diff mbox

KVM: protect concurrent make_all_cpus_request

Message ID 20090617135347.GC14583@amt.cnet (mailing list archive)
State New, archived
Headers show

Commit Message

Marcelo Tosatti June 17, 2009, 1:53 p.m. UTC
make_all_cpus_request contains a race condition which can 
trigger false request completed status, as follows:

CPU0                                              CPU1

if (test_and_set_bit(req,&vcpu->requests))
   ....                                        	   if (test_and_set_bit(req,&vcpu->requests))
   ..                                                  return
proceed to smp_call_function_many(wait=1)

Use a spinlock to serialize concurrent CPUs.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Andrea Arcangeli June 17, 2009, 4:52 p.m. UTC | #1
On Wed, Jun 17, 2009 at 10:53:47AM -0300, Marcelo Tosatti wrote:
> 
> 
> make_all_cpus_request contains a race condition which can 
> trigger false request completed status, as follows:
> 
> CPU0                                              CPU1
> 
> if (test_and_set_bit(req,&vcpu->requests))
>    ....                                        	   if (test_and_set_bit(req,&vcpu->requests))
>    ..                                                  return
> proceed to smp_call_function_many(wait=1)
> 
> Use a spinlock to serialize concurrent CPUs.

Agreed, this is the safest...
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Avi Kivity June 18, 2009, 8:45 a.m. UTC | #2
On 06/17/2009 04:53 PM, Marcelo Tosatti wrote:
> make_all_cpus_request contains a race condition which can
> trigger false request completed status, as follows:
>
> CPU0                                              CPU1
>
> if (test_and_set_bit(req,&vcpu->requests))
>     ....                                        	   if (test_and_set_bit(req,&vcpu->requests))
>     ..                                                  return
> proceed to smp_call_function_many(wait=1)
>
> Use a spinlock to serialize concurrent CPUs.
>
>    

Applied, thanks.
diff mbox

Patch

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 1b48092..2451f48 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -124,6 +124,7 @@  struct kvm_kernel_irq_routing_entry {
 
 struct kvm {
 	spinlock_t mmu_lock;
+	spinlock_t requests_lock;
 	struct rw_semaphore slots_lock;
 	struct mm_struct *mm; /* userspace tied to this vm */
 	int nmemslots;
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 8939ffa..6847bb7 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -739,6 +739,7 @@  static bool make_all_cpus_request(struct kvm *kvm, unsigned int req)
 		cpumask_clear(cpus);
 
 	me = get_cpu();
+	spin_lock(&kvm->requests_lock);
 	kvm_for_each_vcpu(i, vcpu, kvm) {
 		if (test_and_set_bit(req, &vcpu->requests))
 			continue;
@@ -752,6 +753,7 @@  static bool make_all_cpus_request(struct kvm *kvm, unsigned int req)
 		smp_call_function_many(cpus, ack_flush, NULL, 1);
 	else
 		called = false;
+	spin_unlock(&kvm->requests_lock);
 	put_cpu();
 	free_cpumask_var(cpus);
 	return called;
@@ -972,6 +974,7 @@  static struct kvm *kvm_create_vm(void)
 	kvm->mm = current->mm;
 	atomic_inc(&kvm->mm->mm_count);
 	spin_lock_init(&kvm->mmu_lock);
+	spin_lock_init(&kvm->requests_lock);
 	kvm_io_bus_init(&kvm->pio_bus);
 	kvm_irqfd_init(kvm);
 	mutex_init(&kvm->lock);