From: Omar Sandoval <osandov@osandov.com>
To: Sun Ke <sunke32@huawei.com>
Cc: linux-block@vger.kernel.org, osandov@fb.com
Subject: Re: [PATCH blktests v3] nbd/003:add mount and clear_sock test for nbd
Date: Thu, 19 Dec 2019 14:35:19 -0800 [thread overview]
Message-ID: <20191219223519.GB830111@vader> (raw)
In-Reply-To: <1575115540-69845-1-git-send-email-sunke32@huawei.com>
On Sat, Nov 30, 2019 at 08:05:40PM +0800, Sun Ke wrote:
> Add the test case to check nbd device. This test case catches regressions
> fixed by commit 92b5c8f0063e4 "nbd: replace kill_bdev() with
> __invalidate_device() again".
>
> Establish the nbd connection. Run two processes. The first one do mount
> and umount, and the other one do clear_sock ioctl.
>
> Signed-off-by: Sun Ke <sunke32@huawei.com>
> ---
> v2 -> v3
> 1. Now only build nbd0 connection, not 15 connections.
> 2. Add the run_cnt to 225.
> 3. Modify some variable names.
mount_clear_sock still seems overly complicated, so I simplified it.
Additionally, we can loop from the C program instead of looping in the
shell script and execing a bunch of times. And, it seems like we can use
the existing _start_nbd_server/_stop_nbd_server helpers now that we're
only using one connection. I made those changes here:
https://github.com/osandov/blktests/commit/ab2472ee1c54f7fc69011815e15515846ae40eea
Would you mind taking a look and ensuring that it still does what you're
trying to do? I seem to have triggered some other sort of deadlock in
NBD with the optimized test.
> ---
> src/Makefile | 3 ++-
> src/mount_clear_sock.c | 68 ++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/nbd/003 | 55 ++++++++++++++++++++++++++++++++++++++++
> tests/nbd/003.out | 1 +
> 4 files changed, 126 insertions(+), 1 deletion(-)
> create mode 100644 src/mount_clear_sock.c
> create mode 100644 tests/nbd/003
> create mode 100644 tests/nbd/003.out
>
> diff --git a/src/Makefile b/src/Makefile
> index 917d6f4..acd7327 100644
> --- a/src/Makefile
> +++ b/src/Makefile
> @@ -10,7 +10,8 @@ C_TARGETS := \
> sg/syzkaller1 \
> nbdsetsize \
> loop_change_fd \
> - zbdioctl
> + zbdioctl \
> + mount_clear_sock
>
> CXX_TARGETS := \
> discontiguous-io
> diff --git a/src/mount_clear_sock.c b/src/mount_clear_sock.c
> new file mode 100644
> index 0000000..c76cbfe
> --- /dev/null
> +++ b/src/mount_clear_sock.c
> @@ -0,0 +1,68 @@
> +// SPDX-License-Identifier: GPL-3.0+
> +// Copyright (C) 2019 Sun Ke
> +
> +#include <stdio.h>
> +#include <stdlib.h>
> +
> +#include <sys/types.h>
> +#include <sys/stat.h>
> +#include <fcntl.h>
> +
> +#include <linux/nbd.h>
> +#include <assert.h>
> +#include <sys/wait.h>
> +#include <unistd.h>
> +#include <string.h>
> +#include <sys/ioctl.h>
> +#include <sys/mount.h>
> +#include <linux/fs.h>
> +
> +void clear_sock(int fd)
> +{
> + int err;
> +
> + err = ioctl(fd, NBD_CLEAR_SOCK, 0);
> + if (err) {
> + perror("ioctl");
> + }
> +}
> +
> +void mount_nbd(char *dev, char *mp, char *fs)
> +{
> + mount(dev, mp, fs, MS_NOSUID | MS_SYNCHRONOUS, 0);
> + umount(mp);
> +}
> +
> +int main(int argc, char **argv)
> +{
> + if (argc != 4) {
> + fprintf(stderr, "usage: $0 MOUNTPOINT DEV FS");
> + return EXIT_FAILURE;
> + }
> +
> + char *mp = argv[1];
> + char *dev = argv[2];
> + char *fs = argv[3];
> +
> + static int fd = -1;
> +
> + fd = open(dev, O_RDWR);
> + if (fd < 0 ) {
> + perror("open");
> + }
> +
> + if (fork() == 0) {
> + mount_nbd(dev, mp, fs);
> + exit(0);
> + }
> + if (fork() == 0) {
> + clear_sock(fd);
> + exit(0);
> + }
> + while(wait(NULL) > 0)
> + continue;
> +
> + close(fd);
> +
> + return 0;
> +}
> diff --git a/tests/nbd/003 b/tests/nbd/003
> new file mode 100644
> index 0000000..738928e
> --- /dev/null
> +++ b/tests/nbd/003
> @@ -0,0 +1,55 @@
> +#!/bin/bash
> +
> +# SPDX-License-Identifier: GPL-3.0+
> +# Copyright (C) 2019 Sun Ke
> +#
> +# Test nbd device resizing. Regression test for patch
> +#
> +# 2b5c8f0063e4 ("nbd: replace kill_bdev() with __invalidate_device() again")
> +
> +
> +DESCRIPTION="resize a connected nbd device"
> +QUICK=1
> +
> +fs_type=ext4
> +disk_capacity=256M
> +run_cnt=225
> +
> +requires() {
> + _have_nbd && _have_src_program mount_clear_sock
> +}
> +
> +_start_nbd_mount_server() {
> +
> + fallocate -l $1 "${TMPDIR}/disk"
> +
> + if [[ "$2"x = "ext4"x ]]; then
> + mkfs.ext4 "${TMPDIR}/disk" >> "$FULL" 2>&1
> + else
> + mkdosfs "${TMPDIR}/disk" >> "$FULL" 2>&1
> + fi
> + nbd-server 8000 "${TMPDIR}/disk" >> "$FULL" 2>&1
> +
> + mkdir -p "${TMPDIR}/mount_point"
> +}
> +
> +_stop_nbd_mount_server() {
> + pkill -9 -f 8000
> + rm -f "${TMPDIR}/disk"
> + rm -rf "${TMPDIR}/mount_point"
> +}
> +
> +test() {
> + echo "Running ${TEST_NAME}"
> +
> + _start_nbd_mount_server $disk_capacity $fs_type
> + nbd-client localhost 8000 /dev/nbd0 >> "$FULL" 2>&1
> +
> + for ((i = 0; i < $run_cnt; i++))
> + do
> + src/mount_clear_sock "${TMPDIR}/mount_point" /dev/nbd0 $fs_type
> + done
> +
> + nbd-client -d /dev/nbd0
> + _stop_nbd_mount_server
> +}
> diff --git a/tests/nbd/003.out b/tests/nbd/003.out
> new file mode 100644
> index 0000000..aa340db
> --- /dev/null
> +++ b/tests/nbd/003.out
> @@ -0,0 +1 @@
> +Running nbd/003
> --
> 2.13.6
>
prev parent reply other threads:[~2019-12-19 22:35 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-30 12:05 [PATCH blktests v3] nbd/003:add mount and clear_sock test for nbd Sun Ke
2019-11-30 12:04 ` sunke (E)
2019-12-19 22:35 ` Omar Sandoval [this message]
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=20191219223519.GB830111@vader \
--to=osandov@osandov.com \
--cc=linux-block@vger.kernel.org \
--cc=osandov@fb.com \
--cc=sunke32@huawei.com \
/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