[Intel-gfx] [Patch 1/5]: DRM/I915: Add the crt save/restore function for VT switch

[Jesse Barnes]

On Friday, March 6, 2009 5:55:52 pm Keith Packard wrote:
> On Fri, 2009-03-06 at 15:19 -0800, Eric Anholt wrote:
> > Given that these functions aren't actually hooked up to anything, I'm
> > not sure if there's any value here.  It sounds like we're going to go to
> > using the resume_force_mode function for suspend/resume, which means
> > we'd never use these.
>
> I'm concerned that our mode setting code doesn't hit all of the same
> registers that the suspend/resume code does. There are a pile of
> registers which the BIOS may whack, and which will not be reset
> correctly unless we save their values before the system is suspended the
> first time.
>
> Otherwise, I love the 'just set the mode' plan. Any chance we can strip
> the suspend/resume code down to registers which our mode setting code
> *doesn't* save/restore? Or merge suitable 'restorish' bits into the mode
> setting path so that the other registers are set in the normal mode
> setting path?

Yeah, I think that's the direction we should be heading.  The top leve 
suspend/resume routines should handle global GPU state (and possibly 
save/restore a GPU context), but beyond that we just set a mode, which should 
take care of the rest.  That means we can get rid of all the output specific 
save/restore routines as a bonus.

-- 
Jesse Barnes, Intel Open Source Technology Center