[v2] KVM: s390: Add new reset vcpu API
diff mbox series

Message ID 20191203162055.3519-1-frankja@linux.ibm.com
State New
Headers show
Series
  • [v2] KVM: s390: Add new reset vcpu API
Related show

Commit Message

Janosch Frank Dec. 3, 2019, 4:20 p.m. UTC
The architecture states that we need to reset local IRQs for all CPU
resets. Because the old reset interface did not support the normal CPU
reset we never did that.

Now that we have a new interface, let's properly clear out local IRQs
and let this commit be a reminder.

Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
---
 arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
 include/uapi/linux/kvm.h |  4 ++++
 2 files changed, 17 insertions(+)

Comments

David Hildenbrand Dec. 3, 2019, 4:24 p.m. UTC | #1
On 03.12.19 17:20, Janosch Frank wrote:
> The architecture states that we need to reset local IRQs for all CPU
> resets. Because the old reset interface did not support the normal CPU
> reset we never did that.
> 
> Now that we have a new interface, let's properly clear out local IRQs
> and let this commit be a reminder.
> 
> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
> ---
>  arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
>  include/uapi/linux/kvm.h |  4 ++++
>  2 files changed, 17 insertions(+)
> 
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index d9e6bf3d54f0..602214c5616c 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -529,6 +529,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>  	case KVM_CAP_S390_CMMA_MIGRATION:
>  	case KVM_CAP_S390_AIS:
>  	case KVM_CAP_S390_AIS_MIGRATION:
> +	case KVM_CAP_S390_VCPU_RESETS:
>  		r = 1;
>  		break;
>  	case KVM_CAP_S390_HPAGE_1M:
> @@ -3287,6 +3288,13 @@ static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>  	return r;
>  }
>  
> +static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
> +{
> +	kvm_clear_async_pf_completion_queue(vcpu);
> +	kvm_s390_clear_local_irqs(vcpu);
> +	return 0;
> +}
> +
>  static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
>  {
>  	kvm_s390_vcpu_initial_reset(vcpu);
> @@ -4363,7 +4371,12 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
>  		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
>  		break;
>  	}
> +	case KVM_S390_NORMAL_RESET:
> +		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
> +		break;
>  	case KVM_S390_INITIAL_RESET:
> +		/* fallthrough */
> +	case KVM_S390_CLEAR_RESET:
>  		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
>  		break;
>  	case KVM_SET_ONE_REG:
> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
> index 52641d8ca9e8..f4fc865775a5 100644
> --- a/include/uapi/linux/kvm.h
> +++ b/include/uapi/linux/kvm.h
> @@ -1000,6 +1000,7 @@ struct kvm_ppc_resize_hpt {
>  #define KVM_CAP_PMU_EVENT_FILTER 173
>  #define KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 174
>  #define KVM_CAP_HYPERV_DIRECT_TLBFLUSH 175
> +#define KVM_CAP_S390_VCPU_RESETS 180
>  
>  #ifdef KVM_CAP_IRQ_ROUTING
>  
> @@ -1461,6 +1462,9 @@ struct kvm_enc_region {
>  /* Available with KVM_CAP_ARM_SVE */
>  #define KVM_ARM_VCPU_FINALIZE	  _IOW(KVMIO,  0xc2, int)
>  

Maybe also a comment like

"Available with KVM_CAP_S390_VCPU_RESETS" ?

> +#define KVM_S390_NORMAL_RESET	_IO(KVMIO,   0xc3)
> +#define KVM_S390_CLEAR_RESET    _IO(KVMIO,   0xc4)
> +
>  /* Secure Encrypted Virtualization command */
>  enum sev_cmd_id {
>  	/* Guest initialization commands */
> 

Reviewed-by: David Hildenbrand <david@redhat.com>
Cornelia Huck Dec. 3, 2019, 4:28 p.m. UTC | #2
On Tue,  3 Dec 2019 11:20:55 -0500
Janosch Frank <frankja@linux.ibm.com> wrote:

> The architecture states that we need to reset local IRQs for all CPU
> resets. Because the old reset interface did not support the normal CPU
> reset we never did that.
> 
> Now that we have a new interface, let's properly clear out local IRQs
> and let this commit be a reminder.

Well, the new interface did not appear magically, this patch introduces
it :) (The whole sentence is a bit confusing; what is this commit
supposed to remind us of?)

> 
> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
> ---
>  arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
>  include/uapi/linux/kvm.h |  4 ++++
>  2 files changed, 17 insertions(+)
> 
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index d9e6bf3d54f0..602214c5616c 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -529,6 +529,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>  	case KVM_CAP_S390_CMMA_MIGRATION:
>  	case KVM_CAP_S390_AIS:
>  	case KVM_CAP_S390_AIS_MIGRATION:
> +	case KVM_CAP_S390_VCPU_RESETS:
>  		r = 1;
>  		break;
>  	case KVM_CAP_S390_HPAGE_1M:
> @@ -3287,6 +3288,13 @@ static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>  	return r;
>  }
>  
> +static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
> +{
> +	kvm_clear_async_pf_completion_queue(vcpu);
> +	kvm_s390_clear_local_irqs(vcpu);
> +	return 0;
> +}
> +
>  static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
>  {
>  	kvm_s390_vcpu_initial_reset(vcpu);
> @@ -4363,7 +4371,12 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
>  		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
>  		break;
>  	}
> +	case KVM_S390_NORMAL_RESET:
> +		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
> +		break;
>  	case KVM_S390_INITIAL_RESET:
> +		/* fallthrough */
> +	case KVM_S390_CLEAR_RESET:
>  		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);

The cover letter probably should also mention that clear reset falls
back to initial reset for now?

>  		break;
>  	case KVM_SET_ONE_REG:
> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
> index 52641d8ca9e8..f4fc865775a5 100644
> --- a/include/uapi/linux/kvm.h
> +++ b/include/uapi/linux/kvm.h
> @@ -1000,6 +1000,7 @@ struct kvm_ppc_resize_hpt {
>  #define KVM_CAP_PMU_EVENT_FILTER 173
>  #define KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 174
>  #define KVM_CAP_HYPERV_DIRECT_TLBFLUSH 175
> +#define KVM_CAP_S390_VCPU_RESETS 180
>  
>  #ifdef KVM_CAP_IRQ_ROUTING
>  
> @@ -1461,6 +1462,9 @@ struct kvm_enc_region {
>  /* Available with KVM_CAP_ARM_SVE */
>  #define KVM_ARM_VCPU_FINALIZE	  _IOW(KVMIO,  0xc2, int)
>  
> +#define KVM_S390_NORMAL_RESET	_IO(KVMIO,   0xc3)
> +#define KVM_S390_CLEAR_RESET    _IO(KVMIO,   0xc4)
> +
>  /* Secure Encrypted Virtualization command */
>  enum sev_cmd_id {
>  	/* Guest initialization commands */

Looks sane, but please also document the new ioctls and the new cap.
Thomas Huth Dec. 4, 2019, 9:32 a.m. UTC | #3
On 03/12/2019 17.20, Janosch Frank wrote:
> The architecture states that we need to reset local IRQs for all CPU
> resets. Because the old reset interface did not support the normal CPU
> reset we never did that.
> 
> Now that we have a new interface, let's properly clear out local IRQs
> and let this commit be a reminder.
> 
> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
> ---
>  arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
>  include/uapi/linux/kvm.h |  4 ++++
>  2 files changed, 17 insertions(+)
> 
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index d9e6bf3d54f0..602214c5616c 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -529,6 +529,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>  	case KVM_CAP_S390_CMMA_MIGRATION:
>  	case KVM_CAP_S390_AIS:
>  	case KVM_CAP_S390_AIS_MIGRATION:
> +	case KVM_CAP_S390_VCPU_RESETS:
>  		r = 1;
>  		break;
>  	case KVM_CAP_S390_HPAGE_1M:
> @@ -3287,6 +3288,13 @@ static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>  	return r;
>  }
>  
> +static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
> +{
> +	kvm_clear_async_pf_completion_queue(vcpu);
> +	kvm_s390_clear_local_irqs(vcpu);
> +	return 0;
> +}
> +
>  static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
>  {
>  	kvm_s390_vcpu_initial_reset(vcpu);
> @@ -4363,7 +4371,12 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
>  		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
>  		break;
>  	}
> +	case KVM_S390_NORMAL_RESET:
> +		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
> +		break;
>  	case KVM_S390_INITIAL_RESET:
> +		/* fallthrough */
> +	case KVM_S390_CLEAR_RESET:
>  		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
>  		break;

Isn't clear reset supposed to do more than the initial reset? If so, I'd
rather expect it the other way round:

	case KVM_S390_CLEAR_RESET:
		/* fallthrough */
	case KVM_S390_INITIAL_RESET:
		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
		break;

... so you can later add the additional stuff before the "/* fallthrough
*/"?

 Thomas
Janosch Frank Dec. 4, 2019, 10:06 a.m. UTC | #4
On 12/4/19 10:32 AM, Thomas Huth wrote:
> On 03/12/2019 17.20, Janosch Frank wrote:
>> The architecture states that we need to reset local IRQs for all CPU
>> resets. Because the old reset interface did not support the normal CPU
>> reset we never did that.
>>
>> Now that we have a new interface, let's properly clear out local IRQs
>> and let this commit be a reminder.
>>
>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>> ---
>>  arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
>>  include/uapi/linux/kvm.h |  4 ++++
>>  2 files changed, 17 insertions(+)
>>
>> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
>> index d9e6bf3d54f0..602214c5616c 100644
>> --- a/arch/s390/kvm/kvm-s390.c
>> +++ b/arch/s390/kvm/kvm-s390.c
>> @@ -529,6 +529,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>>  	case KVM_CAP_S390_CMMA_MIGRATION:
>>  	case KVM_CAP_S390_AIS:
>>  	case KVM_CAP_S390_AIS_MIGRATION:
>> +	case KVM_CAP_S390_VCPU_RESETS:
>>  		r = 1;
>>  		break;
>>  	case KVM_CAP_S390_HPAGE_1M:
>> @@ -3287,6 +3288,13 @@ static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>>  	return r;
>>  }
>>  
>> +static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
>> +{
>> +	kvm_clear_async_pf_completion_queue(vcpu);
>> +	kvm_s390_clear_local_irqs(vcpu);
>> +	return 0;
>> +}
>> +
>>  static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
>>  {
>>  	kvm_s390_vcpu_initial_reset(vcpu);
>> @@ -4363,7 +4371,12 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
>>  		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
>>  		break;
>>  	}
>> +	case KVM_S390_NORMAL_RESET:
>> +		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
>> +		break;
>>  	case KVM_S390_INITIAL_RESET:
>> +		/* fallthrough */
>> +	case KVM_S390_CLEAR_RESET:
>>  		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
>>  		break;
> 
> Isn't clear reset supposed to do more than the initial reset? If so, I'd
> rather expect it the other way round:

In the PV patch I remove the fallthrough and add a function for the
clear reset. I add the UV reset calls into the
kvm_arch_vcpu_ioctl_*_reset() functions, therefore I can't fallthrough
because the Ultravisor does currently not allow staged resets (i.e.
first initial then clear if a clear reset is needed)

> 
> 	case KVM_S390_CLEAR_RESET:
> 		/* fallthrough */
> 	case KVM_S390_INITIAL_RESET:
> 		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
> 		break;
> 
> ... so you can later add the additional stuff before the "/* fallthrough
> */"?
> 
>  Thomas
>
Thomas Huth Dec. 4, 2019, 10:37 a.m. UTC | #5
On 04/12/2019 11.06, Janosch Frank wrote:
> On 12/4/19 10:32 AM, Thomas Huth wrote:
>> On 03/12/2019 17.20, Janosch Frank wrote:
>>> The architecture states that we need to reset local IRQs for all CPU
>>> resets. Because the old reset interface did not support the normal CPU
>>> reset we never did that.
>>>
>>> Now that we have a new interface, let's properly clear out local IRQs
>>> and let this commit be a reminder.
>>>
>>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>>> ---
>>>  arch/s390/kvm/kvm-s390.c | 13 +++++++++++++
>>>  include/uapi/linux/kvm.h |  4 ++++
>>>  2 files changed, 17 insertions(+)
>>>
>>> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
>>> index d9e6bf3d54f0..602214c5616c 100644
>>> --- a/arch/s390/kvm/kvm-s390.c
>>> +++ b/arch/s390/kvm/kvm-s390.c
>>> @@ -529,6 +529,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>>>  	case KVM_CAP_S390_CMMA_MIGRATION:
>>>  	case KVM_CAP_S390_AIS:
>>>  	case KVM_CAP_S390_AIS_MIGRATION:
>>> +	case KVM_CAP_S390_VCPU_RESETS:
>>>  		r = 1;
>>>  		break;
>>>  	case KVM_CAP_S390_HPAGE_1M:
>>> @@ -3287,6 +3288,13 @@ static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>>>  	return r;
>>>  }
>>>  
>>> +static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
>>> +{
>>> +	kvm_clear_async_pf_completion_queue(vcpu);
>>> +	kvm_s390_clear_local_irqs(vcpu);
>>> +	return 0;
>>> +}
>>> +
>>>  static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
>>>  {
>>>  	kvm_s390_vcpu_initial_reset(vcpu);
>>> @@ -4363,7 +4371,12 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
>>>  		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
>>>  		break;
>>>  	}
>>> +	case KVM_S390_NORMAL_RESET:
>>> +		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
>>> +		break;
>>>  	case KVM_S390_INITIAL_RESET:
>>> +		/* fallthrough */
>>> +	case KVM_S390_CLEAR_RESET:
>>>  		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
>>>  		break;
>>
>> Isn't clear reset supposed to do more than the initial reset? If so, I'd
>> rather expect it the other way round:
> 
> In the PV patch I remove the fallthrough and add a function for the
> clear reset. I add the UV reset calls into the
> kvm_arch_vcpu_ioctl_*_reset() functions, therefore I can't fallthrough
> because the Ultravisor does currently not allow staged resets (i.e.
> first initial then clear if a clear reset is needed)

Ok, then it's fine. But in case you respin your patch, it's maybe less
confusing if you swap it.

 Thomas

Patch
diff mbox series

diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index d9e6bf3d54f0..602214c5616c 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -529,6 +529,7 @@  int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_S390_CMMA_MIGRATION:
 	case KVM_CAP_S390_AIS:
 	case KVM_CAP_S390_AIS_MIGRATION:
+	case KVM_CAP_S390_VCPU_RESETS:
 		r = 1;
 		break;
 	case KVM_CAP_S390_HPAGE_1M:
@@ -3287,6 +3288,13 @@  static int kvm_arch_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
 	return r;
 }
 
+static int kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
+{
+	kvm_clear_async_pf_completion_queue(vcpu);
+	kvm_s390_clear_local_irqs(vcpu);
+	return 0;
+}
+
 static int kvm_arch_vcpu_ioctl_initial_reset(struct kvm_vcpu *vcpu)
 {
 	kvm_s390_vcpu_initial_reset(vcpu);
@@ -4363,7 +4371,12 @@  long kvm_arch_vcpu_ioctl(struct file *filp,
 		r = kvm_arch_vcpu_ioctl_set_initial_psw(vcpu, psw);
 		break;
 	}
+	case KVM_S390_NORMAL_RESET:
+		r = kvm_arch_vcpu_ioctl_normal_reset(vcpu);
+		break;
 	case KVM_S390_INITIAL_RESET:
+		/* fallthrough */
+	case KVM_S390_CLEAR_RESET:
 		r = kvm_arch_vcpu_ioctl_initial_reset(vcpu);
 		break;
 	case KVM_SET_ONE_REG:
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 52641d8ca9e8..f4fc865775a5 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -1000,6 +1000,7 @@  struct kvm_ppc_resize_hpt {
 #define KVM_CAP_PMU_EVENT_FILTER 173
 #define KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 174
 #define KVM_CAP_HYPERV_DIRECT_TLBFLUSH 175
+#define KVM_CAP_S390_VCPU_RESETS 180
 
 #ifdef KVM_CAP_IRQ_ROUTING
 
@@ -1461,6 +1462,9 @@  struct kvm_enc_region {
 /* Available with KVM_CAP_ARM_SVE */
 #define KVM_ARM_VCPU_FINALIZE	  _IOW(KVMIO,  0xc2, int)
 
+#define KVM_S390_NORMAL_RESET	_IO(KVMIO,   0xc3)
+#define KVM_S390_CLEAR_RESET    _IO(KVMIO,   0xc4)
+
 /* Secure Encrypted Virtualization command */
 enum sev_cmd_id {
 	/* Guest initialization commands */