Re: [PATCH v3 12/12] arm64: dts: h3ulcb: Sound SSI support

2016-09-02 Thread Vladimir Barinov

Hi Sergei,

Thank you for the review.
Lemme fix it in v5 patchset.

Regards,
Vladimir

On 02.09.2016 15:38, Sergei Shtylyov wrote:

Hello.

On 9/2/2016 3:05 PM, Vladimir Barinov wrote:


This supports SSI sound for H3ULCB board.
SSI DMA mode used. CS2000 used as AUDIO_CLK_B.

Signed-off-by: Vladimir Barinov 
Acked-by: Kuninori Morimoto 
---
Changes in version 3:
- initially added
Changes in version 4:
- added comment for audio_clkout

 arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts | 115 
+

 1 file changed, 115 insertions(+)

diff --git a/arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts 
b/arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts

index bbd7928..a254231 100644
--- a/arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts
+++ b/arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts
@@ -45,6 +45,12 @@
 };
 };

+x12_clk: x12_clk {


   If the crystal name is X12, name the node "x12" too, this way the 
output clock will be named just "x12".



+compatible = "fixed-clock";
+#clock-cells = <0>;
+clock-frequency = <24576000>;
+};
+
 vcc_sdhi0: regulator-vcc-sdhi0 {
 compatible = "regulator-fixed";


[...]

@@ -136,6 +176,73 @@

[...]

+
+cs2000: clk_multiplier@4f {


   Don't use underscores in the node names, use hyphens instead.


+#clock-cells = <0>;
+compatible = "cirrus,cs2000-cp";
+reg = <0x4f>;
+clocks = <_clkout>, <_clk>;
+clock-names = "clk_in", "ref_clk";
+
+assigned-clocks = <>;
+assigned-clock-rates = <24576000>; /* 1/1 divide */
+};
+};

[...]

MBR, Sergei





Re: [PATCH v3 12/12] arm64: dts: h3ulcb: Sound SSI support

2016-09-02 Thread Vladimir Barinov

On 02.09.2016 15:05, Vladimir Barinov wrote:

This supports SSI sound for H3ULCB board.
SSI DMA mode used. CS2000 used as AUDIO_CLK_B.

Signed-off-by: Vladimir Barinov 
Acked-by: Kuninori Morimoto 
---
Changes in version 3:
- initially added
Changes in version 4:
- added comment for audio_clkout

  arch/arm64/boot/dts/renesas/r8a7795-h3ulcb.dts | 115 +
  1 file changed, 115 insertions(+)


Sorry for the subject typo. The proper one is v4:
[PATCH v4 4/4] arm64: dts: h3ulcb: Sound SSI support



Re: [PATCH v3 12/12] arm64: dts: h3ulcb: Sound SSI support

2016-09-02 Thread Simon Horman
On Fri, Sep 02, 2016 at 01:19:49AM +, Kuninori Morimoto wrote:
> 
> Hi Vladimir
> 
> > This supports SSI sound for H3ULCB board.
> > SSI DMA mode used. CS2000 used as AUDIO_CLK_B.
> > 
> > Signed-off-by: Vladimir Barinov 
> > ---
> 
> Acked-by: Kuninori Morimoto 
> 
> > +   audio_clkout: audio_clkout {
> > +   compatible = "fixed-clock";
> > +   #clock-cells = <0>;
> > +   clock-frequency = <11289600>;
> > +   };
> 
> I think having commont for AUDIO_CLKOUT is better for user, like Salvator-X.
> This "audio_clkout" is dummy. Actually, we don't need this dummy clock,
> and we should use "<_sound 0>" as AUDIO_CLKOUT,
> like AUDIO_CLKOUT3 as "<_sound 3>" which is used on ak4613. 
> But Salvator-X and this ULCB board has very complex connection,
> especially around CS2000.
> This means cs2000 requests rcar_sound, and rcar_sound requests cs2000.
> This audio_clkout exist to avoid probe dead-lock.
> This is the reason why audio_clkout and rcar_sound::clock-frequency
> has same value.

Vladimir,

please address the above and include it when you repost the patches
that I haven't queued up.


Re: [PATCH v3 12/12] arm64: dts: h3ulcb: Sound SSI support

2016-09-01 Thread Kuninori Morimoto

Hi Vladimir

> This supports SSI sound for H3ULCB board.
> SSI DMA mode used. CS2000 used as AUDIO_CLK_B.
> 
> Signed-off-by: Vladimir Barinov 
> ---

Acked-by: Kuninori Morimoto 

> + audio_clkout: audio_clkout {
> + compatible = "fixed-clock";
> + #clock-cells = <0>;
> + clock-frequency = <11289600>;
> + };

I think having commont for AUDIO_CLKOUT is better for user, like Salvator-X.
This "audio_clkout" is dummy. Actually, we don't need this dummy clock,
and we should use "<_sound 0>" as AUDIO_CLKOUT,
like AUDIO_CLKOUT3 as "<_sound 3>" which is used on ak4613. 
But Salvator-X and this ULCB board has very complex connection,
especially around CS2000.
This means cs2000 requests rcar_sound, and rcar_sound requests cs2000.
This audio_clkout exist to avoid probe dead-lock.
This is the reason why audio_clkout and rcar_sound::clock-frequency
has same value.