[RFC] [PATCH] Device Tree on ARM platform

Grant Likely grant.likely at secretlab.ca
Thu May 28 23:34:02 EST 2009


On Thu, May 28, 2009 at 12:34 AM, Wolfram Sang <w.sang at pengutronix.de> wrote:
>
>> It is inherent in the value "atmel,24c08".  Because the exact device
>> model is specified, the kernel can know what the page size and flags
>> are for that device.
>
> I am afraid this is not the case. I remember that we once had two revisions of
> the same eeprom, and the vendor decided to increase the page size without
> changing the naming of the eeprom. (Can't point to any documentation yet, would
> have to dig this up.) Those eeproms can be really weird.
>
> Also, there are _a lot_ of eeprom manufacturers out there, so the match table
> will be enormous....

... which is the whole point of "compatible".  Every possible variant
doesn't get listed in the device driver; device nodes in the tree
claim compatibility with older, already known parts.

g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.



More information about the devicetree-discuss mailing list