[PATCH v7 06/11] drm/i915/psr: Modify intel_dp_get_su_granularity to support panel replay
Manna, Animesh
animesh.manna at intel.com
Mon Apr 29 11:34:26 UTC 2024
> -----Original Message-----
> From: Hogander, Jouni <jouni.hogander at intel.com>
> Sent: Monday, April 29, 2024 4:46 PM
> To: Manna, Animesh <animesh.manna at intel.com>; intel-
> gfx at lists.freedesktop.org
> Subject: Re: [PATCH v7 06/11] drm/i915/psr: Modify
> intel_dp_get_su_granularity to support panel replay
>
> On Mon, 2024-04-29 at 11:02 +0000, Manna, Animesh wrote:
> >
> >
> > > -----Original Message-----
> > > From: Hogander, Jouni <jouni.hogander at intel.com>
> > > Sent: Friday, April 19, 2024 5:42 PM
> > > To: intel-gfx at lists.freedesktop.org
> > > Cc: Manna, Animesh <animesh.manna at intel.com>; Hogander, Jouni
> > > <jouni.hogander at intel.com>
> > > Subject: [PATCH v7 06/11] drm/i915/psr: Modify
> > > intel_dp_get_su_granularity
> > > to support panel replay
> > >
> > > Currently intel_dp_get_su_granularity doesn't support panel replay.
> > > This fix modifies it to support panel replay as well.
> > >
> > > v2: rely on PSR definitions on common bits
> > >
> > > Signed-off-by: Jouni Högander <jouni.hogander at intel.com>
> > > Reviewed-by: Animesh Manna <animesh.manna at intel.com>
> > > ---
> > > drivers/gpu/drm/i915/display/intel_psr.c | 62
> > > +++++++++++++++++++++---
> > > 1 file changed, 55 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/i915/display/intel_psr.c
> > > b/drivers/gpu/drm/i915/display/intel_psr.c
> > > index b94f8e33ed1f..29400fac13c2 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_psr.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> > > @@ -466,6 +466,40 @@ static u8
> > > intel_dp_get_sink_sync_latency(struct
> > > intel_dp *intel_dp)
> > > return val;
> > > }
> > >
> > > +static u8 intel_dp_get_su_capability(struct intel_dp *intel_dp) {
> > > + u8 su_capability;
> > > +
> > > + if (intel_dp->psr.sink_panel_replay_su_support)
> >
> > While relooking found that good to add a check for
> > DP_PANEL_PANEL_REPLAY_SU_GRANULARITY_REQUIRED (6th bit of dpcd
> 0xb1).
> > What if it is zero means granularity not needed but will continue to
> > use x-granularity and y-granularity.
>
> Please note it is 5th bit in 0xb1. See my further comment below...
Same I was referring, 6th bit index 5 as it starts from 0.
>
> > Regards,
> > Animesh
> >
> > > + drm_dp_dpcd_read(&intel_dp->aux,
> > > +
> > > DP_PANEL_PANEL_REPLAY_X_GRANULARITY,
> > > + &su_capability, 1);
> > > + else
> > > + su_capability = intel_dp->psr_dpcd[1];
> > > +
> > > + return su_capability;
> > > +}
> > > +
> > > +static unsigned int
> > > +intel_dp_get_su_x_granularity_offset(struct intel_dp *intel_dp) {
> > > + return intel_dp->psr.sink_panel_replay_su_support ?
> > > + DP_PANEL_PANEL_REPLAY_X_GRANULARITY :
> > > + DP_PSR2_SU_X_GRANULARITY;
> > > +}
> > > +
> > > +static unsigned int
> > > +intel_dp_get_su_y_granularity_offset(struct intel_dp *intel_dp) {
> > > + return intel_dp->psr.sink_panel_replay_su_support ?
> > > + DP_PANEL_PANEL_REPLAY_Y_GRANULARITY :
> > > + DP_PSR2_SU_Y_GRANULARITY;
> > > +}
> > > +
> > > +/*
> > > + * Note: Bits related to granularity are same in panel replay and
> > > psr
> > > + * registers. Rely on PSR definitions on these "common" bits.
> > > + */
>
> Check this comment. See my further comment below...
Agree bit position is same.
>
> > > static void intel_dp_get_su_granularity(struct intel_dp
> > > *intel_dp) {
> > > struct drm_i915_private *i915 = dp_to_i915(intel_dp); @@ -
> > > 473,18
> > > +507,29 @@ static void intel_dp_get_su_granularity(struct intel_dp
> > > *intel_dp)
> > > u16 w;
> > > u8 y;
> > >
> > > - /* If sink don't have specific granularity requirements set
> > > legacy ones
> > > */
> > > - if (!(intel_dp->psr_dpcd[1] &
> > > DP_PSR2_SU_GRANULARITY_REQUIRED)) {
> > > + /*
> > > + * TODO: Do we need to take into account panel supporting
> > > both PSR
> > > and
> > > + * Panel replay?
> > > + */
> > > +
> > > + /*
> > > + * If sink don't have specific granularity requirements set
> > > legacy
> > > + * ones.
> > > + */
> > > + if (!(intel_dp_get_su_capability(intel_dp) &
> > > + DP_PSR2_SU_GRANULARITY_REQUIRED)) {
> > > /* As PSR2 HW sends full lines, we do not care
> > > about x
> > > granularity */
> > > w = 4;
> > > y = 4;
> > > goto exit;
> > > }
>
> This block is taking care of checking bit 5 in 0xb1 and using legacy
> ones if no requirements.
intel_dp_get_su_capability is reading from DP_PANEL_PANEL_REPLAY_X_GRANULARITY which Is dpcd 0xb2. Am I missing something?
Regards,
Animesh
>
> BR,
>
> Jouni Högander
>
> > >
> > > - r = drm_dp_dpcd_read(&intel_dp->aux,
> > > DP_PSR2_SU_X_GRANULARITY, &w, 2);
> > > + r = drm_dp_dpcd_read(&intel_dp->aux,
> > > +
> > > intel_dp_get_su_x_granularity_offset(intel_dp),
> > > + &w, 2);
> > > if (r != 2)
> > > drm_dbg_kms(&i915->drm,
> > > - "Unable to read
> > > DP_PSR2_SU_X_GRANULARITY\n");
> > > + "Unable to read selective update x
> > > granularity\n");
> > > /*
> > > * Spec says that if the value read is 0 the default
> > > granularity should
> > > * be used instead.
> > > @@ -492,10 +537,12 @@ static void
> > > intel_dp_get_su_granularity(struct
> > > intel_dp *intel_dp)
> > > if (r != 2 || w == 0)
> > > w = 4;
> > >
> > > - r = drm_dp_dpcd_read(&intel_dp->aux,
> > > DP_PSR2_SU_Y_GRANULARITY, &y, 1);
> > > + r = drm_dp_dpcd_read(&intel_dp->aux,
> > > +
> > > intel_dp_get_su_y_granularity_offset(intel_dp),
> > > + &y, 1);
> > > if (r != 1) {
> > > drm_dbg_kms(&i915->drm,
> > > - "Unable to read
> > > DP_PSR2_SU_Y_GRANULARITY\n");
> > > + "Unable to read selective update y
> > > granularity\n");
> > > y = 4;
> > > }
> > > if (y == 0)
> > > @@ -588,7 +635,8 @@ void intel_psr_init_dpcd(struct intel_dp
> > > *intel_dp)
> > > if (intel_dp->psr_dpcd[0])
> > > _psr_init_dpcd(intel_dp);
> > >
> > > - if (intel_dp->psr.sink_psr2_support)
> > > + if (intel_dp->psr.sink_psr2_support ||
> > > + intel_dp->psr.sink_panel_replay_su_support)
> > > intel_dp_get_su_granularity(intel_dp);
> > > }
> > >
> > > --
> > > 2.34.1
> >
More information about the Intel-gfx
mailing list