Merge tag 'nvme-6.9-2024-03-21' of git://git.infradead.org/nvme into block-6.9

Pull NVMe fixes from Keith:

"nvme updates for Linux 6.9

 - Make an informative message less ominous (Keith)
 - Enhanced trace decoding (Guixin)
 - TCP updates (Hannes, Li)
 - Fabrics connect deadlock fix (Chunguang)
 - Platform API migration update (Uwe)
 - A new device quirk (Jiawei)"

* tag 'nvme-6.9-2024-03-21' of git://git.infradead.org/nvme:
  nvmet-rdma: remove NVMET_RDMA_REQ_INVALIDATE_RKEY flag
  nvme: remove redundant BUILD_BUG_ON check
  nvme/tcp: Add wq_unbound modparam for nvme_tcp_wq
  nvme-tcp: Export the nvme_tcp_wq to sysfs
  drivers/nvme: Add quirks for device 126f:2262
  nvme: parse format command's lbafu when tracing
  nvme: add tracing of reservation commands
  nvme: parse zns command's zsa and zrasf to string
  nvme: use nvme_disk_is_ns_head helper
  nvme: fix reconnection fail due to reserved tag allocation
  nvmet: add tracing of zns commands
  nvmet: add tracing of authentication commands
  nvme-apple: Convert to platform remove callback returning void
  nvmet-tcp: do not continue for invalid icreq
  nvme: change shutdown timeout setting message

Author: axboe

drivers/nvme/host/apple.c

@@ -1532,7 +1532,7 @@ static int apple_nvme_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int apple_nvme_remove(struct platform_device *pdev)
+static void apple_nvme_remove(struct platform_device *pdev)
 {
 	struct apple_nvme *anv = platform_get_drvdata(pdev);
 
@@ -1547,8 +1547,6 @@ static int apple_nvme_remove(struct platform_device *pdev)
 		apple_rtkit_shutdown(anv->rtk);
 
 	apple_nvme_detach_genpd(anv);
-
-	return 0;
 }
 
 static void apple_nvme_shutdown(struct platform_device *pdev)
@@ -1598,7 +1596,7 @@ static struct platform_driver apple_nvme_driver = {
 		.pm = pm_sleep_ptr(&apple_nvme_pm_ops),
 	},
 	.probe = apple_nvme_probe,
-	.remove = apple_nvme_remove,
+	.remove_new = apple_nvme_remove,
 	.shutdown = apple_nvme_shutdown,
 };
 module_platform_driver(apple_nvme_driver);

drivers/nvme/host/core.c

@@ -1807,9 +1807,6 @@ static void nvme_config_discard(struct nvme_ns *ns, struct queue_limits *lim)
 {
 	struct nvme_ctrl *ctrl = ns->ctrl;
 
-	BUILD_BUG_ON(PAGE_SIZE / sizeof(struct nvme_dsm_range) <
-			NVME_DSM_MAX_RANGES);
-
 	if (ctrl->dmrsl && ctrl->dmrsl <= nvme_sect_to_lba(ns->head, UINT_MAX))
 		lim->max_hw_discard_sectors =
 			nvme_lba_to_sect(ns->head, ctrl->dmrsl);
@@ -3237,7 +3234,7 @@ static int nvme_init_identify(struct nvme_ctrl *ctrl)
 
 		if (ctrl->shutdown_timeout != shutdown_timeout)
 			dev_info(ctrl->device,
-				 "Shutdown timeout set to %u seconds\n",
+				 "D3 entry latency set to %u seconds\n",
 				 ctrl->shutdown_timeout);
 	} else
 		ctrl->shutdown_timeout = shutdown_timeout;
@@ -4391,7 +4388,8 @@ int nvme_alloc_admin_tag_set(struct nvme_ctrl *ctrl, struct blk_mq_tag_set *set,
 	set->ops = ops;
 	set->queue_depth = NVME_AQ_MQ_TAG_DEPTH;
 	if (ctrl->ops->flags & NVME_F_FABRICS)
-		set->reserved_tags = NVMF_RESERVED_TAGS;
+		/* Reserved for fabric connect and keep alive */
+		set->reserved_tags = 2;
 	set->numa_node = ctrl->numa_node;
 	set->flags = BLK_MQ_F_NO_SCHED;
 	if (ctrl->ops->flags & NVME_F_BLOCKING)
@@ -4460,7 +4458,8 @@ int nvme_alloc_io_tag_set(struct nvme_ctrl *ctrl, struct blk_mq_tag_set *set,
 	if (ctrl->quirks & NVME_QUIRK_SHARED_TAGS)
 		set->reserved_tags = NVME_AQ_DEPTH;
 	else if (ctrl->ops->flags & NVME_F_FABRICS)
-		set->reserved_tags = NVMF_RESERVED_TAGS;
+		/* Reserved for fabric connect */
+		set->reserved_tags = 1;
 	set->numa_node = ctrl->numa_node;
 	set->flags = BLK_MQ_F_SHOULD_MERGE;
 	if (ctrl->ops->flags & NVME_F_BLOCKING)

drivers/nvme/host/fabrics.h

@@ -18,13 +18,6 @@
 /* default is -1: the fail fast mechanism is disabled  */
 #define NVMF_DEF_FAIL_FAST_TMO		-1
 
-/*
- * Reserved one command for internal usage.  This command is used for sending
- * the connect command, as well as for the keep alive command on the admin
- * queue once live.
- */
-#define NVMF_RESERVED_TAGS	1
-
 /*
  * Define a host as seen by the target.  We allocate one at boot, but also
  * allow the override it when creating controllers.  This is both to provide

drivers/nvme/host/pci.c

@@ -3363,6 +3363,9 @@ static const struct pci_device_id nvme_id_table[] = {
 				NVME_QUIRK_BOGUS_NID, },
 	{ PCI_VDEVICE(REDHAT, 0x0010),	/* Qemu emulated controller */
 		.driver_data = NVME_QUIRK_BOGUS_NID, },
+	{ PCI_DEVICE(0x126f, 0x2262),	/* Silicon Motion generic */
+		.driver_data = NVME_QUIRK_NO_DEEPEST_PS |
+				NVME_QUIRK_BOGUS_NID, },
 	{ PCI_DEVICE(0x126f, 0x2263),	/* Silicon Motion unidentified */
 		.driver_data = NVME_QUIRK_NO_NS_DESC_LIST |
 				NVME_QUIRK_BOGUS_NID, },

drivers/nvme/host/pr.c

@@ -97,8 +97,7 @@ static int nvme_sc_to_pr_err(int nvme_sc)
 static int nvme_send_pr_command(struct block_device *bdev,
 		struct nvme_command *c, void *data, unsigned int data_len)
 {
-	if (IS_ENABLED(CONFIG_NVME_MULTIPATH) &&
-	    nvme_disk_is_ns_head(bdev->bd_disk))
+	if (nvme_disk_is_ns_head(bdev->bd_disk))
 		return nvme_send_ns_head_pr_command(bdev, c, data, data_len);
 
 	return nvme_send_ns_pr_command(bdev->bd_disk->private_data, c, data,

drivers/nvme/host/sysfs.c

@@ -236,8 +236,7 @@ static ssize_t nuse_show(struct device *dev, struct device_attribute *attr,
 	struct block_device *bdev = disk->part0;
 	int ret;
 
-	if (IS_ENABLED(CONFIG_NVME_MULTIPATH) &&
-	    bdev->bd_disk->fops == &nvme_ns_head_ops)
+	if (nvme_disk_is_ns_head(bdev->bd_disk))
 		ret = ns_head_update_nuse(head);
 	else
 		ret = ns_update_nuse(bdev->bd_disk->private_data);

drivers/nvme/host/tcp.c

@@ -36,6 +36,14 @@ static int so_priority;
 module_param(so_priority, int, 0644);
 MODULE_PARM_DESC(so_priority, "nvme tcp socket optimize priority");
 
+/*
+ * Use the unbound workqueue for nvme_tcp_wq, then we can set the cpu affinity
+ * from sysfs.
+ */
+static bool wq_unbound;
+module_param(wq_unbound, bool, 0644);
+MODULE_PARM_DESC(wq_unbound, "Use unbound workqueue for nvme-tcp IO context (default false)");
+
 /*
  * TLS handshake timeout
  */
@@ -1546,7 +1554,10 @@ static void nvme_tcp_set_queue_io_cpu(struct nvme_tcp_queue *queue)
 	else if (nvme_tcp_poll_queue(queue))
 		n = qid - ctrl->io_queues[HCTX_TYPE_DEFAULT] -
 				ctrl->io_queues[HCTX_TYPE_READ] - 1;
-	queue->io_cpu = cpumask_next_wrap(n - 1, cpu_online_mask, -1, false);
+	if (wq_unbound)
+		queue->io_cpu = WORK_CPU_UNBOUND;
+	else
+		queue->io_cpu = cpumask_next_wrap(n - 1, cpu_online_mask, -1, false);
 }
 
 static void nvme_tcp_tls_done(void *data, int status, key_serial_t pskid)
@@ -2785,6 +2796,8 @@ static struct nvmf_transport_ops nvme_tcp_transport = {
 
 static int __init nvme_tcp_init_module(void)
 {
+	unsigned int wq_flags = WQ_MEM_RECLAIM | WQ_HIGHPRI | WQ_SYSFS;
+
 	BUILD_BUG_ON(sizeof(struct nvme_tcp_hdr) != 8);
 	BUILD_BUG_ON(sizeof(struct nvme_tcp_cmd_pdu) != 72);
 	BUILD_BUG_ON(sizeof(struct nvme_tcp_data_pdu) != 24);
@@ -2794,8 +2807,10 @@ static int __init nvme_tcp_init_module(void)
 	BUILD_BUG_ON(sizeof(struct nvme_tcp_icresp_pdu) != 128);
 	BUILD_BUG_ON(sizeof(struct nvme_tcp_term_pdu) != 24);
 
-	nvme_tcp_wq = alloc_workqueue("nvme_tcp_wq",
-			WQ_MEM_RECLAIM | WQ_HIGHPRI, 0);
+	if (wq_unbound)
+		wq_flags |= WQ_UNBOUND;
+
+	nvme_tcp_wq = alloc_workqueue("nvme_tcp_wq", wq_flags, 0);
 	if (!nvme_tcp_wq)
 		return -ENOMEM;
 

drivers/nvme/host/trace.c

@@ -119,7 +119,10 @@ static const char *nvme_trace_get_lba_status(struct trace_seq *p,
 static const char *nvme_trace_admin_format_nvm(struct trace_seq *p, u8 *cdw10)
 {
 	const char *ret = trace_seq_buffer_ptr(p);
-	u8 lbaf = cdw10[0] & 0xF;
+	/*
+	 * lbafu(bit 13:12) is already in the upper 4 bits, lbafl: bit 03:00.
+	 */
+	u8 lbaf = (cdw10[1] & 0x30) | (cdw10[0] & 0xF);
 	u8 mset = (cdw10[0] >> 4) & 0x1;
 	u8 pi = (cdw10[0] >> 5) & 0x7;
 	u8 pil = cdw10[1] & 0x1;
@@ -164,28 +167,114 @@ static const char *nvme_trace_dsm(struct trace_seq *p, u8 *cdw10)
 
 static const char *nvme_trace_zone_mgmt_send(struct trace_seq *p, u8 *cdw10)
 {
+	static const char * const zsa_strs[] = {
+		[0x01] = "close zone",
+		[0x02] = "finish zone",
+		[0x03] = "open zone",
+		[0x04] = "reset zone",
+		[0x05] = "offline zone",
+		[0x10] = "set zone descriptor extension"
+	};
 	const char *ret = trace_seq_buffer_ptr(p);
 	u64 slba = get_unaligned_le64(cdw10);
+	const char *zsa_str;
 	u8 zsa = cdw10[12];
 	u8 all = cdw10[13];
 
-	trace_seq_printf(p, "slba=%llu, zsa=%u, all=%u", slba, zsa, all);
+	if (zsa < ARRAY_SIZE(zsa_strs) && zsa_strs[zsa])
+		zsa_str = zsa_strs[zsa];
+	else
+		zsa_str = "reserved";
+
+	trace_seq_printf(p, "slba=%llu, zsa=%u:%s, all=%u",
+		slba, zsa, zsa_str, all);
 	trace_seq_putc(p, 0);
 
 	return ret;
 }
 
 static const char *nvme_trace_zone_mgmt_recv(struct trace_seq *p, u8 *cdw10)
 {
+	static const char * const zrasf_strs[] = {
+		[0x00] = "list all zones",
+		[0x01] = "list the zones in the ZSE: Empty state",
+		[0x02] = "list the zones in the ZSIO: Implicitly Opened state",
+		[0x03] = "list the zones in the ZSEO: Explicitly Opened state",
+		[0x04] = "list the zones in the ZSC: Closed state",
+		[0x05] = "list the zones in the ZSF: Full state",
+		[0x06] = "list the zones in the ZSRO: Read Only state",
+		[0x07] = "list the zones in the ZSO: Offline state",
+		[0x09] = "list the zones that have the zone attribute"
+	};
 	const char *ret = trace_seq_buffer_ptr(p);
 	u64 slba = get_unaligned_le64(cdw10);
 	u32 numd = get_unaligned_le32(cdw10 + 8);
 	u8 zra = cdw10[12];
 	u8 zrasf = cdw10[13];
+	const char *zrasf_str;
 	u8 pr = cdw10[14];
 
-	trace_seq_printf(p, "slba=%llu, numd=%u, zra=%u, zrasf=%u, pr=%u",
-			 slba, numd, zra, zrasf, pr);
+	if (zrasf < ARRAY_SIZE(zrasf_strs) && zrasf_strs[zrasf])
+		zrasf_str = zrasf_strs[zrasf];
+	else
+		zrasf_str = "reserved";
+
+	trace_seq_printf(p, "slba=%llu, numd=%u, zra=%u, zrasf=%u:%s, pr=%u",
+		slba, numd, zra, zrasf, zrasf_str, pr);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_resv_reg(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u8 rrega = cdw10[0] & 0x7;
+	u8 iekey = (cdw10[0] >> 3) & 0x1;
+	u8 ptpl = (cdw10[3] >> 6) & 0x3;
+
+	trace_seq_printf(p, "rrega=%u, iekey=%u, ptpl=%u",
+			 rrega, iekey, ptpl);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_resv_acq(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u8 racqa = cdw10[0] & 0x7;
+	u8 iekey = (cdw10[0] >> 3) & 0x1;
+	u8 rtype = cdw10[1];
+
+	trace_seq_printf(p, "racqa=%u, iekey=%u, rtype=%u",
+			 racqa, iekey, rtype);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_resv_rel(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u8 rrela = cdw10[0] & 0x7;
+	u8 iekey = (cdw10[0] >> 3) & 0x1;
+	u8 rtype = cdw10[1];
+
+	trace_seq_printf(p, "rrela=%u, iekey=%u, rtype=%u",
+			 rrela, iekey, rtype);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_resv_report(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u32 numd = get_unaligned_le32(cdw10);
+	u8 eds = cdw10[4] & 0x1;
+
+	trace_seq_printf(p, "numd=%u, eds=%u", numd, eds);
 	trace_seq_putc(p, 0);
 
 	return ret;
@@ -243,6 +332,14 @@ const char *nvme_trace_parse_nvm_cmd(struct trace_seq *p,
 		return nvme_trace_zone_mgmt_send(p, cdw10);
 	case nvme_cmd_zone_mgmt_recv:
 		return nvme_trace_zone_mgmt_recv(p, cdw10);
+	case nvme_cmd_resv_register:
+		return nvme_trace_resv_reg(p, cdw10);
+	case nvme_cmd_resv_acquire:
+		return nvme_trace_resv_acq(p, cdw10);
+	case nvme_cmd_resv_release:
+		return nvme_trace_resv_rel(p, cdw10);
+	case nvme_cmd_resv_report:
+		return nvme_trace_resv_report(p, cdw10);
 	default:
 		return nvme_trace_common(p, cdw10);
 	}

drivers/nvme/target/rdma.c

@@ -53,7 +53,6 @@ struct nvmet_rdma_cmd {
 
 enum {
 	NVMET_RDMA_REQ_INLINE_DATA	= (1 << 0),
-	NVMET_RDMA_REQ_INVALIDATE_RKEY	= (1 << 1),
 };
 
 struct nvmet_rdma_rsp {
@@ -722,7 +721,7 @@ static void nvmet_rdma_queue_response(struct nvmet_req *req)
 	struct rdma_cm_id *cm_id = rsp->queue->cm_id;
 	struct ib_send_wr *first_wr;
 
-	if (rsp->flags & NVMET_RDMA_REQ_INVALIDATE_RKEY) {
+	if (rsp->invalidate_rkey) {
 		rsp->send_wr.opcode = IB_WR_SEND_WITH_INV;
 		rsp->send_wr.ex.invalidate_rkey = rsp->invalidate_rkey;
 	} else {
@@ -905,10 +904,8 @@ static u16 nvmet_rdma_map_sgl_keyed(struct nvmet_rdma_rsp *rsp,
 		goto error_out;
 	rsp->n_rdma += ret;
 
-	if (invalidate) {
+	if (invalidate)
 		rsp->invalidate_rkey = key;
-		rsp->flags |= NVMET_RDMA_REQ_INVALIDATE_RKEY;
-	}
 
 	return 0;
 
@@ -1047,6 +1044,7 @@ static void nvmet_rdma_recv_done(struct ib_cq *cq, struct ib_wc *wc)
 	rsp->req.cmd = cmd->nvme_cmd;
 	rsp->req.port = queue->port;
 	rsp->n_rdma = 0;
+	rsp->invalidate_rkey = 0;
 
 	if (unlikely(queue->state != NVMET_RDMA_Q_LIVE)) {
 		unsigned long flags;

drivers/nvme/target/tcp.c

@@ -898,6 +898,7 @@ static int nvmet_tcp_handle_icreq(struct nvmet_tcp_queue *queue)
 		pr_err("bad nvme-tcp pdu length (%d)\n",
 			le32_to_cpu(icreq->hdr.plen));
 		nvmet_tcp_fatal_error(queue);
+		return -EPROTO;
 	}
 
 	if (icreq->pfv != NVME_TCP_PFV_1_0) {