Description of problem: Xorg crashes, session closed erratically Version-Release number of selected component (if applicable): FC5 in Xen Domain0 xorg-x11-server-Xorg-1.0.1-9 xen-3.0.1-4 kernel-2.6.15-1.2054_FC5xen0 01:00.0 VGA compatible controller: Matrox Graphics, Inc. G400/G450 (rev 82) Running in Xinerama mode with 2 screens How reproducible: Erratically happened twice since I reinstalled my workstation under FC5, 3 days ago Steps to Reproduce: 1. work 2. 3. Actual results: Xorg crashes after a while Expected results: Xorg does not crash Additional info: -------------------------------------------- bottom of Xorg.0.log.old hints it's a segfault in GCCopy SetGrabKeysState - disabled SetGrabKeysState - enabled *** If unresolved symbols were reported above, they might not *** be the reason for the server aborting. Backtrace: 0: /usr/bin/Xorg(xf86SigHandler+0x87) [0x80b86c7] 1: [0x93d420] 2: /usr/bin/Xorg(CopyGC+0x3f8) [0x8098508] 3: [0x86d85bf] Fatal server error: Caught signal 11. Server aborting (II) Screen 0 shares mem & io resources (II) Screen 1 shares mem & io resources ---------------------------------------- xorg.conf attached
Created attachment 126547 [details] /etc/X11/xorg.conf
Using the following driver I guess: xorg-x11-drv-mga-1.2.1.3-1.2 Daniel
Crashed again after less than 6 hours of work. I actually reading a web page at that point, no mouse or keyboard involved at that precise point in time. Different backtrace: SetGrabKeysState - disabled SetGrabKeysState - enabled *** If unresolved symbols were reported above, they might not *** be the reason for the server aborting. Backtrace: 0: /usr/bin/Xorg(xf86SigHandler+0x87) [0x80b86c7] 1: [0x764420] 2: /usr/bin/Xorg(AllocatePixmap+0x22) [0x8070682] 3: /usr/lib/xorg/modules/libfb.so(fbCreatePixmapBpp+0x82) [0xdbc682] 4: /usr/lib/xorg/modules/libfb.so(fbCreatePixmap+0x52) [0xdbc772] 5: /usr/lib/xorg/modules/libxaa.so [0x36bbf7] 6: /usr/bin/Xorg(ProcCreatePixmap+0x126) [0x8083556] 7: /usr/bin/Xorg(PanoramiXCreatePixmap+0x146) [0x8135676] 8: /usr/bin/Xorg(Dispatch+0x19b) [0x80888eb] 9: /usr/bin/Xorg(main+0x487) [0x80701d7] 10: /lib/i686/nosegneg/libc.so.6(__libc_start_main+0xdc) [0x1167e4] 11: /usr/bin/Xorg(FontFileCompleteXLFD+0xb1) [0x806f511] Fatal server error: Caught signal 4. Server aborting (II) Screen 0 shares mem & io resources (II) Screen 1 shares mem & io resources
The xorg.conf used is the exact same one as what I used previously on RHEL4 and which never ever crashed... Daniel
Please attach the full X server log as an attachment. Also, please try booting with the non-Xen kernel and reproducing. As I mentioned in IRC last night, there are known issues with X running on top of Xen, of which some appear to be kernel bugs so far. It'd be nice to rule Xen out of the equation first if possible. Thanks in advance.
Created attachment 126734 [details] full log from the second crash reported
Well X didn't crashed again yet, I will reboot on non-Xen when this happens, note that don't have a guest system running ATM, while this was the case when X crashed previously... this seems to hint at a Xen interraction. Daniel
Okay Xen crashed again after 5 days when running on Xen in Domain 0 but while avoiding any other DomU domain. The bug still exist, but just way harder to trigger apparently. I now rebooted the same kernel version but without Xen I will report if it breaks without Xen or not within a week. Attached is the Xorg log when booted from the non-Xen kernel. Daniel
Created attachment 126937 [details] the Xorg log when booted from the non-Xen kernel
Okay I have been running for 6 days now with the normal non Xen kernel and didn't got any trouble. So I believe the Xorg crash is really related to Xen :-\ Daniel
I'm running into an X related crash at boot time with my Dom0. I saw a workaround relating to FC4 and libint10.a. The thing is, for FC5 this library is now a .so /usr/lib/xorg/modules/libint10.so and the workaround of simply moving it to the side no longer works. X failed to launch when I moved that file aside. Chaning /etc/inittab to default to runlevel 3 is a partial workaround but now I have no X. Changing to runlevel 5 via the normal methods causes a hard crash. See 140847 for more info on the FC4 workaround. I'd be happy enough if someone could provide an FC5 workaround to disable the X probing. # lspci | grep -i vga 06:05.0 VGA compatible controller: XGI - Xabre Graphics Inc Volari Z7 Thanks.
I'm getting Xorg crashing several times a day on FC5 with a Xen0 kernel also, this is the last timg in the log: Fatal server error: Caught signal 11. Server aborting I'll attach the whole log and the config file.
Created attachment 128780 [details] Xorg log file
Created attachment 128781 [details] Xorg configuration file
Oh, I forgot to mention, I'm running with these packages: xorg-x11-server-Xorg-1.0.1-9.fc5 xorg-x11-drv-nv-1.0.1.5-3.1 kernel-xen0-2.6.16-1.2096_FC5 xen-3.0.1-4 And selinux=0 in /etc/grub.conf
Please report separate bugs if you are experience problems under Xen, even if it appears they are the exact same problem as Daniel has reported here. This allows us to track each issue individually, without each bug being cluttered by information which may or may not be part of the the same issue. Later on if we determine they are the same issue, we can always close them as duplicates of each other. Thanks.
I don't see the problem anymore with FC6 or rawhide, Daniel