On Mon, Oct 28, 2024, Nam Cao wrote:
> The hrtimer is initialized in the KVM_XEN_VCPU_SET_ATTR ioctl. That caused
> problem in the past, because the hrtimer can be initialized multiple times,
> which was fixed by commit af735db31285 ("KVM: x86/xen: Initialize Xen timer
> only once"). This commit avoids initializing the timer multiple times by
> checking the field 'function' of struct hrtimer to determine if it has
> already been initialized.
>
> Instead of "abusing" the 'function' field, move the hrtimer initialization
> into kvm_xen_init_vcpu() so that it will only be initialized once.
>
> Signed-off-by: Nam Cao <namcao@xxxxxxxxxxxxx>
> ---
> Cc: Sean Christopherson <seanjc@xxxxxxxxxx>
> Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> Cc: x86@xxxxxxxxxx
> Cc: kvm@xxxxxxxxxxxxxxx
> ---
> arch/x86/kvm/xen.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/arch/x86/kvm/xen.c b/arch/x86/kvm/xen.c
> index 622fe24da910..c386fbe6b58d 100644
> --- a/arch/x86/kvm/xen.c
> +++ b/arch/x86/kvm/xen.c
> @@ -1070,9 +1070,6 @@ int kvm_xen_vcpu_set_attr(struct kvm_vcpu *vcpu, struct kvm_xen_vcpu_attr *data)
> break;
> }
>
> - if (!vcpu->arch.xen.timer.function)
> - kvm_xen_init_timer(vcpu);
> -
> /* Stop the timer (if it's running) before changing the vector */
> kvm_xen_stop_timer(vcpu);
> vcpu->arch.xen.timer_virq = data->u.timer.port;
> @@ -2235,6 +2232,7 @@ void kvm_xen_init_vcpu(struct kvm_vcpu *vcpu)
> vcpu->arch.xen.poll_evtchn = 0;
>
> timer_setup(&vcpu->arch.xen.poll_timer, cancel_evtchn_poll, 0);
> + kvm_xen_init_timer(vcpu);
I vote for opportunistically dropping kvm_xen_init_timer() and open coding its
contents here.
If the intent is for subsystems to grab their relevant patches, I can fixup when
applying.
> kvm_gpc_init(&vcpu->arch.xen.runstate_cache, vcpu->kvm);
> kvm_gpc_init(&vcpu->arch.xen.runstate2_cache, vcpu->kvm);
> --
> 2.39.5
>
