[PATCHSET 0/3] fstests: random fixes for v2023.12.10

[PATCHSET 0/3] fstests: random fixes for v2023.12.10

[Darrick J. Wong]

Hi all,

Here's the usual odd fixes for fstests.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=random-fixes

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=random-fixes

fstests git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=random-fixes
---
 tests/generic/410 |    2 +-
 tests/generic/615 |    6 ++++--
 tests/generic/735 |    1 +
 3 files changed, 6 insertions(+), 3 deletions(-)

[PATCH 1/3] generic/615: fix loop termination failures

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

On 6.7-rc2, I've noticed that this test hangs unpredictably because the
stat loop fails to exit.  While the kill $loop_pid command /should/ take
care of it, it clearly isn't.

Set up an additional safety factor by checking for the existence of a
sentinel flag before starting the loop body.  In bash, "[" is a builtin
so the loop should run almost as tightly as it did before.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 tests/generic/615 |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)


diff --git a/tests/generic/615 b/tests/generic/615
index 4979306d56..9411229874 100755
--- a/tests/generic/615
+++ b/tests/generic/615
@@ -21,11 +21,10 @@ _require_odirect
 
 stat_loop()
 {
-	trap "wait; exit" SIGTERM
 	local filepath=$1
 	local blocks
 
-	while :; do
+	while [ -e "$loop_file" ]; do
 		blocks=$(stat -c %b $filepath)
 		if [ $blocks -eq 0 ]; then
 		    echo "error: stat(2) reported zero blocks"
@@ -39,6 +38,8 @@ _scratch_mount
 
 $XFS_IO_PROG -f -s -c "pwrite -b 64K 0 64K" $SCRATCH_MNT/foo > /dev/null
 
+loop_file=$tmp.loopfile
+touch $loop_file
 stat_loop $SCRATCH_MNT/foo &
 loop_pid=$!
 
@@ -64,6 +65,7 @@ for ((i = 0; i < 2000; i++)); do
 	$XFS_IO_PROG -d -c "pwrite -b 64K 0 64K" $SCRATCH_MNT/foo > /dev/null
 done
 
+rm -f $loop_file
 kill $loop_pid &> /dev/null
 wait
 

[PATCH 2/3] generic/410: don't blow away seqres.full during test

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Don't truncate $seqres.full every time we format a new filesystem; this
makes debugging of this weird failure:

--- /tmp/fstests/tests/generic/410.out	2023-07-11 12:18:21.642971022 -0700
+++ /var/tmp/fstests/generic/410.out.bad	2023-11-29 01:13:00.020000000 -0800
@@ -107,6 +107,9 @@ mpB/dir SCRATCH_DEV
 mpC SCRATCH_DEV
 mpC/dir SCRATCH_DEV
 ======
+mkdir: cannot create directory '/mnt/410/3871733_mpA': File exists
+mkdir: cannot create directory '/mnt/410/3871733_mpB': File exists
+mkdir: cannot create directory '/mnt/410/3871733_mpC': File exists
 make-shared a slave shared mount
 before make-shared run on slave shared
 ------

nearly impossible.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 tests/generic/410 |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


diff --git a/tests/generic/410 b/tests/generic/410
index 8cc36d9f38..5fb5441a0c 100755
--- a/tests/generic/410
+++ b/tests/generic/410
@@ -93,7 +93,7 @@ start_test()
 {
 	local type=$1
 
-	_scratch_mkfs >$seqres.full 2>&1
+	_scratch_mkfs >>$seqres.full 2>&1
 	_get_mount -t $FSTYP $SCRATCH_DEV $MNTHEAD
 	$MOUNT_PROG --make-"${type}" $MNTHEAD
 	mkdir $mpA $mpB $mpC

[PATCH 3/3] generic/735: skip this test if we cannot finsert at pos 1M

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add a _require_congruent_file_oplen to screen out filesystem
configurations that can't start a finsert operation at file pos 1M
because the fs block size isn't congruent with 1048576.  For example,
xfs realtime with 28k rt extents.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 tests/generic/735 |    1 +
 1 file changed, 1 insertion(+)


diff --git a/tests/generic/735 b/tests/generic/735
index 44b454771d..75b23d5efc 100755
--- a/tests/generic/735
+++ b/tests/generic/735
@@ -25,6 +25,7 @@ dev_size=$((80 * 1024 * 1024))
 _scratch_mkfs_sized $dev_size >>$seqres.full 2>&1 || _fail "mkfs failed"
 
 _scratch_mount
+_require_congruent_file_oplen $SCRATCH_MNT 1048576	# finsert at 1M
 file_blksz="$(_get_file_block_size ${SCRATCH_MNT})"
 
 # Reserve 1M space

Re: [PATCH 1/3] generic/615: fix loop termination failures

[Christoph Hellwig]

On Wed, Dec 13, 2023 at 02:34:33PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> On 6.7-rc2, I've noticed that this test hangs unpredictably because the
> stat loop fails to exit.  While the kill $loop_pid command /should/ take
> care of it, it clearly isn't.
> 
> Set up an additional safety factor by checking for the existence of a
> sentinel flag before starting the loop body.  In bash, "[" is a builtin
> so the loop should run almost as tightly as it did before.

Loks good:

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

Re: [PATCH 2/3] generic/410: don't blow away seqres.full during test

[Christoph Hellwig]

Looks good:

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

Re: [PATCH 3/3] generic/735: skip this test if we cannot finsert at pos 1M

[Christoph Hellwig]

On Wed, Dec 13, 2023 at 02:34:45PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> Add a _require_congruent_file_oplen to screen out filesystem
> configurations that can't start a finsert operation at file pos 1M
> because the fs block size isn't congruent with 1048576.  For example,
> xfs realtime with 28k rt extents.

Looks good:

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