Skip to contentSkip to conversation

btrfs-progs: convert: Ensure the data chunks size never exceed device size

PATCH In review Qu Wenruo 5 participants 14 messages Jun 24, 2020 – Feb 17, 2021 Btrfs file system

lore
Share & tools
lore

Qu Wenruo

author
#0 lore raw

PATCH 1/2 btrfs-progs: convert: Ensure the data chunks size never exceed device size

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

---
 convert/main.c | 2 ++
 1 file changed, 2 insertions(+)
convert/main.c +2 -0
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
 			cur_off = cache->start;
 		cur_len = max(cache->start + cache->size - cur_off,
 			      min_stripe_size);
+		/* data chunks should never exceed device boundary */
+		cur_len = min(cctx->total_bytes - cur_off, cur_len);
 		ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
 		if (ret < 0)
 			goto out;
Signature · 1 line
2.27.0
author
#1 lore raw

PATCH 2/2 btrfs-progs: tests/convert: Add test case to make sure we won't allocate dev extents beyond device boundary

Add a test case to check if the converted fs has device extent beyond boundary.

The disk layout of source ext4 fs needs some extents to make them allocated at the very end of the fs.
The script is from the original reporter.

Also, since the existing convert tests always uses 512M as device size, which is not suitable for this test case, make it to grab the existing device size to co-operate with this test case.

---
 tests/common.convert                         | 14 ++++++++-
 tests/convert-tests/017-fs-near-full/test.sh | 30 ++++++++++++++++++++
 2 files changed, 43 insertions(+), 1 deletion(-)
 create mode 100755 tests/convert-tests/017-fs-near-full/test.sh
tests/common.convert +13 -1
index f24ceb0d6a64..0c918387758d 100644
@@ -53,6 +53,16 @@ convert_test_preamble() {
 	echo "creating test image with: $@" >> "$RESULTS"
 }
 
+get_test_file_size() {
+	local path="$1"
+	local ret
+
+	ret=$(ls -l "$path" | cut -f5 -d\ )
+	if [ -z $ret ]; then
+		ret=512M
+	fi
+	echo $ret
+}
 #  prepare TEST_DEV before conversion, create filesystem and mount it, image
 #  size is 512MB
 #  $1: type of the filesystem
@@ -61,14 +71,16 @@ convert_test_prep_fs() {
 	local fstype
 	local force
 	local mountopts
+	local oldsize
 
 	fstype="$1"
 	shift
+	oldsize=$(get_test_file_size "$TEST_DEV")
 	# TEST_DEV not removed as the file might have special permissions, eg.
 	# when test image is on NFS and would not be writable for root
 	run_check truncate -s 0 "$TEST_DEV"
 	# 256MB is the smallest acceptable btrfs image.
-	run_check truncate -s 512M "$TEST_DEV"
+	run_check truncate -s $oldsize "$TEST_DEV"
 	force=
 	mountopts=
 	case "$fstype" in
tests/convert-tests/017-fs-near-full/test.sh new file +30 -0
new file mode 100755
index 000000000000..9459729ee87f
@@ -0,0 +1,30 @@
+#!/bin/bash
+# Check if btrfs-convert creates fs with dev extents boundary device boundary
+
+source "$TEST_TOP/common"
+source "$TEST_TOP/common.convert"
+
+setup_root_helper
+prepare_test_dev 1G
+check_prereq btrfs-convert
+check_global_prereq mke2fs
+check_global_prereq fallocate
+
+convert_test_prep_fs ext4 mke2fs -t ext4 -b 4096
+
+# Use up 800MiB first
+for i in $(seq 1 4); do
+	run_check $SUDO_HELPER fallocate -l 200M "$TEST_MNT/file$i"
+done
+
+# Then add 5MiB for above files. These 5 MiB will be allocated near the very
+# end of the fs, to confuse btrfs-convert
+for i in $(seq 1 4); do
+	run_check $SUDO_HELPER fallocate -l 205M "$TEST_MNT/file$i"
+done
+
+run_check_umount_test_dev
+
+# convert_test_do_convert() will call btrfs check, which should expose any
+# invalid inline extent with too large size
+convert_test_do_convert
Signature · 1 line
2.27.0
#2 lore raw

Re: [PATCH 2/2] btrfs-progs: tests/convert: Add test case to make sure we won't allocate dev extents beyond device boundary

Quoted 24 lines · Qu Wenruo

Add a test case to check if the converted fs has device extent beyond boundary.

The disk layout of source ext4 fs needs some extents to make them allocated at the very end of the fs.
The script is from the original reporter.

Also, since the existing convert tests always uses 512M as device size, which is not suitable for this test case, make it to grab the existing device size to co-operate with this test case.

---
 tests/common.convert                         | 14 ++++++++-
 tests/convert-tests/017-fs-near-full/test.sh | 30 ++++++++++++++++++++
 2 files changed, 43 insertions(+), 1 deletion(-)
 create mode 100755 tests/convert-tests/017-fs-near-full/test.sh
tests/common.convert +0 -0
index f24ceb0d6a64..0c918387758d 100644

tests/common.convert convert_test_preamble() { · line 60

Quoted 8 lines
tests/common.convert +5 -0
@@ -53,6 +53,16 @@ convert_test_preamble() {
 	echo "creating test image with: $@" >> "$RESULTS"
 }
 
+get_test_file_size() {
+	local path="$1"
+	local ret
+
+	ret=$(ls -l "$path" | cut -f5 -d\ )

Parsing ls output is quite fragile, and we have 'stat --format=%s'.

#3 lore raw

Re: [PATCH 2/2] btrfs-progs: tests/convert: Add test case to make sure we won't allocate dev extents beyond device boundary

In reply to Qu Wenruo
Quoted 41 lines · Qu Wenruo

Add a test case to check if the converted fs has device extent beyond boundary.

The disk layout of source ext4 fs needs some extents to make them allocated at the very end of the fs.
The script is from the original reporter.

Also, since the existing convert tests always uses 512M as device size, which is not suitable for this test case, make it to grab the existing device size to co-operate with this test case.

---
 tests/common.convert                         | 14 ++++++++-
 tests/convert-tests/017-fs-near-full/test.sh | 30 ++++++++++++++++++++
 2 files changed, 43 insertions(+), 1 deletion(-)
 create mode 100755 tests/convert-tests/017-fs-near-full/test.sh
tests/common.convert +10 -0
index f24ceb0d6a64..0c918387758d 100644
@@ -53,6 +53,16 @@ convert_test_preamble() {
 	echo "creating test image with: $@" >> "$RESULTS"
 }
 
+get_test_file_size() {
+	local path="$1"
+	local ret
+
+	ret=$(ls -l "$path" | cut -f5 -d\ )
+	if [ -z $ret ]; then
+		ret=512M
+	fi
+	echo $ret
+}
 #  prepare TEST_DEV before conversion, create filesystem and mount it, image
 #  size is 512MB
 #  $1: type of the filesystem

tests/common.convert convert_test_prep_fs() { · line 78

Quoted 8 lines
tests/common.convert +2 -0
@@ -61,14 +71,16 @@ convert_test_prep_fs() {
 	local fstype
 	local force
 	local mountopts
+	local oldsize
 
 	fstype="$1"
 	shift
+	oldsize=$(get_test_file_size "$TEST_DEV")

The helper should be here, with a comment why the return size needs to be 512M in some cases.

#4 lore raw
In reply to Qu Wenruo
Quoted 50 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

With the minor tweaks, added to devel, thanks.

#5 lore raw
In reply to Qu Wenruo

Before this patch the converted ext4 was failing to mount due to 'beyond device boundary' error.

After this patch

# ./btrfs-convert /dev/sda7
create btrfs filesystem:
	blocksize: 4096
	nodesize:  16384
	features:  extref, skinny-metadata (default)
	checksum:  crc32c
creating ext2 image file
ERROR: data bytenr 1644167168 is covered by non-data block group 1644167168 flags 0x4
ERROR: failed to create ext2_saved/image: -22
WARNING: an error occurred during conversion, filesystem is partially created but not finalized and not mountable

Any idea?

Quoted 74 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936

    referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

---
  convert/main.c | 2 ++
  1 file changed, 2 insertions(+)
convert/main.c +2 -0
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
  			cur_off = cache->start;
  		cur_len = max(cache->start + cache->size - cur_off,
  			      min_stripe_size);
+		/* data chunks should never exceed device boundary */
+		cur_len = min(cctx->total_bytes - cur_off, cur_len);
  		ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
  		if (ret < 0)
  			goto out;
#6 lore raw
Quoted 3 lines · Anand Jain

Before this patch the converted ext4 was failing to mount due to 'beyond device boundary' error.

e2image -Q please.

Thanks,
Qu

Quoted 98 lines

After this patch

# ./btrfs-convert /dev/sda7
create btrfs filesystem:
    blocksize: 4096
    nodesize:  16384
    features:  extref, skinny-metadata (default)
    checksum:  crc32c
creating ext2 image file
ERROR: data bytenr 1644167168 is covered by non-data block group
1644167168 flags 0x4
ERROR: failed to create ext2_saved/image: -22
WARNING: an error occurred during conversion, filesystem is partially
created but not finalized and not mountable

Any idea?

Quoted 76 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

   fallocate -l 1G test.img
   mkfs.ext4 test.img
   mount test.img $mnt
   fallocate -l 200m $mnt/file1
   fallocate -l 200m $mnt/file2
   fallocate -l 200m $mnt/file3
   fallocate -l 200m $mnt/file4
   fallocate -l 205m $mnt/file1
   fallocate -l 205m $mnt/file2
   fallocate -l 205m $mnt/file3
   fallocate -l 205m $mnt/file4
   umount $mnt
   btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
   pening filesystem to check...
   Checking filesystem on test.img
   UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
   [1/7] checking root items
   [2/7] checking extents
   ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
   ERROR: errors found in extent allocation tree or chunk allocation [3/7] checking free space cache
   [4/7] checking fs roots
   [5/7] checking only csums items (without verifying data) [6/7] checking root refs
   [7/7] checking quota groups skipped (not enabled on this FS) found 913960960 bytes used, error(s) found
   total csum bytes: 891500
   total tree bytes: 1064960
   total fs tree bytes: 49152
   total extent tree bytes: 16384
   btree space waste bytes: 144885
   file data blocks allocated: 2129063936
    referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

---
  convert/main.c | 2 ++

  1 file changed, 2 insertions(+)
convert/main.c +2 -0
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct
 btrfs_convert_context *cctx)
               cur_off = cache->start;
           cur_len = max(cache->start + cache->size - cur_off,
                     min_stripe_size);
+        /* data chunks should never exceed device boundary */
+        cur_len = min(cctx->total_bytes - cur_off, cur_len);
           ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
           if (ret < 0)

              goto out;

#7 lore raw
In reply to Qu Wenruo
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

    --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
    +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
    @@ -1,2 +1,3 @@
     QA output created by 136
    -Silence is golden
    +btrfs-convert failed
    +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
    ...
    (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Thanks.

Quoted 20 lines
---
 convert/main.c | 2 ++
 1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                        cur_off = cache->start;
                cur_len = max(cache->start + cache->size - cur_off,
                              min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                if (ret < 0)
                        goto out;
--

2.27.0

Signature · 3 lines
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”
#8 lore raw
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936

    referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

     --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
     +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
     @@ -1,2 +1,3 @@
      QA output created by 136
     -Silence is golden
     +btrfs-convert failed
     +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
     ...
     (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

 $ which btrfs
 /usr/bin/btrfs
 $ btrfs --version
 btrfs-progs v5.10.1
 $ sudo ./check btrfs/136
 FSTYP         -- btrfs
 PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
 MKFS_OPTIONS  -- /dev/mapper/test-scratch1
 MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch

btrfs/136 6s ... 10s
Ran: btrfs/136
Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
  convert/main.c | 2 ++
  1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                         cur_off = cache->start;
                 cur_len = max(cache->start + cache->size - cur_off,
                               min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                 ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                 if (ret < 0)
                         goto out;
--

2.27.0

#9 lore raw
Quoted 108 lines · Qu Wenruo
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936

    referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

     --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
     +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
     @@ -1,2 +1,3 @@
      QA output created by 136
     -Silence is golden
     +btrfs-convert failed
     +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
     ...
     (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

  $ which btrfs
  /usr/bin/btrfs
  $ btrfs --version
  btrfs-progs v5.10.1
  $ sudo ./check btrfs/136
  FSTYP         -- btrfs
  PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
  MKFS_OPTIONS  -- /dev/mapper/test-scratch1
  MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch
Quoted 5 lines

btrfs/136 6s ... 10s
Ran: btrfs/136
Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Sure, here it is (also at https://pastebin.com/XhEX2dju):

root 10:06:39 /home/fdmanana/git/hub/xfstests (master)> cat /home/fdmanana/git/hub/xfstests/results//btrfs/136.full mke2fs 1.45.6 (20-Mar-2020)
Discarding device blocks: done
Creating filesystem with 5242880 4k blocks and 1310720 inodes Filesystem UUID: 9519afc8-8ea0-42da-8ac5-3b4c20469dd1 Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000

Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext3 tune2fs 1.45.6 (20-Mar-2020)
e2fsck 1.45.6 (20-Mar-2020)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information

/dev/sdc: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sdc: 235/1310720 files (22.1% non-contiguous), 129757/5242880 blocks Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext4
ERROR: missing data block for bytenr 1048576
ERROR: failed to create ext2_saved/image: -2
WARNING: an error occurred during conversion, filesystem is partially created but not finalized and not mountable
create btrfs filesystem:
blocksize: 4096
nodesize: 16384
features: extref, skinny-metadata (default)
checksum: crc32c
creating ext2 image file
btrfs-convert failed
root 10:14:18 /home/fdmanana/git/hub/xfstests (master)>

Thanks

Quoted 28 lines

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
  convert/main.c | 2 ++
  1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                         cur_off = cache->start;
                 cur_len = max(cache->start + cache->size - cur_off,
                               min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                 ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                 if (ret < 0)
                         goto out;
--

2.27.0

Signature · 3 lines
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”
#10 lore raw
Quoted 143 lines · Filipe Manana
Quoted 108 lines · Qu Wenruo
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

      --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
      +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
      @@ -1,2 +1,3 @@
       QA output created by 136
      -Silence is golden
      +btrfs-convert failed
      +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
      ...
      (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

   $ which btrfs
   /usr/bin/btrfs
   $ btrfs --version
   btrfs-progs v5.10.1
   $ sudo ./check btrfs/136
   FSTYP         -- btrfs
   PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
   MKFS_OPTIONS  -- /dev/mapper/test-scratch1
   MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch
Quoted 5 lines
   btrfs/136 6s ...  10s
   Ran: btrfs/136
   Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Sure, here it is (also at https://pastebin.com/XhEX2dju):

root 10:06:39 /home/fdmanana/git/hub/xfstests (master)> cat /home/fdmanana/git/hub/xfstests/results//btrfs/136.full mke2fs 1.45.6 (20-Mar-2020)
Discarding device blocks: done
Creating filesystem with 5242880 4k blocks and 1310720 inodes Filesystem UUID: 9519afc8-8ea0-42da-8ac5-3b4c20469dd1 Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000

Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext3 tune2fs 1.45.6 (20-Mar-2020)
e2fsck 1.45.6 (20-Mar-2020)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information

Quoted 7 lines

/dev/sdc: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sdc: 235/1310720 files (22.1% non-contiguous), 129757/5242880 blocks Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext4
ERROR: missing data block for bytenr 1048576
ERROR: failed to create ext2_saved/image: -2
WARNING: an error occurred during conversion, filesystem is partially

The bytenr is 1M, which looks related to the super block relocation code.

Thus it maybe disk layout related.

Your disk used here is 20G, although I tried the same size disk, it still passes.

If you could reproduce it reliably (as you can do the bisect), mind to take a e2image -Q dump?

Thanks,
Qu

Quoted 45 lines

created but not finalized and not mountable create btrfs filesystem:
blocksize: 4096
nodesize: 16384
features: extref, skinny-metadata (default)
checksum: crc32c
creating ext2 image file
btrfs-convert failed
root 10:14:18 /home/fdmanana/git/hub/xfstests (master)>

Thanks

Quoted 28 lines

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
   convert/main.c | 2 ++
   1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                          cur_off = cache->start;
                  cur_len = max(cache->start + cache->size - cur_off,
                                min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                  ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                  if (ret < 0)
                          goto out;
--

2.27.0

#11 lore raw
Quoted 157 lines · Qu Wenruo
Quoted 143 lines · Filipe Manana
Quoted 108 lines · Qu Wenruo
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

      --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
      +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
      @@ -1,2 +1,3 @@
       QA output created by 136
      -Silence is golden
      +btrfs-convert failed
      +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
      ...
      (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

   $ which btrfs
   /usr/bin/btrfs
   $ btrfs --version
   btrfs-progs v5.10.1
   $ sudo ./check btrfs/136
   FSTYP         -- btrfs
   PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
   MKFS_OPTIONS  -- /dev/mapper/test-scratch1
   MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch
Quoted 5 lines
   btrfs/136 6s ...  10s
   Ran: btrfs/136
   Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Sure, here it is (also at https://pastebin.com/XhEX2dju):

root 10:06:39 /home/fdmanana/git/hub/xfstests (master)> cat /home/fdmanana/git/hub/xfstests/results//btrfs/136.full mke2fs 1.45.6 (20-Mar-2020)
Discarding device blocks: done
Creating filesystem with 5242880 4k blocks and 1310720 inodes Filesystem UUID: 9519afc8-8ea0-42da-8ac5-3b4c20469dd1 Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000

Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext3 tune2fs 1.45.6 (20-Mar-2020)
e2fsck 1.45.6 (20-Mar-2020)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information

Quoted 7 lines

/dev/sdc: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sdc: 235/1310720 files (22.1% non-contiguous), 129757/5242880 blocks Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext4
ERROR: missing data block for bytenr 1048576
ERROR: failed to create ext2_saved/image: -2
WARNING: an error occurred during conversion, filesystem is partially

The bytenr is 1M, which looks related to the super block relocation code.

Quoted 7 lines

Thus it maybe disk layout related.

Your disk used here is 20G, although I tried the same size disk, it still passes.

If you could reproduce it reliably (as you can do the bisect), mind to take a e2image -Q dump?

Yep, I can reproduce it reliably, always fails with btrfs-progs v5.7+.

Image attached. I took it after the test converts the ext3 fs to ext4 and right before the test calls btrfs-convert.

Thanks.

Quoted 49 lines

Thanks,
Qu

Quoted 45 lines

created but not finalized and not mountable create btrfs filesystem:
blocksize: 4096
nodesize: 16384
features: extref, skinny-metadata (default)
checksum: crc32c
creating ext2 image file
btrfs-convert failed
root 10:14:18 /home/fdmanana/git/hub/xfstests (master)>

Thanks

Quoted 28 lines

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
   convert/main.c | 2 ++
   1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                          cur_off = cache->start;
                  cur_len = max(cache->start + cache->size - cur_off,
                                min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                  ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                  if (ret < 0)
                          goto out;
--

2.27.0

Signature · 3 lines
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”
#12 lore raw
Quoted 163 lines · Filipe Manana
Quoted 162 lines · Qu Wenruo
Quoted 143 lines · Filipe Manana
Quoted 108 lines · Qu Wenruo
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

       --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
       +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
       @@ -1,2 +1,3 @@
        QA output created by 136
       -Silence is golden
       +btrfs-convert failed
       +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
       ...
       (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

    $ which btrfs
    /usr/bin/btrfs
    $ btrfs --version
    btrfs-progs v5.10.1
    $ sudo ./check btrfs/136
    FSTYP         -- btrfs
    PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
    MKFS_OPTIONS  -- /dev/mapper/test-scratch1
    MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch
Quoted 5 lines
    btrfs/136 6s ...  10s
    Ran: btrfs/136
    Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Sure, here it is (also at https://pastebin.com/XhEX2dju):

root 10:06:39 /home/fdmanana/git/hub/xfstests (master)> cat /home/fdmanana/git/hub/xfstests/results//btrfs/136.full mke2fs 1.45.6 (20-Mar-2020)
Discarding device blocks: done
Creating filesystem with 5242880 4k blocks and 1310720 inodes Filesystem UUID: 9519afc8-8ea0-42da-8ac5-3b4c20469dd1 Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000

Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext3 tune2fs 1.45.6 (20-Mar-2020)
e2fsck 1.45.6 (20-Mar-2020)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information

Quoted 7 lines

/dev/sdc: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sdc: 235/1310720 files (22.1% non-contiguous), 129757/5242880 blocks Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext4
ERROR: missing data block for bytenr 1048576
ERROR: failed to create ext2_saved/image: -2
WARNING: an error occurred during conversion, filesystem is partially

The bytenr is 1M, which looks related to the super block relocation code.

Thus it maybe disk layout related.

Your disk used here is 20G, although I tried the same size disk, it still passes.

Quoted 7 lines
Quoted 2 lines

If you could reproduce it reliably (as you can do the bisect), mind to take a e2image -Q dump?

Yep, I can reproduce it reliably, always fails with btrfs-progs v5.7+.

Image attached. I took it after the test converts the ext3 fs to ext4 and right before the test calls btrfs-convert.

Something doesn't go correct.

v5.10.1 btrfs-convert still passes on the image in my test env.

  $ ./btrfs-convert  /dev/vdc
  create btrfs filesystem:
          blocksize: 4096
          nodesize:  16384
          features:  extref, skinny-metadata (default)
          checksum:  crc32c
  free space report:
          total:     21474836480
          free:      15690694656 (73.07%)
  creating ext2 image file
  creating btrfs metadata
  copy inodes [o] [         4/       527]
  conversion complete
  $ ./btrfs --version
  btrfs-progs v5.10.1

I have no idea what's going on now...

I'm afraid you have to dig the bug by yourself, as I really can't reproduce the bug, even with the image dump...

Thanks,
Qu

Quoted 55 lines

Thanks.

Quoted 49 lines

Thanks,
Qu

Quoted 45 lines

created but not finalized and not mountable create btrfs filesystem:
blocksize: 4096
nodesize: 16384
features: extref, skinny-metadata (default)
checksum: crc32c
creating ext2 image file
btrfs-convert failed
root 10:14:18 /home/fdmanana/git/hub/xfstests (master)>

Thanks

Quoted 28 lines

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
    convert/main.c | 2 ++
    1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                           cur_off = cache->start;
                   cur_len = max(cache->start + cache->size - cur_off,
                                 min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                   ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                   if (ret < 0)
                           goto out;
--

2.27.0

#13 lore raw
Quoted 196 lines · Qu Wenruo
Quoted 163 lines · Filipe Manana
Quoted 162 lines · Qu Wenruo
Quoted 143 lines · Filipe Manana
Quoted 108 lines · Qu Wenruo
Quoted 81 lines · Filipe Manana
Quoted 51 lines · Qu Wenruo

[BUG]
The following script could lead to corrupted btrfs fs after btrfs-convert:

fallocate -l 1G test.img
mkfs.ext4 test.img
mount test.img $mnt
fallocate -l 200m $mnt/file1
fallocate -l 200m $mnt/file2
fallocate -l 200m $mnt/file3
fallocate -l 200m $mnt/file4
fallocate -l 205m $mnt/file1
fallocate -l 205m $mnt/file2
fallocate -l 205m $mnt/file3
fallocate -l 205m $mnt/file4
umount $mnt
btrfs-convert test.img

The result btrfs will have a device extent beyond its boundary:
pening filesystem to check...
Checking filesystem on test.img
UUID: bbcd7399-fd5b-41a7-81ae-d48bc6935e43
[1/7] checking root items
[2/7] checking extents
ERROR: dev extent devid 1 physical offset 993198080 len 85786624 is beyond device boundary 1073741824
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 913960960 bytes used, error(s) found
total csum bytes: 891500
total tree bytes: 1064960
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 144885
file data blocks allocated: 2129063936
referenced 1772728320

[CAUSE]
Btrfs-convert first collect all used blocks in the original fs, then slightly enlarge the used blocks range as new btrfs data chunks.

However the enlarge part has a problem, that it doesn't take the device boundary into consideration.

Thus it caused device extents and data chunks to go beyond device boundary.

Quoted 6 lines

[FIX]
Just to extra check before inserting data chunks into btrfs_convert_context::data_chunk.

So, having upgraded a test box from btrfs-progs v5.6.1 to v5.10.1, I now have btrfs/136 (fstests) failing:

$ ./check btrfs/136
FSTYP         -- btrfs
PLATFORM      -- Linux/x86_64 debian8 5.11.0-rc7-btrfs-next-81 #1 SMP
PREEMPT Tue Feb 16 12:29:07 WET 2021
MKFS_OPTIONS  -- /dev/sdc
MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1

btrfs/136 7s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad)

       --- tests/btrfs/136.out 2020-06-10 19:29:03.818519162 +0100
       +++ /home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad
2021-02-16 14:31:30.669559188 +0000
       @@ -1,2 +1,3 @@
        QA output created by 136
       -Silence is golden
       +btrfs-convert failed
       +(see /home/fdmanana/git/hub/xfstests/results//btrfs/136.full for details)
       ...
       (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/136.out
Quoted 8 lines

/home/fdmanana/git/hub/xfstests/results//btrfs/136.out.bad' to see the entire diff)
Ran: btrfs/136
Failures: btrfs/136
Failed 1 of 1 tests

A bisect pointed to this patch.
Did you get this failure on your test box as well?

Nope.

Just tested with btrfs-progs v5.10.1, it passes:

    $ which btrfs
    /usr/bin/btrfs
    $ btrfs --version
    btrfs-progs v5.10.1
    $ sudo ./check btrfs/136
    FSTYP         -- btrfs
    PLATFORM      -- Linux/x86_64 btrfs-desktop-vm 5.11.0-rc4-custom+ #4
SMP PREEMPT Mon Jan 25 18:35:22 CST 2021
    MKFS_OPTIONS  -- /dev/mapper/test-scratch1
    MOUNT_OPTIONS -- /dev/mapper/test-scratch1 /mnt/scratch
Quoted 5 lines
    btrfs/136 6s ...  10s
    Ran: btrfs/136
    Passed all 1 tests

Would you mind to provide the 136.full to help debugging the failure?

Sure, here it is (also at https://pastebin.com/XhEX2dju):

root 10:06:39 /home/fdmanana/git/hub/xfstests (master)> cat /home/fdmanana/git/hub/xfstests/results//btrfs/136.full mke2fs 1.45.6 (20-Mar-2020)
Discarding device blocks: done
Creating filesystem with 5242880 4k blocks and 1310720 inodes Filesystem UUID: 9519afc8-8ea0-42da-8ac5-3b4c20469dd1 Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000

Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext3 tune2fs 1.45.6 (20-Mar-2020)
e2fsck 1.45.6 (20-Mar-2020)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information

Quoted 7 lines

/dev/sdc: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sdc: 235/1310720 files (22.1% non-contiguous), 129757/5242880 blocks Run fsstress -p 20 -n 100 -d /home/fdmanana/btrfs-tests/scratch_1/ext3_ext4_data/ext4
ERROR: missing data block for bytenr 1048576
ERROR: failed to create ext2_saved/image: -2
WARNING: an error occurred during conversion, filesystem is partially

The bytenr is 1M, which looks related to the super block relocation code.

Thus it maybe disk layout related.

Your disk used here is 20G, although I tried the same size disk, it still passes.

Quoted 7 lines
Quoted 2 lines

If you could reproduce it reliably (as you can do the bisect), mind to take a e2image -Q dump?

Yep, I can reproduce it reliably, always fails with btrfs-progs v5.7+.

Image attached. I took it after the test converts the ext3 fs to ext4 and right before the test calls btrfs-convert.

Something doesn't go correct.

v5.10.1 btrfs-convert still passes on the image in my test env.

   $ ./btrfs-convert  /dev/vdc
   create btrfs filesystem:
           blocksize: 4096
           nodesize:  16384
           features:  extref, skinny-metadata (default)
           checksum:  crc32c
   free space report:
           total:     21474836480
           free:      15690694656 (73.07%)
   creating ext2 image file
   creating btrfs metadata
   copy inodes [o] [         4/       527]
   conversion complete
   $ ./btrfs --version
   btrfs-progs v5.10.1
Quoted 4 lines

I have no idea what's going on now...

I'm afraid you have to dig the bug by yourself, as I really can't reproduce the bug, even with the image dump...

I'll add it to my list of things to check.

Thanks for looking into it!

Quoted 59 lines

Thanks,
Qu

Quoted 55 lines

Thanks.

Quoted 49 lines

Thanks,
Qu

Quoted 45 lines

created but not finalized and not mountable create btrfs filesystem:
blocksize: 4096
nodesize: 16384
features: extref, skinny-metadata (default)
checksum: crc32c
creating ext2 image file
btrfs-convert failed
root 10:14:18 /home/fdmanana/git/hub/xfstests (master)>

Thanks

Quoted 28 lines

Thanks,
Qu

Quoted 25 lines

Thanks.

Quoted 20 lines
---
    convert/main.c | 2 ++
    1 file changed, 2 insertions(+)
convert/main.c +2 -1
index c86ddd988c63..7709e9a6c085 100644
@@ -669,6 +669,8 @@ static int calculate_available_space(struct btrfs_convert_context *cctx)
                           cur_off = cache->start;
                   cur_len = max(cache->start + cache->size - cur_off,
                                 min_stripe_size);
+               /* data chunks should never exceed device boundary */
+               cur_len = min(cctx->total_bytes - cur_off, cur_len);
                   ret = add_merge_cache_extent(data_chunks, cur_off, cur_len);
                   if (ret < 0)
                           goto out;
--

2.27.0

Signature · 3 lines
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”