[Intel-gfx] Small series to enable OpRegion->VBT

[Chris Wilson]

On Tue, 24 Aug 2010 10:22:48 -0400, Adam Jackson <ajax at redhat.com> wrote:
> I'd be a little more comfortable with this series if we tried to find
> VBT in VBIOS first.  The way around you have it in patch 3 we'll likely
> be changing behaviour on existing machines, since we've not been paying
> any attention to what OpRegion's VBT contains before now.  I mean, I'm
> not going to insist on it, but I'll reserve the right to say I told you
> so.

That would be the conservative approach. The docs are adamant that
OpRegion should be used in preference to any legacy approach. But nobody
reads those docs, at least not whilst writing the BIOS or during review.

One of the goals here is to solve the mystery of what happens to the VBT
on docking and whether new devices appear via the OpRegion, through the
ROM or never. If you do have one of the affected HP systems, can you at
try the patch and grab the /sys/kernel/debug/dri/0/i915_opregion before
and after docking?

It's a fair point though, with either approach we are without a doubt
going to need quirks at some point. Does anybody have a strong preference
for the precedence order? My vote is to rid the world of BIOS and ACPI but
the problem seems to be only getting worse over time.

-- 
Chris Wilson, Intel Open Source Technology Centre