Date:   Mon, 6 Dec 2021 14:32:31 +0500
From:   Ameer Hamza <amhamza.mgc@gmail.com>
To:     Vitaly Kuznetsov <vkuznets@redhat.com>
Cc:     kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
        pbonzini@redhat.com, seanjc@google.com, wanpengli@tencent.com,
        jmattson@google.com, joro@8bytes.org, tglx@linutronix.de,
        mingo@redhat.com, bp@alien8.de, dave.hansen@linux.intel.com,
        x86@kernel.org, hpa@zytor.com
Subject: Re: [PATCH] KVM: x86: fix for missing initialization of return
 status variable
Message-ID: <20211206093231.GB5609@hamza-OptiPlex-7040>
References: <20211205194719.16987-1-amhamza.mgc@gmail.com>
 <87ee6q6r1p.fsf@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87ee6q6r1p.fsf@redhat.com>
Precedence: bulk

On Mon, Dec 06, 2021 at 10:06:26AM +0100, Vitaly Kuznetsov wrote:
> Ameer Hamza <amhamza.mgc@gmail.com> writes:
> 
> > If undefined ioctl number is passed to the kvm_vcpu_ioctl_device_attr
> > function, it should return with error status.
> >
> > Addresses-Coverity: 1494124 ("Uninitialized scalar variable")
> >
> > Signed-off-by: Ameer Hamza <amhamza.mgc@gmail.com>
> > ---
> >  arch/x86/kvm/x86.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index e0aa4dd53c7f..55b90c185717 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -5001,7 +5001,7 @@ static int kvm_vcpu_ioctl_device_attr(struct kvm_vcpu *vcpu,
> >  				      void __user *argp)
> >  {
> >  	struct kvm_device_attr attr;
> > -	int r;
> > +	int r = -EINVAL;
> >  
> >  	if (copy_from_user(&attr, argp, sizeof(attr)))
> >  		return -EFAULT;
> 
> The reported issue is not real, kvm_vcpu_ioctl_device_attr() is never
> called with anything but [KVM_HAS_DEVICE_ATTR, KVM_GET_DEVICE_ATTR,
> KVM_SET_DEVICE_ATTR] as 'ioctl' and the switch below covers all
> three. Instead of initializing 'r' we could've added a 'default' case to
> the switch, either returning something like EINVAL or just BUG(). Hope
> it'll silence coverity.
Thank you for your kind response. I agree with you and I think its
logical to add default case here. Let me update the patch.

Best Regards,
Hamza