Bug 13132 - vbetool stops X from initialising
Summary: vbetool stops X from initialising
Status: CLOSED WILL_NOT_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(Other) (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-04-17 22:43 UTC by Darren Salt
Modified: 2012-05-30 16:17 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.29.1, 2.6.30-rc
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments

Description Darren Salt 2009-04-17 22:43:20 UTC
Running "vbetool vbestate save" is sufficient to prevent X from properly starting up, leaving a blank screen. Killing X "works" but leaves a static cursor at the top left.

inteldrmfb is in use on 945-based hw.

Leaving X "running", backtraces such as the following start appearing. I've also observed unrelated tasks getting stuck (the scheduler can get caught up in this too). Something not being unlocked...?

INFO: task events/0:5 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
events/0      D f722b4e0     0     5      2
 f706bf60 00000046 f7040e00 f722b4e0 f62cc0ac f7040f90 f64f3228 c011849e
 f62cc080 c04dfd3c f7350d58 f7040e00 f70d6000 f706bf7c c03a6fbf f7350d5c
 f7350d5c f7040e00 f7350d58 f70d6bbc f706bf8c c03a6f31 000001f4 f7350d48
Call Trace:
 [<c011849e>] ? set_next_entity+0x14/0x39
 [<c03a6fbf>] __mutex_lock_slowpath+0x42/0x64
 [<c03a6f31>] mutex_lock+0x19/0x1d
 [<c026b10a>] i915_gem_retire_work_handler+0x22/0x57
 [<c026b0e8>] ? i915_gem_retire_work_handler+0x0/0x57
 [<c01284dc>] worker_thread+0xe4/0x153
 [<c012a82a>] ? autoremove_wake_function+0x0/0x30
 [<c01283f8>] ? worker_thread+0x0/0x153
 [<c012a6bd>] kthread+0x42/0x69
 [<c012a67b>] ? kthread+0x0/0x69
 [<c0102fd3>] kernel_thread_helper+0x7/0x10
Comment 1 Andrew Morton 2009-04-18 00:26:24 UTC
Reassigned this regression to drivers-video-dri.

Which kernels were OK?  2.6.28?
Comment 2 Darren Salt 2009-04-19 14:06:24 UTC
2.6.28 is fine. I've not (yet) checked 2.6.29-rc* for this.

Note You need to log in before you can comment on or make changes to this bug.