KVM: PPC: Book3S HV: Speed up wakeups of CPUs on HV KVM
authorBenjamin Herrenschmidt <benh@kernel.crashing.org>
Wed, 17 Apr 2013 20:30:50 +0000 (20:30 +0000)
committerAlexander Graf <agraf@suse.de>
Fri, 26 Apr 2013 18:27:31 +0000 (20:27 +0200)
commit54695c3088a74e25474db8eb6b490b45d1aeb0ca
tree4ecaa9c41857cf4b380f54017830e2aeb8dc56fe
parentbc5ad3f3701116e7db57268e6f89010ec714697e
KVM: PPC: Book3S HV: Speed up wakeups of CPUs on HV KVM

Currently, we wake up a CPU by sending a host IPI with
smp_send_reschedule() to thread 0 of that core, which will take all
threads out of the guest, and cause them to re-evaluate their
interrupt status on the way back in.

This adds a mechanism to differentiate real host IPIs from IPIs sent
by KVM for guest threads to poke each other, in order to target the
guest threads precisely when possible and avoid that global switch of
the core to host state.

We then use this new facility in the in-kernel XICS code.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
arch/powerpc/include/asm/kvm_book3s_asm.h
arch/powerpc/include/asm/kvm_ppc.h
arch/powerpc/kernel/asm-offsets.c
arch/powerpc/kvm/book3s_hv.c
arch/powerpc/kvm/book3s_hv_rmhandlers.S
arch/powerpc/kvm/book3s_xics.c
arch/powerpc/sysdev/xics/icp-native.c