With delayed logging, we can get inode allocation buffers in the
same transaction inode unlink buffers. We don't currently mark inode
allocation buffers in the log, so inode unlink buffers take
precedence over allocation buffers.

The result is that when they are combined into the same checkpoint,
only the unlinked inode chain fields are replayed, resulting in
uninitialised inode buffers being detected when the next inode
modification is replayed.

To fix this, we need to ensure that we do not set the inode buffer
flag in the buffer log item format flags if the inode allocation has
not already hit the log. To avoid requiring a change to log
recovery, we really need to make this a modification that relies
only on in-memory sate.

We can do this by checking during buffer log formatting (while the
CIL cannot be flushed) if we are still in the same sequence when we
commit the unlink transaction as the inode allocation transaction.
If we are, then we do not add the inode buffer flag to the buffer
log format item flags. This means the entire buffer will be
replayed, not just the unlinked fields. We do this while
CIL flusheѕ are locked out to ensure that we don't race with the
sequence numbers changing and hence fail to put the inode buffer
flag in the buffer format flags when we really need to.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

If we let the CIL grow without bound, it will grow large enough to violate
recovery constraints (must be at least one complete transaction in the log at
all times) or take forever to write out through the log buffers. Hence we need
a check during asynchronous transactions as to whether the CIL needs to be
pushed.

We track the amount of log space the CIL consumes, so it is relatively simple
to limit it on a pure size basis. Make the limit the minimum of just under half
the log size (recovery constraint) or 8MB of log space (which is an awful lot
of metadata).
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

If the filesystem is being shut down and the there is no log error,
the current code forces out the current log buffers. This code now needs
to push the CIL before it forces out the log buffers to acheive the same
result.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The delayed logging code only changes in-memory structures and as
such can be enabled and disabled with a mount option. Add the mount
option and emit a warning that this is an experimental feature that
should not be used in production yet.

We also need infrastructure to track committed items that have not
yet been written to the log. This is what the Committed Item List
(CIL) is for.

The log item also needs to be extended to track the current log
vector, the associated memory buffer and it's location in the Commit
Item List. Extend the log item and log vector structures to enable
this tracking.

To maintain the current log format for transactions with delayed
logging, we need to introduce a checkpoint transaction and a context
for tracking each checkpoint from initiation to transaction
completion.  This includes adding a log ticket for tracking space
log required/used by the context checkpoint.

To track all the changes we need an io vector array per log item,
rather than a single array for the entire transaction. Using the new
log vector structure for this requires two passes - the first to
allocate the log vector structures and chain them together, and the
second to fill them out.  This log vector chain can then be passed
to the CIL for formatting, pinning and insertion into the CIL.

Formatting of the log vector chain is relatively simple - it's just
a loop over the iovecs on each log vector, but it is made slightly
more complex because we re-write the iovec after the copy to point
back at the memory buffer we just copied into.

This code also needs to pin log items. If the log item is not
already tracked in this checkpoint context, then it needs to be
pinned. Otherwise it is already pinned and we don't need to pin it
again.

The only other complexity is calculating the amount of new log space
the formatting has consumed. This needs to be accounted to the
transaction in progress, and the accounting is made more complex
becase we need also to steal space from it for log metadata in the
checkpoint transaction. Calculate all this at insert time and update
all the tickets, counters, etc correctly.

Once we've formatted all the log items in the transaction, attach
the busy extents to the checkpoint context so the busy extents live
until checkpoint completion and can be processed at that point in
time. Transactions can then be freed at this point in time.

Now we need to issue checkpoints - we are tracking the amount of log space
used by the items in the CIL, so we can trigger background checkpoints when the
space usage gets to a certain threshold. Otherwise, checkpoints need ot be
triggered when a log synchronisation point is reached - a log force event.

Because the log write code already handles chained log vectors, writing the
transaction is trivial, too. Construct a transaction header, add it
to the head of the chain and write it into the log, then issue a
commit record write. Then we can release the checkpoint log ticket
and attach the context to the log buffer so it can be called during
Io completion to complete the checkpoint.

We also need to allow for synchronising multiple in-flight
checkpoints. This is needed for two things - the first is to ensure
that checkpoint commit records appear in the log in the correct
sequence order (so they are replayed in the correct order). The
second is so that xfs_log_force_lsn() operates correctly and only
flushes and/or waits for the specific sequence it was provided with.

To do this we need a wait variable and a list tracking the
checkpoint commits in progress. We can walk this list and wait for
the checkpoints to change state or complete easily, an this provides
the necessary synchronisation for correct operation in both cases.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Document the design of the delayed logging implementation. This
includes assumptions made, dead ends followed, the reasoning behind
the structuring of the code, the layout of various structures, how
things fit together, traps and pit-falls avoided, etc. This is all
too much to document in the code itself, so do it in a separate
file.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NAlex Elder <aelder@sgi.com>

When we free a metadata extent, we record it in the per-AG busy
extent array so that it is not re-used before the freeing
transaction hits the disk. This array is fixed size, so when it
overflows we make further allocation transactions synchronous
because we cannot track more freed extents until those transactions
hit the disk and are completed. Under heavy mixed allocation and
freeing workloads with large log buffers, we can overflow this array
quite easily.

Further, the array is sparsely populated, which means that inserts
need to search for a free slot, and array searches often have to
search many more slots that are actually used to check all the
busy extents. Quite inefficient, really.

To enable this aspect of extent freeing to scale better, we need
a structure that can grow dynamically. While in other areas of
XFS we have used radix trees, the extents being freed are at random
locations on disk so are better suited to being indexed by an rbtree.

So, use a per-AG rbtree indexed by block number to track busy
extents.  This incures a memory allocation when marking an extent
busy, but should not occur too often in low memory situations. This
should scale to an arbitrary number of extents so should not be a
limitation for features such as in-memory aggregation of
transactions.

However, there are still situations where we can't avoid allocating
busy extents (such as allocation from the AGFL). To minimise the
overhead of such occurences, we need to avoid doing a synchronous
log force while holding the AGF locked to ensure that the previous
transactions are safely on disk before we use the extent. We can do
this by marking the transaction doing the allocation as synchronous
rather issuing a log force.

Because of the locking involved and the ordering of transactions,
the synchronous transaction provides the same guarantees as a
synchronous log force because it ensures that all the prior
transactions are already on disk when the synchronous transaction
hits the disk. i.e. it preserves the free->allocate order of the
extent correctly in recovery.

By doing this, we avoid holding the AGF locked while log writes are
in progress, hence reducing the length of time the lock is held and
therefore we increase the rate at which we can allocate and free
from the allocation group, thereby increasing overall throughput.

The only problem with this approach is that when a metadata buffer is
marked stale (e.g. a directory block is removed), then buffer remains
pinned and locked until the log goes to disk. The issue here is that
if that stale buffer is reallocated in a subsequent transaction, the
attempt to lock that buffer in the transaction will hang waiting
the log to go to disk to unlock and unpin the buffer. Hence if
someone tries to lock a pinned, stale, locked buffer we need to
push on the log to get it unlocked ASAP. Effectively we are trading
off a guaranteed log force for a much less common trigger for log
force to occur.

Ideally we should not reallocate busy extents. That is a much more
complex fix to the problem as it involves direct intervention in the
allocation btree searches in many places. This is left to a future
set of modifications.

Finally, now that we track busy extents in allocated memory, we
don't need the descriptors in the transaction structure to point to
them. We can replace the complex busy chunk infrastructure with a
simple linked list of busy extents. This allows us to remove a large
chunk of code, making the overall change a net reduction in code
size.
Signed-off-by: NDave Chinner <david@fromorbit.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The ticket ID is needed to uniquely identify transactions when doing busy
extent matching. Delayed logging changes the lifecycle of busy extents with
respect to the transaction structure lifecycle. Hence we can no longer use
the transaction structure as a means of determining the owner of the busy
extent as it may be freed and reused while the busy extent is still active.

This commit provides the infrastructure to access the xlog_tid_t held in the
ticket from a transaction handle. This avoids the need for callers to peek
into the transaction and log structures to find this out.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Push the error message output when a ticket overrun is detected
into the ticket printing functions. Also remove the debug version
of the code as the production version will still panic just as
effectively on a debug kernel via the panic mask being set.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Clean up the buffer log format (XFS_BLI_*) flags because they have a
polluted namespace. They XFS_BLI_ prefix is used for both in-memory
and on-disk flag feilds, but have overlapping values for different
flags. Rename the buffer log format flags to use the XFS_BLF_*
prefix to avoid confusing them with the in-memory XFS_BLI_* prefixed
flags.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The buffer log item reference counts used to take referenceѕ for every
transaction, similar to the pin counting. This is symmetric (like the
pin/unpin) with respect to transaction completion, but with dleayed logging
becomes assymetric as the pinning becomes assymetric w.r.t. transaction
completion.

To make both cases the same, allow the buffer pinning to take a reference to
the buffer log item and always drop the reference the transaction has on it
when being unlocked. This is balanced correctly because the unpin operation
always drops a reference to the log item. Hence reference counting becomes
symmetric w.r.t. item pinning as well as w.r.t active transactions and as a
result the reference counting model remain consistent between normal and
delayed logging.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Delayed logging currently requires ticket allocation to succeed, so
we need to be able to sleep on allocation. It also should not allow
memory allocation to recurse into the filesystem. hence we need to
pass allocation flags directing the type of allocation the caller
requires.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The transaction ID is written into the log as the unique identifier
for transactions during recover. When duplicating a transaction, we
reuse the log ticket, which means it has the same transaction ID as
the previous transaction.

Rather than regenerating a random transaction ID for the duplicated
transaction, just add one to the current ID so that duplicated
transaction can be easily spotted in the log and during recovery
during problem diagnosis.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

And also drop a useless argument to xfs_iomap_write_direct.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Rename all iomap_valid identifiers to imap_valid to fit the new
world order, and clean up xfs_iomap_valid to convert the passed in
offset to blocks instead of the imap values to bytes.  Use the
simpler inode->i_blkbits instead of the XFS macros for this.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The IOMAP_ flags are now only used inside xfs_aops.c for extent
probing and I/O completion tracking, so more them here, and rename
them to IO_* as there's no mapping involved at all.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Now that struct xfs_iomap contains exactly the same units as struct
xfs_bmbt_irec we can just use the latter directly in the aops code.
Replace the missing IOMAP_NEW flag with a new boolean output
parameter to xfs_iomap.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Report the iomap_bn field of struct xfs_iomap in terms of filesystem
blocks instead of in terms of bytes.  Shift the byte conversions
into the caller, and replace the IOMAP_DELAY and IOMAP_HOLE flag
checks with checks for HOLESTARTBLOCK and DELAYSTARTBLOCK.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Report the iomap_offset and iomap_bsize fields of struct xfs_iomap
in terms of fsblocks instead of in terms of disk blocks.  Shift the
byte conversions into the callers temporarily, but they will
disappear or get cleaned up later.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The iomap_delta field in struct xfs_iomap just contains the
difference between the offset passed to xfs_iomap and the
iomap_offset.  Just calculate it in the only caller that cares.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Instead of using the iomap_target field in struct xfs_iomap
and the IOMAP_REALTIME flag just use the already existing
xfs_find_bdev_for_inode helper.  There's some fallout as we
need to pass the inode in a few more places, which we also
use to sanitize some calling conventions.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

We only call xfs_iomap for single mappings anyway, so remove all
code dealing with multiple mappings from xfs_imap_to_bmap and add
asserts that we never get results that we do not expect.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

We currenly have a routine xfs_trans_buf_item_match_all which checks
if any log item in a transaction contains a given buffer, and a
second one that only does this check for the first, embedded chunk
of log items.  We only use the second routine if we know we only
have that log item chunk, so get rid of the limited routine and
always use the more complete one.

Also rename the old xfs_trans_buf_item_match_all to
xfs_trans_buf_item_match and update various surrounding comments,
and move the remaining xfs_trans_buf_item_match on top of the file
to avoid a forward prototype.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

According to Documentation/filesystems/fiemap.txt, If fm_extent_count
is zero, then the fm_extents[] array is ignored (no extents will be
returned), and the fm_mapped_extents count will hold the number of
extents needed.

But as the commit 97db39a1 has changed
bmv_count to the caller's input buffer, this number query function can't
work any more. As this commit is written to change bmv_count from
MAXEXTNUM because of ENOMEM.

This patch just try to  set bm.bmv_count to something sane.
Thanks to Dave Chinner <david@fromorbit.com> for the suggestion.

Cc: Eric Sandeen <sandeen@redhat.com>
Cc: Alex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NTao Ma <tao.ma@oracle.com>

There remains only one user of the l_sectbb_mask field in the log
structure.  Just kill it off and compute the mask where needed from
the power-of-2 sector size.

(Only update from last post is to accomodate the changes in the
previous patch in the series.)
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Change struct log so it keeps track of the size (in basic blocks) of
a log sector in l_sectBBsize rather than the log-base-2 of that
value (previously, l_sectbb_log).  The name was chosen for
consistency with the other fields in the structure that represent
a number of basic blocks.

(Updated so that a variable used in computing and verifying a log's
sector size is named "log2_size".  Also added the "BB" to the
structure field name, based on feedback from Eric Sandeen.  Also
dropped some superfluous parentheses.)
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NEric Sandeen <sandeen@sandeen.net>

This can't be enabled through the build system and has been dead for
ages.  Note that the CRC patches add back log checksumming, but the
code is quite different from the version removed here anyway.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <david@fromorbit.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <david@fromorbit.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <david@fromorbit.com>

Access fields in m_quotainfo directly instead of hiding them behind the
XFS_QI_* macros.  Add local variables for the quotainfo pointer in places
where we have lots of them.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <david@fromorbit.com>

We need to wait for all pending direct I/O requests before taking care of
metadata in fsync and write_inode.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <david@fromorbit.com>

fs/xfs/linux-2.6/xfs_trace.c: xfs_attr_sf.h is included more than once.
Signed-off-by: NAndrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: NAlex Elder <aelder@sgi.com>

Odds and ends in "xfs_log_recover.c".  This patch just contains some
minor things that didn't seem to warrant their own individual
patches:
- In xlog_bread_noalign(), drop an assertion that a pointer is
  non-null (the crash will tell us it was a bad pointer).
- Add a more descriptive header comment for xlog_find_verify_cycle().
- Make a few additions to the comments in xlog_find_head().  Also
  rearrange some expressions in a few spots to produce the same
  result, but in a way that seems more clear what's being computed.

(Updated in response to Dave's review comments.  Note I did not
split this patch like I said I would.)
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

In xlog_find_cycle_start() use a local variable for some repeated
operations rather than constantly accessing the memory location
whose address is passed in.

(This version drops an assertion that a pointer is non-null.)
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Rename a label used in xlog_find_head() that I thought was poorly
chosen.  Also combine two adjacent labels xlog_find_tail() into a
single label, and give it a more generic name.

(Now using Dave's suggested "validate_head" name for first label.)
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

xfs_bwrite is used with the intention of synchronously writing out
buffers, but currently it does not actually clear the async flag if
that's left from previous writes but instead implements async
behaviour if it finds it.  Remove the code handling asynchronous
writes as we've got rid of those entirely outside of the log and
delwri buffers, and make sure that we clear the async and read flags
before writing the buffer.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

All modifications to the superblock are done transactional through
xfs_trans_log_buf, so there is no reason to initiate periodic
asynchronous writeback.  This only removes the superblock from the
delwri list and will lead to sub-optimal I/O scheduling.

Cut down xfs_sync_fsdata now that it's only used for synchronous
superblock writes and move the log coverage checks into the two
callers.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAlex Elder <aelder@sgi.com>

The transaction ID that is written to the log for a transaction is
currently set by taking the lower 32 bits of the memory address of
the ticket structure.  This is not guaranteed to be unique as
tickets comes from a slab and slots can be reallocated immediately
after being freed. As a result, there is no guarantee of uniqueness
in the ticket ID value.

Fix this by assigning a random number to the ticket ID field so that
it is extremely unlikely that duplicates will occur and remove the
possibility of transactions being mixed up during recovery due to
duplicate IDs.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

There are a number of places where a log sector size of 1 uses
special case code.  The round_up() and round_down() macros
produce the correct result even when the log sector size is 1, and
this eliminates the need for treating this as a special case.
Signed-off-by: NAlex Elder <aelder@sgi.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>