[PATCH v4 27/28] gcc-sdm845: Add general purpose clock ops
Dzmitry Sankouski
dsankouski at gmail.com
Fri Sep 13 15:06:36 UTC 2024
вт, 10 сент. 2024 г. в 12:59, Uwe Kleine-König <u.kleine-koenig at baylibre.com>:
>
> Hello,
>
> On Fri, Jul 19, 2024 at 03:55:04PM +0300, Dzmitry Sankouski wrote:
> > SDM845 has "General Purpose" clocks that can be muxed to
> > SoC pins to clock various external devices.
> > Those clocks may be used as e.g. PWM sources for external peripherals.
> >
> > GPCLK can in theory have arbitrary value depending on the use case, so
> > the concept of frequency tables, used in rcg2 clock driver, is not
> > efficient, because it allows only defined frequencies.
> >
> > Introduce clk_rcg2_gp_ops, which automatically calculate clock
> > mnd values for arbitrary clock rate.
> >
> > Signed-off-by: Dzmitry Sankouski <dsankouski at gmail.com>
> > ---
> > drivers/clk/qcom/clk-rcg.h | 1 +
> > drivers/clk/qcom/clk-rcg2.c | 162 ++++++++++++++++++++++++++++++++++++++++--
> > drivers/clk/qcom/gcc-sdm845.c | 19 ++---
> > drivers/pwm/pwm-clk.c | 5 ++
>
> I don't understand why a change to some qcom clk implementation detail
> needs a change to drivers/pwm/pwm-clk.c in the same commit. I guess if
> the change to drivers/pwm/pwm-clk.c is needed it should better go into a
> separate patch with an appropriate commit log?!
>
You're right, I'll fix that. Also I heavily rewrote this patch, and
included it in
starqltechn patch series. Since this patch has the wrong version (sorry for my
mistake) and order numbering, please consider the next patch a successor of it.
--
Best regards,
Dzmitry
More information about the dri-devel
mailing list