Description of problem: X crashed on me with this traceback: [ 189.149] 0: Xorg (xorg_backtrace+0x4a) [0x80aba9a] [ 189.149] 1: Xorg (0x8048000+0x68a18) [0x80b0a18] [ 189.149] 2: (vdso) (__kernel_rt_sigreturn+0x0) [0xb9140c] [ 189.149] 3: /usr/lib/xorg/modules/drivers/qxl_drv.so (0x904000+0x8b42) [0x90cb42] [ 189.150] 4: /usr/lib/xorg/modules/drivers/qxl_drv.so (0x904000+0x596a) [0x90996a] [ 189.150] 5: /usr/lib/xorg/modules/drivers/qxl_drv.so (0x904000+0x13dbe) [0x917dbe] [ 189.150] 6: Xorg (CompositeRects+0x79) [0x8147df9] [ 189.150] 7: Xorg (0x8048000+0x1045c4) [0x814c5c4] [ 189.150] 8: Xorg (0x8048000+0x1003bf) [0x81483bf] [ 189.151] 9: Xorg (0x8048000+0x2f2fd) [0x80772fd] [ 189.151] 10: Xorg (0x8048000+0x1cc05) [0x8064c05] [ 189.151] 11: /lib/libc.so.6 (__libc_start_main+0xf5) [0x1ce785] [ 189.151] 12: Xorg (0x8048000+0x1cf49) [0x8064f49] [ 189.152] [ 189.152] Segmentation fault at address 0xccccccd0 [ 189.152] Fatal server error: [ 189.152] Caught signal 11 (Segmentation fault). Server aborting I have seen this several times in a VM using Spice+qxl. I know it happens when switching ttys (graphical and text). But it does not happen often and I can't surely reproduce it. I haven't seen this problem in an installed VM (but it's true I don't switch ttys much there). Version-Release number of selected component (if applicable): F17 Beta TC2 it seems that xorg-x11-drv-qxl-0.0.22-0.fc17 is inside How reproducible: seen several times, don't know how to trigger Steps to Reproduce: 1. switch from tty7 to tty1 and back (possibly many times) using a spice VM
Created attachment 571399 [details] anaconda.log
Created attachment 571400 [details] program.log
Created attachment 571401 [details] syslog
Created attachment 571402 [details] X.log
I'm not really sure we want to have this as a blocker, but I'm proposing it anyway so that we know about this and can discuss it.
There is one more thing I need to mention. Before I started switching ttys, anaconda seemed to be stuck. I clicked Next on "install or upgrade" screen and nothing happened for a minute or two. Then I started switching ttys and it crashed. I think I vaguely remember that the other occasions where I saw it crashed was also anaconda stuck (or X stuck?). But it might not be the cause of the issue, but I rather mention it.
I've had a todo list item to file this forever, so thanks for doing this. I don't believe it's at all anaconda-specific: it just seems to be the case that X is very prone to crashing when you switch to a VT in a Spice/QXL VM. I've seen this a lot. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
I'm -1 on this as a blocker, though, it's not really serious enough to hit any criteria and it's specific to VMs. It was present in F16.
-1 blocker.
-1 blocker, ambivalent about NTH
Three -1s, this is rejected as a Beta blocker. It might be plausible to raise it as a final blocker, you can get in some nasty situations in VMs with it (like X respawning continuously so you can barely use the VT...), but it's too restricted in scope and impact to really be a Beta blocker. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
proposing as beta NTH, as the issue has been raised. Since this has an impact on lives, I'm +1 on NTH.
-1 to blocker -1 to NTH.
-1 blocker +0 NTH. Unless the fix is very simple, I think waiting for an update is better.
Has anyone verified that this bug still exists with the 0.0.22 version? It could be fixed with the rebase to a new upstream version.
I hadn't checked, but I will now. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
just tried a dozen VT switches, no crash. looks promising. like the virt-manager hang bug, though, i'd like to leave it open for observation...
(In reply to comment #15) > Has anyone verified that this bug still exists with the 0.0.22 version? > > It could be fixed with the rebase to a new upstream version. See the bug description, F17 Beta TC2 crashes for sure. I don't know how to check inside anaconda, but qxl driver 0.0.22 seems to be present in the package repo, so I assume it was used.
ah, I didn't notice the report was from TC2. yeah, TC2 had 0.0.22. so the bug's still present :/ -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
Discussed at 2012-04-06 NTH review meeting: http://meetbot.fedoraproject.org/fedora-bugzappers/2012-04-06/f17-beta-blocker-review-5.2012-04-06-17.01.html . Accepted as NTH as it's an annoying problem in testing anaconda and live images in a VM. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
Beta is out, moving to F17 Final NTH
Discussed at 2012-05-04 NTH review meeting. Accepted as NTH as it's an annoying bug that would get 'locked in' on live images if we don't fix it before release, and we'd have another six months of lives that crash when you switch VTs in a VM. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
I don't think this was ever addressed in F17, and given that end-of-life is approaching I don't think it ever will be fixed. I haven't seen this issue on F18+, but if someone can reproduce on something newer, please reopen.
yeah, I think it still happens in F17 when I have to use it to check something, but it seems to have been fixed since F18.