[PATCH v4.3 0/3] xfstests: minor fixes for the reflink/dedupe tests

[PATCH v4.3 0/3] xfstests: minor fixes for the reflink/dedupe tests

[Darrick J. Wong]

This is a patch set against the reflink/dedupe test cases in xfstests.
All three patches fix some minor problems in the giant testcase merge
last weekend.

The zeroth patch ought to be Filipe Manana's patch ("fstests:
generic/158, test dedupe with destination offset past EOF") but as he
just resent that this morning, I won't repost it here.

The first patch, then, corrects the error output in the high offset
dedupe test (generic/304).

The second patch changes punch-alternating.c to use the filesystem
accounting block size to determine where to punch holes in the file
because that value is the internal filesystem block size.  This makes
the program behave as expected when testing with non-page-size block
sizes on XFS.

The third patch adds to _require_xfs_io_command the ability to check
that a subcommand supports a particular command line argument and adds
to xfs/24[356] the ability to check for the -c switch to the xfs_io
bmap command.

Comments and questions are, as always, welcome.  The fixes are
available on github[1].

--D

[1] https://github.com/djwong/xfstests/tree/for-dave

[PATCH 1/3] generic/304: fix high offset

[Darrick J. Wong]

"Invalid argument" is a better response to an impossibly high offset
dedupe request than "extents don't match", so change the test.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 tests/generic/304.out |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


diff --git a/tests/generic/304.out b/tests/generic/304.out
index 4af2a39..7b5ff0e 100644
--- a/tests/generic/304.out
+++ b/tests/generic/304.out
@@ -8,7 +8,7 @@ dedupe: Invalid argument
 Dedupe past maximum file size in dest file (should fail)
 dedupe: Invalid argument
 Dedupe high offset to low offset
-dedupe: Extents did not match.
+dedupe: Invalid argument
 Dedupe past source file EOF (should fail)
 dedupe: Invalid argument
 Dedupe max size at nonzero offset (should fail)

Re: [PATCH 1/3] generic/304: fix high offset

[Christoph Hellwig]

On Mon, Feb 15, 2016 at 12:46:14PM -0800, Darrick J. Wong wrote:
> "Invalid argument" is a better response to an impossibly high offset
> dedupe request than "extents don't match", so change the test.

What does btrfs return here?

Re: [PATCH 1/3] generic/304: fix high offset

[Filipe Manana]

On Wed, Feb 17, 2016 at 9:56 AM, Christoph Hellwig <hch@infradead.org> wrote:
> On Mon, Feb 15, 2016 at 12:46:14PM -0800, Darrick J. Wong wrote:
>> "Invalid argument" is a better response to an impossibly high offset
>> dedupe request than "extents don't match", so change the test.
>
> What does btrfs return here?

-EINVAL, so this makes the test pass on btrfs.

> --
> 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



-- 
Filipe David Manana,

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."

Re: [PATCH 1/3] generic/304: fix high offset

[Christoph Hellwig]

On Wed, Feb 17, 2016 at 10:04:23AM +0000, Filipe Manana wrote:
> > What does btrfs return here?
> 
> -EINVAL, so this makes the test pass on btrfs.

Ok, looks fine to me then:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 2/3] punch-alternating: use the block size reported by the fs for punching

[Darrick J. Wong]

When we're trying to punch alternating blocks out of a file, use the
bsize reported by fstatfs so that we can punch out single blocks.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 src/punch-alternating.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)


diff --git a/src/punch-alternating.c b/src/punch-alternating.c
index 9566310..4148622 100644
--- a/src/punch-alternating.c
+++ b/src/punch-alternating.c
@@ -14,6 +14,7 @@
 int main(int argc, char *argv[])
 {
 	struct stat	s;
+	struct statfs	sf;
 	off_t		offset;
 	int		fd;
 	blksize_t	blksz;
@@ -35,8 +36,12 @@ int main(int argc, char *argv[])
 	if (error)
 		goto err;
 
+	error = fstatfs(fd, &sf);
+	if (error)
+		goto err;
+
 	sz = s.st_size;
-	blksz = s.st_blksize;
+	blksz = sf.f_bsize;
 
 	mode = FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE;
 	for (offset = 0; offset < sz; offset += blksz * 2) {

Re: [PATCH 2/3] punch-alternating: use the block size reported by the fs for punching

[Christoph Hellwig]

Looks fine,

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 3/3] xfs/24[356]: check for -c switch to xfs_io bmap command

[Darrick J. Wong]

Extend _require_xfs_io_command so that we can pass it a command line
argument to look for, and then use new capabililty in the relevant
tests to ensure that bmap knows how to dump CoW fork contents (bmap -c)

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 common/rc     |    9 +++++++--
 tests/xfs/243 |    1 +
 tests/xfs/245 |    1 +
 tests/xfs/246 |    1 +
 4 files changed, 10 insertions(+), 2 deletions(-)


diff --git a/common/rc b/common/rc
index b11f170..e810062 100644
--- a/common/rc
+++ b/common/rc
@@ -1831,12 +1831,13 @@ _user_do()
 
 _require_xfs_io_command()
 {
-	if [ $# -ne 1 ]
+	if [ -z "$1" ]
 	then
-		echo "Usage: _require_xfs_io_command command" 1>&2
+		echo "Usage: _require_xfs_io_command command [switch]" 1>&2
 		exit 1
 	fi
 	command=$1
+	param=$2
 
 	testfile=$TEST_DIR/$$.xfs_io
 	case $command in
@@ -1866,6 +1867,10 @@ _require_xfs_io_command()
 		_notrun "xfs_io $command support is missing"
 	echo $testio | grep -q "Operation not supported" && \
 		_notrun "xfs_io $command failed (old kernel/wrong fs?)"
+
+	test -z "$param" && return
+	$XFS_IO_PROG -c "help $command" | grep -q "^ $param --" || \
+		_notrun "xfs_io $command doesn't support $param"
 }
 
 # check that xfs_db supports a specific command
diff --git a/tests/xfs/243 b/tests/xfs/243
index 629173e..d75cfd8 100755
--- a/tests/xfs/243
+++ b/tests/xfs/243
@@ -56,6 +56,7 @@ _supported_fs xfs
 _require_scratch_reflink
 _require_xfs_io_command "falloc"
 _require_xfs_io_command "fpunch"
+_require_xfs_io_command "bmap" "-c"
 _require_cp_reflink
 
 rm -f $seqres.full
diff --git a/tests/xfs/245 b/tests/xfs/245
index 6da701d..7f80260 100755
--- a/tests/xfs/245
+++ b/tests/xfs/245
@@ -49,6 +49,7 @@ _supported_os Linux
 _supported_fs xfs
 _require_scratch_reflink
 _require_xfs_io_command "falloc"
+_require_xfs_io_command "bmap" "-c"
 _require_cp_reflink
 
 rm -f $seqres.full
diff --git a/tests/xfs/246 b/tests/xfs/246
index 13b0c98..bf471cc 100755
--- a/tests/xfs/246
+++ b/tests/xfs/246
@@ -42,6 +42,7 @@ _cleanup()
 # real QA test starts here
 _supported_os Linux
 _supported_fs xfs
+_require_xfs_io_command "bmap" "-c"
 
 rm -f $seqres.full
 

Re: [PATCH 3/3] xfs/24[356]: check for -c switch to xfs_io bmap command

[Christoph Hellwig]

Looks fine,

Reviewed-by: Christoph Hellwig <hch@lst.de>