[PATCH] bcachefs: fix deadlock in journal_entry_open()

[PATCH] bcachefs: fix deadlock in journal_entry_open()

[Jeongjun Park]

In the previous commit b3d82c2f2761, code was added to prevent journal sequence
overflow. Among them, the code added to journal_entry_open() uses the
bch2_fs_fatal_err_on() function to handle errors.

However, __journal_res_get() , which calls journal_entry_open() , calls
journal_entry_open() while holding journal->lock , but bch2_fs_fatal_err_on()
internally tries to acquire journal->lock , which results in a deadlock.

Therefore, we need to use bch_err() instead of bch2_fs_fatal_err_on() to
prevent deadlock.

Fixes: b3d82c2f2761 ("bcachefs: Guard against journal seq overflow")
Signed-off-by: Jeongjun Park <aha310510@gmail.com>
---
 fs/bcachefs/journal.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
index 2cd20114b74b..38ba5ab785c3 100644
--- a/fs/bcachefs/journal.c
+++ b/fs/bcachefs/journal.c
@@ -382,9 +382,10 @@ static int journal_entry_open(struct journal *j)
 	if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf))
 		return JOURNAL_ERR_max_in_flight;
 
-	if (bch2_fs_fatal_err_on(journal_cur_seq(j) >= JOURNAL_SEQ_MAX,
-				 c, "cannot start: journal seq overflow"))
+	if (journal_cur_seq(j) >= JOURNAL_SEQ_MAX) {
+		bch_err(c, "cannot start: journal seq overflow");
 		return JOURNAL_ERR_insufficient_devices; /* -EROFS */
+	}
 
 	BUG_ON(!j->cur_entry_sectors);
 
--

Re: [PATCH] bcachefs: fix deadlock in journal_entry_open()

[Kent Overstreet]

On Thu, Jan 30, 2025 at 02:17:44AM +0900, Jeongjun Park wrote:
> In the previous commit b3d82c2f2761, code was added to prevent journal sequence
> overflow. Among them, the code added to journal_entry_open() uses the
> bch2_fs_fatal_err_on() function to handle errors.
> 
> However, __journal_res_get() , which calls journal_entry_open() , calls
> journal_entry_open() while holding journal->lock , but bch2_fs_fatal_err_on()
> internally tries to acquire journal->lock , which results in a deadlock.
> 
> Therefore, we need to use bch_err() instead of bch2_fs_fatal_err_on() to
> prevent deadlock.

We do the emergency shutdown though...

Perhaps add a bch2_journal_halt_locked() helper?

> Fixes: b3d82c2f2761 ("bcachefs: Guard against journal seq overflow")
> Signed-off-by: Jeongjun Park <aha310510@gmail.com>
> ---
>  fs/bcachefs/journal.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
> index 2cd20114b74b..38ba5ab785c3 100644
> --- a/fs/bcachefs/journal.c
> +++ b/fs/bcachefs/journal.c
> @@ -382,9 +382,10 @@ static int journal_entry_open(struct journal *j)
>  	if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf))
>  		return JOURNAL_ERR_max_in_flight;
>  
> -	if (bch2_fs_fatal_err_on(journal_cur_seq(j) >= JOURNAL_SEQ_MAX,
> -				 c, "cannot start: journal seq overflow"))
> +	if (journal_cur_seq(j) >= JOURNAL_SEQ_MAX) {
> +		bch_err(c, "cannot start: journal seq overflow");
>  		return JOURNAL_ERR_insufficient_devices; /* -EROFS */
> +	}
>  
>  	BUG_ON(!j->cur_entry_sectors);
>  
> --