linux-kernel - Re: [PATCH v3 2/4] KVM: SVM: Clear MSR_TSC

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: <YIcYCAANV6o3EI1n@google.com>
Date:   Mon, 26 Apr 2021 19:44:08 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Vitaly Kuznetsov <vkuznets@...hat.com>
Cc:     Paolo Bonzini <pbonzini@...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, Reiji Watanabe <reijiw@...gle.com>
Subject: Re: [PATCH v3 2/4] KVM: SVM: Clear MSR_TSC_AUX[63:32] on write

On Mon, Apr 26, 2021, Sean Christopherson wrote:
> On Mon, Apr 26, 2021, Vitaly Kuznetsov wrote:
> > Actually, shouldn't we just move wrmsrl() here? Assuming we're not sure
> > how (and if) upper 32 bits are going to be used, it would probably make
> > sense to not write them to the actual MSR...
> 
> Argh.  I got too clever in trying to minimize the patch deltas and "broke" this
> intermediate patch.  Once the user return framework is used, the result of
> setting the MSR is checked before setting svm->tsc_aux, i.e. don't set the
> virtual state if the real state could not be set.
> 
> I'm very tempted to add yet another patch to this mess to do:
> 
> 		if (wrmsrl_safe(MSR_TSC_AUX, data))
> 			return 1;
> 
> 		svm->tsc_aux = data;
> 
> And then this patch becomes:
> 
> 		data = (u32)data;
> 
> 		if (wrmsrl_safe(MSR_TSC_AUX, data))
> 			return 1;
> 
> 		svm->tsc_aux = data;
> 
> The above will also make patch 3 cleaner as it will preserve the ordering of
> truncating data and the wrmsr.

Ah, never mind, Paolo already pushed this to kvm/next with your above fix.