Separate out the bufferhead based mapping from the writepage code so
that we have a clear separation of the page operations and the
bufferhead state.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDave Chinner <david@fromorbit.com>

xfs_cluster_write() is not necessary now that xfs_vm_writepages()
aggregates writepage calls across a single mapping. This means we no
longer need to do page lookups in xfs_cluster_write, so writeback
only needs to look up th epage cache once per page being written.
This also removes a large amount of mostly duplicate code between
xfs_do_writepage() and xfs_convert_page().
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDave Chinner <david@fromorbit.com>

xfs_vm_writepages() calls generic_writepages to writeback a range of
a file, but then xfs_vm_writepage() clusters pages itself as it does
not have any context it can pass between->writepage calls from
__write_cache_pages().

Introduce a writeback context for xfs_vm_writepages() and call
__write_cache_pages directly with our own writepage callback so that
we can pass that context to each writepage invocation. This
encapsulates the current mapping, whether it is valid or not, the
current ioend and it's IO type and the ioend chain being built.

This requires us to move the ioend submission up to the level where
the writepage context is declared. This does mean we do not submit
IO until we packaged the entire writeback range, but with the block
plugging in the writepages call this is the way IO is submitted,
anyway.

It also means that we need to handle discontiguous page ranges.  If
the pages sent down by write_cache_pages to the writepage callback
are discontiguous, we need to detect this and put each discontiguous
page range into individual ioends. This is needed to ensure that the
ioend accurately represents the range of the file that it covers so
that file size updates during IO completion set the size correctly.
Failure to take into account the discontiguous ranges results in
files being too small when writeback patterns are non-sequential.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDave Chinner <david@fromorbit.com>

We currently have code to cancel ioends being built because we
change bufferhead state as we build the ioend. On error, this needs
to be unwound and so we have cancelling code that walks the buffers
on the ioend chain and undoes these state changes.

However, the IO submission path already handles state changes for
buffers when a submission error occurs, so we don't really need a
separate cancel function to do this - we can simply submit the
ioend chain with the specific error and it will be cancelled rather
than submitted.

Hence we can remove the explicit cancel code and just rely on
submission to deal with the error correctly.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Remove the nonblocking optimisation done for mapping lookups during
writeback. It's not clear that leaving a hole in the writeback range
just because we couldn't get a lock is really a win, as it makes us
do another small random IO later on rather than a large sequential
IO now.

As this gets in the way of sane error handling later on, just remove
for the moment and we can re-introduce an equivalent optimisation in
future if we see problems due to extent map lock contention.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDave Chinner <david@fromorbit.com>

This allows us to see page cache driven readahead in action as it
passes through XFS. This helps to understand buffered read
throughput problems such as readahead IO IO sizes being too small
for the underlying device to reach max throughput.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

For DAX, we are now doing block zeroing during allocation. This
means we no longer need a special DAX fault IO completion callback
to do unwritten extent conversion. Because mmap never extends the
file size (it SEGVs the process) we don't need a callback to update
the file size, either. Hence we can remove the completion callbacks
from the __dax_fault and __dax_mkwrite calls.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

DAX has a page fault serialisation problem with block allocation.
Because it allows concurrent page faults and does not have a page
lock to serialise faults to the same page, it can get two concurrent
faults to the page that race.

When two read faults race, this isn't a huge problem as the data
underlying the page is not changing and so "detect and drop" works
just fine. The issues are to do with write faults.

When two write faults occur, we serialise block allocation in
get_blocks() so only one faul will allocate the extent. It will,
however, be marked as an unwritten extent, and that is where the
problem lies - the DAX fault code cannot differentiate between a
block that was just allocated and a block that was preallocated and
needs zeroing. The result is that both write faults end up zeroing
the block and attempting to convert it back to written.

The problem is that the first fault can zero and convert before the
second fault starts zeroing, resulting in the zeroing for the second
fault overwriting the data that the first fault wrote with zeros.
The second fault then attempts to convert the unwritten extent,
which is then a no-op because it's already written. Data loss occurs
as a result of this race.

Because there is no sane locking construct in the page fault code
that we can use for serialisation across the page faults, we need to
ensure block allocation and zeroing occurs atomically in the
filesystem. This means we can still take concurrent page faults and
the only time they will serialise is in the filesystem
mapping/allocation callback. The page fault code will always see
written, initialised extents, so we will be able to remove the
unwritten extent handling from the DAX code when all filesystems are
converted.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Both direct IO and DAX pass an offset and count into get_blocks that
will overflow a s64 variable when an IO goes into the last supported
block in a file (i.e. at offset 2^63 - 1FSB bytes). This can be seen
from the tracing:

xfs_get_blocks_alloc: [...] offset 0x7ffffffffffff000 count 4096
xfs_gbmap_direct:     [...] offset 0x7ffffffffffff000 count 4096
xfs_gbmap_direct_none:[...] offset 0x7ffffffffffff000 count 4096

0x7ffffffffffff000 + 4096 = 0x8000000000000000, and hence that
overflows the s64 offset and we fail to detect the need for a
filesize update and an ioend is not allocated.

This is *mostly* avoided for direct IO because such extending IOs
occur with full block allocation, and so the "IS_UNWRITTEN()" check
still evaluates as true and we get an ioend that way. However, doing
single sector extending IOs to this last block will expose the fact
that file size updates will not occur after the first allocating
direct IO as the overflow will then be exposed.

There is one further complexity: the DAX page fault path also
exposes the same issue in block allocation. However, page faults
cannot extend the file size, so in this case we want to allocate the
block but do not want to allocate an ioend to enable file size
update at IO completion. Hence we now need to distinguish between
the direct IO patch allocation and dax fault path allocation to
avoid leaking ioend structures.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

The iomap codepath (via get_blocks()) acquires and release the inode
lock in the case of a direct write that requires block allocation. This
is because xfs_iomap_write_direct() allocates a transaction, which means
the ilock must be dropped and reacquired after the transaction is
allocated and reserved.

xfs_iomap_write_direct() invokes xfs_iomap_eof_align_last_fsb() before
the transaction is created and thus before the ilock is reacquired. This
can lead to calls to xfs_iread_extents() and reads of the in-core extent
list without any synchronization (via xfs_bmap_eof() and
xfs_bmap_last_extent()). xfs_iread_extents() assert fails if the ilock
is not held, but this is not currently seen in practice as the current
callers had already invoked xfs_bmapi_read().

What has been seen in practice are reports of crashes down in the
xfs_bmap_eof() codepath on direct writes due to seemingly bogus pointer
references from xfs_iext_get_ext(). While an explicit reproducer is not
currently available to confirm the cause of the problem, crash analysis
and code inspection from David Jeffrey had identified the insufficient
locking.

xfs_iomap_eof_align_last_fsb() is called from other contexts with the
inode lock already held, so we cannot acquire it therein.
__xfs_get_blocks() acquires and drops the ilock with variable flags to
cover the event that the extent list must be read in. The common case is
that __xfs_get_blocks() acquires the shared ilock. To provide locking
around the last extent alignment call without adding more lock cycles to
the dio path, update xfs_iomap_write_direct() to expect the shared ilock
held on entry and do the extent alignment under its protection. Demote
the lock, if necessary, from __xfs_get_blocks() and push the
xfs_qm_dqattach() call outside of the shared lock critical section.
Also, add an assert to document that the extent list is always expected
to be present in this path. Otherwise, we risk a call to
xfs_iread_extents() while under the shared ilock. This is safe as all
current callers have executed an xfs_bmapi_read() call under the current
iolock context.
Reported-by: NDavid Jeffery <djeffery@redhat.com>
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

When I ran xfstest/073 case, the remount process was blocked to wait
transactions to be zero. I found there was a io error happened, and
the setfilesize transaction was not released properly. We should add
the changes to cancel the io error in this case.

Reproduction steps:
1. dd if=/dev/zero of=xfs1.img bs=1M count=2048
2. mkfs.xfs xfs1.img
3. losetup -f ./xfs1.img /dev/loop0
4. mount -t xfs /dev/loop0 /home/test_dir/
5. mkdir /home/test_dir/test
6. mkfs.xfs -dfile,name=image,size=2g
7. mount -t xfs -o loop image /home/test_dir/test
8. cp a file bigger than 2g to /home/test_dir/test
9. mount -t xfs -o remount,ro /home/test_dir/test

[ dchinner: moved io error detection to xfs_setfilesize_ioend() after
  transaction context restoration. ]
Signed-off-by: NZhao Hongjiang <zhaohongjiang@huawei.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

There is an issue with xfs's error reporting in some cases of I/O partially
failing and partially succeeding. Calls like fsync() can report success even
though not all I/O was successful in partial-failure cases such as one disk of
a RAID0 array being offline.

The issue can occur when there are more than one bio per xfs_ioend struct.
Each call to xfs_end_bio() for a bio completing will write a value to
ioend->io_error.  If a successful bio completes after any failed bio, no
error is reported do to it writing 0 over the error code set by any failed bio.
The I/O error information is now lost and when the ioend is completed
only success is reported back up the filesystem stack.

xfs_end_bio() should only set ioend->io_error in the case of BIO_UPTODATE
being clear.  ioend->io_error is initialized to 0 at allocation so only needs
to be updated by a failed bio. Also check that ioend->io_error is 0 so that
the first error reported will be the error code returned.

Cc: stable@vger.kernel.org
Signed-off-by: NDavid Jeffery <djeffery@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Preparation to hide the sb->s_writers internals from xfs and btrfs.
Add 2 trivial define's they can use rather than play with ->s_writers
directly. No changes in btrfs/transaction.o and xfs/xfs_aops.o.
Signed-off-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NJan Kara <jack@suse.com>

We can always fill up the bio now, no need to estimate the possible
size based on queue parameters.
Acked-by: NSteven Whitehouse <swhiteho@redhat.com>
Signed-off-by: NKent Overstreet <kent.overstreet@gmail.com>
[hch: rebased and wrote a changelog]
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMing Lin <ming.l@ssi.samsung.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

Currently we have two different ways to signal an I/O error on a BIO:

 (1) by clearing the BIO_UPTODATE flag
 (2) by returning a Linux errno value to the bi_end_io callback

The first one has the drawback of only communicating a single possible
error (-EIO), and the second one has the drawback of not beeing persistent
when bios are queued up, and are not passed along from child to parent
bio in the ever more popular chaining scenario.  Having both mechanisms
available has the additional drawback of utterly confusing driver authors
and introducing bugs where various I/O submitters only deal with one of
them, and the others have to add boilerplate code to deal with both kinds
of error returns.

So add a new bi_error field to store an errno value directly in struct
bio and remove the existing mechanisms to clean all this up.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

The flags argument to xfs_trans_commit is not useful for most callers, as
a commit of a transaction without a permanent log reservation must pass
0 here, and all callers for a transaction with a permanent log reservation
except for xfs_trans_roll must pass XFS_TRANS_RELEASE_LOG_RES.  So remove
the flags argument from the public xfs_trans_commit interfaces, and
introduce low-level __xfs_trans_commit variant just for xfs_trans_roll
that regrants a log reservation instead of releasing it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

xfs_trans_cancel takes two flags arguments: XFS_TRANS_RELEASE_LOG_RES and
XFS_TRANS_ABORT.  Both of them are a direct product of the transaction
state, and can be deducted:

 - any dirty transaction needs XFS_TRANS_ABORT to be properly canceled,
   and XFS_TRANS_ABORT is a noop for a transaction that is not dirty.
 - any transaction with a permanent log reservation needs
   XFS_TRANS_RELEASE_LOG_RES to be properly canceled, and passing
   XFS_TRANS_RELEASE_LOG_RES for a transaction without a permanent
   log reservation is invalid.

So just remove the flags argument and do the right thing.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

DAX does not do buffered IO (can't buffer direct access!) and hence
all read/write IO is vectored through the direct IO path.  Hence we
need to add the DAX IO path callouts to the direct IO
infrastructure.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Add the initial support for DAX file operations to XFS. This
includes the necessary block allocation and mmap page fault hooks
for DAX to function.

Note that there are changes to the splice interfaces to ensure that
for DAX splice avoids direct page cache manipulations and instead
takes the DAX IO paths for read/write operations.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

When modifying PG_Dirty on cached file pages, update the new
MEM_CGROUP_STAT_DIRTY counter.  This is done in the same places where
global NR_FILE_DIRTY is managed.  The new memcg stat is visible in the
per memcg memory.stat cgroupfs file.  The most recent past attempt at
this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632

The new accounting supports future efforts to add per cgroup dirty
page throttling and writeback.  It also helps an administrator break
down a container's memory usage and provides evidence to understand
memcg oom kills (the new dirty count is included in memcg oom kill
messages).

The ability to move page accounting between memcg
(memory.move_charge_at_immigrate) makes this accounting more
complicated than the global counter.  The existing
mem_cgroup_{begin,end}_page_stat() lock is used to serialize move
accounting with stat updates.
Typical update operation:
	memcg = mem_cgroup_begin_page_stat(page)
	if (TestSetPageDirty()) {
		[...]
		mem_cgroup_update_page_stat(memcg)
	}
	mem_cgroup_end_page_stat(memcg)

Summary of mem_cgroup_end_page_stat() overhead:
- Without CONFIG_MEMCG it's a no-op
- With CONFIG_MEMCG and no inter memcg task movement, it's just
  rcu_read_lock()
- With CONFIG_MEMCG and inter memcg  task movement, it's
  rcu_read_lock() + spin_lock_irqsave()

A memcg parameter is added to several routines because their callers
now grab mem_cgroup_begin_page_stat() which returns the memcg later
needed by for mem_cgroup_update_page_stat().

Because mem_cgroup_begin_page_stat() may disable interrupts, some
adjustments are needed:
- move __mark_inode_dirty() from __set_page_dirty() to its caller.
  __mark_inode_dirty() locking does not want interrupts disabled.
- use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in
  __delete_from_page_cache(), replace_page_cache_page(),
  invalidate_complete_page2(), and __remove_mapping().

   text    data     bss      dec    hex filename
8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before
8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after
                            +192 text bytes
8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before
8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after
                            +773 text bytes

Performance tests run on v4.0-rc1-36-g4f671fe2.  Lower is better for
all metrics, they're all wall clock or cycle counts.  The read and write
fault benchmarks just measure fault time, they do not include I/O time.

* CONFIG_MEMCG not set:
                            baseline                              patched
  kbuild                 1m25.030000(+-0.088% 3 samples)       1m25.426667(+-0.120% 3 samples)
  dd write 100 MiB          0.859211561 +-15.10%                  0.874162885 +-15.03%
  dd write 200 MiB          1.670653105 +-17.87%                  1.669384764 +-11.99%
  dd write 1000 MiB         8.434691190 +-14.15%                  8.474733215 +-14.77%
  read fault cycles       254.0(+-0.000% 10 samples)            253.0(+-0.000% 10 samples)
  write fault cycles     2021.2(+-3.070% 10 samples)           1984.5(+-1.036% 10 samples)

* CONFIG_MEMCG=y root_memcg:
                            baseline                              patched
  kbuild                 1m25.716667(+-0.105% 3 samples)       1m25.686667(+-0.153% 3 samples)
  dd write 100 MiB          0.855650830 +-14.90%                  0.887557919 +-14.90%
  dd write 200 MiB          1.688322953 +-12.72%                  1.667682724 +-13.33%
  dd write 1000 MiB         8.418601605 +-14.30%                  8.673532299 +-15.00%
  read fault cycles       266.0(+-0.000% 10 samples)            266.0(+-0.000% 10 samples)
  write fault cycles     2051.7(+-1.349% 10 samples)           2049.6(+-1.686% 10 samples)

* CONFIG_MEMCG=y non-root_memcg:
                            baseline                              patched
  kbuild                 1m26.120000(+-0.273% 3 samples)       1m25.763333(+-0.127% 3 samples)
  dd write 100 MiB          0.861723964 +-15.25%                  0.818129350 +-14.82%
  dd write 200 MiB          1.669887569 +-13.30%                  1.698645885 +-13.27%
  dd write 1000 MiB         8.383191730 +-14.65%                  8.351742280 +-14.52%
  read fault cycles       265.7(+-0.172% 10 samples)            267.0(+-0.000% 10 samples)
  write fault cycles     2070.6(+-1.512% 10 samples)           2084.4(+-2.148% 10 samples)

As expected anon page faults are not affected by this patch.

tj: Updated to apply on top of the recent cancel_dirty_page() changes.
Signed-off-by: NSha Zhengju <handai.szj@gmail.com>
Signed-off-by: NGreg Thelen <gthelen@google.com>
Signed-off-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NJens Axboe <axboe@fb.com>

Struct bio has a reference count that controls when it can be freed.
Most uses cases is allocating the bio, which then returns with a
single reference to it, doing IO, and then dropping that single
reference. We can remove this atomic_dec_and_test() in the completion
path, if nobody else is holding a reference to the bio.

If someone does call bio_get() on the bio, then we flag the bio as
now having valid count and that we must properly honor the reference
count when it's being put.
Tested-by: NRobert Elliott <elliott@hp.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

xfs_end_io_direct_write() can race with other IO completions when
updating the in-core inode size. The IO completion processing is not
serialised for direct IO - they are done either under the
IOLOCK_SHARED for non-AIO DIO, and without any IOLOCK held at all
during AIO DIO completion. Hence the non-atomic test-and-set update
of the in-core inode size is racy and can result in the in-core
inode size going backwards if the race if hit just right.

If the inode size goes backwards, this can trigger the EOF zeroing
code to run incorrectly on the next IO, which then will zero data
that has successfully been written to disk by a previous DIO.

To fix this bug, we need to serialise the test/set updates of the
in-core inode size. This first patch introduces locking around the
relevant updates and checks in the DIO path. Because we now have an
ioend in xfs_end_io_direct_write(), we know exactly then we are
doing an IO that requires an in-core EOF update, and we know that
they are not running in interrupt context. As such, we do not need to
use irqsave() spinlock variants to protect against interrupts while
the lock is held.

Hence we can use an existing spinlock in the inode to do this
serialisation and so not need to grow the struct xfs_inode just to
work around this problem.

This patch does not address the test/set EOF update in
generic_file_write_direct() for various reasons - that will be done
as a followup with separate explanation.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

DIO writes that lie entirely within EOF have nothing to do in IO
completion. In this case, we don't need no steekin' ioend, and so we
can avoid allocating an ioend until we have a mapping that spans
EOF.

This means that IO completion has two contexts - deferred completion
to the dio workqueue that uses an ioend, and interrupt completion
that does nothing because there is nothing that can be done in this
context.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Currently a DIO overwrite that extends the EOF (e.g sub-block IO or
write into allocated blocks beyond EOF) requires a transaction for
the EOF update. Thi is done in IO completion context, but we aren't
explicitly handling this situation properly and so it can run in
interrupt context. Ensure that we defer IO that spans EOF correctly
to the DIO completion workqueue, and now that we have an ioend in IO
completion we can use the common ioend completion path to do all the
work.

Note: we do not preallocate the append transaction as we can have
multiple mapping and allocation calls per direct IO. hence
preallocating can still leave us with nested transactions by
attempting to map and allocate more blocks after we've preallocated
an append transaction.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Currently we can only tell DIO completion that an IO requires
unwritten extent completion. This is done by a hacky non-null
private pointer passed to Io completion, but the private pointer
does not actually contain any information that is used.

We also need to pass to IO completion the fact that the IO may be
beyond EOF and so a size update transaction needs to be done. This
is currently determined by checks in the io completion, but we need
to determine if this is necessary at block mapping time as we need
to defer the size update transactions to a completion workqueue,
just like unwritten extent conversion.

To do this, first we need to allocate and pass an ioend to to IO
completion. Add this for unwritten extent conversion; we'll do the
EOF updates in the next commit.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

The mapping size calculation is done last in __xfs_get_blocks(), but
we are going to need the actual mapping size we will use to map the
direct IO correctly in xfs_map_direct(). Factor out the calculation
for code clarity, and move the call to be the first operation in
mapping the extent to the returned buffer.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Clarify and separate the buffer mapping logic so that the direct IO mapping is
not tangled up in propagating the extent status to teh mapping buffer. This
makes it easier to extend the direct IO mapping to use an ioend in future.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Now that no one is using rw, remove it completely.
Signed-off-by: NOmar Sandoval <osandov@osandov.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The rw parameter to direct_IO is redundant with iov_iter->type, and
treated slightly differently just about everywhere it's used: some users
do rw & WRITE, and others do rw == WRITE where they should be doing a
bitwise check. Simplify this with the new iov_iter_rw() helper, which
always returns either READ or WRITE.
Signed-off-by: NOmar Sandoval <osandov@osandov.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Most filesystems call through to these at some point, so we'll start
here.
Signed-off-by: NOmar Sandoval <osandov@osandov.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

struct kiocb now is a generic I/O container, so move it to fs.h.
Also do a #include diet for aio.h while we're at it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Back in the days when the direct I/O ->end_io callback could be called
from interrupt context for AIO we needed a structure to hand off to the
workqueue, and reused the ioend structure for this purpose.  These days
->end_io is always called from user or workqueue context, which allows us
to avoid this memory allocation and simplify the code significantly.

[dchinner: removed now unused xfs_finish_ioend_sync() function after
	   Brian Foster did an initial review. ]
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

More on-disk format consolidation.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

More on-disk format consolidation.  A few declarations that weren't on-disk
format related move into better suitable spots.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

More consolidatation for the on-disk format defintions.  Note that the
XFS_IS_REALTIME_INODE moves to xfs_linux.h instead as it is not related
to the on disk format, but depends on a CONFIG_ option.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

xfs_vm_writepage() walks each buffer_head on the page, maps to the block
on disk and attaches to a running ioend structure that represents the
I/O submission. A new ioend is created when the type of I/O (unwritten,
delayed allocation or overwrite) required for a particular buffer_head
differs from the previous. If a buffer_head is a delalloc or unwritten
buffer, the associated bits are cleared by xfs_map_at_offset() once the
buffer_head is added to the ioend.

The process of mapping each buffer_head occurs in xfs_map_blocks() and
acquires the ilock in blocking or non-blocking mode, depending on the
type of writeback in progress. If the lock cannot be acquired for
non-blocking writeback, we cancel the ioend, redirty the page and
return. Writeback will revisit the page at some later point.

Note that we acquire the ilock for each buffer on the page. Therefore
during non-blocking writeback, it is possible to add an unwritten buffer
to the ioend, clear the unwritten state, fail to acquire the ilock when
mapping a subsequent buffer and cancel the ioend. If this occurs, the
unwritten status of the buffer sitting in the ioend has been lost. The
page will eventually hit writeback again, but xfs_vm_writepage() submits
overwrite I/O instead of unwritten I/O and does not perform unwritten
extent conversion at I/O completion. This leads to data corruption
because unwritten extents are treated as holes on reads and zeroes are
returned instead of reading from disk.

Modify xfs_cancel_ioend() to restore the buffer unwritten bit for ioends
of type XFS_IO_UNWRITTEN. This ensures that unwritten extent conversion
occurs once the page is eventually written back.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

XFS has been having trouble with stray delayed allocation extents
beyond EOF for a long time. Recent changes to the collapse range
code has triggered erroneous EBUSY errors on page invalidtion for
block size smaller than page size filesystems. These
have been caused by dirty buffers beyond EOF on a partial page which
do not get written to disk during a sync.

The issue is that write-ahead in xfs_cluster_write() finds such a
partial page and handles it by leaving the page dirty but pushing it
into a writeback state. This used to work just fine, as the
write_cache_pages() code would then find the dirty partial page in
the next mapping tree lookup as the dirty tag is still set.

Unfortunately, when we moved to a mark and sweep approach to
writeback to fix other writeback sync issues, we broken this. THe
act of marking the page as under writeback now clears the TOWRITE
tag in the radix tree, even though the page is still dirty. This
causes the TOWRITE tag to be cleared, and hence the next lookup on
the mapping tree does not find the dirty partial page and so doesn't
try to write it again.

This same writeback bug was found recently in ext4 and fixed in
commit 1c8349a1 ("ext4: fix data integrity sync in ordered mode")
without communication to the wider filesystem community. We can use
exactly the same fix here so the TOWRITE flag is not cleared on
partial page writes.

cc: stable@vger.kernel.org # dependent on 1c8349a1Root-cause-found-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

generic/263 is failing fsx at this point with a page spanning
EOF that cannot be invalidated. The operations are:

1190 mapwrite   0x52c00 thru    0x5e569 (0xb96a bytes)
1191 mapread    0x5c000 thru    0x5d636 (0x1637 bytes)
1192 write      0x5b600 thru    0x771ff (0x1bc00 bytes)

where 1190 extents EOF from 0x54000 to 0x5e569. When the direct IO
write attempts to invalidate the cached page over this range, it
fails with -EBUSY and so any attempt to do page invalidation fails.

The real question is this: Why can't that page be invalidated after
it has been written to disk and cleaned?

Well, there's data on the first two buffers in the page (1k block
size, 4k page), but the third buffer on the page (i.e. beyond EOF)
is failing drop_buffers because it's bh->b_state == 0x3, which is
BH_Uptodate | BH_Dirty.  IOWs, there's dirty buffers beyond EOF. Say
what?

OK, set_buffer_dirty() is called on all buffers from
__set_page_buffers_dirty(), regardless of whether the buffer is
beyond EOF or not, which means that when we get to ->writepage,
we have buffers marked dirty beyond EOF that we need to clean.
So, we need to implement our own .set_page_dirty method that
doesn't dirty buffers beyond EOF.

This is messy because the buffer code is not meant to be shared
and it has interesting locking issues on the buffer dirty bits.
So just copy and paste it and then modify it to suit what we need.

Note: the solutions the other filesystems and generic block code use
of marking the buffers clean in ->writepage does not work for XFS.
It still leaves dirty buffers beyond EOF and invalidations still
fail. Hence rather than play whack-a-mole, this patch simply
prevents those buffers from being dirtied in the first place.

cc: <stable@kernel.org>
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Convert all the errors the core XFs code to negative error signs
like the rest of the kernel and remove all the sign conversion we
do in the interface layers.

Errors for conversion (and comparison) found via searches like:

$ git grep " E" fs/xfs
$ git grep "return E" fs/xfs
$ git grep " E[A-Z].*;$" fs/xfs

Negation points found via searches like:

$ git grep "= -[a-z,A-Z]" fs/xfs
$ git grep "return -[a-z,A-D,F-Z]" fs/xfs
$ git grep " -[a-z].*;" fs/xfs

[ with some bits I missed from Brian Foster ]
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

XFS_ERROR was designed long ago to trap return values, but it's not
runtime configurable, it's not consistently used, and we can do
similar error trapping with ftrace scripts and triggers from
userspace.

Just nuke XFS_ERROR and associated bits.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>