mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-07-28 21:40:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Improve memory safety

Browse source 
This commit makes numerous refinements to cosmopolitan memory handling.

The default stack size has been reduced from 2mb to 128kb. A new macro
is now provided so you can easily reconfigure the stack size to be any
value you want. Work around the breaking change by adding to your main:

    STATIC_STACK_SIZE(0x00200000);  // 2mb stack

If you're not sure how much stack you need, then you can use:

    STATIC_YOINK("stack_usage_logging");

After which you can `sort -nr o/$MODE/stack.log`. Based on the unit test
suite, nothing in the Cosmopolitan repository (except for Python) needs
a stack size greater than 30kb. There are also new macros for detecting
the size and address of the stack at runtime, e.g. GetStackAddr(). We
also now support sigaltstack() so if you want to see nice looking crash
reports whenever a stack overflow happens, you can put this in main():

    ShowCrashReports();

Under `make MODE=dbg` and `make MODE=asan` the unit testing framework
will now automatically print backtraces of memory allocations when
things like memory leaks happen. Bugs are now fixed in ASAN global
variable overrun detection. The memtrack and asan runtimes also handle
edge cases now. The new tools helped to identify a few memory leaks,
which are fixed by this change.

This change should fix an issue reported in #288 with ARG_MAX limits.
Fixing this doubled the performance of MKDEPS.COM and AR.COM yet again.
This commit is contained in:
Justine Tunney 2021-10-13 17:27:13 -07:00
parent a0b39f886c
commit 226aaf3547
317 changed files with 6474 additions and 3993 deletions
Download patch file Download diff file Expand all files Collapse all files

184
third_party/gdtoa/strtodg.c vendored
View file

@ -39,7 +39,7 @@ fivesbits[] = { 0, 3, 5, 7, 10, 12, 14, 17, 19, 21,
47, 49, 52 };
Bigint *
increment(Bigint *b)
__gdtoa_increment(Bigint *b)
{
ULong *x, *xe;
Bigint *b1;
@ -54,9 +54,9 @@ increment(Bigint *b)
} while(x < xe);
{
if (b->wds >= b->maxwds) {
b1 = Balloc(b->k+1);
b1 = __gdtoa_Balloc(b->k+1);
Bcopy(b1,b);
Bfree(b);
__gdtoa_Bfree(b);
b = b1;
}
b->x[b->wds++] = 1;
@ -65,7 +65,7 @@ increment(Bigint *b)
}
void
decrement(Bigint *b)
__gdtoa_decrement(Bigint *b)
{
ULong *x, *xe;
x = b->x;
@ -95,14 +95,14 @@ all_on(Bigint *b, int n)
}
Bigint *
set_ones(Bigint *b, int n)
__gdtoa_set_ones(Bigint *b, int n)
{
int k;
ULong *x, *xe;
k = (n + ((1 << kshift) - 1)) >> kshift;
if (b->k < k) {
Bfree(b);
b = Balloc(k);
__gdtoa_Bfree(b);
b = __gdtoa_Balloc(k);
}
k = n >> kshift;
if (n &= kmask)
@ -124,7 +124,7 @@ rvOK(U *d, const FPI *fpi, Long *exp, ULong *bits, int exact, int rd, int *irv)
ULong carry, inex, lostbits;
int bdif, e, j, k, k1, nb, rv;
carry = rv = 0;
b = d2b(dval(d), &e, &bdif);
b = __gdtoa_d2b(dval(d), &e, &bdif);
bdif -= nb = fpi->nbits;
e += bdif;
if (bdif <= 0) {
@ -161,24 +161,24 @@ rvOK(U *d, const FPI *fpi, Long *exp, ULong *bits, int exact, int rd, int *irv)
trunc:
inex = lostbits = 0;
if (bdif > 0) {
if ( (lostbits = any_on(b, bdif)) !=0)
if ( (lostbits = __gdtoa_any_on(b, bdif)) !=0)
inex = STRTOG_Inexlo;
rshift(b, bdif);
__gdtoa_rshift(b, bdif);
if (carry) {
inex = STRTOG_Inexhi;
b = increment(b);
b = __gdtoa_increment(b);
if ( (j = nb & kmask) !=0)
j = ULbits - j;
if (hi0bits(b->x[b->wds - 1]) != j) {
if (!lostbits)
lostbits = b->x[0] & 1;
rshift(b, 1);
__gdtoa_rshift(b, 1);
e++;
}
}
}
else if (bdif < 0)
b = lshift(b, -bdif);
b = __gdtoa_lshift(b, -bdif);
if (e < fpi->emin) {
k = fpi->emin - e;
e = fpi->emin;
@ -189,15 +189,15 @@ trunc:
else {
k1 = k - 1;
if (k1 > 0 && !lostbits)
lostbits = any_on(b, k1);
lostbits = __gdtoa_any_on(b, k1);
if (!lostbits && !exact)
goto ret;
lostbits |=
carry = b->x[k1>>kshift] & (1 << (k1 & kmask));
rshift(b, k);
__gdtoa_rshift(b, k);
*irv = STRTOG_Denormal;
if (carry) {
b = increment(b);
b = __gdtoa_increment(b);
inex = STRTOG_Inexhi | STRTOG_Underflow;
}
else if (lostbits)
@ -211,11 +211,11 @@ trunc:
b->wds = inex = 0;
}
*exp = e;
copybits(bits, nb, b);
__gdtoa_copybits(bits, nb, b);
*irv |= inex;
rv = 1;
ret:
Bfree(b);
__gdtoa_Bfree(b);
return rv;
}
@ -276,7 +276,7 @@ break2:
switch(s[1]) {
case 'x':
case 'X':
irv = gethex(&s, fpi, exp, &rvb, sign);
irv = __gdtoa_gethex(&s, fpi, exp, &rvb, sign);
if (irv == STRTOG_NoNumber) {
s = s00;
sign = 0;
@ -376,9 +376,9 @@ dig_done:
switch(c) {
case 'i':
case 'I':
if (match(&s,"nf")) {
if (__gdtoa_match(&s,"nf")) {
--s;
if (!match(&s,"inity"))
if (!__gdtoa_match(&s,"inity"))
++s;
irv = STRTOG_Infinite;
goto infnanexp;
@ -386,11 +386,11 @@ dig_done:
break;
case 'n':
case 'N':
if (match(&s, "an")) {
if (__gdtoa_match(&s, "an")) {
irv = STRTOG_NaN;
*exp = fpi->emax + 1;
if (*s == '(')
irv = hexnan(&s, fpi, bits);
irv = __gdtoa_hexnan(&s, fpi, bits);
goto infnanexp;
}
}
@ -421,7 +421,7 @@ dig_done:
k = nd < DBL_DIG + 2 ? nd : DBL_DIG + 2;
dval(&rv) = y;
if (k > 9)
dval(&rv) = tens[k - 9] * dval(&rv) + z;
dval(&rv) = __gdtoa_tens[k - 9] * dval(&rv) + z;
bd0 = 0;
if (nbits <= P && nd <= DBL_DIG) {
if (!e) {
@ -431,7 +431,7 @@ dig_done:
else if (e > 0) {
if (e <= Ten_pmax) {
i = fivesbits[e] + mantbits(&rv) <= P;
/* rv = */ rounded_product(dval(&rv), tens[e]);
/* rv = */ rounded_product(dval(&rv), __gdtoa_tens[e]);
if (rvOK(&rv, fpi, exp, bits, i, rd, &irv))
goto ret;
e1 -= e;
@ -444,15 +444,15 @@ dig_done:
*/
e2 = e - i;
e1 -= i;
dval(&rv) *= tens[i];
/* rv = */ rounded_product(dval(&rv), tens[e2]);
dval(&rv) *= __gdtoa_tens[i];
/* rv = */ rounded_product(dval(&rv), __gdtoa_tens[e2]);
if (rvOK(&rv, fpi, exp, bits, 0, rd, &irv))
goto ret;
e1 -= e2;
}
}
else if (e >= -Ten_pmax) {
/* rv = */ rounded_quotient(dval(&rv), tens[-e]);
/* rv = */ rounded_quotient(dval(&rv), __gdtoa_tens[-e]);
if (rvOK(&rv, fpi, exp, bits, 0, rd, &irv))
goto ret;
e1 -= e;
@ -464,51 +464,51 @@ rv_notOK:
e2 = 0;
if (e1 > 0) {
if ( (i = e1 & 15) !=0)
dval(&rv) *= tens[i];
dval(&rv) *= __gdtoa_tens[i];
if (e1 &= ~15) {
e1 >>= 4;
while(e1 >= (1 << (n_bigtens-1))) {
while(e1 >= (1 << (n___gdtoa_bigtens-1))) {
e2 += ((word0(&rv) & Exp_mask)
>> Exp_shift1) - Bias;
word0(&rv) &= ~Exp_mask;
word0(&rv) |= Bias << Exp_shift1;
dval(&rv) *= bigtens[n_bigtens-1];
e1 -= 1 << (n_bigtens-1);
dval(&rv) *= __gdtoa_bigtens[n___gdtoa_bigtens-1];
e1 -= 1 << (n___gdtoa_bigtens-1);
}
e2 += ((word0(&rv) & Exp_mask) >> Exp_shift1) - Bias;
word0(&rv) &= ~Exp_mask;
word0(&rv) |= Bias << Exp_shift1;
for(j = 0; e1 > 0; j++, e1 >>= 1)
if (e1 & 1)
dval(&rv) *= bigtens[j];
dval(&rv) *= __gdtoa_bigtens[j];
}
}
else if (e1 < 0) {
e1 = -e1;
if ( (i = e1 & 15) !=0)
dval(&rv) /= tens[i];
dval(&rv) /= __gdtoa_tens[i];
if (e1 &= ~15) {
e1 >>= 4;
while(e1 >= (1 << (n_bigtens-1))) {
while(e1 >= (1 << (n___gdtoa_bigtens-1))) {
e2 += ((word0(&rv) & Exp_mask)
>> Exp_shift1) - Bias;
word0(&rv) &= ~Exp_mask;
word0(&rv) |= Bias << Exp_shift1;
dval(&rv) *= tinytens[n_bigtens-1];
e1 -= 1 << (n_bigtens-1);
dval(&rv) *= __gdtoa_tinytens[n___gdtoa_bigtens-1];
e1 -= 1 << (n___gdtoa_bigtens-1);
}
e2 += ((word0(&rv) & Exp_mask) >> Exp_shift1) - Bias;
word0(&rv) &= ~Exp_mask;
word0(&rv) |= Bias << Exp_shift1;
for(j = 0; e1 > 0; j++, e1 >>= 1)
if (e1 & 1)
dval(&rv) *= tinytens[j];
dval(&rv) *= __gdtoa_tinytens[j];
}
}
rvb = d2b(dval(&rv), &rve, &rvbits); /* rv = rvb * 2^rve */
rvb = __gdtoa_d2b(dval(&rv), &rve, &rvbits); /* rv = rvb * 2^rve */
rve += e2;
if ((j = rvbits - nbits) > 0) {
rshift(rvb, j);
__gdtoa_rshift(rvb, j);
rvbits = nbits;
rve += j;
}
@ -521,7 +521,7 @@ rv_notOK:
denorm = 1;
j = rve - emin;
if (j > 0) {
rvb = lshift(rvb, j);
rvb = __gdtoa_lshift(rvb, j);
rvbits += j;
}
else if (j < 0) {
@ -549,7 +549,7 @@ rv_notOK:
rvb->x[0] = rvb->wds = rvbits = 1;
}
else
rshift(rvb, -j);
__gdtoa_rshift(rvb, -j);
}
rve = rve1 = emin;
if (sudden_underflow && e2 + 1 < emin)
@ -557,15 +557,15 @@ rv_notOK:
}
/* Now the hard part -- adjusting rv to the correct value.*/
/* Put digits into bd: true value = bd * 10^e */
bd0 = s2b(s0, nd0, nd, y, 1);
bd0 = __gdtoa_s2b(s0, nd0, nd, y, 1);
for(;;) {
bd = Balloc(bd0->k);
bd = __gdtoa_Balloc(bd0->k);
Bcopy(bd, bd0);
bb = Balloc(rvb->k);
bb = __gdtoa_Balloc(rvb->k);
Bcopy(bb, rvb);
bbbits = rvbits - bb0;
bbe = rve + bb0;
bs = i2b(1);
bs = __gdtoa_i2b(1);
if (e >= 0) {
bb2 = bb5 = 0;
bd2 = bd5 = e;
@ -594,31 +594,31 @@ rv_notOK:
bs2 -= i;
}
if (bb5 > 0) {
bs = pow5mult(bs, bb5);
bb1 = mult(bs, bb);
Bfree(bb);
bs = __gdtoa_pow5mult(bs, bb5);
bb1 = __gdtoa_mult(bs, bb);
__gdtoa_Bfree(bb);
bb = bb1;
}
bb2 -= bb0;
if (bb2 > 0)
bb = lshift(bb, bb2);
bb = __gdtoa_lshift(bb, bb2);
else if (bb2 < 0)
rshift(bb, -bb2);
__gdtoa_rshift(bb, -bb2);
if (bd5 > 0)
bd = pow5mult(bd, bd5);
bd = __gdtoa_pow5mult(bd, bd5);
if (bd2 > 0)
bd = lshift(bd, bd2);
bd = __gdtoa_lshift(bd, bd2);
if (bs2 > 0)
bs = lshift(bs, bs2);
bs = __gdtoa_lshift(bs, bs2);
asub = 1;
inex = STRTOG_Inexhi;
delta = diff(bb, bd);
delta = __gdtoa_diff(bb, bd);
if (delta->wds <= 1 && !delta->x[0])
break;
dsign = delta->sign;
delta->sign = finished = 0;
L = 0;
i = cmp(delta, bs);
i = __gdtoa_cmp(delta, bs);
if (rd && i <= 0) {
irv = STRTOG_Normal;
if ( (finished = dsign ^ (rd&1)) !=0) {
@ -637,14 +637,14 @@ rv_notOK:
if (j > 1 && lo0bits(rvb->x + i) < j - 1)
goto adj1;
rve = rve1 - 1;
rvb = set_ones(rvb, rvbits = nbits);
rvb = __gdtoa_set_ones(rvb, rvbits = nbits);
break;
}
irv |= dsign ? STRTOG_Inexlo : STRTOG_Inexhi;
break;
}
if (i < 0) {
/* Error is less than half an ulp -- check for
/* Error is less than half an __gdtoa_ulp -- check for
* special case of mantissa a power of two.
*/
irv = dsign
@ -652,8 +652,8 @@ rv_notOK:
: STRTOG_Normal | STRTOG_Inexhi;
if (dsign || bbbits > 1 || denorm || rve1 == emin)
break;
delta = lshift(delta,1);
if (cmp(delta, bs) > 0) {
delta = __gdtoa_lshift(delta,1);
if (__gdtoa_cmp(delta, bs) > 0) {
irv = STRTOG_Normal | STRTOG_Inexlo;
goto drop_down;
}
@ -663,7 +663,7 @@ rv_notOK:
/* exactly half-way between */
if (dsign) {
if (denorm && all_on(rvb, rvbits)) {
/*boundary case -- increment exponent*/
/*boundary case -- __gdtoa_increment exponent*/
rvb->wds = 1;
rvb->x[0] = 1;
rve = emin + nbits - (rvbits = 1);
@ -676,7 +676,7 @@ rv_notOK:
else if (bbbits == 1) {
irv = STRTOG_Normal;
drop_down:
/* boundary case -- decrement exponent */
/* boundary case -- __gdtoa_decrement exponent */
if (rve1 == emin) {
irv = STRTOG_Normal | STRTOG_Inexhi;
if (rvb->wds == 1 && rvb->x[0] == 1)
@ -684,7 +684,7 @@ rv_notOK:
break;
}
rve -= nbits;
rvb = set_ones(rvb, rvbits = nbits);
rvb = __gdtoa_set_ones(rvb, rvbits = nbits);
break;
}
else
@ -692,7 +692,7 @@ rv_notOK:
if ((bbbits < nbits && !denorm) || !(rvb->x[0] & 1))
break;
if (dsign) {
rvb = increment(rvb);
rvb = __gdtoa_increment(rvb);
j = kmask & (ULbits - (rvbits & kmask));
if (hi0bits(rvb->x[rvb->wds - 1]) != j)
rvbits++;
@ -701,12 +701,12 @@ rv_notOK:
else {
if (bbbits == 1)
goto undfl;
decrement(rvb);
__gdtoa_decrement(rvb);
irv = STRTOG_Normal | STRTOG_Inexlo;
}
break;
}
if ((dval(&adj) = ratio(delta, bs)) <= 2.) {
if ((dval(&adj) = __gdtoa_ratio(delta, bs)) <= 2.) {
adj1:
inex = STRTOG_Inexlo;
if (dsign) {
@ -756,23 +756,23 @@ rv_notOK:
}
}
y = rve + rvbits;
/* adj *= ulp(dval(&rv)); */
/* adj *= __gdtoa_ulp(dval(&rv)); */
/* if (asub) rv -= adj; else rv += adj; */
if (!denorm && rvbits < nbits) {
rvb = lshift(rvb, j = nbits - rvbits);
rvb = __gdtoa_lshift(rvb, j = nbits - rvbits);
rve -= j;
rvbits = nbits;
}
ab = d2b(dval(&adj), &abe, &abits);
ab = __gdtoa_d2b(dval(&adj), &abe, &abits);
if (abe < 0)
rshift(ab, -abe);
__gdtoa_rshift(ab, -abe);
else if (abe > 0)
ab = lshift(ab, abe);
ab = __gdtoa_lshift(ab, abe);
rvb0 = rvb;
if (asub) {
/* rv -= adj; */
j = hi0bits(rvb->x[rvb->wds-1]);
rvb = diff(rvb, ab);
rvb = __gdtoa_diff(rvb, ab);
k = rvb0->wds - 1;
if (denorm)
/* do nothing */;
@ -785,7 +785,7 @@ rv_notOK:
denorm = 1;
}
else {
rvb = lshift(rvb, 1);
rvb = __gdtoa_lshift(rvb, 1);
--rve;
--rve1;
L = finished = 0;
@ -793,7 +793,7 @@ rv_notOK:
}
}
else {
rvb = sum(rvb, ab);
rvb = __gdtoa_sum(rvb, ab);
k = rvb->wds - 1;
if (k >= rvb0->wds
|| hi0bits(rvb->x[k]) < hi0bits(rvb0->x[k])) {
@ -802,15 +802,15 @@ rv_notOK:
denorm = 0;
}
else {
rshift(rvb, 1);
__gdtoa_rshift(rvb, 1);
rve++;
rve1++;
L = 0;
}
}
}
Bfree(ab);
Bfree(rvb0);
__gdtoa_Bfree(ab);
__gdtoa_Bfree(rvb0);
if (finished)
break;
z = rve + rvbits;
@ -829,28 +829,28 @@ rv_notOK:
break;
}
}
bb0 = denorm ? 0 : trailz(rvb);
Bfree(bb);
Bfree(bd);
Bfree(bs);
Bfree(delta);
bb0 = denorm ? 0 : __gdtoa_trailz(rvb);
__gdtoa_Bfree(bb);
__gdtoa_Bfree(bd);
__gdtoa_Bfree(bs);
__gdtoa_Bfree(delta);
}
if (!denorm && (j = nbits - rvbits)) {
if (j > 0)
rvb = lshift(rvb, j);
rvb = __gdtoa_lshift(rvb, j);
else
rshift(rvb, -j);
__gdtoa_rshift(rvb, -j);
rve -= j;
}
*exp = rve;
Bfree(bb);
Bfree(bd);
Bfree(bs);
Bfree(bd0);
Bfree(delta);
__gdtoa_Bfree(bb);
__gdtoa_Bfree(bd);
__gdtoa_Bfree(bs);
__gdtoa_Bfree(bd0);
__gdtoa_Bfree(delta);
if (rve > fpi->emax) {
huge:
Bfree(rvb);
__gdtoa_Bfree(rvb);
rvb = 0;
errno = ERANGE;
switch(fpi->rounding & 3) {
@ -903,8 +903,8 @@ ret:
if (sign)
irv |= STRTOG_Neg;
if (rvb) {
copybits(bits, nbits, rvb);
Bfree(rvb);
__gdtoa_copybits(bits, nbits, rvb);
__gdtoa_Bfree(rvb);
}
return irv;
}