Re: [PATCH net-next 0/2] net/sched: don't wait for RCU when destroying qdisc lockdep keys

[Boqun Feng]

On Tue, Oct 20, 2026 at 10:26:31AM -0700, Bart Van Assche wrote:
> On 10/19/26 12:20 PM, Boqun Feng wrote:
> > I think my concern is still valid: for lockdep PoV,
> > lockdep_unregister_key_nosync() is an API that users have to use it
> > carefully, because it leaves the unregistration process in a stage that
> > you need to guarantee that the lockdep key reclamation is delayed by an
> > RCU grace period. Or in other words, do you have a way to prevent a user
> > from doing lockdep_unregister_key_nosync() but kfree(key), or some sort
> > of debugging to find a mis-use?
> 
> Kernel developers are assumed to know what they are doing. A clear comment

Well, if that's true, why do you need lockdep :) If developers know what
they are doing, they should write no deadlock or prove that in a
comment.

Is there any context that I'm missing that you want to shut up this
discussion and rush the change in?

Regards,
Boqun

> above lockdep_unregister_key_nosync() should be sufficient.
> 
> Thanks,
> 
> Bart.