On 07.07.2008 07:50, Stefan Reinauer wrote: > [EMAIL PROTECTED] wrote: > >> Author: hailfinger >> Date: 2008-07-07 01:04:01 +0200 (Mon, 07 Jul 2008) >> New Revision: 3416 >> >> Modified: >> trunk/util/flashrom/flash.h >> trunk/util/flashrom/flashchips.c >> Log: >> r3415 removed symbolic constants for device IDs by accident. >> flash.h is a database of known IDs, whereas flashchips.c is a database >> of chips for which support has been implemented. Keep it that way. >> >> Trivial. >> >> >> > NACK. This was done on purpose, not by accident. As discussed before, we > will not pursue the useless facility of defines for those IDs anymore > because they serve no purpose. >
Thanks for the hint to search the archives. I finally found the mail you are referring to: http://coreboot.org/pipermail/coreboot/2008-January/029760.html On Sun 2008-01-27 03:15:58, Stefan Reinauer wrote: > flash.h is the only place where the VENDOR_ID defines are > defined and the array in flashchips.c is the only place where they are > used. That reasoning may have been valid back then, but now there are multiple places outside flashchips.c where we use vendor and device IDs. Peter's latest patch to add prettyprinting for SST25VF040B illustrates this quite nicely. Please reconsider your reasoning based on the new situation. Besides that, flash.h often has additional info about a chip which is totally missing in flashchips.c (and the lines there are growing more and more and barely fit into my 227 char wide terminal), so someone would have to merge the info. Regards, Carl-Daniel -- http://www.hailfinger.org/ -- coreboot mailing list [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
