From: Lai Jiangshan <jiangshan.ljs@antgroup.com>

Although there is no harm, but there is no point to clear write
flooding for direct SP.

Signed-off-by: Lai Jiangshan <jiangshan.ljs@antgroup.com>
---
 arch/x86/kvm/mmu/paging_tmpl.h | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kvm/mmu/paging_tmpl.h b/arch/x86/kvm/mmu/paging_tmpl.h
index 79da423ee4cb..11f17efbec97 100644
--- a/arch/x86/kvm/mmu/paging_tmpl.h
+++ b/arch/x86/kvm/mmu/paging_tmpl.h
@@ -649,12 +649,12 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault,
 	if (WARN_ON(!VALID_PAGE(vcpu->arch.mmu->root.hpa)))
 		goto out_gpte_changed;
 
-	for (shadow_walk_init(&it, vcpu, fault->addr);
-	     shadow_walk_okay(&it) && it.level > gw->level;
-	     shadow_walk_next(&it)) {
+	for_each_shadow_entry(vcpu, fault->addr, it) {
 		gfn_t table_gfn;
 
 		clear_sp_write_flooding_count(it.sptep);
+		if (it.level == gw->level)
+			break;
 
 		table_gfn = gw->table_gfn[it.level - 2];
 		access = gw->pt_access[it.level - 2];
@@ -706,8 +706,6 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault,
 	trace_kvm_mmu_spte_requested(fault);
 
 	for (; shadow_walk_okay(&it); shadow_walk_next(&it)) {
-		clear_sp_write_flooding_count(it.sptep);
-
 		/*
 		 * We cannot overwrite existing page tables with an NX
 		 * large page, as the leaf could be executable.
-- 
2.19.1.6.gb485710b

On Thu, 05 Jan 2023 18:03:10 +0800, Lai Jiangshan wrote:
> Although there is no harm, but there is no point to clear write
> flooding for direct SP.
> 
> 

Applied to kvm-x86 mmu, thanks!

[1/1] kvm: x86/mmu: Don't clear write flooding for direct SP
      https://github.com/kvm-x86/linux/commit/5ee0c3718540

--
https://github.com/kvm-x86/linux/tree/next
https://github.com/kvm-x86/linux/tree/fixes