On Tue, Oct 17, 2023 at 11:13:53AM +0200, Uwe Kleine-König wrote: > I don't understand what you intend to say here. What is "that"? Are you > talking about > > owner && !try_module_get(owner) > > vs > > !try_module_get(owner) > > or the following line with kobject_get_unless_zero? Do you doubt the > validity of my patch, or is it about something try_module_get() > could/should do more than it currently does? I'm saying that it would be a good idea to turn try_module_get() into an inline in all cases, including the CONFIG_MODULE_UNLOAD one. Turning it into something like !module || __try_module_get(module), with the latter being out of line. With that done, your patch would be entirely unobjectionable...
