In the calc_code_bit() function, we must find all powers of two beneath
the code bit number, *after* it's shifted by those powers of two.  This
requires a loop to see where it ends up.

We can optimize it by starting at its most significant bit.  This shaves
32% off the time, for a total of 67.6% shaved off of the original, naive
implementation.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

When I wrote ocfs2_hamming_encode(), I was following documentation of
the algorithm and didn't have quite the (possibly still imperfect) grasp
of it I do now.  As part of this, I literally hand-coded xor.  I would
test a bit, and then add that bit via xor to the parity word.

I can, of course, just do a single xor of the parity word and the source
word (the code buffer bit offset).  This cuts CPU usage by 53% on a
mostly populated buffer (an inode containing utmp.h inline).

Joel
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Add OCFS2_FEATURE_INCOMPAT_META_ECC to the list of supported features.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The superblock is read via a raw call.  Validate it after we find it
from its signature.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Use the db_check field of ocfs2_dir_block_trailer to crc/ecc the
dirblocks.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Future ocfs2 features metaecc and indexed directories need to store a
little bit of data in each dirblock.  For compatibility, we place this
in a trailer at the end of the dirblock.  The trailer plays itself as an
empty dirent, so that if the features are turned off, it can be reused
without requiring a tunefs scan.

This code adds the trailer and validates it when the block is read in.

[ Mark is the original author, but I reinserted this code before his
  dir index work.  -- Joel ]
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Change the rest of the naked ocfs2_journal_access() calls in
fs/ocfs2/xattr.c to use the appropriate ocfs2_journal_access_*() call
for their metadata type.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_remove_value_outside() needs to know the type of buffer it is
looking at.  Pass in an ocfs2_xattr_value_buf.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_xattr_set_entry is the function that knows what type of block it
is setting into.  This is what we wanted from ocfs2_xattr_value_buf.
Plus, moving the value buf up into ocfs2_xattr_set_entry() allows us to
pass it into ocfs2_xattr_set_value_outside() and ocfs2_xattr_cleanup().
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_xattr_update_entry() updates the entry portion of an xattr buffer.
This can be part of multiple metadata block types, so pass the buffer in
via an ocfs2_xattr_value_buf.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The callers of ocfs2_xattr_value_truncate() now pass in
ocfs2_xattr_value_bufs.  These callers are the ones that calculated the
xv location, so they are the right starting point.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Place an ocfs2_xattr_value_buf in ocfs2_xattr_value_truncate() and pass
it down to ocfs2_xattr_shrink_size().  We can also pass it into
ocfs2_xattr_extend_allocation(), replacing its ocfs2_xattr_value_buf.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Place an ocfs2_xattr_value_buf in __ocfs2_xattr_shrink_size() and pass
it down to __ocfs2_remove_xattr_range().
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

When an ocfs2 extended attribute is large enough to require its own
allocation tree, we root it with an ocfs2_xattr_value_root.  However,
these roots can be a part of inodes, xattr blocks, or xattr buckets.
Thus, they need a different journal access function for each container.

We wrap the bh, its journal access function, and the value root (xv) in
a structure called ocfs2_xattr_valu_buf.  This is a package that can
be passed around.  In this first pass, we simply pass it to the
extent tree code.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The xattr bucket can span multiple blocks on disk.  We have wrappers
for this structure in the code.  We use the new multi-block ecc calls to
calculate and validate the bucket.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The per-metadata-type ocfs2_journal_access_*() functions hook up jbd2
commit triggers and allow us to compute metadata ecc right before the
buffers are written out.  This commit provides ecc for inodes, extent
blocks, group descriptors, and quota blocks.  It is not safe to use
extened attributes and metaecc at the same time yet.

The ocfs2_extent_tree and ocfs2_path abstractions in alloc.c both hide
the type of block at their root.  Before, it didn't matter, but now the
root block must use the appropriate ocfs2_journal_access_*() function.
To keep this abstract, the structures now have a pointer to the matching
journal_access function and a wrapper call to call it.

A few places use naked ocfs2_write_block() calls instead of adding the
blocks to the journal.  We make sure to calculate their checksum and ecc
before the write.

Since we pass around the journal_access functions.  Let's typedef them
in ocfs2.h.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The majority of ocfs2_new_path() calls are:

	ocfs2_new_path(path_root_bh(otherpath),
		       path_root_el(otherpath));

Let's call that ocfs2_new_path_from_path().  The rest do similar things
from struct ocfs2_extent_tree.  Let's call those
ocfs2_new_path_from_et().  This will make the next change easier.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

We create wrappers for ocfs2_journal_access() that are specific to the
type of metadata block.  This allows us to associate jbd2 commit
triggers with the block.  The triggers will compute metadata ecc in a
future commit.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Add block check calls to the read_block validate functions.  This is the
almost all of the read-side checking of metaecc.  xattr buckets are not checked
yet.   Writes are also unchecked, and so a read-write mount will quickly fail.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Add a currently-returns-success hook for quota block reads.  We'll be
adding checks to this.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

This is the code that computes crc32 and ecc for ocfs2 metadata blocks.
There are high-level functions that check whether the filesystem has the
ecc feature, mid-level functions that work on a single block or array of
buffer_heads, and the low-level ecc hamming code that can handle
multiple buffers like crc32_le().

It's not hooked up to the filesystem yet.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Define struct ocfs2_block_check, an 8-byte structure containing a 32bit
crc32_le and a 16bit hamming code ecc.  This will be used for metadata
checksums.  Add the structure to free spaces in the various metadata
structures.

Add the OCFS2_FEATURE_INCOMPAT_META_ECC bit.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Filesystems often to do compute intensive operation on some
metadata.  If this operation is repeated many times, it can be very
expensive.  It would be much nicer if the operation could be performed
once before a buffer goes to disk.

This adds triggers to jbd2 buffer heads.  Just before writing a metadata
buffer to the journal, jbd2 will optionally call a commit trigger associated
with the buffer.  If the journal is aborted, an abort trigger will be
called on any dirty buffers as they are dropped from pending
transactions.

ocfs2 will use this feature.

Initially I tried to come up with a more generic trigger that could be
used for non-buffer-related events like transaction completion.  It
doesn't tie nicely, because the information a buffer trigger needs
(specific to a journal_head) isn't the same as what a transaction
trigger needs (specific to a tranaction_t or perhaps journal_t).  So I
implemented a buffer set, with the understanding that
journal/transaction wide triggers should be implemented separately.

There is only one trigger set allowed per buffer.  I can't think of any
reason to attach more than one set.  Contrast this with a journal or
transaction in which multiple places may want to watch the entire
transaction separately.

The trigger sets are considered static allocation from the jbd2
perspective.  ocfs2 will just have one trigger set per block type,
setting the same set on every bh of the same type.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

A new mlog mask has to be added into mlog_attribute before it can
be really used in mlog. ML_QUOTA is only added in masklog.h, so
add it to the array to enable it.
Signed-off-by: NTao Ma <tao.ma@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Pass the actual target bucket for insert through to
ocfs2_add_new_xattr_bucket().  Now growing a bucket has no buffer_head
knowledge.

ocfs2_add_new_xattr_bucket() leavs xs->bucket in the proper state for
insert.  However, it doesn't update the rest of the search fields in xs,
so we still have to relse() and re-find.  That's OK, because everything
is cached.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Lift the buckets from ocfs2_add_new_xattr_cluster() up into
ocfs2_add_new_xattr_bucket().  Now ocfs2_add_new_xattr_cluster()
doesn't deal with buffer_heads.  In fact, we no longer have to play
get_bh() tricks at all.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Lift the buckets from ocfs2_adjust_xattr_cross_cluster() up into
ocfs2_add_new_xattr_cluster().  Now ocfs2_adjust_xattr_cross_cluster()
doesn't deal with buffer_heads.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Now that ocfs2_adjust_xattr_cross_cluster() has buckets, it can pass
them into ocfs2_mv_xattr_bucket_cross_cluster().  It no longer has to
care about buffer_heads.  The manipulation of first_bh and header_bh
moves up to ocfs2_adjust_xattr_cross_cluster().
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

We want to be passing around buckets instead of buffer_heads.  Let's get
them into ocfs2_adjust_xattr_cross_cluster.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Now that ocfs2_mv_xattr_buckets() can move a partial cluster's worth of
buckets, ocfs2_mv_xattr_bucket_cross_cluster() can use it.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

If you look at ocfs2_mv_xattr_bucket_cross_cluster(), you'll notice that
two-thirds of the code is almost identical to ocfs2_mv_xattr_buckets().
The only difference is that ocfs2_mv_xattr_buckets() moves a whole
cluster's worth, while ocfs2_mv_xattr_bucket_cross_cluster() moves half
the cluster.

We change ocfs2_mv_xattr_buckets() to allow moving partial clusters.
The original caller of ocfs2_mv_xattr_buckets() still moves the whole
cluster's worth - it just passes a start_bucket of 0.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_cp_xattr_cluster() takes the last cluster of an xattr extent,
copies its buckets to the front of a new extent, and then shrinks the bucket
count of the original extent.  So it's really moving the data, not
copying it.

While we're here, the function doesn't need a buffer_head for the old
extent, just the block number.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

The buffer copy loop of ocfs2_mv_xattr_bucket_cross_cluster() actually
looks a lot like ocfs2_cp_xattr_bucket().  Let's just use that instead.
We also use bucket operations to update the buckets at the start of each
extent.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

I was unsure of the JOURNAL_ACCESS parameters in
ocfs2_cp_xattr_cluster().  They're based on the function argument
't_is_new', but I couldn't quite figure out how t_is_new mapped to
allocation.  ocfs2_cp_xattr_cluster() actually overwrites the target,
regardless of t_is_new.

Well, I just figured it out.  So I'm adding a big fat comment for those
who come after me.  ocfs2_divide_xattr_cluster() has the same behavior.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_cp_xattr_cluster() takes the last bucket of a full extent and
copies it over to a new extent.  It then updates the headers of both
extents to reflect the new state.  It is passed the first bh of
the first bucket in order to update that first extent's bucket count.
It reads and dirties the first bh of the new extent for the same reason.

However, future code wants to always dirty the entire bucket when it
is changed.  So it is changed to read the entire bucket it is updating
for both extents.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_extend_xattr_bucket() takes an extent of buckets and shifts some
of them down to make room for a new xattr.  It is passed the first bh of
the first bucket, because that is where we store the number of buckets
in the extent.

However, future code wants to always dirty the entire bucket when it
is changed.  So let's pass the entire bucket into this function, skip
any block reads (we have them), and add the access/dirty logic.  We also
can skip passing in the target bucket bh - we only need its block
number.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

We move the transaction into the loop because in
ocfs2_remove_extent, we will double the credits in function
ocfs2_extend_rotate_transaction. So if we have a large loop
number, we will soon waste much the journal space.
Signed-off-by: NTao Ma <tao.ma@oracle.com>
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

ocfs2_bucket_value_truncate() currently takes the first bh of the
bucket, and magically plays around with the value bh - even though
the bucket structure in the calling function already has it.

In addition, future code wants to always dirty the entire bucket when it
is changed.  So let's pass the entire bucket into this function, skip
any block reads (we have them), and add the access/dirty logic.

ocfs2_xattr_update_value_size() is no longer necessary, as it only did
one thing other than journal access/dirty.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Fix 2 minor things in quota. They are both found by sparse check.
1. an endian bug in ocfs2_local_quota_add_chunk.
2. change olq_alloc_dquot to static.
Signed-off-by: NTao Ma <tao.ma@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>

Signed-off-by: NTao Ma <tao.ma@oracle.com>
Signed-off-by: NMark Fasheh <mfasheh@suse.com>