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

[caiyinyu]

在 2026/3/20 下午6:50, Florian Weimer 写道:
> * 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.

Gentle ping.


>
> Thanks,
> Florian