On 9/7/23 19:05, Philippe Mathieu-Daudé wrote:
On 7/9/23 02:35, Gavin Shan wrote:
The names of supported CPU models instead of CPU types should be
printed when the user specified CPU type isn't supported, to be
consistent with the output from '-cpu ?'.

Correct the error messages to print CPU model names instead of CPU
type names.

Signed-off-by: Gavin Shan <gs...@redhat.com>
---
  hw/core/machine.c | 16 ++++++++++++----
  1 file changed, 12 insertions(+), 4 deletions(-)


@@ -1373,11 +1374,18 @@ static void is_cpu_type_supported(MachineState 
*machine, Error **errp)
          /* The user specified CPU type isn't valid */
          if (!mc->valid_cpu_types[i]) {
-            error_setg(errp, "Invalid CPU type: %s", machine->cpu_type);
-            error_append_hint(errp, "The valid types are: %s",
-                              mc->valid_cpu_types[0]);
+            model = cpu_model_from_type(machine->cpu_type);
+            error_setg(errp, "Invalid CPU type: %s", model);
+            g_free(model);
+
+            model = cpu_model_from_type(mc->valid_cpu_types[0]);
+            error_append_hint(errp, "The valid types are: %s", model);
+            g_free(model);
+
              for (i = 1; mc->valid_cpu_types[i]; i++) {
-                error_append_hint(errp, ", %s", mc->valid_cpu_types[i]);
+                model = cpu_model_from_type(mc->valid_cpu_types[i]);

cpu_model_from_type() can return NULL:

  char *cpu_model_from_type(const char *typename)
  {
      const char *suffix = "-" CPU_RESOLVING_TYPE;

      if (!object_class_by_name(typename)) {
          return NULL;
      }

Don't we want to skip that case?

                    if (!model) {
                        continue;
                    }


No, it's intentional. "(null)" will be printed in this specific case so that
it can be identified quickly and mc->valid_cpu_types[] need to be fixed by
developers.

+                error_append_hint(errp, ", %s", model);
+                g_free(model);
              }
              error_append_hint(errp, "\n");


Thanks,
Gavin


Reply via email to