[Intel-gfx] [Mesa3d-dev] [PATCH] Remove Intel pageflipping support from mesa

Owain Ainsworth zerooa at googlemail.com
Sun Jan 18 17:16:46 CET 2009


On Sun, Jan 18, 2009 at 04:48:03PM +0200, Timo Aaltonen wrote:
> On Sat, 17 Jan 2009, Owain Ainsworth wrote:
>
>> On Sat, Jan 17, 2009 at 08:57:49PM +0200, Timo Aaltonen wrote:
>>> On Sat, 17 Jan 2009, Owain Ainsworth wrote:
>>>
>>>> That's very odd. It 73.rcs are building again kernel headers on openbsd
>>>> here (with only that pageflipping patch applied). The linux kernel
>>>> doesn't have conpat defines for planeA -> pipeA, so the following patch
>>>> should solve that:
>>>
>>> So that's equivalent to applying the compat stuff to the kernel headers?
>>> It's a lot easier for me to fix mesa instead ;)
>>>
>>
>> That should be fine. after all mesa depends on the libdrm version being
>> high enough anyway.
>
> Well the build failed exactly the same way (the warnings were fixed  
> though)..

Yes, the build is having problems with dereferencing the sarea pointers.
in the following code:

  drm_clip_rect_t planeA_rect = { .x1 = sarea->planeA_x, .y1 = sarea->planeA_y,
                                  .x2 = sarea->planeA_x + sarea->planeA_w,
                                  .y2 = sarea->planeA_y + sarea->planeA_h };
  drm_clip_rect_t planeB_rect = { .x1 = sarea->planeB_x, .y1 = sarea->planeB_y,
                                  .x2 = sarea->planeB_x + sarea->planeB_w,
                                  .y2 = sarea->planeB_y + sarea->planeB_h };



My patch did not change that behaviour at all. I wonder if you're
building things correctly, I have no such problem with the sarea being
detected.

-0-
-- 
I am not an Economist.  I am an honest man!
		-- Paul McCracken



More information about the Intel-gfx mailing list