[FFmpeg-devel] [PATCH 3/3] avcodec/bitstream: Consistently treat symbol as VLC_TYPE

Andreas Rheinhardt andreas.rheinhardt at gmail.com
Tue Oct 27 13:37:53 EET 2020


Anton Khirnov:
> Quoting Andreas Rheinhardt (2020-10-27 11:30:32)
>> Andreas Rheinhardt:
>>> If a static VLC table gets initialized a second time (or concurrently by
>>> two threads) and if said VLC table uses symbols that have the sign bit
>>> of VLC_TYPE (a typedef for int16_t) set, initializing the VLC fails. The
>>> reason is that the type of the symbol in the temporary array is an
>>> uint16_t and so comparing it to the symbol read from the VLC table will
>>> fail, because only the lower 16bits coincide. Said failure triggers an
>>> assert.
>>>
>>> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt at gmail.com>
>>> ---
>>> Found when playing around a bit with making the ClearVideo VLCs static.
>>>
>>>  libavcodec/bitstream.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/libavcodec/bitstream.c b/libavcodec/bitstream.c
>>> index 39749c6092..a908c10980 100644
>>> --- a/libavcodec/bitstream.c
>>> +++ b/libavcodec/bitstream.c
>>> @@ -129,7 +129,7 @@ static int alloc_table(VLC *vlc, int size, int use_static)
>>>  
>>>  typedef struct VLCcode {
>>>      uint8_t bits;
>>> -    uint16_t symbol;
>>> +    VLC_TYPE symbol;
>>>      /** codeword, with the first bit-to-be-read in the msb
>>>       * (even if intended for a little-endian bitstream reader) */
>>>      uint32_t code;
>>>
>> LGTM'ed by Lynne via email and pushed.
> 
> This is the second time I see this, why does that not go through the ML?
> 
I don't know. Seems like Lynne has some trouble with her MUA.

- Andreas


More information about the ffmpeg-devel mailing list