feedc0de/wolfssl
1
0
Fork 0
forked from wolfSSL/wolfssl
Code Pull Requests Activity

Adds TFM_MIPS (tested on PIC32MZ2048EFM144 with XC32 v1.44). Adds check for mp_read_radix to return failure for invalid character. Adds checking for s_mp_sub response code. Improved mp_init_copy error handling. Improved init to use size where applicable. Asm PPC 32 changes to add r to register. Asm formatting cleanups.

Browse Source
This commit is contained in:
David Garske
2017-09-22 09:48:35 -07:00
parent 824eb55648
commit a2ed22286f
3 changed files with 363 additions and 172 deletions
Download Patch File Download Diff File Expand all files Collapse all files

260
wolfcrypt/src/asm.c
View File

@ -156,7 +156,7 @@ __asm__( \
#define MONT_FINI
#define LOOP_END
#define LOOP_START \
mu = c[x] * mp;
mu = c[x] * mp
#define INNERMUL \
__asm__( \
@ -333,7 +333,7 @@ __asm__( \
\
:"=r"(_c), "=r"(cy) \
: "0"(_c), "1"(cy), "g"(mu), "r"(tmpm)\
: "%rax", "%rdx", "%r10", "%r11", "cc")\
: "%rax", "%rdx", "%r10", "%r11", "cc")
#define PROPCARRY \
__asm__( \
@ -560,24 +560,24 @@ __asm__( \
#define INNERMUL \
__asm__( \
" mulld 16,%3,%4 \n\t" \
" mulhdu 17,%3,%4 \n\t" \
" addc 16,16,%0 \n\t" \
" addze 17,17 \n\t" \
" ldx 18,0,%1 \n\t" \
" addc 16,16,18 \n\t" \
" addze %0,17 \n\t" \
" sdx 16,0,%1 \n\t" \
:"=r"(cy),"=m"(_c[0]):"0"(cy),"r"(mu),"r"(tmpm[0]),"1"(_c[0]):"16", "17", "18","cc"); ++tmpm;
" mulld r16,%3,%4 \n\t" \
" mulhdu r17,%3,%4 \n\t" \
" addc r16,16,%0 \n\t" \
" addze r17,r17 \n\t" \
" ldx r18,0,%1 \n\t" \
" addc r16,r16,r18 \n\t" \
" addze %0,r17 \n\t" \
" sdx r16,0,%1 \n\t" \
:"=r"(cy),"=m"(_c[0]):"0"(cy),"r"(mu),"r"(tmpm[0]),"1"(_c[0]):"r16", "r17", "r18","cc"); ++tmpm;
#define PROPCARRY \
__asm__( \
" ldx 16,0,%1 \n\t" \
" addc 16,16,%0 \n\t" \
" sdx 16,0,%1 \n\t" \
" ldx r16,0,%1 \n\t" \
" addc r16,r16,%0 \n\t" \
" sdx r16,0,%1 \n\t" \
" xor %0,%0,%0 \n\t" \
" addze %0,%0 \n\t" \
:"=r"(cy),"=m"(_c[0]):"0"(cy),"1"(_c[0]):"16","cc");
:"=r"(cy),"=m"(_c[0]):"0"(cy),"1"(_c[0]):"r16","cc");
/******************************************************************/
@ -610,6 +610,40 @@ __asm__( \
" acr %0 \n\t" \
:"=r"(cy),"=r"(&_c[0]):"0"(cy),"1"(&_c[0]):"r2","cc");
/******************************************************************/
#elif defined(TFM_MIPS)
/* MIPS */
#define MONT_START
#define MONT_FINI
#define LOOP_END
#define LOOP_START \
mu = c[x] * mp
#define INNERMUL \
__asm__( \
" multu %3,%4 \n\t" \
" mflo $12 \n\t" \
" mfhi $13 \n\t" \
" addu $12,$12,%0 \n\t" \
" sltu $10,$12,%0 \n\t" \
" addu $13,$13,$10 \n\t" \
" lw $10,%1 \n\t" \
" addu $12,$12,$10 \n\t" \
" sltu $10,$12,$10 \n\t" \
" addu %0,$13,$10 \n\t" \
" sw $12,%1 \n\t" \
:"+r"(cy),"+m"(_c[0]):""(cy),"r"(mu),"r"(tmpm[0]),""(_c[0]):"$10","$12","$13"); ++tmpm;
#define PROPCARRY \
__asm__( \
" lw $10,%1 \n\t" \
" addu $10,$10,%0 \n\t" \
" sw $10,%1 \n\t" \
" sltu %0,$10,%0 \n\t" \
:"+r"(cy),"+m"(_c[0]):""(cy),""(_c[0]):"$10");
/******************************************************************/
#else
/* ISO C code */
@ -689,8 +723,6 @@ __asm__( \
"xorl %2,%2 \n\t" \
:"=r"(sc0), "=r"(sc1), "=r"(sc2): "g"(i), "g"(j) :"%eax","%edx","cc");
/* TAO removed sc0,1,2 as input to remove warning so %6,%7 become %3,%4 */
#define SQRADDAC(i, j) \
__asm__( \
"movl %6,%%eax \n\t" \
@ -759,8 +791,6 @@ __asm__( \
"xorq %2,%2 \n\t" \
:"=r"(sc0), "=r"(sc1), "=r"(sc2): "g"(i), "g"(j) :"%rax","%rdx","cc");
/* TAO removed sc0,1,2 as input to remove warning so %6,%7 become %3,%4 */
#define SQRADDAC(i, j) \
__asm__( \
"movq %6,%%rax \n\t" \
@ -1022,25 +1052,25 @@ __asm__( \
/* multiplies point i and j, updates carry "c1" and digit c2 */
#define SQRADD(i, j) \
__asm__( \
" mulld 16,%6,%6 \n\t" \
" addc %0,%0,16 \n\t" \
" mulhdu 16,%6,%6 \n\t" \
" adde %1,%1,16 \n\t" \
" mulld r16,%6,%6 \n\t" \
" addc %0,%0,r16 \n\t" \
" mulhdu r16,%6,%6 \n\t" \
" adde %1,%1,r16 \n\t" \
" addze %2,%2 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i):"16","cc");
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i):"r16","cc");
/* for squaring some of the terms are doubled... */
#define SQRADD2(i, j) \
__asm__( \
" mulld 16,%6,%7 \n\t" \
" mulhdu 17,%6,%7 \n\t" \
" addc %0,%0,16 \n\t" \
" adde %1,%1,17 \n\t" \
" mulld r16,%6,%7 \n\t" \
" mulhdu r17,%6,%7 \n\t" \
" addc %0,%0,r16 \n\t" \
" adde %1,%1,r17 \n\t" \
" addze %2,%2 \n\t" \
" addc %0,%0,16 \n\t" \
" adde %1,%1,17 \n\t" \
" addc %0,%0,r16 \n\t" \
" adde %1,%1,r17 \n\t" \
" addze %2,%2 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"16", "17","cc");
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"r16", "r17","cc");
#define SQRADDSC(i, j) \
__asm__( \
@ -1051,12 +1081,12 @@ __asm__( \
#define SQRADDAC(i, j) \
__asm__( \
" mulld 16,%6,%7 \n\t" \
" addc %0,%0,16 \n\t" \
" mulhdu 16,%6,%7 \n\t" \
" adde %1,%1,16 \n\t" \
" mulld r16,%6,%7 \n\t" \
" addc %0,%0,r16 \n\t" \
" mulhdu r16,%6,%7 \n\t" \
" adde %1,%1,r16 \n\t" \
" addze %2,%2 \n\t" \
:"=r"(sc0), "=r"(sc1), "=r"(sc2):"0"(sc0), "1"(sc1), "2"(sc2), "r"(i), "r"(j):"16", "cc");
:"=r"(sc0), "=r"(sc1), "=r"(sc2):"0"(sc0), "1"(sc1), "2"(sc2), "r"(i), "r"(j):"r16", "cc");
#define SQRADDDB \
__asm__( \
@ -1136,6 +1166,112 @@ __asm__( \
" adc %2,%2,%5 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2) : "r"(sc0), "r"(sc1), "r"(sc2), "0"(c0), "1"(c1), "2"(c2) : "cc");
#elif defined(TFM_MIPS)
/* MIPS */
#define COMBA_START
#define CLEAR_CARRY \
c0 = c1 = c2 = 0;
#define COMBA_STORE(x) \
x = c0;
#define COMBA_STORE2(x) \
x = c1;
#define CARRY_FORWARD \
do { c0 = c1; c1 = c2; c2 = 0; } while (0);
#define COMBA_FINI
/* multiplies point i and j, updates carry "c1" and digit c2 */
#define SQRADD(i, j) \
__asm__( \
" multu %6,%6 \n\t" \
" mflo $12 \n\t" \
" mfhi $13 \n\t" \
" addu %0,%0,$12 \n\t" \
" sltu $12,%0,$12 \n\t" \
" addu %1,%1,$13 \n\t" \
" sltu $13,%1,$13 \n\t" \
" addu %1,%1,$12 \n\t" \
" sltu $12,%1,$12 \n\t" \
" addu %2,%2,$13 \n\t" \
" addu %2,%2,$12 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i):"$12","$13");
/* for squaring some of the terms are doubled... */
#define SQRADD2(i, j) \
__asm__( \
" multu %6,%7 \n\t" \
" mflo $12 \n\t" \
" mfhi $13 \n\t" \
\
" addu %0,%0,$12 \n\t" \
" sltu $14,%0,$12 \n\t" \
" addu %1,%1,$13 \n\t" \
" sltu $15,%1,$13 \n\t" \
" addu %1,%1,$14 \n\t" \
" sltu $14,%1,$14 \n\t" \
" addu %2,%2,$15 \n\t" \
" addu %2,%2,$14 \n\t" \
\
" addu %0,%0,$12 \n\t" \
" sltu $14,%0,$12 \n\t" \
" addu %1,%1,$13 \n\t" \
" sltu $15,%1,$13 \n\t" \
" addu %1,%1,$14 \n\t" \
" sltu $14,%1,$14 \n\t" \
" addu %2,%2,$15 \n\t" \
" addu %2,%2,$14 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"$12", "$13", "$14", "$15");
#define SQRADDSC(i, j) \
__asm__( \
" multu %6,%7 \n\t" \
" mflo %0 \n\t" \
" mfhi %1 \n\t" \
" xor %2,%2,%2 \n\t" \
:"=r"(sc0), "=r"(sc1), "=r"(sc2):"0"(sc0), "1"(sc1), "2"(sc2), "r"(i),"r"(j) : "cc");
#define SQRADDAC(i, j) \
__asm__( \
" multu %6,%7 \n\t" \
" mflo $12 \n\t" \
" mfhi $13 \n\t" \
" addu %0,%0,$12 \n\t" \
" sltu $12,%0,$12 \n\t" \
" addu %1,%1,$13 \n\t" \
" sltu $13,%1,$13 \n\t" \
" addu %1,%1,$12 \n\t" \
" sltu $12,%1,$12 \n\t" \
" addu %2,%2,$13 \n\t" \
" addu %2,%2,$12 \n\t" \
:"=r"(sc0), "=r"(sc1), "=r"(sc2):"0"(sc0), "1"(sc1), "2"(sc2), "r"(i), "r"(j):"$12", "$13", "$14");
#define SQRADDDB \
__asm__( \
" addu %0,%0,%3 \n\t" \
" sltu $10,%0,%3 \n\t" \
" addu %1,%1,$10 \n\t" \
" sltu $10,%1,$10 \n\t" \
" addu %1,%1,%4 \n\t" \
" sltu $11,%1,%4 \n\t" \
" addu %2,%2,$10 \n\t" \
" addu %2,%2,$11 \n\t" \
" addu %2,%2,%5 \n\t" \
\
" addu %0,%0,%3 \n\t" \
" sltu $10,%0,%3 \n\t" \
" addu %1,%1,$10 \n\t" \
" sltu $10,%1,$10 \n\t" \
" addu %1,%1,%4 \n\t" \
" sltu $11,%1,%4 \n\t" \
" addu %2,%2,$10 \n\t" \
" addu %2,%2,$11 \n\t" \
" addu %2,%2,%5 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2) : "r"(sc0), "r"(sc1), "r"(sc2), "0"(c0), "1"(c1), "2"(c2) : "$10", "$11");
#else
@ -1174,10 +1310,10 @@ __asm__( \
t = ((fp_word)i) * ((fp_word)j); \
tt = (fp_word)c0 + t; c0 = (fp_digit)tt; \
tt = (fp_word)c1 + (tt >> DIGIT_BIT); c1 = (fp_digit)tt; \
c2 +=(fp_digit)( tt >> DIGIT_BIT); \
c2 +=(fp_digit)(tt >> DIGIT_BIT); \
tt = (fp_word)c0 + t; c0 = (fp_digit)tt; \
tt = (fp_word)c1 + (tt >> DIGIT_BIT); c1 = (fp_digit)tt; \
c2 +=(fp_digit) (tt >> DIGIT_BIT); \
c2 +=(fp_digit)(tt >> DIGIT_BIT); \
} while (0);
#define SQRADDSC(i, j) \
@ -1497,15 +1633,15 @@ __asm__( \
#define COMBA_FINI
/* untested: will mulhwu change the flags? Docs say no */
/* untested: will mulhdu change the flags? Docs say no */
#define MULADD(i, j) \
____asm__( \
" mulld 16,%6,%7 \n\t" \
" mulld r16,%6,%7 \n\t" \
" addc %0,%0,16 \n\t" \
" mulhdu 16,%6,%7 \n\t" \
" mulhdu r16,%6,%7 \n\t" \
" adde %1,%1,16 \n\t" \
" addze %2,%2 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"16");
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"r16");
#elif defined(TFM_AVR32)
@ -1535,6 +1671,40 @@ ____asm__( \
" acr %2 \n\t"\
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"r2","r3");
#elif defined(TFM_MIPS)
/* MIPS */
#define COMBA_START
#define COMBA_CLEAR \
c0 = c1 = c2 = 0;
#define COMBA_FORWARD \
do { c0 = c1; c1 = c2; c2 = 0; } while (0);
#define COMBA_STORE(x) \
x = c0;
#define COMBA_STORE2(x) \
x = c1;
#define COMBA_FINI
#define MULADD(i, j) \
__asm__( \
" multu %6,%7 \n\t" \
" mflo $12 \n\t" \
" mfhi $13 \n\t" \
" addu %0,%0,$12 \n\t" \
" sltu $12,%0,$12 \n\t" \
" addu %1,%1,$13 \n\t" \
" sltu $13,%1,$13 \n\t" \
" addu %1,%1,$12 \n\t" \
" sltu $12,%1,$12 \n\t" \
" addu %2,%2,$13 \n\t" \
" addu %2,%2,$12 \n\t" \
:"=r"(c0), "=r"(c1), "=r"(c2):"0"(c0), "1"(c1), "2"(c2), "r"(i), "r"(j):"$12","$13");
#else
/* ISO C code */
@ -1556,9 +1726,11 @@ ____asm__( \
#define MULADD(i, j) \
do { fp_word t; \
t = (fp_word)c0 + ((fp_word)i) * ((fp_word)j); c0 = (fp_digit)t; \
t = (fp_word)c0 + ((fp_word)i) * ((fp_word)j); \
c0 = (fp_digit)t; \
t = (fp_word)c1 + (t >> DIGIT_BIT); \
c1 = (fp_digit)t; c2 += (fp_digit)(t >> DIGIT_BIT); \
c1 = (fp_digit)t; \
c2 += (fp_digit)(t >> DIGIT_BIT); \
} while (0);
#endif

2
wolfcrypt/src/fp_mont_small.i
View File

@ -29,7 +29,7 @@ void fp_montgomery_reduce_small(fp_int *a, fp_int *m, fp_digit mp)
int oldused, x, y, pa;
/* now zero the buff */
XMEMSET(c, 0, sizeof c);
XMEMSET(c, 0, sizeof(c));
pa = m->used;

57
wolfcrypt/src/integer.c
View File

@ -325,10 +325,15 @@ int mp_init_copy (mp_int * a, mp_int * b)
{
int res;
if ((res = mp_init (a)) != MP_OKAY) {
if ((res = mp_init_size (a, b->used)) != MP_OKAY) {
return res;
}
return mp_copy (b, a);
if((res = mp_copy (b, a)) != MP_OKAY) {
mp_clear(a);
}
return res;
}
@ -929,13 +934,15 @@ int mp_invmod (mp_int * a, mp_int * b, mp_int * c)
#ifdef BN_FAST_MP_INVMOD_C
/* if the modulus is odd we can use a faster routine instead */
if (mp_isodd (b) == MP_YES) {
if ((mp_isodd(b) == MP_YES) && (mp_cmp_d(b, 1) != MP_EQ)) {
return fast_mp_invmod (a, b, c);
}
#endif
#ifdef BN_MP_INVMOD_SLOW_C
return mp_invmod_slow(a, b, c);
#else
return MP_VAL;
#endif
}
@ -1379,7 +1386,7 @@ int mp_mod (mp_int * a, mp_int * b, mp_int * c)
mp_int t;
int res;
if ((res = mp_init (&t)) != MP_OKAY) {
if ((res = mp_init_size (&t, b->used)) != MP_OKAY) {
return res;
}
@ -1388,11 +1395,11 @@ int mp_mod (mp_int * a, mp_int * b, mp_int * c)
return res;
}
if (t.sign != b->sign) {
res = mp_add (b, &t, c);
} else {
if ((mp_iszero(&t) != MP_NO) || (t.sign == b->sign)) {
res = MP_OKAY;
mp_exch (&t, c);
} else {
res = mp_add (b, &t, c);
}
mp_clear (&t);
@ -1891,7 +1898,7 @@ int mp_exptmod_fast (mp_int * G, mp_int * X, mp_int * P, mp_int * Y,
/* init M array */
/* init first cell */
if ((err = mp_init(&M[1])) != MP_OKAY) {
if ((err = mp_init_size(&M[1], P->alloc)) != MP_OKAY) {
#ifdef WOLFSSL_SMALL_STACK
XFREE(M, NULL, DYNAMIC_TYPE_TMP_BUFFER);
#endif
@ -1901,7 +1908,7 @@ int mp_exptmod_fast (mp_int * G, mp_int * X, mp_int * P, mp_int * Y,
/* now init the second half of the array */
for (x = 1<<(winsize-1); x < (1 << winsize); x++) {
if ((err = mp_init(&M[x])) != MP_OKAY) {
if ((err = mp_init_size(&M[x], P->alloc)) != MP_OKAY) {
for (y = 1<<(winsize-1); y < x; y++) {
mp_clear (&M[y]);
}
@ -1967,7 +1974,7 @@ int mp_exptmod_fast (mp_int * G, mp_int * X, mp_int * P, mp_int * Y,
}
/* setup result */
if ((err = mp_init (&res)) != MP_OKAY) {
if ((err = mp_init_size (&res, P->alloc)) != MP_OKAY) {
goto LBL_M;
}
@ -1984,15 +1991,15 @@ int mp_exptmod_fast (mp_int * G, mp_int * X, mp_int * P, mp_int * Y,
if ((err = mp_montgomery_calc_normalization (&res, P)) != MP_OKAY) {
goto LBL_RES;
}
#else
err = MP_VAL;
goto LBL_RES;
#endif
/* now set M[1] to G * R mod m */
if ((err = mp_mulmod (G, &res, P, &M[1])) != MP_OKAY) {
goto LBL_RES;
}
#else
err = MP_VAL;
goto LBL_RES;
#endif
} else {
if ((err = mp_set(&res, 1)) != MP_OKAY) {
goto LBL_RES;
@ -2536,7 +2543,9 @@ top:
* Each successive "recursion" makes the input smaller and smaller.
*/
if (mp_cmp_mag (x, n) != MP_LT) {
s_mp_sub(x, n, x);
if ((err = s_mp_sub(x, n, x)) != MP_OKAY) {
return err;
}
goto top;
}
return MP_OKAY;
@ -2573,7 +2582,9 @@ top:
}
if (mp_cmp_mag(a, n) != MP_LT) {
s_mp_sub(a, n, a);
if ((res = s_mp_sub(a, n, a)) != MP_OKAY) {
goto ERR;
}
goto top;
}
@ -2711,7 +2722,7 @@ int mp_mulmod (mp_int * a, mp_int * b, mp_int * c, mp_int * d)
int res;
mp_int t;
if ((res = mp_init (&t)) != MP_OKAY) {
if ((res = mp_init_size (&t, c->used)) != MP_OKAY) {
return res;
}
@ -3732,7 +3743,9 @@ top:
}
if (mp_cmp_mag(a, n) != MP_LT) {
s_mp_sub(a, n, a);
if ((res = s_mp_sub(a, n, a)) != MP_OKAY) {
goto ERR;
}
goto top;
}
@ -4751,7 +4764,7 @@ int mp_read_radix (mp_int * a, const char *str, int radix)
mp_zero (a);
/* process each digit of the string */
while (*str) {
while (*str != '\0') {
/* if the radix <= 36 the conversion is case insensitive
* this allows numbers like 1AB and 1ab to represent the same value
* [e.g. in hex]
@ -4780,6 +4793,12 @@ int mp_read_radix (mp_int * a, const char *str, int radix)
++str;
}
/* if digit in isn't null term, then invalid character was found */
if (*str != '\0') {
mp_zero (a);
return MP_VAL;
}
/* set the sign only if a != 0 */
if (mp_iszero(a) != MP_YES) {
a->sign = neg;