[Intel-gfx] [PATCH 1/2] drm/i915: introduce & use i915_gem_object_mark_dirty()

[Dave Gordon]

On 28/04/16 17:34, Chris Wilson wrote:
> On Thu, Apr 28, 2016 at 05:26:20PM +0100, Dave Gordon wrote:
>> This just hides the existing obj->dirty flag inside a trivial inline
>> setter, to discourage non-GEM code from looking too closely.
>>
>> Existing code that sets obj->dirty is then changed to use the function
>> instead.
>
> I prefer set_dirty, unset_dirty, is_dirty which is what I used in my
> patches.
> -Chris

I wasn't going to abstract the test and clear operation because only GEM 
code needs those. In fact (apart from debugfs and error capture), only 
put_pages() implementations ever test or clear them.

Anyway, the real reason for sending this patchset was to see whether my 
local result was reproducible, namely that it will expose at least one 
path where an object is marked dirty while not pinned, in defiance of 
your previous comment about the invalidity of such an operation.

.Dave.