qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Daniel P. Berrangé" <berrange@redhat.com>
To: Roy Hopkins <roy.hopkins@suse.com>
Cc: qemu-devel@nongnu.org, "Paolo Bonzini" <pbonzini@redhat.com>,
	"Stefano Garzarella" <sgarzare@redhat.com>,
	"Marcelo Tosatti" <mtosatti@redhat.com>,
	"Michael S . Tsirkin" <mst@redhat.com>,
	"Cornelia Huck" <cohuck@redhat.com>,
	"Marcel Apfelbaum" <marcel.apfelbaum@gmail.com>,
	"Sergio Lopez" <slp@redhat.com>,
	"Eduardo Habkost" <eduardo@habkost.net>,
	"Alistair Francis" <alistair@alistair23.me>,
	"Peter Xu" <peterx@redhat.com>,
	"David Hildenbrand" <david@redhat.com>,
	"Igor Mammedov" <imammedo@redhat.com>,
	"Tom Lendacky" <thomas.lendacky@amd.com>,
	"Michael Roth" <michael.roth@amd.com>,
	"Ani Sinha" <anisinha@redhat.com>,
	"Jörg Roedel" <jroedel@suse.com>
Subject: Re: [PATCH v3 15/15] sev: Provide sev_features flags from IGVM VMSA to KVM_SEV_INIT2
Date: Mon, 24 Jun 2024 15:14:47 +0100	[thread overview]
Message-ID: <Znl_VyXoV30-Cgob@redhat.com> (raw)
In-Reply-To: <8e49a1aed06d4f98c13a98fbbe3a48cdeb6e0070.1718979106.git.roy.hopkins@suse.com>

On Fri, Jun 21, 2024 at 03:29:18PM +0100, Roy Hopkins wrote:
> IGVM files can contain an initial VMSA that should be applied to each
> vcpu as part of the initial guest state. The sev_features flags are
> provided as part of the VMSA structure. However, KVM only allows
> sev_features to be set during initialization and not as the guest is
> being prepared for launch.
> 
> This patch queries KVM for the supported set of sev_features flags and
> processes the IGVM file during kvm_init to determine any sev_features
> flags set in the IGVM file. These are then provided in the call to
> KVM_SEV_INIT2 to ensure the guest state matches that specified in the
> IGVM file.
> 
> This does cause the IGVM file to be processed twice. Firstly to extract
> the sev_features then secondly to actually configure the guest. However,
> the first pass is largely ignored meaning the overhead is minimal.
> 
> Signed-off-by: Roy Hopkins <roy.hopkins@suse.com>
> ---
>  target/i386/sev.c | 145 ++++++++++++++++++++++++++++++++++++++++------
>  1 file changed, 126 insertions(+), 19 deletions(-)
> 

>  static int sev_common_kvm_init(ConfidentialGuestSupport *cgs, Error **errp)
>  {
>      char *devname;
> @@ -1743,6 +1783,10 @@ static int sev_common_kvm_init(ConfidentialGuestSupport *cgs, Error **errp)
>          }
>      }
>  
> +    if (sev_init_supported_features(sev_common, errp) < 0) {
> +        return -1;
> +    }
> +
>      trace_kvm_sev_init();
>      if (x86_klass->kvm_type(X86_CONFIDENTIAL_GUEST(sev_common)) == KVM_X86_DEFAULT_VM) {
>          cmd = sev_es_enabled() ? KVM_SEV_ES_INIT : KVM_SEV_INIT;
> @@ -1750,6 +1794,38 @@ static int sev_common_kvm_init(ConfidentialGuestSupport *cgs, Error **errp)
>          ret = sev_ioctl(sev_common->sev_fd, cmd, NULL, &fw_error);
>      } else {
>          struct kvm_sev_init args = { 0 };
> +        MachineState *machine = MACHINE(qdev_get_machine());
> +
> +        /*
> +         * If configuration is provided via an IGVM file then the IGVM file
> +         * might contain configuration of the initial vcpu context. For SEV
> +         * the vcpu context includes the sev_features which should be applied
> +         * to the vcpu.
> +         *
> +         * KVM does not synchronize sev_features from CPU state. Instead it
> +         * requires sev_features to be provided as part of this initialization
> +         * call which is subsequently automatically applied to the VMSA of
> +         * each vcpu.
> +         *
> +         * The IGVM file is normally processed after initialization. Therefore
> +         * we need to pre-process it here to extract sev_features in order to
> +         * provide it to KVM_SEV_INIT2. Each cgs_* function that is called by
> +         * the IGVM processor detects this pre-process by observing the state
> +         * as SEV_STATE_UNINIT.
> +         */
> +        if (machine->igvm) {
> +            if (IGVM_CFG_GET_CLASS(machine->igvm)
> +                    ->process(machine->igvm, machine->cgs, errp) == -1) {
> +                return -1;
> +            }
> +            /*
> +             * KVM maintains a bitmask of allowed sev_features. This does not
> +             * include SVM_SEV_FEAT_SNP_ACTIVE which is set accordingly by KVM
> +             * itself. Therefore we need to clear this flag.
> +             */
> +            args.vmsa_features = sev_common->sev_features &
> +                                 ~SVM_SEV_FEAT_SNP_ACTIVE;
> +        }
>  
>          ret = sev_ioctl(sev_common->sev_fd, KVM_SEV_INIT2, &args, &fw_error);
>      }

What happens if the code path takes us down the KVM_SEV_INIT
route, rather than KVM_SEV_INIT2 ?  Should we be reporting an
error indicating that IGVM usage is incompatible with the legacy
KVM_SEV_INIT path ?


With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|



  reply	other threads:[~2024-06-24 14:15 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-21 14:29 [PATCH v3 00/15] Introduce support for IGVM files Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 01/15] meson: Add optional dependency on IGVM library Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 02/15] backends/confidential-guest-support: Add functions to support IGVM Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 03/15] backends/igvm: Add IGVM loader and configuration Roy Hopkins
2024-06-24 13:29   ` Daniel P. Berrangé
2024-06-28 10:59     ` Roy Hopkins
2024-06-27  9:06   ` Stefano Garzarella
2024-06-27  9:14     ` Daniel P. Berrangé
2024-06-28 11:00       ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 04/15] hw/core/machine: Add igvm-cfg object and processing for IGVM files Roy Hopkins
2024-06-24 14:00   ` Daniel P. Berrangé
2024-06-28 11:09     ` Roy Hopkins
2024-06-28 11:23       ` Daniel P. Berrangé
2024-07-01 11:59         ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 05/15] i386/pc_sysfw: Ensure sysfw flash configuration does not conflict with IGVM Roy Hopkins
2024-06-27 12:38   ` Stefano Garzarella
2024-06-28 11:10     ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 06/15] sev: Update launch_update_data functions to use Error handling Roy Hopkins
2024-06-27 12:48   ` Stefano Garzarella
2024-06-28 11:20     ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 07/15] i386/sev: Refactor setting of reset vector and initial CPU state Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 08/15] i386/sev: Implement ConfidentialGuestSupport functions for SEV Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 09/15] docs/system: Add documentation on support for IGVM Roy Hopkins
2024-06-24 14:09   ` Daniel P. Berrangé
2024-07-01 14:28     ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 10/15] docs/interop/firmware.json: Add igvm to FirmwareDevice Roy Hopkins
2024-06-27 12:53   ` Stefano Garzarella
2024-07-02 10:36     ` Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 11/15] backends/confidential-guest-support: Add set_guest_policy() function Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 12/15] backends/igvm: Process initialization sections in IGVM file Roy Hopkins
2024-06-21 14:29 ` [PATCH v3 13/15] backends/igvm: Handle policy for SEV guests Roy Hopkins
2024-06-24 14:56   ` Daniel P. Berrangé
2024-06-21 14:29 ` [PATCH v3 14/15] i386/sev: Add implementation of CGS set_guest_policy() Roy Hopkins
2024-06-24 14:53   ` Daniel P. Berrangé
2024-06-21 14:29 ` [PATCH v3 15/15] sev: Provide sev_features flags from IGVM VMSA to KVM_SEV_INIT2 Roy Hopkins
2024-06-24 14:14   ` Daniel P. Berrangé [this message]
2024-07-01 13:50     ` Roy Hopkins
2024-06-24 13:50 ` [PATCH v3 00/15] Introduce support for IGVM files Daniel P. Berrangé
2024-06-28 10:56   ` Roy Hopkins

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Znl_VyXoV30-Cgob@redhat.com \
    --to=berrange@redhat.com \
    --cc=alistair@alistair23.me \
    --cc=anisinha@redhat.com \
    --cc=cohuck@redhat.com \
    --cc=david@redhat.com \
    --cc=eduardo@habkost.net \
    --cc=imammedo@redhat.com \
    --cc=jroedel@suse.com \
    --cc=marcel.apfelbaum@gmail.com \
    --cc=michael.roth@amd.com \
    --cc=mst@redhat.com \
    --cc=mtosatti@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peterx@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=roy.hopkins@suse.com \
    --cc=sgarzare@redhat.com \
    --cc=slp@redhat.com \
    --cc=thomas.lendacky@amd.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).