Hi, I can reliably crash the Xorg server by calling xine under KDE4 with desktop effects enabled. If Xorg was called via startx I am left with a black screen with an immobile mouse cursor and and an immobile fb cursor in the upper left. If it was called via kdm I am thrown back to the logon screen (fully functional). Afterwards the system will crash more easily, eg. when bringing up the kickoff launcher. Usually there are no traces in the logs. Only once did I capture the following in /var/log/syslog: kernel: [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 24 of 128, total 17690624 bytes: -9 kernel: [drm:i915_gem_do_execbuffer] *ERROR* 1409 objects [28 pinned], 97095680 object bytes [30773248 pinned], 89694208/234881024 gtt bytes kdm[2251]: X server for display :0 terminated unexpectedly Xorg.log is inconclusive. The crash can be avoided by switching off KDE4 desktop effects (which is a pity as they work really well if it wasn't for the crash). My system is a shiny new i3 530, H55 chipset, kernel 2.6.33.3, Xorg 1.7.6, intel_drv.so 2.11.0. Note: I do compile kernels, but I have no build environment for Xorg. Thanks, Martin
I'm adding the lspci info here and I'll attach a recent Xorg.log (note: regular session, no crash). 00:00.0 Host bridge [0600]: Intel Corporation Core Processor DRAM Controller [8086:0040] (rev 12) Subsystem: ASUSTeK Computer Inc. Device [1043:8383] Flags: bus master, fast devsel, latency 0 Capabilities: [e0] Vendor Specific Information <?> Kernel driver in use: agpgart-intel 00:02.0 VGA compatible controller [0300]: Intel Corporation Core Processor Integrated Graphics Controller [8086:0042] (rev 12) (prog-if 00 [VGA controller]) Subsystem: ASUSTeK Computer Inc. Device [1043:8383] Flags: bus master, fast devsel, latency 0, IRQ 16 Memory at fb800000 (64-bit, non-prefetchable) [size=4M] Memory at d0000000 (64-bit, prefetchable) [size=256M] I/O ports at cc00 [size=8] Expansion ROM at <unassigned> [disabled] Capabilities: [90] MSI: Enable- Count=1/1 Maskable- 64bit- Capabilities: [d0] Power Management version 2 Capabilities: [a4] PCI Advanced Features Kernel driver in use: i915
Created attachment 35652 [details] Xorg.log (regular session w/o crash)
I have just upgraded to 2.6.33.4 which contains a few i915 updates. Unfortunately I got a crash again after enabling desktop effects. There is nothing in the logs other than the line kdm[1348]: X server for display :0 terminated unexpectedly in /var/log/syslog. ---- Btw, I made another observation. This is just for info and not part of this ticket. The bug logged in this ticket is triggered when running the system within specification. I have observed that my Xorg/intel_drv is very fragile when overclocking. This could be halfway expected since the integrated graphics is interwoven with CPU, northbridge, DRAM and the system of BCLK, QPI etc. Odd is only the fact that a substantial overclock of BCLK works when stress testing on the-other-OS, whereas it locks the screen under Xorg/intel_drv. The symptoms are different from the bug logged in this ticket in the sense that the X server starts with black or corrupted screen, and that there are errors and warnings in Xorg.log.
I have just upgraded the kernel to 2.6.34. However, the instability persists: 30 seconds after activating KDE desktop effects the session crashes back to the kdm greeter.
After the latest distro upgrade I now have Xorg 1.7.7 and intel_drv.so 2.11.0 running on kernel 2.6.34. The instability when switching on KDE desktop effects persists. I did not catch any syslog entries relating to the crash.
I have noticed a few i915 drm patches in kernel 2.6.35-rc2, so I tried that one. However, the instability still occurs. No further hint in the logs (other than: X server died unexpectedly).
I have re-classified the ticket as DRI since it appears kernel related to me.
*** Bug 27911 has been marked as a duplicate of this bug. ***
update: it appears, even without KDE4 desktop effects I can reliably freeze the display simply by maximizing flash videos (found that out thanks to the football world cup). Although I am not sure it is the same bug since the mouse pointer still moves and I hve to telnet in and kill X (no automatic crash back to kdm greeter). software: firefox 3.6.3, flash plugin 10.0 r45
OK, regarding the last comment: it appears that one is a different bug, and it is already covered by ticket no. 26937.
This is in connection with https://bugs.freedesktop.org/show_bug.cgi?id=27911 I just updated X11 packages (to the best of my knowledge only libdrm 2.4.20 to 2.4.21 was major change; though I can be wrong) and everything seem to work fine, at least for cases where it was crashing earlier. Can someone verify this?
(In reply to comment #11) > This is in connection with https://bugs.freedesktop.org/show_bug.cgi?id=27911 > > I just updated X11 packages (to the best of my knowledge only libdrm 2.4.20 to > 2.4.21 was major change; though I can be wrong) and everything seem to work > fine, at least for cases where it was crashing earlier. Can someone verify > this? Good find. I compiled libdrm-2.4.21 from source (or rather, created a Slackware package from source) and installed it over 2.4.20. It needs more testing, but my initial impression is that the crash when starting xine wit KDE4 desktop effects enabled is indeed gone. That is the good news. I have had a few X crashes, though, that might be related to the flash plugin (and thus belong to other bug reports more than this on). I am currently testing. Martin
Hi, me again. After a few days' testing I can confirm that the bug reported here (a certain type of freeze when KDE desktop effects are enabled) is gone when switching to libdrm-2.4.21. However, a different type of freeze is still present, usually associated with firefox and/or flash player, irrespective of KDE desktop effects. Therefore I'd be OK with closing this bug report and moving the firefox/flash related crashes over to bug no. 26937. Thanks, Martin
Optimistically closing then, thanks for the update.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.