Re: [PATCH RESEND] kvm: Fix nonsense handling of compat ioctl

[Avi Kivity <avi@redhat.com>]

On 08/20/2012 05:43 PM, Alan Cox wrote:
> From: Alan Cox <alan@linux.intel.com>
> 
> KVM_SET_SIGNAL_MASK passed a NULL argument leaves the on stack signal
> sets uninitialized. It then passes them through to
> kvm_vcpu_ioctl_set_sigmask.
> 
> We should be passing a NULL in this case not translated garbage.
> 
> Signed-off-by: Alan Cox <alan@linux.intel.com>
> ---
> 
>  virt/kvm/kvm_main.c |    7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index a2e85af..e47a7ca 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -1975,9 +1975,10 @@ static long kvm_vcpu_compat_ioctl(struct file *filp,
>  			if (copy_from_user(&csigset, sigmask_arg->sigset,
>  					   sizeof csigset))
>  				goto out;
> -		}
> -		sigset_from_compat(&sigset, &csigset);
> -		r = kvm_vcpu_ioctl_set_sigmask(vcpu, &sigset);
> +			sigset_from_compat(&sigset, &csigset);
> +			r = kvm_vcpu_ioctl_set_sigmask(vcpu, &sigset);
> +		} else
> +			kvm_vcpu_ioctl_set_sigmask(vcpu, NULL);
>  		break;
>  	}

Now r is uninitiali[sz]ed.


-- 
error compiling committee.c: too many arguments to function