From: James Almer <jamrial@gmail.com>
To: ffmpeg-devel@ffmpeg.org
Subject: Re: [FFmpeg-devel] [PATCH 2/3] avcodec/wmalosslessdec: Check channel mask against num channels
Date: Thu, 17 Mar 2022 22:00:18 -0300
Message-ID: <dafef74e-4bcf-f45d-8905-cb547a7a4b8b@gmail.com> (raw)
In-Reply-To: <74943f9e-b07a-7fe4-fa83-09a9cbda21a0@gmail.com>
On 3/17/2022 9:07 PM, James Almer wrote:
>
>
> On 3/17/2022 8:52 PM, James Almer wrote:
>> On 3/17/2022 8:30 PM, Michael Niedermayer wrote:
>>> Fixes: Out of array write
>>> Fixes:
>>> 45613/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_WMALOSSLESS_fuzzer-4539073606320128
>>>
>>>
>>> Found-by: continuous fuzzing process
>>> https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
>>> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
>>> ---
>>> libavcodec/wmalosslessdec.c | 3 +++
>>> 1 file changed, 3 insertions(+)
>>>
>>> diff --git a/libavcodec/wmalosslessdec.c b/libavcodec/wmalosslessdec.c
>>> index cd05b22689..1728920729 100644
>>> --- a/libavcodec/wmalosslessdec.c
>>> +++ b/libavcodec/wmalosslessdec.c
>>> @@ -281,6 +281,9 @@ static av_cold int decode_init(AVCodecContext
>>> *avctx)
>>> av_channel_layout_uninit(&avctx->ch_layout);
>>> av_channel_layout_from_mask(&avctx->ch_layout, channel_mask);
>>> + if (s->num_channels != avctx->ch_layout.nb_channels)
>>> + return AVERROR_PATCHWELCOME; //are there non fuzzed files
>>> with this or is it an error ?
>>
>> s->num_channels at this point is set to the channels count the user
>> set before calling avcodec_open2() (Normally from lavf), but it could
>> be anything.
>> If channel_mask is taken from extradata, maybe it should be used to
>> set s->num_channels instead of aborting because the user set value and
>> extradata disagreed.
>>
>> Also, can you reproduce this crash before 3c933af493? s->num_channels
>> was being set to the user set channel count too, same as now.
>
> Right, before that commit s->num_channels and avctx->channels were
> always the same, but avctx->channel_layout was whatever came from
> extradata, and its popcnt could be != avctx->channels.
> After it, avctx->ch_layout.nb_channels is always the same as
> popcnt(avctx->ch_layout.u.mask), which can be different than
> s->num_channels.
>
> I think my suggestion above to use the extradata channel mask and
> ignoring the user set channel count is the best approach for this.
Like this maybe (channel_mask could in theory be zero, so in that case
the user set value should be used).
> diff --git a/libavcodec/wmalosslessdec.c b/libavcodec/wmalosslessdec.c
> index cd05b22689..915add1962 100644
> --- a/libavcodec/wmalosslessdec.c
> +++ b/libavcodec/wmalosslessdec.c
> @@ -197,15 +197,6 @@ static av_cold int decode_init(AVCodecContext *avctx)
> return AVERROR_PATCHWELCOME;
> }
>
> - s->max_frame_size = MAX_FRAMESIZE * avctx->ch_layout.nb_channels;
> - s->frame_data = av_mallocz(s->max_frame_size + AV_INPUT_BUFFER_PADDING_SIZE);
> - if (!s->frame_data)
> - return AVERROR(ENOMEM);
> -
> - s->avctx = avctx;
> - ff_llauddsp_init(&s->dsp);
> - init_put_bits(&s->pb, s->frame_data, s->max_frame_size);
> -
> if (avctx->extradata_size >= 18) {
> s->decode_flags = AV_RL16(edata_ptr + 14);
> channel_mask = AV_RL32(edata_ptr + 2);
> @@ -230,6 +221,33 @@ static av_cold int decode_init(AVCodecContext *avctx)
> return AVERROR_PATCHWELCOME;
> }
>
> + if (channel_mask) {
> + av_channel_layout_uninit(&avctx->ch_layout);
> + av_channel_layout_from_mask(&avctx->ch_layout, channel_mask);
> + } else
> + avctx->ch_layout.order = AV_CHANNEL_ORDER_UNSPEC;
> +
> + s->num_channels = avctx->ch_layout.nb_channels;
> +
> + /* extract lfe channel position */
> + s->lfe_channel = -1;
> +
> + if (channel_mask & 8) {
> + unsigned int mask;
> + for (mask = 1; mask < 16; mask <<= 1)
> + if (channel_mask & mask)
> + ++s->lfe_channel;
> + }
> +
> + s->max_frame_size = MAX_FRAMESIZE * avctx->ch_layout.nb_channels;
> + s->frame_data = av_mallocz(s->max_frame_size + AV_INPUT_BUFFER_PADDING_SIZE);
> + if (!s->frame_data)
> + return AVERROR(ENOMEM);
> +
> + s->avctx = avctx;
> + ff_llauddsp_init(&s->dsp);
> + init_put_bits(&s->pb, s->frame_data, s->max_frame_size);
> +
> /* generic init */
> s->log2_frame_size = av_log2(avctx->block_align) + 4;
>
> @@ -263,24 +281,10 @@ static av_cold int decode_init(AVCodecContext *avctx)
> return AVERROR_INVALIDDATA;
> }
>
> - s->num_channels = avctx->ch_layout.nb_channels;
> -
> - /* extract lfe channel position */
> - s->lfe_channel = -1;
> -
> - if (channel_mask & 8) {
> - unsigned int mask;
> - for (mask = 1; mask < 16; mask <<= 1)
> - if (channel_mask & mask)
> - ++s->lfe_channel;
> - }
> -
> s->frame = av_frame_alloc();
> if (!s->frame)
> return AVERROR(ENOMEM);
>
> - av_channel_layout_uninit(&avctx->ch_layout);
> - av_channel_layout_from_mask(&avctx->ch_layout, channel_mask);
> return 0;
> }
_______________________________________________
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".
next prev parent reply other threads:[~2022-03-18 1:00 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-17 23:30 [FFmpeg-devel] [PATCH 1/3] avcodec/dfpwmdec: Check packet size more completely Michael Niedermayer
2022-03-17 23:30 ` [FFmpeg-devel] [PATCH 2/3] avcodec/wmalosslessdec: Check channel mask against num channels Michael Niedermayer
2022-03-17 23:52 ` James Almer
2022-03-18 0:07 ` James Almer
2022-03-18 1:00 ` James Almer [this message]
2022-03-18 11:27 ` Michael Niedermayer
2022-03-18 11:38 ` James Almer
2022-03-19 17:52 ` Michael Niedermayer
2022-03-29 19:02 ` Michael Niedermayer
2022-03-29 19:35 ` James Almer
2022-03-17 23:30 ` [FFmpeg-devel] [PATCH 3/3] avcodec/alsdec: Set channels from data after data is set Michael Niedermayer
2022-03-17 23:40 ` James Almer
2022-03-18 11:28 ` Michael Niedermayer
2022-03-18 8:31 ` [FFmpeg-devel] [PATCH 1/3] avcodec/dfpwmdec: Check packet size more completely Paul B Mahol
2022-03-18 11:28 ` Michael Niedermayer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=dafef74e-4bcf-f45d-8905-cb547a7a4b8b@gmail.com \
--to=jamrial@gmail.com \
--cc=ffmpeg-devel@ffmpeg.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git