Date: Wed, 11 May 2005 11:49:47 GMT From: Steven Jurczyk <steve@post.pl> To: freebsd-gnats-submit@FreeBSD.org Subject: threads/80887: ULE with SMP broke libpthread/libthr on 5.4-RELEASE Message-ID: <200505111149.j4BBnlG7046863@www.freebsd.org> Resent-Message-ID: <200505111150.j4BBo2CE061130@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 80887 >Category: threads >Synopsis: ULE with SMP broke libpthread/libthr on 5.4-RELEASE >Confidential: no >Severity: critical >Priority: high >Responsible: freebsd-threads >State: open >Quarter: >Keywords: >Date-Required: >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Wed May 11 11:50:02 GMT 2005 >Closed-Date: >Last-Modified: >Originator: Steven Jurczyk >Release: 5.4-RELEASE >Organization: >Environment: FreeBSD meritum2.home.net.pl 5.4-RELEASE FreeBSD 5.4-RELEASE #17: Wed May 11 11:11:01 CEST 2005 steve@meritum2.home.net.pl:/data/src/bsd/sys/i386/compile/HOME i386 >Description: Using ULE on SMP machine broke applications which use libpthread/libthr, probably at context switching between threads... The applications simply hang and they aren't killable (kill -9 <pid> don't work).. Usually this also break kernel shutdown (can't flush some inodes or blocks)... >How-To-Repeat: Compile kernel with options: options SMP options SCHED_ULE and try to start mysqld_server on dual dual processor (may be on Pentium IV with enabled HT) The problem also is repeatable at 50% success rate with simple /usr/bin/host util. Setting setenv LIBPTHREAD_SYSTEM_SCOPE usually helps to hang any pthreads application at their initialization. >Fix: sorry :) >Release-Note: >Audit-Trail: >Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200505111149.j4BBnlG7046863>