From 5d1646d90e1f2cceb9f0828f4b28318cd0ec7744 Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Sat, 27 Apr 2024 12:05:51 +0200 Subject: Adding upstream version 5.10.209. Signed-off-by: Daniel Baumann --- drivers/scsi/storvsc_drv.c | 2197 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 2197 insertions(+) create mode 100644 drivers/scsi/storvsc_drv.c (limited to 'drivers/scsi/storvsc_drv.c') diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c new file mode 100644 index 000000000..37ad5f525 --- /dev/null +++ b/drivers/scsi/storvsc_drv.c @@ -0,0 +1,2197 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2009, Microsoft Corporation. + * + * Authors: + * Haiyang Zhang + * Hank Janssen + * K. Y. Srinivasan + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * All wire protocol details (storage protocol between the guest and the host) + * are consolidated here. + * + * Begin protocol definitions. + */ + +/* + * Version history: + * V1 Beta: 0.1 + * V1 RC < 2008/1/31: 1.0 + * V1 RC > 2008/1/31: 2.0 + * Win7: 4.2 + * Win8: 5.1 + * Win8.1: 6.0 + * Win10: 6.2 + */ + +#define VMSTOR_PROTO_VERSION(MAJOR_, MINOR_) ((((MAJOR_) & 0xff) << 8) | \ + (((MINOR_) & 0xff))) + +#define VMSTOR_PROTO_VERSION_WIN6 VMSTOR_PROTO_VERSION(2, 0) +#define VMSTOR_PROTO_VERSION_WIN7 VMSTOR_PROTO_VERSION(4, 2) +#define VMSTOR_PROTO_VERSION_WIN8 VMSTOR_PROTO_VERSION(5, 1) +#define VMSTOR_PROTO_VERSION_WIN8_1 VMSTOR_PROTO_VERSION(6, 0) +#define VMSTOR_PROTO_VERSION_WIN10 VMSTOR_PROTO_VERSION(6, 2) + +/* Packet structure describing virtual storage requests. */ +enum vstor_packet_operation { + VSTOR_OPERATION_COMPLETE_IO = 1, + VSTOR_OPERATION_REMOVE_DEVICE = 2, + VSTOR_OPERATION_EXECUTE_SRB = 3, + VSTOR_OPERATION_RESET_LUN = 4, + VSTOR_OPERATION_RESET_ADAPTER = 5, + VSTOR_OPERATION_RESET_BUS = 6, + VSTOR_OPERATION_BEGIN_INITIALIZATION = 7, + VSTOR_OPERATION_END_INITIALIZATION = 8, + VSTOR_OPERATION_QUERY_PROTOCOL_VERSION = 9, + VSTOR_OPERATION_QUERY_PROPERTIES = 10, + VSTOR_OPERATION_ENUMERATE_BUS = 11, + VSTOR_OPERATION_FCHBA_DATA = 12, + VSTOR_OPERATION_CREATE_SUB_CHANNELS = 13, + VSTOR_OPERATION_MAXIMUM = 13 +}; + +/* + * WWN packet for Fibre Channel HBA + */ + +struct hv_fc_wwn_packet { + u8 primary_active; + u8 reserved1[3]; + u8 primary_port_wwn[8]; + u8 primary_node_wwn[8]; + u8 secondary_port_wwn[8]; + u8 secondary_node_wwn[8]; +}; + + + +/* + * SRB Flag Bits + */ + +#define SRB_FLAGS_QUEUE_ACTION_ENABLE 0x00000002 +#define SRB_FLAGS_DISABLE_DISCONNECT 0x00000004 +#define SRB_FLAGS_DISABLE_SYNCH_TRANSFER 0x00000008 +#define SRB_FLAGS_BYPASS_FROZEN_QUEUE 0x00000010 +#define SRB_FLAGS_DISABLE_AUTOSENSE 0x00000020 +#define SRB_FLAGS_DATA_IN 0x00000040 +#define SRB_FLAGS_DATA_OUT 0x00000080 +#define SRB_FLAGS_NO_DATA_TRANSFER 0x00000000 +#define SRB_FLAGS_UNSPECIFIED_DIRECTION (SRB_FLAGS_DATA_IN | SRB_FLAGS_DATA_OUT) +#define SRB_FLAGS_NO_QUEUE_FREEZE 0x00000100 +#define SRB_FLAGS_ADAPTER_CACHE_ENABLE 0x00000200 +#define SRB_FLAGS_FREE_SENSE_BUFFER 0x00000400 + +/* + * This flag indicates the request is part of the workflow for processing a D3. + */ +#define SRB_FLAGS_D3_PROCESSING 0x00000800 +#define SRB_FLAGS_IS_ACTIVE 0x00010000 +#define SRB_FLAGS_ALLOCATED_FROM_ZONE 0x00020000 +#define SRB_FLAGS_SGLIST_FROM_POOL 0x00040000 +#define SRB_FLAGS_BYPASS_LOCKED_QUEUE 0x00080000 +#define SRB_FLAGS_NO_KEEP_AWAKE 0x00100000 +#define SRB_FLAGS_PORT_DRIVER_ALLOCSENSE 0x00200000 +#define SRB_FLAGS_PORT_DRIVER_SENSEHASPORT 0x00400000 +#define SRB_FLAGS_DONT_START_NEXT_PACKET 0x00800000 +#define SRB_FLAGS_PORT_DRIVER_RESERVED 0x0F000000 +#define SRB_FLAGS_CLASS_DRIVER_RESERVED 0xF0000000 + +#define SP_UNTAGGED ((unsigned char) ~0) +#define SRB_SIMPLE_TAG_REQUEST 0x20 + +/* + * Platform neutral description of a scsi request - + * this remains the same across the write regardless of 32/64 bit + * note: it's patterned off the SCSI_PASS_THROUGH structure + */ +#define STORVSC_MAX_CMD_LEN 0x10 + +#define POST_WIN7_STORVSC_SENSE_BUFFER_SIZE 0x14 +#define PRE_WIN8_STORVSC_SENSE_BUFFER_SIZE 0x12 + +#define STORVSC_SENSE_BUFFER_SIZE 0x14 +#define STORVSC_MAX_BUF_LEN_WITH_PADDING 0x14 + +/* + * Sense buffer size changed in win8; have a run-time + * variable to track the size we should use. This value will + * likely change during protocol negotiation but it is valid + * to start by assuming pre-Win8. + */ +static int sense_buffer_size = PRE_WIN8_STORVSC_SENSE_BUFFER_SIZE; + +/* + * The storage protocol version is determined during the + * initial exchange with the host. It will indicate which + * storage functionality is available in the host. +*/ +static int vmstor_proto_version; + +#define STORVSC_LOGGING_NONE 0 +#define STORVSC_LOGGING_ERROR 1 +#define STORVSC_LOGGING_WARN 2 + +static int logging_level = STORVSC_LOGGING_ERROR; +module_param(logging_level, int, S_IRUGO|S_IWUSR); +MODULE_PARM_DESC(logging_level, + "Logging level, 0 - None, 1 - Error (default), 2 - Warning."); + +static inline bool do_logging(int level) +{ + return logging_level >= level; +} + +#define storvsc_log(dev, level, fmt, ...) \ +do { \ + if (do_logging(level)) \ + dev_warn(&(dev)->device, fmt, ##__VA_ARGS__); \ +} while (0) + +struct vmscsi_win8_extension { + /* + * The following were added in Windows 8 + */ + u16 reserve; + u8 queue_tag; + u8 queue_action; + u32 srb_flags; + u32 time_out_value; + u32 queue_sort_ey; +} __packed; + +struct vmscsi_request { + u16 length; + u8 srb_status; + u8 scsi_status; + + u8 port_number; + u8 path_id; + u8 target_id; + u8 lun; + + u8 cdb_length; + u8 sense_info_length; + u8 data_in; + u8 reserved; + + u32 data_transfer_length; + + union { + u8 cdb[STORVSC_MAX_CMD_LEN]; + u8 sense_data[STORVSC_SENSE_BUFFER_SIZE]; + u8 reserved_array[STORVSC_MAX_BUF_LEN_WITH_PADDING]; + }; + /* + * The following was added in win8. + */ + struct vmscsi_win8_extension win8_extension; + +} __attribute((packed)); + + +/* + * The size of the vmscsi_request has changed in win8. The + * additional size is because of new elements added to the + * structure. These elements are valid only when we are talking + * to a win8 host. + * Track the correction to size we need to apply. This value + * will likely change during protocol negotiation but it is + * valid to start by assuming pre-Win8. + */ +static int vmscsi_size_delta = sizeof(struct vmscsi_win8_extension); + +/* + * The list of storage protocols in order of preference. + */ +struct vmstor_protocol { + int protocol_version; + int sense_buffer_size; + int vmscsi_size_delta; +}; + + +static const struct vmstor_protocol vmstor_protocols[] = { + { + VMSTOR_PROTO_VERSION_WIN10, + POST_WIN7_STORVSC_SENSE_BUFFER_SIZE, + 0 + }, + { + VMSTOR_PROTO_VERSION_WIN8_1, + POST_WIN7_STORVSC_SENSE_BUFFER_SIZE, + 0 + }, + { + VMSTOR_PROTO_VERSION_WIN8, + POST_WIN7_STORVSC_SENSE_BUFFER_SIZE, + 0 + }, + { + VMSTOR_PROTO_VERSION_WIN7, + PRE_WIN8_STORVSC_SENSE_BUFFER_SIZE, + sizeof(struct vmscsi_win8_extension), + }, + { + VMSTOR_PROTO_VERSION_WIN6, + PRE_WIN8_STORVSC_SENSE_BUFFER_SIZE, + sizeof(struct vmscsi_win8_extension), + } +}; + + +/* + * This structure is sent during the initialization phase to get the different + * properties of the channel. + */ + +#define STORAGE_CHANNEL_SUPPORTS_MULTI_CHANNEL 0x1 + +struct vmstorage_channel_properties { + u32 reserved; + u16 max_channel_cnt; + u16 reserved1; + + u32 flags; + u32 max_transfer_bytes; + + u64 reserved2; +} __packed; + +/* This structure is sent during the storage protocol negotiations. */ +struct vmstorage_protocol_version { + /* Major (MSW) and minor (LSW) version numbers. */ + u16 major_minor; + + /* + * Revision number is auto-incremented whenever this file is changed + * (See FILL_VMSTOR_REVISION macro above). Mismatch does not + * definitely indicate incompatibility--but it does indicate mismatched + * builds. + * This is only used on the windows side. Just set it to 0. + */ + u16 revision; +} __packed; + +/* Channel Property Flags */ +#define STORAGE_CHANNEL_REMOVABLE_FLAG 0x1 +#define STORAGE_CHANNEL_EMULATED_IDE_FLAG 0x2 + +struct vstor_packet { + /* Requested operation type */ + enum vstor_packet_operation operation; + + /* Flags - see below for values */ + u32 flags; + + /* Status of the request returned from the server side. */ + u32 status; + + /* Data payload area */ + union { + /* + * Structure used to forward SCSI commands from the + * client to the server. + */ + struct vmscsi_request vm_srb; + + /* Structure used to query channel properties. */ + struct vmstorage_channel_properties storage_channel_properties; + + /* Used during version negotiations. */ + struct vmstorage_protocol_version version; + + /* Fibre channel address packet */ + struct hv_fc_wwn_packet wwn_packet; + + /* Number of sub-channels to create */ + u16 sub_channel_count; + + /* This will be the maximum of the union members */ + u8 buffer[0x34]; + }; +} __packed; + +/* + * Packet Flags: + * + * This flag indicates that the server should send back a completion for this + * packet. + */ + +#define REQUEST_COMPLETION_FLAG 0x1 + +/* Matches Windows-end */ +enum storvsc_request_type { + WRITE_TYPE = 0, + READ_TYPE, + UNKNOWN_TYPE, +}; + +/* + * SRB status codes and masks. In the 8-bit field, the two high order bits + * are flags, while the remaining 6 bits are an integer status code. The + * definitions here include only the subset of the integer status codes that + * are tested for in this driver. + */ +#define SRB_STATUS_AUTOSENSE_VALID 0x80 +#define SRB_STATUS_QUEUE_FROZEN 0x40 + +/* SRB status integer codes */ +#define SRB_STATUS_SUCCESS 0x01 +#define SRB_STATUS_ABORTED 0x02 +#define SRB_STATUS_ERROR 0x04 +#define SRB_STATUS_INVALID_REQUEST 0x06 +#define SRB_STATUS_DATA_OVERRUN 0x12 +#define SRB_STATUS_INVALID_LUN 0x20 + +#define SRB_STATUS(status) \ + (status & ~(SRB_STATUS_AUTOSENSE_VALID | SRB_STATUS_QUEUE_FROZEN)) +/* + * This is the end of Protocol specific defines. + */ + +static int storvsc_ringbuffer_size = (128 * 1024); +static u32 max_outstanding_req_per_channel; +static int storvsc_change_queue_depth(struct scsi_device *sdev, int queue_depth); + +static int storvsc_vcpus_per_sub_channel = 4; + +module_param(storvsc_ringbuffer_size, int, S_IRUGO); +MODULE_PARM_DESC(storvsc_ringbuffer_size, "Ring buffer size (bytes)"); + +module_param(storvsc_vcpus_per_sub_channel, int, S_IRUGO); +MODULE_PARM_DESC(storvsc_vcpus_per_sub_channel, "Ratio of VCPUs to subchannels"); + +static int ring_avail_percent_lowater = 10; +module_param(ring_avail_percent_lowater, int, S_IRUGO); +MODULE_PARM_DESC(ring_avail_percent_lowater, + "Select a channel if available ring size > this in percent"); + +/* + * Timeout in seconds for all devices managed by this driver. + */ +static int storvsc_timeout = 180; + +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) +static struct scsi_transport_template *fc_transport_template; +#endif + +static void storvsc_on_channel_callback(void *context); + +#define STORVSC_MAX_LUNS_PER_TARGET 255 +#define STORVSC_MAX_TARGETS 2 +#define STORVSC_MAX_CHANNELS 8 + +#define STORVSC_FC_MAX_LUNS_PER_TARGET 255 +#define STORVSC_FC_MAX_TARGETS 128 +#define STORVSC_FC_MAX_CHANNELS 8 + +#define STORVSC_IDE_MAX_LUNS_PER_TARGET 64 +#define STORVSC_IDE_MAX_TARGETS 1 +#define STORVSC_IDE_MAX_CHANNELS 1 + +struct storvsc_cmd_request { + struct scsi_cmnd *cmd; + + struct hv_device *device; + + /* Synchronize the request/response if needed */ + struct completion wait_event; + + struct vmbus_channel_packet_multipage_buffer mpb; + struct vmbus_packet_mpb_array *payload; + u32 payload_sz; + + struct vstor_packet vstor_packet; +}; + + +/* A storvsc device is a device object that contains a vmbus channel */ +struct storvsc_device { + struct hv_device *device; + + bool destroy; + bool drain_notify; + atomic_t num_outstanding_req; + struct Scsi_Host *host; + + wait_queue_head_t waiting_to_drain; + + /* + * Each unique Port/Path/Target represents 1 channel ie scsi + * controller. In reality, the pathid, targetid is always 0 + * and the port is set by us + */ + unsigned int port_number; + unsigned char path_id; + unsigned char target_id; + + /* + * Max I/O, the device can support. + */ + u32 max_transfer_bytes; + /* + * Number of sub-channels we will open. + */ + u16 num_sc; + struct vmbus_channel **stor_chns; + /* + * Mask of CPUs bound to subchannels. + */ + struct cpumask alloced_cpus; + /* + * Serializes modifications of stor_chns[] from storvsc_do_io() + * and storvsc_change_target_cpu(). + */ + spinlock_t lock; + /* Used for vsc/vsp channel reset process */ + struct storvsc_cmd_request init_request; + struct storvsc_cmd_request reset_request; + /* + * Currently active port and node names for FC devices. + */ + u64 node_name; + u64 port_name; +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + struct fc_rport *rport; +#endif +}; + +struct hv_host_device { + struct hv_device *dev; + unsigned int port; + unsigned char path; + unsigned char target; + struct workqueue_struct *handle_error_wq; + struct work_struct host_scan_work; + struct Scsi_Host *host; +}; + +struct storvsc_scan_work { + struct work_struct work; + struct Scsi_Host *host; + u8 lun; + u8 tgt_id; +}; + +static void storvsc_device_scan(struct work_struct *work) +{ + struct storvsc_scan_work *wrk; + struct scsi_device *sdev; + + wrk = container_of(work, struct storvsc_scan_work, work); + + sdev = scsi_device_lookup(wrk->host, 0, wrk->tgt_id, wrk->lun); + if (!sdev) + goto done; + scsi_rescan_device(&sdev->sdev_gendev); + scsi_device_put(sdev); + +done: + kfree(wrk); +} + +static void storvsc_host_scan(struct work_struct *work) +{ + struct Scsi_Host *host; + struct scsi_device *sdev; + struct hv_host_device *host_device = + container_of(work, struct hv_host_device, host_scan_work); + + host = host_device->host; + /* + * Before scanning the host, first check to see if any of the + * currrently known devices have been hot removed. We issue a + * "unit ready" command against all currently known devices. + * This I/O will result in an error for devices that have been + * removed. As part of handling the I/O error, we remove the device. + * + * When a LUN is added or removed, the host sends us a signal to + * scan the host. Thus we are forced to discover the LUNs that + * may have been removed this way. + */ + mutex_lock(&host->scan_mutex); + shost_for_each_device(sdev, host) + scsi_test_unit_ready(sdev, 1, 1, NULL); + mutex_unlock(&host->scan_mutex); + /* + * Now scan the host to discover LUNs that may have been added. + */ + scsi_scan_host(host); +} + +static void storvsc_remove_lun(struct work_struct *work) +{ + struct storvsc_scan_work *wrk; + struct scsi_device *sdev; + + wrk = container_of(work, struct storvsc_scan_work, work); + if (!scsi_host_get(wrk->host)) + goto done; + + sdev = scsi_device_lookup(wrk->host, 0, wrk->tgt_id, wrk->lun); + + if (sdev) { + scsi_remove_device(sdev); + scsi_device_put(sdev); + } + scsi_host_put(wrk->host); + +done: + kfree(wrk); +} + + +/* + * We can get incoming messages from the host that are not in response to + * messages that we have sent out. An example of this would be messages + * received by the guest to notify dynamic addition/removal of LUNs. To + * deal with potential race conditions where the driver may be in the + * midst of being unloaded when we might receive an unsolicited message + * from the host, we have implemented a mechanism to gurantee sequential + * consistency: + * + * 1) Once the device is marked as being destroyed, we will fail all + * outgoing messages. + * 2) We permit incoming messages when the device is being destroyed, + * only to properly account for messages already sent out. + */ + +static inline struct storvsc_device *get_out_stor_device( + struct hv_device *device) +{ + struct storvsc_device *stor_device; + + stor_device = hv_get_drvdata(device); + + if (stor_device && stor_device->destroy) + stor_device = NULL; + + return stor_device; +} + + +static inline void storvsc_wait_to_drain(struct storvsc_device *dev) +{ + dev->drain_notify = true; + wait_event(dev->waiting_to_drain, + atomic_read(&dev->num_outstanding_req) == 0); + dev->drain_notify = false; +} + +static inline struct storvsc_device *get_in_stor_device( + struct hv_device *device) +{ + struct storvsc_device *stor_device; + + stor_device = hv_get_drvdata(device); + + if (!stor_device) + goto get_in_err; + + /* + * If the device is being destroyed; allow incoming + * traffic only to cleanup outstanding requests. + */ + + if (stor_device->destroy && + (atomic_read(&stor_device->num_outstanding_req) == 0)) + stor_device = NULL; + +get_in_err: + return stor_device; + +} + +static void storvsc_change_target_cpu(struct vmbus_channel *channel, u32 old, + u32 new) +{ + struct storvsc_device *stor_device; + struct vmbus_channel *cur_chn; + bool old_is_alloced = false; + struct hv_device *device; + unsigned long flags; + int cpu; + + device = channel->primary_channel ? + channel->primary_channel->device_obj + : channel->device_obj; + stor_device = get_out_stor_device(device); + if (!stor_device) + return; + + /* See storvsc_do_io() -> get_og_chn(). */ + spin_lock_irqsave(&stor_device->lock, flags); + + /* + * Determines if the storvsc device has other channels assigned to + * the "old" CPU to update the alloced_cpus mask and the stor_chns + * array. + */ + if (device->channel != channel && device->channel->target_cpu == old) { + cur_chn = device->channel; + old_is_alloced = true; + goto old_is_alloced; + } + list_for_each_entry(cur_chn, &device->channel->sc_list, sc_list) { + if (cur_chn == channel) + continue; + if (cur_chn->target_cpu == old) { + old_is_alloced = true; + goto old_is_alloced; + } + } + +old_is_alloced: + if (old_is_alloced) + WRITE_ONCE(stor_device->stor_chns[old], cur_chn); + else + cpumask_clear_cpu(old, &stor_device->alloced_cpus); + + /* "Flush" the stor_chns array. */ + for_each_possible_cpu(cpu) { + if (stor_device->stor_chns[cpu] && !cpumask_test_cpu( + cpu, &stor_device->alloced_cpus)) + WRITE_ONCE(stor_device->stor_chns[cpu], NULL); + } + + WRITE_ONCE(stor_device->stor_chns[new], channel); + cpumask_set_cpu(new, &stor_device->alloced_cpus); + + spin_unlock_irqrestore(&stor_device->lock, flags); +} + +static void handle_sc_creation(struct vmbus_channel *new_sc) +{ + struct hv_device *device = new_sc->primary_channel->device_obj; + struct device *dev = &device->device; + struct storvsc_device *stor_device; + struct vmstorage_channel_properties props; + int ret; + + stor_device = get_out_stor_device(device); + if (!stor_device) + return; + + memset(&props, 0, sizeof(struct vmstorage_channel_properties)); + + ret = vmbus_open(new_sc, + storvsc_ringbuffer_size, + storvsc_ringbuffer_size, + (void *)&props, + sizeof(struct vmstorage_channel_properties), + storvsc_on_channel_callback, new_sc); + + /* In case vmbus_open() fails, we don't use the sub-channel. */ + if (ret != 0) { + dev_err(dev, "Failed to open sub-channel: err=%d\n", ret); + return; + } + + new_sc->change_target_cpu_callback = storvsc_change_target_cpu; + + /* Add the sub-channel to the array of available channels. */ + stor_device->stor_chns[new_sc->target_cpu] = new_sc; + cpumask_set_cpu(new_sc->target_cpu, &stor_device->alloced_cpus); +} + +static void handle_multichannel_storage(struct hv_device *device, int max_chns) +{ + struct device *dev = &device->device; + struct storvsc_device *stor_device; + int num_sc; + struct storvsc_cmd_request *request; + struct vstor_packet *vstor_packet; + int ret, t; + + /* + * If the number of CPUs is artificially restricted, such as + * with maxcpus=1 on the kernel boot line, Hyper-V could offer + * sub-channels >= the number of CPUs. These sub-channels + * should not be created. The primary channel is already created + * and assigned to one CPU, so check against # CPUs - 1. + */ + num_sc = min((int)(num_online_cpus() - 1), max_chns); + if (!num_sc) + return; + + stor_device = get_out_stor_device(device); + if (!stor_device) + return; + + stor_device->num_sc = num_sc; + request = &stor_device->init_request; + vstor_packet = &request->vstor_packet; + + /* + * Establish a handler for dealing with subchannels. + */ + vmbus_set_sc_create_callback(device->channel, handle_sc_creation); + + /* + * Request the host to create sub-channels. + */ + memset(request, 0, sizeof(struct storvsc_cmd_request)); + init_completion(&request->wait_event); + vstor_packet->operation = VSTOR_OPERATION_CREATE_SUB_CHANNELS; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + vstor_packet->sub_channel_count = num_sc; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + (sizeof(struct vstor_packet) - + vmscsi_size_delta), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + + if (ret != 0) { + dev_err(dev, "Failed to create sub-channel: err=%d\n", ret); + return; + } + + t = wait_for_completion_timeout(&request->wait_event, 10*HZ); + if (t == 0) { + dev_err(dev, "Failed to create sub-channel: timed out\n"); + return; + } + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) { + dev_err(dev, "Failed to create sub-channel: op=%d, sts=%d\n", + vstor_packet->operation, vstor_packet->status); + return; + } + + /* + * We need to do nothing here, because vmbus_process_offer() + * invokes channel->sc_creation_callback, which will open and use + * the sub-channel(s). + */ +} + +static void cache_wwn(struct storvsc_device *stor_device, + struct vstor_packet *vstor_packet) +{ + /* + * Cache the currently active port and node ww names. + */ + if (vstor_packet->wwn_packet.primary_active) { + stor_device->node_name = + wwn_to_u64(vstor_packet->wwn_packet.primary_node_wwn); + stor_device->port_name = + wwn_to_u64(vstor_packet->wwn_packet.primary_port_wwn); + } else { + stor_device->node_name = + wwn_to_u64(vstor_packet->wwn_packet.secondary_node_wwn); + stor_device->port_name = + wwn_to_u64(vstor_packet->wwn_packet.secondary_port_wwn); + } +} + + +static int storvsc_execute_vstor_op(struct hv_device *device, + struct storvsc_cmd_request *request, + bool status_check) +{ + struct vstor_packet *vstor_packet; + int ret, t; + + vstor_packet = &request->vstor_packet; + + init_completion(&request->wait_event); + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + (sizeof(struct vstor_packet) - + vmscsi_size_delta), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret != 0) + return ret; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) + return -ETIMEDOUT; + + if (!status_check) + return ret; + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO || + vstor_packet->status != 0) + return -EINVAL; + + return ret; +} + +static int storvsc_channel_init(struct hv_device *device, bool is_fc) +{ + struct storvsc_device *stor_device; + struct storvsc_cmd_request *request; + struct vstor_packet *vstor_packet; + int ret, i; + int max_chns; + bool process_sub_channels = false; + + stor_device = get_out_stor_device(device); + if (!stor_device) + return -ENODEV; + + request = &stor_device->init_request; + vstor_packet = &request->vstor_packet; + + /* + * Now, initiate the vsc/vsp initialization protocol on the open + * channel + */ + memset(request, 0, sizeof(struct storvsc_cmd_request)); + vstor_packet->operation = VSTOR_OPERATION_BEGIN_INITIALIZATION; + ret = storvsc_execute_vstor_op(device, request, true); + if (ret) + return ret; + /* + * Query host supported protocol version. + */ + + for (i = 0; i < ARRAY_SIZE(vmstor_protocols); i++) { + /* reuse the packet for version range supported */ + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = + VSTOR_OPERATION_QUERY_PROTOCOL_VERSION; + + vstor_packet->version.major_minor = + vmstor_protocols[i].protocol_version; + + /* + * The revision number is only used in Windows; set it to 0. + */ + vstor_packet->version.revision = 0; + ret = storvsc_execute_vstor_op(device, request, false); + if (ret != 0) + return ret; + + if (vstor_packet->operation != VSTOR_OPERATION_COMPLETE_IO) + return -EINVAL; + + if (vstor_packet->status == 0) { + vmstor_proto_version = + vmstor_protocols[i].protocol_version; + + sense_buffer_size = + vmstor_protocols[i].sense_buffer_size; + + vmscsi_size_delta = + vmstor_protocols[i].vmscsi_size_delta; + + break; + } + } + + if (vstor_packet->status != 0) + return -EINVAL; + + + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_QUERY_PROPERTIES; + ret = storvsc_execute_vstor_op(device, request, true); + if (ret != 0) + return ret; + + /* + * Check to see if multi-channel support is there. + * Hosts that implement protocol version of 5.1 and above + * support multi-channel. + */ + max_chns = vstor_packet->storage_channel_properties.max_channel_cnt; + + /* + * Allocate state to manage the sub-channels. + * We allocate an array based on the numbers of possible CPUs + * (Hyper-V does not support cpu online/offline). + * This Array will be sparseley populated with unique + * channels - primary + sub-channels. + * We will however populate all the slots to evenly distribute + * the load. + */ + stor_device->stor_chns = kcalloc(num_possible_cpus(), sizeof(void *), + GFP_KERNEL); + if (stor_device->stor_chns == NULL) + return -ENOMEM; + + device->channel->change_target_cpu_callback = storvsc_change_target_cpu; + + stor_device->stor_chns[device->channel->target_cpu] = device->channel; + cpumask_set_cpu(device->channel->target_cpu, + &stor_device->alloced_cpus); + + if (vmstor_proto_version >= VMSTOR_PROTO_VERSION_WIN8) { + if (vstor_packet->storage_channel_properties.flags & + STORAGE_CHANNEL_SUPPORTS_MULTI_CHANNEL) + process_sub_channels = true; + } + stor_device->max_transfer_bytes = + vstor_packet->storage_channel_properties.max_transfer_bytes; + + if (!is_fc) + goto done; + + /* + * For FC devices retrieve FC HBA data. + */ + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_FCHBA_DATA; + ret = storvsc_execute_vstor_op(device, request, true); + if (ret != 0) + return ret; + + /* + * Cache the currently active port and node ww names. + */ + cache_wwn(stor_device, vstor_packet); + +done: + + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + vstor_packet->operation = VSTOR_OPERATION_END_INITIALIZATION; + ret = storvsc_execute_vstor_op(device, request, true); + if (ret != 0) + return ret; + + if (process_sub_channels) + handle_multichannel_storage(device, max_chns); + + return ret; +} + +static void storvsc_handle_error(struct vmscsi_request *vm_srb, + struct scsi_cmnd *scmnd, + struct Scsi_Host *host, + u8 asc, u8 ascq) +{ + struct storvsc_scan_work *wrk; + void (*process_err_fn)(struct work_struct *work); + struct hv_host_device *host_dev = shost_priv(host); + + switch (SRB_STATUS(vm_srb->srb_status)) { + case SRB_STATUS_ERROR: + case SRB_STATUS_ABORTED: + case SRB_STATUS_INVALID_REQUEST: + if (vm_srb->srb_status & SRB_STATUS_AUTOSENSE_VALID) { + /* Check for capacity change */ + if ((asc == 0x2a) && (ascq == 0x9)) { + process_err_fn = storvsc_device_scan; + /* Retry the I/O that triggered this. */ + set_host_byte(scmnd, DID_REQUEUE); + goto do_work; + } + + /* + * Check for "Operating parameters have changed" + * due to Hyper-V changing the VHD/VHDX BlockSize + * when adding/removing a differencing disk. This + * causes discard_granularity to change, so do a + * rescan to pick up the new granularity. We don't + * want scsi_report_sense() to output a message + * that a sysadmin wouldn't know what to do with. + */ + if ((asc == 0x3f) && (ascq != 0x03) && + (ascq != 0x0e)) { + process_err_fn = storvsc_device_scan; + set_host_byte(scmnd, DID_REQUEUE); + goto do_work; + } + + /* + * Otherwise, let upper layer deal with the + * error when sense message is present + */ + return; + } + + /* + * If there is an error; offline the device since all + * error recovery strategies would have already been + * deployed on the host side. However, if the command + * were a pass-through command deal with it appropriately. + */ + switch (scmnd->cmnd[0]) { + case ATA_16: + case ATA_12: + set_host_byte(scmnd, DID_PASSTHROUGH); + break; + /* + * On some Hyper-V hosts TEST_UNIT_READY command can + * return SRB_STATUS_ERROR. Let the upper level code + * deal with it based on the sense information. + */ + case TEST_UNIT_READY: + break; + default: + set_host_byte(scmnd, DID_ERROR); + } + return; + + case SRB_STATUS_INVALID_LUN: + set_host_byte(scmnd, DID_NO_CONNECT); + process_err_fn = storvsc_remove_lun; + goto do_work; + + } + return; + +do_work: + /* + * We need to schedule work to process this error; schedule it. + */ + wrk = kmalloc(sizeof(struct storvsc_scan_work), GFP_ATOMIC); + if (!wrk) { + set_host_byte(scmnd, DID_TARGET_FAILURE); + return; + } + + wrk->host = host; + wrk->lun = vm_srb->lun; + wrk->tgt_id = vm_srb->target_id; + INIT_WORK(&wrk->work, process_err_fn); + queue_work(host_dev->handle_error_wq, &wrk->work); +} + + +static void storvsc_command_completion(struct storvsc_cmd_request *cmd_request, + struct storvsc_device *stor_dev) +{ + struct scsi_cmnd *scmnd = cmd_request->cmd; + struct scsi_sense_hdr sense_hdr; + struct vmscsi_request *vm_srb; + u32 data_transfer_length; + struct Scsi_Host *host; + u32 payload_sz = cmd_request->payload_sz; + void *payload = cmd_request->payload; + + host = stor_dev->host; + + vm_srb = &cmd_request->vstor_packet.vm_srb; + data_transfer_length = vm_srb->data_transfer_length; + + scmnd->result = vm_srb->scsi_status; + + if (scmnd->result) { + if (scsi_normalize_sense(scmnd->sense_buffer, + SCSI_SENSE_BUFFERSIZE, &sense_hdr) && + !(sense_hdr.sense_key == NOT_READY && + sense_hdr.asc == 0x03A) && + do_logging(STORVSC_LOGGING_ERROR)) + scsi_print_sense_hdr(scmnd->device, "storvsc", + &sense_hdr); + } + + if (vm_srb->srb_status != SRB_STATUS_SUCCESS) { + storvsc_handle_error(vm_srb, scmnd, host, sense_hdr.asc, + sense_hdr.ascq); + /* + * The Windows driver set data_transfer_length on + * SRB_STATUS_DATA_OVERRUN. On other errors, this value + * is untouched. In these cases we set it to 0. + */ + if (vm_srb->srb_status != SRB_STATUS_DATA_OVERRUN) + data_transfer_length = 0; + } + + /* Validate data_transfer_length (from Hyper-V) */ + if (data_transfer_length > cmd_request->payload->range.len) + data_transfer_length = cmd_request->payload->range.len; + + scsi_set_resid(scmnd, + cmd_request->payload->range.len - data_transfer_length); + + scmnd->scsi_done(scmnd); + + if (payload_sz > + sizeof(struct vmbus_channel_packet_multipage_buffer)) + kfree(payload); +} + +static void storvsc_on_io_completion(struct storvsc_device *stor_device, + struct vstor_packet *vstor_packet, + struct storvsc_cmd_request *request) +{ + struct vstor_packet *stor_pkt; + struct hv_device *device = stor_device->device; + + stor_pkt = &request->vstor_packet; + + /* + * The current SCSI handling on the host side does + * not correctly handle: + * INQUIRY command with page code parameter set to 0x80 + * MODE_SENSE command with cmd[2] == 0x1c + * + * Setup srb and scsi status so this won't be fatal. + * We do this so we can distinguish truly fatal failues + * (srb status == 0x4) and off-line the device in that case. + */ + + if ((stor_pkt->vm_srb.cdb[0] == INQUIRY) || + (stor_pkt->vm_srb.cdb[0] == MODE_SENSE)) { + vstor_packet->vm_srb.scsi_status = 0; + vstor_packet->vm_srb.srb_status = SRB_STATUS_SUCCESS; + } + + + /* Copy over the status...etc */ + stor_pkt->vm_srb.scsi_status = vstor_packet->vm_srb.scsi_status; + stor_pkt->vm_srb.srb_status = vstor_packet->vm_srb.srb_status; + + /* Validate sense_info_length (from Hyper-V) */ + if (vstor_packet->vm_srb.sense_info_length > sense_buffer_size) + vstor_packet->vm_srb.sense_info_length = sense_buffer_size; + + stor_pkt->vm_srb.sense_info_length = + vstor_packet->vm_srb.sense_info_length; + + if (vstor_packet->vm_srb.scsi_status != 0 || + vstor_packet->vm_srb.srb_status != SRB_STATUS_SUCCESS) + storvsc_log(device, STORVSC_LOGGING_WARN, + "cmd 0x%x scsi status 0x%x srb status 0x%x\n", + stor_pkt->vm_srb.cdb[0], + vstor_packet->vm_srb.scsi_status, + vstor_packet->vm_srb.srb_status); + + if ((vstor_packet->vm_srb.scsi_status & 0xFF) == 0x02) { + /* CHECK_CONDITION */ + if (vstor_packet->vm_srb.srb_status & + SRB_STATUS_AUTOSENSE_VALID) { + /* autosense data available */ + + storvsc_log(device, STORVSC_LOGGING_WARN, + "stor pkt %p autosense data valid - len %d\n", + request, vstor_packet->vm_srb.sense_info_length); + + memcpy(request->cmd->sense_buffer, + vstor_packet->vm_srb.sense_data, + vstor_packet->vm_srb.sense_info_length); + + } + } + + stor_pkt->vm_srb.data_transfer_length = + vstor_packet->vm_srb.data_transfer_length; + + storvsc_command_completion(request, stor_device); + + if (atomic_dec_and_test(&stor_device->num_outstanding_req) && + stor_device->drain_notify) + wake_up(&stor_device->waiting_to_drain); + + +} + +static void storvsc_on_receive(struct storvsc_device *stor_device, + struct vstor_packet *vstor_packet, + struct storvsc_cmd_request *request) +{ + struct hv_host_device *host_dev; + switch (vstor_packet->operation) { + case VSTOR_OPERATION_COMPLETE_IO: + storvsc_on_io_completion(stor_device, vstor_packet, request); + break; + + case VSTOR_OPERATION_REMOVE_DEVICE: + case VSTOR_OPERATION_ENUMERATE_BUS: + host_dev = shost_priv(stor_device->host); + queue_work( + host_dev->handle_error_wq, &host_dev->host_scan_work); + break; + + case VSTOR_OPERATION_FCHBA_DATA: + cache_wwn(stor_device, vstor_packet); +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + fc_host_node_name(stor_device->host) = stor_device->node_name; + fc_host_port_name(stor_device->host) = stor_device->port_name; +#endif + break; + default: + break; + } +} + +static void storvsc_on_channel_callback(void *context) +{ + struct vmbus_channel *channel = (struct vmbus_channel *)context; + const struct vmpacket_descriptor *desc; + struct hv_device *device; + struct storvsc_device *stor_device; + + if (channel->primary_channel != NULL) + device = channel->primary_channel->device_obj; + else + device = channel->device_obj; + + stor_device = get_in_stor_device(device); + if (!stor_device) + return; + + foreach_vmbus_pkt(desc, channel) { + void *packet = hv_pkt_data(desc); + struct storvsc_cmd_request *request; + + request = (struct storvsc_cmd_request *) + ((unsigned long)desc->trans_id); + + if (request == &stor_device->init_request || + request == &stor_device->reset_request) { + memcpy(&request->vstor_packet, packet, + (sizeof(struct vstor_packet) - vmscsi_size_delta)); + complete(&request->wait_event); + } else { + storvsc_on_receive(stor_device, packet, request); + } + } +} + +static int storvsc_connect_to_vsp(struct hv_device *device, u32 ring_size, + bool is_fc) +{ + struct vmstorage_channel_properties props; + int ret; + + memset(&props, 0, sizeof(struct vmstorage_channel_properties)); + + ret = vmbus_open(device->channel, + ring_size, + ring_size, + (void *)&props, + sizeof(struct vmstorage_channel_properties), + storvsc_on_channel_callback, device->channel); + + if (ret != 0) + return ret; + + ret = storvsc_channel_init(device, is_fc); + + return ret; +} + +static int storvsc_dev_remove(struct hv_device *device) +{ + struct storvsc_device *stor_device; + + stor_device = hv_get_drvdata(device); + + stor_device->destroy = true; + + /* Make sure flag is set before waiting */ + wmb(); + + /* + * At this point, all outbound traffic should be disable. We + * only allow inbound traffic (responses) to proceed so that + * outstanding requests can be completed. + */ + + storvsc_wait_to_drain(stor_device); + + /* + * Since we have already drained, we don't need to busy wait + * as was done in final_release_stor_device() + * Note that we cannot set the ext pointer to NULL until + * we have drained - to drain the outgoing packets, we need to + * allow incoming packets. + */ + hv_set_drvdata(device, NULL); + + /* Close the channel */ + vmbus_close(device->channel); + + kfree(stor_device->stor_chns); + kfree(stor_device); + return 0; +} + +static struct vmbus_channel *get_og_chn(struct storvsc_device *stor_device, + u16 q_num) +{ + u16 slot = 0; + u16 hash_qnum; + const struct cpumask *node_mask; + int num_channels, tgt_cpu; + + if (stor_device->num_sc == 0) { + stor_device->stor_chns[q_num] = stor_device->device->channel; + return stor_device->device->channel; + } + + /* + * Our channel array is sparsley populated and we + * initiated I/O on a processor/hw-q that does not + * currently have a designated channel. Fix this. + * The strategy is simple: + * I. Ensure NUMA locality + * II. Distribute evenly (best effort) + */ + + node_mask = cpumask_of_node(cpu_to_node(q_num)); + + num_channels = 0; + for_each_cpu(tgt_cpu, &stor_device->alloced_cpus) { + if (cpumask_test_cpu(tgt_cpu, node_mask)) + num_channels++; + } + if (num_channels == 0) { + stor_device->stor_chns[q_num] = stor_device->device->channel; + return stor_device->device->channel; + } + + hash_qnum = q_num; + while (hash_qnum >= num_channels) + hash_qnum -= num_channels; + + for_each_cpu(tgt_cpu, &stor_device->alloced_cpus) { + if (!cpumask_test_cpu(tgt_cpu, node_mask)) + continue; + if (slot == hash_qnum) + break; + slot++; + } + + stor_device->stor_chns[q_num] = stor_device->stor_chns[tgt_cpu]; + + return stor_device->stor_chns[q_num]; +} + + +static int storvsc_do_io(struct hv_device *device, + struct storvsc_cmd_request *request, u16 q_num) +{ + struct storvsc_device *stor_device; + struct vstor_packet *vstor_packet; + struct vmbus_channel *outgoing_channel, *channel; + unsigned long flags; + int ret = 0; + const struct cpumask *node_mask; + int tgt_cpu; + + vstor_packet = &request->vstor_packet; + stor_device = get_out_stor_device(device); + + if (!stor_device) + return -ENODEV; + + + request->device = device; + /* + * Select an appropriate channel to send the request out. + */ + /* See storvsc_change_target_cpu(). */ + outgoing_channel = READ_ONCE(stor_device->stor_chns[q_num]); + if (outgoing_channel != NULL) { + if (outgoing_channel->target_cpu == q_num) { + /* + * Ideally, we want to pick a different channel if + * available on the same NUMA node. + */ + node_mask = cpumask_of_node(cpu_to_node(q_num)); + for_each_cpu_wrap(tgt_cpu, + &stor_device->alloced_cpus, q_num + 1) { + if (!cpumask_test_cpu(tgt_cpu, node_mask)) + continue; + if (tgt_cpu == q_num) + continue; + channel = READ_ONCE( + stor_device->stor_chns[tgt_cpu]); + if (channel == NULL) + continue; + if (hv_get_avail_to_write_percent( + &channel->outbound) + > ring_avail_percent_lowater) { + outgoing_channel = channel; + goto found_channel; + } + } + + /* + * All the other channels on the same NUMA node are + * busy. Try to use the channel on the current CPU + */ + if (hv_get_avail_to_write_percent( + &outgoing_channel->outbound) + > ring_avail_percent_lowater) + goto found_channel; + + /* + * If we reach here, all the channels on the current + * NUMA node are busy. Try to find a channel in + * other NUMA nodes + */ + for_each_cpu(tgt_cpu, &stor_device->alloced_cpus) { + if (cpumask_test_cpu(tgt_cpu, node_mask)) + continue; + channel = READ_ONCE( + stor_device->stor_chns[tgt_cpu]); + if (channel == NULL) + continue; + if (hv_get_avail_to_write_percent( + &channel->outbound) + > ring_avail_percent_lowater) { + outgoing_channel = channel; + goto found_channel; + } + } + } + } else { + spin_lock_irqsave(&stor_device->lock, flags); + outgoing_channel = stor_device->stor_chns[q_num]; + if (outgoing_channel != NULL) { + spin_unlock_irqrestore(&stor_device->lock, flags); + goto found_channel; + } + outgoing_channel = get_og_chn(stor_device, q_num); + spin_unlock_irqrestore(&stor_device->lock, flags); + } + +found_channel: + vstor_packet->flags |= REQUEST_COMPLETION_FLAG; + + vstor_packet->vm_srb.length = (sizeof(struct vmscsi_request) - + vmscsi_size_delta); + + + vstor_packet->vm_srb.sense_info_length = sense_buffer_size; + + + vstor_packet->vm_srb.data_transfer_length = + request->payload->range.len; + + vstor_packet->operation = VSTOR_OPERATION_EXECUTE_SRB; + + if (request->payload->range.len) { + + ret = vmbus_sendpacket_mpb_desc(outgoing_channel, + request->payload, request->payload_sz, + vstor_packet, + (sizeof(struct vstor_packet) - + vmscsi_size_delta), + (unsigned long)request); + } else { + ret = vmbus_sendpacket(outgoing_channel, vstor_packet, + (sizeof(struct vstor_packet) - + vmscsi_size_delta), + (unsigned long)request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + } + + if (ret != 0) + return ret; + + atomic_inc(&stor_device->num_outstanding_req); + + return ret; +} + +static int storvsc_device_alloc(struct scsi_device *sdevice) +{ + /* + * Set blist flag to permit the reading of the VPD pages even when + * the target may claim SPC-2 compliance. MSFT targets currently + * claim SPC-2 compliance while they implement post SPC-2 features. + * With this flag we can correctly handle WRITE_SAME_16 issues. + * + * Hypervisor reports SCSI_UNKNOWN type for DVD ROM device but + * still supports REPORT LUN. + */ + sdevice->sdev_bflags = BLIST_REPORTLUN2 | BLIST_TRY_VPD_PAGES; + + return 0; +} + +static int storvsc_device_configure(struct scsi_device *sdevice) +{ + blk_queue_rq_timeout(sdevice->request_queue, (storvsc_timeout * HZ)); + + /* storvsc devices don't support MAINTENANCE_IN SCSI cmd */ + sdevice->no_report_opcodes = 1; + sdevice->no_write_same = 1; + + /* + * If the host is WIN8 or WIN8 R2, claim conformance to SPC-3 + * if the device is a MSFT virtual device. If the host is + * WIN10 or newer, allow write_same. + */ + if (!strncmp(sdevice->vendor, "Msft", 4)) { + switch (vmstor_proto_version) { + case VMSTOR_PROTO_VERSION_WIN8: + case VMSTOR_PROTO_VERSION_WIN8_1: + sdevice->scsi_level = SCSI_SPC_3; + break; + } + + if (vmstor_proto_version >= VMSTOR_PROTO_VERSION_WIN10) + sdevice->no_write_same = 0; + } + + return 0; +} + +static int storvsc_get_chs(struct scsi_device *sdev, struct block_device * bdev, + sector_t capacity, int *info) +{ + sector_t nsect = capacity; + sector_t cylinders = nsect; + int heads, sectors_pt; + + /* + * We are making up these values; let us keep it simple. + */ + heads = 0xff; + sectors_pt = 0x3f; /* Sectors per track */ + sector_div(cylinders, heads * sectors_pt); + if ((sector_t)(cylinders + 1) * heads * sectors_pt < nsect) + cylinders = 0xffff; + + info[0] = heads; + info[1] = sectors_pt; + info[2] = (int)cylinders; + + return 0; +} + +static int storvsc_host_reset_handler(struct scsi_cmnd *scmnd) +{ + struct hv_host_device *host_dev = shost_priv(scmnd->device->host); + struct hv_device *device = host_dev->dev; + + struct storvsc_device *stor_device; + struct storvsc_cmd_request *request; + struct vstor_packet *vstor_packet; + int ret, t; + + + stor_device = get_out_stor_device(device); + if (!stor_device) + return FAILED; + + request = &stor_device->reset_request; + vstor_packet = &request->vstor_packet; + memset(vstor_packet, 0, sizeof(struct vstor_packet)); + + init_completion(&request->wait_event); + + vstor_packet->operation = VSTOR_OPERATION_RESET_BUS; + vstor_packet->flags = REQUEST_COMPLETION_FLAG; + vstor_packet->vm_srb.path_id = stor_device->path_id; + + ret = vmbus_sendpacket(device->channel, vstor_packet, + (sizeof(struct vstor_packet) - + vmscsi_size_delta), + (unsigned long)&stor_device->reset_request, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret != 0) + return FAILED; + + t = wait_for_completion_timeout(&request->wait_event, 5*HZ); + if (t == 0) + return TIMEOUT_ERROR; + + + /* + * At this point, all outstanding requests in the adapter + * should have been flushed out and return to us + * There is a potential race here where the host may be in + * the process of responding when we return from here. + * Just wait for all in-transit packets to be accounted for + * before we return from here. + */ + storvsc_wait_to_drain(stor_device); + + return SUCCESS; +} + +/* + * The host guarantees to respond to each command, although I/O latencies might + * be unbounded on Azure. Reset the timer unconditionally to give the host a + * chance to perform EH. + */ +static enum blk_eh_timer_return storvsc_eh_timed_out(struct scsi_cmnd *scmnd) +{ + return BLK_EH_RESET_TIMER; +} + +static bool storvsc_scsi_cmd_ok(struct scsi_cmnd *scmnd) +{ + bool allowed = true; + u8 scsi_op = scmnd->cmnd[0]; + + switch (scsi_op) { + /* the host does not handle WRITE_SAME, log accident usage */ + case WRITE_SAME: + /* + * smartd sends this command and the host does not handle + * this. So, don't send it. + */ + case SET_WINDOW: + scmnd->result = ILLEGAL_REQUEST << 16; + allowed = false; + break; + default: + break; + } + return allowed; +} + +static int storvsc_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *scmnd) +{ + int ret; + struct hv_host_device *host_dev = shost_priv(host); + struct hv_device *dev = host_dev->dev; + struct storvsc_cmd_request *cmd_request = scsi_cmd_priv(scmnd); + int i; + struct scatterlist *sgl; + unsigned int sg_count = 0; + struct vmscsi_request *vm_srb; + struct scatterlist *cur_sgl; + struct vmbus_packet_mpb_array *payload; + u32 payload_sz; + u32 length; + + if (vmstor_proto_version <= VMSTOR_PROTO_VERSION_WIN8) { + /* + * On legacy hosts filter unimplemented commands. + * Future hosts are expected to correctly handle + * unsupported commands. Furthermore, it is + * possible that some of the currently + * unsupported commands maybe supported in + * future versions of the host. + */ + if (!storvsc_scsi_cmd_ok(scmnd)) { + scmnd->scsi_done(scmnd); + return 0; + } + } + + /* Setup the cmd request */ + cmd_request->cmd = scmnd; + + memset(&cmd_request->vstor_packet, 0, sizeof(struct vstor_packet)); + vm_srb = &cmd_request->vstor_packet.vm_srb; + vm_srb->win8_extension.time_out_value = 60; + + vm_srb->win8_extension.srb_flags |= + SRB_FLAGS_DISABLE_SYNCH_TRANSFER; + + if (scmnd->device->tagged_supported) { + vm_srb->win8_extension.srb_flags |= + (SRB_FLAGS_QUEUE_ACTION_ENABLE | SRB_FLAGS_NO_QUEUE_FREEZE); + vm_srb->win8_extension.queue_tag = SP_UNTAGGED; + vm_srb->win8_extension.queue_action = SRB_SIMPLE_TAG_REQUEST; + } + + /* Build the SRB */ + switch (scmnd->sc_data_direction) { + case DMA_TO_DEVICE: + vm_srb->data_in = WRITE_TYPE; + vm_srb->win8_extension.srb_flags |= SRB_FLAGS_DATA_OUT; + break; + case DMA_FROM_DEVICE: + vm_srb->data_in = READ_TYPE; + vm_srb->win8_extension.srb_flags |= SRB_FLAGS_DATA_IN; + break; + case DMA_NONE: + vm_srb->data_in = UNKNOWN_TYPE; + vm_srb->win8_extension.srb_flags |= SRB_FLAGS_NO_DATA_TRANSFER; + break; + default: + /* + * This is DMA_BIDIRECTIONAL or something else we are never + * supposed to see here. + */ + WARN(1, "Unexpected data direction: %d\n", + scmnd->sc_data_direction); + return -EINVAL; + } + + + vm_srb->port_number = host_dev->port; + vm_srb->path_id = scmnd->device->channel; + vm_srb->target_id = scmnd->device->id; + vm_srb->lun = scmnd->device->lun; + + vm_srb->cdb_length = scmnd->cmd_len; + + memcpy(vm_srb->cdb, scmnd->cmnd, vm_srb->cdb_length); + + sgl = (struct scatterlist *)scsi_sglist(scmnd); + sg_count = scsi_sg_count(scmnd); + + length = scsi_bufflen(scmnd); + payload = (struct vmbus_packet_mpb_array *)&cmd_request->mpb; + payload_sz = 0; + + if (sg_count) { + unsigned int hvpgoff = 0; + unsigned long offset_in_hvpg = sgl->offset & ~HV_HYP_PAGE_MASK; + unsigned int hvpg_count = HVPFN_UP(offset_in_hvpg + length); + u64 hvpfn; + + payload_sz = (hvpg_count * sizeof(u64) + + sizeof(struct vmbus_packet_mpb_array)); + + if (hvpg_count > MAX_PAGE_BUFFER_COUNT) { + payload = kzalloc(payload_sz, GFP_ATOMIC); + if (!payload) + return SCSI_MLQUEUE_DEVICE_BUSY; + } + + /* + * sgl is a list of PAGEs, and payload->range.pfn_array + * expects the page number in the unit of HV_HYP_PAGE_SIZE (the + * page size that Hyper-V uses, so here we need to divide PAGEs + * into HV_HYP_PAGE in case that PAGE_SIZE > HV_HYP_PAGE_SIZE. + * Besides, payload->range.offset should be the offset in one + * HV_HYP_PAGE. + */ + payload->range.len = length; + payload->range.offset = offset_in_hvpg; + hvpgoff = sgl->offset >> HV_HYP_PAGE_SHIFT; + + cur_sgl = sgl; + for (i = 0; i < hvpg_count; i++) { + /* + * 'i' is the index of hv pages in the payload and + * 'hvpgoff' is the offset (in hv pages) of the first + * hv page in the the first page. The relationship + * between the sum of 'i' and 'hvpgoff' and the offset + * (in hv pages) in a payload page ('hvpgoff_in_page') + * is as follow: + * + * |------------------ PAGE -------------------| + * | NR_HV_HYP_PAGES_IN_PAGE hvpgs in total | + * |hvpg|hvpg| ... |hvpg|... |hvpg| + * ^ ^ ^ ^ + * +-hvpgoff-+ +-hvpgoff_in_page-+ + * ^ | + * +--------------------- i ---------------------------+ + */ + unsigned int hvpgoff_in_page = + (i + hvpgoff) % NR_HV_HYP_PAGES_IN_PAGE; + + /* + * Two cases that we need to fetch a page: + * 1) i == 0, the first step or + * 2) hvpgoff_in_page == 0, when we reach the boundary + * of a page. + */ + if (hvpgoff_in_page == 0 || i == 0) { + hvpfn = page_to_hvpfn(sg_page(cur_sgl)); + cur_sgl = sg_next(cur_sgl); + } + + payload->range.pfn_array[i] = hvpfn + hvpgoff_in_page; + } + } + + cmd_request->payload = payload; + cmd_request->payload_sz = payload_sz; + + /* Invokes the vsc to start an IO */ + ret = storvsc_do_io(dev, cmd_request, get_cpu()); + put_cpu(); + + if (ret == -EAGAIN) { + if (payload_sz > sizeof(cmd_request->mpb)) + kfree(payload); + /* no more space */ + return SCSI_MLQUEUE_DEVICE_BUSY; + } + + return 0; +} + +static struct scsi_host_template scsi_driver = { + .module = THIS_MODULE, + .name = "storvsc_host_t", + .cmd_size = sizeof(struct storvsc_cmd_request), + .bios_param = storvsc_get_chs, + .queuecommand = storvsc_queuecommand, + .eh_host_reset_handler = storvsc_host_reset_handler, + .proc_name = "storvsc_host", + .eh_timed_out = storvsc_eh_timed_out, + .slave_alloc = storvsc_device_alloc, + .slave_configure = storvsc_device_configure, + .cmd_per_lun = 2048, + .this_id = -1, + /* Make sure we dont get a sg segment crosses a page boundary */ + .dma_boundary = PAGE_SIZE-1, + /* Ensure there are no gaps in presented sgls */ + .virt_boundary_mask = PAGE_SIZE-1, + .no_write_same = 1, + .track_queue_depth = 1, + .change_queue_depth = storvsc_change_queue_depth, +}; + +enum { + SCSI_GUID, + IDE_GUID, + SFC_GUID, +}; + +static const struct hv_vmbus_device_id id_table[] = { + /* SCSI guid */ + { HV_SCSI_GUID, + .driver_data = SCSI_GUID + }, + /* IDE guid */ + { HV_IDE_GUID, + .driver_data = IDE_GUID + }, + /* Fibre Channel GUID */ + { + HV_SYNTHFC_GUID, + .driver_data = SFC_GUID + }, + { }, +}; + +MODULE_DEVICE_TABLE(vmbus, id_table); + +static const struct { guid_t guid; } fc_guid = { HV_SYNTHFC_GUID }; + +static bool hv_dev_is_fc(struct hv_device *hv_dev) +{ + return guid_equal(&fc_guid.guid, &hv_dev->dev_type); +} + +static int storvsc_probe(struct hv_device *device, + const struct hv_vmbus_device_id *dev_id) +{ + int ret; + int num_cpus = num_online_cpus(); + struct Scsi_Host *host; + struct hv_host_device *host_dev; + bool dev_is_ide = ((dev_id->driver_data == IDE_GUID) ? true : false); + bool is_fc = ((dev_id->driver_data == SFC_GUID) ? true : false); + int target = 0; + struct storvsc_device *stor_device; + int max_luns_per_target; + int max_targets; + int max_channels; + int max_sub_channels = 0; + + /* + * Based on the windows host we are running on, + * set state to properly communicate with the host. + */ + + if (vmbus_proto_version < VERSION_WIN8) { + max_luns_per_target = STORVSC_IDE_MAX_LUNS_PER_TARGET; + max_targets = STORVSC_IDE_MAX_TARGETS; + max_channels = STORVSC_IDE_MAX_CHANNELS; + } else { + max_luns_per_target = STORVSC_MAX_LUNS_PER_TARGET; + max_targets = STORVSC_MAX_TARGETS; + max_channels = STORVSC_MAX_CHANNELS; + /* + * On Windows8 and above, we support sub-channels for storage + * on SCSI and FC controllers. + * The number of sub-channels offerred is based on the number of + * VCPUs in the guest. + */ + if (!dev_is_ide) + max_sub_channels = + (num_cpus - 1) / storvsc_vcpus_per_sub_channel; + } + + scsi_driver.can_queue = max_outstanding_req_per_channel * + (max_sub_channels + 1) * + (100 - ring_avail_percent_lowater) / 100; + + host = scsi_host_alloc(&scsi_driver, + sizeof(struct hv_host_device)); + if (!host) + return -ENOMEM; + + host_dev = shost_priv(host); + memset(host_dev, 0, sizeof(struct hv_host_device)); + + host_dev->port = host->host_no; + host_dev->dev = device; + host_dev->host = host; + + + stor_device = kzalloc(sizeof(struct storvsc_device), GFP_KERNEL); + if (!stor_device) { + ret = -ENOMEM; + goto err_out0; + } + + stor_device->destroy = false; + init_waitqueue_head(&stor_device->waiting_to_drain); + stor_device->device = device; + stor_device->host = host; + spin_lock_init(&stor_device->lock); + hv_set_drvdata(device, stor_device); + + stor_device->port_number = host->host_no; + ret = storvsc_connect_to_vsp(device, storvsc_ringbuffer_size, is_fc); + if (ret) + goto err_out1; + + host_dev->path = stor_device->path_id; + host_dev->target = stor_device->target_id; + + switch (dev_id->driver_data) { + case SFC_GUID: + host->max_lun = STORVSC_FC_MAX_LUNS_PER_TARGET; + host->max_id = STORVSC_FC_MAX_TARGETS; + host->max_channel = STORVSC_FC_MAX_CHANNELS - 1; +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + host->transportt = fc_transport_template; +#endif + break; + + case SCSI_GUID: + host->max_lun = max_luns_per_target; + host->max_id = max_targets; + host->max_channel = max_channels - 1; + break; + + default: + host->max_lun = STORVSC_IDE_MAX_LUNS_PER_TARGET; + host->max_id = STORVSC_IDE_MAX_TARGETS; + host->max_channel = STORVSC_IDE_MAX_CHANNELS - 1; + break; + } + /* max cmd length */ + host->max_cmd_len = STORVSC_MAX_CMD_LEN; + + /* + * set the table size based on the info we got + * from the host. + */ + host->sg_tablesize = (stor_device->max_transfer_bytes >> PAGE_SHIFT); + /* + * For non-IDE disks, the host supports multiple channels. + * Set the number of HW queues we are supporting. + */ + if (!dev_is_ide) + host->nr_hw_queues = num_present_cpus(); + + /* + * Set the error handler work queue. + */ + host_dev->handle_error_wq = + alloc_ordered_workqueue("storvsc_error_wq_%d", + 0, + host->host_no); + if (!host_dev->handle_error_wq) { + ret = -ENOMEM; + goto err_out2; + } + INIT_WORK(&host_dev->host_scan_work, storvsc_host_scan); + /* Register the HBA and start the scsi bus scan */ + ret = scsi_add_host(host, &device->device); + if (ret != 0) + goto err_out3; + + if (!dev_is_ide) { + scsi_scan_host(host); + } else { + target = (device->dev_instance.b[5] << 8 | + device->dev_instance.b[4]); + ret = scsi_add_device(host, 0, target, 0); + if (ret) + goto err_out4; + } +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + if (host->transportt == fc_transport_template) { + struct fc_rport_identifiers ids = { + .roles = FC_PORT_ROLE_FCP_DUMMY_INITIATOR, + }; + + fc_host_node_name(host) = stor_device->node_name; + fc_host_port_name(host) = stor_device->port_name; + stor_device->rport = fc_remote_port_add(host, 0, &ids); + if (!stor_device->rport) { + ret = -ENOMEM; + goto err_out4; + } + } +#endif + return 0; + +err_out4: + scsi_remove_host(host); + +err_out3: + destroy_workqueue(host_dev->handle_error_wq); + +err_out2: + /* + * Once we have connected with the host, we would need to + * to invoke storvsc_dev_remove() to rollback this state and + * this call also frees up the stor_device; hence the jump around + * err_out1 label. + */ + storvsc_dev_remove(device); + goto err_out0; + +err_out1: + kfree(stor_device->stor_chns); + kfree(stor_device); + +err_out0: + scsi_host_put(host); + return ret; +} + +/* Change a scsi target's queue depth */ +static int storvsc_change_queue_depth(struct scsi_device *sdev, int queue_depth) +{ + if (queue_depth > scsi_driver.can_queue) + queue_depth = scsi_driver.can_queue; + + return scsi_change_queue_depth(sdev, queue_depth); +} + +static int storvsc_remove(struct hv_device *dev) +{ + struct storvsc_device *stor_device = hv_get_drvdata(dev); + struct Scsi_Host *host = stor_device->host; + struct hv_host_device *host_dev = shost_priv(host); + +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + if (host->transportt == fc_transport_template) { + fc_remote_port_delete(stor_device->rport); + fc_remove_host(host); + } +#endif + destroy_workqueue(host_dev->handle_error_wq); + scsi_remove_host(host); + storvsc_dev_remove(dev); + scsi_host_put(host); + + return 0; +} + +static int storvsc_suspend(struct hv_device *hv_dev) +{ + struct storvsc_device *stor_device = hv_get_drvdata(hv_dev); + struct Scsi_Host *host = stor_device->host; + struct hv_host_device *host_dev = shost_priv(host); + + storvsc_wait_to_drain(stor_device); + + drain_workqueue(host_dev->handle_error_wq); + + vmbus_close(hv_dev->channel); + + kfree(stor_device->stor_chns); + stor_device->stor_chns = NULL; + + cpumask_clear(&stor_device->alloced_cpus); + + return 0; +} + +static int storvsc_resume(struct hv_device *hv_dev) +{ + int ret; + + ret = storvsc_connect_to_vsp(hv_dev, storvsc_ringbuffer_size, + hv_dev_is_fc(hv_dev)); + return ret; +} + +static struct hv_driver storvsc_drv = { + .name = KBUILD_MODNAME, + .id_table = id_table, + .probe = storvsc_probe, + .remove = storvsc_remove, + .suspend = storvsc_suspend, + .resume = storvsc_resume, + .driver = { + .probe_type = PROBE_PREFER_ASYNCHRONOUS, + }, +}; + +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) +static struct fc_function_template fc_transport_functions = { + .show_host_node_name = 1, + .show_host_port_name = 1, +}; +#endif + +static int __init storvsc_drv_init(void) +{ + int ret; + + /* + * Divide the ring buffer data size (which is 1 page less + * than the ring buffer size since that page is reserved for + * the ring buffer indices) by the max request size (which is + * vmbus_channel_packet_multipage_buffer + struct vstor_packet + u64) + */ + max_outstanding_req_per_channel = + ((storvsc_ringbuffer_size - PAGE_SIZE) / + ALIGN(MAX_MULTIPAGE_BUFFER_PACKET + + sizeof(struct vstor_packet) + sizeof(u64) - + vmscsi_size_delta, + sizeof(u64))); + +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + fc_transport_template = fc_attach_transport(&fc_transport_functions); + if (!fc_transport_template) + return -ENODEV; +#endif + + ret = vmbus_driver_register(&storvsc_drv); + +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + if (ret) + fc_release_transport(fc_transport_template); +#endif + + return ret; +} + +static void __exit storvsc_drv_exit(void) +{ + vmbus_driver_unregister(&storvsc_drv); +#if IS_ENABLED(CONFIG_SCSI_FC_ATTRS) + fc_release_transport(fc_transport_template); +#endif +} + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("Microsoft Hyper-V virtual storage driver"); +module_init(storvsc_drv_init); +module_exit(storvsc_drv_exit); -- cgit v1.2.3