Hi Christoffer,
On 07/18/2015 11:09 AM, Christoffer Dall wrote:
> On Mon, Jul 06, 2015 at 02:49:55PM +0200, Eric Auger wrote:
>> The kvm_vcpu_arch pause field is renamed into power_off to prepare
>> for the introduction of a new pause field.
>>
>> Signed-off-by: Eric Auger <[email protected]>
>>
>> v4 -> v5:
>> - fix compilation issue on arm64 (add power_off field in kvm_host.h)
>> ---
>> arch/arm/include/asm/kvm_host.h | 4 ++--
>> arch/arm/kvm/arm.c | 10 +++++-----
>> arch/arm/kvm/psci.c | 10 +++++-----
>> arch/arm64/include/asm/kvm_host.h | 4 ++--
>> 4 files changed, 14 insertions(+), 14 deletions(-)
>>
>> diff --git a/arch/arm/include/asm/kvm_host.h
>> b/arch/arm/include/asm/kvm_host.h
>> index e896d2c..304004d 100644
>> --- a/arch/arm/include/asm/kvm_host.h
>> +++ b/arch/arm/include/asm/kvm_host.h
>> @@ -129,8 +129,8 @@ struct kvm_vcpu_arch {
>> * here.
>> */
>>
>> - /* Don't run the guest on this vcpu */
>> - bool pause;
>> + /* vcpu power-off state */
>> + bool power_off;
>>
>> /* IO related fields */
>> struct kvm_decode mmio_decode;
>> diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
>> index bcdf799..7537e68 100644
>> --- a/arch/arm/kvm/arm.c
>> +++ b/arch/arm/kvm/arm.c
>> @@ -475,7 +475,7 @@ static void vcpu_pause(struct kvm_vcpu *vcpu)
>> {
>> wait_queue_head_t *wq = kvm_arch_vcpu_wq(vcpu);
>>
>> - wait_event_interruptible(*wq, !vcpu->arch.pause);
>> + wait_event_interruptible(*wq, !vcpu->arch.power_off);
>
> would there be any benefit to simply calling kvm_vcpu_block() instead of
> vcpu_pause, and rewrite kvm_arch_vcpu_runnable to:
Wouldn't it somehow change the known behavior or kvm_vcpu_block which is
expected/used to exit on IRQ/FIQ (WFI). Here it would exit when
power_off changes to false (or maybe you meant pause below in the new
context?).
>
> int kvm_arch_vcpu_runnable(struct kvm_vcpu *v)
> {
> ▸ return !vcpu->arch.power_off &&
> (!!v->arch.irq_lines || kvm_vgic_vcpu_pending_irq(v));
> }
>
> Not sure really, certainly the runnable function does not become more
> readable.
To me the usage of kvm_vcpu_block looks more complex than this code and
I would prefer keeping that version if you don't mind.
>
>> }
>>
>> static int kvm_vcpu_initialized(struct kvm_vcpu *vcpu)
>> @@ -525,7 +525,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu,
>> struct kvm_run *run)
>>
>> update_vttbr(vcpu->kvm);
>>
>> - if (vcpu->arch.pause)
>> + if (vcpu->arch.power_off)
>> vcpu_pause(vcpu);
>
> looking back over this code, how does this actually guarantee that we
> don't run a powered-off cpu?
>
> vcpu_pause() just does a wait_event_interruptible(), so if we get
> scheduled again, we'll just proceed running.
actually it also checks the !vcpu->arch.power_off condition, right?
Is there any case where we
> could get scheduled without signal_pending() being true and therefore
> inadvertedly run the vcpu?
kvm_arm_halt_guest can happen at any time, including after the execution
of above vcpu_pause(vcpu) call. This is the reason why I added the
second check below, once we entered the critical section and just before
running the vcpu.
With regard to renamed power_off boolean my understanding is:
power_off is set
- on kvm_arch_vcpu_ioctl_vcpu_init/KVM_ARM_VCPU_POWER_OFF
- on PSCI calls from guests through traps from HVC instructions
in that case I don't think this can happen
- on KVM_SET_MP_STATE ioctl: I think in that case, what you describe can
happen.
Do you share the same understanding?
Best Regards
Eric
>
> if so, we should change the line below like this:
>
> diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
> index bc738d2..98f31e6 100644
> --- a/arch/arm/kvm/arm.c
> +++ b/arch/arm/kvm/arm.c
> @@ -542,7 +542,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct
> kvm_run *run)
> run->exit_reason = KVM_EXIT_INTR;
> }
>
> - if (ret <= 0 || need_new_vmid_gen(vcpu->kvm)) {
> + if (ret <= 0 || need_new_vmid_gen(vcpu->kvm) ||
> + vcpu->arch.power_off) {
> local_irq_enable();
> preempt_enable();
> kvm_timer_sync_hwstate(vcpu);
>
>
> Sorry for polluting your patch with these questions, I'm otherwise fine
> with the rename.
>
> Thanks,
> -Christoffer
>
>>
>> /*
>> @@ -766,12 +766,12 @@ static int kvm_arch_vcpu_ioctl_vcpu_init(struct
>> kvm_vcpu *vcpu,
>> vcpu_reset_hcr(vcpu);
>>
>> /*
>> - * Handle the "start in power-off" case by marking the VCPU as paused.
>> + * Handle the "start in power-off" case.
>> */
>> if (test_bit(KVM_ARM_VCPU_POWER_OFF, vcpu->arch.features))
>> - vcpu->arch.pause = true;
>> + vcpu->arch.power_off = true;
>> else
>> - vcpu->arch.pause = false;
>> + vcpu->arch.power_off = false;
>>
>> return 0;
>> }
>> diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
>> index 4b94b51..134971a 100644
>> --- a/arch/arm/kvm/psci.c
>> +++ b/arch/arm/kvm/psci.c
>> @@ -63,7 +63,7 @@ static unsigned long kvm_psci_vcpu_suspend(struct kvm_vcpu
>> *vcpu)
>>
>> static void kvm_psci_vcpu_off(struct kvm_vcpu *vcpu)
>> {
>> - vcpu->arch.pause = true;
>> + vcpu->arch.power_off = true;
>> }
>>
>> static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu *source_vcpu)
>> @@ -87,7 +87,7 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu
>> *source_vcpu)
>> */
>> if (!vcpu)
>> return PSCI_RET_INVALID_PARAMS;
>> - if (!vcpu->arch.pause) {
>> + if (!vcpu->arch.power_off) {
>> if (kvm_psci_version(source_vcpu) != KVM_ARM_PSCI_0_1)
>> return PSCI_RET_ALREADY_ON;
>> else
>> @@ -115,7 +115,7 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu
>> *source_vcpu)
>> * the general puspose registers are undefined upon CPU_ON.
>> */
>> *vcpu_reg(vcpu, 0) = context_id;
>> - vcpu->arch.pause = false;
>> + vcpu->arch.power_off = false;
>> smp_mb(); /* Make sure the above is visible */
>>
>> wq = kvm_arch_vcpu_wq(vcpu);
>> @@ -152,7 +152,7 @@ static unsigned long kvm_psci_vcpu_affinity_info(struct
>> kvm_vcpu *vcpu)
>> kvm_for_each_vcpu(i, tmp, kvm) {
>> mpidr = kvm_vcpu_get_mpidr_aff(tmp);
>> if (((mpidr & target_affinity_mask) == target_affinity) &&
>> - !tmp->arch.pause) {
>> + !tmp->arch.power_off) {
>> return PSCI_0_2_AFFINITY_LEVEL_ON;
>> }
>> }
>> @@ -175,7 +175,7 @@ static void kvm_prepare_system_event(struct kvm_vcpu
>> *vcpu, u32 type)
>> * re-initialized.
>> */
>> kvm_for_each_vcpu(i, tmp, vcpu->kvm) {
>> - tmp->arch.pause = true;
>> + tmp->arch.power_off = true;
>> kvm_vcpu_kick(tmp);
>> }
>>
>> diff --git a/arch/arm64/include/asm/kvm_host.h
>> b/arch/arm64/include/asm/kvm_host.h
>> index 2709db2..009da6b 100644
>> --- a/arch/arm64/include/asm/kvm_host.h
>> +++ b/arch/arm64/include/asm/kvm_host.h
>> @@ -122,8 +122,8 @@ struct kvm_vcpu_arch {
>> * here.
>> */
>>
>> - /* Don't run the guest */
>> - bool pause;
>> + /* vcpu power-off state */
>> + bool power_off;
>>
>> /* IO related fields */
>> struct kvm_decode mmio_decode;
>> --
>> 1.9.1
>>
_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
