On Mon, Oct 05, 2020 at 11:46:17PM +0300, Imre Deak wrote: > On Mon, Oct 05, 2020 at 11:30:55PM +0300, Ville Syrjälä wrote: > > On Sat, Oct 03, 2020 at 03:18:45AM +0300, Imre Deak wrote: > > > Atm, if a full modeset is performed during the initial modeset the link > > > training will happen with uninitialized max DP rate and lane count. Make > > > sure the corresponding encoder state is initialized by adding an encoder > > > hook called during driver init and system resume. > > > > > > A better alternative would be to store all states in the CRTC state and > > > make this state available for the link re-training code. Also instead of > > > the DPCD read in the hook there should be really a proper sink HW > > > readout in place. Both of these require a bigger rework, so for now opting > > > for this minimal fix to make at least full initial modesets work. > > > > > > The patch is based on > > > https://patchwork.freedesktop.org/patch/101473/?series=10354&rev=3 > > > > > > Cc: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > > Signed-off-by: Imre Deak <imre.deak@xxxxxxxxx> > > > --- > > > drivers/gpu/drm/i915/display/intel_ddi.c | 8 +++++ > > > drivers/gpu/drm/i915/display/intel_display.c | 4 +++ > > > .../drm/i915/display/intel_display_types.h | 7 +++++ > > > drivers/gpu/drm/i915/display/intel_dp.c | 31 +++++++++++++++++++ > > > drivers/gpu/drm/i915/display/intel_dp.h | 2 ++ > > > drivers/gpu/drm/i915/display/intel_dp_mst.c | 10 ++++++ > > > 6 files changed, 62 insertions(+) > > > > > > diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c > > > index 4e54c55ec99f..a0805260b224 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_ddi.c > > > +++ b/drivers/gpu/drm/i915/display/intel_ddi.c > > > @@ -4564,6 +4564,13 @@ void intel_ddi_get_config(struct intel_encoder *encoder, > > > intel_read_dp_sdp(encoder, pipe_config, DP_SDP_VSC); > > > } > > > > > > +static void intel_ddi_sanitize_state(struct intel_encoder *encoder, > > > + const struct intel_crtc_state *crtc_state) > > > +{ > > > + if (crtc_state && intel_crtc_has_dp_encoder(crtc_state)) > > > + intel_dp_sanitize_state(encoder, crtc_state); > > > +} > > > > I think we usually use 'sanitize' to mean "hw state is garbage -> must > > take steps to sanitize it". This one is just filling in our intel_dp > > sidechannel state. So the name isn't super consistnet with existing > > practies. > > It is called during init/resume time when encoders are sanitized as > well, but yea it's a separate step from HW readout. So I can rename it > for instance (back) to sync_state, or any better idea? All I know is that I suck at naming things. > > > > > > + > > > static bool intel_ddi_initial_fastset_check(struct intel_encoder *encoder, > > > struct intel_crtc_state *crtc_state) > > > { > > > @@ -5182,6 +5189,7 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port) > > > encoder->update_pipe = intel_ddi_update_pipe; > > > encoder->get_hw_state = intel_ddi_get_hw_state; > > > encoder->get_config = intel_ddi_get_config; > > > + encoder->sanitize_state = intel_ddi_sanitize_state; > > > encoder->initial_fastset_check = intel_ddi_initial_fastset_check; > > > encoder->suspend = intel_dp_encoder_suspend; > > > encoder->get_power_domains = intel_ddi_get_power_domains; > > > diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c > > > index 31be63225b10..e61311ee8b8c 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_display.c > > > +++ b/drivers/gpu/drm/i915/display/intel_display.c > > > @@ -18725,8 +18725,12 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) > > > > > > encoder->base.crtc = &crtc->base; > > > encoder->get_config(encoder, crtc_state); > > > + if (encoder->sanitize_state) > > > + encoder->sanitize_state(encoder, crtc_state); > > > } else { > > > encoder->base.crtc = NULL; > > > + if (encoder->sanitize_state) > > > + encoder->sanitize_state(encoder, NULL); > > > > I wonder if we should even bother calling it in this case. > > Yes, it would be just a nop atm, and can't think what state would need > to be updated, so will remove it. > > > > > > } > > > > > > drm_dbg_kms(&dev_priv->drm, > > > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h > > > index 5297b2f08ff9..b2b458144f5a 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_display_types.h > > > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h > > > @@ -188,6 +188,13 @@ struct intel_encoder { > > > void (*get_config)(struct intel_encoder *, > > > struct intel_crtc_state *pipe_config); > > > > > > + /* > > > + * Optional hook called during init/resume to sanitize any state > > > + * stored in the encoder (eg. DP link parameters). > > > + */ > > > + void (*sanitize_state)(struct intel_encoder *encoder, > > > + const struct intel_crtc_state *crtc_state); > > > + > > > /* > > > * Optional hook, returning true if this encoder allows a fastset > > > * during the initial commit, false otherwise. > > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c > > > index df5277c2b9ba..9b6fe3b3b5b2 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_dp.c > > > +++ b/drivers/gpu/drm/i915/display/intel_dp.c > > > @@ -3703,6 +3703,36 @@ static void intel_dp_get_config(struct intel_encoder *encoder, > > > } > > > } > > > > > > +static bool > > > +intel_dp_get_dpcd(struct intel_dp *intel_dp); > > > + > > > +/** > > > + * intel_dp_sanitize_state - sanitize the encoder state during init/resume > > > + * @encoder: intel encoder to sanitize > > > + * @crtc_state: state for the CRTC connected to the encoder > > > + * > > > + * Sanitize any state stored in the encoder during driver init and system > > > + * resume. > > > + */ > > > +void intel_dp_sanitize_state(struct intel_encoder *encoder, > > > + const struct intel_crtc_state *crtc_state) > > > +{ > > > + struct intel_dp *intel_dp = enc_to_intel_dp(encoder); > > > + > > > + if (!crtc_state) > > > + return; > > > + > > > + /* > > > + * Don't clobber DPCD if it's been already read out during output > > > + * setup (eDP) or detect. > > > + */ > > > + if (!memchr_inv(intel_dp->dpcd, 0, sizeof(intel_dp->dpcd))) > > > + intel_dp_get_dpcd(intel_dp); > > > + > > > + intel_dp->max_link_lane_count = intel_dp_max_common_lane_count(intel_dp); > > > + intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp); > > > +} > > > + > > > bool intel_dp_initial_fastset_check(struct intel_encoder *encoder, > > > struct intel_crtc_state *crtc_state) > > > { > > > @@ -8090,6 +8120,7 @@ bool intel_dp_init(struct drm_i915_private *dev_priv, > > > intel_encoder->compute_config = intel_dp_compute_config; > > > intel_encoder->get_hw_state = intel_dp_get_hw_state; > > > intel_encoder->get_config = intel_dp_get_config; > > > + intel_encoder->sanitize_state = intel_dp_sanitize_state; > > > intel_encoder->initial_fastset_check = intel_dp_initial_fastset_check; > > > intel_encoder->update_pipe = intel_panel_update_backlight; > > > intel_encoder->suspend = intel_dp_encoder_suspend; > > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h > > > index 977585aea3c8..1ab741e0be67 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_dp.h > > > +++ b/drivers/gpu/drm/i915/display/intel_dp.h > > > @@ -143,5 +143,7 @@ int intel_dp_init_hdcp(struct intel_digital_port *dig_port, > > > > > > bool intel_dp_initial_fastset_check(struct intel_encoder *encoder, > > > struct intel_crtc_state *crtc_state); > > > +void intel_dp_sanitize_state(struct intel_encoder *encoder, > > > + const struct intel_crtc_state *crtc_state); > > > > > > #endif /* __INTEL_DP_H__ */ > > > diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.c b/drivers/gpu/drm/i915/display/intel_dp_mst.c > > > index e948aacbd4ab..0831d1ee7978 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_dp_mst.c > > > +++ b/drivers/gpu/drm/i915/display/intel_dp_mst.c > > > @@ -591,6 +591,15 @@ static void intel_dp_mst_enc_get_config(struct intel_encoder *encoder, > > > intel_ddi_get_config(&dig_port->base, pipe_config); > > > } > > > > > > +static void intel_dp_mst_sync_state(struct intel_encoder *encoder, > > > + const struct intel_crtc_state *crtc_state) > > > +{ > > > + struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder); > > > + struct intel_digital_port *dig_port = intel_mst->primary; > > > + > > > + return intel_dp_sanitize_state(&dig_port->base, crtc_state); > > > +} > > > + > > > static bool intel_dp_mst_initial_fastset_check(struct intel_encoder *encoder, > > > struct intel_crtc_state *crtc_state) > > > { > > > @@ -906,6 +915,7 @@ intel_dp_create_fake_mst_encoder(struct intel_digital_port *dig_port, enum pipe > > > intel_encoder->enable = intel_mst_enable_dp; > > > intel_encoder->get_hw_state = intel_dp_mst_enc_get_hw_state; > > > intel_encoder->get_config = intel_dp_mst_enc_get_config; > > > + intel_encoder->sanitize_state = intel_dp_mst_sync_state; > > > intel_encoder->initial_fastset_check = intel_dp_mst_initial_fastset_check; > > > > > > return intel_mst; > > > -- > > > 2.25.1 > > > > -- > > Ville Syrjälä > > Intel -- Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
