Re: [Intel-gfx] [PATCH 1/3] drm/i915/display: Move dc3co_exitline variable to struct intel_psr - Mun, Gwan-gyeong

From: "Mun, Gwan-gyeong" <gwan-gyeong.mun@intel.com>
To: "ville.syrjala@linux.intel.com" <ville.syrjala@linux.intel.com>
Cc: "intel-gfx@lists.freedesktop.org" <intel-gfx@lists.freedesktop.org>
Subject: Re: [Intel-gfx] [PATCH 1/3] drm/i915/display: Move dc3co_exitline variable to struct intel_psr
Date: Fri, 5 Mar 2021 20:00:09 +0000	[thread overview]
Message-ID: <93deb32a47a36f95f1e1847a23fe3f10037d541a.camel@intel.com> (raw)
In-Reply-To: <YD+/Q45My7lyi0ow@intel.com>

On Wed, 2021-03-03 at 18:54 +0200, Ville Syrjälä wrote:
> On Wed, Mar 03, 2021 at 06:41:59PM +0200, Gwan-gyeong Mun wrote:
> > dc3co_exitline is indirectly called by intel_psr_compute_config().
> > And it will not be changed until the next calling of
> > intel_psr_compute_config(). So in order to use dc3co_exitline without
> > intel_crtc_state on other psr internal function, it moves
> > dc3co_exitline
> > variable to struct intel_psr.
> > And it removes a dc3co_enabled variable from struct intel_psr.
> > 
> > Cc: José Roberto de Souza <jose.souza@intel.com>
> > Cc: Anshuman Gupta <anshuman.gupta@intel.com>
> > Signed-off-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_display_types.h |  3 +--
> >  drivers/gpu/drm/i915/display/intel_psr.c           | 11 +++++------
> >  2 files changed, 6 insertions(+), 8 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h
> > b/drivers/gpu/drm/i915/display/intel_display_types.h
> > index 1a76e1d9de7a..f69bd1caebbf 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_types.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h
> > @@ -1002,7 +1002,6 @@ struct intel_crtc_state {
> >         bool has_psr;
> >         bool has_psr2;
> >         bool enable_psr2_sel_fetch;
> > -       u32 dc3co_exitline;
> >  
> >         /*
> >          * Frequence the dpll for the port should run at. Differs
> > from the
> > @@ -1453,7 +1452,7 @@ struct intel_psr {
> >         bool sink_not_reliable;
> >         bool irq_aux_error;
> >         u16 su_x_granularity;
> > -       bool dc3co_enabled;
> > +       u32 dc3co_exitline;
> >         u32 dc3co_exit_delay;
> >         struct delayed_work dc3co_work;
> >         struct drm_dp_vsc_sdp vsc;
> > diff --git a/drivers/gpu/drm/i915/display/intel_psr.c
> > b/drivers/gpu/drm/i915/display/intel_psr.c
> > index cd434285e3b7..976826653143 100644
> > --- a/drivers/gpu/drm/i915/display/intel_psr.c
> > +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> > @@ -637,7 +637,7 @@ static void tgl_dc3co_disable_work(struct
> > work_struct *work)
> >  
> >  static void tgl_disallow_dc3co_on_psr2_exit(struct intel_dp
> > *intel_dp)
> >  {
> > -       if (!intel_dp->psr.dc3co_enabled)
> > +       if (!intel_dp->psr.dc3co_exitline)
> >                 return;
> >  
> >         cancel_delayed_work(&intel_dp->psr.dc3co_work);
> > @@ -679,7 +679,7 @@ tgl_dc3co_exitline_compute_config(struct intel_dp
> > *intel_dp,
> >         if (drm_WARN_ON(&dev_priv->drm, exit_scanlines >
> > crtc_vdisplay))
> >                 return;
> >  
> > -       crtc_state->dc3co_exitline = crtc_vdisplay - exit_scanlines;
> > +       intel_dp->psr.dc3co_exitline = crtc_vdisplay -
> > exit_scanlines;
> 
> Thou shalt not mutate externally visible state in .compute_config().
> You either want to make a copy of it or just compute it on the spot in
> the psr_enable(). The first option is a good choice when you can also
> hook into into the readout+state checker.
> 
Hi, thanks for pointing out where I was approached incorrectly.
I'll float new patch that makes copy crtc_state's dc3co_exitline to
intel_psr.dc3co_exitline.

> >  }
> >  
> >  static bool intel_psr2_sel_fetch_config_valid(struct intel_dp
> > *intel_dp,
> > @@ -938,7 +938,7 @@ static void intel_psr_enable_source(struct
> > intel_dp *intel_dp,
> >  
> >         psr_irq_control(intel_dp);
> >  
> > -       if (crtc_state->dc3co_exitline) {
> > +       if (intel_dp->psr.dc3co_exitline) {
> >                 u32 val;
> >  
> >                 /*
> > @@ -947,7 +947,7 @@ static void intel_psr_enable_source(struct
> > intel_dp *intel_dp,
> >                  */
> >                 val = intel_de_read(dev_priv,
> > EXITLINE(cpu_transcoder));
> >                 val &= ~EXITLINE_MASK;
> > -               val |= crtc_state->dc3co_exitline << EXITLINE_SHIFT;
> > +               val |= intel_dp->psr.dc3co_exitline <<
> > EXITLINE_SHIFT;
> >                 val |= EXITLINE_ENABLE;
> >                 intel_de_write(dev_priv, EXITLINE(cpu_transcoder),
> > val);
> >         }
> > @@ -972,7 +972,6 @@ static void intel_psr_enable_locked(struct
> > intel_dp *intel_dp,
> >         intel_dp->psr.psr2_enabled = crtc_state->has_psr2;
> >         intel_dp->psr.busy_frontbuffer_bits = 0;
> >         intel_dp->psr.pipe = to_intel_crtc(crtc_state->uapi.crtc)-
> > >pipe;
> > -       intel_dp->psr.dc3co_enabled = !!crtc_state->dc3co_exitline;
> >         intel_dp->psr.transcoder = crtc_state->cpu_transcoder;
> >         /* DC5/DC6 requires at least 6 idle frames */
> >         val = usecs_to_jiffies(intel_get_frame_time_us(crtc_state) *
> > 6);
> > @@ -1761,7 +1760,7 @@ tgl_dc3co_flush(struct intel_dp *intel_dp,
> > unsigned int frontbuffer_bits,
> >  {
> >         mutex_lock(&intel_dp->psr.lock);
> >  
> > -       if (!intel_dp->psr.dc3co_enabled)
> > +       if (!intel_dp->psr.dc3co_exitline)
> >                 goto unlock;
> >  
> >         if (!intel_dp->psr.psr2_enabled || !intel_dp->psr.active)
> > -- 
> > 2.30.1
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx