Bug 55331 - Commits leading to 3.8.3 break rc6 on Arrandale
Summary: Commits leading to 3.8.3 break rc6 on Arrandale
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - Intel) (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: intel-gfx-bugs@lists.freedesktop.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-03-16 12:25 UTC by Daniel
Modified: 2013-03-18 20:53 UTC (History)
2 users (show)

See Also:
Kernel Version: 3.8.3
Subsystem:
Regression: Yes
Bisected commit-id:

Attachments
Kernel built configuration (86.36 KB, application/octet-stream)
2013-03-16 12:25 UTC, Daniel 		Details
3.8.3: dmesg with rc6=0 (56.88 KB, text/plain)
2013-03-16 12:30 UTC, Daniel 		Details
Add an attachment (proposed patch, testcase, etc.)

Description Daniel 2013-03-16 12:25:12 UTC 
Created attachment 95511 [details]
Kernel built configuration

Hi everyone,

for me, rc6 worked with 3.8.2 and prior but is broken in 3.8.3.
With previous versions I used bitmask 7,
with 3.8.3 all bitmasks (except 0) hang the machine immediately after kernel load; hard reset needed.

Additional information attached.
Comment 1 Daniel 2013-03-16 12:30:28 UTC 
Created attachment 95521 [details]
3.8.3: dmesg with rc6=0
Comment 2 Daniel Vetter 2013-03-18 09:51:30 UTC 
Can you please bisect where this regression has been introduced exactly?
Comment 3 Daniel 2013-03-18 20:30:35 UTC 
git checkout f49b7242e6220ace73138ccae5caa85c59499ed7
make; reboot
--> Seems fine

git checkout 2a9810441fcc26cf3f006f015f8a62094fe57a90
make; reboot
--> Problem described above
Comment 4 Daniel Vetter 2013-03-18 20:53:29 UTC 
2a9810441fcc26cf3f006f015f8a62094fe57a90 will be reverted in stable, it blew up in way more places than just on your machine. So your regression should be fixed in the next 3.8.x release.

Thanks for reporting this issue and please reopen if it's again broken.
Note You need to log in before you can comment on or make changes to this bug.