Re: [Intel-gfx] [PATCH v3 3/4] drm/i915/psr: Check that vblank is long enough for psr2

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Mon, Mar 20, 2023 at 06:59:44PM +0200, Jouni Högander wrote:
> Ensure vblank >= psr2 vblank
> where
> Psr2 vblank = PSR2_CTL Block Count Number maximum line count.
> 
> Bspec: 71580, 49274
> 
> Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_psr.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
> index 1050d777a108..1b40d9c73c18 100644
> --- a/drivers/gpu/drm/i915/display/intel_psr.c
> +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> @@ -958,6 +958,14 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp,
>  		return false;
>  	}
>  
> +	/* Vblank >= PSR2_CTL Block Count Number maximum line count */
> +	if (crtc_state->hw.adjusted_mode.crtc_vblank_end -
> +	    crtc_state->hw.adjusted_mode.crtc_vblank_start < 12) {

Why 12? Shouldn't it be based on the wake_lines/BLOCK_COUNT_NUM stuff?


If so I would suggest we try someting like this:

psr2_block_count_lines()
{
	return ...wake_lines... ? 12 : 8;
}

psr2_block_count()
{
	return psr2_block_count_lines() / 4;
}

if (vblank_lengh < psr2_block_count_lines())
	fail;

if (psr_block_count() > 2)
	val |= BLOCK_COUNT_NUM_3;
else
	val |= BLOCK_COUNT_NUM_2;

> +		drm_dbg_kms(&dev_priv->drm,
> +			    "PSR2 not enabled, too short vblank time\n");
> +		return false;
> +	}
> +
>  	if (HAS_PSR2_SEL_FETCH(dev_priv)) {
>  		if (!intel_psr2_sel_fetch_config_valid(intel_dp, crtc_state) &&
>  		    !HAS_PSR_HW_TRACKING(dev_priv)) {
> -- 
> 2.34.1

-- 
Ville Syrjälä
Intel