dlm for 6.11

- New flag DLM_LSFL_SOFTIRQ_SAFE can be set by code using dlm
   to indicate callbacks can be run from softirq.
 - Change md-cluster to set DLM_LSFL_SOFTIRQ_SAFE.
 - Clean up for previous changes, e.g. unused code and parameters.
 - Remove custom pre-allocation of rsb structs which is unnecessary
   with kmem caches.
 - Change idr to xarray for lkb structs in use.
 - Change idr to xarray for rsb structs being recovered.
 - Change outdated naming related to internal rsb states.
 - Fix some incorrect add/remove of rsb on scan list.
 - Use rcu to free rsb structs.
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCAAdFiEEcGkeEvkvjdvlR90nOBtzx/yAaaoFAmaVUlgACgkQOBtzx/yA
 aapCfQ//eqs19no6+TUagkzboIGxGbrPEqmJNj4Vu1sCSH3tVC4IrkI2IqqPJL9N
 tYHUQvp3BYOdenBZzw6tmbs6cvoA7Fps7YMqqkEKYfBCHcV9KtejqvwBdJfqiN6A
 RniImAph0qvvI6GK4Y+6nDyxU2n8enOhgnZMRDUS/KYV8frc70SxreqzPSkPMWLh
 ZnDgTIF4zahUBFEkILlXYArbbRk5FKL+SMkSDZyDd78bVnjX24KgtOt7HpDX9X70
 /9DrDz3uI+XShXzpIint4Ee4ghZr1lM9g9LXDazuY62SBDknhGTzY0BYVxZ2U3NG
 ocUh2KbJoP29sncNxLf9Nev5JPc+Wx3iCTEgLKkOEc4Yf0jAZg+1xbopWDT+qjRV
 djsgTCQ1gjpHgQxrlUUo7N5ilo5ocgSXSHGJ8b886tG5eZaxiN1y3TB4T4JtO+FH
 Q4IkFJiaYDL44xYR85wpfOcct/5mR7kPvhuxouexKobO+lKXaUONP9Wj7pRgG/M5
 qhrWY4EU8jcO/nPunPxvhJdL68T3WoHDN42tWt/7kYQqY2svvfmr6NEImde6GxqX
 PB3hW20cvD4wULumLM+h0rQacIWuuMQ5ahIX9og6jM7Yx/ucks1pgnRo0M0R1aUc
 OopoTAekSdRtgbRXr5IQPRxpKB6BFUp3Va/Yo+2g0fi5QywcVZc=
 =dDCi
 -----END PGP SIGNATURE-----

Merge tag 'dlm-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm

Pull dlm updates from David Teigland:

 - New flag DLM_LSFL_SOFTIRQ_SAFE can be set by code using dlm to
   indicate callbacks can be run from softirq

 - Change md-cluster to set DLM_LSFL_SOFTIRQ_SAFE

 - Clean up for previous changes, e.g. unused code and parameters

 - Remove custom pre-allocation of rsb structs which is unnecessary with
   kmem caches

 - Change idr to xarray for lkb structs in use

 - Change idr to xarray for rsb structs being recovered

 - Change outdated naming related to internal rsb states

 - Fix some incorrect add/remove of rsb on scan list

 - Use rcu to free rsb structs

* tag 'dlm-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm:
  dlm: add rcu_barrier before destroy kmem cache
  dlm: remove DLM_LSFL_SOFTIRQ from exflags
  fs: dlm: remove unused struct 'dlm_processed_nodes'
  md-cluster: use DLM_LSFL_SOFTIRQ for dlm_new_lockspace()
  dlm: implement LSFL_SOFTIRQ_SAFE
  dlm: introduce DLM_LSFL_SOFTIRQ_SAFE
  dlm: use LSFL_FS to check for kernel lockspace
  dlm: use rcu to avoid an extra rsb struct lookup
  dlm: fix add_scan and del_scan usage
  dlm: change list and timer names
  dlm: move recover idr to xarray datastructure
  dlm: move lkb idr to xarray datastructure
  dlm: drop own rsb pre allocation mechanism
  dlm: remove ls_local_handle from struct dlm_ls
  dlm: remove unused parameter in dlm_midcomms_addr
  dlm: don't kref_init rsbs created for toss list
  dlm: remove scand leftovers

drivers/md/md-cluster.c

@@ -887,7 +887,7 @@ static int join(struct mddev *mddev, int nodes)
 	memset(str, 0, 64);
 	sprintf(str, "%pU", mddev->uuid);
 	ret = dlm_new_lockspace(str, mddev->bitmap_info.cluster_name,
-				0, LVB_SIZE, &md_ls_ops, mddev,
+				DLM_LSFL_SOFTIRQ, LVB_SIZE, &md_ls_ops, mddev,
 				&ops_rv, &cinfo->lockspace);
 	if (ret)
 		goto err;

fs/dlm/ast.c

@@ -18,35 +18,52 @@
 #include "user.h"
 #include "ast.h"
 
+static void dlm_run_callback(uint32_t ls_id, uint32_t lkb_id, int8_t mode,
+			     uint32_t flags, uint8_t sb_flags, int sb_status,
+			     struct dlm_lksb *lksb,
+			     void (*astfn)(void *astparam),
+			     void (*bastfn)(void *astparam, int mode),
+			     void *astparam, const char *res_name,
+			     size_t res_length)
+{
+	if (flags & DLM_CB_BAST) {
+		trace_dlm_bast(ls_id, lkb_id, mode, res_name, res_length);
+		bastfn(astparam, mode);
+	} else if (flags & DLM_CB_CAST) {
+		trace_dlm_ast(ls_id, lkb_id, sb_status, sb_flags, res_name,
+			      res_length);
+		lksb->sb_status = sb_status;
+		lksb->sb_flags = sb_flags;
+		astfn(astparam);
+	}
+}
+
+static void dlm_do_callback(struct dlm_callback *cb)
+{
+	dlm_run_callback(cb->ls_id, cb->lkb_id, cb->mode, cb->flags,
+			 cb->sb_flags, cb->sb_status, cb->lkb_lksb,
+			 cb->astfn, cb->bastfn, cb->astparam,
+			 cb->res_name, cb->res_length);
+	dlm_free_cb(cb);
+}
+
 static void dlm_callback_work(struct work_struct *work)
 {
 	struct dlm_callback *cb = container_of(work, struct dlm_callback, work);
 
-	if (cb->flags & DLM_CB_BAST) {
+	dlm_do_callback(cb);
-		trace_dlm_bast(cb->ls_id, cb->lkb_id, cb->mode, cb->res_name,
-			       cb->res_length);
-		cb->bastfn(cb->astparam, cb->mode);
-	} else if (cb->flags & DLM_CB_CAST) {
-		trace_dlm_ast(cb->ls_id, cb->lkb_id, cb->sb_status,
-			      cb->sb_flags, cb->res_name, cb->res_length);
-		cb->lkb_lksb->sb_status = cb->sb_status;
-		cb->lkb_lksb->sb_flags = cb->sb_flags;
-		cb->astfn(cb->astparam);
-	}
-
-	dlm_free_cb(cb);
 }
 
-int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+bool dlm_may_skip_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
-			   int status, uint32_t sbflags,
+			   int status, uint32_t sbflags, int *copy_lvb)
-			   struct dlm_callback **cb)
 {
 	struct dlm_rsb *rsb = lkb->lkb_resource;
-	int rv = DLM_ENQUEUE_CALLBACK_SUCCESS;
 	struct dlm_ls *ls = rsb->res_ls;
-	int copy_lvb = 0;
 	int prev_mode;
 
+	if (copy_lvb)
+		*copy_lvb = 0;
+
 	if (flags & DLM_CB_BAST) {
 		/* if cb is a bast, it should be skipped if the blocking mode is
 		 * compatible with the last granted mode
@@ -56,7 +73,7 @@ int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 				log_debug(ls, "skip %x bast mode %d for cast mode %d",
 					  lkb->lkb_id, mode,
 					  lkb->lkb_last_cast_cb_mode);
-				goto out;
+				return true;
 			}
 		}
 
@@ -74,7 +91,7 @@ int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 			    (prev_mode > mode && prev_mode > DLM_LOCK_PR)) {
 				log_debug(ls, "skip %x add bast mode %d for bast mode %d",
 					  lkb->lkb_id, mode, prev_mode);
-				goto out;
+				return true;
 			}
 		}
 
@@ -85,8 +102,10 @@ int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 			prev_mode = lkb->lkb_last_cast_cb_mode;
 
 			if (!status && lkb->lkb_lksb->sb_lvbptr &&
-			    dlm_lvb_operations[prev_mode + 1][mode + 1])
+			    dlm_lvb_operations[prev_mode + 1][mode + 1]) {
-				copy_lvb = 1;
+				if (copy_lvb)
+					*copy_lvb = 1;
+			}
 		}
 
 		lkb->lkb_last_cast_cb_mode = mode;
@@ -96,11 +115,19 @@ int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 	lkb->lkb_last_cb_mode = mode;
 	lkb->lkb_last_cb_flags = flags;
 
+	return false;
+}
+
+int dlm_get_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
+	       int status, uint32_t sbflags,
+	       struct dlm_callback **cb)
+{
+	struct dlm_rsb *rsb = lkb->lkb_resource;
+	struct dlm_ls *ls = rsb->res_ls;
+
 	*cb = dlm_allocate_cb();
-	if (!*cb) {
+	if (WARN_ON_ONCE(!*cb))
-		rv = DLM_ENQUEUE_CALLBACK_FAILURE;
+		return -ENOMEM;
-		goto out;
-	}
 
 	/* for tracing */
 	(*cb)->lkb_id = lkb->lkb_id;
@@ -112,19 +139,34 @@ int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 	(*cb)->mode = mode;
 	(*cb)->sb_status = status;
 	(*cb)->sb_flags = (sbflags & 0x000000FF);
-	(*cb)->copy_lvb = copy_lvb;
 	(*cb)->lkb_lksb = lkb->lkb_lksb;
 
-	rv = DLM_ENQUEUE_CALLBACK_NEED_SCHED;
+	return 0;
+}
 
-out:
+static int dlm_get_queue_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
-	return rv;
+			    int status, uint32_t sbflags,
+			    struct dlm_callback **cb)
+{
+	int rv;
+
+	rv = dlm_get_cb(lkb, flags, mode, status, sbflags, cb);
+	if (rv)
+		return rv;
+
+	(*cb)->astfn = lkb->lkb_astfn;
+	(*cb)->bastfn = lkb->lkb_bastfn;
+	(*cb)->astparam = lkb->lkb_astparam;
+	INIT_WORK(&(*cb)->work, dlm_callback_work);
+
+	return 0;
 }
 
 void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
-		  uint32_t sbflags)
+		uint32_t sbflags)
 {
-	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
+	struct dlm_rsb *rsb = lkb->lkb_resource;
+	struct dlm_ls *ls = rsb->res_ls;
 	struct dlm_callback *cb;
 	int rv;
 
@@ -133,34 +175,36 @@ void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
 		return;
 	}
 
-	rv = dlm_queue_lkb_callback(lkb, flags, mode, status, sbflags,
+	if (dlm_may_skip_callback(lkb, flags, mode, status, sbflags, NULL))
-				    &cb);
+		return;
-	switch (rv) {
-	case DLM_ENQUEUE_CALLBACK_NEED_SCHED:
-		cb->astfn = lkb->lkb_astfn;
-		cb->bastfn = lkb->lkb_bastfn;
-		cb->astparam = lkb->lkb_astparam;
-		INIT_WORK(&cb->work, dlm_callback_work);
 
-		spin_lock_bh(&ls->ls_cb_lock);
+	spin_lock_bh(&ls->ls_cb_lock);
-		if (test_bit(LSFL_CB_DELAY, &ls->ls_flags))
+	if (test_bit(LSFL_CB_DELAY, &ls->ls_flags)) {
+		rv = dlm_get_queue_cb(lkb, flags, mode, status, sbflags, &cb);
+		if (!rv)
 			list_add(&cb->list, &ls->ls_cb_delay);
-		else
+	} else {
-			queue_work(ls->ls_callback_wq, &cb->work);
+		if (test_bit(LSFL_SOFTIRQ, &ls->ls_flags)) {
-		spin_unlock_bh(&ls->ls_cb_lock);
+			dlm_run_callback(ls->ls_global_id, lkb->lkb_id, mode, flags,
-		break;
+					 sbflags, status, lkb->lkb_lksb,
-	case DLM_ENQUEUE_CALLBACK_SUCCESS:
+					 lkb->lkb_astfn, lkb->lkb_bastfn,
-		break;
+					 lkb->lkb_astparam, rsb->res_name,
-	case DLM_ENQUEUE_CALLBACK_FAILURE:
+					 rsb->res_length);
-		fallthrough;
+		} else {
-	default:
+			rv = dlm_get_queue_cb(lkb, flags, mode, status, sbflags, &cb);
-		WARN_ON_ONCE(1);
+			if (!rv)
-		break;
+				queue_work(ls->ls_callback_wq, &cb->work);
+		}
 	}
+	spin_unlock_bh(&ls->ls_cb_lock);
 }
 
 int dlm_callback_start(struct dlm_ls *ls)
 {
+	if (!test_bit(LSFL_FS, &ls->ls_flags) ||
+	    test_bit(LSFL_SOFTIRQ, &ls->ls_flags))
+		return 0;
+
 	ls->ls_callback_wq = alloc_ordered_workqueue("dlm_callback",
 						     WQ_HIGHPRI | WQ_MEM_RECLAIM);
 	if (!ls->ls_callback_wq) {
@@ -178,13 +222,15 @@ void dlm_callback_stop(struct dlm_ls *ls)
 
 void dlm_callback_suspend(struct dlm_ls *ls)
 {
-	if (ls->ls_callback_wq) {
+	if (!test_bit(LSFL_FS, &ls->ls_flags))
-		spin_lock_bh(&ls->ls_cb_lock);
+		return;
-		set_bit(LSFL_CB_DELAY, &ls->ls_flags);
-		spin_unlock_bh(&ls->ls_cb_lock);
 
+	spin_lock_bh(&ls->ls_cb_lock);
+	set_bit(LSFL_CB_DELAY, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_cb_lock);
+
+	if (ls->ls_callback_wq)
 		flush_workqueue(ls->ls_callback_wq);
-	}
 }
 
 #define MAX_CB_QUEUE 25
@@ -195,14 +241,18 @@ void dlm_callback_resume(struct dlm_ls *ls)
 	int count = 0, sum = 0;
 	bool empty;
 
-	if (!ls->ls_callback_wq)
+	if (!test_bit(LSFL_FS, &ls->ls_flags))
 		return;
 
 more:
 	spin_lock_bh(&ls->ls_cb_lock);
 	list_for_each_entry_safe(cb, safe, &ls->ls_cb_delay, list) {
 		list_del(&cb->list);
-		queue_work(ls->ls_callback_wq, &cb->work);
+		if (test_bit(LSFL_SOFTIRQ, &ls->ls_flags))
+			dlm_do_callback(cb);
+		else
+			queue_work(ls->ls_callback_wq, &cb->work);
+
 		count++;
 		if (count == MAX_CB_QUEUE)
 			break;

fs/dlm/ast.h

@@ -11,12 +11,11 @@
 #ifndef __ASTD_DOT_H__
 #define __ASTD_DOT_H__
 
-#define DLM_ENQUEUE_CALLBACK_NEED_SCHED	1
+bool dlm_may_skip_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
-#define DLM_ENQUEUE_CALLBACK_SUCCESS	0
+			   int status, uint32_t sbflags, int *copy_lvb);
-#define DLM_ENQUEUE_CALLBACK_FAILURE	-1
+int dlm_get_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
-int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+	       int status, uint32_t sbflags,
-			   int status, uint32_t sbflags,
+	       struct dlm_callback **cb);
-			   struct dlm_callback **cb);
 void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
                 uint32_t sbflags);
 

fs/dlm/config.c

@@ -672,7 +672,7 @@ static ssize_t comm_addr_store(struct config_item *item, const char *buf,
 
 	memcpy(addr, buf, len);
 
-	rv = dlm_midcomms_addr(cm->nodeid, addr, len);
+	rv = dlm_midcomms_addr(cm->nodeid, addr);
 	if (rv) {
 		kfree(addr);
 		return rv;

fs/dlm/debug_fs.c

@@ -380,7 +380,7 @@ static const struct seq_operations format4_seq_ops;
 
 static int table_seq_show(struct seq_file *seq, void *iter_ptr)
 {
-	struct dlm_rsb *rsb = list_entry(iter_ptr, struct dlm_rsb, res_rsbs_list);
+	struct dlm_rsb *rsb = list_entry(iter_ptr, struct dlm_rsb, res_slow_list);
 
 	if (seq->op == &format1_seq_ops)
 		print_format1(rsb, seq);
@@ -409,9 +409,9 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 	}
 
 	if (seq->op == &format4_seq_ops)
-		list = &ls->ls_toss;
+		list = &ls->ls_slow_inactive;
 	else
-		list = &ls->ls_keep;
+		list = &ls->ls_slow_active;
 
 	read_lock_bh(&ls->ls_rsbtbl_lock);
 	return seq_list_start(list, *pos);
@@ -423,9 +423,9 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 	struct list_head *list;
 
 	if (seq->op == &format4_seq_ops)
-		list = &ls->ls_toss;
+		list = &ls->ls_slow_inactive;
 	else
-		list = &ls->ls_keep;
+		list = &ls->ls_slow_active;
 
 	return seq_list_next(iter_ptr, list, pos);
 }

fs/dlm/dlm_internal.h

@@ -36,7 +36,7 @@
 #include <linux/miscdevice.h>
 #include <linux/rhashtable.h>
 #include <linux/mutex.h>
-#include <linux/idr.h>
+#include <linux/xarray.h>
 #include <linux/ratelimit.h>
 #include <linux/uaccess.h>
 
@@ -316,26 +316,24 @@ struct dlm_rsb {
 	int			res_nodeid;
 	int			res_master_nodeid;
 	int			res_dir_nodeid;
-	int			res_id;		/* for ls_recover_idr */
+	unsigned long		res_id;		/* for ls_recover_xa */
 	uint32_t                res_lvbseq;
 	uint32_t		res_hash;
 	unsigned long		res_toss_time;
 	uint32_t		res_first_lkid;
 	struct list_head	res_lookup;	/* lkbs waiting on first */
-	union {
+	struct rhash_head	res_node;	/* rsbtbl */
-		struct list_head	res_hashchain;
-		struct rhash_head	res_node; /* rsbtbl */
-	};
 	struct list_head	res_grantqueue;
 	struct list_head	res_convertqueue;
 	struct list_head	res_waitqueue;
 
-	struct list_head	res_rsbs_list;
+	struct list_head	res_slow_list;      /* ls_slow_* */
+	struct list_head	res_scan_list;
 	struct list_head	res_root_list;	    /* used for recovery */
 	struct list_head	res_masters_list;   /* used for recovery */
 	struct list_head	res_recover_list;   /* used for recovery */
-	struct list_head	res_toss_q_list;
 	int			res_recover_locks_count;
+	struct rcu_head		rcu;
 
 	char			*res_lvbptr;
 	char			res_name[DLM_RESNAME_MAXLEN+1];
@@ -368,7 +366,8 @@ enum rsb_flags {
 	RSB_RECOVER_CONVERT,
 	RSB_RECOVER_GRANT,
 	RSB_RECOVER_LVB_INVAL,
-	RSB_TOSS,
+	RSB_INACTIVE,
+	RSB_HASHED, /* set while rsb is on ls_rsbtbl */
 };
 
 static inline void rsb_set_flag(struct dlm_rsb *r, enum rsb_flags flag)
@@ -559,16 +558,8 @@ struct rcom_lock {
 	char			rl_lvb[];
 };
 
-/*
- * The max number of resources per rsbtbl bucket that shrink will attempt
- * to remove in each iteration.
- */
-
-#define DLM_REMOVE_NAMES_MAX 8
-
 struct dlm_ls {
 	struct list_head	ls_list;	/* list of lockspaces */
-	dlm_lockspace_t		*ls_local_handle;
 	uint32_t		ls_global_id;	/* global unique lockspace ID */
 	uint32_t		ls_generation;
 	uint32_t		ls_exflags;
@@ -578,26 +569,21 @@ struct dlm_ls {
 	wait_queue_head_t	ls_count_wait;
 	int			ls_create_count; /* create/release refcount */
 	unsigned long		ls_flags;	/* LSFL_ */
-	unsigned long		ls_scan_time;
 	struct kobject		ls_kobj;
 
-	struct idr		ls_lkbidr;
+	struct xarray		ls_lkbxa;
-	rwlock_t		ls_lkbidr_lock;
+	rwlock_t		ls_lkbxa_lock;
 
+	/* an rsb is on rsbtl for primary locking functions,
+	   and on a slow list for recovery/dump iteration  */
 	struct rhashtable	ls_rsbtbl;
-	rwlock_t		ls_rsbtbl_lock;
+	rwlock_t		ls_rsbtbl_lock; /* for ls_rsbtbl and ls_slow */
+	struct list_head	ls_slow_inactive; /* to iterate rsbtbl */
+	struct list_head	ls_slow_active;   /* to iterate rsbtbl */
 
-	struct list_head	ls_toss;
+	struct timer_list	ls_scan_timer; /* based on first scan_list rsb toss_time */
-	struct list_head	ls_keep;
+	struct list_head	ls_scan_list;  /* rsbs ordered by res_toss_time */
-
+	spinlock_t		ls_scan_lock;
-	struct timer_list	ls_timer;
-	/* this queue is ordered according the
-	 * absolute res_toss_time jiffies time
-	 * to mod_timer() with the first element
-	 * if necessary.
-	 */
-	struct list_head	ls_toss_q;
-	spinlock_t		ls_toss_q_lock;
 
 	spinlock_t		ls_waiters_lock;
 	struct list_head	ls_waiters;	/* lkbs needing a reply */
@@ -605,10 +591,6 @@ struct dlm_ls {
 	spinlock_t		ls_orphans_lock;
 	struct list_head	ls_orphans;
 
-	spinlock_t		ls_new_rsb_spin;
-	int			ls_new_rsb_count;
-	struct list_head	ls_new_rsb;	/* new rsb structs */
-
 	struct list_head	ls_nodes;	/* current nodes in ls */
 	struct list_head	ls_nodes_gone;	/* dead node list, recovery */
 	int			ls_num_nodes;	/* number of nodes in ls */
@@ -664,8 +646,8 @@ struct dlm_ls {
 	struct list_head	ls_recover_list;
 	spinlock_t		ls_recover_list_lock;
 	int			ls_recover_list_count;
-	struct idr		ls_recover_idr;
+	struct xarray		ls_recover_xa;
-	spinlock_t		ls_recover_idr_lock;
+	spinlock_t		ls_recover_xa_lock;
 	wait_queue_head_t	ls_wait_general;
 	wait_queue_head_t	ls_recover_lock_wait;
 	spinlock_t		ls_clear_proc_locks;
@@ -716,6 +698,8 @@ struct dlm_ls {
 #define LSFL_CB_DELAY		9
 #define LSFL_NODIR		10
 #define LSFL_RECV_MSG_BLOCKED	11
+#define LSFL_FS			12
+#define LSFL_SOFTIRQ		13
 
 #define DLM_PROC_FLAGS_CLOSING 1
 #define DLM_PROC_FLAGS_COMPAT  2

fs/dlm/lock.h

@@ -11,7 +11,6 @@
 #ifndef __LOCK_DOT_H__
 #define __LOCK_DOT_H__
 
-void dlm_rsb_toss_timer(struct timer_list *timer);
 void dlm_dump_rsb(struct dlm_rsb *r);
 void dlm_dump_rsb_name(struct dlm_ls *ls, const char *name, int len);
 void dlm_print_lkb(struct dlm_lkb *lkb);
@@ -19,15 +18,15 @@ void dlm_receive_message_saved(struct dlm_ls *ls, const struct dlm_message *ms,
 			       uint32_t saved_seq);
 void dlm_receive_buffer(const union dlm_packet *p, int nodeid);
 int dlm_modes_compat(int mode1, int mode2);
-void free_toss_rsb(struct dlm_rsb *r);
+void free_inactive_rsb(struct dlm_rsb *r);
 void dlm_put_rsb(struct dlm_rsb *r);
 void dlm_hold_rsb(struct dlm_rsb *r);
 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_lock_recovery(struct dlm_ls *ls);
 void dlm_unlock_recovery(struct dlm_ls *ls);
-void dlm_timer_resume(struct dlm_ls *ls);
+void dlm_rsb_scan(struct timer_list *timer);
+void resume_scan_timer(struct dlm_ls *ls);
 
 int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 		      int len, unsigned int flags, int *r_nodeid, int *result);

fs/dlm/lockspace.c

@@ -38,7 +38,7 @@ static ssize_t dlm_control_store(struct dlm_ls *ls, const char *buf, size_t len)
 
 	if (rc)
 		return rc;
-	ls = dlm_find_lockspace_local(ls->ls_local_handle);
+	ls = dlm_find_lockspace_local(ls);
 	if (!ls)
 		return -EINVAL;
 
@@ -265,18 +265,9 @@ struct dlm_ls *dlm_find_lockspace_global(uint32_t id)
 
 struct dlm_ls *dlm_find_lockspace_local(dlm_lockspace_t *lockspace)
 {
-	struct dlm_ls *ls;
+	struct dlm_ls *ls = lockspace;
 
-	spin_lock_bh(&lslist_lock);
+	atomic_inc(&ls->ls_count);
-	list_for_each_entry(ls, &lslist, ls_list) {
-		if (ls->ls_local_handle == lockspace) {
-			atomic_inc(&ls->ls_count);
-			goto out;
-		}
-	}
-	ls = NULL;
- out:
-	spin_unlock_bh(&lslist_lock);
 	return ls;
 }
 
@@ -410,37 +401,37 @@ static int new_lockspace(const char *name, const char *cluster,
 	atomic_set(&ls->ls_count, 0);
 	init_waitqueue_head(&ls->ls_count_wait);
 	ls->ls_flags = 0;
-	ls->ls_scan_time = jiffies;
 
 	if (ops && dlm_config.ci_recover_callbacks) {
 		ls->ls_ops = ops;
 		ls->ls_ops_arg = ops_arg;
 	}
 
+	if (flags & DLM_LSFL_SOFTIRQ)
+		set_bit(LSFL_SOFTIRQ, &ls->ls_flags);
+
 	/* ls_exflags are forced to match among nodes, and we don't
 	 * need to require all nodes to have some flags set
 	 */
-	ls->ls_exflags = (flags & ~(DLM_LSFL_FS | DLM_LSFL_NEWEXCL));
+	ls->ls_exflags = (flags & ~(DLM_LSFL_FS | DLM_LSFL_NEWEXCL |
+				    DLM_LSFL_SOFTIRQ));
 
-	INIT_LIST_HEAD(&ls->ls_toss);
+	INIT_LIST_HEAD(&ls->ls_slow_inactive);
-	INIT_LIST_HEAD(&ls->ls_keep);
+	INIT_LIST_HEAD(&ls->ls_slow_active);
 	rwlock_init(&ls->ls_rsbtbl_lock);
 
 	error = rhashtable_init(&ls->ls_rsbtbl, &dlm_rhash_rsb_params);
 	if (error)
 		goto out_lsfree;
 
-	idr_init(&ls->ls_lkbidr);
+	xa_init_flags(&ls->ls_lkbxa, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_BH);
-	rwlock_init(&ls->ls_lkbidr_lock);
+	rwlock_init(&ls->ls_lkbxa_lock);
 
 	INIT_LIST_HEAD(&ls->ls_waiters);
 	spin_lock_init(&ls->ls_waiters_lock);
 	INIT_LIST_HEAD(&ls->ls_orphans);
 	spin_lock_init(&ls->ls_orphans_lock);
 
-	INIT_LIST_HEAD(&ls->ls_new_rsb);
-	spin_lock_init(&ls->ls_new_rsb_spin);
-
 	INIT_LIST_HEAD(&ls->ls_nodes);
 	INIT_LIST_HEAD(&ls->ls_nodes_gone);
 	ls->ls_num_nodes = 0;
@@ -484,7 +475,7 @@ static int new_lockspace(const char *name, const char *cluster,
 	ls->ls_recover_buf = kmalloc(DLM_MAX_SOCKET_BUFSIZE, GFP_NOFS);
 	if (!ls->ls_recover_buf) {
 		error = -ENOMEM;
-		goto out_lkbidr;
+		goto out_lkbxa;
 	}
 
 	ls->ls_slot = 0;
@@ -494,32 +485,31 @@ static int new_lockspace(const char *name, const char *cluster,
 
 	INIT_LIST_HEAD(&ls->ls_recover_list);
 	spin_lock_init(&ls->ls_recover_list_lock);
-	idr_init(&ls->ls_recover_idr);
+	xa_init_flags(&ls->ls_recover_xa, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_BH);
-	spin_lock_init(&ls->ls_recover_idr_lock);
+	spin_lock_init(&ls->ls_recover_xa_lock);
 	ls->ls_recover_list_count = 0;
-	ls->ls_local_handle = ls;
 	init_waitqueue_head(&ls->ls_wait_general);
 	INIT_LIST_HEAD(&ls->ls_masters_list);
 	rwlock_init(&ls->ls_masters_lock);
 	INIT_LIST_HEAD(&ls->ls_dir_dump_list);
 	rwlock_init(&ls->ls_dir_dump_lock);
 
-	INIT_LIST_HEAD(&ls->ls_toss_q);
+	INIT_LIST_HEAD(&ls->ls_scan_list);
-	spin_lock_init(&ls->ls_toss_q_lock);
+	spin_lock_init(&ls->ls_scan_lock);
-	timer_setup(&ls->ls_timer, dlm_rsb_toss_timer,
+	timer_setup(&ls->ls_scan_timer, dlm_rsb_scan, TIMER_DEFERRABLE);
-		    TIMER_DEFERRABLE);
 
 	spin_lock_bh(&lslist_lock);
 	ls->ls_create_count = 1;
 	list_add(&ls->ls_list, &lslist);
 	spin_unlock_bh(&lslist_lock);
 
-	if (flags & DLM_LSFL_FS) {
+	if (flags & DLM_LSFL_FS)
-		error = dlm_callback_start(ls);
+		set_bit(LSFL_FS, &ls->ls_flags);
-		if (error) {
+
-			log_error(ls, "can't start dlm_callback %d", error);
+	error = dlm_callback_start(ls);
-			goto out_delist;
+	if (error) {
-		}
+		log_error(ls, "can't start dlm_callback %d", error);
+		goto out_delist;
 	}
 
 	init_waitqueue_head(&ls->ls_recover_lock_wait);
@@ -584,10 +574,10 @@ static int new_lockspace(const char *name, const char *cluster,
 	spin_lock_bh(&lslist_lock);
 	list_del(&ls->ls_list);
 	spin_unlock_bh(&lslist_lock);
-	idr_destroy(&ls->ls_recover_idr);
+	xa_destroy(&ls->ls_recover_xa);
 	kfree(ls->ls_recover_buf);
- out_lkbidr:
+ out_lkbxa:
-	idr_destroy(&ls->ls_lkbidr);
+	xa_destroy(&ls->ls_lkbxa);
 	rhashtable_destroy(&ls->ls_rsbtbl);
  out_lsfree:
 	if (do_unreg)
@@ -643,26 +633,15 @@ int dlm_new_user_lockspace(const char *name, const char *cluster,
 			   void *ops_arg, int *ops_result,
 			   dlm_lockspace_t **lockspace)
 {
+	if (flags & DLM_LSFL_SOFTIRQ)
+		return -EINVAL;
+
 	return __dlm_new_lockspace(name, cluster, flags, lvblen, ops,
 				   ops_arg, ops_result, lockspace);
 }
 
-static int lkb_idr_is_local(int id, void *p, void *data)
+static int lkb_idr_free(struct dlm_lkb *lkb)
 {
-	struct dlm_lkb *lkb = p;
-
-	return lkb->lkb_nodeid == 0 && lkb->lkb_grmode != DLM_LOCK_IV;
-}
-
-static int lkb_idr_is_any(int id, void *p, void *data)
-{
-	return 1;
-}
-
-static int lkb_idr_free(int id, void *p, void *data)
-{
-	struct dlm_lkb *lkb = p;
-
 	if (lkb->lkb_lvbptr && test_bit(DLM_IFL_MSTCPY_BIT, &lkb->lkb_iflags))
 		dlm_free_lvb(lkb->lkb_lvbptr);
 
@@ -670,23 +649,34 @@ static int lkb_idr_free(int id, void *p, void *data)
 	return 0;
 }
 
-/* NOTE: We check the lkbidr here rather than the resource table.
+/* NOTE: We check the lkbxa here rather than the resource table.
    This is because there may be LKBs queued as ASTs that have been unlinked
    from their RSBs and are pending deletion once the AST has been delivered */
 
 static int lockspace_busy(struct dlm_ls *ls, int force)
 {
-	int rv;
+	struct dlm_lkb *lkb;
+	unsigned long id;
+	int rv = 0;
 
-	read_lock_bh(&ls->ls_lkbidr_lock);
+	read_lock_bh(&ls->ls_lkbxa_lock);
 	if (force == 0) {
-		rv = idr_for_each(&ls->ls_lkbidr, lkb_idr_is_any, ls);
+		xa_for_each(&ls->ls_lkbxa, id, lkb) {
+			rv = 1;
+			break;
+		}
 	} else if (force == 1) {
-		rv = idr_for_each(&ls->ls_lkbidr, lkb_idr_is_local, ls);
+		xa_for_each(&ls->ls_lkbxa, id, lkb) {
+			if (lkb->lkb_nodeid == 0 &&
+			    lkb->lkb_grmode != DLM_LOCK_IV) {
+				rv = 1;
+				break;
+			}
+		}
 	} else {
 		rv = 0;
 	}
-	read_unlock_bh(&ls->ls_lkbidr_lock);
+	read_unlock_bh(&ls->ls_lkbxa_lock);
 	return rv;
 }
 
@@ -699,7 +689,8 @@ static void rhash_free_rsb(void *ptr, void *arg)
 
 static int release_lockspace(struct dlm_ls *ls, int force)
 {
-	struct dlm_rsb *rsb;
+	struct dlm_lkb *lkb;
+	unsigned long id;
 	int busy, rv;
 
 	busy = lockspace_busy(ls, force);
@@ -739,7 +730,7 @@ static int release_lockspace(struct dlm_ls *ls, int force)
 	 * time_shutdown_sync(), we don't care anymore
 	 */
 	clear_bit(LSFL_RUNNING, &ls->ls_flags);
-	timer_shutdown_sync(&ls->ls_timer);
+	timer_shutdown_sync(&ls->ls_scan_timer);
 
 	if (ls_count == 1) {
 		dlm_clear_members(ls);
@@ -752,28 +743,22 @@ static int release_lockspace(struct dlm_ls *ls, int force)
 
 	dlm_delete_debug_file(ls);
 
-	idr_destroy(&ls->ls_recover_idr);
+	xa_destroy(&ls->ls_recover_xa);
 	kfree(ls->ls_recover_buf);
 
 	/*
-	 * Free all lkb's in idr
+	 * Free all lkb's in xa
 	 */
-
+	xa_for_each(&ls->ls_lkbxa, id, lkb) {
-	idr_for_each(&ls->ls_lkbidr, lkb_idr_free, ls);
+		lkb_idr_free(lkb);
-	idr_destroy(&ls->ls_lkbidr);
+	}
+	xa_destroy(&ls->ls_lkbxa);
 
 	/*
 	 * Free all rsb's on rsbtbl
 	 */
 	rhashtable_free_and_destroy(&ls->ls_rsbtbl, rhash_free_rsb, NULL);
 
-	while (!list_empty(&ls->ls_new_rsb)) {
-		rsb = list_first_entry(&ls->ls_new_rsb, struct dlm_rsb,
-				       res_hashchain);
-		list_del(&rsb->res_hashchain);
-		dlm_free_rsb(rsb);
-	}
-
 	/*
 	 * Free structures on any other lists
 	 */

fs/dlm/lowcomms.c

@@ -461,7 +461,7 @@ static bool dlm_lowcomms_con_has_addr(const struct connection *con,
 	return false;
 }
 
-int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr, int len)
+int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr)
 {
 	struct connection *con;
 	bool ret, idx;
@@ -858,12 +858,6 @@ static void free_processqueue_entry(struct processqueue_entry *pentry)
 	kfree(pentry);
 }
 
-struct dlm_processed_nodes {
-	int nodeid;
-
-	struct list_head list;
-};
-
 static void process_dlm_messages(struct work_struct *work)
 {
 	struct processqueue_entry *pentry;

fs/dlm/lowcomms.h

@@ -46,7 +46,7 @@ void dlm_lowcomms_put_msg(struct dlm_msg *msg);
 int dlm_lowcomms_resend_msg(struct dlm_msg *msg);
 int dlm_lowcomms_connect_node(int nodeid);
 int dlm_lowcomms_nodes_set_mark(int nodeid, unsigned int mark);
-int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr, int len);
+int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr);
 void dlm_midcomms_receive_done(int nodeid);
 struct kmem_cache *dlm_lowcomms_writequeue_cache_create(void);
 struct kmem_cache *dlm_lowcomms_msg_cache_create(void);

fs/dlm/member.c

@@ -642,7 +642,7 @@ int dlm_ls_stop(struct dlm_ls *ls)
 	set_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
 	new = test_and_clear_bit(LSFL_RUNNING, &ls->ls_flags);
 	if (new)
-		timer_delete_sync(&ls->ls_timer);
+		timer_delete_sync(&ls->ls_scan_timer);
 	ls->ls_recover_seq++;
 
 	/* activate requestqueue and stop processing */

fs/dlm/memory.c

@@ -72,6 +72,8 @@ out:
 
 void dlm_memory_exit(void)
 {
+	rcu_barrier();
+
 	kmem_cache_destroy(writequeue_cache);
 	kmem_cache_destroy(mhandle_cache);
 	kmem_cache_destroy(msg_cache);
@@ -101,13 +103,19 @@ struct dlm_rsb *dlm_allocate_rsb(struct dlm_ls *ls)
 	return r;
 }
 
-void dlm_free_rsb(struct dlm_rsb *r)
+static void __free_rsb_rcu(struct rcu_head *rcu)
 {
+	struct dlm_rsb *r = container_of(rcu, struct dlm_rsb, rcu);
 	if (r->res_lvbptr)
 		dlm_free_lvb(r->res_lvbptr);
 	kmem_cache_free(rsb_cache, r);
 }
 
+void dlm_free_rsb(struct dlm_rsb *r)
+{
+	call_rcu(&r->rcu, __free_rsb_rcu);
+}
+
 struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls)
 {
 	struct dlm_lkb *lkb;

fs/dlm/midcomms.c

@@ -334,12 +334,12 @@ static struct midcomms_node *nodeid2node(int nodeid)
 	return __find_node(nodeid, nodeid_hash(nodeid));
 }
 
-int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len)
+int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr)
 {
 	int ret, idx, r = nodeid_hash(nodeid);
 	struct midcomms_node *node;
 
-	ret = dlm_lowcomms_addr(nodeid, addr, len);
+	ret = dlm_lowcomms_addr(nodeid, addr);
 	if (ret)
 		return ret;
 

fs/dlm/midcomms.h

@@ -19,7 +19,7 @@ int dlm_process_incoming_buffer(int nodeid, unsigned char *buf, int buflen);
 struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len, char **ppc);
 void dlm_midcomms_commit_mhandle(struct dlm_mhandle *mh, const void *name,
 				 int namelen);
-int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len);
+int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr);
 void dlm_midcomms_version_wait(void);
 int dlm_midcomms_close(int nodeid);
 int dlm_midcomms_start(void);

fs/dlm/recover.c

@@ -293,73 +293,78 @@ static void recover_list_clear(struct dlm_ls *ls)
 	spin_unlock_bh(&ls->ls_recover_list_lock);
 }
 
-static int recover_idr_empty(struct dlm_ls *ls)
+static int recover_xa_empty(struct dlm_ls *ls)
 {
 	int empty = 1;
 
-	spin_lock_bh(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_xa_lock);
 	if (ls->ls_recover_list_count)
 		empty = 0;
-	spin_unlock_bh(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
 
 	return empty;
 }
 
-static int recover_idr_add(struct dlm_rsb *r)
+static int recover_xa_add(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
+	struct xa_limit limit = {
+		.min = 1,
+		.max = UINT_MAX,
+	};
+	uint32_t id;
 	int rv;
 
-	spin_lock_bh(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_xa_lock);
 	if (r->res_id) {
 		rv = -1;
 		goto out_unlock;
 	}
-	rv = idr_alloc(&ls->ls_recover_idr, r, 1, 0, GFP_NOWAIT);
+	rv = xa_alloc(&ls->ls_recover_xa, &id, r, limit, GFP_ATOMIC);
 	if (rv < 0)
 		goto out_unlock;
 
-	r->res_id = rv;
+	r->res_id = id;
 	ls->ls_recover_list_count++;
 	dlm_hold_rsb(r);
 	rv = 0;
 out_unlock:
-	spin_unlock_bh(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
 	return rv;
 }
 
-static void recover_idr_del(struct dlm_rsb *r)
+static void recover_xa_del(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock_bh(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_xa_lock);
-	idr_remove(&ls->ls_recover_idr, r->res_id);
+	xa_erase_bh(&ls->ls_recover_xa, r->res_id);
 	r->res_id = 0;
 	ls->ls_recover_list_count--;
-	spin_unlock_bh(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
 
 	dlm_put_rsb(r);
 }
 
-static struct dlm_rsb *recover_idr_find(struct dlm_ls *ls, uint64_t id)
+static struct dlm_rsb *recover_xa_find(struct dlm_ls *ls, uint64_t id)
 {
 	struct dlm_rsb *r;
 
-	spin_lock_bh(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_xa_lock);
-	r = idr_find(&ls->ls_recover_idr, (int)id);
+	r = xa_load(&ls->ls_recover_xa, (int)id);
-	spin_unlock_bh(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
 	return r;
 }
 
-static void recover_idr_clear(struct dlm_ls *ls)
+static void recover_xa_clear(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r;
-	int id;
+	unsigned long id;
 
-	spin_lock_bh(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_xa_lock);
 
-	idr_for_each_entry(&ls->ls_recover_idr, r, id) {
+	xa_for_each(&ls->ls_recover_xa, id, r) {
-		idr_remove(&ls->ls_recover_idr, id);
+		xa_erase_bh(&ls->ls_recover_xa, id);
 		r->res_id = 0;
 		r->res_recover_locks_count = 0;
 		ls->ls_recover_list_count--;
@@ -372,7 +377,7 @@ static void recover_idr_clear(struct dlm_ls *ls)
 			  ls->ls_recover_list_count);
 		ls->ls_recover_list_count = 0;
 	}
-	spin_unlock_bh(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
 }
 
 
@@ -470,7 +475,7 @@ static int recover_master(struct dlm_rsb *r, unsigned int *count, uint64_t seq)
 		set_new_master(r);
 		error = 0;
 	} else {
-		recover_idr_add(r);
+		recover_xa_add(r);
 		error = dlm_send_rcom_lookup(r, dir_nodeid, seq);
 	}
 
@@ -551,10 +556,10 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
 
 	log_rinfo(ls, "dlm_recover_masters %u of %u", count, total);
 
-	error = dlm_wait_function(ls, &recover_idr_empty);
+	error = dlm_wait_function(ls, &recover_xa_empty);
  out:
 	if (error)
-		recover_idr_clear(ls);
+		recover_xa_clear(ls);
 	return error;
 }
 
@@ -563,7 +568,7 @@ int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc)
 	struct dlm_rsb *r;
 	int ret_nodeid, new_master;
 
-	r = recover_idr_find(ls, le64_to_cpu(rc->rc_id));
+	r = recover_xa_find(ls, le64_to_cpu(rc->rc_id));
 	if (!r) {
 		log_error(ls, "dlm_recover_master_reply no id %llx",
 			  (unsigned long long)le64_to_cpu(rc->rc_id));
@@ -582,9 +587,9 @@ int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc)
 	r->res_nodeid = new_master;
 	set_new_master(r);
 	unlock_rsb(r);
-	recover_idr_del(r);
+	recover_xa_del(r);
 
-	if (recover_idr_empty(ls))
+	if (recover_xa_empty(ls))
 		wake_up(&ls->ls_wait_general);
  out:
 	return 0;
@@ -877,29 +882,26 @@ void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list)
 		log_rinfo(ls, "dlm_recover_rsbs %d done", count);
 }
 
-/* Create a single list of all root rsb's to be used during recovery */
+void dlm_clear_inactive(struct dlm_ls *ls)
-
-void dlm_clear_toss(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r, *safe;
 	unsigned int count = 0;
 
 	write_lock_bh(&ls->ls_rsbtbl_lock);
-	list_for_each_entry_safe(r, safe, &ls->ls_toss, res_rsbs_list) {
+	list_for_each_entry_safe(r, safe, &ls->ls_slow_inactive, res_slow_list) {
-		list_del(&r->res_rsbs_list);
+		list_del(&r->res_slow_list);
 		rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
 				       dlm_rhash_rsb_params);
 
-		/* remove it from the toss queue if its part of it */
+		if (!list_empty(&r->res_scan_list))
-		if (!list_empty(&r->res_toss_q_list))
+			list_del_init(&r->res_scan_list);
-			list_del_init(&r->res_toss_q_list);
 
-		free_toss_rsb(r);
+		free_inactive_rsb(r);
 		count++;
 	}
 	write_unlock_bh(&ls->ls_rsbtbl_lock);
 
 	if (count)
-		log_rinfo(ls, "dlm_clear_toss %u done", count);
+		log_rinfo(ls, "dlm_clear_inactive %u done", count);
 }
 

fs/dlm/recover.h

@@ -25,7 +25,7 @@ int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc);
 int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
 		      const struct list_head *root_list);
 void dlm_recovered_lock(struct dlm_rsb *r);
-void dlm_clear_toss(struct dlm_ls *ls);
+void dlm_clear_inactive(struct dlm_ls *ls);
 void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list);
 
 #endif				/* __RECOVER_DOT_H__ */

fs/dlm/recoverd.c

@@ -33,7 +33,7 @@ static int dlm_create_masters_list(struct dlm_ls *ls)
 	}
 
 	read_lock_bh(&ls->ls_rsbtbl_lock);
-	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
+	list_for_each_entry(r, &ls->ls_slow_active, res_slow_list) {
 		if (r->res_nodeid)
 			continue;
 
@@ -63,12 +63,12 @@ static void dlm_create_root_list(struct dlm_ls *ls, struct list_head *root_list)
 	struct dlm_rsb *r;
 
 	read_lock_bh(&ls->ls_rsbtbl_lock);
-	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
+	list_for_each_entry(r, &ls->ls_slow_active, res_slow_list) {
 		list_add(&r->res_root_list, root_list);
 		dlm_hold_rsb(r);
 	}
 
-	WARN_ON_ONCE(!list_empty(&ls->ls_toss));
+	WARN_ON_ONCE(!list_empty(&ls->ls_slow_inactive));
 	read_unlock_bh(&ls->ls_rsbtbl_lock);
 }
 
@@ -98,16 +98,16 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq)
 	spin_lock_bh(&ls->ls_recover_lock);
 	if (ls->ls_recover_seq == seq) {
 		set_bit(LSFL_RUNNING, &ls->ls_flags);
-		/* Schedule next timer if recovery put something on toss.
+		/* Schedule next timer if recovery put something on inactive.
 		 *
 		 * The rsbs that was queued while recovery on toss hasn't
 		 * started yet because LSFL_RUNNING was set everything
 		 * else recovery hasn't started as well because ls_in_recovery
 		 * is still hold. So we should not run into the case that
-		 * dlm_timer_resume() queues a timer that can occur in
+		 * resume_scan_timer() queues a timer that can occur in
 		 * a no op.
 		 */
-		dlm_timer_resume(ls);
+		resume_scan_timer(ls);
 		/* unblocks processes waiting to enter the dlm */
 		up_write(&ls->ls_in_recovery);
 		clear_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
@@ -131,7 +131,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	dlm_callback_suspend(ls);
 
-	dlm_clear_toss(ls);
+	dlm_clear_inactive(ls);
 
 	/*
 	 * This list of root rsb's will be the basis of most of the recovery

fs/dlm/user.c

@@ -182,7 +182,7 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
 	struct dlm_user_args *ua;
 	struct dlm_user_proc *proc;
 	struct dlm_callback *cb;
-	int rv;
+	int rv, copy_lvb;
 
 	if (test_bit(DLM_DFL_ORPHAN_BIT, &lkb->lkb_dflags) ||
 	    test_bit(DLM_IFL_DEAD_BIT, &lkb->lkb_iflags))
@@ -213,28 +213,22 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
 
 	spin_lock_bh(&proc->asts_spin);
 
-	rv = dlm_queue_lkb_callback(lkb, flags, mode, status, sbflags, &cb);
+	if (!dlm_may_skip_callback(lkb, flags, mode, status, sbflags,
-	switch (rv) {
+				   &copy_lvb)) {
-	case DLM_ENQUEUE_CALLBACK_NEED_SCHED:
+		rv = dlm_get_cb(lkb, flags, mode, status, sbflags, &cb);
-		cb->ua = *ua;
+		if (!rv) {
-		cb->lkb_lksb = &cb->ua.lksb;
+			cb->copy_lvb = copy_lvb;
-		if (cb->copy_lvb) {
+			cb->ua = *ua;
-			memcpy(cb->lvbptr, ua->lksb.sb_lvbptr,
+			cb->lkb_lksb = &cb->ua.lksb;
-			       DLM_USER_LVB_LEN);
+			if (copy_lvb) {
-			cb->lkb_lksb->sb_lvbptr = cb->lvbptr;
+				memcpy(cb->lvbptr, ua->lksb.sb_lvbptr,
-		}
+				       DLM_USER_LVB_LEN);
+				cb->lkb_lksb->sb_lvbptr = cb->lvbptr;
+			}
 
-		list_add_tail(&cb->list, &proc->asts);
+			list_add_tail(&cb->list, &proc->asts);
-		wake_up_interruptible(&proc->wait);
+			wake_up_interruptible(&proc->wait);
-		break;
+		}
-	case DLM_ENQUEUE_CALLBACK_SUCCESS:
-		break;
-	case DLM_ENQUEUE_CALLBACK_FAILURE:
-		fallthrough;
-	default:
-		spin_unlock_bh(&proc->asts_spin);
-		WARN_ON_ONCE(1);
-		goto out;
 	}
 	spin_unlock_bh(&proc->asts_spin);
 
@@ -454,7 +448,7 @@ static int device_remove_lockspace(struct dlm_lspace_params *params)
 	if (params->flags & DLM_USER_LSFLG_FORCEFREE)
 		force = 2;
 
-	lockspace = ls->ls_local_handle;
+	lockspace = ls;
 	dlm_put_lockspace(ls);
 
 	/* The final dlm_release_lockspace waits for references to go to
@@ -657,7 +651,7 @@ static int device_open(struct inode *inode, struct file *file)
 		return -ENOMEM;
 	}
 
-	proc->lockspace = ls->ls_local_handle;
+	proc->lockspace = ls;
 	INIT_LIST_HEAD(&proc->asts);
 	INIT_LIST_HEAD(&proc->locks);
 	INIT_LIST_HEAD(&proc->unlocking);

include/linux/dlm.h

@@ -35,6 +35,9 @@ struct dlm_lockspace_ops {
 			      int num_slots, int our_slot, uint32_t generation);
 };
 
+/* only relevant for kernel lockspaces, will be removed in future */
+#define DLM_LSFL_SOFTIRQ __DLM_LSFL_RESERVED0
+
 /*
  * dlm_new_lockspace
  *
@@ -55,6 +58,11 @@ struct dlm_lockspace_ops {
  *   used to select the directory node.  Must be the same on all nodes.
  * DLM_LSFL_NEWEXCL
  *   dlm_new_lockspace() should return -EEXIST if the lockspace exists.
+ * DLM_LSFL_SOFTIRQ
+ *   dlm request callbacks (ast, bast) are softirq safe. Flag should be
+ *   preferred by users. Will be default in some future. If set the
+ *   strongest context for ast, bast callback is softirq as it avoids
+ *   an additional context switch.
  *
  * lvblen: length of lvb in bytes.  Must be multiple of 8.
  *   dlm_new_lockspace() returns an error if this does not match
@@ -121,7 +129,14 @@ int dlm_release_lockspace(dlm_lockspace_t *lockspace, int force);
  * call.
  *
  * AST routines should not block (at least not for long), but may make
- * any locking calls they please.
+ * any locking calls they please. If DLM_LSFL_SOFTIRQ for kernel
+ * users of dlm_new_lockspace() is passed the ast and bast callbacks
+ * can be processed in softirq context. Also some of the callback
+ * contexts are in the same context as the DLM lock request API, users
+ * must not hold locks while calling dlm lock request API and trying
+ * to acquire this lock in the callback again, this will end in a
+ * lock recursion. For newer implementation the DLM_LSFL_SOFTIRQ
+ * should be used.
  */
 
 int dlm_lock(dlm_lockspace_t *lockspace,

include/uapi/linux/dlm.h

@@ -71,6 +71,8 @@ struct dlm_lksb {
 /* DLM_LSFL_TIMEWARN is deprecated and reserved. DO NOT USE! */
 #define DLM_LSFL_TIMEWARN	0x00000002
 #define DLM_LSFL_NEWEXCL     	0x00000008
+/* currently reserved due in-kernel use */
+#define __DLM_LSFL_RESERVED0	0x00000010
 
 
 #endif /* _UAPI__DLM_DOT_H__ */