Merge tag 'dm-3.20-changes-2' of...

Merge tag 'dm-3.20-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull more device mapper changes from Mike Snitzer:

- Significant dm-crypt CPU scalability performance improvements thanks
  to changes that enable effective use of an unbound workqueue across
  all available CPUs.  A large battery of tests were performed to
  validate these changes, summary of results is available here:
  https://www.redhat.com/archives/dm-devel/2015-February/msg00106.html

- A few additional stable fixes (to DM core, dm-snapshot and dm-mirror)
  and a small fix to the dm-space-map-disk.

* tag 'dm-3.20-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm snapshot: fix a possible invalid memory access on unload
  dm: fix a race condition in dm_get_md
  dm crypt: sort writes
  dm crypt: add 'submit_from_crypt_cpus' option
  dm crypt: offload writes to thread
  dm crypt: remove unused io_pool and _crypt_io_pool
  dm crypt: avoid deadlock in mempools
  dm crypt: don't allocate pages for a partial request
  dm crypt: use unbound workqueue for request processing
  dm io: reject unsupported DISCARD requests with EOPNOTSUPP
  dm mirror: do not degrade the mirror on discard error
  dm space map disk: fix sm_disk_count_is_more_than_one()

Documentation/device-mapper/dm-crypt.txt

@@ -51,7 +51,7 @@ Parameters: <cipher> <key> <iv_offset> <device path> \
     Otherwise #opt_params is the number of following arguments.
 
     Example of optional parameters section:
-        1 allow_discards
+        3 allow_discards same_cpu_crypt submit_from_crypt_cpus
 
 allow_discards
     Block discard requests (a.k.a. TRIM) are passed through the crypt device.
@@ -63,6 +63,19 @@ allow_discards
     used space etc.) if the discarded blocks can be located easily on the
     device later.
 
+same_cpu_crypt
+    Perform encryption using the same cpu that IO was submitted on.
+    The default is to use an unbound workqueue so that encryption work
+    is automatically balanced between available CPUs.
+
+submit_from_crypt_cpus
+    Disable offloading writes to a separate thread after encryption.
+    There are some situations where offloading write bios from the
+    encryption threads to a single thread degrades performance
+    significantly.  The default is to offload write bios to the same
+    thread because it benefits CFQ to have writes submitted using the
+    same context.
+
 Example scripts
 ===============
 LUKS (Linux Unified Key Setup) is now the preferred way to set up disk

drivers/md/dm-io.c

@@ -290,6 +290,12 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
 	unsigned short logical_block_size = queue_logical_block_size(q);
 	sector_t num_sectors;
 
+	/* Reject unsupported discard requests */
+	if ((rw & REQ_DISCARD) && !blk_queue_discard(q)) {
+		dec_count(io, region, -EOPNOTSUPP);
+		return;
+	}
+
 	/*
 	 * where->count may be zero if rw holds a flush and we need to
 	 * send a zero-sized flush.

drivers/md/dm-raid1.c

@@ -604,6 +604,15 @@ static void write_callback(unsigned long error, void *context)
 		return;
 	}
 
+	/*
+	 * If the bio is discard, return an error, but do not
+	 * degrade the array.
+	 */
+	if (bio->bi_rw & REQ_DISCARD) {
+		bio_endio(bio, -EOPNOTSUPP);
+		return;
+	}
+
 	for (i = 0; i < ms->nr_mirrors; i++)
 		if (test_bit(i, &error))
 			fail_mirror(ms->mirror + i, DM_RAID1_WRITE_ERROR);

drivers/md/dm-snap.c

@@ -1432,8 +1432,6 @@ static void pending_complete(struct dm_snap_pending_exception *pe, int success)
 		full_bio->bi_private = pe->full_bio_private;
 		atomic_inc(&full_bio->bi_remaining);
 	}
-	free_pending_exception(pe);
-
 	increment_pending_exceptions_done_count();
 
 	up_write(&s->lock);
@@ -1450,6 +1448,8 @@ static void pending_complete(struct dm_snap_pending_exception *pe, int success)
 	}
 
 	retry_origin_bios(s, origin_bios);
+
+	free_pending_exception(pe);
 }
 
 static void commit_callback(void *context, int success)

drivers/md/dm.c

@@ -2571,7 +2571,7 @@ int dm_setup_md_queue(struct mapped_device *md)
 	return 0;
 }
 
-static struct mapped_device *dm_find_md(dev_t dev)
+struct mapped_device *dm_get_md(dev_t dev)
 {
 	struct mapped_device *md;
 	unsigned minor = MINOR(dev);
@@ -2582,12 +2582,15 @@ static struct mapped_device *dm_find_md(dev_t dev)
 	spin_lock(&_minor_lock);
 
 	md = idr_find(&_minor_idr, minor);
-	if (md && (md == MINOR_ALLOCED ||
-		   (MINOR(disk_devt(dm_disk(md))) != minor) ||
-		   dm_deleting_md(md) ||
-		   test_bit(DMF_FREEING, &md->flags))) {
-		md = NULL;
-		goto out;
+	if (md) {
+		if ((md == MINOR_ALLOCED ||
+		     (MINOR(disk_devt(dm_disk(md))) != minor) ||
+		     dm_deleting_md(md) ||
+		     test_bit(DMF_FREEING, &md->flags))) {
+			md = NULL;
+			goto out;
+		}
+		dm_get(md);
 	}
 
 out:
@@ -2595,16 +2598,6 @@ static struct mapped_device *dm_find_md(dev_t dev)
 
 	return md;
 }
-
-struct mapped_device *dm_get_md(dev_t dev)
-{
-	struct mapped_device *md = dm_find_md(dev);
-
-	if (md)
-		dm_get(md);
-
-	return md;
-}
 EXPORT_SYMBOL_GPL(dm_get_md);
 
 void *dm_get_mdptr(struct mapped_device *md)

drivers/md/persistent-data/dm-space-map-disk.c

@@ -78,7 +78,9 @@ static int sm_disk_count_is_more_than_one(struct dm_space_map *sm, dm_block_t b,
 	if (r)
 		return r;
 
-	return count > 1;
+	*result = count > 1;
+
+	return 0;
 }
 
 static int sm_disk_set_count(struct dm_space_map *sm, dm_block_t b,