From: Petr Vorel <pvorel@suse.cz>
To: Andrea Cervesato <andrea.cervesato@suse.de>
Cc: ltp@lists.linux.it
Subject: Re: [LTP] [PATCH v5 4/5] Add fchmodat2_01 test
Date: Fri, 2 Aug 2024 10:47:27 +0200 [thread overview]
Message-ID: <20240802084727.GA1602775@pevik> (raw)
In-Reply-To: <20240802-fchmodat2-v5-4-bff2ec1a4f06@suse.com>
> From: Andrea Cervesato <andrea.cervesato@suse.com>
> This test verifies that fchmodat2() syscall is properly working with
> AT_SYMLINK_NOFOLLOW on regular files. When AT_SYMLINK_NOFOLLOW is used
> on symbolic links instead, we check for EOPNOTSUPP, since the feature is
> not implemented for VFS.
> Signed-off-by: Andrea Cervesato <andrea.cervesato@suse.com>
> ---
> runtest/syscalls | 2 +
> testcases/kernel/syscalls/fchmodat2/.gitignore | 1 +
> testcases/kernel/syscalls/fchmodat2/Makefile | 7 ++
> testcases/kernel/syscalls/fchmodat2/fchmodat2_01.c | 123 +++++++++++++++++++++
> 4 files changed, 133 insertions(+)
> diff --git a/runtest/syscalls b/runtest/syscalls
> index 9b3cba667..50298d01e 100644
> --- a/runtest/syscalls
> +++ b/runtest/syscalls
> @@ -270,6 +270,8 @@ fchmod06 fchmod06
> fchmodat01 fchmodat01
> fchmodat02 fchmodat02
> +fchmodat2_01 fchmodat2_01
> +
> fchown01 fchown01
> fchown01_16 fchown01_16
> fchown02 fchown02
> diff --git a/testcases/kernel/syscalls/fchmodat2/.gitignore b/testcases/kernel/syscalls/fchmodat2/.gitignore
> new file mode 100644
> index 000000000..47d5e2427
> --- /dev/null
> +++ b/testcases/kernel/syscalls/fchmodat2/.gitignore
> @@ -0,0 +1 @@
> +fchmodat2_01
> diff --git a/testcases/kernel/syscalls/fchmodat2/Makefile b/testcases/kernel/syscalls/fchmodat2/Makefile
> new file mode 100644
> index 000000000..8cf1b9024
> --- /dev/null
> +++ b/testcases/kernel/syscalls/fchmodat2/Makefile
> @@ -0,0 +1,7 @@
> +# SPDX-License-Identifier: GPL-2.0-or-later
> +# Copyright (C) 2024 SUSE LLC Andrea Cervesato <andrea.cervesato@suse.com>
> +
> +top_srcdir ?= ../../../..
> +
> +include $(top_srcdir)/include/mk/testcases.mk
> +include $(top_srcdir)/include/mk/generic_leaf_target.mk
> diff --git a/testcases/kernel/syscalls/fchmodat2/fchmodat2_01.c b/testcases/kernel/syscalls/fchmodat2/fchmodat2_01.c
> new file mode 100644
> index 000000000..d12c4a8fd
> --- /dev/null
> +++ b/testcases/kernel/syscalls/fchmodat2/fchmodat2_01.c
> @@ -0,0 +1,123 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/*
> + * Copyright (C) 2024 SUSE LLC Andrea Cervesato <andrea.cervesato@suse.com>
> + */
> +
> +/*\
> + * [Description]
> + *
> + * This test verifies that fchmodat2() syscall is properly working with
> + * regular files, symbolic links and directories. AT_SYMLINK_NOFOLLOW is a
> + * special feature that is not enabled in VFS for symbolic links, so we only
> + * verify that EOPNOTSUPP is correctly raised when used on those particular
> + * files.
If you don't mind, I'll add here explicit note that change done in 5d1f903f75a8.
> + */
> +
> +#include "tst_test.h"
> +#include "tst_safe_file_at.h"
> +#include "lapi/fcntl.h"
> +#include "lapi/stat.h"
> +
> +#define MNTPOINT "mntpoint"
> +#define FNAME "myfile"
> +#define SNAME "symlink"
> +#define DNAME "mydir"
> +#define DNAME_PATH MNTPOINT"/"DNAME
> +
> +static int fd_dir = -1;
> +
> +static void verify_mode(int dirfd, const char *path, mode_t mode)
> +{
> + struct stat st;
> +
> + SAFE_FSTATAT(dirfd, path, &st, AT_SYMLINK_NOFOLLOW);
> + TST_EXP_EQ_LI(st.st_mode, mode);
> +}
> +
> +static void test_regular_file(void)
> +{
> + tst_res(TINFO, "Using regular files");
> +
> + SAFE_CHMOD(MNTPOINT"/"FNAME, 0640);
> +
> + SAFE_FCHMODAT2(fd_dir, FNAME, 0700, 0);
> + verify_mode(fd_dir, FNAME, S_IFREG | 0700);
> +
> + SAFE_FCHMODAT2(fd_dir, FNAME, 0700, AT_SYMLINK_NOFOLLOW);
> + verify_mode(fd_dir, FNAME, S_IFREG | 0700);
> +}
> +
> +static void test_symbolic_link(void)
> +{
> + tst_res(TINFO, "Using symbolic link");
> +
> + SAFE_FCHMODAT2(fd_dir, SNAME, 0700, 0);
> + verify_mode(fd_dir, FNAME, S_IFREG | 0700);
> + verify_mode(fd_dir, SNAME, S_IFLNK | 0777);
> +
> + if (tst_kvercmp(6, 6, 0)) {
> + TST_EXP_FAIL(tst_syscall(__NR_fchmodat2,
> + fd_dir, SNAME, 0640, AT_SYMLINK_NOFOLLOW),
> + EOPNOTSUPP);
> + }
I wrote later this is wrong, because the change was backported to all active
stable/LTS trees. We should expect this functionality on all kernels
(tested on 6.6.15, 6.9.9, 6.10.1).
Also, I hesitated, whether we should check AT_SYMLINK_NOFOLLOW also in
safe_fchmodat2(), but we can probably ignore that.
I can change it before merge.
> +}
> +
> +static void test_empty_folder(void)
> +{
> + tst_res(TINFO, "Using empty folder");
> +
> + int fd;
> +
> + SAFE_CHMOD(DNAME_PATH, 0640);
> + fd = SAFE_OPEN(DNAME_PATH, O_PATH | O_DIRECTORY, 0640);
> +
> + SAFE_FCHMODAT2(fd, "", 0777, AT_EMPTY_PATH);
> + verify_mode(fd_dir, DNAME, S_IFDIR | 0777);
> +
> + SAFE_CLOSE(fd);
> +}
> +
> +static void run(void)
> +{
> + test_regular_file();
> + test_empty_folder();
> + test_symbolic_link();
> +}
> +
> +static void setup(void)
> +{
> + fd_dir = SAFE_OPEN(MNTPOINT, O_PATH | O_DIRECTORY, 0640);
> +
> + if (access(DNAME_PATH, F_OK) == -1)
> + SAFE_MKDIR(DNAME_PATH, 0640);
> +
> + SAFE_TOUCH(MNTPOINT"/"FNAME, 0640, NULL);
> + SAFE_SYMLINKAT(FNAME, fd_dir, SNAME);
> +}
> +
> +static void cleanup(void)
> +{
> + SAFE_UNLINKAT(fd_dir, SNAME, 0);
> + SAFE_RMDIR(DNAME_PATH);
> +
> + if (fd_dir != -1)
> + SAFE_CLOSE(fd_dir);
> +}
> +
> +static struct tst_test test = {
> + .test_all = run,
> + .setup = setup,
> + .cleanup = cleanup,
> + .needs_root = 1,
> + .mntpoint = MNTPOINT,
> + .format_device = 1,
> + .all_filesystems = 1,
> + .skip_filesystems = (const char *const []) {
> + "fuse",
> + NULL
> + },
You were faster than I managed to write. IMHO skipping fuse is not needed.
IMHO it was broken due EOPNOTSUPP.
With these 2 changes:
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Kind regards,
Petr
> + .tags = (const struct tst_tag[]) {
> + {"linux-git", "5d1f903f75a8"},
> + {}
> + }
> +};
--
Mailing list info: https://lists.linux.it/listinfo/ltp
next prev parent reply other threads:[~2024-08-02 8:47 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-02 8:23 [LTP] [PATCH v5 0/5] Add fchmodat2 testing suite Andrea Cervesato
2024-08-02 8:23 ` [LTP] [PATCH v5 1/5] Add SAFE_SYMLINKAT macro Andrea Cervesato
2024-08-02 9:47 ` Petr Vorel
2024-08-02 9:48 ` Cyril Hrubis
2024-08-02 10:02 ` Petr Vorel
2024-08-02 10:39 ` Andrea Cervesato via ltp
2024-08-02 8:23 ` [LTP] [PATCH v5 2/5] Add fchmodat2 syscalls definitions Andrea Cervesato
2024-08-02 9:49 ` Petr Vorel
2024-08-02 8:23 ` [LTP] [PATCH v5 3/5] Add fchmodat2 fallback definition Andrea Cervesato
2024-08-02 9:55 ` Petr Vorel
2024-08-02 10:38 ` Andrea Cervesato via ltp
2024-08-02 10:57 ` Petr Vorel
2024-08-02 8:23 ` [LTP] [PATCH v5 4/5] Add fchmodat2_01 test Andrea Cervesato
2024-08-02 8:47 ` Petr Vorel [this message]
2024-08-02 8:23 ` [LTP] [PATCH v5 5/5] Add fchmodat2_02 test Andrea Cervesato
2024-08-02 10:33 ` Petr Vorel
2024-08-02 9:58 ` [LTP] [PATCH v5 0/5] Add fchmodat2 testing suite Petr Vorel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240802084727.GA1602775@pevik \
--to=pvorel@suse.cz \
--cc=andrea.cervesato@suse.de \
--cc=ltp@lists.linux.it \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox