[PATCH v2] LoongArch: fix missing trap for enabled exceptions on narrowing operation

[Florian Weimer]

* Xi Ruoyao:

> diff --git a/math/test-narrowing-trap.c b/math/test-narrowing-trap.c
> new file mode 100644
> index 0000000000..27c31245c1
> --- /dev/null
> +++ b/math/test-narrowing-trap.c
> @@ -0,0 +1,60 @@
> +/* Copyright (C) 2026 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#ifndef _GNU_SOURCE
> +# define _GNU_SOURCE
> +#endif

We automatically define _GNU_SOURCE for tests, so that's unnecessary.

> +#include <fenv.h>
> +#include <math.h>
> +#include <math-tests.h>
> +#include <signal.h>
> +#include <stdlib.h>
> +#include <support/check.h>
> +#include <support/xunistd.h>
> +
> +static int
> +do_test (void)
> +{
> +  pid_t pid;
> +
> +  if (!EXCEPTION_ENABLE_SUPPORTED (FE_INVALID))
> +    FAIL_UNSUPPORTED ("feenableexcept (FE_INVALID) not supported");
> +
> +  pid = xfork ();
> +  if (pid == 0)
> +    {
> +      int r = feenableexcept (FE_INVALID);
> +      if (r == -1)
> +	exit (2);

Tabs vs spaces inconsistency, I think.  And you could use
TEST_VERIFY (r != -1).  The test framework uses a shared memory
segment, so that subtest failures after fork are communicated
outwards even if the exit status is zero.

> +      fdiv (0.0, 0.0);
> +      exit (0);

This should be _exit (0), so that atexit handlers are not run twice.
(We have protection against this internally, but I think test
code should show how to use fork correctly if possible.)

Sorry, should have mentioned that before.

Thanks,
Florian