Re: [PATCH 1/9] generic: test XATTR_REPLACE doesn't take the fs down

[Eryu Guan <guaneryu@gmail.com>]

On Tue, May 01, 2018 at 08:39:26AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Kanda Motohiro reported that expanding a tiny xattr into a large xattr
> fails on XFS because we remove the tiny xattr from a shortform fork and
> then try to re-add it after converting the fork to extents format having
> not removed the ATTR_REPLACE flag.  This fails because the attr is no
> longer present, causing a fs shutdown.
> 
> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=199119
> Reported-by: kanda.motohiro@gmail.com
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Thanks for the revision! Because I found more issues in v1 than I
expected, so I drop the patch from last fstests update. I should have
made that clear..

> ---
>  .gitignore              |    1 +
>  src/Makefile            |    3 +-
>  src/attr_replace_test.c |   60 +++++++++++++++++++++++++++++++++++++++++
>  tests/generic/706       |   68 +++++++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/706.out   |    2 +
>  tests/generic/group     |    1 +
>  6 files changed, 134 insertions(+), 1 deletion(-)
>  create mode 100644 src/attr_replace_test.c
>  create mode 100755 tests/generic/706
>  create mode 100644 tests/generic/706.out
> 
> 
> diff --git a/.gitignore b/.gitignore
> index 192ca35e..af9743f9 100644
> --- a/.gitignore
> +++ b/.gitignore
> @@ -53,6 +53,7 @@
>  /src/append_reader
>  /src/append_writer
>  /src/attr-list-by-handle-cursor-test
> +/src/attr_replace_test

Seems this should be put before attr-list-by-handle-cursor-test,
according to the result of sort.

>  /src/bstat
>  /src/bulkstat_unlink_test
>  /src/bulkstat_unlink_test_modified
> diff --git a/src/Makefile b/src/Makefile
> index 6ca56366..c42d3bb1 100644
> --- a/src/Makefile
> +++ b/src/Makefile
> @@ -25,7 +25,8 @@ LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
>  	seek_copy_test t_readdir_1 t_readdir_2 fsync-tester nsexec cloner \
>  	renameat2 t_getcwd e4compact test-nextquota punch-alternating \
>  	attr-list-by-handle-cursor-test listxattr dio-interleaved t_dir_type \
> -	dio-invalidate-cache stat_test t_encrypted_d_revalidate
> +	dio-invalidate-cache stat_test t_encrypted_d_revalidate \
> +	attr_replace_test
>  
>  SUBDIRS = log-writes perf
>  
> diff --git a/src/attr_replace_test.c b/src/attr_replace_test.c
> new file mode 100644
> index 00000000..c870d165
> --- /dev/null
> +++ b/src/attr_replace_test.c
> @@ -0,0 +1,60 @@
> +// setattr.c by kanda.motohiro@gmail.com
> +// xfs extended attribute corruption bug reproducer
> +#include <stdio.h>
> +#include <string.h>
> +#include <fcntl.h>
> +#include <errno.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +#include <sys/types.h>
> +#include <sys/xattr.h>
> +#include <sys/stat.h>
> +
> +#define die() do { perror(""); \
> +fprintf(stderr, "error=%d at line %d\n", errno, __LINE__); \
> +exit(1); } while (0)
> +
> +int main(int argc, char *argv[])
> +{
> +	int ret;
> +	int fd;
> +	char *path;
> +	char *name = "user.world";
> +	char *value;
> +	struct stat sbuf;
> +	size_t size = sizeof(value);
> +
> +	if (argc != 2) die();

die() doesn't seem like the right thing to do (and some other places),
it'll print something like:

Success
error=0 at line NN

I'd define a new "fail()" macro and use it where appropriate.

+#define fail(...) do { \
+fprintf(stderr, __VA_ARGS__); exit (1); \
+} while (0)

I can fix the minor issues on commit, for real this time :)

Thanks,
Eryu

> +	path = argv[1];
> +
> +	fd = open(path, O_RDWR | O_CREAT, S_IRUSR | S_IWUSR);
> +	if (fd < 0) die();
> +
> +	/*
> +	 * The value should be 3/4 the size of a fs block to ensure that we
> +	 * get to extents format.
> +	 */
> +	ret = fstat(fd, &sbuf);
> +	if (ret < 0) die();
> +	size = sbuf.st_blksize * 3 / 4;
> +	if (!size) die();
> +	value = malloc(size);
> +	if (!value) die();
> +
> +	// First, create a small xattr.
> +	memset(value, '0', 1);
> +	ret = fsetxattr(fd, name, value, 1, XATTR_CREATE);
> +	if (ret < 0) die();
> +	close(fd);
> +
> +	fd = open(path, O_RDWR);
> +	if (fd < 0) die();
> +
> +	// Then, replace it with bigger one, forcing short form to leaf conversion.
> +	memset(value, '1', size);
> +	ret = fsetxattr(fd, name, value, size, XATTR_REPLACE);
> +	if (ret < 0) die();
> +	close(fd);
> +
> +	return 0;
> +}
> diff --git a/tests/generic/706 b/tests/generic/706
> new file mode 100755
> index 00000000..9f98c3b1
> --- /dev/null
> +++ b/tests/generic/706
> @@ -0,0 +1,68 @@
> +#! /bin/bash
> +# FS QA Test No. 706
> +#
> +# Ensure that we can XATTR_REPLACE a tiny attr into a large attr.
> +# Kanda Motohiro <kanda.motohiro@gmail.com> reports that XATTR_REPLACE'ing
> +# a single-byte attr with a 2048-byte attr causes a fs shutdown because we
> +# remove the shortform attr, convert the attr fork to long format, and then
> +# try to re-add the attr having not cleared ATTR_REPLACE.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2018 Oracle.  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"
> +
> +status=1	# failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> +	cd /
> +	rm -f $testfile
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/attr
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +_supported_fs generic
> +_supported_os Linux
> +_require_test_program "attr_replace_test"
> +_require_attrs
> +_require_scratch
> +
> +rm -f $seqres.full
> +_scratch_mkfs >>$seqres.full 2>&1
> +_scratch_mount >>$seqres.full 2>&1
> +
> +filter_attr_output() {
> +	_filter_scratch | sed -e 's/has a [0-9]* byte value/has a NNNN byte value/g'
> +}
> +
> +./src/attr_replace_test $SCRATCH_MNT/hello
> +$ATTR_PROG -l $SCRATCH_MNT/hello | filter_attr_output
> +
> +status=0
> +exit
> diff --git a/tests/generic/706.out b/tests/generic/706.out
> new file mode 100644
> index 00000000..61c8419a
> --- /dev/null
> +++ b/tests/generic/706.out
> @@ -0,0 +1,2 @@
> +QA output created by 706
> +Attribute "world" has a NNNN byte value for SCRATCH_MNT/hello
> diff --git a/tests/generic/group b/tests/generic/group
> index 19be9267..54a85ea0 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -487,3 +487,4 @@
>  482 auto metadata replay
>  483 auto quick log metadata
>  484 auto quick
> +706 auto quick attr
>