2015-10-05 13:59:40

by Jan Kara

Subject: [PATCH] jbd2: Fix checkpoint list cleanup

Unlike comments and expectation of callers journal_clean_one_cp_list()
returned 1 not only if it freed the transaction but also if it freed
some buffers in the transaction. That could make
__jbd2_journal_clean_checkpoint_list() skip processing
t_checkpoint_io_list and continue with processing the next transaction.
This is mostly a cosmetic issue since the only result is we can
sometimes free less memory than we could. But it's still worth fixing.
Fix journal_clean_one_cp_list() to return 1 only if the transaction was
really freed.

Fixes: 50849db32a9f529235a84bcc84a6b8e631b1d0ec
Signed-off-by: Jan Kara <[email protected]>
---
fs/jbd2/checkpoint.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 8c44654ce274..684996c8a3a4 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -427,7 +427,6 @@ static int journal_clean_one_cp_list(struct journal_head *jh, bool destroy)
struct journal_head *last_jh;
struct journal_head *next_jh = jh;
int ret;
- int freed = 0;

if (!jh)
return 0;
@@ -441,10 +440,9 @@ static int journal_clean_one_cp_list(struct journal_head *jh, bool destroy)
else
ret = __jbd2_journal_remove_checkpoint(jh) + 1;
if (!ret)
- return freed;
+ return 0;
if (ret == 2)
return 1;
- freed = 1;
/*
* This function only frees up some memory
* if possible so we dont have an obligation
@@ -452,10 +450,10 @@ static int journal_clean_one_cp_list(struct journal_head *jh, bool destroy)
* requested:
*/
if (need_resched())
- return freed;
+ return 0;
} while (jh != last_jh);

- return freed;
+ return 0;
}

/*
--
2.1.4

2015-10-18 02:42:56

by Theodore Ts'o

Subject: Re: [PATCH] jbd2: Fix checkpoint list cleanup

On Mon, Oct 05, 2015 at 03:59:32PM +0200, Jan Kara wrote:
> Unlike comments and expectation of callers journal_clean_one_cp_list()
> returned 1 not only if it freed the transaction but also if it freed
> some buffers in the transaction. That could make
> __jbd2_journal_clean_checkpoint_list() skip processing
> t_checkpoint_io_list and continue with processing the next transaction.
> This is mostly a cosmetic issue since the only result is we can
> sometimes free less memory than we could. But it's still worth fixing.
> Fix journal_clean_one_cp_list() to return 1 only if the transaction was
> really freed.
>
> Fixes: 50849db32a9f529235a84bcc84a6b8e631b1d0ec
> Signed-off-by: Jan Kara <[email protected]>

Applied, thanks.

- Ted