diff mbox series

KVM: x86/mmu: Release the pfn in handle_abnormal_pfn() error path

Message ID 20221021111932.3291560-1-tabba@google.com (mailing list archive)
State New, archived
Headers show
Series KVM: x86/mmu: Release the pfn in handle_abnormal_pfn() error path | expand

Commit Message

Fuad Tabba Oct. 21, 2022, 11:19 a.m. UTC 
Currently in case of error it returns without releasing the pfn
faulted in earlier.

Signed-off-by: Fuad Tabba <tabba@google.com>
---

Applies to 6.1-rc1. I think that kvm_release_pfn_clean() has been
replaced with kvm_mmu_release_fault() in development branches,
but the same issue is still there.
---
 arch/x86/kvm/mmu/mmu.c         | 3 ++-
 arch/x86/kvm/mmu/paging_tmpl.h | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)


base-commit: 9abf2313adc1ca1b6180c508c25f22f9395cc780

Comments

Sean Christopherson Oct. 21, 2022, 3:11 p.m. UTC | #1 
On Fri, Oct 21, 2022, Fuad Tabba wrote:
> Currently in case of error it returns without releasing the pfn
> faulted in earlier.

handle_abnormal_pfn() returns something other than RET_PF_CONTINUE if and only if
there was no pfn to fault-in, i.e. it only handles error and no-slot pfns.
diff mbox series

Patch

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 6f81539061d6..9adae837ccdd 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -4250,7 +4250,7 @@  static int direct_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 
 	r = handle_abnormal_pfn(vcpu, fault, ACC_ALL);
 	if (r != RET_PF_CONTINUE)
-		return r;
+		goto out_release;
 
 	r = RET_PF_RETRY;
 
@@ -4276,6 +4276,7 @@  static int direct_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 		read_unlock(&vcpu->kvm->mmu_lock);
 	else
 		write_unlock(&vcpu->kvm->mmu_lock);
+out_release:
 	kvm_release_pfn_clean(fault->pfn);
 	return r;
 }
diff --git a/arch/x86/kvm/mmu/paging_tmpl.h b/arch/x86/kvm/mmu/paging_tmpl.h
index 5ab5f94dcb6f..4a1e4e460990 100644
--- a/arch/x86/kvm/mmu/paging_tmpl.h
+++ b/arch/x86/kvm/mmu/paging_tmpl.h
@@ -847,7 +847,7 @@  static int FNAME(page_fault)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 
 	r = handle_abnormal_pfn(vcpu, fault, walker.pte_access);
 	if (r != RET_PF_CONTINUE)
-		return r;
+		goto out_release;
 
 	/*
 	 * Do not change pte_access if the pfn is a mmio page, otherwise
@@ -881,6 +881,7 @@  static int FNAME(page_fault)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
 
 out_unlock:
 	write_unlock(&vcpu->kvm->mmu_lock);
+out_release:
 	kvm_release_pfn_clean(fault->pfn);
 	return r;
 }