LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Vitaly Kuznetsov <vkuznets@redhat.com>
To: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>
Cc: Sean Christopherson <seanjc@google.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Nitesh Narayan Lal <nitesh@redhat.com>,
	linux-kernel@vger.kernel.org
Subject: [PATCH v2 3/4] KVM: Optimize kvm_make_vcpus_request_mask() a bit
Date: Mon, 23 Aug 2021 16:30:27 +0200	[thread overview]
Message-ID: <20210823143028.649818-4-vkuznets@redhat.com> (raw)
In-Reply-To: <20210823143028.649818-1-vkuznets@redhat.com>

Iterating over set bits in 'vcpu_bitmap' should be faster than going
through all vCPUs, especially when just a few bits are set.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 virt/kvm/kvm_main.c | 83 ++++++++++++++++++++++++++-------------------
 1 file changed, 49 insertions(+), 34 deletions(-)

diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 82c5280dd5ce..e9c2ce2d273f 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -257,49 +257,64 @@ static inline bool kvm_kick_many_cpus(const struct cpumask *cpus, bool wait)
 	return true;
 }
 
+static void kvm_make_vcpu_request(struct kvm *kvm, struct kvm_vcpu *vcpu,
+				  unsigned int req, cpumask_var_t tmp,
+				  int current_cpu)
+{
+	int cpu = vcpu->cpu;
+
+	kvm_make_request(req, vcpu);
+
+	if (!(req & KVM_REQUEST_NO_WAKEUP) && kvm_vcpu_wake_up(vcpu))
+		return;
+
+	/*
+	 * tmp can be NULL if cpumasks are allocated off stack, as allocation of
+	 * the mask is deliberately not fatal and is handled by falling back to
+	 * kicking all online CPUs.
+	 */
+	if (IS_ENABLED(CONFIG_CPUMASK_OFFSTACK) && !tmp)
+		return;
+
+	/*
+	 * Note, the vCPU could get migrated to a different pCPU at any point
+	 * after kvm_request_needs_ipi(), which could result in sending an IPI
+	 * to the previous pCPU.  But, that's ok because the purpose of the IPI
+	 * is to ensure the vCPU returns to OUTSIDE_GUEST_MODE, which is
+	 * satisfied if the vCPU migrates. Entering READING_SHADOW_PAGE_TABLES
+	 * after this point is also ok, as the requirement is only that KVM wait
+	 * for vCPUs that were reading SPTEs _before_ any changes were
+	 * finalized. See kvm_vcpu_kick() for more details on handling requests.
+	 */
+	if (kvm_request_needs_ipi(vcpu, req)) {
+		cpu = READ_ONCE(vcpu->cpu);
+		if (cpu != -1 && cpu != current_cpu)
+			__cpumask_set_cpu(cpu, tmp);
+	}
+}
+
 bool kvm_make_vcpus_request_mask(struct kvm *kvm, unsigned int req,
 				 struct kvm_vcpu *except,
 				 unsigned long *vcpu_bitmap, cpumask_var_t tmp)
 {
-	int i, cpu, me;
+	int i, me;
 	struct kvm_vcpu *vcpu;
 	bool called;
 
 	me = get_cpu();
 
-	kvm_for_each_vcpu(i, vcpu, kvm) {
-		if ((vcpu_bitmap && !test_bit(i, vcpu_bitmap)) ||
-		    vcpu == except)
-			continue;
-
-		kvm_make_request(req, vcpu);
-
-		if (!(req & KVM_REQUEST_NO_WAKEUP) && kvm_vcpu_wake_up(vcpu))
-			continue;
-
-		/*
-		 * tmp can be NULL if cpumasks are allocated off stack, as
-		 * allocation of the mask is deliberately not fatal and is
-		 * handled by falling back to kicking all online CPUs.
-		 */
-		if (IS_ENABLED(CONFIG_CPUMASK_OFFSTACK) && !tmp)
-			continue;
-
-		/*
-		 * Note, the vCPU could get migrated to a different pCPU at any
-		 * point after kvm_request_needs_ipi(), which could result in
-		 * sending an IPI to the previous pCPU.  But, that's ok because
-		 * the purpose of the IPI is to ensure the vCPU returns to
-		 * OUTSIDE_GUEST_MODE, which is satisfied if the vCPU migrates.
-		 * Entering READING_SHADOW_PAGE_TABLES after this point is also
-		 * ok, as the requirement is only that KVM wait for vCPUs that
-		 * were reading SPTEs _before_ any changes were finalized.  See
-		 * kvm_vcpu_kick() for more details on handling requests.
-		 */
-		if (kvm_request_needs_ipi(vcpu, req)) {
-			cpu = READ_ONCE(vcpu->cpu);
-			if (cpu != -1 && cpu != me)
-				__cpumask_set_cpu(cpu, tmp);
+	if (vcpu_bitmap) {
+		for_each_set_bit(i, vcpu_bitmap, KVM_MAX_VCPUS) {
+			vcpu = kvm_get_vcpu(kvm, i);
+			if (!vcpu || vcpu == except)
+				continue;
+			kvm_make_vcpu_request(kvm, vcpu, req, tmp, me);
+		}
+	} else {
+		kvm_for_each_vcpu(i, vcpu, kvm) {
+			if (vcpu == except)
+				continue;
+			kvm_make_vcpu_request(kvm, vcpu, req, tmp, me);
 		}
 	}
 
-- 
2.31.1


  parent reply	other threads:[~2021-08-23 14:30 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-23 14:30 [PATCH v2 0/4] KVM: Various fixes and improvements around kicking vCPUs Vitaly Kuznetsov
2021-08-23 14:30 ` [PATCH v2 1/4] KVM: Clean up benign vcpu->cpu data races when " Vitaly Kuznetsov
2021-08-23 14:30 ` [PATCH v2 2/4] KVM: Guard cpusmask NULL check with CONFIG_CPUMASK_OFFSTACK Vitaly Kuznetsov
2021-08-23 14:30 ` Vitaly Kuznetsov [this message]
2021-08-23 14:30 ` [PATCH v2 4/4] KVM: x86: Fix stack-out-of-bounds memory access from ioapic_write_indirect() Vitaly Kuznetsov
2021-08-23 18:58   ` Eduardo Habkost
2021-08-24  7:13     ` Vitaly Kuznetsov
2021-08-24 14:23       ` Eduardo Habkost
2021-08-24 14:42         ` Vitaly Kuznetsov
2021-08-24 16:07           ` Maxim Levitsky
2021-08-24 17:40             ` Sean Christopherson
2021-08-25  8:26               ` Vitaly Kuznetsov
2021-08-25  8:21             ` Vitaly Kuznetsov
2021-08-25  9:11               ` Maxim Levitsky
2021-08-25  9:43                 ` Vitaly Kuznetsov
2021-08-25 10:41                   ` Maxim Levitsky
2021-08-25 13:19                   ` Eduardo Habkost
2021-08-26 12:40                     ` Vitaly Kuznetsov
2021-08-26 14:52                       ` Eduardo Habkost
2021-08-26 18:01                         ` Sean Christopherson
2021-08-26 18:13                           ` Eduardo Habkost
2021-08-26 19:27             ` Eduardo Habkost
2021-08-30 19:47             ` Nitesh Lal

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=20210823143028.649818-4-vkuznets@redhat.com \
    --to=vkuznets@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=jmattson@google.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=nitesh@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=seanjc@google.com \
    --cc=wanpengli@tencent.com \
    --subject='Re: [PATCH v2 3/4] KVM: Optimize kvm_make_vcpus_request_mask() a bit' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).