Re: Regression found (Stop-marking-clocks-as-CLK_IS_CRITICAL)

[Pierre-Louis Bossart]

On 10/30/18 10:03 AM, Andy Shevchenko wrote:
> On Tue, Oct 30, 2018 at 4:48 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
> > On 30-10-18 15:38, Dean Wallace wrote:
> > > Excellent work Hans.  Compiled 4.19 with
> > > 0001-ASoC-intel-cht_bsw_max98090_ti-Use-pmc_plt_clk_0-ins.patch, sound
> > > works as before.
> > >
> > > for i in /sys/kernel/debug/clk/pmc_plt_clk_?; do echo -n "$i: "; cat $i/clk_flags; echo; done
> > > /sys/kernel/debug/clk/pmc_plt_clk_0:
> > > /sys/kernel/debug/clk/pmc_plt_clk_1:
> > > /sys/kernel/debug/clk/pmc_plt_clk_2:
> > > /sys/kernel/debug/clk/pmc_plt_clk_3:
> > > /sys/kernel/debug/clk/pmc_plt_clk_4:
> > > /sys/kernel/debug/clk/pmc_plt_clk_5:
> > Ok, so as I expected the Swanky is using pmc_plt_clk_0
> > as mclk instead of pmc_plt_clk_3. Now the question
> > becomes is this true for all the designs using the
> > max98090 codec?
> > 1) Unconditionally use pmc_plt_clk_0 as mclk (as my test patch does)
> > in the cht_bsw_max98090_ti.c machine driver
> I don't know the details, but the main question here indeed do we ever
> had a working example of that machine with CLK #3 in use?
yes, Rambi/Orco (Baytrail based).
> P.S. I would go your first proposal until the opposite will be proved.