Re: [PATCH 1/4] documentation: add palmas dts definition

[Stephen Warren]

On 02/28/2013 02:58 AM, Graeme Gregory wrote:
> On 28/02/13 08:52, Laxman Dewangan wrote:
>> On Thursday 28 February 2013 12:02 AM, Stephen Warren wrote:
>>> On 02/17/2013 10:11 PM, J Keerthy wrote:
>>> > +- interrupt-parent : The parent interrupt controller.
>>> > +
>>> > +>Optional node:
>>> > +- Child nodes contain in the palmas. The palmas family is made of several
>>> > +  variants that support a different number of features.
>>> > +  The child nodes will thus depend of the capability of the variant.
...
>>> Are there DT bindings for those child nodes anywhere?
>>>
>>> Representing each internal component as a separate DT node feels a
>>> little like designing the DT bindings to model the Linux-internal MFD
>>> structure. DT bindings should be driven by the HW design and
>>> OS-agnostic.
>>>
>>> From a DT perspective, is there any need at all to create a separate DT
>>> node for each component? This would only be needed or useful if the
>>> child IP blocks (and hence DT bindings for those blocks) could be
>>> re-used in other top-level devices that aren't represented by this
>>> top-level ti,palmas DT binding. Are the HW IP blocks here re-used
>>> anywhere, or will they be?
...
> I wonder why break good software principles of keeping data and code
> localised? Just because there is no current case where a block is
> re-used does not mean it shall not be so in the future. The original
> information I got from TI when designing this was blocks may be re-used
> in future products.
> 
> This structure also makes it much neater when dealing with palmas
> varients with different IP blocks which already exist.

Given the current existence of chips that mix/match IP blocks and the
guidance you quote from TI about future chips doing so too, having these
child nodes in the DT makes sense to me.

>>> +    palmas_rtc {
>>> +        compatible = "ti,palmas_rtc";
>>> +        interrupts = <8 9>;
>>> Are the interrupt outputs of the RTC fed directly to the GIC interrupt
>>> mentioned in the top-level Palmas node, or do these interrupts feed into
>>> a top-level IRQ controller in the Palmas device, which then feeds into
>>> the external IRQ controller?
>>
>> The interrupt goes to the chip-internal irq, not to external of chip. 
>> We have only one int line from chip which can be connected to
>> processor/GIC.
>> yes, interrupt parent need to be define here to get the proper
>> interrupt number.
>
> Those interrupt lines are un-needed in the newer version of driver, I
> forgot to remove them. The regmap-irq takes care of this for us without
> needing this information in the DT at all.
> 
> But actually the OF handles this without requiring a parent in this
> case. These interrupts are fed to the child nodes inside io_resource
> entries.

That doesn't make sense.

The driver for the child node should parse the DT to extract the IRQ
number/handle, and IRQ must be fully represented in DT in the standard
fashion. Put another way, the driver for the child node should not
expect the top-level driver to have created platform-device style
resources for it when instantiated from DT.

If you don't do this, there will be no possibility for the driver for
the top-level Palmas node to completely generically support arbitrary IP
blocks being mixed/matched into it, and hence there will be no point at
all representing the child IP blocks as separate DT nodes.

Even ignoring any of that, the DT content still needs to correctly
represent the HW, using the existing DT standards for interrupts, and it
doesn't as written.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/