Re: [RFC PATCH 1/8] KVM: Document KVM_MAP_MEMORY ioctl

[Isaku Yamahata <isaku.yamahata@linux.intel.com>]

On Thu, Mar 07, 2024 at 12:30:04PM +0000,
"Huang, Kai" <kai.huang@intel.com> wrote:

> On Fri, 2024-03-01 at 09:28 -0800, isaku.yamahata@intel.com wrote:
> > From: Isaku Yamahata <isaku.yamahata@intel.com>
> > 
> > Adds documentation of KVM_MAP_MEMORY ioctl.
> > 
> > It pre-populates guest memory. And potentially do initialized memory
> > contents with encryption and measurement depending on underlying
> > technology.
> > 
> > Suggested-by: Sean Christopherson <seanjc@google.com>
> > Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > ---
> >  Documentation/virt/kvm/api.rst | 36 ++++++++++++++++++++++++++++++++++
> >  1 file changed, 36 insertions(+)
> > 
> > diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst
> > index 0b5a33ee71ee..33d2b63f7dbf 100644
> > --- a/Documentation/virt/kvm/api.rst
> > +++ b/Documentation/virt/kvm/api.rst
> > @@ -6352,6 +6352,42 @@ a single guest_memfd file, but the bound ranges must not overlap).
> >  
> >  See KVM_SET_USER_MEMORY_REGION2 for additional details.
> >  
> > +4.143 KVM_MAP_MEMORY
> > +------------------------
> > +
> > +:Capability: KVM_CAP_MAP_MEMORY
> > +:Architectures: none
> > +:Type: vcpu ioctl
> 
> I think "vcpu ioctl" means theoretically it can be called on multiple vcpus.
> 
> What happens in that case?

Each vcpu can handle the ioctl simaltaneously.  If we assume tdp_mmu, each vcpu
calls the kvm fault handler simultaneously with read spinlock.
If gfn ranges overlap, vcpu will get 0 (success) or EAGAIN.


> > +:Parameters: struct kvm_memory_mapping(in/out)
> > +:Returns: 0 on success, <0 on error
> > +
> > +KVM_MAP_MEMORY populates guest memory without running vcpu.
> > +
> > +::
> > +
> > +  struct kvm_memory_mapping {
> > +	__u64 base_gfn;
> > +	__u64 nr_pages;
> > +	__u64 flags;
> > +	__u64 source;
> > +  };
> > +
> > +  /* For kvm_memory_mapping:: flags */
> > +  #define KVM_MEMORY_MAPPING_FLAG_WRITE         _BITULL(0)
> > +  #define KVM_MEMORY_MAPPING_FLAG_EXEC          _BITULL(1)
> > +  #define KVM_MEMORY_MAPPING_FLAG_USER          _BITULL(2)
> 
> I am not sure what's the good of having "FLAG_USER"?
> 
> This ioctl is called from userspace, thus I think we can just treat this always
> as user-fault?

The point is how to emulate kvm page fault as if vcpu caused the kvm page
fault.  Not we call the ioctl as user context.
-- 
Isaku Yamahata <isaku.yamahata@linux.intel.com>