[LTP] [RFC PATCH 1/1] readahead01: Pass also on memfd secret

[Li Wang]

Hi Petr, All,

I'm not sure but one point I guess that should be similar to 'TST_FD_MEMFD'
since the file is RAM-based so readahead() is not needed.

Your patch looks good, however, I was unable to get the clue in kernel code
to prove that.


On Fri, Jan 19, 2024 at 3:14 AM Petr Vorel <pvorel@suse.cz> wrote:

> It fixes failure on 6.6 kernels:
>
> TFAIL: readahead() on memfd secret succeeded
>
> Fixes: ecf81d729 ("syscalls: readahead01: Make use of tst_fd")
> Reported-by: Dominique Leuenberger <dimstar@opensuse.org>
> Signed-off-by: Petr Vorel <pvorel@suse.cz>
> ---
> Hi,
>
> I wonder if this is a bug or just changed functionality.
>
> I also tested on 5.14 SLES and 5.10 Debian and both TCONF due ENOSYS:
>
> tst_fd.c:260: TCONF: Skipping memfd secret: ENOSYS (38)
>
> Kind regards,
> Petr
>
>  testcases/kernel/syscalls/readahead/readahead01.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/testcases/kernel/syscalls/readahead/readahead01.c
> b/testcases/kernel/syscalls/readahead/readahead01.c
> index e86a73e3e..a1f313605 100644
> --- a/testcases/kernel/syscalls/readahead/readahead01.c
> +++ b/testcases/kernel/syscalls/readahead/readahead01.c
> @@ -51,6 +51,7 @@ static void test_invalid_fd(struct tst_fd *fd)
>         /* These succeed */
>         case TST_FD_FILE:
>         case TST_FD_MEMFD:
> +       case TST_FD_MEMFD_SECRET:
>         case TST_FD_PROC_MAPS:
>                 return;
>         default:
> --
> 2.43.0
>
>

-- 
Regards,
Li Wang