opj_t1_encode_cblk(): avoid uint32 overflow when numbps = 0 (which is well defined behaviour, and is properly handled here, but better avoid it to detect real issues)
This commit is contained in:
parent
1e387de742
commit
8e6c371e66
|
@ -2155,6 +2155,10 @@ static void opj_t1_encode_cblk(opj_t1_t *t1,
|
||||||
|
|
||||||
cblk->numbps = max ? (OPJ_UINT32)((opj_int_floorlog2(max) + 1) -
|
cblk->numbps = max ? (OPJ_UINT32)((opj_int_floorlog2(max) + 1) -
|
||||||
T1_NMSEDEC_FRACBITS) : 0;
|
T1_NMSEDEC_FRACBITS) : 0;
|
||||||
|
if (cblk->numbps == 0) {
|
||||||
|
cblk->totalpasses = 0;
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
|
||||||
bpno = (OPJ_INT32)(cblk->numbps - 1);
|
bpno = (OPJ_INT32)(cblk->numbps - 1);
|
||||||
passtype = 2;
|
passtype = 2;
|
||||||
|
|
Loading…
Reference in New Issue