Merge tag 'dlm-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm

Pull dlm updates from David Teigland:
 "This set includes some small fixes, and some big internal changes:

   - Fix a long standing race between the unlock callback for the last
     lkb struct, and removing the rsb that became unused after the final
     unlock. This could lead different nodes to inconsistent info about
     the rsb master node.

   - Remove unnecessary refcounting on callback structs, returning to
     the way things were done in the past.

   - Do message processing in softirq context. This allows dlm messages
     to be cleared more quickly and efficiently, reducing long lists of
     incomplete requests. A future change to run callbacks directly from
     this context will make this more effective.

   - The softirq message processing involved a number of patches
     changing mutexes to spinlocks and rwlocks, and a fair amount of
     code re-org in preparation.

   - Use an rhashtable for rsb structs, rather than our old internal
     hash table implementation. This also required some re-org of lists
     and locks preparation for the change.

   - Drop the dlm_scand kthread, and use timers to clear unused rsb
     structs. Scanning all rsb's periodically was a lot of wasted work.

   - Fix recent regression in logic for copying LVB data in user space
     lock requests"

* tag 'dlm-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm: (34 commits)
  dlm: return -ENOMEM if ls_recover_buf fails
  dlm: fix sleep in atomic context
  dlm: use rwlock for lkbidr
  dlm: use rwlock for rsb hash table
  dlm: drop dlm_scand kthread and use timers
  dlm: do not use ref counts for rsb in the toss state
  dlm: switch to use rhashtable for rsbs
  dlm: add rsb lists for iteration
  dlm: merge toss and keep hash table lists into one list
  dlm: change to single hashtable lock
  dlm: increment ls_count for dlm_scand
  dlm: do message processing in softirq context
  dlm: use spin_lock_bh for message processing
  dlm: remove schedule in receive path
  dlm: convert ls_recv_active from rw_semaphore to rwlock
  dlm: avoid blocking receive at the end of recovery
  dlm: convert res_lock to spinlock
  dlm: convert ls_waiters_mutex to spinlock
  dlm: drop mutex use in waiters recovery
  dlm: add new struct to save position in dlm_copy_master_names
  ...

1 files changed, 48 insertions, 101 deletions

diff --git a/fs/dlm/recover.c b/fs/dlm/recover.c
index 53917c0aa3c0..f493d5f30c58 100644
--- a/fs/dlm/recover.c
+++ b/fs/dlm/recover.c
@@ -74,9 +74,9 @@ int dlm_wait_function(struct dlm_ls *ls, int (*testfn) (struct dlm_ls *ls))
 uint32_t dlm_recover_status(struct dlm_ls *ls)
 {
 	uint32_t status;
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	status = ls->ls_recover_status;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 	return status;
 }
 
@@ -87,9 +87,9 @@ static void _set_recover_status(struct dlm_ls *ls, uint32_t status)
 
 void dlm_set_recover_status(struct dlm_ls *ls, uint32_t status)
 {
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	_set_recover_status(ls, status);
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 }
 
 static int wait_status_all(struct dlm_ls *ls, uint32_t wait_status,
@@ -188,13 +188,13 @@ int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq)
 
 		rv = dlm_slots_assign(ls, &num_slots, &slots_size, &slots, &gen);
 		if (!rv) {
-			spin_lock(&ls->ls_recover_lock);
+			spin_lock_bh(&ls->ls_recover_lock);
 			_set_recover_status(ls, DLM_RS_NODES_ALL);
 			ls->ls_num_slots = num_slots;
 			ls->ls_slots_size = slots_size;
 			ls->ls_slots = slots;
 			ls->ls_generation = gen;
-			spin_unlock(&ls->ls_recover_lock);
+			spin_unlock_bh(&ls->ls_recover_lock);
 		} else {
 			dlm_set_recover_status(ls, DLM_RS_NODES_ALL);
 		}
@@ -241,9 +241,9 @@ static int recover_list_empty(struct dlm_ls *ls)
 {
 	int empty;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	empty = list_empty(&ls->ls_recover_list);
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 
 	return empty;
 }
@@ -252,23 +252,23 @@ static void recover_list_add(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	if (list_empty(&r->res_recover_list)) {
 		list_add_tail(&r->res_recover_list, &ls->ls_recover_list);
 		ls->ls_recover_list_count++;
 		dlm_hold_rsb(r);
 	}
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 }
 
 static void recover_list_del(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	list_del_init(&r->res_recover_list);
 	ls->ls_recover_list_count--;
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 
 	dlm_put_rsb(r);
 }
@@ -277,7 +277,7 @@ static void recover_list_clear(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r, *s;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	list_for_each_entry_safe(r, s, &ls->ls_recover_list, res_recover_list) {
 		list_del_init(&r->res_recover_list);
 		r->res_recover_locks_count = 0;
@@ -290,17 +290,17 @@ static void recover_list_clear(struct dlm_ls *ls)
 			  ls->ls_recover_list_count);
 		ls->ls_recover_list_count = 0;
 	}
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 }
 
 static int recover_idr_empty(struct dlm_ls *ls)
 {
 	int empty = 1;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	if (ls->ls_recover_list_count)
 		empty = 0;
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 
 	return empty;
 }
@@ -310,8 +310,7 @@ static int recover_idr_add(struct dlm_rsb *r)
 	struct dlm_ls *ls = r->res_ls;
 	int rv;
 
-	idr_preload(GFP_NOFS);
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	if (r->res_id) {
 		rv = -1;
 		goto out_unlock;
@@ -325,8 +324,7 @@ static int recover_idr_add(struct dlm_rsb *r)
 	dlm_hold_rsb(r);
 	rv = 0;
 out_unlock:
-	spin_unlock(&ls->ls_recover_idr_lock);
-	idr_preload_end();
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 	return rv;
 }
 
@@ -334,11 +332,11 @@ static void recover_idr_del(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	idr_remove(&ls->ls_recover_idr, r->res_id);
 	r->res_id = 0;
 	ls->ls_recover_list_count--;
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 
 	dlm_put_rsb(r);
 }
@@ -347,9 +345,9 @@ static struct dlm_rsb *recover_idr_find(struct dlm_ls *ls, uint64_t id)
 {
 	struct dlm_rsb *r;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	r = idr_find(&ls->ls_recover_idr, (int)id);
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 	return r;
 }
 
@@ -358,7 +356,7 @@ static void recover_idr_clear(struct dlm_ls *ls)
 	struct dlm_rsb *r;
 	int id;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 
 	idr_for_each_entry(&ls->ls_recover_idr, r, id) {
 		idr_remove(&ls->ls_recover_idr, id);
@@ -374,7 +372,7 @@ static void recover_idr_clear(struct dlm_ls *ls)
 			  ls->ls_recover_list_count);
 		ls->ls_recover_list_count = 0;
 	}
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 }
 
 
@@ -521,7 +519,8 @@ static int recover_master_static(struct dlm_rsb *r, unsigned int *count)
  * the correct dir node.
  */
 
-int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
+int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
+			const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	unsigned int total = 0;
@@ -531,10 +530,8 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
 
 	log_rinfo(ls, "dlm_recover_masters");
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		if (dlm_recovery_stopped(ls)) {
-			up_read(&ls->ls_root_sem);
 			error = -EINTR;
 			goto out;
 		}
@@ -548,12 +545,9 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
 		cond_resched();
 		total++;
 
-		if (error) {
-			up_read(&ls->ls_root_sem);
+		if (error)
 			goto out;
-		}
 	}
-	up_read(&ls->ls_root_sem);
 
 	log_rinfo(ls, "dlm_recover_masters %u of %u", count, total);
 
@@ -658,13 +652,13 @@ static int recover_locks(struct dlm_rsb *r, uint64_t seq)
 	return error;
 }
 
-int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq)
+int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
+		      const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	int error, count = 0;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		if (is_master(r)) {
 			rsb_clear_flag(r, RSB_NEW_MASTER);
 			continue;
@@ -675,19 +669,15 @@ int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq)
 
 		if (dlm_recovery_stopped(ls)) {
 			error = -EINTR;
-			up_read(&ls->ls_root_sem);
 			goto out;
 		}
 
 		error = recover_locks(r, seq);
-		if (error) {
-			up_read(&ls->ls_root_sem);
+		if (error)
 			goto out;
-		}
 
 		count += r->res_recover_locks_count;
 	}
-	up_read(&ls->ls_root_sem);
 
 	log_rinfo(ls, "dlm_recover_locks %d out", count);
 
@@ -856,13 +846,12 @@ static void recover_grant(struct dlm_rsb *r)
 		rsb_set_flag(r, RSB_RECOVER_GRANT);
 }
 
-void dlm_recover_rsbs(struct dlm_ls *ls)
+void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	unsigned int count = 0;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		lock_rsb(r);
 		if (is_master(r)) {
 			if (rsb_flag(r, RSB_RECOVER_CONVERT))
@@ -883,7 +872,6 @@ void dlm_recover_rsbs(struct dlm_ls *ls)
 		rsb_clear_flag(r, RSB_NEW_MASTER2);
 		unlock_rsb(r);
 	}
-	up_read(&ls->ls_root_sem);
 
 	if (count)
 		log_rinfo(ls, "dlm_recover_rsbs %d done", count);
@@ -891,66 +879,25 @@ void dlm_recover_rsbs(struct dlm_ls *ls)
 
 /* Create a single list of all root rsb's to be used during recovery */
 
-int dlm_create_root_list(struct dlm_ls *ls)
-{
-	struct rb_node *n;
-	struct dlm_rsb *r;
-	int i, error = 0;
-
-	down_write(&ls->ls_root_sem);
-	if (!list_empty(&ls->ls_root_list)) {
-		log_error(ls, "root list not empty");
-		error = -EINVAL;
-		goto out;
-	}
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		spin_lock(&ls->ls_rsbtbl[i].lock);
-		for (n = rb_first(&ls->ls_rsbtbl[i].keep); n; n = rb_next(n)) {
-			r = rb_entry(n, struct dlm_rsb, res_hashnode);
-			list_add(&r->res_root_list, &ls->ls_root_list);
-			dlm_hold_rsb(r);
-		}
-
-		if (!RB_EMPTY_ROOT(&ls->ls_rsbtbl[i].toss))
-			log_error(ls, "dlm_create_root_list toss not empty");
-		spin_unlock(&ls->ls_rsbtbl[i].lock);
-	}
- out:
-	up_write(&ls->ls_root_sem);
-	return error;
-}
-
-void dlm_release_root_list(struct dlm_ls *ls)
+void dlm_clear_toss(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r, *safe;
+	unsigned int count = 0;
 
-	down_write(&ls->ls_root_sem);
-	list_for_each_entry_safe(r, safe, &ls->ls_root_list, res_root_list) {
-		list_del_init(&r->res_root_list);
-		dlm_put_rsb(r);
-	}
-	up_write(&ls->ls_root_sem);
-}
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry_safe(r, safe, &ls->ls_toss, res_rsbs_list) {
+		list_del(&r->res_rsbs_list);
+		rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
+				       dlm_rhash_rsb_params);
 
-void dlm_clear_toss(struct dlm_ls *ls)
-{
-	struct rb_node *n, *next;
-	struct dlm_rsb *r;
-	unsigned int count = 0;
-	int i;
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		spin_lock(&ls->ls_rsbtbl[i].lock);
-		for (n = rb_first(&ls->ls_rsbtbl[i].toss); n; n = next) {
-			next = rb_next(n);
-			r = rb_entry(n, struct dlm_rsb, res_hashnode);
-			rb_erase(n, &ls->ls_rsbtbl[i].toss);
-			dlm_free_rsb(r);
-			count++;
-		}
-		spin_unlock(&ls->ls_rsbtbl[i].lock);
+		/* remove it from the toss queue if its part of it */
+		if (!list_empty(&r->res_toss_q_list))
+			list_del_init(&r->res_toss_q_list);
+
+		free_toss_rsb(r);
+		count++;
 	}
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
 
 	if (count)
 		log_rinfo(ls, "dlm_clear_toss %u done", count);