From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <ffmpeg-devel-bounces@ffmpeg.org>
Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100])
	by master.gitmailbox.com (Postfix) with ESMTP id 9071241A61
	for <ffmpegdev@gitmailbox.com>; Sun, 16 Jan 2022 00:59:57 +0000 (UTC)
Received: from [127.0.1.1] (localhost [127.0.0.1])
	by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 0698F68A990;
	Sun, 16 Jan 2022 02:59:55 +0200 (EET)
Received: from iq.passwd.hu (iq.passwd.hu [217.27.212.140])
 by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 9C9E268073D
 for <ffmpeg-devel@ffmpeg.org>; Sun, 16 Jan 2022 02:59:48 +0200 (EET)
Received: from localhost (localhost [127.0.0.1])
 by iq.passwd.hu (Postfix) with ESMTP id 3A8CBE6666
 for <ffmpeg-devel@ffmpeg.org>; Sun, 16 Jan 2022 01:59:48 +0100 (CET)
X-Virus-Scanned: amavisd-new at passwd.hu
Received: from iq.passwd.hu ([127.0.0.1])
 by localhost (iq.passwd.hu [127.0.0.1]) (amavisd-new, port 10024)
 with ESMTP id yDvghJnBHIK8 for <ffmpeg-devel@ffmpeg.org>;
 Sun, 16 Jan 2022 01:59:44 +0100 (CET)
Received: from iq (iq [217.27.212.140])
 by iq.passwd.hu (Postfix) with ESMTPS id 158D8E65DA
 for <ffmpeg-devel@ffmpeg.org>; Sun, 16 Jan 2022 01:59:44 +0100 (CET)
Date: Sun, 16 Jan 2022 01:59:44 +0100 (CET)
From: Marton Balint <cus@passwd.hu>
To: ffmpeg-devel@ffmpeg.org
Message-ID: <552433aa-e5a5-5821-fc10-3f5525ee468@passwd.hu>
MIME-Version: 1.0
Subject: [FFmpeg-devel] AVFrame data alignment issues
X-BeenThere: ffmpeg-devel@ffmpeg.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: FFmpeg development discussions and patches <ffmpeg-devel.ffmpeg.org>
List-Unsubscribe: <https://ffmpeg.org/mailman/options/ffmpeg-devel>,
 <mailto:ffmpeg-devel-request@ffmpeg.org?subject=unsubscribe>
List-Archive: <https://ffmpeg.org/pipermail/ffmpeg-devel>
List-Post: <mailto:ffmpeg-devel@ffmpeg.org>
List-Help: <mailto:ffmpeg-devel-request@ffmpeg.org?subject=help>
List-Subscribe: <https://ffmpeg.org/mailman/listinfo/ffmpeg-devel>,
 <mailto:ffmpeg-devel-request@ffmpeg.org?subject=subscribe>
Reply-To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset="us-ascii"; Format="flowed"
Errors-To: ffmpeg-devel-bounces@ffmpeg.org
Sender: "ffmpeg-devel" <ffmpeg-devel-bounces@ffmpeg.org>
Archived-At: <https://master.gitmailbox.com/ffmpegdev/552433aa-e5a5-5821-fc10-3f5525ee468@passwd.hu/>
List-Archive: <https://master.gitmailbox.com/ffmpegdev/>
List-Post: <mailto:ffmpegdev@gitmailbox.com>

Hi,

I wonder if there are strict requirements for alignment of AVFrame 
buffers. Yes, it is usually aligned by default, but there are filters, 
like crop or pad which tries to avoid copying data even at the cost of 
breaking the alignment. This unfortunately causes crashes for code which 
requires it, some examples:

- the zscale filter only checks input buffer alignment but not output
   buffer alignment, therefore might crash if a later pad filter provides
   an unaligned get_buffer callback for a filter chain.

- the blend/tblend filter has a similar problem, it's ASM code is using
   aligned movs for output. (not for input however).

- the dnxhd encoder depends on aligment of its input buffers.

How this should be fixed?

- Change pad/crop filters to only do "smart" pad/crop if alignment can be 
guaranteed? Maybe a user option should be added to enable "unsafe" mode?

- Add checks for aligment where it is actually required, e.g. in dnxhd 
encoder or blend/zscale filter?

- Change blend code to use unaligned movs even for output?

Thanks,
Marton
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".