ext4: fix corruption during on-line resize

  We observed a corruption during on-line resize of a file system that is
larger than 16 TiB with 4k block size. With having more then 2^32 blocks
resize_inode is turned off by default by mke2fs. The issue can be
reproduced on a smaller file system for convenience by explicitly
turning off resize_inode. An on-line resize across an 8 GiB boundary (the
size of a meta block group in this setup) then leads to a corruption:

  dev=/dev/<some_dev> # should be >= 16 GiB
  mkdir -p /corruption
  /sbin/mke2fs -t ext4 -b 4096 -O ^resize_inode $dev $((2 * 2**21 - 2**15))
  mount -t ext4 $dev /corruption

  dd if=/dev/zero bs=4096 of=/corruption/test count=$((2*2**21 - 4*2**15))
  sha1sum /corruption/test
  # 79d2658b39dcfd77274e435b0934028adafaab11  /corruption/test

  /sbin/resize2fs $dev $((2*2**21))
  # drop page cache to force reload the block from disk
  echo 1 > /proc/sys/vm/drop_caches

  sha1sum /corruption/test
  # 3c2abc63cbf1a94c9e6977e0fbd72cd832c4d5c3  /corruption/test

2^21 = 2^15*2^6 equals 8 GiB whereof 2^15 is the number of blocks per
block group and 2^6 are the number of block groups that make a meta
block group.

The last checksum might be different depending on how the file is laid
out across the physical blocks. The actual corruption occurs at physical
block 63*2^15 = 2064384 which would be the location of the backup of the
meta block group's block descriptor. During the on-line resize the file
system will be converted to meta_bg starting at s_first_meta_bg which is
2 in the example - meaning all block groups after 16 GiB. However, in
ext4_flex_group_add we might add block groups that are not part of the
first meta block group yet. In the reproducer we achieved this by
substracting the size of a whole block group from the point where the
meta block group would start. This must be considered when updating the
backup block group descriptors to follow the non-meta_bg layout. The fix
is to add a test whether the group to add is already part of the meta
block group or not.

Signed-off-by: Maximilian Heyne <mheyne@amazon.de>
---
 fs/ext4/resize.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
  

On Tue, 2024-02-06 at 12:05 +0000, Maximilian Heyne wrote:
> We observed a corruption during on-line resize of a file system that is
> larger than 16 TiB with 4k block size. With having more then 2^32 blocks
> resize_inode is turned off by default by mke2fs. The issue can be
> reproduced on a smaller file system for convenience by explicitly
> turning off resize_inode. An on-line resize across an 8 GiB boundary (the
> size of a meta block group in this setup) then leads to a corruption:
> 
>   dev=/dev/<some_dev> # should be >= 16 GiB
>   mkdir -p /corruption
>   /sbin/mke2fs -t ext4 -b 4096 -O ^resize_inode $dev $((2 * 2**21 - 2**15))
>   mount -t ext4 $dev /corruption
> 
>   dd if=/dev/zero bs=4096 of=/corruption/test count=$((2*2**21 - 4*2**15))
>   sha1sum /corruption/test
>   # 79d2658b39dcfd77274e435b0934028adafaab11  /corruption/test
> 
>   /sbin/resize2fs $dev $((2*2**21))
>   # drop page cache to force reload the block from disk
>   echo 1 > /proc/sys/vm/drop_caches
> 
>   sha1sum /corruption/test
>   # 3c2abc63cbf1a94c9e6977e0fbd72cd832c4d5c3  /corruption/test
> 
> 2^21 = 2^15*2^6 equals 8 GiB whereof 2^15 is the number of blocks per
> block group and 2^6 are the number of block groups that make a meta
> block group.
> 
> The last checksum might be different depending on how the file is laid
> out across the physical blocks. The actual corruption occurs at physical
> block 63*2^15 = 2064384 which would be the location of the backup of the
> meta block group's block descriptor. During the on-line resize the file
> system will be converted to meta_bg starting at s_first_meta_bg which is
> 2 in the example - meaning all block groups after 16 GiB. However, in
> ext4_flex_group_add we might add block groups that are not part of the
> first meta block group yet. In the reproducer we achieved this by
> substracting the size of a whole block group from the point where the
> meta block group would start. This must be considered when updating the
> backup block group descriptors to follow the non-meta_bg layout. The fix
> is to add a test whether the group to add is already part of the meta
> block group or not.
> 
> Signed-off-by: Maximilian Heyne <mheyne@amazon.de>

Perhaps also:

Fixes: 01f795f9e0d67 ("ext4: add online resizing support for meta_bg and 64-bit file systems")
Cc: stable@vger.kernel.org 

> ---
>  fs/ext4/resize.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
> index 4d4a5a32e310..3433e83bb7a3 100644
> --- a/fs/ext4/resize.c
> +++ b/fs/ext4/resize.c
> @@ -1602,7 +1602,9 @@ static int ext4_flex_group_add(struct super_block *sb,
>                 int gdb_num = group / EXT4_DESC_PER_BLOCK(sb);
>                 int gdb_num_end = ((group + flex_gd->count - 1) /
>                                    EXT4_DESC_PER_BLOCK(sb));
> -               int meta_bg = ext4_has_feature_meta_bg(sb);
> +               unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
> +               int meta_bg = ext4_has_feature_meta_bg(sb) &&
> +                             metagroup >= le32_to_cpu(es->s_first_meta_bg);
>                 sector_t padding_blocks = meta_bg ? 0 : sbi->s_sbh->b_blocknr -
>                                          ext4_group_first_block_no(sb, 0);
>
  

On Tue, 2024-02-06 at 12:05 +0000, Maximilian Heyne wrote:
> We observed a corruption during on-line resize of a file system that is
> larger than 16 TiB with 4k block size. With having more then 2^32 blocks
> resize_inode is turned off by default by mke2fs. The issue can be
> reproduced on a smaller file system for convenience by explicitly
> turning off resize_inode. An on-line resize across an 8 GiB boundary (the
> size of a meta block group in this setup) then leads to a corruption:
>
>   dev=/dev/<some_dev> # should be >= 16 GiB
>   mkdir -p /corruption
>   /sbin/mke2fs -t ext4 -b 4096 -O ^resize_inode $dev $((2 * 2**21 - 2**15))
>   mount -t ext4 $dev /corruption
>
>   dd if=/dev/zero bs=4096 of=/corruption/test count=$((2*2**21 - 4*2**15))
Tip: Use fallocate for faster allocations.

>   sha1sum /corruption/test
>   # 79d2658b39dcfd77274e435b0934028adafaab11  /corruption/test
>
>   /sbin/resize2fs $dev $((2*2**21))
>   # drop page cache to force reload the block from disk
>   echo 1 > /proc/sys/vm/drop_caches
>
>   sha1sum /corruption/test
>   # 3c2abc63cbf1a94c9e6977e0fbd72cd832c4d5c3  /corruption/test
>
> 2^21 = 2^15*2^6 equals 8 GiB whereof 2^15 is the number of blocks per
> block group and 2^6 are the number of block groups that make a meta
> block group.
>
> The last checksum might be different depending on how the file is laid
> out across the physical blocks. The actual corruption occurs at physical
> block 63*2^15 = 2064384 which would be the location of the backup of the
> meta block group's block descriptor. During the on-line resize the file
> system will be converted to meta_bg starting at s_first_meta_bg which is
> 2 in the example - meaning all block groups after 16 GiB. However, in
> ext4_flex_group_add we might add block groups that are not part of the
> first meta block group yet. In the reproducer we achieved this by
> substracting the size of a whole block group from the point where the
> meta block group would start. This must be considered when updating the
> backup block group descriptors to follow the non-meta_bg layout. The fix
> is to add a test whether the group to add is already part of the meta
> block group or not.
>
> Signed-off-by: Maximilian Heyne <mheyne@amazon.de>
> ---
>  fs/ext4/resize.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
> index 4d4a5a32e310..3433e83bb7a3 100644
> --- a/fs/ext4/resize.c
> +++ b/fs/ext4/resize.c
> @@ -1602,7 +1602,9 @@ static int ext4_flex_group_add(struct super_block *sb,
>                 int gdb_num = group / EXT4_DESC_PER_BLOCK(sb);
>                 int gdb_num_end = ((group + flex_gd->count - 1) /
>                                    EXT4_DESC_PER_BLOCK(sb));
> -               int meta_bg = ext4_has_feature_meta_bg(sb);
> +               unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
> +               int meta_bg = ext4_has_feature_meta_bg(sb) &&
> +                             metagroup >= le32_to_cpu(es->s_first_meta_bg);

Instead of creating a new variable, please use existing variable `gdb_num`.

>                 sector_t padding_blocks = meta_bg ? 0 : sbi->s_sbh->b_blocknr -
>                                          ext4_group_first_block_no(sb, 0);
> --
> 2.40.1

Have tested it on different filesystem setups, it does stop corruption.

>
>
>
>
> Amazon Development Center Germany GmbH
> Krausenstr. 38
> 10117 Berlin
> Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss
> Eingetragen am Amtsgericht Charlottenburg unter HRB 149173 B
> Sitz: Berlin
> Ust-ID: DE 289 237 879