Blob Blame History Raw
From 5da844a2c7df642de2618fc3efe9a92eec40899d Mon Sep 17 00:00:00 2001
From: Gang He <ghe@suse.com>
Date: Thu, 11 Jul 2019 20:53:09 -0700
Subject: [PATCH 3/3] ocfs2: add first lock wait time in locking_state
Git-commit: 5da844a2c7df642de2618fc3efe9a92eec40899d
Patch-mainline: v5.3-rc1
References: bsc#1134390

ocfs2 file system uses locking_state file under debugfs to dump each
ocfs2 file system's dlm lock resources, but the users ever encountered
some hang(deadlock) problems in ocfs2 file system.  I'd like to add
first lock wait time in locking_state file, which can help the upper
scripts detect these deadlock problems via comparing the first lock wait
time with the current time.

Link: http://lkml.kernel.org/r/20190611015414.27754-3-ghe@suse.com
Signed-off-by: Gang He <ghe@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 fs/ocfs2/dlmglue.c | 32 +++++++++++++++++++++++++++++---
 fs/ocfs2/ocfs2.h   |  1 +
 2 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 4089daba4c6f..cf90688ff2ed 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -426,6 +426,7 @@ static void ocfs2_remove_lockres_tracking(struct ocfs2_lock_res *res)
 static void ocfs2_init_lock_stats(struct ocfs2_lock_res *res)
 {
 	res->l_lock_refresh = 0;
+	res->l_lock_wait = 0;
 	memset(&res->l_lock_prmode, 0, sizeof(struct ocfs2_lock_stats));
 	memset(&res->l_lock_exmode, 0, sizeof(struct ocfs2_lock_stats));
 }
@@ -469,6 +470,21 @@ static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
 	lockres->l_lock_refresh++;
 }
 
+static inline void ocfs2_track_lock_wait(struct ocfs2_lock_res *lockres)
+{
+	struct ocfs2_mask_waiter *mw;
+
+	if (list_empty(&lockres->l_mask_waiters)) {
+		lockres->l_lock_wait = 0;
+		return;
+	}
+
+	mw = list_first_entry(&lockres->l_mask_waiters,
+				struct ocfs2_mask_waiter, mw_item);
+	lockres->l_lock_wait =
+			ktime_to_us(ktime_mono_to_real(mw->mw_lock_start));
+}
+
 static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
 {
 	mw->mw_lock_start = ktime_get();
@@ -484,6 +500,9 @@ static inline void ocfs2_update_lock_stats(struct ocfs2_lock_res *res,
 static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
 {
 }
+static inline void ocfs2_track_lock_wait(struct ocfs2_lock_res *lockres)
+{
+}
 static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
 {
 }
@@ -877,6 +896,7 @@ static void lockres_set_flags(struct ocfs2_lock_res *lockres,
 		list_del_init(&mw->mw_item);
 		mw->mw_status = 0;
 		complete(&mw->mw_complete);
+		ocfs2_track_lock_wait(lockres);
 	}
 }
 static void lockres_or_flags(struct ocfs2_lock_res *lockres, unsigned long or)
@@ -1388,6 +1408,7 @@ static void lockres_add_mask_waiter(struct ocfs2_lock_res *lockres,
 	list_add_tail(&mw->mw_item, &lockres->l_mask_waiters);
 	mw->mw_mask = mask;
 	mw->mw_goal = goal;
+	ocfs2_track_lock_wait(lockres);
 }
 
 /* returns 0 if the mw that was removed was already satisfied, -EBUSY
@@ -1404,6 +1425,7 @@ static int __lockres_remove_mask_waiter(struct ocfs2_lock_res *lockres,
 
 		list_del_init(&mw->mw_item);
 		init_completion(&mw->mw_complete);
+		ocfs2_track_lock_wait(lockres);
 	}
 
 	return ret;
@@ -3084,7 +3106,7 @@ static void *ocfs2_dlm_seq_next(struct seq_file *m, void *v, loff_t *pos)
  * New in version 3
  *	- Max time in lock stats is in usecs (instead of nsecs)
  * New in version 4
- *	- Add last pr/ex unlock times in usecs
+ *	- Add last pr/ex unlock times and first lock wait time in usecs
  */
 #define OCFS2_DLM_DEBUG_STR_VERSION 4
 static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
@@ -3102,7 +3124,7 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
 		return -EINVAL;
 
 #ifdef CONFIG_OCFS2_FS_STATS
-	if (dlm_debug->d_filter_secs) {
+	if (!lockres->l_lock_wait && dlm_debug->d_filter_secs) {
 		now = ktime_to_us(ktime_get_real());
 		if (lockres->l_lock_prmode.ls_last >
 		    lockres->l_lock_exmode.ls_last)
@@ -3163,6 +3185,7 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
 # define lock_refresh(_l)		((_l)->l_lock_refresh)
 # define lock_last_prmode(_l)		((_l)->l_lock_prmode.ls_last)
 # define lock_last_exmode(_l)		((_l)->l_lock_exmode.ls_last)
+# define lock_wait(_l)			((_l)->l_lock_wait)
 #else
 # define lock_num_prmode(_l)		(0)
 # define lock_num_exmode(_l)		(0)
@@ -3175,6 +3198,7 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
 # define lock_refresh(_l)		(0)
 # define lock_last_prmode(_l)		(0ULL)
 # define lock_last_exmode(_l)		(0ULL)
+# define lock_wait(_l)			(0ULL)
 #endif
 	/* The following seq_print was added in version 2 of this output */
 	seq_printf(m, "%u\t"
@@ -3187,6 +3211,7 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
 		   "%u\t"
 		   "%u\t"
 		   "%llu\t"
+		   "%llu\t"
 		   "%llu\t",
 		   lock_num_prmode(lockres),
 		   lock_num_exmode(lockres),
@@ -3198,7 +3223,8 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
 		   lock_max_exmode(lockres),
 		   lock_refresh(lockres),
 		   lock_last_prmode(lockres),
-		   lock_last_exmode(lockres));
+		   lock_last_exmode(lockres),
+		   lock_wait(lockres));
 
 	/* End the line */
 	seq_printf(m, "\n");
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index c7539601555b..fddbbd60f434 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -192,6 +192,7 @@ struct ocfs2_lock_res {
 #ifdef CONFIG_OCFS2_FS_STATS
 	struct ocfs2_lock_stats  l_lock_prmode;		/* PR mode stats */
 	u32                      l_lock_refresh;	/* Disk refreshes */
+	u64                      l_lock_wait;	/* First lock wait time */
 	struct ocfs2_lock_stats  l_lock_exmode;		/* EX mode stats */
 #endif
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
-- 
2.12.3