i2c-mux-pca954x ACPI case

i2c-mux-pca954x ACPI case

[Andy Shevchenko]

The commit bbf9d262a147aeaeee0bf4e1c121166d69e556d4 ("i2c: mux: pca954x:
Add ACPI support for pca954x") adds a so called "ACPI support" for the
driver and thus I have few questions (besides obvious typo in it):

0. Had it ever been tested?
1. Is there *real* DSDT / registered ACPI IDs for a such device(s)?
2. If "yes" on 1, can you provide Documentation with *real* DSDT
excerpt?

If the answer is "no" on 2, I'm about to revert this, because
ACPI is *not* like Device Tree chaotic mess. Any ID, property and
related stuff *must* be officially registered and carefully chosen.

To maintainers of the drivers, including but not limited to I2C
subsystem, please, Cc ACPI guys (Rafael, Mika, me, etc) *before*
applying any ACPI IDs if there no clear and *real* DSDT excerpt.

It's disregard if this case (pca954x) valid or not.

Thank you for understanding.

-- 
Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Intel Finland Oy

Re: i2c-mux-pca954x ACPI case

[Peter Rosin]

On 2017-03-20 17:04, Andy Shevchenko wrote:
> The commit bbf9d262a147aeaeee0bf4e1c121166d69e556d4 ("i2c: mux: pca954x:
> Add ACPI support for pca954x") adds a so called "ACPI support" for the
> driver and thus I have few questions (besides obvious typo in it):
> 
> 0. Had it ever been tested?
> 1. Is there *real* DSDT / registered ACPI IDs for a such device(s)?
> 2. If "yes" on 1, can you provide Documentation with *real* DSDT
> excerpt?
> 
> If the answer is "no" on 2, I'm about to revert this, because
> ACPI is *not* like Device Tree chaotic mess. Any ID, property and
> related stuff *must* be officially registered and carefully chosen.
> 
> To maintainers of the drivers, including but not limited to I2C
> subsystem, please, Cc ACPI guys (Rafael, Mika, me, etc) *before*
> applying any ACPI IDs if there no clear and *real* DSDT excerpt.
> 
> It's disregard if this case (pca954x) valid or not.
> 
> Thank you for understanding.

Hi Andy,

I have no idea if PCA9540 etc are valid DSDT entries. Going forward,
I will Cc: ACPI people before taking ACPI patches. Sorry for the
inconvenience and sorry for my ignorance.

But before reverting, let's give Tin a chance to speak up...

Cheers,
peda

Re: i2c-mux-pca954x ACPI case

[Andy Shevchenko]

On Mon, 2017-03-20 at 17:38 +0100, Peter Rosin wrote:
> On 2017-03-20 17:04, Andy Shevchenko wrote:
> > The commit bbf9d262a147aeaeee0bf4e1c121166d69e556d4 ("i2c: mux:
> > pca954x:
> > Add ACPI support for pca954x") adds a so called "ACPI support" for
> > the
> > driver and thus I have few questions (besides obvious typo in it):
> > 
> > 0. Had it ever been tested?
> > 1. Is there *real* DSDT / registered ACPI IDs for a such device(s)?
> > 2. If "yes" on 1, can you provide Documentation with *real* DSDT
> > excerpt?
> > 
> > If the answer is "no" on 2, I'm about to revert this, because
> > ACPI is *not* like Device Tree chaotic mess. Any ID, property and
> > related stuff *must* be officially registered and carefully chosen.
> > 
> > To maintainers of the drivers, including but not limited to I2C
> > subsystem, please, Cc ACPI guys (Rafael, Mika, me, etc) *before*
> > applying any ACPI IDs if there no clear and *real* DSDT excerpt.
> > 
> > It's disregard if this case (pca954x) valid or not.
> > 
> > Thank you for understanding.
> 
> Hi Andy,
> 
> I have no idea if PCA9540 etc are valid DSDT entries. Going forward,
> I will Cc: ACPI people before taking ACPI patches. Sorry for the
> inconvenience and sorry for my ignorance.

Thank you, Peter.

> But before reverting, let's give Tin a chance to speak up...

Of course! That's why I wrote this message beforehand.

-- 
Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Intel Finland Oy