss_*_endnetgrent +- s390x: Update math: redirect roundeven function +- posix: Add _Fork [BZ #4737] + +* Mon Jun 28 2021 Florian Weimer - 2.33.9000-34 +- Move librt.a to glibc-devel (#1977058) + +* Mon Jun 28 2021 Florian Weimer - 2.33.9000-33 +- Dropped patches glibc-nosymlink-*.patch, glibc-iconvconfig-corruption.patch, + glibc-libthread_db-dynsym-*.patch; applied upstream. +- Auto-sync with upstream branch master, + commit dd45734e322a03287d34d8af9b7da7b35cfddb8e: +- nptl: Add glibc.pthread.stack_cache_size tunable +- nptl: Export libthread_db-used symbols under GLIBC_PRIVATE +- nptl: Rename nptl_version to __nptl_version +- nptl_db: Clean up main/rtld variable handling +- arm: align stack in clone [BZ 28020] +- Linux: Cleanups after librt move +- Linux: Move timer_settime, __timer_settime64 from librt to libc +- Linux: Move timer_gettime, __timer_gettime64 from librt to libc +- Linux: Move timer_getoverrun from librt to libc +- Linux: Move timer_create, timer_delete from librt to libc +- Linux: Define TIMER_T_WAS_INT_COMPAT in kernel-posix-timers.h +- Install shared objects under their ABI names +- elf: Generalize name-based DSO recognition in ldconfig +- Makerules: Remove lib-version, $(subdir-version) +- nptl_db: Install libthread_db under a regular implementation name +- iconvconfig: Fix multiple issues +- wordexp: handle overflow in positional parameter number (bug 28011) +- Update math: redirect roundeven function +- Use GCC builtins for roundeven functions if desired. +- x86_64: roundeven with sse4.1 support +- math: redirect roundeven function + +* Mon Jun 28 2021 Florian Weimer - 2.33.9000-32 +- Switch to new version of libthread_db .dynsym patch + +* Mon Jun 28 2021 Florian Weimer - 2.33.9000-31 +- Further .symtab adjustment: Keep all __GI_* symbols (#1975895) + +* Mon Jun 28 2021 Florian Weimer - 2.33.9000-30 +- Keep most of .symtab in libc.so.6 (#1975895) + +* Sun Jun 27 2021 Florian Weimer - 2.33.9000-29 +- Apply emergency patch to fix iconvconfig corruption + +* Sun Jun 27 2021 Florian Weimer - 2.33.9000-28 +- Adjust glibc.req so that egrep does not cause eu-readelf to fail + +* Sun Jun 27 2021 Florian Weimer - 2.33.9000-27 +- Drop glibc-revert-dtv-gap-reuse.patch, applied upstream. +- Auto-sync with upstream branch master, + commit 2c16cb88a6e5ace0fb7cedca86860ea7bde522a7: +- Linux: Move timer helper routines from librt to libc +- Linux: Move mq_unlink from librt to libc +- Linux: Move mq_send, mq_timedsend, __mq_timedsend_time64 to libc +- Linux: Move mq_receive, mq_timedreceive, __mq_timedreceive_time64 to libc +- Linux: Move mq_open, __mq_open_2 from librt to libc +- Linux: Move mq_notify from librt to libc +- Linux: Move mq_getattr from librt to libc +- Linux: Move mq_setattr from librt to libc +- Linux: Move mq_close from librt to libc +- Linux: Move lio_listio, lio_listio64 from librt to libc +- rt: Rework lio_listio implementation +- Linux: Move aio_write, aio_write64 into libc +- Linux: Move aio_suspend, aio_suspend64, __aio_suspend_time64 to libc +- Linux: Move aio_return, aio_return64 into libc +- Linux: Move aio_read, aio_read64 into libc +- Linux: Move aio_fsync, aio_fsync64 into libc +- Linux: Move aio_error, aio_error64 into libc +- Linux: Move aio_cancel, aio_cancel64 into libc +- Linux: Move aio_init from librt into libc +- support: Fix xclone build failures on ia64 and hppa +- elf: Disable most of TLS modid gaps processing [BZ #27135] +- elf: Fix glibc-hwcaps priorities with cache flags mismatches [BZ #27046] +- * NEWS: Clarify _TIME_BITS change. +- x86: Remove unnecessary overflow check from wcsnlen-sse4_1.S +- String: Add three more overflow tests cases to test-strnlen.c +- Consolidate pthread_atfork +- posix: Do not clobber errno by atfork handlers +- posix: Consolidate fork implementation +- support: Add xclone +- x86: Fix tst-cpu-features-cpuinfo on Ryzen 9 (BZ #27873) +- x86: Copy IBT and SHSTK usable only if CET is enabled +- x86: Fix overflow bug in wcsnlen-sse4_1 and wcsnlen-avx2 [BZ #27974] +- x86: Fix overflow bug with wmemchr-sse2 and wmemchr-avx2 [BZ #27974] +- String: Add overflow tests for strnlen, memchr, and strncat [BZ #27974] +- x86-64: Add wcslen optimize for sse4.1 +- x86-64: Move strlen.S to multiarch/strlen-vec.S +- hurd: Fix build after 52a5fe70a2 +- nptl: Use SA_RESTART for SIGCANCEL handler +- doc: _TIME_BITS defaults may change +- More mcheck -> malloc-check refactoring +- Add NEWS item for gconv-modules.d change +- Handle DT_UNKNOWN in gconv-modules.d +- iconvconfig: Use common gconv module parsing function +- gconv_conf: Split out configuration file processing +- gconv_conf: Remove unused variables +- iconv: Remove alloca use in gconv-modules configuration parsing +- Remove unsused symbols from nptl/Versions +- linux: Only use 64-bit syscall if required for clock_nanosleep +- linux: Only use 64-bit syscall if required for internal futex +- linux: Only use 64-bit syscall if required for utimensat family +- linux: Only use 64-bit syscall if required for sigtimedwait +- linux: Only use 64-bit syscall if required for mq_timedsend +- linux: Only use 64-bit syscall if required for mq_timedreceive +- linux: Only use 64-bit syscall if required for timerfd_settime +- linux: Only use 64-bit syscall if required for semtimedop +- linux: timerfd_gettime minor cleanup +- linux: Remove time64-support +- linux: Remove supports_time64 () from clock_gettime +- linux: Remove supports_time64 () from clock_getres +- linux: Only use 64-bit syscall if required for select +- linux: Only use 64-bit syscall if required for pselect +- linux: Only use 64-bit syscall if required for ppoll +- support: Add support_create_timer +- Use 64 bit time_t stat internally +- malloc: Drop __malloc_initialized from Versions +- tst-mcheck: Rename to tst-malloc-check +- Add hidden prototypes for fsync, fdatasync +- nptl: Move pthreadP.h into sysdeps directory +- rt: Move generic implementation from sysdeps/pthread to rt +- rt: Move shm_unlink into libc +- rt: Move shm_open into libc +- rt: Replace generic stub of shm_unlink with the posix version +- rt: Replace generic stub of shm_open with the posix version +- Fix librt-routines-var issues for !PTHREAD_IN_LIBC +- rt: Lexicographically sort Versions file; librt-routines in Makefile +- elf: Use _dl_catch_error from base namespace in dl-libc.c [BZ #27646] +- Makeconfig: Fix time64-compat.mk target + +* Sun Jun 27 2021 Florian Weimer - 2.33.9000-26 +- Add automatic requires if building against glibc development snapshots + +* Thu Jun 24 2021 Carlos O'Donell - 2.33.9000-25 +- Fix thread local storage corruption (#1974970) + +* Tue Jun 22 2021 Siddhesh Poyarekar - 2.33.9000-24 +- Strengthen dependency on glibc-gconv-extra. + +* Fri Jun 18 2021 Florian Weimer - 2.33.9000-23 +- Make glibc-all-langpacks require glibc-gconv-extra in buildroots (#1973663) + +* Thu Jun 17 2021 Florian Weimer - 2.33.9000-22 +- Export libthread_db symbols under GLBIC_PRIVATE (#1965374) + +* Thu Jun 17 2021 Florian Weimer - 2.33.9000-21 +- Redo the crafted libc.so.6 symbol table for valgrind (#1965374) + +* Thu Jun 17 2021 Florian Weimer - 2.33.9000-20 +- Remove .symtab from libc.so.6 again (#1965374) + +* Thu Jun 17 2021 Florian Weimer - 2.33.9000-19 +- Drop glibc-rh697421.patch: The ISO-10646-UCS-2// alias for UTF-8 is incorrect + (#1972520) + +* Wed Jun 16 2021 Florian Weimer - 2.33.9000-18 +- Rebuild against rawhide gcc + +* Tue Jun 15 2021 Florian Weimer - 2.33.9000-17 +- Dropped glibc-gconv-modules-revert.patch, applied upstream. +- Auto-sync with upstream branch master, + commit aa9a7f629632c9180de89632d4f2c1e6039f7781: +- nptl: Export _pthread_cleanup_push, _pthread_cleanup_pop again +- s390x: Align child stack while clone. [BZ #27968] +- y2038: Add test coverage +- libsupport: Add 64-bit time_t support for stat functions +- libsupport: Add 64-bit time_t support for time functions +- io: Add ftw64 with 64-bit time_t support +- io: Add fts64 with 64-bit time_t support +- posix: Add glob64 with 64-bit time_t support +- y2038: Add support for 64-bit time on legacy ABIs +- time: Add 64-bit time support for getdate +- y2038: Add __USE_TIME_BITS64 support for socket-constants.h +- y2038: Use a common definition for shmid_ds +- y2038: Use a common definition for semid_ds +- y2038: Use a common definition for msqid_ds +- y2038: Use a common definition for stat +- y2038: linux: Add __USE_TIME_BITS64 support for struct timex +- y2038: Add __USE_TIME_BITS64 support for struct utimbuf +- y2038: Add __USE_TIME_BITS64 support for struct timespec +- y2038: Add __USE_TIME_BITS64 support for struct timeval +- y2038: Add __USE_TIME_BITS64 support for time_t +- linux: Add recvvmsg fallback for 64-bit time_t SO_TIMESTAMP{NS} +- linux: Add fallback for 64-bit time_t SO_TIMESTAMP{NS} +- linux: Add fallback for 64-bit time_t SO_{RCV,SND}TIMEO +- linux: s390: Add libanl.abilist in s390 and s390x +- linux: mips: Split libanl.abilist in n32 and n64 +- linux: mips: Split librt.abilist in n32 and n64 +- Reinstate gconv-modules as the default configuration file + +* Tue Jun 15 2021 Florian Weimer - 2.33.9000-16 +- Preserve some symbols in libc.so.6's symtab (#1965374) + +* Tue Jun 15 2021 Florian Weimer - 2.33.9000-15 +- Install shared objects under their ABI names, avoiding symlinks (#1652867) + +* Mon Jun 14 2021 Siddhesh Poyarekar - 2.33.9000-14 +- Add a conditional dependency for glibc-gconv-extra.i686 in x86_64. + +* Mon Jun 14 2021 Siddhesh Poyarekar - 2.33.9000-13 +- Auto-sync with upstream branch master, + commit ebae2f5a6f971a8f0b6c99e00f9c45ef7433924a. +- Revert gconv configuration file name to gconv-modules. + +* Thu Jun 03 2021 Florian Weimer - 2.33.9000-12 +- libdl is no longer a separate shared object. +- CVE-2021-33574: Use-after-free via mq_notify (#1965410) +- Auto-sync with upstream branch master, + commit 466c1ea15f461edb8e3ffaf5d86d708876343bbf: +- dlfcn: Rework static dlopen hooks +- dlfcn: Eliminate GLIBC_PRIVATE dependency from tststatic2 +- dlfcn: Cleanups after -ldl is no longer required +- dlfcn: Move dlopen into libc +- dlfcn: Move dlvsym into libc +- dlfcn: Move dlinfo into libc +- dlfcn: Move dladdr1 into libc +- dlfcn: Move dlmopen into libc +- dlfcn: Move dlsym into libc +- dlfcn: Move dladdr into libc +- dlfcn: Move dlclose into libc +- Improve test coverage of strlen function +- fix typo +- dlfcn: Move dlerror into libc +- Add libc ABI extension kludge for baseline-violating libdl symbols +- scripts/versions.awk: Add local: * to all version nodes +- Add missing symbols to Version files +- Fix use of __pthread_attr_copy in mq_notify (bug 27896) +- Use __pthread_attr_copy in mq_notify (bug 27896) +- Update floating-point feature test macro handling for C2X +- stdio-common: Remove _IO_vfwscanf +- aarch64: align stack in clone [BZ #27939] +- powerpc: Optimized memcmp for power10 +- x86-64: Align child stack to 16 bytes [BZ #27902] + +* Mon May 31 2021 Florian Weimer - 2.33.9000-11 +- glibc-sigsetxid-sa_onstack.patch was applied upstream +- Auto-sync with upstream branch master, + commit 271ec55d0ae795f03d92e3aa61bff69a31a19e3a: +- support: Do not build xpthread_attr_setaffinity_np for hurd +- nptl: Add pthread_attr_setaffinity_np failure test +- support: Add xpthread_attr_setaffinity_np wrapper +- nptl: Move createthread to pthread_create +- nptl: Move Linux createthread to nptl +- nptl: Install SIGSETXID handler with SA_ONSTACK [BZ #27914] +- aarch64: Added optimized memset for A64FX +- aarch64: Added optimized memcpy and memmove for A64FX +- benchtests: Fixed bench-memcpy-random: buf1: mprotect failed +- aarch64: Added Vector Length Set test helper script +- aarch64: define BTI_C and BTI_J macros as NOP unless HAVE_AARCH64_BTI +- config: Added HAVE_AARCH64_SVE_ASM for aarch64 +- tst-mallinfo2.c: Use correct multiple for total variable + +* Wed May 26 2021 Florian Weimer - 2.33.9000-10 +- nptl: Install SIGSETXID handler with SA_ONSTACK [BZ #27914] + +* Tue May 25 2021 Florian Weimer - 2.33.9000-9 +- Auto-sync with upstream branch master, + commit ac0353af81a23535f517586a5d04427120a157ac. +- This removes libpthread as a separate shared object. New programs + will depend on the GLIBC_2.34 symbol version. Upstream development + brings in the following noteworthy bug fixes: +- ppc64le: scv ABI error handling fails to check IS_ERR_VALUE (#1962971) +- CVE-2021-27645 glibc: Use-after-free in addgetnetgrentX function in + netgroupcache.c (#1932590) +- Linking the main program with jemalloc causes sysconf to deadlock in + audit mode (#1909920) + +* Fri May 21 2021 Florian Weimer - 2.33.9000-8 +- Switch back to a unified glibc-headers package for downstream (#1940686) + +* Fri May 21 2021 Florian Weimer - 2.33.9000-7 +- aarch64: Enable optional memory tagging support + +* Tue May 11 2021 Florian Weimer - 2.33.9000-6 +- Use distribution mechanism for debuginfo (#1661510, #1886295, #1905611) + +* Thu May 6 2021 Florian Weimer - 2.33.9000-5 +- Build locales in parallel again + +* Tue May 4 2021 Florian Weimer - 2.33.9000-4 +- Various changes to get glibc building again, using selected upstream + backports. +- Re-enable -Werror by default. +- This release introduces __libc_start_main@@GLIBC_2.34, so binaries + are not compatible with glibc 2.33. Building binaries against this + glibc version is not recommended because of the partial libpthread + transition. +- glibc-upstream-amx-detection.patch: Fix build failure after GCC 11.1 update. +- glibc-upstream-malloc-test-hang.patch: Avoid test hang due to + annobin/binutils bug (#1951492) + +* Wed Mar 03 2021 Arjun Shankar - 2.33.9000-3 +- Drop glibc-rh819430.patch; fixed upstream. +- Auto-sync with upstream branch master, + commit f01a61e13872109b3b233158ab664364bd1879bc: +- i386: Regenerate ulps +- x86: Add CPU-specific diagnostics to ld.so --list-diagnostics +- x86: Automate generation of PREFERRED_FEATURE_INDEX_1 bitfield +- ld.so: Implement the --list-diagnostics option +- powerpc: Update libm-test-ulps +- tst: Add test for utimes +- tst: Add test for utime +- tst: Add test for futimens +- nptl: __libc_cleanup_push/__libc_cleanup_pop require -fexceptions +- elf: Build __dl_iterate_phdr with unwinding support [BZ #27498] +- nptl: Use for accessing the libgcc_s unwinder +- Implement _Unwind_Resume in libc on top of +- Move sysdeps/gnu/unwind-resume.c to sysdeps/generic/unwind-resume.c +- __frame_state_for: Use for unwinder access +- sparc: Implement backtrace on top +- m68k: Implement backtrace on top of +- i386: Implement backtrace on top of +- arm: Implement backtrace on top of +- backtrace: Implement on top of +- Implement for dynamically loading the libgcc_s unwinder +- Correct buffer end pointer in IO_wdefault_doallocate (BZ #26874) +- aarch64: update ulps. +- Add inputs that generate larger error bounds +- Reduce the statically linked startup code [BZ #23323] +- posix: Falling back to non wide mode in case of encoding error [BZ #14185] +- nptl: Move elision implementations into libc +- NEWS: Add missing bug closures +- added rt to malloc/Depend [BZ #27132] +- x86: Use x86/nptl/pthreaddef.h +- nptl: Move futex-internal into libc +- nptl: Move lowlevellock into libc [BZ #15648] +- nptl: Move futex-internal.c into main nptl directory +- nptl: Reformat Versions +- nptl: Split libpthread-routines into one routine per line +- x86: Remove unused variables for raw cache sizes from cacheinfo.h +- Use Linux 5.11 in build-many-glibcs.py. +- : Correct x86_cpu_TBM +- x86_64/clone.S: Upate comments +- i386/clone.S: Remove redundant EBX load +- aarch64: Remove the unused __read_tp symbol +- build-many-glibcs.py: Use make -O for more consistent log output + +* Sun Feb 21 2021 Carlos O'Donell - 2.33.9000-2 +- Auto-sync with upstream branch master, + commit e9e7f24543e6d1b0a31641f144697e261df6ccd7: +- configure: Replace obsolete AC_TRY_LINK with AC_LINK_IFELSE +- configure: Remove obsolete AC_CHECK_TOOL_PREFIX +- configure: Replace obsoleted AC_HELP_STRING with AS_HELP_STRING +- Update syscall lists for Linux 5.11. +- Correct hppa EFD_NONBLOCK, IN_NONBLOCK, SFD_NONBLOCK and TFD_NONBLOCK defines. +- string: Work around GCC PR 98512 in rawmemchr +- tst: time: Provide Y2038 tests for mktime (tst-mktime4.c) +- tst: Provide test for difftime +- tst: Provide test for ctime +- tst: Provide test for sched_rr_get_interval +- S390: Add new hwcap values. +- aarch64: Fix sys/ptrace.h if linux headers are included +- elf: Do not copy vDSO soname when setting up link map +- x86: Remove the extra space between "# endif" +- hurd: Fix fstatfs build failure +- linux: Remove stat-check.c +- linux: Remove overflow.h +- linux: Consolidate internal_statvfs +- linux: Consolidate statvfs implementations +- linux: Consolidate fstatvfs implementations +- linux: Consolidate statfs implementations +- linux: Consolidate fstatfs implementations +- linux: Set LFS statfs as default +- linux: Set default kernel_stat.h to LFS +- linux: Fix STATFS_IS_STATFS64 definition + +* Fri Feb 12 2021 Florian Weimer - 2.33.9000-1 +- Auto-sync with upstream branch master, + commit 228f30ab4724d4087d5f52018873fde22efea6e2: +- tunables: Disallow negative values for some tunables +- x86: Use SIZE_MAX instead of (long int)-1 for tunable range value +- tunables: Simplify TUNABLE_SET interface +- setrlimit/getrlimit: Use __nonnull to avoid null pointer +- benchtests: Updated json bench-variant attribute +- regex: stop using alloca +- regexec: remove alloca usage in build_trtable +- regex: remove alloca usage on regex set_regs +- malloc: Sync dynarray with gnulib +- misc: Sync cdefs.h with gnulib +- linux: Fix __sem_check_add_mapping search_sem +- linux: Fix __sem_check_add_mapping name length +- Add more ptrace constants for AArch64 and PowerPC. +- strchr: Add additional benchmarks and tests +- x86-64: Refactor and improve performance of strchr-avx2.S +- pthread: Remove alloca usage from __sem_check_add_mapping +- pthread: Refactor semaphore code +- linux: Require /dev/shm as the shared memory file system +- sunrpc: Fix typo in xdr_string comment +- tst: Provide test for ppoll +- tst: Provide test for timerfd related functions +- x86: Add PTWRITE feature detection [BZ #27346] +- nsswitch: return result when nss database is locked [BZ #27343] +- printf: Add smoke tests for long double +- Add NT_ARM_TAGGED_ADDR_CTRL from Linux 5.10 to elf.h. +- argp: Avoid undefined behaviour when invoking qsort(). +- argp: Improve comments. +- argp: Don't pass invalid arguments to isspace, isalnum, isalpha, isdigit. +- argp: Don't rely on undefined behaviour of _tolower(). +- argp: fix pointer-subtraction bug +- Use binutils 2.36 branch in build-many-glibcs.py. +- manual: Correct description of ENTRY [BZ #17183] +- nptl: Remove private futex optimization [BZ #27304] +- stdio-common: Add a few double formatting tests [BZ #27245] +- posix/tst-rfc3484: Fix compile failure linking to local __stat64 +- i686: Regenerate ULPs +- tst-rtld-list-tunables.sh: Unset glibc tunables +- linux: Remove shmmax check from tst-sysvshm-linux +- x86: Adding an upper bound for Enhanced REP MOVSB. +- Fix version.h for glibc 2.34 development +- Add MS_NOSYMFOLLOW from Linux 5.10 to . +- Move _SC_MINSIGSTKSZ/_SC_SIGSTKSZ entry in NEWS +- libSegFault: Fix printing signal number [BZ #27249] +- hurd TIOCFLUSH: fix fixing argument +- sysconf: Add _SC_MINSIGSTKSZ/_SC_SIGSTKSZ [BZ #20305] +- hurd TIOCFLUSH: Cope BSD 4.1 semantic +- tst-mallinfo2.c: Remove useless trailing semicolon for macro +- elf: Replace a --defsym trick with an object file to be compatible with LLD +- Open master branch for glibc 2.34 development +- Prepare for glibc 2.33 release +- Update NEWS with bugs +- Update translations +- NEWS: Fix typo in CVE-2021-3326 entry +- elf: Fix tests that rely on ld.so.cache for cross-compiling +- NEWS: Mention CVE-2021-3326 (iconv assertion with ISO-20220-JP-3) +- NEWS: Add entry for glibc-hwcaps and deprecate legacy hwcaps +- x86: Properly set usable CET feature bits [BZ #26625] +- Update translations +- Update libc.pot for 2.33 release +- Update ia64 libm-test-ulps +- sh: Update libm-tests-ulps +- ia64: Fix brk call on statup +- Update sparc libm-test-ulps +- Update alpha libm-test-ulps +- powerpc64: Workaround sigtramp vdso return call +- Fix nss/tst-reload2 for systems without PATH_MAX +- nsswitch: do not reload if "/" changes +- elf: Limit tst-prelink-cmp target archs +- CVE-2021-3326: gconv: Fix assertion failure in ISO-2022-JP-3 module (#1921917) + +* Wed Jan 27 2021 Arjun Shankar - 2.32.9000-29 +- Auto-sync with upstream branch master, + commit df359a25ba6f6bda06104229fbfe284c1fb30915: +- Revert "Make libc symbols hidden in static PIE" [BZ #27237] +- benchtests: Do not build bench-timing-type with MODULE_NAME=libc +- aarch64: Fix the list of tested IFUNC variants [BZ #26818] +- Update INSTALL with package versions that are known to work +- aarch64: Move and update the definition of MTE_ENABLED +- Fix misplaced const +- Update C-SKY libm-test-ulps +- manual: Correct argument order in mount examples [BZ #27207] +- linux: mips: Fix getdents64 fallback on mips64-n32 +- x86: Properly match CPU features in /proc/cpuinfo [BZ #27222] +- x86-64: Update tst-glibc-hwcaps-2.c for x86-64 baseline +- powerpc64: Select POWER9 machine for the scv instruction +- x86: Check ifunc resolver with CPU_FEATURE_USABLE [BZ #27072] +- Revert "linux: Move {f}xstat{at} to compat symbols" for static build +- aarch64: revert memcpy optimze for kunpeng to avoid performance degradation +- Make libc symbols hidden in static PIE +- csu: Move static pie self relocation later [BZ #27072] +- Use hidden visibility for early static PIE code +- csu: Avoid weak ref for __ehdr_start in static PIE +- configure: Check for static PIE support +- elf: Avoid RELATIVE relocs in __tunables_init +- elf: Make the tunable struct definition internal only +- : Remove the C preprocessor magic +- posix: Fix fnmatch.c on bootstrap +- stdlib: Add testcase for BZ #26241 +- posix: Fix regex_internal.h on bootstrap +- Use in __libc_init_secure +- elf: Avoid RELATIVE relocation for _dl_sysinfo +- libmvec: Add extra-test-objs to test-extras +- Hurd: Add rtld-strncpy-c.c +- Update MIPS libm-test-ulps. +- Update arm libm-test-ulps. +- Update powerpc-nofpu libm-test-ulps. +- Update hppa libm-test-ulps +- ARC: nofpu: Regenerate ulps +- ld.so: Add --list-tunables to print tunable values +- math/test-tgmath2: Fix fabs failure when no long double +- x86: Move x86 processor cache info to cpu_features +- Fix x86 build with --enable-tunable=no +- ifuncmain6pie: Remove the circular IFUNC dependency [BZ #20019] + +* Tue Jan 26 2021 Fedora Release Engineering - 2.32.9000-28 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_34_Mass_Rebuild + +* Sat Jan 23 2021 Arjun Shankar - 2.32.9000-27 +- Introduce new glibc-doc.noarch subpackage (#1346925) +- Move the reference manual info pages from glibc-devel to glibc-doc +- Move debugger interface documentation from glibc to glibc-doc +- Remove unnecessary README, INSTALL, NEWS files from glibc +- Remove unnecessary README.timezone and gai.conf files from glibc-common + +* Thu Jan 14 2021 Arjun Shankar - 2.32.9000-26 +- Deprecate nscd (#1905135) +- https://fedoraproject.org/wiki/Changes/DeprecateNSCD + +* Wed Jan 13 2021 Carlos O'Donell - 2.32.9000-25 +- Auto-sync with upstream branch master, + commit cf1290064598def8dfeddec3d86d98495aee1fba: +- Use the right argument code in unnormal tests +- ldconfig/x86: Store ISA level in cache and aux cache +- elf: work around a gcc bug in elf_get_dynamic_info +- x86: Set header.feature_1 in TCB for always-on CET [BZ #27177] +- posix: consume less entropy on tempname +- Makerules: Do not require startup files for format.lds probe object +- install: Replace scripts/output-format.sed with objdump -f [BZ #26559] +- math: Add BZ#18980 fix back on dbl-64 cosh +- posix: Sync tempname with gnulib [BZ #26648] +- posix: Fix return value of system if shell can not be executed [BZ #27053] +- support: Add xchmod wrapper +- Update STATX_ATTR_DAX value from Linux 5.10. +- riscv: Initialize $gp before resolving the IRELATIVE relocation +- riscv: support GNU indirect function +- posix: Correct attribute access mode on readlinkat [BZ #27024]. +- Add xfchmod to libsupport +- Add xchdir to libsupport. +- POSIX locale: Fix typo in comment +- ARC: Regenerate ulps +- mntent: Use __putc_unlocked instead of fputc_unlocked +- aarch64: define PI_STATIC_AND_HIDDEN +- Update NEWS for CVE-2019-25013. +- x86: Support GNU_PROPERTY_X86_ISA_1_V[234] marker [BZ #26717] +- Remove dbl-64/wordsize-64 (part 2) +- Remove dbl-64/wordsize-64 +- Add SEGV_MTEAERR and SEGV_MTESERR from Linux 5.10. +- support: Add support_small_thread_stack_size +- stdlib: Sync canonicalize with gnulib [BZ #10635] [BZ #26592] [BZ #26341] [BZ #24970] +- malloc: Add scratch_buffer_dupfree +- Import filename.h from gnulib +- Import idx.h from gnulib +- alpha: Provide wait4 for static library [BZ #27150] +- aarch64: push the set of rules before falling into slow path +- nptl: Fix comment typo in pthread_cond_wait.c +- x86: Check IFUNC definition in unrelocated executable [BZ #20019] +- hurd: Fix mmap(!MAP_FIXED) on bogus address + +* Mon Jan 04 2021 Florian Weimer - 2.32.9000-24 +- Drop glibc-fedora-__libc_multiple_libcs.patch. Replaced by upstream's + __libc_initial flag. +- Adjust glibc-rh819430.patch to upstream's gnulib merge of fnmatch. +- Disable -Werror again due to GCC PR98512. +- Auto-sync with upstream branch master, + commit 3ec5d83d2a237d39e7fd6ef7a0bc8ac4c171a4a5: +- x86-64: Avoid rep movsb with short distance [BZ #27130] +- aarch64: fix stack missing after sp is updated +- nptl: Remove set*id, set*gid files which are not built +- Drop nan-pseudo-number.h usage from tests +- posix: Sync fnmatch with gnulib +- Sync flexmember.h with gnulib +- Sync intprops.h with gnulib +- posix: Sync glob code with gnulib +- posix: Sync regex code with gnulib +- Move generic nan-pseudo-number.h to ldbl-96 +- Sync FDL from https://www.gnu.org/licenses/fdl-1.3.texi +- Sync move-if-change from Gnulib +- Update automatically-generated copyright dates +- Update copyright dates not handled by scripts/update-copyrights. +- Update copyright dates with scripts/update-copyrights +- aarch64: use PTR_ARG and SIZE_ARG instead of DELOUSE +- nonstring: Enable __FORTIFY_LEVEL=3 +- string: Enable __FORTIFY_LEVEL=3 +- Introduce _FORTIFY_SOURCE=3 +- Warn on unsupported fortification levels +- powerpc: Use scv instruction on clone when available +- powerpc: Runtime selection between sc and scv for syscalls +- malloc: preserve errno on mcheck hooks [BZ #17924] +- x86 long double: Add tests for pseudo normal numbers +- x86 long double: Consider pseudo numbers as signaling +- io: Remove xmknod{at} implementations +- io: Remove xstat implementations +- free: preserve errno [BZ#17924] +- hurd: Accept including hurd/version.h +- hurd: Add WSTOPPED/WCONTINUED/WEXITED/WNOWAIT support [BZ #23091] +- support: Make support_process_state_wait wait less +- hurd: set sigaction for signal preemptors in arch-independent file +- hurd: Fix spawni SPAWN_XFLAGS_TRY_SHELL with empty argv +- hurd: Try shell in posix_spawn* only in compat mode +- Remove _ISOMAC check from +- x86: Remove the duplicated CPU_FEATURE_CPU_P +- Partially revert 681900d29683722b1cb0a8e565a0585846ec5a61 +- x86 long double: Support pseudo numbers in isnanl +- x86 long double: Support pseudo numbers in fpclassifyl +- MTE: Do not pad size in realloc_check +- tests-mcheck: New variable to run tests with MALLOC_CHECK_=3 +- elf: Account for glibc-hwcaps/ prefix in _dl_important_hwcaps +- misc: Use __ferror_unlocked instead of ferror +- s390x: Regenerate ulps +- powerpc: Regenerate ulps +- addmntent: Remove unbounded alloca usage from getmntent [BZ#27083] +- : Add Intel LAM support +- i386: Regenerate ulps +- aarch64: update ulps. +- aarch64: Add aarch64-specific files for memory tagging support +- aarch64: Add sysv specific enabling code for memory tagging +- linux: Add compatibility definitions to sys/prctl.h for MTE +- malloc: Basic support for memory tagging in the malloc() family +- elf: Add a tunable to control use of tagged memory +- config: Allow memory tagging to be enabled when configuring glibc +- alpha: Remove anonymous union in struct stat [BZ #27042] +- add inputs to auto-libm-test-in yielding larger errors (binary64, x86_64) +- m68k: fix clobbering a5 in setjmp() [BZ #24202] +- iconv add iconv_close before the function returned with bad value. +- iconv: use iconv_close after iconv_open +- Fix buffer overrun in EUC-KR conversion module (bz #24973) +- hurd: Make trampoline fill siginfo ss_sp from sc_uesp +- Hurd: make sigstates hold a reference on thread ports +- profil-counter: Add missing SIGINFO case +- hurd: implement SA_SIGINFO signal handlers. +- hurd: Fix ELF_MACHINE_USER_ADDRESS_MASK value +- hurd: Note when the vm_map kernel bug was fixed +- hurd: Also turn KERN_INVALID_ADDRESS to EINVAL +- ieee754: Remove unused __sin32 and __cos32 +- ieee754: Remove slow paths from asin and acos +- getenv: Move call to strlen to the branch it's used in. +- Update kernel version to 5.10 in tst-mman-consts.py. +- s390x: Require GCC 7.1 or later to build glibc. +- malloc: Use __libc_initial to detect an inner libc +- Replace __libc_multiple_libcs with __libc_initial flag +- {nptl,htl}/semaphoreP.h: clean up +- htl: Get sem_open/sem_close/sem_unlink support [BZ #25524] +- pthread: Move semaphore initialization for open to semaphoreP.h +- Mark __libc_freeres_fn as used [BZ #27002] +- Update syscall lists for Linux 5.10. +- htl: Add pshared semaphore support +- hurd: Add LLL_PRIVATE and LLL_SHARED +- hurd: Add __libc_open and __libc_close +- htl: Add futex-internal.h +- hurd: Add __lll_abstimed_wait_intr +- hurd: make lll_* take a variable instead of a ptr +- hurd: Rename LLL_INITIALIZER to LLL_LOCK_INITIALIZER +- Use Linux 5.10 in build-many-glibcs.py. + +* Wed Dec 16 2020 DJ Delorie - 2.32.9000-23 +- Fix conditionals for _enable_debug_packages and benchtests [BZ #1902514] + +* Tue Dec 15 2020 Patsy Griffin - 2.32.9000-22 +- Auto-sync with upstream branch master, + commit 4d0985543f479a6f421d4d8a9e0d1dc71c9c2c53. +- elf: Record libc.so link map when it is the main program (bug 20972) +- Use GMP 6.2.1 in build-many-glibcs.py. +- aarch64: remove the strlen_asimd symbol +- aarch64: fix static PIE start code for BTI [BZ #27068] +- elf: Fix failure handling in _dl_map_object_from_fd +- elf: inline lose for error handling +- Remove strtoimax, strtoumax, wcstoimax, wcstoumax inlines +- nsswitch: handle missing actions properly +- x86: Remove the default REP MOVSB threshold tunable value [BZ #27061] +- elf.h: Remove SHF_GNU_BUILD_NOTE. +- elf.h: fix spelling typos in comments +- Fix spelling and grammar in several comments +- malloc: Detect infinite-loop in _int_free when freeing tcache [BZ#27052] +- elf: Fix dl-load.c +- elf: Include libc.so.6 as main program in dependency sort (bug 20972) +- support: Add support_slibdir_prefix variable +- aarch64: Use mmap to add PROT_BTI instead of mprotect [BZ #26831] +- elf: Pass the fd to note processing +- elf: Move note processing after l_phdr is updated +- aarch64: align address for BTI protection [BZ #26988] +- aarch64: Fix missing BTI protection from dependencies [BZ #26926] +- Fix linknamespace errors in nss_database.c if build with -Os. +- treewide: fix incorrect spelling of indices in comments +- linux: Consolidate brk implementation +- elf: Include in cache.c +- s390x: Add glibc-hwcaps support +- elf: Fix run-time dependencies of tst-dlopen-fail-2 +- Handle out-of-memory case in svc_tcp.c/svc_unix.c:rendezvous_request. +- elf: Fix incorrect comparison in sort_priorities_by_name +- S390: Derive float_t from FLT_EVAL_METHOD +- Fix parsing of /sys/devices/system/cpu/online (bug 25859) +- Make strtoimax, strtoumax, wcstoimax, wcstoumax into aliases +- Fixed typos in "NEWS for version 2.32" +- Add NEWS entry for CVE-2020-29562 (BZ #26923) +- iconv: Fix incorrect UCS4 inner loop bounds (BZ#26923) +- Drop glibc-rh1906066 and glibc-rh741105 patches fixed by sync. + +* Mon Dec 14 2020 Florian Weimer - 2.32.9000-21 +- Re-enable -Werror everywhere (#1888246) + +* Wed Dec 09 2020 DJ Delorie - 2.32.9000-20 +- nsswitch: handle missing actions properly (temporary fix for 1906066) + +* Mon Dec 07 2020 Arjun Shankar - 2.32.9000-19 +- Auto-sync with upstream branch master, + commit 088e9625378f25607acff3daf7a79cbdee497043: +- x86: Rename readelflib.c +- nsswitch: use new internal API (callers) +- nsswitch: user new internal API (tests) +- nsswitch: use new internal API (core) +- nss: Implement +- : New abstraction for combining NSS modules and NSS actions +- nss: Introduce +- Add scripts/move-symbol-to-libc.py + +* Fri Dec 04 2020 Arjun Shankar - 2.32.9000-18 +- Drop glibc-revert-fxstat-compat.patch; applied upstream. +- Drop glibc-revert-mknod-compat.patch; applied upstream. +- Auto-sync with upstream branch master, + commit 4c38c1a229bc3628269ad98bd7e8d31d118d91f6: +- powerpc64le: Add glibc-hwcaps support +- x86: Adjust tst-cpu-features-supports.c for GCC 11 +- x86: Set RDRAND usable if CPU supports RDRAND +- elf: Add missing header to elf/dl-hwcaps.h +- lowlevellock-futex: Remove not used macros +- futex: Remove not used futex_reltimed_wait{_cancelable} +- y2038: Convert gai_suspend to support 64 bit time +- symbols: Add defines for libanl's libanl_hidden_{def|proto} +- x86_64: Add glibc-hwcaps support +- elf: Add glibc-hwcaps subdirectory support to ld.so cache processing +- elf: Process glibc-hwcaps subdirectories in ldconfig +- elf: Implement tail merging of strings in ldconfig +- elf: Implement a string table for ldconfig, with tail merging +- elf: Add extension mechanism to ld.so.cache +- elf: Add endianness markup to ld.so.cache (bug 27008) +- elf: Add glibc-hwcaps support for LD_LIBRARY_PATH +- elf: Synchronize section header flags with binutils +- x86: Fix THREAD_SELF definition to avoid ld.so crash (bug 27004) +- htl: Add hidden def for __pthread_create/detach +- manual: Clarify File Access Modes section and add O_PATH +- htl: Add missing symbols +- Revert "linux: Move xmknod{at} to compat symbols" +- Revert "linux: Move {f}xstat{at} to compat symbols" +- elf.h: Fix spelling of EM_TILE64 comment +- nptl: Fix __futex_clocklock64 return error check [BZ #26964] +- powerpc64le: ifunc select *f128 routines in multiarch mode +- y2038: Convert aio_suspend to support 64 bit time +- Fix typo in NEWS file +- nptl: Add EOVERFLOW checks for futex calls +- nptl: Fix PTHREAD_PRIO_PROTECT timed lock +- sh: Add sh4 fpu Implies folder +- io: nftw/ftw: Fix stack overflow with large nopenfd [BZ #26353] +- elf: Introduce enum opt_format in the ldconfig implementation +- support: Add support_copy_file +- NEWS entry for commit b4f020c9b408fb3d1d3d4901c4a71839145f8791 +- timezone: Change zdump installation to bin directory +- nptl: Return EINVAL for invalid clock for pthread_clockjoin_np +- nptl: Return EINVAL for pthread_mutex_clocklock/PI with CLOCK_MONOTONIC [BZ #26801] +- nptl: Replace lll_futex_wake with futex-internal.h +- nptl: Replace lll_futex_supported_clockid with futex-internal.h +- nptl: Replace lll_futex_{timed_}wait by futex-internal.h +- nptl: Replace lll_timedwait with __futex_abstimed_wait64 +- nptl: Replace __futex_clocklock_wait64 with __futex_abstimed_wait64 +- nptl: Remove _futex_clock_wait_bitset64 +- nptl: Consolidate __futex_abstimed_wait_{cancelable}64 +- nptl: Extend __futex_abstimed_wait_cancelable64 comment +- nptl: Remove clockwait_tid +- nptl: Remove futex_wait_cancelable +- nptl: Remove unused internal futex functions +- Mark mtrace tests UNSUPPORTED if bug-ga2.mtrace or tst-leaks2.mtrace are missing +- elf: Fix uninitialized variable for _dl_write +- powerpc: Make PT_THREAD_POINTER available to assembly code +- Use libnss_files.so for tests posix/bug-ga2 and resolv/tst-leaks2 [BZ #26821] +- hurd report-wait: Fix stpcpy usage +- hurd S_msg_report_wait: Fix detecting fd ports +- hurd S_msg_report_wait: Fix reporting ports +- hurd: Fix strcpy calls +- hurd: Fix _S_msg_get/set_env_variable prototype +- hurd: Enable using ifunc +- Add {,sysdep-}ld-library-path make variable +- nptl: Move stack list variables into _rtld_global +- hurd: let _dl_argv and __libc_stack_end be relro +- hurd: Remove some remnants of cthreads +- nanosleep: Pass NULL when rem == NULL on ports with __TIMESIZE != 64 +- y2038: Convert thrd_sleep to support 64 bit time +- y2038: Convert mtx_timedlock to support 64 bit time +- y2038: Convert cnd_timedwait to support 64 bit time +- hurd: Drop CLOCK_MONOTONIC change which slipped in +- hurd: make ptsname fail with ENOTTY on non-master-pty +- mach: Add missing assert.h include +- hurd: break relocation loop between libc.so and lib{mach,hurd}user.so +- Remove obsolete defines for HPUX support from fcntl.h and update O_NONBLOCK. +- Remove tls.h inclusion from internal errno.h +- nptl: Eliminate and __is_smp +- powerpc: Eliminate UP macro conditionals +- x86: Remove UP macro. Define LOCK_PREFIX unconditionally. +- alpha: Remove UP preprocessor conditionals +- hurd: Make sure signals get started +- hurd: initialize libpthread before starting the signal thread +- hurd: Make _hurd_libc_proc_init idempotent +- powerpc: Add optimized stpncpy for POWER9 +- powerpc: Add optimized strncpy for POWER9 +- Don't use nested function in test-ffs +- Use __builtin___stpncpy_chk when available +- tests: Remove NULL check for an array +- hurd: Move {,f,l}xstat{,at} and xmknod{at} to compat symbols +- hurd: Notify the proc server later during initialization +- htl: Initialize later +- htl: Keep thread signals blocked during its initialization +- htl: Fix spurious symbols in namespaces +- Use O_CLOEXEC in sysconf [BZ #26791] +- struct _Unwind_Exception alignment should not depend on compiler flags +- hurd: keep only required PLTs in ld.so +- hurd: Add missing startup calls +- riscv: Get cache information through sysconf +- RISC-V: Add _dl_start_user. + +* Thu Nov 26 2020 Florian Weimer - 2.32.9000-17 +- s390x: Do not rewrite program interpreter symlink (make install is enough) + +* Tue Nov 10 2020 Carlos O'Donell - 2.32.9000-16 +- Remove the work around for systemd-nspawn (#1869030). + +* Mon Nov 09 2020 DJ Delorie - 2.32.9000-15 +- Auto-sync with upstream branch master, + commit 75a193b7611bade31a150dfcc528b973e3d46231. +- linux: Allow adjtime with NULL argument [BZ #26833] +- aarch64: Add unwind information to _start (bug 26853) +- bsd unlockpt: unlockpt needs to fail with EINVAL, not ENOTTY +- Rearrange bsd_getpt vs bsd_openpt and implement posix_openpt on BSD +- Remove __warndecl +- Remove __warn_memset_zero_len [BZ #25399] +- iconv: Accept redundant shift sequences in IBM1364 [BZ #26224] +- msg: Remove redundant #include header +- tst-setuid1-static-ENV: Add $(common-objpfx)nss [BZ #26820] + +* Tue Nov 03 2020 Patsy Griffin - 2.32.9000-14 +- Auto-sync with upstream branch master, + commit e156dabc766d6f6f99ce9402999eae380a3ec1f2. +- aarch64: Add variant PCS lazy binding test [BZ #26798] +- aarch64: Fix DT_AARCH64_VARIANT_PCS handling [BZ #26798] +- hurd: Correct 'ethenet' spelling +- Avoid -Wstringop-overflow warning in pthread_cleanup_push macros +- Disable spurious -Warray-bounds for ypclnt.c (bug 26687) +- Do not use array parameter to new_composite_name (bug 26726) +- Disable spurious -Wstringop-overflow for setjmp/longjmp (bug 26647) +- malloc debug: fix compile error when enable macro MALLOC_DEBUG > 1 +- tst-tcfree2: adjust coding style. +- elf: In ldconfig, extract the new_sub_entry function from search_dir +- Use MPC 1.2.1 in build-many-glibcs.py. +- Argument Syntax: Use "option", @option, and @command. +- elf: Unify old and new format cache handling code in ld.so +- x86: Restore processing of cache size tunables in init_cacheinfo +- Make elf.h header self contained. +- x86: Optimizing memcpy for AMD Zen architecture. +- Hurd: Fix ftime build +- Add IP_RECVERR_RFC4884 and IPV6_RECVERR_RFC4884 from Linux 5.9. +- misc: Add internal __getauxval2 function +- Remove NEWS entry about ftime removal +- time: Add 64-bit time_t support for ftime +- Reinstate ftime and add deprecate message on ftime usage +- Update kernel version to 5.9 in tst-mman-consts.py. +- Amend grammar and add a description +- Fix typo in NEWS file +- Remove timing related checks of time/tst-cpuclock1 +- Update syscall lists for Linux 5.9. +- Use Linux 5.9 in build-many-glibcs.py. +- Reword description of SXID_* tunable properties +- New benchtest: pthread locks +- y2038: nptl: Provide __futex_clock_wait_bitset64 to support 64 bit bitset +- C-SKY: Make dynamic linker's name compitable with the older gcc. +- Revert "C-SKY:Fix dynamic linker's name when mfloat-abi=softfp." +- Move vtimes to a compatibility symbol +- y2038: linux: Provide __time64 implementation +- rt: Fix typos in comments in +- C-SKY:Fix dynamic linker's name when mfloat-abi=softfp. +- Drop the glibc-revert-ftime-compat.patch. + +* Thu Oct 29 2020 DJ Delorie - 2.32.9000-13 +- Add BuildRequires for perl (malloc/mtrace) if running the testsuite. + +* Wed Oct 21 2020 Siddhesh Poyarekar - 2.32.9000-12 +- Revert __xstat64 symbol removal. +- Revert xmknod* symbol removal. +- Revert ftime symbol removal. + +* Sun Oct 18 2020 Patsy Griffin - 2.32.9000-11 +- Auto-sync with upstream branch master, + commit 0f09154c64005e78b61484ae87b5ea2028051ea0. +- x86: Initialize CPU info via IFUNC relocation [BZ 26203] +- Add NEWS entry for ftime compatibility move +- support: Add create_temp_file_in_dir +- linux: Add __readdir_unlocked +- linux: Simplify opendir buffer allocation +- linux: Move posix dir implementations to Linux +- linux: Add 64-bit time_t support for wait3 +- Move ftime to a compatibility symbol +- linux: Fix time64 support for futimesat +- linux: Use INTERNAL_SYSCALL on fstatat{64} +- shm tests: Append PID to names passed to shm_open [BZ #26737] +- sysvipc: Fix tst-sysvshm-linux on x32 +- x86/CET: Update vfork to prevent child return +- resolv: Serialize processing in resolv/tst-resolv-txnid-collision +- statfs: add missing f_flags assignment +- y2038: Remove not used __fstatat_time64 define +- y2038: nptl: Convert pthread_mutex_{clock|timed}lock to support 64 bit +- sysvipc: Return EINVAL for invalid shmctl commands +- sysvipc: Fix IPC_INFO and SHM_INFO handling [BZ #26636] +- AArch64: Use __memcpy_simd on Neoverse N2/V1 +- resolv: Handle transaction ID collisions in parallel queries (bug 26600) +- support: Provide a way to clear the RA bit in DNS server responses +- support: Provide a way to reorder responses within the DNS test server +- Add missing stat/mknod symbol on libc.abilist some ABIs +- manual: correct the spelling of "MALLOC_PERTURB_" [BZ #23015] +- manual: replace an obsolete collation example with a valid one +- rtld: fix typo in comment +- elf: Add missing header to elf/dl-usage.c +- hurd: support clock_gettime(CLOCK_PROCESS/THREAD_CPUTIME_ID) +- linux: Move xmknod{at} to compat symbols +- linux: Add {f}stat{at} y2038 support +- linux: Move {f}xstat{at} to compat symbols +- linux: Disentangle fstatat from fxstatat +- linux: Implement {l}fstat{at} in terms of fstatat +- linux: Move the struct stat{64} to struct_stat.h +- Remove mknod wrapper functions, move them to symbols +- Remove stat wrapper functions, move them to exported symbols +- : Add FSRCS/FSRS/FZLRM support +- : Add Intel HRESET support +- : Add AVX-VNNI support +- : Add AVX512_FP16 support +- : Add Intel UINTR support +- elf: Do not pass GLRO(dl_platform), GLRO(dl_platformlen) to _dl_important_hwcaps +- elf: Enhance ld.so --help to print HWCAP subdirectories +- elf: Add library search path information to ld.so --help +- sunrpc: Adjust RPC function declarations to match Sun's (bug 26686] +- Avoid GCC 11 -Warray-parameter warnings [BZ #26686]. +- elf: Make __rtld_env_path_list and __rtld_search_dirs global variables +- elf: Print the full name of the dynamic loader in the ld.so help message +- elf: Use the term "program interpreter" in the ld.so help message +- scripts/update-copyrights: Update csu/version.c, elf/dl-usage.c +- elf: Implement ld.so --version +- nptl: Add missing cancellation flags on lockf +- Update mips64 libm-test-ulps +- Update alpha libm-test-ulps +- elf: Implement ld.so --help +- elf: Record whether paths come from LD_LIBRARY_PATH or --library-path +- elf: Move ld.so error/help output to _dl_usage +- elf: Extract command-line/environment variables state from rtld.c + +* Wed Oct 14 2020 Florian Weimer - 2.32.9000-10 +- Disable -Werror on ELN (#1888246) + +* Wed Oct 14 2020 Florian Weimer - 2.32.9000-9 +- Make glibc.spec self-contained (#1887097) + +* Thu Oct 08 2020 Arjun Shankar - 2.32.9000-8 +- Drop glibc-fix-float128-benchtests.patch; applied upstream. +- Auto-sync with upstream branch master, + commit 72d36ffd7db55ae599f4c77feb0eae25a0f3714e: +- elf: Implement __rtld_malloc_is_complete +- __vfscanf_internal: fix aliasing violation (bug 26690) +- Revert "Fix missing redirects in testsuite targets" +- nptl: Add missing cancellation flags on futex_internal and pselect32 +- elf: Implement _dl_write +- elf: Do not search HWCAP subdirectories in statically linked binaries +- Linux: Require properly configured /dev/pts for PTYs +- Linux: unlockpt needs to fail with EINVAL, not ENOTTY (bug 26053) +- login/tst-grantpt: Convert to support framework, more error checking +- posix: Fix -Warray-bounds instances building timer_create [BZ #26687] +- Replace Minumum/minumum with Minimum/minimum +- Optimize scripts/merge-test-results.sh +- Fix GCC 11 -Warray-parameter warning for __sigsetjmp (bug 26647) +- manual: Fix typo +- y2038: nptl: Convert pthread_rwlock_{clock|timed}{rd|wr}lock to support 64 + bit time +- Y2038: nptl: Provide futex_abstimed_wait64 supporting 64 bit time +- sysvipc: Return EINVAL for invalid msgctl commands +- sysvipc: Fix IPC_INFO and MSG_INFO handling [BZ #26639] +- sysvipc: Return EINVAL for invalid semctl commands +- sysvipc: Fix SEM_STAT_ANY kernel argument pass [BZ #26637] +- aarch64: enforce >=64K guard size [BZ #26691] +- sysvipc: Fix semtimedop for Linux < 5.1 for 64-bit ABI +- nptl: futex: Move __NR_futex_time64 alias to beginning of futex-internal.h +- nptl: Provide proper spelling for 32 bit version of futex_abstimed_wait +- string: Fix strerrorname_np return value [BZ #26555] +- Set tunable value as well as min/max values +- ld.so: add an --argv0 option [BZ #16124] +- Reversing calculation of __x86_shared_non_temporal_threshold +- linux: Add time64 recvmmsg support +- linux: Add time64 support for nanosleep +- linux: Consolidate utimes +- linux: Use 64-bit time_t syscall on clock_getcputclockid +- linux: Add time64 sigtimedwait support +- linux: Add time64 select support +- nptl: Fix __futex_abstimed_wait_cancellable32 +- sysvipc: Fix semtimeop for !__ASSUME_DIRECT_SYSVIPC_SYSCALLS +- hurd: add ST_RELATIME +- intl: Handle translation output codesets with suffixes [BZ #26383] +- bench-strcmp.c: Add workloads on page boundary +- bench-strncmp.c: Add workloads on page boundary +- strcmp: Add a testcase for page boundary +- strncmp: Add a testcase for page boundary [BZ #25933] +- Set locale related environment variables in debugglibc.sh +- benchtests: Run _Float128 tests only on architectures that support it +- powerpc: Protect dl_powerpc_cpu_features on INIT_ARCH() [BZ #26615] +- x86: Harden printf against non-normal long double values (bug 26649) +- x86: Use one ldbl2mpn.c file for both i386 and x86_64 +- Define __THROW to noexcept for C++11 and later + +* Mon Sep 21 2020 Arjun Shankar - 2.32.9000-7 +- Adjust glibc-rh741105.patch. +- Add glibc-fix-float128-benchtests.patch to allow building on armv7hl. +- Auto-sync with upstream branch master, + commit cdf645427d176197b82f44308a5e131d69fb53ad: +- Update mallinfo2 ABI, and test +- Allow memset local PLT reference for RISC-V. +- powerpc: fix ifunc implementation list for POWER9 strlen and stpcpy +- nscd: bump GC cycle during cache pruning (bug 26130) +- x86: Use HAS_CPU_FEATURE with IBT and SHSTK [BZ #26625] +- : Add Intel Key Locker support +- Fix handling of collating symbols in fnmatch (bug 26620) +- pselect.c: Pass a pointer to SYSCALL_CANCEL [BZ #26606] +- y2038: nptl: Convert sem_{clock|timed}wait to support 64 bit time +- hurd: Add __x86_get_cpu_features to ld.abilist +- x86: Install [BZ #26124] +- linux: Add time64 pselect support +- linux: Add time64 semtimedop support +- linux: Add ppoll time64 optimization +- linux: Simplify clock_getres +- Update sparc libm-test-ulps +- Remove internal usage of extensible stat functions +- Linux: Consolidate xmknod +- linux: Consolidate fxstatat{64} +- linux: Consolidate fxstat{64} +- linux: Consolidate lxstat{64} +- linux: Consolidate xstat{64} +- linux: Define STAT64_IS_KERNEL_STAT64 +- linux: Always define STAT_IS_KERNEL_STAT +- Update powerpc libm-test-ulps +- benchtests: Add "workload" traces for sinf128 +- benchtests: Add "workload" traces for sinf +- benchtests: Add "workload" traces for sin +- benchtests: Add "workload" traces for powf128 +- benchtests: Add "workload" traces for pow +- benchtests: Add "workload" traces for expf128 +- benchtests: Add "workload" traces for exp +- nptl: futex: Provide correct indentation for part of + __futex_abstimed_wait_cancelable64 + +* Tue Sep 08 2020 DJ Delorie - 2.32.9000-6 +- Auto-sync with upstream branch master, + commit e74b61c09a2a2ab52153e731225ccba5078659b1. +- Disable -Wstringop-overread for some string tests +- string: Fix GCC 11 `-Werror=stringop-overread' error +- C11 threads: Fix inaccuracies in testsuite +- elf.h: Add aarch64 bti/pac dynamic tag constants +- x86: Set CPU usable feature bits conservatively [BZ #26552] + +* Wed Sep 02 2020 Patsy Griffin - 2.32.9000-5 +- Auto-sync with upstream branch master, + commit 86a912c8634f581ea42ec6973553dde7f058cfbf. +- Update i686 ulps. +- Use LFS readdir in generic POSIX getcwd [BZ# 22899] +- linux: Remove __ASSUME_ATFCTS +- Sync getcwd with gnulib +- x86-64: Fix FMA4 detection in ifunc [BZ #26534] +- y2038: nptl: Convert pthread_cond_{clock|timed}wait to support 64 bit time +- malloc: Fix mallinfo deprecation declaration +- x32: Add and regenerate arch-syscall.h +- Add mallinfo2 function that support sizes >= 4GB. +- Remove obsolete default/nss code +- AArch64: Improve backwards memmove performance +- Add RISC-V 32-bit target to build-many-glibcs.py +- Documentation for the RISC-V 32-bit port +- RISC-V: Build infrastructure for 32-bit port +- RISC-V: Add rv32 path to RTLDLIST in ldd +- riscv32: Specify the arch_minimum_kernel as 5.4 +- RISC-V: Fix llrint and llround missing exceptions on RV32 +- RISC-V: Add the RV32 libm-test-ulps +- RISC-V: Add 32-bit ABI lists +- RISC-V: Add hard float support for 32-bit CPUs +- RISC-V: Support the 32-bit ABI implementation +- RISC-V: Add arch-syscall.h for RV32 +- RISC-V: Add path of library directories for the 32-bit +- RISC-V: Support dynamic loader for the 32-bit +- RISC-V: Add support for 32-bit vDSO calls +- RISC-V: Use 64-bit-time syscall numbers with the 32-bit port +- RISC-V: Cleanup some of the sysdep.h code +- RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 +- io/lockf: Include bits/types.h before __OFF_T_MATCHES_OFF64_T check +- elf/tst-libc_dlvsym: Add a TEST_COMPAT around some symbol tests +- hurd: define BSD 4.3 ioctls only under __USE_MISC +- string: test strncasecmp and strncpy near page boundaries +- linux: Simplify utimensat +- linux: Simplify timerfd_settime +- linux: Simplify timer_gettime +- linux: Simplify sched_rr_get_interval +- linux: Simplify ppoll +- linux: Simplify mq_timedsend +- linux: Simplify mq_timedreceive +- linux: Simplify clock_settime +- linux: Simplify clock_nanosleep +- linux: Simplify clock_gettime +- linux: Simplify clock_adjtime +- linux: Add helper function to optimize 64-bit time_t fallback support +- S390: Sync HWCAP names with kernel by adding aliases [BZ #25971] +- [vcstocl] Import ProjectQuirks from its own file +- build-many-glibcs.py: Add a s390x -O3 glibc variant. +- Fix namespace violation in stdio.h and sys/stat.h if build with optimization. [BZ #26376] +- Add C2x BOOL_MAX and BOOL_WIDTH to limits.h. +- Use MPC 1.2.0 in build-many-glibcs.py. +- Add new STATX_* constants from Linux 5.8 to bits/statx-generic.h. +- Correct locking and cancellation cleanup in syslog functions (bug 26100) + +* Thu Aug 20 2020 Carlos O'Donell - 2.32.9000-4 +- Support building glibc in a mock chroot using older systemd-nspawn (#1869030). + +* Tue Aug 18 2020 Carlos O'Donell - 2.32.9000-3 +- Suggest installing minimal localization e.g. C, POSIX, C.UTF-8. + +* Mon Aug 17 2020 DJ Delorie - 2.32.9000-2 +- Auto-sync with upstream branch master, + commit cb7e7a5ca1d6d25d59bc038bdc09630e507c41e5. +- nptl: Handle NULL abstime [BZ #26394] +- Update build-many-glibcs.py for binutils ia64 obsoletion. +- Update kernel version to 5.8 in tst-mman-consts.py. +- y2038: nptl: Convert pthread_{clock|timed}join_np to support 64 bit time +- aarch64: update ulps. + +* Wed Aug 12 2020 Patsy Griffin - 2.32.9000-1 +- Auto-sync with upstream branch master, + commit 0be0845b7a674dbfb996f66cd03d675f0f6028dc: +- S390: Regenerate ULPs. +- manual: Fix sigdescr_np and sigabbrev_np return type (BZ #26343) +- math: Update x86_64 ulps +- math: Regenerate auto-libm-test-out-j0 +- manual: Put the istrerrorname_np and strerrordesc_np return type in braces +- Linux: Use faccessat2 to implement faccessat (bug 18683) +- manual: Fix strerrorname_np and strerrordesc_np return type (BZ #26343) +- math: Fix inaccuracy of j0f for x >= 2^127 when sin(x)+cos(x) is tiny +- Update syscall lists for Linux 5.8. +- Use Linux 5.8 in build-many-glibcs.py. +- htl: Enable tst-cancelx?[45] +- tst-cancel4: Make blocking on write more portable +- hurd: Add missing hidden def +- hurd: Rework sbrk +- hurd: Implement basic sched_get/setscheduler +- x86: Rename Intel CPU feature names +- manual: Fix some @code/@var formatting glitches chapter Date And Time +- Copy regex_internal.h from Gnulib +- Copy regex BITSET_WORD_BITS porting from Gnulib +- Sync regex.h from Gnulib +- Sync mktime.c from Gnulib +- Sync intprops.h from Gnulib +- Open master branch for glibc 2.33 development. + +* Thu Aug 06 2020 Arjun Shankar - 2.32-1 +- Auto-sync with upstream branch release/2.32/master, + commit 3de512be7ea6053255afed6154db9ee31d4e557a: +- Prepare for glibc 2.32 release. +- Regenerate configure scripts. +- Update NEWS with bugs. +- Update translations. +- Don't mix linker error messages into edited scripts +- benchtests/README update. +- RISC-V: Update lp64d libm-test-ulps according to HiFive Unleashed +- aarch64: update NEWS about branch protection +- Add NEWS entry for CVE-2016-10228 (bug 19519) +- powerpc: Fix incorrect cache line size load in memset (bug 26332) +- Update Nios II libm-test-ulps file. + +* Fri Jul 31 2020 Patsy Griffin - 2.31.9000-24 +- Auto-sync with upstream branch master, + commit 7f1a08cff82255cd4252a2c75fd65b80a6a170bf. +- Move NEWS entry for CVE-2020-1751 to the 2.31 section +- NEWS: Deprecate weak libpthread symbols for single-threaded checks +- NEWS: Deprecate nss_hesiod +- nptl: Zero-extend arguments to SETXID syscalls [BZ #26248] +- Use binutils 2.35 branch in build-many-glibcs.py. +- aarch64: Use future HWCAP2_MTE in ifunc resolver +- Update x86-64 libm-test-ulps +- aarch64: Respect p_flags when protecting code with PROT_BTI +- Disable warnings due to deprecated libselinux symbols used by nss and nscd +- Regenerate INSTALL for ARC port updates. +- Update libc.pot for 2.32 release. +- powerpc: Fix POWER10 selection +- powerpc64le: guarantee a .gnu.attributes section [BZ #26220] + +* Wed Jul 29 2020 Florian Weimer - 2.31.9000-23 +- Inherit -mbranch-protection=standard from redhat-rpm-config (for aarch64) + +* Mon Jul 27 2020 Fedora Release Engineering +- Rebuilt for https://fedoraproject.org/wiki/Fedora_33_Mass_Rebuild + +* Wed Jul 22 2020 Carlos O'Donell - 2.31.9000-21 +- Use make macros +- https://fedoraproject.org/wiki/Changes/UseMakeBuildInstallMacro + +* Tue Jul 21 2020 Arjun Shankar - 2.31.9000-20 +- Add glibc-deprecated-selinux-makedb.patch and + glibc-deprecated-selinux-nscd.patch to work around libselinux API + deprecations. +- Drop glibc-rseq-disable.patch; rseq support removed upstream. (#1855729) +- Auto-sync with upstream branch master, + commit ec2f1fddf29053957d061dfe310f106388472a4f: +- libio: Remove __libc_readline_unlocked +- shadow: Implement fgetspent_r using __nss_fgetent_r +- pwd: Implement fgetpwent_r using __nss_fgetent_r +- gshadow: Implement fgetsgent_r using __nss_fgetent_r (bug 20338) +- grp: Implement fgetgrent_r using __nss_fgetent_r +- nss: Add __nss_fgetent_r +- libio: Add fseterr_unlocked for internal use +- nss_files: Use generic result pointer in parse_line +- nss_files: Consolidate line parse declarations in +- nss_compat: Do not use mmap to read database files (bug 26258) +- nss_files: Consolidate file opening in __nss_files_fopen +- Update powerpc-nofpu libm-test-ulps. +- Use MPFR 4.1.0 in build-many-glibcs.py. +- elf: Change TLS static surplus default back to 1664 +- hurd: Fix longjmp check for sigstate +- hurd: Fix longjmp early in initialization +- manual: New signal and errno string functions are AS-safe +- AArch64: Improve strlen_asimd performance (bug 25824) +- Move from sunrpc to inet +- en_US: Minimize changes to date_fmt (Bug 25923) +- Linux: Remove rseq support +- manual: Use Unicode instead HTML entities for characters (bug 19737) +- Add NEWS entry for CVE-2020-6096 (bug 25620) +- arm: remove string/tst-memmove-overflow XFAIL +- AArch64: Rename IS_ARES to IS_NEOVERSE_N1 +- AArch64: Add optimized Q-register memcpy +- AArch64: Align ENTRY to a cacheline +- Correct timespec implementation [BZ #26232] +- Remove --enable-obsolete-rpc configure flag +- hurd: Fix build-many-glibcs.py +- x86: Support usable check for all CPU features +- string: Make tst-strerror/tst-strsignal unsupported if msgfmt is not installed +- malloc: Deprecate more hook-related functionality +- elf: Support at least 32-byte alignment in static dlopen +- x86: Remove __ASSEMBLER__ check in init-arch.h +- x86: Remove the unused __x86_prefetchw +- Documentation for ARC port +- build-many-glibcs.py: Enable ARC builds +- ARC: Build Infrastructure +- ARC: ABI lists +- ARC: Linux Startup and Dynamic Loading +- ARC: Linux ABI +- ARC: Linux Syscall Interface +- ARC: hardware floating point support +- ARC: math soft float support +- ARC: Atomics and Locking primitives +- ARC: Thread Local Storage support +- ARC: startup and dynamic linking code +- ARC: ABI Implementation +- Fix time/tst-cpuclock1 intermitent failures +- powerpc64: Fix calls when r2 is not used [BZ #26173] +- Add NEWS entry for Update to Unicode 13.0.0 [BZ #25819] +- Update i686 libm-test-ulps +- Fix memory leak in __printf_fp_l (bug 26215). +- Fix double free in __printf_fp_l (bug 26214). +- linux: Fix syscall list generation instructions +- sysv: linux: Add 64-bit time_t variant for shmctl +- sysvipc: Remove the linux shm-pad.h file +- sysvipc: Split out linux struct shmid_ds +- sysv: linux: Add 64-bit time_t variant for msgctl +- sysvipc: Remove the linux msq-pad.h file +- sysvipc: Split out linux struct semid_ds +- sysv: linux: Add 64-bit time_t variant for semctl + +* Fri Jul 10 2020 Florian Weimer - 2.31.9000-19 +- Disable rseq registration by default to help Firefox (#1855729) + +* Thu Jul 09 2020 Florian Weimer - 2.31.9000-18 +- Auto-sync with upstream branch master, + commit ffb17e7ba3a5ba9632cee97330b325072fbe41dd: +- rtld: Avoid using up static TLS surplus for optimizations [BZ #25051] +- rtld: Account static TLS surplus for audit modules +- rtld: Add rtld.nns tunable for the number of supported namespaces +- Remove --enable-obsolete-nsl configure flag +- Move non-deprecated RPC-related functions from sunrpc to inet +- aarch64: add NEWS entry about branch protection support +- aarch64: redefine RETURN_ADDRESS to strip PAC +- aarch64: fix pac-ret support in _mcount +- aarch64: Add pac-ret support to assembly files +- aarch64: configure check for pac-ret code generation +- aarch64: ensure objects are BTI compatible +- aarch64: enable BTI at runtime +- aarch64: fix RTLD_START for BTI +- aarch64: fix swapcontext for BTI +- aarch64: Add BTI support to assembly files +- aarch64: Rename place holder .S files to .c +- aarch64: configure test for BTI support +- Rewrite abi-note.S in C. +- rtld: Clean up PT_NOTE and add PT_GNU_PROPERTY handling +- string: Move tst-strsignal tst-strerror to tests-container +- string: Fix prototype mismatch in sigabbrev_np, __sigdescr_np +- arm: CVE-2020-6096: Fix multiarch memcpy for negative length (#1820332) +- arm: CVE-2020-6096: fix memcpy and memmove for negative length (#1820332) +- sunrpc: Remove hidden aliases for global data symbols (bug 26210) +- hurd: Fix strerror not setting errno +- tst-strsignal: fix checking for RT signals support +- hurd: Evaluate fd before entering the critical section +- CVE-2016-10228: Rewrite iconv option parsing (#1428292) +- nss: Remove cryptographic key support from nss_files, nss_nis, nss_nisplus +- sunrpc: Do not export getrpcport by default +- sunrpc: Do not export key handling hooks by default +- sunrpc: Turn clnt_sperrno into a libc_hidden_nolink_sunrpc symbol +- string: Add strerrorname_np and strerrordesc_np +- string: Add sigabbrev_np and sigdescr_np +- string: Add strerror_l on test-strerror-errno +- string: Add strerror, strerror_r, and strerror_l test +- string: Add strsignal test +- string: Simplify strerror_r +- string: Use tls-internal on strerror_l +- string: Implement strerror in terms of strerror_l +- string: Remove old TLS usage on strsignal +- linux: Fix __NSIG_WORDS and add __NSIG_BYTES +- signal: Move sys_errlist to a compat symbol +- signal: Move sys_siglist to a compat symbol +- signal: Add signum-{generic,arch}.h +- Remove most vfprintf width/precision-dependent allocations (bug 14231, bug 26211). +- elf: Do not signal LA_ACT_CONSISTENT for an empty namespace [BZ #26076] +- Fix stringop-overflow errors from gcc 10 in iconv. +- x86: Add thresholds for "rep movsb/stosb" to tunables +- Use C2x return value from getpayload of non-NaN (bug 26073). +- x86: Detect Extended Feature Disable (XFD) +- x86: Correct bit_cpu_CLFSH [BZ #26208] +- manual: Document __libc_single_threaded +- Add the __libc_single_threaded variable +- Linux: rseq registration tests +- Linux: Use rseq in sched_getcpu if available +- Linux: Perform rseq registration at C startup and thread creation +- tst-cancel4: deal with ENOSYS errors +- manual: Show copyright information not just in the printed manual + + +* Thu Jul 02 2020 Carlos O'Donell - 2.31.9000-17 +- Auto-sync with upstream branch master, + commit c6aac3bf3663709cdefde5f5d5e9e875d607be5e. +- Fix typo in comment in bug 26137 fix. +- Fix strtod multiple-precision division bug (bug 26137). +- Linux: Fix UTC offset setting in settimeofday for __TIMESIZE != 64 +- random: range is not portably RAND_MAX [BZ #7003] +- Update kernel version to 5.7 in tst-mman-consts.py. +- powerpc: Add support for POWER10 +- hurd: Simplify usleep timeout computation +- htl: Enable cancel*16 an cancel*20 tests +- hurd: Add remaining cancelation points +- hurd: fix usleep(ULONG_MAX) +- hurd: Make fcntl(F_SETLKW*) cancellation points +- hurd: make wait4 a cancellation point +- hurd: Fix port definition in HURD_PORT_USE_CANCEL +- hurd: make close a cancellation point +- hurd: make open and openat cancellation points +- hurd: clean fd and port on thread cancel +- htl: Move cleanup handling to non-private libc-lock +- htl: Fix includes for lockfile +- htl: avoid cancelling threads inside critical sections +- tst-cancel4-common.c: fix calling socketpair +- x86: Detect Intel Advanced Matrix Extensions +- Set width of JUNGSEONG/JONGSEONG characters from UD7B0 to UD7FB to 0 [BZ #26120] +- S390: Optimize __memset_z196. +- S390: Optimize __memcpy_z196. +- elf: Include (for size_t), in +- nptl: Don't madvise user provided stack +- S390: Regenerate ULPs. +- htl: Add wrapper header for with hidden __sem_post +- elf: Include in because bool is used +- htl: Fix case when sem_*wait is canceled while holding a token +- htl: Make sem_*wait cancellations points +- htl: Simplify non-cancel path of __pthread_cond_timedwait_internal +- htl: Enable tst-cancel25 test +- powerpc: Add new hwcap values +- aarch64: MTE compatible strncmp +- aarch64: MTE compatible strcmp +- aarch64: MTE compatible strrchr +- aarch64: MTE compatible memrchr +- aarch64: MTE compatible memchr +- aarch64: MTE compatible strcpy +- Add MREMAP_DONTUNMAP from Linux 5.7 +- x86: Update CPU feature detection [BZ #26149] + +* Mon Jun 22 2020 DJ Delorie - 2.31.9000-16 +- Auto-sync with upstream branch master, + commit ea04f0213135b13d80f568ca2c4127c2ec112537. +- aarch64: Remove fpu Makefile +- m68k: Use sqrt{f} builtin for coldfire +- arm: Use sqrt{f} builtin +- riscv: Use sqrt{f} builtin +- s390: Use sqrt{f} builtin +- sparc: Use sqrt{f} builtin +- mips: Use sqrt{f} builtin +- alpha: Use builtin sqrt{f} +- i386: Use builtin sqrtl +- x86_64: Use builtin sqrt{f,l} +- powerpc: Use sqrt{f} builtin +- s390x: Use fma{f} builtin +- aarch64: Use math-use-builtins for ceil{f} +- math: Decompose math-use-builtins.h +- hurd: Add mremap +- ia64: Use generic exp10f +- New exp10f version without SVID compat wrapper +- i386: Use generic exp10f +- math: Optimized generic exp10f with wrappers +- benchtests: Add exp10f benchmark + +* Fri Jun 19 2020 Patsy Franklin - 2.31.9000-15 +- Auto-sync with upstream branch master, + commit 27f8864bd41f0f1b61e8e947d9a030b1a0d23df9. +- x86: Update F16C detection [BZ #26133] +- Fix avx2 strncmp offset compare condition check [BZ #25933] +- nptl: Remove now-spurious tst-cancelx9 references +- x86_64: Use %xmmN with vpxor to clear a vector register +- x86: Correct bit_cpu_CLFLUSHOPT [BZ #26128] +- powerpc64le: refactor e_sqrtf128.c +- Update syscall-names.list for Linux 5.7. +- ieee754/dbl-64: Reduce the scope of temporary storage variables +- manual: Add pthread_attr_setsigmask_np, pthread_attr_getsigmask_np +- ld.so: Check for new cache format first and enhance corruption check +- hurd: Fix __writev_nocancel_nostatus +- hurd: Make send* cancellation points +- htl: Enable more cancellation tests +- hurd: Make write and pwrite64 cancellation points +- htl: Fix cleanup support for IO locking +- htl: Move cleanup stack to variable shared between libc and pthread +- htl: initialize first and prevent from unloading +- htl: Add noreturn attribute on __pthread_exit forward +- hurd: Make recv* cancellation points +- powerpc: Automatic CPU detection in preconfigure +- Use Linux 5.7 in build-many-glibcs.py. +- htl: Enable more cancel tests +- htl: Fix linking static tests by factorizing the symbols list +- Add "%d" support to _dl_debug_vdprintf +- aarch64: MTE compatible strlen +- aarch64: MTE compatible strchr +- aarch64: MTE compatible strchrnul +- AArch64: Merge Falkor memcpy and memmove implementations +- hurd: document that gcc&gdb look at the trampoline code +- pthread: Move back linking rules to nptl and htl +- htl: Enable more tests +- htl: Fix registration of atfork handlers in modules +- htl: Fix tls initialization for already-created threads +- hurd: Make read and pread64 cancellable +- hurd: Fix unwinding over interruptible RPC +- htl: Enable but XFAIL tst-flock2, tst-signal1, tst-signal2 +- hurd: XFAIL more tests that require setpshared support +- hurd: Briefly document in xfails the topics of the bugzilla entries +- htl: Enable more tests +- htl: Add sem_clockwait support +- htl: fix register-atfork ordering +- hurd: Fix hang in _hurd_raise_signal from pthread_kill +- hurd: Reject raising invalid signals +- hurd: fix clearing SS_ONSTACK when longjmp-ing from sighandler +- hurd: Add pointer guard support +- hurd: Add stack guard support +- dl-runtime: reloc_{offset,index} now functions arch overide'able +- powerpc64le: add optimized strlen for P9 +- powerpc64le: use common fmaf128 implementation + +* Fri Jun 05 2020 Patsy Griffin - 2.31.9000-14 +- Auto-sync with upstream branch master, + commit e52434a2e4d1105272daaef87678da950fbec73f. +- benchtests: Restore the clock_gettime option +- Update HP_TIMING_NOW for _ISOMAC in sysdeps/generic/hp-timing.h +- Replace val with __val in TUNABLE_SET_VAL_IF_VALID_RANGE +- support: Fix detecting hole support on >2KB-block filesystems +- powerpc: Fix powerpc64le due a7a3435c9a +- manual/jobs.texi: remove unused var from example code +- powerpc/fpu: use generic fma functions +- aarch/fpu: use generic builtins based math functions +- ieee754: provide gcc builtins based generic fma functions +- ieee754: provide gcc builtins based generic sqrt functions +- Linux: Use __pthread_attr_setsigmask_internal for timer helper thread +- nptl: Add pthread_attr_setsigmask_np, pthread_attr_getsigmask_np +- nptl: Make pthread_attr_t dynamically extensible +- nptl: Destroy the default thread attribute as part of freeres +- nptl: Change type of __default_pthread_attr +- nptl: Use __pthread_attr_setaffinity_np in pthread_getattr_np +- nptl: Use __pthread_getattr_default_np in pthread_create +- nptl: Add internal alias __pthread_getattr_default_np +- htl: Fix gsync_wait symbol exposition +- htl: Make pthread_cond_destroy wait for threads to be woken +- htl: Enable more cond tests +- tst-cond11: Fix build with _SC_MONOTONIC_CLOCK > 0 +- mbstowcs: Document, test, and fix null pointer dst semantics (Bug 25219) +- build: Use FAIL_EXIT1 () on failure to exec child [BZ #23990] +- manual: Fix backtraces code example [BZ #10441] +- hurd: Fix fexecve +- i386: Remove unused file sysdeps/unix/i386/sysdep.S +- hurd: fix ptsname error when called on a non-tty +- hurd: Fix fdopendir checking for directory type +- i386: Remove NO_TLS_DIRECT_SEG_REFS handling +- Hurd: Move internals into wrapper header +- Hurd: Use __sigmask in favor of deprecated sigmask +- hurd: Fix pselect atomicity +- elf: Remove extra hwcap mechanism from ldconfig +- elf: Do not read hwcaps from the vDSO in ld.so +- linux: Use internal DIR locks when accessing filepos on telldir +- Update i386 libm-test-ulps +- htl: Add clock variants +- signal: Deprecate additional legacy signal handling functions +- elf: Turn _dl_printf, _dl_error_printf, _dl_fatal_printf into functions +- x86: Update Intel Atom processor family optimization +- elf.h: add aarch64 property definitions +- elf.h: Add PT_GNU_PROPERTY +- : Add libpthread hidden alias support +- nptl: Use __pthread_attr_copy in pthread_setattr_default_np +- nptl: Use __pthread_attr_copy in pthread_getattr_default_np (bug 25999) +- nptl: Add __pthread_attr_copy for copying pthread_attr_t objects +- nptl: Make __pthread_attr_init, __pthread_attr_destroy available internally +- nptl: Move pthread_gettattr_np into libc +- nptl: Move pthread_getaffinity_np into libc +- nptl: Move pthread_attr_setaffinity_np into libc +- nptl: Replace some stubs with the Linux implementation +- Linux: Add missing handling of tai field to __ntp_gettime64 +- Mention GCC 10 attribute access. +- y2038: Replace __clock_gettime with __clock_gettime64 +- manual: Add missing section and node for clockid_t wait functions +- y2038: linux: Provide __ntp_gettimex64 implementation +- y2038: linux: Provide __ntp_gettime64 implementation +- y2038: Provide conversion helpers for struct __ntptimeval64 +- y2038: Introduce struct __ntptimeval64 - new internal glibc type +- y2038: linux: Provide __adjtime64 implementation +- y2038: linux: Provide ___adjtimex64 implementation +- y2038: linux: Provide __clock_adjtime64 implementation +- ldconfig: Default to the new format for ld.so.cache +- nss_compat: internal_end*ent may clobber errno, hiding ERANGE [BZ #25976] +- powerpc: Optimized rawmemchr for POWER9 +- x86: Add --enable-cet=permissive +- Remove NO_CTORS_DTORS_SECTIONS macro +- elf: Assert that objects are relocated before their constructors run +- powerpc: Optimized stpcpy for POWER9 +- powerpc: Optimized strcpy for POWER9 +- x86: Move CET control to _dl_x86_feature_control [BZ #25887] +- sunrpc/tst-udp-*: Fix timeout value +- Linux: Remove remnants of the getcpu cache +- Update timezone code from tzcode 2020a +- aarch64: fix strcpy and strnlen for big-endian [BZ #25824] +- locale: Add transliteration for Geresh, Gershayim (U+05F3, U+05F4) +- string: Fix string/tst-memmove-overflow to compile with GCC 7 +- Add arch-syscall.h dependency for generating sysd-syscalls file +- arm: XFAIL string/tst-memmove-overflow due to bug 25620 +- elf: Remove redundant add_to_global_resize_failure call from dl_open_args +- string: Add string/tst-memmove-overflow, a test case for bug 25620 +- support: Add support_blob_repeat_allocate_shared +- nptl: wait for pending setxid request also in detached thread (bug 25942) +- aarch64: Accept PLT calls to __getauxval within libc.so +- Use unsigned constants for ICMP6 filters [BZ #22489] +- Linux: Enhance glibcsyscalls.py to support listing system calls + +* Mon May 11 2020 DJ Delorie - 2.31.9000-13 +- Auto-sync with upstream branch master, + commit 978e8ac39f8ba2d694031e521511da1ae803ccfc. +- Suppress GCC 10 true positive warnings [BZ #25967] +- POWER: Add context-synchronizing instructions to pkey_write [BZ #25954] +- hurd: Add missing sigstate members initialization +- x86-64: Use RDX_LP on __x86_shared_non_temporal_threshold [BZ #25966] +- linux: Remove assembly umount2 implementation +- signal: Use for sigemptyset, sigfillset +- ckb_IQ, or_IN locales: Add missing reorder-end keywords +- semaphore: consolidate arch headers into a generic one +- Use GCC 10 branch in build-many-glibcs.py. +- Document the internal _ and N_ macros +- y2038: Provide conversion helpers for struct __timex64 +- y2038: Introduce struct __timex64 - new internal glibc type +- y2038: include: Move struct __timeval64 definition to a separate file +- y2038: nscd: Modify nscd_helper to use __clock_gettime64 +- y2038: inet: Convert inet deadline to support 64 bit time +- y2038: hurd: Provide __clock_gettime64 function +- y2038: Export __clock_gettime64 to be usable in other libraries +- manual: Document the O_NOFOLLOW open flag +- powerpc64le/power9: guard power9 strcmp against rtld usage [BZ# 25905] +- float128: use builtin_signbitf128 always +- improve out-of-bounds checking with GCC 10 attribute access [BZ #25219] +- nios2: delete sysdeps/unix/sysv/linux/nios2/kernel-features.h +- powerpc: Rename argN to _argN in LOADARGS_N [BZ #25902] +- locale/tst-localedef-path-norm: Don't create $(complocaledir) +- support: Set errno before testing it. +- localedef: Add tests-container test for --no-hard-links. +- test-container: Support $(complocaledir) and mkdirp. +- i386: Remove unused variable in sysdeps/x86/cacheinfo.c +- Add a C wrapper for prctl [BZ #25896] +- powerpc64le: Enable support for IEEE long double +- powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0) +- powerpc64le: bump binutils version requirement to >= 2.26 +- powerpc64le: raise GCC requirement to 7.4 for long double transition +- Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI +- ldbl-128ibm-compat: workaround GCC 9 C++ PR90731 +- x86: Add the test case of __get_cpu_features support for Zhaoxin processors +- x86: Add cache information support for Zhaoxin processors +- x86: Add CPU Vendor ID detection support for Zhaoxin processors +- Update translations +- Add C wrappers for process_vm_readv/process_vm_writev [BZ #25810] +- generic/typesizes.h: Add support for 32-bit arches with 64-bit types +- semctl: Remove the sem-pad.h file +- bits/sem.h: Split out struct semid_ds +- Mark unsigned long arguments with U in more syscalls [BZ #25810] +- elf: Add initial flag argument to __libc_early_init +- Add SYSCALL_ULONG_ARG_[12] to pass long to syscall [BZ #25810] +- Makeconfig: Use $(error ...) to output error message +- manual: Fix typos in the fexecve description +- misc: Remove sstk from the autogenerated system call list +- Remove unused floating-point configuration from gmp-impl.h. +- support: Implement key create/delete +- nptl/tst-setuid1-static: Improve isolation from system objects +- Increase the timeout of locale/tst-localedef-path-norm +- Use 2020 as copyright year. +- misc: Turn sstk into a compat symbol +- manual: Document the fexecve function +- nptl: Start new threads with all signals blocked [BZ #25098] +- localedef: Add verbose messages for failure paths. +- Remove most gmp-mparam.h headers. +- elf: Implement __libc_early_init +- elf: Introduce +- Add a syscall test for [BZ #25810] +- elf: Support lld-style link map for librtld.map +- signal: Only handle on NSIG signals on signal functions (BZ #25657) +- linux: Use pthread_sigmask on sigprocmask +- ia64: Remove sigprocmask/sigblock objects from libpthread +- nptl: Move pthread_sigmask implementation to libc +- Bug 25819: Update to Unicode 13.0.0 + +* Wed Apr 29 2020 Florian Weimer - 2.31.9000-12 +- nss_db.x86_64 should install nss_db.i686 if glibc.i686 is installed (#1807821) +- Likewise for nss_hesiod. + +* Mon Apr 27 2020 Florian Weimer - 2.31.9000-11 +- Introduce glibc-headers-x86, glibc-headers-s390 packages (#1828332) +- Remove the glibc-headers package + +* Mon Apr 20 2020 DJ Delorie - 2.31.9000-10 +- Auto-sync with upstream branch master, + commit 0798b8ecc8da8667362496c1217d18635106c609. +- ARC: Update syscall-names.list for ARC specific syscalls +- Revert "x86_64: Add SSE sfp-exceptions" +- provide y2038 safe socket constants for default/asm-generic ABI +- x86_64: Add SSE sfp-exceptions +- Remove __NO_MATH_INLINES +- i686: Add INTERNAL_SYSCALL_NCS 6 argument support +- Reset converter state after second wchar_t output (Bug 25734) +- Fix typo in posix/tst-fnmatch.input (Bug 25790) + +* Wed Apr 15 2020 Patsy Griffin - 2.31.9000-9 +- Auto-sync with upstream branch master, + commit 076f09afbac1aa57756faa7a8feadb7936a724e4. +- Linux: Remove and the sysctl function +- posix: Add wait4 test case +- linux: wait4: Fix incorrect return value comparison +- hurd: add mach_print function +- x32: Properly pass long to syscall [BZ #25810] +- Add GRND_INSECURE from Linux 5.6 to sys/random.h +- Update kernel version to 5.6 in tst-mman-consts.py. + +* Wed Apr 15 2020 Florian Weimer - 2.31.9000-8 +- nsswitch.conf: don't add sss to shadow line + +* Wed Apr 08 2020 Carlos O'Donell - 2.31.9000-7 +- Auto-sync with upstream branch master, + commit b1caa144c74678097cada5a54eda2996bb459d8f. +- Update mips libm-test-ulps +- Update alpha libm-test-ulps +- Update ia64 libm-test-ulps +- Update sparc libm-test-ulps +- Update arm libm-test-ulps +- Update aarch64 libm-test-ulps +- Updates to the shn_MM locale [BZ #25532] +- powerpc: Update ULPs and xfail more ibm128 outputs +- i386: Remove build support for GCC older than GCC 6 +- oc_FR locale: Fix spelling of April (bug 25639) +- Update hppa libm-test-ulps +- y2038: linux: Provide __mq_timedreceive_time64 implementation +- y2038: linux: Provide __mq_timedsend_time64 implementation +- y2038: include: Move struct __timespec64 definition to a separate file +- malloc: ensure set_max_fast never stores zero [BZ #25733] +- powerpc64le: enforce non-specific long double in .gnu.attributes section +- powerpc64le: workaround ieee long double / _Float128 stdc++ bug +- powerpc64le: Enforce -mabi=ibmlongdouble when -mfloat128 used +- powerpc64le/multiarch: don't generate strong aliases for fmaf128-ppc64 +- ldbl-128ibm: simplify iscanonical.h +- i386: Disable check_consistency for GCC 5 and above [BZ #25788] +- Add IPPROTO_ETHERNET and IPPROTO_MPTCP from Linux 5.6 to netinet/in.h. +- Update syscall lists for Linux 5.6. +- elf: Implement DT_AUDIT, DT_DEPAUDIT support [BZ #24943] +- elf: Simplify handling of lists of audit strings +- support: Change xgetline to return 0 on EOF +- nptl: Remove x86_64 cancellation assembly implementations [BZ #25765] +- aarch64: update bits/hwcap.h +- Add tests for Safe-Linking +- S390: Regenerate ULPs. +- sysv/alpha: Use generic __timeval32 and helpers +- linux: Use long time_t for wait4/getrusage +- resource: Add a __rusage64 struct +- linux: Use long time_t __getitimer/__setitimer +- sysv: Define __KERNEL_OLD_TIMEVAL_MATCHES_TIMEVAL64 +- math: Add inputs that yield larger errors for float type (x86_64) + +* Tue Mar 31 2020 DJ Delorie - 2.31.9000-6 +- Auto-sync with upstream branch master, + commit 49c3c37651e2d2ec4ff8ce21252bbbc08a9d6639. +- Fix alignment bug in Safe-Linking +- Typo fixes and CR cleanup in Safe-Linking +- Use Linux 5.6 and GMP 6.2.0 in build-many-glibcs.py. +- Add new file missed in previous hppa commit. +- powerpc: Add support for fmaf128() in hardware +- Fix data race in setting function descriptors during lazy binding on hppa. +- sparc: Move __fenv_{ld,st}fsr to fenv-private.h +- x86: Remove feraiseexcept optimization +- math: Remove fenvinline.h +- hurd: Make O_TRUNC update mtime/ctime +- Add Safe-Linking to fastbins and tcache +- Add benchtests for roundeven and roundevenf. +- time: Add a __itimerval64 struct +- time: Add a timeval with a 32-bit tv_sec and tv_usec +- sysv/linux: Rename alpha functions to be alpha specific +- ARC: add definitions to elf/elf.h +- powerpc64: apply -mabi=ibmlongdouble to special files +- powerpc64le: add -mno-gnu-attribute to *f128 objects and difftime +- Makeconfig: sandwich gnulib-tests between libc/ld linking of tests +- powerpc64le: Ensure correct ldouble compiler flags are used +- Fix tests which expose ldbl -> _Float128 redirects +- ldbl-128ibm-compat: PLT redirects for using ldbl redirects internally + +* Wed Mar 25 2020 Patsy Franklin - 2.31.9000-5 +- Auto-sync with upstream branch master, + commit 4eda036f5b897fa8bc20ddd2099b5a6ed4239dc9. +- stdlib: Move tst-system to tests-container +- support/shell-container.c: Add builtin kill +- support/shell-container.c: Add builtin exit +- support/shell-container.c: Return 127 if execve fails +- Add NEWS entry for CVE-2020-1751 (bug 25423) +- posix: Fix system error return value [BZ #25715] +- y2038: fix: Add missing libc_hidden_def attribute for some syscall wrappers +- Extended Char Intro: Use getwc in example (Bug 25626) +- stdio: Add tests for printf multibyte convertion leak [BZ#25691] +- stdio: Remove memory leak from multibyte convertion [BZ#25691] +- Add NEWS entry for CVE-2020-1752 (bug 25414) +- math: Remove inline math tests +- Remove __LIBC_INTERNAL_MATH_INLINES +- math: Remove mathinline +- m68k: Remove mathinline.h +- oc_FR locale: Fix spelling of Thursday (bug 25639) +- x86: Remove ARCH_CET_LEGACY_BITMAP [BZ #25397] +- Fix build with GCC 10 when long double = double. +- nscd/cachedumper.c : fix whitespace +- Fix nscd/cachedumper.c compile errors +- manual: Fix inconsistent declaration of wcsrchr [BZ #24655] +- nscd: add cache dumper + +* Fri Mar 13 2020 Patsy Franklin - 2.31.9000-4 +- Auto-sync with upstream branch master, + commit 2de7fe62534b7a6461c633114f03e9dff394f5f7. +- parse_tunables: Fix typo in comment +- ldconfig: trace origin paths with -v +- test-container: print errno when execvp fails +- [AArch64] Improve integer memcpy +- Add NEWS entry for CVE-2020-10029 (bug 25487) +- gcc PR 89877: miscompilation due to missing cc clobber in longlong.h macros +- mips: Fix wrong INTERNAL_SYSCALL_ERROR_P check from bc2eb9321e +- elf: Fix wrong indentation from commit eb447b7b4b +- y2038: linux: Provide __futimesat64 implementation +- y2038: linux: Provide __lutimes64 implementation +- y2038: linux: Provide __futimes64 implementation +- y2038: fix: Add missing libc_hidden_def for __futimens64 +- sparc: Move sigreturn stub to assembly +- ldbl-128ibm: Let long double files have specific compiler flags +- ldbl-128ibm-compat: Add tests for IBM long double functions +- powerpc: Fix feraiseexcept and feclearexcept macros +- arm: Fix softp-fp Implies (BZ #25635) +- Remove reference of --without-fp on configure +- linux/sysipc: Include linux/posix_types.h for __kernel_mode_t +- Improve IFUNC check [BZ #25506] +- linux: Clear mode_t padding bits (BZ#25623) +- linux: Remove aarch64 ipc_priv.h +- Linux: Use __fstatat64 in fchmodat implementation +- Linux: Use AT_FDCWD in utime, utimes when calling utimensat +- S390: Remove backchain-based fallback and use generic backtrace.c. +- manual: Fix wrong declaration of wcschr [BZ #24654] +- manual: Fix typo in parse_printf_format example [BZ #24638] + +* Thu Mar 5 2020 Florian Weimer - 2.31.9000-3 +- Emergency patch for broken utimes/utime functions + +* Tue Mar 03 2020 Patsy Franklin - 2.31.9000-2 +- Auto-sync with upstream branch master, + commit 78c9d0c6efabe2067ef7f93cd36325f54c60adc2. +- Update translations +- Convert Python scripts to Python 3 +- alpha: Do not build with -fpic +- y2038: linux: Provide __utime64 implementation +- y2038: linux: Provide __utimes64 implementation +- y2038: Introduce struct __utimbuf64 - new internal glibc type +- microblaze: vfork is always available +- m68k: getpagesize syscall number is always available +- Linux: epoll_pwait syscall number is always available +- x86_64: Do not define __NR_semtimedop in +- ia64: Do not define __NR_semtimedop in +- Linux: open_by_handle_at syscall number is always available +- Linux: set_robust_list syscall number is always available +- Linux: pciconfig_iobase syscall number is always available on alpha +- Linux: getdents64 syscall number is always available on MIPS +- Linux: Clean up preadv2, pwritev2 system call names +- Linux: exit_group syscall number is always available +- Linux: set_tid_address syscall number is always available +- Linux: pkey_mprotect syscall number is always available +- Linux: rt_sigqueueinfo syscall number is always available +- Linux: getrandom syscall number is always available +- Linux: Clean up preadv, pwritev system call names +- Linux: Clean up pread64/pwrite64 system call names +- Linux: sigaltstack syscall number is always available +- Linux: sched_getaffinity syscall number is always available +- Linux: sched_setaffinity syscall number is always available +- Linux: statx syscall number is always available +- Linux: mq_* syscall numbers are always available +- Linux: mlock2 syscall number is always available +- Linux: copy_file_range syscall number is always available +- Linux: renameat2 syscall number is always available +- build-many-glibcs.py: Add list-compilers, list-glibcs commands +- build-many-glibcs.py: Add --shallow option +- Fixed typo in run_command_array() in support/shell-container.c +- Add missing libc_hidden_def for __utimensat64 +- elf: Add elf/check-wx-segment, a test for the presence of WX segments +- i386: Use comdat instead of .gnu.linkonce for i386 setup pic register (BZ #20543) +- ldbl-128ibm-compat: link tst-ldbl-efgcvt against loader too +- ldbl-128ibm-compat: enforce ibm128 on compat tests +- ldbl-128ibm-compat: Provide nexttoward functions +- ldbl-128ibm-compat: Provide a significand implementation +- ldbl-128ibm-compat: Redirect complex math functions +- ldbl-128ibm-compat: Redirect long double functions to f128/ieee128 functions +- posix: Remove posix waitid +- posix: Refactor tst-waitid (BZ #14666) +- support: Add support_process_state_wait +- malloc/tst-mallocfork2: Kill lingering process for unexpected failures + +* Wed Feb 26 2020 Patsy Franklin - 2.31.9000-1 +- Auto-sync with upstream branch master, + commit 758599bc9dcc5764e862bd9e1613c5d1e6efc5d3. +- elf: Apply attribute_relro to pointers in elf/dl-minimal.c +- powerpc: Refactor fenvinline.h +- nss_nis: Use NSS_DECLARE_MODULE_FUNCTIONS +- csu: Use ELF constructor instead of _init in libc.so +- ldbl-128ibm: make ieee754.h work with IEEE 128 long double +- ldbl-128ibm-compat: fixup subdir location of several funcs +- ldbl-128ibm-compat: enforce correct abi flags on internal file +- ldbl-128ibm-compat: Provide ieee128 symbols to narrow functions +- Undefine redirections after long double definition on __LDBL_COMPAT [BZ #23294] +- nios2: Fix Linux kABI for syscall return +- Fix use-after-free in glob when expanding ~user (bug 25414) +- nptl: Move pthread_setschedparam implementation into libc +- nptl: Move pthread_getschedparam implementation into libc +- Add hidden prototypes for __sched_getparam, __sched_getscheduler +- nptl: Move pthread_cond_init implementation into libc +- nptl: Move pthread_cond_destroy implementation into libc +- nptl: Move pthread_condattr_init implementation into libc +- nptl: Move pthread_condattr_destroy implementation into libc +- nptl: Move pthread_attr_setscope implementation into libc +- nptl: Move pthread_attr_getscope implementation into libc +- nptl: Move pthread_attr_setschedpolicy implementation into libc +- nptl: Move pthread_attr_getschedpolicy implementation into libc +- nptl: Sort routines list in Makefile alphabetically +- nptl: Use .NOTPARALLEL in Makefile only if actually running tests +- Block all signals on timer_create thread (BZ#10815) +- Fix tst-pkey expectations on pkey_get [BZ #23202] +- y2038: linux: Provide __gettimeofday64 implementation +- Linux: Work around kernel bugs in chmod on /proc/self/fd paths [BZ #14578] +- Introduce and ELF_INITFINI for all architectures +- mips: Fix bracktrace result for signal frames +- Move implementation of into a C file +- : Add type safety and port to Hurd +- Prepare redirections for IEEE long double on powerpc64le +- conform/conformtest.py: Extend tokenizer to cover character constants +- stdlib: Reduce namespace pollution in +- x86: Avoid single-argument _Static_assert in +- x86 tls: Use _Static_assert for TLS access size assertion +- htl: Link internal htl tests against libpthread +- pthread: Fix building tst-robust8 with nptl +- pthread: Move robust mutex tests from nptl to sysdeps/pthread +- htl: Remove stub warning for pthread_mutexattr_setpshared +- htl: Add missing functions and defines for robust mutexes +- htl: Only check pthread_self coherency when DEBUG is set +- hurd: Add THREAD_GET/SETMEM/_NC +- hurd tls: update comment about fields at the end of tcbhead +- ld.so: Do not export free/calloc/malloc/realloc functions [BZ #25486] +- Remove weak declaration of free from +- elf: Extract _dl_sym_post, _dl_sym_find_caller_map from elf/dl-sym.c +- elf: Introduce the rtld-stubbed-symbols makefile variable +- arm: fix use of INTERNAL_SYSCALL_CALL +- linux: Remove INTERNAL_SYSCALL_DECL +- nptl: Remove ununsed pthread-errnos.h rule +- linux: Consolidate INLINE_SYSCALL +- s390: Consolidate Linux syscall definition +- riscv: Avoid clobbering register parameters in syscall +- microblaze: Avoid clobbering register parameters in syscall +- nios2: Use Linux kABI for syscall return +- mips: Use Linux kABI for syscall return +- mips64: Consolidate Linux sysdep.h +- ia64: Use Linux kABI for syscall return +- alpha: Refactor syscall and Use Linux kABI for syscall return +- sparc: Avoid clobbering register parameters in syscall +- sparc: Use Linux kABI for syscall return +- powerpc: Use Linux kABI for syscall return +- powerpc: Consolidate Linux syscall definition +- i386: Enable CET support in ucontext functions +- tst-clone3: Use __NR_futex_time64 if we don't have __NR_futex +- powerpc64: Add memory protection key support [BZ #23202] +- ldbl-128ibm-compat: Provide a scalb implementation +- Add a generic scalb implementation +- Adjust thresholds in Bessel function implementations (bug 14469). +- resolv: Fix ABA race in /etc/resolv.conf change detection [BZ #25420] +- resolv: Enhance __resolv_conf_load to capture file change data +- resolv: Fix file handle leak in __resolv_conf_load [BZ #25429] +- resolv: Use in __resolv_conf_get_current +- Add STATX_ATTR_VERITY from Linux 5.5 to bits/statx-generic.h. +- Use gcc -finput-charset=ascii for check-installed-headers. +- math/test-sinl-pseudo: Use stack protector only if available +- alpha: Fix static gettimeofday symbol +- nss_nisplus: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss_dns: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss_files: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss_db: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss_compat: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss_hesiod: Use NSS_DECLARE_MODULE_FUNCTIONS +- nss: Add function types and NSS_DECLARE_MODULE_FUNCTIONS macro to +- nss_compat: Do not use nss_* names for function pointers +- Avoid ldbl-96 stack corruption from range reduction of pseudo-zero (bug 25487). +- mips: Fix argument passing for inlined syscalls on Linux [BZ #25523] +- mips: Use 'long int' and 'long long int' in linux syscall code +- alpha: Use generic gettimeofday implementation +- sunrpc: Properly clean up if tst-udp-timeout fails +- elf: avoid stack allocation in dl_open_worker +- elf: avoid redundant sort in dlopen +- elf: Allow dlopen of filter object to work [BZ #16272] +- Update translations +- Rename RWF_WRITE_LIFE_NOT_SET to RWH_WRITE_LIFE_NOT_SET following Linux 5.5. +- S390: Fix non-ascii character in fenv.h. +- io: Add io/tst-lchmod covering lchmod and fchmodat +- Linux: Emulate fchmodat with AT_SYMLINK_NOFOLLOW using O_PATH [BZ #14578] +- io: Implement lchmod using fchmodat [BZ #14578] +- Add internal header file +- elf.h: Add R_RISCV_IRELATIVE +- Fix typo in the name for Wednesday in Kurdish [BZ #9809] +- debug: Add missing locale dependencies of fortify tests +- htl C11 threads: Avoid pthread_ symbols visibility in static library +- hurd: Add __pthread_spin_wait and use it +- ldbl-128ibm-compat: set PRINTF_CHK flag in {,v}sprintf_chk +- Use --disable-gdbserver in build-many-glibcs.py. +- Improve random memcpy benchmark +- nptl: update default pthread-offsets.h +- nptl: add missing pthread-offsets.h +- htl: Avoid a local plt for pthread_self +- pthread: Move some join tests from nptl to sysdeps/pthread +- htl: Make joining self return EDEADLK +- pthread: Move most barrier tests from nptl to sysdeps/pthread +- htl: Fix barrier_wait with one thread +- pthread: Move most sem tests from nptl to sysdeps/pthread +- htl: Make sem_wait/sem_timedwait interruptible +- htl: Make sem_open return ENOSYS +- htl: Add support for semaphore maximum value +- pthread: Move key tests from nptl to sysdeps/pthread +- hurd: Make nanosleep a cancellation point +- htl: Add support for libc cancellation points +- htl: clean __pthread_get_cleanup_stack hidden proto +- htl: XFAIL rwlock tests which need pshared support +- pthread: Move some rwlock tests from nptl to sysdeps/pthread +- pthread: Move most once tests from nptl to sysdeps/pthread +- htl: support cancellation during pthread_once +- pthread: Move most cond tests from nptl to sysdeps/pthread +- htl: make pthread_cond_destroy return EBUSY on waiters +- htl: Report missing mutex lock on pthread_cond_*wait +- htl: Fix linking static testcases +- htl: Move __register_atfork from forward to own file +- pthread: Move some attr tests from nptl to sysdeps/pthread +- htl: Fix default guard size +- pthread: Move most mutex tests from nptl to sysdeps/pthread +- pthread: Move spin tests from nptl to sysdeps/pthread +- htl: make pthread_spin_lock really spin +- htl: Avoid check-installed-headers looking at inlines +- htl: Do not put spin_lock inlines in public headers +- pthread: Move basic tests from nptl to sysdeps/pthread +- htl: Fix calling pthread_exit in the child of a fork +- x86: Remove and use the generic version +- C11 threads: Move implementation to sysdeps/pthread +- htl: Add C11 threads types definitions +- C11 threads: make thrd_join more portable +- C11 threads: Fix thrd_t / pthread_t compatibility assertion +- C11 threads: do not require PTHREAD_DESTRUCTOR_ITERATIONS +- nptl: Move nptl-specific types to separate header +- htl: Make __PTHREAD_ONCE_INIT more flexible +- htl: Add support for C11 threads behavior +- htl: Add missing internal functions declarations +- htl: Rename _pthread_mutex_init/destroy to __pthread_mutex_init/destroy +- htl: Move internal mutex/rwlock symbols to GLIBC_PRIVATE +- Linux: Add io/tst-o_path-locks test +- support: Add the xlstat function +- htl: Remove duplicate files +- htl: Remove unused files +- resolv: Fix CNAME chaining in resolv/tst-resolv-ai_idn-common.c +- Remove a comment claiming that sin/cos round correctly. +- y2038: linux: Provide __settimeofday64 implementation +- y2038: Provide conversion helpers for struct __timeval64 +- y2038: alpha: Rename valid_timeval64_to_timeval to valid_timeval_to_timeval32 +- y2038: alpha: Rename valid_timeval_to_timeval64 to valid_timeval32_to_timeval +- y2038: Introduce struct __timeval64 - new internal glibc type +- y2038: Define __suseconds64_t type to be used with struct __timeval64 +- Update kernel version to 5.5 in tst-mman-consts.py. +- Update syscall lists for Linux 5.5. +- NEWS: Set fill-column hint to 72 +- y2038: linux: Provide __timespec_get64 implementation +- Use binutils 2.34 branch in build-many-glibcs.py. +- Run nptl/tst-pthread-getattr in a container +- test-container: add exec, cwd +- Use Linux 5.5 in build-many-glibcs.py. +- rt: avoid PLT setup in timer_[sg]ettime +- Update or_IN collation [BZ #22525] +- Fix ckb_IQ [BZ #9809] +- Add new locale: ckb_IQ (Kurdish/Sorani spoken in Iraq) [BZ #9809] +- list-fixed-bugs.py: Wrap at 72 chars +- y2038: linux: Provide __sched_rr_get_interval64 implementation +- y2038: linux: Provide __timerfd_settime64 implementation +- y2038: linux: Provide __timerfd_gettime64 implementation +- i386: Remove _exit.S +- i386: Use ENTRY/END in assembly codes +- i386-mcount.S: Add _CET_ENDBR to _mcount and __fentry__ +- i386/sub_n.S: Add a missing _CET_ENDBR to indirect jump target +- i386: Don't unnecessarily save and restore EAX, ECX and EDX [BZ# 25262] +- x86: Don't make 2 calls to dlerror () in a row +- Open master for 2.32 development + +* Mon Feb 03 2020 DJ Delorie - 2.31-1 +- Auto-sync with upstream branch release/2.31/master, + commit 9ea3686266dca3f004ba874745a4087a89682617. +- glibc 2.31 release +- Generate ChangeLog.old/ChangeLog.20 for 2.31 +- Add bugs fixed in 2.31 in NEWS +- Update newest tested versions of dependencies in install.texi +- Add more contributors to the manual +- Add note to NEWS about kernel headers dependency on risc-v +- Add Portuguese (Portugal) translation +- Add NEWS entry about 64-bit time_t syscall use on 32-bit targets +- nptl: Avoid using PTHREAD_MUTEX_DEFAULT in macro definition [BZ #25271] + +* Thu Jan 30 2020 Patsy Franklin - 2.30.9000-33 +- Auto-sync with upstream branch master, + commit 352bb99754ae7c83ff1b974f9c52244e974c9410. +- Build raise with -fasynchronous-unwind-tables. +- Fix locale/tst-locale-locpath cross-testing when sshd sets LANG. +- Fix elf/tst-rtld-preload cross-testing. +- Fix cross-testing of tst-ifunc-fault-* tests. +- gitlog-to-changelog: Drop scripts in favour of gnulib version +- Add NEWS entry about the change in handling of PT_GNU_STACK on MIPS +- Fix array overflow in backtrace on PowerPC (bug 25423) +- getaddrinfo: Fix resource leak after strdup failure in gethosts (swbz#25425) + +* Tue Jan 28 2020 Fedora Release Engineering +- Rebuilt for https://fedoraproject.org/wiki/Fedora_32_Mass_Rebuild + +* Mon Jan 20 2020 Patsy Franklin - 2.30.9000-31 +- Auto-sync with upstream branch master, + commit 92ce43eef7ac844782d50a8015d977d216fbadec. +- Run bench-timing-type with newly built libc. +- Get rid of Werror=maybe-uninitialized in res_send.c. +- translations: Update translations +- translations: Trim po files using msgattrib +- Update translations +- translations: Run msgmerge when downloading translations +- Fix maybe-uninitialized error on powerpc +- powerpc32: Fix syntax error in __GLRO macro +- Remove incorrect alloc_size attribute from pvalloc (swbz#25401) + +* Fri Jan 17 2020 Florian Weimer - 2.30.9000-30 +- Auto-sync with upstream branch master, + commit 70ba28f7ab2923d4e36ffc9d5d2e32357353b25c: +- Fix tst-pkey.c pkey_alloc return checks and manual +- powerpc: Move cache line size to rtld_global_ro +- powerpc: Initialize rtld_global_ro for static dlopen (swbz#20802) +- Revert outdated translations +- vcs-to-changelog: Add quirk for __nonnull +- elf: Add elf/tst-dlopenfail-2 (swbz#25396, #1395758) +- Clear GL(dl_initfirst) when freeing its link_map (swbz#25396, #1395758) +- Update Translations +- Fix "elf: Add tst-ldconfig-ld_so_conf-update test" on 32bit. +- elf: Add tst-ldconfig-ld_so_conf-update test +- sl_SI locale: Use "." as the thousands separator (swbz#25233) + +* Mon Jan 06 2020 Arjun Shankar - 2.30.9000-29 +- Auto-sync with upstream branch master, + commit cbce69e70dc4b04fefcc7257e593733b8b03856c: +- Multiple locales: Add date_fmt (bug 24054) +- Update libc.pot for 2.31 release +- Add libm_alias_finite for _finite symbols +- Linux: Fix clock_nanosleep time64 check +- linux: Fix vDSO macros build with time64 interfaces +- x86: Make x32 use x86 time implementation +- Remove vDSO support from make-syscall.sh +- linux: Update x86 vDSO symbols +- linux: Update mips vDSO symbols +- linux: Consolidate Linux gettimeofday +- linux: Consolidate time implementation +- elf: Enable relro for static build +- elf: Move vDSO setup to rtld (BZ#24967) +- linux: Add support for clock_gettime64 vDSO +- linux: Optimize fallback 32-bit clock_gettime +- linux: Enable vDSO clock_gettime64 for i386 +- linux: Enable vDSO clock_gettime64 for arm +- linux: Enable vDSO clock_gettime64 for mips +- linux: Add support for clock_getres64 vDSO +- linux: Optimize fallback 32-bit clock_getres +- htl: Use dso_handle.h +- htl: Drop common tcbhead_t definition +- htl: Move pthread_atfork to libc_nonshared.a +- htl: Add __errno_location and __h_errno_location +- hurd: Fix message reception for timer_thread + +* Thu Jan 02 2020 Florian Weimer - 2.30.9000-28 +- Auto-sync with upstream branch master, + commit cc47d5c5f53f6d845ac54698ae8929af15662c44: +- Linux: Use built-in system call tables +- lv_LV locale: Correct the time part of d_t_fmt (swbz#25324) +- km_KH locale: Use "%M" instead of "m" in d_t_fmt (swbz#25323) +- ldbl-128ibm-compat: Do not mix -mabi=*longdouble and -mlong-double-128 +- ldbl-128ibm-compat: Compiler flags for stdio functions +- Do not redirect calls to __GI_* symbols, when redirecting to *ieee128 +- aarch64: add default memcpy version for kunpeng920 +- aarch64: ifunc rename for kunpeng +- aarch64: Modify error-shown comments for strcpy +- linux: Consolidate sigprocmask +- Fix return code for __libc_signal_* functions +- nptl: Remove duplicate internal __SIZEOF_PTHREAD_MUTEX_T (swbz#25241) + +* Thu Dec 26 2019 Carlos O'Donell - 2.30.9000-27 +- Auto-sync with upstream branch master, + commit b8c210bcc74840d24c61d39bde15bea9daf3e271. +- mnw_MM, my_MM, and shn_MM locales: Do not use %Op +- Avoid compat symbols for totalorder in powerpc64le IEEE long double +- ldbl-128ibm-compat: Add *cvt functions +- Refactor *cvt functions implementation (2/2) +- Refactor *cvt functions implementation (1/2) +- Add exception-based flags for wait4 +- aarch64: Optimized memset for Kunpeng processor. +- aarch64: Optimized strlen for strlen_asimd +- aarch64: Add Huawei Kunpeng to tunable cpu list +- aarch64: Optimized implementation of memrchr +- aarch64: Optimized implementation of strnlen +- aarch64: Optimized implementation of strcpy +- aarch64: Optimized implementation of memcmp +- Consolidate wait3 implementations +- Implement waitpid in terms of wait4 +- linux: Use waitid on wait4 if __NR_wait4 is not defined +- Implement wait in terms of waitpid +- nptl: Move waitpid implementation to libc +- nptl: Move wait implementation to libc +- Remove __waitpid_nocancel +- Fix test isolation for elf/tst-ifunc-fault-lazy, elf/tst-ifunc-fault-bindnow +- Fix __libc_signal_block_all on sparc64 +- powerpc: Do not run IFUNC resolvers for LD_DEBUG=unused [BZ #24214] + +* Thu Dec 19 2019 Patsy Franklin - 2.30.9000-26 +- Auto-sync with upstream branch master, + commit 3dcad8158f43d71d5b8f6f317f82952ddf3468f3. +- hurd: Do not make sigprocmask available in ld.so +- build-many-glibcs.py: Do not build C++ PCHs by default +- hurd: Make getrandom honour GRND_NONBLOCK +- tunables: report sbrk() failure +- build-many-glibcs.py: Add mipsisa64r6el-linux-gnu target +- mips: Do not include hi and lo in __SYSCALL_CLOBBERS for R6 +- ldbl-128ibm-compat: Add ISO C99 versions of scanf functions +- ldbl-128ibm-compat: Fix selection of GNU and ISO C99 scanf +- hurd: Fix local PLT +- dlopen: Do not block signals +- dlopen: Rework handling of pending NODELETE status +- dlopen: Fix issues related to NODELETE handling and relocations +- hurd: Fix __close_nocancel_nostatus availability +- hurd: add getrandom and getentropy implementations +- hurd: Implement __close_nocancel_nostatus +- manual: clarify fopen with the x flag +- S390: Use sysdeps/ieee754/dbl-64/wordsize-64 on s390x. +- S390: Implement roundtoint and converttoint and define TOINT_INTRINSICS. +- S390: Implement math-barriers math_opt_barrier and math_force_eval. +- S390: Use libc_fe* macros in fe* functions. +- S390: Implement libc_fe* macros. +- S390: Use convert-to-fixed instruction for llround functions. +- S390: Use convert-to-fixed instruction for lround functions. +- S390: Use convert-to-fixed instruction for llrint functions. +- S390: Use convert-to-fixed instruction for lrint functions. +- S390: Use load-fp-integer instruction for roundeven functions. +- Adjust s_copysignl.c regarding code style. +- Adjust s_ceilf.c and s_ceill.c regarding code style. +- Adjust s_floorf.c and s_floorl.c regarding code style. +- Adjust s_rintf.c and s_rintl.c regarding code style. +- Adjust s_nearbyintf.c and s_nearbyintl.c regarding code style. +- Use GCC builtins for copysign functions if desired. +- Use GCC builtins for round functions if desired. +- Use GCC builtins for trunc functions if desired. +- Use GCC builtins for ceil functions if desired. +- Use GCC builtins for floor functions if desired. +- Use GCC builtins for rint functions if desired. +- Use GCC builtins for nearbyint functions if desired. +- Always use wordsize-64 version of s_round.c. +- Always use wordsize-64 version of s_trunc.c. +- Always use wordsize-64 version of s_ceil.c. +- Always use wordsize-64 version of s_floor.c. +- Always use wordsize-64 version of s_rint.c. +- Always use wordsize-64 version of s_nearbyint.c. +- ldconfig: Do not print a warning for a missing ld.so.conf file +- hurd: Fix using altstack while in an RPC call to be aborted +- Fix failure when CFLAGS contains -DNDEBUG (Bug 25251) + +* Mon Dec 09 2019 DJ Delorie - 2.30.9000-25 +- Auto-sync with upstream branch master, + commit 0487ebed2278b20971af4cabf186fd3681adccf0. +- nptl: Add more missing placeholder abi symbol from nanosleep move +- sysdeps/riscv/start.S: rename .Lload_gp to load_gp (bug 24376) +- y2038: linux: Provide __timer_settime64 implementation +- y2038: linux: Provide __timer_gettime64 implementation +- timer: Decouple x86_64 specific timer_settime from generic Linux implementation +- timer: Decouple x86_64 specific timer_gettime from generic Linux implementation +- time: Introduce glibc's internal struct __itimerspec64 +- Correct range checking in mallopt/mxfast/tcache [BZ #25194] +- misc/test-errno-linux: Handle EINVAL from quotactl +- : Define __CORRECT_ISO_CPP_STRING_H_PROTO for Clang [BZ #25232] +- build-many-glibcs.py: Move sparcv8 to extra_glibcs + +* Thu Dec 5 2019 Florian Weimer - 2.30.9000-24 +- Upstream patches for fallout from dlopen NODELETE changes (#1778344, #1778366) + +* Wed Dec 04 2019 Patsy Franklin - 2.30.9000-23 +- Auto-sync with upstream branch master, + commit ec138c67cbda8b5826a0a2a7ba456408117996dc. +- sysdeps: Add clock_gettime64 vDSO +- Do not use ld.so to open statically linked programs in debugglibc.sh +- Attach to test in container from debugglibc.sh +- Expand $(as-needed) and $(no-as-needed) throughout the build system +- x86: Assume --enable-cet if GCC defaults to CET [BZ #25225] +- ldbl-128ibm-compat: Add tests for strfroml, strtold, and wcstold +- ldbl-128ibm-compat: Add tests for strfmon and strfmon_l +- ldbl-128ibm-compat: Add strfmon_l with IEEE long double format +- ldbl-128ibm-compat: Replace http with https in new files +- elf: Do not run IFUNC resolvers for LD_DEBUG=unused [BZ #24214] +- elf/tst-dlopenfail: Disable --no-as-needed for tst-dlopenfailmod1.so +- hurd: Fix ld.so __access override from libc +- hurd: Fix ld.so __getcwd override from libc +- hurd: Make __sigprocmask GLIBC_PRIVATE +- hurd: Fix renameat2 error +- hurd: make strerror(0) coherent with other ports +- hurd: Fix ld.so link +- Update kernel version to 5.4 in tst-mman-consts.py. +- Update SOMAXCONN value from Linux 5.4. +- Update syscall-names.list for Linux 5.4. +- Fix syntax error in build-many-glibcs.py. +- Define MADV_COLD and MADV_PAGEOUT from Linux 5.4. + +* Mon Dec 2 2019 Florian Weimer - 2.30.9000-22 +- dlopen: Remove incorrect assert in activate_nodelete (#1778344) + +* Thu Nov 28 2019 Florian Weimer - 2.30.9000-21 +- Auto-sync with upstream branch master, + commit e37c2cf299b61ce18f62852f6c5624c27829b610: +- Move _dl_open_check to its original place in dl_open_worker +- Block signals during the initial part of dlopen +- Remove all loaded objects if dlopen fails, ignoring NODELETE (#1395758) +- Avoid late dlopen failure due to scope, TLS slotinfo updates (swbz#25112) +- Avoid late failure in dlopen in global scope update (swbz#25112) +- Lazy binding failures during dlopen/dlclose must be fatal (swbz#24304) +- resolv: Implement trust-ad option for /etc/resolv.conf (#1164339) +- dlsym: Do not determine caller link map if not needed +- libio: Disable vtable validation for pre-2.1 interposed handles (swbz#25203) +- ldbl-128ibm-compat: Add syslog functions +- ldbl-128ibm-compat: Add obstack printing functions +- ldbl-128ibm-compat: Reuse tests for err.h and error.h functions +- ldbl-128ibm-compat: Add error.h functions +- ldbl-128ibm-compat: Add err.h functions +- ldbl-128ibm-compat: Add argp_error and argp_failure +- sparc: Use atomic compiler builtins on sparc +- Remove 32 bit sparc v7 support + +* Wed Nov 27 2019 Arjun Shankar - 2.30.9000-20 +- Auto-sync with upstream branch master, + commit bfdb731438206b0f70fe7afa890681155c30b419: +- rtld: Check __libc_enable_secure for LD_PREFER_MAP_32BIT_EXEC (CVE-2019-19126) +- Introduce DL_LOOKUP_FOR_RELOCATE flag for _dl_lookup_symbol_x +- Enable inlining issignalingf within glibc +- Don't use a custom wrapper macro around __has_include (bug 25189). +- Remove duplicate inline implementation of issignalingf +- misc: Set generic pselect as ENOSYS +- Use DEPRECATED_SCANF macro for remaining C99-compliant scanf functions +- ldbl-128ibm-compat: Add regular/wide character printing printing functions +- ldbl-128ibm-compat: Test double values and positional arguments +- ldbl-128ibm-compat: Add regular/wide character scanning functions +- arm: Fix armv7 selection after 'Split BE/LE abilist' +- Use Linux 5.4 in build-many-glibcs.py. +- sysdeps/posix: Simplify if expression in getaddrinfo +- sysdeps/posix/getaddrinfo: Return early on invalid address family +- ru_UA locale: use copy "ru_RU" in LC_TIME (bug 25044) +- locale: Greek -> ASCII transliteration table [BZ #12031] +- nptl: Cleanup mutex internal offset tests +- nptl: Add tests for internal pthread_rwlock_t offsets +- nptl: Remove rwlock elision definitions +- nptl: Add struct_mutex.h and struct_rwlock.h +- nptl: Add default pthreadtypes-arch.h and pthread-offsets.h +- Compile elf/rtld.c with -fno-tree-loop-distribute-patterns. +- nptl: Fix __PTHREAD_MUTEX_INITIALIZER for !__PTHREAD_MUTEX_HAVE_PREV +- S390: Fix handling of needles crossing a page in strstr z15 ifunc [BZ #25226] + +* Mon Nov 18 2019 Patsy Griffin - 2.30.9000-19 +- Auto-sync with upstream branch master, + commit 2a764c6ee848dfe92cb2921ed3b14085f15d9e79. +- Enhance _dl_catch_exception to allow disabling exception handling +- hurd: Suppress GCC 10 -Warray-bounds warning in init-first.c [BZ #25097] +- linux: Add comment on affinity set sizes to tst-skeleton-affinity.c +- Avoid zero-length array at the end of struct link_map [BZ #25097] +- Introduce link_map_audit_state accessor function +- Properly initialize audit cookie for the dynamic loader [BZ #25157] +- nios2: Work around backend bug triggered by csu/libc-tls.c (GCC PR 92499) +- Redefine _IO_iconv_t to store a single gconv step pointer [BZ #25097] +- Add new script for plotting string benchmark JSON output +- support: Fix support_set_small_thread_stack_size to build on Hurd +- login: Use pread64 in utmp implementation +- Clarify purpose of assert in _dl_lookup_symbol_x +- aarch64: Increase small and medium cases for __memcpy_generic +- login: Introduce matches_last_entry to utmp processing + +* Tue Nov 12 2019 Arjun Shankar - 2.30.9000-18 +- Auto-sync with upstream branch master, + commit cba932a5a9e91cffd7f4172d7e91f9b2efb1f84b: +- nptl: Move nanosleep implementation to libc +- Refactor nanosleep in terms of clock_nanosleep +- nptl: Refactor thrd_sleep in terms of clock_nanosleep +- math: enhance the endloop condition of function handle_input_flag +- hurd: Remove lingering references to the time function +- hurd: Use __clock_gettime in _hurd_select +- login: Remove double-assignment of fl.l_whence in try_file_lock +- nptl: Add missing placeholder abi symbol from nanosleep move +- login: Acquire write lock early in pututline [BZ #24882] +- Remove hppa pthreadP.h +- sysdeps/clock_nanosleep: Use clock_nanosleep_time64 if avaliable +- Fix array bounds violation in regex matcher (bug 25149) +- support: Add support_set_small_thread_stack_size +- linux: Reduce stack size for nptl/tst-thread-affinity-pthread +- y2038: linux: Provide __ppoll64 implementation +- Declare asctime_r, ctime_r, gmtime_r, localtime_r for C2X. +- support: Add xsetlocale function +- libio/tst-fopenloc: Use xsetlocale, xfopen, and xfclose +- Fix clock_nanosleep when interrupted by a signal +- slotinfo in struct dtv_slotinfo_list should be flexible array [BZ #25097] + +* Wed Nov 06 2019 Patsy Franklin - 2.30.9000-17 +- Auto-sync with upstream branch master, + commit 2a0356e1191804d57005e1cfe2a72f019b7a8cce. +- posix: Sync regex with gnulib +- Add mnw language code [BZ #25139] +- Add new locale: mnw_MM (Mon language spoken in Myanmar) [BZ #25139] +- S390: Fp comparison are now raising FE_INVALID with gcc 10. +- linux: pselect: Remove CALL_PSELECT6 macro +- Fix run-one-test so that it runs elf tests +- nptl: Fix niggles with pthread_clockjoin_np +- hppa: Align __clone stack argument to 8 bytes (Bug 25066) +- y2038: linux: Provide __futimens64 implementation +- y2038: linux: Provide __utimensat64 implementation +- nptl: Add pthread_timedjoin_np, pthread_clockjoin_np NULL timeout test +- nptl: Add pthread_clockjoin_np +- manual: Add documentation for pthread_tryjoin_np and pthread_timedjoin_np +- nptl: Convert tst-join3 to use libsupport +- Sync time/mktime.c with gnulib +- Sync timespec-{add,sub} with gnulib +- Sync intprops.h with gnulib +- Refactor adjtimex based on clock_adjtime +- Refactor PI mutexes internal definitions +- Remove pause and nanosleep not cancel wrappers +- nptl: Replace non cancellable pause/nanosleep with futex +- Consolidate lowlevellock-futex.h +- Consolidate futex-internal.h +- Base max_fast on alignment, not width, of bins (Bug 24903) +- Revise the documentation of simple calendar time. +- Make second argument of gettimeofday as 'void *' +- Use clock_gettime to implement gettimeofday. +- Use clock_gettime to implement timespec_get. +- Consolidate and deprecate ftime +- Change most internal uses of time to __clock_gettime. +- Use clock_gettime to implement time. +- Use clock_settime to implement settimeofday. +- Use clock_settime to implement stime; withdraw stime. +- Change most internal uses of __gettimeofday to __clock_gettime. +- Linux/Alpha: don't use timeval32 system calls. +- resolv/tst-idna_name_classify: Isolate from system libraries +- hurd: Support for file record locking +- Comment out initgroups from example nsswitch.conf (Bug 25146) + +* Mon Oct 28 2019 DJ Delorie - 2.30.9000-16 +- Auto-sync with upstream branch master, + commit 177a3d48a1c74d7b2cd6bfd48901519d25a5ecad. +- y2038: linux: Provide __clock_getres64 implementation +- time: Introduce function to check correctness of nanoseconds value +- Add Transliterations for Unicode Misc. Mathematical Symbols-A/B [BZ #23132] +- Install charmaps uncompressed in testroot +- Add wait-for-debugger test harness hooks +- Define __STATFS_MATCHES_STATFS64 +- hurd: Fix build after __pread64 usage in the dynamic loader +- sysdeps/stat: Handle 64-bit ino_t types on 32-bit hosts +- S390: Remove not needed stack frame in syscall function. + +* Fri Oct 25 2019 DJ Delorie - 2.30.9000-15 +- Add *.mo files to all-langpacks (#1624528) + +* Thu Oct 24 2019 DJ Delorie - 2.30.9000-14 +- Add Requires on basesystem for main package (#1757267) +- Add Requires on coreutils for glibc-headers (uses rm) + +* Wed Oct 23 2019 Arjun Shankar - 2.30.9000-13 +- Auto-sync with upstream branch master, + commit 7db1fe38de21831d53ceab9ae83493d8d1aec601: +- Include explicitly in Linux clock_settime.c +- Remove math-finite.h +- Remove finite-math tests +- Remove x64 _finite tests and references +- Fix testroot.pristine creation copying dynamic linker + +* Fri Oct 18 2019 Patsy Franklin - 2.30.9000-12 +- Auto-sync with upstream branch master, + commit ef21bd2d8c6805c0c186a01f7c5039189f51b8c4. +- loadarchive: guard against locale-archive corruption (Bug #25115) +- Undo accidental commit to ChangeLog.19. +- nptl: Document AS-safe functions in cancellation.c. +- elf: Use nocancel pread64() instead of lseek()+read() +- Add nocancel version of pread64() +- Add run-one-test convenience target and makefile help text +- Update sysvipc kernel-features.h files for Linux 5.1 +- S390: Add new s390 platform z15. +- nptl: SIGCANCEL, SIGTIMER, SIGSETXID are always defined +- nptl/tst-cancel25 needs to be an internal test +- Remove libc_hidden_def from __semtimedop stub +- sysvipc: Implement semop based on semtimedop +- ipc: Refactor sysvipc internal definitions +- Rename and split elf/tst-dlopen-aout collection of tests +- dlfcn: Remove remnants of caller sensitivity from dlinfo +- ldconfig: handle .dynstr located in separate segment (bug 25087) +- ldd: Print "not a dynamic executable" on standard error [BZ #24150] +- Add PTRACE_GET_SYSCALL_INFO from Linux 5.3 to sys/ptrace.h. +- Move ChangeLog to ChangeLog.old/ChangeLog.19 +- manual: Remove warning in the documentation of the abort function +- sysvipc: Set ipc_perm mode as mode_t (BZ#18231) +- Simplify note processing +- syscall-names.list: fix typos in comment +- y2038: linux: Provide __clock_settime64 implementation +- posix: Use posix_spawn for wordexp +- mips: Do not malloc on getdents64 fallback +- sparc: Assume GOTDATA support in the toolchain +- : Remove wrong comment about getdents64 declaration +- ChangeLog: Remove leading spaces before tabs and trailing whitespace +- Make tst-strftime2 and tst-strftime3 depend on locale generation +- posix/tst-wordexp-nocmd: Fix diagnostics output in test +- wordexp: Split out command execution tests from posix/wordexp-test + +* Tue Oct 08 2019 Arjun Shankar - 2.30.9000-11 +- Adjust glibc-rh741105.patch. +- Auto-sync with upstream branch master, + commit ca602c1536ce2777f95c07525f3c42d78812e665: +- Add TCP_TX_DELAY from Linux 5.3 to netinet/tcp.h +- [powerpc] fenv_private.h clean up +- [powerpc] libc_feupdateenv_test: optimize FPSCR access +- [powerpc] __fesetround_inline optimizations +- [powerpc] Rename fegetenv_status to fegetenv_control +- [powerpc] libc_feholdsetround_noex_ppc_ctx: optimize FPSCR write +- [powerpc] Rename fesetenv_mode to fesetenv_control +- Add helper script for glibc debugging +- Update bits/mman.h constants and tst-mman-consts.py for Linux 5.3. +- y2038: Provide conversion helpers for struct __timespec64 +- Use binutils 2.33 branch in build-many-glibcs.py. +- Sync "language", "lang_name", "territory", "country_name" with CLDR/langtable +- Split up endian.h to minimize exposure of BYTE_ORDER. +- time: Add padding for the timespec if required +- Enable passing arguments to the inferior in debugglibc.sh +- [powerpc] No need to enter "Ignore Exceptions Mode" +- Y2038: Include proper header to provide support for struct timeval on HURD +- Disable warnings in string/tester.c at top level. +- string/endian.h: Restore the __USE_MISC conditionals +- Disable -Wmaybe-uninitialized for total_deadline in sunrpc/clnt_udp.c. +- ChangeLog update from my last commit +- nptl: Move pthread_attr_setinheritsched implementation into libc. +- elf: Never use the file ID of the main executable [BZ #24900] +- elf: Assign TLS modid later during dlopen [BZ #24930] +- nptl: Move pthread_attr_getschedparam implementation into libc +- riscv: Remove support for variable page sizes +- nptl: Move pthread_attr_setschedparam implementation into libc + +* Fri Sep 27 2019 Zbigniew Jędrzejewski-Szmek - 2.30.9000-10 +- Use full locale names in langpack descriptions (#1651375) + +* Thu Sep 26 2019 Patsy Franklin - 2.30.9000-9 +- Auto-sync with upstream branch master, + commit 464cd3a9d5f505d92bae9a941bb75b0d91ac14ee. +- y2038: Introduce struct __timespec64 - new internal glibc type +- auto-changelog: Remove latin1 from codecs +- Set the expects flags to clock_nanosleep +- Fix tst-sigcontext-get_pc rule name from a43565ac447b1 +- inet/net-internal.h: Fix uninitalised clntudp_call() variable +- Fix vDSO initialization on arm and mips +- Script to generate ChangeLog-like output from git log +- [powerpc] SET_RESTORE_ROUND optimizations and bug fix +- Fix building support_ptrace.c on i686-gnu. +- S390: Use _HP_TIMING_S390_H instead of _HP_TIMING_H. +- Update syscall-names.list for Linux 5.3. +- Use Linux 5.3 in build-many-glibcs.py. +- S390: Add support for HP_TIMING_NOW. +- Fix RISC-V vfork build with Linux 5.3 kernel headers. +- Add UNSUPPORTED check in elf/tst-pldd. +- sparc64: Use linux generic time implementation +- mips: Consolidate INTERNAL_VSYSCALL_CALL +- powerpc: Simplify vsyscall internal macros +- Refactor vDSO initialization code +- Remove PREPARE_VERSION and PREPARE_VERSION_KNOW +- Fix small error in HP_TIMING_PRINT trailing null char setting + +* Mon Sep 16 2019 Parag Nemade - 2.30.9000-8 +- Change Supplements "langpacks-" to "langpacks-core-" (#1729992) + +* Mon Sep 16 2019 DJ Delorie - 2.30.9000-7 +- Auto-sync with upstream branch master, + commit 1a6566094d3097f4a3037ab5555cddc6cb11c3a3. +- alpha: force old OSF1 syscalls for getegid, geteuid and getppid [BZ #24986] +- Fix http: URL in 'configure' +- Regenerate charmap-kw.h, locfile-kw.h +- Fix three GNU license URLs, along with trailing-newline issues. +- Prefer https to http for gnu.org and fsf.org URLs + +* Fri Sep 06 2019 Patsy Franklin - 2.30.9000-6 +- Auto-sync with upstream branch master, + commit 1b7f04070bd94f259e2ed24d6fb76309d64fb164. +- locale: Avoid zero-length array in _nl_category_names [BZ #24962] +- math: Replace const attribute with pure in totalorder* functions +- y2038: Introduce the __ASSUME_TIME64_SYSCALLS define +- Finish move of clock_* functions to libc. [BZ #24959] +- Update Alpha libm-test-ulps +- localedef: Use initializer for flexible array member [BZ #24950] +- Add misc/tst-mntent-autofs, testing autofs "ignore" filtering +- Use autofs "ignore" mount hint in getmntent_r/getmntent +- hurd: Fix build +- Use generic memset/memcpy/memmove in benchtests +- nptl: Move pthread_attr_getinheritsched implementation into libc +- hurd: Fix SS_ONSTACK support +- hurd: Remove optimizing anonymous maps as __vm_allocate. +- hurd: Fix poll and select POSIX compliancy details about errors +- hurd: Fix timeout handling in _hurd_select +- hurd getcwd: Allow unknown root directory +- hurd: Fix implementation of setitimer. +- hurd: Fix _hurd_select for single fd sets +- MIPS support for GNU hash +- sh: Split BE/LE abilist +- microblaze: Split BE/LE abilist +- arm: Split BE/LE abilist +- Correct the spelling of more contributors +- Fix posix/tst-regex by using UTF-8 and own test input +- [powerpc] fegetenv_status: simplify instruction generation +- [powerpc] fesetenv: optimize FPSCR access +- [powerpc] SET_RESTORE_ROUND improvements +- [powerpc] fe{en,dis}ableexcept, fesetmode: optimize FPSCR accesses +- [powerpc] fe{en,dis}ableexcept optimize bit translations +- misc: Use allocate_once in getmntent +- nptl: Move pthread_attr_setdetachstate implementation into libc +- login: pututxline could fail to overwrite existing entries [BZ #24902] +- Fix posix/tst-regex by using a dedicated input-file. + +* Tue Aug 27 2019 DJ Delorie - 2.30.9000-5 +- Move makedb from glibc-common to nss_db (#1704334) + +* Mon Aug 26 2019 DJ Delorie - 2.30.9000-4 +- Auto-sync with upstream branch master, + commit 1bced8cadc82077f0201801239e89eb24b68e9aa. +- Don't put non-ASCII into installed headers +- Fix spellings of contributor names in comments and doc +- [MIPS] Raise highest supported EI_ABIVERSION value [SWBZ #24916] +- mips: Force RWX stack for hard-float builds that can run on pre-4.8 kernels +- linux: Make profil_counter a compat_symbol (SWBZ#17726) +- Refactor sigcontextinfo.h +- Add RTLD_SINGLE_THREAD_P on generic single-thread.h +- Chinese locales: Set first_weekday to 2 (swbug 24682). +- powerpc: Fix typos and field name in comments +- Mark IDN tests unsupported with libidn2 before 2.0.5. +- Document strftime %Ob and %OB as C2X features. +- Remove dead regex code +- Fix bad pointer / leak in regex code +- Don't use the argument to time. +- Add tgmath.h macros for narrowing functions. +- Update i386 libm-test-ulps + +* Mon Aug 19 2019 Carlos O'Donell - 2.30.9000-3 +- Drop glibc-fedora-nscd-warnings.patch; applied upstream. +- Drop Source7: nsswitch.conf; applying patch to upstream. +- Add glibc-fedora-nsswitch.patch for Fedora customizations. +- Auto-sync with upstream branch master, + commit d34d4c80226b3f5a1b51a8e5b005a52fba07d7ba: +- Do not print backtraces on fatal glibc errors. +- elf: Self-dlopen failure with explict loader invocation (swbz#24900) +- login: Add nonstring attributes to struct utmp, struct utmpx (swbz#24899) +- login: Use struct flock64 in utmp (swbz#24880) +- login: Disarm timer after utmp lock acquisition (swbz#24879) + +* Fri Aug 16 2019 Carlos O'Donell - 2.30.9000-2 +- Fix C.UTF-8 to use full code ranges. + +* Thu Aug 15 2019 Florian Weimer - 2.30.9000-1 +- Auto-sync with upstream branch master, + commit 341da5b4b6253de9a7581a066f33f89cacb44dec. + +* Fri Aug 02 2019 Florian Weimer - 2.30-1 +- Drop glibc-rh1734680.patch, applied upstream. +- Auto-sync with upstream branch release/2.30/master, + commit be9a328c93834648e0bec106a1f86357d1a8c7e1: +- malloc: Remove unwanted leading whitespace in malloc_info (swbz#24867) +- glibc 2.30 release +- iconv: Revert steps array reference counting changes (#1734680) +- Restore r31 setting in powerpc32 swapcontext + +* Wed Jul 31 2019 Florian Weimer - 2.29.9000-37 +- Fix memory leak in iconv_open (#1734680) + +* Tue Jul 30 2019 Florian Weimer - 2.29.9000-36 +- Drop glibc-rh1732406.patch, fix for the regression applied upstream. +- Auto-sync with upstream branch master, + commit 8a814e20d443adc460a1030fa1a66aa9ae817483: +- nptl: Use uintptr_t for address diagnostic in nptl/tst-pthread-getattr +- Linux: Move getdents64 to +- test-container: Install with $(sorted-subdirs) (swbz#24794) +- gconv: Check reference count in __gconv_release_cache (#1732406) +- x86-64: Compile branred.c with -mprefer-vector-width=128 (swbz#24603) +- build-many-glibcs.py: Use Linux 5.2 by default +- Linux: Use in-tree copy of SO_ constants for !__USE_MISC (swbz#24532) +- test-container: Avoid copying unintended system libraries + +* Thu Jul 25 2019 Fedora Release Engineering +- Rebuilt for https://fedoraproject.org/wiki/Fedora_31_Mass_Rebuild + +* Tue Jul 23 2019 Florian Weimer - 2.29.9000-34 +- Revert libio change that causes crashes (#1732406) + +* Mon Jul 22 2019 DJ Delorie - 2.29.9000-33 +- Auto-sync with upstream branch master, + commit dcf36bcad3f283f77893d3b157ef7bb2c99419f2. +- Add NEWS entry about the new AArch64 IFUNC resolver call ABI +- locale/C-translit.h.in: Cyrillic -> ASCII transliteration [BZ #2872] +- Linux: Update syscall-names.list to Linux 5.2 + +* Thu Jul 18 2019 DJ Delorie - 2.29.9000-32 +- Auto-sync with upstream branch master, + commit 3556658c5b8765480711b265abc901c67d5fc060. +- Regenerate po/libc.pot for 2.30 release. +- nptl: Add POSIX-proposed _clock functions to hppa pthread.h +- nptl: Remove unnecessary forwarding of pthread_cond_clockwait from libc +- Afar locales: Months and days updated from CLDR (bug 21897). +- nl_BE locale: Use "copy "nl_NL"" in LC_NAME (bug 23996). +- nl_BE and nl_NL locales: Dutch salutations (bug 23996). +- ga_IE and en_IE locales: Revert first_weekday removal (bug 24200). +- nptl: Remove futex_supports_exact_relative_timeouts +- Update NEWS for new _clockwait and _clocklock functions +- nptl: Add POSIX-proposed pthread_mutex_clocklock +- nptl: Rename lll_timedlock to lll_clocklock and add clockid parameter +- nptl: Add POSIX-proposed pthread_rwlock_clockrdlock & pthread_rwlock_clockwrlock +- nptl: pthread_rwlock: Move timeout validation into _full functions +- nptl: Add POSIX-proposed pthread_cond_clockwait +- nptl: Add POSIX-proposed sem_clockwait +- nptl: Add clockid parameter to futex timed wait calls +- posix: Fix large mmap64 offset for mips64n32 (BZ#24699) +- nss_db: fix endent wrt NULL mappings [BZ #24695] [BZ #24696] + +* Wed Jul 10 2019 Carlos O'Donell - 2.29.9000-31 +- Auto-sync with upstream branch master, + commit 30ba0375464f34e4bf8129f3d3dc14d0c09add17. +- Don't declare __malloc_check_init in (bug 23352) +- nftw: fill in stat buf for dangling links [BZ #23501] +- dl-vdso: Add LINUX_4 HASH CODE to support nds32 vdso mechanism +- riscv: restore ABI compatibility (bug 24484) +- aarch64: new ifunc resolver ABI +- nptl: Remove vfork IFUNC-based forwarder from libpthread [BZ #20188] +- malloc: Add nptl, htl dependency for the subdirectory [BZ #24757] +- Call _dl_open_check after relocation [BZ #24259] +- Linux: Use mmap instead of malloc in dirent/tst-getdents64 +- ld.so: Support moving versioned symbols between sonames [BZ #24741] +- io: Remove copy_file_range emulation [BZ #24744] +- Linux: Adjust gedents64 buffer size to int range [BZ #24740] +- powerpc: Use generic e_expf +- Linux: Add nds32 specific syscalls to syscall-names.list +- szl_PL locale: Fix a typo in the previous commit (bug 24652). + +* Mon Jun 24 2019 DJ Delorie - 2.29.9000-30 +- Auto-sync with upstream branch master, + commit 2bd81b60d6ffdf7e0d22006d69f4b812b1c80513. +- szl_PL locale: Spelling corrections (swbz 24652). +- nl_{AW,NL}: Correct the thousands separator and grouping (swbz 23831). +- Add missing VDSO_{NAME,HASH}_* macros and use them for PREPARE_VERSION_KNOWN +- nptl: Convert various tests to use libsupport +- support: Invent verbose_printf macro +- support: Add xclock_now helper function. + +* Fri Jun 21 2019 Florian Weimer - 2.29.9000-29 +- Auto-sync with upstream branch master, + commit 21cc130b78a4db9113fb6695e2b951e697662440: +- During exit, skip wide buffer handling for legacy stdio handles (#1722216) +- powerpc: add 'volatile' to asm +- powerpc: Fix static-linked version of __ppc_get_timebase_freq (swbz#24640) +- nl_AW locale: Correct the negative monetary format (swb#z24614) +- Fix gcc 9 build errors for make xcheck. (swbz#24556) +- dlfcn: Avoid one-element flexible array in Dl_serinfo (swbz#24166) +- elf: Refuse to dlopen PIE objects (swbz#24323) +- nl_NL locale: Correct the negative monetary format (swbz#24614) +- powerpc: Refactor powerpc64 lround/lroundf/llround/llroundf +- powerpc: refactor powerpc64 lrint/lrintf/llrint/llrintf + +* Mon Jun 17 2019 Florian Weimer - 2.29.9000-28 +- Auto-sync with upstream branch master, + commit 48c3c1238925410b4e777dc94e2fde4cc9132d44. +- Linux: Fix __glibc_has_include use for and statx (#1721129) +- : Inhibit macro expansion for __glibc_has_include +- Add IPV6_ROUTER_ALERT_ISOLATE from Linux 5.1 to bits/in.h +- aarch64: handle STO_AARCH64_VARIANT_PCS +- aarch64: add STO_AARCH64_VARIANT_PCS and DT_AARCH64_VARIANT_PCS +- powerpc: Remove optimized finite +- math: Use wordsize-64 version for finite +- powerpc: Remove optimized isinf +- math: Use wordsize-64 version for isinf +- powerpc: Remove optimized isnan +- math: Use wordsize-64 version for isnan +- benchtests: Add isnan/isinf/isfinite benchmark +- powerpc: copysign cleanup +- powerpc: consolidate rint +- libio: freopen of default streams crashes in old programs (swbz#24632) +- Linux: Deprecate and sysctl +- : Use Linux UAPI header for statx if available and useful + (#1721129) +- : Add __glibc_has_include macro +- Improve performance of memmem +- Improve performance of strstr +- Benchmark strstr hard needles +- Fix malloc tests build with GCC 10 + +* Mon Jun 10 2019 Patsy Franklin - 2.29.9000-27 +- Auto-sync with upstream branch master, + commit 51ea67d54882318c4fa5394c386f4816ddc22408. +- powerpc: get_rounding_mode: utilize faster method to get rounding mode +- riscv: Do not use __has_include__ +- powerpc: fegetexcept: utilize function instead of duplicating code +- iconv: Use __twalk_r in __gconv_release_shlib +- Fix iconv buffer handling with IGNORE error handler (swbz#18830) + +* Wed Jun 5 2019 Florian Weimer - 2.29.9000-26 +- Restore /usr/lib/locale/locale-archive under its original name (#1716710) + +* Tue Jun 4 2019 Florian Weimer - 2.29.9000-25 +- Add glibc version to locale-archive name (#1716710) + +* Mon Jun 03 2019 Carlos O'Donell - 2.29.9000-24 +- Auto-sync with upstream branch master, + commit dc91a19e6f71e1523f4ac179191a29b2131d74bb: +- Linux: Add oddly-named arm syscalls to syscall-names.list. +- arm: Remove ioperm/iopl/inb/inw/inl/outb/outw/outl support. +- Add INADDR_ALLSNOOPERS_GROUP from Linux 5.1 to netinet/in.h. + +* Sat Jun 01 2019 Carlos O'Donell - 2.29.9000-23 +- Convert glibc_post_upgrade to lua. + +* Sat Jun 01 2019 Florian Weimer - 2.29.9000-22 +- Remove support for filtering glibc-all-langpacks (#1715891) +- Auto-sync with upstream branch master, + commit 9250e6610fdb0f3a6f238d2813e319a41fb7a810: +- powerpc: Fix build failures with current GCC +- Remove unused get_clockfreq files +- powerpc: generic nearbyint/nearbyintf +- tt_RU: Add lang_name (swbz#24370) +- tt_RU: Fix orthographic mistakes in mon and abmon sections (swbz#24369) +- Add IGMP_MRDISC_ADV from Linux 5.1 to netinet/igmp.h. + +* Mon May 27 2019 Arjun Shankar - 2.29.9000-21 +- Auto-sync with upstream branch master, + commit 85188d8211698d1a255f0aec6529546db5c56de3: +- Remove support for PowerPC SPE extension +- elf: Add tst-ldconfig-bad-aux-cache test +- Add F_SEAL_FUTURE_WRITE from Linux 5.1 to bits/fcntl-linux.h +- nss_dns: Check for proper A/AAAA address alignment + +* Tue May 21 2019 DJ Delorie - 2.29.9000-20 +- Auto-sync with upstream branch master, + commit 46ae07324b1cd50fbf8f37a076d6babcfca7c510. +- Improve string benchtest timing +- sysvipc: Add missing bit of semtimedop s390 consolidation +- wcsmbs: Fix data race in __wcsmbs_clone_conv [swbz #24584] +- libio: Fix gconv-related memory leak [swbz #24583] +- libio: Remove codecvt vtable [swbz #24588] +- support: Expose sbindir as support_sbindir_prefix +- support: Add missing EOL terminators on timespec +- support: Correct confusing comment +- sysvipc: Consolidate semtimedop s390 +- sysvipc: Fix compat msgctl (swbz#24570) +- Add NT_ARM_PACA_KEYS and NT_ARM_PACG_KEYS from Linux 5.1 to elf.h. +- Small tcache improvements +- manual: Document O_DIRECTORY +- Update kernel-features.h files for Linux 5.1. +- nss_nis, nss_nisplus: Remove RES_USE_INET6 handling +- nss_files: Remove RES_USE_INET6 from hosts processing +- support: Report NULL blobs explicitly in TEST_COMPARE +- dlfcn: Guard __dlerror_main_freeres with __libc_once_get (once) [swbz# 24476] +- Add missing Changelog entry + +* Wed May 15 2019 Florian Weimer - 2.29.9000-19 +- Auto-sync with upstream branch master, + commit 32ff397533715988c19cbf3675dcbd727ec13e18: +- Fix crash in _IO_wfile_sync (#1710460) +- nss: Turn __nss_database_lookup into a compatibility symbol +- support: Add support_install_rootsbindir +- iconv: Remove public declaration of __gconv_transliterate +- Linux: Add the tgkill function +- manual: Adjust twalk_r documentation. +- elf: Fix tst-pldd for non-default --prefix and/or --bindir (swbz#24544) +- support: Export bindir path on support_path +- configure: Make --bindir effective +- x86: Remove arch-specific low level lock implementation +- nptl: Assume LLL_LOCK_INITIALIZER is 0 +- nptl: Small optimization for lowlevellock +- Add single-thread.h header +- locale: Update to Unicode 12.1.0 (swbz#24535) +- malloc: Fix tcache count maximum (swbz#24531) +- sem_close: Use __twalk_r +- support: Fix timespec printf +- nptl/tst-abstime: Use libsupport +- nptl: Convert some rwlock tests to use libsupport +- nptl: Use recent additions to libsupport in tst-sem5 +- nptl: Convert tst-cond11.c to use libsupport +- support: Add timespec.h +- Move nptl/tst-eintr1 to xtests (swbz#24537) +- powerpc: trunc/truncf refactor +- powerpc: round/roundf refactor +- powerpc: floor/floorf refactor +- support: Add xclock_gettime +- malloc/tst-mallocfork2: Use process-shared barriers +- Update syscall-names.list for Linux 5.1 +- Use GCC 9 in build-many-glibcs.py +- aarch64: thunderx2 memmove performance improvements +- misc/tst-tsearch: Additional explicit error checking +- elf: Fix elf/tst-pldd with --enable-hardcoded-path-in-tests (swbz#24506) +- misc: Add twalk_r function + +* Thu May 02 2019 Arjun Shankar - 2.29.9000-18 +- Auto-sync with upstream branch master, + commit 20aa5819586ac7ad11f711bab64feda307965191: +- semaphore.h: Add nonnull attributes +- powerpc: Remove power4 mpa optimization +- powerpc: Refactor ceil/ceilf +- Fix -O1 compilation errors with `__ddivl' and `__fdivl' [BZ #19444] +- Make mktime etc. compatible with __time64_t + +* Fri Apr 26 2019 Florian Weimer - 2.29.9000-17 +- Auto-sync with upstream branch master, + commit c57afec0a9b318bb691e0f5fa4e9681cf30df7a4: +- Increase BIND_NOW coverage (#1702671) +- Fix pldd hang (#1361689) +- riscv: remove DL_RO_DYN_SECTION (swbz#24484) +- locale: Add LOCPATH diagnostics to the locale program +- Reduce benchtests time + +* Mon Apr 22 2019 DJ Delorie - 2.29.9000-16 +- Auto-sync with upstream branch master, + commit 25f7a3c96116a9102df8bf7b04ef160faa32416d. +- malloc: make malloc fail with requests larger than PTRDIFF_MAX (BZ#23741) +- powerpc: Fix format issue from 3a16dd780eeba602 +- powerpc: fma using builtins +- powerpc: Use generic fabs{f} implementations +- mips: Remove rt_sigreturn usage on context function +- powerpc: Remove rt_sigreturn usage on context function +- support: Add support_capture_subprogram +- stdlib/tst-secure-getenv: handle >64 groups + +* Mon Apr 15 2019 Florian Weimer - 2.29.9000-15 +- Auto-sync with upstream branch master, + commit e3f454bac0f968216699ca405c127c858f0657c7: +- nss_dns: Do not replace root domain with empty string +- alloc_buffer: Return unqualified pointer type in alloc_buffer_next +- malloc: Set and reset all hooks for tracing (swbz#16573) + +* Thu Apr 11 2019 Florian Weimer - 2.29.9000-14 +- Run valgrind smoke test against the install tree + +* Thu Apr 11 2019 Florian Weimer - 2.29.9000-13 +- Do not use --g-libs with find-debuginfo.sh; it breaks valgrind (#1698824) + +* Wed Apr 10 2019 Florian Weimer - 2.29.9000-12 +- Strip debugging information from installed programs again (#1661510) + +* Tue Apr 09 2019 Carlos O'Donell - 2.29.9000-11 +- Drop glibc-warning-fix.patch. Microbenchmark code fixed upstream. +- Auto-sync with upstream branch master, + commit 648279f4af423c4783ec1dfa63cb7b46a7640217: +- powerpc: Use generic wcscpy optimization +- powerpc: Use generic wcschr optimization +- powerpc: Use generic wcsrchr optimization +- aarch64: thunderx2 memcpy implementation cleanup and streamlining +- resolv: Remove support for RES_USE_INET6 and the inet6 option +- resolv: Remove RES_INSECURE1, RES_INSECURE2 + +* Thu Apr 04 2019 Arjun Shankar - 2.29.9000-10 +- Auto-sync with upstream branch master, + commit 8260f23616c1a2a4e609f989a195fba7690a42ca: +- Fix strptime era handling, add more strftime tests [BZ #24394] +- time/tst-strftime2.c: Make the file easier to maintain +- time: Add tests for Minguo calendar [BZ #24293] +- ja_JP locale: Add entry for the new Japanese era [BZ #22964] +- Add Reiwa era tests to time/tst-strftime3.c + +* Mon Apr 01 2019 Arjun Shankar - 2.29.9000-9 +- Auto-sync with upstream branch master, + commit 993e3107af67edefcfc79a62ae55f7b98aa5151e: +- Add AArch64 HWCAPs from Linux 5.0 +- tt_RU: Fix orthographic mistakes in day and abday sections [BZ #24296] +- iconv, localedef: avoid floating point rounding differences [BZ #24372] +- Fix parentheses error in iconvconfig.c and ld-collate.c [BZ #24372] +- S390: New configure check and hwcap values for new CPU architecture arch13 +- S390: Add memmove, strstr, and memmem ifunc variants for arch13 +- nptl: Remove pthread_clock_gettime pthread_clock_settime +- linux: Assume clock_getres CLOCK_{PROCESS,THREAD}_CPUTIME_ID +- Remove __get_clockfreq +- Do not use HP_TIMING_NOW for random bits +- hp-timing: Refactor rtld usage, add generic support +- Add NT_ARM_PAC_MASK and NT_MIPS_MSA from Linux 5.0 to elf.h +- Add UDP_GRO from Linux 5.0 to netinet/udp.h +- nptl: Convert tst-sem5 & tst-sem13 to use libsupport +- nptl/tst-rwlock14: Test pthread_rwlock_timedwrlock correctly +- nss/tst-nss-files-alias-leak: add missing opening quote in printf +- math: Enable some math builtins for clang +- powerpc: Use __builtin_{mffs,mtfsf} +- RISC-V: Fix `test' operand error with soft-float ABI being configured + +* Wed Mar 20 2019 Carlos O'Donell - 2.29.9000-8 +- Add warnings and notes to /etc/nsswitch.conf and /etc/nscd.conf. + +* Mon Mar 18 2019 DJ Delorie - 2.29.9000-7 +- Auto-sync with upstream branch master, + commit 78919d3886c9543279ec755a701e279c62b44164. + +* Thu Mar 14 2019 Florian Weimer - 2.29.9000-6 +- Drop glibc-fedora-streams-rh436349.patch. STREAMS was removed upstream. +- Auto-sync with upstream branch master, + commit a0a0dc83173ce11ff45105fd32e5d14356cdfb9c: +- Remove obsolete, never-implemented XSI STREAMS declarations +- nss: Fix tst-nss-files-alias-truncated for default --as-needed linking +- scripts/check-obsolete-constructs.py: Process all headers as UTF-8. +- Use Linux 5.0 in build-many-glibcs.py. +- hurd: Add no-op version of __res_enable_icmp [BZ #24047] +- Move inttypes.h and stdint.h to stdlib. +- Use a proper C tokenizer to implement the obsolete typedefs test. +- Fix output of LD_SHOW_AUXV=1. + +* Wed Mar 13 2019 Florian Weimer - 2.29.9000-5 +- Drop glibc-rh1670028.patch, applied upstream +- Auto-sync with upstream branch master, + commit 38b52865d4ccfee3647f27e969e539a4396a73b1: +- elf: Add DF_1_KMOD, DF_1_WEAKFILTER, DF_1_NOCOMMON to +- resolv: Enable full ICMP errors for UDP DNS sockets [BZ #24047] +- C-SKY: add elf header definition for elfutils +- C-SKY: mark lr as undefined to stop unwinding +- C-SKY: remove user_regs definition +- C-SKY: fix sigcontext miss match +- Bug 24307: Update to Unicode 12.0.0 +- Break lines before not after operators, batch 4. +- check-wrapper-headers test: Adjust Fortran include file directory +- Fix location where math-vector-fortran.h is installed. + +* Wed Mar 06 2019 DJ Delorie - 2.29.9000-4 +- Auto-sync with upstream branch master, + commit 0ddb7ea842abf63516b74d4b057c052afc6ba863. +- nptl: Assume __ASSUME_FUTEX_CLOCK_REALTIME support +- powerpc: Fix build of wcscpy with --disable-multi-arch +- elf: Remove remnants of MAP_ANON emulation +- S390: Increase function alignment to 16 bytes. +- ja_JP: Change the offset for Taisho gan-nen from 2 to 1 [BZ #24162] +- ldbl-opt: Reuse test cases from misc/ that check long double +- ldbl-opt: Add error and error_at_line (bug 23984) +- ldbl-opt: Add err, errx, verr, verrx, warn, warnx, vwarn, and vwarnx (bug 23984) +- ldbl-opt: Reuse argp tests that print long double +- ldbl-opt: Add argp_error and argp_failure (bug 23983) +- elf/tst-big-note: Improve accuracy of test [BZ #20419] +- S390: Fix introduction of __wcscpy and weak wcscpy symbols. +- __netlink_assert_response: Add more __libc_fatal newlines [BZ #20271] +- Add more spaces before '('. +- elf: Add tests with a local IFUNC resolver [BZ #23937] +- elf/Makefile: Run IFUNC tests if binutils supports IFUNC +- powerpc: Fix linknamespace introduced by 4d8015639a75 +- hurd: Add renameat2 support for RENAME_NOREPLACE +- Fix -Wempty-body warnings in Hurd-specific code. +- Add some spaces before '('. +- wcsmbs: optimize wcsnlen +- wcsmbs: optimize wcsncpy +- wcsmbs: optimize wcsncat +- wcsmbs: optimize wcscpy +- wcsmbs: optimize wcscat +- wcsmbs: optimize wcpncpy +- wcsmbs: optimize wcpcpy +- Break further lines before not after operators. +- Add and move fall-through comments in system-specific code. + +* Fri Mar 1 2019 DJ Delorie - 2.29.9000-3 +- Add .gdb_index to debug information (rhbz#1680765) + +* Wed Feb 27 2019 Carlos O'Donell - 2.29.9000-2 +- Fix build failure related to microbenchmarks. + +* Tue Feb 26 2019 Carlos O'Donell - 2.29.9000-1 +- Auto-sync with upstream branch master, + commit e0cb7b6131ee5f2dca2938069b8b9590304e6f6b: +- nss_files: Fix /etc/aliases null pointer dereference (swbz#24059) +- regex: fix read overrun (swbz#24114) +- libio: use stdout in puts and putchar, etc (swbz#24051) +- aarch64: Add AmpereComputing emag to tunable cpu list +- aarch64: Optimized memset specific to AmpereComputing emag +- aarch64: Optimized memchr specific to AmpereComputing emag +- Require GCC 6.2 or later to build glibc +- manual: Document lack of conformance of sched_* functions (swbz#14829) +- libio: Use stdin consistently for input functions (swbz#24153) +- x86-64 memcmp: Use unsigned Jcc instructions on size (swbz#24155) +- Fix handling of collating elements in fnmatch (swbz#17396,swbz#16976) +- arm: Use "nr" constraint for Systemtap probes (swbz#24164) +- Fix alignment of TLS variables for tls variant TLS_TCB_AT_TP (swbz#23403) +- Add compiler barriers for pthread_mutex_trylock (swbz#24180) +- rt: Turn forwards from librt to libc into compat symbols (swbz#24194) +- Linux: Add gettid system call wrapper (swbz#6399) +- nptl: Avoid fork handler lock for async-signal-safe fork (swbz#24161) +- elf: Ignore LD_AUDIT interfaces if la_version returns 0 (swbz#24122) +- nptl: Reinstate pthread_timedjoin_np as a cancellation point (swbz#24215) +- nptl: Fix invalid Systemtap probe in pthread_join (swbz#24211) + +* Tue Feb 19 2019 Florian Weimer - 2.29-8 +- Drop glibc-rh1674280.patch. Different fix applied upstream. (#1674280) +- Auto-sync with upstream branch release/2.29/master, + commit 067fc32968b601493f4b247a3ac00caeea3f3d61: +- nptl: Fix invalid Systemtap probe in pthread_join (#1674280) + +* Mon Feb 11 2019 Florian Weimer - 2.29-7 +- Hotfix for invalid Systemtap probe in pthread_join (#1674280) + +* Mon Feb 11 2019 Florian Weimer - 2.29-6 +- Remove LRA bug on POWER workaround, fixed in gcc-9.0.1-0.4.fc30 (#1673018) + +* Mon Feb 11 2019 Florian Weimer - 2.29-5 +- Auto-sync with upstream branch release/2.29/master, + commit c096b008d2671028c21ac8cf01f18a2083e73c44: +- nptl: Avoid fork handler lock for async-signal-safe fork (swbz#24161) +- nptl: Add compiler barriers in pthread_mutex_trylock (swbz#24180) + +* Thu Feb 7 2019 Florian Weimer - 2.29-4 +- Work around LRA hang on ppc64le (#1673018) + +* Wed Feb 06 2019 Florian Weimer - 2.29-3 +- Auto-sync with upstream branch release/2.29/master, + commit 2de15ac95713a238dc258eb8977ecdfca811fc19: +- arm: Use "nr" constraint for Systemtap probes (#1196181) + +* Fri Feb 1 2019 Florian Weimer - 2.29-2 +- Eliminate %%glibcrelease macro. +- Switch to regular Release: pattern. + +* Thu Jan 31 2019 Carlos O'Donell - 2.29-1 +- Auto-sync with upstream branch release/2.29/master, + commit 86013ef5cea322b8f4b9c22f230c22cce369e947. +- nptl: Fix pthread_rwlock_try*lock stalls (swbz#23844) + +* Thu Jan 31 2019 Fedora Release Engineering +- Rebuilt for https://fedoraproject.org/wiki/Fedora_30_Mass_Rebuild + +* Mon Jan 28 2019 DJ Delorie - 2.28.9000-37 +- Auto-sync with upstream branch master, + commit e1e47c912a8e557508362715f7468091def3ec4f. +- Update translations. +* Mon Jan 28 2019 Florian Weimer - 2.28.9000-36 +- resolv: Enable full ICMP error reporting in stub resolver (#1670028) + +* Mon Jan 28 2019 Florian Weimer - 2.28.9000-35 +- Remove obsolete scriptlets + +* Fri Jan 25 2019 Florian Weimer - 2.28.9000-34 +- Auto-sync with upstream branch master, + commit 83e6b59625f45db1eee93e5684091f740c52a083: +- elf: Revert LD_AUDIT fix for modules with invalid version (swbz#24122) +- strftime: Pass the additional flags from "%EY" to "%Ey" (swbz#24096) +- strftime: Set the default width of "%Ey" to 2 (swbz#23758) + +* Thu Jan 24 2019 Florian Weimer - 2.28.9000-33 +- Auto-sync with upstream branch master, + commit 3367acdb344a1d7fcf8f53748d301d652c8911dd: +- elf: Fix LD_AUDIT for modules with invalid version (swbz#24122) +- CVE-2016-10739: getaddrinfo: Fully parse IPv4 address strings (#1331390) +- resolv: Do not send queries for non-host-names in nss_dns (swbz#24112) +- malloc: Revert fastbins to old-style atomics + +* Wed Jan 23 2019 Florian Weimer - 2.28.9000-32 +- Use assembler to produce annobin notes for nonshared libraries (#1668822) + +* Wed Jan 16 2019 Carlos O'Donell - 2.28.9000-31 +- Auto-sync with upstream branch master, + commit 008b598e2a495024f9777006716cfd8668f3db33. +- x86-64: Optimize strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2 +- powerpc: Fix VSCR position in ucontext (swbz#24088) +- AArch64: Add ifunc support for Ares +- soft-fp: Properly check _FP_W_TYPE_SIZE (swbz#24066) + +* Thu Jan 10 2019 Florian Weimer - 2.28.9000-30 +- Inherit -march=haswell flag from redhat-rpm-config + +* Mon Jan 07 2019 Arjun Shankar - 2.28.9000-29 +- Auto-sync with upstream branch master, + commit 2ef427168818ce04b03cecb7b739f9db0156e3e4. +- Require GCC 5 or later to build glibc (swbz#23993) +- Only build libm with -fno-math-errno (swbz#24024) +- sysdeps/ieee754: prevent maybe-uninitialized errors with -O (swbz#19444) +- Multiple locales: Use the correct 12-hour time formats (swbz#10496) +- sq_AL: Use the correct date and time formats (swbz#10496, swbz#23724) +- en_US: define date_fmt (swbz#24046) +- Remove executable bit from localedata/locales/bi_VU (swbz#23995) +- malloc: Always call memcpy in _int_realloc (swbz#24027) +- ARM: fix kernel assisted atomics with GCC 8 (swbz#24034) +- S390: Unify 31/64bit mem{set,cmp,cpy,pcpy}, bzero with ifunc handling +- S390: Refactor ifunc handling for several str* and wc* functions +- posix: Clear close-on-exec for posix_spawn adddup2 (swbz#23640) +- termios: Define TIOCSER_TEMT with __USE_MISC (swbz#17783) +- termios: Consolidate Baud Rate Selection definitions (swbz#23783) +- Y2038: add __{localtime64,gmttime64,ctime64}[_r] functions +- Y2038: make __difftime compatible with 64-bit time + +* Mon Dec 17 2018 DJ Delorie - 2.28.9000-28 +- Auto-sync with upstream branch master, + commit 6bbfc5c09fc5b5e3d4a0cddbbd4e2e457767dae7. +- Add statx conditionals for wordsize-32 *xstat.c +- regex: fix storage-exhaustion error (swbz#18040) +- regex: fix heap-use-after-free error (swbz#18040) +- manual: Document thread/task IDs for Linux + +* Thu Dec 13 2018 Carlos O'Donell - 2.28.9000-27 +- Auto-sync with upstream branch master, + commit ade8b817fead73b302d08c88cd44ea2ea56793d4. +- powerpc: missing CFI register information in __mpn_* functions (swbz#23614) +- rdlock stalls indefinitely on an unlocked pthread rwlock (swbz#23861) + +* Tue Dec 11 2018 Carlos O'Donell - 2.28.9000-26 +- Auto-sync with upstream branch master, + commit 505b5b292293a5d6bd4046a6bc7f8c2381a33da4. +- Fix powf overflow handling in non-nearest rounding mode (swbz#23961) +- test-container: move postclean outside of namespace changes (swbz#23948) +- Enable VDSO for static linking on mips (swbz#19767) + +* Mon Dec 10 2018 Florian Weimer - 2.28.9000-25 +- Auto-sync with upstream branch master, + commit 8d20a2f414fa52aceef8a0e3675415df54a840db: +- compat getdents64: Use correct offset for retry (swbz#23972) +- x86: Fix static analysis warning in tst-cet-property-2 (swbz#23490) +- malloc: Add another test for tcache double free check (swbz#23907) +- getcpu: New system call wrapper +- posix_spawn_file_actions_addfchdir_np: New function (swbz#17405) +- Preparations for the IBM long double transition +- Preparations for 64-bit time_t on 32-bit architectures +- Enable VDSO for static linking on arm (swbz#19767) + +* Mon Dec 03 2018 Florian Weimer - 2.28.9000-24 +- Auto-sync with upstream branch master, + commit 7b36d26b22d147ffc347f427f9fd584700578a94: +- CVE-2018-19591: if_nametoindex: Fix descriptor for overlong name (#1654000) +- Mutex: Add pthread mutex tunables +- stdlib: assert on NULL function pointer in atexit etc. (swbz#20544) +- Enable VDSO on i386 statically linked programs +- posix: Use posix_spawn on system +- posix: Use posix_spawn on popen (swbz#17490, swbz#22834) +- Fix _dl_profile_fixup data-dependency issue (swbz#23690) +- Enable VDSO for static linking on aarch64 + +* Thu Nov 29 2018 Carlos O'Donell - 2.28.9000-23 +- Move requirement on libgcc from glibc-devel to glibc (#1352973) + +* Tue Nov 27 2018 Carlos O'Donell - 2.28.9000-22 +- Add requires on explicit glibc version for glibc-nss-devel (#1651260) + +* Tue Nov 27 2018 Arjun Shankar - 2.28.9000-21 +- Drop glibc-rh1652495.patch. Applied upstream. (#1652495) +- Auto-sync with upstream branch master, + commit ce035c6e909ad20ef2fe13c92eab4e69f6495b61. + +* Mon Nov 26 2018 Florian Weimer - 2.28.9000-20 +- Do not use parallel make for building locales (#1652228) + +* Thu Nov 22 2018 Florian Weimer - 2.28.9000-19 +- malloc: Revert tcache double-free check (#1652495) + +* Tue Nov 20 2018 DJ Delorie - 2.28.9000-18 +- Auto-sync with upstream branch master, + commit bcdaad21d4635931d1bd3b54a7894276925d081d. +- malloc: tcache double free check +- [AArch64] Adjust writeback in non-zero memset +- Update config.guess and config.sub to current versions. +- support: Print timestamps in timeout handler +- Use STRFMON_LDBL_IS_DBL instead of __ldbl_is_dbl. + +* Fri Nov 16 2018 Florian Weimer - 2.28.9000-17 +- Auto-sync with upstream branch master, + commit 346ef23f197a0c8ba807c344bd39101b711050ee. + +* Fri Nov 09 2018 Florian Weimer - 2.28.9000-16 +- Auto-sync with upstream branch master, + commit 43257c335ad6b9e05fb882975e1776ff763164ee: +- Disable CET for binaries created by older link editors (#1648297) + +* Wed Nov 07 2018 Florian Weimer - 2.28.9000-15 +- Auto-sync with upstream branch master, + commit 00c86a37d1b63044e3169d1f2ebec23447c73f79. + +* Wed Nov 07 2018 Florian Weimer - 2.28.9000-14 +- Auto-sync with upstream branch master, + commit 1df872fd74f730bcae3df201a229195445d2e18a: +- libanl: Fix crash if first helper thread creation failed (#1646381) + +* Thu Nov 1 2018 Mike FABIAN - 2.28.9000-13 +- Include Esperanto (eo) in glibc-all-langpacks (#1643756) + +* Mon Oct 29 2018 DJ Delorie - 2.28.9000-12 +- Auto-sync with upstream branch master, + commit c6982f7efc1c70fe2d6160a87ee44d871ac85ab0. + +* Fri Oct 26 2018 Arjun Shankar - 2.28.9000-11 +- Auto-sync with upstream branch master, + commit fe61f17cfc18f17befca3280e828bb40e8c772b0. + +* Wed Oct 24 2018 Florian Weimer - 2.28.9000-10 +- Auto-sync with upstream branch master, + commit a27a4f4721837a5fb36ace833764b06a64c5af1c. + +* Thu Oct 18 2018 Florian Weimer - 2.28.9000-9 +- Auto-sync with upstream branch master, + commit 72771e53753647111d31c5c4bf43d8901e6baf7e. + +* Thu Sep 27 2018 Florian Weimer - 2.28.9000-8 +- Auto-sync with upstream branch master, + commit f841c97e515a1673485a2b12b3c280073d737890: +- stdlib/tst-setcontext9 test suite failure on ppc64le (#1623519) + +* Thu Sep 20 2018 Florian Weimer - 2.28.9000-7 +- Auto-sync with upstream branch master, + commit db9a8ad4ff3fc58e3773a9a4d0cabe3c1bc9c94c: +- gethostid: Missing NULL check for gethostbyname_r result (#1631338) +- stdlib/tst-setcontext9 test suite failure (#1623519) + +* Wed Sep 5 2018 Florian Weimer - 2.28.9000-6 +- Add python3-devel build dependency for downstream benefit + +* Wed Sep 05 2018 Carlos O'Donell - 2.28.9000-5 +- Provide compatibility support for linking against libpthread_nonshared.a + (#1625507) + +* Wed Aug 29 2018 Florian Weimer - 2.28.9000-4 +- Remove workaround for valgrind bug (#1600034) + +* Wed Aug 29 2018 Florian Weimer - 2.28.9000-3 +- Auto-sync with upstream branch master, + commit ff6b24501f70da7d6375d6f5929262b9509db39e. + +* Mon Aug 27 2018 Florian Weimer - 2.28.9000-2 +- Auto-sync with upstream branch master, + commit 99ea93ca31795469d2a1f1570f17a5c39c2eb7e2: +- nptl: Fix waiters-after-spinning case in pthread_cond_broadcast (#1622669) +- regex: Fix memory corruption when pattern and input contain NUL (#1622674) + +* Tue Aug 21 2018 Carlos O'Donell - 2.28.9000-1 +- Drop glibc-asflags.patch. Applied upstream. +- Drop glibc-rh1614705.patch. Applied upstream. +- Drop glibc-with-nonshared-cflags.patch. Applied upstream. +- Auto-sync with upstream branch master, + commit aa42b3dbcb0326badf377fec2c7fb2f34fdabecd. + +* Mon Aug 13 2018 Carlos O'Donell - 2.28-5 +- Remove abort() warning in manual (#1615608) + +* Fri Aug 10 2018 Florian Weimer - 2.28-4 +- Fix regression in readdir64@GLIBC_2.1 compat symbol (#1614705) + +* Thu Aug 2 2018 Florian Weimer - 2.28-3 +- Log /proc/sysinfo if available (on s390x) + +* Thu Aug 2 2018 Florian Weimer - 2.28-2 +- Honor %%{valgrind_arches} + +* Wed Aug 01 2018 Florian Weimer - 2.27.9000-43 +- Update to glibc 2.28 release tarball: +- Translation updates +- x86/CET: Fix property note parser (swbz#23467) +- x86: Add tst-get-cpu-features-static to $(tests) (swbz#23458) + +* Mon Jul 30 2018 Florian Weimer - 2.27.9000-42 +- Auto-sync with upstream branch master, + commit af86087f02a5522d8801a11d8381e04f95e33162: +- x86/CET: Don't parse beyond the note end +- Fix Linux fcntl OFD locks tests on unsupported kernels +- x86: Populate COMMON_CPUID_INDEX_80000001 for Intel CPUs (swbz#23459) +- x86: Correct index_cpu_LZCNT (swbz#23456) +- Fix string/tst-xbzero-opt if build with gcc head + +* Thu Jul 26 2018 Florian Weimer - 2.27.9000-41 +- Build with --enable-cet on x86_64, i686 +- Auto-sync with upstream branch master, + commit cfba5dbb10cc3abde632b46c60c10b2843917035: +- Keep expected behaviour for [a-z] and [A-z] (#1607286) +- Additional ucontext tests +- Intel CET enhancements +- ISO C11 threads support +- Fix out-of-bounds access in IBM-1390 converter (swbz#23448) +- New locale Yakut (Sakha) for Russia (sah_RU) (swbz#22241) +- os_RU: Add alternative month names (swbz#23140) +- powerpc64: Always restore TOC on longjmp (swbz#21895) +- dsb_DE locale: Fix syntax error and add tests (swbz#23208) +- Improve performance of the generic strstr implementation +- regcomp: Fix off-by-one bug in build_equiv_class (swbz#23396) +- Fix out of bounds access in findidxwc (swbz#23442) + +* Fri Jul 13 2018 Carlos O'Donell - 2.27.9000-40 +- Fix file list for glibc RPM packaging (#1601011). + +* Wed Jul 11 2018 Florian Weimer - 2.27.9000-39 +- Add POWER9 multilib (downstream only) + +* Wed Jul 11 2018 Florian Weimer - 2.27.9000-38 +- Auto-sync with upstream branch master, + commit 93304f5f7a32f73b551266c5a181db51d97a71e4: +- Install header +- Put the correct Unicode version number 11.0.0 into the generated files + +* Wed Jul 11 2018 Florian Weimer - 2.27.9000-37 +- Work around valgrind issue on i686 (#1600034) + +* Tue Jul 10 2018 Florian Weimer - 2.27.9000-36 +- Auto-sync with upstream branch master, + commit fd70af45528d59a00eb3190ef6706cb299488fcd: +- Add the statx function +- regexec: Fix off-by-one bug in weight comparison (#1582229) +- nss_files: Fix re-reading of long lines (swbz#18991) +- aarch64: add HWCAP_ATOMICS to HWCAP_IMPORTANT +- aarch64: Remove HWCAP_CPUID from HWCAP_IMPORTANT +- conform/conformtest.pl: Escape literal braces in regular expressions +- x86: Use AVX_Fast_Unaligned_Load from Zen onwards. + +* Fri Jul 6 2018 Florian Weimer - 2.27.9000-35 +- Remove ppc64 multilibs + +* Fri Jul 06 2018 Florian Weimer - 2.27.9000-34 +- Auto-sync with upstream branch master, + commit 3a885c1f51b18852869a91cf59a1b39da1595c7a. + +* Thu Jul 5 2018 Florian Weimer - 2.27.9000-33 +- Enable build flags inheritance for nonshared flags + +* Wed Jul 4 2018 Florian Weimer - 2.27.9000-32 +- Add annobin annotations to assembler code (#1548438) + +* Wed Jul 4 2018 Florian Weimer - 2.27.9000-31 +- Enable -D_FORTIFY_SOURCE=2 for nonshared code + +* Mon Jul 02 2018 Florian Weimer - 2.27.9000-30 +- Auto-sync with upstream branch master, + commit b7b88cea4151d85eafd7ababc2e4b7ae1daeedf5: +- New locale: dsb_DE (Lower Sorbian) + +* Fri Jun 29 2018 Florian Weimer - 2.27.9000-29 +- Drop glibc-deprecate_libcrypt.patch. Variant applied upstream. (#1566464) +- Drop glibc-linux-timespec-header-compat.patch. Upstreamed. +- Auto-sync with upstream branch master, + commit e69d994a63afc2d367f286a2a7df28cbf710f0fe. + +* Thu Jun 28 2018 Florian Weimer - 2.27.9000-28 +- Drop glibc-rh1315108.patch. extend_alloca was removed upstream. (#1315108) +- Auto-sync with upstream branch master, + commit c49e18222e4c40f21586dabced8a49732d946917. + +* Thu Jun 21 2018 Florian Weimer - 2.27.9000-27 +- Compatibility fix for and + +* Thu Jun 21 2018 Florian Weimer - 2.27.9000-26 +- Auto-sync with upstream branch master, + commit f496b28e61d0342f579bf794c71b80e9c7d0b1b5. + +* Mon Jun 18 2018 Florian Weimer - 2.27.9000-25 +- Auto-sync with upstream branch master, + commit f2857da7cdb65bfad75ee30981f5b2fde5bbb1dc. + +* Mon Jun 18 2018 Florian Weimer - 2.27.9000-24 +- Auto-sync with upstream branch master, + commit 14beef7575099f6373f9a45b4656f1e3675f7372: +- iconv: Make IBM273 equivalent to ISO-8859-1 (#1592270) + +* Mon Jun 18 2018 Florian Weimer - 2.27.9000-23 +- Inherit the -msse2 build flag as well (#1592212) + +* Fri Jun 01 2018 Florian Weimer - 2.27.9000-22 +- Modernise nsswitch.conf defaults (#1581809) +- Adjust build flags inheritence from redhat-rpm-config +- Auto-sync with upstream branch master, + commit 104502102c6fa322515ba0bb3c95c05c3185da7a. + +* Fri May 25 2018 Florian Weimer - 2.27.9000-21 +- Auto-sync with upstream branch master, + commit c1dc1e1b34873db79dfbfa8f2f0a2abbe28c0514. + +* Wed May 23 2018 Florian Weimer - 2.27.9000-20 +- Auto-sync with upstream branch master, + commit 7f9f1ecb710eac4d65bb02785ddf288cac098323: +- CVE-2018-11237: Buffer overflow in __mempcpy_avx512_no_vzeroupper (#1581275) +- Drop glibc-rh1452750-allocate_once.patch, + glibc-rh1452750-libidn2.patch. Applied upstream. + +* Wed May 23 2018 Florian Weimer - 2.27.9000-19 +- Auto-sync with upstream branch master, + commit 8f145c77123a565b816f918969e0e35ee5b89153. + +* Thu May 17 2018 Florian Weimer - 2.27.9000-18 +- Do not run telinit u on upgrades (#1579225) +- Auto-sync with upstream branch master, + commit 632a6cbe44cdd41dba7242887992cdca7b42922a. + +* Fri May 11 2018 Florian Weimer - 2.27.9000-17 +- Avoid exporting some Sun RPC symbols with default versions (#1577210) +- Inherit the -mstackrealign flag if it is set +- Inherit compiler flags in the original order +- Auto-sync with upstream branch master, + commit 89aacb513eb77549a29df2638913a0f8178cf3f5: +- CVE-2018-11236: realpath: Fix path length overflow (#1581270, swbz#22786) + +* Fri May 11 2018 Florian Weimer - 2.27.9000-16 +- Use /usr/bin/python3 for benchmarks scripts (#1577223) + +* Thu Apr 19 2018 Florian Weimer - 2.27.9000-15 +- Auto-sync with upstream branch master, + commit 0085be1415a38b40a5a1a12e49368498f1687380. + +* Mon Apr 09 2018 Florian Weimer - 2.27.9000-14 +- Auto-sync with upstream branch master, + commit 583a27d525ae189bdfaa6784021b92a9a1dae12e. + +* Thu Mar 29 2018 Florian Weimer - 2.27.9000-13 +- Auto-sync with upstream branch master, + commit d39c0a459ef32a41daac4840859bf304d931adab: +- CVE-2017-18269: memory corruption in i386 memmove (#1580934) + +* Mon Mar 19 2018 Florian Weimer - 2.27.9000-12 +- Auto-sync with upstream branch master, + commit fbce6f7260c3847f14dfa38f60c9111978fb33a5. + +* Fri Mar 16 2018 Florian Weimer - 2.27.9000-11 +- Auto-sync with upstream branch master, + commit 700593fdd7aef1e36cfa8bad969faab76a6facda. + +* Wed Mar 14 2018 Florian Weimer - 2.27.9000-10 +- Auto-sync with upstream branch master, + commit 7108f1f944792ac68332967015d5e6418c5ccc88. + +* Mon Mar 12 2018 Florian Weimer - 2.27.9000-9 +- Auto-sync with upstream branch master, + commit da6d4404ecfd7eacba8c096b0761a5758a59da4b. + +* Tue Mar 6 2018 Florian Weimer - 2.27.9000-8 +- Enable annobin annotations (#1548438) + +* Thu Mar 01 2018 Florian Weimer - 2.27.9000-7 +- Auto-sync with upstream branch master, + commit 1a2f44a848663036c8a14671fe0faa3fed0b2a25: +- Remove spurios reference to libpthread_nonshared.a + +* Thu Mar 01 2018 Florian Weimer - 2.27.9000-6 +- Switch back to upstream master branch +- Drop glibc-rh1013801.patch, applied upstream. +- Drop glibc-fedora-nptl-linklibc.patch, no longer needed. +- Auto-sync with upstream branch master, + commit bd60ce86520b781ca24b99b2555e2ad389bbfeaa. + +* Wed Feb 28 2018 Florian Weimer - 2.27-5 +- Inherit as many flags as possible from redhat-rpm-config (#1550914) + +* Mon Feb 19 2018 Richard W.M. Jones - 2.27-4 +- riscv64: Add symlink from /usr/lib64/lp64d -> /usr/lib64 for ABI compat. +- riscv64: Disable valgrind smoke test on this architecture. + +* Wed Feb 14 2018 Florian Weimer - 2.27-3 +- Spec file cleanups: + - Remove %%defattr(-,root,root) + - Use shell to run ldconfig %%transfiletrigger + - Move %%transfiletrigger* to the glibc-common subpackage + - Trim changelog + - Include ChangeLog.old in the source RPM + +* Wed Feb 7 2018 Florian Weimer - 2.27-2.1 +- Linux: use reserved name __key in pkey_get (#1542643) +- Auto-sync with upstream branch release/2.27/master, + commit 56170e064e2b21ce204f0817733e92f1730541ea. + +* Wed Feb 07 2018 Fedora Release Engineering +- Rebuilt for https://fedoraproject.org/wiki/Fedora_28_Mass_Rebuild + +* Mon Feb 05 2018 Carlos O'Donell - 2.27-1 +- Update to released glibc 2.27. +- Auto-sync with upstream branch master, + commit 23158b08a0908f381459f273a984c6fd328363cb. + +* Tue Jan 30 2018 Richard W.M. Jones - 2.26.9000-52 +- Disable -fstack-clash-protection on riscv64: + not supported even by GCC 7.3.1 on this architecture. + +* Mon Jan 29 2018 Florian Weimer - 2.26.9000-51 +- Explicitly run ldconfig in the buildroot +- Do not run ldconfig from scriptlets +- Put triggers into the glibc-common package, do not pass arguments to ldconfig + +* Mon Jan 29 2018 Florian Weimer - 2.26.9000-50 +- Auto-sync with upstream branch master, + commit cdd14619a713ab41e26ba700add4880604324dbb: +- libnsl: Turn remaining symbols into compat symbols (swbz#22701) +- be_BY, be_BY@latin, lt_LT, el_CY, el_GR, ru_RU, ru_UA, uk_UA: + Add alternative month names (swbz#10871) +- x86: Revert Intel CET changes to __jmp_buf_tag (swbz#22743) +- aarch64: Revert the change of the __reserved member of mcontext_t + +* Mon Jan 29 2018 Igor Gnatenko - 2.26.9000-49 +- Add file triggers to do ldconfig calls automatically + +* Mon Jan 22 2018 Florian Weimer - 2.26.9000-48 +- Auto-sync with upstream branch master, + commit 21c0696cdef617517de6e25711958c40455c554f: +- locale: Implement alternative month names (swbz#10871) +- locale: Change month names for pl_PL (swbz#10871) + +* Mon Jan 22 2018 Florian Weimer - 2.26.9000-47 +- Unconditionally build without libcrypt + +* Fri Jan 19 2018 Björn Esser - 2.26.9000-46 +- Remove deprecated libcrypt, gets replaced by libxcrypt +- Add applicable Requires on libxcrypt + +* Fri Jan 19 2018 Florian Weimer - 2.26.9000-45 +- Drop static PIE support on aarch64. It leads to crashes at run time. +- Remove glibc-rpcgen subpackage. See rpcsvc-proto. (#1531540) + +* Fri Jan 19 2018 Florian Weimer - 2.26.9000-44 +- Correct the list of static PIE architectures (#1247050) +- glibc_post_upgrade: Remove process restart logic +- glibc_post_upgrade: Integrate into the build process +- glibc_post_upgrade: Do not clean up tls subdirectories +- glibc_post_upgrade: Drop ia64 support +- Remove architecture-specific symbolic link for iconvconfig +- Auto-sync with upstream branch master, + commit 4612268a0ad8e3409d8ce2314dd2dd8ee0af5269: +- powerpc: Fix syscalls during early process initialization (swbz#22685) + +* Fri Jan 19 2018 Florian Weimer - 2.26.9000-43 +- Enable static PIE support on i386, x86_64 (#1247050) +- Remove add-on support (already gone upstream) +- Rework test suite status reporting +- Auto-sync with upstream branch master, + commit 64f63cb4583ecc1ba16c7253aacc192b6d088511: +- malloc: Fix integer overflows in memalign and malloc functions (swbz#22343) +- x86-64: Properly align La_x86_64_retval to VEC_SIZE (swbz#22715) +- aarch64: Update bits/hwcap.h for Linux 4.15 +- Add NT_ARM_SVE to elf.h + +* Wed Jan 17 2018 Florian Weimer - 2.26.9000-42 +- CVE-2017-14062, CVE-2016-6261, CVE-2016-6263: + Use libidn2 for IDNA support (#1452750) + +* Mon Jan 15 2018 Florian Weimer - 2.26.9000-41 +- CVE-2018-1000001: Make getcwd fail if it cannot obtain an absolute path + (#1533837) +- elf: Synchronize DF_1_* flags with binutils (#1439328) +- Auto-sync with upstream branch master, + commit 860b0240a5645edd6490161de3f8d1d1f2786025: +- aarch64: fix static pie enabled libc when main is in a shared library +- malloc: Ensure that the consolidated fast chunk has a sane size + +* Fri Jan 12 2018 Florian Weimer - 2.26.9000-40 +- libnsl: Do not install libnsl.so, libnsl.a (#1531540) +- Use unversioned Supplements: for langpacks (#1490725) +- Auto-sync with upstream branch master, + commit 9a08a366a7e7ddffe62113a9ffe5e50605ea0924: +- hu_HU locale: Avoid double space (swbz#22657) +- math: Make default libc_feholdsetround_noex_ctx use __feholdexcept + (swbz#22702) + +* Thu Jan 11 2018 Florian Weimer - 2.26.9000-39 +- nptl: Open libgcc.so with RTLD_NOW during pthread_cancel (#1527887) +- Introduce libnsl subpackage and remove NIS headers (#1531540) +- Use versioned Obsoletes: for libcrypt-nss. +- Auto-sync with upstream branch master, + commit 08c6e95234c60a5c2f37532d1111acf084f39345: +- nptl: Add tst-minstack-cancel, tst-minstack-exit (swbz#22636) +- math: ldbl-128ibm log1pl (-qNaN) spurious "invalid" exception (swbz#22693) + +* Wed Jan 10 2018 Florian Weimer - 2.26.9000-38 +- nptl: Fix stack guard size accounting (#1527887) +- Remove invalid Obsoletes: on glibc-header provides +- Require python3 instead of python during builds +- Auto-sync with upstream branch master, + commit 09085ede12fb9650f286bdcd805609ae69f80618: +- math: ldbl-128ibm lrintl/lroundl missing "invalid" exceptions (swbz#22690) +- x86-64: Add sincosf with vector FMA + +* Mon Jan 8 2018 Florian Weimer - 2.26.9000-37 +- Add glibc-rpcgen subpackage, until the replacement is packaged (#1531540) + +* Mon Jan 08 2018 Florian Weimer - 2.26.9000-36 +- Auto-sync with upstream branch master, + commit 579396ee082565ab5f42ff166a264891223b7b82: +- nptl: Add test for callee-saved register restore in pthread_exit +- getrlimit64: fix for 32-bit configurations with default version >= 2.2 +- elf: Add linux-4.15 VDSO hash for RISC-V +- elf: Add RISC-V dynamic relocations to elf.h +- powerpc: Fix error message during relocation overflow +- prlimit: Replace old_rlimit RLIM64_INFINITY with RLIM_INFINITY (swbz#22678) + +* Fri Jan 05 2018 Florian Weimer - 2.26.9000-35 +- Remove sln (#1531546) +- Remove Sun RPC interfaces (#1531540) +- Rebuild with newer GCC to fix pthread_exit stack unwinding issue (#1529549) +- Auto-sync with upstream branch master, + commit f1a844ac6389ea4e111afc019323ca982b5b027d: +- CVE-2017-16997: elf: Check for empty tokens before DST expansion (#1526866) +- i386: In makecontext, align the stack before calling exit (swbz#22667) +- x86, armhfp: sync sys/ptrace.h with Linux 4.15 (swbz#22433) +- elf: check for rpath emptiness before making a copy of it +- elf: remove redundant is_path argument +- elf: remove redundant code from is_dst +- elf: remove redundant code from _dl_dst_substitute +- scandir: fix wrong assumption about errno (swbz#17804) +- Deprecate external use of libio.h and _G_config.h + +* Fri Dec 22 2017 Florian Weimer - 2.26.9000-34 +- Auto-sync with upstream branch master, + commit bad7a0c81f501fbbcc79af9eaa4b8254441c4a1f: +- copy_file_range: New function to copy file data +- nptl: Consolidate pthread_{timed,try}join{_np} +- nptl: Implement pthread_self in libc.so (swbz#22635) +- math: Provide a C++ version of iseqsig (swbz#22377) +- elf: remove redundant __libc_enable_secure check from fillin_rpath +- math: Avoid signed shift overflow in pow (swbz#21309) +- x86: Add feature_1 to tcbhead_t (swbz#22563) +- x86: Update cancel_jmp_buf to match __jmp_buf_tag (swbz#22563) +- ld.so: Examine GLRO to detect inactive loader (swbz#20204) +- nscd: Fix nscd readlink argument aliasing (swbz#22446) +- elf: do not substitute dst in $LD_LIBRARY_PATH twice (swbz#22627) +- ldconfig: set LC_COLLATE to C (swbz#22505) +- math: New generic sincosf +- powerpc: st{r,p}cpy optimization for aligned strings +- CVE-2017-1000409: Count in expanded path in _dl_init_path (#1524867) +- CVE-2017-1000408: Compute correct array size in _dl_init_paths (#1524867) +- x86-64: Remove sysdeps/x86_64/fpu/s_cosf.S +- aarch64: Improve strcmp unaligned performance + +* Wed Dec 13 2017 Florian Weimer - 2.26.9000-33 +- Remove power6 platform directory (#1522675) + +* Wed Dec 13 2017 Florian Weimer - 2.26.9000-32 +- Obsolete the libcrypt-nss subpackage (#1525396) +- armhfp: Disable -fstack-clash-protection due to GCC bug (#1522678) +- ppc64: Disable power6 multilib due to GCC bug (#1522675) +- Auto-sync with upstream branch master, + commit 243b63337c2c02f30ec3a988ecc44bc0f6ffa0ad: +- libio: Free backup area when it not required (swbz#22415) +- math: Fix nextafter and nexttoward declaration (swbz#22593) +- math: New generic cosf +- powerpc: POWER8 memcpy optimization for cached memory +- x86-64: Add sinf with FMA +- x86-64: Remove sysdeps/x86_64/fpu/s_sinf.S +- math: Fix ctanh (0 + i NaN), ctanh (0 + i Inf) (swbz#22568) +- lt_LT locale: Base collation on copy "iso14651_t1" (swbz#22524) +- math: Add _Float32 function aliases +- math: Make cacosh (0 + iNaN) return NaN + i pi/2 (swbz#22561) +- hsb_DE locale: Base collation on copy "iso14651_t1" (swbz#22515) + +* Wed Dec 06 2017 Florian Weimer - 2.26.9000-31 +- Add elision tunables. Drop related configure flag. (#1383986) +- Auto-sync with upstream branch master, + commit 37ac8e635a29810318f6d79902102e2e96b2b5bf: +- Linux: Implement interfaces for memory protection keys +- math: Add _Float64, _Float32x function aliases +- math: Use sign as double for reduced case in sinf +- math: fix sinf(NAN) +- math: s_sinf.c: Replace floor with simple casts +- et_EE locale: Base collation on iso14651_t1 (swbz#22517) +- tr_TR locale: Base collation on iso14651_t1 (swbz#22527) +- hr_HR locale: Avoid single code points for digraphs in LC_TIME (swbz#10580) +- S390: Fix backtrace in vdso functions + +* Mon Dec 04 2017 Florian Weimer - 2.26.9000-30 +- Add build dependency on bison +- Auto-sync with upstream branch master, + commit 7863a7118112fe502e8020a0db0fa74fef281f29: +- math: New generic sinf (swbz#5997) +- is_IS locale: Base collation on iso14651_t1 (swbz#22519) +- intl: Improve reproducibility by using bison (swbz#22432) +- sr_RS, bs_BA locales: make collation rules the same as for hr_HR (wbz#22534) +- hr_HR locale: various updates (swbz#10580) +- x86: Make a space in jmpbuf for shadow stack pointer +- CVE-2017-17426: malloc: Fix integer overflow in tcache (swbz#22375) +- locale: make forward accent sorting the default in collating (swbz#17750) + +* Wed Nov 29 2017 Florian Weimer - 2.26.9000-29 +- Enable -fstack-clash-protection (#1512531) +- Auto-sync with upstream branch master, + commit a55430cb0e261834ce7a4e118dd9e0f2b7fb14bc: +- elf: Properly compute offsets of note descriptor and next note (swbz#22370) +- cs_CZ locale: Base collation on iso14651_t1 (swbz#22336) +- Implement the mlock2 function +- Add _Float64x function aliases +- elf: Consolidate link map sorting +- pl_PL locale: Base collation on iso14651_t1 (swbz#22469) +- nss: Export nscd hash function as __nss_hash (swbz#22459) + +* Thu Nov 23 2017 Florian Weimer - 2.26.9000-28 +- Auto-sync with upstream branch master, + commit cccb6d4e87053ed63c74aee063fa84eb63ebf7b8: +- sigwait can fail with EINTR (#1516394) +- Add memfd_create function +- resolv: Fix p_secstodate overflow handling (swbz#22463) +- resolv: Obsolete p_secstodate +- Avoid use of strlen in getlogin_r (swbz#22447) +- lv_LV locale: fix collation (swbz#15537) +- S390: Add cfi information for start routines in order to stop unwinding +- aarch64: Optimized memset for falkor + +* Sun Nov 19 2017 Florian Weimer - 2.26.9000-27 +- Auto-sync with upstream branch master, + commit f6e965ee94b37289f64ecd3253021541f7c214c3: +- powerpc: AT_HWCAP2 bit PPC_FEATURE2_HTM_NO_SUSPEND +- aarch64: Add HWCAP_DCPOP bit +- ttyname, ttyname_r: Don't bail prematurely (swbz#22145) +- signal: Optimize sigrelse implementation +- inet: Check length of ifname in if_nametoindex (swbz#22442) +- malloc: Account for all heaps in an arena in malloc_info (swbz#22439) +- malloc: Add missing arena lock in malloc_info (swbz#22408) +- malloc: Use __builtin_tgmath in tgmath.h with GCC 8 (swbz#21660) +- locale: Replaced unicode sequences in the ASCII printable range +- resolv: More precise checks in res_hnok, res_dnok (swbz#22409, swbz#22412) +- resolv: ns_name_pton should report trailing \ as error (swbz#22413) +- locale: mfe_MU, miq_NI, an_ES, kab_DZ, om_ET: Escape / in d_fmt (swbz#22403) + +* Tue Nov 07 2017 Florian Weimer - 2.26.9000-26 +- Auto-sync with upstream branch master, + commit 6b86036452b9ac47b4ee7789a50f2f37df7ecc4f: +- CVE-2017-15804: glob: Fix buffer overflow during GLOB_TILDE unescaping +- powerpc: Use latest string function optimization for internal function calls +- math: No _Float128 support for ppc64le -mlong-double-64 (swbz#22402) +- tpi_PG locale: Fix wrong d_fmt +- aarch64: Disable lazy symbol binding of TLSDESC +- tpi_PG locale: fix syntax error (swbz#22382) +- i586: Use conditional branches in strcpy.S (swbz#22353) +- ffsl, ffsll: Declare under __USE_MISC, not just __USE_GNU +- csb_PL locale: Fix abmon/mon for March (swbz#19485) +- locale: Various yesstr/nostr/yesexpr/noexpr fixes (swbz#15260, swbz#15261) +- localedef: Add --no-warnings/--warnings option +- powerpc: Replace lxvd2x/stxvd2x with lvx/stvx in P7's memcpy/memmove +- locale: Use ASCII as much as possible in LC_MESSAGES +- Add new locale yuw_PG (swbz#20952) +- malloc: Add single-threaded path to malloc/realloc/calloc/memalloc +- i386: Replace assembly versions of e_powf with generic e_powf.c +- i386: Replace assembly versions of e_log2f with generic e_log2f.c +- x86-64: Add powf with FMA +- x86-64: Add logf with FMA +- i386: Replace assembly versions of e_logf with generic e_logf.c +- i386: Replace assembly versions of e_exp2f with generic e_exp2f.c +- x86-64: Add exp2f with FMA +- i386: Replace assembly versions of e_expf with generic e_expf.c + +* Sat Oct 21 2017 Florian Weimer - 2.26.9000-25 +- Auto-sync with upstream branch master, + commit 797ba44ba27521261f94cc521f1c2ca74f650147: +- math: Add bits/floatn.h defines for more _FloatN / _FloatNx types +- posix: Fix improper assert in Linux posix_spawn (swbz#22273) +- x86-64: Use fxsave/xsave/xsavec in _dl_runtime_resolve (swbz#21265) +- CVE-2017-15670: glob: Fix one-byte overflow (#1504807) +- malloc: Add single-threaded path to _int_free +- locale: Add new locale kab_DZ (swbz#18812) +- locale: Add new locale shn_MM (swbz#13605) + +* Fri Oct 20 2017 Florian Weimer - 2.26.9000-24 +- Use make -O to serialize make output +- Auto-sync with upstream branch master, + commit 63b4baa44e8d22501c433c4093aa3310f91b6aa2: +- sysconf: Fix missing definition of UIO_MAXIOV on Linux (#1504165) +- Install correct bits/long-double.h for MIPS64 (swbz#22322) +- malloc: Fix deadlock in _int_free consistency check +- x86-64: Don't set GLRO(dl_platform) to NULL (swbz#22299) +- math: Add _Float128 function aliases +- locale: Add new locale mjw_IN (swbz#13994) +- aarch64: Rewrite elf_machine_load_address using _DYNAMIC symbol +- powerpc: fix check-before-set in SET_RESTORE_ROUND +- locale: Use U+202F as thousands separators in pl_PL locale (swbz#16777) +- math: Use __f128 to define FLT128_* constants in include/float.h for old GCC +- malloc: Improve malloc initialization sequence (swbz#22159) +- malloc: Use relaxed atomics for malloc have_fastchunks +- locale: New locale ca_ES@valencia (swbz#2522) +- math: Let signbit use the builtin in C++ mode with gcc < 6.x (swbz#22296) +- locale: Place monetary symbol in el_GR, el_CY after the amount (swbz#22019) + +* Tue Oct 17 2017 Florian Weimer - 2.26.9000-23 +- Switch to .9000 version numbers during development + +* Tue Oct 17 2017 Florian Weimer - 2.26.90-22 +- Auto-sync with upstream branch master, + commit c38a4bfd596db2be2b9c1f96715bdc833eab760a: +- malloc: Use compat_symbol_reference in libmcheck (swbz#22050) + +* Mon Oct 16 2017 Florian Weimer - 2.26.90-21 +- Auto-sync with upstream branch master, + commit 596f70134a8f11967c65c1d55a94a3a2718c731d: +- Silence -O3 -Wall warning in malloc/hooks.c with GCC 7 (swbz#22052) +- locale: No warning for non-symbolic character (swbz#22295) +- locale: Allow "" int_curr_Symbol (swbz#22294) +- locale: Fix localedef exit code (swbz#22292) +- nptl: Preserve error in setxid thread broadcast in coredumps (swbz#22153) +- powerpc: Avoid putting floating point values in memory (swbz#22189) +- powerpc: Fix the carry bit on mpn_[add|sub]_n on POWER7 (swbz#22142) +- Support profiling PIE (swbz#22284) + +* Wed Oct 11 2017 Florian Weimer - 2.26.90-20 +- Auto-sync with upstream branch master, + commit d8425e116cdd954fea0c04c0f406179b5daebbb3: +- nss_files performance issue in multi mode (swbz#22078) +- Ensure C99 and C11 interfaces are available for C++ (swbz#21326) + +* Mon Oct 09 2017 Florian Weimer - 2.26.90-19 +- Move /var/db/Makefile to nss_db (#1498900) +- Auto-sync with upstream branch master, + commit 645ac9aaf89e3311949828546df6334322f48933: +- openpty: use TIOCGPTPEER to open slave side fd + +* Fri Oct 06 2017 Carlos O'Donell - 2.26.90-18 +- Auto-sync with upstream master, + commit 1e26d35193efbb29239c710a4c46a64708643320. +- malloc: Fix tcache leak after thread destruction (swbz#22111) +- powerpc: Fix IFUNC for memrchr. +- aarch64: Optimized implementation of memmove for Qualcomm Falkor +- Always do locking when iterating over list of streams (swbz#15142) +- abort: Do not flush stdio streams (swbz#15436) + +* Wed Oct 04 2017 Florian Weimer - 2.26.90-17 +- Move nss_compat to the main glibc package (#1400538) +- Auto-sync with upstream master, + commit 11c4f5010c58029e73e656d5df4f8f42c9b8e877: +- crypt: Use NSPR header files in addition to NSS header files (#1489339) +- math: Fix yn(n,0) without SVID wrapper (swbz#22244) +- math: Fix log2(0) and log(10) in downward rounding (swbz#22243) +- math: Add C++ versions of iscanonical for ldbl-96, ldbl-128ibm (swbz#22235) +- powerpc: Optimize memrchr for power8 +- Hide various internal functions (swbz#18822) + +* Sat Sep 30 2017 Florian Weimer - 2.26.90-16 +- Auto-sync with upstream master, + commit 1e2bffd05c36a9be30d7092d6593a9e9aa009ada: +- Add IBM858 charset (#1416405) +- Update kernel version in syscall-names.list to 4.13 +- Add Linux 4.13 constants to bits/fcntl-linux.h +- Add fcntl sealing interfaces from Linux 3.17 to bits/fcntl-linux.h +- math: New generic powf, log2f, logf +- Fix nearbyint arithmetic moved before feholdexcept (swbz#22225) +- Mark __dso_handle as hidden (swbz#18822) +- Skip PT_DYNAMIC segment with p_filesz == 0 (swbz#22101) +- glob now matches dangling symbolic links (swbz#866, swbz#22183) +- nscd: Release read lock after resetting timeout (swbz#22161) +- Avoid __MATH_TG in C++ mode with -Os for fpclassify (swbz#22146) +- Fix dlclose/exit race (swbz#22180) +- x86: Add SSE4.1 trunc, truncf (swbz#20142) +- Fix atexit/exit race (swbz#14333) +- Use execveat syscall in fexecve (swbz#22134) +- Enable unwind info in libc-start.c and backtrace.c +- powerpc: Avoid misaligned stores in memset +- powerpc: build some IFUNC math functions for libc and libm (swbz#21745) +- Removed redundant data (LC_TIME and LC_MESSAGES) for niu_NZ (swbz#22023) +- Fix LC_TELEPHONE for az_AZ (swbz#22112) +- x86: Add MathVec_Prefer_No_AVX512 to cpu-features (swbz#21967) +- x86: Add x86_64 to x86-64 HWCAP (swbz#22093) +- Finish change from “Bengali” to “Bangla” (swbz#14925) +- posix: fix glob bugs with long login names (swbz#1062) +- posix: Fix getpwnam_r usage (swbz#1062) +- posix: accept inode 0 is a valid inode number (swbz#19971) +- Remove redundant LC_TIME data in om_KE (swbz#22100) +- Remove remaining _HAVE_STRING_ARCH_* definitions (swbz#18858) +- resolv: Fix memory leak with OOM during resolv.conf parsing (swbz#22095) +- Add miq_NI locale for Miskito (swbz#20498) +- Fix bits/math-finite.h exp10 condition (swbz#22082) + +* Mon Sep 04 2017 Florian Weimer - 2.26.90-15 +- Auto-sync with upstream master, + commit b38042f51430974642616a60afbbf96fd0b98659: +- Implement tmpfile with O_TMPFILE (swbz#21530) +- Obsolete pow10 functions +- math.h: Warn about an already-defined log macro + +* Fri Sep 01 2017 Florian Weimer - 2.26.90-14 +- Build glibc with -O2 (following the upstream default). +- Auto-sync with upstream master, + commit f4a6be2582b8dfe8adfa68da3dd8decf566b3983: +- malloc: Abort on heap corruption, without a backtrace (swbz#21754) +- getaddrinfo: Return EAI_NODATA for gethostbyname2_r with NO_DATA (swbz#21922) +- getaddrinfo: Fix error handling in gethosts (swbz#21915) (swbz#21922) +- Place $(elf-objpfx)sofini.os last (swbz#22051) +- Various locale fixes (swbz#15332, swbz#22044) + +* Wed Aug 30 2017 Florian Weimer - 2.26.90-13 +- Drop glibc-rh952799.patch, applied upstream (#952799, swbz#22025) +- Auto-sync with upstream master, + commit 5f9409b787c5758fc277f8d1baf7478b752b775d: +- Various locale fixes (swbz#22022, swbz#22038, swbz#21951, swbz#13805, + swbz#21971, swbz#21959) +- MIPS/o32: Fix internal_syscall5/6/7 (swbz#21956) +- AArch64: Fix procfs.h not to expose stdint.h types +- iconv_open: Fix heap corruption on gconv_init failure (swbz#22026) +- iconv: Mangle __btowc_fct even without __init_fct (swbz#22025) +- Fix bits/math-finite.h _MSUF_ expansion namespace (swbz#22028) +- Provide a C++ version of iszero that does not use __MATH_TG (swbz#21930) + +* Mon Aug 28 2017 Florian Weimer - 2.26.90-12 +- Auto-sync with upstream master, + commit 2dba5ce7b8115d6a2789bf279892263621088e74. + +* Fri Aug 25 2017 Florian Weimer - 2.26.90-11 +- Auto-sync with upstream master, + commit 3d7b66f66cb223e899a7ebc0f4c20f13e711c9e0: +- string/stratcliff.c: Replace int with size_t (swbz#21982) +- Fix tgmath.h handling of complex integers (swbz#21684) + +* Thu Aug 24 2017 Florian Weimer - 2.26.90-10 +- Use an architecture-independent system call list (#1484729) +- Drop glibc-fedora-include-bits-ldbl.patch (#1482105) + +* Tue Aug 22 2017 Florian Weimer - 2.26.90-9 +- Auto-sync with upstream master, + commit 80f91666fed71fa3dd5eb5618739147cc731bc89. + +* Mon Aug 21 2017 Florian Weimer - 2.26.90-8 +- Auto-sync with upstream master, + commit a8410a5fc9305c316633a5a3033f3927b759be35: +- Obsolete matherr, _LIB_VERSION, libieee.a. + +* Mon Aug 21 2017 Florian Weimer - 2.26.90-7 +- Auto-sync with upstream master, + commit 4504783c0f65b7074204c6126c6255ed89d6594e. + +* Mon Aug 21 2017 Florian Weimer - 2.26.90-6 +- Auto-sync with upstream master, + commit b5889d25e9bf944a89fdd7bcabf3b6c6f6bb6f7c: +- assert: Support types without operator== (int) (#1483005) + +* Mon Aug 21 2017 Florian Weimer - 2.26.90-5 +- Auto-sync with upstream master, + commit 2585d7b839559e665d5723734862fbe62264b25d: +- Do not use generic selection in C++ mode +- Do not use __builtin_types_compatible_p in C++ mode (#1481205) +- x86-64: Check FMA_Usable in ifunc-mathvec-avx2.h (swbz#21966) +- Various locale fixes (swbz#21750, swbz#21960, swbz#21959, swbz#19852) +- Fix sigval namespace (swbz#21944) +- x86-64: Optimize e_expf with FMA (swbz#21912) +- Adjust glibc-rh827510.patch. + +* Wed Aug 16 2017 Tomasz Kłoczko - 2.26-4 +- Remove 'Buildroot' tag, 'Group' tag, and '%%clean' section, and don't + remove the buildroot in '%%install', all per Fedora Packaging Guidelines + (#1476839) + +* Wed Aug 16 2017 Florian Weimer - 2.26.90-3 +- Auto-sync with upstream master, + commit 403143e1df85dadd374f304bd891be0cd7573e3b: +- x86-64: Align L(SP_RANGE)/L(SP_INF_0) to 8 bytes (swbz#21955) +- powerpc: Add values from Linux 4.8 to +- S390: Add new s390 platform z14. +- Various locale fixes (swbz#14925, swbz#20008, swbz#20482, swbz#12349 + swbz#19982, swbz#20756, swbz#20756, swbz#21836, swbz#17563, swbz#16905, + swbz#21920, swbz#21854) +- NSS: Replace exported NSS lookup functions with stubs (swbz#21962) +- i386: Do not set internal_function +- assert: Suppress pedantic warning caused by statement expression (swbz#21242) +- powerpc: Restrict xssqrtqp operands to Vector Registers (swbz#21941) +- sys/ptrace.h: remove obsolete PTRACE_SEIZE_DEVEL constant (swbz#21928) +- Remove __qaddr_t, __long_double_t +- Fix uc_* namespace (swbz#21457) +- nss: Call __resolv_context_put before early return in get*_r (swbz#21932) +- aarch64: Optimized memcpy for Qualcomm Falkor processor +- manual: Document getcontext uc_stack value on Linux (swbz#759) +- i386: Add (swbz#21913) +- Don't use IFUNC resolver for longjmp or system in libpthread (swbz#21041) +- Fix XPG4.2 bits/sigaction.h namespace (swbz#21899) +- x86-64: Add FMA multiarch functions to libm +- i386: Support static PIE in start.S +- Compile tst-prelink.c without PIE (swbz#21815) +- x86-64: Use _dl_runtime_resolve_opt only with AVX512F (swbz#21871) +- x86: Remove __memset_zero_constant_len_parameter (swbz#21790) + +* Wed Aug 16 2017 Florian Weimer - 2.26-2 +- Disable multi-arch (IFUNC string functions) on i686 (#1471427) +- Remove nosegneg 32-bit Xen PV support libraries (#1482027) +- Adjust spec file to RPM changes + +* Thu Aug 03 2017 Carlos O'Donell - 2.26-1 +- Update to released glibc 2.26. +- Auto-sync with upstream master, + commit 2aad4b04ad7b17a2e6b0e66d2cb4bc559376617b. +- getaddrinfo: Release resolver context on error in gethosts (swbz#21885) + +* Wed Aug 02 2017 Fedora Release Engineering - 2.25.90-30.1 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_27_Binutils_Mass_Rebuild + +* Sat Jul 29 2017 Florian Weimer - 2.25.90-30 +- Auto-sync with upstream master, + commit 5920a4a624b1f4db310d1c44997b640e2a4653e5: +- mutex: Fix robust mutex lock acquire (swbz#21778) + +* Fri Jul 28 2017 Florian Weimer - 2.25.90-29 +- Auto-sync with upstream master, + commit d95fcb2df478efbf4f8537ba898374043ac4561f: +- rwlock: Fix explicit hand-over (swbz#21298) +- tunables: Use direct syscall for access (swbz#21744) +- Avoid accessing corrupted stack from __stack_chk_fail (swbz#21752) +- Remove extra semicolons in struct pthread_mutex (swbz#21804) +- grp: Fix cast-after-dereference (another big-endian group merge issue) +- S390: fix sys/ptrace.h to make it includible again after asm/ptrace.h +- Don't add stack_chk_fail_local.o to libc.a (swbz#21740) +- i386: Test memmove_chk and memset_chk only in libc.so (swbz#21741) +- Add new locales az_IR, mai_NP (swbz#14172) +- Various locale improvements + +* Thu Jul 27 2017 Carlos O'Donell - 2.25.90-28 +- Adjust to new rpm debuginfo generation (#1475009). + +* Wed Jul 26 2017 Fedora Release Engineering - 2.25.90-27.1 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_27_Mass_Rebuild + +* Wed Jul 19 2017 Florian Weimer - 2.25.90-27 +- Auto-sync with upstream master, + commit 00d7a3777369bac3d8d44152dde2bb7381984ef6: +- aarch64: Fix out of bound array access in _dl_hwcap_string + +* Mon Jul 17 2017 Florian Weimer - 2.25.90-26 +- Drop glibc-rh1467518.patch in favor of upstream patch (#1467518) +- Auto-sync with upstream master, + commit 91ac3a7d8474480685632cd25f844d3154c69fdf: +- Fix pointer alignment in NSS group merge result construction (#1471985) +- Various locale fixes + +* Fri Jul 14 2017 Carlos O'Donell - 2.25.90-25 +- armv7hl: Drop 32-bit ARM build fix, already in upstream master. +- s390x: Apply glibc fix again, removing PTRACE_GETREGS etc. (#1469536). +- Auto-sync with upstream master, + commit de895ddcd7fc45caeeeb0ae312311b8bd31d82c5: +- Added Fiji Hindi language locale for Fiji (swbz#21694). +- Added yesstr/nostr for nds_DE and nds_NL (swbz#21756). +- Added yesstr and nostr for Tigrinya (swbz#21759). +- Fix LC_MESSAGES and LC_ADDRESS for anp_IN (swbz#21760). +- Added yesstr/nostr and fix yesexpr for pap_AW and pap_CW (swbz#21757). +- Added Tongan language locale for Tonga (swbz#21728). +- [ARM] Fix ld.so crash when built using Binutils 2.29. +- Added yesstr and nostr for aa_ET (swbz#21768). +- New locale for bi_VU (swbz#21767). +- Disable single thread optimization for open_memstream + +* Wed Jul 12 2017 Carlos O'Donell - 2.25.90-24 +- Fix IFUNC crash in early startup for ppc64le static binaries (#1467518). +- Enable building with BIND_NOW on ppc64le (#1467518). +- Fix 32-bit ARM builds in presence of new binutils. + +* Wed Jul 12 2017 Florian Weimer - 2.25.90-23 +- malloc: Tell GCC optimizers about MAX_FAST_SIZE in _int_malloc (#1470060) +- Auto-sync with upstream master, + commit 30200427a99e5ddac9bad08599418d44d54aa9aa: +- Add per-thread cache to malloc +- Add Samoan language locale for Samoa +- Add Awajún / Aguaruna locale for Peru +- CVE-2010-3192: Avoid backtrace from __stack_chk_fail (swbz#12189) +- Add preadv2, writev2 RWF_NOWAIT flag (swbz#21738) +- Fix abday strings for ar_JO/ar_LB/ar_SY locales (swbz#21749) +- Fix abday strings for ar_SA locale (swbz#21748, swbz#19066) +- Set data_fmt for da_DK locale (swbz#17297) +- Add yesstr and nostr for the zh_HK locale (swbz#21733) +- Fix abday strings for the ksIN@devanagari locale (swbz#21743) +- Do not include _dl_resolv_conflicts in libc.a (swbz#21742) +- Test __memmove_chk, __memset_chk only in libc.so (swbz#21741) +- Add iI and eE to yesexpr and noexpr respectively for ts_ZA locale +- Add yesstr/nostr for kw_GB locale (swbz#21734) +- Add yesstr and nostr for the ts_ZA locale (swbz#21727) +- Fix LC_NAME for hi_IN locale (swbz#21729) +- Add yesstr and nostr for the xh_ZA locale (swbz#21724) +- Add yesstr and nostr for the zh_CN locale (swbz#21723) +- Fix full weekday names for the ks_IN@devanagari locale (swbz#21721) +- Various fixes to Arabic locales after CLDR import + +* Tue Jul 11 2017 Florian Weimer - 2.25.90-22 +- Reinstantiate stack_t cleanup (#1468904) +- s390x: Restore PTRACE_GETREGS etc. to get GCC to build (#1469536) + +* Sun Jul 9 2017 Florian Weimer - 2.25.90-21 +- Back out stack_t cleanup (#1468904) + +* Thu Jul 06 2017 Florian Weimer - 2.25.90-20 +- Auto-sync with upstream master, + commit 031e519c95c069abe4e4c7c59e2b4b67efccdee5: +- x86-64: Align the stack in __tls_get_addr (#1440287) +- Add Tok-Pisin (tpi_PG) locale. +- Add missing yesstr/nostr for Pashto locale (swbz#21711) +- Add missing yesstr/nostr for Breton locale (swbz#21706) +- Single threaded stdio optimization +- sysconf: Use conservative default for _SC_NPROCESSORS_ONLN (swbz#21542) + +* Tue Jul 04 2017 Florian Weimer - 2.25.90-19 +- Auto-sync with upstream master, + commit 4446a885f3aeb3a33b95c72bae1f115bed77f0cb. + +* Tue Jul 04 2017 Florian Weimer - 2.25.90-18 +- Auto-sync with upstream master, + commit 89f6307c5d270ed4f11cee373031fa9f2222f2b9. + +* Tue Jul 4 2017 Florian Weimer - 2.25.90-17 +- Disable building with BIND_NOW on ppc64le (#1467518) + +* Mon Jul 03 2017 Florian Weimer - 2.25.90-16 +- Auto-sync with upstream master, + commit e237357a5a0559dee92261f1914d1fa2cd43a1a8: +- Support an arbitrary number of search domains in the stub resolver (#168253) +- Detect and apply /etc/resolv.conf changes in libresolv (#1374239) +- Increase malloc alignment on i386 to 16 (swbz#21120) +- Make RES_ROTATE start with a random name server (swbz#19570) +- Fix tgmath.h totalorder, totalordermag return type (swbz#21687) +- Miscellaneous sys/ucontext.h namespace fixes (swbz#21457) +- Rename struct ucontext tag (swbz#21457) +- Call exit system call directly in clone (swbz#21512) +- powerpc64le: Enable float128 +- getaddrinfo: Merge IPv6 addresses and IPv4 addresses (swbz#21295) +- Avoid .symver on common symbols (swbz#21666) +- inet_pton: Reject IPv6 addresses with many leading zeros (swbz#16637) + +* Fri Jun 23 2017 Florian Weimer - 2.25.90-15 +- Auto-sync with upstream master, + commit 3ec7c02cc3e922b9364dc8cfd1d4546671b91003, fixing: +- memcmp-avx2-movbe.S incorrect results for lengths 2/3 (#1464403) + +* Fri Jun 23 2017 Florian Weimer - 2.25.90-14 +- Auto-sync with upstream master, + commit 12f50337ae80672c393c2317d471d097ad92c492, changing: +- localedata: fur_IT: Fix spelling of Wednesday (Miercus) +- Update to Unicode 10.0.0 +- inet: __inet6_scopeid_pton should accept node-local addresses (swbz#21657) + +* Fri Jun 23 2017 Florian Weimer - 2.25.90-13 +- Reenable valgrind on aarch64 + +* Thu Jun 22 2017 Florian Weimer - 2.25.90-12 +- Log auxiliary vector during build + +* Thu Jun 22 2017 Florian Weimer - 2.25.90-11 +- Auto-sync with upstream master, + commit 0a47d031e44f15236bcef8aeba80e737bd013c6f. + +* Thu Jun 22 2017 Florian Weimer - 2.25.90-10 +- Disable valgrind on aarch64 + +* Wed Jun 21 2017 Florian Weimer - 2.25.90-9 +- Drop historic aarch64 TLS patches +- Drop workaround for GCC PR69537 +- Auto-sync with upstream master, + commit 9649350d2ee47fae00794d57e2526aa5d67d900e. + +* Wed Jun 21 2017 Florian Weimer - 2.25.90-8 +- Adjust build requirements for gcc, binutils, kernel-headers. +- Auto-sync with upstream master, + commit 43e0ac24c836eed627a75ca932eb7e64698407c6, changing: +- Remove + +* Mon Jun 19 2017 Florian Weimer - 2.25.90-7 +- Drop glibc-Disable-buf-NULL-in-login-tst-ptsname.c, applied upstream. +- Auto-sync with upstream master, + commit 37e9dc814636915afb88d0779e5e897e90e7b8c0, fixing: +- CVE-2017-1000366: Avoid large allocas in the dynamic linker (#1462820) +- wait3 namespace (swbz#21625) +- S390: Sync ptrace.h with kernel (swbz#21539) +- Another x86 sys/ucontext.h namespace issue (swbz#21457) +- siginterrupt namespace (swbz#21597) +- Signal stack namespace (swbz#21584) +- Define struct rusage in sys/wait.h when required (swbz#21575) +- S390: Fix build with gcc configured with --enable-default-pie (swbz#21537) +- Update timezone code from tzcode 2017b +- nptl: Invert the mmap/mprotect logic on allocated stacks (swbz#18988) +- PowerPC64 ELFv2 PPC64_OPT_LOCALENTRY +- Make copy of from GCC (swbz#21573) +- localedata: ce_RU: update weekdays from CLDR (swbz#21207) +- localedata: Remove trailing spaces (swbz#20275) +- XPG4 bsd_signal namespace (swbz#21552) +- Correct collation rules for Malayalam (swbz#19922, swbz#19919) +- waitid namespace (swbz#21561) +- Condition signal.h inclusion in sys/wait.h (swbz#21560) +- ld.so: Consolidate 2 strtouls into _dl_strtoul (swbz#21528) +- tst-timezone race (swbz#14096) +- Define SIG_HOLD for XPG4 (swbz#21538) +- struct sigaltstack namespace (swbz#21517) +- sigevent namespace (swbz#21543) +- Add shim header for bits/syscall.h (swbz#21514) +- namespace issues in sys/ucontext.h (swbz#21457) +- posix: Implement preadv2 and pwritev2 +- Various float128 and tunables improvements + +* Tue Jun 06 2017 Stephen Gallagher - 2.25.90-6 +- Reduce libcrypt-nss dependency to 'Suggests:' + +* Wed May 31 2017 Arjun Shankar - 2.25.90-5 +- Auto-sync with upstream master, + commit cfa9bb61cd09c40def96f042a3123ec0093c4ad0. +- Fix sys/ucontext.h namespace from signal.h etc. inclusion (swbz#21457) +- Fix sigstack namespace (swbz#21511) + +* Wed May 31 2017 Arjun Shankar - 2.25.90-4 +- Disable the NULL buffer test in login/tst-ptsname.c. It leads to a build + failure during 'make check'. A permanent solution is being discussed + upstream. + +* Tue May 23 2017 Arjun Shankar - 2.25.90-3 +- Auto-sync with upstream master, + commit 231a59ce2c5719d2d77752c21092960e28837b4a. +- Add el_GR@euro support (swbz#20686) +- Set dl_platform and dl_hwcap from CPU features (swbz#21391) +- Use __glibc_reserved convention in mcontext, sigcontext (swbz#21457) +- Fix signal.h bsd_signal namespace (swbz#21445) +- Fix network headers stdint.h namespace (swbz#21455) +- resolv: Use RES_DFLRETRY consistently (swbz#21474) +- Condition some sys/ucontext.h contents on __USE_MISC (swbz#21457) +- Consolidate Linux read syscall (swbz#21428) +- fork: Remove bogus parent PID assertions (swbz#21386) +- Reduce value of LD_HWCAP_MASK for tst-env-setuid test case (swbz#21502) +- libio: Avoid dup already opened file descriptor (swbz#21393) + +* Mon May 01 2017 Carlos O'Donell - 2.25.90-2 +- Auto-sync with upstream master, + commit 25e39b4229fb365a605dc4c8f5d6426a77bc08a6. +- logbl for POWER7 return incorrect results (swbz#21280) +- sys/socket.h uio.h namespace (swbz#21426) +- Support POSIX_SPAWN_SETSID (swbz#21340) +- Document how to provide a malloc replacement (swbz#20424) +- Verify that all internal sockets opened with SOCK_CLOEXEC (swbz#15722) +- Use AVX2 memcpy/memset on Skylake server (swbz#21396) +- unwind-dw2-fde deadlock when using AddressSanitizer (swbz#21357) +- resolv: Reduce advertised EDNS0 buffer size to guard against + fragmentation attacks (swbz#21361) +- mmap64 silently truncates large offset values (swbz#21270) +- _dl_map_segments does not test for __mprotect failures consistently + (swbz#20831) + +* Thu Mar 02 2017 Florian Weimer - 2.25.90-1 +- Switch back to upstream master branch. +- Drop Unicode 9 patch, merged upstream. +- Auto-sync with upstream master, + commit a10e9c4e53fc652b79abf838f7f837589d2c84db, fixing: +- Build all DSOs with BIND_NOW (#1406731) + +* Wed Mar 1 2017 Jakub Hrozek - 2.25-3 +- NSS: Prefer sss service for passwd, group databases (#1427646) + +* Tue Feb 28 2017 Florian Weimer - 2.25-2 +- Auto-sync with upstream release/2.25/master, + commit 93cf93e06ce123439e41d3d62790601c313134cb, fixing: +- sunrpc: Improvements for UDP client timeout handling (#1346406) +- sunrpc: Avoid use-after-free read access in clntudp_call (swbz#21115) +- Fix getting tunable values on big-endian (swbz#21109) + +* Wed Feb 08 2017 Carlos O'Donell - 2.25-1 +- Update to final released glibc 2.25. + +* Wed Feb 08 2017 Carlos O'Donell - 2.24.90-31 +- Fix builds with GCC 7.0. + +* Wed Feb 01 2017 Carlos O'Donell - 2.24.90-30 +- Optimize IBM z System builds for zEC12. + +* Wed Jan 25 2017 Florian Weimer - 2.24.90-29 +- Use vpath in crypt-glibc/Makefile to obtain the test input file. +- Auto-sync with upstream master, + commit 5653ab12b4ae15b32d41de7c56b2a4626cd0437a, fixing: +- ARM fpu_control.h for assemblers requiring VFP insn names (swbz#21047) +- FAIL in test string/tst-xbzero-opt (swbz#21006) +- Make soft-float powerpc swapcontext restore the signal mask (swbz#21045) +- Clear list of acquired robust mutexes in the child after fork (swbz#19402) + +* Thu Jan 12 2017 Carlos O'Donell - 2.24.90-28 +- Auto-sync with upstream master, + commit 468e525c81a4af10f2e613289b6ff7c950773a9e: +- Drop rwlock related patches applied upstream. +- Fix i686 memchr for large input sizes (swbz#21014) +- Fix x86 strncat for large input sizes (swbz#19390) +- powerpc: Fix write-after-destroy in lock elision (swbz#20822) +- New pthread rwlock that is more scalable. +- Fix testsuite build for GCC 7 -Wformat-truncation. + +* Mon Jan 02 2017 Florian Weimer - 2.24.90-27 +- Auto-sync with upstream master, + commit 73dfd088936b9237599e4ab737c7ae2ea7d710e1: +- Enable tunables. +- Drop condvar-related patches applied upstream. +- Update DNS RR type definitions (swbz#20593) +- CVE-2015-5180: resolv: Fix crash with internal QTYPE (#1249603) +- sunrpc: Always obtain AF_INET addresses from NSS (swbz#20964) + +* Mon Dec 26 2016 Florian Weimer - 2.24.90-26 +- Auto-sync with upstream master, + commit cecbc7967f0bcac718b6f8f8942b58403c0e917c +- Enable stack protector for most of glibc (#1406731) + +* Fri Dec 23 2016 Carlos O'Donell - 2.24.90-25 +- Auto-sync with upstream master, + commit 81e0662e5f2c342ffa413826b7b100d56677b613, fixing: +- Shared object unload assert when calling dlclose (#1398370, swbz#11941) +- Fix nss_nisplus build with mainline GCC (swbz#20978) +- Add Intel TSX blacklist for silicon with known errata. +- Add fmax, fmin, fmaxf, fminf microbenchmarks. +- Robust mutexes: Fix lost wake-up (swbz#20973). +- Add fmaxmag, fminmag, roundeven, roundevenf, roundevenl functions. + +* Sun Dec 18 2016 Florian Weimer - 2.24.90-24 +- Auto-sync with upstream master, + commit e077349ce589466eecd47213db4fae6b80ec18c4, fixing: +- Warn about assignment in assertions (#1105335) +- powerpc64/power7 memchr for large input sizes (swbz#20971) +- fmax, fmin sNaN handling (swbz#20947) + +* Mon Dec 12 2016 Florian Weimer - 2.24.90-23 +- Auto-sync with upstream master, + commit 92dcaa3e2f7bf0f7f1c04cd2fb6a317df1a4e225, fixing: +- Add getrandom, getentropy (#1172273) +- Add additional compiler barriers to backtrace tests (swbz#20956) + +* Fri Dec 09 2016 Florian Weimer - 2.24.90-22 +- Auto-sync with upstream master, + commit 0abbe7cd700951082b314182a0958d65238297ef, changing: +- IN6_IS_ADDR_ does not require enabling non-standard extensions (#1138893) +- Install libm.a as linker script (swbz#20539) +- Fix writes past the allocated array bounds in execvpe (swbz#20847) +- Fix hypot sNaN handling (swbz#20940) +- Fix x86_64/x86 powl handling of sNaN arguments (swbz#20916) +- Fix sysdeps/ieee754 pow handling of sNaN arguments (swbz#20916) +- Fix pow (qNaN, 0) result with -lieee (swbz#20919) +- Fix --enable-nss-crypt failure of tst-linkall-static (swbz#20918) + +* Fri Dec 02 2016 Florian Weimer - 2.24.90-21 +- Auto-sync with upstream master, + commit 01b23a30b42a90b1ebd882a0d81110a1542e504a, fixing: +- aarch64: Incorrect dynamic TLS resolution (#1400347) + +* Wed Nov 30 2016 Florian Weimer - 2.24.90-20 +- Auto-sync with upstream master, + commit 9e78f6f6e7134a5f299cc8de77370218f8019237, fixing: +- stdio buffering with certain network file systems (#1400144) +- libpthread initialization breaks ld.so exceptions (#1393909) +- x86_64: Use of PLT and GOT in static archives (swbz#20750) +- localedata, iconvdata: 0x80->Euro sign mapping for GBK (swbz#20864) +- math: x86_64 -mfpmath=387 float_t, double_t (swbz#20787) + +* Wed Nov 23 2016 Florian Weimer - 2.24.90-19 +- Auto-sync with upstream master, + commit 7a5e3d9d633c828d84a9535f26b202a6179978e7: +- Fix default float_t definition (swbz#20855) +- Fix writes past the allocated array bounds in execvpe (swbz#20847) + +* Tue Nov 22 2016 Florian Weimer - 2.24.90-18 +- Auto-sync with upstream master, + commit 5ee1a4443a3eb0868cef1fe506ae6fb6af33d4ad. + +* Wed Nov 16 2016 Carlos O'Donell - 2.24.90-17 +* Add new scalable implementation of POSIX read-write locks. + +* Wed Nov 16 2016 Florian Weimer - 2.24.90-16 +- Do not try to link libcrypt statically during tests + +* Wed Nov 16 2016 Florian Weimer - 2.24.90-15 +- Auto-sync with upstream master, + commit 530862a63e0929128dc98fbbd463b120934434fb, fixing: +- Fix rpcgen buffer overrun (swbz#20790) +- Fix ppc64 build failure to swbz#20729 fix attempt + +* Wed Nov 2 2016 Florian Weimer - 2.24.90-14 +- Drop glibc-swbz20019.patch, applied upstream. +- dlerror returns NULL after dlsym (RTLD_NEXT) lookup failure (#1333945) + (fixed by dropping the revert) +- Auto-sync with upstream master, + commit 9032070deaa03431921315f973c548c2c403fecc, fixing: +- Correct clog10 documentation (swbz#19673) +- Fix building with -Os (swbz#20729) +- Properly initialize glob structure with GLOB_BRACE|GLOB_DOOFFS (swbz#20707) +- powerpc: Fix TOC stub on powerpc64 clone (swbz#20728) +- math: Make strtod raise "inexact" exceptions (swbz#19380) +- malloc: Remove malloc_get_state, malloc_set_state (swbz#19473) + +* Sat Oct 22 2016 Florian Weimer - 2.24.90-13 +- Auto-sync with upstream master, + commit e37208ce86916af9510ffb9ce7b3c187986f07de, changing: +- Restore compatbility with extern "C" wrappers + +* Fri Oct 21 2016 Florian Weimer - 2.24.90-12 +- Auto-sync with upstream master, + commit b3918c44db615637b26d919ce599cd86592316b3, fixing: +- math: Turn iszero into a function template (#1387415) +- ARM: Use VSQRT instruction (swbz#20660) +- math: Stop powerpc copysignl raising "invalid" for sNaN (swbz#20718) +- x86: Fix FMA and AVX2 detection (swbz#20689) +- x86: Avoid assertion failure on older Intel CPus (swbz#20647) + +* Mon Oct 17 2016 Carlos O'Donell - 2.24.90-11 +- Add prototype support for detecting invalid IFUNC calls (swbz#20019). +- New POSIX thread condition variable implementation (swbz#13165). + +* Fri Oct 07 2016 Florian Weimer - 2.24.90-10 +- Auto-sync with upstream master, + commit 5140d036f9c16585448b5908c3a219bd96842161, fixing: +- resolv: Remove RES_USEBSTRING and its implementation (swbz#20629) +- Refactor ifunc resolvers due to false debuginfo (swbz#20478) + +* Tue Oct 04 2016 Florian Weimer - 2.24.90-9 +- Auto-sync with upstream master, + commit ff88ee7edfaa439e23c42fccaf3a36cd5f041894, fixing: +- LONG_WIDTH is incorrectly set to the 64 on 32-bit platforms (#1381582) +- libio: Multiple fixes for open_{w}memstream (swbz#18241, swbz#20181) +- Simplify and test _dl_addr_inside_object (swbz#20292) + +* Thu Sep 22 2016 Florian Weimer - 2.24.90-8 +- Add support for MIPS (#1377795) +- Drop glibc-rh1315476-1.patch (sln pre-processor cleanup), it was + applied upstream. +- Auto-sync with upstream master, + commit 17af5da98cd2c9ec958421ae2108f877e0945451, fixing the following bugs: +- Fix non-LE TLS in static programs (swbz#19826) +- resolv: Remove unsupported hook functions from the API (swbz#20016) +- Remove RR type classification macros (swbz#20592) +- Remove obsolete DNSSEC support (swbz#20591) +- manual: Clarify the documentation of strverscmp (swbz#20524) + +* Tue Sep 20 2016 Carlos O'Donell - 2.24.90-7 +- Auto-sync with upstream master. + +* Thu Sep 01 2016 Florian Weimer - 2.24.90-6 +- Auto-sync with upstream master, + commit 4d728087ef8cc826b05bd21d0c74d4eca9b1a27d, fixing: +- Base on Linux headers (#1360480) +- Simplify static malloc interposition (swbz#20432) + +* Fri Aug 26 2016 Florian Weimer - 2.24.90-5 +- Auto-sync with upstream master, + commit 7e625f7e85b4e88f10dbde35a0641742af581806, fixing: +- lt_LT locale: use hyphens in d_fmt (swbz#20497) +- nptl test time reductions (swbz#19946) + +* Sun Aug 21 2016 Florian Weimer - 2.24.90-4 +- Auto-sync with upstream master, + commit 66abf9bfbe24ac1e7207d26ccad725ed938dc52c, fixing: +- argp: Do not override GCC keywords with macros (#1366830) + +* Wed Aug 17 2016 Florian Weimer - 2.24.90-3 +- Auto-sync with upstream master, + commit d9067fca40b8aac156d73cfa44d6875813555a6c, with these changes: +- Avoid duplicating object files already in libc.a (#1352625) +- CVE-2016-6323: Backtraces can hang on ARM EABI (32-bit) (swbz#20435) +- et_EE: locale has wrong {p,n}_cs_precedes value (swbz#20459 + +* Thu Aug 11 2016 Florian Weimer - 2.24.90-2 +- Auto-sync with upstream master, + commit f79211792127f38d5954419bb3784c8eb7f5e4e5 + +* Mon Aug 08 2016 Carlos O'Donell - 2.24.90-1 +- Set version to 2.24.90 to match upstream development. + +* Mon Aug 08 2016 Carlos O'Donell - 2.23.90-31 +- Auto-sync with upstream master. + +* Thu Jul 21 2016 Florian Weimer - 2.23.90-30 +- Drop sendmsg/recvmsg compatibility patch (#1344830) +- glibc-devel depends on libgcc%%{_isa} (#1289356) +- Drop Requires(pre) on libgcc +- Introduce libcrypt and libcrypt-nss (#1324623) +- Do not try to install mtrace when bootstrapping + +* Wed Jul 20 2016 Florian Weimer - 2.23.90-29 +- Move NSS modules to subpackages (#1338889) + +* Wed Jul 13 2016 Florian Weimer - 2.23.90-28 +- Auto-sync with upstream master, commit + f531f93056b34800383c5154280e7ba5112563c7. +- Add de_LI.UTF-8 locale. +- Make ldconfig and sln the same binary. (#1315476) + +* Fri Jul 08 2016 Mike FABIAN - 2.23.90-27 +- Unicode 9.0.0 updates (ctype, charmap, transliteration) (#1351108) + +* Tue Jul 05 2016 Florian Weimer - 2.23.90-26 +- Auto-sync with upstream master, up to commit + 30e4cc5413f72c2c728a544389da0c48500d9904, fixing these bug: +- strcasecmp failure on ppc64le (#nscd breaks initgroups with nis (initgroups are empty) (#1294574) + +* Fri Jun 24 2016 Carlos O'Donell - 2.23.90-25 +- Properly handle more invalid --install-langs arguments (#1349906). + +* Tue Jun 21 2016 Florian Weimer - 2.23.90-24 +- Auto-sync with upstream master, commit + a3b473373ee43a292f5ec68a7fda6b9cfb26a9b0, fixing these bugs: +- Unnecessary mmap fallback in malloc (#1348620) +- pwritev system call passes incorrect offset to kernel (#1346070) + +* Sat Jun 18 2016 Carlos O'Donell - 2.23.90-23 +- Use scriptlet expansion in all-langpacks posttrans script to expand + _install_langes macro. + +* Mon Jun 13 2016 Florian Weimer - 2.23.90-22 +- Remove glibc-fedora-uname-getrlimit.patch. This patch was + introduced to fix bug rhbz#579086 (Preloading a replacement uname + is causing environment to be cleaned if libpthread is loaded). + UTS namespaces should now offer a cleaner way yo do this. +- Drop sendmmsg/recvmmsg compat symbols on 32-bit architectures (#1344830) +* Sat Jun 11 2016 Florian Weimer - 2.23.90-21 +- First phase of sendmsg/recvmsg/sendmmsg/recvmmsg ABI revert: + GLIBC_2.24 compatibility symbols (#1344830) +- Auto-sync with upstream master + (commit 31d0a4fa646db8b8c97ce24e0ec0a7b73de4fca1), + fixing the following bugs: +- Add eo locale +- Crash in the nss_db NSS service module during iteration (#1344480) + +* Thu Jun 09 2016 Florian Weimer - 2.23.90-20 +- Auto-sync with upstream master, fixing this bug: +- Emacs crashes on startup (#1342976) + +* Wed Jun 01 2016 Florian Weimer - 2.23.90-19 +- Auto-sync with upstream master. +- Adjust glibc-rh1315108.patch accordingly. +- Fix fork redirection in libpthread (#1326903) +- CVE-2016-4429: stack overflow in Sun RPC clntudp_call (#1337140) +- Do not disable assertions in release builds (#1338887) + +* Wed May 11 2016 Carlos O'Donell - 2.23.90-18 +- Move support for building GCC 2.96 into compat-gcc-296. + +* Wed May 11 2016 Florian Weimer - 2.23.90-17 +- Temporily revert dlsym (RTLD_NEXT)/dlerror change, to unbreak + ASAN until it is fixed (#1335011) + +* Mon May 9 2016 Florian Weimer - 2.23.90-16 +- Drop the “fix” for fork/vfork NULL symbols in libpthread. It does + not work because ld.so apparently supports some variant of direct + binding. + +* Mon May 09 2016 Florian Weimer - 2.23.90-15 +- Auto-sync with upstream master. +- Drop glibc-nsswitch-Add-group-merging-support.patch, applied upstream. +- Drop glibc-rh1252570.patch, alternative fixes applied upstream. +- Adjust glibc-rh1315108.patch to minor upstream change. +- Update SUPPORTED file. +- Experimental fix for NULL fork/vfork symbols in libpthread (#1326903) + +* Tue May 03 2016 Carlos O'Donell - 2.23.90-14 +- Require libselinux for nscd in non-bootstrap configuration. + +* Fri Apr 29 2016 Carlos O'Donell - 2.23.90-13 +- Auto-sync with upstream master. + +* Thu Apr 28 2016 Carlos O'Donell - 2.23.90-12 +- Move spec file system information logging to the build stage. + +* Thu Apr 14 2016 Florian Weimer - 2.23.90-11 +- Auto-sync with upstream master. +- Unbreak pread/pread64 on armhfp (#1327277) + +* Thu Apr 14 2016 Florian Weimer - 2.23.90-10 +- Auto-sync with upstream master. + +* Thu Apr 14 2016 Florian Weimer - 2.23.90-9 +- Auto-sync with upstream master. Removes type union wait. +- Update SUPPORTED locales file. + +* Fri Apr 08 2016 Florian Weimer - 2.23.90-8 +- Auto-sync with upstream master. + +* Tue Mar 29 2016 Florian Weimer - 2.23.90-7 +- Auto-sync with upstream master. +- Adjust glibc-rh1252570.patch to partial upstream fix. +- Drop glibc-fix-an_ES.patch, now included upstream. + +* Wed Mar 16 2016 Carlos O'Donell - 2.23.90-6 +- Use 'an' as language abbreviation for an_ES. + +* Mon Mar 07 2016 Carlos O'Donell - 2.23.90-5 +- Auto-sync with upstream master. + +* Sun Mar 6 2016 Florian Weimer - 2.23.90-4 +- Remove extend_alloca (#1315108) + +* Mon Feb 29 2016 Carlos O'Donell - 2.23.90-3 +- Enhance support for upgrading from a non-language-pack system. + +* Fri Feb 26 2016 Mike FABIAN - 2.23.90-2 +- Create new language packages for all supported languages. + Locales, translations, and locale sources are split into + distinct sub-packages. A meta-package is created for users + to install all languages. Transparent installation support + is provided via dnf langpacks. + +* Fri Feb 26 2016 Carlos O'Donell - 2.23.90-1 +- Upstream development version is now 2.23.90. + +* Thu Feb 25 2016 Carlos O'Donell - 2.22.90-38 +- Auto-sync with upstream master. + +* Fri Feb 19 2016 Florian Weimer - 2.22.90-37 +- Remove stray newline from Serbian locales (#1114591). + +* Tue Feb 16 2016 CArlos O'Donell - 2.22.90-36 +- Fix CVE-2015-7547: getaddrinfo() stack-based buffer overflow (#1308943). + +* Mon Feb 15 2016 Florian Weimer - 2.22.90-35 +- Revert may_alias attribute for struct sockaddr (#1306511). +- Revert upstream commit 2212c1420c92a33b0e0bd9a34938c9814a56c0f7 (#1252570). + +* Sat Feb 13 2016 Florian Weimer - 2.22.90-34 +- Auto-sync with upstream master. +- Support aliasing with struct sockaddr pointers (#1306511). + +* Tue Feb 09 2016 Carlos O'Donell - 2.22.90-33 +- Use --with-cpu=power8 for ppc64le default runtime (#1227361). + +* Tue Feb 02 2016 Florian Weimer - 2.22.90-32 +- Auto-sync with upstream master. +- Add glibc-isinf-cxx11.patch to improve C++11 compatibility. + +* Thu Jan 28 2016 Florian Weimer - 2.22.90-31 +- Add workaround for GCC PR69537. + +* Thu Jan 28 2016 Florian Weimer - 2.22.90-30 +- Auto-sync with upstream master. + +* Wed Jan 13 2016 Carlos O'Donell - 2.22.90-29 +- New pthread_barrier algorithm with improved standards compliance. + +* Wed Jan 13 2016 Carlos O'Donell - 2.22.90-28 +- Add group merging support for distributed management (#1146822). + +* Tue Jan 12 2016 Carlos O'Donell - 2.22.90-27 +- Remove 32-bit POWER support. +- Add 64-bit POWER7 BE and 64-bit POWER8 BE optimized libraries. + +* Mon Dec 21 2015 Florian Weimer - 2.22.90-26 +- Auto-sync with upstream master. + +* Wed Dec 16 2015 Florian Weimer - 2.22.90-25 +- Auto-sync with upstream master. +- Includes fix for malloc assertion failure in get_free_list. (#1281714) +- Drop Unicode 8.0 patches (now merged upstream). + +* Sat Dec 5 2015 Florian Weimer - 2.22.90-24 +- Put libmvec_nonshared.a into the -devel package. (#1288738) + +* Sat Dec 05 2015 Florian Weimer - 2.22.90-23 +- Auto-sync with upstream master. + +* Thu Nov 26 2015 Carlos O'Donell - 2.22.90-22 +- The generic hidden directive support is already used for + preinit/init/fini-array symbols so we drop the Fedora-specific + patch that does the same thing. + Reported by Dmitry V. Levin + +* Thu Nov 26 2015 DJ Delorie - 2.22.90-22 +- Require glibc-static for C++ tests. +- Require gcc-c++, libstdc++-static, and glibc-static only when needed. +- Fix --without docs to not leave info files. + +* Fri Nov 20 2015 Florian Weimer - 2.22.90-21 +- Auto-sync with upstream master. + +* Wed Nov 18 2015 Florian Weimer - 2.22.90-20 +- Auto-sync with upstream master. + +* Wed Nov 18 2015 Florian Weimer - 2.22.90-19 +- Disable -Werror on s390 (#1283184). + +* Mon Nov 16 2015 Florian Weimer - 2.22.90-18 +- Auto-sync with upstream master. + +* Mon Nov 16 2015 Florian Weimer - 2.22.90-17 +- Revert temporary armhfp build fix. + +* Mon Nov 9 2015 Florian Weimer - 2.22.90-16 +- Apply temporary fix for armhfp build issue. + +* Mon Nov 09 2015 Florian Weimer - 2.22.90-15 +- Auto-sync with upstream master. + +* Tue Nov 3 2015 Florian Weimer - 2.22.90-14 +- Log uname, cpuinfo, meminfo during build (#1276636) + +* Fri Oct 30 2015 Florian Weimer - 2.22.90-13 +- Auto-sync with upstream master. + +* Fri Oct 30 2015 Florian Weimer - 2.22.90-12 +- Revert to upstream implementation of condition variables (#1229659) + +* Wed Oct 28 2015 Florian Weimer - 2.22.90-11 +- Disable valgrind test on ppc64p7, too. + +* Mon Oct 26 2015 Carlos O'Donell - 2.22.90-10 +- Disable valgrind test for ppc64. + +* Wed Oct 21 2015 Carlos O'Donell - 2.22.90-9 +- Sync with upstream master. +- Update new condvar implementation. + +* Fri Oct 9 2015 Carlos O'Donell - 2.22.90-8 +- Remove libbsd.a (#1193168). + +* Wed Sep 16 2015 Mike FABIAN - 2.22.90-7 +- Add the C.UTF-8 locale (#902094). + +* Wed Sep 16 2015 Carlos O'Donell - 2.22.90-6 +- Fix GCC 5 and -Werror related build failures. +- Fix --install-langs bug which causes SIGABRT (#1262040). + +* Fri Aug 28 2015 Carlos O'Donell - 2.22.90-5 +- Auto-sync with upstream master. + +* Thu Aug 27 2015 Carlos O'Donell - 2.22.90-4 +- Build require gcc-c++ for the C++ tests. +- Support --without testsuite option to disable testing after build. +- Support --without benchtests option to disable microbenchmarks. +- Update --with bootstrap to disable benchtests, valgrind, documentation, + selinux, and nss-crypt during bootstrap. +- Support --without werror to disable building with -Werror. +- Support --without docs to disable build requirement on texinfo. +- Support --without valgrind to disable testing with valgrind. +- Remove c_stubs add-on and enable fuller support for static binaries. +- Remove librtkaio support (#1227855). + +* Sun Aug 16 2015 Siddhesh Poyarekar - 2.22.90-3 +- Auto-sync with upstream master. + +* Fri Aug 14 2015 Siddhesh Poyarekar - 2.22.90-2 +- Remove initgroups from the default nsswitch.conf (#751450). + +* Fri Aug 14 2015 Siddhesh Poyarekar - 2.22.90-1 +- Sync with upstream master. + +* Tue Jul 28 2015 Siddhesh Poyarekar - 2.21.90-20 +- Sync with upstream master. + +* Thu Jul 23 2015 Mike FABIAN - 2.21.90-19 +- some more additions to the translit_neutral file by Marko Myllynen + +* Tue Jul 14 2015 Mike FABIAN - 2.21.90-18 +- Unicode 8.0.0 updates, including the transliteration files (#1238412). + +* Sun Jun 21 2015 Carlos O'Donell - 2.21.90-17 +- Remove all linuxthreads handling from glibc spec file. + +* Wed Jun 17 2015 Carlos O'Donell - 2.21.90-16 +- Move split out architecture-dependent header files into devel package + and keep generic variant in headers package, thus keeping headers package + content and file list identical across multilib rpms. + +* Wed Jun 17 2015 Fedora Release Engineering - 2.21.90-15.1 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_23_Mass_Rebuild + +* Wed Jun 3 2015 Carlos O'Donell - 2.21.90-15 +- Remove patch to increase DTV surplus which is no longer needed after + upstream commit f8aeae347377f3dfa8cbadde057adf1827fb1d44. + +* Sat May 30 2015 Siddhesh Poyarekar - 2.21.90-14 +- Fix build failure on aarch64 (#1226459). + +* Mon May 18 2015 Siddhesh Poyarekar - 2.21.90-13 +- Sync with upstream master. +- Install new condvar implementation. + +* Fri May 08 2015 Siddhesh Poyarekar - 2.21.90-12 +- Add benchmark comparison scripts. + +* Thu May 07 2015 Siddhesh Poyarekar - 2.21.90-11 +- Auto-sync with upstream master. +- Revert arena threshold fix to work around #1209451. + +* Tue Apr 07 2015 Siddhesh Poyarekar - 2.21.90-10 +- Revert last auto-sync (#1209451). + +* Mon Apr 06 2015 Siddhesh Poyarekar - 2.21.90-9 +- Auto-sync with upstream master. + +* Tue Mar 24 2015 Siddhesh Poyarekar - 2.21.90-8 +- Auto-sync with upstream master. + +* Tue Mar 17 2015 Carlos O'Donell - 2.21.90-7 +- Use rpm.expand in scripts to reduce set of required RPM features. + +* Thu Mar 12 2015 Siddhesh Poyarekar - 2.21.90-6 +- Auto-sync with upstream master. + +* Tue Mar 3 2015 Mike Fabian - 2.21.90-5 +- Support installing only those locales specified by the RPM macro + %%_install_langs (#156477). + +* Mon Feb 23 2015 Siddhesh Poyarekar - 2.21.90-4 +- Auto-sync with upstream master. + +* Sat Feb 21 2015 Till Maas - 2.21.90-3.1 +- Rebuilt for Fedora 23 Change + https://fedoraproject.org/wiki/Changes/Harden_all_packages_with_position-independent_code + +* Thu Feb 12 2015 Carlos O'Donell - 2.21.90-3 +- Fix missing clock_* IFUNCs in librtkaio. + +* Thu Feb 12 2015 Carlos O'Donell - 2.21.90-2 +- Auto-sync with upstream master. + +* Wed Feb 11 2015 Carlos O'Donell - 2.21.90-1 +- Add back x86 vDSO support. +- Fix rtkaio build to reference clock_* functions from libc. + +* Wed Jan 21 2015 Siddhesh Poyarekar - 2.20.90-20 +- Sync with upstream master. +- Disable werror on s390x. +- Revert x86 vDSO support since it breaks i686 rtkaio build. + +* Tue Jan 20 2015 Peter Robinson 2.20.90-19 +- Drop large ancient ChangeLogs (rhbz #1169546) + +* Mon Jan 12 2015 Siddhesh Poyarekar - 2.20.90-18 +- Pass address of main_arena.mutex to mutex_lock/unlock. + +* Thu Jan 08 2015 Siddhesh Poyarekar - 2.20.90-17 +- Define a __tls_get_addr macro to avoid a conflicting declaration. + +* Wed Jan 07 2015 Siddhesh Poyarekar - 2.20.90-16 +- Disable -Werror for s390 as well. + +* Wed Jan 07 2015 Siddhesh Poyarekar - 2.20.90-14 +- Sync with upstream master. +- Disable -Werror on powerpc and armv7hl. +- Temporarily disable valgrind test on ppc64. + +* Sun Dec 28 2014 Dan Horák +- valgrind available only on selected arches (missing on s390) + +* Wed Dec 10 2014 Kyle McMartin +- aarch64: Drop strchrnul.S revert, apply fix from Richard Earnshaw. + +* Fri Dec 05 2014 Carlos O'Donell - 2.20.90-13 +- Fix permission of debuginfo source files to allow multiarch + debuginfo packages to be installed and upgraded. + +* Fri Dec 05 2014 Siddhesh Poyarekar - 2.20.90-12 +- Remove LIB_LANG since we don't install locales in /usr/lib/locale anymore. +- Don't own any directories in /usr/share/locale (#1167445). +- Use the %%find_lang macro to get the *.mo files (#1167445). +- Add %%lang tags to language locale files in /usr/share/i18n/locale (#1169044). + +* Wed Dec 03 2014 Kyle McMartin - 2.20.90-11 +- aarch64: revert optimized strchrnul.S implementation (rhbz#1167501) + until it can be debugged. + +* Fri Nov 28 2014 Carlos O'Donell - 2.20.90-10 +- Auto-sync with upstream master. + +* Wed Nov 19 2014 Carlos O'Donell - 2.20.90-9 +- Sync with upstream master. + +* Wed Nov 05 2014 Siddhesh Poyarekar - 2.20.90-8 +- Make getconf return only /usr/bin (#1138835). +- Sync with upstream master. + +* Tue Nov 04 2014 Arjun Shankar - 2.20.90-7 +- Add patch that modifies several tests to use test-skeleton.c. + The patch is accepted but not yet committed upstream. + https://sourceware.org/ml/libc-alpha/2014-10/msg00744.html + +* Tue Sep 30 2014 Siddhesh Poyarekar - 2.20.90-6 +- Sync with upstream master. +- Disable more Intel TSX usage in rwlocks (#1146967). +- Enable lock elision again on s390 and s390x. +- Enable Systemtap SDT probes for all architectures (#985109). + +* Fri Sep 26 2014 Carlos O'Donell - 2.20.90-5 +- Disable lock elision support for Intel hardware until microcode + updates can be done in early bootup (#1146967). +- Fix building test tst-strtod-round for ARM. + +* Tue Sep 23 2014 Siddhesh Poyarekar - 2.20.90-4 +- Sync with upstream master. +- Don't own the common debuginfo directories (#1144853). +- Run valgrind in the %%check section to ensure that it does not break. + +* Tue Sep 16 2014 Siddhesh Poyarekar - 2.20.90-3 +- Sync with upstream master. +- Revert patch for #737223. + +* Mon Sep 08 2014 Siddhesh Poyarekar - 2.20.90-2 +- Build build-locale-archive statically again. + +* Mon Sep 08 2014 Siddhesh Poyarekar - 2.20.90-1 +- Sync with upstream master. + +* Thu Sep 4 2014 Carlos O'Donell - 2.19.90-36 +- Allow up to 32 dlopened modules to use static TLS (#1124987). +- Run glibc tests in %%check section of RPM spec file. +- Do not run tests with `-k` and fail if any test fails to build. + +* Tue Aug 26 2014 Siddhesh Poyarekar - 2.19.90-35 +- Sync with upstream master. +- Use INTERNAL_SYSCALL in TLS_INIT_TP (#1133134). +- Remove gconv loadable module transliteration support (CVE-2014-5119, #1119128). + +* Fri Aug 22 2014 Dennis Gilmore - 2.19.90-34 +- add back sss to nsswitch.conf we have added workarounds in the tools + +* Thu Aug 21 2014 Kevin Fenzi - 2.19.90-33.1 +- Rebuild for rpm bug 1131960 + +* Tue Aug 19 2014 Dennis Gilmore - 2.19.90-33 +- remove sss from default nsswitch.conf it causes issues with live image composing + +* Wed Aug 13 2014 Siddhesh Poyarekar - 2.19.90-32 +- Auto-sync with upstream master. +- Revert to only defining __extern_always_inline for g++-4.3+. +- Fix build failure in compat-gcc-32 (#186410). + +* Mon Jul 28 2014 Siddhesh Poyarekar - 2.19.90-31 +- Auto-sync with upstream master. + +* Wed Jul 23 2014 Siddhesh Poyarekar - 2.19.90-30 +- Undo last master sync to fix up rawhide. + +* Tue Jul 15 2014 Siddhesh Poyarekar - 2.19.90-29 +- Auto-sync with upstream master. + +* Sat Jul 12 2014 Tom Callaway - 2.19.90-28 +- fix license handling + +* Mon Jul 07 2014 Siddhesh Poyarekar - 2.19.90-27 +- Auto-sync with upstream master. + +* Fri Jul 04 2014 Siddhesh Poyarekar - 2.19.90-26 +- Sync with upstream roland/nptl branch. +- Improve testsuite failure outputs in build.log + +* Thu Jul 03 2014 Siddhesh Poyarekar - 2.19.90-25 +- Sync with upstream roland/nptl branch. + +* Wed Jul 02 2014 Siddhesh Poyarekar - 2.19.90-24 +- Sync with upstream master. + +* Tue Jun 24 2014 Siddhesh Poyarekar - 2.19.90-23 +- Sync with upstream master. +- Add fix to unbreak i386 ABI breakage due to a change in scalbn. + +* Fri Jun 20 2014 Kyle McMartin - 2.19.90-22 +- AArch64: Save & restore NZCV (flags) upon entry to _dl_tlsdesc_dynamic + in order to work around GCC reordering compares across the TLS + descriptor sequence (GCC PR61545.) Committing a (temporary) fix here + allows us to avoid rebuilding the world with gcc 4.9.0-11.fc21. + +* Mon Jun 16 2014 Kyle McMartin - 2.19.90-21 +- Auto-sync with upstream master. + +* Thu Jun 12 2014 Siddhesh Poyarekar - 2.19.90-20 +- Auto-sync with upstream master. + +* Sat Jun 07 2014 Fedora Release Engineering - 2.19.90-19.1 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_21_Mass_Rebuild + +* Tue Jun 03 2014 Siddhesh Poyarekar - 2.19.90-19 +- Sync with upstream master. + +* Mon May 26 2014 Siddhesh Poyarekar - 2.19.90-18 +- Sync with upstream master. +- Adjust rtkaio patches to build with upstream master. + +* Wed May 21 2014 Kyle McMartin - 2.19.90-17 +- Backport some upstream-wards patches to fix TLS issues on AArch64. + +* Wed May 21 2014 Kyle McMartin - 2.19.90-16 +- AArch64: Fix handling of nocancel syscall failures (#1098327) + +* Thu May 15 2014 Siddhesh Poyarekar - 2.19.90-15 +- Sync with upstream master. + +* Wed May 14 2014 Carlos O'Donell - 2.19.90-14 +- Add support for displaying all test results in build logs. + +* Wed May 14 2014 Carlos O'Donell - 2.19.90-13 +- Add initial support for ppc64le. + +* Tue Apr 29 2014 Siddhesh Poyarekar - 2.19.90-12 +- Auto-sync with upstream master. +- Remove ports addon. + +* Fri Apr 18 2014 Siddhesh Poyarekar - 2.19.90-11 +- Sync with upstream master. + +* Thu Apr 10 2014 Siddhesh Poyarekar - 2.19.90-10 +- Sync with upstream master. + +* Thu Apr 03 2014 Siddhesh Poyarekar - 2.19.90-9 +- Sync with upstream master. + +* Wed Mar 26 2014 Siddhesh Poyarekar - 2.19.90-8 +- Sync with upstream master. + +* Wed Mar 19 2014 Siddhesh Poyarekar - 2.19.90-7 +- Sync with upstream master. +- Fix offset computation for append+ mode on switching from read (#1078355). + +* Wed Mar 12 2014 Carlos O'Donell - 2.19.90-6 +- Sync with upstream master. +- Use cleaner upstream solution for -ftree-loop-distribute-patterns (#911307). + +* Tue Mar 04 2014 Siddhesh Poyarekar - 2.19.90-5 +- Sync with upstream master. + +* Thu Feb 27 2014 Siddhesh Poyarekar - 2.19.90-4 +- Use nscd service files from glibc sources. +- Make nscd service forking in systemd service file. + +* Tue Feb 25 2014 Siddhesh Poyarekar - 2.19.90-3 +- Sync with upstream master. +- Separate ftell from fseek logic and avoid modifying FILE data (#1069559). + +* Mon Feb 24 2014 Carlos O'Donell - 2.19.90-2 +- Fix build-locale-archive failure to open default template. + +* Tue Feb 18 2014 Siddhesh Poyarekar - 2.19.90-1 +- Sync with upstream master. + +* Tue Feb 04 2014 Siddhesh Poyarekar - 2.18.90-27 +- Sync with upstream master. + +* Wed Jan 29 2014 Siddhesh Poyarekar - 2.18.90-26 +- Modify regular expressions to include powerpcle stubs-*.h (#1058258). + +* Wed Jan 29 2014 Siddhesh Poyarekar - 2.18.90-25 +- Sync with upstream master. + +* Sat Jan 25 2014 Ville Skyttä - 2.18.90-24 +- Own the %%{_prefix}/lib/locale dir. + +* Thu Jan 23 2014 Siddhesh Poyarekar - 2.18.90-23 +- Sync with upstream master. + +* Thu Jan 16 2014 Siddhesh Poyarekar - 2.18.90-22 +- Back out ftell test case (#1052846). + +* Tue Jan 14 2014 Siddhesh Poyarekar - 2.18.90-21 +- Sync with upstream master. +- Fix infinite loop in ftell when writing wide char data (#1052846). + +* Tue Jan 7 2014 Siddhesh Poyarekar - 2.18.90-20 +- Sync with upstream master. +- Enable systemtap probes on Power and S/390. + +* Fri Dec 27 2013 Siddhesh Poyarekar - 2.18.90-19 +- Sync with upstream master. + +* Fri Dec 20 2013 Siddhesh Poyarekar - 2.18.90-18 +- Sync with upstream master. + +* Wed Dec 4 2013 Siddhesh Poyarekar - 2.18.90-17 +- Sync with upstream master. + - Fix shm_open validation (#1037787); + +* Thu Nov 28 2013 Siddhesh Poyarekar - 2.18.90-16 +- Sync with upstream master. + +* Wed Nov 20 2013 Siddhesh Poyarekar - 2.18.90-15 +- Sync with upstream master. + +* Fri Nov 8 2013 Carlos O'Donell - 2.18.90-14 +- Enhance NSCD's SELinux support to use dynamic permission names (#1025126). + +* Mon Oct 28 2013 Siddhesh Poyarekar - 2.18.90-13 +- Sync with upstream master. + - Skip over unimplemented timezone format specifier in strptime (#947722). + +* Mon Oct 21 2013 Siddhesh Poyarekar - 2.18.90-12 +- Allow fill_archive to be called with NULL fname. +- Sync with upstream master. + +* Tue Oct 15 2013 Siddhesh Poyarekar - 2.18.90-11 +- Sync with upstream master. + +* Thu Oct 3 2013 Carlos O'Donell - 2.18.90-10 +- Allow applications to use pthread_atfork without explicitly + requiring libpthread.so. (#1013801) +- Support `--list-archive FILE' in localedef utility. + +* Thu Oct 3 2013 Siddhesh Poyarekar - 2.18.90-9 +- Define swap_endianness_p in build-locale-archive. + +* Wed Oct 2 2013 Carlos O'Donell - 2.18.90-8 +- Allow ldconfig cached objects previously marked as hard or soft + ABI to now become unmarked without raising an error. This works + around a binutils bug that caused objects to become unmarked. + (#1009145) + +* Tue Oct 1 2013 Siddhesh Poyarekar - 2.18.90-7 +- Fix check for PI mutex on non-x86 systems (#1007590). +- Resync with upstream master. + +* Tue Sep 24 2013 Carlos O'Donell - 2.18.90-6 +- Avoid the use of __block which is a reserved keyword for clang++ + (#1009623). + +* Mon Sep 23 2013 Siddhesh Poyarekar - 2.18.90-5 +- Resync with upstream master. + +* Sun Sep 22 2013 Carlos O'Donell - 2.18.90-4 +- Fix CVE-2013-4788: Static applications now support pointer mangling. + Existing static applications must be recompiled (#985625). + +* Wed Sep 18 2013 Patsy Franklin - 2.18.90-3 +- Fix conditional requiring specific binutils for s390/s390x. + +* Mon Sep 16 2013 Siddhesh Poyarekar - 2.18.90-2 +- Resync with upstream master. +- Fix CVE-2013-4332 (#1008299). + +* Thu Sep 5 2013 Siddhesh Poyarekar - 2.18.90-1 +- Resync with upstream master. +- Drop patch for #800224. + +* Thu Aug 29 2013 Carlos O'Donell - 2.18-6 +- Fix Power build (#997531). + +* Wed Aug 28 2013 Carlos O'Donell - 2.18-5 +- Fix indirect function support to avoid calling optimized routines + for the wrong hardware (#985342). + +* Mon Aug 26 2013 Siddhesh Poyarekar - 2.18-4 +- Initialize res_hconf in nscd. (#1000924). + +* Tue Aug 20 2013 Siddhesh Poyarekar - 2.18-3 +- Remove non-ELF support in rtkaio. +- Avoid inlining of cleanup function for kaio_suspend. +- Expand sizes of some types in strcoll (#855399, CVE-2012-4424). +- Fix tst-aiod2 and tst-aiod3 test failures (#970865). + +* Mon Aug 19 2013 Siddhesh Poyarekar - 2.18-2 +- Fix buffer overflow in readdir_r (#995841, CVE-2013-4237). +- Remove releng tarball. + +* Fri Aug 16 2013 Siddhesh Poyarekar - 2.18-1 +- Upstream release 2.18. +- Pull in systemd during build and use the tmpfilesdir macro. + +* Wed Aug 14 2013 Carlos O'Donell - 2.17.90-14 +- Update spec file to use rpm prefix everywhere. + +* Tue Aug 13 2013 Carlos O'Donell - 2.17.90-13 +- Revert `Move to /usr' transition. + +* Tue Aug 13 2013 Carlos O'Donell - 2.17.90-12 +- Complete `Move to /usr' transition. All relevant files are now + installed into `/usr'. + +* Wed Aug 07 2013 Karsten Hopp 2.17.90-11 +- rebuild with the latest rpm to fix missing ld64.so provides on PPC + +* Mon Jul 29 2013 Carlos O'Donell - 2.17.90-10 +- Fix missing libbsd.a in debuginfo packages. + +* Mon Jul 29 2013 Siddhesh Poyarekar - 2.17.90-9 +- Fix strcoll flaws (#855399, CVE-2012-4412, CVE-2012-4424). + +* Mon Jul 29 2013 Siddhesh Poyarekar - 2.17.90-8 +- Resync with upstream master. +- Disable pt_chown (CVE-2013-2207). + +* Thu Jul 25 2013 Carlos O'Donell - 2.17.90-7 +- Correctly name the 240-bit slow path sytemtap probe slowpow_p10 for slowpow. + +* Wed Jul 24 2013 Carlos O'Donell - 2.17.90-6 +- Add build requirement on static libstdc++ library to fix testsuite failures + for static C++ tests. + +* Fri Jul 12 2013 Siddhesh Poyarekar - 2.17.90-5 +- Enable lock elision support (#982363). +- Depend on systemd instead of systemd-units (#983760). + +* Tue Jul 9 2013 Siddhesh Poyarekar - 2.17.90-4 +- Resync with upstream master. + +* Thu Jun 20 2013 Siddhesh Poyarekar - 2.17.90-3 +- Resync with upstream master. + +* Tue Jun 11 2013 Remi Collet - 2.17.90-2 +- rebuild for new GD 2.1.0 + +* Tue Jun 4 2013 Siddhesh Poyarekar - 2.17.90-1 +- Resync with upstream master. + +* Tue May 14 2013 Siddhesh Poyarekar - 2.17-9 +- Avoid crashing in LD_DEBUG when program name is unavailable (#961238). + +* Sun May 5 2013 Patsy Franklin - 2.17-8 +- Fix _nl_find_msg malloc failure case, and callers. (#959034). + +* Tue Apr 23 2013 Patsy Franklin - 2.17-7 +- Test init_fct for NULL, not result->__init_fct, after demangling (#952799). + +* Tue Apr 23 2013 Patsy Franklin - 2.17-6 +- Increase limits on xdr name and record requests (#892777). +- Consistently MANGLE/DEMANGLE init_fct, end_fct and btow_fct (#952799). + +* Thu Mar 28 2013 Siddhesh Poyarekar - 2.17-5 +- Don't add input group during initgroups_dyn in hesiod (#921760). + +* Sun Mar 17 2013 Carlos O'Donell - 2.17-4 +- Fixed i386 glibc builds (#917161). +- Fixed multibyte character processing crash in regexp (#905877, CVE-2013-0242) + +* Wed Feb 27 2013 Carlos O'Donell - 2.17-3 +- Renamed release engineering directory to `releng' (#903754). +- Fix building with gcc 4.8.0 (#911307). + +* Thu Feb 7 2013 Carlos O'Donell - 2.17-2 +- Fix ownership of /usr/lib[64]/audit (#894307). +- Support unmarked ARM objects in ld.so.cache and aux cache (#905184). + +* Tue Jan 1 2013 Jeff Law - 2.17-1 +- Resync with official glibc-2.17 release +* Fri Dec 21 2012 Jeff Law - 2.16.90-40 +- Resync with master + +* Wed Dec 19 2012 Jeff Law - 2.16.90-39 +- Add rtld-debugger-interface.txt as documentation. (#872242) + +* Fri Dec 7 2012 Jeff Law - 2.16.90-38 +- Resync with master +- Drop patch for 731228 that is no longer needed. + +* Thu Dec 6 2012 Jeff Law - 2.16.90-37 +- Resync with master +- Patch for 697421 has been submitted upstream. +- Drop local patch for 691912 that is no longer needed. + +* Mon Dec 3 2012 Jeff Law - 2.16.90-36 +- Resync with master +- Drop local patch for 657588 that is no longer needed. +- Drop local patch for 740682 that is no longer needed. +- Drop local patch for 770439 that is no longer needed. +- Drop local patch for 789209 that is no longer needed. +- Drop local patch for nss-files-overflow that seems + useless. +- Drop localedata-locales-fixes as they were rejected + upstream. +- Drop test-debug-gnuc-hack.patch that seems useless now. +- Repack patchlist. + +* Fri Nov 30 2012 Jeff Law - 2.16.90-35 +- Resync with master (#882137). +- Remove local patch for strict-aliasing warnings that + is no longer needed. +- Remove local patch for 730856 that is no longer needed. +- Repack patchlist. + +* Thu Nov 29 2012 Jeff Law - 2.16.90-34 +- Remove local patch which "temporarily" re-added currences + obsoleted by the Euro. +- Remove hunks from strict-aliasing patch that are no longer + needed. + +* Thu Nov 29 2012 Jeff Law - 2.16.90-33 +- Resync with master. +- Drop local patch for 788989. +- Repack patchlist. + +* Wed Nov 28 2012 Jeff Law - 2.16.90-32 +- Resync with master. +- Drop local patch for 878913. +- Drop local patch for 880666. +- Drop local patch for 767693. +- Repack patchlist. + +* Tue Nov 27 2012 Siddhesh Poyarekar - 2.16.90-31 +- Ensure that hashtable size is greater than 3 (#878913). +- fwrite returns 0 on EOF (#880666). + +* Mon Nov 26 2012 Jeff Law - 2.16.90-30 +- Resync with upstream sources +- Drop local patch for getconf. +- Repack patchlist. + +* Fri Nov 16 2012 Jeff Law - 2.16.90-29 +- Rsync with upstream sources +- Drop local patches for 803286, 791161, 790292, 790298 + +* Wed Nov 7 2012 Jeff Law - 2.16.90-28 +- Resync with upstream sources (#873397) + +* Mon Nov 5 2012 Jeff Law - 2.16.90-27 +- Resync with upstream sources. +- Don't use distinct patches for 770869, 787201 and 688948 + as they all modify stuff under fedora/ +- Repack patchlist + +* Thu Nov 1 2012 Jeff Law - 2.16.90-26 +- Resync with upstream sources (#872336) + +* Mon Oct 22 2012 Jeff Law - 2.16.90-25 +- Rsync with upstream sources +- Drop 864820 patch as now that it's upstream. +- Add sss to /etc/nsswitch.conf (#867473) + +* Thu Oct 11 2012 Jeff Law - 2.16.90-24 +- Rsync with upstream sources +- Drop local 552960-2 patch now that it's upstream. +- Drop local 858274 patch now that the root problem is fixed upstream. +- Repack patchlist. + +* Wed Oct 10 2012 Siddhesh Poyarekar - 2.16.90-23 +- Fix Marathi names for Wednesday, September and October (#rh864820). + +* Fri Oct 5 2012 Jeff Law - 2.16.90-22 +- Resync with upstream sources +- Drop local 552960 patch now that it's upstream +- Drop local stap patch now obsolete +- Drop local s390 patch which avoided problems with old assemblers +- Drop old fortify source patch to deal with old compilers + +* Thu Oct 4 2012 Siddhesh Poyarekar - 2.16.90-21 +- Take mutex in cleanup only if it is not already taken. + +* Tue Oct 2 2012 Jeff Law - 2.16.90-20 +- Resync with upstream sources. +- Repack patchlist. + +* Mon Oct 1 2012 Jeff Law - 2.16.90-19 +- Resync with upstream sources to pick up fma fixes + +* Fri Sep 28 2012 Jeff Law - 2.16.90-18 +- Resync with upstream sources. +- Drop fedora-cdefs-gnuc.patch, it's not needed anymore. +- Drop fedora-gai-rfc1918.patch, it's upstream now. +- Drop fedora-localedata-no_NO.patch, it was supposed to be + temporary -- that was back in 2003. This should have been + sorted out long ago. We'll just have to deal with the + fallout. +- Drop fedora-vfprintf-sw6530.patch, it's upstream now. +- Drop rh769421.patch; Siddhesh has fixed this properly with 552960. + +* Fri Sep 28 2012 Siddhesh Poyarekar - 2.16.90-17 +- Release mutex before going back to wait for PI mutexes (#552960). + +* Tue Sep 25 2012 Jeff Law - 2.16.90-16 +- Resync with upstream sources. + +* Fri Sep 21 2012 Jeff Law - 2.16.90-15 +- Remove most of fedora-nscd patch as we no longer use the + old init files, but systemd instead. +- Remove path-to-vi patch. With the usr-move changes that + patch is totally unnecessary. +- Remove i686-nopl patch. Gas was changed back in 2011 to + avoid nopl. +- Move gai-rfc1918 patch to submitted upstream status + +* Fri Sep 21 2012 Jeff Law - 2.16.90-14 +- Revert patch for 816647, it's blatently broken. + +* Fri Sep 21 2012 Siddhesh Poyarekar - 2.16.90-13 +- Bring back byteswap-16.h (#859268). + +* Thu Sep 20 2012 Jeff Law - 2.16.90-12 +- Revert recent upstream strstr changes (#858274) +- Demangle function pointers before testing them (#816647) +- Remove handling of /etc/localtime and /var/spool/postfix/etc/localtime + as systemd will be handling them from now on (#858735). + +* Fri Sep 14 2012 Jeff Law - 2.16.90-11 +- Resync with upstream sources (#857236). + +* Sat Sep 8 2012 Peter Robinson - 2.16.90-10 +- Enable ports to fix FTBFS on ARM + +* Wed Sep 5 2012 Jeff Law - 2.16.90-9 +- Resync with upstream sources. + +* Tue Sep 4 2012 Jeff Law - 2.16.90-8 +- Incorporate ppc64p7 arch changes (#854250) + +* Thu Aug 30 2012 Jeff Law - 2.16.90-7 +- Resync with upstream sources. + +* Wed Aug 22 2012 Jeff Law - 2.16.90-6 +- Resync with upstream sources. + +* Tue Aug 21 2012 Jeff Law - 2.16.90-5 +- Replace manual systemd scriptlets with macroized scriptlets (#850129) + +* Mon Aug 20 2012 Jeff Law - 2.16.90-4 +- Move /etc/localtime into glibc-common package since glibc-common + owns the scriptlets which update it. + +* Mon Aug 20 2012 Jeff Law - 2.16.90-3 +- Remove obsolete patches from glibc-fedora.patch. Explode + remaining patches into distinct patchfiles. Thanks to + Dmitry V. Levin for identifying them! + Drop ia64 specific patches and specfile fragments + +* Wed Aug 15 2012 Jeff Law - 2.16.90-2 +- Fix integer overflow leading to buffer overflow in strto* (#847718) + +* Mon Aug 13 2012 Jeff Law - 2.16.90-1 +- Resync with upstream sources, drop obsolete patches. +- Drop glibc-ports bits as they're part of the master + sources now. + +* Mon Aug 13 2012 Jeff Law - 2.16-9 +- Replace patch for 179072 with official version from upstream. + +* Fri Aug 10 2012 Jeff Law - 2.16-8 +- Replace patch for 789238 with official version from upstream. + +* Wed Jul 25 2012 Jeff Law - 2.16-7 +- Pack IPv4 servers at the start of nsaddr_list and + only track the number of IPV4 servers in EXT(statp->nscounti (#808147) +- Mark set*uid, set*gid as __wur (warn unused result) (#845960) + +* Wed Jul 25 2012 Jeff Law - 2.16-6 +- Revert patch for BZ696143, it made it impossible to use IPV6 + addresses explicitly in getaddrinfo, which in turn broke + ssh, apache and other code. (#808147) +- Avoid another unbound alloca in vfprintf (#841318) +- Remove /etc/localtime.tzupdate in lua scriptlets +- Revert back to using posix.symlink as posix.link with a 3rd + argument isn't supported in the lua version embedded in rpm. +- Revert recent changes to res_send (804630, 835090). +- Fix memcpy args in res_send (#841787). + +* Thu Jul 19 2012 Fedora Release Engineering - 2.16-3 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_18_Mass_Rebuild + +* Thu Jul 5 2012 Jeff Law - 2.16-2 +- Use posix.link rather than posix.symlink in scriptlet to + update /etc/localtime (#837344). + +* Mon Jul 2 2012 Jeff Law - 2.16-1 +- Resync with upstream glibc-2.16 release. + +* Fri Jun 22 2012 Jeff Law - 2.15.90-16 +- Resync with upstream sources, drop obsolete patch. + +* Thu Jun 21 2012 Jeff Law - 2.15.90-15 +- Resync with upstream sources (#834447). +- Fix use-after-free in dcigettext.c (#816647). + +* Fri Jun 15 2012 Jeff Law - 2.15.90-14 +- Resync with master. + +* Thu Jun 14 2012 Jeff Law - 2.15.90-13 +- Delay setting DECIDED field in locale file structure until + we have read the file's data (#827510). + +* Mon Jun 11 2012 Dennis Gilmore - 2.15.90-12 +- actually apply the arm linker hack + +* Mon Jun 11 2012 Dennis Gilmore - 2.15.90-11 +- only deal with the arm linker compat hack on armhfp arches +- armsfp arches do not have a linker change +- Backward compat hack for armhf binaries. + +* Thu Jun 7 2012 Jeff Law - 2.15.90-10 +- Fix parsing of /etc/sysconfig/clock when ZONE has spaces. (#828291) + +* Tue Jun 5 2012 Jeff Law - 2.15.90-9 +- Resync with upstream sources, drop unnecessary patches. +- Fix DoS in RPC implementation (#767693) +- Remove deprecated alpha support. +- Remove redundant hunk from patch. (#823905) + +* Fri Jun 1 2012 Patsy Franklin - 2.15.90-8 +- Fix iconv() segfault when the invalid multibyte character 0xffff is input + when converting from IBM930 (#823905) + +* Fri Jun 1 2012 Jeff Law - 2.15.90-7 +- Resync with upstream sources. (#827040) + +* Thu May 31 2012 Patsy Franklin - 2.15.90-6 +- Fix fnmatch() when '*' wildcard is applied on a file name containing + multibyte chars. (#819430) + +* Wed May 30 2012 Jeff Law - 2.15.90-5 +- Resync with upstream sources, drop unnecessary patches. + +* Tue May 29 2012 Jeff Law - 2.15.90-4 +- Build info files in the source dir, then move to objdir + to avoid multilib conflicts (#825061) + +* Fri May 25 2012 Jeff Law - 2.15.90-3 +- Work around RPM dropping the contents of /etc/localtime + when it turns into a symlink with %post common script (#825159). + +* Wed May 23 2012 Jeff Law - 2.15.90-2 +- Fix option rotate when one IPV6 server is enabled (#804630) +- Reenable slow/uberslow path taps slowpow/slowexp. + +* Wed May 23 2012 Jeff Law - 2.15.90-1 +- Resync with upstream sources, drop unnecessary patches. + +* Tue May 22 2012 Patsy Franklin - 2.15-41 +- Fix tzdata trigger (#822200) +- Make the symlink relative rather than linking into the buildroot (#822200). +- Changed /etc/localtime to a symlink. 8222000 (#822200) + +* Tue May 15 2012 Jeff Law - 2.15-40 +- Update to upstream patch for 806070 (#806070) + +* Mon May 14 2012 Jeff Law - 2.15-39 +- Update upstream patch for AVX testing (#801650) + +* Fri May 11 2012 Jeff Law - 2.15-38 +- Upstream patch to fix AVX testing (#801650) + +* Thu May 10 2012 Jeff Law - 2.15-37 +- Try again to fix AVX testing (#801650) + +* Mon May 7 2012 Jeff Law - 2.15-36 +- Improve fortification disabled warning. +- Change location of dynamic linker for armhf. + +* Mon Apr 30 2012 Jeff Law - 2.15-35 +- Implement context routines for ARM (#817276) + +* Fri Apr 13 2012 Jeff Law - 2.15-34 +- Issue a warning if FORTIFY_CHECKING is requested, but disabled. + +* Thu Apr 12 2012 Jeff Law - 2.15-33 +- Fix another unbound alloca in nscd groups (#788989) + +* Tue Apr 3 2012 Jeff Law - 2.15-32 +- Fix first day of week for lv_LV (#682500) + +* Mon Apr 2 2012 Jeff Law - 2.15-31 +- When retrying after main arena failure, always retry in a + different arena. (#789238) + +* Tue Mar 27 2012 Jeff Law - 2.15-30 +- Avoid unbound alloca usage in *-crypt routines (#804792) +- Fix data race in nscd (#806070) + +* Fri Mar 23 2012 Jeff Law - 2.15-29 +- Fix typo in __nss_getent (#806403). + +* Wed Mar 14 2012 Jeff Law - 2.15-28 +- Add doi_IN, sat_IN and mni_IN to SUPPORTED locals (#803286) +- Add stap probes in slowpow and slowexp. + +* Fri Mar 09 2012 Jeff Law - 2.15-27 +- Fix AVX checks (#801650) + +* Wed Feb 29 2012 Jeff Law - 2.15-26 +- Set errno properly in vfprintf (#794797) +- Don't kill application when LD_PROFILE is set. (#800224) + +* Wed Feb 29 2012 Jeff Law - 2.15-25 +- Fix out of bounds memory access in resolver (#798471) +- Always mark vDSO as used (#758888) + +* Fri Feb 24 2012 Jeff Law - 2.15-24 +- Fix bogus underflow (#760935) +- Correctly handle dns request where large numbers of A and AAA records + are returned (#795498) +- Fix nscd crash when group has many members (#788989) + +* Mon Feb 20 2012 Jeff Law - 2.15-23 +- Avoid "nargs" integer overflow which could be used to bypass FORTIFY_SOURCE (#794797) + +* Mon Feb 20 2012 Jeff Law - 2.15-22 +- Fix main arena locking in malloc/calloc retry path (#789238) + +* Fri Feb 17 2012 Jeff Law - 2.15-21 +- Correctly identify all 127.x.y.z addresses (#739743) +- Don't assign native result if result has no associated interface (#739743) + +* Fri Feb 17 2012 Jeff Law - 2.15-20 +- Ignore link-local IPV6 addresses for AI_ADDRCONFIG (#697149) + +* Fri Feb 17 2012 Jeff Law - 2.15-19 +- Fix reply buffer mismanagement in resolver (#730856) + +* Thu Feb 16 2012 Jeff Law - 2.15-18 +- Revert 552960/769421 changes again, still causing problems. +- Add doi_IN (#791161) +- Add sat_IN (#790292) +- Add mni_IN (#790298) + +* Thu Feb 9 2012 Jeff Law - 2.15-17 +- Fix lost wakeups in pthread_cond_*. (#552960, #769421) +- Clarify info page for snprintf (#564528) +- Fix first_weekday and first_workday for ru_UA (#624296) + +* Tue Feb 7 2012 Jeff Law - 2.15-16 +- Fix currency_symbol for uk_UA (#789209) +- Fix weekday names in Kashmiri locale (#770439) + +* Tue Feb 7 2012 Jeff Law - 2.15-15 +- Remove change for 787662, correct fix is in gcc. + +* Mon Feb 6 2012 Jeff Law - 2.15-13 +- More accurately detect if we're in a chroot (#688948) + +* Fri Feb 3 2012 Jeff Law - 2.15-12 +- Add fedfs to /etc/rpc (#691912) +- Run nscd in the foreground w/ syslogging, fix systemd config (#770869) +- Avoid mapping past end of shared object (#741105) +- Turn off -mno-minimal-toc on PPC (#787201) +- Remove hunk from glibc-rh657588.patch that didn't belong + +* Wed Feb 1 2012 Jeff Law - 2.15-8 +- Prevent erroneous inline optimization of initfini.s on PowerPC64 (#783979) +- Use upstream variant of fix for 740506. +- Fix month abbreviations for zh_CN (#657588) + +* Sun Jan 29 2012 Jeff Law - 2.15-7 +- Sort objects before relocations (sw#13618) +- Fix bogus sort code that was copied from dl-deps.c. + +* Thu Jan 26 2012 Jeff Law - 2.15-6 +- First argument to settimeofday can be null (#740682) +- Add aliases for ISO-10646-UCS-2 (#697421) + +* Tue Jan 24 2012 Jeff Law - 2.15-4 +- Update ports from master. +- Fix first workday/weekday for it_IT (#622499) +- Fix type to uint16_t based on upstream comments (729661) +- Do not cache negative results in nscd if these are transient (#784402) + +* Mon Jan 23 2012 Jeff Law - 2.15-3 +- Fix cycle detection (#729661) +- Fix first workday/weekday for it_IT (#446078) +- Fix first workday/weekday for ca_ES (#454629) + +* Fri Jan 13 2012 Fedora Release Engineering - 2.15-2 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_17_Mass_Rebuild + +* Sun Jan 1 2012 Jeff Law - 2.15-1.fc17 +- Update from master (a316c1f) + +* Thu Dec 22 2011 Jeff Law - 2.14.90-26.fc17 +- Update from master (16c6f99) +- Fix typo in recent tzfile change (#769476) +- Make MALLOC_ARENA_MAX and MALLOC_ARENA_TEST match documentation (#740506) +- Revert "fix" to pthread_cond_wait (#769421) +- Extract patch for 730856 from fedora-patch into a distinct patchfile + +* Mon Dec 19 2011 Jeff Law - 2.14.90-25.fc17 +- Update from master (a4647e7). + +* Sun Dec 18 2011 Jeff Law - 2.14.90-24.fc16.3 +- Check values from TZ file header (#767696) +- Handle EAGAIN from FUTEX_WAIT_REQUEUE_PI (#552960) +- Add {dist}.# +- Correct return value from pthread_create when stack alloction fails. + (#767746) + +* Wed Dec 7 2011 Jeff Law - 2.14.90-23 +- Fix a wrong constant in powerpc hypot implementation (#750811) + #13534 in python bug database + #13472 in glibc bug database +- Truncate time values in Linux futimes when falling back to utime + +* Mon Dec 5 2011 Jeff Law - 2.14.90-22 +- Mark fortified __FD_ELT as extension (#761021) +- Fix typo in manual (#708455) + +* Wed Nov 30 2011 Jeff Law - 2.14.90-21 +- Don't fail in makedb if SELinux is disabled (#750858) +- Fix access after end of search string in regex matcher (#757887) + +* Mon Nov 28 2011 Jeff Law - 2.14.90-20 +- Drop lock before calling malloc_printerr (#757881) + +* Fri Nov 18 2011 Jeff Law - 2.14.90-19 +- Check malloc arena atomically (BZ#13071) +- Don't call reused_arena when _int_new_arena failed (#753601) + +* Wed Nov 16 2011 Jeff Law - 2.14.90-18 +- Fix grouping and reuse other locales in various locales (BZ#13147) + +* Tue Nov 15 2011 Jeff Law - 2.14.90-17 +- Revert bogus commits/rebasing of Nov 14, Nov 11 and Nov 8. Sources + should be equivalent to Fedora 16's initial release. + +* Wed Oct 26 2011 Fedora Release Engineering - 2.14.90-15 +- Rebuilt for glibc bug#747377 + +* Wed Oct 19 2011 Jim Meyering - 2.14.90-14 +- Revert the upstream patch that added the leaf attribute, since it + caused gcc -O2 to move code past thread primitives and sometimes + even out of critical sections. See http://bugzilla.redhat.com/747377 + +* Wed Oct 19 2011 Andreas Schwab - 2.14.90-13 +- Update from master + - Fix linkage conflict with feraiseexcept (#746753) + - More libm optimisations + +* Mon Oct 17 2011 Andreas Schwab - 2.14.90-12 +- Update from master + - Correctly handle missing initgroups database (#745675) + - Optimize many libm functions + - Optimize access to isXYZ and toXYZ tables + - Optimized memcmp and wmemcmp for x86-64 and x86-32 + - Add parameter annotation to modf (BZ#13268) + - Support optimized isXXX functions in C++ code + - Check for zero size in memrchr for x86_64 (#745739) + - Optimized memchr, memrchr, rawmemchr for x86-32 + +* Tue Oct 11 2011 Andreas Schwab - 2.14.90-11 +- Update from master + - Clean up locarchive mmap reservation code + - Fix netname2host (BZ#13179) + - Fix remainder (NaN, 0) (BZ#6779, BZ#6783) + - S/390: Fix longlong.h inline asms for zarch + - Improve 64 bit memchr, memrchr, rawmemchr with SSE2 + - Update translations + - Implement caching of netgroups in nscd + - Handle OOM in NSS + - Don't call ifunc functions in trace mode +- Convert tzdata-update to lua (#729796) +- Horrible workaround for horribly broken software (#737223) + +* Wed Sep 28 2011 Andreas Schwab - 2.14.90-10 +- Update from master + - Correctly reparse group line after enlarging the buffer (#739360) + - Fix parse error in bits/mathinline.h with --std=c99 (#740235) +- Update nscd service file (#740284) +- Drop nscd init file (#740196) + +* Fri Sep 16 2011 Andreas Schwab - 2.14.90-9 +- Update from master + - Define IP_MULTICAST_ALL (BZ#13192) + - Add fmax and fmin inlines for x86-64 + - Avoid race between {,__de}allocate_stack and __reclaim_stacks + during fork (#737387) + - Optimized lrint and llrint for x86-64 + - Also relocate in dependency order when doing symbol dependency + testing (#737459) + - Optimize logb code for 64-bit machines + - Fix jn precision (BZ#11589) + - Fix boundary conditions in scanf (BZ#13138) + - Don't lock string streams in stream cleanup code (BZ#12847) + - Define ELFOSABI_GNU + - Fix lround loss of precision + - Add range checking for FD_SET, FD_CLR, and FD_ISSET +- Make sure AVC thread has capabilities + +* Thu Sep 8 2011 Andreas Schwab - 2.14.90-8 +- Update from master + - Use O_CLOEXEC when loading objects and cache in ld.so (BZ#13068) + - Fix memory leak in case of failed dlopen (BZ#13123) + - Optimizations for POWER + - Prefer real syscalls instead of vsyscalls on x86-64 outside libc.so + - Add Atom-optimized strchr and strrchr for x86-64 + - Try shell in posix_spawn* only in compat mode (BZ#13134) + - Fix glob.h header by removing gcc 1.x support (BZ#13150) + - Optimized strchr and strrchr with SSE2 on x86-32 + - Add optimized x86 wcscmp + - Fixes and optimizations for 32-bit sparc fabs + - Fix nptl semaphore cleanup invocation + - Sanitize HWCAP_SPARC_* defines/usage, and add new entries + +* Thu Sep 1 2011 Andreas Schwab - 2.14.90-7 +- Update from master + - Relocate objects in dependency order (#733462) +- Avoid assertion failure when first DNS answer was empty (#730856) +- Don't treat tls_offset == 1 as forced dynamic (#731228) + +* Wed Aug 24 2011 Andreas Schwab - 2.14.90-6 +- Update from master + - Correct cycle detection during dependency sorting + - Use ifuncs for time and gettimeofday on x86-64 + - Fix fopen (non-existing-file, "re") errno + - Fix CFI info in x86-64 trampolines for non-AVX code + - Build libresolv with SSP flags + - Avoid executable stack in makedb (#731063) + - Align x86 TCB to 64 bytes (cache line size), important for Atom + +* Mon Aug 15 2011 Andreas Schwab - 2.14.90-5 +- Update from master + - Implement LD_DEBUG=scopes + - Locale-independent parsing in libintl (#726536) + - Fix stack alignment on x86_64 (#728762) + - Implement scandirat function + +* Tue Aug 9 2011 Andreas Schwab - 2.14.90-4 +- Update from master + - Properly tokenize nameserver line for servers with IPv6 address + - Fix encoding name for IDN in getaddrinfo (#725755) + - Fix inline strncat/strncmp on x86 + - Define SEEK_DATA and SEEK_HOLE + - Define AF_NFC and PF_NFC + - Update ptrace constants + - Add read barriers in cancellation initialization + - Add read barrier protecting DES initialization + - Fix overflow bug in optimized strncat for x86-64 + - Check for overflows in expressions (BZ#12852) + - Fix check for AVX enablement (#720176, BZ#13007) + - Force La_x86_64_ymm to be 16-byte aligned + - Add const attr to gnu_dev_{major,minor,makedev} +- Filter out GLIBC_PRIVATE symbols again + +* Wed Jul 20 2011 Andreas Schwab - 2.14.90-3 +- Update from master + - S/390: Don't use r11 in INTERNAL_VSYSCALL_NCS macro + - Avoid warning in nscd config file parsing code + - Improve 64 bit strcat functions with SSE2/SSSE3 + - Fix alloca accounting in strxfrm + - Avoid possible crashes in anormal nscd exits + - Updated Swedish and Dutch translations + +* Thu Jul 14 2011 Andreas Schwab - 2.14.90-2 +- Update from master + - Generalize framework to register monitoring of files in nscd + - Handle ext4 in {,f}pathconf + - Handle Lustre filesystem (BZ#12868) + - Handle W; without long options in getopt (BZ#12922) + - Change error code for underflows in strtod (BZ#9696) + - Fix handling of chained netgroups + - Optimize long-word additions in SHA implementation + - Handle nscd negtimeout==0 + - nss_compat: query NIS domain only when needed + - Fix robust mutex handling after fork + - Make sure RES_USE_INET6 is always restored +- Add systemd configuration for nscd +- Be more careful running build-locale-archive + +* Thu Jun 30 2011 Andreas Schwab - 2.14.90-1 +- Update from master + - Fix quoting in some installed shell scripts (BZ#12935) + - Fix missing .ctors/.dtors lead word in soinit + - Improved st{r,p}{,n}cpy for SSE2 and SSSE3 on x86 + - Avoid __check_pf calls in getaddrinfo unless really needed + (BZ#12907) + - Rate limit expensive _SC_NPROCESSORS_ONLN computation + - Add initgroups lookup support to getent + - Reenable nss_db with a completely new implementation + - Rewrite makedb to avoid using db library + - Add pldd program +- Obsolete nss_db +- Don't build tzdata-update and build-locale-archive statically + +* Tue Jun 28 2011 Andreas Schwab - 2.14-4 +- Update from 2.14 branch + - Fix crash in GB18030 encoder (#712901) +- Fix more bugs in GB18030 charmap +- Don't use gethostbyaddr to determine canonical name + +* Tue Jun 21 2011 Andreas Schwab - 2.14-3 +- Update from 2.14 branch + - Fix typo in recent resolver change which causes segvs (#710279) + - Fix memory leak in getaddrinfo (#712178) + - Fix for C++ (BZ#12841) + - Assume Intel Core i3/i5/i7 processor if AVX is available +- Filter results from gethostbyname4_r according to request flags + (#711827) +- Repair GB18030 charmap (#712901) +- Revert "Use .machine to prevent AS from complaining about z9-109 + instructions in iconv modules" (#711330) + +* Fri Jun 3 2011 Andreas Schwab - 2.14-2 +- Revert "Handle DNS server failures in case of AF_UNSPEC lookups + correctly" (#710279) + +* Tue May 31 2011 Andreas Schwab - 2.14-1 +- Update to 2.14 release + - Handle DNS server failures in case of AF_UNSPEC lookups correctly + (BZ#12684) + - Prevent loader from loading itself + - Restore _res correctly (BZ#12350) + - Interpret numeric values in shadow file as signed (BZ#11099) + - Recognize use-vc option in resolv.conf (BZ#11558) + - Mark malloc hook variables as deprecated + - Declare malloc hook variables as volatile (BZ#11781) + - Don't document si_code used for raise (BZ#11799) + - Fix unnecessary overallocation due to incomplete character (BZ#12811) + - Handle failure of _nl_explode_name in all cases + - Add support for time syscall in vDSO (BZ#12813) + - Add sendmmsg and setns syscalls + - Use getcpu definition from vDSO on x86-64 (BZ#12813) +- Don't free non-malloced memory and fix memory leak (#709267) + +* Fri May 27 2011 Andreas Schwab - 2.13.90-14 +- Update from master + - Fix conversion to ISO-2022-JP-2 with ISO-8859-7 designation + (BZ#12814) + - Undo accidental change in x86-64 user.h + - Update Japanese translation + - Define RLIMIT_RTTIME (BZ#12795) + - Update longlong.h from GCC + - Add a few more alloca size checks (BZ#12671) + - Fix flags parameter value passed to pltenter and pltexit + - Define CLOCK_REALTIME_ALARM and CLOCK_BOOTTIME_ALARM + - Always fill output buffer in XPG strerror function (BZ#12782) + - Nicer output for negative error numbers in strerror_r + - Fix CP1258 conversion (BZ#12777) + - Fix handling of LC_CTYPE in locale name handling (BZ#12788) + - Set stream errors in more cases (BZ#12792) + - Don't unconditionally use alloca in gaih_inet (BZ#11869) + - Update documentation in regex.h (BZ#11857) + - Prevent Altivec and VSX insns on PowerPC64 when no FPRs or VRs are + available + - Fix typo in x86-64 powl (BZ#12775) +- Avoid overriding CFLAGS (#703880) + +* Wed May 18 2011 Andreas Schwab - 2.13.90-13 +- Update from master + - Update GB18030 to 2005 version (BZ#11837) + - Update RE_SYNTAX*_AWK constants in regex.h + - Handle long variable names in putenv (BZ#11892) + - Fix test for error_one_per_line mode in error (BZ#12766) + - Cleanup x86-64 sys/user.h (BZ#11820) + - Several locale data updates (BZ#11987, BZ#9732, BZ#9730, BZ#4357, + BZ#12582) + - Avoid potential deadlock in mtrace (BZ#6420) + - Fix a few problems in fopen and freopen + - Provide more helpful error message in getopt (BZ#7101) + - Make stack canary value harder to read through read overflow (BZ#10149) + - Use mmap for allocation of buffers used for __abort_msg (BZ#11901) + - Fix handling of static TLS in dlopen'ed objects (BZ#12453) + - Fix initialization of optimization values for AIO (BZ#12083) + - Fix handling of conversion problem in CP932 module (BZ#12601) + - Fix potential problem with out-of-scope buffer (BZ#12626) + - Handle recursive calls in backtrace better (BZ#12432) + - Fix handling of incomplete character storage in state + - Fix file descriptor position after fclose (BZ#12724) +- Reinstall NIS RPC headers + +* Fri May 13 2011 Andreas Schwab - 2.13.90-12 +- Update from master + - Fix resizing table for unique symbols when adding symbol for copy + relocation (BZ#12511) + - Fix sched_setscheduler call in spawn implementation (BZ#12052) + - Report write error in addmnt even for cached streams (BZ#12625) + - Translate kernel error into what pthread_create should return + (BZ#386) + - More configurability for secondary group lookup (BZ#11257) + - Several locale data updates (BZ#11258, BZ#11487, BZ#11532, + BZ#11578, BZ#11653, BZ#11668, BZ#11945, BZ#11947, BZ#12158, + BZ#12200, BZ#12178, BZ#12178, BZ#12346, BZ#12449, BZ#12545, + BZ#12551, BZ#12611, BZ#12660, BZ#12681, BZ#12541, BZ#12711, + BZ#12738) + - Fix Linux getcwd for long paths (BZ#12713) + - static tls memory leak on TLS_DTV_AT_TP archs + - Actually undefine ARG_MAX from + - Backport BIND code to query name as TLD (BZ#12734) + - Allow $ORIGIN to reference trusted directoreis in SUID binaries + (BZ #12393) + - Add missing {__BEGIN,__END}_DECLS to sys/sysmacros.h + - Report if no record is found by initgroups in nss_files +- Never leave $ORIGIN unexpanded +- Revert "Ignore origin of privileged program" +- Reexport RPC interface + +* Thu May 5 2011 Andreas Schwab - 2.13.90-11 +- Update from master + - Don't use removed rpc headers +- Install rpc/netdb.h again + +* Wed May 4 2011 Andreas Schwab - 2.13.90-10 +- Update from master + - ldconfig: don't crash on empty path in config file (#699784) + - getaddrinfo(AF_INET6) does not return scope_id info provided by + NSS modules (BZ#12714) + - Fix pathconf(_PC_BUF_SIZE) (BZ#12723) + - Fix getnameinfo flags parameter type (BZ#12717) + - Add finer grained control for initgroups lookups to NSS + - Use all possible bytes from fopen mode string (BZ#12685, #698025) + - Define initgroups callback for nss_files + - elf.h: Define R_ARM_IRELATIVE reloc type + - Fix static linking with checking x86/x86-64 memcpy (BZ#12653) + - Fix POWER4/POWER7 optimized strncmp to not read past differing bytes + - Fix FPU context handling in getcontext on x86-64 (BZ#12420) + - Skip extra zeroes when searching auxv on s390 + - Obsolete RPC implementation in libc + - Fix memory leak in TLS of loaded objects (BZ#12650) + - Don't leave empty element in rpath when skipping an element + - Make ppc sync_file_range cancelable + - Maintain stack alignment in ____longjmp_chk on x86_64 + +* Thu Apr 7 2011 Andreas Schwab - 2.13.90-9 +- Update from master + - Fix typo in cache information table for x86-{32,64} + - Define CLOCK_BOOTTIME, O_PATH, AT_EMPTY_PATH + - Work around old buggy program which cannot cope with memcpy + semantics (BZ#12518) + - Fix visibility of declarations of wcpcpy and wcpncpy (BZ#12631) + - Add clock_adjtime, name_to_handle_at, open_by_handle_at, syncfs + syscalls + - Really implement fallocate{,64} and sync_file_range as + cancellation points +- Enable systemtap support (#690281) + +* Thu Mar 24 2011 Andreas Schwab - 2.13.90-8 +- Update from master + - Fix infinite loop (#690323) + +* Mon Mar 21 2011 Andreas Schwab - 2.13.90-7 +- Update from master + - Handle page boundaries in x86 SSE4.2 strncmp (BZ#12597) + - Implement x86 cpuid handling of leaf4 for cache information (BZ#12587) + - Check size of pattern in wide character representation in fnmatch + (BZ #12583) + - Remove __restrict quals from wmemcmp prototype + - Fix copy relocations handling of unique objects (BZ#12510) +- ldd: never run file directly +- Ignore rpath elements containing non-isolated use of $ORIGIN when + privileged +- Don't leave empty element in rpath when skipping the first element +- Revert "Don't crash when dependencies are missing" (#688990) + +* Mon Mar 7 2011 Andreas Schwab - 2.13.90-6 +- Update from master + - Fix loading first object along a path when tracing + - Enable SSE2 memset for AMD'supcoming Orochi processor + - Don't read past end of buffer in fmemopen +- Revert broken changes (#682307) + +* Wed Mar 2 2011 Andreas Schwab - 2.13.90-5 +- Update from master + - Fix memory leak in dlopen with RTLD_NOLOAD (BZ#12509) + - Don't crash when dependencies are missing (BZ#12454) + - Fix allocation when handling positional parameters in printf + (BZ#12445) + - Fix two printf handler issues +- Fix false assertion (BZ#12454, #673014) + +* Mon Feb 14 2011 Andreas Schwab - 2.13.90-4 +- Update from master + - Update sysdeps/unix/sysv/linux/sparc/bits/socket.h + - Synchronize generic bits/sched.h cpu_set_t with Linux implementation + - Schedule nscd cache pruning more accurately from re-added values + - Fix passing symbol value to pltexit callbacks when ld.so auditing + - Fix range error handling in sgetspent +- Revert "Fix ordering of DSO constructors and destructors" (#673014) +- Create debuginfo-common on biarch archs +- Reinstall assembler workaround. +- Replace setuid by file capabilities (#646469) + +* Tue Jan 25 2011 Andreas Schwab - 2.13.90-1 +- Update from master + - Fix ordering of DSO constructors and destructors (BZ#11724) +- Remove no longer needed assembler workaround + +* Tue Jan 18 2011 Andreas Schwab - 2.13-1 +- Update to 2.13 release + - Define AT_NO_AUTOMOUNT + - Define MADV_HUGEPAGE and MADV_NOHUGEPAGE + - Add definitions for new socket protocols + - Signal temporary host lookup errors in nscd as such to the + requester (BZ#6812) + - Change setgroups to affect all the threads in the process + (BZ#10563) + - FIx handling of unterminated [ expression in fnmatch (BZ#12378) + - Relax requirement on close in child created by posix_spawn + - Fix handling of missing syscall in Linux mkdirat (BZ#12397) + - Handle long lines in host lookups in the right place (BZ#10484) + - Fix assertion when handling DSTs during auditing + - Fix alignment in x86 destructor calls + - Fix grouping when rounding increases number of integer digits + (BZ#12394) + - Update Japanese translations + - Fix infloop on persistent failing calloc in regex (BZ#12348) + - Use prlimit64 for 32-bit [gs]etrlimit64 implementation (BZ#12201) + - Change XPG-compliant strerror_r function to return error code + (BZ#12204) + - Always allow overwriting printf modifiers etc. + - Make PowerPC64 default to nonexecutable stack + +* Tue Dec 14 2010 Andreas Schwab - 2.12.90-21 +- Revert bogus change + +* Mon Dec 13 2010 Andreas Schwab - 2.12.90-20 +- Update from master + - Declare wcpcpy and wcpncpy only under _GNU_SOURCE + - Fix use of restrict in wchar.h and string.h + - Fix race in qsort_r initialization (BZ#11655) + - Don't ignore zero TTL in DNS answers + - Allow aux_cache_file open()ing to fail silently even in the chroot + mode (BZ#11149) + - Fix multiple nss_compat initgroups() bugs (BZ#10085) + - Define MAP_HUGETLB and SWAP_FLAG_DISCARD +- Remove .UTF-8 suffix from locale names when it is the only supported + codeset (#657556) +- Don't ignore $ORIGIN in libraries + +* Fri Nov 12 2010 Andreas Schwab - 2.12.90-19 +- Update from master + - Fix memory leak in fnmatch + - Support Intel processor model 6 and model 0x2c + - Fix comparison in sqrtl for IBM long double + - Fix one exit path in x86-64 SSE4.2 str{,n}casecmp (BZ#12205, #651638) + - Fix warnings in __bswap_16 (BZ#12194) + - Use IFUNC on x86-64 memset + - Power7-optimized mempcpy + - Handle uneven cache size in 32bit SSE2 memset (BZ#12191) + - Verify in ttyname that the symlink is valid (BZ#12167) + - Update Danish translations + - Fix concurrency problem between dl_open and dl_iterate_phdr + - Fix x86-64 strchr propagation of search byte into all bytes of SSE + register (BZ#12159) + - Fix perturbing in malloc on free (BZ#12140) + - PPC/A2 optimized memcpy function + - Add C99 FP_FAST_FMA{,F,L} macros to +- Check that the running kernel is new enough (#649589) + +* Fri Oct 22 2010 Andreas Schwab - 2.12.90-18 +- Require suid bit on audit objects in privileged programs (CVE-2010-3856) + +* Tue Oct 19 2010 Andreas Schwab - 2.12.90-17 +- Update from master + - Fix some fma issues, implement fmal (BZ#3268, #43358) + - Expect PLT call to _Unwind_Find_FDE on s390*-linux +- Never expand $ORIGIN in privileged programs (#643306, CVE-2010-3847) + +* Thu Oct 14 2010 Andreas Schwab - 2.12.90-16 +- Update from master + - Implement accurate fma (BZ#3268, #43358) + - Fix alignment of AVX save area on x86-64 (BZ#12113) + - Fix regex memory leaks (BZ#12078) + - Improve output of psiginfo (BZ#12107, BZ#12108) + - Don't return NULL address in getifaddrs (BZ#12093) + - Fix strstr and memmem algorithm (BZ#12092, #641124) +- Don't discard result of decoding ACE if AI_CANONIDN (#636642) +- Remove /etc/gai.conf from glibc-common and mark it %%ghost in glibc +- Require exact glibc version in nscd + +* Mon Oct 4 2010 Andreas Schwab - 2.12.90-15 +- Update from master + - Handle large requests in debugging hooks for malloc (BZ#12005) + - Fix handling of remaining bytes in buffer for strncmp and + strncasecmp (BZ#12077) + - Handle cgroup and btrfs filesystems in statvfs + - S/390: Fix highgprs check in startup code (BZ#12067) + - Properly convert f_fsid in statvfs (BZ#11611) + +* Tue Sep 28 2010 Andreas Schwab - 2.12.90-14 +- Don't try to write to _rtld_global_ro after performing relro + protection (#638091) + +* Mon Sep 27 2010 Andreas Schwab - 2.12.90-13 +- Update from master + - Add two forgotten licence exceptions + - getdents64 fallback d_type support + - Move freeres function from ld.so to libc.so + - Undo feature selection for ftruncate (BZ#12037) + - Fix namespace pollution in pthread_cleanup_push + - Fix limit detection in x86-64 SSE2 strncasecmp (#632560) + - Add support for fanotify_mark on sparc32 and s390 + - Fix register conflict in s390 ____longjmp_chk (#629970) + - Don't try to free rpath strings allocated during startup (#629976) + - Actually make it possible to user the default name server +- Fix memory leak on init/fini dependency list (#632936) +- Fix handling of collating symbols in regexps (BZ#11561) +- Don't parse %%s format argument as multibyte string (BZ#6530) +- Fix overflow in nss files parser +- Fix spurious nop at start of __strspn_ia32 + +* Wed Sep 15 2010 Dennis Gilmore - 2.12.90-12 +- dont build sparcv9v and sparc64v anymore + +* Mon Sep 13 2010 Andreas Schwab - 2.12.90-11 +- Update from master + - Fix _FORITY_SOURCE version of longjmp for Linux/x86-64 (BZ#11968) +- Work around shortest-stem feature in make 3.82+ + +* Mon Sep 6 2010 Andreas Schwab - 2.12.90-10 +- Update from master + - Remove invalid iconv aliases (BZ#11979) + - Update x86-64 mpn routines from GMP 5.0.1 + - Fix array overflow in floating point parser (BZ#7066) + - Support fanotify_mark syscall on powerpc32 + - Unroll x86-64 strlen + - Unroll 32bit SSE strlen and handle slow bsf + - Missing server address again leads to localhost being used (BZ#10851) +- Revert last change +- Remove or don't install unpackaged files for auxarches + +* Sat Sep 04 2010 Dennis Gilmore - 2.12.90-9 +- disable unpackaged file check on auxarches + +* Mon Aug 23 2010 Andreas Schwab - 2.12.90-8 +- Update from master + - Fix static strspn on x86 (#624852) + - Various POWER7 optimized string functions + - Fix x86 pthread_cond_signal() FUTEX_WAKE_OP fallback + - Add optimized strncasecmp versions for x86-64 + - PowerPC64 ABI fixes + - Properly quote output of locale (BZ#11904) + - f_flags in statfs implementation + - Add support for fanotify_init and fanotify_mask syscalls + - Add support for prlimit and prlimit64 + - Fix IPTOS_CLASS definition (BZ#11903) + - Avoid too much stack use in fnmatch (BZ#11883) + - x86: Add support for frame pointer less mcount +- Disable asynchronous-unwind-tables during configure run + +* Mon Aug 2 2010 Andreas Schwab - 2.12.90-7 +- Update from master + - Add optimized x86-64 implementation of strnlen and strcaecmp + - Document M_PERTURB + - Fix vDSO synthetic hwcap handling so they are not masked out from + ld.so.cache matching + - POWER6/7 optimizations for copysign +- Build with ports addon on alpha and armv5tel +- Add conflict with kernel < 2.6.32 (#619538) +- Switch to xz compressed tar files +- build-locale-archive: process only directories matching *_* + +* Wed Jul 21 2010 Andreas Schwab - 2.12.90-6 +- Bump minimum kernel version to 2.6.32 + +* Mon Jul 12 2010 Andreas Schwab - 2.12.90-5 +- Update from master + - Don't pass NULL occation to dl_signal_cerror + - Implement _PC_PIPE_BUF. +- Add glibc-ports tarball + +* Fri Jul 2 2010 Andreas Schwab - 2.12.90-4 +- Update from master + - Work around kernel rejecting valid absolute timestamps + - Improve 64bit memcpy/memmove for Atom, Core 2 and Core i7 + - Fix error handling in Linux getlogin* +- Workaround assembler bug sneaking in nopl (#579838) +- Fix scope handling during dl_close +- Fix setxid race handling exiting threads + +* Tue Jun 15 2010 Andreas Schwab - 2.12.90-3 +- Update from master + - Power7 string compare optimizations + - Properly resize buffer in NIS initgroups + - Define F_SETPIPE_SZ and F_GETPIPE_SZ + - Fix more C++ incompatibility problems in headers +- Properly set __libc_multiple_libcs +- Don't assume AT_PAGESIZE is always available (#597578) +- Don't call uname or getrlimit in libpthread init function (#579086) +- Mark /etc/rpc as %%config (#587050) + +* Mon May 31 2010 Andreas Schwab - 2.12.90-2 +- Update from master + - Small fix to POWER7 32-bit memcpy + - Correct x86 CPU family and model check (BZ#11640, #596554) + - Fix iov size in SH register_dump + - Don't crash on unresolved weak symbol reference + - Implement recvmmsg also as socketcall + - sunrpc: Fix spurious fall-through + - Make compatible with C++ (#593762) +- Fix users and groups creation in nscd %%post script + +* Wed May 19 2010 Andreas Schwab - 2.12.90-1 +- Update from master + - POWER7 optimized memset + - Fix typo in es_CR locale + - Enable IDN support in getent + - Fix race in free sanity check + - Fix lookup of collation sequence value during regexp matching + - Fix name of tt_RU.UTF-8@iqtelif locale (#589138) + - Handle too-small buffers in Linux getlogin_r (BZ#11571, #589946) + +* Tue May 4 2010 Roland McGrath - 2.12-1 +- Update to 2.12 release. + - Fix ldconfig chroot handling. + - Don't deadlock in __dl_iterate_phdr while (un)loading objects. + - Fix handling of newline in addmntent. + - Fix AIO when thread creation failed. + +* Fri Apr 16 2010 Andreas Schwab - 2.11.90-20 +- Update from master + - Fix bugs in x86-32 strcmp-sse4.S and strcmp-ssse3.S + - Add x86-32 FMA support + - Don't crash in trace mode when dependencies are missing + - x86-64 SSE4 optimized memcmp + - Fix makecontext on s390/s390x + +* Tue Apr 13 2010 Andreas Schwab - 2.11.90-19 +- Avoid multiarch memcmp in tzdata-update (#581677) + +* Mon Apr 12 2010 Andreas Schwab - 2.11.90-18 +- Update from master + - Implement interfaces to set and get names of threads (BZ#11390) + - Locale data updates (BZ#10824, BZ#10936, BZ#11470, BZ#11471) + - Print reload count in nscd statistics (BZ#10915) + - Fix reading loginuid file in getlogin{,_r} + - Fix fallocate error return on i386 + - Fix cproj implmentation (BZ#10401) + - Fix getopt handing (BZ#11039, BZ#11040, BZ#11041) + - Implement new mode for NIS passwd.adjunct.byname table (BZ#11134) + - Obey LD_HWCAP_MASK in ld.so.cache lookups + +* Tue Apr 6 2010 Andreas Schwab - 2.11.90-17 +- Update from master + - Locale data updates (BZ#11007, BZ#11258, BZ#11272, BZ#10554) + - Handle DNS timeouts in old-style lookup code (BZ#11010) + - Fix aux cache handling in ldconfig with chroot (BZ#11149) + - Fix printing error messages in getopt (BZ#11043) + - Declare iruserok and iruserok_af (BZ#11070) + - Fix option aliasing in argp (BZ#11254) + - Handle POSIX-compliant errno value of unlink in remove (BZ#11276) + - Fix definition and testing of S_ISSOCK (BZ#11279) + - Fix retrieving of kernel header version (BZ#11287) + - Fix concurrent handling of __cpu_features (BZ#11292) + - Handle unnecessary padding in getdents64 (BZ#11333) + - Fix changes to interface list during getifaddrs calls (BZ#11387) + - Missing memory barrier in DES initialization (BZ#11449) + - Fix spurious UNAVAIL status is getaddrinfo + - Add support for new clocks (BZ#11389) + - Fix Linux getlogin{_r,} implementation + - Fix missing zero-termination in cuserid (BZ#11397) + - Fix glob with empty pattern + - Fix handling of STB_GNU_UNIQUE in LD_TRACE_PRELINKING + - Unify wint_t handling in wchar.h and wctype.h (BZ#11410) + - Implement handling of libc ABI in ELF header + - Don't underestimate length of DST substitution in rpath + - Power7-optimized 64-bit and 32-bit memcpy +- Assign global scope to RFC 1918 addresses (#577626) + +* Thu Mar 18 2010 Andreas Schwab - 2.11.90-16 +- Fix SSSE3 memcmp (#574210) + +* Tue Mar 9 2010 Andreas Schwab - 2.11.90-15 +- Update from master + - sparc64: Fix handling of R_SPARC_TLS_LE_* relocations (#571551) + - Handle ext4 and logfs in statvfs functions + - Fix setxid race with thread creation + - Pass -mtune=i686 to assembler when compiling for i686 + - Fix R_X86_64_PC32 overflow detection + - Fix msgrcv on sparc64 + - Fix unwind info in x86 strcmp-sse4.S (BZ#11332) + - sparc: Add multiarch support for memset/bzero/memcpy +- Remove directories owned by filesystem (#569414) +- Add %%ghost /etc/gai.conf to glibc-common (#567748) + +* Tue Feb 23 2010 Andreas Schwab - 2.11.90-14 +- Update from master + - Sparc updates +- Fix SSSE3 memcpy (#556584) + +* Mon Feb 22 2010 Andreas Schwab - 2.11.90-13 +- Update from master + - Use CPUID_OFFSET instead of FEATURE_OFFSET + - Add 32bit memcmp/strcmp/strncmp optimized for SSSE3/SSS4.2 + - Fix file descriotor leak in nftw with FTW_CHDIR (BZ#11271) + - Add Sparc STT_GNU_IFUNC support + - Add power7-optimized classification functions +- Reapply "Optimize 32bit memset/memcpy with SSE2/SSSE3." +- Use unsigned comparison in sse memcpy/memset + +* Mon Feb 8 2010 Andreas Schwab - 2.11.90-12 +- Update from master + - Update constants in for current kernels (#11235) + - Fix endless loop with invalid /etc/shells file (#11242) + - Fix sorting of malayalam letter 'na' (#10414) + - Add kok_IN locale + - Use common collation data in as_IN locale + - Avoid alloca in setenv for long strings +- Use shared mapping to reserve memory when creating locale archive (#10855) +- Fix fstat on Linux/sparc64 (#11155) + +* Mon Feb 1 2010 Andreas Schwab - 2.11.90-11 +- Update from master + - Fix error checking in iconv (#558053) + - Don't map U00DF to U1E9E in toupper table + - _nl_load_locale() incorrectly handles mmap() failures (BZ#11200) + - Fix various issues in regex matcher (BZ#11183, BZ#11184, BZ#11185, + BZ#11186, BZ#11187, BZ#11188, BZ#11189, BZ#11190, BZ#11191, + BZ#11192, BZ#11193) + +* Tue Jan 19 2010 Andreas Schwab - 2.11.90-10 +- Update from master + - Fix ____longjmp_chk for s390/s390x + - Remove duplicate definitions of O_DSYNC and O_RSYNC for Linux/sparc + - Ignore negative dynamic entry types (#546890) + - Fix pthread_cond_*wait with requeue-PI on i386 (#548989) + - Fix _XOPEN_SOURCE_EXTENDED handling +- Revert "Optimize 32bit memset/memcpy with SSE2/SSSE3." + +* Fri Jan 15 2010 Andreas Schwab - 2.11.90-9 +- Update from master. + - Define IPTOS_CLASS_* macros according to RFC 2474 (BZ#11027) + - Always use IPv4 sockets for IPv4 addresses (BZ#11141) + - regcomp.c: do not ignore memory allocation failure (BZ#11127) + - Fix malloc_info without prior allocations (BZ#11126) + - Optimize 32bit memset/memcpy with SSE2/SSSE3 + - Relax feature tests in headers + +* Tue Jan 12 2010 Andreas Schwab - 2.11.90-8 +- Update from master. + - More POSIX conformance fixes. + +* Mon Jan 11 2010 Andreas Schwab - 2.11.90-7 +- Fix build failure. + +* Mon Jan 11 2010 Andreas Schwab - 2.11.90-6 +- Update from master. + - POSIX conformance fixes (BZ#11125). + +* Mon Jan 4 2010 Andreas Schwab - 2.11.90-5 +- Update from master. + - Additional setcontext(), etc. conformance tests (BZ#11115). + - Handle AT_FDCWD in futimens (BZ#10992). + - Update poll.h header for POSIX 2008 (BZ#11093). + - Avoid ELF lookup race. + +* Mon Dec 14 2009 Andreas Schwab - 2.11.90-4 +- Update from master. + - Add Requeue-PI support for x86 arch. + - Redefine O_SYNC and O_DSYNC to match 2.6.33+ kernels. + - Fix a few error cases in *name4_r lookup handling (BZ#11000). + - Fix kernel version check in recent ptsname change (BZ#11046). + - Add more warnings to exec functions (BZ#11056). + - Add recvmmsg interface. + - Define SCHED_IDLE and SCHED_RESET_ON_FORK for Linux. + +* Mon Nov 30 2009 Andreas Schwab - 2.11.90-3 +- Update from master. + - Fix infloop in __pthread_disable_asynccancel on x86_64 (#537690). + - Prevent unintended file desriptor leak in grantpt (#530558). + - Fix startup to security-relevant statically linked binaries (#528631). +- Re-install CFI in x86/x86_64 clone (#491542). + +* Tue Nov 24 2009 Andreas Schwab - 2.11.90-2 +- Update from master. + - Define week, first_weekday, and first_workday for en_DK locale (#525126). + - Use struct timespec for timestamps in struct stat also if + __USE_XOPEN2K8 (#539870). + - Fix week information for nl_NL locale (#499748). + - Update ntp_gettime for Linux (#479558). + - Fix getwc* and putwc* on non-wide streams (BZ#10958). + - Avoid warnings in CPU_* macros when using const bitsets (BZ#10918). + - Handle LC_GLOBAL_LOCALE in duplocale (BZ#10969). + - Fix _NC_LOCALE_NAME definition (BZ#10968). + - Add missing Linux MADV_* definitions (BZ#10972). + - Add support for new Linux error ERFKILL (BZ#10939). +- Enable multi-arch support on ppc and ppc64. + +* Thu Nov 12 2009 Andreas Schwab - 2.11.90-1 +- Update from master. + +* Thu Nov 5 2009 Andreas Schwab - 2.11-2 +- Fix readahead on powerpc32. +- Fix R_PPC64_{JMP_IREL,IRELATIVE} handling. +- Fix preadv, pwritev and fallocate for -D_FILE_OFFSET_BITS=64 (#533063). + +* Mon Nov 2 2009 Andreas Schwab - 2.11-1 +- Update to 2.11 release. +- Disable multi-arch support on PowerPC again since binutils is too old. +- Fix crash in tzdata-update due to use of multi-arch symbol (#532128). + +* Fri Oct 30 2009 Andreas Schwab - 2.10.90-27 +- Update from master. + - Fix races in setXid implementation (BZ#3270). + - Implement IFUNC for PPC and enable multi-arch support. + - Implement mkstemps/mkstemps64 and mkostemps/mkostemps64 (BZ#10349). + - Fix IA-64 and S390 sigevent definitions (BZ#10446). + - Fix memory leak in NIS grp database handling (BZ#10713). + - Print timestamp in nscd debug messages (BZ#10742). + - Fix mixing IPv4 and IPv6 name server in resolv.conf. + - Fix range checks in coshl. + - Implement SSE4.2 optimized strchr and strrchr. + - Handle IFUNC symbols in dlsym (#529965). + - Misc fixes (BZ#10312, BZ#10315, BZ#10319, BZ#10391, BZ#10425, + BZ#10540, BZ#10553, BZ#10564, BZ#10609, BZ#10692, BZ#10780, + BZ#10717, BZ#10784, BZ#10789, BZ#10847 +- No longer build with -fno-var-tracking-assignments. + +* Mon Oct 19 2009 Andreas Schwab - 2.10.90-26 +- Update from master. + - Add ____longjmp_chk for sparc. +- Avoid installing the same libraries twice. + +* Mon Oct 12 2009 Andreas Schwab - 2.10.90-25 +- Update from master + - Fix descriptor leak when calling dlopen with RTLD_NOLOAD (#527409). + - Fix week-1stday in C locale. + - Check for integer overflows in formatting functions. + - Fix locale program error handling (#525363). + +* Mon Sep 28 2009 Andreas Schwab - 2.10.90-24 +- Update from master. + - Fix missing reloc dependency (#517001). + +* Mon Sep 21 2009 Andreas Schwab - 2.10.90-23 +- Update from master. + +* Mon Sep 14 2009 Andreas Schwab - 2.10.90-22 +- Update from master. + - Fix endless loop in localedef. + - Fix __longjmp_chk on s390/s390x. +- Fix exit codes in nscd start script (#521848). +- Build with -fno-var-tracking-assignments for now (#523172). + +* Mon Sep 7 2009 Andreas Schwab - 2.10.90-21 +- Update from master. + - Fix strstr/strcasestr on i386 (#519226). + +* Thu Sep 3 2009 Andreas Schwab - 2.10.90-20 +- Update from master. + - Fix strstr/strcasestr/fma/fmaf on x86_64 (#519226). + - Fix lookup of group names in hesiod initgroups (#520472). + +* Wed Sep 2 2009 Andreas Schwab - 2.10.90-19 +- Update from master. + - Fix x86_64 bits/mathinline.h for -m32 compilation. + +* Tue Sep 1 2009 Andreas Schwab - 2.10.90-18 +- Update from master. + - fix parse error in (#520209). + +* Thu Aug 27 2009 Roland McGrath - 2.10.90-17 +- Update from master. + +* Wed Aug 26 2009 Andreas Schwab - 2.10.90-16 +- Update from master. + - handle AVX saving on x86-64 in interrupted symbol lookups (#519081). + +* Mon Aug 24 2009 Andreas Schwab - 2.10.90-15 +- Update from master. + - fix fortify failure with longjmp from alternate stack (#512103). +- Add conflict with prelink (#509655). + +* Mon Aug 17 2009 Andreas Schwab - 2.10.90-14 +- Update from master. + - fix pthread_cond_signal (#516469) + +* Mon Aug 10 2009 Andreas Schwab - 2.10.90-13 +- Update from master. + - fix rehashing of unique symbols (#515677) +- Fix spurious messages with --excludedocs (#515948) + +* Mon Aug 3 2009 Andreas Schwab - 2.10.90-12 +- Update from master. + - fix fortify failure with longjmp from alternate stack (#512103) + +* Thu Jul 30 2009 Andreas Schwab - 2.10.90-11 +- Update from master. +- Don't package debuginfo files in glibc-devel. + +* Tue Jul 28 2009 Andreas Schwab - 2.10.90-10 +- Update from master. + * fix memory ordering in pthread_mutex_unlock (BZ#10418) + * implement RES_USE_DNSSEC option in resolver (#205842) + * fix hang in ldd -r (#513945) + +* Mon Jul 27 2009 Andreas Schwab - 2.10.90-9 +- Update from master. + +* Fri Jul 24 2009 Fedora Release Engineering - 2.10.90-8.1 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_12_Mass_Rebuild + +* Fri Jul 24 2009 Jakub Jelinek - 2.10.90-7.1 +- Fix up pthread_cond_timedwait on x86_64 with old kernels. + +* Thu Jul 23 2009 Andreas Schwab - 2.10.90-7 +- Update from master. +- Build with -DNDEBUG unless using a prerelease. + +* Thu Jul 23 2009 Andreas Schwab - 2.10.90-6 +- Rebuilt with binutils-2.19.51.0.14-29.fc12 to fix static binaries + +* Wed Jul 22 2009 Andreas Schwab - 2.10.90-5 +- Update from master. +- Undefine __i686 on x86 to fix build. + +* Mon Jul 20 2009 Andreas Schwab - 2.10.90-4 +- Update from master. +- Don't build separate i686 package. + +* Wed Jul 8 2009 Andreas Schwab 2.10.90-3 +- Reenable setuid on pt_chown. + +* Thu Jul 2 2009 Andreas Schwab 2.10.90-2 +- Update from master. + +* Fri Jun 26 2009 Andreas Schwab 2.10.90-1 +- Update from master. +- Enable multi-arch support on x86/x86-64. +- Add requires glibc-headers to glibc-devel (#476295). +- Implement second fallback mode for DNS requests (#505105). +- Don't generate invalid POSIX TZ string for Asia/Dhaka timezone (#506941). +- Allow backtrace through __longjmp_chk on powerpc. + +* Fri May 22 2009 Jakub Jelinek 2.10.1-2 +- fix accept4 on architectures other than i?86/x86_64 +- robustify nscd client code during server GC +- fix up nscd segfaults during daemon shutdown +- fix memchr on ia64 (BZ#10162) +- replace the Sun RPC license with the BSD license, with the explicit + permission of Sun Microsystems +- fix up powerpc long double errno reporting + +* Sun May 10 2009 Jakub Jelinek 2.10.1-1 +- fix up getsgent_r and getsgnam_r exports on i?86 and ppc + +* Sat May 9 2009 Jakub Jelinek 2.10-2 +- update from trunk + - glibc 2.10 release + - fix memchr on x86_64 (#499689) + +* Mon Apr 27 2009 Jakub Jelinek 2.9.90-22 +- update from trunk + - further localedef fixes +- fix build-locale-archive + +* Fri Apr 24 2009 Jakub Jelinek 2.9.90-21 +- update from trunk + - fix localedef + - fix SHIFT_JIS iconv EILSEQ handling (#497267) + - misc fixes (BZ#10093, BZ#10100) + +* Fri Apr 24 2009 Jakub Jelinek 2.9.90-20 +- update from trunk + - fix p{read,write}v{,64} (#497429, #497434) + - fix strfmon (#496386) + +* Thu Apr 16 2009 Jakub Jelinek 2.9.90-19 +- update from trunk + - fix dlopen from statically linked binaries (#495830) + +* Thu Apr 16 2009 Jakub Jelinek 2.9.90-18 +- update from trunk + - fix fallocate + +* Wed Apr 15 2009 Jakub Jelinek 2.9.90-17 +- update from trunk + - if threads have very small stack sizes, use much smaller buffer + in __get_nprocs when called from within malloc (#494631) + +* Tue Apr 14 2009 Jakub Jelinek 2.9.90-16 +- update from trunk + +* Thu Apr 9 2009 Jakub Jelinek 2.9.90-15 +- rebuilt with fixed gcc to avoid miscompilation of i586 memmove +- reenable experimental malloc again + +* Wed Apr 8 2009 Jakub Jelinek 2.9.90-14 +- update from trunk +- temporarily disable experimental malloc + +* Tue Apr 7 2009 Jakub Jelinek 2.9.90-13 +- update from trunk + - fix strverscmp (#494457) +- configure with --enable-nss-crypt + +* Wed Apr 1 2009 Jakub Jelinek 2.9.90-12 +- update from trunk +- configure with --enable-experimental-malloc + +* Fri Mar 20 2009 Jakub Jelinek 2.9.90-11 +- update from trunk + - POSIX 2008 prototype adjustments for scandir{,64}, alphasort{,64} and + versionsort{,64} + - fix libthread_db (#491197) + +* Tue Mar 10 2009 Jakub Jelinek 2.9.90-10 +- update from trunk + - fix atexit/__cxa_atexit + +* Mon Mar 9 2009 Jakub Jelinek 2.9.90-9 +- update from trunk + - POSIX 2008 support: -D_XOPEN_SOURCE=700 and -D_POSIX_C_SOURCE=200809L +- move libnldbl_nonshared.a on ppc*/s390*/sparc* back to glibc-devel + +* Fri Feb 27 2009 Roland McGrath - 2.9.90-8.1 +- fix libthread_db (#487212) + +* Tue Feb 24 2009 Fedora Release Engineering - 2.9.90-8 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_11_Mass_Rebuild + +* Wed Feb 18 2009 Jakub Jelinek 2.9.90-7 +- update from trunk +- adjust for i586 + i686 from i386 + i686 build +- split static libraries into glibc-static subpackage +- ld -r the whole libpthread.a together to avoid endless issues with + -static ... -lpthread +- require 2.6.18 and later kernel + +* Wed Feb 4 2009 Jakub Jelinek 2.9.90-3 +- update from trunk + - ISO C++ compliant strchr etc. with GCC 4.4+ + - AT_RANDOM support + +* Thu Jan 8 2009 Jakub Jelinek 2.9.90-2 +- update from trunk + +* Fri Jan 2 2009 Jakub Jelinek 2.9.90-1 +- update from trunk (#478314) + +* Mon Dec 8 2008 Jakub Jelinek 2.9-3 +- temporarily disable _nss_dns_gethostbyname4_r (#459756) +- NIS hostname lookup fixes (#473073, #474800, BZ#7058) +- fix unsetenv (#472941) + +* Thu Nov 13 2008 Jakub Jelinek 2.9-2 +- glibc 2.9 release +- fix CPU_ALLOC_SIZE on 32-bit arches (BZ#7029) + +* Wed Nov 12 2008 Jakub Jelinek 2.8.90-17 +- update from trunk + - don't abort on broken DNS replies (#469299, BZ#7009) + - misc fixes (BZ#6966, BZ#7008, BZ#6955, BZ#6843) + +* Fri Oct 31 2008 Jakub Jelinek 2.8.90-16 +- update from trunk + - further resolver fixes + - another dynamic TLS handling fix (#469263) + - misc fixes (BZ#6867, BZ#6875, BZ#6919, BZ#6920, BZ#6942, BZ#6947, + BZ#6968, BZ#6974, BZ#6980, BZ#6995) +- rebuild with newer rpm to avoid stripping + shared libraries when they shouldn't be (#468129) + +* Tue Oct 28 2008 Jakub Jelinek 2.8.90-15 +- update from trunk + - __libc_res_nquery fixes (#466786) + +* Sun Oct 19 2008 Jakub Jelinek 2.8.90-14 +- update from trunk + - fix dynamic TLS handling (#467309) + - fix sys/signalfd.h for C++ (#467172) + - fix sprof (#458861) + - fix _mcount and socket syscalls on s390x (#464146) + - try harder to allocate memory in valloc and pvalloc (#461481) +- fix power6 32-bit libs (#467311) + +* Fri Oct 10 2008 Dennis Gilmore 2.8.90-13 +- apply sparcv9v memset patch from jakub and davem + +* Fri Aug 29 2008 Jakub Jelinek 2.8.90-12 +- update from trunk + - revert origin changes (#457849) + - use MAP_STACK for thread stacks + - misc fixes (BZ#6845, BZ#6544, BZ#6634, BZ#6589, BZ#6790, BZ#6791, + BZ#6824) + - power7 bits (BZ#6817) + - fix expm1 on i?86/x86_64 (#43354, BZ#5794) + +* Sat Aug 2 2008 Jakub Jelinek 2.8.90-11 +- update from trunk + - fix non-absolute $ORIGIN handling (#457560) + - exported some further libresolv APIs (#453325) + - misc fixes + +* Tue Jul 29 2008 Jakub Jelinek 2.8.90-10 +- update from trunk + - resolver fixes + - misc fixes (BZ#6771, BZ#6763, BZ#6698, BZ#6712) + - s390{,x} utmp/utmpx bi-arch support (BZ#6724) + - popen "e" flag +- fr_FR locale changes reenabled + +* Wed Jul 16 2008 Jakub Jelinek 2.8.90-9 +- update from trunk + - fix unbuffered vfprintf if writing to the stream fails (#455360) + - remove useless "malloc: using debugging hooks" message (#455355) + - nscd fixes + - fix resolver alignment issues (#454500) + - fix setvbuf (BZ#6719) + +* Thu Jul 3 2008 Jakub Jelinek 2.8.90-8 +- update from trunk + - watch even resolv.conf in nscd using inotify + - some nscd fixes + +* Fri Jun 13 2008 Jakub Jelinek 2.8.90-7 +- update from trunk + - avoid *lround* on ppc* clobbering cr3/cr4 registers (#450790) + - further nscd fixes (#450704) + - use inotify in nscd to watch files + +* Thu Jun 12 2008 Jakub Jelinek 2.8.90-6 +- update from trunk + - nscd fixes (#450704) + - fix getservbyport (#449358) + - fix regexp.h (#446406) + - avoid crashing on T_DNAME in DNS responses (#450766) + +* Sun May 25 2008 Jakub Jelinek 2.8.90-5 +- update from trunk + +* Tue May 20 2008 Jakub Jelinek 2.8.90-4 +- further getaddrinfo and nscd fixes + +* Sun May 18 2008 Jakub Jelinek 2.8.90-3 +- getaddrinfo and nscd fixes +- reenable assertion checking in rawhide + +* Fri May 16 2008 Jakub Jelinek 2.8.90-2 +- fix getaddrinfo (#446801, #446808) + +* Thu May 15 2008 Jakub Jelinek 2.8.90-1 +- update to trunk + - O(n) memmem/strstr/strcasestr + - i386/x86_64 TLS descriptors support + - concurrent IPv4 and IPv6 DNS lookups by getaddrinfo + +* Mon May 5 2008 Jakub Jelinek 2.8-3 +- don't run telinit u in %%post if both /dev/initctl and + /sbin/initctl exist (#444978) +- workaround GCC ppc64 miscompilation of c{log{,10},acosh,atan}l + (#444996) + +* Wed Apr 30 2008 Jakub Jelinek 2.8-2 +- fix nscd races during GC (BZ#5381) +- rebuilt with fixed GCC to fix regex miscompilation on power6 +- SPARC fixes + +* Sat Apr 12 2008 Jakub Jelinek 2.8-1 +- 2.8 release + +* Fri Apr 11 2008 Jakub Jelinek 2.7.90-16 +- update to trunk + - misc fixes (BZ#4997, BZ#5741) + - make sure all users of __libc_setlocale_lock know it is + now a rwlock + - fix ppc/ppc64 compatibility _sys_errlist and _sys_siglist + symbols + +* Thu Apr 10 2008 Jakub Jelinek 2.7.90-15 +- update to trunk + - misc fixes (BZ#4314, BZ#4407, BZ#5209, BZ#5436, BZ#5768, BZ#5998, + BZ#6024) +- restart sshd in %%post when upstart is used - it doesn't have + /dev/initctl (#441763) +- disable assert checking again + +* Tue Apr 8 2008 Jakub Jelinek 2.7.90-14 +- update to trunk + - misc fixes (BZ#5443, BZ#5475, BZ#5478, BZ#5939, BZ#5979, BZ#5995, + BZ#6004, BZ#6007, BZ#6020, BZ#6021, BZ#6042) + - change mtrace to keep perl 5.10 quiet (#441082) + - don't share conversion state between mbtowc and wctomb (#438687) + - if st_blksize is too large and malloc fails, retry with smaller + buffer in opendir (#430768) + - correct *printf overflow test (#358111) + +* Fri Mar 28 2008 Jakub Jelinek 2.7.90-13 +- update to trunk + - don't define ARG_MAX in , as it is no longer + constant - use sysconf (_SC_ARG_MAX) to get the current + argument size limit + - fix build on sparc64 +- only service sshd condrestart if /etc/rc.d/init.d/sshd exists + (#428859) + +* Wed Mar 26 2008 Jakub Jelinek 2.7.90-12 +- update to trunk + - new CLONE_* flags in (#438542) + - nis+ errno clobbering fix (#437945) + - fix adjtime (#437974) + +* Fri Mar 14 2008 Jakub Jelinek 2.7.90-11 +- update to trunk +- remove , define _XOPEN_STREAMS -1 (#436349) + +* Wed Mar 5 2008 Jakub Jelinek 2.7.90-8 +- update to trunk + - {,v}{as,d}printf and obstack_{,v}printf fortification (#435905) + - fix getnameinfo/gethostbyaddr (#428067, BZ#5790) + - fix yp_order (#435519, BZ#5854) + - misc fixes (BZ#5779, BZ#5736, BZ#5627, BZ#5818, BZ#5012) +- merge review cleanup (Tom Callaway, #225806) + +* Sat Feb 16 2008 Jakub Jelinek 2.7.90-7 +- update to trunk + - make NI_MAXHOST and NI_MAXSERV available even in BSDish + namespaces (BZ#5737) + - timerfd_* syscalls + +* Fri Feb 1 2008 Jakub Jelinek 2.7.90-6 +- fix build + +* Thu Jan 31 2008 Jakub Jelinek 2.7.90-5 +- update to trunk +- rebuild with gcc 4.3 + +* Fri Jan 11 2008 Jakub Jelinek 2.7.90-4 +- update to trunk + - misc fixes (BZ#5541, BZ#5545, BZ#5553, BZ#5112, BZ#5520) + - getaddrinfo fixes + - signalize EOVERFLOW from sem_post instead of overflowing + the counter + - fix i?86 makecontext + - fix iconv for iso-2022-jp//translit (#397021) + +* Thu Jan 3 2008 Jakub Jelinek 2.7.90-3 +- update to trunk + - fix recognition of interface family (#425768) + - add __THROW to __ctype_{b,tolower,toupper}_loc prototypes + +* Thu Dec 27 2007 Jakub Jelinek 2.7.90-2 +- update to trunk + - nsswitch fix (#425768) +- temporarily enable assert checking + +* Wed Dec 12 2007 Jakub Jelinek 2.7.90-1 +- update to trunk + - fix __USE_STRING_INLINES on i?86 (#408731, #371711) + - fix *scanf (#388751) + +* Wed Oct 17 2007 Jakub Jelinek 2.7-1 +- glibc 2.7 release +- fix tzfile.c for times after last transition (#333561) +- fix sem_post@GLIBC_2.0 on i?86 +- appease valgrind in libpthread.so initialization +- misc fixes (BZ#3425, BZ#5184, BZ#5186) + +* Mon Oct 15 2007 Jakub Jelinek 2.6.90-21 +- fix getgr{name,gid}{,_r} with nscd + +* Sun Oct 14 2007 Jakub Jelinek 2.6.90-20 +- install (#330031) +- disable -D_FORTIFY_SOURCE{,=2} support (with a warning) for + GCC 3.4.x and earlier(#327641) +- pl_PL locale changes (BZ#4098, #242296) +- misc fixes (BZ#1140, BZ#3195, BZ#3242, BZ#4359) + +* Thu Oct 11 2007 Jakub Jelinek 2.6.90-19 +- fix +- simple preprocessor in localedef, fix de_DE collation with it + +* Wed Oct 10 2007 Jakub Jelinek 2.6.90-18 +- add signalfd, eventfd, eventfd_read, eventfd_write +- qsort speedups +- workaround for cpuid bugs (#324081) +- make sure gettext's conversion_lock is initialized even if + program isn't linked against libpthread.so.0, only dlopens it (#321761) +- misc fixes (BZ#5112, BZ#5113, BZ#5104, BZ#5063, BZ#5010, BZ#4407, + BZ#3924, BZ#5103, BZ#2633, BZ#181, BZ#73, #321901) + +* Wed Oct 3 2007 Jakub Jelinek 2.6.90-17 +- fix {,v}swprintf with -D_FORTIFY_SOURCE=1 -mlong-double-64 on ppc*/s390*/sparc* +- strcoll fixes +- misc fixes (BZ#645, BZ#5071) +- locale fixes (BZ#4941, #299321, #203364, #196711, #236212) + +* Sat Sep 29 2007 Jakub Jelinek 2.6.90-16 +- misc fixes (BZ#4963, BZ#4972, BZ#5028, BZ#5043, BZ#5058) +- improve -D_FORTIFY_SOURCE{,=2} diagnostic through warning/error + attributes +- fix wcscpy, wcpcpy, fgetws, fgetws_unlocked, swprintf and vswprintf + fortification inlines +- fix a scalability issue with lazy binding in heavily multithreaded + programs + +* Thu Sep 20 2007 Jakub Jelinek 2.6.90-15 +- $5$ (SHA-256) and $6$ (SHA-512) support in crypt + (#228697, #249477, #173834) + +* Tue Sep 18 2007 Jakub Jelinek 2.6.90-14 +- -D_FORTIFY_SOURCE{,=2} support for C++ +- fortification of fread{,_unlocked} +- support *scanf m allocation modifier (%%ms, %%mls, %%mc, ...) +- in -std=c99 or -D_XOPEN_SOURCE=600 mode don't recognize + %%as, %%aS and %%a[ as a GNU extension for *scanf +- fix splice, vmsplice, tee return value, make them cancellation + points +- mq_open checking +- use inline function rather than function-like macro + for open{,at}{,64} checking +- IFA_F_OPTIMISTIC handling in getaddrinfo (#259681) +- fix an ABBA deadlock in ld.so (#284171) +- remove sparc{32,64} unwind info from _start and clone + +* Mon Aug 27 2007 Jakub Jelinek 2.6.90-13 +- fix personality on x86_64/ppc/ppc64 (#256281) + +* Sat Aug 25 2007 Jakub Jelinek 2.6.90-12 +- readd x86_64 gettimeofday stuff, initialize it earlier +- nis_list fix (#254115) +- workaround for bugs in ia64 silly /emul/ia32-linux hack (#253961) +- misc fixes (BZ#3924, BZ#4566, BZ#4582, BZ#4588, BZ#4726, BZ#4946, + BZ#4905, BZ#4814, BZ#4925, BZ#4936, BZ#4896, BZ#4937, BZ#3842, + BZ#4554, BZ#4557, BZ#4938) + +* Fri Aug 17 2007 Jakub Jelinek 2.6.90-11 +- remove __strtold_internal and __wcstold_internal from ppc*/s390*/sparc* + *-ldbl.h headers +- temporarily backout x86_64 gettimeofday.S changes (#252453) +- some further sparc, sparc64 and alpha fixes + +* Wed Aug 15 2007 Jakub Jelinek 2.6.90-10 +- don't open /etc/ld.so.{cache,preload} with O_NOATIME (#252146) +- s390{,x}, alpha and sparc fixes +- sparcv9 is no longer an aux arch, as we expect + to not build sparc.rpm glibc any longer, only sparcv9.rpm, + sparc64.rpm and new two aux arches sparcv9v.rpm and sparc64v.rpm + +* Tue Aug 14 2007 Jakub Jelinek 2.6.90-9 +- private futex even for mutexes and condvars +- some further O_CLOEXEC changes +- use vDSO on x86_64 if available +- ia64 build fixes (#251983) + +* Fri Aug 10 2007 Roland McGrath 2.6.90-8 +- update to trunk + - fix missing strtold_l export on ppc64 + +* Thu Aug 9 2007 Roland McGrath 2.6.90-6 +- update to trunk + - fix local PLT regressions +- spec file revamp for new find-debuginfo.sh + +* Sun Aug 5 2007 Jakub Jelinek 2.6.90-4 +- fix librt.so and librtkaio.so on ppc32, so that it is not using + bss PLT + +* Sat Aug 4 2007 Jakub Jelinek 2.6.90-3 +- fix open{,at}{,64} macro for -pedantic (#250897) +- add transliteration for l with stroke (#250492) +- fix strtod ("-0", NULL) +- update License tag + +* Wed Aug 1 2007 Jakub Jelinek 2.6.90-2 +- make aux-cache purely optional performance optimization in ldconfig, + don't issue any errors if it can't be created (#250430) +- remove override_headers hack, BuildRequire >= 2.6.22 kernel-headers + and rely on its content + +* Tue Jul 31 2007 Jakub Jelinek 2.6.90-1 +- update to trunk + - private futex optimizations + - open{,at}{,64} argument checking +- ldconfig speedups + +* Sun Jul 8 2007 Jakub Jelinek 2.6-4 +- filter pseudo-files from debuginfo source lists (#245714) +- fix sscanf when errno is EINTR before the call (BZ#4745) +- save/restore errno around reading /etc/default/nss (BZ#4702) +- fix LD_HWCAP_MASK handling +- disable workaround for #210748, instead backport + ld.so locking fixes from the trunk (#235026) +- new x86_64 memcpy +- don't write uninitialized padding bytes to nscd socket +- fix dl{,v}sym, dl_iterate_phdr and dlopen if some library is + mapped into ld.so's inter-segment hole on x86_64 (#245035, #244545) +- fix LD_AUDIT=a:b program (#180432) +- don't crash on pseudo-zero long double values passed to + *printf on i?86/x86_64/ia64 (BZ#4586) +- fix *printf %%La and strtold with some hexadecimal floating point + constants on ppc/ppc64 +- fix nextafterl on ppc/ppc64 +- fix sem_timedwait on i?86 and x86_64 + +* Thu May 24 2007 Jakub Jelinek 2.6-3 +- don't use %%config(missingok) for locale-archive.tmpl, + instead of removing it altogether truncate it to zero + size (#240697) +- add a workaround for #210748 + +* Mon May 21 2007 Jakub Jelinek 2.6-2 +- restore malloc_set_state backwards compatibility (#239344) +- fix epoll_pwait (BZ#4525) +- fix printf with unknown format spec or positional arguments + and large width and/or precision (BZ#4514) +- robust mutexes fix (BZ#4512) + +* Tue May 15 2007 Roland McGrath 2.6-1 +- glibc 2.6 release + +* Fri May 11 2007 Jakub Jelinek 2.5.90-24 +- utimensat, futimens and lutimes support + +* Thu May 10 2007 Jakub Jelinek 2.5.90-23 +- use madvise MADV_DONTNEED in malloc +- fix ia64 feraiseexcept +- fix s390{,x} feholdexcept (BZ#3427) +- ppc fenv fixes +- make fdatasync a cancellation point (BZ#4465) +- fix *printf for huge precisions with wide char code and multi-byte + strings +- fix dladdr (#232224, BZ#4131) + +* Fri May 4 2007 Jakub Jelinek 2.5.90-22 +- add transliteration for (BZ#3213) +- fix *scanf with %%f on hexadecimal floats without exponent (BZ#4342) +- fix *printf with very large precisions for %%s (#238406, BZ#4438) +- fix inet_ntop size checking for AF_INET (BZ#4439) +- for *printf %%e avoid 1.000e-00, for exponent 0 always use + sign (#238431) +- fix a regression introduced in #223467 changes +- gethostby*_r alignment fixes (BZ#4381) +- fix ifaddrs error handling + +* Mon Apr 16 2007 Jakub Jelinek 2.5.90-21 +- don't include individual locale files in glibc-common, + rather include prepared locale-archive template and let + build-locale-archive create locale-archive from the template + and any user supplied /usr/lib/locale/*_* directories, + then unlink the locale-archive template - this should save + > 80MB of glibc-common occupied disk space +- fix _XOPEN_VERSION (BZ#4364) +- fix printf with %%g and values tiny bit smaller than 1.e-4 (#235864, + BZ#4362) +- fix NIS+ __nisfind_server (#235229) + +* Sat Mar 31 2007 Jakub Jelinek 2.5.90-20 +- assorted NIS+ speedups (#223467) +- fix HAVE_LIBCAP configure detection (#178934) +- remove %%{_prefix}/sbin/rpcinfo from glibc-common (#228894) +- nexttoward*/nextafter* fixes (BZ#3306) +- feholdexcept/feupdateenv fixes (BZ#3427) +- speed up fnmatch with two or more * in the pattern + +* Sat Mar 17 2007 Jakub Jelinek 2.5.90-19 +- fix power6 libm compat symbols on ppc32 (#232633) +- fix child refcntr in NPTL fork (#230198) +- fix ifaddrs with many net devices on > 4KB page size arches (#230151) +- fix pthread_mutex_timedlock on x86_64 (#228103) +- various fixes (BZ#3919, BZ#4101, BZ#4130, BZ#4181, BZ#4069, BZ#3458) + +* Wed Feb 21 2007 Jakub Jelinek 2.5.90-18 +- fix nftw with FTW_CHDIR on / (BZ#4076) +- nscd fixes (BZ#4074) +- fix fmod{,f,l} on i?86 (BZ#3325) +- support localized digits for fp values in *scanf (BZ#2211) +- namespaces fixes (BZ#2633) +- fix euidaccess (BZ#3842) +- glob fixes (BZ#3996) +- assorted locale data fixes (BZ#1430, BZ#672, BZ#58, BZ#3156, + BZ#2692, BZ#2648, BZ#3363, BZ#3334, BZ#3326, BZ#3322, BZ#3995, + BZ#3885, BZ#3884, BZ#3851) + +* Sun Feb 11 2007 Jakub Jelinek 2.5.90-17 +- RFC2671 support in resolver (#205842) +- fix strptime (BZ#3944) +- fix regcomp with REG_NEWLINE (BZ#3957) +- fix pthread_mutex_timedlock on x86_64 (#228103) + +* Fri Feb 2 2007 Jakub Jelinek 2.5.90-16 +- add strerror_l +- fix application crashes when doing NSS lookups through nscd + mmapped databases and nscd decides to start garbage collection + during the lookups (#219145, #225315) +- fix %%0lld printing of 0LL on 32-bit architectures (BZ#3902) +- ignore errors from install-info in glibc-devel scriptlets + (#223691) + +* Wed Jan 17 2007 Jakub Jelinek 2.5.90-15 +- fix NIS getservbyname when proto is NULL +- fix nss_compat +group handling (#220658) +- cache services in nscd +- fix double free in fts_close (#222089) +- fix vfork+execvp memory leak (#221187) +- soft-fp fixes (BZ#2749) +- further strtod fixes (BZ#3855) +- make sure pthread_kill doesn't return EINVAL even if + the target thread exits in between pthread_kill ESRCH check + and the actual tgkill syscall (#220420) +- fix ABBA deadlock possibility in ld.so scope locking code + +* Tue Dec 19 2006 Jakub Jelinek 2.5.90-14 +- fix {j,m}rand48{,_r} on 64-bit arches (BZ#3747) +- handle power6x AT_PLATFORM (#216970) +- fix a race condition in getXXbyYY_r (#219145) +- fix tst-pselect testcase + +* Thu Dec 14 2006 Jakub Jelinek 2.5.90-13 +- fix setcontext on ppc32 (#219107) +- fix wide stdio after setvbuf (#217064, BZ#2337) +- handle relatime mount option in statvfs +- revert i?86/x86_64 clone CFI temporarily + +* Sun Dec 10 2006 Jakub Jelinek 2.5.90-12 +- fix hasmntopt (#218802) +- fix setusershell and getusershell (#218782) +- strtod fixes (BZ#3664, BZ#3673, BZ#3674) +- fix memusage with realloc (x, 0) + +* Tue Dec 5 2006 Jakub Jelinek 2.5.90-11 +- allow suid apps to setenv NIS_PATH and influence through that + nis_list and nis_lookup (#209155) +- fix ttyname and ttyname_r with invalid file descriptor (#218276) +- cs_CZ LC_TIME fixes (#218438) +- fix build with 2.6.19+ headers (#217723) + +* Fri Dec 1 2006 Jakub Jelinek 2.5.90-10 +- fix x86-64 restore_rt unwind info + +* Thu Nov 30 2006 Jakub Jelinek 2.5.90-9 +- fix last svc_run change (#217850) +- on ppc64 build __libc_start_main without unwind info, + as it breaks MD_FROB_UPDATE_CONTEXT (#217729, #217775; in the + future that could be fixable just by providing .cfi_undefined r2 + in __libc_start_main instead) +- add unwind info for x86-64 restore_rt signal return landing pad + (#217087) +- add power6x subdir to /%%{_lib}/ and /%%{_lib}/rtkaio/, + link all libs from ../power6/* into them + +* Tue Nov 28 2006 Jakub Jelinek 2.5.90-8 +- fix svc_run (#216834, BZ#3559) +- add -fasynchronous-unwind-tables to CFLAGS (#216518) +- make sure there is consistent timestamp for /etc/ld.so.conf, + /etc/localtime and /etc/rpc between multilib glibc rpms + +* Mon Nov 20 2006 Jakub Jelinek 2.5.90-7 +- handle IPv6 addresses in /etc/hosts that are mappable to + IPv4 addresses in IPv4 host lookups (#215283) +- fix :include: /etc/alias handling (#215572) +- handle new tzdata format to cope with year > 2037 transitions + on 64-bit architectures + +* Fri Nov 10 2006 Jakub Jelinek 2.5.90-6 +- fix strxfrm fix +- fix i?86 floor and ceil inlines (BZ#3451) + +* Thu Nov 9 2006 Jakub Jelinek 2.5.90-5 +- fix sysconf (_SC_LEVEL{2,3}_CACHE_SIZE) on Intel Core Duo + CPUs +- fix libthread_db.so on TLS_DTV_AT_TP architectures +- fix --inhibit-rpath (#214569) +- fix _r_debug content when prelinked ld.so executes + a program as its argument +- fix strxfrm +- powerpc-cpu add-on updates + +* Fri Nov 3 2006 Jakub Jelinek 2.5.90-4 +- fix atexit backwards compatibility (#213388) +- add mai_IN locale (#213415) +- remove bogus %%{_libdir}/librt.so.1 symlink (#213555) +- fix memusage (#213656) +- change libc.info category (#209493) + +* Sun Oct 29 2006 Jakub Jelinek 2.5.90-3 +- fix suid/sgid binaries on i?86/x86_64 (#212723) + +* Fri Oct 27 2006 Jakub Jelinek 2.5.90-2 +- fix ia64 build +- don't call _dl_close outside of dl_load_lock critical section + if dlopen failed (BZ#3426) +- add rtld scope locking (#211133) + +* Wed Oct 25 2006 Jakub Jelinek 2.5.90-1 +- fix i?86 6 argument syscalls (e.g. splice) +- fix rtld minimal realloc (BZ#3352) +- fix RFC3484 getaddrinfo sorting according to rules 4 and 7 (BZ#3369) +- fix xdrmem_setpos (#211452) +- bump __GLIBC_MINOR__ +- increase PTHREAD_STACK_MIN on ppc{,64} to 128K to allow + 64K pagesize kernels (#209877) +- speed up initgroups on NIS+ (#208203) + +* Mon Oct 2 2006 Jakub Jelinek 2.5-2 +- fix nscd database growing (#207928) +- bypass prelinking when LD_DYNAMIC_WEAK=1 is in the environment + +* Fri Sep 29 2006 Jakub Jelinek 2.5-1 +- glibc 2.5 release + +* Wed Sep 27 2006 Jakub Jelinek 2.4.90-36 +- rebuilt with gcc-4.1.1-26 to fix unwind info + +* Mon Sep 25 2006 Jakub Jelinek 2.4.90-35 +- fix glob with large number of matches (BZ#3253) +- fix fchownat on kernels that don't support that syscall (BZ#3252) +- fix lrintl on s390{,64} + +* Sat Sep 23 2006 Jakub Jelinek 2.4.90-34 +- fix ppc{32,64} longjmp (BZ#3225) +- fix user visible spelling errors (BZ#3137) +- fix l{,l}rint{,f,l} around zero (BZ#2592) +- avoid stack trampoline in s390{,x} makecontext + +* Tue Sep 19 2006 Jakub Jelinek 2.4.90-33 +- fix dlclose (#206639) +- don't load platform optimized libraries if kernel doesn't set + AT_PLATFORM +- fix ppc{32,64} libSegFault.so +- use -mtune=generic even for glibc-devel.i386 (#206437) +- fix /%%{_lib}/librt.so.1 symlink + +* Fri Sep 15 2006 Jakub Jelinek 2.4.90-32 +- on ppc* use just AT_PLATFORM and altivec AT_HWCAP bit for library selection +- fix lrintl and lroundl on ppc{,64} +- use hidden visibility on fstatat{,64} and mknodat in libc_nonshared.a + +* Sun Sep 10 2006 Jakub Jelinek 2.4.90-31 +- fix pthread_cond_{,timed}wait cancellation (BZ#3123) +- fix lrint on ppc32 (BZ#3155) +- fix malloc allocating more than half of address space (BZ#2775) +- fix mktime on 32-bit arches a few years after 2038 (BZ#2821) + +* Thu Sep 7 2006 Jakub Jelinek 2.4.90-30 +- add librtkaio, to use it add /%%{lib}/rtkaio to your + LD_LIBRARY_PATH or /etc/ld.so.conf +- fix or_IN February name (#204730) +- fix pthread_create called from cancellation handlers (BZ#3124) +- fix regex case insensitive searches with characters where upper + and lower case multibyte representations have different length + (e.g. I and dotless i, #202991) + +* Tue Sep 5 2006 Jakub Jelinek 2.4.90-29 +- randomize resolver query ids before use instead after use (#205113) +- fix resolver symver checking with DT_GNU_HASH (#204909) +- put .hash section in glibc libraries at the end of RO segment + when .gnu.hash is present + +* Thu Aug 31 2006 Jakub Jelinek 2.4.90-28 +- another malloc doubly linked list corruption problem fix (#204653) + +* Thu Aug 31 2006 Jakub Jelinek 2.4.90-27 +- allow $LIB and $PLATFORM in dlopen parameters even in suid/sgid (#204399) +- handle $LIB/$PLATFORM in LD_LIBRARY_PATH +- fix splice prototype (#204530) + +* Mon Aug 28 2006 Jakub Jelinek 2.4.90-26 +- real fix for the doubly linked list corruption problem +- try harder in realloc to allocate memory (BZ#2684) +- fix getnameinfo error reporting (#204122) +- make localedef more robust on invalid input (#203728) + +* Fri Aug 25 2006 Jakub Jelinek 2.4.90-25 +- temporarily back out code to limit number of unsorted block + sort iterations (#203735, #204027) +- handle PLT symbols in dladdr properly (BZ#2683) +- avoid malloc infinite looping for allocations larger than + the system can allocate (#203915) + +* Tue Aug 22 2006 Jakub Jelinek 2.4.90-23 +- malloc fixes, especially for 32-bit arches (#202309) +- further *_IN locale fixes (#200230) +- fix get{serv,rpc}ent{,_r} if NIS map is empty (#203237) +- fix /usr/bin/iconv (#203400) + +* Fri Aug 18 2006 Jakub Jelinek 2.4.90-22 +- rebuilt with latest binutils to pick up 64K -z commonpagesize + on ppc/ppc64 (#203001) + +* Tue Aug 15 2006 Jakub Jelinek 2.4.90-21 +- if some test gets stuck, kill the tee process after make check + finishes +- build with -mtune=generic on i686 and x86_64 + +* Tue Aug 15 2006 Jakub Jelinek 2.4.90-20 +- PTHREAD_PRIO_PROTECT support +- fix errno if nice() fails (#201826) + +* Thu Aug 10 2006 Jakub Jelinek 2.4.90-19 +- adaptive malloc brk/mmap threshold +- fix fchownat to use kernel syscall (if available) on many arches (#201870) +- only define O_DIRECT with -D_GNU_SOURCE on ia64 to match all + other arches (#201748) + +* Mon Aug 7 2006 Jakub Jelinek 2.4.90-18 +- NIS+ fixes +- fix memusage and xtrace scripts (#200736) +- redirect /sbin/service sshd condrestart std{out,err} to /dev/null + when executed from glibc_post_upgrade + +* Wed Aug 2 2006 Jakub Jelinek 2.4.90-17 +- typo fix for the dladdr patch +- build i?86 glibc with -mno-tls-direct-seg-refs (#200469) + +* Wed Aug 2 2006 Jakub Jelinek 2.4.90-16 +- fix dladdr on binaries/libraries with only DT_GNU_HASH and no + DT_HASH (#200635) +- fix early timeout of initgroups data in nscd (#173019) +- add am/pm display to es_PE and es_NI locales (#167101) +- fix nss_compat failures when nis/nis+ unavailable (#192072) + +* Mon Jul 31 2006 Roland McGrath 2.4.90-15 +- fix missing destructor calls in dlclose (#197932) +- enable transliteration support in all locales (#196713) +- disallow RTLD_GLOBAL flag for dlmopen in secondary namespaces (#197462) +- PI mutex support + +* Mon Jul 10 2006 Jakub Jelinek 2.4.90-13 +- DT_GNU_HASH support + +* Fri Jun 30 2006 Jakub Jelinek 2.4.90-12 +- buildrequire gettext +- enable fstatat64/newfstatat syscalls even on ppc*/s390*/ia64 (#196494) +- fix out of memory behavior in gettext (#194321) +- fix regex on multi-byte non-UTF-8 charsets (#193873) +- minor NIS+ fixes (#190803) +- don't use cancellable calls in posix_spawn* and only set{u,g}id + current thread if requested (#193631) + +* Wed May 31 2006 Jakub Jelinek 2.4.90-11 +- don't exit from nscd -i before the database is + actually invalidated, add locking to prune_cache (#191464) +- build glibc-devel.i386 static libraries with + -mno-tls-direct-seg-refs -DNO_TLS_DIRECT_SEG_REFS +- RFC3542 support (advanced API for IPv6; #191001, BZ##2693) + +* Wed May 24 2006 Jakub Jelinek 2.4.90-10 +- on i686 make glibc owner of /lib/i686 directory (#192597) +- search parent NIS+ domains (#190803) + +* Sun May 21 2006 Jakub Jelinek 2.4.90-9 +- update from CVS + - big NIS+ changes + +* Fri May 19 2006 Jakub Jelinek 2.4.90-8 +- update from CVS + - fix nss_compat when SETENT_BATCH_READ=TRUE is in /etc/default/nss + - fix RFC3484 precedence table for site-local and ULA addresses (#188364) + - fix a sunrpc memory leak + +* Thu May 11 2006 Jakub Jelinek 2.4.90-7 +- update from CVS + - fix tcgetattr (#177965) + - fix (#191264) + +* Fri May 5 2006 Jakub Jelinek 2.4.90-6 +- update from CVS +- rebuilt using fixed rpm + +* Fri May 5 2006 Jakub Jelinek 2.4.90-5 +- update from CVS + - some NIS+ fixes + - allow overriding rfc3484 address sorting tables for getaddrinfo + through /etc/gai.conf (sample config file included in %%doc directory) + +* Mon May 1 2006 Jakub Jelinek 2.4.90-4 +- update from CVS + - SETENT_BATCH_READ /etc/default/nss option for speeding up + some usages of NIS+ (#188246) + - move debug state change notification (#179208) + - fix ldd script if one of the dynamic linkers is not installed (#190259) + +* Thu Apr 27 2006 Jakub Jelinek 2.4.90-3 +- update from CVS + - fix a typo in nscd.conf (#190085) + - fix handling of SIGHUP in nscd when some caches are disabled (#189978) + - make nscd paranoia mode working with non-root server-user (#189779) + +* Wed Apr 26 2006 Jakub Jelinek 2.4.90-2 +- update from CVS + - fix getaddrinfo (#190002) + - add auto-propagate nscd.conf options (#177154) + - fix nscd auditing (#169148) + +* Tue Apr 25 2006 Jakub Jelinek 2.4.90-1 +- update from CVS + +* Mon Apr 24 2006 Jakub Jelinek 2.4-6 +- update from CVS + - NIS+ fixes + - don't segfault on too large argp key values (#189545) + - getaddrinfo fixes for RFC3484 (#188364) + +* Tue Mar 28 2006 Jakub Jelinek 2.4-5 +- update from CVS + - pshared robust mutex support + - fix btowc and bwtoc in C++ (#186410) + - fix NIS+ (#186592) + - don't declare __wcsto*l_internal for non-GCC or if not -O1+ (#185667) +- don't mention nscd failures on 2.0 kernels (#185335) + +* Tue Mar 7 2006 Roland McGrath 2.4-4 +- back up %%{ix86} gdb conflicts to < 6.3.0.0-1.111 + +* Tue Mar 7 2006 Jakub Jelinek 2.4-3 +- really fix rintl on ppc64 + +* Tue Mar 7 2006 Jakub Jelinek 2.4-2 +- accurate unwind info for lowlevellock.h stubs on %%{ix86} +- fix ppc/ppc64 ceill, floorl, rintl, roundl and truncl (BZ#2423) + +* Mon Mar 6 2006 Jakub Jelinek 2.4-1 +- update from CVS + - glibc 2.4 release + +* Mon Mar 6 2006 Jakub Jelinek 2.3.91-2 +- update from CVS + - fix sYSMALLOc for MALLOC_ALIGNMENT > 2 * SIZE_SZ (#183895) + - revert ppc32 malloc alignment patch, it breaks malloc_set_state + and needs some further thoughts and time (#183894) +- provide accurate unwind info for lowlevellock.h stubs on x86_64 + +* Thu Mar 2 2006 Jakub Jelinek 2.3.91-1 +- update from CVS + - fixes for various arches +- ensure malloc returns pointers aligned to at least + MIN (2 * sizeof (size_t), __alignof__ (long double)) + (only on ppc32 this has not been the case lately with addition + of 128-bit long double, #182742) + +* Wed Mar 1 2006 Jakub Jelinek 2.3.90-39 +- update from CVS + +* Fri Feb 17 2006 Jakub Jelinek 2.3.90-38 +- update from CVS + - robust mutexes rewrite + +* Mon Feb 13 2006 Jakub Jelinek 2.3.90-37 +- update from CVS + - *at fixes + - unshare syscall wrapper + +* Sat Feb 4 2006 Jakub Jelinek 2.3.90-36 +- update from CVS + - fix frequency setting for ITIMER_PROF (#179938, BZ#2268) + - fix powerpc inline fegetround () + - fix nptl_db (#179946) + +* Fri Feb 3 2006 Jakub Jelinek 2.3.90-35 +- update from CVS + - handle futimesat (fd, NULL, tvp) as futimes (fd, tvp) +- fix q{e,f,g}cvt{,_r} for -mlong-double-64 + +* Thu Feb 2 2006 Jakub Jelinek 2.3.90-34 +- fix with C++ and -mlong-double-64 (#179742) +- add nexttowardl redirect for -mlong-double-64 + +* Thu Feb 2 2006 Jakub Jelinek 2.3.90-33 +- update from CVS + - long double support fixes + +* Wed Feb 1 2006 Jakub Jelinek 2.3.90-32 +- update from CVS + - 128-bit long double fixes for ppc{,64}, s390{,x} and sparc{,v9}, + alpha 128-bit long double support +- add inotify syscall numbers to the override headers + (#179366) + +* Mon Jan 30 2006 Jakub Jelinek 2.3.90-31 +- update from CVS + - 128-bit long double on ppc, ppc64, s390, s390x and sparc{,v9} +- add some new syscall numbers to the override + headers + +* Mon Jan 9 2006 Jakub Jelinek 2.3.90-30 +- update from CVS + - initializer fixes for -std=c{8,9}9 on 32-bit + arches +- avoid writable .rodata (#177121) + +* Fri Jan 6 2006 Jakub Jelinek 2.3.90-29 +- update from CVS + - make pthread_mutex_t an unnamed union again, as it affects + libstdc++ ABI mangling + +* Fri Jan 6 2006 Jakub Jelinek 2.3.90-28 +- update from CVS + - make aio_suspend interruptible by signals (#171968) + +* Fri Jan 6 2006 Jakub Jelinek 2.3.90-27 +- only rely on d_type in 32-bit getdents on s390 for 2.6.11+ + +* Wed Jan 4 2006 Jakub Jelinek 2.3.90-26 +- update from CVS + - for newly linked lio_listio* callers, send per request + notifications (#170116) + - fixup nscd -S option removal changes (#176860) + - remove nonnull attribute from ctermid (#176753) + - fix PTHREAD_*_INITIALIZER{,_NP} on 64-bit arches + - SPARC NPTL support for pre-v9 CPUs +- drop support for 2.4.xx and < 2.6.9 kernels + +* Mon Jan 2 2006 Jakub Jelinek 2.3.90-25 +- update from CVS + - s390{,x} and sparc{,64} pointer mangling fixes +- install a sanitized LinuxThreads + +* Mon Jan 2 2006 Jakub Jelinek 2.3.90-24 +- update from CVS + - nscd audit changes (#174422) + - ppc{32,64} vDSO support and ppc32 hp-timing + +* Tue Dec 27 2005 Jakub Jelinek 2.3.90-23 +- update from CVS + - robust mutexes +- fix transliteration segfaults (#176573, #176583) +- ignore prelink temporaries in ldconfig (#176570) + +* Wed Dec 21 2005 Jakub Jelinek 2.3.90-22 +- update from CVS + - minor fts fixes +- revert broken _Pragma () workaround +- fix ldconfig on bi-arch architectures (#176316) + +* Tue Dec 20 2005 Jakub Jelinek 2.3.90-21 +- update from CVS + - fix pointer (de)mangling in gconv_cache.c + +* Tue Dec 20 2005 Jakub Jelinek 2.3.90-20 +- update from CVS + - time ((void *) 1) should segfault, not return -EFAULT (#174856, BZ#1952) + - fix errlist generation +- update ulps for GCC 4.1 on IA-64 + +* Mon Dec 19 2005 Jakub Jelinek 2.3.90-19 +- update from CVS + - sysdeps/generic reorg + - setjmp/longjmp jump pointer mangling +- rebuilt with GCC 4.1-RH prerelease, worked around broken _Pragma () + handling in it +- remove glibc-profile subpackage +- use non-PLT calls for malloc/free/realloc/memalign invocations in + mtrace and mcheck hooks (#175261) +- setjmp/longjmp jump pointer mangling on ppc{,64}/ia64/s390{,x} + +* Sat Nov 19 2005 Jakub Jelinek 2.3.90-18 +- update from CVS + - change for broken apps that #define const /**/, + handle non-GCC compilers + - fix ppc{32,64} strncmp (BZ#1877, #173643, IT#83510) + - provide shmatt_t typedef in ia64 2.3.90-17 +- update from CVS + - fix in C++ + - {fstat,fchown,rename,unlink}at fixes + - epoll_wait is now a cancellation point + +* Tue Nov 15 2005 Jakub Jelinek 2.3.90-16 +- update from CVS +- make sure waitid syscall is used on ppc*/s390* + +* Thu Oct 20 2005 Jakub Jelinek 2.3.90-15 +- update from CVS + - be permissive in %%n check because of kernel bug #165351 (#171240) + - don't misalign stack in pthread_once on x86_64 (#170786, IT#81521) + - many locale fixes + +* Mon Oct 10 2005 Jakub Jelinek 2.3.90-14 +- update from CVS + - fix malloc bug after fork introduced in the last update + - fix getent hosts IP for IPv4 IPs (#169831) + +* Mon Oct 3 2005 Jakub Jelinek 2.3.90-13 +- update from CVS + - fix setuid etc. hangs if some thread exits during the call (#167766) + - fix innetgr memory leak (#169051) + - support > 2GB nscd log files (#168851) + - too many other changes to list here +- include errno in nscd message if audit_open failed (#169148) + +* Mon Sep 12 2005 Jakub Jelinek 2.3.90-12 +- update from CVS + - netgrp handling fixes (#167728) + - fix memory leak in setlocale (BZ#1318) + - fix hwcaps computation + - several regex portability improvements (#167019) + - hypotf fix + - fix *printf return code if underlying write fails (BZ#1146) + - PPC64 dl{,v}sym fixes for new ABI .opd symbols +- fix calloc with MALLOC_PERTURB_ in environment on 64-bit architectures + (#166719) +- source /etc/sysconfig/nscd (if it exists) in /etc/rc.d/init.d/nscd + (#167083) +- add %%triggerin for tzdata to glibc-common, so that tzdata updates + update /etc/localtime and /var/spool/postfix/etc/localtime if they + exist (#167787) + +* Mon Aug 29 2005 Jakub Jelinek 2.3.90-11 +- FUTEX_WAKE_OP support to speed up pthread_cond_signal + +* Wed Aug 24 2005 Jakub Jelinek 2.3.90-10 +- update from CVS + - fix growing of nscd persistent database (BZ#1204) + - fix _FORTIFY_SOURCE mbstowcs and wcstombs if destination size + is known at compile time, but length argument is not + +* Mon Aug 22 2005 Jakub Jelinek 2.3.90-9 +- update from CVS + - fix resolving over TCP (#161181, #165802) + - on ia64 don't abort on unhandled math function exception codes + (#165693) + +* Mon Aug 8 2005 Jakub Jelinek 2.3.90-8 +- update from CVS + - nscd persistent database verifier (#164001) + - cleanup _FORTIFY_SOURCE bits/*.h headers (#165000) + - handle EINTR in sigwait properly +- make sure poor man's stack guard randomization keeps first + byte 0 even on big-endian 32-bit arches +- fix {elf,nptl}/tst-stackguard1 +- obsolete linuxthreads-devel in glibc-devel + +* Fri Jul 29 2005 Jakub Jelinek 2.3.90-7 +- update from CVS +- do some poor man's stack guard randomization even without + the costly --enable-stackguard-randomization +- rebuilt with new GCC to make it use -msecure-plt on PPC32 + +* Mon Jul 25 2005 Jakub Jelinek 2.3.90-6 +- update from CVS + - fix execvp if PATH is not in environment and the call is going + to fail (BZ#1125) + - another bits/wchar2.h fix (#163990) + +* Fri Jul 22 2005 Jakub Jelinek 2.3.90-5 +- update from CVS + - fix stubs.h generation +- don't use _G_va_list in bits/wchar2.h + +* Fri Jul 22 2005 Jakub Jelinek 2.3.90-4 +- update from CVS + - make sure bits/wchar2.h header is installed + - fix __getgroups_chk return type + +* Thu Jul 21 2005 Jakub Jelinek 2.3.90-3 +- update from CVS + - make sure nscd cmsg buffers aren't misaligned, handle EINTR from + poll when contacting nscd more gracefully + - remove malloc attribute from posix_memalign + - correctly size nscd buffer for grpcache key (#163538) + - fix atan2f + - fix error memory leaks + - some more _FORTIFY_SOURCE protection + +* Fri Jul 8 2005 Jakub Jelinek 2.3.90-2 +- update from CVS + - ia64 stack protector support + - handle DNS referral results as server errors (#162625) + - ctan{,h}{,f,l} fixes (#160759) + - pass argc, argv and envp also to executable's *ni_array + functions (BZ#974) + - add ellipsis to clone prototype (#161593) + - fix glibc-profile (#162601) + - nss_compat fixes +- use sysdeps/generic version of in installed + headers instead of NPTL version (#162634) + +* Mon Jun 27 2005 Jakub Jelinek 2.3.90-1 +- update from CVS + - stack protector support + - fix xdr_{,u_}{longlong_t,hyper} on 64-bit arches (#161583) +- enable @GLIBC_2.4 symbols +- remove linuxthreads + +* Mon Jun 20 2005 Jakub Jelinek 2.3.5-11 +- update from CVS + - PPC32 -msecure-plt support + - support classes keyword in /etc/hesiod.conf (#150350) + - add RLIMIT_NICE and RLIMIT_RTPRIO to (#157049) + - decrease number of .plt relocations in libc.so + - use -laudit in nscd (#159217) + - handle big amounts of networking interfaces in getifaddrs/if_nameindex + (#159399) + - fix pa_IN locale's am_pm (#158715, BZ#622) + - fix debugging of PIEs + +* Mon May 30 2005 Jakub Jelinek 2.3.5-10 +- fix LD_ASSUME_KERNEL (since 2.3.5-8 GLRO(dl_osversion) + has been always overwritten with the version of currently + running kernel) +- remove linuxthreads man pages other than those covered in + 3p section, as 3p man pages are far better quality and describe + POSIX behaviour that NPTL implements (#159084) + +* Tue May 24 2005 Jakub Jelinek 2.3.5-9 +- update from CVS + - increase bindresvport's LOWPORT to 512, apparently some + broken daemons don't think 0 .. 511 ports are reserved + +* Mon May 23 2005 Jakub Jelinek 2.3.5-8 +- update from CVS + - fix kernel version check in ld.so +- fix sendfile{,64} prototypes (BZ#961) +- try more ports in bindresvport if all 600..1023 are + used, don't use priviledged ports when talking to portmap + (#141773) + +* Fri May 20 2005 Jakub Jelinek 2.3.5-7 +- update from CVS + - make regexec thread safe (BZ#934) +- fix statically linked programs on i?86, x86_64, s390* and + sparc* (#158027) +- fix IBM939 iconv module (BZ#955) + +* Wed May 4 2005 Jakub Jelinek 2.3.5-6 +- update from CVS + - fix cancellation on i?86 + - add call frame information to i?86 assembly + +* Tue May 3 2005 Jakub Jelinek 2.3.5-5 +- update from CVS + - add some more UTF-8 locales (#156115) +- clean up /lib64/tls instead of /lib/tls on x86-64, s390x and + ppc64 in glibc_post_upgrade (#156656) +- fix posix_fallocate{,64} (#156289) + +* Thu Apr 28 2005 Jakub Jelinek 2.3.5-4 +- update from CVS + - fix nscd cache pruning (#150748) + +* Wed Apr 27 2005 Jakub Jelinek 2.3.5-3 +- update from CVS + - fix linuxthreads clocks +- put xen libs into the glibc-2*.i686 package instead of a separate one +- fix librt.so symlink in linuxthreads-devel +- do not include linuxthreads-devel on %%{auxarches}, + just on the base architectures + +* Wed Apr 27 2005 Jakub Jelinek 2.3.5-2 +- update from CVS + - with MALLOC_CHECK_=N N>0 (#153003) + - fix recursive dlclose (#154641) + - handle %%z in strptime (#154804) + - automatically append /%%{_lib}/obsolete/linuxthreads/ + to standard library search path if LD_ASSUME_KERNEL=N N <= 2.4.19 + or for glibc 2.0 binaries (or broken ones that don't use errno/h_errno + properly). Warning: all those will stop working when LinuxThreads + is finally nuked, which is not very far away + - remove nonnull attribute from acct prototype (BZ#877) + - kernel CPU clocks support + - fix *scanf in locales with multi-byte decimal point + +* Wed Apr 27 2005 Roland McGrath +- glibc-xen subpackage for i686 + +* Fri Apr 15 2005 Roland McGrath 2.3.5-1 +- update from CVS + - fix execvp regression (BZ#851) + - ia64 libm updates + - sparc updates + - fix initstate{,_r}/strfry (#154504) + - grok PT_NOTE in vDSO for kernel version and extra hwcap dirs, + support "hwcap" keyword in ld.so.conf files + +* Tue Apr 5 2005 Jakub Jelinek 2.3.4-21 +- update from CVS + - fix xdr_rmtcall_args on 64-bit arches (#151686) +- fix and with -std=c89 -fexceptions (#153774) + +* Mon Apr 4 2005 Jakub Jelinek 2.3.4-20 +- move LinuxThreads libraries to /%%{_lib}/obsolete/linuxthreads/ + and NPTL libraries to /%%{_lib}. To run a program against LinuxThreads, + LD_ASSUME_KERNEL=2.4.xx LD_LIBRARY_PATH=/%%{_lib}/obsolete/linuxthreads/ + is now needed +- bzip2 ChangeLog* files instead of gzipping them + +* Sat Apr 2 2005 Jakub Jelinek 2.3.4-19 +- update from CVS + - fix nextafterl and several other libm routines on ia64 + - fix initgroups (BZ#661) +- kill nptl-devel subpackage, add linuxthreads-devel, + compile and link by default against NPTL and only with + -I/usr/include/linuxthreads -L/usr/%%{_lib}/linuxthreads + against LinuxThreads +- package /usr/lib/debug/%%{_lib}/tls/i{5,6}86 symlinks in + i386 glibc-debuginfo +- limit number of ChangeLog* files in glibc-common %%doc + to last 2.5 years of changes only to save space + +* Fri Mar 25 2005 Jakub Jelinek 2.3.4-18 +- fix build on 64-bit arches with new GCC + +* Thu Mar 24 2005 Jakub Jelinek 2.3.4-17 +- update from CVS + - fix LD_AUDIT in LinuxThreads ld.so + - fix calloc with M_PERTURB + - fix error handling in pthread_create with PTHREAD_EXPLICIT_SCHED + on ppc*/ia64/alpha/mips (BZ#801) + - fix a typo in WINDOWS-31J charmap (#151739) + - fix NIS ypprot_err (#151469) + +* Sun Mar 20 2005 Jakub Jelinek 2.3.4-16 +- fix pread with -D_FILE_OFFSET_BITS=64 (#151573) + +* Sat Mar 19 2005 Jakub Jelinek 2.3.4-15 +- update from CVS + - better fix for the dlclose bug (#145810, #150414) + - fix regex crash on case insensitive search in zh_CN locale + (#151215) + - fix malloc_trim (BZ#779) + - with -D_FORTIFY_SOURCE=*, avoid defining read and a bunch of others + as function-like macros, there are too many broken programs + out there +- add %%dir %%{_prefix}/%%{_lib}/gconv to glibc's file list (#151372) + +* Sun Mar 6 2005 Roland McGrath 2.3.4-14 +- fix bits/socket2.h macro typos + +* Sat Mar 5 2005 Jakub Jelinek 2.3.4-12 +- fix tst-chk{2,3} +- fix up AS_NEEDED directive in /usr/%%{_lib}/libc.so +- BuildReq binutils >= 2.15.94.0.2-1 for AS_NEEDED, in + glibc-devel Conflict with binutils < 2.15.94.0.2-1 + +* Thu Mar 3 2005 Jakub Jelinek 2.3.4-11 +- update from CVS + - fix execvp (#149290) + - fix dlclose (#145810) + - clear padding in gconv-modules.cache (#146614, BZ#776) +- rebuilt with GCC4 +- changed __GLIBC_MINOR__ for now back to 3 +- back out the newly added GLIBC_2.4 *_chk routines, instead + do the checking in macros + +* Sat Feb 12 2005 Jakub Jelinek 2.3.4-10 +- hopefully fix interaction with prelink (#147655) + +* Fri Feb 11 2005 Jakub Jelinek 2.3.4-9 +- update from CVS + - bi-arch (BZ#715) + +* Fri Feb 11 2005 Jakub Jelinek 2.3.4-8 +- update from CVS + - bi-arch (BZ#632) + - fix libdl on s390 and maybe other platforms + - fix initstate{,_r} (BZ#710) + - fix generation (BZ#157) +- define CMSPAR in bits/termios.h (#147533) + +* Tue Feb 8 2005 Jakub Jelinek 2.3.4-7 +- update from CVS + - fix TLS handling in linuxthreads + +* Tue Feb 8 2005 Jakub Jelinek 2.3.4-6 +- update from CVS + - ld.so auditing + - fix segfault if chrooted app attempts to dlopen a library + and no standard library directory exists at all (#147067, #144303) + - fix initgroups when nscd is running, but has group caching disabled + (#146588) + - fix pthread_key_{create,destroy} in LinuxThreads when pthread_create + has not been called yet (#146710) + - fix ppc64 swapcontext and setcontext (#146736, BZ#700) + - service nscd cosmetic fixes (#146776) + - fix IA-32 and x86-64 stack alignment in DSO constructors (#145689) + - fix zdump -v segfaults on x86-64 (#146210) + - avoid calling sigaction (SIGPIPE, ...) inside syslog (#146021, IT#56686) + - fix errno values for futimes (BZ#633) + - unconditionally include in malloc.h (BZ#650) + - change regex \B handling to match old GNU regex as well as perl/grep's dfa + (from empty string inside of word to empty string not at a word boundary, + BZ#693) + - slightly optimize i686 TLS accesses, use direct TLS %%gs access in sem_* + and allow building -mno-tls-direct-seg-refs glibc that is free of direct TLS + %%gs access with negative offsets + - fix addseverity + - fix fmemopen + - fix rewinddir + - increase svc{tcp,unix}_create listen backlog + +* Thu Jan 6 2005 Jakub Jelinek 2.3.4-5 +- update from CVS + - add some warn_unused_result marking + - make ftruncate available even for just -D_POSIX_C_SOURCE=200112L + (BZ#640) + +* Thu Jan 6 2005 Jakub Jelinek 2.3.4-4 +- update from CVS + - fix IA-32 stack alignment for LinuxThreads thread functions + and functions passed to clone(2) directly + - fix ecvt{,_r} on denormals (#143279) + - fix __tls_get_addr typo + - fix rounding in IA-64 alarm (#143710) + - don't reinitialize __environ in __libc_start_main, so that + effects of setenv/putenv done in DSO initializers are preserved + (#144037, IT#57403) + - fix fmemopen + - fix vDSO l_map_end and l_text_end values + - IA64 libm update (#142494) +- fix ppc rint/ceil etc. (BZ#602) + +* Tue Dec 21 2004 Jakub Jelinek 2.3.4-3 +- rebuilt + +* Mon Dec 20 2004 Jakub Jelinek 2.3.4-2 +- work around rpm bug some more, this time by copying + iconvconfig to iconvconfig.%%{_target_cpu}. + +* Mon Dec 20 2004 Jakub Jelinek 2.3.4-1 +- update from CVS + - glibc 2.3.4 release + - add -o and --nostdlib options to iconvconfig +- if /sbin/ldconfig doesn't exist when running + glibc_post_upgrade.%%{_target_cpu}, just don't attempt to run it. + This can happen during first install of bi-arch glibc and the + other arch glibc's %%post wil run /sbin/ldconfig (#143326) +- use -o and --nostdlib options to create all needed + gconv-modules.cache files on bi-arch setups + +* Sun Dec 19 2004 Jakub Jelinek 2.3.3-99 +- rebuilt + +* Sat Dec 18 2004 Jakub Jelinek 2.3.3-98 +- add .%%{_target_cpu} to glibc_post_upgrade, only run telinit u + if /sbin/init is the same ELF class and machine as + glibc_post_upgrade.%%{_target_cpu} and similarly with + condrestarting sshd (#143046) + +* Fri Dec 17 2004 Jakub Jelinek 2.3.3-97 +- update from CVS + - fix ppc64 getcontext and swapcontext (BZ#610) + - sparc/sparc64 fixes + +* Wed Dec 15 2004 Jakub Jelinek 2.3.3-96 +- update from CVS + - fix i686 __USE_STRING_INLINES strncat + - make sure ppc/ppc64 maintain correct stack alignment + across clone + +* Wed Dec 15 2004 Jakub Jelinek 2.3.3-95 +- export nis_domain_of_r from libnsl.so again which was + unintentionally lost + +* Wed Dec 15 2004 Jakub Jelinek 2.3.3-93 +- update from CVS + - ppc/ppc64 clone without CLONE_THREAD getpid () adjustement + - fix MALLOC_CHECK_={1,2,3} for non-contiguous main arena + (BZ#457) + - fix sysconf (_POSIX_V6_*) for other ABI environments in + bi-arch setups +- s390/s390x clone without CLONE_THREAD getpid () adjustement + +* Tue Dec 14 2004 Jakub Jelinek 2.3.3-92 +- update from CVS +- fix %%{_prefix}/libexec/getconf filenames generation + +* Tue Dec 14 2004 Jakub Jelinek 2.3.3-91 +- update from CVS + - double buffer size in getXXbyYY or getXXent on ERANGE + instead of adding BUFLEN (#142617) + - avoid busy loop in malloc if another thread is doing fork + (#142214) + - some more realloc corruption checks + - fix getconf _POSIX_V6_WIDTH_RESTRICTED_ENVS output, + tweak %%{_prefix}/libexec/getconf/ filenames + +* Fri Dec 10 2004 Jakub Jelinek 2.3.3-90 +- update from CVS + - regex speedups + - use | cat in ldd if running under bash3+ to allow running + it on binaries that are not through SELinux allowed to access + console or tty +- add __NR_waitid defines for alpha and ia64 + +* Wed Dec 8 2004 Jakub Jelinek 2.3.3-89 +- update from CVS + - fix clone2 on ia64 + - avoid tst-timer5 failing with linuxthreads implementation +- if __libc_enable_secure, disallow mode != normal +- change ldd script to imply -r when -u is used, properly + propagate return value and handle suid binaries + +* Tue Dec 7 2004 Jakub Jelinek 2.3.3-88 +- update from CVS + - disregard LD_SHOW_AUXV and LD_DYNAMIC_WEAK if __libc_enable_secure + - disregard LD_DEBUG if __libc_enable_secure in normal mode + if /suid-debug doesn't exist + - fix fseekpos after ungetc + - avoid reading bytes before start of buffers in regex's + check_dst_limits_calc_pos_1 (#142060) + - make getpid () working with clone/clone2 without CLONE_THREAD + (so far on i386/x86_64/ia64 only) +- move %%{_prefix}/libexec/getconf/* to glibc from glibc-common +- make %%{_prefix}/libexec/getconf directory owned by glibc package + +* Fri Dec 3 2004 Jakub Jelinek 2.3.3-87 +- update from CVS + - build libpthread_nonshared.a objects with -fPIC on s390/s390x + - fix mktime with < 0 or > 59 tm_sec on entry + - remove nonnull attribute for realpath + - add $(make-target-directory) for errlist-compat.c rule + (hopefully fix #141404) +- add testcase for ungetc bug +- define _POSIX_{,THREAD_}CPUTIME to 0 on all Linux arches + +* Tue Nov 30 2004 Jakub Jelinek 2.3.3-86 +- update from CVS + - some posix_opt.h fixes +- fix strtold use of unitialized memory (#141000) +- some more bugfixes for bugs detected by valgrind +- rebuilt with GCC >= 3.4.3-5 to avoid packed stack layout + on s390{,x} (#139678) + +* Fri Nov 26 2004 Jakub Jelinek 2.3.3-85 +- update from CVS + - support -v specification in getconf + - fix sysconf (_SC_LFS64_CFLAGS) etc. + - avoid thread stack aliasing issues on EM64T (#140803) +- move %%{_prefix}/include/nptl headers from nptl-devel + to glibc-headers, so that even NPTL specific programs + can be built bi-arch without problems + +* Wed Nov 24 2004 Jakub Jelinek 2.3.3-84 +- update from CVS + - fix memory leak in getaddrinfo if using nscd (#139559) + - handle large lines in /etc/hosts and /etc/networks + (#140378) + - add nonnull attributes to selected dirent.h and dlfcn.h + functions + +* Sun Nov 21 2004 Jakub Jelinek 2.3.3-83 +- update from CVS + - add deprecated and/or nonnull attribute to some signal.h + functions + - speed up tzset () by only using stat instead of open/fstat + when calling tzset for the second and following time if + /etc/localtime has not changed +- fix tgamma (BZ #552) + +* Sat Nov 20 2004 Jakub Jelinek 2.3.3-82 +- update from CVS + - some malloc () checking + - libpthread.a object dependency cleanups (#115157) + - fix for -std=c89 -pedantic-errors (#140132) + +* Fri Nov 19 2004 Jakub Jelinek 2.3.3-81 +- don't use chunksize in <= 2 * SIZE_SZ free () checks + +* Fri Nov 19 2004 Jakub Jelinek 2.3.3-80 +- update from CVS + - with -D_FORTIFY_SOURCE=2, prevent missing %%N$ formats + - for -D_FORTIFY_SOURCE=2 and %%n in writable format string, + issue special error message instead of using the buffer overflow + detected one + - speedup regex searching with REG_NOSUB, add RE_NO_SUB, + speedup searching with nested subexps (BZ #544) + - block SIGCANCEL in NPTL timer_* helper thread +- further free () checking + +* Tue Nov 16 2004 Jakub Jelinek 2.3.3-79 +- update from CVS +- fix free () checking +- move /etc/default/nss into glibc-common (hopefully fix #132392) + +* Mon Nov 15 2004 Jakub Jelinek 2.3.3-78 +- update from CVS + - fix LD_DEBUG=statistics + - issue error message before aborting in __chk_fail () +- some more free () checking + +* Fri Nov 12 2004 Jakub Jelinek 2.3.3-77 +- update from CVS + - speedup regex on palindromes (BZ #429) + - fix NPTL set{,e,re,res}[ug]id, so that even if making process + less priviledged all threads change their credentials successfully + +* Wed Nov 10 2004 Jakub Jelinek 2.3.3-76 +- update from CVS + - fix regcomp crash (#138439) + - fix ftell{,o,o64} (#137885) + - robustification of nscd to cope with corrupt databases (#137140) + - fix NPTL with pthread_exit immediately after pthread_create (BZ #530) + - some regex optimizations + +* Tue Nov 2 2004 Jakub Jelinek 2.3.3-75 +- update from CVS + - mktime cleanups (BZ #487, #473) + - unique comments in free(3) check error messages +- adjust some x86_64 headers for -m32 (#129712) +- object size checking support even with GCC-3.4.2-RH >= 3.4.2-8 + +* Wed Oct 27 2004 Jakub Jelinek 2.3.3-74 +- fix header +- fix globfree (#137176) +- fix exiting if there are dlmopened libraries in namespaces + other than main one not closed yet +- export again _res_opcodes and __p_{class,type}_syms from + libresolv.so that were lost in -69 + +* Thu Oct 21 2004 Jakub Jelinek 2.3.3-73 +- remove setaltroot and key{_add,_request,ctl} also from Versions +- back out _sys_errlist changes + +* Thu Oct 21 2004 Jakub Jelinek 2.3.3-72 +- back out setaltroot and key{_add,_request,ctl} addition +- fix severe x86-64 symbol versioning regressions that breaks + e.g. java binaries + +* Wed Oct 20 2004 Jakub Jelinek 2.3.3-71 +- update from CVS + - fix minor catchsegv temp file handling vulnerability + (CAN-2004-0968, #136319) + - add 4 new errno codes + - setaltroot, key{_add,_request,ctl} syscalls on some arches + - export _dl_debug_state@GLIBC_PRIVATE from ld.so again for + gdb purpose + - use inet_pton to decide what is address and what is hostname + in getent (#135422) + - change dladdr/dladdr1, so that dli_saddr is the same kind + of value as dlsym/dlvsym return (makes difference on ia64/hppa only) + - fix catchsegv script so that it works with both 32-bit and 64-bit + programs on multi-arch platforms + +* Tue Oct 19 2004 Jakub Jelinek 2.3.3-70 +- update from CVS +- require newer selinux-policy (#135978) +- add %%dir for /var/run/nscd and /var/db/nscd and %%ghost + files in it +- conflict with gcc4 4.0.0-0.6 and earlier (needs __builtin_object_size) + +* Mon Oct 18 2004 Jakub Jelinek 2.3.3-69 +- update from CVS + - object size checking support (-D_FORTIFY_SOURCE={1,2}) + +* Thu Oct 14 2004 Jakub Jelinek 2.3.3-68 +- update from CVS + - support for namespaces in the dynamic linker + - fix dlclose (BZ #77) + - libSegFault.so uses now backtrace() to work on IA-64, x86-64 + and s390 (#130254) + +* Tue Oct 12 2004 Jakub Jelinek 2.3.3-67 +- update from CVS + - use non-blocking sockets in resolver (#135234) + - reset pd->res options on thread exit, so that threads + reusing cached stacks get resolver state properly initialized + (BZ #434) + +* Wed Oct 6 2004 Jakub Jelinek 2.3.3-66 +- update from CVS +- avoid using perl in the spec file, buildrequire sed >= 3.95 + (#127671) +- export TIMEOUTFACTOR=16 +- fix _JMPBUF_CFA_UNWINDS_ADJ on s390{,x} + +* Tue Oct 5 2004 Jakub Jelinek 2.3.3-65 +- update from CVS + - define _POSIX_THREAD_PROCESS_SHARED and _POSIX_CLOCK_SELECTION + to -1 in LinuxThreads + - define _POSIX_CPUTIME and _POSIX_THREAD_CPUTIME to 0 + on i?86/ia64 and make sure sysconf (_SC_{,THREAD_}CPUTIME) + returns correct value +- if _POSIX_CLOCK_SELECTION == -1 in nscd, still try + sysconf (_SC_CLOCK_SELECTION) and if it returns true, + dlopen libpthread.so and dlsym pthread_condattr_setclock +- build nscd with -z relro and -z now + +* Mon Oct 4 2004 Jakub Jelinek 2.3.3-64 +- update from CVS + - stop using __builtin_expect in assert and assert_perror + (#127606) + - try to avoid too much VA fragmentation with malloc + on flexmap layout (#118574) + - nscd robustification + - change valloc to use debugging hooks (#134385) +- make glibc_post_upgrade more verbose on errors (Fergal Daly, + #125700) + +* Fri Oct 1 2004 Jakub Jelinek 2.3.3-63 +- update from CVS + - fix __nscd_getgrouplist + - fix a typo in x86_64 pthread_mutex_timedwait fix + +* Fri Oct 1 2004 Jakub Jelinek 2.3.3-62 +- update from CVS + - fix NPTL pthread_mutex_timedwait on i386/x86_64 (BZ #417) + +* Thu Sep 30 2004 Jakub Jelinek 2.3.3-61 +- update from CVS + - some nscd fixes (#134193) + - cache initgroups in nscd (#132850) + - reread /etc/localtime in tzset () even if just mtime changed + (#133481) + - fix glob (#126460) + - another get_myaddress fix + +* Wed Sep 29 2004 Jakub Jelinek 2.3.3-60 +- update from CVS + - fix get_myaddress (#133982) + - remove nonnull attribute from second utime argument (#133866) + - handle SIGSETXID the same way as SIGCANCEL in + sigaction/pthread_kill/sigwait/sigwaitinfo etc. + - add __extension__ to long long types in NPTL + +* Mon Sep 27 2004 Jakub Jelinek 2.3.3-59 +- update from CVS + - fix BZ #151, #362, #381, #407 + - fdim fix for +inf/+inf (BZ #376) + +* Sun Sep 26 2004 Jakub Jelinek 2.3.3-58 +- update from CVS + - vasprintf fix (BZ #346) + - gettext locking (BZ #322) +- change linuxthreads useldt.h inclusion login again, the last + one failed all linuxthreads FLOATING_STACKS tests + +* Sat Sep 25 2004 Jakub Jelinek 2.3.3-57 +- update from CVS + - fix setuid in LD_ASSUME_KERNEL=2.2.5 libc (#133558) + - fix nis locking (#132204) + - RTLD_DEEPBIND support + - fix pthread_create bugs (BZ #401, #405) + +* Wed Sep 22 2004 Roland McGrath 2.3.3-56 +- migrated CVS to fedora-branch in sources.redhat.com glibc repository + - source tarballs renamed + - redhat/ moved to fedora/, some old cruft removed +- update from trunk + - some __nonnull annotations + +* Wed Sep 22 2004 Jakub Jelinek 2.3.3-55 +- update from CVS + - set{re,e,res}[ug]id now affect the whole process in NPTL + - return EAGAIN instead of ENOMEM when not enough memory + in pthread_create + +* Fri Sep 17 2004 Jakub Jelinek 2.3.3-54 +- update from CVS + - nscd getaddrinfo caching + +* Tue Sep 14 2004 Jakub Jelinek 2.3.3-53 +- restore temporarily old definition of __P()/__PMT() + for third party apps + +* Tue Sep 14 2004 Jakub Jelinek 2.3.3-52 +- update from CVS + - nscd bi-arch fix + - remove all uses of __P()/__PMT() from glibc headers +- update and reenable nscd SELinux patch +- remove libnss1* and libnss*.so.1 compatibility NSS modules + on IA-32, SPARC and Alpha + +* Fri Sep 10 2004 Jakub Jelinek 2.3.3-51 +- update from CVS + - disable one of the malloc double free checks for non-contiguous + arenas where it doesn't have to be true even for non-broken + apps + +* Thu Sep 9 2004 Jakub Jelinek 2.3.3-50 +- update from CVS + - pwd/grp/host loops with nscd speed up by sharing the + nscd cache r/o with applications + - inexpensive double free check in free(3) + - make NPTL pthread.h initializers usable even from C++ + (BZ #375) +- use atomic instructions even in i386 nscd on i486+ CPUs + (conditionally) + +* Fri Sep 3 2004 Jakub Jelinek 2.3.3-49 +- update from CVS +- fix linuxthreads tst-cancel{[45],-static} + +* Fri Sep 3 2004 Jakub Jelinek 2.3.3-48 +- update from CVS + - fix pthread_cond_destroy (BZ #342) + - fix fnmatch without FNM_NOESCAPE (BZ #361) + - fix ppc32 setcontext (BZ #357) +- add NPTL support for i386 glibc (only if run on i486 or higher CPU) +- add __NR_waitid defines for i386, x86_64 and sparc* + +* Tue Aug 31 2004 Jakub Jelinek 2.3.3-47 +- update from CVS + - persistent nscd caching + - ppc64 32-bit atomicity fix + - fix x86-64 nptl-devel headers for -m32 compilation +- %%ghost /etc/ld.so.cache (#130597) +- edit /etc/ld.so.conf in glibc_post_upgrade if + include ld.so.conf.d/*.conf line is missing (#120588) +- ugly hacks for the IA-64 /emul braindamage (#124996, #128267) + +* Sat Aug 21 2004 Jakub Jelinek 2.3.3-46 +- update from CVS + +* Thu Aug 19 2004 Jakub Jelinek 2.3.3-45 +- update from CVS + - fix nss_compat's initgroups handling (#130363) + - fix getaddrinfo ai_canonname setting + +* Thu Aug 19 2004 Jakub Jelinek 2.3.3-44 +- update from CVS + - add ip6-dotint resolv.conf option, make + no-ip6-dotint the default +- BuildPrereq libselinux-devel (#129946) +- on ppc64, build without dot symbols + +* Thu Aug 12 2004 Jakub Jelinek 2.3.3-43 +- update from CVS + - remove debugging printout (#129747) + - make usable in C++ (IT#45148) +- update RLIMIT_* constants in , make + POSIX compliant (#129740) + +* Wed Aug 11 2004 Jakub Jelinek 2.3.3-42 +- fix last tzset () fixes, disable rereading of /etc/localtime + every time for now +- really enable SELinux support for NSCD + +* Wed Aug 11 2004 Jakub Jelinek 2.3.3-41 +- update from CVS + - fread_unlocked/fwrite_unlocked macro fixes (BZ #309, #316) + - tzset () fixes (BZ #154) +- speed up pthread_rwlock_unlock on arches other than i386 and + x86_64 (#129455) +- fix compilation with -ansi (resp. -std=c89 or -std=c99) and + -D_XOPEN_SOURCE=[56]00 but no -D_POSIX_SOURCE* or -D_POSIX_C_SOURCE* + (BZ #284) +- add SELinux support for NSCD + +* Fri Aug 6 2004 Jakub Jelinek 2.3.3-40 +- update from CVS + - change res_init to force all threads to re-initialize + resolver before they use it next time (#125712) + - various getaddrinfo and related fixes (BZ #295, #296) + - fix IBM{932,943} iconv modules (#128674) + - some nscd fixes (e.g. BZ #292) + - RFC 3678 support (Multicast Source Filters) +- handle /lib/i686/librtkaio-* in i386 glibc_post_upgrade + the same as /lib/i686/librt-* + +* Fri Jul 23 2004 Jakub Jelinek 2.3.3-39 +- update from CVS + - conformance related changes in headers +- remove -finline-limit=2000 for GCC 3.4.x+ + +* Thu Jul 22 2004 Jakub Jelinek 2.3.3-38 +- update from CVS + - fix res_init leaks + - fix newlocale races + - fix ppc64 setjmp +- fix strtold (BZ #274) + +* Fri Jul 16 2004 Jakub Jelinek 2.3.3-37 +- update from CVS + - allow pthread_cancel in DSO destructors run at exit time +- fix pow{f,,l} on IA-32 and powl on x86-64 +- allow PIEs on IA-32 to have main in a shared library they depend on + +* Mon Jul 5 2004 Jakub Jelinek 2.3.3-36 +- s390* .plt slot reduction +- fix pthread_rwlock_timedrdlock on x86_64 + +* Wed Jun 30 2004 Jakub Jelinek 2.3.3-35 +- tweak spec file for the libpthread-0.61.so -> libpthread-2.3.3.so + NPTL changes + +* Wed Jun 30 2004 Jakub Jelinek 2.3.3-34 +- update from CVS + - if_nameindex using preferably netlink + - printf_parsemb initialization fix + - NPTL version is now the same as glibc version + +* Mon Jun 28 2004 Jakub Jelinek 2.3.3-33 +- update from CVS + - reread resolv.conf for nscd --invalidate=hosts + - fix F_GETLK/F_SETLK/F_SETLKW constants on x86_64 for + -m32 -D_FILE_OFFSET_BITS=64 compilations + - avoid calling non-existing fcntl64 syscall on ppc64 + +* Mon Jun 14 2004 Jakub Jelinek 2.3.3-32 +- update from CVS + - FUTEX_CMP_REQUEUE support (fix pthread_cond_* deadlocks) + - fix backtrace in statically linked programs +- rebuilt with GCC 3.4, adjusted ulps and i386 + +* Fri May 28 2004 Jakub Jelinek 2.3.3-31 +- update from CVS +- and changes for GCC 3.{2,4,5}+ +- make c_stubs buildable even with GCC 3.2.x (#123042) + +* Fri May 21 2004 Jakub Jelinek 2.3.3-30 +- fix pthread_cond_wait on architectures other than IA-32 and + x86_64 + +* Thu May 20 2004 Jakub Jelinek 2.3.3-29 +- use lib64 instead of lib on ia64 if %%{_lib} is defined to lib64 + +* Wed May 19 2004 Jakub Jelinek 2.3.3-28 +- update from CVS + - FUTEX_REQUEUE fixes (#115349) + - SPARC GCC 3.4 build fix + - fix handling of undefined TLS symbols on IA32 (RELA only), + SPARC and SH + - regex translate fix + - speed up sprintf + - x86_64 makecontext alignment fix + - make POSIX sigpause the default sigpause, unless BSD sigpause + requested + +* Tue May 11 2004 Jakub Jelinek 2.3.3-27 +- remove /lib64/tls/librtkaio-2.3.[23].so in glibc_post_upgrade + on x86-64, s390x and ppc64 instead of /lib/tls/librtkaio-2.3.[23].so +- build mq_{send,receive} with -fexceptions + +* Fri May 7 2004 Jakub Jelinek 2.3.3-26 +- update from CVS + - fix + - fix memory leaks in nis, getifaddrs, etc. caused by incorrect + use of realloc +- remove /lib/{tls,i686}/librtkaio-2.3.[23].so in glibc_post_upgrade + and rerun ldconfig if needed, otherwise after glibc upgrade librt.so.1 + might be a stale symlink + +* Wed May 5 2004 Jakub Jelinek 2.3.3-25 +- update from CVS +- disable FUTEX_REQUEUE (work around #115349) +- mq for sparc/sparc64/ia64 + +* Tue May 4 2004 Jakub Jelinek 2.3.3-24 +- update from CVS + - define S_ISSOCK in -D_XOPEN_SOURCE=600 and S_I[FS]SOCK + plus F_[SG]ETOWN also in -D_XOPEN_SOURCE=500 (both + included already in XNS5) + - reorder dlopen checks, so that dlopening ET_REL objects + complains about != ET_DYN != ET_EXEC, not about phentsize + (#121606) + - fix strpbrk macro for GCC 3.4+ (BZ #130) + - fix (BZ #140) + - sched_[gs]etaffinity documentation fix (BZ #131) + - fix sparc64 build (BZ #139) + - change linuxthreads back to use non-cancellable writes + to manager pipes etc. + - fix sem_timedwait return value in linuxthreads (BZ #133) + - ia64 unnecessary PLT relocs removal + +* Thu Apr 22 2004 Jakub Jelinek 2.3.3-23 +- update from CVS + - fix *scanf + - fix shm_unlink, sem_unlink and mq_unlink errno values + - avoid memory leaks in error + - execstack fixes on s390 + +* Mon Apr 19 2004 Jakub Jelinek 2.3.3-22 +- update from CVS + - mq and timer fixes +- rebuilt with binutils >= 2.15.90.0.3-2 to fix IA-64 statically + linked binaries +- fix linuxthreads librt.so on s390{,x}, so it is no longer DT_TEXTREL + +* Sat Apr 17 2004 Jakub Jelinek 2.3.3-21 +- disable rtkaio +- update from CVS + - POSIX message passing support + - fixed SIGEV_THREAD support for POSIX timers + - fix free on non-malloced memory in syslog + - fix ffsl on some 64-bit arches + - fix sched_setaffinity on x86-64, ia64 + - fix ppc64 umount + - NETID_AUTHORITATIVE, SERVICES_AUTHORITATIVE support + - various NIS speedups + - fix fwrite with > 2GB sizes on 64-bit arches + - fix pthread_getattr_np guardsize reporting in NPTL +- report PLT relocations in ld.so and libc.so during the build + +* Thu Mar 25 2004 Jakub Jelinek 2.3.3-20 +- update from CVS + - change NPTL PTHREAD_MUTEX_ADAPTIVE_NP mutexes to spin on SMP + - strtol speed optimization + - don't try to use certainly unimplemented syscalls on ppc64 +- kill -debug subpackage, move the libs to glibc-debuginfo{,-common} + into /usr/lib/debug/usr/%%{_lib}/ directory +- fix c_stubs with gcc 3.4 +- move all the up to 3 builds into %%build scriptlet and + leave only installation in the %%install scriptlet + +* Mon Mar 22 2004 Jakub Jelinek 2.3.3-19 +- update from CVS + - affinity API changes + +* Thu Mar 18 2004 Jakub Jelinek 2.3.3-18 +- update from CVS + - fix ia64 iopl (#118591) + - add support for /etc/ld.so.conf.d/*.conf + - fix x86-64 LD_DEBUG=statistics +- fix hwcap handling when using ld.so.cache (#118518) + +* Mon Mar 15 2004 Jakub Jelinek 2.3.3-17 +- update from CVS + - implement non-_l function on top of _l functions + +* Thu Mar 11 2004 Jakub Jelinek 2.3.3-16 +- update from CVS +- fix s390{,x} TLS handling + +* Wed Mar 10 2004 Jakub Jelinek 2.3.3-15 +- update from CVS + - special section for compatibility code + - make getpid () work even in vfork () child +- configure with --enable-bind-now to avoid lazy binding in ld.so + and libc.so + +* Fri Mar 5 2004 Jakub Jelinek 2.3.3-14 +- update from CVS + - fix iconv -c (#117021) + - fix PIEs on sparc/sparc64 + - fix posix_fadvise on 64-bit architectures +- add locale-archive as %%ghost file (#117014) + +* Mon Mar 1 2004 Jakub Jelinek 2.3.3-13 +- update from CVS + +* Fri Feb 27 2004 Jakub Jelinek 2.3.3-12 +- update from CVS + +* Fri Feb 27 2004 Jakub Jelinek 2.3.3-11 +- update from CVS + - fix ld.so when vDSO is randomized + +* Fri Feb 20 2004 Jakub Jelinek 2.3.3-10 +- update from CVS + +* Fri Feb 20 2004 Jakub Jelinek 2.3.3-9 +- update from CVS + +* Tue Feb 10 2004 Jakub Jelinek 2.3.3-8 +- update from CVS + +* Tue Jan 27 2004 Jakub Jelinek 2.3.3-7 +- update from CVS + - dl_iterate_phdr extension to signal number of added/removed + libraries +- fix PT_GNU_RELRO support on ppc* with prelinking + +* Fri Jan 23 2004 Jakub Jelinek 2.3.3-6 +- rebuilt with fixed GCC on IA-64 + +* Thu Jan 22 2004 Jakub Jelinek 2.3.3-5 +- fix PT_GNU_RELRO support + +* Wed Jan 21 2004 Jakub Jelinek 2.3.3-4 +- update from CVS + - some further regex speedups + - fix re.translate handling in regex (#112869) + - change regfree to match old regex behaviour (what is freed + and clearing of freed pointers) + - fix accesses to unitialized memory in regex (#113507, #113425, + #113421) + - PT_GNU_RELRO support + +* Tue Dec 30 2003 Jakub Jelinek 2.3.3-3 +- update from CVS + - fix pmap_set fd and memory leak (#112726) +- fix backreference handling in regex +- rebuilt under glibc without the above bug to fix + libc.so linker script (#112738) + +* Mon Dec 29 2003 Jakub Jelinek 2.3.3-2 +- update from CVS + - faster getpid () in NPTL builds + - fix to make pthread_setcancelstate (PTHREAD_CANCEL_DISABLE, ) + really disable cancellation (#112512) + - more regex fixes and speedups + - fix nextafter*/nexttoward* + - handle 6th syscall(3) argument on AMD64 + - handle memalign/posix_memalign in mtrace + - fix linuxthreads memory leak (#112208) + - remove throw () from cancellation points in linuxthreads (#112602) + - fix NPTL unregister_atfork + - fix unwinding through alternate signal stacks + +* Mon Dec 1 2003 Jakub Jelinek 2.3.3-1 +- update from CVS + - 2.3.3 release + - lots of regex fixes and speedups (#110401) + - fix atan2 + - fix pshared condvars in NPTL + - fix pthread_attr_destroy for attributes created with + pthread_attr_init@GLIBC_2.0 +- for the time being, include both nb_NO* and no_NO* as locales + so that the distribution can catch up with the no_NO->nb_NO + transition +- add BuildPrereq texinfo (#110252) + +* Tue Nov 18 2003 Jakub Jelinek 2.3.2-102 +- update from CVS + - fix getifaddrs (CAN-2003-0859) + - fix ftw fd leak + - fix linuxthreads sigaction (#108634) + - fix glibc 2.0 stdio compatibility + - fix uselocale (LC_GLOBAL_LOCALE) + - speed up stdio locking in non-threaded programs on IA-32 + - try to maintain correct order of cleanups between those + registered with __attribute__((cleanup)) + and with LinuxThreads style pthread_cleanup_push/pop (#108631) + - fix segfault in regex (#109606) + - fix RE_ICASE multi-byte handling in regex + - fix pthread_exit in libpthread.a (#109790) + - FTW_ACTIONRETVAL support + - lots of regex fixes and speedups + - fix ceill/floorl on AMD64 + +* Mon Oct 27 2003 Jakub Jelinek 2.3.2-101 +- update from CVS + - fix ld.so --verify (and ldd) + +* Mon Oct 27 2003 Jakub Jelinek 2.3.2-100 +- update from CVS + - fix sprof (#103727) + - avoid infinite loops in {,f}statvfs{,64} with hosed mounts file + - prevent dlopening of executables + - fix glob with GLOB_BRACE and without GLOB_NOESCAPE + - fix locale printing of word values on 64-bit big-endian arches + (#107846) + - fix getnameinfo and getaddrinfo with reverse IPv6 lookups + (#101261) + +* Wed Oct 22 2003 Jakub Jelinek 2.3.2-99 +- update from CVS + - dl_iterate_phdr in libc.a on arches other than IA-64 + - LD_DEBUG=statistics prints number of relative relocations + - fix hwcap computation +- NPTL is now part of upstream glibc CVS +- include {st,xh,zu}_ZA{,.UTF-8} locales + +* Sat Oct 4 2003 Jakub Jelinek 2.3.2-98 +- update from CVS + - fix close, pause and fsync (#105348) + - fix pthread_once on IA-32 +- implement backtrace () on IA-64, handle -fomit-frame-pointer + in AMD64 backtrace () (#90402) + +* Tue Sep 30 2003 Jakub Jelinek 2.3.2-97 +- update from CVS + - fix with C++ or -ansi or -pedantic C + - fix mknod/ustat return value when given bogus device number (#105768) + +* Fri Sep 26 2003 Jakub Jelinek 2.3.2-96 +- rebuilt + +* Fri Sep 26 2003 Jakub Jelinek 2.3.2-95 +- fix IA-64 getcontext + +* Thu Sep 25 2003 Jakub Jelinek 2.3.2-94 +- update from CVS +- fix syslog with non-C non-en_* locales (#61296, #104979) +- filter GLIBC_PRIVATE symbols from glibc provides +- fix NIS+ + +* Thu Sep 25 2003 Jakub Jelinek 2.3.2-93 +- update from CVS +- assume 2.4.21 kernel features on RHEL/ppc*, so that + {make,set,get,swap}context works +- backout execstack support for RHEL +- build rtkaio on amd64 too + +* Wed Sep 24 2003 Jakub Jelinek 2.3.2-92 +- update from CVS + - execstack/noexecstack support + - build nscd as PIE +- move __libc_stack_end back to @GLIBC_2.1 +- build against elfutils >= 0.86 to fix stripping on s390x + +* Mon Sep 22 2003 Jakub Jelinek 2.3.2-91 +- rebuilt + +* Mon Sep 22 2003 Jakub Jelinek 2.3.2-90 +- update from CVS + - NPTL locking change (#102682) +- don't jump around lock on amd64 + +* Thu Sep 18 2003 Jakub Jelinek 2.3.2-89 +- fix open_memstream/syslog (#104661) + +* Thu Sep 18 2003 Jakub Jelinek 2.3.2-88 +- update from CVS + - retrieve affinity in pthread_getattr_np + - fix pthread_attr_[gs]etaffinity_np + - handle hex and octal in wordexp + +* Wed Sep 17 2003 Jakub Jelinek 2.3.2-87 +- update from CVS + - truncate instead of round in utimes when utimes syscall is not available + - don't align stack in every glibc function unnecessarily on IA-32 + - make sure threads have their stack 16 byte aligned on IA-32 + - move sched_[sg]etaffinity to GLIBC_2.3.3 symbol version (#103231) + - fix pthread_getattr_np for the initial thread (#102683) + - avoid linuxthreads signal race (#104368) +- ensure all gzip invocations are done with -n option + +* Fri Sep 12 2003 Jakub Jelinek 2.3.2-86 +- update from CVS +- avoid linking in libgcc_eh.a unnecessarily +- change ssize_t back to long int on s390 -m31, unless + gcc 2.95.x is used + +* Wed Sep 10 2003 Jakub Jelinek 2.3.2-85 +- update from CVS + - fix IA-64 memccpy (#104114) + +* Tue Sep 9 2003 Jakub Jelinek 2.3.2-84 +- update from CVS + - undo broken amd64 signal context changes + +* Tue Sep 9 2003 Jakub Jelinek 2.3.2-83 +- update from CVS +- change *nlink_t, *ssize_t and *intptr_t types on s390 -m31 to + {unsigned,} int +- change *u_quad_t, *quad_t, *qaddr_t, *dev_t, *ino64_t, *loff_t, + *off64_t, *rlim64_t, *blkcnt64_t, *fsblkcnt64_t, *fsfilcnt64_t + on 64-bit arches from {unsigned,} long long int {,*} to + {unsigned,} long int {,*} to restore binary compatibility + for C++ functions using these types as arguments + +* Sun Sep 7 2003 Jakub Jelinek 2.3.2-82 +- rebuilt + +* Sat Sep 6 2003 Jakub Jelinek 2.3.2-81 +- update from CVS + - fix tc[gs]etattr/cf[gs]et[io]speed on ppc (#102732) + - libio fixes + +* Thu Sep 4 2003 Jakub Jelinek 2.3.2-80 +- update from CVS + - fix IA-64 cancellation when mixing __attribute__((cleanup ())) + and old-style pthread_cleanup_push cleanups + +* Tue Sep 2 2003 Jakub Jelinek 2.3.2-79 +- updated from CVS + - lots of cancellation fixes + - fix posix_fadvise* on ppc32 + - TLS layout fix + - optimize stdio cleanups (#103354) + - sparcv9 NPTL + - include sigset, sighold, sigrelse, sigpause and sigignore prototypes + in signal.h even if -D_XOPEN_SOURCE_EXTENDED (#103269) + - fix svc_getreqset on 64-bit big-endian arches + - return ENOSYS in linuxthreads pthread_barrierattr_setpshared for + PTHREAD_PROCESS_SHARED + - add pthread_cond_timedwait stubs to libc.so (#102709) +- split glibc-devel into glibc-devel and glibc-headers to ensure + amd64 /usr/include always wins on amd64/i386 bi-arch installs +- increase PTHREAD_STACK_MIN on alpha, ia64 and sparc* +- get rid of __syscall_* prototypes and stubs in sysdeps/unix/sysv/linux +- run make check also with linuxthreads (on IA-32 non-FLOATING_STACKS) + ld.so and NPTL (on IA-32 also FLOATING_STACKS linuxthreads) libraries + and tests + +* Mon Aug 25 2003 Jakub Jelinek 2.3.2-78 +- include dl-osinfo.h only in glibc-debuginfo-2*.rpm, not + in glibc-debuginfo-common* + +* Mon Aug 25 2003 Jakub Jelinek 2.3.2-77 +- update from CVS + - fix glibc 2.0 libio compatibility (#101385) + - fix ldconfig with /usr/lib/lib*.so symlinks (#102853) + - fix assert.h (#102916, #103017) + - make ld.so.cache identical between IA-32 and AMD64 (#102887) + - fix static linking of large IA-64 binaries (#102586) +- avoid using floating point regs in lazy binding code on ppc64 (#102763) + +* Fri Aug 22 2003 Roland McGrath 2.3.2-76 +- add td_thr_tls_get_addr changes missed in initial nptl_db rewrite + +* Sun Aug 17 2003 Roland McGrath 2.3.2-74 +- nptl_db rewrite not yet in CVS + +* Thu Aug 14 2003 Jakub Jelinek 2.3.2-72 +- update from CVS + - fix rtkaio aio_fsync{,64} + - update rtkaio for !BROKEN_THREAD_SIGNALS + - fix assert macro when used on pointers + +* Wed Aug 13 2003 Jakub Jelinek 2.3.2-71 +- update from CVS + +* Tue Aug 12 2003 Jakub Jelinek 2.3.2-70 +- update from CVS +- disable CLONE_STOPPED for now until it is resolved +- strip crt files +- fix libio on arches with no < GLIBC_2.2 support (#102102, #102105) +- fix glibc-debuginfo to include all nptl and nptl_db sources + +* Thu Aug 7 2003 Jakub Jelinek 2.3.2-69 +- update from CVS + - fix pthread_create@GLIBC_2.0 (#101767) +- __ASSUME_CLONE_STOPPED on all arches but s390* in RHEL + +* Sun Aug 3 2003 Jakub Jelinek 2.3.2-68 +- update from CVS + - only use CLONE_STOPPED if kernel supports it, fix setting of thread + explicit scheduling (#101457) + +* Fri Aug 1 2003 Jakub Jelinek 2.3.2-67 +- update from CVS + - fix utimes and futimes if kernel doesn't support utimes syscall + - fix s390 ssize_t type + - fix dlerror when called before any dlopen/dlsym + - update IA-64 bits/sigcontext.h (#101344) + - various warning fixes + - fix pthread.h comment typos (#101363) + +* Wed Jul 30 2003 Jakub Jelinek 2.3.2-66 +- update from CVS +- fix dlopen of libraries using TLS IE/LE models + +* Tue Jul 29 2003 Jakub Jelinek 2.3.2-65 +- update from CVS + - fix timer_create + - use __extension__ before long long typedefs in (#100718) + +* Mon Jul 28 2003 Jakub Jelinek 2.3.2-64 +- update from CVS + - fix wcpncpy (#99462) + - export _res@GLIBC_2.0 even from NPTL libc.so (__res_state () + unlike __errno_location () or __h_errno_location () was introduced + in glibc 2.2) + - fix zic bug on 64-bit platforms + - some TLS handling fixes + - make ldconfig look into alternate ABI dirs by default (#99402) +- move %%{_datadir}/zoneinfo to tzdata package, so that it can be + errataed separately from glibc +- new add-on - rtkaio +- prereq libgcc, as glibc now relies on libgcc_s.so.1 for pthread_cancel + +* Tue Jul 15 2003 Jakub Jelinek 2.3.2-63 +- fix thread cancellation on ppc64 + +* Sat Jul 12 2003 Jakub Jelinek 2.3.2-62 +- update from CVS + - fix thread cancellation on ppc32, s390 and s390x + +* Thu Jul 10 2003 Jakub Jelinek 2.3.2-61 +- update from CVS + - build libc_nonshared.a with -fPIC instead of -fpic +- fix ppc64 PIE support +- add cfi directives to NPTL sysdep-cancel.h on ppc/ppc64/s390/s390x + +* Tue Jul 8 2003 Jakub Jelinek 2.3.2-60 +- update from CVS + +* Thu Jul 3 2003 Jakub Jelinek 2.3.2-59 +- update from CVS +- on IA-64 use different symbols for cancellation portion of syscall + handlers to make gdb happier + +* Thu Jun 26 2003 Jakub Jelinek 2.3.2-58 +- update from CVS + - nss_compat supporting LDAP etc. + +* Tue Jun 24 2003 Jakub Jelinek 2.3.2-57 +- update from CVS + +* Thu Jun 19 2003 Jakub Jelinek 2.3.2-56 +- fix condvars and semaphores in ppc* NPTL +- fix test-skeleton.c reporting of timed-out tests (#91269) +- increase timeouts for tests during make check + +* Wed Jun 18 2003 Jakub Jelinek 2.3.2-55 +- make ldconfig default to both /lib+/usr/lib and /lib64+/usr/lib64 + on bi-ABI architectures (#97557) +- disable FUTEX_REQUEUE on ppc* temporarily + +* Wed Jun 18 2003 Jakub Jelinek 2.3.2-54 +- update from CVS +- fix glibc_post_upgrade on ppc + +* Tue Jun 17 2003 Jakub Jelinek 2.3.2-53 +- update from CVS +- fix localedef (#90659) +- tweak linuxthreads for librt cancellation + +* Mon Jun 16 2003 Jakub Jelinek 2.3.2-52 +- update from CVS + +* Thu Jun 12 2003 Jakub Jelinek 2.3.2-51 +- update from CVS +- fix (#97169) + +* Wed Jun 11 2003 Jakub Jelinek 2.3.2-50 +- update from CVS + +* Tue Jun 10 2003 Jakub Jelinek 2.3.2-49 +- update from CVS + - fix pthread_cond_signal on IA-32 (#92080, #92253) + - fix setegid (#91567) +- don't prelink -R libc.so on any architecture, it prohibits + address randomization + +* Thu Jun 5 2003 Jakub Jelinek 2.3.2-48 +- update from CVS + - fix IA-64 NPTL build + +* Thu Jun 5 2003 Jakub Jelinek 2.3.2-47 +- update from CVS +- PT_GNU_STACK segment in binaries/executables and .note.GNU-stack + section in *.[oa] + +* Sun Jun 1 2003 Jakub Jelinek 2.3.2-46 +- update from CVS +- enable NPTL on AMD64 +- avoid using trampolines in localedef + +* Thu May 29 2003 Jakub Jelinek 2.3.2-45 +- enable NPTL on IA-64 + +* Thu May 29 2003 Jakub Jelinek 2.3.2-44 +- update from CVS +- enable NPTL on s390 and s390x +- make __init_array_start etc. symbols in elf-init.oS hidden undefined + +* Thu May 29 2003 Jakub Jelinek 2.3.2-43 +- update from CVS + +* Fri May 23 2003 Jakub Jelinek 2.3.2-42 +- update from CVS + +* Tue May 20 2003 Jakub Jelinek 2.3.2-41 +- update from CVS +- use NPTL libs if uname -r contains nptl substring or is >= 2.5.69 + or set_tid_address syscall is available instead of checking + AT_SYSINFO dynamic tag + +* Thu May 15 2003 Jakub Jelinek 2.3.2-40 +- update from CVS + +* Wed May 14 2003 Jakub Jelinek 2.3.2-39 +- update from CVS + - fix for prelinking of libraries with no dependencies + +* Tue May 13 2003 Jakub Jelinek 2.3.2-38 +- update from CVS +- enable NPTL on ppc and ppc64 + +* Tue May 6 2003 Matt Wilson 2.3.2-37 +- rebuild + +* Sun May 4 2003 Jakub Jelinek 2.3.2-36 +- update from CVS + +* Sat May 3 2003 Jakub Jelinek 2.3.2-35 +- update from CVS + - make -jN build fixes + +* Fri May 2 2003 Jakub Jelinek 2.3.2-34 +- update from CVS +- avoid using trampolines in iconvconfig for now + +* Sat Apr 26 2003 Jakub Jelinek 2.3.2-33 +- update from CVS + +* Fri Apr 25 2003 Jakub Jelinek 2.3.2-32 +- update from CVS +- more ppc TLS fixes + +* Wed Apr 23 2003 Jakub Jelinek 2.3.2-31 +- update from CVS + - nscd fixes + - fix Bahrain spelling (#56298) + - fix Ukrainian collation (#83973) + - accept trailing spaces in /etc/ld.so.conf (#86032) + - perror fix (#85994) + - fix localedef (#88978) + - fix getifaddrs (#89026) + - fix strxfrm (#88409) +- fix ppc TLS +- fix getaddrinfo (#89448) +- don't print warning about errno, h_errno or _res if + LD_ASSUME_KERNEL=2.4.1 or earlier + +* Tue Apr 15 2003 Jakub Jelinek 2.3.2-30 +- update from CVS +- fix prelink on ppc32 +- add TLS support on ppc32 and ppc64 +- make sure on -m64 arches all helper binaries are built with this + option + +* Mon Apr 14 2003 Jakub Jelinek 2.3.2-29 +- update from CVS + - fix strxfrm (#88409) +- use -m64 -mno-minimal-toc on ppc64 +- conflict with kernels < 2.4.20 on ppc64 and < 2.4.0 on x86_64 +- link glibc_post_upgrade against newly built libc.a + +* Sun Apr 13 2003 Jakub Jelinek 2.3.2-28 +- update from CVS + - fix NPTL pthread_detach and already terminated, but not yet + joined thread (#88219) + - fix bug-regex4 testcase (#88118) + - reenable prelink support broken in 2.3.2-13 + - fix register_printf_function (#88052) + - fix double free with fopen using ccs= (#88056) + - fix potential access below $esp in {set,swap}context (#88093) + - fix buffer underrun in gencat -H (#88099) + - avoid using unitialized variable in tst-tgmath (#88101) + - fix gammal (#88104) + - fix iconv -c + - fix xdr_string (PR libc/4999) + - fix /usr/lib/nptl/librt.so symlink + - avoid running NPTL cleanups twice in some cases + - unblock __pthread_signal_cancel in linuxthreads, so that + linuxthreads threaded programs work correctly if spawned + from NPTL threaded programs + - fix sysconf _SC_{NPROCESSORS_{CONF,ONLN},{,AV}PHYS_PAGES} +- remove /lib/i686 directory before running ldconfig in glibc post + during i686 -> i386 glibc "upgrades" (#88456) + +* Wed Apr 2 2003 Jakub Jelinek 2.3.2-22 +- update from CVS + - add pthread_atfork to libpthread.a + +* Tue Apr 1 2003 Jakub Jelinek 2.3.2-21 +- update from CVS +- make sure linuxthreads pthread_mutex_lock etc. is not a cancellation + point + +* Sat Mar 29 2003 Jakub Jelinek 2.3.2-20 +- update from CVS +- if kernel >= 2.4.1 doesn't support NPTL, fall back to + /lib/i686 libs on i686, not stright to /lib + +* Fri Mar 28 2003 Jakub Jelinek 2.3.2-19 +- update from CVS + - timers fixes + +* Thu Mar 27 2003 Jakub Jelinek 2.3.2-18 +- update from CVS +- fix NPTL pthread_cond_timedwait +- fix sysconf (_SC_MONOTONIC_CLOCK) +- use /%%{_lib}/tls instead of /lib/tls on x86-64 +- add /%%{_lib}/tls/librt*so* and /%%{_lib}/i686/librt*so* +- display content of .out files for all make check failures + +* Wed Mar 26 2003 Jakub Jelinek 2.3.2-17 +- update from CVS + - kernel POSIX timers support + +* Sat Mar 22 2003 Jakub Jelinek 2.3.2-16 +- update from CVS + - export __fork from glibc again +- fix glibc-compat build in NPTL +- fix c_stubs +- fix some more atomic.h problems +- don't check abi in glibc-compat libs + +* Fri Mar 21 2003 Jakub Jelinek 2.3.2-15 +- update from CVS +- build glibc-compat (for glibc 2.0 compatibility) and c_stubs add-ons +- condrestart sshd in glibc_post_upgrade so that the user can + log in remotely and handle the rest (#86339) +- fix a typo in glibc_post_upgrade on sparc + +* Tue Mar 18 2003 Jakub Jelinek 2.3.2-14 +- update from CVS +- change i686/athlon libc.so.6 base to 0x00e80000 + +* Mon Mar 17 2003 Jakub Jelinek 2.3.2-13 +- update from CVS + - hopefully last fix for condvar problems + +* Fri Mar 14 2003 Jakub Jelinek 2.3.2-12 +- fix bits/syscall.h creation on x86-64 + +* Thu Mar 13 2003 Jakub Jelinek 2.3.2-11 +- update from CVS + +* Wed Mar 12 2003 Jakub Jelinek 2.3.2-10 +- update from CVS + +* Tue Mar 11 2003 Jakub Jelinek 2.3.2-9 +- update from CVS +- fix glibc-debug description (#85111) +- make librt.so a symlink again, not linker script + +* Tue Mar 4 2003 Jakub Jelinek 2.3.2-8 +- update from CVS +- remove the workarounds for broken software accessing GLIBC_PRIVATE + symbols + +* Mon Mar 3 2003 Jakub Jelinek 2.3.2-7 +- update from CVS + +* Sun Mar 2 2003 Jakub Jelinek 2.3.2-6 +- fix TLS IE/LE model handling in dlopened libraries + on TCB_AT_TP arches + +* Tue Feb 25 2003 Jakub Jelinek 2.3.2-5 +- update from CVS + +* Tue Feb 25 2003 Jakub Jelinek 2.3.2-4 +- update from CVS + +* Mon Feb 24 2003 Jakub Jelinek 2.3.2-3 +- update from CVS +- only warn about errno, h_errno or _res for binaries, never + libraries +- rebuilt with gcc-3.2.2-4 to use direct %%gs TLS access insn sequences + +* Sun Feb 23 2003 Jakub Jelinek 2.3.2-2 +- update from CVS + +* Sat Feb 22 2003 Jakub Jelinek 2.3.2-1 +- update from CVS + +* Thu Feb 20 2003 Jakub Jelinek 2.3.1-51 +- update from CVS + +* Wed Feb 19 2003 Jakub Jelinek 2.3.1-50 +- update from CVS + +* Wed Feb 19 2003 Jakub Jelinek 2.3.1-49 +- update from CVS +- remove nisplus and nis from the default nsswitch.conf (#67401, #9952) + +* Tue Feb 18 2003 Jakub Jelinek 2.3.1-48 +- update from CVS + +* Sat Feb 15 2003 Jakub Jelinek 2.3.1-47 +- update from CVS + +* Fri Feb 14 2003 Jakub Jelinek 2.3.1-46 +- update from CVS + - pthread_cond* NPTL fixes, new NPTL testcases + +* Thu Feb 13 2003 Jakub Jelinek 2.3.1-45 +- update from CVS +- include also linuxthreads FLOATING_STACKS libs on i686 and athlon: + LD_ASSUME_KERNEL=2.2.5 to LD_ASSUME_KERNEL=2.4.0 is non-FLOATING_STACKS lt, + LD_ASSUME_KERNEL=2.4.1 to LD_ASSUME_KERNEL=2.4.19 is FLOATING_STACKS lt, + later is NPTL +- enable TLS on alpha/alphaev6 +- add BuildPreReq: /usr/bin/readlink + +* Tue Feb 11 2003 Jakub Jelinek 2.3.1-44 +- update from CVS + - pthread_once fix + +* Mon Feb 10 2003 Jakub Jelinek 2.3.1-43 +- update from CVS +- vfork fix on s390 +- rebuilt with binutils 2.13.90.0.18-5 so that accesses to errno + don't bind locally (#83325) + +* Thu Feb 06 2003 Jakub Jelinek 2.3.1-42 +- update from CVS +- fix pthread_create after vfork+exec in linuxthreads + +* Wed Feb 05 2003 Jakub Jelinek 2.3.1-41 +- update from CVS + +* Thu Jan 30 2003 Jakub Jelinek 2.3.1-40 +- update from CVS + +* Wed Jan 29 2003 Jakub Jelinek 2.3.1-39 +- update from CVS +- enable TLS on s390{,x} and sparc{,v9} + +* Fri Jan 17 2003 Jakub Jelinek 2.3.1-38 +- update from CVS +- initialize __environ in glibc_post_upgrade to empty array, + so that it is not NULL +- compat symlink for s390x /lib/ld64.so.1 +- enable glibc-profile on x86-64 +- only include libNoVersion.so on IA-32, Alpha and Sparc 32-bit + +* Thu Jan 16 2003 Jakub Jelinek 2.3.1-37 +- update from CVS + - nscd fixes, *scanf fix +- fix %%nptlarches noarch build (#81909) +- IA-64 TLS fixes + +* Tue Jan 14 2003 Jakub Jelinek 2.3.1-36 +- update from CVS +- rework -debuginfo subpackage, add -debuginfo-common + subpackage on IA-32, Alpha and Sparc (ie. auxiliary arches) +- fix vfork in libc.a on PPC32, Alpha, Sparc +- fix libio locks in linuxthreads libc.so if libpthread.so + is dlopened later (#81374) + +* Mon Jan 13 2003 Jakub Jelinek 2.3.1-35 +- update from CVS + - dlclose bugfixes +- fix NPTL libpthread.a +- fix glibc_post_upgrade on several arches + +* Sat Jan 11 2003 Jakub Jelinek 2.3.1-34 +- update from CVS +- TLS support on IA-64 + +* Wed Jan 8 2003 Jakub Jelinek 2.3.1-33 +- fix vfork in linuxthreads (#81377, #81363) + +* Tue Jan 7 2003 Jakub Jelinek 2.3.1-32 +- update from CVS +- don't use TLS libs if kernel doesn't set AT_SYSINFO + (#80921, #81212) +- add ntp_adjtime on alpha (#79996) +- fix nptl_db (#81116) + +* Sun Jan 5 2003 Jakub Jelinek 2.3.1-31 +- update from CVS +- support all architectures again + +* Fri Jan 3 2003 Jakub Jelinek 2.3.1-30 +- fix condvar compatibility wrappers +- add ugly hack to use non-TLS libs if a binary is seen + to have errno, h_errno or _res symbols in .dynsym + +* Fri Jan 3 2003 Jakub Jelinek 2.3.1-29 +- update from CVS + - fixes for new condvar + +* Thu Jan 2 2003 Jakub Jelinek 2.3.1-28 +- new NPTL condvar implementation plus related linuxthreads + symbol versioning updates + +* Thu Jan 2 2003 Jakub Jelinek 2.3.1-27 +- update from CVS +- fix #include with -D_BSD_SOURCE or without + feature set macros +- make *sigaction, sigwait and raise the same between + -lpthread -lc and -lc -lpthread in linuxthreads builds + +* Tue Dec 31 2002 Jakub Jelinek 2.3.1-26 +- fix dlclose + +* Sun Dec 29 2002 Jakub Jelinek 2.3.1-25 +- enable sysenter by default for now +- fix endless loop in ldconfig + +* Sat Dec 28 2002 Jakub Jelinek 2.3.1-24 +- update from CVS + +* Fri Dec 27 2002 Jakub Jelinek 2.3.1-23 +- update from CVS + - fix ptmalloc_init after clearenv (#80370) + +* Sun Dec 22 2002 Jakub Jelinek 2.3.1-22 +- update from CVS +- add IA-64 back +- move TLS libraries from /lib/i686 to /lib/tls + +* Thu Dec 19 2002 Jakub Jelinek 2.3.1-21 +- system(3) fix for linuxthreads +- don't segfault in pthread_attr_init from libc.so +- add cancellation tests from nptl to linuxthreads + +* Wed Dec 18 2002 Jakub Jelinek 2.3.1-20 +- fix up lists of exported symbols + their versions + from the libraries + +* Wed Dec 18 2002 Jakub Jelinek 2.3.1-19 +- fix --with-tls --enable-kernel=2.2.5 libc on IA-32 + +* Wed Dec 18 2002 Jakub Jelinek 2.3.1-18 +- update from CVS + - fix NPTL hanging mozilla + - initialize malloc in mALLOPt (fixes problems with squid, #79957) + - make linuxthreads work with dl_dynamic_weak 0 + - clear dl_dynamic_weak everywhere + +* Tue Dec 17 2002 Jakub Jelinek 2.3.1-17 +- update from CVS + - NPTL socket fixes, flockfile/ftrylockfile/funlockfile fix + - kill -debug sub-package, rename -debug-static to -debug + - clear dl_dynamic_weak for NPTL + +* Mon Dec 16 2002 Jakub Jelinek 2.3.1-16 +- fix and for C++ +- automatically generate NPTL libpthread wrappers + +* Mon Dec 16 2002 Jakub Jelinek 2.3.1-15 +- update from CVS + - all functions which need cancellation should now be cancellable + both in libpthread.so and libc.so + - removed @@GLIBC_2.3.2 cancellation wrappers + +* Fri Dec 13 2002 Jakub Jelinek 2.3.1-14 +- update from CVS + - replace __libc_lock_needed@GOTOFF(%%ebx) with + %%gs:offsetof(tcbhead_t, multiple_threads) + - start of new NPTL cancellation wrappers + +* Thu Dec 12 2002 Jakub Jelinek 2.3.1-13 +- update from CVS +- use inline locks in malloc + +* Tue Dec 10 2002 Jakub Jelinek 2.3.1-12 +- update from CVS + - support LD_ASSUME_KERNEL=2.2.5 in statically linked programs + +* Mon Dec 9 2002 Jakub Jelinek 2.3.1-11 +- update from CVS +- rebuilt with gcc-3.2.1-2 + +* Fri Dec 6 2002 Jakub Jelinek 2.3.1-10 +- update from CVS +- non-nptl --with-tls --without-__thread FLOATING_STACKS libpthread + should work now +- faster libc locking when using nptl +- add OUTPUT_FORMAT to linker scripts +- fix x86_64 sendfile (#79111) + +* Wed Dec 4 2002 Jakub Jelinek 2.3.1-9 +- update from CVS + - RUSCII support (#78906) +- for nptl builds add BuildRequires +- fix byteswap.h for non-gcc (#77689) +- add nptl-devel package + +* Tue Dec 3 2002 Jakub Jelinek 2.3.1-8 +- update from CVS + - make --enable-kernel=2.2.5 --with-tls --without-__thread + ld.so load nptl and other --with-__thread libs +- disable nptl by default for now + +* Wed Nov 27 2002 Jakub Jelinek 2.3.1-7 +- update from CVS +- restructured redhat/Makefile and spec, so that src.rpm contains + glibc-.tar.bz2, glibc-redhat-.tar.bz2 and glibc-redhat.patch +- added nptl + +* Fri Nov 8 2002 Jakub Jelinek 2.3.1-6 +- update from CVS + - even more regex fixes +- run sed testsuite to check glibc regex + +* Thu Oct 24 2002 Jakub Jelinek 2.3.1-5 +- fix LD_DEBUG=statistics and LD_TRACE_PRELINKING in programs + using libpthread.so. + +* Thu Oct 24 2002 Jakub Jelinek 2.3.1-4 +- update from CVS + - fixed %%a and %%A in *printf (#75821) + - fix re_comp memory leaking (#76594) + +* Tue Oct 22 2002 Jakub Jelinek 2.3.1-3 +- update from CVS + - some more regex fixes +- fix libpthread.a (#76484) +- fix locale-archive enlarging + +* Fri Oct 18 2002 Jakub Jelinek 2.3.1-2 +- update from CVS + - don't need to use 128K of stacks for DNS lookups + - regex fixes + - updated timezone data e.g. for this year's Brasil DST + changes + - expand ${LIB} in RPATH/RUNPATH/dlopen filenames + +* Fri Oct 11 2002 Jakub Jelinek 2.3.1-1 +- update to 2.3.1 final + - support really low thread stack sizes (#74073) +- tzdata update + +* Wed Oct 9 2002 Jakub Jelinek 2.3-2 +- update from CVS + - handle low stack limits + - move s390x into */lib64 + +* Thu Oct 3 2002 Jakub Jelinek 2.3-1 +- update to 2.3 final + - fix freopen on libstdc++ <= 2.96 stdin/stdout/stderr (#74800) + +* Sun Sep 29 2002 Jakub Jelinek 2.2.94-3 +- don't prelink -r libc.so on ppc/x86-64/sparc*, it doesn't + speed things up, because they are neither REL arches, nor + ELF_MACHINE_REL_RELATIVE +- fix sparc64 build + +* Sun Sep 29 2002 Jakub Jelinek 2.2.94-2 +- update from CVS + +* Sat Sep 28 2002 Jakub Jelinek 2.2.94-1 +- update from CVS +- prelink on ppc and x86-64 too +- don't remove ppc memset +- instead of listing on which arches to remove glibc-compat + list where it should stay + +* Fri Sep 6 2002 Jakub Jelinek 2.2.93-5 +- fix wcsmbs functions with invalid character sets (or malloc + failures) +- make sure __ctype_b etc. compat vars are updated even if + they are copy relocs in the main program + +* Thu Sep 5 2002 Jakub Jelinek 2.2.93-4 +- fix /lib/libnss1_dns.so.1 (missing __set_h_errno definition + leading to unresolved __set_h_errno symbol) + +* Wed Sep 4 2002 Jakub Jelinek 2.2.93-3 +- security fix - increase dns-network.c MAXPACKET to at least + 65536 to avoid buffer overrun. Likewise glibc-compat + dns-{host,network}.c. + +* Tue Sep 3 2002 Jakub Jelinek 2.2.93-2 +- temporarily add back __ctype_b, __ctype_tolower and __ctype_toupper to + libc.a and export them as @@GLIBC_2.0 symbols, not @GLIBC_2.0 + from libc.so - we have still lots of .a libraries referencing + __ctype_{b,tolower,toupper} out there... + +* Tue Sep 3 2002 Jakub Jelinek 2.2.93-1 +- update from CVS + - 2.2.93 release + - use double instead of single indirection in isXXX macros + - per-locale wcsmbs conversion state + +* Sat Aug 31 2002 Jakub Jelinek 2.2.92-2 +- update from CVS + - fix newlocale/duplocale/uselocale +- disable profile on x86_64 for now + +* Sat Aug 31 2002 Jakub Jelinek 2.2.92-1 +- update from CVS + - 2.2.92 release + - fix gettext after uselocale + - fix locales in statically linked threaded programs + - fix NSS + +* Thu Aug 29 2002 Jakub Jelinek 2.2.91-1 +- update from CVS + - 2.2.91 release + - fix fd leaks in locale-archive reader (#72043) +- handle EROFS in build-locale-archive gracefully (#71665) + +* Wed Aug 28 2002 Jakub Jelinek 2.2.90-27 +- update from CVS + - fix re_match (#72312) +- support more than 1024 threads + +* Fri Aug 23 2002 Jakub Jelinek 2.2.90-26 +- update from CVS + - fix i386 build + +* Thu Aug 22 2002 Jakub Jelinek 2.2.90-25 +- update from CVS + - fix locale-archive loading hang on some (non-primary) locales + (#72122, #71878) + - fix umount problems with locale-archives when /usr is a separate + partition (#72043) +- add LICENSES file + +* Fri Aug 16 2002 Jakub Jelinek 2.2.90-24 +- update from CVS + - only mmap up to 2MB of locale-archive on 32-bit machines + initially + - fix fseek past end + fread segfault with mmaped stdio +- include which is mistakenly not included + in glibc-devel on IA-32 + +* Fri Aug 16 2002 Jakub Jelinek 2.2.90-23 +- don't return normalized locale name in setlocale when using + locale-archive + +* Thu Aug 15 2002 Jakub Jelinek 2.2.90-22 +- update from CVS + - optimize for primary system locale +- localedef fixes (#71552, #67705) + +* Wed Aug 14 2002 Jakub Jelinek 2.2.90-21 +- fix path to locale-archive in libc reader +- build locale archive at glibc-common %%post time +- export __strtold_internal and __wcstold_internal on Alpha again +- workaround some localedata problems + +* Tue Aug 13 2002 Jakub Jelinek 2.2.90-20 +- update from CVS +- patch out set_thread_area for now + +* Fri Aug 9 2002 Jakub Jelinek 2.2.90-19 +- update from CVS +- GB18030 patch from Yu Shao +- applied Debian patch for getaddrinfo IPv4 vs. IPv6 +- fix regcomp (#71039) + +* Sun Aug 4 2002 Jakub Jelinek 2.2.90-18 +- update from CVS +- use /usr/sbin/prelink, not prelink (#70376) + +* Thu Jul 25 2002 Jakub Jelinek 2.2.90-17 +- update from CVS + +* Thu Jul 25 2002 Jakub Jelinek 2.2.90-16 +- update from CVS + - ungetc fix (#69586) + - fseek errno fix (#69589) + - change *etrlimit prototypes for C++ (#68588) +- use --without-tls instead of --disable-tls + +* Thu Jul 11 2002 Jakub Jelinek 2.2.90-15 +- set nscd user's shell to /sbin/nologin (#68369) +- fix glibc-compat buffer overflows (security) +- buildrequire prelink, don't build glibc's own copy of it (#67567) +- update from CVS + - regex fix (#67734) + - fix unused warnings (#67706) + - fix freopen with mmap stdio (#67552) + - fix realloc (#68499) + +* Tue Jun 25 2002 Bill Nottingham 2.2.90-14 +- update from CVS + - fix argp on long words + - update atime in libio + +* Sat Jun 22 2002 Jakub Jelinek 2.2.90-13 +- update from CVS + - a thread race fix + - fix readdir on invalid dirp + +* Wed Jun 19 2002 Jakub Jelinek 2.2.90-12 +- update from CVS + - don't use __thread in headers +- fix system(3) in threaded apps +- update prelink, so that it is possible to prelink -u libc.so.6.1 + on Alpha + +* Fri Jun 7 2002 Jakub Jelinek 2.2.90-11 +- update from CVS + - fix __moddi3 (#65612, #65695) + - fix ether_line (#64427) +- fix setvbuf with mmap stdio (#65864) +- --disable-tls for now, waiting for kernel +- avoid duplication of __divtf3 etc. on IA-64 +- make sure get*ent_r and _IO_wfile_jumps are exported (#62278) + +* Tue May 21 2002 Jakub Jelinek 2.2.90-10 +- update from CVS + - fix Alpha pthread bug with gcc 3.1 + +* Fri Apr 19 2002 Jakub Jelinek 2.2.5-35 +- fix nice + +* Mon Apr 15 2002 Jakub Jelinek 2.2.5-34 +- add relocation dependencies even for weak symbols (#63422) +- stricter check_fds check for suid/sgid binaries +- run make check at %%install time + +* Sat Apr 13 2002 Jakub Jelinek 2.2.5-33 +- handle Dec 31 1969 in mktime for timezones west of GMT (#63369) +- back out do-lookup.h change (#63261, #63305) +- use "memory" clobber instead all the fancy stuff in i386/i686/bits/string.h + since lots of compilers break on it +- fix sparc build with gcc 3.1 +- fix spec file for athlon + +* Tue Apr 9 2002 Jakub Jelinek 2.2.5-32 +- fix debugging of threaded apps (#62804) +- fix DST for Estonia (#61494) +- document that pthread_mutexattr_?etkind_np are deprecated + and pthread_mutexattr_?ettype should be used instead in man + pages (#61485) +- fix libSegFault.so undefined externals + +* Fri Apr 5 2002 Jakub Jelinek 2.2.5-31 +- temporarily disable prelinking ld.so, as some statically linked + binaries linked against debugging versions of old glibcs die on it + (#62352) +- fix for -std=c99 (#62516) +- fix ether_ntohost segfault (#62397) +- remove in glibc_post_upgrade on i386 all /lib/i686/libc-*.so, + /lib/i686/libm-*.so and /lib/i686/libpthread-*.so, not just current + version (#61633) +- prelink -r on alpha too + +* Thu Mar 28 2002 Jakub Jelinek 2.2.5-30 +- update GB18030 iconv module (Yu Shao) + +* Tue Mar 26 2002 Jakub Jelinek 2.2.5-29 +- features.h fix + +* Tue Mar 26 2002 Jakub Jelinek 2.2.5-28 +- update from CVS + - fix nscd with huge groups + - fix nis to not close fds it shouldn't +- rebuilt against newer glibc-kernheaders to use the correct + PATH_MAX +- handle .athlon.rpm glibc the same way as .i686.rpm +- add a couple of .ISO-8859-15 locales (#61908) +- readd temporarily currencies which were superceeded by Euro + into the list of accepted currencies by localedef to make + standard conformance testsuites happy +- temporarily moved __libc_waitpid back to make Sun JDK happy +- use old malloc code +- prelink i686/athlon ld.so and prelink -r i686/athlon libc.so + +* Thu Mar 14 2002 Jakub Jelinek 2.2.5-27 +- update from CVS + - fix DST handling for southern hemisphere (#60747) + - fix daylight setting for tzset (#59951) + - fix ftime (#60350) + - fix nice return value + - fix a malloc segfault +- temporarily moved __libc_wait, __libc_fork and __libc_stack_end + back to what they used to be exported at +- censorship (#60758) + +* Thu Feb 28 2002 Jakub Jelinek 2.2.5-26 +- update from CVS +- use __attribute__((visibility(...))) if supported, use _rtld_local + for ld.so only objects +- provide libc's own __{,u}{div,mod}di3 + +* Wed Feb 27 2002 Jakub Jelinek 2.2.5-25 +- switch back to 2.2.5, mmap stdio needs work + +* Mon Feb 25 2002 Jakub Jelinek 2.2.90-8 +- fix two other mmap stdio bugs (#60228) + +* Thu Feb 21 2002 Jakub Jelinek 2.2.90-7 +- fix yet another mmap stdio bug (#60145) + +* Tue Feb 19 2002 Jakub Jelinek 2.2.90-6 +- fix mmap stdio bug (seen on ld as File truncated error, #60043) +- apply Andreas Schwab's fix for pthread sigwait +- remove /lib/i686/ libraries in glibc_post_upgrade when + performing i386 glibc install + +* Thu Feb 14 2002 Jakub Jelinek 2.2.90-5 +- update to CVS +- added glibc-utils subpackage +- disable autoreq in glibc-debug +- readd %%lang() to locale files + +* Thu Feb 7 2002 Jakub Jelinek 2.2.90-4 +- update to CVS +- move glibc private symbols to GLIBC_PRIVATE symbol version + +* Wed Jan 9 2002 Jakub Jelinek 2.2.90-3 +- fix a sqrt bug on alpha which caused SHN_UNDEF $__full_ieee754_sqrt..ng + symbol in libm + +* Tue Jan 8 2002 Jakub Jelinek 2.2.90-2 +- add debug-static package + +* Mon Dec 31 2001 Jakub Jelinek 2.2.90-1 +- update from CVS +- remove -D__USE_STRING_INLINES +- add debug subpackage to trim glibc and glibc-devel size + +* Wed Oct 3 2001 Jakub Jelinek 2.2.4-19 +- fix strsep + +* Fri Sep 28 2001 Jakub Jelinek 2.2.4-18 +- fix a ld.so bug with duplicate searchlists in l_scope +- fix erfcl(-inf) +- turn /usr/lib/librt.so into linker script + +* Wed Sep 26 2001 Jakub Jelinek 2.2.4-17 +- fix a ld.so lookup bug after lots of dlopen calls +- fix CMSG_DATA for non-gcc non-ISOC99 compilers (#53984) +- prelinking support for Sparc64 + +* Fri Sep 21 2001 Jakub Jelinek 2.2.4-16 +- update from CVS to fix DT_SYMBOLIC +- prelinking support for Alpha and Sparc + +* Tue Sep 18 2001 Jakub Jelinek 2.2.4-15 +- update from CVS + - linuxthreads now retries if -1/EINTR is returned from + reading or writing to thread manager pipe (#43742) +- use DT_FILTER in librt.so (#53394) + - update glibc prelink patch so that it handles filters +- fix timer_* with SIGEV_NONE (#53494) +- make glibc_post_upgrade work on PPC (patch from Franz Sirl) + +* Mon Sep 10 2001 Jakub Jelinek 2.2.4-14 +- fix build on sparc32 +- 2.2.4-13 build for some reason missed some locales + on alpha/ia64 + +* Mon Sep 3 2001 Jakub Jelinek 2.2.4-13 +- fix iconvconfig + +* Mon Sep 3 2001 Jakub Jelinek 2.2.4-12 +- add fam to /etc/rpc (#52863) +- fix for C++ (#52960) +- fix perror + +* Mon Aug 27 2001 Jakub Jelinek 2.2.4-11 +- fix strnlen(x, -1) + +* Mon Aug 27 2001 Jakub Jelinek 2.2.4-10 +- doh, should only define __libc_rwlock_t + if __USE_UNIX98. + +* Mon Aug 27 2001 Jakub Jelinek 2.2.4-9 +- fix bits/libc-lock.h so that gcc can compile +- fix s390 build + +* Fri Aug 24 2001 Jakub Jelinek 2.2.4-8 +- kill stale library symlinks in ldconfig (#52350) +- fix inttypes.h for G++ < 3.0 +- use DT_REL*COUNT + +* Wed Aug 22 2001 Jakub Jelinek 2.2.4-7 +- fix strnlen on IA-64 (#50077) + +* Thu Aug 16 2001 Jakub Jelinek 2.2.4-6 +- glibc 2.2.4 final +- fix -lpthread -static (#51672) + +* Fri Aug 10 2001 Jakub Jelinek 2.2.4-5 +- doh, include libio/tst-swscanf.c + +* Fri Aug 10 2001 Jakub Jelinek 2.2.4-4 +- don't crash on catclose(-1) +- fix wscanf %%[] handling +- fix return value from swprintf +- handle year + %%U/%%W week + week day in strptime + +* Thu Aug 9 2001 Jakub Jelinek 2.2.4-3 +- update from CVS to + - fix strcoll (#50548) + - fix seekdir (#51132) + - fix memusage (#50606) +- don't make gconv-modules.cache %%config file, just don't verify + its content. + +* Mon Aug 6 2001 Jakub Jelinek +- fix strtod and *scanf (#50723, #50724) + +* Sat Aug 4 2001 Jakub Jelinek +- update from CVS + - fix iconv cache handling +- glibc should not own %%{_infodir}, %%{_mandir} nor %%{_mandir}/man3 (#50673) +- add gconv-modules.cache as emtpy config file (#50699) +- only run iconvconfig if /usr is mounted read-write (#50667) + +* Wed Jul 25 2001 Jakub Jelinek +- move iconvconfig from glibc-common into glibc subpackage, + call it from glibc_post_upgrade instead of common's post. + +* Tue Jul 24 2001 Jakub Jelinek +- turn off debugging printouts in iconvconfig + +* Tue Jul 24 2001 Jakub Jelinek +- update from CVS + - fix IA-32 makecontext + - make fflush(0) thread-safe (#46446) + +* Mon Jul 23 2001 Jakub Jelinek +- adjust prelinking DT_* and SHT_* values in elf.h +- update from CVS + - iconv cache + - make iconv work in SUID/SGID programs (#34611) + +* Fri Jul 20 2001 Jakub Jelinek +- update from CVS + - kill non-pic code in libm.so + - fix getdate + - fix some locales (#49402) +- rebuilt with binutils-2.11.90.0.8-5 to place .interp section + properly in libBrokenLocale.so, libNoVersion.so and libanl.so +- add floating stacks on IA-64, Alpha, Sparc (#49308) + +* Mon Jul 16 2001 Jakub Jelinek +- make /lib/i686 directory owned by glibc*.i686.rpm + +* Mon Jul 9 2001 Jakub Jelinek +- remove rquota.[hx] headers which are now provided by quota (#47141) +- add prelinking patch + +* Thu Jul 5 2001 Jakub Jelinek +- require sh-utils for nscd + +* Mon Jun 25 2001 Jakub Jelinek +- update from CVS (#43681, #43350, #44663, #45685) +- fix ro_RO bug (#44644) + +* Wed Jun 6 2001 Jakub Jelinek +- fix a bunch of math bugs (#43210, #43345, #43346, #43347, #43348, #43355) +- make rpc headers -ansi compilable (#42390) +- remove alphaev6 optimized memcpy, since there are still far too many + broken apps which call memcpy where they should call memmove +- update from CVS to (among other things): + - fix tanhl bug (#43352) + +* Tue May 22 2001 Jakub Jelinek +- fix #include with -D_XOPEN_SOURCE=500 on ia64 (#35968) +- fix a dlclose reldeps handling bug +- some more profiling fixes +- fix tgmath.h + +* Thu May 17 2001 Jakub Jelinek +- make ldconfig more quiet +- fix LD_PROFILE on i686 (#41030) + +* Wed May 16 2001 Jakub Jelinek +- fix the hardlink program, so that it really catches all files with + identical content +- add a s390x clone fix + +* Wed May 16 2001 Jakub Jelinek +- fix rpc for non-threaded apps using svc_fdset and similar variables (#40409) +- fix nss compatibility DSO versions for alphaev6 +- add a hardlink program instead of the shell 3x for plus cmp -s/link + which takes a lot of time during build +- rework BuildPreReq and Conflicts with gcc, so that + it applies only where it has to + +* Fri May 11 2001 Jakub Jelinek +- fix locale name of ja_JP in UTF-8 (#39783) +- fix re_search_2 (#40244) +- fix memusage script (#39138, #39823) +- fix dlsym(RTLD_NEXT, ) from main program (#39803) +- fix xtrace script (#39609) +- make glibc conflict with glibc-devel 2.2.2 and below (to make sure + libc_nonshared.a has atexit) +- fix getconf LFS_CFLAGS on 64bitters +- recompile with gcc-2.96-84 or above to fix binary compatibility problem + with __frame_state_for function (#37933) + +* Fri Apr 27 2001 Jakub Jelinek +- glibc 2.2.3 release + - fix strcoll (#36539) +- add BuildPreReqs (#36378) + +* Wed Apr 25 2001 Jakub Jelinek +- update from CVS + +* Fri Apr 20 2001 Jakub Jelinek +- update from CVS + - fix sparc64, ia64 + - fix some locale syntax errors (#35982) + +* Wed Apr 18 2001 Jakub Jelinek +- update from CVS + +* Wed Apr 11 2001 Jakub Jelinek +- update from CVS + +* Fri Apr 6 2001 Jakub Jelinek +- support even 2.4.0 kernels on ia64, sparc64 and s390x +- include UTF-8 locales +- make gconv-modules %%config(noreplace) + +* Fri Mar 23 2001 Jakub Jelinek +- back out sunrpc changes + +* Wed Mar 21 2001 Jakub Jelinek +- update from CVS + - fix ia64 build + - fix pthread_getattr_np + +* Fri Mar 16 2001 Jakub Jelinek +- update from CVS + - run atexit() registered functions at dlclose time if they are in shared + libraries (#28625) + - add pthread_getattr_np API to make JVM folks happy + +* Wed Mar 14 2001 Jakub Jelinek +- require 2.4.1 instead of 2.4.0 on platforms where it required 2.4 kernel +- fix ldd behaviour on unresolved symbols +- remove nonsensical ldconfig warning, update osversion for the most + recent library with the same soname in the same directory instead (#31703) +- apply selected patches from CVS +- s390x spec file changes from Florian La Roche + +* Wed Mar 7 2001 Jakub Jelinek +- fix gencat (#30894) +- fix ldconfig changes from yesterday, fix LD_ASSUME_KERNEL handling + +* Tue Mar 6 2001 Jakub Jelinek +- update from CVS +- make pthread_attr_setstacksize consistent before and after pthread manager + is started (#28194) +- pass back struct sigcontext from pthread signal wrapper (on ia32 only so + far, #28493) +- on i686 ship both --enable-kernel 2.2.5 and 2.4.0 libc/libm/libpthread, + make ld.so pick the right one + +* Sat Feb 17 2001 Preston Brown +- glib-common doesn't require glibc, until we can figure out how to get out of dependency hell. + +* Sat Feb 17 2001 Jakub Jelinek +- make glibc require particular version of glibc-common + and glibc-common prerequire glibc. + +* Fri Feb 16 2001 Jakub Jelinek +- glibc 2.2.2 release + - fix regex REG_ICASE bug seen in ksymoops + +* Sat Feb 10 2001 Jakub Jelinek +- fix regexec leaking memory (#26864) + +* Fri Feb 9 2001 Jakub Jelinek +- update from CVS + - fix ia64 build with gnupro + - make regex 64bit clean + - fix tgmath make check failures on alpha + +* Tue Feb 6 2001 Jakub Jelinek +- update again for ia64 DF_1_INITFIRST + +* Fri Feb 2 2001 Jakub Jelinek +- update from CVS + - fix getaddrinfo (#25437) + - support DF_1_INITFIRST (#25029) + +* Wed Jan 24 2001 Jakub Jelinek +- build all auxiliary arches with --enablekernel 2.4.0, those wanting + to run 2.2 kernels can downgrade to the base architecture glibc. + +* Sat Jan 20 2001 Jakub Jelinek +- remove %%lang() flags from %%{_prefix}/lib/locale files temporarily + +* Sun Jan 14 2001 Jakub Jelinek +- update to 2.2.1 final + - fix a pthread_kill_other_threads_np breakage (#23966) + - make static binaries using dlopen work on ia64 again +- fix a typo in glibc-common group + +* Wed Jan 10 2001 Bernhard Rosenkraenzer +- devel requires glibc = %%{version} +- noreplace /etc/nscd.conf + +* Wed Jan 10 2001 Jakub Jelinek +- some more security fixes: + - don't look up LD_PRELOAD libs in cache for SUID apps + (because that bypasses SUID bit checking on the library) + - place output files for profiling SUID apps into /var/profile, + use O_NOFOLLOW for them + - add checks for $MEMUSAGE_OUTPUT and $SEGFAULT_OUTPUT_NAME +- hardlink identical locale files together +- add %%lang() tags to locale stuff +- remove ko_KR.utf8 for now, it is provided by locale-utf8 package + +* Mon Jan 8 2001 Jakub Jelinek +- add glibc-common subpackage +- fix alphaev6 memcpy (#22494) +- fix sys/cdefs.h (#22908) +- don't define stdin/stdout/stderr as macros for -traditional (#22913) +- work around a bug in IBM JDK (#22932, #23012) +- fix pmap_unset when network is down (#23176) +- move nscd in rc.d before netfs on shutdown +- fix $RESOLV_HOST_CONF in SUID apps (#23562) + +* Fri Dec 15 2000 Jakub Jelinek +- fix ftw and nftw + +* Wed Dec 13 2000 Jakub Jelinek +- fix fcvt (#22184) +- ldd /lib/ld-linux.so.2 is not crashing any longer again (#22197) +- fix gencat + +* Mon Dec 11 2000 Jakub Jelinek +- fix alpha htonl and alphaev6 stpcpy + +* Sat Dec 9 2000 Jakub Jelinek +- update to CVS to: + - fix getnameinfo (#21934) + - don't stomp on memory in rpath handling (#21544) + - fix setlocale (#21507) +- fix libNoVersion.so.1 loading code (#21579) +- use auxarches define in spec file for auxiliary + architectures (#21219) +- remove /usr/share directory from filelist (#21218) + +* Sun Nov 19 2000 Jakub Jelinek +- update to CVS to fix getaddrinfo + +* Fri Nov 17 2000 Jakub Jelinek +- update to CVS to fix freopen +- remove all alpha workarounds, not needed anymore + +* Wed Nov 15 2000 Jakub Jelinek +- fix dladdr bug on alpha/sparc32/sparc64 +- fix Makefiles so that they run static tests properly + +* Tue Nov 14 2000 Jakub Jelinek +- update to CVS to fix ldconfig + +* Thu Nov 9 2000 Jakub Jelinek +- update to glibc 2.2 release + +* Mon Nov 6 2000 Jakub Jelinek +- update to CVS to: + - export __sysconf@@GLIBC_2.2 (#20417) + +* Fri Nov 3 2000 Jakub Jelinek +- merge to 2.1.97 + +* Mon Oct 30 2000 Jakub Jelinek +- update to CVS, including: + - fix WORD_BIT/LONG_BIT definition in limits.h (#19088) + - fix hesiod (#19375) + - set LC_MESSAGES in zic/zdump for proper error message output (#19495) + - fix LFS fcntl when used with non-LFS aware kernels (#19730) + +* Thu Oct 19 2000 Jakub Jelinek +- fix alpha semctl (#19199) +- update to CVS, including: + - fix glibc headers for Compaq non-gcc compilers + - fix locale alias handling code (#18832) + - fix rexec on little endian machines (#18886) +- started writing changelog again + +* Thu Aug 10 2000 Adrian Havill +- added ja ujis alias for backwards compatibility diff --git a/SOURCES/RHEL-18039-1.patch b/SOURCES/RHEL-18039-1.patch new file mode 100644 index 0000000..549b26f --- /dev/null +++ b/SOURCES/RHEL-18039-1.patch @@ -0,0 +1,79 @@ +commit 95f61610f3e481d191b6184432342236fd59186d +Author: Florian Weimer +Date: Wed Jul 24 12:06:47 2024 +0200 + + resolv: Support clearing option flags with a “-” prefix (bug 14799) + + I think using a “-” prefix is less confusing than introducing + double-negation construct (“no-no-tld-query”). + + Reviewed-by: DJ Delorie + +diff --git a/resolv/res_init.c b/resolv/res_init.c +index 263263d474721545..243532b3ade338d8 100644 +--- a/resolv/res_init.c ++++ b/resolv/res_init.c +@@ -682,27 +682,29 @@ res_setoptions (struct resolv_conf_parser *parser, const char *options) + { + char str[22]; + uint8_t len; +- uint8_t clear; + unsigned long int flag; + } options[] = { + #define STRnLEN(str) str, sizeof (str) - 1 +- { STRnLEN ("rotate"), 0, RES_ROTATE }, +- { STRnLEN ("edns0"), 0, RES_USE_EDNS0 }, +- { STRnLEN ("single-request-reopen"), 0, RES_SNGLKUPREOP }, +- { STRnLEN ("single-request"), 0, RES_SNGLKUP }, +- { STRnLEN ("no_tld_query"), 0, RES_NOTLDQUERY }, +- { STRnLEN ("no-tld-query"), 0, RES_NOTLDQUERY }, +- { STRnLEN ("no-reload"), 0, RES_NORELOAD }, +- { STRnLEN ("use-vc"), 0, RES_USEVC }, +- { STRnLEN ("trust-ad"), 0, RES_TRUSTAD }, +- { STRnLEN ("no-aaaa"), 0, RES_NOAAAA }, ++ { STRnLEN ("rotate"), RES_ROTATE }, ++ { STRnLEN ("edns0"), RES_USE_EDNS0 }, ++ { STRnLEN ("single-request-reopen"), RES_SNGLKUPREOP }, ++ { STRnLEN ("single-request"), RES_SNGLKUP }, ++ { STRnLEN ("no_tld_query"), RES_NOTLDQUERY }, ++ { STRnLEN ("no-tld-query"), RES_NOTLDQUERY }, ++ { STRnLEN ("no-reload"), RES_NORELOAD }, ++ { STRnLEN ("use-vc"), RES_USEVC }, ++ { STRnLEN ("trust-ad"), RES_TRUSTAD }, ++ { STRnLEN ("no-aaaa"), RES_NOAAAA }, + }; + #define noptions (sizeof (options) / sizeof (options[0])) ++ bool negate_option = *cp == '-'; ++ if (negate_option) ++ ++cp; + for (int i = 0; i < noptions; ++i) + if (strncmp (cp, options[i].str, options[i].len) == 0) + { +- if (options[i].clear) +- parser->template.options &= options[i].flag; ++ if (negate_option) ++ parser->template.options &= ~options[i].flag; + else + parser->template.options |= options[i].flag; + break; +diff --git a/resolv/tst-resolv-res_init-skeleton.c b/resolv/tst-resolv-res_init-skeleton.c +index 6bef62cde2cbf8cd..d3a19eb305d41467 100644 +--- a/resolv/tst-resolv-res_init-skeleton.c ++++ b/resolv/tst-resolv-res_init-skeleton.c +@@ -679,6 +679,16 @@ struct test_case test_cases[] = + "; nameserver[0]: [192.0.2.1]:53\n", + .res_options = "attempts:5 ndots:3 edns0 ", + }, ++ {.name = "RES_OPTIONS can clear flags", ++ .conf = "options ndots:2 use-vc no-aaaa edns0\n" ++ "nameserver 192.0.2.1\n", ++ .expected = "options ndots:3 use-vc\n" ++ "search example.com\n" ++ "; search[0]: example.com\n" ++ "nameserver 192.0.2.1\n" ++ "; nameserver[0]: [192.0.2.1]:53\n", ++ .res_options = "ndots:3 -edns0 -no-aaaa", ++ }, + {.name = "many search list entries (bug 19569)", + .conf = "nameserver 192.0.2.1\n" + "search corp.example.com support.example.com" diff --git a/SOURCES/RHEL-18039-2.patch b/SOURCES/RHEL-18039-2.patch new file mode 100644 index 0000000..9834166 --- /dev/null +++ b/SOURCES/RHEL-18039-2.patch @@ -0,0 +1,204 @@ +commit 765325951ac5c7d072278c9424930b29657e9758 +Author: Florian Weimer +Date: Wed Jul 24 12:06:47 2024 +0200 + + resolv: Implement strict-error stub resolver option (bug 27929) + + For now, do not enable this mode by default due to the potential + impact on compatibility with existing deployments. + + Reviewed-by: DJ Delorie + +diff --git a/resolv/res_init.c b/resolv/res_init.c +index 243532b3ade338d8..b838dc70642e1935 100644 +--- a/resolv/res_init.c ++++ b/resolv/res_init.c +@@ -695,6 +695,7 @@ res_setoptions (struct resolv_conf_parser *parser, const char *options) + { STRnLEN ("use-vc"), RES_USEVC }, + { STRnLEN ("trust-ad"), RES_TRUSTAD }, + { STRnLEN ("no-aaaa"), RES_NOAAAA }, ++ { STRnLEN ("strict-error"), RES_STRICTERR }, + }; + #define noptions (sizeof (options) / sizeof (options[0])) + bool negate_option = *cp == '-'; +diff --git a/resolv/res_send.c b/resolv/res_send.c +index 9c77613f374e5469..9a284ed44aa8cc2e 100644 +--- a/resolv/res_send.c ++++ b/resolv/res_send.c +@@ -1234,21 +1234,38 @@ send_dg(res_state statp, + + if (thisansp_error) { + next_ns: +- if (recvresp1 || (buf2 != NULL && recvresp2)) { +- *resplen2 = 0; +- return resplen; +- } +- if (buf2 != NULL && !single_request) ++ /* Outside of strict-error mode, use the first ++ response even if the second response is an ++ error. This allows parallel resolution to ++ succeed even if the recursive resolver ++ always answers with SERVFAIL for AAAA ++ queries (which still happens in practice ++ unfortunately). ++ ++ In strict-error mode, always switch to the ++ next server and try to get a response from ++ there. */ ++ if ((statp->options & RES_STRICTERR) == 0) + { +- /* No data from the first reply. */ +- resplen = 0; +- /* We are waiting for a possible second reply. */ +- if (matching_query == 1) +- recvresp1 = 1; +- else +- recvresp2 = 1; +- +- goto wait; ++ if (recvresp1 || (buf2 != NULL && recvresp2)) ++ { ++ *resplen2 = 0; ++ return resplen; ++ } ++ ++ if (buf2 != NULL && !single_request) ++ { ++ /* No data from the first reply. */ ++ resplen = 0; ++ /* We are waiting for a possible ++ second reply. */ ++ if (matching_query == 1) ++ recvresp1 = 1; ++ else ++ recvresp2 = 1; ++ ++ goto wait; ++ } + } + + /* don't retry if called from dig */ +diff --git a/resolv/resolv.h b/resolv/resolv.h +index f40d6c58cee0f585..b8a0f66a5fd50e22 100644 +--- a/resolv/resolv.h ++++ b/resolv/resolv.h +@@ -133,6 +133,7 @@ struct res_sym { + #define RES_NORELOAD 0x02000000 /* No automatic configuration reload. */ + #define RES_TRUSTAD 0x04000000 /* Request AD bit, keep it in responses. */ + #define RES_NOAAAA 0x08000000 /* Suppress AAAA queries. */ ++#define RES_STRICTERR 0x10000000 /* Report more DNS errors as errors. */ + + #define RES_DEFAULT (RES_RECURSE|RES_DEFNAMES|RES_DNSRCH) + +diff --git a/resolv/tst-resolv-res_init-skeleton.c b/resolv/tst-resolv-res_init-skeleton.c +index d3a19eb305d41467..e41bcebd9d9a8024 100644 +--- a/resolv/tst-resolv-res_init-skeleton.c ++++ b/resolv/tst-resolv-res_init-skeleton.c +@@ -129,6 +129,7 @@ print_resp (FILE *fp, res_state resp) + print_option_flag (fp, &options, RES_NORELOAD, "no-reload"); + print_option_flag (fp, &options, RES_TRUSTAD, "trust-ad"); + print_option_flag (fp, &options, RES_NOAAAA, "no-aaaa"); ++ print_option_flag (fp, &options, RES_STRICTERR, "strict-error"); + fputc ('\n', fp); + if (options != 0) + fprintf (fp, "; error: unresolved option bits: 0x%x\n", options); +@@ -741,6 +742,15 @@ struct test_case test_cases[] = + "nameserver 192.0.2.1\n" + "; nameserver[0]: [192.0.2.1]:53\n" + }, ++ {.name = "strict-error flag", ++ .conf = "options strict-error\n" ++ "nameserver 192.0.2.1\n", ++ .expected = "options strict-error\n" ++ "search example.com\n" ++ "; search[0]: example.com\n" ++ "nameserver 192.0.2.1\n" ++ "; nameserver[0]: [192.0.2.1]:53\n" ++ }, + { NULL } + }; + +diff --git a/resolv/tst-resolv-semi-failure.c b/resolv/tst-resolv-semi-failure.c +index aa9798b5a7dfaa88..b7681210f450bb5a 100644 +--- a/resolv/tst-resolv-semi-failure.c ++++ b/resolv/tst-resolv-semi-failure.c +@@ -67,6 +67,9 @@ response (const struct resolv_response_context *ctx, + resolv_response_close_record (b); + } + ++/* Set to 1 if strict error checking is enabled. */ ++static int do_strict_error; ++ + static void + check_one (void) + { +@@ -83,7 +86,10 @@ check_one (void) + struct addrinfo *ai; + int ret = getaddrinfo ("www.example", "80", &hints, &ai); + const char *expected; +- if (ret == 0 && ai->ai_next != NULL) ++ /* In strict-error mode, a switch to the second name server ++ happens, and both responses are received, so a single ++ response is a bug. */ ++ if (do_strict_error || (ret == 0 && ai->ai_next != NULL)) + expected = ("address: STREAM/TCP 192.0.2.17 80\n" + "address: STREAM/TCP 2001:db8::1 80\n"); + else +@@ -99,33 +105,36 @@ check_one (void) + static int + do_test (void) + { +- for (int do_single_lookup = 0; do_single_lookup < 2; ++do_single_lookup) +- { +- struct resolv_test *aux = resolv_test_start +- ((struct resolv_redirect_config) +- { +- .response_callback = response, +- }); ++ for (do_strict_error = 0; do_strict_error < 2; ++do_strict_error) ++ for (int do_single_lookup = 0; do_single_lookup < 2; ++do_single_lookup) ++ { ++ struct resolv_test *aux = resolv_test_start ++ ((struct resolv_redirect_config) ++ { ++ .response_callback = response, ++ }); + +- if (do_single_lookup) +- _res.options |= RES_SNGLKUP; ++ if (do_strict_error) ++ _res.options |= RES_STRICTERR; ++ if (do_single_lookup) ++ _res.options |= RES_SNGLKUP; + +- for (int do_fail_aaaa = 0; do_fail_aaaa < 2; ++do_fail_aaaa) +- { +- fail_aaaa = do_fail_aaaa; ++ for (int do_fail_aaaa = 0; do_fail_aaaa < 2; ++do_fail_aaaa) ++ { ++ fail_aaaa = do_fail_aaaa; + +- rcode = 2; /* SERVFAIL. */ +- check_one (); ++ rcode = 2; /* SERVFAIL. */ ++ check_one (); + +- rcode = 4; /* NOTIMP. */ +- check_one (); ++ rcode = 4; /* NOTIMP. */ ++ check_one (); + +- rcode = 5; /* REFUSED. */ +- check_one (); +- } ++ rcode = 5; /* REFUSED. */ ++ check_one (); ++ } + +- resolv_test_end (aux); +- } ++ resolv_test_end (aux); ++ } + + return 0; + } diff --git a/SOURCES/bench.mk b/SOURCES/bench.mk new file mode 100644 index 0000000..dfe46bd --- /dev/null +++ b/SOURCES/bench.mk @@ -0,0 +1,77 @@ +objpfx = $(prefix)/$(ver)/usr/libexec/glibc-benchtests/ + +bench-math := acos acosh asin asinh atan atanh cos cosh exp exp2 ffs ffsll \ + log log2 modf pow rint sin sincos sinh sqrt tan tanh + +bench-pthread := pthread_once + +bench := $(bench-math) $(bench-pthread) + +run-bench := $(prefix)/$(ver)/lib64/ld-linux-x86-64.so.2 --library-path $(prefix)/$(ver)/lib64 $${run} + +# String function benchmarks. +string-bench := bcopy bzero memccpy memchr memcmp memcpy memmem memmove \ + mempcpy memset rawmemchr stpcpy stpncpy strcasecmp strcasestr \ + strcat strchr strchrnul strcmp strcpy strcspn strlen \ + strncasecmp strncat strncmp strncpy strnlen strpbrk strrchr \ + strspn strstr strcpy_chk stpcpy_chk memrchr strsep strtok +string-bench-all := $(string-bench) + +stdlib-bench := strtod + +benchset := $(string-bench-all) $(stdlib-bench) + +bench-malloc := malloc-thread + +binaries-bench := $(addprefix $(objpfx)bench-,$(bench)) +binaries-benchset := $(addprefix $(objpfx)bench-,$(benchset)) +binaries-bench-malloc := $(addprefix $(objpfx)bench-,$(bench-malloc)) + +DETAILED_OPT := + +ifdef DETAILED + DETAILED_OPT := -d +endif + +bench: bench-set bench-func bench-malloc + +bench-set: $(binaries-benchset) + for run in $^; do \ + outfile=$(prefix)/$$(basename $${run}.$(ver).out); \ + echo "Running $${run}"; \ + $(run-bench) > $${outfile}.tmp; \ + mv $${outfile}{.tmp,}; \ + done + +bench-malloc: $(binaries-bench-malloc) + run=$(objpfx)bench-malloc-thread; \ + outfile=$(prefix)/$$(basename $${run}.$(ver).out); \ + for thr in 1 8 16 32; do \ + echo "Running $${run} $${thr}"; \ + $(run-bench) $${thr} > $${outfile}.tmp; \ + mv $${outfile}{.tmp,}; \ + done + +# Build and execute the benchmark functions. This target generates JSON +# formatted bench.out. Each of the programs produce independent JSON output, +# so one could even execute them individually and process it using any JSON +# capable language or tool. +bench-func: $(binaries-bench) + { echo "{\"timing_type\": \"hp-timing\","; \ + echo " \"functions\": {"; \ + for run in $^; do \ + if ! [ "x$${run}" = "x$<" ]; then \ + echo ","; \ + fi; \ + echo "Running $${run}" >&2; \ + $(run-bench) $(DETAILED_OPT); \ + done; \ + echo; \ + echo " }"; \ + echo "}"; } > $(prefix)/bench.$(ver).out-tmp; \ + if [ -f $(prefix)/bench.$(ver).out ]; then \ + mv -f $(prefix)/bench.$(ver).out{,.old}; \ + fi; \ + mv -f $(prefix)/bench.$(ver).out{-tmp,} +# scripts/validate_benchout.py bench.out \ +# scripts/benchout.schema.json diff --git a/SOURCES/glibc-RHEL-22226.patch b/SOURCES/glibc-RHEL-22226.patch new file mode 100644 index 0000000..3d09f70 --- /dev/null +++ b/SOURCES/glibc-RHEL-22226.patch @@ -0,0 +1,121 @@ +commit 5361ad3910c257bc327567be76fde532ed238e42 +Author: Florian Weimer +Date: Fri Apr 19 14:38:17 2024 +0200 + + login: Use unsigned 32-bit types for seconds-since-epoch + + These fields store timestamps when the system was running. No Linux + systems existed before 1970, so these values are unused. Switching + to unsigned types allows continued use of the existing struct layouts + beyond the year 2038. + + The intent is to give distributions more time to switch to improved + interfaces that also avoid locking/data corruption issues. + + Reviewed-by: Adhemerval Zanella + +diff --git a/bits/utmp.h b/bits/utmp.h +index f2d1c13d8cd205b2..27cb536800c46d67 100644 +--- a/bits/utmp.h ++++ b/bits/utmp.h +@@ -36,7 +36,7 @@ + struct lastlog + { + #if __WORDSIZE_TIME64_COMPAT32 +- int32_t ll_time; ++ __uint32_t ll_time; + #else + __time_t ll_time; + #endif +@@ -76,7 +76,7 @@ struct utmp + int32_t ut_session; /* Session ID, used for windowing. */ + struct + { +- int32_t tv_sec; /* Seconds. */ ++ __uint32_t tv_sec; /* Seconds. */ + int32_t tv_usec; /* Microseconds. */ + } ut_tv; /* Time entry was made. */ + #else +diff --git a/login/Makefile b/login/Makefile +index f91190e3dcd1e6c6..84563230ef665f9c 100644 +--- a/login/Makefile ++++ b/login/Makefile +@@ -44,9 +44,11 @@ subdir-dirs = programs + vpath %.c programs + + tests := tst-utmp tst-utmpx tst-grantpt tst-ptsname tst-getlogin tst-updwtmpx \ +- tst-pututxline-lockfail tst-pututxline-cache tst-utmp-size tst-utmp-size-64 ++ tst-pututxline-lockfail tst-pututxline-cache tst-utmp-size tst-utmp-size-64 \ ++ tst-utmp-unsigned tst-utmp-unsigned-64 + + CFLAGS-tst-utmp-size-64.c += -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 ++CFLAGS-tst-utmp-unsigned-64.c += -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 + + # Empty compatibility library for old binaries. + extra-libs := libutil +diff --git a/login/tst-utmp-unsigned-64.c b/login/tst-utmp-unsigned-64.c +new file mode 100644 +index 0000000000000000..940e7654f8dc5fd6 +--- /dev/null ++++ b/login/tst-utmp-unsigned-64.c +@@ -0,0 +1 @@ ++#include "tst-utmp-unsigned.c" +diff --git a/login/tst-utmp-unsigned.c b/login/tst-utmp-unsigned.c +new file mode 100644 +index 0000000000000000..27ad03a7d608e83d +--- /dev/null ++++ b/login/tst-utmp-unsigned.c +@@ -0,0 +1,40 @@ ++/* Check that struct utmp, struct utmpx, struct lastlog use unsigned epoch. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++ ++/* Undefined. Used to check that the conditions below are optimized away. */ ++void link_failure_utmp (void); ++void link_failure_utmpx (void); ++void link_failure_lastlog (void); ++ ++static int ++do_test (void) ++{ ++ if ((struct utmp) { .ut_tv = { 0x80000000U, }, }.ut_tv.tv_sec <= 0) ++ link_failure_utmp (); ++ if ((struct utmpx) { .ut_tv = { 0x80000000U, }, }.ut_tv.tv_sec <= 0) ++ link_failure_utmpx (); ++ if ((struct lastlog) { .ll_time = 0x80000000U, }.ll_time <= 0) ++ link_failure_lastlog (); ++ return 0; ++} ++ ++#include +diff --git a/sysdeps/gnu/bits/utmpx.h b/sysdeps/gnu/bits/utmpx.h +index 34b4afbc6ac25968..ed0df9bd8141d4e6 100644 +--- a/sysdeps/gnu/bits/utmpx.h ++++ b/sysdeps/gnu/bits/utmpx.h +@@ -74,7 +74,7 @@ struct utmpx + __int32_t ut_session; /* Session ID, used for windowing. */ + struct + { +- __int32_t tv_sec; /* Seconds. */ ++ __uint32_t tv_sec; /* Seconds. */ + __int32_t tv_usec; /* Microseconds. */ + } ut_tv; /* Time entry was made. */ + #else diff --git a/SOURCES/glibc-bench-compare b/SOURCES/glibc-bench-compare new file mode 100755 index 0000000..84e3aba --- /dev/null +++ b/SOURCES/glibc-bench-compare @@ -0,0 +1,153 @@ +#!/usr/bin/bash +# This script can be invoked as follows: +# +# glibc-bench-compare [options] [BUILD] +# +# Options may be one of the following: +# +# -t The BUILD arguments are task ids and not a version-release string +# -a ARCH Do comparison for ARCH architecture +# +# If any of the above options are given, both BUILD arguments must be given. +# Otherwise, if only one BUILD is specified, then it is compared against the +# installed glibc. + +# Silence the pushd/popd messages +pushd() { + command pushd "$@" > /dev/null 2>&1 +} + +popd() { + command popd "$@" > /dev/null 2>&1 +} + +# Clean up any downloaded files before we exit +trap "rm -rf /tmp/glibc-bench-compare.$BASHPID.*" EXIT + +task=0 +arch=$(uname -i) +options=0 +path=0 +installed= + +# Look for any commandline options +while getopts ":tpa:" opt; do + case $opt in + p) + path=1 + ;; + t) + task=1 + options=1 + echo "Not implemented." + exit 1 + ;; + a) + arch=$OPTARG + options=1 + ;; + *) + ;; + esac +done + +# Done, now shift all option arguments out. +shift $((OPTIND-1)) + +if [ $# -gt 2 ] || [ $# -eq 0 ] || [ $# -lt 2 -a $options -eq 1 ]; then + echo "Usage: $0 [OPTIONS] [new]" + echo + echo "OPTIONS:" + echo -e "\t-t\tCompare two brew tasks" + echo -e "\t-a ARCH\tGet rpms for the ARCH architecture" + echo -e "\t-p\tCompare built rpms in two paths." + echo -e "\t\tThis minimally needs glibc, glibc-common and glibc-benchtests" + exit 1 +fi + +if [ -z $2 ]; then + new="$1" + old=$(rpm --queryformat "%{VERSION}-%{RELEASE}\n" -q glibc | head -1) + installed=$old +else + new="$2" + old="$1" +fi + +decompress_rpms() { + # We were given a path to the rpms. Figure out the version-release and + # decompress the rpms. + if [ -n $1 ]; then + vr=$(rpm --queryformat="%{VERSION}-%{RELEASE}" -qp $1/glibc-2*.rpm | head -1) + mkdir $vr && pushd $vr + fi + + for r in $1*.rpm; do + ( rpm2cpio $r | cpio -di ) > /dev/null + done + + if [ -n $1 ]; then + popd + echo $vr + fi +} + +# Get rpms for a build and decompress them +get_build() { + echo "Processing build $1" + mkdir $1 && pushd $1 + brew buildinfo "glibc-$1" | + sed -n -e "s|/mnt/koji\(.\+$arch.\+\)|http://kojipkgs.fedoraproject.org\1|p" | + while read url; do + echo "Downloading $url" + wget -q $url + done + decompress_rpms + + echo "Removing rpms" + rm -f $1/*.rpm + + popd +} + +# Run benchmarks for a build +run_bench() { + if [ -z $1 ]; then + make DETAILED=1 ver=$installed prefix= -f /usr/libexec/glibc-benchtests/bench.mk bench + else + make DETAILED=1 ver=$1 prefix=$PWD -f $1/usr/libexec/glibc-benchtests/bench.mk bench + fi +} + +# Get absolute paths if needed, since we will change into the working directory +# next. +if [ $path -eq 1 ]; then + old_path=$(realpath $old)/ + new_path=$(realpath $new)/ +fi + +tmpdir=$(mktemp -p /tmp -d glibc-bench-compare.$$.XXXX) +pushd $tmpdir + +# Get both builds. +if [ $path -eq 0 ]; then + if [ -z $installed ]; then + get_build $old + fi + get_build $new +else + old=$(decompress_rpms $old_path) + new=$(decompress_rpms $new_path) +fi + +# make bench for each of those. +if [ -z $installed ]; then + run_bench $old +else + run_bench +fi +run_bench $new + +# Now run the comparison script. +$old/usr/libexec/glibc-benchtests/compare_bench.py $old/usr/libexec/glibc-benchtests/benchout.schema.json \ + bench.$old.out bench.$new.out diff --git a/SOURCES/glibc-cs-path.patch b/SOURCES/glibc-cs-path.patch new file mode 100644 index 0000000..40835d2 --- /dev/null +++ b/SOURCES/glibc-cs-path.patch @@ -0,0 +1,44 @@ +Short description: Adjust CS_PATH and the test container layout. +Author(s): Fedora glibc team +Origin: PATCH +Upstream status: not-needed + +In Fedora we should return only /usr/bin as CS_PATH because /bin is just +a symlink to /usr/bin after MoveToUsr transition (which glibc has not +really completed). + +We also create /{bin,lib,lib64,sbin} in the test container as symbolic +links. This brings the test container in line with Fedora's filesystem +layout and avoids some test failures. For example, because Fedora's +CS_PATH is /usr/bin, tst-vfork3 will try to execute /usr/bin/echo in the +container. Without this change the container installs `echo' in /bin +not /usr/bin, causing the test to fail. + +diff --git a/Makefile b/Makefile +index a49870d3d1e636a9..feb2599203b10098 100644 +--- a/Makefile ++++ b/Makefile +@@ -598,9 +598,13 @@ $(tests-container) $(addsuffix /tests,$(subdirs)) : \ + $(objpfx)testroot.pristine/install.stamp : + test -d $(objpfx)testroot.pristine || \ + mkdir $(objpfx)testroot.pristine +- # We need a working /bin/sh for some of the tests. +- test -d $(objpfx)testroot.pristine/bin || \ +- mkdir $(objpfx)testroot.pristine/bin ++ # Set up symlinks to directories whose contents got moved to /usr ++ for moved in bin lib lib64 sbin; do \ ++ test -d $(objpfx)testroot.pristine/usr/$$moved || \ ++ mkdir -p $(objpfx)testroot.pristine/usr/$$moved ;\ ++ test -e $(objpfx)testroot.pristine/$$moved || \ ++ ln -s usr/$$moved $(objpfx)testroot.pristine/$$moved ;\ ++ done + # We need the compiled locale dir for localedef tests. + test -d $(objpfx)testroot.pristine/$(complocaledir) || \ + mkdir -p $(objpfx)testroot.pristine/$(complocaledir) +diff --git a/sysdeps/unix/confstr.h b/sysdeps/unix/confstr.h +index 15859c3b2759878e..9b63b7f8069866fd 100644 +--- a/sysdeps/unix/confstr.h ++++ b/sysdeps/unix/confstr.h +@@ -1 +1 @@ +-#define CS_PATH "/bin:/usr/bin" ++#define CS_PATH "/usr/bin" diff --git a/SOURCES/glibc-fedora-linux-tcsetattr.patch b/SOURCES/glibc-fedora-linux-tcsetattr.patch new file mode 100644 index 0000000..3ae7e27 --- /dev/null +++ b/SOURCES/glibc-fedora-linux-tcsetattr.patch @@ -0,0 +1,61 @@ +Short description: Fedora-specific workaround for kernel pty bug. +Author(s): Fedora glibc team +Origin: PATCH +Upstream status: not-submitted + +This is a Fedora-specific workaround for a kernel bug where calling +ioctl on a pty will silently ignore the invalid c_cflag. The +workaround is to use TCGETS to verify the setting matches. This is +not upstream and needs to either be removed or submitted upstream +after analysis. + +Index: b/sysdeps/unix/sysv/linux/tcsetattr.c +=================================================================== +--- a/sysdeps/unix/sysv/linux/tcsetattr.c ++++ b/sysdeps/unix/sysv/linux/tcsetattr.c +@@ -45,6 +45,7 @@ __tcsetattr (int fd, int optional_action + { + struct __kernel_termios k_termios; + unsigned long int cmd; ++ int retval; + + switch (optional_actions) + { +@@ -75,7 +76,36 @@ __tcsetattr (int fd, int optional_action + memcpy (&k_termios.c_cc[0], &termios_p->c_cc[0], + __KERNEL_NCCS * sizeof (cc_t)); + +- return INLINE_SYSCALL (ioctl, 3, fd, cmd, &k_termios); ++ retval = INLINE_SYSCALL (ioctl, 3, fd, cmd, &k_termios); ++ ++ if (retval == 0 && cmd == TCSETS) ++ { ++ /* The Linux kernel has a bug which silently ignore the invalid ++ c_cflag on pty. We have to check it here. */ ++ int save = errno; ++ retval = INLINE_SYSCALL (ioctl, 3, fd, TCGETS, &k_termios); ++ if (retval) ++ { ++ /* We cannot verify if the setting is ok. We don't return ++ an error (?). */ ++ __set_errno (save); ++ retval = 0; ++ } ++ else if ((termios_p->c_cflag & (PARENB | CREAD)) ++ != (k_termios.c_cflag & (PARENB | CREAD)) ++ || ((termios_p->c_cflag & CSIZE) ++ && ((termios_p->c_cflag & CSIZE) ++ != (k_termios.c_cflag & CSIZE)))) ++ { ++ /* It looks like the Linux kernel silently changed the ++ PARENB/CREAD/CSIZE bits in c_cflag. Report it as an ++ error. */ ++ __set_errno (EINVAL); ++ retval = -1; ++ } ++ } ++ ++ return retval; + } + weak_alias (__tcsetattr, tcsetattr) + libc_hidden_def (tcsetattr) diff --git a/SOURCES/glibc-fedora-localedata-rh61908.patch b/SOURCES/glibc-fedora-localedata-rh61908.patch new file mode 100644 index 0000000..d62d7a2 --- /dev/null +++ b/SOURCES/glibc-fedora-localedata-rh61908.patch @@ -0,0 +1,50 @@ +Short description: Add 4 ISO-8859-15 locales to SUPPORTED for Euro symbol. +Author(s): Fedora glibc team +Origin: PATCH +Bug-RHEL: #61908 +Upstream status: not-needed + +Very early RHL 7.3 requirement to add these locales so users can +get access to Euro symbol. We should review this bug and decide if +the UTF-8 locales are now serving the same purpose and drop the +additional locales. + +* Tue Mar 26 2002 Jakub Jelinek 2.2.5-28 +- add a couple of .ISO-8859-15 locales (#61908) + +diff --git a/localedata/SUPPORTED b/localedata/SUPPORTED +index c8b63cc2fe2b4547..32088fdef06a14a3 100644 +--- a/localedata/SUPPORTED ++++ b/localedata/SUPPORTED +@@ -103,6 +103,7 @@ cy_GB.UTF-8/UTF-8 \ + cy_GB/ISO-8859-14 \ + da_DK.UTF-8/UTF-8 \ + da_DK/ISO-8859-1 \ ++da_DK.ISO-8859-15/ISO-8859-15 \ + de_AT.UTF-8/UTF-8 \ + de_AT/ISO-8859-1 \ + de_AT@euro/ISO-8859-15 \ +@@ -140,6 +141,7 @@ en_DK.UTF-8/UTF-8 \ + en_DK/ISO-8859-1 \ + en_GB.UTF-8/UTF-8 \ + en_GB/ISO-8859-1 \ ++en_GB.ISO-8859-15/ISO-8859-15 \ + en_HK.UTF-8/UTF-8 \ + en_HK/ISO-8859-1 \ + en_IE.UTF-8/UTF-8 \ +@@ -157,6 +159,7 @@ en_SG.UTF-8/UTF-8 \ + en_SG/ISO-8859-1 \ + en_US.UTF-8/UTF-8 \ + en_US/ISO-8859-1 \ ++en_US.ISO-8859-15/ISO-8859-15 \ + en_ZA.UTF-8/UTF-8 \ + en_ZA/ISO-8859-1 \ + en_ZM/UTF-8 \ +@@ -425,6 +428,7 @@ sv_FI/ISO-8859-1 \ + sv_FI@euro/ISO-8859-15 \ + sv_SE.UTF-8/UTF-8 \ + sv_SE/ISO-8859-1 \ ++sv_SE.ISO-8859-15/ISO-8859-15 \ + sw_KE/UTF-8 \ + sw_TZ/UTF-8 \ + syr/UTF-8 \ diff --git a/SOURCES/glibc-fedora-manual-dircategory.patch b/SOURCES/glibc-fedora-manual-dircategory.patch new file mode 100644 index 0000000..11c2656 --- /dev/null +++ b/SOURCES/glibc-fedora-manual-dircategory.patch @@ -0,0 +1,31 @@ +Short description: Place glibc info into "Libraries" category. +Author(s): Fedora glibc team +Origin: PATCH +Upstream status: not-needed + +The category names for libraries is completely random including +"Libraries", "GNU Libraries", "GNU libraries", and "Software libraries." +In the GNU info manual the "Software libraries" category is given as an +example, but really we need to standardize on a category for upstream. +I suggest we drop this change after some upstream discussion. + +From 4820b9175535e13df79ce816106016040014916e Mon Sep 17 00:00:00 2001 +From: Jakub Jelinek +Date: Fri, 3 Nov 2006 16:31:21 +0000 +Subject: [PATCH] Change @dircategory. + +--- + manual/libc.texinfo | 2 +- + 1 files changed, 1 insertions(+), 1 deletions(-) + +--- a/manual/libc.texinfo ++++ b/manual/libc.texinfo +@@ -7,7 +7,7 @@ + @include macros.texi + + @comment Tell install-info what to do. +-@dircategory Software libraries ++@dircategory Libraries + @direntry + * Libc: (libc). C library. + @end direntry diff --git a/SOURCES/glibc-python3.patch b/SOURCES/glibc-python3.patch new file mode 100644 index 0000000..ae36c51 --- /dev/null +++ b/SOURCES/glibc-python3.patch @@ -0,0 +1,30 @@ +Use python3 for installed executable python scripts. + +Fedora is a Python3-only distribution: +https://fedoraproject.org/wiki/FinalizingFedoraSwitchtoPython3 + +This fixes build failures where builders may strictly enforce only +python3 during a transitional phase. + +Author: Carlos O'Donell + +diff --git a/benchtests/scripts/compare_bench.py b/benchtests/scripts/compare_bench.py +index d4370d723f313a0e..2bbb91d6ea9f6c18 100755 +--- a/benchtests/scripts/compare_bench.py ++++ b/benchtests/scripts/compare_bench.py +@@ -1,4 +1,4 @@ +-#!/usr/bin/python ++#!/usr/bin/python3 + # Copyright (C) 2015-2024 Free Software Foundation, Inc. + # This file is part of the GNU C Library. + # +diff --git a/benchtests/scripts/import_bench.py b/benchtests/scripts/import_bench.py +index fac2945f6d36db13..ac40878934490380 100644 +--- a/benchtests/scripts/import_bench.py ++++ b/benchtests/scripts/import_bench.py +@@ -1,4 +1,4 @@ +-#!/usr/bin/python ++#!/usr/bin/python3 + # Copyright (C) 2015-2024 Free Software Foundation, Inc. + # This file is part of the GNU C Library. + # diff --git a/SOURCES/glibc-rh2292195-1.patch b/SOURCES/glibc-rh2292195-1.patch new file mode 100644 index 0000000..fc5bbfd --- /dev/null +++ b/SOURCES/glibc-rh2292195-1.patch @@ -0,0 +1,72 @@ +commit 086910fc41655152812b515dc324d2ac0dc36e67 +Author: Florian Weimer +Date: Thu Jun 20 10:32:16 2024 +0200 + + malloc: Always install mtrace (bug 31892) + + Generation of the Perl script does not depend on Perl, so we can + always install it even if $(PERL) is not set during the build. + + Change the malloc/mtrace.pl text substition not to rely on $(PERL). + Instead use PATH at run time to find the Perl interpreter. The Perl + interpreter cannot execute directly a script that starts with + “#! /bin/sh”: it always executes it with /bin/sh. There is no + perl command line switch to disable this behavior. Instead, use + the Perl require function to execute the script. The additional + shift calls remove the “.” shell arguments. Perl interprets the + “.” as a string concatenation operator, making the expression + syntactically valid. + + Reviewed-by: Carlos O'Donell + +diff --git a/malloc/Makefile b/malloc/Makefile +index cc14cf66c9661f99..02aff1bd1dc664c3 100644 +--- a/malloc/Makefile ++++ b/malloc/Makefile +@@ -250,7 +250,6 @@ others-extras = mcheck-init.o + aux := set-freeres thread-freeres + + # The Perl script to analyze the output of the mtrace functions. +-ifneq ($(PERL),no) + install-bin-script = mtrace + generated += mtrace + +@@ -261,7 +260,6 @@ address-width=10 + else + address-width=18 + endif +-endif + + # Unless we get a test for the availability of libgd which also works + # for cross-compiling we disable the memusagestat generation in this +@@ -349,7 +347,7 @@ sLIBdir := $(shell echo $(slibdir) | sed 's,lib\(\|64\)$$,\\\\$$LIB,') + + $(objpfx)mtrace: mtrace.pl + rm -f $@.new +- sed -e 's|@PERL@|$(PERL)|' -e 's|@XXX@|$(address-width)|' \ ++ sed -e 's|@XXX@|$(address-width)|' \ + -e 's|@VERSION@|$(version)|' \ + -e 's|@PKGVERSION@|$(PKGVERSION)|' \ + -e 's|@REPORT_BUGS_TO@|$(REPORT_BUGS_TO)|' $^ > $@.new \ +diff --git a/malloc/mtrace.pl b/malloc/mtrace.pl +index 075da0d9f11da15c..dc6085820e62092c 100644 +--- a/malloc/mtrace.pl ++++ b/malloc/mtrace.pl +@@ -1,6 +1,6 @@ +-#! @PERL@ +-eval "exec @PERL@ -S $0 $@" +- if 0; ++#! /bin/sh ++eval exec "perl -e 'shift; \$progname=shift; shift; require \$progname'" . "$0" . "$@" ++ if 0; + # Copyright (C) 1997-2024 Free Software Foundation, Inc. + # This file is part of the GNU C Library. + # Based on the mtrace.awk script. +@@ -22,7 +22,6 @@ eval "exec @PERL@ -S $0 $@" + $VERSION = "@VERSION@"; + $PKGVERSION = "@PKGVERSION@"; + $REPORT_BUGS_TO = '@REPORT_BUGS_TO@'; +-$progname = $0; + + sub usage { + print "Usage: mtrace [OPTION]... [Binary] MtraceData\n"; diff --git a/SOURCES/glibc-rh2292195-2.patch b/SOURCES/glibc-rh2292195-2.patch new file mode 100644 index 0000000..61436e8 --- /dev/null +++ b/SOURCES/glibc-rh2292195-2.patch @@ -0,0 +1,84 @@ +commit dd144dce21c864781fade4561581d50fb4549956 +Author: Florian Weimer +Date: Thu Jun 20 20:55:10 2024 +0200 + + malloc: Replace shell/Perl gate in mtrace + + The previous version expanded $0 and $@ twice. + + The new version defines a q no-op shell command. The Perl syntax + error is masked by the eval Perl function. The q { … } construct + is executed by the shell without errors because the q shell function + was defined, but treated as a non-expanding quoted string by Perl, + effectively hiding its context from the Perl interpreter. As before + the script is read by require instead of executed directly, to avoid + infinite recursion because the #! line contains /bin/sh. + + Introduce the “fatal” function to produce diagnostics that are not + suppressed by “do”. Use “do” instead of “require” because it has + fewer requirements on the executed script than “require”. + + Prefix relative paths with './' because “do” (and “require“ before) + searches for the script in @INC if the path is relative and does not + start with './'. Use $_ to make the trampoline shorter. + + Add an Emacs mode marker to indentify the script as a Perl script. + +diff --git a/malloc/mtrace.pl b/malloc/mtrace.pl +index dc6085820e62092c..0a631a07bc4cfbb6 100644 +--- a/malloc/mtrace.pl ++++ b/malloc/mtrace.pl +@@ -1,6 +1,12 @@ + #! /bin/sh +-eval exec "perl -e 'shift; \$progname=shift; shift; require \$progname'" . "$0" . "$@" +- if 0; ++# -*- perl -*- ++eval "q () { ++ : ++}"; ++q { ++ exec perl -e '$_ = shift; $_ = "./$_" unless m,^/,; do $_' "$0" "$@" ++} ++; + # Copyright (C) 1997-2024 Free Software Foundation, Inc. + # This file is part of the GNU C Library. + # Based on the mtrace.awk script. +@@ -22,6 +28,7 @@ eval exec "perl -e 'shift; \$progname=shift; shift; require \$progname'" . "$0" + $VERSION = "@VERSION@"; + $PKGVERSION = "@PKGVERSION@"; + $REPORT_BUGS_TO = '@REPORT_BUGS_TO@'; ++$progname = $_; + + sub usage { + print "Usage: mtrace [OPTION]... [Binary] MtraceData\n"; +@@ -33,6 +40,11 @@ sub usage { + exit 0; + } + ++sub fatal { ++ print STDERR "$_[0]\n"; ++ exit 1; ++} ++ + # We expect two arguments: + # #1: the complete path to the binary + # #2: the mtrace data filename +@@ -86,7 +98,7 @@ if ($#ARGV == 0) { + close (LOCS); + } + } else { +- die "Wrong number of arguments, run $progname --help for help."; ++ fatal "Wrong number of arguments, run $progname --help for help."; + } + + sub addr2line { +@@ -148,7 +160,8 @@ sub location { + } + + $nr=0; +-open(DATA, "<$data") || die "Cannot open mtrace data file"; ++open(DATA, "<$data") ++ or fatal "$progname: Cannot open mtrace data file $data: $!"; + while () { + my @cols = split (' '); + my $n, $where; diff --git a/SOURCES/glibc-rh2292195-3.patch b/SOURCES/glibc-rh2292195-3.patch new file mode 100644 index 0000000..ef2d7f1 --- /dev/null +++ b/SOURCES/glibc-rh2292195-3.patch @@ -0,0 +1,31 @@ +commit 2a6c922f09e7a1c206e0cbdb4424f1cf101a5bda +Author: Andreas Schwab +Date: Thu Jun 20 14:13:01 2024 +0200 + + mtrace: make shell commands robust against meta characters + + Use the list form of the open function to avoid interpreting meta + characters in the arguments. + +diff --git a/malloc/mtrace.pl b/malloc/mtrace.pl +index 0a631a07bc4cfbb6..32b4da935f7c7c4a 100644 +--- a/malloc/mtrace.pl ++++ b/malloc/mtrace.pl +@@ -87,7 +87,7 @@ if ($#ARGV == 0) { + } + # Set the environment variable LD_TRACE_LOADED_OBJECTS to 2 so the + # executable is also printed. +- if (open (locs, "env LD_TRACE_LOADED_OBJECTS=2 $prog |")) { ++ if (open (locs, "-|", "env", "LD_TRACE_LOADED_OBJECTS=2", $prog)) { + while () { + chop; + if (/^.*=> (.*) .(0x[0123456789abcdef]*).$/) { +@@ -104,7 +104,7 @@ if ($#ARGV == 0) { + sub addr2line { + my $addr = pop(@_); + my $prog = pop(@_); +- if (open (ADDR, "addr2line -e $prog $addr|")) { ++ if (open (ADDR, "-|", "addr2line", "-e", $prog, $addr)) { + my $line = ; + chomp $line; + close (ADDR); diff --git a/SOURCES/glibc-rh827510.patch b/SOURCES/glibc-rh827510.patch new file mode 100644 index 0000000..6115891 --- /dev/null +++ b/SOURCES/glibc-rh827510.patch @@ -0,0 +1,37 @@ +Short description: Fix newlocale error return. +Author(s): Fedora glibc team +Origin: PATCH +Bug-RHEL: #832516 +Bug-Fedora: #827510 +Bug-Upstream: #14247 +Upstream status: not-submitted + +This needs to go upstream right away to fix the error case for +newlocale not correctly returning an error. + +2012-06-14 Jeff Law + + * locale/loadlocale.c (_nl_load_locale): Delay setting + file->decided until we have successfully loaded the file's + data. + +diff --git a/locale/loadlocale.c b/locale/loadlocale.c +index e3fa187..9fd9216 100644 +--- a/locale/loadlocale.c ++++ b/locale/loadlocale.c +@@ -169,7 +169,6 @@ _nl_load_locale (struct loaded_l10nfile *file, int category) + int save_err; + int alloc = ld_mapped; + +- file->decided = 1; + file->data = NULL; + + fd = __open_nocancel (file->filename, O_RDONLY | O_CLOEXEC); +@@ -278,6 +277,7 @@ _nl_load_locale (struct loaded_l10nfile *file, int category) + newdata->alloc = alloc; + + file->data = newdata; ++ file->decided = 1; + } + + void diff --git a/SOURCES/glibc-upstream-2.39-1.patch b/SOURCES/glibc-upstream-2.39-1.patch new file mode 100644 index 0000000..487c192 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-1.patch @@ -0,0 +1,264 @@ +commit 6d1e3fb07b45e2e31e469b16cf21b24bccf8914c +Author: Andreas K. Hüttel +Date: Wed Jan 31 02:12:43 2024 +0100 + + Replace advisories directory + + Signed-off-by: Andreas K. Hüttel + +diff --git a/ADVISORIES b/ADVISORIES +new file mode 100644 +index 0000000000000000..d4e33f2df3d74cd8 +--- /dev/null ++++ b/ADVISORIES +@@ -0,0 +1,2 @@ ++For the GNU C Library Security Advisories, see the git master branch: ++https://sourceware.org/git/?p=glibc.git;a=tree;f=advisories;hb=HEAD +diff --git a/advisories/GLIBC-SA-2023-0001 b/advisories/GLIBC-SA-2023-0001 +deleted file mode 100644 +index 3d19c91b6a676ffd..0000000000000000 +--- a/advisories/GLIBC-SA-2023-0001 ++++ /dev/null +@@ -1,14 +0,0 @@ +-printf: incorrect output for integers with thousands separator and width field +- +-When the printf family of functions is called with a format specifier +-that uses an (enable grouping) and a minimum width +-specifier, the resulting output could be larger than reasonably expected +-by a caller that computed a tight bound on the buffer size. The +-resulting larger than expected output could result in a buffer overflow +-in the printf family of functions. +- +-CVE-Id: CVE-2023-25139 +-Public-Date: 2023-02-02 +-Vulnerable-Commit: e88b9f0e5cc50cab57a299dc7efe1a4eb385161d (2.37) +-Fix-Commit: c980549cc6a1c03c23cc2fe3e7b0fe626a0364b0 (2.38) +-Fix-Commit: 07b9521fc6369d000216b96562ff7c0ed32a16c4 (2.37-4) +diff --git a/advisories/GLIBC-SA-2023-0002 b/advisories/GLIBC-SA-2023-0002 +deleted file mode 100644 +index 5122669a6451f803..0000000000000000 +--- a/advisories/GLIBC-SA-2023-0002 ++++ /dev/null +@@ -1,15 +0,0 @@ +-getaddrinfo: Stack read overflow in no-aaaa mode +- +-If the system is configured in no-aaaa mode via /etc/resolv.conf, +-getaddrinfo is called for the AF_UNSPEC address family, and a DNS +-response is received over TCP that is larger than 2048 bytes, +-getaddrinfo may potentially disclose stack contents via the returned +-address data, or crash. +- +-CVE-Id: CVE-2023-4527 +-Public-Date: 2023-09-12 +-Vulnerable-Commit: f282cdbe7f436c75864e5640a409a10485e9abb2 (2.36) +-Fix-Commit: bd77dd7e73e3530203be1c52c8a29d08270cb25d (2.39) +-Fix-Commit: 4ea972b7edd7e36610e8cde18bf7a8149d7bac4f (2.36-113) +-Fix-Commit: b7529346025a130fee483d42178b5c118da971bb (2.37-38) +-Fix-Commit: b25508dd774b617f99419bdc3cf2ace4560cd2d6 (2.38-19) +diff --git a/advisories/GLIBC-SA-2023-0003 b/advisories/GLIBC-SA-2023-0003 +deleted file mode 100644 +index d3aef803480bc298..0000000000000000 +--- a/advisories/GLIBC-SA-2023-0003 ++++ /dev/null +@@ -1,15 +0,0 @@ +-getaddrinfo: Potential use-after-free +- +-When an NSS plugin only implements the _gethostbyname2_r and +-_getcanonname_r callbacks, getaddrinfo could use memory that was freed +-during buffer resizing, potentially causing a crash or read or write to +-arbitrary memory. +- +-CVE-Id: CVE-2023-4806 +-Public-Date: 2023-09-12 +-Fix-Commit: 973fe93a5675c42798b2161c6f29c01b0e243994 (2.39) +-Fix-Commit: e09ee267c03e3150c2c9ba28625ab130705a485e (2.34-420) +-Fix-Commit: e3ccb230a961b4797510e6a1f5f21fd9021853e7 (2.35-270) +-Fix-Commit: a9728f798ec7f05454c95637ee6581afaa9b487d (2.36-115) +-Fix-Commit: 6529a7466c935f36e9006b854d6f4e1d4876f942 (2.37-39) +-Fix-Commit: 00ae4f10b504bc4564e9f22f00907093f1ab9338 (2.38-20) +diff --git a/advisories/GLIBC-SA-2023-0004 b/advisories/GLIBC-SA-2023-0004 +deleted file mode 100644 +index 5286a7aa545cd942..0000000000000000 +--- a/advisories/GLIBC-SA-2023-0004 ++++ /dev/null +@@ -1,16 +0,0 @@ +-tunables: local privilege escalation through buffer overflow +- +-If a tunable of the form NAME=NAME=VAL is passed in the environment of a +-setuid program and NAME is valid, it may result in a buffer overflow, +-which could be exploited to achieve escalated privileges. This flaw was +-introduced in glibc 2.34. +- +-CVE-Id: CVE-2023-4911 +-Public-Date: 2023-10-03 +-Vulnerable-Commit: 2ed18c5b534d9e92fc006202a5af0df6b72e7aca (2.34) +-Fix-Commit: 1056e5b4c3f2d90ed2b4a55f96add28da2f4c8fa (2.39) +-Fix-Commit: dcc367f148bc92e7f3778a125f7a416b093964d9 (2.34-423) +-Fix-Commit: c84018a05aec80f5ee6f682db0da1130b0196aef (2.35-274) +-Fix-Commit: 22955ad85186ee05834e47e665056148ca07699c (2.36-118) +-Fix-Commit: b4e23c75aea756b4bddc4abcf27a1c6dca8b6bd3 (2.37-45) +-Fix-Commit: 750a45a783906a19591fb8ff6b7841470f1f5701 (2.38-27) +diff --git a/advisories/GLIBC-SA-2023-0005 b/advisories/GLIBC-SA-2023-0005 +deleted file mode 100644 +index cc4eb90b8283dc14..0000000000000000 +--- a/advisories/GLIBC-SA-2023-0005 ++++ /dev/null +@@ -1,18 +0,0 @@ +-getaddrinfo: DoS due to memory leak +- +-The fix for CVE-2023-4806 introduced a memory leak when an application +-calls getaddrinfo for AF_INET6 with AI_CANONNAME, AI_ALL and AI_V4MAPPED +-flags set. +- +-CVE-Id: CVE-2023-5156 +-Public-Date: 2023-09-25 +-Vulnerable-Commit: e09ee267c03e3150c2c9ba28625ab130705a485e (2.34-420) +-Vulnerable-Commit: e3ccb230a961b4797510e6a1f5f21fd9021853e7 (2.35-270) +-Vulnerable-Commit: a9728f798ec7f05454c95637ee6581afaa9b487d (2.36-115) +-Vulnerable-Commit: 6529a7466c935f36e9006b854d6f4e1d4876f942 (2.37-39) +-Vulnerable-Commit: 00ae4f10b504bc4564e9f22f00907093f1ab9338 (2.38-20) +-Fix-Commit: 8006457ab7e1cd556b919f477348a96fe88f2e49 (2.34-421) +-Fix-Commit: 17092c0311f954e6f3c010f73ce3a78c24ac279a (2.35-272) +-Fix-Commit: 856bac55f98dc840e7c27cfa82262b933385de90 (2.36-116) +-Fix-Commit: 4473d1b87d04b25cdd0e0354814eeaa421328268 (2.37-42) +-Fix-Commit: 5ee59ca371b99984232d7584fe2b1a758b4421d3 (2.38-24) +diff --git a/advisories/GLIBC-SA-2024-0001 b/advisories/GLIBC-SA-2024-0001 +deleted file mode 100644 +index 28931c75ae018cc3..0000000000000000 +--- a/advisories/GLIBC-SA-2024-0001 ++++ /dev/null +@@ -1,15 +0,0 @@ +-syslog: Heap buffer overflow in __vsyslog_internal +- +-__vsyslog_internal did not handle a case where printing a SYSLOG_HEADER +-containing a long program name failed to update the required buffer +-size, leading to the allocation and overflow of a too-small buffer on +-the heap. +- +-CVE-Id: CVE-2023-6246 +-Public-Date: 2024-01-30 +-Vulnerable-Commit: 52a5be0df411ef3ff45c10c7c308cb92993d15b1 (2.37) +-Fix-Commit: 6bd0e4efcc78f3c0115e5ea9739a1642807450da (2.39) +-Fix-Commit: 23514c72b780f3da097ecf33a793b7ba9c2070d2 (2.38-42) +-Fix-Commit: 97a4292aa4a2642e251472b878d0ec4c46a0e59a (2.37-57) +-Vulnerable-Commit: b0e7888d1fa2dbd2d9e1645ec8c796abf78880b9 (2.36-16) +-Fix-Commit: d1a83b6767f68b3cb5b4b4ea2617254acd040c82 (2.36-126) +diff --git a/advisories/GLIBC-SA-2024-0002 b/advisories/GLIBC-SA-2024-0002 +deleted file mode 100644 +index 940bfcf2fcb76050..0000000000000000 +--- a/advisories/GLIBC-SA-2024-0002 ++++ /dev/null +@@ -1,15 +0,0 @@ +-syslog: Heap buffer overflow in __vsyslog_internal +- +-__vsyslog_internal used the return value of snprintf/vsnprintf to +-calculate buffer sizes for memory allocation. If these functions (for +-any reason) failed and returned -1, the resulting buffer would be too +-small to hold output. +- +-CVE-Id: CVE-2023-6779 +-Public-Date: 2024-01-30 +-Vulnerable-Commit: 52a5be0df411ef3ff45c10c7c308cb92993d15b1 (2.37) +-Fix-Commit: 7e5a0c286da33159d47d0122007aac016f3e02cd (2.39) +-Fix-Commit: d0338312aace5bbfef85e03055e1212dd0e49578 (2.38-43) +-Fix-Commit: 67062eccd9a65d7fda9976a56aeaaf6c25a80214 (2.37-58) +-Vulnerable-Commit: b0e7888d1fa2dbd2d9e1645ec8c796abf78880b9 (2.36-16) +-Fix-Commit: 2bc9d7c002bdac38b5c2a3f11b78e309d7765b83 (2.36-127) +diff --git a/advisories/GLIBC-SA-2024-0003 b/advisories/GLIBC-SA-2024-0003 +deleted file mode 100644 +index b43a5150ab1b0cc4..0000000000000000 +--- a/advisories/GLIBC-SA-2024-0003 ++++ /dev/null +@@ -1,13 +0,0 @@ +-syslog: Integer overflow in __vsyslog_internal +- +-__vsyslog_internal calculated a buffer size by adding two integers, but +-did not first check if the addition would overflow. +- +-CVE-Id: CVE-2023-6780 +-Public-Date: 2024-01-30 +-Vulnerable-Commit: 52a5be0df411ef3ff45c10c7c308cb92993d15b1 (2.37) +-Fix-Commit: ddf542da94caf97ff43cc2875c88749880b7259b (2.39) +-Fix-Commit: d37c2b20a4787463d192b32041c3406c2bd91de0 (2.38-44) +-Fix-Commit: 2b58cba076e912961ceaa5fa58588e4b10f791c0 (2.37-59) +-Vulnerable-Commit: b0e7888d1fa2dbd2d9e1645ec8c796abf78880b9 (2.36-16) +-Fix-Commit: b9b7d6a27aa0632f334352fa400771115b3c69b7 (2.36-128) +diff --git a/advisories/README b/advisories/README +deleted file mode 100644 +index 94e68b1350bbad0e..0000000000000000 +--- a/advisories/README ++++ /dev/null +@@ -1,73 +0,0 @@ +-GNU C Library Security Advisory Format +-====================================== +- +-Security advisories in this directory follow a simple git commit log +-format, with a heading and free-format description augmented with tags +-to allow parsing key information. References to code changes are +-specific to the glibc repository and follow a specific format: +- +- Tag-name: (release-version) +- +-The indicates a specific commit in the repository. The +-release-version indicates the publicly consumable release in which this +-commit is known to exist. The release-version is derived from the +-git-describe format, (i.e. stripped out from glibc-2.34.NNN-gxxxx) and +-is of the form 2.34-NNN. If the -NNN suffix is absent, it means that +-the change is in that release tarball, otherwise the change is on the +-release/2.YY/master branch and not in any released tarball. +- +-The following tags are currently being used: +- +-CVE-Id: +-This is the CVE-Id assigned under the CVE Program +-(https://www.cve.org/). +- +-Public-Date: +-The date this issue became publicly known. +- +-Vulnerable-Commit: +-The commit that introduced this vulnerability. There could be multiple +-entries, one for each release branch in the glibc repository; the +-release-version portion of this tag should tell you which branch this is +-on. +- +-Fix-Commit: +-The commit that fixed this vulnerability. There could be multiple +-entries for each release branch in the glibc repository, indicating that +-all of those commits contributed to fixing that issue in each of those +-branches. +- +-Adding an Advisory +------------------- +- +-An advisory for a CVE needs to be added on the master branch in two steps: +- +-1. Add the text of the advisory without any Fix-Commit tags along with +- the fix for the CVE. Add the Vulnerable-Commit tag, if applicable. +- The advisories directory does not exist in release branches, so keep +- the advisory text commit distinct from the code changes, to ease +- backports. Ask for the GLIBC-SA advisory number from the security +- team. +- +-2. Finish all backports on release branches and then back on the msater +- branch, add all commit refs to the advisory using the Fix-Commit +- tags. Don't bother adding the release-version subscript since the +- next step will overwrite it. +- +-3. Run the process-advisories.sh script in the scripts directory on the +- advisory: +- +- scripts/process-advisories.sh update GLIBC-SA-YYYY-NNNN +- +- (replace YYYY-NNNN with the actual advisory number). +- +-4. Verify the updated advisory and push the result. +- +-Getting a NEWS snippet from advisories +--------------------------------------- +- +-Run: +- +- scripts/process-advisories.sh news +- +-and copy the content into the NEWS file. diff --git a/SOURCES/glibc-upstream-2.39-10.patch b/SOURCES/glibc-upstream-2.39-10.patch new file mode 100644 index 0000000..858f28e --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-10.patch @@ -0,0 +1,16 @@ +commit 983f34a1252de3ca6f2305c211d86530ea42010e +Author: caiyinyu +Date: Mon Mar 11 16:07:48 2024 +0800 + + LoongArch: Correct {__ieee754, _}_scalb -> {__ieee754, _}_scalbf + +diff --git a/sysdeps/loongarch/fpu/e_scalbf.c b/sysdeps/loongarch/fpu/e_scalbf.c +index 9f054852362e2d76..7c0395fbb5afbc58 100644 +--- a/sysdeps/loongarch/fpu/e_scalbf.c ++++ b/sysdeps/loongarch/fpu/e_scalbf.c +@@ -57,4 +57,4 @@ __ieee754_scalbf (float x, float fn) + + return x; + } +-libm_alias_finite (__ieee754_scalb, __scalb) ++libm_alias_finite (__ieee754_scalbf, __scalbf) diff --git a/SOURCES/glibc-upstream-2.39-11.patch b/SOURCES/glibc-upstream-2.39-11.patch new file mode 100644 index 0000000..d816c3c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-11.patch @@ -0,0 +1,80 @@ +commit 7fc8242bf87828c935ac5df5cafb9dc7ab635fd9 +Author: H.J. Lu +Date: Fri Feb 16 07:17:10 2024 -0800 + + x86-64: Save APX registers in ld.so trampoline + + Add APX registers to STATE_SAVE_MASK so that APX registers are saved in + ld.so trampoline. This fixes BZ #31371. + + Also update STATE_SAVE_OFFSET and STATE_SAVE_MASK for i386 which will + be used by i386 _dl_tlsdesc_dynamic. + Reviewed-by: Noah Goldstein + + (cherry picked from commit dfb05f8e704edac70db38c4c8ee700769d91a413) + +diff --git a/sysdeps/x86/sysdep.h b/sysdeps/x86/sysdep.h +index 85d0a8c943cbb218..837fd28734914a1c 100644 +--- a/sysdeps/x86/sysdep.h ++++ b/sysdeps/x86/sysdep.h +@@ -21,14 +21,54 @@ + + #include + ++/* The extended state feature IDs in the state component bitmap. */ ++#define X86_XSTATE_X87_ID 0 ++#define X86_XSTATE_SSE_ID 1 ++#define X86_XSTATE_AVX_ID 2 ++#define X86_XSTATE_BNDREGS_ID 3 ++#define X86_XSTATE_BNDCFG_ID 4 ++#define X86_XSTATE_K_ID 5 ++#define X86_XSTATE_ZMM_H_ID 6 ++#define X86_XSTATE_ZMM_ID 7 ++#define X86_XSTATE_PKRU_ID 9 ++#define X86_XSTATE_TILECFG_ID 17 ++#define X86_XSTATE_TILEDATA_ID 18 ++#define X86_XSTATE_APX_F_ID 19 ++ ++#ifdef __x86_64__ + /* Offset for fxsave/xsave area used by _dl_runtime_resolve. Also need + space to preserve RCX, RDX, RSI, RDI, R8, R9 and RAX. It must be +- aligned to 16 bytes for fxsave and 64 bytes for xsave. */ +-#define STATE_SAVE_OFFSET (8 * 7 + 8) +- +-/* Save SSE, AVX, AVX512, mask and bound registers. */ +-#define STATE_SAVE_MASK \ +- ((1 << 1) | (1 << 2) | (1 << 3) | (1 << 5) | (1 << 6) | (1 << 7)) ++ aligned to 16 bytes for fxsave and 64 bytes for xsave. ++ ++ NB: Is is non-zero because of the 128-byte red-zone. Some registers ++ are saved on stack without adjusting stack pointer first. When we ++ update stack pointer to allocate more space, we need to take the ++ red-zone into account. */ ++# define STATE_SAVE_OFFSET (8 * 7 + 8) ++ ++/* Save SSE, AVX, AVX512, mask, bound and APX registers. Bound and APX ++ registers are mutually exclusive. */ ++# define STATE_SAVE_MASK \ ++ ((1 << X86_XSTATE_SSE_ID) \ ++ | (1 << X86_XSTATE_AVX_ID) \ ++ | (1 << X86_XSTATE_BNDREGS_ID) \ ++ | (1 << X86_XSTATE_K_ID) \ ++ | (1 << X86_XSTATE_ZMM_H_ID) \ ++ | (1 << X86_XSTATE_ZMM_ID) \ ++ | (1 << X86_XSTATE_APX_F_ID)) ++#else ++/* Offset for fxsave/xsave area used by _dl_tlsdesc_dynamic. Since i386 ++ doesn't have red-zone, use 0 here. */ ++# define STATE_SAVE_OFFSET 0 ++ ++/* Save SSE, AVX, AXV512, mask and bound registers. */ ++# define STATE_SAVE_MASK \ ++ ((1 << X86_XSTATE_SSE_ID) \ ++ | (1 << X86_XSTATE_AVX_ID) \ ++ | (1 << X86_XSTATE_BNDREGS_ID) \ ++ | (1 << X86_XSTATE_K_ID) \ ++ | (1 << X86_XSTATE_ZMM_H_ID)) ++#endif + + /* Constants for bits in __x86_string_control: */ + diff --git a/SOURCES/glibc-upstream-2.39-12.patch b/SOURCES/glibc-upstream-2.39-12.patch new file mode 100644 index 0000000..74e0b72 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-12.patch @@ -0,0 +1,1453 @@ +commit a364304718725a31ab141936322855c76c73e35e +Author: H.J. Lu +Date: Mon Feb 26 06:37:03 2024 -0800 + + x86: Update _dl_tlsdesc_dynamic to preserve caller-saved registers + + Compiler generates the following instruction sequence for GNU2 dynamic + TLS access: + + leaq tls_var@TLSDESC(%rip), %rax + call *tls_var@TLSCALL(%rax) + + or + + leal tls_var@TLSDESC(%ebx), %eax + call *tls_var@TLSCALL(%eax) + + CALL instruction is transparent to compiler which assumes all registers, + except for EFLAGS and RAX/EAX, are unchanged after CALL. When + _dl_tlsdesc_dynamic is called, it calls __tls_get_addr on the slow + path. __tls_get_addr is a normal function which doesn't preserve any + caller-saved registers. _dl_tlsdesc_dynamic saved and restored integer + caller-saved registers, but didn't preserve any other caller-saved + registers. Add _dl_tlsdesc_dynamic IFUNC functions for FNSAVE, FXSAVE, + XSAVE and XSAVEC to save and restore all caller-saved registers. This + fixes BZ #31372. + + Add GLRO(dl_x86_64_runtime_resolve) with GLRO(dl_x86_tlsdesc_dynamic) + to optimize elf_machine_runtime_setup. + Reviewed-by: Noah Goldstein + + (cherry picked from commit 0aac205a814a8511e98d02b91a8dc908f1c53cde) + +diff --git a/elf/Makefile b/elf/Makefile +index 5d78b659ce813ff9..c5c37a9147e69d83 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -424,6 +424,7 @@ tests += \ + tst-glibc-hwcaps-prepend \ + tst-global1 \ + tst-global2 \ ++ tst-gnu2-tls2 \ + tst-initfinilazyfail \ + tst-initorder \ + tst-initorder2 \ +@@ -846,6 +847,9 @@ modules-names += \ + tst-filterobj-flt \ + tst-finilazyfailmod \ + tst-globalmod2 \ ++ tst-gnu2-tls2mod0 \ ++ tst-gnu2-tls2mod1 \ ++ tst-gnu2-tls2mod2 \ + tst-initlazyfailmod \ + tst-initorder2a \ + tst-initorder2b \ +@@ -3044,8 +3048,22 @@ $(objpfx)tst-tlsgap.out: \ + $(objpfx)tst-tlsgap-mod0.so \ + $(objpfx)tst-tlsgap-mod1.so \ + $(objpfx)tst-tlsgap-mod2.so ++ ++$(objpfx)tst-gnu2-tls2: $(shared-thread-library) ++$(objpfx)tst-gnu2-tls2.out: \ ++ $(objpfx)tst-gnu2-tls2mod0.so \ ++ $(objpfx)tst-gnu2-tls2mod1.so \ ++ $(objpfx)tst-gnu2-tls2mod2.so ++ + ifeq (yes,$(have-mtls-dialect-gnu2)) ++# This test fails if dl_tlsdesc_dynamic doesn't preserve all caller-saved ++# registers. See https://sourceware.org/bugzilla/show_bug.cgi?id=31372 ++test-xfail-tst-gnu2-tls2 = yes ++ + CFLAGS-tst-tlsgap-mod0.c += -mtls-dialect=gnu2 + CFLAGS-tst-tlsgap-mod1.c += -mtls-dialect=gnu2 + CFLAGS-tst-tlsgap-mod2.c += -mtls-dialect=gnu2 ++CFLAGS-tst-gnu2-tls2mod0.c += -mtls-dialect=gnu2 ++CFLAGS-tst-gnu2-tls2mod1.c += -mtls-dialect=gnu2 ++CFLAGS-tst-gnu2-tls2mod2.c += -mtls-dialect=gnu2 + endif +diff --git a/elf/tst-gnu2-tls2.c b/elf/tst-gnu2-tls2.c +new file mode 100644 +index 0000000000000000..7ac04d7f3312033e +--- /dev/null ++++ b/elf/tst-gnu2-tls2.c +@@ -0,0 +1,122 @@ ++/* Test TLSDESC relocation. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include "tst-gnu2-tls2.h" ++ ++#ifndef IS_SUPPORTED ++# define IS_SUPPORTED() true ++#endif ++ ++/* An architecture can define it to clobber caller-saved registers in ++ malloc below to verify that the implicit TLSDESC call won't change ++ caller-saved registers. */ ++#ifndef PREPARE_MALLOC ++# define PREPARE_MALLOC() ++#endif ++ ++extern void * __libc_malloc (size_t); ++ ++size_t malloc_counter = 0; ++ ++void * ++malloc (size_t n) ++{ ++ PREPARE_MALLOC (); ++ malloc_counter++; ++ return __libc_malloc (n); ++} ++ ++static void *mod[3]; ++#ifndef MOD ++# define MOD(i) "tst-gnu2-tls2mod" #i ".so" ++#endif ++static const char *modname[3] = { MOD(0), MOD(1), MOD(2) }; ++#undef MOD ++ ++static void ++open_mod (int i) ++{ ++ mod[i] = xdlopen (modname[i], RTLD_LAZY); ++ printf ("open %s\n", modname[i]); ++} ++ ++static void ++close_mod (int i) ++{ ++ xdlclose (mod[i]); ++ mod[i] = NULL; ++ printf ("close %s\n", modname[i]); ++} ++ ++static void ++access_mod (int i, const char *sym) ++{ ++ struct tls var = { -1, -1, -1, -1 }; ++ struct tls *(*f) (struct tls *) = xdlsym (mod[i], sym); ++ /* Check that our malloc is called. */ ++ malloc_counter = 0; ++ struct tls *p = f (&var); ++ TEST_VERIFY (malloc_counter != 0); ++ printf ("access %s: %s() = %p\n", modname[i], sym, p); ++ TEST_VERIFY_EXIT (memcmp (p, &var, sizeof (var)) == 0); ++ ++(p->a); ++} ++ ++static void * ++start (void *arg) ++{ ++ /* The DTV generation is at the last dlopen of mod0 and the ++ entry for mod1 is NULL. */ ++ ++ open_mod (1); /* Reuse modid of mod1. Uses dynamic TLS. */ ++ ++ /* Force the slow path in GNU2 TLS descriptor call. */ ++ access_mod (1, "apply_tls"); ++ ++ return arg; ++} ++ ++static int ++do_test (void) ++{ ++ if (!IS_SUPPORTED ()) ++ return EXIT_UNSUPPORTED; ++ ++ open_mod (0); ++ open_mod (1); ++ open_mod (2); ++ close_mod (0); ++ close_mod (1); /* Create modid gap at mod1. */ ++ open_mod (0); /* Reuse modid of mod0, bump generation count. */ ++ ++ /* Create a thread where DTV of mod1 is NULL. */ ++ pthread_t t = xpthread_create (NULL, start, NULL); ++ xpthread_join (t); ++ return 0; ++} ++ ++#include +diff --git a/elf/tst-gnu2-tls2.h b/elf/tst-gnu2-tls2.h +new file mode 100644 +index 0000000000000000..77964a57a352e6a4 +--- /dev/null ++++ b/elf/tst-gnu2-tls2.h +@@ -0,0 +1,36 @@ ++/* Test TLSDESC relocation. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++ ++struct tls ++{ ++ int64_t a, b, c, d; ++}; ++ ++extern struct tls *apply_tls (struct tls *); ++ ++/* An architecture can define them to verify that clobber caller-saved ++ registers aren't changed by the implicit TLSDESC call. */ ++#ifndef BEFORE_TLSDESC_CALL ++# define BEFORE_TLSDESC_CALL() ++#endif ++ ++#ifndef AFTER_TLSDESC_CALL ++# define AFTER_TLSDESC_CALL() ++#endif +diff --git a/elf/tst-gnu2-tls2mod0.c b/elf/tst-gnu2-tls2mod0.c +new file mode 100644 +index 0000000000000000..45556a0e173922cc +--- /dev/null ++++ b/elf/tst-gnu2-tls2mod0.c +@@ -0,0 +1,31 @@ ++/* DSO used by tst-gnu2-tls2. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include "tst-gnu2-tls2.h" ++ ++__thread struct tls tls_var0 __attribute__ ((visibility ("hidden"))); ++ ++struct tls * ++apply_tls (struct tls *p) ++{ ++ BEFORE_TLSDESC_CALL (); ++ tls_var0 = *p; ++ struct tls *ret = &tls_var0; ++ AFTER_TLSDESC_CALL (); ++ return ret; ++} +diff --git a/elf/tst-gnu2-tls2mod1.c b/elf/tst-gnu2-tls2mod1.c +new file mode 100644 +index 0000000000000000..e10b9dbc0a7573c7 +--- /dev/null ++++ b/elf/tst-gnu2-tls2mod1.c +@@ -0,0 +1,31 @@ ++/* DSO used by tst-gnu2-tls2. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include "tst-gnu2-tls2.h" ++ ++__thread struct tls tls_var1[100] __attribute__ ((visibility ("hidden"))); ++ ++struct tls * ++apply_tls (struct tls *p) ++{ ++ BEFORE_TLSDESC_CALL (); ++ tls_var1[1] = *p; ++ struct tls *ret = &tls_var1[1]; ++ AFTER_TLSDESC_CALL (); ++ return ret; ++} +diff --git a/elf/tst-gnu2-tls2mod2.c b/elf/tst-gnu2-tls2mod2.c +new file mode 100644 +index 0000000000000000..141af51e55b8bf34 +--- /dev/null ++++ b/elf/tst-gnu2-tls2mod2.c +@@ -0,0 +1,31 @@ ++/* DSO used by tst-gnu2-tls2. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include "tst-gnu2-tls2.h" ++ ++__thread struct tls tls_var2 __attribute__ ((visibility ("hidden"))); ++ ++struct tls * ++apply_tls (struct tls *p) ++{ ++ BEFORE_TLSDESC_CALL (); ++ tls_var2 = *p; ++ struct tls *ret = &tls_var2; ++ AFTER_TLSDESC_CALL (); ++ return ret; ++} +diff --git a/sysdeps/i386/dl-machine.h b/sysdeps/i386/dl-machine.h +index fc1ef96587e1992e..50d74fe6e9aaa7bb 100644 +--- a/sysdeps/i386/dl-machine.h ++++ b/sysdeps/i386/dl-machine.h +@@ -347,7 +347,7 @@ and creates an unsatisfiable circular dependency.\n", + { + td->arg = _dl_make_tlsdesc_dynamic + (sym_map, sym->st_value + (ElfW(Word))td->arg); +- td->entry = _dl_tlsdesc_dynamic; ++ td->entry = GLRO(dl_x86_tlsdesc_dynamic); + } + else + # endif +diff --git a/sysdeps/i386/dl-tlsdesc-dynamic.h b/sysdeps/i386/dl-tlsdesc-dynamic.h +new file mode 100644 +index 0000000000000000..36270285775016ff +--- /dev/null ++++ b/sysdeps/i386/dl-tlsdesc-dynamic.h +@@ -0,0 +1,190 @@ ++/* Thread-local storage handling in the ELF dynamic linker. i386 version. ++ Copyright (C) 2004-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#undef REGISTER_SAVE_AREA ++ ++#if !defined USE_FNSAVE && (STATE_SAVE_ALIGNMENT % 16) != 0 ++# error STATE_SAVE_ALIGNMENT must be multiple of 16 ++#endif ++ ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++# ifdef USE_FNSAVE ++# error USE_FNSAVE shouldn't be defined ++# endif ++# ifdef USE_FXSAVE ++/* Use fxsave to save all registers. */ ++# define REGISTER_SAVE_AREA 512 ++# endif ++#else ++# ifdef USE_FNSAVE ++/* Use fnsave to save x87 FPU stack registers. */ ++# define REGISTER_SAVE_AREA 108 ++# else ++# ifndef USE_FXSAVE ++# error USE_FXSAVE must be defined ++# endif ++/* Use fxsave to save all registers. Add 12 bytes to align the stack ++ to 16 bytes. */ ++# define REGISTER_SAVE_AREA (512 + 12) ++# endif ++#endif ++ ++ .hidden _dl_tlsdesc_dynamic ++ .global _dl_tlsdesc_dynamic ++ .type _dl_tlsdesc_dynamic,@function ++ ++ /* This function is used for symbols that need dynamic TLS. ++ ++ %eax points to the TLS descriptor, such that 0(%eax) points to ++ _dl_tlsdesc_dynamic itself, and 4(%eax) points to a struct ++ tlsdesc_dynamic_arg object. It must return in %eax the offset ++ between the thread pointer and the object denoted by the ++ argument, without clobbering any registers. ++ ++ The assembly code that follows is a rendition of the following ++ C code, hand-optimized a little bit. ++ ++ptrdiff_t ++__attribute__ ((__regparm__ (1))) ++_dl_tlsdesc_dynamic (struct tlsdesc *tdp) ++{ ++ struct tlsdesc_dynamic_arg *td = tdp->arg; ++ dtv_t *dtv = *(dtv_t **)((char *)__thread_pointer + DTV_OFFSET); ++ if (__builtin_expect (td->gen_count <= dtv[0].counter ++ && (dtv[td->tlsinfo.ti_module].pointer.val ++ != TLS_DTV_UNALLOCATED), ++ 1)) ++ return dtv[td->tlsinfo.ti_module].pointer.val + td->tlsinfo.ti_offset ++ - __thread_pointer; ++ ++ return ___tls_get_addr (&td->tlsinfo) - __thread_pointer; ++} ++*/ ++ cfi_startproc ++ .align 16 ++_dl_tlsdesc_dynamic: ++ /* Like all TLS resolvers, preserve call-clobbered registers. ++ We need two scratch regs anyway. */ ++ subl $32, %esp ++ cfi_adjust_cfa_offset (32) ++ movl %ecx, 20(%esp) ++ movl %edx, 24(%esp) ++ movl TLSDESC_ARG(%eax), %eax ++ movl %gs:DTV_OFFSET, %edx ++ movl TLSDESC_GEN_COUNT(%eax), %ecx ++ cmpl (%edx), %ecx ++ ja 2f ++ movl TLSDESC_MODID(%eax), %ecx ++ movl (%edx,%ecx,8), %edx ++ cmpl $-1, %edx ++ je 2f ++ movl TLSDESC_MODOFF(%eax), %eax ++ addl %edx, %eax ++1: ++ movl 20(%esp), %ecx ++ subl %gs:0, %eax ++ movl 24(%esp), %edx ++ addl $32, %esp ++ cfi_adjust_cfa_offset (-32) ++ ret ++ .p2align 4,,7 ++2: ++ cfi_adjust_cfa_offset (32) ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++ movl %ebx, -28(%esp) ++ movl %esp, %ebx ++ cfi_def_cfa_register(%ebx) ++ and $-STATE_SAVE_ALIGNMENT, %esp ++#endif ++#ifdef REGISTER_SAVE_AREA ++ subl $REGISTER_SAVE_AREA, %esp ++# if !DL_RUNTIME_RESOLVE_REALIGN_STACK ++ cfi_adjust_cfa_offset(REGISTER_SAVE_AREA) ++# endif ++#else ++# if !DL_RUNTIME_RESOLVE_REALIGN_STACK ++# error DL_RUNTIME_RESOLVE_REALIGN_STACK must be true ++# endif ++ /* Allocate stack space of the required size to save the state. */ ++ LOAD_PIC_REG (cx) ++ subl RTLD_GLOBAL_RO_DL_X86_CPU_FEATURES_OFFSET+XSAVE_STATE_SIZE_OFFSET+_rtld_local_ro@GOTOFF(%ecx), %esp ++#endif ++#ifdef USE_FNSAVE ++ fnsave (%esp) ++#elif defined USE_FXSAVE ++ fxsave (%esp) ++#else ++ /* Save the argument for ___tls_get_addr in EAX. */ ++ movl %eax, %ecx ++ movl $TLSDESC_CALL_STATE_SAVE_MASK, %eax ++ xorl %edx, %edx ++ /* Clear the XSAVE Header. */ ++# ifdef USE_XSAVE ++ movl %edx, (512)(%esp) ++ movl %edx, (512 + 4 * 1)(%esp) ++ movl %edx, (512 + 4 * 2)(%esp) ++ movl %edx, (512 + 4 * 3)(%esp) ++# endif ++ movl %edx, (512 + 4 * 4)(%esp) ++ movl %edx, (512 + 4 * 5)(%esp) ++ movl %edx, (512 + 4 * 6)(%esp) ++ movl %edx, (512 + 4 * 7)(%esp) ++ movl %edx, (512 + 4 * 8)(%esp) ++ movl %edx, (512 + 4 * 9)(%esp) ++ movl %edx, (512 + 4 * 10)(%esp) ++ movl %edx, (512 + 4 * 11)(%esp) ++ movl %edx, (512 + 4 * 12)(%esp) ++ movl %edx, (512 + 4 * 13)(%esp) ++ movl %edx, (512 + 4 * 14)(%esp) ++ movl %edx, (512 + 4 * 15)(%esp) ++# ifdef USE_XSAVE ++ xsave (%esp) ++# else ++ xsavec (%esp) ++# endif ++ /* Restore the argument for ___tls_get_addr in EAX. */ ++ movl %ecx, %eax ++#endif ++ call HIDDEN_JUMPTARGET (___tls_get_addr) ++ /* Get register content back. */ ++#ifdef USE_FNSAVE ++ frstor (%esp) ++#elif defined USE_FXSAVE ++ fxrstor (%esp) ++#else ++ /* Save and retore ___tls_get_addr return value stored in EAX. */ ++ movl %eax, %ecx ++ movl $TLSDESC_CALL_STATE_SAVE_MASK, %eax ++ xorl %edx, %edx ++ xrstor (%esp) ++ movl %ecx, %eax ++#endif ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++ mov %ebx, %esp ++ cfi_def_cfa_register(%esp) ++ movl -28(%esp), %ebx ++ cfi_restore(%ebx) ++#else ++ addl $REGISTER_SAVE_AREA, %esp ++ cfi_adjust_cfa_offset(-REGISTER_SAVE_AREA) ++#endif ++ jmp 1b ++ cfi_endproc ++ .size _dl_tlsdesc_dynamic, .-_dl_tlsdesc_dynamic ++ ++#undef STATE_SAVE_ALIGNMENT +diff --git a/sysdeps/i386/dl-tlsdesc.S b/sysdeps/i386/dl-tlsdesc.S +index 90d93caa0cdb7442..f002feee56e43f71 100644 +--- a/sysdeps/i386/dl-tlsdesc.S ++++ b/sysdeps/i386/dl-tlsdesc.S +@@ -18,8 +18,27 @@ + + #include + #include ++#include ++#include + #include "tlsdesc.h" + ++#ifndef DL_STACK_ALIGNMENT ++/* Due to GCC bug: ++ ++ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58066 ++ ++ __tls_get_addr may be called with 4-byte stack alignment. Although ++ this bug has been fixed in GCC 4.9.4, 5.3 and 6, we can't assume ++ that stack will be always aligned at 16 bytes. */ ++# define DL_STACK_ALIGNMENT 4 ++#endif ++ ++/* True if _dl_tlsdesc_dynamic should align stack for STATE_SAVE or align ++ stack to MINIMUM_ALIGNMENT bytes before calling ___tls_get_addr. */ ++#define DL_RUNTIME_RESOLVE_REALIGN_STACK \ ++ (STATE_SAVE_ALIGNMENT > DL_STACK_ALIGNMENT \ ++ || MINIMUM_ALIGNMENT > DL_STACK_ALIGNMENT) ++ + .text + + /* This function is used to compute the TP offset for symbols in +@@ -65,69 +84,35 @@ _dl_tlsdesc_undefweak: + .size _dl_tlsdesc_undefweak, .-_dl_tlsdesc_undefweak + + #ifdef SHARED +- .hidden _dl_tlsdesc_dynamic +- .global _dl_tlsdesc_dynamic +- .type _dl_tlsdesc_dynamic,@function +- +- /* This function is used for symbols that need dynamic TLS. +- +- %eax points to the TLS descriptor, such that 0(%eax) points to +- _dl_tlsdesc_dynamic itself, and 4(%eax) points to a struct +- tlsdesc_dynamic_arg object. It must return in %eax the offset +- between the thread pointer and the object denoted by the +- argument, without clobbering any registers. +- +- The assembly code that follows is a rendition of the following +- C code, hand-optimized a little bit. +- +-ptrdiff_t +-__attribute__ ((__regparm__ (1))) +-_dl_tlsdesc_dynamic (struct tlsdesc *tdp) +-{ +- struct tlsdesc_dynamic_arg *td = tdp->arg; +- dtv_t *dtv = *(dtv_t **)((char *)__thread_pointer + DTV_OFFSET); +- if (__builtin_expect (td->gen_count <= dtv[0].counter +- && (dtv[td->tlsinfo.ti_module].pointer.val +- != TLS_DTV_UNALLOCATED), +- 1)) +- return dtv[td->tlsinfo.ti_module].pointer.val + td->tlsinfo.ti_offset +- - __thread_pointer; +- +- return ___tls_get_addr (&td->tlsinfo) - __thread_pointer; +-} +-*/ +- cfi_startproc +- .align 16 +-_dl_tlsdesc_dynamic: +- /* Like all TLS resolvers, preserve call-clobbered registers. +- We need two scratch regs anyway. */ +- subl $28, %esp +- cfi_adjust_cfa_offset (28) +- movl %ecx, 20(%esp) +- movl %edx, 24(%esp) +- movl TLSDESC_ARG(%eax), %eax +- movl %gs:DTV_OFFSET, %edx +- movl TLSDESC_GEN_COUNT(%eax), %ecx +- cmpl (%edx), %ecx +- ja .Lslow +- movl TLSDESC_MODID(%eax), %ecx +- movl (%edx,%ecx,8), %edx +- cmpl $-1, %edx +- je .Lslow +- movl TLSDESC_MODOFF(%eax), %eax +- addl %edx, %eax +-.Lret: +- movl 20(%esp), %ecx +- subl %gs:0, %eax +- movl 24(%esp), %edx +- addl $28, %esp +- cfi_adjust_cfa_offset (-28) +- ret +- .p2align 4,,7 +-.Lslow: +- cfi_adjust_cfa_offset (28) +- call HIDDEN_JUMPTARGET (___tls_get_addr) +- jmp .Lret +- cfi_endproc +- .size _dl_tlsdesc_dynamic, .-_dl_tlsdesc_dynamic ++# define USE_FNSAVE ++# define MINIMUM_ALIGNMENT 4 ++# define STATE_SAVE_ALIGNMENT 4 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_fnsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef MINIMUM_ALIGNMENT ++# undef USE_FNSAVE ++ ++# define MINIMUM_ALIGNMENT 16 ++ ++# define USE_FXSAVE ++# define STATE_SAVE_ALIGNMENT 16 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_fxsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_FXSAVE ++ ++# define USE_XSAVE ++# define STATE_SAVE_ALIGNMENT 64 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_xsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_XSAVE ++ ++# define USE_XSAVEC ++# define STATE_SAVE_ALIGNMENT 64 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_xsavec ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_XSAVEC + #endif /* SHARED */ +diff --git a/sysdeps/x86/Makefile b/sysdeps/x86/Makefile +index 4d50b327b55ffd65..992aabe43ec60abf 100644 +--- a/sysdeps/x86/Makefile ++++ b/sysdeps/x86/Makefile +@@ -1,5 +1,5 @@ + ifeq ($(subdir),csu) +-gen-as-const-headers += cpu-features-offsets.sym ++gen-as-const-headers += cpu-features-offsets.sym features-offsets.sym + endif + + ifeq ($(subdir),elf) +@@ -86,6 +86,11 @@ endif + tst-ifunc-isa-2-ENV = GLIBC_TUNABLES=glibc.cpu.hwcaps=-SSE4_2,-AVX,-AVX2,-AVX512F + tst-ifunc-isa-2-static-ENV = $(tst-ifunc-isa-2-ENV) + tst-hwcap-tunables-ARGS = -- $(host-test-program-cmd) ++ ++CFLAGS-tst-gnu2-tls2.c += -msse ++CFLAGS-tst-gnu2-tls2mod0.c += -msse2 -mtune=haswell ++CFLAGS-tst-gnu2-tls2mod1.c += -msse2 -mtune=haswell ++CFLAGS-tst-gnu2-tls2mod2.c += -msse2 -mtune=haswell + endif + + ifeq ($(subdir),math) +diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c +index 25e6622a79bb969f..835113b42f924b83 100644 +--- a/sysdeps/x86/cpu-features.c ++++ b/sysdeps/x86/cpu-features.c +@@ -27,8 +27,13 @@ + extern void TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t *) + attribute_hidden; + +-#if defined SHARED && defined __x86_64__ +-# include ++#if defined SHARED ++extern void _dl_tlsdesc_dynamic_fxsave (void) attribute_hidden; ++extern void _dl_tlsdesc_dynamic_xsave (void) attribute_hidden; ++extern void _dl_tlsdesc_dynamic_xsavec (void) attribute_hidden; ++ ++# ifdef __x86_64__ ++# include + + static void + TUNABLE_CALLBACK (set_plt_rewrite) (tunable_val_t *valp) +@@ -47,6 +52,15 @@ TUNABLE_CALLBACK (set_plt_rewrite) (tunable_val_t *valp) + : plt_rewrite_jmp); + } + } ++# else ++extern void _dl_tlsdesc_dynamic_fnsave (void) attribute_hidden; ++# endif ++#endif ++ ++#ifdef __x86_64__ ++extern void _dl_runtime_resolve_fxsave (void) attribute_hidden; ++extern void _dl_runtime_resolve_xsave (void) attribute_hidden; ++extern void _dl_runtime_resolve_xsavec (void) attribute_hidden; + #endif + + #ifdef __LP64__ +@@ -1130,6 +1144,44 @@ no_cpuid: + TUNABLE_CALLBACK (set_x86_shstk)); + #endif + ++ if (GLRO(dl_x86_cpu_features).xsave_state_size != 0) ++ { ++ if (CPU_FEATURE_USABLE_P (cpu_features, XSAVEC)) ++ { ++#ifdef __x86_64__ ++ GLRO(dl_x86_64_runtime_resolve) = _dl_runtime_resolve_xsavec; ++#endif ++#ifdef SHARED ++ GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_xsavec; ++#endif ++ } ++ else ++ { ++#ifdef __x86_64__ ++ GLRO(dl_x86_64_runtime_resolve) = _dl_runtime_resolve_xsave; ++#endif ++#ifdef SHARED ++ GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_xsave; ++#endif ++ } ++ } ++ else ++ { ++#ifdef __x86_64__ ++ GLRO(dl_x86_64_runtime_resolve) = _dl_runtime_resolve_fxsave; ++# ifdef SHARED ++ GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; ++# endif ++#else ++# ifdef SHARED ++ if (CPU_FEATURE_USABLE_P (cpu_features, FXSR)) ++ GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; ++ else ++ GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fnsave; ++# endif ++#endif ++ } ++ + #ifdef SHARED + # ifdef __x86_64__ + TUNABLE_GET (plt_rewrite, tunable_val_t *, +diff --git a/sysdeps/x86/dl-procinfo.c b/sysdeps/x86/dl-procinfo.c +index ee957b4d70050b1c..5920d4b32034501d 100644 +--- a/sysdeps/x86/dl-procinfo.c ++++ b/sysdeps/x86/dl-procinfo.c +@@ -86,3 +86,19 @@ PROCINFO_CLASS const char _dl_x86_platforms[4][9] + #else + , + #endif ++ ++#if defined SHARED && !IS_IN (ldconfig) ++# if !defined PROCINFO_DECL ++ ._dl_x86_tlsdesc_dynamic ++# else ++PROCINFO_CLASS void * _dl_x86_tlsdesc_dynamic ++# endif ++# ifndef PROCINFO_DECL ++= NULL ++# endif ++# ifdef PROCINFO_DECL ++; ++# else ++, ++# endif ++#endif +diff --git a/sysdeps/x86_64/features-offsets.sym b/sysdeps/x86/features-offsets.sym +similarity index 89% +rename from sysdeps/x86_64/features-offsets.sym +rename to sysdeps/x86/features-offsets.sym +index 9e4be3393a60fd92..77e990c7053a38bf 100644 +--- a/sysdeps/x86_64/features-offsets.sym ++++ b/sysdeps/x86/features-offsets.sym +@@ -3,4 +3,6 @@ + #include + + RTLD_GLOBAL_RO_DL_X86_CPU_FEATURES_OFFSET offsetof (struct rtld_global_ro, _dl_x86_cpu_features) ++#ifdef __x86_64__ + RTLD_GLOBAL_DL_X86_FEATURE_1_OFFSET offsetof (struct rtld_global, _dl_x86_feature_1) ++#endif +diff --git a/sysdeps/x86/sysdep.h b/sysdeps/x86/sysdep.h +index 837fd28734914a1c..485cad9c0283b334 100644 +--- a/sysdeps/x86/sysdep.h ++++ b/sysdeps/x86/sysdep.h +@@ -70,6 +70,12 @@ + | (1 << X86_XSTATE_ZMM_H_ID)) + #endif + ++/* States which should be saved for TLSDESC_CALL and TLS_DESC_CALL. ++ Compiler assumes that all registers, including x87 FPU stack registers, ++ are unchanged after CALL, except for EFLAGS and RAX/EAX. */ ++#define TLSDESC_CALL_STATE_SAVE_MASK \ ++ (STATE_SAVE_MASK | (1 << X86_XSTATE_X87_ID)) ++ + /* Constants for bits in __x86_string_control: */ + + /* Avoid short distance REP MOVSB. */ +diff --git a/sysdeps/x86/tst-gnu2-tls2.c b/sysdeps/x86/tst-gnu2-tls2.c +new file mode 100644 +index 0000000000000000..de900a423bb70321 +--- /dev/null ++++ b/sysdeps/x86/tst-gnu2-tls2.c +@@ -0,0 +1,20 @@ ++#ifndef __x86_64__ ++#include ++ ++#define IS_SUPPORTED() CPU_FEATURE_ACTIVE (SSE2) ++#endif ++ ++/* Clear XMM0...XMM7 */ ++#define PREPARE_MALLOC() \ ++{ \ ++ asm volatile ("xorps %%xmm0, %%xmm0" : : : "xmm0" ); \ ++ asm volatile ("xorps %%xmm1, %%xmm1" : : : "xmm1" ); \ ++ asm volatile ("xorps %%xmm2, %%xmm2" : : : "xmm2" ); \ ++ asm volatile ("xorps %%xmm3, %%xmm3" : : : "xmm3" ); \ ++ asm volatile ("xorps %%xmm4, %%xmm4" : : : "xmm4" ); \ ++ asm volatile ("xorps %%xmm5, %%xmm5" : : : "xmm5" ); \ ++ asm volatile ("xorps %%xmm6, %%xmm6" : : : "xmm6" ); \ ++ asm volatile ("xorps %%xmm7, %%xmm7" : : : "xmm7" ); \ ++} ++ ++#include +diff --git a/sysdeps/x86_64/Makefile b/sysdeps/x86_64/Makefile +index 90f4ecfd262cfb87..e8babc9a4edbf90b 100644 +--- a/sysdeps/x86_64/Makefile ++++ b/sysdeps/x86_64/Makefile +@@ -10,7 +10,7 @@ LDFLAGS-rtld += -Wl,-z,nomark-plt + endif + + ifeq ($(subdir),csu) +-gen-as-const-headers += features-offsets.sym link-defines.sym ++gen-as-const-headers += link-defines.sym + endif + + ifeq ($(subdir),gmon) +diff --git a/sysdeps/x86_64/dl-machine.h b/sysdeps/x86_64/dl-machine.h +index 6d605d0d3293bcd6..ff5d45f7cb7cd81d 100644 +--- a/sysdeps/x86_64/dl-machine.h ++++ b/sysdeps/x86_64/dl-machine.h +@@ -71,9 +71,6 @@ elf_machine_runtime_setup (struct link_map *l, struct r_scope_elem *scope[], + int lazy, int profile) + { + Elf64_Addr *got; +- extern void _dl_runtime_resolve_fxsave (ElfW(Word)) attribute_hidden; +- extern void _dl_runtime_resolve_xsave (ElfW(Word)) attribute_hidden; +- extern void _dl_runtime_resolve_xsavec (ElfW(Word)) attribute_hidden; + extern void _dl_runtime_profile_sse (ElfW(Word)) attribute_hidden; + extern void _dl_runtime_profile_avx (ElfW(Word)) attribute_hidden; + extern void _dl_runtime_profile_avx512 (ElfW(Word)) attribute_hidden; +@@ -96,8 +93,6 @@ elf_machine_runtime_setup (struct link_map *l, struct r_scope_elem *scope[], + /* Identify this shared object. */ + *(ElfW(Addr) *) (got + 1) = (ElfW(Addr)) l; + +- const struct cpu_features* cpu_features = __get_cpu_features (); +- + #ifdef SHARED + /* The got[2] entry contains the address of a function which gets + called to get the address of a so far unresolved function and +@@ -107,6 +102,7 @@ elf_machine_runtime_setup (struct link_map *l, struct r_scope_elem *scope[], + end in this function. */ + if (__glibc_unlikely (profile)) + { ++ const struct cpu_features* cpu_features = __get_cpu_features (); + if (X86_ISA_CPU_FEATURE_USABLE_P (cpu_features, AVX512F)) + *(ElfW(Addr) *) (got + 2) = (ElfW(Addr)) &_dl_runtime_profile_avx512; + else if (X86_ISA_CPU_FEATURE_USABLE_P (cpu_features, AVX)) +@@ -126,15 +122,8 @@ elf_machine_runtime_setup (struct link_map *l, struct r_scope_elem *scope[], + /* This function will get called to fix up the GOT entry + indicated by the offset on the stack, and then jump to + the resolved address. */ +- if (MINIMUM_X86_ISA_LEVEL >= AVX_X86_ISA_LEVEL +- || GLRO(dl_x86_cpu_features).xsave_state_size != 0) +- *(ElfW(Addr) *) (got + 2) +- = (CPU_FEATURE_USABLE_P (cpu_features, XSAVEC) +- ? (ElfW(Addr)) &_dl_runtime_resolve_xsavec +- : (ElfW(Addr)) &_dl_runtime_resolve_xsave); +- else +- *(ElfW(Addr) *) (got + 2) +- = (ElfW(Addr)) &_dl_runtime_resolve_fxsave; ++ *(ElfW(Addr) *) (got + 2) ++ = (ElfW(Addr)) GLRO(dl_x86_64_runtime_resolve); + } + } + +@@ -383,7 +372,7 @@ and creates an unsatisfiable circular dependency.\n", + { + td->arg = _dl_make_tlsdesc_dynamic + (sym_map, sym->st_value + reloc->r_addend); +- td->entry = _dl_tlsdesc_dynamic; ++ td->entry = GLRO(dl_x86_tlsdesc_dynamic); + } + else + # endif +diff --git a/sysdeps/x86_64/dl-procinfo.c b/sysdeps/x86_64/dl-procinfo.c +index 4d1d790fbb2f2992..06637a8154d8648f 100644 +--- a/sysdeps/x86_64/dl-procinfo.c ++++ b/sysdeps/x86_64/dl-procinfo.c +@@ -41,5 +41,21 @@ + + #include + ++#if !IS_IN (ldconfig) ++# if !defined PROCINFO_DECL && defined SHARED ++ ._dl_x86_64_runtime_resolve ++# else ++PROCINFO_CLASS void * _dl_x86_64_runtime_resolve ++# endif ++# ifndef PROCINFO_DECL ++= NULL ++# endif ++# if !defined SHARED || defined PROCINFO_DECL ++; ++# else ++, ++# endif ++#endif ++ + #undef PROCINFO_DECL + #undef PROCINFO_CLASS +diff --git a/sysdeps/x86_64/dl-tlsdesc-dynamic.h b/sysdeps/x86_64/dl-tlsdesc-dynamic.h +new file mode 100644 +index 0000000000000000..0c2e8d5320d0bd26 +--- /dev/null ++++ b/sysdeps/x86_64/dl-tlsdesc-dynamic.h +@@ -0,0 +1,166 @@ ++/* Thread-local storage handling in the ELF dynamic linker. x86_64 version. ++ Copyright (C) 2004-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#ifndef SECTION ++# define SECTION(p) p ++#endif ++ ++#undef REGISTER_SAVE_AREA ++#undef LOCAL_STORAGE_AREA ++#undef BASE ++ ++#include "dl-trampoline-state.h" ++ ++ .section SECTION(.text),"ax",@progbits ++ ++ .hidden _dl_tlsdesc_dynamic ++ .global _dl_tlsdesc_dynamic ++ .type _dl_tlsdesc_dynamic,@function ++ ++ /* %rax points to the TLS descriptor, such that 0(%rax) points to ++ _dl_tlsdesc_dynamic itself, and 8(%rax) points to a struct ++ tlsdesc_dynamic_arg object. It must return in %rax the offset ++ between the thread pointer and the object denoted by the ++ argument, without clobbering any registers. ++ ++ The assembly code that follows is a rendition of the following ++ C code, hand-optimized a little bit. ++ ++ptrdiff_t ++_dl_tlsdesc_dynamic (register struct tlsdesc *tdp asm ("%rax")) ++{ ++ struct tlsdesc_dynamic_arg *td = tdp->arg; ++ dtv_t *dtv = *(dtv_t **)((char *)__thread_pointer + DTV_OFFSET); ++ if (__builtin_expect (td->gen_count <= dtv[0].counter ++ && (dtv[td->tlsinfo.ti_module].pointer.val ++ != TLS_DTV_UNALLOCATED), ++ 1)) ++ return dtv[td->tlsinfo.ti_module].pointer.val + td->tlsinfo.ti_offset ++ - __thread_pointer; ++ ++ return __tls_get_addr_internal (&td->tlsinfo) - __thread_pointer; ++} ++*/ ++ cfi_startproc ++ .align 16 ++_dl_tlsdesc_dynamic: ++ _CET_ENDBR ++ /* Preserve call-clobbered registers that we modify. ++ We need two scratch regs anyway. */ ++ movq %rsi, -16(%rsp) ++ mov %fs:DTV_OFFSET, %RSI_LP ++ movq %rdi, -8(%rsp) ++ movq TLSDESC_ARG(%rax), %rdi ++ movq (%rsi), %rax ++ cmpq %rax, TLSDESC_GEN_COUNT(%rdi) ++ ja 2f ++ movq TLSDESC_MODID(%rdi), %rax ++ salq $4, %rax ++ movq (%rax,%rsi), %rax ++ cmpq $-1, %rax ++ je 2f ++ addq TLSDESC_MODOFF(%rdi), %rax ++1: ++ movq -16(%rsp), %rsi ++ sub %fs:0, %RAX_LP ++ movq -8(%rsp), %rdi ++ ret ++2: ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++ movq %rbx, -24(%rsp) ++ mov %RSP_LP, %RBX_LP ++ cfi_def_cfa_register(%rbx) ++ and $-STATE_SAVE_ALIGNMENT, %RSP_LP ++#endif ++#ifdef REGISTER_SAVE_AREA ++# if DL_RUNTIME_RESOLVE_REALIGN_STACK ++ /* STATE_SAVE_OFFSET has space for 8 integer registers. But we ++ need space for RCX, RDX, RSI, RDI, R8, R9, R10 and R11, plus ++ RBX above. */ ++ sub $(REGISTER_SAVE_AREA + STATE_SAVE_ALIGNMENT), %RSP_LP ++# else ++ sub $REGISTER_SAVE_AREA, %RSP_LP ++ cfi_adjust_cfa_offset(REGISTER_SAVE_AREA) ++# endif ++#else ++ /* Allocate stack space of the required size to save the state. */ ++ sub _rtld_local_ro+RTLD_GLOBAL_RO_DL_X86_CPU_FEATURES_OFFSET+XSAVE_STATE_SIZE_OFFSET(%rip), %RSP_LP ++#endif ++ /* Besides rdi and rsi, saved above, save rcx, rdx, r8, r9, ++ r10 and r11. */ ++ movq %rcx, REGISTER_SAVE_RCX(%rsp) ++ movq %rdx, REGISTER_SAVE_RDX(%rsp) ++ movq %r8, REGISTER_SAVE_R8(%rsp) ++ movq %r9, REGISTER_SAVE_R9(%rsp) ++ movq %r10, REGISTER_SAVE_R10(%rsp) ++ movq %r11, REGISTER_SAVE_R11(%rsp) ++#ifdef USE_FXSAVE ++ fxsave STATE_SAVE_OFFSET(%rsp) ++#else ++ movl $TLSDESC_CALL_STATE_SAVE_MASK, %eax ++ xorl %edx, %edx ++ /* Clear the XSAVE Header. */ ++# ifdef USE_XSAVE ++ movq %rdx, (STATE_SAVE_OFFSET + 512)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8)(%rsp) ++# endif ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 2)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 3)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 4)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 5)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 6)(%rsp) ++ movq %rdx, (STATE_SAVE_OFFSET + 512 + 8 * 7)(%rsp) ++# ifdef USE_XSAVE ++ xsave STATE_SAVE_OFFSET(%rsp) ++# else ++ xsavec STATE_SAVE_OFFSET(%rsp) ++# endif ++#endif ++ /* %rdi already points to the tlsinfo data structure. */ ++ call HIDDEN_JUMPTARGET (__tls_get_addr) ++ # Get register content back. ++#ifdef USE_FXSAVE ++ fxrstor STATE_SAVE_OFFSET(%rsp) ++#else ++ /* Save and retore __tls_get_addr return value stored in RAX. */ ++ mov %RAX_LP, %RCX_LP ++ movl $TLSDESC_CALL_STATE_SAVE_MASK, %eax ++ xorl %edx, %edx ++ xrstor STATE_SAVE_OFFSET(%rsp) ++ mov %RCX_LP, %RAX_LP ++#endif ++ movq REGISTER_SAVE_R11(%rsp), %r11 ++ movq REGISTER_SAVE_R10(%rsp), %r10 ++ movq REGISTER_SAVE_R9(%rsp), %r9 ++ movq REGISTER_SAVE_R8(%rsp), %r8 ++ movq REGISTER_SAVE_RDX(%rsp), %rdx ++ movq REGISTER_SAVE_RCX(%rsp), %rcx ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++ mov %RBX_LP, %RSP_LP ++ cfi_def_cfa_register(%rsp) ++ movq -24(%rsp), %rbx ++ cfi_restore(%rbx) ++#else ++ add $REGISTER_SAVE_AREA, %RSP_LP ++ cfi_adjust_cfa_offset(-REGISTER_SAVE_AREA) ++#endif ++ jmp 1b ++ cfi_endproc ++ .size _dl_tlsdesc_dynamic, .-_dl_tlsdesc_dynamic ++ ++#undef STATE_SAVE_ALIGNMENT +diff --git a/sysdeps/x86_64/dl-tlsdesc.S b/sysdeps/x86_64/dl-tlsdesc.S +index f748af2ece8de09a..ea69f5223a77e0c0 100644 +--- a/sysdeps/x86_64/dl-tlsdesc.S ++++ b/sysdeps/x86_64/dl-tlsdesc.S +@@ -18,7 +18,19 @@ + + #include + #include ++#include ++#include + #include "tlsdesc.h" ++#include "dl-trampoline-save.h" ++ ++/* Area on stack to save and restore registers used for parameter ++ passing when calling _dl_tlsdesc_dynamic. */ ++#define REGISTER_SAVE_RCX 0 ++#define REGISTER_SAVE_RDX (REGISTER_SAVE_RCX + 8) ++#define REGISTER_SAVE_R8 (REGISTER_SAVE_RDX + 8) ++#define REGISTER_SAVE_R9 (REGISTER_SAVE_R8 + 8) ++#define REGISTER_SAVE_R10 (REGISTER_SAVE_R9 + 8) ++#define REGISTER_SAVE_R11 (REGISTER_SAVE_R10 + 8) + + .text + +@@ -67,80 +79,24 @@ _dl_tlsdesc_undefweak: + .size _dl_tlsdesc_undefweak, .-_dl_tlsdesc_undefweak + + #ifdef SHARED +- .hidden _dl_tlsdesc_dynamic +- .global _dl_tlsdesc_dynamic +- .type _dl_tlsdesc_dynamic,@function +- +- /* %rax points to the TLS descriptor, such that 0(%rax) points to +- _dl_tlsdesc_dynamic itself, and 8(%rax) points to a struct +- tlsdesc_dynamic_arg object. It must return in %rax the offset +- between the thread pointer and the object denoted by the +- argument, without clobbering any registers. +- +- The assembly code that follows is a rendition of the following +- C code, hand-optimized a little bit. +- +-ptrdiff_t +-_dl_tlsdesc_dynamic (register struct tlsdesc *tdp asm ("%rax")) +-{ +- struct tlsdesc_dynamic_arg *td = tdp->arg; +- dtv_t *dtv = *(dtv_t **)((char *)__thread_pointer + DTV_OFFSET); +- if (__builtin_expect (td->gen_count <= dtv[0].counter +- && (dtv[td->tlsinfo.ti_module].pointer.val +- != TLS_DTV_UNALLOCATED), +- 1)) +- return dtv[td->tlsinfo.ti_module].pointer.val + td->tlsinfo.ti_offset +- - __thread_pointer; +- +- return __tls_get_addr_internal (&td->tlsinfo) - __thread_pointer; +-} +-*/ +- cfi_startproc +- .align 16 +-_dl_tlsdesc_dynamic: +- _CET_ENDBR +- /* Preserve call-clobbered registers that we modify. +- We need two scratch regs anyway. */ +- movq %rsi, -16(%rsp) +- mov %fs:DTV_OFFSET, %RSI_LP +- movq %rdi, -8(%rsp) +- movq TLSDESC_ARG(%rax), %rdi +- movq (%rsi), %rax +- cmpq %rax, TLSDESC_GEN_COUNT(%rdi) +- ja .Lslow +- movq TLSDESC_MODID(%rdi), %rax +- salq $4, %rax +- movq (%rax,%rsi), %rax +- cmpq $-1, %rax +- je .Lslow +- addq TLSDESC_MODOFF(%rdi), %rax +-.Lret: +- movq -16(%rsp), %rsi +- sub %fs:0, %RAX_LP +- movq -8(%rsp), %rdi +- ret +-.Lslow: +- /* Besides rdi and rsi, saved above, save rdx, rcx, r8, r9, +- r10 and r11. Also, align the stack, that's off by 8 bytes. */ +- subq $72, %rsp +- cfi_adjust_cfa_offset (72) +- movq %rdx, 8(%rsp) +- movq %rcx, 16(%rsp) +- movq %r8, 24(%rsp) +- movq %r9, 32(%rsp) +- movq %r10, 40(%rsp) +- movq %r11, 48(%rsp) +- /* %rdi already points to the tlsinfo data structure. */ +- call HIDDEN_JUMPTARGET (__tls_get_addr) +- movq 8(%rsp), %rdx +- movq 16(%rsp), %rcx +- movq 24(%rsp), %r8 +- movq 32(%rsp), %r9 +- movq 40(%rsp), %r10 +- movq 48(%rsp), %r11 +- addq $72, %rsp +- cfi_adjust_cfa_offset (-72) +- jmp .Lret +- cfi_endproc +- .size _dl_tlsdesc_dynamic, .-_dl_tlsdesc_dynamic ++# define USE_FXSAVE ++# define STATE_SAVE_ALIGNMENT 16 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_fxsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_FXSAVE ++ ++# define USE_XSAVE ++# define STATE_SAVE_ALIGNMENT 64 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_xsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_XSAVE ++ ++# define USE_XSAVEC ++# define STATE_SAVE_ALIGNMENT 64 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_xsavec ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_XSAVEC + #endif /* SHARED */ +diff --git a/sysdeps/x86_64/dl-trampoline-save.h b/sysdeps/x86_64/dl-trampoline-save.h +new file mode 100644 +index 0000000000000000..84eac4a8ac13ad86 +--- /dev/null ++++ b/sysdeps/x86_64/dl-trampoline-save.h +@@ -0,0 +1,34 @@ ++/* x86-64 PLT trampoline register save macros. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#ifndef DL_STACK_ALIGNMENT ++/* Due to GCC bug: ++ ++ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58066 ++ ++ __tls_get_addr may be called with 8-byte stack alignment. Although ++ this bug has been fixed in GCC 4.9.4, 5.3 and 6, we can't assume ++ that stack will be always aligned at 16 bytes. */ ++# define DL_STACK_ALIGNMENT 8 ++#endif ++ ++/* True if _dl_runtime_resolve should align stack for STATE_SAVE or align ++ stack to 16 bytes before calling _dl_fixup. */ ++#define DL_RUNTIME_RESOLVE_REALIGN_STACK \ ++ (STATE_SAVE_ALIGNMENT > DL_STACK_ALIGNMENT \ ++ || 16 > DL_STACK_ALIGNMENT) +diff --git a/sysdeps/x86_64/dl-trampoline-state.h b/sysdeps/x86_64/dl-trampoline-state.h +new file mode 100644 +index 0000000000000000..575f120797860583 +--- /dev/null ++++ b/sysdeps/x86_64/dl-trampoline-state.h +@@ -0,0 +1,51 @@ ++/* x86-64 PLT dl-trampoline state macros. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#if (STATE_SAVE_ALIGNMENT % 16) != 0 ++# error STATE_SAVE_ALIGNMENT must be multiple of 16 ++#endif ++ ++#if (STATE_SAVE_OFFSET % STATE_SAVE_ALIGNMENT) != 0 ++# error STATE_SAVE_OFFSET must be multiple of STATE_SAVE_ALIGNMENT ++#endif ++ ++#if DL_RUNTIME_RESOLVE_REALIGN_STACK ++/* Local stack area before jumping to function address: RBX. */ ++# define LOCAL_STORAGE_AREA 8 ++# define BASE rbx ++# ifdef USE_FXSAVE ++/* Use fxsave to save XMM registers. */ ++# define REGISTER_SAVE_AREA (512 + STATE_SAVE_OFFSET) ++# if (REGISTER_SAVE_AREA % 16) != 0 ++# error REGISTER_SAVE_AREA must be multiple of 16 ++# endif ++# endif ++#else ++# ifndef USE_FXSAVE ++# error USE_FXSAVE must be defined ++# endif ++/* Use fxsave to save XMM registers. */ ++# define REGISTER_SAVE_AREA (512 + STATE_SAVE_OFFSET + 8) ++/* Local stack area before jumping to function address: All saved ++ registers. */ ++# define LOCAL_STORAGE_AREA REGISTER_SAVE_AREA ++# define BASE rsp ++# if (REGISTER_SAVE_AREA % 16) != 8 ++# error REGISTER_SAVE_AREA must be odd multiple of 8 ++# endif ++#endif +diff --git a/sysdeps/x86_64/dl-trampoline.S b/sysdeps/x86_64/dl-trampoline.S +index b2e7e0f69b709ffd..87c5137837f01a63 100644 +--- a/sysdeps/x86_64/dl-trampoline.S ++++ b/sysdeps/x86_64/dl-trampoline.S +@@ -22,25 +22,7 @@ + #include + #include + #include +- +-#ifndef DL_STACK_ALIGNMENT +-/* Due to GCC bug: +- +- https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58066 +- +- __tls_get_addr may be called with 8-byte stack alignment. Although +- this bug has been fixed in GCC 4.9.4, 5.3 and 6, we can't assume +- that stack will be always aligned at 16 bytes. We use unaligned +- 16-byte move to load and store SSE registers, which has no penalty +- on modern processors if stack is 16-byte aligned. */ +-# define DL_STACK_ALIGNMENT 8 +-#endif +- +-/* True if _dl_runtime_resolve should align stack for STATE_SAVE or align +- stack to 16 bytes before calling _dl_fixup. */ +-#define DL_RUNTIME_RESOLVE_REALIGN_STACK \ +- (STATE_SAVE_ALIGNMENT > DL_STACK_ALIGNMENT \ +- || 16 > DL_STACK_ALIGNMENT) ++#include "dl-trampoline-save.h" + + /* Area on stack to save and restore registers used for parameter + passing when calling _dl_fixup. */ +diff --git a/sysdeps/x86_64/dl-trampoline.h b/sysdeps/x86_64/dl-trampoline.h +index f55c6ea040b9d319..d9ccfb40d46d3312 100644 +--- a/sysdeps/x86_64/dl-trampoline.h ++++ b/sysdeps/x86_64/dl-trampoline.h +@@ -27,39 +27,7 @@ + # undef LOCAL_STORAGE_AREA + # undef BASE + +-# if (STATE_SAVE_ALIGNMENT % 16) != 0 +-# error STATE_SAVE_ALIGNMENT must be multiple of 16 +-# endif +- +-# if (STATE_SAVE_OFFSET % STATE_SAVE_ALIGNMENT) != 0 +-# error STATE_SAVE_OFFSET must be multiple of STATE_SAVE_ALIGNMENT +-# endif +- +-# if DL_RUNTIME_RESOLVE_REALIGN_STACK +-/* Local stack area before jumping to function address: RBX. */ +-# define LOCAL_STORAGE_AREA 8 +-# define BASE rbx +-# ifdef USE_FXSAVE +-/* Use fxsave to save XMM registers. */ +-# define REGISTER_SAVE_AREA (512 + STATE_SAVE_OFFSET) +-# if (REGISTER_SAVE_AREA % 16) != 0 +-# error REGISTER_SAVE_AREA must be multiple of 16 +-# endif +-# endif +-# else +-# ifndef USE_FXSAVE +-# error USE_FXSAVE must be defined +-# endif +-/* Use fxsave to save XMM registers. */ +-# define REGISTER_SAVE_AREA (512 + STATE_SAVE_OFFSET + 8) +-/* Local stack area before jumping to function address: All saved +- registers. */ +-# define LOCAL_STORAGE_AREA REGISTER_SAVE_AREA +-# define BASE rsp +-# if (REGISTER_SAVE_AREA % 16) != 8 +-# error REGISTER_SAVE_AREA must be odd multiple of 8 +-# endif +-# endif ++# include "dl-trampoline-state.h" + + .globl _dl_runtime_resolve + .hidden _dl_runtime_resolve diff --git a/SOURCES/glibc-upstream-2.39-13.patch b/SOURCES/glibc-upstream-2.39-13.patch new file mode 100644 index 0000000..0a52189 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-13.patch @@ -0,0 +1,496 @@ +commit 853e915fdd6ae6c5f1a7a68d2594ec8dbfef1286 +Author: H.J. Lu +Date: Wed Feb 28 12:08:03 2024 -0800 + + x86-64: Update _dl_tlsdesc_dynamic to preserve AMX registers + + _dl_tlsdesc_dynamic should also preserve AMX registers which are + caller-saved. Add X86_XSTATE_TILECFG_ID and X86_XSTATE_TILEDATA_ID + to x86-64 TLSDESC_CALL_STATE_SAVE_MASK. Compute the AMX state size + and save it in xsave_state_full_size which is only used by + _dl_tlsdesc_dynamic_xsave and _dl_tlsdesc_dynamic_xsavec. This fixes + the AMX part of BZ #31372. Tested on AMX processor. + + AMX test is enabled only for compilers with the fix for + + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114098 + + GCC 14 and GCC 11/12/13 branches have the bug fix. + Reviewed-by: Sunil K Pandey + + (cherry picked from commit 9b7091415af47082664717210ac49d51551456ab) + +diff --git a/sysdeps/unix/sysv/linux/x86_64/Makefile b/sysdeps/unix/sysv/linux/x86_64/Makefile +index 4223feb95f6fe2f5..9a1e7aa6461725af 100644 +--- a/sysdeps/unix/sysv/linux/x86_64/Makefile ++++ b/sysdeps/unix/sysv/linux/x86_64/Makefile +@@ -63,6 +63,33 @@ $(objpfx)libx86-64-isa-level%.os: $(..)/sysdeps/unix/sysv/linux/x86_64/x86-64-is + $(objpfx)libx86-64-isa-level.so: $(objpfx)libx86-64-isa-level-1.so + cp $< $@ + endif ++ ++ifeq (yes,$(have-mamx-tile)) ++tests += \ ++ tst-gnu2-tls2-amx \ ++# tests ++ ++modules-names += \ ++ tst-gnu2-tls2-amx-mod0 \ ++ tst-gnu2-tls2-amx-mod1 \ ++ tst-gnu2-tls2-amx-mod2 \ ++# modules-names ++ ++$(objpfx)tst-gnu2-tls2-amx: $(shared-thread-library) ++$(objpfx)tst-gnu2-tls2-amx.out: \ ++ $(objpfx)tst-gnu2-tls2-amx-mod0.so \ ++ $(objpfx)tst-gnu2-tls2-amx-mod1.so \ ++ $(objpfx)tst-gnu2-tls2-amx-mod2.so ++$(objpfx)tst-gnu2-tls2-amx-mod0.so: $(libsupport) ++$(objpfx)tst-gnu2-tls2-amx-mod1.so: $(libsupport) ++$(objpfx)tst-gnu2-tls2-amx-mod2.so: $(libsupport) ++ ++CFLAGS-tst-gnu2-tls2-amx.c += -mamx-tile ++CFLAGS-tst-gnu2-tls2-amx-mod0.c += -mamx-tile -mtls-dialect=gnu2 ++CFLAGS-tst-gnu2-tls2-amx-mod1.c += -mamx-tile -mtls-dialect=gnu2 ++CFLAGS-tst-gnu2-tls2-amx-mod2.c += -mamx-tile -mtls-dialect=gnu2 ++endif ++ + endif # $(subdir) == elf + + ifneq ($(enable-cet),no) +diff --git a/sysdeps/unix/sysv/linux/x86_64/include/asm/prctl.h b/sysdeps/unix/sysv/linux/x86_64/include/asm/prctl.h +index 2f511321ad3b3ac1..ef4631bf4b2fd9aa 100644 +--- a/sysdeps/unix/sysv/linux/x86_64/include/asm/prctl.h ++++ b/sysdeps/unix/sysv/linux/x86_64/include/asm/prctl.h +@@ -20,3 +20,8 @@ + # define ARCH_SHSTK_SHSTK 0x1 + # define ARCH_SHSTK_WRSS 0x2 + #endif ++ ++#ifndef ARCH_GET_XCOMP_PERM ++# define ARCH_GET_XCOMP_PERM 0x1022 ++# define ARCH_REQ_XCOMP_PERM 0x1023 ++#endif +diff --git a/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod0.c b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod0.c +new file mode 100644 +index 0000000000000000..2e0c7b91b7caf3ab +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod0.c +@@ -0,0 +1,2 @@ ++#include "tst-gnu2-tls2-amx.h" ++#include +diff --git a/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod1.c b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod1.c +new file mode 100644 +index 0000000000000000..b8a8ccf1c119d443 +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod1.c +@@ -0,0 +1,2 @@ ++#include "tst-gnu2-tls2-amx.h" ++#include +diff --git a/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod2.c b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod2.c +new file mode 100644 +index 0000000000000000..cdf4a8f3635b327c +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx-mod2.c +@@ -0,0 +1,2 @@ ++#include "tst-gnu2-tls2-amx.h" ++#include +diff --git a/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.c b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.c +new file mode 100644 +index 0000000000000000..ae4dd82556c9b2ef +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.c +@@ -0,0 +1,83 @@ ++/* Test TLSDESC relocation with AMX. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include "tst-gnu2-tls2-amx.h" ++ ++extern int arch_prctl (int, ...); ++ ++#define X86_XSTATE_TILECFG_ID 17 ++#define X86_XSTATE_TILEDATA_ID 18 ++ ++/* Initialize tile config. */ ++__attribute__ ((noinline, noclone)) ++static void ++init_tile_config (__tilecfg *tileinfo) ++{ ++ int i; ++ tileinfo->palette_id = 1; ++ tileinfo->start_row = 0; ++ ++ tileinfo->colsb[0] = MAX_ROWS; ++ tileinfo->rows[0] = MAX_ROWS; ++ ++ for (i = 1; i < 4; ++i) ++ { ++ tileinfo->colsb[i] = MAX_COLS; ++ tileinfo->rows[i] = MAX_ROWS; ++ } ++ ++ _tile_loadconfig (tileinfo); ++} ++ ++static bool ++enable_amx (void) ++{ ++ uint64_t bitmask; ++ if (arch_prctl (ARCH_GET_XCOMP_PERM, &bitmask) != 0) ++ return false; ++ ++ if ((bitmask & (1 << X86_XSTATE_TILECFG_ID)) == 0) ++ return false; ++ ++ if (arch_prctl (ARCH_REQ_XCOMP_PERM, X86_XSTATE_TILEDATA_ID) != 0) ++ return false; ++ ++ /* Load tile configuration. */ ++ __tilecfg tile_data = { 0 }; ++ init_tile_config (&tile_data); ++ ++ return true; ++} ++ ++/* An architecture can define it to clobber caller-saved registers in ++ malloc below to verify that the implicit TLSDESC call won't change ++ caller-saved registers. */ ++static void ++clear_tile_register (void) ++{ ++ _tile_zero (2); ++} ++ ++#define MOD(i) "tst-gnu2-tls2-amx-mod" #i ".so" ++#define IS_SUPPORTED() enable_amx () ++#define PREPARE_MALLOC() clear_tile_register () ++ ++#include +diff --git a/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.h b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.h +new file mode 100644 +index 0000000000000000..1845a3caba43a0f1 +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/x86_64/tst-gnu2-tls2-amx.h +@@ -0,0 +1,63 @@ ++/* Test TLSDESC relocation with AMX. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++ ++#define MAX_ROWS 16 ++#define MAX_COLS 64 ++#define MAX 1024 ++#define STRIDE 64 ++ ++typedef struct __tile_config ++{ ++ uint8_t palette_id; ++ uint8_t start_row; ++ uint8_t reserved_0[14]; ++ uint16_t colsb[16]; ++ uint8_t rows[16]; ++} __tilecfg __attribute__ ((aligned (64))); ++ ++/* Initialize int8_t buffer */ ++static inline void ++init_buffer (int8_t *buf, int8_t value) ++{ ++ int rows, colsb, i, j; ++ rows = MAX_ROWS; ++ colsb = MAX_COLS; ++ ++ for (i = 0; i < rows; i++) ++ for (j = 0; j < colsb; j++) ++ buf[i * colsb + j] = value; ++} ++ ++#define BEFORE_TLSDESC_CALL() \ ++ int8_t src[MAX]; \ ++ int8_t res[MAX]; \ ++ /* Initialize src with data */ \ ++ init_buffer (src, 2); \ ++ /* Load tile rows from memory. */ \ ++ _tile_loadd (2, src, STRIDE); ++ ++#define AFTER_TLSDESC_CALL() \ ++ /* Store the tile data to memory. */ \ ++ _tile_stored (2, res, STRIDE); \ ++ _tile_release (); \ ++ TEST_VERIFY_EXIT (memcmp (src, res, sizeof (res)) == 0); +diff --git a/sysdeps/x86/cpu-features-offsets.sym b/sysdeps/x86/cpu-features-offsets.sym +index 6a8fd298137b7f23..21fc88d6510840e6 100644 +--- a/sysdeps/x86/cpu-features-offsets.sym ++++ b/sysdeps/x86/cpu-features-offsets.sym +@@ -3,3 +3,4 @@ + #include + + XSAVE_STATE_SIZE_OFFSET offsetof (struct cpu_features, xsave_state_size) ++XSAVE_STATE_FULL_SIZE_OFFSET offsetof (struct cpu_features, xsave_state_full_size) +diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c +index 835113b42f924b83..d71e8d3d2e0e49f9 100644 +--- a/sysdeps/x86/cpu-features.c ++++ b/sysdeps/x86/cpu-features.c +@@ -307,6 +307,8 @@ update_active (struct cpu_features *cpu_features) + __cpuid_count (0xd, 0, eax, ebx, ecx, edx); + if (ebx != 0) + { ++ /* NB: On AMX capable processors, ebx always includes AMX ++ states. */ + unsigned int xsave_state_full_size + = ALIGN_UP (ebx + STATE_SAVE_OFFSET, 64); + +@@ -320,6 +322,11 @@ update_active (struct cpu_features *cpu_features) + { + unsigned int xstate_comp_offsets[32]; + unsigned int xstate_comp_sizes[32]; ++#ifdef __x86_64__ ++ unsigned int xstate_amx_comp_offsets[32]; ++ unsigned int xstate_amx_comp_sizes[32]; ++ unsigned int amx_ecx; ++#endif + unsigned int i; + + xstate_comp_offsets[0] = 0; +@@ -327,16 +334,39 @@ update_active (struct cpu_features *cpu_features) + xstate_comp_offsets[2] = 576; + xstate_comp_sizes[0] = 160; + xstate_comp_sizes[1] = 256; ++#ifdef __x86_64__ ++ xstate_amx_comp_offsets[0] = 0; ++ xstate_amx_comp_offsets[1] = 160; ++ xstate_amx_comp_offsets[2] = 576; ++ xstate_amx_comp_sizes[0] = 160; ++ xstate_amx_comp_sizes[1] = 256; ++#endif + + for (i = 2; i < 32; i++) + { +- if ((STATE_SAVE_MASK & (1 << i)) != 0) ++ if ((FULL_STATE_SAVE_MASK & (1 << i)) != 0) + { + __cpuid_count (0xd, i, eax, ebx, ecx, edx); +- xstate_comp_sizes[i] = eax; ++#ifdef __x86_64__ ++ /* Include this in xsave_state_full_size. */ ++ amx_ecx = ecx; ++ xstate_amx_comp_sizes[i] = eax; ++ if ((AMX_STATE_SAVE_MASK & (1 << i)) != 0) ++ { ++ /* Exclude this from xsave_state_size. */ ++ ecx = 0; ++ xstate_comp_sizes[i] = 0; ++ } ++ else ++#endif ++ xstate_comp_sizes[i] = eax; + } + else + { ++#ifdef __x86_64__ ++ amx_ecx = 0; ++ xstate_amx_comp_sizes[i] = 0; ++#endif + ecx = 0; + xstate_comp_sizes[i] = 0; + } +@@ -349,6 +379,15 @@ update_active (struct cpu_features *cpu_features) + if ((ecx & (1 << 1)) != 0) + xstate_comp_offsets[i] + = ALIGN_UP (xstate_comp_offsets[i], 64); ++#ifdef __x86_64__ ++ xstate_amx_comp_offsets[i] ++ = (xstate_amx_comp_offsets[i - 1] ++ + xstate_amx_comp_sizes[i - 1]); ++ if ((amx_ecx & (1 << 1)) != 0) ++ xstate_amx_comp_offsets[i] ++ = ALIGN_UP (xstate_amx_comp_offsets[i], ++ 64); ++#endif + } + } + +@@ -357,6 +396,18 @@ update_active (struct cpu_features *cpu_features) + = xstate_comp_offsets[31] + xstate_comp_sizes[31]; + if (size) + { ++#ifdef __x86_64__ ++ unsigned int amx_size ++ = (xstate_amx_comp_offsets[31] ++ + xstate_amx_comp_sizes[31]); ++ amx_size = ALIGN_UP (amx_size + STATE_SAVE_OFFSET, ++ 64); ++ /* Set xsave_state_full_size to the compact AMX ++ state size for XSAVEC. NB: xsave_state_full_size ++ is only used in _dl_tlsdesc_dynamic_xsave and ++ _dl_tlsdesc_dynamic_xsavec. */ ++ cpu_features->xsave_state_full_size = amx_size; ++#endif + cpu_features->xsave_state_size + = ALIGN_UP (size + STATE_SAVE_OFFSET, 64); + CPU_FEATURE_SET (cpu_features, XSAVEC); +diff --git a/sysdeps/x86/include/cpu-features.h b/sysdeps/x86/include/cpu-features.h +index b9bf3115b616f05f..cd7bd27cf35959fd 100644 +--- a/sysdeps/x86/include/cpu-features.h ++++ b/sysdeps/x86/include/cpu-features.h +@@ -934,6 +934,8 @@ struct cpu_features + /* The full state size for XSAVE when XSAVEC is disabled by + + GLIBC_TUNABLES=glibc.cpu.hwcaps=-XSAVEC ++ ++ and the AMX state size when XSAVEC is available. + */ + unsigned int xsave_state_full_size; + /* Data cache size for use in memory and string routines, typically +diff --git a/sysdeps/x86/sysdep.h b/sysdeps/x86/sysdep.h +index 485cad9c0283b334..db8e576e91767db5 100644 +--- a/sysdeps/x86/sysdep.h ++++ b/sysdeps/x86/sysdep.h +@@ -56,6 +56,14 @@ + | (1 << X86_XSTATE_ZMM_H_ID) \ + | (1 << X86_XSTATE_ZMM_ID) \ + | (1 << X86_XSTATE_APX_F_ID)) ++ ++/* AMX state mask. */ ++# define AMX_STATE_SAVE_MASK \ ++ ((1 << X86_XSTATE_TILECFG_ID) | (1 << X86_XSTATE_TILEDATA_ID)) ++ ++/* States to be included in xsave_state_full_size. */ ++# define FULL_STATE_SAVE_MASK \ ++ (STATE_SAVE_MASK | AMX_STATE_SAVE_MASK) + #else + /* Offset for fxsave/xsave area used by _dl_tlsdesc_dynamic. Since i386 + doesn't have red-zone, use 0 here. */ +@@ -68,13 +76,17 @@ + | (1 << X86_XSTATE_BNDREGS_ID) \ + | (1 << X86_XSTATE_K_ID) \ + | (1 << X86_XSTATE_ZMM_H_ID)) ++ ++/* States to be included in xsave_state_size. */ ++# define FULL_STATE_SAVE_MASK STATE_SAVE_MASK + #endif + + /* States which should be saved for TLSDESC_CALL and TLS_DESC_CALL. +- Compiler assumes that all registers, including x87 FPU stack registers, +- are unchanged after CALL, except for EFLAGS and RAX/EAX. */ ++ Compiler assumes that all registers, including AMX and x87 FPU ++ stack registers, are unchanged after CALL, except for EFLAGS and ++ RAX/EAX. */ + #define TLSDESC_CALL_STATE_SAVE_MASK \ +- (STATE_SAVE_MASK | (1 << X86_XSTATE_X87_ID)) ++ (FULL_STATE_SAVE_MASK | (1 << X86_XSTATE_X87_ID)) + + /* Constants for bits in __x86_string_control: */ + +diff --git a/sysdeps/x86_64/configure b/sysdeps/x86_64/configure +index 418cc4a9b862f7e0..04a534fa126a7bf7 100755 +--- a/sysdeps/x86_64/configure ++++ b/sysdeps/x86_64/configure +@@ -134,6 +134,34 @@ fi + config_vars="$config_vars + enable-cet = $enable_cet" + ++# Check if -mamx-tile works properly. ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking whether -mamx-tile works properly" >&5 ++printf %s "checking whether -mamx-tile works properly... " >&6; } ++if test ${libc_cv_x86_have_amx_tile+y} ++then : ++ printf %s "(cached) " >&6 ++else $as_nop ++ cat > conftest.c < ++EOF ++ libc_cv_x86_have_amx_tile=no ++ if { ac_try='${CC-cc} -E $CFLAGS -mamx-tile conftest.c > conftest.i' ++ { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5 ++ (eval $ac_try) 2>&5 ++ ac_status=$? ++ printf "%s\n" "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5 ++ test $ac_status = 0; }; }; then ++ if grep -q __builtin_ia32_ldtilecfg conftest.i; then ++ libc_cv_x86_have_amx_tile=yes ++ fi ++ fi ++ rm -rf conftest* ++fi ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_x86_have_amx_tile" >&5 ++printf "%s\n" "$libc_cv_x86_have_amx_tile" >&6; } ++config_vars="$config_vars ++have-mamx-tile = $libc_cv_x86_have_amx_tile" ++ + test -n "$critic_missing" && as_fn_error $? " + *** $critic_missing" "$LINENO" 5 + +diff --git a/sysdeps/x86_64/configure.ac b/sysdeps/x86_64/configure.ac +index d1f803c02ee67fc5..c714c47351e70390 100644 +--- a/sysdeps/x86_64/configure.ac ++++ b/sysdeps/x86_64/configure.ac +@@ -61,5 +61,20 @@ elif test $enable_cet = permissive; then + fi + LIBC_CONFIG_VAR([enable-cet], [$enable_cet]) + ++# Check if -mamx-tile works properly. ++AC_CACHE_CHECK(whether -mamx-tile works properly, ++ libc_cv_x86_have_amx_tile, [dnl ++cat > conftest.c < ++EOF ++ libc_cv_x86_have_amx_tile=no ++ if AC_TRY_COMMAND(${CC-cc} -E $CFLAGS -mamx-tile conftest.c > conftest.i); then ++ if grep -q __builtin_ia32_ldtilecfg conftest.i; then ++ libc_cv_x86_have_amx_tile=yes ++ fi ++ fi ++ rm -rf conftest*]) ++LIBC_CONFIG_VAR([have-mamx-tile], [$libc_cv_x86_have_amx_tile]) ++ + test -n "$critic_missing" && AC_MSG_ERROR([ + *** $critic_missing]) +diff --git a/sysdeps/x86_64/dl-tlsdesc-dynamic.h b/sysdeps/x86_64/dl-tlsdesc-dynamic.h +index 0c2e8d5320d0bd26..9f02cfc3eb297ed2 100644 +--- a/sysdeps/x86_64/dl-tlsdesc-dynamic.h ++++ b/sysdeps/x86_64/dl-tlsdesc-dynamic.h +@@ -99,7 +99,7 @@ _dl_tlsdesc_dynamic: + # endif + #else + /* Allocate stack space of the required size to save the state. */ +- sub _rtld_local_ro+RTLD_GLOBAL_RO_DL_X86_CPU_FEATURES_OFFSET+XSAVE_STATE_SIZE_OFFSET(%rip), %RSP_LP ++ sub _rtld_local_ro+RTLD_GLOBAL_RO_DL_X86_CPU_FEATURES_OFFSET+XSAVE_STATE_FULL_SIZE_OFFSET(%rip), %RSP_LP + #endif + /* Besides rdi and rsi, saved above, save rcx, rdx, r8, r9, + r10 and r11. */ diff --git a/SOURCES/glibc-upstream-2.39-14.patch b/SOURCES/glibc-upstream-2.39-14.patch new file mode 100644 index 0000000..a49f788 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-14.patch @@ -0,0 +1,250 @@ +commit 354cabcb2634abe16da7a2ba5e648aac1204b58e +Author: H.J. Lu +Date: Mon Mar 18 06:40:16 2024 -0700 + + x86-64: Allocate state buffer space for RDI, RSI and RBX + + _dl_tlsdesc_dynamic preserves RDI, RSI and RBX before realigning stack. + After realigning stack, it saves RCX, RDX, R8, R9, R10 and R11. Define + TLSDESC_CALL_REGISTER_SAVE_AREA to allocate space for RDI, RSI and RBX + to avoid clobbering saved RDI, RSI and RBX values on stack by xsave to + STATE_SAVE_OFFSET(%rsp). + + +==================+<- stack frame start aligned at 8 or 16 bytes + | |<- RDI saved in the red zone + | |<- RSI saved in the red zone + | |<- RBX saved in the red zone + | |<- paddings for stack realignment of 64 bytes + |------------------|<- xsave buffer end aligned at 64 bytes + | |<- + | |<- + | |<- + |------------------|<- xsave buffer start at STATE_SAVE_OFFSET(%rsp) + | |<- 8-byte padding for 64-byte alignment + | |<- 8-byte padding for 64-byte alignment + | |<- R11 + | |<- R10 + | |<- R9 + | |<- R8 + | |<- RDX + | |<- RCX + +==================+<- RSP aligned at 64 bytes + + Define TLSDESC_CALL_REGISTER_SAVE_AREA, the total register save area size + for all integer registers by adding 24 to STATE_SAVE_OFFSET since RDI, RSI + and RBX are saved onto stack without adjusting stack pointer first, using + the red-zone. This fixes BZ #31501. + Reviewed-by: Sunil K Pandey + + (cherry picked from commit 717ebfa85c8240d32d0d19d86a484c31c55c9617) + +diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c +index d71e8d3d2e0e49f9..6fe1b728c607f39e 100644 +--- a/sysdeps/x86/cpu-features.c ++++ b/sysdeps/x86/cpu-features.c +@@ -310,7 +310,7 @@ update_active (struct cpu_features *cpu_features) + /* NB: On AMX capable processors, ebx always includes AMX + states. */ + unsigned int xsave_state_full_size +- = ALIGN_UP (ebx + STATE_SAVE_OFFSET, 64); ++ = ALIGN_UP (ebx + TLSDESC_CALL_REGISTER_SAVE_AREA, 64); + + cpu_features->xsave_state_size + = xsave_state_full_size; +@@ -400,8 +400,10 @@ update_active (struct cpu_features *cpu_features) + unsigned int amx_size + = (xstate_amx_comp_offsets[31] + + xstate_amx_comp_sizes[31]); +- amx_size = ALIGN_UP (amx_size + STATE_SAVE_OFFSET, +- 64); ++ amx_size ++ = ALIGN_UP ((amx_size ++ + TLSDESC_CALL_REGISTER_SAVE_AREA), ++ 64); + /* Set xsave_state_full_size to the compact AMX + state size for XSAVEC. NB: xsave_state_full_size + is only used in _dl_tlsdesc_dynamic_xsave and +@@ -409,7 +411,8 @@ update_active (struct cpu_features *cpu_features) + cpu_features->xsave_state_full_size = amx_size; + #endif + cpu_features->xsave_state_size +- = ALIGN_UP (size + STATE_SAVE_OFFSET, 64); ++ = ALIGN_UP (size + TLSDESC_CALL_REGISTER_SAVE_AREA, ++ 64); + CPU_FEATURE_SET (cpu_features, XSAVEC); + } + } +diff --git a/sysdeps/x86/sysdep.h b/sysdeps/x86/sysdep.h +index db8e576e91767db5..7359149e17ccf341 100644 +--- a/sysdeps/x86/sysdep.h ++++ b/sysdeps/x86/sysdep.h +@@ -38,14 +38,59 @@ + #ifdef __x86_64__ + /* Offset for fxsave/xsave area used by _dl_runtime_resolve. Also need + space to preserve RCX, RDX, RSI, RDI, R8, R9 and RAX. It must be +- aligned to 16 bytes for fxsave and 64 bytes for xsave. +- +- NB: Is is non-zero because of the 128-byte red-zone. Some registers +- are saved on stack without adjusting stack pointer first. When we +- update stack pointer to allocate more space, we need to take the +- red-zone into account. */ ++ aligned to 16 bytes for fxsave and 64 bytes for xsave. It is non-zero ++ because MOV, instead of PUSH, is used to save registers onto stack. ++ ++ +==================+<- stack frame start aligned at 8 or 16 bytes ++ | |<- paddings for stack realignment of 64 bytes ++ |------------------|<- xsave buffer end aligned at 64 bytes ++ | |<- ++ | |<- ++ | |<- ++ |------------------|<- xsave buffer start at STATE_SAVE_OFFSET(%rsp) ++ | |<- 8-byte padding for 64-byte alignment ++ | |<- R9 ++ | |<- R8 ++ | |<- RDI ++ | |<- RSI ++ | |<- RDX ++ | |<- RCX ++ | |<- RAX ++ +==================+<- RSP aligned at 64 bytes ++ ++ */ + # define STATE_SAVE_OFFSET (8 * 7 + 8) + ++/* _dl_tlsdesc_dynamic preserves RDI, RSI and RBX before realigning ++ stack. After realigning stack, it saves RCX, RDX, R8, R9, R10 and ++ R11. Allocate space for RDI, RSI and RBX to avoid clobbering saved ++ RDI, RSI and RBX values on stack by xsave. ++ ++ +==================+<- stack frame start aligned at 8 or 16 bytes ++ | |<- RDI saved in the red zone ++ | |<- RSI saved in the red zone ++ | |<- RBX saved in the red zone ++ | |<- paddings for stack realignment of 64 bytes ++ |------------------|<- xsave buffer end aligned at 64 bytes ++ | |<- ++ | |<- ++ | |<- ++ |------------------|<- xsave buffer start at STATE_SAVE_OFFSET(%rsp) ++ | |<- 8-byte padding for 64-byte alignment ++ | |<- 8-byte padding for 64-byte alignment ++ | |<- R11 ++ | |<- R10 ++ | |<- R9 ++ | |<- R8 ++ | |<- RDX ++ | |<- RCX ++ +==================+<- RSP aligned at 64 bytes ++ ++ Define the total register save area size for all integer registers by ++ adding 24 to STATE_SAVE_OFFSET since RDI, RSI and RBX are saved onto ++ stack without adjusting stack pointer first, using the red-zone. */ ++# define TLSDESC_CALL_REGISTER_SAVE_AREA (STATE_SAVE_OFFSET + 24) ++ + /* Save SSE, AVX, AVX512, mask, bound and APX registers. Bound and APX + registers are mutually exclusive. */ + # define STATE_SAVE_MASK \ +@@ -66,8 +111,9 @@ + (STATE_SAVE_MASK | AMX_STATE_SAVE_MASK) + #else + /* Offset for fxsave/xsave area used by _dl_tlsdesc_dynamic. Since i386 +- doesn't have red-zone, use 0 here. */ ++ uses PUSH to save registers onto stack, use 0 here. */ + # define STATE_SAVE_OFFSET 0 ++# define TLSDESC_CALL_REGISTER_SAVE_AREA 0 + + /* Save SSE, AVX, AXV512, mask and bound registers. */ + # define STATE_SAVE_MASK \ +diff --git a/sysdeps/x86_64/tst-gnu2-tls2mod1.S b/sysdeps/x86_64/tst-gnu2-tls2mod1.S +new file mode 100644 +index 0000000000000000..1d636669ba255724 +--- /dev/null ++++ b/sysdeps/x86_64/tst-gnu2-tls2mod1.S +@@ -0,0 +1,87 @@ ++/* Check if TLSDESC relocation preserves %rdi, %rsi and %rbx. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++ ++/* On AVX512 machines, OFFSET == 40 caused _dl_tlsdesc_dynamic_xsavec ++ to clobber %rdi, %rsi and %rbx. On Intel AVX CPUs, the state size ++ is 960 bytes and this test didn't fail. It may be due to the unused ++ last 128 bytes. On AMD AVX CPUs, the state size is 832 bytes and ++ this test might fail without the fix. */ ++#ifndef OFFSET ++# define OFFSET 40 ++#endif ++ ++ .text ++ .p2align 4 ++ .globl apply_tls ++ .type apply_tls, @function ++apply_tls: ++ cfi_startproc ++ _CET_ENDBR ++ pushq %rbp ++ cfi_def_cfa_offset (16) ++ cfi_offset (6, -16) ++ movdqu (%RDI_LP), %xmm0 ++ lea tls_var1@TLSDESC(%rip), %RAX_LP ++ mov %RSP_LP, %RBP_LP ++ cfi_def_cfa_register (6) ++ /* Align stack to 64 bytes. */ ++ and $-64, %RSP_LP ++ sub $OFFSET, %RSP_LP ++ pushq %rbx ++ /* Set %ebx to 0xbadbeef. */ ++ movl $0xbadbeef, %ebx ++ movl $0xbadbeef, %esi ++ movq %rdi, saved_rdi(%rip) ++ movq %rsi, saved_rsi(%rip) ++ call *tls_var1@TLSCALL(%RAX_LP) ++ /* Check if _dl_tlsdesc_dynamic preserves %rdi, %rsi and %rbx. */ ++ cmpq saved_rdi(%rip), %rdi ++ jne L(hlt) ++ cmpq saved_rsi(%rip), %rsi ++ jne L(hlt) ++ cmpl $0xbadbeef, %ebx ++ jne L(hlt) ++ add %fs:0, %RAX_LP ++ movups %xmm0, 32(%RAX_LP) ++ movdqu 16(%RDI_LP), %xmm1 ++ mov %RAX_LP, %RBX_LP ++ movups %xmm1, 48(%RAX_LP) ++ lea 32(%RBX_LP), %RAX_LP ++ pop %rbx ++ leave ++ cfi_def_cfa (7, 8) ++ ret ++L(hlt): ++ hlt ++ cfi_endproc ++ .size apply_tls, .-apply_tls ++ .hidden tls_var1 ++ .globl tls_var1 ++ .section .tbss,"awT",@nobits ++ .align 16 ++ .type tls_var1, @object ++ .size tls_var1, 3200 ++tls_var1: ++ .zero 3200 ++ .local saved_rdi ++ .comm saved_rdi,8,8 ++ .local saved_rsi ++ .comm saved_rsi,8,8 ++ .section .note.GNU-stack,"",@progbits diff --git a/SOURCES/glibc-upstream-2.39-15.patch b/SOURCES/glibc-upstream-2.39-15.patch new file mode 100644 index 0000000..037a4d3 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-15.patch @@ -0,0 +1,38 @@ +commit 15aebdbada54098787715448c94701f17033fc92 +Author: Adhemerval Zanella +Date: Tue Mar 12 13:21:18 2024 -0300 + + Ignore undefined symbols for -mtls-dialect=gnu2 + + So it does not fail for arm config that defaults to -mtp=soft (which + issues a call to __aeabi_read_tp). + Reviewed-by: H.J. Lu + + (cherry picked from commit 968b0ca9440040a2b31248a572891f0e55c1ab10) + +diff --git a/configure b/configure +index 59ff1e415dda4fbf..117b48a421792eda 100755 +--- a/configure ++++ b/configure +@@ -7020,7 +7020,7 @@ void foo (void) + } + EOF + if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fPIC -mtls-dialect=gnu2 -nostdlib -nostartfiles +- conftest.c -o conftest 1>&5' ++ -shared conftest.c -o conftest 1>&5' + { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5 + (eval $ac_try) 2>&5 + ac_status=$? +diff --git a/configure.ac b/configure.ac +index 65799e56852a5356..19b88a47a52508a1 100644 +--- a/configure.ac ++++ b/configure.ac +@@ -1297,7 +1297,7 @@ void foo (void) + } + EOF + if AC_TRY_COMMAND([${CC-cc} $CFLAGS $CPPFLAGS -fPIC -mtls-dialect=gnu2 -nostdlib -nostartfiles +- conftest.c -o conftest 1>&AS_MESSAGE_LOG_FD]) ++ -shared conftest.c -o conftest 1>&AS_MESSAGE_LOG_FD]) + then + libc_cv_mtls_dialect_gnu2=yes + else diff --git a/SOURCES/glibc-upstream-2.39-16.patch b/SOURCES/glibc-upstream-2.39-16.patch new file mode 100644 index 0000000..e5fa640 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-16.patch @@ -0,0 +1,446 @@ +commit a8ba52bde58c69f2b31da62ad2311f119adf6cb9 +Author: Adhemerval Zanella +Date: Tue Mar 12 13:21:19 2024 -0300 + + arm: Update _dl_tlsdesc_dynamic to preserve caller-saved registers (BZ 31372) + + ARM _dl_tlsdesc_dynamic slow path has two issues: + + * The ip/r12 is defined by AAPCS as a scratch register, and gcc is + used to save the stack pointer before on some function calls. So it + should also be saved/restored as well. It fixes the tst-gnu2-tls2. + + * None of the possible VFP registers are saved/restored. ARM has the + additional complexity to have different VFP bank sizes (depending of + VFP support by the chip). + + The tst-gnu2-tls2 test is extended to check for VFP registers, although + only for hardfp builds. Different than setcontext, _dl_tlsdesc_dynamic + does not have HWCAP_ARM_IWMMXT (I don't have a way to properly test + it and it is almost a decade since newer hardware was released). + + With this patch there is no need to mark tst-gnu2-tls2 as XFAIL. + + Checked on arm-linux-gnueabihf. + Reviewed-by: H.J. Lu + + (cherry picked from commit 64c7e344289ed085517c2227d8e3b06388242c13) + +diff --git a/config.h.in b/config.h.in +index 44a34072a47aa008..4d33c63a841d3d6d 100644 +--- a/config.h.in ++++ b/config.h.in +@@ -141,6 +141,9 @@ + /* LOONGARCH floating-point ABI for ld.so. */ + #undef LOONGARCH_ABI_FRLEN + ++/* Define whether ARM used hard-float and support VFPvX-D32. */ ++#undef HAVE_ARM_PCS_VFP_D32 ++ + /* Linux specific: minimum supported kernel version. */ + #undef __LINUX_KERNEL_VERSION + +diff --git a/elf/Makefile b/elf/Makefile +index c5c37a9147e69d83..030db4d207d3491e 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -3056,10 +3056,6 @@ $(objpfx)tst-gnu2-tls2.out: \ + $(objpfx)tst-gnu2-tls2mod2.so + + ifeq (yes,$(have-mtls-dialect-gnu2)) +-# This test fails if dl_tlsdesc_dynamic doesn't preserve all caller-saved +-# registers. See https://sourceware.org/bugzilla/show_bug.cgi?id=31372 +-test-xfail-tst-gnu2-tls2 = yes +- + CFLAGS-tst-tlsgap-mod0.c += -mtls-dialect=gnu2 + CFLAGS-tst-tlsgap-mod1.c += -mtls-dialect=gnu2 + CFLAGS-tst-tlsgap-mod2.c += -mtls-dialect=gnu2 +diff --git a/elf/tst-gnu2-tls2.h b/elf/tst-gnu2-tls2.h +index 77964a57a352e6a4..1ade8151e200af68 100644 +--- a/elf/tst-gnu2-tls2.h ++++ b/elf/tst-gnu2-tls2.h +@@ -27,6 +27,10 @@ extern struct tls *apply_tls (struct tls *); + + /* An architecture can define them to verify that clobber caller-saved + registers aren't changed by the implicit TLSDESC call. */ ++#ifndef INIT_TLSDESC_CALL ++# define INIT_TLSDESC_CALL() ++#endif ++ + #ifndef BEFORE_TLSDESC_CALL + # define BEFORE_TLSDESC_CALL() + #endif +diff --git a/elf/tst-gnu2-tls2mod0.c b/elf/tst-gnu2-tls2mod0.c +index 45556a0e173922cc..3fe3c142777abe04 100644 +--- a/elf/tst-gnu2-tls2mod0.c ++++ b/elf/tst-gnu2-tls2mod0.c +@@ -16,13 +16,14 @@ + License along with the GNU C Library; if not, see + . */ + +-#include "tst-gnu2-tls2.h" ++#include + + __thread struct tls tls_var0 __attribute__ ((visibility ("hidden"))); + + struct tls * + apply_tls (struct tls *p) + { ++ INIT_TLSDESC_CALL (); + BEFORE_TLSDESC_CALL (); + tls_var0 = *p; + struct tls *ret = &tls_var0; +diff --git a/elf/tst-gnu2-tls2mod1.c b/elf/tst-gnu2-tls2mod1.c +index e10b9dbc0a7573c7..e2105384689e2d2e 100644 +--- a/elf/tst-gnu2-tls2mod1.c ++++ b/elf/tst-gnu2-tls2mod1.c +@@ -16,13 +16,14 @@ + License along with the GNU C Library; if not, see + . */ + +-#include "tst-gnu2-tls2.h" ++#include + + __thread struct tls tls_var1[100] __attribute__ ((visibility ("hidden"))); + + struct tls * + apply_tls (struct tls *p) + { ++ INIT_TLSDESC_CALL (); + BEFORE_TLSDESC_CALL (); + tls_var1[1] = *p; + struct tls *ret = &tls_var1[1]; +diff --git a/elf/tst-gnu2-tls2mod2.c b/elf/tst-gnu2-tls2mod2.c +index 141af51e55b8bf34..6d3031dc5fbc1041 100644 +--- a/elf/tst-gnu2-tls2mod2.c ++++ b/elf/tst-gnu2-tls2mod2.c +@@ -16,13 +16,14 @@ + License along with the GNU C Library; if not, see + . */ + +-#include "tst-gnu2-tls2.h" ++#include + + __thread struct tls tls_var2 __attribute__ ((visibility ("hidden"))); + + struct tls * + apply_tls (struct tls *p) + { ++ INIT_TLSDESC_CALL (); + BEFORE_TLSDESC_CALL (); + tls_var2 = *p; + struct tls *ret = &tls_var2; +diff --git a/sysdeps/arm/configure b/sysdeps/arm/configure +index 35e2918922300956..4ef4d46cbd5384e9 100644 +--- a/sysdeps/arm/configure ++++ b/sysdeps/arm/configure +@@ -187,6 +187,38 @@ else + default-abi = soft" + fi + ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking whether VFP supports 32 registers" >&5 ++printf %s "checking whether VFP supports 32 registers... " >&6; } ++if test ${libc_cv_arm_pcs_vfp_d32+y} ++then : ++ printf %s "(cached) " >&6 ++else $as_nop ++ ++cat confdefs.h - <<_ACEOF >conftest.$ac_ext ++/* end confdefs.h. */ ++ ++void foo (void) ++{ ++ asm volatile ("vldr d16,=17" : : : "d16"); ++} ++ ++_ACEOF ++if ac_fn_c_try_compile "$LINENO" ++then : ++ libc_cv_arm_pcs_vfp_d32=yes ++else $as_nop ++ libc_cv_arm_pcs_vfp_d32=no ++fi ++rm -f core conftest.err conftest.$ac_objext conftest.beam conftest.$ac_ext ++fi ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_arm_pcs_vfp_d32" >&5 ++printf "%s\n" "$libc_cv_arm_pcs_vfp_d32" >&6; } ++if test "$libc_cv_arm_pcs_vfp_d32" = yes ; ++then ++ printf "%s\n" "#define HAVE_ARM_PCS_VFP_D32 1" >>confdefs.h ++ ++fi ++ + { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking whether PC-relative relocs in movw/movt work properly" >&5 + printf %s "checking whether PC-relative relocs in movw/movt work properly... " >&6; } + if test ${libc_cv_arm_pcrel_movw+y} +diff --git a/sysdeps/arm/configure.ac b/sysdeps/arm/configure.ac +index 5172e30bbe79f995..cd00ddc9d9ade5d7 100644 +--- a/sysdeps/arm/configure.ac ++++ b/sysdeps/arm/configure.ac +@@ -21,6 +21,21 @@ else + LIBC_CONFIG_VAR([default-abi], [soft]) + fi + ++AC_CACHE_CHECK([whether VFP supports 32 registers], ++ libc_cv_arm_pcs_vfp_d32, [ ++AC_COMPILE_IFELSE([AC_LANG_SOURCE([[ ++void foo (void) ++{ ++ asm volatile ("vldr d16,=17" : : : "d16"); ++} ++]])], ++ [libc_cv_arm_pcs_vfp_d32=yes], ++ [libc_cv_arm_pcs_vfp_d32=no])]) ++if test "$libc_cv_arm_pcs_vfp_d32" = yes ; ++then ++ AC_DEFINE(HAVE_ARM_PCS_VFP_D32) ++fi ++ + AC_CACHE_CHECK([whether PC-relative relocs in movw/movt work properly], + libc_cv_arm_pcrel_movw, [ + cat > conftest.s <<\EOF +diff --git a/sysdeps/arm/dl-tlsdesc.S b/sysdeps/arm/dl-tlsdesc.S +index 764c56e70f046b03..ada106521da8971a 100644 +--- a/sysdeps/arm/dl-tlsdesc.S ++++ b/sysdeps/arm/dl-tlsdesc.S +@@ -19,6 +19,7 @@ + #include + #include + #include ++#include + #include "tlsdesc.h" + + .text +@@ -83,14 +84,20 @@ _dl_tlsdesc_dynamic(struct tlsdesc *tdp) + .align 2 + _dl_tlsdesc_dynamic: + /* Our calling convention is to clobber r0, r1 and the processor +- flags. All others that are modified must be saved */ +- eabi_save ({r2,r3,r4,lr}) +- push {r2,r3,r4,lr} +- cfi_adjust_cfa_offset (16) ++ flags. All others that are modified must be saved. r5 is ++ used as the hwcap value to avoid reload after __tls_get_addr ++ call. If required we will save the vector register on the slow ++ path. */ ++ eabi_save ({r2,r3,r4,r5,ip,lr}) ++ push {r2,r3,r4,r5,ip,lr} ++ cfi_adjust_cfa_offset (24) + cfi_rel_offset (r2,0) + cfi_rel_offset (r3,4) + cfi_rel_offset (r4,8) +- cfi_rel_offset (lr,12) ++ cfi_rel_offset (r5,12) ++ cfi_rel_offset (ip,16) ++ cfi_rel_offset (lr,20) ++ + ldr r1, [r0] /* td */ + GET_TLS (lr) + mov r4, r0 /* r4 = tp */ +@@ -113,22 +120,69 @@ _dl_tlsdesc_dynamic: + rsbne r0, r4, r3 + bne 2f + 1: mov r0, r1 ++ ++ /* Load the hwcap to check for vector support. */ ++ ldr r2, 3f ++ ldr r1, .Lrtld_global_ro ++0: add r2, pc, r2 ++ ldr r2, [r2, r1] ++ ldr r5, [r2, #RTLD_GLOBAL_RO_DL_HWCAP_OFFSET] ++ ++#ifdef __SOFTFP__ ++ tst r5, #HWCAP_ARM_VFP ++ beq .Lno_vfp ++#endif ++ ++ /* Store the VFP registers. Don't use VFP instructions directly ++ because this code is used in non-VFP multilibs. */ ++#define VFP_STACK_REQ (32*8 + 8) ++ sub sp, sp, VFP_STACK_REQ ++ cfi_adjust_cfa_offset (VFP_STACK_REQ) ++ mov r3, sp ++ .inst 0xeca30b20 /* vstmia r3!, {d0-d15} */ ++ tst r5, #HWCAP_ARM_VFPD32 ++ beq 4f ++ .inst 0xece30b20 /* vstmia r3!, {d16-d31} */ ++ /* Store the floating-point status register. */ ++4: .inst 0xeef12a10 /* vmrs r2, fpscr */ ++ str r2, [r3] ++.Lno_vfp: + bl __tls_get_addr + rsb r0, r4, r0 ++#ifdef __SOFTFP__ ++ tst r5, #HWCAP_ARM_VFP ++ beq 2f ++#endif ++ mov r3, sp ++ .inst 0xecb30b20 /* vldmia r3!, {d0-d15} */ ++ tst r5, #HWCAP_ARM_VFPD32 ++ beq 5f ++ .inst 0xecf30b20 /* vldmia r3!, {d16-d31} */ ++ ldr r4, [r3] ++5: .inst 0xeee14a10 /* vmsr fpscr, r4 */ ++ add sp, sp, VFP_STACK_REQ ++ cfi_adjust_cfa_offset (-VFP_STACK_REQ) ++ + 2: + #if ((defined (__ARM_ARCH_4T__) && defined (__THUMB_INTERWORK__)) \ + || defined (ARM_ALWAYS_BX)) +- pop {r2,r3,r4, lr} +- cfi_adjust_cfa_offset (-16) ++ pop {r2,r3,r4,r5,ip, lr} ++ cfi_adjust_cfa_offset (-20) + cfi_restore (lr) ++ cfi_restore (ip) ++ cfi_restore (r5) + cfi_restore (r4) + cfi_restore (r3) + cfi_restore (r2) + bx lr + #else +- pop {r2,r3,r4, pc} ++ pop {r2,r3,r4,r5,ip, pc} + #endif + eabi_fnend + cfi_endproc + .size _dl_tlsdesc_dynamic, .-_dl_tlsdesc_dynamic ++ ++3: .long _GLOBAL_OFFSET_TABLE_ - 0b - PC_OFS ++.Lrtld_global_ro: ++ .long C_SYMBOL_NAME(_rtld_global_ro)(GOT) + #endif /* SHARED */ +diff --git a/sysdeps/arm/tst-gnu2-tls2.h b/sysdeps/arm/tst-gnu2-tls2.h +new file mode 100644 +index 0000000000000000..e413ac21fb9ed9bf +--- /dev/null ++++ b/sysdeps/arm/tst-gnu2-tls2.h +@@ -0,0 +1,128 @@ ++/* Test TLSDESC relocation. ARM version. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++#include ++ ++#ifndef __SOFTFP__ ++ ++# ifdef HAVE_ARM_PCS_VFP_D32 ++# define SAVE_VFP_D32 \ ++ asm volatile ("vldr d16,=17" : : : "d16"); \ ++ asm volatile ("vldr d17,=18" : : : "d17"); \ ++ asm volatile ("vldr d18,=19" : : : "d18"); \ ++ asm volatile ("vldr d19,=20" : : : "d19"); \ ++ asm volatile ("vldr d20,=21" : : : "d20"); \ ++ asm volatile ("vldr d21,=22" : : : "d21"); \ ++ asm volatile ("vldr d22,=23" : : : "d22"); \ ++ asm volatile ("vldr d23,=24" : : : "d23"); \ ++ asm volatile ("vldr d24,=25" : : : "d24"); \ ++ asm volatile ("vldr d25,=26" : : : "d25"); \ ++ asm volatile ("vldr d26,=27" : : : "d26"); \ ++ asm volatile ("vldr d27,=28" : : : "d27"); \ ++ asm volatile ("vldr d28,=29" : : : "d28"); \ ++ asm volatile ("vldr d29,=30" : : : "d29"); \ ++ asm volatile ("vldr d30,=31" : : : "d30"); \ ++ asm volatile ("vldr d31,=32" : : : "d31"); ++# else ++# define SAVE_VFP_D32 ++# endif ++ ++# define INIT_TLSDESC_CALL() \ ++ unsigned long hwcap = getauxval (AT_HWCAP) ++ ++/* Set each vector register to a value from 1 to 32 before the TLS access, ++ dump to memory after TLS access, and compare with the expected values. */ ++ ++# define BEFORE_TLSDESC_CALL() \ ++ if (hwcap & HWCAP_ARM_VFP) \ ++ { \ ++ asm volatile ("vldr d0,=1" : : : "d0"); \ ++ asm volatile ("vldr d1,=2" : : : "d1"); \ ++ asm volatile ("vldr d2,=3" : : : "d1"); \ ++ asm volatile ("vldr d3,=4" : : : "d3"); \ ++ asm volatile ("vldr d4,=5" : : : "d4"); \ ++ asm volatile ("vldr d5,=6" : : : "d5"); \ ++ asm volatile ("vldr d6,=7" : : : "d6"); \ ++ asm volatile ("vldr d7,=8" : : : "d7"); \ ++ asm volatile ("vldr d8,=9" : : : "d8"); \ ++ asm volatile ("vldr d9,=10" : : : "d9"); \ ++ asm volatile ("vldr d10,=11" : : : "d10"); \ ++ asm volatile ("vldr d11,=12" : : : "d11"); \ ++ asm volatile ("vldr d12,=13" : : : "d12"); \ ++ asm volatile ("vldr d13,=14" : : : "d13"); \ ++ asm volatile ("vldr d14,=15" : : : "d14"); \ ++ asm volatile ("vldr d15,=16" : : : "d15"); \ ++ } \ ++ if (hwcap & HWCAP_ARM_VFPD32) \ ++ { \ ++ SAVE_VFP_D32 \ ++ } ++ ++# define VFP_STACK_REQ (16*8) ++# if __BYTE_ORDER == __BIG_ENDIAN ++# define DISP 7 ++# else ++# define DISP 0 ++# endif ++ ++# ifdef HAVE_ARM_PCS_VFP_D32 ++# define CHECK_VFP_D32 \ ++ char vfp[VFP_STACK_REQ]; \ ++ asm volatile ("vstmia %0, {d16-d31}\n" \ ++ : \ ++ : "r" (vfp) \ ++ : "memory"); \ ++ \ ++ char expected[VFP_STACK_REQ] = { 0 }; \ ++ for (int i = 0; i < 16; ++i) \ ++ expected[i * 8 + DISP] = i + 17; \ ++ \ ++ if (memcmp (vfp, expected, VFP_STACK_REQ) != 0) \ ++ abort (); ++# else ++# define CHECK_VFP_D32 ++# endif ++ ++# define AFTER_TLSDESC_CALL() \ ++ if (hwcap & HWCAP_ARM_VFP) \ ++ { \ ++ char vfp[VFP_STACK_REQ]; \ ++ asm volatile ("vstmia %0, {d0-d15}\n" \ ++ : \ ++ : "r" (vfp) \ ++ : "memory"); \ ++ \ ++ char expected[VFP_STACK_REQ] = { 0 }; \ ++ for (int i = 0; i < 16; ++i) \ ++ expected[i * 8 + DISP] = i + 1; \ ++ \ ++ if (memcmp (vfp, expected, VFP_STACK_REQ) != 0) \ ++ abort (); \ ++ } \ ++ if (hwcap & HWCAP_ARM_VFPD32) \ ++ { \ ++ CHECK_VFP_D32 \ ++ } ++ ++#endif /* __SOFTFP__ */ ++ ++#include_next diff --git a/SOURCES/glibc-upstream-2.39-17.patch b/SOURCES/glibc-upstream-2.39-17.patch new file mode 100644 index 0000000..050355c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-17.patch @@ -0,0 +1,221 @@ +commit aded2fc004e7ee85cf0b45b1382552d41e555a23 +Author: Adhemerval Zanella +Date: Tue Mar 12 13:21:20 2024 -0300 + + elf: Enable TLS descriptor tests on aarch64 + + The aarch64 uses 'trad' for traditional tls and 'desc' for tls + descriptors, but unlike other targets it defaults to 'desc'. The + gnutls2 configure check does not set aarch64 as an ABI that uses + TLS descriptors, which then disable somes stests. + + Also rename the internal machinery fron gnu2 to tls descriptors. + + Checked on aarch64-linux-gnu. + Reviewed-by: H.J. Lu + + (cherry picked from commit 3d53d18fc71c5d9ef4773b8bce04d54b80181926) + +diff --git a/configure b/configure +index 117b48a421792eda..432e40a59295cffd 100755 +--- a/configure ++++ b/configure +@@ -653,7 +653,7 @@ LIBGD + libc_cv_cc_loop_to_function + libc_cv_cc_submachine + libc_cv_cc_nofma +-libc_cv_mtls_dialect_gnu2 ++libc_cv_mtls_descriptor + libc_cv_has_glob_dat + libc_cv_fpie + libc_cv_z_execstack +@@ -4760,6 +4760,9 @@ libc_config_ok=no + # whether to use such directories. + with_fp_cond=1 + ++# A preconfigure script may define another name to TLS descriptor variant ++mtls_descriptor=gnu2 ++ + if frags=`ls -d $srcdir/sysdeps/*/preconfigure 2> /dev/null` + then + { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for sysdeps preconfigure fragments" >&5 +@@ -7006,9 +7009,9 @@ fi + printf "%s\n" "$libc_cv_has_glob_dat" >&6; } + + +-{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for -mtls-dialect=gnu2" >&5 +-printf %s "checking for -mtls-dialect=gnu2... " >&6; } +-if test ${libc_cv_mtls_dialect_gnu2+y} ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for tls descriptor support" >&5 ++printf %s "checking for tls descriptor support... " >&6; } ++if test ${libc_cv_mtls_descriptor+y} + then : + printf %s "(cached) " >&6 + else $as_nop +@@ -7019,7 +7022,7 @@ void foo (void) + i = 10; + } + EOF +-if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fPIC -mtls-dialect=gnu2 -nostdlib -nostartfiles ++if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fPIC -mtls-dialect=$mtls_descriptor -nostdlib -nostartfiles + -shared conftest.c -o conftest 1>&5' + { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5 + (eval $ac_try) 2>&5 +@@ -7027,17 +7030,17 @@ if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fPIC -mtls-dialect=gnu2 -nostdlib -nost + printf "%s\n" "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5 + test $ac_status = 0; }; } + then +- libc_cv_mtls_dialect_gnu2=yes ++ libc_cv_mtls_descriptor=$mtls_descriptor + else +- libc_cv_mtls_dialect_gnu2=no ++ libc_cv_mtls_descriptor=no + fi + rm -f conftest* + fi +-{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_mtls_dialect_gnu2" >&5 +-printf "%s\n" "$libc_cv_mtls_dialect_gnu2" >&6; } ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_mtls_descriptor" >&5 ++printf "%s\n" "$libc_cv_mtls_descriptor" >&6; } + + config_vars="$config_vars +-have-mtls-dialect-gnu2 = $libc_cv_mtls_dialect_gnu2" ++have-mtls-descriptor = $libc_cv_mtls_descriptor" + + { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking if -Wno-ignored-attributes is required for aliases" >&5 + printf %s "checking if -Wno-ignored-attributes is required for aliases... " >&6; } +diff --git a/configure.ac b/configure.ac +index 19b88a47a52508a1..bdc385d03c3dc7f5 100644 +--- a/configure.ac ++++ b/configure.ac +@@ -442,6 +442,9 @@ libc_config_ok=no + # whether to use such directories. + with_fp_cond=1 + ++# A preconfigure script may define another name to TLS descriptor variant ++mtls_descriptor=gnu2 ++ + dnl Let sysdeps/*/preconfigure act here. + LIBC_PRECONFIGURE([$srcdir], [for sysdeps]) + +@@ -1287,7 +1290,7 @@ fi + rm -f conftest*]) + AC_SUBST(libc_cv_has_glob_dat) + +-AC_CACHE_CHECK([for -mtls-dialect=gnu2], libc_cv_mtls_dialect_gnu2, ++AC_CACHE_CHECK([for tls descriptor support], libc_cv_mtls_descriptor, + [dnl + cat > conftest.c <&AS_MESSAGE_LOG_FD]) + then +- libc_cv_mtls_dialect_gnu2=yes ++ libc_cv_mtls_descriptor=$mtls_descriptor + else +- libc_cv_mtls_dialect_gnu2=no ++ libc_cv_mtls_descriptor=no + fi + rm -f conftest*]) +-AC_SUBST(libc_cv_mtls_dialect_gnu2) +-LIBC_CONFIG_VAR([have-mtls-dialect-gnu2], [$libc_cv_mtls_dialect_gnu2]) ++AC_SUBST(libc_cv_mtls_descriptor) ++LIBC_CONFIG_VAR([have-mtls-descriptor], [$libc_cv_mtls_descriptor]) + + dnl clang emits an warning for a double alias redirection, to warn the + dnl original symbol is sed even when weak definition overrides it. +diff --git a/elf/Makefile b/elf/Makefile +index 030db4d207d3491e..69aa423c4b90127d 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -999,13 +999,13 @@ modules-names-tests = $(filter-out ifuncmod% tst-tlsmod%,\ + # For +depfiles in Makerules. + extra-test-objs += tst-auditmod17.os + +-ifeq (yes,$(have-mtls-dialect-gnu2)) ++ifneq (no,$(have-mtls-descriptor)) + tests += tst-gnu2-tls1 + modules-names += tst-gnu2-tls1mod + $(objpfx)tst-gnu2-tls1: $(objpfx)tst-gnu2-tls1mod.so + tst-gnu2-tls1mod.so-no-z-defs = yes +-CFLAGS-tst-gnu2-tls1mod.c += -mtls-dialect=gnu2 +-endif # $(have-mtls-dialect-gnu2) ++CFLAGS-tst-gnu2-tls1mod.c += -mtls-dialect=$(have-mtls-descriptor) ++endif # $(have-mtls-descriptor) + + ifeq (yes,$(have-protected-data)) + modules-names += tst-protected1moda tst-protected1modb +@@ -2972,11 +2972,11 @@ $(objpfx)tst-tls-allocation-failure-static-patched.out: \ + $(objpfx)tst-audit-tlsdesc: $(objpfx)tst-audit-tlsdesc-mod1.so \ + $(objpfx)tst-audit-tlsdesc-mod2.so \ + $(shared-thread-library) +-ifeq (yes,$(have-mtls-dialect-gnu2)) ++ifneq (no,$(have-mtls-descriptor)) + # The test is valid for all TLS types, but we want to exercise GNU2 + # TLS if possible. +-CFLAGS-tst-audit-tlsdesc-mod1.c += -mtls-dialect=gnu2 +-CFLAGS-tst-audit-tlsdesc-mod2.c += -mtls-dialect=gnu2 ++CFLAGS-tst-audit-tlsdesc-mod1.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-audit-tlsdesc-mod2.c += -mtls-dialect=$(have-mtls-descriptor) + endif + $(objpfx)tst-audit-tlsdesc-dlopen: $(shared-thread-library) + $(objpfx)tst-audit-tlsdesc-dlopen.out: $(objpfx)tst-audit-tlsdesc-mod1.so \ +@@ -3055,11 +3055,11 @@ $(objpfx)tst-gnu2-tls2.out: \ + $(objpfx)tst-gnu2-tls2mod1.so \ + $(objpfx)tst-gnu2-tls2mod2.so + +-ifeq (yes,$(have-mtls-dialect-gnu2)) +-CFLAGS-tst-tlsgap-mod0.c += -mtls-dialect=gnu2 +-CFLAGS-tst-tlsgap-mod1.c += -mtls-dialect=gnu2 +-CFLAGS-tst-tlsgap-mod2.c += -mtls-dialect=gnu2 +-CFLAGS-tst-gnu2-tls2mod0.c += -mtls-dialect=gnu2 +-CFLAGS-tst-gnu2-tls2mod1.c += -mtls-dialect=gnu2 +-CFLAGS-tst-gnu2-tls2mod2.c += -mtls-dialect=gnu2 ++ifneq (no,$(have-mtls-descriptor)) ++CFLAGS-tst-tlsgap-mod0.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-tlsgap-mod1.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-tlsgap-mod2.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-gnu2-tls2mod0.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-gnu2-tls2mod1.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-gnu2-tls2mod2.c += -mtls-dialect=$(have-mtls-descriptor) + endif +diff --git a/sysdeps/aarch64/preconfigure b/sysdeps/aarch64/preconfigure +index d9bd1f8558a079cb..19657b627bc84c4e 100644 +--- a/sysdeps/aarch64/preconfigure ++++ b/sysdeps/aarch64/preconfigure +@@ -2,5 +2,6 @@ case "$machine" in + aarch64*) + base_machine=aarch64 + machine=aarch64 ++ mtls_descriptor=desc + ;; + esac +diff --git a/sysdeps/arm/Makefile b/sysdeps/arm/Makefile +index d5cea717a9c201aa..619474eca94fe8e4 100644 +--- a/sysdeps/arm/Makefile ++++ b/sysdeps/arm/Makefile +@@ -13,15 +13,15 @@ $(objpfx)libgcc-stubs.a: $(objpfx)aeabi_unwind_cpp_pr1.os + lib-noranlib: $(objpfx)libgcc-stubs.a + + ifeq ($(build-shared),yes) +-ifeq (yes,$(have-mtls-dialect-gnu2)) ++ifneq (no,$(have-mtls-descriptor)) + tests += tst-armtlsdescloc tst-armtlsdescextnow tst-armtlsdescextlazy + modules-names += tst-armtlsdesclocmod + modules-names += tst-armtlsdescextlazymod tst-armtlsdescextnowmod + CPPFLAGS-tst-armtlsdescextnowmod.c += -Dstatic= + CPPFLAGS-tst-armtlsdescextlazymod.c += -Dstatic= +-CFLAGS-tst-armtlsdesclocmod.c += -mtls-dialect=gnu2 +-CFLAGS-tst-armtlsdescextnowmod.c += -mtls-dialect=gnu2 +-CFLAGS-tst-armtlsdescextlazymod.c += -mtls-dialect=gnu2 ++CFLAGS-tst-armtlsdesclocmod.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-armtlsdescextnowmod.c += -mtls-dialect=$(have-mtls-descriptor) ++CFLAGS-tst-armtlsdescextlazymod.c += -mtls-dialect=$(have-mtls-descriptor) + LDFLAGS-tst-armtlsdescextnowmod.so += -Wl,-z,now + tst-armtlsdescloc-ENV = LD_BIND_NOW=1 + tst-armtlsdescextnow-ENV = LD_BIND_NOW=1 diff --git a/SOURCES/glibc-upstream-2.39-18.patch b/SOURCES/glibc-upstream-2.39-18.patch new file mode 100644 index 0000000..60a31ff --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-18.patch @@ -0,0 +1,24 @@ +commit 5a461f2949ded98d8211939f84988bc464c7b4fe +Author: Andreas Schwab +Date: Tue Mar 19 13:49:50 2024 +0100 + + Add tst-gnu2-tls2mod1 to test-internal-extras + + That allows sysdeps/x86_64/tst-gnu2-tls2mod1.S to use internal headers. + + Fixes: 717ebfa85c ("x86-64: Allocate state buffer space for RDI, RSI and RBX") + (cherry picked from commit fd7ee2e6c5eb49e4a630a9978b4d668bff6354ee) + +diff --git a/sysdeps/x86_64/Makefile b/sysdeps/x86_64/Makefile +index e8babc9a4edbf90b..9d374a329916fc45 100644 +--- a/sysdeps/x86_64/Makefile ++++ b/sysdeps/x86_64/Makefile +@@ -210,6 +210,8 @@ tst-plt-rewrite2-ENV = GLIBC_TUNABLES=glibc.cpu.plt_rewrite=2 + $(objpfx)tst-plt-rewrite2: $(objpfx)tst-plt-rewritemod2.so + endif + ++test-internal-extras += tst-gnu2-tls2mod1 ++ + endif # $(subdir) == elf + + ifeq ($(subdir),csu) diff --git a/SOURCES/glibc-upstream-2.39-19.patch b/SOURCES/glibc-upstream-2.39-19.patch new file mode 100644 index 0000000..fa9220e --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-19.patch @@ -0,0 +1,146 @@ +commit aa4249266e9906c4bc833e4847f4d8feef59504f +Author: Adhemerval Zanella +Date: Thu Feb 8 10:08:38 2024 -0300 + + x86: Fix Zen3/Zen4 ERMS selection (BZ 30994) + + The REP MOVSB usage on memcpy/memmove does not show much performance + improvement on Zen3/Zen4 cores compared to the vectorized loops. Also, + as from BZ 30994, if the source is aligned and the destination is not + the performance can be 20x slower. + + The performance difference is noticeable with small buffer sizes, closer + to the lower bounds limits when memcpy/memmove starts to use ERMS. The + performance of REP MOVSB is similar to vectorized instruction on the + size limit (the L2 cache). Also, there is no drawback to multiple cores + sharing the cache. + + Checked on x86_64-linux-gnu on Zen3. + Reviewed-by: H.J. Lu + + (cherry picked from commit 0c0d39fe4aeb0f69b26e76337c5dfd5530d5d44e) + +diff --git a/sysdeps/x86/dl-cacheinfo.h b/sysdeps/x86/dl-cacheinfo.h +index d5101615e348e5c2..f34d12846caf9422 100644 +--- a/sysdeps/x86/dl-cacheinfo.h ++++ b/sysdeps/x86/dl-cacheinfo.h +@@ -791,7 +791,6 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + long int data = -1; + long int shared = -1; + long int shared_per_thread = -1; +- long int core = -1; + unsigned int threads = 0; + unsigned long int level1_icache_size = -1; + unsigned long int level1_icache_linesize = -1; +@@ -809,7 +808,6 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + if (cpu_features->basic.kind == arch_kind_intel) + { + data = handle_intel (_SC_LEVEL1_DCACHE_SIZE, cpu_features); +- core = handle_intel (_SC_LEVEL2_CACHE_SIZE, cpu_features); + shared = handle_intel (_SC_LEVEL3_CACHE_SIZE, cpu_features); + shared_per_thread = shared; + +@@ -822,7 +820,8 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + = handle_intel (_SC_LEVEL1_DCACHE_ASSOC, cpu_features); + level1_dcache_linesize + = handle_intel (_SC_LEVEL1_DCACHE_LINESIZE, cpu_features); +- level2_cache_size = core; ++ level2_cache_size ++ = handle_intel (_SC_LEVEL2_CACHE_SIZE, cpu_features); + level2_cache_assoc + = handle_intel (_SC_LEVEL2_CACHE_ASSOC, cpu_features); + level2_cache_linesize +@@ -835,12 +834,12 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + level4_cache_size + = handle_intel (_SC_LEVEL4_CACHE_SIZE, cpu_features); + +- get_common_cache_info (&shared, &shared_per_thread, &threads, core); ++ get_common_cache_info (&shared, &shared_per_thread, &threads, ++ level2_cache_size); + } + else if (cpu_features->basic.kind == arch_kind_zhaoxin) + { + data = handle_zhaoxin (_SC_LEVEL1_DCACHE_SIZE); +- core = handle_zhaoxin (_SC_LEVEL2_CACHE_SIZE); + shared = handle_zhaoxin (_SC_LEVEL3_CACHE_SIZE); + shared_per_thread = shared; + +@@ -849,19 +848,19 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + level1_dcache_size = data; + level1_dcache_assoc = handle_zhaoxin (_SC_LEVEL1_DCACHE_ASSOC); + level1_dcache_linesize = handle_zhaoxin (_SC_LEVEL1_DCACHE_LINESIZE); +- level2_cache_size = core; ++ level2_cache_size = handle_zhaoxin (_SC_LEVEL2_CACHE_SIZE); + level2_cache_assoc = handle_zhaoxin (_SC_LEVEL2_CACHE_ASSOC); + level2_cache_linesize = handle_zhaoxin (_SC_LEVEL2_CACHE_LINESIZE); + level3_cache_size = shared; + level3_cache_assoc = handle_zhaoxin (_SC_LEVEL3_CACHE_ASSOC); + level3_cache_linesize = handle_zhaoxin (_SC_LEVEL3_CACHE_LINESIZE); + +- get_common_cache_info (&shared, &shared_per_thread, &threads, core); ++ get_common_cache_info (&shared, &shared_per_thread, &threads, ++ level2_cache_size); + } + else if (cpu_features->basic.kind == arch_kind_amd) + { + data = handle_amd (_SC_LEVEL1_DCACHE_SIZE); +- core = handle_amd (_SC_LEVEL2_CACHE_SIZE); + shared = handle_amd (_SC_LEVEL3_CACHE_SIZE); + + level1_icache_size = handle_amd (_SC_LEVEL1_ICACHE_SIZE); +@@ -869,7 +868,7 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + level1_dcache_size = data; + level1_dcache_assoc = handle_amd (_SC_LEVEL1_DCACHE_ASSOC); + level1_dcache_linesize = handle_amd (_SC_LEVEL1_DCACHE_LINESIZE); +- level2_cache_size = core; ++ level2_cache_size = handle_amd (_SC_LEVEL2_CACHE_SIZE);; + level2_cache_assoc = handle_amd (_SC_LEVEL2_CACHE_ASSOC); + level2_cache_linesize = handle_amd (_SC_LEVEL2_CACHE_LINESIZE); + level3_cache_size = shared; +@@ -880,12 +879,12 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + if (shared <= 0) + { + /* No shared L3 cache. All we have is the L2 cache. */ +- shared = core; ++ shared = level2_cache_size; + } + else if (cpu_features->basic.family < 0x17) + { + /* Account for exclusive L2 and L3 caches. */ +- shared += core; ++ shared += level2_cache_size; + } + + shared_per_thread = shared; +@@ -987,6 +986,12 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + if (CPU_FEATURE_USABLE_P (cpu_features, FSRM)) + rep_movsb_threshold = 2112; + ++ /* For AMD CPUs that support ERMS (Zen3+), REP MOVSB is in a lot of ++ cases slower than the vectorized path (and for some alignments, ++ it is really slow, check BZ #30994). */ ++ if (cpu_features->basic.kind == arch_kind_amd) ++ rep_movsb_threshold = non_temporal_threshold; ++ + /* The default threshold to use Enhanced REP STOSB. */ + unsigned long int rep_stosb_threshold = 2048; + +@@ -1028,16 +1033,9 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + SIZE_MAX); + + unsigned long int rep_movsb_stop_threshold; +- /* ERMS feature is implemented from AMD Zen3 architecture and it is +- performing poorly for data above L2 cache size. Henceforth, adding +- an upper bound threshold parameter to limit the usage of Enhanced +- REP MOVSB operations and setting its value to L2 cache size. */ +- if (cpu_features->basic.kind == arch_kind_amd) +- rep_movsb_stop_threshold = core; + /* Setting the upper bound of ERMS to the computed value of +- non-temporal threshold for architectures other than AMD. */ +- else +- rep_movsb_stop_threshold = non_temporal_threshold; ++ non-temporal threshold for all architectures. */ ++ rep_movsb_stop_threshold = non_temporal_threshold; + + cpu_features->data_cache_size = data; + cpu_features->shared_cache_size = shared; diff --git a/SOURCES/glibc-upstream-2.39-2.patch b/SOURCES/glibc-upstream-2.39-2.patch new file mode 100644 index 0000000..77deb04 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-2.patch @@ -0,0 +1,35 @@ +commit 63295e4fda1f6dab4bf7442706fe303bf283036c +Author: Adhemerval Zanella +Date: Mon Feb 5 16:10:24 2024 +0000 + + arm: Remove wrong ldr from _dl_start_user (BZ 31339) + + The commit 49d877a80b29d3002887b084eec6676d9f5fec18 (arm: Remove + _dl_skip_args usage) removed the _SKIP_ARGS literal, which was + previously loader to r4 on loader _start. However, the cleanup did not + remove the following 'ldr r4, [sl, r4]' on _dl_start_user, used to check + to skip the arguments after ld self-relocations. + + In my testing, the kernel initially set r4 to 0, which makes the + ldr instruction just read the _GLOBAL_OFFSET_TABLE_. However, since r4 + is a callee-saved register; a different runtime might not zero + initialize it and thus trigger an invalid memory access. + + Checked on arm-linux-gnu. + + Reported-by: Adrian Ratiu + Reviewed-by: Szabolcs Nagy + (cherry picked from commit 1e25112dc0cb2515d27d8d178b1ecce778a9d37a) + +diff --git a/sysdeps/arm/dl-machine.h b/sysdeps/arm/dl-machine.h +index b857bbc868ea2ce4..dd1a0f6b6e7a5ea8 100644 +--- a/sysdeps/arm/dl-machine.h ++++ b/sysdeps/arm/dl-machine.h +@@ -139,7 +139,6 @@ _start:\n\ + _dl_start_user:\n\ + adr r6, .L_GET_GOT\n\ + add sl, sl, r6\n\ +- ldr r4, [sl, r4]\n\ + @ save the entry point in another register\n\ + mov r6, r0\n\ + @ get the original arg count\n\ diff --git a/SOURCES/glibc-upstream-2.39-20.patch b/SOURCES/glibc-upstream-2.39-20.patch new file mode 100644 index 0000000..3aac665 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-20.patch @@ -0,0 +1,30 @@ +commit 6484a92698039c4a7a510f0214e22d067b0d78b3 +Author: Adhemerval Zanella +Date: Thu Feb 8 10:08:39 2024 -0300 + + x86: Do not prefer ERMS for memset on Zen3+ + + For AMD Zen3+ architecture, the performance of the vectorized loop is + slightly better than ERMS. + + Checked on x86_64-linux-gnu on Zen3. + Reviewed-by: H.J. Lu + + (cherry picked from commit 272708884cb750f12f5c74a00e6620c19dc6d567) + +diff --git a/sysdeps/x86/dl-cacheinfo.h b/sysdeps/x86/dl-cacheinfo.h +index f34d12846caf9422..5a98f70364220da4 100644 +--- a/sysdeps/x86/dl-cacheinfo.h ++++ b/sysdeps/x86/dl-cacheinfo.h +@@ -1021,6 +1021,11 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) + minimum value is fixed. */ + rep_stosb_threshold = TUNABLE_GET (x86_rep_stosb_threshold, + long int, NULL); ++ if (cpu_features->basic.kind == arch_kind_amd ++ && !TUNABLE_IS_INITIALIZED (x86_rep_stosb_threshold)) ++ /* For AMD Zen3+ architecture, the performance of the vectorized loop is ++ slightly better than ERMS. */ ++ rep_stosb_threshold = SIZE_MAX; + + TUNABLE_SET_WITH_BOUNDS (x86_data_cache_size, data, 0, SIZE_MAX); + TUNABLE_SET_WITH_BOUNDS (x86_shared_cache_size, shared, 0, SIZE_MAX); diff --git a/SOURCES/glibc-upstream-2.39-21.patch b/SOURCES/glibc-upstream-2.39-21.patch new file mode 100644 index 0000000..6ceb90c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-21.patch @@ -0,0 +1,24 @@ +commit 5d070d12b3a52bc44dd1b71743abc4b6243862ae +Author: Adhemerval Zanella +Date: Thu Feb 8 10:08:40 2024 -0300 + + x86: Expand the comment on when REP STOSB is used on memset + + Reviewed-by: H.J. Lu + (cherry picked from commit 491e55beab7457ed310a4a47496f4a333c5d1032) + +diff --git a/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S b/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S +index 9984c3ca0fafab6a..97839a22483b0613 100644 +--- a/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S ++++ b/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S +@@ -21,7 +21,9 @@ + 2. If size is less than VEC, use integer register stores. + 3. If size is from VEC_SIZE to 2 * VEC_SIZE, use 2 VEC stores. + 4. If size is from 2 * VEC_SIZE to 4 * VEC_SIZE, use 4 VEC stores. +- 5. If size is more to 4 * VEC_SIZE, align to 4 * VEC_SIZE with ++ 5. On machines ERMS feature, if size is greater or equal than ++ __x86_rep_stosb_threshold then REP STOSB will be used. ++ 6. If size is more to 4 * VEC_SIZE, align to 4 * VEC_SIZE with + 4 VEC stores and store 4 * VEC at a time until done. */ + + #include diff --git a/SOURCES/glibc-upstream-2.39-22.patch b/SOURCES/glibc-upstream-2.39-22.patch new file mode 100644 index 0000000..fb7281f --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-22.patch @@ -0,0 +1,32 @@ +commit 31c7d69af59da0da80caa74b2ec6ae149013384d +Author: Florian Weimer +Date: Fri Feb 16 07:40:37 2024 +0100 + + i386: Use generic memrchr in libc (bug 31316) + + Before this change, we incorrectly used the SSE2 variant in the + implementation, without checking that the system actually supports + SSE2. + + Tested-by: Sam James + (cherry picked from commit 0d9166c2245cad4ac520b337dee40c9a583872b6) + +diff --git a/sysdeps/i386/i686/multiarch/memrchr-c.c b/sysdeps/i386/i686/multiarch/memrchr-c.c +index ef7bbbe792afc78e..20bfdf3af35b66e4 100644 +--- a/sysdeps/i386/i686/multiarch/memrchr-c.c ++++ b/sysdeps/i386/i686/multiarch/memrchr-c.c +@@ -5,3 +5,4 @@ extern void *__memrchr_ia32 (const void *, int, size_t); + #endif + + #include "string/memrchr.c" ++strong_alias (__memrchr_ia32, __GI___memrchr) +diff --git a/sysdeps/i386/i686/multiarch/memrchr-sse2.S b/sysdeps/i386/i686/multiarch/memrchr-sse2.S +index d9dae04171bfedff..e123f87435b3c6a6 100644 +--- a/sysdeps/i386/i686/multiarch/memrchr-sse2.S ++++ b/sysdeps/i386/i686/multiarch/memrchr-sse2.S +@@ -720,5 +720,4 @@ L(ret_null): + ret + + END (__memrchr_sse2) +-strong_alias (__memrchr_sse2, __GI___memrchr) + #endif diff --git a/SOURCES/glibc-upstream-2.39-23.patch b/SOURCES/glibc-upstream-2.39-23.patch new file mode 100644 index 0000000..11773cf --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-23.patch @@ -0,0 +1,669 @@ +commit b0e0a07018098c2c5927796be5681a298c312626 +Author: Joe Ramsay +Date: Tue Feb 20 16:44:13 2024 +0000 + + aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR + + This includes a fix for big-endian in AdvSIMD log, some cosmetic + changes, and numerous small optimisations mainly around inlining and + using indexed variants of MLA intrinsics. + Reviewed-by: Adhemerval Zanella + + (cherry picked from commit e302e1021391d13a9611ba3a910df128830bd19e) + +diff --git a/sysdeps/aarch64/fpu/acos_advsimd.c b/sysdeps/aarch64/fpu/acos_advsimd.c +index a8eabb5e7155360f..0a86c9823a279766 100644 +--- a/sysdeps/aarch64/fpu/acos_advsimd.c ++++ b/sysdeps/aarch64/fpu/acos_advsimd.c +@@ -40,8 +40,8 @@ static const struct data + }; + + #define AllMask v_u64 (0xffffffffffffffff) +-#define Oneu (0x3ff0000000000000) +-#define Small (0x3e50000000000000) /* 2^-53. */ ++#define Oneu 0x3ff0000000000000 ++#define Small 0x3e50000000000000 /* 2^-53. */ + + #if WANT_SIMD_EXCEPT + static float64x2_t VPCS_ATTR NOINLINE +diff --git a/sysdeps/aarch64/fpu/asin_advsimd.c b/sysdeps/aarch64/fpu/asin_advsimd.c +index 141646e954aa8ba1..2de6eff40782bc79 100644 +--- a/sysdeps/aarch64/fpu/asin_advsimd.c ++++ b/sysdeps/aarch64/fpu/asin_advsimd.c +@@ -39,8 +39,8 @@ static const struct data + }; + + #define AllMask v_u64 (0xffffffffffffffff) +-#define One (0x3ff0000000000000) +-#define Small (0x3e50000000000000) /* 2^-12. */ ++#define One 0x3ff0000000000000 ++#define Small 0x3e50000000000000 /* 2^-12. */ + + #if WANT_SIMD_EXCEPT + static float64x2_t VPCS_ATTR NOINLINE +diff --git a/sysdeps/aarch64/fpu/atan2_sve.c b/sysdeps/aarch64/fpu/atan2_sve.c +index 09a4c559b8afbe95..04fa71fa37c3de29 100644 +--- a/sysdeps/aarch64/fpu/atan2_sve.c ++++ b/sysdeps/aarch64/fpu/atan2_sve.c +@@ -37,9 +37,6 @@ static const struct data + .pi_over_2 = 0x1.921fb54442d18p+0, + }; + +-/* Useful constants. */ +-#define SignMask sv_u64 (0x8000000000000000) +- + /* Special cases i.e. 0, infinity, nan (fall back to scalar calls). */ + static svfloat64_t NOINLINE + special_case (svfloat64_t y, svfloat64_t x, svfloat64_t ret, +@@ -72,14 +69,15 @@ svfloat64_t SV_NAME_D2 (atan2) (svfloat64_t y, svfloat64_t x, const svbool_t pg) + svbool_t cmp_y = zeroinfnan (iy, pg); + svbool_t cmp_xy = svorr_z (pg, cmp_x, cmp_y); + +- svuint64_t sign_x = svand_x (pg, ix, SignMask); +- svuint64_t sign_y = svand_x (pg, iy, SignMask); +- svuint64_t sign_xy = sveor_x (pg, sign_x, sign_y); +- + svfloat64_t ax = svabs_x (pg, x); + svfloat64_t ay = svabs_x (pg, y); ++ svuint64_t iax = svreinterpret_u64 (ax); ++ svuint64_t iay = svreinterpret_u64 (ay); ++ ++ svuint64_t sign_x = sveor_x (pg, ix, iax); ++ svuint64_t sign_y = sveor_x (pg, iy, iay); ++ svuint64_t sign_xy = sveor_x (pg, sign_x, sign_y); + +- svbool_t pred_xlt0 = svcmplt (pg, x, 0.0); + svbool_t pred_aygtax = svcmpgt (pg, ay, ax); + + /* Set up z for call to atan. */ +@@ -88,8 +86,9 @@ svfloat64_t SV_NAME_D2 (atan2) (svfloat64_t y, svfloat64_t x, const svbool_t pg) + svfloat64_t z = svdiv_x (pg, n, d); + + /* Work out the correct shift. */ +- svfloat64_t shift = svsel (pred_xlt0, sv_f64 (-2.0), sv_f64 (0.0)); +- shift = svsel (pred_aygtax, svadd_x (pg, shift, 1.0), shift); ++ svfloat64_t shift = svreinterpret_f64 (svlsr_x (pg, sign_x, 1)); ++ shift = svsel (pred_aygtax, sv_f64 (1.0), shift); ++ shift = svreinterpret_f64 (svorr_x (pg, sign_x, svreinterpret_u64 (shift))); + shift = svmul_x (pg, shift, data_ptr->pi_over_2); + + /* Use split Estrin scheme for P(z^2) with deg(P)=19. */ +@@ -109,10 +108,10 @@ svfloat64_t SV_NAME_D2 (atan2) (svfloat64_t y, svfloat64_t x, const svbool_t pg) + ret = svadd_m (pg, ret, shift); + + /* Account for the sign of x and y. */ +- ret = svreinterpret_f64 (sveor_x (pg, svreinterpret_u64 (ret), sign_xy)); +- + if (__glibc_unlikely (svptest_any (pg, cmp_xy))) +- return special_case (y, x, ret, cmp_xy); +- +- return ret; ++ return special_case ( ++ y, x, ++ svreinterpret_f64 (sveor_x (pg, svreinterpret_u64 (ret), sign_xy)), ++ cmp_xy); ++ return svreinterpret_f64 (sveor_x (pg, svreinterpret_u64 (ret), sign_xy)); + } +diff --git a/sysdeps/aarch64/fpu/atan2f_sve.c b/sysdeps/aarch64/fpu/atan2f_sve.c +index b92f83cdea4fb2da..9ea197147ccca0ac 100644 +--- a/sysdeps/aarch64/fpu/atan2f_sve.c ++++ b/sysdeps/aarch64/fpu/atan2f_sve.c +@@ -32,10 +32,8 @@ static const struct data + .pi_over_2 = 0x1.921fb6p+0f, + }; + +-#define SignMask sv_u32 (0x80000000) +- + /* Special cases i.e. 0, infinity, nan (fall back to scalar calls). */ +-static inline svfloat32_t ++static svfloat32_t NOINLINE + special_case (svfloat32_t y, svfloat32_t x, svfloat32_t ret, + const svbool_t cmp) + { +@@ -67,14 +65,15 @@ svfloat32_t SV_NAME_F2 (atan2) (svfloat32_t y, svfloat32_t x, const svbool_t pg) + svbool_t cmp_y = zeroinfnan (iy, pg); + svbool_t cmp_xy = svorr_z (pg, cmp_x, cmp_y); + +- svuint32_t sign_x = svand_x (pg, ix, SignMask); +- svuint32_t sign_y = svand_x (pg, iy, SignMask); +- svuint32_t sign_xy = sveor_x (pg, sign_x, sign_y); +- + svfloat32_t ax = svabs_x (pg, x); + svfloat32_t ay = svabs_x (pg, y); ++ svuint32_t iax = svreinterpret_u32 (ax); ++ svuint32_t iay = svreinterpret_u32 (ay); ++ ++ svuint32_t sign_x = sveor_x (pg, ix, iax); ++ svuint32_t sign_y = sveor_x (pg, iy, iay); ++ svuint32_t sign_xy = sveor_x (pg, sign_x, sign_y); + +- svbool_t pred_xlt0 = svcmplt (pg, x, 0.0); + svbool_t pred_aygtax = svcmpgt (pg, ay, ax); + + /* Set up z for call to atan. */ +@@ -83,11 +82,12 @@ svfloat32_t SV_NAME_F2 (atan2) (svfloat32_t y, svfloat32_t x, const svbool_t pg) + svfloat32_t z = svdiv_x (pg, n, d); + + /* Work out the correct shift. */ +- svfloat32_t shift = svsel (pred_xlt0, sv_f32 (-2.0), sv_f32 (0.0)); +- shift = svsel (pred_aygtax, svadd_x (pg, shift, 1.0), shift); ++ svfloat32_t shift = svreinterpret_f32 (svlsr_x (pg, sign_x, 1)); ++ shift = svsel (pred_aygtax, sv_f32 (1.0), shift); ++ shift = svreinterpret_f32 (svorr_x (pg, sign_x, svreinterpret_u32 (shift))); + shift = svmul_x (pg, shift, sv_f32 (data_ptr->pi_over_2)); + +- /* Use split Estrin scheme for P(z^2) with deg(P)=7. */ ++ /* Use pure Estrin scheme for P(z^2) with deg(P)=7. */ + svfloat32_t z2 = svmul_x (pg, z, z); + svfloat32_t z4 = svmul_x (pg, z2, z2); + svfloat32_t z8 = svmul_x (pg, z4, z4); +@@ -101,10 +101,12 @@ svfloat32_t SV_NAME_F2 (atan2) (svfloat32_t y, svfloat32_t x, const svbool_t pg) + ret = svadd_m (pg, ret, shift); + + /* Account for the sign of x and y. */ +- ret = svreinterpret_f32 (sveor_x (pg, svreinterpret_u32 (ret), sign_xy)); + + if (__glibc_unlikely (svptest_any (pg, cmp_xy))) +- return special_case (y, x, ret, cmp_xy); ++ return special_case ( ++ y, x, ++ svreinterpret_f32 (sveor_x (pg, svreinterpret_u32 (ret), sign_xy)), ++ cmp_xy); + +- return ret; ++ return svreinterpret_f32 (sveor_x (pg, svreinterpret_u32 (ret), sign_xy)); + } +diff --git a/sysdeps/aarch64/fpu/cos_advsimd.c b/sysdeps/aarch64/fpu/cos_advsimd.c +index 2897e8b909bbcb66..3924c9ce44c30d4d 100644 +--- a/sysdeps/aarch64/fpu/cos_advsimd.c ++++ b/sysdeps/aarch64/fpu/cos_advsimd.c +@@ -63,8 +63,7 @@ float64x2_t VPCS_ATTR V_NAME_D1 (cos) (float64x2_t x) + special-case handler later. */ + r = vbslq_f64 (cmp, v_f64 (1.0), r); + #else +- cmp = vcageq_f64 (d->range_val, x); +- cmp = vceqzq_u64 (cmp); /* cmp = ~cmp. */ ++ cmp = vcageq_f64 (x, d->range_val); + r = x; + #endif + +diff --git a/sysdeps/aarch64/fpu/cosf_advsimd.c b/sysdeps/aarch64/fpu/cosf_advsimd.c +index 60abc8dfcfff9ed4..d0c285b03a8bfe22 100644 +--- a/sysdeps/aarch64/fpu/cosf_advsimd.c ++++ b/sysdeps/aarch64/fpu/cosf_advsimd.c +@@ -64,8 +64,7 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (cos) (float32x4_t x) + special-case handler later. */ + r = vbslq_f32 (cmp, v_f32 (1.0f), r); + #else +- cmp = vcageq_f32 (d->range_val, x); +- cmp = vceqzq_u32 (cmp); /* cmp = ~cmp. */ ++ cmp = vcageq_f32 (x, d->range_val); + r = x; + #endif + +diff --git a/sysdeps/aarch64/fpu/exp10_advsimd.c b/sysdeps/aarch64/fpu/exp10_advsimd.c +index fe7149b19176da21..eeb31ca839c1f671 100644 +--- a/sysdeps/aarch64/fpu/exp10_advsimd.c ++++ b/sysdeps/aarch64/fpu/exp10_advsimd.c +@@ -57,7 +57,7 @@ const static struct data + # define BigBound v_u64 (0x4070000000000000) /* asuint64 (0x1p8). */ + # define Thres v_u64 (0x2070000000000000) /* BigBound - TinyBound. */ + +-static inline float64x2_t VPCS_ATTR ++static float64x2_t VPCS_ATTR NOINLINE + special_case (float64x2_t x, float64x2_t y, uint64x2_t cmp) + { + /* If fenv exceptions are to be triggered correctly, fall back to the scalar +@@ -72,7 +72,7 @@ special_case (float64x2_t x, float64x2_t y, uint64x2_t cmp) + # define SpecialBias1 v_u64 (0x7000000000000000) /* 0x1p769. */ + # define SpecialBias2 v_u64 (0x3010000000000000) /* 0x1p-254. */ + +-static float64x2_t VPCS_ATTR NOINLINE ++static inline float64x2_t VPCS_ATTR + special_case (float64x2_t s, float64x2_t y, float64x2_t n, + const struct data *d) + { +diff --git a/sysdeps/aarch64/fpu/exp10f_advsimd.c b/sysdeps/aarch64/fpu/exp10f_advsimd.c +index 7ee0c909487c32b7..ab117b69da23e5f3 100644 +--- a/sysdeps/aarch64/fpu/exp10f_advsimd.c ++++ b/sysdeps/aarch64/fpu/exp10f_advsimd.c +@@ -25,7 +25,8 @@ + static const struct data + { + float32x4_t poly[5]; +- float32x4_t shift, log10_2, log2_10_hi, log2_10_lo; ++ float32x4_t log10_2_and_inv, shift; ++ + #if !WANT_SIMD_EXCEPT + float32x4_t scale_thresh; + #endif +@@ -38,9 +39,9 @@ static const struct data + .poly = { V4 (0x1.26bb16p+1f), V4 (0x1.5350d2p+1f), V4 (0x1.04744ap+1f), + V4 (0x1.2d8176p+0f), V4 (0x1.12b41ap-1f) }, + .shift = V4 (0x1.8p23f), +- .log10_2 = V4 (0x1.a934fp+1), +- .log2_10_hi = V4 (0x1.344136p-2), +- .log2_10_lo = V4 (-0x1.ec10cp-27), ++ ++ /* Stores constants 1/log10(2), log10(2)_high, log10(2)_low, 0. */ ++ .log10_2_and_inv = { 0x1.a934fp+1, 0x1.344136p-2, -0x1.ec10cp-27, 0 }, + #if !WANT_SIMD_EXCEPT + .scale_thresh = V4 (ScaleBound) + #endif +@@ -98,24 +99,22 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (exp10) (float32x4_t x) + #if WANT_SIMD_EXCEPT + /* asuint(x) - TinyBound >= BigBound - TinyBound. */ + uint32x4_t cmp = vcgeq_u32 ( +- vsubq_u32 (vandq_u32 (vreinterpretq_u32_f32 (x), v_u32 (0x7fffffff)), +- TinyBound), +- Thres); ++ vsubq_u32 (vreinterpretq_u32_f32 (vabsq_f32 (x)), TinyBound), Thres); + float32x4_t xm = x; + /* If any lanes are special, mask them with 1 and retain a copy of x to allow + special case handler to fix special lanes later. This is only necessary if + fenv exceptions are to be triggered correctly. */ + if (__glibc_unlikely (v_any_u32 (cmp))) +- x = vbslq_f32 (cmp, v_f32 (1), x); ++ x = v_zerofy_f32 (x, cmp); + #endif + + /* exp10(x) = 2^n * 10^r = 2^n * (1 + poly (r)), + with poly(r) in [1/sqrt(2), sqrt(2)] and + x = r + n * log10 (2), with r in [-log10(2)/2, log10(2)/2]. */ +- float32x4_t z = vfmaq_f32 (d->shift, x, d->log10_2); ++ float32x4_t z = vfmaq_laneq_f32 (d->shift, x, d->log10_2_and_inv, 0); + float32x4_t n = vsubq_f32 (z, d->shift); +- float32x4_t r = vfmsq_f32 (x, n, d->log2_10_hi); +- r = vfmsq_f32 (r, n, d->log2_10_lo); ++ float32x4_t r = vfmsq_laneq_f32 (x, n, d->log10_2_and_inv, 1); ++ r = vfmsq_laneq_f32 (r, n, d->log10_2_and_inv, 2); + uint32x4_t e = vshlq_n_u32 (vreinterpretq_u32_f32 (z), 23); + + float32x4_t scale = vreinterpretq_f32_u32 (vaddq_u32 (e, ExponentBias)); +diff --git a/sysdeps/aarch64/fpu/exp2_advsimd.c b/sysdeps/aarch64/fpu/exp2_advsimd.c +index 391a93180c93427d..ae1e63d5030633d5 100644 +--- a/sysdeps/aarch64/fpu/exp2_advsimd.c ++++ b/sysdeps/aarch64/fpu/exp2_advsimd.c +@@ -24,6 +24,7 @@ + #define IndexMask (N - 1) + #define BigBound 1022.0 + #define UOFlowBound 1280.0 ++#define TinyBound 0x2000000000000000 /* asuint64(0x1p-511). */ + + static const struct data + { +@@ -48,14 +49,13 @@ lookup_sbits (uint64x2_t i) + + #if WANT_SIMD_EXCEPT + +-# define TinyBound 0x2000000000000000 /* asuint64(0x1p-511). */ + # define Thres 0x2080000000000000 /* asuint64(512.0) - TinyBound. */ + + /* Call scalar exp2 as a fallback. */ + static float64x2_t VPCS_ATTR NOINLINE +-special_case (float64x2_t x) ++special_case (float64x2_t x, float64x2_t y, uint64x2_t is_special) + { +- return v_call_f64 (exp2, x, x, v_u64 (0xffffffffffffffff)); ++ return v_call_f64 (exp2, x, y, is_special); + } + + #else +@@ -65,7 +65,7 @@ special_case (float64x2_t x) + # define SpecialBias1 0x7000000000000000 /* 0x1p769. */ + # define SpecialBias2 0x3010000000000000 /* 0x1p-254. */ + +-static float64x2_t VPCS_ATTR ++static inline float64x2_t VPCS_ATTR + special_case (float64x2_t s, float64x2_t y, float64x2_t n, + const struct data *d) + { +@@ -94,10 +94,10 @@ float64x2_t V_NAME_D1 (exp2) (float64x2_t x) + #if WANT_SIMD_EXCEPT + uint64x2_t ia = vreinterpretq_u64_f64 (vabsq_f64 (x)); + cmp = vcgeq_u64 (vsubq_u64 (ia, v_u64 (TinyBound)), v_u64 (Thres)); +- /* If any special case (inf, nan, small and large x) is detected, +- fall back to scalar for all lanes. */ +- if (__glibc_unlikely (v_any_u64 (cmp))) +- return special_case (x); ++ /* Mask special lanes and retain a copy of x for passing to special-case ++ handler. */ ++ float64x2_t xc = x; ++ x = v_zerofy_f64 (x, cmp); + #else + cmp = vcagtq_f64 (x, d->scale_big_bound); + #endif +@@ -120,9 +120,11 @@ float64x2_t V_NAME_D1 (exp2) (float64x2_t x) + float64x2_t y = v_pairwise_poly_3_f64 (r, r2, d->poly); + y = vmulq_f64 (r, y); + +-#if !WANT_SIMD_EXCEPT + if (__glibc_unlikely (v_any_u64 (cmp))) ++#if !WANT_SIMD_EXCEPT + return special_case (s, y, n, d); ++#else ++ return special_case (xc, vfmaq_f64 (s, s, y), cmp); + #endif + return vfmaq_f64 (s, s, y); + } +diff --git a/sysdeps/aarch64/fpu/exp2f_sve.c b/sysdeps/aarch64/fpu/exp2f_sve.c +index 9a5a523a10280d1d..8a686e3e054cb7f5 100644 +--- a/sysdeps/aarch64/fpu/exp2f_sve.c ++++ b/sysdeps/aarch64/fpu/exp2f_sve.c +@@ -20,6 +20,8 @@ + #include "sv_math.h" + #include "poly_sve_f32.h" + ++#define Thres 0x1.5d5e2ap+6f ++ + static const struct data + { + float poly[5]; +@@ -33,7 +35,7 @@ static const struct data + .shift = 0x1.903f8p17f, + /* Roughly 87.3. For x < -Thres, the result is subnormal and not handled + correctly by FEXPA. */ +- .thres = 0x1.5d5e2ap+6f, ++ .thres = Thres, + }; + + static svfloat32_t NOINLINE +diff --git a/sysdeps/aarch64/fpu/exp_advsimd.c b/sysdeps/aarch64/fpu/exp_advsimd.c +index fd215f1d2c5e9eea..5e3a9a0d44a43656 100644 +--- a/sysdeps/aarch64/fpu/exp_advsimd.c ++++ b/sysdeps/aarch64/fpu/exp_advsimd.c +@@ -54,7 +54,7 @@ const static volatile struct + # define BigBound v_u64 (0x4080000000000000) /* asuint64 (0x1p9). */ + # define SpecialBound v_u64 (0x2080000000000000) /* BigBound - TinyBound. */ + +-static inline float64x2_t VPCS_ATTR ++static float64x2_t VPCS_ATTR NOINLINE + special_case (float64x2_t x, float64x2_t y, uint64x2_t cmp) + { + /* If fenv exceptions are to be triggered correctly, fall back to the scalar +@@ -69,7 +69,7 @@ special_case (float64x2_t x, float64x2_t y, uint64x2_t cmp) + # define SpecialBias1 v_u64 (0x7000000000000000) /* 0x1p769. */ + # define SpecialBias2 v_u64 (0x3010000000000000) /* 0x1p-254. */ + +-static float64x2_t VPCS_ATTR NOINLINE ++static inline float64x2_t VPCS_ATTR + special_case (float64x2_t s, float64x2_t y, float64x2_t n) + { + /* 2^(n/N) may overflow, break it up into s1*s2. */ +diff --git a/sysdeps/aarch64/fpu/expm1_advsimd.c b/sysdeps/aarch64/fpu/expm1_advsimd.c +index 0b85bd06f3c10068..3628398674468131 100644 +--- a/sysdeps/aarch64/fpu/expm1_advsimd.c ++++ b/sysdeps/aarch64/fpu/expm1_advsimd.c +@@ -23,7 +23,7 @@ + static const struct data + { + float64x2_t poly[11]; +- float64x2_t invln2, ln2_lo, ln2_hi, shift; ++ float64x2_t invln2, ln2, shift; + int64x2_t exponent_bias; + #if WANT_SIMD_EXCEPT + uint64x2_t thresh, tiny_bound; +@@ -38,8 +38,7 @@ static const struct data + V2 (0x1.71ddf82db5bb4p-19), V2 (0x1.27e517fc0d54bp-22), + V2 (0x1.af5eedae67435p-26), V2 (0x1.1f143d060a28ap-29) }, + .invln2 = V2 (0x1.71547652b82fep0), +- .ln2_hi = V2 (0x1.62e42fefa39efp-1), +- .ln2_lo = V2 (0x1.abc9e3b39803fp-56), ++ .ln2 = { 0x1.62e42fefa39efp-1, 0x1.abc9e3b39803fp-56 }, + .shift = V2 (0x1.8p52), + .exponent_bias = V2 (0x3ff0000000000000), + #if WANT_SIMD_EXCEPT +@@ -83,7 +82,7 @@ float64x2_t VPCS_ATTR V_NAME_D1 (expm1) (float64x2_t x) + x = v_zerofy_f64 (x, special); + #else + /* Large input, NaNs and Infs. */ +- uint64x2_t special = vceqzq_u64 (vcaltq_f64 (x, d->oflow_bound)); ++ uint64x2_t special = vcageq_f64 (x, d->oflow_bound); + #endif + + /* Reduce argument to smaller range: +@@ -93,8 +92,8 @@ float64x2_t VPCS_ATTR V_NAME_D1 (expm1) (float64x2_t x) + where 2^i is exact because i is an integer. */ + float64x2_t n = vsubq_f64 (vfmaq_f64 (d->shift, d->invln2, x), d->shift); + int64x2_t i = vcvtq_s64_f64 (n); +- float64x2_t f = vfmsq_f64 (x, n, d->ln2_hi); +- f = vfmsq_f64 (f, n, d->ln2_lo); ++ float64x2_t f = vfmsq_laneq_f64 (x, n, d->ln2, 0); ++ f = vfmsq_laneq_f64 (f, n, d->ln2, 1); + + /* Approximate expm1(f) using polynomial. + Taylor expansion for expm1(x) has the form: +diff --git a/sysdeps/aarch64/fpu/expm1f_advsimd.c b/sysdeps/aarch64/fpu/expm1f_advsimd.c +index 8d4c9a21936d31dd..93db200f618379be 100644 +--- a/sysdeps/aarch64/fpu/expm1f_advsimd.c ++++ b/sysdeps/aarch64/fpu/expm1f_advsimd.c +@@ -23,7 +23,8 @@ + static const struct data + { + float32x4_t poly[5]; +- float32x4_t invln2, ln2_lo, ln2_hi, shift; ++ float32x4_t invln2_and_ln2; ++ float32x4_t shift; + int32x4_t exponent_bias; + #if WANT_SIMD_EXCEPT + uint32x4_t thresh; +@@ -34,9 +35,8 @@ static const struct data + /* Generated using fpminimax with degree=5 in [-log(2)/2, log(2)/2]. */ + .poly = { V4 (0x1.fffffep-2), V4 (0x1.5554aep-3), V4 (0x1.555736p-5), + V4 (0x1.12287cp-7), V4 (0x1.6b55a2p-10) }, +- .invln2 = V4 (0x1.715476p+0f), +- .ln2_hi = V4 (0x1.62e4p-1f), +- .ln2_lo = V4 (0x1.7f7d1cp-20f), ++ /* Stores constants: invln2, ln2_hi, ln2_lo, 0. */ ++ .invln2_and_ln2 = { 0x1.715476p+0f, 0x1.62e4p-1f, 0x1.7f7d1cp-20f, 0 }, + .shift = V4 (0x1.8p23f), + .exponent_bias = V4 (0x3f800000), + #if !WANT_SIMD_EXCEPT +@@ -80,7 +80,7 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (expm1) (float32x4_t x) + x = v_zerofy_f32 (x, special); + #else + /* Handles very large values (+ve and -ve), +/-NaN, +/-Inf. */ +- uint32x4_t special = vceqzq_u32 (vcaltq_f32 (x, d->oflow_bound)); ++ uint32x4_t special = vcagtq_f32 (x, d->oflow_bound); + #endif + + /* Reduce argument to smaller range: +@@ -88,10 +88,11 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (expm1) (float32x4_t x) + and f = x - i * ln2, then f is in [-ln2/2, ln2/2]. + exp(x) - 1 = 2^i * (expm1(f) + 1) - 1 + where 2^i is exact because i is an integer. */ +- float32x4_t j = vsubq_f32 (vfmaq_f32 (d->shift, d->invln2, x), d->shift); ++ float32x4_t j = vsubq_f32 ( ++ vfmaq_laneq_f32 (d->shift, x, d->invln2_and_ln2, 0), d->shift); + int32x4_t i = vcvtq_s32_f32 (j); +- float32x4_t f = vfmsq_f32 (x, j, d->ln2_hi); +- f = vfmsq_f32 (f, j, d->ln2_lo); ++ float32x4_t f = vfmsq_laneq_f32 (x, j, d->invln2_and_ln2, 1); ++ f = vfmsq_laneq_f32 (f, j, d->invln2_and_ln2, 2); + + /* Approximate expm1(f) using polynomial. + Taylor expansion for expm1(x) has the form: +diff --git a/sysdeps/aarch64/fpu/log_advsimd.c b/sysdeps/aarch64/fpu/log_advsimd.c +index 067ae7961300c66b..21df61728ca87374 100644 +--- a/sysdeps/aarch64/fpu/log_advsimd.c ++++ b/sysdeps/aarch64/fpu/log_advsimd.c +@@ -58,8 +58,13 @@ lookup (uint64x2_t i) + uint64_t i1 = (i[1] >> (52 - V_LOG_TABLE_BITS)) & IndexMask; + float64x2_t e0 = vld1q_f64 (&__v_log_data.table[i0].invc); + float64x2_t e1 = vld1q_f64 (&__v_log_data.table[i1].invc); ++#if __BYTE_ORDER == __LITTLE_ENDIAN + e.invc = vuzp1q_f64 (e0, e1); + e.logc = vuzp2q_f64 (e0, e1); ++#else ++ e.invc = vuzp1q_f64 (e1, e0); ++ e.logc = vuzp2q_f64 (e1, e0); ++#endif + return e; + } + +diff --git a/sysdeps/aarch64/fpu/sin_advsimd.c b/sysdeps/aarch64/fpu/sin_advsimd.c +index efce183e86e319f1..a0d9d3b81965db76 100644 +--- a/sysdeps/aarch64/fpu/sin_advsimd.c ++++ b/sysdeps/aarch64/fpu/sin_advsimd.c +@@ -75,8 +75,7 @@ float64x2_t VPCS_ATTR V_NAME_D1 (sin) (float64x2_t x) + r = vbslq_f64 (cmp, vreinterpretq_f64_u64 (cmp), x); + #else + r = x; +- cmp = vcageq_f64 (d->range_val, x); +- cmp = vceqzq_u64 (cmp); /* cmp = ~cmp. */ ++ cmp = vcageq_f64 (x, d->range_val); + #endif + + /* n = rint(|x|/pi). */ +diff --git a/sysdeps/aarch64/fpu/sinf_advsimd.c b/sysdeps/aarch64/fpu/sinf_advsimd.c +index 60cf3f2ca102619c..375dfc3331fa6a9c 100644 +--- a/sysdeps/aarch64/fpu/sinf_advsimd.c ++++ b/sysdeps/aarch64/fpu/sinf_advsimd.c +@@ -67,8 +67,7 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (sin) (float32x4_t x) + r = vbslq_f32 (cmp, vreinterpretq_f32_u32 (cmp), x); + #else + r = x; +- cmp = vcageq_f32 (d->range_val, x); +- cmp = vceqzq_u32 (cmp); /* cmp = ~cmp. */ ++ cmp = vcageq_f32 (x, d->range_val); + #endif + + /* n = rint(|x|/pi) */ +diff --git a/sysdeps/aarch64/fpu/tan_advsimd.c b/sysdeps/aarch64/fpu/tan_advsimd.c +index d7e5ba7b1ab941e8..0459821ab25487a8 100644 +--- a/sysdeps/aarch64/fpu/tan_advsimd.c ++++ b/sysdeps/aarch64/fpu/tan_advsimd.c +@@ -23,7 +23,7 @@ + static const struct data + { + float64x2_t poly[9]; +- float64x2_t half_pi_hi, half_pi_lo, two_over_pi, shift; ++ float64x2_t half_pi, two_over_pi, shift; + #if !WANT_SIMD_EXCEPT + float64x2_t range_val; + #endif +@@ -34,8 +34,7 @@ static const struct data + V2 (0x1.226e5e5ecdfa3p-7), V2 (0x1.d6c7ddbf87047p-9), + V2 (0x1.7ea75d05b583ep-10), V2 (0x1.289f22964a03cp-11), + V2 (0x1.4e4fd14147622p-12) }, +- .half_pi_hi = V2 (0x1.921fb54442d18p0), +- .half_pi_lo = V2 (0x1.1a62633145c07p-54), ++ .half_pi = { 0x1.921fb54442d18p0, 0x1.1a62633145c07p-54 }, + .two_over_pi = V2 (0x1.45f306dc9c883p-1), + .shift = V2 (0x1.8p52), + #if !WANT_SIMD_EXCEPT +@@ -56,15 +55,15 @@ special_case (float64x2_t x) + + /* Vector approximation for double-precision tan. + Maximum measured error is 3.48 ULP: +- __v_tan(0x1.4457047ef78d8p+20) got -0x1.f6ccd8ecf7dedp+37 +- want -0x1.f6ccd8ecf7deap+37. */ ++ _ZGVnN2v_tan(0x1.4457047ef78d8p+20) got -0x1.f6ccd8ecf7dedp+37 ++ want -0x1.f6ccd8ecf7deap+37. */ + float64x2_t VPCS_ATTR V_NAME_D1 (tan) (float64x2_t x) + { + const struct data *dat = ptr_barrier (&data); +- /* Our argument reduction cannot calculate q with sufficient accuracy for very +- large inputs. Fall back to scalar routine for all lanes if any are too +- large, or Inf/NaN. If fenv exceptions are expected, also fall back for tiny +- input to avoid underflow. */ ++ /* Our argument reduction cannot calculate q with sufficient accuracy for ++ very large inputs. Fall back to scalar routine for all lanes if any are ++ too large, or Inf/NaN. If fenv exceptions are expected, also fall back for ++ tiny input to avoid underflow. */ + #if WANT_SIMD_EXCEPT + uint64x2_t iax = vreinterpretq_u64_f64 (vabsq_f64 (x)); + /* iax - tiny_bound > range_val - tiny_bound. */ +@@ -82,8 +81,8 @@ float64x2_t VPCS_ATTR V_NAME_D1 (tan) (float64x2_t x) + /* Use q to reduce x to r in [-pi/4, pi/4], by: + r = x - q * pi/2, in extended precision. */ + float64x2_t r = x; +- r = vfmsq_f64 (r, q, dat->half_pi_hi); +- r = vfmsq_f64 (r, q, dat->half_pi_lo); ++ r = vfmsq_laneq_f64 (r, q, dat->half_pi, 0); ++ r = vfmsq_laneq_f64 (r, q, dat->half_pi, 1); + /* Further reduce r to [-pi/8, pi/8], to be reconstructed using double angle + formula. */ + r = vmulq_n_f64 (r, 0.5); +@@ -106,14 +105,15 @@ float64x2_t VPCS_ATTR V_NAME_D1 (tan) (float64x2_t x) + and reciprocity around pi/2: + tan(x) = 1 / (tan(pi/2 - x)) + to assemble result using change-of-sign and conditional selection of +- numerator/denominator, dependent on odd/even-ness of q (hence quadrant). */ ++ numerator/denominator, dependent on odd/even-ness of q (hence quadrant). ++ */ + float64x2_t n = vfmaq_f64 (v_f64 (-1), p, p); + float64x2_t d = vaddq_f64 (p, p); + + uint64x2_t no_recip = vtstq_u64 (vreinterpretq_u64_s64 (qi), v_u64 (1)); + + #if !WANT_SIMD_EXCEPT +- uint64x2_t special = vceqzq_u64 (vcaleq_f64 (x, dat->range_val)); ++ uint64x2_t special = vcageq_f64 (x, dat->range_val); + if (__glibc_unlikely (v_any_u64 (special))) + return special_case (x); + #endif +diff --git a/sysdeps/aarch64/fpu/tanf_advsimd.c b/sysdeps/aarch64/fpu/tanf_advsimd.c +index 1f16103f8a7668f8..5a7489390a9692c6 100644 +--- a/sysdeps/aarch64/fpu/tanf_advsimd.c ++++ b/sysdeps/aarch64/fpu/tanf_advsimd.c +@@ -23,7 +23,8 @@ + static const struct data + { + float32x4_t poly[6]; +- float32x4_t neg_half_pi_1, neg_half_pi_2, neg_half_pi_3, two_over_pi, shift; ++ float32x4_t pi_consts; ++ float32x4_t shift; + #if !WANT_SIMD_EXCEPT + float32x4_t range_val; + #endif +@@ -31,10 +32,9 @@ static const struct data + /* Coefficients generated using FPMinimax. */ + .poly = { V4 (0x1.55555p-2f), V4 (0x1.11166p-3f), V4 (0x1.b88a78p-5f), + V4 (0x1.7b5756p-6f), V4 (0x1.4ef4cep-8f), V4 (0x1.0e1e74p-7f) }, +- .neg_half_pi_1 = V4 (-0x1.921fb6p+0f), +- .neg_half_pi_2 = V4 (0x1.777a5cp-25f), +- .neg_half_pi_3 = V4 (0x1.ee59dap-50f), +- .two_over_pi = V4 (0x1.45f306p-1f), ++ /* Stores constants: (-pi/2)_high, (-pi/2)_mid, (-pi/2)_low, and 2/pi. */ ++ .pi_consts ++ = { -0x1.921fb6p+0f, 0x1.777a5cp-25f, 0x1.ee59dap-50f, 0x1.45f306p-1f }, + .shift = V4 (0x1.8p+23f), + #if !WANT_SIMD_EXCEPT + .range_val = V4 (0x1p15f), +@@ -58,10 +58,11 @@ eval_poly (float32x4_t z, const struct data *d) + { + float32x4_t z2 = vmulq_f32 (z, z); + #if WANT_SIMD_EXCEPT +- /* Tiny z (<= 0x1p-31) will underflow when calculating z^4. If fp exceptions +- are to be triggered correctly, sidestep this by fixing such lanes to 0. */ ++ /* Tiny z (<= 0x1p-31) will underflow when calculating z^4. ++ If fp exceptions are to be triggered correctly, ++ sidestep this by fixing such lanes to 0. */ + uint32x4_t will_uflow +- = vcleq_u32 (vreinterpretq_u32_f32 (vabsq_f32 (z)), TinyBound); ++ = vcleq_u32 (vreinterpretq_u32_f32 (vabsq_f32 (z)), TinyBound); + if (__glibc_unlikely (v_any_u32 (will_uflow))) + z2 = vbslq_f32 (will_uflow, v_f32 (0), z2); + #endif +@@ -94,16 +95,16 @@ float32x4_t VPCS_ATTR NOINLINE V_NAME_F1 (tan) (float32x4_t x) + #endif + + /* n = rint(x/(pi/2)). */ +- float32x4_t q = vfmaq_f32 (d->shift, d->two_over_pi, x); ++ float32x4_t q = vfmaq_laneq_f32 (d->shift, x, d->pi_consts, 3); + float32x4_t n = vsubq_f32 (q, d->shift); + /* Determine if x lives in an interval, where |tan(x)| grows to infinity. */ + uint32x4_t pred_alt = vtstq_u32 (vreinterpretq_u32_f32 (q), v_u32 (1)); + + /* r = x - n * (pi/2) (range reduction into -pi./4 .. pi/4). */ + float32x4_t r; +- r = vfmaq_f32 (x, d->neg_half_pi_1, n); +- r = vfmaq_f32 (r, d->neg_half_pi_2, n); +- r = vfmaq_f32 (r, d->neg_half_pi_3, n); ++ r = vfmaq_laneq_f32 (x, n, d->pi_consts, 0); ++ r = vfmaq_laneq_f32 (r, n, d->pi_consts, 1); ++ r = vfmaq_laneq_f32 (r, n, d->pi_consts, 2); + + /* If x lives in an interval, where |tan(x)| + - is finite, then use a polynomial approximation of the form diff --git a/SOURCES/glibc-upstream-2.39-24.patch b/SOURCES/glibc-upstream-2.39-24.patch new file mode 100644 index 0000000..61d3f48 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-24.patch @@ -0,0 +1,54 @@ +commit 395a89f61e19fa916ae4cc93fc10d81a28ce3039 +Author: Szabolcs Nagy +Date: Wed Mar 13 14:34:14 2024 +0000 + + aarch64: fix check for SVE support in assembler + + Due to GCC bug 110901 -mcpu can override -march setting when compiling + asm code and thus a compiler targetting a specific cpu can fail the + configure check even when binutils gas supports SVE. + + The workaround is that explicit .arch directive overrides both -mcpu + and -march, and since that's what the actual SVE memcpy uses the + configure check should use that too even if the GCC issue is fixed + independently. + + Reviewed-by: Florian Weimer + (cherry picked from commit 73c26018ed0ecd9c807bb363cc2c2ab4aca66a82) + +diff --git a/sysdeps/aarch64/configure b/sysdeps/aarch64/configure +old mode 100644 +new mode 100755 +index ca57edce472e4507..9606137e8ddae545 +--- a/sysdeps/aarch64/configure ++++ b/sysdeps/aarch64/configure +@@ -325,9 +325,10 @@ then : + printf %s "(cached) " >&6 + else $as_nop + cat > conftest.s <<\EOF +- ptrue p0.b ++ .arch armv8.2-a+sve ++ ptrue p0.b + EOF +-if { ac_try='${CC-cc} -c -march=armv8.2-a+sve conftest.s 1>&5' ++if { ac_try='${CC-cc} -c conftest.s 1>&5' + { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5 + (eval $ac_try) 2>&5 + ac_status=$? +diff --git a/sysdeps/aarch64/configure.ac b/sysdeps/aarch64/configure.ac +index 27874eceb44911e4..56d12d661d726892 100644 +--- a/sysdeps/aarch64/configure.ac ++++ b/sysdeps/aarch64/configure.ac +@@ -90,9 +90,10 @@ LIBC_CONFIG_VAR([aarch64-variant-pcs], [$libc_cv_aarch64_variant_pcs]) + # Check if asm support armv8.2-a+sve + AC_CACHE_CHECK([for SVE support in assembler], [libc_cv_aarch64_sve_asm], [dnl + cat > conftest.s <<\EOF +- ptrue p0.b ++ .arch armv8.2-a+sve ++ ptrue p0.b + EOF +-if AC_TRY_COMMAND(${CC-cc} -c -march=armv8.2-a+sve conftest.s 1>&AS_MESSAGE_LOG_FD); then ++if AC_TRY_COMMAND(${CC-cc} -c conftest.s 1>&AS_MESSAGE_LOG_FD); then + libc_cv_aarch64_sve_asm=yes + else + libc_cv_aarch64_sve_asm=no diff --git a/SOURCES/glibc-upstream-2.39-25.patch b/SOURCES/glibc-upstream-2.39-25.patch new file mode 100644 index 0000000..f168f15 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-25.patch @@ -0,0 +1,144 @@ +commit 9d92452c70805a2e2dbbdb2b1ffc34bd86e1c8df +Author: Wilco Dijkstra +Date: Thu Mar 21 16:48:33 2024 +0000 + + AArch64: Check kernel version for SVE ifuncs + + Old Linux kernels disable SVE after every system call. Calling the + SVE-optimized memcpy afterwards will then cause a trap to reenable SVE. + As a result, applications with a high use of syscalls may run slower with + the SVE memcpy. This is true for kernels between 4.15.0 and before 6.2.0, + except for 5.14.0 which was patched. Avoid this by checking the kernel + version and selecting the SVE ifunc on modern kernels. + + Parse the kernel version reported by uname() into a 24-bit kernel.major.minor + value without calling any library functions. If uname() is not supported or + if the version format is not recognized, assume the kernel is modern. + + Tested-by: Florian Weimer + Reviewed-by: Szabolcs Nagy + (cherry picked from commit 2e94e2f5d2bf2de124c8ad7da85463355e54ccb2) + +diff --git a/sysdeps/aarch64/cpu-features.h b/sysdeps/aarch64/cpu-features.h +index 77a782422af1b6e4..5f2da91ebbd0adaf 100644 +--- a/sysdeps/aarch64/cpu-features.h ++++ b/sysdeps/aarch64/cpu-features.h +@@ -71,6 +71,7 @@ struct cpu_features + /* Currently, the GLIBC memory tagging tunable only defines 8 bits. */ + uint8_t mte_state; + bool sve; ++ bool prefer_sve_ifuncs; + bool mops; + }; + +diff --git a/sysdeps/aarch64/multiarch/init-arch.h b/sysdeps/aarch64/multiarch/init-arch.h +index c52860efb22d70eb..61dc40088f4d9e5e 100644 +--- a/sysdeps/aarch64/multiarch/init-arch.h ++++ b/sysdeps/aarch64/multiarch/init-arch.h +@@ -36,5 +36,7 @@ + MTE_ENABLED (); \ + bool __attribute__((unused)) sve = \ + GLRO(dl_aarch64_cpu_features).sve; \ ++ bool __attribute__((unused)) prefer_sve_ifuncs = \ ++ GLRO(dl_aarch64_cpu_features).prefer_sve_ifuncs; \ + bool __attribute__((unused)) mops = \ + GLRO(dl_aarch64_cpu_features).mops; +diff --git a/sysdeps/aarch64/multiarch/memcpy.c b/sysdeps/aarch64/multiarch/memcpy.c +index d12eccfca51f4bcf..ce53567dab33c2f0 100644 +--- a/sysdeps/aarch64/multiarch/memcpy.c ++++ b/sysdeps/aarch64/multiarch/memcpy.c +@@ -47,7 +47,7 @@ select_memcpy_ifunc (void) + { + if (IS_A64FX (midr)) + return __memcpy_a64fx; +- return __memcpy_sve; ++ return prefer_sve_ifuncs ? __memcpy_sve : __memcpy_generic; + } + + if (IS_THUNDERX (midr)) +diff --git a/sysdeps/aarch64/multiarch/memmove.c b/sysdeps/aarch64/multiarch/memmove.c +index 2081eeb4d40e0240..fe95037be391896c 100644 +--- a/sysdeps/aarch64/multiarch/memmove.c ++++ b/sysdeps/aarch64/multiarch/memmove.c +@@ -47,7 +47,7 @@ select_memmove_ifunc (void) + { + if (IS_A64FX (midr)) + return __memmove_a64fx; +- return __memmove_sve; ++ return prefer_sve_ifuncs ? __memmove_sve : __memmove_generic; + } + + if (IS_THUNDERX (midr)) +diff --git a/sysdeps/unix/sysv/linux/aarch64/cpu-features.c b/sysdeps/unix/sysv/linux/aarch64/cpu-features.c +index b1a3f673f067280b..c0b047bc0dbeae42 100644 +--- a/sysdeps/unix/sysv/linux/aarch64/cpu-features.c ++++ b/sysdeps/unix/sysv/linux/aarch64/cpu-features.c +@@ -21,6 +21,7 @@ + #include + #include + #include ++#include + #include + + #define DCZID_DZP_MASK (1 << 4) +@@ -62,6 +63,46 @@ get_midr_from_mcpu (const struct tunable_str_t *mcpu) + return UINT64_MAX; + } + ++#if __LINUX_KERNEL_VERSION < 0x060200 ++ ++/* Return true if we prefer using SVE in string ifuncs. Old kernels disable ++ SVE after every system call which results in unnecessary traps if memcpy ++ uses SVE. This is true for kernels between 4.15.0 and before 6.2.0, except ++ for 5.14.0 which was patched. For these versions return false to avoid using ++ SVE ifuncs. ++ Parse the kernel version into a 24-bit kernel.major.minor value without ++ calling any library functions. If uname() is not supported or if the version ++ format is not recognized, assume the kernel is modern and return true. */ ++ ++static inline bool ++prefer_sve_ifuncs (void) ++{ ++ struct utsname buf; ++ const char *p = &buf.release[0]; ++ int kernel = 0; ++ int val; ++ ++ if (__uname (&buf) < 0) ++ return true; ++ ++ for (int shift = 16; shift >= 0; shift -= 8) ++ { ++ for (val = 0; *p >= '0' && *p <= '9'; p++) ++ val = val * 10 + *p - '0'; ++ kernel |= (val & 255) << shift; ++ if (*p++ != '.') ++ break; ++ } ++ ++ if (kernel >= 0x060200 || kernel == 0x050e00) ++ return true; ++ if (kernel >= 0x040f00) ++ return false; ++ return true; ++} ++ ++#endif ++ + static inline void + init_cpu_features (struct cpu_features *cpu_features) + { +@@ -126,6 +167,13 @@ init_cpu_features (struct cpu_features *cpu_features) + /* Check if SVE is supported. */ + cpu_features->sve = GLRO (dl_hwcap) & HWCAP_SVE; + ++ cpu_features->prefer_sve_ifuncs = cpu_features->sve; ++ ++#if __LINUX_KERNEL_VERSION < 0x060200 ++ if (cpu_features->sve) ++ cpu_features->prefer_sve_ifuncs = prefer_sve_ifuncs (); ++#endif ++ + /* Check if MOPS is supported. */ + cpu_features->mops = GLRO (dl_hwcap2) & HWCAP2_MOPS; + } diff --git a/SOURCES/glibc-upstream-2.39-26.patch b/SOURCES/glibc-upstream-2.39-26.patch new file mode 100644 index 0000000..278d238 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-26.patch @@ -0,0 +1,108 @@ +commit 9883f4304cfb1558d0f1e6d9f48c4ab0a35355fe +Author: H.J. Lu +Date: Wed Feb 28 09:51:14 2024 -0800 + + x86-64: Don't use SSE resolvers for ISA level 3 or above + + When glibc is built with ISA level 3 or above enabled, SSE resolvers + aren't available and glibc fails to build: + + ld: .../elf/librtld.os: in function `init_cpu_features': + .../elf/../sysdeps/x86/cpu-features.c:1200:(.text+0x1445f): undefined reference to `_dl_runtime_resolve_fxsave' + ld: .../elf/librtld.os: relocation R_X86_64_PC32 against undefined hidden symbol `_dl_runtime_resolve_fxsave' can not be used when making a shared object + /usr/local/bin/ld: final link failed: bad value + + For ISA level 3 or above, don't use _dl_runtime_resolve_fxsave nor + _dl_tlsdesc_dynamic_fxsave. + + This fixes BZ #31429. + Reviewed-by: Noah Goldstein + + (cherry picked from commit befe2d3c4dec8be2cdd01a47132e47bdb7020922) + +diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c +index 6fe1b728c607f39e..b8abe733abe54fc4 100644 +--- a/sysdeps/x86/cpu-features.c ++++ b/sysdeps/x86/cpu-features.c +@@ -18,6 +18,7 @@ + + #include + #include ++#include + #include + #include + #include +@@ -1198,7 +1199,9 @@ no_cpuid: + TUNABLE_CALLBACK (set_x86_shstk)); + #endif + ++#if MINIMUM_X86_ISA_LEVEL < AVX_X86_ISA_LEVEL + if (GLRO(dl_x86_cpu_features).xsave_state_size != 0) ++#endif + { + if (CPU_FEATURE_USABLE_P (cpu_features, XSAVEC)) + { +@@ -1219,22 +1222,24 @@ no_cpuid: + #endif + } + } ++#if MINIMUM_X86_ISA_LEVEL < AVX_X86_ISA_LEVEL + else + { +-#ifdef __x86_64__ ++# ifdef __x86_64__ + GLRO(dl_x86_64_runtime_resolve) = _dl_runtime_resolve_fxsave; +-# ifdef SHARED ++# ifdef SHARED + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; +-# endif +-#else +-# ifdef SHARED ++# endif ++# else ++# ifdef SHARED + if (CPU_FEATURE_USABLE_P (cpu_features, FXSR)) + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; + else + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fnsave; ++# endif + # endif +-#endif + } ++#endif + + #ifdef SHARED + # ifdef __x86_64__ +diff --git a/sysdeps/x86_64/dl-tlsdesc.S b/sysdeps/x86_64/dl-tlsdesc.S +index ea69f5223a77e0c0..057a10862afd6208 100644 +--- a/sysdeps/x86_64/dl-tlsdesc.S ++++ b/sysdeps/x86_64/dl-tlsdesc.S +@@ -20,6 +20,7 @@ + #include + #include + #include ++#include + #include "tlsdesc.h" + #include "dl-trampoline-save.h" + +@@ -79,12 +80,14 @@ _dl_tlsdesc_undefweak: + .size _dl_tlsdesc_undefweak, .-_dl_tlsdesc_undefweak + + #ifdef SHARED +-# define USE_FXSAVE +-# define STATE_SAVE_ALIGNMENT 16 +-# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_fxsave +-# include "dl-tlsdesc-dynamic.h" +-# undef _dl_tlsdesc_dynamic +-# undef USE_FXSAVE ++# if MINIMUM_X86_ISA_LEVEL < AVX_X86_ISA_LEVEL ++# define USE_FXSAVE ++# define STATE_SAVE_ALIGNMENT 16 ++# define _dl_tlsdesc_dynamic _dl_tlsdesc_dynamic_fxsave ++# include "dl-tlsdesc-dynamic.h" ++# undef _dl_tlsdesc_dynamic ++# undef USE_FXSAVE ++# endif + + # define USE_XSAVE + # define STATE_SAVE_ALIGNMENT 64 diff --git a/SOURCES/glibc-upstream-2.39-27.patch b/SOURCES/glibc-upstream-2.39-27.patch new file mode 100644 index 0000000..8f136ce --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-27.patch @@ -0,0 +1,61 @@ +commit 7b92f46f04c6cbce19d19ae1099628431858996c +Author: Sunil K Pandey +Date: Thu Feb 29 17:57:02 2024 -0800 + + x86-64: Simplify minimum ISA check ifdef conditional with if + + Replace minimum ISA check ifdef conditional with if. Since + MINIMUM_X86_ISA_LEVEL and AVX_X86_ISA_LEVEL are compile time constants, + compiler will perform constant folding optimization, getting same + results. + + Reviewed-by: H.J. Lu + (cherry picked from commit b6e3898194bbae78910bbe9cd086937014961e45) + +diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c +index b8abe733abe54fc4..3d7c2819d7cc6643 100644 +--- a/sysdeps/x86/cpu-features.c ++++ b/sysdeps/x86/cpu-features.c +@@ -1199,9 +1199,8 @@ no_cpuid: + TUNABLE_CALLBACK (set_x86_shstk)); + #endif + +-#if MINIMUM_X86_ISA_LEVEL < AVX_X86_ISA_LEVEL +- if (GLRO(dl_x86_cpu_features).xsave_state_size != 0) +-#endif ++ if (MINIMUM_X86_ISA_LEVEL >= AVX_X86_ISA_LEVEL ++ || (GLRO(dl_x86_cpu_features).xsave_state_size != 0)) + { + if (CPU_FEATURE_USABLE_P (cpu_features, XSAVEC)) + { +@@ -1222,24 +1221,22 @@ no_cpuid: + #endif + } + } +-#if MINIMUM_X86_ISA_LEVEL < AVX_X86_ISA_LEVEL + else + { +-# ifdef __x86_64__ ++#ifdef __x86_64__ + GLRO(dl_x86_64_runtime_resolve) = _dl_runtime_resolve_fxsave; +-# ifdef SHARED ++# ifdef SHARED + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; +-# endif +-# else +-# ifdef SHARED ++# endif ++#else ++# ifdef SHARED + if (CPU_FEATURE_USABLE_P (cpu_features, FXSR)) + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fxsave; + else + GLRO(dl_x86_tlsdesc_dynamic) = _dl_tlsdesc_dynamic_fnsave; +-# endif + # endif +- } + #endif ++ } + + #ifdef SHARED + # ifdef __x86_64__ diff --git a/SOURCES/glibc-upstream-2.39-28.patch b/SOURCES/glibc-upstream-2.39-28.patch new file mode 100644 index 0000000..b5546e0 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-28.patch @@ -0,0 +1,50 @@ +commit edb9a76e3008725e9dc035d38a58e849a3bde0f1 +Author: Florian Weimer +Date: Sun Apr 14 08:24:51 2024 +0200 + + powerpc: Fix ld.so address determination for PCREL mode (bug 31640) + + This seems to have stopped working with some GCC 14 versions, + which clobber r2. With other compilers, the kernel-provided + r2 value is still available at this point. + + Reviewed-by: Peter Bergner + (cherry picked from commit 14e56bd4ce15ac2d1cc43f762eb2e6b83fec1afe) + +diff --git a/sysdeps/powerpc/powerpc64/dl-machine.h b/sysdeps/powerpc/powerpc64/dl-machine.h +index c6682f3445615636..2b6f5d2b08cb10b8 100644 +--- a/sysdeps/powerpc/powerpc64/dl-machine.h ++++ b/sysdeps/powerpc/powerpc64/dl-machine.h +@@ -78,6 +78,7 @@ elf_host_tolerates_class (const Elf64_Ehdr *ehdr) + static inline Elf64_Addr + elf_machine_load_address (void) __attribute__ ((const)); + ++#ifndef __PCREL__ + static inline Elf64_Addr + elf_machine_load_address (void) + { +@@ -105,6 +106,24 @@ elf_machine_dynamic (void) + /* Then subtract off the load address offset. */ + return runtime_dynamic - elf_machine_load_address() ; + } ++#else /* __PCREL__ */ ++/* In PCREL mode, r2 may have been clobbered. Rely on relative ++ relocations instead. */ ++ ++static inline ElfW(Addr) ++elf_machine_load_address (void) ++{ ++ extern const ElfW(Ehdr) __ehdr_start attribute_hidden; ++ return (ElfW(Addr)) &__ehdr_start; ++} ++ ++static inline ElfW(Addr) ++elf_machine_dynamic (void) ++{ ++ extern ElfW(Dyn) _DYNAMIC[] attribute_hidden; ++ return (ElfW(Addr)) _DYNAMIC - elf_machine_load_address (); ++} ++#endif /* __PCREL__ */ + + /* The PLT uses Elf64_Rela relocs. */ + #define elf_machine_relplt elf_machine_rela diff --git a/SOURCES/glibc-upstream-2.39-29.patch b/SOURCES/glibc-upstream-2.39-29.patch new file mode 100644 index 0000000..41a85b1 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-29.patch @@ -0,0 +1,246 @@ +commit 04df8652eb1919da18d54b3dcd6db1675993d45d +Author: H.J. Lu +Date: Thu Feb 15 11:19:56 2024 -0800 + + Apply the Makefile sorting fix + + Apply the Makefile sorting fix generated by sort-makefile-lines.py. + + (cherry picked from commit ef7f4b1fef67430a8f3cfc77fa6aada2add851d7) + +diff --git a/sysdeps/loongarch/lp64/multiarch/Makefile b/sysdeps/loongarch/lp64/multiarch/Makefile +index fe863e1ba411cc4b..01762ef526854d54 100644 +--- a/sysdeps/loongarch/lp64/multiarch/Makefile ++++ b/sysdeps/loongarch/lp64/multiarch/Makefile +@@ -1,52 +1,52 @@ + ifeq ($(subdir),string) + sysdep_routines += \ +- strlen-aligned \ +- strlen-lsx \ +- strlen-lasx \ +- strnlen-aligned \ +- strnlen-lsx \ +- strnlen-lasx \ ++ memchr-aligned \ ++ memchr-lasx \ ++ memchr-lsx \ ++ memcmp-aligned \ ++ memcmp-lasx \ ++ memcmp-lsx \ ++ memcpy-aligned \ ++ memcpy-unaligned \ ++ memmove-lasx \ ++ memmove-lsx \ ++ memmove-unaligned \ ++ memrchr-generic \ ++ memrchr-lasx \ ++ memrchr-lsx \ ++ memset-aligned \ ++ memset-lasx \ ++ memset-lsx \ ++ memset-unaligned \ ++ rawmemchr-aligned \ ++ rawmemchr-lasx \ ++ rawmemchr-lsx \ ++ stpcpy-aligned \ ++ stpcpy-lasx \ ++ stpcpy-lsx \ ++ stpcpy-unaligned \ + strchr-aligned \ +- strchr-lsx \ + strchr-lasx \ +- strrchr-aligned \ +- strrchr-lsx \ +- strrchr-lasx \ ++ strchr-lsx \ + strchrnul-aligned \ +- strchrnul-lsx \ + strchrnul-lasx \ ++ strchrnul-lsx \ + strcmp-aligned \ + strcmp-lsx \ +- strncmp-aligned \ +- strncmp-lsx \ + strcpy-aligned \ +- strcpy-unaligned \ +- strcpy-lsx \ + strcpy-lasx \ +- stpcpy-aligned \ +- stpcpy-unaligned \ +- stpcpy-lsx \ +- stpcpy-lasx \ +- memcpy-aligned \ +- memcpy-unaligned \ +- memmove-unaligned \ +- memmove-lsx \ +- memmove-lasx \ +- rawmemchr-aligned \ +- rawmemchr-lsx \ +- rawmemchr-lasx \ +- memchr-aligned \ +- memchr-lsx \ +- memchr-lasx \ +- memrchr-generic \ +- memrchr-lsx \ +- memrchr-lasx \ +- memset-aligned \ +- memset-unaligned \ +- memset-lsx \ +- memset-lasx \ +- memcmp-aligned \ +- memcmp-lsx \ +- memcmp-lasx \ ++ strcpy-lsx \ ++ strcpy-unaligned \ ++ strlen-aligned \ ++ strlen-lasx \ ++ strlen-lsx \ ++ strncmp-aligned \ ++ strncmp-lsx \ ++ strnlen-aligned \ ++ strnlen-lasx \ ++ strnlen-lsx \ ++ strrchr-aligned \ ++ strrchr-lasx \ ++ strrchr-lsx \ + # sysdep_routines + endif +diff --git a/sysdeps/x86/Makefile b/sysdeps/x86/Makefile +index 992aabe43ec60abf..5311b594aff62f7c 100644 +--- a/sysdeps/x86/Makefile ++++ b/sysdeps/x86/Makefile +@@ -15,18 +15,18 @@ CFLAGS-dl-get-cpu-features.os += $(rtld-early-cflags) + CFLAGS-get-cpuid-feature-leaf.o += $(no-stack-protector) + + tests += \ +- tst-get-cpu-features \ +- tst-get-cpu-features-static \ + tst-cpu-features-cpuinfo \ + tst-cpu-features-cpuinfo-static \ + tst-cpu-features-supports \ + tst-cpu-features-supports-static \ ++ tst-get-cpu-features \ ++ tst-get-cpu-features-static \ + tst-hwcap-tunables \ + # tests + tests-static += \ +- tst-get-cpu-features-static \ + tst-cpu-features-cpuinfo-static \ + tst-cpu-features-supports-static \ ++ tst-get-cpu-features-static \ + # tests-static + ifeq (yes,$(have-ifunc)) + ifeq (yes,$(have-gcc-ifunc)) +diff --git a/sysdeps/x86_64/Makefile b/sysdeps/x86_64/Makefile +index 9d374a329916fc45..0ede447405d549b5 100644 +--- a/sysdeps/x86_64/Makefile ++++ b/sysdeps/x86_64/Makefile +@@ -252,6 +252,10 @@ sysdep-dl-routines += dl-cet + + tests += \ + tst-cet-legacy-1 \ ++ tst-cet-legacy-10 \ ++ tst-cet-legacy-10-static \ ++ tst-cet-legacy-10a \ ++ tst-cet-legacy-10a-static \ + tst-cet-legacy-1a \ + tst-cet-legacy-2 \ + tst-cet-legacy-2a \ +@@ -263,15 +267,11 @@ tests += \ + tst-cet-legacy-8 \ + tst-cet-legacy-9 \ + tst-cet-legacy-9-static \ +- tst-cet-legacy-10 \ +- tst-cet-legacy-10-static \ +- tst-cet-legacy-10a \ +- tst-cet-legacy-10a-static \ + # tests + tests-static += \ +- tst-cet-legacy-9-static \ + tst-cet-legacy-10-static \ + tst-cet-legacy-10a-static \ ++ tst-cet-legacy-9-static \ + # tests-static + tst-cet-legacy-1a-ARGS = -- $(host-test-program-cmd) + +diff --git a/sysdeps/x86_64/fpu/multiarch/Makefile b/sysdeps/x86_64/fpu/multiarch/Makefile +index ea81753b708fcb8d..e1a490dd98b4be07 100644 +--- a/sysdeps/x86_64/fpu/multiarch/Makefile ++++ b/sysdeps/x86_64/fpu/multiarch/Makefile +@@ -4,10 +4,10 @@ libm-sysdep_routines += \ + s_ceilf-c \ + s_floor-c \ + s_floorf-c \ +- s_rint-c \ +- s_rintf-c \ + s_nearbyint-c \ + s_nearbyintf-c \ ++ s_rint-c \ ++ s_rintf-c \ + s_roundeven-c \ + s_roundevenf-c \ + s_trunc-c \ +@@ -21,10 +21,10 @@ libm-sysdep_routines += \ + s_floorf-sse4_1 \ + s_nearbyint-sse4_1 \ + s_nearbyintf-sse4_1 \ +- s_roundeven-sse4_1 \ +- s_roundevenf-sse4_1 \ + s_rint-sse4_1 \ + s_rintf-sse4_1 \ ++ s_roundeven-sse4_1 \ ++ s_roundevenf-sse4_1 \ + s_trunc-sse4_1 \ + s_truncf-sse4_1 \ + # libm-sysdep_routines +@@ -84,12 +84,12 @@ CFLAGS-s_cosf-fma.c = -mfma -mavx2 + CFLAGS-s_sincosf-fma.c = -mfma -mavx2 + + libm-sysdep_routines += \ ++ e_asin-fma4 \ ++ e_atan2-fma4 \ + e_exp-fma4 \ + e_log-fma4 \ + e_pow-fma4 \ +- e_asin-fma4 \ + s_atan-fma4 \ +- e_atan2-fma4 \ + s_sin-fma4 \ + s_sincos-fma4 \ + s_tan-fma4 \ +@@ -106,10 +106,10 @@ CFLAGS-s_tan-fma4.c = -mfma4 + CFLAGS-s_sincos-fma4.c = -mfma4 + + libm-sysdep_routines += \ ++ e_atan2-avx \ + e_exp-avx \ + e_log-avx \ + s_atan-avx \ +- e_atan2-avx \ + s_sin-avx \ + s_sincos-avx \ + s_tan-avx \ +diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile +index e1e894c963afb8b1..d3d2270394bd635d 100644 +--- a/sysdeps/x86_64/multiarch/Makefile ++++ b/sysdeps/x86_64/multiarch/Makefile +@@ -4,8 +4,8 @@ sysdep_routines += \ + memchr-avx2 \ + memchr-avx2-rtm \ + memchr-evex \ +- memchr-evex512 \ + memchr-evex-rtm \ ++ memchr-evex512 \ + memchr-sse2 \ + memcmp-avx2-movbe \ + memcmp-avx2-movbe-rtm \ +@@ -37,8 +37,8 @@ sysdep_routines += \ + rawmemchr-avx2 \ + rawmemchr-avx2-rtm \ + rawmemchr-evex \ +- rawmemchr-evex512 \ + rawmemchr-evex-rtm \ ++ rawmemchr-evex512 \ + rawmemchr-sse2 \ + stpcpy-avx2 \ + stpcpy-avx2-rtm \ diff --git a/SOURCES/glibc-upstream-2.39-3.patch b/SOURCES/glibc-upstream-2.39-3.patch new file mode 100644 index 0000000..c98c21c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-3.patch @@ -0,0 +1,78 @@ +commit 312e159626b67fe11f39e83e222cf4348a3962f3 +Author: Adhemerval Zanella +Date: Thu Feb 1 14:29:53 2024 -0300 + + mips: FIx clone3 implementation (BZ 31325) + + For o32 we need to setup a minimal stack frame to allow cprestore + on __thread_start_clone3 (which instruct the linker to save the + gp for PIC). Also, there is no guarantee by kABI that $8 will be + preserved after syscall execution, so we need to save it on the + provided stack. + + Checked on mipsel-linux-gnu. + + Reported-by: Khem Raj + Tested-by: Khem Raj + (cherry picked from commit bbd248ac0d75efdef8fe61ea69b1fb25fb95b6e7) + +diff --git a/sysdeps/unix/sysv/linux/mips/clone3.S b/sysdeps/unix/sysv/linux/mips/clone3.S +index e9fec2fa471b99ba..481b8ae96366fc70 100644 +--- a/sysdeps/unix/sysv/linux/mips/clone3.S ++++ b/sysdeps/unix/sysv/linux/mips/clone3.S +@@ -37,11 +37,6 @@ + + .text + .set nomips16 +-#if _MIPS_SIM == _ABIO32 +-# define EXTRA_LOCALS 1 +-#else +-# define EXTRA_LOCALS 0 +-#endif + #define FRAMESZ ((NARGSAVE*SZREG)+ALSZ)&ALMASK + GPOFF= FRAMESZ-(1*SZREG) + NESTED(__clone3, SZREG, sp) +@@ -68,8 +63,31 @@ NESTED(__clone3, SZREG, sp) + beqz a0, L(error) /* No NULL cl_args pointer. */ + beqz a2, L(error) /* No NULL function pointer. */ + ++#if _MIPS_SIM == _ABIO32 ++ /* Both stack and stack_size on clone_args are defined as uint64_t, and ++ there is no need to handle values larger than to 32 bits for o32. */ ++# if __BYTE_ORDER == __BIG_ENDIAN ++# define CL_STACKPOINTER_OFFSET 44 ++# define CL_STACKSIZE_OFFSET 52 ++# else ++# define CL_STACKPOINTER_OFFSET 40 ++# define CL_STACKSIZE_OFFSET 48 ++# endif ++ ++ /* For o32 we need to setup a minimal stack frame to allow cprestore ++ on __thread_start_clone3. Also there is no guarantee by kABI that ++ $8 will be preserved after syscall execution (so we need to save it ++ on the provided stack). */ ++ lw t0, CL_STACKPOINTER_OFFSET(a0) /* Load the stack pointer. */ ++ lw t1, CL_STACKSIZE_OFFSET(a0) /* Load the stack_size. */ ++ addiu t1, -32 /* Update the stack size. */ ++ addu t2, t1, t0 /* Calculate the thread stack. */ ++ sw a3, 0(t2) /* Save argument pointer. */ ++ sw t1, CL_STACKSIZE_OFFSET(a0) /* Save the new stack size. */ ++#else + move $8, a3 /* a3 is set to 0/1 for syscall success/error + while a4/$8 is returned unmodified. */ ++#endif + + /* Do the system call, the kernel expects: + v0: system call number +@@ -125,7 +143,11 @@ L(thread_start_clone3): + + /* Restore the arg for user's function. */ + move t9, a2 /* Function pointer. */ ++#if _MIPS_SIM == _ABIO32 ++ PTR_L a0, 0(sp) ++#else + move a0, $8 /* Argument pointer. */ ++#endif + + /* Call the user's function. */ + jal t9 diff --git a/SOURCES/glibc-upstream-2.39-30.patch b/SOURCES/glibc-upstream-2.39-30.patch new file mode 100644 index 0000000..3796da6 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-30.patch @@ -0,0 +1,2143 @@ +commit 423099a03264ea28298f47355d7811b8efe03c97 +Author: Sunil K Pandey +Date: Tue Feb 13 12:23:14 2024 -0800 + + x86_64: Exclude SSE, AVX and FMA4 variants in libm multiarch + + When glibc is built with ISA level 3 or higher by default, the resulting + glibc binaries won't run on SSE or FMA4 processors. Exclude SSE, AVX and + FMA4 variants in libm multiarch when ISA level 3 or higher is enabled by + default. + + When glibc is built with ISA level 2 enabled by default, only keep SSE4.1 + variant. + + Fixes BZ 31335. + + NB: elf/tst-valgrind-smoke test fails with ISA level 4, because valgrind + doesn't support AVX512 instructions: + + https://bugs.kde.org/show_bug.cgi?id=383010 + + Reviewed-by: H.J. Lu + (cherry picked from commit 9f78a7c1d0963282608da836b840f0d5ae1c478e) + +diff --git a/sysdeps/x86/configure b/sysdeps/x86/configure +index 1f4c2d67fd2693ed..2a5421bb31e9efe4 100644 +--- a/sysdeps/x86/configure ++++ b/sysdeps/x86/configure +@@ -98,6 +98,7 @@ printf "%s\n" "$libc_cv_have_x86_lahf_sahf" >&6; } + if test $libc_cv_have_x86_lahf_sahf = yes; then + printf "%s\n" "#define HAVE_X86_LAHF_SAHF 1" >>confdefs.h + ++ ISAFLAG="-DHAVE_X86_LAHF_SAHF" + fi + { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for MOVBE instruction support" >&5 + printf %s "checking for MOVBE instruction support... " >&6; } +@@ -120,9 +121,41 @@ printf "%s\n" "$libc_cv_have_x86_movbe" >&6; } + if test $libc_cv_have_x86_movbe = yes; then + printf "%s\n" "#define HAVE_X86_MOVBE 1" >>confdefs.h + ++ ISAFLAG="$ISAFLAG -DHAVE_X86_MOVBE" + fi ++ ++ # Check for ISA level support. ++ { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for ISA level support" >&5 ++printf %s "checking for ISA level support... " >&6; } ++if test ${libc_cv_have_x86_isa_level+y} ++then : ++ printf %s "(cached) " >&6 ++else $as_nop ++ cat > conftest.c < ++#if MINIMUM_X86_ISA_LEVEL >= 4 ++libc_cv_have_x86_isa_level=4 ++#elif MINIMUM_X86_ISA_LEVEL == 3 ++libc_cv_have_x86_isa_level=3 ++#elif MINIMUM_X86_ISA_LEVEL == 2 ++libc_cv_have_x86_isa_level=2 ++#else ++libc_cv_have_x86_isa_level=baseline ++#endif ++EOF ++ eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` ++ rm -rf conftest* ++fi ++{ printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_have_x86_isa_level" >&5 ++printf "%s\n" "$libc_cv_have_x86_isa_level" >&6; } ++else ++ libc_cv_have_x86_isa_level=baseline + fi + config_vars="$config_vars ++have-x86-isa-level = $libc_cv_have_x86_isa_level" ++config_vars="$config_vars ++x86-isa-level-3-or-above = 3 4" ++config_vars="$config_vars + enable-x86-isa-level = $libc_cv_include_x86_isa_level" + + printf "%s\n" "#define SUPPORT_STATIC_PIE 1" >>confdefs.h +diff --git a/sysdeps/x86/configure.ac b/sysdeps/x86/configure.ac +index 437a50623b35163a..78ff7c8f41c552bc 100644 +--- a/sysdeps/x86/configure.ac ++++ b/sysdeps/x86/configure.ac +@@ -72,6 +72,7 @@ if test $libc_cv_include_x86_isa_level = yes; then + fi]) + if test $libc_cv_have_x86_lahf_sahf = yes; then + AC_DEFINE(HAVE_X86_LAHF_SAHF) ++ ISAFLAG="-DHAVE_X86_LAHF_SAHF" + fi + AC_CACHE_CHECK([for MOVBE instruction support], + libc_cv_have_x86_movbe, [dnl +@@ -81,8 +82,31 @@ if test $libc_cv_include_x86_isa_level = yes; then + fi]) + if test $libc_cv_have_x86_movbe = yes; then + AC_DEFINE(HAVE_X86_MOVBE) ++ ISAFLAG="$ISAFLAG -DHAVE_X86_MOVBE" + fi ++ ++ # Check for ISA level support. ++ AC_CACHE_CHECK([for ISA level support], ++ libc_cv_have_x86_isa_level, [dnl ++cat > conftest.c < ++#if MINIMUM_X86_ISA_LEVEL >= 4 ++libc_cv_have_x86_isa_level=4 ++#elif MINIMUM_X86_ISA_LEVEL == 3 ++libc_cv_have_x86_isa_level=3 ++#elif MINIMUM_X86_ISA_LEVEL == 2 ++libc_cv_have_x86_isa_level=2 ++#else ++libc_cv_have_x86_isa_level=baseline ++#endif ++EOF ++ eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` ++ rm -rf conftest*]) ++else ++ libc_cv_have_x86_isa_level=baseline + fi ++LIBC_CONFIG_VAR([have-x86-isa-level], [$libc_cv_have_x86_isa_level]) ++LIBC_CONFIG_VAR([x86-isa-level-3-or-above], [3 4]) + LIBC_CONFIG_VAR([enable-x86-isa-level], [$libc_cv_include_x86_isa_level]) + + dnl Static PIE is supported. +diff --git a/sysdeps/x86_64/fpu/multiarch/Makefile b/sysdeps/x86_64/fpu/multiarch/Makefile +index e1a490dd98b4be07..6ddd50240ce33d22 100644 +--- a/sysdeps/x86_64/fpu/multiarch/Makefile ++++ b/sysdeps/x86_64/fpu/multiarch/Makefile +@@ -1,49 +1,4 @@ + ifeq ($(subdir),math) +-libm-sysdep_routines += \ +- s_ceil-c \ +- s_ceilf-c \ +- s_floor-c \ +- s_floorf-c \ +- s_nearbyint-c \ +- s_nearbyintf-c \ +- s_rint-c \ +- s_rintf-c \ +- s_roundeven-c \ +- s_roundevenf-c \ +- s_trunc-c \ +- s_truncf-c \ +-# libm-sysdep_routines +- +-libm-sysdep_routines += \ +- s_ceil-sse4_1 \ +- s_ceilf-sse4_1 \ +- s_floor-sse4_1 \ +- s_floorf-sse4_1 \ +- s_nearbyint-sse4_1 \ +- s_nearbyintf-sse4_1 \ +- s_rint-sse4_1 \ +- s_rintf-sse4_1 \ +- s_roundeven-sse4_1 \ +- s_roundevenf-sse4_1 \ +- s_trunc-sse4_1 \ +- s_truncf-sse4_1 \ +-# libm-sysdep_routines +- +-libm-sysdep_routines += \ +- e_asin-fma \ +- e_atan2-fma \ +- e_exp-fma \ +- e_log-fma \ +- e_log2-fma \ +- e_pow-fma \ +- s_atan-fma \ +- s_expm1-fma \ +- s_log1p-fma \ +- s_sin-fma \ +- s_sincos-fma \ +- s_tan-fma \ +-# libm-sysdep_routines +- + CFLAGS-e_asin-fma.c = -mfma -mavx2 + CFLAGS-e_atan2-fma.c = -mfma -mavx2 + CFLAGS-e_exp-fma.c = -mfma -mavx2 +@@ -57,23 +12,6 @@ CFLAGS-s_sin-fma.c = -mfma -mavx2 + CFLAGS-s_tan-fma.c = -mfma -mavx2 + CFLAGS-s_sincos-fma.c = -mfma -mavx2 + +-libm-sysdep_routines += \ +- s_cosf-sse2 \ +- s_sincosf-sse2 \ +- s_sinf-sse2 \ +-# libm-sysdep_routines +- +-libm-sysdep_routines += \ +- e_exp2f-fma \ +- e_expf-fma \ +- e_log2f-fma \ +- e_logf-fma \ +- e_powf-fma \ +- s_cosf-fma \ +- s_sincosf-fma \ +- s_sinf-fma \ +-# libm-sysdep_routines +- + CFLAGS-e_exp2f-fma.c = -mfma -mavx2 + CFLAGS-e_expf-fma.c = -mfma -mavx2 + CFLAGS-e_log2f-fma.c = -mfma -mavx2 +@@ -83,17 +21,93 @@ CFLAGS-s_sinf-fma.c = -mfma -mavx2 + CFLAGS-s_cosf-fma.c = -mfma -mavx2 + CFLAGS-s_sincosf-fma.c = -mfma -mavx2 + ++# Check if ISA level is 3 or above. ++ifneq (,$(filter $(have-x86-isa-level),$(x86-isa-level-3-or-above))) + libm-sysdep_routines += \ ++ s_ceil-avx \ ++ s_ceilf-avx \ ++ s_floor-avx \ ++ s_floorf-avx \ ++ s_nearbyint-avx \ ++ s_nearbyintf-avx \ ++ s_rint-avx \ ++ s_rintf-avx \ ++ s_roundeven-avx \ ++ s_roundevenf-avx \ ++ s_trunc-avx \ ++ s_truncf-avx \ ++# libm-sysdep_routines ++else ++libm-sysdep_routines += \ ++ e_asin-fma \ + e_asin-fma4 \ ++ e_atan2-avx \ ++ e_atan2-fma \ + e_atan2-fma4 \ ++ e_exp-avx \ ++ e_exp-fma \ + e_exp-fma4 \ ++ e_exp2f-fma \ ++ e_expf-fma \ ++ e_log-avx \ ++ e_log-fma \ + e_log-fma4 \ ++ e_log2-fma \ ++ e_log2f-fma \ ++ e_logf-fma \ ++ e_pow-fma \ + e_pow-fma4 \ ++ e_powf-fma \ ++ s_atan-avx \ ++ s_atan-fma \ + s_atan-fma4 \ ++ s_ceil-sse4_1 \ ++ s_ceilf-sse4_1 \ ++ s_cosf-fma \ ++ s_cosf-sse2 \ ++ s_expm1-fma \ ++ s_floor-sse4_1 \ ++ s_floorf-sse4_1 \ ++ s_log1p-fma \ ++ s_nearbyint-sse4_1 \ ++ s_nearbyintf-sse4_1 \ ++ s_rint-sse4_1 \ ++ s_rintf-sse4_1 \ ++ s_roundeven-sse4_1 \ ++ s_roundevenf-sse4_1 \ ++ s_sin-avx \ ++ s_sin-fma \ + s_sin-fma4 \ ++ s_sincos-avx \ ++ s_sincos-fma \ + s_sincos-fma4 \ ++ s_sincosf-fma \ ++ s_sincosf-sse2 \ ++ s_sinf-fma \ ++ s_sinf-sse2 \ ++ s_tan-avx \ ++ s_tan-fma \ + s_tan-fma4 \ ++ s_trunc-sse4_1 \ ++ s_truncf-sse4_1 \ + # libm-sysdep_routines ++ifeq ($(have-x86-isa-level),baseline) ++libm-sysdep_routines += \ ++ s_ceil-c \ ++ s_ceilf-c \ ++ s_floor-c \ ++ s_floorf-c \ ++ s_nearbyint-c \ ++ s_nearbyintf-c \ ++ s_rint-c \ ++ s_rintf-c \ ++ s_roundeven-c \ ++ s_roundevenf-c \ ++ s_trunc-c \ ++ s_truncf-c \ ++# libm-sysdep_routines ++endif ++endif + + CFLAGS-e_asin-fma4.c = -mfma4 + CFLAGS-e_atan2-fma4.c = -mfma4 +@@ -105,16 +119,6 @@ CFLAGS-s_sin-fma4.c = -mfma4 + CFLAGS-s_tan-fma4.c = -mfma4 + CFLAGS-s_sincos-fma4.c = -mfma4 + +-libm-sysdep_routines += \ +- e_atan2-avx \ +- e_exp-avx \ +- e_log-avx \ +- s_atan-avx \ +- s_sin-avx \ +- s_sincos-avx \ +- s_tan-avx \ +-# libm-sysdep_routines +- + CFLAGS-e_atan2-avx.c = -msse2avx -DSSE2AVX + CFLAGS-e_exp-avx.c = -msse2avx -DSSE2AVX + CFLAGS-e_log-avx.c = -msse2avx -DSSE2AVX +diff --git a/sysdeps/x86_64/fpu/multiarch/e_asin.c b/sysdeps/x86_64/fpu/multiarch/e_asin.c +index 2eaa6c2c0490d8fc..d64fca2586d43f03 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_asin.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_asin.c +@@ -16,26 +16,29 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_ieee754_asin (double); + extern double __redirect_ieee754_acos (double); + +-#define SYMBOL_NAME ieee754_asin +-#include "ifunc-fma4.h" ++# define SYMBOL_NAME ieee754_asin ++# include "ifunc-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_asin, __ieee754_asin, + IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_asin, __asin) + +-#undef SYMBOL_NAME +-#define SYMBOL_NAME ieee754_acos +-#include "ifunc-fma4.h" ++# undef SYMBOL_NAME ++# define SYMBOL_NAME ieee754_acos ++# include "ifunc-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_acos, __ieee754_acos, + IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_acos, __acos) + +-#define __ieee754_acos __ieee754_acos_sse2 +-#define __ieee754_asin __ieee754_asin_sse2 ++# define __ieee754_acos __ieee754_acos_sse2 ++# define __ieee754_asin __ieee754_asin_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_atan2.c b/sysdeps/x86_64/fpu/multiarch/e_atan2.c +index 17ee4f3c366b57c4..8a86c14ded1784dd 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_atan2.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_atan2.c +@@ -16,16 +16,19 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_ieee754_atan2 (double, double); + +-#define SYMBOL_NAME ieee754_atan2 +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME ieee754_atan2 ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_atan2, + __ieee754_atan2, IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_atan2, __atan2) + +-#define __ieee754_atan2 __ieee754_atan2_sse2 ++# define __ieee754_atan2 __ieee754_atan2_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_exp.c b/sysdeps/x86_64/fpu/multiarch/e_exp.c +index 406b7ebd44d8eba1..d56329291a204ecf 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_exp.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_exp.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern double __redirect_ieee754_exp (double); + +-#define SYMBOL_NAME ieee754_exp +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME ieee754_exp ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_exp, __ieee754_exp, + IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_exp, __exp) + +-#define __exp __ieee754_exp_sse2 ++# define __exp __ieee754_exp_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_exp2f.c b/sysdeps/x86_64/fpu/multiarch/e_exp2f.c +index 804fd6be850926b3..06fe5028d6e72122 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_exp2f.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_exp2f.c +@@ -16,25 +16,28 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern float __redirect_exp2f (float); + +-#define SYMBOL_NAME exp2f +-#include "ifunc-fma.h" ++# define SYMBOL_NAME exp2f ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_exp2f, __exp2f, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + versioned_symbol (libm, __ieee754_exp2f, exp2f, GLIBC_2_27); + libm_alias_float_other (__exp2, exp2) +-#else ++# else + libm_alias_float (__exp2, exp2) +-#endif ++# endif + + strong_alias (__exp2f, __ieee754_exp2f) + libm_alias_finite (__exp2f, __exp2f) + +-#define __exp2f __exp2f_sse2 ++# define __exp2f __exp2f_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_expf.c b/sysdeps/x86_64/fpu/multiarch/e_expf.c +index 4a7e2a5bce697d60..19d767f636126c8d 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_expf.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_expf.c +@@ -16,28 +16,31 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern float __redirect_expf (float); + +-#define SYMBOL_NAME expf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME expf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_expf, __expf, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + __hidden_ver1 (__expf, __GI___expf, __redirect_expf) + __attribute__ ((visibility ("hidden"))); + + versioned_symbol (libm, __ieee754_expf, expf, GLIBC_2_27); + libm_alias_float_other (__exp, exp) +-#else ++# else + libm_alias_float (__exp, exp) +-#endif ++# endif + + strong_alias (__expf, __ieee754_expf) + libm_alias_finite (__expf, __expf) + +-#define __expf __expf_sse2 ++# define __expf __expf_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_log.c b/sysdeps/x86_64/fpu/multiarch/e_log.c +index 067fbf58c3f8e0f5..d80c1b1463954af8 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_log.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_log.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern double __redirect_ieee754_log (double); + +-#define SYMBOL_NAME ieee754_log +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME ieee754_log ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_log, __ieee754_log, + IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_log, __log) + +-#define __log __ieee754_log_sse2 ++# define __log __ieee754_log_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_log2.c b/sysdeps/x86_64/fpu/multiarch/e_log2.c +index 9c57a2f6ccfcbdfd..9686782c09b1f343 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_log2.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_log2.c +@@ -16,28 +16,31 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern double __redirect_log2 (double); + +-#define SYMBOL_NAME log2 +-#include "ifunc-fma.h" ++# define SYMBOL_NAME log2 ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_log2, __log2, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + __hidden_ver1 (__log2, __GI___log2, __redirect_log2) + __attribute__ ((visibility ("hidden"))); + + versioned_symbol (libm, __ieee754_log2, log2, GLIBC_2_29); + libm_alias_double_other (__log2, log2) +-#else ++# else + libm_alias_double (__log2, log2) +-#endif ++# endif + + strong_alias (__log2, __ieee754_log2) + libm_alias_finite (__log2, __log2) + +-#define __log2 __log2_sse2 ++# define __log2 __log2_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_log2f.c b/sysdeps/x86_64/fpu/multiarch/e_log2f.c +index 2b45c87f38afea62..8ada46e11e62726c 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_log2f.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_log2f.c +@@ -16,28 +16,31 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern float __redirect_log2f (float); + +-#define SYMBOL_NAME log2f +-#include "ifunc-fma.h" ++# define SYMBOL_NAME log2f ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_log2f, __log2f, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + __hidden_ver1 (__log2f, __GI___log2f, __redirect_log2f) + __attribute__ ((visibility ("hidden"))); + + versioned_symbol (libm, __ieee754_log2f, log2f, GLIBC_2_27); + libm_alias_float_other (__log2, log2) +-#else ++# else + libm_alias_float (__log2, log2) +-#endif ++# endif + + strong_alias (__log2f, __ieee754_log2f) + libm_alias_finite (__log2f, __log2f) + +-#define __log2f __log2f_sse2 ++# define __log2f __log2f_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_logf.c b/sysdeps/x86_64/fpu/multiarch/e_logf.c +index 97e23c8fea39f37e..a3978d9a8e40c588 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_logf.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_logf.c +@@ -16,28 +16,31 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern float __redirect_logf (float); + +-#define SYMBOL_NAME logf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME logf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_logf, __logf, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + __hidden_ver1 (__logf, __GI___logf, __redirect_logf) + __attribute__ ((visibility ("hidden"))); + + versioned_symbol (libm, __ieee754_logf, logf, GLIBC_2_27); + libm_alias_float_other (__log, log) +-#else ++# else + libm_alias_float (__log, log) +-#endif ++# endif + + strong_alias (__logf, __ieee754_logf) + libm_alias_finite (__logf, __logf) + +-#define __logf __logf_sse2 ++# define __logf __logf_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_pow.c b/sysdeps/x86_64/fpu/multiarch/e_pow.c +index 42618e7112fac769..f8f17aff9ffb9bf2 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_pow.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_pow.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + + extern double __redirect_ieee754_pow (double, double); + +-#define SYMBOL_NAME ieee754_pow +-#include "ifunc-fma4.h" ++# define SYMBOL_NAME ieee754_pow ++# include "ifunc-fma4.h" + + libc_ifunc_redirected (__redirect_ieee754_pow, + __ieee754_pow, IFUNC_SELECTOR ()); + libm_alias_finite (__ieee754_pow, __pow) + +-#define __pow __ieee754_pow_sse2 ++# define __pow __ieee754_pow_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/e_powf.c b/sysdeps/x86_64/fpu/multiarch/e_powf.c +index 8e6ce13cc1a2470b..8b1a4c7d047115f5 100644 +--- a/sysdeps/x86_64/fpu/multiarch/e_powf.c ++++ b/sysdeps/x86_64/fpu/multiarch/e_powf.c +@@ -16,31 +16,34 @@ + License along with the GNU C Library; if not, see + . */ + +-#include +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include ++# include + +-#define powf __redirect_powf +-#define __DECL_SIMD___redirect_powf +-#include +-#undef powf ++# define powf __redirect_powf ++# define __DECL_SIMD___redirect_powf ++# include ++# undef powf + +-#define SYMBOL_NAME powf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME powf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_powf, __powf, IFUNC_SELECTOR ()); + +-#ifdef SHARED ++# ifdef SHARED + __hidden_ver1 (__powf, __GI___powf, __redirect_powf) + __attribute__ ((visibility ("hidden"))); + + versioned_symbol (libm, __ieee754_powf, powf, GLIBC_2_27); + libm_alias_float_other (__pow, pow) +-#else ++# else + libm_alias_float (__pow, pow) +-#endif ++# endif + + strong_alias (__powf, __ieee754_powf) + libm_alias_finite (__powf, __powf) + +-#define __powf __powf_sse2 ++# define __powf __powf_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_atan.c b/sysdeps/x86_64/fpu/multiarch/s_atan.c +index 71bad096a9da60fe..4d2c6ce0060b6d8f 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_atan.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_atan.c +@@ -16,15 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_atan (double); + +-#define SYMBOL_NAME atan +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME atan ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_atan, __atan, IFUNC_SELECTOR ()); + libm_alias_double (__atan, atan) + +-#define __atan __atan_sse2 ++# define __atan __atan_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceil-avx.S b/sysdeps/x86_64/fpu/multiarch/s_ceil-avx.S +new file mode 100644 +index 0000000000000000..e6c1106753f29600 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceil-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of ceil function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__ceil) ++ vroundsd $10, %xmm0, %xmm0, %xmm0 ++ ret ++END(__ceil) ++ ++libm_alias_double (__ceil, ceil) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceil-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_ceil-sse4_1.S +index 64119011add00832..dba756c38f47b871 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_ceil-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceil-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __ceil_sse41 __ceil ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__ceil_sse41) + roundsd $10, %xmm0, %xmm0 + ret + END(__ceil_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__ceil, ceil) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceil.c b/sysdeps/x86_64/fpu/multiarch/s_ceil.c +index cc028addee82f19c..46c8e91e199311db 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_ceil.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceil.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define ceil __redirect_ceil +-#define __ceil __redirect___ceil +-#include +-#undef ceil +-#undef __ceil ++# define ceil __redirect_ceil ++# define __ceil __redirect___ceil ++# include ++# undef ceil ++# undef __ceil + +-#define SYMBOL_NAME ceil +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME ceil ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_ceil, __ceil, IFUNC_SELECTOR ()); + libm_alias_double (__ceil, ceil) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceilf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_ceilf-avx.S +new file mode 100644 +index 0000000000000000..b4d8ac045569900c +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceilf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of ceilf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__ceilf) ++ vroundss $10, %xmm0, %xmm0, %xmm0 ++ ret ++END(__ceilf) ++ ++libm_alias_float (__ceil, ceil) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceilf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_ceilf-sse4_1.S +index dd9a9f6b71449fc4..9abc87b91afafcbb 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_ceilf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceilf-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __ceilf_sse41 __ceilf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__ceilf_sse41) + roundss $10, %xmm0, %xmm0 + ret + END(__ceilf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__ceil, ceil) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_ceilf.c b/sysdeps/x86_64/fpu/multiarch/s_ceilf.c +index 97a0ca7d19a0d8dd..bb53108f73c00fd0 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_ceilf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_ceilf.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define ceilf __redirect_ceilf +-#define __ceilf __redirect___ceilf +-#include +-#undef ceilf +-#undef __ceilf ++# define ceilf __redirect_ceilf ++# define __ceilf __redirect___ceilf ++# include ++# undef ceilf ++# undef __ceilf + +-#define SYMBOL_NAME ceilf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME ceilf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_ceilf, __ceilf, IFUNC_SELECTOR ()); + libm_alias_float (__ceil, ceil) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_cosf.c b/sysdeps/x86_64/fpu/multiarch/s_cosf.c +index 2703c576dfa715ce..8a02e04538841602 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_cosf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_cosf.c +@@ -16,13 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern float __redirect_cosf (float); + +-#define SYMBOL_NAME cosf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME cosf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_cosf, __cosf, IFUNC_SELECTOR ()); + + libm_alias_float (__cos, cos) ++#else ++# include ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_expm1.c b/sysdeps/x86_64/fpu/multiarch/s_expm1.c +index 8a2d69f9b28fb03d..d58ef3d8f5e7933f 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_expm1.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_expm1.c +@@ -16,21 +16,24 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_expm1 (double); + +-#define SYMBOL_NAME expm1 +-#include "ifunc-fma.h" ++# define SYMBOL_NAME expm1 ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_expm1, __expm1, IFUNC_SELECTOR ()); + libm_alias_double (__expm1, expm1) + +-#define __expm1 __expm1_sse2 ++# define __expm1 __expm1_sse2 + + /* NB: __expm1 may be expanded to __expm1_sse2 in the following + prototypes. */ + extern long double __expm1l (long double); + extern long double __expm1f128 (long double); + ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floor-avx.S b/sysdeps/x86_64/fpu/multiarch/s_floor-avx.S +new file mode 100644 +index 0000000000000000..ff74b5a8bfe69423 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_floor-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of floor function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__floor) ++ vroundsd $9, %xmm0, %xmm0, %xmm0 ++ ret ++END(__floor) ++ ++libm_alias_double (__floor, floor) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floor-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_floor-sse4_1.S +index 2f7521f39f5e1c63..c9b9b0639bd24182 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_floor-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_floor-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __floor_sse41 __floor ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__floor_sse41) + roundsd $9, %xmm0, %xmm0 + ret + END(__floor_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__floor, floor) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floor.c b/sysdeps/x86_64/fpu/multiarch/s_floor.c +index 8cebd48e1008af6e..2c87dd0056419492 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_floor.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_floor.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define floor __redirect_floor +-#define __floor __redirect___floor +-#include +-#undef floor +-#undef __floor ++# define floor __redirect_floor ++# define __floor __redirect___floor ++# include ++# undef floor ++# undef __floor + +-#define SYMBOL_NAME floor +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME floor ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_floor, __floor, IFUNC_SELECTOR ()); + libm_alias_double (__floor, floor) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floorf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_floorf-avx.S +new file mode 100644 +index 0000000000000000..c378baae8e9c7ac7 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_floorf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of floorf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__floorf) ++ vroundss $9, %xmm0, %xmm0, %xmm0 ++ ret ++END(__floorf) ++ ++libm_alias_float (__floor, floor) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floorf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_floorf-sse4_1.S +index 5f6020d27daceb80..c2216899db7b71a3 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_floorf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_floorf-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __floorf_sse41 __floorf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__floorf_sse41) + roundss $9, %xmm0, %xmm0 + ret + END(__floorf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__floor, floor) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_floorf.c b/sysdeps/x86_64/fpu/multiarch/s_floorf.c +index a14e18b03c1a1dc9..a277802b6ddc1dfd 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_floorf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_floorf.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define floorf __redirect_floorf +-#define __floorf __redirect___floorf +-#include +-#undef floorf +-#undef __floorf ++# define floorf __redirect_floorf ++# define __floorf __redirect___floorf ++# include ++# undef floorf ++# undef __floorf + +-#define SYMBOL_NAME floorf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME floorf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_floorf, __floorf, IFUNC_SELECTOR ()); + libm_alias_float (__floor, floor) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_log1p.c b/sysdeps/x86_64/fpu/multiarch/s_log1p.c +index a8e1a3f21b17236c..3fa1185d81af33e7 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_log1p.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_log1p.c +@@ -16,14 +16,17 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_log1p (double); + +-#define SYMBOL_NAME log1p +-#include "ifunc-fma.h" ++# define SYMBOL_NAME log1p ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_log1p, __log1p, IFUNC_SELECTOR ()); + +-#define __log1p __log1p_sse2 ++# define __log1p __log1p_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyint-avx.S b/sysdeps/x86_64/fpu/multiarch/s_nearbyint-avx.S +new file mode 100644 +index 0000000000000000..5bfdf73c28b34350 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyint-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of nearbyint function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__nearbyint) ++ vroundsd $0xc, %xmm0, %xmm0, %xmm0 ++ ret ++END(__nearbyint) ++ ++libm_alias_double (__nearbyint, nearbyint) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyint-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_nearbyint-sse4_1.S +index 674f7eb40abb33d9..9d84410a1f347b7a 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_nearbyint-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyint-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __nearbyint_sse41 __nearbyint ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__nearbyint_sse41) + roundsd $0xc, %xmm0, %xmm0 + ret + END(__nearbyint_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__nearbyint, nearbyint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyint.c b/sysdeps/x86_64/fpu/multiarch/s_nearbyint.c +index 693e42dd4edffcc1..057a7ca60f0853cf 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_nearbyint.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyint.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# include + +-#define nearbyint __redirect_nearbyint +-#define __nearbyint __redirect___nearbyint +-#include +-#undef nearbyint +-#undef __nearbyint ++# define nearbyint __redirect_nearbyint ++# define __nearbyint __redirect___nearbyint ++# include ++# undef nearbyint ++# undef __nearbyint + +-#define SYMBOL_NAME nearbyint +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME nearbyint ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_nearbyint, __nearbyint, + IFUNC_SELECTOR ()); + libm_alias_double (__nearbyint, nearbyint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-avx.S +new file mode 100644 +index 0000000000000000..1dbaed0324daa024 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implmentation of nearbyintf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__nearbyintf) ++ vroundss $0xc, %xmm0, %xmm0, %xmm0 ++ ret ++END(__nearbyintf) ++ ++libm_alias_float (__nearbyint, nearbyint) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-sse4_1.S +index 5892bd756366f076..3cf35f92d6a0ea9b 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __nearbyintf_sse41 __nearbyintf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__nearbyintf_sse41) + roundss $0xc, %xmm0, %xmm0 + ret + END(__nearbyintf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__nearbyint, nearbyint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_nearbyintf.c b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf.c +index a0ac009f4bd43baa..41f374ba72902bcf 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_nearbyintf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_nearbyintf.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# include + +-#define nearbyintf __redirect_nearbyintf +-#define __nearbyintf __redirect___nearbyintf +-#include +-#undef nearbyintf +-#undef __nearbyintf ++# define nearbyintf __redirect_nearbyintf ++# define __nearbyintf __redirect___nearbyintf ++# include ++# undef nearbyintf ++# undef __nearbyintf + +-#define SYMBOL_NAME nearbyintf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME nearbyintf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_nearbyintf, __nearbyintf, + IFUNC_SELECTOR ()); + libm_alias_float (__nearbyint, nearbyint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rint-avx.S b/sysdeps/x86_64/fpu/multiarch/s_rint-avx.S +new file mode 100644 +index 0000000000000000..2b403b331f145221 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_rint-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of rint function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__rint) ++ vroundsd $4, %xmm0, %xmm0, %xmm0 ++ ret ++END(__rint) ++ ++libm_alias_double (__rint, rint) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rint-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_rint-sse4_1.S +index 405372991bfac3d0..8cd9cf759f3ac70c 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_rint-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_rint-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __rint_sse41 __rint ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__rint_sse41) + roundsd $4, %xmm0, %xmm0 + ret + END(__rint_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__rint, rint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rint.c b/sysdeps/x86_64/fpu/multiarch/s_rint.c +index 754c87e004f4b006..18623b7d99d1a26e 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_rint.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_rint.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define rint __redirect_rint +-#define __rint __redirect___rint +-#include +-#undef rint +-#undef __rint ++# define rint __redirect_rint ++# define __rint __redirect___rint ++# include ++# undef rint ++# undef __rint + +-#define SYMBOL_NAME rint +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME rint ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_rint, __rint, IFUNC_SELECTOR ()); + libm_alias_double (__rint, rint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rintf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_rintf-avx.S +new file mode 100644 +index 0000000000000000..171c2867f4ab4e3c +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_rintf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of rintf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__rintf) ++ vroundss $4, %xmm0, %xmm0, %xmm0 ++ ret ++END(__rintf) ++ ++libm_alias_float (__rint, rint) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rintf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_rintf-sse4_1.S +index 8ac67ce7673d2aca..fc1e70f0c9a068e4 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_rintf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_rintf-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __rintf_sse41 __rintf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__rintf_sse41) + roundss $4, %xmm0, %xmm0 + ret + END(__rintf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__rint, rint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_rintf.c b/sysdeps/x86_64/fpu/multiarch/s_rintf.c +index e9d6b7a5f2080b8b..e275368decba8359 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_rintf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_rintf.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define rintf __redirect_rintf +-#define __rintf __redirect___rintf +-#include +-#undef rintf +-#undef __rintf ++# define rintf __redirect_rintf ++# define __rintf __redirect___rintf ++# include ++# undef rintf ++# undef __rintf + +-#define SYMBOL_NAME rintf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME rintf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_rintf, __rintf, IFUNC_SELECTOR ()); + libm_alias_float (__rint, rint) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundeven-avx.S b/sysdeps/x86_64/fpu/multiarch/s_roundeven-avx.S +new file mode 100644 +index 0000000000000000..576790355c6b493f +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundeven-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of roundeven function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__roundeven) ++ vroundsd $8, %xmm0, %xmm0, %xmm0 ++ ret ++END(__roundeven) ++ ++libm_alias_double (__roundeven, roundeven) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundeven-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_roundeven-sse4_1.S +index 5ef102336bfa9e2b..f00be56c592a614d 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_roundeven-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundeven-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __roundeven_sse41 __roundeven ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__roundeven_sse41) + roundsd $8, %xmm0, %xmm0 + ret + END(__roundeven_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__roundeven, roundeven) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundeven.c b/sysdeps/x86_64/fpu/multiarch/s_roundeven.c +index 8737b32e2604290a..139aad088f59c957 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_roundeven.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundeven.c +@@ -16,16 +16,19 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# include + +-#define roundeven __redirect_roundeven +-#define __roundeven __redirect___roundeven +-#include +-#undef roundeven +-#undef __roundeven ++# define roundeven __redirect_roundeven ++# define __roundeven __redirect___roundeven ++# include ++# undef roundeven ++# undef __roundeven + +-#define SYMBOL_NAME roundeven +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME roundeven ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_roundeven, __roundeven, IFUNC_SELECTOR ()); + libm_alias_double (__roundeven, roundeven) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundevenf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_roundevenf-avx.S +new file mode 100644 +index 0000000000000000..42c359f4cd04fc56 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundevenf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of roundevenf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__roundevenf) ++ vroundss $8, %xmm0, %xmm0, %xmm0 ++ ret ++END(__roundevenf) ++ ++libm_alias_float (__roundeven, roundeven) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundevenf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_roundevenf-sse4_1.S +index 792c90ba07a3f985..6b148e435316816f 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_roundevenf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundevenf-sse4_1.S +@@ -17,8 +17,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __roundevenf_sse41 __roundevenf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__roundevenf_sse41) + roundss $8, %xmm0, %xmm0 + ret + END(__roundevenf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__roundeven, roundeven) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_roundevenf.c b/sysdeps/x86_64/fpu/multiarch/s_roundevenf.c +index e96016a4d5ba7b39..2fb090075d328ae8 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_roundevenf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_roundevenf.c +@@ -16,16 +16,19 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# include + +-#define roundevenf __redirect_roundevenf +-#define __roundevenf __redirect___roundevenf +-#include +-#undef roundevenf +-#undef __roundevenf ++# define roundevenf __redirect_roundevenf ++# define __roundevenf __redirect___roundevenf ++# include ++# undef roundevenf ++# undef __roundevenf + +-#define SYMBOL_NAME roundevenf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME roundevenf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_roundevenf, __roundevenf, IFUNC_SELECTOR ()); + libm_alias_float (__roundeven, roundeven) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_sin.c b/sysdeps/x86_64/fpu/multiarch/s_sin.c +index 355cc0092eda46a6..21e77943a3662cd6 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_sin.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_sin.c +@@ -16,24 +16,27 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_sin (double); + extern double __redirect_cos (double); + +-#define SYMBOL_NAME sin +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME sin ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_sin, __sin, IFUNC_SELECTOR ()); + libm_alias_double (__sin, sin) + +-#undef SYMBOL_NAME +-#define SYMBOL_NAME cos +-#include "ifunc-avx-fma4.h" ++# undef SYMBOL_NAME ++# define SYMBOL_NAME cos ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_cos, __cos, IFUNC_SELECTOR ()); + libm_alias_double (__cos, cos) + +-#define __cos __cos_sse2 +-#define __sin __sin_sse2 ++# define __cos __cos_sse2 ++# define __sin __sin_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_sincos.c b/sysdeps/x86_64/fpu/multiarch/s_sincos.c +index 70107e999c39da91..b35757f8de4f9d2d 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_sincos.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_sincos.c +@@ -16,15 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern void __redirect_sincos (double, double *, double *); + +-#define SYMBOL_NAME sincos +-#include "ifunc-fma4.h" ++# define SYMBOL_NAME sincos ++# include "ifunc-fma4.h" + + libc_ifunc_redirected (__redirect_sincos, __sincos, IFUNC_SELECTOR ()); + libm_alias_double (__sincos, sincos) + +-#define __sincos __sincos_sse2 ++# define __sincos __sincos_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_sincosf.c b/sysdeps/x86_64/fpu/multiarch/s_sincosf.c +index 80bc028451e5153c..0ea9b40e844a2b22 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_sincosf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_sincosf.c +@@ -16,13 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern void __redirect_sincosf (float, float *, float *); + +-#define SYMBOL_NAME sincosf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME sincosf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_sincosf, __sincosf, IFUNC_SELECTOR ()); + + libm_alias_float (__sincos, sincos) ++#else ++# include ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_sinf.c b/sysdeps/x86_64/fpu/multiarch/s_sinf.c +index a32b9e955052c296..c61624e3eed441dd 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_sinf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_sinf.c +@@ -16,13 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern float __redirect_sinf (float); + +-#define SYMBOL_NAME sinf +-#include "ifunc-fma.h" ++# define SYMBOL_NAME sinf ++# include "ifunc-fma.h" + + libc_ifunc_redirected (__redirect_sinf, __sinf, IFUNC_SELECTOR ()); + + libm_alias_float (__sin, sin) ++#else ++# include ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_tan.c b/sysdeps/x86_64/fpu/multiarch/s_tan.c +index f9a2474a139b0807..125d992ba1fb01ea 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_tan.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_tan.c +@@ -16,15 +16,18 @@ + License along with the GNU C Library; if not, see + . */ + +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < AVX2_X86_ISA_LEVEL ++# include + + extern double __redirect_tan (double); + +-#define SYMBOL_NAME tan +-#include "ifunc-avx-fma4.h" ++# define SYMBOL_NAME tan ++# include "ifunc-avx-fma4.h" + + libc_ifunc_redirected (__redirect_tan, __tan, IFUNC_SELECTOR ()); + libm_alias_double (__tan, tan) + +-#define __tan __tan_sse2 ++# define __tan __tan_sse2 ++#endif + #include +diff --git a/sysdeps/x86_64/fpu/multiarch/s_trunc-avx.S b/sysdeps/x86_64/fpu/multiarch/s_trunc-avx.S +new file mode 100644 +index 0000000000000000..b3e87e96068e5404 +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_trunc-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of trunc function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__trunc) ++ vroundsd $11, %xmm0, %xmm0, %xmm0 ++ ret ++END(__trunc) ++ ++libm_alias_double (__trunc, trunc) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_trunc-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_trunc-sse4_1.S +index b496a6ef49ac5d1c..2b79174eeda798f8 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_trunc-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_trunc-sse4_1.S +@@ -18,8 +18,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __trunc_sse41 __trunc ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__trunc_sse41) + roundsd $11, %xmm0, %xmm0 + ret + END(__trunc_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_double (__trunc, trunc) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_trunc.c b/sysdeps/x86_64/fpu/multiarch/s_trunc.c +index 9bc9df8744dd56b0..ea89c4f85d9c9dbd 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_trunc.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_trunc.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define trunc __redirect_trunc +-#define __trunc __redirect___trunc +-#include +-#undef trunc +-#undef __trunc ++# define trunc __redirect_trunc ++# define __trunc __redirect___trunc ++# include ++# undef trunc ++# undef __trunc + +-#define SYMBOL_NAME trunc +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME trunc ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_trunc, __trunc, IFUNC_SELECTOR ()); + libm_alias_double (__trunc, trunc) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_truncf-avx.S b/sysdeps/x86_64/fpu/multiarch/s_truncf-avx.S +new file mode 100644 +index 0000000000000000..f31ac7d7f7980f2e +--- /dev/null ++++ b/sysdeps/x86_64/fpu/multiarch/s_truncf-avx.S +@@ -0,0 +1,28 @@ ++/* AVX implementation of truncf function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++ .text ++ENTRY(__truncf) ++ vroundss $11, %xmm0, %xmm0, %xmm0 ++ ret ++END(__truncf) ++ ++libm_alias_float (__trunc, trunc) +diff --git a/sysdeps/x86_64/fpu/multiarch/s_truncf-sse4_1.S b/sysdeps/x86_64/fpu/multiarch/s_truncf-sse4_1.S +index 22e9a8330746da2f..60498b2cb21a97fa 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_truncf-sse4_1.S ++++ b/sysdeps/x86_64/fpu/multiarch/s_truncf-sse4_1.S +@@ -18,8 +18,20 @@ + + #include + ++#include ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++# include ++# define __truncf_sse41 __truncf ++ .text ++#else + .section .text.sse4.1,"ax",@progbits ++#endif ++ + ENTRY(__truncf_sse41) + roundss $11, %xmm0, %xmm0 + ret + END(__truncf_sse41) ++ ++#if MINIMUM_X86_ISA_LEVEL == SSE4_1_X86_ISA_LEVEL ++libm_alias_float (__trunc, trunc) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/s_truncf.c b/sysdeps/x86_64/fpu/multiarch/s_truncf.c +index dae01d166a0dfa13..92435ce39dbd7cca 100644 +--- a/sysdeps/x86_64/fpu/multiarch/s_truncf.c ++++ b/sysdeps/x86_64/fpu/multiarch/s_truncf.c +@@ -16,17 +16,20 @@ + License along with the GNU C Library; if not, see + . */ + +-#define NO_MATH_REDIRECT +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL < SSE4_1_X86_ISA_LEVEL ++# define NO_MATH_REDIRECT ++# include + +-#define truncf __redirect_truncf +-#define __truncf __redirect___truncf +-#include +-#undef truncf +-#undef __truncf ++# define truncf __redirect_truncf ++# define __truncf __redirect___truncf ++# include ++# undef truncf ++# undef __truncf + +-#define SYMBOL_NAME truncf +-#include "ifunc-sse4_1.h" ++# define SYMBOL_NAME truncf ++# include "ifunc-sse4_1.h" + + libc_ifunc_redirected (__redirect_truncf, __truncf, IFUNC_SELECTOR ()); + libm_alias_float (__trunc, trunc) ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/w_exp.c b/sysdeps/x86_64/fpu/multiarch/w_exp.c +index 27eee98a0a3e16f0..3584187e0e223195 100644 +--- a/sysdeps/x86_64/fpu/multiarch/w_exp.c ++++ b/sysdeps/x86_64/fpu/multiarch/w_exp.c +@@ -1 +1,6 @@ +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL >= AVX2_X86_ISA_LEVEL ++# include ++#else ++# include ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/w_log.c b/sysdeps/x86_64/fpu/multiarch/w_log.c +index 9b2b0187116bc48e..414ca3ca3ddf530a 100644 +--- a/sysdeps/x86_64/fpu/multiarch/w_log.c ++++ b/sysdeps/x86_64/fpu/multiarch/w_log.c +@@ -1 +1,6 @@ +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL >= AVX2_X86_ISA_LEVEL ++# include ++#else ++# include ++#endif +diff --git a/sysdeps/x86_64/fpu/multiarch/w_pow.c b/sysdeps/x86_64/fpu/multiarch/w_pow.c +index b50c1988de597731..d5fcc4f871bd7eea 100644 +--- a/sysdeps/x86_64/fpu/multiarch/w_pow.c ++++ b/sysdeps/x86_64/fpu/multiarch/w_pow.c +@@ -1 +1,6 @@ +-#include ++#include ++#if MINIMUM_X86_ISA_LEVEL >= AVX2_X86_ISA_LEVEL ++# include ++#else ++# include ++#endif diff --git a/SOURCES/glibc-upstream-2.39-31.patch b/SOURCES/glibc-upstream-2.39-31.patch new file mode 100644 index 0000000..bf21203 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-31.patch @@ -0,0 +1,207 @@ +commit 31da30f23cddd36db29d5b6a1c7619361b271fb4 +Author: Charles Fol +Date: Thu Mar 28 12:25:38 2024 -0300 + + iconv: ISO-2022-CN-EXT: fix out-of-bound writes when writing escape sequence (CVE-2024-2961) + + ISO-2022-CN-EXT uses escape sequences to indicate character set changes + (as specified by RFC 1922). While the SOdesignation has the expected + bounds checks, neither SS2designation nor SS3designation have its; + allowing a write overflow of 1, 2, or 3 bytes with fixed values: + '$+I', '$+J', '$+K', '$+L', '$+M', or '$*H'. + + Checked on aarch64-linux-gnu. + + Co-authored-by: Adhemerval Zanella + Reviewed-by: Carlos O'Donell + Tested-by: Carlos O'Donell + + (cherry picked from commit f9dc609e06b1136bb0408be9605ce7973a767ada) + +diff --git a/iconvdata/Makefile b/iconvdata/Makefile +index ea019ce5c0e67e98..7196a8744bb66e8c 100644 +--- a/iconvdata/Makefile ++++ b/iconvdata/Makefile +@@ -75,7 +75,8 @@ ifeq (yes,$(build-shared)) + tests = bug-iconv1 bug-iconv2 tst-loading tst-e2big tst-iconv4 bug-iconv4 \ + tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7 bug-iconv8 bug-iconv9 \ + bug-iconv10 bug-iconv11 bug-iconv12 tst-iconv-big5-hkscs-to-2ucs4 \ +- bug-iconv13 bug-iconv14 bug-iconv15 ++ bug-iconv13 bug-iconv14 bug-iconv15 \ ++ tst-iconv-iso-2022-cn-ext + ifeq ($(have-thread-library),yes) + tests += bug-iconv3 + endif +@@ -330,6 +331,8 @@ $(objpfx)bug-iconv14.out: $(addprefix $(objpfx), $(gconv-modules)) \ + $(addprefix $(objpfx),$(modules.so)) + $(objpfx)bug-iconv15.out: $(addprefix $(objpfx), $(gconv-modules)) \ + $(addprefix $(objpfx),$(modules.so)) ++$(objpfx)tst-iconv-iso-2022-cn-ext.out: $(addprefix $(objpfx), $(gconv-modules)) \ ++ $(addprefix $(objpfx),$(modules.so)) + + $(objpfx)iconv-test.out: run-iconv-test.sh \ + $(addprefix $(objpfx), $(gconv-modules)) \ +diff --git a/iconvdata/iso-2022-cn-ext.c b/iconvdata/iso-2022-cn-ext.c +index b34c8a36f4564c11..cce29b19692263d6 100644 +--- a/iconvdata/iso-2022-cn-ext.c ++++ b/iconvdata/iso-2022-cn-ext.c +@@ -574,6 +574,12 @@ DIAG_IGNORE_Os_NEEDS_COMMENT (5, "-Wmaybe-uninitialized"); + { \ + const char *escseq; \ + \ ++ if (outptr + 4 > outend) \ ++ { \ ++ result = __GCONV_FULL_OUTPUT; \ ++ break; \ ++ } \ ++ \ + assert (used == CNS11643_2_set); /* XXX */ \ + escseq = "*H"; \ + *outptr++ = ESC; \ +@@ -587,6 +593,12 @@ DIAG_IGNORE_Os_NEEDS_COMMENT (5, "-Wmaybe-uninitialized"); + { \ + const char *escseq; \ + \ ++ if (outptr + 4 > outend) \ ++ { \ ++ result = __GCONV_FULL_OUTPUT; \ ++ break; \ ++ } \ ++ \ + assert ((used >> 5) >= 3 && (used >> 5) <= 7); \ + escseq = "+I+J+K+L+M" + ((used >> 5) - 3) * 2; \ + *outptr++ = ESC; \ +diff --git a/iconvdata/tst-iconv-iso-2022-cn-ext.c b/iconvdata/tst-iconv-iso-2022-cn-ext.c +new file mode 100644 +index 0000000000000000..96a8765fd5369681 +--- /dev/null ++++ b/iconvdata/tst-iconv-iso-2022-cn-ext.c +@@ -0,0 +1,128 @@ ++/* Verify ISO-2022-CN-EXT does not write out of the bounds. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++ ++#include ++#include ++#include ++ ++#include ++#include ++#include ++ ++/* The test sets up a two memory page buffer with the second page marked ++ PROT_NONE to trigger a fault if the conversion writes beyond the exact ++ expected amount. Then we carry out various conversions and precisely ++ place the start of the output buffer in order to trigger a SIGSEGV if the ++ process writes anywhere between 1 and page sized bytes more (only one ++ PROT_NONE page is setup as a canary) than expected. These tests exercise ++ all three of the cases in ISO-2022-CN-EXT where the converter must switch ++ character sets and may run out of buffer space while doing the ++ operation. */ ++ ++static int ++do_test (void) ++{ ++ iconv_t cd = iconv_open ("ISO-2022-CN-EXT", "UTF-8"); ++ TEST_VERIFY_EXIT (cd != (iconv_t) -1); ++ ++ char *ntf; ++ size_t ntfsize; ++ char *outbufbase; ++ { ++ int pgz = getpagesize (); ++ TEST_VERIFY_EXIT (pgz > 0); ++ ntfsize = 2 * pgz; ++ ++ ntf = xmmap (NULL, ntfsize, PROT_READ | PROT_WRITE, MAP_PRIVATE ++ | MAP_ANONYMOUS, -1); ++ xmprotect (ntf + pgz, pgz, PROT_NONE); ++ ++ outbufbase = ntf + pgz; ++ } ++ ++ /* Check if SOdesignation escape sequence does not trigger an OOB write. */ ++ { ++ char inbuf[] = "\xe4\xba\xa4\xe6\x8d\xa2"; ++ ++ for (int i = 0; i < 9; i++) ++ { ++ char *inp = inbuf; ++ size_t inleft = sizeof (inbuf) - 1; ++ ++ char *outp = outbufbase - i; ++ size_t outleft = i; ++ ++ TEST_VERIFY_EXIT (iconv (cd, &inp, &inleft, &outp, &outleft) ++ == (size_t) -1); ++ TEST_COMPARE (errno, E2BIG); ++ ++ TEST_VERIFY_EXIT (iconv (cd, NULL, NULL, NULL, NULL) == 0); ++ } ++ } ++ ++ /* Same as before for SS2designation. */ ++ { ++ char inbuf[] = "㴽 \xe3\xb4\xbd"; ++ ++ for (int i = 0; i < 14; i++) ++ { ++ char *inp = inbuf; ++ size_t inleft = sizeof (inbuf) - 1; ++ ++ char *outp = outbufbase - i; ++ size_t outleft = i; ++ ++ TEST_VERIFY_EXIT (iconv (cd, &inp, &inleft, &outp, &outleft) ++ == (size_t) -1); ++ TEST_COMPARE (errno, E2BIG); ++ ++ TEST_VERIFY_EXIT (iconv (cd, NULL, NULL, NULL, NULL) == 0); ++ } ++ } ++ ++ /* Same as before for SS3designation. */ ++ { ++ char inbuf[] = "劄 \xe5\x8a\x84"; ++ ++ for (int i = 0; i < 14; i++) ++ { ++ char *inp = inbuf; ++ size_t inleft = sizeof (inbuf) - 1; ++ ++ char *outp = outbufbase - i; ++ size_t outleft = i; ++ ++ TEST_VERIFY_EXIT (iconv (cd, &inp, &inleft, &outp, &outleft) ++ == (size_t) -1); ++ TEST_COMPARE (errno, E2BIG); ++ ++ TEST_VERIFY_EXIT (iconv (cd, NULL, NULL, NULL, NULL) == 0); ++ } ++ } ++ ++ TEST_VERIFY_EXIT (iconv_close (cd) != -1); ++ ++ xmunmap (ntf, ntfsize); ++ ++ return 0; ++} ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-32.patch b/SOURCES/glibc-upstream-2.39-32.patch new file mode 100644 index 0000000..f5cf970 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-32.patch @@ -0,0 +1,49 @@ +commit e828914cf9f2fc2caa5bced0fc6a03cb78324979 +Author: Florian Weimer +Date: Tue Apr 23 21:16:32 2024 +0200 + + nptl: Fix tst-cancel30 on kernels without ppoll_time64 support + + Fall back to ppoll if ppoll_time64 fails with ENOSYS. + Fixes commit 370da8a121c3ba9eeb2f13da15fc0f21f4136b25 ("nptl: Fix + tst-cancel30 on sparc64"). + + Reviewed-by: Adhemerval Zanella + (cherry picked from commit f4724843ada64a51d66f65d3199fe431f9d4c254) + +diff --git a/sysdeps/pthread/tst-cancel30.c b/sysdeps/pthread/tst-cancel30.c +index 3030660e5fd93e2c..94ad6281bcf080f4 100644 +--- a/sysdeps/pthread/tst-cancel30.c ++++ b/sysdeps/pthread/tst-cancel30.c +@@ -18,6 +18,7 @@ + License along with the GNU C Library; if not, see + . */ + ++#include + #include + #include + #include +@@ -46,13 +47,19 @@ tf (void *arg) + + /* Wait indefinitely for cancellation, which only works if asynchronous + cancellation is enabled. */ +-#if defined SYS_ppoll || defined SYS_ppoll_time64 +-# ifndef SYS_ppoll_time64 +-# define SYS_ppoll_time64 SYS_ppoll ++#ifdef SYS_ppoll_time64 ++ long int ret = syscall (SYS_ppoll_time64, NULL, 0, NULL, NULL); ++ (void) ret; ++# ifdef SYS_ppoll ++ if (ret == -1 && errno == ENOSYS) ++ syscall (SYS_ppoll, NULL, 0, NULL, NULL); + # endif +- syscall (SYS_ppoll_time64, NULL, 0, NULL, NULL); + #else ++# ifdef SYS_ppoll ++ syscall (SYS_ppoll, NULL, 0, NULL, NULL); ++# else + for (;;); ++# endif + #endif + + return 0; diff --git a/SOURCES/glibc-upstream-2.39-33.patch b/SOURCES/glibc-upstream-2.39-33.patch new file mode 100644 index 0000000..a13dbfa --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-33.patch @@ -0,0 +1,20 @@ +commit e701c7d761f6e5c48d8e9dd5da88cbe2e94943f4 +Author: Florian Weimer +Date: Thu Apr 25 12:56:48 2024 +0200 + + i386: ulp update for SSE2 --disable-multi-arch configurations + + (cherry picked from commit 3a3a4497421422aa854c855cbe5110ca7d598ffc) + +diff --git a/sysdeps/i386/fpu/libm-test-ulps b/sysdeps/i386/fpu/libm-test-ulps +index 84e6686eba5fe79a..f2139fc172ceef7b 100644 +--- a/sysdeps/i386/fpu/libm-test-ulps ++++ b/sysdeps/i386/fpu/libm-test-ulps +@@ -1232,6 +1232,7 @@ ldouble: 6 + + Function: "hypot": + double: 1 ++float: 1 + float128: 1 + ldouble: 1 + diff --git a/SOURCES/glibc-upstream-2.39-34.patch b/SOURCES/glibc-upstream-2.39-34.patch new file mode 100644 index 0000000..961fedd --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-34.patch @@ -0,0 +1,89 @@ +commit 2f8f157eb0cc7f1d8d9a3fcaa8c55bed53b092a8 +Author: H.J. Lu +Date: Tue Apr 23 13:59:50 2024 -0700 + + x86: Define MINIMUM_X86_ISA_LEVEL in config.h [BZ #31676] + + Define MINIMUM_X86_ISA_LEVEL at configure time to avoid + + /usr/bin/ld: …/build/elf/librtld.os: in function `init_cpu_features': + …/git/elf/../sysdeps/x86/cpu-features.c:1202: undefined reference to `_dl_runtime_resolve_fxsave' + /usr/bin/ld: …/build/elf/librtld.os: relocation R_X86_64_PC32 against undefined hidden symbol `_dl_runtime_resolve_fxsave' can not be used when making a shared object + /usr/bin/ld: final link failed: bad value + collect2: error: ld returned 1 exit status + + when glibc is built with -march=x86-64-v3 and configured with + --with-rtld-early-cflags=-march=x86-64, which is used to allow ld.so to + print an error message on unsupported CPUs: + + Fatal glibc error: CPU does not support x86-64-v3 + + This fixes BZ #31676. + Reviewed-by: Sunil K Pandey + + (cherry picked from commit 46c999741340ea559784c20a45077955b50aca43) + +diff --git a/config.h.in b/config.h.in +index 4d33c63a841d3d6d..1e647de58580bc2d 100644 +--- a/config.h.in ++++ b/config.h.in +@@ -286,6 +286,9 @@ + /* Define if x86 ISA level should be included in shared libraries. */ + #undef INCLUDE_X86_ISA_LEVEL + ++/* The x86 ISA level. 1 for baseline. Undefined on non-x86. */ ++#undef MINIMUM_X86_ISA_LEVEL ++ + /* Define if -msahf is enabled by default on x86. */ + #undef HAVE_X86_LAHF_SAHF + +diff --git a/sysdeps/x86/configure b/sysdeps/x86/configure +index 2a5421bb31e9efe4..d28d9bcb296c6380 100644 +--- a/sysdeps/x86/configure ++++ b/sysdeps/x86/configure +@@ -151,6 +151,13 @@ printf "%s\n" "$libc_cv_have_x86_isa_level" >&6; } + else + libc_cv_have_x86_isa_level=baseline + fi ++if test $libc_cv_have_x86_isa_level = baseline; then ++ printf "%s\n" "#define MINIMUM_X86_ISA_LEVEL 1" >>confdefs.h ++ ++else ++ printf "%s\n" "#define MINIMUM_X86_ISA_LEVEL $libc_cv_have_x86_isa_level" >>confdefs.h ++ ++fi + config_vars="$config_vars + have-x86-isa-level = $libc_cv_have_x86_isa_level" + config_vars="$config_vars +diff --git a/sysdeps/x86/configure.ac b/sysdeps/x86/configure.ac +index 78ff7c8f41c552bc..5b0acd03d2a30c9b 100644 +--- a/sysdeps/x86/configure.ac ++++ b/sysdeps/x86/configure.ac +@@ -105,6 +105,11 @@ EOF + else + libc_cv_have_x86_isa_level=baseline + fi ++if test $libc_cv_have_x86_isa_level = baseline; then ++ AC_DEFINE_UNQUOTED(MINIMUM_X86_ISA_LEVEL, 1) ++else ++ AC_DEFINE_UNQUOTED(MINIMUM_X86_ISA_LEVEL, $libc_cv_have_x86_isa_level) ++fi + LIBC_CONFIG_VAR([have-x86-isa-level], [$libc_cv_have_x86_isa_level]) + LIBC_CONFIG_VAR([x86-isa-level-3-or-above], [3 4]) + LIBC_CONFIG_VAR([enable-x86-isa-level], [$libc_cv_include_x86_isa_level]) +diff --git a/sysdeps/x86/isa-level.h b/sysdeps/x86/isa-level.h +index 11fe1ca90c5bfedf..2c7f74212b9a27e5 100644 +--- a/sysdeps/x86/isa-level.h ++++ b/sysdeps/x86/isa-level.h +@@ -61,8 +61,10 @@ + # define __X86_ISA_V4 0 + #endif + +-#define MINIMUM_X86_ISA_LEVEL \ ++#ifndef MINIMUM_X86_ISA_LEVEL ++# define MINIMUM_X86_ISA_LEVEL \ + (__X86_ISA_V1 + __X86_ISA_V2 + __X86_ISA_V3 + __X86_ISA_V4) ++#endif + + /* Depending on the minimum ISA level, a feature check result can be a + compile-time constant.. */ diff --git a/SOURCES/glibc-upstream-2.39-35.patch b/SOURCES/glibc-upstream-2.39-35.patch new file mode 100644 index 0000000..41555b4 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-35.patch @@ -0,0 +1,32 @@ +commit 1263d583d2e28afb8be53f8d6922f0842036f35d +Author: Florian Weimer +Date: Thu Apr 25 15:00:45 2024 +0200 + + CVE-2024-33599: nscd: Stack-based buffer overflow in netgroup cache (bug 31677) + + Using alloca matches what other caches do. The request length is + bounded by MAXKEYLEN. + + Reviewed-by: Carlos O'Donell + (cherry picked from commit 87801a8fd06db1d654eea3e4f7626ff476a9bdaa) + +diff --git a/nscd/netgroupcache.c b/nscd/netgroupcache.c +index 0c6e46f15c5d7139..f227dc7fa2856e38 100644 +--- a/nscd/netgroupcache.c ++++ b/nscd/netgroupcache.c +@@ -502,12 +502,13 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + = (struct indataset *) mempool_alloc (db, + sizeof (*dataset) + req->key_len, + 1); +- struct indataset dataset_mem; + bool cacheable = true; + if (__glibc_unlikely (dataset == NULL)) + { + cacheable = false; +- dataset = &dataset_mem; ++ /* The alloca is safe because nscd_run_worker verfies that ++ key_len is not larger than MAXKEYLEN. */ ++ dataset = alloca (sizeof (*dataset) + req->key_len); + } + + datahead_init_pos (&dataset->head, sizeof (*dataset) + req->key_len, diff --git a/SOURCES/glibc-upstream-2.39-36.patch b/SOURCES/glibc-upstream-2.39-36.patch new file mode 100644 index 0000000..2851845 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-36.patch @@ -0,0 +1,53 @@ +commit 5a508e0b508c8ad53bd0d2fb48fd71b242626341 +Author: Florian Weimer +Date: Thu Apr 25 15:01:07 2024 +0200 + + CVE-2024-33600: nscd: Do not send missing not-found response in addgetnetgrentX (bug 31678) + + If we failed to add a not-found response to the cache, the dataset + point can be null, resulting in a null pointer dereference. + + Reviewed-by: Siddhesh Poyarekar + (cherry picked from commit 7835b00dbce53c3c87bbbb1754a95fb5e58187aa) + +diff --git a/nscd/netgroupcache.c b/nscd/netgroupcache.c +index f227dc7fa2856e38..c18fe111f37496d7 100644 +--- a/nscd/netgroupcache.c ++++ b/nscd/netgroupcache.c +@@ -147,7 +147,7 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + /* No such service. */ + cacheable = do_notfound (db, fd, req, key, &dataset, &total, &timeout, + &key_copy); +- goto writeout; ++ goto maybe_cache_add; + } + + memset (&data, '\0', sizeof (data)); +@@ -348,7 +348,7 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + { + cacheable = do_notfound (db, fd, req, key, &dataset, &total, &timeout, + &key_copy); +- goto writeout; ++ goto maybe_cache_add; + } + + total = buffilled; +@@ -410,14 +410,12 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + } + + if (he == NULL && fd != -1) +- { +- /* We write the dataset before inserting it to the database +- since while inserting this thread might block and so would +- unnecessarily let the receiver wait. */ +- writeout: ++ /* We write the dataset before inserting it to the database since ++ while inserting this thread might block and so would ++ unnecessarily let the receiver wait. */ + writeall (fd, &dataset->resp, dataset->head.recsize); +- } + ++ maybe_cache_add: + if (cacheable) + { + /* If necessary, we also propagate the data to disk. */ diff --git a/SOURCES/glibc-upstream-2.39-37.patch b/SOURCES/glibc-upstream-2.39-37.patch new file mode 100644 index 0000000..b0a31a8 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-37.patch @@ -0,0 +1,54 @@ +commit c99f886de54446cd4447db6b44be93dabbdc2f8b +Author: Florian Weimer +Date: Thu Apr 25 15:01:07 2024 +0200 + + CVE-2024-33600: nscd: Avoid null pointer crashes after notfound response (bug 31678) + + The addgetnetgrentX call in addinnetgrX may have failed to produce + a result, so the result variable in addinnetgrX can be NULL. + Use db->negtimeout as the fallback value if there is no result data; + the timeout is also overwritten below. + + Also avoid sending a second not-found response. (The client + disconnects after receiving the first response, so the data stream did + not go out of sync even without this fix.) It is still beneficial to + add the negative response to the mapping, so that the client can get + it from there in the future, instead of going through the socket. + + Reviewed-by: Siddhesh Poyarekar + (cherry picked from commit b048a482f088e53144d26a61c390bed0210f49f2) + +diff --git a/nscd/netgroupcache.c b/nscd/netgroupcache.c +index c18fe111f37496d7..e22ffa5884e36260 100644 +--- a/nscd/netgroupcache.c ++++ b/nscd/netgroupcache.c +@@ -511,14 +511,15 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + + datahead_init_pos (&dataset->head, sizeof (*dataset) + req->key_len, + sizeof (innetgroup_response_header), +- he == NULL ? 0 : dh->nreloads + 1, result->head.ttl); ++ he == NULL ? 0 : dh->nreloads + 1, ++ result == NULL ? db->negtimeout : result->head.ttl); + /* Set the notfound status and timeout based on the result from + getnetgrent. */ +- dataset->head.notfound = result->head.notfound; ++ dataset->head.notfound = result == NULL || result->head.notfound; + dataset->head.timeout = timeout; + + dataset->resp.version = NSCD_VERSION; +- dataset->resp.found = result->resp.found; ++ dataset->resp.found = result != NULL && result->resp.found; + /* Until we find a matching entry the result is 0. */ + dataset->resp.result = 0; + +@@ -566,7 +567,9 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + goto out; + } + +- if (he == NULL) ++ /* addgetnetgrentX may have already sent a notfound response. Do ++ not send another one. */ ++ if (he == NULL && dataset->resp.found) + { + /* We write the dataset before inserting it to the database + since while inserting this thread might block and so would diff --git a/SOURCES/glibc-upstream-2.39-38.patch b/SOURCES/glibc-upstream-2.39-38.patch new file mode 100644 index 0000000..5bda08b --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-38.patch @@ -0,0 +1,384 @@ +commit a9a8d3eebb145779a18d90e3966009a1daa63cd8 +Author: Florian Weimer +Date: Thu Apr 25 15:01:07 2024 +0200 + + CVE-2024-33601, CVE-2024-33602: nscd: netgroup: Use two buffers in addgetnetgrentX (bug 31680) + + This avoids potential memory corruption when the underlying NSS + callback function does not use the buffer space to store all strings + (e.g., for constant strings). + + Instead of custom buffer management, two scratch buffers are used. + This increases stack usage somewhat. + + Scratch buffer allocation failure is handled by return -1 + (an invalid timeout value) instead of terminating the process. + This fixes bug 31679. + + Reviewed-by: Siddhesh Poyarekar + (cherry picked from commit c04a21e050d64a1193a6daab872bca2528bda44b) + +diff --git a/nscd/netgroupcache.c b/nscd/netgroupcache.c +index e22ffa5884e36260..e8fe041846b75cb9 100644 +--- a/nscd/netgroupcache.c ++++ b/nscd/netgroupcache.c +@@ -23,6 +23,7 @@ + #include + #include + #include ++#include + + #include "../nss/netgroup.h" + #include "nscd.h" +@@ -65,6 +66,16 @@ struct dataset + char strdata[0]; + }; + ++/* Send a notfound response to FD. Always returns -1 to indicate an ++ ephemeral error. */ ++static time_t ++send_notfound (int fd) ++{ ++ if (fd != -1) ++ TEMP_FAILURE_RETRY (send (fd, ¬found, sizeof (notfound), MSG_NOSIGNAL)); ++ return -1; ++} ++ + /* Sends a notfound message and prepares a notfound dataset to write to the + cache. Returns true if there was enough memory to allocate the dataset and + returns the dataset in DATASETP, total bytes to write in TOTALP and the +@@ -83,8 +94,7 @@ do_notfound (struct database_dyn *db, int fd, request_header *req, + total = sizeof (notfound); + timeout = time (NULL) + db->negtimeout; + +- if (fd != -1) +- TEMP_FAILURE_RETRY (send (fd, ¬found, total, MSG_NOSIGNAL)); ++ send_notfound (fd); + + dataset = mempool_alloc (db, sizeof (struct dataset) + req->key_len, 1); + /* If we cannot permanently store the result, so be it. */ +@@ -109,11 +119,78 @@ do_notfound (struct database_dyn *db, int fd, request_header *req, + return cacheable; + } + ++struct addgetnetgrentX_scratch ++{ ++ /* This is the result that the caller should use. It can be NULL, ++ point into buffer, or it can be in the cache. */ ++ struct dataset *dataset; ++ ++ struct scratch_buffer buffer; ++ ++ /* Used internally in addgetnetgrentX as a staging area. */ ++ struct scratch_buffer tmp; ++ ++ /* Number of bytes in buffer that are actually used. */ ++ size_t buffer_used; ++}; ++ ++static void ++addgetnetgrentX_scratch_init (struct addgetnetgrentX_scratch *scratch) ++{ ++ scratch->dataset = NULL; ++ scratch_buffer_init (&scratch->buffer); ++ scratch_buffer_init (&scratch->tmp); ++ ++ /* Reserve space for the header. */ ++ scratch->buffer_used = sizeof (struct dataset); ++ static_assert (sizeof (struct dataset) < sizeof (scratch->tmp.__space), ++ "initial buffer space"); ++ memset (scratch->tmp.data, 0, sizeof (struct dataset)); ++} ++ ++static void ++addgetnetgrentX_scratch_free (struct addgetnetgrentX_scratch *scratch) ++{ ++ scratch_buffer_free (&scratch->buffer); ++ scratch_buffer_free (&scratch->tmp); ++} ++ ++/* Copy LENGTH bytes from S into SCRATCH. Returns NULL if SCRATCH ++ could not be resized, otherwise a pointer to the copy. */ ++static char * ++addgetnetgrentX_append_n (struct addgetnetgrentX_scratch *scratch, ++ const char *s, size_t length) ++{ ++ while (true) ++ { ++ size_t remaining = scratch->buffer.length - scratch->buffer_used; ++ if (remaining >= length) ++ break; ++ if (!scratch_buffer_grow_preserve (&scratch->buffer)) ++ return NULL; ++ } ++ char *copy = scratch->buffer.data + scratch->buffer_used; ++ memcpy (copy, s, length); ++ scratch->buffer_used += length; ++ return copy; ++} ++ ++/* Copy S into SCRATCH, including its null terminator. Returns false ++ if SCRATCH could not be resized. */ ++static bool ++addgetnetgrentX_append (struct addgetnetgrentX_scratch *scratch, const char *s) ++{ ++ if (s == NULL) ++ s = ""; ++ return addgetnetgrentX_append_n (scratch, s, strlen (s) + 1) != NULL; ++} ++ ++/* Caller must initialize and free *SCRATCH. If the return value is ++ negative, this function has sent a notfound response. */ + static time_t + addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + const char *key, uid_t uid, struct hashentry *he, +- struct datahead *dh, struct dataset **resultp, +- void **tofreep) ++ struct datahead *dh, struct addgetnetgrentX_scratch *scratch) + { + if (__glibc_unlikely (debug_level > 0)) + { +@@ -132,14 +209,10 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + + char *key_copy = NULL; + struct __netgrent data; +- size_t buflen = MAX (1024, sizeof (*dataset) + req->key_len); +- size_t buffilled = sizeof (*dataset); +- char *buffer = NULL; + size_t nentries = 0; + size_t group_len = strlen (key) + 1; + struct name_list *first_needed + = alloca (sizeof (struct name_list) + group_len); +- *tofreep = NULL; + + if (netgroup_database == NULL + && !__nss_database_get (nss_database_netgroup, &netgroup_database)) +@@ -151,8 +224,6 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + } + + memset (&data, '\0', sizeof (data)); +- buffer = xmalloc (buflen); +- *tofreep = buffer; + first_needed->next = first_needed; + memcpy (first_needed->name, key, group_len); + data.needed_groups = first_needed; +@@ -195,8 +266,8 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + while (1) + { + int e; +- status = getfct.f (&data, buffer + buffilled, +- buflen - buffilled - req->key_len, &e); ++ status = getfct.f (&data, scratch->tmp.data, ++ scratch->tmp.length, &e); + if (status == NSS_STATUS_SUCCESS) + { + if (data.type == triple_val) +@@ -204,68 +275,10 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + const char *nhost = data.val.triple.host; + const char *nuser = data.val.triple.user; + const char *ndomain = data.val.triple.domain; +- +- size_t hostlen = strlen (nhost ?: "") + 1; +- size_t userlen = strlen (nuser ?: "") + 1; +- size_t domainlen = strlen (ndomain ?: "") + 1; +- +- if (nhost == NULL || nuser == NULL || ndomain == NULL +- || nhost > nuser || nuser > ndomain) +- { +- const char *last = nhost; +- if (last == NULL +- || (nuser != NULL && nuser > last)) +- last = nuser; +- if (last == NULL +- || (ndomain != NULL && ndomain > last)) +- last = ndomain; +- +- size_t bufused +- = (last == NULL +- ? buffilled +- : last + strlen (last) + 1 - buffer); +- +- /* We have to make temporary copies. */ +- size_t needed = hostlen + userlen + domainlen; +- +- if (buflen - req->key_len - bufused < needed) +- { +- buflen += MAX (buflen, 2 * needed); +- /* Save offset in the old buffer. We don't +- bother with the NULL check here since +- we'll do that later anyway. */ +- size_t nhostdiff = nhost - buffer; +- size_t nuserdiff = nuser - buffer; +- size_t ndomaindiff = ndomain - buffer; +- +- char *newbuf = xrealloc (buffer, buflen); +- /* Fix up the triplet pointers into the new +- buffer. */ +- nhost = (nhost ? newbuf + nhostdiff +- : NULL); +- nuser = (nuser ? newbuf + nuserdiff +- : NULL); +- ndomain = (ndomain ? newbuf + ndomaindiff +- : NULL); +- *tofreep = buffer = newbuf; +- } +- +- nhost = memcpy (buffer + bufused, +- nhost ?: "", hostlen); +- nuser = memcpy ((char *) nhost + hostlen, +- nuser ?: "", userlen); +- ndomain = memcpy ((char *) nuser + userlen, +- ndomain ?: "", domainlen); +- } +- +- char *wp = buffer + buffilled; +- wp = memmove (wp, nhost ?: "", hostlen); +- wp += hostlen; +- wp = memmove (wp, nuser ?: "", userlen); +- wp += userlen; +- wp = memmove (wp, ndomain ?: "", domainlen); +- wp += domainlen; +- buffilled = wp - buffer; ++ if (!(addgetnetgrentX_append (scratch, nhost) ++ && addgetnetgrentX_append (scratch, nuser) ++ && addgetnetgrentX_append (scratch, ndomain))) ++ return send_notfound (fd); + ++nentries; + } + else +@@ -317,8 +330,8 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + } + else if (status == NSS_STATUS_TRYAGAIN && e == ERANGE) + { +- buflen *= 2; +- *tofreep = buffer = xrealloc (buffer, buflen); ++ if (!scratch_buffer_grow (&scratch->tmp)) ++ return send_notfound (fd); + } + else if (status == NSS_STATUS_RETURN + || status == NSS_STATUS_NOTFOUND +@@ -351,10 +364,17 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + goto maybe_cache_add; + } + +- total = buffilled; ++ /* Capture the result size without the key appended. */ ++ total = scratch->buffer_used; ++ ++ /* Make a copy of the key. The scratch buffer must not move after ++ this point. */ ++ key_copy = addgetnetgrentX_append_n (scratch, key, req->key_len); ++ if (key_copy == NULL) ++ return send_notfound (fd); + + /* Fill in the dataset. */ +- dataset = (struct dataset *) buffer; ++ dataset = scratch->buffer.data; + timeout = datahead_init_pos (&dataset->head, total + req->key_len, + total - offsetof (struct dataset, resp), + he == NULL ? 0 : dh->nreloads + 1, +@@ -363,11 +383,7 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + dataset->resp.version = NSCD_VERSION; + dataset->resp.found = 1; + dataset->resp.nresults = nentries; +- dataset->resp.result_len = buffilled - sizeof (*dataset); +- +- assert (buflen - buffilled >= req->key_len); +- key_copy = memcpy (buffer + buffilled, key, req->key_len); +- buffilled += req->key_len; ++ dataset->resp.result_len = total - sizeof (*dataset); + + /* Now we can determine whether on refill we have to create a new + record or not. */ +@@ -398,7 +414,7 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + if (__glibc_likely (newp != NULL)) + { + /* Adjust pointer into the memory block. */ +- key_copy = (char *) newp + (key_copy - buffer); ++ key_copy = (char *) newp + (key_copy - (char *) dataset); + + dataset = memcpy (newp, dataset, total + req->key_len); + cacheable = true; +@@ -439,7 +455,7 @@ addgetnetgrentX (struct database_dyn *db, int fd, request_header *req, + } + + out: +- *resultp = dataset; ++ scratch->dataset = dataset; + + return timeout; + } +@@ -460,6 +476,9 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + if (user != NULL) + key = strchr (key, '\0') + 1; + const char *domain = *key++ ? key : NULL; ++ struct addgetnetgrentX_scratch scratch; ++ ++ addgetnetgrentX_scratch_init (&scratch); + + if (__glibc_unlikely (debug_level > 0)) + { +@@ -475,12 +494,8 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + group, group_len, + db, uid); + time_t timeout; +- void *tofree; + if (result != NULL) +- { +- timeout = result->head.timeout; +- tofree = NULL; +- } ++ timeout = result->head.timeout; + else + { + request_header req_get = +@@ -489,7 +504,10 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + .key_len = group_len + }; + timeout = addgetnetgrentX (db, -1, &req_get, group, uid, NULL, NULL, +- &result, &tofree); ++ &scratch); ++ result = scratch.dataset; ++ if (timeout < 0) ++ goto out; + } + + struct indataset +@@ -603,7 +621,7 @@ addinnetgrX (struct database_dyn *db, int fd, request_header *req, + } + + out: +- free (tofree); ++ addgetnetgrentX_scratch_free (&scratch); + return timeout; + } + +@@ -613,11 +631,12 @@ addgetnetgrentX_ignore (struct database_dyn *db, int fd, request_header *req, + const char *key, uid_t uid, struct hashentry *he, + struct datahead *dh) + { +- struct dataset *ignore; +- void *tofree; +- time_t timeout = addgetnetgrentX (db, fd, req, key, uid, he, dh, +- &ignore, &tofree); +- free (tofree); ++ struct addgetnetgrentX_scratch scratch; ++ addgetnetgrentX_scratch_init (&scratch); ++ time_t timeout = addgetnetgrentX (db, fd, req, key, uid, he, dh, &scratch); ++ addgetnetgrentX_scratch_free (&scratch); ++ if (timeout < 0) ++ timeout = 0; + return timeout; + } + +@@ -661,5 +680,9 @@ readdinnetgr (struct database_dyn *db, struct hashentry *he, + .key_len = he->len + }; + +- return addinnetgrX (db, -1, &req, db->data + he->key, he->owner, he, dh); ++ int timeout = addinnetgrX (db, -1, &req, db->data + he->key, he->owner, ++ he, dh); ++ if (timeout < 0) ++ timeout = 0; ++ return timeout; + } diff --git a/SOURCES/glibc-upstream-2.39-39.patch b/SOURCES/glibc-upstream-2.39-39.patch new file mode 100644 index 0000000..4dee6ea --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-39.patch @@ -0,0 +1,49 @@ +commit fd658f026f25cf59e8db243bc3b3e09cd5a20ba0 +Author: H.J. Lu +Date: Thu Apr 25 08:06:52 2024 -0700 + + elf: Also compile dl-misc.os with $(rtld-early-cflags) + + Also compile dl-misc.os with $(rtld-early-cflags) to avoid + + Program received signal SIGILL, Illegal instruction. + 0x00007ffff7fd36ea in _dl_strtoul (nptr=nptr@entry=0x7fffffffe2c9 "2", + endptr=endptr@entry=0x7fffffffd728) at dl-misc.c:156 + 156 bool positive = true; + (gdb) bt + #0 0x00007ffff7fd36ea in _dl_strtoul (nptr=nptr@entry=0x7fffffffe2c9 "2", + endptr=endptr@entry=0x7fffffffd728) at dl-misc.c:156 + #1 0x00007ffff7fdb1a9 in tunable_initialize ( + cur=cur@entry=0x7ffff7ffbc00 , + strval=strval@entry=0x7fffffffe2c9 "2", len=len@entry=1) + at dl-tunables.c:131 + #2 0x00007ffff7fdb3a2 in parse_tunables (valstring=) + at dl-tunables.c:258 + #3 0x00007ffff7fdb5d9 in __GI___tunables_init (envp=0x7fffffffdd58) + at dl-tunables.c:288 + #4 0x00007ffff7fe44c3 in _dl_sysdep_start ( + start_argptr=start_argptr@entry=0x7fffffffdcb0, + dl_main=dl_main@entry=0x7ffff7fe5f80 ) + at ../sysdeps/unix/sysv/linux/dl-sysdep.c:110 + #5 0x00007ffff7fe5cae in _dl_start_final (arg=0x7fffffffdcb0) at rtld.c:494 + #6 _dl_start (arg=0x7fffffffdcb0) at rtld.c:581 + #7 0x00007ffff7fe4b38 in _start () + (gdb) + + when setting GLIBC_TUNABLES in glibc compiled with APX. + Reviewed-by: Florian Weimer + + (cherry picked from commit 049b7684c912dd32b67b1b15b0f43bf07d5f512e) + +diff --git a/elf/Makefile b/elf/Makefile +index 69aa423c4b90127d..a50a988e7362cf3b 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -170,6 +170,7 @@ CFLAGS-.op += $(call elide-stack-protector,.op,$(elide-routines.os)) + CFLAGS-.os += $(call elide-stack-protector,.os,$(all-rtld-routines)) + + # Add the requested compiler flags to the early startup code. ++CFLAGS-dl-misc.os += $(rtld-early-cflags) + CFLAGS-dl-printf.os += $(rtld-early-cflags) + CFLAGS-dl-setup_hash.os += $(rtld-early-cflags) + CFLAGS-dl-sysdep.os += $(rtld-early-cflags) diff --git a/SOURCES/glibc-upstream-2.39-4.patch b/SOURCES/glibc-upstream-2.39-4.patch new file mode 100644 index 0000000..cc3350f --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-4.patch @@ -0,0 +1,42 @@ +commit d0724994de40934c552f1f68de89053848a44927 +Author: Xi Ruoyao +Date: Thu Feb 22 21:26:55 2024 +0100 + + math: Update mips64 ulps + + Signed-off-by: Andreas K. Hüttel + (cherry picked from commit e2a65ecc4b30a797df7dc6529f09b712aa256029) + +diff --git a/sysdeps/mips/mips64/libm-test-ulps b/sysdeps/mips/mips64/libm-test-ulps +index 78969745b245d094..933aba47350b777d 100644 +--- a/sysdeps/mips/mips64/libm-test-ulps ++++ b/sysdeps/mips/mips64/libm-test-ulps +@@ -1066,17 +1066,17 @@ double: 1 + ldouble: 1 + + Function: "j0": +-double: 2 ++double: 3 + float: 9 + ldouble: 2 + + Function: "j0_downward": +-double: 5 ++double: 6 + float: 9 + ldouble: 9 + + Function: "j0_towardzero": +-double: 6 ++double: 7 + float: 9 + ldouble: 9 + +@@ -1146,6 +1146,7 @@ float: 6 + ldouble: 8 + + Function: "log": ++double: 1 + float: 1 + ldouble: 1 + diff --git a/SOURCES/glibc-upstream-2.39-40.patch b/SOURCES/glibc-upstream-2.39-40.patch new file mode 100644 index 0000000..c3070ad --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-40.patch @@ -0,0 +1,209 @@ +commit 9831f98c266a8d56d1bf729b709c08e40375540c +Author: Florian Weimer +Date: Fri Apr 19 14:38:17 2024 +0200 + + login: Check default sizes of structs utmp, utmpx, lastlog + + The default is for ports with a 64-bit time_t. + Ports with a 32-bit time_t or with __WORDSIZE_TIME64_COMPAT32=1 + need to override it. + + Reviewed-by: Adhemerval Zanella + (cherry picked from commit 4d4da5aab936504b2d3eca3146e109630d9093c4) + +diff --git a/login/Makefile b/login/Makefile +index 1e22008a61e99083..b26ac42bfceadf89 100644 +--- a/login/Makefile ++++ b/login/Makefile +@@ -44,7 +44,7 @@ subdir-dirs = programs + vpath %.c programs + + tests := tst-utmp tst-utmpx tst-grantpt tst-ptsname tst-getlogin tst-updwtmpx \ +- tst-pututxline-lockfail tst-pututxline-cache ++ tst-pututxline-lockfail tst-pututxline-cache tst-utmp-size + + # Empty compatibility library for old binaries. + extra-libs := libutil +diff --git a/login/tst-utmp-size.c b/login/tst-utmp-size.c +new file mode 100644 +index 0000000000000000..1b7f7ff04224efb5 +--- /dev/null ++++ b/login/tst-utmp-size.c +@@ -0,0 +1,33 @@ ++/* Check expected sizes of struct utmp, struct utmpx, struct lastlog. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++ ++static int ++do_test (void) ++{ ++ _Static_assert (sizeof (struct utmp) == UTMP_SIZE, "struct utmp size"); ++ _Static_assert (sizeof (struct utmpx) == UTMP_SIZE, "struct utmpx size"); ++ _Static_assert (sizeof (struct lastlog) == LASTLOG_SIZE, ++ "struct lastlog size"); ++ return 0; ++} ++ ++#include +diff --git a/sysdeps/arc/utmp-size.h b/sysdeps/arc/utmp-size.h +new file mode 100644 +index 0000000000000000..a247fcd3dab15f81 +--- /dev/null ++++ b/sysdeps/arc/utmp-size.h +@@ -0,0 +1,3 @@ ++/* arc has less padding than other architectures with 64-bit time_t. */ ++#define UTMP_SIZE 392 ++#define LASTLOG_SIZE 296 +diff --git a/sysdeps/arm/utmp-size.h b/sysdeps/arm/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/arm/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/csky/utmp-size.h b/sysdeps/csky/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/csky/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/generic/utmp-size.h b/sysdeps/generic/utmp-size.h +new file mode 100644 +index 0000000000000000..89dbe878b02301e9 +--- /dev/null ++++ b/sysdeps/generic/utmp-size.h +@@ -0,0 +1,23 @@ ++/* Expected sizes of utmp-related structures stored in files. 64-bit version. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++/* Expected size, in bytes, of struct utmp and struct utmpx. */ ++#define UTMP_SIZE 400 ++ ++/* Expected size, in bytes, of struct lastlog. */ ++#define LASTLOG_SIZE 296 +diff --git a/sysdeps/hppa/utmp-size.h b/sysdeps/hppa/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/hppa/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/m68k/utmp-size.h b/sysdeps/m68k/utmp-size.h +new file mode 100644 +index 0000000000000000..5946685819d60289 +--- /dev/null ++++ b/sysdeps/m68k/utmp-size.h +@@ -0,0 +1,3 @@ ++/* m68k has 2-byte alignment. */ ++#define UTMP_SIZE 382 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/microblaze/utmp-size.h b/sysdeps/microblaze/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/microblaze/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/mips/utmp-size.h b/sysdeps/mips/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/mips/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/nios2/utmp-size.h b/sysdeps/nios2/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/nios2/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/or1k/utmp-size.h b/sysdeps/or1k/utmp-size.h +new file mode 100644 +index 0000000000000000..6b3653aa4dccd59d +--- /dev/null ++++ b/sysdeps/or1k/utmp-size.h +@@ -0,0 +1,3 @@ ++/* or1k has less padding than other architectures with 64-bit time_t. */ ++#define UTMP_SIZE 392 ++#define LASTLOG_SIZE 296 +diff --git a/sysdeps/powerpc/utmp-size.h b/sysdeps/powerpc/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/powerpc/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/riscv/utmp-size.h b/sysdeps/riscv/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/riscv/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/sh/utmp-size.h b/sysdeps/sh/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/sh/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/sparc/utmp-size.h b/sysdeps/sparc/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/sparc/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 +diff --git a/sysdeps/x86/utmp-size.h b/sysdeps/x86/utmp-size.h +new file mode 100644 +index 0000000000000000..8f21ebe1b6c26ea1 +--- /dev/null ++++ b/sysdeps/x86/utmp-size.h +@@ -0,0 +1,2 @@ ++#define UTMP_SIZE 384 ++#define LASTLOG_SIZE 292 diff --git a/SOURCES/glibc-upstream-2.39-41.patch b/SOURCES/glibc-upstream-2.39-41.patch new file mode 100644 index 0000000..659acb3 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-41.patch @@ -0,0 +1,368 @@ +commit 836d43b98973e0845b739ff5d3aad3af09dc7d0f +Author: Florian Weimer +Date: Fri Apr 19 14:38:17 2024 +0200 + + login: structs utmp, utmpx, lastlog _TIME_BITS independence (bug 30701) + + These structs describe file formats under /var/log, and should not + depend on the definition of _TIME_BITS. This is achieved by + defining __WORDSIZE_TIME64_COMPAT32 to 1 on 32-bit ports that + support 32-bit time_t values (where __time_t is 32 bits). + + Reviewed-by: Adhemerval Zanella + (cherry picked from commit 9abdae94c7454c45e02e97e4ed1eb1b1915d13d8) + +diff --git a/bits/wordsize.h b/bits/wordsize.h +index 14edae3a11d01c97..53013a9275c7c81e 100644 +--- a/bits/wordsize.h ++++ b/bits/wordsize.h +@@ -21,7 +21,9 @@ + #define __WORDSIZE32_PTRDIFF_LONG + + /* Set to 1 in order to force time types to be 32 bits instead of 64 bits in +- struct lastlog and struct utmp{,x} on 64-bit ports. This may be done in ++ struct lastlog and struct utmp{,x}. This may be done in + order to make 64-bit ports compatible with 32-bit ports. Set to 0 for +- 64-bit ports where the time types are 64-bits or for any 32-bit ports. */ ++ 64-bit ports where the time types are 64-bits and new 32-bit ports ++ where time_t is 64 bits, and there is no companion architecture with ++ 32-bit time_t. */ + #define __WORDSIZE_TIME64_COMPAT32 +diff --git a/login/Makefile b/login/Makefile +index b26ac42bfceadf89..f91190e3dcd1e6c6 100644 +--- a/login/Makefile ++++ b/login/Makefile +@@ -44,7 +44,9 @@ subdir-dirs = programs + vpath %.c programs + + tests := tst-utmp tst-utmpx tst-grantpt tst-ptsname tst-getlogin tst-updwtmpx \ +- tst-pututxline-lockfail tst-pututxline-cache tst-utmp-size ++ tst-pututxline-lockfail tst-pututxline-cache tst-utmp-size tst-utmp-size-64 ++ ++CFLAGS-tst-utmp-size-64.c += -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 + + # Empty compatibility library for old binaries. + extra-libs := libutil +diff --git a/login/tst-utmp-size-64.c b/login/tst-utmp-size-64.c +new file mode 100644 +index 0000000000000000..7a581a4c1254644a +--- /dev/null ++++ b/login/tst-utmp-size-64.c +@@ -0,0 +1,2 @@ ++/* The on-disk layout must not change in time64 mode. */ ++#include "tst-utmp-size.c" +diff --git a/sysdeps/arm/bits/wordsize.h b/sysdeps/arm/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/arm/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/csky/bits/wordsize.h b/sysdeps/csky/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/csky/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/m68k/bits/wordsize.h b/sysdeps/m68k/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/m68k/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/microblaze/bits/wordsize.h b/sysdeps/microblaze/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/microblaze/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/mips/bits/wordsize.h b/sysdeps/mips/bits/wordsize.h +index 57f0f2a22f81745c..30dd3fd85db1f966 100644 +--- a/sysdeps/mips/bits/wordsize.h ++++ b/sysdeps/mips/bits/wordsize.h +@@ -19,11 +19,7 @@ + + #define __WORDSIZE _MIPS_SZPTR + +-#if _MIPS_SIM == _ABI64 +-# define __WORDSIZE_TIME64_COMPAT32 1 +-#else +-# define __WORDSIZE_TIME64_COMPAT32 0 +-#endif ++#define __WORDSIZE_TIME64_COMPAT32 1 + + #if __WORDSIZE == 32 + #define __WORDSIZE32_SIZE_ULONG 0 +diff --git a/sysdeps/nios2/bits/wordsize.h b/sysdeps/nios2/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/nios2/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/powerpc/powerpc32/bits/wordsize.h b/sysdeps/powerpc/powerpc32/bits/wordsize.h +index 04ca9debf00d7ee9..6993fb6b29ea3f74 100644 +--- a/sysdeps/powerpc/powerpc32/bits/wordsize.h ++++ b/sysdeps/powerpc/powerpc32/bits/wordsize.h +@@ -2,10 +2,9 @@ + + #if defined __powerpc64__ + # define __WORDSIZE 64 +-# define __WORDSIZE_TIME64_COMPAT32 1 + #else + # define __WORDSIZE 32 +-# define __WORDSIZE_TIME64_COMPAT32 0 + # define __WORDSIZE32_SIZE_ULONG 0 + # define __WORDSIZE32_PTRDIFF_LONG 0 + #endif ++#define __WORDSIZE_TIME64_COMPAT32 1 +diff --git a/sysdeps/powerpc/powerpc64/bits/wordsize.h b/sysdeps/powerpc/powerpc64/bits/wordsize.h +index 04ca9debf00d7ee9..6993fb6b29ea3f74 100644 +--- a/sysdeps/powerpc/powerpc64/bits/wordsize.h ++++ b/sysdeps/powerpc/powerpc64/bits/wordsize.h +@@ -2,10 +2,9 @@ + + #if defined __powerpc64__ + # define __WORDSIZE 64 +-# define __WORDSIZE_TIME64_COMPAT32 1 + #else + # define __WORDSIZE 32 +-# define __WORDSIZE_TIME64_COMPAT32 0 + # define __WORDSIZE32_SIZE_ULONG 0 + # define __WORDSIZE32_PTRDIFF_LONG 0 + #endif ++#define __WORDSIZE_TIME64_COMPAT32 1 +diff --git a/sysdeps/sh/bits/wordsize.h b/sysdeps/sh/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/sh/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/sparc/sparc32/bits/wordsize.h b/sysdeps/sparc/sparc32/bits/wordsize.h +index 4bbd2e63b49bb2e2..a2e79e0fa9dc41d9 100644 +--- a/sysdeps/sparc/sparc32/bits/wordsize.h ++++ b/sysdeps/sparc/sparc32/bits/wordsize.h +@@ -1,6 +1,6 @@ + /* Determine the wordsize from the preprocessor defines. */ + + #define __WORDSIZE 32 +-#define __WORDSIZE_TIME64_COMPAT32 0 ++#define __WORDSIZE_TIME64_COMPAT32 1 + #define __WORDSIZE32_SIZE_ULONG 0 + #define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/sparc/sparc64/bits/wordsize.h b/sysdeps/sparc/sparc64/bits/wordsize.h +index 2f66f10d7206731a..ea103e5970829abc 100644 +--- a/sysdeps/sparc/sparc64/bits/wordsize.h ++++ b/sysdeps/sparc/sparc64/bits/wordsize.h +@@ -2,10 +2,9 @@ + + #if defined __arch64__ || defined __sparcv9 + # define __WORDSIZE 64 +-# define __WORDSIZE_TIME64_COMPAT32 1 + #else + # define __WORDSIZE 32 +-# define __WORDSIZE_TIME64_COMPAT32 0 + # define __WORDSIZE32_SIZE_ULONG 0 + # define __WORDSIZE32_PTRDIFF_LONG 0 + #endif ++#define __WORDSIZE_TIME64_COMPAT32 1 +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/wordsize.h b/sysdeps/unix/sysv/linux/hppa/bits/wordsize.h +new file mode 100644 +index 0000000000000000..6ecbfe7c863f3a17 +--- /dev/null ++++ b/sysdeps/unix/sysv/linux/hppa/bits/wordsize.h +@@ -0,0 +1,21 @@ ++/* Copyright (C) 1999-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#define __WORDSIZE 32 ++#define __WORDSIZE_TIME64_COMPAT32 1 ++#define __WORDSIZE32_SIZE_ULONG 0 ++#define __WORDSIZE32_PTRDIFF_LONG 0 +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/wordsize.h b/sysdeps/unix/sysv/linux/powerpc/bits/wordsize.h +index 04ca9debf00d7ee9..6993fb6b29ea3f74 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/wordsize.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/wordsize.h +@@ -2,10 +2,9 @@ + + #if defined __powerpc64__ + # define __WORDSIZE 64 +-# define __WORDSIZE_TIME64_COMPAT32 1 + #else + # define __WORDSIZE 32 +-# define __WORDSIZE_TIME64_COMPAT32 0 + # define __WORDSIZE32_SIZE_ULONG 0 + # define __WORDSIZE32_PTRDIFF_LONG 0 + #endif ++#define __WORDSIZE_TIME64_COMPAT32 1 +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/wordsize.h b/sysdeps/unix/sysv/linux/sparc/bits/wordsize.h +index 7562875ee23ba8c5..ea103e5970829abc 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/wordsize.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/wordsize.h +@@ -2,10 +2,9 @@ + + #if defined __arch64__ || defined __sparcv9 + # define __WORDSIZE 64 +-# define __WORDSIZE_TIME64_COMPAT32 1 + #else + # define __WORDSIZE 32 + # define __WORDSIZE32_SIZE_ULONG 0 + # define __WORDSIZE32_PTRDIFF_LONG 0 +-# define __WORDSIZE_TIME64_COMPAT32 0 + #endif ++#define __WORDSIZE_TIME64_COMPAT32 1 +diff --git a/sysdeps/x86/bits/wordsize.h b/sysdeps/x86/bits/wordsize.h +index 70f652bca14d65c1..3f40aa76f99c75e5 100644 +--- a/sysdeps/x86/bits/wordsize.h ++++ b/sysdeps/x86/bits/wordsize.h +@@ -8,10 +8,9 @@ + #define __WORDSIZE32_PTRDIFF_LONG 0 + #endif + ++#define __WORDSIZE_TIME64_COMPAT32 1 ++ + #ifdef __x86_64__ +-# define __WORDSIZE_TIME64_COMPAT32 1 + /* Both x86-64 and x32 use the 64-bit system call interface. */ + # define __SYSCALL_WORDSIZE 64 +-#else +-# define __WORDSIZE_TIME64_COMPAT32 0 + #endif diff --git a/SOURCES/glibc-upstream-2.39-42.patch b/SOURCES/glibc-upstream-2.39-42.patch new file mode 100644 index 0000000..6be162f --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-42.patch @@ -0,0 +1,31 @@ +commit acc56074b0a5127631a64640aef1b7c5c103ebd8 +Author: Florian Weimer +Date: Thu May 2 17:06:19 2024 +0200 + + nscd: Use time_t for return type of addgetnetgrentX + + Using int may give false results for future dates (timeouts after the + year 2028). + + Fixes commit 04a21e050d64a1193a6daab872bca2528bda44b ("CVE-2024-33601, + CVE-2024-33602: nscd: netgroup: Use two buffers in addgetnetgrentX + (bug 31680)"). + + Reviewed-by: Carlos O'Donell + (cherry picked from commit 4bbca1a44691a6e9adcee5c6798a707b626bc331) + +diff --git a/nscd/netgroupcache.c b/nscd/netgroupcache.c +index e8fe041846b75cb9..01d554af9c407739 100644 +--- a/nscd/netgroupcache.c ++++ b/nscd/netgroupcache.c +@@ -680,8 +680,8 @@ readdinnetgr (struct database_dyn *db, struct hashentry *he, + .key_len = he->len + }; + +- int timeout = addinnetgrX (db, -1, &req, db->data + he->key, he->owner, +- he, dh); ++ time_t timeout = addinnetgrX (db, -1, &req, db->data + he->key, he->owner, ++ he, dh); + if (timeout < 0) + timeout = 0; + return timeout; diff --git a/SOURCES/glibc-upstream-2.39-43.patch b/SOURCES/glibc-upstream-2.39-43.patch new file mode 100644 index 0000000..093670f --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-43.patch @@ -0,0 +1,73 @@ +commit 273a835fe7c685cc54266bb8b502787bad5e9bae +Author: Carlos O'Donell +Date: Tue Apr 23 13:30:37 2024 -0400 + + time: Allow later version licensing. + + The FSF's Licensing and Compliance Lab noted a discrepancy in the + licensing of several files in the glibc package. + + When timespect_get.c was impelemented the license did not include + the standard ", or (at your option) any later version." text. + + Change the license in timespec_get.c and all copied files to match + the expected license. + + This change was previously approved in principle by the FSF in + RT ticket #1316403. And a similar instance was fixed in + commit 46703efa02f6ddebce5ee54c92f7c32598de0de6. + + (cherry picked from commit 91695ee4598b39d181ab8df579b888a8863c4cab) + +diff --git a/sysdeps/unix/sysv/linux/timespec_get.c b/sysdeps/unix/sysv/linux/timespec_get.c +index c6e5e6628928523b..778d1e33548d2369 100644 +--- a/sysdeps/unix/sysv/linux/timespec_get.c ++++ b/sysdeps/unix/sysv/linux/timespec_get.c +@@ -5,7 +5,7 @@ + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either +- version 2.1 of the License. ++ version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of +diff --git a/sysdeps/unix/sysv/linux/timespec_getres.c b/sysdeps/unix/sysv/linux/timespec_getres.c +index 5acebe2a2cb99ccf..2eef9e512c6f650e 100644 +--- a/sysdeps/unix/sysv/linux/timespec_getres.c ++++ b/sysdeps/unix/sysv/linux/timespec_getres.c +@@ -5,7 +5,7 @@ + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either +- version 2.1 of the License. ++ version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of +diff --git a/time/timespec_get.c b/time/timespec_get.c +index b031e42ca20c1eea..26a044bca6e7f9fe 100644 +--- a/time/timespec_get.c ++++ b/time/timespec_get.c +@@ -4,7 +4,7 @@ + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either +- version 2.1 of the License. ++ version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of +diff --git a/time/timespec_getres.c b/time/timespec_getres.c +index edb397507cdfc2fa..2e18b8bcacfec498 100644 +--- a/time/timespec_getres.c ++++ b/time/timespec_getres.c +@@ -5,7 +5,7 @@ + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either +- version 2.1 of the License. ++ version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of diff --git a/SOURCES/glibc-upstream-2.39-44.patch b/SOURCES/glibc-upstream-2.39-44.patch new file mode 100644 index 0000000..92b587a --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-44.patch @@ -0,0 +1,34 @@ +commit 3148714ab61ad61281bae5a30f530d637034ac3b +Author: Gabi Falk +Date: Tue Apr 30 20:05:02 2024 +0000 + + i586: Fix multiple definitions of __memcpy_chk and __mempcpy_chk + + /home/bmg/install/compilers/x86_64-linux-gnu/lib/gcc/x86_64-glibc-linux-gnu/13.2.1/../../../../x86_64-glibc-linux-gnu/bin/ld: /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(memcpy_chk.o): in function `__memcpy_chk': + /home/bmg/src/glibc/debug/../sysdeps/i386/memcpy_chk.S:29: multiple definition of `__memcpy_chk';/home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(memcpy.o):/home/bmg/src/glibc/string/../sysdeps/i386/i586/memcpy.S:31: first defined here /home/bmg/install/compilers/x86_64-linux-gnu/lib/gcc/x86_64-glibc-linux-gnu/13.2.1/../../../../x86_64-glibc-linux-gnu/bin/ld: /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(mempcpy_chk.o): in function `__mempcpy_chk': /home/bmg/src/glibc/debug/../sysdeps/i386/mempcpy_chk.S:28: multiple definition of `__mempcpy_chk'; /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(mempcpy.o):/home/bmg/src/glibc/string/../sysdeps/i386/i586/memcpy.S:31: first defined here + + After this change, the static library built for i586, regardless of PIC + options, contains implementations of these functions respectively from + sysdeps/i386/memcpy_chk.S and sysdeps/i386/mempcpy_chk.S. This ensures + that memcpy and mempcpy won't pull in __chk_fail and the routines it + calls. + + Reported-by: Florian Weimer + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit 789894a2f554d4503ecb2f13b2b4e93e43414f33) + +diff --git a/sysdeps/i386/i586/memcpy.S b/sysdeps/i386/i586/memcpy.S +index 3e26f112d685e148..79856d498af90f66 100644 +--- a/sysdeps/i386/i586/memcpy.S ++++ b/sysdeps/i386/i586/memcpy.S +@@ -26,7 +26,7 @@ + #define LEN SRC+4 + + .text +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY (__memcpy_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-45.patch b/SOURCES/glibc-upstream-2.39-45.patch new file mode 100644 index 0000000..208555e --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-45.patch @@ -0,0 +1,57 @@ +commit ad92c483a4bd34db1cfb3eb625212ea64848244f +Author: Gabi Falk +Date: Tue Apr 30 20:05:03 2024 +0000 + + i686: Fix multiple definitions of __memmove_chk and __memset_chk + + Commit c73c96a4a1af1326df7f96eec58209e1e04066d8 updated memcpy.S and + mempcpy.S, but omitted memmove.S and memset.S. As a result, the static + library built as PIC, whether with or without multiarch support, + contains two definitions for each of the __memmove_chk and __memset_chk + symbols. + + /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../i686-pc-linux-gnu/bin/ld: /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../lib/libc.a(memset-ia32.o): in function `__memset_chk': + /var/tmp/portage/sys-libs/glibc-2.39-r3/work/glibc-2.39/string/../sysdeps/i386/i686/memset.S:32: multiple definition of `__memset_chk'; /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../lib/libc.a(memset_chk.o):/var/tmp/portage/sys-libs/glibc-2.39-r3/work/glibc-2.39/debug/../sysdeps/i386/i686/multiarch/memset_chk.c:24: first defined here + + After this change, regardless of PIC options, the static library, built + for i686 with multiarch contains implementations of these functions + respectively from debug/memmove_chk.c and debug/memset_chk.c, and + without multiarch contains implementations of these functions + respectively from sysdeps/i386/memmove_chk.S and + sysdeps/i386/memset_chk.S. This ensures that memmove and memset won't + pull in __chk_fail and the routines it calls. + + Reported-by: Sam James + Tested-by: Sam James + Fixes: c73c96a4a1 ("i686: Fix build with --disable-multiarch") + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit 5a2cf833f5772d6c37c7adac388dd9af9cc1c4b9) + +diff --git a/sysdeps/i386/i686/memmove.S b/sysdeps/i386/i686/memmove.S +index f230359ad62b2443..effd958120082b04 100644 +--- a/sysdeps/i386/i686/memmove.S ++++ b/sysdeps/i386/i686/memmove.S +@@ -29,7 +29,7 @@ + #define SRC DEST+4 + #define LEN SRC+4 + +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY_CHK (__memmove_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) +diff --git a/sysdeps/i386/i686/memset.S b/sysdeps/i386/i686/memset.S +index f02f5a6df763d4e9..ab06771ea0ca071f 100644 +--- a/sysdeps/i386/i686/memset.S ++++ b/sysdeps/i386/i686/memset.S +@@ -27,7 +27,7 @@ + #define LEN CHR+4 + + .text +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY_CHK (__memset_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-46.patch b/SOURCES/glibc-upstream-2.39-46.patch new file mode 100644 index 0000000..d2f28d5 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-46.patch @@ -0,0 +1,34 @@ +commit ff110b2591f0bdeccd121c3726af19c62d6fb184 +Author: Gabi Falk +Date: Tue Apr 30 20:05:04 2024 +0000 + + Add a test to check for duplicate definitions in the static library + + This change follows two previous fixes addressing multiple definitions + of __memcpy_chk and __mempcpy_chk functions on i586, and __memmove_chk + and __memset_chk functions on i686. The test is intended to prevent + such issues from occurring in the future. + + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit ded2e0753e9c46debeb2e0d26c5e560d2581d314) + +diff --git a/Makefile b/Makefile +index f7e4eb9ff2cc464c..37bf70aa4ad4403f 100644 +--- a/Makefile ++++ b/Makefile +@@ -577,6 +577,13 @@ $(objpfx)lint-makefiles.out: scripts/lint-makefiles.sh + $(SHELL) $< "$(PYTHON)" `pwd` > $@ ; \ + $(evaluate-test) + ++# Link libc.a as a whole to verify that it does not contain multiple ++# definitions of any symbols. ++tests-special += $(objpfx)link-static-libc.out ++$(objpfx)link-static-libc.out: ++ $(LINK.o) $(whole-archive) -r $(objpfx)libc.a -o /dev/null > $@ 2>&1; \ ++ $(evaluate-test) ++ + # Print test summary for tests in $1 .sum file; + # $2 is optional test identifier. + # Fail if there are unexpected failures in the test results. diff --git a/SOURCES/glibc-upstream-2.39-47.patch b/SOURCES/glibc-upstream-2.39-47.patch new file mode 100644 index 0000000..0ab4156 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-47.patch @@ -0,0 +1,32 @@ +commit fa616ea3730cb42046d19f28d611be0bc390af7c +Author: Sam James +Date: Sat May 4 13:28:13 2024 +0100 + + Revert "Add a test to check for duplicate definitions in the static library" + + This reverts commit ff110b2591f0bdeccd121c3726af19c62d6fb184. + + I had the wrong cherry-pick reference (the commit content is right; it's + just referring to a base that isn't upstream), but let's revert and reapply + for clarity. + + Signed-off-by: Sam James + +diff --git a/Makefile b/Makefile +index 37bf70aa4ad4403f..f7e4eb9ff2cc464c 100644 +--- a/Makefile ++++ b/Makefile +@@ -577,13 +577,6 @@ $(objpfx)lint-makefiles.out: scripts/lint-makefiles.sh + $(SHELL) $< "$(PYTHON)" `pwd` > $@ ; \ + $(evaluate-test) + +-# Link libc.a as a whole to verify that it does not contain multiple +-# definitions of any symbols. +-tests-special += $(objpfx)link-static-libc.out +-$(objpfx)link-static-libc.out: +- $(LINK.o) $(whole-archive) -r $(objpfx)libc.a -o /dev/null > $@ 2>&1; \ +- $(evaluate-test) +- + # Print test summary for tests in $1 .sum file; + # $2 is optional test identifier. + # Fail if there are unexpected failures in the test results. diff --git a/SOURCES/glibc-upstream-2.39-48.patch b/SOURCES/glibc-upstream-2.39-48.patch new file mode 100644 index 0000000..133d6a0 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-48.patch @@ -0,0 +1,40 @@ +commit c16871e662cd0f3370173d916864b19e69f1bc9a +Author: Sam James +Date: Sat May 4 13:28:51 2024 +0100 + + Revert "i686: Fix multiple definitions of __memmove_chk and __memset_chk" + + This reverts commit ad92c483a4bd34db1cfb3eb625212ea64848244f. + + I had the wrong cherry-pick reference (the commit content is right; it's + just referring to a base that isn't upstream), but let's revert and reapply + for clarity. + + Signed-off-by: Sam James + +diff --git a/sysdeps/i386/i686/memmove.S b/sysdeps/i386/i686/memmove.S +index effd958120082b04..f230359ad62b2443 100644 +--- a/sysdeps/i386/i686/memmove.S ++++ b/sysdeps/i386/i686/memmove.S +@@ -29,7 +29,7 @@ + #define SRC DEST+4 + #define LEN SRC+4 + +-#if defined SHARED && IS_IN (libc) ++#if defined PIC && IS_IN (libc) + ENTRY_CHK (__memmove_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) +diff --git a/sysdeps/i386/i686/memset.S b/sysdeps/i386/i686/memset.S +index ab06771ea0ca071f..f02f5a6df763d4e9 100644 +--- a/sysdeps/i386/i686/memset.S ++++ b/sysdeps/i386/i686/memset.S +@@ -27,7 +27,7 @@ + #define LEN CHR+4 + + .text +-#if defined SHARED && IS_IN (libc) ++#if defined PIC && IS_IN (libc) + ENTRY_CHK (__memset_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-49.patch b/SOURCES/glibc-upstream-2.39-49.patch new file mode 100644 index 0000000..a3c6b88 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-49.patch @@ -0,0 +1,27 @@ +commit 5141d4d83c17406f0eaea3e345ef2b52e10f386e +Author: Sam James +Date: Sat May 4 13:28:54 2024 +0100 + + Revert "i586: Fix multiple definitions of __memcpy_chk and __mempcpy_chk" + + This reverts commit 3148714ab61ad61281bae5a30f530d637034ac3b. + + I had the wrong cherry-pick reference (the commit content is right; it's + just referring to a base that isn't upstream), but let's revert and reapply + for clarity. + + Signed-off-by: Sam James + +diff --git a/sysdeps/i386/i586/memcpy.S b/sysdeps/i386/i586/memcpy.S +index 79856d498af90f66..3e26f112d685e148 100644 +--- a/sysdeps/i386/i586/memcpy.S ++++ b/sysdeps/i386/i586/memcpy.S +@@ -26,7 +26,7 @@ + #define LEN SRC+4 + + .text +-#if defined SHARED && IS_IN (libc) ++#if defined PIC && IS_IN (libc) + ENTRY (__memcpy_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-5.patch b/SOURCES/glibc-upstream-2.39-5.patch new file mode 100644 index 0000000..92661f6 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-5.patch @@ -0,0 +1,143 @@ +commit e0910f1d3278f05439fb434ee528fc9be1b6bd5e +Author: Stefan Liebler +Date: Thu Feb 22 15:03:27 2024 +0100 + + S390: Do not clobber r7 in clone [BZ #31402] + + Starting with commit e57d8fc97b90127de4ed3e3a9cdf663667580935 + "S390: Always use svc 0" + clone clobbers the call-saved register r7 in error case: + function or stack is NULL. + + This patch restores the saved registers also in the error case. + Furthermore the existing test misc/tst-clone is extended to check + all error cases and that clone does not clobber registers in this + error case. + + (cherry picked from commit 02782fd12849b6673cb5c2728cb750e8ec295aa3) + +diff --git a/sysdeps/unix/sysv/linux/s390/s390-32/clone.S b/sysdeps/unix/sysv/linux/s390/s390-32/clone.S +index 4c882ef2ee3f5b81..a7a863242c4ebd84 100644 +--- a/sysdeps/unix/sysv/linux/s390/s390-32/clone.S ++++ b/sysdeps/unix/sysv/linux/s390/s390-32/clone.S +@@ -53,6 +53,7 @@ ENTRY(__clone) + br %r14 + error: + lhi %r2,-EINVAL ++ lm %r6,%r7,24(%r15) /* Load registers. */ + j SYSCALL_ERROR_LABEL + PSEUDO_END (__clone) + +diff --git a/sysdeps/unix/sysv/linux/s390/s390-64/clone.S b/sysdeps/unix/sysv/linux/s390/s390-64/clone.S +index 4eb104be71ee565a..c552a6b8decb63de 100644 +--- a/sysdeps/unix/sysv/linux/s390/s390-64/clone.S ++++ b/sysdeps/unix/sysv/linux/s390/s390-64/clone.S +@@ -54,6 +54,7 @@ ENTRY(__clone) + br %r14 + error: + lghi %r2,-EINVAL ++ lmg %r6,%r7,48(%r15) /* Restore registers. */ + jg SYSCALL_ERROR_LABEL + PSEUDO_END (__clone) + +diff --git a/sysdeps/unix/sysv/linux/tst-clone.c b/sysdeps/unix/sysv/linux/tst-clone.c +index 470676ab2bb3fc31..2bc71249837fbb66 100644 +--- a/sysdeps/unix/sysv/linux/tst-clone.c ++++ b/sysdeps/unix/sysv/linux/tst-clone.c +@@ -16,12 +16,16 @@ + License along with the GNU C Library; if not, see + . */ + +-/* BZ #2386 */ ++/* BZ #2386, BZ #31402 */ + #include + #include + #include + #include + #include ++#include /* For _STACK_GROWS_{UP,DOWN}. */ ++#include ++ ++volatile unsigned v = 0xdeadbeef; + + int child_fn(void *arg) + { +@@ -30,22 +34,67 @@ int child_fn(void *arg) + } + + static int +-do_test (void) ++__attribute__((noinline)) ++do_clone (int (*fn)(void *), void *stack) + { + int result; ++ unsigned int a = v; ++ unsigned int b = v; ++ unsigned int c = v; ++ unsigned int d = v; ++ unsigned int e = v; ++ unsigned int f = v; ++ unsigned int g = v; ++ unsigned int h = v; ++ unsigned int i = v; ++ unsigned int j = v; ++ unsigned int k = v; ++ unsigned int l = v; ++ unsigned int m = v; ++ unsigned int n = v; ++ unsigned int o = v; ++ ++ result = clone (fn, stack, 0, NULL); ++ ++ /* Check that clone does not clobber call-saved registers. */ ++ TEST_VERIFY (a == v && b == v && c == v && d == v && e == v && f == v ++ && g == v && h == v && i == v && j == v && k == v && l == v ++ && m == v && n == v && o == v); ++ ++ return result; ++} ++ ++static void ++__attribute__((noinline)) ++do_test_single (int (*fn)(void *), void *stack) ++{ ++ printf ("%s (fn=%p, stack=%p)\n", __FUNCTION__, fn, stack); ++ errno = 0; ++ ++ int result = do_clone (fn, stack); ++ ++ TEST_COMPARE (errno, EINVAL); ++ TEST_COMPARE (result, -1); ++} + +- result = clone (child_fn, NULL, 0, NULL); ++static int ++do_test (void) ++{ ++ char st[128 * 1024] __attribute__ ((aligned)); ++ void *stack = NULL; ++#if _STACK_GROWS_DOWN ++ stack = st + sizeof (st); ++#elif _STACK_GROWS_UP ++ stack = st; ++#else ++# error "Define either _STACK_GROWS_DOWN or _STACK_GROWS_UP" ++#endif + +- if (errno != EINVAL || result != -1) +- { +- printf ("FAIL: clone()=%d (wanted -1) errno=%d (wanted %d)\n", +- result, errno, EINVAL); +- return 1; +- } ++ do_test_single (child_fn, NULL); ++ do_test_single (NULL, stack); ++ do_test_single (NULL, NULL); + +- puts ("All OK"); + return 0; + } + +-#define TEST_FUNCTION do_test () +-#include "../test-skeleton.c" ++#include diff --git a/SOURCES/glibc-upstream-2.39-50.patch b/SOURCES/glibc-upstream-2.39-50.patch new file mode 100644 index 0000000..d79e2db --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-50.patch @@ -0,0 +1,34 @@ +commit 8323a83abd73446dc434aceff66219712c09140b +Author: Gabi Falk +Date: Tue Apr 30 20:05:02 2024 +0000 + + i586: Fix multiple definitions of __memcpy_chk and __mempcpy_chk + + /home/bmg/install/compilers/x86_64-linux-gnu/lib/gcc/x86_64-glibc-linux-gnu/13.2.1/../../../../x86_64-glibc-linux-gnu/bin/ld: /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(memcpy_chk.o): in function `__memcpy_chk': + /home/bmg/src/glibc/debug/../sysdeps/i386/memcpy_chk.S:29: multiple definition of `__memcpy_chk';/home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(memcpy.o):/home/bmg/src/glibc/string/../sysdeps/i386/i586/memcpy.S:31: first defined here /home/bmg/install/compilers/x86_64-linux-gnu/lib/gcc/x86_64-glibc-linux-gnu/13.2.1/../../../../x86_64-glibc-linux-gnu/bin/ld: /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(mempcpy_chk.o): in function `__mempcpy_chk': /home/bmg/src/glibc/debug/../sysdeps/i386/mempcpy_chk.S:28: multiple definition of `__mempcpy_chk'; /home/bmg/build/glibcs/i586-linux-gnu/glibc/libc.a(mempcpy.o):/home/bmg/src/glibc/string/../sysdeps/i386/i586/memcpy.S:31: first defined here + + After this change, the static library built for i586, regardless of PIC + options, contains implementations of these functions respectively from + sysdeps/i386/memcpy_chk.S and sysdeps/i386/mempcpy_chk.S. This ensures + that memcpy and mempcpy won't pull in __chk_fail and the routines it + calls. + + Reported-by: Florian Weimer + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit 0fdf4ba48ccce5abf567340b0ab8fa8ed8a9bc6e) + +diff --git a/sysdeps/i386/i586/memcpy.S b/sysdeps/i386/i586/memcpy.S +index 3e26f112d685e148..79856d498af90f66 100644 +--- a/sysdeps/i386/i586/memcpy.S ++++ b/sysdeps/i386/i586/memcpy.S +@@ -26,7 +26,7 @@ + #define LEN SRC+4 + + .text +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY (__memcpy_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-51.patch b/SOURCES/glibc-upstream-2.39-51.patch new file mode 100644 index 0000000..4c9718c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-51.patch @@ -0,0 +1,57 @@ +commit 8b005d7869debac4d5cd67f65e49a0fad89da9ad +Author: Gabi Falk +Date: Tue Apr 30 20:05:03 2024 +0000 + + i686: Fix multiple definitions of __memmove_chk and __memset_chk + + Commit c73c96a4a1af1326df7f96eec58209e1e04066d8 updated memcpy.S and + mempcpy.S, but omitted memmove.S and memset.S. As a result, the static + library built as PIC, whether with or without multiarch support, + contains two definitions for each of the __memmove_chk and __memset_chk + symbols. + + /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../i686-pc-linux-gnu/bin/ld: /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../lib/libc.a(memset-ia32.o): in function `__memset_chk': + /var/tmp/portage/sys-libs/glibc-2.39-r3/work/glibc-2.39/string/../sysdeps/i386/i686/memset.S:32: multiple definition of `__memset_chk'; /usr/lib/gcc/i686-pc-linux-gnu/14/../../../../lib/libc.a(memset_chk.o):/var/tmp/portage/sys-libs/glibc-2.39-r3/work/glibc-2.39/debug/../sysdeps/i386/i686/multiarch/memset_chk.c:24: first defined here + + After this change, regardless of PIC options, the static library, built + for i686 with multiarch contains implementations of these functions + respectively from debug/memmove_chk.c and debug/memset_chk.c, and + without multiarch contains implementations of these functions + respectively from sysdeps/i386/memmove_chk.S and + sysdeps/i386/memset_chk.S. This ensures that memmove and memset won't + pull in __chk_fail and the routines it calls. + + Reported-by: Sam James + Tested-by: Sam James + Fixes: c73c96a4a1 ("i686: Fix build with --disable-multiarch") + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit 5a2cf833f5772d6c37c7adac388dd9af9cc1c4b9) + +diff --git a/sysdeps/i386/i686/memmove.S b/sysdeps/i386/i686/memmove.S +index f230359ad62b2443..effd958120082b04 100644 +--- a/sysdeps/i386/i686/memmove.S ++++ b/sysdeps/i386/i686/memmove.S +@@ -29,7 +29,7 @@ + #define SRC DEST+4 + #define LEN SRC+4 + +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY_CHK (__memmove_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) +diff --git a/sysdeps/i386/i686/memset.S b/sysdeps/i386/i686/memset.S +index f02f5a6df763d4e9..ab06771ea0ca071f 100644 +--- a/sysdeps/i386/i686/memset.S ++++ b/sysdeps/i386/i686/memset.S +@@ -27,7 +27,7 @@ + #define LEN CHR+4 + + .text +-#if defined PIC && IS_IN (libc) ++#if defined SHARED && IS_IN (libc) + ENTRY_CHK (__memset_chk) + movl 12(%esp), %eax + cmpl %eax, 16(%esp) diff --git a/SOURCES/glibc-upstream-2.39-52.patch b/SOURCES/glibc-upstream-2.39-52.patch new file mode 100644 index 0000000..c03d7cc --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-52.patch @@ -0,0 +1,34 @@ +commit f8e462342189525e4605cf233b8f798d1c7f398d +Author: Gabi Falk +Date: Tue Apr 30 20:05:04 2024 +0000 + + Add a test to check for duplicate definitions in the static library + + This change follows two previous fixes addressing multiple definitions + of __memcpy_chk and __mempcpy_chk functions on i586, and __memmove_chk + and __memset_chk functions on i686. The test is intended to prevent + such issues from occurring in the future. + + Signed-off-by: Gabi Falk + Reviewed-by: H.J. Lu + Reviewed-by: Dmitry V. Levin + (cherry picked from commit ded2e0753e9c46debeb2e0d26c5e560d2581d314) + +diff --git a/Makefile b/Makefile +index f7e4eb9ff2cc464c..37bf70aa4ad4403f 100644 +--- a/Makefile ++++ b/Makefile +@@ -577,6 +577,13 @@ $(objpfx)lint-makefiles.out: scripts/lint-makefiles.sh + $(SHELL) $< "$(PYTHON)" `pwd` > $@ ; \ + $(evaluate-test) + ++# Link libc.a as a whole to verify that it does not contain multiple ++# definitions of any symbols. ++tests-special += $(objpfx)link-static-libc.out ++$(objpfx)link-static-libc.out: ++ $(LINK.o) $(whole-archive) -r $(objpfx)libc.a -o /dev/null > $@ 2>&1; \ ++ $(evaluate-test) ++ + # Print test summary for tests in $1 .sum file; + # $2 is optional test identifier. + # Fail if there are unexpected failures in the test results. diff --git a/SOURCES/glibc-upstream-2.39-53.patch b/SOURCES/glibc-upstream-2.39-53.patch new file mode 100644 index 0000000..94ec662 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-53.patch @@ -0,0 +1,207 @@ +commit 71149c2a2e85a8233631cc816030d449f021bb2a +Author: Adhemerval Zanella +Date: Mon May 6 13:18:45 2024 -0300 + + elf: Only process multiple tunable once (BZ 31686) + + The 680c597e9c3 commit made loader reject ill-formatted strings by + first tracking all set tunables and then applying them. However, it does + not take into consideration if the same tunable is set multiple times, + where parse_tunables_string appends the found tunable without checking + if it was already in the list. It leads to a stack-based buffer overflow + if the tunable is specified more than the total number of tunables. For + instance: + + GLIBC_TUNABLES=glibc.malloc.check=2:... (repeat over the number of + total support for different tunable). + + Instead, use the index of the tunable list to get the expected tunable + entry. Since now the initial list is zero-initialized, the compiler + might emit an extra memset and this requires some minor adjustment + on some ports. + + Checked on x86_64-linux-gnu and aarch64-linux-gnu. + + Reported-by: Yuto Maeda + Reported-by: Yutaro Shimizu + Reviewed-by: Siddhesh Poyarekar + (cherry picked from commit bcae44ea8536b30a7119c0986ff5692bddacb672) + +diff --git a/elf/dl-tunables.c b/elf/dl-tunables.c +index 03e1a68675d65224..614ac9c0471c5963 100644 +--- a/elf/dl-tunables.c ++++ b/elf/dl-tunables.c +@@ -32,6 +32,7 @@ + #include + #include + #include ++#include + + #define TUNABLES_INTERNAL 1 + #include "dl-tunables.h" +@@ -223,6 +224,7 @@ parse_tunables_string (const char *valstring, struct tunable_toset_t *tunables) + { + tunables[ntunables++] = + (struct tunable_toset_t) { cur, value, p - value }; ++ + break; + } + } +@@ -234,23 +236,27 @@ parse_tunables_string (const char *valstring, struct tunable_toset_t *tunables) + static void + parse_tunables (const char *valstring) + { +- struct tunable_toset_t tunables[tunables_list_size]; +- int ntunables = parse_tunables_string (valstring, tunables); +- if (ntunables == -1) ++ struct tunable_toset_t tunables[tunables_list_size] = { 0 }; ++ if (parse_tunables_string (valstring, tunables) == -1) + { + _dl_error_printf ( + "WARNING: ld.so: invalid GLIBC_TUNABLES `%s': ignored.\n", valstring); + return; + } + +- for (int i = 0; i < ntunables; i++) +- if (!tunable_initialize (tunables[i].t, tunables[i].value, +- tunables[i].len)) +- _dl_error_printf ("WARNING: ld.so: invalid GLIBC_TUNABLES value `%.*s' " +- "for option `%s': ignored.\n", +- (int) tunables[i].len, +- tunables[i].value, +- tunables[i].t->name); ++ for (int i = 0; i < tunables_list_size; i++) ++ { ++ if (tunables[i].t == NULL) ++ continue; ++ ++ if (!tunable_initialize (tunables[i].t, tunables[i].value, ++ tunables[i].len)) ++ _dl_error_printf ("WARNING: ld.so: invalid GLIBC_TUNABLES value `%.*s' " ++ "for option `%s': ignored.\n", ++ (int) tunables[i].len, ++ tunables[i].value, ++ tunables[i].t->name); ++ } + } + + /* Initialize the tunables list from the environment. For now we only use the +diff --git a/elf/tst-tunables.c b/elf/tst-tunables.c +index 095b5c81d95c8760..dff34ed748b4ae83 100644 +--- a/elf/tst-tunables.c ++++ b/elf/tst-tunables.c +@@ -17,6 +17,10 @@ + . */ + + #include ++/* The test uses the tunable_list size, which is only exported for ++ ld.so. This will result in a copy of tunable_list, which is ununsed by ++ the test itself. */ ++#define TUNABLES_INTERNAL 1 + #include + #include + #include +@@ -24,12 +28,13 @@ + #include + #include + #include ++#include + + static int restart; + #define CMDLINE_OPTIONS \ + { "restart", no_argument, &restart, 1 }, + +-static const struct test_t ++static struct test_t + { + const char *name; + const char *value; +@@ -284,6 +289,29 @@ static const struct test_t + 0, + 0, + }, ++ /* Also check for repeated tunables with a count larger than the total number ++ of tunables. */ ++ { ++ "GLIBC_TUNABLES", ++ NULL, ++ 2, ++ 0, ++ 0, ++ }, ++ { ++ "GLIBC_TUNABLES", ++ NULL, ++ 1, ++ 0, ++ 0, ++ }, ++ { ++ "GLIBC_TUNABLES", ++ NULL, ++ 0, ++ 0, ++ 0, ++ }, + }; + + static int +@@ -327,6 +355,37 @@ do_test (int argc, char *argv[]) + spargv[i] = NULL; + } + ++ /* Create a tunable line with the duplicate values with a total number ++ larger than the different number of tunables. */ ++ { ++ enum { tunables_list_size = array_length (tunable_list) }; ++ const char *value = ""; ++ for (int i = 0; i < tunables_list_size; i++) ++ value = xasprintf ("%sglibc.malloc.check=2%c", ++ value, ++ i == (tunables_list_size - 1) ? '\0' : ':'); ++ tests[33].value = value; ++ } ++ /* Same as before, but the last tunable values is differen than the ++ rest. */ ++ { ++ enum { tunables_list_size = array_length (tunable_list) }; ++ const char *value = ""; ++ for (int i = 0; i < tunables_list_size - 1; i++) ++ value = xasprintf ("%sglibc.malloc.check=2:", value); ++ value = xasprintf ("%sglibc.malloc.check=1", value); ++ tests[34].value = value; ++ } ++ /* Same as before, but with an invalid last entry. */ ++ { ++ enum { tunables_list_size = array_length (tunable_list) }; ++ const char *value = ""; ++ for (int i = 0; i < tunables_list_size - 1; i++) ++ value = xasprintf ("%sglibc.malloc.check=2:", value); ++ value = xasprintf ("%sglibc.malloc.check=1=1", value); ++ tests[35].value = value; ++ } ++ + for (int i = 0; i < array_length (tests); i++) + { + snprintf (nteststr, sizeof nteststr, "%d", i); +diff --git a/sysdeps/aarch64/multiarch/memset_generic.S b/sysdeps/aarch64/multiarch/memset_generic.S +index 81748bdbce53e636..e125a5ed853301e1 100644 +--- a/sysdeps/aarch64/multiarch/memset_generic.S ++++ b/sysdeps/aarch64/multiarch/memset_generic.S +@@ -33,3 +33,7 @@ + #endif + + #include <../memset.S> ++ ++#if IS_IN (rtld) ++strong_alias (memset, __memset_generic) ++#endif +diff --git a/sysdeps/sparc/sparc64/rtld-memset.c b/sysdeps/sparc/sparc64/rtld-memset.c +index 55f38357902933b5..a19202a620fb8f7d 100644 +--- a/sysdeps/sparc/sparc64/rtld-memset.c ++++ b/sysdeps/sparc/sparc64/rtld-memset.c +@@ -1 +1,4 @@ + #include ++#if IS_IN(rtld) ++strong_alias (memset, __memset_ultra1) ++#endif diff --git a/SOURCES/glibc-upstream-2.39-54.patch b/SOURCES/glibc-upstream-2.39-54.patch new file mode 100644 index 0000000..7ed1911 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-54.patch @@ -0,0 +1,45 @@ +commit 97bb89668d7171164975f3dc895e38343a2f3a95 +Author: H.J. Lu +Date: Thu May 9 20:07:01 2024 -0700 + + Force DT_RPATH for --enable-hardcoded-path-in-tests + + On Fedora 40/x86-64, linker enables --enable-new-dtags by default which + generates DT_RUNPATH instead of DT_RPATH. Unlike DT_RPATH, DT_RUNPATH + only applies to DT_NEEDED entries in the executable and doesn't applies + to DT_NEEDED entries in shared libraries which are loaded via DT_NEEDED + entries in the executable. Some glibc tests have libstdc++.so.6 in + DT_NEEDED, which has libm.so.6 in DT_NEEDED. When DT_RUNPATH is generated, + /lib64/libm.so.6 is loaded for such tests. If the newly built glibc is + older than glibc 2.36, these tests fail with + + assert/tst-assert-c++: /export/build/gnu/tools-build/glibc-gitlab-release/build-x86_64-linux/libc.so.6: version `GLIBC_2.36' not found (required by /lib64/libm.so.6) + assert/tst-assert-c++: /export/build/gnu/tools-build/glibc-gitlab-release/build-x86_64-linux/libc.so.6: version `GLIBC_ABI_DT_RELR' not found (required by /lib64/libm.so.6) + + Pass -Wl,--disable-new-dtags to linker when building glibc tests with + --enable-hardcoded-path-in-tests. This fixes BZ #31719. + + Signed-off-by: H.J. Lu + (cherry picked from commit 2dcaf70643710e22f92a351e36e3cff8b48c60dc) + +diff --git a/Makeconfig b/Makeconfig +index 85e00cef94e9deb4..522182abdc426e70 100644 +--- a/Makeconfig ++++ b/Makeconfig +@@ -586,10 +586,13 @@ link-libc-rpath-link = -Wl,-rpath-link=$(rpath-link) + # before the expansion of LDLIBS-* variables). + + # Tests use -Wl,-rpath instead of -Wl,-rpath-link for +-# build-hardcoded-path-in-tests. ++# build-hardcoded-path-in-tests. Add -Wl,--disable-new-dtags to force ++# DT_RPATH instead of DT_RUNPATH which only applies to DT_NEEDED entries ++# in the executable and doesn't applies to DT_NEEDED entries in shared ++# libraries which are loaded via DT_NEEDED entries in the executable. + ifeq (yes,$(build-hardcoded-path-in-tests)) +-link-libc-tests-rpath-link = $(link-libc-rpath) +-link-test-modules-rpath-link = $(link-libc-rpath) ++link-libc-tests-rpath-link = $(link-libc-rpath) -Wl,--disable-new-dtags ++link-test-modules-rpath-link = $(link-libc-rpath) -Wl,--disable-new-dtags + else + link-libc-tests-rpath-link = $(link-libc-rpath-link) + link-test-modules-rpath-link = diff --git a/SOURCES/glibc-upstream-2.39-55.patch b/SOURCES/glibc-upstream-2.39-55.patch new file mode 100644 index 0000000..0bb52db --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-55.patch @@ -0,0 +1,38 @@ +commit ab4ef4421f85ea7afeb482ded51003658b08de7e +Author: Gabi Falk +Date: Tue May 7 18:25:00 2024 +0000 + + x86_64: Fix missing wcsncat function definition without multiarch (x86-64-v4) + + This code expects the WCSCAT preprocessor macro to be predefined in case + the evex implementation of the function should be defined with a name + different from __wcsncat_evex. However, when glibc is built for + x86-64-v4 without multiarch support, sysdeps/x86_64/wcsncat.S defines + WCSNCAT variable instead of WCSCAT to build it as wcsncat. Rename the + variable to WCSNCAT, as it is actually a better naming choice for the + variable in this case. + + Reported-by: Kenton Groombridge + Link: https://bugs.gentoo.org/921945 + Fixes: 64b8b6516b ("x86: Add evex optimized functions for the wchar_t strcpy family") + Signed-off-by: Gabi Falk + Reviewed-by: Sunil K Pandey + (cherry picked from commit dd5f891c1ad9f1b43b9db93afe2a55cbb7a6194e) + +diff --git a/sysdeps/x86_64/multiarch/wcsncat-evex.S b/sysdeps/x86_64/multiarch/wcsncat-evex.S +index 392215950afd56be..10bfb0a5314130cf 100644 +--- a/sysdeps/x86_64/multiarch/wcsncat-evex.S ++++ b/sysdeps/x86_64/multiarch/wcsncat-evex.S +@@ -1,9 +1,9 @@ +-#ifndef WCSCAT +-# define WCSCAT __wcsncat_evex ++#ifndef WCSNCAT ++# define WCSNCAT __wcsncat_evex + #endif + + #define USE_AS_WCSCPY + #define USE_AS_STRCAT + +-#define STRNCAT WCSCAT ++#define STRNCAT WCSNCAT + #include "strncat-evex.S" diff --git a/SOURCES/glibc-upstream-2.39-56.patch b/SOURCES/glibc-upstream-2.39-56.patch new file mode 100644 index 0000000..1acf952 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-56.patch @@ -0,0 +1,71 @@ +commit 2db79c96baa1256b8fc2656596143da92fabd074 +Author: Sergey Kolosov +Date: Wed Apr 10 17:58:04 2024 +0200 + + libsupport: Add xgetpeername + + The patch adds redirections for getpeername. + + Reviewed-by: Arjun Shankar + (cherry picked from commit 6687a6e3f962759536a8019d31c68c1009ccd6eb) + +diff --git a/support/Makefile b/support/Makefile +index 362a51f882787f2b..aa57207bdccc852d 100644 +--- a/support/Makefile ++++ b/support/Makefile +@@ -131,6 +131,7 @@ libsupport-routines = \ + xfreopen \ + xftruncate \ + xgetline \ ++ xgetpeername \ + xgetsockname \ + xlisten \ + xlseek \ +diff --git a/support/xgetpeername.c b/support/xgetpeername.c +new file mode 100644 +index 0000000000000000..6f448e456a1d9e1e +--- /dev/null ++++ b/support/xgetpeername.c +@@ -0,0 +1,30 @@ ++/* getpeername with error checking. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++ ++#include ++#include ++#include ++ ++void ++xgetpeername (int fd, struct sockaddr *sa, socklen_t *plen) ++{ ++ if (getpeername (fd, sa, plen) != 0) ++ FAIL_EXIT1 ("getpeername (%d): %m", fd); ++} +diff --git a/support/xsocket.h b/support/xsocket.h +index 3e4410354676cb2a..4ac0e1f5ffe35dc2 100644 +--- a/support/xsocket.h ++++ b/support/xsocket.h +@@ -26,6 +26,7 @@ + int xsocket (int, int, int); + void xsetsockopt (int, int, int, const void *, socklen_t); + void xgetsockname (int, struct sockaddr *, socklen_t *); ++void xgetpeername (int, struct sockaddr *, socklen_t *); + void xconnect (int, const struct sockaddr *, socklen_t); + void xbind (int, const struct sockaddr *, socklen_t); + void xlisten (int, int); diff --git a/SOURCES/glibc-upstream-2.39-57.patch b/SOURCES/glibc-upstream-2.39-57.patch new file mode 100644 index 0000000..aee72fa --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-57.patch @@ -0,0 +1,143 @@ +commit 32969a2b36b8cc74343182b768b3babe6f81c3aa +Author: Sergey Kolosov +Date: Wed Apr 10 17:58:05 2024 +0200 + + socket: Add new test for connect + + This commit adds a simple bind/accept/connect test for an IPv4 TCP + connection to a local process via the loopback interface. + + Reviewed-by: Arjun Shankar + (cherry picked from commit 3a83f79024cc023a74c3892a1673542e8e972485) + +diff --git a/socket/Makefile b/socket/Makefile +index 74ca5b8452cd15d3..fc1bd0a2608f04bc 100644 +--- a/socket/Makefile ++++ b/socket/Makefile +@@ -70,6 +70,7 @@ tests := \ + tst-accept4 \ + tst-cmsg_cloexec \ + tst-cmsghdr \ ++ tst-connect \ + tst-sockopt \ + # tests + +diff --git a/socket/tst-connect.c b/socket/tst-connect.c +new file mode 100644 +index 0000000000000000..ec2fdd92c0a6f1be +--- /dev/null ++++ b/socket/tst-connect.c +@@ -0,0 +1,113 @@ ++/* Test the connect function. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++static struct sockaddr_in server_address; ++ ++int ++open_socket_inet_tcp (void) ++{ ++ int fd = socket (AF_INET, SOCK_STREAM, IPPROTO_TCP); ++ if (fd < 0) ++ { ++ if (errno == EAFNOSUPPORT) ++ FAIL_UNSUPPORTED ("The host does not support IPv4"); ++ else ++ FAIL_EXIT1 ("socket (AF_INET, SOCK_STREAM, IPPROTO_TCP): %m\n"); ++ } ++ return fd; ++} ++ ++static pid_t ++start_server (void) ++{ ++ server_address.sin_family = AF_INET; ++ server_address.sin_port = 0; ++ server_address.sin_addr.s_addr = htonl (INADDR_LOOPBACK); ++ ++ int server_sock = open_socket_inet_tcp (); ++ ++ xbind (server_sock, (struct sockaddr *) &server_address, ++ sizeof (server_address)); ++ ++ socklen_t sa_len = sizeof (server_address); ++ xgetsockname (server_sock, (struct sockaddr *) &server_address, &sa_len); ++ xlisten (server_sock, 5); ++ ++ pid_t my_pid = xfork (); ++ if (my_pid > 0) ++ { ++ xclose (server_sock); ++ return my_pid; ++ } ++ ++ struct sockaddr_in client_address; ++ socklen_t ca_len = sizeof (server_address); ++ int client_sock = xaccept (server_sock, (struct sockaddr *) &client_address, ++ &ca_len); ++ printf ("socket accepted %d\n", client_sock); ++ ++ _exit (0); ++} ++ ++static int ++do_test (void) ++{ ++ pid_t serv_pid; ++ struct sockaddr_in peer; ++ socklen_t peer_len; ++ ++ serv_pid = start_server (); ++ int client_sock = open_socket_inet_tcp (); ++ xconnect (client_sock, (const struct sockaddr *) &server_address, ++ sizeof (server_address)); ++ ++ /* A second connect with same arguments should fail with EISCONN. */ ++ int result = connect (client_sock, ++ (const struct sockaddr *) &server_address, ++ sizeof (server_address)); ++ if (result == 0 || errno != EISCONN) ++ FAIL_EXIT1 ("Second connect (%d), should fail with EISCONN: %m", ++ client_sock); ++ ++ peer_len = sizeof (peer); ++ xgetpeername (client_sock, (struct sockaddr *) &peer, &peer_len); ++ TEST_COMPARE (peer_len, sizeof (peer)); ++ TEST_COMPARE (peer.sin_port, server_address.sin_port); ++ TEST_COMPARE_BLOB (&peer.sin_addr, sizeof (peer.sin_addr), ++ &server_address.sin_addr, ++ sizeof (server_address.sin_addr)); ++ ++ int status; ++ xwaitpid (serv_pid, &status, 0); ++ TEST_COMPARE (status, 0); ++ ++ return 0; ++} ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-58.patch b/SOURCES/glibc-upstream-2.39-58.patch new file mode 100644 index 0000000..cbe201a --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-58.patch @@ -0,0 +1,23 @@ +commit c7c3f5bf80ae86b34501f473f1a9fc545c911b7f +Author: caiyinyu +Date: Sat May 11 10:25:54 2024 +0800 + + LoongArch: Fix undefined `__memset_aligned` reference in ld.so linking. + + This patch from 095067efdf68c8061d6f99a21a0300841bede999 (LoongArch: Add + glibc.cpu.hwcap support.) + +diff --git a/sysdeps/loongarch/lp64/multiarch/dl-symbol-redir-ifunc.h b/sysdeps/loongarch/lp64/multiarch/dl-symbol-redir-ifunc.h +index cb640d77b7695b93..a73390b12f67a3fc 100644 +--- a/sysdeps/loongarch/lp64/multiarch/dl-symbol-redir-ifunc.h ++++ b/sysdeps/loongarch/lp64/multiarch/dl-symbol-redir-ifunc.h +@@ -19,6 +19,9 @@ + #ifndef _DL_IFUNC_GENERIC_H + #define _DL_IFUNC_GENERIC_H + ++#ifndef SHARED + asm ("memset = __memset_aligned"); ++asm ("memcmp = __memcmp_aligned"); ++#endif + + #endif diff --git a/SOURCES/glibc-upstream-2.39-59.patch b/SOURCES/glibc-upstream-2.39-59.patch new file mode 100644 index 0000000..0f4fb1c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-59.patch @@ -0,0 +1,40 @@ +commit 9f2b100d6705b9bbb25206b53e80d7759644e06e +Author: H.J. Lu +Date: Sat May 25 05:13:41 2024 -0700 + + parse_fdinfo: Don't advance pointer twice [BZ #31798] + + pidfd_getpid.c has + + /* Ignore invalid large values. */ + if (INT_MULTIPLY_WRAPV (10, n, &n) + || INT_ADD_WRAPV (n, *l++ - '0', &n)) + return -1; + + For GCC older than GCC 7, INT_ADD_WRAPV(a, b, r) is defined as + + _GL_INT_OP_WRAPV (a, b, r, +, _GL_INT_ADD_RANGE_OVERFLOW) + + and *l++ - '0' is evaluated twice. Fix BZ #31798 by moving "l++" out of + the if statement. Tested with GCC 6.4 and GCC 14.1. + + Signed-off-by: H.J. Lu + Reviewed-by: Adhemerval Zanella + (cherry picked from commit f981bf6b9db87e0732b46bfe92fdad4d363225e8) + +diff --git a/sysdeps/unix/sysv/linux/pidfd_getpid.c b/sysdeps/unix/sysv/linux/pidfd_getpid.c +index 8567b413dd2c210b..30025e5863f7b956 100644 +--- a/sysdeps/unix/sysv/linux/pidfd_getpid.c ++++ b/sysdeps/unix/sysv/linux/pidfd_getpid.c +@@ -74,8 +74,10 @@ parse_fdinfo (const char *l, void *arg) + + /* Ignore invalid large values. */ + if (INT_MULTIPLY_WRAPV (10, n, &n) +- || INT_ADD_WRAPV (n, *l++ - '0', &n)) ++ || INT_ADD_WRAPV (n, *l - '0', &n)) + return -1; ++ ++ l++; + } + + /* -1 indicates that the process is terminated. */ diff --git a/SOURCES/glibc-upstream-2.39-6.patch b/SOURCES/glibc-upstream-2.39-6.patch new file mode 100644 index 0000000..61f4133 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-6.patch @@ -0,0 +1,1052 @@ +commit 1b9c1a0047fb26a65a9b2a7b8cd977243f7d353c +Author: Jakub Jelinek +Date: Wed Jan 31 19:17:27 2024 +0100 + + Use gcc __builtin_stdc_* builtins in stdbit.h if possible + + The following patch uses the GCC 14 __builtin_stdc_* builtins in stdbit.h + for the type-generic macros, so that when compiled with GCC 14 or later, + it supports not just 8/16/32/64-bit unsigned integers, but also 128-bit + (if target supports them) and unsigned _BitInt (any supported precision). + And so that the macros don't expand arguments multiple times and can be + evaluated in constant expressions. + + The new testcase is gcc's gcc/testsuite/gcc.dg/builtin-stdc-bit-1.c + adjusted to test stdbit.h and the type-generic macros in there instead + of the builtins and adjusted to use glibc test framework rather than + gcc style tests with __builtin_abort (). + + Signed-off-by: Jakub Jelinek + Reviewed-by: Joseph Myers + (cherry picked from commit da89496337b97e6a2aaf1e81d55cf998f6db1070) + +diff --git a/manual/stdbit.texi b/manual/stdbit.texi +index fe41c671d8f8f7d9..6c75ed9a20444132 100644 +--- a/manual/stdbit.texi ++++ b/manual/stdbit.texi +@@ -32,7 +32,13 @@ and @code{unsigned long long int}. In addition, there is a + corresponding type-generic macro (not listed below), named the same as + the functions but without any suffix such as @samp{_uc}. The + type-generic macro can only be used with an argument of an unsigned +-integer type with a width of 8, 16, 32 or 64 bits. ++integer type with a width of 8, 16, 32 or 64 bits, or when using ++a compiler with support for ++@uref{https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html,@code{__builtin_stdc_bit_ceil}}, ++etc.@:, built-in functions such as GCC 14.1 or later ++any unsigned integer type those built-in functions support. ++In GCC 14.1 that includes support for @code{unsigned __int128} and ++@code{unsigned _BitInt(@var{n})} if supported by the target. + + @deftypefun {unsigned int} stdc_leading_zeros_uc (unsigned char @var{x}) + @deftypefunx {unsigned int} stdc_leading_zeros_us (unsigned short @var{x}) +diff --git a/stdlib/Makefile b/stdlib/Makefile +index d587f054d198a173..9898cc5d8a560625 100644 +--- a/stdlib/Makefile ++++ b/stdlib/Makefile +@@ -308,6 +308,7 @@ tests := \ + tst-setcontext10 \ + tst-setcontext11 \ + tst-stdbit-Wconversion \ ++ tst-stdbit-builtins \ + tst-stdc_bit_ceil \ + tst-stdc_bit_floor \ + tst-stdc_bit_width \ +diff --git a/stdlib/stdbit.h b/stdlib/stdbit.h +index f334eb174d209c82..2801590c63006ad2 100644 +--- a/stdlib/stdbit.h ++++ b/stdlib/stdbit.h +@@ -64,9 +64,13 @@ extern unsigned int stdc_leading_zeros_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_leading_zeros_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_leading_zeros(x) \ ++#if __glibc_has_builtin (__builtin_stdc_leading_zeros) ++# define stdc_leading_zeros(x) (__builtin_stdc_leading_zeros (x)) ++#else ++# define stdc_leading_zeros(x) \ + (stdc_leading_zeros_ull (x) \ + - (unsigned int) (8 * (sizeof (0ULL) - sizeof (x)))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline unsigned int +@@ -116,9 +120,13 @@ extern unsigned int stdc_leading_ones_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_leading_ones_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_leading_ones(x) \ ++#if __glibc_has_builtin (__builtin_stdc_leading_ones) ++# define stdc_leading_ones(x) (__builtin_stdc_leading_ones (x)) ++#else ++# define stdc_leading_ones(x) \ + (stdc_leading_ones_ull ((unsigned long long int) (x) \ + << 8 * (sizeof (0ULL) - sizeof (x)))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline unsigned int +@@ -168,11 +176,15 @@ extern unsigned int stdc_trailing_zeros_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_trailing_zeros_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_trailing_zeros(x) \ ++#if __glibc_has_builtin (__builtin_stdc_trailing_zeros) ++# define stdc_trailing_zeros(x) (__builtin_stdc_trailing_zeros (x)) ++#else ++# define stdc_trailing_zeros(x) \ + (sizeof (x) == 8 ? stdc_trailing_zeros_ull (x) \ + : sizeof (x) == 4 ? stdc_trailing_zeros_ui (x) \ + : sizeof (x) == 2 ? stdc_trailing_zeros_us (__pacify_uint16 (x)) \ + : stdc_trailing_zeros_uc (__pacify_uint8 (x))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_ctzll) + static __always_inline unsigned int +@@ -222,7 +234,11 @@ extern unsigned int stdc_trailing_ones_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_trailing_ones_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_trailing_ones(x) (stdc_trailing_ones_ull (x)) ++#if __glibc_has_builtin (__builtin_stdc_trailing_ones) ++# define stdc_trailing_ones(x) (__builtin_stdc_trailing_ones (x)) ++#else ++# define stdc_trailing_ones(x) (stdc_trailing_ones_ull (x)) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_ctzll) + static __always_inline unsigned int +@@ -272,11 +288,15 @@ extern unsigned int stdc_first_leading_zero_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_first_leading_zero_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_first_leading_zero(x) \ ++#if __glibc_has_builtin (__builtin_stdc_first_leading_zero) ++# define stdc_first_leading_zero(x) (__builtin_stdc_first_leading_zero (x)) ++#else ++# define stdc_first_leading_zero(x) \ + (sizeof (x) == 8 ? stdc_first_leading_zero_ull (x) \ + : sizeof (x) == 4 ? stdc_first_leading_zero_ui (x) \ + : sizeof (x) == 2 ? stdc_first_leading_zero_us (__pacify_uint16 (x)) \ + : stdc_first_leading_zero_uc (__pacify_uint8 (x))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline unsigned int +@@ -326,11 +346,15 @@ extern unsigned int stdc_first_leading_one_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_first_leading_one_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_first_leading_one(x) \ ++#if __glibc_has_builtin (__builtin_stdc_first_leading_one) ++# define stdc_first_leading_one(x) (__builtin_stdc_first_leading_one (x)) ++#else ++# define stdc_first_leading_one(x) \ + (sizeof (x) == 8 ? stdc_first_leading_one_ull (x) \ + : sizeof (x) == 4 ? stdc_first_leading_one_ui (x) \ + : sizeof (x) == 2 ? stdc_first_leading_one_us (__pacify_uint16 (x)) \ + : stdc_first_leading_one_uc (__pacify_uint8 (x))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline unsigned int +@@ -380,11 +404,15 @@ extern unsigned int stdc_first_trailing_zero_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_first_trailing_zero_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_first_trailing_zero(x) \ ++#if __glibc_has_builtin (__builtin_stdc_first_trailing_zero) ++# define stdc_first_trailing_zero(x) (__builtin_stdc_first_trailing_zero (x)) ++#else ++# define stdc_first_trailing_zero(x) \ + (sizeof (x) == 8 ? stdc_first_trailing_zero_ull (x) \ + : sizeof (x) == 4 ? stdc_first_trailing_zero_ui (x) \ + : sizeof (x) == 2 ? stdc_first_trailing_zero_us (__pacify_uint16 (x)) \ + : stdc_first_trailing_zero_uc (__pacify_uint8 (x))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_ctzll) + static __always_inline unsigned int +@@ -434,11 +462,15 @@ extern unsigned int stdc_first_trailing_one_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_first_trailing_one_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_first_trailing_one(x) \ ++#if __glibc_has_builtin (__builtin_stdc_first_trailing_one) ++# define stdc_first_trailing_one(x) (__builtin_stdc_first_trailing_one (x)) ++#else ++# define stdc_first_trailing_one(x) \ + (sizeof (x) == 8 ? stdc_first_trailing_one_ull (x) \ + : sizeof (x) == 4 ? stdc_first_trailing_one_ui (x) \ + : sizeof (x) == 2 ? stdc_first_trailing_one_us (__pacify_uint16 (x)) \ + : stdc_first_trailing_one_uc (__pacify_uint8 (x))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_ctzll) + static __always_inline unsigned int +@@ -488,9 +520,13 @@ extern unsigned int stdc_count_zeros_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_count_zeros_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_count_zeros(x) \ ++#if __glibc_has_builtin (__builtin_stdc_count_zeros) ++# define stdc_count_zeros(x) (__builtin_stdc_count_zeros (x)) ++#else ++# define stdc_count_zeros(x) \ + (stdc_count_zeros_ull (x) \ + - (unsigned int) (8 * (sizeof (0ULL) - sizeof (x)))) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_popcountll) + static __always_inline unsigned int +@@ -540,7 +576,11 @@ extern unsigned int stdc_count_ones_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_count_ones_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_count_ones(x) (stdc_count_ones_ull (x)) ++#if __glibc_has_builtin (__builtin_stdc_count_ones) ++# define stdc_count_ones(x) (__builtin_stdc_count_ones (x)) ++#else ++# define stdc_count_ones(x) (stdc_count_ones_ull (x)) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_popcountll) + static __always_inline unsigned int +@@ -590,10 +630,14 @@ extern bool stdc_has_single_bit_ul (unsigned long int __x) + __extension__ + extern bool stdc_has_single_bit_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_has_single_bit(x) \ ++#if __glibc_has_builtin (__builtin_stdc_has_single_bit) ++# define stdc_has_single_bit(x) (__builtin_stdc_has_single_bit (x)) ++#else ++# define stdc_has_single_bit(x) \ + ((bool) (sizeof (x) <= sizeof (unsigned int) \ + ? stdc_has_single_bit_ui (x) \ + : stdc_has_single_bit_ull (x))) ++#endif + + static __always_inline bool + __hsb64_inline (uint64_t __x) +@@ -641,7 +685,11 @@ extern unsigned int stdc_bit_width_ul (unsigned long int __x) + __extension__ + extern unsigned int stdc_bit_width_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_bit_width(x) (stdc_bit_width_ull (x)) ++#if __glibc_has_builtin (__builtin_stdc_bit_width) ++# define stdc_bit_width(x) (__builtin_stdc_bit_width (x)) ++#else ++# define stdc_bit_width(x) (stdc_bit_width_ull (x)) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline unsigned int +@@ -691,7 +739,11 @@ extern unsigned long int stdc_bit_floor_ul (unsigned long int __x) + __extension__ + extern unsigned long long int stdc_bit_floor_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_bit_floor(x) ((__typeof (x)) stdc_bit_floor_ull (x)) ++#if __glibc_has_builtin (__builtin_stdc_bit_floor) ++# define stdc_bit_floor(x) (__builtin_stdc_bit_floor (x)) ++#else ++# define stdc_bit_floor(x) ((__typeof (x)) stdc_bit_floor_ull (x)) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline uint64_t +@@ -743,7 +795,11 @@ extern unsigned long int stdc_bit_ceil_ul (unsigned long int __x) + __extension__ + extern unsigned long long int stdc_bit_ceil_ull (unsigned long long int __x) + __THROW __attribute_const__; +-#define stdc_bit_ceil(x) ((__typeof (x)) stdc_bit_ceil_ull (x)) ++#if __glibc_has_builtin (__builtin_stdc_bit_ceil) ++# define stdc_bit_ceil(x) (__builtin_stdc_bit_ceil (x)) ++#else ++# define stdc_bit_ceil(x) ((__typeof (x)) stdc_bit_ceil_ull (x)) ++#endif + + #if __GNUC_PREREQ (3, 4) || __glibc_has_builtin (__builtin_clzll) + static __always_inline uint64_t +diff --git a/stdlib/tst-stdbit-builtins.c b/stdlib/tst-stdbit-builtins.c +new file mode 100644 +index 0000000000000000..536841ca8ac3425b +--- /dev/null ++++ b/stdlib/tst-stdbit-builtins.c +@@ -0,0 +1,778 @@ ++/* Test type-generic macros with compiler __builtin_stdc_* support. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++ ++#if __glibc_has_builtin (__builtin_stdc_leading_zeros) \ ++ && __glibc_has_builtin (__builtin_stdc_leading_ones) \ ++ && __glibc_has_builtin (__builtin_stdc_trailing_zeros) \ ++ && __glibc_has_builtin (__builtin_stdc_trailing_ones) \ ++ && __glibc_has_builtin (__builtin_stdc_first_leading_zero) \ ++ && __glibc_has_builtin (__builtin_stdc_first_leading_one) \ ++ && __glibc_has_builtin (__builtin_stdc_first_trailing_zero) \ ++ && __glibc_has_builtin (__builtin_stdc_first_trailing_one) \ ++ && __glibc_has_builtin (__builtin_stdc_count_zeros) \ ++ && __glibc_has_builtin (__builtin_stdc_count_ones) \ ++ && __glibc_has_builtin (__builtin_stdc_has_single_bit) \ ++ && __glibc_has_builtin (__builtin_stdc_bit_width) \ ++ && __glibc_has_builtin (__builtin_stdc_bit_floor) \ ++ && __glibc_has_builtin (__builtin_stdc_bit_ceil) ++ ++# if !defined (BITINT_MAXWIDTH) && defined (__BITINT_MAXWIDTH__) ++# define BITINT_MAXWIDTH __BITINT_MAXWIDTH__ ++# endif ++ ++typedef unsigned char uc; ++typedef unsigned short us; ++typedef unsigned int ui; ++typedef unsigned long int ul; ++typedef unsigned long long int ull; ++ ++# define expr_has_type(e, t) _Generic (e, default : 0, t : 1) ++ ++static int ++do_test (void) ++{ ++ TEST_COMPARE (stdc_leading_zeros ((uc) 0), CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros ((us) 0), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros (0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros (0U), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros (0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros (0UL), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros (0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros (0ULL), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_leading_zeros ((us) ~0U), 0); ++ TEST_COMPARE (stdc_leading_zeros (~0U), 0); ++ TEST_COMPARE (stdc_leading_zeros (~0UL), 0); ++ TEST_COMPARE (stdc_leading_zeros (~0ULL), 0); ++ TEST_COMPARE (stdc_leading_zeros ((uc) 3), CHAR_BIT - 2); ++ TEST_COMPARE (stdc_leading_zeros ((us) 9), sizeof (short) * CHAR_BIT - 4); ++ TEST_COMPARE (stdc_leading_zeros (34U), sizeof (int) * CHAR_BIT - 6); ++ TEST_COMPARE (stdc_leading_zeros (130UL), sizeof (long int) * CHAR_BIT - 8); ++ TEST_COMPARE (stdc_leading_zeros (512ULL), ++ sizeof (long long int) * CHAR_BIT - 10); ++ TEST_COMPARE (stdc_leading_ones ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_leading_ones ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_leading_ones (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones (0U), ui), 1); ++ TEST_COMPARE (stdc_leading_ones (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones (0UL), ui), 1); ++ TEST_COMPARE (stdc_leading_ones (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones (0ULL), ui), 1); ++ TEST_COMPARE (stdc_leading_ones ((uc) ~0U), CHAR_BIT); ++ TEST_COMPARE (stdc_leading_ones ((us) ~0U), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (stdc_leading_ones (~0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (stdc_leading_ones (~0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_leading_ones (~0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_leading_ones ((uc) ~3), CHAR_BIT - 2); ++ TEST_COMPARE (stdc_leading_ones ((us) ~9), sizeof (short) * CHAR_BIT - 4); ++ TEST_COMPARE (stdc_leading_ones (~34U), sizeof (int) * CHAR_BIT - 6); ++ TEST_COMPARE (stdc_leading_ones (~130UL), sizeof (long int) * CHAR_BIT - 8); ++ TEST_COMPARE (stdc_leading_ones (~512ULL), ++ sizeof (long long int) * CHAR_BIT - 10); ++ TEST_COMPARE (stdc_trailing_zeros ((uc) 0), CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros ((us) 0), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros (0U), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros (0UL), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros (0ULL), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_trailing_zeros ((us) ~0U), 0); ++ TEST_COMPARE (stdc_trailing_zeros (~0U), 0); ++ TEST_COMPARE (stdc_trailing_zeros (~0UL), 0); ++ TEST_COMPARE (stdc_trailing_zeros (~0ULL), 0); ++ TEST_COMPARE (stdc_trailing_zeros ((uc) 2), 1); ++ TEST_COMPARE (stdc_trailing_zeros ((us) 24), 3); ++ TEST_COMPARE (stdc_trailing_zeros (32U), 5); ++ TEST_COMPARE (stdc_trailing_zeros (128UL), 7); ++ TEST_COMPARE (stdc_trailing_zeros (512ULL), 9); ++ TEST_COMPARE (stdc_trailing_ones ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones (0U), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones (0UL), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones (0ULL), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones ((uc) ~0U), CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_ones ((us) ~0U), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_ones (~0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_ones (~0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_ones (~0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_ones ((uc) 5), 1); ++ TEST_COMPARE (stdc_trailing_ones ((us) 15), 4); ++ TEST_COMPARE (stdc_trailing_ones (127U), 7); ++ TEST_COMPARE (stdc_trailing_ones (511UL), 9); ++ TEST_COMPARE (stdc_trailing_ones (~0ULL >> 2), ++ sizeof (long long int) * CHAR_BIT - 2); ++ TEST_COMPARE (stdc_first_leading_zero ((uc) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero ((us) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (0U), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero (0U), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (0UL), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero (0UL), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (0ULL), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero (0ULL), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_first_leading_zero ((us) ~0U), 0); ++ TEST_COMPARE (stdc_first_leading_zero (~0U), 0); ++ TEST_COMPARE (stdc_first_leading_zero (~0UL), 0); ++ TEST_COMPARE (stdc_first_leading_zero (~0ULL), 0); ++ TEST_COMPARE (stdc_first_leading_zero ((uc) ~3U), CHAR_BIT - 1); ++ TEST_COMPARE (stdc_first_leading_zero ((us) ~15U), ++ sizeof (short) * CHAR_BIT - 3); ++ TEST_COMPARE (stdc_first_leading_zero (~63U), sizeof (int) * CHAR_BIT - 5); ++ TEST_COMPARE (stdc_first_leading_zero (~255UL), ++ sizeof (long int) * CHAR_BIT - 7); ++ TEST_COMPARE (stdc_first_leading_zero (~1023ULL), ++ sizeof (long long int) * CHAR_BIT - 9); ++ TEST_COMPARE (stdc_first_leading_one ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one (0U), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one (0UL), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one (0ULL), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one ((uc) ~0U), 1); ++ TEST_COMPARE (stdc_first_leading_one ((us) ~0U), 1); ++ TEST_COMPARE (stdc_first_leading_one (~0U), 1); ++ TEST_COMPARE (stdc_first_leading_one (~0UL), 1); ++ TEST_COMPARE (stdc_first_leading_one (~0ULL), 1); ++ TEST_COMPARE (stdc_first_leading_one ((uc) 3), CHAR_BIT - 1); ++ TEST_COMPARE (stdc_first_leading_one ((us) 9), ++ sizeof (short) * CHAR_BIT - 3); ++ TEST_COMPARE (stdc_first_leading_one (34U), sizeof (int) * CHAR_BIT - 5); ++ TEST_COMPARE (stdc_first_leading_one (130UL), ++ sizeof (long int) * CHAR_BIT - 7); ++ TEST_COMPARE (stdc_first_leading_one (512ULL), ++ sizeof (long long int) * CHAR_BIT - 9); ++ TEST_COMPARE (stdc_first_trailing_zero ((uc) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero ((us) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (0U), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero (0U), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (0UL), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero (0UL), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (0ULL), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero (0ULL), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_first_trailing_zero ((us) ~0U), 0); ++ TEST_COMPARE (stdc_first_trailing_zero (~0U), 0); ++ TEST_COMPARE (stdc_first_trailing_zero (~0UL), 0); ++ TEST_COMPARE (stdc_first_trailing_zero (~0ULL), 0); ++ TEST_COMPARE (stdc_first_trailing_zero ((uc) 2), 1); ++ TEST_COMPARE (stdc_first_trailing_zero ((us) 15), 5); ++ TEST_COMPARE (stdc_first_trailing_zero (63U), 7); ++ TEST_COMPARE (stdc_first_trailing_zero (128UL), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (511ULL), 10); ++ TEST_COMPARE (stdc_first_trailing_one ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one (0U), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one (0UL), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one (0ULL), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one ((uc) ~0U), 1); ++ TEST_COMPARE (stdc_first_trailing_one ((us) ~0U), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~0U), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~0UL), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~0ULL), 1); ++ TEST_COMPARE (stdc_first_trailing_one ((uc) 4), 3); ++ TEST_COMPARE (stdc_first_trailing_one ((us) 96), 6); ++ TEST_COMPARE (stdc_first_trailing_one (127U), 1); ++ TEST_COMPARE (stdc_first_trailing_one (511UL), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~0ULL << 12), 13); ++ TEST_COMPARE (stdc_count_zeros ((uc) 0), CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_count_zeros ((us) 0), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_count_zeros (0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros (0U), ui), 1); ++ TEST_COMPARE (stdc_count_zeros (0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros (0UL), ui), 1); ++ TEST_COMPARE (stdc_count_zeros (0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros (0ULL), ui), 1); ++ TEST_COMPARE (stdc_count_zeros ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_count_zeros ((us) ~0U), 0); ++ TEST_COMPARE (stdc_count_zeros (~0U), 0); ++ TEST_COMPARE (stdc_count_zeros (~0UL), 0); ++ TEST_COMPARE (stdc_count_zeros (~0ULL), 0); ++ TEST_COMPARE (stdc_count_zeros ((uc) 1U), CHAR_BIT - 1); ++ TEST_COMPARE (stdc_count_zeros ((us) 42), sizeof (short) * CHAR_BIT - 3); ++ TEST_COMPARE (stdc_count_zeros (291U), sizeof (int) * CHAR_BIT - 4); ++ TEST_COMPARE (stdc_count_zeros (~1315UL), 5); ++ TEST_COMPARE (stdc_count_zeros (3363ULL), ++ sizeof (long long int) * CHAR_BIT - 6); ++ TEST_COMPARE (stdc_count_ones ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_count_ones ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_count_ones (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones (0U), ui), 1); ++ TEST_COMPARE (stdc_count_ones (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones (0UL), ui), 1); ++ TEST_COMPARE (stdc_count_ones (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones (0ULL), ui), 1); ++ TEST_COMPARE (stdc_count_ones ((uc) ~0U), CHAR_BIT); ++ TEST_COMPARE (stdc_count_ones ((us) ~0U), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (stdc_count_ones (~0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (stdc_count_ones (~0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_count_ones (~0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_count_ones ((uc) ~1U), CHAR_BIT - 1); ++ TEST_COMPARE (stdc_count_ones ((us) ~42), sizeof (short) * CHAR_BIT - 3); ++ TEST_COMPARE (stdc_count_ones (~291U), sizeof (int) * CHAR_BIT - 4); ++ TEST_COMPARE (stdc_count_ones (1315UL), 5); ++ TEST_COMPARE (stdc_count_ones (~3363ULL), ++ sizeof (long long int) * CHAR_BIT - 6); ++ TEST_COMPARE (stdc_has_single_bit ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit ((uc) 0), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit ((us) 0), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit (0U), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit (0UL), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit (0ULL), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit ((uc) 2), 1); ++ TEST_COMPARE (stdc_has_single_bit ((us) 8), 1); ++ TEST_COMPARE (stdc_has_single_bit (32U), 1); ++ TEST_COMPARE (stdc_has_single_bit (128UL), 1); ++ TEST_COMPARE (stdc_has_single_bit (512ULL), 1); ++ TEST_COMPARE (stdc_has_single_bit ((uc) 7), 0); ++ TEST_COMPARE (stdc_has_single_bit ((us) 96), 0); ++ TEST_COMPARE (stdc_has_single_bit (513U), 0); ++ TEST_COMPARE (stdc_has_single_bit (1022UL), 0); ++ TEST_COMPARE (stdc_has_single_bit (12ULL), 0); ++ TEST_COMPARE (stdc_bit_width ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width ((uc) 0), ui), 1); ++ TEST_COMPARE (stdc_bit_width ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width ((us) 0), ui), 1); ++ TEST_COMPARE (stdc_bit_width (0U), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width (0U), ui), 1); ++ TEST_COMPARE (stdc_bit_width (0UL), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width (0UL), ui), 1); ++ TEST_COMPARE (stdc_bit_width (0ULL), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width (0ULL), ui), 1); ++ TEST_COMPARE (stdc_bit_width ((uc) ~0U), CHAR_BIT); ++ TEST_COMPARE (stdc_bit_width ((us) ~0U), sizeof (short) * CHAR_BIT); ++ TEST_COMPARE (stdc_bit_width (~0U), sizeof (int) * CHAR_BIT); ++ TEST_COMPARE (stdc_bit_width (~0UL), sizeof (long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_bit_width (~0ULL), sizeof (long long int) * CHAR_BIT); ++ TEST_COMPARE (stdc_bit_width ((uc) ((uc) ~0U >> 1)), CHAR_BIT - 1); ++ TEST_COMPARE (stdc_bit_width ((uc) 6), 3); ++ TEST_COMPARE (stdc_bit_width ((us) 12U), 4); ++ TEST_COMPARE (stdc_bit_width ((us) ((us) ~0U >> 5)), ++ sizeof (short) * CHAR_BIT - 5); ++ TEST_COMPARE (stdc_bit_width (137U), 8); ++ TEST_COMPARE (stdc_bit_width (269U), 9); ++ TEST_COMPARE (stdc_bit_width (39UL), 6); ++ TEST_COMPARE (stdc_bit_width (~0UL >> 2), sizeof (long int) * CHAR_BIT - 2); ++ TEST_COMPARE (stdc_bit_width (1023ULL), 10); ++ TEST_COMPARE (stdc_bit_width (1024ULL), 11); ++ TEST_COMPARE (stdc_bit_floor ((uc) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor ((uc) 0), uc), 1); ++ TEST_COMPARE (stdc_bit_floor ((us) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor ((us) 0), us), 1); ++ TEST_COMPARE (stdc_bit_floor (0U), 0U); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor (0U), ui), 1); ++ TEST_COMPARE (stdc_bit_floor (0UL), 0UL); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor (0UL), ul), 1); ++ TEST_COMPARE (stdc_bit_floor (0ULL), 0ULL); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor (0ULL), ull), 1); ++ TEST_COMPARE (stdc_bit_floor ((uc) ~0U), (1U << (CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_floor ((us) ~0U), ++ (1U << (sizeof (short) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_floor (~0U), (1U << (sizeof (int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_floor (~0UL), ++ (1UL << (sizeof (long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_floor (~0ULL), ++ (1ULL << (sizeof (long long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_floor ((uc) 4), 4); ++ TEST_COMPARE (stdc_bit_floor ((uc) 7), 4); ++ TEST_COMPARE (stdc_bit_floor ((us) 8U), 8); ++ TEST_COMPARE (stdc_bit_floor ((us) 31U), 16); ++ TEST_COMPARE (stdc_bit_floor (137U), 128U); ++ TEST_COMPARE (stdc_bit_floor (269U), 256U); ++ TEST_COMPARE (stdc_bit_floor (511UL), 256UL); ++ TEST_COMPARE (stdc_bit_floor (512UL), 512UL); ++ TEST_COMPARE (stdc_bit_floor (513UL), 512ULL); ++ TEST_COMPARE (stdc_bit_floor (1024ULL), 1024ULL); ++ TEST_COMPARE (stdc_bit_ceil ((uc) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil ((uc) 0), uc), 1); ++ TEST_COMPARE (stdc_bit_ceil ((us) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil ((us) 0), us), 1); ++ TEST_COMPARE (stdc_bit_ceil (0U), 1U); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil (0U), ui), 1); ++ TEST_COMPARE (stdc_bit_ceil (0UL), 1UL); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil (0UL), ul), 1); ++ TEST_COMPARE (stdc_bit_ceil (0ULL), 1ULL); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil (0ULL), ull), 1); ++ TEST_COMPARE (stdc_bit_ceil ((uc) ~0U), 0); ++ TEST_COMPARE (stdc_bit_ceil ((us) ~0U), 0); ++ TEST_COMPARE (stdc_bit_ceil (~0U), 0U); ++ TEST_COMPARE (stdc_bit_ceil (~0UL), 0UL); ++ TEST_COMPARE (stdc_bit_ceil (~0ULL), 0ULL); ++ TEST_COMPARE (stdc_bit_ceil ((uc) ((uc) ~0U >> 1)), (1U << (CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil ((uc) ((uc) ~0U >> 1)), (1U << (CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil ((us) ((us) ~0U >> 1)), ++ (1U << (sizeof (short) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil ((us) ((us) ~0U >> 1)), ++ (1U << (sizeof (short) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (~0U >> 1), ++ (1U << (sizeof (int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (1U << (sizeof (int) * CHAR_BIT - 1)), ++ (1U << (sizeof (int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (~0UL >> 1), ++ (1UL << (sizeof (long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (~0UL >> 1), ++ (1UL << (sizeof (long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (1ULL ++ << (sizeof (long long int) * CHAR_BIT - 1)), ++ (1ULL << (sizeof (long long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil (~0ULL >> 1), ++ (1ULL << (sizeof (long long int) * CHAR_BIT - 1))); ++ TEST_COMPARE (stdc_bit_ceil ((uc) 1), 1); ++ TEST_COMPARE (stdc_bit_ceil ((uc) 2), 2); ++ TEST_COMPARE (stdc_bit_ceil ((us) 3U), 4); ++ TEST_COMPARE (stdc_bit_ceil ((us) 4U), 4); ++ TEST_COMPARE (stdc_bit_ceil (5U), 8U); ++ TEST_COMPARE (stdc_bit_ceil (269U), 512U); ++ TEST_COMPARE (stdc_bit_ceil (511UL), 512UL); ++ TEST_COMPARE (stdc_bit_ceil (512UL), 512UL); ++ TEST_COMPARE (stdc_bit_ceil (513ULL), 1024ULL); ++ TEST_COMPARE (stdc_bit_ceil (1025ULL), 2048ULL); ++# ifdef __SIZEOF_INT128__ ++ TEST_COMPARE (stdc_leading_zeros ((unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros ((unsigned __int128) 0), ui), ++ 1); ++ TEST_COMPARE (stdc_leading_zeros (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_leading_ones ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones ((unsigned __int128) 0), ui), ++ 1); ++ TEST_COMPARE (stdc_leading_ones (~(unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (stdc_trailing_zeros ((unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros ((unsigned __int128) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_trailing_ones ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones ((unsigned __int128) 0), ui), ++ 1); ++ TEST_COMPARE (stdc_trailing_ones (~(unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (stdc_first_leading_zero ((unsigned __int128) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero ((unsigned __int128) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_first_leading_one ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one ((unsigned __int128) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (~(unsigned __int128) 0), 1); ++ TEST_COMPARE (stdc_first_trailing_zero ((unsigned __int128) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero ((unsigned __int128) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_first_trailing_one ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one ((unsigned __int128) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~(unsigned __int128) 0), 1); ++ TEST_COMPARE (stdc_count_zeros ((unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros ((unsigned __int128) 0), ui), ++ 1); ++ TEST_COMPARE (stdc_count_zeros (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_count_ones ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones ((unsigned __int128) 0), ui), ++ 1); ++ TEST_COMPARE (stdc_count_ones (~(unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (stdc_has_single_bit ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit ((unsigned __int128) 0), ++ _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (~(unsigned __int128) 0), 0); ++ TEST_COMPARE (stdc_bit_width ((unsigned __int128) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width ((unsigned __int128) 0), ui), 1); ++ TEST_COMPARE (stdc_bit_width (~(unsigned __int128) 0), ++ sizeof (__int128) * CHAR_BIT); ++ TEST_COMPARE (stdc_bit_floor ((unsigned __int128) 0) != 0, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor ((unsigned __int128) 0), ++ unsigned __int128), 1); ++ TEST_COMPARE (stdc_bit_floor (~(unsigned __int128) 0) ++ != ((unsigned __int128) 1) << (sizeof (__int128) ++ * CHAR_BIT - 1), 0); ++ TEST_COMPARE (stdc_bit_ceil ((unsigned __int128) 0) != 1, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil ((unsigned __int128) 0), ++ unsigned __int128), 1); ++ TEST_COMPARE (stdc_bit_ceil ((unsigned __int128) 1) != 1, 0); ++ TEST_COMPARE (stdc_bit_ceil ((~(unsigned __int128) 0) >> 1) ++ != ((unsigned __int128) 1) << (sizeof (__int128) ++ * CHAR_BIT - 1), 0); ++ TEST_COMPARE (stdc_bit_ceil (~(unsigned __int128) 0) != 0, 0); ++# endif ++ uc a = 0; ++ TEST_COMPARE (stdc_bit_width (a++), 0); ++ TEST_COMPARE (a, 1); ++ ull b = 0; ++ TEST_COMPARE (stdc_bit_width (b++), 0); ++ TEST_COMPARE (b, 1); ++ TEST_COMPARE (stdc_bit_floor (a++), 1); ++ TEST_COMPARE (a, 2); ++ TEST_COMPARE (stdc_bit_floor (b++), 1); ++ TEST_COMPARE (b, 2); ++ TEST_COMPARE (stdc_bit_ceil (a++), 2); ++ TEST_COMPARE (a, 3); ++ TEST_COMPARE (stdc_bit_ceil (b++), 2); ++ TEST_COMPARE (b, 3); ++ TEST_COMPARE (stdc_leading_zeros (a++), CHAR_BIT - 2); ++ TEST_COMPARE (a, 4); ++ TEST_COMPARE (stdc_leading_zeros (b++), ++ sizeof (long long int) * CHAR_BIT - 2); ++ TEST_COMPARE (b, 4); ++ TEST_COMPARE (stdc_leading_ones (a++), 0); ++ TEST_COMPARE (a, 5); ++ TEST_COMPARE (stdc_leading_ones (b++), 0); ++ TEST_COMPARE (b, 5); ++ TEST_COMPARE (stdc_trailing_zeros (a++), 0); ++ TEST_COMPARE (a, 6); ++ TEST_COMPARE (stdc_trailing_zeros (b++), 0); ++ TEST_COMPARE (b, 6); ++ TEST_COMPARE (stdc_trailing_ones (a++), 0); ++ TEST_COMPARE (a, 7); ++ TEST_COMPARE (stdc_trailing_ones (b++), 0); ++ TEST_COMPARE (b, 7); ++ TEST_COMPARE (stdc_first_leading_zero (a++), 1); ++ TEST_COMPARE (a, 8); ++ TEST_COMPARE (stdc_first_leading_zero (b++), 1); ++ TEST_COMPARE (b, 8); ++ TEST_COMPARE (stdc_first_leading_one (a++), CHAR_BIT - 3); ++ TEST_COMPARE (a, 9); ++ TEST_COMPARE (stdc_first_leading_one (b++), ++ sizeof (long long int) * CHAR_BIT - 3); ++ TEST_COMPARE (b, 9); ++ TEST_COMPARE (stdc_first_trailing_zero (a++), 2); ++ TEST_COMPARE (a, 10); ++ TEST_COMPARE (stdc_first_trailing_zero (b++), 2); ++ TEST_COMPARE (b, 10); ++ TEST_COMPARE (stdc_first_trailing_one (a++), 2); ++ TEST_COMPARE (a, 11); ++ TEST_COMPARE (stdc_first_trailing_one (b++), 2); ++ TEST_COMPARE (b, 11); ++ TEST_COMPARE (stdc_count_zeros (a++), CHAR_BIT - 3); ++ TEST_COMPARE (a, 12); ++ TEST_COMPARE (stdc_count_zeros (b++), ++ sizeof (long long int) * CHAR_BIT - 3); ++ TEST_COMPARE (b, 12); ++ TEST_COMPARE (stdc_count_ones (a++), 2); ++ TEST_COMPARE (a, 13); ++ TEST_COMPARE (stdc_count_ones (b++), 2); ++ TEST_COMPARE (b, 13); ++ TEST_COMPARE (stdc_has_single_bit (a++), 0); ++ TEST_COMPARE (a, 14); ++ TEST_COMPARE (stdc_has_single_bit (b++), 0); ++ TEST_COMPARE (b, 14); ++# ifdef BITINT_MAXWIDTH ++# if BITINT_MAXWIDTH >= 64 ++ TEST_COMPARE (stdc_leading_zeros (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros (0uwb), ui), 1); ++ TEST_COMPARE (stdc_leading_zeros (1uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros (1uwb), ui), 1); ++ TEST_COMPARE (stdc_leading_ones (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones (0uwb), ui), 1); ++ TEST_COMPARE (stdc_leading_ones (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones (1uwb), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros (0uwb), ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (1uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros (1uwb), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones (0uwb), ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones (1uwb), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero (0uwb), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (1uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero (1uwb), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one (0uwb), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one (1uwb), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero (0uwb), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (1uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero (1uwb), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one (0uwb), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one (1uwb), ui), 1); ++ TEST_COMPARE (stdc_count_zeros (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros (0uwb), ui), 1); ++ TEST_COMPARE (stdc_count_zeros (1uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros (1uwb), ui), 1); ++ TEST_COMPARE (stdc_count_ones (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones (0uwb), ui), 1); ++ TEST_COMPARE (stdc_count_ones (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_count_ones (1uwb), ui), 1); ++ TEST_COMPARE (stdc_has_single_bit (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit (0uwb), _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit (1uwb), _Bool), 1); ++ TEST_COMPARE (stdc_bit_width (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width (0uwb), ui), 1); ++ TEST_COMPARE (stdc_bit_width (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_width (1uwb), ui), 1); ++ TEST_COMPARE (stdc_bit_floor (0uwb), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor (0uwb), unsigned _BitInt(1)), 1); ++ TEST_COMPARE (stdc_bit_floor (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor (1uwb), unsigned _BitInt(1)), 1); ++ TEST_COMPARE (stdc_bit_ceil (0uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil (0uwb), unsigned _BitInt(1)), 1); ++ TEST_COMPARE (stdc_bit_ceil (1uwb), 1); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil (1uwb), unsigned _BitInt(1)), 1); ++ unsigned _BitInt(1) c = 0; ++ TEST_COMPARE (stdc_bit_floor (c++), 0); ++ TEST_COMPARE (c, 1); ++ TEST_COMPARE (stdc_bit_floor (c++), 1); ++ TEST_COMPARE (c, 0); ++ TEST_COMPARE (stdc_bit_ceil (c++), 1); ++ TEST_COMPARE (c, 1); ++ TEST_COMPARE (stdc_bit_ceil (c++), 1); ++ TEST_COMPARE (c, 0); ++# endif ++# if BITINT_MAXWIDTH >= 512 ++ TEST_COMPARE (stdc_leading_zeros ((unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_leading_zeros ((unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (expr_has_type (stdc_leading_zeros ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_leading_zeros (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_leading_zeros (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_leading_zeros ((unsigned _BitInt(512)) 275), 512 - 9); ++ TEST_COMPARE (stdc_leading_zeros ((unsigned _BitInt(373)) 512), 373 - 10); ++ TEST_COMPARE (stdc_leading_ones ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_leading_ones ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_leading_ones ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_leading_ones (~(unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (stdc_leading_ones (~(unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (stdc_leading_ones (~(unsigned _BitInt(512)) 275), 512 - 9); ++ TEST_COMPARE (stdc_leading_ones (~(unsigned _BitInt(373)) 512), 373 - 10); ++ TEST_COMPARE (stdc_trailing_zeros ((unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros ((unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (expr_has_type (stdc_trailing_zeros ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_trailing_zeros (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_trailing_zeros (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_trailing_zeros ((unsigned _BitInt(512)) 256), 8); ++ TEST_COMPARE (stdc_trailing_zeros ((unsigned _BitInt(373)) 512), 9); ++ TEST_COMPARE (stdc_trailing_ones ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_trailing_ones ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_trailing_ones ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_trailing_ones (~(unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (stdc_trailing_ones (~(unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (stdc_trailing_ones ((unsigned _BitInt(512)) 255), 8); ++ TEST_COMPARE (stdc_trailing_ones ((~(unsigned _BitInt(373)) 0) >> 2), ++ 373 - 2); ++ TEST_COMPARE (stdc_first_leading_zero ((unsigned _BitInt(512)) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero ((unsigned _BitInt(512)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero ((unsigned _BitInt(373)) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_zero ((unsigned _BitInt(373)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_zero (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_first_leading_zero (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_first_leading_zero (~(unsigned _BitInt(512)) 511), ++ 512 - 8); ++ TEST_COMPARE (stdc_first_leading_zero (~(unsigned _BitInt(373)) 1023), ++ 373 - 9); ++ TEST_COMPARE (stdc_first_leading_one ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one ((unsigned _BitInt(512)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_leading_one ((unsigned _BitInt(373)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_leading_one (~(unsigned _BitInt(512)) 0), 1); ++ TEST_COMPARE (stdc_first_leading_one (~(unsigned _BitInt(373)) 0), 1); ++ TEST_COMPARE (stdc_first_leading_one ((unsigned _BitInt(512)) 275), 512 - 8); ++ TEST_COMPARE (stdc_first_leading_one ((unsigned _BitInt(373)) 512), 373 - 9); ++ TEST_COMPARE (stdc_first_trailing_zero ((unsigned _BitInt(512)) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero ((unsigned ++ _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero ((unsigned _BitInt(373)) 0), 1); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_zero ((unsigned ++ _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_first_trailing_zero (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_first_trailing_zero (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_first_trailing_zero ((unsigned _BitInt(512)) 255), 9); ++ TEST_COMPARE (stdc_first_trailing_zero ((unsigned _BitInt(373)) 511), 10); ++ TEST_COMPARE (stdc_first_trailing_one ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one ((unsigned _BitInt(512)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_first_trailing_one ((unsigned _BitInt(373)) ++ 0), ui), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~(unsigned _BitInt(512)) 0), 1); ++ TEST_COMPARE (stdc_first_trailing_one (~(unsigned _BitInt(373)) 0), 1); ++ TEST_COMPARE (stdc_first_trailing_one (((unsigned _BitInt(512)) 255) << 175), ++ 176); ++ TEST_COMPARE (stdc_first_trailing_one ((~(unsigned _BitInt(373)) 0) << 311), ++ 312); ++ TEST_COMPARE (stdc_count_zeros ((unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_count_zeros ((unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (expr_has_type (stdc_count_zeros ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_count_zeros (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_count_zeros (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_count_zeros ((unsigned _BitInt(512)) 1315), 512 - 5); ++ TEST_COMPARE (stdc_count_zeros ((unsigned _BitInt(373)) 3363), 373 - 6); ++ TEST_COMPARE (stdc_count_ones ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_count_ones ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_count_ones ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_count_ones (~(unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (stdc_count_ones (~(unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (stdc_count_ones (~(unsigned _BitInt(512)) 1315), 512 - 5); ++ TEST_COMPARE (stdc_count_ones (~(unsigned _BitInt(373)) 3363), 373 - 6); ++ TEST_COMPARE (stdc_has_single_bit ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit ((unsigned _BitInt(512)) 0), ++ _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_has_single_bit ((unsigned _BitInt(373)) 0), ++ _Bool), 1); ++ TEST_COMPARE (stdc_has_single_bit (~(unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (stdc_has_single_bit (~(unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (stdc_has_single_bit (((unsigned _BitInt(512)) 1022) << 279), ++ 0); ++ TEST_COMPARE (stdc_has_single_bit (((unsigned _BitInt(373)) 12) << 305), 0); ++ TEST_COMPARE (stdc_bit_width ((unsigned _BitInt(512)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width ((unsigned _BitInt(512)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_bit_width ((unsigned _BitInt(373)) 0), 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_width ((unsigned _BitInt(373)) 0), ++ ui), 1); ++ TEST_COMPARE (stdc_bit_width (~(unsigned _BitInt(512)) 0), 512); ++ TEST_COMPARE (stdc_bit_width (~(unsigned _BitInt(373)) 0), 373); ++ TEST_COMPARE (stdc_bit_width (((unsigned _BitInt(512)) 1023) << 405), ++ 405 + 10); ++ TEST_COMPARE (stdc_bit_width (((unsigned _BitInt(373)) 1024) << 242), ++ 242 + 11); ++ TEST_COMPARE (stdc_bit_floor ((unsigned _BitInt(512)) 0) != 0, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor ((unsigned _BitInt(512)) 0), ++ unsigned _BitInt(512)), 1); ++ TEST_COMPARE (stdc_bit_floor ((unsigned _BitInt(373)) 0) != 0, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_floor ((unsigned _BitInt(373)) 0), ++ unsigned _BitInt(373)), 1); ++ TEST_COMPARE (stdc_bit_floor (~(unsigned _BitInt(512)) 0) ++ != ((unsigned _BitInt(512)) 1) << (512 - 1), 0); ++ TEST_COMPARE (stdc_bit_floor (~(unsigned _BitInt(373)) 0) ++ != ((unsigned _BitInt(373)) 1) << (373 - 1), 0); ++ TEST_COMPARE (stdc_bit_floor (((unsigned _BitInt(512)) 511) << 405) ++ != (((unsigned _BitInt(512)) 256) << 405), 0); ++ TEST_COMPARE (stdc_bit_floor (((unsigned _BitInt(373)) 512) << 242) ++ != (((unsigned _BitInt(512)) 512) << 242), 0); ++ TEST_COMPARE (stdc_bit_ceil ((unsigned _BitInt(512)) 0) != 1, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil ((unsigned _BitInt(512)) 0), ++ unsigned _BitInt(512)), 1); ++ TEST_COMPARE (stdc_bit_ceil ((unsigned _BitInt(373)) 0) != 1, 0); ++ TEST_COMPARE (expr_has_type (stdc_bit_ceil ((unsigned _BitInt(373)) 0), ++ unsigned _BitInt(373)), 1); ++ TEST_COMPARE (stdc_bit_ceil (~(unsigned _BitInt(512)) 0) != 0, 0); ++ TEST_COMPARE (stdc_bit_ceil (~(unsigned _BitInt(373)) 0) != 0, 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(512)) 1) << (512 - 1)) ++ != ((unsigned _BitInt(512)) 1) << (512 - 1), 0); ++ TEST_COMPARE (stdc_bit_ceil ((~(unsigned _BitInt(373)) 0) >> 1) ++ != ((unsigned _BitInt(373)) 1) << (373 - 1), 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(512)) 512) << 405) ++ != (((unsigned _BitInt(512)) 512) << 405), 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(373)) 513) << 242) ++ != (((unsigned _BitInt(512)) 1024) << 242), 0); ++ TEST_COMPARE (stdc_bit_floor ((unsigned _BitInt(BITINT_MAXWIDTH)) 0) != 0, ++ 0); ++ TEST_COMPARE (stdc_bit_floor (~(unsigned _BitInt(BITINT_MAXWIDTH)) 0) ++ != ((unsigned _BitInt(BITINT_MAXWIDTH)) 1) << (BITINT_MAXWIDTH ++ - 1), 0); ++ TEST_COMPARE (stdc_bit_floor (((unsigned _BitInt(BITINT_MAXWIDTH)) 511) ++ << 405) ++ != (((unsigned _BitInt(BITINT_MAXWIDTH)) 256) << 405), 0); ++ TEST_COMPARE (stdc_bit_floor (((unsigned _BitInt(BITINT_MAXWIDTH)) 512) ++ << 405) ++ != (((unsigned _BitInt(BITINT_MAXWIDTH)) 512) << 405), 0); ++ TEST_COMPARE (stdc_bit_ceil ((unsigned _BitInt(BITINT_MAXWIDTH)) 0) != 1, 0); ++ TEST_COMPARE (stdc_bit_ceil (~(unsigned _BitInt(BITINT_MAXWIDTH)) 0) != 0, ++ 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(BITINT_MAXWIDTH)) 1) ++ << (BITINT_MAXWIDTH - 1)) ++ != ((unsigned _BitInt(BITINT_MAXWIDTH)) 1) << (BITINT_MAXWIDTH ++ - 1), 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(BITINT_MAXWIDTH)) 512) ++ << 405) ++ != (((unsigned _BitInt(BITINT_MAXWIDTH)) 512) << 405), 0); ++ TEST_COMPARE (stdc_bit_ceil (((unsigned _BitInt(BITINT_MAXWIDTH)) 513) ++ << 405) ++ != (((unsigned _BitInt(BITINT_MAXWIDTH)) 1024) << 405), 0); ++# endif ++# endif ++ return 0; ++} ++#else ++static int ++do_test (void) ++{ ++ return 0; ++} ++#endif ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-60.patch b/SOURCES/glibc-upstream-2.39-60.patch new file mode 100644 index 0000000..8765056 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-60.patch @@ -0,0 +1,195 @@ +commit 26e7005728f0eea2972474e6be2905c467661237 +Author: Florian Weimer +Date: Sat May 18 09:33:19 2024 +0200 + + socket: Use may_alias on sockaddr structs (bug 19622) + + This supports common coding patterns. The GCC C front end before + version 7 rejects the may_alias attribute on a struct definition + if it was not present in a previous forward declaration, so this + attribute can only be conditionally applied. + + This implements the spirit of the change in Austin Group issue 1641. + + Suggested-by: Marek Polacek + Suggested-by: Jakub Jelinek + Reviewed-by: Sam James + Reviewed-by: Carlos O'Donell + (cherry picked from commit 8d7b6b4cb27d4dec1dd5f7960298c1699275f962) + +diff --git a/bits/socket.h b/bits/socket.h +index 13de124bac67c217..772074d52ab957df 100644 +--- a/bits/socket.h ++++ b/bits/socket.h +@@ -149,7 +149,7 @@ enum __socket_type + #include + + /* Structure describing a generic socket address. */ +-struct sockaddr ++struct __attribute_struct_may_alias__ sockaddr + { + __SOCKADDR_COMMON (sa_); /* Common data: address family and length. */ + char sa_data[14]; /* Address data. */ +@@ -166,7 +166,7 @@ struct sockaddr + #define _SS_PADSIZE \ + (_SS_SIZE - __SOCKADDR_COMMON_SIZE - sizeof (__ss_aligntype)) + +-struct sockaddr_storage ++struct __attribute_struct_may_alias__ sockaddr_storage + { + __SOCKADDR_COMMON (ss_); /* Address family, etc. */ + char __ss_padding[_SS_PADSIZE]; +diff --git a/inet/netinet/in.h b/inet/netinet/in.h +index fa57b6107908590c..f684be5beb26fc62 100644 +--- a/inet/netinet/in.h ++++ b/inet/netinet/in.h +@@ -244,7 +244,7 @@ extern const struct in6_addr in6addr_loopback; /* ::1 */ + + + /* Structure describing an Internet socket address. */ +-struct sockaddr_in ++struct __attribute_struct_may_alias__ sockaddr_in + { + __SOCKADDR_COMMON (sin_); + in_port_t sin_port; /* Port number. */ +@@ -257,9 +257,11 @@ struct sockaddr_in + - sizeof (struct in_addr)]; + }; + +-#if !__USE_KERNEL_IPV6_DEFS ++#if __USE_KERNEL_IPV6_DEFS ++struct __attribute_struct_may_alias__ sockaddr_in6; ++#else + /* Ditto, for IPv6. */ +-struct sockaddr_in6 ++struct __attribute_struct_may_alias__ sockaddr_in6 + { + __SOCKADDR_COMMON (sin6_); + in_port_t sin6_port; /* Transport layer port # */ +diff --git a/misc/sys/cdefs.h b/misc/sys/cdefs.h +index 520231dbea3a4e24..399ee7d24cae9446 100644 +--- a/misc/sys/cdefs.h ++++ b/misc/sys/cdefs.h +@@ -720,4 +720,13 @@ _Static_assert (0, "IEEE 128-bits long double requires redirection on this platf + # define __attribute_returns_twice__ /* Ignore. */ + #endif + ++/* Mark struct types as aliasable. Restricted to compilers that ++ support forward declarations of structs in the presence of the ++ attribute. */ ++#if __GNUC_PREREQ (7, 1) || defined __clang__ ++# define __attribute_struct_may_alias__ __attribute__ ((__may_alias__)) ++#else ++# define __attribute_struct_may_alias__ ++#endif ++ + #endif /* sys/cdefs.h */ +diff --git a/socket/sys/un.h b/socket/sys/un.h +index bf03b7d6ce959065..ff9cbd6efa7693d5 100644 +--- a/socket/sys/un.h ++++ b/socket/sys/un.h +@@ -26,7 +26,7 @@ + __BEGIN_DECLS + + /* Structure describing the address of an AF_LOCAL (aka AF_UNIX) socket. */ +-struct sockaddr_un ++struct __attribute_struct_may_alias__ sockaddr_un + { + __SOCKADDR_COMMON (sun_); + char sun_path[108]; /* Path name. */ +diff --git a/sysdeps/mach/hurd/bits/socket.h b/sysdeps/mach/hurd/bits/socket.h +index 3e72f9fa93add888..b5eeac373111e694 100644 +--- a/sysdeps/mach/hurd/bits/socket.h ++++ b/sysdeps/mach/hurd/bits/socket.h +@@ -153,7 +153,7 @@ enum __socket_type + #include + + /* Structure describing a generic socket address. */ +-struct sockaddr ++struct __attribute_struct_may_alias__ sockaddr + { + __SOCKADDR_COMMON (sa_); /* Common data: address family and length. */ + char sa_data[14]; /* Address data. */ +@@ -170,7 +170,7 @@ struct sockaddr + #define _SS_PADSIZE \ + (_SS_SIZE - __SOCKADDR_COMMON_SIZE - sizeof (__ss_aligntype)) + +-struct sockaddr_storage ++struct __attribute_struct_may_alias__ sockaddr_storage + { + __SOCKADDR_COMMON (ss_); /* Address family, etc. */ + char __ss_padding[_SS_PADSIZE]; +diff --git a/sysdeps/unix/sysv/linux/bits/socket.h b/sysdeps/unix/sysv/linux/bits/socket.h +index 0d86feb4cadc13e3..6dab283a2ebeaed1 100644 +--- a/sysdeps/unix/sysv/linux/bits/socket.h ++++ b/sysdeps/unix/sysv/linux/bits/socket.h +@@ -180,7 +180,7 @@ typedef __socklen_t socklen_t; + #include + + /* Structure describing a generic socket address. */ +-struct sockaddr ++struct __attribute_struct_may_alias__ sockaddr + { + __SOCKADDR_COMMON (sa_); /* Common data: address family and length. */ + char sa_data[14]; /* Address data. */ +@@ -193,7 +193,7 @@ struct sockaddr + #define _SS_PADSIZE \ + (_SS_SIZE - __SOCKADDR_COMMON_SIZE - sizeof (__ss_aligntype)) + +-struct sockaddr_storage ++struct __attribute_struct_may_alias__ sockaddr_storage + { + __SOCKADDR_COMMON (ss_); /* Address family, etc. */ + char __ss_padding[_SS_PADSIZE]; +diff --git a/sysdeps/unix/sysv/linux/net/if_packet.h b/sysdeps/unix/sysv/linux/net/if_packet.h +index 9ffb69b508cc8ca7..c17e1c23c5cf9169 100644 +--- a/sysdeps/unix/sysv/linux/net/if_packet.h ++++ b/sysdeps/unix/sysv/linux/net/if_packet.h +@@ -26,7 +26,7 @@ + From Linux 2.1 the AF_PACKET interface is preferred and you should + consider using it in place of this one. */ + +-struct sockaddr_pkt ++struct __attribute_struct_may_alias__ sockaddr_pkt + { + __SOCKADDR_COMMON (spkt_); + unsigned char spkt_device[14]; +diff --git a/sysdeps/unix/sysv/linux/netash/ash.h b/sysdeps/unix/sysv/linux/netash/ash.h +index 7d885d17cc9ec313..7a6ff50b17b370c4 100644 +--- a/sysdeps/unix/sysv/linux/netash/ash.h ++++ b/sysdeps/unix/sysv/linux/netash/ash.h +@@ -22,7 +22,7 @@ + #include + #include + +-struct sockaddr_ash ++struct __attribute_struct_may_alias__ sockaddr_ash + { + __SOCKADDR_COMMON (sash_); /* Common data: address family etc. */ + int sash_ifindex; /* Interface to use. */ +diff --git a/sysdeps/unix/sysv/linux/neteconet/ec.h b/sysdeps/unix/sysv/linux/neteconet/ec.h +index b07a10796196e28e..f3132f06ff9f8dee 100644 +--- a/sysdeps/unix/sysv/linux/neteconet/ec.h ++++ b/sysdeps/unix/sysv/linux/neteconet/ec.h +@@ -28,7 +28,7 @@ struct ec_addr + unsigned char net; /* Network number. */ + }; + +-struct sockaddr_ec ++struct __attribute_struct_may_alias__ sockaddr_ec + { + __SOCKADDR_COMMON (sec_); + unsigned char port; /* Port number. */ +diff --git a/sysdeps/unix/sysv/linux/netiucv/iucv.h b/sysdeps/unix/sysv/linux/netiucv/iucv.h +index f5fad817513f3c46..27151e8bbe1e7fa2 100644 +--- a/sysdeps/unix/sysv/linux/netiucv/iucv.h ++++ b/sysdeps/unix/sysv/linux/netiucv/iucv.h +@@ -23,7 +23,7 @@ + + __BEGIN_DECLS + +-struct sockaddr_iucv ++struct __attribute_struct_may_alias__ sockaddr_iucv + { + __SOCKADDR_COMMON (siucv_); + unsigned short siucv_port; /* Reserved */ diff --git a/SOURCES/glibc-upstream-2.39-61.patch b/SOURCES/glibc-upstream-2.39-61.patch new file mode 100644 index 0000000..8c4aa98 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-61.patch @@ -0,0 +1,1824 @@ +commit dd535f4f19ef2b5c367a362af445ecadcf45401e +Author: Adhemerval Zanella +Date: Thu Jan 18 10:18:01 2024 -0300 + + Always define __USE_TIME_BITS64 when 64 bit time_t is used + + It was raised on libc-help [1] that some Linux kernel interfaces expect + the libc to define __USE_TIME_BITS64 to indicate the time_t size for the + kABI. Different than defined by the initial y2038 design document [2], + the __USE_TIME_BITS64 is only defined for ABIs that support more than + one time_t size (by defining the _TIME_BITS for each module). + + The 64 bit time_t redirects are now enabled using a different internal + define (__USE_TIME64_REDIRECTS). There is no expected change in semantic + or code generation. + + Checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu, and + arm-linux-gnueabi + + [1] https://sourceware.org/pipermail/libc-help/2024-January/006557.html + [2] https://sourceware.org/glibc/wiki/Y2038ProofnessDesign + + Reviewed-by: DJ Delorie + (cherry picked from commit a4ed0471d71739928a0d0fa3258b3ff3b158e9b9) + +diff --git a/io/bits/poll2.h b/io/bits/poll2.h +index 6152a8c5e4536d68..efc8b85403e91ba2 100644 +--- a/io/bits/poll2.h ++++ b/io/bits/poll2.h +@@ -43,7 +43,7 @@ poll (struct pollfd *__fds, nfds_t __nfds, int __timeout) + + + #ifdef __USE_GNU +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + extern int __REDIRECT (__ppoll64_alias, (struct pollfd *__fds, nfds_t __nfds, + const struct timespec *__timeout, + const __sigset_t *__ss), __ppoll64); +diff --git a/io/fcntl.h b/io/fcntl.h +index 9cee0b5900cb4f7c..666b7e5eb616b950 100644 +--- a/io/fcntl.h ++++ b/io/fcntl.h +@@ -172,7 +172,7 @@ typedef __pid_t pid_t; + + This function is a cancellation point and therefore not marked with + __THROW. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + # ifndef __USE_FILE_OFFSET64 + extern int fcntl (int __fd, int __cmd, ...); + # else +@@ -185,7 +185,7 @@ extern int __REDIRECT (fcntl, (int __fd, int __cmd, ...), fcntl64); + # ifdef __USE_LARGEFILE64 + extern int fcntl64 (int __fd, int __cmd, ...); + # endif +-#else /* __USE_TIME_BITS64 */ ++#else /* __USE_TIME64_REDIRECTS */ + # ifdef __REDIRECT + extern int __REDIRECT_NTH (fcntl, (int __fd, int __request, ...), + __fcntl_time64); +diff --git a/io/fts.h b/io/fts.h +index 61f95bb4415f77e2..97a031ebbd040080 100644 +--- a/io/fts.h ++++ b/io/fts.h +@@ -187,7 +187,7 @@ FTSENT *fts_read (FTS *); + int fts_set (FTS *, FTSENT *, int) __THROW; + #else + # ifdef __REDIRECT +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + FTSENT *__REDIRECT (fts_children, (FTS *, int), fts64_children); + int __REDIRECT (fts_close, (FTS *), fts64_close); + FTS *__REDIRECT (fts_open, (char * const *, int, +@@ -206,7 +206,7 @@ int __REDIRECT_NTH (fts_set, (FTS *, FTSENT *, int), + __fts64_set_time64); + # endif + # else +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + # define fts_children fts64_children + # define fts_close fts64_close + # define fts_open fts64_open +@@ -217,7 +217,7 @@ int __REDIRECT_NTH (fts_set, (FTS *, FTSENT *, int), + # endif + #endif + #ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + FTSENT64 *fts64_children (FTS64 *, int); + int fts64_close (FTS64 *); + FTS64 *fts64_open (char * const *, int, +diff --git a/io/ftw.h b/io/ftw.h +index e4d1b84d53550d26..39cf595b277a0049 100644 +--- a/io/ftw.h ++++ b/io/ftw.h +@@ -137,7 +137,7 @@ extern int ftw (const char *__dir, __ftw_func_t __func, int __descriptors) + __nonnull ((1, 2)); + #else + # ifdef __REDIRECT +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int __REDIRECT (ftw, (const char *__dir, __ftw_func_t __func, + int __descriptors), ftw64) __nonnull ((1, 2)); + # else +@@ -146,7 +146,7 @@ extern int __REDIRECT (ftw, (const char *__dir, __ftw_func_t __func, + __nonnull ((1, 2)); + # endif + # else +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + # define ftw ftw64 + # else + # define ftw __ftw64_time64 +@@ -154,7 +154,7 @@ extern int __REDIRECT (ftw, (const char *__dir, __ftw_func_t __func, + # endif + #endif + #ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int ftw64 (const char *__dir, __ftw64_func_t __func, + int __descriptors) __nonnull ((1, 2)); + # else +@@ -180,7 +180,7 @@ extern int nftw (const char *__dir, __nftw_func_t __func, int __descriptors, + int __flag) __nonnull ((1, 2)); + # else + # ifdef __REDIRECT +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int __REDIRECT (nftw, (const char *__dir, __nftw_func_t __func, + int __descriptors, int __flag), nftw64) + __nonnull ((1, 2)); +@@ -190,7 +190,7 @@ extern int __REDIRECT (nftw, (const char *__dir, __nftw_func_t __func, + __nonnull ((1, 2)); + # endif + # else +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + # define nftw nftw64 + # else + # define nftw __nftw64_time64 +@@ -198,7 +198,7 @@ extern int __REDIRECT (nftw, (const char *__dir, __nftw_func_t __func, + # endif + # endif + # ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int nftw64 (const char *__dir, __nftw64_func_t __func, + int __descriptors, int __flag) __nonnull ((1, 2)); + # else +diff --git a/io/sys/poll.h b/io/sys/poll.h +index 7858fad6b9b53725..c324ff5dad0d5a81 100644 +--- a/io/sys/poll.h ++++ b/io/sys/poll.h +@@ -66,7 +66,7 @@ extern int ppoll (struct pollfd *__fds, nfds_t __nfds, + const __sigset_t *__ss) + __fortified_attr_access (__write_only__, 1, 2); + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # ifdef __REDIRECT + extern int __REDIRECT (ppoll, (struct pollfd *__fds, nfds_t __nfds, + const struct timespec *__timeout, +diff --git a/io/sys/stat.h b/io/sys/stat.h +index 1fa6d6e62ecb2e4b..3b4ba80132d0b703 100644 +--- a/io/sys/stat.h ++++ b/io/sys/stat.h +@@ -209,7 +209,7 @@ extern int stat (const char *__restrict __file, + that file descriptor FD is open on and put them in BUF. */ + extern int fstat (int __fd, struct stat *__buf) __THROW __nonnull ((2)); + #else +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # ifdef __REDIRECT_NTH + extern int __REDIRECT_NTH (stat, (const char *__restrict __file, + struct stat *__restrict __buf), +@@ -236,7 +236,7 @@ extern int __REDIRECT_NTH (fstat, (int __fd, struct stat *__buf), fstat64) + # endif + #endif + #ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int stat64 (const char *__restrict __file, + struct stat64 *__restrict __buf) __THROW __nonnull ((1, 2)); + extern int fstat64 (int __fd, struct stat64 *__buf) __THROW __nonnull ((2)); +@@ -265,7 +265,7 @@ extern int fstatat (int __fd, const char *__restrict __file, + struct stat *__restrict __buf, int __flag) + __THROW __nonnull ((2, 3)); + # else +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # ifdef __REDIRECT_NTH + extern int __REDIRECT_NTH (fstatat, (int __fd, const char *__restrict __file, + struct stat *__restrict __buf, +@@ -287,7 +287,7 @@ extern int __REDIRECT_NTH (fstatat, (int __fd, const char *__restrict __file, + # endif + + # ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int fstatat64 (int __fd, const char *__restrict __file, + struct stat64 *__restrict __buf, int __flag) + __THROW __nonnull ((2, 3)); +@@ -313,7 +313,7 @@ extern int __REDIRECT_NTH (fstatat64, (int __fd, + extern int lstat (const char *__restrict __file, + struct stat *__restrict __buf) __THROW __nonnull ((1, 2)); + # else +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # ifdef __REDIRECT_NTH + extern int __REDIRECT_NTH (lstat, + (const char *__restrict __file, +@@ -334,7 +334,7 @@ extern int __REDIRECT_NTH (lstat, + # endif + # endif + # ifdef __USE_LARGEFILE64 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int lstat64 (const char *__restrict __file, + struct stat64 *__restrict __buf) + __THROW __nonnull ((1, 2)); +@@ -427,7 +427,7 @@ extern int mkfifoat (int __fd, const char *__path, __mode_t __mode) + #endif + + #ifdef __USE_ATFILE +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Set file access and modification times relative to directory file + descriptor. */ + extern int utimensat (int __fd, const char *__path, +@@ -447,7 +447,7 @@ extern int __REDIRECT_NTH (utimensat, (int fd, const char *__path, + #endif + + #ifdef __USE_XOPEN2K8 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Set file access and modification times of the file associated with FD. */ + extern int futimens (int __fd, const struct timespec __times[2]) __THROW; + +diff --git a/io/utime.h b/io/utime.h +index c5eacedd6a7c9925..1c7587d9c1475348 100644 +--- a/io/utime.h ++++ b/io/utime.h +@@ -35,7 +35,7 @@ __BEGIN_DECLS + /* Structure describing file times. */ + struct utimbuf + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + __time64_t actime; /* Access time. */ + __time64_t modtime; /* Modification time. */ + #else +@@ -46,7 +46,7 @@ struct utimbuf + + /* Set the access and modification times of FILE to those given in + *FILE_TIMES. If FILE_TIMES is NULL, set them to the current time. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int utime (const char *__file, + const struct utimbuf *__file_times) + __THROW __nonnull ((1)); +diff --git a/manual/creature.texi b/manual/creature.texi +index fa5dd9d6ad9081f6..8d19e9dd73560607 100644 +--- a/manual/creature.texi ++++ b/manual/creature.texi +@@ -192,6 +192,12 @@ Linux kernel version on which the system is running. For Linux kernel + version above @b{5.1} syscalls supporting 64-bit time are used. Otherwise, + a fallback code is used with legacy (i.e. 32-bit) syscalls. + ++On such platforms, @theglibc{} will also define @code{__USE_TIME64_REDIRECTS} ++to indicate whether the declarations are expanded to different ones ++(either by redefiniding the symbol name or using symbol aliais). ++For instance, if the symbol @code{clock_gettime} expands to ++@code{__glock_gettime64}. ++ + @item + If @code{_TIME_BITS} is defined to be 32, @code{time_t} is defined to + be a 32-bit integer where that is supported. This is not recommended, +diff --git a/manual/maint.texi b/manual/maint.texi +index 89da704f45912ecb..04faa222e2bd2fc4 100644 +--- a/manual/maint.texi ++++ b/manual/maint.texi +@@ -491,6 +491,10 @@ derived as in the dual-time configuration case, and which expands to + the symbol's name. For instance, the macro @code{__clock_gettime64} + expands to @code{clock_gettime}. + ++When @code{__TIMESIZE} is set to 64, @theglibc{} will also define ++the@code{__USE_TIME_BITS64} macro. It is used by the Linux kernel ABI ++to set the expected @code{time_t} size used on some syscalls. ++ + These macros are purely internal to @theglibc{} and exist only so that + a single definition of the 64-bit time functions can be used on both + single-time and dual-time configurations, and so that glibc code can +diff --git a/misc/sys/ioctl.h b/misc/sys/ioctl.h +index 3f2338ddd32ee2c5..ea6583e12251266d 100644 +--- a/misc/sys/ioctl.h ++++ b/misc/sys/ioctl.h +@@ -38,7 +38,7 @@ __BEGIN_DECLS + /* Perform the I/O control operation specified by REQUEST on FD. + One argument may follow; its presence and type depend on REQUEST. + Return value depends on REQUEST. Usually -1 indicates error. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int ioctl (int __fd, unsigned long int __request, ...) __THROW; + #else + # ifdef __REDIRECT +diff --git a/misc/sys/select.h b/misc/sys/select.h +index e6a0c1b8b278ec22..2e45e94bc19b932e 100644 +--- a/misc/sys/select.h ++++ b/misc/sys/select.h +@@ -98,7 +98,7 @@ __BEGIN_DECLS + + This function is a cancellation point and therefore not marked with + __THROW. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int select (int __nfds, fd_set *__restrict __readfds, + fd_set *__restrict __writefds, + fd_set *__restrict __exceptfds, +@@ -123,7 +123,7 @@ extern int __REDIRECT (select, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pselect (int __nfds, fd_set *__restrict __readfds, + fd_set *__restrict __writefds, + fd_set *__restrict __exceptfds, +diff --git a/posix/glob.h b/posix/glob.h +index 3406662840997c68..b6bba0fbcd0ecce8 100644 +--- a/posix/glob.h ++++ b/posix/glob.h +@@ -150,7 +150,7 @@ extern int glob (const char *__restrict __pattern, int __flags, + /* Free storage allocated in PGLOB by a previous `glob' call. */ + extern void globfree (glob_t *__pglob) __THROW; + #else +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + extern int __REDIRECT_NTHNL (glob, (const char *__restrict __pattern, + int __flags, + int (*__errfunc) (const char *, int), +@@ -170,7 +170,7 @@ extern void __REDIRECT_NTH (globfree, (glob_t *__pglob), globfree64); + #endif + + #ifdef __USE_LARGEFILE64 +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + extern int __REDIRECT_NTHNL (glob64, (const char *__restrict __pattern, + int __flags, + int (*__errfunc) (const char *, int), +diff --git a/posix/sched.h b/posix/sched.h +index 3eac206f359090e9..49f504a488a08727 100644 +--- a/posix/sched.h ++++ b/posix/sched.h +@@ -74,7 +74,7 @@ extern int sched_get_priority_max (int __algorithm) __THROW; + extern int sched_get_priority_min (int __algorithm) __THROW; + + /* Get the SCHED_RR interval for the named process. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int sched_rr_get_interval (__pid_t __pid, struct timespec *__t) __THROW; + #else + # ifdef __REDIRECT_NTH +diff --git a/posix/sys/wait.h b/posix/sys/wait.h +index edbb7b43920dffe7..1f44ee114566b9c9 100644 +--- a/posix/sys/wait.h ++++ b/posix/sys/wait.h +@@ -139,7 +139,7 @@ struct rusage; + nil, store information about the child's resource usage there. If the + WUNTRACED bit is set in OPTIONS, return status for stopped children; + otherwise don't. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern __pid_t wait3 (int *__stat_loc, int __options, + struct rusage * __usage) __THROWNL; + # else +@@ -154,7 +154,7 @@ extern __pid_t __REDIRECT_NTHNL (wait3, (int *__stat_loc, int __options, + #endif + + #ifdef __USE_MISC +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* PID is like waitpid. Other args are like wait3. */ + extern __pid_t wait4 (__pid_t __pid, int *__stat_loc, int __options, + struct rusage *__usage) __THROWNL; +diff --git a/resolv/netdb.h b/resolv/netdb.h +index 14228b0d952664ea..b7f473fafe801b02 100644 +--- a/resolv/netdb.h ++++ b/resolv/netdb.h +@@ -701,7 +701,7 @@ extern int getaddrinfo_a (int __mode, struct gaicb *__list[__restrict_arr], + extern int gai_suspend (const struct gaicb *const __list[], int __ent, + const struct timespec *__timeout); + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # if defined(__REDIRECT) + extern int __REDIRECT (gai_suspend, (const struct gaicb *const __list[], + int __ent, +diff --git a/resource/sys/resource.h b/resource/sys/resource.h +index a5634ba7153c069c..b0bf751e929af810 100644 +--- a/resource/sys/resource.h ++++ b/resource/sys/resource.h +@@ -88,7 +88,7 @@ extern int setrlimit64 (__rlimit_resource_t __resource, + and put it in *USAGE. Returns 0 for success, -1 for failure. */ + extern int getrusage (__rusage_who_t __who, struct rusage *__usage) __THROW; + +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # if defined(__REDIRECT_NTH) + extern int __REDIRECT_NTH (getrusage, (__rusage_who_t __who, + struct rusage *__usage), +diff --git a/rt/aio.h b/rt/aio.h +index 37d19abf16ee98e3..e71435733f988f85 100644 +--- a/rt/aio.h ++++ b/rt/aio.h +@@ -193,7 +193,7 @@ extern __ssize_t __REDIRECT_NTH (aio_return, (struct aiocb *__aiocbp), + extern int __REDIRECT_NTH (aio_cancel, + (int __fildes, struct aiocb *__aiocbp), + aio_cancel64); +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + extern int __REDIRECT_NTH (aio_suspend, + (const struct aiocb *const __list[], int __nent, + const struct timespec *__restrict __timeout), +@@ -215,7 +215,7 @@ extern int __REDIRECT_NTH (aio_fsync, + # define aio_error aio_error64 + # define aio_return aio_return64 + # define aio_cancel aio_cancel64 +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define aio_suspend __aio_suspend_time64 + # else + # define aio_suspend aio_suspend64 +diff --git a/rt/mqueue.h b/rt/mqueue.h +index 787cc36df2eaf9e8..fd6fff4bb2326717 100644 +--- a/rt/mqueue.h ++++ b/rt/mqueue.h +@@ -71,7 +71,7 @@ extern int mq_send (mqd_t __mqdes, const char *__msg_ptr, size_t __msg_len, + unsigned int __msg_prio) __nonnull ((2)); + + #ifdef __USE_XOPEN2K +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Receive the oldest from highest priority messages in message queue + MQDES, stop waiting if ABS_TIMEOUT expires. */ + extern ssize_t mq_timedreceive (mqd_t __mqdes, char *__restrict __msg_ptr, +diff --git a/signal/signal.h b/signal/signal.h +index f37499ce607b998a..8e07b041b1c07fa9 100644 +--- a/signal/signal.h ++++ b/signal/signal.h +@@ -269,7 +269,7 @@ extern int sigwaitinfo (const sigset_t *__restrict __set, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int sigtimedwait (const sigset_t *__restrict __set, + siginfo_t *__restrict __info, + const struct timespec *__restrict __timeout) +diff --git a/socket/sys/socket.h b/socket/sys/socket.h +index 366eaab845a6f326..463cf3291b714427 100644 +--- a/socket/sys/socket.h ++++ b/socket/sys/socket.h +@@ -170,7 +170,7 @@ extern ssize_t recvfrom (int __fd, void *__restrict __buf, size_t __n, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern ssize_t sendmsg (int __fd, const struct msghdr *__message, + int __flags); + #else +@@ -191,7 +191,7 @@ extern ssize_t __sendmsg64 (int __fd, const struct msghdr *__message, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int sendmmsg (int __fd, struct mmsghdr *__vmessages, + unsigned int __vlen, int __flags); + # else +@@ -204,7 +204,7 @@ extern int __sendmmsg64 (int __fd, struct mmsghdr *__vmessages, + unsigned int __vlen, int __flags); + # define sendmmsg __sendmmsg64 + # endif +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + #endif /* __USE_GNU */ + + /* Receive a message as described by MESSAGE from socket FD. +@@ -212,7 +212,7 @@ extern int __sendmmsg64 (int __fd, struct mmsghdr *__vmessages, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern ssize_t recvmsg (int __fd, struct msghdr *__message, int __flags); + #else + # ifdef __REDIRECT +@@ -231,7 +231,7 @@ extern ssize_t __recvmsg64 (int __fd, struct msghdr *__message, int __flags); + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int recvmmsg (int __fd, struct mmsghdr *__vmessages, + unsigned int __vlen, int __flags, + struct timespec *__tmo); +@@ -251,7 +251,7 @@ extern int __REDIRECT (recvmmsg, (int __fd, struct mmsghdr *__vmessages, + /* Put the current value for socket FD's option OPTNAME at protocol level LEVEL + into OPTVAL (which is *OPTLEN bytes long), and set *OPTLEN to the value's + actual length. Returns 0 on success, -1 for errors. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int getsockopt (int __fd, int __level, int __optname, + void *__restrict __optval, + socklen_t *__restrict __optlen) __THROW; +@@ -273,7 +273,7 @@ extern int __getsockopt64 (int __fd, int __level, int __optname, + /* Set socket FD's option OPTNAME at protocol level LEVEL + to *OPTVAL (which is OPTLEN bytes long). + Returns 0 on success, -1 for errors. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int setsockopt (int __fd, int __level, int __optname, + const void *__optval, socklen_t __optlen) __THROW; + #else +diff --git a/support/timespec.h b/support/timespec.h +index 42f32bcc2c30eaac..5ccc7163a50198a5 100644 +--- a/support/timespec.h ++++ b/support/timespec.h +@@ -35,7 +35,7 @@ make_timespec (time_t s, long int ns) + + enum { TIMESPEC_HZ = 1000000000 }; + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + struct timespec timespec_add (struct timespec, struct timespec) + __attribute__((const)); + struct timespec timespec_sub (struct timespec, struct timespec) +diff --git a/support/xtime.h b/support/xtime.h +index 5200eef2e017d380..9961899634e47a1b 100644 +--- a/support/xtime.h ++++ b/support/xtime.h +@@ -30,7 +30,7 @@ __BEGIN_DECLS + /* The following functions call the corresponding libc functions and + terminate the process on error. */ + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + void xclock_gettime (clockid_t clock, struct timespec *ts); + void xclock_settime (clockid_t clock, const struct timespec *ts); + #else +diff --git a/support/xunistd.h b/support/xunistd.h +index b5e8c2f420021c42..13be9a46a3e6b07b 100644 +--- a/support/xunistd.h ++++ b/support/xunistd.h +@@ -36,7 +36,7 @@ pid_t xwaitpid (pid_t, int *status, int flags); + void xpipe (int[2]); + void xdup2 (int, int); + int xopen (const char *path, int flags, mode_t); +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + # ifdef __USE_FILE_OFFSET64 + void xstat (const char *path, struct stat *); + void xlstat (const char *path, struct stat *); +diff --git a/sysdeps/generic/features-time64.h b/sysdeps/generic/features-time64.h +deleted file mode 100644 +index 4d38b8ba766b34f5..0000000000000000 +--- a/sysdeps/generic/features-time64.h ++++ /dev/null +@@ -1,19 +0,0 @@ +-/* Features part to handle 64-bit time_t support. Generic version. +- Copyright (C) 2021-2024 Free Software Foundation, Inc. +- This file is part of the GNU C Library. +- +- The GNU C Library is free software; you can redistribute it and/or +- modify it under the terms of the GNU Lesser General Public +- License as published by the Free Software Foundation; either +- version 2.1 of the License, or (at your option) any later version. +- +- The GNU C Library is distributed in the hope that it will be useful, +- but WITHOUT ANY WARRANTY; without even the implied warranty of +- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +- Lesser General Public License for more details. +- +- You should have received a copy of the GNU Lesser General Public +- License along with the GNU C Library; if not, see +- . */ +- +-/* The generic configuration only support _TIME_BITS=32. */ +diff --git a/sysdeps/nptl/pthread.h b/sysdeps/nptl/pthread.h +index 1607fdf29a5ed964..3d4f4a756c66750d 100644 +--- a/sysdeps/nptl/pthread.h ++++ b/sysdeps/nptl/pthread.h +@@ -223,7 +223,7 @@ extern int pthread_join (pthread_t __th, void **__thread_return); + the thread in *THREAD_RETURN, if THREAD_RETURN is not NULL. */ + extern int pthread_tryjoin_np (pthread_t __th, void **__thread_return) __THROW; + +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Make calling thread wait for termination of the thread TH, but only + until TIMEOUT. The exit status of the thread is stored in + *THREAD_RETURN, if THREAD_RETURN is not NULL. +@@ -796,7 +796,7 @@ extern int pthread_mutex_lock (pthread_mutex_t *__mutex) + + #ifdef __USE_XOPEN2K + /* Wait until lock becomes available, or specified time passes. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_mutex_timedlock (pthread_mutex_t *__restrict __mutex, + const struct timespec *__restrict + __abstime) __THROWNL __nonnull ((1, 2)); +@@ -813,7 +813,7 @@ extern int __REDIRECT_NTHNL (pthread_mutex_timedlock, + #endif + + #ifdef __USE_GNU +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_mutex_clocklock (pthread_mutex_t *__restrict __mutex, + clockid_t __clockid, + const struct timespec *__restrict +@@ -982,7 +982,7 @@ extern int pthread_rwlock_tryrdlock (pthread_rwlock_t *__rwlock) + + # ifdef __USE_XOPEN2K + /* Try to acquire read lock for RWLOCK or return after specified time. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_rwlock_timedrdlock (pthread_rwlock_t *__restrict __rwlock, + const struct timespec *__restrict + __abstime) __THROWNL __nonnull ((1, 2)); +@@ -1000,7 +1000,7 @@ extern int __REDIRECT_NTHNL (pthread_rwlock_timedrdlock, + # endif + + # ifdef __USE_GNU +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_rwlock_clockrdlock (pthread_rwlock_t *__restrict __rwlock, + clockid_t __clockid, + const struct timespec *__restrict +@@ -1029,7 +1029,7 @@ extern int pthread_rwlock_trywrlock (pthread_rwlock_t *__rwlock) + + # ifdef __USE_XOPEN2K + /* Try to acquire write lock for RWLOCK or return after specified time. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_rwlock_timedwrlock (pthread_rwlock_t *__restrict __rwlock, + const struct timespec *__restrict + __abstime) __THROWNL __nonnull ((1, 2)); +@@ -1047,7 +1047,7 @@ extern int __REDIRECT_NTHNL (pthread_rwlock_timedwrlock, + # endif + + # ifdef __USE_GNU +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_rwlock_clockwrlock (pthread_rwlock_t *__restrict __rwlock, + clockid_t __clockid, + const struct timespec *__restrict +@@ -1141,7 +1141,7 @@ extern int pthread_cond_wait (pthread_cond_t *__restrict __cond, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_cond_timedwait (pthread_cond_t *__restrict __cond, + pthread_mutex_t *__restrict __mutex, + const struct timespec *__restrict __abstime) +@@ -1167,7 +1167,7 @@ extern int __REDIRECT (pthread_cond_timedwait, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int pthread_cond_clockwait (pthread_cond_t *__restrict __cond, + pthread_mutex_t *__restrict __mutex, + __clockid_t __clock_id, +diff --git a/sysdeps/pthread/semaphore.h b/sysdeps/pthread/semaphore.h +index c75ca4ce6d12e081..8a096336991730a8 100644 +--- a/sysdeps/pthread/semaphore.h ++++ b/sysdeps/pthread/semaphore.h +@@ -59,7 +59,7 @@ extern int sem_wait (sem_t *__sem) __nonnull ((1)); + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int sem_timedwait (sem_t *__restrict __sem, + const struct timespec *__restrict __abstime) + __nonnull ((1, 2)); +@@ -77,7 +77,7 @@ extern int __REDIRECT (sem_timedwait, + #endif + + #ifdef __USE_GNU +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int sem_clockwait (sem_t *__restrict __sem, + clockid_t clock, + const struct timespec *__restrict __abstime) +diff --git a/sysdeps/pthread/threads.h b/sysdeps/pthread/threads.h +index c5410b5c3af84883..7cade24e1f4b9e7a 100644 +--- a/sysdeps/pthread/threads.h ++++ b/sysdeps/pthread/threads.h +@@ -90,7 +90,7 @@ extern thrd_t thrd_current (void); + __TIME_POINT. The current thread may resume if receives a signal. In + that case, if __REMAINING is not NULL, the remaining time is stored in + the object pointed by it. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int thrd_sleep (const struct timespec *__time_point, + struct timespec *__remaining); + #else +@@ -143,7 +143,7 @@ extern int mtx_lock (mtx_t *__mutex); + /* Block the current thread until the mutex pointed by __MUTEX is unlocked + or time pointed by __TIME_POINT is reached. In case the mutex is unlock, + the current thread will not be blocked. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int mtx_timedlock (mtx_t *__restrict __mutex, + const struct timespec *__restrict __time_point); + #else +@@ -194,7 +194,7 @@ extern int cnd_wait (cnd_t *__cond, mtx_t *__mutex); + /* Block current thread on the condition variable until condition variable + pointed by __COND is signaled or time pointed by __TIME_POINT is + reached. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int cnd_timedwait (cnd_t *__restrict __cond, + mtx_t *__restrict __mutex, + const struct timespec *__restrict __time_point); +diff --git a/sysdeps/unix/sysv/linux/arm/bits/struct_stat.h b/sysdeps/unix/sysv/linux/arm/bits/struct_stat.h +index 255feaa8cbf048b5..89534fea85100202 100644 +--- a/sysdeps/unix/sysv/linux/arm/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/arm/bits/struct_stat.h +@@ -28,7 +28,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -83,13 +83,13 @@ struct stat + # else + __ino64_t st_ino; /* File serial number. */ + # endif +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -125,7 +125,7 @@ struct stat64 + unsigned long int st_ctimensec; /* Nsecs of last status change. */ + # endif + __ino64_t st_ino; /* File serial number. */ +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/bits/socket-constants.h b/sysdeps/unix/sysv/linux/bits/socket-constants.h +index d997dbf5943aa90f..b2102d3abfe5a386 100644 +--- a/sysdeps/unix/sysv/linux/bits/socket-constants.h ++++ b/sysdeps/unix/sysv/linux/bits/socket-constants.h +@@ -64,7 +64,7 @@ + # define SO_TIMESTAMPNS_NEW 64 + # define SO_TIMESTAMPING_NEW 65 + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define SO_RCVTIMEO SO_RCVTIMEO_NEW + # define SO_SNDTIMEO SO_SNDTIMEO_NEW + # define SO_TIMESTAMP SO_TIMESTAMP_NEW +diff --git a/sysdeps/unix/sysv/linux/bits/time.h b/sysdeps/unix/sysv/linux/bits/time.h +index b70ba58a7dd58f72..cab405797130387d 100644 +--- a/sysdeps/unix/sysv/linux/bits/time.h ++++ b/sysdeps/unix/sysv/linux/bits/time.h +@@ -77,7 +77,7 @@ __BEGIN_DECLS + /* Tune a POSIX clock. */ + extern int clock_adjtime (__clockid_t __clock_id, struct timex *__utx) __THROW __nonnull((2)); + +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # if defined(__REDIRECT_NTH) + extern int __REDIRECT_NTH (clock_adjtime, (__clockid_t __clock_id, + struct timex *__utx), +diff --git a/sysdeps/unix/sysv/linux/bits/timex.h b/sysdeps/unix/sysv/linux/bits/timex.h +index 398d8094f23f0b5d..03ccde6d0a0d086c 100644 +--- a/sysdeps/unix/sysv/linux/bits/timex.h ++++ b/sysdeps/unix/sysv/linux/bits/timex.h +@@ -25,7 +25,7 @@ + + struct timex + { +-# if defined __USE_TIME_BITS64 || (__TIMESIZE == 64 && __WORDSIZE == 32) ++# if defined __USE_TIME64_REDIRECTS || (__TIMESIZE == 64 && __WORDSIZE == 32) + unsigned int modes; /* mode selector */ + int :32; /* pad */ + long long offset; /* time offset (usec) */ +diff --git a/sysdeps/unix/sysv/linux/bits/types/struct_msqid_ds.h b/sysdeps/unix/sysv/linux/bits/types/struct_msqid_ds.h +index fae50281c7fccdfc..86296ca92299c114 100644 +--- a/sysdeps/unix/sysv/linux/bits/types/struct_msqid_ds.h ++++ b/sysdeps/unix/sysv/linux/bits/types/struct_msqid_ds.h +@@ -26,7 +26,7 @@ + The type `struct msg' is opaque. */ + struct msqid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm msg_perm; /* structure describing operation permission */ +diff --git a/sysdeps/unix/sysv/linux/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/bits/types/struct_semid_ds.h +index 3c277ed1d8a3a956..2ac89b3ce456fd39 100644 +--- a/sysdeps/unix/sysv/linux/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/bits/types/struct_shmid_ds.h b/sysdeps/unix/sysv/linux/bits/types/struct_shmid_ds.h +index 09de0b4e3a587c52..1012ed031702df04 100644 +--- a/sysdeps/unix/sysv/linux/bits/types/struct_shmid_ds.h ++++ b/sysdeps/unix/sysv/linux/bits/types/struct_shmid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a shared memory segment. */ + struct shmid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm shm_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/csky/bits/struct_stat.h b/sysdeps/unix/sysv/linux/csky/bits/struct_stat.h +index 4b4822d6d0c00081..47a9f0aaff8fce00 100644 +--- a/sysdeps/unix/sysv/linux/csky/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/csky/bits/struct_stat.h +@@ -43,7 +43,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -88,7 +88,7 @@ struct stat + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +diff --git a/sysdeps/unix/sysv/linux/features-time64.h b/sysdeps/unix/sysv/linux/features-time64.h +index dd3442c2eab58cbe..8d573cd23e48c0f7 100644 +--- a/sysdeps/unix/sysv/linux/features-time64.h ++++ b/sysdeps/unix/sysv/linux/features-time64.h +@@ -24,9 +24,8 @@ + # if _TIME_BITS == 64 + # if ! defined (_FILE_OFFSET_BITS) || _FILE_OFFSET_BITS != 64 + # error "_TIME_BITS=64 is allowed only with _FILE_OFFSET_BITS=64" +-# elif __TIMESIZE == 32 +-# define __USE_TIME_BITS64 1 + # endif ++# define __USE_TIME_BITS64 1 + # elif _TIME_BITS == 32 + # if __TIMESIZE > 32 + # error "_TIME_BITS=32 is not compatible with __TIMESIZE > 32" +@@ -34,4 +33,10 @@ + # else + # error Invalid _TIME_BITS value (can only be 32 or 64-bit) + # endif ++#elif __TIMESIZE == 64 ++# define __USE_TIME_BITS64 1 ++#endif ++ ++#if defined __USE_TIME_BITS64 && __TIMESIZE == 32 ++# define __USE_TIME64_REDIRECTS 1 + #endif +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/socket-constants.h b/sysdeps/unix/sysv/linux/hppa/bits/socket-constants.h +index 58b523d03a234a76..80a76a17dc0d75c9 100644 +--- a/sysdeps/unix/sysv/linux/hppa/bits/socket-constants.h ++++ b/sysdeps/unix/sysv/linux/hppa/bits/socket-constants.h +@@ -54,7 +54,7 @@ + # define SO_TIMESTAMPNS_NEW 0x4039 + # define SO_TIMESTAMPING_NEW 0x403A + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define SO_RCVTIMEO SO_RCVTIMEO_NEW + # define SO_SNDTIMEO SO_SNDTIMEO_NEW + # define SO_TIMESTAMP SO_TIMESTAMP_NEW +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/struct_stat.h b/sysdeps/unix/sysv/linux/hppa/bits/struct_stat.h +index 069efdbc5579bfde..09ea40054bcd12da 100644 +--- a/sysdeps/unix/sysv/linux/hppa/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/hppa/bits/struct_stat.h +@@ -28,7 +28,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -83,13 +83,13 @@ struct stat + # else + __ino64_t st_ino; /* File serial number. */ + # endif +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -125,7 +125,7 @@ struct stat64 + unsigned long int st_ctimensec; /* Nsecs of last status change. */ + # endif + __ino64_t st_ino; /* File serial number. */ +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_msqid_ds.h b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_msqid_ds.h +index 2bc7cac06da36e27..4995e0a4a5017f1a 100644 +--- a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_msqid_ds.h ++++ b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_msqid_ds.h +@@ -26,7 +26,7 @@ + The type `struct msg' is opaque. */ + struct msqid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm msg_perm; /* structure describing operation permission */ +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_semid_ds.h +index dd8fbebcf48891f9..df88949dc20c9821 100644 +--- a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_shmid_ds.h b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_shmid_ds.h +index 58ac572b6e9b13f9..cb3b0303aa28d192 100644 +--- a/sysdeps/unix/sysv/linux/hppa/bits/types/struct_shmid_ds.h ++++ b/sysdeps/unix/sysv/linux/hppa/bits/types/struct_shmid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a shared memory segment. */ + struct shmid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm shm_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/m68k/bits/struct_stat.h b/sysdeps/unix/sysv/linux/m68k/bits/struct_stat.h +index 3b1db157bc0a76a7..b7f20189b1600585 100644 +--- a/sysdeps/unix/sysv/linux/m68k/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/m68k/bits/struct_stat.h +@@ -25,7 +25,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -80,13 +80,13 @@ struct stat + # else + __ino64_t st_ino; /* File serial number. */ + # endif +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -122,7 +122,7 @@ struct stat64 + unsigned long int st_ctimensec; /* Nsecs of last status change. */ + # endif + __ino64_t st_ino; /* File serial number. */ +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/microblaze/bits/struct_stat.h b/sysdeps/unix/sysv/linux/microblaze/bits/struct_stat.h +index 0f4693fb1fa6750e..ff1e269f1461a50e 100644 +--- a/sysdeps/unix/sysv/linux/microblaze/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/microblaze/bits/struct_stat.h +@@ -26,7 +26,7 @@ + #ifndef __USE_FILE_OFFSET64 + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -64,7 +64,7 @@ struct stat + # endif + unsigned int __glibc_reserved4; + unsigned int __glibc_reserved5; +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + #else /* __USE_FILE_OFFSET64 */ + /* MS: If __USE_FILE_OFFSET64 is setup then struct stat should match stat64 +@@ -74,7 +74,7 @@ struct stat + * create one ifdef to separate stats structures. */ + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + unsigned long long st_dev; /* Device. */ +@@ -112,14 +112,14 @@ struct stat + # endif + unsigned int __glibc_reserved4; + unsigned int __glibc_reserved5; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif /* __USE_FILE_OFFSET64 */ + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + unsigned long long st_dev; /* Device. */ +@@ -157,7 +157,7 @@ struct stat64 + # endif + unsigned int __glibc_reserved4; + unsigned int __glibc_reserved5; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/mips/bits/socket-constants.h b/sysdeps/unix/sysv/linux/mips/bits/socket-constants.h +index c6908a2793db1e6f..77ffc8b890311c23 100644 +--- a/sysdeps/unix/sysv/linux/mips/bits/socket-constants.h ++++ b/sysdeps/unix/sysv/linux/mips/bits/socket-constants.h +@@ -54,7 +54,7 @@ + # define SO_TIMESTAMPNS_NEW 64 + # define SO_TIMESTAMPING_NEW 65 + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define SO_RCVTIMEO SO_RCVTIMEO_NEW + # define SO_SNDTIMEO SO_SNDTIMEO_NEW + # define SO_TIMESTAMP SO_TIMESTAMP_NEW +diff --git a/sysdeps/unix/sysv/linux/mips/bits/struct_stat.h b/sysdeps/unix/sysv/linux/mips/bits/struct_stat.h +index 277ebad9b6e66e42..50a4b367f6d8d8da 100644 +--- a/sysdeps/unix/sysv/linux/mips/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/mips/bits/struct_stat.h +@@ -29,7 +29,7 @@ + /* Structure describing file characteristics. */ + struct stat + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + unsigned long int st_dev; +@@ -82,13 +82,13 @@ struct stat + __blkcnt64_t st_blocks; /* Number of 512-byte blocks allocated. */ + # endif + long int st_pad5[14]; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + + # ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + unsigned long int st_dev; +@@ -123,7 +123,7 @@ struct stat64 + long int st_pad3; + __blkcnt64_t st_blocks; /* Number of 512-byte blocks allocated. */ + long int st_pad4[14]; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + # endif /* __USE_LARGEFILE64 */ + +@@ -131,7 +131,7 @@ struct stat64 + + struct stat + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; +@@ -189,7 +189,7 @@ struct stat + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; +@@ -224,7 +224,7 @@ struct stat64 + unsigned int st_pad3; + __blkcnt64_t st_blocks; + int st_pad4[14]; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/mips/bits/types/struct_msqid_ds.h b/sysdeps/unix/sysv/linux/mips/bits/types/struct_msqid_ds.h +index 4e8bd51b0aa59929..09c53648b7451ac4 100644 +--- a/sysdeps/unix/sysv/linux/mips/bits/types/struct_msqid_ds.h ++++ b/sysdeps/unix/sysv/linux/mips/bits/types/struct_msqid_ds.h +@@ -26,7 +26,7 @@ + The type `struct msg' is opaque. */ + struct msqid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm msg_perm; /* structure describing operation permission */ +diff --git a/sysdeps/unix/sysv/linux/mips/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/mips/bits/types/struct_semid_ds.h +index d1a30e31645ee2a4..0746684a7d4c0dc4 100644 +--- a/sysdeps/unix/sysv/linux/mips/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/mips/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/mips/bits/types/struct_shmid_ds.h b/sysdeps/unix/sysv/linux/mips/bits/types/struct_shmid_ds.h +index 8771164b5762c48f..c665af187489c675 100644 +--- a/sysdeps/unix/sysv/linux/mips/bits/types/struct_shmid_ds.h ++++ b/sysdeps/unix/sysv/linux/mips/bits/types/struct_shmid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a shared memory segment. */ + struct shmid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm shm_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/nios2/bits/struct_stat.h b/sysdeps/unix/sysv/linux/nios2/bits/struct_stat.h +index 672c7c6bb8f20314..dab466d881c94b61 100644 +--- a/sysdeps/unix/sysv/linux/nios2/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/nios2/bits/struct_stat.h +@@ -43,7 +43,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -88,7 +88,7 @@ struct stat + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/socket-constants.h b/sysdeps/unix/sysv/linux/powerpc/bits/socket-constants.h +index 6be5055e65ab0b88..0d3e095c5d4bb94c 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/socket-constants.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/socket-constants.h +@@ -54,7 +54,7 @@ + # define SO_TIMESTAMPNS_NEW 64 + # define SO_TIMESTAMPING_NEW 65 + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define SO_RCVTIMEO SO_RCVTIMEO_NEW + # define SO_SNDTIMEO SO_SNDTIMEO_NEW + # define SO_TIMESTAMP SO_TIMESTAMP_NEW +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/struct_stat.h b/sysdeps/unix/sysv/linux/powerpc/bits/struct_stat.h +index f6328399cd1f2985..2cf331544a805f06 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/struct_stat.h +@@ -28,7 +28,7 @@ + #if __WORDSIZE == 32 + struct stat + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -79,13 +79,13 @@ struct stat + # endif + unsigned long int __glibc_reserved4; + unsigned long int __glibc_reserved5; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + + # ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -122,7 +122,7 @@ struct stat64 + # endif + unsigned long int __glibc_reserved4; + unsigned long int __glibc_reserved5; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + # endif /* __USE_LARGEFILE64 */ + +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_msqid_ds.h b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_msqid_ds.h +index 830629cd37e4fccf..44ae08265d41502c 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_msqid_ds.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_msqid_ds.h +@@ -26,7 +26,7 @@ + The type `struct msg' is opaque. */ + struct msqid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm msg_perm; /* structure describing operation permission */ +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_semid_ds.h +index 43702575570a7d54..ccee57c28bf70729 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_shmid_ds.h b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_shmid_ds.h +index da1b4b3c56383000..58145d0a5d2e8203 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_shmid_ds.h ++++ b/sysdeps/unix/sysv/linux/powerpc/bits/types/struct_shmid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a shared memory segment. */ + struct shmid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm shm_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/s390/bits/struct_stat.h b/sysdeps/unix/sysv/linux/s390/bits/struct_stat.h +index 9911c47bb268e207..e5c9024fb241fad9 100644 +--- a/sysdeps/unix/sysv/linux/s390/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/s390/bits/struct_stat.h +@@ -65,7 +65,7 @@ struct stat + #else + struct stat + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -166,7 +166,7 @@ struct stat64 + # else + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +diff --git a/sysdeps/unix/sysv/linux/sh/bits/struct_stat.h b/sysdeps/unix/sysv/linux/sh/bits/struct_stat.h +index cbd4bc0f31ac0c97..d47cd7be1eeb9f17 100644 +--- a/sysdeps/unix/sysv/linux/sh/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/sh/bits/struct_stat.h +@@ -28,7 +28,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -83,13 +83,13 @@ struct stat + # else + __ino64_t st_ino; /* File serial number. */ + # endif +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -125,7 +125,7 @@ struct stat64 + unsigned long int st_ctimensec; /* Nsecs of last status change. */ + # endif + __ino64_t st_ino; /* File serial number. */ +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/socket-constants.h b/sysdeps/unix/sysv/linux/sparc/bits/socket-constants.h +index 59958611c42ed1f6..e7a6b684ccc51a1e 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/socket-constants.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/socket-constants.h +@@ -54,7 +54,7 @@ + # define SO_TIMESTAMPNS_NEW 0x0042 + # define SO_TIMESTAMPING_NEW 0x0043 + +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # define SO_RCVTIMEO SO_RCVTIMEO_NEW + # define SO_SNDTIMEO SO_SNDTIMEO_NEW + # define SO_TIMESTAMP SO_TIMESTAMP_NEW +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/struct_stat.h b/sysdeps/unix/sysv/linux/sparc/bits/struct_stat.h +index 4e48634edc7f16f6..fcab5f480478768b 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/struct_stat.h +@@ -28,7 +28,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -79,13 +79,13 @@ struct stat + # endif + unsigned long int __glibc_reserved4; + unsigned long int __glibc_reserved5; +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -126,7 +126,7 @@ struct stat64 + # endif + unsigned long int __glibc_reserved4; + unsigned long int __glibc_reserved5; +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_msqid_ds.h b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_msqid_ds.h +index db783c28d414000f..ed8d47c9b61f182b 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_msqid_ds.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_msqid_ds.h +@@ -26,7 +26,7 @@ + The type `struct msg' is opaque. */ + struct msqid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm msg_perm; /* structure describing operation permission */ +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_semid_ds.h +index 1c8a3693dbd6977f..b9e729b8b46a232e 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_shmid_ds.h b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_shmid_ds.h +index 35a0cc36ab764e52..7885d2ab25d9c679 100644 +--- a/sysdeps/unix/sysv/linux/sparc/bits/types/struct_shmid_ds.h ++++ b/sysdeps/unix/sysv/linux/sparc/bits/types/struct_shmid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a shared memory segment. */ + struct shmid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm shm_perm; /* operation permission struct */ +diff --git a/sysdeps/unix/sysv/linux/sys/epoll.h b/sysdeps/unix/sysv/linux/sys/epoll.h +index 9b51e3bd1471c25c..fc8dce45c8947a03 100644 +--- a/sysdeps/unix/sysv/linux/sys/epoll.h ++++ b/sysdeps/unix/sysv/linux/sys/epoll.h +@@ -140,7 +140,7 @@ extern int epoll_pwait (int __epfd, struct epoll_event *__events, + + This function is a cancellation point and therefore not marked with + __THROW. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int epoll_pwait2 (int __epfd, struct epoll_event *__events, + int __maxevents, const struct timespec *__timeout, + const __sigset_t *__ss) +diff --git a/sysdeps/unix/sysv/linux/sys/prctl.h b/sysdeps/unix/sysv/linux/sys/prctl.h +index b13b84626149450e..79a9b31273b33bca 100644 +--- a/sysdeps/unix/sysv/linux/sys/prctl.h ++++ b/sysdeps/unix/sysv/linux/sys/prctl.h +@@ -38,7 +38,7 @@ + __BEGIN_DECLS + + /* Control process execution. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int prctl (int __option, ...) __THROW; + #else + # ifdef __REDIRECT +diff --git a/sysdeps/unix/sysv/linux/sys/timerfd.h b/sysdeps/unix/sysv/linux/sys/timerfd.h +index 19d0cbfae020d7a6..f66ece306ade51dd 100644 +--- a/sysdeps/unix/sysv/linux/sys/timerfd.h ++++ b/sysdeps/unix/sysv/linux/sys/timerfd.h +@@ -47,7 +47,7 @@ extern int timerfd_settime (int __ufd, int __flags, + const struct itimerspec *__utmr, + struct itimerspec *__otmr) __THROW; + +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # if defined(__REDIRECT_NTH) + extern int __REDIRECT_NTH (timerfd_settime, + (int __ufd, int __flags, +@@ -62,7 +62,7 @@ extern int __REDIRECT_NTH (timerfd_settime, + /* Return the next expiration time of UFD. */ + extern int timerfd_gettime (int __ufd, struct itimerspec *__otmr) __THROW; + +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # if defined(__REDIRECT_NTH) + extern int __REDIRECT_NTH (timerfd_gettime, (int __ufd, + struct itimerspec *__otmr), +diff --git a/sysdeps/unix/sysv/linux/sys/timex.h b/sysdeps/unix/sysv/linux/sys/timex.h +index 28ce022253e057ce..568748d7679211fd 100644 +--- a/sysdeps/unix/sysv/linux/sys/timex.h ++++ b/sysdeps/unix/sysv/linux/sys/timex.h +@@ -54,7 +54,7 @@ struct ntptimeval + + __BEGIN_DECLS + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int adjtimex (struct timex *__ntx) __THROW __nonnull ((1)); + extern int ntp_gettimex (struct ntptimeval *__ntv) __THROW __nonnull ((1)); + +diff --git a/sysdeps/unix/sysv/linux/x86/bits/struct_stat.h b/sysdeps/unix/sysv/linux/x86/bits/struct_stat.h +index 21aa315d8dc15681..810d6566f05683f3 100644 +--- a/sysdeps/unix/sysv/linux/x86/bits/struct_stat.h ++++ b/sysdeps/unix/sysv/linux/x86/bits/struct_stat.h +@@ -25,7 +25,7 @@ + + struct stat + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + __dev_t st_dev; /* Device. */ +@@ -95,14 +95,14 @@ struct stat + __ino64_t st_ino; /* File serial number. */ + # endif + # endif +-#endif /* __USE_TIME_BITS64 */ ++#endif /* __USE_TIME64_REDIRECTS */ + }; + + #ifdef __USE_LARGEFILE64 + /* Note stat64 has the same shape as stat for x86-64. */ + struct stat64 + { +-# ifdef __USE_TIME_BITS64 ++# ifdef __USE_TIME64_REDIRECTS + # include + # else + __dev_t st_dev; /* Device. */ +@@ -152,7 +152,7 @@ struct stat64 + # else + __ino64_t st_ino; /* File serial number. */ + # endif +-# endif /* __USE_TIME_BITS64 */ ++# endif /* __USE_TIME64_REDIRECTS */ + }; + #endif + +diff --git a/sysdeps/unix/sysv/linux/x86/bits/types/struct_semid_ds.h b/sysdeps/unix/sysv/linux/x86/bits/types/struct_semid_ds.h +index 9f3d170b65269927..81867c0316080913 100644 +--- a/sysdeps/unix/sysv/linux/x86/bits/types/struct_semid_ds.h ++++ b/sysdeps/unix/sysv/linux/x86/bits/types/struct_semid_ds.h +@@ -23,7 +23,7 @@ + /* Data structure describing a set of semaphores. */ + struct semid_ds + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + # include + #else + struct ipc_perm sem_perm; /* operation permission struct */ +diff --git a/sysvipc/sys/msg.h b/sysvipc/sys/msg.h +index d0388b0522fb4183..4178ad99550e8b12 100644 +--- a/sysvipc/sys/msg.h ++++ b/sysvipc/sys/msg.h +@@ -58,7 +58,7 @@ struct msgbuf + __BEGIN_DECLS + + /* Message queue control operation. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int msgctl (int __msqid, int __cmd, struct msqid_ds *__buf) __THROW; + #else + # ifdef __REDIRECT_NTH +diff --git a/sysvipc/sys/sem.h b/sysvipc/sys/sem.h +index 5d9ec39296af6db9..812f1303b3d93360 100644 +--- a/sysvipc/sys/sem.h ++++ b/sysvipc/sys/sem.h +@@ -48,7 +48,7 @@ struct sembuf + __BEGIN_DECLS + + /* Semaphore control operation. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int semctl (int __semid, int __semnum, int __cmd, ...) __THROW; + #else + # ifdef __REDIRECT_NTH +@@ -68,7 +68,7 @@ extern int semop (int __semid, struct sembuf *__sops, size_t __nsops) __THROW; + + #ifdef __USE_GNU + /* Operate on semaphore with timeout. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int semtimedop (int __semid, struct sembuf *__sops, size_t __nsops, + const struct timespec *__timeout) __THROW; + # else +diff --git a/sysvipc/sys/shm.h b/sysvipc/sys/shm.h +index 04191656d54d3c89..496e57ef4528af2e 100644 +--- a/sysvipc/sys/shm.h ++++ b/sysvipc/sys/shm.h +@@ -46,7 +46,7 @@ __BEGIN_DECLS + facility. The definition is found in XPG4.2. */ + + /* Shared memory control operation. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int shmctl (int __shmid, int __cmd, struct shmid_ds *__buf) __THROW; + #else + # ifdef __REDIRECT_NTH +diff --git a/time/bits/types/struct_timespec.h b/time/bits/types/struct_timespec.h +index 489e81136d7f4ad5..1141015f276cb330 100644 +--- a/time/bits/types/struct_timespec.h ++++ b/time/bits/types/struct_timespec.h +@@ -10,14 +10,14 @@ + has nanoseconds instead of microseconds. */ + struct timespec + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + __time64_t tv_sec; /* Seconds. */ + #else + __time_t tv_sec; /* Seconds. */ + #endif + #if __WORDSIZE == 64 \ + || (defined __SYSCALL_WORDSIZE && __SYSCALL_WORDSIZE == 64) \ +- || (__TIMESIZE == 32 && !defined __USE_TIME_BITS64) ++ || (__TIMESIZE == 32 && !defined __USE_TIME64_REDIRECTS) + __syscall_slong_t tv_nsec; /* Nanoseconds. */ + #else + # if __BYTE_ORDER == __BIG_ENDIAN +diff --git a/time/bits/types/struct_timeval.h b/time/bits/types/struct_timeval.h +index 3466137c35fb5d0e..0c8e88c82c400d68 100644 +--- a/time/bits/types/struct_timeval.h ++++ b/time/bits/types/struct_timeval.h +@@ -7,7 +7,7 @@ + microsecond but also has a range of years. */ + struct timeval + { +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + __time64_t tv_sec; /* Seconds. */ + __suseconds64_t tv_usec; /* Microseconds. */ + #else +diff --git a/time/bits/types/time_t.h b/time/bits/types/time_t.h +index 84d67f6ac346e249..00cde92c623375f3 100644 +--- a/time/bits/types/time_t.h ++++ b/time/bits/types/time_t.h +@@ -4,7 +4,7 @@ + #include + + /* Returned by `time'. */ +-#ifdef __USE_TIME_BITS64 ++#ifdef __USE_TIME64_REDIRECTS + typedef __time64_t time_t; + #else + typedef __time_t time_t; +diff --git a/time/sys/time.h b/time/sys/time.h +index c8708198a5fa2e58..8c3d0c30222d15d1 100644 +--- a/time/sys/time.h ++++ b/time/sys/time.h +@@ -63,7 +63,7 @@ struct timezone + use localtime etc. instead. + This function itself is semi-obsolete; + most callers should use time or clock_gettime instead. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern int gettimeofday (struct timeval *__restrict __tv, + void *__restrict __tz) __THROW __nonnull ((1)); + #else +@@ -77,7 +77,7 @@ extern int __REDIRECT_NTH (gettimeofday, (struct timeval *__restrict __tv, + #endif + + #ifdef __USE_MISC +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Set the current time of day and timezone information. + This call is restricted to the super-user. + Setting the timezone in this way is obsolete, but we don't yet +@@ -143,7 +143,7 @@ typedef enum __itimer_which __itimer_which_t; + typedef int __itimer_which_t; + #endif + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + /* Set *VALUE to the current setting of timer WHICH. + Return 0 on success, -1 on errors. */ + extern int getitimer (__itimer_which_t __which, +@@ -184,7 +184,7 @@ extern int __REDIRECT_NTH (utimes, (const char *__file, + #endif + + #ifdef __USE_MISC +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Same as `utimes', but does not follow symbolic links. */ + extern int lutimes (const char *__file, const struct timeval __tvp[2]) + __THROW __nonnull ((1)); +@@ -207,7 +207,7 @@ extern int __REDIRECT_NTH (futimes, (int __fd, const struct timeval __tvp[2]), + #endif + + #ifdef __USE_GNU +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Change the access time of FILE relative to FD to TVP[0] and the + modification time of FILE to TVP[1]. If TVP is a null pointer, use + the current time instead. Returns 0 on success, -1 on errors. */ +diff --git a/time/time.h b/time/time.h +index 1609aaeffaac5713..3785dc608f6c6ed6 100644 +--- a/time/time.h ++++ b/time/time.h +@@ -71,7 +71,7 @@ __BEGIN_DECLS + The result / CLOCKS_PER_SEC is program time in seconds. */ + extern clock_t clock (void) __THROW; + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + /* Return the current time and put it in *TIMER if TIMER is not NULL. */ + extern time_t time (time_t *__timer) __THROW; + +@@ -127,7 +127,7 @@ extern char *strptime_l (const char *__restrict __s, + #endif + + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + /* Return the `struct tm' representation of *TIMER + in Universal Coordinated Time (aka Greenwich Mean Time). */ + extern struct tm *gmtime (const time_t *__timer) __THROW; +@@ -149,7 +149,7 @@ extern struct tm *__REDIRECT_NTH (localtime, (const time_t *__timer), + + + #if defined __USE_POSIX || __GLIBC_USE (ISOC2X) +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Return the `struct tm' representation of *TIMER in UTC, + using *TP to store the result. */ + extern struct tm *gmtime_r (const time_t *__restrict __timer, +@@ -180,7 +180,7 @@ extern struct tm*__REDIRECT_NTH (localtime_r, (const time_t *__restrict __t, + extern char *asctime (const struct tm *__tp) __THROW; + + /* Equivalent to `asctime (localtime (timer))'. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern char *ctime (const time_t *__timer) __THROW; + #else + # ifdef __REDIRECT_NTH +@@ -199,7 +199,7 @@ extern char *asctime_r (const struct tm *__restrict __tp, + char *__restrict __buf) __THROW; + + /* Equivalent to `asctime_r (localtime_r (timer, *TMP*), buf)'. */ +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + extern char *ctime_r (const time_t *__restrict __timer, + char *__restrict __buf) __THROW; + #else +@@ -242,7 +242,7 @@ extern long int timezone; + + + #if defined __USE_MISC || __GLIBC_USE (ISOC2X) +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Like `mktime', but for TP represents Universal Time, not local time. */ + extern time_t timegm (struct tm *__tp) __THROW; + # else +@@ -259,7 +259,7 @@ extern time_t __REDIRECT_NTH (timegm, (struct tm *__tp), __timegm64); + /* Miscellaneous functions many Unices inherited from the public domain + localtime package. These are included only for compatibility. */ + +-#ifndef __USE_TIME_BITS64 ++#ifndef __USE_TIME64_REDIRECTS + /* Another name for `mktime'. */ + extern time_t timelocal (struct tm *__tp) __THROW; + #else +@@ -274,7 +274,7 @@ extern int dysize (int __year) __THROW __attribute__ ((__const__)); + + + #ifdef __USE_POSIX199309 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Pause execution for a number of nanoseconds. + + This function is a cancellation point and therefore not marked with +@@ -320,7 +320,7 @@ extern int __REDIRECT_NTH (clock_settime, (clockid_t __clock_id, const struct + + This function is a cancellation point and therefore not marked with + __THROW. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int clock_nanosleep (clockid_t __clock_id, int __flags, + const struct timespec *__req, + struct timespec *__rem); +@@ -349,7 +349,7 @@ extern int timer_create (clockid_t __clock_id, + extern int timer_delete (timer_t __timerid) __THROW; + + /* Set timer TIMERID to VALUE, returning old value in OVALUE. */ +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + extern int timer_settime (timer_t __timerid, int __flags, + const struct itimerspec *__restrict __value, + struct itimerspec *__restrict __ovalue) __THROW; +@@ -379,7 +379,7 @@ extern int timer_getoverrun (timer_t __timerid) __THROW; + + + #ifdef __USE_ISOC11 +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Set TS to calendar time based in time base BASE. */ + extern int timespec_get (struct timespec *__ts, int __base) + __THROW __nonnull ((1)); +@@ -395,7 +395,7 @@ extern int __REDIRECT_NTH (timespec_get, (struct timespec *__ts, int __base), + + + #if __GLIBC_USE (ISOC2X) +-# ifndef __USE_TIME_BITS64 ++# ifndef __USE_TIME64_REDIRECTS + /* Set TS to resolution of time base BASE. */ + extern int timespec_getres (struct timespec *__ts, int __base) + __THROW; diff --git a/SOURCES/glibc-upstream-2.39-62.patch b/SOURCES/glibc-upstream-2.39-62.patch new file mode 100644 index 0000000..e928935 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-62.patch @@ -0,0 +1,38 @@ +commit aee37de299af95007a1f00f2cdf6c1452f39e682 +Author: Adhemerval Zanella +Date: Fri Apr 5 09:02:36 2024 -0300 + + Reinstate generic features-time64.h + + The a4ed0471d7 removed the generic version which is included by + features.h and used by Hurd. + + Checked by building i686-gnu and x86_64-gnu with build-many-glibc.py. + + (cherry picked from commit c27f8763cffbb7db9b3f1f5e09ef24d26cbb63f4) + +diff --git a/sysdeps/generic/features-time64.h b/sysdeps/generic/features-time64.h +new file mode 100644 +index 0000000000000000..4d38b8ba766b34f5 +--- /dev/null ++++ b/sysdeps/generic/features-time64.h +@@ -0,0 +1,19 @@ ++/* Features part to handle 64-bit time_t support. Generic version. ++ Copyright (C) 2021-2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++/* The generic configuration only support _TIME_BITS=32. */ diff --git a/SOURCES/glibc-upstream-2.39-63.patch b/SOURCES/glibc-upstream-2.39-63.patch new file mode 100644 index 0000000..4db9709 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-63.patch @@ -0,0 +1,62 @@ +commit c9d8534406ab69b9bc1fd3fdfb9e88c9580d3f24 +Author: Sunil K Pandey +Date: Mon May 27 10:08:18 2024 -0700 + + i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782) + + This patch disables Intel Xeon Phi tests for GCC 15 and above. + + GCC 15 removed Intel Xeon Phi ISA support. + commit e1a7e2c54d52d0ba374735e285b617af44841ace + Author: Haochen Jiang + Date: Mon May 20 10:43:44 2024 +0800 + + i386: Remove Xeon Phi ISA support + + Fixes BZ 31782. + + Reviewed-by: H.J. Lu + (cherry picked from commit 1b713c9a5349ef3cd1a8ccf9de017c7865713c67) + +diff --git a/sysdeps/x86/tst-cpu-features-supports.c b/sysdeps/x86/tst-cpu-features-supports.c +index 93008dac703e762f..0f43ef2b2da0d3db 100644 +--- a/sysdeps/x86/tst-cpu-features-supports.c ++++ b/sysdeps/x86/tst-cpu-features-supports.c +@@ -65,7 +65,7 @@ do_test (int argc, char **argv) + #endif + fails += CHECK_FEATURE_ACTIVE (avx, AVX); + fails += CHECK_FEATURE_ACTIVE (avx2, AVX2); +-#if __GNUC_PREREQ (7, 0) ++#if __GNUC_PREREQ (7, 0) && !__GNUC_PREREQ (15, 0) + fails += CHECK_FEATURE_ACTIVE (avx5124fmaps, AVX512_4FMAPS); + fails += CHECK_FEATURE_ACTIVE (avx5124vnniw, AVX512_4VNNIW); + #endif +@@ -92,14 +92,18 @@ do_test (int argc, char **argv) + #if __GNUC_PREREQ (6, 0) + fails += CHECK_FEATURE_ACTIVE (avx512bw, AVX512BW); + fails += CHECK_FEATURE_ACTIVE (avx512cd, AVX512CD); ++# if !__GNUC_PREREQ (15, 0) + fails += CHECK_FEATURE_ACTIVE (avx512er, AVX512ER); ++# endif + fails += CHECK_FEATURE_ACTIVE (avx512dq, AVX512DQ); + #endif + #if __GNUC_PREREQ (5, 0) + fails += CHECK_FEATURE_ACTIVE (avx512f, AVX512F); + #endif + #if __GNUC_PREREQ (6, 0) ++# if !__GNUC_PREREQ (15, 0) + fails += CHECK_FEATURE_ACTIVE (avx512pf, AVX512PF); ++# endif + fails += CHECK_FEATURE_ACTIVE (avx512vl, AVX512VL); + #endif + #if __GNUC_PREREQ (5, 0) +@@ -148,7 +152,9 @@ do_test (int argc, char **argv) + #endif + fails += CHECK_FEATURE_ACTIVE (popcnt, POPCNT); + #if __GNUC_PREREQ (11, 0) ++# if !__GNUC_PREREQ (15, 0) + fails += CHECK_FEATURE_ACTIVE (prefetchwt1, PREFETCHWT1); ++# endif + fails += CHECK_FEATURE_ACTIVE (ptwrite, PTWRITE); + fails += CHECK_FEATURE_ACTIVE (rdpid, RDPID); + fails += CHECK_FEATURE_ACTIVE (rdrnd, RDRAND); diff --git a/SOURCES/glibc-upstream-2.39-64.patch b/SOURCES/glibc-upstream-2.39-64.patch new file mode 100644 index 0000000..bbba766 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-64.patch @@ -0,0 +1,68 @@ +commit 70f560fc22212f733647c9121c26bbb2307f2e10 +Author: Stafford Horne +Date: Wed Apr 3 06:40:37 2024 +0100 + + misc: Add support for Linux uio.h RWF_NOAPPEND flag + + In Linux 6.9 a new flag is added to allow for Per-io operations to + disable append mode even if a file was opened with the flag O_APPEND. + This is done with the new RWF_NOAPPEND flag. + + This caused two test failures as these tests expected the flag 0x00000020 + to be unused. Adding the flag definition now fixes these tests on Linux + 6.9 (v6.9-rc1). + + FAIL: misc/tst-preadvwritev2 + FAIL: misc/tst-preadvwritev64v2 + + This patch adds the flag, adjusts the test and adds details to + documentation. + + Link: https://lore.kernel.org/all/20200831153207.GO3265@brightrain.aerifal.cx/ + Reviewed-by: Adhemerval Zanella + (cherry picked from commit 3db9d208dd5f30b12900989c6d2214782b8e2011) + +diff --git a/manual/llio.texi b/manual/llio.texi +index 0b61d491f50b763a..fae49d14332db675 100644 +--- a/manual/llio.texi ++++ b/manual/llio.texi +@@ -1339,6 +1339,10 @@ will fail and set @code{errno} to @code{EAGAIN} if the operation would block. + + @item RWF_APPEND + Per-IO synchronization as if the file was opened with @code{O_APPEND} flag. ++ ++@item RWF_NOAPPEND ++This flag allows an offset to be honored, even if the file was opened with ++@code{O_APPEND} flag. + @end vtable + + When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} the +diff --git a/misc/tst-preadvwritev2-common.c b/misc/tst-preadvwritev2-common.c +index b5f19f002cd1653e..8e04ff7282e727b6 100644 +--- a/misc/tst-preadvwritev2-common.c ++++ b/misc/tst-preadvwritev2-common.c +@@ -34,8 +34,11 @@ + #ifndef RWF_APPEND + # define RWF_APPEND 0 + #endif ++#ifndef RWF_NOAPPEND ++# define RWF_NOAPPEND 0 ++#endif + #define RWF_SUPPORTED (RWF_HIPRI | RWF_DSYNC | RWF_SYNC | RWF_NOWAIT \ +- | RWF_APPEND) ++ | RWF_APPEND | RWF_NOAPPEND) + + /* Generic uio_lim.h does not define IOV_MAX. */ + #ifndef IOV_MAX +diff --git a/sysdeps/unix/sysv/linux/bits/uio-ext.h b/sysdeps/unix/sysv/linux/bits/uio-ext.h +index 7854cccef3e0a618..ead7a091566aa295 100644 +--- a/sysdeps/unix/sysv/linux/bits/uio-ext.h ++++ b/sysdeps/unix/sysv/linux/bits/uio-ext.h +@@ -47,6 +47,7 @@ extern ssize_t process_vm_writev (pid_t __pid, const struct iovec *__lvec, + #define RWF_SYNC 0x00000004 /* per-IO O_SYNC. */ + #define RWF_NOWAIT 0x00000008 /* per-IO nonblocking mode. */ + #define RWF_APPEND 0x00000010 /* per-IO O_APPEND. */ ++#define RWF_NOAPPEND 0x00000020 /* per-IO negation of O_APPEND */ + + __END_DECLS + diff --git a/SOURCES/glibc-upstream-2.39-65.patch b/SOURCES/glibc-upstream-2.39-65.patch new file mode 100644 index 0000000..60b9efb --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-65.patch @@ -0,0 +1,44 @@ +commit 6ade91c21140d8c803c289932dbfc74537f65a1f +Author: Florian Weimer +Date: Mon Jun 3 10:49:40 2024 +0200 + + elf: Avoid some free (NULL) calls in _dl_update_slotinfo + + This has been confirmed to work around some interposed mallocs. Here + is a discussion of the impact test ust/libc-wrapper/test_libc-wrapper + in lttng-tools: + + New TLS usage in libgcc_s.so.1, compatibility impact + + + Reportedly, this patch also papers over a similar issue when tcmalloc + 2.9.1 is not compiled with -ftls-model=initial-exec. Of course the + goal really should be to compile mallocs with the initial-exec TLS + model, but this commit appears to be a useful interim workaround. + + Fixes commit d2123d68275acc0f061e73d5f86ca504e0d5a344 ("elf: Fix slow + tls access after dlopen [BZ #19924]"). + + Reviewed-by: Carlos O'Donell + (cherry picked from commit afe42e935b3ee97bac9a7064157587777259c60e) + +diff --git a/elf/dl-tls.c b/elf/dl-tls.c +index 7b3dd9ab60e89ae9..670dbc42fc2e3334 100644 +--- a/elf/dl-tls.c ++++ b/elf/dl-tls.c +@@ -819,7 +819,14 @@ _dl_update_slotinfo (unsigned long int req_modid, size_t new_gen) + dtv entry free it. Note: this is not AS-safe. */ + /* XXX Ideally we will at some point create a memory + pool. */ +- free (dtv[modid].pointer.to_free); ++ /* Avoid calling free on a null pointer. Some mallocs ++ incorrectly use dynamic TLS, and depending on how the ++ free function was compiled, it could call ++ __tls_get_addr before the null pointer check in the ++ free implementation. Checking here papers over at ++ least some dynamic TLS usage by interposed mallocs. */ ++ if (dtv[modid].pointer.to_free != NULL) ++ free (dtv[modid].pointer.to_free); + dtv[modid].pointer.val = TLS_DTV_UNALLOCATED; + dtv[modid].pointer.to_free = NULL; + diff --git a/SOURCES/glibc-upstream-2.39-66.patch b/SOURCES/glibc-upstream-2.39-66.patch new file mode 100644 index 0000000..5c83c5b --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-66.patch @@ -0,0 +1,57 @@ +commit 00899eba260ff3edb62bc6f45c1860bc64fd59e0 +Author: Michael Jeanson +Date: Mon Feb 5 15:22:39 2024 -0500 + + x86/cet: fix shadow stack test scripts + + Some shadow stack test scripts use the '==' operator with the 'test' + command to validate exit codes resulting in the following error: + + sysdeps/x86_64/tst-shstk-legacy-1e.sh: 31: test: 139: unexpected operator + + The '==' operator is invalid for the 'test' command, use '-eq' like the + previous call to 'test'. + + Signed-off-by: Michael Jeanson + Reviewed-by: H.J. Lu + (cherry picked from commit 155bb9d036646138348fee0ac045de601811e0c5) + +diff --git a/sysdeps/x86_64/tst-shstk-legacy-1e-static.sh b/sysdeps/x86_64/tst-shstk-legacy-1e-static.sh +index 46f12337571127c6..0a9a164a3e8f4610 100755 +--- a/sysdeps/x86_64/tst-shstk-legacy-1e-static.sh ++++ b/sysdeps/x86_64/tst-shstk-legacy-1e-static.sh +@@ -26,7 +26,7 @@ ${common_objpfx}elf/tst-shstk-legacy-1e-static + status=$? + if test $status -eq 77; then + exit 77 +-elif test $status == 139; then ++elif test $status -eq 139; then + exit 0 + else + exit 1 +diff --git a/sysdeps/x86_64/tst-shstk-legacy-1e.sh b/sysdeps/x86_64/tst-shstk-legacy-1e.sh +index 31212453d9374013..3dec5623e41dc3c8 100755 +--- a/sysdeps/x86_64/tst-shstk-legacy-1e.sh ++++ b/sysdeps/x86_64/tst-shstk-legacy-1e.sh +@@ -28,7 +28,7 @@ ${test_program_prefix} \ + status=$? + if test $status -eq 77; then + exit 77 +-elif test $status == 139; then ++elif test $status -eq 139; then + exit 0 + else + exit 1 +diff --git a/sysdeps/x86_64/tst-shstk-legacy-1g.sh b/sysdeps/x86_64/tst-shstk-legacy-1g.sh +index e84087068e6b7b15..249831e816f6589c 100755 +--- a/sysdeps/x86_64/tst-shstk-legacy-1g.sh ++++ b/sysdeps/x86_64/tst-shstk-legacy-1g.sh +@@ -28,7 +28,7 @@ ${test_program_prefix} \ + status=$? + if test $status -eq 77; then + exit 77 +-elif test $status == 139; then ++elif test $status -eq 139; then + exit 0 + else + exit 1 diff --git a/SOURCES/glibc-upstream-2.39-67.patch b/SOURCES/glibc-upstream-2.39-67.patch new file mode 100644 index 0000000..30bd3b2 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-67.patch @@ -0,0 +1,232 @@ +commit 305ee48826961b9b2ad2b2fb36ec19622d8e1d77 +Author: H.J. Lu +Date: Thu May 9 20:27:39 2024 -0700 + + malloc/Makefile: Split and sort tests + + Put each test on a separate line and sort tests. + + Signed-off-by: H.J. Lu + Reviewed-by: Carlos O'Donell + (cherry picked from commit d49cd6a1913da9744b9a0ffbefb3f7958322382e) + +diff --git a/malloc/Makefile b/malloc/Makefile +index c83ade5f10063113..77ba1a91093ebc55 100644 +--- a/malloc/Makefile ++++ b/malloc/Makefile +@@ -24,60 +24,92 @@ include ../Makeconfig + + dist-headers := malloc.h + headers := $(dist-headers) obstack.h mcheck.h +-tests := mallocbug tst-malloc tst-valloc tst-calloc tst-obstack \ +- tst-malloc-check tst-mallocfork tst-trim1 \ +- tst-malloc-usable tst-realloc tst-reallocarray tst-posix_memalign \ +- tst-pvalloc tst-pvalloc-fortify tst-memalign tst-mallopt \ +- tst-malloc-backtrace tst-malloc-thread-exit \ +- tst-malloc-thread-fail tst-malloc-fork-deadlock \ +- tst-mallocfork2 \ +- tst-mallocfork3 \ +- tst-interpose-nothread \ +- tst-interpose-thread \ +- tst-alloc_buffer \ +- tst-free-errno \ +- tst-malloc-tcache-leak \ +- tst-malloc_info tst-mallinfo2 \ +- tst-malloc-too-large \ +- tst-malloc-stats-cancellation \ +- tst-tcfree1 tst-tcfree2 tst-tcfree3 \ +- tst-safe-linking \ +- tst-mallocalign1 \ +- tst-memalign-2 \ +- tst-memalign-3 \ +- tst-aligned-alloc ++tests := \ ++ mallocbug \ ++ tst-aligned-alloc \ ++ tst-alloc_buffer \ ++ tst-calloc \ ++ tst-free-errno \ ++ tst-interpose-nothread \ ++ tst-interpose-thread \ ++ tst-malloc \ ++ tst-malloc-backtrace \ ++ tst-malloc-check \ ++ tst-malloc-fork-deadlock \ ++ tst-malloc-stats-cancellation \ ++ tst-malloc-tcache-leak \ ++ tst-malloc-thread-exit \ ++ tst-malloc-thread-fail \ ++ tst-malloc-too-large \ ++ tst-malloc-usable \ ++ tst-malloc_info tst-mallinfo2 \ ++ tst-mallocalign1 \ ++ tst-mallocfork \ ++ tst-mallocfork2 \ ++ tst-mallocfork3 \ ++ tst-mallopt \ ++ tst-memalign \ ++ tst-memalign-2 \ ++ tst-memalign-3 \ ++ tst-obstack \ ++ tst-posix_memalign \ ++ tst-pvalloc \ ++ tst-pvalloc-fortify \ ++ tst-realloc \ ++ tst-reallocarray \ ++ tst-safe-linking \ ++ tst-tcfree1 tst-tcfree2 tst-tcfree3 \ ++ tst-trim1 \ ++ tst-valloc \ ++# tests + + tests-static := \ +- tst-interpose-static-nothread \ +- tst-interpose-static-thread \ +- tst-aligned-alloc-static ++ tst-aligned-alloc-static \ ++ tst-interpose-static-nothread \ ++ tst-interpose-static-thread \ ++# tests-static + + # Test for the malloc_set_state symbol removed in glibc 2.25. + ifeq ($(have-GLIBC_2.23)$(build-shared),yesyes) +-tests += tst-mallocstate tst-compathooks-off tst-compathooks-on ++tests += \ ++ tst-compathooks-off \ ++ tst-compathooks-on \ ++ tst-mallocstate \ ++# tests + endif + + tests-internal := tst-scratch_buffer + + # The dynarray framework is only available inside glibc. + tests-internal += \ +- tst-dynarray \ +- tst-dynarray-fail \ +- tst-dynarray-at-fail \ ++ tst-dynarray \ ++ tst-dynarray-at-fail \ ++ tst-dynarray-fail \ ++# tests-internal + +-tests += tst-malloc-usable-tunables tst-mxfast ++tests += \ ++ tst-malloc-usable-tunables \ ++ tst-mxfast \ ++# tests + + tests += $(tests-static) + test-srcs = tst-mtrace + + # These tests either are run with MALLOC_CHECK_=3 by default or do not work + # with MALLOC_CHECK_=3 because they expect a specific failure. +-tests-exclude-malloc-check = tst-malloc-check tst-malloc-usable \ +- tst-mxfast tst-safe-linking \ +- tst-compathooks-off tst-compathooks-on tst-memalign-2 tst-memalign-3 \ +- tst-mallocfork2 \ +- tst-mallocfork3 \ +- tst-malloc-tcache-leak ++tests-exclude-malloc-check = \ ++ tst-compathooks-off \ ++ tst-compathooks-on \ ++ tst-malloc-check \ ++ tst-malloc-tcache-leak \ ++ tst-malloc-usable \ ++ tst-mallocfork2 \ ++ tst-mallocfork3 \ ++ tst-memalign-2 \ ++ tst-memalign-3 \ ++ tst-mxfast \ ++ tst-safe-linking \ ++# tests-exclude-malloc-check + + # Run all tests with MALLOC_CHECK_=3 + tests-malloc-check = $(filter-out $(tests-exclude-malloc-check) \ +@@ -87,18 +119,19 @@ tests-malloc-check = $(filter-out $(tests-exclude-malloc-check) \ + # the Transparent Huge Pages support (1) or automatic huge page support (2). + # We need exclude some tests that define the ENV vars. + tests-exclude-hugetlb1 = \ +- tst-compathooks-off \ +- tst-compathooks-on \ +- tst-interpose-nothread \ +- tst-interpose-thread \ +- tst-interpose-static-nothread \ +- tst-interpose-static-thread \ +- tst-malloc-usable \ +- tst-malloc-usable-tunables \ +- tst-mallocstate \ +- tst-malloc-tcache-leak \ +- tst-mallocfork2 \ +- tst-mallocfork3 ++ tst-compathooks-off \ ++ tst-compathooks-on \ ++ tst-interpose-nothread \ ++ tst-interpose-static-nothread \ ++ tst-interpose-static-thread \ ++ tst-interpose-thread \ ++ tst-malloc-tcache-leak \ ++ tst-malloc-usable \ ++ tst-malloc-usable-tunables \ ++ tst-mallocfork2 \ ++ tst-mallocfork3 \ ++ tst-mallocstate \ ++# tests-exclude-hugetlb1 + # The tst-free-errno relies on the used malloc page size to mmap an + # overlapping region. + tests-exclude-hugetlb2 = \ +@@ -114,22 +147,25 @@ ifeq ($(have-GLIBC_2.23)$(build-shared),yesyes) + # Tests that don't play well with mcheck. They are either bugs in mcheck or + # the tests expect specific internal behavior that is changed due to linking to + # libmcheck.a. +-tests-exclude-mcheck = tst-mallocstate \ +- tst-safe-linking \ +- tst-malloc-backtrace \ +- tst-malloc-fork-deadlock \ +- tst-malloc-stats-cancellation \ +- tst-malloc-tcache-leak \ +- tst-malloc-thread-exit \ +- tst-malloc-thread-fail \ +- tst-malloc-usable-tunables \ +- tst-malloc_info \ +- tst-compathooks-off tst-compathooks-on \ +- tst-memalign-2 \ +- tst-memalign-3 \ +- tst-mxfast \ +- tst-mallocfork2 \ +- tst-mallocfork3 ++tests-exclude-mcheck = \ ++ tst-compathooks-off \ ++ tst-compathooks-on \ ++ tst-malloc-backtrace \ ++ tst-malloc-fork-deadlock \ ++ tst-malloc-stats-cancellation \ ++ tst-malloc-tcache-leak \ ++ tst-malloc-thread-exit \ ++ tst-malloc-thread-fail \ ++ tst-malloc-usable-tunables \ ++ tst-malloc_info \ ++ tst-mallocfork2 \ ++ tst-mallocfork3 \ ++ tst-mallocstate \ ++ tst-memalign-2 \ ++ tst-memalign-3 \ ++ tst-mxfast \ ++ tst-safe-linking \ ++# tests-exclude-mcheck + + tests-mcheck = $(filter-out $(tests-exclude-mcheck) $(tests-static), $(tests)) + endif +@@ -159,10 +195,12 @@ extra-libs-others = $(extra-libs) + extra-test-objs += \ + tst-interpose-aux-nothread.o \ + tst-interpose-aux-thread.o \ ++# extra-test-objs + + test-extras = \ + tst-interpose-aux-nothread \ + tst-interpose-aux-thread \ ++# test-extras + + libmemusage-routines = memusage + libmemusage-inhibit-o = $(filter-out .os,$(object-suffixes)) diff --git a/SOURCES/glibc-upstream-2.39-68.patch b/SOURCES/glibc-upstream-2.39-68.patch new file mode 100644 index 0000000..979f9e0 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-68.patch @@ -0,0 +1,244 @@ +commit a2da98aa2a13b1bde7fd3a98f31b04ddecfafa8f +Author: Joe Simmons-Talbott +Date: Tue May 14 14:36:50 2024 +0000 + + malloc: Improve aligned_alloc and calloc test coverage. + + Add a DSO (malloc/tst-aligned_alloc-lib.so) that can be used during + testing to interpose malloc with a call that randomly uses either + aligned_alloc, __libc_malloc, or __libc_calloc in the place of malloc. + Use LD_PRELOAD with the DSO to mirror malloc/tst-malloc.c testing as an + example in malloc/tst-malloc-random.c. Add malloc/tst-aligned-alloc-random.c + as another example that does a number of malloc calls with randomly sized, + but limited to 0xffff, requests. + + The intention is to be able to utilize existing malloc testing to ensure + that similar allocation APIs are also exposed to the same rigors. + + Reviewed-by: DJ Delorie + (cherry picked from commit 3395157ff2b0657d70c36169156f67440205c8bf) + +diff --git a/malloc/Makefile b/malloc/Makefile +index 77ba1a91093ebc55..c1a03f3cb02c464d 100644 +--- a/malloc/Makefile ++++ b/malloc/Makefile +@@ -27,6 +27,7 @@ headers := $(dist-headers) obstack.h mcheck.h + tests := \ + mallocbug \ + tst-aligned-alloc \ ++ tst-aligned-alloc-random \ + tst-alloc_buffer \ + tst-calloc \ + tst-free-errno \ +@@ -36,6 +37,7 @@ tests := \ + tst-malloc-backtrace \ + tst-malloc-check \ + tst-malloc-fork-deadlock \ ++ tst-malloc-random \ + tst-malloc-stats-cancellation \ + tst-malloc-tcache-leak \ + tst-malloc-thread-exit \ +@@ -193,6 +195,7 @@ extra-libs-others = $(extra-libs) + + # Helper objects for some tests. + extra-test-objs += \ ++ tst-aligned_alloc-lib.so \ + tst-interpose-aux-nothread.o \ + tst-interpose-aux-thread.o \ + # extra-test-objs +@@ -202,6 +205,9 @@ test-extras = \ + tst-interpose-aux-thread \ + # test-extras + ++modules-names = \ ++ tst-aligned_alloc-lib ++ + libmemusage-routines = memusage + libmemusage-inhibit-o = $(filter-out .os,$(object-suffixes)) + +@@ -408,3 +414,9 @@ tst-mallocstate-malloc-check-ENV = LD_PRELOAD=$(objpfx)libc_malloc_debug.so + # libc_malloc_debug.so. + $(objpfx)tst-mallocstate: $(objpfx)libc_malloc_debug.so + $(objpfx)tst-mallocstate-malloc-check: $(objpfx)libc_malloc_debug.so ++ ++$(objpfx)tst-aligned-alloc-random.out: $(objpfx)tst-aligned_alloc-lib.so ++$(objpfx)tst-malloc-random.out: $(objpfx)tst-aligned_alloc-lib.so ++ ++tst-aligned-alloc-random-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so ++tst-malloc-random-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so +diff --git a/malloc/tst-aligned-alloc-random.c b/malloc/tst-aligned-alloc-random.c +new file mode 100644 +index 0000000000000000..f2825ce38f04e04e +--- /dev/null ++++ b/malloc/tst-aligned-alloc-random.c +@@ -0,0 +1,43 @@ ++/* Test for randomized malloc that calls aligned_alloc ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++ ++static int ++do_test (void) ++{ ++ void *p1; ++ int i; ++ ++ srandom (time (NULL)); ++ ++ for (i = 0; i < 1024; i++) ++ { ++ size_t size = random () & 0xffff; ++ ++ p1 = malloc (size); ++ TEST_VERIFY (p1 != NULL); ++ } ++ ++ return 0; ++} ++ ++ ++#include +diff --git a/malloc/tst-aligned_alloc-lib.c b/malloc/tst-aligned_alloc-lib.c +new file mode 100644 +index 0000000000000000..0205df5acf6297a5 +--- /dev/null ++++ b/malloc/tst-aligned_alloc-lib.c +@@ -0,0 +1,72 @@ ++/* Module used for improved aligned_alloc testing. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ Copyright The GNU Toolchain Authors. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public License as ++ published by the Free Software Foundation; either version 2.1 of the ++ License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; see the file COPYING.LIB. If ++ not, see . */ ++ ++#include ++#include ++#include ++ ++extern void *__libc_malloc (size_t size); ++extern void *__libc_calloc (size_t n, size_t size); ++ ++int aligned_alloc_count = 0; ++int libc_malloc_count = 0; ++int libc_calloc_count = 0; ++ ++/* Get a random alignment value. Biased towards the smaller values. Must be ++ a power of 2. */ ++static size_t get_random_alignment (void) ++{ ++ size_t aligns[] = { ++ 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384 ++ }; ++ ++ return aligns[random () % array_length (aligns)]; ++} ++ ++static void *get_random_alloc (size_t size) ++{ ++ void *retval; ++ size_t align; ++ ++ switch (random() % 3) ++ { ++ case 1: ++ align = get_random_alignment (); ++ retval = aligned_alloc (align, size); ++ aligned_alloc_count++; ++ break; ++ case 2: ++ retval = __libc_calloc (1, size); ++ libc_calloc_count++; ++ break; ++ default: ++ retval = __libc_malloc (size); ++ libc_malloc_count++; ++ break; ++ } ++ ++ return retval; ++} ++ ++ ++void * __random_malloc (size_t size) ++{ ++ return get_random_alloc (size); ++} ++strong_alias (__random_malloc, malloc) +diff --git a/malloc/tst-malloc-random.c b/malloc/tst-malloc-random.c +new file mode 100644 +index 0000000000000000..762b70c918cc6004 +--- /dev/null ++++ b/malloc/tst-malloc-random.c +@@ -0,0 +1,20 @@ ++/* Test malloc with random calls to aligned_alloc and calloc. ++ ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include "tst-malloc.c" +diff --git a/malloc/tst-malloc.c b/malloc/tst-malloc.c +index a7491d3d00f9525f..f7a6e4654c374d01 100644 +--- a/malloc/tst-malloc.c ++++ b/malloc/tst-malloc.c +@@ -18,7 +18,9 @@ + #include + #include + #include ++#include + #include ++#include + + static int errors = 0; + +@@ -35,6 +37,8 @@ do_test (void) + void *p, *q; + int save; + ++ srandom (time (NULL)); ++ + errno = 0; + + DIAG_PUSH_NEEDS_COMMENT; diff --git a/SOURCES/glibc-upstream-2.39-69.patch b/SOURCES/glibc-upstream-2.39-69.patch new file mode 100644 index 0000000..34018ce --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-69.patch @@ -0,0 +1,107 @@ +commit 9de9cd17e73db0ba9af9ef11dc12d490fb59720c +Author: sayan paul +Date: Wed May 29 15:31:04 2024 +0530 + + malloc: New test to check malloc alternate path using memory obstruction + + The test aims to ensure that malloc uses the alternate path to + allocate memory when sbrk() or brk() fails.To achieve this, + the test first creates an obstruction at current program break, + tests that obstruction with a failing sbrk(), then checks if malloc + is still returning a valid ptr thus inferring that malloc() used + mmap() instead of brk() or sbrk() to allocate the memory. + Reviewed-by: Arjun Shankar + Reviewed-by: Zack Weinberg + + (cherry picked from commit 127fc56152347d73cb7c1c283e60e1cb1f15e9f9) + +diff --git a/malloc/Makefile b/malloc/Makefile +index c1a03f3cb02c464d..cc14cf66c9661f99 100644 +--- a/malloc/Makefile ++++ b/malloc/Makefile +@@ -34,6 +34,7 @@ tests := \ + tst-interpose-nothread \ + tst-interpose-thread \ + tst-malloc \ ++ tst-malloc-alternate-path \ + tst-malloc-backtrace \ + tst-malloc-check \ + tst-malloc-fork-deadlock \ +diff --git a/malloc/tst-malloc-alternate-path.c b/malloc/tst-malloc-alternate-path.c +new file mode 100644 +index 0000000000000000..43ae916815d6ff47 +--- /dev/null ++++ b/malloc/tst-malloc-alternate-path.c +@@ -0,0 +1,72 @@ ++/* Test that malloc uses mmap when sbrk or brk fails. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++/* This test sets up an obstruction to ensure that brk/sbrk fails to ++ grow the heap, then verifies that malloc uses mmap for allocations ++ instead. */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++#define LARGE_SIZE (10 * (1 << 20)) // 10 MB ++static long page_size; ++ ++static int ++do_test (void) ++{ ++ /* Get current program break. */ ++ void *current_brk = sbrk (0); ++ ++ page_size = sysconf (_SC_PAGESIZE); ++ ++ /* Round up to the next page boundary. */ ++ void *next_page_boundary = PTR_ALIGN_UP (current_brk, page_size); ++ ++ /* Place a mapping using mmap at the next page boundary. */ ++ void *obstruction_addr ++ = mmap (next_page_boundary, page_size, PROT_READ, ++ MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -1, 0); ++ ++ /* Check if memory obstruction is set up correctly. */ ++ TEST_VERIFY_EXIT (obstruction_addr == next_page_boundary); ++ ++ /* Try to extend the heap beyond the obstruction using sbrk */ ++ int *ptr = sbrk (page_size); ++ TEST_VERIFY_EXIT (ptr == (void *) -1); ++ ++ /* Attempt multiple small allocations using malloc. */ ++ for (size_t i = 0; i < page_size / alignof (max_align_t); i++) ++ { ++ TEST_VERIFY (malloc (alignof (max_align_t))); ++ } ++ ++ /* Attempt to allocate a large block of memory using malloc. */ ++ TEST_VERIFY_EXIT (malloc (LARGE_SIZE) != NULL); ++ ++ /* Check if malloc changed current program break. */ ++ TEST_VERIFY_EXIT (current_brk == sbrk (0)); ++ ++ return 0; ++} ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-7.patch b/SOURCES/glibc-upstream-2.39-7.patch new file mode 100644 index 0000000..5932145 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-7.patch @@ -0,0 +1,46 @@ +commit 71fcdba577884627c3ee4e43beb915da752efb1f +Author: Florian Weimer +Date: Fri Mar 15 19:08:24 2024 +0100 + + linux: Use rseq area unconditionally in sched_getcpu (bug 31479) + + Originally, nptl/descr.h included , but we removed that + in commit 2c6b4b272e6b4d07303af25709051c3e96288f2d ("nptl: + Unconditionally use a 32-byte rseq area"). After that, it was + not ensured that the RSEQ_SIG macro was defined during sched_getcpu.c + compilation that provided a definition. This commit always checks + the rseq area for CPU number information before using the other + approaches. + + This adds an unnecessary (but well-predictable) branch on + architectures which do not define RSEQ_SIG, but its cost is small + compared to the system call. Most architectures that have vDSO + acceleration for getcpu also have rseq support. + + Fixes: 2c6b4b272e6b4d07303af25709051c3e96288f2d + Fixes: 1d350aa06091211863e41169729cee1bca39f72f + Reviewed-by: Arjun Shankar + (cherry picked from commit 7a76f218677d149d8b7875b336722108239f7ee9) + +diff --git a/sysdeps/unix/sysv/linux/sched_getcpu.c b/sysdeps/unix/sysv/linux/sched_getcpu.c +index dfb884568d154537..72a3360550b8667a 100644 +--- a/sysdeps/unix/sysv/linux/sched_getcpu.c ++++ b/sysdeps/unix/sysv/linux/sched_getcpu.c +@@ -33,17 +33,9 @@ vsyscall_sched_getcpu (void) + return r == -1 ? r : cpu; + } + +-#ifdef RSEQ_SIG + int + sched_getcpu (void) + { + int cpu_id = THREAD_GETMEM_VOLATILE (THREAD_SELF, rseq_area.cpu_id); + return __glibc_likely (cpu_id >= 0) ? cpu_id : vsyscall_sched_getcpu (); + } +-#else /* RSEQ_SIG */ +-int +-sched_getcpu (void) +-{ +- return vsyscall_sched_getcpu (); +-} +-#endif /* RSEQ_SIG */ diff --git a/SOURCES/glibc-upstream-2.39-70.patch b/SOURCES/glibc-upstream-2.39-70.patch new file mode 100644 index 0000000..78fef11 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-70.patch @@ -0,0 +1,21 @@ +commit e1d0040a6d1b0714e1ad14abeb7b90bfbd86a8dc +Author: Mike FABIAN +Date: Wed Feb 7 18:41:02 2024 +0100 + + localedata: ssy_ER: Fix syntax error + + (cherry picked from commit 07fd072caff50bca2a7e9f5737a5b38280d2ffda) + +diff --git a/localedata/locales/ssy_ER b/localedata/locales/ssy_ER +index 05d94c9f10398c45..43a16e0f5452ad6a 100644 +--- a/localedata/locales/ssy_ER ++++ b/localedata/locales/ssy_ER +@@ -106,7 +106,7 @@ country_ab3 "ERI" + country_num 232 + country_car "ER" + % country_isbn unknown, Need ISO 2108 +-# https://en.wikipedia.org/wiki/Saho_language has "Saaho" as the endonym but CLDR has "Saho" ++% https://en.wikipedia.org/wiki/Saho_language has "Saaho" as the endonym but CLDR has "Saho" + lang_name "Saho" + lang_ab "" + lang_term "ssy" diff --git a/SOURCES/glibc-upstream-2.39-71.patch b/SOURCES/glibc-upstream-2.39-71.patch new file mode 100644 index 0000000..6e73e1a --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-71.patch @@ -0,0 +1,69 @@ +commit b7f5b0a7114e29577daf64e68970673b61e5fcba +Author: H.J. Lu +Date: Mon Jun 10 13:02:06 2024 -0700 + + x86: Properly set MINIMUM_X86_ISA_LEVEL for i386 [BZ #31867] + + On i386, set the default minimum ISA level to 0, not 1 (baseline which + includes SSE2). There are no changes in config.h nor in config.make on + x86-64. This fixes BZ #31867. + + Signed-off-by: H.J. Lu + Tested-by: Ian Jordan + Reviewed-by: Sam James + Reviewed-by: Florian Weimer + (cherry picked from commit 09bc68b0ac26331a0109f0578c9368e09176da18) + +diff --git a/sysdeps/x86/configure b/sysdeps/x86/configure +index d28d9bcb296c6380..1e2325d0d7212d67 100644 +--- a/sysdeps/x86/configure ++++ b/sysdeps/x86/configure +@@ -139,8 +139,10 @@ libc_cv_have_x86_isa_level=4 + libc_cv_have_x86_isa_level=3 + #elif MINIMUM_X86_ISA_LEVEL == 2 + libc_cv_have_x86_isa_level=2 +-#else ++#elif defined __x86_64__ + libc_cv_have_x86_isa_level=baseline ++#else ++libc_cv_have_x86_isa_level=MINIMUM_X86_ISA_LEVEL + #endif + EOF + eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` +@@ -148,8 +150,10 @@ EOF + fi + { printf "%s\n" "$as_me:${as_lineno-$LINENO}: result: $libc_cv_have_x86_isa_level" >&5 + printf "%s\n" "$libc_cv_have_x86_isa_level" >&6; } +-else ++elif test $base_machine = x86_64; then + libc_cv_have_x86_isa_level=baseline ++else ++ libc_cv_have_x86_isa_level=0 + fi + if test $libc_cv_have_x86_isa_level = baseline; then + printf "%s\n" "#define MINIMUM_X86_ISA_LEVEL 1" >>confdefs.h +diff --git a/sysdeps/x86/configure.ac b/sysdeps/x86/configure.ac +index 5b0acd03d2a30c9b..0b32fdfd4f1bb115 100644 +--- a/sysdeps/x86/configure.ac ++++ b/sysdeps/x86/configure.ac +@@ -96,14 +96,18 @@ libc_cv_have_x86_isa_level=4 + libc_cv_have_x86_isa_level=3 + #elif MINIMUM_X86_ISA_LEVEL == 2 + libc_cv_have_x86_isa_level=2 +-#else ++#elif defined __x86_64__ + libc_cv_have_x86_isa_level=baseline ++#else ++libc_cv_have_x86_isa_level=MINIMUM_X86_ISA_LEVEL + #endif + EOF + eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` + rm -rf conftest*]) +-else ++elif test $base_machine = x86_64; then + libc_cv_have_x86_isa_level=baseline ++else ++ libc_cv_have_x86_isa_level=0 + fi + if test $libc_cv_have_x86_isa_level = baseline; then + AC_DEFINE_UNQUOTED(MINIMUM_X86_ISA_LEVEL, 1) diff --git a/SOURCES/glibc-upstream-2.39-72.patch b/SOURCES/glibc-upstream-2.39-72.patch new file mode 100644 index 0000000..0f06f29 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-72.patch @@ -0,0 +1,82 @@ +commit f05638731eeaae70c53b4fb30bfc58bc3c52a6d5 +Author: H.J. Lu +Date: Tue Jun 11 20:14:56 2024 -0700 + + x86: Properly set x86 minimum ISA level [BZ #31883] + + Properly set libc_cv_have_x86_isa_level in shell for MINIMUM_X86_ISA_LEVEL + defined as + + (__X86_ISA_V1 + __X86_ISA_V2 + __X86_ISA_V3 + __X86_ISA_V4) + + Also set __X86_ISA_V2 to 1 for i386 if __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 + is defined. There are no changes in config.h nor in config.make on x86-64. + On i386, -march=x86-64-v2 with GCC generates + + #define MINIMUM_X86_ISA_LEVEL 2 + + in config.h and + + have-x86-isa-level = 2 + + in config.make. This fixes BZ #31883. + + Signed-off-by: H.J. Lu + Reviewed-by: Noah Goldstein + (cherry picked from commit 29807a271edca3e47195bda0c69ae45e245551a9) + +diff --git a/sysdeps/x86/configure b/sysdeps/x86/configure +index 1e2325d0d7212d67..04c6ba3e6ca53fb6 100644 +--- a/sysdeps/x86/configure ++++ b/sysdeps/x86/configure +@@ -141,8 +141,10 @@ libc_cv_have_x86_isa_level=3 + libc_cv_have_x86_isa_level=2 + #elif defined __x86_64__ + libc_cv_have_x86_isa_level=baseline ++#elif MINIMUM_X86_ISA_LEVEL == 1 ++libc_cv_have_x86_isa_level=1 + #else +-libc_cv_have_x86_isa_level=MINIMUM_X86_ISA_LEVEL ++libc_cv_have_x86_isa_level=0 + #endif + EOF + eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` +diff --git a/sysdeps/x86/configure.ac b/sysdeps/x86/configure.ac +index 0b32fdfd4f1bb115..8a259d3971488e4b 100644 +--- a/sysdeps/x86/configure.ac ++++ b/sysdeps/x86/configure.ac +@@ -98,8 +98,10 @@ libc_cv_have_x86_isa_level=3 + libc_cv_have_x86_isa_level=2 + #elif defined __x86_64__ + libc_cv_have_x86_isa_level=baseline ++#elif MINIMUM_X86_ISA_LEVEL == 1 ++libc_cv_have_x86_isa_level=1 + #else +-libc_cv_have_x86_isa_level=MINIMUM_X86_ISA_LEVEL ++libc_cv_have_x86_isa_level=0 + #endif + EOF + eval `${CC-cc} $CFLAGS $CPPFLAGS $ISAFLAG -I$srcdir -E conftest.c | grep libc_cv_have_x86_isa_level` +diff --git a/sysdeps/x86/isa-level.h b/sysdeps/x86/isa-level.h +index 2c7f74212b9a27e5..03c1fe2bf54e0673 100644 +--- a/sysdeps/x86/isa-level.h ++++ b/sysdeps/x86/isa-level.h +@@ -35,7 +35,17 @@ + # define __X86_ISA_V1 0 + #endif + +-#if __X86_ISA_V1 && defined __GCC_HAVE_SYNC_COMPARE_AND_SWAP_16 \ ++#ifdef __x86_64__ ++# ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_16 ++# define __GCC_HAVE_SYNC_COMPARE_AND_SWAP ++# endif ++#else ++# ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 ++# define __GCC_HAVE_SYNC_COMPARE_AND_SWAP ++# endif ++#endif ++ ++#if __X86_ISA_V1 && defined __GCC_HAVE_SYNC_COMPARE_AND_SWAP \ + && defined HAVE_X86_LAHF_SAHF && defined __POPCNT__ && defined __SSE3__ \ + && defined __SSSE3__ && defined __SSE4_1__ && defined __SSE4_2__ + /* NB: ISAs in x86-64 ISA level v2 are used. */ diff --git a/SOURCES/glibc-upstream-2.39-75.patch b/SOURCES/glibc-upstream-2.39-75.patch new file mode 100644 index 0000000..49b5c42 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-75.patch @@ -0,0 +1,28 @@ +commit 7f9f25f255ee2c00178779fbce502f4b94b848b9 +Author: Florian Weimer +Date: Tue Jun 18 10:56:34 2024 +0200 + + Linux: Include in dl-sysdep.c + + The _dl_sysdep_parse_arguments function contains initalization + of a large on-stack variable: + + dl_parse_auxv_t auxv_values = { 0, }; + + This uses a non-inline version of memset on powerpc64le-linux-gnu, + so it must use the baseline memset. + + (cherry picked from commit f6ea5d1291cf3f264514d03872ebae84e0293b69) + +diff --git a/sysdeps/unix/sysv/linux/dl-sysdep.c b/sysdeps/unix/sysv/linux/dl-sysdep.c +index e1b14e9eb34ff5cb..a8ec2d7c18cc423e 100644 +--- a/sysdeps/unix/sysv/linux/dl-sysdep.c ++++ b/sysdeps/unix/sysv/linux/dl-sysdep.c +@@ -40,6 +40,7 @@ + #include + #include + #include ++#include + + #include + #include diff --git a/SOURCES/glibc-upstream-2.39-76.patch b/SOURCES/glibc-upstream-2.39-76.patch new file mode 100644 index 0000000..6203daa --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-76.patch @@ -0,0 +1,155 @@ +commit 74630b1bb717fb98f4692261f2be8d5c84851fa3 +Author: Adhemerval Zanella +Date: Mon May 6 13:20:56 2024 -0300 + + posix: Fix pidfd_spawn/pidfd_spawnp leak if execve fails (BZ 31695) + + If the pidfd_spawn/pidfd_spawnp helper process succeeds, but evecve + fails for some reason (either with an invalid/non-existent, memory + allocation, etc.) the resulting pidfd is never closed, nor returned + to caller (so it can call close). + + Since the process creation failed, it should be up to posix_spawn to + also, close the file descriptor in this case (similar to what it + does to reap the process). + + This patch also changes the waitpid with waitid (P_PIDFD) for pidfd + case, to avoid a possible pid re-use. + + Checked on x86_64-linux-gnu. + Reviewed-by: Carlos O'Donell + + (cherry picked from commit c90cfce849d010474e8cccf3e5bff49a2c8b141f) + +diff --git a/posix/tst-spawn2.c b/posix/tst-spawn2.c +index bb507204a2b25271..b2bad3f1f7e026fd 100644 +--- a/posix/tst-spawn2.c ++++ b/posix/tst-spawn2.c +@@ -26,6 +26,7 @@ + #include + + #include ++#include + #include + + int +@@ -38,38 +39,53 @@ do_test (void) + char * const args[] = { 0 }; + PID_T_TYPE pid = -1; + +- int ret = POSIX_SPAWN (&pid, program, 0, 0, args, environ); +- if (ret != ENOENT) +- { +- errno = ret; +- FAIL_EXIT1 ("posix_spawn: %m"); +- } +- +- /* POSIX states the value returned on pid variable in case of an error +- is not specified. GLIBC will update the value iff the child +- execution is successful. */ +- if (pid != -1) +- FAIL_EXIT1 ("posix_spawn returned pid != -1 (%i)", (int) pid); +- +- /* Check if no child is actually created. */ +- TEST_COMPARE (WAITID (P_ALL, 0, NULL, WEXITED), -1); +- TEST_COMPARE (errno, ECHILD); +- +- /* Same as before, but with posix_spawnp. */ +- char *args2[] = { (char*) program, 0 }; +- +- ret = POSIX_SPAWNP (&pid, args2[0], 0, 0, args2, environ); +- if (ret != ENOENT) +- { +- errno = ret; +- FAIL_EXIT1 ("posix_spawnp: %m"); +- } +- +- if (pid != -1) +- FAIL_EXIT1 ("posix_spawnp returned pid != -1 (%i)", (int) pid); +- +- TEST_COMPARE (WAITID (P_ALL, 0, NULL, WEXITED), -1); +- TEST_COMPARE (errno, ECHILD); ++ { ++ struct support_descriptors *descrs = support_descriptors_list (); ++ ++ int ret = POSIX_SPAWN (&pid, program, 0, 0, args, environ); ++ if (ret != ENOENT) ++ { ++ errno = ret; ++ FAIL_EXIT1 ("posix_spawn: %m"); ++ } ++ ++ /* POSIX states the value returned on pid variable in case of an error ++ is not specified. GLIBC will update the value iff the child ++ execution is successful. */ ++ if (pid != -1) ++ FAIL_EXIT1 ("posix_spawn returned pid != -1 (%i)", (int) pid); ++ ++ /* Check if no child is actually created. */ ++ TEST_COMPARE (WAITID (P_ALL, 0, NULL, WEXITED), -1); ++ TEST_COMPARE (errno, ECHILD); ++ ++ /* Also check if there is no leak descriptors. */ ++ support_descriptors_check (descrs); ++ support_descriptors_free (descrs); ++ } ++ ++ { ++ /* Same as before, but with posix_spawnp. */ ++ char *args2[] = { (char*) program, 0 }; ++ ++ struct support_descriptors *descrs = support_descriptors_list (); ++ ++ int ret = POSIX_SPAWNP (&pid, args2[0], 0, 0, args2, environ); ++ if (ret != ENOENT) ++ { ++ errno = ret; ++ FAIL_EXIT1 ("posix_spawnp: %m"); ++ } ++ ++ if (pid != -1) ++ FAIL_EXIT1 ("posix_spawnp returned pid != -1 (%i)", (int) pid); ++ ++ TEST_COMPARE (WAITID (P_ALL, 0, NULL, WEXITED), -1); ++ TEST_COMPARE (errno, ECHILD); ++ ++ support_descriptors_check (descrs); ++ support_descriptors_free (descrs); ++ } + + return 0; + } +diff --git a/sysdeps/unix/sysv/linux/spawni.c b/sysdeps/unix/sysv/linux/spawni.c +index e8ed2babb9b13969..f57e92815eaf3c7b 100644 +--- a/sysdeps/unix/sysv/linux/spawni.c ++++ b/sysdeps/unix/sysv/linux/spawni.c +@@ -449,13 +449,22 @@ __spawnix (int *pid, const char *file, + caller to actually collect it. */ + ec = args.err; + if (ec > 0) +- /* There still an unlikely case where the child is cancelled after +- setting args.err, due to a positive error value. Also there is +- possible pid reuse race (where the kernel allocated the same pid +- to an unrelated process). Unfortunately due synchronization +- issues where the kernel might not have the process collected +- the waitpid below can not use WNOHANG. */ +- __waitpid (new_pid, NULL, 0); ++ { ++ /* There still an unlikely case where the child is cancelled after ++ setting args.err, due to a positive error value. Also there is ++ possible pid reuse race (where the kernel allocated the same pid ++ to an unrelated process). Unfortunately due synchronization ++ issues where the kernel might not have the process collected ++ the waitpid below can not use WNOHANG. */ ++ __waitid (use_pidfd ? P_PIDFD : P_PID, ++ use_pidfd ? args.pidfd : new_pid, ++ NULL, ++ WEXITED); ++ /* For pidfd we need to also close the file descriptor for the case ++ where execve fails. */ ++ if (use_pidfd) ++ __close_nocancel_nostatus (args.pidfd); ++ } + } + else + ec = errno; diff --git a/SOURCES/glibc-upstream-2.39-77.patch b/SOURCES/glibc-upstream-2.39-77.patch new file mode 100644 index 0000000..f4c45e7 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-77.patch @@ -0,0 +1,107 @@ +commit 6cb25aff8583421cf7a55ddd40abd7bd49176c60 +Author: Adhemerval Zanella +Date: Thu Mar 14 14:59:35 2024 -0300 + + math: Fix i386 and m68k fmod/fmodf on static build (BZ 31488) + + The commit 16439f419b removed the static fmod/fmodf on i386 and m68k + with and empty w_fmod.c (required for the ABIs that uses the newly + implementation). This patch fixes by adding the required symbols on + the arch-specific w_fmod{f}_compat.c implementation. + + To statically build fmod fails on some ABI (alpha, s390, sparc) because + it does not export the ldexpf128, this is also fixed by this patch. + + Checked on i686-linux-gnu and with a build for m68k-linux-gnu. + + Reviewed-by: Aurelien Jarno + Tested-by: Aurelien Jarno + (cherry picked from commit 0b716305dfb48c2d13ed4f7d06c082b90c1d226f) + +diff --git a/sysdeps/i386/fpu/w_fmod_compat.c b/sysdeps/i386/fpu/w_fmod_compat.c +index 5ac9995ffd862a7f..528bfc2a135b5251 100644 +--- a/sysdeps/i386/fpu/w_fmod_compat.c ++++ b/sysdeps/i386/fpu/w_fmod_compat.c +@@ -7,8 +7,9 @@ + # define LIBM_SVID_COMPAT 1 + # undef compat_symbol + # define compat_symbol(a, b, c, d) +-#endif +-#include +-#ifdef SHARED ++# include + libm_alias_double (__fmod_compat, fmod) ++#else ++#include ++#include + #endif +diff --git a/sysdeps/i386/fpu/w_fmodf_compat.c b/sysdeps/i386/fpu/w_fmodf_compat.c +index cc417e07d39b271d..5a61693e51f98f8f 100644 +--- a/sysdeps/i386/fpu/w_fmodf_compat.c ++++ b/sysdeps/i386/fpu/w_fmodf_compat.c +@@ -7,8 +7,9 @@ + # define LIBM_SVID_COMPAT 1 + # undef compat_symbol + # define compat_symbol(a, b, c, d) +-#endif +-#include +-#ifdef SHARED ++# include + libm_alias_float (__fmod_compat, fmod) ++#else ++#include ++#include + #endif +diff --git a/sysdeps/ieee754/ldbl-opt/s_ldexpl.c b/sysdeps/ieee754/ldbl-opt/s_ldexpl.c +index 1afbe7d8adbb00ae..932cc4341c53dec3 100644 +--- a/sysdeps/ieee754/ldbl-opt/s_ldexpl.c ++++ b/sysdeps/ieee754/ldbl-opt/s_ldexpl.c +@@ -17,13 +17,13 @@ + License along with the GNU C Library; if not, see + . */ + +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + # define declare_mgen_alias(f,t) + #endif + #include + #include + +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + long_double_symbol (libc, __ldexpl, ldexpl); + long_double_symbol (libc, __wrap_scalbnl, scalbnl); + #endif +diff --git a/sysdeps/m68k/m680x0/fpu/w_fmod_compat.c b/sysdeps/m68k/m680x0/fpu/w_fmod_compat.c +index 527d4fbed201d4b4..57f38091e6724848 100644 +--- a/sysdeps/m68k/m680x0/fpu/w_fmod_compat.c ++++ b/sysdeps/m68k/m680x0/fpu/w_fmod_compat.c +@@ -7,8 +7,9 @@ + # define LIBM_SVID_COMPAT 1 + # undef compat_symbol + # define compat_symbol(a, b, c, d) +-#endif + #include +-#ifdef SHARED + libm_alias_double (__fmod_compat, fmod) ++#else ++#include ++#include + #endif +diff --git a/sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c b/sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c +index 5043586b910e765f..88db07f443b0c339 100644 +--- a/sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c ++++ b/sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c +@@ -7,8 +7,9 @@ + # define LIBM_SVID_COMPAT 1 + # undef compat_symbol + # define compat_symbol(a, b, c, d) +-#endif +-#include +-#ifdef SHARED ++# include + libm_alias_float (__fmod_compat, fmod) ++#else ++#include ++#include + #endif diff --git a/SOURCES/glibc-upstream-2.39-78.patch b/SOURCES/glibc-upstream-2.39-78.patch new file mode 100644 index 0000000..2f3022c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-78.patch @@ -0,0 +1,46 @@ +commit 3950cbd7a18b48c0f272ca41fa60cc135ff3175a +Author: Adhemerval Zanella +Date: Wed Mar 20 11:31:54 2024 -0300 + + math: Fix i386 and m68k exp10 on static build (BZ 31775) + + The commit 08ddd26814 removed the static exp10 on i386 and m68k with an + empty w_exp10.c (required for the ABIs that uses the newly + implementation). This patch fixes by adding the required symbols on the + arch-specific w_exp{f}_compat.c implementation. + + Checked on i686-linux-gnu and with a build for m68k-linux-gnu. + + Reviewed-by: Aurelien Jarno + (cherry picked from commit 1f09aae36aa185b8b1100dfa6b776442279bf580) + +diff --git a/sysdeps/i386/fpu/w_exp10_compat.c b/sysdeps/i386/fpu/w_exp10_compat.c +index b53455386ed8e189..49a0e03385e4bfe3 100644 +--- a/sysdeps/i386/fpu/w_exp10_compat.c ++++ b/sysdeps/i386/fpu/w_exp10_compat.c +@@ -1,3 +1,8 @@ + /* i386 provides an optimized __ieee754_exp10. */ +-#define NO_COMPAT_NEEDED 1 +-#include ++#ifdef SHARED ++# define NO_COMPAT_NEEDED 1 ++# include ++#else ++# include ++# include ++#endif +diff --git a/sysdeps/m68k/m680x0/fpu/w_exp10_compat.c b/sysdeps/m68k/m680x0/fpu/w_exp10_compat.c +index 0d3e7186261041e3..350f2e4b4d37e569 100644 +--- a/sysdeps/m68k/m680x0/fpu/w_exp10_compat.c ++++ b/sysdeps/m68k/m680x0/fpu/w_exp10_compat.c +@@ -1,3 +1,8 @@ + /* m68k provides an optimized __ieee754_exp10. */ +-#define NO_COMPAT_NEEDED 1 +-#include ++#ifdef SHARED ++# define NO_COMPAT_NEEDED 1 ++# include ++#else ++# include ++# include ++#endif diff --git a/SOURCES/glibc-upstream-2.39-79.patch b/SOURCES/glibc-upstream-2.39-79.patch new file mode 100644 index 0000000..fe3ef5c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-79.patch @@ -0,0 +1,30 @@ +commit d473c9bb3b6b0448985cd195c558de2a49fd2dc2 +Author: Adhemerval Zanella +Date: Tue May 21 13:52:54 2024 -0300 + + math: Fix isnanf128 static build (BZ 31774) + + Some static implementation of float128 routines might call __isnanf128, + which is not provided by the static object. + + Checked on x86_64-linux-gnu. + Reviewed-by: H.J. Lu + + (cherry picked from commit 5d4999e519ec77e75bef920e2540e8605015680a) + +diff --git a/sysdeps/ieee754/float128/s_isnanf128.c b/sysdeps/ieee754/float128/s_isnanf128.c +index 59f71533cedc752e..b73a4e80d707eeb8 100644 +--- a/sysdeps/ieee754/float128/s_isnanf128.c ++++ b/sysdeps/ieee754/float128/s_isnanf128.c +@@ -11,7 +11,11 @@ + #include "../ldbl-128/s_isnanl.c" + #if !IS_IN (libm) + #include ++#ifdef SHARED + hidden_ver (__isnanf128_impl, __isnanf128) ++#else ++strong_alias (__isnanf128_impl, __isnanf128) ++#endif + _weak_alias (__isnanf128_impl, isnanl) + versioned_symbol (libc, __isnanf128_impl, __isnanf128, GLIBC_2_34); + #if (SHLIB_COMPAT (libc, FLOAT128_VERSION_M, GLIBC_2_34)) diff --git a/SOURCES/glibc-upstream-2.39-8.patch b/SOURCES/glibc-upstream-2.39-8.patch new file mode 100644 index 0000000..653d5a1 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-8.patch @@ -0,0 +1,284 @@ +commit ee7f4c54e19738c2c27d3846e1e9b3595c89221f +Author: Manjunath Matti +Date: Tue Mar 19 15:29:48 2024 -0500 + + powerpc: Add HWCAP3/HWCAP4 data to TCB for Power Architecture. + + This patch adds a new feature for powerpc. In order to get faster + access to the HWCAP3/HWCAP4 masks, similar to HWCAP/HWCAP2 (i.e. for + implementing __builtin_cpu_supports() in GCC) without the overhead of + reading them from the auxiliary vector, we now reserve space for them + in the TCB. + + Suggested-by: Peter Bergner + Reviewed-by: Peter Bergner + (cherry picked from commit 3ab9b88e2ac91062b6d493fe32bd101a55006c6a) + +diff --git a/elf/dl-diagnostics.c b/elf/dl-diagnostics.c +index 7345ebc4e586883f..aaf67b87e81b04c8 100644 +--- a/elf/dl-diagnostics.c ++++ b/elf/dl-diagnostics.c +@@ -235,6 +235,8 @@ _dl_print_diagnostics (char **environ) + _dl_diagnostics_print_labeled_value ("dl_hwcap", GLRO (dl_hwcap)); + _dl_diagnostics_print_labeled_value ("dl_hwcap_important", HWCAP_IMPORTANT); + _dl_diagnostics_print_labeled_value ("dl_hwcap2", GLRO (dl_hwcap2)); ++ _dl_diagnostics_print_labeled_value ("dl_hwcap3", GLRO (dl_hwcap3)); ++ _dl_diagnostics_print_labeled_value ("dl_hwcap4", GLRO (dl_hwcap4)); + _dl_diagnostics_print_labeled_string + ("dl_hwcaps_subdirs", _dl_hwcaps_subdirs); + _dl_diagnostics_print_labeled_value +diff --git a/elf/dl-support.c b/elf/dl-support.c +index 2f502c8b0d27b784..451932dd03e971b8 100644 +--- a/elf/dl-support.c ++++ b/elf/dl-support.c +@@ -158,6 +158,8 @@ const ElfW(Phdr) *_dl_phdr; + size_t _dl_phnum; + uint64_t _dl_hwcap; + uint64_t _dl_hwcap2; ++uint64_t _dl_hwcap3; ++uint64_t _dl_hwcap4; + + enum dso_sort_algorithm _dl_dso_sort_algo; + +diff --git a/elf/elf.h b/elf/elf.h +index 455731663c6ed339..1c394c64cd5c66ed 100644 +--- a/elf/elf.h ++++ b/elf/elf.h +@@ -1234,6 +1234,10 @@ typedef struct + #define AT_RSEQ_FEATURE_SIZE 27 /* rseq supported feature size. */ + #define AT_RSEQ_ALIGN 28 /* rseq allocation alignment. */ + ++/* More machine-dependent hints about processor capabilities. */ ++#define AT_HWCAP3 29 /* extension of AT_HWCAP. */ ++#define AT_HWCAP4 30 /* extension of AT_HWCAP. */ ++ + #define AT_EXECFN 31 /* Filename of executable. */ + + /* Pointer to the global system page used for system calls and other +diff --git a/sysdeps/generic/ldsodefs.h b/sysdeps/generic/ldsodefs.h +index 117c901ccc5c5f0b..50f58a60e3f02330 100644 +--- a/sysdeps/generic/ldsodefs.h ++++ b/sysdeps/generic/ldsodefs.h +@@ -646,6 +646,8 @@ struct rtld_global_ro + /* Mask for more hardware capabilities that are available on some + platforms. */ + EXTERN uint64_t _dl_hwcap2; ++ EXTERN uint64_t _dl_hwcap3; ++ EXTERN uint64_t _dl_hwcap4; + + EXTERN enum dso_sort_algorithm _dl_dso_sort_algo; + +diff --git a/sysdeps/powerpc/dl-procinfo.c b/sysdeps/powerpc/dl-procinfo.c +index a76bb6e5b0895e3f..8cf00aa7e359bb6a 100644 +--- a/sysdeps/powerpc/dl-procinfo.c ++++ b/sysdeps/powerpc/dl-procinfo.c +@@ -38,6 +38,10 @@ + needed. + */ + ++/* The total number of available bits (including those prior to ++ _DL_HWCAP_FIRST). Some of these bits might not be used. */ ++#define _DL_HWCAP_COUNT 128 ++ + #ifndef PROCINFO_CLASS + # define PROCINFO_CLASS + #endif +@@ -61,7 +65,7 @@ PROCINFO_CLASS struct cpu_features _dl_powerpc_cpu_features + #if !defined PROCINFO_DECL && defined SHARED + ._dl_powerpc_cap_flags + #else +-PROCINFO_CLASS const char _dl_powerpc_cap_flags[64][15] ++PROCINFO_CLASS const char _dl_powerpc_cap_flags[_DL_HWCAP_COUNT][15] + #endif + #ifndef PROCINFO_DECL + = { +diff --git a/sysdeps/powerpc/dl-procinfo.h b/sysdeps/powerpc/dl-procinfo.h +index 68f424109501aaef..f8cb343877386402 100644 +--- a/sysdeps/powerpc/dl-procinfo.h ++++ b/sysdeps/powerpc/dl-procinfo.h +@@ -22,16 +22,17 @@ + #include + #include /* This defines the PPC_FEATURE[2]_* macros. */ + +-/* The total number of available bits (including those prior to +- _DL_HWCAP_FIRST). Some of these bits might not be used. */ +-#define _DL_HWCAP_COUNT 64 ++/* Feature masks are all 32-bits in size. */ ++#define _DL_HWCAP_SIZE 32 + +-/* Features started at bit 31 and decremented as new features were added. */ +-#define _DL_HWCAP_LAST 31 ++/* AT_HWCAP2 feature strings follow the AT_HWCAP feature strings. */ ++#define _DL_HWCAP2_OFFSET _DL_HWCAP_SIZE + +-/* AT_HWCAP2 features started at bit 31 and decremented as new features were +- added. HWCAP2 feature bits start at bit 0. */ +-#define _DL_HWCAP2_LAST 31 ++/* AT_HWCAP3 feature strings follow the AT_HWCAP2 feature strings. */ ++#define _DL_HWCAP3_OFFSET (_DL_HWCAP2_OFFSET + _DL_HWCAP_SIZE) ++ ++/* AT_HWCAP4 feature strings follow the AT_HWCAP3 feature strings. */ ++#define _DL_HWCAP4_OFFSET (_DL_HWCAP3_OFFSET + _DL_HWCAP_SIZE) + + /* These bits influence library search. */ + #define HWCAP_IMPORTANT (PPC_FEATURE_HAS_ALTIVEC \ +@@ -187,21 +188,42 @@ _dl_procinfo (unsigned int type, unsigned long int word) + case AT_HWCAP: + _dl_printf ("AT_HWCAP: "); + +- for (int i = 0; i <= _DL_HWCAP_LAST; ++i) ++ for (int i = 0; i < _DL_HWCAP_SIZE; ++i) + if (word & (1 << i)) + _dl_printf (" %s", _dl_hwcap_string (i)); + break; + case AT_HWCAP2: + { +- unsigned int offset = _DL_HWCAP_LAST + 1; + + _dl_printf ("AT_HWCAP2: "); + +- /* We have to go through them all because the kernel added the +- AT_HWCAP2 features starting with the high bits. */ +- for (int i = 0; i <= _DL_HWCAP2_LAST; ++i) +- if (word & (1 << i)) +- _dl_printf (" %s", _dl_hwcap_string (offset + i)); ++ /* We have to go through them all because the kernel added the ++ AT_HWCAP2 features starting with the high bits. */ ++ for (int i = 0; i < _DL_HWCAP_SIZE; ++i) ++ if (word & (1 << i)) ++ _dl_printf (" %s", _dl_hwcap_string (_DL_HWCAP2_OFFSET + i)); ++ break; ++ } ++ case AT_HWCAP3: ++ { ++ _dl_printf ("AT_HWCAP3: "); ++ ++ /* We have to go through them all because the kernel added the ++ AT_HWCAP3 features starting with the high bits. */ ++ for (int i = 0; i < _DL_HWCAP_SIZE; ++i) ++ if (word & (1 << i)) ++ _dl_printf (" %s", _dl_hwcap_string (_DL_HWCAP3_OFFSET + i)); ++ break; ++ } ++ case AT_HWCAP4: ++ { ++ _dl_printf ("AT_HWCAP4: "); ++ ++ /* We have to go through them all because the kernel added the ++ AT_HWCAP4 features starting with the high bits. */ ++ for (int i = 0; i <= _DL_HWCAP_SIZE; ++i) ++ if (word & (1 << i)) ++ _dl_printf (" %s", _dl_hwcap_string (_DL_HWCAP4_OFFSET + i)); + break; + } + case AT_L1I_CACHEGEOMETRY: +diff --git a/sysdeps/powerpc/hwcapinfo.c b/sysdeps/powerpc/hwcapinfo.c +index 76344f285a903858..f6fede15a7dfbf6c 100644 +--- a/sysdeps/powerpc/hwcapinfo.c ++++ b/sysdeps/powerpc/hwcapinfo.c +@@ -31,7 +31,7 @@ void + __tcb_parse_hwcap_and_convert_at_platform (void) + { + +- uint64_t h1, h2; ++ uint64_t h1, h2, h3, h4; + + /* Read AT_PLATFORM string from auxv and convert it to a number. */ + __tcb.at_platform = _dl_string_platform (GLRO (dl_platform)); +@@ -39,6 +39,8 @@ __tcb_parse_hwcap_and_convert_at_platform (void) + /* Read HWCAP and HWCAP2 from auxv. */ + h1 = GLRO (dl_hwcap); + h2 = GLRO (dl_hwcap2); ++ h3 = GLRO (dl_hwcap3); ++ h4 = GLRO (dl_hwcap4); + + /* hwcap contains only the latest supported ISA, the code checks which is + and fills the previous supported ones. */ +@@ -64,13 +66,16 @@ __tcb_parse_hwcap_and_convert_at_platform (void) + else if (h1 & PPC_FEATURE_POWER5) + h1 |= PPC_FEATURE_POWER4; + +- uint64_t array_hwcaps[] = { h1, h2 }; ++ uint64_t array_hwcaps[] = { h1, h2, h3, h4 }; + init_cpu_features (&GLRO(dl_powerpc_cpu_features), array_hwcaps); + + /* Consolidate both HWCAP and HWCAP2 into a single doubleword so that + we can read both in a single load later. */ + __tcb.hwcap = (h1 << 32) | (h2 & 0xffffffff); +- __tcb.hwcap_extn = 0x0; ++ ++ /* Consolidate both HWCAP3 and HWCAP4 into a single doubleword so that ++ we can read both in a single load later. */ ++ __tcb.hwcap_extn = (h3 << 32) | (h4 & 0xffffffff); + + } + #if IS_IN (rtld) +diff --git a/sysdeps/unix/sysv/linux/dl-parse_auxv.h b/sysdeps/unix/sysv/linux/dl-parse_auxv.h +index e3d758b163c619df..ea2a58ecb1668774 100644 +--- a/sysdeps/unix/sysv/linux/dl-parse_auxv.h ++++ b/sysdeps/unix/sysv/linux/dl-parse_auxv.h +@@ -47,6 +47,8 @@ void _dl_parse_auxv (ElfW(auxv_t) *av, dl_parse_auxv_t auxv_values) + GLRO(dl_platform) = (void *) auxv_values[AT_PLATFORM]; + GLRO(dl_hwcap) = auxv_values[AT_HWCAP]; + GLRO(dl_hwcap2) = auxv_values[AT_HWCAP2]; ++ GLRO(dl_hwcap3) = auxv_values[AT_HWCAP3]; ++ GLRO(dl_hwcap4) = auxv_values[AT_HWCAP4]; + GLRO(dl_clktck) = auxv_values[AT_CLKTCK]; + GLRO(dl_fpu_control) = auxv_values[AT_FPUCW]; + _dl_random = (void *) auxv_values[AT_RANDOM]; +diff --git a/sysdeps/unix/sysv/linux/dl-sysdep.c b/sysdeps/unix/sysv/linux/dl-sysdep.c +index ad3692d73839d7a3..e1b14e9eb34ff5cb 100644 +--- a/sysdeps/unix/sysv/linux/dl-sysdep.c ++++ b/sysdeps/unix/sysv/linux/dl-sysdep.c +@@ -197,6 +197,8 @@ _dl_show_auxv (void) + [AT_SYSINFO_EHDR - 2] = { "SYSINFO_EHDR: 0x", hex }, + [AT_RANDOM - 2] = { "RANDOM: 0x", hex }, + [AT_HWCAP2 - 2] = { "HWCAP2: 0x", hex }, ++ [AT_HWCAP3 - 2] = { "HWCAP3: 0x", hex }, ++ [AT_HWCAP4 - 2] = { "HWCAP4: 0x", hex }, + [AT_MINSIGSTKSZ - 2] = { "MINSIGSTKSZ: ", dec }, + [AT_L1I_CACHESIZE - 2] = { "L1I_CACHESIZE: ", dec }, + [AT_L1I_CACHEGEOMETRY - 2] = { "L1I_CACHEGEOMETRY: 0x", hex }, +diff --git a/sysdeps/unix/sysv/linux/powerpc/cpu-features.c b/sysdeps/unix/sysv/linux/powerpc/cpu-features.c +index 8e8a5ec2eab7e8c6..a947d62db63965b1 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/cpu-features.c ++++ b/sysdeps/unix/sysv/linux/powerpc/cpu-features.c +@@ -94,6 +94,8 @@ init_cpu_features (struct cpu_features *cpu_features, uint64_t hwcaps[]) + which are set by __tcb_parse_hwcap_and_convert_at_platform. */ + cpu_features->hwcap = hwcaps[0]; + cpu_features->hwcap2 = hwcaps[1]; ++ cpu_features->hwcap3 = hwcaps[2]; ++ cpu_features->hwcap4 = hwcaps[3]; + /* Default is to use aligned memory access on optimized function unless + tunables is enable, since for this case user can explicit disable + unaligned optimizations. */ +diff --git a/sysdeps/unix/sysv/linux/powerpc/cpu-features.h b/sysdeps/unix/sysv/linux/powerpc/cpu-features.h +index 1294f0b601ebf54f..e9eb6a13c8ab11d7 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/cpu-features.h ++++ b/sysdeps/unix/sysv/linux/powerpc/cpu-features.h +@@ -26,6 +26,8 @@ struct cpu_features + bool use_cached_memopt; + unsigned long int hwcap; + unsigned long int hwcap2; ++ unsigned long int hwcap3; ++ unsigned long int hwcap4; + }; + + static const char hwcap_names[] = { +diff --git a/sysdeps/unix/sysv/linux/powerpc/libc-start.c b/sysdeps/unix/sysv/linux/powerpc/libc-start.c +index a4705daf1cdea4de..6a00cd88cd64b992 100644 +--- a/sysdeps/unix/sysv/linux/powerpc/libc-start.c ++++ b/sysdeps/unix/sysv/linux/powerpc/libc-start.c +@@ -87,6 +87,12 @@ __libc_start_main_impl (int argc, char **argv, + case AT_HWCAP2: + _dl_hwcap2 = (unsigned long int) av->a_un.a_val; + break; ++ case AT_HWCAP3: ++ _dl_hwcap3 = (unsigned long int) av->a_un.a_val; ++ break; ++ case AT_HWCAP4: ++ _dl_hwcap4 = (unsigned long int) av->a_un.a_val; ++ break; + case AT_PLATFORM: + _dl_platform = (void *) av->a_un.a_val; + break; diff --git a/SOURCES/glibc-upstream-2.39-80.patch b/SOURCES/glibc-upstream-2.39-80.patch new file mode 100644 index 0000000..793f597 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-80.patch @@ -0,0 +1,66 @@ +commit d2cbfcf1d9a4b539007fe04d33e0bdb82d02a2f2 +Author: Adhemerval Zanella +Date: Wed May 22 15:07:12 2024 -0300 + + math: Provide missing math symbols on libc.a (BZ 31781) + + The libc.a for alpha, s390, and sparcv9 does not provide + copysignf64x, copysignf128, frexpf64x, frexpf128, modff64x, and + modff128. + + Checked with a static build for the affected ABIs. + Reviewed-by: H.J. Lu + + (cherry picked from commit eaa8113bf0eb599025e3efdbe1bb214ee8dc645a) + +diff --git a/sysdeps/ieee754/ldbl-64-128/s_copysignl.c b/sysdeps/ieee754/ldbl-64-128/s_copysignl.c +index 11b42d04ba61b446..80137847d32c7c08 100644 +--- a/sysdeps/ieee754/ldbl-64-128/s_copysignl.c ++++ b/sysdeps/ieee754/ldbl-64-128/s_copysignl.c +@@ -1,10 +1,10 @@ + #include + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + # undef libm_alias_ldouble + # define libm_alias_ldouble(from, to) + #endif + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + long_double_symbol (libc, __copysignl, copysignl); + #endif +diff --git a/sysdeps/ieee754/ldbl-64-128/s_frexpl.c b/sysdeps/ieee754/ldbl-64-128/s_frexpl.c +index 73ac41e40c214394..f5f7d349f78ab64e 100644 +--- a/sysdeps/ieee754/ldbl-64-128/s_frexpl.c ++++ b/sysdeps/ieee754/ldbl-64-128/s_frexpl.c +@@ -1,10 +1,10 @@ + #include + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + # undef libm_alias_ldouble + # define libm_alias_ldouble(from, to) + #endif + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + long_double_symbol (libc, __frexpl, frexpl); + #endif +diff --git a/sysdeps/ieee754/ldbl-64-128/s_modfl.c b/sysdeps/ieee754/ldbl-64-128/s_modfl.c +index 7d7aeae111fc6e7a..ba3d31334a0fefa0 100644 +--- a/sysdeps/ieee754/ldbl-64-128/s_modfl.c ++++ b/sysdeps/ieee754/ldbl-64-128/s_modfl.c +@@ -1,10 +1,10 @@ + #include + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + # undef libm_alias_ldouble + # define libm_alias_ldouble(from, to) + #endif + #include +-#if IS_IN (libc) ++#if IS_IN (libc) && defined SHARED + long_double_symbol (libc, __modfl, modfl); + #endif diff --git a/SOURCES/glibc-upstream-2.39-81.patch b/SOURCES/glibc-upstream-2.39-81.patch new file mode 100644 index 0000000..9f070f2 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-81.patch @@ -0,0 +1,32 @@ +commit a03631124602f2dcef40d46660b96d2e51c44bfd +Author: Adhemerval Zanella +Date: Fri Apr 5 10:27:29 2024 -0300 + + stdlib: fix arc4random fallback to /dev/urandom (BZ 31612) + + The __getrandom_nocancel used by __arc4random_buf uses + INLINE_SYSCALL_CALL (which returns -1/errno) and the loop checks for + the return value instead of errno to fallback to /dev/urandom. + + The malloc code now uses __getrandom_nocancel_nostatus, which uses + INTERNAL_SYSCALL_CALL, so there is no need to use the variant that does + not set errno (BZ#29624). + + Checked on x86_64-linux-gnu. + + Reviewed-by: Xi Ruoyao + (cherry picked from commit 184b9e530e6326e668709826903b6d30dc6cac3f) + +diff --git a/stdlib/arc4random.c b/stdlib/arc4random.c +index 3ae8fc130234b04d..7818cb9cf66e0f3b 100644 +--- a/stdlib/arc4random.c ++++ b/stdlib/arc4random.c +@@ -51,7 +51,7 @@ __arc4random_buf (void *p, size_t n) + n -= l; + continue; /* Interrupted by a signal; keep going. */ + } +- else if (l == -ENOSYS) ++ else if (l < 0 && errno == ENOSYS) + break; /* No syscall, so fallback to /dev/urandom. */ + arc4random_getrandom_failure (); + } diff --git a/SOURCES/glibc-upstream-2.39-82.patch b/SOURCES/glibc-upstream-2.39-82.patch new file mode 100644 index 0000000..26e6b1a --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-82.patch @@ -0,0 +1,53 @@ +commit 5c46e6b66636be0010e9a732d5ba1e65ebd54687 +Author: Stefan Liebler +Date: Thu Jul 11 11:28:53 2024 +0200 + + s390x: Fix segfault in wcsncmp [BZ #31934] + + The z13/vector-optimized wcsncmp implementation segfaults if n=1 + and there is only one character (equal on both strings) before + the page end. Then it loads and compares one character and misses + to check n again. The following load fails. + + This patch removes the extra load and compare of the first character + and just start with the loop which uses vector-load-to-block-boundary. + This code-path also checks n. + + With this patch both tests are passing: + - the simplified one mentioned in the bugzilla 31934 + - the full one in Florian Weimer's patch: + "manual: Document a GNU extension for strncmp/wcsncmp" + (https://patchwork.sourceware.org/project/glibc/patch/874j9eml6y.fsf@oldenburg.str.redhat.com/): + On s390x-linux-gnu (z16), the new wcsncmp test fails due to bug 31934. + Reviewed-by: Carlos O'Donell + + (cherry picked from commit 9b7651410375ec8848a1944992d663d514db4ba7) + +diff --git a/sysdeps/s390/wcsncmp-vx.S b/sysdeps/s390/wcsncmp-vx.S +index bf6dfa6bc2b904b5..8b081567a2a5732c 100644 +--- a/sysdeps/s390/wcsncmp-vx.S ++++ b/sysdeps/s390/wcsncmp-vx.S +@@ -59,14 +59,7 @@ ENTRY(WCSNCMP_Z13) + sllg %r4,%r4,2 /* Convert character-count to byte-count. */ + locgrne %r4,%r1 /* Use max byte-count, if bit 0/1 was one. */ + +- /* Check first character without vector load. */ +- lghi %r5,4 /* current_len = 4 bytes. */ +- /* Check s1/2[0]. */ +- lt %r0,0(%r2) +- l %r1,0(%r3) +- je .Lend_cmp_one_char +- crjne %r0,%r1,.Lend_cmp_one_char +- ++ lghi %r5,0 /* current_len = 0 bytes. */ + .Lloop: + vlbb %v17,0(%r5,%r3),6 /* Load s2 to block boundary. */ + vlbb %v16,0(%r5,%r2),6 /* Load s1 to block boundary. */ +@@ -167,7 +160,6 @@ ENTRY(WCSNCMP_Z13) + srl %r4,2 /* And convert it to character-index. */ + vlgvf %r0,%v16,0(%r4) /* Load character-values. */ + vlgvf %r1,%v17,0(%r4) +-.Lend_cmp_one_char: + cr %r0,%r1 + je .Lend_equal + lghi %r2,1 diff --git a/SOURCES/glibc-upstream-2.39-83.patch b/SOURCES/glibc-upstream-2.39-83.patch new file mode 100644 index 0000000..6b531df --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-83.patch @@ -0,0 +1,153 @@ +commit 1062ebbd1911ec9efe909765ca3ee3809ec3fd67 +Author: Michael Jeanson +Date: Wed Jul 3 12:35:34 2024 -0400 + + nptl: fix potential merge of __rseq_* relro symbols + + While working on a patch to add support for the extensible rseq ABI, we + came across an issue where a new 'const' variable would be merged with + the existing '__rseq_size' variable. We tracked this to the use of + '-fmerge-all-constants' which allows the compiler to merge identical + constant variables. This means that all 'const' variables in a compile + unit that are of the same size and are initialized to the same value can + be merged. + + In this specific case, on 32 bit systems 'unsigned int' and 'ptrdiff_t' + are both 4 bytes and initialized to 0 which should trigger the merge. + However for reasons we haven't delved into when the attribute 'section + (".data.rel.ro")' is added to the mix, only variables of the same exact + types are merged. As far as we know this behavior is not specified + anywhere and could change with a new compiler version, hence this patch. + + Move the definitions of these variables into an assembler file and add + hidden writable aliases for internal use. This has the added bonus of + removing the asm workaround to set the values on rseq registration. + + Tested on Debian 12 with GCC 12.2. + + Signed-off-by: Michael Jeanson + Reviewed-by: Mathieu Desnoyers + Reviewed-by: Florian Weimer + (cherry picked from commit 2b92982e2369d292560793bee8e730f695f48ff3) + +diff --git a/elf/Makefile b/elf/Makefile +index a50a988e7362cf3b..0049ffa13c8d3e51 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -73,6 +73,7 @@ dl-routines = \ + dl-origin \ + dl-printf \ + dl-reloc \ ++ dl-rseq-symbols \ + dl-runtime \ + dl-scope \ + dl-setup_hash \ +diff --git a/elf/dl-rseq-symbols.S b/elf/dl-rseq-symbols.S +new file mode 100644 +index 0000000000000000..b4bba06a99b0a486 +--- /dev/null ++++ b/elf/dl-rseq-symbols.S +@@ -0,0 +1,64 @@ ++/* Define symbols used by rseq. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++ ++#if __WORDSIZE == 64 ++#define RSEQ_OFFSET_SIZE 8 ++#else ++#define RSEQ_OFFSET_SIZE 4 ++#endif ++ ++/* Some targets define a macro to denote the zero register. */ ++#undef zero ++ ++/* Define 2 symbols: '__rseq_size' is public const and '_rseq_size' (an ++ alias of '__rseq_size') is hidden and writable for internal use by the ++ dynamic linker which will initialize the value both symbols point to ++ before copy relocations take place. */ ++ ++ .globl __rseq_size ++ .type __rseq_size, %object ++ .size __rseq_size, 4 ++ .hidden _rseq_size ++ .globl _rseq_size ++ .type _rseq_size, %object ++ .size _rseq_size, 4 ++ .section .data.rel.ro ++ .balign 4 ++__rseq_size: ++_rseq_size: ++ .zero 4 ++ ++/* Define 2 symbols: '__rseq_offset' is public const and '_rseq_offset' (an ++ alias of '__rseq_offset') is hidden and writable for internal use by the ++ dynamic linker which will initialize the value both symbols point to ++ before copy relocations take place. */ ++ ++ .globl __rseq_offset ++ .type __rseq_offset, %object ++ .size __rseq_offset, RSEQ_OFFSET_SIZE ++ .hidden _rseq_offset ++ .globl _rseq_offset ++ .type _rseq_offset, %object ++ .size _rseq_offset, RSEQ_OFFSET_SIZE ++ .section .data.rel.ro ++ .balign RSEQ_OFFSET_SIZE ++__rseq_offset: ++_rseq_offset: ++ .zero RSEQ_OFFSET_SIZE +diff --git a/sysdeps/nptl/dl-tls_init_tp.c b/sysdeps/nptl/dl-tls_init_tp.c +index 092c274f369dd046..7eb35fb13384f6ac 100644 +--- a/sysdeps/nptl/dl-tls_init_tp.c ++++ b/sysdeps/nptl/dl-tls_init_tp.c +@@ -45,8 +45,10 @@ rtld_mutex_dummy (pthread_mutex_t *lock) + #endif + + const unsigned int __rseq_flags; +-const unsigned int __rseq_size attribute_relro; +-const ptrdiff_t __rseq_offset attribute_relro; ++ ++/* The variables are in .data.relro but are not yet write-protected. */ ++extern unsigned int _rseq_size attribute_hidden; ++extern ptrdiff_t _rseq_offset attribute_hidden; + + void + __tls_pre_init_tp (void) +@@ -105,10 +107,7 @@ __tls_init_tp (void) + do_rseq = TUNABLE_GET (rseq, int, NULL); + if (rseq_register_current_thread (pd, do_rseq)) + { +- /* We need a writable view of the variables. They are in +- .data.relro and are not yet write-protected. */ +- extern unsigned int size __asm__ ("__rseq_size"); +- size = sizeof (pd->rseq_area); ++ _rseq_size = sizeof (pd->rseq_area); + } + + #ifdef RSEQ_SIG +@@ -117,8 +116,7 @@ __tls_init_tp (void) + all targets support __thread_pointer, so set __rseq_offset only + if the rseq registration may have happened because RSEQ_SIG is + defined. */ +- extern ptrdiff_t offset __asm__ ("__rseq_offset"); +- offset = (char *) &pd->rseq_area - (char *) __thread_pointer (); ++ _rseq_offset = (char *) &pd->rseq_area - (char *) __thread_pointer (); + #endif + } + diff --git a/SOURCES/glibc-upstream-2.39-84.patch b/SOURCES/glibc-upstream-2.39-84.patch new file mode 100644 index 0000000..1e5251c --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-84.patch @@ -0,0 +1,43 @@ +commit 143a7a06235cf091b63f58a739c8367ffe54a722 +Author: Adhemerval Zanella +Date: Thu Jul 4 10:09:07 2024 -0300 + + elf: Make dl-rseq-symbols Linux only + + And avoid a Hurd build failures. + + Checked on x86_64-linux-gnu. + + (cherry picked from commit 9fc639f654dc004736836613be703e6bed0c36a8) + +diff --git a/elf/Makefile b/elf/Makefile +index 0049ffa13c8d3e51..a50a988e7362cf3b 100644 +--- a/elf/Makefile ++++ b/elf/Makefile +@@ -73,7 +73,6 @@ dl-routines = \ + dl-origin \ + dl-printf \ + dl-reloc \ +- dl-rseq-symbols \ + dl-runtime \ + dl-scope \ + dl-setup_hash \ +diff --git a/sysdeps/unix/sysv/linux/Makefile b/sysdeps/unix/sysv/linux/Makefile +index 415aa1f14dd20ba6..6ab9b901234dc72e 100644 +--- a/sysdeps/unix/sysv/linux/Makefile ++++ b/sysdeps/unix/sysv/linux/Makefile +@@ -615,6 +615,10 @@ tests += \ + endif + + ifeq ($(subdir),elf) ++dl-routines += \ ++ dl-rseq-symbols \ ++ # dl-routines ++ + sysdep-rtld-routines += \ + dl-brk \ + dl-getcwd \ +diff --git a/elf/dl-rseq-symbols.S b/sysdeps/unix/sysv/linux/dl-rseq-symbols.S +similarity index 100% +rename from elf/dl-rseq-symbols.S +rename to sysdeps/unix/sysv/linux/dl-rseq-symbols.S diff --git a/SOURCES/glibc-upstream-2.39-85.patch b/SOURCES/glibc-upstream-2.39-85.patch new file mode 100644 index 0000000..eb2394a --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-85.patch @@ -0,0 +1,137 @@ +commit e8f521709731ce3ae8d6f1eca30135d5c0606f02 +Author: Florian Weimer +Date: Mon Jul 8 21:14:00 2024 +0200 + + Linux: Make __rseq_size useful for feature detection (bug 31965) + + The __rseq_size value is now the active area of struct rseq + (so 20 initially), not the full struct size including padding + at the end (32 initially). + + Update misc/tst-rseq to print some additional diagnostics. + + Reviewed-by: Michael Jeanson + Reviewed-by: Mathieu Desnoyers + (cherry picked from commit 2e456ccf0c34a056e3ccafac4a0c7effef14d918) + +diff --git a/manual/threads.texi b/manual/threads.texi +index e5544ff3da8733f4..25e99c9606dcad77 100644 +--- a/manual/threads.texi ++++ b/manual/threads.texi +@@ -1007,8 +1007,12 @@ This variable is either zero (if restartable sequence registration + failed or has been disabled) or the size of the restartable sequence + registration. This can be different from the size of @code{struct rseq} + if the kernel has extended the size of the registration. If +-registration is successful, @code{__rseq_size} is at least 32 (the +-initial size of @code{struct rseq}). ++registration is successful, @code{__rseq_size} is at least 20 (the ++initially active size of @code{struct rseq}). ++ ++Previous versions of @theglibc{} set this to 32 even if the kernel only ++supported the initial area of 20 bytes because the value included unused ++padding at the end of the restartable sequence area. + @end deftypevar + + @deftypevar {unsigned int} __rseq_flags +diff --git a/sysdeps/nptl/dl-tls_init_tp.c b/sysdeps/nptl/dl-tls_init_tp.c +index 7eb35fb13384f6ac..7803e19fd16ad803 100644 +--- a/sysdeps/nptl/dl-tls_init_tp.c ++++ b/sysdeps/nptl/dl-tls_init_tp.c +@@ -46,10 +46,6 @@ rtld_mutex_dummy (pthread_mutex_t *lock) + + const unsigned int __rseq_flags; + +-/* The variables are in .data.relro but are not yet write-protected. */ +-extern unsigned int _rseq_size attribute_hidden; +-extern ptrdiff_t _rseq_offset attribute_hidden; +- + void + __tls_pre_init_tp (void) + { +@@ -106,9 +102,7 @@ __tls_init_tp (void) + bool do_rseq = true; + do_rseq = TUNABLE_GET (rseq, int, NULL); + if (rseq_register_current_thread (pd, do_rseq)) +- { +- _rseq_size = sizeof (pd->rseq_area); +- } ++ _rseq_size = RSEQ_AREA_SIZE_INITIAL_USED; + + #ifdef RSEQ_SIG + /* This should be a compile-time constant, but the current +diff --git a/sysdeps/unix/sysv/linux/rseq-internal.h b/sysdeps/unix/sysv/linux/rseq-internal.h +index 48eebc1e168fad1e..7ea935b4adab8c20 100644 +--- a/sysdeps/unix/sysv/linux/rseq-internal.h ++++ b/sysdeps/unix/sysv/linux/rseq-internal.h +@@ -25,15 +25,34 @@ + #include + #include + ++/* 32 is the initially required value for the area size. The ++ actually used rseq size may be less (20 bytes initially). */ ++#define RSEQ_AREA_SIZE_INITIAL 32 ++#define RSEQ_AREA_SIZE_INITIAL_USED 20 ++ ++/* The variables are in .data.relro but are not yet write-protected. */ ++extern unsigned int _rseq_size attribute_hidden; ++extern ptrdiff_t _rseq_offset attribute_hidden; ++ + #ifdef RSEQ_SIG + static inline bool + rseq_register_current_thread (struct pthread *self, bool do_rseq) + { + if (do_rseq) + { ++ unsigned int size; ++#if IS_IN (rtld) ++ /* Use the hidden symbol in ld.so. */ ++ size = _rseq_size; ++#else ++ size = __rseq_size; ++#endif ++ if (size < RSEQ_AREA_SIZE_INITIAL) ++ /* The initial implementation used only 20 bytes out of 32, ++ but still expected size 32. */ ++ size = RSEQ_AREA_SIZE_INITIAL; + int ret = INTERNAL_SYSCALL_CALL (rseq, &self->rseq_area, +- sizeof (self->rseq_area), +- 0, RSEQ_SIG); ++ size, 0, RSEQ_SIG); + if (!INTERNAL_SYSCALL_ERROR_P (ret)) + return true; + } +diff --git a/sysdeps/unix/sysv/linux/tst-rseq.c b/sysdeps/unix/sysv/linux/tst-rseq.c +index 2c90409ba02182e7..08a95331306b2a12 100644 +--- a/sysdeps/unix/sysv/linux/tst-rseq.c ++++ b/sysdeps/unix/sysv/linux/tst-rseq.c +@@ -29,6 +29,7 @@ + # include + # include + # include ++# include + # include + # include + # include "tst-rseq.h" +@@ -42,7 +43,8 @@ do_rseq_main_test (void) + TEST_COMPARE (__rseq_flags, 0); + TEST_VERIFY ((char *) __thread_pointer () + __rseq_offset + == (char *) &pd->rseq_area); +- TEST_COMPARE (__rseq_size, sizeof (pd->rseq_area)); ++ /* The current implementation only supports the initial size. */ ++ TEST_COMPARE (__rseq_size, 20); + } + + static void +@@ -52,6 +54,12 @@ do_rseq_test (void) + { + FAIL_UNSUPPORTED ("kernel does not support rseq, skipping test"); + } ++ printf ("info: __rseq_size: %u\n", __rseq_size); ++ printf ("info: __rseq_offset: %td\n", __rseq_offset); ++ printf ("info: __rseq_flags: %u\n", __rseq_flags); ++ printf ("info: getauxval (AT_RSEQ_FEATURE_SIZE): %ld\n", ++ getauxval (AT_RSEQ_FEATURE_SIZE)); ++ printf ("info: getauxval (AT_RSEQ_ALIGN): %ld\n", getauxval (AT_RSEQ_ALIGN)); + do_rseq_main_test (); + } + #else /* RSEQ_SIG */ diff --git a/SOURCES/glibc-upstream-2.39-86.patch b/SOURCES/glibc-upstream-2.39-86.patch new file mode 100644 index 0000000..b15d1ac --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-86.patch @@ -0,0 +1,58 @@ +commit f6a75fddf4e71545c63dfcad99cc2df9bac38093 +Author: John David Anglin +Date: Fri Jul 19 10:10:17 2024 -0400 + + Fix usage of _STACK_GROWS_DOWN and _STACK_GROWS_UP defines [BZ 31989] + + Signed-off-by: John David Anglin + Reviewed-By: Andreas K. Hüttel + (cherry picked from commit 8cfa4ecff21adf226984f135aa576dd8063bbba3) + +diff --git a/malloc/memusage.c b/malloc/memusage.c +index e8ae80dc74af4585..f80225b95a36707b 100644 +--- a/malloc/memusage.c ++++ b/malloc/memusage.c +@@ -172,7 +172,7 @@ update_data (struct header *result, size_t len, size_t old_len) + start_sp = __thread_stack_pointer (); + + uintptr_t sp = __thread_stack_pointer (); +-#ifdef _STACK_GROWS_UP ++#if _STACK_GROWS_UP + /* This can happen in threads where we didn't catch the thread's + stack early enough. */ + if (__glibc_unlikely (sp < start_sp)) +diff --git a/stdlib/tst-swapcontext2.c b/stdlib/tst-swapcontext2.c +index f679755649809653..a9c1dc827cb897f3 100644 +--- a/stdlib/tst-swapcontext2.c ++++ b/stdlib/tst-swapcontext2.c +@@ -85,7 +85,7 @@ do_test (void) + { + /* ____longjmp_chk has */ + #if 0 +-#ifdef _STACK_GROWS_DOWN ++#if _STACK_GROWS_DOWN + #define called_from(this, saved) ((this) < (saved)) + #else + #define called_from(this, saved) ((this) > (saved)) +@@ -98,7 +98,7 @@ do_test (void) + /* Arrange stacks for uctx_func1 and uctx_func2 so that called_from + is true when setjmp is called from uctx_func1 and longjmp is called + from uctx_func2. */ +-#ifdef _STACK_GROWS_DOWN ++#if _STACK_GROWS_DOWN + # define UCTX_FUNC1_STACK 1 + # define UCTX_FUNC2_STACK 0 + #else +diff --git a/sysdeps/unix/sysv/linux/____longjmp_chk.c b/sysdeps/unix/sysv/linux/____longjmp_chk.c +index 0896dc5755dfa1db..3c66a4638eedbbea 100644 +--- a/sysdeps/unix/sysv/linux/____longjmp_chk.c ++++ b/sysdeps/unix/sysv/linux/____longjmp_chk.c +@@ -23,7 +23,7 @@ + #include + #include + +-#ifdef _STACK_GROWS_DOWN ++#if _STACK_GROWS_DOWN + #define called_from(this, saved) ((this) < (saved)) + #else + #define called_from(this, saved) ((this) > (saved)) diff --git a/SOURCES/glibc-upstream-2.39-87.patch b/SOURCES/glibc-upstream-2.39-87.patch new file mode 100644 index 0000000..311203b --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-87.patch @@ -0,0 +1,81 @@ +commit cb19cef087eaa551568739aa5b59cf97b8d5da1e +Author: Adhemerval Zanella +Date: Tue Jul 16 07:57:45 2024 -0300 + + elf: Fix localplt.awk for DT_RELR-enabled builds (BZ 31978) + + For each input readelf output, localplt.awk parses each 'Relocation + section' entry, checks its offset against the dynamic section entry, and + saves each DT_JMPREL, DT_RELA, and DT_REL offset value it finds. After + all lines are read, the script checks if any segment offset differed + from 0, meaning at least one 'Relocation section' was matched. + + However, if the shared object was built with RELR support and the static + linker could place all the relocation on DT_RELR, there would be no + DT_JMPREL, DT_RELA, and DT_REL entries; only a DT_RELR. + + For the current three ABIs that support (aarch64, x86, and powerpc64), + the powerpc64 ld.so shows the behavior above. Both x86_64 and aarch64 + show extra relocations on '.rela.dyn', which makes the script check to + succeed. + + This patch fixes by handling DT_RELR, where the offset is checked + against the dynamic section entries and if the shared object contains an + entry it means that there are no extra PLT entries (since all + relocations are relative). + + It fixes the elf/check-localplt failure on powerpc. + + Checked with a build/check for aarch64-linux-gnu, x86_64-linux-gnu, + i686-linux-gnu, arm-linux-gnueabihf, s390x-linux-gnu, powerpc-linux-gnu, + powerpc64-linux-gnu, and powerpc64le-linux-gnu. + Reviewed-by: Carlos O'Donell + + (cherry picked from commit 4f047d9edecb1a9b796a9a904dcd42bd3cc3d3b6) + +diff --git a/scripts/localplt.awk b/scripts/localplt.awk +index fe79ca01abcb1a75..621ae7d8e815a4e5 100644 +--- a/scripts/localplt.awk ++++ b/scripts/localplt.awk +@@ -10,7 +10,8 @@ BEGIN { + } + + FILENAME != lastfile { +- if (lastfile && jmprel_offset == 0 && rela_offset == 0 && rel_offset == 0) { ++ if (lastfile && jmprel_offset == 0 && rela_offset == 0 && rel_offset == 0 \ ++ && relr_offset == 0) { + print FILENAME ": *** failed to find expected output (readelf -WSdr)"; + result = 2; + } +@@ -22,6 +23,7 @@ FILENAME != lastfile { + jmprel_offset = 0; + rela_offset = 0; + rel_offset = 0; ++ relr_offset = 0; + pltrelsz = -1; + delete section_offset_by_address; + } +@@ -77,6 +79,8 @@ in_relocs && relocs_offset == rel_offset && NF >= 5 { + } + } + ++# No need to handle DT_RELR (all packed relocations are relative). ++ + in_relocs { next } + + $1 == "Relocation" && $2 == "section" && $5 == "offset" { +@@ -121,4 +125,14 @@ $2 == "(REL)" { + } + next + } ++ ++$2 == "(RELR)" { ++ relr_addr = strtonum($3); ++ if (relr_addr in section_offset_by_address) { ++ relr_offset = section_offset_by_address[relr_addr]; ++ } else { ++ print FILENAME ": *** DT_RELR does not match any section's address"; ++ result = 2; ++ } ++} + END { exit(result) } diff --git a/SOURCES/glibc-upstream-2.39-88.patch b/SOURCES/glibc-upstream-2.39-88.patch new file mode 100644 index 0000000..1afdee1 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-88.patch @@ -0,0 +1,201 @@ +commit fc1dc399841051862a93ed458897e3d8ea9d4ff4 +Author: Florian Weimer +Date: Wed Jul 24 12:06:47 2024 +0200 + + resolv: Allow short error responses to match any query (bug 31890) + + Reviewed-by: DJ Delorie + (cherry picked from commit 691a3b2e9bfaba842e46a5ccb7f5e6ea144c3ade) + +diff --git a/resolv/Makefile b/resolv/Makefile +index 5f44f5896bde9958..d927e337d9e2b496 100644 +--- a/resolv/Makefile ++++ b/resolv/Makefile +@@ -106,6 +106,7 @@ tests += \ + tst-resolv-nondecimal \ + tst-resolv-res_init-multi \ + tst-resolv-search \ ++ tst-resolv-short-response \ + tst-resolv-trailing \ + + # This test calls __res_context_send directly, which is not exported +@@ -299,6 +300,8 @@ $(objpfx)tst-resolv-nondecimal: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-qtypes: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-rotate: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-search: $(objpfx)libresolv.so $(shared-thread-library) ++$(objpfx)tst-resolv-short-response: $(objpfx)libresolv.so \ ++ $(shared-thread-library) + $(objpfx)tst-resolv-trailing: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-threads: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-txnid-collision: $(objpfx)libresolv.a \ +diff --git a/resolv/res_send.c b/resolv/res_send.c +index fb0217204a6c913f..cef606a6b1523224 100644 +--- a/resolv/res_send.c ++++ b/resolv/res_send.c +@@ -1197,19 +1197,30 @@ send_dg(res_state statp, + } + + /* Check for the correct header layout and a matching +- question. */ ++ question. Some recursive resolvers send REFUSED ++ without copying back the question section ++ (producing a response that is only HFIXEDSZ bytes ++ long). Skip query matching in this case. */ ++ bool thisansp_error = (anhp->rcode == SERVFAIL || ++ anhp->rcode == NOTIMP || ++ anhp->rcode == REFUSED); ++ bool skip_query_match = (*thisresplenp == HFIXEDSZ ++ && ntohs (anhp->qdcount) == 0 ++ && thisansp_error); + int matching_query = 0; /* Default to no matching query. */ + if (!recvresp1 + && anhp->id == hp->id +- && __libc_res_queriesmatch (buf, buf + buflen, +- *thisansp, +- *thisansp + *thisanssizp)) ++ && (skip_query_match ++ || __libc_res_queriesmatch (buf, buf + buflen, ++ *thisansp, ++ *thisansp + *thisanssizp))) + matching_query = 1; + if (!recvresp2 + && anhp->id == hp2->id +- && __libc_res_queriesmatch (buf2, buf2 + buflen2, +- *thisansp, +- *thisansp + *thisanssizp)) ++ && (skip_query_match ++ || __libc_res_queriesmatch (buf2, buf2 + buflen2, ++ *thisansp, ++ *thisansp + *thisanssizp))) + matching_query = 2; + if (matching_query == 0) + /* Spurious UDP packet. Drop it and continue +@@ -1219,9 +1230,7 @@ send_dg(res_state statp, + goto wait; + } + +- if (anhp->rcode == SERVFAIL || +- anhp->rcode == NOTIMP || +- anhp->rcode == REFUSED) { ++ if (thisansp_error) { + next_ns: + if (recvresp1 || (buf2 != NULL && recvresp2)) { + *resplen2 = 0; +diff --git a/resolv/tst-resolv-short-response.c b/resolv/tst-resolv-short-response.c +new file mode 100644 +index 0000000000000000..cf1e39876f8fb470 +--- /dev/null ++++ b/resolv/tst-resolv-short-response.c +@@ -0,0 +1,112 @@ ++/* Test for spurious timeouts with short 12-byte responses (bug 31890). ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++ ++/* The rcode in the initial response. */ ++static volatile int rcode; ++ ++static void ++response (const struct resolv_response_context *ctx, ++ struct resolv_response_builder *b, ++ const char *qname, uint16_t qclass, uint16_t qtype) ++{ ++ switch (ctx->server_index) ++ { ++ case 0: ++ /* First server times out. */ ++ struct resolv_response_flags flags = {.rcode = rcode}; ++ resolv_response_init (b, flags); ++ break; ++ case 1: ++ /* Second server sends reply. */ ++ resolv_response_init (b, (struct resolv_response_flags) {}); ++ resolv_response_add_question (b, qname, qclass, qtype); ++ resolv_response_section (b, ns_s_an); ++ resolv_response_open_record (b, qname, qclass, qtype, 0); ++ switch (qtype) ++ { ++ case T_A: ++ { ++ char ipv4[4] = {192, 0, 2, 17}; ++ resolv_response_add_data (b, &ipv4, sizeof (ipv4)); ++ } ++ break; ++ case T_AAAA: ++ { ++ char ipv6[16] ++ = {0x20, 0x01, 0xd, 0xb8, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1}; ++ resolv_response_add_data (b, &ipv6, sizeof (ipv6)); ++ } ++ break; ++ default: ++ FAIL_EXIT1 ("unexpected TYPE%d query", qtype); ++ } ++ resolv_response_close_record (b); ++ break; ++ default: ++ FAIL_EXIT1 ("unexpected query to server %d", ctx->server_index); ++ } ++} ++ ++static void ++check_one (void) ++{ ++ ++ /* The buggy 1-second query timeout results in 30 seconds of delay, ++ which triggers a test timeout failure. */ ++ for (int i = 0; i < 10; ++i) ++ { ++ check_hostent ("www.example", gethostbyname ("www.example"), ++ "name: www.example\n" ++ "address: 192.0.2.17\n"); ++ check_hostent ("www.example", gethostbyname2 ("www.example", AF_INET6), ++ "name: www.example\n" ++ "address: 2001:db8::1\n"); ++ } ++} ++ ++static int ++do_test (void) ++{ ++ struct resolv_test *aux = resolv_test_start ++ ((struct resolv_redirect_config) ++ { ++ .response_callback = response, ++ }); ++ ++ _res.options |= RES_SNGLKUP; ++ ++ rcode = 2; /* SERVFAIL. */ ++ check_one (); ++ ++ rcode = 4; /* NOTIMP. */ ++ check_one (); ++ ++ rcode = 5; /* REFUSED. */ ++ check_one (); ++ ++ resolv_test_end (aux); ++ ++ return 0; ++} ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-89.patch b/SOURCES/glibc-upstream-2.39-89.patch new file mode 100644 index 0000000..a6c7115 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-89.patch @@ -0,0 +1,209 @@ +commit 8c9fbc6651b992e759f6f0b4af21c1a8769bbb2c +Author: Florian Weimer +Date: Wed Jul 24 12:06:47 2024 +0200 + + resolv: Do not wait for non-existing second DNS response after error (bug 30081) + + In single-request mode, there is no second response after an error + because the second query has not been sent yet. Waiting for it + introduces an unnecessary timeout. + + Reviewed-by: DJ Delorie + (cherry picked from commit af625987d619388a100b153520d3ee308bda9889) + +diff --git a/resolv/Makefile b/resolv/Makefile +index d927e337d9e2b496..abff7fc0074e893b 100644 +--- a/resolv/Makefile ++++ b/resolv/Makefile +@@ -106,6 +106,7 @@ tests += \ + tst-resolv-nondecimal \ + tst-resolv-res_init-multi \ + tst-resolv-search \ ++ tst-resolv-semi-failure \ + tst-resolv-short-response \ + tst-resolv-trailing \ + +@@ -300,6 +301,8 @@ $(objpfx)tst-resolv-nondecimal: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-qtypes: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-rotate: $(objpfx)libresolv.so $(shared-thread-library) + $(objpfx)tst-resolv-search: $(objpfx)libresolv.so $(shared-thread-library) ++$(objpfx)tst-resolv-semi-failure: $(objpfx)libresolv.so \ ++ $(shared-thread-library) + $(objpfx)tst-resolv-short-response: $(objpfx)libresolv.so \ + $(shared-thread-library) + $(objpfx)tst-resolv-trailing: $(objpfx)libresolv.so $(shared-thread-library) +diff --git a/resolv/res_send.c b/resolv/res_send.c +index cef606a6b1523224..3a4a20684fcb9a63 100644 +--- a/resolv/res_send.c ++++ b/resolv/res_send.c +@@ -1236,7 +1236,7 @@ send_dg(res_state statp, + *resplen2 = 0; + return resplen; + } +- if (buf2 != NULL) ++ if (buf2 != NULL && !single_request) + { + /* No data from the first reply. */ + resplen = 0; +diff --git a/resolv/tst-resolv-semi-failure.c b/resolv/tst-resolv-semi-failure.c +new file mode 100644 +index 0000000000000000..aa9798b5a7dfaa88 +--- /dev/null ++++ b/resolv/tst-resolv-semi-failure.c +@@ -0,0 +1,133 @@ ++/* Test parallel failure/success responses (bug 30081). ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public ++ License as published by the Free Software Foundation; either ++ version 2.1 of the License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; if not, see ++ . */ ++ ++#include ++#include ++#include ++#include ++ ++/* The rcode in the initial response. */ ++static volatile int rcode; ++ ++/* Whether to fail the initial A query (!fail_aaaa) or the initial ++ AAAA query (fail_aaaa). */ ++static volatile bool fail_aaaa; ++ ++static void ++response (const struct resolv_response_context *ctx, ++ struct resolv_response_builder *b, ++ const char *qname, uint16_t qclass, uint16_t qtype) ++{ ++ /* Handle the failing query. */ ++ if ((fail_aaaa && qtype == T_AAAA) && ctx->server_index == 0) ++ { ++ struct resolv_response_flags flags = {.rcode = rcode}; ++ resolv_response_init (b, flags); ++ return; ++ } ++ ++ /* Otherwise produce a response. */ ++ resolv_response_init (b, (struct resolv_response_flags) {}); ++ resolv_response_add_question (b, qname, qclass, qtype); ++ resolv_response_section (b, ns_s_an); ++ resolv_response_open_record (b, qname, qclass, qtype, 0); ++ switch (qtype) ++ { ++ case T_A: ++ { ++ char ipv4[4] = {192, 0, 2, 17}; ++ resolv_response_add_data (b, &ipv4, sizeof (ipv4)); ++ } ++ break; ++ case T_AAAA: ++ { ++ char ipv6[16] ++ = {0x20, 0x01, 0xd, 0xb8, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1}; ++ resolv_response_add_data (b, &ipv6, sizeof (ipv6)); ++ } ++ break; ++ default: ++ FAIL_EXIT1 ("unexpected TYPE%d query", qtype); ++ } ++ resolv_response_close_record (b); ++} ++ ++static void ++check_one (void) ++{ ++ ++ /* The buggy 1-second query timeout results in 30 seconds of delay, ++ which triggers are test timeout failure. */ ++ for (int i = 0; i < 30; ++i) ++ { ++ static const struct addrinfo hints = ++ { ++ .ai_family = AF_UNSPEC, ++ .ai_socktype = SOCK_STREAM, ++ }; ++ struct addrinfo *ai; ++ int ret = getaddrinfo ("www.example", "80", &hints, &ai); ++ const char *expected; ++ if (ret == 0 && ai->ai_next != NULL) ++ expected = ("address: STREAM/TCP 192.0.2.17 80\n" ++ "address: STREAM/TCP 2001:db8::1 80\n"); ++ else ++ /* Only one response because the AAAA lookup failure is ++ treated as an ignoreable error. */ ++ expected = "address: STREAM/TCP 192.0.2.17 80\n"; ++ check_addrinfo ("www.example", ai, ret, expected); ++ if (ret == 0) ++ freeaddrinfo (ai); ++ } ++} ++ ++static int ++do_test (void) ++{ ++ for (int do_single_lookup = 0; do_single_lookup < 2; ++do_single_lookup) ++ { ++ struct resolv_test *aux = resolv_test_start ++ ((struct resolv_redirect_config) ++ { ++ .response_callback = response, ++ }); ++ ++ if (do_single_lookup) ++ _res.options |= RES_SNGLKUP; ++ ++ for (int do_fail_aaaa = 0; do_fail_aaaa < 2; ++do_fail_aaaa) ++ { ++ fail_aaaa = do_fail_aaaa; ++ ++ rcode = 2; /* SERVFAIL. */ ++ check_one (); ++ ++ rcode = 4; /* NOTIMP. */ ++ check_one (); ++ ++ rcode = 5; /* REFUSED. */ ++ check_one (); ++ } ++ ++ resolv_test_end (aux); ++ } ++ ++ return 0; ++} ++ ++#include +diff --git a/resolv/tst-resolv-short-response.c b/resolv/tst-resolv-short-response.c +index cf1e39876f8fb470..be354ae1c7f2a81a 100644 +--- a/resolv/tst-resolv-short-response.c ++++ b/resolv/tst-resolv-short-response.c +@@ -81,6 +81,18 @@ check_one (void) + check_hostent ("www.example", gethostbyname2 ("www.example", AF_INET6), + "name: www.example\n" + "address: 2001:db8::1\n"); ++ static const struct addrinfo hints = ++ { ++ .ai_family = AF_UNSPEC, ++ .ai_socktype = SOCK_STREAM, ++ }; ++ struct addrinfo *ai; ++ int ret = getaddrinfo ("www.example", "80", &hints, &ai); ++ check_addrinfo ("www.example", ai, ret, ++ "address: STREAM/TCP 192.0.2.17 80\n" ++ "address: STREAM/TCP 2001:db8::1 80\n"); ++ if (ret == 0) ++ freeaddrinfo (ai); + } + } + diff --git a/SOURCES/glibc-upstream-2.39-9.patch b/SOURCES/glibc-upstream-2.39-9.patch new file mode 100644 index 0000000..4289fa2 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-9.patch @@ -0,0 +1,172 @@ +commit aad45c8ac30aa1072e54903ce6aead22702f244a +Author: Amrita H S +Date: Tue Mar 19 19:08:47 2024 -0500 + + powerpc: Placeholder and infrastructure/build support to add Power11 related changes. + + The following three changes have been added to provide initial Power11 support. + 1. Add the directories to hold Power11 files. + 2. Add support to select Power11 libraries based on AT_PLATFORM. + 3. Let submachine=power11 be set automatically. + + Reviewed-by: Florian Weimer + Reviewed-by: Peter Bergner + (cherry picked from commit 1ea051145612f199d8716ecdf78b084b00b5a727) + +diff --git a/sysdeps/powerpc/dl-procinfo.h b/sysdeps/powerpc/dl-procinfo.h +index f8cb343877386402..b36697ba440654be 100644 +--- a/sysdeps/powerpc/dl-procinfo.h ++++ b/sysdeps/powerpc/dl-procinfo.h +@@ -38,7 +38,7 @@ + #define HWCAP_IMPORTANT (PPC_FEATURE_HAS_ALTIVEC \ + + PPC_FEATURE_HAS_DFP) + +-#define _DL_PLATFORMS_COUNT 16 ++#define _DL_PLATFORMS_COUNT 17 + + #define _DL_FIRST_PLATFORM 32 + /* Mask to filter out platforms. */ +@@ -62,6 +62,7 @@ + #define PPC_PLATFORM_POWER8 13 + #define PPC_PLATFORM_POWER9 14 + #define PPC_PLATFORM_POWER10 15 ++#define PPC_PLATFORM_POWER11 16 + + static inline const char * + __attribute__ ((unused)) +@@ -89,6 +90,11 @@ _dl_string_platform (const char *str) + ret = _DL_FIRST_PLATFORM + PPC_PLATFORM_POWER10; + str++; + } ++ else if (str[1] == '1') ++ { ++ ret = _DL_FIRST_PLATFORM + PPC_PLATFORM_POWER11; ++ str++; ++ } + else + return -1; + break; +diff --git a/sysdeps/powerpc/powerpc32/power11/Implies b/sysdeps/powerpc/powerpc32/power11/Implies +new file mode 100644 +index 0000000000000000..051cbe0f7911c93c +--- /dev/null ++++ b/sysdeps/powerpc/powerpc32/power11/Implies +@@ -0,0 +1,2 @@ ++powerpc/powerpc32/power10/fpu ++powerpc/powerpc32/power10 +diff --git a/sysdeps/powerpc/powerpc32/power11/fpu/multiarch/Implies b/sysdeps/powerpc/powerpc32/power11/fpu/multiarch/Implies +new file mode 100644 +index 0000000000000000..58edb2861d17f504 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc32/power11/fpu/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc32/power10/fpu/multiarch +diff --git a/sysdeps/powerpc/powerpc32/power11/multiarch/Implies b/sysdeps/powerpc/powerpc32/power11/multiarch/Implies +new file mode 100644 +index 0000000000000000..c70f0428badbaf14 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc32/power11/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc32/power10/multiarch +diff --git a/sysdeps/powerpc/powerpc64/be/power11/Implies b/sysdeps/powerpc/powerpc64/be/power11/Implies +new file mode 100644 +index 0000000000000000..de481d1c13db695e +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/be/power11/Implies +@@ -0,0 +1,2 @@ ++powerpc/powerpc64/be/power10/fpu ++powerpc/powerpc64/be/power10 +diff --git a/sysdeps/powerpc/powerpc64/be/power11/fpu/Implies b/sysdeps/powerpc/powerpc64/be/power11/fpu/Implies +new file mode 100644 +index 0000000000000000..dff0e13064ce8238 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/be/power11/fpu/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/be/power10/fpu +diff --git a/sysdeps/powerpc/powerpc64/be/power11/fpu/multiarch/Implies b/sysdeps/powerpc/powerpc64/be/power11/fpu/multiarch/Implies +new file mode 100644 +index 0000000000000000..c3f259e0097386a5 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/be/power11/fpu/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/be/power10/fpu/multiarch +diff --git a/sysdeps/powerpc/powerpc64/be/power11/multiarch/Implies b/sysdeps/powerpc/powerpc64/be/power11/multiarch/Implies +new file mode 100644 +index 0000000000000000..9491a394c9519d01 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/be/power11/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/be/power10/multiarch +diff --git a/sysdeps/powerpc/powerpc64/le/power11/Implies b/sysdeps/powerpc/powerpc64/le/power11/Implies +new file mode 100644 +index 0000000000000000..e18182dcc1f4c25f +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/le/power11/Implies +@@ -0,0 +1,2 @@ ++powerpc/powerpc64/le/power10/fpu ++powerpc/powerpc64/le/power10 +diff --git a/sysdeps/powerpc/powerpc64/le/power11/fpu/Implies b/sysdeps/powerpc/powerpc64/le/power11/fpu/Implies +new file mode 100644 +index 0000000000000000..e41bd55684dbb5b8 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/le/power11/fpu/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/le/power10/fpu +diff --git a/sysdeps/powerpc/powerpc64/le/power11/fpu/multiarch/Implies b/sysdeps/powerpc/powerpc64/le/power11/fpu/multiarch/Implies +new file mode 100644 +index 0000000000000000..c838d5093140eae3 +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/le/power11/fpu/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/le/power10/fpu/multiarch +diff --git a/sysdeps/powerpc/powerpc64/le/power11/multiarch/Implies b/sysdeps/powerpc/powerpc64/le/power11/multiarch/Implies +new file mode 100644 +index 0000000000000000..687248c3c267cd8c +--- /dev/null ++++ b/sysdeps/powerpc/powerpc64/le/power11/multiarch/Implies +@@ -0,0 +1 @@ ++powerpc/powerpc64/le/power10/multiarch +diff --git a/sysdeps/powerpc/powerpc64/le/tst-glibc-hwcaps.c b/sysdeps/powerpc/powerpc64/le/tst-glibc-hwcaps.c +index 77465d9133410267..65d3e69303a1c963 100644 +--- a/sysdeps/powerpc/powerpc64/le/tst-glibc-hwcaps.c ++++ b/sysdeps/powerpc/powerpc64/le/tst-glibc-hwcaps.c +@@ -36,9 +36,11 @@ compute_level (void) + return 9; + if (strcmp (platform, "power10") == 0) + return 10; ++ if (strcmp (platform, "power11") == 0) ++ return 11; + printf ("warning: unrecognized AT_PLATFORM value: %s\n", platform); +- /* Assume that the new platform supports POWER10. */ +- return 10; ++ /* Assume that the new platform supports POWER11. */ ++ return 11; + } + + static int +diff --git a/sysdeps/powerpc/preconfigure b/sysdeps/powerpc/preconfigure +index 4de94089a3f68532..9e5a07ab6d6767cd 100644 +--- a/sysdeps/powerpc/preconfigure ++++ b/sysdeps/powerpc/preconfigure +@@ -58,7 +58,7 @@ fi + + ;; + +- a2|970|power[4-9]|power5x|power6+|power10) ++ a2|970|power[4-9]|power5x|power6+|power10|power11) + submachine=${archcpu} + if test ${libc_cv_cc_submachine+y} + then : +diff --git a/sysdeps/powerpc/preconfigure.ac b/sysdeps/powerpc/preconfigure.ac +index 6c63bd8257b7e40a..14b6dafd4a895c3b 100644 +--- a/sysdeps/powerpc/preconfigure.ac ++++ b/sysdeps/powerpc/preconfigure.ac +@@ -46,7 +46,7 @@ case "${machine}:${submachine}" in + AC_CACHE_VAL(libc_cv_cc_submachine,libc_cv_cc_submachine="") + ;; + +- a2|970|power[[4-9]]|power5x|power6+|power10) ++ a2|970|power[[4-9]]|power5x|power6+|power10|power11) + submachine=${archcpu} + AC_CACHE_VAL(libc_cv_cc_submachine,libc_cv_cc_submachine="") + ;; diff --git a/SOURCES/glibc-upstream-2.39-90.patch b/SOURCES/glibc-upstream-2.39-90.patch new file mode 100644 index 0000000..2c61b36 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-90.patch @@ -0,0 +1,61 @@ +commit 51db012c9408d0ae08ea5f6dd8e663fb3a5a5dfd +Author: Florian Weimer +Date: Thu Jun 13 18:56:30 2024 +0200 + + resolv: Track single-request fallback via _res._flags (bug 31476) + + This avoids changing _res.options, which inteferes with change + detection as part of automatic reloading of /etc/resolv.conf. + + Reviewed-by: DJ Delorie + (cherry picked from commit 868ab8923a2ec977faafec97ecafac0c3159c1b2) + +diff --git a/resolv/res_send.c b/resolv/res_send.c +index 3a4a20684fcb9a63..9c77613f374e5469 100644 +--- a/resolv/res_send.c ++++ b/resolv/res_send.c +@@ -947,9 +947,11 @@ send_dg(res_state statp, + seconds /= statp->nscount; + if (seconds <= 0) + seconds = 1; +- bool single_request_reopen = (statp->options & RES_SNGLKUPREOP) != 0; +- bool single_request = (((statp->options & RES_SNGLKUP) != 0) +- | single_request_reopen); ++ bool single_request_reopen = ((statp->options & RES_SNGLKUPREOP) ++ || (statp->_flags & RES_F_SNGLKUPREOP)); ++ bool single_request = ((statp->options & RES_SNGLKUP) ++ || (statp->_flags & RES_F_SNGLKUP) ++ || single_request_reopen); + int save_gotsomewhere = *gotsomewhere; + + int retval; +@@ -1006,14 +1008,14 @@ send_dg(res_state statp, + have received the first answer. */ + if (!single_request) + { +- statp->options |= RES_SNGLKUP; ++ statp->_flags |= RES_F_SNGLKUP; + single_request = true; + *gotsomewhere = save_gotsomewhere; + goto retry; + } + else if (!single_request_reopen) + { +- statp->options |= RES_SNGLKUPREOP; ++ statp->_flags |= RES_F_SNGLKUPREOP; + single_request_reopen = true; + *gotsomewhere = save_gotsomewhere; + __res_iclose (statp, false); +diff --git a/resolv/resolv-internal.h b/resolv/resolv-internal.h +index 24b164f6b5df4c99..944af3ee7613b3c4 100644 +--- a/resolv/resolv-internal.h ++++ b/resolv/resolv-internal.h +@@ -26,6 +26,8 @@ + #define RES_F_VC 0x00000001 /* Socket is TCP. */ + #define RES_F_CONN 0x00000002 /* Socket is connected. */ + #define RES_F_EDNS0ERR 0x00000004 /* EDNS0 caused errors. */ ++#define RES_F_SNGLKUP 0x00200000 /* Private version of RES_SNGLKUP. */ ++#define RES_F_SNGLKUPREOP 0x00400000 /* Private version of RES_SNGLKUPREOP. */ + + /* The structure HEADER is normally aligned on a word boundary. In + some code, we need to access this structure when it may be aligned diff --git a/SOURCES/glibc-upstream-2.39-91.patch b/SOURCES/glibc-upstream-2.39-91.patch new file mode 100644 index 0000000..bc0c1a3 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-91.patch @@ -0,0 +1,88 @@ +commit f11b7178a12be4f26fef157cdee7bf6a944693d9 +Author: Miguel Martín +Date: Tue Jul 16 17:14:56 2024 +0200 + + malloc: avoid global locks in tst-aligned_alloc-lib.c + + Make sure the DSO used by aligned_alloc/calloc/malloc tests does not get + a global lock on multithreaded tests. + Reviewed-by: Arjun Shankar + + (cherry picked from commit 9a27b566b2048f599048f2f4afe1cce06c4ef43d) + +diff --git a/malloc/tst-aligned_alloc-lib.c b/malloc/tst-aligned_alloc-lib.c +index 0205df5acf6297a5..9ef1f839c101a6ae 100644 +--- a/malloc/tst-aligned_alloc-lib.c ++++ b/malloc/tst-aligned_alloc-lib.c +@@ -17,37 +17,38 @@ + License along with the GNU C Library; see the file COPYING.LIB. If + not, see . */ + +-#include + #include + #include ++#include + + extern void *__libc_malloc (size_t size); + extern void *__libc_calloc (size_t n, size_t size); + ++__thread unsigned int seed = 0; ++ + int aligned_alloc_count = 0; + int libc_malloc_count = 0; + int libc_calloc_count = 0; + +-/* Get a random alignment value. Biased towards the smaller values. Must be +- a power of 2. */ +-static size_t get_random_alignment (void) +-{ +- size_t aligns[] = { +- 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384 +- }; +- +- return aligns[random () % array_length (aligns)]; +-} +- +-static void *get_random_alloc (size_t size) ++static void * ++get_random_alloc (size_t size) + { + void *retval; + size_t align; ++ struct timespec tp; ++ ++ if (seed == 0) ++ { ++ clock_gettime (CLOCK_REALTIME, &tp); ++ seed = tp.tv_nsec; ++ } + +- switch (random() % 3) +- { ++ switch (rand_r (&seed) % 3) ++ { + case 1: +- align = get_random_alignment (); ++ /* Get a random alignment value. Biased towards the smaller ++ * values up to 16384. Must be a power of 2. */ ++ align = 1 << rand_r (&seed) % 15; + retval = aligned_alloc (align, size); + aligned_alloc_count++; + break; +@@ -59,13 +60,13 @@ static void *get_random_alloc (size_t size) + retval = __libc_malloc (size); + libc_malloc_count++; + break; +- } ++ } + + return retval; + } + +- +-void * __random_malloc (size_t size) ++void * ++__random_malloc (size_t size) + { + return get_random_alloc (size); + } diff --git a/SOURCES/glibc-upstream-2.39-92.patch b/SOURCES/glibc-upstream-2.39-92.patch new file mode 100644 index 0000000..1847f4f --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-92.patch @@ -0,0 +1,247 @@ +commit 946006d37cfc66c162877fc678405525a50f20df +Author: Miguel Martín +Date: Tue Jul 16 17:14:57 2024 +0200 + + malloc: add multi-threaded tests for aligned_alloc/calloc/malloc + + Improve aligned_alloc/calloc/malloc test coverage by adding + multi-threaded tests with random memory allocations and with/without + cross-thread memory deallocations. + + Perform a number of memory allocation calls with random sizes limited + to 0xffff. + + Use the existing DSO ('malloc/tst-aligned_alloc-lib.c') to randomize + allocator selection. + + The multi-threaded allocation/deallocation is staged as described below: + + - Stage 1: Half of the threads will be allocating memory and the + other half will be waiting for them to finish the allocation. + - Stage 2: Half of the threads will be allocating memory and the + other half will be deallocating memory. + - Stage 3: Half of the threads will be deallocating memory and the + second half waiting on them to finish. + + Add 'malloc/tst-aligned-alloc-random-thread.c' where each thread will + deallocate only the memory that was previously allocated by itself. + + Add 'malloc/tst-aligned-alloc-random-thread-cross.c' where each thread + will deallocate memory that was previously allocated by another thread. + + The intention is to be able to utilize existing malloc testing to ensure + that similar allocation APIs are also exposed to the same rigors. + Reviewed-by: Arjun Shankar + + (cherry picked from commit b0fbcb7d0051a68baf26b2aed51a8a31c34d68e5) + +diff --git a/malloc/Makefile b/malloc/Makefile +index 02aff1bd1dc664c3..98d507a6eb61f4fe 100644 +--- a/malloc/Makefile ++++ b/malloc/Makefile +@@ -28,6 +28,8 @@ tests := \ + mallocbug \ + tst-aligned-alloc \ + tst-aligned-alloc-random \ ++ tst-aligned-alloc-random-thread \ ++ tst-aligned-alloc-random-thread-cross \ + tst-alloc_buffer \ + tst-calloc \ + tst-free-errno \ +@@ -151,6 +153,8 @@ ifeq ($(have-GLIBC_2.23)$(build-shared),yesyes) + # the tests expect specific internal behavior that is changed due to linking to + # libmcheck.a. + tests-exclude-mcheck = \ ++ tst-aligned-alloc-random-thread \ ++ tst-aligned-alloc-random-thread-cross \ + tst-compathooks-off \ + tst-compathooks-on \ + tst-malloc-backtrace \ +@@ -415,7 +419,11 @@ $(objpfx)tst-mallocstate: $(objpfx)libc_malloc_debug.so + $(objpfx)tst-mallocstate-malloc-check: $(objpfx)libc_malloc_debug.so + + $(objpfx)tst-aligned-alloc-random.out: $(objpfx)tst-aligned_alloc-lib.so ++$(objpfx)tst-aligned-alloc-random-thread.out: $(objpfx)tst-aligned_alloc-lib.so ++$(objpfx)tst-aligned-alloc-random-thread-cross.out: $(objpfx)tst-aligned_alloc-lib.so + $(objpfx)tst-malloc-random.out: $(objpfx)tst-aligned_alloc-lib.so + + tst-aligned-alloc-random-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so ++tst-aligned-alloc-random-thread-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so ++tst-aligned-alloc-random-thread-cross-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so + tst-malloc-random-ENV = LD_PRELOAD=$(objpfx)tst-aligned_alloc-lib.so +diff --git a/malloc/tst-aligned-alloc-random-thread-cross.c b/malloc/tst-aligned-alloc-random-thread-cross.c +new file mode 100644 +index 0000000000000000..360ecc56ee7c8c06 +--- /dev/null ++++ b/malloc/tst-aligned-alloc-random-thread-cross.c +@@ -0,0 +1,19 @@ ++/* multi-threaded memory allocation and cross-thread deallocation test. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public License as ++ published by the Free Software Foundation; either version 2.1 of the ++ License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; see the file COPYING.LIB. If ++ not, see . */ ++#define CROSS_THREAD_DEALLOC ++#include "tst-aligned-alloc-random-thread.c" +diff --git a/malloc/tst-aligned-alloc-random-thread.c b/malloc/tst-aligned-alloc-random-thread.c +new file mode 100644 +index 0000000000000000..e95f79250abec85e +--- /dev/null ++++ b/malloc/tst-aligned-alloc-random-thread.c +@@ -0,0 +1,145 @@ ++/* multi-threaded memory allocation/deallocation test. ++ Copyright (C) 2024 Free Software Foundation, Inc. ++ This file is part of the GNU C Library. ++ ++ The GNU C Library is free software; you can redistribute it and/or ++ modify it under the terms of the GNU Lesser General Public License as ++ published by the Free Software Foundation; either version 2.1 of the ++ License, or (at your option) any later version. ++ ++ The GNU C Library is distributed in the hope that it will be useful, ++ but WITHOUT ANY WARRANTY; without even the implied warranty of ++ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ Lesser General Public License for more details. ++ ++ You should have received a copy of the GNU Lesser General Public ++ License along with the GNU C Library; see the file COPYING.LIB. If ++ not, see . */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++#ifndef ITERATIONS ++# define ITERATIONS 16 ++#endif ++ ++#ifndef NUM_THREADS ++# define NUM_THREADS 8 ++#endif ++ ++#ifndef NUM_ALLOCATIONS ++# define NUM_ALLOCATIONS 2048 ++#endif ++ ++static pthread_barrier_t barrier; ++ ++__thread unsigned int seed; ++ ++typedef struct ++{ ++ int id; ++ pthread_t thread; ++} thread; ++ ++thread threads[NUM_THREADS]; ++ ++void *allocations[NUM_THREADS][NUM_ALLOCATIONS]; ++ ++void ++run_thread_dealloc (int id) ++{ ++ for (int i = 0; i < NUM_ALLOCATIONS; i++) ++ { ++ free (allocations[id][i]); ++ allocations[id][i] = NULL; ++ } ++} ++ ++void ++run_thread_alloc (int id) ++{ ++ size_t msb, size; ++ for (int i = 0; i < NUM_ALLOCATIONS; i++) ++ { ++ msb = 1 << rand_r (&seed) % 16; ++ size = msb + rand_r (&seed) % msb; ++ allocations[id][i] = malloc (size); ++ TEST_VERIFY_EXIT (allocations[id][i] != NULL); ++ } ++} ++ ++void * ++run_allocations (void *arg) ++{ ++ int id = *((int *) arg); ++ seed = time (NULL) + id; ++ ++ /* Stage 1: First half o the threads allocating memory and the second ++ * half waiting for them to finish ++ */ ++ if (id < NUM_THREADS / 2) ++ run_thread_alloc (id); ++ ++ xpthread_barrier_wait (&barrier); ++ ++ /* Stage 2: Half of the threads allocationg memory and the other ++ * half deallocating: ++ * - In the non cross-thread dealloc scenario the first half will be ++ * deallocating the memory allocated by themselves in stage 1 and the ++ * second half will be allocating memory. ++ * - In the cross-thread dealloc scenario the first half will continue ++ * to allocate memory and the second half will deallocate the memory ++ * allocated by the first half in stage 1. ++ */ ++ if (id < NUM_THREADS / 2) ++#ifndef CROSS_THREAD_DEALLOC ++ run_thread_dealloc (id); ++#else ++ run_thread_alloc (id + NUM_THREADS / 2); ++#endif ++ else ++#ifndef CROSS_THREAD_DEALLOC ++ run_thread_alloc (id); ++#else ++ run_thread_dealloc (id - NUM_THREADS / 2); ++#endif ++ ++ xpthread_barrier_wait (&barrier); ++ ++ // Stage 3: Second half of the threads deallocating and the first half ++ // waiting for them to finish. ++ if (id >= NUM_THREADS / 2) ++ run_thread_dealloc (id); ++ ++ return NULL; ++} ++ ++static int ++do_test (void) ++{ ++ xpthread_barrier_init (&barrier, NULL, NUM_THREADS); ++ ++ for (int i = 0; i < ITERATIONS; i++) ++ { ++ for (int t = 0; t < NUM_THREADS; t++) ++ { ++ threads[t].id = t; ++ threads[t].thread ++ = xpthread_create (NULL, run_allocations, &threads[t].id); ++ } ++ ++ for (int t = 0; t < NUM_THREADS; t++) ++ xpthread_join (threads[t].thread); ++ } ++ ++ return 0; ++} ++ ++#include diff --git a/SOURCES/glibc-upstream-2.39-93.patch b/SOURCES/glibc-upstream-2.39-93.patch new file mode 100644 index 0000000..b78c675 --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-93.patch @@ -0,0 +1,52 @@ +commit b71a51189d3e1a2f515d93c0444916293bd5221f +Author: Arjun Shankar +Date: Mon Jul 29 14:30:59 2024 +0200 + + manual/stdio: Clarify putc and putwc + + The manual entry for `putc' described what "most systems" do instead of + describing the glibc implementation and its guarantees. This commit + fixes that by warning that putc may be implemented as a macro that + double-evaluates `stream', and removing the performance claim. + + Even though the current `putc' implementation does not double-evaluate + `stream', offering this obscure guarantee as an extension to what + POSIX allows does not seem very useful. + + The entry for `putwc' is also edited to bring it in line with `putc'. + Reviewed-by: Florian Weimer + + (cherry picked from commit 10de4a47ef3f481592e3c62eb07bcda23e9fde4d) + +diff --git a/manual/stdio.texi b/manual/stdio.texi +index 002fce7a10e97107..c11d37b363385531 100644 +--- a/manual/stdio.texi ++++ b/manual/stdio.texi +@@ -903,21 +903,21 @@ This function is a GNU extension. + @deftypefun int putc (int @var{c}, FILE *@var{stream}) + @standards{ISO, stdio.h} + @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{} @aculock{}}} +-This is just like @code{fputc}, except that most systems implement it as ++This is just like @code{fputc}, except that it may be implemented as + a macro, making it faster. One consequence is that it may evaluate the + @var{stream} argument more than once, which is an exception to the +-general rule for macros. @code{putc} is usually the best function to +-use for writing a single character. ++general rule for macros. Therefore, @var{stream} should never be an ++expression with side-effects. + @end deftypefun + + @deftypefun wint_t putwc (wchar_t @var{wc}, FILE *@var{stream}) + @standards{ISO, wchar.h} + @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{} @aculock{}}} +-This is just like @code{fputwc}, except that it can be implement as ++This is just like @code{fputwc}, except that it may be implemented as + a macro, making it faster. One consequence is that it may evaluate the + @var{stream} argument more than once, which is an exception to the +-general rule for macros. @code{putwc} is usually the best function to +-use for writing a single wide character. ++general rule for macros. Therefore, @var{stream} should never be an ++expression with side-effects. + @end deftypefun + + @deftypefun int putc_unlocked (int @var{c}, FILE *@var{stream}) diff --git a/SOURCES/glibc-upstream-2.39-94.patch b/SOURCES/glibc-upstream-2.39-94.patch new file mode 100644 index 0000000..22cf29d --- /dev/null +++ b/SOURCES/glibc-upstream-2.39-94.patch @@ -0,0 +1,31 @@ +commit 4bdcc1963bc2b5ba5f8e319e402d9eb2cb6096c1 +Author: Lukas Bulwahn +Date: Mon Jul 29 11:08:17 2024 +0200 + + manual: make setrlimit() description less ambiguous + + The existing description for setrlimit() has some ambiguity. It could be + understood to have the semantics of getrlimit(), i.e., the limits from the + process are stored in the provided rlp pointer. + + Make the description more explicit that rlp are the input values, and that + the limits of the process is changed with this function. + + Reviewed-by: Florian Weimer + (cherry picked from commit aedbf08891069fc029ed021e4dba933eb877b394) + +diff --git a/manual/resource.texi b/manual/resource.texi +index 37462abc9e467690..743cc9439665b9d5 100644 +--- a/manual/resource.texi ++++ b/manual/resource.texi +@@ -192,8 +192,8 @@ If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a + @standards{BSD, sys/resource.h} + @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} + @c Direct syscall on most systems; lock-taking critical section on HURD. +-Store the current and maximum limits for the resource @var{resource} +-in @code{*@var{rlp}}. ++Change the current and maximum limits of the process for the resource ++@var{resource} to the values provided in @code{*@var{rlp}}. + + The return value is @code{0} on success and @code{-1} on failure. The + following @code{errno} error condition is possible: diff --git a/SOURCES/glibc.attr b/SOURCES/glibc.attr new file mode 100644 index 0000000..fddfd91 --- /dev/null +++ b/SOURCES/glibc.attr @@ -0,0 +1,3 @@ +%__glibc_requires %{_rpmconfigdir}/glibc.req +%__glibc_magic ELF +%__glibc_flags exeonly diff --git a/SOURCES/glibc.req.in b/SOURCES/glibc.req.in new file mode 100644 index 0000000..b82c5a1 --- /dev/null +++ b/SOURCES/glibc.req.in @@ -0,0 +1,42 @@ +#!/bin/bash +# Auto-generating dependencies for glibc development snapshots. +# +# A glibc development snapshot (say version 2.33.9000) may define +# symbols in its under-development symbol version (GLIBC_2.34). RPM +# automatically derives RPM dependencies such as +# libc.so.6(GLIBC_2.34)(64bit) from that. While the GLIBC_2.34 +# version is under development, these dependencies may be inaccurate +# and could be satisfied by glibc RPM package versions that lack the +# symbols because they were created from an earlier development +# snapshot that had some other GLIBC_2.34 symbols. Therefore, if the +# latest, under-development ELF symbol version is detected, this +# dependency generator adds an explicit RPM dependencies on the glibc +# packaging version against which an RPM package is built. +# +# This script runs for the glibc build itself. In this case, it may +# produce a >= dependency on the build-time glibc, but there will also +# be an (potentially indirect) = dependency, which takes precedence. + +set -e +set -o pipefail + +searching=true +# Pre-filter using eu-elfclassify, to skip kernel modules. +eu-elfclassify --loadable --file --stdin --print | while read path; do + # Assume that all dynamically linked objects depend on glibc in + # some way. + if $searching; then + # Undefined symbols within latest, under-development + # (changing) symbol versions trigger the versioned RPM + # dependency. Do not use "grep -Eq" to keep reading from the + # pipe, avoiding a spurious EPIPE error in eu-readelf. + if eu-readelf -s "$path" \ + | grep -E '\sUNDEF\s.*@''@SYMVER@(\s|$)' >/dev/null + then + echo 'glibc >= @VERSION@-@RELEASE@' + # Stop searching after the first match, but keep reading from + # the pipe. + searching=false + fi + fi +done diff --git a/SOURCES/parse-SUPPORTED.py b/SOURCES/parse-SUPPORTED.py new file mode 100644 index 0000000..cf512de --- /dev/null +++ b/SOURCES/parse-SUPPORTED.py @@ -0,0 +1,40 @@ +#!/usr/bin/python3 +# +# This script turns localedata/SUPPORTED (whose path is passed as the +# first argument) into a normalized list of LANGUAGE "_" REGION pairs. +# (If there is no REGION defined, only LANGUAGE is used.) The list +# is written to standard output, with one element per line. + +import sys + +supported, = sys.argv[1:] + +# Pairs seen so far. Used to suppress duplicates. +seen = set() +with open(supported) as inp: + for line in inp: + if line.startswith("#") or line == "SUPPORTED-LOCALES=\\\n": + # Comment or prefix. + continue + if not line.endswith(" \\\n"): + raise IOError("line without continuation: " + repr(line)) + try: + slash = line.index("/") + except ValueError: + raise IOError("line without slash: " + repr(line)) + spec = line[:slash] + for separator in ".@": + try: + # Strip charset, variant specifiers. + spec = spec[:spec.index(separator)] + except ValueError: + pass + seen.add(spec) + +# The C locale does not correspond to a language. +seen.remove("C") + +# The glibc source file is not sorted. +for spec in sorted(seen): + print(spec) +print() # The Lua generator produces a trailing newline. diff --git a/SOURCES/wrap-find-debuginfo.sh b/SOURCES/wrap-find-debuginfo.sh new file mode 100644 index 0000000..5257de7 --- /dev/null +++ b/SOURCES/wrap-find-debuginfo.sh @@ -0,0 +1,147 @@ +#!/bin/bash +# Wrapper script for find-debuginfo.sh +# +# Usage: +# wrap-find-debuginfo.sh SYSROOT-PATH SCRIPT-PATH SCRIPT-ARGS... +# +# The wrapper saves the original version of ld.so found in SYSROOT-PATH, +# invokes SCRIPT-PATH with SCRIPT-ARGS, and then restores the +# LDSO-PATH file, followed by note merging and DWZ compression. +# As a result, ld.so has (mostly) unchanged debuginfo even +# after debuginfo extraction. +# +# For libc.so.6, a set of strategic symbols is preserved in .symtab +# that are frequently used in valgrind suppressions and elsewhere. + +set -ex + +workdir="$(mktemp -d -t find_debuginfo.XXXXXX)" + +ldso_tmp="$workdir/ld.so" +libc_tmp_dir="$workdir/" + +# Return the path where a libc should be saved temporarily. This path is +# based on its original path received in $1. +libc_tmp_path() { + echo "$libc_tmp_dir"`dirname "$1"`"/libc.so" +} + +# Prefer a separately installed debugedit over the RPM-integrated one. +if command -v debugedit >/dev/null ; then + debugedit=debugedit +else + debugedit=/usr/lib/rpm/debugedit +fi + +cleanup () { + rm -rf "$workdir" +} +trap cleanup 0 + +sysroot_path="$1" +shift +# Resolve symbolic link, so that the activities below only alter the +# file it points to. +ldso_path="$(readlink -f "$sysroot_path/$1")" +shift +script_path="$1" +shift + +# libc.so.6 always uses this name, so it is simpler to locate. +# This can result in multiple paths, hence the loop below. +libc_path=`find "$sysroot_path" -name libc.so.6` + + +# Preserve the original files. +cp "$ldso_path" "$ldso_tmp" +for lib in $libc_path ; do + libtmp=`libc_tmp_path $lib` + mkdir -p `dirname "$libtmp"` + cp "$lib" "$libtmp" +done + +# Run the debuginfo extraction. +"$script_path" "$@" + +for lib in $libc_path ; do + libtmp=`libc_tmp_path "$lib"` + # libc.so.6: Extract the .gnu_debuglink section + objcopy -j.gnu_debuglink --set-section-flags .gnu_debuglink=alloc \ + -O binary "$lib" "$libtmp.debuglink" + # Restore the original files. + cp "$libtmp" "$lib" + + # Reduce the size of libc notes. Primarily for annobin. + objcopy --merge-notes "$lib" + + # libc.so.6: Restore the .gnu_debuglink section + objcopy --add-section .gnu_debuglink="$libtmp.debuglink" "$lib" + + # libc.so.6: Reduce to valuable symbols. Eliminate file symbols, + # annobin symbols, and symbols used by the glibc build to implement + # hidden aliases (__EI_*). We would also like to remove __GI_* + # symbols, but even listing them explicitly (as in -K __GI_strlen) + # still causes strip to remove them, so there is no filtering of + # __GI_* here. (Debuginfo is gone after this, so no need to optimize + # it.) + strip -w \ + -K '*' \ + -K '!*.c' \ + -K '!*.os' \ + -K '!.annobin_*' \ + -K '!__EI_*' \ + -K '!__PRETTY_FUNCTION__*' \ + "$lib" +done + +# Restore the original ld.so. +cp "$ldso_tmp" "$ldso_path" + +# Reduce the size of notes. Primarily for annobin. +objcopy --merge-notes "$ldso_path" + +# ld.so does not have separated debuginfo and so the debuginfo file +# generated by find-debuginfo is redundant. Therefore, remove it. +ldso_debug= +for ldso_debug_candidate in `find "$sysroot_path" -maxdepth 2 \ + -regextype posix-extended \ + -regex '.*/ld(-.*|64|)\.so\.[0-9]+.*debug$' -type f` ; do + if test -z "$ldso_debug" ; then + ldso_debug="$ldso_debug_candidate" + else + echo "error: multiple ld.so debug candidates: $ldso_debug, $ldso_debug_candidate" + exit 1 + fi +done +rm -f "$ldso_debug" + +# ld.so: Rewrite the source file paths to match the extracted +# locations. First compute the arguments for invoking debugedit. +# See find-debuginfo.sh. +debug_dest_name="/usr/src/debug" +last_arg= +while true ; do + arg="$1" + shift || break + case "$arg" in + (--unique-debug-src-base) + debug_dest_name="/usr/src/debug/$1" + shift + ;; + (-*) + ;; + (*) + last_arg="$arg" + ;; + esac +done +debug_base_name=${last_arg:-$RPM_BUILD_ROOT} +$debugedit -b "$debug_base_name" -d "$debug_dest_name" -n $ldso_path +# Remove the .annobin* symbols (and only them). +if nm --format=posix "$ldso_path" | cut -d' ' -f1 \ + | grep '^\.annobin' > "$ldso_tmp.annobin-symbols"; then + objcopy --strip-symbols="$ldso_tmp.annobin-symbols" "$ldso_path" +fi + +# Apply single-file DWARF optimization. +dwz $ldso_path diff --git a/SPECS/glibc.spec b/SPECS/glibc.spec new file mode 100644 index 0000000..3e465a1 --- /dev/null +++ b/SPECS/glibc.spec @@ -0,0 +1,4765 @@ +%global glibcsrcdir glibc-2.39 +%global glibcversion 2.39 +# Pre-release tarballs are pulled in from git using a command that is +# effectively: +# +# git archive HEAD --format=tar --prefix=$(git describe --match 'glibc-*')/ \ +# > $(git describe --match 'glibc-*').tar +# gzip -9 $(git describe --match 'glibc-*').tar +# +# glibc_release_url is only defined when we have a release tarball. +# Conversly, glibc_autorequires is set for development snapshots, where +# dependencies based on symbol versions are inaccurate. +%{lua: if string.match(rpm.expand("%glibcsrcdir"), "^glibc%-[0-9.]+$") then + rpm.define("glibc_release_url https://ftp.gnu.org/gnu/glibc/") + end + local major, minor = string.match(rpm.expand("%glibcversion"), + "^([0-9]+)%.([0-9]+)%.9000$") + if major and minor then + rpm.define("glibc_autorequires 1") + -- The minor version in a .9000 development version lags the actual + -- symbol version by one. + local symver = "GLIBC_" .. major .. "." .. (minor + 1) + rpm.define("glibc_autorequires_symver " .. symver) + else + rpm.define("glibc_autorequires 0") + end} +############################################################################## +# We support the following options: +# --with/--without, +# * testsuite - Running the testsuite. +# * benchtests - Running and building benchmark subpackage. +# * bootstrap - Bootstrapping the package. +# * werror - Build with -Werror +# * docs - Build with documentation and the required dependencies. +# * valgrind - Run smoke tests with valgrind to verify dynamic loader. +# +# You must always run the testsuite for production builds. +# Default: Always run the testsuite. +%bcond_without testsuite +# Default: Always build the benchtests. +%bcond_without benchtests +# Default: Not bootstrapping. +%bcond_with bootstrap +# Default: Enable using -Werror +%bcond_without werror +# Default: Always build documentation. +%bcond_without docs + +# Default: Always run valgrind tests if there is architecture support. +%ifarch %{valgrind_arches} +%bcond_without valgrind +%else +%bcond_with valgrind +%endif +# Restrict %%{valgrind_arches} further in case there are problems with +# the smoke test. +%if %{with valgrind} +%ifarch ppc64 ppc64p7 +# The valgrind smoke test does not work on ppc64, ppc64p7 (bug 1273103). +%undefine with_valgrind +%endif +%endif + +# Build the POWER10 multilib. +%ifarch ppc64le +%define buildpower10 1 +%else +%define buildpower10 0 +%endif + +%if %{with bootstrap} +# Disable benchtests, -Werror, docs, and valgrind if we're bootstrapping +%undefine with_benchtests +%undefine with_werror +%undefine with_docs +%undefine with_valgrind +%endif + +# We do our own build flags management. In particular, see +# glibc_shell_* below. +%undefine _auto_set_build_flags + +############################################################################## +# Any architecture/kernel combination that supports running 32-bit and 64-bit +# code in userspace is considered a biarch arch. +%global biarcharches %{ix86} x86_64 s390 s390x + +# Avoid generating a glibc-headers package on architectures which are +# not biarch. +%ifarch %{biarcharches} +%global need_headers_package 1 +%if 0%{?rhel} > 0 +%global headers_package_name glibc-headers +%else +%ifarch %{ix86} x86_64 +%global headers_package_name glibc-headers-x86 +%endif +%ifarch s390 s390x +%global headers_package_name glibc-headers-s390 +%endif +%dnl !rhel +%endif +%else +%global need_headers_package 0 +%dnl !biarcharches +%endif + +############################################################################## +# Utility functions for pre/post scripts. Stick them at the beginning of +# any lua %pre, %post, %postun, etc. sections to have them expand into +# those scripts. It only works in lua sections and not anywhere else. +%global glibc_post_funcs %{expand: +-- We use lua because there may be no shell that we can run during +-- glibc upgrade. We used to implement much of %%post as a C program, +-- but from an overall maintenance perspective the lua in the spec +-- file was simpler and safer given the operations required. +-- All lua code will be ignored by rpm-ostree; see: +-- https://github.com/projectatomic/rpm-ostree/pull/1869 +-- If we add new lua actions to the %%post code we should coordinate +-- with rpm-ostree and ensure that their glibc install is functional. +-- We must not use rpm.execute because this is a RPM 4.15 features and +-- we must still support downstream bootstrap with RPM 4.14 and missing +-- containerized boostrap. + +-- Open-code rpm.execute with error message handling. +function post_exec (msg, program, ...) + local pid = posix.fork () + if pid == 0 then + posix.exec (program, ...) + io.stdout:write (msg) + assert (nil) + elseif pid > 0 then + posix.wait (pid) + end +end + +-- Note: We use _prefix because Fedora's UsrMove says so. +function call_ldconfig () + post_exec("Error: call to %{_prefix}/sbin/ldconfig failed.\n", + "%{_prefix}/sbin/ldconfig") +end + +function update_gconv_modules_cache () + local iconv_dir = "%{_libdir}/gconv" + local iconv_cache = iconv_dir .. "/gconv-modules.cache" + local iconv_modules = iconv_dir .. "/gconv-modules" + if posix.utime(iconv_modules) == 0 then + if posix.utime (iconv_cache) == 0 then + post_exec ("Error: call to %{_prefix}/sbin/iconvconfig failed.\n", + "%{_prefix}/sbin/iconvconfig", + "-o", iconv_cache, + "--nostdlib", + iconv_dir) + else + io.stdout:write ("Error: Missing " .. iconv_cache .. " file.\n") + end + end +end} + +############################################################################## +# %%package glibc - The GNU C Library (glibc) core package. +############################################################################## +Summary: The GNU libc libraries +Name: glibc +Version: %{glibcversion} + +# We'll use baserelease here for two reasons: +# - It is known to rpmdev-bumpspec, so it will be properly handled for mass- +# rebuilds +# - It allows using the Release number without the %%dist tag in the dependency +# generator to make the generated requires interchangeable between Rawhide +# and ELN (.elnYY < .fcXX). +%global baserelease 22 +Release: %{baserelease}%{?dist} + +# Licenses: +# +# High level license status of the glibc source tree: +# +# * In general, GPLv2+ is used by programs, LGPLv2+ is used for +# libraries. +# +# * LGPLv2+ with exceptions is used for things that are linked directly +# into dynamically linked programs and shared libraries (e.g. crt +# files, lib*_nonshared.a). Historically, this exception also applies +# to parts of libio. +# +# * LGPLv2 is used in a couple of places (e.g. time/timespec_get.c, by +# mistake). +# +# * GPLv2+ with exceptions is used for parts of the Arm unwinder. +# +# * GFDL is used for the documentation. +# +# * Some other licenses are used in various places (BSD, Inner-Net, +# ISC, Public Domain, etc.). +# +# Licenses that make an appearance in the source tree but are not used: +# +# * HSRL and FSFAP are only used in test cases, which currently do not +# ship in binary RPMs, so they are not listed here. +# +# * GPLv3+ is used by manual/texinfo.tex, which we do not use and a test and +# some scripts that we do not ship, and so it is not listed here. +# +# * LGPLv3+ is used by some Hurd code, which we do not build. +# +# * A copyleft license is used in posix/runtests.c, but it is only a test +# case and so the license is not listed here. +# +# * A "PCRE License" is used by PCRE.tests, but it is only a test case and +# so the license is not listed here. +# +# * BSL-1.0 is only used by a test from boost and so the license is not +# listed here. +# +# * Unlicense is used in an OpenRISC 1000 file which we don't support. +# +# SPDX references: +# https://spdx.org/licenses +# https://docs.fedoraproject.org/en-US/legal/allowed-licenses +# https://gitlab.com/fedora/legal/fedora-license-data +# +# SPDX license string based on evaluation of glibc-2.39 sources by +# ScanCode toolkit (https://github.com/nexB/scancode-toolkit), +# and accounting for exceptions listed above: +License: LGPL-2.1-or-later AND SunPro AND LGPL-2.1-or-later WITH GCC-exception-2.0 AND BSD-3-Clause AND GPL-2.0-or-later AND LGPL-2.1-or-later WITH GNU-compiler-exception AND GPL-2.0-only AND ISC AND LicenseRef-Fedora-Public-Domain AND HPND AND CMU-Mach AND LGPL-2.1-only AND LGPL-2.0-or-later AND Unicode-DFS-2015 AND GFDL-1.1-or-later AND GPL-1.0-or-later AND FSFUL AND MIT AND Inner-Net-2.0 AND X11 AND GPL-2.0-or-later WITH GCC-exception-2.0 AND GFDL-1.3-only AND GFDL-1.1-only + +URL: http://www.gnu.org/software/glibc/ +Source0: %{?glibc_release_url}%{glibcsrcdir}.tar.xz +Source1: bench.mk +Source2: glibc-bench-compare +Source3: glibc.req.in +Source4: glibc.attr +Source10: wrap-find-debuginfo.sh +Source11: parse-SUPPORTED.py +# Include in the source RPM for reference. +Source12: ChangeLog.old + +# glibc_ldso: ABI-specific program interpreter name. Used for debuginfo +# extraction (wrap-find-debuginfo.sh) and smoke testing ($run_ldso below). +# +# glibc_rtld_early_cflags: The ABI baseline for architectures with +# potentially a later baseline. The --with-rtld-early-cflags= +# configure option is passed to the main glibc build if this macro is +# defined. +%ifarch %{ix86} +%global glibc_ldso /lib/ld-linux.so.2 +%endif +%ifarch aarch64 +%global glibc_ldso /lib/ld-linux-aarch64.so.1 +%endif +%ifarch ppc +%global glibc_ldso /lib/ld.so.1 +%endif +%ifarch ppc64 +%global glibc_ldso /lib64/ld64.so.1 +%endif +%ifarch ppc64le +%global glibc_ldso /lib64/ld64.so.2 +%define glibc_rtld_early_cflags -mcpu=power8 +%endif +%ifarch riscv64 +%global glibc_ldso /lib/ld-linux-riscv64-lp64d.so.1 +%endif +%ifarch s390 +%global glibc_ldso /lib/ld.so.1 +%define glibc_rtld_early_cflags -march=z13 +%endif +%ifarch s390x +%global glibc_ldso /lib/ld64.so.1 +%define glibc_rtld_early_cflags -march=z13 +%endif +%ifarch x86_64 x86_64_v2 x86_64_v3 x86_64_v4 +%global glibc_ldso /lib64/ld-linux-x86-64.so.2 +%define glibc_rtld_early_cflags -march=x86-64 +%endif + +###################################################################### +# Activate the wrapper script for debuginfo generation, by rewriting +# the definition of __debug_install_post. +%{lua: +local wrapper = rpm.expand("%{SOURCE10}") +local sysroot = rpm.expand("%{glibc_sysroot}") +local original = rpm.expand("%{macrobody:__debug_install_post}") +-- Strip leading newline. It confuses the macro redefinition. +-- Avoid embedded newlines that confuse the macro definition. +original = original:match("^%s*(.-)%s*$"):gsub("\\\n", "") +rpm.define("__debug_install_post bash " .. wrapper + .. " " .. sysroot .. " %{glibc_ldso} " .. original) +} + +# sysroot package support. These contain arch-specific packages, so +# turn off the rpmbuild check. +%global _binaries_in_noarch_packages_terminate_build 0 +# Variant of %%dist that contains just the distribution release, no affixes. +%{?fedora:%global sysroot_dist fc%{fedora}} +%{?rhel:%global sysroot_dist el%{rhel}} +%{?!sysroot_dist:%global sysroot_dist root} +# The name of the sysroot package. +%global sysroot_package_arch sysroot-%{_arch}-%{sysroot_dist}-%{name} +# Installed path for the sysroot tree. Must contain /sys-root/, which +# triggers filtering. +%global sysroot_prefix /usr/%{_arch}-redhat-linux/sys-root/%{sysroot_dist} + +# The wrapper script relies on the fact that debugedit does not change +# build IDs. +%global _no_recompute_build_ids 1 +%undefine _unique_build_ids + +%ifarch %{ix86} +# The memory tracing tools (like mtrace, memusage) in glibc-utils only work +# when the corresponding memory tracing libraries are preloaded. So we ship +# memory allocation tracing/checking libraries in glibc-utils, except on +# i686 where we need to ship them in glibc.i686. This is because +# glibc-utils.x86_64 will contain only the 64-bit version of these +# libraries. +%global glibc_ship_tracelibs_in_utils 0 +%else +%global glibc_ship_tracelibs_in_utils 1 +%endif + +############################################################################## +# Patches: +# - See each individual patch file for origin and upstream status. +# - For new patches follow template.patch format. +############################################################################## +Patch4: glibc-fedora-linux-tcsetattr.patch +Patch8: glibc-fedora-manual-dircategory.patch +Patch9: glibc-rh827510.patch +Patch13: glibc-fedora-localedata-rh61908.patch +Patch17: glibc-cs-path.patch +Patch23: glibc-python3.patch +Patch24: glibc-upstream-2.39-1.patch +Patch25: glibc-upstream-2.39-2.patch +Patch26: glibc-upstream-2.39-3.patch +Patch27: glibc-upstream-2.39-4.patch +Patch28: glibc-upstream-2.39-5.patch +Patch29: glibc-upstream-2.39-6.patch +Patch30: glibc-upstream-2.39-7.patch +Patch31: glibc-upstream-2.39-8.patch +Patch32: glibc-upstream-2.39-9.patch +Patch33: glibc-upstream-2.39-10.patch +Patch34: glibc-upstream-2.39-11.patch +Patch35: glibc-upstream-2.39-12.patch +Patch36: glibc-upstream-2.39-13.patch +Patch37: glibc-upstream-2.39-14.patch +Patch38: glibc-upstream-2.39-15.patch +Patch39: glibc-upstream-2.39-16.patch +Patch40: glibc-upstream-2.39-17.patch +Patch41: glibc-upstream-2.39-18.patch +Patch42: glibc-upstream-2.39-19.patch +Patch43: glibc-upstream-2.39-20.patch +Patch44: glibc-upstream-2.39-21.patch +Patch45: glibc-upstream-2.39-22.patch +Patch46: glibc-upstream-2.39-23.patch +Patch47: glibc-upstream-2.39-24.patch +Patch48: glibc-upstream-2.39-25.patch +Patch49: glibc-upstream-2.39-26.patch +Patch50: glibc-upstream-2.39-27.patch +Patch51: glibc-upstream-2.39-28.patch +Patch52: glibc-upstream-2.39-29.patch +Patch53: glibc-upstream-2.39-30.patch +Patch54: glibc-upstream-2.39-31.patch +Patch55: glibc-upstream-2.39-32.patch +Patch56: glibc-upstream-2.39-33.patch +Patch57: glibc-upstream-2.39-34.patch +Patch58: glibc-upstream-2.39-35.patch +Patch59: glibc-upstream-2.39-36.patch +Patch60: glibc-upstream-2.39-37.patch +Patch61: glibc-upstream-2.39-38.patch +Patch62: glibc-upstream-2.39-39.patch +Patch63: glibc-upstream-2.39-40.patch +Patch64: glibc-upstream-2.39-41.patch +Patch65: glibc-upstream-2.39-42.patch +Patch66: glibc-upstream-2.39-43.patch +Patch67: glibc-upstream-2.39-44.patch +Patch68: glibc-upstream-2.39-45.patch +Patch69: glibc-upstream-2.39-46.patch +Patch70: glibc-upstream-2.39-47.patch +Patch71: glibc-upstream-2.39-48.patch +Patch72: glibc-upstream-2.39-49.patch +Patch73: glibc-upstream-2.39-50.patch +Patch74: glibc-upstream-2.39-51.patch +Patch75: glibc-upstream-2.39-52.patch +Patch76: glibc-upstream-2.39-53.patch +Patch77: glibc-upstream-2.39-54.patch +Patch78: glibc-RHEL-22226.patch +Patch79: glibc-upstream-2.39-55.patch +Patch80: glibc-upstream-2.39-56.patch +Patch81: glibc-upstream-2.39-57.patch +Patch82: glibc-upstream-2.39-58.patch +Patch83: glibc-upstream-2.39-59.patch +Patch84: glibc-upstream-2.39-60.patch +Patch85: glibc-upstream-2.39-61.patch +Patch86: glibc-upstream-2.39-62.patch +Patch87: glibc-upstream-2.39-63.patch +Patch88: glibc-upstream-2.39-64.patch +Patch89: glibc-upstream-2.39-65.patch +Patch90: glibc-upstream-2.39-66.patch +Patch91: glibc-upstream-2.39-67.patch +Patch92: glibc-upstream-2.39-68.patch +Patch93: glibc-upstream-2.39-69.patch +Patch94: glibc-upstream-2.39-70.patch +Patch95: glibc-upstream-2.39-71.patch +Patch96: glibc-upstream-2.39-72.patch +# NEWS update: glibc-upstream-2.39-73.patch +# NEWS update: glibc-upstream-2.39-74.patch +Patch97: glibc-upstream-2.39-75.patch +Patch98: glibc-rh2292195-1.patch +Patch99: glibc-rh2292195-2.patch +Patch100: glibc-rh2292195-3.patch +Patch101: glibc-upstream-2.39-76.patch +Patch102: glibc-upstream-2.39-77.patch +Patch103: glibc-upstream-2.39-78.patch +Patch104: glibc-upstream-2.39-79.patch +Patch105: glibc-upstream-2.39-80.patch +Patch106: glibc-upstream-2.39-81.patch +Patch107: glibc-upstream-2.39-82.patch +Patch108: glibc-upstream-2.39-83.patch +Patch109: glibc-upstream-2.39-84.patch +Patch110: glibc-upstream-2.39-85.patch +Patch111: glibc-upstream-2.39-86.patch +Patch112: glibc-upstream-2.39-87.patch +Patch113: glibc-upstream-2.39-88.patch +Patch114: glibc-upstream-2.39-89.patch +Patch115: glibc-upstream-2.39-90.patch +Patch116: glibc-upstream-2.39-91.patch +Patch117: glibc-upstream-2.39-92.patch +Patch118: glibc-upstream-2.39-93.patch +Patch119: glibc-upstream-2.39-94.patch +Patch120: RHEL-18039-1.patch +Patch121: RHEL-18039-2.patch + +############################################################################## +# Continued list of core "glibc" package information: +############################################################################## +Obsoletes: glibc-profile < 2.4 +Obsoletes: nscd < 2.35 +Provides: ldconfig +Provides: /sbin/ldconfig +Provides: /usr/sbin/ldconfig + +# The dynamic linker supports DT_GNU_HASH +Provides: rtld(GNU_HASH) + +# We need libgcc for cancellation support in POSIX threads. +Requires: libgcc%{_isa} + +Requires: glibc-common = %{version}-%{release} + +# Various components (regex, glob) have been imported from gnulib. +Provides: bundled(gnulib) + +Requires(pre): basesystem +Requires: basesystem + +%ifarch %{ix86} +# Automatically install the 32-bit variant if the 64-bit variant has +# been installed. This covers the case when glibc.i686 is installed +# after nss_*.x86_64. (See below for the other ordering.) +Recommends: (nss_db(x86-32) if nss_db(x86-64)) +Recommends: (nss_hesiod(x86-32) if nss_hesiod(x86-64)) +# Deinstall the glibc32 package if present. This helps tests that do +# not run against the compose. +Conflicts: glibc32 <= %{version}-%{release} +Obsoletes: glibc32 <= %{version}-%{release} +%endif + +# This is for building auxiliary programs like memusage +# For initial glibc bootstraps it can be commented out +%if %{without bootstrap} +BuildRequires: gd-devel libpng-devel zlib-devel +%endif +%if %{with docs} +%endif +%if %{without bootstrap} +BuildRequires: libselinux-devel >= 1.33.4-3 +%endif +BuildRequires: audit-libs-devel >= 1.1.3, sed >= 3.95, libcap-devel, gettext +# We need procps-ng (/bin/ps), util-linux (/bin/kill), and gawk (/bin/awk), +# but it is more flexible to require the actual programs and let rpm infer +# the packages. However, until bug 1259054 is widely fixed we avoid the +# following: +# BuildRequires: /bin/ps, /bin/kill, /bin/awk +# And use instead (which should be reverted some time in the future): +BuildRequires: procps-ng, util-linux, gawk +BuildRequires: systemtap-sdt-devel + +%if %{with valgrind} +# Require valgrind for smoke testing the dynamic loader to make sure we +# have not broken valgrind. +BuildRequires: valgrind +%endif + +# We use python for the microbenchmarks and locale data regeneration +# from unicode sources (carried out manually). We choose python3 +# explicitly because it supports both use cases. On some +# distributions, python3 does not actually install /usr/bin/python3, +# so we also depend on python3-devel. +BuildRequires: python3 python3-devel + +# This GCC version is needed for -fstack-clash-protection support. +BuildRequires: gcc >= 7.2.1-6 +%global enablekernel 3.2 +Conflicts: kernel < %{enablekernel} +%define target %{_target_cpu}-redhat-linux +%ifarch ppc64le +%global target ppc64le-redhat-linux +%endif + +# GNU make 4.0 introduced the -O option. +BuildRequires: make >= 4.0 + +# The intl subsystem generates a parser using bison. +BuildRequires: bison >= 2.7 + +# binutils 2.30-17 is needed for --generate-missing-build-notes. +BuildRequires: binutils >= 2.30-17 + +# Earlier releases have broken support for IRELATIVE relocations +Conflicts: prelink < 0.4.2 + +%if %{without bootstrap} +%if %{with testsuite} +BuildRequires: diffutils +# The testsuite builds static C++ binaries that require a C++ compiler, +# static C++ runtime from libstdc++-static, and lastly static glibc. +BuildRequires: gcc-c++ +BuildRequires: libstdc++-static +# A configure check tests for the ability to create static C++ binaries +# before glibc is built and therefore we need a glibc-static for that +# check to pass even if we aren't going to use any of those objects to +# build the tests. +BuildRequires: glibc-static + +# libidn2 (but not libidn2-devel) is needed for testing AI_IDN/NI_IDN. +BuildRequires: libidn2 + +# The testsuite runs mtrace, which is a perl script +BuildRequires: perl-interpreter +%endif +%endif + +# Filter out all GLIBC_PRIVATE symbols since they are internal to +# the package and should not be examined by any other tool. +%global __filter_GLIBC_PRIVATE 1 +%global __provides_exclude ^libc_malloc_debug\\.so.*$ + +# For language packs we have glibc require a virtual dependency +# "glibc-langpack" wich gives us at least one installed langpack. +# If no langpack providing 'glibc-langpack' was installed you'd +# get language-neutral support e.g. C, POSIX, and C.UTF-8 locales. +# In the past we used to install the glibc-all-langpacks by default +# but we no longer do this to minimize container and VM sizes. +# Today you must actively use the language packs infrastructure to +# install language support. +Requires: glibc-langpack = %{version}-%{release} +Suggests: glibc-minimal-langpack = %{version}-%{release} + +# Suggest extra gconv modules so that they are installed by default but can be +# removed if needed to build a minimal OS image. +Recommends: glibc-gconv-extra%{_isa} = %{version}-%{release} +# Use redhat-rpm-config as a marker for a buildroot configuration, and +# unconditionally pull in glibc-gconv-extra in that case. +Requires: (glibc-gconv-extra%{_isa} = %{version}-%{release} if redhat-rpm-config) + +%description +The glibc package contains standard libraries which are used by +multiple programs on the system. In order to save disk space and +memory, as well as to make upgrading easier, common system code is +kept in one place and shared between programs. This particular package +contains the most important sets of shared libraries: the standard C +library and the standard math library. Without these two libraries, a +Linux system will not function. + +###################################################################### +# libnsl subpackage +###################################################################### + +%package -n libnsl +Summary: Legacy support library for NIS +Requires: %{name}%{_isa} = %{version}-%{release} + +%description -n libnsl +This package provides the legacy version of libnsl library, for +accessing NIS services. + +This library is provided for backwards compatibility only; +applications should use libnsl2 instead to gain IPv6 support. + +############################################################################## +# glibc "devel" sub-package +############################################################################## +%package devel +Summary: Object files for development using standard C libraries. +Requires: %{name} = %{version}-%{release} +Requires: libxcrypt-devel%{_isa} >= 4.0.0 +Requires: kernel-headers >= 3.2 +BuildRequires: kernel-headers >= 3.2 +%if %{need_headers_package} +Requires: %{headers_package_name} = %{version}-%{release} +%endif +%if !(0%{?rhel} > 0 && %{need_headers_package}) +# For backwards compatibility, when the glibc-headers package existed. +Provides: glibc-headers = %{version}-%{release} +Provides: glibc-headers(%{_target_cpu}) +Obsoletes: glibc-headers < %{version}-%{release} +%endif + +%description devel +The glibc-devel package contains the object files necessary +for developing programs which use the standard C libraries (which are +used by nearly all programs). If you are developing programs which +will use the standard C libraries, your system needs to have these +standard object files available in order to create the +executables. + +Install glibc-devel if you are going to develop programs which will +use the standard C libraries. + +############################################################################## +# glibc "doc" sub-package +############################################################################## +%if %{with docs} +%package doc +Summary: Documentation for GNU libc +BuildArch: noarch +Requires: %{name} = %{version}-%{release} + +# Removing texinfo will cause check-safety.sh test to fail because it seems to +# trigger documentation generation based on dependencies. We need to fix this +# upstream in some way that doesn't depend on generating docs to validate the +# texinfo. I expect it's simply the wrong dependency for that target. +BuildRequires: texinfo >= 5.0 + +%description doc +The glibc-doc package contains The GNU C Library Reference Manual in info +format. Additional package documentation is also provided. +%endif + +############################################################################## +# glibc "static" sub-package +############################################################################## +%package static +Summary: C library static libraries for -static linking. +Requires: %{name}-devel = %{version}-%{release} +Requires: libxcrypt-static%{?_isa} >= 4.0.0 + +%description static +The glibc-static package contains the C library static libraries +for -static linking. You don't need these, unless you link statically, +which is highly discouraged. + +############################################################################## +# glibc "headers" sub-package +# - The headers package includes all common headers that are shared amongst +# the multilib builds. It avoids file conflicts between the architecture- +# specific glibc-devel variants. +# Files like gnu/stubs.h which have gnu/stubs-32.h (i686) and gnu/stubs-64.h +# are included in glibc-headers, but the -32 and -64 files are in their +# respective i686 and x86_64 devel packages. +############################################################################## +%if %{need_headers_package} +%package -n %{headers_package_name} +Summary: Additional internal header files for glibc-devel. +Requires: %{name} = %{version}-%{release} +%if 0%{?rhel} > 0 +Provides: %{name}-headers(%{_target_cpu}) +Obsoletes: glibc-headers-x86 < %{version}-%{release} +Obsoletes: glibc-headers-s390 < %{version}-%{release} +%else +BuildArch: noarch +%endif + +%description -n %{headers_package_name} +The %{headers_package_name} package contains the architecture-specific +header files which cannot be included in glibc-devel package. +%endif + +############################################################################## +# glibc "common" sub-package +############################################################################## +%package common +Summary: Common binaries and locale data for glibc +Requires: %{name} = %{version}-%{release} +Recommends: tzdata >= 2003a + +%description common +The glibc-common package includes common binaries for the GNU libc +libraries, as well as national language (locale) support. + +###################################################################### +# File triggers to do ldconfig calls automatically (see rhbz#1380878) +###################################################################### + +# File triggers for when libraries are added or removed in standard +# paths. +%transfiletriggerin common -P 2000000 -p -- /lib /usr/lib /lib64 /usr/lib64 +%glibc_post_funcs +call_ldconfig() +%end + +%transfiletriggerpostun common -P 2000000 -p -- /lib /usr/lib /lib64 /usr/lib64 +%glibc_post_funcs +call_ldconfig() +%end + +# We need to run ldconfig manually because __brp_ldconfig assumes that +# glibc itself is always installed in $RPM_BUILD_ROOT, but with sysroots +# we may be installed into a subdirectory of that path. Therefore we +# unset __brp_ldconfig and run ldconfig by hand with the sysroots path +# passed to -r. +%undefine __brp_ldconfig + +###################################################################### + +%package locale-source +Summary: The sources for the locales +Requires: %{name} = %{version}-%{release} +Requires: %{name}-common = %{version}-%{release} + +%description locale-source +The sources for all locales provided in the language packs. +If you are building custom locales you will most likely use +these sources as the basis for your new locale. + +# We define a global regular expression to capture all of the locale +# sources. We use it later when constructing the various packages. +%global locale_rx eo syr tok *_* + +%{lua: +-- To make lua-mode happy: ' + +-- List of supported locales. This is used to generate the langpack +-- subpackages below. This table needs adjustments if the set of +-- glibc locales changes. "code" is the glibc code for the language +-- (before the "_". "name" is the English translation of the language +-- name (for use in subpackage descriptions). "regions" is a table of +-- variant specifiers (after the "_", excluding "@" and "." +-- variants/charset specifiers). The table must be sorted by the code +-- field, and the regions table must be sorted as well. +-- +-- English translations of language names can be obtained using (for +-- the "aa" language in this example): +-- +-- python3 -c 'import langtable; print(langtable.language_name("aa", languageIdQuery="en"))' + +local locales = { + { code="aa", name="Afar", regions={ "DJ", "ER", "ET" } }, + { code="af", name="Afrikaans", regions={ "ZA" } }, + { code="agr", name="Aguaruna", regions={ "PE" } }, + { code="ak", name="Akan", regions={ "GH" } }, + { code="am", name="Amharic", regions={ "ET" } }, + { code="an", name="Aragonese", regions={ "ES" } }, + { code="anp", name="Angika", regions={ "IN" } }, + { + code="ar", + name="Arabic", + regions={ + "AE", + "BH", + "DZ", + "EG", + "IN", + "IQ", + "JO", + "KW", + "LB", + "LY", + "MA", + "OM", + "QA", + "SA", + "SD", + "SS", + "SY", + "TN", + "YE" + } + }, + { code="as", name="Assamese", regions={ "IN" } }, + { code="ast", name="Asturian", regions={ "ES" } }, + { code="ayc", name="Southern Aymara", regions={ "PE" } }, + { code="az", name="Azerbaijani", regions={ "AZ", "IR" } }, + { code="be", name="Belarusian", regions={ "BY" } }, + { code="bem", name="Bemba", regions={ "ZM" } }, + { code="ber", name="Berber", regions={ "DZ", "MA" } }, + { code="bg", name="Bulgarian", regions={ "BG" } }, + { code="bhb", name="Bhili", regions={ "IN" } }, + { code="bho", name="Bhojpuri", regions={ "IN", "NP" } }, + { code="bi", name="Bislama", regions={ "VU" } }, + { code="bn", name="Bangla", regions={ "BD", "IN" } }, + { code="bo", name="Tibetan", regions={ "CN", "IN" } }, + { code="br", name="Breton", regions={ "FR" } }, + { code="brx", name="Bodo", regions={ "IN" } }, + { code="bs", name="Bosnian", regions={ "BA" } }, + { code="byn", name="Blin", regions={ "ER" } }, + { code="ca", name="Catalan", regions={ "AD", "ES", "FR", "IT" } }, + { code="ce", name="Chechen", regions={ "RU" } }, + { code="chr", name="Cherokee", regions={ "US" } }, + { code="ckb", name="Central Kurdish", regions={ "IQ" } }, + { code="cmn", name="Mandarin Chinese", regions={ "TW" } }, + { code="crh", name="Crimean Turkish", regions={ "RU", "UA" } }, + { code="cs", name="Czech", regions={ "CZ" } }, + { code="csb", name="Kashubian", regions={ "PL" } }, + { code="cv", name="Chuvash", regions={ "RU" } }, + { code="cy", name="Welsh", regions={ "GB" } }, + { code="da", name="Danish", regions={ "DK" } }, + { + code="de", + name="German", + regions={ "AT", "BE", "CH", "DE", "IT", "LI", "LU" } + }, + { code="doi", name="Dogri", regions={ "IN" } }, + { code="dsb", name="Lower Sorbian", regions={ "DE" } }, + { code="dv", name="Divehi", regions={ "MV" } }, + { code="dz", name="Dzongkha", regions={ "BT" } }, + { code="el", name="Greek", regions={ "CY", "GR" } }, + { + code="en", + name="English", + regions={ + "AG", + "AU", + "BW", + "CA", + "DK", + "GB", + "HK", + "IE", + "IL", + "IN", + "NG", + "NZ", + "PH", + "SC", + "SG", + "US", + "ZA", + "ZM", + "ZW" + } + }, + { code="eo", name="Esperanto", regions={} }, + { + code="es", + name="Spanish", + regions={ + "AR", + "BO", + "CL", + "CO", + "CR", + "CU", + "DO", + "EC", + "ES", + "GT", + "HN", + "MX", + "NI", + "PA", + "PE", + "PR", + "PY", + "SV", + "US", + "UY", + "VE" + } + }, + { code="et", name="Estonian", regions={ "EE" } }, + { code="eu", name="Basque", regions={ "ES" } }, + { code="fa", name="Persian", regions={ "IR" } }, + { code="ff", name="Fulah", regions={ "SN" } }, + { code="fi", name="Finnish", regions={ "FI" } }, + { code="fil", name="Filipino", regions={ "PH" } }, + { code="fo", name="Faroese", regions={ "FO" } }, + { code="fr", name="French", regions={ "BE", "CA", "CH", "FR", "LU" } }, + { code="fur", name="Friulian", regions={ "IT" } }, + { code="fy", name="Western Frisian", regions={ "DE", "NL" } }, + { code="ga", name="Irish", regions={ "IE" } }, + { code="gbm", name="Garhwali", regions={ "IN" } }, + { code="gd", name="Scottish Gaelic", regions={ "GB" } }, + { code="gez", name="Geez", regions={ "ER", "ET" } }, + { code="gl", name="Galician", regions={ "ES" } }, + { code="gu", name="Gujarati", regions={ "IN" } }, + { code="gv", name="Manx", regions={ "GB" } }, + { code="ha", name="Hausa", regions={ "NG" } }, + { code="hak", name="Hakka Chinese", regions={ "TW" } }, + { code="he", name="Hebrew", regions={ "IL" } }, + { code="hi", name="Hindi", regions={ "IN" } }, + { code="hif", name="Fiji Hindi", regions={ "FJ" } }, + { code="hne", name="Chhattisgarhi", regions={ "IN" } }, + { code="hr", name="Croatian", regions={ "HR" } }, + { code="hsb", name="Upper Sorbian", regions={ "DE" } }, + { code="ht", name="Haitian Creole", regions={ "HT" } }, + { code="hu", name="Hungarian", regions={ "HU" } }, + { code="hy", name="Armenian", regions={ "AM" } }, + { code="ia", name="Interlingua", regions={ "FR" } }, + { code="id", name="Indonesian", regions={ "ID" } }, + { code="ig", name="Igbo", regions={ "NG" } }, + { code="ik", name="Inupiaq", regions={ "CA" } }, + { code="is", name="Icelandic", regions={ "IS" } }, + { code="it", name="Italian", regions={ "CH", "IT" } }, + { code="iu", name="Inuktitut", regions={ "CA" } }, + { code="ja", name="Japanese", regions={ "JP" } }, + { code="ka", name="Georgian", regions={ "GE" } }, + { code="kab", name="Kabyle", regions={ "DZ" } }, + { code="kk", name="Kazakh", regions={ "KZ" } }, + { code="kl", name="Kalaallisut", regions={ "GL" } }, + { code="km", name="Khmer", regions={ "KH" } }, + { code="kn", name="Kannada", regions={ "IN" } }, + { code="ko", name="Korean", regions={ "KR" } }, + { code="kok", name="Konkani", regions={ "IN" } }, + { code="ks", name="Kashmiri", regions={ "IN" } }, + { code="ku", name="Kurdish", regions={ "TR" } }, + { code="kv", name="Komi", regions={ "RU" } }, + { code="kw", name="Cornish", regions={ "GB" } }, + { code="ky", name="Kyrgyz", regions={ "KG" } }, + { code="lb", name="Luxembourgish", regions={ "LU" } }, + { code="lg", name="Ganda", regions={ "UG" } }, + { code="li", name="Limburgish", regions={ "BE", "NL" } }, + { code="lij", name="Ligurian", regions={ "IT" } }, + { code="ln", name="Lingala", regions={ "CD" } }, + { code="lo", name="Lao", regions={ "LA" } }, + { code="lt", name="Lithuanian", regions={ "LT" } }, + { code="lv", name="Latvian", regions={ "LV" } }, + { code="lzh", name="Literary Chinese", regions={ "TW" } }, + { code="mag", name="Magahi", regions={ "IN" } }, + { code="mai", name="Maithili", regions={ "IN", "NP" } }, + { code="mfe", name="Morisyen", regions={ "MU" } }, + { code="mg", name="Malagasy", regions={ "MG" } }, + { code="mhr", name="Meadow Mari", regions={ "RU" } }, + { code="mi", name="Maori", regions={ "NZ" } }, + { code="miq", name="Miskito", regions={ "NI" } }, + { code="mjw", name="Karbi", regions={ "IN" } }, + { code="mk", name="Macedonian", regions={ "MK" } }, + { code="ml", name="Malayalam", regions={ "IN" } }, + { code="mn", name="Mongolian", regions={ "MN" } }, + { code="mni", name="Manipuri", regions={ "IN" } }, + { code="mnw", name="Mon", regions={ "MM" } }, + { code="mr", name="Marathi", regions={ "IN" } }, + { code="ms", name="Malay", regions={ "MY" } }, + { code="mt", name="Maltese", regions={ "MT" } }, + { code="my", name="Burmese", regions={ "MM" } }, + { code="nan", name="Min Nan Chinese", regions={ "TW" } }, + { code="nb", name="Norwegian Bokmål", regions={ "NO" } }, + { code="nds", name="Low German", regions={ "DE", "NL" } }, + { code="ne", name="Nepali", regions={ "NP" } }, + { code="nhn", name="Tlaxcala-Puebla Nahuatl", regions={ "MX" } }, + { code="niu", name="Niuean", regions={ "NU", "NZ" } }, + { code="nl", name="Dutch", regions={ "AW", "BE", "NL" } }, + { code="nn", name="Norwegian Nynorsk", regions={ "NO" } }, + { code="nr", name="South Ndebele", regions={ "ZA" } }, + { code="nso", name="Northern Sotho", regions={ "ZA" } }, + { code="oc", name="Occitan", regions={ "FR" } }, + { code="om", name="Oromo", regions={ "ET", "KE" } }, + { code="or", name="Odia", regions={ "IN" } }, + { code="os", name="Ossetic", regions={ "RU" } }, + { code="pa", name="Punjabi", regions={ "IN", "PK" } }, + { code="pap", name="Papiamento", regions={ "AW", "CW" } }, + { code="pl", name="Polish", regions={ "PL" } }, + { code="ps", name="Pashto", regions={ "AF" } }, + { code="pt", name="Portuguese", regions={ "BR", "PT" } }, + { code="quz", name="Cusco Quechua", regions={ "PE" } }, + { code="raj", name="Rajasthani", regions={ "IN" } }, + { code="rif", name="Tarifit", regions={ "MA" } }, + { code="ro", name="Romanian", regions={ "RO" } }, + { code="ru", name="Russian", regions={ "RU", "UA" } }, + { code="rw", name="Kinyarwanda", regions={ "RW" } }, + { code="sa", name="Sanskrit", regions={ "IN" } }, + { code="sah", name="Sakha", regions={ "RU" } }, + { code="sat", name="Santali", regions={ "IN" } }, + { code="sc", name="Sardinian", regions={ "IT" } }, + { code="sd", name="Sindhi", regions={ "IN" } }, + { code="se", name="Northern Sami", regions={ "NO" } }, + { code="sgs", name="Samogitian", regions={ "LT" } }, + { code="shn", name="Shan", regions={ "MM" } }, + { code="shs", name="Shuswap", regions={ "CA" } }, + { code="si", name="Sinhala", regions={ "LK" } }, + { code="sid", name="Sidamo", regions={ "ET" } }, + { code="sk", name="Slovak", regions={ "SK" } }, + { code="sl", name="Slovenian", regions={ "SI" } }, + { code="sm", name="Samoan", regions={ "WS" } }, + { code="so", name="Somali", regions={ "DJ", "ET", "KE", "SO" } }, + { code="sq", name="Albanian", regions={ "AL", "MK" } }, + { code="sr", name="Serbian", regions={ "ME", "RS" } }, + { code="ss", name="Swati", regions={ "ZA" } }, + { code="ssy", name="Saho", regions={ "ER" } }, + { code="st", name="Southern Sotho", regions={ "ZA" } }, + { code="su", name="Sudanese", regions={ "ID" } }, + { code="sv", name="Swedish", regions={ "FI", "SE" } }, + { code="sw", name="Swahili", regions={ "KE", "TZ" } }, + { code="syr", name="Syriac", regions={} }, + { code="szl", name="Silesian", regions={ "PL" } }, + { code="ta", name="Tamil", regions={ "IN", "LK" } }, + { code="tcy", name="Tulu", regions={ "IN" } }, + { code="te", name="Telugu", regions={ "IN" } }, + { code="tg", name="Tajik", regions={ "TJ" } }, + { code="th", name="Thai", regions={ "TH" } }, + { code="the", name="Chitwania Tharu", regions={ "NP" } }, + { code="ti", name="Tigrinya", regions={ "ER", "ET" } }, + { code="tig", name="Tigre", regions={ "ER" } }, + { code="tk", name="Turkmen", regions={ "TM" } }, + { code="tl", name="Tagalog", regions={ "PH" } }, + { code="tn", name="Tswana", regions={ "ZA" } }, + { code="to", name="Tongan", regions={ "TO" } }, + { code="tok", name="Toki Pona", regions={} }, + { code="tpi", name="Tok Pisin", regions={ "PG" } }, + { code="tr", name="Turkish", regions={ "CY", "TR" } }, + { code="ts", name="Tsonga", regions={ "ZA" } }, + { code="tt", name="Tatar", regions={ "RU" } }, + { code="ug", name="Uyghur", regions={ "CN" } }, + { code="uk", name="Ukrainian", regions={ "UA" } }, + { code="unm", name="Unami language", regions={ "US" } }, + { code="ur", name="Urdu", regions={ "IN", "PK" } }, + { code="uz", name="Uzbek", regions={ "UZ" } }, + { code="ve", name="Venda", regions={ "ZA" } }, + { code="vi", name="Vietnamese", regions={ "VN" } }, + { code="wa", name="Walloon", regions={ "BE" } }, + { code="wae", name="Walser", regions={ "CH" } }, + { code="wal", name="Wolaytta", regions={ "ET" } }, + { code="wo", name="Wolof", regions={ "SN" } }, + { code="xh", name="Xhosa", regions={ "ZA" } }, + { code="yi", name="Yiddish", regions={ "US" } }, + { code="yo", name="Yoruba", regions={ "NG" } }, + { code="yue", name="Cantonese", regions={ "HK" } }, + { code="yuw", name="Yau", regions={ "PG" } }, + { code="zgh", name="Tamazight", regions={ "MA" } }, + { code="zh", name="Mandarin Chinese", regions={ "CN", "HK", "SG", "TW" } }, + { code="zu", name="Zulu", regions={ "ZA" } } +} + +-- Prints a list of LANGUAGE "_" REGION pairs. The output is expected +-- to be identical to parse-SUPPORTED.py. Called from the %%prep section. +function print_locale_pairs() + for i = 1, #locales do + local locale = locales[i] + if #locale.regions == 0 then + print(locale.code .. "\n") + else + for j = 1, #locale.regions do + print(locale.code .. "_" .. locale.regions[j] .. "\n") + end + end + end +end + +local function compute_supplements(locale) + local lang = locale.code + local regions = locale.regions + result = "langpacks-core-" .. lang + for i = 1, #regions do + result = result .. " or langpacks-core-" .. lang .. "_" .. regions[i] + end + return result +end + +-- Emit the definition of a language pack package. +local function lang_package(locale) + local lang = locale.code + local langname = locale.name + local suppl = compute_supplements(locale) + print(rpm.expand([[ + +%package langpack-]]..lang..[[ + +Summary: Locale data for ]]..langname..[[ + +Provides: glibc-langpack = %{version}-%{release} +Requires: %{name} = %{version}-%{release} +Requires: %{name}-common = %{version}-%{release} +Supplements: ((glibc and (]]..suppl..[[)) unless glibc-all-langpacks) +%description langpack-]]..lang..[[ + +The glibc-langpack-]]..lang..[[ package includes the basic information required +to support the ]]..langname..[[ language in your applications. +%files -f langpack-]]..lang..[[.filelist langpack-]]..lang..[[ +]])) +end + +for i = 1, #locales do + lang_package(locales[i]) +end +} + +# The glibc-all-langpacks provides the virtual glibc-langpack, +# and thus satisfies glibc's requirement for installed locales. +# Users can add one more other langauge packs and then eventually +# uninstall all-langpacks to save space. +%package all-langpacks +Summary: All language packs for %{name}. +Requires: %{name} = %{version}-%{release} +Requires: %{name}-common = %{version}-%{release} +Provides: %{name}-langpack = %{version}-%{release} +%description all-langpacks + +# No %files, this is an empty package. The C/POSIX and +# C.UTF-8 files are already installed by glibc. We create +# minimal-langpack because the virtual provide of +# glibc-langpack needs at least one package installed +# to satisfy it. Given that no-locales installed is a valid +# use case we support it here with this package. +%package minimal-langpack +Summary: Minimal language packs for %{name}. +Provides: glibc-langpack = %{version}-%{release} +Requires: %{name} = %{version}-%{release} +Requires: %{name}-common = %{version}-%{release} +%description minimal-langpack +This is a Meta package that is used to install minimal language packs. +This package ensures you can use C, POSIX, or C.UTF-8 locales, but +nothing else. It is designed for assembling a minimal system. +%files minimal-langpack + +# Infrequently used iconv converter modules. +%package gconv-extra +Summary: All iconv converter modules for %{name}. +Requires: %{name}%{_isa} = %{version}-%{release} +Requires: %{name}-common = %{version}-%{release} + +%description gconv-extra +This package contains all iconv converter modules built in %{name}. + +############################################################################## +# Subpackages for NSS modules except nss_files, nss_compat, nss_dns +############################################################################## + +# This should remain it's own subpackage or "Provides: nss_db" to allow easy +# migration from old systems that previously had the old nss_db package +# installed. Note that this doesn't make the migration that smooth, the +# databases still need rebuilding because the formats were different. +# The nss_db package was deprecated in F16 and onwards: +# https://lists.fedoraproject.org/pipermail/devel/2011-July/153665.html +# The different database format does cause some issues for users: +# https://lists.fedoraproject.org/pipermail/devel/2011-December/160497.html +%package -n nss_db +Summary: Name Service Switch (NSS) module using hash-indexed files +Requires: %{name}%{_isa} = %{version}-%{release} +%ifarch x86_64 +# Automatically install the 32-bit variant if the 64-bit variant has +# been installed. This covers the case when glibc.i686 is installed +# before nss_db.x86_64. (See above for the other ordering.) +Recommends: (nss_db(x86-32) if glibc(x86-32)) +%endif + +%description -n nss_db +The nss_db Name Service Switch module uses hash-indexed files in /var/db +to speed up user, group, service, host name, and other NSS-based lookups. + +%package -n nss_hesiod +Summary: Name Service Switch (NSS) module using Hesiod +Requires: %{name}%{_isa} = %{version}-%{release} +%ifarch x86_64 +# Automatically install the 32-bit variant if the 64-bit variant has +# been installed. This covers the case when glibc.i686 is installed +# before nss_hesiod.x86_64. (See above for the other ordering.) +Recommends: (nss_hesiod(x86-32) if glibc(x86-32)) +%endif + +%description -n nss_hesiod +The nss_hesiod Name Service Switch module uses the Domain Name System +(DNS) as a source for user, group, and service information, following +the Hesiod convention of Project Athena. + +%package nss-devel +Summary: Development files for directly linking NSS service modules +Requires: %{name}%{_isa} = %{version}-%{release} +Requires: nss_db%{_isa} = %{version}-%{release} +Requires: nss_hesiod%{_isa} = %{version}-%{release} + +%description nss-devel +The glibc-nss-devel package contains the object files necessary to +compile applications and libraries which directly link against NSS +modules supplied by glibc. + +This is a rare and special use case; regular development has to use +the glibc-devel package instead. + +############################################################################## +# glibc "utils" sub-package +############################################################################## +%package utils +Summary: Development utilities from GNU C library +Requires: %{name} = %{version}-%{release} +Requires: /usr/bin/perl + +%description utils +The glibc-utils package contains memusage, a memory usage profiler, +mtrace, a memory leak tracer and xtrace, a function call tracer +which can be helpful during program debugging. + +If unsure if you need this, don't install this package. + +%if %{with benchtests} +%package benchtests +Summary: Benchmarking binaries and scripts for %{name} +%description benchtests +This package provides built benchmark binaries and scripts to run +microbenchmark tests on the system. +%endif + +############################################################################## +# compat-libpthread-nonshared +# See: https://sourceware.org/bugzilla/show_bug.cgi?id=23500 +############################################################################## +%package -n compat-libpthread-nonshared +Summary: Compatibility support for linking against libpthread_nonshared.a. + +%description -n compat-libpthread-nonshared +This package provides compatibility support for applications that expect +libpthread_nonshared.a to exist. The support provided is in the form of +an empty libpthread_nonshared.a that allows dynamic links to succeed. +Such applications should be adjusted to avoid linking against +libpthread_nonshared.a which is no longer used. The static library +libpthread_nonshared.a is an internal implementation detail of the C +runtime and should not be expected to exist. + +%if %{without bootstrap} +%package -n %sysroot_package_arch +Summary: Sysroot package for glibc, %{_arch} architecture +BuildArch: noarch +Provides: sysroot-%{_arch}-%{name} +# The files are not usable for execution, so do not provide nor +# require anything. +AutoReqProv: no + +%description -n %sysroot_package_arch +This package contains development files for the glibc package +that can be installed across architectures. +%dnl %%{without bootstrap} +%endif + +############################################################################## +# glibc32 (only for use in building GCC, not shipped) +############################################################################## +%ifarch x86_64 +%package -n glibc32 +Summary: The GNU libc libraries (32-bit) +Conflicts: glibc(x86-32) +%dnl The gcc package does not use ELF dependencies to install glibc32: +%dnl BuildRequires: (glibc32 or glibc-devel(%{__isa_name}-32)) +%dnl Not generating the ELF dependencies for glibc32 makes it less likely +%dnl that the package is selected by accident over glibc.i686. +AutoReqProv: no + +%description -n glibc32 +This package is only used for internal building of multilib aware +packages, like gcc, due to a technical limitation in the distribution +build environment. Any package which needs both 32-bit and 64-bit +runtimes at the same time must install glibc32 (marked as a 64-bit +package) to access the 32-bit development files during a 64-bit build. + +This package is not supported or intended for use outside of the +distribution build enviroment. Regular users can install both 32-bit and +64-bit runtimes and development files without any problems. + +%endif + +############################################################################## +# Prepare for the build. +############################################################################## +%prep +%autosetup -n %{glibcsrcdir} -p1 + +############################################################################## +# %%prep - Additional prep required... +############################################################################## +# Make benchmark scripts executable +chmod +x benchtests/scripts/*.py scripts/pylint + +# Remove all files generated from patching. +find . -type f -size 0 -o -name "*.orig" -exec rm -f {} \; + +# Ensure timestamps on configure files are current to prevent +# regenerating them. +touch `find . -name configure` + +# Ensure *-kw.h files are current to prevent regenerating them. +touch locale/programs/*-kw.h + +# Verify that our locales table is compatible with the locales table +# in the spec file. +set +x +echo '%{lua: print_locale_pairs()}' > localedata/SUPPORTED.spec +set -x +python3 %{SOURCE11} localedata/SUPPORTED > localedata/SUPPORTED.glibc +diff -u \ + --label "spec file" localedata/SUPPORTED.spec \ + --label "glibc localedata/SUPPORTED" localedata/SUPPORTED.glibc +rm localedata/SUPPORTED.spec localedata/SUPPORTED.glibc + +############################################################################## +# Build glibc... +############################################################################## +%build +# Log osystem information +uname -a +LD_SHOW_AUXV=1 /bin/true +ld.so --list-diagnostics || true +ld.so --list-tunables || true +cat /proc/cpuinfo +cat /proc/sysinfo 2>/dev/null || true +cat /proc/meminfo +df + +# Propgate select compiler flags from redhat-rpm-config. These flags +# are target-dependent, so we use only those which are specified in +# redhat-rpm-config. We keep the -m32/-m32/-m64 flags to support +# multilib builds. + +%{lua: +-- Split the string argument into keys of an associate array. +-- The values are set to true. +local function string_to_array(s) + local result = {} + for e in string.gmatch(s, "%S+") do + result[e] = true + end + return result +end + +local inherit_flags = {} + +-- These flags are put into the CC and CXX arguments to configure. +-- Alternate builds do not use the flags listed here, only the main build does. +inherit_flags.cc_main = string_to_array [[ +-march=armv8-a+lse +-march=armv8.1-a +-march=haswell +-march=i686 +-march=x86-64 +-march=x86-64-v2 +-march=x86-64-v3 +-march=x86-64-v4 +-march=z13 +-march=z14 +-march=z15 +-march=zEC12 +-mcpu=power10 +-mcpu=power8 +-mcpu=power9 +-mtune=generic +-mtune=power10 +-mtune=power8 +-mtune=power9 +-mtune=z13 +-mtune=z14 +-mtune=z15 +-mtune=zEC12 +]] + +-- Like inherit_flags_cc_main, but also used for alternate builds. +inherit_flags.cc = string_to_array [[ +-m31 +-m32 +-m64 +]] + +-- These flags are passed through CFLAGS and CXXFLAGS. +inherit_flags.cflags = string_to_array [[ +-O2 +-O3 +-Wall +-Wp,-D_GLIBCXX_ASSERTIONS +-fasynchronous-unwind-tables +-fno-omit-frame-pointer +-fstack-clash-protection +-funwind-tables +-g +-mbackchain +-mbranch-protection=standard +-mfpmath=sse +-mno-omit-leaf-frame-pointer +-msse2 +-mstackrealign +-specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 +]] + +-- Iterate over the build_cflags RPM variable and emit a shell +-- variable that contains the inherited flags of the indicated variant. +local function shell_build_flags(variant) + local result = {} + local inherit = assert(inherit_flags[variant]) + for f in string.gmatch(rpm.expand("%build_cflags"), "%S+") do + if inherit[f] then + result[#result + 1] = f + end + end + print("glibc_flags_" .. variant .. "=\"" .. table.concat(result, " ") + .. "\"\n") +end + +shell_build_flags('cc_main') -- Set $glibc_flags_cc_main. +shell_build_flags('cc') -- Set $glibc_flags_cc. +shell_build_flags('cflags') -- Set $glibc_flags_cflags. +} + +%if 0%{?_annotated_build} > 0 +# libc_nonshared.a cannot be built with the default hardening flags +# because the glibc build system is incompatible with +# -D_FORTIFY_SOURCE. The object files need to be marked as to be +# skipped in annobin annotations. (The -specs= variant of activating +# annobin does not work here because of flag ordering issues.) +# See . +BuildFlagsNonshared="-fplugin=annobin -fplugin-arg-annobin-disable -Wa,--generate-missing-build-notes=yes" +%endif + +# Special flag to enable annobin annotations for statically linked +# assembler code. Needs to be passed to make; not preserved by +# configure. +%global glibc_make_flags_as ASFLAGS="-g -Wa,--generate-missing-build-notes=yes" +%global glibc_make_flags %{glibc_make_flags_as} + +############################################################################## +# %%build - Generic options. +############################################################################## +EnableKernel="--enable-kernel=%{enablekernel}" + +############################################################################## +# build() +# Build glibc in the directory $1, passing the rest of the arguments +# as additional configure arguments. Several +# global values are used to determine build flags, kernel version, +# system tap support, etc. +############################################################################## +build() +{ + local builddir=$1 + shift + rm -rf $builddir + mkdir $builddir + pushd $builddir + ../configure "$@" \ + --prefix=%{_prefix} \ + --with-headers=%{_prefix}/include $EnableKernel \ + --with-nonshared-cflags="$BuildFlagsNonshared" \ + --enable-bind-now \ + --build=%{target} \ + --enable-stack-protector=strong \ + --enable-systemtap \ +%ifarch %{ix86} + --disable-multi-arch \ +%endif +%if %{without werror} + --disable-werror \ +%endif + --disable-profile \ +%if %{with bootstrap} + --without-selinux \ +%endif +%ifarch aarch64 + --enable-memory-tagging \ +%endif + --disable-crypt \ + --disable-build-nscd \ + --disable-nscd \ + --enable-fortify-source || + { cat config.log; false; } + + # We enable DT_GNU_HASH and DT_HASH for ld.so and DSOs to improve + # compatibility with applications that expect DT_HASH e.g. Epic Games + # Easy Anti-Cheat. This is temporary as applications move to + # supporting only DT_GNU_HASH. This was initially enabled in Fedora + # 37. We must use 'env' because it is the only way to pass, via the + # environment, two variables that set the initial Makefile values for + # LDFLAGS used to build shared objects and the dynamic loader. + env LDFLAGS.so="-Wl,--hash-style=both" \ + LDFLAGS-rtld="-Wl,--hash-style=both" \ + %make_build -r %{glibc_make_flags} + popd +} + +%ifarch x86_64 +# Build for the glibc32 package. +build build-%{target}-32 \ + CC="gcc -m32" \ + CXX="g++ -m32" \ + CFLAGS="${glibc_flags_cflags/-m64/-m32}" \ + --host=i686-linux-gnu \ +# +%endif + +# Default set of compiler options. +build build-%{target} \ + CC="gcc $glibc_flags_cc $glibc_flags_cc_main" \ + CXX="g++ $glibc_flags_cc $glibc_flags_cc_main" \ + CFLAGS="$glibc_flags_cflags" \ + %{?glibc_rtld_early_cflags:--with-rtld-early-cflags=%glibc_rtld_early_cflags} \ +%ifarch x86_64 + --enable-cet \ +%endif +# + +# POWER10 build variant. +%if %{buildpower10} +build build-%{target}-power10 \ + CC="gcc $glibc_flags_cc" \ + CXX="g++ $glibc_flags_cc" \ + CFLAGS="$glibc_flags_cflags" \ + --with-cpu=power10 \ +# +%endif + + +############################################################################## +# Install glibc... +############################################################################## +%install + +# The built glibc is installed into a subdirectory of $RPM_BUILD_ROOT. +# For a system glibc that subdirectory is "/" (the root of the filesystem). +# This is called a sysroot (system root) and can be changed if we have a +# distribution that supports multiple installed glibc versions. +%global glibc_sysroot $RPM_BUILD_ROOT + +# Remove existing file lists. +find . -type f -name '*.filelist' -exec rm -rf {} \; + +%ifarch riscv64 +# RISC-V ABI wants to install everything in /lib64/lp64d or /usr/lib64/lp64d. +# Make these be symlinks to /lib64 or /usr/lib64 respectively. See: +# https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/thread/DRHT5YTPK4WWVGL3GIN5BF2IKX2ODHZ3/ +for d in %{glibc_sysroot}%{_libdir} %{glibc_sysroot}/%{_lib}; do + mkdir -p $d + (cd $d && ln -sf . lp64d) +done +%endif + +%ifarch x86_64 +# Install for the glibc32 package. +pushd build-%{target}-32 +%make_build install_root=%{glibc_sysroot} install +popd +pushd %{glibc_sysroot} +rm -rf etc sbin var usr/bin usr/lib/gconv usr/libexec usr/sbin usr/share +rm -f lib/libnss_db* lib/libnss_hesiod* lib/libnsl* usr/lib/libnsl* usr/lib/libnss* +rm usr/lib/libc_malloc_debug.so +strip -g usr/lib/*.o +mv lib/{libmemusage,libpcprofile}.so usr/lib/ +popd +%endif + +# Build and install: +pushd build-%{target} +%make_build install_root=%{glibc_sysroot} install +%make_build install_root=%{glibc_sysroot} \ + install-locale-files -C ../localedata objdir=`pwd` +popd +# Locale creation via install-locale-files does not group identical files +# via hardlinks, so we must group them ourselves. +hardlink -c %{glibc_sysroot}/usr/lib/locale + +# install_different: +# Install all core libraries into DESTDIR/SUBDIR. Either the file is +# installed as a copy or a symlink to the default install (if it is the +# same). The path SUBDIR_UP is the prefix used to go from +# DESTDIR/SUBDIR to the default installed libraries e.g. +# ln -s SUBDIR_UP/foo.so DESTDIR/SUBDIR/foo.so. +# When you call this function it is expected that you are in the root +# of the build directory, and that the default build directory is: +# "../build-%{target}" (relatively). +# The primary use of this function is to install alternate runtimes +# into the build directory and avoid duplicating this code for each +# runtime. +install_different() +{ + local lib libbase libbaseso dlib + local destdir="$1" + local subdir="$2" + local subdir_up="$3" + local libdestdir="$destdir/$subdir" + # All three arguments must be non-zero paths. + if ! [ "$destdir" \ + -a "$subdir" \ + -a "$subdir_up" ]; then + echo "One of the arguments to install_different was emtpy." + exit 1 + fi + # Create the destination directory and the multilib directory. + mkdir -p "$destdir" + mkdir -p "$libdestdir" + # Walk all of the libraries we installed... + for lib in libc math/libm + do + libbase=${lib#*/} + # Take care that `libbaseso' has a * that needs expanding so + # take care with quoting. + libbaseso=$(basename %{glibc_sysroot}/%{_lib}/${libbase}.so.*) + # Only install if different from default build library. + if cmp -s ${lib}.so ../build-%{target}/${lib}.so; then + ln -sf "$subdir_up"/$libbaseso $libdestdir/$libbaseso + else + cp -a ${lib}.so $libdestdir/$libbaseso + fi + done +} + +%if %{buildpower10} +pushd build-%{target}-power10 +install_different "$RPM_BUILD_ROOT/%{_libdir}/glibc-hwcaps" power10 .. +popd +%endif + + +############################################################################## +# Remove the files we don't want to distribute +############################################################################## + +# Remove the libNoVersion files. +# XXX: This looks like a bug in glibc that accidentally installed these +# wrong files. We probably don't need this today. +rm -f %{glibc_sysroot}/%{_libdir}/libNoVersion* +rm -f %{glibc_sysroot}/%{_lib}/libNoVersion* + +# Remove the old nss modules. +rm -f %{glibc_sysroot}/%{_lib}/libnss1-* +rm -f %{glibc_sysroot}/%{_lib}/libnss-*.so.1 + +# This statically linked binary is no longer necessary in a world where +# the default Fedora install uses an initramfs, and further we have rpm-ostree +# which captures the whole userspace FS tree. +# Further, see https://github.com/projectatomic/rpm-ostree/pull/1173#issuecomment-355014583 +rm -f %{glibc_sysroot}/{usr/,}sbin/sln + +###################################################################### +# Run ldconfig to create all the symbolic links we need +###################################################################### + +# Note: This has to happen before creating /etc/ld.so.conf. + +mkdir -p %{glibc_sysroot}/var/cache/ldconfig +truncate -s 0 %{glibc_sysroot}/var/cache/ldconfig/aux-cache + +# ldconfig is statically linked, so we can use the new version. +%{glibc_sysroot}/sbin/ldconfig -N -r %{glibc_sysroot} + +############################################################################## +# Install info files +############################################################################## + +%if %{with docs} +# Move the info files if glibc installed them into the wrong location. +if [ -d %{glibc_sysroot}%{_prefix}/info -a "%{_infodir}" != "%{_prefix}/info" ]; then + mkdir -p %{glibc_sysroot}%{_infodir} + mv -f %{glibc_sysroot}%{_prefix}/info/* %{glibc_sysroot}%{_infodir} + rm -rf %{glibc_sysroot}%{_prefix}/info +fi + +# Compress all of the info files. +gzip -9nvf %{glibc_sysroot}%{_infodir}/libc* + +# Copy the debugger interface documentation over to the right location +mkdir -p %{glibc_sysroot}%{_docdir}/glibc +cp elf/rtld-debugger-interface.txt %{glibc_sysroot}%{_docdir}/glibc +cp posix/gai.conf %{glibc_sysroot}%{_docdir}/glibc +%else +rm -f %{glibc_sysroot}%{_infodir}/dir +rm -f %{glibc_sysroot}%{_infodir}/libc.info* +%endif + +############################################################################## +# Create locale sub-package file lists +############################################################################## + +olddir=`pwd` +pushd %{glibc_sysroot}%{_prefix}/lib/locale +rm -f locale-archive +$olddir/build-%{target}/elf/ld.so \ + --library-path $olddir/build-%{target}/ \ + $olddir/build-%{target}/locale/localedef \ + --alias-file=$olddir/intl/locale.alias \ + --prefix %{glibc_sysroot} --add-to-archive \ + %locale_rx +# Historically, glibc-all-langpacks deleted the file on updates (sic), +# so we need to restore it in the posttrans scriptlet (like the old +# glibc-all-langpacks versions) +ln locale-archive locale-archive.real + +# Almost half the LC_CTYPE files in langpacks are identical to the C.utf8 +# variant which is installed by default. When we keep them as hardlinks, +# each langpack ends up retaining a copy. If we convert these to symbolic +# links instead, we save ~350K each when they get installed that way. +# +# LC_MEASUREMENT and LC_PAPER also have several duplicates but we don't +# bother with these because they are only ~30 bytes each. +pushd %{glibc_sysroot}/usr/lib/locale +for f in $(find %locale_rx -samefile C.utf8/LC_CTYPE); do + rm $f && ln -s '../C.utf8/LC_CTYPE' $f +done +popd + +# Create the file lists for the language specific sub-packages: +for i in %locale_rx +do + lang=${i%%_*} + if [ ! -e langpack-${lang}.filelist ]; then + echo "%dir %{_prefix}/lib/locale" >> langpack-${lang}.filelist + fi + echo "%dir %{_prefix}/lib/locale/$i" >> langpack-${lang}.filelist + echo "%{_prefix}/lib/locale/$i/*" >> langpack-${lang}.filelist +done +popd +pushd %{glibc_sysroot}%{_prefix}/share/locale +for i in */LC_MESSAGES/libc.mo +do + locale=${i%%%%/*} + lang=${locale%%%%_*} + echo "%lang($lang) %{_prefix}/share/locale/${i}" \ + >> %{glibc_sysroot}%{_prefix}/lib/locale/langpack-${lang}.filelist +done +popd +mv %{glibc_sysroot}%{_prefix}/lib/locale/*.filelist . + +############################################################################## +# Install configuration files for services +############################################################################## + +# Include ld.so.conf +echo 'include ld.so.conf.d/*.conf' > %{glibc_sysroot}/etc/ld.so.conf +truncate -s 0 %{glibc_sysroot}/etc/ld.so.cache +chmod 644 %{glibc_sysroot}/etc/ld.so.conf +mkdir -p %{glibc_sysroot}/etc/ld.so.conf.d +truncate -s 0 %{glibc_sysroot}/etc/gai.conf + +# Include %{_libdir}/gconv/gconv-modules.cache +truncate -s 0 %{glibc_sysroot}%{_libdir}/gconv/gconv-modules.cache +chmod 644 %{glibc_sysroot}%{_libdir}/gconv/gconv-modules.cache + +# Remove any zoneinfo files; they are maintained by tzdata. +rm -rf %{glibc_sysroot}%{_prefix}/share/zoneinfo + +# Make sure %config files have the same timestamp across multilib packages. +# +# XXX: Ideally ld.so.conf should have the timestamp of the spec file, but there +# doesn't seem to be any macro to give us that. So we do the next best thing, +# which is to at least keep the timestamp consistent. The choice of using +# SOURCE0 is arbitrary. +touch -r %{SOURCE0} %{glibc_sysroot}/etc/ld.so.conf +touch -r inet/etc.rpc %{glibc_sysroot}/etc/rpc + +%if %{with benchtests} +# Build benchmark binaries. Ignore the output of the benchmark runs. +pushd build-%{target} +make BENCH_DURATION=1 bench-build +popd + +# Copy over benchmark binaries. +mkdir -p %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests +cp $(find build-%{target}/benchtests -type f -executable) %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +# ... and the makefile. +for b in %{SOURCE1} %{SOURCE2}; do + cp $b %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +done +# .. and finally, the comparison scripts. +cp benchtests/scripts/benchout.schema.json %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +cp benchtests/scripts/compare_bench.py %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +cp benchtests/scripts/import_bench.py %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +cp benchtests/scripts/validate_benchout.py %{glibc_sysroot}%{_prefix}/libexec/glibc-benchtests/ +%endif + +# The #line directives gperf generates do not give the proper +# file name relative to the build directory. +pushd locale +ln -s programs/*.gperf . +popd +pushd iconv +ln -s ../locale/programs/charmap-kw.gperf . +popd + +%if %{with docs} +# Remove the `dir' info-heirarchy file which will be maintained +# by the system as it adds info files to the install. +rm -f %{glibc_sysroot}%{_infodir}/dir +%endif + +# Move libpcprofile.so and libmemusage.so into the proper library directory. +# They can be moved without any real consequences because users would not use +# them directly. +mkdir -p %{glibc_sysroot}%{_libdir} +mv -f %{glibc_sysroot}/%{_lib}/lib{pcprofile,memusage}.so \ + %{glibc_sysroot}%{_libdir} + +# Disallow linking against libc_malloc_debug. +rm %{glibc_sysroot}%{_libdir}/libc_malloc_debug.so + +# Strip all of the installed object files. +strip -g %{glibc_sysroot}%{_libdir}/*.o + +# The xtrace and memusage scripts have hard-coded paths that need to be +# translated to a correct set of paths using the $LIB token which is +# dynamically translated by ld.so as the default lib directory. +for i in %{glibc_sysroot}%{_prefix}/bin/{xtrace,memusage}; do +%if %{with bootstrap} + test -w $i || continue +%endif + sed -e 's~=/%{_lib}/libpcprofile.so~=%{_libdir}/libpcprofile.so~' \ + -e 's~=/%{_lib}/libmemusage.so~=%{_libdir}/libmemusage.so~' \ + -e 's~='\''/\\\$LIB/libpcprofile.so~='\''%{_prefix}/\\$LIB/libpcprofile.so~' \ + -e 's~='\''/\\\$LIB/libmemusage.so~='\''%{_prefix}/\\$LIB/libmemusage.so~' \ + -i $i +done + +############################################################################## +# Build an empty libpthread_nonshared.a for compatiliby with applications +# that have old linker scripts that reference this file. We ship this only +# in compat-libpthread-nonshared sub-package. +############################################################################## +ar cr %{glibc_sysroot}%{_prefix}/%{_lib}/libpthread_nonshared.a + +############################################################################### +# Sysroot package creation. +############################################################################### + +%if %{without bootstrap} +mkdir -p %{glibc_sysroot}/%{sysroot_prefix} +pushd %{glibc_sysroot}/%{sysroot_prefix} +mkdir -p usr/lib usr/lib64 + +cp -a %{glibc_sysroot}/%{_prefix}/include usr/. +%ifarch x86_64 +# 32-bit headers for glibc32 don't go in the sysroot. +rm usr/include/gnu/*-32.h +%endif +for lib in lib lib64; do +%ifarch x86_64 + if [ "$lib" = "lib" ]; then + # 32-bit libraries built for glibc32 don't go in the sysroot. + continue + fi +%endif + for pfx in "" %{_prefix}/; do + if test -d %{glibc_sysroot}/$pfx$lib ; then + # Implement UsrMove: everything goes into usr/$lib. Only + # copy files directly in $lib. + find %{glibc_sysroot}/$pfx$lib -maxdepth 1 -type f \ + | xargs -I '{}' cp '{}' usr/$lib/. + # Symbolic links need to be adjusted for UsrMove: They + # need to stay within the same directory. + for sl in `find %{glibc_sysroot}/$pfx$lib -maxdepth 1 -type l`; do + set +x + slbase=$(basename $sl) + sltarget=$(basename $(readlink $sl)) + if test "$sltarget" = . ; then + # This is the lp64d symbolic link on riscv64, see above. + continue + fi + if ! test -r usr/$lib/$sltarget; then + echo "$sl: inferred $sltarget ($(readlink $sl)) missing" + exit 1 + fi + set -x + ln -s $sltarget usr/$lib/$slbase + done + fi + done +done + +# Workaround for the lack of a kernel sysroot package. Copy the +# kernel headers into the sysroot. +rpm -ql kernel-headers | grep "^/usr/include" | while read f ; do + if test -f "$f" ; then + install -D "$f" "./$f" + fi +done + +# Remove the executable bit from files in the sysroot. This prevents +# debuginfo extraction. +find -type f | xargs chmod a-x + +# Use sysroot-relative paths in linker script. Ignore symbolic links. +sed -e 's,\([^0-9a-zA-Z=*]/lib\),/usr/lib,g' \ + -e 's,\([^0-9a-zA-Z=*]\)/,\1/,g' \ + -i $(find -type f -name 'lib[cm].so') + +popd +%dnl %%{without bootstrap} +%endif + +############################################################################## +# Beyond this point in the install process we no longer modify the set of +# installed files. +############################################################################## + +############################################################################## +# Build the file lists used for describing the package and subpackages. +############################################################################## +# There are several main file lists (and many more for +# the langpack sub-packages (langpack-${lang}.filelist)): +# * master.filelist +# - Master file list from which all other lists are built. +# * glibc.filelist +# - Files for the glibc packages. +# * common.filelist +# - Flies for the common subpackage. +# * utils.filelist +# - Files for the utils subpackage. +# * devel.filelist +# - Files for the devel subpackage. +# * doc.filelist +# - Files for the documentation subpackage. +# * headers.filelist +# - Files for the headers subpackage. +# * static.filelist +# - Files for the static subpackage. +# * libnsl.filelist +# - Files for the libnsl subpackage +# * nss_db.filelist +# * nss_hesiod.filelist +# - File lists for nss_* NSS module subpackages. +# * nss-devel.filelist +# - File list with the .so symbolic links for NSS packages. +# * compat-libpthread-nonshared.filelist. +# - File list for compat-libpthread-nonshared subpackage. +# * glibc32.filelist +# - Files for the glibc32 packages. + +# Create the main file lists. This way we can append to any one of them later +# wihtout having to create it. Note these are removed at the start of the +# install phase. +touch master.filelist +touch glibc.filelist +touch common.filelist +touch utils.filelist +touch gconv.filelist +touch devel.filelist +touch doc.filelist +touch headers.filelist +touch static.filelist +touch libnsl.filelist +touch nss_db.filelist +touch nss_hesiod.filelist +touch nss-devel.filelist +touch compat-libpthread-nonshared.filelist +%ifarch x86_64 +touch glibc32.filelist +%endif + +############################################################################### +# Master file list, excluding a few things. +############################################################################### +{ + # List all files or links that we have created during install. + # Files with 'etc' are configuration files, likewise 'gconv-modules' + # and 'gconv-modules.cache' are caches, and we exclude them. + find %{glibc_sysroot} \( -type f -o -type l \) \ + \( \ + -name etc -printf "%%%%config " -o \ + -name gconv-modules.cache \ + -printf "%%%%verify(not md5 size mtime) " -o \ + -name gconv-modules* \ + -printf "%%%%verify(not md5 size mtime) %%%%config(noreplace) " \ + , \ + ! -path "*/lib/debug/*" -printf "/%%P\n" \) + # List all directories with a %%dir prefix. We omit the info directory and + # all directories in (and including) /usr/share/locale. + find %{glibc_sysroot} -type d \ + \( -path '*%{_prefix}/share/locale' -prune -o \ + \( -path '*%{_prefix}/share/*' \ +%if %{with docs} + ! -path '*%{_infodir}' -o \ +%endif + -path "*%{_prefix}/include/*" \ + \) -printf "%%%%dir /%%P\n" \) +} | { + # Also remove the *.mo entries. We will add them to the + # language specific sub-packages. + # libnss_ files go into subpackages related to NSS modules. + # and .*/share/i18n/charmaps/.*), they go into the sub-package + # "locale-source". /sys-root/ files are put into the sysroot package. + sed -e '\,.*/share/locale/\([^/_]\+\).*/LC_MESSAGES/.*\.mo,d' \ + -e '\,.*/share/i18n/locales/.*,d' \ + -e '\,.*/share/i18n/charmaps/.*,d' \ + -e '\,.*/etc/\(localtime\|nsswitch.conf\|ld\.so\.conf\|ld\.so\.cache\|default\|rpc\|gai\.conf\),d' \ + -e '\,.*/%{_libdir}/lib\(pcprofile\|memusage\)\.so,d' \ + -e '\,.*/bin/\(memusage\|mtrace\|xtrace\|pcprofiledump\),d' \ + -e '\,.*/sys-root,d' +} | sort > master.filelist + +# The master file list is now used by each subpackage to list their own +# files. We go through each package and subpackage now and create their lists. +# Each subpackage picks the files from the master list that they need. +# The order of the subpackage list generation does not matter. + +# Make the master file list read-only after this point to avoid accidental +# modification. +chmod 0444 master.filelist + +############################################################################### +# glibc +############################################################################### + +# Add all files with the following exceptions: +# - The info files '%{_infodir}/dir' +# - The partial (lib*_p.a) static libraries, include files. +# - The static files, objects, and unversioned DSOs. +# - The bin, locale, some sbin, and share. +# - We want iconvconfig in the main package and we do this by using +# a double negation of -v and [^i] so it removes all files in +# sbin *but* iconvconfig. +# - All the libnss files (we add back the ones we want later). +# - libc_malloc_debug.so, since it is a debugging library. +# - All bench test binaries. +# - The aux-cache, since it's handled specially in the files section. +# - Extra gconv modules. We add the required modules later. +cat master.filelist \ + | grep -v \ + -e '%{_infodir}' \ + -e '%{_libdir}/lib.*_p.a' \ + -e '%{_prefix}/include' \ + -e '%{_libdir}/lib.*\.a' \ + -e '%{_libdir}/.*\.o' \ + -e '%{_libdir}/lib.*\.so' \ + -e '%{_libdir}/gconv/.*\.so$' \ + -e '%{_libdir}/gconv/gconv-modules.d/gconv-modules-extra\.conf$' \ + -e '%{_prefix}/bin' \ + -e '%{_prefix}/lib/locale' \ + -e '%{_prefix}/sbin/[^i]' \ + -e '%{_prefix}/share' \ + -e '/var/db/Makefile' \ + -e '/libnss_.*\.so[0-9.]*$' \ + -e '/libnsl' \ +%if %{glibc_ship_tracelibs_in_utils} + -e 'libc_malloc_debug\.so' \ +%endif + -e 'glibc-benchtests' \ + -e 'aux-cache' \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + -e '/lib/.*\.o' \ + -e '/lib/.*\.a' \ + -e '/lib/.*\.so.*' \ +%endif + > glibc.filelist + +# Add specific files: +# - The nss_files, nss_compat, and nss_db files. +# - The libmemusage.so and libpcprofile.so used by utils. +for module in compat files dns; do + cat master.filelist \ + | grep -E \ + -e "/libnss_$module(\.so\.[0-9.]+|-[0-9.]+\.so)$" \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + | grep -v -e /lib/lib \ +%endif + >> glibc.filelist +done +grep \ +%if ! %{glibc_ship_tracelibs_in_utils} + -e "libmemusage.so" \ +%endif + -e "libpcprofile.so" \ + master.filelist \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + | grep -v -e /lib/lib \ +%endif + >> glibc.filelist + +############################################################################### +# glibc-gconv-extra +############################################################################### + +grep -e "gconv-modules-extra.conf" master.filelist > gconv.filelist + +# Put the essential gconv modules into the main package. +GconvBaseModules="ISO8859-15 ISO8859-1 CP1252" +GconvBaseModules="$GconvBaseModules UNICODE UTF-16 UTF-32 UTF-7" +%ifarch s390 s390x +GconvBaseModules="$GconvBaseModules ISO-8859-1_CP037_Z900 UTF8_UTF16_Z9" +GconvBaseModules="$GconvBaseModules UTF16_UTF32_Z9 UTF8_UTF32_Z9" +%endif +GconvAllModules=$(cat master.filelist | + sed -n 's|%{_libdir}/gconv/\(.*\)\.so|\1|p') + +# Put the base modules into glibc and the rest into glibc-gconv-extra +for conv in $GconvAllModules; do + if echo $GconvBaseModules | grep -q $conv; then + grep -E -e "%{_libdir}/gconv/$conv.so$" \ + master.filelist >> glibc.filelist + else + grep -E -e "%{_libdir}/gconv/$conv.so$" \ + master.filelist >> gconv.filelist + fi +done + +############################################################################### +# glibc-devel +############################################################################### + +# Static libraries that land in glibc-devel, not glibc-static. +devel_static_library_pattern='/lib\(\(c\|nldbl\|mvec\)_nonshared\|g\|ieee\|mcheck\|pthread\|dl\|rt\|util\|anl\)\.a$' +# Static libraries neither in glibc-devel nor in glibc-static. +other_static_library_pattern='/libpthread_nonshared\.a' + +grep '%{_libdir}/lib.*\.a' master.filelist \ + | grep "$devel_static_library_pattern" \ + | grep -v "$other_static_library_pattern" \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + | grep -v '/lib/.*\.a' \ +%endif + > devel.filelist + +# Put all of the object files and *.so (not the versioned ones) into the +# devel package. +grep '%{_libdir}/.*\.o' < master.filelist >> devel.filelist +grep '%{_libdir}/lib.*\.so' < master.filelist >> devel.filelist + +# The exceptions are: +# - libmemusage.so and libpcprofile.so in glibc used by utils. +# - libnss_*.so which are in nss-devel. +sed -i -e '\,libmemusage.so,d' \ + -e '\,libpcprofile.so,d' \ + -e '\,/libnss_[a-z]*\.so$,d' \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + -e '\,/lib/.*\.so$,d' \ +%endif + devel.filelist + +%if %{glibc_autorequires} +mkdir -p %{glibc_sysroot}/%{_rpmconfigdir} %{glibc_sysroot}/%{_fileattrsdir} +sed < %{SOURCE3} \ + -e s/@VERSION@/%{version}/ \ + -e s/@RELEASE@/%{baserelease}/ \ + -e s/@SYMVER@/%{glibc_autorequires_symver}/ \ + > %{glibc_sysroot}/%{_rpmconfigdir}/glibc.req +cp %{SOURCE4} %{glibc_sysroot}/%{_fileattrsdir}/glibc.attr +%endif + +############################################################################### +# glibc-doc +############################################################################### + +%if %{with docs} +# Put the info files into the doc file list, but exclude the generated dir. +grep '%{_infodir}' master.filelist | grep -v '%{_infodir}/dir' > doc.filelist +grep '%{_docdir}' master.filelist >> doc.filelist +%endif + +############################################################################### +# glibc-headers +############################################################################### + +%if %{need_headers_package} +# The glibc-headers package includes only common files which are identical +# across all multilib packages. We must keep gnu/stubs.h and gnu/lib-names.h +# in the glibc-headers package, but the -32, -64, -64-v1, and -64-v2 versions +# go into glibc-devel. +%ifarch x86_64 +# Hardcode 32-bit and 64-bit header names here, so that 32-bit headers end up +# in glibc32. +grep '%{_prefix}/include/gnu/stubs-64\.h$' < master.filelist >> devel.filelist || : +grep '%{_prefix}/include/gnu/lib-names-64\.h$' < master.filelist >> devel.filelist || : +grep '%{_prefix}/include/gnu/stubs-32\.h$' < master.filelist >> glibc32.filelist || : +grep '%{_prefix}/include/gnu/lib-names-32\.h$' < master.filelist >> glibc32.filelist || : +%else +grep '%{_prefix}/include/gnu/stubs-.*\.h$' < master.filelist >> devel.filelist || : +grep '%{_prefix}/include/gnu/lib-names-.*\.h$' < master.filelist >> devel.filelist || : +%endif +# Put the include files into headers file list. +grep '%{_prefix}/include' < master.filelist \ + | grep -E -v '%{_prefix}/include/gnu/stubs-.*\.h$' \ + | grep -E -v '%{_prefix}/include/gnu/lib-names-.*\.h$' \ + > headers.filelist +%else +# If there is no glibc-headers package, all header files go into the +# glibc-devel package. +grep '%{_prefix}/include' < master.filelist >> devel.filelist +%endif + +############################################################################### +# glibc-static +############################################################################### + +# Put the rest of the static files into the static package. +grep '%{_libdir}/lib.*\.a' < master.filelist \ + | grep -v "$devel_static_library_pattern" \ + | grep -v "$other_static_library_pattern" \ +%ifarch x86_64 +%dnl Exclude 32-bit libraries built for glibc32. + | grep -v '/lib/.*\.a' \ +%endif + > static.filelist + +############################################################################### +# glibc-common +############################################################################### + +# All of the bin and certain sbin files go into the common package except +# iconvconfig which needs to go in glibc. The iconvconfig binary is kept in +# the main glibc package because we use it in the post-install scriptlet to +# rebuild the gconv-modules.cache. The makedb binary is in nss_db. +grep '%{_prefix}/bin' master.filelist \ + | grep -v '%{_prefix}/bin/makedb' \ + >> common.filelist +grep '%{_prefix}/sbin' master.filelist \ + | grep -v '%{_prefix}/sbin/iconvconfig' >> common.filelist +# All of the files under share go into the common package since they should be +# multilib-independent. +# Exceptions: +# - The actual share directory, not owned by us. +# - The info files which go into doc, and the info directory. +# - All documentation files, which go into doc. +grep '%{_prefix}/share' master.filelist \ + | grep -v \ + -e '%{_prefix}/share/info/libc.info.*' \ + -e '%%dir %{prefix}/share/info' \ + -e '%%dir %{prefix}/share' \ + -e '%{_docdir}' \ + >> common.filelist + +############################################################################### +# glibc-utils +############################################################################### + +# Add the utils scripts and programs to the utils subpackage. +cat > utils.filelist <> utils.filelist +%endif + +############################################################################### +# nss_db, nss_hesiod +############################################################################### + +# Move the NSS-related files to the NSS subpackages. Be careful not +# to pick up .debug files, and the -devel symbolic links. +for module in db hesiod; do + grep -E "/libnss_$module\\.so\\.[0-9.]+\$" \ + master.filelist > nss_$module.filelist +done +grep -E "%{_prefix}/bin/makedb$" master.filelist >> nss_db.filelist + +############################################################################### +# nss-devel +############################################################################### + +# Symlinks go into the nss-devel package (instead of the main devel +# package). +grep '/libnss_[a-z]*\.so$' master.filelist > nss-devel.filelist + +############################################################################### +# libnsl +############################################################################### + +# Prepare the libnsl-related file lists. +grep -E '/libnsl\.so\.[0-9]+$' master.filelist > libnsl.filelist +test $(wc -l < libnsl.filelist) -eq 1 + +%if %{with benchtests} +############################################################################### +# glibc-benchtests +############################################################################### + +# List of benchmarks. +find build-%{target}/benchtests -type f -executable | while read b; do + echo "%{_prefix}/libexec/glibc-benchtests/$(basename $b)" +done >> benchtests.filelist +# ... and the makefile. +for b in %{SOURCE1} %{SOURCE2}; do + echo "%{_prefix}/libexec/glibc-benchtests/$(basename $b)" >> benchtests.filelist +done +# ... and finally, the comparison scripts. +echo "%{_prefix}/libexec/glibc-benchtests/benchout.schema.json" >> benchtests.filelist +echo "%{_prefix}/libexec/glibc-benchtests/compare_bench.py*" >> benchtests.filelist +echo "%{_prefix}/libexec/glibc-benchtests/import_bench.py*" >> benchtests.filelist +echo "%{_prefix}/libexec/glibc-benchtests/validate_benchout.py*" >> benchtests.filelist +%endif + +############################################################################### +# compat-libpthread-nonshared +############################################################################### +echo "%{_libdir}/libpthread_nonshared.a" >> compat-libpthread-nonshared.filelist + +############################################################################### +# glibc32 +############################################################################### +%ifarch x86_64 +grep '/lib/.*\.o$' master.filelist >> glibc32.filelist +grep '/lib/.*\.a$' master.filelist >> glibc32.filelist +grep '/lib/.*\.so' master.filelist >> glibc32.filelist +%endif + +############################################################################## +# Run the glibc testsuite +############################################################################## +%check +%if %{with testsuite} + +# Run the glibc tests. If any tests fail to build we exit %check with +# an error, otherwise we print the test failure list and the failed +# test output and continue. Write to standard error to avoid +# synchronization issues with make and shell tracing output if +# standard output and standard error are different pipes. +run_tests () { + # This hides a test suite build failure, which should be fatal. We + # check "Summary of test results:" below to verify that all tests + # were built and run. + %make_build check |& tee rpmbuild.check.log >&2 + test -n tests.sum + if ! grep -Eq '^\s+=== Summary of results ===$' rpmbuild.check.log ; then + echo "FAIL: test suite build of target: $(basename "$(pwd)")" >& 2 + exit 1 + fi + set +x + grep -v ^PASS: tests.sum > rpmbuild.tests.sum.not-passing || true + if test -n rpmbuild.tests.sum.not-passing ; then + echo ===================FAILED TESTS===================== >&2 + echo "Target: $(basename "$(pwd)")" >& 2 + cat rpmbuild.tests.sum.not-passing >&2 + while read failed_code failed_test ; do + for suffix in out test-result ; do + if test -e "$failed_test.$suffix"; then + echo >&2 + echo "=====$failed_code $failed_test.$suffix=====" >&2 + cat -- "$failed_test.$suffix" >&2 + echo >&2 + fi + done + done &2 + cat misc/tst-syscall-list.out >&2 + set -x +} + +# Increase timeouts +export TIMEOUTFACTOR=16 +parent=$$ +echo ====================TESTING========================= + +# Default libraries. +pushd build-%{target} +run_tests +popd + +%if %{buildpower10} +# Run this test only if the server supports Power10 instructions. +if LD_SHOW_AUXV=1 /bin/true | grep -E "AT_HWCAP2:[^$]*arch_3_1" > /dev/null; then + echo ====================TESTING -mcpu=power10============= + pushd build-%{target}-power10 + run_tests + popd +fi +%endif + +echo ====================TESTING END===================== +PLTCMD='/^Relocation section .*\(\.rela\?\.plt\|\.rela\.IA_64\.pltoff\)/,/^$/p' +echo ====================PLT RELOCS LD.SO================ +readelf -Wr %{glibc_sysroot}/%{_lib}/ld-*.so | sed -n -e "$PLTCMD" +echo ====================PLT RELOCS LIBC.SO============== +readelf -Wr %{glibc_sysroot}/%{_lib}/libc-*.so | sed -n -e "$PLTCMD" +echo ====================PLT RELOCS END================== + +run_ldso="%{glibc_sysroot}/%{glibc_ldso} --library-path %{glibc_sysroot}/%{_lib}" + +# Show the auxiliary vector as seen by the new library +# (even if we do not perform the valgrind test). +LD_SHOW_AUXV=1 $run_ldso /bin/true + +%if 0%{?_enable_debug_packages} +# Finally, check if valgrind runs with the new glibc. +# We want to fail building if valgrind is not able to run with this glibc so +# that we can then coordinate with valgrind to get it fixed before we update +# glibc. +%if %{with valgrind} +$run_ldso /usr/bin/valgrind --error-exitcode=1 \ + $run_ldso /usr/bin/true +# true --help performs some memory allocations. +$run_ldso /usr/bin/valgrind --error-exitcode=1 \ + $run_ldso /usr/bin/true --help >/dev/null +%endif +%endif + +%endif + + +%pre -p +-- Check that the running kernel is new enough +required = '%{enablekernel}' +rel = posix.uname("%r") +if rpm.vercmp(rel, required) < 0 then + error("FATAL: kernel too old", 0) +end + +-- (1) Remove multilib libraries from previous installs. +-- In order to support in-place upgrades, we must immediately remove +-- all platform directories before installing a new glibc +-- version. RPM only deletes files removed by updates near the end +-- of the transaction. If we did not remove all platform +-- directories here, they may be preferred by the dynamic linker +-- during the execution of subsequent RPM scriptlets, likely +-- resulting in process startup failures. + +-- Full set of libraries glibc may install. +install_libs = { "anl", "BrokenLocale", "c", "dl", "m", "mvec", + "nss_compat", "nss_db", "nss_dns", "nss_files", + "nss_hesiod", "pthread", "resolv", "rt", "SegFault", + "thread_db", "util" } + +-- We are going to remove these libraries. Generally speaking we remove +-- all core libraries in the multilib directory. +-- For the versioned install names, the version are [2.0,9.9*], so we +-- match "libc-2.0.so" and so on up to "libc-9.9*". +-- For the unversioned install names, we match the library plus ".so." +-- followed by digests. +remove_regexps = {} +for i = 1, #install_libs do + -- Versioned install name. + remove_regexps[#remove_regexps + 1] = ("lib" .. install_libs[i] + .. "%%-[2-9]%%.[0-9]+%%.so$") + -- Unversioned install name. + remove_regexps[#remove_regexps + 1] = ("lib" .. install_libs[i] + .. "%%.so%%.[0-9]+$") +end + +-- Two exceptions: +remove_regexps[#install_libs + 1] = "libthread_db%%-1%%.0%%.so" +remove_regexps[#install_libs + 2] = "libSegFault%%.so" + +-- We are going to search these directories. +local remove_dirs = { "%{_libdir}/i686", + "%{_libdir}/i686/nosegneg", + "%{_libdir}/power6", + "%{_libdir}/power7", + "%{_libdir}/power8", + "%{_libdir}/power9", + } + +-- Add all the subdirectories of the glibc-hwcaps subdirectory. +repeat + local iter = posix.files("%{_libdir}/glibc-hwcaps") + if iter ~= nil then + for entry in iter do + if entry ~= "." and entry ~= ".." then + local path = "%{_libdir}/glibc-hwcaps/" .. entry + if posix.access(path .. "/.", "x") then + remove_dirs[#remove_dirs + 1] = path + end + end + end + end +until true + +-- Walk all the directories with files we need to remove... +for _, rdir in ipairs (remove_dirs) do + if posix.access (rdir) then + -- If the directory exists we look at all the files... + local remove_files = posix.files (rdir) + for rfile in remove_files do + for _, rregexp in ipairs (remove_regexps) do + -- Does it match the regexp? + local dso = string.match (rfile, rregexp) + if (dso ~= nil) then + -- Removing file... + os.remove (rdir .. '/' .. rfile) + end + end + end + end +end + +%post -p +%glibc_post_funcs +-- (1) Update /etc/ld.so.conf +-- Next we update /etc/ld.so.conf to ensure that it starts with +-- a literal "include ld.so.conf.d/*.conf". + +local ldsoconf = "/etc/ld.so.conf" +local ldsoconf_tmp = "/etc/glibc_post_upgrade.ld.so.conf" + +if posix.access (ldsoconf) then + + -- We must have a "include ld.so.conf.d/*.conf" line. + local have_include = false + for line in io.lines (ldsoconf) do + -- This must match, and we don't ignore whitespace. + if string.match (line, "^include ld.so.conf.d/%%*%%.conf$") ~= nil then + have_include = true + end + end + + if not have_include then + -- Insert "include ld.so.conf.d/*.conf" line at the start of the + -- file. We only support one of these post upgrades running at + -- a time (temporary file name is fixed). + local tmp_fd = io.open (ldsoconf_tmp, "w") + if tmp_fd ~= nil then + tmp_fd:write ("include ld.so.conf.d/*.conf\n") + for line in io.lines (ldsoconf) do + tmp_fd:write (line .. "\n") + end + tmp_fd:close () + local res = os.rename (ldsoconf_tmp, ldsoconf) + if res == nil then + io.stdout:write ("Error: Unable to update configuration file (rename).\n") + end + else + io.stdout:write ("Error: Unable to update configuration file (open).\n") + end + end +end + +-- (2) Rebuild ld.so.cache early. +-- If the format of the cache changes then we need to rebuild +-- the cache early to avoid any problems running binaries with +-- the new glibc. + +call_ldconfig() + +-- (3) Update gconv modules cache. +-- If the /usr/lib/gconv/gconv-modules.cache exists, then update it +-- with the latest set of modules that were just installed. +-- We assume that the cache is in _libdir/gconv and called +-- "gconv-modules.cache". + +update_gconv_modules_cache() + +-- (4) On upgrades, restart systemd if installed. "systemctl -q" does +-- not suppress the error message (which is common in chroots), so +-- open-code rpm.execute with standard error suppressed. +if tonumber(arg[2]) >= 2 + and posix.access("%{_prefix}/bin/systemctl", "x") +then + local pid = posix.fork() + if pid == 0 then + posix.redirect2null(2) + posix.exec("%{_prefix}/bin/systemctl", "daemon-reexec") + elseif pid > 0 then + posix.wait(pid) + end +end + +%posttrans all-langpacks -e -p +-- The old glibc-all-langpacks postun scriptlet deleted the locale-archive +-- file, so we may have to resurrect it on upgrades. +local archive_path = "%{_prefix}/lib/locale/locale-archive" +local real_path = "%{_prefix}/lib/locale/locale-archive.real" +local stat_archive = posix.stat(archive_path) +local stat_real = posix.stat(real_path) +-- If the hard link was removed, restore it. +if stat_archive ~= nil and stat_real ~= nil + and (stat_archive.ino ~= stat_real.ino + or stat_archive.dev ~= stat_real.dev) then + posix.unlink(archive_path) + stat_archive = nil +end +-- If the file is gone, restore it. +if stat_archive == nil then + posix.link(real_path, archive_path) +end +-- Remove .rpmsave file potentially created due to config file change. +local save_path = archive_path .. ".rpmsave" +if posix.access(save_path) then + posix.unlink(save_path) +end + +%post gconv-extra -p +%glibc_post_funcs +update_gconv_modules_cache () + +%postun gconv-extra -p +%glibc_post_funcs +update_gconv_modules_cache () + +%files -f glibc.filelist +%dir %{_prefix}/%{_lib}/audit +%if %{buildpower10} +%dir /%{_libdir}/glibc-hwcaps/power10 +%endif +%verify(not md5 size mtime) %config(noreplace) /etc/ld.so.conf +%verify(not md5 size mtime) %config(noreplace) /etc/rpc +%dir /etc/ld.so.conf.d +%dir %{_prefix}/libexec/getconf +%dir %{_libdir}/gconv +%dir %{_libdir}/gconv/gconv-modules.d +%dir %attr(0700,root,root) /var/cache/ldconfig +%attr(0600,root,root) %verify(not md5 size mtime) %ghost %config(missingok,noreplace) /var/cache/ldconfig/aux-cache +%attr(0644,root,root) %verify(not md5 size mtime) %ghost %config(missingok,noreplace) /etc/ld.so.cache +%attr(0644,root,root) %verify(not md5 size mtime) %ghost %config(missingok,noreplace) /etc/gai.conf +# If rpm doesn't support %license, then use %doc instead. +%{!?_licensedir:%global license %%doc} +%license COPYING COPYING.LIB LICENSES + +%files -f common.filelist common +%dir %{_prefix}/lib/locale +%dir %{_prefix}/lib/locale/C.utf8 +%{_prefix}/lib/locale/C.utf8/* + +%files all-langpacks +%{_prefix}/lib/locale/locale-archive +%{_prefix}/lib/locale/locale-archive.real +%{_prefix}/share/locale/*/LC_MESSAGES/libc.mo + +%files locale-source +%dir %{_prefix}/share/i18n/locales +%{_prefix}/share/i18n/locales/* +%dir %{_prefix}/share/i18n/charmaps +%{_prefix}/share/i18n/charmaps/* + +%files -f devel.filelist devel +%if %{glibc_autorequires} +%attr(0755,root,root) %{_rpmconfigdir}/glibc.req +%{_fileattrsdir}/glibc.attr +%endif + +%if %{with docs} +%files -f doc.filelist doc +%endif + +%files -f static.filelist static + +%if %{need_headers_package} +%files -f headers.filelist -n %{headers_package_name} +%endif + +%files -f utils.filelist utils + +%files -f gconv.filelist gconv-extra + +%files -f nss_db.filelist -n nss_db +/var/db/Makefile +%files -f nss_hesiod.filelist -n nss_hesiod +%doc hesiod/README.hesiod +%files -f nss-devel.filelist nss-devel + +%files -f libnsl.filelist -n libnsl +/%{_lib}/libnsl.so.1 + +%if %{with benchtests} +%files benchtests -f benchtests.filelist +%endif + +%files -f compat-libpthread-nonshared.filelist -n compat-libpthread-nonshared + +%if %{without bootstrap} +%files -n sysroot-%{_arch}-%{sysroot_dist}-glibc +%{sysroot_prefix} +%endif + +%ifarch x86_64 +%files -f glibc32.filelist -n glibc32 +%endif + +%changelog +* Tue Nov 26 2024 MSVSphere Packaging Team - 2.39-22 +- Rebuilt for MSVSphere 10 + +* Wed Jul 31 2024 Florian Weimer - 2.39-22 +- Support clearing options in /etc/resolv.conf, RES_OPTIONS with a - prefix +- Introduce the strict-error/RES_STRICTERR stub resolver option (RHEL-18039) + +* Wed Jul 31 2024 Arjun Shankar - 2.39-21 +- Sync with upstream branch release/2.39/master, + commit 4bdcc1963bc2b5ba5f8e319e402d9eb2cb6096c1: +- manual: make setrlimit() description less ambiguous +- manual/stdio: Clarify putc and putwc +- malloc: add multi-threaded tests for aligned_alloc/calloc/malloc +- malloc: avoid global locks in tst-aligned_alloc-lib.c +- resolv: Track single-request fallback via _res._flags (bug 31476) +- resolv: Do not wait for non-existing second DNS response after error (bug 30081) +- resolv: Allow short error responses to match any query (bug 31890) +- elf: Fix localplt.awk for DT_RELR-enabled builds (BZ 31978) +- Fix usage of _STACK_GROWS_DOWN and _STACK_GROWS_UP defines [BZ 31989] +- Linux: Make __rseq_size useful for feature detection (bug 31965) +- elf: Make dl-rseq-symbols Linux only +- nptl: fix potential merge of __rseq_* relro symbols +- s390x: Fix segfault in wcsncmp [BZ #31934] +- stdlib: fix arc4random fallback to /dev/urandom (BZ 31612) +- math: Provide missing math symbols on libc.a (BZ 31781) +- math: Fix isnanf128 static build (BZ 31774) +- math: Fix i386 and m68k exp10 on static build (BZ 31775) +- math: Fix i386 and m68k fmod/fmodf on static build (BZ 31488) +- posix: Fix pidfd_spawn/pidfd_spawnp leak if execve fails (BZ 31695) + +* Mon Jul 22 2024 Florian Weimer - 2.39-20 +- Support --without testsuite builds without perl installed (#2292195) + +* Fri Jul 19 2024 Florian Weimer - 2.39-19 +- Add Conflicts:/Obsoletes: for glibc32 to glibc.i686 + +* Thu Jul 18 2024 Arjun Shankar - 2.39-18 +- ppc64le: Build early startup code with -mcpu=power8 + +* Tue Jul 02 2024 Patsy Griffin - 2.39-17 +- Move ANSI_X3.110-1983 support from main package to glibc-gconv-extra. + +* Fri Jun 21 2024 Arjun Shankar - 2.39-16 +- Sync with upstream branch release/2.39/master, + commit 7f9f25f255ee2c00178779fbce502f4b94b848b9: +- Linux: Include in dl-sysdep.c +- x86: Properly set x86 minimum ISA level [BZ #31883] +- x86: Properly set MINIMUM_X86_ISA_LEVEL for i386 [BZ #31867] +- localedata: ssy_ER: Fix syntax error +- malloc: New test to check malloc alternate path using memory obstruction +- malloc: Improve aligned_alloc and calloc test coverage. +- malloc/Makefile: Split and sort tests +- x86/cet: fix shadow stack test scripts + +* Wed Jun 05 2024 Arjun Shankar - 2.39-15 +- Sync with upstream branch release/2.39/master, + commit 6ade91c21140d8c803c289932dbfc74537f65a1f: +- elf: Avoid some free (NULL) calls in _dl_update_slotinfo +- misc: Add support for Linux uio.h RWF_NOAPPEND flag +- i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782) +- Reinstate generic features-time64.h +- Always define __USE_TIME_BITS64 when 64 bit time_t is used +- socket: Use may_alias on sockaddr structs (bug 19622) +- parse_fdinfo: Don't advance pointer twice [BZ #31798] +- LoongArch: Fix undefined `__memset_aligned` reference in ld.so linking. +- socket: Add new test for connect +- libsupport: Add xgetpeername +- x86_64: Fix missing wcsncat function definition without multiarch (x86-64-v4) + +* Fri May 31 2024 Florian Weimer - 2.39-14 +- Enable CPU compatibility diagnostics in ld.so (#2276631, RHEL-31738) + +* Wed May 15 2024 Arjun Shankar - 2.39-13 +- Move memory tracing libraries to glibc-utils + +* Fri May 10 2024 Florian Weimer - 2.39-12 +- Use unsigned types in / (RHEL-22226) + +* Fri May 10 2024 Florian Weimer - 2.39-11 +- Sync with upstream branch release/2.39/master, + commit 97bb89668d7171164975f3dc895e38343a2f3a95: +- Force DT_RPATH for --enable-hardcoded-path-in-tests +- elf: Only process multiple tunable once (BZ 31686) +- Add a test to check for duplicate definitions in the static library +- i686: Fix multiple definitions of __memmove_chk and __memset_chk +- i586: Fix multiple definitions of __memcpy_chk and __mempcpy_chk +- time: Allow later version licensing. +- nscd: Use time_t for return type of addgetnetgrentX +- login: structs utmp, utmpx, lastlog _TIME_BITS independence (bug 30701) +- login: Check default sizes of structs utmp, utmpx, lastlog + +* Fri May 03 2024 Florian Weimer - 2.39-10 +- Build POWER10 multilib + +* Fri Apr 26 2024 Florian Weimer - 2.39-9 +- nscd is currently not build, so the security fixes below are not relevant. +- Sync with upstream branch release/2.39/master, + commit fd658f026f25cf59e8db243bc3b3e09cd5a20ba0: +- elf: Also compile dl-misc.os with $(rtld-early-cflags) +- CVE-2024-33601, CVE-2024-33602: nscd: netgroup: Use two buffers in addgetnetgrentX (bug 31680) +- CVE-2024-33600: nscd: Avoid null pointer crashes after notfound response (bug 31678) +- CVE-2024-33600: nscd: Do not send missing not-found response in addgetnetgrentX (bug 31678) +- CVE-2024-33599: nscd: Stack-based buffer overflow in netgroup cache (bug 31677) +- x86: Define MINIMUM_X86_ISA_LEVEL in config.h [BZ #31676] +- i386: ulp update for SSE2 --disable-multi-arch configurations +- nptl: Fix tst-cancel30 on kernels without ppoll_time64 support + +* Thu Apr 18 2024 Florian Weimer - 2.39-8 +- Sync with upstream branch release/2.39/master, + commit 31da30f23cddd36db29d5b6a1c7619361b271fb4: +- iconv: ISO-2022-CN-EXT: fix out-of-bound writes when writing escape sequence (CVE-2024-2961) +- x86_64: Exclude SSE, AVX and FMA4 variants in libm multiarch +- Apply the Makefile sorting fix +- powerpc: Fix ld.so address determination for PCREL mode (bug 31640) +- x86-64: Simplify minimum ISA check ifdef conditional with if +- x86-64: Don't use SSE resolvers for ISA level 3 or above +- AArch64: Check kernel version for SVE ifuncs +- aarch64: fix check for SVE support in assembler +- aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR +- i386: Use generic memrchr in libc (bug 31316) + +* Thu Apr 04 2024 Arjun Shankar - 2.39-7 +- Sync with upstream branch release/2.39/master, + commit 5d070d12b3a52bc44dd1b71743abc4b6243862ae: +- x86: Expand the comment on when REP STOSB is used on memset +- x86: Do not prefer ERMS for memset on Zen3+ +- x86: Fix Zen3/Zen4 ERMS selection (BZ 30994) +- Add tst-gnu2-tls2mod1 to test-internal-extras +- elf: Enable TLS descriptor tests on aarch64 +- arm: Update _dl_tlsdesc_dynamic to preserve caller-saved registers (BZ 31372) +- Ignore undefined symbols for -mtls-dialect=gnu2 +- x86-64: Allocate state buffer space for RDI, RSI and RBX +- x86-64: Update _dl_tlsdesc_dynamic to preserve AMX registers +- x86: Update _dl_tlsdesc_dynamic to preserve caller-saved registers +- x86-64: Save APX registers in ld.so trampoline +- LoongArch: Correct {__ieee754, _}_scalb -> {__ieee754, _}_scalbf +- powerpc: Placeholder and infrastructure/build support to add Power11 related changes. +- powerpc: Add HWCAP3/HWCAP4 data to TCB for Power Architecture. + +* Tue Mar 26 2024 Florian Weimer - 2.39-6 +- Do not generate ELF dependency information for glibc32 + +* Tue Mar 26 2024 Joseph Myers - 2.39-5 +- Build glibc32 binary package from glibc sources as part of x86_64 build, + not from glibc32 SRPM that contains binaries from i686 RPM build. + +* Tue Mar 19 2024 Arjun Shankar - 2.39-4 +- Sync with upstream branch release/2.39/master, + commit: 71fcdba577884627c3ee4e43beb915da752efb1f: +- linux: Use rseq area unconditionally in sched_getcpu (bug 31479) +- Use gcc __builtin_stdc_* builtins in stdbit.h if possible +- S390: Do not clobber r7 in clone [BZ #31402] +- math: Update mips64 ulps +- mips: FIx clone3 implementation (BZ 31325) +- arm: Remove wrong ldr from _dl_start_user (BZ 31339) +- Replace advisories directory + +* Wed Feb 28 2024 Arjun Shankar - 2.39-3 +- Analyse glibc-2.39 sources for license information +- Migrate License field to SPDX identifiers for + https://docs.fedoraproject.org/en-US/legal/allowed-licenses/ + https://docs.fedoraproject.org/en-US/legal/update-existing-packages + (#2222074) + +* Wed Feb 7 2024 Florian Weimer - 2.39-2 +- Ignore symbolic links to . in sysroot construction + +* Fri Feb 02 2024 Carlos O'Donell - 2.39-1 +- Switch to upstream 2.39 release, + commit ef321e23c20eebc6d6fb4044425c00e6df27b05f +- Document CVE-2023-6246, CVE-2023-6779, and CVE-2023-6780 +- Update advisory format and introduce some automation +- manual/io: Fix swapped reading and writing phrase. +- Fix typo +- S390: Fix building with --disable-mutli-arch [BZ #31196] +- NEWS: insert advisories and fixed bugs for 2.39 +- contrib.texi: update +- INSTALL, install.texi: minor updates, regenerate +- libc.pot: regenerate +- version.h, include/features.h: Bump version to 2.39 +- Create ChangeLog.old/ChangeLog.28 + +* Wed Jan 31 2024 Florian Weimer - 2.38.9000-39 +- Add noarch sysroot subpackages + +* Tue Jan 30 2024 Patsy Griffin - 2.38.9000-38 +- Auto-sync with upstream branch master, + commit ddf542da94caf97ff43cc2875c88749880b7259b: +- syslog: Fix integer overflow in __vsyslog_internal (CVE-2023-6780) +- syslog: Fix heap buffer overflow in __vsyslog_internal (CVE-2023-6779) +- syslog: Fix heap buffer overflow in __vsyslog_internal (CVE-2023-6246) +- Use binutils 2.42 branch in build-many-glibcs.py +- elf: correct relocation statistics for !ELF_MACHINE_START_ADDRESS + +* Mon Jan 29 2024 Arjun Shankar - 2.38.9000-37 +- Auto-sync with upstream branch master, + commit ae49a7b29acc184b03c2a6bd6ac01b5e08efd54f: +- Relicense IBM portions of resolv/base64.c resolv/res_debug.c. +- localedata: Use consistent values for grouping and mon_grouping +- manual: fix order of arguments of memalign and aligned_alloc (Bug 27547) + +* Thu Jan 25 2024 Jens Petersen - 2.38.9000-36 +- no longer supplement langpacks if all-langpacks installed + +* Wed Jan 24 2024 Florian Weimer - 2.38.9000-35 +- Add crh_RU, gbm_IN, ssy_ER locales. +- Auto-sync with upstream branch master, + commit 486452affbac684db739b7fcca1e84e8a7ce33d1: +- manual, NEWS: Document malloc side effect of dynamic TLS changes +- NEWS: Update temporary files ignored by ldconfig +- po: Incorporate translations (sr) +- string: Disable stack protector for memset in early static initialization +- qsort: Fix a typo causing unnecessary malloc/free (BZ 31276) +- riscv: add support for static PIE +- localedata: renamed: aa_ER@saaho -> ssy_ER +- Define ISO 639-3 "ssy" (Saho) +- localedata: add crh_RU, Crimean Tartar language in the Cyrillic script as used in Russia. +- localedata: tr_TR, ku_TR: Sync with CLDR: “Turkey” -> “Türkiye” +- localedata: miq_NI: Shorten month names in abmon +- Update kernel version to 6.7 in header constant tests +- localedata: add gbm_IN locale +- Define ISO 639-3 "gbm" (Garhwali) +- Update syscall lists for Linux 6.7 +- stdlib: Remove unused is_aligned function from qsort.c + +* Fri Jan 19 2024 Fedora Release Engineering - 2.38.9000-34 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_40_Mass_Rebuild + +* Tue Jan 16 2024 DJ Delorie - 2.38.9000-33 +- Auto-sync with upstream branch master, + commit e2803cfd8bb00b54816d1a2c381d2cf82b136da6. +- NEWS: Mention PLT rewrite on x86-64 +- stdlib: Verify heapsort for two-element cases +- stdlib: Fix heapsort for cases with exactly two elements +- localedata: anp_IN: Fix abbreviated month names + +* Mon Jan 15 2024 DJ Delorie - 2.38.9000-32 +- Auto-sync with upstream branch master, + commit 709fbd3ec3595f2d1076b4fec09a739327459288. +- stdlib: Reinstate stable mergesort implementation on qsort +- x86-64: Check if mprotect works before rewriting PLT +- aarch64: Add NEWS entry about libmvec for 2.39 + +* Mon Jan 15 2024 Carlos O'Donell - 2.38.9000-31 +- Add new locales for kv_RU, su_ID, tok, and zgh_MA. +- Drop glibc-rh2255506.patch; fix applied upstream. +- Enable Intel CET only on x86_64. +- Auto-sync with upstream branch master, + commit 064c708c78cc2a6b5802dce73108fc0c1c6bfc80: +- localedata/unicode-gen/utf8_gen.py: fix Hangul syllable name +- x86_64: Optimize ffsll function code size. +- localedata: Remove redundant comments +- RISC-V: Enable static-pie. +- linux: Fix fstat64 on alpha and sparc64 +- math: remove exp10 wrappers +- Benchtests: Increase benchmark iterations +- debug/getwd_chk.c: warning should be emitted for the __getwd_chk symbol. +- Make __getrandom_nocancel set errno and add a _nostatus version +- x86-64/cet: Make CET feature check specific to Linux/x86 +- Incorporate translations (zh_CN) +- Define ISO 639-3 "glk" (Gilaki) +- resolv: Fix endless loop in __res_context_query +- localedata: revert all the remaining locale sources to UTF-8 +- localedata: am_ET ber_DZ en_GB en_PH en_US fil_PH kab_DZ om_ET om_KE ti_ET tl_PH: convert to UTF-8 +- localedata: resolve cyclic dependencies +- localedata: kv_RU: convert to UTF-8 +- localedata: add new locale kv_RU +- elf: Fix tst-nodeps2 test failure. +- localedata: Sort Makefile variables. +- locale: Sort Makefile variables. +- i386: Remove CET support bits +- x86-64/cet: Move check-cet.awk to x86_64 +- x86-64/cet: Move dl-cet.[ch] to x86_64 directories +- x86: Move x86-64 shadow stack startup codes +- Fix deprecated utcnow() usage in build-many-glibcs.py +- Fix invalid escape sequence in build-many-glibcs.py +- math: Fix test-fenv.c feupdateenv tests +- Remove installed header rule on $(..)include/%.h +- i386: Fail if configured with --enable-cet +- i386: Remove CET support +- x86: Move CET infrastructure to x86_64 +- localedata: su_ID: make lang_name agree with CLDR +- localedata: add new locale su_ID +- localedata: add new locale zgh_MA +- INSTALL: regenerate +- localedata: add tok/UTF-8 to SUPPORTED +- localedata: tok: add yY and nN to yesexpr and noexpr +- localedata: tok: convert to UTF-8 +- localedata: add data for tok (Toki Pona) +- Remove ia64-linux-gnu +- localedata: dz_BT, bo_CN: convert to UTF-8 +- localedata: dz_BT, bo_CN: Fix spelling of "phur bu" in both Tibetan and Dzongkha +- localedata: bo_CN: Fix spelling errors in Tibetan data +- localedata: bo_CN: Fix incomplete edit in Tibetan yesexpr +- localedata: dz_BT: Fix spelling errors in Dzongha data +- localedata: unicode-gen: Remove redundant \s* from regexp, fix comments +- localedata: convert the remaining *_RU locales to UTF-8 +- Incorporate translations +- x32: Handle displacement overflow in PLT rewrite [BZ #31218] +- x86: Fixup some nits in longjmp asm implementation +- stdlib: Fix stdbit.h with -Wconversion for clang +- stdlib: Fix stdbit.h with -Wconversion for older gcc +- elf: Add ELF_DYNAMIC_AFTER_RELOC to rewrite PLT +- aarch64: Make cpu-features definitions not Linux-specific +- hurd: Initializy _dl_pagesize early in static builds +- hurd: Only init early static TLS if it's used to store stack or pointer guards +- hurd: Make init-first.c no longer x86-specific +- hurd: Drop x86-specific assembly from init-first.c +- hurd: Pass the data pointer to _hurd_stack_setup explicitly +- x86-64/cet: Check the restore token in longjmp +- localedata: ru_RU, ru_UA: convert to UTF-8 +- localedata: es_??: convert to UTF-8 +- localedata: miq_NI: convert to UTF-8 +- i386: Ignore --enable-cet +- mach: Drop SNARF_ARGS macro +- mach: Drop some unnecessary vm_param.h includes +- hurd: Declare _hurd_intr_rpc_msg* with protected visibility +- hurd: Add some missing includes +- localedata: fy_DE: make this "Western Frisian" to agree with the language code "fy" +- localedata: fy_DE, fy_NL: convert to UTF-8 +- localedata: ast_ES: convert to UTF-8 +- localedata: ast_ES: Remove wrong copyright text +- localedata: de_{AT,BE,CH,IT,LU}: convert to UTF-8 +- localedata: lv_LV, it_IT, it_CH: convert to UTF-8 +- localedata: it_IT, lv_LV: currency symbol should follow the amount +- Implement C23 +- localedata: ms_MY should not use 12-hour format +- localedata: es_ES: convert to UTF-8 +- localedata: es_ES: Add am_pm strings +- aarch64: Add longjmp test for SME +- aarch64: Add setcontext support for SME +- aarch64: Add longjmp support for SME +- aarch64: Add SME runtime support +- localedata: convert uz_UZ and uz_UZ@cyrillic to UTF-8 +- localedata: uz_UZ and uz_UZ@cyrillic: Fix decimal point and thousands separator +- libio: Check remaining buffer size in _IO_wdo_write (bug 31183) +- getaddrinfo: translate ENOMEM to EAI_MEMORY (bug 31163) +- string: Add additional output in test-strchr failure +- Add a setjmp/longjmp test between user contexts +- x86/cet: Add -fcf-protection=none before -fcf-protection=branch +- Regenerate libc.pot +- Omit regex.c pragmas no longer needed +- Update copyright dates not handled by scripts/update-copyrights +- Update copyright in generated files by running "make" +- Update copyright dates with scripts/update-copyrights +- x86/cet: Run some CET tests with shadow stack +- x86/cet: Don't set CET active by default +- x86/cet: Check feature_1 in TCB for active IBT and SHSTK +- x86/cet: Enable shadow stack during startup +- elf: Always provide _dl_get_dl_main_map in libc.a +- x86/cet: Sync with Linux kernel 6.6 shadow stack interface +- RISC-V: Add support for dl_runtime_profile (BZ #31151) +- debug: Add fortify wprintf tests +- debug: Add fortify syslog tests +- debug: Add fortify dprintf tests +- debug: Increase tst-fortify checks for compiler without __va_arg_pack support +- debug: Adapt fortify tests to libsupport +- localedata: yo_NT: remove redundant comments +- localedata: convert en_AU, en_NZ, mi_NZ, niu_NZ to UTF-8 +- localedata: First day of the week in AU is Monday, LC_TIME in en_NZ is identical to LC_TIME in en_AU then +- localedata: convert yo_NG to UTF-8, check that language name in Yoruba agrees with CLDR +- x86-64: Fix the tcb field load for x32 [BZ #31185] +- x86-64: Fix the dtv field load for x32 [BZ #31184] + +* Wed Jan 3 2024 Florian Weimer - 2.38.9000-30 +- Infinite loop in res_mkquery with malformed domain name (#2255506) + +* Fri Dec 22 2023 Florian Weimer - 2.38.9000-29 +- Auto-sync with upstream branch master, + commit 61bac1a9d2ab80ebcbc51484722e6ea43414bec7: +- nss: Remove unused allocation from get_nscd_addresses in getaddrinfo +- x86/cet: Don't disable CET if not single threaded +- x86: Modularize sysdeps/x86/dl-cet.c +- x86/cet: Update tst-cet-vfork-1 +- elf: Add TLS modid reuse test for bug 29039 +- aarch64: Add SIMD attributes to math functions with vector versions +- aarch64: Add half-width versions of AdvSIMD f32 libmvec routines +- Fix elf/tst-env-setuid[-static] if test needs to be rerun. +- Fix elf: Do not duplicate the GLIBC_TUNABLES string +- tst-setcontext10.c: Undef _FORTIFY_SOURCE +- Fix elf: Do not duplicate the GLIBC_TUNABLES string +- riscv: Fix feenvupdate with FE_DFL_ENV (BZ 31022) +- manual: Clarify undefined behavior of feenableexcept (BZ 31019) +- x86: Do not raises floating-point exception traps on fesetexceptflag (BZ 30990) +- i686: Do not raise exception traps on fesetexcept (BZ 30989) +- powerpc: Do not raise exception traps for fesetexcept/fesetexceptflag (BZ 30988) +- elf: Do not set invalid tunables values +- elf: Do not duplicate the GLIBC_TUNABLES string +- x86/cet: Check CPU_FEATURE_ACTIVE in permissive mode +- x86/cet: Check legacy shadow stack code in .init_array section +- x86/cet: Add tests for GLIBC_TUNABLES=glibc.cpu.hwcaps=-SHSTK +- x86/cet: Check CPU_FEATURE_ACTIVE when CET is disabled +- x86/cet: Check legacy shadow stack applications +- localedata: id_ID: change first weekday to Sunday +- s390: Set psw addr field in getcontext and friends. +- x86: Unifies 'strlen-evex' and 'strlen-evex512' implementations. +- x86/cet: Don't assume that SHSTK implies IBT +- id_ID: Update Time Locales +- Update code to handle the new ABI for sending inlined port rights. +- x86/cet: Check user_shstk in /proc/cpuinfo +- Add a test for setjmp/longjmp within user context +- Add a test for longjmp from user context +- powerpc: Add space for HWCAP3/HWCAP4 in the TCB for future Power. +- powerpc: Fix performance issues of strcmp power10 +- localedata: Convert el_GR and el_CY locales to UTF-8 +- localedata: el_GR: Greece now uses the 24h format for time +- powerpc : Add optimized memchr for POWER10 +- intl: Treat C.UTF-8 locale like C locale, part 2 (BZ# 16621) +- resolv: Fix a few unaligned accesses to fields in HEADER +- x86: Check PT_GNU_PROPERTY early + +* Wed Dec 13 2023 Carlos O'Donell - 2.38.9000-28 +- Depend only on RPM 4.14 features (RHEL-19045) + +* Fri Dec 08 2023 Carlos O'Donell - 2.38.9000-27 +- Drop glibc-rh2248502.patch; fix applied upstream, and +- Auto-sync with upstream branch master, + commit b3bee76c5f59498b9c189608f0a3132e2013fa1a: +- elf: Initialize GLRO(dl_lazy) before relocating libc in dynamic startup +- Move CVE information into advisories directory +- powerpc: Optimized strcmp for power10 +- elf: Fix wrong break removal from 8ee878592c +- localedata: Convert day names in nn_NO locale to UTF-8 +- localedata: Remove trailing whitespace in weekday names in nn_NO locale +- elf: Refactor process_envvars +- elf: Ignore LD_BIND_NOW and LD_BIND_NOT for setuid binaries +- elf: Ignore loader debug env vars for setuid +- Adapt the security policy for the security page +- aarch64: correct CFI in rawmemchr (bug 31113) +- math: Add new exp10 implementation +- aarch64: fix tested ifunc variants +- stdlib: Fix array bounds protection in insertion sort phase of qsort +- Revert "Update code to handle the new ABI for sending inlined port rights." +- Revert "hurd: Fix build" +- hurd: Fix build +- Update code to handle the new ABI for sending inlined port rights. +- hurd: [!__USE_MISC] Do not #undef BSD macros in ioctls +- linux: Make fdopendir fail with O_PATH (BZ 30373) +- Avoid padding in _init and _fini. [BZ #31042] +- aarch64: Improve special-case handling in AdvSIMD double-precision libmvec routines +- malloc: Improve MAP_HUGETLB with glibc.malloc.hugetlb=2 +- elf: Add a way to check if tunable is set (BZ 27069) + +* Tue Nov 28 2023 Arjun Shankar - 2.38.9000-26 +- Drop glibc-benchtests-aarch64.patch; fix applied upstream, and +- Auto-sync with upstream branch master, + commit 9469261cf1924d350feeec64d2c80cafbbdcdd4d: +- x86: Only align destination to 1x VEC_SIZE in memset 4x loop +- elf: Fix TLS modid reuse generation assignment (BZ 29039) +- Add TCP_MD5SIG_FLAG_IFINDEX from Linux 5.6 to netinet/tcp.h. +- elf: Relocate libc.so early during startup and dlmopen (bug 31083) +- elf: Introduce the _dl_open_relocate_one_object function +- elf: In _dl_relocate_object, skip processing if object is relocated +- Remove __access_noerrno +- malloc: Use __get_nprocs on arena_get2 (BZ 30945) +- aarch64: Fix libmvec benchmarks + +* Mon Nov 27 2023 Florian Weimer - 2.38.9000-25 +- Fix qsort workaround (#2248502) + +* Thu Nov 23 2023 Florian Weimer - 2.38.9000-24 +- Restore qsort workaround for 389-ds-base. (#2248502) + +* Wed Nov 22 2023 Florian Weimer - 2.38.9000-23 +- Apply glibc-benchtests-aarch64.patch to fix an aarch64 build failure. +- Drop glibc-rh2244688.patch revert. Fix applied upstream. +- Drop glibc-rh2244992.patch, glibc-rh2248915.patch, glibc-rh2248502-3.patch. + All applied upstream. +- Auto-sync with upstream branch master, + commit 5d7f1bce7d8eea31f4baeb68bcc3124b35acc751: +- posix: Revert the removal of the crypt prototype from +- elf: Add comments on how LD_AUDIT and LD_PRELOAD handle __libc_enable_secure +- elf: Ignore LD_LIBRARY_PATH and debug env var for setuid for static +- elf: Remove any_debug from dl_main_state +- elf: Remove LD_PROFILE for static binaries +- elf: Ignore LD_PROFILE for setuid binaries +- s390: Use dl-symbol-redir-ifunc.h on cpu-tunables +- x86: Use dl-symbol-redir-ifunc.h on cpu-tunables +- elf: Emit warning if tunable is ill-formatted +- elf: Fix _dl_debug_vdprintf to work before self-relocation +- elf: Do not parse ill-formatted strings +- elf: Do not process invalid tunable format +- elf: Add all malloc tunable to unsecvars +- elf: Ignore GLIBC_TUNABLES for setuid/setgid binaries +- elf: Add GLIBC_TUNABLES to unsecvars +- elf: Remove /etc/suid-debug support +- stdlib: The qsort implementation needs to use heapsort in more cases +- stdlib: Handle various corner cases in the fallback heapsort for qsort +- stdlib: Avoid another self-comparison in qsort +- hurd: fix restarting reauth_dtable on signal +- hurd: Prevent the final file_exec_paths call from signals +- manual: Fix termios.c example. (Bug 31078) +- aarch64: Add vector implementations of expm1 routines +- linux: Use fchmodat2 on fchmod for flags different than 0 (BZ 26401) +- intl: Add test case for bug 16621 +- resolv: free only initialized items from gai pool +- ldconfig: Fixes for skipping temporary files. +- nptl: Link tst-execstack-threads-mod.so with -z execstack +- nptl: Rename tst-execstack to tst-execstack-threads +- localedata: Convert oc_FR locale to UTF-8 +- localedata: Add information for Occitan +- elf: Fix force_first handling in dlclose (bug 30981) +- elf: Handle non-directory name in search path (BZ 31035) +- New Zealand locales (en_NZ & mi_NZ) first day of week should be Monday +- x86: Fix unchecked AVX512-VBMI2 usage in strrchr-evex-base.S +- posix: Check pidfd_spawn with tst-spawn7-pid +- y2038: Fix support for 64-bit time on legacy ABIs +- AArch64: Remove Falkor memcpy +- AArch64: Add memset_zva64 +- AArch64: Cleanup emag memset +- test: Run the tst-tls-allocation-failure-static-patched with test-wrapper. +- aarch64: Add vector implementations of log1p routines +- aarch64: Add vector implementations of atan2 routines +- aarch64: Add vector implementations of atan routines +- aarch64: Add vector implementations of acos routines +- aarch64: Add vector implementations of asin routines + +* Wed Nov 15 2023 Florian Weimer - 2.38.9000-22 +- Work around another self-comparison application issue in qsort (#2248502) + +* Sat Nov 11 2023 Florian Weimer - 2.38.9000-21 +- Fix missing entries in /etc/ld.so.cache (#2248915) + +* Sat Nov 11 2023 Florian Weimer - 2.38.9000-20 +- Drop glibc-rh2248502-*.patch, workaround applied upstream +- Auto-sync with upstream branch master, + commit d1dcb565a1fb5829f9476a1438c30eccc4027d04: +- Fix type typo in “String/Array Conventions” doc +- stdlib: Avoid element self-comparisons in qsort (#2248502) +- elf: Add glibc.mem.decorate_maps tunable +- linux: Decorate __libc_fatal error buffer +- assert: Decorate error message buffer +- malloc: Decorate malloc maps +- nptl: Decorate thread stack on pthread_create +- support: Add support_set_vma_name +- linux: Add PR_SET_VMA_ANON_NAME support + +* Wed Nov 8 2023 Florian Weimer - 2.38.9000-19 +- Fix force-first handling in dlclose, take two (#2244992, #2246048) + +* Tue Nov 07 2023 Florian Weimer - 2.38.9000-18 +- Revert back to old qsort/qsort_r implementation (#2248502) +- Adjust test build completion check to match new DejaGnu-style message. +- Auto-sync with upstream branch master, + commit 5dd3bda59c2d9da138f0d98808d087cdb95cdc17: +- sysdeps: sem_open: Clear O_CREAT when semaphore file is expected to exist [BZ #30789] +- Add SEGV_CPERR from Linux 6.6 to bits/siginfo-consts.h +- linux: Sync Linux 6.6 elf.h +- linux: Add HWCAP2_HBC from Linux 6.6 to AArch64 bits/hwcap.h +- linux: Add FSCONFIG_CMD_CREATE_EXCL from Linux 6.6 to sys/mount.h +- linux: Add MMAP_ABOVE4G from Linux 6.6 to sys/mman.h +- Update kernel version to 6.6 in header constant tests +- Update syscall lists for Linux 6.6 +- Format test results closer to what DejaGnu does +- AArch64: Cleanup ifuncs +- Use correct subdir when building tst-rfc3484* for mach and arm +- stdlib: Add more qsort{_r} coverage +- stdlib: qsort: Move some macros to inline function +- stdlib: Move insertion sort out qsort +- stdlib: Optimization qsort{_r} swap implementation +- string: Add internal memswap implementation +- crypt: Remove manul entry for --enable-crypt +- Use Linux 6.6 in build-many-glibcs.py +- crypt: Remove libcrypt support +- sparc: Remove optimize md5, sha256, and sha512 +- build-many-glibcs: Fix traililing whitespace +- AArch64: Add support for MOPS memcpy/memmove/memset +- Move getnameinfo from 'inet' to 'nss' +- Move getaddrinfo from 'posix' into 'nss' +- Move 'services' routines from 'inet' into 'nss' +- Move 'rpc' routines from 'inet' into 'nss' +- Move 'protocols' routines from 'inet' into 'nss' +- Move 'networks' routines from 'inet' into 'nss' +- Move 'netgroup' routines from 'inet' into 'nss' +- Move 'hosts' routines from 'inet' into 'nss' +- Move 'ethers' routines from 'inet' into 'nss' +- Move 'aliases' routines from 'inet' into 'nss' +- Remove 'shadow' and merge into 'nss' +- Remove 'pwd' and merge into 'nss' +- Remove 'gshadow' and merge into 'nss' +- Remove 'grp' and merge into 'nss' and 'posix' +- malloc: Fix tst-tcfree3 build csky-linux-gnuabiv2 with fortify source +- test-container: disable ld.so system cache on DSO detection +- aarch64: Add vector implementations of exp10 routines +- aarch64: Add vector implementations of log10 routines +- aarch64: Add vector implementations of log2 routines +- aarch64: Add vector implementations of exp2 routines +- aarch64: Add vector implementations of tan routines +- elf: ldconfig should skip temporary files created by package managers +- tst-spawn-cgroup.c: Fix argument order of UNSUPPORTED message. +- Add NT_PPC_DEXCR and NT_PPC_HASHKEYR from Linux 6.5 to elf.h +- s390: Fix undefined behaviour in feenableexcept, fedisableexcept [BZ #30960] +- elf: Do not print the cache entry if --inhibit-cache is used + +* Thu Oct 26 2023 Carlos O'Donell - 2.38.9000-17 +- Revert "Fix force-first handling in dlclose" (#2246048) + +* Tue Oct 24 2023 Arjun Shankar - 2.38.9000-16 +- Provide template gai.conf in glibc-doc + +* Thu Oct 19 2023 Florian Weimer - 2.38.9000-15 +- Fix force-first handling in dlclose (#2244992) + +* Wed Oct 18 2023 Florian Weimer - 2.38.9000-14 +- Revert "x86: Prepare `strrchr-evex` and `strrchr-evex512` for AVX10" (#2244688) +- Auto-sync with upstream branch master, + commit dd32e1db386c77c61850a7cbd0c126b7b3c63ece: +- Revert "elf: Always call destructors in reverse constructor order (bug 30785)" +- Revert "elf: Fix compile error with -DNDEBUG [BZ #18755]" +- Add strlcat/wcslcat testcase. +- Add strlcpy/wcslcpy testcase +- Add LE DSCP code point from RFC-8622. +- Add HWCAP2_MOPS from Linux 6.5 to AArch64 bits/hwcap.h +- Add SCM_SECURITY, SCM_PIDFD to bits/socket.h +- Add AT_HANDLE_FID from Linux 6.5 to bits/fcntl-linux.h +- Avoid maybe-uninitialized warning in __kernel_rem_pio2 +- Fix WAIT_FOR_DEBUGGER for container tests. + +* Thu Oct 12 2023 Florian Weimer - 2.38.9000-13 +- Drop glibc-disable-werror-tst-realloc.patch, GCC was fixed long ago. +- Auto-sync with upstream branch master, + commit 69239bd7a216007692470aa9d5f3658024638742: +- stdlib: fix grouping verification with multi-byte thousands separator (bug 30964) +- build-many-glibcs: Check for required system tools +- x86: Prepare `strrchr-evex` and `strrchr-evex512` for AVX10 +- aarch64: Optimise vecmath logs +- aarch64: Cosmetic change in SVE exp routines +- aarch64: Optimize SVE cos & cosf +- aarch64: Improve vecmath sin routines +- nss: Get rid of alloca usage in makedb's write_output. +- debug: Add regression tests for BZ 30932 +- Fix FORTIFY_SOURCE false positive +- nss: Rearrange and sort Makefile variables +- inet: Rearrange and sort Makefile variables +- Fix off-by-one OOB write in iconv/tst-iconv-mt + +* Tue Oct 03 2023 Arjun Shankar - 2.38.9000-12 +- Auto-sync with upstream branch master, + 1056e5b4c3f2d90ed2b4a55f96add28da2f4c8fa: +- tunables: Terminate if end of input is reached (CVE-2023-4911) +- Propagate GLIBC_TUNABLES in setxid binaries + +* Tue Oct 03 2023 Arjun Shankar - 2.38.9000-11 +- Auto-sync with upstream branch master, + 9e4e896f0f5a19a16c1a77567463b013a0f4952d: +- Linux: add ST_NOSYMFOLLOW +- resolve: Remove __res_context_query alloca usage +- mips: dl-machine-reject-phdr: Get rid of alloca. +- x86: Add support for AVX10 preset and vec size in cpu-features +- resolv: Fix a comment typo in __resolv_conf_load +- Remove unused -DRESOLVER getaddrinfo build flag +- C2x scanf %wN, %wfN support +- test-container: Use nftw instead of rm -rf + +* Thu Sep 28 2023 Patsy Griffin - 2.38.9000-10 +- Auto-sync with upstream branch master, + commit 29d4591b07a4da53320e949557c6946c62c26bde. +- hurd: Drop REG_GSFS and REG_ESDS from x86_64's ucontext +- elf: Fix compile error with -DNDEBUG [BZ #18755] +- MIPS: Add relocation types +- MIPS: Add new section type SHT_MIPS_ABIFLAGS +- MIPS: Add ELF file header flags +- fegetenv_and_set_rn now uses the builtins provided by GCC. +- io: Do not implement fstat with fstatat +- libio: Add nonnull attribute for most FILE * arguments in stdio.h +- AArch64: Remove -0.0 check from vector sin +- Document CVE-2023-4806 and CVE-2023-5156 in NEWS +- elf: Add dummy declaration of _dl_audit_objclose for !SHARED +- Fix leak in getaddrinfo introduced by the fix for CVE-2023-4806 [BZ #30843] +- elf: dl-lookup: Remove unused alloca.h include +- Remove unused localedata/th_TH.in +- Adapt collation in th_TH locale to use the iso14651_t1_common file and sync the collation with CLDR +- Revert "LoongArch: Add glibc.cpu.hwcap support." +- Update kernel version to 6.5 in header constant tests +- LoongArch: Add glibc.cpu.hwcap support. +- math: Add a no-mathvec flag for sin (-0.0) + +* Mon Sep 18 2023 Arjun Shankar - 2.38.9000-9 +- Auto-sync with upstream branch master, + commit bb5bbc20702981c287aa3e44640e7d2f2b9a28cf: +- Update to Unicode 15.1.0 [BZ #30854] +- localedata/unicode-gen/utf8_gen.py: adapt regexp to get relevant lines from EastAsianWidth.txt +- Fix regexp syntax warnings in localedata/unicode-gen/ctype_compatibility.py +- getaddrinfo: Fix use after free in getcanonname (CVE-2023-4806) +- LoongArch: Change to put magic number to .rodata section +- LoongArch: Add ifunc support for strrchr{aligned, lsx, lasx} +- LoongArch: Add ifunc support for strcpy, stpcpy{aligned, unaligned, lsx, lasx} +- LoongArch: Replace deprecated $v0 with $a0 to eliminate 'as' Warnings. +- LoongArch: Add lasx/lsx support for _dl_runtime_profile. +- Add MOVE_MOUNT_BENEATH from Linux 6.5 to sys/mount.h +- CVE-2023-4527: Stack read overflow with large TCP responses in no-aaaa mode +- resolv: Fix some unaligned accesses in resolver [BZ #30750] +- Update syscall lists for Linux 6.5 + +* Mon Sep 11 2023 Patsy Griffin - 2.38.9000-8 +- Auto-sync with upstream branch master, + commit 073edbdfabaad4786e974a451efe4b6b3f7a5a61. +- ia64: Work around miscompilation and fix build on ia64's gcc-10 and later +- stdio: Remove __libc_message alloca usage +- htl: avoid exposing the vm_region symbol +- libio: Fix oversized __io_vtables +- Use Linux 6.5 in build-many-glibcs.py +- elf: Remove unused l_text_end field from struct link_map + +* Fri Sep 08 2023 Florian Weimer - 2.38.9000-7 +- Auto-sync with upstream branch master, + commit 6985865bc3ad5b23147ee73466583dd7fdf65892: +- elf: Always call destructors in reverse constructor order (bug 30785) +- io: Fix record locking contants for powerpc64 with __USE_FILE_OFFSET64 +- manual: Fix ld.so diagnostics menu/section structure +- getaddrinfo: Get rid of alloca +- riscv: Add support for XTheadBb in string-fz[a,i].h +- getcanonname: Fix a typo +- linux: Add pidfd_getpid +- posix: Add pidfd_spawn and pidfd_spawnp (BZ 30349) +- linux: Add posix_spawnattr_{get, set}cgroup_np (BZ 26371) +- linux: Define __ASSUME_CLONE3 to 0 for alpha, ia64, nios2, sh, and sparc +- __call_tls_dtors: Use call_function_static_weak +- intl: Treat C.UTF-8 locale like C locale (BZ# 16621) +- htl: Fix stack information for main thread +- htl: thread_local destructors support +- elf: Fix slow tls access after dlopen [BZ #19924] +- x86: Check the lower byte of EAX of CPUID leaf 2 [BZ #30643] + +* Tue Aug 29 2023 DJ Delorie - 2.38.9000-6 +- Auto-sync with upstream branch master, + commit e1d3312015e8f70344620375aedf91afe7e7e7a4. +- add GB18030-2022 charmap and test the entire GB18030 charmap [BZ #30243] +- Use GMP 6.3.0, MPFR 4.2.1 in build-many-glibcs.py +- localedata: Translit common emojis to smileys [BZ #30649] +- nscd: Skip unusable entries in first pass in prune_cache (bug 30800) +- LoongArch: Change loongarch to LoongArch in comments +- LoongArch: Add ifunc support for memcmp{aligned, lsx, lasx} +- LoongArch: Add ifunc support for memset{aligned, unaligned, lsx, lasx} +- LoongArch: Add ifunc support for memrchr{lsx, lasx} +- LoongArch: Add ifunc support for memchr{aligned, lsx, lasx} +- LoongArch: Add ifunc support for rawmemchr{aligned, lsx, lasx} +- LoongArch: Micro-optimize LD_PCREL +- LoongArch: Remove support code for old linker in start.S +- LoongArch: Simplify the autoconf check for static PIE +- Add F_SEAL_EXEC from Linux 6.3 to bits/fcntl-linux.h. +- argp-parse: Get rid of alloca +- gencat: Get rid of alloca. +- m68k: Use M68K_SCALE_AVAILABLE on __mpn_lshift and __mpn_rshift +- m68k: Fix build with -mcpu=68040 or higher (BZ 30740) +- elf: Check that --list-diagnostics output has the expected syntax +- manual: Document ld.so --list-diagnostics output +- manual/jobs.texi: Add missing @item EPERM for getpgid +- LoongArch: Add ifunc support for strncmp{aligned, lsx} +- LoongArch: Add ifunc support for strcmp{aligned, lsx} +- LoongArch: Add ifunc support for strnlen{aligned, lsx, lasx} +- htl: move pthread_attr_setdetachstate into libc +- htl: move pthread_attr_getdetachstate into libc +- htl: move pthread_attr_setschedpolicy into libc +- htl: move pthread_attr_getschedpolicy into libc +- htl: move pthread_attr_setinheritsched into libc +- htl: move pthread_attr_getinheritsched into libc +- htl: move pthread_attr_getschedparam into libc +- htl: move pthread_setschedparam into libc +- htl: move pthread_getschedparam into libc +- htl: move pthread_equal into libc +- Linux: Avoid conflicting types in ld.so --list-diagnostics + +* Tue Aug 22 2023 Arjun Shankar - 2.38.9000-5 +- Auto-sync with upstream branch master, + commit f6c8204fd7fabf0cf4162eaf10ccf23258e4d10e: +- elf: Do not run constructors for proxy objects +- x86_64: Add log1p with FMA +- Remove references to the defunct db2 subdir +- string: Fix tester build with fortify enable with gcc < 12 +- s390x: Fix static PIE condition for toolchain bootstrapping. +- m68k: fix __mpn_lshift and __mpn_rshift for non-68020 +- sysdeps: tst-bz21269: fix -Wreturn-type +- Loongarch: Add ifunc support for memcpy{aligned, unaligned, lsx, lasx} and memmove{aligned, unaligned, lsx, lasx} +- Loongarch: Add ifunc support for strchr{aligned, lsx, lasx} and strchrnul{aligned, lsx, lasx} +- sysdeps: tst-bz21269: handle ENOSYS & skip appropriately +- sysdeps: tst-bz21269: fix test parameter +- hurd: Fix strictness of +- hurd: Add prototype for and thus fix _hurdsig_abort_rpcs call +- io/tst-statvfs: fix statfs().f_type comparison test on some arches +- fxprintf: Get rid of alloca + +* Tue Aug 15 2023 Carlos O'Donell - 2.38-4 +- Collect dynamic loader diagnostics from the build system. + +* Tue Aug 15 2023 Florian Weimer - 2.38-3 +- Auto-sync with upstream branch master, + commit d6fe19facc61caffb25383d9c25eff86a0e115c8: +- configure: Add -Wall again to the default CFLAGS +- malloc: Remove bin scanning from memalign (bug 30723) +- resolv/nss_dns/dns-host: Get rid of alloca. +- x86_64: Add expm1 with FMA +- elf: Add new LoongArch reloc types (101 to 108) into elf.h +- x86: Fix incorrect scope of setting `shared_per_thread` [BZ# 30745] +- x86_64: Add log2 with FMA +- malloc: Enable merging of remainders in memalign (bug 30723) +- nscd: Do not rebuild getaddrinfo (bug 30709) +- x86_64: Sort fpu/multiarch/Makefile +- i686: Fix build with --disable-multiarch +- x86_64: Fix build with --disable-multiarch (BZ 30721) +- Add PTRACE_SET_SYSCALL_USER_DISPATCH_CONFIG etc. from Linux 6.4 to sys/ptrace.h +- Add PACKET_VNET_HDR_SZ from Linux 6.4 to netpacket/packet.h +- linux: statvfs: allocate spare for f_type +- x86: Fix for cache computation on AMD legacy cpus. +- powerpc longjmp: Fix build after chk hidden builtin fix +- LoongArch: Fix static PIE condition for toolchain bootstrapping. +- chk: Add and fix hidden builtin definitions for *_chk +- tst-realpath-toolong: return "unsupported" when PATH_MAX is undefined +- tst-*glob*: Do not check d_name size +- iconv: restore verbosity with unrecognized encoding names (bug 30694) +- configure: Remove --enable-all-warnings option +- Add IP_PROTOCOL from Linux 6.4 to bits/in.h +- Update kernel version to 6.4 in header constant tests +- PowerPC: Influence cpu/arch hwcap features via GLIBC_TUNABLES +- vfprintf-internal: Get rid of alloca. +- stdlib: Improve tst-realpath compatibility with source fortification +- Open master branch for glibc 2.39 development + +* Tue Aug 1 2023 Siddhesh Poyarekar - 2.38-2 +- Drop downstream glibc shadow stack userspace support patches. + +* Tue Aug 1 2023 Florian Weimer - 2.38-1 +- Switch to upstream 2.38 release +- : Add APX support +- translations: update cs, nl, vi +- string: Fix tester build with fortify enable with gcc 6 +- linux: Fix i686 with gcc6 +- i386: Remove memset_chk-nonshared.S +- i386: Fix build with --enable-fortify=3 +- posix: Fix test-errno build with fortify enable +- powerpc: Fix powerpc64 strchrnul build with old gcc +- sunrpc: Fix netname build with older gcc +- malloc: Fix set-freeres.c with gcc 6 +- nscd: cleanup obsolete _FORTIFY_SOURCE setting + +* Mon Jul 31 2023 Patsy Griffin - 2.37.9000-20 +- Allow for the optional removal of tzdata. +- Rebuilt for https://fedoraproject.org/wiki/Changes/AllowRemovalOfTzdata + +* Tue Jul 25 2023 DJ Delorie - 2.37.9000-19 +- Auto-sync with upstream branch master, + commit 637aac2ae3980de31a6baab236a9255fe853cc76. +- Include sys/rseq.h in tst-rseq-disable.c +- string: Fix tester with fortify enabled +- string: Fix bug-strncat1 with fortify enabled +- nscd: Use errval, not errno to guide cache update (bug 30662) +- Restore lookup of IPv4 mapped addresses in files database (bug 25457) +- Revert "MIPS: Sync elf.h from binutils" +- riscv: Update rvd libm test ulps +- MIPS: Sync elf.h from binutils +- Merge translations (bg, hr, pl, sv) +- nptl: Unconditionally use a 32-byte rseq area +- hurd: Fix tst-openloc +- scripts: Fix fortify checks if compiler does not support _FORTIFY_SOURCE=3 +- configure: Disable building libcrypt by default +- nptl: Make tst-tls3mod.so explicitly lazy +- make ‘struct pthread’ a complete type +- scripts: Add fortify checks on installed headers +- Update x86_64 libm-test-ulps (x32 ABI) +- Fix getting return address in elf/tst-audit28.c. +- [PATCH v1] x86: Use `3/4*sizeof(per-thread-L3)` as low bound for NT threshold. +- x86: Fix slight bug in `shared_per_thread` cache size calculation. +- Update i686 libm-test-ulps (again) +- Update i686 libm-test-ulps +- Merge translations (de, ro, uk, zh_TW) +- Regenerate libc.pot +- configure: Use autoconf 2.71 +- Update sparc libm-test-ulps +- s390: Add the clone3 wrapper +- manual: Fix typos in struct dl_find_object +- sparc: Fix la_symbind for bind-now (BZ 23734) +- i386: make debug wrappers compatible with static PIE +- LoongArch: Fix soft-float bug about _dl_runtime_resolve{,lsx,lasx} + +* Wed Jul 19 2023 Fedora Release Engineering - 2.37.9000-18 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_39_Mass_Rebuild + +* Tue Jul 11 2023 Arjun Shankar - 2.37.9000-17 +- Auto-sync with upstream branch master, + commit 7f079fdc16e88ebb8020e17b2fd900e8924da29a: +- LoongArch: Add vector implementation for _dl_runtime_resolve. +- LoongArch: config: Added HAVE_LOONGARCH_VEC_ASM. +- sysdeps: Add missing hidden definitions for i386 +- sysdeps/s390: Exclude fortified routines from being built with _FORTIFY_SOURCE +- Translations: Add new ro support and update others. +- elf: _dl_find_object may return 1 during early startup (bug 30515) +- LoongArch: config: Rewrite check on static PIE. +- Revert "hppa: Drop 16-byte pthread lock alignment" +- realloc: Limit chunk reuse to only growing requests [BZ #30579] +- vfscanf-internal: Remove potentially unbounded allocas + +* Thu Jul 06 2023 Frédéric Bérat - 2.37.9000-16 +- Add "--enable-fortify-source" option to configure +- Auto-sync with upstream branch master, + commit 5324d258427fd11ca0f4f595c94016e568b26d6b. +- fileops: Don't process ,ccs= as individual mode flags (BZ#18906) +- sysdeps/ieee754/ldbl-128ibm-compat: Fix warn unused result +- libio/bits/stdio2-decl.h: Avoid PLT entries with _FORTIFY_SOURCE +- libio/bits/stdio2.h: Clearly separate declaration from definitions +- misc/bits/syslog.h: Clearly separate declaration from definition +- misc/bits/select2.h: Clearly separate declaration from definitions +- unistd: Avoid PLT entries with _FORTIFY_SOURCE +- posix/bits/unistd.h: Clearly separate declaration from definitions +- wchar: Avoid PLT entries with _FORTIFY_SOURCE +- misc/sys/cdefs.h: Create FORTIFY redirects for internal calls +- stdio: Ensure *_chk routines have their hidden builtin definition available +- string: Ensure *_chk routines have their hidden builtin definition available +- sysdeps: Ensure ieee128*_chk routines to be properly named +- Exclude routines from fortification +- Allow glibc to be built with _FORTIFY_SOURCE +- manual: Update documentation of strerror and related functions +- manual: Enhance documentation of the functions +- Always do locking when accessing streams (bug 15142, bug 14697) +- hurd: Implement MAP_EXCL +- hurd: Fix mapping at address 0 with MAP_FIXED +- hurd: Fix calling vm_deallocate (NULL) +- hurd: Map brk non-executable +- htl: Let Mach place thread stacks +- mach: strerror must not return NULL (bug 30555) +- hppa: xfail debug/tst-ssp-1 when have-ssp is yes (gcc-12 and later) +- support: Build with exceptions and asynchronous unwind tables [BZ #30587] +- hurd: Make getrandom return ENOSYS when /dev/random is not set up +- Stop applying a GCC-specific workaround on clang [BZ #30550] +- ld.so: Always use MAP_COPY to map the first segment [BZ #30452] +- setenv.c: Get rid of alloca. +- Add checks for wday, yday and new date formats +- aarch64: Add vector implementations of exp routines +- aarch64: Add vector implementations of log routines +- aarch64: Add vector implementations of sin routines +- aarch64: Add vector implementations of cos routines +- Switch to UTF-8 for INSTALL +- Make sure INSTALL is ASCII plaintext +- Update syscall lists for Linux 6.4 + +* Wed Jun 28 2023 Carlos O'Donell - 2.37.9000-15 +- Auto-sync with upstream branch master, + commit d35fbd3e684e6bb5e5ec452ad8dac6ada8424bdd: +- linux: Return unsupported if procfs can not be mount on tst-ttyname-namespace +- linux: Split tst-ttyname +- Use Linux 6.4 in build-many-glibcs.py +- x86: Adjust Linux x32 dl-cache inclusion path +- elf: Update list of RISC-V relocations +- Fix tests-clean Makefile target (bug 30545) +- check_native: Get rid of alloca +- ifaddrs: Get rid of alloca +- x86: Make dl-cache.h and readelflib.c not Linux-specific +- elf: Port ldconfig away from stack-allocated paths +- Call "CST" a time zone abbreviation, not a name +- benchtests: fix warn unused result +- sysdeps/powerpc/fpu/tst-setcontext-fpscr.c: Fix warn unused result +- rt/tst-mqueue4.c: Fix wrong number of argument for mq_open +- debug/readlink{, at}_chk.c: Harmonize declaration and definition +- wcsmbs/bits/wchar2{, -decl}.h: Clearly separate declaration from definitions +- stdio-common: tests: Incorrect maxlen parameter for swprintf +- sysdeps/{i386, x86_64}/mempcpy_chk.S: fix linknamespace for __mempcpy_chk +- hurd: readv: Get rid of alloca +- hurd: writev: Add back cleanup handler +- Fix misspellings -- BZ 25337 +- C2x scanf %b support +- C2x printf %wN, %wfN support (bug 24466) +- tests: replace system by xsystem +- tests: replace read by xread +- hurd: writev: Get rid of alloca +- grantpt: Get rid of alloca +- string: strerror must not return NULL (bug 30555) +- hurd: Add strlcpy, strlcat, wcslcpy, wcslcat to libc.abilist +- manual: Manual update for strlcat, strlcpy, wcslcat, wclscpy +- Add the wcslcpy, wcslcat functions +- Implement strlcpy and strlcat [BZ #178] +- tests: replace fgets by xfgets +- tests: replace fread by xfread +- posix: Add test case for gai_strerror() +- posix: Handle success in gai_strerror() +- LoongArch: Add support for dl_runtime_profile +- malloc: Decrease resource usage for malloc tests +- stdlib: Tune down fork arc4random tests +- tst-getdate: Improve testcase flexibility and add test. +- x86: Make the divisor in setting `non_temporal_threshold` cpu specific +- x86: Refactor Intel `init_cpu_features` +- x86: Increase `non_temporal_threshold` to roughly `sizeof_L3 / 4` +- Remove unused DATEMSK file for tst-getdate +- resolv_conf: release lock on allocation failure (bug 30527) + +* Thu Jun 08 2023 Siddhesh Poyarekar - 2.37.9000-14 +- Shadow stack userspace support, downstream only and disabled by default. + +* Wed Jun 07 2023 Arjun Shankar - 2.37.9000-13 +- Auto-sync with upstream branch master, + commit 85e6d8b4175fcb195011a0a1bad37d6f3b2355db: +- time: Fix use-after-free in getdate +- Move {read,write}_all functions to a dedicated header +- tests: Replace various function calls with their x variant +- tests: fix warn unused result on asprintf calls +- pthreads: Use _exit to terminate the tst-stdio1 test +- support: Add delayed__exit (with two underscores) + +* Mon Jun 05 2023 Arjun Shankar - 2.37.9000-12 +- Auto-sync with upstream branch master, + commit e3622a8f391deea3b75a577dce70d023dfa3f1c7. +- time: Also check for EPERM while trying to clock_settime +- linux: Fail as unsupported if personality call is filtered +- Remove MAP_VARIABLE from hppa bits/mman.h +- hurd: Fix x86_64 sigreturn restoring bogus reply_port +- Add lint-makefiles Makefile linting test. +- elf: Sort Makefile variables. +- Fix a few more typos I missed in previous round -- BZ 25337 +- Fix all the remaining misspellings -- BZ 25337 +- Use __nonnull for the epoll_wait(2) family of syscalls +- Fix invalid use of NULL in epoll_pwait2(2) test +- getipv4sourcefilter: Get rid of alloca +- getsourcefilter: Get rid of alloca. +- tests: fix warn unused results +- nptl_db/thread_dbP.h: fix warn unused result +- malloc/{memusage.c, memusagestat.c}: fix warn unused result +- catgets/gencat.c: fix warn unused result +- tests: replace ftruncate by xftruncate +- tests: replace write by xwrite +- x86-64: Use YMM registers in memcmpeq-evex.S + +* Thu Jun 01 2023 Patsy Griffin - 2.37.9000-11 +- Auto-sync with upstream branch master, + commit 6286cca2cb8389dcffec39238a8bf15ffea96396. +- support: Don't fail on fchown when spawning sgid processes +- io: Fix F_GETLK, F_SETLK, and F_SETLKW for powerpc64 +- elf: Remove spurios SHARED conditional from elf/rtld.c +- Fix misspellings in sysdeps/ -- BZ 25337 +- io: Fix record locking contants on 32 bit arch with 64 bit default time_t (BZ#30477) +- io: Re-flow and sort multiline Makefile definitions +- elf: Make more functions available for binding during dlclose (bug 30425) +- LoongArch: Fix inconsistency in SHMLBA macro values between glibc and kernel +- Fix misspellings in elf/ -- BZ 25337 +- riscv: Add the clone3 wrapper +- posix: Add error message for EAI_OVERFLOW +- setsourcefilter: Replace alloca with a scratch_buffer. +- time: strftime_l: Avoid an unbounded alloca. +- x86: Use 64MB as nt-store threshold if no cacheinfo [BZ #30429] +- hurd: Fix setting up signal thread stack alignment +- mach: Fix startup with stack protector +- Fix misspellings in manual/ -- BZ 25337 +- Fix misspellings in iconv/ and iconvdata/ -- BZ 25337 +- Add MFD_NOEXEC_SEAL, MFD_EXEC from Linux 6.3 to bits/mman-shared.h +- Add IP_LOCAL_PORT_RANGE from Linux 6.3 to bits/in.h +- Add AT_RSEQ_* from Linux 6.3 to elf.h +- setipv4sourcefilter: Avoid using alloca. +- wchar: Define va_list for POSIX (BZ #30035) +- elf: add test for dl-printf +- elf: fix handling of negative numbers in dl-printf +- elf: Update comment in open_path +- elf: Add test for locating libraries in root dir (bug 30435) +- io: Re-flow and sort multiline Makefile definitions +- Fix special case for C2x strtol binary constant handling (BZ# 30371) +- rtld: properly handle root directory in load path (bug 30435) +- sysdeps/pthread/eintr.c: fix warn unused result +- sunrpc/netname.c: fix warn unused result +- locale/programs/locarchive.c: fix warn unused result +- support: Reformat Makefile. +- Regenerate configure fragment -- BZ 25337. +- Fix misspellings in sysdeps/powerpc -- BZ 25337 +- Fix misspellings in sysdeps/unix -- BZ 25337 +- Fix misspellings in sysdeps/x86_64 -- BZ 25337. +- mach: Fix accessing mach_i386.h +- Fix misspellings in sysdeps/x86_64/fpu/multiarch -- BZ 25337. +- mach: Fix installing mach_i386.h +- hurd: Fix making ld.so run static binaries with retry +- Add voice-admit DSCP code point from RFC-5865 +- mach: Fix mach_setup_thread_impl with NULL stack_base +- Remove last remnants of have-protected +- S390: Use compile-only instead of also link-tests in configure. +- Fix build for hurd/thread-self.c for i386. +- io: Fix a typo +- htl: Use __hurd_fail () instead of assigning errno +- hurd: Use __hurd_fail () instead of assigning errno +- powerpc:GCC(<10) doesn't allow -mlong-double-64 after -mabi=ieeelongdouble +- hurd: Fix using interposable hurd_thread_self +- hurd 64bit: Re-introduce gai_suspend symbol +- hurd: Fix __TIMESIZE on x86_64 +- posix: Reformat Makefile. +- hurd: Fix expected c++ types +- catgets: Reformat Makefile. +- benchtests: Reformat Makefile. +- assert: Reformat Makefile. +- nptl: Reformat Makefile. +- wcsmbs: Reformat Makefile. +- misc: Reformat Makefile. +- stdio-common: Adjust tests in Makefile +- elf: Adjust tests in Makefile. +- scripts: sort-makefile-lines.py +- Add a SECURITY.md +- Add HWCAP2_SME* from Linux 6.3 to AArch64 bits/hwcap.h +- hurd: Also make it possible to call strlen very early +- hurd: Fix setting up pthreads +- hurd: Fix x86_64 _hurd_tls_fork +- hurd: Make sure to not use tcb->self +- hurd: Use __mach_setup_thread_call () +- mach: Add __mach_setup_thread_call () +- hurd: Use MACHINE_THREAD_STATE_SETUP_CALL +- mach: Define MACHINE_THREAD_STATE_SETUP_CALL +- Use TASK_THREAD_TIMES_INFO_COUNT when calling task_info with TASK_THREAD_TIMES_INFO +- argp: Reformat Makefile. +- stdlib: Avoid undefined behavior in stdlib/tst-labs +- stdlib: Use long long int in stdlib/tst-llabs +- Update kernel version to 6.3 in header constant tests +- i386: Use pthread_barrier for synchronization on tst-bz21269 +- stdlib: Add testcases for llabs(). (BZ #30263) +- stdlib: Add testcases for labs(). (BZ #30263) +- stdlib: Add testcases for abs(). (BZ #30263) +- hurd: Fix computing user stack pointer +- hurd: Fix sc_i386_thread_state layout +- hurd: Align signal stack pointer after allocating stackframe +- hurd: Fix aligning signal stack pointer + +* Tue May 16 2023 Carlos O'Donell - 2.37.9000-10 +- Auto-sync with upstream branch master, + commit 40b68e8cc00ca40348bc084b651c0561d31feb46: +- wcsmbs: Reformat Makefile. +- linux: Reformat Makefile. +- stdlib: Reformat Makefile. +- stdio-common: Reformat Makefile. +- socket: Reformat Makefile. +- misc: Reformat Makefile. +- debug: Reformat Makefile. +- elf: Reformat Makefile. +- libio: Add __nonnull for FILE * arguments of fclose and freopen +- nss: Reconcile conditional declaration and use of `is_nscd' +- Update syscall lists for Linux 6.3 +- hurd: rule out some mach headers when generating errno.h +- Stop checking if MiG supports retcode. +- Added Redirects to longdouble error functions [BZ #29033] +- nptl: Reformat Makefile. +- scripts: Add sort-makefile-lines.py to sort Makefile variables. +- dlopen: skip debugger notification for DSO loaded from sprof (bug 30258) + +* Tue May 09 2023 Arjun Shankar - 2.37.9000-9 +- Auto-sync with upstream branch master, + commit d1417176a35d27ffb8da0ffb1e33154163b6eeb2: +- aligned_alloc: conform to C17 +- testsuite: stdlib/isomac.c: fix REQUIREMENTS +- manual: Remove unsupported line breaks in waiting-with-clock section +- Enable new device_open_new RPC in libmachuser. +- Revert "riscv: Resolve symbols directly for symbols with STO_RISCV_VARIANT_CC." +- Update hurd/hurdselect.c to be more portable. +- hurd: Fix ld.so name +- hurd: Add ioperm symbol on x86_64 +- time: Remove alloca() from getdate +- aarch64: More configure checks for libmvec +- aarch64: SVE ACLE configure test cleanups +- hppa: Fix 'concurrency' typo in comment +- Update hurd/intr-msg.c to be more portable +- Update sysdeps/mach/hurd/ioctl.c to make it more portable +- aarch64: fix SVE ACLE check for bootstrap glibc builds +- Enable libmvec support for AArch64 +- hurd: Enable x86_64 build script +- hurd 64bit: Make dev_t word type +- malloc: Really fix tst-memalign-3 link against threads +- malloc: Fix tst-memalign-3 link against threads +- Use GCC 13 branch, Linux 6.3 in build-many-glibcs.py +- Mark various cold functions as __COLD +- Fix regex type usage +- hurd 64bit: Fix struct msqid_ds and shmid_ds fields +- hurd 64bit: Fix ipc_perm fields types +- hurd 64bit: Fix flock fields types +- hurd 64bit: Add data for check-c++-types +- hurd 64bit: Fix pthread_t/thread_t type to long +- socket: Fix tst-cmsghdr-skeleton.c use of cmsg_len +- hurd 64bit: Add missing data file for check-localplt test +- hurd 64bit: Add missing libanl +- hurd: Also XFAIL missing SA_NOCLDWAIT on 64bit +- hurd: Fix tst-writev test +- nptl: move tst-x86-64-tls-1 to nptl-only tests +- hurd: Add expected abilist files for x86_64 +- hurd: Replace reply port with a dead name on failed interruption +- Define __mig_strlen to support dynamically sized strings in hurd RPCs +- mach: Disable 32bit compatibility mode +- hurd: Make it possible to call memcpy very early +- hurd: Implement longjmp for x86_64 +- hurd: Implement sigreturn for x86_64 +- Make __mach_msg_destroy portable for x86_64 +- hurd: Mark error functions as __COLD +- cdefs.h: Define __COLD +- hurd: Fix FS_RETRY_MAGICAL "machtype" handling +- hurd: Respect existing FD_CLOEXEC in S_msg_set_fd +- hurd: Don't leak the auth port in msg* RPCs +- hurd: Make _exit work during early boot-up +- hurd: Mark various conditions as unlikely +- hurd: Move libc_hidden_def's around +- hurd: Simplify _hurd_critical_section_lock a bit +- __check_pf: Add a cancellation cleanup handler [BZ #20975] +- Remap __GLIBC_FLT_EVAL_METHOD to 0 if __FLT_EVAL_METHOD__ is -1 +- riscv: Resolve symbols directly for symbols with STO_RISCV_VARIANT_CC. +- elf.h: add PT_GNU_SFRAME +- Fix Hurd getcwd build with GCC >= 13 +- Regenerate sysdeps/mach/hurd/bits/errno.h +- locale/programs/locarchive.c: Remove unnecessary check in add_locale_archive +- manual: document posix_openpt (bug 17010) +- if_index: Remove unneeded alloca.h include +- gethostid: Do not include alloca.h + +* Tue Apr 25 2023 Patsy Griffin - 2.37.9000-8 +- Auto-sync with upstream branch master, + commit 904b94c07af84b7e4c98de3bbb822ccffcaf8c40. +- socket: Add a test for MSG_CMSG_CLOEXEC +- hurd: Do not take any flag from the CMSG_DATA +- hurd: Implement MSG_CMSG_CLOEXEC +- hurd: Don't pass FD_CLOEXEC in CMSG_DATA +- hurd: Implement prefer_map_32bit_exec tunable +- hurd: Don't attempt to deallocate MACH_PORT_DEAD +- hurd: Only deallocate addrport when it's valid +- hurd: Implement MAP_32BIT +- Use O_CLOEXEC in more places (BZ #15722) +- misc: Convert daemon () to GNU coding style +- wcsmbs: Add wcsdup() tests. (BZ #30266) +- string: Add tests for strndup (BZ #30266) +- string: Add tests for strdup (BZ #30266) +- string: Allow use of test-string.h for non-ifunc implementations. +- hurd: Don't migrate reply port into __init1_tcbhead +- hurd: Make dl-sysdep's open () cope with O_IGNORE_CTTY +- Created tunable to force small pages on stack allocation. +- malloc: Add missing shared thread library flags +- linux: Re-flow and sort multiline Makefile definitions +- posix: Re-flow and sort multiline Makefile definitions + +* Mon Apr 24 2023 Florian Weimer - 2.37.9000-7 +- Explicitly provide ldconfig paths (#2188550) + +* Thu Apr 20 2023 Florian Weimer - 2.37.9000-6 +- Auto-sync with upstream branch master, + commit 65cbd52174f5bc211dd655727c2239e25e55bfce: +- build-many-glibcs.py: --disable-gcov for gcc-first +- malloc: set NON_MAIN_ARENA flag for reclaimed memalign chunk (BZ #30101) +- rcmd.c: Fix indentation in last commit +- inet/rcmd.c: fix warn unused result +- wcsmbs: Re-flow and sort routines, tests variables in Makefile +- debug: Re-flow and sort routines variable in Makefile +- math: Improve fmod(f) performance +- Benchtests: Adjust timing +- malloc: Assure that THP mode read do write OOB end of stringt +- malloc: Assure that THP mode is always null terminated +- aio: Fix freeing memory +- elf: Stop including tls.h in ldsodefs.h +- manual: update AddressSanitizer discussion +- manual: document snprintf truncation better +- manual: improve string section wording +- manual: fix texinfo typo +- : Make fopencookie, vasprintf, asprintf available by default +- : Make strchrnul, strcasestr, memmem available by default +- : Add PREFETCHI support +- : Add AMX-COMPLEX support +- : Add AVX-NE-CONVERT support +- : Add AVX-VNNI-INT8 support +- : Add MSRLIST support +- : Add AVX-IFMA support +- : Add AMX-FP16 support +- : Add WRMSRNS support +- : Add ArchPerfmonExt support +- : Add CMPCCXADD support +- : Add LASS support +- : Add RAO-INT support +- : Add LBR support +- : Add RTM_FORCE_ABORT support +- : Add SGX-KEYS support +- : Add BUS_LOCK_DETECT support +- : Add LA57 support +- platform.texi: Move LAM after LAHF64_SAHF64 +- : Rename to x86_cpu_INDEX_7_ECX_15 +- hppa: Update struct __pthread_rwlock_arch_t comment. +- hppa: Revise __TIMESIZE define to use __WORDSIZE +- libio: Remove unused pragma weak on vtable +- malloc: Only set pragma weak for rpc freemem if required +- compare_strings.py : Add --gmean flag +- x86/dl-cacheinfo: remove unsused parameter from handle_amd +- powerpc: Disable stack protector in early static initialization +- nptl: Fix tst-cancel30 on sparc64 +- math: Remove the error handling wrapper from fmod and fmodf +- math: Improve fmodf +- math: Improve fmod +- benchtests: Add fmodf benchmark +- benchtests: Add fmod benchmark +- x86: Set FSGSBASE to active if enabled by kernel +- x86_64: Fix asm constraints in feraiseexcept (bug 30305) +- manual: Document __wur usage under _FORTIFY_SOURCE +- x86_64: Add rtld-stpncpy & rtld-strncpy +- stdio-common: Fix building when !IS_IN (libc) +- time: Fix strftime(3) API regarding nullability +- Update arm libm-tests-ulps +- getlogin_r: fix missing fallback if loginuid is unset (bug 30235) +- memalign: Support scanning for aligned chunks. +- malloc: Use C11 atomics on memusage +- Remove --enable-tunables configure option +- Remove --disable-experimental-malloc option + +* Fri Mar 31 2023 Arjun Shankar - 2.37.9000-5 +- Apply glibc-disable-werror-tst-realloc.patch to disable spurious GCC + warning; and +- Auto-sync with upstream branch master, + commit 885d3cda907d0dee54b13cbbf61b040c9951d5a2: +- Allow building with --disable-nscd again +- system: Add "--" after "-c" for sh (BZ #28519) +- posix: Fix some crashes in wordexp [BZ #18096] +- LoongArch: ldconfig: Add comments for using EF_LARCH_OBJABI_V1 +- elf: Take into account ${sysconfdir} in elf/tst-ldconfig-p.sh +- Fix tst-glibc-hwcaps-prepend-cache with custom configure prefix value +- Fix tst-ldconfig-ld_so_conf-update with custom configure prefix value +- support: introduce support_sysconfdir_prefix +- Remove set-hooks.h from generic includes +- Remove --with-default-link configure option +- libio: Remove the usage of __libc_IO_vtables +- libio: Do not autogenerate stdio_lim.h +- Move libc_freeres_ptrs and libc_subfreeres to hidden/weak functions +- benchtests: Move libmvec benchtest inputs to benchtests directory +- stdio-common: tests: don't double-define _FORTIFY_SOURCE +- LoongArch: ldconfig: Ignore EF_LARCH_OBJABI_V1 in shared objects +- _dl_map_object_from_fd: Remove unnecessary debugger notification in error path +- hppa: Drop 16-byte pthread lock alignment +- Minor: don't call _dl_debug_update (which can have side effects) inside assert +- x86: Don't check PREFETCHWT1 in tst-cpu-features-cpuinfo.c +- Declare wcstofN, wcstofNx for C2x +- Update printf %b/%B C2x support +- ARC: run child from the separate start block in __clone +- ARC: Add the clone3 wrapper + +* Mon Mar 13 2023 Florian Weimer - 2.37.9000-4 +- Auto-sync with upstream branch master, + commit 90233f113cc941ef88ce03b7f73221a964dcaca8: +- LoongArch: Add get_rounding_mode. +- LoongArch: Add support for ldconfig. +- linux: fix ntp_gettime abi break (BZ# 30156) +- elf: Add missing dependency between resolvfail and testobj1.so +- elf: Add -z lazy to some more tests +- Benchtests: Remove simple_str(r)chr +- Benchtests: Remove simple_str(n)casecmp +- Benchtests: Remove simple_memcmp +- Benchtests: Remove simple_strcspn/strpbrk/strsep +- Benchtests: Remove memchr_strnlen +- Benchtests: Remove simple_mem(r)chr +- Benchtests: Remove simple_strcpy_chk +- Benchtests: Remove simple_str(n)cmp +- malloc: Fix transposed arguments in sysmalloc_mmap_fallback call +- rt: fix shm_open not set ENAMETOOLONG when name exceeds {_POSIX_PATH_MAX} +- posix: Ensure the initial signal disposition for tst-spawn7 +- hurd: fix build of tst-system.c +- x86: Fix bug about glibc.cpu.hwcaps. +- posix: Fix system blocks SIGCHLD erroneously [BZ #30163] +- gshadow: Matching sgetsgent, sgetsgent_r ERANGE handling (bug 30151) + +* Mon Mar 06 2023 DJ Delorie - 2.37.9000-3 +- Auto-sync with upstream branch master, + commit 8390014c2320f94ffd8a8f6088c10c1f64567954. +- Update kernel version to 6.2 in header constant tests +- arm: Remove __builtin_arm_uqsub8 usage on string-fza.h +- alpha: Remove strncmp optimization +- powerpc: Remove powerpc64 strncmp variants +- powerpc: Remove strncmp variants +- C2x scanf binary constant handling +- Fix stringop-overflow warning in test-strncat. +- nis: Fix stringop-truncation warning with -O3 in nis_local_host. +- support: use 64-bit time_t (bug 30111) +- LoongArch: Update libm-test-ulps. +- LoongArch: Further refine the condition to enable static PIE +- hurd: Fix some broken indentation +- hurd: Remove the ecx kludge + +* Wed Mar 01 2023 Carlos O'Donell - 2.37.9000-2 +- Auto-sync with upstream branch master, + commit 59a6d5e9477695c41d6feef7ef8636f8f744f3c5: +- Add AArch64 HWCAP2 values from Linux 6.2 to bits/hwcap.h +- crypt: Remove invalid end of page test badsalttest +- S390: Fix _FPU_SETCW/GETCW when compiling with Clang [BZ #30130] +- s390x: Regenerate ULPs. +- Add Arm HWCAP values from Linux 6.2 to bits/hwcap.h +- htl: Add pthreadtypes-arch.h for x86_64 +- hurd: Implement TLS for x86_64 +- htl: Make pthread_mutex_t pointer-aligned +- x86_64: Update libm test ulps +- localedata: de_DE should not use Fräulein +- LoongArch: Add math-barriers.h +- cdefs.h: fix "__clang_major" typo +- hppa: Drop old parisc-specific MADV_* constants +- hurd: Generalize init-first.c to support x86_64 +- hurd: Simplify init-first.c further +- hurd: Mark some audit tests as unsupported +- htl: Mark select loop test as unsupported +- hurd: Mark RLIMIT_AS tests as unsupported +- aarch64: update libm test ulps +- powerpc:Regenerate ulps for hypot +- Update syscall lists for Linux 6.2 +- tunables.texi: Change \code{1} to @code{1} +- x86-64: Add glibc.cpu.prefer_map_32bit_exec [BZ #28656] +- gmon: fix memory corruption issues [BZ# 30101] +- gmon: improve mcount overflow handling [BZ# 27576] +- gmon: Fix allocated buffer overflow (bug 29444) +- malloc: remove redundant check of unsorted bin corruption +- Use Linux 6.2 in build-many-glibcs.py +- Ignore MAP_VARIABLE in tst-mman-consts.py +- AArch64: Fix HP_TIMING_DIFF computation [BZ# 29329] + +* Mon Feb 20 2023 Arjun Shankar - 2.37.9000-1 +- Drop glibc-printf-grouping-swbz30068.patch; fix applied upstream, and +- Auto-sync with upstream branch master, + commit 8b014a1b1f7aee1e3348db108aeea396359d481e: +- s390: Fix build for -march=z13 +- arm: Support gcc older than 10 for find_zero_all +- Linux: Remove generic Implies +- Linux: Remove unused generic Makefile +- Linux: Assume and consolidate getpeername wire-up syscall +- Linux: Assume and consolidate getsockname wire-up syscall +- Linux: Move wordsize-32 Version to default +- __glob64_time64: Fix typo for stub_warning call (BZ #30146) +- elf: Restore ldconfig libc6 implicit soname logic [BZ #30125] +- stdlib: Undo post review change to 16adc58e73f3 [BZ #27749] +- Define PC, SP and SYSRETURN for hurd x86_64 +- mach: Use PAGE_SIZE +- hurd: Simplify init-first.c a bit +- hurd: Make timer_t pointer-sized +- hurd: Fix xattr function return type +- hurd: Use proper integer types +- hurd: Move thread state manipulation into _hurd_tls_new () +- glob64_time64: Fix typo for stub_warning call (BZ #30146) +- Use uintptr_t instead of performing pointer subtraction with a null pointer +- ARC:fpu: add extra capability check before use of sqrt and fma builtins +- ARC: align child stack in clone +- string: Remove string_private.h +- iconv: Remove _STRING_ARCH_unaligned usage +- iconv: Remove _STRING_ARCH_unaligned usage for get/set macros +- resolv: Remove _STRING_ARCH_unaligned usage +- nscd: Remove _STRING_ARCH_unaligned usage +- stdlib: Simplify getenv +- crypto: Remove _STRING_ARCH_unaligned usage +- Fix ifunc-impl-list.c build for s390 +- [hurd] Fix i686 build breakage caused by 4fedebc91108 +- C2x strtol binary constant handling +- [hurd] Add MTU_DISCOVER values +- hurd: Fix unwinding over INTR_MSG_TRAP in shared too +- mach: undef ENTRY2 +- hurd: i386 TLS tweaks +- stdio: Do not ignore posix_spawn error on popen (BZ #29016) +- update auto-libm-test-out-hypot +- added pair of inputs for hypotf in binary32 +- Naming the parameter of dummy_sa_handler +- hurd: Fix tcflag_t and speed_t types on 64-bit +- htl: Remove ./sysdeps/htl/bits/types/struct___pthread_mutex.h +- hurd, htl: Add some x86_64-specific code +- Fix typos in comments +- htl: Generalize i386 pt-machdep.h to x86 +- hurd: Set up the basic tree for x86_64-gnu +- mach: Look for mach_i386.defs on x86_64 too +- htl: Fix semaphore reference +- hurd: Fix xattr error value +- mach, hurd: Cast through uintptr_t +- hurd: Use mach_msg_type_number_t where appropriate +- hurd: Refactor readlinkat() +- Use __builtin_FILE instead of __FILE__ in assert in C++. +- hurd: Fix unwinding over INTR_MSG_TRAP +- powerpc64: Add the clone3 wrapper +- string: Disable stack protector in early static initialization +- string: Add libc_hidden_proto for memrchr +- string: Add libc_hidden_proto for strchrnul +- elf: Smoke-test ldconfig -p against system /etc/ld.so.cache +- NEWS: Document CVE-2023-25139. +- Use 64-bit time_t interfaces in strftime and strptime (bug 30053) +- C-SKY: Strip hard float abi from hard float feature. +- S390: Influence hwcaps/stfle via GLIBC_TUNABLES. +- string: Hook up the default implementation on test-strrchr +- string: Hook up the default implementation on test-memrchr +- string: Hook up the default implementation on test-memchr +- string: Hook up the default implementation on test-strcpy +- string: Hook up the default implementation on test-stpcpy +- string: Hook up the default implementation on test-strncmp +- string: Hook up the default implementation on test-strcmp +- string: Hook up the default implementation on test-strchr +- string: Hook up the default implementation on test-strnlen +- string: Hook up the default implementation on test-strlen +- riscv: Add string-fza.h and string-fzi.h +- sh: Add string-fzb.h +- powerpc: Add string-fza.h +- arm: Add string-fza.h +- alpha: Add string-fza, string-fzb.h, string-fzi.h, and string-shift.h +- hppa: Add string-fza.h, string-fzc.h, and string-fzi.h +- hppa: Add memcopy.h +- string: Improve generic strrchr with memrchr and strlen +- string: Improve generic memrchr +- string: Improve generic strnlen with memchr +- string: Improve generic memchr +- string: Improve generic strcpy +- string: Improve generic stpcpy +- string: Improve generic strncmp +- string: Improve generic strcmp +- string: Improve generic strchr +- string: Improve generic strchrnul +- string: Improve generic strlen +- Add string vectorized find and detection functions +- Parameterize OP_T_THRES from memcopy.h +- Parameterize op_t from memcopy.h +- Replace rawmemchr (s, '\0') with strchr +- AArch64: Improve SVE memcpy and memmove +- Account for grouping in printf width (bug 30068) +- Move RETURN_TO to x86/sysdep.h and implement x86_64 version. +- Remove pthread-pi-defines.sym +- stdlib: tests: don't double-define _FORTIFY_SOURCE +- LoongArch: Add new relocation types. +- Remove sysdeps/mach/i386/machine-sp.h +- cdefs: Limit definition of fortification macros +- hurd: Move some i386 bits to x86 +- Remove support setting custom demuxers during signal handling. +- hurd: Implement SHM_ANON +- hurd: Implement O_TMPFILE +- hurd: Consolidate file_name_lookup implementation +- Linux: optimize clone3 internal usage +- aarch64: Add the clone3 wrapper +- linux: Add clone3 CLONE_CLEAR_SIGHAND optimization to posix_spawn +- Linux: Do not align the stack for __clone3 +- linux: Extend internal clone3 documentation +- linux: Do not reset signal handler in posix_spawn if it is already SIG_DFL +- Open master branch for glibc 2.38 development + +* Sat Feb 04 2023 Carlos O'Donell - 2.37-1 +- Drop already included glibc-dprintf-length.patch patch. +- Apply glibc-printf-grouping-swbz30068.patch to fix swbz#30068. +- Auto-sync with upstream branch release/2.37/master, + commit a704fd9a133bfb10510e18702f48a6a9c88dbbd5: +- Create ChangeLog.old/ChangeLog.26. (tag: glibc-2.37) +- Prepare for glibc 2.37 release. +- x86: Fix strncat-avx2.S reading past length [BZ #30065] +- Update install.texi, and regenerate INSTALL. +- Update manual/contrib.texi. +- Update NEWS file with bug fixes. +- Regenerate configure. +- Update all PO files in preparation for release. +- doc: correct _FORTIFY_SOURCE doc in features.h +- libio: Update number of written bytes in dprintf implementation + +* Tue Jan 31 2023 Florian Weimer - 2.36.9000-25 +- Apply glibc-dprintf-length.patch to fix dprintf return value regression. +- Auto-sync with upstream branch master, + commit 2f39e44a8417b4186a7f15bfeac5d0b557e63e03: +- Account for octal marker in %#o format (rhbz#2165869) +- Use binutils 2.40 branch in build-many-glibcs.py +- Use MPFR 4.2.0, MPC 1.3.1 in build-many-glibcs.py + +* Wed Jan 25 2023 Florian Weimer - 2.36.9000-24 +- Auto-sync with upstream branch master, + commit 0d50f477f47ba637b54fb03ac48d769ec4543e8d: +- stdio-common: Handle -1 buffer size in __sprintf_chk & co (bug 30039) +- Document '%F' format specifier +- sparc (64bit): Regenerate ulps +- ia64: Regenerate ulps +- Update libc.pot for 2.37 release. +- x86: Cache computation for AMD architecture. +- manual: Fix typo +- Add STATX_DIOALIGN from Linux 6.1 to bits/statx-generic.h +- Add IPPROTO_L2TP from Linux 6.1 to netinet/in.h +- AArch64: Improve strrchr +- AArch64: Optimize strnlen +- AArch64: Optimize strlen +- AArch64: Optimize strcpy +- AArch64: Improve strchrnul +- AArch64: Optimize strchr +- AArch64: Improve strlen_asimd +- AArch64: Optimize memrchr +- AArch64: Optimize memchr + +* Thu Jan 19 2023 Fedora Release Engineering - 2.36.9000-23 +- Rebuilt for https://fedoraproject.org/wiki/Fedora_38_Mass_Rebuild + +* Tue Jan 17 2023 Florian Weimer - 2.36.9000-22 +- Auto-sync with upstream branch master, + commit 569cfcc6bf35c28112ca8d7112e9eb4a22bed5b8: +- hurd: Fix _NOFLSH value +- elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864] +- string: Suppress -Wmaybe-unitialized for wordcopy [BZ #19444] +- scripts/build-many-glibcs.py: Remove unused RANLIB and STRIP option +- configure: Move nm, objdump, and readelf to LIBC_PROG_BINUTILS + +* Wed Jan 11 2023 Patsy Griffin - 2.36.9000-21 +- Auto-sync with upstream branch master, + commit 2d2d7e1a8f2e62b442ae8978f0a6c17f385575c4. +- configure: Allow user override LD, AR, OBJCOPY, and GPROF +- math: Suppress -O0 warnings for soft-fp fsqrt [BZ #19444] +- sunrpc: Suppress GCC -O1 warning on user2netname [BZ #19444] +- locale: Use correct buffer size for utf8_sequence_error [BZ #19444] +- Add HWCAP2_SVE_EBF16 from Linux 6.1 to AArch64 bits/hwcap.h +- Add _FORTIFY_SOURCE implementation documentation [BZ #28998] +- Update copyright dates not handled by scripts/update-copyrights +- Update copyright dates with scripts/update-copyrights +- Remove trailing whitespace in gmp.h +- Remove trailing whitespace +- C2x semantics for +- time: Set daylight to 1 for matching DST/offset change (bug 29951) +- Fix ldbl-128 built-in function use +- x86: Check minimum/maximum of non_temporal_threshold [BZ #29953] +- i686: Regenerate ulps + +* Mon Jan 02 2023 Arjun Shankar - 2.36.9000-20 +- Drop glibc-rh2155825.patch; fix applied upstream, and +- Auto-sync with upstream branch master, + commit 5f55b22f4b3ea14c777a60f239d25dc4555eb804: +- hurd getcwd: Fix memory leak on error +- hurd fcntl: Make LOCKED macro more robust +- hurd: Make dl-sysdep __sbrk check __vm_allocate call +- htl: Drop duplicate check in __pthread_stack_alloc +- hurd hurdstartup: Initialize remaining fields of hurd_startup_data +- hurd _S_msg_add_auth: Initialize new arrays to 0 +- htl: Check error returned by __getrlimit +- getdelim: ensure error indicator is set on error (bug 29917) +- htl: Fix sem_wait race between read and gsync_wait +- Avoid use of atoi in malloc +- Linux: Pass size argument of epoll_create to the kernel +- Simplify scripts/cross-test-ssh.sh configuration. +- Define MADV_COLLAPSE from Linux 6.1 +- powerpc64: Increase SIGSTKSZ and MINSIGSTKSZ +- Update all PO files in preparation for release. +- Update kernel version to 6.1 in header constant tests +- Update syscall lists for Linux 6.1 +- libio: Convert __vswprintf_internal to buffers (bug 27857) +- libio: Convert __obstack_vprintf_internal to buffers (bug 27124) +- libio: Convert __vdprintf_internal to buffers +- libio: Convert __vasprintf_internal to buffers +- libio: Convert __vsprintf_internal to buffers +- stdio-common: Add lock optimization to vfprintf and vfwprintf +- stdio-common: Convert vfprintf and related functions to buffers +- stdio-common: Add __translated_number_width +- stdio-common: Add __printf_function_invoke +- stdio-common: Introduce buffers for implementing printf +- locale: Implement struct grouping_iterator +- Use Linux 6.1 in build-many-glibcs.py +- Avoid use of atoi in some places in libc + +* Thu Dec 22 2022 Florian Weimer - 2.36.9000-19 +- Fix epoll_create regression (#2155825) + +* Mon Dec 19 2022 Florian Weimer - 2.36.9000-18 +- Auto-sync with upstream branch master, + commit c1c0dea38833751f36a145c322ce53c9a08332e1: +- Linux: Remove epoll_create, inotify_init from syscalls.list (#2154747) +- Linux: Reflow and sort some Makefile variables +- mach: Drop remnants of old_CFLAGS +- mach: Fix passing -ffreestanding when checking for gnumach headers +- Force use of -ffreestanding when checking for gnumach headers +- elf: Fix tst-relro-symbols.py argument passing +- x86: Prevent SIGSEGV in memcmp-sse2 when data is concurrently modified [BZ #29863] +- Allow _Qp_fgt in sparc64 localplt.data + +* Mon Dec 12 2022 DJ Delorie - 2.36.9000-17 +- Auto-sync with upstream branch master, + commit 5dcd2d0ad02ff12c76355ef4f40947c1857ac482. +- stdlib: Move _IO_cleanup to call_function_static_weak +- elf: Do not assume symbol order on tst-audit25{a,b} +- time: Use 64 bit time on tzfile +- nscd: Use 64 bit time_t on libc nscd routines (BZ# 29402) +- nis: Build libnsl with 64 bit time_t +- realloc: Return unchanged if request is within usable size +- Linux: Consolidate typesizes.h +- Linux: Make generic fcntl.h the default one +- Linux: make generic xstatver.h the default one +- Linux: Remove generic sysdep +- Linux: Assume and consolidate shutdown wire-up syscall +- Linux: Assume and consolidate listen wire-up syscall +- Linux: Assume and consolidate socketpair wire-up syscall +- Linux: Assume and consolidate socket wire-up syscall +- Linux: Assume and consolidate bind wire-up syscall +- Linux: consolidate ____longjmp_chk +- Linux: consolidate sendfile implementation +- Linux: consolidate unlink implementation +- Linux: consolidate symlink implementation +- Linux: consolidate rmdir implementation +- Linux: consolidate readlink implementation +- Linux: consolidate mkdir implementation +- Linux: consolidate link implementation +- Linux: consolidate lchown implementation +- Linux: consolidate inotify_init implementation +- Lninux: consolidate epoll_create implementation +- Linux: consolidate dup2 implementation +- Linux: consolidate chown implementation +- Linux: consolidate chmod implementation +- linux: Consolidate dl-origin.c +- linux: Use long int for syscall return value +- LoongArch: Use medium cmodel build libc_nonshared.a. +- x86_64: State assembler is being tested on sysdeps/x86/configure +- configure: Remove AS check +- configure: Remove check if ld is GNU +- configure: Remove check if as is GNU +- configure: Move locale tools early + +* Mon Dec 05 2022 Arjun Shankar - 2.36.9000-16 +- Auto-sync with upstream branch master, + commit 8fb923ddc38dd5f4bfac4869d70fd80483fdb87a: +- hurd: Make getrandom cache the server port +- powerpc64: Remove old strncmp optimization +- x86-64 strncpy: Properly handle the length parameter [BZ# 29839] +- x86-64 strncat: Properly handle the length parameter [BZ# 24097] +- ARC: update definitions in elf/elf.h +- scripts: Add "|" operator support to glibcpp's parsing +- Apply asm redirections in syslog.h before first use [BZ #27087] +- LoongArch: Add support for ilogb[f] +- LoongArch: Add support for scalb[f] +- LoongArch: Add support for scalbn[f] +- LoongArch: Use __builtin_logb{,f} with GCC >= 13 +- Use GCC builtins for logb functions if desired. +- LoongArch: Use __builtin_llrint{,f} with GCC >= 13 +- Use GCC builtins for llrint functions if desired. +- LoongArch: Use __builtin_lrint{,f} with GCC >= 13 +- Use GCC builtins for lrint functions if desired. +- LoongArch: Use __builtin_rint{,f} with GCC >= 13 + +* Mon Nov 28 2022 Florian Weimer - 2.36.9000-15 +- Auto-sync with upstream branch master, + commit f704192911c6c7b65a54beab3ab369fca7609a5d: +- x86/fpu: Factor out shared avx2/avx512 code in svml_{s|d}_wrapper_impl.h +- x86/fpu: Cleanup code in svml_{s|d}_wrapper_impl.h +- x86/fpu: Reformat svml_{s|d}_wrapper_impl.h +- x86/fpu: Fix misspelled evex512 section in variety of svml files +- x86/fpu: Add missing ISA sections to variety of svml files +- stdio-common: Add missing dependencies (bug 29780) +- i386: Avoid rely on linker optimization to avoid relocation +- elf: Fix rtld-audit trampoline for aarch64 +- Define in_int32_t_range to check if the 64 bit time_t syscall should be used + +* Mon Nov 14 2022 Arjun Shankar - 2.36.9000-14 +- Auto-sync with upstream branch master, + commit 94628de77888c3292fc103840731ff85f283368e: +- elf/tst-tlsopt-powerpc fails when compiled with -mcpu=power10 (BZ# 29776) +- LoongArch: Hard Float Support for fmaximum_mag_num{f/ }, fminimum_mag_num{f/ }. +- LoongArch: Hard Float Support for fmaximum_mag{f/ }, fminimum_mag{f/ }. +- LoongArch: Hard Float Support for fmaxmag{f/ }, fminmag{f/ }. +- LoongArch: Hard Float Support for fmaximum_num{f/ }, fminimum_num{f/ }. +- LoongArch: Hard Float Support for fmaximum{f/ }, fminimum{f/ }. +- LoongArch: Hard Float Support for float-point classification functions. +- LoongArch: Use __builtin_{fma, fmaf} to implement function {fma, fmaf}. + +* Thu Nov 10 2022 Florian Weimer - 2.36.9000-13 +- Auto-sync with upstream branch master, + commit 22a46dee24351fd5f4f188ad80554cad79c82524: +- Linux: Support __IPC_64 in sysvctl *ctl command arguments (bug 29771) +- riscv: Get level 3 cache's information +- debug: Fix typo in tests-unsupported rule +- iconvdata/tst-table-charmap.sh: remove handling of old, borrowed format +- Makerules: Generate shlib.lds with -fuse-ld=bfd +- x86: Add avx2 optimized functions for the wchar_t strcpy family +- x86: Add evex optimized functions for the wchar_t strcpy family +- x86: Optimize and shrink st{r|p}{n}{cat|cpy}-avx2 functions +- x86: Optimize and shrink st{r|p}{n}{cat|cpy}-evex functions +- benchtests: Make str{n}{cat|cpy} benchmarks output json +- x86: Use VMM API in memcmpeq-evex.S and minor changes +- x86: Use VMM API in memcmp-evex-movbe.S and minor changes +- string: Add len=0 to {w}memcmp{eq} tests and benchtests +- Linux: Add ppoll fortify symbol for 64 bit time_t (BZ# 29746) +- hurd: Add sigtimedwait and sigwaitinfo support + +* Mon Nov 07 2022 DJ Delorie - 2.36.9000-12 +- Auto-sync with upstream branch master, + commit 8d291eabd541029d7ac705cc1ea112c58dfbb05f. +- Apply asm redirection in gmp.h before first use +- Rewrite find_cxx_header config configure.ac +- elf/tlsdeschtab.h: Add the Malloc return value check in _dl_make_tlsdesc_dynamic() +- elf: Disable some subtests of ifuncmain1, ifuncmain5 for !PIE +- posix: Make posix_spawn extensions available by default +- x86_64: Implement evex512 version of strrchr and wcsrchr +- elf: Introduce and call_tls_init_tp (bug 29249) +- LoongArch: Fix ABI related macros in elf.h to keep consistent with binutils[1]. +- scripts/glibcelf.py: Properly report parsing failures +- elf: Rework exception handling in the dynamic loader [BZ #25486] +- linux: Drop useless include from fstatat.c +- Fix OOB read in stdlib thousand grouping parsing [BZ #29727] +- linux: Fix fstatat on MIPSn64 (BZ #29730) +- elf: Remove allocate use on _dl_debug_printf +- nptl: Fix pthread_create.c build with clang +- allocate_once: Apply asm redirection before first use +- alloc_buffer: Apply asm redirection before first use +- configure: Use -Wno-ignored-attributes if compiler warns about multiple aliases +- Disable use of -fsignaling-nans if compiler does not support it +- intl: Fix clang -Wunused-but-set-variable on plural.c +- Apply asm redirection in not-cancel before first use +- malloc: Use uintptr_t for pointer alignment +- Use uintptr_t in fts for pointer alignment +- Fix build with GCC 13 _FloatN, _FloatNx built-in functions +- elf: Build tst-relr-mod[34]a.so with $(LDFLAGS-rpath-ORIGIN) +- x86-64: Improve evex512 version of strlen functions +- Correctly determine libc.so 'OUTPUT_FORMAT' when cross-compiling. +- Remove unused scratch_buffer_dupfree +- Fix elf/tst-dlmopen-twice not to exhaust static TLS +- Use uintptr_t in string/tester for pointer alignment +- stdlib/strfrom: Add copysign to fix NAN issue on riscv (BZ #29501) +- Fix resource/bug-ulimit1 test +- Fix missing NUL terminator in stdio-common/scanf13 test +- Fix off-by-one OOB read in elf/tst-tls20 +- elf: Fix alloca size in _dl_debug_vdprintf +- malloc: Use uintptr_t in alloc_buffer +- Fix invalid pointer dereference in wcpcpy_chk +- Fix invalid pointer dereference in wcscpy_chk +- aarch64: Fix the extension header write in getcontext and swapcontext +- aarch64: Don't build wordcopy +- scripts: Use bool in tunables initializer +- longlong.h: update from GCC for LoongArch clz/ctz support + +* Thu Oct 27 2022 Patsy Griffin - 2.36.9000-11 +- Auto-sync with upstream branch master, + commit 6f360366f7f76b158a0f4bf20d42f2854ad56264. +- elf: Introduce to _dl_call_fini +- ld.so: Export tls_init_tp_called as __rtld_tls_init_tp_called +- scripts/localplt.awk: Handle DT_JMPREL with empty PLT (for C-SKY) +- Remove lingering libSegfault Makefile entries +- aarch64: Use memcpy_simd as the default memcpy +- aarch64: Cleanup memset ifunc +- elf: Reinstate on DL_DEBUG_BINDINGS _dl_lookup_symbol_x +- x86_64: Implement evex512 version of strchrnul, strchr and wcschr +- linux: Fix generic struct_stat for 64 bit time (BZ# 29657) +- manual: Add missing % in int conversion list +- Avoid undefined behaviour in ibm128 implementation of llroundl (BZ #29488) +- Remove all assembly optimizations for htonl and htons +- Remove htonl.S for i386/x86_64 +- Fix BZ #29463 in the ibm128 implementation of y1l too +- Add ADDRB from Linux 6.0 to bits/termios-c_cflag.h +- x86: Use `testb` for FSRM check in memmove-vec-unaligned-erms +- x86: Use `testb` for case-locale check in str{n}casecmp-sse42 +- x86: Use `testb` for case-locale check in str{n}casecmp-sse2 +- x86: Use `testb` for case-locale check in str{n}casecmp-avx2 +- x86: Add support for VEC_SIZE == 64 in strcmp-evex.S impl +- x86: Remove AVX512-BVMI2 instruction from strrchr-evex.S +- sysdeps: arm: Fix preconfigure script for ARMv8/v9 targets [BZ #29698] +- nis: Fix nis_print_directory +- linux: Avoid shifting a negative signed on POSIX timer interface +- Bench: Improve benchtests for memchr, strchr, strnlen, strrchr +- x86: Optimize strrchr-evex.S and implement with VMM headers +- x86: Optimize memrchr-evex.S +- x86: Optimize strnlen-evex.S and implement with VMM headers +- x86: Shrink / minorly optimize strchr-evex and implement with VMM headers +- x86: Optimize memchr-evex.S and implement with VMM headers +- x86_64: Implement evex512 version of memchr, rawmemchr and wmemchr +- String: Improve test coverage for memchr +- Use PTR_MANGLE and PTR_DEMANGLE unconditionally in C sources +- Introduce , extracted from +- x86-64: Move LP_SIZE definition to its own header +- math: Fix asin and acos invalid exception with old gcc +- x86: Update strlen-evex-base to use new reg/vec macros. +- x86: Remove now unused vec header macros. +- x86: Update memset to use new VEC macros +- x86: Update memmove to use new VEC macros +- x86: Update memrchr to use new VEC macros +- x86: Update VEC macros to complete API for evex/evex512 impls +- elf: Do not completely clear reused namespace in dlmopen (bug 29600) +- malloc: Switch global_max_fast to uint8_t +- Add NT_S390_PV_CPU_DATA from Linux 6.0 to elf.h +- Add AArch64 HWCAP2_EBF16 from Linux 6.0 to bits/hwcap.h +- String: Improve test coverage for memchr +- elf: Remove -fno-tree-loop-distribute-patterns usage on dl-support +- socket: Use offsetof in SUN_LEN (bug 29578) +- Expose all MAP_ constants in unconditionally (bug 29375) +- LoongArch: Fix the condition to use PC-relative addressing in start.S +- arm: Enable USE_ATOMIC_COMPILER_BUILTINS (BZ #24774) +- csu: Disable stack protector for static-reloc for static-pie +- NEWS: Fix grammar +- elf: Simplify output of hwcap subdirectories in ld.so help +- elf: Remove _dl_string_hwcap +- Add NEWS entry for legacy hwcaps removal +- elf: Remove hwcap and bits_hwcap fields from struct cache_entry +- elf: Remove hwcap parameter from add_to_cache signature +- elf: Remove legacy hwcaps support from ldconfig +- elf: Remove legacy hwcaps support from the dynamic loader +- x86_64: Remove platform directory library loading test +- Update to Unicode 15.0.0 [BZ #29604] +- Update kernel version to 6.0 in header constant tests +- x86: Fix -Os build (BZ #29576) +- sunrpc: Suppress GCC -Os warning on user2netname +- rt: Initialize mq_send input on tst-mqueue{5,6} +- posix: Suppress -Os may be used uninitialized warnings on regexec +- posix: Suppress -Os warnings on fnmatch +- locale: prevent maybe-uninitialized errors with -Os [BZ #19444] +- Regenerate sysdeps/mach/hurd/bits/errno.h +- Update syscall lists for Linux 6.0 +- nscd: Drop local address tuple variable [BZ #29607] +- Use Linux 6.0 in build-many-glibcs.py +- x86-64: Require BMI1/BMI2 for AVX2 strrchr and wcsrchr implementations +- x86-64: Require BMI2 and LZCNT for AVX2 memrchr implementation +- x86-64: Require BMI2 for AVX2 (raw|w)memchr implementations +- x86-64: Require BMI2 for AVX2 wcs(n)cmp implementations +- x86-64: Require BMI2 for AVX2 strncmp implementation +- x86-64: Require BMI2 for AVX2 strcmp implementation +- x86-64: Require BMI2 for AVX2 str(n)casecmp implementations +- x86: include BMI1 and BMI2 in x86-64-v3 level +- x86: Cleanup pthread_spin_{try}lock.S +- Benchtests: Add bench for pthread_spin_{try}lock and mutex_trylock + +* Mon Oct 17 2022 Carlos O'Donell - 2.36.9000-10 +- Enable ELF DT_HASH for shared objects and the dynamic loader (#2129358) + +* Mon Oct 03 2022 DJ Delorie - 2.36.9000-9 +- Auto-sync with upstream branch master, + commit 114e299ca66353fa7be1ee45bb4e1307d3de1fa2. +- x86: Remove .tfloat usage +- nptl: Convert tst-setuid2 to test-driver +- support: Add xpthread_cond_signal wrapper +- hppa: Fix initialization of dp register [BZ 29635] +- Fix iseqsig for _FloatN and _FloatNx in C++ with GCC 13 +- malloc: Do not clobber errno on __getrandom_nocancel (BZ #29624) +- stdlib: Fix __getrandom_nocancel type and arc4random usage (BZ #29638) +- LoongArch: Add static PIE support +- Benchtest: Add additional benchmarks for strlen and strnlen +- x86: Fix wcsnlen-avx2 page cross length comparison [BZ #29591] +- Update _FloatN header support for C++ in GCC 13 +- hurd: Fix typo +- get_nscd_addresses: Fix subscript typos [BZ #29605] +- hurd: Increase SOMAXCONN to 4096 +- Use atomic_exchange_release/acquire + +* Fri Sep 23 2022 Patsy Griffin - 2.36.9000-8 +- Auto-sync with upstream branch master, + commit c02e29a0ba47d636281e1a026444a1a0a254aa12. +- nss: Use shared prefix in IPv4 address in tst-reload1 +- nss: Enhance tst-reload1 coverage and logging +- Use C11 atomics instead of atomic_decrement_and_test +- Use C11 atomics instead of atomic_increment(_val) +- Use C11 atomics instead of atomic_and/or +- malloc: Print error when oldsize is not equal to the current size. +- Use '%z' instead of '%Z' on printf functions +- elf: Extract glibcelf constants from +- scripts: Enhance glibcpp to do basic macro processing +- scripts: Extract glibcpp.py from check-obsolete-constructs.py +- riscv: Remove RV32 floating point functions +- riscv: Consolidate the libm-test-ulps +- hurd: Fix SIOCADD/DELRT ioctls +- hurd: Drop struct rtentry and in6_rtmsg +- hurd: Add _IOT_ifrtreq to +- elf: Use C11 atomics on _dl_mcount +- hurd: Use IF_NAMESIZE rather than IFNAMSIZ +- hurd: Add ifrtreq structure to net/route.h +- hppa: undef __ASSUME_SET_ROBUST_LIST +- linux: Use same type for MMAP2_PAGE_UNIT +- m68k: Enforce 4-byte alignment on internal locks (BZ #29537) +- nss: Fix tst-nss-files-hosts-long on single-stack hosts (bug 24816) +- nss: Implement --no-addrconfig option for getent +- gconv: Use 64-bit interfaces in gconv_parseconfdir (bug 29583) +- elf: Implement force_first handling in _dl_sort_maps_dfs (bug 28937) +- Linux: Do not skip d_ino == 0 entries in readdir, readdir64 (bug 12165) +- hurd: Factorize at/non-at functions +- tst-sprintf-errno: Update Hurd message length +- RISC-V: Allow long jumps to __syscall_error +- hurd: Make readlink* just reopen the file used for stat +- hurd: Fix readlink() hanging on fifo +- Fix BRE typos in check-safety.sh +- Makerules: fix MAKEFLAGS assignment for upcoming make-4.4 [BZ# 29564] +- Use relaxed atomics since there is no MO dependence + +* Wed Sep 14 2022 Florian Weimer - 2.36.9000-7 +- Remove .annobin* symbols from ld.so (#2126477) + +* Tue Sep 13 2022 Florian Weimer - 2.36.9000-6 +- Drop glibc-deprecated-selinux-makedb.patch. Upstream has been ported + to new libselinux. +- Auto-sync with upstream branch master, + commit f278835f594740f5913001430641cf1da4878670: +- makedb: fix build with libselinux >= 3.1 (Bug 26233) +- tst-sprintf-errno: Update Hurd message output +- Use C11 atomics instead of atomic_decrement(_val) +- Use C11 atomics instead atomic_add(_zero) +- mktime: improve heuristic for ca-1986 Indiana DST +- Assume HAVE_TZSET in time/mktime.c +- elf: Fix hwcaps string size overestimation +- errlist: add missing entry for EDEADLOCK (bug 29545) +- Do not define static_assert or thread_local in headers for C2x +- malloc: Use C11 atomics rather than atomic_exchange_and_add +- Add NEWS entry for CVE-2022-39046 +- elf: Rename _dl_sort_maps parameter from skip to force_first +- scripts/dso-ordering-test.py: Generate program run-time dependencies +- math: x86: Use prefix for FP_INIT_ROUNDMODE +- scripts/build-many-glibcs.py: Use https:// for sourceware.org Git clones +- debug: test for more required cacellation points (BZ# 29274) +- elf.h: Remove duplicate definition of VER_FLG_WEAK +- syslog: Remove extra whitespace between timestamp and message (BZ#29544) +- LoongArch: Add soft float support. +- elf: Restore how vDSO dependency is printed with LD_TRACE_LOADED_OBJECTS (BZ #29539) +- nptl: x86_64: Use same code for CURRENT_STACK_FRAME and stackinfo_get_sp + +* Mon Sep 05 2022 Arjun Shankar - 2.36.9000-5 +- Co-Authored-By: Benjamin Herrenschmidt +- Retain .gnu_debuglink section in libc.so.6 (#2090744) +- Remove redundant ld.so debuginfo file (#2090744) + +* Tue Aug 30 2022 DJ Delorie - 2.36.9000-4 +- Auto-sync with upstream branch master, + commit c7509d49c4e8fa494120c5ead21338559dad16f5. +- Apply asm redirections in wchar.h before first use +- resolv: Fix building tst-resolv-invalid-cname for earlier C standards +- syslog: Fix large messages (BZ#29536) +- posix: Fix macro expansion producing 'defined' has undefined behavior +- stdlib: Fix macro expansion producing 'defined' has undefined behavior +- S390: Always use svc 0 +- nss_dns: Rewrite _nss_dns_gethostbyname4_r using current interfaces +- resolv: Add new tst-resolv-invalid-cname +- nss_dns: In gaih_getanswer_slice, skip strange aliases (bug 12154) +- nss_dns: Rewrite getanswer_r to match getanswer_ptr (bug 12154, bug 29305) +- nss_dns: Remove remnants of IPv6 address mapping +- nss_dns: Rewrite _nss_dns_gethostbyaddr2_r and getanswer_ptr +- nss_dns: Split getanswer_ptr from getanswer_r +- resolv: Add DNS packet parsing helpers geared towards wire format +- resolv: Add internal __ns_name_length_uncompressed function +- resolv: Add the __ns_samebinaryname function +- resolv: Add internal __res_binary_hnok function +- resolv: Add tst-resolv-aliases +- resolv: Add tst-resolv-byaddr for testing reverse lookup +- LoongArch: Use __builtin_{fmax,fmaxf,fmin,fminf} with GCC >= 13 +- LoongArch: Fix ptr mangling/demangling features. +- nscd: Fix netlink cache invalidation if epoll is used [BZ #29415] +- Add test for bug 29530 +- Makeconfig: Set pie-ccflag to -fPIE by default [BZ# 29514] +- hurd: Fix vm_size_t incoherencies +- mach: Make xpg_strerror_r set a message on error +- mach: Fix incoherency between perror and strerror +- elf: Call __libc_early_init for reused namespaces (bug 29528) +- csu: Change start code license to have link exception +- s390: Move hwcaps/platform names out of _rtld_global_ro +- Revert "Detect ld.so and libc.so version inconsistency during startup" +- Add NT_LOONGARCH_* from Linux 5.19 to elf.h +- Detect ld.so and libc.so version inconsistency during startup +- Merge getopt patch from Gnulib +- Merge _GL_UNUSED C23 patch from Gnulib +- LoongArch: Fix dl-machine.h code formatting. +- scripts/glibcelf.py: Add hashing support +- hurd: Fix starting static binaries with stack protection enabled +- htl: Make pthread*_cond_timedwait register wref before releasing mutex +- htl: make __pthread_hurd_cond_timedwait_internal check mutex is held +- Add AArch64 HWCAP2_* constants from Linux 5.19 + +* Mon Aug 22 2022 Arjun Shankar - 2.36.9000-3 +- Auto-sync with upstream branch master, + commit a727220b37efc9d4d558a77c5fc57f3af99a4829: +- Add AGROUP from Linux 5.19 to sys/acct.h, remove Alpha version (bug 29502) +- alpha: Fix generic brk system call emulation in __brk_call (bug 29490) +- hurd: Assume non-suid during bootstrap + +* Thu Aug 18 2022 Patsy Griffin - 2.36.9000-2 +- Auto-sync with upstream branch master, + commit 9125e43daf92e3d4e69044a54b9fe9ed88c861ad. +- Use binutils 2.39 branch in build-many-glibcs.py +- S390: Fix werror=unused-variable in ifunc-impl-list.c. +- Ensure calculations happen with desired rounding mode in y1lf128 +- localedata: Convert French language locales (fr_*) to UTF-8 +- Linux: Fix enum fsconfig_command detection in +- elf: Run tst-audit-tlsdesc, tst-audit-tlsdesc-dlopen everywhere +- Move ip_mreqn structure from Linux to generic +- malloc: Do not use MAP_NORESERVE to allocate heap segments +- Linux: Terminate subprocess on late failure in tst-pidfd (bug 29485) +- non-linux: bits/in.h: Add more RFC options + +* Mon Aug 15 2022 Florian Weimer - 2.36.9000-1 +- Auto-sync with upstream branch master, + commit 453b88efe6fa79f5c7c6fccc3a520c75fdd43074: +- arm: Remove nested functionf rom relocate_pc24 +- linux: Fix sys/mount.h usage with kernel headers +- linux: Use compile_c_snippet to check linux/mount.h availability +- linux: Mimic kernel defition for BLOCK_SIZE +- linux: Use compile_c_snippet to check linux/pidfd.h availability +- glibcextract.py: Add compile_c_snippet +- LoongArch: Add pointer mangling support. +- AArch64: Fix typo in sve configure check (BZ# 29394) +- libio: Improve performance of IO locks +- tst-process_madvise: Check process_madvise-syscall support. +- elf.h: Add ELFCOMPRESS_ZSTD +- inet: Turn __ivaliduser into a compatibility symbol +- x86: Fix `#define STRCPY` guard in strcpy-sse2.S +- elf: Replace `strcpy` call with `memcpy` [BZ #29454] +- soft-fp: Add fixhf[uns][di|si] and float[uns][di|si]hf +- i386: Use cmpl instead of cmp +- i386: Use fldt instead of fld on e_logl.S +- i386: Replace movzx with movzbl +- dlfcn: Pass caller pointer to static dlopen implementation (bug 29446) +- malloc: Correct the documentation of the top_pad default +- i386: Remove RELA support +- arm: Remove RELA support +- Remove ldd libc4 support +- Assume only FLAG_ELF_LIBC6 suport +- Remove left over LD_LIBRARY_VERSION usages +- Linux: Remove exit system call from _exit +- LoongArch: Add vdso support for gettimeofday. +- Update kernel version to 5.19 in header constant tests +- assert: Do not use stderr in libc-internal assert +- nptl: Remove uses of assert_perror +- stdio: Clean up __libc_message after unconditional abort +- Update syscall lists for Linux 5.19 +- Use Linux 5.19 in build-many-glibcs.py +- socket: Check lengths before advancing pointer in CMSG_NXTHDR +- Don't use unsupported format string in ld.so (bug 29427) +- htl: Let pthread_self and cancellability called early +- stdlib: Simplify arc4random_uniform +- malloc: Use __getrandom_nocancel during tcache initiailization +- Remove spurious references to _dl_open_hook +- Open master branch for glibc 2.37 development + +* Wed Aug 03 2022 Carlos O'Donell - 2.36-1 +- Auto-sync with upstream branch release/2.36/master, + commit 33f1b4c1452b33991e670f636ebe98b90a405e10: +- wcsmbs: Add missing test-c8rtomb/test-mbrtoc8 dependency +- stdlib: Suppress gcc diagnostic that char8_t is a keyword in C++20 in uchar.h. +- Create ChangeLog.old/ChangeLog.25. (tag: glibc-2.36) +- Prepare for glibc 2.36 release. +- Update install.texi, and regenerate INSTALL. +- Update NEWS bug list. +- Update libc.pot for 2.36 release. +- tst-pidfd.c: UNSUPPORTED if we get EPERM on valid pidfd_getfd call +- stdlib: Tuned down tst-arc4random-thread internal parameters +- LoongArch: Add greg_t and gregset_t. +- LoongArch: Fix VDSO_HASH and VDSO_NAME. +- riscv: Update rv64 libm test ulps +- riscv: Update nofpu libm test ulps