linux-kernel - Re: [PATCH v5 08/10] media: Quit parsing stream if doesn't start with SOI

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2dfd7785da904e9b518a0e7579d4db79a0b19276.camel@pengutronix.de>
Date:   Wed, 02 Dec 2020 16:18:44 +0100
From:   Philipp Zabel <p.zabel@...gutronix.de>
To:     "Mirela Rabulea (OSS)" <mirela.rabulea@....nxp.com>,
        mchehab@...nel.org, hverkuil-cisco@...all.nl, shawnguo@...nel.org,
        robh+dt@...nel.org
Cc:     paul.kocialkowski@...tlin.com, linux-media@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-imx@....com,
        s.hauer@...gutronix.de, aisheng.dong@....com,
        daniel.baluta@....com, robert.chiras@....com,
        laurentiu.palcu@....com, mark.rutland@....com,
        devicetree@...r.kernel.org, ezequiel@...labora.com,
        laurent.pinchart+renesas@...asonboard.com,
        niklas.soderlund+renesas@...natech.se,
        dafna.hirschfeld@...labora.com,
        Mirela Rabulea <mirela.rabulea@....com>
Subject: Re: [PATCH v5 08/10] media: Quit parsing stream if doesn't start
 with SOI

On Thu, 2020-11-12 at 05:05 +0200, Mirela Rabulea (OSS) wrote:
> From: Mirela Rabulea <mirela.rabulea@....com>
> 
> In the case we get an invalid stream, such as from v4l2-compliance
> streaming test, jpeg_next_marker will end up parsing the entire
> stream. The standard describes the high level syntax of a jpeg
> as starting with SOI, ending with EOI, so return error if the very
> first 2 bytes are not SOI.
> 
> Signed-off-by: Mirela Rabulea <mirela.rabulea@....com>
> ---
>  drivers/media/v4l2-core/v4l2-jpeg.c | 7 ++-----
>  1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/media/v4l2-core/v4l2-jpeg.c b/drivers/media/v4l2-core/v4l2-jpeg.c
> index 3181ce544f79..d77e04083d57 100644
> --- a/drivers/media/v4l2-core/v4l2-jpeg.c
> +++ b/drivers/media/v4l2-core/v4l2-jpeg.c
> @@ -499,11 +499,8 @@ int v4l2_jpeg_parse_header(void *buf, size_t len, struct v4l2_jpeg_header *out)
>  	out->num_dht = 0;
>  	out->num_dqt = 0;
>  
> -	/* the first marker must be SOI */
> -	marker = jpeg_next_marker(&stream);
> -	if (marker < 0)
> -		return marker;
> -	if (marker != SOI)
> +	/* the first bytes must be SOI, B.2.1 High-level syntax */
> +	if (jpeg_get_word_be(&stream) != SOI)
>  		return -EINVAL;
>  
>  	/* init value to signal if this marker is not present */

Yes, shorter, potentially faster code, and it adheres to the
specification more strictly.

Reviewed-by: Philipp Zabel <p.zabel@...gutronix.de>

regards
Philipp