linux-kernel - Re: [PATCH] KVM: apic: avoid calculating pending eoi from an uninitialized val

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Thu, 20 Feb 2020 09:20:38 -0800
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Vitaly Kuznetsov <vkuznets@...hat.com>
Cc:     linmiaohe <linmiaohe@...wei.com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, x86@...nel.org, pbonzini@...hat.com,
        rkrcmar@...hat.com, wanpengli@...cent.com, jmattson@...gle.com,
        joro@...tes.org, tglx@...utronix.de, mingo@...hat.com,
        bp@...en8.de, hpa@...or.com
Subject: Re: [PATCH] KVM: apic: avoid calculating pending eoi from an
 uninitialized val

On Thu, Feb 20, 2020 at 05:33:17PM +0100, Vitaly Kuznetsov wrote:
> linmiaohe <linmiaohe@...wei.com> writes:
> 
> > From: Miaohe Lin <linmiaohe@...wei.com>
> >
> > When get user eoi value failed, var val would be uninitialized and result
> > in calculating pending eoi from an uninitialized val. Initialize var val
> > to 0 to fix this case.
> 
> Let me try to suggest an alternative wording,
> 
> "When pv_eoi_get_user() fails, 'val' may remain uninitialized and the
> return value of pv_eoi_get_pending() becomes random. Fix the issue by
> initializing the variable."
> 
> >
> > Signed-off-by: Miaohe Lin <linmiaohe@...wei.com>
> > ---
> >  arch/x86/kvm/lapic.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> > index 4f14ec7525f6..7e77e94f3176 100644
> > --- a/arch/x86/kvm/lapic.c
> > +++ b/arch/x86/kvm/lapic.c
> > @@ -626,7 +626,7 @@ static inline bool pv_eoi_enabled(struct kvm_vcpu *vcpu)
> >  
> >  static bool pv_eoi_get_pending(struct kvm_vcpu *vcpu)
> >  {
> > -	u8 val;
> > +	u8 val = 0;

Rather than initialize @val, I'd prefer to explicitly handle the error,
similar to pv_eoi_clr_pending() and pv_eoi_set_pending(), e.g.

	u8 val;

	if (pv_eoi_get_user(vcpu, &val) < 0) {
		printk(KERN_WARNING "Can't read EOI MSR value: 0x%llx\n",
			   (unsigned long long)vcpu->arch.pv_eoi.msr_val);
		return false;
	}
	return val & 0x1;

> >  	if (pv_eoi_get_user(vcpu, &val) < 0)
> >  		printk(KERN_WARNING "Can't read EOI MSR value: 0x%llx\n",
> >  			   (unsigned long long)vcpu->arch.pv_eoi.msr_val);
> 
> Reviewed-by: Vitaly Kuznetsov <vkuznets@...hat.com>
> 
> But why compilers don't complain?

Clang might?