Re: [Qemu-devel] [QEMU PATCH v4 10/10] target/i386: kvm: Add nested migration blocker only when kernel lacks required capabilities

[Maran Wilson]

On 6/19/2019 9:21 AM, Liran Alon wrote:

Previous commits have added support for migration of nested virtualization
workloads. This was done by utilising two new KVM capabilities:
KVM_CAP_NESTED_STATE and KVM_CAP_EXCEPTION_PAYLOAD. Both which are
required in order to correctly migrate such workloads.

Therefore, change code to add a migration blocker for vCPUs exposed with
Intel VMX or AMD SVM in case one of these kernel capabilities is
missing.

Signed-off-by: Liran Alon 
---
  target/i386/kvm.c | 9 +++--
  1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/target/i386/kvm.c b/target/i386/kvm.c
index 99480a52ad33..a3d0fbed3b35 100644
--- a/target/i386/kvm.c
+++ b/target/i386/kvm.c
@@ -1313,9 +1313,14 @@ int kvm_arch_init_vcpu(CPUState *cs)
!!(c->ecx & CPUID_EXT_SMX);
  }
  
-if (cpu_has_nested_virt(env) && !nested_virt_mig_blocker) {

+if (cpu_has_nested_virt(env) && !nested_virt_mig_blocker &&
+((kvm_max_nested_state_length() <= 0) || !has_exception_payload)) {
  error_setg(_virt_mig_blocker,
-   "Nested virtualization does not support live migration 
yet");
+   "Kernel do not provide required capabilities for "


s/do/does/

And with that change:

Reviewed-by: Maran Wilson 

Thanks,
-Maran



+   "nested virtualization migration. "
+   "(CAP_NESTED_STATE=%d, CAP_EXCEPTION_PAYLOAD=%d)",
+   kvm_max_nested_state_length() > 0,
+   has_exception_payload);
  r = migrate_add_blocker(nested_virt_mig_blocker, _err);
  if (local_err) {
  error_report_err(local_err);

[Qemu-devel] [QEMU PATCH v4 10/10] target/i386: kvm: Add nested migration blocker only when kernel lacks required capabilities

[Liran Alon]

Previous commits have added support for migration of nested virtualization
workloads. This was done by utilising two new KVM capabilities:
KVM_CAP_NESTED_STATE and KVM_CAP_EXCEPTION_PAYLOAD. Both which are
required in order to correctly migrate such workloads.

Therefore, change code to add a migration blocker for vCPUs exposed with
Intel VMX or AMD SVM in case one of these kernel capabilities is
missing.

Signed-off-by: Liran Alon 
---
 target/i386/kvm.c | 9 +++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/target/i386/kvm.c b/target/i386/kvm.c
index 99480a52ad33..a3d0fbed3b35 100644
--- a/target/i386/kvm.c
+++ b/target/i386/kvm.c
@@ -1313,9 +1313,14 @@ int kvm_arch_init_vcpu(CPUState *cs)
   !!(c->ecx & CPUID_EXT_SMX);
 }
 
-if (cpu_has_nested_virt(env) && !nested_virt_mig_blocker) {
+if (cpu_has_nested_virt(env) && !nested_virt_mig_blocker &&
+((kvm_max_nested_state_length() <= 0) || !has_exception_payload)) {
 error_setg(_virt_mig_blocker,
-   "Nested virtualization does not support live migration 
yet");
+   "Kernel do not provide required capabilities for "
+   "nested virtualization migration. "
+   "(CAP_NESTED_STATE=%d, CAP_EXCEPTION_PAYLOAD=%d)",
+   kvm_max_nested_state_length() > 0,
+   has_exception_payload);
 r = migrate_add_blocker(nested_virt_mig_blocker, _err);
 if (local_err) {
 error_report_err(local_err);
-- 
2.20.1