Re: [PATCH v8 07/13] confidential guest support: Introduce cgs "ready" flag

[Greg Kurz <groug@kaod.org>]

On Tue,  2 Feb 2021 15:13:09 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> The platform specific details of mechanisms for implementing
> confidential guest support may require setup at various points during
> initialization.  Thus, it's not really feasible to have a single cgs
> initialization hook, but instead each mechanism needs its own
> initialization calls in arch or machine specific code.
> 
> However, to make it harder to have a bug where a mechanism isn't
> properly initialized under some circumstances, we want to have a
> common place, late in boot, where we verify that cgs has been
> initialized if it was requested.
> 
> This patch introduces a ready flag to the ConfidentialGuestSupport
> base type to accomplish this, which we verify in
> qemu_machine_creation_done().
> 
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>  include/exec/confidential-guest-support.h | 24 +++++++++++++++++++++++
>  softmmu/vl.c                              | 10 ++++++++++
>  target/i386/sev.c                         |  2 ++
>  3 files changed, 36 insertions(+)
> 
> diff --git a/include/exec/confidential-guest-support.h b/include/exec/confidential-guest-support.h
> index 3db6380e63..5dcf602047 100644
> --- a/include/exec/confidential-guest-support.h
> +++ b/include/exec/confidential-guest-support.h
> @@ -27,6 +27,30 @@ OBJECT_DECLARE_SIMPLE_TYPE(ConfidentialGuestSupport, CONFIDENTIAL_GUEST_SUPPORT)
>  
>  struct ConfidentialGuestSupport {
>      Object parent;
> +
> +    /*
> +     * ready: flag set by CGS initialization code once it's ready to
> +     *        start executing instructions in a potentially-secure
> +     *        guest
> +     *
> +     * The definition here is a bit fuzzy, because this is essentially
> +     * part of a self-sanity-check, rather than a strict mechanism.
> +     *
> +     * It's not fasible to have a single point in the common machine

s/fasible/feasible

Anyway,

Reviewed-by: Greg Kurz <groug@kaod.org>

> +     * init path to configure confidential guest support, because
> +     * different mechanisms have different interdependencies requiring
> +     * initialization in different places, often in arch or machine
> +     * type specific code.  It's also usually not possible to check
> +     * for invalid configurations until that initialization code.
> +     * That means it would be very easy to have a bug allowing CGS
> +     * init to be bypassed entirely in certain configurations.
> +     *
> +     * Silently ignoring a requested security feature would be bad, so
> +     * to avoid that we check late in init that this 'ready' flag is
> +     * set if CGS was requested.  If the CGS init hasn't happened, and
> +     * so 'ready' is not set, we'll abort.
> +     */
> +    bool ready;
>  };
>  
>  typedef struct ConfidentialGuestSupportClass {
> diff --git a/softmmu/vl.c b/softmmu/vl.c
> index 1b464e3474..1869ed54a9 100644
> --- a/softmmu/vl.c
> +++ b/softmmu/vl.c
> @@ -101,6 +101,7 @@
>  #include "qemu/plugin.h"
>  #include "qemu/queue.h"
>  #include "sysemu/arch_init.h"
> +#include "exec/confidential-guest-support.h"
>  
>  #include "ui/qemu-spice.h"
>  #include "qapi/string-input-visitor.h"
> @@ -2497,6 +2498,8 @@ static void qemu_create_cli_devices(void)
>  
>  static void qemu_machine_creation_done(void)
>  {
> +    MachineState *machine = MACHINE(qdev_get_machine());
> +
>      /* Did we create any drives that we failed to create a device for? */
>      drive_check_orphaned();
>  
> @@ -2516,6 +2519,13 @@ static void qemu_machine_creation_done(void)
>  
>      qdev_machine_creation_done();
>  
> +    if (machine->cgs) {
> +        /*
> +         * Verify that Confidential Guest Support has actually been initialized
> +         */
> +        assert(machine->cgs->ready);
> +    }
> +
>      if (foreach_device_config(DEV_GDB, gdbserver_start) < 0) {
>          exit(1);
>      }
> diff --git a/target/i386/sev.c b/target/i386/sev.c
> index 590cb31fa8..f9e9b5d8ae 100644
> --- a/target/i386/sev.c
> +++ b/target/i386/sev.c
> @@ -737,6 +737,8 @@ int sev_kvm_init(ConfidentialGuestSupport *cgs, Error **errp)
>      qemu_add_machine_init_done_notifier(&sev_machine_done_notify);
>      qemu_add_vm_change_state_handler(sev_vm_state_change, sev);
>  
> +    cgs->ready = true;
> +
>      return 0;
>  err:
>      sev_guest = NULL;