On Sep 11, 2005, at 8:38 AM, Greg Lindahl wrote:
On Sat, Sep 10, 2005 at 04:45:48PM -0500, Brian Barrett wrote:
I think that this is a fairly easy fix - Irix identifies any MIPS
chip as a mips-* from config.guess, but Linux apparently makes a
distinction between mips and mips64.
That's because there is a difference between mips and mpis64. The
assembly syntax looks similar, but there are subtle pitfalls involved
in using the same code for both. You can look at the Linux kernel code
for atomics in mips and mips64 to see if there are any differences.
-- greg, who used to play around with MIPS64 in the good old days
Darn. I was really hoping that it was like OS X always reporting
architecture as ppc, while Linux will give ppc64 on a G5. Ah well,
can't always get lucky :).
Jonathan - you might want to try the small change I sent anyway - it
might work, and the unit tests actually do a relatively good job at
catching when things aren't right.
If they don't work, we'll need to write atomic operations for the
mips64 platform. Since the Linux kernel has mips64 support, that's
not an impossible task (one can usually figure out calling
conventions and the like from there), but it would be much easier
with a mips64 assembly reference manual and ABI documentation. You
wouldn't happen to know where to find such things, would you? I'll
admit that my only interaction with the MIPS architectures are the
R4K and above in SGI systems, so I'm a bit out of my league.
Brian
--
Brian Barrett
Open MPI developer
http://www.open-mpi.org/
