Message ID | 1379701772-12818-1-git-send-email-jbarnes@virtuousgeek.org (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, Sep 20, 2013 at 11:29:32AM -0700, Jesse Barnes wrote: > Calculation is a little different than other platforms. > > v2: update to use port_clock instead > rebase on top of Ville's changes > v3: update to new port_clock semantics - don't divide by > pixel_multiplier (Ville) > > References: https://bugs.freedesktop.org/show_bug.cgi?id=67345 > Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> > --- > drivers/gpu/drm/i915/intel_display.c | 33 ++++++++++++++++++++++++++++++++- > 1 file changed, 32 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c > index 7eecf37..e5c9c1c 100644 > --- a/drivers/gpu/drm/i915/intel_display.c > +++ b/drivers/gpu/drm/i915/intel_display.c > @@ -5048,6 +5048,34 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc, > I915_READ(LVDS) & LVDS_BORDER_ENABLE; > } > > +static void vlv_crtc_clock_get(struct intel_crtc *crtc, > + struct intel_crtc_config *pipe_config) > +{ > + struct drm_device *dev = crtc->base.dev; > + struct drm_i915_private *dev_priv = dev->dev_private; > + int pipe = pipe_config->cpu_transcoder; > + intel_clock_t clock; > + u32 mdiv; > + int refclk = 100000, fastclk, update_rate; > + > + mutex_lock(&dev_priv->dpio_lock); > + mdiv = vlv_dpio_read(dev_priv, pipe, DPIO_DIV(pipe)); > + mutex_unlock(&dev_priv->dpio_lock); > + > + clock.m1 = (mdiv >> DPIO_M1DIV_SHIFT) & 7; > + clock.m2 = mdiv & DPIO_M2DIV_MASK; > + clock.n = (mdiv >> DPIO_N_SHIFT) & 0xf; > + clock.p1 = (mdiv >> DPIO_P1_SHIFT) & 7; > + clock.p2 = (mdiv >> DPIO_P2_SHIFT) & 0x1f; > + > + update_rate = refclk / clock.n; > + clock.vco = update_rate * clock.m1 * clock.m2; > + fastclk = clock.vco / clock.p1 / clock.p2; > + clock.dot = (2 * fastclk); > + > + pipe_config->port_clock = clock.dot / 10; Looks like it should get roughly the right answer, but I don't see much point in all the intermediate results. If you want to keep some of them for clarity, then I think this should be enough: clock.vco = refclk * clock.m / clock.n; clock.dot = clock.vco / clock.p; /* fast clock */ pipe_config->port_clock = clock.dot / 5; Although calling the fast clock "dot" is a bit wrong, but I think it's fine here, especially as it matches what I have in mind for vlv_find_best_dpll(). I had to rewrite that sucker a bit to make it compatible with my brain ;) I'll send a patch for that ASAP. > +} > + > static bool i9xx_get_pipe_config(struct intel_crtc *crtc, > struct intel_crtc_config *pipe_config) > { > @@ -5113,7 +5141,10 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc, > DPLL_PORTB_READY_MASK); > } > > - i9xx_crtc_clock_get(crtc, pipe_config); > + if (IS_VALLEYVIEW(dev)) > + vlv_crtc_clock_get(crtc, pipe_config); > + else > + i9xx_crtc_clock_get(crtc, pipe_config); > > return true; > } > -- > 1.7.9.5 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
On Mon, 23 Sep 2013 21:01:44 +0300 Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: > On Fri, Sep 20, 2013 at 11:29:32AM -0700, Jesse Barnes wrote: > > Calculation is a little different than other platforms. > > > > v2: update to use port_clock instead > > rebase on top of Ville's changes > > v3: update to new port_clock semantics - don't divide by > > pixel_multiplier (Ville) > > > > References: https://bugs.freedesktop.org/show_bug.cgi?id=67345 > > Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> > > --- > > drivers/gpu/drm/i915/intel_display.c | 33 ++++++++++++++++++++++++++++++++- > > 1 file changed, 32 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c > > index 7eecf37..e5c9c1c 100644 > > --- a/drivers/gpu/drm/i915/intel_display.c > > +++ b/drivers/gpu/drm/i915/intel_display.c > > @@ -5048,6 +5048,34 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc, > > I915_READ(LVDS) & LVDS_BORDER_ENABLE; > > } > > > > +static void vlv_crtc_clock_get(struct intel_crtc *crtc, > > + struct intel_crtc_config *pipe_config) > > +{ > > + struct drm_device *dev = crtc->base.dev; > > + struct drm_i915_private *dev_priv = dev->dev_private; > > + int pipe = pipe_config->cpu_transcoder; > > + intel_clock_t clock; > > + u32 mdiv; > > + int refclk = 100000, fastclk, update_rate; > > + > > + mutex_lock(&dev_priv->dpio_lock); > > + mdiv = vlv_dpio_read(dev_priv, pipe, DPIO_DIV(pipe)); > > + mutex_unlock(&dev_priv->dpio_lock); > > + > > + clock.m1 = (mdiv >> DPIO_M1DIV_SHIFT) & 7; > > + clock.m2 = mdiv & DPIO_M2DIV_MASK; > > + clock.n = (mdiv >> DPIO_N_SHIFT) & 0xf; > > + clock.p1 = (mdiv >> DPIO_P1_SHIFT) & 7; > > + clock.p2 = (mdiv >> DPIO_P2_SHIFT) & 0x1f; > > + > > + update_rate = refclk / clock.n; > > + clock.vco = update_rate * clock.m1 * clock.m2; > > + fastclk = clock.vco / clock.p1 / clock.p2; > > + clock.dot = (2 * fastclk); > > + > > + pipe_config->port_clock = clock.dot / 10; > > Looks like it should get roughly the right answer, but I don't see much > point in all the intermediate results. > > If you want to keep some of them for clarity, then I think this should > be enough: > > clock.vco = refclk * clock.m / clock.n; > clock.dot = clock.vco / clock.p; /* fast clock */ > pipe_config->port_clock = clock.dot / 5; > > Although calling the fast clock "dot" is a bit wrong, but I think it's > fine here, especially as it matches what I have in mind for > vlv_find_best_dpll(). I had to rewrite that sucker a bit to make it > compatible with my brain ;) I'll send a patch for that ASAP. The other advantage of having it all spelled out like this is that it matches the freq calculation spreadsheet I sourced it from. So I'd prefer to keep this as is, if I can have your r-b anyway. :) Thanks,
On Wed, Sep 25, 2013 at 10:38:33AM -0700, Jesse Barnes wrote: > On Mon, 23 Sep 2013 21:01:44 +0300 > Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: > > > On Fri, Sep 20, 2013 at 11:29:32AM -0700, Jesse Barnes wrote: > > > Calculation is a little different than other platforms. > > > > > > v2: update to use port_clock instead > > > rebase on top of Ville's changes > > > v3: update to new port_clock semantics - don't divide by > > > pixel_multiplier (Ville) > > > > > > References: https://bugs.freedesktop.org/show_bug.cgi?id=67345 > > > Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> > > > --- > > > drivers/gpu/drm/i915/intel_display.c | 33 ++++++++++++++++++++++++++++++++- > > > 1 file changed, 32 insertions(+), 1 deletion(-) > > > > > > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c > > > index 7eecf37..e5c9c1c 100644 > > > --- a/drivers/gpu/drm/i915/intel_display.c > > > +++ b/drivers/gpu/drm/i915/intel_display.c > > > @@ -5048,6 +5048,34 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc, > > > I915_READ(LVDS) & LVDS_BORDER_ENABLE; > > > } > > > > > > +static void vlv_crtc_clock_get(struct intel_crtc *crtc, > > > + struct intel_crtc_config *pipe_config) > > > +{ > > > + struct drm_device *dev = crtc->base.dev; > > > + struct drm_i915_private *dev_priv = dev->dev_private; > > > + int pipe = pipe_config->cpu_transcoder; > > > + intel_clock_t clock; > > > + u32 mdiv; > > > + int refclk = 100000, fastclk, update_rate; > > > + > > > + mutex_lock(&dev_priv->dpio_lock); > > > + mdiv = vlv_dpio_read(dev_priv, pipe, DPIO_DIV(pipe)); > > > + mutex_unlock(&dev_priv->dpio_lock); > > > + > > > + clock.m1 = (mdiv >> DPIO_M1DIV_SHIFT) & 7; > > > + clock.m2 = mdiv & DPIO_M2DIV_MASK; > > > + clock.n = (mdiv >> DPIO_N_SHIFT) & 0xf; > > > + clock.p1 = (mdiv >> DPIO_P1_SHIFT) & 7; > > > + clock.p2 = (mdiv >> DPIO_P2_SHIFT) & 0x1f; > > > + > > > + update_rate = refclk / clock.n; > > > + clock.vco = update_rate * clock.m1 * clock.m2; > > > + fastclk = clock.vco / clock.p1 / clock.p2; > > > + clock.dot = (2 * fastclk); > > > + > > > + pipe_config->port_clock = clock.dot / 10; > > > > Looks like it should get roughly the right answer, but I don't see much > > point in all the intermediate results. > > > > If you want to keep some of them for clarity, then I think this should > > be enough: > > > > clock.vco = refclk * clock.m / clock.n; > > clock.dot = clock.vco / clock.p; /* fast clock */ > > pipe_config->port_clock = clock.dot / 5; > > > > Although calling the fast clock "dot" is a bit wrong, but I think it's > > fine here, especially as it matches what I have in mind for > > vlv_find_best_dpll(). I had to rewrite that sucker a bit to make it > > compatible with my brain ;) I'll send a patch for that ASAP. > > The other advantage of having it all spelled out like this is that it > matches the freq calculation spreadsheet I sourced it from. > > So I'd prefer to keep this as is, if I can have your r-b anyway. :) Queued for -next with Ville's irc-r-b, thanks for the patch. -Daniel
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7eecf37..e5c9c1c 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -5048,6 +5048,34 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc, I915_READ(LVDS) & LVDS_BORDER_ENABLE; } +static void vlv_crtc_clock_get(struct intel_crtc *crtc, + struct intel_crtc_config *pipe_config) +{ + struct drm_device *dev = crtc->base.dev; + struct drm_i915_private *dev_priv = dev->dev_private; + int pipe = pipe_config->cpu_transcoder; + intel_clock_t clock; + u32 mdiv; + int refclk = 100000, fastclk, update_rate; + + mutex_lock(&dev_priv->dpio_lock); + mdiv = vlv_dpio_read(dev_priv, pipe, DPIO_DIV(pipe)); + mutex_unlock(&dev_priv->dpio_lock); + + clock.m1 = (mdiv >> DPIO_M1DIV_SHIFT) & 7; + clock.m2 = mdiv & DPIO_M2DIV_MASK; + clock.n = (mdiv >> DPIO_N_SHIFT) & 0xf; + clock.p1 = (mdiv >> DPIO_P1_SHIFT) & 7; + clock.p2 = (mdiv >> DPIO_P2_SHIFT) & 0x1f; + + update_rate = refclk / clock.n; + clock.vco = update_rate * clock.m1 * clock.m2; + fastclk = clock.vco / clock.p1 / clock.p2; + clock.dot = (2 * fastclk); + + pipe_config->port_clock = clock.dot / 10; +} + static bool i9xx_get_pipe_config(struct intel_crtc *crtc, struct intel_crtc_config *pipe_config) { @@ -5113,7 +5141,10 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc, DPLL_PORTB_READY_MASK); } - i9xx_crtc_clock_get(crtc, pipe_config); + if (IS_VALLEYVIEW(dev)) + vlv_crtc_clock_get(crtc, pipe_config); + else + i9xx_crtc_clock_get(crtc, pipe_config); return true; }
Calculation is a little different than other platforms. v2: update to use port_clock instead rebase on top of Ville's changes v3: update to new port_clock semantics - don't divide by pixel_multiplier (Ville) References: https://bugs.freedesktop.org/show_bug.cgi?id=67345 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> --- drivers/gpu/drm/i915/intel_display.c | 33 ++++++++++++++++++++++++++++++++- 1 file changed, 32 insertions(+), 1 deletion(-)