Hide Forgot
Description of problem: X segufault intermittently while logout. When this symptom happen, we can not type user name or password any more. All we can do is restart X, or restart computer. Version-Release number of selected component (if applicable): xorg-x11-server-Xorg-1.1.1-48.76.el5_5.2 How reproducible: Intermittently. This symptom happen once in 20-30 logout. Steps to Reproduce: 1. Install OS with GNOME. 2. Install following additional RPMs. kdebase-3.5.4-24.el5 Wed 29 Jun 2011 07:40:45 PM PDT htdig-3.2.0b6-11.el5 Wed 29 Jun 2011 07:40:15 PM PDT lm_sensors-2.10.7-9.el5 Wed 29 Jun 2011 07:40:13 PM PDT kdelibs-3.5.4-25.el5_4.1 Wed 29 Jun 2011 07:39:59 PM PDT arts-1.5.4-1 Wed 29 Jun 2011 07:39:29 PM PDT qt-3.3.6-23.el5 Wed 29 Jun 2011 07:39:25 PM PDT libmng-1.0.9-5.1 Wed 29 Jun 2011 07:39:23 PM PDT lcms-1.18-0.1.beta1.el5_3.2 Wed 29 Jun 2011 07:39:21 PM PDT libXcomposite-0.3-5.1 Wed 29 Jun 2011 07:39:08 PM PDT 3. Edit /etc/sysconfig/desktop file DISPLAYMANAGER="KDE" DESKTOP="GNOME" 4. login and logout from X environment. Actual results: X server segfault and logged following error while logout the user. X[5987]: segfault at ffffffffc4000f10 rip 000000000048b252 rsp 00007fff35efc3b0 error 4 No core file is created even though I set ulimit -c unlimited in /etc/profile. No backtrace is written in /var/log/xorg.0.log. Only 2 lines (backtrace) are written in the log, but no additional lines. Expected results: X keep running even after logout users. Additional info: Workaround: Enable TerminateServer=true in /usr/share/config/kdm/kdmrc
Created attachment 513952 [details] strace output while X got segfault (See line 70-80. )
This symptom can be reproduced on bare metal system, as well as RHEL5 running on VMware Player.
Created attachment 513979 [details] Capture segfault using gdb Run X on gdb and try to capture the backtrace of the point where segfault happen. I have installed glibc-debuginfo and xorg-x11-server debuginfo RPMs.
Additional information. When this symptom happen, following messages are logged in /var/log/messages. rhel5 kdm: :0[14741]: Hung in XOpenDisplay(:0), aborting rhel5 kdm: :0[14741]: Cannot connect to :0, giving up rhel5 kdm[8167]: Display :0 cannot be opened
Created attachment 514146 [details] Capture segfault using gdb with all debuginfo RPMS
Same test on RHEL5.7(64bit), the same symptom still exist. The backtrace of the X process segfault. #0 0x00000033c78708df in malloc_consolidate (av=0x33c7b539e0) at malloc.c:5083 #1 0x00000033c787239c in _int_free (av=0x33c7b539e0, p=0x14e892c0, have_lock=0) at malloc.c:4962 #2 0x00000033c78728bb in __libc_free (mem=<value optimized out>) at malloc.c:3670 #3 0x00002b856337b19f in XAADestroyInfoRec (infoRec=0x14e86120) at xaaInit.c:93 #4 0x00002b85624fc37c in vmwareXAACloseScreen () from /usr/lib64/xorg/modules/drivers/vmware_drv.so #5 0x00002b85624facd5 in ?? () from /usr/lib64/xorg/modules/drivers/vmware_drv.so #6 0x00000000004f108f in CursorCloseScreen (index=0, pScreen=0x14e84830) at cursor.c:173 #7 0x0000000000514eee in AnimCurCloseScreen (index=-944424480, pScreen=0x1) at animcur.c:130 #8 0x00000000004337de in main (argc=6, argv=0x7fff8d09e9d8, envp=<value optimized out>) at main.c:472 similar backtrace, (hit segfault free => malloc_consolidate).
http://people.redhat.com/tumeya/.sf00518164_x_hang/ Masanari, I haven't been able to reproduce this. While trying this, is it possible for you to try the test packages above? Thank you for your cooperation.
Umeya-san Thanks for update. I have downloaded the RPMs and started to test. I will provide the feed back of the logout/login test result on early next week.
*** This bug has been marked as a duplicate of bug 674741 ***