I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1746 bytes --]

On Fri, Dec 18, 2015 at 12:02:39PM +0900, Simon Horman wrote:
> On Mon, Dec 14, 2015 at 10:39:47AM +0100, Geert Uytterhoeven wrote:
> > On Mon, Dec 14, 2015 at 8:11 AM, Wolfram Sang <wsa@the-dreams.de> wrote:
> > >> > >+  eeprom@50 {
> > >> > >+          compatible = "renesas,24c02";
> > >> >
> > >> >    This is not a valid value -- the Renesas chip model is different from 24c02.
> > >>
> > >> I copied this value from r8a7791.dtsi.
> > >>
> > >> Looking at the schematic for gose (v100), koelsch (rev024) and porter (v300)
> > >> I see the following "R1EX24002ATAS0G#U0". Shall we update r8a7791 and
> > >> this patch to "renesas,24002" or leave things as is?
> > >
> > > I wouldn't like to update the at24 driver with all namings from all
> > > vendors for chips which in large cases are simple 24c02 devices.
> > >
> > > So, if Sergei insists on the change, I'd propose
> > >
> > >         compatible = "renesas,24002", "24c02";
> > 
> > "renesas,r1ex24002"?
> > 
> > I don't think the "A" is relevant (<= 64 is A, >= 128 is B).
> > 
> > http://www.renesas.com/products/memory/eeprom/product_selector.jsp
> 
> Thanks.
> 
> It seems to me that we have some consensus around:
> 
> 	compatible = "renesas,r1ex24002", "24c02";

Thinking again, "generic,24c02" or "generic-24c02" could also be an
option.

> Should this be added to Documentation/devicetree/bindings/eeprom/eeprom.txt ?
> Or documented elsewhere?

Probably we need a DT maintainers advice here? I don't mind vendor
specific compatibles being documented, but I'm reluctant to add all
these compatibles for the myriads of I2C eeproms to the at24 driver. 99%
are covered by the generic case.

Adding DT to CC.

Thanks,

   Wolfram


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

[Arnd Bergmann]

On Friday 18 December 2015 08:35:32 Wolfram Sang wrote:
> > 
> > It seems to me that we have some consensus around:
> > 
> >       compatible = "renesas,r1ex24002", "24c02";
> 
> Thinking again, "generic,24c02" or "generic-24c02" could also be an
> option.
> 
> > Should this be added to Documentation/devicetree/bindings/eeprom/eeprom.txt ?
> > Or documented elsewhere?
> 
> Probably we need a DT maintainers advice here? I don't mind vendor
> specific compatibles being documented, but I'm reluctant to add all
> these compatibles for the myriads of I2C eeproms to the at24 driver. 99%
> are covered by the generic case.
> 
> Adding DT to CC.

I'd rather use some vendor string in addition to 24c02. Isn't this originally
an Atmel part? In that case, using "atmel,24c02" as the most generic string
would be appropriate, and IIRC the i2c framework will just match that with
the "24c02" entry in the i2c_device_id list.

	Arnd

Re: I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1434 bytes --]

On Fri, Dec 18, 2015 at 10:06:48AM +0100, Arnd Bergmann wrote:
> On Friday 18 December 2015 08:35:32 Wolfram Sang wrote:
> > > 
> > > It seems to me that we have some consensus around:
> > > 
> > >       compatible = "renesas,r1ex24002", "24c02";
> > 
> > Thinking again, "generic,24c02" or "generic-24c02" could also be an
> > option.
> > 
> > > Should this be added to Documentation/devicetree/bindings/eeprom/eeprom.txt ?
> > > Or documented elsewhere?
> > 
> > Probably we need a DT maintainers advice here? I don't mind vendor
> > specific compatibles being documented, but I'm reluctant to add all
> > these compatibles for the myriads of I2C eeproms to the at24 driver. 99%
> > are covered by the generic case.
> > 
> > Adding DT to CC.
> 
> I'd rather use some vendor string in addition to 24c02. Isn't this originally
> an Atmel part? In that case, using "atmel,24c02" as the most generic string
> would be appropriate,

Yeah, the at24 driver is named after Atmel chips AFAIR. Having "atmel,*"
as the generic fallback sounds like a good solution to me, too.

> and IIRC the i2c framework will just match that with
> the "24c02" entry in the i2c_device_id list.

True, although this behaviour is often complained about. There have been
attempts to make i2c/spi behave like the rest of the DT world and to
deprecate the current way. It didn't happen because of lots gory details,
however :/


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

[Simon Horman]

On Fri, Dec 18, 2015 at 11:32:15AM +0100, Wolfram Sang wrote:
> On Fri, Dec 18, 2015 at 10:06:48AM +0100, Arnd Bergmann wrote:
> > On Friday 18 December 2015 08:35:32 Wolfram Sang wrote:
> > > > 
> > > > It seems to me that we have some consensus around:
> > > > 
> > > >       compatible = "renesas,r1ex24002", "24c02";
> > > 
> > > Thinking again, "generic,24c02" or "generic-24c02" could also be an
> > > option.
> > > 
> > > > Should this be added to Documentation/devicetree/bindings/eeprom/eeprom.txt ?
> > > > Or documented elsewhere?
> > > 
> > > Probably we need a DT maintainers advice here? I don't mind vendor
> > > specific compatibles being documented, but I'm reluctant to add all
> > > these compatibles for the myriads of I2C eeproms to the at24 driver. 99%
> > > are covered by the generic case.
> > > 
> > > Adding DT to CC.
> > 
> > I'd rather use some vendor string in addition to 24c02. Isn't this originally
> > an Atmel part? In that case, using "atmel,24c02" as the most generic string
> > would be appropriate,
> 
> Yeah, the at24 driver is named after Atmel chips AFAIR. Having "atmel,*"
> as the generic fallback sounds like a good solution to me, too.
> 
> > and IIRC the i2c framework will just match that with
> > the "24c02" entry in the i2c_device_id list.
> 
> True, although this behaviour is often complained about. There have been
> attempts to make i2c/spi behave like the rest of the DT world and to
> deprecate the current way. It didn't happen because of lots gory details,
> however :/

My reading of the above is that we are currently leaning towards:

	compatible = "renesas,r1ex24002", "atmel,24c02";

And we are unsure what to do about the documentation.

Re: I2C eeprom compatibles? (was Re: [PATCH/RFC 03/19] ARM: shmobile: gose: add i2c2 bus to device tree)

[Geert Uytterhoeven]

On Mon, Dec 21, 2015 at 5:33 AM, Simon Horman <horms@verge.net.au> wrote:
> On Fri, Dec 18, 2015 at 11:32:15AM +0100, Wolfram Sang wrote:
>> On Fri, Dec 18, 2015 at 10:06:48AM +0100, Arnd Bergmann wrote:
>> > On Friday 18 December 2015 08:35:32 Wolfram Sang wrote:
>> > > >
>> > > > It seems to me that we have some consensus around:
>> > > >
>> > > >       compatible = "renesas,r1ex24002", "24c02";
>> > >
>> > > Thinking again, "generic,24c02" or "generic-24c02" could also be an
>> > > option.
>> > >
>> > > > Should this be added to Documentation/devicetree/bindings/eeprom/eeprom.txt ?
>> > > > Or documented elsewhere?
>> > >
>> > > Probably we need a DT maintainers advice here? I don't mind vendor
>> > > specific compatibles being documented, but I'm reluctant to add all
>> > > these compatibles for the myriads of I2C eeproms to the at24 driver. 99%
>> > > are covered by the generic case.
>> > >
>> > > Adding DT to CC.
>> >
>> > I'd rather use some vendor string in addition to 24c02. Isn't this originally
>> > an Atmel part? In that case, using "atmel,24c02" as the most generic string
>> > would be appropriate,
>>
>> Yeah, the at24 driver is named after Atmel chips AFAIR. Having "atmel,*"
>> as the generic fallback sounds like a good solution to me, too.
>>
>> > and IIRC the i2c framework will just match that with
>> > the "24c02" entry in the i2c_device_id list.
>>
>> True, although this behaviour is often complained about. There have been
>> attempts to make i2c/spi behave like the rest of the DT world and to
>> deprecate the current way. It didn't happen because of lots gory details,
>> however :/
>
> My reading of the above is that we are currently leaning towards:
>
>         compatible = "renesas,r1ex24002", "atmel,24c02";
>
> And we are unsure what to do about the documentation.

To avoid checkpatch.pl complaining about our DTS patches, "renesas,r1ex24002"
should be added to Documentation/devicetree/bindings/eeprom/eeprom.txt.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds