1 files changed, 618 insertions, 388 deletions

diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 4fea3c33af2a..6664ea73ccf8 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -171,6 +171,13 @@ static void ceph_osd_data_bvecs_init(struct ceph_osd_data *osd_data,
 	osd_data->num_bvecs = num_bvecs;
 }
 
+static void ceph_osd_iter_init(struct ceph_osd_data *osd_data,
+			       struct iov_iter *iter)
+{
+	osd_data->type = CEPH_OSD_DATA_TYPE_ITER;
+	osd_data->iter = *iter;
+}
+
 static struct ceph_osd_data *
 osd_req_op_raw_data_in(struct ceph_osd_request *osd_req, unsigned int which)
 {
@@ -213,16 +220,6 @@ void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
 }
 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);
 
-void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
-			unsigned int which, struct ceph_pagelist *pagelist)
-{
-	struct ceph_osd_data *osd_data;
-
-	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
-	ceph_osd_data_pagelist_init(osd_data, pagelist);
-}
-EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);
-
 #ifdef CONFIG_BLOCK
 void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
 				    unsigned int which,
@@ -264,28 +261,31 @@ void osd_req_op_extent_osd_data_bvec_pos(struct ceph_osd_request *osd_req,
 }
 EXPORT_SYMBOL(osd_req_op_extent_osd_data_bvec_pos);
 
-static void osd_req_op_cls_request_info_pagelist(
-			struct ceph_osd_request *osd_req,
-			unsigned int which, struct ceph_pagelist *pagelist)
+/**
+ * osd_req_op_extent_osd_iter - Set up an operation with an iterator buffer
+ * @osd_req: The request to set up
+ * @which: Index of the operation in which to set the iter
+ * @iter: The buffer iterator
+ */
+void osd_req_op_extent_osd_iter(struct ceph_osd_request *osd_req,
+				unsigned int which, struct iov_iter *iter)
 {
 	struct ceph_osd_data *osd_data;
 
-	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
-	ceph_osd_data_pagelist_init(osd_data, pagelist);
+	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
+	ceph_osd_iter_init(osd_data, iter);
 }
+EXPORT_SYMBOL(osd_req_op_extent_osd_iter);
 
-void osd_req_op_cls_request_data_pagelist(
+static void osd_req_op_cls_request_info_pagelist(
 			struct ceph_osd_request *osd_req,
 			unsigned int which, struct ceph_pagelist *pagelist)
 {
 	struct ceph_osd_data *osd_data;
 
-	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
+	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
 	ceph_osd_data_pagelist_init(osd_data, pagelist);
-	osd_req->r_ops[which].cls.indata_len += pagelist->length;
-	osd_req->r_ops[which].indata_len += pagelist->length;
 }
-EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);
 
 void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req,
 			unsigned int which, struct page **pages, u64 length,
@@ -346,6 +346,8 @@ static u64 ceph_osd_data_length(struct ceph_osd_data *osd_data)
 #endif /* CONFIG_BLOCK */
 	case CEPH_OSD_DATA_TYPE_BVECS:
 		return osd_data->bvec_pos.iter.bi_size;
+	case CEPH_OSD_DATA_TYPE_ITER:
+		return iov_iter_count(&osd_data->iter);
 	default:
 		WARN(true, "unrecognized data type %d\n", (int)osd_data->type);
 		return 0;
@@ -376,8 +378,10 @@ static void osd_req_op_data_release(struct ceph_osd_request *osd_req,
 
 	switch (op->op) {
 	case CEPH_OSD_OP_READ:
+	case CEPH_OSD_OP_SPARSE_READ:
 	case CEPH_OSD_OP_WRITE:
 	case CEPH_OSD_OP_WRITEFULL:
+		kfree(op->extent.sparse_ext);
 		ceph_osd_data_release(&op->extent.osd_data);
 		break;
 	case CEPH_OSD_OP_CALL:
@@ -445,8 +449,10 @@ static void target_copy(struct ceph_osd_request_target *dest,
 	dest->size = src->size;
 	dest->min_size = src->min_size;
 	dest->sort_bitwise = src->sort_bitwise;
+	dest->recovery_deletes = src->recovery_deletes;
 
 	dest->flags = src->flags;
+	dest->used_replica = src->used_replica;
 	dest->paused = src->paused;
 
 	dest->epoch = src->epoch;
@@ -523,7 +529,7 @@ EXPORT_SYMBOL(ceph_osdc_put_request);
 
 static void request_init(struct ceph_osd_request *req)
 {
-	/* req only, each op is zeroed in _osd_req_op_init() */
+	/* req only, each op is zeroed in osd_req_op_init() */
 	memset(req, 0, sizeof(*req));
 
 	kref_init(&req->r_kref);
@@ -535,43 +541,6 @@ static void request_init(struct ceph_osd_request *req)
 	target_init(&req->r_t);
 }
 
-/*
- * This is ugly, but it allows us to reuse linger registration and ping
- * requests, keeping the structure of the code around send_linger{_ping}()
- * reasonable.  Setting up a min_nr=2 mempool for each linger request
- * and dealing with copying ops (this blasts req only, watch op remains
- * intact) isn't any better.
- */
-static void request_reinit(struct ceph_osd_request *req)
-{
-	struct ceph_osd_client *osdc = req->r_osdc;
-	bool mempool = req->r_mempool;
-	unsigned int num_ops = req->r_num_ops;
-	u64 snapid = req->r_snapid;
-	struct ceph_snap_context *snapc = req->r_snapc;
-	bool linger = req->r_linger;
-	struct ceph_msg *request_msg = req->r_request;
-	struct ceph_msg *reply_msg = req->r_reply;
-
-	dout("%s req %p\n", __func__, req);
-	WARN_ON(kref_read(&req->r_kref) != 1);
-	request_release_checks(req);
-
-	WARN_ON(kref_read(&request_msg->kref) != 1);
-	WARN_ON(kref_read(&reply_msg->kref) != 1);
-	target_destroy(&req->r_t);
-
-	request_init(req);
-	req->r_osdc = osdc;
-	req->r_mempool = mempool;
-	req->r_num_ops = num_ops;
-	req->r_snapid = snapid;
-	req->r_snapc = snapc;
-	req->r_linger = linger;
-	req->r_request = request_msg;
-	req->r_reply = reply_msg;
-}
-
 struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
 					       struct ceph_snap_context *snapc,
 					       unsigned int num_ops,
@@ -704,6 +673,7 @@ static void get_num_data_items(struct ceph_osd_request *req,
 		/* reply */
 		case CEPH_OSD_OP_STAT:
 		case CEPH_OSD_OP_READ:
+		case CEPH_OSD_OP_SPARSE_READ:
 		case CEPH_OSD_OP_LIST_WATCHERS:
 			*num_reply_data_items += 1;
 			break;
@@ -744,8 +714,8 @@ EXPORT_SYMBOL(ceph_osdc_alloc_messages);
  * other information associated with them.  It also serves as a
  * common init routine for all the other init functions, below.
  */
-static struct ceph_osd_req_op *
-_osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
+struct ceph_osd_req_op *
+osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
 		 u16 opcode, u32 flags)
 {
 	struct ceph_osd_req_op *op;
@@ -760,12 +730,6 @@ _osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
 
 	return op;
 }
-
-void osd_req_op_init(struct ceph_osd_request *osd_req,
-		     unsigned int which, u16 opcode, u32 flags)
-{
-	(void)_osd_req_op_init(osd_req, which, opcode, flags);
-}
 EXPORT_SYMBOL(osd_req_op_init);
 
 void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
@@ -773,13 +737,13 @@ void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
 				u64 offset, u64 length,
 				u64 truncate_size, u32 truncate_seq)
 {
-	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
-						      opcode, 0);
+	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which,
+						     opcode, 0);
 	size_t payload_len = 0;
 
 	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
 	       opcode != CEPH_OSD_OP_WRITEFULL && opcode != CEPH_OSD_OP_ZERO &&
-	       opcode != CEPH_OSD_OP_TRUNCATE);
+	       opcode != CEPH_OSD_OP_TRUNCATE && opcode != CEPH_OSD_OP_SPARSE_READ);
 
 	op->extent.offset = offset;
 	op->extent.length = length;
@@ -820,7 +784,7 @@ void osd_req_op_extent_dup_last(struct ceph_osd_request *osd_req,
 	BUG_ON(which + 1 >= osd_req->r_num_ops);
 
 	prev_op = &osd_req->r_ops[which];
-	op = _osd_req_op_init(osd_req, which + 1, prev_op->op, prev_op->flags);
+	op = osd_req_op_init(osd_req, which + 1, prev_op->op, prev_op->flags);
 	/* dup previous one */
 	op->indata_len = prev_op->indata_len;
 	op->outdata_len = prev_op->outdata_len;
@@ -843,7 +807,7 @@ int osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
 	size_t size;
 	int ret;
 
-	op = _osd_req_op_init(osd_req, which, CEPH_OSD_OP_CALL, 0);
+	op = osd_req_op_init(osd_req, which, CEPH_OSD_OP_CALL, 0);
 
 	pagelist = ceph_pagelist_alloc(GFP_NOFS);
 	if (!pagelist)
@@ -881,8 +845,8 @@ int osd_req_op_xattr_init(struct ceph_osd_request *osd_req, unsigned int which,
 			  u16 opcode, const char *name, const void *value,
 			  size_t size, u8 cmp_op, u8 cmp_mode)
 {
-	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
-						      opcode, 0);
+	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which,
+						     opcode, 0);
 	struct ceph_pagelist *pagelist;
 	size_t payload_len;
 	int ret;
@@ -922,14 +886,30 @@ EXPORT_SYMBOL(osd_req_op_xattr_init);
  * @watch_opcode: CEPH_OSD_WATCH_OP_*
  */
 static void osd_req_op_watch_init(struct ceph_osd_request *req, int which,
-				  u64 cookie, u8 watch_opcode)
+				  u8 watch_opcode, u64 cookie, u32 gen)
 {
 	struct ceph_osd_req_op *op;
 
-	op = _osd_req_op_init(req, which, CEPH_OSD_OP_WATCH, 0);
+	op = osd_req_op_init(req, which, CEPH_OSD_OP_WATCH, 0);
 	op->watch.cookie = cookie;
 	op->watch.op = watch_opcode;
-	op->watch.gen = 0;
+	op->watch.gen = gen;
+}
+
+/*
+ * prot_ver, timeout and notify payload (may be empty) should already be
+ * encoded in @request_pl
+ */
+static void osd_req_op_notify_init(struct ceph_osd_request *req, int which,
+				   u64 cookie, struct ceph_pagelist *request_pl)
+{
+	struct ceph_osd_req_op *op;
+
+	op = osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY, 0);
+	op->notify.cookie = cookie;
+
+	ceph_osd_data_pagelist_init(&op->notify.request_data, request_pl);
+	op->indata_len = request_pl->length;
 }
 
 /*
@@ -941,10 +921,9 @@ void osd_req_op_alloc_hint_init(struct ceph_osd_request *osd_req,
 				u64 expected_write_size,
 				u32 flags)
 {
-	struct ceph_osd_req_op *op = _osd_req_op_init(osd_req, which,
-						      CEPH_OSD_OP_SETALLOCHINT,
-						      0);
+	struct ceph_osd_req_op *op;
 
+	op = osd_req_op_init(osd_req, which, CEPH_OSD_OP_SETALLOCHINT, 0);
 	op->alloc_hint.expected_object_size = expected_object_size;
 	op->alloc_hint.expected_write_size = expected_write_size;
 	op->alloc_hint.flags = flags;
@@ -977,6 +956,8 @@ static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
 #endif
 	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BVECS) {
 		ceph_msg_data_add_bvecs(msg, &osd_data->bvec_pos);
+	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_ITER) {
+		ceph_msg_data_add_iter(msg, &osd_data->iter);
 	} else {
 		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
 	}
@@ -989,6 +970,7 @@ static u32 osd_req_encode_op(struct ceph_osd_op *dst,
 	case CEPH_OSD_OP_STAT:
 		break;
 	case CEPH_OSD_OP_READ:
+	case CEPH_OSD_OP_SPARSE_READ:
 	case CEPH_OSD_OP_WRITE:
 	case CEPH_OSD_OP_WRITEFULL:
 	case CEPH_OSD_OP_ZERO:
@@ -1043,6 +1025,10 @@ static u32 osd_req_encode_op(struct ceph_osd_op *dst,
 		dst->copy_from.src_fadvise_flags =
 			cpu_to_le32(src->copy_from.src_fadvise_flags);
 		break;
+	case CEPH_OSD_OP_ASSERT_VER:
+		dst->assert_ver.unused = cpu_to_le64(0);
+		dst->assert_ver.ver = cpu_to_le64(src->assert_ver.ver);
+		break;
 	default:
 		pr_err("unsupported osd opcode %s\n",
 			ceph_osd_op_name(src->op));
@@ -1085,7 +1071,8 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 
 	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE &&
 	       opcode != CEPH_OSD_OP_ZERO && opcode != CEPH_OSD_OP_TRUNCATE &&
-	       opcode != CEPH_OSD_OP_CREATE && opcode != CEPH_OSD_OP_DELETE);
+	       opcode != CEPH_OSD_OP_CREATE && opcode != CEPH_OSD_OP_DELETE &&
+	       opcode != CEPH_OSD_OP_SPARSE_READ);
 
 	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
 					GFP_NOFS);
@@ -1117,24 +1104,39 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 				       truncate_size, truncate_seq);
 	}
 
-	req->r_flags = flags;
 	req->r_base_oloc.pool = layout->pool_id;
 	req->r_base_oloc.pool_ns = ceph_try_get_string(layout->pool_ns);
 	ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum);
+	req->r_flags = flags | osdc->client->options->read_from_replica;
 
 	req->r_snapid = vino.snap;
 	if (flags & CEPH_OSD_FLAG_WRITE)
 		req->r_data_offset = off;
 
-	if (num_ops > 1)
+	if (num_ops > 1) {
+		int num_req_ops, num_rep_ops;
+
 		/*
-		 * This is a special case for ceph_writepages_start(), but it
-		 * also covers ceph_uninline_data().  If more multi-op request
-		 * use cases emerge, we will need a separate helper.
+		 * If this is a multi-op write request, assume that we'll need
+		 * request ops. If it's a multi-op read then assume we'll need
+		 * reply ops. Anything else and call it -EINVAL.
 		 */
-		r = __ceph_osdc_alloc_messages(req, GFP_NOFS, num_ops, 0);
-	else
+		if (flags & CEPH_OSD_FLAG_WRITE) {
+			num_req_ops = num_ops;
+			num_rep_ops = 0;
+		} else if (flags & CEPH_OSD_FLAG_READ) {
+			num_req_ops = 0;
+			num_rep_ops = num_ops;
+		} else {
+			r = -EINVAL;
+			goto fail;
+		}
+
+		r = __ceph_osdc_alloc_messages(req, GFP_NOFS, num_req_ops,
+					       num_rep_ops);
+	} else {
 		r = ceph_osdc_alloc_messages(req, GFP_NOFS);
+	}
 	if (r)
 		goto fail;
 
@@ -1146,6 +1148,20 @@ fail:
 }
 EXPORT_SYMBOL(ceph_osdc_new_request);
 
+int __ceph_alloc_sparse_ext_map(struct ceph_osd_req_op *op, int cnt)
+{
+	WARN_ON(op->op != CEPH_OSD_OP_SPARSE_READ);
+
+	op->extent.sparse_ext_cnt = cnt;
+	op->extent.sparse_ext = kmalloc_array(cnt,
+					      sizeof(*op->extent.sparse_ext),
+					      GFP_NOFS);
+	if (!op->extent.sparse_ext)
+		return -ENOMEM;
+	return 0;
+}
+EXPORT_SYMBOL(__ceph_alloc_sparse_ext_map);
+
 /*
  * We keep osd requests in an rbtree, sorted by ->r_tid.
  */
@@ -1203,6 +1219,7 @@ static void osd_init(struct ceph_osd *osd)
 {
 	refcount_set(&osd->o_ref, 1);
 	RB_CLEAR_NODE(&osd->o_node);
+	spin_lock_init(&osd->o_requests_lock);
 	osd->o_requests = RB_ROOT;
 	osd->o_linger_requests = RB_ROOT;
 	osd->o_backoff_mappings = RB_ROOT;
@@ -1213,6 +1230,13 @@ static void osd_init(struct ceph_osd *osd)
 	mutex_init(&osd->lock);
 }
 
+static void ceph_init_sparse_read(struct ceph_sparse_read *sr)
+{
+	kfree(sr->sr_extent);
+	memset(sr, '\0', sizeof(*sr));
+	sr->sr_state = CEPH_SPARSE_READ_HDR;
+}
+
 static void osd_cleanup(struct ceph_osd *osd)
 {
 	WARN_ON(!RB_EMPTY_NODE(&osd->o_node));
@@ -1223,6 +1247,8 @@ static void osd_cleanup(struct ceph_osd *osd)
 	WARN_ON(!list_empty(&osd->o_osd_lru));
 	WARN_ON(!list_empty(&osd->o_keepalive_item));
 
+	ceph_init_sparse_read(&osd->o_sparse_read);
+
 	if (osd->o_auth.authorizer) {
 		WARN_ON(osd_homeless(osd));
 		ceph_auth_destroy_authorizer(osd->o_auth.authorizer);
@@ -1242,6 +1268,9 @@ static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
 	osd_init(osd);
 	osd->o_osdc = osdc;
 	osd->o_osd = onum;
+	osd->o_sparse_op_idx = -1;
+
+	ceph_init_sparse_read(&osd->o_sparse_read);
 
 	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
 
@@ -1432,7 +1461,9 @@ static void link_request(struct ceph_osd *osd, struct ceph_osd_request *req)
 		atomic_inc(&osd->o_osdc->num_homeless);
 
 	get_osd(osd);
+	spin_lock(&osd->o_requests_lock);
 	insert_request(&osd->o_requests, req);
+	spin_unlock(&osd->o_requests_lock);
 	req->r_osd = osd;
 }
 
@@ -1444,7 +1475,9 @@ static void unlink_request(struct ceph_osd *osd, struct ceph_osd_request *req)
 	     req, req->r_tid);
 
 	req->r_osd = NULL;
+	spin_lock(&osd->o_requests_lock);
 	erase_request(&osd->o_requests, req);
+	spin_unlock(&osd->o_requests_lock);
 	put_osd(osd);
 
 	if (!osd_homeless(osd))
@@ -1505,7 +1538,7 @@ static bool target_should_be_paused(struct ceph_osd_client *osdc,
 
 static int pick_random_replica(const struct ceph_osds *acting)
 {
-	int i = prandom_u32() % acting->size;
+	int i = get_random_u32_below(acting->size);
 
 	dout("%s picked osd%d, primary osd%d\n", __func__,
 	     acting->osds[i], acting->primary);
@@ -2042,6 +2075,7 @@ static void setup_request_data(struct ceph_osd_request *req)
 					       &op->raw_data_in);
 			break;
 		case CEPH_OSD_OP_READ:
+		case CEPH_OSD_OP_SPARSE_READ:
 			ceph_osdc_msg_data_add(reply_msg,
 					       &op->extent.osd_data);
 			break;
@@ -2390,7 +2424,11 @@ again:
 		if (ceph_test_opt(osdc->client, ABORT_ON_FULL)) {
 			err = -ENOSPC;
 		} else {
-			pr_warn_ratelimited("FULL or reached pool quota\n");
+			if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL))
+				pr_warn_ratelimited("cluster is full (osdmap FULL)\n");
+			else
+				pr_warn_ratelimited("pool %lld is full or reached quota\n",
+						    req->r_t.base_oloc.pool);
 			req->r_t.paused = true;
 			maybe_request_map(osdc);
 		}
@@ -2431,14 +2469,11 @@ promote:
 
 static void account_request(struct ceph_osd_request *req)
 {
-	struct ceph_osd_client *osdc = req->r_osdc;
-
 	WARN_ON(req->r_flags & (CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK));
 	WARN_ON(!(req->r_flags & (CEPH_OSD_FLAG_READ | CEPH_OSD_FLAG_WRITE)));
 
 	req->r_flags |= CEPH_OSD_FLAG_ONDISK;
-	req->r_flags |= osdc->client->options->osd_req_flags;
-	atomic_inc(&osdc->num_requests);
+	atomic_inc(&req->r_osdc->num_requests);
 
 	req->r_start_stamp = jiffies;
 	req->r_start_latency = ktime_get();
@@ -2460,8 +2495,10 @@ static void finish_request(struct ceph_osd_request *req)
 
 	req->r_end_latency = ktime_get();
 
-	if (req->r_osd)
+	if (req->r_osd) {
+		ceph_init_sparse_read(&req->r_osd->o_sparse_read);
 		unlink_request(req->r_osd, req);
+	}
 	atomic_dec(&osdc->num_requests);
 
 	/*
@@ -2735,10 +2772,13 @@ static void linger_release(struct kref *kref)
 	WARN_ON(!list_empty(&lreq->pending_lworks));
 	WARN_ON(lreq->osd);
 
-	if (lreq->reg_req)
-		ceph_osdc_put_request(lreq->reg_req);
-	if (lreq->ping_req)
-		ceph_osdc_put_request(lreq->ping_req);
+	if (lreq->request_pl)
+		ceph_pagelist_release(lreq->request_pl);
+	if (lreq->notify_id_pages)
+		ceph_release_page_vector(lreq->notify_id_pages, 1);
+
+	ceph_osdc_put_request(lreq->reg_req);
+	ceph_osdc_put_request(lreq->ping_req);
 	target_destroy(&lreq->t);
 	kfree(lreq);
 }
@@ -3007,6 +3047,12 @@ static void linger_commit_cb(struct ceph_osd_request *req)
 	struct ceph_osd_linger_request *lreq = req->r_priv;
 
 	mutex_lock(&lreq->lock);
+	if (req != lreq->reg_req) {
+		dout("%s lreq %p linger_id %llu unknown req (%p != %p)\n",
+		     __func__, lreq, lreq->linger_id, req, lreq->reg_req);
+		goto out;
+	}
+
 	dout("%s lreq %p linger_id %llu result %d\n", __func__, lreq,
 	     lreq->linger_id, req->r_result);
 	linger_reg_commit_complete(lreq, req->r_result);
@@ -3030,6 +3076,7 @@ static void linger_commit_cb(struct ceph_osd_request *req)
 		}
 	}
 
+out:
 	mutex_unlock(&lreq->lock);
 	linger_put(lreq);
 }
@@ -3052,6 +3099,12 @@ static void linger_reconnect_cb(struct ceph_osd_request *req)
 	struct ceph_osd_linger_request *lreq = req->r_priv;
 
 	mutex_lock(&lreq->lock);
+	if (req != lreq->reg_req) {
+		dout("%s lreq %p linger_id %llu unknown req (%p != %p)\n",
+		     __func__, lreq, lreq->linger_id, req, lreq->reg_req);
+		goto out;
+	}
+
 	dout("%s lreq %p linger_id %llu result %d last_error %d\n", __func__,
 	     lreq, lreq->linger_id, req->r_result, lreq->last_error);
 	if (req->r_result < 0) {
@@ -3061,48 +3114,64 @@ static void linger_reconnect_cb(struct ceph_osd_request *req)
 		}
 	}
 
+out:
 	mutex_unlock(&lreq->lock);
 	linger_put(lreq);
 }
 
 static void send_linger(struct ceph_osd_linger_request *lreq)
 {
-	struct ceph_osd_request *req = lreq->reg_req;
-	struct ceph_osd_req_op *op = &req->r_ops[0];
+	struct ceph_osd_client *osdc = lreq->osdc;
+	struct ceph_osd_request *req;
+	int ret;
 
-	verify_osdc_wrlocked(req->r_osdc);
+	verify_osdc_wrlocked(osdc);
+	mutex_lock(&lreq->lock);
 	dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
 
-	if (req->r_osd)
-		cancel_linger_request(req);
+	if (lreq->reg_req) {
+		if (lreq->reg_req->r_osd)
+			cancel_linger_request(lreq->reg_req);
+		ceph_osdc_put_request(lreq->reg_req);
+	}
 
-	request_reinit(req);
-	ceph_oid_copy(&req->r_base_oid, &lreq->t.base_oid);
-	ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
-	req->r_flags = lreq->t.flags;
+	req = ceph_osdc_alloc_request(osdc, NULL, 1, true, GFP_NOIO);
+	BUG_ON(!req);
+
+	target_copy(&req->r_t, &lreq->t);
 	req->r_mtime = lreq->mtime;
 
-	mutex_lock(&lreq->lock);
 	if (lreq->is_watch && lreq->committed) {
-		WARN_ON(op->op != CEPH_OSD_OP_WATCH ||
-			op->watch.cookie != lreq->linger_id);
-		op->watch.op = CEPH_OSD_WATCH_OP_RECONNECT;
-		op->watch.gen = ++lreq->register_gen;
+		osd_req_op_watch_init(req, 0, CEPH_OSD_WATCH_OP_RECONNECT,
+				      lreq->linger_id, ++lreq->register_gen);
 		dout("lreq %p reconnect register_gen %u\n", lreq,
-		     op->watch.gen);
+		     req->r_ops[0].watch.gen);
 		req->r_callback = linger_reconnect_cb;
 	} else {
-		if (!lreq->is_watch)
+		if (lreq->is_watch) {
+			osd_req_op_watch_init(req, 0, CEPH_OSD_WATCH_OP_WATCH,
+					      lreq->linger_id, 0);
+		} else {
 			lreq->notify_id = 0;
-		else
-			WARN_ON(op->watch.op != CEPH_OSD_WATCH_OP_WATCH);
+
+			refcount_inc(&lreq->request_pl->refcnt);
+			osd_req_op_notify_init(req, 0, lreq->linger_id,
+					       lreq->request_pl);
+			ceph_osd_data_pages_init(
+			    osd_req_op_data(req, 0, notify, response_data),
+			    lreq->notify_id_pages, PAGE_SIZE, 0, false, false);
+		}
 		dout("lreq %p register\n", lreq);
 		req->r_callback = linger_commit_cb;
 	}
-	mutex_unlock(&lreq->lock);
+
+	ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
+	BUG_ON(ret);
 
 	req->r_priv = linger_get(lreq);
 	req->r_linger = true;
+	lreq->reg_req = req;
+	mutex_unlock(&lreq->lock);
 
 	submit_request(req, true);
 }
@@ -3112,6 +3181,12 @@ static void linger_ping_cb(struct ceph_osd_request *req)
 	struct ceph_osd_linger_request *lreq = req->r_priv;
 
 	mutex_lock(&lreq->lock);
+	if (req != lreq->ping_req) {
+		dout("%s lreq %p linger_id %llu unknown req (%p != %p)\n",
+		     __func__, lreq, lreq->linger_id, req, lreq->ping_req);
+		goto out;
+	}
+
 	dout("%s lreq %p linger_id %llu result %d ping_sent %lu last_error %d\n",
 	     __func__, lreq, lreq->linger_id, req->r_result, lreq->ping_sent,
 	     lreq->last_error);
@@ -3127,6 +3202,7 @@ static void linger_ping_cb(struct ceph_osd_request *req)
 		     lreq->register_gen, req->r_ops[0].watch.gen);
 	}
 
+out:
 	mutex_unlock(&lreq->lock);
 	linger_put(lreq);
 }
@@ -3134,8 +3210,8 @@ static void linger_ping_cb(struct ceph_osd_request *req)
 static void send_linger_ping(struct ceph_osd_linger_request *lreq)
 {
 	struct ceph_osd_client *osdc = lreq->osdc;
-	struct ceph_osd_request *req = lreq->ping_req;
-	struct ceph_osd_req_op *op = &req->r_ops[0];
+	struct ceph_osd_request *req;
+	int ret;
 
 	if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD)) {
 		dout("%s PAUSERD\n", __func__);
@@ -3147,19 +3223,26 @@ static void send_linger_ping(struct ceph_osd_linger_request *lreq)
 	     __func__, lreq, lreq->linger_id, lreq->ping_sent,
 	     lreq->register_gen);
 
-	if (req->r_osd)
-		cancel_linger_request(req);
+	if (lreq->ping_req) {
+		if (lreq->ping_req->r_osd)
+			cancel_linger_request(lreq->ping_req);
+		ceph_osdc_put_request(lreq->ping_req);
+	}
 
-	request_reinit(req);
-	target_copy(&req->r_t, &lreq->t);
+	req = ceph_osdc_alloc_request(osdc, NULL, 1, true, GFP_NOIO);
+	BUG_ON(!req);
 
-	WARN_ON(op->op != CEPH_OSD_OP_WATCH ||
-		op->watch.cookie != lreq->linger_id ||
-		op->watch.op != CEPH_OSD_WATCH_OP_PING);
-	op->watch.gen = lreq->register_gen;
+	target_copy(&req->r_t, &lreq->t);
+	osd_req_op_watch_init(req, 0, CEPH_OSD_WATCH_OP_PING, lreq->linger_id,
+			      lreq->register_gen);
 	req->r_callback = linger_ping_cb;
+
+	ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
+	BUG_ON(ret);
+
 	req->r_priv = linger_get(lreq);
 	req->r_linger = true;
+	lreq->ping_req = req;
 
 	ceph_osdc_get_request(req);
 	account_request(req);
@@ -3175,12 +3258,6 @@ static void linger_submit(struct ceph_osd_linger_request *lreq)
 
 	down_write(&osdc->lock);
 	linger_register(lreq);
-	if (lreq->is_watch) {
-		lreq->reg_req->r_ops[0].watch.cookie = lreq->linger_id;
-		lreq->ping_req->r_ops[0].watch.cookie = lreq->linger_id;
-	} else {
-		lreq->reg_req->r_ops[0].notify.cookie = lreq->linger_id;
-	}
 
 	calc_target(osdc, &lreq->t, false);
 	osd = lookup_create_osd(osdc, lreq->t.osd, true);
@@ -3212,9 +3289,9 @@ static void cancel_linger_map_check(struct ceph_osd_linger_request *lreq)
  */
 static void __linger_cancel(struct ceph_osd_linger_request *lreq)
 {
-	if (lreq->is_watch && lreq->ping_req->r_osd)
+	if (lreq->ping_req && lreq->ping_req->r_osd)
 		cancel_linger_request(lreq->ping_req);
-	if (lreq->reg_req->r_osd)
+	if (lreq->reg_req && lreq->reg_req->r_osd)
 		cancel_linger_request(lreq->reg_req);
 	cancel_linger_map_check(lreq);
 	unlink_linger(lreq->osd, lreq);
@@ -3319,17 +3396,24 @@ static int linger_reg_commit_wait(struct ceph_osd_linger_request *lreq)
 	int ret;
 
 	dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
-	ret = wait_for_completion_interruptible(&lreq->reg_commit_wait);
+	ret = wait_for_completion_killable(&lreq->reg_commit_wait);
 	return ret ?: lreq->reg_commit_error;
 }
 
-static int linger_notify_finish_wait(struct ceph_osd_linger_request *lreq)
+static int linger_notify_finish_wait(struct ceph_osd_linger_request *lreq,
+				     unsigned long timeout)
 {
-	int ret;
+	long left;
 
 	dout("%s lreq %p linger_id %llu\n", __func__, lreq, lreq->linger_id);
-	ret = wait_for_completion_interruptible(&lreq->notify_finish_wait);
-	return ret ?: lreq->notify_finish_error;
+	left = wait_for_completion_killable_timeout(&lreq->notify_finish_wait,
+						ceph_timeout_jiffies(timeout));
+	if (left <= 0)
+		left = left ?: -ETIMEDOUT;
+	else
+		left = lreq->notify_finish_error; /* completed */
+
+	return left;
 }
 
 /*
@@ -3773,6 +3857,7 @@ static void handle_reply(struct ceph_osd *osd, struct ceph_msg *msg)
 	 * one (type of) reply back.
 	 */
 	WARN_ON(!(m.flags & CEPH_OSD_FLAG_ONDISK));
+	req->r_version = m.user_version;
 	req->r_result = m.result ?: data_len;
 	finish_request(req);
 	mutex_unlock(&osd->lock);
@@ -3864,7 +3949,7 @@ static void scan_requests(struct ceph_osd *osd,
 			if (!force_resend && !force_resend_writes)
 				break;
 
-			/* fall through */
+			fallthrough;
 		case CALC_TARGET_NEED_RESEND:
 			cancel_linger_map_check(lreq);
 			/*
@@ -3901,7 +3986,7 @@ static void scan_requests(struct ceph_osd *osd,
 			     !force_resend_writes))
 				break;
 
-			/* fall through */
+			fallthrough;
 		case CALC_TARGET_NEED_RESEND:
 			cancel_map_check(req);
 			unlink_request(osd, req);
@@ -3928,9 +4013,11 @@ static int handle_one_map(struct ceph_osd_client *osdc,
 	set_pool_was_full(osdc);
 
 	if (incremental)
-		newmap = osdmap_apply_incremental(&p, end, osdc->osdmap);
+		newmap = osdmap_apply_incremental(&p, end,
+						  ceph_msgr2(osdc->client),
+						  osdc->osdmap);
 	else
-		newmap = ceph_osdmap_decode(&p, end);
+		newmap = ceph_osdmap_decode(&p, end, ceph_msgr2(osdc->client));
 	if (IS_ERR(newmap))
 		return PTR_ERR(newmap);
 
@@ -4561,21 +4648,23 @@ bad:
 /*
  * Register request, send initial attempt.
  */
-int ceph_osdc_start_request(struct ceph_osd_client *osdc,
-			    struct ceph_osd_request *req,
-			    bool nofail)
+void ceph_osdc_start_request(struct ceph_osd_client *osdc,
+			     struct ceph_osd_request *req)
 {
 	down_read(&osdc->lock);
 	submit_request(req, false);
 	up_read(&osdc->lock);
-
-	return 0;
 }
 EXPORT_SYMBOL(ceph_osdc_start_request);
 
 /*
- * Unregister a registered request.  The request is not completed:
- * ->r_result isn't set and __complete_request() isn't called.
+ * Unregister request.  If @req was registered, it isn't completed:
+ * r_result isn't set and __complete_request() isn't invoked.
+ *
+ * If @req wasn't registered, this call may have raced with
+ * handle_reply(), in which case r_result would already be set and
+ * __complete_request() would be getting invoked, possibly even
+ * concurrently with this call.
  */
 void ceph_osdc_cancel_request(struct ceph_osd_request *req)
 {
@@ -4661,43 +4750,6 @@ again:
 }
 EXPORT_SYMBOL(ceph_osdc_sync);
 
-static struct ceph_osd_request *
-alloc_linger_request(struct ceph_osd_linger_request *lreq)
-{
-	struct ceph_osd_request *req;
-
-	req = ceph_osdc_alloc_request(lreq->osdc, NULL, 1, false, GFP_NOIO);
-	if (!req)
-		return NULL;
-
-	ceph_oid_copy(&req->r_base_oid, &lreq->t.base_oid);
-	ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
-	return req;
-}
-
-static struct ceph_osd_request *
-alloc_watch_request(struct ceph_osd_linger_request *lreq, u8 watch_opcode)
-{
-	struct ceph_osd_request *req;
-
-	req = alloc_linger_request(lreq);
-	if (!req)
-		return NULL;
-
-	/*
-	 * Pass 0 for cookie because we don't know it yet, it will be
-	 * filled in by linger_submit().
-	 */
-	osd_req_op_watch_init(req, 0, 0, watch_opcode);
-
-	if (ceph_osdc_alloc_messages(req, GFP_NOIO)) {
-		ceph_osdc_put_request(req);
-		return NULL;
-	}
-
-	return req;
-}
-
 /*
  * Returns a handle, caller owns a ref.
  */
@@ -4727,18 +4779,6 @@ ceph_osdc_watch(struct ceph_osd_client *osdc,
 	lreq->t.flags = CEPH_OSD_FLAG_WRITE;
 	ktime_get_real_ts64(&lreq->mtime);
 
-	lreq->reg_req = alloc_watch_request(lreq, CEPH_OSD_WATCH_OP_WATCH);
-	if (!lreq->reg_req) {
-		ret = -ENOMEM;
-		goto err_put_lreq;
-	}
-
-	lreq->ping_req = alloc_watch_request(lreq, CEPH_OSD_WATCH_OP_PING);
-	if (!lreq->ping_req) {
-		ret = -ENOMEM;
-		goto err_put_lreq;
-	}
-
 	linger_submit(lreq);
 	ret = linger_reg_commit_wait(lreq);
 	if (ret) {
@@ -4776,14 +4816,14 @@ int ceph_osdc_unwatch(struct ceph_osd_client *osdc,
 	ceph_oloc_copy(&req->r_base_oloc, &lreq->t.base_oloc);
 	req->r_flags = CEPH_OSD_FLAG_WRITE;
 	ktime_get_real_ts64(&req->r_mtime);
-	osd_req_op_watch_init(req, 0, lreq->linger_id,
-			      CEPH_OSD_WATCH_OP_UNWATCH);
+	osd_req_op_watch_init(req, 0, CEPH_OSD_WATCH_OP_UNWATCH,
+			      lreq->linger_id, 0);
 
 	ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
 	if (ret)
 		goto out_put_req;
 
-	ceph_osdc_start_request(osdc, req, false);
+	ceph_osdc_start_request(osdc, req);
 	linger_cancel(lreq);
 	linger_put(lreq);
 	ret = wait_request_timeout(req, opts->mount_timeout);
@@ -4802,7 +4842,7 @@ static int osd_req_op_notify_ack_init(struct ceph_osd_request *req, int which,
 	struct ceph_pagelist *pl;
 	int ret;
 
-	op = _osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY_ACK, 0);
+	op = osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY_ACK, 0);
 
 	pl = ceph_pagelist_alloc(GFP_NOIO);
 	if (!pl)
@@ -4854,7 +4894,7 @@ int ceph_osdc_notify_ack(struct ceph_osd_client *osdc,
 	if (ret)
 		goto out_put_req;
 
-	ceph_osdc_start_request(osdc, req, false);
+	ceph_osdc_start_request(osdc, req);
 	ret = ceph_osdc_wait_request(osdc, req);
 
 out_put_req:
@@ -4863,35 +4903,6 @@ out_put_req:
 }
 EXPORT_SYMBOL(ceph_osdc_notify_ack);
 
-static int osd_req_op_notify_init(struct ceph_osd_request *req, int which,
-				  u64 cookie, u32 prot_ver, u32 timeout,
-				  void *payload, u32 payload_len)
-{
-	struct ceph_osd_req_op *op;
-	struct ceph_pagelist *pl;
-	int ret;
-
-	op = _osd_req_op_init(req, which, CEPH_OSD_OP_NOTIFY, 0);
-	op->notify.cookie = cookie;
-
-	pl = ceph_pagelist_alloc(GFP_NOIO);
-	if (!pl)
-		return -ENOMEM;
-
-	ret = ceph_pagelist_encode_32(pl, 1); /* prot_ver */
-	ret |= ceph_pagelist_encode_32(pl, timeout);
-	ret |= ceph_pagelist_encode_32(pl, payload_len);
-	ret |= ceph_pagelist_append(pl, payload, payload_len);
-	if (ret) {
-		ceph_pagelist_release(pl);
-		return -ENOMEM;
-	}
-
-	ceph_osd_data_pagelist_init(&op->notify.request_data, pl);
-	op->indata_len = pl->length;
-	return 0;
-}
-
 /*
  * @timeout: in seconds
  *
@@ -4910,7 +4921,6 @@ int ceph_osdc_notify(struct ceph_osd_client *osdc,
 		     size_t *preply_len)
 {
 	struct ceph_osd_linger_request *lreq;
-	struct page **pages;
 	int ret;
 
 	WARN_ON(!timeout);
@@ -4923,46 +4933,41 @@ int ceph_osdc_notify(struct ceph_osd_client *osdc,
 	if (!lreq)
 		return -ENOMEM;
 
-	lreq->preply_pages = preply_pages;
-	lreq->preply_len = preply_len;
-
-	ceph_oid_copy(&lreq->t.base_oid, oid);
-	ceph_oloc_copy(&lreq->t.base_oloc, oloc);
-	lreq->t.flags = CEPH_OSD_FLAG_READ;
-
-	lreq->reg_req = alloc_linger_request(lreq);
-	if (!lreq->reg_req) {
+	lreq->request_pl = ceph_pagelist_alloc(GFP_NOIO);
+	if (!lreq->request_pl) {
 		ret = -ENOMEM;
 		goto out_put_lreq;
 	}
 
-	/*
-	 * Pass 0 for cookie because we don't know it yet, it will be
-	 * filled in by linger_submit().
-	 */
-	ret = osd_req_op_notify_init(lreq->reg_req, 0, 0, 1, timeout,
-				     payload, payload_len);
-	if (ret)
+	ret = ceph_pagelist_encode_32(lreq->request_pl, 1); /* prot_ver */
+	ret |= ceph_pagelist_encode_32(lreq->request_pl, timeout);
+	ret |= ceph_pagelist_encode_32(lreq->request_pl, payload_len);
+	ret |= ceph_pagelist_append(lreq->request_pl, payload, payload_len);
+	if (ret) {
+		ret = -ENOMEM;
 		goto out_put_lreq;
+	}
 
 	/* for notify_id */
-	pages = ceph_alloc_page_vector(1, GFP_NOIO);
-	if (IS_ERR(pages)) {
-		ret = PTR_ERR(pages);
+	lreq->notify_id_pages = ceph_alloc_page_vector(1, GFP_NOIO);
+	if (IS_ERR(lreq->notify_id_pages)) {
+		ret = PTR_ERR(lreq->notify_id_pages);
+		lreq->notify_id_pages = NULL;
 		goto out_put_lreq;
 	}
-	ceph_osd_data_pages_init(osd_req_op_data(lreq->reg_req, 0, notify,
-						 response_data),
-				 pages, PAGE_SIZE, 0, false, true);
 
-	ret = ceph_osdc_alloc_messages(lreq->reg_req, GFP_NOIO);
-	if (ret)
-		goto out_put_lreq;
+	lreq->preply_pages = preply_pages;
+	lreq->preply_len = preply_len;
+
+	ceph_oid_copy(&lreq->t.base_oid, oid);
+	ceph_oloc_copy(&lreq->t.base_oloc, oloc);
+	lreq->t.flags = CEPH_OSD_FLAG_READ;
 
 	linger_submit(lreq);
 	ret = linger_reg_commit_wait(lreq);
 	if (!ret)
-		ret = linger_notify_finish_wait(lreq);
+		ret = linger_notify_finish_wait(lreq,
+				 msecs_to_jiffies(2 * timeout * MSEC_PER_SEC));
 	else
 		dout("lreq %p failed to initiate notify %d\n", lreq, ret);
 
@@ -4973,40 +4978,6 @@ out_put_lreq:
 }
 EXPORT_SYMBOL(ceph_osdc_notify);
 
-/*
- * Return the number of milliseconds since the watch was last
- * confirmed, or an error.  If there is an error, the watch is no
- * longer valid, and should be destroyed with ceph_osdc_unwatch().
- */
-int ceph_osdc_watch_check(struct ceph_osd_client *osdc,
-			  struct ceph_osd_linger_request *lreq)
-{
-	unsigned long stamp, age;
-	int ret;
-
-	down_read(&osdc->lock);
-	mutex_lock(&lreq->lock);
-	stamp = lreq->watch_valid_thru;
-	if (!list_empty(&lreq->pending_lworks)) {
-		struct linger_work *lwork =
-		    list_first_entry(&lreq->pending_lworks,
-				     struct linger_work,
-				     pending_item);
-
-		if (time_before(lwork->queued_stamp, stamp))
-			stamp = lwork->queued_stamp;
-	}
-	age = jiffies - stamp;
-	dout("%s lreq %p linger_id %llu age %lu last_error %d\n", __func__,
-	     lreq, lreq->linger_id, age, lreq->last_error);
-	/* we are truncating to msecs, so return a safe upper bound */
-	ret = lreq->last_error ?: 1 + jiffies_to_msecs(age);
-
-	mutex_unlock(&lreq->lock);
-	up_read(&osdc->lock);
-	return ret;
-}
-
 static int decode_watcher(void **p, void *end, struct ceph_watch_item *item)
 {
 	u8 struct_v;
@@ -5106,7 +5077,7 @@ int ceph_osdc_list_watchers(struct ceph_osd_client *osdc,
 	if (ret)
 		goto out_put_req;
 
-	ceph_osdc_start_request(osdc, req, false);
+	ceph_osdc_start_request(osdc, req);
 	ret = ceph_osdc_wait_request(osdc, req);
 	if (ret >= 0) {
 		void *p = page_address(pages[0]);
@@ -5183,7 +5154,7 @@ int ceph_osdc_call(struct ceph_osd_client *osdc,
 	if (ret)
 		goto out_put_req;
 
-	ceph_osdc_start_request(osdc, req, false);
+	ceph_osdc_start_request(osdc, req);
 	ret = ceph_osdc_wait_request(osdc, req);
 	if (ret >= 0) {
 		ret = req->r_ops[0].rval;
@@ -5318,14 +5289,14 @@ void ceph_osdc_stop(struct ceph_osd_client *osdc)
 	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
 }
 
-static int osd_req_op_copy_from_init(struct ceph_osd_request *req,
-				     u64 src_snapid, u64 src_version,
-				     struct ceph_object_id *src_oid,
-				     struct ceph_object_locator *src_oloc,
-				     u32 src_fadvise_flags,
-				     u32 dst_fadvise_flags,
-				     u32 truncate_seq, u64 truncate_size,
-				     u8 copy_from_flags)
+int osd_req_op_copy_from_init(struct ceph_osd_request *req,
+			      u64 src_snapid, u64 src_version,
+			      struct ceph_object_id *src_oid,
+			      struct ceph_object_locator *src_oloc,
+			      u32 src_fadvise_flags,
+			      u32 dst_fadvise_flags,
+			      u32 truncate_seq, u64 truncate_size,
+			      u8 copy_from_flags)
 {
 	struct ceph_osd_req_op *op;
 	struct page **pages;
@@ -5335,8 +5306,8 @@ static int osd_req_op_copy_from_init(struct ceph_osd_request *req,
 	if (IS_ERR(pages))
 		return PTR_ERR(pages);
 
-	op = _osd_req_op_init(req, 0, CEPH_OSD_OP_COPY_FROM2,
-			      dst_fadvise_flags);
+	op = osd_req_op_init(req, 0, CEPH_OSD_OP_COPY_FROM2,
+			     dst_fadvise_flags);
 	op->copy_from.snapid = src_snapid;
 	op->copy_from.src_version = src_version;
 	op->copy_from.flags = copy_from_flags;
@@ -5354,49 +5325,7 @@ static int osd_req_op_copy_from_init(struct ceph_osd_request *req,
 				 op->indata_len, 0, false, true);
 	return 0;
 }
-
-int ceph_osdc_copy_from(struct ceph_osd_client *osdc,
-			u64 src_snapid, u64 src_version,
-			struct ceph_object_id *src_oid,
-			struct ceph_object_locator *src_oloc,
-			u32 src_fadvise_flags,
-			struct ceph_object_id *dst_oid,
-			struct ceph_object_locator *dst_oloc,
-			u32 dst_fadvise_flags,
-			u32 truncate_seq, u64 truncate_size,
-			u8 copy_from_flags)
-{
-	struct ceph_osd_request *req;
-	int ret;
-
-	req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_KERNEL);
-	if (!req)
-		return -ENOMEM;
-
-	req->r_flags = CEPH_OSD_FLAG_WRITE;
-
-	ceph_oloc_copy(&req->r_t.base_oloc, dst_oloc);
-	ceph_oid_copy(&req->r_t.base_oid, dst_oid);
-
-	ret = osd_req_op_copy_from_init(req, src_snapid, src_version, src_oid,
-					src_oloc, src_fadvise_flags,
-					dst_fadvise_flags, truncate_seq,
-					truncate_size, copy_from_flags);
-	if (ret)
-		goto out;
-
-	ret = ceph_osdc_alloc_messages(req, GFP_KERNEL);
-	if (ret)
-		goto out;
-
-	ceph_osdc_start_request(osdc, req, false);
-	ret = ceph_osdc_wait_request(osdc, req);
-
-out:
-	ceph_osdc_put_request(req);
-	return ret;
-}
-EXPORT_SYMBOL(ceph_osdc_copy_from);
+EXPORT_SYMBOL(osd_req_op_copy_from_init);
 
 int __init ceph_osdc_setup(void)
 {
@@ -5420,7 +5349,7 @@ void ceph_osdc_cleanup(void)
 /*
  * handle incoming message
  */
-static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
+static void osd_dispatch(struct ceph_connection *con, struct ceph_msg *msg)
 {
 	struct ceph_osd *osd = con->private;
 	struct ceph_osd_client *osdc = osd->o_osdc;
@@ -5448,6 +5377,24 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
 	ceph_msg_put(msg);
 }
 
+/* How much sparse data was requested? */
+static u64 sparse_data_requested(struct ceph_osd_request *req)
+{
+	u64 len = 0;
+
+	if (req->r_flags & CEPH_OSD_FLAG_READ) {
+		int i;
+
+		for (i = 0; i < req->r_num_ops; ++i) {
+			struct ceph_osd_req_op *op = &req->r_ops[i];
+
+			if (op->op == CEPH_OSD_OP_SPARSE_READ)
+				len += op->extent.length;
+		}
+	}
+	return len;
+}
+
 /*
  * Lookup and return message for incoming reply.  Don't try to do
  * anything about a larger than preallocated data portion of the
@@ -5464,6 +5411,7 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 	int front_len = le32_to_cpu(hdr->front_len);
 	int data_len = le32_to_cpu(hdr->data_len);
 	u64 tid = le64_to_cpu(hdr->tid);
+	u64 srlen;
 
 	down_read(&osdc->lock);
 	if (!osd_registered(osd)) {
@@ -5496,7 +5444,8 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 		req->r_reply = m;
 	}
 
-	if (data_len > req->r_reply->data_length) {
+	srlen = sparse_data_requested(req);
+	if (!srlen && data_len > req->r_reply->data_length) {
 		pr_warn("%s osd%d tid %llu data %d > preallocated %zu, skipping\n",
 			__func__, osd->o_osd, req->r_tid, data_len,
 			req->r_reply->data_length);
@@ -5506,6 +5455,8 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 	}
 
 	m = ceph_msg_get(req->r_reply);
+	m->sparse_read_total = srlen;
+
 	dout("get_reply tid %lld %p\n", tid, m);
 
 out_unlock_session:
@@ -5542,9 +5493,9 @@ static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 	return m;
 }
 
-static struct ceph_msg *alloc_msg(struct ceph_connection *con,
-				  struct ceph_msg_header *hdr,
-				  int *skip)
+static struct ceph_msg *osd_alloc_msg(struct ceph_connection *con,
+				      struct ceph_msg_header *hdr,
+				      int *skip)
 {
 	struct ceph_osd *osd = con->private;
 	int type = le16_to_cpu(hdr->type);
@@ -5568,7 +5519,7 @@ static struct ceph_msg *alloc_msg(struct ceph_connection *con,
 /*
  * Wrappers to refcount containing ceph_osd struct
  */
-static struct ceph_connection *get_osd_con(struct ceph_connection *con)
+static struct ceph_connection *osd_get_con(struct ceph_connection *con)
 {
 	struct ceph_osd *osd = con->private;
 	if (get_osd(osd))
@@ -5576,7 +5527,7 @@ static struct ceph_connection *get_osd_con(struct ceph_connection *con)
 	return NULL;
 }
 
-static void put_osd_con(struct ceph_connection *con)
+static void osd_put_con(struct ceph_connection *con)
 {
 	struct ceph_osd *osd = con->private;
 	put_osd(osd);
@@ -5585,39 +5536,29 @@ static void put_osd_con(struct ceph_connection *con)
 /*
  * authentication
  */
+
 /*
  * Note: returned pointer is the address of a structure that's
  * managed separately.  Caller must *not* attempt to free it.
  */
-static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
-					int *proto, int force_new)
+static struct ceph_auth_handshake *
+osd_get_authorizer(struct ceph_connection *con, int *proto, int force_new)
 {
 	struct ceph_osd *o = con->private;
 	struct ceph_osd_client *osdc = o->o_osdc;
 	struct ceph_auth_client *ac = osdc->client->monc.auth;
 	struct ceph_auth_handshake *auth = &o->o_auth;
+	int ret;
 
-	if (force_new && auth->authorizer) {
-		ceph_auth_destroy_authorizer(auth->authorizer);
-		auth->authorizer = NULL;
-	}
-	if (!auth->authorizer) {
-		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
-						      auth);
-		if (ret)
-			return ERR_PTR(ret);
-	} else {
-		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
-						     auth);
-		if (ret)
-			return ERR_PTR(ret);
-	}
-	*proto = ac->protocol;
+	ret = __ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_OSD,
+					 force_new, proto, NULL, NULL);
+	if (ret)
+		return ERR_PTR(ret);
 
 	return auth;
 }
 
-static int add_authorizer_challenge(struct ceph_connection *con,
+static int osd_add_authorizer_challenge(struct ceph_connection *con,
 				    void *challenge_buf, int challenge_buf_len)
 {
 	struct ceph_osd *o = con->private;
@@ -5628,16 +5569,19 @@ static int add_authorizer_challenge(struct ceph_connection *con,
 					    challenge_buf, challenge_buf_len);
 }
 
-static int verify_authorizer_reply(struct ceph_connection *con)
+static int osd_verify_authorizer_reply(struct ceph_connection *con)
 {
 	struct ceph_osd *o = con->private;
 	struct ceph_osd_client *osdc = o->o_osdc;
 	struct ceph_auth_client *ac = osdc->client->monc.auth;
+	struct ceph_auth_handshake *auth = &o->o_auth;
 
-	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer);
+	return ceph_auth_verify_authorizer_reply(ac, auth->authorizer,
+		auth->authorizer_reply_buf, auth->authorizer_reply_buf_len,
+		NULL, NULL, NULL, NULL);
 }
 
-static int invalidate_authorizer(struct ceph_connection *con)
+static int osd_invalidate_authorizer(struct ceph_connection *con)
 {
 	struct ceph_osd *o = con->private;
 	struct ceph_osd_client *osdc = o->o_osdc;
@@ -5647,6 +5591,80 @@ static int invalidate_authorizer(struct ceph_connection *con)
 	return ceph_monc_validate_auth(&osdc->client->monc);
 }
 
+static int osd_get_auth_request(struct ceph_connection *con,
+				void *buf, int *buf_len,
+				void **authorizer, int *authorizer_len)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_auth_client *ac = o->o_osdc->client->monc.auth;
+	struct ceph_auth_handshake *auth = &o->o_auth;
+	int ret;
+
+	ret = ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_OSD,
+				       buf, buf_len);
+	if (ret)
+		return ret;
+
+	*authorizer = auth->authorizer_buf;
+	*authorizer_len = auth->authorizer_buf_len;
+	return 0;
+}
+
+static int osd_handle_auth_reply_more(struct ceph_connection *con,
+				      void *reply, int reply_len,
+				      void *buf, int *buf_len,
+				      void **authorizer, int *authorizer_len)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_auth_client *ac = o->o_osdc->client->monc.auth;
+	struct ceph_auth_handshake *auth = &o->o_auth;
+	int ret;
+
+	ret = ceph_auth_handle_svc_reply_more(ac, auth, reply, reply_len,
+					      buf, buf_len);
+	if (ret)
+		return ret;
+
+	*authorizer = auth->authorizer_buf;
+	*authorizer_len = auth->authorizer_buf_len;
+	return 0;
+}
+
+static int osd_handle_auth_done(struct ceph_connection *con,
+				u64 global_id, void *reply, int reply_len,
+				u8 *session_key, int *session_key_len,
+				u8 *con_secret, int *con_secret_len)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_auth_client *ac = o->o_osdc->client->monc.auth;
+	struct ceph_auth_handshake *auth = &o->o_auth;
+
+	return ceph_auth_handle_svc_reply_done(ac, auth, reply, reply_len,
+					       session_key, session_key_len,
+					       con_secret, con_secret_len);
+}
+
+static int osd_handle_auth_bad_method(struct ceph_connection *con,
+				      int used_proto, int result,
+				      const int *allowed_protos, int proto_cnt,
+				      const int *allowed_modes, int mode_cnt)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_mon_client *monc = &o->o_osdc->client->monc;
+	int ret;
+
+	if (ceph_auth_handle_bad_authorizer(monc->auth, CEPH_ENTITY_TYPE_OSD,
+					    used_proto, result,
+					    allowed_protos, proto_cnt,
+					    allowed_modes, mode_cnt)) {
+		ret = ceph_monc_validate_auth(monc);
+		if (ret)
+			return ret;
+	}
+
+	return -EACCES;
+}
+
 static void osd_reencode_message(struct ceph_msg *msg)
 {
 	int type = le16_to_cpu(msg->hdr.type);
@@ -5671,17 +5689,229 @@ static int osd_check_message_signature(struct ceph_msg *msg)
 	return ceph_auth_check_message_signature(auth, msg);
 }
 
+static void advance_cursor(struct ceph_msg_data_cursor *cursor, size_t len,
+			   bool zero)
+{
+	while (len) {
+		struct page *page;
+		size_t poff, plen;
+
+		page = ceph_msg_data_next(cursor, &poff, &plen);
+		if (plen > len)
+			plen = len;
+		if (zero)
+			zero_user_segment(page, poff, poff + plen);
+		len -= plen;
+		ceph_msg_data_advance(cursor, plen);
+	}
+}
+
+static int prep_next_sparse_read(struct ceph_connection *con,
+				 struct ceph_msg_data_cursor *cursor)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_sparse_read *sr = &o->o_sparse_read;
+	struct ceph_osd_request *req;
+	struct ceph_osd_req_op *op;
+
+	spin_lock(&o->o_requests_lock);
+	req = lookup_request(&o->o_requests, le64_to_cpu(con->in_msg->hdr.tid));
+	if (!req) {
+		spin_unlock(&o->o_requests_lock);
+		return -EBADR;
+	}
+
+	if (o->o_sparse_op_idx < 0) {
+		dout("%s: [%d] starting new sparse read req\n",
+		     __func__, o->o_osd);
+	} else {
+		u64 end;
+
+		op = &req->r_ops[o->o_sparse_op_idx];
+
+		WARN_ON_ONCE(op->extent.sparse_ext);
+
+		/* hand back buffer we took earlier */
+		op->extent.sparse_ext = sr->sr_extent;
+		sr->sr_extent = NULL;
+		op->extent.sparse_ext_cnt = sr->sr_count;
+		sr->sr_ext_len = 0;
+		dout("%s: [%d] completed extent array len %d cursor->resid %zd\n",
+		     __func__, o->o_osd, op->extent.sparse_ext_cnt, cursor->resid);
+		/* Advance to end of data for this operation */
+		end = ceph_sparse_ext_map_end(op);
+		if (end < sr->sr_req_len)
+			advance_cursor(cursor, sr->sr_req_len - end, false);
+	}
+
+	ceph_init_sparse_read(sr);
+
+	/* find next op in this request (if any) */
+	while (++o->o_sparse_op_idx < req->r_num_ops) {
+		op = &req->r_ops[o->o_sparse_op_idx];
+		if (op->op == CEPH_OSD_OP_SPARSE_READ)
+			goto found;
+	}
+
+	/* reset for next sparse read request */
+	spin_unlock(&o->o_requests_lock);
+	o->o_sparse_op_idx = -1;
+	return 0;
+found:
+	sr->sr_req_off = op->extent.offset;
+	sr->sr_req_len = op->extent.length;
+	sr->sr_pos = sr->sr_req_off;
+	dout("%s: [%d] new sparse read op at idx %d 0x%llx~0x%llx\n", __func__,
+	     o->o_osd, o->o_sparse_op_idx, sr->sr_req_off, sr->sr_req_len);
+
+	/* hand off request's sparse extent map buffer */
+	sr->sr_ext_len = op->extent.sparse_ext_cnt;
+	op->extent.sparse_ext_cnt = 0;
+	sr->sr_extent = op->extent.sparse_ext;
+	op->extent.sparse_ext = NULL;
+
+	spin_unlock(&o->o_requests_lock);
+	return 1;
+}
+
+#ifdef __BIG_ENDIAN
+static inline void convert_extent_map(struct ceph_sparse_read *sr)
+{
+	int i;
+
+	for (i = 0; i < sr->sr_count; i++) {
+		struct ceph_sparse_extent *ext = &sr->sr_extent[i];
+
+		ext->off = le64_to_cpu((__force __le64)ext->off);
+		ext->len = le64_to_cpu((__force __le64)ext->len);
+	}
+}
+#else
+static inline void convert_extent_map(struct ceph_sparse_read *sr)
+{
+}
+#endif
+
+static int osd_sparse_read(struct ceph_connection *con,
+			   struct ceph_msg_data_cursor *cursor,
+			   char **pbuf)
+{
+	struct ceph_osd *o = con->private;
+	struct ceph_sparse_read *sr = &o->o_sparse_read;
+	u32 count = sr->sr_count;
+	u64 eoff, elen, len = 0;
+	int i, ret;
+
+	switch (sr->sr_state) {
+	case CEPH_SPARSE_READ_HDR:
+next_op:
+		ret = prep_next_sparse_read(con, cursor);
+		if (ret <= 0)
+			return ret;
+
+		/* number of extents */
+		ret = sizeof(sr->sr_count);
+		*pbuf = (char *)&sr->sr_count;
+		sr->sr_state = CEPH_SPARSE_READ_EXTENTS;
+		break;
+	case CEPH_SPARSE_READ_EXTENTS:
+		/* Convert sr_count to host-endian */
+		count = le32_to_cpu((__force __le32)sr->sr_count);
+		sr->sr_count = count;
+		dout("[%d] got %u extents\n", o->o_osd, count);
+
+		if (count > 0) {
+			if (!sr->sr_extent || count > sr->sr_ext_len) {
+				/* no extent array provided, or too short */
+				kfree(sr->sr_extent);
+				sr->sr_extent = kmalloc_array(count,
+							      sizeof(*sr->sr_extent),
+							      GFP_NOIO);
+				if (!sr->sr_extent) {
+					pr_err("%s: failed to allocate %u extents\n",
+					       __func__, count);
+					return -ENOMEM;
+				}
+				sr->sr_ext_len = count;
+			}
+			ret = count * sizeof(*sr->sr_extent);
+			*pbuf = (char *)sr->sr_extent;
+			sr->sr_state = CEPH_SPARSE_READ_DATA_LEN;
+			break;
+		}
+		/* No extents? Read data len */
+		fallthrough;
+	case CEPH_SPARSE_READ_DATA_LEN:
+		convert_extent_map(sr);
+		ret = sizeof(sr->sr_datalen);
+		*pbuf = (char *)&sr->sr_datalen;
+		sr->sr_state = CEPH_SPARSE_READ_DATA_PRE;
+		break;
+	case CEPH_SPARSE_READ_DATA_PRE:
+		/* Convert sr_datalen to host-endian */
+		sr->sr_datalen = le32_to_cpu((__force __le32)sr->sr_datalen);
+		for (i = 0; i < count; i++)
+			len += sr->sr_extent[i].len;
+		if (sr->sr_datalen != len) {
+			pr_warn_ratelimited("data len %u != extent len %llu\n",
+					    sr->sr_datalen, len);
+			return -EREMOTEIO;
+		}
+		sr->sr_state = CEPH_SPARSE_READ_DATA;
+		fallthrough;
+	case CEPH_SPARSE_READ_DATA:
+		if (sr->sr_index >= count) {
+			sr->sr_state = CEPH_SPARSE_READ_HDR;
+			goto next_op;
+		}
+
+		eoff = sr->sr_extent[sr->sr_index].off;
+		elen = sr->sr_extent[sr->sr_index].len;
+
+		dout("[%d] ext %d off 0x%llx len 0x%llx\n",
+		     o->o_osd, sr->sr_index, eoff, elen);
+
+		if (elen > INT_MAX) {
+			dout("Sparse read extent length too long (0x%llx)\n",
+			     elen);
+			return -EREMOTEIO;
+		}
+
+		/* zero out anything from sr_pos to start of extent */
+		if (sr->sr_pos < eoff)
+			advance_cursor(cursor, eoff - sr->sr_pos, true);
+
+		/* Set position to end of extent */
+		sr->sr_pos = eoff + elen;
+
+		/* send back the new length and nullify the ptr */
+		cursor->sr_resid = elen;
+		ret = elen;
+		*pbuf = NULL;
+
+		/* Bump the array index */
+		++sr->sr_index;
+		break;
+	}
+	return ret;
+}
+
 static const struct ceph_connection_operations osd_con_ops = {
-	.get = get_osd_con,
-	.put = put_osd_con,
-	.dispatch = dispatch,
-	.get_authorizer = get_authorizer,
-	.add_authorizer_challenge = add_authorizer_challenge,
-	.verify_authorizer_reply = verify_authorizer_reply,
-	.invalidate_authorizer = invalidate_authorizer,
-	.alloc_msg = alloc_msg,
+	.get = osd_get_con,
+	.put = osd_put_con,
+	.sparse_read = osd_sparse_read,
+	.alloc_msg = osd_alloc_msg,
+	.dispatch = osd_dispatch,
+	.fault = osd_fault,
 	.reencode_message = osd_reencode_message,
+	.get_authorizer = osd_get_authorizer,
+	.add_authorizer_challenge = osd_add_authorizer_challenge,
+	.verify_authorizer_reply = osd_verify_authorizer_reply,
+	.invalidate_authorizer = osd_invalidate_authorizer,
 	.sign_message = osd_sign_message,
 	.check_message_signature = osd_check_message_signature,
-	.fault = osd_fault,
+	.get_auth_request = osd_get_auth_request,
+	.handle_auth_reply_more = osd_handle_auth_reply_more,
+	.handle_auth_done = osd_handle_auth_done,
+	.handle_auth_bad_method = osd_handle_auth_bad_method,
 };