Bug 56333

Summary: [ivb] kernel freeze with i915 and Intel HD 4000 (ivy bridge)
Product: DRI Reporter: Per Foreby <per>
Component: DRM/IntelAssignee: Daniel Vetter <daniel>
Status: CLOSED WONTFIX QA Contact:
Severity: critical    
Priority: medium CC: ben, chris, daniel, jbarnes, jrnieder, per
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Xorg log
none
dmesg output none

Description Per Foreby 2012-10-23 17:45:31 UTC 
Created attachment 68963 [details]
Xorg log

chipset: Intel GT2 / HD 4000
system environment: x86_64
libdrm version: 2.4.33
mesa version: 8.0.4
xf86-video-intel version: 2.19.0
xserver version: 1.12.4
kernel version: 3.2.0
distribution: debian wheezy
hardware: Asus P8Z77-V LE, i7 3770
display connector: displayport

Running with 64 MB GPU RAM on the 3.2.0 kernel, the computer totally freezes within minutes or hours. The freezes have happened on keyboard or mouse input, and the computer can run unattended with heavy load (folding@home) for weeks without problems.

With 256 MB GPU RAM comfigured in BIOS, the freeze took days to occur, and with a 3.5 there have been no freezes so far.

If I blacklist i915 in the kernel to use the Vesa driver, the system is stable, so it looks like the problem in in the i915 module, and probably has something to do with memory handling.

The whole system freezes, and power cycling is the only way out.

Nothing shows up in the logs, even with netconsole and drm.debug=0xe.

Please see the debian bug at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=689268 for a long discussion on the subject (which ended in a request to report the problem upstream). As you can see there, I'm not the only one with this problem.
Comment 1 Per Foreby 2012-10-23 17:46:47 UTC 
Created attachment 68964 [details]
dmesg output
Comment 2 Per Foreby 2012-10-23 17:51:11 UTC 
Oops, typo.

What I meant to say is that the system is stable on a 3.5 *kernel*. I tried the one from debian experimental and hade no freezes for 11 days (when we unfortunately had a power outage).
Comment 3 Daniel Vetter 2012-10-23 18:26:20 UTC 
I'm confused a bit: To clarify: 3.2 is broken and 3.5 works?
Comment 4 Jonathan Nieder 2012-10-23 18:55:36 UTC 
(In reply to comment #3)
> I'm confused a bit: To clarify: 3.2 is broken and 3.5 works?

Yep. This report is a request to help find the commit to backport to stable kernels so others don't run into the same freeze.
Comment 5 Daniel Vetter 2012-10-23 19:10:43 UTC 
In that case I highly suggest to do a reverse bisect and submit the patch that fixes things to stable. We generally don't bother to support anything which isn't the latest and greatest release, simply because we fix too many bugs.
Comment 6 Jonathan Nieder 2012-10-23 20:32:58 UTC 
Ok, thanks.  Of course if the symptoms (ivb, system locks up when
interacting with the machine e.g. by webbrowsing for a few hours or
so, fan spins down, some noninteractive gpu stress testing didn't
trigger it) ring a bell to anyone, please don't hesitate to let me
know.

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.