[Intel-gfx] [PATCH 3/4] drm/i915: Use wm0 only during async flips for DG2

[Ville Syrjälä]

On Fri, Jan 21, 2022 at 10:06:14AM +0200, Stanislav Lisovskiy wrote:
> This optimization allows to achieve higher perfomance
> during async flips.
> For the first async flip we have to still temporarily
> switch to sync flip, in order to reprogram plane
> watermarks, so this requires taking into account
> old plane state's do_async_flip flag.
> 
> v2: - Removed redundant new_plane_state->do_async_flip
>       check from needs_async_flip_wm_override condition
>       (Ville Syrjälä)
>     - Extract dg2_async_flip_optimization to separate
>       function(Ville Syrjälä)
>     - Check for plane->async_flip instead of plane_id
>       (Ville Syrjälä)
> 
> v3: - Rename "needs_async_flip_wm_override" to
>       "intel_plane_do_async_flip" and move all the required
>       checks there (Ville Syrjälä)
>     - Rename "dg2_async_flip_optimization" to
>       "use_minimal_wm0_only" (Ville Syrjälä)
> 
> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy at intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_display.c | 24 +++++++++++++++++++-
>  drivers/gpu/drm/i915/intel_pm.c              | 12 +++++++++-
>  2 files changed, 34 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> index 9996daa036a0..3b86ede01b57 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.c
> +++ b/drivers/gpu/drm/i915/display/intel_display.c
> @@ -4910,6 +4910,28 @@ static bool needs_scaling(const struct intel_plane_state *state)
>  	return (src_w != dst_w || src_h != dst_h);
>  }
>  
> +static bool intel_plane_do_async_flip(struct intel_plane *plane,
> +				      const struct intel_crtc_state *new_crtc_state,
> +				      const struct intel_crtc_state *old_crtc_state)

I think typically we put the old state before the new state.
Sadly the compiler can't help us with these so we should try
to be consistent to avoid accidental mishaps.

> +{
> +	struct drm_i915_private *i915 = to_i915(new_crtc_state->uapi.crtc->dev);

Would be a bit shorter to grab this from plane->base.dev

> +
> +	if (!plane->async_flip)
> +		return false;
> +
> +	if (!new_crtc_state->uapi.async_flip)
> +		return false;
> +
> +	/*
> +	 * In platforms after DISPLAY13, we might need to override
> +	 * first async flip in order to change watermark levels
> +	 * as part of optimization.
> +	 * So for those, we are checking if this is a first async flip.
> +	 * For platforms earlier than DISPLAY13 we always do async flip.
> +	 */
> +	return DISPLAY_VER(i915) < 13 || old_crtc_state->uapi.async_flip;
> +}
> +
>  int intel_plane_atomic_calc_changes(const struct intel_crtc_state *old_crtc_state,
>  				    struct intel_crtc_state *new_crtc_state,
>  				    const struct intel_plane_state *old_plane_state,
> @@ -5029,7 +5051,7 @@ int intel_plane_atomic_calc_changes(const struct intel_crtc_state *old_crtc_stat
>  			 needs_scaling(new_plane_state))))
>  		new_crtc_state->disable_lp_wm = true;
>  
> -	if (new_crtc_state->uapi.async_flip && plane->async_flip)
> +	if (intel_plane_do_async_flip(plane, new_crtc_state, old_crtc_state))
>  		new_plane_state->do_async_flip = true;
>  
>  	return 0;
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 35d0bd8c6e57..5fb022a2a4d7 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -5510,6 +5510,15 @@ static int skl_wm_max_lines(struct drm_i915_private *dev_priv)
>  		return 31;
>  }
>  
> +static bool use_minimal_wm0_only(struct drm_i915_private *i915,

We can dig out 'i915' from eg. the plane, so no need for the
caller to pass it in.

> +				 const struct intel_crtc_state *crtc_state,
> +				 const struct intel_plane *plane)
> +{

Atypical 'const' still on the plane pointer here.

Apart from those lgtm
Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com>

> +	return DISPLAY_VER(i915) >= 13 &&
> +	       crtc_state->uapi.async_flip &&
> +	       plane->async_flip;
> +}
> +
>  static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
>  				 const struct intel_plane *plane,
>  				 int level,
> @@ -5523,7 +5532,8 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
>  	uint_fixed_16_16_t selected_result;
>  	u32 blocks, lines, min_ddb_alloc = 0;
>  
> -	if (latency == 0) {
> +	if (latency == 0 ||
> +	    (use_minimal_wm0_only(dev_priv, crtc_state, plane) && level > 0)) {
>  		/* reject it */
>  		result->min_ddb_alloc = U16_MAX;
>  		return;
> -- 
> 2.24.1.485.gad05a3d8e5

-- 
Ville Syrjälä
Intel