diff options
Diffstat (limited to 'drivers/infiniband/ulp/srp')
| -rw-r--r-- | drivers/infiniband/ulp/srp/ib_srp.c | 2376 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/srp/ib_srp.h | 185 |
2 files changed, 1853 insertions, 708 deletions
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index cfc1d65c457..e3c2c5b4297 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -30,6 +30,8 @@ * SOFTWARE. */ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include <linux/module.h> #include <linux/init.h> #include <linux/slab.h> @@ -39,11 +41,12 @@ #include <linux/random.h> #include <linux/jiffies.h> -#include <asm/atomic.h> +#include <linux/atomic.h> #include <scsi/scsi.h> #include <scsi/scsi_device.h> #include <scsi/scsi_dbg.h> +#include <scsi/scsi_tcq.h> #include <scsi/srp.h> #include <scsi/scsi_transport_srp.h> @@ -51,32 +54,74 @@ #define DRV_NAME "ib_srp" #define PFX DRV_NAME ": " -#define DRV_VERSION "0.2" -#define DRV_RELDATE "November 1, 2005" +#define DRV_VERSION "1.0" +#define DRV_RELDATE "July 1, 2013" MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("InfiniBand SCSI RDMA Protocol initiator " "v" DRV_VERSION " (" DRV_RELDATE ")"); MODULE_LICENSE("Dual BSD/GPL"); -static int srp_sg_tablesize = SRP_DEF_SG_TABLESIZE; -static int srp_max_iu_len; +static unsigned int srp_sg_tablesize; +static unsigned int cmd_sg_entries; +static unsigned int indirect_sg_entries; +static bool allow_ext_sg; +static bool prefer_fr; +static bool register_always; +static int topspin_workarounds = 1; -module_param(srp_sg_tablesize, int, 0444); -MODULE_PARM_DESC(srp_sg_tablesize, - "Max number of gather/scatter entries per I/O (default is 12, max 255)"); +module_param(srp_sg_tablesize, uint, 0444); +MODULE_PARM_DESC(srp_sg_tablesize, "Deprecated name for cmd_sg_entries"); -static int topspin_workarounds = 1; +module_param(cmd_sg_entries, uint, 0444); +MODULE_PARM_DESC(cmd_sg_entries, + "Default number of gather/scatter entries in the SRP command (default is 12, max 255)"); + +module_param(indirect_sg_entries, uint, 0444); +MODULE_PARM_DESC(indirect_sg_entries, + "Default max number of gather/scatter entries (default is 12, max is " __stringify(SCSI_MAX_SG_CHAIN_SEGMENTS) ")"); + +module_param(allow_ext_sg, bool, 0444); +MODULE_PARM_DESC(allow_ext_sg, + "Default behavior when there are more than cmd_sg_entries S/G entries after mapping; fails the request when false (default false)"); module_param(topspin_workarounds, int, 0444); MODULE_PARM_DESC(topspin_workarounds, "Enable workarounds for Topspin/Cisco SRP target bugs if != 0"); -static int mellanox_workarounds = 1; - -module_param(mellanox_workarounds, int, 0444); -MODULE_PARM_DESC(mellanox_workarounds, - "Enable workarounds for Mellanox SRP target bugs if != 0"); +module_param(prefer_fr, bool, 0444); +MODULE_PARM_DESC(prefer_fr, +"Whether to use fast registration if both FMR and fast registration are supported"); + +module_param(register_always, bool, 0444); +MODULE_PARM_DESC(register_always, + "Use memory registration even for contiguous memory regions"); + +static struct kernel_param_ops srp_tmo_ops; + +static int srp_reconnect_delay = 10; +module_param_cb(reconnect_delay, &srp_tmo_ops, &srp_reconnect_delay, + S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(reconnect_delay, "Time between successive reconnect attempts"); + +static int srp_fast_io_fail_tmo = 15; +module_param_cb(fast_io_fail_tmo, &srp_tmo_ops, &srp_fast_io_fail_tmo, + S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(fast_io_fail_tmo, + "Number of seconds between the observation of a transport" + " layer error and failing all I/O. \"off\" means that this" + " functionality is disabled."); + +static int srp_dev_loss_tmo = 600; +module_param_cb(dev_loss_tmo, &srp_tmo_ops, &srp_dev_loss_tmo, + S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(dev_loss_tmo, + "Maximum number of seconds that the SRP transport should" + " insulate transport layer errors. After this time has been" + " exceeded the SCSI host is removed. Should be" + " between 1 and " __stringify(SCSI_DEVICE_BLOCK_MAX_TIMEOUT) + " if fast_io_fail_tmo has not been set. \"off\" means that" + " this functionality is disabled."); static void srp_add_one(struct ib_device *device); static void srp_remove_one(struct ib_device *device); @@ -94,6 +139,48 @@ static struct ib_client srp_client = { static struct ib_sa_client srp_sa_client; +static int srp_tmo_get(char *buffer, const struct kernel_param *kp) +{ + int tmo = *(int *)kp->arg; + + if (tmo >= 0) + return sprintf(buffer, "%d", tmo); + else + return sprintf(buffer, "off"); +} + +static int srp_tmo_set(const char *val, const struct kernel_param *kp) +{ + int tmo, res; + + if (strncmp(val, "off", 3) != 0) { + res = kstrtoint(val, 0, &tmo); + if (res) + goto out; + } else { + tmo = -1; + } + if (kp->arg == &srp_reconnect_delay) + res = srp_tmo_valid(tmo, srp_fast_io_fail_tmo, + srp_dev_loss_tmo); + else if (kp->arg == &srp_fast_io_fail_tmo) + res = srp_tmo_valid(srp_reconnect_delay, tmo, srp_dev_loss_tmo); + else + res = srp_tmo_valid(srp_reconnect_delay, srp_fast_io_fail_tmo, + tmo); + if (res) + goto out; + *(int *)kp->arg = tmo; + +out: + return res; +} + +static struct kernel_param_ops srp_tmo_ops = { + .get = srp_tmo_get, + .set = srp_tmo_set, +}; + static inline struct srp_target_port *host_to_target(struct Scsi_Host *host) { return (struct srp_target_port *) host->hostdata; @@ -114,14 +201,6 @@ static int srp_target_is_topspin(struct srp_target_port *target) !memcmp(&target->ioc_guid, cisco_oui, sizeof cisco_oui)); } -static int srp_target_is_mellanox(struct srp_target_port *target) -{ - static const u8 mellanox_oui[3] = { 0x00, 0x02, 0xc9 }; - - return mellanox_workarounds && - !memcmp(&target->ioc_guid, mellanox_oui, sizeof mellanox_oui); -} - static struct srp_iu *srp_alloc_iu(struct srp_host *host, size_t size, gfp_t gfp_mask, enum dma_data_direction direction) @@ -167,7 +246,7 @@ static void srp_free_iu(struct srp_host *host, struct srp_iu *iu) static void srp_qp_event(struct ib_event *event, void *context) { - printk(KERN_ERR PFX "QP event %d\n", event->event); + pr_debug("QP event %d\n", event->event); } static int srp_init_qp(struct srp_target_port *target, @@ -219,80 +298,288 @@ static int srp_new_cm_id(struct srp_target_port *target) return 0; } +static struct ib_fmr_pool *srp_alloc_fmr_pool(struct srp_target_port *target) +{ + struct srp_device *dev = target->srp_host->srp_dev; + struct ib_fmr_pool_param fmr_param; + + memset(&fmr_param, 0, sizeof(fmr_param)); + fmr_param.pool_size = target->scsi_host->can_queue; + fmr_param.dirty_watermark = fmr_param.pool_size / 4; + fmr_param.cache = 1; + fmr_param.max_pages_per_fmr = dev->max_pages_per_mr; + fmr_param.page_shift = ilog2(dev->mr_page_size); + fmr_param.access = (IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_WRITE | + IB_ACCESS_REMOTE_READ); + + return ib_create_fmr_pool(dev->pd, &fmr_param); +} + +/** + * srp_destroy_fr_pool() - free the resources owned by a pool + * @pool: Fast registration pool to be destroyed. + */ +static void srp_destroy_fr_pool(struct srp_fr_pool *pool) +{ + int i; + struct srp_fr_desc *d; + + if (!pool) + return; + + for (i = 0, d = &pool->desc[0]; i < pool->size; i++, d++) { + if (d->frpl) + ib_free_fast_reg_page_list(d->frpl); + if (d->mr) + ib_dereg_mr(d->mr); + } + kfree(pool); +} + +/** + * srp_create_fr_pool() - allocate and initialize a pool for fast registration + * @device: IB device to allocate fast registration descriptors for. + * @pd: Protection domain associated with the FR descriptors. + * @pool_size: Number of descriptors to allocate. + * @max_page_list_len: Maximum fast registration work request page list length. + */ +static struct srp_fr_pool *srp_create_fr_pool(struct ib_device *device, + struct ib_pd *pd, int pool_size, + int max_page_list_len) +{ + struct srp_fr_pool *pool; + struct srp_fr_desc *d; + struct ib_mr *mr; + struct ib_fast_reg_page_list *frpl; + int i, ret = -EINVAL; + + if (pool_size <= 0) + goto err; + ret = -ENOMEM; + pool = kzalloc(sizeof(struct srp_fr_pool) + + pool_size * sizeof(struct srp_fr_desc), GFP_KERNEL); + if (!pool) + goto err; + pool->size = pool_size; + pool->max_page_list_len = max_page_list_len; + spin_lock_init(&pool->lock); + INIT_LIST_HEAD(&pool->free_list); + + for (i = 0, d = &pool->desc[0]; i < pool->size; i++, d++) { + mr = ib_alloc_fast_reg_mr(pd, max_page_list_len); + if (IS_ERR(mr)) { + ret = PTR_ERR(mr); + goto destroy_pool; + } + d->mr = mr; + frpl = ib_alloc_fast_reg_page_list(device, max_page_list_len); + if (IS_ERR(frpl)) { + ret = PTR_ERR(frpl); + goto destroy_pool; + } + d->frpl = frpl; + list_add_tail(&d->entry, &pool->free_list); + } + +out: + return pool; + +destroy_pool: + srp_destroy_fr_pool(pool); + +err: + pool = ERR_PTR(ret); + goto out; +} + +/** + * srp_fr_pool_get() - obtain a descriptor suitable for fast registration + * @pool: Pool to obtain descriptor from. + */ +static struct srp_fr_desc *srp_fr_pool_get(struct srp_fr_pool *pool) +{ + struct srp_fr_desc *d = NULL; + unsigned long flags; + + spin_lock_irqsave(&pool->lock, flags); + if (!list_empty(&pool->free_list)) { + d = list_first_entry(&pool->free_list, typeof(*d), entry); + list_del(&d->entry); + } + spin_unlock_irqrestore(&pool->lock, flags); + + return d; +} + +/** + * srp_fr_pool_put() - put an FR descriptor back in the free list + * @pool: Pool the descriptor was allocated from. + * @desc: Pointer to an array of fast registration descriptor pointers. + * @n: Number of descriptors to put back. + * + * Note: The caller must already have queued an invalidation request for + * desc->mr->rkey before calling this function. + */ +static void srp_fr_pool_put(struct srp_fr_pool *pool, struct srp_fr_desc **desc, + int n) +{ + unsigned long flags; + int i; + + spin_lock_irqsave(&pool->lock, flags); + for (i = 0; i < n; i++) + list_add(&desc[i]->entry, &pool->free_list); + spin_unlock_irqrestore(&pool->lock, flags); +} + +static struct srp_fr_pool *srp_alloc_fr_pool(struct srp_target_port *target) +{ + struct srp_device *dev = target->srp_host->srp_dev; + + return srp_create_fr_pool(dev->dev, dev->pd, + target->scsi_host->can_queue, + dev->max_pages_per_mr); +} + static int srp_create_target_ib(struct srp_target_port *target) { + struct srp_device *dev = target->srp_host->srp_dev; struct ib_qp_init_attr *init_attr; + struct ib_cq *recv_cq, *send_cq; + struct ib_qp *qp; + struct ib_fmr_pool *fmr_pool = NULL; + struct srp_fr_pool *fr_pool = NULL; + const int m = 1 + dev->use_fast_reg; int ret; init_attr = kzalloc(sizeof *init_attr, GFP_KERNEL); if (!init_attr) return -ENOMEM; - target->recv_cq = ib_create_cq(target->srp_host->srp_dev->dev, - srp_recv_completion, NULL, target, SRP_RQ_SIZE, 0); - if (IS_ERR(target->recv_cq)) { - ret = PTR_ERR(target->recv_cq); + recv_cq = ib_create_cq(dev->dev, srp_recv_completion, NULL, target, + target->queue_size, target->comp_vector); + if (IS_ERR(recv_cq)) { + ret = PTR_ERR(recv_cq); goto err; } - target->send_cq = ib_create_cq(target->srp_host->srp_dev->dev, - srp_send_completion, NULL, target, SRP_SQ_SIZE, 0); - if (IS_ERR(target->send_cq)) { - ret = PTR_ERR(target->send_cq); + send_cq = ib_create_cq(dev->dev, srp_send_completion, NULL, target, + m * target->queue_size, target->comp_vector); + if (IS_ERR(send_cq)) { + ret = PTR_ERR(send_cq); goto err_recv_cq; } - ib_req_notify_cq(target->recv_cq, IB_CQ_NEXT_COMP); + ib_req_notify_cq(recv_cq, IB_CQ_NEXT_COMP); init_attr->event_handler = srp_qp_event; - init_attr->cap.max_send_wr = SRP_SQ_SIZE; - init_attr->cap.max_recv_wr = SRP_RQ_SIZE; + init_attr->cap.max_send_wr = m * target->queue_size; + init_attr->cap.max_recv_wr = target->queue_size; init_attr->cap.max_recv_sge = 1; init_attr->cap.max_send_sge = 1; - init_attr->sq_sig_type = IB_SIGNAL_ALL_WR; + init_attr->sq_sig_type = IB_SIGNAL_REQ_WR; init_attr->qp_type = IB_QPT_RC; - init_attr->send_cq = target->send_cq; - init_attr->recv_cq = target->recv_cq; + init_attr->send_cq = send_cq; + init_attr->recv_cq = recv_cq; - target->qp = ib_create_qp(target->srp_host->srp_dev->pd, init_attr); - if (IS_ERR(target->qp)) { - ret = PTR_ERR(target->qp); + qp = ib_create_qp(dev->pd, init_attr); + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); goto err_send_cq; } - ret = srp_init_qp(target, target->qp); + ret = srp_init_qp(target, qp); if (ret) goto err_qp; + if (dev->use_fast_reg && dev->has_fr) { + fr_pool = srp_alloc_fr_pool(target); + if (IS_ERR(fr_pool)) { + ret = PTR_ERR(fr_pool); + shost_printk(KERN_WARNING, target->scsi_host, PFX + "FR pool allocation failed (%d)\n", ret); + goto err_qp; + } + if (target->fr_pool) + srp_destroy_fr_pool(target->fr_pool); + target->fr_pool = fr_pool; + } else if (!dev->use_fast_reg && dev->has_fmr) { + fmr_pool = srp_alloc_fmr_pool(target); + if (IS_ERR(fmr_pool)) { + ret = PTR_ERR(fmr_pool); + shost_printk(KERN_WARNING, target->scsi_host, PFX + "FMR pool allocation failed (%d)\n", ret); + goto err_qp; + } + if (target->fmr_pool) + ib_destroy_fmr_pool(target->fmr_pool); + target->fmr_pool = fmr_pool; + } + + if (target->qp) + ib_destroy_qp(target->qp); + if (target->recv_cq) + ib_destroy_cq(target->recv_cq); + if (target->send_cq) + ib_destroy_cq(target->send_cq); + + target->qp = qp; + target->recv_cq = recv_cq; + target->send_cq = send_cq; + kfree(init_attr); return 0; err_qp: - ib_destroy_qp(target->qp); + ib_destroy_qp(qp); err_send_cq: - ib_destroy_cq(target->send_cq); + ib_destroy_cq(send_cq); err_recv_cq: - ib_destroy_cq(target->recv_cq); + ib_destroy_cq(recv_cq); err: kfree(init_attr); return ret; } +/* + * Note: this function may be called without srp_alloc_iu_bufs() having been + * invoked. Hence the target->[rt]x_ring checks. + */ static void srp_free_target_ib(struct srp_target_port *target) { + struct srp_device *dev = target->srp_host->srp_dev; int i; + if (dev->use_fast_reg) { + if (target->fr_pool) + srp_destroy_fr_pool(target->fr_pool); + } else { + if (target->fmr_pool) + ib_destroy_fmr_pool(target->fmr_pool); + } ib_destroy_qp(target->qp); ib_destroy_cq(target->send_cq); ib_destroy_cq(target->recv_cq); - for (i = 0; i < SRP_RQ_SIZE; ++i) - srp_free_iu(target->srp_host, target->rx_ring[i]); - for (i = 0; i < SRP_SQ_SIZE; ++i) - srp_free_iu(target->srp_host, target->tx_ring[i]); + target->qp = NULL; + target->send_cq = target->recv_cq = NULL; + + if (target->rx_ring) { + for (i = 0; i < target->queue_size; ++i) + srp_free_iu(target->srp_host, target->rx_ring[i]); + kfree(target->rx_ring); + target->rx_ring = NULL; + } + if (target->tx_ring) { + for (i = 0; i < target->queue_size; ++i) + srp_free_iu(target->srp_host, target->tx_ring[i]); + kfree(target->tx_ring); + target->tx_ring = NULL; + } } static void srp_path_rec_completion(int status, @@ -312,6 +599,8 @@ static void srp_path_rec_completion(int status, static int srp_lookup_path(struct srp_target_port *target) { + int ret; + target->path.numb_path = 1; init_completion(&target->done); @@ -332,7 +621,9 @@ static int srp_lookup_path(struct srp_target_port *target) if (target->path_query_id < 0) return target->path_query_id; - wait_for_completion(&target->done); + ret = wait_for_completion_interruptible(&target->done); + if (ret < 0) + return ret; if (target->status < 0) shost_printk(KERN_WARNING, target->scsi_host, @@ -372,13 +663,13 @@ static int srp_send_req(struct srp_target_port *target) req->param.responder_resources = 4; req->param.remote_cm_response_timeout = 20; req->param.local_cm_response_timeout = 20; - req->param.retry_count = 7; + req->param.retry_count = target->tl_retry_count; req->param.rnr_retry_count = 7; req->param.max_cm_retries = 15; req->priv.opcode = SRP_LOGIN_REQ; req->priv.tag = 0; - req->priv.req_it_iu_len = cpu_to_be32(srp_max_iu_len); + req->priv.req_it_iu_len = cpu_to_be32(target->max_iu_len); req->priv.req_buf_fmt = cpu_to_be16(SRP_BUF_FORMAT_DIRECT | SRP_BUF_FORMAT_INDIRECT); /* @@ -428,48 +719,193 @@ static int srp_send_req(struct srp_target_port *target) return status; } +static bool srp_queue_remove_work(struct srp_target_port *target) +{ + bool changed = false; + + spin_lock_irq(&target->lock); + if (target->state != SRP_TARGET_REMOVED) { + target->state = SRP_TARGET_REMOVED; + changed = true; + } + spin_unlock_irq(&target->lock); + + if (changed) + queue_work(system_long_wq, &target->remove_work); + + return changed; +} + +static bool srp_change_conn_state(struct srp_target_port *target, + bool connected) +{ + bool changed = false; + + spin_lock_irq(&target->lock); + if (target->connected != connected) { + target->connected = connected; + changed = true; + } + spin_unlock_irq(&target->lock); + + return changed; +} + static void srp_disconnect_target(struct srp_target_port *target) { - /* XXX should send SRP_I_LOGOUT request */ + if (srp_change_conn_state(target, false)) { + /* XXX should send SRP_I_LOGOUT request */ - init_completion(&target->done); - if (ib_send_cm_dreq(target->cm_id, NULL, 0)) { - shost_printk(KERN_DEBUG, target->scsi_host, - PFX "Sending CM DREQ failed\n"); - return; + if (ib_send_cm_dreq(target->cm_id, NULL, 0)) { + shost_printk(KERN_DEBUG, target->scsi_host, + PFX "Sending CM DREQ failed\n"); + } } - wait_for_completion(&target->done); } -static void srp_remove_work(struct work_struct *work) +static void srp_free_req_data(struct srp_target_port *target) { - struct srp_target_port *target = - container_of(work, struct srp_target_port, work); + struct srp_device *dev = target->srp_host->srp_dev; + struct ib_device *ibdev = dev->dev; + struct srp_request *req; + int i; - spin_lock_irq(target->scsi_host->host_lock); - if (target->state != SRP_TARGET_DEAD) { - spin_unlock_irq(target->scsi_host->host_lock); + if (!target->req_ring) return; + + for (i = 0; i < target->req_ring_size; ++i) { + req = &target->req_ring[i]; + if (dev->use_fast_reg) + kfree(req->fr_list); + else + kfree(req->fmr_list); + kfree(req->map_page); + if (req->indirect_dma_addr) { + ib_dma_unmap_single(ibdev, req->indirect_dma_addr, + target->indirect_size, + DMA_TO_DEVICE); + } + kfree(req->indirect_desc); } - target->state = SRP_TARGET_REMOVED; - spin_unlock_irq(target->scsi_host->host_lock); - spin_lock(&target->srp_host->target_lock); - list_del(&target->list); - spin_unlock(&target->srp_host->target_lock); + kfree(target->req_ring); + target->req_ring = NULL; +} + +static int srp_alloc_req_data(struct srp_target_port *target) +{ + struct srp_device *srp_dev = target->srp_host->srp_dev; + struct ib_device *ibdev = srp_dev->dev; + struct srp_request *req; + void *mr_list; + dma_addr_t dma_addr; + int i, ret = -ENOMEM; + + INIT_LIST_HEAD(&target->free_reqs); + target->req_ring = kzalloc(target->req_ring_size * + sizeof(*target->req_ring), GFP_KERNEL); + if (!target->req_ring) + goto out; + + for (i = 0; i < target->req_ring_size; ++i) { + req = &target->req_ring[i]; + mr_list = kmalloc(target->cmd_sg_cnt * sizeof(void *), + GFP_KERNEL); + if (!mr_list) + goto out; + if (srp_dev->use_fast_reg) + req->fr_list = mr_list; + else + req->fmr_list = mr_list; + req->map_page = kmalloc(srp_dev->max_pages_per_mr * + sizeof(void *), GFP_KERNEL); + if (!req->map_page) + goto out; + req->indirect_desc = kmalloc(target->indirect_size, GFP_KERNEL); + if (!req->indirect_desc) + goto out; + + dma_addr = ib_dma_map_single(ibdev, req->indirect_desc, + target->indirect_size, + DMA_TO_DEVICE); + if (ib_dma_mapping_error(ibdev, dma_addr)) + goto out; + + req->indirect_dma_addr = dma_addr; + req->index = i; + list_add_tail(&req->list, &target->free_reqs); + } + ret = 0; + +out: + return ret; +} + +/** + * srp_del_scsi_host_attr() - Remove attributes defined in the host template. + * @shost: SCSI host whose attributes to remove from sysfs. + * + * Note: Any attributes defined in the host template and that did not exist + * before invocation of this function will be ignored. + */ +static void srp_del_scsi_host_attr(struct Scsi_Host *shost) +{ + struct device_attribute **attr; + + for (attr = shost->hostt->shost_attrs; attr && *attr; ++attr) + device_remove_file(&shost->shost_dev, *attr); +} + +static void srp_remove_target(struct srp_target_port *target) +{ + WARN_ON_ONCE(target->state != SRP_TARGET_REMOVED); + + srp_del_scsi_host_attr(target->scsi_host); + srp_rport_get(target->rport); srp_remove_host(target->scsi_host); scsi_remove_host(target->scsi_host); + srp_stop_rport_timers(target->rport); + srp_disconnect_target(target); ib_destroy_cm_id(target->cm_id); srp_free_target_ib(target); + cancel_work_sync(&target->tl_err_work); + srp_rport_put(target->rport); + srp_free_req_data(target); + + spin_lock(&target->srp_host->target_lock); + list_del(&target->list); + spin_unlock(&target->srp_host->target_lock); + scsi_host_put(target->scsi_host); } +static void srp_remove_work(struct work_struct *work) +{ + struct srp_target_port *target = + container_of(work, struct srp_target_port, remove_work); + + WARN_ON_ONCE(target->state != SRP_TARGET_REMOVED); + + srp_remove_target(target); +} + +static void srp_rport_delete(struct srp_rport *rport) +{ + struct srp_target_port *target = rport->lld_data; + + srp_queue_remove_work(target); +} + static int srp_connect_target(struct srp_target_port *target) { int retries = 3; int ret; + WARN_ON_ONCE(target->connected); + + target->qp_in_error = false; + ret = srp_lookup_path(target); if (ret) return ret; @@ -479,7 +915,9 @@ static int srp_connect_target(struct srp_target_port *target) ret = srp_send_req(target); if (ret) return ret; - wait_for_completion(&target->done); + ret = wait_for_completion_interruptible(&target->done); + if (ret < 0) + return ret; /* * The CM event handling code will set status to @@ -489,6 +927,7 @@ static int srp_connect_target(struct srp_target_port *target) */ switch (target->status) { case 0: + srp_change_conn_state(target, true); return 0; case SRP_PORT_REDIRECT: @@ -521,199 +960,436 @@ static int srp_connect_target(struct srp_target_port *target) } } +static int srp_inv_rkey(struct srp_target_port *target, u32 rkey) +{ + struct ib_send_wr *bad_wr; + struct ib_send_wr wr = { + .opcode = IB_WR_LOCAL_INV, + .wr_id = LOCAL_INV_WR_ID_MASK, + .next = NULL, + .num_sge = 0, + .send_flags = 0, + .ex.invalidate_rkey = rkey, + }; + + return ib_post_send(target->qp, &wr, &bad_wr); +} + static void srp_unmap_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, struct srp_request *req) { + struct srp_device *dev = target->srp_host->srp_dev; + struct ib_device *ibdev = dev->dev; + int i, res; + if (!scsi_sglist(scmnd) || (scmnd->sc_data_direction != DMA_TO_DEVICE && scmnd->sc_data_direction != DMA_FROM_DEVICE)) return; - if (req->fmr) { - ib_fmr_pool_unmap(req->fmr); - req->fmr = NULL; + if (dev->use_fast_reg) { + struct srp_fr_desc **pfr; + + for (i = req->nmdesc, pfr = req->fr_list; i > 0; i--, pfr++) { + res = srp_inv_rkey(target, (*pfr)->mr->rkey); + if (res < 0) { + shost_printk(KERN_ERR, target->scsi_host, PFX + "Queueing INV WR for rkey %#x failed (%d)\n", + (*pfr)->mr->rkey, res); + queue_work(system_long_wq, + &target->tl_err_work); + } + } + if (req->nmdesc) + srp_fr_pool_put(target->fr_pool, req->fr_list, + req->nmdesc); + } else { + struct ib_pool_fmr **pfmr; + + for (i = req->nmdesc, pfmr = req->fmr_list; i > 0; i--, pfmr++) + ib_fmr_pool_unmap(*pfmr); } - ib_dma_unmap_sg(target->srp_host->srp_dev->dev, scsi_sglist(scmnd), - scsi_sg_count(scmnd), scmnd->sc_data_direction); + ib_dma_unmap_sg(ibdev, scsi_sglist(scmnd), scsi_sg_count(scmnd), + scmnd->sc_data_direction); } -static void srp_remove_req(struct srp_target_port *target, struct srp_request *req) +/** + * srp_claim_req - Take ownership of the scmnd associated with a request. + * @target: SRP target port. + * @req: SRP request. + * @sdev: If not NULL, only take ownership for this SCSI device. + * @scmnd: If NULL, take ownership of @req->scmnd. If not NULL, only take + * ownership of @req->scmnd if it equals @scmnd. + * + * Return value: + * Either NULL or a pointer to the SCSI command the caller became owner of. + */ +static struct scsi_cmnd *srp_claim_req(struct srp_target_port *target, + struct srp_request *req, + struct scsi_device *sdev, + struct scsi_cmnd *scmnd) { - srp_unmap_data(req->scmnd, target, req); - list_move_tail(&req->list, &target->free_reqs); + unsigned long flags; + + spin_lock_irqsave(&target->lock, flags); + if (req->scmnd && + (!sdev || req->scmnd->device == sdev) && + (!scmnd || req->scmnd == scmnd)) { + scmnd = req->scmnd; + req->scmnd = NULL; + } else { + scmnd = NULL; + } + spin_unlock_irqrestore(&target->lock, flags); + + return scmnd; } -static void srp_reset_req(struct srp_target_port *target, struct srp_request *req) +/** + * srp_free_req() - Unmap data and add request to the free request list. + * @target: SRP target port. + * @req: Request to be freed. + * @scmnd: SCSI command associated with @req. + * @req_lim_delta: Amount to be added to @target->req_lim. + */ +static void srp_free_req(struct srp_target_port *target, + struct srp_request *req, struct scsi_cmnd *scmnd, + s32 req_lim_delta) { - req->scmnd->result = DID_RESET << 16; - req->scmnd->scsi_done(req->scmnd); - srp_remove_req(target, req); + unsigned long flags; + + srp_unmap_data(scmnd, target, req); + + spin_lock_irqsave(&target->lock, flags); + target->req_lim += req_lim_delta; + list_add_tail(&req->list, &target->free_reqs); + spin_unlock_irqrestore(&target->lock, flags); } -static int srp_reconnect_target(struct srp_target_port *target) +static void srp_finish_req(struct srp_target_port *target, + struct srp_request *req, struct scsi_device *sdev, + int result) { - struct ib_qp_attr qp_attr; - struct srp_request *req, *tmp; - struct ib_wc wc; - int ret; + struct scsi_cmnd *scmnd = srp_claim_req(target, req, sdev, NULL); - spin_lock_irq(target->scsi_host->host_lock); - if (target->state != SRP_TARGET_LIVE) { - spin_unlock_irq(target->scsi_host->host_lock); - return -EAGAIN; + if (scmnd) { + srp_free_req(target, req, scmnd, 0); + scmnd->result = result; + scmnd->scsi_done(scmnd); } - target->state = SRP_TARGET_CONNECTING; - spin_unlock_irq(target->scsi_host->host_lock); +} + +static void srp_terminate_io(struct srp_rport *rport) +{ + struct srp_target_port *target = rport->lld_data; + struct Scsi_Host *shost = target->scsi_host; + struct scsi_device *sdev; + int i; + + /* + * Invoking srp_terminate_io() while srp_queuecommand() is running + * is not safe. Hence the warning statement below. + */ + shost_for_each_device(sdev, shost) + WARN_ON_ONCE(sdev->request_queue->request_fn_active); + + for (i = 0; i < target->req_ring_size; ++i) { + struct srp_request *req = &target->req_ring[i]; + srp_finish_req(target, req, NULL, DID_TRANSPORT_FAILFAST << 16); + } +} + +/* + * It is up to the caller to ensure that srp_rport_reconnect() calls are + * serialized and that no concurrent srp_queuecommand(), srp_abort(), + * srp_reset_device() or srp_reset_host() calls will occur while this function + * is in progress. One way to realize that is not to call this function + * directly but to call srp_reconnect_rport() instead since that last function + * serializes calls of this function via rport->mutex and also blocks + * srp_queuecommand() calls before invoking this function. + */ +static int srp_rport_reconnect(struct srp_rport *rport) +{ + struct srp_target_port *target = rport->lld_data; + int i, ret; srp_disconnect_target(target); /* - * Now get a new local CM ID so that we avoid confusing the - * target in case things are really fouled up. + * Now get a new local CM ID so that we avoid confusing the target in + * case things are really fouled up. Doing so also ensures that all CM + * callbacks will have finished before a new QP is allocated. */ ret = srp_new_cm_id(target); - if (ret) - goto err; - qp_attr.qp_state = IB_QPS_RESET; - ret = ib_modify_qp(target->qp, &qp_attr, IB_QP_STATE); - if (ret) - goto err; + for (i = 0; i < target->req_ring_size; ++i) { + struct srp_request *req = &target->req_ring[i]; + srp_finish_req(target, req, NULL, DID_RESET << 16); + } - ret = srp_init_qp(target, target->qp); - if (ret) - goto err; + /* + * Whether or not creating a new CM ID succeeded, create a new + * QP. This guarantees that all callback functions for the old QP have + * finished before any send requests are posted on the new QP. + */ + ret += srp_create_target_ib(target); - while (ib_poll_cq(target->recv_cq, 1, &wc) > 0) - ; /* nothing */ - while (ib_poll_cq(target->send_cq, 1, &wc) > 0) - ; /* nothing */ + INIT_LIST_HEAD(&target->free_tx); + for (i = 0; i < target->queue_size; ++i) + list_add(&target->tx_ring[i]->list, &target->free_tx); - spin_lock_irq(target->scsi_host->host_lock); - list_for_each_entry_safe(req, tmp, &target->req_queue, list) - srp_reset_req(target, req); - spin_unlock_irq(target->scsi_host->host_lock); + if (ret == 0) + ret = srp_connect_target(target); - target->rx_head = 0; - target->tx_head = 0; - target->tx_tail = 0; + if (ret == 0) + shost_printk(KERN_INFO, target->scsi_host, + PFX "reconnect succeeded\n"); - target->qp_in_error = 0; - ret = srp_connect_target(target); - if (ret) - goto err; + return ret; +} - spin_lock_irq(target->scsi_host->host_lock); - if (target->state == SRP_TARGET_CONNECTING) { - ret = 0; - target->state = SRP_TARGET_LIVE; - } else - ret = -EAGAIN; - spin_unlock_irq(target->scsi_host->host_lock); +static void srp_map_desc(struct srp_map_state *state, dma_addr_t dma_addr, + unsigned int dma_len, u32 rkey) +{ + struct srp_direct_buf *desc = state->desc; - return ret; + desc->va = cpu_to_be64(dma_addr); + desc->key = cpu_to_be32(rkey); + desc->len = cpu_to_be32(dma_len); -err: - shost_printk(KERN_ERR, target->scsi_host, - PFX "reconnect failed (%d), removing target port.\n", ret); + state->total_len += dma_len; + state->desc++; + state->ndesc++; +} - /* - * We couldn't reconnect, so kill our target port off. - * However, we have to defer the real removal because we might - * be in the context of the SCSI error handler now, which - * would deadlock if we call scsi_remove_host(). - */ - spin_lock_irq(target->scsi_host->host_lock); - if (target->state == SRP_TARGET_CONNECTING) { - target->state = SRP_TARGET_DEAD; - INIT_WORK(&target->work, srp_remove_work); - schedule_work(&target->work); +static int srp_map_finish_fmr(struct srp_map_state *state, + struct srp_target_port *target) +{ + struct ib_pool_fmr *fmr; + u64 io_addr = 0; + + fmr = ib_fmr_pool_map_phys(target->fmr_pool, state->pages, + state->npages, io_addr); + if (IS_ERR(fmr)) + return PTR_ERR(fmr); + + *state->next_fmr++ = fmr; + state->nmdesc++; + + srp_map_desc(state, 0, state->dma_len, fmr->fmr->rkey); + + return 0; +} + +static int srp_map_finish_fr(struct srp_map_state *state, + struct srp_target_port *target) +{ + struct srp_device *dev = target->srp_host->srp_dev; + struct ib_send_wr *bad_wr; + struct ib_send_wr wr; + struct srp_fr_desc *desc; + u32 rkey; + + desc = srp_fr_pool_get(target->fr_pool); + if (!desc) + return -ENOMEM; + + rkey = ib_inc_rkey(desc->mr->rkey); + ib_update_fast_reg_key(desc->mr, rkey); + + memcpy(desc->frpl->page_list, state->pages, + sizeof(state->pages[0]) * state->npages); + + memset(&wr, 0, sizeof(wr)); + wr.opcode = IB_WR_FAST_REG_MR; + wr.wr_id = FAST_REG_WR_ID_MASK; + wr.wr.fast_reg.iova_start = state->base_dma_addr; + wr.wr.fast_reg.page_list = desc->frpl; + wr.wr.fast_reg.page_list_len = state->npages; + wr.wr.fast_reg.page_shift = ilog2(dev->mr_page_size); + wr.wr.fast_reg.length = state->dma_len; + wr.wr.fast_reg.access_flags = (IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE); + wr.wr.fast_reg.rkey = desc->mr->lkey; + + *state->next_fr++ = desc; + state->nmdesc++; + + srp_map_desc(state, state->base_dma_addr, state->dma_len, + desc->mr->rkey); + + return ib_post_send(target->qp, &wr, &bad_wr); +} + +static int srp_finish_mapping(struct srp_map_state *state, + struct srp_target_port *target) +{ + int ret = 0; + + if (state->npages == 0) + return 0; + + if (state->npages == 1 && !register_always) + srp_map_desc(state, state->base_dma_addr, state->dma_len, + target->rkey); + else + ret = target->srp_host->srp_dev->use_fast_reg ? + srp_map_finish_fr(state, target) : + srp_map_finish_fmr(state, target); + + if (ret == 0) { + state->npages = 0; + state->dma_len = 0; } - spin_unlock_irq(target->scsi_host->host_lock); return ret; } -static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat, - int sg_cnt, struct srp_request *req, - struct srp_direct_buf *buf) +static void srp_map_update_start(struct srp_map_state *state, + struct scatterlist *sg, int sg_index, + dma_addr_t dma_addr) +{ + state->unmapped_sg = sg; + state->unmapped_index = sg_index; + state->unmapped_addr = dma_addr; +} + +static int srp_map_sg_entry(struct srp_map_state *state, + struct srp_target_port *target, + struct scatterlist *sg, int sg_index, + bool use_mr) { - u64 io_addr = 0; - u64 *dma_pages; - u32 len; - int page_cnt; - int i, j; - int ret; struct srp_device *dev = target->srp_host->srp_dev; struct ib_device *ibdev = dev->dev; - struct scatterlist *sg; - - if (!dev->fmr_pool) - return -ENODEV; + dma_addr_t dma_addr = ib_sg_dma_address(ibdev, sg); + unsigned int dma_len = ib_sg_dma_len(ibdev, sg); + unsigned int len; + int ret; - if (srp_target_is_mellanox(target) && - (ib_sg_dma_address(ibdev, &scat[0]) & ~dev->fmr_page_mask)) - return -EINVAL; + if (!dma_len) + return 0; - len = page_cnt = 0; - scsi_for_each_sg(req->scmnd, sg, sg_cnt, i) { - unsigned int dma_len = ib_sg_dma_len(ibdev, sg); + if (!use_mr) { + /* + * Once we're in direct map mode for a request, we don't + * go back to FMR or FR mode, so no need to update anything + * other than the descriptor. + */ + srp_map_desc(state, dma_addr, dma_len, target->rkey); + return 0; + } - if (ib_sg_dma_address(ibdev, sg) & ~dev->fmr_page_mask) { - if (i > 0) - return -EINVAL; - else - ++page_cnt; - } - if ((ib_sg_dma_address(ibdev, sg) + dma_len) & - ~dev->fmr_page_mask) { - if (i < sg_cnt - 1) - return -EINVAL; - else - ++page_cnt; - } + /* + * Since not all RDMA HW drivers support non-zero page offsets for + * FMR, if we start at an offset into a page, don't merge into the + * current FMR mapping. Finish it out, and use the kernel's MR for + * this sg entry. + */ + if ((!dev->use_fast_reg && dma_addr & ~dev->mr_page_mask) || + dma_len > dev->mr_max_size) { + ret = srp_finish_mapping(state, target); + if (ret) + return ret; - len += dma_len; + srp_map_desc(state, dma_addr, dma_len, target->rkey); + srp_map_update_start(state, NULL, 0, 0); + return 0; } - page_cnt += len >> dev->fmr_page_shift; - if (page_cnt > SRP_FMR_SIZE) - return -ENOMEM; + /* + * If this is the first sg that will be mapped via FMR or via FR, save + * our position. We need to know the first unmapped entry, its index, + * and the first unmapped address within that entry to be able to + * restart mapping after an error. + */ + if (!state->unmapped_sg) + srp_map_update_start(state, sg, sg_index, dma_addr); - dma_pages = kmalloc(sizeof (u64) * page_cnt, GFP_ATOMIC); - if (!dma_pages) - return -ENOMEM; + while (dma_len) { + unsigned offset = dma_addr & ~dev->mr_page_mask; + if (state->npages == dev->max_pages_per_mr || offset != 0) { + ret = srp_finish_mapping(state, target); + if (ret) + return ret; + + srp_map_update_start(state, sg, sg_index, dma_addr); + } - page_cnt = 0; - scsi_for_each_sg(req->scmnd, sg, sg_cnt, i) { - unsigned int dma_len = ib_sg_dma_len(ibdev, sg); + len = min_t(unsigned int, dma_len, dev->mr_page_size - offset); - for (j = 0; j < dma_len; j += dev->fmr_page_size) - dma_pages[page_cnt++] = - (ib_sg_dma_address(ibdev, sg) & - dev->fmr_page_mask) + j; + if (!state->npages) + state->base_dma_addr = dma_addr; + state->pages[state->npages++] = dma_addr & dev->mr_page_mask; + state->dma_len += len; + dma_addr += len; + dma_len -= len; } - req->fmr = ib_fmr_pool_map_phys(dev->fmr_pool, - dma_pages, page_cnt, io_addr); - if (IS_ERR(req->fmr)) { - ret = PTR_ERR(req->fmr); - req->fmr = NULL; - goto out; + /* + * If the last entry of the MR wasn't a full page, then we need to + * close it out and start a new one -- we can only merge at page + * boundries. + */ + ret = 0; + if (len != dev->mr_page_size) { + ret = srp_finish_mapping(state, target); + if (!ret) + srp_map_update_start(state, NULL, 0, 0); } + return ret; +} - buf->va = cpu_to_be64(ib_sg_dma_address(ibdev, &scat[0]) & - ~dev->fmr_page_mask); - buf->key = cpu_to_be32(req->fmr->fmr->rkey); - buf->len = cpu_to_be32(len); +static int srp_map_sg(struct srp_map_state *state, + struct srp_target_port *target, struct srp_request *req, + struct scatterlist *scat, int count) +{ + struct srp_device *dev = target->srp_host->srp_dev; + struct ib_device *ibdev = dev->dev; + struct scatterlist *sg; + int i; + bool use_mr; - ret = 0; + state->desc = req->indirect_desc; + state->pages = req->map_page; + if (dev->use_fast_reg) { + state->next_fr = req->fr_list; + use_mr = !!target->fr_pool; + } else { + state->next_fmr = req->fmr_list; + use_mr = !!target->fmr_pool; + } -out: - kfree(dma_pages); + for_each_sg(scat, sg, count, i) { + if (srp_map_sg_entry(state, target, sg, i, use_mr)) { + /* + * Memory registration failed, so backtrack to the + * first unmapped entry and continue on without using + * memory registration. + */ + dma_addr_t dma_addr; + unsigned int dma_len; + +backtrack: + sg = state->unmapped_sg; + i = state->unmapped_index; + + dma_addr = ib_sg_dma_address(ibdev, sg); + dma_len = ib_sg_dma_len(ibdev, sg); + dma_len -= (state->unmapped_addr - dma_addr); + dma_addr = state->unmapped_addr; + use_mr = false; + srp_map_desc(state, dma_addr, dma_len, target->rkey); + } + } - return ret; + if (use_mr && srp_finish_mapping(state, target)) + goto backtrack; + + req->nmdesc = state->nmdesc; + + return 0; } static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, @@ -722,9 +1398,12 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, struct scatterlist *scat; struct srp_cmd *cmd = req->cmd->buf; int len, nents, count; - u8 fmt = SRP_DATA_DESC_DIRECT; struct srp_device *dev; struct ib_device *ibdev; + struct srp_map_state state; + struct srp_indirect_buf *indirect_hdr; + u32 table_len; + u8 fmt; if (!scsi_sglist(scmnd) || scmnd->sc_data_direction == DMA_NONE) return sizeof (struct srp_cmd); @@ -744,11 +1423,13 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, ibdev = dev->dev; count = ib_dma_map_sg(ibdev, scat, nents, scmnd->sc_data_direction); + if (unlikely(count == 0)) + return -EIO; fmt = SRP_DATA_DESC_DIRECT; len = sizeof (struct srp_cmd) + sizeof (struct srp_direct_buf); - if (count == 1) { + if (count == 1 && !register_always) { /* * The midlayer only generated a single gather/scatter * entry, or DMA mapping coalesced everything to a @@ -758,51 +1439,73 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, struct srp_direct_buf *buf = (void *) cmd->add_data; buf->va = cpu_to_be64(ib_sg_dma_address(ibdev, scat)); - buf->key = cpu_to_be32(dev->mr->rkey); + buf->key = cpu_to_be32(target->rkey); buf->len = cpu_to_be32(ib_sg_dma_len(ibdev, scat)); - } else if (srp_map_fmr(target, scat, count, req, - (void *) cmd->add_data)) { + + req->nmdesc = 0; + goto map_complete; + } + + /* + * We have more than one scatter/gather entry, so build our indirect + * descriptor table, trying to merge as many entries as we can. + */ + indirect_hdr = (void *) cmd->add_data; + + ib_dma_sync_single_for_cpu(ibdev, req->indirect_dma_addr, + target->indirect_size, DMA_TO_DEVICE); + + memset(&state, 0, sizeof(state)); + srp_map_sg(&state, target, req, scat, count); + + /* We've mapped the request, now pull as much of the indirect + * descriptor table as we can into the command buffer. If this + * target is not using an external indirect table, we are + * guaranteed to fit into the command, as the SCSI layer won't + * give us more S/G entries than we allow. + */ + if (state.ndesc == 1) { /* - * FMR mapping failed, and the scatterlist has more - * than one entry. Generate an indirect memory - * descriptor. + * Memory registration collapsed the sg-list into one entry, + * so use a direct descriptor. */ - struct srp_indirect_buf *buf = (void *) cmd->add_data; - struct scatterlist *sg; - u32 datalen = 0; - int i; - - fmt = SRP_DATA_DESC_INDIRECT; - len = sizeof (struct srp_cmd) + - sizeof (struct srp_indirect_buf) + - count * sizeof (struct srp_direct_buf); - - scsi_for_each_sg(scmnd, sg, count, i) { - unsigned int dma_len = ib_sg_dma_len(ibdev, sg); - - buf->desc_list[i].va = - cpu_to_be64(ib_sg_dma_address(ibdev, sg)); - buf->desc_list[i].key = - cpu_to_be32(dev->mr->rkey); - buf->desc_list[i].len = cpu_to_be32(dma_len); - datalen += dma_len; - } - - if (scmnd->sc_data_direction == DMA_TO_DEVICE) - cmd->data_out_desc_cnt = count; - else - cmd->data_in_desc_cnt = count; + struct srp_direct_buf *buf = (void *) cmd->add_data; - buf->table_desc.va = - cpu_to_be64(req->cmd->dma + sizeof *cmd + sizeof *buf); - buf->table_desc.key = - cpu_to_be32(target->srp_host->srp_dev->mr->rkey); - buf->table_desc.len = - cpu_to_be32(count * sizeof (struct srp_direct_buf)); + *buf = req->indirect_desc[0]; + goto map_complete; + } - buf->len = cpu_to_be32(datalen); + if (unlikely(target->cmd_sg_cnt < state.ndesc && + !target->allow_ext_sg)) { + shost_printk(KERN_ERR, target->scsi_host, + "Could not fit S/G list into SRP_CMD\n"); + return -EIO; } + count = min(state.ndesc, target->cmd_sg_cnt); + table_len = state.ndesc * sizeof (struct srp_direct_buf); + + fmt = SRP_DATA_DESC_INDIRECT; + len = sizeof(struct srp_cmd) + sizeof (struct srp_indirect_buf); + len += count * sizeof (struct srp_direct_buf); + + memcpy(indirect_hdr->desc_list, req->indirect_desc, + count * sizeof (struct srp_direct_buf)); + + indirect_hdr->table_desc.va = cpu_to_be64(req->indirect_dma_addr); + indirect_hdr->table_desc.key = cpu_to_be32(target->rkey); + indirect_hdr->table_desc.len = cpu_to_be32(table_len); + indirect_hdr->len = cpu_to_be32(state.total_len); + + if (scmnd->sc_data_direction == DMA_TO_DEVICE) + cmd->data_out_desc_cnt = count; + else + cmd->data_in_desc_cnt = count; + + ib_dma_sync_single_for_device(ibdev, req->indirect_dma_addr, table_len, + DMA_TO_DEVICE); + +map_complete: if (scmnd->sc_data_direction == DMA_TO_DEVICE) cmd->buf_fmt = fmt << 4; else @@ -812,9 +1515,23 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, } /* - * Must be called with target->scsi_host->host_lock held to protect - * req_lim and tx_head. Lock cannot be dropped between call here and - * call to __srp_post_send(). + * Return an IU and possible credit to the free pool + */ +static void srp_put_tx_iu(struct srp_target_port *target, struct srp_iu *iu, + enum srp_iu_type iu_type) +{ + unsigned long flags; + + spin_lock_irqsave(&target->lock, flags); + list_add(&iu->list, &target->free_tx); + if (iu_type != SRP_IU_RSP) + ++target->req_lim; + spin_unlock_irqrestore(&target->lock, flags); +} + +/* + * Must be called with target->lock held to protect req_lim and free_tx. + * If IU is not sent, it must be returned using srp_put_tx_iu(). * * Note: * An upper limit for the number of allocated information units for each @@ -833,83 +1550,59 @@ static struct srp_iu *__srp_get_tx_iu(struct srp_target_port *target, srp_send_completion(target->send_cq, target); - if (target->tx_head - target->tx_tail >= SRP_SQ_SIZE) + if (list_empty(&target->free_tx)) return NULL; /* Initiator responses to target requests do not consume credits */ - if (target->req_lim <= rsv && iu_type != SRP_IU_RSP) { - ++target->zero_req_lim; - return NULL; + if (iu_type != SRP_IU_RSP) { + if (target->req_lim <= rsv) { + ++target->zero_req_lim; + return NULL; + } + + --target->req_lim; } - iu = target->tx_ring[target->tx_head & SRP_SQ_MASK]; - iu->type = iu_type; + iu = list_first_entry(&target->free_tx, struct srp_iu, list); + list_del(&iu->list); return iu; } -/* - * Must be called with target->scsi_host->host_lock held to protect - * req_lim and tx_head. - */ -static int __srp_post_send(struct srp_target_port *target, - struct srp_iu *iu, int len) +static int srp_post_send(struct srp_target_port *target, + struct srp_iu *iu, int len) { struct ib_sge list; struct ib_send_wr wr, *bad_wr; - int ret = 0; list.addr = iu->dma; list.length = len; - list.lkey = target->srp_host->srp_dev->mr->lkey; + list.lkey = target->lkey; wr.next = NULL; - wr.wr_id = target->tx_head & SRP_SQ_MASK; + wr.wr_id = (uintptr_t) iu; wr.sg_list = &list; wr.num_sge = 1; wr.opcode = IB_WR_SEND; wr.send_flags = IB_SEND_SIGNALED; - ret = ib_post_send(target->qp, &wr, &bad_wr); - - if (!ret) { - ++target->tx_head; - if (iu->type != SRP_IU_RSP) - --target->req_lim; - } - - return ret; + return ib_post_send(target->qp, &wr, &bad_wr); } -static int srp_post_recv(struct srp_target_port *target) +static int srp_post_recv(struct srp_target_port *target, struct srp_iu *iu) { - unsigned long flags; - struct srp_iu *iu; - struct ib_sge list; struct ib_recv_wr wr, *bad_wr; - unsigned int next; - int ret; - - spin_lock_irqsave(target->scsi_host->host_lock, flags); - - next = target->rx_head & SRP_RQ_MASK; - wr.wr_id = next; - iu = target->rx_ring[next]; + struct ib_sge list; list.addr = iu->dma; list.length = iu->size; - list.lkey = target->srp_host->srp_dev->mr->lkey; + list.lkey = target->lkey; wr.next = NULL; + wr.wr_id = (uintptr_t) iu; wr.sg_list = &list; wr.num_sge = 1; - ret = ib_post_recv(target->qp, &wr, &bad_wr); - if (!ret) - ++target->rx_head; - - spin_unlock_irqrestore(target->scsi_host->host_lock, flags); - - return ret; + return ib_post_recv(target->qp, &wr, &bad_wr); } static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp) @@ -917,28 +1610,30 @@ static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp) struct srp_request *req; struct scsi_cmnd *scmnd; unsigned long flags; - s32 delta; - - delta = (s32) be32_to_cpu(rsp->req_lim_delta); - - spin_lock_irqsave(target->scsi_host->host_lock, flags); - - target->req_lim += delta; - - req = &target->req_ring[rsp->tag & ~SRP_TAG_TSK_MGMT]; if (unlikely(rsp->tag & SRP_TAG_TSK_MGMT)) { - if (be32_to_cpu(rsp->resp_data_len) < 4) - req->tsk_status = -1; - else - req->tsk_status = rsp->data[3]; - complete(&req->done); + spin_lock_irqsave(&target->lock, flags); + target->req_lim += be32_to_cpu(rsp->req_lim_delta); + spin_unlock_irqrestore(&target->lock, flags); + + target->tsk_mgmt_status = -1; + if (be32_to_cpu(rsp->resp_data_len) >= 4) + target->tsk_mgmt_status = rsp->data[3]; + complete(&target->tsk_mgmt_done); } else { - scmnd = req->scmnd; - if (!scmnd) + req = &target->req_ring[rsp->tag]; + scmnd = srp_claim_req(target, req, NULL, NULL); + if (!scmnd) { shost_printk(KERN_ERR, target->scsi_host, "Null scmnd for RSP w/tag %016llx\n", (unsigned long long) rsp->tag); + + spin_lock_irqsave(&target->lock, flags); + target->req_lim += be32_to_cpu(rsp->req_lim_delta); + spin_unlock_irqrestore(&target->lock, flags); + + return; + } scmnd->result = rsp->status; if (rsp->flags & SRP_RSP_FLAG_SNSVALID) { @@ -953,49 +1648,44 @@ static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp) else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER)) scsi_set_resid(scmnd, be32_to_cpu(rsp->data_in_res_cnt)); - if (!req->tsk_mgmt) { - scmnd->host_scribble = (void *) -1L; - scmnd->scsi_done(scmnd); + srp_free_req(target, req, scmnd, + be32_to_cpu(rsp->req_lim_delta)); - srp_remove_req(target, req); - } else - req->cmd_done = 1; + scmnd->host_scribble = NULL; + scmnd->scsi_done(scmnd); } - - spin_unlock_irqrestore(target->scsi_host->host_lock, flags); } static int srp_response_common(struct srp_target_port *target, s32 req_delta, void *rsp, int len) { - struct ib_device *dev; + struct ib_device *dev = target->srp_host->srp_dev->dev; unsigned long flags; struct srp_iu *iu; - int err = 1; + int err; - dev = target->srp_host->srp_dev->dev; - - spin_lock_irqsave(target->scsi_host->host_lock, flags); + spin_lock_irqsave(&target->lock, flags); target->req_lim += req_delta; - iu = __srp_get_tx_iu(target, SRP_IU_RSP); + spin_unlock_irqrestore(&target->lock, flags); + if (!iu) { shost_printk(KERN_ERR, target->scsi_host, PFX "no IU available to send response\n"); - goto out; + return 1; } ib_dma_sync_single_for_cpu(dev, iu->dma, len, DMA_TO_DEVICE); memcpy(iu->buf, rsp, len); ib_dma_sync_single_for_device(dev, iu->dma, len, DMA_TO_DEVICE); - err = __srp_post_send(target, iu, len); - if (err) + err = srp_post_send(target, iu, len); + if (err) { shost_printk(KERN_ERR, target->scsi_host, PFX "unable to post response: %d\n", err); + srp_put_tx_iu(target, iu, SRP_IU_RSP); + } -out: - spin_unlock_irqrestore(target->scsi_host->host_lock, flags); return err; } @@ -1032,14 +1722,11 @@ static void srp_process_aer_req(struct srp_target_port *target, static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc) { - struct ib_device *dev; - struct srp_iu *iu; + struct ib_device *dev = target->srp_host->srp_dev->dev; + struct srp_iu *iu = (struct srp_iu *) (uintptr_t) wc->wr_id; int res; u8 opcode; - iu = target->rx_ring[wc->wr_id]; - - dev = target->srp_host->srp_dev->dev; ib_dma_sync_single_for_cpu(dev, iu->dma, target->max_ti_iu_len, DMA_FROM_DEVICE); @@ -1080,12 +1767,51 @@ static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc) ib_dma_sync_single_for_device(dev, iu->dma, target->max_ti_iu_len, DMA_FROM_DEVICE); - res = srp_post_recv(target); + res = srp_post_recv(target, iu); if (res != 0) shost_printk(KERN_ERR, target->scsi_host, PFX "Recv failed with error code %d\n", res); } +/** + * srp_tl_err_work() - handle a transport layer error + * @work: Work structure embedded in an SRP target port. + * + * Note: This function may get invoked before the rport has been created, + * hence the target->rport test. + */ +static void srp_tl_err_work(struct work_struct *work) +{ + struct srp_target_port *target; + + target = container_of(work, struct srp_target_port, tl_err_work); + if (target->rport) + srp_start_tl_fail_timers(target->rport); +} + +static void srp_handle_qp_err(u64 wr_id, enum ib_wc_status wc_status, + bool send_err, struct srp_target_port *target) +{ + if (target->connected && !target->qp_in_error) { + if (wr_id & LOCAL_INV_WR_ID_MASK) { + shost_printk(KERN_ERR, target->scsi_host, PFX + "LOCAL_INV failed with status %d\n", + wc_status); + } else if (wr_id & FAST_REG_WR_ID_MASK) { + shost_printk(KERN_ERR, target->scsi_host, PFX + "FAST_REG_MR failed status %d\n", + wc_status); + } else { + shost_printk(KERN_ERR, target->scsi_host, + PFX "failed %s status %d for iu %p\n", + send_err ? "send" : "receive", + wc_status, (void *)(uintptr_t)wr_id); + } + queue_work(system_long_wq, &target->tl_err_work); + } + target->qp_in_error = true; +} + static void srp_recv_completion(struct ib_cq *cq, void *target_ptr) { struct srp_target_port *target = target_ptr; @@ -1093,15 +1819,11 @@ static void srp_recv_completion(struct ib_cq *cq, void *target_ptr) ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); while (ib_poll_cq(cq, 1, &wc) > 0) { - if (wc.status) { - shost_printk(KERN_ERR, target->scsi_host, - PFX "failed receive status %d\n", - wc.status); - target->qp_in_error = 1; - break; + if (likely(wc.status == IB_WC_SUCCESS)) { + srp_handle_recv(target, &wc); + } else { + srp_handle_qp_err(wc.wr_id, wc.status, false, target); } - - srp_handle_recv(target, &wc); } } @@ -1109,53 +1831,57 @@ static void srp_send_completion(struct ib_cq *cq, void *target_ptr) { struct srp_target_port *target = target_ptr; struct ib_wc wc; + struct srp_iu *iu; while (ib_poll_cq(cq, 1, &wc) > 0) { - if (wc.status) { - shost_printk(KERN_ERR, target->scsi_host, - PFX "failed send status %d\n", - wc.status); - target->qp_in_error = 1; - break; + if (likely(wc.status == IB_WC_SUCCESS)) { + iu = (struct srp_iu *) (uintptr_t) wc.wr_id; + list_add(&iu->list, &target->free_tx); + } else { + srp_handle_qp_err(wc.wr_id, wc.status, true, target); } - - ++target->tx_tail; } } -static int srp_queuecommand(struct scsi_cmnd *scmnd, - void (*done)(struct scsi_cmnd *)) +static int srp_queuecommand(struct Scsi_Host *shost, struct scsi_cmnd *scmnd) { - struct srp_target_port *target = host_to_target(scmnd->device->host); + struct srp_target_port *target = host_to_target(shost); + struct srp_rport *rport = target->rport; struct srp_request *req; struct srp_iu *iu; struct srp_cmd *cmd; struct ib_device *dev; - int len; + unsigned long flags; + int len, ret; + const bool in_scsi_eh = !in_interrupt() && current == shost->ehandler; - if (target->state == SRP_TARGET_CONNECTING) - goto err; + /* + * The SCSI EH thread is the only context from which srp_queuecommand() + * can get invoked for blocked devices (SDEV_BLOCK / + * SDEV_CREATED_BLOCK). Avoid racing with srp_reconnect_rport() by + * locking the rport mutex if invoked from inside the SCSI EH. + */ + if (in_scsi_eh) + mutex_lock(&rport->mutex); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) { - scmnd->result = DID_BAD_TARGET << 16; - done(scmnd); - return 0; - } + scmnd->result = srp_chkready(target->rport); + if (unlikely(scmnd->result)) + goto err; + spin_lock_irqsave(&target->lock, flags); iu = __srp_get_tx_iu(target, SRP_IU_CMD); if (!iu) - goto err; + goto err_unlock; + + req = list_first_entry(&target->free_reqs, struct srp_request, list); + list_del(&req->list); + spin_unlock_irqrestore(&target->lock, flags); dev = target->srp_host->srp_dev->dev; - ib_dma_sync_single_for_cpu(dev, iu->dma, srp_max_iu_len, + ib_dma_sync_single_for_cpu(dev, iu->dma, target->max_iu_len, DMA_TO_DEVICE); - req = list_first_entry(&target->free_reqs, struct srp_request, list); - - scmnd->scsi_done = done; - scmnd->result = 0; - scmnd->host_scribble = (void *) (long) req->index; + scmnd->host_scribble = (void *) req; cmd = iu->buf; memset(cmd, 0, sizeof *cmd); @@ -1167,40 +1893,85 @@ static int srp_queuecommand(struct scsi_cmnd *scmnd, req->scmnd = scmnd; req->cmd = iu; - req->cmd_done = 0; - req->tsk_mgmt = NULL; len = srp_map_data(scmnd, target, req); if (len < 0) { shost_printk(KERN_ERR, target->scsi_host, - PFX "Failed to map data\n"); - goto err; + PFX "Failed to map data (%d)\n", len); + /* + * If we ran out of memory descriptors (-ENOMEM) because an + * application is queuing many requests with more than + * max_pages_per_mr sg-list elements, tell the SCSI mid-layer + * to reduce queue depth temporarily. + */ + scmnd->result = len == -ENOMEM ? + DID_OK << 16 | QUEUE_FULL << 1 : DID_ERROR << 16; + goto err_iu; } - ib_dma_sync_single_for_device(dev, iu->dma, srp_max_iu_len, + ib_dma_sync_single_for_device(dev, iu->dma, target->max_iu_len, DMA_TO_DEVICE); - if (__srp_post_send(target, iu, len)) { + if (srp_post_send(target, iu, len)) { shost_printk(KERN_ERR, target->scsi_host, PFX "Send failed\n"); goto err_unmap; } - list_move_tail(&req->list, &target->req_queue); + ret = 0; + +unlock_rport: + if (in_scsi_eh) + mutex_unlock(&rport->mutex); - return 0; + return ret; err_unmap: srp_unmap_data(scmnd, target, req); +err_iu: + srp_put_tx_iu(target, iu, SRP_IU_CMD); + + /* + * Avoid that the loops that iterate over the request ring can + * encounter a dangling SCSI command pointer. + */ + req->scmnd = NULL; + + spin_lock_irqsave(&target->lock, flags); + list_add(&req->list, &target->free_reqs); + +err_unlock: + spin_unlock_irqrestore(&target->lock, flags); + err: - return SCSI_MLQUEUE_HOST_BUSY; + if (scmnd->result) { + scmnd->scsi_done(scmnd); + ret = 0; + } else { + ret = SCSI_MLQUEUE_HOST_BUSY; + } + + goto unlock_rport; } +/* + * Note: the resources allocated in this function are freed in + * srp_free_target_ib(). + */ static int srp_alloc_iu_bufs(struct srp_target_port *target) { int i; - for (i = 0; i < SRP_RQ_SIZE; ++i) { + target->rx_ring = kzalloc(target->queue_size * sizeof(*target->rx_ring), + GFP_KERNEL); + if (!target->rx_ring) + goto err_no_ring; + target->tx_ring = kzalloc(target->queue_size * sizeof(*target->tx_ring), + GFP_KERNEL); + if (!target->tx_ring) + goto err_no_ring; + + for (i = 0; i < target->queue_size; ++i) { target->rx_ring[i] = srp_alloc_iu(target->srp_host, target->max_ti_iu_len, GFP_KERNEL, DMA_FROM_DEVICE); @@ -1208,30 +1979,138 @@ static int srp_alloc_iu_bufs(struct srp_target_port *target) goto err; } - for (i = 0; i < SRP_SQ_SIZE; ++i) { + for (i = 0; i < target->queue_size; ++i) { target->tx_ring[i] = srp_alloc_iu(target->srp_host, - srp_max_iu_len, + target->max_iu_len, GFP_KERNEL, DMA_TO_DEVICE); if (!target->tx_ring[i]) goto err; + + list_add(&target->tx_ring[i]->list, &target->free_tx); } return 0; err: - for (i = 0; i < SRP_RQ_SIZE; ++i) { + for (i = 0; i < target->queue_size; ++i) { srp_free_iu(target->srp_host, target->rx_ring[i]); - target->rx_ring[i] = NULL; - } - - for (i = 0; i < SRP_SQ_SIZE; ++i) { srp_free_iu(target->srp_host, target->tx_ring[i]); - target->tx_ring[i] = NULL; } + +err_no_ring: + kfree(target->tx_ring); + target->tx_ring = NULL; + kfree(target->rx_ring); + target->rx_ring = NULL; + return -ENOMEM; } +static uint32_t srp_compute_rq_tmo(struct ib_qp_attr *qp_attr, int attr_mask) +{ + uint64_t T_tr_ns, max_compl_time_ms; + uint32_t rq_tmo_jiffies; + + /* + * According to section 11.2.4.2 in the IBTA spec (Modify Queue Pair, + * table 91), both the QP timeout and the retry count have to be set + * for RC QP's during the RTR to RTS transition. + */ + WARN_ON_ONCE((attr_mask & (IB_QP_TIMEOUT | IB_QP_RETRY_CNT)) != + (IB_QP_TIMEOUT | IB_QP_RETRY_CNT)); + + /* + * Set target->rq_tmo_jiffies to one second more than the largest time + * it can take before an error completion is generated. See also + * C9-140..142 in the IBTA spec for more information about how to + * convert the QP Local ACK Timeout value to nanoseconds. + */ + T_tr_ns = 4096 * (1ULL << qp_attr->timeout); + max_compl_time_ms = qp_attr->retry_cnt * 4 * T_tr_ns; + do_div(max_compl_time_ms, NSEC_PER_MSEC); + rq_tmo_jiffies = msecs_to_jiffies(max_compl_time_ms + 1000); + + return rq_tmo_jiffies; +} + +static void srp_cm_rep_handler(struct ib_cm_id *cm_id, + struct srp_login_rsp *lrsp, + struct srp_target_port *target) +{ + struct ib_qp_attr *qp_attr = NULL; + int attr_mask = 0; + int ret; + int i; + + if (lrsp->opcode == SRP_LOGIN_RSP) { + target->max_ti_iu_len = be32_to_cpu(lrsp->max_ti_iu_len); + target->req_lim = be32_to_cpu(lrsp->req_lim_delta); + + /* + * Reserve credits for task management so we don't + * bounce requests back to the SCSI mid-layer. + */ + target->scsi_host->can_queue + = min(target->req_lim - SRP_TSK_MGMT_SQ_SIZE, + target->scsi_host->can_queue); + target->scsi_host->cmd_per_lun + = min_t(int, target->scsi_host->can_queue, + target->scsi_host->cmd_per_lun); + } else { + shost_printk(KERN_WARNING, target->scsi_host, + PFX "Unhandled RSP opcode %#x\n", lrsp->opcode); + ret = -ECONNRESET; + goto error; + } + + if (!target->rx_ring) { + ret = srp_alloc_iu_bufs(target); + if (ret) + goto error; + } + + ret = -ENOMEM; + qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); + if (!qp_attr) + goto error; + + qp_attr->qp_state = IB_QPS_RTR; + ret = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); + if (ret) + goto error_free; + + ret = ib_modify_qp(target->qp, qp_attr, attr_mask); + if (ret) + goto error_free; + + for (i = 0; i < target->queue_size; i++) { + struct srp_iu *iu = target->rx_ring[i]; + ret = srp_post_recv(target, iu); + if (ret) + goto error_free; + } + + qp_attr->qp_state = IB_QPS_RTS; + ret = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); + if (ret) + goto error_free; + + target->rq_tmo_jiffies = srp_compute_rq_tmo(qp_attr, attr_mask); + + ret = ib_modify_qp(target->qp, qp_attr, attr_mask); + if (ret) + goto error_free; + + ret = ib_send_cm_rtu(cm_id, NULL, 0); + +error_free: + kfree(qp_attr); + +error: + target->status = ret; +} + static void srp_cm_rej_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event, struct srp_target_port *target) @@ -1291,8 +2170,10 @@ static void srp_cm_rej_handler(struct ib_cm_id *cm_id, shost_printk(KERN_WARNING, shost, PFX "SRP_LOGIN_REJ: requested max_it_iu_len too large\n"); else - shost_printk(KERN_WARNING, shost, - PFX "SRP LOGIN REJECTED, reason 0x%08x\n", reason); + shost_printk(KERN_WARNING, shost, PFX + "SRP LOGIN from %pI6 to %pI6 REJECTED, reason 0x%08x\n", + target->path.sgid.raw, + target->orig_dgid, reason); } else shost_printk(KERN_WARNING, shost, " REJ reason: IB_CM_REJ_CONSUMER_DEFINED," @@ -1315,11 +2196,7 @@ static void srp_cm_rej_handler(struct ib_cm_id *cm_id, static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) { struct srp_target_port *target = cm_id->context; - struct ib_qp_attr *qp_attr = NULL; - int attr_mask = 0; int comp = 0; - int opcode = 0; - int i; switch (event->event) { case IB_CM_REQ_ERROR: @@ -1331,70 +2208,7 @@ static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) case IB_CM_REP_RECEIVED: comp = 1; - opcode = *(u8 *) event->private_data; - - if (opcode == SRP_LOGIN_RSP) { - struct srp_login_rsp *rsp = event->private_data; - - target->max_ti_iu_len = be32_to_cpu(rsp->max_ti_iu_len); - target->req_lim = be32_to_cpu(rsp->req_lim_delta); - - /* - * Reserve credits for task management so we don't - * bounce requests back to the SCSI mid-layer. - */ - target->scsi_host->can_queue - = min(target->req_lim - SRP_TSK_MGMT_SQ_SIZE, - target->scsi_host->can_queue); - } else { - shost_printk(KERN_WARNING, target->scsi_host, - PFX "Unhandled RSP opcode %#x\n", opcode); - target->status = -ECONNRESET; - break; - } - - if (!target->rx_ring[0]) { - target->status = srp_alloc_iu_bufs(target); - if (target->status) - break; - } - - qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); - if (!qp_attr) { - target->status = -ENOMEM; - break; - } - - qp_attr->qp_state = IB_QPS_RTR; - target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); - if (target->status) - break; - - target->status = ib_modify_qp(target->qp, qp_attr, attr_mask); - if (target->status) - break; - - for (i = 0; i < SRP_RQ_SIZE; i++) { - target->status = srp_post_recv(target); - if (target->status) - break; - } - if (target->status) - break; - - qp_attr->qp_state = IB_QPS_RTS; - target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); - if (target->status) - break; - - target->status = ib_modify_qp(target->qp, qp_attr, attr_mask); - if (target->status) - break; - - target->status = ib_send_cm_rtu(cm_id, NULL, 0); - if (target->status) - break; - + srp_cm_rep_handler(cm_id, event->private_data, target); break; case IB_CM_REJ_RECEIVED: @@ -1407,16 +2221,18 @@ static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) case IB_CM_DREQ_RECEIVED: shost_printk(KERN_WARNING, target->scsi_host, PFX "DREQ received - connection closed\n"); + srp_change_conn_state(target, false); if (ib_send_cm_drep(cm_id, NULL, 0)) shost_printk(KERN_ERR, target->scsi_host, PFX "Sending CM DREP failed\n"); + queue_work(system_long_wq, &target->tl_err_work); break; case IB_CM_TIMEWAIT_EXIT: shost_printk(KERN_ERR, target->scsi_host, PFX "connection closed\n"); - comp = 1; + target->status = 0; break; @@ -1434,31 +2250,87 @@ static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) if (comp) complete(&target->done); - kfree(qp_attr); - return 0; } +/** + * srp_change_queue_type - changing device queue tag type + * @sdev: scsi device struct + * @tag_type: requested tag type + * + * Returns queue tag type. + */ +static int +srp_change_queue_type(struct scsi_device *sdev, int tag_type) +{ + if (sdev->tagged_supported) { + scsi_set_tag_type(sdev, tag_type); + if (tag_type) + scsi_activate_tcq(sdev, sdev->queue_depth); + else + scsi_deactivate_tcq(sdev, sdev->queue_depth); + } else + tag_type = 0; + + return tag_type; +} + +/** + * srp_change_queue_depth - setting device queue depth + * @sdev: scsi device struct + * @qdepth: requested queue depth + * @reason: SCSI_QDEPTH_DEFAULT/SCSI_QDEPTH_QFULL/SCSI_QDEPTH_RAMP_UP + * (see include/scsi/scsi_host.h for definition) + * + * Returns queue depth. + */ +static int +srp_change_queue_depth(struct scsi_device *sdev, int qdepth, int reason) +{ + struct Scsi_Host *shost = sdev->host; + int max_depth; + if (reason == SCSI_QDEPTH_DEFAULT || reason == SCSI_QDEPTH_RAMP_UP) { + max_depth = shost->can_queue; + if (!sdev->tagged_supported) + max_depth = 1; + if (qdepth > max_depth) + qdepth = max_depth; + scsi_adjust_queue_depth(sdev, scsi_get_tag_type(sdev), qdepth); + } else if (reason == SCSI_QDEPTH_QFULL) + scsi_track_queue_full(sdev, qdepth); + else + return -EOPNOTSUPP; + + return sdev->queue_depth; +} + static int srp_send_tsk_mgmt(struct srp_target_port *target, - struct srp_request *req, u8 func) + u64 req_tag, unsigned int lun, u8 func) { + struct srp_rport *rport = target->rport; struct ib_device *dev = target->srp_host->srp_dev->dev; struct srp_iu *iu; struct srp_tsk_mgmt *tsk_mgmt; - spin_lock_irq(target->scsi_host->host_lock); - - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) { - req->scmnd->result = DID_BAD_TARGET << 16; - goto out; - } + if (!target->connected || target->qp_in_error) + return -1; - init_completion(&req->done); + init_completion(&target->tsk_mgmt_done); + /* + * Lock the rport mutex to avoid that srp_create_target_ib() is + * invoked while a task management function is being sent. + */ + mutex_lock(&rport->mutex); + spin_lock_irq(&target->lock); iu = __srp_get_tx_iu(target, SRP_IU_TSK_MGMT); - if (!iu) - goto out; + spin_unlock_irq(&target->lock); + + if (!iu) { + mutex_unlock(&rport->mutex); + + return -1; + } ib_dma_sync_single_for_cpu(dev, iu->dma, sizeof *tsk_mgmt, DMA_TO_DEVICE); @@ -1466,70 +2338,48 @@ static int srp_send_tsk_mgmt(struct srp_target_port *target, memset(tsk_mgmt, 0, sizeof *tsk_mgmt); tsk_mgmt->opcode = SRP_TSK_MGMT; - tsk_mgmt->lun = cpu_to_be64((u64) req->scmnd->device->lun << 48); - tsk_mgmt->tag = req->index | SRP_TAG_TSK_MGMT; + tsk_mgmt->lun = cpu_to_be64((u64) lun << 48); + tsk_mgmt->tag = req_tag | SRP_TAG_TSK_MGMT; tsk_mgmt->tsk_mgmt_func = func; - tsk_mgmt->task_tag = req->index; + tsk_mgmt->task_tag = req_tag; ib_dma_sync_single_for_device(dev, iu->dma, sizeof *tsk_mgmt, DMA_TO_DEVICE); - if (__srp_post_send(target, iu, sizeof *tsk_mgmt)) - goto out; - - req->tsk_mgmt = iu; - - spin_unlock_irq(target->scsi_host->host_lock); + if (srp_post_send(target, iu, sizeof *tsk_mgmt)) { + srp_put_tx_iu(target, iu, SRP_IU_TSK_MGMT); + mutex_unlock(&rport->mutex); - if (!wait_for_completion_timeout(&req->done, - msecs_to_jiffies(SRP_ABORT_TIMEOUT_MS))) return -1; + } + mutex_unlock(&rport->mutex); - return 0; - -out: - spin_unlock_irq(target->scsi_host->host_lock); - return -1; -} - -static int srp_find_req(struct srp_target_port *target, - struct scsi_cmnd *scmnd, - struct srp_request **req) -{ - if (scmnd->host_scribble == (void *) -1L) + if (!wait_for_completion_timeout(&target->tsk_mgmt_done, + msecs_to_jiffies(SRP_ABORT_TIMEOUT_MS))) return -1; - *req = &target->req_ring[(long) scmnd->host_scribble]; - return 0; } static int srp_abort(struct scsi_cmnd *scmnd) { struct srp_target_port *target = host_to_target(scmnd->device->host); - struct srp_request *req; - int ret = SUCCESS; + struct srp_request *req = (struct srp_request *) scmnd->host_scribble; + int ret; shost_printk(KERN_ERR, target->scsi_host, "SRP abort called\n"); - if (target->qp_in_error) - return FAILED; - if (srp_find_req(target, scmnd, &req)) - return FAILED; - if (srp_send_tsk_mgmt(target, req, SRP_TSK_ABORT_TASK)) - return FAILED; - - spin_lock_irq(target->scsi_host->host_lock); - - if (req->cmd_done) { - srp_remove_req(target, req); - scmnd->scsi_done(scmnd); - } else if (!req->tsk_status) { - srp_remove_req(target, req); - scmnd->result = DID_ABORT << 16; - } else + if (!req || !srp_claim_req(target, req, NULL, scmnd)) + return SUCCESS; + if (srp_send_tsk_mgmt(target, req->index, scmnd->device->lun, + SRP_TSK_ABORT_TASK) == 0) + ret = SUCCESS; + else if (target->rport->state == SRP_RPORT_LOST) + ret = FAST_IO_FAIL; + else ret = FAILED; - - spin_unlock_irq(target->scsi_host->host_lock); + srp_free_req(target, req, scmnd, 0); + scmnd->result = DID_ABORT << 16; + scmnd->scsi_done(scmnd); return ret; } @@ -1537,26 +2387,20 @@ static int srp_abort(struct scsi_cmnd *scmnd) static int srp_reset_device(struct scsi_cmnd *scmnd) { struct srp_target_port *target = host_to_target(scmnd->device->host); - struct srp_request *req, *tmp; + int i; shost_printk(KERN_ERR, target->scsi_host, "SRP reset_device called\n"); - if (target->qp_in_error) + if (srp_send_tsk_mgmt(target, SRP_TAG_NO_REQ, scmnd->device->lun, + SRP_TSK_LUN_RESET)) return FAILED; - if (srp_find_req(target, scmnd, &req)) + if (target->tsk_mgmt_status) return FAILED; - if (srp_send_tsk_mgmt(target, req, SRP_TSK_LUN_RESET)) - return FAILED; - if (req->tsk_status) - return FAILED; - - spin_lock_irq(target->scsi_host->host_lock); - list_for_each_entry_safe(req, tmp, &target->req_queue, list) - if (req->scmnd->device == scmnd->device) - srp_reset_req(target, req); - - spin_unlock_irq(target->scsi_host->host_lock); + for (i = 0; i < target->req_ring_size; ++i) { + struct srp_request *req = &target->req_ring[i]; + srp_finish_req(target, req, scmnd->device, DID_RESET << 16); + } return SUCCESS; } @@ -1564,14 +2408,25 @@ static int srp_reset_device(struct scsi_cmnd *scmnd) static int srp_reset_host(struct scsi_cmnd *scmnd) { struct srp_target_port *target = host_to_target(scmnd->device->host); - int ret = FAILED; shost_printk(KERN_ERR, target->scsi_host, PFX "SRP reset_host called\n"); - if (!srp_reconnect_target(target)) - ret = SUCCESS; + return srp_reconnect_rport(target->rport) == 0 ? SUCCESS : FAILED; +} - return ret; +static int srp_slave_configure(struct scsi_device *sdev) +{ + struct Scsi_Host *shost = sdev->host; + struct srp_target_port *target = host_to_target(shost); + struct request_queue *q = sdev->request_queue; + unsigned long timeout; + + if (sdev->type == TYPE_DISK) { + timeout = max_t(unsigned, 30 * HZ, target->rq_tmo_jiffies); + blk_queue_rq_timeout(q, timeout); + } + + return 0; } static ssize_t show_id_ext(struct device *dev, struct device_attribute *attr, @@ -1579,10 +2434,6 @@ static ssize_t show_id_ext(struct device *dev, struct device_attribute *attr, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "0x%016llx\n", (unsigned long long) be64_to_cpu(target->id_ext)); } @@ -1592,10 +2443,6 @@ static ssize_t show_ioc_guid(struct device *dev, struct device_attribute *attr, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "0x%016llx\n", (unsigned long long) be64_to_cpu(target->ioc_guid)); } @@ -1605,10 +2452,6 @@ static ssize_t show_service_id(struct device *dev, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "0x%016llx\n", (unsigned long long) be64_to_cpu(target->service_id)); } @@ -1618,21 +2461,21 @@ static ssize_t show_pkey(struct device *dev, struct device_attribute *attr, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "0x%04x\n", be16_to_cpu(target->path.pkey)); } -static ssize_t show_dgid(struct device *dev, struct device_attribute *attr, +static ssize_t show_sgid(struct device *dev, struct device_attribute *attr, char *buf) { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; + return sprintf(buf, "%pI6\n", target->path.sgid.raw); +} + +static ssize_t show_dgid(struct device *dev, struct device_attribute *attr, + char *buf) +{ + struct srp_target_port *target = host_to_target(class_to_shost(dev)); return sprintf(buf, "%pI6\n", target->path.dgid.raw); } @@ -1642,10 +2485,6 @@ static ssize_t show_orig_dgid(struct device *dev, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "%pI6\n", target->orig_dgid); } @@ -1654,10 +2493,6 @@ static ssize_t show_req_lim(struct device *dev, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "%d\n", target->req_lim); } @@ -1666,10 +2501,6 @@ static ssize_t show_zero_req_lim(struct device *dev, { struct srp_target_port *target = host_to_target(class_to_shost(dev)); - if (target->state == SRP_TARGET_DEAD || - target->state == SRP_TARGET_REMOVED) - return -ENODEV; - return sprintf(buf, "%d\n", target->zero_req_lim); } @@ -1689,28 +2520,70 @@ static ssize_t show_local_ib_device(struct device *dev, return sprintf(buf, "%s\n", target->srp_host->srp_dev->dev->name); } +static ssize_t show_comp_vector(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct srp_target_port *target = host_to_target(class_to_shost(dev)); + + return sprintf(buf, "%d\n", target->comp_vector); +} + +static ssize_t show_tl_retry_count(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct srp_target_port *target = host_to_target(class_to_shost(dev)); + + return sprintf(buf, "%d\n", target->tl_retry_count); +} + +static ssize_t show_cmd_sg_entries(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct srp_target_port *target = host_to_target(class_to_shost(dev)); + + return sprintf(buf, "%u\n", target->cmd_sg_cnt); +} + +static ssize_t show_allow_ext_sg(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct srp_target_port *target = host_to_target(class_to_shost(dev)); + + return sprintf(buf, "%s\n", target->allow_ext_sg ? "true" : "false"); +} + static DEVICE_ATTR(id_ext, S_IRUGO, show_id_ext, NULL); static DEVICE_ATTR(ioc_guid, S_IRUGO, show_ioc_guid, NULL); static DEVICE_ATTR(service_id, S_IRUGO, show_service_id, NULL); static DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL); +static DEVICE_ATTR(sgid, S_IRUGO, show_sgid, NULL); static DEVICE_ATTR(dgid, S_IRUGO, show_dgid, NULL); static DEVICE_ATTR(orig_dgid, S_IRUGO, show_orig_dgid, NULL); static DEVICE_ATTR(req_lim, S_IRUGO, show_req_lim, NULL); static DEVICE_ATTR(zero_req_lim, S_IRUGO, show_zero_req_lim, NULL); static DEVICE_ATTR(local_ib_port, S_IRUGO, show_local_ib_port, NULL); static DEVICE_ATTR(local_ib_device, S_IRUGO, show_local_ib_device, NULL); +static DEVICE_ATTR(comp_vector, S_IRUGO, show_comp_vector, NULL); +static DEVICE_ATTR(tl_retry_count, S_IRUGO, show_tl_retry_count, NULL); +static DEVICE_ATTR(cmd_sg_entries, S_IRUGO, show_cmd_sg_entries, NULL); +static DEVICE_ATTR(allow_ext_sg, S_IRUGO, show_allow_ext_sg, NULL); static struct device_attribute *srp_host_attrs[] = { &dev_attr_id_ext, &dev_attr_ioc_guid, &dev_attr_service_id, &dev_attr_pkey, + &dev_attr_sgid, &dev_attr_dgid, &dev_attr_orig_dgid, &dev_attr_req_lim, &dev_attr_zero_req_lim, &dev_attr_local_ib_port, &dev_attr_local_ib_device, + &dev_attr_comp_vector, + &dev_attr_tl_retry_count, + &dev_attr_cmd_sg_entries, + &dev_attr_allow_ext_sg, NULL }; @@ -1718,14 +2591,19 @@ static struct scsi_host_template srp_template = { .module = THIS_MODULE, .name = "InfiniBand SRP initiator", .proc_name = DRV_NAME, + .slave_configure = srp_slave_configure, .info = srp_target_info, .queuecommand = srp_queuecommand, + .change_queue_depth = srp_change_queue_depth, + .change_queue_type = srp_change_queue_type, .eh_abort_handler = srp_abort, .eh_device_reset_handler = srp_reset_device, .eh_host_reset_handler = srp_reset_host, - .can_queue = SRP_CMD_SQ_SIZE, + .skip_settle_delay = true, + .sg_tablesize = SRP_DEF_SG_TABLESIZE, + .can_queue = SRP_DEFAULT_CMD_SQ_SIZE, .this_id = -1, - .cmd_per_lun = SRP_CMD_SQ_SIZE, + .cmd_per_lun = SRP_DEFAULT_CMD_SQ_SIZE, .use_clustering = ENABLE_CLUSTERING, .shost_attrs = srp_host_attrs }; @@ -1750,6 +2628,9 @@ static int srp_add_target(struct srp_host *host, struct srp_target_port *target) return PTR_ERR(rport); } + rport->lld_data = target; + target->rport = rport; + spin_lock(&host->target_lock); list_add_tail(&target->list, &host->target_list); spin_unlock(&host->target_lock); @@ -1775,6 +2656,38 @@ static struct class srp_class = { .dev_release = srp_release_dev }; +/** + * srp_conn_unique() - check whether the connection to a target is unique + * @host: SRP host. + * @target: SRP target port. + */ +static bool srp_conn_unique(struct srp_host *host, + struct srp_target_port *target) +{ + struct srp_target_port *t; + bool ret = false; + + if (target->state == SRP_TARGET_REMOVED) + goto out; + + ret = true; + + spin_lock(&host->target_lock); + list_for_each_entry(t, &host->target_list, list) { + if (t != target && + target->id_ext == t->id_ext && + target->ioc_guid == t->ioc_guid && + target->initiator_ext == t->initiator_ext) { + ret = false; + break; + } + } + spin_unlock(&host->target_lock); + +out: + return ret; +} + /* * Target ports are added by writing * @@ -1794,6 +2707,12 @@ enum { SRP_OPT_MAX_CMD_PER_LUN = 1 << 6, SRP_OPT_IO_CLASS = 1 << 7, SRP_OPT_INITIATOR_EXT = 1 << 8, + SRP_OPT_CMD_SG_ENTRIES = 1 << 9, + SRP_OPT_ALLOW_EXT_SG = 1 << 10, + SRP_OPT_SG_TABLESIZE = 1 << 11, + SRP_OPT_COMP_VECTOR = 1 << 12, + SRP_OPT_TL_RETRY_COUNT = 1 << 13, + SRP_OPT_QUEUE_SIZE = 1 << 14, SRP_OPT_ALL = (SRP_OPT_ID_EXT | SRP_OPT_IOC_GUID | SRP_OPT_DGID | @@ -1811,6 +2730,12 @@ static const match_table_t srp_opt_tokens = { { SRP_OPT_MAX_CMD_PER_LUN, "max_cmd_per_lun=%d" }, { SRP_OPT_IO_CLASS, "io_class=%x" }, { SRP_OPT_INITIATOR_EXT, "initiator_ext=%s" }, + { SRP_OPT_CMD_SG_ENTRIES, "cmd_sg_entries=%u" }, + { SRP_OPT_ALLOW_EXT_SG, "allow_ext_sg=%u" }, + { SRP_OPT_SG_TABLESIZE, "sg_tablesize=%u" }, + { SRP_OPT_COMP_VECTOR, "comp_vector=%u" }, + { SRP_OPT_TL_RETRY_COUNT, "tl_retry_count=%u" }, + { SRP_OPT_QUEUE_SIZE, "queue_size=%d" }, { SRP_OPT_ERR, NULL } }; @@ -1865,7 +2790,7 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target) goto out; } if (strlen(p) != 32) { - printk(KERN_WARNING PFX "bad dest GID parameter '%s'\n", p); + pr_warn("bad dest GID parameter '%s'\n", p); kfree(p); goto out; } @@ -1880,7 +2805,7 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target) case SRP_OPT_PKEY: if (match_hex(args, &token)) { - printk(KERN_WARNING PFX "bad P_Key parameter '%s'\n", p); + pr_warn("bad P_Key parameter '%s'\n", p); goto out; } target->path.pkey = cpu_to_be16(token); @@ -1899,30 +2824,43 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target) case SRP_OPT_MAX_SECT: if (match_int(args, &token)) { - printk(KERN_WARNING PFX "bad max sect parameter '%s'\n", p); + pr_warn("bad max sect parameter '%s'\n", p); goto out; } target->scsi_host->max_sectors = token; break; + case SRP_OPT_QUEUE_SIZE: + if (match_int(args, &token) || token < 1) { + pr_warn("bad queue_size parameter '%s'\n", p); + goto out; + } + target->scsi_host->can_queue = token; + target->queue_size = token + SRP_RSP_SQ_SIZE + + SRP_TSK_MGMT_SQ_SIZE; + if (!(opt_mask & SRP_OPT_MAX_CMD_PER_LUN)) + target->scsi_host->cmd_per_lun = token; + break; + case SRP_OPT_MAX_CMD_PER_LUN: - if (match_int(args, &token)) { - printk(KERN_WARNING PFX "bad max cmd_per_lun parameter '%s'\n", p); + if (match_int(args, &token) || token < 1) { + pr_warn("bad max cmd_per_lun parameter '%s'\n", + p); goto out; } - target->scsi_host->cmd_per_lun = min(token, SRP_CMD_SQ_SIZE); + target->scsi_host->cmd_per_lun = token; break; case SRP_OPT_IO_CLASS: if (match_hex(args, &token)) { - printk(KERN_WARNING PFX "bad IO class parameter '%s' \n", p); + pr_warn("bad IO class parameter '%s'\n", p); goto out; } if (token != SRP_REV10_IB_IO_CLASS && token != SRP_REV16A_IB_IO_CLASS) { - printk(KERN_WARNING PFX "unknown IO class parameter value" - " %x specified (use %x or %x).\n", - token, SRP_REV10_IB_IO_CLASS, SRP_REV16A_IB_IO_CLASS); + pr_warn("unknown IO class parameter value %x specified (use %x or %x).\n", + token, SRP_REV10_IB_IO_CLASS, + SRP_REV16A_IB_IO_CLASS); goto out; } target->io_class = token; @@ -1938,9 +2876,53 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target) kfree(p); break; + case SRP_OPT_CMD_SG_ENTRIES: + if (match_int(args, &token) || token < 1 || token > 255) { + pr_warn("bad max cmd_sg_entries parameter '%s'\n", + p); + goto out; + } + target->cmd_sg_cnt = token; + break; + + case SRP_OPT_ALLOW_EXT_SG: + if (match_int(args, &token)) { + pr_warn("bad allow_ext_sg parameter '%s'\n", p); + goto out; + } + target->allow_ext_sg = !!token; + break; + + case SRP_OPT_SG_TABLESIZE: + if (match_int(args, &token) || token < 1 || + token > SCSI_MAX_SG_CHAIN_SEGMENTS) { + pr_warn("bad max sg_tablesize parameter '%s'\n", + p); + goto out; + } + target->sg_tablesize = token; + break; + + case SRP_OPT_COMP_VECTOR: + if (match_int(args, &token) || token < 0) { + pr_warn("bad comp_vector parameter '%s'\n", p); + goto out; + } + target->comp_vector = token; + break; + + case SRP_OPT_TL_RETRY_COUNT: + if (match_int(args, &token) || token < 2 || token > 7) { + pr_warn("bad tl_retry_count parameter '%s' (must be a number between 2 and 7)\n", + p); + goto out; + } + target->tl_retry_count = token; + break; + default: - printk(KERN_WARNING PFX "unknown parameter or missing value " - "'%s' in target creation request\n", p); + pr_warn("unknown parameter or missing value '%s' in target creation request\n", + p); goto out; } } @@ -1951,9 +2933,14 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target) for (i = 0; i < ARRAY_SIZE(srp_opt_tokens); ++i) if ((srp_opt_tokens[i].token & SRP_OPT_ALL) && !(srp_opt_tokens[i].token & opt_mask)) - printk(KERN_WARNING PFX "target creation request is " - "missing parameter '%s'\n", - srp_opt_tokens[i].pattern); + pr_warn("target creation request is missing parameter '%s'\n", + srp_opt_tokens[i].pattern); + + if (target->scsi_host->cmd_per_lun > target->scsi_host->can_queue + && (opt_mask & SRP_OPT_MAX_CMD_PER_LUN)) + pr_warn("cmd_per_lun = %d > queue_size = %d\n", + target->scsi_host->cmd_per_lun, + target->scsi_host->can_queue); out: kfree(options); @@ -1968,55 +2955,85 @@ static ssize_t srp_create_target(struct device *dev, container_of(dev, struct srp_host, dev); struct Scsi_Host *target_host; struct srp_target_port *target; + struct srp_device *srp_dev = host->srp_dev; + struct ib_device *ibdev = srp_dev->dev; int ret; - int i; target_host = scsi_host_alloc(&srp_template, sizeof (struct srp_target_port)); if (!target_host) return -ENOMEM; - target_host->transportt = ib_srp_transport_template; + target_host->transportt = ib_srp_transport_template; + target_host->max_channel = 0; + target_host->max_id = 1; target_host->max_lun = SRP_MAX_LUN; target_host->max_cmd_len = sizeof ((struct srp_cmd *) (void *) 0L)->cdb; target = host_to_target(target_host); - target->io_class = SRP_REV16A_IB_IO_CLASS; - target->scsi_host = target_host; - target->srp_host = host; + target->io_class = SRP_REV16A_IB_IO_CLASS; + target->scsi_host = target_host; + target->srp_host = host; + target->lkey = host->srp_dev->mr->lkey; + target->rkey = host->srp_dev->mr->rkey; + target->cmd_sg_cnt = cmd_sg_entries; + target->sg_tablesize = indirect_sg_entries ? : cmd_sg_entries; + target->allow_ext_sg = allow_ext_sg; + target->tl_retry_count = 7; + target->queue_size = SRP_DEFAULT_QUEUE_SIZE; - INIT_LIST_HEAD(&target->free_reqs); - INIT_LIST_HEAD(&target->req_queue); - for (i = 0; i < SRP_CMD_SQ_SIZE; ++i) { - target->req_ring[i].index = i; - list_add_tail(&target->req_ring[i].list, &target->free_reqs); - } + mutex_lock(&host->add_target_mutex); ret = srp_parse_options(buf, target); if (ret) goto err; - ib_query_gid(host->srp_dev->dev, host->port, 0, &target->path.sgid); + target->req_ring_size = target->queue_size - SRP_TSK_MGMT_SQ_SIZE; - shost_printk(KERN_DEBUG, target->scsi_host, PFX - "new target: id_ext %016llx ioc_guid %016llx pkey %04x " - "service_id %016llx dgid %pI6\n", - (unsigned long long) be64_to_cpu(target->id_ext), - (unsigned long long) be64_to_cpu(target->ioc_guid), - be16_to_cpu(target->path.pkey), - (unsigned long long) be64_to_cpu(target->service_id), - target->path.dgid.raw); + if (!srp_conn_unique(target->srp_host, target)) { + shost_printk(KERN_INFO, target->scsi_host, + PFX "Already connected to target port with id_ext=%016llx;ioc_guid=%016llx;initiator_ext=%016llx\n", + be64_to_cpu(target->id_ext), + be64_to_cpu(target->ioc_guid), + be64_to_cpu(target->initiator_ext)); + ret = -EEXIST; + goto err; + } + + if (!srp_dev->has_fmr && !srp_dev->has_fr && !target->allow_ext_sg && + target->cmd_sg_cnt < target->sg_tablesize) { + pr_warn("No MR pool and no external indirect descriptors, limiting sg_tablesize to cmd_sg_cnt\n"); + target->sg_tablesize = target->cmd_sg_cnt; + } + + target_host->sg_tablesize = target->sg_tablesize; + target->indirect_size = target->sg_tablesize * + sizeof (struct srp_direct_buf); + target->max_iu_len = sizeof (struct srp_cmd) + + sizeof (struct srp_indirect_buf) + + target->cmd_sg_cnt * sizeof (struct srp_direct_buf); + + INIT_WORK(&target->tl_err_work, srp_tl_err_work); + INIT_WORK(&target->remove_work, srp_remove_work); + spin_lock_init(&target->lock); + INIT_LIST_HEAD(&target->free_tx); + ret = srp_alloc_req_data(target); + if (ret) + goto err_free_mem; + + ret = ib_query_gid(ibdev, host->port, 0, &target->path.sgid); + if (ret) + goto err_free_mem; ret = srp_create_target_ib(target); if (ret) - goto err; + goto err_free_mem; ret = srp_new_cm_id(target); if (ret) - goto err_free; + goto err_free_ib; - target->qp_in_error = 0; ret = srp_connect_target(target); if (ret) { shost_printk(KERN_ERR, target->scsi_host, @@ -2028,7 +3045,19 @@ static ssize_t srp_create_target(struct device *dev, if (ret) goto err_disconnect; - return count; + shost_printk(KERN_DEBUG, target->scsi_host, PFX + "new target: id_ext %016llx ioc_guid %016llx pkey %04x service_id %016llx sgid %pI6 dgid %pI6\n", + be64_to_cpu(target->id_ext), + be64_to_cpu(target->ioc_guid), + be16_to_cpu(target->path.pkey), + be64_to_cpu(target->service_id), + target->path.sgid.raw, target->path.dgid.raw); + + ret = count; + +out: + mutex_unlock(&host->add_target_mutex); + return ret; err_disconnect: srp_disconnect_target(target); @@ -2036,13 +3065,15 @@ err_disconnect: err_cm_id: ib_destroy_cm_id(target->cm_id); -err_free: +err_free_ib: srp_free_target_ib(target); +err_free_mem: + srp_free_req_data(target); + err: scsi_host_put(target_host); - - return ret; + goto out; } static DEVICE_ATTR(add_target, S_IWUSR, NULL, srp_create_target); @@ -2078,6 +3109,7 @@ static struct srp_host *srp_add_port(struct srp_device *device, u8 port) INIT_LIST_HEAD(&host->target_list); spin_lock_init(&host->target_lock); init_completion(&host->released); + mutex_init(&host->add_target_mutex); host->srp_dev = device; host->port = port; @@ -2109,17 +3141,16 @@ static void srp_add_one(struct ib_device *device) { struct srp_device *srp_dev; struct ib_device_attr *dev_attr; - struct ib_fmr_pool_param fmr_param; struct srp_host *host; - int s, e, p; + int mr_page_shift, s, e, p; + u64 max_pages_per_mr; dev_attr = kmalloc(sizeof *dev_attr, GFP_KERNEL); if (!dev_attr) return; if (ib_query_device(device, dev_attr)) { - printk(KERN_WARNING PFX "Query device failed for %s\n", - device->name); + pr_warn("Query device failed for %s\n", device->name); goto free_attr; } @@ -2127,14 +3158,39 @@ static void srp_add_one(struct ib_device *device) if (!srp_dev) goto free_attr; + srp_dev->has_fmr = (device->alloc_fmr && device->dealloc_fmr && + device->map_phys_fmr && device->unmap_fmr); + srp_dev->has_fr = (dev_attr->device_cap_flags & + IB_DEVICE_MEM_MGT_EXTENSIONS); + if (!srp_dev->has_fmr && !srp_dev->has_fr) + dev_warn(&device->dev, "neither FMR nor FR is supported\n"); + + srp_dev->use_fast_reg = (srp_dev->has_fr && + (!srp_dev->has_fmr || prefer_fr)); + /* * Use the smallest page size supported by the HCA, down to a - * minimum of 512 bytes (which is the smallest sector that a - * SCSI command will ever carry). + * minimum of 4096 bytes. We're unlikely to build large sglists + * out of smaller entries. */ - srp_dev->fmr_page_shift = max(9, ffs(dev_attr->page_size_cap) - 1); - srp_dev->fmr_page_size = 1 << srp_dev->fmr_page_shift; - srp_dev->fmr_page_mask = ~((u64) srp_dev->fmr_page_size - 1); + mr_page_shift = max(12, ffs(dev_attr->page_size_cap) - 1); + srp_dev->mr_page_size = 1 << mr_page_shift; + srp_dev->mr_page_mask = ~((u64) srp_dev->mr_page_size - 1); + max_pages_per_mr = dev_attr->max_mr_size; + do_div(max_pages_per_mr, srp_dev->mr_page_size); + srp_dev->max_pages_per_mr = min_t(u64, SRP_MAX_PAGES_PER_MR, + max_pages_per_mr); + if (srp_dev->use_fast_reg) { + srp_dev->max_pages_per_mr = + min_t(u32, srp_dev->max_pages_per_mr, + dev_attr->max_fast_reg_page_list_len); + } + srp_dev->mr_max_size = srp_dev->mr_page_size * + srp_dev->max_pages_per_mr; + pr_debug("%s: mr_page_shift = %d, dev_attr->max_mr_size = %#llx, dev_attr->max_fast_reg_page_list_len = %u, max_pages_per_mr = %d, mr_max_size = %#x\n", + device->name, mr_page_shift, dev_attr->max_mr_size, + dev_attr->max_fast_reg_page_list_len, + srp_dev->max_pages_per_mr, srp_dev->mr_max_size); INIT_LIST_HEAD(&srp_dev->dev_list); @@ -2150,20 +3206,6 @@ static void srp_add_one(struct ib_device *device) if (IS_ERR(srp_dev->mr)) goto err_pd; - memset(&fmr_param, 0, sizeof fmr_param); - fmr_param.pool_size = SRP_FMR_POOL_SIZE; - fmr_param.dirty_watermark = SRP_FMR_DIRTY_SIZE; - fmr_param.cache = 1; - fmr_param.max_pages_per_fmr = SRP_FMR_SIZE; - fmr_param.page_shift = srp_dev->fmr_page_shift; - fmr_param.access = (IB_ACCESS_LOCAL_WRITE | - IB_ACCESS_REMOTE_WRITE | - IB_ACCESS_REMOTE_READ); - - srp_dev->fmr_pool = ib_create_fmr_pool(srp_dev->pd, &fmr_param); - if (IS_ERR(srp_dev->fmr_pool)) - srp_dev->fmr_pool = NULL; - if (device->node_type == RDMA_NODE_IB_SWITCH) { s = 0; e = 0; @@ -2196,10 +3238,11 @@ static void srp_remove_one(struct ib_device *device) { struct srp_device *srp_dev; struct srp_host *host, *tmp_host; - LIST_HEAD(target_list); - struct srp_target_port *target, *tmp_target; + struct srp_target_port *target; srp_dev = ib_get_client_data(device, &srp_client); + if (!srp_dev) + return; list_for_each_entry_safe(host, tmp_host, &srp_dev->dev_list, list) { device_unregister(&host->dev); @@ -2210,39 +3253,21 @@ static void srp_remove_one(struct ib_device *device) wait_for_completion(&host->released); /* - * Mark all target ports as removed, so we stop queueing - * commands and don't try to reconnect. + * Remove all target ports. */ spin_lock(&host->target_lock); - list_for_each_entry(target, &host->target_list, list) { - spin_lock_irq(target->scsi_host->host_lock); - target->state = SRP_TARGET_REMOVED; - spin_unlock_irq(target->scsi_host->host_lock); - } + list_for_each_entry(target, &host->target_list, list) + srp_queue_remove_work(target); spin_unlock(&host->target_lock); /* - * Wait for any reconnection tasks that may have - * started before we marked our target ports as - * removed, and any target port removal tasks. + * Wait for target port removal tasks. */ - flush_scheduled_work(); - - list_for_each_entry_safe(target, tmp_target, - &host->target_list, list) { - srp_remove_host(target->scsi_host); - scsi_remove_host(target->scsi_host); - srp_disconnect_target(target); - ib_destroy_cm_id(target->cm_id); - srp_free_target_ib(target); - scsi_host_put(target->scsi_host); - } + flush_workqueue(system_long_wq); kfree(host); } - if (srp_dev->fmr_pool) - ib_destroy_fmr_pool(srp_dev->fmr_pool); ib_dereg_mr(srp_dev->mr); ib_dealloc_pd(srp_dev->pd); @@ -2250,18 +3275,42 @@ static void srp_remove_one(struct ib_device *device) } static struct srp_function_template ib_srp_transport_functions = { + .has_rport_state = true, + .reset_timer_if_blocked = true, + .reconnect_delay = &srp_reconnect_delay, + .fast_io_fail_tmo = &srp_fast_io_fail_tmo, + .dev_loss_tmo = &srp_dev_loss_tmo, + .reconnect = srp_rport_reconnect, + .rport_delete = srp_rport_delete, + .terminate_rport_io = srp_terminate_io, }; static int __init srp_init_module(void) { int ret; - BUILD_BUG_ON_NOT_POWER_OF_2(SRP_SQ_SIZE); - BUILD_BUG_ON_NOT_POWER_OF_2(SRP_RQ_SIZE); + BUILD_BUG_ON(FIELD_SIZEOF(struct ib_wc, wr_id) < sizeof(void *)); - if (srp_sg_tablesize > 255) { - printk(KERN_WARNING PFX "Clamping srp_sg_tablesize to 255\n"); - srp_sg_tablesize = 255; + if (srp_sg_tablesize) { + pr_warn("srp_sg_tablesize is deprecated, please use cmd_sg_entries\n"); + if (!cmd_sg_entries) + cmd_sg_entries = srp_sg_tablesize; + } + + if (!cmd_sg_entries) + cmd_sg_entries = SRP_DEF_SG_TABLESIZE; + + if (cmd_sg_entries > 255) { + pr_warn("Clamping cmd_sg_entries to 255\n"); + cmd_sg_entries = 255; + } + + if (!indirect_sg_entries) + indirect_sg_entries = cmd_sg_entries; + else if (indirect_sg_entries < cmd_sg_entries) { + pr_warn("Bumping up indirect_sg_entries to match cmd_sg_entries (%u)\n", + cmd_sg_entries); + indirect_sg_entries = cmd_sg_entries; } ib_srp_transport_template = @@ -2269,14 +3318,9 @@ static int __init srp_init_module(void) if (!ib_srp_transport_template) return -ENOMEM; - srp_template.sg_tablesize = srp_sg_tablesize; - srp_max_iu_len = (sizeof (struct srp_cmd) + - sizeof (struct srp_indirect_buf) + - srp_sg_tablesize * 16); - ret = class_register(&srp_class); if (ret) { - printk(KERN_ERR PFX "couldn't register class infiniband_srp\n"); + pr_err("couldn't register class infiniband_srp\n"); srp_release_transport(ib_srp_transport_template); return ret; } @@ -2285,7 +3329,7 @@ static int __init srp_init_module(void) ret = ib_register_client(&srp_client); if (ret) { - printk(KERN_ERR PFX "couldn't register IB client\n"); + pr_err("couldn't register IB client\n"); srp_release_transport(ib_srp_transport_template); ib_sa_unregister_client(&srp_sa_client); class_unregister(&srp_class); diff --git a/drivers/infiniband/ulp/srp/ib_srp.h b/drivers/infiniband/ulp/srp/ib_srp.h index ed0dce9e479..e46ecb15aa0 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.h +++ b/drivers/infiniband/ulp/srp/ib_srp.h @@ -57,29 +57,24 @@ enum { SRP_MAX_LUN = 512, SRP_DEF_SG_TABLESIZE = 12, - SRP_RQ_SHIFT = 6, - SRP_RQ_SIZE = 1 << SRP_RQ_SHIFT, - SRP_RQ_MASK = SRP_RQ_SIZE - 1, - - SRP_SQ_SIZE = SRP_RQ_SIZE, - SRP_SQ_MASK = SRP_SQ_SIZE - 1, + SRP_DEFAULT_QUEUE_SIZE = 1 << 6, SRP_RSP_SQ_SIZE = 1, - SRP_REQ_SQ_SIZE = SRP_SQ_SIZE - SRP_RSP_SQ_SIZE, SRP_TSK_MGMT_SQ_SIZE = 1, - SRP_CMD_SQ_SIZE = SRP_REQ_SQ_SIZE - SRP_TSK_MGMT_SQ_SIZE, + SRP_DEFAULT_CMD_SQ_SIZE = SRP_DEFAULT_QUEUE_SIZE - SRP_RSP_SQ_SIZE - + SRP_TSK_MGMT_SQ_SIZE, + + SRP_TAG_NO_REQ = ~0U, + SRP_TAG_TSK_MGMT = 1U << 31, - SRP_TAG_TSK_MGMT = 1 << (SRP_RQ_SHIFT + 1), + SRP_MAX_PAGES_PER_MR = 512, - SRP_FMR_SIZE = 256, - SRP_FMR_POOL_SIZE = 1024, - SRP_FMR_DIRTY_SIZE = SRP_FMR_POOL_SIZE / 4 + LOCAL_INV_WR_ID_MASK = 1, + FAST_REG_WR_ID_MASK = 2, }; enum srp_target_state { SRP_TARGET_LIVE, - SRP_TARGET_CONNECTING, - SRP_TARGET_DEAD, - SRP_TARGET_REMOVED + SRP_TARGET_REMOVED, }; enum srp_iu_type { @@ -88,15 +83,24 @@ enum srp_iu_type { SRP_IU_RSP, }; +/* + * @mr_page_mask: HCA memory registration page mask. + * @mr_page_size: HCA memory registration page size. + * @mr_max_size: Maximum size in bytes of a single FMR / FR registration + * request. + */ struct srp_device { struct list_head dev_list; struct ib_device *dev; struct ib_pd *pd; struct ib_mr *mr; - struct ib_fmr_pool *fmr_pool; - int fmr_page_shift; - int fmr_page_size; - u64 fmr_page_mask; + u64 mr_page_mask; + int mr_page_size; + int mr_max_size; + int max_pages_per_mr; + bool has_fmr; + bool has_fr; + bool use_fast_reg; }; struct srp_host { @@ -107,21 +111,51 @@ struct srp_host { spinlock_t target_lock; struct completion released; struct list_head list; + struct mutex add_target_mutex; }; struct srp_request { struct list_head list; struct scsi_cmnd *scmnd; struct srp_iu *cmd; - struct srp_iu *tsk_mgmt; - struct ib_pool_fmr *fmr; - struct completion done; + union { + struct ib_pool_fmr **fmr_list; + struct srp_fr_desc **fr_list; + }; + u64 *map_page; + struct srp_direct_buf *indirect_desc; + dma_addr_t indirect_dma_addr; + short nmdesc; short index; - u8 cmd_done; - u8 tsk_status; }; struct srp_target_port { + /* These are RW in the hot path, and commonly used together */ + struct list_head free_tx; + struct list_head free_reqs; + spinlock_t lock; + s32 req_lim; + + /* These are read-only in the hot path */ + struct ib_cq *send_cq ____cacheline_aligned_in_smp; + struct ib_cq *recv_cq; + struct ib_qp *qp; + union { + struct ib_fmr_pool *fmr_pool; + struct srp_fr_pool *fr_pool; + }; + u32 lkey; + u32 rkey; + enum srp_target_state state; + unsigned int max_iu_len; + unsigned int cmd_sg_cnt; + unsigned int indirect_size; + bool allow_ext_sg; + + /* Everything above this point is used in the hot path of + * command processing. Try to keep them packed into cachelines. + */ + __be64 id_ext; __be64 ioc_guid; __be64 service_id; @@ -129,50 +163,117 @@ struct srp_target_port { u16 io_class; struct srp_host *srp_host; struct Scsi_Host *scsi_host; + struct srp_rport *rport; char target_name[32]; unsigned int scsi_id; + unsigned int sg_tablesize; + int queue_size; + int req_ring_size; + int comp_vector; + int tl_retry_count; struct ib_sa_path_rec path; __be16 orig_dgid[8]; struct ib_sa_query *path_query; int path_query_id; + u32 rq_tmo_jiffies; + bool connected; + struct ib_cm_id *cm_id; - struct ib_cq *recv_cq; - struct ib_cq *send_cq; - struct ib_qp *qp; int max_ti_iu_len; - s32 req_lim; int zero_req_lim; - unsigned rx_head; - struct srp_iu *rx_ring[SRP_RQ_SIZE]; + struct srp_iu **tx_ring; + struct srp_iu **rx_ring; + struct srp_request *req_ring; - unsigned tx_head; - unsigned tx_tail; - struct srp_iu *tx_ring[SRP_SQ_SIZE]; - - struct list_head free_reqs; - struct list_head req_queue; - struct srp_request req_ring[SRP_CMD_SQ_SIZE]; - - struct work_struct work; + struct work_struct tl_err_work; + struct work_struct remove_work; struct list_head list; struct completion done; int status; - enum srp_target_state state; - int qp_in_error; + bool qp_in_error; + + struct completion tsk_mgmt_done; + u8 tsk_mgmt_status; }; struct srp_iu { + struct list_head list; u64 dma; void *buf; size_t size; enum dma_data_direction direction; - enum srp_iu_type type; +}; + +/** + * struct srp_fr_desc - fast registration work request arguments + * @entry: Entry in srp_fr_pool.free_list. + * @mr: Memory region. + * @frpl: Fast registration page list. + */ +struct srp_fr_desc { + struct list_head entry; + struct ib_mr *mr; + struct ib_fast_reg_page_list *frpl; +}; + +/** + * struct srp_fr_pool - pool of fast registration descriptors + * + * An entry is available for allocation if and only if it occurs in @free_list. + * + * @size: Number of descriptors in this pool. + * @max_page_list_len: Maximum fast registration work request page list length. + * @lock: Protects free_list. + * @free_list: List of free descriptors. + * @desc: Fast registration descriptor pool. + */ +struct srp_fr_pool { + int size; + int max_page_list_len; + spinlock_t lock; + struct list_head free_list; + struct srp_fr_desc desc[0]; +}; + +/** + * struct srp_map_state - per-request DMA memory mapping state + * @desc: Pointer to the element of the SRP buffer descriptor array + * that is being filled in. + * @pages: Array with DMA addresses of pages being considered for + * memory registration. + * @base_dma_addr: DMA address of the first page that has not yet been mapped. + * @dma_len: Number of bytes that will be registered with the next + * FMR or FR memory registration call. + * @total_len: Total number of bytes in the sg-list being mapped. + * @npages: Number of page addresses in the pages[] array. + * @nmdesc: Number of FMR or FR memory descriptors used for mapping. + * @ndesc: Number of SRP buffer descriptors that have been filled in. + * @unmapped_sg: First element of the sg-list that is mapped via FMR or FR. + * @unmapped_index: Index of the first element mapped via FMR or FR. + * @unmapped_addr: DMA address of the first element mapped via FMR or FR. + */ +struct srp_map_state { + union { + struct ib_pool_fmr **next_fmr; + struct srp_fr_desc **next_fr; + }; + struct srp_direct_buf *desc; + u64 *pages; + dma_addr_t base_dma_addr; + u32 dma_len; + u32 total_len; + unsigned int npages; + unsigned int nmdesc; + unsigned int ndesc; + struct scatterlist *unmapped_sg; + int unmapped_index; + dma_addr_t unmapped_addr; }; #endif /* IB_SRP_H */ |