[Intel-gfx] [PATCH] drm/i915/cnl: Don't try to manage Port F power wells on all CNL.

[Rodrigo Vivi]

On Tue, Jan 23, 2018 at 03:03:28AM +0000, Pandiyan, Dhinakaran wrote:
> 
> 
> 
> On Mon, 2018-01-22 at 15:48 -0800, Rodrigo Vivi wrote:
> > SKUs that lacks on the full port F split will just time out
> > when touching this power well bits, causing a noisy warn.
> > 
> > v2: Suggested-by: Imre. Temporarily remove the aux pw id after setting
> >     it instead of duplicating and redefining everything.
> > 
> > Cc: Lucas De Marchi <lucas.demarchi at intel.com>
> > Cc: Imre Deak <imre.deak at intel.com>
> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_runtime_pm.c | 25 +++++++++++++++++++------
> >  1 file changed, 19 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > index 433048ffa5c6..7cee63860a7b 100644
> > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > @@ -1861,18 +1861,20 @@ void intel_display_power_put(struct drm_i915_private *dev_priv,
> >  #define CNL_DISPLAY_AUX_D_POWER_DOMAINS (		\
> >  	BIT_ULL(POWER_DOMAIN_AUX_D) |			\
> >  	BIT_ULL(POWER_DOMAIN_INIT))
> > -#define CNL_DISPLAY_AUX_F_POWER_DOMAINS (		\
> > -	BIT_ULL(POWER_DOMAIN_AUX_F) |			\
> > -	BIT_ULL(POWER_DOMAIN_INIT))
> > -#define CNL_DISPLAY_DDI_F_IO_POWER_DOMAINS (		\
> > -	BIT_ULL(POWER_DOMAIN_PORT_DDI_F_IO) |		\
> > -	BIT_ULL(POWER_DOMAIN_INIT))
> >  #define CNL_DISPLAY_DC_OFF_POWER_DOMAINS (		\
> >  	CNL_DISPLAY_POWERWELL_2_POWER_DOMAINS |		\
> >  	BIT_ULL(POWER_DOMAIN_GT_IRQ) |			\
> >  	BIT_ULL(POWER_DOMAIN_MODESET) |			\
> >  	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
> >  	BIT_ULL(POWER_DOMAIN_INIT))
> > +/* Power wells for CNL with port F after this */
> > +#define CNL_FIRST_PORT_F_PW CNL_DISP_PW_AUX_F
> > +#define CNL_DISPLAY_AUX_F_POWER_DOMAINS (		\
> > +	BIT_ULL(POWER_DOMAIN_AUX_F) |			\
> > +	BIT_ULL(POWER_DOMAIN_INIT))
> > +#define CNL_DISPLAY_DDI_F_IO_POWER_DOMAINS (		\
> > +	BIT_ULL(POWER_DOMAIN_PORT_DDI_F_IO) |		\
> > +	BIT_ULL(POWER_DOMAIN_INIT))
> >  
> >  static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
> >  	.sync_hw = i9xx_power_well_sync_hw_noop,
> > @@ -2544,6 +2546,17 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
> >  		set_power_wells(power_domains, skl_power_wells);
> >  	} else if (IS_CANNONLAKE(dev_priv)) {
> >  		set_power_wells(power_domains, cnl_power_wells);
> > +
> > +		if (!IS_CNL_WITH_PORT_F(dev_priv)) {
> > +			int i;
> > +
> > +			for (i = 0; i < power_domains->power_well_count; i++)
> > +				if (power_domains->power_wells[i].id ==
> > +				    CNL_FIRST_PORT_F_PW)
> > +					break;
> > +			WARN_ON(power_domains->power_well_count == i - 1);
> > +			power_domains->power_well_count = i - 1;
> Shouldn't this be
> 	WARN_ON(power_domains->power_well_count == i);

oh yeap... c&p from below without thinking. :/

> 	power_domains->power_well_count = i; ?

this is also what Imre suggested, but I don't think so.

The first-of-non-port-f - 1 is the last of non-port-f.

> 
> 			 
> > +		}
> >  	} else if (IS_BROXTON(dev_priv)) {
> >  		set_power_wells(power_domains, bxt_power_wells);
> >  	} else if (IS_GEMINILAKE(dev_priv)) {