[REGRESSION]: Kernel crash with "Asynchronous wait on fence"
Problem
On an Nvidia Optimus dual-GPU system, the Nvidia-connected display suddenly turns off and the internal Intel-connected display freezes. X stops responding to keyboard/mouse input. If Ctrl+Alt+Fn is pressed repeatedly right away, it may be possible to switch to a virtual TTY, but in any case the system is in a broken state now.
It takes up to a week of "regular workstation use" to reproduce the problem.
Testing
- culprit commit is 6eaa1f3c
- v6.6 (with these patches for an ACPI regression) fails
- v6.6 with culprit commit reverted does not fail
- Compiled with attached v6.6-config1
- Decoded kernel log: async-wait-on-fence-180.log
- This patch reverts the culprit commit: 0001-Revert-d386a4b54607cf6f76e23815c2c9a3abc1d66882.patch
Hardware
- ThinkPad W530 2438-52U
- Dock with Nvidia-connected DVI ports
- Secondary display connected via DVI
- Nvidia Optimus GPU switching system
- Internal and external displays both in use
$ lspci | grep -i vga
00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)
01:00.0 VGA compatible controller: NVIDIA Corporation GK107GLM [Quadro K2000M] (rev a1)See also
Edited by Owen T. Heisler