2022-01-22 00:36:00

by Joseph Qi

[permalink] [raw]
Subject: [PATCH 0/2] ocfs2: fix a deadlock case

This is trying to fix a deadlock case in ocfs2.
We firstly export jbd2 symbols jbd2_journal_[grab|put]_journal_head as
preparation and later use them in ocfs2 insread of
jbd_[lock|unlock]_bh_journal_head to fix the deadlock.

Joseph Qi (2):
jbd2: export jbd2_journal_[grab|put]_journal_head
ocfs2: fix a deadlock when commit trans

fs/jbd2/journal.c | 2 ++
fs/ocfs2/suballoc.c | 25 +++++++++++--------------
2 files changed, 13 insertions(+), 14 deletions(-)

--
2.19.1.6.gb485710b


2022-01-22 00:36:00

by Joseph Qi

[permalink] [raw]
Subject: [PATCH 1/2] jbd2: export jbd2_journal_[grab|put]_journal_head

This exports symbols jbd2_journal_[grab|put]_journal_head, which will be
used outside modules, e.g. ocfs2.

Signed-off-by: Joseph Qi <[email protected]>
---
fs/jbd2/journal.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 0b86a4365b66..e9f0c72f6664 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -2972,6 +2972,7 @@ struct journal_head *jbd2_journal_grab_journal_head(struct buffer_head *bh)
jbd_unlock_bh_journal_head(bh);
return jh;
}
+EXPORT_SYMBOL(jbd2_journal_grab_journal_head);

static void __journal_remove_journal_head(struct buffer_head *bh)
{
@@ -3024,6 +3025,7 @@ void jbd2_journal_put_journal_head(struct journal_head *jh)
jbd_unlock_bh_journal_head(bh);
}
}
+EXPORT_SYMBOL(jbd2_journal_put_journal_head);

/*
* Initialize jbd inode head
--
2.19.1.6.gb485710b

2022-01-22 00:36:49

by Joseph Qi

[permalink] [raw]
Subject: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

commit 6f1b228529ae introduces a regression which can deadlock as
follows:

Task1: Task2:
jbd2_journal_commit_transaction ocfs2_test_bg_bit_allocatable
spin_lock(&jh->b_state_lock) jbd_lock_bh_journal_head
__jbd2_journal_remove_checkpoint spin_lock(&jh->b_state_lock)
jbd2_journal_put_journal_head
jbd_lock_bh_journal_head

Task1 and Task2 lock bh->b_state and jh->b_state_lock in different
order, which finally result in a deadlock.

So use jbd2_journal_[grab|put]_journal_head instead in
ocfs2_test_bg_bit_allocatable() to fix it.

Reported-by: Gautham Ananthakrishna <[email protected]>
Fixes: 6f1b228529ae ("ocfs2: fix race between searching chunks and release journal_head from buffer_head")
Cc: <[email protected]>
Signed-off-by: Joseph Qi <[email protected]>
---
fs/ocfs2/suballoc.c | 25 +++++++++++--------------
1 file changed, 11 insertions(+), 14 deletions(-)

diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index 481017e1dac5..166c8918c825 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -1251,26 +1251,23 @@ static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
{
struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
struct journal_head *jh;
- int ret = 1;
+ int ret;

if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
return 0;

- if (!buffer_jbd(bg_bh))
+ jh = jbd2_journal_grab_journal_head(bg_bh);
+ if (!jh)
return 1;

- jbd_lock_bh_journal_head(bg_bh);
- if (buffer_jbd(bg_bh)) {
- jh = bh2jh(bg_bh);
- spin_lock(&jh->b_state_lock);
- bg = (struct ocfs2_group_desc *) jh->b_committed_data;
- if (bg)
- ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
- else
- ret = 1;
- spin_unlock(&jh->b_state_lock);
- }
- jbd_unlock_bh_journal_head(bg_bh);
+ spin_lock(&jh->b_state_lock);
+ bg = (struct ocfs2_group_desc *) jh->b_committed_data;
+ if (bg)
+ ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
+ else
+ ret = 1;
+ spin_unlock(&jh->b_state_lock);
+ jbd2_journal_put_journal_head(jh);

return ret;
}
--
2.19.1.6.gb485710b

2022-01-24 05:50:55

by Gautham Ananthakrishna

[permalink] [raw]
Subject: RE: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

Hi,
This deadlock was originally reported by [email protected] Could you please add Saeed as the reportedby.

Thanks,
Gautham.

-----Original Message-----
From: Joseph Qi <[email protected]>
Sent: Friday, January 21, 2022 12:42 PM
To: [email protected]; [email protected]; [email protected]
Cc: Gautham Ananthakrishna <[email protected]>; [email protected]; [email protected]
Subject: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

commit 6f1b228529ae introduces a regression which can deadlock as
follows:

Task1: Task2:
jbd2_journal_commit_transaction ocfs2_test_bg_bit_allocatable
spin_lock(&jh->b_state_lock) jbd_lock_bh_journal_head
__jbd2_journal_remove_checkpoint spin_lock(&jh->b_state_lock)
jbd2_journal_put_journal_head
jbd_lock_bh_journal_head

Task1 and Task2 lock bh->b_state and jh->b_state_lock in different order, which finally result in a deadlock.

So use jbd2_journal_[grab|put]_journal_head instead in
ocfs2_test_bg_bit_allocatable() to fix it.

Reported-by: Gautham Ananthakrishna <[email protected]>
Fixes: 6f1b228529ae ("ocfs2: fix race between searching chunks and release journal_head from buffer_head")
Cc: <[email protected]>
Signed-off-by: Joseph Qi <[email protected]>
---
fs/ocfs2/suballoc.c | 25 +++++++++++--------------
1 file changed, 11 insertions(+), 14 deletions(-)

diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 481017e1dac5..166c8918c825 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -1251,26 +1251,23 @@ static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh, {
struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
struct journal_head *jh;
- int ret = 1;
+ int ret;

if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
return 0;

- if (!buffer_jbd(bg_bh))
+ jh = jbd2_journal_grab_journal_head(bg_bh);
+ if (!jh)
return 1;

- jbd_lock_bh_journal_head(bg_bh);
- if (buffer_jbd(bg_bh)) {
- jh = bh2jh(bg_bh);
- spin_lock(&jh->b_state_lock);
- bg = (struct ocfs2_group_desc *) jh->b_committed_data;
- if (bg)
- ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
- else
- ret = 1;
- spin_unlock(&jh->b_state_lock);
- }
- jbd_unlock_bh_journal_head(bg_bh);
+ spin_lock(&jh->b_state_lock);
+ bg = (struct ocfs2_group_desc *) jh->b_committed_data;
+ if (bg)
+ ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
+ else
+ ret = 1;
+ spin_unlock(&jh->b_state_lock);
+ jbd2_journal_put_journal_head(jh);

return ret;
}
--
2.19.1.6.gb485710b

2022-01-24 16:02:01

by Joseph Qi

[permalink] [raw]
Subject: Re: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

Sure, will do it in v2.
So could this patch resolve your issue?

Thanks,
Joseph

On 1/23/22 1:31 PM, Gautham Ananthakrishna wrote:
> Hi,
> This deadlock was originally reported by [email protected] Could you please add Saeed as the reportedby.
>
> Thanks,
> Gautham.
>
> -----Original Message-----
> From: Joseph Qi <[email protected]>
> Sent: Friday, January 21, 2022 12:42 PM
> To: [email protected]; [email protected]; [email protected]
> Cc: Gautham Ananthakrishna <[email protected]>; [email protected]; [email protected]
> Subject: [PATCH 2/2] ocfs2: fix a deadlock when commit trans
>
> commit 6f1b228529ae introduces a regression which can deadlock as
> follows:
>
> Task1: Task2:
> jbd2_journal_commit_transaction ocfs2_test_bg_bit_allocatable
> spin_lock(&jh->b_state_lock) jbd_lock_bh_journal_head
> __jbd2_journal_remove_checkpoint spin_lock(&jh->b_state_lock)
> jbd2_journal_put_journal_head
> jbd_lock_bh_journal_head
>
> Task1 and Task2 lock bh->b_state and jh->b_state_lock in different order, which finally result in a deadlock.
>
> So use jbd2_journal_[grab|put]_journal_head instead in
> ocfs2_test_bg_bit_allocatable() to fix it.
>
> Reported-by: Gautham Ananthakrishna <[email protected]>
> Fixes: 6f1b228529ae ("ocfs2: fix race between searching chunks and release journal_head from buffer_head")
> Cc: <[email protected]>
> Signed-off-by: Joseph Qi <[email protected]>
> ---
> fs/ocfs2/suballoc.c | 25 +++++++++++--------------
> 1 file changed, 11 insertions(+), 14 deletions(-)
>
> diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 481017e1dac5..166c8918c825 100644
> --- a/fs/ocfs2/suballoc.c
> +++ b/fs/ocfs2/suballoc.c
> @@ -1251,26 +1251,23 @@ static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh, {
> struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
> struct journal_head *jh;
> - int ret = 1;
> + int ret;
>
> if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
> return 0;
>
> - if (!buffer_jbd(bg_bh))
> + jh = jbd2_journal_grab_journal_head(bg_bh);
> + if (!jh)
> return 1;
>
> - jbd_lock_bh_journal_head(bg_bh);
> - if (buffer_jbd(bg_bh)) {
> - jh = bh2jh(bg_bh);
> - spin_lock(&jh->b_state_lock);
> - bg = (struct ocfs2_group_desc *) jh->b_committed_data;
> - if (bg)
> - ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
> - else
> - ret = 1;
> - spin_unlock(&jh->b_state_lock);
> - }
> - jbd_unlock_bh_journal_head(bg_bh);
> + spin_lock(&jh->b_state_lock);
> + bg = (struct ocfs2_group_desc *) jh->b_committed_data;
> + if (bg)
> + ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
> + else
> + ret = 1;
> + spin_unlock(&jh->b_state_lock);
> + jbd2_journal_put_journal_head(jh);
>
> return ret;
> }
> --
> 2.19.1.6.gb485710b

2022-01-27 23:15:08

by Gautham Ananthakrishna

[permalink] [raw]
Subject: RE: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

Yes. The patch has resolved the issue.

Thanks,
Gautham.

-----Original Message-----
From: Joseph Qi <[email protected]>
Sent: Monday, January 24, 2022 8:08 AM
To: Gautham Ananthakrishna <[email protected]>; [email protected]; [email protected]; [email protected]
Cc: [email protected]; [email protected]; Saeed Mirzamohammadi <[email protected]>
Subject: Re: [PATCH 2/2] ocfs2: fix a deadlock when commit trans

Sure, will do it in v2.
So could this patch resolve your issue?

Thanks,
Joseph

On 1/23/22 1:31 PM, Gautham Ananthakrishna wrote:
> Hi,
> This deadlock was originally reported by [email protected] Could you please add Saeed as the reportedby.
>
> Thanks,
> Gautham.
>
> -----Original Message-----
> From: Joseph Qi <[email protected]>
> Sent: Friday, January 21, 2022 12:42 PM
> To: [email protected]; [email protected]; [email protected]
> Cc: Gautham Ananthakrishna <[email protected]>; [email protected]; [email protected]
> Subject: [PATCH 2/2] ocfs2: fix a deadlock when commit trans
>
> commit 6f1b228529ae introduces a regression which can deadlock as
> follows:
>
> Task1: Task2:
> jbd2_journal_commit_transaction ocfs2_test_bg_bit_allocatable
> spin_lock(&jh->b_state_lock) jbd_lock_bh_journal_head
> __jbd2_journal_remove_checkpoint spin_lock(&jh->b_state_lock)
> jbd2_journal_put_journal_head
> jbd_lock_bh_journal_head
>
> Task1 and Task2 lock bh->b_state and jh->b_state_lock in different order, which finally result in a deadlock.
>
> So use jbd2_journal_[grab|put]_journal_head instead in
> ocfs2_test_bg_bit_allocatable() to fix it.
>
> Reported-by: Gautham Ananthakrishna <[email protected]>
> Fixes: 6f1b228529ae ("ocfs2: fix race between searching chunks and release journal_head from buffer_head")
> Cc: <[email protected]>
> Signed-off-by: Joseph Qi <[email protected]>
> ---
> fs/ocfs2/suballoc.c | 25 +++++++++++--------------
> 1 file changed, 11 insertions(+), 14 deletions(-)
>
> diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 481017e1dac5..166c8918c825 100644
> --- a/fs/ocfs2/suballoc.c
> +++ b/fs/ocfs2/suballoc.c
> @@ -1251,26 +1251,23 @@ static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh, {
> struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
> struct journal_head *jh;
> - int ret = 1;
> + int ret;
>
> if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
> return 0;
>
> - if (!buffer_jbd(bg_bh))
> + jh = jbd2_journal_grab_journal_head(bg_bh);
> + if (!jh)
> return 1;
>
> - jbd_lock_bh_journal_head(bg_bh);
> - if (buffer_jbd(bg_bh)) {
> - jh = bh2jh(bg_bh);
> - spin_lock(&jh->b_state_lock);
> - bg = (struct ocfs2_group_desc *) jh->b_committed_data;
> - if (bg)
> - ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
> - else
> - ret = 1;
> - spin_unlock(&jh->b_state_lock);
> - }
> - jbd_unlock_bh_journal_head(bg_bh);
> + spin_lock(&jh->b_state_lock);
> + bg = (struct ocfs2_group_desc *) jh->b_committed_data;
> + if (bg)
> + ret = !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
> + else
> + ret = 1;
> + spin_unlock(&jh->b_state_lock);
> + jbd2_journal_put_journal_head(jh);
>
> return ret;
> }
> --
> 2.19.1.6.gb485710b