[Intel-gfx] [PATCH] drm/i915: Addin-offset is an unreliable indicator of LVDS present

[Adam Jackson]

On Mon, 2010-08-23 at 16:23 +0100, Chris Wilson wrote:
> On Mon, 23 Aug 2010 10:46:33 -0400, Adam Jackson <ajax at redhat.com> wrote:
> > AIM seems to be structured such that you have to have an addin to make
> > LVDS work.  I think the opregion change you posted is sufficient to find
> > the real VBT; having done that, the addin offset check is still valid.
> 
> Sadly, I discovered this after enabling the OpRegion->VBT. The Samsung
> went from finding no VBT to declaring the LVDS absent. :(
> 
> I am on the scrounge for documentation on how this all meant to hang
> together. Reality will be different, but it would be nice to know what the
> fields are supposed to mean.

addin_offset is (afaict) intended to be the offset to the setup ROM
embedded on SDVO devices once it's glued into the final image at boot
time, or the offset to the equivalent table for non-SDVO outputs.  Which
is why I considered it pretty reliable; if it points to the vtable for
output setup, then it's only going to exist if the output exists.

But I suppose some enterprising BIOS engineer could have looked at the
whole vtable thing and thought "that's too big, I can do it better".  In
which case, if we have to resort to opregion to find the VBT, then maybe
we should believe whatever it claims to have.

IWBNI we had the VBT in a sysfs or debugfs file, if we're not going to
be able to pull it out of the legacy ROM space reliably.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20100823/0503641f/attachment.sig>