[Intel-gfx] [PATCH 50/67] drm/i915/gen10+: use the SKL code for reading WM latencies

[Ville Syrjälä]

On Mon, Apr 24, 2017 at 04:10:41PM -0300, Paulo Zanoni wrote:
> Em Seg, 2017-04-24 às 21:22 +0300, Ville Syrjälä escreveu:
> > On Thu, Apr 06, 2017 at 12:15:46PM -0700, Rodrigo Vivi wrote:
> > > 
> > > From: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > > 
> > > Gen 10 should use the exact same code as Gen 9, so change the check
> > > to
> > > take this into consideration, and also assume that future platforms
> > > will run this code.
> > > 
> > > Also add a MISSING_CASE(), just in case we do something wrong,
> > > instead
> > > of silently failing.
> > > 
> > > Signed-off-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > > Signed-off-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/intel_pm.c | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/intel_pm.c
> > > b/drivers/gpu/drm/i915/intel_pm.c
> > > index 4c07b91..a2b2509 100644
> > > --- a/drivers/gpu/drm/i915/intel_pm.c
> > > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > > @@ -2291,7 +2291,7 @@ static void ilk_compute_wm_level(const struct
> > > drm_i915_private *dev_priv,
> > >  static void intel_read_wm_latency(struct drm_i915_private
> > > *dev_priv,
> > >  				  uint16_t wm[8])
> > >  {
> > > -	if (IS_GEN9(dev_priv)) {
> > > +	if (INTEL_GEN(dev_priv) >= 9) {
> > >  		uint32_t val;
> > >  		int ret, i;
> > >  		int level, max_level = ilk_wm_max_level(dev_priv);
> > > @@ -2351,7 +2351,7 @@ static void intel_read_wm_latency(struct
> > > drm_i915_private *dev_priv,
> > >  		}
> > >  
> > >  		/*
> > > -		 * WaWmMemoryReadLatency:skl,glk
> > > +		 * WaWmMemoryReadLatency:skl+,glk
> > 
> > When we did we start to use the '+' notation in the w/a notes?
> 
> The + is because this code, when written, was going to be run for every
> gen >= 9. Stuff changed, rebases happened, so I guess I'll need to re-
> check to see that's still the case.
> 
> > 
> > What would it mean to say 'skl+,glk'? Is BXT included or not?
> 
> When I originally wrote this patch, I changed from "skl" to "skl+",
> there was no glk, and it was added by not-me during some rebase at some
> point.

Well, up to now we've never used + in these. So maybe just list ever
platform explicitly. Makes it clear where it really applies. The + stuff
gets hard when the w/a is no longer needed by some more recent platform.

> 
> > 
> > > 
> > >  		 *
> > >  		 * punit doesn't take into account the read
> > > latency so we need
> > >  		 * to add 2us to the various latency levels we
> > > retrieve from the
> > > @@ -2390,6 +2390,8 @@ static void intel_read_wm_latency(struct
> > > drm_i915_private *dev_priv,
> > >  		wm[0] = 7;
> > >  		wm[1] = (mltr >> MLTR_WM1_SHIFT) & ILK_SRLT_MASK;
> > >  		wm[2] = (mltr >> MLTR_WM2_SHIFT) & ILK_SRLT_MASK;
> > > +	} else {
> > > +		MISSING_CASE(INTEL_DEVID(dev_priv));
> > >  	}
> > >  }
> > >  
> > > -- 
> > > 1.9.1
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx at lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > 

-- 
Ville Syrjälä
Intel OTC