Skip to contentSkip to conversation

KVM: x86/mmu: Refactor slot null check in kvm_mmu_hugepage_adjust

PATCH In review David Matlack 3 participants 3 messages Aug 24, 2021 – Sep 23, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 KVM: x86/mmu: Refactor slot null check in kvm_mmu_hugepage_adjust

The current code is correct but relies on is_error_noslot_pfn() to ensure slot is not null. The only reason is_error_noslot_pfn() was checked instead is because we did not have the slot before commit 6574422f913e ("KVM: x86/mmu: Pass the memslot around via struct kvm_page_fault") and looking up the memslot is expensive.

Now that the slot is available, explicitly check if it's null and get rid of the redundant is_error_noslot_pfn() check.

---
 arch/x86/kvm/mmu/mmu.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/kvm/mmu/mmu.c +2 -2
index 4853c033e6ce..9b5424bcb173 100644
@@ -2925,10 +2925,10 @@ 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) || kvm_is_reserved_pfn(fault->pfn))
+	if (!slot || kvm_slot_dirty_track_enabled(slot))
 		return;
 
-	if (kvm_slot_dirty_track_enabled(slot))
+	if (kvm_is_reserved_pfn(fault->pfn))
 		return;
 
 	/*
Signature · 1 line
2.33.0.rc2.250.ged5fa647cd-goog
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 6 lines · David Matlack

The current code is correct but relies on is_error_noslot_pfn() to ensure slot is not null. The only reason is_error_noslot_pfn() was checked instead is because we did not have the slot before commit 6574422f913e ("KVM: x86/mmu: Pass the memslot around via struct kvm_page_fault") and looking up the memslot is expensive.

Quoted 6 lines

Now that the slot is available, explicitly check if it's null and get rid of the redundant is_error_noslot_pfn() check.

---

maintainer
#2 lore raw
In reply to David Matlack
Quoted 23 lines · David Matlack

The current code is correct but relies on is_error_noslot_pfn() to ensure slot is not null. The only reason is_error_noslot_pfn() was checked instead is because we did not have the slot before commit 6574422f913e ("KVM: x86/mmu: Pass the memslot around via struct kvm_page_fault") and looking up the memslot is expensive.

Now that the slot is available, explicitly check if it's null and get rid of the redundant is_error_noslot_pfn() check.

---
  arch/x86/kvm/mmu/mmu.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/kvm/mmu/mmu.c +0 -0
index 4853c033e6ce..9b5424bcb173 100644
@@ -2925,10 +2925,10 @@ void kvm_mmu_hugepage_adjust(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
  	if (unlikely(fault->max_level == PG_LEVEL_4K))
  		return;

arch/x86/kvm/mmu/mmu.c void kvm_mmu_hugepage_adjust(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault · line 2934

Quoted 10 lines
arch/x86/kvm/mmu/mmu.c +2 -2
 
-	if (is_error_noslot_pfn(fault->pfn) || kvm_is_reserved_pfn(fault->pfn))
+	if (!slot || kvm_slot_dirty_track_enabled(slot))
  		return;
 
-	if (kvm_slot_dirty_track_enabled(slot))
+	if (kvm_is_reserved_pfn(fault->pfn))
  		return;
 
  	/*

Squashed, thanks.

Paolo