Re: [PATCH 3/6] drm/i915: Wrap the preallocated BIOS framebuffer and preserve for KMS fbcon v7

[Daniel Vetter <daniel@ffwll.ch>]

On Sat, Dec 14, 2013 at 12:13:45PM +0100, Daniel Vetter wrote:
> On Thu, Dec 12, 2013 at 12:41:54PM -0800, Jesse Barnes wrote:
> > +	ifbdev->helper.funcs->initial_config = intel_fb_initial_config;
> 
> This here is a bit surprising - my model of operation here presumed that
> if we correctly assign the crtc->fb and the ifbdev->fb pointers we could
> fully rely on the fastboot setcrtc logic to eschew the modeset.
> 
> Being the ever-vary of special-purpose logic I'd much prefer this implicit
> approach - otherwise we have one more special case to care about in the
> fastboot=y/n and CONFIG_FB=y/n matrix.
> 
> So have you tried to ditch this special initial_config functions
> (obviously only looks good with fastboot=1) or what precise corner-case
> does this fix?

Ok, I've dug out your old patch from almost a year ago which added the
->initial_config hook. I see the point now of copying exactly the bios
config in the hope that we end up with something that has a higher chance
of working.

But imo this is an issue separate from the "take over bios fb" feature
here, so this should be
- split into a separate patch
- used even when we fail to take over the bios fb
The later point will require some mode-from-pipe_config reconstruction to
work outside of the fastboot=1 hack mode.

I really like the idea though.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch