linux-kernel - Re: [PATCH] KVM: x86: Perform non-canonical checks in 32-bit KVM

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

Message-ID: <20200116155057.GB20561@linux.intel.com>
Date:   Thu, 16 Jan 2020 07:50:57 -0800
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Krish Sadhukhan <krish.sadhukhan@...cle.com>
Cc:     Paolo Bonzini <pbonzini@...hat.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] KVM: x86: Perform non-canonical checks in 32-bit KVM

On Wed, Jan 15, 2020 at 05:37:16PM -0800, Krish Sadhukhan wrote:
> 
> On 01/15/2020 10:36 AM, Sean Christopherson wrote:
> >  arch/x86/kvm/x86.h | 8 --------
> >  1 file changed, 8 deletions(-)
> >
> >diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
> >index cab5e71f0f0f..3ff590ec0238 100644
> >--- a/arch/x86/kvm/x86.h
> >+++ b/arch/x86/kvm/x86.h
> >@@ -166,21 +166,13 @@ static inline u64 get_canonical(u64 la, u8 vaddr_bits)
> >  static inline bool is_noncanonical_address(u64 la, struct kvm_vcpu *vcpu)
> >  {
> >-#ifdef CONFIG_X86_64
> >  	return get_canonical(la, vcpu_virt_addr_bits(vcpu)) != la;
> >-#else
> >-	return false;
> >-#endif
> >  }
> >  static inline bool emul_is_noncanonical_address(u64 la,
> >  						struct x86_emulate_ctxt *ctxt)
> >  {
> >-#ifdef CONFIG_X86_64
> >  	return get_canonical(la, ctxt_virt_addr_bits(ctxt)) != la;
> >-#else
> >-	return false;
> >-#endif
> >  }
> >  static inline void vcpu_cache_mmio_info(struct kvm_vcpu *vcpu,
> 
> nested_vmx_check_host_state() still won't call it on 32-bit because it has
> the CONFIG_X86_64 guard around the callee:
> 
>  #ifdef CONFIG_X86_64
>         if (CC(is_noncanonical_address(vmcs12->host_fs_base, vcpu)) ||
>             CC(is_noncanonical_address(vmcs12->host_gs_base, vcpu)) ||
>  ...

Doh, I was looking at an older version of nested.c.  Nice catch!

> Don't we need to remove these guards in the callers as well ?

Ya, that would be my preference.