MIME-Version: 1.0
References: <20230228175929.7534-1-ubizjak@gmail.com> <20230228175929.7534-4-ubizjak@gmail.com>
 <20230228164346.0691bb11@gandalf.local.home> <CAFULd4aGx=kGYYUz0BkFJz3abz97WOwDEgFHAOocT8SWT2oX-Q@mail.gmail.com>
 <20230301111850.768f8526@gandalf.local.home> <CAFULd4YePGAqxpq0kJMnNHkqqkL9HEnEz1ud9kNrae5wqpRVbA@mail.gmail.com>
 <20230301131831.6c8d4ff5@gandalf.local.home> <20230301132839.43b51b4e@gandalf.local.home>
In-Reply-To: <20230301132839.43b51b4e@gandalf.local.home>
From:   Uros Bizjak <ubizjak@gmail.com>
Date:   Wed, 1 Mar 2023 19:37:06 +0100
Message-ID: <CAFULd4bu07FnVEtRPB77hAEdm_Qyx9dH26JYx9U8q+3w6BBo8A@mail.gmail.com>
Subject: Re: [PATCH 3/3] ring_buffer: Use try_cmpxchg instead of cmpxchg
To:     Steven Rostedt <rostedt@goodmis.org>
Cc:     linux-trace-kernel@vger.kernel.org, linux-kernel@vger.kernel.org,
        Masami Hiramatsu <mhiramat@kernel.org>,
        "Paul E. McKenney" <paulmck@kernel.org>,
        Joel Fernandes <joel@joelfernandes.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Mar 1, 2023 at 7:28=E2=80=AFPM Steven Rostedt <rostedt@goodmis.org>=
 wrote:
>
> On Wed, 1 Mar 2023 13:18:31 -0500
> Steven Rostedt <rostedt@goodmis.org> wrote:
>
> > I am leaning to just taking this, and hope that it doesn't cause proble=
ms.
> > And I would really love to change all the local_cmpxchg() to
> > local_try_cmpxchg(). Hmm, I wonder how much of an effort that would be =
to
> > implement those?
>
> I see that you were the one that added the generic support for
> try_cmpxchg64() and messed with all those generated files. Care to add on=
e
> for local_try_cmpxchg() ;-)

I already have a half-written patch that implements local_try_cmpxchg.
Half-written in the sense that above-mentioned scripts generate
correct locking primitives, but the fallback code for local_cmpxchg is
a bit different that the fallback code for cmpxchg. There is an effort
to unify all cmpxchg stuff (cmpxchg128 will be introduced) and I think
that local_cmpxchg should also be handled in the same way. So, if
there is interest, I can find some time to finish the infrastructure
patch and convert the uses. But this is quite an undertaking that will
take some time.

Uros.