https://gcc.gnu.org/g:7c317ad17a2c808f5465fbf56e3c2e7569c45a0b
commit r17-5205-g7c317ad17a2c808f5465fbf56e3c2e7569c45a0b Author: Richard Sandiford <[email protected]> Date: Wed Oct 14 19:43:51 2026 +0100 testsuite: Tweak gcc.target/aarch64/neon/discarded_result.c The testcase passed for me even with r17-4816-gb24418a2212866 reverted. That's because the test adds -fno-trapping-math, which means that vadd_f32 is treated as having no side-effects. The call can then be optimised away even without folding. The test does ICE with unpatched sources if the dg-option is removed. However, the proper default for trapping-math is controversial, so it seems better to add an explicit -ftrapping-math instead. gcc/testsuite/ * gcc.target/aarch64/neon/discarded_result.c: Use -ftrapping-math. Diff: --- gcc/testsuite/gcc.target/aarch64/neon/discarded_result.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/gcc/testsuite/gcc.target/aarch64/neon/discarded_result.c b/gcc/testsuite/gcc.target/aarch64/neon/discarded_result.c index 183b162f2ee3..1adec7fe8d3b 100644 --- a/gcc/testsuite/gcc.target/aarch64/neon/discarded_result.c +++ b/gcc/testsuite/gcc.target/aarch64/neon/discarded_result.c @@ -1,6 +1,6 @@ /* { dg-do compile } */ /* { dg-final { check-function-bodies "**" "" } } */ -/* { dg-additional-options "-fno-trapping-math" } */ +/* { dg-additional-options "-ftrapping-math" } */ #include "arm_neon_test.h"
