From b1b8f39c2475a1df597e6d1970e25bd64c89d774 Mon Sep 17 00:00:00 2001 From: Ye Bin Date: Thu, 17 Mar 2022 22:21:37 +0800 Subject: jbd2: fix a potential race while discarding reserved buffers after an abort 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 --- fs/jbd2/commit.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index d188fa913a07..34b1406c06fd 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -501,7 +501,6 @@ void jbd2_journal_commit_transaction(journal_t *journal) } spin_unlock(&commit_transaction->t_handle_lock); commit_transaction->t_state = T_SWITCH; - write_unlock(&journal->j_state_lock); J_ASSERT (atomic_read(&commit_transaction->t_outstanding_credits) <= journal->j_max_transaction_buffers); @@ -521,6 +520,8 @@ void jbd2_journal_commit_transaction(journal_t *journal) * has reserved. This is consistent with the existing behaviour * that multiple jbd2_journal_get_write_access() calls to the same * buffer are perfectly permissible. + * We use journal->j_state_lock here to serialize processing of + * t_reserved_list with eviction of buffers from journal_unmap_buffer(). */ while (commit_transaction->t_reserved_list) { jh = commit_transaction->t_reserved_list; @@ -540,6 +541,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) jbd2_journal_refile_buffer(journal, jh); } + write_unlock(&journal->j_state_lock); /* * Now try to drop any written-back buffers from the journal's * checkpoint lists. We do this *before* commit because it potentially -- cgit v1.2.3