[Intel-gfx] [PATCH 2/3] drm/i915/display: Convert link bitrate to corresponding PLL clock

Sripada, Radhakrishna radhakrishna.sripada at intel.com
Tue Dec 5 18:09:23 UTC 2023


Hi Mika,

> -----Original Message-----
> From: Kahola, Mika <mika.kahola at intel.com>
> Sent: Tuesday, December 5, 2023 12:28 AM
> To: Sripada, Radhakrishna <radhakrishna.sripada at intel.com>; intel-
> gfx at lists.freedesktop.org
> Subject: RE: [Intel-gfx] [PATCH 2/3] drm/i915/display: Convert link bitrate to
> corresponding PLL clock
> 
> > -----Original Message-----
> > From: Sripada, Radhakrishna <radhakrishna.sripada at intel.com>
> > Sent: Tuesday, December 5, 2023 3:36 AM
> > To: Kahola, Mika <mika.kahola at intel.com>; intel-gfx at lists.freedesktop.org
> > Subject: RE: [Intel-gfx] [PATCH 2/3] drm/i915/display: Convert link bitrate to
> corresponding PLL clock
> >
> > Hi Mika,
> >
> > > -----Original Message-----
> > > From: Intel-gfx <intel-gfx-bounces at lists.freedesktop.org> On Behalf Of
> > > Mika Kahola
> > > Sent: Monday, December 4, 2023 3:59 AM
> > > To: intel-gfx at lists.freedesktop.org
> > > Subject: [Intel-gfx] [PATCH 2/3] drm/i915/display: Convert link
> > > bitrate to corresponding PLL clock
> > >
> > > Compute clock during PLL readout. This prevents warn when only c20 phy
> > > is connected during modprobe. The intel_c20pll_calc_port_clock()
> > > function returns link bitrate which in DP2.0 and HDMI cases does not
> > > match with the clock rate. Hence, conversion function is needed to
> > > convert link bitrate to corresponding PLL clock rate.
> > >
> > > while at it, update clock on C10 pll state as well.
> > >
> > > Signed-off-by: Clint Taylor <Clinton.A.Taylor at intel.com>
> > > Signed-off-by: Mika Kahola <mika.kahola at intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_cx0_phy.c | 38
> > > ++++++++++++++++++--  drivers/gpu/drm/i915/display/intel_cx0_phy.h |  1 +
> > >  drivers/gpu/drm/i915/display/intel_ddi.c     |  2 +-
> > >  3 files changed, 37 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> > > b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> > > index 2e6412fc2258..02efe2786c6a 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> > > @@ -1871,6 +1871,7 @@ static int intel_c10pll_calc_state(struct
> > > intel_crtc_state *crtc_state,  }
> > >
> > >  static void intel_c10pll_readout_hw_state(struct intel_encoder
> > > *encoder,
> > > +					  struct intel_crtc_state *crtc_state,
> > >  					  struct intel_c10pll_state *pll_state)  {
> > >  	struct drm_i915_private *i915 = to_i915(encoder->base.dev); @@
> > > -1894,6 +1895,7 @@ static void intel_c10pll_readout_hw_state(struct
> > > intel_encoder *encoder,
> > >
> > >  	pll_state->cmn = intel_cx0_read(i915, encoder->port, lane,
> > > PHY_C10_VDR_CMN(0));
> > >  	pll_state->tx = intel_cx0_read(i915, encoder->port, lane,
> > > PHY_C10_VDR_TX(0));
> > > +	pll_state->clock = crtc_state->port_clock;
> > >
> > >  	intel_cx0_phy_transaction_end(encoder, wakeref);  } @@ -2445,12
> > > +2447,33 @@ static void intel_program_port_clock_ctl(struct
> > > intel_encoder *encoder,
> > >  		     XELPDP_SSC_ENABLE_PLLB, val);
> > >  }
> > >
> > > +static int intel_link_bitrate_to_clock(struct intel_encoder *encoder,
> > > +				       struct intel_crtc_state *crtc_state,
> > > +				       int link_bit_rate)
> > > +{
> > > +	const struct intel_c20pll_state * const *tables;
> > > +	int i;
> > > +
> > > +	tables = intel_c20_pll_tables_get(crtc_state, encoder);
> > This will produce incorrect result. intel_c20_pll_tables_get depends on
> intel_crtc_has_{dp_encoder,hdmi..} which depends
> > crtc_state->output_types to determine edp/dp/hdmi table which is not
> initialized until later in mtl_ddi_init_config under
> > intel_ddi_get_config -> intel_ddi_read_func_ctl
> >
> > We might be needing a separate sanitization of initial pll state to be done after
> intel_ddi_get_config. Or a special case to handle
> > initial modeset.
> I actually noticed this while testing it that at first we don't even get the tables and
> function returns with -EINVAL. Eventually we do get the correct table and clock.
> Maybe it would be better to simply use the if-else ladder for those cases that
> differs from link bitrate vs. clock?
I am skeptical about making 2 different entries for the same data. Why don’t we make
intel_c20_get_dp_rate work on link bit rate numbers which is what intel_c20pll_calc_port_clock
returns and that is what is stored in crtc_state/pipe_config->port_clock and use this field to program
instead of relying on pll_state->clock in step5 of pll programming sequence?

--Radhakrishna
> 
> 
> >
> > --Radhakrishna Sripada
> > > +	if (!tables)
> > > +		return -EINVAL;
> > > +
> > > +	for (i = 0; tables[i]; i++) {
> > > +		if (link_bit_rate == tables[i]->link_bit_rate)
> > > +			return tables[i]->clock;
> > > +	}
> > > +
> > > +	return -EINVAL;
> > > +}
> > > +
> > >  static void intel_c20pll_readout_hw_state(struct intel_encoder
> > > *encoder,
> > > +					  struct intel_crtc_state *crtc_state,
> > >  					  struct intel_c20pll_state *pll_state)  {
> > >  	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
> > >  	bool cntx;
> > >  	intel_wakeref_t wakeref;
> > > +	int clock;
> > >  	int i;
> > >
> > >  	wakeref = intel_cx0_phy_transaction_begin(encoder);
> > > @@ -2500,6 +2523,13 @@ static void
> > > intel_c20pll_readout_hw_state(struct
> > > intel_encoder *encoder,
> > >  		}
> > >  	}
> > >
> > > +	pll_state->link_bit_rate = intel_c20pll_calc_port_clock(encoder,
> > > pll_state);
> > > +	clock = intel_link_bitrate_to_clock(encoder, crtc_state,
> > > +					    pll_state->link_bit_rate);
> > > +
> > > +	if (clock >= 0)
> > > +		pll_state->clock = clock;
> > > +
> > >  	intel_cx0_phy_transaction_end(encoder, wakeref);  }
> > >
> > > @@ -3053,15 +3083,16 @@ static void intel_c10pll_state_verify(const
> > > struct intel_crtc_state *state,  }
> > >
> > >  void intel_cx0pll_readout_hw_state(struct intel_encoder *encoder,
> > > +				   struct intel_crtc_state *crtc_state,
> > >  				   struct intel_cx0pll_state *pll_state)  {
> > >  	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
> > >  	enum phy phy = intel_port_to_phy(i915, encoder->port);
> > >
> > >  	if (intel_is_c10phy(i915, phy))
> > > -		intel_c10pll_readout_hw_state(encoder, &pll_state->c10);
> > > +		intel_c10pll_readout_hw_state(encoder, crtc_state, &pll_state-
> > > >c10);
> > >  	else
> > > -		intel_c20pll_readout_hw_state(encoder, &pll_state->c20);
> > > +		intel_c20pll_readout_hw_state(encoder, crtc_state, &pll_state-
> > > >c20);
> > >  }
> > >
> > >  int intel_cx0pll_calc_port_clock(struct intel_encoder *encoder, @@
> > > -3145,7 +3176,8 @@ void intel_cx0pll_state_verify(struct
> > > intel_atomic_state *state,
> > >  	if (intel_tc_port_in_tbt_alt_mode(enc_to_dig_port(encoder)))
> > >  		return;
> > >
> > > -	intel_cx0pll_readout_hw_state(encoder, &mpll_hw_state);
> > > +	intel_cx0pll_readout_hw_state(encoder, (struct
> > > intel_crtc_state*)new_crtc_state,
> > > +				      &mpll_hw_state);
> > >
> > >  	if (intel_is_c10phy(i915, phy))
> > >  		intel_c10pll_state_verify(new_crtc_state, crtc, encoder,
> > > &mpll_hw_state.c10); diff --git
> > > a/drivers/gpu/drm/i915/display/intel_cx0_phy.h
> > > b/drivers/gpu/drm/i915/display/intel_cx0_phy.h
> > > index c6682677253a..eac7354e9a4e 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_cx0_phy.h
> > > +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.h
> > > @@ -32,6 +32,7 @@ intel_mtl_port_pll_type(struct intel_encoder
> > > *encoder,
> > >
> > >  int intel_cx0pll_calc_state(struct intel_crtc_state *crtc_state,
> > > struct intel_encoder *encoder);  void
> > > intel_cx0pll_readout_hw_state(struct intel_encoder *encoder,
> > > +				   struct intel_crtc_state *crtc_state,
> > >  				   struct intel_cx0pll_state *pll_state);  int
> > > intel_cx0pll_calc_port_clock(struct intel_encoder *encoder,
> > >  				 const struct intel_cx0pll_state *pll_state); diff -
> -git
> > > a/drivers/gpu/drm/i915/display/intel_ddi.c
> > > b/drivers/gpu/drm/i915/display/intel_ddi.c
> > > index 38f28c480b38..508d3363e89f 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> > > @@ -3953,7 +3953,7 @@ static void mtl_ddi_get_config(struct
> > > intel_encoder *encoder,
> > >  	if (intel_tc_port_in_tbt_alt_mode(dig_port)) {
> > >  		crtc_state->port_clock =
> > > intel_mtl_tbt_calc_port_clock(encoder);
> > >  	} else {
> > > -		intel_cx0pll_readout_hw_state(encoder, &crtc_state-
> > > >cx0pll_state);
> > > +		intel_cx0pll_readout_hw_state(encoder, crtc_state, &crtc_state-
> > > >cx0pll_state);
> > >  		crtc_state->port_clock = intel_cx0pll_calc_port_clock(encoder,
> > > &crtc_state->cx0pll_state);
> > >  	}
> > >
> > > --
> > > 2.34.1



More information about the Intel-gfx mailing list