[Intel-gfx] [PATCH 4/4] drm/atomic-helpers: Recover full cursor plane behaviour

[Dave Gordon]

On 21/01/15 17:45, Daniel Vetter wrote:
> On Wed, Jan 21, 2015 at 04:59:08PM +0000, Dave Gordon wrote:
>> On 20/01/15 22:09, Daniel Vetter wrote:
>>> Cursor plane updates have historically been fully async and mutliple
>>> updates batched together for the next vsync. And userspace relies upon
>>> that. Since implementing a full queue of async atomic updates is a bit
>>> of work lets just recover the cursor specific behaviour with a hint
>>> flag and some hacks to drop the vblank wait.
>>
>> I tried this patch 4/4 alone, and then the whole set of four, on
>> -internal, which is currently also suffering from the laggy-mouse
>> problem, and it didn't fix the issue. Was it expected that this patchset
>> would cure that problem, or is there something else required?
> 
> Nah, since I didn't bother to hack up the plane transitional helpers that
> we're currently using. The laggy cursor /should/ be mostly fixed though
> with
> 
> commit d99b70ce7d73d78a88311453ccdd0fa0a670dd50
> Author: Matt Roper <matthew.d.roper at intel.com>
> Date:   Mon Jan 19 08:31:49 2015 -0800
> 
>     drm/plane-helper: Skip prepare_fb/cleanup_fb when newfb==oldfb
> 
> Can you pls double-check that you have that patch?
> 
> This patch series here is only effective one Matt Roper's latest series to
> switch to atomic proper has landed. But if there's a need I can extend it
> a bit to the plane transitional helpers.
> -Daniel

Thanks Daniel, the -internal baseline that Alex & I are working with was
based on -nightly of 2015y-01m-08d-11h-43m-40s and therefore didn't have
that patch. Adding that seems to fix all the mouse lag :)

.Dave.