Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull SCSI updates from James Bottomley:
 "This is mostly update of the usual drivers: arcmsr, qla2xxx, lpfc,
  hisi_sas, target/iscsi and target/core.

  Additionally Christoph refactored gdth as part of the dma changes. The
  major mid-layer change this time is the removal of bidi commands and
  with them the whole of the osd/exofs driver and filesystem. This is a
  major simplification for block and mq in particular"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (240 commits)
  scsi: cxgb4i: validate tcp sequence number only if chip version <= T5
  scsi: cxgb4i: get pf number from lldi->pf
  scsi: core: replace GFP_ATOMIC with GFP_KERNEL in scsi_scan.c
  scsi: mpt3sas: Add missing breaks in switch statements
  scsi: aacraid: Fix missing break in switch statement
  scsi: kill command serial number
  scsi: csiostor: drop serial_number usage
  scsi: mvumi: use request tag instead of serial_number
  scsi: dpt_i2o: remove serial number usage
  scsi: st: osst: Remove negative constant left-shifts
  scsi: ufs-bsg: Allow reading descriptors
  scsi: ufs: Allow reading descriptor via raw upiu
  scsi: ufs-bsg: Change the calling convention for write descriptor
  scsi: ufs: Remove unused device quirks
  Revert "scsi: ufs: disable vccq if it's not needed by UFS device"
  scsi: megaraid_sas: Remove a bunch of set but not used variables
  scsi: clean obsolete return values of eh_timed_out
  scsi: sd: Optimal I/O size should be a multiple of physical block size
  scsi: MAINTAINERS: SCSI initiator and target tweaks
  scsi: fcoe: make use of fip_mode enum complete
  ...

5 files changed, 1233 insertions, 125 deletions

diff --git a/drivers/scsi/hisi_sas/hisi_sas.h b/drivers/scsi/hisi_sas/hisi_sas.h
index af291947a54d..6c87bd34509a 100644
--- a/drivers/scsi/hisi_sas/hisi_sas.h
+++ b/drivers/scsi/hisi_sas/hisi_sas.h
@@ -14,6 +14,7 @@
 
 #include <linux/acpi.h>
 #include <linux/clk.h>
+#include <linux/debugfs.h>
 #include <linux/dmapool.h>
 #include <linux/iopoll.h>
 #include <linux/lcm.h>
@@ -29,7 +30,7 @@
 
 #define HISI_SAS_MAX_PHYS	9
 #define HISI_SAS_MAX_QUEUES	32
-#define HISI_SAS_QUEUE_SLOTS 512
+#define HISI_SAS_QUEUE_SLOTS	4096
 #define HISI_SAS_MAX_ITCT_ENTRIES 1024
 #define HISI_SAS_MAX_DEVICES HISI_SAS_MAX_ITCT_ENTRIES
 #define HISI_SAS_RESET_BIT	0
@@ -40,20 +41,25 @@
 #define HISI_SAS_COMMAND_TABLE_SZ (sizeof(union hisi_sas_command_table))
 
 #define hisi_sas_status_buf_addr(buf) \
-	(buf + offsetof(struct hisi_sas_slot_buf_table, status_buffer))
-#define hisi_sas_status_buf_addr_mem(slot) hisi_sas_status_buf_addr(slot->buf)
+	((buf) + offsetof(struct hisi_sas_slot_buf_table, status_buffer))
+#define hisi_sas_status_buf_addr_mem(slot) hisi_sas_status_buf_addr((slot)->buf)
 #define hisi_sas_status_buf_addr_dma(slot) \
-	hisi_sas_status_buf_addr(slot->buf_dma)
+	hisi_sas_status_buf_addr((slot)->buf_dma)
 
 #define hisi_sas_cmd_hdr_addr(buf) \
-	(buf + offsetof(struct hisi_sas_slot_buf_table, command_header))
-#define hisi_sas_cmd_hdr_addr_mem(slot) hisi_sas_cmd_hdr_addr(slot->buf)
-#define hisi_sas_cmd_hdr_addr_dma(slot) hisi_sas_cmd_hdr_addr(slot->buf_dma)
+	((buf) + offsetof(struct hisi_sas_slot_buf_table, command_header))
+#define hisi_sas_cmd_hdr_addr_mem(slot) hisi_sas_cmd_hdr_addr((slot)->buf)
+#define hisi_sas_cmd_hdr_addr_dma(slot) hisi_sas_cmd_hdr_addr((slot)->buf_dma)
 
 #define hisi_sas_sge_addr(buf) \
-	(buf + offsetof(struct hisi_sas_slot_buf_table, sge_page))
-#define hisi_sas_sge_addr_mem(slot) hisi_sas_sge_addr(slot->buf)
-#define hisi_sas_sge_addr_dma(slot) hisi_sas_sge_addr(slot->buf_dma)
+	((buf) + offsetof(struct hisi_sas_slot_buf_table, sge_page))
+#define hisi_sas_sge_addr_mem(slot) hisi_sas_sge_addr((slot)->buf)
+#define hisi_sas_sge_addr_dma(slot) hisi_sas_sge_addr((slot)->buf_dma)
+
+#define hisi_sas_sge_dif_addr(buf) \
+	((buf) + offsetof(struct hisi_sas_slot_dif_buf_table, sge_dif_page))
+#define hisi_sas_sge_dif_addr_mem(slot) hisi_sas_sge_dif_addr((slot)->buf)
+#define hisi_sas_sge_dif_addr_dma(slot) hisi_sas_sge_dif_addr((slot)->buf_dma)
 
 #define HISI_SAS_MAX_SSP_RESP_SZ (sizeof(struct ssp_frame_hdr) + 1024)
 #define HISI_SAS_MAX_SMP_RESP_SZ 1028
@@ -73,7 +79,13 @@
 				SHOST_DIF_TYPE2_PROTECTION | \
 				SHOST_DIF_TYPE3_PROTECTION)
 
-#define HISI_SAS_PROT_MASK (HISI_SAS_DIF_PROT_MASK)
+#define HISI_SAS_DIX_PROT_MASK (SHOST_DIX_TYPE1_PROTECTION | \
+				SHOST_DIX_TYPE2_PROTECTION | \
+				SHOST_DIX_TYPE3_PROTECTION)
+
+#define HISI_SAS_PROT_MASK (HISI_SAS_DIF_PROT_MASK | HISI_SAS_DIX_PROT_MASK)
+
+#define HISI_SAS_WAIT_PHYUP_TIMEOUT 20
 
 struct hisi_hba;
 
@@ -82,11 +94,6 @@ enum {
 	PORT_TYPE_SATA = (1U << 0),
 };
 
-enum dev_status {
-	HISI_SAS_DEV_NORMAL,
-	HISI_SAS_DEV_EH,
-};
-
 enum {
 	HISI_SAS_INT_ABT_CMD = 0,
 	HISI_SAS_INT_ABT_DEV = 1,
@@ -145,6 +152,7 @@ struct hisi_sas_phy {
 	struct asd_sas_phy	sas_phy;
 	struct sas_identify	identify;
 	struct completion *reset_completion;
+	struct timer_list timer;
 	spinlock_t lock;
 	u64		port_id; /* from hw */
 	u64		frame_rcvd_size;
@@ -165,6 +173,7 @@ struct hisi_sas_port {
 
 struct hisi_sas_cq {
 	struct hisi_hba *hisi_hba;
+	const struct cpumask *pci_irq_mask;
 	struct tasklet_struct tasklet;
 	int	rd_point;
 	int	id;
@@ -187,7 +196,7 @@ struct hisi_sas_device {
 	enum sas_device_type	dev_type;
 	int device_id;
 	int sata_idx;
-	u8 dev_status;
+	spinlock_t lock; /* For protecting slots */
 };
 
 struct hisi_sas_tmf_task {
@@ -203,12 +212,14 @@ struct hisi_sas_slot {
 	struct sas_task *task;
 	struct hisi_sas_port	*port;
 	u64	n_elem;
+	u64	n_elem_dif;
 	int	dlvry_queue;
 	int	dlvry_queue_slot;
 	int	cmplt_queue;
 	int	cmplt_queue_slot;
 	int	abort;
 	int	ready;
+	int	device_id;
 	void	*cmd_hdr;
 	dma_addr_t cmd_hdr_dma;
 	struct timer_list internal_abort_timer;
@@ -220,6 +231,24 @@ struct hisi_sas_slot {
 	u16	idx;
 };
 
+#define HISI_SAS_DEBUGFS_REG(x) {#x, x}
+
+struct hisi_sas_debugfs_reg_lu {
+	char *name;
+	int off;
+};
+
+struct hisi_sas_debugfs_reg {
+	const struct hisi_sas_debugfs_reg_lu *lu;
+	int count;
+	int base_off;
+	union {
+		u32 (*read_global_reg)(struct hisi_hba *hisi_hba, u32 off);
+		u32 (*read_port_reg)(struct hisi_hba *hisi_hba, int port,
+				     u32 off);
+	};
+};
+
 struct hisi_sas_hw {
 	int (*hw_init)(struct hisi_hba *hisi_hba);
 	void (*setup_itct)(struct hisi_hba *hisi_hba,
@@ -227,7 +256,7 @@ struct hisi_sas_hw {
 	int (*slot_index_alloc)(struct hisi_hba *hisi_hba,
 				struct domain_device *device);
 	struct hisi_sas_device *(*alloc_dev)(struct domain_device *device);
-	void (*sl_notify)(struct hisi_hba *hisi_hba, int phy_no);
+	void (*sl_notify_ssp)(struct hisi_hba *hisi_hba, int phy_no);
 	int (*get_free_slot)(struct hisi_hba *hisi_hba, struct hisi_sas_dq *dq);
 	void (*start_delivery)(struct hisi_sas_dq *dq);
 	void (*prep_ssp)(struct hisi_hba *hisi_hba,
@@ -259,11 +288,16 @@ struct hisi_sas_hw {
 	u32 (*get_phys_state)(struct hisi_hba *hisi_hba);
 	int (*write_gpio)(struct hisi_hba *hisi_hba, u8 reg_type,
 				u8 reg_index, u8 reg_count, u8 *write_data);
-	void (*wait_cmds_complete_timeout)(struct hisi_hba *hisi_hba,
-					   int delay_ms, int timeout_ms);
+	int (*wait_cmds_complete_timeout)(struct hisi_hba *hisi_hba,
+					  int delay_ms, int timeout_ms);
+	void (*snapshot_prepare)(struct hisi_hba *hisi_hba);
+	void (*snapshot_restore)(struct hisi_hba *hisi_hba);
 	int max_command_entries;
 	int complete_hdr_size;
 	struct scsi_host_template *sht;
+
+	const struct hisi_sas_debugfs_reg *debugfs_reg_global;
+	const struct hisi_sas_debugfs_reg *debugfs_reg_port;
 };
 
 struct hisi_hba {
@@ -329,9 +363,25 @@ struct hisi_hba {
 	const struct hisi_sas_hw *hw;	/* Low level hw interface */
 	unsigned long sata_dev_bitmap[BITS_TO_LONGS(HISI_SAS_MAX_DEVICES)];
 	struct work_struct rst_work;
+	struct work_struct debugfs_work;
 	u32 phy_state;
 	u32 intr_coal_ticks;	/* Time of interrupt coalesce in us */
 	u32 intr_coal_count;	/* Interrupt count to coalesce */
+
+	int cq_nvecs;
+	unsigned int *reply_map;
+
+	/* debugfs memories */
+	u32 *debugfs_global_reg;
+	u32 *debugfs_port_reg[HISI_SAS_MAX_PHYS];
+	void *debugfs_complete_hdr[HISI_SAS_MAX_QUEUES];
+	struct hisi_sas_cmd_hdr	*debugfs_cmd_hdr[HISI_SAS_MAX_QUEUES];
+	struct hisi_sas_iost *debugfs_iost;
+	struct hisi_sas_itct *debugfs_itct;
+
+	struct dentry *debugfs_dir;
+	struct dentry *debugfs_dump_dentry;
+	bool debugfs_snapshot;
 };
 
 /* Generic HW DMA host memory structures */
@@ -430,6 +480,11 @@ struct hisi_sas_sge_page {
 	struct hisi_sas_sge sge[HISI_SAS_SGE_PAGE_CNT];
 }  __aligned(16);
 
+#define HISI_SAS_SGE_DIF_PAGE_CNT   SG_CHUNK_SIZE
+struct hisi_sas_sge_dif_page {
+	struct hisi_sas_sge sge[HISI_SAS_SGE_DIF_PAGE_CNT];
+}  __aligned(16);
+
 struct hisi_sas_command_table_ssp {
 	struct ssp_frame_hdr hdr;
 	union {
@@ -460,9 +515,18 @@ struct hisi_sas_slot_buf_table {
 	struct hisi_sas_sge_page sge_page;
 };
 
+struct hisi_sas_slot_dif_buf_table {
+	struct hisi_sas_slot_buf_table slot_buf;
+	struct hisi_sas_sge_dif_page sge_dif_page;
+};
+
 extern struct scsi_transport_template *hisi_sas_stt;
+
+extern bool hisi_sas_debugfs_enable;
+extern struct dentry *hisi_sas_debugfs_dir;
+
 extern void hisi_sas_stop_phys(struct hisi_hba *hisi_hba);
-extern int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost);
+extern int hisi_sas_alloc(struct hisi_hba *hisi_hba);
 extern void hisi_sas_free(struct hisi_hba *hisi_hba);
 extern u8 hisi_sas_get_ata_protocol(struct host_to_dev_fis *fis,
 				int direction);
@@ -487,10 +551,14 @@ extern void hisi_sas_init_mem(struct hisi_hba *hisi_hba);
 extern void hisi_sas_rst_work_handler(struct work_struct *work);
 extern void hisi_sas_sync_rst_work_handler(struct work_struct *work);
 extern void hisi_sas_kill_tasklets(struct hisi_hba *hisi_hba);
+extern void hisi_sas_phy_oob_ready(struct hisi_hba *hisi_hba, int phy_no);
 extern bool hisi_sas_notify_phy_event(struct hisi_sas_phy *phy,
 				enum hisi_sas_phy_event event);
 extern void hisi_sas_release_tasks(struct hisi_hba *hisi_hba);
 extern u8 hisi_sas_get_prog_phy_linkrate_mask(enum sas_linkrate max);
 extern void hisi_sas_controller_reset_prepare(struct hisi_hba *hisi_hba);
 extern void hisi_sas_controller_reset_done(struct hisi_hba *hisi_hba);
+extern void hisi_sas_debugfs_init(struct hisi_hba *hisi_hba);
+extern void hisi_sas_debugfs_exit(struct hisi_hba *hisi_hba);
+extern void hisi_sas_debugfs_work_handler(struct work_struct *work);
 #endif
diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index bc17fa0d8375..13ca5a0bdf6b 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -144,7 +144,7 @@ EXPORT_SYMBOL_GPL(hisi_sas_get_ncq_tag);
  */
 u8 hisi_sas_get_prog_phy_linkrate_mask(enum sas_linkrate max)
 {
-	u16 rate = 0;
+	u8 rate = 0;
 	int i;
 
 	max -= SAS_LINK_RATE_1_5_GBPS;
@@ -241,8 +241,9 @@ static void hisi_sas_slot_index_init(struct hisi_hba *hisi_hba)
 void hisi_sas_slot_task_free(struct hisi_hba *hisi_hba, struct sas_task *task,
 			     struct hisi_sas_slot *slot)
 {
-	struct hisi_sas_dq *dq = &hisi_hba->dq[slot->dlvry_queue];
 	unsigned long flags;
+	int device_id = slot->device_id;
+	struct hisi_sas_device *sas_dev = &hisi_hba->devices[device_id];
 
 	if (task) {
 		struct device *dev = hisi_hba->dev;
@@ -252,17 +253,24 @@ void hisi_sas_slot_task_free(struct hisi_hba *hisi_hba, struct sas_task *task,
 
 		task->lldd_task = NULL;
 
-		if (!sas_protocol_ata(task->task_proto))
+		if (!sas_protocol_ata(task->task_proto)) {
+			struct sas_ssp_task *ssp_task = &task->ssp_task;
+			struct scsi_cmnd *scsi_cmnd = ssp_task->cmd;
+
 			if (slot->n_elem)
 				dma_unmap_sg(dev, task->scatter,
 					     task->num_scatter,
 					     task->data_dir);
+			if (slot->n_elem_dif)
+				dma_unmap_sg(dev, scsi_prot_sglist(scsi_cmnd),
+					     scsi_prot_sg_count(scsi_cmnd),
+					     task->data_dir);
+		}
 	}
 
-
-	spin_lock_irqsave(&dq->lock, flags);
+	spin_lock_irqsave(&sas_dev->lock, flags);
 	list_del_init(&slot->entry);
-	spin_unlock_irqrestore(&dq->lock, flags);
+	spin_unlock_irqrestore(&sas_dev->lock, flags);
 
 	memset(slot, 0, offsetof(struct hisi_sas_slot, buf));
 
@@ -380,6 +388,59 @@ prep_out:
 	return rc;
 }
 
+static void hisi_sas_dif_dma_unmap(struct hisi_hba *hisi_hba,
+				   struct sas_task *task, int n_elem_dif)
+{
+	struct device *dev = hisi_hba->dev;
+
+	if (n_elem_dif) {
+		struct sas_ssp_task *ssp_task = &task->ssp_task;
+		struct scsi_cmnd *scsi_cmnd = ssp_task->cmd;
+
+		dma_unmap_sg(dev, scsi_prot_sglist(scsi_cmnd),
+			     scsi_prot_sg_count(scsi_cmnd),
+			     task->data_dir);
+	}
+}
+
+static int hisi_sas_dif_dma_map(struct hisi_hba *hisi_hba,
+				int *n_elem_dif, struct sas_task *task)
+{
+	struct device *dev = hisi_hba->dev;
+	struct sas_ssp_task *ssp_task;
+	struct scsi_cmnd *scsi_cmnd;
+	int rc;
+
+	if (task->num_scatter) {
+		ssp_task = &task->ssp_task;
+		scsi_cmnd = ssp_task->cmd;
+
+		if (scsi_prot_sg_count(scsi_cmnd)) {
+			*n_elem_dif = dma_map_sg(dev,
+						 scsi_prot_sglist(scsi_cmnd),
+						 scsi_prot_sg_count(scsi_cmnd),
+						 task->data_dir);
+
+			if (!*n_elem_dif)
+				return -ENOMEM;
+
+			if (*n_elem_dif > HISI_SAS_SGE_DIF_PAGE_CNT) {
+				dev_err(dev, "task prep: n_elem_dif(%d) too large\n",
+					*n_elem_dif);
+				rc = -EINVAL;
+				goto err_out_dif_dma_unmap;
+			}
+		}
+	}
+
+	return 0;
+
+err_out_dif_dma_unmap:
+	dma_unmap_sg(dev, scsi_prot_sglist(scsi_cmnd),
+		     scsi_prot_sg_count(scsi_cmnd), task->data_dir);
+	return rc;
+}
+
 static int hisi_sas_task_prep(struct sas_task *task,
 			      struct hisi_sas_dq **dq_pointer,
 			      bool is_tmf, struct hisi_sas_tmf_task *tmf,
@@ -394,7 +455,7 @@ static int hisi_sas_task_prep(struct sas_task *task,
 	struct asd_sas_port *sas_port = device->port;
 	struct device *dev = hisi_hba->dev;
 	int dlvry_queue_slot, dlvry_queue, rc, slot_idx;
-	int n_elem = 0, n_elem_req = 0, n_elem_resp = 0;
+	int n_elem = 0, n_elem_dif = 0, n_elem_req = 0, n_elem_resp = 0;
 	struct hisi_sas_dq *dq;
 	unsigned long flags;
 	int wr_q_index;
@@ -410,7 +471,14 @@ static int hisi_sas_task_prep(struct sas_task *task,
 		return -ECOMM;
 	}
 
-	*dq_pointer = dq = sas_dev->dq;
+	if (hisi_hba->reply_map) {
+		int cpu = raw_smp_processor_id();
+		unsigned int dq_index = hisi_hba->reply_map[cpu];
+
+		*dq_pointer = dq = &hisi_hba->dq[dq_index];
+	} else {
+		*dq_pointer = dq = sas_dev->dq;
+	}
 
 	port = to_hisi_sas_port(sas_port);
 	if (port && !port->port_attached) {
@@ -427,6 +495,12 @@ static int hisi_sas_task_prep(struct sas_task *task,
 	if (rc < 0)
 		goto prep_out;
 
+	if (!sas_protocol_ata(task->task_proto)) {
+		rc = hisi_sas_dif_dma_map(hisi_hba, &n_elem_dif, task);
+		if (rc < 0)
+			goto err_out_dma_unmap;
+	}
+
 	if (hisi_hba->hw->slot_index_alloc)
 		rc = hisi_hba->hw->slot_index_alloc(hisi_hba, device);
 	else {
@@ -445,7 +519,7 @@ static int hisi_sas_task_prep(struct sas_task *task,
 		rc  = hisi_sas_slot_index_alloc(hisi_hba, scsi_cmnd);
 	}
 	if (rc < 0)
-		goto err_out_dma_unmap;
+		goto err_out_dif_dma_unmap;
 
 	slot_idx = rc;
 	slot = &hisi_hba->slot_info[slot_idx];
@@ -459,13 +533,17 @@ static int hisi_sas_task_prep(struct sas_task *task,
 	}
 
 	list_add_tail(&slot->delivery, &dq->list);
-	list_add_tail(&slot->entry, &sas_dev->list);
 	spin_unlock_irqrestore(&dq->lock, flags);
+	spin_lock_irqsave(&sas_dev->lock, flags);
+	list_add_tail(&slot->entry, &sas_dev->list);
+	spin_unlock_irqrestore(&sas_dev->lock, flags);
 
 	dlvry_queue = dq->id;
 	dlvry_queue_slot = wr_q_index;
 
+	slot->device_id = sas_dev->device_id;
 	slot->n_elem = n_elem;
+	slot->n_elem_dif = n_elem_dif;
 	slot->dlvry_queue = dlvry_queue;
 	slot->dlvry_queue_slot = dlvry_queue_slot;
 	cmd_hdr_base = hisi_hba->cmd_hdr[dlvry_queue];
@@ -509,6 +587,9 @@ static int hisi_sas_task_prep(struct sas_task *task,
 
 err_out_tag:
 	hisi_sas_slot_index_free(hisi_hba, slot_idx);
+err_out_dif_dma_unmap:
+	if (!sas_protocol_ata(task->task_proto))
+		hisi_sas_dif_dma_unmap(hisi_hba, task, n_elem_dif);
 err_out_dma_unmap:
 	hisi_sas_dma_unmap(hisi_hba, task, n_elem,
 			   n_elem_req, n_elem_resp);
@@ -626,11 +707,11 @@ static struct hisi_sas_device *hisi_sas_alloc_dev(struct domain_device *device)
 
 			hisi_hba->devices[i].device_id = i;
 			sas_dev = &hisi_hba->devices[i];
-			sas_dev->dev_status = HISI_SAS_DEV_NORMAL;
 			sas_dev->dev_type = device->dev_type;
 			sas_dev->hisi_hba = hisi_hba;
 			sas_dev->sas_device = device;
 			sas_dev->dq = dq;
+			spin_lock_init(&sas_dev->lock);
 			INIT_LIST_HEAD(&hisi_hba->devices[i].list);
 			break;
 		}
@@ -778,7 +859,8 @@ static void hisi_sas_phyup_work(struct work_struct *work)
 	struct asd_sas_phy *sas_phy = &phy->sas_phy;
 	int phy_no = sas_phy->id;
 
-	hisi_hba->hw->sl_notify(hisi_hba, phy_no); /* This requires a sleep */
+	if (phy->identify.target_port_protocols == SAS_PROTOCOL_SSP)
+		hisi_hba->hw->sl_notify_ssp(hisi_hba, phy_no);
 	hisi_sas_bytes_dmaed(hisi_hba, phy_no);
 }
 
@@ -808,6 +890,30 @@ bool hisi_sas_notify_phy_event(struct hisi_sas_phy *phy,
 }
 EXPORT_SYMBOL_GPL(hisi_sas_notify_phy_event);
 
+static void hisi_sas_wait_phyup_timedout(struct timer_list *t)
+{
+	struct hisi_sas_phy *phy = from_timer(phy, t, timer);
+	struct hisi_hba *hisi_hba = phy->hisi_hba;
+	struct device *dev = hisi_hba->dev;
+	int phy_no = phy->sas_phy.id;
+
+	dev_warn(dev, "phy%d wait phyup timeout, issuing link reset\n", phy_no);
+	hisi_sas_notify_phy_event(phy, HISI_PHYE_LINK_RESET);
+}
+
+void hisi_sas_phy_oob_ready(struct hisi_hba *hisi_hba, int phy_no)
+{
+	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
+	struct device *dev = hisi_hba->dev;
+
+	if (!timer_pending(&phy->timer)) {
+		dev_dbg(dev, "phy%d OOB ready\n", phy_no);
+		phy->timer.expires = jiffies + HISI_SAS_WAIT_PHYUP_TIMEOUT * HZ;
+		add_timer(&phy->timer);
+	}
+}
+EXPORT_SYMBOL_GPL(hisi_sas_phy_oob_ready);
+
 static void hisi_sas_phy_init(struct hisi_hba *hisi_hba, int phy_no)
 {
 	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
@@ -836,6 +942,8 @@ static void hisi_sas_phy_init(struct hisi_hba *hisi_hba, int phy_no)
 		INIT_WORK(&phy->works[i], hisi_sas_phye_fns[i]);
 
 	spin_lock_init(&phy->lock);
+
+	timer_setup(&phy->timer, hisi_sas_wait_phyup_timedout, 0);
 }
 
 static void hisi_sas_port_notify_formed(struct asd_sas_phy *sas_phy)
@@ -926,7 +1034,7 @@ static void hisi_sas_dev_gone(struct domain_device *device)
 
 	if (!test_bit(HISI_SAS_RESET_BIT, &hisi_hba->flags)) {
 		hisi_sas_internal_task_abort(hisi_hba, device,
-				     HISI_SAS_INT_ABT_DEV, 0);
+					     HISI_SAS_INT_ABT_DEV, 0);
 
 		hisi_sas_dereg_device(hisi_hba, device);
 
@@ -946,7 +1054,7 @@ static int hisi_sas_queue_command(struct sas_task *task, gfp_t gfp_flags)
 	return hisi_sas_task_exec(task, gfp_flags, 0, NULL);
 }
 
-static void hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,
+static int hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,
 			struct sas_phy_linkrates *r)
 {
 	struct sas_phy_linkrates _r;
@@ -955,6 +1063,9 @@ static void hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,
 	struct asd_sas_phy *sas_phy = &phy->sas_phy;
 	enum sas_linkrate min, max;
 
+	if (r->minimum_linkrate > SAS_LINK_RATE_1_5_GBPS)
+		return -EINVAL;
+
 	if (r->maximum_linkrate == SAS_LINK_RATE_UNKNOWN) {
 		max = sas_phy->phy->maximum_linkrate;
 		min = r->minimum_linkrate;
@@ -962,7 +1073,7 @@ static void hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,
 		max = r->maximum_linkrate;
 		min = sas_phy->phy->minimum_linkrate;
 	} else
-		return;
+		return -EINVAL;
 
 	_r.maximum_linkrate = max;
 	_r.minimum_linkrate = min;
@@ -974,6 +1085,8 @@ static void hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,
 	msleep(100);
 	hisi_hba->hw->phy_set_linkrate(hisi_hba, phy_no, &_r);
 	hisi_hba->hw->phy_start(hisi_hba, phy_no);
+
+	return 0;
 }
 
 static int hisi_sas_control_phy(struct asd_sas_phy *sas_phy, enum phy_func func,
@@ -999,8 +1112,7 @@ static int hisi_sas_control_phy(struct asd_sas_phy *sas_phy, enum phy_func func,
 		break;
 
 	case PHY_FUNC_SET_LINK_RATE:
-		hisi_sas_phy_set_linkrate(hisi_hba, phy_no, funcdata);
-		break;
+		return hisi_sas_phy_set_linkrate(hisi_hba, phy_no, funcdata);
 	case PHY_FUNC_GET_EVENTS:
 		if (hisi_hba->hw->get_events) {
 			hisi_hba->hw->get_events(hisi_hba, phy_no);
@@ -1068,7 +1180,7 @@ static int hisi_sas_exec_internal_tmf_task(struct domain_device *device,
 		task->task_done = hisi_sas_task_done;
 
 		task->slow_task->timer.function = hisi_sas_tmf_timedout;
-		task->slow_task->timer.expires = jiffies + TASK_TIMEOUT*HZ;
+		task->slow_task->timer.expires = jiffies + TASK_TIMEOUT * HZ;
 		add_timer(&task->slow_task->timer);
 
 		res = hisi_sas_task_exec(task, GFP_KERNEL, 1, tmf);
@@ -1429,6 +1541,9 @@ static int hisi_sas_controller_reset(struct hisi_hba *hisi_hba)
 	struct Scsi_Host *shost = hisi_hba->shost;
 	int rc;
 
+	if (hisi_sas_debugfs_enable && hisi_hba->debugfs_itct)
+		queue_work(hisi_hba->wq, &hisi_hba->debugfs_work);
+
 	if (!hisi_hba->hw->soft_reset)
 		return -1;
 
@@ -1491,7 +1606,6 @@ static int hisi_sas_abort_task(struct sas_task *task)
 	task->task_state_flags |= SAS_TASK_STATE_ABORTED;
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
 
-	sas_dev->dev_status = HISI_SAS_DEV_EH;
 	if (task->lldd_task && task->task_proto & SAS_PROTOCOL_SSP) {
 		struct scsi_cmnd *cmnd = task->uldd_task;
 		struct hisi_sas_slot *slot = task->lldd_task;
@@ -1527,7 +1641,8 @@ static int hisi_sas_abort_task(struct sas_task *task)
 		task->task_proto & SAS_PROTOCOL_STP) {
 		if (task->dev->dev_type == SAS_SATA_DEV) {
 			rc = hisi_sas_internal_task_abort(hisi_hba, device,
-						HISI_SAS_INT_ABT_DEV, 0);
+							  HISI_SAS_INT_ABT_DEV,
+							  0);
 			if (rc < 0) {
 				dev_err(dev, "abort task: internal abort failed\n");
 				goto out;
@@ -1542,7 +1657,7 @@ static int hisi_sas_abort_task(struct sas_task *task)
 		struct hisi_sas_cq *cq = &hisi_hba->cq[slot->dlvry_queue];
 
 		rc = hisi_sas_internal_task_abort(hisi_hba, device,
-			     HISI_SAS_INT_ABT_CMD, tag);
+						  HISI_SAS_INT_ABT_CMD, tag);
 		if (((rc < 0) || (rc == TMF_RESP_FUNC_FAILED)) &&
 					task->lldd_task) {
 			/*
@@ -1568,7 +1683,7 @@ static int hisi_sas_abort_task_set(struct domain_device *device, u8 *lun)
 	int rc = TMF_RESP_FUNC_FAILED;
 
 	rc = hisi_sas_internal_task_abort(hisi_hba, device,
-					HISI_SAS_INT_ABT_DEV, 0);
+					  HISI_SAS_INT_ABT_DEV, 0);
 	if (rc < 0) {
 		dev_err(dev, "abort task set: internal abort rc=%d\n", rc);
 		return TMF_RESP_FUNC_FAILED;
@@ -1586,8 +1701,8 @@ static int hisi_sas_abort_task_set(struct domain_device *device, u8 *lun)
 
 static int hisi_sas_clear_aca(struct domain_device *device, u8 *lun)
 {
-	int rc = TMF_RESP_FUNC_FAILED;
 	struct hisi_sas_tmf_task tmf_task;
+	int rc;
 
 	tmf_task.tmf = TMF_CLEAR_ACA;
 	rc = hisi_sas_debug_issue_ssp_tmf(device, lun, &tmf_task);
@@ -1635,17 +1750,12 @@ static int hisi_sas_debug_I_T_nexus_reset(struct domain_device *device)
 
 static int hisi_sas_I_T_nexus_reset(struct domain_device *device)
 {
-	struct hisi_sas_device *sas_dev = device->lldd_dev;
 	struct hisi_hba *hisi_hba = dev_to_hisi_hba(device);
 	struct device *dev = hisi_hba->dev;
-	int rc = TMF_RESP_FUNC_FAILED;
-
-	if (sas_dev->dev_status != HISI_SAS_DEV_EH)
-		return TMF_RESP_FUNC_FAILED;
-	sas_dev->dev_status = HISI_SAS_DEV_NORMAL;
+	int rc;
 
 	rc = hisi_sas_internal_task_abort(hisi_hba, device,
-					HISI_SAS_INT_ABT_DEV, 0);
+					  HISI_SAS_INT_ABT_DEV, 0);
 	if (rc < 0) {
 		dev_err(dev, "I_T nexus reset: internal abort (%d)\n", rc);
 		return TMF_RESP_FUNC_FAILED;
@@ -1667,7 +1777,6 @@ static int hisi_sas_lu_reset(struct domain_device *device, u8 *lun)
 	struct device *dev = hisi_hba->dev;
 	int rc = TMF_RESP_FUNC_FAILED;
 
-	sas_dev->dev_status = HISI_SAS_DEV_EH;
 	if (dev_is_sata(device)) {
 		struct sas_phy *phy;
 
@@ -1691,7 +1800,7 @@ static int hisi_sas_lu_reset(struct domain_device *device, u8 *lun)
 		struct hisi_sas_tmf_task tmf_task = { .tmf =  TMF_LU_RESET };
 
 		rc = hisi_sas_internal_task_abort(hisi_hba, device,
-						HISI_SAS_INT_ABT_DEV, 0);
+						  HISI_SAS_INT_ABT_DEV, 0);
 		if (rc < 0) {
 			dev_err(dev, "lu_reset: internal abort failed\n");
 			goto out;
@@ -1777,7 +1886,7 @@ static int hisi_sas_query_task(struct sas_task *task)
 static int
 hisi_sas_internal_abort_task_exec(struct hisi_hba *hisi_hba, int device_id,
 				  struct sas_task *task, int abort_flag,
-				  int task_tag)
+				  int task_tag, struct hisi_sas_dq *dq)
 {
 	struct domain_device *device = task->dev;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
@@ -1786,7 +1895,6 @@ hisi_sas_internal_abort_task_exec(struct hisi_hba *hisi_hba, int device_id,
 	struct hisi_sas_slot *slot;
 	struct asd_sas_port *sas_port = device->port;
 	struct hisi_sas_cmd_hdr *cmd_hdr_base;
-	struct hisi_sas_dq *dq = sas_dev->dq;
 	int dlvry_queue_slot, dlvry_queue, n_elem = 0, rc, slot_idx;
 	unsigned long flags, flags_dq = 0;
 	int wr_q_index;
@@ -1816,10 +1924,14 @@ hisi_sas_internal_abort_task_exec(struct hisi_hba *hisi_hba, int device_id,
 	}
 	list_add_tail(&slot->delivery, &dq->list);
 	spin_unlock_irqrestore(&dq->lock, flags_dq);
+	spin_lock_irqsave(&sas_dev->lock, flags);
+	list_add_tail(&slot->entry, &sas_dev->list);
+	spin_unlock_irqrestore(&sas_dev->lock, flags);
 
 	dlvry_queue = dq->id;
 	dlvry_queue_slot = wr_q_index;
 
+	slot->device_id = sas_dev->device_id;
 	slot->n_elem = n_elem;
 	slot->dlvry_queue = dlvry_queue;
 	slot->dlvry_queue_slot = dlvry_queue_slot;
@@ -1843,7 +1955,6 @@ hisi_sas_internal_abort_task_exec(struct hisi_hba *hisi_hba, int device_id,
 	WRITE_ONCE(slot->ready, 1);
 	/* send abort command to the chip */
 	spin_lock_irqsave(&dq->lock, flags);
-	list_add_tail(&slot->entry, &sas_dev->list);
 	hisi_hba->hw->start_delivery(dq);
 	spin_unlock_irqrestore(&dq->lock, flags);
 
@@ -1858,18 +1969,19 @@ err_out:
 }
 
 /**
- * hisi_sas_internal_task_abort -- execute an internal
+ * _hisi_sas_internal_task_abort -- execute an internal
  * abort command for single IO command or a device
  * @hisi_hba: host controller struct
  * @device: domain device
  * @abort_flag: mode of operation, device or single IO
  * @tag: tag of IO to be aborted (only relevant to single
  *       IO mode)
+ * @dq: delivery queue for this internal abort command
  */
 static int
-hisi_sas_internal_task_abort(struct hisi_hba *hisi_hba,
-			     struct domain_device *device,
-			     int abort_flag, int tag)
+_hisi_sas_internal_task_abort(struct hisi_hba *hisi_hba,
+			      struct domain_device *device, int abort_flag,
+			      int tag, struct hisi_sas_dq *dq)
 {
 	struct sas_task *task;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
@@ -1893,11 +2005,11 @@ hisi_sas_internal_task_abort(struct hisi_hba *hisi_hba,
 	task->task_proto = device->tproto;
 	task->task_done = hisi_sas_task_done;
 	task->slow_task->timer.function = hisi_sas_tmf_timedout;
-	task->slow_task->timer.expires = jiffies + INTERNAL_ABORT_TIMEOUT*HZ;
+	task->slow_task->timer.expires = jiffies + INTERNAL_ABORT_TIMEOUT * HZ;
 	add_timer(&task->slow_task->timer);
 
 	res = hisi_sas_internal_abort_task_exec(hisi_hba, sas_dev->device_id,
-						task, abort_flag, tag);
+						task, abort_flag, tag, dq);
 	if (res) {
 		del_timer(&task->slow_task->timer);
 		dev_err(dev, "internal task abort: executing internal task failed: %d\n",
@@ -1923,6 +2035,7 @@ hisi_sas_internal_task_abort(struct hisi_hba *hisi_hba,
 				slot->task = NULL;
 			}
 			dev_err(dev, "internal task abort: timeout and not done.\n");
+
 			res = -EIO;
 			goto exit;
 		} else
@@ -1953,6 +2066,46 @@ exit:
 	return res;
 }
 
+static int
+hisi_sas_internal_task_abort(struct hisi_hba *hisi_hba,
+			     struct domain_device *device,
+			     int abort_flag, int tag)
+{
+	struct hisi_sas_slot *slot;
+	struct device *dev = hisi_hba->dev;
+	struct hisi_sas_dq *dq;
+	int i, rc;
+
+	switch (abort_flag) {
+	case HISI_SAS_INT_ABT_CMD:
+		slot = &hisi_hba->slot_info[tag];
+		dq = &hisi_hba->dq[slot->dlvry_queue];
+		return _hisi_sas_internal_task_abort(hisi_hba, device,
+						     abort_flag, tag, dq);
+	case HISI_SAS_INT_ABT_DEV:
+		for (i = 0; i < hisi_hba->cq_nvecs; i++) {
+			struct hisi_sas_cq *cq = &hisi_hba->cq[i];
+			const struct cpumask *mask = cq->pci_irq_mask;
+
+			if (mask && !cpumask_intersects(cpu_online_mask, mask))
+				continue;
+			dq = &hisi_hba->dq[i];
+			rc = _hisi_sas_internal_task_abort(hisi_hba, device,
+							   abort_flag, tag,
+							   dq);
+			if (rc)
+				return rc;
+		}
+		break;
+	default:
+		dev_err(dev, "Unrecognised internal abort flag (%d)\n",
+			abort_flag);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static void hisi_sas_port_formed(struct asd_sas_phy *sas_phy)
 {
 	hisi_sas_port_notify_formed(sas_phy);
@@ -2019,7 +2172,7 @@ void hisi_sas_kill_tasklets(struct hisi_hba *hisi_hba)
 {
 	int i;
 
-	for (i = 0; i < hisi_hba->queue_count; i++) {
+	for (i = 0; i < hisi_hba->cq_nvecs; i++) {
 		struct hisi_sas_cq *cq = &hisi_hba->cq[i];
 
 		tasklet_kill(&cq->tasklet);
@@ -2048,14 +2201,18 @@ static struct sas_domain_function_template hisi_sas_transport_ops = {
 
 void hisi_sas_init_mem(struct hisi_hba *hisi_hba)
 {
-	int i, s, max_command_entries = hisi_hba->hw->max_command_entries;
+	int i, s, j, max_command_entries = hisi_hba->hw->max_command_entries;
+	struct hisi_sas_breakpoint *sata_breakpoint = hisi_hba->sata_breakpoint;
 
 	for (i = 0; i < hisi_hba->queue_count; i++) {
 		struct hisi_sas_cq *cq = &hisi_hba->cq[i];
 		struct hisi_sas_dq *dq = &hisi_hba->dq[i];
+		struct hisi_sas_cmd_hdr *cmd_hdr = hisi_hba->cmd_hdr[i];
+
+		s = sizeof(struct hisi_sas_cmd_hdr);
+		for (j = 0; j < HISI_SAS_QUEUE_SLOTS; j++)
+			memset(&cmd_hdr[j], 0, s);
 
-		s = sizeof(struct hisi_sas_cmd_hdr) * HISI_SAS_QUEUE_SLOTS;
-		memset(hisi_hba->cmd_hdr[i], 0, s);
 		dq->wr_point = 0;
 
 		s = hisi_hba->hw->complete_hdr_size * HISI_SAS_QUEUE_SLOTS;
@@ -2072,12 +2229,13 @@ void hisi_sas_init_mem(struct hisi_hba *hisi_hba)
 	s = max_command_entries * sizeof(struct hisi_sas_breakpoint);
 	memset(hisi_hba->breakpoint, 0, s);
 
-	s = HISI_SAS_MAX_ITCT_ENTRIES * sizeof(struct hisi_sas_sata_breakpoint);
-	memset(hisi_hba->sata_breakpoint, 0, s);
+	s = sizeof(struct hisi_sas_sata_breakpoint);
+	for (j = 0; j < HISI_SAS_MAX_ITCT_ENTRIES; j++)
+		memset(&sata_breakpoint[j], 0, s);
 }
 EXPORT_SYMBOL_GPL(hisi_sas_init_mem);
 
-int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost)
+int hisi_sas_alloc(struct hisi_hba *hisi_hba)
 {
 	struct device *dev = hisi_hba->dev;
 	int i, j, s, max_command_entries = hisi_hba->hw->max_command_entries;
@@ -2095,7 +2253,6 @@ int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost)
 	for (i = 0; i < HISI_SAS_MAX_DEVICES; i++) {
 		hisi_hba->devices[i].dev_type = SAS_PHY_UNUSED;
 		hisi_hba->devices[i].device_id = i;
-		hisi_hba->devices[i].dev_status = HISI_SAS_DEV_NORMAL;
 	}
 
 	for (i = 0; i < hisi_hba->queue_count; i++) {
@@ -2131,10 +2288,9 @@ int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost)
 
 	s = HISI_SAS_MAX_ITCT_ENTRIES * sizeof(struct hisi_sas_itct);
 	hisi_hba->itct = dmam_alloc_coherent(dev, s, &hisi_hba->itct_dma,
-					     GFP_KERNEL);
+					     GFP_KERNEL | __GFP_ZERO);
 	if (!hisi_hba->itct)
 		goto err_out;
-	memset(hisi_hba->itct, 0, s);
 
 	hisi_hba->slot_info = devm_kcalloc(dev, max_command_entries,
 					   sizeof(struct hisi_sas_slot),
@@ -2144,19 +2300,24 @@ int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost)
 
 	/* roundup to avoid overly large block size */
 	max_command_entries_ru = roundup(max_command_entries, 64);
-	sz_slot_buf_ru = roundup(sizeof(struct hisi_sas_slot_buf_table), 64);
+	if (hisi_hba->prot_mask & HISI_SAS_DIX_PROT_MASK)
+		sz_slot_buf_ru = sizeof(struct hisi_sas_slot_dif_buf_table);
+	else
+		sz_slot_buf_ru = sizeof(struct hisi_sas_slot_buf_table);
+	sz_slot_buf_ru = roundup(sz_slot_buf_ru, 64);
 	s = lcm(max_command_entries_ru, sz_slot_buf_ru);
 	blk_cnt = (max_command_entries_ru * sz_slot_buf_ru) / s;
 	slots_per_blk = s / sz_slot_buf_ru;
+
 	for (i = 0; i < blk_cnt; i++) {
-		struct hisi_sas_slot_buf_table *buf;
-		dma_addr_t buf_dma;
 		int slot_index = i * slots_per_blk;
+		dma_addr_t buf_dma;
+		void *buf;
 
-		buf = dmam_alloc_coherent(dev, s, &buf_dma, GFP_KERNEL);
+		buf = dmam_alloc_coherent(dev, s, &buf_dma,
+					  GFP_KERNEL | __GFP_ZERO);
 		if (!buf)
 			goto err_out;
-		memset(buf, 0, s);
 
 		for (j = 0; j < slots_per_blk; j++, slot_index++) {
 			struct hisi_sas_slot *slot;
@@ -2166,8 +2327,8 @@ int hisi_sas_alloc(struct hisi_hba *hisi_hba, struct Scsi_Host *shost)
 			slot->buf_dma = buf_dma;
 			slot->idx = slot_index;
 
-			buf++;
-			buf_dma += sizeof(*buf);
+			buf += sz_slot_buf_ru;
+			buf_dma += sz_slot_buf_ru;
 		}
 	}
 
@@ -2365,7 +2526,7 @@ static struct Scsi_Host *hisi_sas_shost_alloc(struct platform_device *pdev,
 			goto err_out;
 	}
 
-	if (hisi_sas_alloc(hisi_hba, shost)) {
+	if (hisi_sas_alloc(hisi_hba)) {
 		hisi_sas_free(hisi_hba);
 		goto err_out;
 	}
@@ -2461,6 +2622,555 @@ err_out_ha:
 }
 EXPORT_SYMBOL_GPL(hisi_sas_probe);
 
+struct dentry *hisi_sas_debugfs_dir;
+
+static void hisi_sas_debugfs_snapshot_cq_reg(struct hisi_hba *hisi_hba)
+{
+	int queue_entry_size = hisi_hba->hw->complete_hdr_size;
+	int i;
+
+	for (i = 0; i < hisi_hba->queue_count; i++)
+		memcpy(hisi_hba->debugfs_complete_hdr[i],
+		       hisi_hba->complete_hdr[i],
+		       HISI_SAS_QUEUE_SLOTS * queue_entry_size);
+}
+
+static void hisi_sas_debugfs_snapshot_dq_reg(struct hisi_hba *hisi_hba)
+{
+	int queue_entry_size = sizeof(struct hisi_sas_cmd_hdr);
+	int i;
+
+	for (i = 0; i < hisi_hba->queue_count; i++) {
+		struct hisi_sas_cmd_hdr	*debugfs_cmd_hdr, *cmd_hdr;
+		int j;
+
+		debugfs_cmd_hdr = hisi_hba->debugfs_cmd_hdr[i];
+		cmd_hdr = hisi_hba->cmd_hdr[i];
+
+		for (j = 0; j < HISI_SAS_QUEUE_SLOTS; j++)
+			memcpy(&debugfs_cmd_hdr[j], &cmd_hdr[j],
+			       queue_entry_size);
+	}
+}
+
+static void hisi_sas_debugfs_snapshot_port_reg(struct hisi_hba *hisi_hba)
+{
+	const struct hisi_sas_debugfs_reg *port =
+		hisi_hba->hw->debugfs_reg_port;
+	int i, phy_cnt;
+	u32 offset;
+	u32 *databuf;
+
+	for (phy_cnt = 0; phy_cnt < hisi_hba->n_phy; phy_cnt++) {
+		databuf = (u32 *)hisi_hba->debugfs_port_reg[phy_cnt];
+		for (i = 0; i < port->count; i++, databuf++) {
+			offset = port->base_off + 4 * i;
+			*databuf = port->read_port_reg(hisi_hba, phy_cnt,
+						       offset);
+		}
+	}
+}
+
+static void hisi_sas_debugfs_snapshot_global_reg(struct hisi_hba *hisi_hba)
+{
+	u32 *databuf = (u32 *)hisi_hba->debugfs_global_reg;
+	const struct hisi_sas_debugfs_reg *global =
+		hisi_hba->hw->debugfs_reg_global;
+	int i;
+
+	for (i = 0; i < global->count; i++, databuf++)
+		*databuf = global->read_global_reg(hisi_hba, 4 * i);
+}
+
+static void hisi_sas_debugfs_snapshot_itct_reg(struct hisi_hba *hisi_hba)
+{
+	void *databuf = hisi_hba->debugfs_itct;
+	struct hisi_sas_itct *itct;
+	int i;
+
+	itct = hisi_hba->itct;
+
+	for (i = 0; i < HISI_SAS_MAX_ITCT_ENTRIES; i++, itct++) {
+		memcpy(databuf, itct, sizeof(struct hisi_sas_itct));
+		databuf += sizeof(struct hisi_sas_itct);
+	}
+}
+
+static void hisi_sas_debugfs_snapshot_iost_reg(struct hisi_hba *hisi_hba)
+{
+	int max_command_entries = hisi_hba->hw->max_command_entries;
+	void *databuf = hisi_hba->debugfs_iost;
+	struct hisi_sas_iost *iost;
+	int i;
+
+	iost = hisi_hba->iost;
+
+	for (i = 0; i < max_command_entries; i++, iost++) {
+		memcpy(databuf, iost, sizeof(struct hisi_sas_iost));
+		databuf += sizeof(struct hisi_sas_iost);
+	}
+}
+
+static const char *
+hisi_sas_debugfs_to_reg_name(int off, int base_off,
+			     const struct hisi_sas_debugfs_reg_lu *lu)
+{
+	for (; lu->name; lu++) {
+		if (off == lu->off - base_off)
+			return lu->name;
+	}
+
+	return NULL;
+}
+
+static void hisi_sas_debugfs_print_reg(u32 *regs_val, const void *ptr,
+				       struct seq_file *s)
+{
+	const struct hisi_sas_debugfs_reg *reg = ptr;
+	int i;
+
+	for (i = 0; i < reg->count; i++) {
+		int off = i * 4;
+		const char *name;
+
+		name = hisi_sas_debugfs_to_reg_name(off, reg->base_off,
+						    reg->lu);
+
+		if (name)
+			seq_printf(s, "0x%08x 0x%08x %s\n", off,
+				   regs_val[i], name);
+		else
+			seq_printf(s, "0x%08x 0x%08x\n", off,
+				   regs_val[i]);
+	}
+}
+
+static int hisi_sas_debugfs_global_show(struct seq_file *s, void *p)
+{
+	struct hisi_hba *hisi_hba = s->private;
+	const struct hisi_sas_hw *hw = hisi_hba->hw;
+	const struct hisi_sas_debugfs_reg *reg_global = hw->debugfs_reg_global;
+
+	hisi_sas_debugfs_print_reg(hisi_hba->debugfs_global_reg,
+				   reg_global, s);
+
+	return 0;
+}
+
+static int hisi_sas_debugfs_global_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_global_show,
+			   inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_global_fops = {
+	.open = hisi_sas_debugfs_global_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int hisi_sas_debugfs_port_show(struct seq_file *s, void *p)
+{
+	struct hisi_sas_phy *phy = s->private;
+	struct hisi_hba *hisi_hba = phy->hisi_hba;
+	const struct hisi_sas_hw *hw = hisi_hba->hw;
+	const struct hisi_sas_debugfs_reg *reg_port = hw->debugfs_reg_port;
+	u32 *databuf = hisi_hba->debugfs_port_reg[phy->sas_phy.id];
+
+	hisi_sas_debugfs_print_reg(databuf, reg_port, s);
+
+	return 0;
+}
+
+static int hisi_sas_debugfs_port_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_port_show, inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_port_fops = {
+	.open = hisi_sas_debugfs_port_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int hisi_sas_show_row_64(struct seq_file *s, int index,
+				int sz, __le64 *ptr)
+{
+	int i;
+
+	/* completion header size not fixed per HW version */
+	seq_printf(s, "index %04d:\n\t", index);
+	for (i = 1; i <= sz / 8; i++, ptr++) {
+		seq_printf(s, " 0x%016llx", le64_to_cpu(*ptr));
+		if (!(i % 2))
+			seq_puts(s, "\n\t");
+	}
+
+	seq_puts(s, "\n");
+
+	return 0;
+}
+
+static int hisi_sas_show_row_32(struct seq_file *s, int index,
+				int sz, __le32 *ptr)
+{
+	int i;
+
+	/* completion header size not fixed per HW version */
+	seq_printf(s, "index %04d:\n\t", index);
+	for (i = 1; i <= sz / 4; i++, ptr++) {
+		seq_printf(s, " 0x%08x", le32_to_cpu(*ptr));
+		if (!(i % 4))
+			seq_puts(s, "\n\t");
+	}
+	seq_puts(s, "\n");
+
+	return 0;
+}
+
+static int hisi_sas_cq_show_slot(struct seq_file *s, int slot, void *cq_ptr)
+{
+	struct hisi_sas_cq *cq = cq_ptr;
+	struct hisi_hba *hisi_hba = cq->hisi_hba;
+	void *complete_queue = hisi_hba->debugfs_complete_hdr[cq->id];
+	__le32 *complete_hdr = complete_queue +
+			(hisi_hba->hw->complete_hdr_size * slot);
+
+	return hisi_sas_show_row_32(s, slot,
+				hisi_hba->hw->complete_hdr_size,
+				complete_hdr);
+}
+
+static int hisi_sas_debugfs_cq_show(struct seq_file *s, void *p)
+{
+	struct hisi_sas_cq *cq = s->private;
+	int slot, ret;
+
+	for (slot = 0; slot < HISI_SAS_QUEUE_SLOTS; slot++) {
+		ret = hisi_sas_cq_show_slot(s, slot, cq);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static int hisi_sas_debugfs_cq_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_cq_show, inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_cq_fops = {
+	.open = hisi_sas_debugfs_cq_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int hisi_sas_dq_show_slot(struct seq_file *s, int slot, void *dq_ptr)
+{
+	struct hisi_sas_dq *dq = dq_ptr;
+	struct hisi_hba *hisi_hba = dq->hisi_hba;
+	void *cmd_queue = hisi_hba->debugfs_cmd_hdr[dq->id];
+	__le32 *cmd_hdr = cmd_queue +
+		sizeof(struct hisi_sas_cmd_hdr) * slot;
+
+	return hisi_sas_show_row_32(s, slot, sizeof(struct hisi_sas_cmd_hdr),
+				    cmd_hdr);
+}
+
+static int hisi_sas_debugfs_dq_show(struct seq_file *s, void *p)
+{
+	int slot, ret;
+
+	for (slot = 0; slot < HISI_SAS_QUEUE_SLOTS; slot++) {
+		ret = hisi_sas_dq_show_slot(s, slot, s->private);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static int hisi_sas_debugfs_dq_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_dq_show, inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_dq_fops = {
+	.open = hisi_sas_debugfs_dq_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int hisi_sas_debugfs_iost_show(struct seq_file *s, void *p)
+{
+	struct hisi_hba *hisi_hba = s->private;
+	struct hisi_sas_iost *debugfs_iost = hisi_hba->debugfs_iost;
+	int i, ret, max_command_entries = hisi_hba->hw->max_command_entries;
+	__le64 *iost = &debugfs_iost->qw0;
+
+	for (i = 0; i < max_command_entries; i++, debugfs_iost++) {
+		ret = hisi_sas_show_row_64(s, i, sizeof(*debugfs_iost),
+					   iost);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int hisi_sas_debugfs_iost_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_iost_show, inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_iost_fops = {
+	.open = hisi_sas_debugfs_iost_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int hisi_sas_debugfs_itct_show(struct seq_file *s, void *p)
+{
+	int i, ret;
+	struct hisi_hba *hisi_hba = s->private;
+	struct hisi_sas_itct *debugfs_itct = hisi_hba->debugfs_itct;
+	__le64 *itct = &debugfs_itct->qw0;
+
+	for (i = 0; i < HISI_SAS_MAX_ITCT_ENTRIES; i++, debugfs_itct++) {
+		ret = hisi_sas_show_row_64(s, i, sizeof(*debugfs_itct),
+					   itct);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int hisi_sas_debugfs_itct_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, hisi_sas_debugfs_itct_show, inode->i_private);
+}
+
+static const struct file_operations hisi_sas_debugfs_itct_fops = {
+	.open = hisi_sas_debugfs_itct_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static void hisi_sas_debugfs_create_files(struct hisi_hba *hisi_hba)
+{
+	struct dentry *dump_dentry;
+	struct dentry *dentry;
+	char name[256];
+	int p;
+	int c;
+	int d;
+
+	/* Create dump dir inside device dir */
+	dump_dentry = debugfs_create_dir("dump", hisi_hba->debugfs_dir);
+	hisi_hba->debugfs_dump_dentry = dump_dentry;
+
+	debugfs_create_file("global", 0400, dump_dentry, hisi_hba,
+			    &hisi_sas_debugfs_global_fops);
+
+	/* Create port dir and files */
+	dentry = debugfs_create_dir("port", dump_dentry);
+	for (p = 0; p < hisi_hba->n_phy; p++) {
+		snprintf(name, 256, "%d", p);
+
+		debugfs_create_file(name, 0400, dentry, &hisi_hba->phy[p],
+				    &hisi_sas_debugfs_port_fops);
+	}
+
+	/* Create CQ dir and files */
+	dentry = debugfs_create_dir("cq", dump_dentry);
+	for (c = 0; c < hisi_hba->queue_count; c++) {
+		snprintf(name, 256, "%d", c);
+
+		debugfs_create_file(name, 0400, dentry, &hisi_hba->cq[c],
+				    &hisi_sas_debugfs_cq_fops);
+	}
+
+	/* Create DQ dir and files */
+	dentry = debugfs_create_dir("dq", dump_dentry);
+	for (d = 0; d < hisi_hba->queue_count; d++) {
+		snprintf(name, 256, "%d", d);
+
+		debugfs_create_file(name, 0400, dentry, &hisi_hba->dq[d],
+				    &hisi_sas_debugfs_dq_fops);
+	}
+
+	debugfs_create_file("iost", 0400, dump_dentry, hisi_hba,
+			    &hisi_sas_debugfs_iost_fops);
+
+	debugfs_create_file("itct", 0400, dump_dentry, hisi_hba,
+			    &hisi_sas_debugfs_itct_fops);
+
+	return;
+}
+
+static void hisi_sas_debugfs_snapshot_regs(struct hisi_hba *hisi_hba)
+{
+	hisi_hba->hw->snapshot_prepare(hisi_hba);
+
+	hisi_sas_debugfs_snapshot_global_reg(hisi_hba);
+	hisi_sas_debugfs_snapshot_port_reg(hisi_hba);
+	hisi_sas_debugfs_snapshot_cq_reg(hisi_hba);
+	hisi_sas_debugfs_snapshot_dq_reg(hisi_hba);
+	hisi_sas_debugfs_snapshot_itct_reg(hisi_hba);
+	hisi_sas_debugfs_snapshot_iost_reg(hisi_hba);
+
+	hisi_sas_debugfs_create_files(hisi_hba);
+
+	hisi_hba->hw->snapshot_restore(hisi_hba);
+}
+
+static ssize_t hisi_sas_debugfs_trigger_dump_write(struct file *file,
+						   const char __user *user_buf,
+						   size_t count, loff_t *ppos)
+{
+	struct hisi_hba *hisi_hba = file->f_inode->i_private;
+	char buf[8];
+
+	/* A bit racy, but don't care too much since it's only debugfs */
+	if (hisi_hba->debugfs_snapshot)
+		return -EFAULT;
+
+	if (count > 8)
+		return -EFAULT;
+
+	if (copy_from_user(buf, user_buf, count))
+		return -EFAULT;
+
+	if (buf[0] != '1')
+		return -EFAULT;
+
+	queue_work(hisi_hba->wq, &hisi_hba->debugfs_work);
+
+	return count;
+}
+
+static const struct file_operations hisi_sas_debugfs_trigger_dump_fops = {
+	.write = &hisi_sas_debugfs_trigger_dump_write,
+	.owner = THIS_MODULE,
+};
+
+void hisi_sas_debugfs_work_handler(struct work_struct *work)
+{
+	struct hisi_hba *hisi_hba =
+		container_of(work, struct hisi_hba, debugfs_work);
+
+	if (hisi_hba->debugfs_snapshot)
+		return;
+	hisi_hba->debugfs_snapshot = true;
+
+	hisi_sas_debugfs_snapshot_regs(hisi_hba);
+}
+EXPORT_SYMBOL_GPL(hisi_sas_debugfs_work_handler);
+
+void hisi_sas_debugfs_init(struct hisi_hba *hisi_hba)
+{
+	int max_command_entries = hisi_hba->hw->max_command_entries;
+	struct device *dev = hisi_hba->dev;
+	int p, i, c, d;
+	size_t sz;
+
+	hisi_hba->debugfs_dir = debugfs_create_dir(dev_name(dev),
+						   hisi_sas_debugfs_dir);
+	debugfs_create_file("trigger_dump", 0600,
+			    hisi_hba->debugfs_dir,
+			    hisi_hba,
+			    &hisi_sas_debugfs_trigger_dump_fops);
+
+	/* Alloc buffer for global */
+	sz = hisi_hba->hw->debugfs_reg_global->count * 4;
+	hisi_hba->debugfs_global_reg =
+		devm_kmalloc(dev, sz, GFP_KERNEL);
+
+	if (!hisi_hba->debugfs_global_reg)
+		goto fail_global;
+
+	/* Alloc buffer for port */
+	sz = hisi_hba->hw->debugfs_reg_port->count * 4;
+	for (p = 0; p < hisi_hba->n_phy; p++) {
+		hisi_hba->debugfs_port_reg[p] =
+			devm_kmalloc(dev, sz, GFP_KERNEL);
+
+		if (!hisi_hba->debugfs_port_reg[p])
+			goto fail_port;
+	}
+
+	/* Alloc buffer for cq */
+	sz = hisi_hba->hw->complete_hdr_size * HISI_SAS_QUEUE_SLOTS;
+	for (c = 0; c < hisi_hba->queue_count; c++) {
+		hisi_hba->debugfs_complete_hdr[c] =
+			devm_kmalloc(dev, sz, GFP_KERNEL);
+
+		if (!hisi_hba->debugfs_complete_hdr[c])
+			goto fail_cq;
+	}
+
+	/* Alloc buffer for dq */
+	sz = sizeof(struct hisi_sas_cmd_hdr) * HISI_SAS_QUEUE_SLOTS;
+	for (d = 0; d < hisi_hba->queue_count; d++) {
+		hisi_hba->debugfs_cmd_hdr[d] =
+			devm_kmalloc(dev, sz, GFP_KERNEL);
+
+		if (!hisi_hba->debugfs_cmd_hdr[d])
+			goto fail_iost_dq;
+	}
+
+	/* Alloc buffer for iost */
+	sz = max_command_entries * sizeof(struct hisi_sas_iost);
+
+	hisi_hba->debugfs_iost = devm_kmalloc(dev, sz, GFP_KERNEL);
+	if (!hisi_hba->debugfs_iost)
+		goto fail_iost_dq;
+
+	/* Alloc buffer for itct */
+	/* New memory allocation must be locate before itct */
+	sz = HISI_SAS_MAX_ITCT_ENTRIES * sizeof(struct hisi_sas_itct);
+
+	hisi_hba->debugfs_itct = devm_kmalloc(dev, sz, GFP_KERNEL);
+	if (!hisi_hba->debugfs_itct)
+		goto fail_itct;
+
+	return;
+fail_itct:
+	devm_kfree(dev, hisi_hba->debugfs_iost);
+fail_iost_dq:
+	for (i = 0; i < d; i++)
+		devm_kfree(dev, hisi_hba->debugfs_cmd_hdr[i]);
+fail_cq:
+	for (i = 0; i < c; i++)
+		devm_kfree(dev, hisi_hba->debugfs_complete_hdr[i]);
+fail_port:
+	for (i = 0; i < p; i++)
+		devm_kfree(dev, hisi_hba->debugfs_port_reg[i]);
+	devm_kfree(dev, hisi_hba->debugfs_global_reg);
+fail_global:
+	debugfs_remove_recursive(hisi_hba->debugfs_dir);
+	dev_dbg(dev, "failed to init debugfs!\n");
+}
+EXPORT_SYMBOL_GPL(hisi_sas_debugfs_init);
+
+void hisi_sas_debugfs_exit(struct hisi_hba *hisi_hba)
+{
+	debugfs_remove_recursive(hisi_hba->debugfs_dir);
+}
+EXPORT_SYMBOL_GPL(hisi_sas_debugfs_exit);
+
 int hisi_sas_remove(struct platform_device *pdev)
 {
 	struct sas_ha_struct *sha = platform_get_drvdata(pdev);
@@ -2479,18 +3189,28 @@ int hisi_sas_remove(struct platform_device *pdev)
 }
 EXPORT_SYMBOL_GPL(hisi_sas_remove);
 
+bool hisi_sas_debugfs_enable;
+EXPORT_SYMBOL_GPL(hisi_sas_debugfs_enable);
+module_param_named(debugfs_enable, hisi_sas_debugfs_enable, bool, 0444);
+MODULE_PARM_DESC(hisi_sas_debugfs_enable, "Enable driver debugfs (default disabled)");
+
 static __init int hisi_sas_init(void)
 {
 	hisi_sas_stt = sas_domain_attach_transport(&hisi_sas_transport_ops);
 	if (!hisi_sas_stt)
 		return -ENOMEM;
 
+	if (hisi_sas_debugfs_enable)
+		hisi_sas_debugfs_dir = debugfs_create_dir("hisi_sas", NULL);
+
 	return 0;
 }
 
 static __exit void hisi_sas_exit(void)
 {
 	sas_release_transport(hisi_sas_stt);
+
+	debugfs_remove(hisi_sas_debugfs_dir);
 }
 
 module_init(hisi_sas_init);
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
index 28ab52a021cf..293807443480 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
@@ -835,7 +835,7 @@ static void phys_init_v1_hw(struct hisi_hba *hisi_hba)
 	mod_timer(timer, jiffies + HZ);
 }
 
-static void sl_notify_v1_hw(struct hisi_hba *hisi_hba, int phy_no)
+static void sl_notify_ssp_v1_hw(struct hisi_hba *hisi_hba, int phy_no)
 {
 	u32 sl_control;
 
@@ -1749,6 +1749,8 @@ static int interrupt_init_v1_hw(struct hisi_hba *hisi_hba)
 		}
 	}
 
+	hisi_hba->cq_nvecs = hisi_hba->queue_count;
+
 	return 0;
 }
 
@@ -1826,7 +1828,7 @@ static struct scsi_host_template sht_v1_hw = {
 static const struct hisi_sas_hw hisi_sas_v1_hw = {
 	.hw_init = hisi_sas_v1_init,
 	.setup_itct = setup_itct_v1_hw,
-	.sl_notify = sl_notify_v1_hw,
+	.sl_notify_ssp = sl_notify_ssp_v1_hw,
 	.clear_itct = clear_itct_v1_hw,
 	.prep_smp = prep_smp_v1_hw,
 	.prep_ssp = prep_ssp_v1_hw,
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index c8ebff3ba559..e40cc6b3b67b 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -868,12 +868,12 @@ hisi_sas_device *alloc_dev_quirk_v2_hw(struct domain_device *device)
 
 			hisi_hba->devices[i].device_id = i;
 			sas_dev = &hisi_hba->devices[i];
-			sas_dev->dev_status = HISI_SAS_DEV_NORMAL;
 			sas_dev->dev_type = device->dev_type;
 			sas_dev->hisi_hba = hisi_hba;
 			sas_dev->sas_device = device;
 			sas_dev->sata_idx = sata_idx;
 			sas_dev->dq = dq;
+			spin_lock_init(&sas_dev->lock);
 			INIT_LIST_HEAD(&hisi_hba->devices[i].list);
 			break;
 		}
@@ -1589,7 +1589,7 @@ static void phys_init_v2_hw(struct hisi_hba *hisi_hba)
 	}
 }
 
-static void sl_notify_v2_hw(struct hisi_hba *hisi_hba, int phy_no)
+static void sl_notify_ssp_v2_hw(struct hisi_hba *hisi_hba, int phy_no)
 {
 	u32 sl_control;
 
@@ -2677,6 +2677,8 @@ static int phy_up_v2_hw(int phy_no, struct hisi_hba *hisi_hba)
 	if (is_sata_phy_v2_hw(hisi_hba, phy_no))
 		goto end;
 
+	del_timer(&phy->timer);
+
 	if (phy_no == 8) {
 		u32 port_state = hisi_sas_read32(hisi_hba, PORT_STATE);
 
@@ -2756,6 +2758,7 @@ static int phy_down_v2_hw(int phy_no, struct hisi_hba *hisi_hba)
 	struct hisi_sas_port *port = phy->port;
 	struct device *dev = hisi_hba->dev;
 
+	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_NOT_RDY_MSK, 1);
 
 	phy_state = hisi_sas_read32(hisi_hba, PHY_STATE);
@@ -2944,6 +2947,9 @@ static irqreturn_t int_chnl_int_v2_hw(int irq_no, void *p)
 			if (irq_value0 & CHL_INT0_SL_RX_BCST_ACK_MSK)
 				phy_bcast_v2_hw(phy_no, hisi_hba);
 
+			if (irq_value0 & CHL_INT0_PHY_RDY_MSK)
+				hisi_sas_phy_oob_ready(hisi_hba, phy_no);
+
 			hisi_sas_phy_write32(hisi_hba, phy_no,
 					CHL_INT0, irq_value0
 					& (~CHL_INT0_HOTPLUG_TOUT_MSK)
@@ -3227,6 +3233,8 @@ static irqreturn_t sata_int_v2_hw(int irq_no, void *p)
 	unsigned long flags;
 	int phy_no, offset;
 
+	del_timer(&phy->timer);
+
 	phy_no = sas_phy->id;
 	initial_fis = &hisi_hba->initial_fis[phy_no];
 	fis = &initial_fis->fis;
@@ -3393,6 +3401,8 @@ static int interrupt_init_v2_hw(struct hisi_hba *hisi_hba)
 		tasklet_init(t, cq_tasklet_v2_hw, (unsigned long)cq);
 	}
 
+	hisi_hba->cq_nvecs = hisi_hba->queue_count;
+
 	return 0;
 
 free_cq_int_irqs:
@@ -3542,8 +3552,8 @@ static int write_gpio_v2_hw(struct hisi_hba *hisi_hba, u8 reg_type,
 	return 0;
 }
 
-static void wait_cmds_complete_timeout_v2_hw(struct hisi_hba *hisi_hba,
-					     int delay_ms, int timeout_ms)
+static int wait_cmds_complete_timeout_v2_hw(struct hisi_hba *hisi_hba,
+					    int delay_ms, int timeout_ms)
 {
 	struct device *dev = hisi_hba->dev;
 	int entries, entries_old = 0, time;
@@ -3557,7 +3567,12 @@ static void wait_cmds_complete_timeout_v2_hw(struct hisi_hba *hisi_hba,
 		msleep(delay_ms);
 	}
 
+	if (time >= timeout_ms)
+		return -ETIMEDOUT;
+
 	dev_dbg(dev, "wait commands complete %dms\n", time);
+
+	return 0;
 }
 
 static struct device_attribute *host_attrs_v2_hw[] = {
@@ -3590,7 +3605,7 @@ static const struct hisi_sas_hw hisi_sas_v2_hw = {
 	.setup_itct = setup_itct_v2_hw,
 	.slot_index_alloc = slot_index_alloc_quirk_v2_hw,
 	.alloc_dev = alloc_dev_quirk_v2_hw,
-	.sl_notify = sl_notify_v2_hw,
+	.sl_notify_ssp = sl_notify_ssp_v2_hw,
 	.get_wideport_bitmap = get_wideport_bitmap_v2_hw,
 	.clear_itct = clear_itct_v2_hw,
 	.free_device = free_device_v2_hw,
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index e0570fd8466e..9ec8848ec541 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -11,7 +11,7 @@
 #include "hisi_sas.h"
 #define DRV_NAME "hisi_sas_v3_hw"
 
-/* global registers need init*/
+/* global registers need init */
 #define DLVRY_QUEUE_ENABLE		0x0
 #define IOST_BASE_ADDR_LO		0x8
 #define IOST_BASE_ADDR_HI		0xc
@@ -186,6 +186,7 @@
 #define CHL_INT0_MSK			(PORT_BASE + 0x1c0)
 #define CHL_INT1_MSK			(PORT_BASE + 0x1c4)
 #define CHL_INT2_MSK			(PORT_BASE + 0x1c8)
+#define SAS_EC_INT_COAL_TIME		(PORT_BASE + 0x1cc)
 #define CHL_INT_COAL_EN			(PORT_BASE + 0x1d0)
 #define SAS_RX_TRAIN_TIMER		(PORT_BASE + 0x2a4)
 #define PHY_CTRL_RDY_MSK		(PORT_BASE + 0x2b0)
@@ -205,6 +206,7 @@
 #define ERR_CNT_DWS_LOST		(PORT_BASE + 0x380)
 #define ERR_CNT_RESET_PROB		(PORT_BASE + 0x384)
 #define ERR_CNT_INVLD_DW		(PORT_BASE + 0x390)
+#define ERR_CNT_CODE_ERR		(PORT_BASE + 0x394)
 #define ERR_CNT_DISP_ERR		(PORT_BASE + 0x398)
 
 #define DEFAULT_ITCT_HW		2048 /* reset value, not reprogrammed */
@@ -397,6 +399,11 @@ struct hisi_sas_err_record_v3 {
 #define USR_DATA_BLOCK_SZ_OFF	20
 #define USR_DATA_BLOCK_SZ_MSK	(0x3 << USR_DATA_BLOCK_SZ_OFF)
 #define T10_CHK_MSK_OFF	    16
+#define T10_CHK_REF_TAG_MSK (0xf0 << T10_CHK_MSK_OFF)
+#define T10_CHK_APP_TAG_MSK (0xc << T10_CHK_MSK_OFF)
+
+#define BASE_VECTORS_V3_HW  16
+#define MIN_AFFINE_VECTORS_V3_HW  (BASE_VECTORS_V3_HW + 1)
 
 static bool hisi_sas_intr_conv;
 MODULE_PARM_DESC(intr_conv, "interrupt converge enable (0-1)");
@@ -406,6 +413,11 @@ static int prot_mask;
 module_param(prot_mask, int, 0);
 MODULE_PARM_DESC(prot_mask, " host protection capabilities mask, def=0x0 ");
 
+static bool auto_affine_msi_experimental;
+module_param(auto_affine_msi_experimental, bool, 0444);
+MODULE_PARM_DESC(auto_affine_msi_experimental, "Enable auto-affinity of MSI IRQs as experimental:\n"
+		 "default is off");
+
 static u32 hisi_sas_read32(struct hisi_hba *hisi_hba, u32 off)
 {
 	void __iomem *regs = hisi_hba->regs + off;
@@ -716,7 +728,7 @@ static void clear_itct_v3_hw(struct hisi_hba *hisi_hba,
 		hisi_sas_write32(hisi_hba, ENT_INT_SRC3,
 				 ENT_INT_SRC3_ITC_INT_MSK);
 
-	/* clear the itct table*/
+	/* clear the itct table */
 	reg_val = ITCT_CLR_EN_MSK | (dev_id & ITCT_DEV_MSK);
 	hisi_sas_write32(hisi_hba, ITCT_CLR, reg_val);
 
@@ -868,7 +880,7 @@ static void phys_init_v3_hw(struct hisi_hba *hisi_hba)
 	}
 }
 
-static void sl_notify_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
+static void sl_notify_ssp_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
 {
 	u32 sl_control;
 
@@ -967,19 +979,44 @@ static void prep_prd_sge_v3_hw(struct hisi_hba *hisi_hba,
 
 	hdr->prd_table_addr = cpu_to_le64(hisi_sas_sge_addr_dma(slot));
 
-	hdr->sg_len = cpu_to_le32(n_elem << CMD_HDR_DATA_SGL_LEN_OFF);
+	hdr->sg_len |= cpu_to_le32(n_elem << CMD_HDR_DATA_SGL_LEN_OFF);
+}
+
+static void prep_prd_sge_dif_v3_hw(struct hisi_hba *hisi_hba,
+				   struct hisi_sas_slot *slot,
+				   struct hisi_sas_cmd_hdr *hdr,
+				   struct scatterlist *scatter,
+				   int n_elem)
+{
+	struct hisi_sas_sge_dif_page *sge_dif_page;
+	struct scatterlist *sg;
+	int i;
+
+	sge_dif_page = hisi_sas_sge_dif_addr_mem(slot);
+
+	for_each_sg(scatter, sg, n_elem, i) {
+		struct hisi_sas_sge *entry = &sge_dif_page->sge[i];
+
+		entry->addr = cpu_to_le64(sg_dma_address(sg));
+		entry->page_ctrl_0 = 0;
+		entry->page_ctrl_1 = 0;
+		entry->data_len = cpu_to_le32(sg_dma_len(sg));
+		entry->data_off = 0;
+	}
+
+	hdr->dif_prd_table_addr =
+		cpu_to_le64(hisi_sas_sge_dif_addr_dma(slot));
+
+	hdr->sg_len |= cpu_to_le32(n_elem << CMD_HDR_DIF_SGL_LEN_OFF);
 }
 
 static u32 get_prot_chk_msk_v3_hw(struct scsi_cmnd *scsi_cmnd)
 {
 	unsigned char prot_flags = scsi_cmnd->prot_flags;
 
-	if (prot_flags & SCSI_PROT_TRANSFER_PI) {
-		if (prot_flags & SCSI_PROT_REF_CHECK)
-			return 0xc << 16;
-		return 0xfc << 16;
-	}
-	return 0;
+	if (prot_flags & SCSI_PROT_REF_CHECK)
+		return T10_CHK_APP_TAG_MSK;
+	return T10_CHK_REF_TAG_MSK | T10_CHK_APP_TAG_MSK;
 }
 
 static void fill_prot_v3_hw(struct scsi_cmnd *scsi_cmnd,
@@ -990,15 +1027,33 @@ static void fill_prot_v3_hw(struct scsi_cmnd *scsi_cmnd,
 	u32 lbrt_chk_val = t10_pi_ref_tag(scsi_cmnd->request);
 
 	switch (prot_op) {
+	case SCSI_PROT_READ_INSERT:
+		prot->dw0 |= T10_INSRT_EN_MSK;
+		prot->lbrtgv = lbrt_chk_val;
+		break;
 	case SCSI_PROT_READ_STRIP:
 		prot->dw0 |= (T10_RMV_EN_MSK | T10_CHK_EN_MSK);
 		prot->lbrtcv = lbrt_chk_val;
 		prot->dw4 |= get_prot_chk_msk_v3_hw(scsi_cmnd);
 		break;
+	case SCSI_PROT_READ_PASS:
+		prot->dw0 |= T10_CHK_EN_MSK;
+		prot->lbrtcv = lbrt_chk_val;
+		prot->dw4 |= get_prot_chk_msk_v3_hw(scsi_cmnd);
+		break;
 	case SCSI_PROT_WRITE_INSERT:
 		prot->dw0 |= T10_INSRT_EN_MSK;
 		prot->lbrtgv = lbrt_chk_val;
 		break;
+	case SCSI_PROT_WRITE_STRIP:
+		prot->dw0 |= (T10_RMV_EN_MSK | T10_CHK_EN_MSK);
+		prot->lbrtcv = lbrt_chk_val;
+		break;
+	case SCSI_PROT_WRITE_PASS:
+		prot->dw0 |= T10_CHK_EN_MSK;
+		prot->lbrtcv = lbrt_chk_val;
+		prot->dw4 |= get_prot_chk_msk_v3_hw(scsi_cmnd);
+		break;
 	default:
 		WARN(1, "prot_op(0x%x) is not valid\n", prot_op);
 		break;
@@ -1033,8 +1088,8 @@ static void prep_ssp_v3_hw(struct hisi_hba *hisi_hba,
 	struct sas_ssp_task *ssp_task = &task->ssp_task;
 	struct scsi_cmnd *scsi_cmnd = ssp_task->cmd;
 	struct hisi_sas_tmf_task *tmf = slot->tmf;
-	unsigned char prot_op = scsi_get_prot_op(scsi_cmnd);
 	int has_data = 0, priority = !!tmf;
+	unsigned char prot_op;
 	u8 *buf_cmd;
 	u32 dw1 = 0, dw2 = 0, len = 0;
 
@@ -1049,6 +1104,7 @@ static void prep_ssp_v3_hw(struct hisi_hba *hisi_hba,
 		dw1 |= 2 << CMD_HDR_FRAME_TYPE_OFF;
 		dw1 |= DIR_NO_DATA << CMD_HDR_DIR_OFF;
 	} else {
+		prot_op = scsi_get_prot_op(scsi_cmnd);
 		dw1 |= 1 << CMD_HDR_FRAME_TYPE_OFF;
 		switch (scsi_cmnd->sc_data_direction) {
 		case DMA_TO_DEVICE:
@@ -1074,9 +1130,15 @@ static void prep_ssp_v3_hw(struct hisi_hba *hisi_hba,
 	hdr->dw2 = cpu_to_le32(dw2);
 	hdr->transfer_tags = cpu_to_le32(slot->idx);
 
-	if (has_data)
+	if (has_data) {
 		prep_prd_sge_v3_hw(hisi_hba, slot, hdr, task->scatter,
-					slot->n_elem);
+				   slot->n_elem);
+
+		if (scsi_prot_sg_count(scsi_cmnd))
+			prep_prd_sge_dif_v3_hw(hisi_hba, slot, hdr,
+					       scsi_prot_sglist(scsi_cmnd),
+					       slot->n_elem_dif);
+	}
 
 	hdr->cmd_table_addr = cpu_to_le64(hisi_sas_cmd_hdr_addr_dma(slot));
 	hdr->sts_buffer_addr = cpu_to_le64(hisi_sas_status_buf_addr_dma(slot));
@@ -1117,18 +1179,19 @@ static void prep_ssp_v3_hw(struct hisi_hba *hisi_hba,
 		fill_prot_v3_hw(scsi_cmnd, &prot);
 		memcpy(buf_cmd_prot, &prot,
 		       sizeof(struct hisi_sas_protect_iu_v3_hw));
-
 		/*
 		 * For READ, we need length of info read to memory, while for
 		 * WRITE we need length of data written to the disk.
 		 */
-		if (prot_op == SCSI_PROT_WRITE_INSERT) {
+		if (prot_op == SCSI_PROT_WRITE_INSERT ||
+		    prot_op == SCSI_PROT_READ_INSERT ||
+		    prot_op == SCSI_PROT_WRITE_PASS ||
+		    prot_op == SCSI_PROT_READ_PASS) {
 			unsigned int interval = scsi_prot_interval(scsi_cmnd);
 			unsigned int ilog2_interval = ilog2(interval);
 
 			len = (task->total_xfer_len >> ilog2_interval) * 8;
 		}
-
 	}
 
 	hdr->dw1 = cpu_to_le32(dw1);
@@ -1288,6 +1351,7 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 	struct device *dev = hisi_hba->dev;
 	unsigned long flags;
 
+	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_PHY_ENA_MSK, 1);
 
 	port_id = hisi_sas_read32(hisi_hba, PHY_PORT_NUM_MA);
@@ -1381,9 +1445,11 @@ end:
 
 static irqreturn_t phy_down_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 {
+	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
 	u32 phy_state, sl_ctrl, txid_auto;
 	struct device *dev = hisi_hba->dev;
 
+	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_NOT_RDY_MSK, 1);
 
 	phy_state = hisi_sas_read32(hisi_hba, PHY_STATE);
@@ -1552,6 +1618,19 @@ static void handle_chl_int2_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
 	hisi_sas_phy_write32(hisi_hba, phy_no, CHL_INT2, irq_value);
 }
 
+static void handle_chl_int0_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
+{
+	u32 irq_value0 = hisi_sas_phy_read32(hisi_hba, phy_no, CHL_INT0);
+
+	if (irq_value0 & CHL_INT0_PHY_RDY_MSK)
+		hisi_sas_phy_oob_ready(hisi_hba, phy_no);
+
+	hisi_sas_phy_write32(hisi_hba, phy_no, CHL_INT0,
+			     irq_value0 & (~CHL_INT0_SL_RX_BCST_ACK_MSK)
+			     & (~CHL_INT0_SL_PHY_ENABLE_MSK)
+			     & (~CHL_INT0_NOT_RDY_MSK));
+}
+
 static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 {
 	struct hisi_hba *hisi_hba = p;
@@ -1562,8 +1641,8 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 				& 0xeeeeeeee;
 
 	while (irq_msk) {
-		u32 irq_value0 = hisi_sas_phy_read32(hisi_hba, phy_no,
-						     CHL_INT0);
+		if (irq_msk & (2 << (phy_no * 4)))
+			handle_chl_int0_v3_hw(hisi_hba, phy_no);
 
 		if (irq_msk & (4 << (phy_no * 4)))
 			handle_chl_int1_v3_hw(hisi_hba, phy_no);
@@ -1571,13 +1650,6 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 		if (irq_msk & (8 << (phy_no * 4)))
 			handle_chl_int2_v3_hw(hisi_hba, phy_no);
 
-		if (irq_msk & (2 << (phy_no * 4)) && irq_value0) {
-			hisi_sas_phy_write32(hisi_hba, phy_no,
-					CHL_INT0, irq_value0
-					& (~CHL_INT0_SL_RX_BCST_ACK_MSK)
-					& (~CHL_INT0_SL_PHY_ENABLE_MSK)
-					& (~CHL_INT0_NOT_RDY_MSK));
-		}
 		irq_msk &= ~(0xe << (phy_no * 4));
 		phy_no++;
 	}
@@ -1644,6 +1716,7 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 	u32 irq_value, irq_msk;
 	struct hisi_hba *hisi_hba = p;
 	struct device *dev = hisi_hba->dev;
+	struct pci_dev *pdev = hisi_hba->pci_dev;
 	int i;
 
 	irq_msk = hisi_sas_read32(hisi_hba, ENT_INT_SRC_MSK3);
@@ -1675,6 +1748,17 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 				error->msg, irq_value);
 			queue_work(hisi_hba->wq, &hisi_hba->rst_work);
 		}
+
+		if (pdev->revision < 0x21) {
+			u32 reg_val;
+
+			reg_val = hisi_sas_read32(hisi_hba,
+						  AXI_MASTER_CFG_BASE +
+						  AM_CTRL_GLOBAL);
+			reg_val |= AM_CTRL_SHUTDOWN_REQ_MSK;
+			hisi_sas_write32(hisi_hba, AXI_MASTER_CFG_BASE +
+					 AM_CTRL_GLOBAL, reg_val);
+		}
 	}
 
 	if (irq_value & BIT(ENT_INT_SRC3_ITC_INT_OFF)) {
@@ -1959,21 +2043,68 @@ static irqreturn_t cq_interrupt_v3_hw(int irq_no, void *p)
 	return IRQ_HANDLED;
 }
 
+static void setup_reply_map_v3_hw(struct hisi_hba *hisi_hba, int nvecs)
+{
+	const struct cpumask *mask;
+	int queue, cpu;
+
+	for (queue = 0; queue < nvecs; queue++) {
+		struct hisi_sas_cq *cq = &hisi_hba->cq[queue];
+
+		mask = pci_irq_get_affinity(hisi_hba->pci_dev, queue +
+					    BASE_VECTORS_V3_HW);
+		if (!mask)
+			goto fallback;
+		cq->pci_irq_mask = mask;
+		for_each_cpu(cpu, mask)
+			hisi_hba->reply_map[cpu] = queue;
+	}
+	return;
+
+fallback:
+	for_each_possible_cpu(cpu)
+		hisi_hba->reply_map[cpu] = cpu % hisi_hba->queue_count;
+	/* Don't clean all CQ masks */
+}
+
 static int interrupt_init_v3_hw(struct hisi_hba *hisi_hba)
 {
 	struct device *dev = hisi_hba->dev;
 	struct pci_dev *pdev = hisi_hba->pci_dev;
 	int vectors, rc;
 	int i, k;
-	int max_msi = HISI_SAS_MSI_COUNT_V3_HW;
-
-	vectors = pci_alloc_irq_vectors(hisi_hba->pci_dev, 1,
-					max_msi, PCI_IRQ_MSI);
-	if (vectors < max_msi) {
-		dev_err(dev, "could not allocate all msi (%d)\n", vectors);
-		return -ENOENT;
+	int max_msi = HISI_SAS_MSI_COUNT_V3_HW, min_msi;
+
+	if (auto_affine_msi_experimental) {
+		struct irq_affinity desc = {
+			.pre_vectors = BASE_VECTORS_V3_HW,
+		};
+
+		min_msi = MIN_AFFINE_VECTORS_V3_HW;
+
+		hisi_hba->reply_map = devm_kcalloc(dev, nr_cpu_ids,
+						   sizeof(unsigned int),
+						   GFP_KERNEL);
+		if (!hisi_hba->reply_map)
+			return -ENOMEM;
+		vectors = pci_alloc_irq_vectors_affinity(hisi_hba->pci_dev,
+							 min_msi, max_msi,
+							 PCI_IRQ_MSI |
+							 PCI_IRQ_AFFINITY,
+							 &desc);
+		if (vectors < 0)
+			return -ENOENT;
+		setup_reply_map_v3_hw(hisi_hba, vectors - BASE_VECTORS_V3_HW);
+	} else {
+		min_msi = max_msi;
+		vectors = pci_alloc_irq_vectors(hisi_hba->pci_dev, min_msi,
+						max_msi, PCI_IRQ_MSI);
+		if (vectors < 0)
+			return vectors;
 	}
 
+	hisi_hba->cq_nvecs = vectors - BASE_VECTORS_V3_HW;
+
 	rc = devm_request_irq(dev, pci_irq_vector(pdev, 1),
 			      int_phy_up_down_bcast_v3_hw, 0,
 			      DRV_NAME " phy", hisi_hba);
@@ -2002,7 +2133,7 @@ static int interrupt_init_v3_hw(struct hisi_hba *hisi_hba)
 	}
 
 	/* Init tasklets for cq only */
-	for (i = 0; i < hisi_hba->queue_count; i++) {
+	for (i = 0; i < hisi_hba->cq_nvecs; i++) {
 		struct hisi_sas_cq *cq = &hisi_hba->cq[i];
 		struct tasklet_struct *t = &cq->tasklet;
 		int nr = hisi_sas_intr_conv ? 16 : 16 + i;
@@ -2201,8 +2332,8 @@ static int write_gpio_v3_hw(struct hisi_hba *hisi_hba, u8 reg_type,
 	return 0;
 }
 
-static void wait_cmds_complete_timeout_v3_hw(struct hisi_hba *hisi_hba,
-					     int delay_ms, int timeout_ms)
+static int wait_cmds_complete_timeout_v3_hw(struct hisi_hba *hisi_hba,
+					    int delay_ms, int timeout_ms)
 {
 	struct device *dev = hisi_hba->dev;
 	int entries, entries_old = 0, time;
@@ -2216,7 +2347,12 @@ static void wait_cmds_complete_timeout_v3_hw(struct hisi_hba *hisi_hba,
 		msleep(delay_ms);
 	}
 
+	if (time >= timeout_ms)
+		return -ETIMEDOUT;
+
 	dev_dbg(dev, "wait commands complete %dms\n", time);
+
+	return 0;
 }
 
 static ssize_t intr_conv_v3_hw_show(struct device *dev,
@@ -2332,6 +2468,159 @@ static struct device_attribute *host_attrs_v3_hw[] = {
 	NULL
 };
 
+static const struct hisi_sas_debugfs_reg_lu debugfs_port_reg_lu[] = {
+	HISI_SAS_DEBUGFS_REG(PHY_CFG),
+	HISI_SAS_DEBUGFS_REG(HARD_PHY_LINKRATE),
+	HISI_SAS_DEBUGFS_REG(PROG_PHY_LINK_RATE),
+	HISI_SAS_DEBUGFS_REG(PHY_CTRL),
+	HISI_SAS_DEBUGFS_REG(SL_CFG),
+	HISI_SAS_DEBUGFS_REG(AIP_LIMIT),
+	HISI_SAS_DEBUGFS_REG(SL_CONTROL),
+	HISI_SAS_DEBUGFS_REG(RX_PRIMS_STATUS),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD0),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD1),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD2),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD3),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD4),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD5),
+	HISI_SAS_DEBUGFS_REG(TX_ID_DWORD6),
+	HISI_SAS_DEBUGFS_REG(TXID_AUTO),
+	HISI_SAS_DEBUGFS_REG(RX_IDAF_DWORD0),
+	HISI_SAS_DEBUGFS_REG(RXOP_CHECK_CFG_H),
+	HISI_SAS_DEBUGFS_REG(STP_LINK_TIMER),
+	HISI_SAS_DEBUGFS_REG(STP_LINK_TIMEOUT_STATE),
+	HISI_SAS_DEBUGFS_REG(CON_CFG_DRIVER),
+	HISI_SAS_DEBUGFS_REG(SAS_SSP_CON_TIMER_CFG),
+	HISI_SAS_DEBUGFS_REG(SAS_SMP_CON_TIMER_CFG),
+	HISI_SAS_DEBUGFS_REG(SAS_STP_CON_TIMER_CFG),
+	HISI_SAS_DEBUGFS_REG(CHL_INT0),
+	HISI_SAS_DEBUGFS_REG(CHL_INT1),
+	HISI_SAS_DEBUGFS_REG(CHL_INT2),
+	HISI_SAS_DEBUGFS_REG(CHL_INT0_MSK),
+	HISI_SAS_DEBUGFS_REG(CHL_INT1_MSK),
+	HISI_SAS_DEBUGFS_REG(CHL_INT2_MSK),
+	HISI_SAS_DEBUGFS_REG(SAS_EC_INT_COAL_TIME),
+	HISI_SAS_DEBUGFS_REG(CHL_INT_COAL_EN),
+	HISI_SAS_DEBUGFS_REG(SAS_RX_TRAIN_TIMER),
+	HISI_SAS_DEBUGFS_REG(PHY_CTRL_RDY_MSK),
+	HISI_SAS_DEBUGFS_REG(PHYCTRL_NOT_RDY_MSK),
+	HISI_SAS_DEBUGFS_REG(PHYCTRL_DWS_RESET_MSK),
+	HISI_SAS_DEBUGFS_REG(PHYCTRL_PHY_ENA_MSK),
+	HISI_SAS_DEBUGFS_REG(SL_RX_BCAST_CHK_MSK),
+	HISI_SAS_DEBUGFS_REG(PHYCTRL_OOB_RESTART_MSK),
+	HISI_SAS_DEBUGFS_REG(DMA_TX_STATUS),
+	HISI_SAS_DEBUGFS_REG(DMA_RX_STATUS),
+	HISI_SAS_DEBUGFS_REG(COARSETUNE_TIME),
+	HISI_SAS_DEBUGFS_REG(ERR_CNT_DWS_LOST),
+	HISI_SAS_DEBUGFS_REG(ERR_CNT_RESET_PROB),
+	HISI_SAS_DEBUGFS_REG(ERR_CNT_INVLD_DW),
+	HISI_SAS_DEBUGFS_REG(ERR_CNT_CODE_ERR),
+	HISI_SAS_DEBUGFS_REG(ERR_CNT_DISP_ERR),
+	{}
+};
+
+static const struct hisi_sas_debugfs_reg debugfs_port_reg = {
+	.lu = debugfs_port_reg_lu,
+	.count = 0x100,
+	.base_off = PORT_BASE,
+	.read_port_reg = hisi_sas_phy_read32,
+};
+
+static const struct hisi_sas_debugfs_reg_lu debugfs_global_reg_lu[] = {
+	HISI_SAS_DEBUGFS_REG(DLVRY_QUEUE_ENABLE),
+	HISI_SAS_DEBUGFS_REG(PHY_CONTEXT),
+	HISI_SAS_DEBUGFS_REG(PHY_STATE),
+	HISI_SAS_DEBUGFS_REG(PHY_PORT_NUM_MA),
+	HISI_SAS_DEBUGFS_REG(PHY_CONN_RATE),
+	HISI_SAS_DEBUGFS_REG(ITCT_CLR),
+	HISI_SAS_DEBUGFS_REG(IO_SATA_BROKEN_MSG_ADDR_LO),
+	HISI_SAS_DEBUGFS_REG(IO_SATA_BROKEN_MSG_ADDR_HI),
+	HISI_SAS_DEBUGFS_REG(SATA_INITI_D2H_STORE_ADDR_LO),
+	HISI_SAS_DEBUGFS_REG(SATA_INITI_D2H_STORE_ADDR_HI),
+	HISI_SAS_DEBUGFS_REG(CFG_MAX_TAG),
+	HISI_SAS_DEBUGFS_REG(HGC_SAS_TX_OPEN_FAIL_RETRY_CTRL),
+	HISI_SAS_DEBUGFS_REG(HGC_SAS_TXFAIL_RETRY_CTRL),
+	HISI_SAS_DEBUGFS_REG(HGC_GET_ITV_TIME),
+	HISI_SAS_DEBUGFS_REG(DEVICE_MSG_WORK_MODE),
+	HISI_SAS_DEBUGFS_REG(OPENA_WT_CONTI_TIME),
+	HISI_SAS_DEBUGFS_REG(I_T_NEXUS_LOSS_TIME),
+	HISI_SAS_DEBUGFS_REG(MAX_CON_TIME_LIMIT_TIME),
+	HISI_SAS_DEBUGFS_REG(BUS_INACTIVE_LIMIT_TIME),
+	HISI_SAS_DEBUGFS_REG(REJECT_TO_OPEN_LIMIT_TIME),
+	HISI_SAS_DEBUGFS_REG(CQ_INT_CONVERGE_EN),
+	HISI_SAS_DEBUGFS_REG(CFG_AGING_TIME),
+	HISI_SAS_DEBUGFS_REG(HGC_DFX_CFG2),
+	HISI_SAS_DEBUGFS_REG(CFG_ABT_SET_QUERY_IPTT),
+	HISI_SAS_DEBUGFS_REG(CFG_ABT_SET_IPTT_DONE),
+	HISI_SAS_DEBUGFS_REG(HGC_IOMB_PROC1_STATUS),
+	HISI_SAS_DEBUGFS_REG(CHNL_INT_STATUS),
+	HISI_SAS_DEBUGFS_REG(HGC_AXI_FIFO_ERR_INFO),
+	HISI_SAS_DEBUGFS_REG(INT_COAL_EN),
+	HISI_SAS_DEBUGFS_REG(OQ_INT_COAL_TIME),
+	HISI_SAS_DEBUGFS_REG(OQ_INT_COAL_CNT),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_COAL_TIME),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_COAL_CNT),
+	HISI_SAS_DEBUGFS_REG(OQ_INT_SRC),
+	HISI_SAS_DEBUGFS_REG(OQ_INT_SRC_MSK),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC1),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC2),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC3),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC_MSK1),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC_MSK2),
+	HISI_SAS_DEBUGFS_REG(ENT_INT_SRC_MSK3),
+	HISI_SAS_DEBUGFS_REG(CHNL_PHYUPDOWN_INT_MSK),
+	HISI_SAS_DEBUGFS_REG(CHNL_ENT_INT_MSK),
+	HISI_SAS_DEBUGFS_REG(HGC_COM_INT_MSK),
+	HISI_SAS_DEBUGFS_REG(SAS_ECC_INTR),
+	HISI_SAS_DEBUGFS_REG(SAS_ECC_INTR_MSK),
+	HISI_SAS_DEBUGFS_REG(HGC_ERR_STAT_EN),
+	HISI_SAS_DEBUGFS_REG(CQE_SEND_CNT),
+	HISI_SAS_DEBUGFS_REG(DLVRY_Q_0_DEPTH),
+	HISI_SAS_DEBUGFS_REG(DLVRY_Q_0_WR_PTR),
+	HISI_SAS_DEBUGFS_REG(DLVRY_Q_0_RD_PTR),
+	HISI_SAS_DEBUGFS_REG(HYPER_STREAM_ID_EN_CFG),
+	HISI_SAS_DEBUGFS_REG(OQ0_INT_SRC_MSK),
+	HISI_SAS_DEBUGFS_REG(COMPL_Q_0_DEPTH),
+	HISI_SAS_DEBUGFS_REG(COMPL_Q_0_WR_PTR),
+	HISI_SAS_DEBUGFS_REG(COMPL_Q_0_RD_PTR),
+	HISI_SAS_DEBUGFS_REG(AWQOS_AWCACHE_CFG),
+	HISI_SAS_DEBUGFS_REG(ARQOS_ARCACHE_CFG),
+	HISI_SAS_DEBUGFS_REG(HILINK_ERR_DFX),
+	HISI_SAS_DEBUGFS_REG(SAS_GPIO_CFG_0),
+	HISI_SAS_DEBUGFS_REG(SAS_GPIO_CFG_1),
+	HISI_SAS_DEBUGFS_REG(SAS_GPIO_TX_0_1),
+	HISI_SAS_DEBUGFS_REG(SAS_CFG_DRIVE_VLD),
+	{}
+};
+
+static const struct hisi_sas_debugfs_reg debugfs_global_reg = {
+	.lu = debugfs_global_reg_lu,
+	.count = 0x800,
+	.read_global_reg = hisi_sas_read32,
+};
+
+static void debugfs_snapshot_prepare_v3_hw(struct hisi_hba *hisi_hba)
+{
+	struct device *dev = hisi_hba->dev;
+
+	set_bit(HISI_SAS_REJECT_CMD_BIT, &hisi_hba->flags);
+
+	hisi_sas_write32(hisi_hba, DLVRY_QUEUE_ENABLE, 0);
+
+	if (wait_cmds_complete_timeout_v3_hw(hisi_hba, 100, 5000) == -ETIMEDOUT)
+		dev_dbg(dev, "Wait commands complete timeout!\n");
+
+	hisi_sas_kill_tasklets(hisi_hba);
+}
+
+static void debugfs_snapshot_restore_v3_hw(struct hisi_hba *hisi_hba)
+{
+	hisi_sas_write32(hisi_hba, DLVRY_QUEUE_ENABLE,
+			 (u32)((1ULL << hisi_hba->queue_count) - 1));
+
+	clear_bit(HISI_SAS_REJECT_CMD_BIT, &hisi_hba->flags);
+}
+
 static struct scsi_host_template sht_v3_hw = {
 	.name			= DRV_NAME,
 	.module			= THIS_MODULE,
@@ -2344,6 +2633,7 @@ static struct scsi_host_template sht_v3_hw = {
 	.bios_param		= sas_bios_param,
 	.this_id		= -1,
 	.sg_tablesize		= HISI_SAS_SGE_PAGE_CNT,
+	.sg_prot_tablesize	= HISI_SAS_SGE_PAGE_CNT,
 	.max_sectors		= SCSI_DEFAULT_MAX_SECTORS,
 	.eh_device_reset_handler = sas_eh_device_reset_handler,
 	.eh_target_reset_handler = sas_eh_target_reset_handler,
@@ -2360,7 +2650,7 @@ static const struct hisi_sas_hw hisi_sas_v3_hw = {
 	.get_wideport_bitmap = get_wideport_bitmap_v3_hw,
 	.complete_hdr_size = sizeof(struct hisi_sas_complete_v3_hdr),
 	.clear_itct = clear_itct_v3_hw,
-	.sl_notify = sl_notify_v3_hw,
+	.sl_notify_ssp = sl_notify_ssp_v3_hw,
 	.prep_ssp = prep_ssp_v3_hw,
 	.prep_smp = prep_smp_v3_hw,
 	.prep_stp = prep_ata_v3_hw,
@@ -2380,6 +2670,10 @@ static const struct hisi_sas_hw hisi_sas_v3_hw = {
 	.get_events = phy_get_events_v3_hw,
 	.write_gpio = write_gpio_v3_hw,
 	.wait_cmds_complete_timeout = wait_cmds_complete_timeout_v3_hw,
+	.debugfs_reg_global = &debugfs_global_reg,
+	.debugfs_reg_port = &debugfs_port_reg,
+	.snapshot_prepare = debugfs_snapshot_prepare_v3_hw,
+	.snapshot_restore = debugfs_snapshot_restore_v3_hw,
 };
 
 static struct Scsi_Host *
@@ -2397,6 +2691,7 @@ hisi_sas_shost_alloc_pci(struct pci_dev *pdev)
 	hisi_hba = shost_priv(shost);
 
 	INIT_WORK(&hisi_hba->rst_work, hisi_sas_rst_work_handler);
+	INIT_WORK(&hisi_hba->debugfs_work, hisi_sas_debugfs_work_handler);
 	hisi_hba->hw = &hisi_sas_v3_hw;
 	hisi_hba->pci_dev = pdev;
 	hisi_hba->dev = dev;
@@ -2414,7 +2709,7 @@ hisi_sas_shost_alloc_pci(struct pci_dev *pdev)
 	if (hisi_sas_get_fw_info(hisi_hba) < 0)
 		goto err_out;
 
-	if (hisi_sas_alloc(hisi_hba, shost)) {
+	if (hisi_sas_alloc(hisi_hba)) {
 		hisi_sas_free(hisi_hba);
 		goto err_out;
 	}
@@ -2513,8 +2808,14 @@ hisi_sas_v3_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		dev_info(dev, "Registering for DIF/DIX prot_mask=0x%x\n",
 			 prot_mask);
 		scsi_host_set_prot(hisi_hba->shost, prot_mask);
+		if (hisi_hba->prot_mask & HISI_SAS_DIX_PROT_MASK)
+			scsi_host_set_guard(hisi_hba->shost,
+					    SHOST_DIX_GUARD_CRC);
 	}
 
+	if (hisi_sas_debugfs_enable)
+		hisi_sas_debugfs_init(hisi_hba);
+
 	rc = scsi_add_host(shost, dev);
 	if (rc)
 		goto err_out_ha;
@@ -2551,7 +2852,7 @@ hisi_sas_v3_destroy_irqs(struct pci_dev *pdev, struct hisi_hba *hisi_hba)
 	free_irq(pci_irq_vector(pdev, 1), hisi_hba);
 	free_irq(pci_irq_vector(pdev, 2), hisi_hba);
 	free_irq(pci_irq_vector(pdev, 11), hisi_hba);
-	for (i = 0; i < hisi_hba->queue_count; i++) {
+	for (i = 0; i < hisi_hba->cq_nvecs; i++) {
 		struct hisi_sas_cq *cq = &hisi_hba->cq[i];
 		int nr = hisi_sas_intr_conv ? 16 : 16 + i;
 
@@ -2567,6 +2868,8 @@ static void hisi_sas_v3_remove(struct pci_dev *pdev)
 	struct hisi_hba *hisi_hba = sha->lldd_ha;
 	struct Scsi_Host *shost = sha->core.shost;
 
+	hisi_sas_debugfs_exit(hisi_hba);
+
 	if (timer_pending(&hisi_hba->timer))
 		del_timer(&hisi_hba->timer);