[PATCH v4 6/7] KVM: arm64: Relax locking for kvm_test_age_gfn and kvm_age_gfn

James Houghton jthoughton at google.com
Wed May 29 11:05:09 PDT 2024


Replace the MMU write locks for read locks.

Grabbing the read lock instead of the write lock is safe because the
only requirement we have is that the stage-2 page tables do not get
deallocated while we are walking them. The stage2_age_walker() callback
is safe to race with itself; update the comment to reflect the
synchronization change.

Signed-off-by: James Houghton <jthoughton at google.com>
---
 arch/arm64/kvm/hyp/pgtable.c | 9 ++++-----
 arch/arm64/kvm/mmu.c         | 8 ++++----
 2 files changed, 8 insertions(+), 9 deletions(-)

diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
index 9e2bbee77491..eabb07c66a07 100644
--- a/arch/arm64/kvm/hyp/pgtable.c
+++ b/arch/arm64/kvm/hyp/pgtable.c
@@ -1319,10 +1319,8 @@ static int stage2_age_walker(const struct kvm_pgtable_visit_ctx *ctx,
 	data->young = true;
 
 	/*
-	 * stage2_age_walker() is always called while holding the MMU lock for
-	 * write, so this will always succeed. Nonetheless, this deliberately
-	 * follows the race detection pattern of the other stage-2 walkers in
-	 * case the locking mechanics of the MMU notifiers is ever changed.
+	 * This walk may not be exclusive; the PTE is permitted to change
+	 * from under us.
 	 */
 	if (data->mkold && !stage2_try_set_pte(ctx, new))
 		return -EAGAIN;
@@ -1345,7 +1343,8 @@ bool kvm_pgtable_stage2_test_clear_young(struct kvm_pgtable *pgt, u64 addr,
 	struct kvm_pgtable_walker walker = {
 		.cb		= stage2_age_walker,
 		.arg		= &data,
-		.flags		= KVM_PGTABLE_WALK_LEAF,
+		.flags		= KVM_PGTABLE_WALK_LEAF |
+				  KVM_PGTABLE_WALK_SHARED,
 	};
 
 	WARN_ON(kvm_pgtable_walk(pgt, addr, size, &walker));
diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c
index 8337009dde77..40e7427462a7 100644
--- a/arch/arm64/kvm/mmu.c
+++ b/arch/arm64/kvm/mmu.c
@@ -1775,7 +1775,7 @@ bool kvm_age_gfn(struct kvm *kvm, struct kvm_gfn_range *range)
 	u64 size = (range->end - range->start) << PAGE_SHIFT;
 	bool young = false;
 
-	write_lock(&kvm->mmu_lock);
+	read_lock(&kvm->mmu_lock);
 
 	if (!kvm->arch.mmu.pgt)
 		goto out;
@@ -1785,7 +1785,7 @@ bool kvm_age_gfn(struct kvm *kvm, struct kvm_gfn_range *range)
 						    size, true);
 
 out:
-	write_unlock(&kvm->mmu_lock);
+	read_unlock(&kvm->mmu_lock);
 	return young;
 }
 
@@ -1794,7 +1794,7 @@ bool kvm_test_age_gfn(struct kvm *kvm, struct kvm_gfn_range *range)
 	u64 size = (range->end - range->start) << PAGE_SHIFT;
 	bool young = false;
 
-	write_lock(&kvm->mmu_lock);
+	read_lock(&kvm->mmu_lock);
 
 	if (!kvm->arch.mmu.pgt)
 		goto out;
@@ -1804,7 +1804,7 @@ bool kvm_test_age_gfn(struct kvm *kvm, struct kvm_gfn_range *range)
 						    size, false);
 
 out:
-	write_unlock(&kvm->mmu_lock);
+	read_unlock(&kvm->mmu_lock);
 	return young;
 }
 
-- 
2.45.1.288.g0e0cd299f1-goog




More information about the kvm-riscv mailing list