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, 9 insertions, 34 deletions

diff --git a/fs/dlm/requestqueue.c b/fs/dlm/requestqueue.c
index 892d6ca21e74..719a5243a069 100644
--- a/fs/dlm/requestqueue.c
+++ b/fs/dlm/requestqueue.c
@@ -37,7 +37,7 @@ void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
 	int length = le16_to_cpu(ms->m_header.h_length) -
 		sizeof(struct dlm_message);
 
-	e = kmalloc(sizeof(struct rq_entry) + length, GFP_NOFS);
+	e = kmalloc(sizeof(struct rq_entry) + length, GFP_ATOMIC);
 	if (!e) {
 		log_print("dlm_add_requestqueue: out of memory len %d", length);
 		return;
@@ -48,10 +48,7 @@ void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
 	memcpy(&e->request, ms, sizeof(*ms));
 	memcpy(&e->request.m_extra, ms->m_extra, length);
 
-	atomic_inc(&ls->ls_requestqueue_cnt);
-	mutex_lock(&ls->ls_requestqueue_mutex);
 	list_add_tail(&e->list, &ls->ls_requestqueue);
-	mutex_unlock(&ls->ls_requestqueue_mutex);
 }
 
 /*
@@ -71,16 +68,14 @@ int dlm_process_requestqueue(struct dlm_ls *ls)
 	struct dlm_message *ms;
 	int error = 0;
 
-	mutex_lock(&ls->ls_requestqueue_mutex);
-
+	write_lock_bh(&ls->ls_requestqueue_lock);
 	for (;;) {
 		if (list_empty(&ls->ls_requestqueue)) {
-			mutex_unlock(&ls->ls_requestqueue_mutex);
+			clear_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
 			error = 0;
 			break;
 		}
-		e = list_entry(ls->ls_requestqueue.next, struct rq_entry, list);
-		mutex_unlock(&ls->ls_requestqueue_mutex);
+		e = list_first_entry(&ls->ls_requestqueue, struct rq_entry, list);
 
 		ms = &e->request;
 
@@ -93,41 +88,23 @@ int dlm_process_requestqueue(struct dlm_ls *ls)
 			  e->recover_seq);
 
 		dlm_receive_message_saved(ls, &e->request, e->recover_seq);
-
-		mutex_lock(&ls->ls_requestqueue_mutex);
 		list_del(&e->list);
-		if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
-			wake_up(&ls->ls_requestqueue_wait);
 		kfree(e);
 
 		if (dlm_locking_stopped(ls)) {
 			log_debug(ls, "process_requestqueue abort running");
-			mutex_unlock(&ls->ls_requestqueue_mutex);
 			error = -EINTR;
 			break;
 		}
+		write_unlock_bh(&ls->ls_requestqueue_lock);
 		schedule();
+		write_lock_bh(&ls->ls_requestqueue_lock);
 	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
 
 	return error;
 }
 
-/*
- * After recovery is done, locking is resumed and dlm_recoverd takes all the
- * saved requests and processes them as they would have been by dlm_recv.  At
- * the same time, dlm_recv will start receiving new requests from remote nodes.
- * We want to delay dlm_recv processing new requests until dlm_recoverd has
- * finished processing the old saved requests.  We don't check for locking
- * stopped here because dlm_ls_stop won't stop locking until it's suspended us
- * (dlm_recv).
- */
-
-void dlm_wait_requestqueue(struct dlm_ls *ls)
-{
-	wait_event(ls->ls_requestqueue_wait,
-		   atomic_read(&ls->ls_requestqueue_cnt) == 0);
-}
-
 static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
 {
 	__le32 type = ms->m_type;
@@ -158,17 +135,15 @@ void dlm_purge_requestqueue(struct dlm_ls *ls)
 	struct dlm_message *ms;
 	struct rq_entry *e, *safe;
 
-	mutex_lock(&ls->ls_requestqueue_mutex);
+	write_lock_bh(&ls->ls_requestqueue_lock);
 	list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
 		ms =  &e->request;
 
 		if (purge_request(ls, ms, e->nodeid)) {
 			list_del(&e->list);
-			if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
-				wake_up(&ls->ls_requestqueue_wait);
 			kfree(e);
 		}
 	}
-	mutex_unlock(&ls->ls_requestqueue_mutex);
+	write_unlock_bh(&ls->ls_requestqueue_lock);
 }