On Wed, Jul 21, 2021 at 11:54:45AM +0200, Gerd Hoffmann wrote:
> > ObjectClass *module_object_class_by_name(const char *typename)
> > {
> > ObjectClass *oc;
> > @@ -1031,8 +1049,20 @@ ObjectClass *module_object_class_by_name(const char
> > *typename)
> > oc = object_class_by_name(typename);
> > #ifdef CONFIG_MODULES
> > if (!oc) {
> > + char *module_name;
> > module_load_qom_one(typename);
> > oc = object_class_by_name(typename);
> > + module_name = get_accel_module_name(typename);
> > + if (module_name) {
> > + if (!module_is_loaded(module_name)) {
> > + fprintf(stderr, "%s module is missing, install the "
> > + "package or config the library path "
> > + "correctly.\n", module_name);
> > + g_free(module_name);
> > + exit(1);
> > + }
> > + g_free(module_name);
> > + }
>
> This error logging should IMHO be moved to util/module.c. Either have a
> helper function to print the error message, or have
> module_load_qom_one() print it.
>
> There is also no need to hard-code the module names. We have the module
> database and module_load_qom_one() uses it to figure which module must
> be loaded for a specific qom object. We can likewise use the database
> for printing the error message.
IIUC, module loading can be triggered from hotplug of backends/devices,
If so, we really shouldn't be printing to stderr, but using Error *errp
throughout, so that QMP can give back useful error messages
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
