[LinuxBIOS] [PATCH] Add initial serial output for Via vt82c686
corey_osgood at verizon.net
Sun Mar 18 04:04:18 CET 2007
Uwe Hermann wrote:
> Committed in r2571, thanks!
> On Sat, Mar 17, 2007 at 12:01:28AM -0400, Corey Osgood wrote:
>>> If it works (and if you don't spot further problems in the code) I'd say
>>> can be committed. Please send an Acked-by if the code looks good to you.
>> Awesome! I was starting to get concerned that I'd done something that
>> only worked on my board. I've got no problem with your changes (although
>> I can't _find_ an io base), and the code works fine on my board.
> Yep, iobase is not used yet. I guess we should make these two lines
> dependant on the value of iobase?
> + vt82c686_sio_write(VT82C686_SP1, 0xfe); /* SP1: 0x3f8 */
> + vt82c686_sio_write(VT82C686_SP2, 0xbe); /* SP2: 0x2f8 */
Oops, my bad, that wasn't what I meant. But yeah, that will get
>> die() in console/console.c does send a post code, but if I try to use
>> it, gcc gives an error about function type pointers that I haven't been
>> able to track down the cause of. The die() src/arch/i386/lib/console.c
>> doesn't send a post code, and that's the one I'm currently using, since
>> that's what whichever motherboard I based my code on was using. So,
>> either that console.c should get fixed up, or I should figure out why I
>> can't use the other one (or even both).
> Hm, strange. I'll have a look later. Shall we hardcode a post_code()
> call in addition of die() for now?
Can't do that either, for the exact same reason, post_code() isn't in
that particular console.c, it needs to be outb(). Why do we have two
different console.c's, anyways? Is it a romcc or cashe-as-ram thing?
More information about the coreboot