[LinuxBIOS] [v3][PATCH] POST_CODE macro

Uwe Hermann uwe at hermann-uwe.de
Mon Jul 23 12:07:53 CEST 2007


On Sun, Jul 22, 2007 at 04:55:01PM +0200, Stefan Reinauer wrote:
> * ron minnich <rminnich at gmail.com> [070720 23:33]:
> > Sorry I have not been paying close attention.
> > 
> > Why is post code a library function? So you can extend it with serial
> > post output, and so on. I don't think this macro is the right thing to
> > do. We've had post_code do much more over the years than outb to port
> > 0x80.
> 
> Do we really need to extend this? I always got the impression that post
> codes are completely useless because that technology has been
> superseeded by serial console a couple of decades ago.
> 
> From that perspective I fail to see why we would print POST numbers if
> we even have the serial console.

I agree that post_code() should not be a macro, and I also agree that it
should not do things other than emit a POST code.

Stafan is right, as soon as you have a serial console there's absolutely
no need for POST codes, just use printk()'s. If we need other fancy
stuff, I'd say we make that an additional function, and not cram all of
it into post_code().

However, we should still use POST codes (i.e., not drop them completely)
as there are legacy-free/embedded/crippled systems out there which don't
have any (easy) means for debugging output other than POST codes
(no serial, no USB Debug port, no JTAG, etc). Also, you need POST codes
for early bringup where the Super I/O (serial port) is not yet working.


Uwe.
-- 
http://www.hermann-uwe.de  | http://www.holsham-traders.de
http://www.crazy-hacks.org | http://www.unmaintained-free-software.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://www.coreboot.org/pipermail/coreboot/attachments/20070723/67326f3b/attachment.sig>


More information about the coreboot mailing list