[PATCH] device-mapper snapshot: metadata reading separation

[PATCH] device-mapper snapshot: metadata reading separation

[Alasdair G Kergon]

More snapshot metadata reading into separate function, to prepare
for changing the place it gets called from.

Signed-Off-By: Alasdair G Kergon <agk@redhat.com>

Index: linux-2.6.14/drivers/md/dm-snap.c
===================================================================
--- linux-2.6.14.orig/drivers/md/dm-snap.c	2005-11-17 22:17:55.000000000 +0000
+++ linux-2.6.14/drivers/md/dm-snap.c	2005-11-17 22:18:25.000000000 +0000
@@ -371,6 +371,20 @@ static inline ulong round_up(ulong n, ul
 	return (n + size) & ~size;
 }
 
+static void read_snapshot_metadata(struct dm_snapshot *s)
+{
+	if (s->have_metadata)
+		return;
+
+	if (s->store.read_metadata(&s->store)) {
+		down_write(&s->lock);
+		s->valid = 0;
+		up_write(&s->lock);
+	}
+
+	s->have_metadata = 1;
+}
+
 /*
  * Construct a snapshot mapping: <origin_dev> <COW-dev> <p/n> <chunk-size>
  */
@@ -848,16 +862,7 @@ static void snapshot_resume(struct dm_ta
 {
 	struct dm_snapshot *s = (struct dm_snapshot *) ti->private;
 
-	if (s->have_metadata)
-		return;
-
-	if (s->store.read_metadata(&s->store)) {
-		down_write(&s->lock);
-		s->valid = 0;
-		up_write(&s->lock);
-	}
-
-	s->have_metadata = 1;
+	read_snapshot_metadata(s);
 }
 
 static int snapshot_status(struct dm_target *ti, status_type_t type,

Re: [PATCH] device-mapper snapshot: metadata reading separation

[Andrew Morton]

Alasdair G Kergon <agk@redhat.com> wrote:
>
> +static void read_snapshot_metadata(struct dm_snapshot *s)
>  +{
>  +	if (s->have_metadata)
>  +		return;
>  +
>  +	if (s->store.read_metadata(&s->store)) {
>  +		down_write(&s->lock);
>  +		s->valid = 0;
>  +		up_write(&s->lock);
>  +	}
>  +
>  +	s->have_metadata = 1;
>  +}
>  +

I always get suspicious when I see a lock around a plain assignment. 
Sometimes it's legitimate, usually when some other user of the LHS expects
its value to be stable across an entire locked region - it is read multiple
times and those reads are expected to return the same thing.  (I can't
think of any other case).

Maybe this is such a case?