You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
openssl/SOURCES/0127-speedup-SSL_add_cert_s...

202 lines
7.4 KiB

From e2e469593a15681983d16e36d856bf8fb7de8589 Mon Sep 17 00:00:00 2001
From: Clemens Lang <cllang@redhat.com>
Date: Wed, 31 Jul 2024 12:45:11 +0200
Subject: [PATCH] Speed up SSL_add_{file,dir}_cert_subjects_to_stack
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The X509_NAME comparison function converts its arguments to DER using
i2d_X509_NAME before comparing the results using memcmp(). For every
invocation of the comparison function (of which there are many when
loading many certificates), it allocates two buffers of the appropriate
size for the DER encoding.
Switching to static buffers (possibly of X509_NAME_MAX size as defined
in crypto/x509/x_name.c) would not work with multithreaded use, e.g.,
when two threads sort two separate STACK_OF(X509_NAME)s at the same
time. A suitable re-usable buffer could have been added to the
STACK_OF(X509_NAME) if sk_X509_NAME_compfunc did have a void* argument,
or a pointer to the STACK_OF(X509_NAME) but it does not.
Instead, copy the solution chosen in SSL_load_client_CA_file() by
filling an LHASH_OF(X509_NAME) with all existing names in the stack and
using that to deduplicate, rather than relying on sk_X509_NAME_find(),
which ends up being very slow.
Adjust SSL_add_dir_cert_subjects_to_stack() to keep a local
LHASH_OF(X509_NAME)s over the complete directory it is processing.
In a small benchmark that calls SSL_add_dir_cert_subjects_to_stack()
twice, once on a directory with one entry, and once with a directory
with 1000 certificates, and repeats this in a loop 10 times, this change
yields a speed-up of 5.32:
| Benchmark 1: ./bench 10 dir-1 dir-1000
| Time (mean ± σ): 6.685 s ± 0.017 s [User: 6.402 s, System: 0.231 s]
| Range (min … max): 6.658 s … 6.711 s 10 runs
|
| Benchmark 2: LD_LIBRARY_PATH=. ./bench 10 dir-1 dir-1000
| Time (mean ± σ): 1.256 s ± 0.013 s [User: 1.034 s, System: 0.212 s]
| Range (min … max): 1.244 s … 1.286 s 10 runs
|
| Summary
| LD_LIBRARY_PATH=. ./bench 10 dir-1 dir-1000 ran
| 5.32 ± 0.06 times faster than ./bench 10 dir-1 dir-1000
In the worst case scenario where many entries are added to a stack that
is then repeatedly used to add more certificates, and with a larger test
size, the speedup is still very significant. With 15000 certificates,
a single pass to load them, followed by attempting to load a subset of
1000 of these 15000 certificates, followed by a single certificate, the
new approach is ~85 times faster:
| Benchmark 1: ./bench 1 dir-15000 dir-1000 dir-1
| Time (mean ± σ): 176.295 s ± 4.147 s [User: 174.593 s, System: 0.448 s]
| Range (min … max): 173.774 s … 185.594 s 10 runs
|
| Benchmark 2: LD_LIBRARY_PATH=. ./bench 1 dir-15000 dir-1000 dir-1
| Time (mean ± σ): 2.087 s ± 0.034 s [User: 1.679 s, System: 0.393 s]
| Range (min … max): 2.057 s … 2.167 s 10 runs
|
| Summary
| LD_LIBRARY_PATH=. ./bench 1 dir-15000 dir-1000 dir-1 ran
| 84.48 ± 2.42 times faster than ./bench 1 dir-15000 dir-1000 dir-1
Signed-off-by: Clemens Lang <cllang@redhat.com>
---
ssl/ssl_cert.c | 74 ++++++++++++++++++++++++++++++++++++++++++++------
1 file changed, 65 insertions(+), 9 deletions(-)
diff --git a/ssl/ssl_cert.c b/ssl/ssl_cert.c
index 0ff407bf55edc..5e5ffe39d0655 100644
--- a/ssl/ssl_cert.c
+++ b/ssl/ssl_cert.c
@@ -813,16 +813,14 @@ STACK_OF(X509_NAME) *SSL_load_client_CA_file(const char *file)
return SSL_load_client_CA_file_ex(file, NULL, NULL);
}
-int SSL_add_file_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
- const char *file)
+static int add_file_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
+ const char *file,
+ LHASH_OF(X509_NAME) *name_hash)
{
BIO *in;
X509 *x = NULL;
X509_NAME *xn = NULL;
int ret = 1;
- int (*oldcmp) (const X509_NAME *const *a, const X509_NAME *const *b);
-
- oldcmp = sk_X509_NAME_set_cmp_func(stack, xname_sk_cmp);
in = BIO_new(BIO_s_file());
@@ -842,12 +840,15 @@ int SSL_add_file_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
xn = X509_NAME_dup(xn);
if (xn == NULL)
goto err;
- if (sk_X509_NAME_find(stack, xn) >= 0) {
+ if (lh_X509_NAME_retrieve(name_hash, xn) != NULL) {
/* Duplicate. */
X509_NAME_free(xn);
} else if (!sk_X509_NAME_push(stack, xn)) {
X509_NAME_free(xn);
goto err;
+ } else {
+ /* Successful insert, add to hash table */
+ lh_X509_NAME_insert(name_hash, xn);
}
}
@@ -859,7 +860,42 @@ int SSL_add_file_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
done:
BIO_free(in);
X509_free(x);
- (void)sk_X509_NAME_set_cmp_func(stack, oldcmp);
+ return ret;
+}
+
+int SSL_add_file_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
+ const char *file)
+{
+ X509_NAME *xn = NULL;
+ int ret = 1;
+ int idx = 0;
+ int num = 0;
+ LHASH_OF(X509_NAME) *name_hash = lh_X509_NAME_new(xname_hash, xname_cmp);
+
+ if (name_hash == NULL) {
+ ERR_raise(ERR_LIB_SSL, ERR_R_CRYPTO_LIB);
+ goto err;
+ }
+
+ /*
+ * Pre-populate the lhash with the existing entries of the stack, since
+ * using the LHASH_OF is much faster for duplicate checking. That's because
+ * xname_cmp converts the X509_NAMEs to DER involving a memory allocation
+ * for every single invocation of the comparison function.
+ */
+ num = sk_X509_NAME_num(stack);
+ for (idx = 0; idx < num; idx++) {
+ xn = sk_X509_NAME_value(stack, idx);
+ lh_X509_NAME_insert(name_hash, xn);
+ }
+
+ ret = add_file_cert_subjects_to_stack(stack, file, name_hash);
+ goto done;
+
+ err:
+ ret = 0;
+ done:
+ lh_X509_NAME_free(name_hash);
return ret;
}
@@ -869,8 +905,27 @@ int SSL_add_dir_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
OPENSSL_DIR_CTX *d = NULL;
const char *filename;
int ret = 0;
+ X509_NAME *xn = NULL;
+ int idx = 0;
+ int num = 0;
+ LHASH_OF(X509_NAME) *name_hash = lh_X509_NAME_new(xname_hash, xname_cmp);
+
+ if (name_hash == NULL) {
+ ERR_raise(ERR_LIB_SSL, ERR_R_CRYPTO_LIB);
+ goto err;
+ }
- /* Note that a side effect is that the CAs will be sorted by name */
+ /*
+ * Pre-populate the lhash with the existing entries of the stack, since
+ * using the LHASH_OF is much faster for duplicate checking. That's because
+ * xname_cmp converts the X509_NAMEs to DER involving a memory allocation
+ * for every single invocation of the comparison function.
+ */
+ num = sk_X509_NAME_num(stack);
+ for (idx = 0; idx < num; idx++) {
+ xn = sk_X509_NAME_value(stack, idx);
+ lh_X509_NAME_insert(name_hash, xn);
+ }
while ((filename = OPENSSL_DIR_read(&d, dir))) {
char buf[1024];
@@ -899,7 +954,7 @@ int SSL_add_dir_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
#endif
if (r <= 0 || r >= (int)sizeof(buf))
goto err;
- if (!SSL_add_file_cert_subjects_to_stack(stack, buf))
+ if (!add_file_cert_subjects_to_stack(stack, buf, name_hash))
goto err;
}
@@ -915,6 +970,7 @@ int SSL_add_dir_cert_subjects_to_stack(STACK_OF(X509_NAME) *stack,
err:
if (d)
OPENSSL_DIR_end(&d);
+ lh_X509_NAME_free(name_hash);
return ret;
}