| Summary: | X segfaults when logging out from kde4 kdm (intel SNA accel). | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | xorg | Reporter: | Pawel Sikora <pawel_sikora> | ||||||||
| Component: | Driver/intel | Assignee: | Chris Wilson <chris> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||
| Severity: | normal | ||||||||||
| Priority: | medium | CC: | arekm | ||||||||
| Version: | unspecified | ||||||||||
| Hardware: | x86-64 (AMD64) | ||||||||||
| OS: | Linux (All) | ||||||||||
| Whiteboard: | |||||||||||
| i915 platform: | i915 features: | ||||||||||
| Attachments: |
|
||||||||||
|
Description
Pawel Sikora
2012-07-19 07:29:38 UTC
commit 7a3b98e05b706548527e73b2008600391c601a62 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Thu Jul 19 08:42:05 2012 +0100 sna: Re-register the SHM funcs every server generation As the SHM layer hooks into the CloseScreen chain to free its privates, we then need to call the registration function again on the next generation to ensure that the private is reallocated before use. Reported-by: Pawel Sikora <pluto@agmk.net> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52255 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Argh. Sorry you updated your email address after I already pushed the commit. :( (In reply to comment #1) > commit 7a3b98e05b706548527e73b2008600391c601a62 > Author: Chris Wilson <chris@chris-wilson.co.uk> > Date: Thu Jul 19 08:42:05 2012 +0100 > > sna: Re-register the SHM funcs every server generation will you plan a quick 2.20.1 bugfix release? I'm planning a fix of the brown paper bag variety this w/e. I'm still gathering the bags... 2.20.0 with 7a3b98e.. patch crashes in different way: Program received signal SIGSEGV, Segmentation fault. 0x00000000023da310 in ?? () #0 0x00000000023da310 in ?? () #1 0x00000000004e2d5b in ChangePicture (pPicture=pPicture@entry=0x26a9290, vmask=0, vmask@entry=4096, vlist=0x7fff0886d5d0, vlist@entry=0x7fff0886d5cc, ulist=ulist@entry=0x0, client=client@entry=0x233fb80) at picture.c:1193 #2 0x00000000004e34ee in CreatePicture (pid=pid@entry=0, pDrawable=0x2cd8e20, pFormat=pFormat@entry=0x26a6dc8, vmask=vmask@entry=4096, vlist=vlist@entry=0x7fff0886d5cc, client=0x233fb80, error=error@entry=0x7fff0886d5c8) at picture.c:763 #3 0x00007fa57127ec73 in sna_glyphs_create (sna=sna@entry=0x7fa5749ad010) at sna_glyphs.c:222 #4 0x00007fa571259481 in sna_accel_create (screen=screen@entry=0x2c99350, sna=sna@entry=0x7fa5749ad010) at sna_accel.c:12963 #5 0x00007fa57127caa2 in sna_create_screen_resources (screen=0x2c99350) at sna_driver.c:171 #6 0x000000000049e6c7 in xf86CrtcCreateScreenResources (screen=0x2c99350) at xf86Crtc.c:704 #7 0x000000000042320c in main (argc=5, argv=0x7fff0886d7b8, envp=<optimized out>) at main.c:215 Too hasty. I thought I understood the problem.... Can you please attach an Xorg.log for failure with --enable-debug=full? Created attachment 64388 [details]
xorg.log with --enable-debug=full
(In reply to comment #5) > Too hasty. I thought I understood the problem.... > > Can you please attach an Xorg.log for failure with --enable-debug=full? this is weird, with --enable-debug=full i can't login into kde session. after entering u/p kdm goes back to login prompt. xorg.log attached. Hmm, odd. There is no clue there as to why KDE shutdown, as the Xorg closed normally. Anything in kdm.log or .xsession-errors? Nothing I can do until I can either reproduce this locally, or I have a debug log showing the sequence triggering the crash. :| Hmm, perhaps a valgrind run would do the trick. Compile with --enable-debug to make the code valgrind clean. Is there anything unusual in your xorg.conf, the list of modules being loaded perhaps? Any peculiarities in how it is compiled? Anything that might give a clue. Created attachment 64402 [details]
xsession log for 2.20.0 with 7a3b98e git patch and enable-debug=full.
Created attachment 64403 [details]
xsession log for pure 2.20.0 without any debug enabled.
(In reply to comment #9) > Nothing I can do until I can either reproduce this locally, i can reproduce this problem (crash on kdm logout) on two different machines: 1). on classic workstation with intel-i3 cpu with integrated gpu and dual lcd display. 2). on asus laptop with intel-i7 cpu with integrated gpu and additional (unused) nvidia-optimus accelerator. > Hmm, perhaps a valgrind run would do the trick. Compile with --enable-debug to > make the code valgrind clean. any hints how to run xserver under valgrind? > Is there anything unusual in your xorg.conf, i haven't an explicit xorg.conf on my laptop, xserver works fine with auto-detected parts. > Anything that might give a clue. i've compared xorg.log, kdm.log and daemon log w/o any obvious diffs but the xsession log shows some diffs... for pure 2.20.0 crashing driver i see sucessfull XSyncBasedPoller attempts: (...) kded(8816) XSyncBasedPoller::XSyncBasedPoller: 3 1 kded(8816) XSyncBasedPoller::XSyncBasedPoller: XSync seems available and ready kded(8816) XSyncBasedPoller::setUpPoller: XSync Inited kded(8816) XSyncBasedPoller::setUpPoller: Supported, init completed (...) akonadi_nepomuk_feeder(8886) XSyncBasedPoller::XSyncBasedPoller: 3 1 akonadi_nepomuk_feeder(8886) XSyncBasedPoller::XSyncBasedPoller: XSync seems available and ready akonadi_nepomuk_feeder(8886) XSyncBasedPoller::setUpPoller: XSync Inited akonadi_nepomuk_feeder(8886) XSyncBasedPoller::setUpPoller: Supported, init completed (...) and for patched 2.20.0 with full debug enabled: (...) kded(6796) XSyncBasedPoller::XSyncBasedPoller: 3 1 kded(6796) XSyncBasedPoller::XSyncBasedPoller: XSync seems available and ready kded(6796) XSyncBasedPoller::setUpPoller: XSync Inited kded(6796) XSyncBasedPoller::setUpPoller: Supported, init completed (...) akonadi_nepomuk_feeder(6865) XSyncBasedPoller::XSyncBasedPoller: 3 1 kded4: Fatal IO error: client killed (...) after this 'client killed' kde goes down and returns into kdm login prompt. In kdmrc there is ServerCmd=/usr/bin/X That would look to be the best place to start X under valgrind, try ServerCmd=valgrind --trace-children=yes --track-origins=yes /usr/bin/X Is there anything in /var/log/kdm.log relating to the errors whilst starting under debug=full? The attached Xorg.log is a normal shutdown, yet the .xsession-errors suggests that it was rather unexpected. i can't start xorg under valgrind in this way but i think this is not needed at this moment beacuse the intel driver crashes on simple null pointer deref. current sources (2.20.2-24-gfd3a123) crashes on kdm logout: (gdb) bt Program received signal SIGSEGV, Segmentation fault. ShmDestroyPixmap (pPixmap=0xd0ac80) at shm.c:272 272 pScreen->DestroyPixmap = screen_priv->destroyPixmap; #0 ShmDestroyPixmap (pPixmap=0xd0ac80) at shm.c:272 #1 0x00007f973c735a48 in sna_glyphs_create (sna=sna@entry=0x7f973fe88010) at sna_glyphs.c:227 #2 0x00007f973c70eafc in sna_accel_create (screen=screen@entry=0x1327410, sna=sna@entry=0x7f973fe88010) at sna_accel.c:13829 #3 0x00007f973c733792 in sna_create_screen_resources (screen=0x1327410) at sna_driver.c:171 #4 0x000000000049e6c7 in xf86CrtcCreateScreenResources (screen=0x1327410) at xf86Crtc.c:704 #5 0x000000000042320c in main (argc=7, argv=0x7ffffaa61b88, envp=<optimized the screen_priv pointer is null: (gdb) p screen_priv $1 = (ShmScrPrivateRec *) 0x0 i'm not sure if the "sna: Re-register the SHM funcs every server generation" commit fixes anything in this area... 2.20.5 still crashes. with the latest 2.20.7 release and the newest xserver it seems to work for me. xorg-driver-video-intel-2.20.7-1.x86_64 xorg-xserver-libglx-1.13.0-1.x86_64 xorg-xserver-server-1.13.0-1.x86_64 (In reply to comment #16) > with the latest 2.20.7 release and the newest xserver it seems to work for me. > > xorg-driver-video-intel-2.20.7-1.x86_64 > xorg-xserver-libglx-1.13.0-1.x86_64 > xorg-xserver-server-1.13.0-1.x86_64 it was too optimistic. now the first kdm login-logout cycle works but the second cycle crashes as usual: Program received signal SIGSEGV, Segmentation fault. 0x0000000000505847 in ChangePicture (pPicture=pPicture@entry=0x7f80480a9010, vmask=0, vmask@entry=4096, vlist=0x7fff792d3050, vlist@entry=0x7fff792d304c, ulist=ulist@entry=0x0, client=client@entry=0x1c04930) at picture.c:1226 1226 (*ps->ChangePicture) (pPicture, maskQ); (gdb) bt #0 0x0000000000505847 in ChangePicture (pPicture=pPicture@entry=0x7f80480a9010, vmask=0, vmask@entry=4096, vlist=0x7fff792d3050, vlist@entry=0x7fff792d304c, ulist=ulist@entry=0x0, client=client@entry=0x1c04930) at picture.c:1226 #1 0x0000000000505fde in CreatePicture (pid=pid@entry=0, pDrawable=0x1f6f1b0, pFormat=pFormat@entry=0x21efa68, vmask=vmask@entry=4096, vlist=vlist@entry=0x7fff792d304c, client=0x1c04930, error=error@entry=0x7fff792d3048) at picture.c:796 #2 0x00007f8049c28099 in sna_glyphs_create (sna=sna@entry=0x7f804c93c010) at sna_glyphs.c:225 #3 0x00007f8049c032f9 in sna_accel_create (sna=sna@entry=0x7f804c93c010) at sna_accel.c:14486 #4 0x00007f8049c25ddd in sna_create_screen_resources (screen=0x21a19f0) at sna_driver.c:171 #5 0x00000000004ab56e in xf86CrtcCreateScreenResources (screen=0x21a19f0) at xf86Crtc.c:706 #6 0x000000000042755d in main (argc=7, argv=0x7fff792d3238, envp=<optimized out>) at main.c:222 commit 1e06d19a00f5a5a05369deeb3c5ae15b282c0f92 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Mon Nov 26 15:30:09 2012 +0000 sna: Disable shadow tracking upon regen |
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.