From: Richard Palethorpe <rpalethorpe@suse.de>
To: Andrea Cervesato <andrea.cervesato@suse.com>
Cc: ltp@lists.linux.it
Subject: Re: [LTP] [PATCH v4 4/6] Refactor mqns_04 using new LTP API
Date: Tue, 28 Feb 2023 15:11:22 +0000 [thread overview]
Message-ID: <87356pu85b.fsf@suse.de> (raw)
In-Reply-To: <20230216085107.7068-5-andrea.cervesato@suse.com>
Hello,
Andrea Cervesato via ltp <ltp@lists.linux.it> writes:
> Signed-off-by: Andrea Cervesato <andrea.cervesato@suse.com>
> ---
> runtest/containers | 4 +-
> testcases/kernel/containers/mqns/mqns_04.c | 273 +++++++++------------
> 2 files changed, 122 insertions(+), 155 deletions(-)
>
> diff --git a/runtest/containers b/runtest/containers
> index 1cc6c838d..318113e19 100644
> --- a/runtest/containers
> +++ b/runtest/containers
> @@ -24,8 +24,8 @@ mqns_02_unshare mqns_02 -m unshare
> mqns_03 mqns_03
> mqns_03_clone mqns_03 -m clone
> mqns_03_unshare mqns_03 -m unshare
> -mqns_04 mqns_04
> -mqns_04_clone mqns_04 -clone
> +mqns_04_clone mqns_04 -m clone
> +mqns_04_unshare mqns_04 -m unshare
>
> netns_netlink netns_netlink
> netns_breakns_ip_ipv4_netlink netns_breakns.sh
> diff --git a/testcases/kernel/containers/mqns/mqns_04.c b/testcases/kernel/containers/mqns/mqns_04.c
> index d07a85c04..93df61ec4 100644
> --- a/testcases/kernel/containers/mqns/mqns_04.c
> +++ b/testcases/kernel/containers/mqns/mqns_04.c
> @@ -1,187 +1,154 @@
> +// SPDX-License-Identifier: GPL-2.0
> /*
> -* Copyright (c) International Business Machines Corp., 2009
> -* This program is free software; you can redistribute it and/or modify
> -* it under the terms of the GNU General Public License as published by
> -* the Free Software Foundation; either version 2 of the License, or
> -* (at your option) any later version.
> -*
> -* This program 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 General Public License for more details.
> -* You should have received a copy of the GNU General Public License
> -* along with this program; if not, write to the Free Software
> -* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
> -*
> -* Author: Serge Hallyn <serue@us.ibm.com>
> -*
> -* Check mqueuefs lifetime
> -* . parent creates /dev/mqueue2
> -* . child mounts mqueue there
> -* . child does mq_open("/ab")
> -* . parent checks for /dev/mqueue2
> -* . child exits
> -* . parent checks for /dev/mqueue2
> -* . parent tries 'touch /dev/mqueue2/dd' -> should fail
> -* . parent umounts /dev/mqueue2
> -
> -***************************************************************************/
> -
> -#ifndef _GNU_SOURCE
> + * Copyright (c) International Business Machines Corp., 2009
> + * Copyright (c) Serge Hallyn <serue@us.ibm.com>
> + * Copyright (C) 2023 SUSE LLC Andrea Cervesato <andrea.cervesato@suse.com>
> + */
> +
> +/*\
> + * [Description]
> + *
> + * Test mqueuefs manipulation from child/parent namespaces.
> + *
> + * [Algorithm]
> + *
> + * - parent creates mqueue folder in <tmpdir>
> + * - child mounts mqueue there
> + * - child creates /MQ1 mqueue
> + * - parent checks for <tmpdir>/mqueue/MQ1 existence
> + * - child exits
> + * - parent checks for <tmpdir>/mqueue/MQ1 existence
> + * - parent tries 'touch <tmpdir>/mqueue/MQ2' -> should fail
> + * - parent umount mqueuefs
> + */
> #define _GNU_SOURCE
> -#endif
> -#include <sys/types.h>
> -#include <sys/stat.h>
> +
> #include <sys/wait.h>
> -#include <assert.h>
> -#include <stdio.h>
> -#include <stdlib.h>
> -#include <unistd.h>
> -#include <string.h>
> -#include <errno.h>
> -#include "mqns.h"
> -#include "mqns_helper.h"
> +#include "tst_test.h"
> +#include "lapi/sched.h"
> +#include "tst_safe_posix_ipc.h"
> +#include "tst_safe_stdio.h"
>
> -char *TCID = "posixmq_namespace_04";
> -int TST_TOTAL = 1;
> +#define CHECK_MQ_OPEN_RET(x) ((x) >= 0 || ((x) == -1 && errno != EMFILE))
>
> -int p1[2];
> -int p2[2];
> +#define MQNAME1 "/MQ1"
> +#define MQNAME2 "/MQ2"
/MQ2 is never created
>
> -#define FNAM1 DEV_MQUEUE2 SLASH_MQ1
> -#define FNAM2 DEV_MQUEUE2 SLASH_MQ2
> +static char *str_op;
> +static char *devdir;
> +static char *mqueue1;
> +static char *mqueue2;
> +static int *mq_freed;
>
> -int check_mqueue(void *vtest)
> +static void check_mqueue(void)
> {
> - char buf[30];
> mqd_t mqd;
> - int rc;
>
> - (void) vtest;
> + mqd = TST_RETRY_FUNC(
> + mq_open(MQNAME1, O_RDWR | O_CREAT | O_EXCL, 0755, NULL),
> + CHECK_MQ_OPEN_RET);
> + if (mqd == -1)
> + tst_brk(TBROK | TERRNO, "mq_open failed");
>
> - close(p1[1]);
> - close(p2[0]);
> + SAFE_MQ_CLOSE(mqd);
>
> - read(p1[0], buf, 3); /* go */
> + SAFE_MOUNT("mqueue", devdir, "mqueue", 0, NULL);
>
> - mqd = tst_syscall(__NR_mq_open, NOSLASH_MQ1, O_RDWR | O_CREAT | O_EXCL,
> - 0755, NULL);
> - if (mqd == -1) {
> - write(p2[1], "mqfail", 7);
> - tst_exit();
> - }
> + TST_CHECKPOINT_WAKE_AND_WAIT(0);
> +}
> +
> +static void run(void)
> +{
> + const struct tst_clone_args clone_args = { CLONE_NEWIPC, SIGCHLD };
> + int status;
> + struct stat statbuf;
> +
> + tst_res(TINFO, "Checking mqueue filesystem lifetime");
> +
> + if (str_op && !strcmp(str_op, "clone")) {
> + tst_res(TINFO, "Spawning isolated process");
>
> - mq_close(mqd);
> + if (!SAFE_CLONE(&clone_args)) {
> + check_mqueue();
> + return;
> + }
> + } else if (str_op && !strcmp(str_op, "unshare")) {
> + tst_res(TINFO, "Spawning unshared process");
>
> - rc = mount("mqueue", DEV_MQUEUE2, "mqueue", 0, NULL);
Possibly this is a mistake in the original test. Do we really need to
remount mqueuefs? Isn't it just the same FS in a different location?
This thing that really makes a difference is that the child is in a new
namespace, so the parent and child should not see each others queues?
> - if (rc == -1) {
> - perror("mount");
> - write(p2[1], "mount", 6);
> - tst_exit();
> + if (!SAFE_FORK()) {
> + SAFE_UNSHARE(CLONE_NEWIPC);
> + check_mqueue();
> + return;
> + }
> }
>
> - write(p2[1], "go", 3);
> - read(p1[0], buf, 3);
> + TST_CHECKPOINT_WAIT(0);
>
> - tst_exit();
> -}
> + SAFE_STAT(mqueue1, &statbuf);
> + tst_res(TPASS, "%s child's mqueue can be accessed from parent", mqueue1);
>
> -static void setup(void)
> -{
> - tst_require_root();
> - check_mqns();
> -}
> + TST_CHECKPOINT_WAKE(0);
>
> -int main(int argc, char *argv[])
> -{
> - int rc;
> - int status;
> - char buf[30];
> - struct stat statbuf;
> - int use_clone = T_UNSHARE;
> + tst_res(TINFO, "Waiting child to exit");
>
> - setup();
> + SAFE_WAIT(&status);
Again we can just use tst_reap_children() which does a thorough check of
the child status.
>
> - if (argc == 2 && strcmp(argv[1], "-clone") == 0) {
> - tst_resm(TINFO,
> - "Testing posix mq namespaces through clone(2).");
> - use_clone = T_CLONE;
> - } else
> - tst_resm(TINFO,
> - "Testing posix mq namespaces through unshare(2).");
> + if (!WIFEXITED(status))
> + tst_brk(TBROK, "Child did not exit normally: %s", tst_strstatus(status));
>
> - if (pipe(p1) == -1) {
> - perror("pipe");
> - exit(EXIT_FAILURE);
> - }
> - if (pipe(p2) == -1) {
> - perror("pipe");
> - exit(EXIT_FAILURE);
> - }
> + SAFE_STAT(mqueue1, &statbuf);
> + tst_res(TPASS, "%s child's mqueue can be accessed from parent after child's dead", mqueue1);
>
> - mkdir(DEV_MQUEUE2, 0755);
> + TST_EXP_FAIL(creat(mqueue2, 0755), EACCES);
>
> - tst_resm(TINFO, "Checking mqueue filesystem lifetime");
> + SAFE_UMOUNT(devdir);
>
> - /* fire off the test */
> - rc = do_clone_unshare_test(use_clone, CLONE_NEWIPC, check_mqueue, NULL);
> - if (rc < 0) {
> - tst_resm(TFAIL, "failed clone/unshare");
> - goto fail;
> - }
> + *mq_freed = 1;
This doesn't appear to be near the point that the MQ is unlinked.
> +}
>
> - close(p1[0]);
> - close(p2[1]);
> - write(p1[1], "go", 3);
> -
> - read(p2[0], buf, 7);
> - if (!strcmp(buf, "mqfail")) {
> - tst_resm(TFAIL, "child process could not create mqueue");
> - goto fail;
> - } else if (!strcmp(buf, "mount")) {
> - tst_resm(TFAIL, "child process could not mount mqueue");
> - goto fail;
> - }
> +static void setup(void)
> +{
> + char *tmpdir;
>
> - rc = stat(FNAM1, &statbuf);
> - if (rc == -1) {
> - perror("stat");
> - write(p1[1], "go", 3);
> - tst_resm(TFAIL, "parent could not see child's created mq");
> - goto fail;
> - }
> - write(p1[1], "go", 3);
> + if (!str_op)
> + tst_brk(TCONF, "Please use clone|unshare execution
> mode");
When this happens the test segfaults because devdir and mq_freed have
not been allocated.
This is also a problem in previous tests as well.
>
> - rc = wait(&status);
> - if (rc == -1) {
> - perror("wait");
> - tst_resm(TFAIL, "error while parent waited on child to exit");
> - goto fail;
> - }
> - if (!WIFEXITED(status)) {
> - tst_resm(TFAIL, "Child did not exit normally (status %d)",
> - status);
> - goto fail;
> - }
> - rc = stat(FNAM1, &statbuf);
> - if (rc == -1) {
> - tst_resm(TFAIL,
> - "parent's view of child's mq died with child");
> - goto fail;
> - }
> + tmpdir = tst_get_tmpdir();
>
> - rc = creat(FNAM2, 0755);
> - if (rc != -1) {
> - tst_resm(TFAIL,
> - "parent was able to create a file in dead child's mqfs");
> - goto fail;
> - }
> + SAFE_ASPRINTF(&devdir, "%s/mqueue", tmpdir);
> + SAFE_MKDIR(devdir, 0755);
> +
> + SAFE_ASPRINTF(&mqueue1, "%s" MQNAME1, devdir);
> + SAFE_ASPRINTF(&mqueue2, "%s" MQNAME2, devdir);
>
> - tst_resm(TPASS, "Child mqueue fs still visible for parent");
> + mq_freed = SAFE_MMAP(NULL,
> + sizeof(int),
> + PROT_READ | PROT_WRITE,
> + MAP_SHARED | MAP_ANONYMOUS,
> + -1, 0);
This starts as 0, but we have not created any MQs yet so we could not
free them in cleanup (also in mqns03).
Try running with -i0
> +}
>
> -fail:
> - umount(DEV_MQUEUE2);
> - rmdir(DEV_MQUEUE2);
> +static void cleanup(void)
> +{
> + if (tst_is_mounted(devdir))
> + SAFE_UMOUNT(devdir);
>
> - tst_exit();
> + if (!*mq_freed) {
> + SAFE_MQ_UNLINK(MQNAME1);
> + SAFE_MQ_UNLINK(MQNAME2);
> + }
> }
> +
> +static struct tst_test test = {
> + .test_all = run,
> + .setup = setup,
> + .cleanup = cleanup,
> + .needs_root = 1,
> + .forks_child = 1,
> + .needs_tmpdir = 1,
> + .needs_checkpoints = 1,
> + .options = (struct tst_option[]) {
> + { "m:", &str_op, "Test execution mode <clone|unshare>" },
> + {},
> + },
> +};
> --
> 2.35.3
--
Thank you,
Richard.
--
Mailing list info: https://lists.linux.it/listinfo/ltp
next prev parent reply other threads:[~2023-02-28 15:22 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-16 8:51 [LTP] [PATCH v4 0/6] Refactor mqns testing suite Andrea Cervesato via ltp
2023-02-16 8:51 ` [LTP] [PATCH v4 1/6] Refactor mqns_01 using new LTP API Andrea Cervesato via ltp
2023-02-16 8:51 ` [LTP] [PATCH v4 2/6] Refactor mqns_02 " Andrea Cervesato via ltp
2023-02-16 8:51 ` [LTP] [PATCH v4 3/6] Refactor mqns_03 " Andrea Cervesato via ltp
2023-02-28 15:06 ` Richard Palethorpe
2023-02-16 8:51 ` [LTP] [PATCH v4 4/6] Refactor mqns_04 " Andrea Cervesato via ltp
2023-02-28 15:11 ` Richard Palethorpe [this message]
2023-03-01 14:03 ` Andrea Cervesato via ltp
2023-02-16 8:51 ` [LTP] [PATCH v4 5/6] Remove deprecated header files from mqns suite Andrea Cervesato via ltp
2023-02-16 8:51 ` [LTP] [PATCH v4 6/6] Remove libclone dependency " Andrea Cervesato via ltp
2023-02-28 15:22 ` [LTP] [PATCH v4 0/6] Refactor mqns testing suite Richard Palethorpe
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=87356pu85b.fsf@suse.de \
--to=rpalethorpe@suse.de \
--cc=andrea.cervesato@suse.com \
--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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.