Am 10.10.23 um 13:12 schrieb Thomas Huth:
On 10/10/2023 13.02, Christian Borntraeger wrote:


Am 09.10.23 um 19:07 schrieb Thomas Huth:
Since we already require at least kernel 3.15 in the s390x KVM code,
we can assume that the KVM_CAP_SYNC_REGS capability is always there.
Thus turn this into a hard requirement now.

Signed-off-by: Thomas Huth <th...@redhat.com>
---
  target/s390x/kvm/kvm.c | 20 ++++++++++++++------
  1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/target/s390x/kvm/kvm.c b/target/s390x/kvm/kvm.c
index bc5c56a305..b3e2eaa2eb 100644
--- a/target/s390x/kvm/kvm.c
+++ b/target/s390x/kvm/kvm.c
@@ -337,21 +337,29 @@ int kvm_arch_get_default_type(MachineState *ms)
  int kvm_arch_init(MachineState *ms, KVMState *s)
  {
+    int required_caps[] = {
+        KVM_CAP_DEVICE_CTRL,
+        KVM_CAP_SYNC_REGS,
+    };
+
+    for (int i = 0; i < ARRAY_SIZE(required_caps); i++) {
+        if (!kvm_check_extension(s, required_caps[i])) {
+            error_report("KVM is missing capability #%d - "
+                         "please use kernel 3.15 or newer", required_caps[i]);
+            return -1;
+        }
+    }
+
      object_class_foreach(ccw_machine_class_foreach, TYPE_S390_CCW_MACHINE,
                           false, NULL);
-    if (!kvm_check_extension(kvm_state, KVM_CAP_DEVICE_CTRL)) {
-        error_report("KVM is missing capability KVM_CAP_DEVICE_CTRL - "
-                     "please use kernel 3.15 or newer");
-        return -1;
-    }
      if (!kvm_check_extension(s, KVM_CAP_S390_COW)) {
          error_report("KVM is missing capability KVM_CAP_S390_COW - "
                       "unsupported environment");
          return -1;
      }

Not sure if we also want to move KVM_CAP_S390_COW somehow. The message would be 
different.

IIRC that error could happen when you ran KVM within an older version of z/VM, so the 
"please use kernel 3.15 or newer" message would be completely misleading there.

Yes, thats what I was trying to say, we would need a different message.
Lets go with this patch.

Aparch from that:
Reviewed-by: Christian Borntraeger <borntrae...@linux.ibm.com>

Thanks,
   Thomas



Reply via email to