Re: [PATCH 4/4 V8] selftest: KVM: Add intra host migration tests

[Marc Orr]

On Tue, Sep 21, 2021 at 7:20 AM Peter Gonda <pgonda@xxxxxxxxxx> wrote:
>
> On Wed, Sep 15, 2021 at 11:28 AM Marc Orr <marcorr@xxxxxxxxxx> wrote:
> >
> > On Tue, Sep 14, 2021 at 9:47 AM Peter Gonda <pgonda@xxxxxxxxxx> wrote:
> > >
> > > Adds testcases for intra host migration for SEV and SEV-ES. Also adds
> > > locking test to confirm no deadlock exists.
> > >
> > > Signed-off-by: Peter Gonda <pgonda@xxxxxxxxxx>
> > > Suggested-by: Sean Christopherson <seanjc@xxxxxxxxxx>
> > > Reviewed-by: Marc Orr <marcorr@xxxxxxxxxx>
> > > Cc: Marc Orr <marcorr@xxxxxxxxxx>
> > > Cc: Sean Christopherson <seanjc@xxxxxxxxxx>
> > > Cc: David Rientjes <rientjes@xxxxxxxxxx>
> > > Cc: Brijesh Singh <brijesh.singh@xxxxxxx>
> > > Cc: kvm@xxxxxxxxxxxxxxx
> > > Cc: linux-kernel@xxxxxxxxxxxxxxx
> > > ---
> > >  tools/testing/selftests/kvm/Makefile          |   1 +
> > >  .../selftests/kvm/x86_64/sev_vm_tests.c       | 203 ++++++++++++++++++
> > >  2 files changed, 204 insertions(+)
> > >  create mode 100644 tools/testing/selftests/kvm/x86_64/sev_vm_tests.c
> > >
> > > diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile
> > > index c103873531e0..44fd3566fb51 100644
> > > --- a/tools/testing/selftests/kvm/Makefile
> > > +++ b/tools/testing/selftests/kvm/Makefile
> > > @@ -72,6 +72,7 @@ TEST_GEN_PROGS_x86_64 += x86_64/vmx_pmu_msrs_test
> > >  TEST_GEN_PROGS_x86_64 += x86_64/xen_shinfo_test
> > >  TEST_GEN_PROGS_x86_64 += x86_64/xen_vmcall_test
> > >  TEST_GEN_PROGS_x86_64 += x86_64/vmx_pi_mmio_test
> > > +TEST_GEN_PROGS_x86_64 += x86_64/sev_vm_tests
> > >  TEST_GEN_PROGS_x86_64 += access_tracking_perf_test
> > >  TEST_GEN_PROGS_x86_64 += demand_paging_test
> > >  TEST_GEN_PROGS_x86_64 += dirty_log_test
> > > diff --git a/tools/testing/selftests/kvm/x86_64/sev_vm_tests.c b/tools/testing/selftests/kvm/x86_64/sev_vm_tests.c
> > > new file mode 100644
> > > index 000000000000..ec3bbc96e73a
> > > --- /dev/null
> > > +++ b/tools/testing/selftests/kvm/x86_64/sev_vm_tests.c
> > > @@ -0,0 +1,203 @@
> > > +// SPDX-License-Identifier: GPL-2.0-only
> > > +#include <linux/kvm.h>
> > > +#include <linux/psp-sev.h>
> > > +#include <stdio.h>
> > > +#include <sys/ioctl.h>
> > > +#include <stdlib.h>
> > > +#include <errno.h>
> > > +#include <pthread.h>
> > > +
> > > +#include "test_util.h"
> > > +#include "kvm_util.h"
> > > +#include "processor.h"
> > > +#include "svm_util.h"
> > > +#include "kselftest.h"
> > > +#include "../lib/kvm_util_internal.h"
> > > +
> > > +#define SEV_POLICY_ES 0b100
> > > +
> > > +#define NR_MIGRATE_TEST_VCPUS 4
> > > +#define NR_MIGRATE_TEST_VMS 3
> > > +#define NR_LOCK_TESTING_THREADS 3
> > > +#define NR_LOCK_TESTING_ITERATIONS 10000
> > > +
> > > +static void sev_ioctl(int vm_fd, int cmd_id, void *data)
> > > +{
> > > +       struct kvm_sev_cmd cmd = {
> > > +               .id = cmd_id,
> > > +               .data = (uint64_t)data,
> > > +               .sev_fd = open_sev_dev_path_or_exit(),
> > > +       };
> > > +       int ret;
> > > +
> > > +       ret = ioctl(vm_fd, KVM_MEMORY_ENCRYPT_OP, &cmd);
> > > +       TEST_ASSERT((ret == 0 || cmd.error == SEV_RET_SUCCESS),
> > > +                   "%d failed: return code: %d, errno: %d, fw error: %d",
> > > +                   cmd_id, ret, errno, cmd.error);
> > > +}
> > > +
> > > +static struct kvm_vm *sev_vm_create(bool es)
> > > +{
> > > +       struct kvm_vm *vm;
> > > +       struct kvm_sev_launch_start start = { 0 };
> > > +       int i;
> > > +
> > > +       vm = vm_create(VM_MODE_DEFAULT, 0, O_RDWR);
> > > +       sev_ioctl(vm->fd, es ? KVM_SEV_ES_INIT : KVM_SEV_INIT, NULL);
> > > +       for (i = 0; i < NR_MIGRATE_TEST_VCPUS; ++i)
> > > +               vm_vcpu_add(vm, i);
> > > +       if (es)
> > > +               start.policy |= SEV_POLICY_ES;
> > > +       sev_ioctl(vm->fd, KVM_SEV_LAUNCH_START, &start);
> > > +       if (es)
> > > +               sev_ioctl(vm->fd, KVM_SEV_LAUNCH_UPDATE_VMSA, NULL);
> > > +       return vm;
> > > +}
> >
> > I should've suggested this in my original review. But is it worth
> > moving `sev_vm_create()` and `sev_ioctl()` into the broader selftests
> > library, so others can leverage this function to write selftests?
>
> This function isn't fully complete. It doesn't get to launch_finish,
> i.e. it only goes far enough for copyless migration ioctls to work. I
> think this would be a good expansion but could happen in follow up
> series, thoughts?

SGTM. Let's leave it here for now then.

>
> >
> > > +
> > > +static struct kvm_vm *__vm_create(void)
> > > +{
> > > +       struct kvm_vm *vm;
> > > +       int i;
> > > +
> > > +       vm = vm_create(VM_MODE_DEFAULT, 0, O_RDWR);
> > > +       for (i = 0; i < NR_MIGRATE_TEST_VCPUS; ++i)
> > > +               vm_vcpu_add(vm, i);
> > > +
> > > +       return vm;
> > > +}
> > > +
> > > +static int __sev_migrate_from(int dst_fd, int src_fd)
> > > +{
> > > +       struct kvm_enable_cap cap = {
> > > +               .cap = KVM_CAP_VM_MIGRATE_PROTECTED_VM_FROM,
> > > +               .args = { src_fd }
> > > +       };
> > > +
> > > +       return ioctl(dst_fd, KVM_ENABLE_CAP, &cap);
> > > +}
> > > +
> > > +
> > > +static void sev_migrate_from(int dst_fd, int src_fd)
> > > +{
> > > +       int ret;
> > > +
> > > +       ret = __sev_migrate_from(dst_fd, src_fd);
> > > +       TEST_ASSERT(!ret, "Migration failed, ret: %d, errno: %d\n", ret, errno);
> > > +}
> > > +
> > > +static void test_sev_migrate_from(bool es)
> > > +{
> > > +       struct kvm_vm *src_vm;
> > > +       struct kvm_vm *dst_vms[NR_MIGRATE_TEST_VMS];
> > > +       int i;
> > > +
> > > +       src_vm = sev_vm_create(es);
> > > +       for (i = 0; i < NR_MIGRATE_TEST_VMS; ++i)
> > > +               dst_vms[i] = __vm_create();
> > > +
> > > +       /* Initial migration from the src to the first dst. */
> > > +       sev_migrate_from(dst_vms[0]->fd, src_vm->fd);
> > > +
> > > +       for (i = 1; i < NR_MIGRATE_TEST_VMS; i++)
> > > +               sev_migrate_from(dst_vms[i]->fd, dst_vms[i - 1]->fd);
> > > +
> > > +       /* Migrate the guest back to the original VM. */
> > > +       sev_migrate_from(src_vm->fd, dst_vms[NR_MIGRATE_TEST_VMS - 1]->fd);
> > > +
> > > +       kvm_vm_free(src_vm);
> > > +       for (i = 0; i < NR_MIGRATE_TEST_VMS; ++i)
> > > +               kvm_vm_free(dst_vms[i]);
> > > +}
> > > +
> > > +struct locking_thread_input {
> > > +       struct kvm_vm *vm;
> > > +       int source_fds[NR_LOCK_TESTING_THREADS];
> > > +};
> > > +
> > > +static void *locking_test_thread(void *arg)
> > > +{
> > > +       int i, j;
> > > +       struct locking_thread_input *input = (struct locking_test_thread *)arg;
> > > +
> > > +       for (i = 0; i < NR_LOCK_TESTING_ITERATIONS; ++i) {
> > > +               j = i % NR_LOCK_TESTING_THREADS;
> > > +               __sev_migrate_from(input->vm->fd, input->source_fds[j]);
> > > +       }
> > > +
> > > +       return NULL;
> > > +}
> > > +
> > > +static void test_sev_migrate_locking(void)
> > > +{
> > > +       struct locking_thread_input input[NR_LOCK_TESTING_THREADS];
> > > +       pthread_t pt[NR_LOCK_TESTING_THREADS];
> > > +       int i;
> > > +
> > > +       for (i = 0; i < NR_LOCK_TESTING_THREADS; ++i) {
> > > +               input[i].vm = sev_vm_create(/* es= */ false);
> > > +               input[0].source_fds[i] = input[i].vm->fd;
> > > +       }
> > > +       for (i = 1; i < NR_LOCK_TESTING_THREADS; ++i)
> > > +               memcpy(input[i].source_fds, input[0].source_fds,
> > > +                      sizeof(input[i].source_fds));
> > > +
> > > +       for (i = 0; i < NR_LOCK_TESTING_THREADS; ++i)
> > > +               pthread_create(&pt[i], NULL, locking_test_thread, &input[i]);
> > > +
> > > +       for (i = 0; i < NR_LOCK_TESTING_THREADS; ++i)
> > > +               pthread_join(pt[i], NULL);
> > > +}
> > > +
> > > +static void test_sev_migrate_parameters(void)
> > > +{
> > > +       struct kvm_vm *sev_vm, *sev_es_vm, *vm_no_vcpu, *vm_no_sev,
> > > +               *sev_es_vm_no_vmsa;
> > > +       int ret;
> > > +
> > > +       sev_vm = sev_vm_create(/* es= */ false);
> > > +       sev_es_vm = sev_vm_create(/* es= */ true);
> > > +       vm_no_vcpu = vm_create(VM_MODE_DEFAULT, 0, O_RDWR);
> > > +       vm_no_sev = __vm_create();
> > > +       sev_es_vm_no_vmsa = vm_create(VM_MODE_DEFAULT, 0, O_RDWR);
> > > +       sev_ioctl(sev_es_vm_no_vmsa->fd, KVM_SEV_ES_INIT, NULL);
> > > +       vm_vcpu_add(sev_es_vm_no_vmsa, 1);
> > > +
> > > +
> > > +       ret = __sev_migrate_from(sev_vm->fd, sev_es_vm->fd);
> > > +       TEST_ASSERT(
> > > +               ret == -1 && errno == EINVAL,
> > > +               "Should not be able migrate to SEV enabled VM. ret: %d, errno: %d\n",
> > > +               ret, errno);
> > > +
> > > +       ret = __sev_migrate_from(sev_es_vm->fd, sev_vm->fd);
> > > +       TEST_ASSERT(
> > > +               ret == -1 && errno == EINVAL,
> > > +               "Should not be able migrate to SEV-ES enabled VM. ret: %d, errno: %d\n",
> > > +               ret, errno);
> > > +
> > > +       ret = __sev_migrate_from(vm_no_vcpu->fd, sev_es_vm->fd);
> > > +       TEST_ASSERT(
> > > +               ret == -1 && errno == EINVAL,
> > > +               "SEV-ES migrations require same number of vCPUS. ret: %d, errno: %d\n",
> > > +               ret, errno);
> >
> > How do we know that this failed because `vm_no_vcpu` has no vCPUs or
> > because it's not a SEV-ES VM?
>
> Actually with V8 we only migrate to none SEV(-ES)? enabled guests.

I think my point is that the test case should be written to treat the
underlying KVM code as a black box. Without looking at the KVM code,
the test case should be setup to be accepted perfectly by KVM and then
mutated in a minimal way to trigger the intended failure case.

Here, we've defined `vm_no_vcpu`, which as far as I can tell is: (1)
not a SEV VM, (2) not a SEV-ES VM, (3) has no vCPUs. Based on the
error message in the TEST_ASSERT, the intention here is to verify that
a migration that would otherwise works, fails because the target has a
different number of vCPUs than the source. Therefore, I think
`vm_no_vcpu` should be defined as a SEV-ES VM, so that the test case
is setup such that it would've otherwise passed if `vm_no_vcpu` had
the correct number of vCPUs added.

>
> >
> > > +
> > > +       ret = __sev_migrate_from(vm_no_vcpu->fd, sev_es_vm_no_vmsa->fd);
> > > +       TEST_ASSERT(
> > > +               ret == -1 && errno == EINVAL,
> > > +               "SEV-ES migrations require UPDATE_VMSA. ret %d, errno: %d\n",
> > > +               ret, errno);
> >
> > Same question. How do we know why this failed? `sev_es_vm_no_vmsa` did
> > not have any vCPUs added. Would it be cleaner to add an additional
> > param to `sev_vm_create()` to skip calling UPDATE_VMSA? Then,
> > `sev_es_vm_no_vmsa` can be created from `sev_vm_create()` and it's
> > obvious to the read that the VMs are identical except for this aspect.
> >
> > > +
> > > +       ret = __sev_migrate_from(vm_no_vcpu->fd, vm_no_sev->fd);
> > > +       TEST_ASSERT(ret == -1 && errno == EINVAL,
> > > +                   "Migrations require SEV enabled. ret %d, errno: %d\n", ret,
> > > +                   errno);
> >
> > `vm_no_sev` has vCPUs. Therefore, how do we know why this failed --
> > (a) differing vCPU counts or (b) no SEV?
>
> Ditto we require dst to be none SEV enabled.

Understood. But I think the test should treat KVM as a black box.
Therefore, I think in this test case, `vm_no_vcpu` should be defined
to have the same number of vCPUs as `vm_no_sev`.

>
> >
> > > +}
> > > +
> > > +int main(int argc, char *argv[])
> > > +{
> > > +       test_sev_migrate_from(/* es= */ false);
> > > +       test_sev_migrate_from(/* es= */ true);
> > > +       test_sev_migrate_locking();
> > > +       test_sev_migrate_parameters();
> > > +       return 0;
> > > +}
> > > --
> > > 2.33.0.309.g3052b89438-goog
> > >