Re: [PATCH] selftests/pidfd: print time_t as long long
From: David Laight
Date: Wed Nov 18 2026 - 08:22:59 EST
On Tue, 3 Nov 2026 10:54:47 -0500
Sasha Levin <sashal@xxxxxxxxxx> wrote:
> On 32-bit architectures with a 64-bit time_t, such as armhf in Debian
> 13, printing the time_t values with %lu warns:
>
> pidfd_test.c:523:50: warning: format '%lu' expects argument of type 'long unsigned int', but argument 2 has type 'time_t' {aka 'long long int'} [-Wformat=]
>
> Cast them to long long and print them with %lld.
>
> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> ---
> Found with a 32-bit build of the selftests with a 64-bit time_t (gcc 16
> and clang 21 with -m32 -D_TIME_BITS=64 -D_FILE_OFFSET_BITS=64);
> KernelCI's 32-bit arm builds show the same warnings. The changed files
> build without format warnings there now, and the x86_64 builds don't
> change. The other selftests get the same kind of fix in separate
> patches.
>
> tools/testing/selftests/pidfd/pidfd_test.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/tools/testing/selftests/pidfd/pidfd_test.c b/tools/testing/selftests/pidfd/pidfd_test.c
> index 932cbd8caa770..a53e72cb256ce 100644
> --- a/tools/testing/selftests/pidfd/pidfd_test.c
> +++ b/tools/testing/selftests/pidfd/pidfd_test.c
> @@ -520,7 +520,7 @@ static void test_pidfd_poll_exec(int use_waitpid)
>
> time_t prog_time = time(NULL) - prog_start;
Why not just change the type to 'unsigned int' and use "%u" ?
It isn't as though the difference is going to be longer than 75 years.
David
>
> - ksft_print_msg("Time waited for child: %lu\n", prog_time);
> + ksft_print_msg("Time waited for child: %lld\n", (long long)prog_time);
>
> close(pidfd);
>
> @@ -599,7 +599,8 @@ static void test_pidfd_poll_leader_exit(int use_waitpid)
>
> time_t since_child_exit = time(NULL) - *child_exit_secs;
>
> - ksft_print_msg("Time since child exit: %lu\n", since_child_exit);
> + ksft_print_msg("Time since child exit: %lld\n",
> + (long long)since_child_exit);
>
> close(pidfd);
>