"Nori, Sekhar" <[email protected]> writes: > From: Troy Kisky [mailto:[email protected]] > Sent: Monday, April 06, 2009 1:05 AM >> >> Nori, Sekhar wrote: >> > From: Troy Kisky >> > Sent: Saturday, April 04, 2009 7:37 AM >> >> If the MACSRCADDR register has been initialized by bootloader, >> >> then use that instead of a random mac. >> > >> > Where was this scenario useful? If intent is to use the initialization >> > done by U- >> Boot, using eth= bootarg would be a better approach like the way it is being >> done >> for SFFSDR board. >> > >> > Thanks, >> > Sekhar >> > >> The same argument (and more) applies to using a random address. To be >> consistent, >> you should be arguing against that as well. > > I think random address is a fallback when EEPROM gives invalid contents. We > shouldn't need two fallbacks for the same issue. > > Also, if EEPROM has invalid contents, trying to use U-Boot initialization of > MAC address won't serve well. >
I agree with Sekhar. In addition, this dependency on platform-specifics does not belong in the driver. Rather, any dependencies on platform-specific settings or bootloader-passed settings should be handled in platform-specific code. Kevin _______________________________________________ Davinci-linux-open-source mailing list [email protected] http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
