Skip to contentSkip to conversation

Sandy Bridge X crash - OpenGL/DPMS-related?

DISCUSSION Ian Pilcher 3 participants 7 messages Mar 26, 2011 – Mar 27, 2011

lore
Share & tools
lore
#0 lore raw

Fedora 15 Alpha on a brand new Sandy Bridge system (Core i7 2600 with "Intel HD" graphics). Running KDE and using GLMatrix screensaver.

When I unlocked the system this morning, the password entry dialog was not displayed correctly. When I pressed a key, only the password entry field became visible. The "Switch User ...", "Unlock", and "Cancel" buttons only appeared when I moved the mouse pointer over them; the dialog background remained black.

When I entered my password, X crashed. The backtrace in the log is:

Backtrace:
[ 49562.907] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x4a117f]
[ 49562.907] 1: /usr/bin/X (0x400000+0x621c6) [0x4621c6]
[ 49562.907] 2: /lib64/libpthread.so.0 (0x37ae800000+0xf4e0) [0x37ae80f4e0]
[ 49562.907] 3: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x84251)
[0x7fd70c43d251]
[ 49562.907] 4: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x6e116)
[0x7fd70c427116]
[ 49562.907] 5: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x5d5cd)
[0x7fd70c4165cd]
[ 49562.907] 6: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x148b93)
[0x7fd70c501b93]
[ 49562.907] 7: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x14690c)
[0x7fd70c4ff90c]
[ 49562.907] 8: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x146b0a)
[0x7fd70c4ffb0a]
[ 49562.907] 9: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x1090cb)
[0x7fd70c4c20cb]
[ 49562.907] 10: /usr/lib64/xorg/modules/extensions/libglx.so
(0x7fd70d839000+0x310b9) [0x7fd70d86a0b9]
[ 49562.907] 11: /usr/lib64/xorg/modules/extensions/libglx.so
(0x7fd70d839000+0x33831) [0x7fd70d86c831]
[ 49562.907] 12: /usr/bin/X (0x400000+0x2ebd1) [0x42ebd1]
[ 49562.907] 13: /usr/bin/X (0x400000+0x22e5a) [0x422e5a]
[ 49562.907] 14: /lib64/libc.so.6 (__libc_start_main+0xed) [0x37adc2131d]
[ 49562.907] 15: /usr/bin/X (0x400000+0x23141) [0x423141]
[ 49562.907] Segmentation fault at address (nil)
[ 49562.907]
Fatal server error:
[ 49562.907] Caught signal 11 (Segmentation fault). Server aborting

(I have installed the relevant debuginfo packages and reproduced the crash, but the backtrace in the X.org log doesn't appear to be using those symbols. Any way I can get a more useful backtrace?)

Based on my attempts to reproduce the crash, it appears to be DPMS- related. Unlocking the screensaver before DPMS turns off my displays works just fine. (The unlock dialog is displayed correctly, and X doesn't crash.) If I let the screensaver run until the displays are powered down, the unlock dialog is display incorrectly and X crashes when I enter my password and press enter.

Any ideas?

Thanks!

Signature · 4 lines
========================================================================
Ian Pilcher                                         [email protected]
"If you're going to shift my paradigm ... at least buy me dinner first."
========================================================================
#1 lore raw

Hi,

Ian Pilcher <[email protected]> (26/03/2011):

Quoted 3 lines

Fedora 15 Alpha on a brand new Sandy Bridge system (Core i7 2600 with "Intel HD" graphics). Running KDE and using GLMatrix screensaver.

Fedora might have a better doc but…

Quoted 3 lines

(I have installed the relevant debuginfo packages and reproduced the crash, but the backtrace in the X.org log doesn't appear to be using those symbols. Any way I can get a more useful backtrace?)

(see man 3 backtrace, which is used within X) the easiest way is to get X to dump a core, or to attach it from gdb (possibly from another machine); here's the doc I'm pointing Debian users to, instructions there should be quite portable across distros:
http://pkg-xorg.alioth.debian.org/howto/use-gdb.html

You may want to post kernel logs (see dmesg), and mention your kernel and X versions (unless everyone else knows what Fedora 15α ships).

KiBi.

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#2 lore raw
Quoted 5 lines · Cyril Brulebois

(see man 3 backtrace, which is used within X) the easiest way is to get X to dump a core, or to attach it from gdb (possibly from another machine); here's the doc I'm pointing Debian users to, instructions there should be quite portable across distros:
http://pkg-xorg.alioth.debian.org/howto/use-gdb.html

I will take a look at this.

Quoted 2 lines

You may want to post kernel logs (see dmesg), and mention your kernel and X versions (unless everyone else knows what Fedora 15α ships).

Most of the gory details can be found at:

    https://bugzilla.redhat.com/show_bug.cgi?id=691084

Software versions are:

    xorg-x11-server-common-1.10.0-4.fc15.x86_64
    xorg-x11-server-Xorg-1.10.0-4.fc15.x86_64
    xorg-x11-drv-intel-2.14.0-4.fc15.x86_64
    mesa-dri-drivers-7.10.1-1.fc15.x86_64

Thanks!

Signature · 9 lines
========================================================================
Ian Pilcher                                         [email protected]
"If you're going to shift my paradigm ... at least buy me dinner first."
========================================================================

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw

Thanks to Cyril Brulebois, I was able to get a (hopefully) useful back- trace from gdb. Attaching to avoid word-wrap borkage.

Signature · 4 lines
========================================================================
Ian Pilcher                                         [email protected]
"If you're going to shift my paradigm ... at least buy me dinner first."
========================================================================

#0 prepare_wm_surfaces (brw=0x2cd33c0) at brw_wm_surface_state.c:602
rb = 0x2c2d8a0
irb = <optimized out>
region = 0x0
ctx = 0x2cd33c0
i = <optimized out>
nr_surfaces = <optimized out>
#1 0x00007f653c0e6116 in brw_validate_state (brw=0x2cd33c0) at brw_state_upload.c:397

        atom = <optimized out>
        ctx = 0x2cd33c0
        intel = 0x2cd33c0
        state = 0x2cea2d0
        i = <optimized out>
        atoms = <optimized out>
        num_atoms = 41
#2  0x00007f653c0d55cd in brw_try_draw_prims (max_index=<optimized out>, min_index=<optimized out>, ib=0x0, nr_prims=1, prim=0x2c63f04, 
    arrays=0x2c657e8, ctx=0x2cd33c0) at brw_draw.c:362
        hw_prim = 6
        brw = 0x2cd33c0
        retval = <optimized out>
        warn = <optimized out>
        first_time = 0 '\000'
        i = <optimized out>
        intel = 0x2cd33c0
#3  brw_draw_prims (ctx=0x2cd33c0, arrays=0x2c657e8, prim=0x2c63f04, nr_prims=1, ib=0x0, index_bounds_valid=<optimized out>, 
    min_index=0, max_index=3) at brw_draw.c:447
        retval = 0 '\000'
#4  0x00007f653c1c0b93 in vbo_exec_vtx_flush (exec=0x2c63c20, unmap=1 '\001') at vbo/vbo_exec_draw.c:381
        ctx = 0x2cd33c0
#5  0x00007f653c1be90c in vbo_exec_FlushVertices_internal (ctx=<optimized out>, unmap=<optimized out>) at vbo/vbo_exec_api.c:911
        exec = 0x2c63c20
#6  0x00007f653c1beb0a in vbo_exec_FlushVertices (ctx=<optimized out>, flags=1) at vbo/vbo_exec_api.c:945
        exec = 0x2c63c20

#7 0x00007f653c1810cb in _mesa_set_scissor (ctx=0x2cd33c0, x=3279, y=<optimized out>, width=<optimized out>, height=<optimized out>)
at main/scissor.c:75
No locals.
#8 0x00007f653d5290b9 in __glXDisp_Render (cl=<optimized out>, pc=<optimized out>) at glxcmds.c:2000

        entry = {bytes = 20, varsize = 0}
        proc = 0x7f653d507cd0 <__glXDisp_Scissor>
        err = 0
        req = <optimized out>
        client = 0x2c1bae0
        left = <optimized out>
        cmdlen = 20
        error = <optimized out>
        commandsDone = <optimized out>
        opcode = <optimized out>
        hdr = 0x53e09ac
        glxc = 0x2c2db90
        sw = <optimized out>

#9 0x00007f653d52b831 in __glXDispatch (client=0x2c1bae0) at glxext.c:583
rendering = <optimized out>
stuff = 0x53e0848
opcode = <optimized out>
proc = 0x7f653d529000 <__glXDisp_Render>
cl = 0x2c1bbf8
retval = <optimized out>
#10 0x000000000042ebd1 in Dispatch () at dispatch.c:431

        clientReady = 0x2bf8f40
        result = <optimized out>
        client = 0x2c1bae0
        nready = 0
        icheck = 0x7e34f0
        start_tick = 1760

#11 0x0000000000422e5a in main (argc=<optimized out>, argv=0x7fffbe6f4ef8, envp=<optimized out>) at main.c:287
i = <optimized out>
alwaysCheckForInput = {0, 1}
A debugging session is active.

Inferior 1 [process 12799] will be detached.

Quit anyway? (y or n) Detaching from program: /usr/bin/Xorg, process 12799

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw
Quoted 7 lines · Ian Pilcher

#0 prepare_wm_surfaces (brw=0x2cd33c0) at brw_wm_surface_state.c:602
rb = 0x2c2d8a0
irb = <optimized out>
region = 0x0
ctx = 0x2cd33c0
i = <optimized out>
nr_surfaces = <optimized out>

This commit looks promising:

http://cgit.freedesktop.org/mesa/mesa/commit/?id=13bab58f04c1ec6d0d52760eab490a0997d9abe2

Signature · 4 lines
========================================================================
Ian Pilcher                                         [email protected]
"If you're going to shift my paradigm ... at least buy me dinner first."
========================================================================
#5 lore raw
In reply to Ian Pilcher
Quoted 34 lines · Ian Pilcher

When I entered my password, X crashed. The backtrace in the log is:

Backtrace:
[ 49562.907] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x4a117f]
[ 49562.907] 1: /usr/bin/X (0x400000+0x621c6) [0x4621c6]
[ 49562.907] 2: /lib64/libpthread.so.0 (0x37ae800000+0xf4e0) [0x37ae80f4e0]
[ 49562.907] 3: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x84251)
[0x7fd70c43d251]
[ 49562.907] 4: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x6e116)
[0x7fd70c427116]
[ 49562.907] 5: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x5d5cd)
[0x7fd70c4165cd]
[ 49562.907] 6: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x148b93)
[0x7fd70c501b93]
[ 49562.907] 7: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x14690c)
[0x7fd70c4ff90c]
[ 49562.907] 8: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x146b0a)
[0x7fd70c4ffb0a]
[ 49562.907] 9: /usr/lib64/dri/i965_dri.so (0x7fd70c3b9000+0x1090cb)
[0x7fd70c4c20cb]
[ 49562.907] 10: /usr/lib64/xorg/modules/extensions/libglx.so
(0x7fd70d839000+0x310b9) [0x7fd70d86a0b9]
[ 49562.907] 11: /usr/lib64/xorg/modules/extensions/libglx.so
(0x7fd70d839000+0x33831) [0x7fd70d86c831]
[ 49562.907] 12: /usr/bin/X (0x400000+0x2ebd1) [0x42ebd1]
[ 49562.907] 13: /usr/bin/X (0x400000+0x22e5a) [0x422e5a]
[ 49562.907] 14: /lib64/libc.so.6 (__libc_start_main+0xed) [0x37adc2131d]
[ 49562.907] 15: /usr/bin/X (0x400000+0x23141) [0x423141]
[ 49562.907] Segmentation fault at address (nil)
[ 49562.907]
Fatal server error:
[ 49562.907] Caught signal 11 (Segmentation fault). Server aborting
Quoted 3 lines

Any ideas?

Thanks!

I had the same problem, as I recall backtrace was similar to this one (i965_dri.so) with mesa 7.10 on ArchLinux. Although the trigger to this bug was somewhat different, exiting UrbanTerror was killing X server. Haven't had that problem after upgrading to mesa 7.10.1

I've installed Fedora 15 on a different machine yesterday and I saw that mesa 7.10.1 was in update list. If you can confirm that mesa 7.10.1 fixes the bug or not would be great.

#6 lore raw
Quoted 3 lines · Ivan Bulatovic

I've installed Fedora 15 on a different machine yesterday and I saw that mesa 7.10.1 was in update list. If you can confirm that mesa 7.10.1 fixes the bug or not would be great.

Unfortunately, I'm already running:

mesa-dri-drivers-7.10.1-1.fc15.x86_64

So doesn't look like mesa 7.10.1 fixes my particular problem. :-(

I do appreciate the response, though. Thanks!

Signature · 4 lines
========================================================================
Ian Pilcher                                         [email protected]
"If you're going to shift my paradigm ... at least buy me dinner first."
========================================================================