Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "A collection of fixes for the nvme over fabrics code"

* 'for-linus' of git://git.kernel.dk/linux-block:
  nvme-rdma: Get rid of redundant defines
  nvme-rdma: Get rid of duplicate variable
  nvme: fabrics drivers don't need the nvme-pci driver
  nvme-fabrics: get a reference when reusing a nvme_host structure
  nvme-fabrics: change NQN UUID to big-endian format
  nvme-loop: set sqsize to 0-based value, per spec
  nvme-rdma: fix sqsize/hsqsize per spec
  fabrics: define admin sqsize min default, per spec
  nvmet-rdma: +1 to *queue_size from hsqsize/hrqsize
  nvmet-rdma: Fix use after free
  nvme-rdma: initialize ret to zero to avoid returning garbage

8 files changed, 54 insertions, 34 deletions

diff --git a/drivers/nvme/host/Kconfig b/drivers/nvme/host/Kconfig
index db39d53cdfb9..0c644f7bdf80 100644
--- a/drivers/nvme/host/Kconfig
+++ b/drivers/nvme/host/Kconfig
@@ -31,7 +31,7 @@ config NVME_FABRICS
 config NVME_RDMA
 	tristate "NVM Express over Fabrics RDMA host driver"
 	depends on INFINIBAND
-	depends on BLK_DEV_NVME
+	select NVME_CORE
 	select NVME_FABRICS
 	select SG_POOL
 	help
diff --git a/drivers/nvme/host/fabrics.c b/drivers/nvme/host/fabrics.c
index dc996761042f..4eff49174466 100644
--- a/drivers/nvme/host/fabrics.c
+++ b/drivers/nvme/host/fabrics.c
@@ -47,8 +47,10 @@ static struct nvmf_host *nvmf_host_add(const char *hostnqn)
 
 	mutex_lock(&nvmf_hosts_mutex);
 	host = __nvmf_host_find(hostnqn);
-	if (host)
+	if (host) {
+		kref_get(&host->ref);
 		goto out_unlock;
+	}
 
 	host = kmalloc(sizeof(*host), GFP_KERNEL);
 	if (!host)
@@ -56,7 +58,7 @@ static struct nvmf_host *nvmf_host_add(const char *hostnqn)
 
 	kref_init(&host->ref);
 	memcpy(host->nqn, hostnqn, NVMF_NQN_SIZE);
-	uuid_le_gen(&host->id);
+	uuid_be_gen(&host->id);
 
 	list_add_tail(&host->list, &nvmf_hosts);
 out_unlock:
@@ -73,9 +75,9 @@ static struct nvmf_host *nvmf_host_default(void)
 		return NULL;
 
 	kref_init(&host->ref);
-	uuid_le_gen(&host->id);
+	uuid_be_gen(&host->id);
 	snprintf(host->nqn, NVMF_NQN_SIZE,
-		"nqn.2014-08.org.nvmexpress:NVMf:uuid:%pUl", &host->id);
+		"nqn.2014-08.org.nvmexpress:NVMf:uuid:%pUb", &host->id);
 
 	mutex_lock(&nvmf_hosts_mutex);
 	list_add_tail(&host->list, &nvmf_hosts);
@@ -363,7 +365,14 @@ int nvmf_connect_admin_queue(struct nvme_ctrl *ctrl)
 	cmd.connect.opcode = nvme_fabrics_command;
 	cmd.connect.fctype = nvme_fabrics_type_connect;
 	cmd.connect.qid = 0;
-	cmd.connect.sqsize = cpu_to_le16(ctrl->sqsize);
+
+	/*
+	 * fabrics spec sets a minimum of depth 32 for admin queue,
+	 * so set the queue with this depth always until
+	 * justification otherwise.
+	 */
+	cmd.connect.sqsize = cpu_to_le16(NVMF_AQ_DEPTH - 1);
+
 	/*
 	 * Set keep-alive timeout in seconds granularity (ms * 1000)
 	 * and add a grace period for controller kato enforcement
@@ -375,7 +384,7 @@ int nvmf_connect_admin_queue(struct nvme_ctrl *ctrl)
 	if (!data)
 		return -ENOMEM;
 
-	memcpy(&data->hostid, &ctrl->opts->host->id, sizeof(uuid_le));
+	memcpy(&data->hostid, &ctrl->opts->host->id, sizeof(uuid_be));
 	data->cntlid = cpu_to_le16(0xffff);
 	strncpy(data->subsysnqn, ctrl->opts->subsysnqn, NVMF_NQN_SIZE);
 	strncpy(data->hostnqn, ctrl->opts->host->nqn, NVMF_NQN_SIZE);
@@ -434,7 +443,7 @@ int nvmf_connect_io_queue(struct nvme_ctrl *ctrl, u16 qid)
 	if (!data)
 		return -ENOMEM;
 
-	memcpy(&data->hostid, &ctrl->opts->host->id, sizeof(uuid_le));
+	memcpy(&data->hostid, &ctrl->opts->host->id, sizeof(uuid_be));
 	data->cntlid = cpu_to_le16(ctrl->cntlid);
 	strncpy(data->subsysnqn, ctrl->opts->subsysnqn, NVMF_NQN_SIZE);
 	strncpy(data->hostnqn, ctrl->opts->host->nqn, NVMF_NQN_SIZE);
diff --git a/drivers/nvme/host/fabrics.h b/drivers/nvme/host/fabrics.h
index 89df52c8be97..46e460aee52d 100644
--- a/drivers/nvme/host/fabrics.h
+++ b/drivers/nvme/host/fabrics.h
@@ -34,7 +34,7 @@ struct nvmf_host {
 	struct kref		ref;
 	struct list_head	list;
 	char			nqn[NVMF_NQN_SIZE];
-	uuid_le			id;
+	uuid_be			id;
 };
 
 /**
diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index 8d2875b4c56d..ab545fb347a0 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -43,10 +43,6 @@
 
 #define NVME_RDMA_MAX_INLINE_SEGMENTS	1
 
-#define NVME_RDMA_MAX_PAGES_PER_MR	512
-
-#define NVME_RDMA_DEF_RECONNECT_DELAY	20
-
 /*
  * We handle AEN commands ourselves and don't even let the
  * block layer know about them.
@@ -77,7 +73,6 @@ struct nvme_rdma_request {
 	u32			num_sge;
 	int			nents;
 	bool			inline_data;
-	bool			need_inval;
 	struct ib_reg_wr	reg_wr;
 	struct ib_cqe		reg_cqe;
 	struct nvme_rdma_queue  *queue;
@@ -286,7 +281,7 @@ static int nvme_rdma_reinit_request(void *data, struct request *rq)
 	struct nvme_rdma_request *req = blk_mq_rq_to_pdu(rq);
 	int ret = 0;
 
-	if (!req->need_inval)
+	if (!req->mr->need_inval)
 		goto out;
 
 	ib_dereg_mr(req->mr);
@@ -298,7 +293,7 @@ static int nvme_rdma_reinit_request(void *data, struct request *rq)
 		req->mr = NULL;
 	}
 
-	req->need_inval = false;
+	req->mr->need_inval = false;
 
 out:
 	return ret;
@@ -645,7 +640,8 @@ static int nvme_rdma_init_io_queues(struct nvme_rdma_ctrl *ctrl)
 	int i, ret;
 
 	for (i = 1; i < ctrl->queue_count; i++) {
-		ret = nvme_rdma_init_queue(ctrl, i, ctrl->ctrl.sqsize);
+		ret = nvme_rdma_init_queue(ctrl, i,
+					   ctrl->ctrl.opts->queue_size);
 		if (ret) {
 			dev_info(ctrl->ctrl.device,
 				"failed to initialize i/o queue: %d\n", ret);
@@ -849,7 +845,7 @@ static void nvme_rdma_unmap_data(struct nvme_rdma_queue *queue,
 	if (!blk_rq_bytes(rq))
 		return;
 
-	if (req->need_inval) {
+	if (req->mr->need_inval) {
 		res = nvme_rdma_inv_rkey(queue, req);
 		if (res < 0) {
 			dev_err(ctrl->ctrl.device,
@@ -935,7 +931,7 @@ static int nvme_rdma_map_sg_fr(struct nvme_rdma_queue *queue,
 			     IB_ACCESS_REMOTE_READ |
 			     IB_ACCESS_REMOTE_WRITE;
 
-	req->need_inval = true;
+	req->mr->need_inval = true;
 
 	sg->addr = cpu_to_le64(req->mr->iova);
 	put_unaligned_le24(req->mr->length, sg->length);
@@ -958,7 +954,7 @@ static int nvme_rdma_map_data(struct nvme_rdma_queue *queue,
 
 	req->num_sge = 1;
 	req->inline_data = false;
-	req->need_inval = false;
+	req->mr->need_inval = false;
 
 	c->common.flags |= NVME_CMD_SGL_METABUF;
 
@@ -1145,7 +1141,7 @@ static int nvme_rdma_process_nvme_rsp(struct nvme_rdma_queue *queue,
 
 	if ((wc->wc_flags & IB_WC_WITH_INVALIDATE) &&
 	    wc->ex.invalidate_rkey == req->mr->rkey)
-		req->need_inval = false;
+		req->mr->need_inval = false;
 
 	blk_mq_complete_request(rq, status);
 
@@ -1278,8 +1274,22 @@ static int nvme_rdma_route_resolved(struct nvme_rdma_queue *queue)
 
 	priv.recfmt = cpu_to_le16(NVME_RDMA_CM_FMT_1_0);
 	priv.qid = cpu_to_le16(nvme_rdma_queue_idx(queue));
-	priv.hrqsize = cpu_to_le16(queue->queue_size);
-	priv.hsqsize = cpu_to_le16(queue->queue_size);
+	/*
+	 * set the admin queue depth to the minimum size
+	 * specified by the Fabrics standard.
+	 */
+	if (priv.qid == 0) {
+		priv.hrqsize = cpu_to_le16(NVMF_AQ_DEPTH);
+		priv.hsqsize = cpu_to_le16(NVMF_AQ_DEPTH - 1);
+	} else {
+		/*
+		 * current interpretation of the fabrics spec
+		 * is at minimum you make hrqsize sqsize+1, or a
+		 * 1's based representation of sqsize.
+		 */
+		priv.hrqsize = cpu_to_le16(queue->queue_size);
+		priv.hsqsize = cpu_to_le16(queue->ctrl->ctrl.sqsize);
+	}
 
 	ret = rdma_connect(queue->cm_id, &param);
 	if (ret) {
@@ -1319,7 +1329,7 @@ out_destroy_queue_ib:
 static int nvme_rdma_device_unplug(struct nvme_rdma_queue *queue)
 {
 	struct nvme_rdma_ctrl *ctrl = queue->ctrl;
-	int ret;
+	int ret = 0;
 
 	/* Own the controller deletion */
 	if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_DELETING))
@@ -1461,7 +1471,7 @@ static int nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
 	if (rq->cmd_type == REQ_TYPE_FS && req_op(rq) == REQ_OP_FLUSH)
 		flush = true;
 	ret = nvme_rdma_post_send(queue, sqe, req->sge, req->num_sge,
-			req->need_inval ? &req->reg_wr.wr : NULL, flush);
+			req->mr->need_inval ? &req->reg_wr.wr : NULL, flush);
 	if (ret) {
 		nvme_rdma_unmap_data(queue, rq);
 		goto err;
@@ -1816,7 +1826,7 @@ static int nvme_rdma_create_io_queues(struct nvme_rdma_ctrl *ctrl)
 
 	memset(&ctrl->tag_set, 0, sizeof(ctrl->tag_set));
 	ctrl->tag_set.ops = &nvme_rdma_mq_ops;
-	ctrl->tag_set.queue_depth = ctrl->ctrl.sqsize;
+	ctrl->tag_set.queue_depth = ctrl->ctrl.opts->queue_size;
 	ctrl->tag_set.reserved_tags = 1; /* fabric connect */
 	ctrl->tag_set.numa_node = NUMA_NO_NODE;
 	ctrl->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
@@ -1914,7 +1924,7 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
 	spin_lock_init(&ctrl->lock);
 
 	ctrl->queue_count = opts->nr_io_queues + 1; /* +1 for admin queue */
-	ctrl->ctrl.sqsize = opts->queue_size;
+	ctrl->ctrl.sqsize = opts->queue_size - 1;
 	ctrl->ctrl.kato = opts->kato;
 
 	ret = -ENOMEM;
diff --git a/drivers/nvme/target/Kconfig b/drivers/nvme/target/Kconfig
index a5c31cbeb481..3a5b9d0576cb 100644
--- a/drivers/nvme/target/Kconfig
+++ b/drivers/nvme/target/Kconfig
@@ -15,8 +15,8 @@ config NVME_TARGET
 
 config NVME_TARGET_LOOP
 	tristate "NVMe loopback device support"
-	depends on BLK_DEV_NVME
 	depends on NVME_TARGET
+	select NVME_CORE
 	select NVME_FABRICS
 	select SG_POOL
 	help
diff --git a/drivers/nvme/target/loop.c b/drivers/nvme/target/loop.c
index 7affd40a6b33..395e60dad835 100644
--- a/drivers/nvme/target/loop.c
+++ b/drivers/nvme/target/loop.c
@@ -556,7 +556,7 @@ static int nvme_loop_create_io_queues(struct nvme_loop_ctrl *ctrl)
 
 	memset(&ctrl->tag_set, 0, sizeof(ctrl->tag_set));
 	ctrl->tag_set.ops = &nvme_loop_mq_ops;
-	ctrl->tag_set.queue_depth = ctrl->ctrl.sqsize;
+	ctrl->tag_set.queue_depth = ctrl->ctrl.opts->queue_size;
 	ctrl->tag_set.reserved_tags = 1; /* fabric connect */
 	ctrl->tag_set.numa_node = NUMA_NO_NODE;
 	ctrl->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
@@ -620,7 +620,7 @@ static struct nvme_ctrl *nvme_loop_create_ctrl(struct device *dev,
 
 	ret = -ENOMEM;
 
-	ctrl->ctrl.sqsize = opts->queue_size;
+	ctrl->ctrl.sqsize = opts->queue_size - 1;
 	ctrl->ctrl.kato = opts->kato;
 
 	ctrl->queues = kcalloc(opts->nr_io_queues + 1, sizeof(*ctrl->queues),
diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c
index b4d648536c3e..1cbe6e053b5b 100644
--- a/drivers/nvme/target/rdma.c
+++ b/drivers/nvme/target/rdma.c
@@ -978,10 +978,11 @@ static void nvmet_rdma_release_queue_work(struct work_struct *w)
 		container_of(w, struct nvmet_rdma_queue, release_work);
 	struct rdma_cm_id *cm_id = queue->cm_id;
 	struct nvmet_rdma_device *dev = queue->dev;
+	enum nvmet_rdma_queue_state state = queue->state;
 
 	nvmet_rdma_free_queue(queue);
 
-	if (queue->state != NVMET_RDMA_IN_DEVICE_REMOVAL)
+	if (state != NVMET_RDMA_IN_DEVICE_REMOVAL)
 		rdma_destroy_id(cm_id);
 
 	kref_put(&dev->ref, nvmet_rdma_free_dev);
@@ -1003,10 +1004,10 @@ nvmet_rdma_parse_cm_connect_req(struct rdma_conn_param *conn,
 	queue->host_qid = le16_to_cpu(req->qid);
 
 	/*
-	 * req->hsqsize corresponds to our recv queue size
+	 * req->hsqsize corresponds to our recv queue size plus 1
 	 * req->hrqsize corresponds to our send queue size
 	 */
-	queue->recv_queue_size = le16_to_cpu(req->hsqsize);
+	queue->recv_queue_size = le16_to_cpu(req->hsqsize) + 1;
 	queue->send_queue_size = le16_to_cpu(req->hrqsize);
 
 	if (!queue->host_qid && queue->recv_queue_size > NVMF_AQ_DEPTH)
diff --git a/include/linux/nvme.h b/include/linux/nvme.h
index d8b37bab2887..7676557ce357 100644
--- a/include/linux/nvme.h
+++ b/include/linux/nvme.h
@@ -794,7 +794,7 @@ struct nvmf_connect_command {
 };
 
 struct nvmf_connect_data {
-	uuid_le		hostid;
+	uuid_be		hostid;
 	__le16		cntlid;
 	char		resv4[238];
 	char		subsysnqn[NVMF_NQN_FIELD_LEN];