On Tue, 2018-08-28 at 11:35 +0800, Jian Hu wrote:
> Hi: Jerome
> 
> On 2018/8/27 21:07, Jerome Brunet wrote:
> > On Fri, 2018-08-24 at 21:34 +0800, Jian Hu wrote:
> > > > 
> > > 
> > > I am confued about aoclk81's parent clocks.
> > > 
> > > I can not get the example of axg audio clock driver, Could you provide
> > > the link? Had it merged into clk-meson.git?
> > 
> > Yes and mainline as well : drivers/clk/meson/axg-audio.c
> > 
> > Basically this driver is creating bypass input clocks (audio_pclk, 
> > mst_in[0-9],
> > etc...) .
> > 
> > This allows to collect input clocks from DT (like any consumer should) will
> > keeping constant in the controller clock tree.
> > 
> > > From what I've seen of your controller drivers, the EE controller should 
> > > have
> > 
> > one input, the AO should have 3.
> > 
> > 
> > 
> > .
> > 
> 
> I still can not get the example meaning in axg audio driver.

here is one specific example:


https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/clk/meson/axg-audio.c?h=v4.19-rc1#n785


> 
> In 26 page of A113D_Datasheet V0.7 20170725-Baylibre.pdf,We can see the 
> aoclk81 has two parents. clk81 and ao_slow_clk. I can not get 3 parents.
> 
>                           clk81|\
>                         -------| \        aoclk81
> src0 |\                       |  |-------------------
> -----| \     ao_slow_clk      |  |
>       |  |---------------------| /
> -----| /                      |/
>   src1|/
> 
> 
> src0 is from xtal, if can generate 32k clock, but it is never used.
> src1 is from gpio clock, It is never used. If necessary, the ao_slow_clk 
> maybe described.
> 
> So why aoclk81 has 3 parents?

I never said that. I said your clock controller driver should have 3 input
clocks. These clock are the one used by the controller but not produced by it:
for the AO controller xtal, ao_xtal, and ee controller clk81.



Reply via email to