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@openbsd.org
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; } }
On Sun, Aug 4, 2013 at 7:12 AM, Mark Kettenis mark.kettenis@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@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@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Date: Sun, 4 Aug 2013 12:21:53 -0400 From: Alex Deucher alexdeucher@gmail.com
On Sun, Aug 4, 2013 at 7:12 AM, Mark Kettenis mark.kettenis@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@openbsd.org
We should just always program the MC as on previous asics.
That one does the trick as well.
Thanks,
Mark
dri-devel@lists.freedesktop.org
-
Alex Deucher -
Mark Kettenis