Skip to contentSkip to conversation

tests for btrfs fsverity

PATCH Superseded Boris Burkov 3 participants 7 messages Apr 8, 2021 – Apr 11, 2021

Subsystems (2)

v3 of 13 Superseded Latest is v13 →

lore
Share & tools
lore
#0 lore raw

v3 cover tests for btrfs fsverity

This patchset provides tests for fsverity support in btrfs.

It includes modifications for generic tests to pass with btrfs as well as new btrfs specific tests.

--
v3: rebase onto xfstests master branch
v2: pass generic tests, add logwrites test

Boris Burkov (3):
btrfs: test btrfs specific fsverity corruption
generic/574: corrupt btrfs merkle tree data
btrfs: test verity orphans with dmlogwrites

 common/config       |   1 +
 common/verity       |  22 ++++-
 tests/btrfs/290     | 190 ++++++++++++++++++++++++++++++++++++++++++++
 tests/btrfs/290.out |  17 ++++
 tests/btrfs/291     | 156 ++++++++++++++++++++++++++++++++++++
 tests/btrfs/291.out |   2 +
 tests/btrfs/group   |   2 +
 7 files changed, 388 insertions(+), 2 deletions(-)
 create mode 100755 tests/btrfs/290
 create mode 100644 tests/btrfs/290.out
 create mode 100755 tests/btrfs/291
 create mode 100644 tests/btrfs/291.out
Signature · 1 line
2.30.2
#1 lore raw

PATCH v3 1/3 btrfs: test btrfs specific fsverity corruption

There are some btrfs specific fsverity scenarios that don't map neatly onto the tests in generic/574 like holes, inline extents, and preallocated extents. Cover those in a btrfs specific test.

This test relies on the btrfs implementation of fsverity in:
<kernel patches>
and it relies on btrfs-corrupt-block for corruption, with the patches:
<corrupt block patches>

---
 common/config       |   1 +
 common/verity       |   7 ++
 tests/btrfs/290     | 190 ++++++++++++++++++++++++++++++++++++++++++++
 tests/btrfs/290.out |  17 ++++
 tests/btrfs/group   |   1 +
 5 files changed, 216 insertions(+)
 create mode 100755 tests/btrfs/290
 create mode 100644 tests/btrfs/290.out
common/config +1 -0
index a47e462c..003b2a88 100644
@@ -256,6 +256,7 @@ export BTRFS_UTIL_PROG=$(type -P btrfs)
 export BTRFS_SHOW_SUPER_PROG=$(type -P btrfs-show-super)
 export BTRFS_CONVERT_PROG=$(type -P btrfs-convert)
 export BTRFS_TUNE_PROG=$(type -P btrfstune)
+export BTRFS_CORRUPT_BLOCK_PROG=$(type -P btrfs-corrupt-block)
 export XFS_FSR_PROG=$(type -P xfs_fsr)
 export MKFS_NFS_PROG="false"
 export MKFS_CIFS_PROG="false"
common/verity +7 -0
index 38eea157..d2c1ea24 100644
@@ -8,6 +8,10 @@ _require_scratch_verity()
 	_require_scratch
 	_require_command "$FSVERITY_PROG" fsverity
 
+	if [ $FSTYP == "btrfs" ]; then
+		_require_command "$BTRFS_CORRUPT_BLOCK_PROG" btrfs_corrupt_block
+	fi
+
 	if ! _scratch_mkfs_verity &>>$seqres.full; then
 		# ext4: need e2fsprogs v1.44.5 or later (but actually v1.45.2+
 		#       is needed for some tests to pass, due to an e2fsck bug)
@@ -147,6 +151,9 @@ _scratch_mkfs_verity()
 	ext4|f2fs)
 		_scratch_mkfs -O verity
 		;;
+	btrfs)
+		_scratch_mkfs
+		;;
 	*)
 		_notrun "No verity support for $FSTYP"
 		;;
tests/btrfs/290 new file +190 -0

190 lines not on this page · raw message

tests/btrfs/290.out new file +17 -0
new file mode 100644
index 00000000..4da61246
@@ -0,0 +1,17 @@
+QA output created by 290
+cat: /mnt/scratch/inl: Input/output error
+dd: error reading '/mnt/scratch/inl': Input/output error
+cat: /mnt/scratch/prealloc: Input/output error
+dd: error reading '/mnt/scratch/prealloc': Input/output error
+cat: /mnt/scratch/reg: Input/output error
+dd: error reading '/mnt/scratch/reg': Input/output error
+cat: /mnt/scratch/punch: Input/output error
+dd: error reading '/mnt/scratch/punch': Input/output error
+cat: /mnt/scratch/plug: Input/output error
+dd: error reading '/mnt/scratch/plug': Input/output error
+cat: /mnt/scratch/desc: Invalid argument
+dd: failed to open '/mnt/scratch/desc': Invalid argument
+cat: /mnt/scratch/roothash: Input/output error
+dd: error reading '/mnt/scratch/roothash': Input/output error
+cat: /mnt/scratch/merkle: Input/output error
+dd: error reading '/mnt/scratch/merkle': Input/output error
tests/btrfs/group +1 -0
index 331dd432..13051562 100644
@@ -238,3 +238,4 @@
 233 auto quick subvolume
 234 auto quick compress rw
 235 auto quick send
+290 auto quick verity
Signature · 1 line
2.30.2
#4 lore raw

PATCH v3 2/3 generic/574: corrupt btrfs merkle tree data

In reply to Boris Burkov

generic/574 has tests for corrupting the merkle tree data stored by the filesystem. Since btrfs uses a different scheme for storing this data, the existing logic for corrupting it doesn't work out of the box. Adapt it to properly corrupt btrfs merkle items.

Note that there is a bit of a kludge here: since btrfs_corrupt_block doesn't handle streaming corruption bytes from stdin (I could change that, but it feels like overkill for this purpose), I just read the first corruption byte and duplicate it for the desired length. That is how the test is using the interface in practice, anyway.

This relies on the following kernel patch for btrfs verity support:
<btrfs-verity-patch>
And the following btrfs-progs patch for btrfs_corrupt_block support:
<btrfs-corrupt-block-patch>

---
 common/verity | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)
common/verity +13 -2
index d2c1ea24..fdd05783 100644
@@ -3,8 +3,7 @@
 #
 # Functions for setting up and testing fs-verity
 
-_require_scratch_verity()
-{
+_require_scratch_verity() {
 	_require_scratch
 	_require_command "$FSVERITY_PROG" fsverity
 
@@ -315,6 +314,18 @@ _fsv_scratch_corrupt_merkle_tree()
 		(( offset += ($(_get_filesize $file) + 65535) & ~65535 ))
 		_fsv_scratch_corrupt_bytes $file $offset
 		;;
+	btrfs)
+		ino=$(ls -i $file | awk '{print $1}')
+		sync
+		cat > $tmp.bytes
+		sz=$(_get_filesize $tmp.bytes)
+		read -n 1 byte < $tmp.bytes
+		ascii=$(printf "%d" "'$byte'")
+		_scratch_unmount
+		$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,37,0 -v $ascii -o $offset -b $sz $SCRATCH_DEV
+		sync
+		_scratch_mount
+		;;
 	*)
 		_fail "_fsv_scratch_corrupt_merkle_tree() unimplemented on $FSTYP"
 		;;
Signature · 1 line
2.30.2
#6 lore raw

PATCH v3 3/3 btrfs: test verity orphans with dmlogwrites

In reply to Boris Burkov

The behavior of orphans is most interesting across mounts, interrupted at arbitrary points during fsverity enable. To cover as many such cases as possible, use dmlogwrites and dmsnapshot as in log-writes/replay-individual.sh. At each log entry, we want to assert a somewhat complicated invariant:

If verity has not yet started: an orphan indicates that verity has started.
If verity has started: mount should handle the orphan and blow away verity data: expect 0 merkle items after mounting the snapshot dev. If we can measure the file, verity has finished.
If verity has finished: the orphan should be gone, so mount should not blow away merkle items. Expect the same number of merkle items before and after mounting the snapshot dev.

Note that this relies on grepping btrfs inspect-internal dump-tree. Until btrfs-progs has the ability to print the new Merkle items, they will show up as UNKNOWN.36/37.

---
 tests/btrfs/291     | 156 ++++++++++++++++++++++++++++++++++++++++++++
 tests/btrfs/291.out |   2 +
 tests/btrfs/group   |   1 +
 3 files changed, 159 insertions(+)
 create mode 100755 tests/btrfs/291
 create mode 100644 tests/btrfs/291.out
tests/btrfs/291 new file +156 -0
new file mode 100755
index 00000000..61f36426
@@ -0,0 +1,156 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (C) 2021 Facebook, Inc. All Rights Reserved.
+#
+# FS QA Test 291
+#
+# Test btrfs consistency after each FUA while enabling verity on a file
+# This test works by following the pattern in log-writes/replay-individual.sh:
+# 1. run a workload (verity + sync) while logging to the log device
+# 2. replay an entry to the replay device
+# 3. snapshot the replay device to the snapshot device
+# 4. run destructive tests on the snapshot device (e.g. mount with orphans)
+# 5. goto 2
+#
+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 /
+	_log_writes_cleanup &> /dev/null
+	rm -f $tmp.*
+	$LVM_PROG vgremove -f -y $vgname >>$seqres.full 2>&1
+	losetup -d $loop_dev >>$seqres.full 2>&1
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+. ./common/attr
+. ./common/dmlogwrites
+. ./common/verity
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+
+_supported_fs btrfs
+
+_require_scratch
+_require_log_writes
+_require_dm_target snapshot
+_require_command $LVM_PROG lvm
+_require_scratch_verity
+
+sync_loop() {
+	i=$1
+	[ -z "$i" ] && _fail "sync loop needs a number of iterations"
+	while [ $i -gt 0 ]
+	do
+		$XFS_IO_PROG -c sync $SCRATCH_MNT
+		let i-=1
+	done
+}
+
+dump_tree() {
+	dev=$1
+	$BTRFS_UTIL_PROG inspect-internal dump-tree $dev
+}
+
+count_item() {
+	dev=$1
+	item=$2
+	dump_tree $dev | grep -c $item
+}
+
+_log_writes_init $SCRATCH_DEV
+_log_writes_mkfs
+_log_writes_mount
+
+f=$SCRATCH_MNT/fsv
+dd if=/dev/zero of=$f bs=1M count=10 >>$seqres.full 2>&1
+$XFS_IO_PROG -c sync $SCRATCH_MNT
+sync_loop 10 &
+_fsv_enable $f
+$XFS_IO_PROG -c sync $SCRATCH_MNT
+
+_log_writes_unmount
+_log_writes_remove
+
+dd if=/dev/zero of=$tmp.loop-file bs=1M count=1 seek=8192 >>$seqres.full 2>&1
+loop_dev=$(losetup -f --show $tmp.loop-file)
+vgname=vg_replay
+lvname=lv_replay
+replay_dev=/dev/mapper/vg_replay-lv_replay
+snapname=lv_snap
+snap_dev=/dev/mapper/vg_replay-$snapname
+
+$LVM_PROG vgcreate -f $vgname $loop_dev >>$seqres.full 2>&1 || _fail "failed to vgcreate $vgname"
+$LVM_PROG lvcreate -L 4G -n $lvname $vgname -y >>$seqres.full 2>&1 || \
+	_fail "failed to lvcreate $lvname"
+$UDEV_SETTLE_PROG >>$seqres.full 2>&1
+
+replay_log_prog=$here/src/log-writes/replay-log
+num_entries=$($replay_log_prog --log $LOGWRITES_DEV --num-entries)
+entry=$($replay_log_prog --log $LOGWRITES_DEV --replay $replay_dev --find --end-mark mkfs | cut -d@ -f1)
+$replay_log_prog --log $LOGWRITES_DEV --replay $replay_dev --limit $entry || \
+	_fail "failed to replay to start entry $entry"
+let entry+=1
+
+# state = 0: verity hasn't started
+# state = 1: verity underway
+# state = 2: verity done
+state=0
+while [ $entry -lt $num_entries ];
+do
+	$replay_log_prog --limit 1 --log $LOGWRITES_DEV --replay $replay_dev --start $entry || \
+		_fail "failed to take replay step at entry: $entry"
+
+	$LVM_PROG lvcreate -s -L 4M -n $snapname $vgname/$lvname >>$seqres.full 2>&1 || \
+		_fail "Failed to create snapshot"
+	$UDEV_SETTLE_PROG >>$seqres.full 2>&1
+
+	orphan=$(count_item $snap_dev ORPHAN)
+	if [ $state -eq 0 ]; then
+		[ $orphan -gt 0 ] && state=1
+	fi
+
+	pre_mount=$(count_item $snap_dev UNKNOWN.3[67])
+	_mount $snap_dev $SCRATCH_MNT || _fail "mount failed at entry $entry"
+	fsverity measure $SCRATCH_MNT/fsv >>$seqres.full 2>&1
+	measured=$?
+	umount $SCRATCH_MNT
+	[ $state -eq 1 ] && [ $measured -eq 0 ] && state=2
+	[ $state -eq 2 ] && ([ $measured -eq 0 ] || _fail "verity done, but measurement failed at entry $entry")
+	post_mount=$(count_item $snap_dev UNKNOWN.3[67])
+
+	echo "entry: $entry, state: $state, orphan: $orphan, pre_mount: $pre_mount, post_mount: $post_mount" >> $seqres.full
+
+	if [ $state -eq 1 ]; then
+		[ $post_mount -eq 0 ] || \
+			_fail "mount failed to clear under-construction merkle items pre: $pre_mount, post: $post_mount at entry $entry";
+	fi
+	if [ $state -eq 2 ]; then
+		[ $pre_mount -gt 0 ] || \
+			_fail "expected to have verity items before mount at entry $entry"
+		[ $pre_mount -eq $post_mount ] || \
+			_fail "mount cleared merkle items after verity was enabled $pre_mount vs $post_mount at entry $entry";
+	fi
+
+	let entry+=1
+	$LVM_PROG lvremove $vgname/$snapname -y >>$seqres.full
+done
+
+echo "Silence is golden"
+
+# success, all done
+status=0
+exit
tests/btrfs/291.out new file +2 -0
new file mode 100644
index 00000000..04605c70
@@ -0,0 +1,2 @@
+QA output created by 291
+Silence is golden
tests/btrfs/group +1 -0
index 13051562..cc5a811e 100644
@@ -239,3 +239,4 @@
 234 auto quick compress rw
 235 auto quick send
 290 auto quick verity
+291 auto verity
Signature · 1 line
2.30.2
maintainer
#2 lore raw

Re: [PATCH v3 1/3] btrfs: test btrfs specific fsverity corruption

In reply to Boris Burkov

tests/btrfs/290 line 41

Quoted 4 lines · Boris Burkov
tests/btrfs/290 new file +4 -0
+get_ino() {
+	file=$1
+	ls -i $file | awk '{print $1}'
+}

Please use 'local' when declaring variables in shell functions.

tests/btrfs/290 line 59

Quoted 8 lines
tests/btrfs/290 new file +8 -0
+# corrupt the data portion of an inline extent
+corrupt_inline() {
+	f=$SCRATCH_MNT/inl
+	head -c 42 /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount

Isn't the sync unnecessary because the filesystem is unmounted anyway?

Likewise for all the other corrupt_* functions.

Otherwise, at a high level this test looks good -- thanks for writing it!

- Eric

#3 lore raw

Re: [PATCH v3 1/3] btrfs: test btrfs specific fsverity corruption

In reply to Boris Burkov
Quoted 8 lines · Boris Burkov

There are some btrfs specific fsverity scenarios that don't map neatly onto the tests in generic/574 like holes, inline extents, and preallocated extents. Cover those in a btrfs specific test.

This test relies on the btrfs implementation of fsverity in:
<kernel patches>
and it relies on btrfs-corrupt-block for corruption, with the patches:
<corrupt block patches>

I assume you'll fill in the patch lists when they're merged.

Quoted 29 lines
---
 common/config       |   1 +
 common/verity       |   7 ++
 tests/btrfs/290     | 190 ++++++++++++++++++++++++++++++++++++++++++++
 tests/btrfs/290.out |  17 ++++
 tests/btrfs/group   |   1 +
 5 files changed, 216 insertions(+)
 create mode 100755 tests/btrfs/290
 create mode 100644 tests/btrfs/290.out
common/config +1 -0
index a47e462c..003b2a88 100644
@@ -256,6 +256,7 @@ export BTRFS_UTIL_PROG=$(type -P btrfs)
 export BTRFS_SHOW_SUPER_PROG=$(type -P btrfs-show-super)
 export BTRFS_CONVERT_PROG=$(type -P btrfs-convert)
 export BTRFS_TUNE_PROG=$(type -P btrfstune)
+export BTRFS_CORRUPT_BLOCK_PROG=$(type -P btrfs-corrupt-block)
 export XFS_FSR_PROG=$(type -P xfs_fsr)
 export MKFS_NFS_PROG="false"
 export MKFS_CIFS_PROG="false"
common/verity +0 -0
index 38eea157..d2c1ea24 100644

common/verity _require_scratch_verity() · line 13

Quoted 6 lines
common/verity +3 -0
@@ -8,6 +8,10 @@ _require_scratch_verity()
 	_require_scratch
 	_require_command "$FSVERITY_PROG" fsverity
 
+	if [ $FSTYP == "btrfs" ]; then
+		_require_command "$BTRFS_CORRUPT_BLOCK_PROG" btrfs_corrupt_block
+	fi

Does the fsverity function depend on btrfs-corrupt-block command? I think btrfs-corrupt-block is only needed in this specific test, other btrfs fsverity tests may not depend on it. So add this require rule in the test if that's the case.

Quoted 42 lines
common/verity +4 -0
+
 	if ! _scratch_mkfs_verity &>>$seqres.full; then
 		# ext4: need e2fsprogs v1.44.5 or later (but actually v1.45.2+
 		#       is needed for some tests to pass, due to an e2fsck bug)
@@ -147,6 +151,9 @@ _scratch_mkfs_verity()
 	ext4|f2fs)
 		_scratch_mkfs -O verity
 		;;
+	btrfs)
+		_scratch_mkfs
+		;;
 	*)
 		_notrun "No verity support for $FSTYP"
 		;;
tests/btrfs/290 new file +22 -0
new file mode 100755
index 00000000..5aff7648
@@ -0,0 +1,190 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (C) 2021 Facebook, Inc. All Rights Reserved.
+#
+# FS QA Test 290
+#
+# Test btrfs support for fsverity.
+# This test extends the generic fsverity testing by corrupting inline extents,
+# preallocated extents, holes, and the Merkle descriptor in a btrfs-aware way.
+#
+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
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter

tests/btrfs/290 line 32

Quoted 10 lines
tests/btrfs/290 new file +10 -0
+. ./common/verity
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+_supported_fs btrfs
+_require_scratch
+_require_scratch_verity
+
+cleanup()

Better to name it as _cleanup(), though usually we name local functions without the leading underscore, but _cleanup function is a bit special, almost all tests name it with "_" and it's in template in 'check' script. And it's easier to do a global s/_cleanup/cleanup/ if it's consistent with other tests.

tests/btrfs/290 line 39

Quoted 7 lines
tests/btrfs/290 new file +7 -0
+{
+	cd /
+	rm -f $tmp.*
+}
+
+get_ino() {
+	file=$1

Declare local variables with local keyword.

tests/btrfs/290 line 40

Quoted 1 line
tests/btrfs/290 new file +1 -0
+	ls -i $file | awk '{print $1}'

stat -c "%i"

tests/btrfs/290 line 45

Quoted 5 lines
tests/btrfs/290 new file +5 -0
+}
+
+validate() {
+	f=$1
+	sz=$2

sz is not used, but xfs_io below would use $sz

tests/btrfs/290 line 49

Quoted 4 lines
tests/btrfs/290 new file +4 -0
+	# buffered io
+	cat $f > /dev/null
+	# direct io
+	dd if=$f iflag=direct of=/dev/null status=none

$XFS_IO_PROG -dc "pread -q 0 $sz" $f

Direct I/O is used, then we need

_require_odirect

tests/btrfs/290 line 55

Quoted 6 lines
tests/btrfs/290 new file +6 -0
+}
+
+# corrupt the data portion of an inline extent
+corrupt_inline() {
+	f=$SCRATCH_MNT/inl
+	head -c 42 /dev/zero | tr '\0' X > $f

$XFS_IO_PROG -fc "pwrite -q -S 0x58 0 42" $f

And all usages in other functions.

tests/btrfs/290 line 64

Quoted 9 lines
tests/btrfs/290 new file +9 -0
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# inline data starts at disk_bytenr
+	# overwrite the first u64 with random bogus junk
+	$BTRFS_CORRUPT_BLOCK_PROG -i $ino -x 0 -f disk_bytenr $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f

Then all validte calls should pass file size as second param.

tests/btrfs/290 line 70

Quoted 6 lines
tests/btrfs/290 new file +6 -0
+}
+
+# preallocate a file, then corrupt it by changing it to a regular file
+corrupt_prealloc_to_reg() {
+	f=$SCRATCH_MNT/prealloc
+	fallocate -l 4k $f

$XFS_IO_PROG -fc "falloc 0 4k" $f

tests/btrfs/290 line 58

Quoted 3 lines
tests/btrfs/290 new file +3 -0
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT

Not sure why sync is needed in all the tests, we do umount in every test which syncs filesystem as well.

tests/btrfs/290 line 76

Quoted 3 lines
tests/btrfs/290 new file +3 -0
+	_scratch_unmount
+	# set extent type from prealloc (2) to reg (1)
+	$BTRFS_CORRUPT_BLOCK_PROG -i $ino -x 0 -f type -v 1 $SCRATCH_DEV 2>/dev/null >/dev/null

Usually do ">/dev/null 2>&1" to discard stdout & stderr

Quoted 16 lines
tests/btrfs/290 new file +16 -0
+	_scratch_mount
+	validate $f
+}
+
+# corrupt a regular file by changing the type to preallocated
+corrupt_reg_to_prealloc() {
+	f=$SCRATCH_MNT/reg
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# set type from reg (1) to prealloc (2)
+	$BTRFS_CORRUPT_BLOCK_PROG -i $ino -x 0 -f type -v 2 $SCRATCH_DEV 2>/dev/null >/dev/null
+	_scratch_mount
+	validate $f

tests/btrfs/290 line 102

Quoted 10 lines
tests/btrfs/290 new file +10 -0
+}
+
+# corrupt a file by punching a hole
+corrupt_punch_hole() {
+	f=$SCRATCH_MNT/punch
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	# make a new extent in the middle
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	head -c 4k /dev/zero | tr '\0' Y | dd of=$f bs=4k count=1 seek=1 conv=notrunc 2>/dev/null

$XFS_IO_PROG -c "pwrite -q -S 0x59 4k 4k" $f

Quoted 75 lines
tests/btrfs/290 new file +75 -0
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# change disk_bytenr to 0, representing a hole
+	$BTRFS_CORRUPT_BLOCK_PROG -i $ino -x 4096 -f disk_bytenr -v 0 $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f
+}
+
+# plug hole
+corrupt_plug_hole() {
+	f=$SCRATCH_MNT/plug
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	fallocate -p -o 4k -l 4k $f
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# change disk_bytenr to some value, plugging the hole
+	$BTRFS_CORRUPT_BLOCK_PROG -i $ino -x 4096 -f disk_bytenr -v 13639680 $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f
+}
+
+# corrupt the fsverity descriptor item indiscriminately (causes EINVAL)
+corrupt_verity_descriptor() {
+	f=$SCRATCH_MNT/desc
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# key for the descriptor item is <inode, BTRFS_VERITY_DESC_ITEM_KEY, 1>,
+	# 88 is X. So we write 5 Xs to the start of the descriptor
+	$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,36,1 -v 88 -o 0 -b 5 $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f
+}
+
+# specifically target the root hash in the descriptor (causes EIO)
+corrupt_root_hash() {
+	f=$SCRATCH_MNT/roothash
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,36,1 -v 88 -o 16 -b 1 $SCRATCH_DEV >> $seqres.full
+	#$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,36,0 -v 88 -o 120 -b 5 $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f
+}
+
+# corrupt the Merkle tree data itself
+corrupt_merkle_tree() {
+	f=$SCRATCH_MNT/merkle
+	head -c 12k /dev/zero | tr '\0' X > $f
+	ino=$(get_ino $f)
+	_fsv_enable $f
+	$XFS_IO_PROG -c sync $SCRATCH_MNT
+	_scratch_unmount
+	# key for the descriptor item is <inode, BTRFS_VERITY_MERKLE_ITEM_KEY, 0>,
+	# 88 is X. So we write 5 Xs to somewhere in the middle of the first
+	# merkle item
+	$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,37,0 -v 88 -o 100 -b 5 $SCRATCH_DEV > /dev/null
+	_scratch_mount
+	validate $f
+}
+
+# real QA test starts here
+_scratch_mkfs >/dev/null
+_scratch_mount
+
+corrupt_inline
+corrupt_prealloc_to_reg

tests/btrfs/290 line 187

Quoted 10 lines
tests/btrfs/290 new file +10 -0
+corrupt_reg_to_prealloc
+corrupt_punch_hole
+corrupt_plug_hole
+corrupt_verity_descriptor
+corrupt_root_hash
+corrupt_merkle_tree
+
+# we intentionally corrupted, re-mkfs to avoid tripping the corrupted fs error
+_scratch_unmount
+_scratch_mkfs >/dev/null

Use _require_scratch_nocheck, then there's no post-test fsck.

tests/btrfs/290.out line 3

Quoted 12 lines
tests/btrfs/290 new file +3 -0
+
+status=0
+exit
tests/btrfs/290.out new file +3 -0
new file mode 100644
index 00000000..4da61246
@@ -0,0 +1,17 @@
+QA output created by 290
+cat: /mnt/scratch/inl: Input/output error
+dd: error reading '/mnt/scratch/inl': Input/output error

The $SCRATCH_MNT path in error message should be filtered, as people may hanve different SCRATCH_MNT defined. Use _filter_scratch, i.e.

cat $f 2>&1 >/dev/null | _filter_sratch

Thanks,
Eryu

Quoted 25 lines
tests/btrfs/290.out new file +14 -0
+cat: /mnt/scratch/prealloc: Input/output error
+dd: error reading '/mnt/scratch/prealloc': Input/output error
+cat: /mnt/scratch/reg: Input/output error
+dd: error reading '/mnt/scratch/reg': Input/output error
+cat: /mnt/scratch/punch: Input/output error
+dd: error reading '/mnt/scratch/punch': Input/output error
+cat: /mnt/scratch/plug: Input/output error
+dd: error reading '/mnt/scratch/plug': Input/output error
+cat: /mnt/scratch/desc: Invalid argument
+dd: failed to open '/mnt/scratch/desc': Invalid argument
+cat: /mnt/scratch/roothash: Input/output error
+dd: error reading '/mnt/scratch/roothash': Input/output error
+cat: /mnt/scratch/merkle: Input/output error
+dd: error reading '/mnt/scratch/merkle': Input/output error
tests/btrfs/group +1 -0
index 331dd432..13051562 100644
@@ -238,3 +238,4 @@
 233 auto quick subvolume
 234 auto quick compress rw
 235 auto quick send
+290 auto quick verity
Signature · 1 line
2.30.2
#5 lore raw

Re: [PATCH v3 2/3] generic/574: corrupt btrfs merkle tree data

In reply to Boris Burkov
Quoted 26 lines · Boris Burkov

generic/574 has tests for corrupting the merkle tree data stored by the filesystem. Since btrfs uses a different scheme for storing this data, the existing logic for corrupting it doesn't work out of the box. Adapt it to properly corrupt btrfs merkle items.

Note that there is a bit of a kludge here: since btrfs_corrupt_block doesn't handle streaming corruption bytes from stdin (I could change that, but it feels like overkill for this purpose), I just read the first corruption byte and duplicate it for the desired length. That is how the test is using the interface in practice, anyway.

This relies on the following kernel patch for btrfs verity support:
<btrfs-verity-patch>
And the following btrfs-progs patch for btrfs_corrupt_block support:
<btrfs-corrupt-block-patch>

---
 common/verity | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)
common/verity +0 -0
index d2c1ea24..fdd05783 100644

common/verity line 6

Quoted 6 lines
common/verity +1 -2
@@ -3,8 +3,7 @@
 #
 # Functions for setting up and testing fs-verity
 
-_require_scratch_verity()
-{
+_require_scratch_verity() {

No need to change this.

common/verity _fsv_scratch_corrupt_merkle_tree() · line 318

Quoted 9 lines
common/verity +2 -0
 	_require_scratch
 	_require_command "$FSVERITY_PROG" fsverity
 
@@ -315,6 +314,18 @@ _fsv_scratch_corrupt_merkle_tree()
 		(( offset += ($(_get_filesize $file) + 65535) & ~65535 ))
 		_fsv_scratch_corrupt_bytes $file $offset
 		;;
+	btrfs)
+		ino=$(ls -i $file | awk '{print $1}')

stat -c %i $1

And declare local variables with local.

common/verity _fsv_scratch_corrupt_merkle_tree() · line 319

Quoted 1 line
common/verity +1 -0
+		sync

Why a system wide sync is needed here?

common/verity _fsv_scratch_corrupt_merkle_tree() · line 320

Quoted 1 line
common/verity +1 -0
+		cat > $tmp.bytes

I think this cat would just hang there.

common/verity _fsv_scratch_corrupt_merkle_tree() · line 325

Quoted 5 lines
common/verity +5 -0
+		sz=$(_get_filesize $tmp.bytes)
+		read -n 1 byte < $tmp.bytes
+		ascii=$(printf "%d" "'$byte'")
+		_scratch_unmount
+		$BTRFS_CORRUPT_BLOCK_PROG -r 5 -I $ino,37,0 -v $ascii -o $offset -b $sz $SCRATCH_DEV

It'd be better to explain this command in comments.

common/verity _fsv_scratch_corrupt_merkle_tree() · line 326

Quoted 1 line
common/verity +1 -0
+		sync

Again, is this sync really needed?

Thanks,
Eryu

Quoted 7 lines
common/verity +2 -0
+		_scratch_mount
+		;;
 	*)
 		_fail "_fsv_scratch_corrupt_merkle_tree() unimplemented on $FSTYP"
 		;;
Signature · 1 line
2.30.2