diff options
author | Jakub Jelinek <jakub@redhat.com> | 2020-06-27 12:38:23 +0200 |
---|---|---|
committer | Jakub Jelinek <jakub@redhat.com> | 2020-06-27 12:42:13 +0200 |
commit | 37995960984ea2222346dd9d168d332cd6f7adf0 (patch) | |
tree | a4fa692125fde0f40042723e52372c67f8911b77 /gcc/c-family | |
parent | daaed0199ee57013ae011421a7e90b7bdd295373 (diff) | |
download | gcc-37995960984ea2222346dd9d168d332cd6f7adf0.zip gcc-37995960984ea2222346dd9d168d332cd6f7adf0.tar.gz gcc-37995960984ea2222346dd9d168d332cd6f7adf0.tar.bz2 |
c-family: Use TYPE_OVERFLOW_UNDEFINED instead of !TYPE_UNSIGNED in pointer_sum [PR95903]
For lp64 targets and int off ... ptr[off + 1]
is lowered in pointer_sum to *(ptr + ((sizetype) off + (sizetype) 1)).
That is fine when signed integer wrapping is undefined (and is not done
already if off has unsigned type), but changes behavior for -fwrapv, where
overflow is well defined. Runtime test could be:
int
main ()
{
char *p = __builtin_malloc (0x100000000UL);
if (!p) return 0;
char *q = p + 0x80000000UL;
int o = __INT_MAX__;
q[o + 1] = 1;
if (q[-__INT_MAX__ - 1] != 1) __builtin_abort ();
return 0;
}
with -fwrapv or so, not included in the testsuite because it requires 4GB
allocation (with some other test it would be enough to have something
slightly above 2GB, but still...).
2020-06-27 Jakub Jelinek <jakub@redhat.com>
PR middle-end/95903
gcc/c-family/
* c-common.c (pointer_int_sum): Use TYPE_OVERFLOW_UNDEFINED instead of
!TYPE_UNSIGNED check to see if we can apply distributive law and handle
smaller precision intop operands separately.
gcc/testsuite/
* c-c++-common/pr95903.c: New test.
Diffstat (limited to 'gcc/c-family')
-rw-r--r-- | gcc/c-family/c-common.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/gcc/c-family/c-common.c b/gcc/c-family/c-common.c index b73ad2e..a9fa0d1 100644 --- a/gcc/c-family/c-common.c +++ b/gcc/c-family/c-common.c @@ -3141,7 +3141,7 @@ pointer_int_sum (location_t loc, enum tree_code resultcode, /* If the constant is unsigned, and smaller than the pointer size, then we must skip this optimization. This is because it could cause an overflow error if the constant is negative but INTOP is not. */ - && (!TYPE_UNSIGNED (TREE_TYPE (intop)) + && (TYPE_OVERFLOW_UNDEFINED (TREE_TYPE (intop)) || (TYPE_PRECISION (TREE_TYPE (intop)) == TYPE_PRECISION (TREE_TYPE (ptrop))))) { |