https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80808
Bug ID: 80808
Summary: [7/8 Regression] gnupg miscompilation on arm starting
with r241660
Product: gcc
Version: 7.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: jakub at gcc dot gnu.org
Target Milestone: ---
The following testcase distilled from gnupg mpih-div.c is miscompiled on arm
with -O2 -march=armv7-a -mfpu=vfpv3-d16 -mfloat-abi=hard options:
typedef unsigned int mpi_limb_t;
typedef mpi_limb_t *mpi_ptr_t;
typedef int mpi_size_t;
typedef mpi_limb_t UWtype;
typedef unsigned int USItype __attribute__ ((mode (SI)));
static __attribute__ ((noinline, noclone)) mpi_limb_t
mpihelp_add_n (mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_ptr_t s2_ptr, mpi_size_t size)
{
mpi_limb_t x, y, cy;
mpi_size_t j;
asm volatile ("":::"memory");
j = -size;
s1_ptr -= j;
s2_ptr -= j;
res_ptr -= j;
cy = 0;
do
{
y = s2_ptr[j];
x = s1_ptr[j];
y += cy;
cy = y < cy;
y += x;
cy += y < x;
res_ptr[j] = y;
}
while (++j);
return cy;
}
static __attribute__ ((noinline, noclone)) mpi_limb_t
mpihelp_submul_1 (mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr,
mpi_size_t s1_size, mpi_limb_t s2_limb)
{
mpi_limb_t cy_limb;
mpi_size_t j;
mpi_limb_t prod_high, prod_low;
mpi_limb_t x;
asm volatile ("":::"memory");
j = -s1_size;
res_ptr -= j;
s1_ptr -= j;
cy_limb = 0;
do
{
__asm__ ("umull %r1, %r0, %r2, %r3": "=&r" ((USItype) (prod_high)), "=r"
((USItype) (prod_low)): "r" ((USItype) (s1_ptr[j])), "r" ((USItype)
(s2_limb)):"r0", "r1");
prod_low += cy_limb;
cy_limb = (prod_low < cy_limb ? 1 : 0) + prod_high;
x = res_ptr[j];
prod_low = x - prod_low;
cy_limb += prod_low > x ? 1 : 0;
res_ptr[j] = prod_low;
}
while (++j);
return cy_limb;
}
static __attribute__ ((noinline, noclone)) mpi_limb_t
mpihelp_sub_n (mpi_ptr_t res_ptr, mpi_ptr_t s1_ptr, mpi_ptr_t s2_ptr,
mpi_size_t size)
{
mpi_limb_t x, y, cy;
mpi_size_t j;
asm volatile ("":::"memory");
j = -size;
s1_ptr -= j;
s2_ptr -= j;
res_ptr -= j;
cy = 0;
do
{
y = s2_ptr[j];
x = s1_ptr[j];
y += cy;
cy = y < cy;
y = x - y;
cy += y > x;
res_ptr[j] = y;
}
while (++j);
return cy;
}
static __attribute__ ((noinline, noclone)) int
mpihelp_cmp (mpi_ptr_t op1_ptr, mpi_ptr_t op2_ptr, mpi_size_t size)
{
mpi_size_t i;
mpi_limb_t op1_word, op2_word;
asm volatile ("":::"memory");
for (i = size - 1; i >= 0; i--)
{
op1_word = op1_ptr[i];
op2_word = op2_ptr[i];
if (op1_word != op2_word)
goto diff;
}
return 0;
diff:
return (op1_word > op2_word) ? 1 : -1;
}
static __attribute__ ((noinline, noclone)) mpi_limb_t
mpihelp_divrem (mpi_ptr_t qp, mpi_size_t qextra_limbs,
mpi_ptr_t np, mpi_size_t nsize,
mpi_ptr_t dp, mpi_size_t dsize)
{
mpi_limb_t most_significant_q_limb = 0;
switch (dsize)
{
case 0:
return 1 / dsize;
case 1:
{
mpi_size_t i;
mpi_limb_t n1;
mpi_limb_t d;
d = dp[0];
n1 = np[nsize - 1];
if (n1 >= d)
{
n1 -= d;
most_significant_q_limb = 1;
}
qp += qextra_limbs;
for (i = nsize - 2; i >= 0; i--)
do
{
UWtype __d1, __d0, __q1, __q0, __r1, __r0, __m;
__d1 = ((UWtype) (d) >> ((8 * (4)) / 2));
__d0 = ((UWtype) (d) & (((UWtype) 1 << ((8 * (4)) / 2)) - 1));
__r1 = (n1) % __d1;
__q1 = (n1) / __d1;
__m = (UWtype) __q1 *__d0;
__r1 =
__r1 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (np[i]) >>
((8 * (4)) / 2));
if (__r1 < __m)
{
__q1--, __r1 += (d);
if (__r1 >= (d))
if (__r1 < __m)
__q1--, __r1 += (d);
}
__r1 -= __m;
__r0 = __r1 % __d1;
__q0 = __r1 / __d1;
__m = (UWtype) __q0 *__d0;
__r0 =
__r0 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (np[i]) &
(((UWtype) 1 <<
((8 * (4)) / 2)) - 1));
if (__r0 < __m)
{
__q0--, __r0 += (d);
if (__r0 >= (d))
if (__r0 < __m)
__q0--, __r0 += (d);
}
__r0 -= __m;
(qp[i]) = (UWtype) __q1 *((UWtype) 1 << ((8 * (4)) / 2)) | __q0;
(n1) = __r0;
}
while (0);
qp -= qextra_limbs;
for (i = qextra_limbs - 1; i >= 0; i--)
do
{
UWtype __d1, __d0, __q1, __q0, __r1, __r0, __m;
__d1 = ((UWtype) (d) >> ((8 * (4)) / 2));
__d0 = ((UWtype) (d) & (((UWtype) 1 << ((8 * (4)) / 2)) - 1));
__r1 = (n1) % __d1;
__q1 = (n1) / __d1;
__m = (UWtype) __q1 *__d0;
__r1 =
__r1 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (0) >>
((8 * (4)) / 2));
if (__r1 < __m)
{
__q1--, __r1 += (d);
if (__r1 >= (d))
if (__r1 < __m)
__q1--, __r1 += (d);
}
__r1 -= __m;
__r0 = __r1 % __d1;
__q0 = __r1 / __d1;
__m = (UWtype) __q0 *__d0;
__r0 =
__r0 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (0) &
(((UWtype) 1 <<
((8 * (4)) / 2)) - 1));
if (__r0 < __m)
{
__q0--, __r0 += (d);
if (__r0 >= (d))
if (__r0 < __m)
__q0--, __r0 += (d);
}
__r0 -= __m;
(qp[i]) = (UWtype) __q1 *((UWtype) 1 << ((8 * (4)) / 2)) | __q0;
(n1) = __r0;
}
while (0);
np[0] = n1;
}
break;
case 2:
{
mpi_size_t i;
mpi_limb_t n1, n0, n2;
mpi_limb_t d1, d0;
np += nsize - 2;
d1 = dp[1];
d0 = dp[0];
n1 = np[1];
n0 = np[0];
if (n1 >= d1 && (n1 > d1 || n0 >= d0))
{
__asm__ ("subs %1, %4, %5\n" "sbc %0, %2, %3": "=r" ((USItype)
(n1)), "=&r" ((USItype) (n0)):"r" ((USItype) (n1)), "rI" ((USItype) (d1)), "r"
((USItype) (n0)), "rI" ((USItype) (d0)));
most_significant_q_limb = 1;
}
for (i = qextra_limbs + nsize - 2 - 1; i >= 0; i--)
{
mpi_limb_t q;
mpi_limb_t r;
if (i >= qextra_limbs)
np--;
else
np[0] = 0;
if (n1 == d1)
{
q = ~(mpi_limb_t) 0;
r = n0 + d1;
if (r < d1)
{
__asm__ ("adds %1, %4, %5\n" "adc %0, %2, %3": "=r"
((USItype) (n1)), "=&r" ((USItype) (n0)):"%r" ((USItype) (r - d0)), "rI"
((USItype) (0)), "%r" ((USItype) (np[0])), "rI" ((USItype) (d0)));
qp[i] = q;
continue;
}
n1 = d0 - (d0 != 0 ? 1 : 0);
n0 = -d0;
}
else
{
do
{
UWtype __d1, __d0, __q1, __q0, __r1, __r0, __m;
__d1 = ((UWtype) (d1) >> ((8 * (4)) / 2));
__d0 =
((UWtype) (d1) & (((UWtype) 1 << ((8 * (4)) / 2)) - 1));
__r1 = (n1) % __d1;
__q1 = (n1) / __d1;
__m = (UWtype) __q1 *__d0;
__r1 =
__r1 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (n0) >>
((8 * (4)) / 2));
if (__r1 < __m)
{
__q1--, __r1 += (d1);
if (__r1 >= (d1))
if (__r1 < __m)
__q1--, __r1 += (d1);
}
__r1 -= __m;
__r0 = __r1 % __d1;
__q0 = __r1 / __d1;
__m = (UWtype) __q0 *__d0;
__r0 =
__r0 *
((UWtype) 1 << ((8 * (4)) / 2)) | ((UWtype) (n0) &
(((UWtype) 1 <<
((8 * (4)) / 2)) -
1));
if (__r0 < __m)
{
__q0--, __r0 += (d1);
if (__r0 >= (d1))
if (__r0 < __m)
__q0--, __r0 += (d1);
}
__r0 -= __m;
(q) =
(UWtype) __q1 *((UWtype) 1 << ((8 * (4)) / 2)) | __q0;
(r) = __r0;
}
while (0);
__asm__ ("umull %r1, %r0, %r2, %r3": "=&r" ((USItype) (n1)),
"=r" ((USItype) (n0)): "r" ((USItype) (d0)), "r" ((USItype) (q)):"r0", "r1");
}
n2 = np[0];
q_test:
if (n1 > r || (n1 == r && n0 > n2))
{
q--;
__asm__ ("subs %1, %4, %5\n" "sbc %0, %2, %3": "=r" ((USItype)
(n1)), "=&r" ((USItype) (n0)):"r" ((USItype) (n1)), "rI" ((USItype) (0)), "r"
((USItype) (n0)), "rI" ((USItype) (d0)));
r += d1;
if (r >= d1)
goto q_test;
}
qp[i] = q;
__asm__ ("subs %1, %4, %5\n" "sbc %0, %2, %3": "=r" ((USItype)
(n1)), "=&r" ((USItype) (n0)):"r" ((USItype) (r)), "rI" ((USItype) (n1)), "r"
((USItype) (n2)), "rI" ((USItype) (n0)));
}
np[1] = n1;
np[0] = n0;
}
break;
default:
{
mpi_size_t i;
mpi_limb_t dX, d1, n0;
np += nsize - dsize;
dX = dp[dsize - 1];
d1 = dp[dsize - 2];
n0 = np[dsize - 1];
if (n0 >= dX)
{
if (n0 > dX || mpihelp_cmp (np, dp, dsize - 1) >= 0)
{
mpihelp_sub_n (np, np, dp, dsize);
n0 = np[dsize - 1];
most_significant_q_limb = 1;
}
}
for (i = qextra_limbs + nsize - dsize - 1; i >= 0; i--)
{
mpi_limb_t q;
mpi_limb_t n1, n2;
mpi_limb_t cy_limb;
if (i >= qextra_limbs)
{
np--;
n2 = np[dsize];
}
else
{
n2 = np[dsize - 1];
do
{
mpi_size_t _i;
for (_i = (dsize - 1) - 1; _i >= 0; _i--)
(np + 1)[_i] = (np)[_i];
}
while (0);
np[0] = 0;
}
if (n0 == dX)
{
q = ~(mpi_limb_t) 0;
}
else
{
mpi_limb_t r;
do
{
UWtype __d1, __d0, __q1, __q0, __r1, __r0, __m;
__d1 = ((UWtype) (dX) >> ((8 * (4)) / 2));
__d0 =
((UWtype) (dX) & (((UWtype) 1 << ((8 * (4)) / 2)) - 1));
__r1 = (n0) % __d1;
__q1 = (n0) / __d1;
__m = (UWtype) __q1 *__d0;
__r1 =
__r1 *
((UWtype) 1 << ((8 * (4)) / 2)) |
((UWtype) (np[dsize - 1]) >> ((8 * (4)) / 2));
if (__r1 < __m)
{
__q1--, __r1 += (dX);
if (__r1 >= (dX))
if (__r1 < __m)
__q1--, __r1 += (dX);
}
__r1 -= __m;
__r0 = __r1 % __d1;
__q0 = __r1 / __d1;
__m = (UWtype) __q0 *__d0;
__r0 =
__r0 *
((UWtype) 1 << ((8 * (4)) / 2)) |
((UWtype) (np[dsize - 1]) &
(((UWtype) 1 << ((8 * (4)) / 2)) - 1));
if (__r0 < __m)
{
__q0--, __r0 += (dX);
if (__r0 >= (dX))
if (__r0 < __m)
__q0--, __r0 += (dX);
}
__r0 -= __m;
(q) =
(UWtype) __q1 *((UWtype) 1 << ((8 * (4)) / 2)) | __q0;
(r) = __r0;
}
while (0);
__asm__ ("umull %r1, %r0, %r2, %r3": "=&r" ((USItype) (n1)),
"=r" ((USItype) (n0)): "r" ((USItype) (d1)), "r" ((USItype) (q)):"r0", "r1");
while (n1 > r || (n1 == r && n0 > np[dsize - 2]))
{
q--;
r += dX;
if (r < dX)
break;
n1 -= n0 < d1;
n0 -= d1;
}
}
cy_limb = mpihelp_submul_1 (np, dp, dsize, q);
if (n2 != cy_limb)
{
mpihelp_add_n (np, np, dp, dsize);
q--;
}
qp[i] = q;
n0 = np[dsize - 1];
}
}
}
return most_significant_q_limb;
}
int
main ()
{
mpi_limb_t qp[1];
mpi_limb_t np[3] = { 0xdaafeaa6, 0x0e77816a, 1 };
mpi_limb_t dp[2] = { 0x6816ec64, 0xb9d5666d };
volatile int l = 0;
mpi_limb_t ret =
mpihelp_divrem (qp + l, 0 + l, np + l, 3 + l, dp + l, 2 + l);
if (ret != 0 || qp[0] != 1 || np[0] != 0x7298fe42 || np[1] != 0x54a21afd)
__builtin_abort ();
return 0;
}
