Date:   Mon, 4 Jul 2022 15:56:46 +0200
From:   Pierre Morel <pmorel@...ux.ibm.com>
To:     Janis Schoetterl-Glausch <scgl@...ux.ibm.com>, kvm@...r.kernel.org
Cc:     linux-s390@...r.kernel.org, linux-kernel@...r.kernel.org,
        borntraeger@...ibm.com, frankja@...ux.ibm.com, cohuck@...hat.com,
        david@...hat.com, thuth@...hat.com, imbrenda@...ux.ibm.com,
        hca@...ux.ibm.com, gor@...ux.ibm.com, wintera@...ux.ibm.com,
        seiden@...ux.ibm.com, nrb@...ux.ibm.com
Subject: Re: [PATCH v11 3/3] KVM: s390: resetting the Topology-Change-Report



On 7/4/22 11:35, Janis Schoetterl-Glausch wrote:
> On 7/1/22 18:25, Pierre Morel wrote:
>> During a subsystem reset the Topology-Change-Report is cleared.
>>
>> Let's give userland the possibility to clear the MTCR in the case
>> of a subsystem reset.
>>
>> To migrate the MTCR, we give userland the possibility to
>> query the MTCR state.
>>
>> We indicate KVM support for the CPU topology facility with a new
>> KVM capability: KVM_CAP_S390_CPU_TOPOLOGY.
>>
>> Signed-off-by: Pierre Morel <pmorel@...ux.ibm.com>
>> ---
>>   Documentation/virt/kvm/api.rst   | 25 +++++++++++++++
>>   arch/s390/include/uapi/asm/kvm.h | 10 ++++++
>>   arch/s390/kvm/kvm-s390.c         | 53 ++++++++++++++++++++++++++++++++
>>   include/uapi/linux/kvm.h         |  1 +
>>   4 files changed, 89 insertions(+)
>>
>> diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst
>> index 11e00a46c610..5e086125d8ad 100644
>> --- a/Documentation/virt/kvm/api.rst
>> +++ b/Documentation/virt/kvm/api.rst
>> @@ -7956,6 +7956,31 @@ should adjust CPUID leaf 0xA to reflect that the PMU is disabled.
>>   When enabled, KVM will exit to userspace with KVM_EXIT_SYSTEM_EVENT of
>>   type KVM_SYSTEM_EVENT_SUSPEND to process the guest suspend request.
>>   
>> +8.37 KVM_CAP_S390_CPU_TOPOLOGY
>> +------------------------------
>> +
>> +:Capability: KVM_CAP_S390_CPU_TOPOLOGY
>> +:Architectures: s390
>> +:Type: vm
>> +
>> +This capability indicates that KVM will provide the S390 CPU Topology
>> +facility which consist of the interpretation of the PTF instruction for
>> +the function code 2 along with interception and forwarding of both the
>> +PTF instruction with function codes 0 or 1 and the STSI(15,1,x)
>> +instruction to the userland hypervisor.
> The latter only if the user STSI capability is also enabled.

Hum, not sure about this.
we can not set facility 11 and return 3 to STSI(15) for valid selectors.

I think that it was right before, KVM_CAP_S390_CPU_TOPOLOGY and 
KVM_CAP_S390_USER_STSI are independent in KVM, userland can turn on one 
and not the other.
But KVM proposes both.

Of course it is stupid to turn on only KVM_CAP_S390_CPU_TOPOLOGY but KVM 
is not responsible for this userland is.

Otherwise, we need to check on KVM_CAP_S390_USER_STSI before authorizing 
  KVM_CAP_S390_CPU_TOPOLOGY and that looks even more complicated for me,
or we suppress the KVM_CAP_S390_CPU_TOPOLOGY and implement the all 
stsi(15) in the kernel what I really do not think is good because of the 
complexity of the userland API

>> +
>> +The stfle facility 11, CPU Topology facility, should not be indicated
>> +to the guest without this capability.
>> +
>> +When this capability is present, KVM provides a new attribute group
>> +on vm fd, KVM_S390_VM_CPU_TOPOLOGY.
>> +This new attribute allows to get, set or clear the Modified Change
>> +Topology Report (MTCR) bit of the SCA through the kvm_device_attr
>> +structure.
>> +
>> +When getting the Modified Change Topology Report value, the attr->addr
>> +must point to a byte where the value will be stored.
>> +
>>   9. Known KVM API problems
>>   =========================
>>   
>> diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
>> index 7a6b14874d65..df5e8279ffd0 100644
>> --- a/arch/s390/include/uapi/asm/kvm.h
>> +++ b/arch/s390/include/uapi/asm/kvm.h
>> @@ -74,6 +74,7 @@ struct kvm_s390_io_adapter_req {
>>   #define KVM_S390_VM_CRYPTO		2
>>   #define KVM_S390_VM_CPU_MODEL		3
>>   #define KVM_S390_VM_MIGRATION		4
>> +#define KVM_S390_VM_CPU_TOPOLOGY	5
>>   
>>   /* kvm attributes for mem_ctrl */
>>   #define KVM_S390_VM_MEM_ENABLE_CMMA	0
>> @@ -171,6 +172,15 @@ struct kvm_s390_vm_cpu_subfunc {
>>   #define KVM_S390_VM_MIGRATION_START	1
>>   #define KVM_S390_VM_MIGRATION_STATUS	2
>>   
>> +/* kvm attributes for cpu topology */
>> +#define KVM_S390_VM_CPU_TOPO_MTCR_CLEAR	0
>> +#define KVM_S390_VM_CPU_TOPO_MTCR_SET	1
>> +
>> +struct kvm_cpu_topology {
>> +	__u16 mtcr : 1;
>> +	__u16 reserved : 15;
>> +};
> 
> This is no longer used, is it?

No, I sent the wrong patch it seems!! Sorry for that.
There is nothing more in kvm.h now but the definition for 
KVM_S390_VM_CPU_TOPOLOGY




>> +
>>   /* for KVM_GET_REGS and KVM_SET_REGS */
>>   struct kvm_regs {
>>   	/* general purpose regs for s390 */
>> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
>> index ee59b03f2e45..5029fe40adbd 100644
>> --- a/arch/s390/kvm/kvm-s390.c
>> +++ b/arch/s390/kvm/kvm-s390.c
>> @@ -606,6 +606,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>>   	case KVM_CAP_S390_PROTECTED:
>>   		r = is_prot_virt_host();
>>   		break;
>> +	case KVM_CAP_S390_CPU_TOPOLOGY:
>> +		r = test_facility(11);
>> +		break;
>>   	default:
>>   		r = 0;
>>   	}
>> @@ -817,6 +820,20 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
>>   		icpt_operexc_on_all_vcpus(kvm);
>>   		r = 0;
>>   		break;
>> +	case KVM_CAP_S390_CPU_TOPOLOGY:
>> +		r = -EINVAL;
>> +		mutex_lock(&kvm->lock);
>> +		if (kvm->created_vcpus) {
>> +			r = -EBUSY;
>> +		} else if (test_facility(11)) {
>> +			set_kvm_facility(kvm->arch.model.fac_mask, 11);
>> +			set_kvm_facility(kvm->arch.model.fac_list, 11);
>> +			r = 0;
>> +		}
>> +		mutex_unlock(&kvm->lock);
>> +		VM_EVENT(kvm, 3, "ENABLE: CPU TOPOLOGY %s",
> 
> I still would go for consistency here, "ENABLE: CAP_S390_CPU_TOPOLOGY %s".

Yes, done.

> 
>> +			 r ? "(not available)" : "(success)");
>> +		break;
>>   	default:
>>   		r = -EINVAL;
>>   		break;
>> @@ -1716,6 +1733,33 @@ static void kvm_s390_update_topology_change_report(struct kvm *kvm, bool val)
>>   	read_unlock(&kvm->arch.sca_lock);
>>   }
>>   
>> +static int kvm_s390_set_topology(struct kvm *kvm, struct kvm_device_attr *attr)
>> +{
>> +	if (!test_kvm_facility(kvm, 11))
>> +		return -ENXIO;
>> +
>> +	kvm_s390_update_topology_change_report(kvm, !!attr->attr);
> 
> Will this not be automatically clamped to 0,1 if the argument has type bool?

I do not know, anyway done like this is sure.

>> +	return 0;
>> +}
>> +
>> +static int kvm_s390_get_topology(struct kvm *kvm, struct kvm_device_attr *attr)
>> +{
>> +	union sca_utility utility;
>> +	struct bsca_block *sca = kvm->arch.sca;
>> +	__u8 topo;
>> +
>> +	if (!test_kvm_facility(kvm, 11))
>> +		return -ENXIO;
>> +
>          read_lock(&kvm->arch.sca_lock);
>          utility.val = READ_ONCE(kvm->arch.sca->utility.val);
>          read_unlock(&kvm->arch.sca_lock); >
> And then get rid of the sca declaration.


OK

>> +	topo = utility.mtcr;
>> +
>> +	if (copy_to_user((void __user *)attr->addr, &topo, sizeof(topo)))
>> +		return -EFAULT;
>> +
>> +	return 0;
>> +}
>> +
> [...]
> 

-- 
Pierre Morel
IBM Lab Boeblingen

Powered by blists - more mailing lists