While investigating interactivity problems it was clear that processes
sometimes stall for long periods of times if an attempt is made to
lock a buffer which is undergoing writeback.  It would stall in
a trace looking something like

[<ffffffff811a39de>] __lock_buffer+0x2e/0x30
[<ffffffff8123a60f>] do_get_write_access+0x43f/0x4b0
[<ffffffff8123a7cb>] jbd2_journal_get_write_access+0x2b/0x50
[<ffffffff81220f79>] __ext4_journal_get_write_access+0x39/0x80
[<ffffffff811f3198>] ext4_reserve_inode_write+0x78/0xa0
[<ffffffff811f3209>] ext4_mark_inode_dirty+0x49/0x220
[<ffffffff811f57d1>] ext4_dirty_inode+0x41/0x60
[<ffffffff8119ac3e>] __mark_inode_dirty+0x4e/0x2d0
[<ffffffff8118b9b9>] update_time+0x79/0xc0
[<ffffffff8118ba98>] file_update_time+0x98/0x100
[<ffffffff81110ffc>] __generic_file_aio_write+0x17c/0x3b0
[<ffffffff811112aa>] generic_file_aio_write+0x7a/0xf0
[<ffffffff811ea853>] ext4_file_write+0x83/0xd0
[<ffffffff81172b23>] do_sync_write+0xa3/0xe0
[<ffffffff811731ae>] vfs_write+0xae/0x180
[<ffffffff8117361d>] sys_write+0x4d/0x90
[<ffffffff8159d62d>] system_call_fastpath+0x1a/0x1f
[<ffffffffffffffff>] 0xffffffffffffffff
Signed-off-by: NMel Gorman <mgorman@suse.de>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The jbd2_alloc_handle() function is only called by new_handle().  So
this commit uses kmem_cache_zalloc() instead of
kmem_cache_alloc()/memset().
Signed-off-by: NZheng Liu <wenqing.lz@taobao.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The following race is possible:

[kjournald2]                              other_task
jbd2_journal_commit_transaction()
  j_state = T_FINISHED;
  spin_unlock(&journal->j_list_lock);
                                         ->jbd2_journal_remove_checkpoint()
					   ->jbd2_journal_free_transaction();
					     ->kmem_cache_free(transaction)
  ->j_commit_callback(journal, transaction);
    -> USE_AFTER_FREE

WARNING: at lib/list_debug.c:62 __list_del_entry+0x1c0/0x250()
Hardware name:
list_del corruption. prev->next should be ffff88019a4ec198, but was 6b6b6b6b6b6b6b6b
Modules linked in: cpufreq_ondemand acpi_cpufreq freq_table mperf coretemp kvm_intel kvm crc32c_intel ghash_clmulni_intel microcode sg xhci_hcd button sd_mod crc_t10dif aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul ahci libahci pata_acpi ata_generic dm_mirror dm_region_hash dm_log dm_mod
Pid: 16400, comm: jbd2/dm-1-8 Tainted: G        W    3.8.0-rc3+ #107
Call Trace:
 [<ffffffff8106fb0d>] warn_slowpath_common+0xad/0xf0
 [<ffffffff8106fc06>] warn_slowpath_fmt+0x46/0x50
 [<ffffffff813637e9>] ? ext4_journal_commit_callback+0x99/0xc0
 [<ffffffff8148cae0>] __list_del_entry+0x1c0/0x250
 [<ffffffff813637bf>] ext4_journal_commit_callback+0x6f/0xc0
 [<ffffffff813ca336>] jbd2_journal_commit_transaction+0x23a6/0x2570
 [<ffffffff8108aa42>] ? try_to_del_timer_sync+0x82/0xa0
 [<ffffffff8108b491>] ? del_timer_sync+0x91/0x1e0
 [<ffffffff813d3ecf>] kjournald2+0x19f/0x6a0
 [<ffffffff810ad630>] ? wake_up_bit+0x40/0x40
 [<ffffffff813d3d30>] ? bit_spin_lock+0x80/0x80
 [<ffffffff810ac6be>] kthread+0x10e/0x120
 [<ffffffff810ac5b0>] ? __init_kthread_worker+0x70/0x70
 [<ffffffff818ff6ac>] ret_from_fork+0x7c/0xb0
 [<ffffffff810ac5b0>] ? __init_kthread_worker+0x70/0x70

In order to demonstrace this issue one should mount ext4 with mount -o
discard option on SSD disk.  This makes callback longer and race
window becomes wider.

In order to fix this we should mark transaction as finished only after
callbacks have completed
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

In the case where an inode has a very stale transaction id (tid) in
i_datasync_tid or i_sync_tid, it's possible that after a very large
(2**31) number of transactions, that the tid number space might wrap,
causing tid_geq()'s calculations to fail.

Commit deeeaf13 "jbd2: fix fsync() tid wraparound bug", later modified
by commit e7b04ac0 "jbd2: don't wake kjournald unnecessarily",
attempted to fix this problem, but it only avoided kjournald spinning
forever by fixing the logic in jbd2_log_start_commit().

Unfortunately, in the codepaths in fs/ext4/fsync.c and fs/ext4/inode.c
that might call jbd2_log_start_commit() with a stale tid, those
functions will subsequently call jbd2_log_wait_commit() with the same
stale tid, and then wait for a very long time.  To fix this, we
replace the calls to jbd2_log_start_commit() and
jbd2_log_wait_commit() with a call to a new function,
jbd2_complete_transaction(), which will correctly handle stale tid's.

As a bonus, jbd2_complete_transaction() will avoid locking
j_state_lock for writing unless a commit needs to be started.  This
should have a small (but probably not measurable) improvement for
ext4's scalability.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Reported-by: NBen Hutchings <ben@decadent.org.uk>
Reported-by: NGeorge Barnett <gbarnett@atlassian.com>
Cc: stable@vger.kernel.org

jbd2_journal_dirty_metadata() didn't get a reference to journal_head it
was working with. This is OK in most of the cases since the journal head
should be attached to a transaction but in rare occasions when we are
journalling data, __ext4_journalled_writepage() can race with
jbd2_journal_invalidatepage() stripping buffers from a page and thus
journal head can be freed under hands of jbd2_journal_dirty_metadata().

Fix the problem by getting own journal head reference in
jbd2_journal_dirty_metadata() (and also in jbd2_journal_set_triggers()
which can possibly have the same issue).
Reported-by: NZheng Liu <gnehzuil.liu@gmail.com>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

If start_this_handle() failed handle will be initialized
to ERR_PTR() and can not be dereferenced.

paging request at fffffffffffffff6
IP: [<ffffffff813c073f>] jbd2__journal_start+0x18f/0x290
PGD 200e067 PUD 200f067 PMD 0
Oops: 0000 [#1] SMP
Modules linked in: cpufreq_ondemand acpi_cpufreq freq_table mperf coretemp kvm_intel kvm crc32c_intel ghash_clmulni_intel microcode sg xhci_hcd button sd_mod crc_t10dif aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul ahci libahci pata_acpi ata_generic dm_mirror dm_region_hash dm_log dm_mod
CPU 0 journal commit I/O error

Pid: 2694, comm: fio Not tainted 3.8.0-rc3+ #79                  /DQ67SW
RIP: 0010:[<ffffffff813c073f>]  [<ffffffff813c073f>] jbd2__journal_start+0x18f/0x290
RSP: 0018:ffff880233b8ba58  EFLAGS: 00010292
RAX: 00000000ffffffe2 RBX: ffffffffffffffe2 RCX: 0000000000000006
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffff82128f48
RBP: ffff880233b8ba98 R08: 0000000000000000 R09: ffff88021440a6e0
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

There are multiple reasons to move away from debugfs.  First of all,
we are only using it for a single parameter, and it is much more
complicated to set up (some 30 lines of code compared to 3), and one
more thing that might fail while loading the jbd2 module.

Secondly, as a module paramter it can be specified as a boot option if
jbd2 is built into the kernel, or as a parameter when the module is
loaded, and it can also be manipulated dynamically under
/sys/module/jbd2/parameters/jbd2_debug.  So it is more flexible.

Ultimately we want to move away from using jbd_debug() towards
tracepoints, but for now this is still a useful simplification of the
code base.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Handles which stay open a long time are problematic when it comes time
to close down a transaction so it can be committed.  These tracepoints
will help us determine which ones are the problematic ones, and to
validate whether changes makes things better or worse.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Track the delay between when we first request that the commit begin
and when it actually begins, so we can see how much of a gap exists.
In theory, this should just be the remaining scheduling quantuum of
the thread which requested the commit (assuming it was not a
synchronous operation which triggered the commit request) plus
scheduling overhead; however, it's possible that real time processes
might get in the way of letting the kjournald thread from executing.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Don't send an extra wakeup to kjournald in the case where we
already have the proper target in j_commit_request, i.e. that
transaction has already been requested for commit.

commit deeeaf13 "jbd2: fix fsync() tid wraparound bug" changed
the logic leading to a wakeup, but it caused some extra wakeups
which were found to lead to a measurable performance regression.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
[tytso@mit.edu: reworked check to make it clearer]
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We cannot wait for transaction commit in journal_unmap_buffer()
because we hold page lock which ranks below transaction start.  We
solve the issue by bailing out of journal_unmap_buffer() and
jbd2_journal_invalidatepage() with -EBUSY.  Caller is then responsible
for waiting for transaction commit to finish and try invalidation
again. Since the issue can happen only for page stradding i_size, it
is simple enough to manually call jbd2_journal_invalidatepage() for
such page from ext4_setattr(), check the return value and wait if
necessary.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The following race is possible between start_this_handle() and someone
calling jbd2_journal_flush().

Process A                              Process B
start_this_handle().
  if (journal->j_barrier_count) # false
  if (!journal->j_running_transaction) { #true
    read_unlock(&journal->j_state_lock);
                                       jbd2_journal_lock_updates()
                                       jbd2_journal_flush()
                                         write_lock(&journal->j_state_lock);
                                         if (journal->j_running_transaction) {
                                           # false
                                         ... wait for committing trans ...
                                         write_unlock(&journal->j_state_lock);
    ...
    write_lock(&journal->j_state_lock);
    if (!journal->j_running_transaction) { # true
      jbd2_get_transaction(journal, new_transaction);
    write_unlock(&journal->j_state_lock);
    goto repeat; # eventually blocks on j_barrier_count > 0
                                         ...
                                         J_ASSERT(!journal->j_running_transaction);
                                           # fails

We fix the race by rechecking j_barrier_count after reacquiring j_state_lock
in exclusive mode.

Reported-by: yjwsignal@empal.com
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

"Whether" is misspelled in various comments across the tree; this
fixes them. No code changes.
Signed-off-by: NAdam Buchbinder <adam.buchbinder@gmail.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

ext4_handle_release_buffer() was intended to remove journal
write access from a buffer, but it doesn't actually do anything
at all other than add a BUFFER_TRACE point, but it's not reliably
used for that either.  Remove all the associated dead code.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: NCarlos Maiolino <cmaiolino@redhat.com>

ext4 users of data=journal mode with blocksize < pagesize were
occasionally hitting assertion failure in
jbd2_journal_commit_transaction() checking whether the transaction has
at least as many credits reserved as buffers attached.  The core of the
problem is that when a file gets truncated, buffers that still need
checkpointing or that are attached to the committing transaction are
left with buffer_mapped set. When this happens to buffers beyond i_size
attached to a page stradding i_size, subsequent write extending the file
will see these buffers and as they are mapped (but underlying blocks
were freed) things go awry from here.

The assertion failure just coincidentally (and in this case luckily as
we would start corrupting filesystem) triggers due to journal_head not
being properly cleaned up as well.

We fix the problem by unmapping buffers if possible (in lots of cases we
just need a buffer attached to a transaction as a place holder but it
must not be written out anyway).  And in one case, we just have to bite
the bullet and wait for transaction commit to finish.

CC: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: NJan Kara <jack@suse.cz>

This sequence:

# truncate --size=1g fsfile
# mkfs.ext4 -F fsfile
# mount -o loop,ro fsfile /mnt
# umount /mnt
# dmesg | tail

results in an IO error when unmounting the RO filesystem:

[  318.020828] Buffer I/O error on device loop1, logical block 196608
[  318.027024] lost page write due to I/O error on loop1
[  318.032088] JBD2: Error -5 detected when updating journal superblock for loop1-8.

This was a regression introduced by commit 24bcc89c: "jbd2: split
updating of journal superblock and marking journal empty".
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org

blkdev_issue_flush() can fail; make sure the error gets properly
propagated.

This is a port of the equivalent jbd patch from commit 349ecd6a.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

After we transfer set the EXT4_ERROR_FS bit in the file system
superblock, it's not enough to call jbd2_journal_clear_err() to clear
the error indication from journal superblock --- we need to call
jbd2_journal_update_sb_errno() as well.  Otherwise, when the root file
system is mounted read-only, the journal is replayed, and the error
indicator is transferred to the superblock --- but the s_errno field
in the jbd2 superblock is left set (since although we cleared it in
memory, we never flushed it out to disk).

This can end up confusing e2fsck.  We should make e2fsck more robust
in this case, but the kernel shouldn't be leaving things in this
confused state, either.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org

The '->write_super' superblock method is gone, and this patch removes all the
references to 'write_super' from various jbd and jbd2.

Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jan Kara <jack@suse.cz>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NArtem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NCong Wang <amwang@redhat.com>

Use kmem_cache_zalloc wrapper instead of flag __GFP_ZERO.
Signed-off-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Calculate and verify checksums of each data block being stored in the journal.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Calculate and verify the checksum of commit blocks.  In checksum v2,
deprecate most of the checksum v1 commit block checksum fields, since
each block has its own checksum.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Calculate and verify a checksum of each descriptor block.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Compute and verify revoke blocks inside the journal.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Calculate and verify a checksum covering the journal superblock.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Obtain a reference to the crc32c driver if needed for the v2 checksum.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Add in the necessary code so that journal clients can enable the new
journal checksumming features.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Define flags and allocate space in on-disk journal structures to support
checksumming of journal metadata.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

flush request is issued in transaction commit code path, so looks using
GFP_KERNEL to allocate memory for flush request bio falls into the classic
deadlock issue.  I saw btrfs and dm get it right, but ext4, xfs and md are
using GFP.
Signed-off-by: NShaohua Li <shli@fusionio.com>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Reviewed-by: NJan Kara <jack@suse.cz>
Cc: stable@vger.kernel.org

Remove all #inclusions of asm/system.h preparatory to splitting and killing
it.  Performed with the following command:

perl -p -i -e 's!^#\s*include\s*<asm/system[.]h>.*\n!!' `grep -Irl '^#\s*include\s*<asm/system[.]h>' *`
Signed-off-by: NDavid Howells <dhowells@redhat.com>

Signed-off-by: NCong Wang <amwang@redhat.com>

Normally, we have to issue a cache flush before we can update journal tail in
journal superblock, effectively wiping out old transactions from the journal.
So use the fact that during transaction commit we issue cache flush anyway and
opportunistically push journal tail as far as we can. Since update of journal
superblock is still costly (we have to use WRITE_FUA), we update log tail only
if we can free significant amount of space.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

All accesses to checkpointing entries in journal_head are protected
by j_list_lock. Thus __jbd2_journal_remove_checkpoint() doesn't really
need bh_state lock.

Also the only part of journal head that the rest of checkpointing code
needs to check is jh->b_transaction which is safe to read under
j_list_lock.

So we can safely remove bh_state lock from all of checkpointing code which
makes it considerably prettier.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The check b_jlist == BJ_None in __journal_try_to_free_buffer() is
always true (__jbd2_journal_temp_unlink_buffer() also checks this in
an assertion) so just remove it.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

BH_JWrite bit should be set when buffer is written to the journal. So
checkpointing shouldn't set this bit when writing out buffer. This didn't
cause any observable bug since BH_JWrite bit is used only for debugging
purposes but it's good to have this consistent.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

When we reach jbd2_cleanup_journal_tail(), there is no guarantee that
checkpointed buffers are on a stable storage - especially if buffers were
written out by jbd2_log_do_checkpoint(), they are likely to be only in disk's
caches. Thus when we update journal superblock effectively removing old
transaction from journal, this write of superblock can get to stable storage
before those checkpointed buffers which can result in filesystem corruption
after a crash. Thus we must unconditionally issue a cache flush before we
update journal superblock in these cases.

A similar problem can also occur if journal superblock is written only in
disk's caches, other transaction starts reusing space of the transaction
cleaned from the log and power failure happens. Subsequent journal replay would
still try to replay the old transaction but some of it's blocks may be already
overwritten by the new transaction. For this reason we must use WRITE_FUA when
updating log tail and we must first write new log tail to disk and update
in-memory information only after that.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

With the latest and greatest changes to the freezer, I started seeing
panics that were caused by jbd2 running post-process freezing and
hitting the canary BUG_ON for non-TuxOnIce I/O submission. I've traced
this back to a lack of set_freezable calls in both jbd and jbd2. Since
they're clearly meant to be frozen (there are tests for freezing()), I
submit the following patch to add the missing calls.
Signed-off-by: NNigel Cunningham <nigel@tuxonice.net>
Acked-by: NJan Kara <jack@suse.cz>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

There are some log tail updates that are not protected by j_checkpoint_mutex.
Some of these are harmless because they happen during startup or shutdown but
updates in jbd2_journal_commit_transaction() and jbd2_journal_flush() can
really race with other log tail updates (e.g. someone doing
jbd2_journal_flush() with someone running jbd2_cleanup_journal_tail()). So
protect all log tail updates with j_checkpoint_mutex.
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>