From: Eryu Guan <eguan@redhat.com>
To: Christoph Hellwig <hch@lst.de>
Cc: fstests@vger.kernel.org, linux-xfs@vger.kernel.org
Subject: Re: add a new test for racing AIO COW writes V2
Date: Mon, 9 Oct 2017 15:18:33 +0800 [thread overview]
Message-ID: <20171009071833.GE10593@eguan.usersys.redhat.com> (raw)
In-Reply-To: <20171004061616.GA20222@lst.de>
On Wed, Oct 04, 2017 at 08:16:16AM +0200, Christoph Hellwig wrote:
> This can be used to trigger an assert in the current XFS code because it
> can't handle the case where there are COW extents on a file, but none
> at or below the range converted by the AIO completion handler.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
I fixed several minor issues on commit (see inline comments).
I haven't reproduced the assert failure yet on my test vm, but as the
commit log in v1 patch, it's not 100% reproducible, maybe I'm just lucky
(or not lucky :)
>
> diff --git a/src/aio-dio-regress/aio-dio-cow-race.c b/src/aio-dio-regress/aio-dio-cow-race.c
> new file mode 100755
> index 0000000..c7f8e89
> --- /dev/null
> +++ b/src/aio-dio-regress/aio-dio-cow-race.c
Added new gitignore entry for this new binary file.
> @@ -0,0 +1,119 @@
> +/*
> + * Copyright (c) Christoph Hellwig. All Rights Reserved.
> + *
> + * 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.
> + *
> + * This program is distributed in the hope that it would 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 the Free Software Foundation,
> + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
> + */
> +#include <sys/stat.h>
> +#include <sys/types.h>
> +#include <errno.h>
> +#include <libgen.h>
> +#include <fcntl.h>
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +#include <ctype.h>
> +#include <sys/ioctl.h>
> +#include <linux/fs.h>
> +
> +#include <libaio.h>
> +
> +#ifndef FICLONE
> +#define FICLONE _IOW(0x94, 9, int)
> +#endif
> +
> +#define IO_PATTERN 0xab
> +
> +int main(int argc, char *argv[])
> +{
> + struct io_context *ctx = NULL;
> + struct io_event evs[4];
> + struct iocb iocb1, iocb2, iocb3, iocb4;
> + struct iocb *iocbs[] = { &iocb1, &iocb2, &iocb3, &iocb4 };
> + void *buf;
> + int fd, clone, err = 0;
> + unsigned long buf_size = getpagesize() * 2;
> + char *filename, *clonename;
> +
> + if (argc != 3) {
> + fprintf(stderr, "usage: %s filename clonename\n", argv[0]);
> + exit(1);
> + }
> +
> + err = posix_memalign(&buf, getpagesize(), buf_size);
> + if (err) {
> + fprintf(stderr, "error %s during %s\n",
> + strerror(err), "posix_memalign");
> + return 1;
> + }
> + memset(buf, IO_PATTERN, buf_size);
> +
> + filename = argv[1];
> + fd = open(filename, O_DIRECT | O_CREAT | O_TRUNC | O_RDWR, 0600);
> + if (fd == -1) {
> + perror("open");
> + return 1;
> + }
> +
> + if (write(fd, buf, buf_size) != buf_size) {
> + perror("write");
> + return 1;
> + }
> +
> + clonename = argv[2];
> + clone = open(clonename, O_DIRECT | O_CREAT | O_TRUNC | O_RDWR, 0600);
> + if (clone == -1) {
> + perror("open clone");
> + return 1;
> + }
> +
> + if (ioctl(clone, FICLONE, fd)) {
> + perror("FICLONE");
> + return 1;
> + }
> +
> + err = io_setup(4, &ctx);
> + if (err) {
> + fprintf(stderr, "error %s during %s\n",
> + strerror(err),
> + "io_setup");
This error message format looks weired, I changed it to
fprintf(stderr, "io_setup error %s\n", strerror(err));
Did the same to io_submit and io_getevents.
> + return 1;
> + }
> +
> + /*
> + * Test overlapping aio writes, where an earlier one clears the whole
> + * range of a later aio, thus leaving nothing to do for the I/O
> + * completion to do. To make things harder also make sure there is
> + * other outstanding COW I/O.
> + */
> + io_prep_pwrite(&iocb1, fd, buf, buf_size, 0);
> + io_prep_pwrite(&iocb2, fd, buf, buf_size / 2, 0);
> + io_prep_pwrite(&iocb3, fd, buf, buf_size / 2, buf_size);
> + io_prep_pwrite(&iocb4, fd, buf, buf_size, 0);
> +
> + err = io_submit(ctx, 4, iocbs);
> + if (err != 4) {
> + fprintf(stderr, "error %s during %s\n",
> + strerror(err), "io_submit");
> + return 1;
> + }
> +
> + err = io_getevents(ctx, 4, 4, evs, NULL);
> + if (err != 4) {
> + fprintf(stderr, "error %s during %s\n",
> + strerror(err), "io_getevents");
> + return 1;
> + }
> +
> + return 0;
> +}
> diff --git a/tests/generic/463 b/tests/generic/463
> new file mode 100755
> index 0000000..b3afbb0
> --- /dev/null
> +++ b/tests/generic/463
> @@ -0,0 +1,60 @@
> +#! /bin/bash
> +# FS QA Test 463
> +#
> +# Test racy COW AIO write completions.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017 Christoph Hellwig. All Rights Reserved.
> +#
> +# 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.
> +#
> +# This program is distributed in the hope that it would 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 the Free Software Foundation,
> +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
> +#-----------------------------------------------------------------------
> +#
> +
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +tmp=/tmp/$$
> +status=1 # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> + cd /
> + rm -f $tmp.*
> + rm -f $TEST_DIR/file $TEST_DIR/clone
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/reflink
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +_supported_fs generic
> +_supported_os Linux
> +
> +_require_test
> +_require_test_reflink
> +_require_aio
> +_require_odirect
Need _require_test_program "aio-dio-regress/aio-dio-cow-race"
But a single '_require_aiodio "aio-dio-cow-race"' did all the three
requirements check above.
> +
> +$here/src/aio-dio-regress/aio-dio-cow-race $TEST_DIR/file $TEST_DIR/clone
With _require_aiodio, we can use $AIO_TEST here for the test binary.
> +
> +echo "Silence is golden"
Added generic/463.out file too.
Thanks,
Eryu
> +status=0
> +exit
> diff --git a/tests/generic/group b/tests/generic/group
> index f2a6cda..014d6ea 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -465,3 +465,4 @@
> 460 auto quick rw
> 461 auto shutdown stress
> 462 auto quick dax
> +463 auto quick clone dangerous
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2017-10-09 7:18 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-03 14:38 add a new test for racing AIO COW writes Christoph Hellwig
2017-10-03 16:02 ` Darrick J. Wong
2017-10-04 6:16 ` add a new test for racing AIO COW writes V2 Christoph Hellwig
2017-10-04 17:30 ` Darrick J. Wong
2017-10-09 7:18 ` Eryu Guan [this message]
2017-10-09 7:28 ` Christoph Hellwig
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=20171009071833.GE10593@eguan.usersys.redhat.com \
--to=eguan@redhat.com \
--cc=fstests@vger.kernel.org \
--cc=hch@lst.de \
--cc=linux-xfs@vger.kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).