[OpenAFS-devel] DRAFT: New sysname standard

[Derek Atkins]

Indeed, this is why the original Linux-AFS ports used i386_linux[123]
because there were 3 ABI versions.  The kernel modules were
independent of the AFS sysname.  OTOH, if you want to support kernel
modules in AFS, you do need to have some way to differentiate the
kernel verson.

Honestly, this is really only a problem on Linux.  I can't think of
any other system that has this same problem where the kernel/OS/ABI
interfaces aren't tied to each other in some way.  Linux is the only
system where you can have multiple primary ABIs with the same "kernel
version".

Do any other (non-Linux) platforms separate out the OS version from
the ABI version?

-derek

Chas Williams <chas@cmf.nrl.navy.mil> writes:

> In message <15077.59397.53018.505814@zappa.ms.com>,Phil.Moore@morganstanley.com
>  writes:
> >	<p>ia64_linux_24_22</p>
> 
> just a comment about the linux sysnames.  it seems redundant to specify the
> kernel version in there since that usually its the important piece.  if i
> had to do it over again, i would have renamed i386_linux22 to just i386_linux
> using the '#if KERNEL_VERSION >' to handle any differences.  this wouldnt
> have taken into account glibc differences though.
> 
> if you have ia64_linux_24_22 and ia64_linux_22_22 you are going to have
> two sets of afs binaries that are essentially the same.  perhaps just
> having i386_linux_22 (were 22 is the abi) would be 'better'?  just
> distribute/build the appropriate different kernel modules.
> _______________________________________________
> OpenAFS-devel mailing list
> OpenAFS-devel@openafs.org
> https://lists.openafs.org/mailman/listinfo.cgi/openafs-devel

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord@MIT.EDU                        PGP key available