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;
Date:   Tue, 5 May 2020 17:18:40 +0200
From:   Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
To:     Neil Armstrong <narmstrong@baylibre.com>
Cc:     linux-media@vger.kernel.org, hverkuil-cisco@xs4all.nl,
        linux-amlogic@lists.infradead.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        Maxime Jourdan <mjourdan@baylibre.com>
Subject: Re: [PATCH 2/3] media: meson: vdec: fix another case of VP9 buffer
 shortage
Message-ID: <20200505171840.48d89ccc@coco.lan>
In-Reply-To: <20200428125036.9401-3-narmstrong@baylibre.com>
References: <20200428125036.9401-1-narmstrong@baylibre.com>
        <20200428125036.9401-3-narmstrong@baylibre.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Em Tue, 28 Apr 2020 14:50:35 +0200
Neil Armstrong <narmstrong@baylibre.com> escreveu:

> From: Maxime Jourdan <mjourdan@baylibre.com>
> 
> - Redo the logic where VP9 gets fresh CAPTURE buffers. The previous code
>   could lead to a hardlock.
> - Reserve 4 margin buffers instead of 3, as apparently there are corner
>   cases where 3 is not enough.
> 
> Fixes: e9a3eb4819ca ("media: meson: vdec: add VP9 input support")
> Fixes: 00c43088aa68 ("media: meson: vdec: add VP9 decoder support")
> Signed-off-by: Maxime Jourdan <mjourdan@baylibre.com>
> Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>

> +static struct vb2_v4l2_buffer *get_free_vbuf(struct amvdec_session *sess)
> +{
> +	struct codec_vp9 *vp9 = sess->priv;
> +	struct vb2_v4l2_buffer *vbuf = v4l2_m2m_dst_buf_remove(sess->m2m_ctx);
> +	struct vb2_v4l2_buffer *vbuf2;
> +
> +	if (!vbuf)
> +		return NULL;
> +
> +	if (!codec_vp9_get_frame_by_idx(vp9, vbuf->vb2_buf.index))
> +		return vbuf;
> +
> +	vbuf2 = get_free_vbuf(sess);

Huh!!!!

Never use recursive functions inside the Kernel! Kernel stack is too
limited.

Also, even if Kernel stack would be unlimited, the above logic
would endlessly be calling get_free_vbuf(sess).

Thanks,
Mauro