Received-SPF: pass (google.com: domain of linux-kernel+bounces-60945-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <20231208073119.29283-1-dinghao.liu@zju.edu.cn>
 <20231210133228.5fd425ea@jic23-huawei> <20231217132800.27e83a01@jic23-huawei>
 <ZckdEDbAqin1Fsgt@surfacebook.localdomain> <20240211193711.518b76d7@jic23-huawei>
In-Reply-To: <20240211193711.518b76d7@jic23-huawei>
From: Andy Shevchenko <andy.shevchenko@gmail.com>
Date: Sun, 11 Feb 2024 22:19:46 +0200
Message-ID: <CAHp75VdEwxVG3qnZJgx_VKZPOkRcqHUiLcvJj3cBAnJh90Wt0g@mail.gmail.com>
Subject: Re: [PATCH] iio: core: fix memleak in iio_device_register_sysfs
To: Jonathan Cameron <jic23@kernel.org>
Cc: Dinghao Liu <dinghao.liu@zju.edu.cn>, Lars-Peter Clausen <lars@metafoo.de>, 
	Alexandru Ardelean <alexandru.ardelean@analog.com>, linux-iio@vger.kernel.org, 
	linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sun, Feb 11, 2024 at 9:37=E2=80=AFPM Jonathan Cameron <jic23@kernel.org>=
 wrote:
> On Sun, 11 Feb 2024 21:16:32 +0200
> andy.shevchenko@gmail.com wrote:
>
> > Sun, Dec 17, 2023 at 01:28:00PM +0000, Jonathan Cameron kirjoitti:
> > > On Sun, 10 Dec 2023 13:32:28 +0000
> > > Jonathan Cameron <jic23@kernel.org> wrote:
> > >
> > > > On Fri,  8 Dec 2023 15:31:19 +0800
> > > > Dinghao Liu <dinghao.liu@zju.edu.cn> wrote:
> > > >
> > > > > When iio_device_register_sysfs_group() fails, we should
> > > > > free iio_dev_opaque->chan_attr_group.attrs to prevent
> > > > > potential memleak.
> > > > >
> > > > > Fixes: 32f171724e5c ("iio: core: rework iio device group creation=
")
> > > > > Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
> > > > Hi.
> > > >
> > > > Looks good to me, but I'd like to leave this one on the list a litt=
le
> > > > longer to see if anyone else has comments.
> > > >
> > > Guess no comments!
> >
> > This patch does not fix anything.
> >
> > Yet, it might be considered as one that increases robustness, but with =
this applied the
> > goto
> > https://elixir.bootlin.com/linux/v6.8-rc3/source/drivers/iio/industrial=
io-core.c#L2007
> > can be amended, right?
> >
>
> I'm lost.  That goto results in a call to
> iio_buffers_free_sysfs_and_mask(indio_dev);
> which continues to undo stuff from before that call.
> Now if it did
> iio_device_unregister_sysfs(indio_dev);
> (as per the label above it in the cleanup) then I'd agree.

Ah, it's me who hasn't noticed the word "buffer" in the goto label.
Yes, you are right!

> Perhaps I'm misreading the code flow here?
>
> All this code is supposed to be side effect free on error so I'm
> keen on the change even if there is another path where it gets cleaned
> up that I'm missing.

Everything is fine, sorry for the noise.

--=20
With Best Regards,
Andy Shevchenko