Re: [PATCH] Input: uinput - fix circular locking dependency with ff-core

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Mon, Mar 23, 2026 at 10:17:01AM +0500, Mikhail Gavrilov wrote:
> On Mon, Mar 23, 2026 at 7:47 AM Dmitry Torokhov
> <dmitry.torokhov@gmail.com> wrote:
> >
> > Thank you for the patch, it looks solid, however I wonder if creating a
> > separate "state_lock" spinlock would not be better than reusing
> > requests_lock?
> 
> Hi Dmitry,
> 
> Thank you for the review.
> 
> A separate spinlock would certainly be cleaner from a naming
> perspective. One thing I'd like to note though: the current
> approach of reusing requests_lock has the benefit of atomically
> checking state and queueing the event in uinput_request_send(),
> and atomically changing state and flushing requests in
> uinput_destroy_device(). With a separate state_lock these become
> two independent locks, so the ordering between them would need to
> be defined.

Hmm, I am not sure I see the issue. We are not going to change state
back to UIST_CREATED until after uinput_destroy_device() returns so we
will not submit more requests...

What am I missing?

Thanks.

-- 
Dmitry