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);