Re: [PATCH 5/5] pwm: pwm-tiehrpwm: Update dt binding document to use generic node name

[Rob Herring <robh@kernel.org>]

On Thu, Mar 17, 2016 at 1:20 PM, Franklin S Cooper Jr. <fcooper@ti.com> wrote:
>
>
> On 03/17/2016 01:00 PM, Rob Herring wrote:
>> On Thu, Mar 17, 2016 at 11:49 AM, Franklin S Cooper Jr. <fcooper@ti.com> wrote:
>>>
>>> On 03/17/2016 10:03 AM, Rob Herring wrote:
>>>> On Mon, Mar 07, 2016 at 01:51:58PM -0600, Franklin S Cooper Jr wrote:
>>>>> Now that the node name has been changed from ehrpwm to pwm the document
>>>>> should show this proper usage. Also change the unit address in the example
>>>>> from 0 to the proper physical address value that should be used.
>>>>>
>>>>> Signed-off-by: Franklin S Cooper Jr <fcooper@ti.com>
>>>>> ---
>>>>>  Documentation/devicetree/bindings/pwm/pwm-tiehrpwm.txt | 4 ++--
>>>>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/pwm/pwm-tiehrpwm.txt b/Documentation/devicetree/bindings/pwm/pwm-tiehrpwm.txt
>>>>> index 9c100b2..20211ed 100644
>>>>> --- a/Documentation/devicetree/bindings/pwm/pwm-tiehrpwm.txt
>>>>> +++ b/Documentation/devicetree/bindings/pwm/pwm-tiehrpwm.txt
>>>>> @@ -15,14 +15,14 @@ Optional properties:
>>>>>
>>>>>  Example:
>>>>>
>>>>> -ehrpwm0: ehrpwm@0 { /* EHRPWM on am33xx */
>>>>> +ehrpwm0: pwm@48300200 { /* EHRPWM on am33xx */
>>>>>      compatible = "ti,am33xx-ehrpwm";
>>>>>      #pwm-cells = <3>;
>>>>>      reg = <0x48300200 0x100>;
>>>>>      ti,hwmods = "ehrpwm0";
>>>>>  };
>>>>>
>>>>> -ehrpwm0: ehrpwm@0 { /* EHRPWM on da850 */
>>>>> +ehrpwm0: pwm@01f00000 { /* EHRPWM on da850 */
>>>> No leading 0s, but more importantly the address is wrong.
>>> I will remove the leading 0. However, this value was taken
>>> from the .dtsi and I just double checked and I see the same
>>> value in the datasheet. I believe DA850,OMAP-L138 and AM18x
>>> all have the same memory mapping. I'm looking at
>>> http://www.ti.com/lit/ds/symlink/am1808.pdf page 233 and the
>>> addresses match up what is seen here and in the .dtsi.
>>>
>>> Can you point me to which document your looking at that
>>> shows a different value?
>> Ummm, ...
>>
>>>>>      compatible = "ti,da850-ehrpwm", "ti,am33xx-ehrpwm";
>>>>>      #pwm-cells = <3>;
>>>>>      reg = <0x300000 0x2000>;
>> right here.
>
> So I don't know the history but the SOC node specifies a
> ranges value of ranges = <0x0 0x01c00000 0x400000>;. It
> seems that all child nodes of SOC have a reg property then
> is based on an offset of 0x01c00000. So this is true for
> UART, rtc, i2c, wdt, mmc, spi etc... So using a base offset
> of 0x01c00000 + 0x300000 (reg value of the pwm) equals the
> physical address of the ehrpwm0 register 0x1f00000.
>
> For the child nodes within the SOC node, the unit-address is
> always based on the physical address not based on the offset
> address.

They are all wrong and should be fixed. Unit address should match the
reg property unless the bus has defined something different (e.g.
PCI).

Rob