| Bug #55345 | mysql_library_init causes segfault if executed after calling mysql_library_end | ||
|---|---|---|---|
| Submitted: | 18 Jul 2010 12:46 | Modified: | 2 Sep 2010 15:57 |
| Reporter: | Oleg Sidorkin | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server: C API (client library) | Severity: | S2 (Serious) |
| Version: | 5.5.5-m3, 5.6.99 | OS: | Any (FreeBSD, Linux) |
| Assigned to: | Christopher Powers | CPU Architecture: | Any |
| Tags: | mysql_library_init, regression | ||
[19 Jul 2010 7:05]
Sveta Smirnova
Thank you for the report. I can not repeat described behavior on FreeBSD6. You wrote you use FreeBSD8. Which connector/C binaries do you use? If you compiled connector yourself please send us configure options you used.
[19 Jul 2010 7:05]
Susanne Ebrecht
Do you have similar problems by using MySQL server 5.1 instead 5.5?
[19 Jul 2010 9:10]
Oleg Sidorkin
Sorry. I've completely forgot about build flags. Both client and server were built with the following flags: WITH_XCHARSET=all BUILD_OPTIMIZED=yes WITHOUT_INNODB=no WITH_ARCHIEVE=yes WITH_FEDERATED=yes WITH_OPENSSL=yes WITH_CHARSET=koi8r WITH_FAST_MUTEXES=yes MySQL 5.1 worked with the same flags with no problems
[19 Jul 2010 9:17]
Sveta Smirnova
Thank you for the feedback. You opened bug in Connector/C category while speaking about compiling server and using server. Do you really use Connector/C or is is regular C API libraries bundled with server?
[19 Jul 2010 10:30]
Oleg Sidorkin
It is client library bug. Should I change the change the category of the bug?
[19 Jul 2010 11:20]
Sveta Smirnova
Thank you for the feedback. Verified as described. Connector/C is not affected.
[23 Aug 2010 1:43]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/116458 3190 Christopher Powers 2010-08-22 Bug #55345, "mysql_library_init causes segfault if executed after calling mysql_library_end" my_thread_global_end() now sets mysql_thread_basic_global_init_done= 0 to ensure destroyed mutexes are not reused. I verified that clearing this flag will not result in the redundant allocation of other resources allocated by my_thread_global_init() and my_thread_basic_global_init(). @ mysys/my_thr_init.c Clear mysql_thread_basic_global_init_done flag at the end of my_thread_global_end()
[25 Aug 2010 9:21]
Bugs System
Pushed into mysql-5.5 5.5.6-m3 (revid:alik@ibmvm-20100825092002-2yvkb3iwu43ycpnm) (version source revid:alik@ibmvm-20100825092002-2yvkb3iwu43ycpnm) (merge vers: 5.5.6-m3) (pib:20)
[27 Aug 2010 20:01]
Paul DuBois
Noted in 5.5.6 changelog. A call to mysql_library_init() following a call to mysql_library_end() caused a client crash. Setting report to Need Merge pending push to 5.6.x.
[30 Aug 2010 8:29]
Bugs System
Pushed into mysql-trunk 5.6.1-m4 (revid:alik@sun.com-20100830082732-n2eyijnv86exc5ci) (version source revid:alik@sun.com-20100830082732-n2eyijnv86exc5ci) (merge vers: 5.6.1-m4) (pib:21)
[30 Aug 2010 8:33]
Bugs System
Pushed into mysql-next-mr (revid:alik@sun.com-20100830082745-n6sh01wlwh3itasv) (version source revid:alik@sun.com-20100830082745-n6sh01wlwh3itasv) (pib:21)
[30 Aug 2010 15:07]
Paul DuBois
Noted in 5.6.1 changelog.
Description: If you invoke mysql_library_init() after calling mysql_library_end() you are getting segmentation fault due to passig already freed object to the libpthread. This happends due to my_thread_basic_global_init_done flag is set to 1 in my_thread_basic_global_init(), but it is not reset to 0 my_thread_global_end() where corresponding objects are deleted. On the second call mysql_library_init() deleted objects are not created again because my_thread_basic_global_init_done is still 1. Tested on FreeBSD 8.1 How to repeat: compile with -lmysqlclient_r and run the following: #include <my_global.h> #include <mysql.h> #include <stdio.h> int main(void). { if (mysql_library_init(0, 0, 0)) { printf("could not initialize MySQL library\n"); return 1; } mysql_library_end(); if (mysql_library_init(0, 0, 0)) { printf("could not initialize MySQL library\n"); return 1; } mysql_library_end(); return 0; } You will get core dump with the following backtrace: (gdb) bt #0 0x0000000801070328 in pthread_mutexattr_init () from /lib/libthr.so.3 #1 0x00000008006b18df in safe_mutex_init (mp=0x800adfa20, attr=0x800adfb70, file=0x800732828 "my_thr_init.c", line=217) at thr_mutex.c:62 #2 0x00000008006b0fe4 in my_thread_global_init () at mysql_thread.h:591 #3 0x00000008006a8d8c in my_basic_init () at my_init.c:103 #4 0x00000008006a8e08 in my_init () at my_init.c:146 #5 0x00000008006a7560 in mysql_server_init (argc=Variable "argc" is not available. ) at libmysql.c:126 #6 0x00000000004007bc in main () at main.c:16 Suggested fix: add my_thread_basic_global_init_done= 0; to the end of my_thread_global_end() or create new function my_thread_basic_global_end() to be able to control objects lifecycle carefully.