Subject: re: new sysctl: hw.cpu_isa
To: Darren Reed <email@example.com>
From: matthew green <firstname.lastname@example.org>
Date: 11/14/2000 11:09:34
What is this sysctl reporting ? The kernel architecture or the system's ?
A 32bit kernel on a 64bit system is sun4u but is NOT sparcv9.
In either case, I'm starting to think that reporting "sun4*" is incorrect,
also, we have the potential to be compatible with isalist/optisa on Solaris
which used sparc* rather than sun4* to report and represent ISA.
using sun4* would be as far wrong as possible. the dumped solaris bits are
pretty much the right names to use:
32bit: sparc, sparcv7 (default?), sparcv8, sparcv8plus, sparcv8plusa, sparcv8a
64bit: sparcv9, sparcv9a, sparcv9plusa (default?)
with the v8a/v8aplus only being available on ultrasparc-class machines. (hmm,
from the solaris names, the call the "a" bit "+vis").
the above also mostly agree with the gcc -m names.