FS#31338 - [linux] 3.5.3 nouveau crashes after suspend/resume

Attached to Project: Arch Linux
Opened by Miha Verlic (fluke571) - Thursday, 30 August 2012, 08:45 GMT
Last edited by Tobias Powalowski (tpowa) - Sunday, 07 October 2012, 11:24 GMT
Task Type Bug Report
Category Packages: Core
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 5
Private No

Details

Description:
After upgrading to kernel 3.5.3 suspend/resume brakes nouveau on my T61. Laptop wakes up OK, however immediatelly after starting firefox nouveau crashes. dmesg complains with "PFIFO: channel 3 unload timeout" and "Failed to idle channel 2." - attached below.

Someone bisected the bug to commit 03bd6efa ("drm/nv50/fifo: use hardware channel kickoff functionality").
http://lkml.indiana.edu/hypermail/linux/kernel/1206.1/01611.html

Additional info:
linux 3.5.3-1
libdrm-nouveau1 2.4.33-1
nouveau-dri 8.0.4-2
xf86-video-nouveau 1.0.1-1
xorg-server 1.12.4-1

01:00.0 VGA compatible controller: NVIDIA Corporation G86 [Quadro NVS 140M] (rev a1)

dmesg:
[60594.245673] [drm] nouveau 0000:01:00.0: Disabling display...
[60595.549676] [drm] nouveau 0000:01:00.0: Disabling fbcon...
[60595.549679] [drm] nouveau 0000:01:00.0: Unpinning framebuffer(s)...
[60595.549736] [drm] nouveau 0000:01:00.0: Evicting buffers...
[60595.607803] [drm] nouveau 0000:01:00.0: Idling channels...
[60595.609130] [drm] nouveau 0000:01:00.0: Suspending GPU objects...
[60598.465651] [drm] nouveau 0000:01:00.0: And we're gone!
[60598.477333] nouveau 0000:01:00.0: power state changed by ACPI to D3
[60599.248619] [drm] nouveau 0000:01:00.0: We're back, enabling device...
[60599.248629] nouveau 0000:01:00.0: power state changed by ACPI to D0
[60599.248632] nouveau 0000:01:00.0: power state changed by ACPI to D0
[60599.248636] nouveau 0000:01:00.0: power state changed by ACPI to D0
[60599.248639] nouveau 0000:01:00.0: power state changed by ACPI to D0
[60599.248645] [drm] nouveau 0000:01:00.0: POSTing device...
[60599.248647] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table 0 at offset 0xDD0F
[60599.272408] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table 1 at offset 0xE04F
[60599.298006] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table 2 at offset 0xEAA4
[60599.298034] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table 3 at offset 0xEB96
[60599.299133] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table 4 at offset 0xED83
[60599.299134] [drm] nouveau 0000:01:00.0: Parsing VBIOS init table at offset 0xEDE8
[60599.319150] [drm] nouveau 0000:01:00.0: Restoring GPU objects...
[60599.433836] [drm] nouveau 0000:01:00.0: Reinitialising engines...
[60599.434496] [drm] nouveau 0000:01:00.0: Restoring mode...
[60613.247281] [drm] nouveau 0000:01:00.0: PFIFO: channel 3 unload timeout
[60648.300629] [drm] nouveau 0000:01:00.0: Failed to idle channel 2.
[60648.480143] [drm] nouveau 0000:01:00.0: PFIFO: channel 2 unload timeout
[60708.263950] [drm] nouveau 0000:01:00.0: Failed to idle channel 2.
[60708.267279] [drm] nouveau 0000:01:00.0: PFIFO: channel 2 unload timeout
[60768.130616] [drm] nouveau 0000:01:00.0: Failed to idle channel 2.
[60768.133939] [drm] nouveau 0000:01:00.0: PFIFO: channel 2 unload timeout
[60827.980612] [drm] nouveau 0000:01:00.0: Failed to idle channel 2.
[60827.983933] [drm] nouveau 0000:01:00.0: PFIFO: channel 2 unload timeout

Steps to reproduce:
suspend, resume, run firefox, crashes every time
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Sunday, 07 October 2012, 11:24 GMT
Reason for closing:  Fixed
Additional comments about closing:  3.5.5-1
Comment by Miha Verlic (fluke571) - Thursday, 30 August 2012, 15:52 GMT Comment by Edgar Merino (emerino) - Saturday, 01 September 2012, 17:17 GMT
I can confirm this, it works fine with the above patch.
Comment by Miha Verlic (fluke571) - Sunday, 07 October 2012, 10:45 GMT
ticket can be closed, as it was fixed in 3.5.5 and 3.6

Loading...