Re: [FFmpeg-devel] [PATCH 3/5] af_hdcd: fix possible integer overflow

[Burt P.]

applied

with fix:
uint64_t sustain_reset = (uint64_t)cdt_ms * rate / 1000;

Thank you, Michael.

-- 
Burt

On Tue, Sep 6, 2016 at 5:00 PM, Michael Niedermayer
 wrote:
> On Mon, Sep 05, 2016 at 06:18:43AM -0500, Burt P wrote:
>> Signed-off-by: Burt P 
>> ---
>>  libavfilter/af_hdcd.c | 9 -
>>  1 file changed, 4 insertions(+), 5 deletions(-)
>>
>> diff --git a/libavfilter/af_hdcd.c b/libavfilter/af_hdcd.c
>> index c8bda82..c249589 100644
>> --- a/libavfilter/af_hdcd.c
>> +++ b/libavfilter/af_hdcd.c
>> @@ -1004,16 +1004,15 @@ AVFILTER_DEFINE_CLASS(hdcd);
>>  static void hdcd_reset(hdcd_state *state, unsigned rate, unsigned cdt_ms)
>>  {
>>  int i;
>> +uint64_t sustain_reset = cdt_ms * rate / 1000;
>
> this can still overflow
> cdt_ms and rate are 32bit their product is 32bit divided by 1000
> its around 22 bit, the 64bit is too late
>
>
>>
>>  state->window = 0;
>>  state->readahead = 32;
>>  state->arg = 0;
>>  state->control = 0;
>> -
>>  state->running_gain = 0;
>> -
>> +state->sustain_reset = sustain_reset;
>>  state->sustain = 0;
>> -state->sustain_reset = cdt_ms*rate/1000;
>
> [...]
>
> --
> Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> The worst form of inequality is to try to make unequal things equal.
> -- Aristotle
>
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>



-- 
Burt
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH 3/5] af_hdcd: fix possible integer overflow

[Michael Niedermayer]

On Mon, Sep 05, 2016 at 06:18:43AM -0500, Burt P wrote:
> Signed-off-by: Burt P 
> ---
>  libavfilter/af_hdcd.c | 9 -
>  1 file changed, 4 insertions(+), 5 deletions(-)
> 
> diff --git a/libavfilter/af_hdcd.c b/libavfilter/af_hdcd.c
> index c8bda82..c249589 100644
> --- a/libavfilter/af_hdcd.c
> +++ b/libavfilter/af_hdcd.c
> @@ -1004,16 +1004,15 @@ AVFILTER_DEFINE_CLASS(hdcd);
>  static void hdcd_reset(hdcd_state *state, unsigned rate, unsigned cdt_ms)
>  {
>  int i;
> +uint64_t sustain_reset = cdt_ms * rate / 1000;

this can still overflow
cdt_ms and rate are 32bit their product is 32bit divided by 1000
its around 22 bit, the 64bit is too late


>  
>  state->window = 0;
>  state->readahead = 32;
>  state->arg = 0;
>  state->control = 0;
> -
>  state->running_gain = 0;
> -
> +state->sustain_reset = sustain_reset;
>  state->sustain = 0;
> -state->sustain_reset = cdt_ms*rate/1000;

[...]

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The worst form of inequality is to try to make unequal things equal.
-- Aristotle


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH 3/5] af_hdcd: fix possible integer overflow

[Burt P]

Signed-off-by: Burt P 
---
 libavfilter/af_hdcd.c | 9 -
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/libavfilter/af_hdcd.c b/libavfilter/af_hdcd.c
index c8bda82..c249589 100644
--- a/libavfilter/af_hdcd.c
+++ b/libavfilter/af_hdcd.c
@@ -1004,16 +1004,15 @@ AVFILTER_DEFINE_CLASS(hdcd);
 static void hdcd_reset(hdcd_state *state, unsigned rate, unsigned cdt_ms)
 {
 int i;
+uint64_t sustain_reset = cdt_ms * rate / 1000;
 
 state->window = 0;
 state->readahead = 32;
 state->arg = 0;
 state->control = 0;
-
 state->running_gain = 0;
-
+state->sustain_reset = sustain_reset;
 state->sustain = 0;
-state->sustain_reset = cdt_ms*rate/1000;
 
 state->code_counterA = 0;
 state->code_counterA_almost = 0;
@@ -1788,8 +1787,8 @@ static av_cold int init(AVFilterContext *ctx)
 hdcd_reset(>state[c], 44100, s->cdt_ms);
 }
 
-av_log(ctx, AV_LOG_VERBOSE, "CDT period: %dms (%d samples @44100Hz)\n",
-s->cdt_ms, s->cdt_ms*44100/1000 );
+av_log(ctx, AV_LOG_VERBOSE, "CDT period: %dms (%u samples @44100Hz)\n",
+s->cdt_ms, s->state[0].sustain_reset );
 av_log(ctx, AV_LOG_VERBOSE, "Process mode: %s\n",
 (s->process_stereo) ? "process stereo channels together" : "process 
each channel separately");
 av_log(ctx, AV_LOG_VERBOSE, "Force PE: %s\n",
-- 
2.7.4

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel