diff options
| author | Marko Mäkelä <marko.makela@mariadb.com> | 2019-03-29 11:38:45 +0200 |
|---|---|---|
| committer | Marko Mäkelä <marko.makela@mariadb.com> | 2019-03-29 11:38:45 +0200 |
| commit | fc168c3a5e58d8b364a2e87e0d876a261ec7fced (patch) | |
| tree | 3624e717802b484c7170267579557877f60a96a0 /mysys_ssl | |
| parent | 8fcd9478cc6fdb1d4cdb84a552d7dcf8e6f596d3 (diff) | |
| download | mariadb-git-fc168c3a5e58d8b364a2e87e0d876a261ec7fced.tar.gz | |
MDEV-15587 AES test fails, segfaults in EVP_CipherInit_ex
When HAVE_YASSL is defined (due to cmake -DWITH_SSL=bundled
or otherwise), mysys_ssl/my_crypt.cc will #include "yassl.cc"
from the same directory.
When MariaDB 10.2 or later is compiled with GCC 8 and optimizations
are enabled, then the check
if (iv)
in EVP_CipherInit_ex() can be wrongly optimized away.
The reason appears to be that __attribute__((nonnull)) is attached
to the variable iv, because there is a (no-op) call
memcpy(oiv, iv, ivlen=0) earlier in the code path.
It is possible that this started failing after the code was
refactored in MDEV-10332 (MariaDB 10.2.6). In MariaDB 10.1,
there is a similar memcpy() call in MyCTX_nopad::init(),
but the code appears to work fine.
Diffstat (limited to 'mysys_ssl')
| -rw-r--r-- | mysys_ssl/my_crypt.cc | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/mysys_ssl/my_crypt.cc b/mysys_ssl/my_crypt.cc index db303f37b0e..2d6f5188034 100644 --- a/mysys_ssl/my_crypt.cc +++ b/mysys_ssl/my_crypt.cc @@ -1,6 +1,6 @@ /* Copyright (c) 2014 Google Inc. - Copyright (c) 2014, 2017 MariaDB Corporation + Copyright (c) 2014, 2019, MariaDB Corporation. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -93,7 +93,8 @@ public: this->key= key; this->klen= klen; this->buf_len= 0; - memcpy(oiv, iv, ivlen); + if (ivlen) + memcpy(oiv, iv, ivlen); DBUG_ASSERT(ivlen == 0 || ivlen == sizeof(oiv)); int res= MyCTX::init(cipher, encrypt, key, klen, iv, ivlen); |