Merge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm:
  dm: detect lost queue
  dm: publish dm_vcalloc
  dm: publish dm_table_unplug_all
  dm: publish dm_get_mapinfo
  dm: export struct dm_dev
  dm crypt: avoid unnecessary wait when splitting bio
  dm crypt: tidy ctx pending
  dm crypt: fix async inc_pending
  dm crypt: move dec_pending on error into write_io_submit
  dm crypt: remove inc_pending from write_io_submit
  dm crypt: tidy write loop pending
  dm crypt: tidy crypt alloc
  dm crypt: tidy inc pending
  dm exception store: use chunk_t for_areas
  dm exception store: introduce area_location function
  dm raid1: kcopyd should stop on error if errors handled
  dm mpath: remove is_active from struct dm_path
  dm mpath: use more error codes

Fixed up trivial conflict in drivers/md/dm-mpath.c manually.

Author: torvalds

drivers/md/dm-crypt.c

@@ -333,7 +333,6 @@ static void crypt_convert_init(struct crypt_config *cc,
 	ctx->idx_out = bio_out ? bio_out->bi_idx : 0;
 	ctx->sector = sector + cc->iv_offset;
 	init_completion(&ctx->restart);
-	atomic_set(&ctx->pending, 1);
 }
 
 static int crypt_convert_block(struct crypt_config *cc,
@@ -408,6 +407,8 @@ static int crypt_convert(struct crypt_config *cc,
 {
 	int r;
 
+	atomic_set(&ctx->pending, 1);
+
 	while(ctx->idx_in < ctx->bio_in->bi_vcnt &&
 	      ctx->idx_out < ctx->bio_out->bi_vcnt) {
 
@@ -456,9 +457,11 @@ static void dm_crypt_bio_destructor(struct bio *bio)
 /*
  * Generate a new unfragmented bio with the given size
  * This should never violate the device limitations
- * May return a smaller bio when running out of pages
+ * May return a smaller bio when running out of pages, indicated by
+ * *out_of_pages set to 1.
  */
-static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size)
+static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size,
+				      unsigned *out_of_pages)
 {
 	struct crypt_config *cc = io->target->private;
 	struct bio *clone;
@@ -472,11 +475,14 @@ static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size)
 		return NULL;
 
 	clone_init(io, clone);
+	*out_of_pages = 0;
 
 	for (i = 0; i < nr_iovecs; i++) {
 		page = mempool_alloc(cc->page_pool, gfp_mask);
-		if (!page)
+		if (!page) {
+			*out_of_pages = 1;
 			break;
+		}
 
 		/*
 		 * if additional pages cannot be allocated without waiting,
@@ -517,6 +523,27 @@ static void crypt_free_buffer_pages(struct crypt_config *cc, struct bio *clone)
 	}
 }
 
+static struct dm_crypt_io *crypt_io_alloc(struct dm_target *ti,
+					  struct bio *bio, sector_t sector)
+{
+	struct crypt_config *cc = ti->private;
+	struct dm_crypt_io *io;
+
+	io = mempool_alloc(cc->io_pool, GFP_NOIO);
+	io->target = ti;
+	io->base_bio = bio;
+	io->sector = sector;
+	io->error = 0;
+	atomic_set(&io->pending, 0);
+
+	return io;
+}
+
+static void crypt_inc_pending(struct dm_crypt_io *io)
+{
+	atomic_inc(&io->pending);
+}
+
 /*
  * One of the bios was finished. Check for completion of
  * the whole request and correctly clean up the buffer.
@@ -591,7 +618,7 @@ static void kcryptd_io_read(struct dm_crypt_io *io)
 	struct bio *base_bio = io->base_bio;
 	struct bio *clone;
 
-	atomic_inc(&io->pending);
+	crypt_inc_pending(io);
 
 	/*
 	 * The block layer might modify the bvec array, so always
@@ -653,6 +680,7 @@ static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io,
 		crypt_free_buffer_pages(cc, clone);
 		bio_put(clone);
 		io->error = -EIO;
+		crypt_dec_pending(io);
 		return;
 	}
 
@@ -664,66 +692,67 @@ static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io,
 
 	if (async)
 		kcryptd_queue_io(io);
-	else {
-		atomic_inc(&io->pending);
+	else
 		generic_make_request(clone);
-	}
 }
 
-static void kcryptd_crypt_write_convert_loop(struct dm_crypt_io *io)
+static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
 {
 	struct crypt_config *cc = io->target->private;
 	struct bio *clone;
+	int crypt_finished;
+	unsigned out_of_pages = 0;
 	unsigned remaining = io->base_bio->bi_size;
 	int r;
 
+	/*
+	 * Prevent io from disappearing until this function completes.
+	 */
+	crypt_inc_pending(io);
+	crypt_convert_init(cc, &io->ctx, NULL, io->base_bio, io->sector);
+
 	/*
 	 * The allocated buffers can be smaller than the whole bio,
 	 * so repeat the whole process until all the data can be handled.
 	 */
 	while (remaining) {
-		clone = crypt_alloc_buffer(io, remaining);
+		clone = crypt_alloc_buffer(io, remaining, &out_of_pages);
 		if (unlikely(!clone)) {
 			io->error = -ENOMEM;
-			return;
+			break;
 		}
 
 		io->ctx.bio_out = clone;
 		io->ctx.idx_out = 0;
 
 		remaining -= clone->bi_size;
 
+		crypt_inc_pending(io);
 		r = crypt_convert(cc, &io->ctx);
+		crypt_finished = atomic_dec_and_test(&io->ctx.pending);
 
-		if (atomic_dec_and_test(&io->ctx.pending)) {
-			/* processed, no running async crypto  */
+		/* Encryption was already finished, submit io now */
+		if (crypt_finished) {
 			kcryptd_crypt_write_io_submit(io, r, 0);
-			if (unlikely(r < 0))
-				return;
-		} else
-			atomic_inc(&io->pending);
 
-		/* out of memory -> run queues */
-		if (unlikely(remaining)) {
-			/* wait for async crypto then reinitialize pending */
-			wait_event(cc->writeq, !atomic_read(&io->ctx.pending));
-			atomic_set(&io->ctx.pending, 1);
-			congestion_wait(WRITE, HZ/100);
+			/*
+			 * If there was an error, do not try next fragments.
+			 * For async, error is processed in async handler.
+			 */
+			if (unlikely(r < 0))
+				break;
 		}
-	}
-}
 
-static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
-{
-	struct crypt_config *cc = io->target->private;
-
-	/*
-	 * Prevent io from disappearing until this function completes.
-	 */
-	atomic_inc(&io->pending);
+		/*
+		 * Out of memory -> run queues
+		 * But don't wait if split was due to the io size restriction
+		 */
+		if (unlikely(out_of_pages))
+			congestion_wait(WRITE, HZ/100);
 
-	crypt_convert_init(cc, &io->ctx, NULL, io->base_bio, io->sector);
-	kcryptd_crypt_write_convert_loop(io);
+		if (unlikely(remaining))
+			wait_event(cc->writeq, !atomic_read(&io->ctx.pending));
+	}
 
 	crypt_dec_pending(io);
 }
@@ -741,7 +770,7 @@ static void kcryptd_crypt_read_convert(struct dm_crypt_io *io)
 	struct crypt_config *cc = io->target->private;
 	int r = 0;
 
-	atomic_inc(&io->pending);
+	crypt_inc_pending(io);
 
 	crypt_convert_init(cc, &io->ctx, io->base_bio, io->base_bio,
 			   io->sector);
@@ -1108,15 +1137,9 @@ static void crypt_dtr(struct dm_target *ti)
 static int crypt_map(struct dm_target *ti, struct bio *bio,
 		     union map_info *map_context)
 {
-	struct crypt_config *cc = ti->private;
 	struct dm_crypt_io *io;
 
-	io = mempool_alloc(cc->io_pool, GFP_NOIO);
-	io->target = ti;
-	io->base_bio = bio;
-	io->sector = bio->bi_sector - ti->begin;
-	io->error = 0;
-	atomic_set(&io->pending, 0);
+	io = crypt_io_alloc(ti, bio, bio->bi_sector - ti->begin);
 
 	if (bio_data_dir(io->base_bio) == READ)
 		kcryptd_queue_io(io);

drivers/md/dm-exception-store.c

@@ -108,12 +108,12 @@ struct pstore {
 	 * Used to keep track of which metadata area the data in
 	 * 'chunk' refers to.
 	 */
-	uint32_t current_area;
+	chunk_t current_area;
 
 	/*
 	 * The next free chunk for an exception.
 	 */
-	uint32_t next_free;
+	chunk_t next_free;
 
 	/*
 	 * The index of next free exception in the current
@@ -175,7 +175,7 @@ static void do_metadata(struct work_struct *work)
 /*
  * Read or write a chunk aligned and sized block of data from a device.
  */
-static int chunk_io(struct pstore *ps, uint32_t chunk, int rw, int metadata)
+static int chunk_io(struct pstore *ps, chunk_t chunk, int rw, int metadata)
 {
 	struct dm_io_region where = {
 		.bdev = ps->snap->cow->bdev,
@@ -208,17 +208,24 @@ static int chunk_io(struct pstore *ps, uint32_t chunk, int rw, int metadata)
 	return req.result;
 }
 
+/*
+ * Convert a metadata area index to a chunk index.
+ */
+static chunk_t area_location(struct pstore *ps, chunk_t area)
+{
+	return 1 + ((ps->exceptions_per_area + 1) * area);
+}
+
 /*
  * Read or write a metadata area.  Remembering to skip the first
  * chunk which holds the header.
  */
-static int area_io(struct pstore *ps, uint32_t area, int rw)
+static int area_io(struct pstore *ps, chunk_t area, int rw)
 {
 	int r;
-	uint32_t chunk;
+	chunk_t chunk;
 
-	/* convert a metadata area index to a chunk index */
-	chunk = 1 + ((ps->exceptions_per_area + 1) * area);
+	chunk = area_location(ps, area);
 
 	r = chunk_io(ps, chunk, rw, 0);
 	if (r)
@@ -228,7 +235,7 @@ static int area_io(struct pstore *ps, uint32_t area, int rw)
 	return 0;
 }
 
-static int zero_area(struct pstore *ps, uint32_t area)
+static int zero_area(struct pstore *ps, chunk_t area)
 {
 	memset(ps->area, 0, ps->snap->chunk_size << SECTOR_SHIFT);
 	return area_io(ps, area, WRITE);
@@ -404,7 +411,7 @@ static int insert_exceptions(struct pstore *ps, int *full)
 
 static int read_exceptions(struct pstore *ps)
 {
-	uint32_t area;
+	chunk_t area;
 	int r, full = 1;
 
 	/*
@@ -517,6 +524,7 @@ static int persistent_prepare(struct exception_store *store,
 {
 	struct pstore *ps = get_info(store);
 	uint32_t stride;
+	chunk_t next_free;
 	sector_t size = get_dev_size(store->snap->cow->bdev);
 
 	/* Is there enough room ? */
@@ -530,7 +538,8 @@ static int persistent_prepare(struct exception_store *store,
 	 * into account the location of the metadata chunks.
 	 */
 	stride = (ps->exceptions_per_area + 1);
-	if ((++ps->next_free % stride) == 1)
+	next_free = ++ps->next_free;
+	if (sector_div(next_free, stride) == 1)
 		ps->next_free++;
 
 	atomic_inc(&ps->pending_count);

drivers/md/dm-ioctl.c

@@ -1131,7 +1131,7 @@ static void retrieve_deps(struct dm_table *table,
 	unsigned int count = 0;
 	struct list_head *tmp;
 	size_t len, needed;
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 	struct dm_target_deps *deps;
 
 	deps = get_result_buffer(param, param_size, &len);
@@ -1157,7 +1157,7 @@ static void retrieve_deps(struct dm_table *table,
 	deps->count = count;
 	count = 0;
 	list_for_each_entry (dd, dm_table_get_devices(table), list)
-		deps->dev[count++] = huge_encode_dev(dd->bdev->bd_dev);
+		deps->dev[count++] = huge_encode_dev(dd->dm_dev.bdev->bd_dev);
 
 	param->data_size = param->data_start + needed;
 }