Hello Geert,

On Mon, May 22, 2017 at 9:30 PM, Geert Uytterhoeven
[off-list ref] wrote:

Hi Javier,

On Mon, May 22, 2017 at 7:15 PM, Javier Martinez Canillas
[off-list ref] wrote:

quoted

quoted

quoted

I also wonder why this is really needed if AFAIU "renesas,24c02" is
compatible with "atmel,24c02". IOW, the driver doesn't need to
differentiate between the two since the devices are the same and will
always match using "atmel,24c02".

It is needed, so that when a difference is found, it can be handled
without updating the DT.

Yes, I understand this. What I tried to ask is if there could really
be a difference for the same chip type between different vendors, or
is just that people were using other manufacturers in the compatible
string as a consequence of the DT binding doc and the I2C core
ignoring the vendor prefix.

The devices from different vendors are not the same. They contain FLASH
ROM of a specific size, and glue logic to expose an i2c slave
interface providing
an AT24-compatible command set.  They should behave similar within
the limits of the AT24 "spec".  But the actual implementation may be different.

I see, really appreciate your explanation. I'm not familiar with these
devices and driver but the patch-series are needed in order to make
sure that no regressions will happen once the I2C core reports a
proper OF modalias.

quoted

I don't mind though, I will leave the manufacturers that are different
than the atmel variants in the mainline DTS as you and Geert asked.

OK, thanks!

Thanks a lot for your feedback!

Gr{oetje,eeting}s,

                        Geert

--

Best regards,
Javier