kris/boringssl
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Check SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER before touching wpend_buf.

Browse Source 
SSL_write has messy semantics around retries. As a sanity-check, it does
pointer and length checks and requires the original and retry SSL_write
pass the same buffer pointer.

In some cases, buffer addresses may change but still include the
original data as a prefix on the retry. Callers then set
SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER to skip the pointer check. But, in
that case, the pointer may have been freed so doing a comparison is
undefined behavior.

Short-circuiting the pointer equality check avoids this problem.

Change-Id: I76cb8f7d45533504cd95287bc53897ca636af51d
Reviewed-on: https://boringssl-review.googlesource.com/11760
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
This commit is contained in:
David Benjamin 2016-10-24 20:13:20 -04:00 committed by CQ bot account: commit-bot@chromium.org
parent 35c8afd314
commit c6722cd6e0
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ssl/s3_pkt.c
View File

@ -245,8 +245,8 @@ int ssl3_write_bytes(SSL *ssl, int type, const void *buf_, int len) {
static int ssl3_write_pending(SSL *ssl, int type, const uint8_t *buf,
unsigned int len) {
if (ssl->s3->wpend_tot > (int)len ||
(ssl->s3->wpend_buf != buf &&
!(ssl->mode & SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER)) ||
(!(ssl->mode & SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER) &&
ssl->s3->wpend_buf != buf) ||
ssl->s3->wpend_type != type) {
OPENSSL_PUT_ERROR(SSL, SSL_R_BAD_WRITE_RETRY);
return -1;