Re: [PATCH 06/11] drm/i915: Add intel_dotclock_calculate()

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Fri, Sep 13, 2013 at 03:30:51PM +0300, Jani Nikula wrote:
> On Fri, 06 Sep 2013, ville.syrjala@linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >
> > Extract the code to calculate the dotclock from the link clock and M/N
> > values into a new function from ironlake_crtc_clock_get().
> >
> > The new function can be used to calculate the dotclock for both FDI and
> > DP cases.
> >
> > Also simplify the code a bit along the way.
> >
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_display.c | 40 +++++++++++++++---------------------
> >  drivers/gpu/drm/i915/intel_drv.h     |  2 ++
> >  2 files changed, 19 insertions(+), 23 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > index b3049a6..c393c8e 100644
> > --- a/drivers/gpu/drm/i915/intel_display.c
> > +++ b/drivers/gpu/drm/i915/intel_display.c
> > @@ -7410,16 +7410,9 @@ static void i9xx_crtc_clock_get(struct intel_crtc *crtc,
> >  	pipe_config->adjusted_mode.clock = clock.dot;
> >  }
> >  
> > -static void ironlake_crtc_clock_get(struct intel_crtc *crtc,
> > -				    struct intel_crtc_config *pipe_config)
> > +int intel_dotclock_calculate(int link_freq,
> > +			     const struct intel_link_m_n *m_n)
> >  {
> > -	struct drm_device *dev = crtc->base.dev;
> > -	struct drm_i915_private *dev_priv = dev->dev_private;
> > -	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
> > -	int link_freq;
> > -	u64 clock;
> > -	u32 link_m, link_n;
> > -
> >  	/*
> >  	 * The calculation for the data clock is:
> >  	 * pixel_clock = ((m/n)*(link_clock * nr_lanes))/bpp
> > @@ -7430,6 +7423,18 @@ static void ironlake_crtc_clock_get(struct intel_crtc *crtc,
> >  	 * link_clock = (m * link_clock) / n
> >  	 */
> >  
> > +	if (!m_n->link_n)
> > +		return 0;
> > +
> > +	return div_u64((u64)m_n->link_m * link_freq, m_n->link_n);
> > +}
> > +
> > +static void ironlake_crtc_clock_get(struct intel_crtc *crtc,
> > +				    struct intel_crtc_config *pipe_config)
> > +{
> > +	struct drm_device *dev = crtc->base.dev;
> > +	int link_freq;
> > +
> >  	/*
> >  	 * We need to get the FDI or DP link clock here to derive
> >  	 * the M/N dividers.
> > @@ -7438,21 +7443,10 @@ static void ironlake_crtc_clock_get(struct intel_crtc *crtc,
> >  	 * For DP, it's either 1.62GHz or 2.7GHz.
> >  	 * We do our calculations in 10*MHz since we don't need much precison.
> >  	 */
> > -	if (pipe_config->has_pch_encoder)
> > -		link_freq = intel_fdi_link_freq(dev) * 10000;
> > -	else
> > -		link_freq = pipe_config->port_clock;
> 
> The new code loses this distinction. I don't know if it matters. Please
> explain.
> 
> > -
> > -	link_m = I915_READ(PIPE_LINK_M1(cpu_transcoder));
> > -	link_n = I915_READ(PIPE_LINK_N1(cpu_transcoder));
> > -
> > -	if (!link_m || !link_n)
> > -		return;
> > -
> > -	clock = ((u64)link_m * (u64)link_freq);
> > -	do_div(clock, link_n);
> > +	link_freq = intel_fdi_link_freq(dev) * 10000;
> >  
> > -	pipe_config->adjusted_mode.clock = clock;
> > +	pipe_config->adjusted_mode.clock =
> > +		intel_dotclock_calculate(link_freq, &pipe_config->fdi_m_n);
> 
> I'm not sure if the fear in unwarranted, but can we always be sure
> fdi_m_n has been set?

I seem to have jumped the gun a bit with this patch. Patch 9 will make
it so that we don't call ironlake_crtc_clock_get() except in the PCH
encoder case (could use a rename of the function at that point maybe).
Before patch 9 we still call this guy for non-PCH cases and then it
might fall apart a bit. As the dp_m_n extraction patch appears before
this patch in the series, I think I can just make a slightly fixed
version of this patch. Stay tuned for v2...

-- 
Ville Syrjälä
Intel OTC