Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: make plock operation killable
  dlm: remove shared message stub for recovery
  dlm: delayed reply message warning
  dlm: Remove superfluous call to recalc_sigpending()

fs/dlm/config.c

@@ -100,6 +100,7 @@ struct dlm_cluster {
 	unsigned int cl_log_debug;
 	unsigned int cl_protocol;
 	unsigned int cl_timewarn_cs;
+	unsigned int cl_waitwarn_us;
 };
 
 enum {
@@ -114,6 +115,7 @@ enum {
 	CLUSTER_ATTR_LOG_DEBUG,
 	CLUSTER_ATTR_PROTOCOL,
 	CLUSTER_ATTR_TIMEWARN_CS,
+	CLUSTER_ATTR_WAITWARN_US,
 };
 
 struct cluster_attribute {
@@ -166,6 +168,7 @@ CLUSTER_ATTR(scan_secs, 1);
 CLUSTER_ATTR(log_debug, 0);
 CLUSTER_ATTR(protocol, 0);
 CLUSTER_ATTR(timewarn_cs, 1);
+CLUSTER_ATTR(waitwarn_us, 0);
 
 static struct configfs_attribute *cluster_attrs[] = {
 	[CLUSTER_ATTR_TCP_PORT] = &cluster_attr_tcp_port.attr,
@@ -179,6 +182,7 @@ static struct configfs_attribute *cluster_attrs[] = {
 	[CLUSTER_ATTR_LOG_DEBUG] = &cluster_attr_log_debug.attr,
 	[CLUSTER_ATTR_PROTOCOL] = &cluster_attr_protocol.attr,
 	[CLUSTER_ATTR_TIMEWARN_CS] = &cluster_attr_timewarn_cs.attr,
+	[CLUSTER_ATTR_WAITWARN_US] = &cluster_attr_waitwarn_us.attr,
 	NULL,
 };
 
@@ -439,6 +443,7 @@ static struct config_group *make_cluster(struct config_group *g,
 	cl->cl_log_debug = dlm_config.ci_log_debug;
 	cl->cl_protocol = dlm_config.ci_protocol;
 	cl->cl_timewarn_cs = dlm_config.ci_timewarn_cs;
+	cl->cl_waitwarn_us = dlm_config.ci_waitwarn_us;
 
 	space_list = &sps->ss_group;
 	comm_list = &cms->cs_group;
@@ -986,6 +991,7 @@ int dlm_our_addr(struct sockaddr_storage *addr, int num)
 #define DEFAULT_LOG_DEBUG          0
 #define DEFAULT_PROTOCOL           0
 #define DEFAULT_TIMEWARN_CS      500 /* 5 sec = 500 centiseconds */
+#define DEFAULT_WAITWARN_US	   0
 
 struct dlm_config_info dlm_config = {
 	.ci_tcp_port = DEFAULT_TCP_PORT,
@@ -998,6 +1004,7 @@ struct dlm_config_info dlm_config = {
 	.ci_scan_secs = DEFAULT_SCAN_SECS,
 	.ci_log_debug = DEFAULT_LOG_DEBUG,
 	.ci_protocol = DEFAULT_PROTOCOL,
-	.ci_timewarn_cs = DEFAULT_TIMEWARN_CS
+	.ci_timewarn_cs = DEFAULT_TIMEWARN_CS,
+	.ci_waitwarn_us = DEFAULT_WAITWARN_US
 };
 

fs/dlm/config.h

@@ -28,6 +28,7 @@ struct dlm_config_info {
 	int ci_log_debug;
 	int ci_protocol;
 	int ci_timewarn_cs;
+	int ci_waitwarn_us;
 };
 
 extern struct dlm_config_info dlm_config;

fs/dlm/dlm_internal.h

@@ -209,6 +209,7 @@ struct dlm_args {
 #define DLM_IFL_WATCH_TIMEWARN	0x00400000
 #define DLM_IFL_TIMEOUT_CANCEL	0x00800000
 #define DLM_IFL_DEADLOCK_CANCEL	0x01000000
+#define DLM_IFL_STUB_MS		0x02000000 /* magic number for m_flags */
 #define DLM_IFL_USER		0x00000001
 #define DLM_IFL_ORPHAN		0x00000002
 
@@ -245,6 +246,7 @@ struct dlm_lkb {
 
 	int8_t			lkb_wait_type;	/* type of reply waiting for */
 	int8_t			lkb_wait_count;
+	int			lkb_wait_nodeid; /* for debugging */
 
 	struct list_head	lkb_idtbl_list;	/* lockspace lkbtbl */
 	struct list_head	lkb_statequeue;	/* rsb g/c/w list */
@@ -254,6 +256,7 @@ struct dlm_lkb {
 	struct list_head	lkb_ownqueue;	/* list of locks for a process */
 	struct list_head	lkb_time_list;
 	ktime_t			lkb_timestamp;
+	ktime_t			lkb_wait_time;
 	unsigned long		lkb_timeout_cs;
 
 	struct dlm_callback	lkb_callbacks[DLM_CALLBACKS_SIZE];

fs/dlm/lock.h

@@ -24,6 +24,7 @@ int dlm_put_lkb(struct dlm_lkb *lkb);
 void dlm_scan_rsbs(struct dlm_ls *ls);
 int dlm_lock_recovery_try(struct dlm_ls *ls);
 void dlm_unlock_recovery(struct dlm_ls *ls);
+void dlm_scan_waiters(struct dlm_ls *ls);
 void dlm_scan_timeout(struct dlm_ls *ls);
 void dlm_adjust_timeouts(struct dlm_ls *ls);
 

fs/dlm/lockspace.c

@@ -243,7 +243,6 @@ static struct dlm_ls *find_ls_to_scan(void)
 static int dlm_scand(void *data)
 {
 	struct dlm_ls *ls;
-	int timeout_jiffies = dlm_config.ci_scan_secs * HZ;
 
 	while (!kthread_should_stop()) {
 		ls = find_ls_to_scan();
@@ -252,13 +251,14 @@ static int dlm_scand(void *data)
 				ls->ls_scan_time = jiffies;
 				dlm_scan_rsbs(ls);
 				dlm_scan_timeout(ls);
+				dlm_scan_waiters(ls);
 				dlm_unlock_recovery(ls);
 			} else {
 				ls->ls_scan_time += HZ;
 			}
-		} else {
-			schedule_timeout_interruptible(timeout_jiffies);
+			continue;
 		}
+		schedule_timeout_interruptible(dlm_config.ci_scan_secs * HZ);
 	}
 	return 0;
 }

fs/dlm/plock.c

@@ -71,6 +71,36 @@ static void send_op(struct plock_op *op)
 	wake_up(&send_wq);
 }
 
+/* If a process was killed while waiting for the only plock on a file,
+   locks_remove_posix will not see any lock on the file so it won't
+   send an unlock-close to us to pass on to userspace to clean up the
+   abandoned waiter.  So, we have to insert the unlock-close when the
+   lock call is interrupted. */
+
+static void do_unlock_close(struct dlm_ls *ls, u64 number,
+			    struct file *file, struct file_lock *fl)
+{
+	struct plock_op *op;
+
+	op = kzalloc(sizeof(*op), GFP_NOFS);
+	if (!op)
+		return;
+
+	op->info.optype		= DLM_PLOCK_OP_UNLOCK;
+	op->info.pid		= fl->fl_pid;
+	op->info.fsid		= ls->ls_global_id;
+	op->info.number		= number;
+	op->info.start		= 0;
+	op->info.end		= OFFSET_MAX;
+	if (fl->fl_lmops && fl->fl_lmops->fl_grant)
+		op->info.owner	= (__u64) fl->fl_pid;
+	else
+		op->info.owner	= (__u64)(long) fl->fl_owner;
+
+	op->info.flags |= DLM_PLOCK_FL_CLOSE;
+	send_op(op);
+}
+
 int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 		   int cmd, struct file_lock *fl)
 {
@@ -114,9 +144,19 @@ int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 
 	send_op(op);
 
-	if (xop->callback == NULL)
-		wait_event(recv_wq, (op->done != 0));
-	else {
+	if (xop->callback == NULL) {
+		rv = wait_event_killable(recv_wq, (op->done != 0));
+		if (rv == -ERESTARTSYS) {
+			log_debug(ls, "dlm_posix_lock: wait killed %llx",
+				  (unsigned long long)number);
+			spin_lock(&ops_lock);
+			list_del(&op->list);
+			spin_unlock(&ops_lock);
+			kfree(xop);
+			do_unlock_close(ls, number, file, fl);
+			goto out;
+		}
+	} else {
 		rv = FILE_LOCK_DEFERRED;
 		goto out;
 	}
@@ -233,6 +273,13 @@ int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	else
 		op->info.owner	= (__u64)(long) fl->fl_owner;
 
+	if (fl->fl_flags & FL_CLOSE) {
+		op->info.flags |= DLM_PLOCK_FL_CLOSE;
+		send_op(op);
+		rv = 0;
+		goto out;
+	}
+
 	send_op(op);
 	wait_event(recv_wq, (op->done != 0));
 
@@ -334,7 +381,10 @@ static ssize_t dev_read(struct file *file, char __user *u, size_t count,
 	spin_lock(&ops_lock);
 	if (!list_empty(&send_list)) {
 		op = list_entry(send_list.next, struct plock_op, list);
-		list_move(&op->list, &recv_list);
+		if (op->info.flags & DLM_PLOCK_FL_CLOSE)
+			list_del(&op->list);
+		else
+			list_move(&op->list, &recv_list);
 		memcpy(&info, &op->info, sizeof(info));
 	}
 	spin_unlock(&ops_lock);
@@ -342,6 +392,13 @@ static ssize_t dev_read(struct file *file, char __user *u, size_t count,
 	if (!op)
 		return -EAGAIN;
 
+	/* there is no need to get a reply from userspace for unlocks
+	   that were generated by the vfs cleaning up for a close
+	   (the process did not make an unlock call). */
+
+	if (op->info.flags & DLM_PLOCK_FL_CLOSE)
+		kfree(op);
+
 	if (copy_to_user(u, &info, sizeof(info)))
 		return -EFAULT;
 	return sizeof(info);

fs/dlm/user.c

@@ -611,7 +611,6 @@ static ssize_t device_write(struct file *file, const char __user *buf,
 
  out_sig:
 	sigprocmask(SIG_SETMASK, &tmpsig, NULL);
-	recalc_sigpending();
  out_free:
 	kfree(kbuf);
 	return error;

include/linux/dlm_plock.h

@@ -14,7 +14,7 @@
 #define DLM_PLOCK_MISC_NAME		"dlm_plock"
 
 #define DLM_PLOCK_VERSION_MAJOR	1
-#define DLM_PLOCK_VERSION_MINOR	1
+#define DLM_PLOCK_VERSION_MINOR	2
 #define DLM_PLOCK_VERSION_PATCH	0
 
 enum {
@@ -23,12 +23,14 @@ enum {
 	DLM_PLOCK_OP_GET,
 };
 
+#define DLM_PLOCK_FL_CLOSE 1
+
 struct dlm_plock_info {
 	__u32 version[3];
 	__u8 optype;
 	__u8 ex;
 	__u8 wait;
-	__u8 pad;
+	__u8 flags;
 	__u32 pid;
 	__s32 nodeid;
 	__s32 rv;