Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Emanuele Giuseppe Esposito <eesposit@redhat.com>
To:     kvm@vger.kernel.org
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        Jonathan Corbet <corbet@lwn.net>,
        Maxim Levitsky <mlevitsk@redhat.com>,
        Sean Christopherson <seanjc@google.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        David Hildenbrand <david@redhat.com>, x86@kernel.org,
        "H. Peter Anvin" <hpa@zytor.com>, linux-doc@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Emanuele Giuseppe Esposito <eesposit@redhat.com>
Subject: [PATCH 3/4] KVM: introduce memory transaction semaphore
Date:   Sat, 22 Oct 2022 11:48:18 -0400
Message-Id: <20221022154819.1823133-4-eesposit@redhat.com>
In-Reply-To: <20221022154819.1823133-1-eesposit@redhat.com>
References: <20221022154819.1823133-1-eesposit@redhat.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Right now the semaphore is only used to signal that a vcpu
entered KVM_RUN (not necessarly in guest mode, could be also
blocked/halted).
Later it will be used by specific ioctls (writers) to wait that
all vcpus (readers) exit from KVM_RUN.

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
---
 virt/kvm/kvm_main.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index c080b93edc0d..ae0240928a4a 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -119,6 +119,8 @@ static const struct file_operations stat_fops_per_vm;
 
 static struct file_operations kvm_chardev_ops;
 
+static DECLARE_RWSEM(memory_transaction);
+
 static long kvm_vcpu_ioctl(struct file *file, unsigned int ioctl,
 			   unsigned long arg);
 #ifdef CONFIG_KVM_COMPAT
@@ -4074,7 +4076,19 @@ static long kvm_vcpu_ioctl(struct file *filp,
 				synchronize_rcu();
 			put_pid(oldpid);
 		}
+		/*
+		 * Notify that a vcpu wants to run, and thus could be reading
+		 * memslots.
+		 * If KVM_KICK_ALL_RUNNING_VCPUS runs afterwards, it will have
+		 * to wait that KVM_RUN exited and up_read() is called.
+		 * If KVM_KICK_ALL_RUNNING_VCPUS already returned but
+		 * KVM_RESUME_ALL_KICKED_VCPUS didn't start yet, then there
+		 * is a request pending for the vcpu that will cause it to
+		 * exit KVM_RUN.
+		 */
+		down_read(&memory_transaction);
 		r = kvm_arch_vcpu_ioctl_run(vcpu);
+		up_read(&memory_transaction);
 		trace_kvm_userspace_exit(vcpu->run->exit_reason, r);
 		break;
 	}
-- 
2.31.1