linux-toradex.git/fs/ext4/mmp.c, branch v3.14.48

ext4: Replace open coded mdata csum feature to helper function

2014-11-14T16:59:58+00:00

commit 9aa5d32ba269bec0e7eaba2697a986a7b0bc8528 upstream.

Besides the fact that this replacement improves code readability
it also protects from errors caused direct EXT4_S(sb)->s_es manipulation
which may result attempt to use uninitialized  csum machinery.

#Testcase_BEGIN
IMG=/dev/ram0
MNT=/mnt
mkfs.ext4 $IMG
mount $IMG $MNT
#Enable feature directly on disk, on mounted fs
tune2fs -O metadata_csum  $IMG
# Provoke metadata update, likey result in OOPS
touch $MNT/test
umount $MNT
#Testcase_END

# Replacement script
@@
expression E;
@@
- EXT4_HAS_RO_COMPAT_FEATURE(E, EXT4_FEATURE_RO_COMPAT_METADATA_CSUM)
+ ext4_has_metadata_csum(E)

https://bugzilla.kernel.org/show_bug.cgi?id=82201

Signed-off-by: Dmitry Monakhov 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Greg Kroah-Hartman

ext4: use prandom_u32() instead of get_random_bytes()

2013-11-08T05:14:53+00:00

Many of the uses of get_random_bytes() do not actually need
cryptographically secure random numbers.  Replace those uses with a
call to prandom_u32(), which is faster and which doesn't consume
entropy from the /dev/random driver.

Signed-off-by: "Theodore Ts'o"

ext4: mark all metadata I/O with REQ_META

2013-04-20T19:46:17+00:00

As Dave Chinner pointed out at the 2013 LSF/MM workshop, it's
important that metadata I/O requests are marked as such to avoid
priority inversions caused by I/O bandwidth throttling.

Signed-off-by: "Theodore Ts'o"

ext4: fix big-endian bug in metadata checksum calculations

2013-04-10T03:56:48+00:00

Signed-off-by: Dmitry Monakhov 
Signed-off-by: "Theodore Ts'o" 
Cc: stable@vger.kernel.org

ext4: use unlikely to improve the efficiency of the kernel

2013-01-12T21:28:47+00:00

Because the function 'sb_getblk' seldomly fails to return NULL
value,it will be better to use 'unlikely' to optimize it.

Signed-off-by: Wang Shilong 
Signed-off-by: "Theodore Ts'o"

ext4: return ENOMEM if sb_getblk() fails

2013-01-12T21:19:36+00:00

The only reason for sb_getblk() failing is if it can't allocate the
buffer_head.  So ENOMEM is more appropriate than EIO.  In addition,
make sure that the file system is marked as being inconsistent if
sb_getblk() fails.

Signed-off-by: "Theodore Ts'o" 
Cc: stable@vger.kernel.org

ext4: Convert to new freezing mechanism

2012-07-31T05:45:48+00:00

We remove most of frozen checks since upper layer takes care of blocking all
writes. We have to handle protection in ext4_page_mkwrite() in a special way
because we cannot use generic block_page_mkwrite(). Also we add a freeze
protection to ext4_evict_inode() so that iput() of unlinked inode cannot modify
a frozen filesystem (we cannot easily instrument ext4_journal_start() /
ext4_journal_stop() with freeze protection because we are missing the
superblock pointer in ext4_journal_stop() in nojournal mode).

CC: linux-ext4@vger.kernel.org
CC: "Theodore Ts'o" 
BugLink: https://bugs.launchpad.net/bugs/897421
Tested-by: Kamal Mostafa 
Tested-by: Peter M. Petrakis 
Tested-by: Dann Frazier 
Tested-by: Massimo Morana 
Acked-by: "Theodore Ts'o" 
Signed-off-by: Jan Kara 
Signed-off-by: Al Viro

ext4: add checksums to the MMP block

2012-04-29T22:47:10+00:00

Compute and verify a checksum for the MMP block.

Signed-off-by: Darrick J. Wong 
Signed-off-by: "Theodore Ts'o"

ext4: Fix endianness bug when reading the MMP block

2012-02-27T06:09:03+00:00

Sparse complained about this endian bug in fs/ext4/mmp.c.

Signed-off-by: Santosh Nayak 
Reviewed-by: Johann Lombardi 
Reviewed-by: Andreas Dilger 
Signed-off-by: "Theodore Ts'o"

ext4: Fix comparison endianness problem in MMP initialization

2011-10-18T14:53:51+00:00

As part of startup, the MMP initialization code does this:

mmp->mmp_seq = seq = cpu_to_le32(mmp_new_seq());

Next, mmp->mmp_seq is written out to disk, a delay happens, and then
the MMP block is read back in and the sequence value is tested:

if (seq != le32_to_cpu(mmp->mmp_seq)) {
	/* fail the mount */

On a LE system such as x86, the *le32* functions do nothing and this
works.  Unfortunately, on a BE system such as ppc64, this comparison
becomes:

if (cpu_to_le32(new_seq) != le32_to_cpu(cpu_to_le32(new_seq)) {
	/* fail the mount */

Except for a few palindromic sequence numbers, this test always causes
the mount to fail, which makes MMP filesystems generally unmountable
on ppc64.  The attached patch fixes this situation.

Signed-off-by: Darrick J. Wong 
Signed-off-by: "Theodore Ts'o"