Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Govindapillai, Vinod" <vinod.govindapillai@intel.com>
To: "ville.syrjala@linux.intel.com" <ville.syrjala@linux.intel.com>
Cc: "Saarinen, Jani" <jani.saarinen@intel.com>,
	"Lisovskiy, Stanislav" <stanislav.lisovskiy@intel.com>,
	"intel-gfx@lists.freedesktop.org"
	<intel-gfx@lists.freedesktop.org>,
	"Syrjala, Ville" <ville.syrjala@intel.com>
Subject: Re: [PATCH v10 6/6] drm/i915/display: force qgv check after the hw state readout
Date: Tue, 23 Apr 2024 07:46:45 +0000	[thread overview]
Message-ID: <e4a47c78906f7dcbb8c9314fdd84671d1821f918.camel@intel.com> (raw)
In-Reply-To: <58490248f4c6d443aea6dbe335c1233542703d58.camel@intel.com>

On Tue, 2024-04-23 at 10:40 +0300, Govindapillai, Vinod wrote:
> On Fri, 2024-04-19 at 20:19 +0300, Ville Syrjälä wrote:
> > On Fri, Apr 05, 2024 at 02:35:33PM +0300, Vinod Govindapillai wrote:
> > > The current intel_bw_atomic_check do not check the possbility
> > > of a sagv configuration change after the hw state readout.
> > > Hence cannot update the sagv configuration until some other
> > > relevant changes like data rates, number of planes etc. happen.
> > > Introduce a flag to force qgv check in such cases.
> > 
> > The correct fix would be to make sure the readout actually
> > works, and thus the software state reflects the state of
> > the hardare accurately.
> 
> Okay! Does that mean, we need to update the qgv points and update to pcode right after the HW
> readout?
> 
> Right now, on bw_init we force disable sagv, so that we have a known sagv state. Then on
>         intel_modeset_readout_hw_state() we update the bw_state parameters. SAGV is still
> disabled. So on
> nextmodesetifthebw_stateparametersaresameastheoldm,thenwewillnotbecalculatingtheQGVpoints.Ithinksu
> chascenariohappensonlyonceaftertheboot.SothatsthereasonIaddedthisflagtoforcecalculateQGV.

Resending the previous reply as somehow the format got messed up!

Okay! Does that mean, we need to update the qgv points and update to pcode right after the HW
readout?

Right now, on bw_init we force disable sagv, so that we have a known sagv state. Then on
intel_modeset_readout_hw_state we update the bw_state parameters. SAGV is still disabled. So on
next modeset if the bw_state parameters are same as the old, then we will not be calculating the QGV
points. I think such a scenario happens only once after the boot. So thats the reason I added this
flag to force calculate QGV.

BR
Vinod

> 
> 
> BR
> Vinod
> 
> > 
> > > 
> > > Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_bw.c | 8 ++++++--
> > >  drivers/gpu/drm/i915/display/intel_bw.h | 6 ++++++
> > >  2 files changed, 12 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/display/intel_bw.c b/drivers/gpu/drm/i915/display/intel_bw.c
> > > index 6fb228a1a28f..1b190be745a0 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_bw.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_bw.c
> > > @@ -755,6 +755,7 @@ void intel_bw_crtc_update(struct intel_bw_state *bw_state,
> > >                 intel_bw_crtc_data_rate(crtc_state);
> > >         bw_state->num_active_planes[crtc->pipe] =
> > >                 intel_bw_crtc_num_active_planes(crtc_state);
> > > +       bw_state->force_check_qgv = true;
> > >  
> > >         drm_dbg_kms(&i915->drm, "pipe %c data rate %u num active planes %u\n",
> > >                     pipe_name(crtc->pipe),
> > > @@ -1339,8 +1340,9 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
> > >         new_bw_state = intel_atomic_get_new_bw_state(state);
> > >  
> > >         if (new_bw_state &&
> > > -           intel_can_enable_sagv(i915, old_bw_state) !=
> > > -           intel_can_enable_sagv(i915, new_bw_state))
> > > +           (intel_can_enable_sagv(i915, old_bw_state) !=
> > > +            intel_can_enable_sagv(i915, new_bw_state) ||
> > > +            new_bw_state->force_check_qgv))
> > >                 changed = true;
> > >  
> > >         /*
> > > @@ -1354,6 +1356,8 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
> > >         if (ret)
> > >                 return ret;
> > >  
> > > +       new_bw_state->force_check_qgv = false;
> > > +
> > >         return 0;
> > >  }
> > >  
> > > diff --git a/drivers/gpu/drm/i915/display/intel_bw.h b/drivers/gpu/drm/i915/display/intel_bw.h
> > > index fa1e924ec961..161813cca473 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_bw.h
> > > +++ b/drivers/gpu/drm/i915/display/intel_bw.h
> > > @@ -47,6 +47,12 @@ struct intel_bw_state {
> > >          */
> > >         u16 qgv_points_mask;
> > >  
> > > +       /*
> > > +        * Flag to force the QGV comparison in atomic check right after the
> > > +        * hw state readout
> > > +        */
> > > +       bool force_check_qgv;
> > > +
> > >         int min_cdclk[I915_MAX_PIPES];
> > >         unsigned int data_rate[I915_MAX_PIPES];
> > >         u8 num_active_planes[I915_MAX_PIPES];
> > > -- 
> > > 2.34.1
> > 
> 


  reply	other threads:[~2024-04-23  7:47 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-04-05 11:35 [PATCH v10 0/6] QGV/SAGV related fixes Vinod Govindapillai
2024-04-05 11:35 ` [PATCH v10 1/6] drm/i915/display: Add meaningful traces for QGV point info error handling Vinod Govindapillai
2024-04-19 17:11   ` Ville Syrjälä
2024-04-05 11:35 ` [PATCH v10 2/6] drm/i915/display: Extract code required to calculate max qgv/psf gv point Vinod Govindapillai
2024-04-08  8:32   ` Lisovskiy, Stanislav
2024-04-08 10:25     ` Govindapillai, Vinod
2024-04-05 11:35 ` [PATCH v10 3/6] drm/i915/display: extract code to prepare qgv points mask Vinod Govindapillai
2024-04-05 11:35 ` [PATCH v10 4/6] drm/i915/display: Disable SAGV on bw init, to force QGV point recalculation Vinod Govindapillai
2024-04-05 11:35 ` [PATCH v10 5/6] drm/i915/display: handle systems with duplicate psf gv points Vinod Govindapillai
2024-04-08  8:29   ` Lisovskiy, Stanislav
2024-04-08 10:18     ` Govindapillai, Vinod
2024-04-05 11:35 ` [PATCH v10 6/6] drm/i915/display: force qgv check after the hw state readout Vinod Govindapillai
2024-04-19 17:19   ` Ville Syrjälä
2024-04-23  7:40     ` Govindapillai, Vinod
2024-04-23  7:46       ` Govindapillai, Vinod [this message]
2024-04-22  7:54   ` Lisovskiy, Stanislav
2024-04-05 12:24 ` ✗ Fi.CI.SPARSE: warning for QGV/SAGV related fixes (rev10) Patchwork
2024-04-05 12:40 ` ✗ Fi.CI.BAT: failure " Patchwork
2024-04-06  5:02 ` ✓ Fi.CI.BAT: success " Patchwork
2024-04-06  7:38 ` ✗ Fi.CI.IGT: failure " Patchwork
2024-04-19  8:42 ` [PATCH v10 0/6] QGV/SAGV related fixes Hogander, Jouni
2024-04-19 13:32   ` Govindapillai, Vinod

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=e4a47c78906f7dcbb8c9314fdd84671d1821f918.camel@intel.com \
    --to=vinod.govindapillai@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jani.saarinen@intel.com \
    --cc=stanislav.lisovskiy@intel.com \
    --cc=ville.syrjala@intel.com \
    --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