2022-11-25 03:42:54

by Rock Li(李宏伟)

Subject: [PATCH] ocfs2: properly clean task pointer before o2hb thread exits abnormally

From: Rock Li <[email protected]>

in case o2hb thread exits abnormally before stop o2cb service,
heartbeat_drop_group_item will try to stop a noexit thread cause
ret->hb_task was not clear properly.

Signed-off-by: Rock Li <[email protected]>
---
fs/ocfs2/cluster/heartbeat.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index b13d344..9db448d 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -1258,6 +1258,8 @@ static int o2hb_thread(void *data)
/* Unpin node */
o2nm_undepend_this_node();

+ reg->hr_task = NULL;
+
mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");

return 0;
--
1.8.3.1

2022-11-25 09:43:55

by Joseph Qi

Subject: Re: [PATCH] ocfs2: properly clean task pointer before o2hb thread exits abnormally

On 11/25/22 11:15 AM, lihongweizz wrote:
> From: Rock Li <[email protected]>
>
> in case o2hb thread exits abnormally before stop o2cb service,
> heartbeat_drop_group_item will try to stop a noexit thread cause
> ret->hb_task was not clear properly.
>
Could you please describe the scenario in detail?

Thanks,
Joseph

> Signed-off-by: Rock Li <[email protected]>
> ---
> fs/ocfs2/cluster/heartbeat.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
> index b13d344..9db448d 100644
> --- a/fs/ocfs2/cluster/heartbeat.c
> +++ b/fs/ocfs2/cluster/heartbeat.c
> @@ -1258,6 +1258,8 @@ static int o2hb_thread(void *data)
> /* Unpin node */
> o2nm_undepend_this_node();
>
> + reg->hr_task = NULL;
> +
> mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");
>
> return 0;