Re: Coredump on quickly disconnecting clients
On Sunday 25 March 2001 18:57, Voytek Lapinski wrote: On Sunday 25 March 2001 16:27, you wrote: Found the problem with your signal 11 with strange traces from inside LinuxThreads. This was a hard concurrency bug - the problem was that a memory reference passed to pthread_create was getting invalidated by the thread that pthread_create was launching before pthread_create would right to that value. Took me a while to figure this out - I have scrutinized the entire thread creation mechanism in LinuxThreads suspecting a bad mmap or munmap somewhere before I realized the problem had nothing to do with LinuxThreads internal code. I believe the problem has existed for as long as MySQL has been around - back then the machines were just not fast enough to make this condition happen frequently enough for someone to be able to create a repeatable case. I have fixed it by extending the critical region. Here is the patch: Fantastic! I take it you won't be wanting the better back trace then? As long as you stop getting them :-) -- MySQL Development Team __ ___ ___ __ / |/ /_ __/ __/ __ \/ / Sasha Pachev [EMAIL PROTECTED] / /|_/ / // /\ \/ /_/ / /__ MySQL AB, http://www.mysql.com/ /_/ /_/\_, /___/\___\_\___/ Provo, Utah, USA ___/ - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: Coredump on quickly disconnecting clients
On Sunday 25 March 2001 16:27, you wrote: Found the problem with your signal 11 with strange traces from inside LinuxThreads. This was a hard concurrency bug - the problem was that a memory reference passed to pthread_create was getting invalidated by the thread that pthread_create was launching before pthread_create would right to that value. Took me a while to figure this out - I have scrutinized the entire thread creation mechanism in LinuxThreads suspecting a bad mmap or munmap somewhere before I realized the problem had nothing to do with LinuxThreads internal code. I believe the problem has existed for as long as MySQL has been around - back then the machines were just not fast enough to make this condition happen frequently enough for someone to be able to create a repeatable case. I have fixed it by extending the critical region. Here is the patch: Fantastic! I take it you won't be wanting the better back trace then? I'll get back to you once I've confirmed that it's fixed it in our case. Thanks a lot for the help. You guys rock. -- Voytek Lapinski [EMAIL PROTECTED] Software Engineer Hotkey Internet ph: 03 9923 3656 mob: 0427 469 891 fax: 03 9923 3388 www.hotkey.net.au - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
