linux-stable.git/fs/ext4/resize.c, branch linux-3.16.y

ext4: add missing brelse() in add_new_gdb_meta_bg()

2019-08-13T11:38:41+00:00

commit d64264d6218e6892edd832dc3a5a5857c2856c53 upstream.

Currently in add_new_gdb_meta_bg() there is a missing brelse of gdb_bh
in case ext4_journal_get_write_access() fails.
Additionally kvfree() is missing in the same error path. Fix it by
moving the ext4_journal_get_write_access() before the ext4 sb update as
Ted suggested and release n_group_desc and gdb_bh in case it fails.

Fixes: 61a9c11e5e7a ("ext4: add missing brelse() add_new_gdb_meta_bg()'s error path")
Signed-off-by: Lukas Czerner 
Signed-off-by: Theodore Ts'o 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

ext4: fix crash during online resizing

2019-07-09T21:03:59+00:00

commit f96c3ac8dfc24b4e38fc4c2eba5fea2107b929d1 upstream.

When computing maximum size of filesystem possible with given number of
group descriptor blocks, we forget to include s_first_data_block into
the number of blocks. Thus for filesystems with non-zero
s_first_data_block it can happen that computed maximum filesystem size
is actually lower than current filesystem size which confuses the code
and eventually leads to a BUG_ON in ext4_alloc_group_tables() hitting on
flex_gd->count == 0. The problem can be reproduced like:

truncate -s 100g /tmp/image
mkfs.ext4 -b 1024 -E resize=262144 /tmp/image 32768
mount -t ext4 -o loop /tmp/image /mnt
resize2fs /dev/loop0 262145
resize2fs /dev/loop0 300000

Fix the problem by properly including s_first_data_block into the
computed number of filesystem blocks.

Fixes: 1c6bd7173d66 "ext4: convert file system to meta_bg if needed..."
Signed-off-by: Jan Kara 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: avoid declaring fs inconsistent due to invalid file handles

2019-04-04T15:13:59+00:00

commit 8a363970d1dc38c4ec4ad575c862f776f468d057 upstream.

If we receive a file handle, either from NFS or open_by_handle_at(2),
and it points at an inode which has not been initialized, and the file
system has metadata checksums enabled, we shouldn't try to get the
inode, discover the checksum is invalid, and then declare the file
system as being inconsistent.

This can be reproduced by creating a test file system via "mke2fs -t
ext4 -O metadata_csum /tmp/foo.img 8M", mounting it, cd'ing into that
directory, and then running the following program.

#define _GNU_SOURCE
#include 

struct handle {
	struct file_handle fh;
	unsigned char fid[MAX_HANDLE_SZ];
};

int main(int argc, char **argv)
{
	struct handle h = {{8, 1 }, { 12, }};

	open_by_handle_at(AT_FDCWD, &h.fh, O_RDONLY);
	return 0;
}

Google-Bug-Id: 120690101
Signed-off-by: Theodore Ts'o 
[bwh: Backported to 3.16:
 - Keep using EIO instead of EFSCORRUPTED and EFSBADCRC
 - Drop inapplicable changes
 - Adjust context]
Signed-off-by: Ben Hutchings

ext4: avoid possible double brelse() in add_new_gdb() on error path

2019-02-11T17:53:50+00:00

commit 4f32c38b4662312dd3c5f113d8bdd459887fb773 upstream.

Fixes: b40971426a83 ("ext4: add error checking to calls to ...")
Reported-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: fix possible inode leak in the retry loop of ext4_resize_fs()

2019-02-11T17:53:49+00:00

commit db6aee62406d9fbb53315fcddd81f1dc271d49fa upstream.

Fixes: 1c6bd7173d66 ("ext4: convert file system to meta_bg if needed ...")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: fix missing cleanup if ext4_alloc_flex_bg_array() fails while resizing

2019-02-11T17:53:49+00:00

commit f348e2241fb73515d65b5d77dd9c174128a7fbf2 upstream.

Fixes: 117fff10d7f1 ("ext4: grow the s_flex_groups array as needed ...")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: add missing brelse() update_backups()'s error path

2019-02-11T17:53:47+00:00

commit ea0abbb648452cdb6e1734b702b6330a7448fcf8 upstream.

Fixes: ac27a0ec112a ("ext4: initial copy of files from ext3")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: add missing brelse() add_new_gdb_meta_bg()'s error path

2019-02-11T17:53:47+00:00

commit 61a9c11e5e7a0dab5381afa5d9d4dd5ebf18f7a0 upstream.

Fixes: 01f795f9e0d6 ("ext4: add online resizing support for meta_bg ...")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings

ext4: add missing brelse() in set_flexbg_block_bitmap()'s error path

2019-02-11T17:53:46+00:00

commit cea5794122125bf67559906a0762186cf417099c upstream.

Fixes: 33afdcc5402d ("ext4: add a function which sets up group blocks ...")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

ext4: avoid potential extra brelse in setup_new_flex_group_blocks()

2019-02-11T17:53:46+00:00

commit 9e4028935cca3f9ef9b6a90df9da6f1f94853536 upstream.

Currently bh is set to NULL only during first iteration of for cycle,
then this pointer is not cleared after end of using.
Therefore rollback after errors can lead to extra brelse(bh) call,
decrements bh counter and later trigger an unexpected warning in __brelse()

Patch moves brelse() calls in body of cycle to exclude requirement of
brelse() call in rollback.

Fixes: 33afdcc5402d ("ext4: add a function which sets up group blocks ...")
Signed-off-by: Vasily Averin 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Ben Hutchings