On Thu, Sep 22, 2022, isaku.yamahata@xxxxxxxxx wrote:
> From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
>
> Factor out the logic on reboot event as arch hook. Later kvm/x86 overrides
> it.
>
> Suggested-by: Sean Christopherson <seanjc@xxxxxxxxxx>
> Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> ---
...
> @@ -5135,6 +5141,8 @@ static void kvm_del_vm(void)
> static int kvm_reboot(struct notifier_block *notifier, unsigned long val,
> void *v)
> {
> + int r;
> +
> /*
> * Some (well, at least mine) BIOSes hang on reboot if
> * in vmx root mode.
> @@ -5143,8 +5151,14 @@ static int kvm_reboot(struct notifier_block *notifier, unsigned long val,
> */
> pr_info("kvm: exiting hardware virtualization\n");
> kvm_rebooting = true;
> - on_each_cpu(hardware_disable_nolock, NULL, 1);
> - return NOTIFY_OK;
> +
> + /* This hook is called without cpuhotplug disabled. */
> + cpus_read_lock();
> + mutex_lock(&kvm_lock);
> + r = kvm_arch_reboot(val);
Unless there's a valid use case for rejecting/stopping reboot, which I'm pretty
sure there isn't, don't allow arch code to return a value. I.e. return NOTIFY_OK
unconditionally from kvm_reboot() and drop the return from kvm_arch_reboot().
> + mutex_unlock(&kvm_lock);
> + cpus_read_unlock();
> + return r;
> }
>
> static struct notifier_block kvm_reboot_notifier = {
> --
> 2.25.1
>
