* Alex G. <[email protected]> [110225 23:26]: > Hi Stefan and Patrick. I Just saw your emails. > > inb(0x80) or post_code(POST_SMBUS_DELAY): > make up your minds :)
The second one seems wrong. It's not a post_code, but a delay that happens to print some garbage on a post card. Changing that into post_code() silently pretends that this is done on purpose. Not good, in my opinion. > > Can we put this in one file together with > > src/include/cpu/amd/geode_post_code.h > > Looks interesting. Looking into that. Using this will totally obsolete > documentation/POSTCODES, which I used as the basis. If you prefer to use > these codes, say "green". > > > src/include/cpu/x86/post_code.h > > No. This would ruin the behavior of post_code() in console.c, which also > outputs to console if the option is selected. Why? Just add #ifdef ASSEMBLY around it. Then it won't be visible in console.c (or console.h for that matter) > On 02/25/2011 11:59 PM, Stefan Reinauer wrote: > > > > refactored... should we move the delay function to a common place? > > > I can move it. Just which file ? Also, wouldn't this make a single patch > too hard to swallow? Yes. I think you should remove all changes to smbus* files from your patch and then we can look at the issue separately. -- coreboot mailing list: [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
