Re: [PATCH 1/5] drm/i915: Update mbus in intel_dbuf_mbus_update and do it properly

From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Cc: intel-gfx@lists.freedesktop.org, jani.saarinen@intel.com
Subject: Re: [PATCH 1/5] drm/i915: Update mbus in intel_dbuf_mbus_update and do it properly
Date: Fri, 22 Mar 2024 19:46:16 +0200	[thread overview]
Message-ID: <Zf3D6IsC1iubhYjJ@intel.com> (raw)
In-Reply-To: <20240322114046.24930-2-stanislav.lisovskiy@intel.com>

On Fri, Mar 22, 2024 at 01:40:42PM +0200, Stanislav Lisovskiy wrote:
> According to BSpec we need to do correspondent MBUS updates before
> or after DBUF reallocation, depending on whether we are enabling
> or disabling mbus joining(typical scenario is swithing between
> multiple and single displays).
> 
> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
> ---
>  drivers/gpu/drm/i915/display/skl_watermark.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/skl_watermark.c b/drivers/gpu/drm/i915/display/skl_watermark.c
> index bc341abcab2fe..8ff69da664807 100644
> --- a/drivers/gpu/drm/i915/display/skl_watermark.c
> +++ b/drivers/gpu/drm/i915/display/skl_watermark.c
> @@ -3574,7 +3574,7 @@ void intel_dbuf_mdclk_cdclk_ratio_update(struct drm_i915_private *i915, u8 ratio
>   * Configure MBUS_CTL and all DBUF_CTL_S of each slice to join_mbus state before
>   * update the request state of all DBUS slices.
>   */
> -static void update_mbus_pre_enable(struct intel_atomic_state *state)
> +static void intel_dbuf_mbus_update(struct intel_atomic_state *state)
>  {
>  	struct drm_i915_private *i915 = to_i915(state->base.dev);
>  	u32 mbus_ctl;
> @@ -3632,7 +3632,9 @@ void intel_dbuf_pre_plane_update(struct intel_atomic_state *state)
>  
>  	WARN_ON(!new_dbuf_state->base.changed);
>  
> -	update_mbus_pre_enable(state);
> +	if (!old_dbuf_state->joined_mbus && new_dbuf_state->joined_mbus)
> +		intel_dbuf_mbus_update(state);
> +

Does this acutally do something sensible on its own?
If not I'd just squash this into the other patch.

>  	gen9_dbuf_slices_update(i915,
>  				old_dbuf_state->enabled_slices |
>  				new_dbuf_state->enabled_slices);
> @@ -3653,6 +3655,9 @@ void intel_dbuf_post_plane_update(struct intel_atomic_state *state)
>  
>  	WARN_ON(!new_dbuf_state->base.changed);
>  
> +	if (old_dbuf_state->joined_mbus && !new_dbuf_state->joined_mbus)
> +		intel_dbuf_mbus_update(state);
> +
>  	gen9_dbuf_slices_update(i915,
>  				new_dbuf_state->enabled_slices);
>  }
> -- 
> 2.37.3

-- 
Ville Syrjälä
Intel