On Thu, Dec 18, 2014 at 12:52 PM, Vladimir Makarov <vmaka...@redhat.com> wrote:
> The following patch solves
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64291
>
> It is a bug in a new rematerialization subpass of LRA.
>
> The patch was bootstrapped on x86/x86-64.
>
> Committed as rev. 218874.
>
> 2014-12-18 Vladimir Makarov <vmaka...@redhat.com>
>
> PR rtl-optimization/64291
> * lra-remat.c (bad_for_rematerialization_p): Add UNPSEC_VLOATILE.
> (create_cands): Process only output reload insn with potential
> cands.
>
> 2014-12-18 Vladimir Makarov <vmaka...@redhat.com>
>
> PR rtl-optimization/64291
> * testsuite/gcc.target/i386/pr64291-[12].c: New tests.
>
There are a couple problems with the testcase:
1. It has
typedef struct
{
int _mp_size;
unsigned long *_mp_d;
} __mpz_struct;
typedef __mpz_struct mpz_t[1];
int main ()
{
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;
For 32-bit targets, like -m32 and -mx32, GCC complains:
/export/gnu/import/git/gcc/gcc/testsuite/gcc.target/i386/pr64291-1.c:33:25:
warning: left shift count >= width of type [-Wshift-count-overflow]
2. This program uses uninitialized stack variable n, d:
void f(void*p,...){}
The behavior of this testcase is totally undefined.
--
H.J.
