On Mon Mar 30, 2026 at 4:19 PM CEST, Bartosz Golaszewski wrote:

On Mon, Mar 30, 2026 at 3:47 PM Danilo Krummrich [off-list ref] wrote:

quoted

On Mon Mar 30, 2026 at 2:40 PM CEST, Bartosz Golaszewski wrote:

quoted

diff --git a/include/linux/kobject.h b/include/linux/kobject.h
index c8219505a79f98bc370e52997efc8af51833cfda..71b9086621c35b7e4ef99b9d3b6707db23faf58c 100644
--- a/include/linux/kobject.h
+++ b/include/linux/kobject.h

@@ -219,4 +219,6 @@ int kobject_synth_uevent(struct kobject *kobj, const char *buf, size_t count);
 __printf(2, 3)
 int add_uevent_var(struct kobj_uevent_env *env, const char *format, ...);

+void ksysfs_init(void);

NIT: I'm aware there's also all the core kobjects in include/linux/kobject.h,
but maybe a separate header would be a better fit.

Do you mean moving all the top-level kobject declarations
(kernel_kobj, firmware_kobj, etc.) out of kobject.h into this new
header (ksysfs.h?) along with their init functions?

I think the top-level kobjects are fine; it's just ksysfs_init() that somehow
feels odd to me being placed in kobject.h.

The top-level kobject do make sense as they are the base for a lot of other
kobjects being created by other core code.

Whereas ksysfs_init() is a ksysfs specific thing that is only ever used by
init/main.c, i.e. other than the top-level kobjects, it has nothing to do with
the kobject API itself.