[RFC PATCH v2 2/3] KVM: arm64: Introduce a dedicated walker for stage2 write-protect

[Leonardo Bras]

When write-protecting a range, clean the writable bit (S2AP[1]) and keep
the DBM bit set for pages, so on HDBSS-enabled systems we can skip the
fault for making it writable again.

On block entries clean both the writable bit and the DBM bit, so we fault
even on HDBSS-enabled systems, in order to have hugepage splitting.

With disabled HDBSS, there should be no change in faulting behavior.

Signed-off-by: Leonardo Bras <leo.bras at arm.com>
---
 arch/arm64/kvm/hyp/pgtable.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
index 8c2320c6a7f1..837546be5c13 100644
--- a/arch/arm64/kvm/hyp/pgtable.c
+++ b/arch/arm64/kvm/hyp/pgtable.c
@@ -1283,17 +1283,34 @@ static int stage2_update_leaf_attrs(struct kvm_pgtable *pgt, u64 addr,
 	if (level)
 		*level = data.level;
 	return 0;
 }
 
+static int stage2_wrprotect_walker(const struct kvm_pgtable_visit_ctx *ctx,
+				   enum kvm_pgtable_walk_flags visit)
+{
+	kvm_pte_t new = ctx->old & ~KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W;
+
+	/* We remove DBM on blocks so they can fault and get split */
+	if (ctx->level < KVM_PGTABLE_LAST_LEVEL)
+		new &= ~KVM_PTE_LEAF_ATTR_HI_S2_DBM;
+
+	if (kvm_pte_valid(ctx->old) && ctx->old != new)
+		WRITE_ONCE(*ctx->ptep, new);
+
+	return 0;
+}
+
+
 int kvm_pgtable_stage2_wrprotect(struct kvm_pgtable *pgt, u64 addr, u64 size)
 {
-	return stage2_update_leaf_attrs(pgt, addr, size, 0,
-					KVM_PTE_LEAF_ATTR_HI_S2_DBM |
-					KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W,
-					NULL, NULL,
-					KVM_PGTABLE_WALK_IGNORE_EAGAIN);
+	struct kvm_pgtable_walker walker = {
+		.cb	= stage2_wrprotect_walker,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+	};
+
+	return kvm_pgtable_walk(pgt, addr, size, &walker);
 }
 
 void kvm_pgtable_stage2_mkyoung(struct kvm_pgtable *pgt, u64 addr,
 				enum kvm_pgtable_walk_flags flags)
 {
-- 
2.56.0