Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <1620706164-28856-1-git-send-email-dillon.minfei@gmail.com> <2193bb5c-89c3-cb1a-f2ef-1389ccfb7d61@xs4all.nl>
In-Reply-To: <2193bb5c-89c3-cb1a-f2ef-1389ccfb7d61@xs4all.nl>
From:   Dillon Min <dillon.minfei@gmail.com>
Date:   Wed, 26 May 2021 23:13:19 +0800
Message-ID: <CAL9mu0+bTrVsUVmZv8jFWxP1Wkcb7VQoj3k3j6Yzhe5W=5sY3Q@mail.gmail.com>
Subject: Re: [PATCH] media: s5p-g2d: Fix a memory leak on ctx->fh.m2m_ctx
To:     Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc:     mchehab+huawei@kernel.org, a.hajda@samsung.com,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        linux-media <linux-media@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Hans,

Thanks for the reply.

On Wed, May 26, 2021 at 9:41 PM Hans Verkuil <hverkuil-cisco@xs4all.nl> wrote:
>
> On 11/05/2021 06:09, dillon.minfei@gmail.com wrote:
> > From: Dillon Min <dillon.minfei@gmail.com>
> >
> > The m2m_ctx resources was allocated by v4l2_m2m_ctx_init() in g2d_open()
> > should be freed from g2d_release() when it's not used.
> >
> > Fix it
> >
> > Fixes: 918847341af0 ("[media] v4l: add G2D driver for s5p device family")
> > Signed-off-by: Dillon Min <dillon.minfei@gmail.com>
> > ---
> >  drivers/media/platform/s5p-g2d/g2d.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/media/platform/s5p-g2d/g2d.c b/drivers/media/platform/s5p-g2d/g2d.c
> > index 15bcb7f6e113..0818fdd3e984 100644
> > --- a/drivers/media/platform/s5p-g2d/g2d.c
> > +++ b/drivers/media/platform/s5p-g2d/g2d.c
> > @@ -279,6 +279,9 @@ static int g2d_release(struct file *file)
> >       v4l2_ctrl_handler_free(&ctx->ctrl_handler);
> >       v4l2_fh_del(&ctx->fh);
> >       v4l2_fh_exit(&ctx->fh);
> > +     mutex_lock(&dev->mutex);
> > +     v4l2_m2m_ctx_release(ctx->fh.m2m_ctx);
> > +     mutex_unlock(&dev->mutex);
>
> This should be moved up to just before the v4l2_ctrl_handler_free() call.

On my mistake, thanks for pointing it out. will be moved up to
v4l2_ctrl_handler_free() in v2.

Best regards.
Dillon

>
> Regards,
>
>         Hans
>
> >       kfree(ctx);
> >       v4l2_info(&dev->v4l2_dev, "instance closed\n");
> >       return 0;
> >
>