Bug 54195

Summary: [sna] crashes sometimes with two KDE sessions
Product: xorg Reporter: Martin Steigerwald <Martin>
Component: Driver/intelAssignee: Chris Wilson <chris>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: 7.7 (2012.06)   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Xorg log with one of these crashes none

Description Martin Steigerwald 2012-08-29 09:00:47 UTC
X.org sometimes crashes when using two KDE sessions together with SNA.

[ 12614.049] (II) intel(0): switch to mode 1920x1080 on crtc 3 (pipe 0)
[ 12614.054] (II) intel(0): EDID vendor "LEN", prod id 16562
[ 12614.054] (II) intel(0): Printing DDC gathered Modelines:
[ 12614.054] (II) intel(0): Modeline "1920x1080"x0.0  139.00  1920 1980 2028 2050  1080 1090 1100 1130 -hsync -vsync (67.8 kHz eP)
[ 12614.054] (II) intel(0): Modeline "1920x1080"x0.0  115.83  1920 1980 2028 2050  1080 1090 1100 1130 -hsync -vsync (56.5 kHz e)
[ 12614.151] (--) synaptics: SynPS/2 Synaptics TouchPad: touchpad found
[ 13064.857] 
[ 13064.857] Backtrace:
[ 13064.867] 0: /usr/bin/X (xorg_backtrace+0x36) [0x7f483e222ca6]
[ 13064.867] 1: /usr/bin/X (0x7f483e0a4000+0x1827f9) [0x7f483e2267f9]
[ 13064.867] 2: /lib/x86_64-linux-gnu/libpthread.so.0 (0x7f483d3cc000+0xf030) [0x7f483d3db030]
[ 13064.867] 3: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f4839fce000+0x5f3de) [0x7f483a02d3de]
[ 13064.867] 4: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f4839fce000+0xb931a) [0x7f483a08731a]
[ 13064.867] 5: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f4839fce000+0x61d6e) [0x7f483a02fd6e]
[ 13064.867] 6: /usr/bin/X (0x7f483e0a4000+0x110a49) [0x7f483e1b4a49]
[ 13064.867] 7: /usr/bin/X (0x7f483e0a4000+0x1099a4) [0x7f483e1ad9a4]
[ 13064.867] 8: /usr/bin/X (0x7f483e0a4000+0x52e31) [0x7f483e0f6e31]
[ 13064.867] 9: /usr/bin/X (0x7f483e0a4000+0x41ec5) [0x7f483e0e5ec5]
[ 13064.867] 10: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xfd) [0x7f483c0f6ead]
[ 13064.867] 11: /usr/bin/X (0x7f483e0a4000+0x4219d) [0x7f483e0e619d]
[ 13064.867] 
[ 13064.867] Segmentation fault at address 0x7a
[ 13064.867] 
Fatal server error:
[ 13064.867] Caught signal 11 (Segmentation fault). Server aborting
[ 13064.867] 
[ 13064.867] 


My X.org config is as follows:


Section "Device"
        Identifier      "Grafikkern"

        Driver          "intel"

        # Monitore
        Option          "Monitor-LVDS1" "Notebook-Display"

        # Optionen
        Option          "AccelMethod"   "sna"

EndSection

Section "Monitor"
        Identifier      "Notebook-Display"

        DisplaySize     340 192
EndSection


I am using:


Linux version 3.5.2-tp520 (martin@merkaba) (gcc version 4.7.1 (Debian 4.7.1-7) ) #1 SMP PREEMPT Sun Aug 19 12:39:04 CEST 2012

(compiled from Greg KH git)

and


libdrm-dev/sid uptodate 2.4.33-3
libdrm-intel1/sid uptodate 2.4.33-3
libdrm-nouveau1a/sid uptodate 2.4.33-3
libdrm-radeon1/sid uptodate 2.4.33-3
libdrm2/sid uptodate 2.4.33-3
libgl1-mesa-dri/sid uptodate 8.0.4-2
xserver-xorg/sid uptodate 1:7.7+1
xserver-xorg-core/sid uptodate 2:1.12.3.902-1
xserver-xorg-input-all/sid uptodate 1:7.7+1
xserver-xorg-input-evdev/sid uptodate 1:2.7.0-1+b1
xserver-xorg-input-synaptics/sid uptodate 1.6.2-1
xserver-xorg-video-intel/experimental uptodate 2:2.20.5-1
xserver-xorg-video-intel-dbg/experimental uptodate 2:2.20.5-1

on Debian Sid.

I will attached X.org log.

I will switch off SNA to see whether it makes a difference.

I could try to get a more complete backtrace, but I´d like to get a hint how to. I think I´d need to attach GDB to running X.org process then.

Thanks,
Martin
Comment 1 Chris Wilson 2012-08-29 09:06:34 UTC
To get the backtrace, install the debug symbols (xserver-xorg-video-intel-dbg) and then call "addr2line -e /usr/lib/xorg/modules/drivers/intel_drv.so 0x5f3de 0xb931a 0x61d6e"

However, I recognise that trace:

commit 8e10a5b348a37feadcf935ec7694e46cc0802bdf
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sun Aug 26 14:53:12 2012 +0100

    sna/gen6+: Do not call sna_blt_composite() after prepping the composite op
    
    As sna_blt_composite() will overwrite parts of the composite op as it
    checks whether or not it can execute that operation, it will lead to a
    crash as the normal render path finds the op corrupt. (The BLT
    conversion functions cater for the cases where we may wish to switch
    pipelines after choosing src/dst bo.)
    
    Reported-by: rei4dan@gmail.com
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Comment 2 Martin Steigerwald 2012-08-29 09:06:47 UTC
Created attachment 66267 [details]
Xorg log with one of these crashes
Comment 3 Martin Steigerwald 2012-08-29 09:07:40 UTC
Wow! That fast fast ;).

Thanks,
Martin
Comment 4 Chris Wilson 2012-08-29 09:10:10 UTC
Or too slow. The first report of that crash arrived just after I made the 2.20.5 release, and now I will have to do another quick release -- just as soon as I can be sure that there aren't any other major surprises.
Comment 5 Martin Steigerwald 2012-08-29 12:16:50 UTC
Hmmm, I reported this as

xserver-xorg-video-intel: [sna] crashes sometimes with two KDE sessions
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=686161

for Debian so that the Debian X.org team is aware of the issue.

AFAIK for Debian Wheezy SNA will not (yet) be enabled by default anyway.

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.