On Fri, May 27, 2005 at 10:48:50AM -0700, Ian Romanick wrote:
> Lukas Hejtmanek wrote:
>
> >drm_stub.c version 1.52 contains check whether device is AGP. I have i915
> >card
> >that is PCIE and it fails to initialize i915 module. If I comment out
> >check for
> >AGP then everything is OK.
>
> That's interesting. I guess the problem is that the i915 driver has
> DRIVER_REQUIRE_AGP set. I'm not sure what the right answer is here.
> Even though my change to drm_stub.c broke the Linux version, the BSD
> version *already* had that same check (and was therefore already broken
> in the same way).
>
> I suspect we may need to change the way drm_device_is_agp works. Right
> now, the drm_driver::device_is_agp function can only fail a device that
> would have otherwise passed. Changing the return code of
> drm_driver::device_is_agp from meaning {"might be AGP", "abosolutely not
> AGP"} to a tri-state meaning {"absolutely is AGP", "might be AGP",
> "abosolutely not AGP"} might do the trick.
>
> Keith, since you're the main i915 guy, what do you think? Would this
> bug also effect Via PCI-e chipsets with integrated graphics?
Ian,
Why does the mga_drv.c set DRIVER_REQUIRE_AGP when it doesn't 'require' it
if there's a PCI Matrox card ?
Couldn't we do....
if (drm_device_is_agp(dev) || DRIVER_REQUIRE_AGP)
dev->agp = drm_agp_init(dev);
and remove DRIVER_REQUIRE_AGP from mga_drv.c ?
Alan.
-------------------------------------------------------
This SF.Net email is sponsored by Yahoo.
Introducing Yahoo! Search Developer Network - Create apps using Yahoo!
Search APIs Find out how you can build Yahoo! directly into your own
Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005
--
_______________________________________________
Dri-devel mailing list
Dri-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dri-devel
