[PATCH] device-mapper raid1: drop mark_region spinlock fix

[Alasdair G Kergon]

The spinlock region_lock is held while calling mark_region which can
sleep.  Drop the spinlock before calling that function.

A region's state and inclusion in the clean list are altered
by rh_inc and rh_dec.  The state variable is set to RH_CLEAN in rh_dec, 
but only if 'pending' is zero.  It is set to RH_DIRTY in rh_inc, but 
not if it is already so.  The changes to 'pending', the state, and 
the region's inclusion in the clean list need to be atomicly.

From: Jonathan E Brassow <jbrassow@xxxxxxxxxx>
Signed-Off-By: Alasdair G Kergon <agk@xxxxxxxxxx>

Index: linux-2.6.14/drivers/md/dm-raid1.c
===================================================================
--- linux-2.6.14.orig/drivers/md/dm-raid1.c	2005-10-26 20:14:27.000000000 +0100
+++ linux-2.6.14/drivers/md/dm-raid1.c	2005-10-26 20:23:03.000000000 +0100
@@ -376,16 +376,18 @@ static void rh_inc(struct region_hash *r
 	read_lock(&rh->hash_lock);
 	reg = __rh_find(rh, region);
 
+	spin_lock_irq(&rh->region_lock);
 	atomic_inc(&reg->pending);
 
-	spin_lock_irq(&rh->region_lock);
 	if (reg->state == RH_CLEAN) {
-		rh->log->type->mark_region(rh->log, reg->key);
-
 		reg->state = RH_DIRTY;
 		list_del_init(&reg->list);	/* take off the clean list */
-	}
-	spin_unlock_irq(&rh->region_lock);
+		spin_unlock_irq(&rh->region_lock);
+
+		rh->log->type->mark_region(rh->log, reg->key);
+	} else
+		spin_unlock_irq(&rh->region_lock);
+
 
 	read_unlock(&rh->hash_lock);
 }
@@ -408,21 +410,17 @@ static void rh_dec(struct region_hash *r
 	reg = __rh_lookup(rh, region);
 	read_unlock(&rh->hash_lock);
 
+	spin_lock_irqsave(&rh->region_lock, flags);
 	if (atomic_dec_and_test(&reg->pending)) {
-		spin_lock_irqsave(&rh->region_lock, flags);
-		if (atomic_read(&reg->pending)) { /* check race */
-			spin_unlock_irqrestore(&rh->region_lock, flags);
-			return;
-		}
 		if (reg->state == RH_RECOVERING) {
 			list_add_tail(&reg->list, &rh->quiesced_regions);
 		} else {
 			reg->state = RH_CLEAN;
 			list_add(&reg->list, &rh->clean_regions);
 		}
-		spin_unlock_irqrestore(&rh->region_lock, flags);
 		should_wake = 1;
 	}
+	spin_unlock_irqrestore(&rh->region_lock, flags);
 
 	if (should_wake)
 		wake();
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/