Bug 28102

Summary:	New display errors in 2.6.38-rc2-00175-g6fb1b30
Product:	Drivers	Reporter:	Maciej Rutecki (maciej.rutecki)
Component:	Video(DRI - Intel)	Assignee:	drivers_video-dri-intel (drivers_video-dri-intel)
Status:	CLOSED UNREPRODUCIBLE
Severity:	normal	CC:	maciej.rutecki, manfred.kitzbichler, nico-bugzilla.kernel.org, rjw
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	2.6.38-rc2	Subsystem:
Regression:	Yes	Bisected commit-id:
Bug Depends on:
Bug Blocks:	27352

Description Maciej Rutecki 2011-02-02 20:27:16 UTC

Subject    : New display errors in 2.6.38-rc2-00175-g6fb1b30
Submitter  : Nico Schottelius <nico-kernel-20110117@schottelius.org>
Date       : 2011-01-28 20:29
Message-ID : 20110128202905.GB3395@schottelius.org
References : http://marc.info/?l=linux-kernel&m=129624657425949&w=2

This entry is being used for tracking a regression from 2.6.37. Please don't
close it until the problem is fixed in the mainline.

Comment 1 Nico Schottelius 2011-02-24 17:48:08 UTC

Still happens in vmlinuz-2.6.38-rc6-00020-gd8204a3.

I tried to bisect, but most versions in between of 2.6.38-rc2-00175-g6fb1b30
and 2.6.37 crash, so I stopped trying.

Comment 2 Rafael J. Wysocki 2011-04-05 06:20:30 UTC

On Monday, April 04, 2011, Nico Schottelius wrote:
> Hey!
> 
> I've tested 2.6.38, which does not have this bug anymore.

Comment 3 Manfred Kitzbichler 2011-06-18 19:27:41 UTC

Please reopen this bug, it seems to be a regression even into 2.6.39 !

This is my kernel version:
2.6.39.1-32-default #1 SMP Sun Jun 12 23:19:51 UTC 2011 (fb46697) x86_64 x86_64 x86_64 GNU/Linux

And this is what I found in my "warn" log file after my display crashed showing scary static:
[drm:ironlake_fdi_link_train] *ERROR* FDI train 2 fail!

My CPU is an Intel core i5 M560 with integrated graphics, so it's using the i915 kernel module.

Xorg.0.log contains this:
[    25.651] (II) intel(0): Integrated Graphics Chipset: Intel(R) Arrandale
[    25.651] (--) intel(0): Chipset: "Arrandale"


I had similar problems with kernel 2.6.38 when I tried it on my openSUSE, so I went back to 2.6.37 where everything seems to work fine. Now that 2.6.39 came out, I hoped these issues were resolved, but this does not seem to be the case.

Based on Chris Wilson's comment on the original bug report it would appear that the unstable power saving features of the Ironlake CPU which had been taken out of the kernel were put back in - again. This seems to have produced regressions already at least twice.