linux-kernel - Re: [PATCH v3] x86/sev: Add SEV-SNP guest feature negotiation support

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]

Date:   Tue, 3 Jan 2023 08:55:35 +0530
From:   "Nikunj A. Dadhania" <nikunj@....com>
To:     Borislav Petkov <bp@...en8.de>
Cc:     David Rientjes <rientjes@...gle.com>, linux-kernel@...r.kernel.org,
        x86@...nel.org, kvm@...r.kernel.org, mingo@...hat.com,
        tglx@...utronix.de, dave.hansen@...ux.intel.com, seanjc@...gle.com,
        pbonzini@...hat.com, thomas.lendacky@....com, michael.roth@....com,
        stable@...nel.org
Subject: Re: [PATCH v3] x86/sev: Add SEV-SNP guest feature negotiation support

On 03/01/23 01:12, Borislav Petkov wrote:
> On Mon, Jan 02, 2023 at 08:50:23PM +0530, Nikunj A. Dadhania wrote:
>> I think the "why" part depends on the user. Whether or not the user needs a
>> certain feature enabled for the confidential guest.
>>
>> If the cloud provider(hypervisor) enables the feature on user request, the
>> guest terminates with GHCB_SNP_FEAT_NOT_IMPLEMENTED when guest kernel does
>> have corresponding code/implementation.
> 
> I think you mean "does not have" here.

Yes, that is correct.

> 
> In any case, I think this whole handling of SEV features could go both ways:
> 
> * Cloud provider could say: we've enabled features X, Y and Z and if the guest
> doesn't have support for them, then it would fail booting.
> 
> There would optimally be some text sowewhere in the cloud provider documentation
> stating why those features are enabled and thus required to be supported by the
> guest.
> 
> * Guest owner could require a minimal subset of features which must be present
> in the HV in order to even boot on that HV.
> 
> Of course, I'm only speculating here. How it ends up really playing out in
> reality we will have to see...
>