Re: [Bug #12806] i915 broken STR

[Eric Anholt]

On Tue, 2009-03-10 at 14:56 +0100, Rafael J. Wysocki wrote:
> On Tuesday 10 March 2009, Eric Anholt wrote:
> > On Sat, 2009-03-07 at 22:28 +0100, Rafael J. Wysocki wrote:
> > > On Saturday 07 March 2009, Harvey Harrison wrote:
> > > > On Tue, 2009-03-03 at 20:25 +0100, Rafael J. Wysocki wrote:
> > > > > This message has been generated automatically as a part of a report
> > > > > of recent regressions.
> > > > > 
> > > > > The following bug entry is on the current list of known regressions
> > > > > from 2.6.28.  Please verify if it still should be listed and let me know
> > > > > (either way).
> > > > > 
> > > > > 
> > > > > Bug-Entry	: http://bugzilla.kernel.org/show_bug.cgi?id=12806
> > > > > Subject		: i915 broken STR
> > > > > Submitter	: Harvey Harrison <harvey.harrison@xxxxxxxxx>
> > > > > Date		: 2009-02-28 4:20 (4 days old)
> > > > > First-Bad-Commit: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=5669fcacc58bf3a7386057addffd280d75380858
> > > > > References	: http://marc.info/?l=linux-kernel&m=123579487801064&w=4
> > > > > 
> > > > > 
> > > > 
> > > > This looks to have been a case of old-userspace..upgrading to
> > > > xorg-intel-video 2.6.1 (ubuntu Jaunty Alpha5) fixes STR here.
> > > 
> > > Well, thanks, but I'm not really sure if this is a good news.
> > > 
> > > Jesse, Eric, does it mean we now break resume on anything earlier than 2.6.1
> > > on the Harvey's hardware?
> > 
> > There's not much information (dmesg, Xorg.0.log) with this bug -- I'm
> > guessing he's using KMS, given that that commit was where the problems
> > started.  If so, yes, you really do need good userland for KMS to work.
> 
> BTW, Is suspend generally supposed to work with KMS and the 2.6.1 (or later) X
> driver?

I pulled a giant pile of KMS bugfixes into 2.6.2, so I'm not sure about
2.6.1.

We've still got some work to do with making our suspend/resume path
really complete for KMS.  We're missing some register setup that UMS
does, so behavior can be a bit random.  The current plan is to switch to
using the KMS modeset paths for resume, so we hopefully have reliable
(or at least consistent) mode setting behavior, and then fill those
remaining register writes in around it.  Basically, a separate
suspend/resume path for KMS, when we have full control of the device.

-- 
Eric Anholt
eric@xxxxxxxxxx                         eric.anholt@xxxxxxxxx

Attachment: signature.asc
Description: This is a digitally signed message part