HZ1000

Signed-off-by: monkeyyang <monkeyyang@tencent.com>

drivers/scsi/smartpqi/smartpqi.h

@@ -21,6 +21,9 @@
 #if !defined(_SMARTPQI_H)
 #define _SMARTPQI_H
 
+#include <scsi/scsi_host.h>
+#include <linux/bsg-lib.h>
+
 #pragma pack(1)
 
 #define PQI_DEVICE_SIGNATURE	"PQI DREG"
@@ -97,6 +100,12 @@ struct pqi_ctrl_registers {
 	struct pqi_device_registers pqi_registers;	/* 4000h */
 };
 
+#if ((HZ) < 1000)
+#define PQI_HZ  1000
+#else
+#define PQI_HZ  (HZ)
+#endif
+
 #define PQI_DEVICE_REGISTERS_OFFSET	0x4000
 
 enum pqi_io_path {
@@ -347,6 +356,10 @@ struct pqi_event_config {
 
 #define PQI_MAX_EVENT_DESCRIPTORS	255
 
+#define PQI_EVENT_OFA_MEMORY_ALLOCATION	0x0
+#define PQI_EVENT_OFA_QUIESCE		0x1
+#define PQI_EVENT_OFA_CANCELLED		0x2
+
 struct pqi_event_response {
 	struct pqi_iu_header header;
 	u8	event_type;
@@ -354,7 +367,17 @@ struct pqi_event_response {
 	u8	request_acknowlege : 1;
 	__le16	event_id;
 	__le32	additional_event_id;
-	u8	data[16];
+	union {
+		struct {
+			__le32	bytes_requested;
+			u8	reserved[12];
+		} ofa_memory_allocation;
+
+		struct {
+			__le16	reason;		/* reason for cancellation */
+			u8	reserved[14];
+		} ofa_cancelled;
+	} data;
 };
 
 struct pqi_event_acknowledge_request {
@@ -389,6 +412,54 @@ struct pqi_task_management_response {
 	u8	response_code;
 };
 
+struct pqi_vendor_general_request {
+	struct pqi_iu_header header;
+	__le16	request_id;
+	__le16	function_code;
+	union {
+		struct {
+			__le16	first_section;
+			__le16	last_section;
+			u8	reserved[48];
+		} config_table_update;
+
+		struct {
+			__le64	buffer_address;
+			__le32	buffer_length;
+			u8	reserved[40];
+		} ofa_memory_allocation;
+	} data;
+};
+
+struct pqi_vendor_general_response {
+	struct pqi_iu_header header;
+	__le16	request_id;
+	__le16	function_code;
+	__le16	status;
+	u8	reserved[2];
+};
+
+#define PQI_VENDOR_GENERAL_CONFIG_TABLE_UPDATE	0
+#define PQI_VENDOR_GENERAL_HOST_MEMORY_UPDATE	1
+
+#define PQI_OFA_VERSION			1
+#define PQI_OFA_SIGNATURE		"OFA_QRM"
+#define PQI_OFA_MAX_SG_DESCRIPTORS	64
+
+#define PQI_OFA_MEMORY_DESCRIPTOR_LENGTH \
+	(offsetof(struct pqi_ofa_memory, sg_descriptor) + \
+	(PQI_OFA_MAX_SG_DESCRIPTORS * sizeof(struct pqi_sg_descriptor)))
+
+struct pqi_ofa_memory {
+	__le64	signature;	/* "OFA_QRM" */
+	__le16	version;	/* version of this struct(1 = 1st version) */
+	u8	reserved[62];
+	__le32	bytes_allocated;	/* total allocated memory in bytes */
+	__le16	num_memory_descriptors;
+	u8	reserved1[2];
+	struct pqi_sg_descriptor sg_descriptor[1];
+};
+
 struct pqi_aio_error_info {
 	u8	status;
 	u8	service_response;
@@ -419,6 +490,7 @@ struct pqi_raid_error_info {
 #define PQI_REQUEST_IU_GENERAL_ADMIN			0x60
 #define PQI_REQUEST_IU_REPORT_VENDOR_EVENT_CONFIG	0x72
 #define PQI_REQUEST_IU_SET_VENDOR_EVENT_CONFIG		0x73
+#define PQI_REQUEST_IU_VENDOR_GENERAL			0x75
 #define PQI_REQUEST_IU_ACKNOWLEDGE_VENDOR_EVENT		0xf6
 
 #define PQI_RESPONSE_IU_GENERAL_MANAGEMENT		0x81
@@ -430,6 +502,7 @@ struct pqi_raid_error_info {
 #define PQI_RESPONSE_IU_AIO_PATH_IO_ERROR		0xf3
 #define PQI_RESPONSE_IU_AIO_PATH_DISABLED		0xf4
 #define PQI_RESPONSE_IU_VENDOR_EVENT			0xf5
+#define PQI_RESPONSE_IU_VENDOR_GENERAL			0xf7
 
 #define PQI_GENERAL_ADMIN_FUNCTION_REPORT_DEVICE_CAPABILITY	0x0
 #define PQI_GENERAL_ADMIN_FUNCTION_CREATE_IQ			0x10
@@ -483,6 +556,8 @@ struct pqi_raid_error_info {
 #define CISS_CMD_STATUS_TMF			0xd
 #define CISS_CMD_STATUS_AIO_DISABLED		0xe
 
+#define PQI_CMD_STATUS_ABORTED	CISS_CMD_STATUS_ABORTED
+
 #define PQI_NUM_EVENT_QUEUE_ELEMENTS	32
 #define PQI_EVENT_OQ_ELEMENT_LENGTH	sizeof(struct pqi_event_response)
 
@@ -490,6 +565,7 @@ struct pqi_raid_error_info {
 #define PQI_EVENT_TYPE_HARDWARE			0x2
 #define PQI_EVENT_TYPE_PHYSICAL_DEVICE		0x4
 #define PQI_EVENT_TYPE_LOGICAL_DEVICE		0x5
+#define PQI_EVENT_TYPE_OFA			0xfb
 #define PQI_EVENT_TYPE_AIO_STATE_CHANGE		0xfd
 #define PQI_EVENT_TYPE_AIO_CONFIG_CHANGE	0xfe
 
@@ -554,6 +630,7 @@ typedef u32 pqi_index_t;
 #define SOP_TASK_ATTRIBUTE_ACA			4
 
 #define SOP_TMF_COMPLETE		0x0
+#define SOP_TMF_REJECTED		0x4
 #define SOP_TMF_FUNCTION_SUCCEEDED	0x8
 
 /* additional CDB bytes usage field codes */
@@ -581,8 +658,8 @@ struct pqi_admin_queues_aligned {
 struct pqi_admin_queues {
 	void		*iq_element_array;
 	void		*oq_element_array;
-	volatile pqi_index_t *iq_ci;
-	volatile pqi_index_t *oq_pi;
+	pqi_index_t	*iq_ci;
+	pqi_index_t __iomem *oq_pi;
 	dma_addr_t	iq_element_array_bus_addr;
 	dma_addr_t	oq_element_array_bus_addr;
 	dma_addr_t	iq_ci_bus_addr;
@@ -606,8 +683,8 @@ struct pqi_queue_group {
 	dma_addr_t	oq_element_array_bus_addr;
 	__le32 __iomem	*iq_pi[2];
 	pqi_index_t	iq_pi_copy[2];
-	volatile pqi_index_t *iq_ci[2];
-	volatile pqi_index_t *oq_pi;
+	pqi_index_t __iomem	*iq_ci[2];
+	pqi_index_t __iomem	*oq_pi;
 	dma_addr_t	iq_ci_bus_addr[2];
 	dma_addr_t	oq_pi_bus_addr;
 	__le32 __iomem	*oq_ci;
@@ -620,7 +697,7 @@ struct pqi_event_queue {
 	u16		oq_id;
 	u16		int_msg_num;
 	void		*oq_element_array;
-	volatile pqi_index_t *oq_pi;
+	pqi_index_t __iomem	*oq_pi;
 	dma_addr_t	oq_element_array_bus_addr;
 	dma_addr_t	oq_pi_bus_addr;
 	__le32 __iomem	*oq_ci;
@@ -642,11 +719,13 @@ struct pqi_encryption_info {
 #define PQI_CONFIG_TABLE_MAX_LENGTH	((u16)~0)
 
 /* configuration table section IDs */
+#define PQI_CONFIG_TABLE_ALL_SECTIONS			(-1)
 #define PQI_CONFIG_TABLE_SECTION_GENERAL_INFO		0
 #define PQI_CONFIG_TABLE_SECTION_FIRMWARE_FEATURES	1
 #define PQI_CONFIG_TABLE_SECTION_FIRMWARE_ERRATA	2
 #define PQI_CONFIG_TABLE_SECTION_DEBUG			3
 #define PQI_CONFIG_TABLE_SECTION_HEARTBEAT		4
+#define PQI_CONFIG_TABLE_SECTION_SOFT_RESET		5
 
 struct pqi_config_table {
 	u8	signature[8];		/* "CFGTABLE" */
@@ -678,6 +757,18 @@ struct pqi_config_table_general_info {
 					/* command */
 };
 
+struct pqi_config_table_firmware_features {
+	struct pqi_config_table_section_header header;
+	__le16	num_elements;
+	u8	features_supported[];
+/*	u8	features_requested_by_host[]; */
+/*	u8	features_enabled[]; */
+};
+
+#define PQI_FIRMWARE_FEATURE_OFA			0
+#define PQI_FIRMWARE_FEATURE_SMP			1
+#define PQI_FIRMWARE_FEATURE_SOFT_RESET_HANDSHAKE	11
+
 struct pqi_config_table_debug {
 	struct pqi_config_table_section_header header;
 	__le32	scratchpad;
@@ -688,6 +779,22 @@ struct pqi_config_table_heartbeat {
 	__le32	heartbeat_counter;
 };
 
+struct pqi_config_table_soft_reset {
+	struct pqi_config_table_section_header header;
+	u8 soft_reset_status;
+};
+
+#define PQI_SOFT_RESET_INITIATE		0x1
+#define PQI_SOFT_RESET_ABORT		0x2
+
+enum pqi_soft_reset_status {
+	RESET_INITIATE_FIRMWARE,
+	RESET_INITIATE_DRIVER,
+	RESET_ABORT,
+	RESET_NORESPONSE,
+	RESET_TIMEDOUT
+};
+
 union pqi_reset_register {
 	struct {
 		u32	reset_type : 3;
@@ -806,8 +913,10 @@ struct pqi_scsi_dev {
 	u8	scsi3addr[8];
 	__be64	wwid;
 	u8	volume_id[16];
+	u8	unique_id[16];
 	u8	is_physical_device : 1;
 	u8	is_external_raid_device : 1;
+	u8	is_expander_smp_device : 1;
 	u8	target_lun_valid : 1;
 	u8	device_gone : 1;
 	u8	new_device : 1;
@@ -815,6 +924,7 @@ struct pqi_scsi_dev {
 	u8	volume_offline : 1;
 	bool	aio_enabled;		/* only valid for physical disks */
 	bool	in_reset;
+	bool	in_remove;
 	bool	device_offline;
 	u8	vendor[8];		/* bytes 8-15 of inquiry data */
 	u8	model[16];		/* bytes 16-31 of inquiry data */
@@ -852,6 +962,8 @@ struct pqi_scsi_dev {
 #define CISS_VPD_LV_DEVICE_GEOMETRY	0xc1	/* vendor-specific page */
 #define CISS_VPD_LV_BYPASS_STATUS	0xc2	/* vendor-specific page */
 #define CISS_VPD_LV_STATUS		0xc3	/* vendor-specific page */
+#define SCSI_VPD_HEADER_SZ		4
+#define SCSI_VPD_DEVICE_ID_IDX		8	/* Index of page id in page */
 
 #define VPD_PAGE	(1 << 8)
 
@@ -914,6 +1026,7 @@ struct pqi_sas_node {
 struct pqi_sas_port {
 	struct list_head port_list_entry;
 	u64	sas_address;
+	struct pqi_scsi_dev *device;
 	struct sas_port *port;
 	int	next_phy_index;
 	struct list_head phy_list_head;
@@ -945,13 +1058,15 @@ struct pqi_io_request {
 	struct list_head request_list_entry;
 };
 
-#define PQI_NUM_SUPPORTED_EVENTS	6
+#define PQI_NUM_SUPPORTED_EVENTS	7
 
 struct pqi_event {
 	bool	pending;
 	u8	event_type;
 	__le16	event_id;
 	__le32	additional_event_id;
+	__le32	ofa_bytes_requested;
+	__le16	ofa_cancel_reason;
 };
 
 #define PQI_RESERVED_IO_SLOTS_LUN_RESET			1
@@ -1012,12 +1127,16 @@ struct pqi_ctrl_info {
 
 	struct mutex	scan_mutex;
 	struct mutex	lun_reset_mutex;
+	struct mutex	ofa_mutex; /* serialize ofa */
 	bool		controller_online;
 	bool		block_requests;
+	bool		in_shutdown;
+	bool		in_ofa;
 	u8		inbound_spanning_supported : 1;
 	u8		outbound_spanning_supported : 1;
 	u8		pqi_mode_enabled : 1;
 	u8		pqi_reset_quiesce_supported : 1;
+	u8		soft_reset_handshake_supported : 1;
 
 	struct list_head scsi_device_list;
 	spinlock_t	scsi_device_list_lock;
@@ -1038,6 +1157,7 @@ struct pqi_ctrl_info {
 	int		previous_num_interrupts;
 	u32		previous_heartbeat_count;
 	__le32 __iomem	*heartbeat_counter;
+	u8 __iomem	*soft_reset_status;
 	struct timer_list heartbeat_timer;
 	struct work_struct ctrl_offline_work;
 
@@ -1049,6 +1169,10 @@ struct pqi_ctrl_info {
 	struct list_head raid_bypass_retry_list;
 	spinlock_t	raid_bypass_retry_list_lock;
 	struct work_struct raid_bypass_retry_work;
+
+	struct          pqi_ofa_memory *pqi_ofa_mem_virt_addr;
+	dma_addr_t      pqi_ofa_mem_dma_handle;
+	void            **pqi_ofa_chunk_virt_addr;
 };
 
 enum pqi_ctrl_mode {
@@ -1078,8 +1202,13 @@ enum pqi_ctrl_mode {
 #define BMIC_WRITE				0x27
 #define BMIC_SENSE_CONTROLLER_PARAMETERS	0x64
 #define BMIC_SENSE_SUBSYSTEM_INFORMATION	0x66
+#define BMIC_CSMI_PASSTHRU			0x68
 #define BMIC_WRITE_HOST_WELLNESS		0xa5
 #define BMIC_FLUSH_CACHE			0xc2
+#define BMIC_SET_DIAG_OPTIONS			0xf4
+#define BMIC_SENSE_DIAG_OPTIONS			0xf5
+
+#define CSMI_CC_SAS_SMP_PASSTHRU		0X17
 
 #define SA_FLUSH_CACHE				0x1
 
@@ -1107,6 +1236,10 @@ struct bmic_identify_controller {
 	u8	reserved3[32];
 };
 
+#define SA_EXPANDER_SMP_DEVICE		0x05
+/*SCSI Invalid Device Type for SAS devices*/
+#define PQI_SAS_SCSI_INVALID_DEVTYPE	0xff
+
 struct bmic_identify_physical_device {
 	u8	scsi_bus;		/* SCSI Bus number on controller */
 	u8	scsi_id;		/* SCSI ID on this bus */
@@ -1187,6 +1320,50 @@ struct bmic_identify_physical_device {
 	u8	padding_to_multiple_of_512[9];
 };
 
+struct bmic_smp_request {
+	u8	frame_type;
+	u8	function;
+	u8	allocated_response_length;
+	u8	request_length;
+	u8	additional_request_bytes[1016];
+};
+
+struct  bmic_smp_response {
+	u8	frame_type;
+	u8	function;
+	u8	function_result;
+	u8	response_length;
+	u8	additional_response_bytes[1016];
+};
+
+struct bmic_csmi_ioctl_header {
+	__le32	header_length;
+	u8	signature[8];
+	__le32	timeout;
+	__le32	control_code;
+	__le32	return_code;
+	__le32	length;
+};
+
+struct bmic_csmi_smp_passthru {
+	u8	phy_identifier;
+	u8	port_identifier;
+	u8	connection_rate;
+	u8	reserved;
+	__be64	destination_sas_address;
+	__le32	request_length;
+	struct bmic_smp_request request;
+	u8	connection_status;
+	u8	reserved1[3];
+	__le32	response_length;
+	struct bmic_smp_response response;
+};
+
+struct bmic_csmi_smp_passthru_buffer {
+	struct bmic_csmi_ioctl_header ioctl_header;
+	struct bmic_csmi_smp_passthru parameters;
+};
+
 struct bmic_flush_cache {
 	u8	disable_flag;
 	u8	system_power_action;
@@ -1204,8 +1381,42 @@ enum bmic_flush_cache_shutdown_event {
 	RESTART = 4
 };
 
+struct bmic_diag_options {
+	__le32 options;
+};
+
 #pragma pack()
 
+static inline struct pqi_ctrl_info *shost_to_hba(struct Scsi_Host *shost)
+{
+	void *hostdata = shost_priv(shost);
+
+	return *((struct pqi_ctrl_info **)hostdata);
+}
+
+static inline bool pqi_ctrl_offline(struct pqi_ctrl_info *ctrl_info)
+{
+	return !ctrl_info->controller_online;
+}
+
+static inline void pqi_ctrl_busy(struct pqi_ctrl_info *ctrl_info)
+{
+	atomic_inc(&ctrl_info->num_busy_threads);
+}
+
+static inline void pqi_ctrl_unbusy(struct pqi_ctrl_info *ctrl_info)
+{
+	atomic_dec(&ctrl_info->num_busy_threads);
+}
+
+static inline bool pqi_ctrl_blocked(struct pqi_ctrl_info *ctrl_info)
+{
+	return ctrl_info->block_requests;
+}
+
+void pqi_sas_smp_handler(struct bsg_job *job, struct Scsi_Host *shost,
+	struct sas_rphy *rphy);
+
 int pqi_add_sas_host(struct Scsi_Host *shost, struct pqi_ctrl_info *ctrl_info);
 void pqi_delete_sas_host(struct pqi_ctrl_info *ctrl_info);
 int pqi_add_sas_device(struct pqi_sas_node *pqi_sas_node,
@@ -1214,6 +1425,9 @@ void pqi_remove_sas_device(struct pqi_scsi_dev *device);
 struct pqi_scsi_dev *pqi_find_device_by_sas_rphy(
 	struct pqi_ctrl_info *ctrl_info, struct sas_rphy *rphy);
 void pqi_prep_for_scsi_done(struct scsi_cmnd *scmd);
+int pqi_csmi_smp_passthru(struct pqi_ctrl_info *ctrl_info,
+	struct bmic_csmi_smp_passthru_buffer *buffer, size_t buffer_length,
+	struct pqi_raid_error_info *error_info);
 
 extern struct sas_function_template pqi_sas_transport_functions;
 

drivers/scsi/smartpqi/smartpqi_sas_transport.c

@@ -17,9 +17,11 @@
  */
 
 #include <linux/kernel.h>
+#include <linux/bsg-lib.h>
 #include <scsi/scsi_host.h>
 #include <scsi/scsi_cmnd.h>
 #include <scsi/scsi_transport_sas.h>
+#include <asm/unaligned.h>
 #include "smartpqi.h"
 
 static struct pqi_sas_phy *pqi_alloc_sas_phy(struct pqi_sas_port *pqi_sas_port)
@@ -97,14 +99,32 @@ static int pqi_sas_port_add_rphy(struct pqi_sas_port *pqi_sas_port,
 
 	identify = &rphy->identify;
 	identify->sas_address = pqi_sas_port->sas_address;
-	identify->initiator_port_protocols = SAS_PROTOCOL_STP;
-	identify->target_port_protocols = SAS_PROTOCOL_STP;
+
+	if (pqi_sas_port->device &&
+		pqi_sas_port->device->is_expander_smp_device) {
+		identify->initiator_port_protocols = SAS_PROTOCOL_SMP;
+		identify->target_port_protocols = SAS_PROTOCOL_SMP;
+	} else {
+		identify->initiator_port_protocols = SAS_PROTOCOL_STP;
+		identify->target_port_protocols = SAS_PROTOCOL_STP;
+	}
 
 	return sas_rphy_add(rphy);
 }
 
+static struct sas_rphy *pqi_sas_rphy_alloc(struct pqi_sas_port *pqi_sas_port)
+{
+	if (pqi_sas_port->device &&
+		pqi_sas_port->device->is_expander_smp_device)
+		return sas_expander_alloc(pqi_sas_port->port,
+				SAS_FANOUT_EXPANDER_DEVICE);
+
+	return sas_end_device_alloc(pqi_sas_port->port);
+}
+
 static struct pqi_sas_port *pqi_alloc_sas_port(
-	struct pqi_sas_node *pqi_sas_node, u64 sas_address)
+	struct pqi_sas_node *pqi_sas_node, u64 sas_address,
+	struct pqi_scsi_dev *device)
 {
 	int rc;
 	struct pqi_sas_port *pqi_sas_port;
@@ -127,6 +147,7 @@ static struct pqi_sas_port *pqi_alloc_sas_port(
 
 	pqi_sas_port->port = port;
 	pqi_sas_port->sas_address = sas_address;
+	pqi_sas_port->device = device;
 	list_add_tail(&pqi_sas_port->port_list_entry,
 		&pqi_sas_node->port_list_head);
 
@@ -146,7 +167,7 @@ static void pqi_free_sas_port(struct pqi_sas_port *pqi_sas_port)
 	struct pqi_sas_phy *next;
 
 	list_for_each_entry_safe(pqi_sas_phy, next,
-			&pqi_sas_port->phy_list_head, phy_list_entry)
+		&pqi_sas_port->phy_list_head, phy_list_entry)
 		pqi_free_sas_phy(pqi_sas_phy);
 
 	sas_port_delete(pqi_sas_port->port);
@@ -176,7 +197,7 @@ static void pqi_free_sas_node(struct pqi_sas_node *pqi_sas_node)
 		return;
 
 	list_for_each_entry_safe(pqi_sas_port, next,
-			&pqi_sas_node->port_list_head, port_list_entry)
+		&pqi_sas_node->port_list_head, port_list_entry)
 		pqi_free_sas_port(pqi_sas_port);
 
 	kfree(pqi_sas_node);
@@ -206,13 +227,14 @@ int pqi_add_sas_host(struct Scsi_Host *shost, struct pqi_ctrl_info *ctrl_info)
 	struct pqi_sas_port *pqi_sas_port;
 	struct pqi_sas_phy *pqi_sas_phy;
 
-	parent_dev = &shost->shost_gendev;
+	parent_dev = &shost->shost_dev;
 
 	pqi_sas_node = pqi_alloc_sas_node(parent_dev);
 	if (!pqi_sas_node)
 		return -ENOMEM;
 
-	pqi_sas_port = pqi_alloc_sas_port(pqi_sas_node, ctrl_info->sas_address);
+	pqi_sas_port = pqi_alloc_sas_port(pqi_sas_node,
+		ctrl_info->sas_address, NULL);
 	if (!pqi_sas_port) {
 		rc = -ENODEV;
 		goto free_sas_node;
@@ -254,11 +276,12 @@ int pqi_add_sas_device(struct pqi_sas_node *pqi_sas_node,
 	struct pqi_sas_port *pqi_sas_port;
 	struct sas_rphy *rphy;
 
-	pqi_sas_port = pqi_alloc_sas_port(pqi_sas_node, device->sas_address);
+	pqi_sas_port = pqi_alloc_sas_port(pqi_sas_node,
+		device->sas_address, device);
 	if (!pqi_sas_port)
 		return -ENOMEM;
 
-	rphy = sas_end_device_alloc(pqi_sas_port->port);
+	rphy = pqi_sas_rphy_alloc(pqi_sas_port);
 	if (!rphy) {
 		rc = -ENODEV;
 		goto free_sas_port;
@@ -329,6 +352,128 @@ static int pqi_sas_phy_speed(struct sas_phy *phy,
 	return -EINVAL;
 }
 
+#define CSMI_IOCTL_TIMEOUT	60
+#define SMP_CRC_FIELD_LENGTH	4
+
+static struct bmic_csmi_smp_passthru_buffer *
+pqi_build_csmi_smp_passthru_buffer(struct sas_rphy *rphy,
+	struct bsg_job *job)
+{
+	struct bmic_csmi_smp_passthru_buffer *smp_buf;
+	struct bmic_csmi_ioctl_header *ioctl_header;
+	struct bmic_csmi_smp_passthru *parameters;
+	u32 req_size;
+	u32 resp_size;
+
+	smp_buf = kzalloc(sizeof(*smp_buf), GFP_KERNEL);
+	if (!smp_buf)
+		return NULL;
+
+	req_size = job->request_payload.payload_len;
+	resp_size = job->reply_payload.payload_len;
+
+	ioctl_header = &smp_buf->ioctl_header;
+	put_unaligned_le32(sizeof(smp_buf->ioctl_header),
+		&ioctl_header->header_length);
+	put_unaligned_le32(CSMI_IOCTL_TIMEOUT, &ioctl_header->timeout);
+	put_unaligned_le32(CSMI_CC_SAS_SMP_PASSTHRU,
+		&ioctl_header->control_code);
+	put_unaligned_le32(sizeof(smp_buf->parameters), &ioctl_header->length);
+
+	parameters = &smp_buf->parameters;
+	parameters->phy_identifier = rphy->identify.phy_identifier;
+	parameters->port_identifier = 0;
+	parameters->connection_rate = 0;
+	put_unaligned_be64(rphy->identify.sas_address,
+		&parameters->destination_sas_address);
+
+	if (req_size > SMP_CRC_FIELD_LENGTH)
+		req_size -= SMP_CRC_FIELD_LENGTH;
+
+	put_unaligned_le32(req_size, &parameters->request_length);
+
+	put_unaligned_le32(resp_size, &parameters->response_length);
+
+	sg_copy_to_buffer(job->request_payload.sg_list,
+		job->reply_payload.sg_cnt, &parameters->request,
+		req_size);
+
+	return smp_buf;
+}
+
+static unsigned int pqi_build_sas_smp_handler_reply(
+	struct bmic_csmi_smp_passthru_buffer *smp_buf, struct bsg_job *job,
+	struct pqi_raid_error_info *error_info)
+{
+	sg_copy_from_buffer(job->reply_payload.sg_list,
+		job->reply_payload.sg_cnt, &smp_buf->parameters.response,
+		le32_to_cpu(smp_buf->parameters.response_length));
+
+	job->reply_len = le16_to_cpu(error_info->sense_data_length);
+	memcpy(job->reply, error_info->data,
+			le16_to_cpu(error_info->sense_data_length));
+
+	return job->reply_payload.payload_len -
+		get_unaligned_le32(&error_info->data_in_transferred);
+}
+
+void pqi_sas_smp_handler(struct bsg_job *job, struct Scsi_Host *shost,
+	struct sas_rphy *rphy)
+{
+	int rc;
+	struct pqi_ctrl_info *ctrl_info = shost_to_hba(shost);
+	struct bmic_csmi_smp_passthru_buffer *smp_buf;
+	struct pqi_raid_error_info error_info;
+	unsigned int reslen = 0;
+
+	pqi_ctrl_busy(ctrl_info);
+
+	if (job->reply_payload.payload_len == 0) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	if (!rphy) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (rphy->identify.device_type != SAS_FANOUT_EXPANDER_DEVICE) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (job->request_payload.sg_cnt > 1 || job->reply_payload.sg_cnt > 1) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (pqi_ctrl_offline(ctrl_info)) {
+		rc = -ENXIO;
+		goto out;
+	}
+
+	if (pqi_ctrl_blocked(ctrl_info)) {
+		rc = -EBUSY;
+		goto out;
+	}
+
+	smp_buf = pqi_build_csmi_smp_passthru_buffer(rphy, job);
+	if (!smp_buf) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	rc = pqi_csmi_smp_passthru(ctrl_info, smp_buf, sizeof(*smp_buf),
+		&error_info);
+	if (rc)
+		goto out;
+
+	reslen = pqi_build_sas_smp_handler_reply(smp_buf, job, &error_info);
+out:
+	bsg_job_done(job, rc, reslen);
+	pqi_ctrl_unbusy(ctrl_info);
+}
 struct sas_function_template pqi_sas_transport_functions = {
 	.get_linkerrors = pqi_sas_get_linkerrors,
 	.get_enclosure_identifier = pqi_sas_get_enclosure_identifier,
@@ -338,4 +483,5 @@ struct sas_function_template pqi_sas_transport_functions = {
 	.phy_setup = pqi_sas_phy_setup,
 	.phy_release = pqi_sas_phy_release,
 	.set_phy_speed = pqi_sas_phy_speed,
+	.smp_handler = pqi_sas_smp_handler,
 };

drivers/scsi/smartpqi/smartpqi_sis.c

@@ -34,6 +34,7 @@
 #define SIS_REENABLE_SIS_MODE			0x1
 #define SIS_ENABLE_MSIX				0x40
 #define SIS_ENABLE_INTX				0x80
+#define SIS_SOFT_RESET				0x100
 #define SIS_CMD_READY				0x200
 #define SIS_TRIGGER_SHUTDOWN			0x800000
 #define SIS_PQI_RESET_QUIESCE			0x1000000
@@ -90,7 +91,7 @@ static int sis_wait_for_ctrl_ready_with_timeout(struct pqi_ctrl_info *ctrl_info,
 	unsigned long timeout;
 	u32 status;
 
-	timeout = (timeout_secs * HZ) + jiffies;
+	timeout = (timeout_secs * PQI_HZ) + jiffies;
 
 	while (1) {
 		status = readl(&ctrl_info->registers->sis_firmware_status);
@@ -202,7 +203,7 @@ static int sis_send_sync_cmd(struct pqi_ctrl_info *ctrl_info,
 	 * the top of the loop in order to give the controller time to start
 	 * processing the command before we start polling.
 	 */
-	timeout = (SIS_CMD_COMPLETE_TIMEOUT_SECS * HZ) + jiffies;
+	timeout = (SIS_CMD_COMPLETE_TIMEOUT_SECS * PQI_HZ) + jiffies;
 	while (1) {
 		msleep(SIS_CMD_COMPLETE_POLL_INTERVAL_MSECS);
 		doorbell = readl(&registers->sis_ctrl_to_host_doorbell);
@@ -316,9 +317,9 @@ int sis_init_base_struct_addr(struct pqi_ctrl_info *ctrl_info)
 	put_unaligned_le32(ctrl_info->max_io_slots,
 		&base_struct->error_buffer_num_elements);
 
-	bus_address = pci_map_single(ctrl_info->pci_dev, base_struct,
-		sizeof(*base_struct), PCI_DMA_TODEVICE);
-	if (pci_dma_mapping_error(ctrl_info->pci_dev, bus_address)) {
+	bus_address = dma_map_single(&ctrl_info->pci_dev->dev, base_struct,
+		sizeof(*base_struct), DMA_TO_DEVICE);
+	if (dma_mapping_error(&ctrl_info->pci_dev->dev, bus_address)) {
 		rc = -ENOMEM;
 		goto out;
 	}
@@ -331,9 +332,8 @@ int sis_init_base_struct_addr(struct pqi_ctrl_info *ctrl_info)
 	rc = sis_send_sync_cmd(ctrl_info, SIS_CMD_INIT_BASE_STRUCT_ADDRESS,
 		&params);
 
-	pci_unmap_single(ctrl_info->pci_dev, bus_address, sizeof(*base_struct),
-		PCI_DMA_TODEVICE);
-
+	dma_unmap_single(&ctrl_info->pci_dev->dev, bus_address,
+			sizeof(*base_struct), DMA_TO_DEVICE);
 out:
 	kfree(base_struct_unaligned);
 
@@ -349,7 +349,7 @@ static int sis_wait_for_doorbell_bit_to_clear(
 	u32 doorbell_register;
 	unsigned long timeout;
 
-	timeout = (SIS_DOORBELL_BIT_CLEAR_TIMEOUT_SECS * HZ) + jiffies;
+	timeout = (SIS_DOORBELL_BIT_CLEAR_TIMEOUT_SECS * PQI_HZ) + jiffies;
 
 	while (1) {
 		doorbell_register =
@@ -421,6 +421,12 @@ u32 sis_read_driver_scratch(struct pqi_ctrl_info *ctrl_info)
 	return readl(&ctrl_info->registers->sis_driver_scratch);
 }
 
+void sis_soft_reset(struct pqi_ctrl_info *ctrl_info)
+{
+	writel(SIS_SOFT_RESET,
+		&ctrl_info->registers->sis_host_to_ctrl_doorbell);
+}
+
 static void __attribute__((unused)) verify_structures(void)
 {
 	BUILD_BUG_ON(offsetof(struct sis_base_struct,

drivers/scsi/smartpqi/smartpqi_sis.h

@@ -33,5 +33,6 @@ int sis_pqi_reset_quiesce(struct pqi_ctrl_info *ctrl_info);
 int sis_reenable_sis_mode(struct pqi_ctrl_info *ctrl_info);
 void sis_write_driver_scratch(struct pqi_ctrl_info *ctrl_info, u32 value);
 u32 sis_read_driver_scratch(struct pqi_ctrl_info *ctrl_info);
+void sis_soft_reset(struct pqi_ctrl_info *ctrl_info);
 
 #endif	/* _SMARTPQI_SIS_H */

include/linux/netfilter/x_tables.h

@@ -373,7 +373,7 @@ static inline unsigned int xt_write_recseq_begin(void)
 	 * since addend is most likely 1
 	 */
 	__this_cpu_add(xt_recseq.sequence, addend);
-	smp_wmb();
+	smp_mb();
 
 	return addend;
 }

kernel/sched/fair.c

@@ -5449,28 +5449,43 @@ static int wake_wide(struct task_struct *p)
  * soonest. For the purpose of speed we only consider the waking and previous
  * CPU.
  *
- * wake_affine_idle() - only considers 'now', it check if the waking CPU is (or
- *			will be) idle.
+ * wake_affine_idle() - only considers 'now', it check if the waking CPU is
+ * 			cache-affine and is (or will be) idle.
  *
  * wake_affine_weight() - considers the weight to reflect the average
  *			  scheduling latency of the CPUs. This seems to work
  *			  for the overloaded case.
  */
 
-static bool
+static int
 wake_affine_idle(struct sched_domain *sd, struct task_struct *p,
 		 int this_cpu, int prev_cpu, int sync)
 {
-	if (idle_cpu(this_cpu))
-		return true;
+	/*
+	 * If this_cpu is idle, it implies the wakeup is from interrupt
+	 * context. Only allow the move if cache is shared. Otherwise an
+	 * interrupt intensive workload could force all tasks onto one
+	 * node depending on the IO topology or IRQ affinity settings.
+	 *
+	 * If the prev_cpu is idle and cache affine then avoid a migration.
+	 * There is no guarantee that the cache hot data from an interrupt
+	 * is more important than cache hot data on the prev_cpu and from
+	 * a cpufreq perspective, it's better to have higher utilisation
+	 * on one CPU.
+	 */
+	if (idle_cpu(this_cpu) && cpus_share_cache(this_cpu, prev_cpu))
+		return idle_cpu(prev_cpu) ? prev_cpu : this_cpu;
 
 	if (sync && cpu_rq(this_cpu)->nr_running == 1)
-		return true;
+		return this_cpu;
 
-	return false;
+	if (idle_cpu(prev_cpu))
+		return prev_cpu;
+
+	return nr_cpumask_bits;
 }
 
-static bool
+static int
 wake_affine_weight(struct sched_domain *sd, struct task_struct *p,
 		   int this_cpu, int prev_cpu, int sync)
 {
@@ -5484,7 +5499,7 @@ wake_affine_weight(struct sched_domain *sd, struct task_struct *p,
 		unsigned long current_load = task_h_load(current);
 
 		if (current_load > this_eff_load)
-			return true;
+			return this_cpu;
 
 		this_eff_load -= current_load;
 	}
@@ -5501,7 +5516,16 @@ wake_affine_weight(struct sched_domain *sd, struct task_struct *p,
 		prev_eff_load *= 100 + (sd->imbalance_pct - 100) / 2;
 	prev_eff_load *= capacity_of(this_cpu);
 
-	return this_eff_load <= prev_eff_load;
+	/*
+	 * If sync, adjust the weight of prev_eff_load such that if
+	 * prev_eff == this_eff that select_idle_sibling() will consider
+	 * stacking the wakee on top of the waker if no other CPU is
+	 * idle.
+	 */
+	if (sync)
+		prev_eff_load += 1;
+
+	return this_eff_load < prev_eff_load ? this_cpu : nr_cpumask_bits;
 }
 
 static int wake_affine(struct sched_domain *sd, struct task_struct *p,
@@ -5509,14 +5533,17 @@ static int wake_affine(struct sched_domain *sd, struct task_struct *p,
 {
 	int this_cpu = smp_processor_id();
 	bool affine = false;
+	int target = nr_cpumask_bits;
 
-	if (sched_feat(WA_IDLE) && !affine)
-		affine = wake_affine_idle(sd, p, this_cpu, prev_cpu, sync);
+	if (sched_feat(WA_IDLE))
+		target = wake_affine_idle(sd, p, this_cpu, prev_cpu, sync);
 
-	if (sched_feat(WA_WEIGHT) && !affine)
-		affine = wake_affine_weight(sd, p, this_cpu, prev_cpu, sync);
+	if (sched_feat(WA_WEIGHT) && target == nr_cpumask_bits)
+		target = wake_affine_weight(sd, p, this_cpu, prev_cpu, sync);
 
 	schedstat_inc(p->se.statistics.nr_wakeups_affine_attempts);
+	if (target == this_cpu)
+		affine = true;
 	if (affine) {
 		schedstat_inc(sd->ttwu_move_affine);
 		schedstat_inc(p->se.statistics.nr_wakeups_affine);

net/ipv4/netfilter/arp_tables.c

@@ -638,6 +638,25 @@ static void get_counters(const struct xt_table_info *t,
 	}
 }
 
+static void get_old_counters(const struct xt_table_info *t,
+                            struct xt_counters counters[])
+{
+	struct arpt_entry *iter;
+	unsigned int cpu, i;
+	for_each_possible_cpu(cpu) {
+		i = 0;
+		xt_entry_foreach(iter, t->entries, t->size) {
+			struct xt_counters *tmp;
+
+			tmp = xt_get_per_cpu_counter(&iter->counters, cpu);
+			ADD_COUNTER(counters[i], tmp->bcnt, tmp->pcnt);
+			++i;
+		}
+		cond_resched();
+	}
+}
+
+
 static struct xt_counters *alloc_counters(const struct xt_table *table)
 {
 	unsigned int countersize;
@@ -916,8 +935,7 @@ static int __do_replace(struct net *net, const char *name,
 	    (newinfo->number <= oldinfo->initial_entries))
 		module_put(t->me);
 
-	/* Get the old counters, and synchronize with replace */
-	get_counters(oldinfo, counters);
+	get_old_counters(oldinfo, counters);
 
 	/* Decrease module usage counts and free resource */
 	loc_cpu_old_entry = oldinfo->entries;

net/ipv4/netfilter/ip_tables.c

@@ -787,6 +787,27 @@ get_counters(const struct xt_table_info *t,
 	}
 }
 
+static void get_old_counters(const struct xt_table_info *t,
+                            struct xt_counters counters[])
+{
+	struct ipt_entry *iter;
+	unsigned int cpu, i;
+
+	for_each_possible_cpu(cpu) {
+		i = 0;
+		xt_entry_foreach(iter, t->entries, t->size) {
+			const struct xt_counters *tmp;
+
+			tmp = xt_get_per_cpu_counter(&iter->counters, cpu);
+			ADD_COUNTER(counters[i], tmp->bcnt, tmp->pcnt);
+			++i; /* macro does multi eval of i */
+		}
+
+		cond_resched();
+	}
+}
+
+
 static struct xt_counters *alloc_counters(const struct xt_table *table)
 {
 	unsigned int countersize;
@@ -1078,8 +1099,7 @@ __do_replace(struct net *net, const char *name, unsigned int valid_hooks,
 	    (newinfo->number <= oldinfo->initial_entries))
 		module_put(t->me);
 
-	/* Get the old counters, and synchronize with replace */
-	get_counters(oldinfo, counters);
+	get_old_counters(oldinfo, counters);
 
 	/* Decrease module usage counts and free resource */
 	xt_entry_foreach(iter, oldinfo->entries, oldinfo->size)

net/ipv4/route.c

@@ -1403,6 +1403,7 @@ static bool rt_bind_exception(struct rtable *rt, struct fib_nh_exception *fnhe,
 	return ret;
 }
 
+static void rt_add_uncached_list(struct rtable *rt);
 static bool rt_cache_route(struct fib_nh *nh, struct rtable *rt)
 {
 	struct rtable *orig, *prev, **p;
@@ -1422,7 +1423,7 @@ static bool rt_cache_route(struct fib_nh *nh, struct rtable *rt)
 	prev = cmpxchg(p, orig, rt);
 	if (prev == orig) {
 		if (orig) {
-			dst_dev_put(&orig->dst);
+			rt_add_uncached_list(orig);
 			dst_release(&orig->dst);
 		}
 	} else {

net/ipv6/netfilter/ip6_tables.c

@@ -805,6 +805,25 @@ get_counters(const struct xt_table_info *t,
 	}
 }
 
+static void get_old_counters(const struct xt_table_info *t,
+                            struct xt_counters counters[])
+{
+	struct ip6t_entry *iter;
+	unsigned int cpu, i;
+
+	for_each_possible_cpu(cpu) {
+		i = 0;
+		xt_entry_foreach(iter, t->entries, t->size) {
+			const struct xt_counters *tmp;
+
+			tmp = xt_get_per_cpu_counter(&iter->counters, cpu);
+			ADD_COUNTER(counters[i], tmp->bcnt, tmp->pcnt);
+			++i;
+		}
+		cond_resched();
+	}
+}
+
 static struct xt_counters *alloc_counters(const struct xt_table *table)
 {
 	unsigned int countersize;
@@ -1097,8 +1116,7 @@ __do_replace(struct net *net, const char *name, unsigned int valid_hooks,
 	    (newinfo->number <= oldinfo->initial_entries))
 		module_put(t->me);
 
-	/* Get the old counters, and synchronize with replace */
-	get_counters(oldinfo, counters);
+	get_old_counters(oldinfo, counters);
 
 	/* Decrease module usage counts and free resource */
 	xt_entry_foreach(iter, oldinfo->entries, oldinfo->size)

net/netfilter/x_tables.c

@@ -1225,6 +1225,7 @@ xt_replace_table(struct xt_table *table,
 {
 	struct xt_table_info *private;
 	int ret;
+	unsigned int cpu;
 
 	ret = xt_jumpstack_alloc(newinfo);
 	if (ret < 0) {
@@ -1253,14 +1254,26 @@ xt_replace_table(struct xt_table *table,
 	smp_wmb();
 	table->private = newinfo;
 
+	/* make sure all cpus see new ->private value */
+	smp_mb();
 	/*
 	 * Even though table entries have now been swapped, other CPU's
-	 * may still be using the old entries. This is okay, because
-	 * resynchronization happens because of the locking done
-	 * during the get_counters() routine.
+	 * may still be using the old entries...
 	 */
 	local_bh_enable();
 
+	/* ... so wait for even xt_recseq on all cpus */
+	for_each_possible_cpu(cpu) {
+		seqcount_t *s = &per_cpu(xt_recseq, cpu);
+		u32 seq = raw_read_seqcount(s);
+		if (seq & 1) {
+			do {
+				cond_resched();
+				cpu_relax();
+			} while (seq == raw_read_seqcount(s));
+		}
+	}
+
 #ifdef CONFIG_AUDIT
 	if (audit_enabled) {
 		audit_log(current->audit_context, GFP_KERNEL,

package/default/config.default

@@ -644,10 +644,10 @@ CONFIG_EFI_STUB=y
 # CONFIG_EFI_MIXED is not set
 CONFIG_SECCOMP=y
 # CONFIG_HZ_100 is not set
-CONFIG_HZ_250=y
+# CONFIG_HZ_250 is not set
 # CONFIG_HZ_300 is not set
-# CONFIG_HZ_1000 is not set
-CONFIG_HZ=250
+CONFIG_HZ_1000=y
+CONFIG_HZ=1000
 CONFIG_SCHED_HRTICK=y
 CONFIG_KEXEC=y
 # CONFIG_KEXEC_FILE is not set
@@ -1715,7 +1715,7 @@ CONFIG_SCSI_ARCMSR=y
 CONFIG_SCSI_MPT3SAS=y
 CONFIG_SCSI_MPT2SAS_MAX_SGE=128
 CONFIG_SCSI_MPT3SAS_MAX_SGE=128
-# CONFIG_SCSI_SMARTPQI is not set
+CONFIG_SCSI_SMARTPQI=y
 CONFIG_SCSI_UFSHCD=y
 CONFIG_SCSI_UFSHCD_PCI=y
 # CONFIG_SCSI_UFS_DWC_TC_PCI is not set