Sent from my mobile phone
On 03.11.2010, at 09:56, Ward Vandewege ward@gnu.org wrote:
On Tue, Nov 02, 2010 at 11:20:43PM +0100, Uwe Hermann wrote:
I don't object to the patch, and we should probably fix this in all other southbridges, I think the same problem applies there.
But: the die() call itself also does a printk(), so that'll still hang if the error path is chosen (at that point it probably doesn't matter much, though, as we die anyway).
Right, I think it does not matter. If the die happens when printk is already functional, great, if not it will hang there which is fine.
I also agree that die() should have a POST code, preferrably something easy to remember. It already has a commented-out "//post_code(0xff);". Not sure why it's disabled, but I think it should be something other than 0xff, that's a bit too "special" for my taste.
We have "0xee: Not supposed to get here" as per documentation/POSTCODES, so maybe we can use 0xdd ("d" as in die), if that's not already used elsewhere.
So, thinking about this a little more, I'm not sure adding a post code to 'die' is a good idea. The problem with doing that is that it would clobber any previous post codes, which might be a better indicator for what's going wrong.
Perhaps a good way to deal with fatal runtime error conditions would be
a) set a unique post code b) call die
in the assumption that die does not clobber the post code.
We could add a post code to the parameters of the die() function.
What do you think?
Thanks, Ward.
-- Ward Vandewege ward@fsf.org Free Software Foundation - Senior Systems Administrator
-- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot