[BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while atomic
 

When I run games under wine, the entire PC usually locks up, requiring a hard reboot. The crash does not always occur at the same point, but usually occurs within five minutes. First the game crashes, then X becomes unresponsive for long periods, and finally the entire PC locks up.

The syslog usually shows "NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context" and sometimes indicates Xid 13 and 39. What do they mean? I can't find documentation for them.

This happens with 280.13, 285.05.09, and also 290.03 and the card is a 540M.

PCI isn't using MMCONFIG.

I've attached the nvidia-bug-report.log file and also the Xorg log file for the nvidia card (because nvidia-bug-report didn't find the right one).

A sample syslog showing the crash is:

Oct 25 15:38:04 sierra kernel: [ 108.594430] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 290.03 Mon Oct 17 10:23:33 PDT 2011
Oct 25 15:42:44 sierra kernel: [ 388.502786] NVRM: Xid (0000:01:00): 13, 0006 00000000 00009197 00002490 00000030 00000000
Oct 25 15:42:44 sierra kernel: [ 388.508071] NVRM: Xid (0000:01:00): 39, CCMDs 00000005 000090b5
Oct 25 15:42:46 sierra kernel: [ 390.507398] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Oct 25 15:43:22 sierra kernel: [ 426.682227] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
Oct 25 15:44:22 sierra kernel: [ 486.683053] INFO: rcu_sched_state detected stall on CPU 2 (t=6000 jiffies)


Other sample logs:

Oct 9 11:09:37 sierra kernel: [39332.964552] NVRM: Xid (0000:01:00): 13, 0006 00000000 00009197 00002144 00004a00 00000000
Oct 9 11:09:37 sierra kernel: [39332.968471] NVRM: Xid (0000:01:00): 39, CCMDs 00000005 000090b5
Oct 9 11:09:39 sierra kernel: [39334.967992] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context

Oct 10 13:18:33 sierra kernel: [11670.045503] NVRM: Xid (0000:01:00): 13, 0006 00000000 00009197 00001f04 0a95ffff 00000000
Oct 10 13:18:33 sierra kernel: [11670.050862] NVRM: Xid (0000:01:00): 39, CCMDs 00000005 000090b5

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

I believe this is similar to my problem in this thread:

http://www.nvnews.net/vbulletin/showthread.php?t=167472

While playing World of Tanks tonight I have received many of the same issues, but with the 275.09.07 driver, which I find to be the most stable driver with Wine. I found with later drivers Xid errors would cause lockups, but with the older driver I currently use it takes many more Xid errors and finally an atomic interrupt error or Fallen off Bus error for the final lockup to take place.

First error tonight:

Oct 26 21:05:42 zod kernel: [19587.761532] NVRM: Xid (0000:03:00): 12, COCOD 0000000c beef9097 00009097 00001080 3a400048
Oct 26 21:05:42 zod kernel: [19587.769123] NVRM: Xid (0000:03:00): 12, COCOD 0000000c beef9097 00009097 00001084 3a400049
Oct 26 21:05:42 zod kernel: [19587.776668] NVRM: Xid (0000:03:00): 12, COCOD 0000000c beef9097 00009097 00001088 3a40004a
Oct 26 21:05:42 zod kernel: [19587.784136] NVRM: Xid (0000:03:00): 12, COCOD 0000000c beef9097 00009097 0000108c 3a40004b
Oct 26 21:18:11 zod kernel: [20336.641788] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 00000d9c 00000030 0000000c
Oct 26 21:31:59 zod kernel: [21164.477317] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 0000035c 00000001 00000004
Oct 26 21:31:59 zod kernel: [21164.485138] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 0000035c 00000001 00000004
Oct 26 21:46:05 zod kernel: [22010.309527] NVRM: Xid (0000:04:00): 31, Ch 00000004, engmask 00000101, intr 10000000
Oct 26 21:46:27 zod kernel: [22031.805043] NVRM: GPU at 0000:03:00.0 has fallen off the bus.

And while playing WoT again after a reboot:

Oct 26 21:51:54 zod kernel: [ 266.235405] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 00002380 3e3f42a9 0000000c
Oct 26 21:51:54 zod kernel: [ 266.242962] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 00002380 be393c1c 0000000c
Oct 26 21:51:54 zod kernel: [ 266.250373] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 00002380 bf3416a0 0000000c
Oct 26 21:51:54 zod kernel: [ 266.257834] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 00002380 43e9ce5f 0000000c
Oct 26 21:53:29 zod kernel: [ 361.032337] NVRM: Xid (0000:02:00): 13, 0006 00000000 00009097 0000192c 00000802 0000000c
Oct 26 21:53:29 zod kernel: [ 361.039871] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 0000192c 00000802 0000000c
Oct 26 21:53:29 zod kernel: [ 361.047294] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 0000192c 00000802 0000000c
Oct 26 21:53:53 zod kernel: [ 384.156466] NVRM: Xid (0000:02:00): 13, 0006 00000000 00009097 00002380 00000000 00000004

The computer has not yet locked up due to these errors.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Yes, yours looks pretty similar. It's a pretty serious bug when it locks up the entire kernel, isn't it!

I've seen threads in this forum about the 'attempted to yield... while atomic' locking-up problem from last year as well, so the bug has been around for a long time. Unfortunately, no-one from nvidia has even posted a reply to any of them, as far as I can see.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

I made a 32-bit installation to see whether this is specific to x86_64, but the nvidia 280.13 driver also crashes and hangs the entire machine on 32 bit when playing games under wine.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

A workaround that I have found is to renice the wine exe process to say -15.

The first time I tried this, I got 25 minutes of gameplay without any crashing: when I reniced the process back to zero it crashed within a minute.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

OK, decreasing the nice level is not a perfect workaround - nvidia just managed to totally crash the system again. But it's much less likely than before.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Ya sorry it took so long for me to reply but I wanted to thoroughly test this reniceing. It does not seem to help very much or at all for me. It maybe slows the errors but it still causes hard freezes with 290.03. Going to revert back to 275.09.07 to get some stability back.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

*bump*

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Bump. I am having this issue with 275.09.07, but it is quite random. Sometimes it will happen after boot, and almost always after running a CUDA application. Then the game will run fine for a few days and I think its good but the cycle repeats in a couple of days. I am using this driver as I find it is the most stable. My latest kern.log exerpt from a crash a few minutes ago:

Nov 3 18:58:02 zod kernel: [ 3539.529967] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 0000142c 00000020 0000000c
Nov 3 18:58:03 zod kernel: [ 3540.035552] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 000003e0 20164010 0000000c
Nov 3 18:58:03 zod kernel: [ 3540.043066] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 000003e0 20164010 0000000c
Nov 3 18:58:05 zod kernel: [ 3542.623257] NVRM: Xid (0000:02:00): 13, 0006 00000000 00009097 000017d8 00000021 0000000c
Nov 3 18:58:05 zod kernel: [ 3542.630806] NVRM: Xid (0000:03:00): 13, 0006 00000000 00009097 000017d8 00000021 0000000c
Nov 3 18:58:05 zod kernel: [ 3542.638240] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 000017d8 00000021 0000000c
Nov 3 18:58:05 zod kernel: [ 3542.654608] NVRM: Xid (0000:02:00): 13, 0006 00000000 00009097 000017d8 00000021 0000000c
Nov 3 18:58:05 zod kernel: [ 3542.662057] NVRM: Xid (0000:04:00): 13, 0006 00000000 00009097 000017d8 00000021 0000000c

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Still happening with 290.06. After about 30 minutes of using this version I crashed while using VDPAU in smplayer.

Nov 8 18:14:03 zod kernel: [ 3348.205833] NVRM: Xid (0000:02:00): 13, 0004 00000000 00009039 00000304 00000000 00000000
Nov 8 18:14:03 zod kernel: [ 3348.335850] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 00001b0c 1000f010 00000000
Nov 8 18:14:03 zod kernel: [ 3348.344875] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 00001b0c 1000f010 00000000
Nov 8 18:14:03 zod kernel: [ 3348.353794] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 00001b0c 1000f010 00000000
Nov 8 18:17:14 zod kernel: [ 3539.307655] NVRM: Xid (0000:04:00): 13, 0004 00000000 00009097 00002398 00000000 00000000
Nov 8 18:17:14 zod kernel: [ 3539.315304] NVRM: Xid (0000:04:00): 31, Ch 00000004, engmask 00000101, intr 10000000
Nov 8 18:17:27 zod kernel: [ 3552.349067] NVRM: Xid (0000:02:00): 8, Channel 00000000

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Yes, I experienced it relatively quickly with 290.06 (5 mins), even with my wine app reniced to -19.

Re: [BUG] nvidia crashes kernel with 'Xid 13' and attempted to yield the CPU while at
 

Bump. Still happening very often with gaming in wine. These Xid's are not always from wine, they happen at random times. Using VDPAU or CUDA then running World of Tanks in wine will almost guarantee these errors will occur and freeze the computer, although ssh still works.

Nov 11 11:15:42 zod kernel: [ 1601.962405] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 00002388 40c34420 00000004
Nov 11 11:16:04 zod kernel: [ 1624.386556] NVRM: Xid (0000:03:00): 32, Channel ID 00000004 intr 00040000
Nov 11 11:30:02 zod kernel: [ 2459.364332] NVRM: Xid (0000:04:00): 13, 0004 00000000 00009097 00001618 00000024 00000005
Nov 11 11:33:13 zod kernel: [ 2649.160048] NVRM: Xid (0000:03:00): 12, COCOD 00000004 beef9097 00009097 00002680 00000001
Nov 11 11:38:17 zod kernel: [ 2952.074928] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 000003e0 20164010 0000000c
Nov 11 11:38:33 zod kernel: [ 2968.412126] NVRM: Xid (0000:03:00): 32, Channel ID 00000004 intr 00040000
Nov 11 11:38:35 zod kernel: [ 2970.337309] NVRM: Xid (0000:03:00): 13, 0004 00000000 00009097 00001168 38400c20 0000000c
Nov 11 11:38:38 zod kernel: [ 2973.701132] NVRM: Xid (0000:02:00): 13, 0004 00000000 00009097 0000131c 0b22c000 00000000

Please let us know if the issue has been repro'd and is being fixed!