Re: [PATCH v5 00/21] kprobes: Unify kretprobe trampoline handlers and make kretprobe lockless

From: Masami Hiramatsu
Date: Wed Sep 09 2020 - 01:28:55 EST

On Wed, 9 Sep 2020 00:09:23 +0900
Masami Hiramatsu <mhiramat@xxxxxxxxxx> wrote:

> > > Of course make it lockless then warning is gone.
> > > But even without the lockless patch, this warning can be false-positive
> > > because we prohibit nested kprobe call, right?
> >
> > Yes, because the actual nesting is avoided by kprobe_busy, but lockdep
> > can't tell. Lockdep sees a regular lock user and an in-nmi lock user and
> > figures that's a bad combination.

Hmm, what about introducing new LOCK_USED_KPROBE bit, which will be set
if the lock is accessed when the current_kprobe is set (including kprobe_busy)?
This means it is in the kprobe user-handler context. If we access the lock always
in the kprobes context, it is never nested.

Thank you,

Masami Hiramatsu <mhiramat@xxxxxxxxxx>