[Intel-gfx] [PATCH] drm/i915: Fix S0ix/S3 suspend stress issue

Souza, Jose jose.souza at intel.com
Wed Sep 18 17:43:56 UTC 2019


On Wed, 2019-09-18 at 11:23 +0530, Gaurav K Singh wrote:
> During S0ix/S3 suspend stress test on Cometlake chromebook,
> after few iterations we are seeing failure wrt PSR link CRC
> Error and stress test stops. This S0ix test is failing only
> when there is a CRC mismatch. In case of CRC mismatch, panel
> generates IRQ_HD and whenever there is CRC mismatch, we are
> disabling PSR2 in driver.
> 
> By not disabling PSR2 in driver and only by writing 1 to clear
> sticky bit 0 in DPCD 0x2006 in panel,issue goes away.
> Completed 2500 S0ix/S3 test cycles on multiple CML chromebooks.
> 
> As per EDP spec for CRC mismatch, nothing has been mentioned
> explicitly for Source device, only by writing 1 to clear
> sticky bit 0 in DPCD 0x2006 in sink is mentioned.
> 
> Signed-off-by: Gaurav K Singh <gaurav.k.singh at intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_psr.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_psr.c
> b/drivers/gpu/drm/i915/display/intel_psr.c
> index b3c7eef53bf3..502e29dbbea9 100644
> --- a/drivers/gpu/drm/i915/display/intel_psr.c
> +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> @@ -1325,15 +1325,11 @@ void intel_psr_short_pulse(struct intel_dp
> *intel_dp)
>  	if (val & DP_PSR_VSC_SDP_UNCORRECTABLE_ERROR)
>  		DRM_DEBUG_KMS("PSR VSC SDP uncorrectable error,
> disabling PSR\n");
>  	if (val & DP_PSR_LINK_CRC_ERROR)
> -		DRM_ERROR("PSR Link CRC error, disabling PSR\n");
> +		DRM_DEBUG_KMS("PSR Link CRC error, clearing PSR error
> status DPCD\n");
>  
>  	if (val & ~errors)
>  		DRM_ERROR("PSR_ERROR_STATUS unhandled errors %x\n",
>  			  val & ~errors);
> -	if (val & errors) {
> -		intel_psr_disable_locked(intel_dp);
> -		psr->sink_not_reliable = true;
> -	}

With this change you are breaking CRC and all the other error handling.

If CRC did not match, it means the image that was received by sink do
not match the expected, it could happen because of problems on sink,
source or flat cable.

It is better have perfect frames than save power, so this is not
acceptable.

>  	/* clear status register */
>  	drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_ERROR_STATUS, val);
>  exit:


More information about the Intel-gfx mailing list