f2fs/011: test to check out-of-space issue
PATCH 1/1 f2fs/011: test to check out-of-space issue
This is a regression testcase to check whether we will handle out-of-space case correctly during fallocate() on pinned file once we disable checkpoint.
Testcase:
1. mount f2fs w/ checkpoint=disable option
2. create fragmented file data
3. set flag w/ pinned flag
4. fallocate space for pinned file, expects panic due to running out of space.
We should apply commit 48ea8b200414 ("f2fs: fix to avoid panic once fallocation fails for pinfile") to avoid system panic. Note that it only fix the issue for regular block device, problem still exist for zoned block device.
--- tests/f2fs/011 | 58 ++++++++++++++++++++++++++++++++++++++++++++++ tests/f2fs/011.out | 2 ++ 2 files changed, 60 insertions(+) create mode 100755 tests/f2fs/011 create mode 100644 tests/f2fs/011.out
tests/f2fs/011 new file +58 -0
@@ -0,0 +1,58 @@ +#! /bin/bash +# SPDX-License-Identifier: GPL-2.0 +# Copyright (c) 2025 Chao Yu. All Rights Reserved. +# +# FS QA Test No. f2fs/012 +# +# This is a regression testcase to check whether we will handle +# out-of-space case correctly during fallocate() on pinned file +# once we disable checkpoint. +# 1. mount f2fs w/ checkpoint=disable option +# 2. create fragmented file data +# 3. set flag w/ pinned flag +# 4. fallocate space for pinned file, expects panic due to running +# out of space +# We should apply commit 48ea8b200414 ("f2fs: fix to avoid panic +# once fallocation fails for pinfile") to avoid system panic. +# Note that it only fix the issue for regular block device, problem +# still exist for zoned block device. +# +. ./common/preamble +_begin_fstest auto quick + +_cleanup() +{ + cd / + rm -r -f $tmp.* +} + +_fixed_by_kernel_commit 48ea8b200414 \ + "f2fs: fix to avoid panic once fallocation fails for pinfile" + +_require_scratch +_require_command "$F2FS_IO_PROG" f2fs_io + +_scratch_mkfs_sized $((1*1024*1024*1024)) >> $seqres.full +_scratch_mount -o checkpoint=disable:10% + +pinfile=$SCRATCH_MNT/file + +# simulate fragment status in f2fs +for ((i=0;i<256;i++)) do + dd if=/dev/zero of=$SCRATCH_MNT/$i bs=1M count=1 >>$seqres.full 2>&1 +done +sync + +for ((i=0;i<256;i+=2)) do + rm $SCRATCH_MNT/$i >> $seqres.full 2>&1 +done +sync + +touch $pinfile +$F2FS_IO_PROG pinfile set $pinfile >> $seqres.full +fallocate -l $((3*256*1024*1024)) $pinfile >> $seqres.full 2>&1 + +echo "Silence is golden" + +status=0 +exit
tests/f2fs/011.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 011 +Silence is golden
Signature · 1 line
2.49.0
Quoted 48 lines · Chao Yu
This is a regression testcase to check whether we will handle out-of-space case correctly during
fallocate()on pinned file once we disable checkpoint.Testcase:
1. mount f2fs w/ checkpoint=disable option
2. create fragmented file data
3. set flag w/ pinned flag
4. fallocate space for pinned file, expects panic due to running out of space.We should apply commit
48ea8b200414("f2fs: fix to avoid panic once fallocation fails for pinfile") to avoid system panic. Note that it only fix the issue for regular block device, problem still exist for zoned block device.Cc Jaegeuk Kim <[email protected]>Signed-off-by Chao Yu <[email protected]>--- tests/f2fs/011 | 58 ++++++++++++++++++++++++++++++++++++++++++++++ tests/f2fs/011.out | 2 ++ 2 files changed, 60 insertions(+) create mode 100755 tests/f2fs/011 create mode 100644 tests/f2fs/011.outtests/f2fs/011 new file +17
-0@@ -0,0 +1,58 @@ +#! /bin/bash +# SPDX-License-Identifier: GPL-2.0 +# Copyright (c) 2025 Chao Yu. All Rights Reserved. +# +# FS QA Test No. f2fs/012 +# +# This is a regression testcase to check whether we will handle +# out-of-space case correctly during fallocate() on pinned file +# once we disable checkpoint. +# 1. mount f2fs w/ checkpoint=disable option +# 2. create fragmented file data +# 3. set flag w/ pinned flag +# 4. fallocate space for pinned file, expects panic due to running +# out of space +# We should apply commit 48ea8b200414 ("f2fs: fix to avoid panic +# once fallocation fails for pinfile") to avoid system panic. +# Note that it only fix the issue for regular block device, problem
tests/f2fs/011 line 27
Quoted 10 lines
tests/f2fs/011 new file +10
-0+# still exist for zoned block device. +# +. ./common/preamble +_begin_fstest auto quick + +_cleanup() +{ + cd / + rm -r -f $tmp.* +}
Default _cleanup does the same things, so above _cleanup can be removed.
Quoted 5 lines
tests/f2fs/011 new file +5
-0+ +_fixed_by_kernel_commit 48ea8b200414 \ + "f2fs: fix to avoid panic once fallocation fails for pinfile" + +_require_scratch
tests/f2fs/011 line 42
Quoted 10 lines
tests/f2fs/011 new file +10
-0+_require_command "$F2FS_IO_PROG" f2fs_io + +_scratch_mkfs_sized $((1*1024*1024*1024)) >> $seqres.full +_scratch_mount -o checkpoint=disable:10% + +pinfile=$SCRATCH_MNT/file + +# simulate fragment status in f2fs +for ((i=0;i<256;i++)) do + dd if=/dev/zero of=$SCRATCH_MNT/$i bs=1M count=1 >>$seqres.full 2>&1
We'd like to use xfs_io when we can:
$XFS_IO_PROG -f -c "pwrite 0 1m" $SCRATCH_MNT/$i >>$seqres.full
tests/f2fs/011 line 47
Quoted 5 lines
tests/f2fs/011 new file +5
-0+done +sync + +for ((i=0;i<256;i+=2)) do + rm $SCRATCH_MNT/$i >> $seqres.full 2>&1
How about "rm -f $SCRATCH_MNT/$i" without ">> $seqres.full 2>&1" ?
tests/f2fs/011 line 53
Quoted 6 lines
tests/f2fs/011 new file +6
-0+done +sync + +touch $pinfile +$F2FS_IO_PROG pinfile set $pinfile >> $seqres.full +fallocate -l $((3*256*1024*1024)) $pinfile >> $seqres.full 2>&1
Same, we'd like to use xfs_io at first, due to it's already a necessary running dependence of fstests.
$XFS_IO_PROG -c "falloc 0 $((3*256*1024*1024))" $pinfile
Thanks,
Zorro
Quoted 16 lines
tests/f2fs/011 new file +5
-0+ +echo "Silence is golden" + +status=0 +exittests/f2fs/011.out new file +2
-0@@ -0,0 +1,2 @@ +QA output created by 011 +Silence is goldenSignature · 1 line
2.49.0