[PATCH v2 04/10] KVM: arm: guest debug, add stub KVM_SET_GUEST_DEBUG ioctl

[Alex Bennée]

David Hildenbrand <dahi at linux.vnet.ibm.com> writes:

>> On Tue, Mar 31, 2015 at 04:08:02PM +0100, Alex Bennée wrote:
>> > This commit adds a stub function to support the KVM_SET_GUEST_DEBUG
>> > ioctl. Currently any operation flag will return EINVAL. Actual
>> > functionality will be added with further patches.
>> > 
>> > Signed-off-by: Alex Bennée <alex.bennee at linaro.org>.
>> > 
>> > ---
>> > v2
>> >   - simplified form of the ioctl (stuff will go into setup_debug)
>> > 
>> > diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
>> > index b112efc..06c5064 100644
>> > --- a/Documentation/virtual/kvm/api.txt
>> > +++ b/Documentation/virtual/kvm/api.txt
>> > @@ -2604,7 +2604,7 @@ handled.
>> >  4.87 KVM_SET_GUEST_DEBUG
>> >  
>> >  Capability: KVM_CAP_SET_GUEST_DEBUG
>> > -Architectures: x86, s390, ppc
>> > +Architectures: x86, s390, ppc, arm64
>> >  Type: vcpu ioctl
>> >  Parameters: struct kvm_guest_debug (in)
>> >  Returns: 0 on success; -1 on error
>> > diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
>> > index 5560f74..445933d 100644
>> > --- a/arch/arm/kvm/arm.c
>> > +++ b/arch/arm/kvm/arm.c
>> > @@ -183,6 +183,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>> >  	case KVM_CAP_ARM_PSCI:
>> >  	case KVM_CAP_ARM_PSCI_0_2:
>> >  	case KVM_CAP_READONLY_MEM:
>> > +	case KVM_CAP_SET_GUEST_DEBUG:
>> >  		r = 1;
>> >  		break;
>> 
>> shouldn't you wait with advertising this capability until you've
>> implemented support for it?
>> 
>
> I think this would work for now, however it's not very practical
> - in the end one has to sense which debug flags are actually supported.
>
> Question is if he wants to add initial support and extend functionality and
> flags with each patch or enable the whole set of features in one shot at the
> end.

This is what in effect I was doing. Testing each feature in turn and
ensuring it failed gracefully when kernel features where not present
(both missing the capability or returning EINVAL).

> Doing the latter seems more practicable to me (especially as the debug features
> are added in the same patch series).

Well in practice the whole series will go in together so this is only
really relevant to what happens when you bisect the series.

>
> David

-- 
Alex Bennée