[PATCH] kill useless SMT code in prom_hold_cpus
ntl at pobox.com
Tue Jul 15 12:24:40 EST 2008
Benjamin Herrenschmidt wrote:
> On Tue, 2008-07-08 at 17:36 -0500, Nathan Lynch wrote:
> > I think this code that counts SMT threads and compares against NR_CPUS
> > is an artifact of pre-powerpc-merge ppc64. We care about starting
> > only primary threads in the OF client code.
> > Signed-off-by: Nathan Lynch <ntl at pobox.com>
> That looks good. I'm not merging it right now because I want to dbl
> check that it's allright on all SMT machines. IE. We compare reg
> against _prom->cpu now instead of interrupt_server and I thus
> want to ensure it's the same everywhere.
Thanks. Looks like prom_find_boot_cpu is setting _prom->cpu to reg
(or 0), so I think it should be fine... a system where reg differed
from interrupt_server would have been broken before this patch
anyway, I think?
More information about the Linuxppc-dev