Bug 25990

Summary:

Random lockups (Mobility HD 3400)

Product:

xorg

Reporter:

Mikhail Gusarov <dottedmag>

Component:

Driver/Radeon

Assignee:

xf86-video-ati maintainers <xorg-driver-ati>

Status:

RESOLVED WORKSFORME

QA Contact:

Xorg Project Team <xorg-team>

Severity:

normal

Priority:

medium

CC:

krummas

Version:

git

Hardware:

x86 (IA32)

OS:

Linux (All)

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
Xorg.0.log	none
lspci	none
dmesg	none

Description Mikhail Gusarov 2010-01-11 07:00:57 UTC

Hardware environment: Thinkpad T400
Software environment: Ubuntu Karmic + kernel from Lucid + xorg-edgers
Software versions:
  kernel 2.6.23-9 (lucid)
  X server 1.7.99.2~git20091130.99d88ef6
  libdrm 2.4.17+git20091230.c5c503b5
  mesa 7.8.0~git20100107.d699b672
  xf86-video-intel 2.10.0+git20100108.4902f546

I'm experiencing random lockups with radeon at unpredictable occasions (mostly after some keypress). X server busyloops trying to ioctl /dev/dri/card0 (which returns with -EBUSY), the following is in dmesg (repeated):

[90143.665813] [drm:radeon_ib_get] *ERROR* radeon: IB(8:0x0000000010181000:0)
[90143.665817] [drm:radeon_ib_get] *ERROR* radeon: GPU lockup detected, fail to get a IB
[90143.665819] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !

And the following backtrace is in Xorg.0.log:

[mi] EQ overflowing. The server is probably stuck in an infinite loop.

Backtrace:
0: /usr/bin/X (xorg_backtrace+0x3b) [0x80a403b]
1: /usr/bin/X (mieqEnqueue+0x1ab) [0x80a1c0b]
2: /usr/bin/X (xf86PostMotionEventP+0xcc) [0x80b2cec]
3: /usr/lib/xorg/modules/input/evdev_drv.so (0xb70b5000+0x3141) [0xb70b8141]
4: /usr/lib/xorg/modules/input/evdev_drv.so (0xb70b5000+0x33e9) [0xb70b83e9]
5: /usr/bin/X (0x8048000+0x692ff) [0x80b12ff]
6: /usr/bin/X (0x8048000+0x120cb4) [0x8168cb4]
7: (vdso) (__kernel_sigreturn+0x0) [0xb78ad400]
8: /usr/lib/libdrm.so.2 (drmCommandWriteRead+0x3b) [0xb744304b]
9: /usr/lib/libdrm_radeon.so.1 (0xb7458000+0x11fa) [0xb74591fa]
10: /usr/lib/libdrm_radeon.so.1 (0xb7458000+0x124d) [0xb745924d]
11: /usr/lib/libdrm_radeon.so.1 (radeon_bo_map+0x1b) [0xb745a99b]
12: /usr/lib/xorg/modules/drivers/radeon_drv.so (0xb734e000+0xa48b6) [0xb73f28b6]
13: /usr/lib/xorg/modules/libexa.so (0xb731c000+0x3713) [0xb731f713]
14: /usr/lib/xorg/modules/libexa.so (0xb731c000+0x652c) [0xb732252c]
15: /usr/lib/xorg/modules/libexa.so (0xb731c000+0x3802) [0xb731f802]
16: /usr/lib/xorg/modules/libexa.so (0xb731c000+0xf1f2) [0xb732b1f2]
17: /usr/bin/X (miImageText8+0x87) [0x81a5d57]
18: /usr/bin/X (0x8048000+0xedf0e) [0x8135f0e]
19: /usr/bin/X (doImageText+0x1ed) [0x80989cd]
20: /usr/bin/X (ImageText+0x7f) [0x8098bff]
21: /usr/bin/X (0x8048000+0x2f02e) [0x807702e]
22: /usr/bin/X (0x8048000+0x30b27) [0x8078b27]
23: /usr/bin/X (0x8048000+0x1f385) [0x8067385]
24: /lib/tls/i686/cmov/libc.so.6 (__libc_start_main+0xe6) [0xb75a2b56]
25: /usr/bin/X (0x8048000+0x1ef71) [0x8066f71]

Full dmesg, Xorg.0.log and lspci are attached.

Comment 1 Mikhail Gusarov 2010-01-11 07:52:07 UTC

Damn, I rebooted and forgot to save dmesg :(

At least the rest is still available

Comment 2 Mikhail Gusarov 2010-01-11 07:52:42 UTC

Created attachment 32567 [details]
Xorg.0.log

Comment 3 Mikhail Gusarov 2010-01-11 07:53:06 UTC

Created attachment 32568 [details]
lspci

Comment 4 Mikhail Gusarov 2010-01-11 13:33:58 UTC

Created attachment 32578 [details]
dmesg

Another lockup, now with dmesg

Comment 5 Alex Deucher 2010-01-11 13:37:09 UTC

This is a GPU lockup.  Can you isolate any particular operation or application that causes it?

Comment 6 Mikhail Gusarov 2010-01-11 13:45:54 UTC

Lockups are pretty random - I'm getting them while opening new windows, closing tabs in browser or maybe anything else (hard to tell, as I figure out that the X is locked up several seconds later, and  Though opening new window seems to be most frequent cause (I have got lot of lockups while starting new session).

PS: Is there way to prepare myself for next random lockup to collect more information?

PPS: I forgot to mention that mouse cursor still moves.

PPPS: No compositing manager is running.

Comment 7 Marcus Eriksson 2010-02-15 10:09:51 UTC

This bug affects me too, anything I can do to help?

Comment 8 Pauli 2010-02-15 10:26:08 UTC

Trying to isolate if there is any give application that has to be open always when GPU hang happens would help a lot. If all applications are causing the lockup then it will be harder to debug.

Also if you had earlier system that with KMS that didn't hang bisecting the problematic commit would help too. http://kernel.osuosl.org/pub/software/scm/git/docs/git-bisect.html

GPU hangs are very hard to debug and fix. Reducing number of components that may have caused it would help a lot.

Comment 9 Mikhail Gusarov 2010-03-04 02:37:28 UTC

Tried again and figured out that lockups are gone. Hence closing this bug.

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.