[PATCH] drm/radeon: program the MC if loading firmware fails on ARUBA

[Alex Deucher]

On Sun, Aug 4, 2013 at 7:12 AM, Mark Kettenis <mark.kettenis at xs4all.nl> wrote:
> Failure to load the firmwares on ARUBA isn't considered fatal.
> However modesetting doesn't quite work in that case.  On an HP
> Pavillion Sleekbook 15 with A4-4335M APU, I end up with a screen full
> of garbage.  Calling evergreen_mc_program() fixes the issue.
>
> Diff below is against the OpenBSD codebase, which corresponds to Linux
> 3.8.13.
>
> Signed-off-by: Mark Kettenis <kettenis at openbsd.org>

We should just always program the MC as on previous asics.

Alex


>
>
> diff --git a/sys/dev/pci/drm/radeon/ni.c b/sys/dev/pci/drm/radeon/ni.c
> index cebaf8a..1b29475 100644
> --- a/sys/dev/pci/drm/radeon/ni.c
> +++ b/sys/dev/pci/drm/radeon/ni.c
> @@ -1519,6 +1519,7 @@ static int cayman_startup(struct radeon_device *rdev)
>                         r = ni_init_microcode(rdev);
>                         if (r) {
>                                 DRM_ERROR("Failed to load firmware!\n");
> +                               evergreen_mc_program(rdev);
>                                 return r;
>                         }
>                 }
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-drm-radeon-always-program-the-MC-on-startup.patch
Type: application/octet-stream
Size: 4769 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20130804/1b1ab876/attachment.obj>