Re: [PATCH v3] New-style I2C and SMBus EEPROM driver (with device_ids)

[Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org>]

On Tue, 10 Jun 2008 15:43:47 +0200, Wolfram Sang wrote:
> > The Atmel 24C01 datasheet says page size is 4 bytes, and the Microchip
> > 24C01A datasheet says 2 bytes. So defaulting to 8 doesn't look safe.
> 
> Will go back to 2 because of Microchip. But 24C01 seems to have lots of
> variants, which makes a generic entry difficult. Some would need

Feel free to not make a generic entry at all if you think it's not
worth it.

> AT24_FLAG_24C00 (doesn't really matter), and AT24C01 needs 128
> addresses?? (please, someone, prove me wrong)

Why do you think so? My personal guess is that they simply forgot to
mention the address in the datasheet. A chip responding to all
addresses would prevent any other chip from being connected to the bus,
that's impractical enough to be reasonably certain that no manufacturer
did this.

> > In the end, the only things that must go in at24.h are the definition
> > of struct at24_platform_data and its flags. All the rest is internal to
> > the driver and should go in at24.c.
> 
> I wanted to have the AT24_SIZE_* flags next to the struct, so I won't
> forget to change their size if anything inside the struct will change.
> Then again, I might work with sizeof here; the result will probably look
> a bit nasty, too...

I see, it makes some sense to keep these flags around then. But then
please add a warning that these are for the driver internal use only
and shouldn't be considered stable.

-- 
Jean Delvare

_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c