On Tue, 01 Aug 2017 21:25:11 +0200, Mathieu Larouche wrote:
On 18/07/2017 10:43 AM, Takashi Iwai wrote:
Hi,
this is a summer cleanup sale, a patchset containing various fixes for mgag200 driver taken from openSUSE / SUSE kernels. They have been in our kernels for ages, so at least they are supposed to be stable.
Most of patches came from Egbert, and one PM patch from me that is a resubmission of the once-post-and-lost patch.
thanks,
Takashi
===
Egbert Eich (13): drm/mgag200: Add doublescan and interlace support drm/mgag200: Add additional limits for certain G200 variants drm/mgag200: Fix memleak in error path in mgag200_bo_create() drm/mgag200: Free container instead of member in mga_user_framebuffer_destroy() drm/mgag200: Initialize data needed to map fbdev memory drm/mgag200: Simplify function mgag200_ttm_placement() drm/mgag200: Add support for MATROX PCI device IDs 0x520 and 0x521 drm/mgag200: Cleanup cursor BOs properly drm/mgag200: Add missing drm_connector_unregister() drm/mgag200: Don't use crtc_* parameters for validation drm/mgag200: Consolidate depth/bpp handling drm/mgag200: Add command line option to specify preferred depth drm/mgag200: Add mode validation debugging code
Takashi Iwai (1): drm/mgag200: Implement basic PM support
drivers/gpu/drm/mgag200/mgag200_drv.c | 54 +++++++++++ drivers/gpu/drm/mgag200/mgag200_drv.h | 22 +++-- drivers/gpu/drm/mgag200/mgag200_fb.c | 14 +-- drivers/gpu/drm/mgag200/mgag200_main.c | 146 +++++++++++++++++++++++++++-- drivers/gpu/drm/mgag200/mgag200_mode.c | 166 ++++++++++++++++++++++++++------- drivers/gpu/drm/mgag200/mgag200_ttm.c | 11 ++- 6 files changed, 357 insertions(+), 56 deletions(-)
Patches were tested against G200eW3, G200e4 & G200eH3 and it's working fine and we haven't seen any issues.
Thanks for testing!
There's one thing though, the patch "[PATCH 01/14] drm/mgag200: Add doublescan and interlace support" may cause problems as doublescan and interlace aren't tested and aren't officially supported on the G200 server line products. So, I'm wondering if it shouldn't be kept disabled for them.
OK, that's good to know. My understanding is that the patch was brought only for adapting the UMS X driver quality, but not for actually fixing the real bugs. Let's drop that patch, then.
Takashi