Re: make_task_dead() & kthread_exit()

[Alice Ryhl <aliceryhl@google.com>]

Christian Brauner <brauner@kernel.org> writes:
> Oh nice.
> I was kinda hoping Tejun would jump on this one and so just pointed to
> one potential way to fix it but didn't really spend time on it.
> 
> Anyway, let's just take what you proposed and slap a commit message on
> it. Fwiw, init_task does have ->worker_private it just gets set later
> during sched_init():
> 
>           /*
>            * The idle task doesn't need the kthread struct to function, but it
>            * is dressed up as a per-CPU kthread and thus needs to play the part
>            * if we want to avoid special-casing it in code that deals with per-CPU
>            * kthreads.
>            */
>           WARN_ON(!set_kthread_struct(current));
> 
> I think that @current here is misleading. When sched_init() runs it
> should be single-threaded still and current == &init_task. So that
> set_kthread_struct(current) call sets @init_task's worker_private iiuc.
> 
> Patch appended. I'll stuff it into vfs.fixes.

welp, I guess you get another one here:

Tested-by: Alice Ryhl <aliceryhl@google.com>

from:
https://lore.kernel.org/all/aaaxK5gFWN70WmMn@google.com/

Alice