H_CEDE should enable the vcpu's MSR:EE bit. It does on "HV" KVM (it's burried in the assembly code though) and as far as I can tell, qemu does it as well. Signed-off-by: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx> --- Alex, you probably want that in -now- (and maybe even in stable). Without that, bad things happen in guests, especially after Linus pulls what I'm about to send him to fix more issues with the way pseries handles H_CEDE vs. lazy IRQ disabling (it will hang in PR KVM without that fix). diff --git a/arch/powerpc/kvm/book3s_pr_papr.c b/arch/powerpc/kvm/book3s_pr_papr.c index 3ff9013..ee02b30 100644 --- a/arch/powerpc/kvm/book3s_pr_papr.c +++ b/arch/powerpc/kvm/book3s_pr_papr.c @@ -241,6 +241,7 @@ int kvmppc_h_pr(struct kvm_vcpu *vcpu, unsigned long cmd) case H_PUT_TCE: return kvmppc_h_pr_put_tce(vcpu); case H_CEDE: + vcpu->arch.shared->msr |= MSR_EE; kvm_vcpu_block(vcpu); clear_bit(KVM_REQ_UNHALT, &vcpu->requests); vcpu->stat.halt_wakeup++; -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
