linux-stable.git/fs/jbd2/commit.c, branch v5.4.239

jbd2: wake up journal waiters in FIFO order, not LIFO

2022-10-26T11:22:17+00:00

commit 34fc8768ec6089565d6d73bad26724083cecf7bd upstream.

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 
Reviewed-by: Ritesh Harjani (IBM) 
Link: https://lore.kernel.org/r/20220907165959.1137482-1-alexey.lyashkov@gmail.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Greg Kroah-Hartman

jbd2: fix outstanding credits assert in jbd2_journal_commit_transaction()

2022-08-25T09:17:59+00:00

[ Upstream commit a89573ce4ad32f19f43ec669771726817e185be0 ]

We catch an assert problem in jbd2_journal_commit_transaction() when
doing fsstress and request falut injection tests. The problem is
happened in a race condition between jbd2_journal_commit_transaction()
and ext4_end_io_end(). Firstly, ext4_writepages() writeback dirty pages
and start reserved handle, and then the journal was aborted due to some
previous metadata IO error, jbd2_journal_abort() start to commit current
running transaction, the committing procedure could be raced by
ext4_end_io_end() and lead to subtract j_reserved_credits twice from
commit_transaction->t_outstanding_credits, finally the
t_outstanding_credits is mistakenly smaller than t_nr_buffers and
trigger assert.

kjournald2           kworker

jbd2_journal_commit_transaction()
 write_unlock(&journal->j_state_lock);
 atomic_sub(j_reserved_credits, t_outstanding_credits); //sub once

     	             jbd2_journal_start_reserved()
     	              start_this_handle()  //detect aborted journal
     	              jbd2_journal_free_reserved()  //get running transaction
                       read_lock(&journal->j_state_lock)
     	                __jbd2_journal_unreserve_handle()
     	               atomic_sub(j_reserved_credits, t_outstanding_credits);
                       //sub again
                       read_unlock(&journal->j_state_lock);

 journal->j_running_transaction = NULL;
 J_ASSERT(t_nr_buffers <= t_outstanding_credits) //bomb!!!

Fix this issue by using journal->j_state_lock to protect the subtraction
in jbd2_journal_commit_transaction().

Fixes: 96f1e0974575 ("jbd2: avoid long hold times of j_state_lock while committing a transaction")
Signed-off-by: Zhang Yi 
Reviewed-by: Jan Kara 
Link: https://lore.kernel.org/r/20220611130426.2013258-1-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

jbd2: fix a potential race while discarding reserved buffers after an abort

2022-04-27T11:50:50+00:00

commit 23e3d7f7061f8682c751c46512718f47580ad8f0 upstream.

we got issue as follows:
[   72.796117] EXT4-fs error (device sda): ext4_journal_check_start:83: comm fallocate: Detected aborted journal
[   72.826847] EXT4-fs (sda): Remounting filesystem read-only
fallocate: fallocate failed: Read-only file system
[   74.791830] jbd2_journal_commit_transaction: jh=0xffff9cfefe725d90 bh=0x0000000000000000 end delay
[   74.793597] ------------[ cut here ]------------
[   74.794203] kernel BUG at fs/jbd2/transaction.c:2063!
[   74.794886] invalid opcode: 0000 [#1] PREEMPT SMP PTI
[   74.795533] CPU: 4 PID: 2260 Comm: jbd2/sda-8 Not tainted 5.17.0-rc8-next-20220315-dirty #150
[   74.798327] RIP: 0010:__jbd2_journal_unfile_buffer+0x3e/0x60
[   74.801971] RSP: 0018:ffffa828c24a3cb8 EFLAGS: 00010202
[   74.802694] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[   74.803601] RDX: 0000000000000001 RSI: ffff9cfefe725d90 RDI: ffff9cfefe725d90
[   74.804554] RBP: ffff9cfefe725d90 R08: 0000000000000000 R09: ffffa828c24a3b20
[   74.805471] R10: 0000000000000001 R11: 0000000000000001 R12: ffff9cfefe725d90
[   74.806385] R13: ffff9cfefe725d98 R14: 0000000000000000 R15: ffff9cfe833a4d00
[   74.807301] FS:  0000000000000000(0000) GS:ffff9d01afb00000(0000) knlGS:0000000000000000
[   74.808338] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   74.809084] CR2: 00007f2b81bf4000 CR3: 0000000100056000 CR4: 00000000000006e0
[   74.810047] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   74.810981] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   74.811897] Call Trace:
[   74.812241]  
[   74.812566]  __jbd2_journal_refile_buffer+0x12f/0x180
[   74.813246]  jbd2_journal_refile_buffer+0x4c/0xa0
[   74.813869]  jbd2_journal_commit_transaction.cold+0xa1/0x148
[   74.817550]  kjournald2+0xf8/0x3e0
[   74.819056]  kthread+0x153/0x1c0
[   74.819963]  ret_from_fork+0x22/0x30

Above issue may happen as follows:
        write                   truncate                   kjournald2
generic_perform_write
 ext4_write_begin
  ext4_walk_page_buffers
   do_journal_get_write_access ->add BJ_Reserved list
 ext4_journalled_write_end
  ext4_walk_page_buffers
   write_end_fn
    ext4_handle_dirty_metadata
                ***************JBD2 ABORT**************
     jbd2_journal_dirty_metadata
 -> return -EROFS, jh in reserved_list
                                                   jbd2_journal_commit_transaction
                                                    while (commit_transaction->t_reserved_list)
                                                      jh = commit_transaction->t_reserved_list;
                        truncate_pagecache_range
                         do_invalidatepage
			  ext4_journalled_invalidatepage
			   jbd2_journal_invalidatepage
			    journal_unmap_buffer
			     __dispose_buffer
			      __jbd2_journal_unfile_buffer
			       jbd2_journal_put_journal_head ->put last ref_count
			        __journal_remove_journal_head
				 bh->b_private = NULL;
				 jh->b_bh = NULL;
				                      jbd2_journal_refile_buffer(journal, jh);
							bh = jh2bh(jh);
							->bh is NULL, later will trigger null-ptr-deref
				 journal_free_journal_head(jh);

After commit 96f1e0974575, we no longer hold the j_state_lock while
iterating over the list of reserved handles in
jbd2_journal_commit_transaction().  This potentially allows the
journal_head to be freed by journal_unmap_buffer while the commit
codepath is also trying to free the BJ_Reserved buffers.  Keeping
j_state_lock held while trying extends hold time of the lock
minimally, and solves this issue.

Fixes: 96f1e0974575("jbd2: avoid long hold times of j_state_lock while committing a transaction")
Signed-off-by: Ye Bin 
Reviewed-by: Jan Kara 
Link: https://lore.kernel.org/r/20220317142137.1821590-1-yebin10@huawei.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Greg Kroah-Hartman

jbd2: improve comments about freeing data buffers whose page mapping is NULL

2020-04-21T07:04:52+00:00

commit 780f66e59231fcf882f36c63f287252ee47cc75a upstream.

Improve comments in jbd2_journal_commit_transaction() to describe why
we don't need to clear the buffer_mapped bit for freeing file mapping
buffers whose page mapping is NULL.

Link: https://lore.kernel.org/r/20200217112706.20085-1-yi.zhang@huawei.com
Fixes: c96dceeabf76 ("jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer")
Suggested-by: Jan Kara 
Reviewed-by: Jan Kara 
Signed-off-by: zhangyi (F) 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Greg Kroah-Hartman

jbd2: switch to use jbd2_journal_abort() when failed to submit the commit record

2020-02-24T07:36:57+00:00

[ Upstream commit d0a186e0d3e7ac05cc77da7c157dae5aa59f95d9 ]

We invoke jbd2_journal_abort() to abort the journal and record errno
in the jbd2 superblock when committing journal transaction besides the
failure on submitting the commit record. But there is no need for the
case and we can also invoke jbd2_journal_abort() instead of
__jbd2_journal_abort_hard().

Fixes: 818d276ceb83a ("ext4: Add the journal checksum feature")
Signed-off-by: zhangyi (F) 
Reviewed-by: Jan Kara 
Link: https://lore.kernel.org/r/20191204124614.45424-2-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

2020-02-19T18:53:09+00:00

[ Upstream commit c96dceeabf765d0b1b1f29c3bf50a5c01315b820 ]

Commit 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from
an older transaction") set the BH_Freed flag when forgetting a metadata
buffer which belongs to the committing transaction, it indicate the
committing process clear dirty bits when it is done with the buffer. But
it also clear the BH_Mapped flag at the same time, which may trigger
below NULL pointer oops when block_size < PAGE_SIZE.

rmdir 1             kjournald2                 mkdir 2
                    jbd2_journal_commit_transaction
		    commit transaction N
jbd2_journal_forget
set_buffer_freed(bh1)
                    jbd2_journal_commit_transaction
                     commit transaction N+1
                     ...
                     clear_buffer_mapped(bh1)
                                               ext4_getblk(bh2 ummapped)
                                               ...
                                               grow_dev_page
                                                init_page_buffers
                                                 bh1->b_private=NULL
                                                 bh2->b_private=NULL
                     jbd2_journal_put_journal_head(jh1)
                      __journal_remove_journal_head(hb1)
		       jh1 is NULL and trigger oops

*) Dir entry block bh1 and bh2 belongs to one page, and the bh2 has
   already been unmapped.

For the metadata buffer we forgetting, we should always keep the mapped
flag and clear the dirty flags is enough, so this patch pick out the
these buffers and keep their BH_Mapped flag.

Link: https://lore.kernel.org/r/20200213063821.30455-3-yi.zhang@huawei.com
Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
Reviewed-by: Jan Kara 
Signed-off-by: zhangyi (F) 
Signed-off-by: Theodore Ts'o 
Cc: stable@kernel.org
Signed-off-by: Sasha Levin

jbd2: move the clearing of b_modified flag to the journal_unmap_buffer()

2020-02-19T18:53:09+00:00

[ Upstream commit 6a66a7ded12baa6ebbb2e3e82f8cb91382814839 ]

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 
Signed-off-by: zhangyi (F) 
Signed-off-by: Theodore Ts'o 
Cc: stable@kernel.org
Signed-off-by: Sasha Levin

jbd2: Fix statistics for the number of logged blocks

2020-01-04T18:17:14+00:00

[ Upstream commit 015c6033068208d6227612c878877919f3fcf6b6 ]

jbd2 statistics counting number of blocks logged in a transaction was
wrong. It didn't count the commit block and more importantly it didn't
count revoke descriptor blocks. Make sure these get properly counted.

Reviewed-by: Theodore Ts'o 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/r/20191105164437.32602-13-jack@suse.cz
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

jbd2: introduce jbd2_inode dirty range scoping

2019-06-20T21:24:56+00:00

Currently both journal_submit_inode_data_buffers() and
journal_finish_inode_data_buffers() operate on the entire address space
of each of the inodes associated with a given journal entry.  The
consequence of this is that if we have an inode where we are constantly
appending dirty pages we can end up waiting for an indefinite amount of
time in journal_finish_inode_data_buffers() while we wait for all the
pages under writeback to be written out.

The easiest way to cause this type of workload is do just dd from
/dev/zero to a file until it fills the entire filesystem.  This can
cause journal_finish_inode_data_buffers() to wait for the duration of
the entire dd operation.

We can improve this situation by scoping each of the inode dirty ranges
associated with a given transaction.  We do this via the jbd2_inode
structure so that the scoping is contained within jbd2 and so that it
follows the lifetime and locking rules for that structure.

This allows us to limit the writeback & wait in
journal_submit_inode_data_buffers() and
journal_finish_inode_data_buffers() respectively to the dirty range for
a given struct jdb2_inode, keeping us from waiting forever if the inode
in question is still being appended to.

Signed-off-by: Ross Zwisler 
Signed-off-by: Theodore Ts'o 
Reviewed-by: Jan Kara 
Cc: stable@vger.kernel.org

jbd2: fix typo in comment of journal_submit_inode_data_buffers

2019-05-30T19:15:57+00:00

delayed/dealyed

Signed-off-by: Liu Song 
Signed-off-by: Theodore Ts'o 
Reviewed-by: Jan Kara