| Bug #98457 | Crash using mysql_real_connect | ||
|---|---|---|---|
| Submitted: | 1 Feb 2020 12:24 | Modified: | 10 Feb 2020 13:50 |
| Reporter: | Adrián G. | Email Updates: | |
| Status: | Not a Bug | Impact on me: | |
| Category: | Connector / C | Severity: | S1 (Critical) |
| Version: | 5.7.27 | OS: | Ubuntu (18.04.1) |
| Assigned to: | CPU Architecture: | x86 (x86_64) | |
| Tags: | linux, multithreaded, mysql_real_connect | ||
[1 Feb 2020 12:24]
Adrián G.
[1 Feb 2020 12:27]
Adrián G.
Description: Hi, the following crash seems to happen sometimes at mysql_real_connect. Call is performed inside code from a custom MySQL C wrapper interface library that's loaded dynamically (dlsym/dlopen) by a game server main library. It may be important to note that it's done on a separate thread, It seems it's caused by malloc. What potential causes could incur into it crashing with a SIGABRT signal? Thanks ---------------------------------------------- CRASH: Sat Feb 1 12:16:33 CET 2020 Start Line: ./srcds_linux -game hl2mp -debug -dumplongticks +maxplayers 32 +mp_teamplay 1 +map rp_c18_v1 [New LWP 7089] [New LWP 7075] [New LWP 7076] [New LWP 7077] [New LWP 7078] [New LWP 7079] [New LWP 7080] [New LWP 7081] [New LWP 7082] [New LWP 7086] [New LWP 7087] [New LWP 7088] [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by `./srcds_linux -game hl2mp -debug -dumplongticks +maxplayers 32 +mp_teamplay 1 +'. Program terminated with signal SIGABRT, Aborted. #0 0xf7f9d939 in __kernel_vsyscall () [Current thread is 1 (Thread 0xe9ff7b40 (LWP 7089))] #0 0xf7f9d939 in __kernel_vsyscall () #1 0xf7be3832 in raise () from /lib/i386-linux-gnu/libc.so.6 #2 0xf7be4cc1 in abort () from /lib/i386-linux-gnu/libc.so.6 #3 0xf7c2c66b in ?? () from /lib/i386-linux-gnu/libc.so.6 #4 0xf7c2fc70 in ?? () from /lib/i386-linux-gnu/libc.so.6 #5 0xf7c30e38 in malloc () from /lib/i386-linux-gnu/libc.so.6 #6 0xe9b23fc3 in my_raw_malloc (my_flags=16, size=<optimized out>) at /export/home/pb2/build/sb_0-34537258-1560178776.88/mysql-5.7.27/mysys/my_malloc.c:191 #7 my_malloc (key=0, size=396, flags=16) at /export/home/pb2/build/sb_0-34537258-1560178776.88/mysql-5.7.27/mysys/my_malloc.c:54 #8 0xe9b2e7c0 in mysql_socket_vio_new (mysql_socket=..., type=VIO_TYPE_TCPIP, flags=2) at /export/home/pb2/build/sb_0-34537258-1560178776.88/mysql-5.7.27/vio/vio.c:274 #9 0xe9b2eaa3 in vio_new (sd=27, type=VIO_TYPE_TCPIP, flags=2) at /export/home/pb2/build/sb_0-34537258-1560178776.88/mysql-5.7.27/vio/vio.c:293 #10 0xe9b04477 in mysql_real_connect (mysql=0xb0bcd0c, host=0xb6d6a90 "remotemysql.com", user=0xb6d64d0 "<my_username>", passwd=0xb6d63b0 "<my_password>", db=0xb6d6280 "fSkSY53EP5", port=3306, unix_socket=0x0, client_flag=0) at /export/home/pb2/build/sb_0-34537258-1560178776.88/mysql-5.7.27/sql-common/client.c:4277 #11 0xe9ace98d in CMySQLConnection::Connect (this=0xb0bcd00, pHostName=0xb6d6a90 "remotemysql.com", pUserName=0xb6d64d0 "<my_username>", pPassword=0xb6d63b0 "<my_password>", pSchemaName=0xb6d6280 "fSkSY53EP5", port=3306) at /home/vsts/work/1/s/mp/src/game/server/hl2rp/dal/mysql_driver/mysql_driver.cpp:344 #12 0xf1e0b4c4 in CDAL::LoadDatabaseConfiguration (this=0xf2667840 <gDAL>) at /home/vsts/work/1/s/mp/src/game/server/hl2rp/dal/dal.cpp:320 #13 0xf1e0aec9 in CDAL::Run (this=0xf2667840 <gDAL>) at /home/vsts/work/1/s/mp/src/game/server/hl2rp/dal/dal.cpp:208 #14 0xf7b82690 in CThread::ThreadProc(void*) () from bin/libtier0_srv.so #15 0xf7d983bd in start_thread () from /lib/i386-linux-gnu/libpthread.so.0 #16 0xf7caee16 in clone () from /lib/i386-linux-gnu/libc.so.6 No symbol table info available. eax 0x0 0 ecx 0xe9ff688c -369137524 edx 0x0 0 ebx 0x2 2 esp 0xe9ff6870 0xe9ff6870 ebp 0xe9ff688c 0xe9ff688c esi 0x8 8 edi 0x0 0 eip 0xf7f9d939 0xf7f9d939 <__kernel_vsyscall+9> eflags 0x286 [ PF SF IF ] cs 0x23 35 ss 0x2b 43 ds 0x2b 43 es 0x2b 43 fs 0x3 3 gs 0x63 99 k0 0x0 0 k1 0x0 0 k2 0x0 0 k3 0x0 0 k4 0x0 0 k5 0x0 0 k6 0x0 0 k7 0x0 0 From To Syms Read Shared Object Library 0xf7e92590 0xf7f528f5 Yes (*) /lib/i386-linux-gnu/libm.so.6 0xf7e84ad0 0xf7e85a84 Yes (*) /lib/i386-linux-gnu/libdl.so.2 0xf7df7914 0xf7e3dc78 Yes bin/libstdc++.so.6 0xf7d969d0 0xf7da620f Yes (*) /lib/i386-linux-gnu/libpthread.so.0 0xf7bce610 0xf7d1b386 Yes (*) /lib/i386-linux-gnu/libc.so.6 0xf7f9eab0 0xf7fb97fb Yes (*) /lib/ld-linux.so.2 0xf7ba1e00 0xf7bb10fc Yes bin/libgcc_s.so.1 0xf7b7b1d0 0xf7b8e44c Yes (*) bin/libtier0_srv.so 0xf7b68910 0xf7b6c3a4 Yes (*) /lib/i386-linux-gnu/librt.so.1 0xf7a6eec0 0xf7a9d944 Yes (*) bin/libvstdlib_srv.so 0xf7435460 0xf75a6a10 Yes (*) bin/dedicated_srv.so 0xf73bc4e0 0xf73cad3c Yes (*) bin/libsteam_api.so 0xf7290e50 0xf729cd44 Yes (*) /lib32/libtinfo.so.5 0xf676ddd0 0xf692d208 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/engine_srv.so 0xf6287d30 0xf62ad3b4 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/soundemittersystem_srv.so 0xf613b1c0 0xf620e324 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/materialsystem_srv.so 0xf3c84cf0 0xf3d2f608 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/studiorender_srv.so 0xf3ae4590 0xf3c03860 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/vphysics_srv.so 0xf3a47a20 0xf3a92ee8 Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/datacache_srv.so 0xf3a0d190 0xf3a21cbc Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/shaderapiempty_srv.so 0xf1831f20 0xf20e44e4 Yes /home/steam/.steam/steamcmd/hl2rp_server/hl2mp/bin/server_srv.so 0xf38a0630 0xf38b5a3c Yes (*) /home/steam/.steam/steamcmd/hl2rp_server/bin/scenefilecache.so 0xec2cc880 0xed4d2fb4 Yes (*) /home/steam/.steam/sdk32/steamclient.so 0xf3844bb0 0xf384b554 Yes (*) /lib/i386-linux-gnu/libnss_files.so.2 0xf3a01b70 0xf3a04ad4 Yes (*) /lib/i386-linux-gnu/libnss_dns.so.2 0xf382d660 0xf38394c4 Yes (*) /lib/i386-linux-gnu/libresolv.so.2 0xeb079440 0xebc547c0 Yes (*) bin/steamclient.so 0xeaa74170 0xeaa9c0a4 Yes (*) bin/crashhandler.so 0xe9acd920 0xe9ba1088 Yes /home/steam/.steam/steamcmd/hl2rp_server/bin/mysql_driver.so (*): Shared library is missing debugging information. Dump of assembler code for function __kernel_vsyscall: 0xf7f9d930 <+0>: push %ecx 0xf7f9d931 <+1>: push %edx 0xf7f9d932 <+2>: push %ebp 0xf7f9d933 <+3>: mov %esp,%ebp 0xf7f9d935 <+5>: sysenter 0xf7f9d937 <+7>: int $0x80 => 0xf7f9d939 <+9>: pop %ebp 0xf7f9d93a <+10>: pop %edx 0xf7f9d93b <+11>: pop %ecx 0xf7f9d93c <+12>: ret End of assembler dump. Stack level 0, frame at 0xe9ff6880: eip = 0xf7f9d939 in __kernel_vsyscall; saved eip = 0xf7be3832 called by frame at 0xe9ff69b0 Arglist at 0xe9ff6878, args: Locals at 0xe9ff6878, Previous frame's sp is 0xe9ff6880 Saved registers: ecx at 0xe9ff6878, edx at 0xe9ff6874, ebp at 0xe9ff6870, eip at 0xe9ff687c End of Source crash report
[1 Feb 2020 12:52]
Adrián G.
Update trailing dash typo in OS distro
[2 Feb 2020 18:56]
MySQL Verification Team
Thank you for the bug report. Please provide C file test case which allow to repeat the issue?. Attach it using the Files tab.
[8 Feb 2020 13:02]
Adrián G.
Hello again, It seems I fixed my issue. Apparently, it was due to the multithreading management. Since this was also causing an unrelated massive gameplay stuttering issue (which required me to add a small periodic sleep), and given that the architecture which involves the MySQL driver wrapper is quite complex, I prefered to concentrate to fix my issues and not attempt to make any test case which would probably be useless without an undeterministic multithreaded context. One of the possible causes regarding my original issue, though, is that I had set my separate thread to make every MySQL operations except mysql_close(), which was done on the main thread when I wanted to. Now everything is done in the I/O thread. Then, I have a final question. Given that my custom dynamically loaded MySQL wrapper library is dlclose()'d along the mysql_close() call in an appropiate destructor, do I also have to call mysql_library_end() right after it at that destructor? Or all resources including possible TCP/IP sockets would be closed automatically along the effective module close? Thanks in advance.
[10 Feb 2020 13:19]
MySQL Verification Team
Hi Mr. G. mysql_close() is enough, for every mysql MYSQL pointer that you are using ......
[10 Feb 2020 13:50]
Adrián G.
Hi Sinisa, Thanks for answering. I decided to call mysql_library_end() still when module is requested to close via dlclose(), for convenience. So I pressume bug wasn't mysql_client library fault. Best regards.
[10 Feb 2020 14:46]
MySQL Verification Team
Best regards from me, Mr. G. !!!!!!
