nouveau driver causing system lockups

[Frode Nordahl]

System locks up consistently whenever display is locked after 5 minutes idle. Other situations might be adding or removing a display.

Jan 16 09:17:52 frode-MacBookPro kernel: [ 2941.179883] nouveau 0000:01:00.0: fifo: read fault at 000039e000 engine 00 [GR] client 10 [PD] reason 02 [PTE] on channel 5 [003f958000 compiz[2469]]
Jan 16 09:17:52 frode-MacBookPro kernel: [ 2941.179887] nouveau 0000:01:00.0: fifo: gr engine fault on channel 5, recovering...
Jan 16 09:18:07 frode-MacBookPro kernel: [ 2956.373531] [TTM] Buffer eviction failed

I have been using mainline kernel 4.9.2 for some days without having any of these issues.

Will try to find related commits and/or bisect to find exact version that improves the behavior.

[Frode Nordahl (fnordahl)]

Feb 2 08:43:51 frode-MacBookPro kernel: [ 5188.344618] nouveau 0000:01:00.0: fifo: read fault at 0012162000 engine 1b [CE2] client 18 [GR_CE] reason 02 [PTE] on channel 2 [003fa63000 Xorg[1753]]
Feb 2 08:43:51 frode-MacBookPro kernel: [ 5188.344624] nouveau 0000:01:00.0: fifo: ce2 engine fault on channel 2, recovering...

[Gareth Bult (gareth-encryptec)]

I'm now getting this on a daily basis, standard Linux PC with GeForce GT 430, causes machine to lock completely.

Mar 15 00:48:57 darkstar kernel: [429543.957731] [TTM] Buffer eviction failed
Mar 15 00:48:57 darkstar kernel: [429543.957755] nouveau 0000:01:00.0: chrome[21695]: fail ttm_validate
Mar 15 00:48:57 darkstar kernel: [429543.957762] nouveau 0000:01:00.0: chrome[21695]: validating bo list
Mar 15 00:48:57 darkstar kernel: [429543.958250] nouveau 0000:01:00.0: chrome[21695]: validate: -12
Mar 15 00:49:12 darkstar kernel: [429558.956617] [TTM] Buffer eviction failed
Mar 15 00:49:27 darkstar kernel: [429574.043483] [TTM] Buffer eviction failed
Mar 15 00:49:27 darkstar kernel: [429574.043506] nouveau 0000:01:00.0: chrome[21695]: fail ttm_validate
Mar 15 00:49:27 darkstar kernel: [429574.043514] nouveau 0000:01:00.0: chrome[21695]: validating bo list
Mar 15 00:49:27 darkstar kernel: [429574.043551] nouveau 0000:01:00.0: chrome[21695]: validate: -12

[Gareth Bult (gareth-encryptec)]

Mar 15 00:48:27 darkstar kernel: [429513.960360] nouveau 0000:01:00.0: fifo: read fault at 0000e9a000 engine 15 [PCE0] client 01 [PCOPY0] reason 02 [PAGE_NOT_PRESENT] on channel 0 [001fe74000 DRM]
Mar 15 00:48:27 darkstar kernel: [429513.960367] nouveau 0000:01:00.0: fifo: ce0 engine fault on channel 0, recovering...

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Gaele Strootman (gaele)]

Same here.

Jul 18 11:40:50 mozzy kernel: [ 5215.664696] nouveau 0000:01:00.0: fifo: read fault at 001b510000 engine 1b [CE2] client 18 [GR_CE] reason 02 [PTE] on channel 2 [007f9a3000 Xorg[1651]]
Jul 18 11:40:50 mozzy kernel: [ 5215.664700] nouveau 0000:01:00.0: fifo: ce2 engine fault on channel 2, recovering...

[Gaele Strootman (gaele)]

This happens when using my machine, BTW. It's not idle.