Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Hogander, Jouni" <jouni.hogander@intel.com>
To: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
	"Nautiyal, Ankit K" <ankit.k.nautiyal@intel.com>,
	"intel-gfx@lists.freedesktop.org"
	<intel-gfx@lists.freedesktop.org>
Cc: "ville.syrjala@linux.intel.com" <ville.syrjala@linux.intel.com>
Subject: Re: [PATCH 1/5] drm/i915/psr: Add helper to get min psr guardband
Date: Fri, 17 Oct 2025 09:58:24 +0000	[thread overview]
Message-ID: <043a61da1973f6bfd5cc9b45c84c1c13e038828e.camel@intel.com> (raw)
In-Reply-To: <ad6cd831-517a-4eb9-b812-1fb4dadea00d@intel.com>

On Fri, 2025-10-17 at 15:07 +0530, Nautiyal, Ankit K wrote:
> 
> On 10/17/2025 2:37 PM, Hogander, Jouni wrote:
> > On Fri, 2025-10-17 at 10:31 +0530, Ankit Nautiyal wrote:
> > > Introduce a helper to compute the max link wake latency when
> > > using
> > > Auxless/Aux wake mechanism for PSR/Panel Replay/LOBF features.
> > > 
> > > This will be used to compute the minimum guardband so that the
> > > link
> > > wake
> > > latencies are accounted and these features work smoothly for
> > > higher
> > > refresh rate panels.
> > > 
> > > Bspec: 70151, 71477
> > > Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
> > > ---
> > >   drivers/gpu/drm/i915/display/intel_psr.c | 12 ++++++++++++
> > >   drivers/gpu/drm/i915/display/intel_psr.h |  1 +
> > >   2 files changed, 13 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/display/intel_psr.c
> > > b/drivers/gpu/drm/i915/display/intel_psr.c
> > > index 703e5f6af04c..a8303b669853 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_psr.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> > > @@ -4416,3 +4416,15 @@ void intel_psr_compute_config_late(struct
> > > intel_dp *intel_dp,
> > >   
> > >   	intel_psr_set_non_psr_pipes(intel_dp, crtc_state);
> > >   }
> > > +
> > > +int intel_psr_min_guardband(struct intel_crtc_state *crtc_state)
> > > +{
> > > +	struct intel_display *display =
> > > to_intel_display(crtc_state);
> > > +	int auxless_wake_lines = crtc_state-
> > > > alpm_state.aux_less_wake_lines;
> > > +	int wake_lines = DISPLAY_VER(display) < 20 ?
> > > +			 psr2_block_count_lines(crtc_state-
> > > > alpm_state.io_wake_lines,
> > > +						crtc_state-
> > > > alpm_state.fast_wake_lines) :
> > > +			 crtc_state->alpm_state.io_wake_lines;
> > > +
> > > +	return max(auxless_wake_lines, wake_lines);
> > hmm, now if you add:
> > 
> > if (crtc_state->req_psr2_sdp_prior_scanline)
> > 		psr_min_guardband++;
> 
> I did not get this part. Do we need to account for 1 more wakelines
> if 
> this flag is set?

If you look at how that flag affects vblank check in 
intel_psr_compute_config_late:

...
static bool _wake_lines_fit_into_vblank(const struct intel_crtc_state *crtc_state,
					int vblank,
					int wake_lines)
{
	if (crtc_state->req_psr2_sdp_prior_scanline)
		vblank -= 1;
...

So to take that into account when calculating minimum guardband needed
by PSR you need to increase by one. Same goes with SCL:

...
	int scl = _intel_psr_min_set_context_latency(crtc_state,
						     needs_panel_replay,
						     needs_sel_update);
	vblank -= scl;
...

You are already partially taking into account PSR needs when
calculating optimized guardband except these two lines which are needed
conditionally.

Also intel_psr_compute_config is run at this point -> you know which
one to use: auxless wake time or aux wake time. no need to use max()
with them. Just choose the one which is relevant.

With these changes you could drop intel_psr_compute_config_late as
vblank would be long enough for PSR mode computed by
intel_psr_compute_config, no?

BR,

Jouni Högander


> 
> What we want to do is to check for min guardband for 
> panel_replay/sel_update based on the required wakelines.
> 
> Whether we can use the auxless_wake_lines and wake_lines as computed 
> above to estimate the max wakelines or instead we should use
> functions 
> from alpm.c :
> 
> io_buffer_wake_time() and _lnl_compute_aux_less_wake_time() to get
> the 
> worst case wakelines.
> 
> 
> > 
> > Whatever is the PSR mode it can be enabled what comes to vblank
> > restrictions and you can drop psr_compute_config_late?
> 
> 
> I think we cannot drop the psr_compute_config_late as it checks
> whether 
> the actual guardband is enough for PSR features.
> 
> intel_psr_min_guardband() is used along with other features to have
> an estimate on the guardband that works for all cases, without a need
> to change the guardband.
> It is bounded by the vblank length available
> 
> Regards,
> 
> Ankit
> 
> > 
> > BR,
> > 
> > Jouni Högander
> > 
> > > +}
> > > diff --git a/drivers/gpu/drm/i915/display/intel_psr.h
> > > b/drivers/gpu/drm/i915/display/intel_psr.h
> > > index b17ce312dc37..620b35928832 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_psr.h
> > > +++ b/drivers/gpu/drm/i915/display/intel_psr.h
> > > @@ -85,5 +85,6 @@ bool intel_psr_needs_alpm_aux_less(struct
> > > intel_dp
> > > *intel_dp,
> > >   				   const struct intel_crtc_state
> > > *crtc_state);
> > >   void intel_psr_compute_config_late(struct intel_dp *intel_dp,
> > >   				   struct intel_crtc_state
> > > *crtc_state);
> > > +int intel_psr_min_guardband(struct intel_crtc_state
> > > *crtc_state);
> > >   
> > >   #endif /* __INTEL_PSR_H__ */


  reply	other threads:[~2025-10-17  9:58 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-17  5:01 [PATCH 0/5] Optimize vrr.guardband Ankit Nautiyal
2025-10-17  5:01 ` [PATCH 1/5] drm/i915/psr: Add helper to get min psr guardband Ankit Nautiyal
2025-10-17  9:07   ` Hogander, Jouni
2025-10-17  9:30     ` Hogander, Jouni
2025-10-17  9:41       ` Nautiyal, Ankit K
2025-10-17  9:37     ` Nautiyal, Ankit K
2025-10-17  9:58       ` Hogander, Jouni [this message]
2025-10-17 10:15         ` Hogander, Jouni
2025-10-17 10:30           ` Hogander, Jouni
2025-10-17 11:11           ` Nautiyal, Ankit K
2025-10-17 11:30             ` Hogander, Jouni
2025-10-17  5:01 ` [PATCH 2/5] drm/i915/dp: Add helper to get min sdp guardband Ankit Nautiyal
2025-10-17 10:50   ` Hogander, Jouni
2025-10-17 11:07     ` Nautiyal, Ankit K
2025-10-17  5:02 ` [PATCH 3/5] drm/i915/dp: Check if guardband can accommodate sdp latencies Ankit Nautiyal
2025-10-17 12:02   ` Ville Syrjälä
2025-10-17  5:02 ` [PATCH 4/5] drm/i915/vrr: Use the min static optimized guardband Ankit Nautiyal
2025-10-17 12:06   ` Ville Syrjälä
2025-10-17  5:02 ` [PATCH 5/5] drm/i915/vrr: Use optimized guardband whenever VRR TG is active Ankit Nautiyal
2025-10-17 12:13   ` Ville Syrjälä
2025-10-17  6:10 ` ✓ i915.CI.BAT: success for Optimize vrr.guardband (rev3) Patchwork
  -- strict thread matches above, loose matches on Subject: below --
2025-10-17 12:34 [PATCH 0/5] Optimize vrr.guardband Ankit Nautiyal
2025-10-17 12:35 ` [PATCH 1/5] drm/i915/psr: Add helper to get min psr guardband Ankit Nautiyal
2025-10-17 13:02   ` Hogander, Jouni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=043a61da1973f6bfd5cc9b45c84c1c13e038828e.camel@intel.com \
    --to=jouni.hogander@intel.com \
    --cc=ankit.k.nautiyal@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=ville.syrjala@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox