[FFmpeg-devel,2/2] avfilter/af_mcompand: check allocation results

Fixes the only remaining part of ticket #8931.

Signed-off-by: Marton Balint <cus@passwd.hu>
---
 libavfilter/af_mcompand.c | 3 +++
 1 file changed, 3 insertions(+)

Marton Balint:
> Fixes the only remaining part of ticket #8931.
> 
> Signed-off-by: Marton Balint <cus@passwd.hu>
> ---
>  libavfilter/af_mcompand.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/libavfilter/af_mcompand.c b/libavfilter/af_mcompand.c
> index ce4f366ad7..d1a3dc123a 100644
> --- a/libavfilter/af_mcompand.c
> +++ b/libavfilter/af_mcompand.c
> @@ -384,6 +384,9 @@ static int config_output(AVFilterLink *outlink)
>          s->bands[i].attack_rate = av_calloc(outlink->channels, sizeof(double));
>          s->bands[i].decay_rate = av_calloc(outlink->channels, sizeof(double));
>          s->bands[i].volume = av_calloc(outlink->channels, sizeof(double));
> +        if (!s->bands[i].attack_rate || !s->bands[i].decay_rate || !s->bands[i].volume)
> +            return AVERROR(ENOMEM);
> +
>          for (k = 0; k < FFMIN(nb_attacks / 2, outlink->channels); k++) {
>              char *tstr3 = av_strtok(p3, ",", &saveptr3);
>  
> 
You are not the first one:
https://ffmpeg.org/pipermail/ffmpeg-devel/2020-October/270956.html (I
haven't applied this as I thought that Paul as the author/maintainer
should do so.)

- Andreas

On Sat, 10 Apr 2021, Andreas Rheinhardt wrote:

> Marton Balint:
>> Fixes the only remaining part of ticket #8931.
>> 
>> Signed-off-by: Marton Balint <cus@passwd.hu>
>> ---
>>  libavfilter/af_mcompand.c | 3 +++
>>  1 file changed, 3 insertions(+)
>> 
>> diff --git a/libavfilter/af_mcompand.c b/libavfilter/af_mcompand.c
>> index ce4f366ad7..d1a3dc123a 100644
>> --- a/libavfilter/af_mcompand.c
>> +++ b/libavfilter/af_mcompand.c
>> @@ -384,6 +384,9 @@ static int config_output(AVFilterLink *outlink)
>>          s->bands[i].attack_rate = av_calloc(outlink->channels, sizeof(double));
>>          s->bands[i].decay_rate = av_calloc(outlink->channels, sizeof(double));
>>          s->bands[i].volume = av_calloc(outlink->channels, sizeof(double));
>> +        if (!s->bands[i].attack_rate || !s->bands[i].decay_rate || !s->bands[i].volume)
>> +            return AVERROR(ENOMEM);
>> +
>>          for (k = 0; k < FFMIN(nb_attacks / 2, outlink->channels); k++) {
>>              char *tstr3 = av_strtok(p3, ",", &saveptr3);
>> 
>> 
> You are not the first one:
> https://ffmpeg.org/pipermail/ffmpeg-devel/2020-October/270956.html (I
> haven't applied this as I thought that Paul as the author/maintainer
> should do so.)

If the maintainer does not seem interested then it is OK to apply, 
especially if it looks trivial. So feel free to apply whichever version 
you prefer.

Thanks,
Marton

On Sat, 10 Apr 2021, Marton Balint wrote:

>
>
> On Sat, 10 Apr 2021, Andreas Rheinhardt wrote:
>
>> Marton Balint:
>>> Fixes the only remaining part of ticket #8931.
>>> 
>>> Signed-off-by: Marton Balint <cus@passwd.hu>
>>> ---
>>>  libavfilter/af_mcompand.c | 3 +++
>>>  1 file changed, 3 insertions(+)
>>> 
>>> diff --git a/libavfilter/af_mcompand.c b/libavfilter/af_mcompand.c
>>> index ce4f366ad7..d1a3dc123a 100644
>>> --- a/libavfilter/af_mcompand.c
>>> +++ b/libavfilter/af_mcompand.c
>>> @@ -384,6 +384,9 @@ static int config_output(AVFilterLink *outlink)
>>>          s->bands[i].attack_rate = av_calloc(outlink->channels, 
> sizeof(double));
>>>          s->bands[i].decay_rate = av_calloc(outlink->channels, 
> sizeof(double));
>>>          s->bands[i].volume = av_calloc(outlink->channels, 
> sizeof(double));
>>> +        if (!s->bands[i].attack_rate || !s->bands[i].decay_rate || 
> !s->bands[i].volume)
>>> +            return AVERROR(ENOMEM);
>>> +
>>>          for (k = 0; k < FFMIN(nb_attacks / 2, outlink->channels); k++) {
>>>              char *tstr3 = av_strtok(p3, ",", &saveptr3);
>>> 
>>> 
>> You are not the first one:
>> https://ffmpeg.org/pipermail/ffmpeg-devel/2020-October/270956.html (I
>> haven't applied this as I thought that Paul as the author/maintainer
>> should do so.)
>
> If the maintainer does not seem interested then it is OK to apply, 
> especially if it looks trivial. So feel free to apply whichever version 
> you prefer.

Will push this series soon.

Regards,
Marton