linux-kernel - Re: [PATCH v2 1/4] KVM: SVM: Get rid of *ghcb_msr

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]

Message-ID: <YUiyg1W1cDxVbgzs@google.com>
Date:   Mon, 20 Sep 2021 16:10:43 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Joerg Roedel <joro@...tes.org>
Cc:     Paolo Bonzini <pbonzini@...hat.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Brijesh Singh <brijesh.singh@....com>,
        Tom Lendacky <thomas.lendacky@....com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-coco@...ts.linux.dev,
        Joerg Roedel <jroedel@...e.de>
Subject: Re: [PATCH v2 1/4] KVM: SVM: Get rid of *ghcb_msr_bits() functions

On Thu, Sep 09, 2021, Joerg Roedel wrote:
> On Wed, Sep 01, 2021 at 09:12:10PM +0000, Sean Christopherson wrote:
> > On Thu, Jul 22, 2021, Joerg Roedel wrote:
> > >  	case GHCB_MSR_TERM_REQ: {
> > >  		u64 reason_set, reason_code;
> > >  
> > > -		reason_set = get_ghcb_msr_bits(svm,
> > > -					       GHCB_MSR_TERM_REASON_SET_MASK,
> > > -					       GHCB_MSR_TERM_REASON_SET_POS);
> > > -		reason_code = get_ghcb_msr_bits(svm,
> > > -						GHCB_MSR_TERM_REASON_MASK,
> > > -						GHCB_MSR_TERM_REASON_POS);
> > > +		reason_set  = GHCB_MSR_TERM_REASON_SET(control->ghcb_gpa);
> > > +		reason_code = GHCB_MSR_TERM_REASON(control->ghcb_gpa);
> > > +
> > >  		pr_info("SEV-ES guest requested termination: %#llx:%#llx\n",
> > >  			reason_set, reason_code);
> > > +
> > >  		fallthrough;
> > 
> > Not related to this patch, but why use fallthrough and more importantly, why is
> > this an -EINVAL return?  Why wouldn't KVM forward the request to userspace instead
> > of returning an opaque -EINVAL?
> 
> I guess it is to signal an error condition up the call-chain to get the
> guest terminated, like requested.

Yes, but it's odd bizarre/unfortunate that KVM doesn't take this opportunity to
forward the termination info to the VMM.  The above pr_info() should not exist.
If that information is relevant then it should be handed to the VMM directly, not
dumped to dmesg.