Message ID | 20160915151402.15992-5-wens@csie.org (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Hi, On Thu, Sep 15, 2016 at 11:14:02PM +0800, Chen-Yu Tsai wrote: > With display pixel clocks we want to have the closest possible clock > rate, to minimize timing and refresh rate skews. Whether the actual > clock rate is higher or lower than the requested rate is less important. > > Also check candidates against the requested rate, rather than the > ideal parent rate, the varying dividers also influence the difference > between the requested rate and the rounded rate. > > Signed-off-by: Chen-Yu Tsai <wens@csie.org> > --- > drivers/gpu/drm/sun4i/sun4i_dotclock.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c > index 3eb99784f371..d401156490f3 100644 > --- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c > +++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c > @@ -90,7 +90,8 @@ static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate, > goto out; > } > > - if ((rounded < ideal) && (rounded > best_parent)) { > + if (abs(rate - rounded / i) < > + abs(rate - best_parent / best_div)) { I'm not sure what you're trying to do here. Why is the divider involved? Maxime
On Mon, Sep 19, 2016 at 3:16 AM, Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > Hi, > > On Thu, Sep 15, 2016 at 11:14:02PM +0800, Chen-Yu Tsai wrote: >> With display pixel clocks we want to have the closest possible clock >> rate, to minimize timing and refresh rate skews. Whether the actual >> clock rate is higher or lower than the requested rate is less important. >> >> Also check candidates against the requested rate, rather than the >> ideal parent rate, the varying dividers also influence the difference >> between the requested rate and the rounded rate. >> >> Signed-off-by: Chen-Yu Tsai <wens@csie.org> >> --- >> drivers/gpu/drm/sun4i/sun4i_dotclock.c | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c >> index 3eb99784f371..d401156490f3 100644 >> --- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c >> +++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c >> @@ -90,7 +90,8 @@ static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate, >> goto out; >> } >> >> - if ((rounded < ideal) && (rounded > best_parent)) { >> + if (abs(rate - rounded / i) < >> + abs(rate - best_parent / best_div)) { > > I'm not sure what you're trying to do here. Why is the divider involved? Say you want the dotclock at X, so you try Y = 6 ~ 127 for the divider. Now you're asking the CCF to round (X*Y). In the original code, you were comparing the result of rounding (X * Y). if ((rounded < ideal) && (rounded > best_parent)) { best_parent = rounded; best_div = i; } where ideal = X * Y (i in the code). Given the divider increases in the loop, you are actually not closing in on the best divider, but the highest divider that doesn't give a higher rate than the ideal rate. Including the divider makes it compare the actual dot clock frequency if a given divider was used. Does this makes sense? Explaining this kind of makes my head spin... Regards ChenYu > > Maxime > > -- > Maxime Ripard, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
On Thu, Sep 15, 2016 at 11:14:02PM +0800, Chen-Yu Tsai wrote: > With display pixel clocks we want to have the closest possible clock > rate, to minimize timing and refresh rate skews. Whether the actual > clock rate is higher or lower than the requested rate is less important. > > Also check candidates against the requested rate, rather than the > ideal parent rate, the varying dividers also influence the difference > between the requested rate and the rounded rate. > > Signed-off-by: Chen-Yu Tsai <wens@csie.org> Applied, thanks! Maxime
On Mon, Sep 19, 2016 at 11:36:18PM +0800, Chen-Yu Tsai wrote: > On Mon, Sep 19, 2016 at 3:16 AM, Maxime Ripard > <maxime.ripard@free-electrons.com> wrote: > > Hi, > > > > On Thu, Sep 15, 2016 at 11:14:02PM +0800, Chen-Yu Tsai wrote: > >> With display pixel clocks we want to have the closest possible clock > >> rate, to minimize timing and refresh rate skews. Whether the actual > >> clock rate is higher or lower than the requested rate is less important. > >> > >> Also check candidates against the requested rate, rather than the > >> ideal parent rate, the varying dividers also influence the difference > >> between the requested rate and the rounded rate. > >> > >> Signed-off-by: Chen-Yu Tsai <wens@csie.org> > >> --- > >> drivers/gpu/drm/sun4i/sun4i_dotclock.c | 3 ++- > >> 1 file changed, 2 insertions(+), 1 deletion(-) > >> > >> diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c > >> index 3eb99784f371..d401156490f3 100644 > >> --- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c > >> +++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c > >> @@ -90,7 +90,8 @@ static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate, > >> goto out; > >> } > >> > >> - if ((rounded < ideal) && (rounded > best_parent)) { > >> + if (abs(rate - rounded / i) < > >> + abs(rate - best_parent / best_div)) { > > > > I'm not sure what you're trying to do here. Why is the divider involved? > > Say you want the dotclock at X, so you try Y = 6 ~ 127 for the divider. > Now you're asking the CCF to round (X*Y). > > In the original code, you were comparing the result of rounding (X * Y). > > if ((rounded < ideal) && (rounded > best_parent)) { > best_parent = rounded; > best_div = i; > } > > where ideal = X * Y (i in the code). Given the divider increases in > the loop, you are actually not closing in on the best divider, but the > highest divider that doesn't give a higher rate than the ideal rate. > > Including the divider makes it compare the actual dot clock frequency > if a given divider was used. > > Does this makes sense? Explaining this kind of makes my head spin... Yes, sorry, I didn't remember rounded was actually the rounded parent rate. Thanks! Maxime
diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c index 3eb99784f371..d401156490f3 100644 --- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c +++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c @@ -90,7 +90,8 @@ static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate, goto out; } - if ((rounded < ideal) && (rounded > best_parent)) { + if (abs(rate - rounded / i) < + abs(rate - best_parent / best_div)) { best_parent = rounded; best_div = i; }
With display pixel clocks we want to have the closest possible clock rate, to minimize timing and refresh rate skews. Whether the actual clock rate is higher or lower than the requested rate is less important. Also check candidates against the requested rate, rather than the ideal parent rate, the varying dividers also influence the difference between the requested rate and the rounded rate. Signed-off-by: Chen-Yu Tsai <wens@csie.org> --- drivers/gpu/drm/sun4i/sun4i_dotclock.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)