[PATCH] KVM: Remove CREATE_IRQCHIP/SET_PIT2 race

[PATCH] KVM: Remove CREATE_IRQCHIP/SET_PIT2 race

[Jon Cargille]

From: Steve Rutherford <srutherford@google.com>

Fixes a NULL pointer dereference, caused by the PIT firing an interrupt
before the interrupt table has been initialized.

SET_PIT2 can race with the creation of the IRQchip. In particular,
if SET_PIT2 is called with a low PIT timer period (after the creation of
the IOAPIC, but before the instantiation of the irq routes), the PIT can
fire an interrupt at an uninitialized table.

Signed-off-by: Steve Rutherford <srutherford@google.com>
Signed-off-by: Jon Cargille <jcargill@google.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
---
 arch/x86/kvm/x86.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 027dfd278a973..3cc3f673785c8 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -5049,10 +5049,13 @@ long kvm_arch_vm_ioctl(struct file *filp,
 		r = -EFAULT;
 		if (copy_from_user(&u.ps, argp, sizeof(u.ps)))
 			goto out;
+		mutex_lock(&kvm->lock);
 		r = -ENXIO;
 		if (!kvm->arch.vpit)
-			goto out;
+			goto set_pit_out;
 		r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
+set_pit_out:
+		mutex_unlock(&kvm->lock);
 		break;
 	}
 	case KVM_GET_PIT2: {
@@ -5072,10 +5075,13 @@ long kvm_arch_vm_ioctl(struct file *filp,
 		r = -EFAULT;
 		if (copy_from_user(&u.ps2, argp, sizeof(u.ps2)))
 			goto out;
+		mutex_lock(&kvm->lock);
 		r = -ENXIO;
 		if (!kvm->arch.vpit)
-			goto out;
+			goto set_pit2_out;
 		r = kvm_vm_ioctl_set_pit2(kvm, &u.ps2);
+set_pit2_out:
+		mutex_unlock(&kvm->lock);
 		break;
 	}
 	case KVM_REINJECT_CONTROL: {
-- 
2.26.0.110.g2183baf09c-goog

Re: [PATCH] KVM: Remove CREATE_IRQCHIP/SET_PIT2 race

[Paolo Bonzini]

On 16/04/20 21:11, Jon Cargille wrote:
> From: Steve Rutherford <srutherford@google.com>
> 
> Fixes a NULL pointer dereference, caused by the PIT firing an interrupt
> before the interrupt table has been initialized.
> 
> SET_PIT2 can race with the creation of the IRQchip. In particular,
> if SET_PIT2 is called with a low PIT timer period (after the creation of
> the IOAPIC, but before the instantiation of the irq routes), the PIT can
> fire an interrupt at an uninitialized table.
> 
> Signed-off-by: Steve Rutherford <srutherford@google.com>
> Signed-off-by: Jon Cargille <jcargill@google.com>
> Reviewed-by: Jim Mattson <jmattson@google.com>
> ---
>  arch/x86/kvm/x86.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 027dfd278a973..3cc3f673785c8 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -5049,10 +5049,13 @@ long kvm_arch_vm_ioctl(struct file *filp,
>  		r = -EFAULT;
>  		if (copy_from_user(&u.ps, argp, sizeof(u.ps)))
>  			goto out;
> +		mutex_lock(&kvm->lock);
>  		r = -ENXIO;
>  		if (!kvm->arch.vpit)
> -			goto out;
> +			goto set_pit_out;
>  		r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
> +set_pit_out:
> +		mutex_unlock(&kvm->lock);
>  		break;
>  	}
>  	case KVM_GET_PIT2: {
> @@ -5072,10 +5075,13 @@ long kvm_arch_vm_ioctl(struct file *filp,
>  		r = -EFAULT;
>  		if (copy_from_user(&u.ps2, argp, sizeof(u.ps2)))
>  			goto out;
> +		mutex_lock(&kvm->lock);
>  		r = -ENXIO;
>  		if (!kvm->arch.vpit)
> -			goto out;
> +			goto set_pit2_out;
>  		r = kvm_vm_ioctl_set_pit2(kvm, &u.ps2);
> +set_pit2_out:
> +		mutex_unlock(&kvm->lock);
>  		break;
>  	}
>  	case KVM_REINJECT_CONTROL: {
> 

Queued, thanks.

Paolo