public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Sean Christopherson <seanjc@google.com>
To: Paolo Bonzini <pbonzini@redhat.com>, Joerg Roedel <joro@8bytes.org>
Cc: Sean Christopherson <seanjc@google.com>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>,
	Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>,
	kvm@vger.kernel.org, iommu@lists.linux-foundation.org,
	linux-kernel@vger.kernel.org,
	Maxim Levitsky <mlevitsk@redhat.com>
Subject: [PATCH v3 21/26] KVM: SVM: Drop AVIC's intermediate avic_set_running() helper
Date: Wed,  8 Dec 2021 01:52:31 +0000	[thread overview]
Message-ID: <20211208015236.1616697-22-seanjc@google.com> (raw)
In-Reply-To: <20211208015236.1616697-1-seanjc@google.com>

Drop avic_set_running() in favor of calling avic_vcpu_{load,put}()
directly, and modify the block+put path to use preempt_disable/enable()
instead of get/put_cpu(), as it doesn't actually care about the current
pCPU associated with the vCPU.  Opportunistically add lockdep assertions
as being preempted in avic_vcpu_put() would lead to consuming stale data,
even though doing so _in the current code base_ would not be fatal.

Add a much needed comment explaining why svm_vcpu_blocking() needs to
unload the AVIC and update the IRTE _before_ the vCPU starts blocking.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 arch/x86/kvm/svm/avic.c | 50 ++++++++++++++++++++++++-----------------
 1 file changed, 30 insertions(+), 20 deletions(-)

diff --git a/arch/x86/kvm/svm/avic.c b/arch/x86/kvm/svm/avic.c
index 88b3c315b34f..dd0d688bc342 100644
--- a/arch/x86/kvm/svm/avic.c
+++ b/arch/x86/kvm/svm/avic.c
@@ -977,6 +977,8 @@ void avic_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
 	int h_physical_id = kvm_cpu_get_apicid(cpu);
 	struct vcpu_svm *svm = to_svm(vcpu);
 
+	lockdep_assert_preemption_disabled();
+
 	/*
 	 * Since the host physical APIC id is 8 bits,
 	 * we can support host APIC ID upto 255.
@@ -1010,6 +1012,8 @@ void avic_vcpu_put(struct kvm_vcpu *vcpu)
 	u64 entry;
 	struct vcpu_svm *svm = to_svm(vcpu);
 
+	lockdep_assert_preemption_disabled();
+
 	entry = READ_ONCE(*(svm->avic_physical_id_cache));
 
 	/* Nothing to do if IsRunning == '0' due to vCPU blocking. */
@@ -1022,31 +1026,37 @@ void avic_vcpu_put(struct kvm_vcpu *vcpu)
 	WRITE_ONCE(*(svm->avic_physical_id_cache), entry);
 }
 
-/*
- * This function is called during VCPU halt/unhalt.
- */
-static void avic_set_running(struct kvm_vcpu *vcpu, bool is_run)
-{
-	int cpu = get_cpu();
-
-	WARN_ON(cpu != vcpu->cpu);
-
-	if (kvm_vcpu_apicv_active(vcpu)) {
-		if (is_run)
-			avic_vcpu_load(vcpu, cpu);
-		else
-			avic_vcpu_put(vcpu);
-	}
-	put_cpu();
-}
-
 void svm_vcpu_blocking(struct kvm_vcpu *vcpu)
 {
-	avic_set_running(vcpu, false);
+	if (!kvm_vcpu_apicv_active(vcpu))
+		return;
+
+	preempt_disable();
+
+	/*
+	 * Unload the AVIC when the vCPU is about to block, _before_ the vCPU
+	 * actually blocks.  The vCPU needs to be marked IsRunning=0 before the
+	 * final pass over the vIRR via kvm_vcpu_check_block().  Any IRQs that
+	 * arrive before IsRunning=0 will not signal the doorbell, i.e. it's
+	 * KVM's responsibility to ensure there are no pending IRQs in the vIRR
+	 * after IsRunning is cleared, prior to scheduling out the vCPU.
+	 */
+	avic_vcpu_put(vcpu);
+
+	preempt_enable();
 }
 
 void svm_vcpu_unblocking(struct kvm_vcpu *vcpu)
 {
+	int cpu;
 
-	avic_set_running(vcpu, true);
+	if (!kvm_vcpu_apicv_active(vcpu))
+		return;
+
+	cpu = get_cpu();
+	WARN_ON(cpu != vcpu->cpu);
+
+	avic_vcpu_load(vcpu, cpu);
+
+	put_cpu();
 }
-- 
2.34.1.400.ga245620fadb-goog


  parent reply	other threads:[~2021-12-08  1:56 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-08  1:52 [PATCH v3 00/26] KVM: x86: Halt and APICv overhaul Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 01/26] KVM: fix avic_set_running for preemptable kernels Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 02/26] KVM: nVMX: Ensure vCPU honors event request if posting nested IRQ fails Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 03/26] KVM: VMX: Clean up PI pre/post-block WARNs Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 04/26] KVM: VMX: Handle PI wakeup shenanigans during vcpu_put/load Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 05/26] KVM: Drop unused kvm_vcpu.pre_pcpu field Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 06/26] KVM: Move x86 VMX's posted interrupt list_head to vcpu_vmx Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 07/26] KVM: VMX: Move preemption timer <=> hrtimer dance to common x86 Sean Christopherson
2023-03-29 12:34   ` Tudor Ambarus
2023-03-29 13:47     ` Paolo Bonzini
2023-03-29 15:22       ` Tudor Ambarus
2023-03-30  7:12         ` Tudor Ambarus
2021-12-08  1:52 ` [PATCH v3 08/26] KVM: x86: Unexport LAPIC's switch_to_{hv,sw}_timer() helpers Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 09/26] KVM: x86: Remove defunct pre_block/post_block kvm_x86_ops hooks Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 10/26] KVM: SVM: Signal AVIC doorbell iff vCPU is in guest mode Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 11/26] KVM: SVM: Don't bother checking for "running" AVIC when kicking for IPIs Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 12/26] KVM: SVM: Remove unnecessary APICv/AVIC update in vCPU unblocking path Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 13/26] KVM: SVM: Use kvm_vcpu_is_blocking() in AVIC load to handle preemption Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 14/26] KVM: SVM: Skip AVIC and IRTE updates when loading blocking vCPU Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 15/26] iommu/amd: KVM: SVM: Use pCPU to infer IsRun state for IRTE Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 16/26] KVM: VMX: Don't do full kick when triggering posted interrupt "fails" Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 17/26] KVM: VMX: Wake vCPU when delivering posted IRQ even if vCPU == this vCPU Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 18/26] KVM: VMX: Pass desired vector instead of bool for triggering posted IRQ Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 19/26] KVM: VMX: Fold fallback path into triggering posted IRQ helper Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 20/26] KVM: VMX: Don't do full kick when handling posted interrupt wakeup Sean Christopherson
2021-12-08  1:52 ` Sean Christopherson [this message]
2021-12-08 14:43   ` [PATCH v3 21/26] KVM: SVM: Drop AVIC's intermediate avic_set_running() helper Paolo Bonzini
2021-12-08 15:03     ` Maxim Levitsky
2021-12-08 15:43     ` Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 22/26] KVM: SVM: Move svm_hardware_setup() and its helpers below svm_x86_ops Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 23/26] KVM: SVM: Nullify vcpu_(un)blocking() hooks if AVIC is disabled Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 24/26] KVM: x86: Skip APICv update if APICv is disable at the module level Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 25/26] KVM: x86: Drop NULL check on kvm_x86_ops.check_apicv_inhibit_reasons Sean Christopherson
2021-12-08  1:52 ` [PATCH v3 26/26] KVM: x86: Unexport __kvm_request_apicv_update() Sean Christopherson
2021-12-08  9:04 ` [PATCH v3 00/26] KVM: x86: Halt and APICv overhaul Paolo Bonzini
2021-12-08 14:51 ` Paolo Bonzini
2021-12-08 23:00 ` Maxim Levitsky
2021-12-08 23:16   ` Maxim Levitsky
2021-12-08 23:34     ` Maxim Levitsky
2021-12-09  0:04       ` Sean Christopherson
2021-12-09  6:36         ` Maxim Levitsky
2021-12-09  0:02     ` Sean Christopherson
2021-12-09 14:29       ` Paolo Bonzini
2021-12-09 14:48         ` Maxim Levitsky
2021-12-09 15:45           ` Sean Christopherson
2021-12-09 16:03             ` Maxim Levitsky
2021-12-09  1:37     ` Sean Christopherson
2021-12-09  6:31       ` Maxim Levitsky
2021-12-08 23:43   ` Sean Christopherson
2021-12-09  6:34     ` Maxim Levitsky

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=20211208015236.1616697-22-seanjc@google.com \
    --to=seanjc@google.com \
    --cc=iommu@lists.linux-foundation.org \
    --cc=jmattson@google.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mlevitsk@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=suravee.suthikulpanit@amd.com \
    --cc=vkuznets@redhat.com \
    --cc=wanpengli@tencent.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