Merge tag 'nvme-5.12-2021-02-11' of git://git.infradead.org/nvme into for-5.12/drivers

Pull NVMe updates from Christoph:

"nvme updates for 5.12:

 - fix multipath handling of ->queue_rq errors (Chao Leng)
 - nvmet cleanups (Chaitanya Kulkarni)
 - add a quirk for buggy Amazon controller (Filippo Sironi)
 - avoid devm allocations in nvme-hwmon that don't interact well with
   fabrics (Hannes Reinecke)
 - sysfs cleanups (Jiapeng Chong)
 - fix nr_zones for multipath (Keith Busch)
 - nvme-tcp crash fix for no-data commands (Sagi Grimberg)
 - nvmet-tcp fixes (Sagi Grimberg)
 - add a missing __rcu annotation (me)"

* tag 'nvme-5.12-2021-02-11' of git://git.infradead.org/nvme: (22 commits)
  nvme-tcp: fix crash triggered with a dataless request submission
  nvme: add 48-bit DMA address quirk for Amazon NVMe controllers
  nvme-hwmon: rework to avoid devm allocation
  nvmet: remove else at the end of the function
  nvmet: add nvmet_req_subsys() helper
  nvmet: use min of device_path and disk len
  nvmet: use invalid cmd opcode helper
  nvmet: use invalid cmd opcode helper
  nvmet: add helper to report invalid opcode
  nvmet: remove extra variable in id-ns handler
  nvmet: make nvmet_find_namespace() req based
  nvmet: return uniform error for invalid ns
  nvmet: set status to 0 in case for invalid nsid
  nvmet-fc: add a missing __rcu annotation to nvmet_fc_tgt_assoc.queues
  nvme-multipath: set nr_zones for zoned namespaces
  nvmet-tcp: fix potential race of tcp socket closing accept_work
  nvmet-tcp: fix receive data digest calculation for multiple h2cdata PDUs
  nvme-rdma: handle nvme_rdma_post_send failures better
  nvme-fabrics: avoid double completions in nvmf_fail_nonready_command
  nvme: introduce a nvme_host_path_error helper
  ...

drivers/nvme/host/core.c

@@ -355,6 +355,21 @@ void nvme_complete_rq(struct request *req)
 }
 EXPORT_SYMBOL_GPL(nvme_complete_rq);
 
+/*
+ * Called to unwind from ->queue_rq on a failed command submission so that the
+ * multipathing code gets called to potentially failover to another path.
+ * The caller needs to unwind all transport specific resource allocations and
+ * must return propagate the return value.
+ */
+blk_status_t nvme_host_path_error(struct request *req)
+{
+	nvme_req(req)->status = NVME_SC_HOST_PATH_ERROR;
+	blk_mq_set_request_complete(req);
+	nvme_complete_rq(req);
+	return BLK_STS_OK;
+}
+EXPORT_SYMBOL_GPL(nvme_host_path_error);
+
 bool nvme_cancel_request(struct request *req, void *data, bool reserved)
 {
 	dev_dbg_ratelimited(((struct nvme_ctrl *) data)->device,
@@ -2848,7 +2863,7 @@ static ssize_t nvme_subsys_show_nqn(struct device *dev,
 	struct nvme_subsystem *subsys =
 		container_of(dev, struct nvme_subsystem, dev);
 
-	return snprintf(buf, PAGE_SIZE, "%s\n", subsys->subnqn);
+	return sysfs_emit(buf, "%s\n", subsys->subnqn);
 }
 static SUBSYS_ATTR_RO(subsysnqn, S_IRUGO, nvme_subsys_show_nqn);
 
@@ -3541,7 +3556,7 @@ static ssize_t nvme_sysfs_show_transport(struct device *dev,
 {
 	struct nvme_ctrl *ctrl = dev_get_drvdata(dev);
 
-	return snprintf(buf, PAGE_SIZE, "%s\n", ctrl->ops->name);
+	return sysfs_emit(buf, "%s\n", ctrl->ops->name);
 }
 static DEVICE_ATTR(transport, S_IRUGO, nvme_sysfs_show_transport, NULL);
 
@@ -3575,7 +3590,7 @@ static ssize_t nvme_sysfs_show_subsysnqn(struct device *dev,
 {
 	struct nvme_ctrl *ctrl = dev_get_drvdata(dev);
 
-	return snprintf(buf, PAGE_SIZE, "%s\n", ctrl->subsys->subnqn);
+	return sysfs_emit(buf, "%s\n", ctrl->subsys->subnqn);
 }
 static DEVICE_ATTR(subsysnqn, S_IRUGO, nvme_sysfs_show_subsysnqn, NULL);
 
@@ -3585,7 +3600,7 @@ static ssize_t nvme_sysfs_show_hostnqn(struct device *dev,
 {
 	struct nvme_ctrl *ctrl = dev_get_drvdata(dev);
 
-	return snprintf(buf, PAGE_SIZE, "%s\n", ctrl->opts->host->nqn);
+	return sysfs_emit(buf, "%s\n", ctrl->opts->host->nqn);
 }
 static DEVICE_ATTR(hostnqn, S_IRUGO, nvme_sysfs_show_hostnqn, NULL);
 
@@ -3595,7 +3610,7 @@ static ssize_t nvme_sysfs_show_hostid(struct device *dev,
 {
 	struct nvme_ctrl *ctrl = dev_get_drvdata(dev);
 
-	return snprintf(buf, PAGE_SIZE, "%pU\n", &ctrl->opts->host->id);
+	return sysfs_emit(buf, "%pU\n", &ctrl->opts->host->id);
 }
 static DEVICE_ATTR(hostid, S_IRUGO, nvme_sysfs_show_hostid, NULL);
 
@@ -4456,6 +4471,7 @@ EXPORT_SYMBOL_GPL(nvme_start_ctrl);
 
 void nvme_uninit_ctrl(struct nvme_ctrl *ctrl)
 {
+	nvme_hwmon_exit(ctrl);
 	nvme_fault_inject_fini(&ctrl->fault_inject);
 	dev_pm_qos_hide_latency_tolerance(ctrl->device);
 	cdev_device_del(&ctrl->cdev, ctrl->device);

drivers/nvme/host/fabrics.c

@@ -552,11 +552,7 @@ blk_status_t nvmf_fail_nonready_command(struct nvme_ctrl *ctrl,
 	    !test_bit(NVME_CTRL_FAILFAST_EXPIRED, &ctrl->flags) &&
 	    !blk_noretry_request(rq) && !(rq->cmd_flags & REQ_NVME_MPATH))
 		return BLK_STS_RESOURCE;
-
-	nvme_req(rq)->status = NVME_SC_HOST_PATH_ERROR;
-	blk_mq_start_request(rq);
-	nvme_complete_rq(rq);
-	return BLK_STS_OK;
+	return nvme_host_path_error(rq);
 }
 EXPORT_SYMBOL_GPL(nvmf_fail_nonready_command);
 

drivers/nvme/host/hwmon.c

@@ -223,12 +223,12 @@ static const struct hwmon_chip_info nvme_hwmon_chip_info = {
 
 int nvme_hwmon_init(struct nvme_ctrl *ctrl)
 {
-	struct device *dev = ctrl->dev;
+	struct device *dev = ctrl->device;
 	struct nvme_hwmon_data *data;
 	struct device *hwmon;
 	int err;
 
-	data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
 	if (!data)
 		return 0;
 
@@ -237,19 +237,30 @@ int nvme_hwmon_init(struct nvme_ctrl *ctrl)
 
 	err = nvme_hwmon_get_smart_log(data);
 	if (err) {
-		dev_warn(ctrl->device,
-			"Failed to read smart log (error %d)\n", err);
-		devm_kfree(dev, data);
+		dev_warn(dev, "Failed to read smart log (error %d)\n", err);
+		kfree(data);
 		return err;
 	}
 
-	hwmon = devm_hwmon_device_register_with_info(dev, "nvme", data,
-						     &nvme_hwmon_chip_info,
-						     NULL);
+	hwmon = hwmon_device_register_with_info(dev, "nvme",
+						data, &nvme_hwmon_chip_info,
+						NULL);
 	if (IS_ERR(hwmon)) {
 		dev_warn(dev, "Failed to instantiate hwmon device\n");
-		devm_kfree(dev, data);
+		kfree(data);
 	}
-
+	ctrl->hwmon_device = hwmon;
 	return 0;
 }
+
+void nvme_hwmon_exit(struct nvme_ctrl *ctrl)
+{
+	if (ctrl->hwmon_device) {
+		struct nvme_hwmon_data *data =
+			dev_get_drvdata(ctrl->hwmon_device);
+
+		hwmon_device_unregister(ctrl->hwmon_device);
+		ctrl->hwmon_device = NULL;
+		kfree(data);
+	}
+}

drivers/nvme/host/multipath.c

@@ -677,6 +677,10 @@ void nvme_mpath_add_disk(struct nvme_ns *ns, struct nvme_id_ns *id)
 	if (blk_queue_stable_writes(ns->queue) && ns->head->disk)
 		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES,
 				   ns->head->disk->queue);
+#ifdef CONFIG_BLK_DEV_ZONED
+	if (blk_queue_is_zoned(ns->queue) && ns->head->disk)
+		ns->head->disk->queue->nr_zones = ns->queue->nr_zones;
+#endif
 }
 
 void nvme_mpath_remove_disk(struct nvme_ns_head *head)

drivers/nvme/host/nvme.h

@@ -144,6 +144,12 @@ enum nvme_quirks {
 	 * NVMe 1.3 compliance.
 	 */
 	NVME_QUIRK_NO_NS_DESC_LIST		= (1 << 15),
+
+	/*
+	 * The controller does not properly handle DMA addresses over
+	 * 48 bits.
+	 */
+	NVME_QUIRK_DMA_ADDRESS_BITS_48		= (1 << 16),
 };
 
 /*
@@ -246,6 +252,9 @@ struct nvme_ctrl {
 	struct rw_semaphore namespaces_rwsem;
 	struct device ctrl_device;
 	struct device *device;	/* char device */
+#ifdef CONFIG_NVME_HWMON
+	struct device *hwmon_device;
+#endif
 	struct cdev cdev;
 	struct work_struct reset_work;
 	struct work_struct delete_work;
@@ -575,6 +584,7 @@ static inline bool nvme_is_aen_req(u16 qid, __u16 command_id)
 }
 
 void nvme_complete_rq(struct request *req);
+blk_status_t nvme_host_path_error(struct request *req);
 bool nvme_cancel_request(struct request *req, void *data, bool reserved);
 void nvme_cancel_tagset(struct nvme_ctrl *ctrl);
 void nvme_cancel_admin_tagset(struct nvme_ctrl *ctrl);
@@ -811,11 +821,16 @@ static inline struct nvme_ns *nvme_get_ns_from_dev(struct device *dev)
 
 #ifdef CONFIG_NVME_HWMON
 int nvme_hwmon_init(struct nvme_ctrl *ctrl);
+void nvme_hwmon_exit(struct nvme_ctrl *ctrl);
 #else
 static inline int nvme_hwmon_init(struct nvme_ctrl *ctrl)
 {
 	return 0;
 }
+
+static inline void nvme_hwmon_exit(struct nvme_ctrl *ctrl)
+{
+}
 #endif
 
 u32 nvme_command_effects(struct nvme_ctrl *ctrl, struct nvme_ns *ns,

drivers/nvme/host/pci.c

@@ -2362,13 +2362,16 @@ static int nvme_pci_enable(struct nvme_dev *dev)
 {
 	int result = -ENOMEM;
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
+	int dma_address_bits = 64;
 
 	if (pci_enable_device_mem(pdev))
 		return result;
 
 	pci_set_master(pdev);
 
-	if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(64)))
+	if (dev->ctrl.quirks & NVME_QUIRK_DMA_ADDRESS_BITS_48)
+		dma_address_bits = 48;
+	if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(dma_address_bits)))
 		goto disable;
 
 	if (readl(dev->bar + NVME_REG_CSTS) == -1) {
@@ -3257,6 +3260,22 @@ static const struct pci_device_id nvme_id_table[] = {
 		.driver_data = NVME_QUIRK_DISABLE_WRITE_ZEROES, },
 	{ PCI_DEVICE(0x15b7, 0x2001),   /*  Sandisk Skyhawk */
 		.driver_data = NVME_QUIRK_DISABLE_WRITE_ZEROES, },
+	{ PCI_DEVICE(0x1d97, 0x2263),   /* SPCC */
+		.driver_data = NVME_QUIRK_DISABLE_WRITE_ZEROES, },
+	{ PCI_DEVICE(0x2646, 0x2263),   /* KINGSTON A2000 NVMe SSD  */
+		.driver_data = NVME_QUIRK_NO_DEEPEST_PS, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0x0061),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0x0065),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0x8061),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0xcd00),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0xcd01),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_AMAZON, 0xcd02),
+		.driver_data = NVME_QUIRK_DMA_ADDRESS_BITS_48, },
 	{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2001),
 		.driver_data = NVME_QUIRK_SINGLE_VECTOR },
 	{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2003) },

drivers/nvme/host/rdma.c

@@ -2098,7 +2098,9 @@ static blk_status_t nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
 err_unmap:
 	nvme_rdma_unmap_data(queue, rq);
 err:
-	if (err == -ENOMEM || err == -EAGAIN)
+	if (err == -EIO)
+		ret = nvme_host_path_error(rq);
+	else if (err == -ENOMEM || err == -EAGAIN)
 		ret = BLK_STS_RESOURCE;
 	else
 		ret = BLK_STS_IOERR;

drivers/nvme/host/tcp.c

@@ -2271,7 +2271,7 @@ static blk_status_t nvme_tcp_setup_cmd_pdu(struct nvme_ns *ns,
 	req->data_len = blk_rq_nr_phys_segments(rq) ?
 				blk_rq_payload_bytes(rq) : 0;
 	req->curr_bio = rq->bio;
-	if (req->curr_bio)
+	if (req->curr_bio && req->data_len)
 		nvme_tcp_init_iter(req, rq_data_dir(rq));
 
 	if (rq_data_dir(rq) == WRITE &&

drivers/nvme/target/admin-cmd.c

@@ -75,15 +75,11 @@ static u16 nvmet_get_smart_log_nsid(struct nvmet_req *req,
 		struct nvme_smart_log *slog)
 {
 	u64 host_reads, host_writes, data_units_read, data_units_written;
+	u16 status;
 
-	req->ns = nvmet_find_namespace(req->sq->ctrl,
-				       req->cmd->get_log_page.nsid);
-	if (!req->ns) {
-		pr_err("Could not find namespace id : %d\n",
-				le32_to_cpu(req->cmd->get_log_page.nsid));
-		req->error_loc = offsetof(struct nvme_rw_command, nsid);
-		return NVME_SC_INVALID_NS;
-	}
+	status = nvmet_req_find_ns(req);
+	if (status)
+		return status;
 
 	/* we don't have the right data for file backed ns */
 	if (!req->ns->bdev)
@@ -466,9 +462,8 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
 
 static void nvmet_execute_identify_ns(struct nvmet_req *req)
 {
-	struct nvmet_ctrl *ctrl = req->sq->ctrl;
 	struct nvme_id_ns *id;
-	u16 status = 0;
+	u16 status;
 
 	if (le32_to_cpu(req->cmd->identify.nsid) == NVME_NSID_ALL) {
 		req->error_loc = offsetof(struct nvme_identify, nsid);
@@ -483,9 +478,9 @@ static void nvmet_execute_identify_ns(struct nvmet_req *req)
 	}
 
 	/* return an all zeroed buffer if we can't find an active namespace */
-	req->ns = nvmet_find_namespace(ctrl, req->cmd->identify.nsid);
-	if (!req->ns) {
-		status = NVME_SC_INVALID_NS;
+	status = nvmet_req_find_ns(req);
+	if (status) {
+		status = 0;
 		goto done;
 	}
 
@@ -527,7 +522,7 @@ static void nvmet_execute_identify_ns(struct nvmet_req *req)
 
 	id->lbaf[0].ds = req->ns->blksize_shift;
 
-	if (ctrl->pi_support && nvmet_ns_has_pi(req->ns)) {
+	if (req->sq->ctrl->pi_support && nvmet_ns_has_pi(req->ns)) {
 		id->dpc = NVME_NS_DPC_PI_FIRST | NVME_NS_DPC_PI_LAST |
 			  NVME_NS_DPC_PI_TYPE1 | NVME_NS_DPC_PI_TYPE2 |
 			  NVME_NS_DPC_PI_TYPE3;
@@ -604,15 +599,12 @@ static u16 nvmet_copy_ns_identifier(struct nvmet_req *req, u8 type, u8 len,
 
 static void nvmet_execute_identify_desclist(struct nvmet_req *req)
 {
-	u16 status = 0;
 	off_t off = 0;
+	u16 status;
 
-	req->ns = nvmet_find_namespace(req->sq->ctrl, req->cmd->identify.nsid);
-	if (!req->ns) {
-		req->error_loc = offsetof(struct nvme_identify, nsid);
-		status = NVME_SC_INVALID_NS | NVME_SC_DNR;
+	status = nvmet_req_find_ns(req);
+	if (status)
 		goto out;
-	}
 
 	if (memchr_inv(&req->ns->uuid, 0, sizeof(req->ns->uuid))) {
 		status = nvmet_copy_ns_identifier(req, NVME_NIDT_UUID,
@@ -691,14 +683,12 @@ static u16 nvmet_write_protect_flush_sync(struct nvmet_req *req)
 static u16 nvmet_set_feat_write_protect(struct nvmet_req *req)
 {
 	u32 write_protect = le32_to_cpu(req->cmd->common.cdw11);
-	struct nvmet_subsys *subsys = req->sq->ctrl->subsys;
-	u16 status = NVME_SC_FEATURE_NOT_CHANGEABLE;
+	struct nvmet_subsys *subsys = nvmet_req_subsys(req);
+	u16 status;
 
-	req->ns = nvmet_find_namespace(req->sq->ctrl, req->cmd->rw.nsid);
-	if (unlikely(!req->ns)) {
-		req->error_loc = offsetof(struct nvme_common_command, nsid);
+	status = nvmet_req_find_ns(req);
+	if (status)
 		return status;
-	}
 
 	mutex_lock(&subsys->lock);
 	switch (write_protect) {
@@ -752,7 +742,7 @@ u16 nvmet_set_feat_async_event(struct nvmet_req *req, u32 mask)
 
 void nvmet_execute_set_features(struct nvmet_req *req)
 {
-	struct nvmet_subsys *subsys = req->sq->ctrl->subsys;
+	struct nvmet_subsys *subsys = nvmet_req_subsys(req);
 	u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
 	u32 cdw11 = le32_to_cpu(req->cmd->common.cdw11);
 	u16 status = 0;
@@ -796,14 +786,13 @@ void nvmet_execute_set_features(struct nvmet_req *req)
 
 static u16 nvmet_get_feat_write_protect(struct nvmet_req *req)
 {
-	struct nvmet_subsys *subsys = req->sq->ctrl->subsys;
+	struct nvmet_subsys *subsys = nvmet_req_subsys(req);
 	u32 result;
 
-	req->ns = nvmet_find_namespace(req->sq->ctrl, req->cmd->common.nsid);
-	if (!req->ns)  {
-		req->error_loc = offsetof(struct nvme_common_command, nsid);
-		return NVME_SC_INVALID_NS | NVME_SC_DNR;
-	}
+	result = nvmet_req_find_ns(req);
+	if (result)
+		return result;
+
 	mutex_lock(&subsys->lock);
 	if (req->ns->readonly == true)
 		result = NVME_NS_WRITE_PROTECT;
@@ -827,7 +816,7 @@ void nvmet_get_feat_async_event(struct nvmet_req *req)
 
 void nvmet_execute_get_features(struct nvmet_req *req)
 {
-	struct nvmet_subsys *subsys = req->sq->ctrl->subsys;
+	struct nvmet_subsys *subsys = nvmet_req_subsys(req);
 	u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
 	u16 status = 0;
 
@@ -934,7 +923,7 @@ u16 nvmet_parse_admin_cmd(struct nvmet_req *req)
 
 	if (nvme_is_fabrics(cmd))
 		return nvmet_parse_fabrics_cmd(req);
-	if (req->sq->ctrl->subsys->type == NVME_NQN_DISC)
+	if (nvmet_req_subsys(req)->type == NVME_NQN_DISC)
 		return nvmet_parse_discovery_cmd(req);
 
 	ret = nvmet_check_ctrl_status(req, cmd);

drivers/nvme/target/core.c

@@ -82,6 +82,15 @@ inline u16 errno_to_nvme_status(struct nvmet_req *req, int errno)
 	return status;
 }
 
+u16 nvmet_report_invalid_opcode(struct nvmet_req *req)
+{
+	pr_debug("unhandled cmd %d on qid %d\n", req->cmd->common.opcode,
+		 req->sq->qid);
+
+	req->error_loc = offsetof(struct nvme_common_command, opcode);
+	return NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
+}
+
 static struct nvmet_subsys *nvmet_find_get_subsys(struct nvmet_port *port,
 		const char *subsysnqn);
 
@@ -417,15 +426,18 @@ void nvmet_stop_keep_alive_timer(struct nvmet_ctrl *ctrl)
 	cancel_delayed_work_sync(&ctrl->ka_work);
 }
 
-struct nvmet_ns *nvmet_find_namespace(struct nvmet_ctrl *ctrl, __le32 nsid)
+u16 nvmet_req_find_ns(struct nvmet_req *req)
 {
-	struct nvmet_ns *ns;
+	u32 nsid = le32_to_cpu(req->cmd->common.nsid);
 
-	ns = xa_load(&ctrl->subsys->namespaces, le32_to_cpu(nsid));
-	if (ns)
-		percpu_ref_get(&ns->ref);
+	req->ns = xa_load(&nvmet_req_subsys(req)->namespaces, nsid);
+	if (unlikely(!req->ns)) {
+		req->error_loc = offsetof(struct nvme_common_command, nsid);
+		return NVME_SC_INVALID_NS | NVME_SC_DNR;
+	}
 
-	return ns;
+	percpu_ref_get(&req->ns->ref);
+	return NVME_SC_SUCCESS;
 }
 
 static void nvmet_destroy_namespace(struct percpu_ref *ref)
@@ -862,11 +874,10 @@ static u16 nvmet_parse_io_cmd(struct nvmet_req *req)
 	if (nvmet_req_passthru_ctrl(req))
 		return nvmet_parse_passthru_io_cmd(req);
 
-	req->ns = nvmet_find_namespace(req->sq->ctrl, cmd->rw.nsid);
-	if (unlikely(!req->ns)) {
-		req->error_loc = offsetof(struct nvme_common_command, nsid);
-		return NVME_SC_INVALID_NS | NVME_SC_DNR;
-	}
+	ret = nvmet_req_find_ns(req);
+	if (unlikely(ret))
+		return ret;
+
 	ret = nvmet_check_ana_state(req->port, req->ns);
 	if (unlikely(ret)) {
 		req->error_loc = offsetof(struct nvme_common_command, nsid);
@@ -880,8 +891,8 @@ static u16 nvmet_parse_io_cmd(struct nvmet_req *req)
 
 	if (req->ns->file)
 		return nvmet_file_parse_io_cmd(req);
-	else
-		return nvmet_bdev_parse_io_cmd(req);
+
+	return nvmet_bdev_parse_io_cmd(req);
 }
 
 bool nvmet_req_init(struct nvmet_req *req, struct nvmet_cq *cq,

drivers/nvme/target/fc.c

@@ -165,7 +165,7 @@ struct nvmet_fc_tgt_assoc {
 	struct nvmet_fc_hostport	*hostport;
 	struct nvmet_fc_ls_iod		*rcv_disconn;
 	struct list_head		a_list;
-	struct nvmet_fc_tgt_queue	*queues[NVMET_NR_QUEUES + 1];
+	struct nvmet_fc_tgt_queue __rcu	*queues[NVMET_NR_QUEUES + 1];
 	struct kref			ref;
 	struct work_struct		del_work;
 	struct rcu_head			rcu;

drivers/nvme/target/io-cmd-bdev.c

@@ -449,9 +449,6 @@ u16 nvmet_bdev_parse_io_cmd(struct nvmet_req *req)
 		req->execute = nvmet_bdev_execute_write_zeroes;
 		return 0;
 	default:
-		pr_err("unhandled cmd %d on qid %d\n", cmd->common.opcode,
-		       req->sq->qid);
-		req->error_loc = offsetof(struct nvme_common_command, opcode);
-		return NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
+		return nvmet_report_invalid_opcode(req);
 	}
 }

drivers/nvme/target/io-cmd-file.c

@@ -400,9 +400,6 @@ u16 nvmet_file_parse_io_cmd(struct nvmet_req *req)
 		req->execute = nvmet_file_execute_write_zeroes;
 		return 0;
 	default:
-		pr_err("unhandled cmd for file ns %d on qid %d\n",
-				cmd->common.opcode, req->sq->qid);
-		req->error_loc = offsetof(struct nvme_common_command, opcode);
-		return NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
+		return nvmet_report_invalid_opcode(req);
 	}
 }

drivers/nvme/target/nvmet.h

@@ -443,7 +443,7 @@ struct nvmet_subsys *nvmet_subsys_alloc(const char *subsysnqn,
 void nvmet_subsys_put(struct nvmet_subsys *subsys);
 void nvmet_subsys_del_ctrls(struct nvmet_subsys *subsys);
 
-struct nvmet_ns *nvmet_find_namespace(struct nvmet_ctrl *ctrl, __le32 nsid);
+u16 nvmet_req_find_ns(struct nvmet_req *req);
 void nvmet_put_namespace(struct nvmet_ns *ns);
 int nvmet_ns_enable(struct nvmet_ns *ns);
 void nvmet_ns_disable(struct nvmet_ns *ns);
@@ -551,6 +551,11 @@ static inline u32 nvmet_dsm_len(struct nvmet_req *req)
 		sizeof(struct nvme_dsm_range);
 }
 
+static inline struct nvmet_subsys *nvmet_req_subsys(struct nvmet_req *req)
+{
+	return req->sq->ctrl->subsys;
+}
+
 #ifdef CONFIG_NVME_TARGET_PASSTHRU
 void nvmet_passthru_subsys_free(struct nvmet_subsys *subsys);
 int nvmet_passthru_ctrl_enable(struct nvmet_subsys *subsys);
@@ -585,10 +590,11 @@ static inline struct nvme_ctrl *nvmet_passthru_ctrl(struct nvmet_subsys *subsys)
 static inline struct nvme_ctrl *
 nvmet_req_passthru_ctrl(struct nvmet_req *req)
 {
-	return nvmet_passthru_ctrl(req->sq->ctrl->subsys);
+	return nvmet_passthru_ctrl(nvmet_req_subsys(req));
 }
 
 u16 errno_to_nvme_status(struct nvmet_req *req, int errno);
+u16 nvmet_report_invalid_opcode(struct nvmet_req *req);
 
 /* Convert a 32-bit number to a 16-bit 0's based number */
 static inline __le16 to0based(u32 a)

drivers/nvme/target/passthru.c

@@ -239,9 +239,9 @@ static void nvmet_passthru_execute_cmd(struct nvmet_req *req)
 		}
 
 		q = ns->queue;
-		timeout = req->sq->ctrl->subsys->io_timeout;
+		timeout = nvmet_req_subsys(req)->io_timeout;
 	} else {
-		timeout = req->sq->ctrl->subsys->admin_timeout;
+		timeout = nvmet_req_subsys(req)->admin_timeout;
 	}
 
 	rq = nvme_alloc_request(q, req->cmd, 0);
@@ -494,7 +494,7 @@ u16 nvmet_parse_passthru_admin_cmd(struct nvmet_req *req)
 		return nvmet_setup_passthru_command(req);
 	default:
 		/* Reject commands not in the allowlist above */
-		return NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
+		return nvmet_report_invalid_opcode(req);
 	}
 }
 

drivers/nvme/target/tcp.c

@@ -378,7 +378,7 @@ static int nvmet_tcp_map_data(struct nvmet_tcp_cmd *cmd)
 	return NVME_SC_INTERNAL;
 }
 
-static void nvmet_tcp_ddgst(struct ahash_request *hash,
+static void nvmet_tcp_send_ddgst(struct ahash_request *hash,
 		struct nvmet_tcp_cmd *cmd)
 {
 	ahash_request_set_crypt(hash, cmd->req.sg,
@@ -386,6 +386,23 @@ static void nvmet_tcp_ddgst(struct ahash_request *hash,
 	crypto_ahash_digest(hash);
 }
 
+static void nvmet_tcp_recv_ddgst(struct ahash_request *hash,
+		struct nvmet_tcp_cmd *cmd)
+{
+	struct scatterlist sg;
+	struct kvec *iov;
+	int i;
+
+	crypto_ahash_init(hash);
+	for (i = 0, iov = cmd->iov; i < cmd->nr_mapped; i++, iov++) {
+		sg_init_one(&sg, iov->iov_base, iov->iov_len);
+		ahash_request_set_crypt(hash, &sg, NULL, iov->iov_len);
+		crypto_ahash_update(hash);
+	}
+	ahash_request_set_crypt(hash, NULL, (void *)&cmd->exp_ddgst, 0);
+	crypto_ahash_final(hash);
+}
+
 static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)
 {
 	struct nvme_tcp_data_pdu *pdu = cmd->data_pdu;
@@ -410,7 +427,7 @@ static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)
 
 	if (queue->data_digest) {
 		pdu->hdr.flags |= NVME_TCP_F_DDGST;
-		nvmet_tcp_ddgst(queue->snd_hash, cmd);
+		nvmet_tcp_send_ddgst(queue->snd_hash, cmd);
 	}
 
 	if (cmd->queue->hdr_digest) {
@@ -1059,7 +1076,7 @@ static void nvmet_tcp_prep_recv_ddgst(struct nvmet_tcp_cmd *cmd)
 {
 	struct nvmet_tcp_queue *queue = cmd->queue;
 
-	nvmet_tcp_ddgst(queue->rcv_hash, cmd);
+	nvmet_tcp_recv_ddgst(queue->rcv_hash, cmd);
 	queue->offset = 0;
 	queue->left = NVME_TCP_DIGEST_LENGTH;
 	queue->rcv_state = NVMET_TCP_RECV_DDGST;
@@ -1080,14 +1097,14 @@ static int nvmet_tcp_try_recv_data(struct nvmet_tcp_queue *queue)
 		cmd->rbytes_done += ret;
 	}
 
+	if (queue->data_digest) {
+		nvmet_tcp_prep_recv_ddgst(cmd);
+		return 0;
+	}
 	nvmet_tcp_unmap_pdu_iovec(cmd);
 
 	if (!(cmd->flags & NVMET_TCP_F_INIT_FAILED) &&
 	    cmd->rbytes_done == cmd->req.transfer_len) {
-		if (queue->data_digest) {
-			nvmet_tcp_prep_recv_ddgst(cmd);
-			return 0;
-		}
 		cmd->req.execute(&cmd->req);
 	}
 
@@ -1467,17 +1484,27 @@ static int nvmet_tcp_set_queue_sock(struct nvmet_tcp_queue *queue)
 	if (inet->rcv_tos > 0)
 		ip_sock_set_tos(sock->sk, inet->rcv_tos);
 
+	ret = 0;
 	write_lock_bh(&sock->sk->sk_callback_lock);
-	sock->sk->sk_user_data = queue;
-	queue->data_ready = sock->sk->sk_data_ready;
-	sock->sk->sk_data_ready = nvmet_tcp_data_ready;
-	queue->state_change = sock->sk->sk_state_change;
-	sock->sk->sk_state_change = nvmet_tcp_state_change;
-	queue->write_space = sock->sk->sk_write_space;
-	sock->sk->sk_write_space = nvmet_tcp_write_space;
+	if (sock->sk->sk_state != TCP_ESTABLISHED) {
+		/*
+		 * If the socket is already closing, don't even start
+		 * consuming it
+		 */
+		ret = -ENOTCONN;
+	} else {
+		sock->sk->sk_user_data = queue;
+		queue->data_ready = sock->sk->sk_data_ready;
+		sock->sk->sk_data_ready = nvmet_tcp_data_ready;
+		queue->state_change = sock->sk->sk_state_change;
+		sock->sk->sk_state_change = nvmet_tcp_state_change;
+		queue->write_space = sock->sk->sk_write_space;
+		sock->sk->sk_write_space = nvmet_tcp_write_space;
+		queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
+	}
 	write_unlock_bh(&sock->sk->sk_callback_lock);
 
-	return 0;
+	return ret;
 }
 
 static int nvmet_tcp_alloc_queue(struct nvmet_tcp_port *port,
@@ -1525,8 +1552,6 @@ static int nvmet_tcp_alloc_queue(struct nvmet_tcp_port *port,
 	if (ret)
 		goto out_destroy_sq;
 
-	queue_work_on(queue_cpu(queue), nvmet_tcp_wq, &queue->io_work);
-
 	return 0;
 out_destroy_sq:
 	mutex_lock(&nvmet_tcp_queue_mutex);

drivers/nvme/target/trace.h

@@ -48,10 +48,13 @@ static inline struct nvmet_ctrl *nvmet_req_to_ctrl(struct nvmet_req *req)
 
 static inline void __assign_req_name(char *name, struct nvmet_req *req)
 {
-	if (req->ns)
-		strncpy(name, req->ns->device_path, DISK_NAME_LEN);
-	else
+	if (!req->ns) {
 		memset(name, 0, DISK_NAME_LEN);
+		return;
+	}
+
+	strncpy(name, req->ns->device_path,
+		min_t(size_t, DISK_NAME_LEN, strlen(req->ns->device_path)));
 }
 #endif
 

include/linux/blk-mq.h

@@ -490,6 +490,18 @@ static inline int blk_mq_request_completed(struct request *rq)
 	return blk_mq_rq_state(rq) == MQ_RQ_COMPLETE;
 }
 
+/*
+ * 
+ * Set the state to complete when completing a request from inside ->queue_rq.
+ * This is used by drivers that want to ensure special complete actions that
+ * need access to the request are called on failure, e.g. by nvme for
+ * multipathing.
+ */
+static inline void blk_mq_set_request_complete(struct request *rq)
+{
+	WRITE_ONCE(rq->state, MQ_RQ_COMPLETE);
+}
+
 void blk_mq_start_request(struct request *rq);
 void blk_mq_end_request(struct request *rq, blk_status_t error);
 void __blk_mq_end_request(struct request *rq, blk_status_t error);