linux-kernel - [PATCH 1/2] KVM: x86/mmu: Don't check old SPTE permissions when trying to unsync

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20260123090304.32286-1-jiangshanlai@gmail.com>
Date: Fri, 23 Jan 2026 17:03:02 +0800
From: Lai Jiangshan <jiangshanlai@...il.com>
To: linux-kernel@...r.kernel.org
Cc: Lai Jiangshan <jiangshan.ljs@...group.com>,
	Sean Christopherson <seanjc@...gle.com>,
	Paolo Bonzini <pbonzini@...hat.com>,
	Thomas Gleixner <tglx@...nel.org>,
	Ingo Molnar <mingo@...hat.com>,
	Borislav Petkov <bp@...en8.de>,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	x86@...nel.org,
	"H. Peter Anvin" <hpa@...or.com>,
	kvm@...r.kernel.org
Subject: [PATCH 1/2] KVM: x86/mmu: Don't check old SPTE permissions when trying to unsync

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

Commit ecc5589f19a5 ("KVM: MMU: optimize set_spte for page sync") added
a writable permission check on the old SPTE to avoid unnecessary calls
to mmu_try_to_unsync_pages() when syncing SPTEs.

Later, commit e6722d9211b2 ("KVM: x86/mmu: Reduce the update to the spte
in FNAME(sync_spte)") indirectly achieves it by avoiding some SPTE
updates altogether, which makes the writable permission check in
make_spte() much less useful.

Remove the old-SPTE writable permission check from make_spte() to
simplify the code.

This may cause mmu_try_to_unsync_pages() to be called in a few
additional cases not covered by commit e6722d9211b2, such as when the
guest toggles the execute bit, which is expected to be rare.

Signed-off-by: Lai Jiangshan <jiangshan.ljs@...group.com>
---
 arch/x86/kvm/mmu/mmu.c         |  2 +-
 arch/x86/kvm/mmu/paging_tmpl.h |  2 +-
 arch/x86/kvm/mmu/spte.c        | 12 ++----------
 arch/x86/kvm/mmu/spte.h        |  2 +-
 arch/x86/kvm/mmu/tdp_mmu.c     |  2 +-
 5 files changed, 6 insertions(+), 14 deletions(-)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 02c450686b4a..4535d2836004 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -3073,7 +3073,7 @@ static int mmu_set_spte(struct kvm_vcpu *vcpu, struct kvm_memory_slot *slot,
 			was_rmapped = 1;
 	}
 
-	wrprot = make_spte(vcpu, sp, slot, pte_access, gfn, pfn, *sptep, prefetch,
+	wrprot = make_spte(vcpu, sp, slot, pte_access, gfn, pfn, prefetch,
 			   false, host_writable, &spte);
 
 	if (*sptep == spte) {
diff --git a/arch/x86/kvm/mmu/paging_tmpl.h b/arch/x86/kvm/mmu/paging_tmpl.h
index 901cd2bd40b8..95fccee63563 100644
--- a/arch/x86/kvm/mmu/paging_tmpl.h
+++ b/arch/x86/kvm/mmu/paging_tmpl.h
@@ -954,7 +954,7 @@ static int FNAME(sync_spte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp, int
 	host_writable = spte & shadow_host_writable_mask;
 	slot = kvm_vcpu_gfn_to_memslot(vcpu, gfn);
 	make_spte(vcpu, sp, slot, pte_access, gfn,
-		  spte_to_pfn(spte), spte, true, true,
+		  spte_to_pfn(spte), true, true,
 		  host_writable, &spte);
 
 	/*
diff --git a/arch/x86/kvm/mmu/spte.c b/arch/x86/kvm/mmu/spte.c
index 85a0473809b0..a8e2606ccd22 100644
--- a/arch/x86/kvm/mmu/spte.c
+++ b/arch/x86/kvm/mmu/spte.c
@@ -186,7 +186,7 @@ bool spte_needs_atomic_update(u64 spte)
 bool make_spte(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 	       const struct kvm_memory_slot *slot,
 	       unsigned int pte_access, gfn_t gfn, kvm_pfn_t pfn,
-	       u64 old_spte, bool prefetch, bool synchronizing,
+	       bool prefetch, bool synchronizing,
 	       bool host_writable, u64 *new_spte)
 {
 	int level = sp->role.level;
@@ -258,16 +258,8 @@ bool make_spte(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 		 * SPTE.  Write-protect the SPTE if the page can't be unsync'd,
 		 * e.g. it's write-tracked (upper-level SPs) or has one or more
 		 * shadow pages and unsync'ing pages is not allowed.
-		 *
-		 * When overwriting an existing leaf SPTE, and the old SPTE was
-		 * writable, skip trying to unsync shadow pages as any relevant
-		 * shadow pages must already be unsync, i.e. the hash lookup is
-		 * unnecessary (and expensive).  Note, this relies on KVM not
-		 * changing PFNs without first zapping the old SPTE, which is
-		 * guaranteed by both the shadow MMU and the TDP MMU.
 		 */
-		if ((!is_last_spte(old_spte, level) || !is_writable_pte(old_spte)) &&
-		    mmu_try_to_unsync_pages(vcpu->kvm, slot, gfn, synchronizing, prefetch))
+		if (mmu_try_to_unsync_pages(vcpu->kvm, slot, gfn, synchronizing, prefetch))
 			wrprot = true;
 		else
 			spte |= PT_WRITABLE_MASK | shadow_mmu_writable_mask |
diff --git a/arch/x86/kvm/mmu/spte.h b/arch/x86/kvm/mmu/spte.h
index 91ce29fd6f1b..cf9cd27bcd4f 100644
--- a/arch/x86/kvm/mmu/spte.h
+++ b/arch/x86/kvm/mmu/spte.h
@@ -543,7 +543,7 @@ bool spte_needs_atomic_update(u64 spte);
 bool make_spte(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 	       const struct kvm_memory_slot *slot,
 	       unsigned int pte_access, gfn_t gfn, kvm_pfn_t pfn,
-	       u64 old_spte, bool prefetch, bool synchronizing,
+	       bool prefetch, bool synchronizing,
 	       bool host_writable, u64 *new_spte);
 u64 make_small_spte(struct kvm *kvm, u64 huge_spte,
 		    union kvm_mmu_page_role role, int index);
diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
index 9c26038f6b77..8dfaab2a4fd9 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.c
+++ b/arch/x86/kvm/mmu/tdp_mmu.c
@@ -1188,7 +1188,7 @@ static int tdp_mmu_map_handle_target_level(struct kvm_vcpu *vcpu,
 		new_spte = make_mmio_spte(vcpu, iter->gfn, ACC_ALL);
 	else
 		wrprot = make_spte(vcpu, sp, fault->slot, ACC_ALL, iter->gfn,
-				   fault->pfn, iter->old_spte, fault->prefetch,
+				   fault->pfn, fault->prefetch,
 				   false, fault->map_writable, &new_spte);
 
 	if (new_spte == iter->old_spte)
-- 
2.19.1.6.gb485710b