database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-07-27 18:02:13 +03:00
Code Activity

Revert MDEV-18464 and MDEV-12009

Browse Source 
This reverts commit 21b2fada7a
and commit 81d71ee6b2.

The MDEV-18464 change introduces a few data race issues. Contrary to
the documentation, the field trx_t::victim is not always being protected
by lock_sys_t::mutex and trx_t::mutex. Most importantly, it seems
that KILL QUERY could wrongly avoid acquiring both mutexes when
invoking lock_trx_handle_wait_low(), in case another thread had
already set trx->victim=true.

We also revert MDEV-12009, because it should depend on the MDEV-18464
fix being present.
This commit is contained in:
Marko Mäkelä
2019-03-28 12:27:06 +02:00
parent 81d71ee6b2
commit d0116e10a5
22 changed files with 178 additions and 194 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
sql/sql_parse.cc
View File

@ -8292,19 +8292,11 @@ kill_one_thread(THD *thd, longlong id, killed_state kill_signal, killed_type typ
It's ok to also kill DELAYED threads with KILL_CONNECTION instead of
KILL_SYSTEM_THREAD; The difference is that KILL_CONNECTION may be
faster and do a harder kill than KILL_SYSTEM_THREAD;
Note that if thread is wsrep Brute Force or applier thread we
allow killing it only when we're SUPER.
*/
if ((thd->security_ctx->master_access & SUPER_ACL) ||
(thd->security_ctx->user_matches(tmp->security_ctx)
#ifdef WITH_WSREP
&&
!tmp->wsrep_applier &&
!wsrep_thd_is_BF(tmp, false)
#endif
))
if (((thd->security_ctx->master_access & SUPER_ACL) ||
thd->security_ctx->user_matches(tmp->security_ctx)) &&
!wsrep_thd_is_BF(tmp, false))
{
tmp->awake(kill_signal);
error=0;