Re: [PATCH v2 05/10] KVM: x86/mmu: Introduce kvm_tdp_map_page() to populate guest memory

[Isaku Yamahata]

On Wed, Apr 17, 2024 at 03:04:08PM +0800,
Chao Gao <chao.gao@xxxxxxxxx> wrote:

> On Wed, Apr 10, 2024 at 03:07:31PM -0700, isaku.yamahata@xxxxxxxxx wrote:
> >From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> >
> >Introduce a helper function to call the KVM fault handler.  It allows a new
> >ioctl to invoke the KVM fault handler to populate without seeing RET_PF_*
> >enums or other KVM MMU internal definitions because RET_PF_* are internal
> >to x86 KVM MMU.  The implementation is restricted to two-dimensional paging
> >for simplicity.  The shadow paging uses GVA for faulting instead of L1 GPA.
> >It makes the API difficult to use.
> >
> >Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> >---
> >v2:
> >- Make the helper function two-dimensional paging specific. (David)
> >- Return error when vcpu is in guest mode. (David)
> >- Rename goal_level to level in kvm_tdp_mmu_map_page(). (Sean)
> >- Update return code conversion. Don't check pfn.
> >  RET_PF_EMULATE => EINVAL, RET_PF_CONTINUE => EIO (Sean)
> >- Add WARN_ON_ONCE on RET_PF_CONTINUE and RET_PF_INVALID. (Sean)
> >- Drop unnecessary EXPORT_SYMBOL_GPL(). (Sean)
> >---
> > arch/x86/kvm/mmu.h     |  3 +++
> > arch/x86/kvm/mmu/mmu.c | 32 ++++++++++++++++++++++++++++++++
> > 2 files changed, 35 insertions(+)
> >
> >diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
> >index e8b620a85627..51ff4f67e115 100644
> >--- a/arch/x86/kvm/mmu.h
> >+++ b/arch/x86/kvm/mmu.h
> >@@ -183,6 +183,9 @@ static inline void kvm_mmu_refresh_passthrough_bits(struct kvm_vcpu *vcpu,
> > 	__kvm_mmu_refresh_passthrough_bits(vcpu, mmu);
> > }
> > 
> >+int kvm_tdp_map_page(struct kvm_vcpu *vcpu, gpa_t gpa, u64 error_code,
> >+		     u8 *level);
> >+
> > /*
> >  * Check if a given access (described through the I/D, W/R and U/S bits of a
> >  * page fault error code pfec) causes a permission fault with the given PTE
> >diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> >index 91dd4c44b7d8..a34f4af44cbd 100644
> >--- a/arch/x86/kvm/mmu/mmu.c
> >+++ b/arch/x86/kvm/mmu/mmu.c
> >@@ -4687,6 +4687,38 @@ int kvm_tdp_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault)
> > 	return direct_page_fault(vcpu, fault);
> > }
> > 
> >+int kvm_tdp_map_page(struct kvm_vcpu *vcpu, gpa_t gpa, u64 error_code,
> >+		     u8 *level)
> >+{
> >+	int r;
> >+
> >+	/* Restrict to TDP page fault. */
> 
> need to explain why. (just as you do in the changelog)

Sure.


> >+	if (vcpu->arch.mmu->page_fault != kvm_tdp_page_fault)
> 
> page fault handlers (i.e., vcpu->arch.mmu->page_fault()) will be called
> finally. why not let page fault handlers reject the request to get rid of
> this ad-hoc check? We just need to plumb a flag indicating this is a
> pre-population request into the handlers. I think this way is clearer.
> 
> What do you think?

__kvm_mmu_do_page_fault() doesn't check if the mmu mode is TDP or not.
If we don't want to check page_fault handler, the alternative check would
be if (!vcpu->arch.mmu->direct).  Or we will require the caller to guarantee
that MMU mode is tdp (direct or tdp_mmu).
-- 
Isaku Yamahata <isaku.yamahata@xxxxxxxxx>