[Intel-gfx] [PATCH 07/17] drm/i915/icl: Fail flip if ddb allocated are less than min display buffer needed

[Paulo Zanoni]

Em Sex, 2018-01-26 às 15:50 -0800, James Ausmus escreveu:
> On Tue, Jan 23, 2018 at 05:05:26PM -0200, Paulo Zanoni wrote:
> > From: Mahesh Kumar <mahesh1.kumar at intel.com>
> > 
> > ICL require DDB allocation of plane to be more than "minimum
> > display
> > buffer needed" for each level in order to enable WM level.
> > 
> > This patch implements and consider the same while allocating DDB
> > and enabling WM.
> > 
> > Changes Since V1:
> >  - rebase
> > Changes Since V2:
> >  - Remove extra parentheses
> >  - Use FP16.16 only when absolutely necessary (Paulo)
> > Changes Since V3:
> >  - Rebase
> > Changes since v4 (from Paulo)
> >  - Coding style issue.
> > 
> > Reviewed-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > Signed-off-by: Mahesh Kumar <mahesh1.kumar at intel.com>
> > Signed-off-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
> 
> Reviewed-by: James Ausmus <james.ausmus at intel.com>
> 
> > ---
> >  drivers/gpu/drm/i915/intel_pm.c | 30
> > +++++++++++++++++++++++++++++-
> >  1 file changed, 29 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_pm.c
> > b/drivers/gpu/drm/i915/intel_pm.c
> > index 44d952a3d9a6..c6d31a5075ad 100644
> > --- a/drivers/gpu/drm/i915/intel_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > @@ -4510,6 +4510,7 @@ static int skl_compute_plane_wm(const struct
> > drm_i915_private *dev_priv,
> >  	struct intel_atomic_state *state =
> >  		to_intel_atomic_state(cstate->base.state);
> >  	bool apply_memory_bw_wa = skl_needs_memory_bw_wa(state);
> > +	uint32_t min_disp_buf_needed;
> >  
> >  	if (latency == 0 ||
> >  	    !intel_wm_plane_visible(cstate, intel_pstate)) {
> > @@ -4568,7 +4569,34 @@ static int skl_compute_plane_wm(const struct
> > drm_i915_private *dev_priv,
> >  		}
> >  	}
> >  
> > -	if (res_blocks >= ddb_allocation || res_lines > 31) {
> > +	if (INTEL_GEN(dev_priv) >= 11) {
> > +		if (wp->y_tiled) {
> > +			uint32_t extra_lines;
> > +			uint_fixed_16_16_t fp_min_disp_buf_needed;
> > +
> > +			if (res_lines % wp->y_min_scanlines == 0)
> > +				extra_lines = wp->y_min_scanlines;
> > +			else
> > +				extra_lines = wp->y_min_scanlines
> > * 2 -
> > +					      res_lines % wp-
> > >y_min_scanlines;
> > +
> > +			fp_min_disp_buf_needed =
> > mul_u32_fixed16(res_lines +
> > +						extra_lines,
> > +						wp-
> > >plane_blocks_per_line);
> > +			min_disp_buf_needed =
> > fixed16_to_u32_round_up(
> > +						fp_min_disp_buf_ne
> > eded);
> > +		} else {
> > +			min_disp_buf_needed =
> > DIV_ROUND_UP(res_blocks * 11, 10);
> > +		}
> > +	} else {
> > +		/*
> > +		 * To enable a WM level ddb_allocation should be
> > +		 * greater than result blocks for GEN-9/10.
> > +		 */
> > +		min_disp_buf_needed = res_blocks + 1;
> > +	}
> > +
> > +	if (min_disp_buf_needed > ddb_allocation || res_lines > 

BSpec says that if min_disp_buf_needed == ddb_allocation we should also
fail, as weird as it sounds.


> > 31) {
> >  		*enabled = false;
> >  
> >  		/*
> > -- 
> > 2.14.3
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx