Re: [E1000-devel] Supporting platforms with no EEPROM?
Lennart Sorensen wrote:
> On Fri, Sep 12, 2008 at 08:54:29PM +0200, Arnaud Patard wrote:
>> I don't know which iop32x configurations are enabled in the default
>> iop32x kernel but there are some boards with e1000 which do have the
>> eeprom so imho unless you do some kind of runtime detection, I fear
>> you'll break theses boards.
>
> Well looking at the datasheet of one of the e1000 chips, it looks like
> if you check the contents of EEPROM_INIT_CONTROL1_REG in the eeprom
> data, then a check of (value & 0xc000 == 0x4000) will tell you that the
> eeprom is valid, and otherwise you have no eeprom or an invalid eeprom.
> So if you detect that no valid eeprom is present, check if the existing
> MAC address is a valid one, and if so, leave the chip alone and assume
> it is a board with no eeprom but some other method of configuring it.
>
Right! So now, all we gotta do is get Intel to sign on to distributing such
code in their mainline driver... ;)
b.g.
--
Bill Gatliff
bgat@billgatliff.com
Reply to: