]> www.pilppa.org Git - linux-2.6-omap-h63xx.git/commit
[POWERPC] Push down or eliminate smp_processor_id calls in xics code
authorMilton Miller <miltonm@bga.com>
Fri, 14 Dec 2007 04:52:09 +0000 (15:52 +1100)
committerPaul Mackerras <paulus@samba.org>
Thu, 20 Dec 2007 05:15:34 +0000 (16:15 +1100)
commitd7cf0edb8f7d53bfc8b5e40893415d48656c0399
treeb868deafc979c1e93b0aa17ec22c0be268e66e7d
parent7e1961ff4954d6ab627d9b606785afdcd0dda84b
[POWERPC] Push down or eliminate smp_processor_id calls in xics code

The per-processor interrupt request register and current processor
priority register are only accessed on the current cpu.  In fact the
hypervisor doesn't even let us choose which cpu's registers to access.

The only function to use cpu twice is xics_migrate_irqs_away, not a fast
path.  But we can cache the result of get_hard_processor_id() instead of
calling get_hard_smp_processor_id(cpu) in a loop across the call to rtas.

Years ago the irq code passed smp_processor_id into get_irq, I thought
we might initialize the CPPR third party at boot as an extra measure of
saftey, and it made the code symmetric with the qirr (queued interrupt
for software generated interrupts), but now it is just extra and
sometimes unneeded work to pass it down.

Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
arch/powerpc/platforms/pseries/xics.c
arch/powerpc/platforms/pseries/xics.h