[Bug 98690] [SKL bisected] System freeze when starting X using kernel 4.9-rc1 or later

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=98690

--- Comment #26 from Gleb Nemshilov <gleb at fastmail.com> ---
Could it be the runpm related issue? Because with "nouveau.runpm=0" kernel
boots fine (as stated in another bug report I reported here with the same
issue). I can see with dmesg the following strings:

===kernel booted===
===these lines below means runpm turned off discrete adapter===
[62233.697650] nouveau 0000:02:00.0: DRM: suspending console...
[62233.697656] nouveau 0000:02:00.0: DRM: suspending display...
[62233.697673] nouveau 0000:02:00.0: DRM: evicting buffers...
[62233.697676] nouveau 0000:02:00.0: DRM: waiting for kernel channels to go
idle...
[62233.697711] nouveau 0000:02:00.0: DRM: suspending client object trees...
[62233.698025] nouveau 0000:02:00.0: DRM: suspending kernel object tree...

===these lines appear when the discrete GPU is tunred on===
[63751.307103] nouveau 0000:02:00.0: DRM: resuming kernel object tree...
[63751.539847] nouveau 0000:02:00.0: DRM: resuming client object trees...
[63751.539975] nouveau 0000:02:00.0: DRM: resuming display...
[63751.539999] nouveau 0000:02:00.0: DRM: resuming console...


It's also strange that those lines appear long after the system started,
because the GPU should be turned off all the time.

As I understand, "suspending" stuff means that discrete GPU is truned off, and
"resuming" means it turns on. So when the affected kernel boots, the last line
is "DRM: resuming console..." and it's stuck, X server won't load. But with not
affected kernel the last lines are "DRM: suspending client object trees..." and
"DRM: suspending kernel object tree...".

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20170105/f4e503b5/attachment.html>