[Intel-gfx] [PATCH v11 06/12] drm/i915: Try to make bigjoiner work in atomic check

Ville Syrjälä ville.syrjala at linux.intel.com
Fri Oct 23 18:30:56 UTC 2020


On Fri, Oct 23, 2020 at 11:13:47AM -0700, Navare, Manasi wrote:
> On Fri, Oct 23, 2020 at 08:42:00PM +0300, Ville Syrjälä wrote:
> > On Wed, Oct 21, 2020 at 10:42:17PM -0700, Manasi Navare wrote:
> > > From: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> > > 
> > >  When the clock is higher than the dotclock, try with 2 pipes enabled.
> > >  If we can enable 2, then we will go into big joiner mode, and steal
> > >  the adjacent crtc.
> > > 
> > >  This only links the crtc's in software, no hardware or plane
> > >  programming is done yet. Blobs are also copied from the master's
> > >  crtc_state, so it doesn't depend at commit time on the other
> > >  crtc_state.
> > > 
> > > v4:
> > > * Fixes in intel_crtc_compute_config (Ville)
> > > v3:
> > > * Manual Rebase (Manasi)
> > >  Changes since v1:
> > >  - Rename pipe timings to transcoder timings, as they are now different.
> > >   Changes since v2:
> > >  - Rework bigjoiner checks; always disable slave when recalculating
> > >    master. No need to have a separate bigjoiner pass any more.
> > >  - Use pipe_mode instead of transcoder_mode, to clean up the code.
> > > 
> > > Signed-off-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> > > Signed-off-by: Manasi Navare <manasi.d.navare at intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_display.c  | 181 ++++++++++++++++--
> > >  .../drm/i915/display/intel_display_types.h    |   9 +
> > >  drivers/gpu/drm/i915/display/intel_dp.c       |  22 ++-
> > >  3 files changed, 191 insertions(+), 21 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> > > index 92cd38964142..ffee6afe4ddb 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_display.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_display.c
> > > @@ -8237,9 +8237,24 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
> > >  				     struct intel_crtc_state *pipe_config)
> > >  {
> > >  	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
> > > -	const struct drm_display_mode *pipe_mode = &pipe_config->hw.pipe_mode;
> > > +	struct drm_display_mode *pipe_mode = &pipe_config->hw.pipe_mode;
> > >  	int clock_limit = dev_priv->max_dotclk_freq;
> > >  
> > > +	*pipe_mode = pipe_config->hw.adjusted_mode;
> > > +
> > > +	/* Adjust pipe_mode for bigjoiner, with half the horizontal mode */
> > > +	if (pipe_config->bigjoiner) {
> > > +		pipe_mode->crtc_clock /= 2;
> > > +		pipe_mode->crtc_hdisplay /= 2;
> > > +		pipe_mode->crtc_hblank_start /= 2;
> > > +		pipe_mode->crtc_hblank_end /= 2;
> > > +		pipe_mode->crtc_hsync_start /= 2;
> > > +		pipe_mode->crtc_hsync_end /= 2;
> > > +		pipe_mode->crtc_htotal /= 2;
> > > +		pipe_mode->crtc_hskew /= 2;
> > > +		pipe_config->pipe_src_w /= 2;
> > > +	}
> > > +
> > >  	if (INTEL_GEN(dev_priv) < 4) {
> > >  		clock_limit = dev_priv->max_cdclk_freq * 9 / 10;
> > >  
> > > @@ -8300,7 +8315,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
> > >  	 * WaPruneModeWithIncorrectHsyncOffset:ctg,elk,ilk,snb,ivb,vlv,hsw.
> > >  	 */
> > >  	if ((INTEL_GEN(dev_priv) > 4 || IS_G4X(dev_priv)) &&
> > > -		pipe_mode->crtc_hsync_start == pipe_mode->crtc_hdisplay)
> > > +	    pipe_mode->crtc_hsync_start == pipe_mode->crtc_hdisplay)
> > >  		return -EINVAL;
> > >  
> > >  	intel_crtc_compute_pixel_rate(pipe_config);
> > > @@ -12657,7 +12672,7 @@ static bool encoders_cloneable(const struct intel_encoder *a,
> > >  			  b->cloneable & (1 << a->type));
> > >  }
> > >  
> > > -static bool check_single_encoder_cloning(struct drm_atomic_state *state,
> > > +static bool check_single_encoder_cloning(struct intel_atomic_state *state,
> > >  					 struct intel_crtc *crtc,
> > >  					 struct intel_encoder *encoder)
> > >  {
> > > @@ -12666,7 +12681,7 @@ static bool check_single_encoder_cloning(struct drm_atomic_state *state,
> > >  	struct drm_connector_state *connector_state;
> > >  	int i;
> > >  
> > > -	for_each_new_connector_in_state(state, connector, connector_state, i) {
> > > +	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
> > 
> > There's still this extra noise in the patch. I thought you were going
> > to split out the calling convention changes to reduce it?
> 
> Hmm I split the calling convention changes that were adding from_crtc_state into a separate patch
> May be I will split this one out too in a diff patch?

Go through all the patches, read them all while keeping the
single thought "which parts can be split out?" foremost in
our mind. When you no longer find any parts that can be split
out then the patch is ready, not before.

> 
> 
> > 
> > >  		if (connector_state->crtc != &crtc->base)
> > >  			continue;
> > >  
> > > @@ -12906,6 +12921,7 @@ static int intel_crtc_atomic_check(struct intel_atomic_state *state,
> > >  
> > >  	if (mode_changed && crtc_state->hw.enable &&
> > >  	    dev_priv->display.crtc_compute_clock &&
> > > +	    !crtc_state->bigjoiner_slave &&
> > >  	    !drm_WARN_ON(&dev_priv->drm, crtc_state->shared_dpll)) {
> > >  		ret = dev_priv->display.crtc_compute_clock(crtc, crtc_state);
> > >  		if (ret)
> > > @@ -13447,6 +13463,15 @@ intel_crtc_copy_uapi_to_hw_state_nomodeset(struct intel_atomic_state *state,
> > >  {
> > >  	const struct intel_crtc_state *from_crtc_state = crtc_state;
> > >  
> > > +	if (crtc_state->bigjoiner_slave) {
> > > +		from_crtc_state = intel_atomic_get_new_crtc_state(state,
> > > +								  crtc_state->bigjoiner_linked_crtc);
> > > +
> > > +		/* No need to copy state if the master state is unchanged */
> > > +		if (!from_crtc_state)
> > > +			return;
> > > +	}
> > > +
> > >  	intel_crtc_copy_color_blobs(crtc_state, from_crtc_state);
> > >  }
> > >  
> > > @@ -13482,6 +13507,47 @@ static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state
> > >  				  crtc_state->hw.ctm);
> > >  }
> > >  
> > > +static int
> > > +copy_bigjoiner_crtc_state(struct intel_crtc_state *crtc_state,
> > > +			  const struct intel_crtc_state *from_crtc_state)
> > > +{
> > > +	struct intel_crtc_state *saved_state;
> > > +	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
> > > +
> > > +	saved_state = kmemdup(from_crtc_state, sizeof(*saved_state), GFP_KERNEL);
> > > +	if (!saved_state)
> > > +		return -ENOMEM;
> > > +
> > > +	saved_state->uapi = crtc_state->uapi;
> > > +	saved_state->scaler_state = crtc_state->scaler_state;
> > > +	saved_state->shared_dpll = crtc_state->shared_dpll;
> > > +	saved_state->dpll_hw_state = crtc_state->dpll_hw_state;
> > > +	saved_state->crc_enabled = crtc_state->crc_enabled;
> > > +
> > > +	intel_crtc_free_hw_state(crtc_state);
> > > +	memcpy(crtc_state, saved_state, sizeof(*crtc_state));
> > > +	kfree(saved_state);
> > > +
> > > +	/* Re-init hw state */
> > > +	memset(&crtc_state->hw, 0, sizeof(saved_state->hw));
> > > +	crtc_state->hw.enable = from_crtc_state->hw.enable;
> > > +	crtc_state->hw.active = from_crtc_state->hw.active;
> > > +	crtc_state->hw.pipe_mode = from_crtc_state->hw.pipe_mode;
> > > +	crtc_state->hw.adjusted_mode = from_crtc_state->hw.adjusted_mode;
> > > +
> > > +	/* Some fixups */
> > > +	crtc_state->uapi.mode_changed = from_crtc_state->uapi.mode_changed;
> > > +	crtc_state->uapi.connectors_changed = from_crtc_state->uapi.connectors_changed;
> > > +	crtc_state->uapi.active_changed = from_crtc_state->uapi.active_changed;
> > > +	crtc_state->nv12_planes = crtc_state->c8_planes = crtc_state->update_planes = 0;
> > > +	crtc_state->bigjoiner_linked_crtc = to_intel_crtc(from_crtc_state->uapi.crtc);
> > > +	crtc_state->bigjoiner_slave = true;
> > > +	crtc_state->cpu_transcoder = (enum transcoder)crtc->pipe;
> > > +	crtc_state->has_audio = false;
> > > +
> > > +	return 0;
> > > +}
> > > +
> > >  static int
> > >  intel_crtc_prepare_cleared_state(struct intel_atomic_state *state,
> > >  				 struct intel_crtc_state *crtc_state)
> > > @@ -13522,10 +13588,10 @@ intel_crtc_prepare_cleared_state(struct intel_atomic_state *state,
> > >  }
> > >  
> > >  static int
> > > -intel_modeset_pipe_config(struct intel_crtc_state *pipe_config)
> > > +intel_modeset_pipe_config(struct intel_atomic_state *state,
> > > +			  struct intel_crtc_state *pipe_config)
> > >  {
> > >  	struct drm_crtc *crtc = pipe_config->uapi.crtc;
> > > -	struct drm_atomic_state *state = pipe_config->uapi.state;
> > >  	struct drm_i915_private *i915 = to_i915(pipe_config->uapi.crtc->dev);
> > >  	struct drm_connector *connector;
> > >  	struct drm_connector_state *connector_state;
> > > @@ -13567,7 +13633,7 @@ intel_modeset_pipe_config(struct intel_crtc_state *pipe_config)
> > >  			       &pipe_config->pipe_src_w,
> > >  			       &pipe_config->pipe_src_h);
> > >  
> > > -	for_each_new_connector_in_state(state, connector, connector_state, i) {
> > > +	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
> > >  		struct intel_encoder *encoder =
> > >  			to_intel_encoder(connector_state->best_encoder);
> > >  
> > > @@ -13605,7 +13671,7 @@ intel_modeset_pipe_config(struct intel_crtc_state *pipe_config)
> > >  	 * adjust it according to limitations or connector properties, and also
> > >  	 * a chance to reject the mode entirely.
> > >  	 */
> > > -	for_each_new_connector_in_state(state, connector, connector_state, i) {
> > > +	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
> > >  		struct intel_encoder *encoder =
> > >  			to_intel_encoder(connector_state->best_encoder);
> > >  
> > > @@ -13657,9 +13723,6 @@ intel_modeset_pipe_config(struct intel_crtc_state *pipe_config)
> > >  		    "hw max bpp: %i, pipe bpp: %i, dithering: %i\n",
> > >  		    base_bpp, pipe_config->pipe_bpp, pipe_config->dither);
> > >  
> > > -	/* without bigjoiner, pipe_mode == adjusted_mode */
> > > -	pipe_config->hw.pipe_mode = pipe_config->hw.adjusted_mode;
> > > -
> > >  	return 0;
> > >  }
> > >  
> > > @@ -15055,6 +15118,75 @@ static bool intel_cpu_transcoders_need_modeset(struct intel_atomic_state *state,
> > >  	return false;
> > >  }
> > >  
> > > +static int intel_atomic_check_bigjoiner(struct intel_atomic_state *state,
> > > +					struct intel_crtc *crtc,
> > > +					struct intel_crtc_state *old_crtc_state,
> > > +					struct intel_crtc_state *new_crtc_state)
> > > +{
> > > +	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
> > > +	struct intel_crtc_state *slave_crtc_state, *master_crtc_state;
> > > +	struct intel_crtc *slave, *master;
> > > +
> > > +	/* slave being enabled, is master is still claiming this crtc? */
> > > +	if (old_crtc_state->bigjoiner_slave) {
> > > +		slave = crtc;
> > > +		master = old_crtc_state->bigjoiner_linked_crtc;
> > > +		master_crtc_state = intel_atomic_get_new_crtc_state(state, master);
> > > +		if (!master_crtc_state || !needs_modeset(master_crtc_state))
> > > +			goto claimed;
> > > +	}
> > > +
> > > +	if (!new_crtc_state->bigjoiner)
> > > +		return 0;
> > > +
> > > +	if (1 + crtc->pipe >= INTEL_NUM_PIPES(dev_priv)) {
> > > +		DRM_DEBUG_KMS("[CRTC:%d:%s] Big joiner configuration requires "
> > > +			      "CRTC + 1 to be used, doesn't exist\n",
> > > +			      crtc->base.base.id, crtc->base.name);
> > > +		return -EINVAL;
> > > +	}
> > > +
> > > +	slave = new_crtc_state->bigjoiner_linked_crtc =
> > > +		intel_get_crtc_for_pipe(dev_priv, crtc->pipe + 1);
> > > +	slave_crtc_state = intel_atomic_get_crtc_state(&state->base, slave);
> > > +	master = crtc;
> > > +	if (IS_ERR(slave_crtc_state))
> > > +		return PTR_ERR(slave_crtc_state);
> > > +
> > > +	/* master being enabled, slave was already configured? */
> > > +	if (slave_crtc_state->uapi.enable)
> > > +		goto claimed;
> > > +
> > > +	DRM_DEBUG_KMS("[CRTC:%d:%s] Used as slave for big joiner\n",
> > > +		      slave->base.base.id, slave->base.name);
> > > +
> > > +	return copy_bigjoiner_crtc_state(slave_crtc_state, new_crtc_state);
> > > +
> > > +claimed:
> > > +	DRM_DEBUG_KMS("[CRTC:%d:%s] Slave is enabled as normal CRTC, but "
> > > +		      "[CRTC:%d:%s] claiming this CRTC for bigjoiner.\n",
> > > +		      slave->base.base.id, slave->base.name,
> > > +		      master->base.base.id, master->base.name);
> > > +	return -EINVAL;
> > > +}
> > > +
> > > +static int kill_bigjoiner_slave(struct intel_atomic_state *state,
> > > +				struct intel_crtc_state *master_crtc_state)
> > > +{
> > > +	struct intel_crtc_state *slave_crtc_state =
> > > +			intel_atomic_get_crtc_state(&state->base,
> > > +						    master_crtc_state->bigjoiner_linked_crtc);
> > > +
> > > +		if (IS_ERR(slave_crtc_state))
> > > +			return PTR_ERR(slave_crtc_state);
> > > +
> > > +		slave_crtc_state->bigjoiner = master_crtc_state->bigjoiner = false;
> > > +		slave_crtc_state->bigjoiner_slave = master_crtc_state->bigjoiner_slave = false;
> > > +		slave_crtc_state->bigjoiner_linked_crtc = master_crtc_state->bigjoiner_linked_crtc = NULL;
> > > +		intel_crtc_copy_uapi_to_hw_state(state, slave_crtc_state);
> > > +		return 0;
> > > +}
> > > +
> > >  /**
> > >   * DOC: asynchronous flip implementation
> > >   *
> > > @@ -15222,14 +15354,31 @@ static int intel_atomic_check(struct drm_device *dev,
> > >  			continue;
> > >  		}
> > >  
> > > +		/* Kill old bigjoiner link, we may re-establish afterwards */
> > > +		if (old_crtc_state->bigjoiner && !old_crtc_state->bigjoiner_slave) {
> > > +			ret = kill_bigjoiner_slave(state, new_crtc_state);
> > > +			if (ret)
> > > +				goto fail;
> > > +		}
> > > +
> > > +		if (!new_crtc_state->uapi.enable) {
> > > +			if (!new_crtc_state->bigjoiner_slave) {
> > > +				intel_crtc_copy_uapi_to_hw_state(state, new_crtc_state);
> > > +				any_ms = true;
> > > +			}
> > > +			continue;
> > > +		}
> > > +
> > >  		ret = intel_crtc_prepare_cleared_state(state, new_crtc_state);
> > >  		if (ret)
> > >  			goto fail;
> > >  
> > > -		if (!new_crtc_state->hw.enable)
> > > -			continue;
> > > +		ret = intel_modeset_pipe_config(state, new_crtc_state);
> > > +		if (ret)
> > > +			goto fail;
> > >  
> > > -		ret = intel_modeset_pipe_config(new_crtc_state);
> > > +		ret = intel_atomic_check_bigjoiner(state, crtc, old_crtc_state,
> > > +						   new_crtc_state);
> > >  		if (ret)
> > >  			goto fail;
> > >  	}
> > > @@ -15567,7 +15716,9 @@ static void intel_update_crtc(struct intel_atomic_state *state,
> > >  
> > >  	commit_pipe_config(state, crtc);
> > >  
> > > -	if (INTEL_GEN(dev_priv) >= 9)
> > > +	if (new_crtc_state->bigjoiner) {
> > > +	/* Not supported yet */
> > > +	} else if (INTEL_GEN(dev_priv) >= 9)
> > >  		skl_update_planes_on_crtc(state, crtc);
> > >  	else
> > >  		i9xx_update_planes_on_crtc(state, crtc);
> > > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
> > > index b526afee595c..59ed94d68b55 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_display_types.h
> > > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h
> > > @@ -1070,6 +1070,15 @@ struct intel_crtc_state {
> > >  	/* enable pipe csc? */
> > >  	bool csc_enable;
> > >  
> > > +	/* enable pipe big joiner? */
> > > +	bool bigjoiner;
> > > +
> > > +	/* big joiner slave crtc? */
> > > +	bool bigjoiner_slave;
> > > +
> > > +	/* linked crtc for bigjoiner, either slave or master */
> > > +	struct intel_crtc *bigjoiner_linked_crtc;
> > > +
> > >  	/* Display Stream compression state */
> > >  	struct {
> > >  		bool compression_enable;
> > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> > > index d123e5a508c1..cb604899a881 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > > @@ -2349,6 +2349,15 @@ static int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
> > >  	pipe_config->port_clock = intel_dp->common_rates[limits->max_clock];
> > >  	pipe_config->lane_count = limits->max_lane_count;
> > >  
> > > +	if (adjusted_mode->crtc_clock > intel_dp_max_dotclock(intel_dp, false)) {
> > > +		if (adjusted_mode->crtc_clock > intel_dp_max_dotclock(intel_dp, true)) {
> > > +			DRM_DEBUG_KMS("Clock rate too high for big joiner\n");
> > > +			return -EINVAL;
> > > +		}
> > 
> > That second check is redundant I think. intel_crtc_compute_config()
> > already takes care of it.
> 
> Why is it redundant? I mean with bigjoiner = true we still need to make sure
> that crtc_clock is not > max dotclock with bigjoiner right?

intel_crtc_compute_config() already takes care of it, like I said.

> 
> > 
> > So seems like we should be good with just something super simple like:
> > 
> > if (crtc_clock > dev_priv->max_dotclock && can_bigjoiner())
> > 	bigjoiner = true;
> 
> Manasi
> 
> > 
> > 
> > > +		pipe_config->bigjoiner = true;
> > > +		DRM_DEBUG_KMS("Using bigjoiner configuration\n");
> > > +	}
> > > +
> > >  	if (intel_dp_is_edp(intel_dp)) {
> > >  		pipe_config->dsc.compressed_bpp =
> > >  			min_t(u16, drm_edp_dsc_sink_output_bpp(intel_dp->dsc_dpcd) >> 4,
> > > @@ -2366,12 +2375,12 @@ static int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
> > >  						    pipe_config->lane_count,
> > >  						    adjusted_mode->crtc_clock,
> > >  						    adjusted_mode->crtc_hdisplay,
> > > -						    false);
> > > +						    pipe_config->bigjoiner);
> > >  		dsc_dp_slice_count =
> > >  			intel_dp_dsc_get_slice_count(intel_dp,
> > >  						     adjusted_mode->crtc_clock,
> > >  						     adjusted_mode->crtc_hdisplay,
> > > -						     false);
> > > +						     pipe_config->bigjoiner);
> > >  		if (!dsc_max_output_bpp || !dsc_dp_slice_count) {
> > >  			drm_dbg_kms(&dev_priv->drm,
> > >  				    "Compressed BPP/Slice Count not supported\n");
> > > @@ -2387,14 +2396,15 @@ static int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
> > >  	 * is greater than the maximum Cdclock and if slice count is even
> > >  	 * then we need to use 2 VDSC instances.
> > >  	 */
> > > -	if (adjusted_mode->crtc_clock > dev_priv->max_cdclk_freq) {
> > > -		if (pipe_config->dsc.slice_count > 1) {
> > > -			pipe_config->dsc.dsc_split = true;
> > > -		} else {
> > > +	if (adjusted_mode->crtc_clock > dev_priv->max_cdclk_freq ||
> > > +	    pipe_config->bigjoiner) {
> > > +		if (pipe_config->dsc.slice_count < 2) {
> > >  			drm_dbg_kms(&dev_priv->drm,
> > >  				    "Cannot split stream to use 2 VDSC instances\n");
> > >  			return -EINVAL;
> > >  		}
> > > +
> > > +		pipe_config->dsc.dsc_split = true;
> > >  	}
> > >  
> > >  	ret = intel_dp_dsc_compute_params(&dig_port->base, pipe_config);
> > > -- 
> > > 2.19.1
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx at lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > 
> > -- 
> > Ville Syrjälä
> > Intel

-- 
Ville Syrjälä
Intel


More information about the Intel-gfx mailing list