https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64427
--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Marc Glisse from comment #2)
> Restricting the testcase to x86_64 would be fine with me.
>
> As far as main is concerned, f could have written in n and d. The
> uninitialized values are only passed around, it doesn't seem relevant.
There are
mpz_t n, d;
long nn, dn;
unsigned long *np, *dup, *dnp, *qp;
long alloc, itch;
f (n);
f (d);
qp = (unsigned long*)__builtin_alloca(4099*8) + 1;
dnp = (unsigned long*)__builtin_alloca (2049*8);
alloc = 1;
for (test = 0; test < 1; test++)
{
dn = d->_mp_size;
^^^^^^^^^^^^^^^^^^
dup = d->_mp_d;
f (dnp, dup, dn);
dnp[dn - 1] |= 1UL<<63;
^^^^^^^^^^
"dn" here contains some random value, which may be > size of dnp and
program may crash which happens with -m32 for me.
