[PATCH v6 2/4] KVM: x86/mmu: use gfn_to_pfn_noref

[David Stevens]

From: David Stevens <stevensd@xxxxxxxxxxxx>

Switch the x86 mmu to the new gfn_to_pfn_noref functions. This allows IO
and PFNMAP mappings backed with valid struct pages but without
refcounting (e.g. tail pages of non-compound higher order allocations)
to be mapped into the guest.

Signed-off-by: David Stevens <stevensd@xxxxxxxxxxxx>
---
 arch/x86/kvm/mmu/mmu.c          | 19 ++++++++++---------
 arch/x86/kvm/mmu/mmu_internal.h |  1 +
 arch/x86/kvm/mmu/paging_tmpl.h  |  7 ++++---
 arch/x86/kvm/x86.c              |  5 +++--
 4 files changed, 18 insertions(+), 14 deletions(-)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 144c5a01cd77..86b74e7bccfa 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -3114,7 +3114,7 @@ void kvm_mmu_hugepage_adjust(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 	if (unlikely(fault->max_level == PG_LEVEL_4K))
 		return;
 
-	if (is_error_noslot_pfn(fault->pfn))
+	if (!fault->page)
 		return;
 
 	if (kvm_slot_dirty_track_enabled(slot))
@@ -4224,6 +4224,7 @@ static int __kvm_faultin_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 		if (is_guest_mode(vcpu)) {
 			fault->slot = NULL;
 			fault->pfn = KVM_PFN_NOSLOT;
+			fault->page = NULL;
 			fault->map_writable = false;
 			return RET_PF_CONTINUE;
 		}
@@ -4239,9 +4240,9 @@ static int __kvm_faultin_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 	}
 
 	async = false;
-	fault->pfn = __gfn_to_pfn_memslot(slot, fault->gfn, false, false, &async,
-					  fault->write, &fault->map_writable,
-					  &fault->hva);
+	fault->pfn = __gfn_to_pfn_noref_memslot(slot, fault->gfn, false, false, &async,
+						fault->write, &fault->map_writable,
+						&fault->hva, &fault->page);
 	if (!async)
 		return RET_PF_CONTINUE; /* *pfn has correct page already */
 
@@ -4261,9 +4262,9 @@ static int __kvm_faultin_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 	 * to wait for IO.  Note, gup always bails if it is unable to quickly
 	 * get a page and a fatal signal, i.e. SIGKILL, is pending.
 	 */
-	fault->pfn = __gfn_to_pfn_memslot(slot, fault->gfn, false, true, NULL,
-					  fault->write, &fault->map_writable,
-					  &fault->hva);
+	fault->pfn = __gfn_to_pfn_noref_memslot(slot, fault->gfn, false, true, NULL,
+						fault->write, &fault->map_writable,
+						&fault->hva, &fault->page);
 	return RET_PF_CONTINUE;
 }
 
@@ -4349,7 +4350,7 @@ static int direct_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 
 out_unlock:
 	write_unlock(&vcpu->kvm->mmu_lock);
-	kvm_release_pfn_clean(fault->pfn);
+	kvm_release_pfn_noref_clean(fault->pfn, fault->page);
 	return r;
 }
 
@@ -4427,7 +4428,7 @@ static int kvm_tdp_mmu_page_fault(struct kvm_vcpu *vcpu,
 
 out_unlock:
 	read_unlock(&vcpu->kvm->mmu_lock);
-	kvm_release_pfn_clean(fault->pfn);
+	kvm_release_pfn_noref_clean(fault->pfn, fault->page);
 	return r;
 }
 #endif
diff --git a/arch/x86/kvm/mmu/mmu_internal.h b/arch/x86/kvm/mmu/mmu_internal.h
index 2cbb155c686c..6ee34a2d0e13 100644
--- a/arch/x86/kvm/mmu/mmu_internal.h
+++ b/arch/x86/kvm/mmu/mmu_internal.h
@@ -239,6 +239,7 @@ struct kvm_page_fault {
 	unsigned long mmu_seq;
 	kvm_pfn_t pfn;
 	hva_t hva;
+	struct page *page;
 	bool map_writable;
 
 	/*
diff --git a/arch/x86/kvm/mmu/paging_tmpl.h b/arch/x86/kvm/mmu/paging_tmpl.h
index a056f2773dd9..e4e54e372721 100644
--- a/arch/x86/kvm/mmu/paging_tmpl.h
+++ b/arch/x86/kvm/mmu/paging_tmpl.h
@@ -525,6 +525,7 @@ FNAME(prefetch_gpte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 	unsigned pte_access;
 	gfn_t gfn;
 	kvm_pfn_t pfn;
+	struct page *page;
 
 	if (FNAME(prefetch_invalid_gpte)(vcpu, sp, spte, gpte))
 		return false;
@@ -540,12 +541,12 @@ FNAME(prefetch_gpte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 	if (!slot)
 		return false;
 
-	pfn = gfn_to_pfn_memslot_atomic(slot, gfn);
+	pfn = gfn_to_pfn_noref_memslot_atomic(slot, gfn, &page);
 	if (is_error_pfn(pfn))
 		return false;
 
 	mmu_set_spte(vcpu, slot, spte, pte_access, gfn, pfn, NULL);
-	kvm_release_pfn_clean(pfn);
+	kvm_release_pfn_noref_clean(pfn, page);
 	return true;
 }
 
@@ -830,7 +831,7 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 
 out_unlock:
 	write_unlock(&vcpu->kvm->mmu_lock);
-	kvm_release_pfn_clean(fault->pfn);
+	kvm_release_pfn_noref_clean(fault->pfn, fault->page);
 	return r;
 }
 
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 237c483b1230..53a8c9e776e5 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8458,6 +8458,7 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
 {
 	gpa_t gpa = cr2_or_gpa;
 	kvm_pfn_t pfn;
+	struct page *page;
 
 	if (!(emulation_type & EMULTYPE_ALLOW_RETRY_PF))
 		return false;
@@ -8487,7 +8488,7 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
 	 * retry instruction -> write #PF -> emulation fail -> retry
 	 * instruction -> ...
 	 */
-	pfn = gfn_to_pfn(vcpu->kvm, gpa_to_gfn(gpa));
+	pfn = gfn_to_pfn_noref(vcpu->kvm, gpa_to_gfn(gpa), &page);
 
 	/*
 	 * If the instruction failed on the error pfn, it can not be fixed,
@@ -8496,7 +8497,7 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
 	if (is_error_noslot_pfn(pfn))
 		return false;
 
-	kvm_release_pfn_clean(pfn);
+	kvm_release_pfn_noref_clean(pfn, page);
 
 	/* The instructions are well-emulated on direct mmu. */
 	if (vcpu->arch.mmu->root_role.direct) {
-- 
2.40.0.348.gf938b09366-goog