From: Yosry Ahmed <yosry.ahmed@linux.dev>
To: Sean Christopherson <seanjc@google.com>,
Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [RFC PATCH 00/13] Optimize nSVM TLB flushes
Date: Wed, 5 Feb 2025 18:23:49 +0000 [thread overview]
Message-ID: <20250205182402.2147495-1-yosry.ahmed@linux.dev> (raw)
Currently KVM does a TLB flush and an MMU sync on every nested
transition (L1 <-> L2), because it uses the same ASID to run both L1 and
L2.
This series addresses that by giving a separate ASID to L2, adding the
necessary TLB management for it, and properly virtualizing TLB flushes
for L1.
Patch 1 introduces a separate ASID for L2, althoug not properly handled
yet, so it keeps the unconditional flushes.
Patches 2 to 6 are some refactoring and groundwork.
Patches 7 to 12 add the actual TLB management for nSVM, some of which
are items on the TODO list in nested_svm_transition_tlb_flush().
Patch 13 finally stops the unconditional flushes on every nested
transition.
I tested this by booting an L2 and running some basic workloads,
including a CPUID microbenchmark to measure the performance improvement
(numbers in the last patch). I sent the RFC to get feedback on the
general approach, and meanwhile I will try to run more tests that could
exercise TLB flushing.
Yosry Ahmed (13):
KVM: nSVM: Track the ASID per-VMCB
KVM: nSVM: Rework svm_flush_tlb_asid() to operate on a given VMCB
KVM: nSVM: Split nested_svm_transition_tlb_flush() into entry/exit fns
KVM: SVM: Introduce helpers for updating TLB_CONTROL
KVM: x86/mmu: rename __kvm_mmu_invalidate_addr()
KVM: x86/mmu: Allow skipping the gva flush in
kvm_mmu_invalidate_addr()
KVM: nSVM: Handle INVLPGA interception correctly
KVM: nSVM: Flush both L1 and L2 ASIDs on KVM_REQ_TLB_FLUSH
KVM: nSVM: Handle nested TLB flush requests through TLB_CONTROL
KVM: nSVM: Flush the TLB if L1 changes L2's ASID
KVM: nSVM: Do not reset TLB_CONTROL in VMCB02 on nested entry
KVM: nSVM: Service local TLB flushes before nested transitions
KVM: nSVM: Stop bombing the TLB on nested transitions
arch/x86/include/asm/kvm_host.h | 2 +
arch/x86/include/asm/svm.h | 6 ---
arch/x86/kvm/mmu/mmu.c | 22 +++++---
arch/x86/kvm/svm/nested.c | 64 +++++++++++++++-------
arch/x86/kvm/svm/sev.c | 4 +-
arch/x86/kvm/svm/svm.c | 95 ++++++++++++++++++++++++++-------
arch/x86/kvm/svm/svm.h | 33 +++++++++++-
7 files changed, 170 insertions(+), 56 deletions(-)
--
2.48.1.362.g079036d154-goog
next reply other threads:[~2025-02-05 18:24 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-05 18:23 Yosry Ahmed [this message]
2025-02-05 18:23 ` [RFC PATCH 01/13] KVM: nSVM: Track the ASID per-VMCB Yosry Ahmed
2025-03-01 0:03 ` Sean Christopherson
2025-03-03 17:51 ` Jim Mattson
2025-03-03 18:53 ` Sean Christopherson
2025-03-03 19:18 ` Yosry Ahmed
2025-03-01 1:23 ` Maxim Levitsky
2025-03-03 19:31 ` Yosry Ahmed
2025-02-05 18:23 ` [RFC PATCH 02/13] KVM: nSVM: Rework svm_flush_tlb_asid() to operate on a given VMCB Yosry Ahmed
2025-03-01 1:29 ` Maxim Levitsky
2025-03-03 21:58 ` Yosry Ahmed
2025-03-05 2:52 ` Maxim Levitsky
2025-02-05 18:23 ` [RFC PATCH 03/13] KVM: nSVM: Split nested_svm_transition_tlb_flush() into entry/exit fns Yosry Ahmed
2025-03-01 1:34 ` Maxim Levitsky
2025-02-05 18:23 ` [RFC PATCH 04/13] KVM: SVM: Introduce helpers for updating TLB_CONTROL Yosry Ahmed
2025-03-01 1:37 ` Maxim Levitsky
2025-02-05 18:23 ` [RFC PATCH 05/13] KVM: x86/mmu: rename __kvm_mmu_invalidate_addr() Yosry Ahmed
2025-02-05 18:23 ` [RFC PATCH 06/13] KVM: x86/mmu: Allow skipping the gva flush in kvm_mmu_invalidate_addr() Yosry Ahmed
2025-02-05 18:23 ` [RFC PATCH 07/13] KVM: nSVM: Handle INVLPGA interception correctly Yosry Ahmed
2025-03-01 1:55 ` Maxim Levitsky
2025-03-03 22:05 ` Yosry Ahmed
2025-03-05 2:54 ` Maxim Levitsky
2025-03-05 6:20 ` Yosry Ahmed
2025-02-05 18:23 ` [RFC PATCH 08/13] KVM: nSVM: Flush both L1 and L2 ASIDs on KVM_REQ_TLB_FLUSH Yosry Ahmed
2025-03-01 1:58 ` Maxim Levitsky
2025-03-03 22:06 ` Yosry Ahmed
2025-02-05 18:23 ` [RFC PATCH 09/13] KVM: nSVM: Handle nested TLB flush requests through TLB_CONTROL Yosry Ahmed
2025-02-05 21:45 ` Yosry Ahmed
2025-03-01 2:06 ` Maxim Levitsky
[not found] ` <Z8Yovz0I3QLuq6VQ@google.com>
2025-03-05 2:57 ` Maxim Levitsky
2025-02-05 18:23 ` [RFC PATCH 10/13] KVM: nSVM: Flush the TLB if L1 changes L2's ASID Yosry Ahmed
2025-03-01 2:13 ` Maxim Levitsky
2025-02-05 18:24 ` [RFC PATCH 11/13] KVM: nSVM: Do not reset TLB_CONTROL in VMCB02 on nested entry Yosry Ahmed
2025-03-01 2:17 ` Maxim Levitsky
2025-03-03 22:14 ` Yosry Ahmed
2025-03-05 3:01 ` Maxim Levitsky
2025-03-05 6:34 ` Yosry Ahmed
2025-02-05 18:24 ` [RFC PATCH 12/13] KVM: nSVM: Service local TLB flushes before nested transitions Yosry Ahmed
2025-03-01 2:20 ` Maxim Levitsky
2025-03-03 22:18 ` Yosry Ahmed
2025-03-05 3:03 ` Maxim Levitsky
2025-03-05 6:37 ` Yosry Ahmed
2025-02-05 18:24 ` [RFC PATCH 13/13] KVM: nSVM: Stop bombing the TLB on " Yosry Ahmed
2025-03-01 2:21 ` Maxim Levitsky
2025-03-03 22:21 ` Yosry Ahmed
2025-03-05 3:14 ` Maxim Levitsky
2025-03-05 6:45 ` Yosry Ahmed
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=20250205182402.2147495-1-yosry.ahmed@linux.dev \
--to=yosry.ahmed@linux.dev \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=seanjc@google.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®