Bug 86849

Summary: [GM965/GL960 bisected] kernel 3.18 machine frozen a few seconds after boot & login
Product: DRI Reporter: YannisDas <yannis.dassiras>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED DUPLICATE QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: major    
Priority: high CC: intel-gfx-bugs
Version: XOrg git   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
git bisect run 1
none
git bisect run 2
none
trace from "good" point
none
trace 2 from "good" point
none
dmesgs logging
none
CPU2 disappears
none
dmesgs logging
none
CPU2 disappears none

Description YannisDas 2014-11-29 14:13:53 UTC
Created attachment 110224 [details]
git bisect run 1

Machine freezes a few seconds after login

I run the bisection twice as the results are strange.

First bisection pointed to [a12624959ad4e3bfa8c344ad71728ffc9a379158] which is strange (I might have messed something up)

Second run points to the merge commit [40d201af0b9e6196a210b97d3b2493b1156564f6] which is strange too.

I have captured too traces from "good", as in not frozen, bisection points attached below.


Machine: Toshiba Portege M700
Distro: Ubuntu 14.04 x86_64
Comment 1 YannisDas 2014-11-29 14:14:18 UTC
Created attachment 110225 [details]
git bisect run 2
Comment 2 YannisDas 2014-11-29 14:15:05 UTC
Created attachment 110226 [details]
trace from "good" point
Comment 3 YannisDas 2014-11-29 14:15:36 UTC
Created attachment 110227 [details]
trace 2 from "good" point
Comment 4 Dave Airlie 2014-11-30 22:07:46 UTC
can you ssh in from another machine?

boot with nomodeset, then modprobe i915 modeset=1 and see it hangs

or at least some stuff might get to logs.

Looks like an i965GM.
Comment 5 YannisDas 2014-11-30 23:01:28 UTC
I managed to ssh in the machine before the freeze but the connection does not survive the freeze. The same trace is there too.

I also set-up a small script taking dmesgs at the default 2sec interval but this too dies after the freeze. I got 3 blank files and then it stopped trying alltogether.

Booting with nomodeset is fine.
Even after "modprobe i915 modeset=1" the machine lives on happily
Comment 6 YannisDas 2014-11-30 23:14:44 UTC
This time, something like a minute after the hang and as I was trying to resurrect the ssh connection, the machine showed signs of life.

I have a few screenlets running on the desktop among which a digital clock and two cpu graph indicators.

The seconds started ticking and the cpu graphs started moving. No mouse or keyboard reaction but cpu1 graph was behaving normally showing fluctuations. cpu2 graph was moving but with at zero usage flat.

Maybe it's just a side effect.
Comment 7 YannisDas 2014-12-02 18:35:24 UTC
Created attachment 110371 [details]
dmesgs logging

Notice that after the freeze file is empty and "watch" has skipped a second
Comment 8 YannisDas 2014-12-02 18:36:27 UTC
Created attachment 110372 [details]
CPU2 disappears

CPU2 disappears
Comment 9 YannisDas 2014-12-02 18:38:46 UTC
Created attachment 110373 [details]
dmesgs logging

Notice that after the freeze he file is empty and "watch" skips a second
Comment 10 YannisDas 2014-12-02 18:39:35 UTC
Created attachment 110374 [details]
CPU2 disappears

CPU 2 is missing?
Comment 11 YannisDas 2014-12-15 06:14:10 UTC
I did a third bisection focusing on the trace and starting with the merge as bad commit this time.

The trace begins after commit d6bbafa183793537d8dca4d4c2e448805e59448a so I am cc'ing Chris.
Comment 12 Chris Wilson 2015-01-06 10:43:02 UTC

*** This bug has been marked as a duplicate of bug 85888 ***

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.