import mdadm-4.2-6.el9

c9 imports/c9/mdadm-4.2-6.el9
CentOS Sources 2 years ago committed by MSVSphere Packaging Team
commit 93f2298110

1
.gitignore vendored

@ -0,0 +1 @@
SOURCES/mdadm-4.2.tar.xz

@ -0,0 +1 @@
27f240cff200e00c28a486a028bcdb14f67f8790 SOURCES/mdadm-4.2.tar.xz

@ -0,0 +1,47 @@
From f1cc8ab9ab6a92c3cd94ab7590b46285e214681e Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Tue, 15 Mar 2022 09:30:30 +0100
Subject: [PATCH 01/12] Unify error message.
Provide the same error message for the same error that can occur in Grow.c and super-intel.c.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Grow.c | 4 ++--
super-intel.c | 4 ++--
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/Grow.c b/Grow.c
index 9c6fc95e..9a947204 100644
--- a/Grow.c
+++ b/Grow.c
@@ -1001,8 +1001,8 @@ int remove_disks_for_takeover(struct supertype *st,
rv = 1;
sysfs_free(arrays);
if (rv) {
- pr_err("Error. Cannot perform operation on /dev/%s\n", st->devnm);
- pr_err("For this operation it MUST be single array in container\n");
+ pr_err("Error. Cannot perform operation on %s- for this operation "
+ "it MUST be single array in container\n", st->devnm);
return rv;
}
}
diff --git a/super-intel.c b/super-intel.c
index d5fad102..5ffa7636 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -11683,8 +11683,8 @@ enum imsm_reshape_type imsm_analyze_change(struct supertype *st,
struct imsm_super *mpb = super->anchor;
if (mpb->num_raid_devs > 1) {
- pr_err("Error. Cannot perform operation on %s- for this operation it MUST be single array in container\n",
- geo->dev_name);
+ pr_err("Error. Cannot perform operation on %s- for this operation "
+ "it MUST be single array in container\n", geo->dev_name);
change = -1;
}
}
--
2.31.1

@ -0,0 +1,33 @@
From 5ce5a15f0bf007e850e15259bba4f53736605fb2 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 25 Mar 2022 12:48:59 +0100
Subject: [PATCH 02/12] mdadm: Fix double free
If there was a size mismatch after creation it would get fixed on grow
in imsm_fix_size_mismatch(), but due to double free "double free or corruption (fasttop)"
error occurs and grow cannot proceed.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/super-intel.c b/super-intel.c
index 5ffa7636..6ff336ee 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -11783,9 +11783,8 @@ static int imsm_fix_size_mismatch(struct supertype *st, int subarray_index)
st->update_tail = &st->updates;
} else {
imsm_sync_metadata(st);
+ free(update);
}
-
- free(update);
}
ret_val = 0;
exit:
--
2.31.1

@ -0,0 +1,83 @@
From fea026b4849182fc8413014c81456e7215af28d9 Mon Sep 17 00:00:00 2001
From: Mateusz Kusiak <mateusz.kusiak@intel.com>
Date: Wed, 23 Mar 2022 15:05:19 +0100
Subject: [PATCH 03/12] Grow_reshape: Add r0 grow size error message and update
man
Grow size on r0 is not supported for imsm and native metadata.
Add proper error message.
Update man for proper use of --size.
Signed-off-by: Mateusz Kusiak <mateusz.kusiak@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Grow.c | 6 ++++++
mdadm.8.in | 19 ++++++++++++-------
2 files changed, 18 insertions(+), 7 deletions(-)
diff --git a/Grow.c b/Grow.c
index 9a947204..aa72490b 100644
--- a/Grow.c
+++ b/Grow.c
@@ -1998,6 +1998,12 @@ int Grow_reshape(char *devname, int fd,
goto release;
}
+ if (array.level == 0) {
+ pr_err("Component size change is not supported for RAID0\n");
+ rv = 1;
+ goto release;
+ }
+
if (reshape_super(st, s->size, UnSet, UnSet, 0, 0, UnSet, NULL,
devname, APPLY_METADATA_CHANGES,
c->verbose > 0)) {
diff --git a/mdadm.8.in b/mdadm.8.in
index be902dba..e2a42425 100644
--- a/mdadm.8.in
+++ b/mdadm.8.in
@@ -459,7 +459,8 @@ number of spare devices.
.TP
.BR \-z ", " \-\-size=
-Amount (in Kilobytes) of space to use from each drive in RAID levels 1/4/5/6.
+Amount (in Kilobytes) of space to use from each drive in RAID levels 1/4/5/6/10
+and for RAID 0 on external metadata.
This must be a multiple of the chunk size, and must leave about 128Kb
of space at the end of the drive for the RAID superblock.
If this is not specified
@@ -478,10 +479,19 @@ To guard against this it can be useful to set the initial size
slightly smaller than the smaller device with the aim that it will
still be larger than any replacement.
+This option can be used with
+.B \-\-create
+for determining initial size of an array. For external metadata,
+it can be used on a volume, but not on a container itself.
+Setting initial size of
+.B RAID 0
+array is only valid for external metadata.
+
This value can be set with
.B \-\-grow
-for RAID level 1/4/5/6 though
+for RAID level 1/4/5/6/10 though
DDF arrays may not be able to support this.
+RAID 0 array size cannot be changed.
If the array was created with a size smaller than the currently
active drives, the extra space can be accessed using
.BR \-\-grow .
@@ -501,11 +511,6 @@ problems the array can be made bigger again with no loss with another
.B "\-\-grow \-\-size="
command.
-This value cannot be used when creating a
-.B CONTAINER
-such as with DDF and IMSM metadata, though it perfectly valid when
-creating an array inside a container.
-
.TP
.BR \-Z ", " \-\-array\-size=
This is only meaningful with
--
2.31.1

@ -0,0 +1,67 @@
From cf9a109209aad285372b67306d54118af6fc522b Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Fri, 14 Jan 2022 16:44:33 +0100
Subject: [PATCH 04/12] udev: adapt rules to systemd v247
New events have been added in kernel 4.14 ("bind" and "unbind").
Systemd maintainer suggests to modify "add|change" branches.
This patches implements their suggestions. There is no issue yet because
new event types are not used in md.
Please see systemd announcement for details[1].
[1] https://lists.freedesktop.org/archives/systemd-devel/2020-November/045646.html
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
udev-md-raid-arrays.rules | 2 +-
udev-md-raid-assembly.rules | 5 +++--
udev-md-raid-safe-timeouts.rules | 2 +-
3 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/udev-md-raid-arrays.rules b/udev-md-raid-arrays.rules
index 13c9076e..2967ace1 100644
--- a/udev-md-raid-arrays.rules
+++ b/udev-md-raid-arrays.rules
@@ -3,7 +3,7 @@
SUBSYSTEM!="block", GOTO="md_end"
# handle md arrays
-ACTION!="add|change", GOTO="md_end"
+ACTION=="remove", GOTO="md_end"
KERNEL!="md*", GOTO="md_end"
# partitions have no md/{array_state,metadata_version}, but should not
diff --git a/udev-md-raid-assembly.rules b/udev-md-raid-assembly.rules
index d668cddd..39b4344b 100644
--- a/udev-md-raid-assembly.rules
+++ b/udev-md-raid-assembly.rules
@@ -30,8 +30,9 @@ LABEL="md_inc"
# remember you can limit what gets auto/incrementally assembled by
# mdadm.conf(5)'s 'AUTO' and selectively whitelist using 'ARRAY'
-ACTION=="add|change", IMPORT{program}="BINDIR/mdadm --incremental --export $devnode --offroot $env{DEVLINKS}"
-ACTION=="add|change", ENV{MD_STARTED}=="*unsafe*", ENV{MD_FOREIGN}=="no", ENV{SYSTEMD_WANTS}+="mdadm-last-resort@$env{MD_DEVICE}.timer"
+ACTION!="remove", IMPORT{program}="BINDIR/mdadm --incremental --export $devnode --offroot $env{DEVLINKS}"
+ACTION!="remove", ENV{MD_STARTED}=="*unsafe*", ENV{MD_FOREIGN}=="no", ENV{SYSTEMD_WANTS}+="mdadm-last-resort@$env{MD_DEVICE}.timer"
+
ACTION=="remove", ENV{ID_PATH}=="?*", RUN+="BINDIR/mdadm -If $name --path $env{ID_PATH}"
ACTION=="remove", ENV{ID_PATH}!="?*", RUN+="BINDIR/mdadm -If $name"
diff --git a/udev-md-raid-safe-timeouts.rules b/udev-md-raid-safe-timeouts.rules
index 12bdcaa8..2e185cee 100644
--- a/udev-md-raid-safe-timeouts.rules
+++ b/udev-md-raid-safe-timeouts.rules
@@ -50,7 +50,7 @@ ENV{DEVTYPE}!="partition", GOTO="md_timeouts_end"
IMPORT{program}="/sbin/mdadm --examine --export $devnode"
-ACTION=="add|change", \
+ACTION!="remove", \
ENV{ID_FS_TYPE}=="linux_raid_member", \
ENV{MD_LEVEL}=="raid[1-9]*", \
TEST=="/sys/block/$parent/device/timeout", \
--
2.31.1

@ -0,0 +1,252 @@
From 83a379cfbd283b387919fe05d44eb4c49e155ad6 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Mon, 21 Feb 2022 13:05:20 +0100
Subject: [PATCH 05/12] Replace error prone signal() with sigaction()
Up to this date signal() was used which implementation could vary [1].
Sigaction() call is preferred. This commit introduces replacement
from signal() to sigaction() by the use of signal_s() wrapper.
Also remove redundant signal.h header includes.
[1] https://man7.org/linux/man-pages/man2/signal.2.html
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Grow.c | 4 ++--
Monitor.c | 5 +++--
managemon.c | 1 -
mdadm.h | 22 ++++++++++++++++++++++
mdmon.c | 1 -
monitor.c | 1 -
probe_roms.c | 6 +++---
raid6check.c | 25 +++++++++++++++----------
util.c | 1 -
9 files changed, 45 insertions(+), 21 deletions(-)
diff --git a/Grow.c b/Grow.c
index aa72490b..18c5719b 100644
--- a/Grow.c
+++ b/Grow.c
@@ -26,7 +26,6 @@
#include <sys/mman.h>
#include <stddef.h>
#include <stdint.h>
-#include <signal.h>
#include <sys/wait.h>
#if ! defined(__BIG_ENDIAN) && ! defined(__LITTLE_ENDIAN)
@@ -3566,7 +3565,8 @@ started:
fd = -1;
mlockall(MCL_FUTURE);
- signal(SIGTERM, catch_term);
+ if (signal_s(SIGTERM, catch_term) == SIG_ERR)
+ goto release;
if (st->ss->external) {
/* metadata handler takes it from here */
diff --git a/Monitor.c b/Monitor.c
index 30c031a2..c0ab5412 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -26,7 +26,6 @@
#include "md_p.h"
#include "md_u.h"
#include <sys/wait.h>
-#include <signal.h>
#include <limits.h>
#include <syslog.h>
#ifndef NO_LIBUDEV
@@ -435,8 +434,10 @@ static void alert(char *event, char *dev, char *disc, struct alert_info *info)
if (mp) {
FILE *mdstat;
char hname[256];
+
gethostname(hname, sizeof(hname));
- signal(SIGPIPE, SIG_IGN);
+ signal_s(SIGPIPE, SIG_IGN);
+
if (info->mailfrom)
fprintf(mp, "From: %s\n", info->mailfrom);
else
diff --git a/managemon.c b/managemon.c
index bb7334cf..0e9bdf00 100644
--- a/managemon.c
+++ b/managemon.c
@@ -106,7 +106,6 @@
#include "mdmon.h"
#include <sys/syscall.h>
#include <sys/socket.h>
-#include <signal.h>
static void close_aa(struct active_array *aa)
{
diff --git a/mdadm.h b/mdadm.h
index c7268a71..26e7e5cd 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -46,6 +46,7 @@ extern __off64_t lseek64 __P ((int __fd, __off64_t __offset, int __whence));
#include <string.h>
#include <syslog.h>
#include <stdbool.h>
+#include <signal.h>
/* Newer glibc requires sys/sysmacros.h directly for makedev() */
#include <sys/sysmacros.h>
#ifdef __dietlibc__
@@ -1729,6 +1730,27 @@ static inline char *to_subarray(struct mdstat_ent *ent, char *container)
return &ent->metadata_version[10+strlen(container)+1];
}
+/**
+ * signal_s() - Wrapper for sigaction() with signal()-like interface.
+ * @sig: The signal to set the signal handler to.
+ * @handler: The signal handler.
+ *
+ * Return: previous handler or SIG_ERR on failure.
+ */
+static inline sighandler_t signal_s(int sig, sighandler_t handler)
+{
+ struct sigaction new_act;
+ struct sigaction old_act;
+
+ new_act.sa_handler = handler;
+ new_act.sa_flags = 0;
+
+ if (sigaction(sig, &new_act, &old_act) == 0)
+ return old_act.sa_handler;
+
+ return SIG_ERR;
+}
+
#ifdef DEBUG
#define dprintf(fmt, arg...) \
fprintf(stderr, "%s: %s: "fmt, Name, __func__, ##arg)
diff --git a/mdmon.c b/mdmon.c
index c71e62c6..5570574b 100644
--- a/mdmon.c
+++ b/mdmon.c
@@ -56,7 +56,6 @@
#include <errno.h>
#include <string.h>
#include <fcntl.h>
-#include <signal.h>
#include <dirent.h>
#ifdef USE_PTHREADS
#include <pthread.h>
diff --git a/monitor.c b/monitor.c
index e0d3be67..b877e595 100644
--- a/monitor.c
+++ b/monitor.c
@@ -22,7 +22,6 @@
#include "mdmon.h"
#include <sys/syscall.h>
#include <sys/select.h>
-#include <signal.h>
static char *array_states[] = {
"clear", "inactive", "suspended", "readonly", "read-auto",
diff --git a/probe_roms.c b/probe_roms.c
index 7ea04c7a..94c80c2c 100644
--- a/probe_roms.c
+++ b/probe_roms.c
@@ -22,7 +22,6 @@
#include "probe_roms.h"
#include "mdadm.h"
#include <unistd.h>
-#include <signal.h>
#include <fcntl.h>
#include <sys/mman.h>
#include <sys/stat.h>
@@ -69,7 +68,8 @@ static int probe_address16(const __u16 *ptr, __u16 *val)
void probe_roms_exit(void)
{
- signal(SIGBUS, SIG_DFL);
+ signal_s(SIGBUS, SIG_DFL);
+
if (rom_fd >= 0) {
close(rom_fd);
rom_fd = -1;
@@ -98,7 +98,7 @@ int probe_roms_init(unsigned long align)
if (roms_init())
return -1;
- if (signal(SIGBUS, sigbus) == SIG_ERR)
+ if (signal_s(SIGBUS, sigbus) == SIG_ERR)
rc = -1;
if (rc == 0) {
fd = open("/dev/mem", O_RDONLY);
diff --git a/raid6check.c b/raid6check.c
index a8e6005b..99477761 100644
--- a/raid6check.c
+++ b/raid6check.c
@@ -24,7 +24,6 @@
#include "mdadm.h"
#include <stdint.h>
-#include <signal.h>
#include <sys/mman.h>
#define CHECK_PAGE_BITS (12)
@@ -130,30 +129,36 @@ void raid6_stats(int *disk, int *results, int raid_disks, int chunk_size)
}
int lock_stripe(struct mdinfo *info, unsigned long long start,
- int chunk_size, int data_disks, sighandler_t *sig) {
+ int chunk_size, int data_disks, sighandler_t *sig)
+{
int rv;
+
+ sig[0] = signal_s(SIGTERM, SIG_IGN);
+ sig[1] = signal_s(SIGINT, SIG_IGN);
+ sig[2] = signal_s(SIGQUIT, SIG_IGN);
+
+ if (sig[0] == SIG_ERR || sig[1] == SIG_ERR || sig[2] == SIG_ERR)
+ return 1;
+
if(mlockall(MCL_CURRENT | MCL_FUTURE) != 0) {
return 2;
}
- sig[0] = signal(SIGTERM, SIG_IGN);
- sig[1] = signal(SIGINT, SIG_IGN);
- sig[2] = signal(SIGQUIT, SIG_IGN);
-
rv = sysfs_set_num(info, NULL, "suspend_lo", start * chunk_size * data_disks);
rv |= sysfs_set_num(info, NULL, "suspend_hi", (start + 1) * chunk_size * data_disks);
return rv * 256;
}
-int unlock_all_stripes(struct mdinfo *info, sighandler_t *sig) {
+int unlock_all_stripes(struct mdinfo *info, sighandler_t *sig)
+{
int rv;
rv = sysfs_set_num(info, NULL, "suspend_lo", 0x7FFFFFFFFFFFFFFFULL);
rv |= sysfs_set_num(info, NULL, "suspend_hi", 0);
rv |= sysfs_set_num(info, NULL, "suspend_lo", 0);
- signal(SIGQUIT, sig[2]);
- signal(SIGINT, sig[1]);
- signal(SIGTERM, sig[0]);
+ signal_s(SIGQUIT, sig[2]);
+ signal_s(SIGINT, sig[1]);
+ signal_s(SIGTERM, sig[0]);
if(munlockall() != 0)
return 3;
diff --git a/util.c b/util.c
index 3d05d074..cc94f96e 100644
--- a/util.c
+++ b/util.c
@@ -35,7 +35,6 @@
#include <poll.h>
#include <ctype.h>
#include <dirent.h>
-#include <signal.h>
#include <dlfcn.h>
--
2.31.1

File diff suppressed because it is too large Load Diff

@ -0,0 +1,48 @@
From c23400377bb3d8e98e810cd92dba478dac1dff82 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 18 Mar 2022 09:26:05 +0100
Subject: [PATCH 07/12] mdadm: Update ReadMe
Instead of hardcoded config file path give reference to config manual.
Add missing monitordelay and homecluster parameters.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
ReadMe.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/ReadMe.c b/ReadMe.c
index 81399765..8f873c48 100644
--- a/ReadMe.c
+++ b/ReadMe.c
@@ -613,7 +613,6 @@ char Help_incr[] =
;
char Help_config[] =
-"The /etc/mdadm.conf config file:\n\n"
" The config file contains, apart from blank lines and comment lines that\n"
" start with a hash(#), array lines, device lines, and various\n"
" configuration lines.\n"
@@ -636,10 +635,12 @@ char Help_config[] =
" than a device must match all of them to be considered.\n"
"\n"
" Other configuration lines include:\n"
-" mailaddr, mailfrom, program used for --monitor mode\n"
-" create, auto used when creating device names in /dev\n"
-" homehost, policy, part-policy used to guide policy in various\n"
-" situations\n"
+" mailaddr, mailfrom, program, monitordelay used for --monitor mode\n"
+" create, auto used when creating device names in /dev\n"
+" homehost, homecluster, policy, part-policy used to guide policy in various\n"
+" situations\n"
+"\n"
+"For more details see mdadm.conf(5).\n"
"\n"
;
--
2.31.1

@ -0,0 +1,203 @@
From 24e075c659d0a8718aabefe5af4c97195a188af7 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 18 Mar 2022 09:26:06 +0100
Subject: [PATCH 08/12] mdadm: Update config man regarding default files and
multi-keyword behavior
Simplify default and alternative config file and directory location references
from mdadm(8) as references to mdadm.conf(5). Add FILE section in config man
and explain order and conditions in which default and alternative config files
and directories are used.
Update config man behavior regarding parsing order when multiple keywords/config
files are involved.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
mdadm.8.in | 30 +++++++++--------------
mdadm.conf.5.in | 65 ++++++++++++++++++++++++++++++++++++++++++++-----
2 files changed, 71 insertions(+), 24 deletions(-)
diff --git a/mdadm.8.in b/mdadm.8.in
index 8b21ffd4..0be02e4a 100644
--- a/mdadm.8.in
+++ b/mdadm.8.in
@@ -266,14 +266,11 @@ the exact meaning of this option in different contexts.
.TP
.BR \-c ", " \-\-config=
-Specify the config file or directory. Default is to use
-.B {CONFFILE}
-and
-.BR {CONFFILE}.d ,
-or if those are missing then
-.B {CONFFILE2}
-and
-.BR {CONFFILE2}.d .
+Specify the config file or directory. If not specified, default config file
+and default conf.d directory will be used. See
+.BR mdadm.conf (5)
+for more details.
+
If the config file given is
.B "partitions"
then nothing will be read, but
@@ -2013,11 +2010,9 @@ The config file is only used if explicitly named with
.B \-\-config
or requested with (a possibly implicit)
.BR \-\-scan .
-In the later case,
-.B {CONFFILE}
-or
-.B {CONFFILE2}
-is used.
+In the later case, default config file is used. See
+.BR mdadm.conf (5)
+for more details.
If
.B \-\-scan
@@ -3346,16 +3341,15 @@ on Monitor mode.
.SS {CONFFILE} (or {CONFFILE2})
-The config file lists which devices may be scanned to see if
-they contain MD super block, and gives identifying information
-(e.g. UUID) about known MD arrays. See
+Default config file. See
.BR mdadm.conf (5)
for more details.
.SS {CONFFILE}.d (or {CONFFILE2}.d)
-A directory containing configuration files which are read in lexical
-order.
+Default directory containing configuration files. See
+.BR mdadm.conf (5)
+for more details.
.SS {MAP_PATH}
When
diff --git a/mdadm.conf.5.in b/mdadm.conf.5.in
index 83edd008..dd331a6a 100644
--- a/mdadm.conf.5.in
+++ b/mdadm.conf.5.in
@@ -88,7 +88,8 @@ but only the major and minor device numbers. It scans
.I /dev
to find the name that matches the numbers.
-If no DEVICE line is present, then "DEVICE partitions containers" is assumed.
+If no DEVICE line is present in any config file,
+then "DEVICE partitions containers" is assumed.
For example:
.IP
@@ -272,6 +273,10 @@ catenated with spaces to form the address.
Note that this value cannot be set via the
.I mdadm
commandline. It is only settable via the config file.
+There should only be one
+.B MAILADDR
+line and it should have only one address. Any subsequent addresses
+are silently ignored.
.TP
.B PROGRAM
@@ -286,7 +291,8 @@ device.
There should only be one
.B program
-line and it should be give only one program.
+line and it should be given only one program. Any subsequent programs
+are silently ignored.
.TP
@@ -295,7 +301,14 @@ The
.B create
line gives default values to be used when creating arrays, new members
of arrays, and device entries for arrays.
-These include:
+
+There should only be one
+.B create
+line. Any subsequent lines will override the previous settings.
+
+Keywords used in the
+.I CREATE
+line and supported values are:
.RS 4
.TP
@@ -475,8 +488,8 @@ The known metadata types are
.B AUTO
should be given at most once. Subsequent lines are silently ignored.
-Thus an earlier config file in a config directory will over-ride
-the setting in a later config file.
+Thus a later config file in a config directory will not overwrite
+the setting in an earlier config file.
.TP
.B POLICY
@@ -594,6 +607,7 @@ The
line lists custom values of MD device's sysfs attributes which will be
stored in sysfs after the array is assembled. Multiple lines are allowed and each
line has to contain the uuid or the name of the device to which it relates.
+Lines are applied in reverse order.
.RS 4
.TP
.B uuid=
@@ -621,7 +635,46 @@ is running in
.B \-\-monitor
mode.
.B \-d/\-\-delay
-command line argument takes precedence over the config file
+command line argument takes precedence over the config file.
+
+If multiple
+.B MINITORDELAY
+lines are provided, only first non-zero value is considered.
+
+.SH FILES
+
+.SS {CONFFILE}
+
+The default config file location, used when
+.I mdadm
+is running without --config option.
+
+.SS {CONFFILE}.d
+
+The default directory with config files. Used when
+.I mdadm
+is running without --config option, after successful reading of the
+.B {CONFFILE}
+default config file. Files in that directory
+are read in lexical order.
+
+
+.SS {CONFFILE2}
+
+Alternative config file that is read, when
+.I mdadm
+is running without --config option and the
+.B {CONFFILE}
+default config file was not opened successfully.
+
+.SS {CONFFILE2}.d
+
+The alternative directory with config files. Used when
+.I mdadm
+is runninng without --config option, after reading the
+.B {CONFFILE2}
+alternative config file whether it was successful or not. Files in
+that directory are read in lexical order.
.SH EXAMPLE
DEVICE /dev/sd[bcdjkl]1
--
2.31.1

@ -0,0 +1,45 @@
From c33bbda5b0e127bb161fd4ad44bcfaa2a5daf153 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 18 Mar 2022 09:26:07 +0100
Subject: [PATCH 09/12] mdadm: Update config manual
Add missing HOMECLUSTER keyword description.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
mdadm.conf.5.in | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/mdadm.conf.5.in b/mdadm.conf.5.in
index dd331a6a..cd4e6a9d 100644
--- a/mdadm.conf.5.in
+++ b/mdadm.conf.5.in
@@ -439,6 +439,23 @@ from any possible local name. e.g.
.B /dev/md/1_1
or
.BR /dev/md/home_0 .
+
+.TP
+.B HOMECLUSTER
+The
+.B homcluster
+line gives a default value for the
+.B \-\-homecluster=
+option to mdadm. It specifies the cluster name for the md device.
+The md device can be assembled only on the cluster which matches
+the name specified. If
+.B homcluster
+is not provided, mdadm tries to detect the cluster name automatically.
+
+There should only be one
+.B homecluster
+line. Any subsequent lines will be silently ignored.
+
.TP
.B AUTO
A list of names of metadata format can be given, each preceded by a
--
2.31.1

@ -0,0 +1,153 @@
From 913f07d1db4a0078acc26d6ccabe1c315cf9273c Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Thu, 20 Jan 2022 13:18:32 +0100
Subject: [PATCH 10/12] Create, Build: use default_layout()
This code is duplicated for Build mode so make default_layout() extern
and use it. Simplify the function structure.
It introduced change for Build mode, now for raid0 RAID0_ORIG_LAYOUT
will be returned same as for Create.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Build.c | 23 +------------------
Create.c | 67 ++++++++++++++++++++++++++++++++++----------------------
mdadm.h | 1 +
3 files changed, 43 insertions(+), 48 deletions(-)
diff --git a/Build.c b/Build.c
index 962c2e37..8d6f6f58 100644
--- a/Build.c
+++ b/Build.c
@@ -71,28 +71,7 @@ int Build(char *mddev, struct mddev_dev *devlist,
}
if (s->layout == UnSet)
- switch(s->level) {
- default: /* no layout */
- s->layout = 0;
- break;
- case 10:
- s->layout = 0x102; /* near=2, far=1 */
- if (c->verbose > 0)
- pr_err("layout defaults to n1\n");
- break;
- case 5:
- case 6:
- s->layout = map_name(r5layout, "default");
- if (c->verbose > 0)
- pr_err("layout defaults to %s\n", map_num(r5layout, s->layout));
- break;
- case LEVEL_FAULTY:
- s->layout = map_name(faultylayout, "default");
-
- if (c->verbose > 0)
- pr_err("layout defaults to %s\n", map_num(faultylayout, s->layout));
- break;
- }
+ s->layout = default_layout(NULL, s->level, c->verbose);
/* We need to create the device. It can have no name. */
map_lock(&map);
diff --git a/Create.c b/Create.c
index 0ff1922d..9ea19de0 100644
--- a/Create.c
+++ b/Create.c
@@ -39,39 +39,54 @@ static int round_size_and_verify(unsigned long long *size, int chunk)
return 0;
}
-static int default_layout(struct supertype *st, int level, int verbose)
+/**
+ * default_layout() - Get default layout for level.
+ * @st: metadata requested, could be NULL.
+ * @level: raid level requested.
+ * @verbose: verbose level.
+ *
+ * Try to ask metadata handler first, otherwise use global defaults.
+ *
+ * Return: Layout or &UnSet, return value meaning depends of level used.
+ */
+int default_layout(struct supertype *st, int level, int verbose)
{
int layout = UnSet;
+ mapping_t *layout_map = NULL;
+ char *layout_name = NULL;
if (st && st->ss->default_geometry)
st->ss->default_geometry(st, &level, &layout, NULL);
- if (layout == UnSet)
- switch(level) {
- default: /* no layout */
- layout = 0;
- break;
- case 0:
- layout = RAID0_ORIG_LAYOUT;
- break;
- case 10:
- layout = 0x102; /* near=2, far=1 */
- if (verbose > 0)
- pr_err("layout defaults to n2\n");
- break;
- case 5:
- case 6:
- layout = map_name(r5layout, "default");
- if (verbose > 0)
- pr_err("layout defaults to %s\n", map_num(r5layout, layout));
- break;
- case LEVEL_FAULTY:
- layout = map_name(faultylayout, "default");
+ if (layout != UnSet)
+ return layout;
- if (verbose > 0)
- pr_err("layout defaults to %s\n", map_num(faultylayout, layout));
- break;
- }
+ switch (level) {
+ default: /* no layout */
+ layout = 0;
+ break;
+ case 0:
+ layout = RAID0_ORIG_LAYOUT;
+ break;
+ case 10:
+ layout = 0x102; /* near=2, far=1 */
+ layout_name = "n2";
+ break;
+ case 5:
+ case 6:
+ layout_map = r5layout;
+ break;
+ case LEVEL_FAULTY:
+ layout_map = faultylayout;
+ break;
+ }
+
+ if (layout_map) {
+ layout = map_name(layout_map, "default");
+ layout_name = map_num(layout_map, layout);
+ }
+ if (layout_name && verbose > 0)
+ pr_err("layout defaults to %s\n", layout_name);
return layout;
}
diff --git a/mdadm.h b/mdadm.h
index 26e7e5cd..cd72e711 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -1512,6 +1512,7 @@ extern int get_linux_version(void);
extern int mdadm_version(char *version);
extern unsigned long long parse_size(char *size);
extern int parse_uuid(char *str, int uuid[4]);
+int default_layout(struct supertype *st, int level, int verbose);
extern int is_near_layout_10(int layout);
extern int parse_layout_10(char *layout);
extern int parse_layout_faulty(char *layout);
--
2.31.1

@ -0,0 +1,382 @@
From 5f21d67472ad08c1e96b4385254adba79aa1c467 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Thu, 20 Jan 2022 13:18:33 +0100
Subject: [PATCH 11/12] mdadm: add map_num_s()
map_num() returns NULL if key is not defined. This patch adds
alternative, non NULL version for cases where NULL is not expected.
There are many printf() calls where map_num() is called on variable
without NULL verification. It works, even if NULL is passed because
gcc is able to ignore NULL argument quietly but the behavior is
undefined. For safety reasons such usages will use map_num_s() now.
It is a potential point of regression.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Assemble.c | 6 ++----
Create.c | 2 +-
Detail.c | 4 ++--
Grow.c | 16 ++++++++--------
Query.c | 4 ++--
maps.c | 24 ++++++++++++++++++++++++
mdadm.c | 20 ++++++++++----------
mdadm.h | 2 +-
super-ddf.c | 6 +++---
super-intel.c | 2 +-
super0.c | 2 +-
super1.c | 2 +-
sysfs.c | 9 +++++----
13 files changed, 61 insertions(+), 38 deletions(-)
diff --git a/Assemble.c b/Assemble.c
index 704b8293..9eac9ce0 100644
--- a/Assemble.c
+++ b/Assemble.c
@@ -63,7 +63,7 @@ static void set_array_assembly_status(struct context *c,
struct assembly_array_info *arr)
{
int raid_disks = arr->preexist_cnt + arr->new_cnt;
- char *status_msg = map_num(assemble_statuses, status);
+ char *status_msg = map_num_s(assemble_statuses, status);
if (c->export && result)
*result |= status;
@@ -77,9 +77,7 @@ static void set_array_assembly_status(struct context *c,
fprintf(stderr, " (%d new)", arr->new_cnt);
if (arr->exp_cnt)
fprintf(stderr, " ( + %d for expansion)", arr->exp_cnt);
- if (status_msg)
- fprintf(stderr, " %s", status_msg);
- fprintf(stderr, ".\n");
+ fprintf(stderr, " %s.\n", status_msg);
}
static int name_matches(char *found, char *required, char *homehost, int require_homehost)
diff --git a/Create.c b/Create.c
index 9ea19de0..c84c1ac8 100644
--- a/Create.c
+++ b/Create.c
@@ -83,7 +83,7 @@ int default_layout(struct supertype *st, int level, int verbose)
if (layout_map) {
layout = map_name(layout_map, "default");
- layout_name = map_num(layout_map, layout);
+ layout_name = map_num_s(layout_map, layout);
}
if (layout_name && verbose > 0)
pr_err("layout defaults to %s\n", layout_name);
diff --git a/Detail.c b/Detail.c
index 95d4cc70..ce7a8445 100644
--- a/Detail.c
+++ b/Detail.c
@@ -495,8 +495,8 @@ int Detail(char *dev, struct context *c)
if (array.state & (1 << MD_SB_CLEAN)) {
if ((array.level == 0) ||
(array.level == LEVEL_LINEAR))
- arrayst = map_num(sysfs_array_states,
- sra->array_state);
+ arrayst = map_num_s(sysfs_array_states,
+ sra->array_state);
else
arrayst = "clean";
} else {
diff --git a/Grow.c b/Grow.c
index 18c5719b..8a242b0f 100644
--- a/Grow.c
+++ b/Grow.c
@@ -547,7 +547,7 @@ int Grow_consistency_policy(char *devname, int fd, struct context *c, struct sha
if (s->consistency_policy != CONSISTENCY_POLICY_RESYNC &&
s->consistency_policy != CONSISTENCY_POLICY_PPL) {
pr_err("Operation not supported for consistency policy %s\n",
- map_num(consistency_policies, s->consistency_policy));
+ map_num_s(consistency_policies, s->consistency_policy));
return 1;
}
@@ -578,14 +578,14 @@ int Grow_consistency_policy(char *devname, int fd, struct context *c, struct sha
if (sra->consistency_policy == (unsigned)s->consistency_policy) {
pr_err("Consistency policy is already %s\n",
- map_num(consistency_policies, s->consistency_policy));
+ map_num_s(consistency_policies, s->consistency_policy));
ret = 1;
goto free_info;
} else if (sra->consistency_policy != CONSISTENCY_POLICY_RESYNC &&
sra->consistency_policy != CONSISTENCY_POLICY_PPL) {
pr_err("Current consistency policy is %s, cannot change to %s\n",
- map_num(consistency_policies, sra->consistency_policy),
- map_num(consistency_policies, s->consistency_policy));
+ map_num_s(consistency_policies, sra->consistency_policy),
+ map_num_s(consistency_policies, s->consistency_policy));
ret = 1;
goto free_info;
}
@@ -704,8 +704,8 @@ int Grow_consistency_policy(char *devname, int fd, struct context *c, struct sha
}
ret = sysfs_set_str(sra, NULL, "consistency_policy",
- map_num(consistency_policies,
- s->consistency_policy));
+ map_num_s(consistency_policies,
+ s->consistency_policy));
if (ret)
pr_err("Failed to change array consistency policy\n");
@@ -2241,7 +2241,7 @@ size_change_error:
info.new_layout = UnSet;
if (info.array.level == 6 && info.new_level == UnSet) {
char l[40], *h;
- strcpy(l, map_num(r6layout, info.array.layout));
+ strcpy(l, map_num_s(r6layout, info.array.layout));
h = strrchr(l, '-');
if (h && strcmp(h, "-6") == 0) {
*h = 0;
@@ -2266,7 +2266,7 @@ size_change_error:
info.new_layout = info.array.layout;
else if (info.array.level == 5 && info.new_level == 6) {
char l[40];
- strcpy(l, map_num(r5layout, info.array.layout));
+ strcpy(l, map_num_s(r5layout, info.array.layout));
strcat(l, "-6");
info.new_layout = map_name(r6layout, l);
} else {
diff --git a/Query.c b/Query.c
index 23fbf8aa..adcd231e 100644
--- a/Query.c
+++ b/Query.c
@@ -93,7 +93,7 @@ int Query(char *dev)
else {
printf("%s: %s %s %d devices, %d spare%s. Use mdadm --detail for more detail.\n",
dev, human_size_brief(larray_size,IEC),
- map_num(pers, level), raid_disks,
+ map_num_s(pers, level), raid_disks,
spare_disks, spare_disks == 1 ? "" : "s");
}
st = guess_super(fd);
@@ -131,7 +131,7 @@ int Query(char *dev)
dev,
info.disk.number, info.array.raid_disks,
activity,
- map_num(pers, info.array.level),
+ map_num_s(pers, info.array.level),
mddev);
if (st->ss == &super0)
put_md_name(mddev);
diff --git a/maps.c b/maps.c
index a4fd2797..20fcf719 100644
--- a/maps.c
+++ b/maps.c
@@ -166,6 +166,30 @@ mapping_t sysfs_array_states[] = {
{ NULL, ARRAY_UNKNOWN_STATE }
};
+/**
+ * map_num_s() - Safer alternative of map_num() function.
+ * @map: map to search.
+ * @num: key to match.
+ *
+ * Shall be used only if key existence is quaranted.
+ *
+ * Return: Pointer to name of the element.
+ */
+char *map_num_s(mapping_t *map, int num)
+{
+ char *ret = map_num(map, num);
+
+ assert(ret);
+ return ret;
+}
+
+/**
+ * map_num() - get element name by key.
+ * @map: map to search.
+ * @num: key to match.
+ *
+ * Return: Pointer to name of the element or NULL.
+ */
char *map_num(mapping_t *map, int num)
{
while (map->name) {
diff --git a/mdadm.c b/mdadm.c
index 26299b2e..be40686c 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -280,8 +280,8 @@ int main(int argc, char *argv[])
else
fprintf(stderr, "-%c", opt);
fprintf(stderr, " would set mdadm mode to \"%s\", but it is already set to \"%s\".\n",
- map_num(modes, newmode),
- map_num(modes, mode));
+ map_num_s(modes, newmode),
+ map_num_s(modes, mode));
exit(2);
} else if (!mode && newmode) {
mode = newmode;
@@ -544,7 +544,7 @@ int main(int argc, char *argv[])
switch(s.level) {
default:
pr_err("layout not meaningful for %s arrays.\n",
- map_num(pers, s.level));
+ map_num_s(pers, s.level));
exit(2);
case UnSet:
pr_err("raid level must be given before layout.\n");
@@ -1248,10 +1248,10 @@ int main(int argc, char *argv[])
if (option_index > 0)
pr_err(":option --%s not valid in %s mode\n",
long_options[option_index].name,
- map_num(modes, mode));
+ map_num_s(modes, mode));
else
pr_err("option -%c not valid in %s mode\n",
- opt, map_num(modes, mode));
+ opt, map_num_s(modes, mode));
exit(2);
}
@@ -1276,7 +1276,7 @@ int main(int argc, char *argv[])
if (s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN &&
s.consistency_policy != CONSISTENCY_POLICY_JOURNAL) {
pr_err("--write-journal is not supported with consistency policy: %s\n",
- map_num(consistency_policies, s.consistency_policy));
+ map_num_s(consistency_policies, s.consistency_policy));
exit(2);
}
}
@@ -1285,12 +1285,12 @@ int main(int argc, char *argv[])
s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN) {
if (s.level <= 0) {
pr_err("--consistency-policy not meaningful with level %s.\n",
- map_num(pers, s.level));
+ map_num_s(pers, s.level));
exit(2);
} else if (s.consistency_policy == CONSISTENCY_POLICY_JOURNAL &&
!s.journaldisks) {
pr_err("--write-journal is required for consistency policy: %s\n",
- map_num(consistency_policies, s.consistency_policy));
+ map_num_s(consistency_policies, s.consistency_policy));
exit(2);
} else if (s.consistency_policy == CONSISTENCY_POLICY_PPL &&
s.level != 5) {
@@ -1300,14 +1300,14 @@ int main(int argc, char *argv[])
(!s.bitmap_file ||
strcmp(s.bitmap_file, "none") == 0)) {
pr_err("--bitmap is required for consistency policy: %s\n",
- map_num(consistency_policies, s.consistency_policy));
+ map_num_s(consistency_policies, s.consistency_policy));
exit(2);
} else if (s.bitmap_file &&
strcmp(s.bitmap_file, "none") != 0 &&
s.consistency_policy != CONSISTENCY_POLICY_BITMAP &&
s.consistency_policy != CONSISTENCY_POLICY_JOURNAL) {
pr_err("--bitmap is not compatible with consistency policy: %s\n",
- map_num(consistency_policies, s.consistency_policy));
+ map_num_s(consistency_policies, s.consistency_policy));
exit(2);
}
}
diff --git a/mdadm.h b/mdadm.h
index cd72e711..09915a00 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -770,7 +770,7 @@ extern int restore_stripes(int *dest, unsigned long long *offsets,
#endif
#define SYSLOG_FACILITY LOG_DAEMON
-
+extern char *map_num_s(mapping_t *map, int num);
extern char *map_num(mapping_t *map, int num);
extern int map_name(mapping_t *map, char *name);
extern mapping_t r0layout[], r5layout[], r6layout[],
diff --git a/super-ddf.c b/super-ddf.c
index 3f304cdc..8cda23a7 100644
--- a/super-ddf.c
+++ b/super-ddf.c
@@ -1477,13 +1477,13 @@ static void examine_vds(struct ddf_super *sb)
printf("\n");
printf(" unit[%d] : %d\n", i, be16_to_cpu(ve->unit));
printf(" state[%d] : %s, %s%s\n", i,
- map_num(ddf_state, ve->state & 7),
+ map_num_s(ddf_state, ve->state & 7),
(ve->state & DDF_state_morphing) ? "Morphing, ": "",
(ve->state & DDF_state_inconsistent)? "Not Consistent" : "Consistent");
printf(" init state[%d] : %s\n", i,
- map_num(ddf_init_state, ve->init_state&DDF_initstate_mask));
+ map_num_s(ddf_init_state, ve->init_state & DDF_initstate_mask));
printf(" access[%d] : %s\n", i,
- map_num(ddf_access, (ve->init_state & DDF_access_mask) >> 6));
+ map_num_s(ddf_access, (ve->init_state & DDF_access_mask) >> 6));
printf(" Name[%d] : %.16s\n", i, ve->name);
examine_vd(i, sb, ve->guid);
}
diff --git a/super-intel.c b/super-intel.c
index 6ff336ee..ba3bd41f 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -5625,7 +5625,7 @@ static int init_super_imsm_volume(struct supertype *st, mdu_array_info_t *info,
free(dev);
free(dv);
pr_err("imsm does not support consistency policy %s\n",
- map_num(consistency_policies, s->consistency_policy));
+ map_num_s(consistency_policies, s->consistency_policy));
return 0;
}
diff --git a/super0.c b/super0.c
index b79b97a9..61c9ec1d 100644
--- a/super0.c
+++ b/super0.c
@@ -288,7 +288,7 @@ static void export_examine_super0(struct supertype *st)
{
mdp_super_t *sb = st->sb;
- printf("MD_LEVEL=%s\n", map_num(pers, sb->level));
+ printf("MD_LEVEL=%s\n", map_num_s(pers, sb->level));
printf("MD_DEVICES=%d\n", sb->raid_disks);
if (sb->minor_version >= 90)
printf("MD_UUID=%08x:%08x:%08x:%08x\n",
diff --git a/super1.c b/super1.c
index a12a5bc8..e3e2f954 100644
--- a/super1.c
+++ b/super1.c
@@ -671,7 +671,7 @@ static void export_examine_super1(struct supertype *st)
int len = 32;
int layout;
- printf("MD_LEVEL=%s\n", map_num(pers, __le32_to_cpu(sb->level)));
+ printf("MD_LEVEL=%s\n", map_num_s(pers, __le32_to_cpu(sb->level)));
printf("MD_DEVICES=%d\n", __le32_to_cpu(sb->raid_disks));
for (i = 0; i < 32; i++)
if (sb->set_name[i] == '\n' || sb->set_name[i] == '\0') {
diff --git a/sysfs.c b/sysfs.c
index 2995713d..0d98a65f 100644
--- a/sysfs.c
+++ b/sysfs.c
@@ -689,7 +689,7 @@ int sysfs_set_array(struct mdinfo *info, int vers)
if (info->array.level < 0)
return 0; /* FIXME */
rv |= sysfs_set_str(info, NULL, "level",
- map_num(pers, info->array.level));
+ map_num_s(pers, info->array.level));
if (info->reshape_active && info->delta_disks != UnSet)
raid_disks -= info->delta_disks;
rv |= sysfs_set_num(info, NULL, "raid_disks", raid_disks);
@@ -724,9 +724,10 @@ int sysfs_set_array(struct mdinfo *info, int vers)
}
if (info->consistency_policy == CONSISTENCY_POLICY_PPL) {
- if (sysfs_set_str(info, NULL, "consistency_policy",
- map_num(consistency_policies,
- info->consistency_policy))) {
+ char *policy = map_num_s(consistency_policies,
+ info->consistency_policy);
+
+ if (sysfs_set_str(info, NULL, "consistency_policy", policy)) {
pr_err("This kernel does not support PPL. Falling back to consistency-policy=resync.\n");
info->consistency_policy = CONSISTENCY_POLICY_RESYNC;
}
--
2.31.1

@ -0,0 +1,122 @@
From 1066ab83dbe9a4cc20f7db44a40aa2cbb9d5eed6 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 13 May 2022 09:19:42 +0200
Subject: [PATCH 13/52] mdmon: Stop parsing duplicate options
Introduce new function is_duplicate_opt() to check if given option
was already used and prevent setting it again along with an error
message.
Move parsing above in_initrd() check to be able to detect --offroot
option duplicates.
Now help option is executed after parsing to prevent executing commands
like: 'mdmon --help --ndlksnlksajndfjksndafasj'.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
mdmon.c | 44 +++++++++++++++++++++++++++++++++++---------
1 file changed, 35 insertions(+), 9 deletions(-)
diff --git a/mdmon.c b/mdmon.c
index 5570574b..c057da63 100644
--- a/mdmon.c
+++ b/mdmon.c
@@ -288,6 +288,15 @@ void usage(void)
exit(2);
}
+static bool is_duplicate_opt(const int opt, const int set_val, const char *long_name)
+{
+ if (opt == set_val) {
+ pr_err("--%s option duplicated!\n", long_name);
+ return true;
+ }
+ return false;
+}
+
static int mdmon(char *devnm, int must_fork, int takeover);
int main(int argc, char *argv[])
@@ -299,6 +308,7 @@ int main(int argc, char *argv[])
int all = 0;
int takeover = 0;
int dofork = 1;
+ bool help = false;
static struct option options[] = {
{"all", 0, NULL, 'a'},
{"takeover", 0, NULL, 't'},
@@ -308,37 +318,50 @@ int main(int argc, char *argv[])
{NULL, 0, NULL, 0}
};
- if (in_initrd()) {
- /*
- * set first char of argv[0] to @. This is used by
- * systemd to signal that the task was launched from
- * initrd/initramfs and should be preserved during shutdown
- */
- argv[0][0] = '@';
- }
-
while ((opt = getopt_long(argc, argv, "thaF", options, NULL)) != -1) {
switch (opt) {
case 'a':
+ if (is_duplicate_opt(all, 1, "all"))
+ exit(1);
container_name = argv[optind-1];
all = 1;
break;
case 't':
+ if (is_duplicate_opt(takeover, 1, "takeover"))
+ exit(1);
takeover = 1;
break;
case 'F':
+ if (is_duplicate_opt(dofork, 0, "foreground"))
+ exit(1);
dofork = 0;
break;
case OffRootOpt:
+ if (is_duplicate_opt(argv[0][0], '@', "offroot"))
+ exit(1);
argv[0][0] = '@';
break;
case 'h':
+ if (is_duplicate_opt(help, true, "help"))
+ exit(1);
+ help = true;
+ break;
default:
usage();
break;
}
}
+
+ if (in_initrd()) {
+ /*
+ * set first char of argv[0] to @. This is used by
+ * systemd to signal that the task was launched from
+ * initrd/initramfs and should be preserved during shutdown
+ */
+ argv[0][0] = '@';
+ }
+
if (all == 0 && container_name == NULL) {
if (argv[optind])
container_name = argv[optind];
@@ -353,6 +376,9 @@ int main(int argc, char *argv[])
if (strcmp(container_name, "/proc/mdstat") == 0)
all = 1;
+ if (help)
+ usage();
+
if (all) {
struct mdstat_ent *mdstat, *e;
int container_len = strlen(container_name);
--
2.31.1

@ -0,0 +1,41 @@
From 20e114e334ed6ed3280c37a9a08fb95578393d1a Mon Sep 17 00:00:00 2001
From: Mateusz Kusiak <mateusz.kusiak@intel.com>
Date: Thu, 19 May 2022 09:16:08 +0200
Subject: [PATCH 14/52] Grow: block -n on external volumes.
Performing --raid-devices on external metadata volume should be blocked
as it causes unwanted behaviour.
Eg. Performing
mdadm -G /dev/md/volume -l10 -n4
on r0_d2 inside 4 disk container, returns
mdadm: Need 2 spares to avoid degraded array, only have 0.
Signed-off-by: Mateusz Kusiak <mateusz.kusiak@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Grow.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/Grow.c b/Grow.c
index 8a242b0f..f6efbc48 100644
--- a/Grow.c
+++ b/Grow.c
@@ -1892,6 +1892,14 @@ int Grow_reshape(char *devname, int fd,
if (retval) {
pr_err("Cannot read superblock for %s\n", devname);
+ close(cfd);
+ free(subarray);
+ return 1;
+ }
+
+ if (s->raiddisks && subarray) {
+ pr_err("--raid-devices operation can be performed on a container only\n");
+ close(cfd);
free(subarray);
return 1;
}
--
2.31.1

@ -0,0 +1,90 @@
From de064c93e3819d72720e4fba6575265ba10e1553 Mon Sep 17 00:00:00 2001
From: Mateusz Grzonka <mateusz.grzonka@intel.com>
Date: Mon, 13 Jun 2022 12:11:25 +0200
Subject: [PATCH 15/52] Incremental: Fix possible memory and resource leaks
map allocated through map_by_uuid() is not freed if mdfd is invalid.
In addition mdfd is not closed, and mdinfo list is not freed too.
Signed-off-by: Mateusz Grzonka <mateusz.grzonka@intel.com>
Change-Id: I25e726f0e2502cf7e8ce80c2bd7944b3b1e2b9dc
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Incremental.c | 32 +++++++++++++++++++++++---------
1 file changed, 23 insertions(+), 9 deletions(-)
diff --git a/Incremental.c b/Incremental.c
index a57fc323..4d0cd9d6 100644
--- a/Incremental.c
+++ b/Incremental.c
@@ -1499,7 +1499,7 @@ static int Incremental_container(struct supertype *st, char *devname,
return 0;
}
for (ra = list ; ra ; ra = ra->next) {
- int mdfd;
+ int mdfd = -1;
char chosen_name[1024];
struct map_ent *mp;
struct mddev_ident *match = NULL;
@@ -1514,6 +1514,12 @@ static int Incremental_container(struct supertype *st, char *devname,
if (mp) {
mdfd = open_dev(mp->devnm);
+ if (!is_fd_valid(mdfd)) {
+ pr_err("failed to open %s: %s.\n",
+ mp->devnm, strerror(errno));
+ rv = 2;
+ goto release;
+ }
if (mp->path)
strcpy(chosen_name, mp->path);
else
@@ -1573,21 +1579,25 @@ static int Incremental_container(struct supertype *st, char *devname,
c->autof,
trustworthy,
chosen_name, 0);
+
+ if (!is_fd_valid(mdfd)) {
+ pr_err("create_mddev failed with chosen name %s: %s.\n",
+ chosen_name, strerror(errno));
+ rv = 2;
+ goto release;
+ }
}
- if (only && (!mp || strcmp(mp->devnm, only) != 0))
- continue;
- if (mdfd < 0) {
- pr_err("failed to open %s: %s.\n",
- chosen_name, strerror(errno));
- return 2;
+ if (only && (!mp || strcmp(mp->devnm, only) != 0)) {
+ close_fd(&mdfd);
+ continue;
}
assemble_container_content(st, mdfd, ra, c,
chosen_name, &result);
map_free(map);
map = NULL;
- close(mdfd);
+ close_fd(&mdfd);
}
if (c->export && result) {
char sep = '=';
@@ -1610,7 +1620,11 @@ static int Incremental_container(struct supertype *st, char *devname,
}
printf("\n");
}
- return 0;
+
+release:
+ map_free(map);
+ sysfs_free(list);
+ return rv;
}
static void run_udisks(char *arg1, char *arg2)
--
2.31.1

@ -0,0 +1,98 @@
From e702f392959d1c2ad2089e595b52235ed97b4e18 Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Mon, 6 Jun 2022 12:32:12 +0200
Subject: [PATCH 16/52] Mdmonitor: Fix segfault
Mdadm with "--monitor" parameter requires md device
as an argument to be monitored. If given argument is
not a md device, error shall be returned. Previously
it was not checked and invalid argument caused
segmentation fault. This commit adds checking
that devices passed to mdmonitor are md devices.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Monitor.c | 10 +++++++++-
mdadm.h | 1 +
mdopen.c | 17 +++++++++++++++++
3 files changed, 27 insertions(+), 1 deletion(-)
diff --git a/Monitor.c b/Monitor.c
index c0ab5412..4e5802b5 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -182,6 +182,7 @@ int Monitor(struct mddev_dev *devlist,
continue;
if (strcasecmp(mdlist->devname, "<ignore>") == 0)
continue;
+
st = xcalloc(1, sizeof *st);
if (mdlist->devname[0] == '/')
st->devname = xstrdup(mdlist->devname);
@@ -190,6 +191,8 @@ int Monitor(struct mddev_dev *devlist,
strcpy(strcpy(st->devname, "/dev/md/"),
mdlist->devname);
}
+ if (!is_mddev(mdlist->devname))
+ return 1;
st->next = statelist;
st->devnm[0] = 0;
st->percent = RESYNC_UNKNOWN;
@@ -203,7 +206,12 @@ int Monitor(struct mddev_dev *devlist,
struct mddev_dev *dv;
for (dv = devlist; dv; dv = dv->next) {
- struct state *st = xcalloc(1, sizeof *st);
+ struct state *st;
+
+ if (!is_mddev(dv->devname))
+ return 1;
+
+ st = xcalloc(1, sizeof *st);
mdlist = conf_get_ident(dv->devname);
st->devname = xstrdup(dv->devname);
st->next = statelist;
diff --git a/mdadm.h b/mdadm.h
index 09915a00..d53df169 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -1636,6 +1636,7 @@ extern int create_mddev(char *dev, char *name, int autof, int trustworthy,
#define FOREIGN 2
#define METADATA 3
extern int open_mddev(char *dev, int report_errors);
+extern int is_mddev(char *dev);
extern int open_container(int fd);
extern int metadata_container_matches(char *metadata, char *devnm);
extern int metadata_subdev_matches(char *metadata, char *devnm);
diff --git a/mdopen.c b/mdopen.c
index 245be537..d18c9319 100644
--- a/mdopen.c
+++ b/mdopen.c
@@ -475,6 +475,23 @@ int open_mddev(char *dev, int report_errors)
return mdfd;
}
+/**
+ * is_mddev() - check that file name passed is an md device.
+ * @dev: file name that has to be checked.
+ * Return: 1 if file passed is an md device, 0 if not.
+ */
+int is_mddev(char *dev)
+{
+ int fd = open_mddev(dev, 1);
+
+ if (fd >= 0) {
+ close(fd);
+ return 1;
+ }
+
+ return 0;
+}
+
char *find_free_devnm(int use_partitions)
{
static char devnm[32];
--
2.31.1

@ -0,0 +1,61 @@
From f5ff2988761625b43eb15555993f2797af29f166 Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Mon, 6 Jun 2022 12:32:13 +0200
Subject: [PATCH 17/52] Mdmonitor: Improve logging method
Change logging, and as a result, mdmonitor in verbose
mode will report its configuration.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Signed-off-by: Oleksandr Shchirskyi <oleksandr.shchirskyi@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Monitor.c | 25 ++++++++++++++-----------
1 file changed, 14 insertions(+), 11 deletions(-)
diff --git a/Monitor.c b/Monitor.c
index 4e5802b5..6ca1ebe5 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -136,24 +136,27 @@ int Monitor(struct mddev_dev *devlist,
struct mddev_ident *mdlist;
int delay_for_event = c->delay;
- if (!mailaddr) {
+ if (!mailaddr)
mailaddr = conf_get_mailaddr();
- if (mailaddr && ! c->scan)
- pr_err("Monitor using email address \"%s\" from config file\n",
- mailaddr);
- }
- mailfrom = conf_get_mailfrom();
- if (!alert_cmd) {
+ if (!alert_cmd)
alert_cmd = conf_get_program();
- if (alert_cmd && !c->scan)
- pr_err("Monitor using program \"%s\" from config file\n",
- alert_cmd);
- }
+
+ mailfrom = conf_get_mailfrom();
+
if (c->scan && !mailaddr && !alert_cmd && !dosyslog) {
pr_err("No mail address or alert command - not monitoring.\n");
return 1;
}
+
+ if (c->verbose) {
+ pr_err("Monitor is started with delay %ds\n", c->delay);
+ if (mailaddr)
+ pr_err("Monitor using email address %s\n", mailaddr);
+ if (alert_cmd)
+ pr_err("Monitor using program %s\n", alert_cmd);
+ }
+
info.alert_cmd = alert_cmd;
info.mailaddr = mailaddr;
info.mailfrom = mailfrom;
--
2.31.1

@ -0,0 +1,73 @@
From 626bc45396c4959f2c4685c2faa7c4f553f4efdf Mon Sep 17 00:00:00 2001
From: Mateusz Grzonka <mateusz.grzonka@intel.com>
Date: Mon, 13 Jun 2022 11:59:34 +0200
Subject: [PATCH 18/52] Fix possible NULL ptr dereferences and memory leaks
In Assemble there was a NULL check for sra variable,
which effectively didn't stop the execution in every case.
That might have resulted in a NULL pointer dereference.
Also in super-ddf, mu variable was set to NULL for some condition,
and then immidiately dereferenced.
Additionally some memory wasn't freed as well.
Signed-off-by: Mateusz Grzonka <mateusz.grzonka@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Assemble.c | 7 ++++++-
super-ddf.c | 9 +++++++--
2 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/Assemble.c b/Assemble.c
index 9eac9ce0..4b213560 100644
--- a/Assemble.c
+++ b/Assemble.c
@@ -1982,7 +1982,12 @@ int assemble_container_content(struct supertype *st, int mdfd,
}
sra = sysfs_read(mdfd, NULL, GET_VERSION|GET_DEVS);
- if (sra == NULL || strcmp(sra->text_version, content->text_version) != 0) {
+ if (sra == NULL) {
+ pr_err("Failed to read sysfs parameters\n");
+ return 1;
+ }
+
+ if (strcmp(sra->text_version, content->text_version) != 0) {
if (content->array.major_version == -1 &&
content->array.minor_version == -2 &&
c->readonly &&
diff --git a/super-ddf.c b/super-ddf.c
index 8cda23a7..abbc8b09 100644
--- a/super-ddf.c
+++ b/super-ddf.c
@@ -5125,13 +5125,16 @@ static struct mdinfo *ddf_activate_spare(struct active_array *a,
*/
vc = find_vdcr(ddf, a->info.container_member, rv->disk.raid_disk,
&n_bvd, &vcl);
- if (vc == NULL)
+ if (vc == NULL) {
+ free(rv);
return NULL;
+ }
mu = xmalloc(sizeof(*mu));
if (posix_memalign(&mu->space, 512, sizeof(struct vcl)) != 0) {
free(mu);
- mu = NULL;
+ free(rv);
+ return NULL;
}
mu->len = ddf->conf_rec_len * 512 * vcl->conf.sec_elmnt_count;
@@ -5161,6 +5164,8 @@ static struct mdinfo *ddf_activate_spare(struct active_array *a,
pr_err("BUG: can't find disk %d (%d/%d)\n",
di->disk.raid_disk,
di->disk.major, di->disk.minor);
+ free(mu);
+ free(rv);
return NULL;
}
vc->phys_refnum[i_prim] = ddf->phys->entries[dl->pdnum].refnum;
--
2.31.1

@ -0,0 +1,301 @@
From 756a15f32338fdf0c562678694bc8991ad6afb90 Mon Sep 17 00:00:00 2001
From: Mateusz Grzonka <mateusz.grzonka@intel.com>
Date: Mon, 13 Jun 2022 12:00:09 +0200
Subject: [PATCH 19/52] imsm: Remove possibility for get_imsm_dev to return
NULL
Returning NULL from get_imsm_dev or __get_imsm_dev will cause segfault.
Guarantee that it never happens.
Signed-off-by: Mateusz Grzonka <mateusz.grzonka@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 153 +++++++++++++++++++++++++-------------------------
1 file changed, 78 insertions(+), 75 deletions(-)
diff --git a/super-intel.c b/super-intel.c
index ba3bd41f..3788feb9 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -851,6 +851,21 @@ static struct disk_info *get_disk_info(struct imsm_update_create_array *update)
return inf;
}
+/**
+ * __get_imsm_dev() - Get device with index from imsm_super.
+ * @mpb: &imsm_super pointer, not NULL.
+ * @index: Device index.
+ *
+ * Function works as non-NULL, aborting in such a case,
+ * when NULL would be returned.
+ *
+ * Device index should be in range 0 up to num_raid_devs.
+ * Function assumes the index was already verified.
+ * Index must be valid, otherwise abort() is called.
+ *
+ * Return: Pointer to corresponding imsm_dev.
+ *
+ */
static struct imsm_dev *__get_imsm_dev(struct imsm_super *mpb, __u8 index)
{
int offset;
@@ -858,30 +873,47 @@ static struct imsm_dev *__get_imsm_dev(struct imsm_super *mpb, __u8 index)
void *_mpb = mpb;
if (index >= mpb->num_raid_devs)
- return NULL;
+ goto error;
/* devices start after all disks */
offset = ((void *) &mpb->disk[mpb->num_disks]) - _mpb;
- for (i = 0; i <= index; i++)
+ for (i = 0; i <= index; i++, offset += sizeof_imsm_dev(_mpb + offset, 0))
if (i == index)
return _mpb + offset;
- else
- offset += sizeof_imsm_dev(_mpb + offset, 0);
-
- return NULL;
+error:
+ pr_err("cannot find imsm_dev with index %u in imsm_super\n", index);
+ abort();
}
+/**
+ * get_imsm_dev() - Get device with index from intel_super.
+ * @super: &intel_super pointer, not NULL.
+ * @index: Device index.
+ *
+ * Function works as non-NULL, aborting in such a case,
+ * when NULL would be returned.
+ *
+ * Device index should be in range 0 up to num_raid_devs.
+ * Function assumes the index was already verified.
+ * Index must be valid, otherwise abort() is called.
+ *
+ * Return: Pointer to corresponding imsm_dev.
+ *
+ */
static struct imsm_dev *get_imsm_dev(struct intel_super *super, __u8 index)
{
struct intel_dev *dv;
if (index >= super->anchor->num_raid_devs)
- return NULL;
+ goto error;
+
for (dv = super->devlist; dv; dv = dv->next)
if (dv->index == index)
return dv->dev;
- return NULL;
+error:
+ pr_err("cannot find imsm_dev with index %u in intel_super\n", index);
+ abort();
}
static inline unsigned long long __le48_to_cpu(const struct bbm_log_block_addr
@@ -4364,8 +4396,7 @@ int check_mpb_migr_compatibility(struct intel_super *super)
for (i = 0; i < super->anchor->num_raid_devs; i++) {
struct imsm_dev *dev_iter = __get_imsm_dev(super->anchor, i);
- if (dev_iter &&
- dev_iter->vol.migr_state == 1 &&
+ if (dev_iter->vol.migr_state == 1 &&
dev_iter->vol.migr_type == MIGR_GEN_MIGR) {
/* This device is migrating */
map0 = get_imsm_map(dev_iter, MAP_0);
@@ -4514,8 +4545,6 @@ static void clear_hi(struct intel_super *super)
}
for (i = 0; i < mpb->num_raid_devs; ++i) {
struct imsm_dev *dev = get_imsm_dev(super, i);
- if (!dev)
- return;
for (n = 0; n < 2; ++n) {
struct imsm_map *map = get_imsm_map(dev, n);
if (!map)
@@ -5836,7 +5865,7 @@ static int add_to_super_imsm_volume(struct supertype *st, mdu_disk_info_t *dk,
struct imsm_dev *_dev = __get_imsm_dev(mpb, 0);
_disk = __get_imsm_disk(mpb, dl->index);
- if (!_dev || !_disk) {
+ if (!_disk) {
pr_err("BUG mpb setup error\n");
return 1;
}
@@ -6171,10 +6200,10 @@ static int write_super_imsm(struct supertype *st, int doclose)
for (i = 0; i < mpb->num_raid_devs; i++) {
struct imsm_dev *dev = __get_imsm_dev(mpb, i);
struct imsm_dev *dev2 = get_imsm_dev(super, i);
- if (dev && dev2) {
- imsm_copy_dev(dev, dev2);
- mpb_size += sizeof_imsm_dev(dev, 0);
- }
+
+ imsm_copy_dev(dev, dev2);
+ mpb_size += sizeof_imsm_dev(dev, 0);
+
if (is_gen_migration(dev2))
clear_migration_record = 0;
}
@@ -9033,29 +9062,26 @@ static int imsm_rebuild_allowed(struct supertype *cont, int dev_idx, int failed)
__u8 state;
dev2 = get_imsm_dev(cont->sb, dev_idx);
- if (dev2) {
- state = imsm_check_degraded(cont->sb, dev2, failed, MAP_0);
- if (state == IMSM_T_STATE_FAILED) {
- map = get_imsm_map(dev2, MAP_0);
- if (!map)
- return 1;
- for (slot = 0; slot < map->num_members; slot++) {
- /*
- * Check if failed disks are deleted from intel
- * disk list or are marked to be deleted
- */
- idx = get_imsm_disk_idx(dev2, slot, MAP_X);
- idisk = get_imsm_dl_disk(cont->sb, idx);
- /*
- * Do not rebuild the array if failed disks
- * from failed sub-array are not removed from
- * container.
- */
- if (idisk &&
- is_failed(&idisk->disk) &&
- (idisk->action != DISK_REMOVE))
- return 0;
- }
+
+ state = imsm_check_degraded(cont->sb, dev2, failed, MAP_0);
+ if (state == IMSM_T_STATE_FAILED) {
+ map = get_imsm_map(dev2, MAP_0);
+ for (slot = 0; slot < map->num_members; slot++) {
+ /*
+ * Check if failed disks are deleted from intel
+ * disk list or are marked to be deleted
+ */
+ idx = get_imsm_disk_idx(dev2, slot, MAP_X);
+ idisk = get_imsm_dl_disk(cont->sb, idx);
+ /*
+ * Do not rebuild the array if failed disks
+ * from failed sub-array are not removed from
+ * container.
+ */
+ if (idisk &&
+ is_failed(&idisk->disk) &&
+ (idisk->action != DISK_REMOVE))
+ return 0;
}
}
return 1;
@@ -10089,7 +10115,6 @@ static void imsm_process_update(struct supertype *st,
int victim = u->dev_idx;
struct active_array *a;
struct intel_dev **dp;
- struct imsm_dev *dev;
/* sanity check that we are not affecting the uuid of
* active arrays, or deleting an active array
@@ -10105,8 +10130,7 @@ static void imsm_process_update(struct supertype *st,
* is active in the container, so checking
* mpb->num_raid_devs is just extra paranoia
*/
- dev = get_imsm_dev(super, victim);
- if (a || !dev || mpb->num_raid_devs == 1) {
+ if (a || mpb->num_raid_devs == 1 || victim >= super->anchor->num_raid_devs) {
dprintf("failed to delete subarray-%d\n", victim);
break;
}
@@ -10140,7 +10164,7 @@ static void imsm_process_update(struct supertype *st,
if (a->info.container_member == target)
break;
dev = get_imsm_dev(super, u->dev_idx);
- if (a || !dev || !check_name(super, name, 1)) {
+ if (a || !check_name(super, name, 1)) {
dprintf("failed to rename subarray-%d\n", target);
break;
}
@@ -10169,10 +10193,6 @@ static void imsm_process_update(struct supertype *st,
struct imsm_update_rwh_policy *u = (void *)update->buf;
int target = u->dev_idx;
struct imsm_dev *dev = get_imsm_dev(super, target);
- if (!dev) {
- dprintf("could not find subarray-%d\n", target);
- break;
- }
if (dev->rwh_policy != u->new_policy) {
dev->rwh_policy = u->new_policy;
@@ -11397,8 +11417,10 @@ static int imsm_create_metadata_update_for_migration(
{
struct intel_super *super = st->sb;
int update_memory_size;
+ int current_chunk_size;
struct imsm_update_reshape_migration *u;
- struct imsm_dev *dev;
+ struct imsm_dev *dev = get_imsm_dev(super, super->current_vol);
+ struct imsm_map *map = get_imsm_map(dev, MAP_0);
int previous_level = -1;
dprintf("(enter) New Level = %i\n", geo->level);
@@ -11415,23 +11437,15 @@ static int imsm_create_metadata_update_for_migration(
u->new_disks[0] = -1;
u->new_chunksize = -1;
- dev = get_imsm_dev(super, u->subdev);
- if (dev) {
- struct imsm_map *map;
+ current_chunk_size = __le16_to_cpu(map->blocks_per_strip) / 2;
- map = get_imsm_map(dev, MAP_0);
- if (map) {
- int current_chunk_size =
- __le16_to_cpu(map->blocks_per_strip) / 2;
-
- if (geo->chunksize != current_chunk_size) {
- u->new_chunksize = geo->chunksize / 1024;
- dprintf("imsm: chunk size change from %i to %i\n",
- current_chunk_size, u->new_chunksize);
- }
- previous_level = map->raid_level;
- }
+ if (geo->chunksize != current_chunk_size) {
+ u->new_chunksize = geo->chunksize / 1024;
+ dprintf("imsm: chunk size change from %i to %i\n",
+ current_chunk_size, u->new_chunksize);
}
+ previous_level = map->raid_level;
+
if (geo->level == 5 && previous_level == 0) {
struct mdinfo *spares = NULL;
@@ -12519,9 +12533,6 @@ static int validate_internal_bitmap_imsm(struct supertype *st)
unsigned long long offset;
struct dl *d;
- if (!dev)
- return -1;
-
if (dev->rwh_policy != RWH_BITMAP)
return 0;
@@ -12567,16 +12578,8 @@ static int add_internal_bitmap_imsm(struct supertype *st, int *chunkp,
return -1;
dev = get_imsm_dev(super, vol_idx);
-
- if (!dev) {
- dprintf("cannot find the device for volume index %d\n",
- vol_idx);
- return -1;
- }
dev->rwh_policy = RWH_BITMAP;
-
*chunkp = calculate_bitmap_chunksize(st, dev);
-
return 0;
}
--
2.31.1

@ -0,0 +1,85 @@
From 190dc029b141c423e724566cbed5d5afbb10b05a Mon Sep 17 00:00:00 2001
From: Nigel Croxon <ncroxon@redhat.com>
Date: Mon, 18 Apr 2022 13:44:23 -0400
Subject: [PATCH 20/52] Revert "mdadm: fix coredump of mdadm --monitor -r"
This reverts commit 546047688e1c64638f462147c755b58119cabdc8.
The change from commit mdadm: fix coredump of mdadm
--monitor -r broke the printing of the return message when
passing -r to mdadm --manage, the removal of a device from
an array.
If the current code reverts this commit, both issues are
still fixed.
The original problem reported that the fix tried to address
was: The --monitor -r option requires a parameter,
otherwise a null pointer will be manipulated when
converting to integer data, and a core dump will appear.
The original problem was really fixed with:
60815698c0a Refactor parse_num and use it to parse optarg.
Which added a check for NULL in 'optarg' before moving it
to the 'increments' variable.
New issue: When trying to remove a device using the short
argument -r, instead of the long argument --remove, the
output is empty. The problem started when commit
546047688e1c was added.
Steps to Reproduce:
1. create/assemble /dev/md0 device
2. mdadm --manage /dev/md0 -r /dev/vdxx
Actual results:
Nothing, empty output, nothing happens, the device is still
connected to the array.
The output should have stated "mdadm: hot remove failed
for /dev/vdxx: Device or resource busy", if the device was
still active. Or it should remove the device and print
a message:
mdadm: set /dev/vdd faulty in /dev/md0
mdadm: hot removed /dev/vdd from /dev/md0
The following commit should be reverted as it breaks
mdadm --manage -r.
commit 546047688e1c64638f462147c755b58119cabdc8
Author: Wu Guanghao <wuguanghao3@huawei.com>
Date: Mon Aug 16 15:24:51 2021 +0800
mdadm: fix coredump of mdadm --monitor -r
-Nigel
Signed-off-by: Nigel Croxon <ncroxon@redhat.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
ReadMe.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/ReadMe.c b/ReadMe.c
index 8f873c48..bec1be9a 100644
--- a/ReadMe.c
+++ b/ReadMe.c
@@ -81,11 +81,11 @@ char Version[] = "mdadm - v" VERSION " - " VERS_DATE EXTRAVERSION "\n";
* found, it is started.
*/
-char short_options[]="-ABCDEFGIQhVXYWZ:vqbc:i:l:p:m:r:n:x:u:c:d:z:U:N:safRSow1tye:k";
+char short_options[]="-ABCDEFGIQhVXYWZ:vqbc:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:k:";
char short_bitmap_options[]=
- "-ABCDEFGIQhVXYWZ:vqb:c:i:l:p:m:r:n:x:u:c:d:z:U:N:sarfRSow1tye:k:";
+ "-ABCDEFGIQhVXYWZ:vqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:k:";
char short_bitmap_auto_options[]=
- "-ABCDEFGIQhVXYWZ:vqb:c:i:l:p:m:r:n:x:u:c:d:z:U:N:sa:rfRSow1tye:k:";
+ "-ABCDEFGIQhVXYWZ:vqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sa:rfRSow1tye:k:";
struct option long_options[] = {
{"manage", 0, 0, ManageOpt},
--
2.31.1

@ -0,0 +1,31 @@
From 953cc7e5a485a91ddec7312c7a5d7779749fad5f Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Tue, 21 Jun 2022 00:10:39 +0800
Subject: [PATCH 21/52] util: replace ioctl use with function
Replace using of ioctl calling to get md array info with
special function prepared to it.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
util.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/util.c b/util.c
index cc94f96e..38f0420e 100644
--- a/util.c
+++ b/util.c
@@ -267,7 +267,7 @@ int md_array_active(int fd)
* GET_ARRAY_INFO doesn't provide access to the proper state
* information, so fallback to a basic check for raid_disks != 0
*/
- ret = ioctl(fd, GET_ARRAY_INFO, &array);
+ ret = md_get_array_info(fd, &array);
}
return !ret;
--
2.31.1

@ -0,0 +1,110 @@
From 63902857b98c37c8ac4b837bb01d006b327a4532 Mon Sep 17 00:00:00 2001
From: Heming Zhao <heming.zhao@suse.com>
Date: Tue, 21 Jun 2022 00:10:40 +0800
Subject: [PATCH 22/52] mdadm/super1: restore commit 45a87c2f31335 to fix
clustered slot issue
Commit 9d67f6496c71 ("mdadm:check the nodes when operate clustered
array") modified assignment logic for st->nodes in write_bitmap1(),
which introduced bitmap slot issue:
load_super1 didn't set up supertype.nodes, which made spare disk only
have one slot info. Then it triggered kernel md_bitmap_load_sb to get
wrong bitmap slot data.
For fixing this issue, there are two methods:
1> revert the related code of commit 9d67f6496c71. and restore the code
from former commit 45a87c2f31335 ("super1: add more checks for
NodeNumUpdate option").
st->nodes value would be 0 & 1 under current code logic. i.e.
When adding a spare disk, there is no place to init st->nodes, and
the value is ZERO.
2> keep 9d67f6496c71, add additional ->nodes handling in load_super1(),
let load_super1 to set st->nodes when bitmap is BITMAP_MAJOR_CLUSTERED.
Under current mdadm code logic, load_super1 will be called many
times, any new code in load_super1 will cost mdadm running more time.
And more reason is I prefer as much as possible to limit clustered
code spreading in every corner.
So I used method <1> to fix this issue.
How to trigger:
dd if=/dev/zero bs=1M count=1 oflag=direct of=/dev/sda
dd if=/dev/zero bs=1M count=1 oflag=direct of=/dev/sdb
dd if=/dev/zero bs=1M count=1 oflag=direct of=/dev/sdc
mdadm -C /dev/md0 -b clustered -e 1.2 -n 2 -l mirror /dev/sda /dev/sdb
mdadm -a /dev/md0 /dev/sdc
mdadm /dev/md0 --fail /dev/sda
mdadm /dev/md0 --remove /dev/sda
mdadm -Ss
mdadm -A /dev/md0 /dev/sdb /dev/sdc
the output of current "mdadm -X /dev/sdc":
(there should be (by default) 4 slot info for correct output)
```
Filename : /dev/sdc
Magic : 6d746962
Version : 5
UUID : a74642f8:a6b1fba8:58e1f8db:cfe7b082
Events : 29
Events Cleared : 0
State : OK
Chunksize : 64 MB
Daemon : 5s flush period
Write Mode : Normal
Sync Size : 306176 (299.00 MiB 313.52 MB)
Bitmap : 5 bits (chunks), 5 dirty (100.0%)
```
And mdadm later operations will trigger kernel output error message:
(triggered by "mdadm -A /dev/md0 /dev/sdb /dev/sdc")
```
kernel: md0: invalid bitmap file superblock: bad magic
kernel: md_bitmap_copy_from_slot can't get bitmap from slot 1
kernel: md-cluster: Could not gather bitmaps from slot 1
kernel: md0: invalid bitmap file superblock: bad magic
kernel: md_bitmap_copy_from_slot can't get bitmap from slot 2
kernel: md-cluster: Could not gather bitmaps from slot 2
kernel: md0: invalid bitmap file superblock: bad magic
kernel: md_bitmap_copy_from_slot can't get bitmap from slot 3
kernel: md-cluster: Could not gather bitmaps from slot 3
kernel: md-cluster: failed to gather all resyn infos
kernel: md0: detected capacity change from 0 to 612352
```
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super1.c | 12 +++++++++++-
1 file changed, 11 insertions(+), 1 deletion(-)
diff --git a/super1.c b/super1.c
index e3e2f954..3a0c69fd 100644
--- a/super1.c
+++ b/super1.c
@@ -2674,7 +2674,17 @@ static int write_bitmap1(struct supertype *st, int fd, enum bitmap_update update
}
if (bms->version == BITMAP_MAJOR_CLUSTERED) {
- if (__cpu_to_le32(st->nodes) < bms->nodes) {
+ if (st->nodes == 1) {
+ /* the parameter for nodes is not valid */
+ pr_err("Warning: cluster-md at least needs two nodes\n");
+ return -EINVAL;
+ } else if (st->nodes == 0) {
+ /*
+ * parameter "--nodes" is not specified, (eg, add a disk to
+ * clustered raid)
+ */
+ break;
+ } else if (__cpu_to_le32(st->nodes) < bms->nodes) {
/*
* Since the nodes num is not increased, no
* need to check the space enough or not,
--
2.31.1

@ -0,0 +1,122 @@
From 76c152ca9851e9fcdf52e8f6e7e6c09b936bdd14 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 21 Jun 2022 00:10:41 +0800
Subject: [PATCH 23/52] imsm: introduce get_disk_slot_in_dev()
The routine was added to remove unnecessary get_imsm_dev() and
get_imsm_map() calls, used only to determine disk slot.
Additionally, enum for IMSM return statues was added for further usage.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 47 ++++++++++++++++++++++++++++++++++++-----------
1 file changed, 36 insertions(+), 11 deletions(-)
diff --git a/super-intel.c b/super-intel.c
index 3788feb9..cd1f1e3d 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -366,6 +366,18 @@ struct migr_record {
};
ASSERT_SIZE(migr_record, 128)
+/**
+ * enum imsm_status - internal IMSM return values representation.
+ * @STATUS_OK: function succeeded.
+ * @STATUS_ERROR: General error ocurred (not specified).
+ *
+ * Typedefed to imsm_status_t.
+ */
+typedef enum imsm_status {
+ IMSM_STATUS_ERROR = -1,
+ IMSM_STATUS_OK = 0,
+} imsm_status_t;
+
struct md_list {
/* usage marker:
* 1: load metadata
@@ -1183,7 +1195,7 @@ static void set_imsm_ord_tbl_ent(struct imsm_map *map, int slot, __u32 ord)
map->disk_ord_tbl[slot] = __cpu_to_le32(ord);
}
-static int get_imsm_disk_slot(struct imsm_map *map, unsigned idx)
+static int get_imsm_disk_slot(struct imsm_map *map, const unsigned int idx)
{
int slot;
__u32 ord;
@@ -1194,7 +1206,7 @@ static int get_imsm_disk_slot(struct imsm_map *map, unsigned idx)
return slot;
}
- return -1;
+ return IMSM_STATUS_ERROR;
}
static int get_imsm_raid_level(struct imsm_map *map)
@@ -1209,6 +1221,23 @@ static int get_imsm_raid_level(struct imsm_map *map)
return map->raid_level;
}
+/**
+ * get_disk_slot_in_dev() - retrieve disk slot from &imsm_dev.
+ * @super: &intel_super pointer, not NULL.
+ * @dev_idx: imsm device index.
+ * @idx: disk index.
+ *
+ * Return: Slot on success, IMSM_STATUS_ERROR otherwise.
+ */
+static int get_disk_slot_in_dev(struct intel_super *super, const __u8 dev_idx,
+ const unsigned int idx)
+{
+ struct imsm_dev *dev = get_imsm_dev(super, dev_idx);
+ struct imsm_map *map = get_imsm_map(dev, MAP_0);
+
+ return get_imsm_disk_slot(map, idx);
+}
+
static int cmp_extent(const void *av, const void *bv)
{
const struct extent *a = av;
@@ -1225,13 +1254,9 @@ static int count_memberships(struct dl *dl, struct intel_super *super)
int memberships = 0;
int i;
- for (i = 0; i < super->anchor->num_raid_devs; i++) {
- struct imsm_dev *dev = get_imsm_dev(super, i);
- struct imsm_map *map = get_imsm_map(dev, MAP_0);
-
- if (get_imsm_disk_slot(map, dl->index) >= 0)
+ for (i = 0; i < super->anchor->num_raid_devs; i++)
+ if (get_disk_slot_in_dev(super, i, dl->index) >= 0)
memberships++;
- }
return memberships;
}
@@ -1941,6 +1966,7 @@ void examine_migr_rec_imsm(struct intel_super *super)
/* first map under migration */
map = get_imsm_map(dev, MAP_0);
+
if (map)
slot = get_imsm_disk_slot(map, super->disks->index);
if (map == NULL || slot > 1 || slot < 0) {
@@ -9655,10 +9681,9 @@ static int apply_update_activate_spare(struct imsm_update_activate_spare *u,
/* count arrays using the victim in the metadata */
found = 0;
for (a = active_array; a ; a = a->next) {
- dev = get_imsm_dev(super, a->info.container_member);
- map = get_imsm_map(dev, MAP_0);
+ int dev_idx = a->info.container_member;
- if (get_imsm_disk_slot(map, victim) >= 0)
+ if (get_disk_slot_in_dev(super, dev_idx, victim) >= 0)
found++;
}
--
2.31.1

@ -0,0 +1,252 @@
From 6d4d9ab295de165e57b5c30e044028dbffb8f297 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 21 Jun 2022 00:10:42 +0800
Subject: [PATCH 24/52] imsm: use same slot across container
Autolayout relies on drives order on super->disks list, but
it is not quaranted by readdir() in sysfs_read(). As a result
drive could be put in different slot in second volume.
Make it consistent by reffering to first volume, if exists.
Use enum imsm_status to unify error handling.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 169 ++++++++++++++++++++++++++++++++------------------
1 file changed, 108 insertions(+), 61 deletions(-)
diff --git a/super-intel.c b/super-intel.c
index cd1f1e3d..deef7c87 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -7522,11 +7522,27 @@ static int validate_geometry_imsm_volume(struct supertype *st, int level,
return 1;
}
-static int imsm_get_free_size(struct supertype *st, int raiddisks,
- unsigned long long size, int chunk,
- unsigned long long *freesize)
+/**
+ * imsm_get_free_size() - get the biggest, common free space from members.
+ * @super: &intel_super pointer, not NULL.
+ * @raiddisks: number of raid disks.
+ * @size: requested size, could be 0 (means max size).
+ * @chunk: requested chunk.
+ * @freesize: pointer for returned size value.
+ *
+ * Return: &IMSM_STATUS_OK or &IMSM_STATUS_ERROR.
+ *
+ * @freesize is set to meaningful value, this can be @size, or calculated
+ * max free size.
+ * super->create_offset value is modified and set appropriately in
+ * merge_extends() for further creation.
+ */
+static imsm_status_t imsm_get_free_size(struct intel_super *super,
+ const int raiddisks,
+ unsigned long long size,
+ const int chunk,
+ unsigned long long *freesize)
{
- struct intel_super *super = st->sb;
struct imsm_super *mpb = super->anchor;
struct dl *dl;
int i;
@@ -7570,12 +7586,10 @@ static int imsm_get_free_size(struct supertype *st, int raiddisks,
/* chunk is in K */
minsize = chunk * 2;
- if (cnt < raiddisks ||
- (super->orom && used && used != raiddisks) ||
- maxsize < minsize ||
- maxsize == 0) {
+ if (cnt < raiddisks || (super->orom && used && used != raiddisks) ||
+ maxsize < minsize || maxsize == 0) {
pr_err("not enough devices with space to create array.\n");
- return 0; /* No enough free spaces large enough */
+ return IMSM_STATUS_ERROR;
}
if (size == 0) {
@@ -7588,37 +7602,69 @@ static int imsm_get_free_size(struct supertype *st, int raiddisks,
}
if (mpb->num_raid_devs > 0 && size && size != maxsize)
pr_err("attempting to create a second volume with size less then remaining space.\n");
- cnt = 0;
- for (dl = super->disks; dl; dl = dl->next)
- if (dl->e)
- dl->raiddisk = cnt++;
-
*freesize = size;
dprintf("imsm: imsm_get_free_size() returns : %llu\n", size);
- return 1;
+ return IMSM_STATUS_OK;
}
-static int reserve_space(struct supertype *st, int raiddisks,
- unsigned long long size, int chunk,
- unsigned long long *freesize)
+/**
+ * autolayout_imsm() - automatically layout a new volume.
+ * @super: &intel_super pointer, not NULL.
+ * @raiddisks: number of raid disks.
+ * @size: requested size, could be 0 (means max size).
+ * @chunk: requested chunk.
+ * @freesize: pointer for returned size value.
+ *
+ * We are being asked to automatically layout a new volume based on the current
+ * contents of the container. If the parameters can be satisfied autolayout_imsm
+ * will record the disks, start offset, and will return size of the volume to
+ * be created. See imsm_get_free_size() for details.
+ * add_to_super() and getinfo_super() detect when autolayout is in progress.
+ * If first volume exists, slots are set consistently to it.
+ *
+ * Return: &IMSM_STATUS_OK on success, &IMSM_STATUS_ERROR otherwise.
+ *
+ * Disks are marked for creation via dl->raiddisk.
+ */
+static imsm_status_t autolayout_imsm(struct intel_super *super,
+ const int raiddisks,
+ unsigned long long size, const int chunk,
+ unsigned long long *freesize)
{
- struct intel_super *super = st->sb;
- struct dl *dl;
- int cnt;
- int rv = 0;
+ int curr_slot = 0;
+ struct dl *disk;
+ int vol_cnt = super->anchor->num_raid_devs;
+ imsm_status_t rv;
- rv = imsm_get_free_size(st, raiddisks, size, chunk, freesize);
- if (rv) {
- cnt = 0;
- for (dl = super->disks; dl; dl = dl->next)
- if (dl->e)
- dl->raiddisk = cnt++;
- rv = 1;
+ rv = imsm_get_free_size(super, raiddisks, size, chunk, freesize);
+ if (rv != IMSM_STATUS_OK)
+ return IMSM_STATUS_ERROR;
+
+ for (disk = super->disks; disk; disk = disk->next) {
+ if (!disk->e)
+ continue;
+
+ if (curr_slot == raiddisks)
+ break;
+
+ if (vol_cnt == 0) {
+ disk->raiddisk = curr_slot;
+ } else {
+ int _slot = get_disk_slot_in_dev(super, 0, disk->index);
+
+ if (_slot == -1) {
+ pr_err("Disk %s is not used in first volume, aborting\n",
+ disk->devname);
+ return IMSM_STATUS_ERROR;
+ }
+ disk->raiddisk = _slot;
+ }
+ curr_slot++;
}
- return rv;
+ return IMSM_STATUS_OK;
}
static int validate_geometry_imsm(struct supertype *st, int level, int layout,
@@ -7654,35 +7700,35 @@ static int validate_geometry_imsm(struct supertype *st, int level, int layout,
}
if (!dev) {
- if (st->sb) {
- struct intel_super *super = st->sb;
- if (!validate_geometry_imsm_orom(st->sb, level, layout,
- raiddisks, chunk, size,
- verbose))
+ struct intel_super *super = st->sb;
+
+ /*
+ * Autolayout mode, st->sb and freesize must be set.
+ */
+ if (!super || !freesize) {
+ pr_vrb("freesize and superblock must be set for autolayout, aborting\n");
+ return 1;
+ }
+
+ if (!validate_geometry_imsm_orom(st->sb, level, layout,
+ raiddisks, chunk, size,
+ verbose))
+ return 0;
+
+ if (super->orom) {
+ imsm_status_t rv;
+ int count = count_volumes(super->hba, super->orom->dpa,
+ verbose);
+ if (super->orom->vphba <= count) {
+ pr_vrb("platform does not support more than %d raid volumes.\n",
+ super->orom->vphba);
return 0;
- /* we are being asked to automatically layout a
- * new volume based on the current contents of
- * the container. If the the parameters can be
- * satisfied reserve_space will record the disks,
- * start offset, and size of the volume to be
- * created. add_to_super and getinfo_super
- * detect when autolayout is in progress.
- */
- /* assuming that freesize is always given when array is
- created */
- if (super->orom && freesize) {
- int count;
- count = count_volumes(super->hba,
- super->orom->dpa, verbose);
- if (super->orom->vphba <= count) {
- pr_vrb("platform does not support more than %d raid volumes.\n",
- super->orom->vphba);
- return 0;
- }
}
- if (freesize)
- return reserve_space(st, raiddisks, size,
- *chunk, freesize);
+
+ rv = autolayout_imsm(super, raiddisks, size, *chunk,
+ freesize);
+ if (rv != IMSM_STATUS_OK)
+ return 0;
}
return 1;
}
@@ -11538,7 +11584,7 @@ enum imsm_reshape_type imsm_analyze_change(struct supertype *st,
unsigned long long current_size;
unsigned long long free_size;
unsigned long long max_size;
- int rv;
+ imsm_status_t rv;
getinfo_super_imsm_volume(st, &info, NULL);
if (geo->level != info.array.level && geo->level >= 0 &&
@@ -11657,9 +11703,10 @@ enum imsm_reshape_type imsm_analyze_change(struct supertype *st,
}
/* check the maximum available size
*/
- rv = imsm_get_free_size(st, dev->vol.map->num_members,
- 0, chunk, &free_size);
- if (rv == 0)
+ rv = imsm_get_free_size(super, dev->vol.map->num_members,
+ 0, chunk, &free_size);
+
+ if (rv != IMSM_STATUS_OK)
/* Cannot find maximum available space
*/
max_size = 0;
--
2.31.1

@ -0,0 +1,122 @@
From 9a7df595bbe360132cb37c8b39aa1fd9ac24b43f Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 21 Jun 2022 00:10:43 +0800
Subject: [PATCH 25/52] imsm: block changing slots during creation
If user specifies drives for array creation, then slot order across
volumes is not preserved.
Ideally, it should be checked in validate_geometry() but it is not
possible in current implementation (order is determined later).
Add verification in add_to_super_imsm_volume() and throw error if
mismatch is detected.
IMSM allows to use only same members within container.
This is not hardware dependency but metadata limitation.
Therefore, 09-imsm-overlap test is removed. Testing it is pointless.
After this patch, creation in this scenario is blocked. Offset
verification is covered in other tests.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 33 ++++++++++++++++++++++-----------
tests/09imsm-overlap | 28 ----------------------------
2 files changed, 22 insertions(+), 39 deletions(-)
delete mode 100644 tests/09imsm-overlap
diff --git a/super-intel.c b/super-intel.c
index deef7c87..8ffe485c 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -5789,6 +5789,10 @@ static int add_to_super_imsm_volume(struct supertype *st, mdu_disk_info_t *dk,
struct imsm_map *map;
struct dl *dl, *df;
int slot;
+ int autolayout = 0;
+
+ if (!is_fd_valid(fd))
+ autolayout = 1;
dev = get_imsm_dev(super, super->current_vol);
map = get_imsm_map(dev, MAP_0);
@@ -5799,25 +5803,32 @@ static int add_to_super_imsm_volume(struct supertype *st, mdu_disk_info_t *dk,
return 1;
}
- if (!is_fd_valid(fd)) {
- /* we're doing autolayout so grab the pre-marked (in
- * validate_geometry) raid_disk
- */
- for (dl = super->disks; dl; dl = dl->next)
+ for (dl = super->disks; dl ; dl = dl->next) {
+ if (autolayout) {
if (dl->raiddisk == dk->raid_disk)
break;
- } else {
- for (dl = super->disks; dl ; dl = dl->next)
- if (dl->major == dk->major &&
- dl->minor == dk->minor)
- break;
+ } else if (dl->major == dk->major && dl->minor == dk->minor)
+ break;
}
if (!dl) {
- pr_err("%s is not a member of the same container\n", devname);
+ if (!autolayout)
+ pr_err("%s is not a member of the same container.\n",
+ devname);
return 1;
}
+ if (!autolayout && super->current_vol > 0) {
+ int _slot = get_disk_slot_in_dev(super, 0, dl->index);
+
+ if (_slot != dk->raid_disk) {
+ pr_err("Member %s is in %d slot for the first volume, but is in %d slot for a new volume.\n",
+ dl->devname, _slot, dk->raid_disk);
+ pr_err("Raid members are in different order than for the first volume, aborting.\n");
+ return 1;
+ }
+ }
+
if (mpb->num_disks == 0)
if (!get_dev_sector_size(dl->fd, dl->devname,
&super->sector_size))
diff --git a/tests/09imsm-overlap b/tests/09imsm-overlap
deleted file mode 100644
index ff5d2093..00000000
--- a/tests/09imsm-overlap
+++ /dev/null
@@ -1,28 +0,0 @@
-
-. tests/env-imsm-template
-
-# create raid arrays with varying degress of overlap
-mdadm -CR $container -e imsm -n 6 $dev0 $dev1 $dev2 $dev3 $dev4 $dev5
-imsm_check container 6
-
-size=1024
-level=1
-num_disks=2
-mdadm -CR $member0 $dev0 $dev1 -n $num_disks -l $level -z $size
-mdadm -CR $member1 $dev1 $dev2 -n $num_disks -l $level -z $size
-mdadm -CR $member2 $dev2 $dev3 -n $num_disks -l $level -z $size
-mdadm -CR $member3 $dev3 $dev4 -n $num_disks -l $level -z $size
-mdadm -CR $member4 $dev4 $dev5 -n $num_disks -l $level -z $size
-
-udevadm settle
-
-offset=0
-imsm_check member $member0 $num_disks $level $size 1024 $offset
-offset=$((offset+size+4096))
-imsm_check member $member1 $num_disks $level $size 1024 $offset
-offset=$((offset+size+4096))
-imsm_check member $member2 $num_disks $level $size 1024 $offset
-offset=$((offset+size+4096))
-imsm_check member $member3 $num_disks $level $size 1024 $offset
-offset=$((offset+size+4096))
-imsm_check member $member4 $num_disks $level $size 1024 $offset
--
2.31.1

@ -0,0 +1,177 @@
From 70f1ff4291b0388adca1f4c91918ce1175e8b360 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Wed, 15 Jun 2022 14:28:39 +0200
Subject: [PATCH 26/52] mdadm: block update=ppl for non raid456 levels
Option ppl should be used only for raid levels 4, 5 and 6. Cancel update
for other levels.
Applied globally for imsm and ddf format.
Additionally introduce is_level456() helper function.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Assemble.c | 11 +++++------
Grow.c | 2 +-
Manage.c | 14 ++++++++++++--
mdadm.h | 11 +++++++++++
super0.c | 2 +-
super1.c | 3 +--
6 files changed, 31 insertions(+), 12 deletions(-)
diff --git a/Assemble.c b/Assemble.c
index 4b213560..6df6bfbc 100644
--- a/Assemble.c
+++ b/Assemble.c
@@ -906,8 +906,7 @@ static int force_array(struct mdinfo *content,
* devices in RAID4 or last devices in RAID4/5/6.
*/
delta = devices[j].i.delta_disks;
- if (devices[j].i.array.level >= 4 &&
- devices[j].i.array.level <= 6 &&
+ if (is_level456(devices[j].i.array.level) &&
i/2 >= content->array.raid_disks - delta)
/* OK */;
else if (devices[j].i.array.level == 4 &&
@@ -1226,8 +1225,7 @@ static int start_array(int mdfd,
fprintf(stderr, ".\n");
}
if (content->reshape_active &&
- content->array.level >= 4 &&
- content->array.level <= 6) {
+ is_level456(content->array.level)) {
/* might need to increase the size
* of the stripe cache - default is 256
*/
@@ -1974,7 +1972,8 @@ int assemble_container_content(struct supertype *st, int mdfd,
int start_reshape;
char *avail;
int err;
- int is_raid456, is_clean, all_disks;
+ int is_clean, all_disks;
+ bool is_raid456;
if (sysfs_init(content, mdfd, NULL)) {
pr_err("Unable to initialize sysfs\n");
@@ -2107,7 +2106,7 @@ int assemble_container_content(struct supertype *st, int mdfd,
content->array.state |= 1;
}
- is_raid456 = (content->array.level >= 4 && content->array.level <= 6);
+ is_raid456 = is_level456(content->array.level);
is_clean = content->array.state & 1;
if (enough(content->array.level, content->array.raid_disks,
diff --git a/Grow.c b/Grow.c
index f6efbc48..8c520d42 100644
--- a/Grow.c
+++ b/Grow.c
@@ -2944,7 +2944,7 @@ static int impose_level(int fd, int level, char *devname, int verbose)
}
md_get_array_info(fd, &array);
- if (level == 0 && (array.level >= 4 && array.level <= 6)) {
+ if (level == 0 && is_level456(array.level)) {
/* To convert to RAID0 we need to fail and
* remove any non-data devices. */
int found = 0;
diff --git a/Manage.c b/Manage.c
index f789e0c1..e5e6abe4 100644
--- a/Manage.c
+++ b/Manage.c
@@ -307,7 +307,7 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
* - unfreeze reshape
* - wait on 'sync_completed' for that point to be reached.
*/
- if (mdi && (mdi->array.level >= 4 && mdi->array.level <= 6) &&
+ if (mdi && is_level456(mdi->array.level) &&
sysfs_attribute_available(mdi, NULL, "sync_action") &&
sysfs_attribute_available(mdi, NULL, "reshape_direction") &&
sysfs_get_str(mdi, NULL, "sync_action", buf, 20) > 0 &&
@@ -1679,6 +1679,7 @@ int Update_subarray(char *dev, char *subarray, char *update, struct mddev_ident
{
struct supertype supertype, *st = &supertype;
int fd, rv = 2;
+ struct mdinfo *info = NULL;
memset(st, 0, sizeof(*st));
@@ -1696,6 +1697,13 @@ int Update_subarray(char *dev, char *subarray, char *update, struct mddev_ident
if (mdmon_running(st->devnm))
st->update_tail = &st->updates;
+ info = st->ss->container_content(st, subarray);
+
+ if (strncmp(update, "ppl", 3) == 0 && !is_level456(info->array.level)) {
+ pr_err("RWH policy ppl is supported only for raid4, raid5 and raid6.\n");
+ goto free_super;
+ }
+
rv = st->ss->update_subarray(st, subarray, update, ident);
if (rv) {
@@ -1711,7 +1719,9 @@ int Update_subarray(char *dev, char *subarray, char *update, struct mddev_ident
pr_err("Updated subarray-%s name from %s, UUIDs may have changed\n",
subarray, dev);
- free_super:
+free_super:
+ if (info)
+ free(info);
st->ss->free_super(st);
close(fd);
diff --git a/mdadm.h b/mdadm.h
index d53df169..974415b9 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -796,6 +796,17 @@ static inline int is_fd_valid(int fd)
return (fd > -1);
}
+/**
+ * is_level456() - check whether given level is between inclusive 4 and 6.
+ * @level: level to check.
+ *
+ * Return: true if condition is met, false otherwise
+ */
+static inline bool is_level456(int level)
+{
+ return (level >= 4 && level <= 6);
+}
+
/**
* close_fd() - verify, close and unset file descriptor.
* @fd: pointer to file descriptor.
diff --git a/super0.c b/super0.c
index 61c9ec1d..37f595ed 100644
--- a/super0.c
+++ b/super0.c
@@ -683,7 +683,7 @@ static int update_super0(struct supertype *st, struct mdinfo *info,
int parity = sb->level == 6 ? 2 : 1;
rv = 0;
- if (sb->level >= 4 && sb->level <= 6 &&
+ if (is_level456(sb->level) &&
sb->reshape_position % (
sb->new_chunk/512 *
(sb->raid_disks - sb->delta_disks - parity))) {
diff --git a/super1.c b/super1.c
index 3a0c69fd..71af860c 100644
--- a/super1.c
+++ b/super1.c
@@ -1530,8 +1530,7 @@ static int update_super1(struct supertype *st, struct mdinfo *info,
* So we reject a revert-reshape unless the
* alignment is good.
*/
- if (__le32_to_cpu(sb->level) >= 4 &&
- __le32_to_cpu(sb->level) <= 6) {
+ if (is_level456(__le32_to_cpu(sb->level))) {
reshape_sectors =
__le64_to_cpu(sb->reshape_position);
reshape_chunk = __le32_to_cpu(sb->new_chunk);
--
2.31.1

@ -0,0 +1,30 @@
From 42e02e613fb0b4a2c0c0d984b9e6e2933875bb44 Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 22 Jul 2022 08:43:47 +0200
Subject: [PATCH 27/52] mdadm: Fix array size mismatch after grow
imsm_fix_size_mismatch() is invoked to fix the problem, but it couldn't
proceed due to migration check. This patch allows for intended behavior.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/super-intel.c b/super-intel.c
index 8ffe485c..76b947f5 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -11854,7 +11854,7 @@ static int imsm_fix_size_mismatch(struct supertype *st, int subarray_index)
unsigned long long d_size = imsm_dev_size(dev);
int u_size;
- if (calc_size == d_size || dev->vol.migr_type == MIGR_GEN_MIGR)
+ if (calc_size == d_size)
continue;
/* There is a difference, confirm that imsm_dev_size is
--
2.31.1

@ -0,0 +1,34 @@
From 751757620afb25a4c02746bf8368a7b5f22352ec Mon Sep 17 00:00:00 2001
From: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Date: Fri, 22 Jul 2022 08:43:48 +0200
Subject: [PATCH 28/52] mdadm: Remove dead code in imsm_fix_size_mismatch
imsm_create_metadata_update_for_size_change() that returns u_size value
could return 0 in the past. As its behavior changed, and returned value
is always the size of imsm_update_size_change structure, check for
u_size is no longer needed.
Signed-off-by: Lukasz Florczak <lukasz.florczak@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
super-intel.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/super-intel.c b/super-intel.c
index 76b947f5..4ddfcf94 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -11869,10 +11869,6 @@ static int imsm_fix_size_mismatch(struct supertype *st, int subarray_index)
geo.size = d_size;
u_size = imsm_create_metadata_update_for_size_change(st, &geo,
&update);
- if (u_size < 1) {
- dprintf("imsm: Cannot prepare size change update\n");
- goto exit;
- }
imsm_update_metadata_locally(st, update, u_size);
if (st->update_tail) {
append_metadata_update(st, update, u_size);
--
2.31.1

@ -0,0 +1,40 @@
From c8d1c398505b62d9129a4e711f17e4469f4327ff Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Thu, 14 Jul 2022 09:02:10 +0200
Subject: [PATCH 29/52] Monitor: use devname as char array instead of pointer
Device name wasn't filled properly due to incorrect use of strcpy.
Strcpy was used twice. Firstly to fill devname with "/dev/md/"
and then to add chosen name. First strcpy result was overwritten by
second one (as a result <device_name> instead of "/dev/md/<device_name>"
was assigned). This commit changes this implementation to use snprintf
and devname with fixed size.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Monitor.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/Monitor.c b/Monitor.c
index 6ca1ebe5..a5b11ae2 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -190,9 +190,11 @@ int Monitor(struct mddev_dev *devlist,
if (mdlist->devname[0] == '/')
st->devname = xstrdup(mdlist->devname);
else {
- st->devname = xmalloc(8+strlen(mdlist->devname)+1);
- strcpy(strcpy(st->devname, "/dev/md/"),
- mdlist->devname);
+ /* length of "/dev/md/" + device name + terminating byte */
+ size_t _len = sizeof("/dev/md/") + strnlen(mdlist->devname, PATH_MAX);
+
+ st->devname = xcalloc(_len, sizeof(char));
+ snprintf(st->devname, _len, "/dev/md/%s", mdlist->devname);
}
if (!is_mddev(mdlist->devname))
return 1;
--
2.31.1

@ -0,0 +1,133 @@
From 84d969be8f6d8a345b75f558fad26e4f62a558f6 Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Thu, 14 Jul 2022 09:02:11 +0200
Subject: [PATCH 30/52] Monitor: use snprintf to fill device name
Safe string functions are propagated in Monitor.c.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Monitor.c | 37 ++++++++++++++-----------------------
1 file changed, 14 insertions(+), 23 deletions(-)
diff --git a/Monitor.c b/Monitor.c
index a5b11ae2..93f36ac0 100644
--- a/Monitor.c
+++ b/Monitor.c
@@ -33,8 +33,8 @@
#endif
struct state {
- char *devname;
- char devnm[32]; /* to sync with mdstat info */
+ char devname[MD_NAME_MAX + sizeof("/dev/md/")]; /* length of "/dev/md/" + device name + terminating byte*/
+ char devnm[MD_NAME_MAX]; /* to sync with mdstat info */
unsigned int utime;
int err;
char *spare_group;
@@ -45,9 +45,9 @@ struct state {
int devstate[MAX_DISKS];
dev_t devid[MAX_DISKS];
int percent;
- char parent_devnm[32]; /* For subarray, devnm of parent.
- * For others, ""
- */
+ char parent_devnm[MD_NAME_MAX]; /* For subarray, devnm of parent.
+ * For others, ""
+ */
struct supertype *metadata;
struct state *subarray;/* for a container it is a link to first subarray
* for a subarray it is a link to next subarray
@@ -187,15 +187,8 @@ int Monitor(struct mddev_dev *devlist,
continue;
st = xcalloc(1, sizeof *st);
- if (mdlist->devname[0] == '/')
- st->devname = xstrdup(mdlist->devname);
- else {
- /* length of "/dev/md/" + device name + terminating byte */
- size_t _len = sizeof("/dev/md/") + strnlen(mdlist->devname, PATH_MAX);
-
- st->devname = xcalloc(_len, sizeof(char));
- snprintf(st->devname, _len, "/dev/md/%s", mdlist->devname);
- }
+ snprintf(st->devname, MD_NAME_MAX + sizeof("/dev/md/"),
+ "/dev/md/%s", basename(mdlist->devname));
if (!is_mddev(mdlist->devname))
return 1;
st->next = statelist;
@@ -218,7 +211,7 @@ int Monitor(struct mddev_dev *devlist,
st = xcalloc(1, sizeof *st);
mdlist = conf_get_ident(dv->devname);
- st->devname = xstrdup(dv->devname);
+ snprintf(st->devname, MD_NAME_MAX + sizeof("/dev/md/"), "%s", dv->devname);
st->next = statelist;
st->devnm[0] = 0;
st->percent = RESYNC_UNKNOWN;
@@ -301,7 +294,6 @@ int Monitor(struct mddev_dev *devlist,
for (stp = &statelist; (st = *stp) != NULL; ) {
if (st->from_auto && st->err > 5) {
*stp = st->next;
- free(st->devname);
free(st->spare_group);
free(st);
} else
@@ -554,7 +546,7 @@ static int check_array(struct state *st, struct mdstat_ent *mdstat,
goto disappeared;
if (st->devnm[0] == 0)
- strcpy(st->devnm, fd2devnm(fd));
+ snprintf(st->devnm, MD_NAME_MAX, "%s", fd2devnm(fd));
for (mse2 = mdstat; mse2; mse2 = mse2->next)
if (strcmp(mse2->devnm, st->devnm) == 0) {
@@ -684,7 +676,7 @@ static int check_array(struct state *st, struct mdstat_ent *mdstat,
strncmp(mse->metadata_version, "external:", 9) == 0 &&
is_subarray(mse->metadata_version+9)) {
char *sl;
- strcpy(st->parent_devnm, mse->metadata_version + 10);
+ snprintf(st->parent_devnm, MD_NAME_MAX, "%s", mse->metadata_version + 10);
sl = strchr(st->parent_devnm, '/');
if (sl)
*sl = 0;
@@ -772,14 +764,13 @@ static int add_new_arrays(struct mdstat_ent *mdstat, struct state **statelist,
continue;
}
- st->devname = xstrdup(name);
+ snprintf(st->devname, MD_NAME_MAX + sizeof("/dev/md/"), "%s", name);
if ((fd = open(st->devname, O_RDONLY)) < 0 ||
md_get_array_info(fd, &array) < 0) {
/* no such array */
if (fd >= 0)
close(fd);
put_md_name(st->devname);
- free(st->devname);
if (st->metadata) {
st->metadata->ss->free_super(st->metadata);
free(st->metadata);
@@ -791,7 +782,7 @@ static int add_new_arrays(struct mdstat_ent *mdstat, struct state **statelist,
st->next = *statelist;
st->err = 1;
st->from_auto = 1;
- strcpy(st->devnm, mse->devnm);
+ snprintf(st->devnm, MD_NAME_MAX, "%s", mse->devnm);
st->percent = RESYNC_UNKNOWN;
st->expected_spares = -1;
if (mse->metadata_version &&
@@ -799,8 +790,8 @@ static int add_new_arrays(struct mdstat_ent *mdstat, struct state **statelist,
"external:", 9) == 0 &&
is_subarray(mse->metadata_version+9)) {
char *sl;
- strcpy(st->parent_devnm,
- mse->metadata_version+10);
+ snprintf(st->parent_devnm, MD_NAME_MAX,
+ "%s", mse->metadata_version + 10);
sl = strchr(st->parent_devnm, '/');
*sl = 0;
} else
--
2.31.1

@ -0,0 +1,42 @@
From 14ae4c37bce9a53da08d59d6c2d7e0946e9c9f47 Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:06 -0600
Subject: [PATCH 31/52] Makefile: Don't build static build with everything and
everything-test
Running the test suite requires building everything, but it seems to be
difficult to build the static version of mdadm now seeing there
is no readily available static udev library.
The test suite doesn't need the static binary so just don't build it
with the everything or everything-test targets.
Leave the mdadm.static and install-static targets in place in case
someone still has a use case for the static binary.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile
index bf126033..ec1f99ed 100644
--- a/Makefile
+++ b/Makefile
@@ -182,9 +182,9 @@ check_rundir:
echo "***** or set CHECK_RUN_DIR=0"; exit 1; \
fi
-everything: all mdadm.static swap_super test_stripe raid6check \
+everything: all swap_super test_stripe raid6check \
mdadm.Os mdadm.O2 man
-everything-test: all mdadm.static swap_super test_stripe \
+everything-test: all swap_super test_stripe \
mdadm.Os mdadm.O2 man
# mdadm.uclibc doesn't work on x86-64
# mdadm.tcc doesn't work..
--
2.31.1

@ -0,0 +1,141 @@
From 679bd9508a30b2a0a1baecc9a21dd6c7d8d8d7dc Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:07 -0600
Subject: [PATCH 32/52] DDF: Cleanup validate_geometry_ddf_container()
Move the function up so that the function declaration is not necessary
and remove the unused arguments to the function.
No functional changes are intended but will help with a bug fix in the
next patch.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
super-ddf.c | 88 ++++++++++++++++++++++++-----------------------------
1 file changed, 39 insertions(+), 49 deletions(-)
diff --git a/super-ddf.c b/super-ddf.c
index abbc8b09..9d867f69 100644
--- a/super-ddf.c
+++ b/super-ddf.c
@@ -503,13 +503,6 @@ struct ddf_super {
static int load_super_ddf_all(struct supertype *st, int fd,
void **sbp, char *devname);
static int get_svd_state(const struct ddf_super *, const struct vcl *);
-static int
-validate_geometry_ddf_container(struct supertype *st,
- int level, int layout, int raiddisks,
- int chunk, unsigned long long size,
- unsigned long long data_offset,
- char *dev, unsigned long long *freesize,
- int verbose);
static int validate_geometry_ddf_bvd(struct supertype *st,
int level, int layout, int raiddisks,
@@ -3322,6 +3315,42 @@ static int reserve_space(struct supertype *st, int raiddisks,
return 1;
}
+static int
+validate_geometry_ddf_container(struct supertype *st,
+ int level, int raiddisks,
+ unsigned long long data_offset,
+ char *dev, unsigned long long *freesize,
+ int verbose)
+{
+ int fd;
+ unsigned long long ldsize;
+
+ if (level != LEVEL_CONTAINER)
+ return 0;
+ if (!dev)
+ return 1;
+
+ fd = dev_open(dev, O_RDONLY|O_EXCL);
+ if (fd < 0) {
+ if (verbose)
+ pr_err("ddf: Cannot open %s: %s\n",
+ dev, strerror(errno));
+ return 0;
+ }
+ if (!get_dev_size(fd, dev, &ldsize)) {
+ close(fd);
+ return 0;
+ }
+ close(fd);
+ if (freesize) {
+ *freesize = avail_size_ddf(st, ldsize >> 9, INVALID_SECTORS);
+ if (*freesize == 0)
+ return 0;
+ }
+
+ return 1;
+}
+
static int validate_geometry_ddf(struct supertype *st,
int level, int layout, int raiddisks,
int *chunk, unsigned long long size,
@@ -3347,11 +3376,9 @@ static int validate_geometry_ddf(struct supertype *st,
level = LEVEL_CONTAINER;
if (level == LEVEL_CONTAINER) {
/* Must be a fresh device to add to a container */
- return validate_geometry_ddf_container(st, level, layout,
- raiddisks, *chunk,
- size, data_offset, dev,
- freesize,
- verbose);
+ return validate_geometry_ddf_container(st, level, raiddisks,
+ data_offset, dev,
+ freesize, verbose);
}
if (!dev) {
@@ -3449,43 +3476,6 @@ static int validate_geometry_ddf(struct supertype *st,
return 1;
}
-static int
-validate_geometry_ddf_container(struct supertype *st,
- int level, int layout, int raiddisks,
- int chunk, unsigned long long size,
- unsigned long long data_offset,
- char *dev, unsigned long long *freesize,
- int verbose)
-{
- int fd;
- unsigned long long ldsize;
-
- if (level != LEVEL_CONTAINER)
- return 0;
- if (!dev)
- return 1;
-
- fd = dev_open(dev, O_RDONLY|O_EXCL);
- if (fd < 0) {
- if (verbose)
- pr_err("ddf: Cannot open %s: %s\n",
- dev, strerror(errno));
- return 0;
- }
- if (!get_dev_size(fd, dev, &ldsize)) {
- close(fd);
- return 0;
- }
- close(fd);
- if (freesize) {
- *freesize = avail_size_ddf(st, ldsize >> 9, INVALID_SECTORS);
- if (*freesize == 0)
- return 0;
- }
-
- return 1;
-}
-
static int validate_geometry_ddf_bvd(struct supertype *st,
int level, int layout, int raiddisks,
int *chunk, unsigned long long size,
--
2.31.1

@ -0,0 +1,49 @@
From 2b93288a5650bb811932836f67f30d63c5ddcfbd Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:08 -0600
Subject: [PATCH 33/52] DDF: Fix NULL pointer dereference in
validate_geometry_ddf()
A relatively recent patch added a call to validate_geometry() in
Manage_add() that has level=LEVEL_CONTAINER and chunk=NULL.
This causes some ddf tests to segfault which aborts the test suite.
To fix this, avoid dereferencing chunk when the level is
LEVEL_CONTAINER or LEVEL_NONE.
Fixes: 1f5d54a06df0 ("Manage: Call validate_geometry when adding drive to external container")
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
super-ddf.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/super-ddf.c b/super-ddf.c
index 9d867f69..949e7d15 100644
--- a/super-ddf.c
+++ b/super-ddf.c
@@ -3369,9 +3369,6 @@ static int validate_geometry_ddf(struct supertype *st,
* If given BVDs, we make an SVD, changing all the GUIDs in the process.
*/
- if (*chunk == UnSet)
- *chunk = DEFAULT_CHUNK;
-
if (level == LEVEL_NONE)
level = LEVEL_CONTAINER;
if (level == LEVEL_CONTAINER) {
@@ -3381,6 +3378,9 @@ static int validate_geometry_ddf(struct supertype *st,
freesize, verbose);
}
+ if (*chunk == UnSet)
+ *chunk = DEFAULT_CHUNK;
+
if (!dev) {
mdu_array_info_t array = {
.level = level,
--
2.31.1

@ -0,0 +1,85 @@
From 548e9b916f86c06e2cdb50d8f49633f9bec66c7e Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:09 -0600
Subject: [PATCH 34/52] mdadm/Grow: Fix use after close bug by closing after
fork
The test 07reshape-grow fails most of the time. But it succeeds around
1 in 5 times. When it does succeed, it causes the tests to die because
mdadm has segfaulted.
The segfault was caused by mdadm attempting to repoen a file
descriptor that was already closed. The backtrace of the segfault
was:
#0 __strncmp_avx2 () at ../sysdeps/x86_64/multiarch/strcmp-avx2.S:101
#1 0x000056146e31d44b in devnm2devid (devnm=0x0) at util.c:956
#2 0x000056146e31dab4 in open_dev_flags (devnm=0x0, flags=0)
at util.c:1072
#3 0x000056146e31db22 in open_dev (devnm=0x0) at util.c:1079
#4 0x000056146e3202e8 in reopen_mddev (mdfd=4) at util.c:2244
#5 0x000056146e329f36 in start_array (mdfd=4,
mddev=0x7ffc55342450 "/dev/md0", content=0x7ffc55342860,
st=0x56146fc78660, ident=0x7ffc55342f70, best=0x56146fc6f5d0,
bestcnt=10, chosen_drive=0, devices=0x56146fc706b0, okcnt=5,
sparecnt=0, rebuilding_cnt=0, journalcnt=0, c=0x7ffc55342e90,
clean=1, avail=0x56146fc78720 "\001\001\001\001\001",
start_partial_ok=0, err_ok=0, was_forced=0)
at Assemble.c:1206
#6 0x000056146e32c36e in Assemble (st=0x56146fc78660,
mddev=0x7ffc55342450 "/dev/md0", ident=0x7ffc55342f70,
devlist=0x56146fc6e2d0, c=0x7ffc55342e90)
at Assemble.c:1914
#7 0x000056146e312ac9 in main (argc=11, argv=0x7ffc55343238)
at mdadm.c:1510
The file descriptor was closed early in Grow_continue(). The noted commit
moved the close() call to close the fd above the fork which caused the
parent process to return with a closed fd.
This meant reshape_array() and Grow_continue() would return in the parent
with the fd forked. The fd would eventually be passed to reopen_mddev()
which returned an unhandled NULL from fd2devnm() which would then be
dereferenced in devnm2devid.
Fix this by moving the close() call below the fork. This appears to
fix the 07revert-grow test. While we're at it, switch to using
close_fd() to invalidate the file descriptor.
Fixes: 77b72fa82813 ("mdadm/Grow: prevent md's fd from being occupied during delayed time")
Cc: Alex Wu <alexwu@synology.com>
Cc: BingJing Chang <bingjingc@synology.com>
Cc: Danny Shih <dannyshih@synology.com>
Cc: ChangSyun Peng <allenpeng@synology.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
Grow.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/Grow.c b/Grow.c
index 8c520d42..97f22c75 100644
--- a/Grow.c
+++ b/Grow.c
@@ -3514,7 +3514,6 @@ started:
return 0;
}
- close(fd);
/* Now we just need to kick off the reshape and watch, while
* handling backups of the data...
* This is all done by a forked background process.
@@ -3535,6 +3534,9 @@ started:
break;
}
+ /* Close unused file descriptor in the forked process */
+ close_fd(&fd);
+
/* If another array on the same devices is busy, the
* reshape will wait for them. This would mean that
* the first section that we suspend will stay suspended
--
2.31.1

@ -0,0 +1,36 @@
From 9ae62977b51dab0f4bb46b1c8ea5ebd1705b2f4d Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:10 -0600
Subject: [PATCH 35/52] monitor: Avoid segfault when calling NULL
get_bad_blocks
Not all struct superswitch implement a get_bad_blocks() function,
yet mdmon seems to call it without checking for NULL and thus
occasionally segfaults in the test 10ddf-geometry.
Fix this by checking for NULL before calling it.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
monitor.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/monitor.c b/monitor.c
index b877e595..820a93d0 100644
--- a/monitor.c
+++ b/monitor.c
@@ -311,6 +311,9 @@ static int check_for_cleared_bb(struct active_array *a, struct mdinfo *mdi)
struct md_bb *bb;
int i;
+ if (!ss->get_bad_blocks)
+ return -1;
+
/*
* Get a list of bad blocks for an array, then read list of
* acknowledged bad blocks from kernel and compare it against metadata
--
2.31.1

@ -0,0 +1,78 @@
From 6c9d9260633f2c8491985b0782cf0fbd7e51651b Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:11 -0600
Subject: [PATCH 36/52] mdadm: Fix mdadm -r remove option regression
The commit noted below globally adds a parameter to the -r option but missed
the fact that -r is used for another purpose: --remove.
After that commit, a command such as:
mdadm /dev/md0 -r /dev/loop0
will do nothing seeing the device parameter will be consumed as a
argument to the -r option; thus, there will only be one device
seen one the command line, devs_found will only be 1 and nothing will
happen.
This caused the 01r5integ and 01raid6integ tests to hang indefinitely
as mdadm did not remove the failed device. With the device not removed,
it would not be readded. Then the loop waiting for the array status to
change would loop forever.
This commit was recently reverted, but the legitimate fix for the
monitor operations was still not fixed. So add specific monitor
short ops to re-fix the --monitor -r option.
Fixes: 546047688e1c ("mdadm: fix coredump of mdadm --monitor -r")
Fixes: 190dc029b141 ("Revert "mdadm: fix coredump of mdadm --monitor -r"")
Cc: Wu Guanghao <wuguanghao3@huawei.com>
Cc: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
ReadMe.c | 1 +
mdadm.c | 1 +
mdadm.h | 1 +
3 files changed, 3 insertions(+)
diff --git a/ReadMe.c b/ReadMe.c
index bec1be9a..7518a32a 100644
--- a/ReadMe.c
+++ b/ReadMe.c
@@ -82,6 +82,7 @@ char Version[] = "mdadm - v" VERSION " - " VERS_DATE EXTRAVERSION "\n";
*/
char short_options[]="-ABCDEFGIQhVXYWZ:vqbc:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:k:";
+char short_monitor_options[]="-ABCDEFGIQhVXYWZ:vqbc:i:l:p:m:r:n:x:u:c:d:z:U:N:safRSow1tye:k:";
char short_bitmap_options[]=
"-ABCDEFGIQhVXYWZ:vqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:k:";
char short_bitmap_auto_options[]=
diff --git a/mdadm.c b/mdadm.c
index be40686c..d0c5e6de 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -227,6 +227,7 @@ int main(int argc, char *argv[])
shortopt = short_bitmap_auto_options;
break;
case 'F': newmode = MONITOR;
+ shortopt = short_monitor_options;
break;
case 'G': newmode = GROW;
shortopt = short_bitmap_options;
diff --git a/mdadm.h b/mdadm.h
index 974415b9..163f4a49 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -419,6 +419,7 @@ enum mode {
};
extern char short_options[];
+extern char short_monitor_options[];
extern char short_bitmap_options[];
extern char short_bitmap_auto_options[];
extern struct option long_options[];
--
2.31.1

@ -0,0 +1,42 @@
From 41edf6f45895193f4a523cb0a08d639c9ff9ccc9 Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:12 -0600
Subject: [PATCH 37/52] mdadm: Fix optional --write-behind parameter
The commit noted below changed the behaviour of --write-behind to
require an argument. This broke the 06wrmostly test with the error:
mdadm: Invalid value for maximum outstanding write-behind writes: (null).
Must be between 0 and 16383.
To fix this, check if optarg is NULL before parising it, as the origial
code did.
Fixes: 60815698c0ac ("Refactor parse_num and use it to parse optarg.")
Cc: Mateusz Grzonka <mateusz.grzonka@intel.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
mdadm.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/mdadm.c b/mdadm.c
index d0c5e6de..56722ed9 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -1201,8 +1201,9 @@ int main(int argc, char *argv[])
case O(BUILD, WriteBehind):
case O(CREATE, WriteBehind):
s.write_behind = DEFAULT_MAX_WRITE_BEHIND;
- if (parse_num(&s.write_behind, optarg) != 0 ||
- s.write_behind < 0 || s.write_behind > 16383) {
+ if (optarg &&
+ (parse_num(&s.write_behind, optarg) != 0 ||
+ s.write_behind < 0 || s.write_behind > 16383)) {
pr_err("Invalid value for maximum outstanding write-behind writes: %s.\n\tMust be between 0 and 16383.\n",
optarg);
exit(2);
--
2.31.1

@ -0,0 +1,38 @@
From 7539254342bc591717b0051734cc6c09c1b88640 Mon Sep 17 00:00:00 2001
From: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Date: Wed, 22 Jun 2022 14:25:13 -0600
Subject: [PATCH 38/52] tests/00raid0: add a test that validates raid0 with
layout fails for 0.9
329dfc28debb disallows the creation of raid0 with layouts for 0.9
metadata. This test confirms the new behavior.
Signed-off-by: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
tests/00raid0 | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/tests/00raid0 b/tests/00raid0
index 8bc18985..e6b21cc4 100644
--- a/tests/00raid0
+++ b/tests/00raid0
@@ -6,11 +6,9 @@ check raid0
testdev $md0 3 $mdsize2_l 512
mdadm -S $md0
-# now with version-0.90 superblock
+# verify raid0 with layouts fail for 0.90
mdadm -CR $md0 -e0.90 -l0 -n4 $dev0 $dev1 $dev2 $dev3
-check raid0
-testdev $md0 4 $mdsize0 512
-mdadm -S $md0
+check opposite_result
# now with no superblock
mdadm -B $md0 -l0 -n5 $dev0 $dev1 $dev2 $dev3 $dev4
--
2.31.1

@ -0,0 +1,99 @@
From 14c2161edb77d7294199e8aa7daa9f9d1d0ad5d7 Mon Sep 17 00:00:00 2001
From: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Date: Wed, 22 Jun 2022 14:25:14 -0600
Subject: [PATCH 39/52] tests: fix raid0 tests for 0.90 metadata
Some of the test cases fail because raid0 creation fails with the error,
"0.90 metadata does not support layouts for RAID0" added by commit,
329dfc28debb. Fix some of the test cases by switching from raid0 to
linear level for 0.9 metadata where possible.
Signed-off-by: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
tests/00raid0 | 4 ++--
tests/00readonly | 4 ++++
tests/03r0assem | 6 +++---
tests/04r0update | 4 ++--
tests/04update-metadata | 2 +-
5 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/tests/00raid0 b/tests/00raid0
index e6b21cc4..9b8896cb 100644
--- a/tests/00raid0
+++ b/tests/00raid0
@@ -20,8 +20,8 @@ mdadm -S $md0
# now same again with different chunk size
for chunk in 4 32 256
do
- mdadm -CR $md0 -e0.90 -l raid0 --chunk $chunk -n3 $dev0 $dev1 $dev2
- check raid0
+ mdadm -CR $md0 -e0.90 -l linear --chunk $chunk -n3 $dev0 $dev1 $dev2
+ check linear
testdev $md0 3 $mdsize0 $chunk
mdadm -S $md0
diff --git a/tests/00readonly b/tests/00readonly
index 28b0fa13..39202487 100644
--- a/tests/00readonly
+++ b/tests/00readonly
@@ -4,6 +4,10 @@ for metadata in 0.9 1.0 1.1 1.2
do
for level in linear raid0 raid1 raid4 raid5 raid6 raid10
do
+ if [[ $metadata == "0.9" && $level == "raid0" ]];
+ then
+ continue
+ fi
mdadm -CR $md0 -l $level -n 4 --metadata=$metadata \
$dev1 $dev2 $dev3 $dev4 --assume-clean
check nosync
diff --git a/tests/03r0assem b/tests/03r0assem
index 6744e322..44df0645 100644
--- a/tests/03r0assem
+++ b/tests/03r0assem
@@ -68,9 +68,9 @@ mdadm -S $md2
### Now for version 0...
mdadm --zero-superblock $dev0 $dev1 $dev2
-mdadm -CR $md2 -l0 --metadata=0.90 -n3 $dev0 $dev1 $dev2
-check raid0
-tst="testdev $md2 3 $mdsize0 512"
+mdadm -CR $md2 -llinear --metadata=0.90 -n3 $dev0 $dev1 $dev2
+check linear
+tst="testdev $md2 3 $mdsize0 1"
$tst
uuid=`mdadm -Db $md2 | sed 's/.*UUID=//'`
diff --git a/tests/04r0update b/tests/04r0update
index 73ee3b9f..b95efb06 100644
--- a/tests/04r0update
+++ b/tests/04r0update
@@ -1,7 +1,7 @@
# create a raid0, re-assemble with a different super-minor
-mdadm -CR -e 0.90 $md0 -l0 -n3 $dev0 $dev1 $dev2
-testdev $md0 3 $mdsize0 512
+mdadm -CR -e 0.90 $md0 -llinear -n3 $dev0 $dev1 $dev2
+testdev $md0 3 $mdsize0 1
minor1=`mdadm -E $dev0 | sed -n -e 's/.*Preferred Minor : //p'`
mdadm -S /dev/md0
diff --git a/tests/04update-metadata b/tests/04update-metadata
index 232fc1ff..08c14af7 100644
--- a/tests/04update-metadata
+++ b/tests/04update-metadata
@@ -8,7 +8,7 @@ set -xe
dlist="$dev0 $dev1 $dev2 $dev3"
-for ls in raid0/4 linear/4 raid1/1 raid5/3 raid6/2
+for ls in linear/4 raid1/1 raid5/3 raid6/2
do
s=${ls#*/} l=${ls%/*}
mdadm -CR --assume-clean -e 0.90 $md0 --level $l -n 4 -c 64 $dlist
--
2.31.1

@ -0,0 +1,39 @@
From de045db607b1ac4b70fc2a8878463e029c2ab1dc Mon Sep 17 00:00:00 2001
From: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Date: Wed, 22 Jun 2022 14:25:15 -0600
Subject: [PATCH 40/52] tests/04update-metadata: avoid passing chunk size to
raid1
'04update-metadata' test fails with error, "specifying chunk size is
forbidden for this level" added by commit, 5b30a34aa4b5e. Hence,
correcting the test to ignore passing chunk size to raid1.
Signed-off-by: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@oracle.com>
[logang@deltatee.com: fix if/then style and dropped unrelated hunk]
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
tests/04update-metadata | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/tests/04update-metadata b/tests/04update-metadata
index 08c14af7..2b72a303 100644
--- a/tests/04update-metadata
+++ b/tests/04update-metadata
@@ -11,7 +11,11 @@ dlist="$dev0 $dev1 $dev2 $dev3"
for ls in linear/4 raid1/1 raid5/3 raid6/2
do
s=${ls#*/} l=${ls%/*}
- mdadm -CR --assume-clean -e 0.90 $md0 --level $l -n 4 -c 64 $dlist
+ if [[ $l == 'raid1' ]]; then
+ mdadm -CR --assume-clean -e 0.90 $md0 --level $l -n 4 $dlist
+ else
+ mdadm -CR --assume-clean -e 0.90 $md0 --level $l -n 4 -c 64 $dlist
+ fi
testdev $md0 $s 19904 64
mdadm -S $md0
mdadm -A $md0 --update=metadata $dlist
--
2.31.1

@ -0,0 +1,31 @@
From a2c832465fc75202e244327b2081231dfa974617 Mon Sep 17 00:00:00 2001
From: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Date: Wed, 22 Jun 2022 14:25:16 -0600
Subject: [PATCH 41/52] tests/02lineargrow: clear the superblock at every
iteration
This fixes 02lineargrow test as prior metadata causes --add operation
to misbehave.
Signed-off-by: Sudhakar Panneerselvam <sudhakar.panneerselvam@oracle.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
tests/02lineargrow | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tests/02lineargrow b/tests/02lineargrow
index e05c219d..595bf9f2 100644
--- a/tests/02lineargrow
+++ b/tests/02lineargrow
@@ -20,4 +20,6 @@ do
testdev $md0 3 $sz 1
mdadm -S $md0
+ mdadm --zero /dev/loop2
+ mdadm --zero /dev/loop3
done
--
2.31.1

@ -0,0 +1,88 @@
From a7bfcc716e235664dfb3b6c5a9590273e611ac72 Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:17 -0600
Subject: [PATCH 42/52] mdadm/test: Add a mode to repeat specified tests
Many tests fail infrequently or rarely. To help find these, add
an option to run the tests multiple times by specifying --loop=N.
If --loop=0 is specified, the test will be looped forever.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
test | 36 ++++++++++++++++++++++++------------
1 file changed, 24 insertions(+), 12 deletions(-)
diff --git a/test b/test
index 711a3c7a..da6db5e0 100755
--- a/test
+++ b/test
@@ -10,6 +10,7 @@ devlist=
savelogs=0
exitonerror=1
+loop=1
prefix='[0-9][0-9]'
# use loop devices by default if doesn't specify --dev
@@ -117,6 +118,7 @@ do_help() {
--logdir=directory Directory to save all logfiles in
--save-logs Usually use with --logdir together
--keep-going | --no-error Don't stop on error, ie. run all tests
+ --loop=N Run tests N times (0 to run forever)
--dev=loop|lvm|ram|disk Use loop devices (default), LVM, RAM or disk
--disks= Provide a bunch of physical devices for test
--volgroup=name LVM volume group for LVM test
@@ -211,6 +213,9 @@ parse_args() {
--keep-going | --no-error )
exitonerror=0
;;
+ --loop=* )
+ loop="${i##*=}"
+ ;;
--disable-multipath )
unset MULTIPATH
;;
@@ -263,19 +268,26 @@ main() {
echo "Testing on linux-$(uname -r) kernel"
[ "$savelogs" == "1" ] &&
echo "Saving logs to $logdir"
- if [ "x$TESTLIST" != "x" ]
- then
- for script in ${TESTLIST[@]}
- do
- do_test $testdir/$script
- done
- else
- for script in $testdir/$prefix $testdir/$prefix*[^~]
- do
- do_test $script
- done
- fi
+ while true; do
+ if [ "x$TESTLIST" != "x" ]
+ then
+ for script in ${TESTLIST[@]}
+ do
+ do_test $testdir/$script
+ done
+ else
+ for script in $testdir/$prefix $testdir/$prefix*[^~]
+ do
+ do_test $script
+ done
+ fi
+
+ let loop=$loop-1
+ if [ "$loop" == "0" ]; then
+ break
+ fi
+ done
exit 0
}
--
2.31.1

@ -0,0 +1,120 @@
From 28520bf114b3b0515a48ff44fff4ecbe9ed6dfad Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:18 -0600
Subject: [PATCH 43/52] mdadm/test: Mark and ignore broken test failures
Add functionality to continue if a test marked as broken fails.
To mark a test as broken, a file with the same name but with the suffix
'.broken' should exist. The first line in the file will be printed with
a KNOWN BROKEN message; the rest of the file can describe the how the
test is broken.
Also adds --skip-broken and --skip-always-broken to skip all the tests
that have a .broken file or to skip all tests whose .broken file's first
line contains the keyword always.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
test | 37 +++++++++++++++++++++++++++++++++++--
1 file changed, 35 insertions(+), 2 deletions(-)
diff --git a/test b/test
index da6db5e0..61d9ee83 100755
--- a/test
+++ b/test
@@ -10,6 +10,8 @@ devlist=
savelogs=0
exitonerror=1
+ctrl_c_error=0
+skipbroken=0
loop=1
prefix='[0-9][0-9]'
@@ -36,6 +38,7 @@ die() {
ctrl_c() {
exitonerror=1
+ ctrl_c_error=1
}
# mdadm always adds --quiet, and we want to see any unexpected messages
@@ -80,8 +83,21 @@ mdadm() {
do_test() {
_script=$1
_basename=`basename $_script`
+ _broken=0
+
if [ -f "$_script" ]
then
+ if [ -f "${_script}.broken" ]; then
+ _broken=1
+ _broken_msg=$(head -n1 "${_script}.broken" | tr -d '\n')
+ if [ "$skipbroken" == "all" ]; then
+ return
+ elif [ "$skipbroken" == "always" ] &&
+ [[ "$_broken_msg" == *always* ]]; then
+ return
+ fi
+ fi
+
rm -f $targetdir/stderr
# this might have been reset: restore the default.
echo 2000 > /proc/sys/dev/raid/speed_limit_max
@@ -98,10 +114,15 @@ do_test() {
else
save_log fail
_fail=1
+ if [ "$_broken" == "1" ]; then
+ echo " (KNOWN BROKEN TEST: $_broken_msg)"
+ fi
fi
[ "$savelogs" == "1" ] &&
mv -f $targetdir/log $logdir/$_basename.log
- [ "$_fail" == "1" -a "$exitonerror" == "1" ] && exit 1
+ [ "$ctrl_c_error" == "1" ] && exit 1
+ [ "$_fail" == "1" -a "$exitonerror" == "1" \
+ -a "$_broken" == "0" ] && exit 1
fi
}
@@ -119,6 +140,8 @@ do_help() {
--save-logs Usually use with --logdir together
--keep-going | --no-error Don't stop on error, ie. run all tests
--loop=N Run tests N times (0 to run forever)
+ --skip-broken Skip tests that are known to be broken
+ --skip-always-broken Skip tests that are known to always fail
--dev=loop|lvm|ram|disk Use loop devices (default), LVM, RAM or disk
--disks= Provide a bunch of physical devices for test
--volgroup=name LVM volume group for LVM test
@@ -216,6 +239,12 @@ parse_args() {
--loop=* )
loop="${i##*=}"
;;
+ --skip-broken )
+ skipbroken=all
+ ;;
+ --skip-always-broken )
+ skipbroken=always
+ ;;
--disable-multipath )
unset MULTIPATH
;;
@@ -279,7 +308,11 @@ main() {
else
for script in $testdir/$prefix $testdir/$prefix*[^~]
do
- do_test $script
+ case $script in
+ *.broken) ;;
+ *)
+ do_test $script
+ esac
done
fi
--
2.31.1

@ -0,0 +1,447 @@
From daa86d6634761796ada1f535c13e47fdd3cc95eb Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 22 Jun 2022 14:25:19 -0600
Subject: [PATCH 44/52] tests: Add broken files for all broken tests
Each broken file contains the rough frequency of brokeness as well
as a brief explanation of what happens when it breaks. Estimates
of failure rates are not statistically significant and can vary
run to run.
This is really just a view from my window. Tests were done on a
small VM with the default loop devices, not real hardware. We've
seen different kernel configurations can cause bugs to appear as well
(ie. different block schedulers). It may also be that different race
conditions will be seen on machines with different performance
characteristics.
These annotations were done with the kernel currently in md/md-next:
facef3b96c5b ("md: Notify sysfs sync_completed in md_reap_sync_thread()")
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
---
tests/01r5integ.broken | 7 ++++
tests/01raid6integ.broken | 7 ++++
tests/04r5swap.broken | 7 ++++
tests/07autoassemble.broken | 8 ++++
tests/07autodetect.broken | 5 +++
tests/07changelevelintr.broken | 9 +++++
tests/07changelevels.broken | 9 +++++
tests/07reshape5intr.broken | 45 ++++++++++++++++++++++
tests/07revert-grow.broken | 31 +++++++++++++++
tests/07revert-shrink.broken | 9 +++++
tests/07testreshape5.broken | 12 ++++++
tests/09imsm-assemble.broken | 6 +++
tests/09imsm-create-fail-rebuild.broken | 5 +++
tests/09imsm-overlap.broken | 7 ++++
tests/10ddf-assemble-missing.broken | 6 +++
tests/10ddf-fail-create-race.broken | 7 ++++
tests/10ddf-fail-two-spares.broken | 5 +++
tests/10ddf-incremental-wrong-order.broken | 9 +++++
tests/14imsm-r1_2d-grow-r1_3d.broken | 5 +++
tests/14imsm-r1_2d-takeover-r0_2d.broken | 6 +++
tests/18imsm-r10_4d-takeover-r0_2d.broken | 5 +++
tests/18imsm-r1_2d-takeover-r0_1d.broken | 6 +++
tests/19raid6auto-repair.broken | 5 +++
tests/19raid6repair.broken | 5 +++
24 files changed, 226 insertions(+)
create mode 100644 tests/01r5integ.broken
create mode 100644 tests/01raid6integ.broken
create mode 100644 tests/04r5swap.broken
create mode 100644 tests/07autoassemble.broken
create mode 100644 tests/07autodetect.broken
create mode 100644 tests/07changelevelintr.broken
create mode 100644 tests/07changelevels.broken
create mode 100644 tests/07reshape5intr.broken
create mode 100644 tests/07revert-grow.broken
create mode 100644 tests/07revert-shrink.broken
create mode 100644 tests/07testreshape5.broken
create mode 100644 tests/09imsm-assemble.broken
create mode 100644 tests/09imsm-create-fail-rebuild.broken
create mode 100644 tests/09imsm-overlap.broken
create mode 100644 tests/10ddf-assemble-missing.broken
create mode 100644 tests/10ddf-fail-create-race.broken
create mode 100644 tests/10ddf-fail-two-spares.broken
create mode 100644 tests/10ddf-incremental-wrong-order.broken
create mode 100644 tests/14imsm-r1_2d-grow-r1_3d.broken
create mode 100644 tests/14imsm-r1_2d-takeover-r0_2d.broken
create mode 100644 tests/18imsm-r10_4d-takeover-r0_2d.broken
create mode 100644 tests/18imsm-r1_2d-takeover-r0_1d.broken
create mode 100644 tests/19raid6auto-repair.broken
create mode 100644 tests/19raid6repair.broken
diff --git a/tests/01r5integ.broken b/tests/01r5integ.broken
new file mode 100644
index 00000000..20737637
--- /dev/null
+++ b/tests/01r5integ.broken
@@ -0,0 +1,7 @@
+fails rarely
+
+Fails about 1 in every 30 runs with a sha mismatch error:
+
+ c49ab26e1b01def7874af9b8a6d6d0c29fdfafe6 /dev/md0 does not match
+ 15dc2f73262f811ada53c65e505ceec9cf025cb9 /dev/md0 with /dev/loop3
+ missing
diff --git a/tests/01raid6integ.broken b/tests/01raid6integ.broken
new file mode 100644
index 00000000..1df735f0
--- /dev/null
+++ b/tests/01raid6integ.broken
@@ -0,0 +1,7 @@
+fails infrequently
+
+Fails about 1 in 5 with a sha mismatch:
+
+ 8286c2bc045ae2cfe9f8b7ae3a898fa25db6926f /dev/md0 does not match
+ a083a0738b58caab37fd568b91b177035ded37df /dev/md0 with /dev/loop2 and
+ /dev/loop3 missing
diff --git a/tests/04r5swap.broken b/tests/04r5swap.broken
new file mode 100644
index 00000000..e38987db
--- /dev/null
+++ b/tests/04r5swap.broken
@@ -0,0 +1,7 @@
+always fails
+
+Fails with errors:
+
+ mdadm: /dev/loop0 has no superblock - assembly aborted
+
+ ERROR: no recovery happening
diff --git a/tests/07autoassemble.broken b/tests/07autoassemble.broken
new file mode 100644
index 00000000..8be09407
--- /dev/null
+++ b/tests/07autoassemble.broken
@@ -0,0 +1,8 @@
+always fails
+
+Prints lots of messages, but the array doesn't assemble. Error
+possibly related to:
+
+ mdadm: /dev/md/1 is busy - skipping
+ mdadm: no recogniseable superblock on /dev/md/testing:0
+ mdadm: /dev/md/2 is busy - skipping
diff --git a/tests/07autodetect.broken b/tests/07autodetect.broken
new file mode 100644
index 00000000..294954a1
--- /dev/null
+++ b/tests/07autodetect.broken
@@ -0,0 +1,5 @@
+always fails
+
+Fails with error:
+
+ ERROR: no resync happening
diff --git a/tests/07changelevelintr.broken b/tests/07changelevelintr.broken
new file mode 100644
index 00000000..284b4906
--- /dev/null
+++ b/tests/07changelevelintr.broken
@@ -0,0 +1,9 @@
+always fails
+
+Fails with errors:
+
+ mdadm: this change will reduce the size of the array.
+ use --grow --array-size first to truncate array.
+ e.g. mdadm --grow /dev/md0 --array-size 56832
+
+ ERROR: no reshape happening
diff --git a/tests/07changelevels.broken b/tests/07changelevels.broken
new file mode 100644
index 00000000..9b930d93
--- /dev/null
+++ b/tests/07changelevels.broken
@@ -0,0 +1,9 @@
+always fails
+
+Fails with errors:
+
+ mdadm: /dev/loop0 is smaller than given size. 18976K < 19968K + metadata
+ mdadm: /dev/loop1 is smaller than given size. 18976K < 19968K + metadata
+ mdadm: /dev/loop2 is smaller than given size. 18976K < 19968K + metadata
+
+ ERROR: /dev/md0 isn't a block device.
diff --git a/tests/07reshape5intr.broken b/tests/07reshape5intr.broken
new file mode 100644
index 00000000..efe52a66
--- /dev/null
+++ b/tests/07reshape5intr.broken
@@ -0,0 +1,45 @@
+always fails
+
+This patch, recently added to md-next causes the test to always fail:
+
+7e6ba434cc60 ("md: don't unregister sync_thread with reconfig_mutex
+held")
+
+The new error is simply:
+
+ ERROR: no reshape happening
+
+Before the patch, the error seen is below.
+
+--
+
+fails infrequently
+
+Fails roughly 1 in 4 runs with errors:
+
+ mdadm: Merging with already-assembled /dev/md/0
+ mdadm: cannot re-read metadata from /dev/loop6 - aborting
+
+ ERROR: no reshape happening
+
+Also have seen a random deadlock:
+
+ INFO: task mdadm:109702 blocked for more than 30 seconds.
+ Not tainted 5.18.0-rc3-eid-vmlocalyes-dbg-00095-g3c2b5427979d #2040
+ "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
+ task:mdadm state:D stack: 0 pid:109702 ppid: 1 flags:0x00004000
+ Call Trace:
+ <TASK>
+ __schedule+0x67e/0x13b0
+ schedule+0x82/0x110
+ mddev_suspend+0x2e1/0x330
+ suspend_lo_store+0xbd/0x140
+ md_attr_store+0xcb/0x130
+ sysfs_kf_write+0x89/0xb0
+ kernfs_fop_write_iter+0x202/0x2c0
+ new_sync_write+0x222/0x330
+ vfs_write+0x3bc/0x4d0
+ ksys_write+0xd9/0x180
+ __x64_sys_write+0x43/0x50
+ do_syscall_64+0x3b/0x90
+ entry_SYSCALL_64_after_hwframe+0x44/0xae
diff --git a/tests/07revert-grow.broken b/tests/07revert-grow.broken
new file mode 100644
index 00000000..9b6db86f
--- /dev/null
+++ b/tests/07revert-grow.broken
@@ -0,0 +1,31 @@
+always fails
+
+This patch, recently added to md-next causes the test to always fail:
+
+7e6ba434cc60 ("md: don't unregister sync_thread with reconfig_mutex held")
+
+The errors are:
+
+ mdadm: No active reshape to revert on /dev/loop0
+ ERROR: active raid5 not found
+
+Before the patch, the error seen is below.
+
+--
+
+fails rarely
+
+Fails about 1 in every 30 runs with errors:
+
+ mdadm: Merging with already-assembled /dev/md/0
+ mdadm: backup file /tmp/md-backup inaccessible: No such file or directory
+ mdadm: failed to add /dev/loop1 to /dev/md/0: Invalid argument
+ mdadm: failed to add /dev/loop2 to /dev/md/0: Invalid argument
+ mdadm: failed to add /dev/loop3 to /dev/md/0: Invalid argument
+ mdadm: failed to add /dev/loop0 to /dev/md/0: Invalid argument
+ mdadm: /dev/md/0 assembled from 1 drive - need all 5 to start it
+ (use --run to insist).
+
+ grep: /sys/block/md*/md/sync_action: No such file or directory
+
+ ERROR: active raid5 not found
diff --git a/tests/07revert-shrink.broken b/tests/07revert-shrink.broken
new file mode 100644
index 00000000..c33c39ec
--- /dev/null
+++ b/tests/07revert-shrink.broken
@@ -0,0 +1,9 @@
+always fails
+
+Fails with errors:
+
+ mdadm: this change will reduce the size of the array.
+ use --grow --array-size first to truncate array.
+ e.g. mdadm --grow /dev/md0 --array-size 53760
+
+ ERROR: active raid5 not found
diff --git a/tests/07testreshape5.broken b/tests/07testreshape5.broken
new file mode 100644
index 00000000..a8ce03e4
--- /dev/null
+++ b/tests/07testreshape5.broken
@@ -0,0 +1,12 @@
+always fails
+
+Test seems to run 'test_stripe' at $dir directory, but $dir is never
+set. If $dir is adjusted to $PWD, the test still fails with:
+
+ mdadm: /dev/loop2 is not suitable for this array.
+ mdadm: create aborted
+ ++ return 1
+ ++ cmp -s -n 8192 /dev/md0 /tmp/RandFile
+ ++ echo cmp failed
+ cmp failed
+ ++ exit 2
diff --git a/tests/09imsm-assemble.broken b/tests/09imsm-assemble.broken
new file mode 100644
index 00000000..a6d4d5cf
--- /dev/null
+++ b/tests/09imsm-assemble.broken
@@ -0,0 +1,6 @@
+fails infrequently
+
+Fails roughly 1 in 10 runs with errors:
+
+ mdadm: /dev/loop2 is still in use, cannot remove.
+ /dev/loop2 removal from /dev/md/container should have succeeded
diff --git a/tests/09imsm-create-fail-rebuild.broken b/tests/09imsm-create-fail-rebuild.broken
new file mode 100644
index 00000000..40c4b294
--- /dev/null
+++ b/tests/09imsm-create-fail-rebuild.broken
@@ -0,0 +1,5 @@
+always fails
+
+Fails with error:
+
+ **Error**: Array size mismatch - expected 3072, actual 16384
diff --git a/tests/09imsm-overlap.broken b/tests/09imsm-overlap.broken
new file mode 100644
index 00000000..e7ccab76
--- /dev/null
+++ b/tests/09imsm-overlap.broken
@@ -0,0 +1,7 @@
+always fails
+
+Fails with errors:
+
+ **Error**: Offset mismatch - expected 15360, actual 0
+ **Error**: Offset mismatch - expected 15360, actual 0
+ /dev/md/vol3 failed check
diff --git a/tests/10ddf-assemble-missing.broken b/tests/10ddf-assemble-missing.broken
new file mode 100644
index 00000000..bfd8d103
--- /dev/null
+++ b/tests/10ddf-assemble-missing.broken
@@ -0,0 +1,6 @@
+always fails
+
+Fails with errors:
+
+ ERROR: /dev/md/vol0 has unexpected state on /dev/loop10
+ ERROR: unexpected number of online disks on /dev/loop10
diff --git a/tests/10ddf-fail-create-race.broken b/tests/10ddf-fail-create-race.broken
new file mode 100644
index 00000000..6c0df023
--- /dev/null
+++ b/tests/10ddf-fail-create-race.broken
@@ -0,0 +1,7 @@
+usually fails
+
+Fails about 9 out of 10 times with many errors:
+
+ mdadm: cannot open MISSING: No such file or directory
+ ERROR: non-degraded array found
+ ERROR: disk 0 not marked as failed in meta data
diff --git a/tests/10ddf-fail-two-spares.broken b/tests/10ddf-fail-two-spares.broken
new file mode 100644
index 00000000..eeea56d9
--- /dev/null
+++ b/tests/10ddf-fail-two-spares.broken
@@ -0,0 +1,5 @@
+fails infrequently
+
+Fails roughly 1 in 3 with error:
+
+ ERROR: /dev/md/vol1 should be optimal in meta data
diff --git a/tests/10ddf-incremental-wrong-order.broken b/tests/10ddf-incremental-wrong-order.broken
new file mode 100644
index 00000000..a5af3bab
--- /dev/null
+++ b/tests/10ddf-incremental-wrong-order.broken
@@ -0,0 +1,9 @@
+always fails
+
+Fails with errors:
+ ERROR: sha1sum of /dev/md/vol0 has changed
+ ERROR: /dev/md/vol0 has unexpected state on /dev/loop10
+ ERROR: unexpected number of online disks on /dev/loop10
+ ERROR: /dev/md/vol0 has unexpected state on /dev/loop8
+ ERROR: unexpected number of online disks on /dev/loop8
+ ERROR: sha1sum of /dev/md/vol0 has changed
diff --git a/tests/14imsm-r1_2d-grow-r1_3d.broken b/tests/14imsm-r1_2d-grow-r1_3d.broken
new file mode 100644
index 00000000..4ef1d406
--- /dev/null
+++ b/tests/14imsm-r1_2d-grow-r1_3d.broken
@@ -0,0 +1,5 @@
+always fails
+
+Fails with error:
+
+ mdadm/tests/func.sh: line 325: dvsize/chunk: division by 0 (error token is "chunk")
diff --git a/tests/14imsm-r1_2d-takeover-r0_2d.broken b/tests/14imsm-r1_2d-takeover-r0_2d.broken
new file mode 100644
index 00000000..89cd4e57
--- /dev/null
+++ b/tests/14imsm-r1_2d-takeover-r0_2d.broken
@@ -0,0 +1,6 @@
+always fails
+
+Fails with error:
+
+ tests/func.sh: line 325: dvsize/chunk: division by 0 (error token
+ is "chunk")
diff --git a/tests/18imsm-r10_4d-takeover-r0_2d.broken b/tests/18imsm-r10_4d-takeover-r0_2d.broken
new file mode 100644
index 00000000..a27399f5
--- /dev/null
+++ b/tests/18imsm-r10_4d-takeover-r0_2d.broken
@@ -0,0 +1,5 @@
+fails rarely
+
+Fails about 1 run in 100 with message:
+
+ ERROR: size is wrong for /dev/md/vol0: 2 * 5120 (chunk=128) = 20480, not 0
diff --git a/tests/18imsm-r1_2d-takeover-r0_1d.broken b/tests/18imsm-r1_2d-takeover-r0_1d.broken
new file mode 100644
index 00000000..aa1982e6
--- /dev/null
+++ b/tests/18imsm-r1_2d-takeover-r0_1d.broken
@@ -0,0 +1,6 @@
+always fails
+
+Fails with error:
+
+ tests/func.sh: line 325: dvsize/chunk: division by 0 (error token
+ is "chunk")
diff --git a/tests/19raid6auto-repair.broken b/tests/19raid6auto-repair.broken
new file mode 100644
index 00000000..e91a1425
--- /dev/null
+++ b/tests/19raid6auto-repair.broken
@@ -0,0 +1,5 @@
+always fails
+
+Fails with:
+
+ "should detect errors"
diff --git a/tests/19raid6repair.broken b/tests/19raid6repair.broken
new file mode 100644
index 00000000..e91a1425
--- /dev/null
+++ b/tests/19raid6repair.broken
@@ -0,0 +1,5 @@
+always fails
+
+Fails with:
+
+ "should detect errors"
--
2.31.1

@ -0,0 +1,316 @@
From 239b3cc0b5da87e966746533b1873c439db54b16 Mon Sep 17 00:00:00 2001
From: Mateusz Grzonka <mateusz.grzonka@intel.com>
Date: Fri, 12 Aug 2022 16:36:02 +0200
Subject: [PATCH 45/52] mdadm: Replace obsolete usleep with nanosleep
According to POSIX.1-2001, usleep is considered obsolete.
Replace it with a wrapper that uses nanosleep, as recommended in man.
Add handy macros for conversions between msec, usec and nsec.
Signed-off-by: Mateusz Grzonka <mateusz.grzonka@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Assemble.c | 2 +-
Grow.c | 4 ++--
Manage.c | 10 +++++-----
managemon.c | 8 ++++----
mdadm.h | 4 ++++
mdmon.c | 4 ++--
super-intel.c | 6 +++---
util.c | 42 +++++++++++++++++++++++++++++++++---------
8 files changed, 54 insertions(+), 26 deletions(-)
diff --git a/Assemble.c b/Assemble.c
index 6df6bfbc..be2160b4 100644
--- a/Assemble.c
+++ b/Assemble.c
@@ -1947,7 +1947,7 @@ out:
break;
close(mdfd);
}
- usleep(usecs);
+ sleep_for(0, USEC_TO_NSEC(usecs), true);
usecs <<= 1;
}
}
diff --git a/Grow.c b/Grow.c
index 97f22c75..5780635a 100644
--- a/Grow.c
+++ b/Grow.c
@@ -954,7 +954,7 @@ int start_reshape(struct mdinfo *sra, int already_running,
err = sysfs_set_str(sra, NULL, "sync_action",
"reshape");
if (err)
- sleep(1);
+ sleep_for(1, 0, true);
} while (err && errno == EBUSY && cnt-- > 0);
}
return err;
@@ -5058,7 +5058,7 @@ int Grow_continue_command(char *devname, int fd,
}
st->ss->getinfo_super(st, content, NULL);
if (!content->reshape_active)
- sleep(3);
+ sleep_for(3, 0, true);
else
break;
} while (cnt-- > 0);
diff --git a/Manage.c b/Manage.c
index e5e6abe4..a142f8bd 100644
--- a/Manage.c
+++ b/Manage.c
@@ -244,7 +244,7 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
"array_state",
"inactive")) < 0 &&
errno == EBUSY) {
- usleep(200000);
+ sleep_for(0, MSEC_TO_NSEC(200), true);
count--;
}
if (err) {
@@ -328,7 +328,7 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
sysfs_get_ll(mdi, NULL, "sync_max", &old_sync_max) == 0) {
/* must be in the critical section - wait a bit */
delay -= 1;
- usleep(100000);
+ sleep_for(0, MSEC_TO_NSEC(100), true);
}
if (sysfs_set_str(mdi, NULL, "sync_action", "frozen") != 0)
@@ -405,7 +405,7 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
* quite started yet. Wait a bit and
* check 'sync_action' to see.
*/
- usleep(10000);
+ sleep_for(0, MSEC_TO_NSEC(10), true);
sysfs_get_str(mdi, NULL, "sync_action", buf, sizeof(buf));
if (strncmp(buf, "reshape", 7) != 0)
break;
@@ -447,7 +447,7 @@ done:
count = 25; err = 0;
while (count && fd >= 0 &&
(err = ioctl(fd, STOP_ARRAY, NULL)) < 0 && errno == EBUSY) {
- usleep(200000);
+ sleep_for(0, MSEC_TO_NSEC(200), true);
count --;
}
if (fd >= 0 && err) {
@@ -1105,7 +1105,7 @@ int Manage_remove(struct supertype *tst, int fd, struct mddev_dev *dv,
ret = sysfs_unique_holder(devnm, rdev);
if (ret < 2)
break;
- usleep(100 * 1000); /* 100ms */
+ sleep_for(0, MSEC_TO_NSEC(100), true);
} while (--count > 0);
if (ret == 0) {
diff --git a/managemon.c b/managemon.c
index 0e9bdf00..a7bfa8f6 100644
--- a/managemon.c
+++ b/managemon.c
@@ -207,7 +207,7 @@ static void replace_array(struct supertype *container,
remove_old();
while (pending_discard) {
while (discard_this == NULL)
- sleep(1);
+ sleep_for(1, 0, true);
remove_old();
}
pending_discard = old;
@@ -568,7 +568,7 @@ static void manage_member(struct mdstat_ent *mdstat,
updates = NULL;
while (update_queue_pending || update_queue) {
check_update_queue(container);
- usleep(15*1000);
+ sleep_for(0, MSEC_TO_NSEC(15), true);
}
replace_array(container, a, newa);
if (sysfs_set_str(&a->info, NULL,
@@ -822,7 +822,7 @@ static void handle_message(struct supertype *container, struct metadata_update *
if (msg->len <= 0)
while (update_queue_pending || update_queue) {
check_update_queue(container);
- usleep(15*1000);
+ sleep_for(0, MSEC_TO_NSEC(15), true);
}
if (msg->len == 0) { /* ping_monitor */
@@ -836,7 +836,7 @@ static void handle_message(struct supertype *container, struct metadata_update *
wakeup_monitor();
while (monitor_loop_cnt - cnt < 0)
- usleep(10 * 1000);
+ sleep_for(0, MSEC_TO_NSEC(10), true);
} else if (msg->len == -1) { /* ping_manager */
struct mdstat_ent *mdstat = mdstat_read(1, 0);
diff --git a/mdadm.h b/mdadm.h
index 163f4a49..add9c0b6 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -1720,6 +1720,10 @@ extern int cluster_get_dlmlock(void);
extern int cluster_release_dlmlock(void);
extern void set_dlm_hooks(void);
+#define MSEC_TO_NSEC(msec) ((msec) * 1000000)
+#define USEC_TO_NSEC(usec) ((usec) * 1000)
+extern void sleep_for(unsigned int sec, long nsec, bool wake_after_interrupt);
+
#define _ROUND_UP(val, base) (((val) + (base) - 1) & ~(base - 1))
#define ROUND_UP(val, base) _ROUND_UP(val, (typeof(val))(base))
#define ROUND_UP_PTR(ptr, base) ((typeof(ptr)) \
diff --git a/mdmon.c b/mdmon.c
index c057da63..e9d035eb 100644
--- a/mdmon.c
+++ b/mdmon.c
@@ -99,7 +99,7 @@ static int clone_monitor(struct supertype *container)
if (rc)
return rc;
while (mon_tid == -1)
- usleep(10);
+ sleep_for(0, USEC_TO_NSEC(10), true);
pthread_attr_destroy(&attr);
mgr_tid = syscall(SYS_gettid);
@@ -209,7 +209,7 @@ static void try_kill_monitor(pid_t pid, char *devname, int sock)
rv = kill(pid, SIGUSR1);
if (rv < 0)
break;
- usleep(200000);
+ sleep_for(0, MSEC_TO_NSEC(200), true);
}
}
diff --git a/super-intel.c b/super-intel.c
index 4ddfcf94..4d82af3d 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -5275,7 +5275,7 @@ static int get_super_block(struct intel_super **super_list, char *devnm, char *d
/* retry the load if we might have raced against mdmon */
if (err == 3 && devnm && mdmon_running(devnm))
for (retry = 0; retry < 3; retry++) {
- usleep(3000);
+ sleep_for(0, MSEC_TO_NSEC(3), true);
err = load_and_parse_mpb(dfd, s, NULL, keep_fd);
if (err != 3)
break;
@@ -5377,7 +5377,7 @@ static int load_super_imsm(struct supertype *st, int fd, char *devname)
if (mdstat && mdmon_running(mdstat->devnm) && getpid() != mdmon_pid(mdstat->devnm)) {
for (retry = 0; retry < 3; retry++) {
- usleep(3000);
+ sleep_for(0, MSEC_TO_NSEC(3), true);
rv = load_and_parse_mpb(fd, super, devname, 0);
if (rv != 3)
break;
@@ -12084,7 +12084,7 @@ int wait_for_reshape_imsm(struct mdinfo *sra, int ndata)
close(fd);
return 1;
}
- usleep(30000);
+ sleep_for(0, MSEC_TO_NSEC(30), true);
} else
break;
} while (retry--);
diff --git a/util.c b/util.c
index 38f0420e..ca48d976 100644
--- a/util.c
+++ b/util.c
@@ -166,7 +166,7 @@ retry:
pr_err("error %d when get PW mode on lock %s\n", errno, str);
/* let's try several times if EAGAIN happened */
if (dlm_lock_res->lksb.sb_status == EAGAIN && retry_count < 10) {
- sleep(10);
+ sleep_for(10, 0, true);
retry_count++;
goto retry;
}
@@ -1085,7 +1085,7 @@ int open_dev_excl(char *devnm)
int i;
int flags = O_RDWR;
dev_t devid = devnm2devid(devnm);
- long delay = 1000;
+ unsigned int delay = 1; // miliseconds
sprintf(buf, "%d:%d", major(devid), minor(devid));
for (i = 0; i < 25; i++) {
@@ -1098,8 +1098,8 @@ int open_dev_excl(char *devnm)
}
if (errno != EBUSY)
return fd;
- usleep(delay);
- if (delay < 200000)
+ sleep_for(0, MSEC_TO_NSEC(delay), true);
+ if (delay < 200)
delay *= 2;
}
return -1;
@@ -1123,7 +1123,7 @@ void wait_for(char *dev, int fd)
{
int i;
struct stat stb_want;
- long delay = 1000;
+ unsigned int delay = 1; // miliseconds
if (fstat(fd, &stb_want) != 0 ||
(stb_want.st_mode & S_IFMT) != S_IFBLK)
@@ -1135,8 +1135,8 @@ void wait_for(char *dev, int fd)
(stb.st_mode & S_IFMT) == S_IFBLK &&
(stb.st_rdev == stb_want.st_rdev))
return;
- usleep(delay);
- if (delay < 200000)
+ sleep_for(0, MSEC_TO_NSEC(delay), true);
+ if (delay < 200)
delay *= 2;
}
if (i == 25)
@@ -1821,7 +1821,7 @@ int hot_remove_disk(int mdfd, unsigned long dev, int force)
while ((ret = ioctl(mdfd, HOT_REMOVE_DISK, dev)) == -1 &&
errno == EBUSY &&
cnt-- > 0)
- usleep(10000);
+ sleep_for(0, MSEC_TO_NSEC(10), true);
return ret;
}
@@ -1834,7 +1834,7 @@ int sys_hot_remove_disk(int statefd, int force)
while ((ret = write(statefd, "remove", 6)) == -1 &&
errno == EBUSY &&
cnt-- > 0)
- usleep(10000);
+ sleep_for(0, MSEC_TO_NSEC(10), true);
return ret == 6 ? 0 : -1;
}
@@ -2375,3 +2375,27 @@ out:
close(fd_zero);
return ret;
}
+
+/**
+ * sleep_for() - Sleeps for specified time.
+ * @sec: Seconds to sleep for.
+ * @nsec: Nanoseconds to sleep for, has to be less than one second.
+ * @wake_after_interrupt: If set, wake up if interrupted.
+ *
+ * Function immediately returns if error different than EINTR occurs.
+ */
+void sleep_for(unsigned int sec, long nsec, bool wake_after_interrupt)
+{
+ struct timespec delay = {.tv_sec = sec, .tv_nsec = nsec};
+
+ assert(nsec < MSEC_TO_NSEC(1000));
+
+ do {
+ errno = 0;
+ nanosleep(&delay, &delay);
+ if (errno != 0 && errno != EINTR) {
+ pr_err("Error sleeping for %us %ldns: %s\n", sec, nsec, strerror(errno));
+ return;
+ }
+ } while (!wake_after_interrupt && errno == EINTR);
+}
--
2.31.1

@ -0,0 +1,36 @@
From 39b381252c32275079344d30de18b76fda4bba26 Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 27 Jul 2022 15:52:45 -0600
Subject: [PATCH 46/52] tests/00readonly: Run udevadm settle before setting ro
In some recent kernel versions, 00readonly fails with:
mdadm: failed to set readonly for /dev/md0: Device or resource busy
ERROR: array is not read-only!
This was traced down to a race condition with udev holding a reference
to the block device at the same time as trying to set it read only.
To fix this, call udevadm settle before setting the array read only.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
tests/00readonly | 1 +
1 file changed, 1 insertion(+)
diff --git a/tests/00readonly b/tests/00readonly
index 39202487..afe243b3 100644
--- a/tests/00readonly
+++ b/tests/00readonly
@@ -12,6 +12,7 @@ do
$dev1 $dev2 $dev3 $dev4 --assume-clean
check nosync
check $level
+ udevadm settle
mdadm -ro $md0
check readonly
state=$(cat /sys/block/md0/md/array_state)
--
2.31.1

@ -0,0 +1,119 @@
From b7671c82010ffc04dfaecff2dd19ef8b2283e2b6 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 19 Jul 2022 14:48:21 +0200
Subject: [PATCH 47/52] tests: add test for names
Current behavior is not documented and tested. This test is a base for
future improvements. It is enough to test it only with native metadata,
because it is generic code. Generated properties are passed to metadata
handler.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
tests/00createnames | 93 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 93 insertions(+)
create mode 100644 tests/00createnames
diff --git a/tests/00createnames b/tests/00createnames
new file mode 100644
index 00000000..64b81b92
--- /dev/null
+++ b/tests/00createnames
@@ -0,0 +1,93 @@
+set -x -e
+
+# Test how <devname> and --name= are handled for create mode.
+# We need to check three properties, generated from those parameters:
+# - devnode name
+# - link in /dev/md/ (MD_DEVNAME property from --detail --export)
+# - name in metadata (MD_NAME property from --examine --export)
+
+function _verify() {
+ local DEVNODE_NAME="$1"
+ local WANTED_LINK="$2"
+ local WANTED_NAME="$3"
+
+ local RES="$(mdadm -D --export $DEVNODE_NAME | grep MD_DEVNAME)"
+ if [[ "$?" != "0" ]]; then
+ echo "Cannot get details for $DEVNODE_NAME - unexpected devnode."
+ exit 1
+ fi
+
+ if [[ "$WANTED_LINK" != "empty" ]]; then
+ local EXPECTED="MD_DEVNAME=$WANTED_LINK"
+ if [[ "$RES" != "$EXPECTED" ]]; then
+ echo "$RES doesn't match $EXPECTED."
+ exit 1
+ fi
+ fi
+
+
+ local RES="$(mdadm -E --export $dev0 | grep MD_NAME)"
+ if [[ "$?" != "0" ]]; then
+ echo "Cannot get metadata from $dev0."
+ exit 1
+ fi
+
+ local EXPECTED="MD_NAME=$(hostname):$WANTED_NAME"
+ if [[ "$RES" != "$EXPECTED" ]]; then
+ echo "$RES doesn't match $EXPECTED."
+ exit 1
+ fi
+}
+
+function _create() {
+ local DEVNAME=$1
+ local NAME=$2
+
+ if [[ -z "$NAME" ]]; then
+ mdadm -CR "$DEVNAME" -l0 -n 1 $dev0 --force
+ else
+ mdadm -CR "$DEVNAME" --name="$NAME" -l0 -n 1 $dev0 --force
+ fi
+
+ if [[ "$?" != "0" ]]; then
+ echo "Cannot create device."
+ exit 1
+ fi
+}
+
+# The most trivial case.
+_create "/dev/md/name"
+_verify "/dev/md127" "name" "name"
+mdadm -S "/dev/md127"
+
+_create "name"
+_verify "/dev/md127" "name" "name"
+mdadm -S "/dev/md127"
+
+# Use 'mdX' as name.
+_create "/dev/md/md0"
+_verify "/dev/md127" "md0" "md0"
+mdadm -S "/dev/md127"
+
+_create "md0"
+_verify "/dev/md127" "md0" "md0"
+mdadm -S "/dev/md127"
+
+# <devnode> is used to create MD_DEVNAME but, name is used to create MD_NAME.
+_create "/dev/md/devnode" "name"
+_verify "/dev/md127" "devnode" "name"
+mdadm -S "/dev/md127"
+
+_create "devnode" "name"
+_verify "/dev/md127" "devnode" "name"
+mdadm -S "/dev/md127"
+
+# Devnode points to /dev/ directory. MD_DEVNAME doesn't exist.
+_create "/dev/md0"
+_verify "/dev/md0" "empty" "0"
+mdadm -S "/dev/md0"
+
+# Devnode points to /dev/ directory and name is set.
+_create "/dev/md0" "name"
+_verify "/dev/md0" "empty" "name"
+mdadm -S "/dev/md0"
--
2.31.1

@ -0,0 +1,176 @@
From e4a030a0d3a953b8e74c118200e58dc83c2fc608 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 19 Jul 2022 14:48:22 +0200
Subject: [PATCH 48/52] mdadm: remove symlink option
The option is not used. Remove it from code.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
ReadMe.c | 1 -
config.c | 7 +------
mdadm.8.in | 9 ---------
mdadm.c | 20 --------------------
mdadm.conf.5.in | 15 ---------------
mdadm.h | 2 --
6 files changed, 1 insertion(+), 53 deletions(-)
diff --git a/ReadMe.c b/ReadMe.c
index 7518a32a..7f94847e 100644
--- a/ReadMe.c
+++ b/ReadMe.c
@@ -147,7 +147,6 @@ struct option long_options[] = {
{"nofailfast",0, 0, NoFailFast},
{"re-add", 0, 0, ReAdd},
{"homehost", 1, 0, HomeHost},
- {"symlinks", 1, 0, Symlinks},
{"data-offset",1, 0, DataOffset},
{"nodes",1, 0, Nodes}, /* also for --assemble */
{"home-cluster",1, 0, ClusterName},
diff --git a/config.c b/config.c
index 9c725457..dc1620c1 100644
--- a/config.c
+++ b/config.c
@@ -194,7 +194,6 @@ struct mddev_dev *load_containers(void)
struct createinfo createinfo = {
.autof = 2, /* by default, create devices with standard names */
- .symlinks = 1,
.names = 0, /* By default, stick with numbered md devices. */
.bblist = 1, /* Use a bad block list by default */
#ifdef DEBIAN
@@ -310,11 +309,7 @@ static void createline(char *line)
if (!createinfo.supertype)
pr_err("metadata format %s unknown, ignoring\n",
w+9);
- } else if (strncasecmp(w, "symlinks=yes", 12) == 0)
- createinfo.symlinks = 1;
- else if (strncasecmp(w, "symlinks=no", 11) == 0)
- createinfo.symlinks = 0;
- else if (strncasecmp(w, "names=yes", 12) == 0)
+ } else if (strncasecmp(w, "names=yes", 12) == 0)
createinfo.names = 1;
else if (strncasecmp(w, "names=no", 11) == 0)
createinfo.names = 0;
diff --git a/mdadm.8.in b/mdadm.8.in
index 0be02e4a..f2736226 100644
--- a/mdadm.8.in
+++ b/mdadm.8.in
@@ -1048,11 +1048,6 @@ simultaneously. If not specified, this defaults to 4.
Specify journal device for the RAID-4/5/6 array. The journal device
should be a SSD with reasonable lifetime.
-.TP
-.BR \-\-symlinks
-Auto creation of symlinks in /dev to /dev/md, option --symlinks must
-be 'no' or 'yes' and work with --create and --build.
-
.TP
.BR \-k ", " \-\-consistency\-policy=
Specify how the array maintains consistency in case of unexpected shutdown.
@@ -1405,10 +1400,6 @@ Reshape can be continued later using the
.B \-\-continue
option for the grow command.
-.TP
-.BR \-\-symlinks
-See this option under Create and Build options.
-
.SH For Manage mode:
.TP
diff --git a/mdadm.c b/mdadm.c
index 56722ed9..180f7a9c 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -59,7 +59,6 @@ int main(int argc, char *argv[])
struct mddev_dev *dv;
mdu_array_info_t array;
int devs_found = 0;
- char *symlinks = NULL;
int grow_continue = 0;
/* autof indicates whether and how to create device node.
* bottom 3 bits are style. Rest (when shifted) are number of parts
@@ -663,13 +662,6 @@ int main(int argc, char *argv[])
case O(ASSEMBLE,Auto): /* auto-creation of device node */
c.autof = parse_auto(optarg, "--auto flag", 0);
continue;
-
- case O(CREATE,Symlinks):
- case O(BUILD,Symlinks):
- case O(ASSEMBLE,Symlinks): /* auto creation of symlinks in /dev to /dev/md */
- symlinks = optarg;
- continue;
-
case O(BUILD,'f'): /* force honouring '-n 1' */
case O(BUILD,Force): /* force honouring '-n 1' */
case O(GROW,'f'): /* ditto */
@@ -1325,18 +1317,6 @@ int main(int argc, char *argv[])
exit(2);
}
- if (symlinks) {
- struct createinfo *ci = conf_get_create_info();
-
- if (strcasecmp(symlinks, "yes") == 0)
- ci->symlinks = 1;
- else if (strcasecmp(symlinks, "no") == 0)
- ci->symlinks = 0;
- else {
- pr_err("option --symlinks must be 'no' or 'yes'\n");
- exit(2);
- }
- }
/* Ok, got the option parsing out of the way
* hopefully it's mostly right but there might be some stuff
* missing
diff --git a/mdadm.conf.5.in b/mdadm.conf.5.in
index cd4e6a9d..bc2295c2 100644
--- a/mdadm.conf.5.in
+++ b/mdadm.conf.5.in
@@ -338,21 +338,6 @@ missing device entries should be created.
The name of the metadata format to use if none is explicitly given.
This can be useful to impose a system-wide default of version-1 superblocks.
-.TP
-.B symlinks=no
-Normally when creating devices in
-.B /dev/md/
-.I mdadm
-will create a matching symlink from
-.B /dev/
-with a name starting
-.B md
-or
-.BR md_ .
-Give
-.B symlinks=no
-to suppress this symlink creation.
-
.TP
.B names=yes
Since Linux 2.6.29 it has been possible to create
diff --git a/mdadm.h b/mdadm.h
index add9c0b6..93e72786 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -394,7 +394,6 @@ struct createinfo {
int gid;
int autof;
int mode;
- int symlinks;
int names;
int bblist;
struct supertype *supertype;
@@ -442,7 +441,6 @@ enum special_options {
BackupFile,
HomeHost,
AutoHomeHost,
- Symlinks,
AutoDetect,
Waitclean,
DetailPlatform,
--
2.31.1

@ -0,0 +1,232 @@
From ae5dfc56b7a96805d5a0b50eaf93b9fec8604298 Mon Sep 17 00:00:00 2001
From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Date: Tue, 19 Jul 2022 14:48:23 +0200
Subject: [PATCH 49/52] mdadm: move data_offset to struct shape
Data offset is a shape property so move it there to remove additional
parameter from some functions.
Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Create.c | 16 ++++++++--------
Grow.c | 7 +++----
mdadm.c | 20 +++++++++-----------
mdadm.h | 5 ++---
4 files changed, 22 insertions(+), 26 deletions(-)
diff --git a/Create.c b/Create.c
index c84c1ac8..e06ec2ae 100644
--- a/Create.c
+++ b/Create.c
@@ -95,7 +95,7 @@ int Create(struct supertype *st, char *mddev,
char *name, int *uuid,
int subdevs, struct mddev_dev *devlist,
struct shape *s,
- struct context *c, unsigned long long data_offset)
+ struct context *c)
{
/*
* Create a new raid array.
@@ -288,7 +288,7 @@ int Create(struct supertype *st, char *mddev,
newsize = s->size * 2;
if (st && ! st->ss->validate_geometry(st, s->level, s->layout, s->raiddisks,
&s->chunk, s->size*2,
- data_offset, NULL,
+ s->data_offset, NULL,
&newsize, s->consistency_policy,
c->verbose >= 0))
return 1;
@@ -323,10 +323,10 @@ int Create(struct supertype *st, char *mddev,
info.array.working_disks = 0;
dnum = 0;
for (dv = devlist; dv; dv = dv->next)
- if (data_offset == VARIABLE_OFFSET)
+ if (s->data_offset == VARIABLE_OFFSET)
dv->data_offset = INVALID_SECTORS;
else
- dv->data_offset = data_offset;
+ dv->data_offset = s->data_offset;
for (dv=devlist; dv && !have_container; dv=dv->next, dnum++) {
char *dname = dv->devname;
@@ -342,7 +342,7 @@ int Create(struct supertype *st, char *mddev,
missing_disks ++;
continue;
}
- if (data_offset == VARIABLE_OFFSET) {
+ if (s->data_offset == VARIABLE_OFFSET) {
doff = strchr(dname, ':');
if (doff) {
*doff++ = 0;
@@ -350,7 +350,7 @@ int Create(struct supertype *st, char *mddev,
} else
dv->data_offset = INVALID_SECTORS;
} else
- dv->data_offset = data_offset;
+ dv->data_offset = s->data_offset;
dfd = open(dname, O_RDONLY);
if (dfd < 0) {
@@ -535,7 +535,7 @@ int Create(struct supertype *st, char *mddev,
if (!st->ss->validate_geometry(st, s->level, s->layout,
s->raiddisks,
&s->chunk, minsize*2,
- data_offset,
+ s->data_offset,
NULL, NULL,
s->consistency_policy, 0)) {
pr_err("devices too large for RAID level %d\n", s->level);
@@ -754,7 +754,7 @@ int Create(struct supertype *st, char *mddev,
}
}
if (!st->ss->init_super(st, &info.array, s, name, c->homehost, uuid,
- data_offset))
+ s->data_offset))
goto abort_locked;
total_slots = info.array.nr_disks;
diff --git a/Grow.c b/Grow.c
index 5780635a..868bdc3a 100644
--- a/Grow.c
+++ b/Grow.c
@@ -1775,7 +1775,6 @@ static int reshape_container(char *container, char *devname,
int Grow_reshape(char *devname, int fd,
struct mddev_dev *devlist,
- unsigned long long data_offset,
struct context *c, struct shape *s)
{
/* Make some changes in the shape of an array.
@@ -1821,7 +1820,7 @@ int Grow_reshape(char *devname, int fd,
return 1;
}
- if (data_offset != INVALID_SECTORS && array.level != 10 &&
+ if (s->data_offset != INVALID_SECTORS && array.level != 10 &&
(array.level < 4 || array.level > 6)) {
pr_err("--grow --data-offset not yet supported\n");
return 1;
@@ -2179,7 +2178,7 @@ size_change_error:
if ((s->level == UnSet || s->level == array.level) &&
(s->layout_str == NULL) &&
(s->chunk == 0 || s->chunk == array.chunk_size) &&
- data_offset == INVALID_SECTORS &&
+ s->data_offset == INVALID_SECTORS &&
(s->raiddisks == 0 || s->raiddisks == array.raid_disks)) {
/* Nothing more to do */
if (!changed && c->verbose >= 0)
@@ -2379,7 +2378,7 @@ size_change_error:
}
sync_metadata(st);
rv = reshape_array(container, fd, devname, st, &info, c->force,
- devlist, data_offset, c->backup_file,
+ devlist, s->data_offset, c->backup_file,
c->verbose, 0, 0, 0);
frozen = 0;
}
diff --git a/mdadm.c b/mdadm.c
index 180f7a9c..845e4466 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -49,7 +49,6 @@ int main(int argc, char *argv[])
int i;
unsigned long long array_size = 0;
- unsigned long long data_offset = INVALID_SECTORS;
struct mddev_ident ident;
char *configfile = NULL;
int devmode = 0;
@@ -79,6 +78,7 @@ int main(int argc, char *argv[])
.layout = UnSet,
.bitmap_chunk = UnSet,
.consistency_policy = CONSISTENCY_POLICY_UNKNOWN,
+ .data_offset = INVALID_SECTORS,
};
char sys_hostname[256];
@@ -479,15 +479,15 @@ int main(int argc, char *argv[])
case O(CREATE,DataOffset):
case O(GROW,DataOffset):
- if (data_offset != INVALID_SECTORS) {
+ if (s.data_offset != INVALID_SECTORS) {
pr_err("data-offset may only be specified one. Second value is %s.\n", optarg);
exit(2);
}
if (mode == CREATE && strcmp(optarg, "variable") == 0)
- data_offset = VARIABLE_OFFSET;
+ s.data_offset = VARIABLE_OFFSET;
else
- data_offset = parse_size(optarg);
- if (data_offset == INVALID_SECTORS) {
+ s.data_offset = parse_size(optarg);
+ if (s.data_offset == INVALID_SECTORS) {
pr_err("invalid data-offset: %s\n",
optarg);
exit(2);
@@ -1416,7 +1416,7 @@ int main(int argc, char *argv[])
exit(1);
}
- if (c.backup_file && data_offset != INVALID_SECTORS) {
+ if (c.backup_file && s.data_offset != INVALID_SECTORS) {
pr_err("--backup-file and --data-offset are incompatible\n");
exit(2);
}
@@ -1587,8 +1587,7 @@ int main(int argc, char *argv[])
rv = Create(ss, devlist->devname,
ident.name, ident.uuid_set ? ident.uuid : NULL,
- devs_found-1, devlist->next,
- &s, &c, data_offset);
+ devs_found - 1, devlist->next, &s, &c);
break;
case MISC:
if (devmode == 'E') {
@@ -1706,10 +1705,9 @@ int main(int argc, char *argv[])
c.verbose);
else if (s.size > 0 || s.raiddisks || s.layout_str ||
s.chunk != 0 || s.level != UnSet ||
- data_offset != INVALID_SECTORS) {
+ s.data_offset != INVALID_SECTORS) {
rv = Grow_reshape(devlist->devname, mdfd,
- devlist->next,
- data_offset, &c, &s);
+ devlist->next, &c, &s);
} else if (s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN) {
rv = Grow_consistency_policy(devlist->devname, mdfd, &c, &s);
} else if (array_size == 0)
diff --git a/mdadm.h b/mdadm.h
index 93e72786..adb7cdaa 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -595,6 +595,7 @@ struct shape {
int assume_clean;
int write_behind;
unsigned long long size;
+ unsigned long long data_offset;
int consistency_policy;
};
@@ -1431,7 +1432,6 @@ extern int Grow_addbitmap(char *devname, int fd,
struct context *c, struct shape *s);
extern int Grow_reshape(char *devname, int fd,
struct mddev_dev *devlist,
- unsigned long long data_offset,
struct context *c, struct shape *s);
extern int Grow_restart(struct supertype *st, struct mdinfo *info,
int *fdlist, int cnt, char *backup_file, int verbose);
@@ -1462,8 +1462,7 @@ extern int Create(struct supertype *st, char *mddev,
char *name, int *uuid,
int subdevs, struct mddev_dev *devlist,
struct shape *s,
- struct context *c,
- unsigned long long data_offset);
+ struct context *c);
extern int Detail(char *dev, struct context *c);
extern int Detail_Platform(struct superswitch *ss, int scan, int verbose, int export, char *controller_path);
--
2.31.1

@ -0,0 +1,162 @@
From 27ad4900501c615b7c6b266bf23948e5606dba53 Mon Sep 17 00:00:00 2001
From: Logan Gunthorpe <logang@deltatee.com>
Date: Wed, 27 Jul 2022 15:52:46 -0600
Subject: [PATCH 50/52] mdadm: Don't open md device for CREATE and ASSEMBLE
The mdadm command tries to open the md device for most modes, first
thing, no matter what. When running to create or assemble an array,
in most cases, the md device will not exist, the open call will fail
and everything will proceed correctly.
However, when running tests, a create or assembly command may be run
shortly after stopping an array and the old md device file may still
be around. Then, if create_on_open is set in the kernel, a new md
device will be created when mdadm does its initial open.
When mdadm gets around to creating the new device with the new_array
parameter it issues this error:
mdadm: Fail to create md0 when using
/sys/module/md_mod/parameters/new_array, fallback to creation via node
This is because an mddev was already created by the kernel with the
earlier open() call and thus the new one being created will fail with
EEXIST. The mdadm command will still successfully be created due to
falling back to the node creation method. However, the error message
itself will fail any test that's running it.
This issue is a race condition that is very rare, but a recent change
in the kernel caused this to happen more frequently: about 1 in 50
times.
To fix this, don't bother trying to open the md device for CREATE,
ASSEMBLE and BUILD commands, as the file descriptor will never be used
anyway even if it is successfully openned. The mdfd has not been used
for these commands since:
7f91af49ad09 ("Delay creation of array devices for assemble/build/create")
The checks that were done on the open device can be changed to being
done with stat.
Side note: it would be nice to disable create_on_open as well to help
solve this, but it seems the work for this was never finished. By default,
mdadm will create using the old node interface when a name is specified
unless the user specifically puts names=yes in a config file, which
doesn't seem to be common or desirable to require this..
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
lib.c | 12 ++++++++++++
mdadm.c | 40 ++++++++++++++++++++--------------------
mdadm.h | 1 +
3 files changed, 33 insertions(+), 20 deletions(-)
diff --git a/lib.c b/lib.c
index 7e3e3d47..e395b28d 100644
--- a/lib.c
+++ b/lib.c
@@ -164,6 +164,18 @@ char *stat2devnm(struct stat *st)
return devid2devnm(st->st_rdev);
}
+bool stat_is_md_dev(struct stat *st)
+{
+ if ((S_IFMT & st->st_mode) != S_IFBLK)
+ return false;
+ if (major(st->st_rdev) == MD_MAJOR)
+ return true;
+ if (major(st->st_rdev) == (unsigned)get_mdp_major())
+ return true;
+
+ return false;
+}
+
char *fd2devnm(int fd)
{
struct stat stb;
diff --git a/mdadm.c b/mdadm.c
index 845e4466..972adb52 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -1329,6 +1329,9 @@ int main(int argc, char *argv[])
if (mode == MANAGE || mode == BUILD || mode == CREATE ||
mode == GROW || (mode == ASSEMBLE && ! c.scan)) {
+ struct stat stb;
+ int ret;
+
if (devs_found < 1) {
pr_err("an md device must be given in this mode\n");
exit(2);
@@ -1341,6 +1344,12 @@ int main(int argc, char *argv[])
mdfd = open_mddev(devlist->devname, 1);
if (mdfd < 0)
exit(1);
+
+ ret = fstat(mdfd, &stb);
+ if (ret) {
+ pr_err("fstat failed on %s.\n", devlist->devname);
+ exit(1);
+ }
} else {
char *bname = basename(devlist->devname);
@@ -1348,30 +1357,21 @@ int main(int argc, char *argv[])
pr_err("Name %s is too long.\n", devlist->devname);
exit(1);
}
- /* non-existent device is OK */
- mdfd = open_mddev(devlist->devname, 0);
- }
- if (mdfd == -2) {
- pr_err("device %s exists but is not an md array.\n", devlist->devname);
- exit(1);
- }
- if ((int)ident.super_minor == -2) {
- struct stat stb;
- if (mdfd < 0) {
+
+ ret = stat(devlist->devname, &stb);
+ if (ident.super_minor == -2 && ret != 0) {
pr_err("--super-minor=dev given, and listed device %s doesn't exist.\n",
- devlist->devname);
+ devlist->devname);
+ exit(1);
+ }
+
+ if (!ret && !stat_is_md_dev(&stb)) {
+ pr_err("device %s exists but is not an md array.\n", devlist->devname);
exit(1);
}
- fstat(mdfd, &stb);
- ident.super_minor = minor(stb.st_rdev);
- }
- if (mdfd >= 0 && mode != MANAGE && mode != GROW) {
- /* We don't really want this open yet, we just might
- * have wanted to check some things
- */
- close(mdfd);
- mdfd = -1;
}
+ if (ident.super_minor == -2)
+ ident.super_minor = minor(stb.st_rdev);
}
if (s.raiddisks) {
diff --git a/mdadm.h b/mdadm.h
index adb7cdaa..8208b81e 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -1672,6 +1672,7 @@ void *super1_make_v0(struct supertype *st, struct mdinfo *info, mdp_super_t *sb0
extern char *stat2kname(struct stat *st);
extern char *fd2kname(int fd);
extern char *stat2devnm(struct stat *st);
+bool stat_is_md_dev(struct stat *st);
extern char *fd2devnm(int fd);
extern void udev_block(char *devnm);
extern void udev_unblock(void);
--
2.31.1

@ -0,0 +1,231 @@
From 7211116c295ba1f9e1fcbdc2dd2d3762855062e1 Mon Sep 17 00:00:00 2001
From: Mateusz Kusiak <mateusz.kusiak@intel.com>
Date: Thu, 28 Jul 2022 20:20:53 +0800
Subject: [PATCH 51/52] Grow: Split Grow_reshape into helper function
Grow_reshape should be split into helper functions given its size.
- Add helper function for preparing reshape on external metadata.
- Close cfd file descriptor.
Signed-off-by: Mateusz Kusiak <mateusz.kusiak@intel.com>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Grow.c | 125 ++++++++++++++++++++++++++++++--------------------------
mdadm.h | 1 +
util.c | 14 +++++++
3 files changed, 81 insertions(+), 59 deletions(-)
diff --git a/Grow.c b/Grow.c
index 868bdc3a..0f07a894 100644
--- a/Grow.c
+++ b/Grow.c
@@ -1773,6 +1773,65 @@ static int reshape_container(char *container, char *devname,
char *backup_file, int verbose,
int forked, int restart, int freeze_reshape);
+/**
+ * prepare_external_reshape() - prepares update on external metadata if supported.
+ * @devname: Device name.
+ * @subarray: Subarray.
+ * @st: Supertype.
+ * @container: Container.
+ * @cfd: Container file descriptor.
+ *
+ * Function checks that the requested reshape is supported on external metadata,
+ * and performs an initial check that the container holds the pre-requisite
+ * spare devices (mdmon owns final validation).
+ *
+ * Return: 0 on success, else 1
+ */
+static int prepare_external_reshape(char *devname, char *subarray,
+ struct supertype *st, char *container,
+ const int cfd)
+{
+ struct mdinfo *cc = NULL;
+ struct mdinfo *content = NULL;
+
+ if (st->ss->load_container(st, cfd, NULL)) {
+ pr_err("Cannot read superblock for %s\n", devname);
+ return 1;
+ }
+
+ if (!st->ss->container_content)
+ return 1;
+
+ cc = st->ss->container_content(st, subarray);
+ for (content = cc; content ; content = content->next) {
+ /*
+ * check if reshape is allowed based on metadata
+ * indications stored in content.array.status
+ */
+ if (is_bit_set(&content->array.state, MD_SB_BLOCK_VOLUME) ||
+ is_bit_set(&content->array.state, MD_SB_BLOCK_CONTAINER_RESHAPE)) {
+ pr_err("Cannot reshape arrays in container with unsupported metadata: %s(%s)\n",
+ devname, container);
+ goto error;
+ }
+ if (content->consistency_policy == CONSISTENCY_POLICY_PPL) {
+ pr_err("Operation not supported when ppl consistency policy is enabled\n");
+ goto error;
+ }
+ if (content->consistency_policy == CONSISTENCY_POLICY_BITMAP) {
+ pr_err("Operation not supported when write-intent bitmap consistency policy is enabled\n");
+ goto error;
+ }
+ }
+ sysfs_free(cc);
+ if (mdmon_running(container))
+ st->update_tail = &st->updates;
+ return 0;
+error:
+ sysfs_free(cc);
+ return 1;
+}
+
int Grow_reshape(char *devname, int fd,
struct mddev_dev *devlist,
struct context *c, struct shape *s)
@@ -1799,7 +1858,7 @@ int Grow_reshape(char *devname, int fd,
struct supertype *st;
char *subarray = NULL;
- int frozen;
+ int frozen = 0;
int changed = 0;
char *container = NULL;
int cfd = -1;
@@ -1808,7 +1867,7 @@ int Grow_reshape(char *devname, int fd,
int added_disks;
struct mdinfo info;
- struct mdinfo *sra;
+ struct mdinfo *sra = NULL;
if (md_get_array_info(fd, &array) < 0) {
pr_err("%s is not an active md array - aborting\n",
@@ -1865,13 +1924,7 @@ int Grow_reshape(char *devname, int fd,
}
}
- /* in the external case we need to check that the requested reshape is
- * supported, and perform an initial check that the container holds the
- * pre-requisite spare devices (mdmon owns final validation)
- */
if (st->ss->external) {
- int retval;
-
if (subarray) {
container = st->container_devnm;
cfd = open_dev_excl(st->container_devnm);
@@ -1887,13 +1940,12 @@ int Grow_reshape(char *devname, int fd,
return 1;
}
- retval = st->ss->load_container(st, cfd, NULL);
-
- if (retval) {
- pr_err("Cannot read superblock for %s\n", devname);
- close(cfd);
+ rv = prepare_external_reshape(devname, subarray, st,
+ container, cfd);
+ if (rv > 0) {
free(subarray);
- return 1;
+ close(cfd);
+ goto release;
}
if (s->raiddisks && subarray) {
@@ -1902,51 +1954,6 @@ int Grow_reshape(char *devname, int fd,
free(subarray);
return 1;
}
-
- /* check if operation is supported for metadata handler */
- if (st->ss->container_content) {
- struct mdinfo *cc = NULL;
- struct mdinfo *content = NULL;
-
- cc = st->ss->container_content(st, subarray);
- for (content = cc; content ; content = content->next) {
- int allow_reshape = 1;
-
- /* check if reshape is allowed based on metadata
- * indications stored in content.array.status
- */
- if (content->array.state &
- (1 << MD_SB_BLOCK_VOLUME))
- allow_reshape = 0;
- if (content->array.state &
- (1 << MD_SB_BLOCK_CONTAINER_RESHAPE))
- allow_reshape = 0;
- if (!allow_reshape) {
- pr_err("cannot reshape arrays in container with unsupported metadata: %s(%s)\n",
- devname, container);
- sysfs_free(cc);
- free(subarray);
- return 1;
- }
- if (content->consistency_policy ==
- CONSISTENCY_POLICY_PPL) {
- pr_err("Operation not supported when ppl consistency policy is enabled\n");
- sysfs_free(cc);
- free(subarray);
- return 1;
- }
- if (content->consistency_policy ==
- CONSISTENCY_POLICY_BITMAP) {
- pr_err("Operation not supported when write-intent bitmap is enabled\n");
- sysfs_free(cc);
- free(subarray);
- return 1;
- }
- }
- sysfs_free(cc);
- }
- if (mdmon_running(container))
- st->update_tail = &st->updates;
}
added_disks = 0;
diff --git a/mdadm.h b/mdadm.h
index 8208b81e..941a5f38 100644
--- a/mdadm.h
+++ b/mdadm.h
@@ -1539,6 +1539,7 @@ extern int stat_is_blkdev(char *devname, dev_t *rdev);
extern bool is_dev_alive(char *path);
extern int get_mdp_major(void);
extern int get_maj_min(char *dev, int *major, int *minor);
+extern bool is_bit_set(int *val, unsigned char index);
extern int dev_open(char *dev, int flags);
extern int open_dev(char *devnm);
extern void reopen_mddev(int mdfd);
diff --git a/util.c b/util.c
index ca48d976..26ffdcea 100644
--- a/util.c
+++ b/util.c
@@ -1027,6 +1027,20 @@ int get_maj_min(char *dev, int *major, int *minor)
*e == 0);
}
+/**
+ * is_bit_set() - get bit value by index.
+ * @val: value.
+ * @index: index of the bit (LSB numbering).
+ *
+ * Return: bit value.
+ */
+bool is_bit_set(int *val, unsigned char index)
+{
+ if ((*val) & (1 << index))
+ return true;
+ return false;
+}
+
int dev_open(char *dev, int flags)
{
/* like 'open', but if 'dev' matches %d:%d, create a temp
--
2.31.1

@ -0,0 +1,36 @@
From 5c3c3df646dd3b7e8df81152f08e9ac4ddccc671 Mon Sep 17 00:00:00 2001
From: Kinga Tanska <kinga.tanska@intel.com>
Date: Fri, 19 Aug 2022 02:55:46 +0200
Subject: [PATCH 52/52] Assemble: check if device is container before
scheduling force-clean update
Up to now using assemble with force flag making each array as clean.
Force-clean should not be done for the container. This commit add
check if device is different than container before cleaning.
Signed-off-by: Kinga Tanska <kinga.tanska@intel.com>
Signed-off-by: Jes Sorensen <jsorensen@fb.com>
---
Assemble.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/Assemble.c b/Assemble.c
index be2160b4..1dd82a8c 100644
--- a/Assemble.c
+++ b/Assemble.c
@@ -1809,10 +1809,9 @@ try_again:
}
#endif
}
- if (c->force && !clean &&
+ if (c->force && !clean && content->array.level != LEVEL_CONTAINER &&
!enough(content->array.level, content->array.raid_disks,
- content->array.layout, clean,
- avail)) {
+ content->array.layout, clean, avail)) {
change += st->ss->update_super(st, content, "force-array",
devices[chosen_drive].devname, c->verbose,
0, NULL);
--
2.31.1

@ -0,0 +1,11 @@
--- mdadm/Makefile.orig 2021-07-28 21:39:23.887433859 +0800
+++ mdadm/Makefile 2021-07-28 21:39:37.989432841 +0800
@@ -50,7 +50,7 @@
CC := $(CROSS_COMPILE)gcc
endif
CXFLAGS ?= -ggdb
-CWFLAGS = -Wall -Werror -Wstrict-prototypes -Wextra -Wno-unused-parameter
+CWFLAGS = -Wall -Wstrict-prototypes -Wextra -Wno-unused-parameter
ifdef WARN_UNUSED
CWFLAGS += -Wp,-D_FORTIFY_SOURCE=2 -O3
endif

@ -0,0 +1,23 @@
--- mdadm-3.2.1/Makefile.static 2021-01-11 15:46:47.292126848 +0800
+++ mdadm-3.2.1/Makefile 2021-01-11 15:46:10.720192519 +0800
@@ -248,16 +248,16 @@
install : install-bin install-man install-udev
install-static : mdadm.static install-man
- $(INSTALL) -D $(STRIP) -m 755 mdadm.static $(DESTDIR)$(BINDIR)/mdadm
+ $(INSTALL) -D $(STRIP) -m 755 mdadm.static $(DESTDIR)$(BINDIR)/mdadm.static
install-tcc : mdadm.tcc install-man
- $(INSTALL) -D $(STRIP) -m 755 mdadm.tcc $(DESTDIR)$(BINDIR)/mdadm
+ $(INSTALL) -D $(STRIP) -m 755 mdadm.tcc $(DESTDIR)$(BINDIR)/mdadm.tcc
install-uclibc : mdadm.uclibc install-man
- $(INSTALL) -D $(STRIP) -m 755 mdadm.uclibc $(DESTDIR)$(BINDIR)/mdadm
+ $(INSTALL) -D $(STRIP) -m 755 mdadm.uclibc $(DESTDIR)$(BINDIR)/mdadm.uclibc
install-klibc : mdadm.klibc install-man
- $(INSTALL) -D $(STRIP) -m 755 mdadm.klibc $(DESTDIR)$(BINDIR)/mdadm
+ $(INSTALL) -D $(STRIP) -m 755 mdadm.klibc $(DESTDIR)$(BINDIR)/mdadm.klibc
install-man: mdadm.8 md.4 mdadm.conf.5 mdmon.8
$(INSTALL) -D -m 644 mdadm.8 $(DESTDIR)$(MAN8DIR)/mdadm.8

@ -0,0 +1,13 @@
--- mdadm/udev-md-raid-assembly.rules~ 2018-07-09 18:24:27.450774446 +0800
+++ mdadm/udev-md-raid-assembly.rules 2018-07-09 18:25:12.630735637 +0800
@@ -5,6 +5,10 @@
ENV{ANACONDA}=="?*", GOTO="md_inc_end"
# assemble md arrays
+# In Fedora we handle the raid components in 65-md-incremental.rules so that
+# we can do things like honor anaconda command line options and such
+GOTO="md_inc_end"
+
SUBSYSTEM!="block", GOTO="md_inc_end"
# skip non-initialized devices

@ -0,0 +1,60 @@
#!/bin/bash
#
# Configuration file for /etc/cron.weekly/raid-check
#
# options:
# ENABLED - must be yes in order for the raid check to proceed
# CHECK - can be either check or repair depending on the type of
# operation the user desires. A check operation will scan
# the drives looking for bad sectors and automatically
# repairing only bad sectors. If it finds good sectors that
# contain bad data (meaning that the data in a sector does
# not agree with what the data from another disk indicates
# the data should be, for example the parity block + the other
# data blocks would cause us to think that this data block
# is incorrect), then it does nothing but increments the
# counter in the file /sys/block/$dev/md/mismatch_count.
# This allows the sysadmin to inspect the data in the sector
# and the data that would be produced by rebuilding the
# sector from redundant information and pick the correct
# data to keep. The repair option does the same thing, but
# when it encounters a mismatch in the data, it automatically
# updates the data to be consistent. However, since we really
# don't know whether it's the parity or the data block that's
# correct (or which data block in the case of raid1), it's
# luck of the draw whether or not the user gets the right
# data instead of the bad data. This option is the default
# option for devices not listed in either CHECK_DEVS or
# REPAIR_DEVS.
# CHECK_DEVS - a space delimited list of devs that the user specifically
# wants to run a check operation on.
# REPAIR_DEVS - a space delimited list of devs that the user
# specifically wants to run a repair on.
# SKIP_DEVS - a space delimited list of devs that should be skipped
# NICE - Change the raid check CPU and IO priority in order to make
# the system more responsive during lengthy checks. Valid
# values are high, normal, low, idle.
# MAXCONCURENT - Limit the number of devices to be checked at a time.
# By default all devices will be checked at the same time.
#
# Note: the raid-check script intentionaly runs last in the cron.weekly
# sequence. This is so we can wait for all the resync operations to complete
# and then check the mismatch_count on each array without unduly delaying
# other weekly cron jobs. If any arrays have a non-0 mismatch_count after
# the check completes, we echo a warning to stdout which will then me emailed
# to the admin as long as mails from cron jobs have not been redirected to
# /dev/null. We do not wait for repair operations to complete as the
# md stack will correct any mismatch_cnts automatically.
#
# Note2: you can not use symbolic names for the raid devices, such as you
# /dev/md/root. The names used in this file must match the names seen in
# /proc/mdstat and in /sys/block.
ENABLED=yes
CHECK=check
NICE=low
# To check devs /dev/md0 and /dev/md3, use "md0 md3"
CHECK_DEVS=""
REPAIR_DEVS=""
SKIP_DEVS=""
MAXCONCURRENT=

@ -0,0 +1 @@
d /run/mdadm 0710 root root -

@ -0,0 +1,69 @@
# This file causes block devices with Linux RAID (mdadm) signatures to
# automatically cause mdadm to be run.
# See udev(8) for syntax
# Don't process any events if anaconda is running as anaconda brings up
# raid devices manually
ENV{ANACONDA}=="?*", GOTO="md_end"
# Also don't process disks that are slated to be a multipath device
ENV{DM_MULTIPATH_DEVICE_PATH}=="1", GOTO="md_end"
# We process add events on block devices (since they are ready as soon as
# they are added to the system), but we must process change events as well
# on any dm devices (like LUKS partitions or LVM logical volumes) and on
# md devices because both of these first get added, then get brought live
# and trigger a change event. The reason we don't process change events
# on bare hard disks is because if you stop all arrays on a disk, then
# run fdisk on the disk to change the partitions, when fdisk exits it
# triggers a change event, and we want to wait until all the fdisks on
# all member disks are done before we do anything. Unfortunately, we have
# no way of knowing that, so we just have to let those arrays be brought
# up manually after fdisk has been run on all of the disks.
# First, process all add events (md and dm devices will not really do
# anything here, just regular disks, and this also won't get any imsm
# array members either)
SUBSYSTEM=="block", ACTION=="add", ENV{ID_FS_TYPE}=="linux_raid_member", \
IMPORT{program}="/sbin/mdadm -I $env{DEVNAME} --export $devnode --offroot $${DEVLINKS}"
SUBSYSTEM=="block", ACTION=="add", ENV{ID_FS_TYPE}=="linux_raid_member", \
ENV{MD_STARTED}=="*unsafe*", ENV{MD_FOREIGN}=="no", ENV{SYSTEMD_WANTS}+="mdadm-last-resort@$env{MD_DEVICE}.timer"
SUBSYSTEM=="block", ACTION=="remove", ENV{ID_PATH}=="?*", \
ENV{ID_FS_TYPE}=="linux_raid_member", \
RUN+="/sbin/mdadm -If $name --path $env{ID_PATH}"
SUBSYSTEM=="block", ACTION=="remove", ENV{ID_PATH}!="?*", \
ENV{ID_FS_TYPE}=="linux_raid_member", \
RUN+="/sbin/mdadm -If $name"
# Next, check to make sure the BIOS raid stuff wasn't turned off via cmdline
IMPORT{cmdline}="noiswmd"
IMPORT{cmdline}="nodmraid"
ENV{noiswmd}=="?*", GOTO="md_imsm_inc_end"
ENV{nodmraid}=="?*", GOTO="md_imsm_inc_end"
SUBSYSTEM=="block", ACTION=="add", ENV{ID_FS_TYPE}=="isw_raid_member", \
RUN+="/sbin/mdadm -I $env{DEVNAME}"
SUBSYSTEM=="block", ACTION=="add", ENV{ID_FS_TYPE}=="ddf_raid_member", \
RUN+="/sbin/mdadm -I $env{DEVNAME}"
SUBSYSTEM=="block", ACTION=="remove", ENV{ID_PATH}=="?*", \
ENV{ID_FS_TYPE}=="isw_raid_member", \
RUN+="/sbin/mdadm -If $name --path $env{ID_PATH}"
SUBSYSTEM=="block", ACTION=="remove", ENV{ID_PATH}!="?*", \
ENV{ID_FS_TYPE}=="isw_raid_member", \
RUN+="/sbin/mdadm -If $name"
LABEL="md_imsm_inc_end"
# Next make sure that this isn't a dm device we should skip for some reason
ENV{DM_UDEV_RULES_VSN}!="?*", GOTO="dm_change_end"
ENV{DM_UDEV_DISABLE_OTHER_RULES_FLAG}=="1", GOTO="dm_change_end"
ENV{DM_SUSPENDED}=="1", GOTO="dm_change_end"
KERNEL=="dm-*", SUBSYSTEM=="block", ENV{ID_FS_TYPE}=="linux_raid_member", \
ACTION=="change", RUN+="/sbin/mdadm -I $env{DEVNAME}"
LABEL="dm_change_end"
# Finally catch any nested md raid arrays. If we brought up an md raid
# array that's part of another md raid array, it won't be ready to be used
# until the change event that occurs when it becomes live
KERNEL=="md*", SUBSYSTEM=="block", ENV{ID_FS_TYPE}=="linux_raid_member", \
ACTION=="change", RUN+="/sbin/mdadm -I $env{DEVNAME}"
LABEL="md_end"

@ -0,0 +1,5 @@
# Save /proc/mdstat in case of crash in mdadm/mdmon
EVENT=post-create component=mdadm
cat /proc/mdstat >> mdstat_data
echo "Saved output of /proc/mdstat"

@ -0,0 +1,12 @@
[Unit]
Description=Software RAID monitoring and management
ConditionPathExists=/etc/mdadm.conf
[Service]
Type=forking
PIDFile=/run/mdadm/mdadm.pid
EnvironmentFile=-/etc/sysconfig/mdmonitor
ExecStart=/sbin/mdadm --monitor --scan --syslog -f --pid-file=/run/mdadm/mdadm.pid
[Install]
WantedBy=multi-user.target

@ -0,0 +1,135 @@
#!/bin/bash
#
# This script reads it's configuration from /etc/sysconfig/raid-check
# Please use that file to enable/disable this script or to set the
# type of check you wish performed.
# We might be on a kernel with no raid support at all, exit if so
[ -f /proc/mdstat ] || exit 0
# and exit if we haven't been set up properly
[ -f /etc/sysconfig/raid-check ] || exit 0
. /etc/sysconfig/raid-check
# Wait until no more than arg1 arrays in arg2 list are busy
waitbusy() {
local threshold=$(($1 + 1))
local dev_list="$2"
while true
do
local busy=0
local dev=""
for dev in $dev_list; do
local sync_action=`cat /sys/block/$dev/md/sync_action`
if [ "$sync_action" != "idle" ]; then
let busy++
fi
done
[ $busy -lt $threshold ] && break
sleep 60
done
}
[ "$ENABLED" != "yes" ] && exit 0
case "$CHECK" in
check) ;;
repair) ;;
*) exit 0;;
esac
ionice=""
renice=""
case $NICE in
high)
renice="-n -5"
;;
low)
renice="-n 5"
ionice="-c2 -n7"
;;
idle)
renice="-n 15"
ionice="-c3"
;;
*)
;;
esac
active_list=`grep "^md.*: active" /proc/mdstat | cut -f 1 -d ' '`
[ -z "$active_list" ] && exit 0
declare -A check
dev_list=""
check_list=""
for dev in $active_list; do
echo $SKIP_DEVS | grep -w $dev >&/dev/null && continue
if [ -f /sys/block/$dev/md/sync_action ]; then
# Only perform the checks on idle, healthy arrays, but delay
# actually writing the check field until the next loop so we
# don't switch currently idle arrays to active, which happens
# when two or more arrays are on the same physical disk
array_state=`cat /sys/block/$dev/md/array_state`
if [ "$array_state" != "clean" -a "$array_state" != "active" ]; then
continue
fi
sync_action=`cat /sys/block/$dev/md/sync_action`
if [ "$sync_action" != idle ]; then
continue
fi
ck=""
echo $REPAIR_DEVS | grep -w $dev >&/dev/null && ck="repair"
echo $CHECK_DEVS | grep -w $dev >&/dev/null && ck="check"
[ -z "$ck" ] && ck=$CHECK
dev_list="$dev_list $dev"
check[$dev]=$ck
[ "$ck" = "check" ] && check_list="$check_list $dev"
fi
done
[ -z "$dev_list" ] && exit 0
for dev in $dev_list; do
#Only run $MAXCONCURRENT checks at a time
if [ -n "$MAXCONCURRENT" ]; then
waitbusy $((MAXCONCURRENT - 1)) "$dev_list"
fi
echo "${check[$dev]}" > /sys/block/$dev/md/sync_action
resync_pid=""
wait=10
while [ $wait -gt 0 -a -z "$resync_pid" ]; do
sleep 6
let wait--
resync_pid=$(ps -ef | awk -v mddev=$dev 'BEGIN { pattern = "^\\[" mddev "_resync]$" } $8 ~ pattern { print $2 }')
done
[ -n "$resync_pid" -a -n "$renice" ] &&
renice $renice -p $resync_pid >&/dev/null
[ -n "$resync_pid" -a -n "$ionice" ] &&
ionice $ionice -p $resync_pid >&/dev/null
done
[ -z "$check_list" ] && exit 0
waitbusy 0 "$check_list"
for dev in $check_list; do
mismatch_cnt=`cat /sys/block/$dev/md/mismatch_cnt`
# Due to the fact that raid1/10 writes in the kernel are unbuffered,
# a raid1 array can have non-0 mismatch counts even when the
# array is healthy. These non-0 counts will only exist in
# transient data areas where they don't pose a problem. However,
# since we can't tell the difference between a non-0 count that
# is just in transient data or a non-0 count that signifies a
# real problem, simply don't check the mismatch_cnt on raid1
# devices as it's providing far too many false positives. But by
# leaving the raid1 device in the check list and performing the
# check, we still catch and correct any bad sectors there might
# be in the device.
raid_lvl=`cat /sys/block/$dev/md/level`
if [ "$raid_lvl" = "raid1" -o "$raid_lvl" = "raid10" ]; then
continue
fi
if [ "$mismatch_cnt" -ne 0 ]; then
echo "WARNING: mismatch_cnt is not 0 on /dev/$dev"
fi
done

@ -0,0 +1,6 @@
[Unit]
Description=RAID setup health check
[Service]
Type=oneshot
ExecStart=/usr/sbin/raid-check

@ -0,0 +1,10 @@
[Unit]
Description=Weekly RAID setup health check
[Timer]
OnCalendar=Sun *-*-* 01:00:00
Persistent=true
AccuracySec=24h
[Install]
WantedBy=timers.target

File diff suppressed because it is too large Load Diff
Loading…
Cancel
Save