diff mbox

[v1] arm: pxa: fix DT node name for PXA27X usb

Message ID 1456139297-5854-1-git-send-email-ynvich@gmail.com (mailing list archive)
State New, archived
Headers show

Commit Message

Sergey Yanovich Feb. 22, 2016, 11:08 a.m. UTC
Without this patch I am getting an error:
---8<---
[   15.709742] ohci-pxa27x: OHCI PXA27x/PXA3x driver
[   15.737150] pxa27x-ohci: probe of 4c000000.ohci failed with error -2
---8<---

The error is generated in drivers/usb/host/ohci-pxa27x.c at line 441:
---8<---
441:	usb_clk = devm_clk_get(&pdev->dev, NULL);
442:	if (IS_ERR(usb_clk))
443:		return PTR_ERR(usb_clk);
---8<---

The error is caused by different names for the same DT node in pxa2xx.dtsi
and pxa27x.dtsi.

Signed-off-by: Sergei Ianovich <ynvich@gmail.com>
CC: Robert Jarzmik <robert.jarzmik@free.fr>
---
 arch/arm/boot/dts/pxa27x.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Robert Jarzmik Feb. 22, 2016, 3:35 p.m. UTC | #1
Sergei Ianovich <ynvich@gmail.com> writes:

> Without this patch I am getting an error:
> ---8<---
> [   15.709742] ohci-pxa27x: OHCI PXA27x/PXA3x driver
> [   15.737150] pxa27x-ohci: probe of 4c000000.ohci failed with error -2
> ---8<---
>
> The error is generated in drivers/usb/host/ohci-pxa27x.c at line 441:
> ---8<---
> 441:	usb_clk = devm_clk_get(&pdev->dev, NULL);
> 442:	if (IS_ERR(usb_clk))
> 443:		return PTR_ERR(usb_clk);
> ---8<---
>
> The error is caused by different names for the same DT node in pxa2xx.dtsi
> and pxa27x.dtsi.
>
> Signed-off-by: Sergei Ianovich <ynvich@gmail.com>
> CC: Robert Jarzmik <robert.jarzmik@free.fr>
Hi Sergei,

You're right, I haven't seen the pxa2xx.dtsi statement before.

As such, could you amend a bit your patch please to :
 - add:
Fixes: 0ec1939668e5 ("ARM: dts: pxa: add the usb host controller")
 - make the same change in pxa3xx.dtsi, as the same error is there
 - and more globally align pxa2xx.dtsi, pxa27x.dtsi and pxa3xx.dtsi
 - remove the compatible, reg, interrupts and status from both pxa27x.dtsi and
   pxa3xx.dtsi as they are redundant with the included pxa2xxx.dsti one

I'm even wondering if the proper change wouldn't be to scrap the ohci
declarations from pxa27x.dtsi and pxa3xx.dtsi, and only add the "clocks"
property to pxa2xx.dtsi ...

Cheers.

--
Robert
Sergey Yanovich Feb. 23, 2016, 12:01 p.m. UTC | #2
Hi Robert, 

On Mon, 2016-02-22 at 16:35 +0100, Robert Jarzmik wrote:
> You're right, I haven't seen the pxa2xx.dtsi statement before.
> 
> As such, could you amend a bit your patch please to :
>  - add:
> Fixes: 0ec1939668e5 ("ARM: dts: pxa: add the usb host controller")
>  - make the same change in pxa3xx.dtsi, as the same error is there
>  - and more globally align pxa2xx.dtsi, pxa27x.dtsi and pxa3xx.dtsi
>  - remove the compatible, reg, interrupts and status from both pxa27x.dtsi and
>    pxa3xx.dtsi as they are redundant with the included pxa2xxx.dsti one

No problem. Let's figure out the best approach.

> I'm even wondering if the proper change wouldn't be to scrap the ohci
> declarations from pxa27x.dtsi and pxa3xx.dtsi, and only add the "clocks"
> property to pxa2xx.dtsi ...

Since clocks are declared in pxa27x.dtsi and pxa3xx.dtsi, there is a
stronger ground to have node clocks properties set in pxa27x.dtsi
and pxa3xx.dtsi as well. However, there is already nodes with clocks
properties in pxa2xx.dtsi.

I cannot choose, you should make a decision.
Robert Jarzmik Feb. 23, 2016, 8:12 p.m. UTC | #3
Sergei Ianovich <ynvich@gmail.com> writes:

> Hi Robert, 
>
> On Mon, 2016-02-22 at 16:35 +0100, Robert Jarzmik wrote:
>> You're right, I haven't seen the pxa2xx.dtsi statement before.
>> 
>> As such, could you amend a bit your patch please to :
>>  - add:
>> Fixes: 0ec1939668e5 ("ARM: dts: pxa: add the usb host controller")
>>  - make the same change in pxa3xx.dtsi, as the same error is there
>>  - and more globally align pxa2xx.dtsi, pxa27x.dtsi and pxa3xx.dtsi
>>  - remove the compatible, reg, interrupts and status from both pxa27x.dtsi and
>>    pxa3xx.dtsi as they are redundant with the included pxa2xxx.dsti one
>
> No problem. Let's figure out the best approach.
>
>> I'm even wondering if the proper change wouldn't be to scrap the ohci
>> declarations from pxa27x.dtsi and pxa3xx.dtsi, and only add the "clocks"
>> property to pxa2xx.dtsi ...
>
> Since clocks are declared in pxa27x.dtsi and pxa3xx.dtsi, there is a
> stronger ground to have node clocks properties set in pxa27x.dtsi
> and pxa3xx.dtsi as well. However, there is already nodes with clocks
> properties in pxa2xx.dtsi.
>
> I cannot choose, you should make a decision.
I have checked my manuals, and pxa25x doesn't have an USB Host controller.

So I'd rather have :
 - remove the usb0: ohci@4c000000 node from pxa2xx.dtsi
 - add it completely to pxa27x.dtsi and pxa3xx.dtsi, with clocks and everything,
   in "disabled" status.

Cheers.
diff mbox

Patch

diff --git a/arch/arm/boot/dts/pxa27x.dtsi b/arch/arm/boot/dts/pxa27x.dtsi
index 7f68a1e..3016a73 100644
--- a/arch/arm/boot/dts/pxa27x.dtsi
+++ b/arch/arm/boot/dts/pxa27x.dtsi
@@ -26,7 +26,7 @@ 
 			clocks = <&clks CLK_NONE>;
 		};
 
-		pxa27x_ohci: usb@4c000000 {
+		usb0: ohci@4c000000 {
 			compatible = "marvell,pxa-ohci";
 			reg = <0x4c000000 0x10000>;
 			interrupts = <3>;