Received-SPF: pass (google.com: domain of linux-kernel+bounces-180395-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240514160926.1309778-1-cmllamas@google.com> <CAH5fLgjzVHSAQBF5-C1BthK1jspAGRr1z4SQXdQepkeVL9Eq2A@mail.gmail.com>
 <ZkTlG-ZNHRYXnHLQ@google.com> <ZkTmNLxhlPlHdQCW@google.com>
In-Reply-To: <ZkTmNLxhlPlHdQCW@google.com>
From: Alice Ryhl <aliceryhl@google.com>
Date: Wed, 15 May 2024 23:42:11 +0200
Message-ID: <CAH5fLgi1YUE7DHV86GcZaaUKX+RDkOXwHqKBRVR3YFyj0T6dLQ@mail.gmail.com>
Subject: Re: [PATCH v2] binder: use bitmap for faster descriptor lookup
To: Carlos Llamas <cmllamas@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>, =?UTF-8?B?QXJ2ZSBIasO4bm5ldsOlZw==?= <arve@android.com>, 
	Todd Kjos <tkjos@android.com>, Martijn Coenen <maco@android.com>, 
	Joel Fernandes <joel@joelfernandes.org>, Christian Brauner <brauner@kernel.org>, 
	Suren Baghdasaryan <surenb@google.com>, linux-kernel@vger.kernel.org, kernel-team@android.com, 
	Tim Murray <timmurray@google.com>, John Stultz <jstultz@google.com>, 
	Steven Moreland <smoreland@google.com>, Nick Chen <chenjia3@oppo.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, May 15, 2024 at 6:43=E2=80=AFPM Carlos Llamas <cmllamas@google.com>=
 wrote:
>
> On Wed, May 15, 2024 at 04:38:51PM +0000, Carlos Llamas wrote:
> > On Wed, May 15, 2024 at 05:29:29PM +0200, Alice Ryhl wrote:
> > > On Tue, May 14, 2024 at 6:09=E2=80=AFPM Carlos Llamas <cmllamas@googl=
e.com> wrote:
> > > > +static inline int
> > > > +dbitmap_get_first_zero_bit(struct dbitmap *dmap, unsigned long *bi=
t)
> > > > +{
> > > > +       unsigned long n;
> > > > +
> > > > +       n =3D find_first_zero_bit(dmap->map, dmap->nbits);
> > > > +       if (unlikely(n =3D=3D dmap->nbits))
> > > > +               return -ENOSPC;
> > > > +
> > > > +       *bit =3D n;
> > > > +       set_bit(n, dmap->map);
> > > > +
> > > > +       return 0;
> > > > +}
> > >
> > > Could we rename this method to something that makes it more clear tha=
t
> > > it's not just a getter, but that it actually also sets the bit?
> > >
> > > Alice
> >
> > Sure, what were you thinking? I had picked "get" and not just "find" to
> > indicate this behavior. However, I'll take any better ideas. The option
> > of dbitmap_find_and_set_first_zero_bit() seemed too long for me.
> >
> > --
> > Carlos Llamas
>
> I like dbitmap_acquire_first_zero_bit(). Sounds better?

acquire sounds good!

Alice