mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Cc: yan.y.zhao@intel.com, isaku.yamahata@intel.com,
	binbin.wu@linux.intel.com, rick.p.edgecombe@intel.com
Subject: [PATCH 01/18] KVM: x86/mmu: Zap invalid roots with mmu_lock holding for write at uninit
Date: Fri, 13 Dec 2024 14:56:54 -0500	[thread overview]
Message-ID: <20241213195711.316050-2-pbonzini@redhat.com> (raw)
In-Reply-To: <20241213195711.316050-1-pbonzini@redhat.com>

From: Rick Edgecombe <rick.p.edgecombe@intel.com>

Add a bool parameter to kvm_tdp_mmu_zap_invalidated_roots() to specify
zapping invalid roots under mmu_lock held for read or write. Hold mmu_lock
for write when kvm_tdp_mmu_zap_invalidated_roots() is called by
kvm_mmu_uninit_tdp_mmu().

kvm_mmu_uninit_tdp_mmu() is invoked either before or after executing any
atomic operations on SPTEs by vCPU threads. Therefore, it will not impact
vCPU threads performance if kvm_tdp_mmu_zap_invalidated_roots() acquires
mmu_lock for write to zap invalid roots.

This is a preparation for future TDX patch which asserts that "Users of
atomic zapping don't operate on mirror roots".

Co-developed-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-2-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kvm/mmu/mmu.c     |  2 +-
 arch/x86/kvm/mmu/tdp_mmu.c | 16 +++++++++++-----
 arch/x86/kvm/mmu/tdp_mmu.h |  2 +-
 3 files changed, 13 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 22e7ad235123..98b94905e231 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -6479,7 +6479,7 @@ static void kvm_mmu_zap_all_fast(struct kvm *kvm)
 	 * lead to use-after-free.
 	 */
 	if (tdp_mmu_enabled)
-		kvm_tdp_mmu_zap_invalidated_roots(kvm);
+		kvm_tdp_mmu_zap_invalidated_roots(kvm, true);
 }
 
 void kvm_mmu_init_vm(struct kvm *kvm)
diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
index 4508d868f1cd..b0bddcb2f781 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.c
+++ b/arch/x86/kvm/mmu/tdp_mmu.c
@@ -38,7 +38,7 @@ void kvm_mmu_uninit_tdp_mmu(struct kvm *kvm)
 	 * ultimately frees all roots.
 	 */
 	kvm_tdp_mmu_invalidate_all_roots(kvm);
-	kvm_tdp_mmu_zap_invalidated_roots(kvm);
+	kvm_tdp_mmu_zap_invalidated_roots(kvm, false);
 
 	WARN_ON(atomic64_read(&kvm->arch.tdp_mmu_pages));
 	WARN_ON(!list_empty(&kvm->arch.tdp_mmu_roots));
@@ -883,11 +883,14 @@ void kvm_tdp_mmu_zap_all(struct kvm *kvm)
  * Zap all invalidated roots to ensure all SPTEs are dropped before the "fast
  * zap" completes.
  */
-void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm)
+void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm, bool shared)
 {
 	struct kvm_mmu_page *root;
 
-	read_lock(&kvm->mmu_lock);
+	if (shared)
+		read_lock(&kvm->mmu_lock);
+	else
+		write_lock(&kvm->mmu_lock);
 
 	for_each_tdp_mmu_root_yield_safe(kvm, root) {
 		if (!root->tdp_mmu_scheduled_root_to_zap)
@@ -905,7 +908,7 @@ void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm)
 		 * that may be zapped, as such entries are associated with the
 		 * ASID on both VMX and SVM.
 		 */
-		tdp_mmu_zap_root(kvm, root, true);
+		tdp_mmu_zap_root(kvm, root, shared);
 
 		/*
 		 * The referenced needs to be put *after* zapping the root, as
@@ -915,7 +918,10 @@ void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm)
 		kvm_tdp_mmu_put_root(kvm, root);
 	}
 
-	read_unlock(&kvm->mmu_lock);
+	if (shared)
+		read_unlock(&kvm->mmu_lock);
+	else
+		write_unlock(&kvm->mmu_lock);
 }
 
 /*
diff --git a/arch/x86/kvm/mmu/tdp_mmu.h b/arch/x86/kvm/mmu/tdp_mmu.h
index f03ca0dd13d9..6d7cdc462f58 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.h
+++ b/arch/x86/kvm/mmu/tdp_mmu.h
@@ -23,7 +23,7 @@ bool kvm_tdp_mmu_zap_leafs(struct kvm *kvm, gfn_t start, gfn_t end, bool flush);
 bool kvm_tdp_mmu_zap_sp(struct kvm *kvm, struct kvm_mmu_page *sp);
 void kvm_tdp_mmu_zap_all(struct kvm *kvm);
 void kvm_tdp_mmu_invalidate_all_roots(struct kvm *kvm);
-void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm);
+void kvm_tdp_mmu_zap_invalidated_roots(struct kvm *kvm, bool shared);
 
 int kvm_tdp_mmu_map(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault);
 
-- 
2.43.5



  reply	other threads:[~2024-12-13 19:57 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-13 19:56 [PATCH v5 00/18] TDX MMU prep series part 1 Paolo Bonzini
2024-12-13 19:56 ` Paolo Bonzini [this message]
2024-12-13 19:56 ` [PATCH 02/18] KVM: Add member to struct kvm_gfn_range for target alias Paolo Bonzini
2024-12-13 19:56 ` [PATCH 03/18] KVM: x86: Add a VM type define for TDX Paolo Bonzini
2024-12-13 19:56 ` [PATCH 04/18] KVM: x86/mmu: Add an external pointer to struct kvm_mmu_page Paolo Bonzini
2024-12-16 12:31   ` Huang, Kai
2024-12-13 19:56 ` [PATCH 05/18] KVM: x86/mmu: Add an is_mirror member for union kvm_mmu_page_role Paolo Bonzini
2024-12-13 19:56 ` [PATCH 06/18] KVM: x86/mmu: Make kvm_tdp_mmu_alloc_root() return void Paolo Bonzini
2024-12-13 19:57 ` [PATCH 07/18] KVM: x86/tdp_mmu: Take struct kvm in iter loops Paolo Bonzini
2024-12-13 19:57 ` [PATCH 08/18] KVM: x86/mmu: Support GFN direct bits Paolo Bonzini
2024-12-13 19:57 ` [PATCH 09/18] KVM: x86/tdp_mmu: Extract root invalid check from tdx_mmu_next_root() Paolo Bonzini
2024-12-16 12:35   ` Huang, Kai
2024-12-13 19:57 ` [PATCH 10/18] KVM: x86/tdp_mmu: Introduce KVM MMU root types to specify page table type Paolo Bonzini
2024-12-13 19:57 ` [PATCH 11/18] KVM: x86/tdp_mmu: Take root in tdp_mmu_for_each_pte() Paolo Bonzini
2024-12-13 19:57 ` [PATCH 12/18] KVM: x86/tdp_mmu: Support mirror root for TDP MMU Paolo Bonzini
2024-12-13 19:57 ` [PATCH 13/18] KVM: x86/tdp_mmu: Propagate attr_filter to MMU notifier callbacks Paolo Bonzini
2024-12-17 12:53   ` Yan Zhao
2024-12-13 19:57 ` [PATCH 14/18] KVM: x86/tdp_mmu: Propagate building mirror page tables Paolo Bonzini
2024-12-13 19:57 ` [PATCH 15/18] KVM: x86/tdp_mmu: Propagate tearing down " Paolo Bonzini
2024-12-13 19:57 ` [PATCH 16/18] KVM: x86/tdp_mmu: Take root types for kvm_tdp_mmu_invalidate_all_roots() Paolo Bonzini
2024-12-13 19:57 ` [PATCH 17/18] KVM: x86/tdp_mmu: Don't zap valid mirror roots in kvm_tdp_mmu_zap_all() Paolo Bonzini
2024-12-18  2:49   ` Yan Zhao
2024-12-13 19:57 ` [PATCH 18/18] KVM: x86/mmu: Prevent aliased memslot GFNs Paolo Bonzini
2024-12-18  0:34 ` [PATCH v5 00/18] TDX MMU prep series part 1 Edgecombe, Rick P
2024-12-20  9:32   ` Yan Zhao

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20241213195711.316050-2-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=binbin.wu@linux.intel.com \
    --cc=isaku.yamahata@intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rick.p.edgecombe@intel.com \
    --cc=yan.y.zhao@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®