Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1711942 > unrolled thread

[PATCH] ASoC: mediatek: Correct the interrupt property

Back to article view | Back to linux.kernel

Contents

  [PATCH] ASoC: mediatek: Correct the interrupt property Ryder Lee <ryder.lee@mediatek.com> - 2017-08-15 10:10 +0200
    Re: [PATCH] ASoC: mediatek: Correct the interrupt property Mark Brown <broonie@kernel.org> - 2017-08-15 14:00 +0200
      Re: [PATCH] ASoC: mediatek: Correct the interrupt property Ryder Lee <ryder.lee@mediatek.com> - 2017-08-15 16:00 +0200
        Re: [PATCH] ASoC: mediatek: Correct the interrupt property Mark Brown <broonie@kernel.org> - 2017-08-15 16:00 +0200
        Re: [PATCH] ASoC: mediatek: Correct the interrupt property Rob Herring <robh@kernel.org> - 2017-08-17 23:30 +0200

#1711942 — [PATCH] ASoC: mediatek: Correct the interrupt property

This patch update interrupt property since we only use ASYS IRQ
in the driver, and this may confuse people.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
CC: Garlic Tseng <garlic.tseng@mediatek.com>
---
 Documentation/devicetree/bindings/sound/mt2701-afe-pcm.txt | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/sound/mt2701-afe-pcm.txt b/Documentation/devicetree/bindings/sound/mt2701-afe-pcm.txt
index 9800a56..9d7b6dc 100644
--- a/Documentation/devicetree/bindings/sound/mt2701-afe-pcm.txt
+++ b/Documentation/devicetree/bindings/sound/mt2701-afe-pcm.txt
@@ -3,7 +3,7 @@ Mediatek AFE PCM controller for mt2701
 Required properties:
 - compatible = "mediatek,mt2701-audio";
 - reg: register location and size
-- interrupts: Should contain AFE interrupt
+- interrupts: Should contain ASYS interrupt
 - power-domains: should define the power domain
 - clock-names: should have these clock names:
 		"infra_sys_audio_clk",
@@ -57,8 +57,7 @@ Example:
 		compatible = "mediatek,mt2701-audio";
 		reg = <0 0x11220000 0 0x2000>,
 		      <0 0x112A0000 0 0x20000>;
-		interrupts = <GIC_SPI 104 IRQ_TYPE_LEVEL_LOW>,
-			     <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
+		interrupts = <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
 		power-domains = <&scpsys MT2701_POWER_DOMAIN_IFR_MSC>;
 		clocks = <&infracfg CLK_INFRA_AUDIO>,
 			 <&topckgen CLK_TOP_AUD_MUX1_SEL>,
-- 
1.9.1

[toc] | [next] | [standalone]

#1712113

[Multipart message — attachments visible in raw view] — view raw

On Tue, Aug 15, 2017 at 04:07:02PM +0800, Ryder Lee wrote:

> This patch update interrupt property since we only use ASYS IRQ
> in the driver, and this may confuse people.

> -		interrupts = <GIC_SPI 104 IRQ_TYPE_LEVEL_LOW>,
> -			     <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
> +		interrupts = <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;

What harm does having the extra information in the bindings do?  If it's
possible there might be a use for the extra interrupt it seems better to
have people describe it.

[toc] | [prev] | [next] | [standalone]

#1712199

On Tue, 2017-08-15 at 12:53 +0100, Mark Brown wrote:
> On Tue, Aug 15, 2017 at 04:07:02PM +0800, Ryder Lee wrote:
> 
> > This patch update interrupt property since we only use ASYS IRQ
> > in the driver, and this may confuse people.
> 
> > -		interrupts = <GIC_SPI 104 IRQ_TYPE_LEVEL_LOW>,
> > -			     <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
> > +		interrupts = <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
> 
> What harm does having the extra information in the bindings do?  If it's
> possible there might be a use for the extra interrupt it seems better to
> have people describe it.

Yes you're right. The current driver get IRQ number by index "0" but
actually it should be "1" (GIC 132). Perhaps we can switch to use the
platform_get_irq_byname() and add interrupt-name in DT so that binding
can be agnostic of the IRQ order. 

[toc] | [prev] | [next] | [standalone]

#1712200

[Multipart message — attachments visible in raw view] — view raw

On Tue, Aug 15, 2017 at 09:51:40PM +0800, Ryder Lee wrote:
> On Tue, 2017-08-15 at 12:53 +0100, Mark Brown wrote:

> > What harm does having the extra information in the bindings do?  If it's
> > possible there might be a use for the extra interrupt it seems better to
> > have people describe it.

> Yes you're right. The current driver get IRQ number by index "0" but
> actually it should be "1" (GIC 132). Perhaps we can switch to use the
> platform_get_irq_byname() and add interrupt-name in DT so that binding
> can be agnostic of the IRQ order. 

Yes, that's generally good - it also means that if future revisions of
the IP have different sets of interrupts the binding can adapt
gracefully.

[toc] | [prev] | [next] | [standalone]

#1714391

On Tue, Aug 15, 2017 at 09:51:40PM +0800, Ryder Lee wrote:
> On Tue, 2017-08-15 at 12:53 +0100, Mark Brown wrote:
> > On Tue, Aug 15, 2017 at 04:07:02PM +0800, Ryder Lee wrote:
> > 
> > > This patch update interrupt property since we only use ASYS IRQ
> > > in the driver, and this may confuse people.
> > 
> > > -		interrupts = <GIC_SPI 104 IRQ_TYPE_LEVEL_LOW>,
> > > -			     <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
> > > +		interrupts = <GIC_SPI 132 IRQ_TYPE_LEVEL_LOW>;
> > 
> > What harm does having the extra information in the bindings do?  If it's
> > possible there might be a use for the extra interrupt it seems better to
> > have people describe it.

Agreed. It should describe interrupts the h/w has, not what the driver 
(currently) uses.
> 
> Yes you're right. The current driver get IRQ number by index "0" but
> actually it should be "1" (GIC 132). Perhaps we can switch to use the
> platform_get_irq_byname() and add interrupt-name in DT so that binding
> can be agnostic of the IRQ order. 

The binding is wrong already in that it says a single interrrupt. The 
example has 2 though. Add the 2nd interrupt and define the order such 
that you don't break the current driver. Probably that means ASYS is 
first.

Rob

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web