jbd2_alloc() allocates a buffer from slab when the block size is smaller
than PAGE_SIZE, and slab may be using a compound page. Before commit
8147c4c454, we set b_page to the precise page containing the buffer
and this code worked well. Now we set b_page to the head page of the
allocation, so we can no longer use offset_in_page(). While we could
do a 1:1 replacement with offset_in_folio(), use the more idiomatic
bh_offset() and the folio APIs to map the buffer.
This isn't enough to support a b_size larger than PAGE_SIZE on HIGHMEM
machines, but this is good enough to fix the actual bug we're seeing.
Fixes: 8147c4c454 ("jbd2: use a folio in jbd2_journal_write_metadata_buffer()")
Reported-by: Zorro Lang <zlang@kernel.org>
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
[converted to be more folio]
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
__journal_try_to_free_buffer() has only one caller and it's logic is
much simple now, so just remove it and open code in
jbd2_journal_try_to_free_buffers().
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230606135928.434610-7-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Before removing checkpoint buffer from the t_checkpoint_list, we have to
check both BH_Dirty and BH_Lock bits together to distinguish buffers
have not been or were being written back. But __cp_buffer_busy() checks
them separately, it first check lock state and then check dirty, the
window between these two checks could be raced by writing back
procedure, which locks buffer and clears buffer dirty before I/O
completes. So it cannot guarantee checkpointing buffers been written
back to disk if some error happens later. Finally, it may clean
checkpoint transactions and lead to inconsistent filesystem.
jbd2_journal_forget() and __journal_try_to_free_buffer() also have the
same problem (journal_unmap_buffer() escape from this issue since it's
running under the buffer lock), so fix them through introducing a new
helper to try holding the buffer lock and remove really clean buffer.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217490
Cc: stable@vger.kernel.org
Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230606135928.434610-6-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When invalidating buffers under the partial tail page,
jbd2_journal_invalidate_folio() returns -EBUSY if the buffer is part of
the committing transaction as we cannot safely modify buffer state.
However if the buffer is already invalidated (due to previous
invalidation attempts from ext4_wait_for_tail_page_commit()), there's
nothing to do and there's no point in returning -EBUSY. This fixes
occasional warnings from ext4_journalled_invalidate_folio() triggered by
generic/051 fstest when blocksize < pagesize.
Fixes: 53e872681f ("ext4: fix deadlock in journal_unmap_buffer()")
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230329154950.19720-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Following process will make data lost and could lead to a filesystem
corrupted problem:
1. jh(bh) is inserted into T1->t_checkpoint_list, bh is dirty, and
jh->b_transaction = NULL
2. T1 is added into journal->j_checkpoint_transactions.
3. Get bh prepare to write while doing checkpoing:
PA PB
do_get_write_access jbd2_log_do_checkpoint
spin_lock(&jh->b_state_lock)
if (buffer_dirty(bh))
clear_buffer_dirty(bh) // clear buffer dirty
set_buffer_jbddirty(bh)
transaction =
journal->j_checkpoint_transactions
jh = transaction->t_checkpoint_list
if (!buffer_dirty(bh))
__jbd2_journal_remove_checkpoint(jh)
// bh won't be flushed
jbd2_cleanup_journal_tail
__jbd2_journal_file_buffer(jh, transaction, BJ_Reserved)
4. Aborting journal/Power-cut before writing latest bh on journal area.
In this way we get a corrupted filesystem with bh's data lost.
Fix it by moving the clearing of buffer_dirty bit just before the call
to __jbd2_journal_file_buffer(), both bit clearing and jh->b_transaction
assignment are under journal->j_list_lock locked, so that
jbd2_log_do_checkpoint() will wait until jh's new transaction fininshed
even bh is currently not dirty. And journal_shrink_one_cp_list() won't
remove jh from checkpoint list if the buffer head is reused in
do_get_write_access().
Fetch a reproducer in [Link].
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216898
Cc: <stable@kernel.org>
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: zhanchengbin <zhanchengbin1@huawei.com>
Suggested-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230110015327.1181863-1-chengzhihao1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
LIFO wakeup order is unfair and sometimes leads to a journal
user not being able to get a journal handle for hundreds of
transactions in a row.
FIFO wakeup can make things more fair.
Cc: stable@kernel.org
Signed-off-by: Alexey Lyashkov <alexey.lyashkov@gmail.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20220907165959.1137482-1-alexey.lyashkov@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The name of jbd_debug() is confusing as all functions inside jbd2 have
jbd2_ prefix. Rename jbd_debug() to jbd2_debug(). No functional changes.
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
Link: https://lore.kernel.org/r/20220608112355.4397-2-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Add the description of @folio and remove @page in function kernel-doc
comment to remove warnings found by running scripts/kernel-doc, which
is caused by using 'make W=1'.
fs/jbd2/transaction.c:2149: warning: Function parameter or member
'folio' not described in 'jbd2_journal_try_to_free_buffers'
fs/jbd2/transaction.c:2149: warning: Excess function parameter 'page'
description in 'jbd2_journal_try_to_free_buffers'
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220512075432.31763-1-yang.lee@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
All but two of the callers already have a folio; pass a folio into
try_to_free_buffers(). This removes the last user of cancel_dirty_page()
so remove that wrapper function too.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Also convert it to return a bool since it's called from release_folio().
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Primarily this series converts some of the address_space operations
to take a folio instead of a page.
->is_partially_uptodate() takes a folio instead of a page and changes the
type of the 'from' and 'count' arguments to make it obvious they're bytes.
->invalidatepage() becomes ->invalidate_folio() and has a similar type change.
->launder_page() becomes ->launder_folio()
->set_page_dirty() becomes ->dirty_folio() and adds the address_space as
an argument.
There are a couple of other misc changes up front that weren't worth
separating into their own pull request.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEEejHryeLBw/spnjHrDpNsjXcpgj4FAmI4hqMACgkQDpNsjXcp
gj7r7Af/fVJ7m8kKqjP/IayX3HiJRuIDQw+vM++BlRNXdjz+IyED6whdmFGxJeOY
BMyT+8ApOAz7ErS4G+7fAv4ScJK/aEgFUsnSeAiCp0PliiEJ5NNJzElp6sVmQ7H5
SX7+Ek444FZUGsQuy0qL7/ELpR3ditnD7x+5U2g0p5TeaHGUQn84crRyfR4xuhNG
EBD9D71BOb7OxUcOHe93pTkK51QsQ0aCrcIsB1tkK5KR0BAthn1HqF7ehL90Rvrr
omx5M7aDWGY4oj7IKrhlAs+55Ah2WaOzrZBp0FXNbr4UENDBKWKyUxErwa4xPkf6
Gm1iQG/CspOHnxN3YWsd5WjtlL3A+A==
=cOiq
-----END PGP SIGNATURE-----
Merge tag 'folio-5.18b' of git://git.infradead.org/users/willy/pagecache
Pull filesystem folio updates from Matthew Wilcox:
"Primarily this series converts some of the address_space operations to
take a folio instead of a page.
Notably:
- a_ops->is_partially_uptodate() takes a folio instead of a page and
changes the type of the 'from' and 'count' arguments to make it
obvious they're bytes.
- a_ops->invalidatepage() becomes ->invalidate_folio() and has a
similar type change.
- a_ops->launder_page() becomes ->launder_folio()
- a_ops->set_page_dirty() becomes ->dirty_folio() and adds the
address_space as an argument.
There are a couple of other misc changes up front that weren't worth
separating into their own pull request"
* tag 'folio-5.18b' of git://git.infradead.org/users/willy/pagecache: (53 commits)
fs: Remove aops ->set_page_dirty
fb_defio: Use noop_dirty_folio()
fs: Convert __set_page_dirty_no_writeback to noop_dirty_folio
fs: Convert __set_page_dirty_buffers to block_dirty_folio
nilfs: Convert nilfs_set_page_dirty() to nilfs_dirty_folio()
mm: Convert swap_set_page_dirty() to swap_dirty_folio()
ubifs: Convert ubifs_set_page_dirty to ubifs_dirty_folio
f2fs: Convert f2fs_set_node_page_dirty to f2fs_dirty_node_folio
f2fs: Convert f2fs_set_data_page_dirty to f2fs_dirty_data_folio
f2fs: Convert f2fs_set_meta_page_dirty to f2fs_dirty_meta_folio
afs: Convert afs_dir_set_page_dirty() to afs_dir_dirty_folio()
btrfs: Convert extent_range_redirty_for_io() to use folios
fs: Convert trivial uses of __set_page_dirty_nobuffers to filemap_dirty_folio
btrfs: Convert from set_page_dirty to dirty_folio
fscache: Convert fscache_set_page_dirty() to fscache_dirty_folio()
fs: Add aops->dirty_folio
fs: Remove aops->launder_page
orangefs: Convert launder_page to launder_folio
nfs: Convert from launder_page to launder_folio
fuse: Convert from launder_page to launder_folio
...
Extensive changes, but fairly mechanical.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Tested-by: Mike Marshall <hubcap@omnibond.com> # orangefs
Tested-by: David Howells <dhowells@redhat.com> # afs
CONFIG_JBD2_DEBUG and jbd2_journal_enable_debug knobs were added in
update_t_max_wait(), since earlier it used to take a spinlock for
updating t_max_wait, which could cause a bottleneck while starting a
txn (start_this_handle()).
Since in previous patch, we have killed t_handle_lock completely, we
could get rid of this debug config and knob to let t_max_wait be
updated by default again.
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/ad7319a601fd501079310747ce87d908e0944763.1644992076.git.riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch kills t_handle_lock transaction spinlock completely from
jbd2.
To explain the reasoning, currently there were three sites at which
this spinlock was used.
1. jbd2_journal_wait_updates()
a. Based on careful code review it can be seen that, we don't need this
lock here. This is since we wait for any currently ongoing updates
based on a atomic variable t_updates. And we anyway don't take any
t_handle_lock while in stop_this_handle().
i.e.
write_lock(&journal->j_state_lock()
jbd2_journal_wait_updates() stop_this_handle()
while (atomic_read(txn->t_updates) { |
DEFINE_WAIT(wait); |
prepare_to_wait(); |
if (atomic_read(txn->t_updates) if (atomic_dec_and_test(txn->t_updates))
write_unlock(&journal->j_state_lock);
schedule(); wake_up()
write_lock(&journal->j_state_lock);
finish_wait();
}
txn->t_state = T_COMMIT
write_unlock(&journal->j_state_lock);
b. Also note that between atomic_inc(&txn->t_updates) in
start_this_handle() and jbd2_journal_wait_updates(), the
synchronization happens via read_lock(journal->j_state_lock) in
start_this_handle();
2. jbd2_journal_extend()
a. jbd2_journal_extend() is called with the handle of each process from
task_struct. So no lock required in updating member fields of handle_t
b. For member fields of h_transaction, all updates happens only via
atomic APIs (which is also within read_lock()).
So, no need of this transaction spinlock.
3. update_t_max_wait()
Based on Jan suggestion, this can be carefully removed using atomic
cmpxchg API.
Note that there can be several processes which are waiting for a new
transaction to be allocated and started. For doing this only one
process will succeed in taking write_lock() and allocating a new txn.
After that all of the process will be updating the t_max_wait (max
transaction wait time). This can be done via below method w/o taking
any locks using atomic cmpxchg.
For more details refer [1]
new = get_new_val();
old = READ_ONCE(ptr->max_val);
while (old < new)
old = cmpxchg(&ptr->max_val, old, new);
[1]: https://lwn.net/Articles/849237/
Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/d89e599658b4a1f3893a48c6feded200073037fc.1644992076.git.riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2_journal_wait_updates() is called with j_state_lock held. But if
there is a commit in progress, then this transaction might get committed
and freed via jbd2_journal_commit_transaction() ->
jbd2_journal_free_transaction(), when we release j_state_lock.
So check for journal->j_running_transaction everytime we release and
acquire j_state_lock to avoid use-after-free issue.
Link: https://lore.kernel.org/r/948c2fed518ae739db6a8f7f83f1d58b504f87d0.1644497105.git.ritesh.list@gmail.com
Fixes: 4f98186848 ("jbd2: refactor wait logic for transaction updates into a common function")
Cc: stable@kernel.org
Reported-and-tested-by: syzbot+afa2ca5171d93e44b348@syzkaller.appspotmail.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
No functionality change as such in this patch. This only refactors the
common piece of code which waits for t_updates to finish into a common
function named as jbd2_journal_wait_updates(journal_t *)
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/8c564f70f4b2591171677a2a74fccb22a7b6c3a4.1642416995.git.riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
JBD2 layer support triggers which are called when journaling layer moves
buffer to a certain state. We can use the frozen trigger, which gets
called when buffer data is frozen and about to be written out to the
journal, to compute block checksums for some buffer types (similarly as
does ocfs2). This avoids unnecessary repeated recomputation of the
checksum (at the cost of larger window where memory corruption won't be
caught by checksumming) and is even necessary when there are
unsynchronized updaters of the checksummed data.
So add superblock and journal trigger type arguments to
ext4_journal_get_write_access() and ext4_journal_get_create_access() so
that frozen triggers can be set accordingly. Also add inode argument to
ext4_walk_page_buffers() and all the callbacks used with that function
for the same purpose. This patch is mostly only a change of prototype of
the above mentioned functions and a few small helpers. Real checksumming
will come later.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20210816095713.16537-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Now that we can be sure the journal is aborted once a buffer has failed
to be written back to disk, we can remove the journal abort logic in
jbd2_journal_try_to_free_buffers() which was introduced in
commit c044f3d836 ("jbd2: abort journal if free a async write error
metadata buffer"), because it may cost and propably is not safe.
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20210610112440.3438139-4-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Assertion checks in jbd2_journal_dirty_metadata() are known to be racy
but we don't want to be grabbing locks just for them. We thus recheck
them under b_state_lock only if it looks like they would fail. Annotate
the checks with data_race().
Cc: stable@kernel.org
Reported-by: Hao Sun <sunhao.th@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20210406161804.20150-2-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Kernel-doc markup should use this format:
identifier - description
They should not have any type before that, as otherwise
the parser won't do the right thing.
Also, some identifiers have different names between their
prototypes and the kernel-doc markup.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Link: https://lore.kernel.org/r/72f5c6628f5f278d67625f60893ffbc2ca28d46e.1605521731.git.mchehab+huawei@kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Add missing __acquires() and __releases() annotations. Also, in an
"this should never happen" WARN_ON check, if it *does* actually
happen, we need to release j_state_lock since this function is always
supposed to release that lock. Otherwise, things will quickly grind
to a halt after the WARN_ON trips.
Fixes: 96f1e09745 ("jbd2: avoid long hold times of j_state_lock...")
Cc: stable@kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Parameter gfp_mask in jbd2_journal_try_to_free_buffers() is no longer
used after commit <536fc240e7147> ("jbd2: clean up
jbd2_journal_try_to_free_buffers()"), so just remove it.
Signed-off-by: zhangyi (F) <yi.zhang@huawei.com>
Link: https://lore.kernel.org/r/20200620025427.1756360-6-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If we free a metadata buffer which has been failed to async write out
in the background, the jbd2 checkpoint procedure will not detect this
failure in jbd2_log_do_checkpoint(), so it may lead to filesystem
inconsistency after cleanup journal tail. This patch abort the journal
if free a buffer has write_io_error flag to prevent potential further
inconsistency.
Signed-off-by: zhangyi (F) <yi.zhang@huawei.com>
Link: https://lore.kernel.org/r/20200620025427.1756360-5-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Callers of __jbd2_journal_unfile_buffer() and
__jbd2_journal_refile_buffer() assume that the b_transaction is set. In
fact if it's not, we can end up with journal_head refcounting errors
leading to crash much later that might be very hard to track down. Add
asserts to make sure that is the case.
We also make sure that b_next_transaction is NULL in
__jbd2_journal_unfile_buffer() since the callers expect that as well and
we should not get into that stage in this state anyway, leading to
problems later on if we do.
Tested with fstests.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20200617092549.6712-1-lczerner@redhat.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When reserved transaction handle is unused, we subtract its reserved
credits in __jbd2_journal_unreserve_handle() called from
jbd2_journal_stop(). However this function forgets to remove reserved
credits from transaction->t_outstanding_credits and thus the transaction
space that was reserved remains effectively leaked. The leaked
transaction space can be quite significant in some cases and leads to
unnecessarily small transactions and thus reducing throughput of the
journalling machinery. E.g. fsmark workload creating lots of 4k files
was observed to have about 20% lower throughput due to this when ext4 is
mounted with dioread_nolock mount option.
Subtract reserved credits from t_outstanding_credits as well.
CC: stable@vger.kernel.org
Fixes: 8f7d89f368 ("jbd2: transaction reservation support")
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20200520133119.1383-3-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
journal_head::b_transaction and journal_head::b_next_transaction could
be accessed concurrently as noticed by KCSAN,
LTP: starting fsync04
/dev/zero: Can't open blockdev
EXT4-fs (loop0): mounting ext3 file system using the ext4 subsystem
EXT4-fs (loop0): mounted filesystem with ordered data mode. Opts: (null)
==================================================================
BUG: KCSAN: data-race in __jbd2_journal_refile_buffer [jbd2] / jbd2_write_access_granted [jbd2]
write to 0xffff99f9b1bd0e30 of 8 bytes by task 25721 on cpu 70:
__jbd2_journal_refile_buffer+0xdd/0x210 [jbd2]
__jbd2_journal_refile_buffer at fs/jbd2/transaction.c:2569
jbd2_journal_commit_transaction+0x2d15/0x3f20 [jbd2]
(inlined by) jbd2_journal_commit_transaction at fs/jbd2/commit.c:1034
kjournald2+0x13b/0x450 [jbd2]
kthread+0x1cd/0x1f0
ret_from_fork+0x27/0x50
read to 0xffff99f9b1bd0e30 of 8 bytes by task 25724 on cpu 68:
jbd2_write_access_granted+0x1b2/0x250 [jbd2]
jbd2_write_access_granted at fs/jbd2/transaction.c:1155
jbd2_journal_get_write_access+0x2c/0x60 [jbd2]
__ext4_journal_get_write_access+0x50/0x90 [ext4]
ext4_mb_mark_diskspace_used+0x158/0x620 [ext4]
ext4_mb_new_blocks+0x54f/0xca0 [ext4]
ext4_ind_map_blocks+0xc79/0x1b40 [ext4]
ext4_map_blocks+0x3b4/0x950 [ext4]
_ext4_get_block+0xfc/0x270 [ext4]
ext4_get_block+0x3b/0x50 [ext4]
__block_write_begin_int+0x22e/0xae0
__block_write_begin+0x39/0x50
ext4_write_begin+0x388/0xb50 [ext4]
generic_perform_write+0x15d/0x290
ext4_buffered_write_iter+0x11f/0x210 [ext4]
ext4_file_write_iter+0xce/0x9e0 [ext4]
new_sync_write+0x29c/0x3b0
__vfs_write+0x92/0xa0
vfs_write+0x103/0x260
ksys_write+0x9d/0x130
__x64_sys_write+0x4c/0x60
do_syscall_64+0x91/0xb05
entry_SYSCALL_64_after_hwframe+0x49/0xbe
5 locks held by fsync04/25724:
#0: ffff99f9911093f8 (sb_writers#13){.+.+}, at: vfs_write+0x21c/0x260
#1: ffff99f9db4c0348 (&sb->s_type->i_mutex_key#15){+.+.}, at: ext4_buffered_write_iter+0x65/0x210 [ext4]
#2: ffff99f5e7dfcf58 (jbd2_handle){++++}, at: start_this_handle+0x1c1/0x9d0 [jbd2]
#3: ffff99f9db4c0168 (&ei->i_data_sem){++++}, at: ext4_map_blocks+0x176/0x950 [ext4]
#4: ffffffff99086b40 (rcu_read_lock){....}, at: jbd2_write_access_granted+0x4e/0x250 [jbd2]
irq event stamp: 1407125
hardirqs last enabled at (1407125): [<ffffffff980da9b7>] __find_get_block+0x107/0x790
hardirqs last disabled at (1407124): [<ffffffff980da8f9>] __find_get_block+0x49/0x790
softirqs last enabled at (1405528): [<ffffffff98a0034c>] __do_softirq+0x34c/0x57c
softirqs last disabled at (1405521): [<ffffffff97cc67a2>] irq_exit+0xa2/0xc0
Reported by Kernel Concurrency Sanitizer on:
CPU: 68 PID: 25724 Comm: fsync04 Tainted: G L 5.6.0-rc2-next-20200221+ #7
Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019
The plain reads are outside of jh->b_state_lock critical section which result
in data races. Fix them by adding pairs of READ|WRITE_ONCE().
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Qian Cai <cai@lca.pw>
Link: https://lore.kernel.org/r/20200222043111.2227-1-cai@lca.pw
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
There is no need to delay the clearing of b_modified flag to the
transaction committing time when unmapping the journalled buffer, so
just move it to the journal_unmap_buffer().
Link: https://lore.kernel.org/r/20200213063821.30455-2-yi.zhang@huawei.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: zhangyi (F) <yi.zhang@huawei.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
* Direct I/O via iomap (required the iomap-for-next branch from Darrick
as a prereq).
* Support for using dioread-nolock where the block size < page size.
* Support for encryption for file systems where the block size < page size.
* Rework of journal credits handling so a revoke-heavy workload will
not cause the journal to run out of space.
* Replace bit-spinlocks with spinlocks in jbd2
Also included were some bug fixes and cleanups, mostly to clean up
corner cases from fuzzed file systems and error path handling.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEK2m5VNv+CHkogTfJ8vlZVpUNgaMFAl3dHxoACgkQ8vlZVpUN
gaMZswf5AbtQhTEJDXO7Pc1ull38GIGFgAv7uAth0TymLC3h1/FEYWW0crEPFsDr
1Eei55UUVOYrMMUKQ4P7wlLX0cIh3XDPMWnRFuqBoV5/ZOsH/ZSbkY//TG2Xze/v
9wXIH/RKQnzbRtXffJ1+DnvmXJk+HFm1R1gjl0nfyUXGrnlSfqJxhLSczyd6bJJq
ehi/tso5UC/4EQsAIdWp7VWsAdaHcZ7ogHqDoy8dXpM1equ408iml7VlKr8R+Nr7
5ANpCISXChSlLLYm0NYN5vhO8upF5uDxWLdCtxVPL5kFdM2m/ELjXw9h9C+78l7C
EWJGlGlxvx07Px+e+bfStEsoixpWBg==
=0eko
-----END PGP SIGNATURE-----
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 updates from Ted Ts'o:
"This merge window saw the the following new featuers added to ext4:
- Direct I/O via iomap (required the iomap-for-next branch from
Darrick as a prereq).
- Support for using dioread-nolock where the block size < page size.
- Support for encryption for file systems where the block size < page
size.
- Rework of journal credits handling so a revoke-heavy workload will
not cause the journal to run out of space.
- Replace bit-spinlocks with spinlocks in jbd2
Also included were some bug fixes and cleanups, mostly to clean up
corner cases from fuzzed file systems and error path handling"
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (59 commits)
ext4: work around deleting a file with i_nlink == 0 safely
ext4: add more paranoia checking in ext4_expand_extra_isize handling
jbd2: make jbd2_handle_buffer_credits() handle reserved handles
ext4: fix a bug in ext4_wait_for_tail_page_commit
ext4: bio_alloc with __GFP_DIRECT_RECLAIM never fails
ext4: code cleanup for get_next_id
ext4: fix leak of quota reservations
ext4: remove unused variable warning in parse_options()
ext4: Enable encryption for subpage-sized blocks
fs/buffer.c: support fscrypt in block_read_full_page()
ext4: Add error handling for io_end_vec struct allocation
jbd2: Fine tune estimate of necessary descriptor blocks
jbd2: Provide trace event for handle restarts
ext4: Reserve revoke credits for freed blocks
jbd2: Make credit checking more strict
jbd2: Rename h_buffer_credits to h_total_credits
jbd2: Reserve space for revoke descriptor blocks
jbd2: Drop jbd2_space_needed()
jbd2: Account descriptor blocks into t_outstanding_credits
jbd2: Factor out common parts of stopping and restarting a handle
...
Currently we reserve j_max_transaction_buffers / 32 for transaction
descriptor blocks. Now that revoke descriptors are accounted for
separately this estimate is unnecessarily high and we can actually
compute much tighter estimate. In the common case of 32k journal blocks
and 4k blocksize this actually reduces the amount of reserved descriptor
blocks from 256 to ~25 which allows us to fit more real data into a
transaction.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-25-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Provide trace event for handle restarts to ease debugging.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-24-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Make checking of available credits in jbd2_journal_dirty_metadata() more
strict. There should be always enough credits in the handle to write all
potential revoke descriptors. Also we warn in case there are not enough
credits since this is a bug in the filesystem.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-22-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The credit counter now contains both buffer and revoke descriptor block
credits. Rename to counter to h_total_credits to reflect that. No
functional change.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-21-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Extend functions for starting, extending, and restarting transaction
handles to take number of revoke records handle must be able to
accommodate. These functions then make sure transaction has enough
credits to be able to store resulting revoke descriptor blocks. Also
revoke code tracks number of revoke records created by a handle to catch
situation where some place didn't reserve enough space for revoke
records. Similarly to standard transaction credits, space for unused
reserved revoke records is released when the handle is stopped.
On the ext4 side we currently take a simplistic approach of reserving
space for 1024 revoke records for any transaction. This grows amount of
credits reserved for each handle only by a few and is enough for any
normal workload so that we don't hit warnings in jbd2. We will refine
the logic in following commits.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-20-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The function is now just a trivial wrapper returning
journal->j_max_transaction_buffers. Drop it.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-19-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, journal descriptor blocks were not accounted in
transaction->t_outstanding_credits and we were just leaving some slack
space in the journal for them (in jbd2_log_space_left() and
jbd2_space_needed()). This is making proper accounting (and reservation
we want to add) of descriptor blocks difficult so switch to accounting
descriptor blocks in transaction->t_outstanding_credits and just reserve
the same amount of credits in t_outstanding credits for journal
descriptor blocks when creating transaction.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-18-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2__journal_restart() has quite some code that is common with
jbd2_journal_stop(). Factor this functionality into stop_this_handle()
helper and use it from both functions. Note that this also drops
t_handle_lock protection from jbd2__journal_restart() as
jbd2_journal_stop() does the same thing without it.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-17-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When we drop last handle from a transaction and journal->j_barrier_count
> 0, jbd2_journal_stop() wakes up journal->j_wait_transaction_locked
wait queue. This looks pointless - wait for outstanding handles always
happens on journal->j_wait_updates waitqueue.
journal->j_wait_transaction_locked is used to wait for transaction state
changes and by start_this_handle() for waiting until
journal->j_barrier_count drops to 0. The first case is clearly
irrelevant here since only jbd2 thread changes transaction state. The
second case looks related but jbd2_journal_unlock_updates() is
responsible for the wakeup in this case. So just drop the wakeup.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-16-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If a transaction is larger than journal->j_max_transaction_buffers, that
is a bug and not a trigger for transaction commit. Also the very next
attempt to start new handle will start transaction commit anyway. So
just remove the pointless check. Arguably, we could start transaction
commit whenever the transaction size is *close* to
journal->j_max_transaction_buffers. This has a potential to reduce
latency of the next jbd2_journal_start() at the cost of somewhat smaller
transactions. However for this to have any effect, it would mean that
there isn't someone already waiting in jbd2_journal_start() which means
metadata load for the fs is pretty light anyway so probably this
optimization is not worth it.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-15-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Move code in jbd2_journal_stop() around a bit. It removes some
unnecessary code duplication and will make factoring out parts common
with jbd2__journal_restart() easier.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-14-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Bit-spinlocks are problematic on PREEMPT_RT if functions which might sleep
on RT, e.g. spin_lock(), alloc/free(), are invoked inside the lock held
region because bit spinlocks disable preemption even on RT.
A first attempt was to replace state lock with a spinlock placed in struct
buffer_head and make the locking conditional on PREEMPT_RT and
DEBUG_BIT_SPINLOCKS.
Jan pointed out that there is a 4 byte hole in struct journal_head where a
regular spinlock fits in and he would not object to convert the state lock
to a spinlock unconditionally.
Aside of solving the RT problem, this also gains lockdep coverage for the
journal head state lock (bit-spinlocks are not covered by lockdep as it's
hard to fit a lockdep map into a single bit).
The trivial change would have been to convert the jbd_*lock_bh_state()
inlines, but that comes with the downside that these functions take a
buffer head pointer which needs to be converted to a journal head pointer
which adds another level of indirection.
As almost all functions which use this lock have a journal head pointer
readily available, it makes more sense to remove the lock helper inlines
and write out spin_*lock() at all call sites.
Fixup all locking comments as well.
Suggested-by: Jan Kara <jack@suse.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jan Kara <jack@suse.com>
Cc: linux-ext4@vger.kernel.org
Link: https://lore.kernel.org/r/20190809124233.13277-7-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2_journal_forget() jumps to 'not_jbd' branch which calls __bforget()
in cases where the buffer is clean which is pointless. In case of failed
assertion, it can be even argued that it is safer not to touch buffer's
dirty bits. Also logically it makes more sense to just jump to 'drop'
and that will make logic also simpler when we switch bh_state_lock to a
spinlock.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20190809124233.13277-6-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
We have cleared both dirty & jbddirty bits from the bh. So there's no
difference between bforget() and brelse(). Thus there's no point jumping
to no_jbd branch.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20190809124233.13277-5-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
__jbd2_journal_unfile_buffer() and __jbd2_journal_refile_buffer() drop
transaction's jh reference when they remove jh from a transaction. This
will be however inconvenient once we move state lock into journal_head
itself as we still need to unlock it and we'd need to grab jh reference
just for that. Move dropping of jh reference out of these functions into
the few callers.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20190809124233.13277-4-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
