[Intel-gfx] [PATCH] drm/i915: Force initial atomic check in all eDP panels
Imre Deak
imre.deak at intel.com
Thu Oct 29 16:12:34 UTC 2020
On Wed, Oct 28, 2020 at 02:07:12PM -0700, José Roberto de Souza wrote:
> After commit 00e5deb5c4f5 ("drm/i915: Fix encoder lookup during PSR
> atomic check") dig_port was not being used but while fixing it I
> realized that would be better to mark all CRTCs that has a eDP
> connector as needing to have their state computed.
> The principal reason is that in future we will support PSR in
> multiple panels.
> And this is only forcing the state compute if no register change is
> need our atomic handling will just ignore this CRTC + connector
> during the atomic commit phase.
>
> Cc: Imre Deak <imre.deak at intel.com>
> Signed-off-by: José Roberto de Souza <jose.souza at intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_psr.c | 10 ++--------
> 1 file changed, 2 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
> index 1576c3722d0b..b5441f0b5b58 100644
> --- a/drivers/gpu/drm/i915/display/intel_psr.c
> +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> @@ -1875,17 +1875,11 @@ void intel_psr_atomic_check(struct drm_connector *connector,
> struct drm_connector_state *new_state)
> {
> struct drm_i915_private *dev_priv = to_i915(connector->dev);
> - struct intel_connector *intel_connector;
Arg, compiler didn't warn for this.
> - struct intel_digital_port *dig_port;
> struct drm_crtc_state *crtc_state;
>
> if (!CAN_PSR(dev_priv) || !new_state->crtc ||
> - !dev_priv->psr.force_mode_changed)
> - return;
> -
> - intel_connector = to_intel_connector(connector);
> - dig_port = enc_to_dig_port(to_intel_encoder(new_state->best_encoder));
> - if (dev_priv->psr.dp != &dig_port->dp)
> + !dev_priv->psr.force_mode_changed ||
> + connector->connector_type != DRM_MODE_CONNECTOR_eDP)
> return;
Can't you simplify even more by an is_edp && CAN_PSR check in
intel_dp_initial_fastset_check() instead of the psr.force_mode_changed
logic?
>
> crtc_state = drm_atomic_get_new_crtc_state(new_state->state,
> --
> 2.29.1
>
More information about the Intel-gfx
mailing list