Hi!
The test is not endianess clean, x[0] is supposed to be ((__int128)0x19)<<32
on little endian - 0x19 is in the second vector elt - but ((__int128)0x19)<<64
on big endian. I've added also verification of int and __int128 sizes just
in case we have say 16-bit or 64-bit int target with __int128 type, or
pdp endian gets __int128 support.
Tested on x86_64-linux and s390x-linux, committed to trunk and 15.2 as
obvious.
2025-07-01 Jakub Jelinek <ja...@redhat.com>
PR ipa/119318
PR testsuite/120082
* gcc.dg/ipa/pr119318.c (main): Expect different result on big endian
from little endian, on unexpected endianness or int/int128 sizes don't
test anything. Formatting fixes.
--- gcc/testsuite/gcc.dg/ipa/pr119318.c.jj 2025-04-15 12:26:06.484323860
+0200
+++ gcc/testsuite/gcc.dg/ipa/pr119318.c 2025-07-01 11:54:40.941287695 +0200
@@ -30,8 +30,14 @@ int
main ()
{
W x = foo (0, (V) { 0, 5 });
- for (unsigned i = 0; i < sizeof(x)/sizeof(x[0]); i++)
+ for (unsigned i = 0; i < sizeof (x) / sizeof (x[0]); i++)
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ && __SIZEOF_INT__ == 4 &&
__SIZEOF_INT128__ == 16
if (x[i] != (i ? 0 : 0x1900000000))
- __builtin_abort();
+#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ && __SIZEOF_INT__ == 4 &&
__SIZEOF_INT128__ == 16
+ if (x[i] != (i ? 0 : ((__int128) 0x19) << 64))
+#else
+ if (0)
+#endif
+ __builtin_abort ();
return 0;
}
Jakub
