Created attachment 360648 [details] log lines mentioning nouveau warnings and errors I've encountered several system crashes related to usage of the nouveau driver in the past few days. I'm attaching the output of grep -h -C1 -E '\] nouveau [^ 0]|Linux version' applied to the currently available kernel logs. I can see no obious pattern there. During boot, it sometimes mentions "unhandled intr 0x00000080", but in some cases everything works fine after that, while in some other cases it crashes even without that error message during boot. The crashes themselves aren't reported as Oops, or maybe they are but the file system is already unavailable by that time, so I don't notice. Unfortunately I don't have a serial terminal attached for logging at all times. I don't recall the exact circumstances of this Oct 08 incident. First and last line of that error: nouveau E[ PFIFO][0000:01:00.0] DMA_PUSHER - ch 2 [X[19399]] get 0x002002965c put 0x002002ae2c ib_get 0x00000015 ib_put 0x0000002f state 0x8000e684 (err: INVALID_CMD) push 0x00406040 nouveau E[ PGRAPH][0000:01:00.0] ch 2 [0x000fb33000 X[19399]] subc 7 class 0x8297 mthd 0x1358 data 0x00004303 The Oct 10 incident was while I was using Stellarium in windowed mode, and Firefox displaying and downsizing a pretty large JPEG of the orion nebula from Wikipedia. So lots of graphics activity, and a good deal of 3D rendering on the part of Stellarium. nouveau E[ PFIFO][0000:01:00.0] DMA_PUSHER - ch 4 [stellarium[29629]] get 0x002016a6bc put 0x002017c714 ib_get 0x00000201 ib_put 0x00000202 state 0x80000000 (err: INVALID_CMD) push 0x00406040 nouveau E[ PGRAPH][0000:01:00.0] ch 2 [0x000fb33000 X[2842]] subc 7 class 0x8297 mthd 0x1288 data 0xffffffff After that crash, I rebooted and just restored my Firefox session, without opening Stellarium. It crashed again. I'm really surprised that I can find no evidence of this boot in my logs. Strange. The Oct 11 incident was mostly Firefox opening a large number of PDF documents with massive use of graphics. Probably kept a number of them buffered somewhere? I'll quote the first and the last TWO lines, since the penultimate one seems similar to the last one of the previous crashes. nouveau E[ PFIFO][0000:01:00.0] DMA_PUSHER - ch 2 [X[3272]] get 0x002002a90c put 0x002002be40 ib_get 0x0000012b ib_put 0x00000141 state 0x80004861 (err: INVALID_CMD) push 0x00406040 nouveau E[ PGRAPH][0000:01:00.0] ch 2 [0x000fb33000 X[3272]] subc 0 class 0x5039 mthd 0x0328 data 0x00000000 nouveau E[ PFB][0000:01:00.0] trapped read at 0x0022733000 on channel 0x0000fb33 [X[3272]] PGRAPH/DISPATCH/M2M_IN reason: PAGE_NOT_PRESENT lspci reports the graphics card as VGA compatible controller: NVIDIA Corporation G84 [GeForce 8400 GS] (rev a1) How should I proceed? File this bug with the main kernel bug tracker?
(In reply to Martin von Gagern from comment #0) > How should I proceed? File this bug with the main kernel bug tracker? Please follow http://nouveau.freedesktop.org/wiki/Bugs/ for that. Can you provide us a link to the upstream bug report when you do? Having done a bit of playing aronud with Nouveau, I might look at this...
(In reply to Tom Wijsman (TomWij) from comment #1) > Please follow http://nouveau.freedesktop.org/wiki/Bugs/ for that. Tried to, as best as I could. dmesg simply isn't an option after a reboot… > Can you provide us a link to the upstream bug report when you do? https://bugs.freedesktop.org/show_bug.cgi?id=70390 > Having done a bit of playing aronud with Nouveau, I might look at this... I'll appreciate any attempt to sched light on this. If I can be of assistance, let me know. In the meantime, I'll see if I can get a network console going, and if I can, will try to reproduce the issue in some way, to see whether I lost any kernel oops.
