On Tue, Jun 22, 2021 at 04:28:14PM -0700, Matt Roper wrote:
> On Tue, Jun 22, 2021 at 02:22:09PM -0700, Lucas De Marchi wrote:
> > We should no longer increment level 0 by 1usec when we have 16Gb DIMMs.
> > Instead spec says to add 3usec (as opposed to 2) to each valid level
> > when punit replies 0 to level 0.
> >
> > So set wm_lv_0_adjust_needed to false for DISPLAY_VER() >= 12 and set
> > the proper adjustement value when handling WaWmMemoryReadLatency.
> >
> > Signed-off-by: Lucas De Marchi <lucas.demarchi@xxxxxxxxx>
>
> Reviewed-by: Matt Roper <matthew.d.roper@xxxxxxxxx>
Might also be worth adding
Bspec: 49326, 4381
to the commit message while applying.
>
> > ---
> > drivers/gpu/drm/i915/intel_dram.c | 3 +--
> > drivers/gpu/drm/i915/intel_pm.c | 13 +++++++------
> > 2 files changed, 8 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_dram.c b/drivers/gpu/drm/i915/intel_dram.c
> > index 50fdea84ba70..879b0f007be3 100644
> > --- a/drivers/gpu/drm/i915/intel_dram.c
> > +++ b/drivers/gpu/drm/i915/intel_dram.c
> > @@ -484,8 +484,7 @@ static int gen11_get_dram_info(struct drm_i915_private *i915)
> >
> > static int gen12_get_dram_info(struct drm_i915_private *i915)
> > {
> > - /* Always needed for GEN12+ */
> > - i915->dram_info.wm_lv_0_adjust_needed = true;
> > + i915->dram_info.wm_lv_0_adjust_needed = false;
> >
> > return icl_pcode_read_mem_global_info(i915);
> > }
> > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> > index ef8d9b2284b3..be2931d54b95 100644
> > --- a/drivers/gpu/drm/i915/intel_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > @@ -2911,18 +2911,20 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv,
> > }
> >
> > /*
> > - * WaWmMemoryReadLatency:skl+,glk
> > + * WaWmMemoryReadLatency
> > *
> > * punit doesn't take into account the read latency so we need
> > - * to add 2us to the various latency levels we retrieve from the
> > - * punit when level 0 response data us 0us.
> > + * to add proper adjustement to each valid level we retrieve
> > + * from the punit when level 0 response data is 0us.
> > */
> > if (wm[0] == 0) {
> > - wm[0] += 2;
> > + u8 adjust = DISPLAY_VER(dev_priv) >= 12 ? 3 : 2;
> > +
> > + wm[0] += adjust;
> > for (level = 1; level <= max_level; level++) {
> > if (wm[level] == 0)
> > break;
> > - wm[level] += 2;
> > + wm[level] += adjust;
> > }
> > }
> >
> > @@ -2934,7 +2936,6 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv,
> > */
> > if (dev_priv->dram_info.wm_lv_0_adjust_needed)
> > wm[0] += 1;
> > -
> > } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
> > u64 sskpd = intel_uncore_read64(uncore, MCH_SSKPD);
> >
> > --
> > 2.31.1
> >
>
> --
> Matt Roper
> Graphics Software Engineer
> VTT-OSGC Platform Enablement
> Intel Corporation
> (916) 356-2795
--
Matt Roper
Graphics Software Engineer
VTT-OSGC Platform Enablement
Intel Corporation
(916) 356-2795
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
