2016-07-01 10:33+0200, Paolo Bonzini: > On 30/06/2016 22:54, Radim Krčmář wrote: >> +static void __kvm_apic_state_fixup(struct kvm_vcpu *vcpu, >> + struct kvm_lapic_state *s, bool set) >> +{ >> + if (apic_x2apic_mode(vcpu->arch.apic)) { >> + u32 *id = (u32 *)(s->regs + APIC_ID); >> + if (set) >> + *id >>= 24; >> + else >> + *id <<= 24; >> + } > > When setting, this should read from the apic_base being set. So I think > you cannot use apic_x2apic_mode.
apic_x2apic_mode uses apic_base MSR, so its value does not depend on LAPIC_SET/GET. I don't like the dependency much, but all combinations of values/orders should work well. > With this change, you don't need the struct kvm_vcpu argument now; add a > struct kvm argument instead in patch 10. I don't think we can use hardware-compatible format in API without knowing the x2apic mode through vcpu. (The big endian format could do without knowing apic mode.) >> @@ -2780,6 +2780,7 @@ static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu >> *vcpu, >> kvm_x86_ops->sync_pir_to_irr(vcpu); >> >> memcpy(s->regs, vcpu->arch.apic->regs, sizeof *s); >> + kvm_apic_state_get_fixup(vcpu, s); > > Instead of kvm_apic_state_get/set_fixup, group the memcpy and > __kvm_apic_state_fixup in a new function kvm_apic_get_state(struct > kvm_lapic *apic, char *regs). Ok. >> return 0; >> } >> @@ -2787,6 +2788,7 @@ static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu >> *vcpu, >> static int kvm_vcpu_ioctl_set_lapic(struct kvm_vcpu *vcpu, >> struct kvm_lapic_state *s) >> { >> + kvm_apic_state_set_fixup(vcpu, s); >> kvm_apic_post_state_restore(vcpu, s); > > ... and likewise merge these two in a refactored > kvm_apic_post_state_restore called kvm_apic_set_state(struct kvm_lapic > *apic, char *regs), by calling __kvm_apic_state_fixup before > kvm_apic_post_state_restore's memcpy. > > With these changes I guess there's no need for the underscores in > __kvm_apic_state_fixup. Yes, good idea. > And you can also change the struct > kvm_lapic_state pointer in the function to a char *. Ok, I'll pass just s->regs. Thanks.