On Tue, Jan 28, 2020 at 03:52:40PM -0800, José Roberto de Souza wrote:
> dGFX have local memory so it do not have aperture and do not support
> CPU fences but even for iGFX it have a small number of fences.
>
> As replacement for fences to track frontbuffer modifications by CPU
> we have a software tracking that is already in used by FBC and PSR.
> PSR don't support fences so it shows that this tracking is reliable.
>
> So lets make fences a nice-to-have to activate FBC for GEN9+(as we
> only have a good CI coverage for GEN9+), this will allow us to enable
> FBC for dGFXs and iGFXs even when there is no available fence.
>
> intel_fbc_hw_tracking_covers_screen() maybe can also have the same
> treatment as fences but BSpec is not clear if the size limitation is
> for hardware tracking or general use of FBC and I don't have a 5K
> display to test it, so keeping as is for safety.
>
> Cc: Daniel Vetter <daniel.vet...@intel.com>
> Cc: Dhinakaran Pandiyan <dhinakaran.pandi...@intel.com>
> Signed-off-by: José Roberto de Souza <jose.so...@intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_fbc.c | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c
> b/drivers/gpu/drm/i915/display/intel_fbc.c
> index 2a3f1333c8ff..1f0d24a1dec1 100644
> --- a/drivers/gpu/drm/i915/display/intel_fbc.c
> +++ b/drivers/gpu/drm/i915/display/intel_fbc.c
> @@ -717,11 +717,15 @@ static bool intel_fbc_can_activate(struct intel_crtc
> *crtc)
> return false;
> }
>
> - /* The use of a CPU fence is mandatory in order to detect writes
> - * by the CPU to the scanout and trigger updates to the FBC.
> + /* The use of a CPU fence is one of two ways to detect writes by the
> + * CPU to the scanout and trigger updates to the FBC.
> + *
> + * The other method is by software tracking(see
> + * intel_fbc_invalidate/flush()), it will manually notify FBC and nuke
> + * the current compressed buffer and recompress it.
> *
> * Note that is possible for a tiled surface to be unmappable (and
> - * so have no fence associated with it) due to aperture constaints
> + * so have no fence associated with it) due to aperture constraints
> * at the time of pinning.
> *
> * FIXME with 90/270 degree rotation we should use the fence on
> @@ -730,7 +734,7 @@ static bool intel_fbc_can_activate(struct intel_crtc
> *crtc)
> * For now this will effecively disable FBC with 90/270 degree
> * rotation.
> */
> - if (cache->fence_id < 0) {
> + if (cache->fence_id < 0 && INTEL_GEN(dev_priv) < 9) {
Not enough. We need to check that the tiling format is actually supported.
Also the tracking stuff is busted in intel ddx so need to get
https://patchwork.freedesktop.org/series/70636/ merged somehow.
> fbc->no_fbc_reason = "framebuffer not tiled or fenced";
> return false;
> }
> --
> 2.25.0
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
