On 7 December 2012 14:23, Martin Storsjö <mar...@martin.st> wrote:
> On Fri, 7 Dec 2012, Mans Rullgard wrote:
>
>> When LOCAL_ALIGNED uses manual alignment initialisation is not
>> possible.
>>

There's another case further down.

diff --git a/libavcodec/takdec.c b/libavcodec/takdec.c
index 0ac870c..d47db48 100644
--- a/libavcodec/takdec.c
+++ b/libavcodec/takdec.c
@@ -601,10 +601,12 @@ static int decorrelate(TAKDecContext *s, int c1,
int c2, int length)
     case 6:
         FFSWAP(int32_t*, p1, p2);
     case 7: {
-        LOCAL_ALIGNED_16(int16_t, filter, [MAX_PREDICTORS]) = { 0 };
+        LOCAL_ALIGNED_16(int16_t, filter, [MAX_PREDICTORS]);
         int length2, order_half, filter_order, dval1, dval2;
         int av_uninit(code_size);

+        memset(filter, 0, MAX_PREDICTORS * sizeof(*filter));
+
         if (length < 256)
             return AVERROR_INVALIDDATA;



>> Signed-off-by: Mans Rullgard <m...@mansr.com>
>> ---
>> libavcodec/takdec.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/libavcodec/takdec.c b/libavcodec/takdec.c
>> index 87fcf83..0ac870c 100644
>> --- a/libavcodec/takdec.c
>> +++ b/libavcodec/takdec.c
>> @@ -420,11 +420,13 @@ static void decode_filter_coeffs(TAKDecContext *s,
>> int filter_order, int size,
>> static int decode_subframe(TAKDecContext *s, int32_t *decoded,
>>                            int subframe_size, int prev_subframe_size)
>> {
>> -    LOCAL_ALIGNED_16(int16_t, filter, [MAX_PREDICTORS]) = { 0, };
>> +    LOCAL_ALIGNED_16(int16_t, filter, [MAX_PREDICTORS]);
>>     GetBitContext *gb = &s->gb;
>>     int i, ret;
>>     int dshift, size, filter_quant, filter_order;
>>
>> +    memset(filter, 0, MAX_PREDICTORS * sizeof(*filter));
>> +
>>     if (!get_bits1(gb))
>>         return decode_residues(s, decoded, subframe_size);
>>
>> --
>> 1.8.0
>
>
> Looks good to me
>
> // Martin
>
> _______________________________________________
> libav-devel mailing list
> libav-devel@libav.org
> https://lists.libav.org/mailman/listinfo/libav-devel
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to