Jay Estabrook schrieb:
On Sat, Mar 17, 2007 at 02:26:18PM +0300, Sergey Tikhonov wrote:I remember tracing that a long time ago). The -p and -i is handled by "uname" and have some altering code for other arches. I'll install modified version of uname and see how other programs would like the new output. :)Let us know!So far, no problems for my usual build procedures.This (building) is BY FAR my biggest concern about changing ANY of the entries returned by "uname -mpi". There's so much perl and python and sh and whatnot code out there that expects certain things for "uname -m", abd possibly both of the others, that I'm very fearful of the impact of changes... :-( I guess it's prolly gonna be "try it and see" time... ;-}
Hm. Jay, that's true, many configures will expect *things*. :-) But building for ev67, ev6, ev68 or whatever is out there should be only done for special packages. I'm thinking of: kernel (that will be the hardest part, as we will have to provide different .config's; But this part make much sense and I believe it will change how uname -m comes up!), glibc (should be no problem to turn on special flags with --target; Maybe will also influence uname!?), openssl (same as glibc). Maybe some other package as well, but these are in my head and I know from x86, that they are build for different sub-arches as well. What I want to say with that: I don't want to build *everything* for sub-arches. That doesn't make sense I believe...
However, what uname -m should return, I will ask tomorrow on lkml... Maybe someone has a *good* hint. :-) I believe it's some .config kernel voodoo. :-)
-of