linux-kernel - Re: [PATCH 7/9] media: hantro: Add core bits to support H264 decoding

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 for Android: free password hash cracker in your pocket

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

Message-ID: <25cc2826-fc8f-570a-07fa-8cbdb11830a7@rasmusvillemoes.dk>
Date:   Thu, 25 Jul 2019 15:31:41 +0200
From:   Rasmus Villemoes <linux@...musvillemoes.dk>
To:     Boris Brezillon <boris.brezillon@...labora.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Hans Verkuil <hans.verkuil@...co.com>,
        Laurent Pinchart <laurent.pinchart@...asonboard.com>,
        Sakari Ailus <sakari.ailus@....fi>, linux-media@...r.kernel.org
Cc:     linux-kernel@...r.kernel.org, Tomasz Figa <tfiga@...omium.org>,
        Nicolas Dufresne <nicolas@...fresne.ca>, kernel@...labora.com,
        Paul Kocialkowski <paul.kocialkowski@...tlin.com>,
        Ezequiel Garcia <ezequiel@...labora.com>,
        Jonas Karlman <jonas@...boo.se>,
        linux-rockchip@...ts.infradead.org,
        Heiko Stuebner <heiko@...ech.de>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        Hertz Wong <hertz.wong@...k-chips.com>
Subject: Re: [PATCH 7/9] media: hantro: Add core bits to support H264 decoding

On 19/06/2019 14.15, Boris Brezillon wrote:
> From: Hertz Wong <hertz.wong@...k-chips.com>
> 
> Add helpers and patch hantro_{drv,v4l2}.c to prepare addition of H264
> decoding support.
> 
> Signed-off-by: Hertz Wong <hertz.wong@...k-chips.com>
> Signed-off-by: Boris Brezillon <boris.brezillon@...labora.com>
> ---
> +
> +	/*
> +	 * Short term pics in descending pic num order, long term ones in
> +	 * ascending order.
> +	 */
> +	if (!(a->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM))
> +		return b->frame_num - a->frame_num;
> +
> +	return a->pic_num - b->pic_num;
> +}

Pet peeve: This works because ->frame_num and ->pic_num are u16, so
their difference is guaranteed to fit in an int.

> +static int b0_ref_list_cmp(const void *ptra, const void *ptrb, const void *data)
> +{
> +	const struct hantro_h264_reflist_builder *builder = data;
> +	const struct v4l2_h264_dpb_entry *a, *b;
> +	s32 poca, pocb;
> +	u8 idxa, idxb;
> +
> +	idxa = *((u8 *)ptra);
> +	idxb = *((u8 *)ptrb);
> +	a = &builder->dpb[idxa];
> +	b = &builder->dpb[idxb];
> +
> +	if ((a->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM) !=
> +	    (b->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM)) {
> +		/* Short term pics firt. */
> +		if (!(a->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM))
> +			return -1;
> +		else
> +			return 1;
> +	}
> +
> +	/* Long term pics in ascending pic num order. */
> +	if (a->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM)
> +		return a->pic_num - b->pic_num;
> +
> +	poca = builder->pocs[idxa];
> +	pocb = builder->pocs[idxb];
> +
> +	/*
> +	 * Short term pics with POC < cur POC first in POC descending order
> +	 * followed by short term pics with POC > cur POC in POC ascending
> +	 * order.
> +	 */
> +	if ((poca < builder->curpoc) != (pocb < builder->curpoc))
> +		return poca - pocb;
> +	else if (poca < builder->curpoc)
> +		return pocb - poca;
> +
> +	return poca - pocb;
> +}

Here, however, poca and pocb are ints. What guarantees that their values
are not more than 2^31 apart? I know absolutely nothing about this code
or what these numbers represent, so it may be obvious that they are
smallish.

Rasmus