linux-kernel - Re: [PATCH v12 32/46] x86/compressed/64: Add support for SEV-SNP CPUID table in #VC handlers

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: <CAMkAt6r==_=U4Ha6ZTmii-JL3htJ3-dD4tc+QBqN7dVt711N2A@mail.gmail.com>
Date:   Mon, 14 Mar 2022 11:34:42 -0600
From:   Peter Gonda <pgonda@...gle.com>
To:     Joerg Roedel <jroedel@...e.de>
Cc:     Michael Roth <michael.roth@....com>,
        Brijesh Singh <brijesh.singh@....com>,
        "the arch/x86 maintainers" <x86@...nel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        kvm list <kvm@...r.kernel.org>, linux-efi@...r.kernel.org,
        platform-driver-x86@...r.kernel.org, linux-coco@...ts.linux.dev,
        linux-mm@...ck.org, Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        Tom Lendacky <thomas.lendacky@....com>,
        "H. Peter Anvin" <hpa@...or.com>, Ard Biesheuvel <ardb@...nel.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Sean Christopherson <seanjc@...gle.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Jim Mattson <jmattson@...gle.com>,
        Andy Lutomirski <luto@...nel.org>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Sergio Lopez <slp@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
        David Rientjes <rientjes@...gle.com>,
        Dov Murik <dovmurik@...ux.ibm.com>,
        Tobin Feldman-Fitzthum <tobin@....com>,
        Borislav Petkov <bp@...en8.de>,
        Vlastimil Babka <vbabka@...e.cz>,
        "Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
        Andi Kleen <ak@...ux.intel.com>,
        "Dr . David Alan Gilbert" <dgilbert@...hat.com>,
        brijesh.ksingh@...il.com, Tony Luck <tony.luck@...el.com>,
        Marc Orr <marcorr@...gle.com>,
        Sathyanarayanan Kuppuswamy 
        <sathyanarayanan.kuppuswamy@...ux.intel.com>
Subject: Re: [PATCH v12 32/46] x86/compressed/64: Add support for SEV-SNP
 CPUID table in #VC handlers

On Fri, Mar 11, 2022 at 10:06 AM Joerg Roedel <jroedel@...e.de> wrote:
>
> On Thu, Mar 10, 2022 at 03:25:04PM -0600, Michael Roth wrote:
> > Joerg, do you have more background on that? Would it make sense, outside
> > of this series, to change it to a terminate? Maybe with a specific set
> > of error codes for ES_{OK,UNSUPPORTED,VMM_ERROR,DECODE_FAILED}?
>
> This seems to be a left over from development of the SEV-ES guest
> patch-set. I wanted to see whether the VM crashed due to a triple fault
> or an error in the #VC handler. The halt loop can be replaced by
> termination request now.
>
> > > I am still working on why the early_printk()s in that function are not
> > > working, it seems that they lead to a different halt.
> >
> > I don't see a different halt. They just don't seem to print anything.
> > (keep in mind you still need to advance the IP or else the guest is
> > still gonna end up spinning here, even if you're removing the halt loop
> > for testing purposes)
>
> The early_printks() also cause #VC exceptions, and if that handling is
> broken for some reason nothing will be printed.
>
> >
> > > working, it seems that they lead to a different halt. Have you tested
> > > any of those error paths manually? For example if you set your CPUID
> > > bits to explicitly fail here do you see the expected printks?
> >
> > I think at that point in the code, when the XSAVE stuff is setup, the
> > console hasn't been enabled yet, so messages would get buffered until they
> > get flushed later (which won't happen since there's halt loop after). I
> > know in some cases devs will dump the log buffer from memory instead to get
> > at the error messages for early failures. (Maybe that's also why Joerg
> > decided to use a halt loop there instead of terminating?)
>
> It is hard to dump the log-buffer from encrypted memory :) But I
> remember having seen messages from these early_printks under SEV-ES for
> different bugs. Not sure why they don't appear in this situation.
>
> > So maybe reworking the error handling in handle_vc_boot_ghcb() to use
> > sev_es_terminate() might be warranted, but probably worth checking with
> > Joerg first, and should be done as a separate series since it is not
> > SNP-related.
>
> I am fine with this change.

I'll send a patch out for that.

I was also thinking about adding a vcpu run exit reason for
termination. It would be nice to get a more informative exit reason
than -EINVAL in userspace. Thoughts?

>
> Regards,
>
> --
> Jörg Rödel
> jroedel@...e.de
>
> SUSE Software Solutions Germany GmbH
> Maxfeldstr. 5
> 90409 Nürnberg
> Germany
>
> (HRB 36809, AG Nürnberg)
> Geschäftsführer: Ivo Totev
>