diff mbox

[v2] ARM: dts: i.MX35: Fix can support.

Message ID 1437640272-5054-1-git-send-email-denis@eukrea.com (mailing list archive)
State New, archived
Headers show

Commit Message

Denis Carikli July 23, 2015, 8:31 a.m. UTC
Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
("can: flexcan: add 2nd clock to support imx53 and newer")
the can driver requires a dt nodes to have a second clock.
Add them to imx35 to fix probing the flex can driver on the
respective platforms.

Signed-off-by: Denis Carikli <denis@eukrea.com>
---
 arch/arm/boot/dts/imx35.dtsi | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Comments

Stefan Wahren July 23, 2015, 8:58 a.m. UTC | #1
Hi Denis,

[ add Marc to CC ]

Am 23.07.2015 um 10:31 schrieb Denis Carikli:
> Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
> ("can: flexcan: add 2nd clock to support imx53 and newer")
> the can driver requires a dt nodes to have a second clock.
> Add them to imx35 to fix probing the flex can driver on the
> respective platforms.

could you describe the problem more in detail or do you have a log in
case the probing failed?

According to Marc's commit this fix shouldn't be necessary.

Regards
Stefan
Marc Kleine-Budde July 23, 2015, 9:29 a.m. UTC | #2
On 07/23/2015 10:58 AM, Stefan Wahren wrote:
> Hi Denis,
> 
> [ add Marc to CC ]
> 
> Am 23.07.2015 um 10:31 schrieb Denis Carikli:
>> Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
>> ("can: flexcan: add 2nd clock to support imx53 and newer")
>> the can driver requires a dt nodes to have a second clock.
>> Add them to imx35 to fix probing the flex can driver on the
>> respective platforms.
> 
> could you describe the problem more in detail or do you have a log in
> case the probing failed?
> 
> According to Marc's commit this fix shouldn't be necessary.

In the pre DT era and without a clock name, a single clock is
sufficient. But with clock name or DT with clock-name you need both a
"ipg" and a "per" clock. The imx35.dtsi was added with one clock only,
so it was probably never working on mx35 with DT.

Marc
Stefan Wahren July 23, 2015, 9:48 a.m. UTC | #3
Am 23.07.2015 um 11:29 schrieb Marc Kleine-Budde:
> On 07/23/2015 10:58 AM, Stefan Wahren wrote:
>> Hi Denis,
>>
>> [ add Marc to CC ]
>>
>> Am 23.07.2015 um 10:31 schrieb Denis Carikli:
>>> Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
>>> ("can: flexcan: add 2nd clock to support imx53 and newer")
>>> the can driver requires a dt nodes to have a second clock.
>>> Add them to imx35 to fix probing the flex can driver on the
>>> respective platforms.
>> could you describe the problem more in detail or do you have a log in
>> case the probing failed?
>>
>> According to Marc's commit this fix shouldn't be necessary.
> In the pre DT era and without a clock name, a single clock is
> sufficient. But with clock name or DT with clock-name you need both a
> "ipg" and a "per" clock. The imx35.dtsi was added with one clock only,
> so it was probably never working on mx35 with DT.

Okay the other plaforms like mx25 and mx28 already had 2 clocks defined.

But why doesn't say the binding doc nothing about those required clocks?

Stefan

>
> Marc
>
Marc Kleine-Budde July 23, 2015, 9:54 a.m. UTC | #4
On 07/23/2015 11:48 AM, Stefan Wahren wrote:
> Am 23.07.2015 um 11:29 schrieb Marc Kleine-Budde:
>> On 07/23/2015 10:58 AM, Stefan Wahren wrote:
>>> Hi Denis,
>>>
>>> [ add Marc to CC ]
>>>
>>> Am 23.07.2015 um 10:31 schrieb Denis Carikli:
>>>> Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
>>>> ("can: flexcan: add 2nd clock to support imx53 and newer")
>>>> the can driver requires a dt nodes to have a second clock.
>>>> Add them to imx35 to fix probing the flex can driver on the
>>>> respective platforms.
>>> could you describe the problem more in detail or do you have a log in
>>> case the probing failed?
>>>
>>> According to Marc's commit this fix shouldn't be necessary.
>> In the pre DT era and without a clock name, a single clock is
>> sufficient. But with clock name or DT with clock-name you need both a
>> "ipg" and a "per" clock. The imx35.dtsi was added with one clock only,
>> so it was probably never working on mx35 with DT.
> 
> Okay the other plaforms like mx25 and mx28 already had 2 clocks defined.
> 
> But why doesn't say the binding doc nothing about those required clocks?

Good question - Dokumentation bug...

Marc
Shawn Guo July 25, 2015, 3:38 a.m. UTC | #5
On Thu, Jul 23, 2015 at 10:31:12AM +0200, Denis Carikli wrote:
> Since commit 3d42a379b6fa5b46058e3302b1802b29f64865bb
> ("can: flexcan: add 2nd clock to support imx53 and newer")
> the can driver requires a dt nodes to have a second clock.
> Add them to imx35 to fix probing the flex can driver on the
> respective platforms.
> 
> Signed-off-by: Denis Carikli <denis@eukrea.com>

Applied, thanks.
diff mbox

Patch

diff --git a/arch/arm/boot/dts/imx35.dtsi b/arch/arm/boot/dts/imx35.dtsi
index b6478e9..e6540b5 100644
--- a/arch/arm/boot/dts/imx35.dtsi
+++ b/arch/arm/boot/dts/imx35.dtsi
@@ -286,8 +286,8 @@ 
 			can1: can@53fe4000 {
 				compatible = "fsl,imx35-flexcan", "fsl,p1010-flexcan";
 				reg = <0x53fe4000 0x1000>;
-				clocks = <&clks 33>;
-				clock-names = "ipg";
+				clocks = <&clks 33>, <&clks 33>;
+				clock-names = "ipg", "per";
 				interrupts = <43>;
 				status = "disabled";
 			};
@@ -295,8 +295,8 @@ 
 			can2: can@53fe8000 {
 				compatible = "fsl,imx35-flexcan", "fsl,p1010-flexcan";
 				reg = <0x53fe8000 0x1000>;
-				clocks = <&clks 34>;
-				clock-names = "ipg";
+				clocks = <&clks 34>, <&clks 34>;
+				clock-names = "ipg", "per";
 				interrupts = <44>;
 				status = "disabled";
 			};