Re: [PATCH v4 10/14] drm/vblank: Add helper to get next vblank time

[Ville Syrjälä]

On Wed, Feb 22, 2023 at 07:44:42AM -0800, Rob Clark wrote:
> On Wed, Feb 22, 2023 at 1:57 AM Pekka Paalanen <ppaalanen@xxxxxxxxx> wrote:
> >
> > On Tue, 21 Feb 2023 09:50:20 -0800
> > Rob Clark <robdclark@xxxxxxxxx> wrote:
> >
> > > On Tue, Feb 21, 2023 at 5:01 AM Ville Syrjälä
> > > <ville.syrjala@xxxxxxxxxxxxxxx> wrote:
> > > >
> > > > On Tue, Feb 21, 2023 at 10:45:51AM +0200, Pekka Paalanen wrote:
> > > > > On Mon, 20 Feb 2023 07:55:41 -0800
> > > > > Rob Clark <robdclark@xxxxxxxxx> wrote:
> > > > >
> > > > > > On Mon, Feb 20, 2023 at 1:08 AM Pekka Paalanen <ppaalanen@xxxxxxxxx> wrote:
> > > > > > >
> > > > > > > On Sat, 18 Feb 2023 13:15:53 -0800
> > > > > > > Rob Clark <robdclark@xxxxxxxxx> wrote:
> > > > > > >
> > > > > > > > From: Rob Clark <robdclark@xxxxxxxxxxxx>
> > > > > > > >
> > > > > > > > Will be used in the next commit to set a deadline on fences that an
> > > > > > > > atomic update is waiting on.
> > > > > > > >
> > > > > > > > Signed-off-by: Rob Clark <robdclark@xxxxxxxxxxxx>
> > > > > > > > ---
> > > > > > > >  drivers/gpu/drm/drm_vblank.c | 32 ++++++++++++++++++++++++++++++++
> > > > > > > >  include/drm/drm_vblank.h     |  1 +
> > > > > > > >  2 files changed, 33 insertions(+)
> > > > > > > >
> > > > > > > > diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
> > > > > > > > index 2ff31717a3de..caf25ebb34c5 100644
> > > > > > > > --- a/drivers/gpu/drm/drm_vblank.c
> > > > > > > > +++ b/drivers/gpu/drm/drm_vblank.c
> > > > > > > > @@ -980,6 +980,38 @@ u64 drm_crtc_vblank_count_and_time(struct drm_crtc *crtc,
> > > > > > > >  }
> > > > > > > >  EXPORT_SYMBOL(drm_crtc_vblank_count_and_time);
> > > > > > > >
> > > > > > > > +/**
> > > > > > > > + * drm_crtc_next_vblank_time - calculate the time of the next vblank
> > > > > > > > + * @crtc: the crtc for which to calculate next vblank time
> > > > > > > > + * @vblanktime: pointer to time to receive the next vblank timestamp.
> > > > > > > > + *
> > > > > > > > + * Calculate the expected time of the next vblank based on time of previous
> > > > > > > > + * vblank and frame duration
> > > > > > >
> > > > > > > Hi,
> > > > > > >
> > > > > > > for VRR this targets the highest frame rate possible for the current
> > > > > > > VRR mode, right?
> > > > > > >
> > > > > >
> > > > > > It is based on vblank->framedur_ns which is in turn based on
> > > > > > mode->crtc_clock.  Presumably for VRR that ends up being a maximum?
> > > > >
> > > > > I don't know. :-)
> > > >
> > > > At least for i915 this will give you the maximum frame
> > > > duration.
> > >
> > > I suppose one could argue that maximum frame duration is the actual
> > > deadline.  Anything less is just moar fps, but not going to involve
> > > stalling until vblank N+1, AFAIU
> > >
> > > > Also this does not calculate the the start of vblank, it
> > > > calculates the start of active video.
> > >
> > > Probably something like end of previous frame's video..  might not be
> > > _exactly_ correct (because some buffering involved), but OTOH on the
> > > GPU side, I expect the driver to set a timer for a few ms or so before
> > > the deadline.  So there is some wiggle room.
> >
> > The vblank timestamp is defined to be the time of the first active
> > pixel of the frame in the video signal. At least that's the one that
> > UAPI carries (when not tearing?). It is not the start of vblank period.
> >
> > With VRR, the front porch before the first active pixel can be multiple
> > milliseconds. The difference between 144 Hz and 60 Hz is 9.7 ms for
> > example.
> 
> What we really want is the deadline for the hw to latch for the next
> frame.. which as Ville pointed out is definitely before the end of
> vblank.
> 
> Honestly this sort of feature is a lot more critical for the non-VRR
> case, and VRR is kind of a minority edge case.  So I'd prefer not to
> get too hung up on VRR.  If there is an easy way for the helpers to
> detect VRR, I'd be perfectly fine not setting a deadline hint in that
> case, and let someone who actually has a VRR display figure out how to
> handle that case.

The formula I gave you earlier works for both VRR and non-VRR.

-- 
Ville Syrjälä
Intel