Check-in [edd0e8d3da]
Not logged in

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:add tcl upstream changes
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: edd0e8d3da7c19f017ca41ad696b4164a9ad5913
User & Date: chw 2019-11-08 05:38:46
Context
2019-11-08
09:22
add selected open62541 upstream changes in topcua subdir check-in: e9e95211a0 user: chw tags: trunk
05:38
add tcl upstream changes check-in: edd0e8d3da user: chw tags: trunk
2019-11-07
17:40
make ffidl compile/link on freebsd 12.1 check-in: dce7402899 user: chw tags: trunk
Changes

Changes to jni/tcl/libtommath/bn_mp_mul.c.

8
9
10
11
12
13
14


15
16
17
18
19
20
21
22
{
   mp_err err;
   int min_len = MP_MIN(a->used, b->used),
       max_len = MP_MAX(a->used, b->used),
       digs = a->used + b->used + 1;
   mp_sign neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;



   if (MP_HAS(S_MP_BALANCE_MUL) &&
       /* Check sizes. The smaller one needs to be larger than the Karatsuba cut-off.
        * The bigger one needs to be at least about one MP_KARATSUBA_MUL_CUTOFF bigger
        * to make some sense, but it depends on architecture, OS, position of the
        * stars... so YMMV.
        * Using it to cut the input into slices small enough for fast_s_mp_mul_digs
        * was actually slower on the author's machine, but YMMV.
        */







>
>
|







8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
{
   mp_err err;
   int min_len = MP_MIN(a->used, b->used),
       max_len = MP_MAX(a->used, b->used),
       digs = a->used + b->used + 1;
   mp_sign neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;

   if (a == b) {
       return mp_sqr(a,c);
   } else if (MP_HAS(S_MP_BALANCE_MUL) &&
       /* Check sizes. The smaller one needs to be larger than the Karatsuba cut-off.
        * The bigger one needs to be at least about one MP_KARATSUBA_MUL_CUTOFF bigger
        * to make some sense, but it depends on architecture, OS, position of the
        * stars... so YMMV.
        * Using it to cut the input into slices small enough for fast_s_mp_mul_digs
        * was actually slower on the author's machine, but YMMV.
        */