earlier implementation of "DDB writing" was using this flag before
wait_for_vblank if we need to expand the DDB & it overlap with other
pipe's DDB.
In that case, I observed on corner case requiring a cd-clock change,
intel_update_watermarks was waiting for vblank, when it was called from
haswell_crtc_enable function. This was leading to "vblank not available
on crtc " warning (since vblank int was not yet enabled on that CRTC).
But as I can see we now no longer use wait for vblank during DDB writing
if it's called from crtc_enable function after Paul's patch series, I
can drop this patch all together.
thanks for review & input
Regards,
-Mahesh
On Wednesday 31 August 2016 04:21 PM, Maarten Lankhorst wrote:
more state needs to be pre-calculated, rather than touching intel_crtc->active.
Something like using the .initial_watermarks callback, we're so close to removing
that non-atomic flag altogether.:-)
~Maarten
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
