This should make it more clear what this structure is used
for, and how some of the (mutually exclusive) fields are
used to keep page cache references.
Signed-off-by: NCurt Wohlgemuth <curtw@google.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We can clear PageWriteback on each page when the IO
completes, but we can't release the references on the page
until we convert any uninitialized extents.

Without this patch, the use of the dioread_nolock mount
option can break buffered writes, because extents may
not be converted by the time a subsequent buffered read
comes in; if the page is not in the page cache, a read
will return zeros if the extent is still uninitialized.

I tested this with a (temporary) patch that adds a call
to msleep(1000) at the start of ext4_end_io_work(), to delay
processing of each DIO-unwritten work queue item.  With this
msleep(), a simple workload of

  fallocate
  write
  fadvise
  read

will fail without this patch, succeeds with it.
Signed-off-by: NCurt Wohlgemuth <curtw@google.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The following command line will leave the aio-stress process unkillable
on an ext4 file system (in my case, mounted on /mnt/test):

aio-stress -t 20 -s 10 -O -S -o 2 -I 1000 /mnt/test/aiostress.3561.4 /mnt/test/aiostress.3561.4.20 /mnt/test/aiostress.3561.4.19 /mnt/test/aiostress.3561.4.18 /mnt/test/aiostress.3561.4.17 /mnt/test/aiostress.3561.4.16 /mnt/test/aiostress.3561.4.15 /mnt/test/aiostress.3561.4.14 /mnt/test/aiostress.3561.4.13 /mnt/test/aiostress.3561.4.12 /mnt/test/aiostress.3561.4.11 /mnt/test/aiostress.3561.4.10 /mnt/test/aiostress.3561.4.9 /mnt/test/aiostress.3561.4.8 /mnt/test/aiostress.3561.4.7 /mnt/test/aiostress.3561.4.6 /mnt/test/aiostress.3561.4.5 /mnt/test/aiostress.3561.4.4 /mnt/test/aiostress.3561.4.3 /mnt/test/aiostress.3561.4.2

This is using the aio-stress program from the xfstests test suite.
That particular command line tells aio-stress to do random writes to
20 files from 20 threads (one thread per file).  The files are NOT
preallocated, so you will get writes to random offsets within the
file, thus creating holes and extending i_size.  It also opens the
file with O_DIRECT and O_SYNC.

On to the problem.  When an I/O requires unwritten extent conversion,
it is queued onto the completed_io_list for the ext4 inode.  Two code
paths will pull work items from this list.  The first is the
ext4_end_io_work routine, and the second is ext4_flush_completed_IO,
which is called via the fsync path (and O_SYNC handling, as well).
There are two issues I've found in these code paths.  First, if the
fsync path beats the work routine to a particular I/O, the work
routine will free the io_end structure!  It does not take into account
the fact that the io_end may still be in use by the fsync path.  I've
fixed this issue by adding yet another IO_END flag, indicating that
the io_end is being processed by the fsync path.

The second problem is that the work routine will make an assignment to
io->flag outside of the lock.  I have witnessed this result in a hang
at umount.  Moving the flag setting inside the lock resolved that
problem.

The problem was introduced by commit b82e384c ("ext4: optimize
locking for end_io extent conversion"), which first appeared in 3.2.
As such, the fix should be backported to that release (probably along
with the unwritten extent conversion race fix).
Signed-off-by: NJeff Moyer <jmoyer@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
CC: stable@kernel.org

For extent-based files, you can perform DIO to holes, as mentioned in
the comments in ext4_ext_direct_IO.  However, that function passes
DIO_SKIP_HOLES to __blockdev_direct_IO, which is *really* confusing to
the uninitiated reader.  The key, here, is that the get_block function
passed in, ext4_get_block_write, completely ignores the create flag
that is passed to it (the create flag is passed in from the direct I/O
code, which uses the DIO_SKIP_HOLES flag to determine whether or not
it should be cleared).

This is a long-winded way of saying that the DIO_SKIP_HOLES flag is
ultimately ignored.  So let's remove it.
Signed-off-by: NJeff Moyer <jmoyer@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

No other file system allows ACL's and extended attributes to be
enabled or disabled via a mount option.  So let's try to deprecate
these options from ext4.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Users who tried to use the ext4 file system driver is being used for
the ext2 or ext3 file systems (via the CONFIG_EXT4_USE_FOR_EXT23
option) could have failed mounts if their /etc/fstab contains options
recognized by ext2 or ext3 but which have since been removed in ext4.

So teach ext4 to recognize them and give a warning that the mount
option was removed.

Report: https://bbs.archlinux.org/profile.php?id=33804Signed-off-by: NTom Gundersen <teg@jklm.no>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: Thomas Baechler <thomas@archlinux.org>
Cc: Tobias Powalowski <tobias.powalowski@googlemail.com>
Cc: Dave Reisner <d@falconindy.com>

Now that /proc/mounts is consistently showing only those mount options
which need to be specified in /etc/fstab or on the mount command line,
it is useful to have file which shows exactly which file system
options are enabled.  This can be useful when debugging a user
problem.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Consistently show mount options which are the non-default, so that
/proc/mounts accurately shows the mount options that would be
necessary to mount the file system in its current mode of operation.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

This commit is strictly a code movement so in preparation of changing
ext4_show_options to be table driven.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

By using a table-drive approach, we shave about 100 lines of code from
ext4, and make the code a bit more regular and factored out.  This
will also make it possible in a future patch to use this table for
displaying the mount options that were specified in /proc/mounts.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

There's no point to have two bits that are set in parallel; so use the
MS_I_VERSION flag that is needed by the VFS anyway, and that way we
free up a bit in sbi->s_mount_opts.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

This is completely unused so let's just get rid of it.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

People complained about removing both of these features, so per
Linus's dictate, we won't be able to remove them.  Sigh...
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Sparse complained about this endian bug in fs/ext4/mmp.c.
Signed-off-by: NSantosh Nayak <santoshprasadnayak@gmail.com>
Reviewed-by: NJohann Lombardi <johann@whamcloud.com>
Reviewed-by: NAndreas Dilger <adilger@dilger.ca>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Fix ext4_warning format flag in dx_probe().

CC: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NZheng Liu <wenqing.lz@taobao.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Processes hang forever on a sync-mounted ext2 file system that
is mounted with the ext4 module (default in Fedora 16).

I can reproduce this reliably by mounting an ext2 partition with
"-o sync" and opening a new file an that partition with vim. vim
will hang in "D" state forever.  The same happens on ext4 without
a journal.

I am attaching a small patch here that solves this issue for me.
In the sync mounted case without a journal,
ext4_handle_dirty_metadata() may call sync_dirty_buffer(), which
can't be called with buffer lock held.

Also move mb_cache_entry_release inside lock to avoid race
fixed previously by 8a2bfdcb ext[34]: EA block reference count racing fix
Note too that ext2 fixed this same problem in 2006 with
b2f49033 [PATCH] fix deadlock in ext2

Signed-off-by: Martin.Wilck@ts.fujitsu.com
[sandeen@redhat.com: move mb_cache_entry_release before unlock, edit commit msg]
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

When resizing file system in the way that the new size of the file
system is still in the same group (no new groups are added), then we can
hit a BUG_ON in ext4_alloc_group_tables()

BUG_ON(flex_gd->count == 0 || group_data == NULL);

because flex_gd->count is zero. The reason is the missing check for such
case, so the code always extend the last group fully and then attempt to
add more groups, but at that time n_blocks_count is actually smaller
than o_blocks_count.

It can be easily reproduced like this:

mkfs.ext4 -b 4096 /dev/sda 30M
mount /dev/sda /mnt/test
resize2fs /dev/sda 50M

Fix this by checking whether the resize happens within the singe group
and only add that many blocks into the last group to satisfy user
request. Then o_blocks_count == n_blocks_count and the resize will exit
successfully without and attempt to add more groups into the fs.

Also fix mixing together block number and blocks count which might be
confusing and can easily lead to off-by-one errors (but it is actually
not the case here since the two occurrence of this mix-up will cancel
each other).
Signed-off-by: NLukas Czerner <lczerner@redhat.com>
Reported-by: NMilan Broz <mbroz@redhat.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The following comment in ext4_end_io_dio caught my attention:

	/* XXX: probably should move into the real I/O completion handler */
        inode_dio_done(inode);

The truncate code takes i_mutex, then calls inode_dio_wait.  Because the
ext4 code path above will end up dropping the mutex before it is
reacquired by the worker thread that does the extent conversion, it
seems to me that the truncate can happen out of order.  Jan Kara
mentioned that this might result in error messages in the system logs,
but that should be the extent of the "damage."

The fix is pretty straight-forward: don't call inode_dio_done until the
extent conversion is complete.
Reviewed-by: NJan Kara <jack@suse.cz>
Signed-off-by: NJeff Moyer <jmoyer@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

Get rid of this one:

fs/ext4/balloc.c: In function 'ext4_wait_block_bitmap':
fs/ext4/balloc.c:405:3: warning: format '%llu' expects argument of
  type 'long long unsigned int', but argument 6 has type 'sector_t' [-Wformat]

Happens because sector_t is u64 (unsigned long long) or unsigned long
dependent on CONFIG_64BIT.
Signed-off-by: NHeiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: NRandy Dunlap <rdunlap@xenotime.net>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The EXT4_MB_BITMAP and EXT4_MB_BUDDY macros obfuscate more than they
provide any abstraction.   So remove them.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

"inode" is a valid pointer here.  "tmp_inode" was intended.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We dereference "bh" unconditionally a couple lines down to find
"by->b_size".  This function is never called with a NULL "bh" so I have
removed the check.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We could return directly from ext4_xattr_check_block(). Thus, we
shouldn't need to define a 'error' variable.
Signed-off-by: NZheng Liu <wenqing.lz@taobao.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The resize mount option seems to be of limited value,
especially in the age of online resize2fs.  Nuke it.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The V2 journal format was introduced around ten years ago,
for ext3. It seems highly unlikely that anyone will need this
migration option for ext4.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The 'orig_size' local variable is only used in a call to
mb_debug().  Mark it with '__maybe_unused'.
Signed-off-by: NCurt Wohlgemuth <curtw@google.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Use the KMEM_CACHE helper macro instead of kmem_cache_create().
Signed-off-by: NYongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

This patch renames functions initializing the slab caches for the
journal head and handle structures to so they are consistent with the
names of the corresponding functions which destroys those slab caches.
Signed-off-by: NYongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

There is normally only a handful of these active at any one time, but
putting them in a separate slab cache makes debugging memory
corruption problems easier.  Manish Katiyar also wanted this make it
easier to test memory failure scenarios in the jbd2 layer.

Cc: Manish Katiyar <mkatiyar@gmail.com>
Signed-off-by: NYongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The per-commit callback was used by mballoc code to manage free space
bitmaps after deleted blocks have been released.  This patch expands
it to support multiple different callbacks, to allow other things to
be done after the commit has been completed.
Signed-off-by: NBobi Jam <bobijam@whamcloud.com>
Signed-off-by: NAndreas Dilger <adilger@whamcloud.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

journal_unmap_buffer()'s zap_buffer: code clears a lot of buffer head
state ala discard_buffer(), but does not touch _Delay or _Unwritten as
discard_buffer() does.

This can be problematic in some areas of the ext4 code which assume
that if they have found a buffer marked unwritten or delay, then it's
a live one.  Perhaps those spots should check whether it is mapped
as well, but if jbd2 is going to tear down a buffer, let's really
tear it down completely.

Without this I get some fsx failures on sub-page-block filesystems
up until v3.2, at which point 4e96b2db
and 189e868f make the failures go
away, because buried within that large change is some more flag
clearing.  I still think it's worth doing in jbd2, since
->invalidatepage leads here directly, and it's the right place
to clear away these flags.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

In commit 9b90e5e0 I incorrectly reserved the wrong bit for
EXT4_FEATURE_INCOMPAT_INLINEDATA per the discussion on the linux-ext4
list on December 7, 2011.  The codepoint 0x2000 should be used for
EXT4_FEATURE_INCOMPAT_USE_META_CSUM, so INLINEDATA will be assigned
the value 0x8000.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

This patch adds trace_jbd2_drop_transaction and
trace_jbd2_update_superblock_end because there are similar tracepoints
in jbd and they are needed in jbd2 as well.
Reviewed-by: NLukas Czerner <lczerner@redhat.com>
Signed-off-by: NSeiji Aguchi <seiji.aguchi@hds.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Ext4 does not support data journalling with delayed allocation enabled.
We even do not allow to mount the file system with delayed allocation
and data journalling enabled, however it can be set via FS_IOC_SETFLAGS
so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file
system mounted with delayed allocation (default) and that's where
problem arises. The easies way to reproduce this problem is with the
following set of commands:

 mkfs.ext4 /dev/sdd
 mount /dev/sdd /mnt/test1
 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4
 chattr +j /mnt/test1/file
 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc
 chattr -j /mnt/test1/file

Additionally it can be reproduced quite reliably with xfstests 272 and
269. In fact the above reproducer is a part of test 272.

To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if
the file system is mounted with delayed allocation. This can be easily
done by fixing ext4_should_*_data() functions do ignore data journal
flag when delalloc is set (suggested by Ted). We also have to set the
appropriate address space operations for the inode (again, ignoring data
journal flag if delalloc enabled).

Additionally this commit introduces ext4_inode_journal_mode() function
because ext4_should_*_data() has already had a lot of common code and
this change is putting it all into one function so it is easier to
read.

Successfully tested with xfstests in following configurations:

delalloc + data=ordered
delalloc + data=writeback
data=journal
nodelalloc + data=ordered
nodelalloc + data=writeback
nodelalloc + data=journal
Signed-off-by: NLukas Czerner <lczerner@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

In ext4_read_{inode,block}_bitmap() we were setting bitmap_uptodate()
before submitting the buffer for read.  The is bad, since we check
bitmap_uptodate() without locking the buffer, and so if another
process is racing with us, it's possible that they will think the
bitmap is uptodate even though the read has not completed yet,
resulting in inodes and blocks potentially getting allocated more than
once if we get really unlucky.

Addresses-Google-Bug: 2828254
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The function ext4_claim_inode() is only called by one function,
ext4_new_inode(), and by folding the functionality into
ext4_new_inode(), we can remove almost 50 lines of code, and put all
of the logic of allocating a new inode into a single place.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

There are few important bug fixes for LogFS

* tag 'for-linus' of git://github.com/prasad-joshi/logfs_upstream:
  Logfs: Allow NULL block_isbad() methods
  logfs: Grow inode in delete path
  logfs: Free areas before calling generic_shutdown_super()
  logfs: remove useless BUG_ON
  MAINTAINERS: Add Prasad Joshi in LogFS maintiners
  logfs: Propagate page parameter to __logfs_write_inode
  logfs: set superblock shutdown flag after generic sb shutdown
  logfs: take write mutex lock during fsync and sync
  logfs: Prevent memory corruption
  logfs: update page reference count for pined pages

Fix up conflict in fs/logfs/dev_mtd.c due to semantic change in what
"mtd->block_isbad" means in commit f2933e86: "Logfs: Allow NULL
block_isbad() methods" clashing with the abstraction changes in the
commits 7086c19d: "mtd: introduce mtd_block_isbad interface" and
d58b27ed: "logfs: do not use 'mtd->block_isbad' directly".

This resolution takes the semantics from commit f2933e86, and just
makes mtd_block_isbad() return zero (false) if the 'block_isbad'
function is NULL.  But that also means that now "mtd_can_have_bb()"
always returns 0.

Now, "mtd_block_markbad()" will obviously return an error if the
low-level driver doesn't support bad blocks, so this is somewhat
non-symmetric, but it actually makes sense if a NULL "block_isbad"
function is considered to mean "I assume that all my blocks are always
good".