diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c index ecc1048de837..4976db56e351 100644 --- a/drivers/nvme/host/fc.c +++ b/drivers/nvme/host/fc.c @@ -19,6 +19,7 @@ #include #include #include +#include #include "nvme.h" #include "fabrics.h" @@ -44,6 +45,8 @@ enum nvme_fc_queue_flags { #define NVMEFC_QUEUE_DELAY 3 /* ms units */ +#define NVME_FC_MAX_CONNECT_ATTEMPTS 1 + struct nvme_fc_queue { struct nvme_fc_ctrl *ctrl; struct device *dev; @@ -65,6 +68,7 @@ enum nvme_fcop_flags { FCOP_FLAGS_TERMIO = (1 << 0), FCOP_FLAGS_RELEASED = (1 << 1), FCOP_FLAGS_COMPLETE = (1 << 2), + FCOP_FLAGS_AEN = (1 << 3), }; struct nvmefc_ls_req_op { @@ -86,6 +90,7 @@ enum nvme_fcpop_state { FCPOP_STATE_IDLE = 1, FCPOP_STATE_ACTIVE = 2, FCPOP_STATE_ABORTED = 3, + FCPOP_STATE_COMPLETE = 4, }; struct nvme_fc_fcp_op { @@ -104,6 +109,7 @@ struct nvme_fc_fcp_op { struct request *rq; atomic_t state; + u32 flags; u32 rqno; u32 nents; @@ -134,19 +140,17 @@ struct nvme_fc_rport { struct kref ref; } __aligned(sizeof(u64)); /* alignment for other things alloc'd with */ -enum nvme_fcctrl_state { - FCCTRL_INIT = 0, - FCCTRL_ACTIVE = 1, +enum nvme_fcctrl_flags { + FCCTRL_TERMIO = (1 << 0), }; struct nvme_fc_ctrl { spinlock_t lock; struct nvme_fc_queue *queues; - u32 queue_count; - struct device *dev; struct nvme_fc_lport *lport; struct nvme_fc_rport *rport; + u32 queue_count; u32 cnum; u64 association_id; @@ -159,8 +163,14 @@ struct nvme_fc_ctrl { struct blk_mq_tag_set tag_set; struct work_struct delete_work; + struct work_struct reset_work; + struct delayed_work connect_work; + int reconnect_delay; + int connect_attempts; + struct kref ref; - int state; + u32 flags; + u32 iocnt; struct nvme_fc_fcp_op aen_ops[NVME_FC_NR_AEN_COMMANDS]; @@ -1132,6 +1142,7 @@ nvme_fc_xmt_disconnect_assoc(struct nvme_fc_ctrl *ctrl) /* *********************** NVME Ctrl Routines **************************** */ +static void __nvme_fc_final_op_cleanup(struct request *rq); static int nvme_fc_reinit_request(void *data, struct request *rq) @@ -1169,21 +1180,84 @@ nvme_fc_exit_request(void *data, struct request *rq, return __nvme_fc_exit_request(data, op); } +static int +__nvme_fc_abort_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_fcp_op *op) +{ + int state; + + state = atomic_xchg(&op->state, FCPOP_STATE_ABORTED); + if (state != FCPOP_STATE_ACTIVE) { + atomic_set(&op->state, state); + return -ECANCELED; + } + + ctrl->lport->ops->fcp_abort(&ctrl->lport->localport, + &ctrl->rport->remoteport, + op->queue->lldd_handle, + &op->fcp_req); + + return 0; +} + static void -nvme_fc_exit_aen_ops(struct nvme_fc_ctrl *ctrl) +nvme_fc_abort_aen_ops(struct nvme_fc_ctrl *ctrl) { struct nvme_fc_fcp_op *aen_op = ctrl->aen_ops; - int i; + unsigned long flags; + int i, ret; for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) { - if (atomic_read(&aen_op->state) == FCPOP_STATE_UNINIT) + if (atomic_read(&aen_op->state) != FCPOP_STATE_ACTIVE) continue; - __nvme_fc_exit_request(ctrl, aen_op); - nvme_fc_ctrl_put(ctrl); + + spin_lock_irqsave(&ctrl->lock, flags); + if (ctrl->flags & FCCTRL_TERMIO) { + ctrl->iocnt++; + aen_op->flags |= FCOP_FLAGS_TERMIO; + } + spin_unlock_irqrestore(&ctrl->lock, flags); + + ret = __nvme_fc_abort_op(ctrl, aen_op); + if (ret) { + /* + * if __nvme_fc_abort_op failed the io wasn't + * active. Thus this call path is running in + * parallel to the io complete. Treat as non-error. + */ + + /* back out the flags/counters */ + spin_lock_irqsave(&ctrl->lock, flags); + if (ctrl->flags & FCCTRL_TERMIO) + ctrl->iocnt--; + aen_op->flags &= ~FCOP_FLAGS_TERMIO; + spin_unlock_irqrestore(&ctrl->lock, flags); + return; + } } } -void +static inline int +__nvme_fc_fcpop_chk_teardowns(struct nvme_fc_ctrl *ctrl, + struct nvme_fc_fcp_op *op) +{ + unsigned long flags; + bool complete_rq = false; + + spin_lock_irqsave(&ctrl->lock, flags); + if (unlikely(op->flags & FCOP_FLAGS_TERMIO)) { + if (ctrl->flags & FCCTRL_TERMIO) + ctrl->iocnt--; + } + if (op->flags & FCOP_FLAGS_RELEASED) + complete_rq = true; + else + op->flags |= FCOP_FLAGS_COMPLETE; + spin_unlock_irqrestore(&ctrl->lock, flags); + + return complete_rq; +} + +static void nvme_fc_fcpio_done(struct nvmefc_fcp_req *req) { struct nvme_fc_fcp_op *op = fcp_req_to_fcp_op(req); @@ -1192,8 +1266,10 @@ nvme_fc_fcpio_done(struct nvmefc_fcp_req *req) struct nvme_fc_ctrl *ctrl = op->ctrl; struct nvme_fc_queue *queue = op->queue; struct nvme_completion *cqe = &op->rsp_iu.cqe; + struct nvme_command *sqe = &op->cmd_iu.sqe; __le16 status = cpu_to_le16(NVME_SC_SUCCESS << 1); union nvme_result result; + bool complete_rq; /* * WARNING: @@ -1274,7 +1350,7 @@ nvme_fc_fcpio_done(struct nvmefc_fcp_req *req) be32_to_cpu(op->rsp_iu.xfrd_len) != freq->transferred_length || op->rsp_iu.status_code || - op->rqno != le16_to_cpu(cqe->command_id))) { + sqe->common.command_id != cqe->command_id)) { status = cpu_to_le16(NVME_SC_FC_TRANSPORT_ERROR << 1); goto done; } @@ -1288,13 +1364,25 @@ nvme_fc_fcpio_done(struct nvmefc_fcp_req *req) } done: - if (!queue->qnum && op->rqno >= AEN_CMDID_BASE) { + if (op->flags & FCOP_FLAGS_AEN) { nvme_complete_async_event(&queue->ctrl->ctrl, status, &result); + complete_rq = __nvme_fc_fcpop_chk_teardowns(ctrl, op); + atomic_set(&op->state, FCPOP_STATE_IDLE); + op->flags = FCOP_FLAGS_AEN; /* clear other flags */ nvme_fc_ctrl_put(ctrl); return; } - nvme_end_request(rq, status, result); + complete_rq = __nvme_fc_fcpop_chk_teardowns(ctrl, op); + if (!complete_rq) { + if (unlikely(op->flags & FCOP_FLAGS_TERMIO)) { + status = cpu_to_le16(NVME_SC_ABORT_REQ); + if (blk_queue_dying(rq->q)) + status |= cpu_to_le16(NVME_SC_DNR); + } + nvme_end_request(rq, status, result); + } else + __nvme_fc_final_op_cleanup(rq); } static int @@ -1375,25 +1463,55 @@ nvme_fc_init_aen_ops(struct nvme_fc_ctrl *ctrl) struct nvme_fc_fcp_op *aen_op; struct nvme_fc_cmd_iu *cmdiu; struct nvme_command *sqe; + void *private; int i, ret; aen_op = ctrl->aen_ops; for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) { + private = kzalloc(ctrl->lport->ops->fcprqst_priv_sz, + GFP_KERNEL); + if (!private) + return -ENOMEM; + cmdiu = &aen_op->cmd_iu; sqe = &cmdiu->sqe; ret = __nvme_fc_init_request(ctrl, &ctrl->queues[0], aen_op, (struct request *)NULL, (AEN_CMDID_BASE + i)); - if (ret) + if (ret) { + kfree(private); return ret; + } + + aen_op->flags = FCOP_FLAGS_AEN; + aen_op->fcp_req.first_sgl = NULL; /* no sg list */ + aen_op->fcp_req.private = private; memset(sqe, 0, sizeof(*sqe)); sqe->common.opcode = nvme_admin_async_event; + /* Note: core layer may overwrite the sqe.command_id value */ sqe->common.command_id = AEN_CMDID_BASE + i; } return 0; } +static void +nvme_fc_term_aen_ops(struct nvme_fc_ctrl *ctrl) +{ + struct nvme_fc_fcp_op *aen_op; + int i; + + aen_op = ctrl->aen_ops; + for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) { + if (!aen_op->fcp_req.private) + continue; + + __nvme_fc_exit_request(ctrl, aen_op); + + kfree(aen_op->fcp_req.private); + aen_op->fcp_req.private = NULL; + } +} static inline void __nvme_fc_init_hctx(struct blk_mq_hw_ctx *hctx, struct nvme_fc_ctrl *ctrl, @@ -1492,15 +1610,6 @@ __nvme_fc_delete_hw_queue(struct nvme_fc_ctrl *ctrl, queue->lldd_handle = NULL; } -static void -nvme_fc_destroy_admin_queue(struct nvme_fc_ctrl *ctrl) -{ - __nvme_fc_delete_hw_queue(ctrl, &ctrl->queues[0], 0); - blk_cleanup_queue(ctrl->ctrl.admin_q); - blk_mq_free_tag_set(&ctrl->admin_tag_set); - nvme_fc_free_queue(&ctrl->queues[0]); -} - static void nvme_fc_free_io_queues(struct nvme_fc_ctrl *ctrl) { @@ -1588,19 +1697,27 @@ nvme_fc_ctrl_free(struct kref *ref) container_of(ref, struct nvme_fc_ctrl, ref); unsigned long flags; - if (ctrl->state != FCCTRL_INIT) { - /* remove from rport list */ - spin_lock_irqsave(&ctrl->rport->lock, flags); - list_del(&ctrl->ctrl_list); - spin_unlock_irqrestore(&ctrl->rport->lock, flags); + if (ctrl->ctrl.tagset) { + blk_cleanup_queue(ctrl->ctrl.connect_q); + blk_mq_free_tag_set(&ctrl->tag_set); } + /* remove from rport list */ + spin_lock_irqsave(&ctrl->rport->lock, flags); + list_del(&ctrl->ctrl_list); + spin_unlock_irqrestore(&ctrl->rport->lock, flags); + + blk_cleanup_queue(ctrl->ctrl.admin_q); + blk_mq_free_tag_set(&ctrl->admin_tag_set); + + kfree(ctrl->queues); + put_device(ctrl->dev); nvme_fc_rport_put(ctrl->rport); - kfree(ctrl->queues); ida_simple_remove(&nvme_fc_ctrl_cnt, ctrl->cnum); - nvmf_free_options(ctrl->ctrl.opts); + if (ctrl->ctrl.opts) + nvmf_free_options(ctrl->ctrl.opts); kfree(ctrl); } @@ -1621,57 +1738,38 @@ nvme_fc_ctrl_get(struct nvme_fc_ctrl *ctrl) * controller. Called after last nvme_put_ctrl() call */ static void -nvme_fc_free_nvme_ctrl(struct nvme_ctrl *nctrl) +nvme_fc_nvme_ctrl_freed(struct nvme_ctrl *nctrl) { struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl); WARN_ON(nctrl != &ctrl->ctrl); - /* - * Tear down the association, which will generate link - * traffic to terminate connections - */ - - if (ctrl->state != FCCTRL_INIT) { - /* send a Disconnect(association) LS to fc-nvme target */ - nvme_fc_xmt_disconnect_assoc(ctrl); - - if (ctrl->ctrl.tagset) { - blk_cleanup_queue(ctrl->ctrl.connect_q); - blk_mq_free_tag_set(&ctrl->tag_set); - nvme_fc_delete_hw_io_queues(ctrl); - nvme_fc_free_io_queues(ctrl); - } - - nvme_fc_exit_aen_ops(ctrl); - - nvme_fc_destroy_admin_queue(ctrl); - } - nvme_fc_ctrl_put(ctrl); } - -static int -__nvme_fc_abort_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_fcp_op *op) +static void +nvme_fc_error_recovery(struct nvme_fc_ctrl *ctrl, char *errmsg) { - int state; + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: transport association error detected: %s\n", + ctrl->cnum, errmsg); + dev_info(ctrl->ctrl.device, + "NVME-FC{%d}: resetting controller\n", ctrl->cnum); - state = atomic_xchg(&op->state, FCPOP_STATE_ABORTED); - if (state != FCPOP_STATE_ACTIVE) { - atomic_set(&op->state, state); - return -ECANCELED; /* fail */ + if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RECONNECTING)) { + dev_err(ctrl->ctrl.device, + "NVME-FC{%d}: error_recovery: Couldn't change state " + "to RECONNECTING\n", ctrl->cnum); + return; } - ctrl->lport->ops->fcp_abort(&ctrl->lport->localport, - &ctrl->rport->remoteport, - op->queue->lldd_handle, - &op->fcp_req); - - return 0; + if (!queue_work(nvme_fc_wq, &ctrl->reset_work)) + dev_err(ctrl->ctrl.device, + "NVME-FC{%d}: error_recovery: Failed to schedule " + "reset work\n", ctrl->cnum); } -enum blk_eh_timer_return +static enum blk_eh_timer_return nvme_fc_timeout(struct request *rq, bool reserved) { struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq); @@ -1687,11 +1785,13 @@ nvme_fc_timeout(struct request *rq, bool reserved) return BLK_EH_HANDLED; /* - * TODO: force a controller reset - * when that happens, queues will be torn down and outstanding - * ios will be terminated, and the above abort, on a single io - * will no longer be needed. + * we can't individually ABTS an io without affecting the queue, + * thus killing the queue, adn thus the association. + * So resolve by performing a controller reset, which will stop + * the host/io stack, terminate the association on the link, + * and recreate an association on the link. */ + nvme_fc_error_recovery(ctrl, "io timeout error"); return BLK_EH_HANDLED; } @@ -1785,6 +1885,13 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue, u32 csn; int ret; + /* + * before attempting to send the io, check to see if we believe + * the target device is present + */ + if (ctrl->rport->remoteport.port_state != FC_OBJSTATE_ONLINE) + return BLK_MQ_RQ_QUEUE_ERROR; + if (!nvme_fc_ctrl_get(ctrl)) return BLK_MQ_RQ_QUEUE_ERROR; @@ -1829,14 +1936,9 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue, sqe->rw.dptr.sgl.length = cpu_to_le32(data_len); sqe->rw.dptr.sgl.addr = 0; - /* odd that we set the command_id - should come from nvme-fabrics */ - WARN_ON_ONCE(sqe->common.command_id != cpu_to_le16(op->rqno)); - - if (op->rq) { /* skipped on aens */ + if (!(op->flags & FCOP_FLAGS_AEN)) { ret = nvme_fc_map_data(ctrl, op->rq, op); if (ret < 0) { - dev_err(queue->ctrl->ctrl.device, - "Failed to map data (%d)\n", ret); nvme_cleanup_cmd(op->rq); nvme_fc_ctrl_put(ctrl); return (ret == -ENOMEM || ret == -EAGAIN) ? @@ -1849,7 +1951,7 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue, atomic_set(&op->state, FCPOP_STATE_ACTIVE); - if (op->rq) + if (!(op->flags & FCOP_FLAGS_AEN)) blk_mq_start_request(op->rq); ret = ctrl->lport->ops->fcp_io(&ctrl->lport->localport, @@ -1857,9 +1959,6 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue, queue->lldd_handle, &op->fcp_req); if (ret) { - dev_err(ctrl->dev, - "Send nvme command failed - lldd returned %d.\n", ret); - if (op->rq) { /* normal request */ nvme_fc_unmap_data(ctrl, op->rq, op); nvme_cleanup_cmd(op->rq); @@ -1929,12 +2028,8 @@ nvme_fc_poll(struct blk_mq_hw_ctx *hctx, unsigned int tag) struct nvme_fc_fcp_op *op; req = blk_mq_tag_to_rq(nvme_fc_tagset(queue), tag); - if (!req) { - dev_err(queue->ctrl->ctrl.device, - "tag 0x%x on QNum %#x not found\n", - tag, queue->qnum); + if (!req) return 0; - } op = blk_mq_rq_to_pdu(req); @@ -1951,11 +2046,21 @@ nvme_fc_submit_async_event(struct nvme_ctrl *arg, int aer_idx) { struct nvme_fc_ctrl *ctrl = to_fc_ctrl(arg); struct nvme_fc_fcp_op *aen_op; + unsigned long flags; + bool terminating = false; int ret; if (aer_idx > NVME_FC_NR_AEN_COMMANDS) return; + spin_lock_irqsave(&ctrl->lock, flags); + if (ctrl->flags & FCCTRL_TERMIO) + terminating = true; + spin_unlock_irqrestore(&ctrl->lock, flags); + + if (terminating) + return; + aen_op = &ctrl->aen_ops[aer_idx]; ret = nvme_fc_start_fcp_op(ctrl, aen_op->queue, aen_op, 0, @@ -1966,13 +2071,14 @@ nvme_fc_submit_async_event(struct nvme_ctrl *arg, int aer_idx) } static void -nvme_fc_complete_rq(struct request *rq) +__nvme_fc_final_op_cleanup(struct request *rq) { struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq); struct nvme_fc_ctrl *ctrl = op->ctrl; - int state; - state = atomic_xchg(&op->state, FCPOP_STATE_IDLE); + atomic_set(&op->state, FCPOP_STATE_IDLE); + op->flags &= ~(FCOP_FLAGS_TERMIO | FCOP_FLAGS_RELEASED | + FCOP_FLAGS_COMPLETE); nvme_cleanup_cmd(rq); nvme_fc_unmap_data(ctrl, rq, op); @@ -1981,108 +2087,31 @@ nvme_fc_complete_rq(struct request *rq) } -static const struct blk_mq_ops nvme_fc_mq_ops = { - .queue_rq = nvme_fc_queue_rq, - .complete = nvme_fc_complete_rq, - .init_request = nvme_fc_init_request, - .exit_request = nvme_fc_exit_request, - .reinit_request = nvme_fc_reinit_request, - .init_hctx = nvme_fc_init_hctx, - .poll = nvme_fc_poll, - .timeout = nvme_fc_timeout, -}; - -static const struct blk_mq_ops nvme_fc_admin_mq_ops = { - .queue_rq = nvme_fc_queue_rq, - .complete = nvme_fc_complete_rq, - .init_request = nvme_fc_init_admin_request, - .exit_request = nvme_fc_exit_request, - .reinit_request = nvme_fc_reinit_request, - .init_hctx = nvme_fc_init_admin_hctx, - .timeout = nvme_fc_timeout, -}; - -static int -nvme_fc_configure_admin_queue(struct nvme_fc_ctrl *ctrl) +static void +nvme_fc_complete_rq(struct request *rq) { - u32 segs; - int error; + struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq); + struct nvme_fc_ctrl *ctrl = op->ctrl; + unsigned long flags; + bool completed = false; - nvme_fc_init_queue(ctrl, 0, NVME_FC_AQ_BLKMQ_DEPTH); + /* + * the core layer, on controller resets after calling + * nvme_shutdown_ctrl(), calls complete_rq without our + * calling blk_mq_complete_request(), thus there may still + * be live i/o outstanding with the LLDD. Means transport has + * to track complete calls vs fcpio_done calls to know what + * path to take on completes and dones. + */ + spin_lock_irqsave(&ctrl->lock, flags); + if (op->flags & FCOP_FLAGS_COMPLETE) + completed = true; + else + op->flags |= FCOP_FLAGS_RELEASED; + spin_unlock_irqrestore(&ctrl->lock, flags); - error = nvme_fc_connect_admin_queue(ctrl, &ctrl->queues[0], - NVME_FC_AQ_BLKMQ_DEPTH, - (NVME_FC_AQ_BLKMQ_DEPTH / 4)); - if (error) - return error; - - memset(&ctrl->admin_tag_set, 0, sizeof(ctrl->admin_tag_set)); - ctrl->admin_tag_set.ops = &nvme_fc_admin_mq_ops; - ctrl->admin_tag_set.queue_depth = NVME_FC_AQ_BLKMQ_DEPTH; - ctrl->admin_tag_set.reserved_tags = 2; /* fabric connect + Keep-Alive */ - ctrl->admin_tag_set.numa_node = NUMA_NO_NODE; - ctrl->admin_tag_set.cmd_size = sizeof(struct nvme_fc_fcp_op) + - (SG_CHUNK_SIZE * - sizeof(struct scatterlist)) + - ctrl->lport->ops->fcprqst_priv_sz; - ctrl->admin_tag_set.driver_data = ctrl; - ctrl->admin_tag_set.nr_hw_queues = 1; - ctrl->admin_tag_set.timeout = ADMIN_TIMEOUT; - - error = blk_mq_alloc_tag_set(&ctrl->admin_tag_set); - if (error) - goto out_free_queue; - - ctrl->ctrl.admin_q = blk_mq_init_queue(&ctrl->admin_tag_set); - if (IS_ERR(ctrl->ctrl.admin_q)) { - error = PTR_ERR(ctrl->ctrl.admin_q); - goto out_free_tagset; - } - - error = __nvme_fc_create_hw_queue(ctrl, &ctrl->queues[0], 0, - NVME_FC_AQ_BLKMQ_DEPTH); - if (error) - goto out_cleanup_queue; - - error = nvmf_connect_admin_queue(&ctrl->ctrl); - if (error) - goto out_delete_hw_queue; - - error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap); - if (error) { - dev_err(ctrl->ctrl.device, - "prop_get NVME_REG_CAP failed\n"); - goto out_delete_hw_queue; - } - - ctrl->ctrl.sqsize = - min_t(int, NVME_CAP_MQES(ctrl->cap), ctrl->ctrl.sqsize); - - error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap); - if (error) - goto out_delete_hw_queue; - - segs = min_t(u32, NVME_FC_MAX_SEGMENTS, - ctrl->lport->ops->max_sgl_segments); - ctrl->ctrl.max_hw_sectors = (segs - 1) << (PAGE_SHIFT - 9); - - error = nvme_init_identify(&ctrl->ctrl); - if (error) - goto out_delete_hw_queue; - - nvme_start_keep_alive(&ctrl->ctrl); - - return 0; - -out_delete_hw_queue: - __nvme_fc_delete_hw_queue(ctrl, &ctrl->queues[0], 0); -out_cleanup_queue: - blk_cleanup_queue(ctrl->ctrl.admin_q); -out_free_tagset: - blk_mq_free_tag_set(&ctrl->admin_tag_set); -out_free_queue: - nvme_fc_free_queue(&ctrl->queues[0]); - return error; + if (completed) + __nvme_fc_final_op_cleanup(rq); } /* @@ -2104,150 +2133,47 @@ nvme_fc_terminate_exchange(struct request *req, void *data, bool reserved) struct nvme_ctrl *nctrl = data; struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl); struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(req); -int status; + unsigned long flags; + int status; if (!blk_mq_request_started(req)) return; - /* this performs an ABTS-LS on the FC exchange for the io */ - status = __nvme_fc_abort_op(ctrl, op); - /* - * if __nvme_fc_abort_op failed: io wasn't active to abort - * consider it done. Assume completion path already completing - * in parallel - */ - if (status) - /* io wasn't active to abort consider it done */ - /* assume completion path already completing in parallel */ - return; -} - - -/* - * This routine stops operation of the controller. Admin and IO queues - * are stopped, outstanding ios on them terminated, and the nvme ctrl - * is shutdown. - */ -static void -nvme_fc_shutdown_ctrl(struct nvme_fc_ctrl *ctrl) -{ - /* - * If io queues are present, stop them and terminate all outstanding - * ios on them. As FC allocates FC exchange for each io, the - * transport must contact the LLDD to terminate the exchange, - * thus releasing the FC exchange. We use blk_mq_tagset_busy_itr() - * to tell us what io's are busy and invoke a transport routine - * to kill them with the LLDD. After terminating the exchange - * the LLDD will call the transport's normal io done path, but it - * will have an aborted status. The done path will return the - * io requests back to the block layer as part of normal completions - * (but with error status). - */ - if (ctrl->queue_count > 1) { - nvme_stop_queues(&ctrl->ctrl); - blk_mq_tagset_busy_iter(&ctrl->tag_set, - nvme_fc_terminate_exchange, &ctrl->ctrl); + spin_lock_irqsave(&ctrl->lock, flags); + if (ctrl->flags & FCCTRL_TERMIO) { + ctrl->iocnt++; + op->flags |= FCOP_FLAGS_TERMIO; } + spin_unlock_irqrestore(&ctrl->lock, flags); - if (ctrl->ctrl.state == NVME_CTRL_LIVE) - nvme_shutdown_ctrl(&ctrl->ctrl); + status = __nvme_fc_abort_op(ctrl, op); + if (status) { + /* + * if __nvme_fc_abort_op failed the io wasn't + * active. Thus this call path is running in + * parallel to the io complete. Treat as non-error. + */ - /* - * now clean up the admin queue. Same thing as above. - * use blk_mq_tagset_busy_itr() and the transport routine to - * terminate the exchanges. - */ - blk_mq_stop_hw_queues(ctrl->ctrl.admin_q); - blk_mq_tagset_busy_iter(&ctrl->admin_tag_set, - nvme_fc_terminate_exchange, &ctrl->ctrl); + /* back out the flags/counters */ + spin_lock_irqsave(&ctrl->lock, flags); + if (ctrl->flags & FCCTRL_TERMIO) + ctrl->iocnt--; + op->flags &= ~FCOP_FLAGS_TERMIO; + spin_unlock_irqrestore(&ctrl->lock, flags); + return; + } } -/* - * Called to teardown an association. - * May be called with association fully in place or partially in place. - */ -static void -__nvme_fc_remove_ctrl(struct nvme_fc_ctrl *ctrl) -{ - nvme_stop_keep_alive(&ctrl->ctrl); - /* stop and terminate ios on admin and io queues */ - nvme_fc_shutdown_ctrl(ctrl); - - /* - * tear down the controller - * This will result in the last reference on the nvme ctrl to - * expire, calling the transport nvme_fc_free_nvme_ctrl() callback. - * From there, the transport will tear down it's logical queues and - * association. - */ - nvme_uninit_ctrl(&ctrl->ctrl); - - nvme_put_ctrl(&ctrl->ctrl); -} - -static void -nvme_fc_del_ctrl_work(struct work_struct *work) -{ - struct nvme_fc_ctrl *ctrl = - container_of(work, struct nvme_fc_ctrl, delete_work); - - __nvme_fc_remove_ctrl(ctrl); -} - -static int -__nvme_fc_del_ctrl(struct nvme_fc_ctrl *ctrl) -{ - if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_DELETING)) - return -EBUSY; - - if (!queue_work(nvme_fc_wq, &ctrl->delete_work)) - return -EBUSY; - - return 0; -} - -/* - * Request from nvme core layer to delete the controller - */ -static int -nvme_fc_del_nvme_ctrl(struct nvme_ctrl *nctrl) -{ - struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl); - struct nvme_fc_rport *rport = ctrl->rport; - unsigned long flags; - int ret; - - spin_lock_irqsave(&rport->lock, flags); - ret = __nvme_fc_del_ctrl(ctrl); - spin_unlock_irqrestore(&rport->lock, flags); - if (ret) - return ret; - - flush_work(&ctrl->delete_work); - - return 0; -} - -static int -nvme_fc_reset_nvme_ctrl(struct nvme_ctrl *nctrl) -{ - return -EIO; -} - -static const struct nvme_ctrl_ops nvme_fc_ctrl_ops = { - .name = "fc", - .module = THIS_MODULE, - .is_fabrics = true, - .reg_read32 = nvmf_reg_read32, - .reg_read64 = nvmf_reg_read64, - .reg_write32 = nvmf_reg_write32, - .reset_ctrl = nvme_fc_reset_nvme_ctrl, - .free_ctrl = nvme_fc_free_nvme_ctrl, - .submit_async_event = nvme_fc_submit_async_event, - .delete_ctrl = nvme_fc_del_nvme_ctrl, - .get_subsysnqn = nvmf_get_subsysnqn, - .get_address = nvmf_get_address, +static const struct blk_mq_ops nvme_fc_mq_ops = { + .queue_rq = nvme_fc_queue_rq, + .complete = nvme_fc_complete_rq, + .init_request = nvme_fc_init_request, + .exit_request = nvme_fc_exit_request, + .reinit_request = nvme_fc_reinit_request, + .init_hctx = nvme_fc_init_hctx, + .poll = nvme_fc_poll, + .timeout = nvme_fc_timeout, }; static int @@ -2323,15 +2249,478 @@ out_free_tag_set: return ret; } +static int +nvme_fc_reinit_io_queues(struct nvme_fc_ctrl *ctrl) +{ + struct nvmf_ctrl_options *opts = ctrl->ctrl.opts; + int ret; + + ret = nvme_set_queue_count(&ctrl->ctrl, &opts->nr_io_queues); + if (ret) { + dev_info(ctrl->ctrl.device, + "set_queue_count failed: %d\n", ret); + return ret; + } + + /* check for io queues existing */ + if (ctrl->queue_count == 1) + return 0; + + dev_info(ctrl->ctrl.device, "Recreating %d I/O queues.\n", + opts->nr_io_queues); + + nvme_fc_init_io_queues(ctrl); + + ret = blk_mq_reinit_tagset(&ctrl->tag_set); + if (ret) + goto out_free_io_queues; + + ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.opts->queue_size); + if (ret) + goto out_free_io_queues; + + ret = nvme_fc_connect_io_queues(ctrl, ctrl->ctrl.opts->queue_size); + if (ret) + goto out_delete_hw_queues; + + return 0; + +out_delete_hw_queues: + nvme_fc_delete_hw_io_queues(ctrl); +out_free_io_queues: + nvme_fc_free_io_queues(ctrl); + return ret; +} + +/* + * This routine restarts the controller on the host side, and + * on the link side, recreates the controller association. + */ +static int +nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) +{ + struct nvmf_ctrl_options *opts = ctrl->ctrl.opts; + u32 segs; + int ret; + bool changed; + + ctrl->connect_attempts++; + + /* + * Create the admin queue + */ + + nvme_fc_init_queue(ctrl, 0, NVME_FC_AQ_BLKMQ_DEPTH); + + ret = __nvme_fc_create_hw_queue(ctrl, &ctrl->queues[0], 0, + NVME_FC_AQ_BLKMQ_DEPTH); + if (ret) + goto out_free_queue; + + ret = nvme_fc_connect_admin_queue(ctrl, &ctrl->queues[0], + NVME_FC_AQ_BLKMQ_DEPTH, + (NVME_FC_AQ_BLKMQ_DEPTH / 4)); + if (ret) + goto out_delete_hw_queue; + + if (ctrl->ctrl.state != NVME_CTRL_NEW) + blk_mq_start_stopped_hw_queues(ctrl->ctrl.admin_q, true); + + ret = nvmf_connect_admin_queue(&ctrl->ctrl); + if (ret) + goto out_disconnect_admin_queue; + + /* + * Check controller capabilities + * + * todo:- add code to check if ctrl attributes changed from + * prior connection values + */ + + ret = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap); + if (ret) { + dev_err(ctrl->ctrl.device, + "prop_get NVME_REG_CAP failed\n"); + goto out_disconnect_admin_queue; + } + + ctrl->ctrl.sqsize = + min_t(int, NVME_CAP_MQES(ctrl->cap) + 1, ctrl->ctrl.sqsize); + + ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap); + if (ret) + goto out_disconnect_admin_queue; + + segs = min_t(u32, NVME_FC_MAX_SEGMENTS, + ctrl->lport->ops->max_sgl_segments); + ctrl->ctrl.max_hw_sectors = (segs - 1) << (PAGE_SHIFT - 9); + + ret = nvme_init_identify(&ctrl->ctrl); + if (ret) + goto out_disconnect_admin_queue; + + /* sanity checks */ + + /* FC-NVME does not have other data in the capsule */ + if (ctrl->ctrl.icdoff) { + dev_err(ctrl->ctrl.device, "icdoff %d is not supported!\n", + ctrl->ctrl.icdoff); + goto out_disconnect_admin_queue; + } + + nvme_start_keep_alive(&ctrl->ctrl); + + /* FC-NVME supports normal SGL Data Block Descriptors */ + + if (opts->queue_size > ctrl->ctrl.maxcmd) { + /* warn if maxcmd is lower than queue_size */ + dev_warn(ctrl->ctrl.device, + "queue_size %zu > ctrl maxcmd %u, reducing " + "to queue_size\n", + opts->queue_size, ctrl->ctrl.maxcmd); + opts->queue_size = ctrl->ctrl.maxcmd; + } + + ret = nvme_fc_init_aen_ops(ctrl); + if (ret) + goto out_term_aen_ops; + + /* + * Create the io queues + */ + + if (ctrl->queue_count > 1) { + if (ctrl->ctrl.state == NVME_CTRL_NEW) + ret = nvme_fc_create_io_queues(ctrl); + else + ret = nvme_fc_reinit_io_queues(ctrl); + if (ret) + goto out_term_aen_ops; + } + + changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE); + WARN_ON_ONCE(!changed); + + ctrl->connect_attempts = 0; + + kref_get(&ctrl->ctrl.kref); + + if (ctrl->queue_count > 1) { + nvme_start_queues(&ctrl->ctrl); + nvme_queue_scan(&ctrl->ctrl); + nvme_queue_async_events(&ctrl->ctrl); + } + + return 0; /* Success */ + +out_term_aen_ops: + nvme_fc_term_aen_ops(ctrl); + nvme_stop_keep_alive(&ctrl->ctrl); +out_disconnect_admin_queue: + /* send a Disconnect(association) LS to fc-nvme target */ + nvme_fc_xmt_disconnect_assoc(ctrl); +out_delete_hw_queue: + __nvme_fc_delete_hw_queue(ctrl, &ctrl->queues[0], 0); +out_free_queue: + nvme_fc_free_queue(&ctrl->queues[0]); + + return ret; +} + +/* + * This routine stops operation of the controller on the host side. + * On the host os stack side: Admin and IO queues are stopped, + * outstanding ios on them terminated via FC ABTS. + * On the link side: the association is terminated. + */ +static void +nvme_fc_delete_association(struct nvme_fc_ctrl *ctrl) +{ + unsigned long flags; + + nvme_stop_keep_alive(&ctrl->ctrl); + + spin_lock_irqsave(&ctrl->lock, flags); + ctrl->flags |= FCCTRL_TERMIO; + ctrl->iocnt = 0; + spin_unlock_irqrestore(&ctrl->lock, flags); + + /* + * If io queues are present, stop them and terminate all outstanding + * ios on them. As FC allocates FC exchange for each io, the + * transport must contact the LLDD to terminate the exchange, + * thus releasing the FC exchange. We use blk_mq_tagset_busy_itr() + * to tell us what io's are busy and invoke a transport routine + * to kill them with the LLDD. After terminating the exchange + * the LLDD will call the transport's normal io done path, but it + * will have an aborted status. The done path will return the + * io requests back to the block layer as part of normal completions + * (but with error status). + */ + if (ctrl->queue_count > 1) { + nvme_stop_queues(&ctrl->ctrl); + blk_mq_tagset_busy_iter(&ctrl->tag_set, + nvme_fc_terminate_exchange, &ctrl->ctrl); + } + + /* + * Other transports, which don't have link-level contexts bound + * to sqe's, would try to gracefully shutdown the controller by + * writing the registers for shutdown and polling (call + * nvme_shutdown_ctrl()). Given a bunch of i/o was potentially + * just aborted and we will wait on those contexts, and given + * there was no indication of how live the controlelr is on the + * link, don't send more io to create more contexts for the + * shutdown. Let the controller fail via keepalive failure if + * its still present. + */ + + /* + * clean up the admin queue. Same thing as above. + * use blk_mq_tagset_busy_itr() and the transport routine to + * terminate the exchanges. + */ + blk_mq_stop_hw_queues(ctrl->ctrl.admin_q); + blk_mq_tagset_busy_iter(&ctrl->admin_tag_set, + nvme_fc_terminate_exchange, &ctrl->ctrl); + + /* kill the aens as they are a separate path */ + nvme_fc_abort_aen_ops(ctrl); + + /* wait for all io that had to be aborted */ + spin_lock_irqsave(&ctrl->lock, flags); + while (ctrl->iocnt) { + spin_unlock_irqrestore(&ctrl->lock, flags); + msleep(1000); + spin_lock_irqsave(&ctrl->lock, flags); + } + ctrl->flags &= ~FCCTRL_TERMIO; + spin_unlock_irqrestore(&ctrl->lock, flags); + + nvme_fc_term_aen_ops(ctrl); + + /* + * send a Disconnect(association) LS to fc-nvme target + * Note: could have been sent at top of process, but + * cleaner on link traffic if after the aborts complete. + * Note: if association doesn't exist, association_id will be 0 + */ + if (ctrl->association_id) + nvme_fc_xmt_disconnect_assoc(ctrl); + + if (ctrl->ctrl.tagset) { + nvme_fc_delete_hw_io_queues(ctrl); + nvme_fc_free_io_queues(ctrl); + } + + __nvme_fc_delete_hw_queue(ctrl, &ctrl->queues[0], 0); + nvme_fc_free_queue(&ctrl->queues[0]); +} + +static void +nvme_fc_delete_ctrl_work(struct work_struct *work) +{ + struct nvme_fc_ctrl *ctrl = + container_of(work, struct nvme_fc_ctrl, delete_work); + + cancel_work_sync(&ctrl->reset_work); + cancel_delayed_work_sync(&ctrl->connect_work); + + /* + * kill the association on the link side. this will block + * waiting for io to terminate + */ + nvme_fc_delete_association(ctrl); + + /* + * tear down the controller + * This will result in the last reference on the nvme ctrl to + * expire, calling the transport nvme_fc_nvme_ctrl_freed() callback. + * From there, the transport will tear down it's logical queues and + * association. + */ + nvme_uninit_ctrl(&ctrl->ctrl); + + nvme_put_ctrl(&ctrl->ctrl); +} + +static int +__nvme_fc_del_ctrl(struct nvme_fc_ctrl *ctrl) +{ + if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_DELETING)) + return -EBUSY; + + if (!queue_work(nvme_fc_wq, &ctrl->delete_work)) + return -EBUSY; + + return 0; +} + +/* + * Request from nvme core layer to delete the controller + */ +static int +nvme_fc_del_nvme_ctrl(struct nvme_ctrl *nctrl) +{ + struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl); + int ret; + + if (!kref_get_unless_zero(&ctrl->ctrl.kref)) + return -EBUSY; + + ret = __nvme_fc_del_ctrl(ctrl); + + if (!ret) + flush_workqueue(nvme_fc_wq); + + nvme_put_ctrl(&ctrl->ctrl); + + return ret; +} + +static void +nvme_fc_reset_ctrl_work(struct work_struct *work) +{ + struct nvme_fc_ctrl *ctrl = + container_of(work, struct nvme_fc_ctrl, reset_work); + int ret; + + /* will block will waiting for io to terminate */ + nvme_fc_delete_association(ctrl); + + ret = nvme_fc_create_association(ctrl); + if (ret) { + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: reset: Reconnect attempt failed (%d)\n", + ctrl->cnum, ret); + if (ctrl->connect_attempts >= NVME_FC_MAX_CONNECT_ATTEMPTS) { + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: Max reconnect attempts (%d) " + "reached. Removing controller\n", + ctrl->cnum, ctrl->connect_attempts); + + if (!nvme_change_ctrl_state(&ctrl->ctrl, + NVME_CTRL_DELETING)) { + dev_err(ctrl->ctrl.device, + "NVME-FC{%d}: failed to change state " + "to DELETING\n", ctrl->cnum); + return; + } + + WARN_ON(!queue_work(nvme_fc_wq, &ctrl->delete_work)); + return; + } + + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: Reconnect attempt in %d seconds.\n", + ctrl->cnum, ctrl->reconnect_delay); + queue_delayed_work(nvme_fc_wq, &ctrl->connect_work, + ctrl->reconnect_delay * HZ); + } else + dev_info(ctrl->ctrl.device, + "NVME-FC{%d}: controller reset complete\n", ctrl->cnum); +} + +/* + * called by the nvme core layer, for sysfs interface that requests + * a reset of the nvme controller + */ +static int +nvme_fc_reset_nvme_ctrl(struct nvme_ctrl *nctrl) +{ + struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl); + + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: admin requested controller reset\n", ctrl->cnum); + + if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RESETTING)) + return -EBUSY; + + if (!queue_work(nvme_fc_wq, &ctrl->reset_work)) + return -EBUSY; + + flush_work(&ctrl->reset_work); + + return 0; +} + +static const struct nvme_ctrl_ops nvme_fc_ctrl_ops = { + .name = "fc", + .module = THIS_MODULE, + .is_fabrics = true, + .reg_read32 = nvmf_reg_read32, + .reg_read64 = nvmf_reg_read64, + .reg_write32 = nvmf_reg_write32, + .reset_ctrl = nvme_fc_reset_nvme_ctrl, + .free_ctrl = nvme_fc_nvme_ctrl_freed, + .submit_async_event = nvme_fc_submit_async_event, + .delete_ctrl = nvme_fc_del_nvme_ctrl, + .get_subsysnqn = nvmf_get_subsysnqn, + .get_address = nvmf_get_address, +}; + +static void +nvme_fc_connect_ctrl_work(struct work_struct *work) +{ + int ret; + + struct nvme_fc_ctrl *ctrl = + container_of(to_delayed_work(work), + struct nvme_fc_ctrl, connect_work); + + ret = nvme_fc_create_association(ctrl); + if (ret) { + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: Reconnect attempt failed (%d)\n", + ctrl->cnum, ret); + if (ctrl->connect_attempts >= NVME_FC_MAX_CONNECT_ATTEMPTS) { + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: Max reconnect attempts (%d) " + "reached. Removing controller\n", + ctrl->cnum, ctrl->connect_attempts); + + if (!nvme_change_ctrl_state(&ctrl->ctrl, + NVME_CTRL_DELETING)) { + dev_err(ctrl->ctrl.device, + "NVME-FC{%d}: failed to change state " + "to DELETING\n", ctrl->cnum); + return; + } + + WARN_ON(!queue_work(nvme_fc_wq, &ctrl->delete_work)); + return; + } + + dev_warn(ctrl->ctrl.device, + "NVME-FC{%d}: Reconnect attempt in %d seconds.\n", + ctrl->cnum, ctrl->reconnect_delay); + queue_delayed_work(nvme_fc_wq, &ctrl->connect_work, + ctrl->reconnect_delay * HZ); + } else + dev_info(ctrl->ctrl.device, + "NVME-FC{%d}: controller reconnect complete\n", + ctrl->cnum); +} + + +static const struct blk_mq_ops nvme_fc_admin_mq_ops = { + .queue_rq = nvme_fc_queue_rq, + .complete = nvme_fc_complete_rq, + .init_request = nvme_fc_init_admin_request, + .exit_request = nvme_fc_exit_request, + .reinit_request = nvme_fc_reinit_request, + .init_hctx = nvme_fc_init_admin_hctx, + .timeout = nvme_fc_timeout, +}; + static struct nvme_ctrl * -__nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts, +nvme_fc_init_ctrl(struct device *dev, struct nvmf_ctrl_options *opts, struct nvme_fc_lport *lport, struct nvme_fc_rport *rport) { struct nvme_fc_ctrl *ctrl; unsigned long flags; int ret, idx; - bool changed; ctrl = kzalloc(sizeof(*ctrl), GFP_KERNEL); if (!ctrl) { @@ -2350,17 +2739,15 @@ __nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts, ctrl->lport = lport; ctrl->rport = rport; ctrl->dev = lport->dev; - ctrl->state = FCCTRL_INIT; ctrl->cnum = idx; - ret = nvme_init_ctrl(&ctrl->ctrl, dev, &nvme_fc_ctrl_ops, 0); - if (ret) - goto out_free_ida; - get_device(ctrl->dev); kref_init(&ctrl->ref); - INIT_WORK(&ctrl->delete_work, nvme_fc_del_ctrl_work); + INIT_WORK(&ctrl->delete_work, nvme_fc_delete_ctrl_work); + INIT_WORK(&ctrl->reset_work, nvme_fc_reset_ctrl_work); + INIT_DELAYED_WORK(&ctrl->connect_work, nvme_fc_connect_ctrl_work); + ctrl->reconnect_delay = opts->reconnect_delay; spin_lock_init(&ctrl->lock); /* io queue count */ @@ -2377,87 +2764,87 @@ __nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts, ctrl->queues = kcalloc(ctrl->queue_count, sizeof(struct nvme_fc_queue), GFP_KERNEL); if (!ctrl->queues) - goto out_uninit_ctrl; + goto out_free_ida; - ret = nvme_fc_configure_admin_queue(ctrl); + memset(&ctrl->admin_tag_set, 0, sizeof(ctrl->admin_tag_set)); + ctrl->admin_tag_set.ops = &nvme_fc_admin_mq_ops; + ctrl->admin_tag_set.queue_depth = NVME_FC_AQ_BLKMQ_DEPTH; + ctrl->admin_tag_set.reserved_tags = 2; /* fabric connect + Keep-Alive */ + ctrl->admin_tag_set.numa_node = NUMA_NO_NODE; + ctrl->admin_tag_set.cmd_size = sizeof(struct nvme_fc_fcp_op) + + (SG_CHUNK_SIZE * + sizeof(struct scatterlist)) + + ctrl->lport->ops->fcprqst_priv_sz; + ctrl->admin_tag_set.driver_data = ctrl; + ctrl->admin_tag_set.nr_hw_queues = 1; + ctrl->admin_tag_set.timeout = ADMIN_TIMEOUT; + + ret = blk_mq_alloc_tag_set(&ctrl->admin_tag_set); if (ret) - goto out_uninit_ctrl; + goto out_free_queues; - /* sanity checks */ - - /* FC-NVME does not have other data in the capsule */ - if (ctrl->ctrl.icdoff) { - dev_err(ctrl->ctrl.device, "icdoff %d is not supported!\n", - ctrl->ctrl.icdoff); - goto out_remove_admin_queue; + ctrl->ctrl.admin_q = blk_mq_init_queue(&ctrl->admin_tag_set); + if (IS_ERR(ctrl->ctrl.admin_q)) { + ret = PTR_ERR(ctrl->ctrl.admin_q); + goto out_free_admin_tag_set; } - /* FC-NVME supports normal SGL Data Block Descriptors */ + /* + * Would have been nice to init io queues tag set as well. + * However, we require interaction from the controller + * for max io queue count before we can do so. + * Defer this to the connect path. + */ - if (opts->queue_size > ctrl->ctrl.maxcmd) { - /* warn if maxcmd is lower than queue_size */ - dev_warn(ctrl->ctrl.device, - "queue_size %zu > ctrl maxcmd %u, reducing " - "to queue_size\n", - opts->queue_size, ctrl->ctrl.maxcmd); - opts->queue_size = ctrl->ctrl.maxcmd; - } - - ret = nvme_fc_init_aen_ops(ctrl); + ret = nvme_init_ctrl(&ctrl->ctrl, dev, &nvme_fc_ctrl_ops, 0); if (ret) - goto out_exit_aen_ops; + goto out_cleanup_admin_q; - if (ctrl->queue_count > 1) { - ret = nvme_fc_create_io_queues(ctrl); - if (ret) - goto out_exit_aen_ops; - } - - spin_lock_irqsave(&ctrl->lock, flags); - ctrl->state = FCCTRL_ACTIVE; - spin_unlock_irqrestore(&ctrl->lock, flags); - - changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE); - WARN_ON_ONCE(!changed); - - dev_info(ctrl->ctrl.device, - "NVME-FC{%d}: new ctrl: NQN \"%s\"\n", - ctrl->cnum, ctrl->ctrl.opts->subsysnqn); - - kref_get(&ctrl->ctrl.kref); + /* at this point, teardown path changes to ref counting on nvme ctrl */ spin_lock_irqsave(&rport->lock, flags); list_add_tail(&ctrl->ctrl_list, &rport->ctrl_list); spin_unlock_irqrestore(&rport->lock, flags); - if (opts->nr_io_queues) { - nvme_queue_scan(&ctrl->ctrl); - nvme_queue_async_events(&ctrl->ctrl); + ret = nvme_fc_create_association(ctrl); + if (ret) { + ctrl->ctrl.opts = NULL; + /* initiate nvme ctrl ref counting teardown */ + nvme_uninit_ctrl(&ctrl->ctrl); + nvme_put_ctrl(&ctrl->ctrl); + + /* as we're past the point where we transition to the ref + * counting teardown path, if we return a bad pointer here, + * the calling routine, thinking it's prior to the + * transition, will do an rport put. Since the teardown + * path also does a rport put, we do an extra get here to + * so proper order/teardown happens. + */ + nvme_fc_rport_get(rport); + + if (ret > 0) + ret = -EIO; + return ERR_PTR(ret); } + dev_info(ctrl->ctrl.device, + "NVME-FC{%d}: new ctrl: NQN \"%s\"\n", + ctrl->cnum, ctrl->ctrl.opts->subsysnqn); + return &ctrl->ctrl; -out_exit_aen_ops: - nvme_fc_exit_aen_ops(ctrl); -out_remove_admin_queue: - /* send a Disconnect(association) LS to fc-nvme target */ - nvme_fc_xmt_disconnect_assoc(ctrl); - nvme_stop_keep_alive(&ctrl->ctrl); - nvme_fc_destroy_admin_queue(ctrl); -out_uninit_ctrl: - nvme_uninit_ctrl(&ctrl->ctrl); - nvme_put_ctrl(&ctrl->ctrl); - if (ret > 0) - ret = -EIO; - /* exit via here will follow ctlr ref point callbacks to free */ - return ERR_PTR(ret); - +out_cleanup_admin_q: + blk_cleanup_queue(ctrl->ctrl.admin_q); +out_free_admin_tag_set: + blk_mq_free_tag_set(&ctrl->admin_tag_set); +out_free_queues: + kfree(ctrl->queues); out_free_ida: + put_device(ctrl->dev); ida_simple_remove(&nvme_fc_ctrl_cnt, ctrl->cnum); out_free_ctrl: kfree(ctrl); out_fail: - nvme_fc_rport_put(rport); /* exit via here doesn't follow ctlr ref points */ return ERR_PTR(ret); } @@ -2529,6 +2916,7 @@ nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts) { struct nvme_fc_lport *lport; struct nvme_fc_rport *rport; + struct nvme_ctrl *ctrl; struct nvmet_fc_traddr laddr = { 0L, 0L }; struct nvmet_fc_traddr raddr = { 0L, 0L }; unsigned long flags; @@ -2560,7 +2948,10 @@ nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts) spin_unlock_irqrestore(&nvme_fc_lock, flags); - return __nvme_fc_create_ctrl(dev, opts, lport, rport); + ctrl = nvme_fc_init_ctrl(dev, opts, lport, rport); + if (IS_ERR(ctrl)) + nvme_fc_rport_put(rport); + return ctrl; } } spin_unlock_irqrestore(&nvme_fc_lock, flags); diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c index de61a4a03d78..e4e4e60b1224 100644 --- a/drivers/nvme/host/lightnvm.c +++ b/drivers/nvme/host/lightnvm.c @@ -483,7 +483,7 @@ static void nvme_nvm_end_io(struct request *rq, int error) { struct nvm_rq *rqd = rq->end_io_data; - rqd->ppa_status = nvme_req(rq)->result.u64; + rqd->ppa_status = le64_to_cpu(nvme_req(rq)->result.u64); rqd->error = nvme_req(rq)->status; nvm_end_io(rqd); diff --git a/drivers/nvme/host/scsi.c b/drivers/nvme/host/scsi.c index f49ae2758bb7..1f7671e631dd 100644 --- a/drivers/nvme/host/scsi.c +++ b/drivers/nvme/host/scsi.c @@ -1609,7 +1609,7 @@ static int nvme_trans_do_nvme_io(struct nvme_ns *ns, struct sg_io_hdr *hdr, struct nvme_command c; u8 opcode = (is_write ? nvme_cmd_write : nvme_cmd_read); u16 control; - u32 max_blocks = queue_max_hw_sectors(ns->queue); + u32 max_blocks = queue_max_hw_sectors(ns->queue) >> (ns->lba_shift - 9); num_cmds = nvme_trans_io_get_num_cmds(hdr, cdb_info, max_blocks); @@ -2138,15 +2138,6 @@ static int nvme_trans_request_sense(struct nvme_ns *ns, struct sg_io_hdr *hdr, return res; } -static int nvme_trans_security_protocol(struct nvme_ns *ns, - struct sg_io_hdr *hdr, - u8 *cmd) -{ - return nvme_trans_completion(hdr, SAM_STAT_CHECK_CONDITION, - ILLEGAL_REQUEST, SCSI_ASC_ILLEGAL_COMMAND, - SCSI_ASCQ_CAUSE_NOT_REPORTABLE); -} - static int nvme_trans_synchronize_cache(struct nvme_ns *ns, struct sg_io_hdr *hdr) { @@ -2414,10 +2405,6 @@ static int nvme_scsi_translate(struct nvme_ns *ns, struct sg_io_hdr *hdr) case REQUEST_SENSE: retcode = nvme_trans_request_sense(ns, hdr, cmd); break; - case SECURITY_PROTOCOL_IN: - case SECURITY_PROTOCOL_OUT: - retcode = nvme_trans_security_protocol(ns, hdr, cmd); - break; case SYNCHRONIZE_CACHE: retcode = nvme_trans_synchronize_cache(ns, hdr); break; diff --git a/drivers/nvme/target/fc.c b/drivers/nvme/target/fc.c index 074bd3743b5f..62eba29c85fb 100644 --- a/drivers/nvme/target/fc.c +++ b/drivers/nvme/target/fc.c @@ -119,7 +119,7 @@ struct nvmet_fc_tgt_queue { u16 qid; u16 sqsize; u16 ersp_ratio; - u16 sqhd; + __le16 sqhd; int cpu; atomic_t connected; atomic_t sqtail; @@ -1058,7 +1058,7 @@ EXPORT_SYMBOL_GPL(nvmet_fc_unregister_targetport); static void -nvmet_fc_format_rsp_hdr(void *buf, u8 ls_cmd, u32 desc_len, u8 rqst_ls_cmd) +nvmet_fc_format_rsp_hdr(void *buf, u8 ls_cmd, __be32 desc_len, u8 rqst_ls_cmd) { struct fcnvme_ls_acc_hdr *acc = buf; @@ -1700,7 +1700,7 @@ nvmet_fc_prep_fcp_rsp(struct nvmet_fc_tgtport *tgtport, xfr_length != fod->total_length || (le16_to_cpu(cqe->status) & 0xFFFE) || cqewd[0] || cqewd[1] || (sqe->flags & (NVME_CMD_FUSE_FIRST | NVME_CMD_FUSE_SECOND)) || - queue_90percent_full(fod->queue, cqe->sq_head)) + queue_90percent_full(fod->queue, le16_to_cpu(cqe->sq_head))) send_ersp = true; /* re-set the fields */ @@ -2055,7 +2055,7 @@ nvmet_fc_fcp_nvme_cmd_done(struct nvmet_req *nvme_req) /* * Actual processing routine for received FC-NVME LS Requests from the LLD */ -void +static void nvmet_fc_handle_fcp_rqst(struct nvmet_fc_tgtport *tgtport, struct nvmet_fc_fcp_iod *fod) { diff --git a/drivers/nvme/target/fcloop.c b/drivers/nvme/target/fcloop.c index aaa3dbe22bd5..15551ef79c8c 100644 --- a/drivers/nvme/target/fcloop.c +++ b/drivers/nvme/target/fcloop.c @@ -666,7 +666,7 @@ fcloop_targetport_delete(struct nvmet_fc_target_port *targetport) #define FCLOOP_SGL_SEGS 256 #define FCLOOP_DMABOUND_4G 0xFFFFFFFF -struct nvme_fc_port_template fctemplate = { +static struct nvme_fc_port_template fctemplate = { .localport_delete = fcloop_localport_delete, .remoteport_delete = fcloop_remoteport_delete, .create_queue = fcloop_create_queue, @@ -686,7 +686,7 @@ struct nvme_fc_port_template fctemplate = { .fcprqst_priv_sz = sizeof(struct fcloop_ini_fcpreq), }; -struct nvmet_fc_target_template tgttemplate = { +static struct nvmet_fc_target_template tgttemplate = { .targetport_delete = fcloop_targetport_delete, .xmt_ls_rsp = fcloop_xmt_ls_rsp, .fcp_op = fcloop_fcp_op, diff --git a/drivers/scsi/lpfc/lpfc.h b/drivers/scsi/lpfc/lpfc.h index 257bbdd0f0b8..6d7840b096e6 100644 --- a/drivers/scsi/lpfc/lpfc.h +++ b/drivers/scsi/lpfc/lpfc.h @@ -56,7 +56,7 @@ struct lpfc_sli2_slim; #define LPFC_MAX_SG_SEG_CNT 4096 /* sg element count per scsi cmnd */ #define LPFC_MAX_SGL_SEG_CNT 512 /* SGL element count per scsi cmnd */ #define LPFC_MAX_BPL_SEG_CNT 4096 /* BPL element count per scsi cmnd */ -#define LPFC_MIN_NVME_SEG_CNT 254 +#define LPFC_MAX_NVME_SEG_CNT 128 /* max SGL element cnt per NVME cmnd */ #define LPFC_MAX_SGE_SIZE 0x80000000 /* Maximum data allowed in a SGE */ #define LPFC_IOCB_LIST_CNT 2250 /* list of IOCBs for fast-path usage. */ @@ -474,6 +474,8 @@ struct lpfc_vport { unsigned long rcv_buffer_time_stamp; uint32_t vport_flag; #define STATIC_VPORT 1 +#define FAWWPN_SET 2 +#define FAWWPN_PARAM_CHG 4 uint16_t fdmi_num_disc; uint32_t fdmi_hba_mask; @@ -781,6 +783,7 @@ struct lpfc_hba { uint32_t cfg_nvmet_fb_size; uint32_t cfg_total_seg_cnt; uint32_t cfg_sg_seg_cnt; + uint32_t cfg_nvme_seg_cnt; uint32_t cfg_sg_dma_buf_size; uint64_t cfg_soft_wwnn; uint64_t cfg_soft_wwpn; diff --git a/drivers/scsi/lpfc/lpfc_attr.c b/drivers/scsi/lpfc/lpfc_attr.c index 22819afbaef5..513fd07715cd 100644 --- a/drivers/scsi/lpfc/lpfc_attr.c +++ b/drivers/scsi/lpfc/lpfc_attr.c @@ -2292,6 +2292,8 @@ lpfc_soft_wwn_enable_store(struct device *dev, struct device_attribute *attr, struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata; struct lpfc_hba *phba = vport->phba; unsigned int cnt = count; + uint8_t vvvl = vport->fc_sparam.cmn.valid_vendor_ver_level; + u32 *fawwpn_key = (uint32_t *)&vport->fc_sparam.un.vendorVersion[0]; /* * We're doing a simple sanity check for soft_wwpn setting. @@ -2305,6 +2307,12 @@ lpfc_soft_wwn_enable_store(struct device *dev, struct device_attribute *attr, * here. The intent is to protect against the random user or * application that is just writing attributes. */ + if (vvvl == 1 && cpu_to_be32(*fawwpn_key) == FAPWWN_KEY_VENDOR) { + lpfc_printf_log(phba, KERN_ERR, LOG_INIT, + "0051 "LPFC_DRIVER_NAME" soft wwpn can not" + " be enabled: fawwpn is enabled\n"); + return -EINVAL; + } /* count may include a LF at end of string */ if (buf[cnt-1] == '\n') @@ -3335,7 +3343,7 @@ LPFC_ATTR_R(enable_fc4_type, LPFC_ENABLE_FCP, * percentage will go to NVME. */ LPFC_ATTR_R(xri_split, 50, 10, 90, - "Division of XRI resources between SCSI and NVME"); + "Division of XRI resources between SCSI and NVME"); /* # lpfc_log_verbose: Only turn this flag on if you are willing to risk being diff --git a/drivers/scsi/lpfc/lpfc_bsg.c b/drivers/scsi/lpfc/lpfc_bsg.c index 18157d2840a3..a1686c2d863c 100644 --- a/drivers/scsi/lpfc/lpfc_bsg.c +++ b/drivers/scsi/lpfc/lpfc_bsg.c @@ -2486,6 +2486,10 @@ static int lpfcdiag_loop_self_reg(struct lpfc_hba *phba, uint16_t *rpi) mbox, *rpi); else { *rpi = lpfc_sli4_alloc_rpi(phba); + if (*rpi == LPFC_RPI_ALLOC_ERROR) { + mempool_free(mbox, phba->mbox_mem_pool); + return -EBUSY; + } status = lpfc_reg_rpi(phba, phba->pport->vpi, phba->pport->fc_myDID, (uint8_t *)&phba->pport->fc_sparam, diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h index 54e6ac42fbcd..944b32ca4931 100644 --- a/drivers/scsi/lpfc/lpfc_crtn.h +++ b/drivers/scsi/lpfc/lpfc_crtn.h @@ -24,6 +24,7 @@ typedef int (*node_filter)(struct lpfc_nodelist *, void *); struct fc_rport; struct fc_frame_header; +struct lpfc_nvmet_rcv_ctx; void lpfc_down_link(struct lpfc_hba *, LPFC_MBOXQ_t *); void lpfc_sli_read_link_ste(struct lpfc_hba *); void lpfc_dump_mem(struct lpfc_hba *, LPFC_MBOXQ_t *, uint16_t, uint16_t); @@ -99,7 +100,7 @@ void lpfc_issue_reg_vpi(struct lpfc_hba *, struct lpfc_vport *); int lpfc_check_sli_ndlp(struct lpfc_hba *, struct lpfc_sli_ring *, struct lpfc_iocbq *, struct lpfc_nodelist *); -void lpfc_nlp_init(struct lpfc_vport *, struct lpfc_nodelist *, uint32_t); +struct lpfc_nodelist *lpfc_nlp_init(struct lpfc_vport *vport, uint32_t did); struct lpfc_nodelist *lpfc_nlp_get(struct lpfc_nodelist *); int lpfc_nlp_put(struct lpfc_nodelist *); int lpfc_nlp_not_used(struct lpfc_nodelist *ndlp); @@ -245,6 +246,10 @@ struct hbq_dmabuf *lpfc_sli4_rb_alloc(struct lpfc_hba *); void lpfc_sli4_rb_free(struct lpfc_hba *, struct hbq_dmabuf *); struct rqb_dmabuf *lpfc_sli4_nvmet_alloc(struct lpfc_hba *phba); void lpfc_sli4_nvmet_free(struct lpfc_hba *phba, struct rqb_dmabuf *dmab); +void lpfc_nvmet_rq_post(struct lpfc_hba *phba, struct lpfc_nvmet_rcv_ctx *ctxp, + struct lpfc_dmabuf *mp); +int lpfc_nvmet_rcv_unsol_abort(struct lpfc_vport *vport, + struct fc_frame_header *fc_hdr); void lpfc_sli4_build_dflt_fcf_record(struct lpfc_hba *, struct fcf_record *, uint16_t); int lpfc_sli4_rq_put(struct lpfc_queue *hq, struct lpfc_queue *dq, @@ -302,6 +307,8 @@ int lpfc_sli_check_eratt(struct lpfc_hba *); void lpfc_sli_handle_slow_ring_event(struct lpfc_hba *, struct lpfc_sli_ring *, uint32_t); void lpfc_sli4_handle_received_buffer(struct lpfc_hba *, struct hbq_dmabuf *); +void lpfc_sli4_seq_abort_rsp(struct lpfc_vport *vport, + struct fc_frame_header *fc_hdr, bool aborted); void lpfc_sli_def_mbox_cmpl(struct lpfc_hba *, LPFC_MBOXQ_t *); void lpfc_sli4_unreg_rpi_cmpl_clr(struct lpfc_hba *, LPFC_MBOXQ_t *); int lpfc_sli_issue_iocb(struct lpfc_hba *, uint32_t, diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c index d3e9af983015..1487406aea77 100644 --- a/drivers/scsi/lpfc/lpfc_ct.c +++ b/drivers/scsi/lpfc/lpfc_ct.c @@ -537,19 +537,53 @@ lpfc_prep_node_fc4type(struct lpfc_vport *vport, uint32_t Did, uint8_t fc4_type) } } +static void +lpfc_ns_rsp_audit_did(struct lpfc_vport *vport, uint32_t Did, uint8_t fc4_type) +{ + struct lpfc_hba *phba = vport->phba; + struct lpfc_nodelist *ndlp = NULL; + struct Scsi_Host *shost = lpfc_shost_from_vport(vport); + + /* + * To conserve rpi's, filter out addresses for other + * vports on the same physical HBAs. + */ + if (Did != vport->fc_myDID && + (!lpfc_find_vport_by_did(phba, Did) || + vport->cfg_peer_port_login)) { + if (!phba->nvmet_support) { + /* FCPI/NVMEI path. Process Did */ + lpfc_prep_node_fc4type(vport, Did, fc4_type); + return; + } + /* NVMET path. NVMET only cares about NVMEI nodes. */ + list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) { + if (ndlp->nlp_type != NLP_NVME_INITIATOR || + ndlp->nlp_state != NLP_STE_UNMAPPED_NODE) + continue; + spin_lock_irq(shost->host_lock); + if (ndlp->nlp_DID == Did) + ndlp->nlp_flag &= ~NLP_NVMET_RECOV; + else + ndlp->nlp_flag |= NLP_NVMET_RECOV; + spin_unlock_irq(shost->host_lock); + } + } +} + static int lpfc_ns_rsp(struct lpfc_vport *vport, struct lpfc_dmabuf *mp, uint8_t fc4_type, uint32_t Size) { - struct lpfc_hba *phba = vport->phba; struct lpfc_sli_ct_request *Response = (struct lpfc_sli_ct_request *) mp->virt; - struct lpfc_nodelist *ndlp = NULL; struct lpfc_dmabuf *mlast, *next_mp; uint32_t *ctptr = (uint32_t *) & Response->un.gid.PortType; uint32_t Did, CTentry; int Cnt; struct list_head head; + struct Scsi_Host *shost = lpfc_shost_from_vport(vport); + struct lpfc_nodelist *ndlp = NULL; lpfc_set_disctmo(vport); vport->num_disc_nodes = 0; @@ -574,19 +608,7 @@ lpfc_ns_rsp(struct lpfc_vport *vport, struct lpfc_dmabuf *mp, uint8_t fc4_type, /* Get next DID from NameServer List */ CTentry = *ctptr++; Did = ((be32_to_cpu(CTentry)) & Mask_DID); - - ndlp = NULL; - - /* - * Check for rscn processing or not - * To conserve rpi's, filter out addresses for other - * vports on the same physical HBAs. - */ - if ((Did != vport->fc_myDID) && - ((lpfc_find_vport_by_did(phba, Did) == NULL) || - vport->cfg_peer_port_login)) - lpfc_prep_node_fc4type(vport, Did, fc4_type); - + lpfc_ns_rsp_audit_did(vport, Did, fc4_type); if (CTentry & (cpu_to_be32(SLI_CT_LAST_ENTRY))) goto nsout1; @@ -596,6 +618,22 @@ lpfc_ns_rsp(struct lpfc_vport *vport, struct lpfc_dmabuf *mp, uint8_t fc4_type, } + /* All GID_FT entries processed. If the driver is running in + * in target mode, put impacted nodes into recovery and drop + * the RPI to flush outstanding IO. + */ + if (vport->phba->nvmet_support) { + list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) { + if (!(ndlp->nlp_flag & NLP_NVMET_RECOV)) + continue; + lpfc_disc_state_machine(vport, ndlp, NULL, + NLP_EVT_DEVICE_RECOVERY); + spin_lock_irq(shost->host_lock); + ndlp->nlp_flag &= ~NLP_NVMET_RECOV; + spin_lock_irq(shost->host_lock); + } + } + nsout1: list_del(&head); return 0; diff --git a/drivers/scsi/lpfc/lpfc_debugfs.c b/drivers/scsi/lpfc/lpfc_debugfs.c index 913eed822cb8..fce549a91911 100644 --- a/drivers/scsi/lpfc/lpfc_debugfs.c +++ b/drivers/scsi/lpfc/lpfc_debugfs.c @@ -745,73 +745,102 @@ lpfc_debugfs_nvmestat_data(struct lpfc_vport *vport, char *buf, int size) { struct lpfc_hba *phba = vport->phba; struct lpfc_nvmet_tgtport *tgtp; + struct lpfc_nvmet_rcv_ctx *ctxp, *next_ctxp; int len = 0; + int cnt; if (phba->nvmet_support) { if (!phba->targetport) return len; tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "\nNVME Targetport Statistics\n"); - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "LS: Rcv %08x Drop %08x Abort %08x\n", atomic_read(&tgtp->rcv_ls_req_in), atomic_read(&tgtp->rcv_ls_req_drop), atomic_read(&tgtp->xmt_ls_abort)); if (atomic_read(&tgtp->rcv_ls_req_in) != atomic_read(&tgtp->rcv_ls_req_out)) { - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "Rcv LS: in %08x != out %08x\n", atomic_read(&tgtp->rcv_ls_req_in), atomic_read(&tgtp->rcv_ls_req_out)); } - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "LS: Xmt %08x Drop %08x Cmpl %08x Err %08x\n", atomic_read(&tgtp->xmt_ls_rsp), atomic_read(&tgtp->xmt_ls_drop), atomic_read(&tgtp->xmt_ls_rsp_cmpl), atomic_read(&tgtp->xmt_ls_rsp_error)); - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "FCP: Rcv %08x Drop %08x\n", atomic_read(&tgtp->rcv_fcp_cmd_in), atomic_read(&tgtp->rcv_fcp_cmd_drop)); if (atomic_read(&tgtp->rcv_fcp_cmd_in) != atomic_read(&tgtp->rcv_fcp_cmd_out)) { - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "Rcv FCP: in %08x != out %08x\n", atomic_read(&tgtp->rcv_fcp_cmd_in), atomic_read(&tgtp->rcv_fcp_cmd_out)); } - len += snprintf(buf+len, size-len, - "FCP Rsp: read %08x readrsp %08x write %08x rsp %08x\n", + len += snprintf(buf + len, size - len, + "FCP Rsp: read %08x readrsp %08x " + "write %08x rsp %08x\n", atomic_read(&tgtp->xmt_fcp_read), atomic_read(&tgtp->xmt_fcp_read_rsp), atomic_read(&tgtp->xmt_fcp_write), atomic_read(&tgtp->xmt_fcp_rsp)); - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "FCP Rsp: abort %08x drop %08x\n", atomic_read(&tgtp->xmt_fcp_abort), atomic_read(&tgtp->xmt_fcp_drop)); - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "FCP Rsp Cmpl: %08x err %08x drop %08x\n", atomic_read(&tgtp->xmt_fcp_rsp_cmpl), atomic_read(&tgtp->xmt_fcp_rsp_error), atomic_read(&tgtp->xmt_fcp_rsp_drop)); - len += snprintf(buf+len, size-len, + len += snprintf(buf + len, size - len, "ABORT: Xmt %08x Err %08x Cmpl %08x", atomic_read(&tgtp->xmt_abort_rsp), atomic_read(&tgtp->xmt_abort_rsp_error), atomic_read(&tgtp->xmt_abort_cmpl)); - len += snprintf(buf+len, size-len, "\n"); + len += snprintf(buf + len, size - len, "\n"); + + cnt = 0; + spin_lock(&phba->sli4_hba.abts_nvme_buf_list_lock); + list_for_each_entry_safe(ctxp, next_ctxp, + &phba->sli4_hba.lpfc_abts_nvmet_ctx_list, + list) { + cnt++; + } + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + if (cnt) { + len += snprintf(buf + len, size - len, + "ABORT: %d ctx entries\n", cnt); + spin_lock(&phba->sli4_hba.abts_nvme_buf_list_lock); + list_for_each_entry_safe(ctxp, next_ctxp, + &phba->sli4_hba.lpfc_abts_nvmet_ctx_list, + list) { + if (len >= (size - LPFC_DEBUG_OUT_LINE_SZ)) + break; + len += snprintf(buf + len, size - len, + "Entry: oxid %x state %x " + "flag %x\n", + ctxp->oxid, ctxp->state, + ctxp->flag); + } + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + } } else { if (!(phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME)) return len; @@ -3128,8 +3157,6 @@ __lpfc_idiag_print_rqpair(struct lpfc_queue *qp, struct lpfc_queue *datqp, datqp->queue_id, datqp->entry_count, datqp->entry_size, datqp->host_index, datqp->hba_index); - len += snprintf(pbuffer + len, LPFC_QUE_INFO_GET_BUF_SIZE - len, "\n"); - return len; } @@ -5700,10 +5727,8 @@ lpfc_debugfs_terminate(struct lpfc_vport *vport) #ifdef CONFIG_SCSI_LPFC_DEBUG_FS struct lpfc_hba *phba = vport->phba; - if (vport->disc_trc) { - kfree(vport->disc_trc); - vport->disc_trc = NULL; - } + kfree(vport->disc_trc); + vport->disc_trc = NULL; debugfs_remove(vport->debug_disc_trc); /* discovery_trace */ vport->debug_disc_trc = NULL; @@ -5770,10 +5795,8 @@ lpfc_debugfs_terminate(struct lpfc_vport *vport) debugfs_remove(phba->debug_readRef); /* readRef */ phba->debug_readRef = NULL; - if (phba->slow_ring_trc) { - kfree(phba->slow_ring_trc); - phba->slow_ring_trc = NULL; - } + kfree(phba->slow_ring_trc); + phba->slow_ring_trc = NULL; /* slow_ring_trace */ debugfs_remove(phba->debug_slow_ring_trc); diff --git a/drivers/scsi/lpfc/lpfc_disc.h b/drivers/scsi/lpfc/lpfc_disc.h index f4ff99d95db3..9d5a379f4b15 100644 --- a/drivers/scsi/lpfc/lpfc_disc.h +++ b/drivers/scsi/lpfc/lpfc_disc.h @@ -157,6 +157,7 @@ struct lpfc_node_rrq { #define NLP_LOGO_SND 0x00000100 /* sent LOGO request for this entry */ #define NLP_RNID_SND 0x00000400 /* sent RNID request for this entry */ #define NLP_ELS_SND_MASK 0x000007e0 /* sent ELS request for this entry */ +#define NLP_NVMET_RECOV 0x00001000 /* NVMET auditing node for recovery. */ #define NLP_DEFER_RM 0x00010000 /* Remove this ndlp if no longer used */ #define NLP_DELAY_TMO 0x00020000 /* delay timeout is running for node */ #define NLP_NPR_2B_DISC 0x00040000 /* node is included in num_disc_nodes */ diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index a5ca37e45fb6..67827e397431 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -603,9 +603,11 @@ lpfc_check_clean_addr_bit(struct lpfc_vport *vport, memcmp(&vport->fabric_portname, &sp->portName, sizeof(struct lpfc_name)) || memcmp(&vport->fabric_nodename, &sp->nodeName, - sizeof(struct lpfc_name))) + sizeof(struct lpfc_name)) || + (vport->vport_flag & FAWWPN_PARAM_CHG)) { fabric_param_changed = 1; - + vport->vport_flag &= ~FAWWPN_PARAM_CHG; + } /* * Word 1 Bit 31 in common service parameter is overloaded. * Word 1 Bit 31 in FLOGI request is multiple NPort request @@ -895,10 +897,9 @@ lpfc_cmpl_els_flogi_nport(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, * Cannot find existing Fabric ndlp, so allocate a * new one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, PT2PT_RemoteID); if (!ndlp) goto fail; - lpfc_nlp_init(vport, ndlp, PT2PT_RemoteID); } else if (!NLP_CHK_NODE_ACT(ndlp)) { ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_UNUSED_NODE); @@ -1364,7 +1365,6 @@ lpfc_els_abort_flogi(struct lpfc_hba *phba) int lpfc_initial_flogi(struct lpfc_vport *vport) { - struct lpfc_hba *phba = vport->phba; struct lpfc_nodelist *ndlp; vport->port_state = LPFC_FLOGI; @@ -1374,10 +1374,9 @@ lpfc_initial_flogi(struct lpfc_vport *vport) ndlp = lpfc_findnode_did(vport, Fabric_DID); if (!ndlp) { /* Cannot find existing Fabric ndlp, so allocate a new one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, Fabric_DID); if (!ndlp) return 0; - lpfc_nlp_init(vport, ndlp, Fabric_DID); /* Set the node type */ ndlp->nlp_type |= NLP_FABRIC; /* Put ndlp onto node list */ @@ -1418,17 +1417,15 @@ lpfc_initial_flogi(struct lpfc_vport *vport) int lpfc_initial_fdisc(struct lpfc_vport *vport) { - struct lpfc_hba *phba = vport->phba; struct lpfc_nodelist *ndlp; /* First look for the Fabric ndlp */ ndlp = lpfc_findnode_did(vport, Fabric_DID); if (!ndlp) { /* Cannot find existing Fabric ndlp, so allocate a new one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, Fabric_DID); if (!ndlp) return 0; - lpfc_nlp_init(vport, ndlp, Fabric_DID); /* Put ndlp onto node list */ lpfc_enqueue_node(vport, ndlp); } else if (!NLP_CHK_NODE_ACT(ndlp)) { @@ -1564,14 +1561,13 @@ lpfc_plogi_confirm_nport(struct lpfc_hba *phba, uint32_t *prsp, phba->active_rrq_pool); return ndlp; } - new_ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_ATOMIC); + new_ndlp = lpfc_nlp_init(vport, ndlp->nlp_DID); if (!new_ndlp) { if (active_rrqs_xri_bitmap) mempool_free(active_rrqs_xri_bitmap, phba->active_rrq_pool); return ndlp; } - lpfc_nlp_init(vport, new_ndlp, ndlp->nlp_DID); } else if (!NLP_CHK_NODE_ACT(new_ndlp)) { rc = memcmp(&ndlp->nlp_portname, name, sizeof(struct lpfc_name)); @@ -2845,10 +2841,9 @@ lpfc_issue_els_scr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry) ndlp = lpfc_findnode_did(vport, nportid); if (!ndlp) { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, nportid); if (!ndlp) return 1; - lpfc_nlp_init(vport, ndlp, nportid); lpfc_enqueue_node(vport, ndlp); } else if (!NLP_CHK_NODE_ACT(ndlp)) { ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_UNUSED_NODE); @@ -2938,10 +2933,9 @@ lpfc_issue_els_farpr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry) ndlp = lpfc_findnode_did(vport, nportid); if (!ndlp) { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, nportid); if (!ndlp) return 1; - lpfc_nlp_init(vport, ndlp, nportid); lpfc_enqueue_node(vport, ndlp); } else if (!NLP_CHK_NODE_ACT(ndlp)) { ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_UNUSED_NODE); @@ -4403,7 +4397,7 @@ lpfc_els_rsp_prli_acc(struct lpfc_vport *vport, struct lpfc_iocbq *oldiocb, pcmd = (uint8_t *) (((struct lpfc_dmabuf *) elsiocb->context2)->virt); memset(pcmd, 0, cmdsize); - *((uint32_t *) (pcmd)) = (ELS_CMD_ACC | (ELS_CMD_PRLI & ~ELS_RSP_MASK)); + *((uint32_t *)(pcmd)) = elsrspcmd; pcmd += sizeof(uint32_t); /* For PRLI, remainder of payload is PRLI parameter page */ @@ -5867,8 +5861,11 @@ lpfc_rscn_recovery_check(struct lpfc_vport *vport) (ndlp->nlp_state == NLP_STE_UNUSED_NODE) || !lpfc_rscn_payload_check(vport, ndlp->nlp_DID)) continue; + + /* NVME Target mode does not do RSCN Recovery. */ if (vport->phba->nvmet_support) continue; + lpfc_disc_state_machine(vport, ndlp, NULL, NLP_EVT_DEVICE_RECOVERY); lpfc_cancel_retry_delay_tmo(vport, ndlp); @@ -6133,7 +6130,6 @@ int lpfc_els_handle_rscn(struct lpfc_vport *vport) { struct lpfc_nodelist *ndlp; - struct lpfc_hba *phba = vport->phba; /* Ignore RSCN if the port is being torn down. */ if (vport->load_flag & FC_UNLOADING) { @@ -6157,22 +6153,16 @@ lpfc_els_handle_rscn(struct lpfc_vport *vport) ndlp = lpfc_findnode_did(vport, NameServer_DID); if (ndlp && NLP_CHK_NODE_ACT(ndlp) && ndlp->nlp_state == NLP_STE_UNMAPPED_NODE) { - /* Good ndlp, issue CT Request to NameServer */ + /* Good ndlp, issue CT Request to NameServer. Need to + * know how many gidfts were issued. If none, then just + * flush the RSCN. Otherwise, the outstanding requests + * need to complete. + */ vport->gidft_inp = 0; - if (lpfc_issue_gidft(vport) == 0) - /* Wait for NameServer query cmpl before we can - * continue - */ + if (lpfc_issue_gidft(vport) > 0) return 1; } else { - /* If login to NameServer does not exist, issue one */ - /* Good status, issue PLOGI to NameServer */ - ndlp = lpfc_findnode_did(vport, NameServer_DID); - if (ndlp && NLP_CHK_NODE_ACT(ndlp)) - /* Wait for NameServer login cmpl before we can - continue */ - return 1; - + /* Nameserver login in question. Revalidate. */ if (ndlp) { ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_PLOGI_ISSUE); @@ -6182,12 +6172,11 @@ lpfc_els_handle_rscn(struct lpfc_vport *vport) } ndlp->nlp_prev_state = NLP_STE_UNUSED_NODE; } else { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, NameServer_DID); if (!ndlp) { lpfc_els_flush_rscn(vport); return 0; } - lpfc_nlp_init(vport, ndlp, NameServer_DID); ndlp->nlp_prev_state = ndlp->nlp_state; lpfc_nlp_set_state(vport, ndlp, NLP_STE_PLOGI_ISSUE); } @@ -7746,11 +7735,9 @@ lpfc_els_unsol_buffer(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, ndlp = lpfc_findnode_did(vport, did); if (!ndlp) { /* Cannot find existing Fabric ndlp, so allocate a new one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, did); if (!ndlp) goto dropit; - - lpfc_nlp_init(vport, ndlp, did); lpfc_nlp_set_state(vport, ndlp, NLP_STE_NPR_NODE); newnode = 1; if ((did & Fabric_DID_MASK) == Fabric_DID_MASK) @@ -8193,7 +8180,6 @@ lpfc_els_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, static void lpfc_start_fdmi(struct lpfc_vport *vport) { - struct lpfc_hba *phba = vport->phba; struct lpfc_nodelist *ndlp; /* If this is the first time, allocate an ndlp and initialize @@ -8202,9 +8188,8 @@ lpfc_start_fdmi(struct lpfc_vport *vport) */ ndlp = lpfc_findnode_did(vport, FDMI_DID); if (!ndlp) { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, FDMI_DID); if (ndlp) { - lpfc_nlp_init(vport, ndlp, FDMI_DID); ndlp->nlp_type |= NLP_FABRIC; } else { return; @@ -8257,7 +8242,7 @@ lpfc_do_scr_ns_plogi(struct lpfc_hba *phba, struct lpfc_vport *vport) ndlp = lpfc_findnode_did(vport, NameServer_DID); if (!ndlp) { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, NameServer_DID); if (!ndlp) { if (phba->fc_topology == LPFC_TOPOLOGY_LOOP) { lpfc_disc_start(vport); @@ -8268,7 +8253,6 @@ lpfc_do_scr_ns_plogi(struct lpfc_hba *phba, struct lpfc_vport *vport) "0251 NameServer login: no memory\n"); return; } - lpfc_nlp_init(vport, ndlp, NameServer_DID); } else if (!NLP_CHK_NODE_ACT(ndlp)) { ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_UNUSED_NODE); if (!ndlp) { @@ -8771,7 +8755,7 @@ lpfc_issue_els_fdisc(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, pcmd += sizeof(uint32_t); /* Node Name */ pcmd += sizeof(uint32_t); /* Node Name */ memcpy(pcmd, &vport->fc_nodename, 8); - + memset(sp->un.vendorVersion, 0, sizeof(sp->un.vendorVersion)); lpfc_set_disctmo(vport); phba->fc_stat.elsXmitFDISC++; diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c index 180b072beef6..0482c5580331 100644 --- a/drivers/scsi/lpfc/lpfc_hbadisc.c +++ b/drivers/scsi/lpfc/lpfc_hbadisc.c @@ -3002,6 +3002,7 @@ lpfc_mbx_cmpl_read_sparam(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) MAILBOX_t *mb = &pmb->u.mb; struct lpfc_dmabuf *mp = (struct lpfc_dmabuf *) pmb->context1; struct lpfc_vport *vport = pmb->vport; + struct Scsi_Host *shost = lpfc_shost_from_vport(vport); struct serv_parm *sp = &vport->fc_sparam; uint32_t ed_tov; @@ -3031,6 +3032,7 @@ lpfc_mbx_cmpl_read_sparam(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) } lpfc_update_vport_wwn(vport); + fc_host_port_name(shost) = wwn_to_u64(vport->fc_portname.u.wwn); if (vport->port_type == LPFC_PHYSICAL_PORT) { memcpy(&phba->wwnn, &vport->fc_nodename, sizeof(phba->wwnn)); memcpy(&phba->wwpn, &vport->fc_portname, sizeof(phba->wwnn)); @@ -3309,6 +3311,7 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) struct lpfc_sli_ring *pring; MAILBOX_t *mb = &pmb->u.mb; struct lpfc_dmabuf *mp = (struct lpfc_dmabuf *) (pmb->context1); + uint8_t attn_type; /* Unblock ELS traffic */ pring = lpfc_phba_elsring(phba); @@ -3325,6 +3328,7 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) } la = (struct lpfc_mbx_read_top *) &pmb->u.mb.un.varReadTop; + attn_type = bf_get(lpfc_mbx_read_top_att_type, la); memcpy(&phba->alpa_map[0], mp->virt, 128); @@ -3337,7 +3341,7 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) if (phba->fc_eventTag <= la->eventTag) { phba->fc_stat.LinkMultiEvent++; - if (bf_get(lpfc_mbx_read_top_att_type, la) == LPFC_ATT_LINK_UP) + if (attn_type == LPFC_ATT_LINK_UP) if (phba->fc_eventTag != 0) lpfc_linkdown(phba); } @@ -3353,7 +3357,7 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) } phba->link_events++; - if ((bf_get(lpfc_mbx_read_top_att_type, la) == LPFC_ATT_LINK_UP) && + if ((attn_type == LPFC_ATT_LINK_UP) && !(phba->sli.sli_flag & LPFC_MENLO_MAINT)) { phba->fc_stat.LinkUp++; if (phba->link_flag & LS_LOOPBACK_MODE) { @@ -3379,8 +3383,8 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) phba->wait_4_mlo_maint_flg); } lpfc_mbx_process_link_up(phba, la); - } else if (bf_get(lpfc_mbx_read_top_att_type, la) == - LPFC_ATT_LINK_DOWN) { + } else if (attn_type == LPFC_ATT_LINK_DOWN || + attn_type == LPFC_ATT_UNEXP_WWPN) { phba->fc_stat.LinkDown++; if (phba->link_flag & LS_LOOPBACK_MODE) lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, @@ -3389,6 +3393,14 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) "Data: x%x x%x x%x\n", la->eventTag, phba->fc_eventTag, phba->pport->port_state, vport->fc_flag); + else if (attn_type == LPFC_ATT_UNEXP_WWPN) + lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, + "1313 Link Down UNEXP WWPN Event x%x received " + "Data: x%x x%x x%x x%x x%x\n", + la->eventTag, phba->fc_eventTag, + phba->pport->port_state, vport->fc_flag, + bf_get(lpfc_mbx_read_top_mm, la), + bf_get(lpfc_mbx_read_top_fa, la)); else lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, "1305 Link Down Event x%x received " @@ -3399,8 +3411,8 @@ lpfc_mbx_cmpl_read_topology(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) bf_get(lpfc_mbx_read_top_fa, la)); lpfc_mbx_issue_link_down(phba); } - if ((phba->sli.sli_flag & LPFC_MENLO_MAINT) && - ((bf_get(lpfc_mbx_read_top_att_type, la) == LPFC_ATT_LINK_UP))) { + if (phba->sli.sli_flag & LPFC_MENLO_MAINT && + attn_type == LPFC_ATT_LINK_UP) { if (phba->link_state != LPFC_LINK_DOWN) { phba->fc_stat.LinkDown++; lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, @@ -4136,7 +4148,6 @@ lpfc_nlp_state_cleanup(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, int old_state, int new_state) { struct Scsi_Host *shost = lpfc_shost_from_vport(vport); - struct lpfc_hba *phba = vport->phba; if (new_state == NLP_STE_UNMAPPED_NODE) { ndlp->nlp_flag &= ~NLP_NODEV_REMOVE; @@ -4155,14 +4166,14 @@ lpfc_nlp_state_cleanup(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, lpfc_unregister_remote_port(ndlp); } - /* Notify the NVME transport of this rport's loss */ - if (((phba->cfg_enable_fc4_type == LPFC_ENABLE_BOTH) || - (phba->cfg_enable_fc4_type == LPFC_ENABLE_NVME)) && - (vport->phba->nvmet_support == 0) && - ((ndlp->nlp_fc4_type & NLP_FC4_NVME) || - (ndlp->nlp_DID == Fabric_DID))) { + /* Notify the NVME transport of this rport's loss on the + * Initiator. For NVME Target, should upcall transport + * in the else clause when API available. + */ + if (ndlp->nlp_fc4_type & NLP_FC4_NVME) { vport->phba->nport_event_cnt++; - lpfc_nvme_unregister_port(vport, ndlp); + if (vport->phba->nvmet_support == 0) + lpfc_nvme_unregister_port(vport, ndlp); } } @@ -4368,10 +4379,17 @@ lpfc_enable_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, uint32_t did; unsigned long flags; unsigned long *active_rrqs_xri_bitmap = NULL; + int rpi = LPFC_RPI_ALLOC_ERROR; if (!ndlp) return NULL; + if (phba->sli_rev == LPFC_SLI_REV4) { + rpi = lpfc_sli4_alloc_rpi(vport->phba); + if (rpi == LPFC_RPI_ALLOC_ERROR) + return NULL; + } + spin_lock_irqsave(&phba->ndlp_lock, flags); /* The ndlp should not be in memory free mode */ if (NLP_CHK_FREE_REQ(ndlp)) { @@ -4381,7 +4399,7 @@ lpfc_enable_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, "usgmap:x%x refcnt:%d\n", (void *)ndlp, ndlp->nlp_usg_map, kref_read(&ndlp->kref)); - return NULL; + goto free_rpi; } /* The ndlp should not already be in active mode */ if (NLP_CHK_NODE_ACT(ndlp)) { @@ -4391,7 +4409,7 @@ lpfc_enable_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, "usgmap:x%x refcnt:%d\n", (void *)ndlp, ndlp->nlp_usg_map, kref_read(&ndlp->kref)); - return NULL; + goto free_rpi; } /* Keep the original DID */ @@ -4409,7 +4427,7 @@ lpfc_enable_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, spin_unlock_irqrestore(&phba->ndlp_lock, flags); if (vport->phba->sli_rev == LPFC_SLI_REV4) { - ndlp->nlp_rpi = lpfc_sli4_alloc_rpi(vport->phba); + ndlp->nlp_rpi = rpi; lpfc_printf_vlog(vport, KERN_INFO, LOG_NODE, "0008 rpi:%x DID:%x flg:%x refcnt:%d " "map:%x %p\n", ndlp->nlp_rpi, ndlp->nlp_DID, @@ -4426,6 +4444,11 @@ lpfc_enable_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, "node enable: did:x%x", ndlp->nlp_DID, 0, 0); return ndlp; + +free_rpi: + if (phba->sli_rev == LPFC_SLI_REV4) + lpfc_sli4_free_rpi(vport->phba, rpi); + return NULL; } void @@ -5104,65 +5127,82 @@ lpfc_setup_disc_node(struct lpfc_vport *vport, uint32_t did) ndlp = lpfc_findnode_did(vport, did); if (!ndlp) { + if (vport->phba->nvmet_support) + return NULL; if ((vport->fc_flag & FC_RSCN_MODE) != 0 && lpfc_rscn_payload_check(vport, did) == 0) return NULL; - ndlp = (struct lpfc_nodelist *) - mempool_alloc(vport->phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, did); if (!ndlp) return NULL; - lpfc_nlp_init(vport, ndlp, did); lpfc_nlp_set_state(vport, ndlp, NLP_STE_NPR_NODE); - if (vport->phba->nvmet_support) - return ndlp; spin_lock_irq(shost->host_lock); ndlp->nlp_flag |= NLP_NPR_2B_DISC; spin_unlock_irq(shost->host_lock); return ndlp; } else if (!NLP_CHK_NODE_ACT(ndlp)) { + if (vport->phba->nvmet_support) + return NULL; ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_NPR_NODE); if (!ndlp) return NULL; - if (vport->phba->nvmet_support) - return ndlp; spin_lock_irq(shost->host_lock); ndlp->nlp_flag |= NLP_NPR_2B_DISC; spin_unlock_irq(shost->host_lock); return ndlp; } + /* The NVME Target does not want to actively manage an rport. + * The goal is to allow the target to reset its state and clear + * pending IO in preparation for the initiator to recover. + */ if ((vport->fc_flag & FC_RSCN_MODE) && !(vport->fc_flag & FC_NDISC_ACTIVE)) { if (lpfc_rscn_payload_check(vport, did)) { + + /* Since this node is marked for discovery, + * delay timeout is not needed. + */ + lpfc_cancel_retry_delay_tmo(vport, ndlp); + + /* NVME Target mode waits until rport is known to be + * impacted by the RSCN before it transitions. No + * active management - just go to NPR provided the + * node had a valid login. + */ + if (vport->phba->nvmet_support) + return ndlp; + /* If we've already received a PLOGI from this NPort * we don't need to try to discover it again. */ if (ndlp->nlp_flag & NLP_RCV_PLOGI) return NULL; - /* Since this node is marked for discovery, - * delay timeout is not needed. - */ - lpfc_cancel_retry_delay_tmo(vport, ndlp); - if (vport->phba->nvmet_support) - return ndlp; spin_lock_irq(shost->host_lock); ndlp->nlp_flag |= NLP_NPR_2B_DISC; spin_unlock_irq(shost->host_lock); } else ndlp = NULL; } else { - /* If we've already received a PLOGI from this NPort, - * or we are already in the process of discovery on it, - * we don't need to try to discover it again. + /* If the initiator received a PLOGI from this NPort or if the + * initiator is already in the process of discovery on it, + * there's no need to try to discover it again. */ if (ndlp->nlp_state == NLP_STE_ADISC_ISSUE || ndlp->nlp_state == NLP_STE_PLOGI_ISSUE || - ndlp->nlp_flag & NLP_RCV_PLOGI) + (!vport->phba->nvmet_support && + ndlp->nlp_flag & NLP_RCV_PLOGI)) return NULL; - lpfc_nlp_set_state(vport, ndlp, NLP_STE_NPR_NODE); + if (vport->phba->nvmet_support) return ndlp; + + /* Moving to NPR state clears unsolicited flags and + * allows for rediscovery + */ + lpfc_nlp_set_state(vport, ndlp, NLP_STE_NPR_NODE); + spin_lock_irq(shost->host_lock); ndlp->nlp_flag |= NLP_NPR_2B_DISC; spin_unlock_irq(shost->host_lock); @@ -5887,16 +5927,31 @@ lpfc_find_vport_by_vpid(struct lpfc_hba *phba, uint16_t vpi) return NULL; } -void -lpfc_nlp_init(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, - uint32_t did) +struct lpfc_nodelist * +lpfc_nlp_init(struct lpfc_vport *vport, uint32_t did) { + struct lpfc_nodelist *ndlp; + int rpi = LPFC_RPI_ALLOC_ERROR; + + if (vport->phba->sli_rev == LPFC_SLI_REV4) { + rpi = lpfc_sli4_alloc_rpi(vport->phba); + if (rpi == LPFC_RPI_ALLOC_ERROR) + return NULL; + } + + ndlp = mempool_alloc(vport->phba->nlp_mem_pool, GFP_KERNEL); + if (!ndlp) { + if (vport->phba->sli_rev == LPFC_SLI_REV4) + lpfc_sli4_free_rpi(vport->phba, rpi); + return NULL; + } + memset(ndlp, 0, sizeof (struct lpfc_nodelist)); lpfc_initialize_node(vport, ndlp, did); INIT_LIST_HEAD(&ndlp->nlp_listp); if (vport->phba->sli_rev == LPFC_SLI_REV4) { - ndlp->nlp_rpi = lpfc_sli4_alloc_rpi(vport->phba); + ndlp->nlp_rpi = rpi; lpfc_printf_vlog(vport, KERN_INFO, LOG_NODE, "0007 rpi:%x DID:%x flg:%x refcnt:%d " "map:%x %p\n", ndlp->nlp_rpi, ndlp->nlp_DID, @@ -5918,7 +5973,7 @@ lpfc_nlp_init(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, "node init: did:x%x", ndlp->nlp_DID, 0, 0); - return; + return ndlp; } /* This routine releases all resources associated with a specifc NPort's ndlp diff --git a/drivers/scsi/lpfc/lpfc_hw.h b/drivers/scsi/lpfc/lpfc_hw.h index 15ca21484150..26a5647e057e 100644 --- a/drivers/scsi/lpfc/lpfc_hw.h +++ b/drivers/scsi/lpfc/lpfc_hw.h @@ -509,6 +509,8 @@ struct class_parms { uint8_t word3Reserved2; /* Fc Word 3, bit 0: 7 */ }; +#define FAPWWN_KEY_VENDOR 0x42524344 /*valid vendor version fawwpn key*/ + struct serv_parm { /* Structure is in Big Endian format */ struct csp cmn; struct lpfc_name portName; @@ -2885,6 +2887,7 @@ struct lpfc_mbx_read_top { #define LPFC_ATT_RESERVED 0x00 /* Reserved - attType */ #define LPFC_ATT_LINK_UP 0x01 /* Link is up */ #define LPFC_ATT_LINK_DOWN 0x02 /* Link is down */ +#define LPFC_ATT_UNEXP_WWPN 0x06 /* Link is down Unexpected WWWPN */ uint32_t word3; #define lpfc_mbx_read_top_alpa_granted_SHIFT 24 #define lpfc_mbx_read_top_alpa_granted_MASK 0x000000FF diff --git a/drivers/scsi/lpfc/lpfc_hw4.h b/drivers/scsi/lpfc/lpfc_hw4.h index 15277705cb6b..1d12f2be36bc 100644 --- a/drivers/scsi/lpfc/lpfc_hw4.h +++ b/drivers/scsi/lpfc/lpfc_hw4.h @@ -2720,6 +2720,9 @@ struct lpfc_mbx_request_features { #define lpfc_mbx_rq_ftr_rq_ifip_SHIFT 7 #define lpfc_mbx_rq_ftr_rq_ifip_MASK 0x00000001 #define lpfc_mbx_rq_ftr_rq_ifip_WORD word2 +#define lpfc_mbx_rq_ftr_rq_iaar_SHIFT 9 +#define lpfc_mbx_rq_ftr_rq_iaar_MASK 0x00000001 +#define lpfc_mbx_rq_ftr_rq_iaar_WORD word2 #define lpfc_mbx_rq_ftr_rq_perfh_SHIFT 11 #define lpfc_mbx_rq_ftr_rq_perfh_MASK 0x00000001 #define lpfc_mbx_rq_ftr_rq_perfh_WORD word2 @@ -3853,6 +3856,7 @@ struct lpfc_acqe_fc_la { #define LPFC_FC_LA_TYPE_NO_HARD_ALPA 0x3 #define LPFC_FC_LA_TYPE_MDS_LINK_DOWN 0x4 #define LPFC_FC_LA_TYPE_MDS_LOOPBACK 0x5 +#define LPFC_FC_LA_TYPE_UNEXP_WWPN 0x6 #define lpfc_acqe_fc_la_port_type_SHIFT 6 #define lpfc_acqe_fc_la_port_type_MASK 0x00000003 #define lpfc_acqe_fc_la_port_type_WORD word0 diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index 6cc561b04211..90ae354a9c45 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -42,6 +42,10 @@ #include #include #include +#include +#include + +#include #include "lpfc_hw4.h" #include "lpfc_hw.h" @@ -52,6 +56,7 @@ #include "lpfc.h" #include "lpfc_scsi.h" #include "lpfc_nvme.h" +#include "lpfc_nvmet.h" #include "lpfc_logmsg.h" #include "lpfc_crtn.h" #include "lpfc_vport.h" @@ -335,6 +340,9 @@ lpfc_dump_wakeup_param_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmboxq) void lpfc_update_vport_wwn(struct lpfc_vport *vport) { + uint8_t vvvl = vport->fc_sparam.cmn.valid_vendor_ver_level; + u32 *fawwpn_key = (u32 *)&vport->fc_sparam.un.vendorVersion[0]; + /* If the soft name exists then update it using the service params */ if (vport->phba->cfg_soft_wwnn) u64_to_wwn(vport->phba->cfg_soft_wwnn, @@ -354,9 +362,25 @@ lpfc_update_vport_wwn(struct lpfc_vport *vport) memcpy(&vport->fc_sparam.nodeName, &vport->fc_nodename, sizeof(struct lpfc_name)); - if (vport->fc_portname.u.wwn[0] == 0 || vport->phba->cfg_soft_wwpn) + /* + * If the port name has changed, then set the Param changes flag + * to unreg the login + */ + if (vport->fc_portname.u.wwn[0] != 0 && + memcmp(&vport->fc_portname, &vport->fc_sparam.portName, + sizeof(struct lpfc_name))) + vport->vport_flag |= FAWWPN_PARAM_CHG; + + if (vport->fc_portname.u.wwn[0] == 0 || + vport->phba->cfg_soft_wwpn || + (vvvl == 1 && cpu_to_be32(*fawwpn_key) == FAPWWN_KEY_VENDOR) || + vport->vport_flag & FAWWPN_SET) { memcpy(&vport->fc_portname, &vport->fc_sparam.portName, sizeof(struct lpfc_name)); + vport->vport_flag &= ~FAWWPN_SET; + if (vvvl == 1 && cpu_to_be32(*fawwpn_key) == FAPWWN_KEY_VENDOR) + vport->vport_flag |= FAWWPN_SET; + } else memcpy(&vport->fc_sparam.portName, &vport->fc_portname, sizeof(struct lpfc_name)); @@ -1003,8 +1027,10 @@ static int lpfc_hba_down_post_s4(struct lpfc_hba *phba) { struct lpfc_scsi_buf *psb, *psb_next; + struct lpfc_nvmet_rcv_ctx *ctxp, *ctxp_next; LIST_HEAD(aborts); LIST_HEAD(nvme_aborts); + LIST_HEAD(nvmet_aborts); unsigned long iflag = 0; struct lpfc_sglq *sglq_entry = NULL; @@ -1027,16 +1053,10 @@ lpfc_hba_down_post_s4(struct lpfc_hba *phba) list_for_each_entry(sglq_entry, &phba->sli4_hba.lpfc_abts_els_sgl_list, list) sglq_entry->state = SGL_FREED; - list_for_each_entry(sglq_entry, - &phba->sli4_hba.lpfc_abts_nvmet_sgl_list, list) - sglq_entry->state = SGL_FREED; list_splice_init(&phba->sli4_hba.lpfc_abts_els_sgl_list, &phba->sli4_hba.lpfc_els_sgl_list); - if (phba->sli4_hba.nvme_wq) - list_splice_init(&phba->sli4_hba.lpfc_abts_nvmet_sgl_list, - &phba->sli4_hba.lpfc_nvmet_sgl_list); spin_unlock(&phba->sli4_hba.sgl_list_lock); /* abts_scsi_buf_list_lock required because worker thread uses this @@ -1053,6 +1073,8 @@ lpfc_hba_down_post_s4(struct lpfc_hba *phba) spin_lock(&phba->sli4_hba.abts_nvme_buf_list_lock); list_splice_init(&phba->sli4_hba.lpfc_abts_nvme_buf_list, &nvme_aborts); + list_splice_init(&phba->sli4_hba.lpfc_abts_nvmet_ctx_list, + &nvmet_aborts); spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); } @@ -1066,13 +1088,20 @@ lpfc_hba_down_post_s4(struct lpfc_hba *phba) list_splice(&aborts, &phba->lpfc_scsi_buf_list_put); spin_unlock_irqrestore(&phba->scsi_buf_list_put_lock, iflag); - list_for_each_entry_safe(psb, psb_next, &nvme_aborts, list) { - psb->pCmd = NULL; - psb->status = IOSTAT_SUCCESS; + if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) { + list_for_each_entry_safe(psb, psb_next, &nvme_aborts, list) { + psb->pCmd = NULL; + psb->status = IOSTAT_SUCCESS; + } + spin_lock_irqsave(&phba->nvme_buf_list_put_lock, iflag); + list_splice(&nvme_aborts, &phba->lpfc_nvme_buf_list_put); + spin_unlock_irqrestore(&phba->nvme_buf_list_put_lock, iflag); + + list_for_each_entry_safe(ctxp, ctxp_next, &nvmet_aborts, list) { + ctxp->flag &= ~(LPFC_NVMET_XBUSY | LPFC_NVMET_ABORT_OP); + lpfc_nvmet_rq_post(phba, ctxp, &ctxp->rqb_buffer->hbuf); + } } - spin_lock_irqsave(&phba->nvme_buf_list_put_lock, iflag); - list_splice(&nvme_aborts, &phba->lpfc_nvme_buf_list_put); - spin_unlock_irqrestore(&phba->nvme_buf_list_put_lock, iflag); lpfc_sli4_free_sp_events(phba); return 0; @@ -2874,34 +2903,38 @@ lpfc_sli4_node_prep(struct lpfc_hba *phba) { struct lpfc_nodelist *ndlp, *next_ndlp; struct lpfc_vport **vports; - int i; + int i, rpi; + unsigned long flags; if (phba->sli_rev != LPFC_SLI_REV4) return; vports = lpfc_create_vport_work_array(phba); - if (vports != NULL) { - for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) { - if (vports[i]->load_flag & FC_UNLOADING) - continue; + if (vports == NULL) + return; - list_for_each_entry_safe(ndlp, next_ndlp, - &vports[i]->fc_nodes, - nlp_listp) { - if (NLP_CHK_NODE_ACT(ndlp)) { - ndlp->nlp_rpi = - lpfc_sli4_alloc_rpi(phba); - lpfc_printf_vlog(ndlp->vport, KERN_INFO, - LOG_NODE, - "0009 rpi:%x DID:%x " - "flg:%x map:%x %p\n", - ndlp->nlp_rpi, - ndlp->nlp_DID, - ndlp->nlp_flag, - ndlp->nlp_usg_map, - ndlp); - } + for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) { + if (vports[i]->load_flag & FC_UNLOADING) + continue; + + list_for_each_entry_safe(ndlp, next_ndlp, + &vports[i]->fc_nodes, + nlp_listp) { + if (!NLP_CHK_NODE_ACT(ndlp)) + continue; + rpi = lpfc_sli4_alloc_rpi(phba); + if (rpi == LPFC_RPI_ALLOC_ERROR) { + spin_lock_irqsave(&phba->ndlp_lock, flags); + NLP_CLR_NODE_ACT(ndlp); + spin_unlock_irqrestore(&phba->ndlp_lock, flags); + continue; } + ndlp->nlp_rpi = rpi; + lpfc_printf_vlog(ndlp->vport, KERN_INFO, LOG_NODE, + "0009 rpi:%x DID:%x " + "flg:%x map:%x %p\n", ndlp->nlp_rpi, + ndlp->nlp_DID, ndlp->nlp_flag, + ndlp->nlp_usg_map, ndlp); } } lpfc_destroy_vport_work_array(phba, vports); @@ -3508,6 +3541,12 @@ lpfc_sli4_scsi_sgl_update(struct lpfc_hba *phba) spin_unlock(&phba->scsi_buf_list_put_lock); spin_unlock_irq(&phba->scsi_buf_list_get_lock); + lpfc_printf_log(phba, KERN_INFO, LOG_SLI, + "6060 Current allocated SCSI xri-sgl count:%d, " + "maximum SCSI xri count:%d (split:%d)\n", + phba->sli4_hba.scsi_xri_cnt, + phba->sli4_hba.scsi_xri_max, phba->cfg_xri_split); + if (phba->sli4_hba.scsi_xri_cnt > phba->sli4_hba.scsi_xri_max) { /* max scsi xri shrinked below the allocated scsi buffers */ scsi_xri_cnt = phba->sli4_hba.scsi_xri_cnt - @@ -4508,9 +4547,15 @@ lpfc_sli4_async_fc_evt(struct lpfc_hba *phba, struct lpfc_acqe_fc_la *acqe_fc) /* Parse and translate link attention fields */ la = (struct lpfc_mbx_read_top *)&pmb->u.mb.un.varReadTop; la->eventTag = acqe_fc->event_tag; - bf_set(lpfc_mbx_read_top_att_type, la, - LPFC_FC_LA_TYPE_LINK_DOWN); + if (phba->sli4_hba.link_state.status == + LPFC_FC_LA_TYPE_UNEXP_WWPN) { + bf_set(lpfc_mbx_read_top_att_type, la, + LPFC_FC_LA_TYPE_UNEXP_WWPN); + } else { + bf_set(lpfc_mbx_read_top_att_type, la, + LPFC_FC_LA_TYPE_LINK_DOWN); + } /* Invoke the mailbox command callback function */ lpfc_mbx_cmpl_read_topology(phba, pmb); @@ -4716,10 +4761,9 @@ lpfc_sli4_perform_vport_cvl(struct lpfc_vport *vport) ndlp = lpfc_findnode_did(vport, Fabric_DID); if (!ndlp) { /* Cannot find existing Fabric ndlp, so allocate a new one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, Fabric_DID); if (!ndlp) return 0; - lpfc_nlp_init(vport, ndlp, Fabric_DID); /* Set the node type */ ndlp->nlp_type |= NLP_FABRIC; /* Put ndlp onto node list */ @@ -5778,6 +5822,7 @@ lpfc_sli4_driver_resource_setup(struct lpfc_hba *phba) /* Initialize the Abort nvme buffer list used by driver */ spin_lock_init(&phba->sli4_hba.abts_nvme_buf_list_lock); INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_nvme_buf_list); + INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_nvmet_ctx_list); /* Fast-path XRI aborted CQ Event work queue list */ INIT_LIST_HEAD(&phba->sli4_hba.sp_nvme_xri_aborted_work_queue); } @@ -5809,6 +5854,12 @@ lpfc_sli4_driver_resource_setup(struct lpfc_hba *phba) INIT_LIST_HEAD(&phba->sli4_hba.lpfc_vfi_blk_list); INIT_LIST_HEAD(&phba->lpfc_vpi_blk_list); + /* Initialize mboxq lists. If the early init routines fail + * these lists need to be correctly initialized. + */ + INIT_LIST_HEAD(&phba->sli.mboxq); + INIT_LIST_HEAD(&phba->sli.mboxq_cmpl); + /* initialize optic_state to 0xFF */ phba->sli4_hba.lnk_info.optic_state = 0xff; @@ -5874,6 +5925,7 @@ lpfc_sli4_driver_resource_setup(struct lpfc_hba *phba) "READ_NV, mbxStatus x%x\n", bf_get(lpfc_mqe_command, &mboxq->u.mqe), bf_get(lpfc_mqe_status, &mboxq->u.mqe)); + mempool_free(mboxq, phba->mbox_mem_pool); rc = -EIO; goto out_free_bsmbx; } @@ -6398,7 +6450,7 @@ lpfc_init_sgl_list(struct lpfc_hba *phba) INIT_LIST_HEAD(&phba->sli4_hba.lpfc_els_sgl_list); INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_els_sgl_list); INIT_LIST_HEAD(&phba->sli4_hba.lpfc_nvmet_sgl_list); - INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_nvmet_sgl_list); + INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_nvmet_ctx_list); /* els xri-sgl book keeping */ phba->sli4_hba.els_xri_cnt = 0; @@ -7799,7 +7851,7 @@ lpfc_alloc_fcp_wq_cq(struct lpfc_hba *phba, int wqidx) /* Create Fast Path FCP WQs */ wqesize = (phba->fcp_embed_io) ? - LPFC_WQE128_SIZE : phba->sli4_hba.wq_esize; + LPFC_WQE128_SIZE : phba->sli4_hba.wq_esize; qdesc = lpfc_sli4_queue_alloc(phba, wqesize, phba->sli4_hba.wq_ecount); if (!qdesc) { lpfc_printf_log(phba, KERN_ERR, LOG_INIT, @@ -7830,7 +7882,7 @@ int lpfc_sli4_queue_create(struct lpfc_hba *phba) { struct lpfc_queue *qdesc; - int idx, io_channel, max; + int idx, io_channel; /* * Create HBA Record arrays. @@ -7991,15 +8043,6 @@ lpfc_sli4_queue_create(struct lpfc_hba *phba) if (lpfc_alloc_nvme_wq_cq(phba, idx)) goto out_error; - /* allocate MRQ CQs */ - max = phba->cfg_nvme_io_channel; - if (max < phba->cfg_nvmet_mrq) - max = phba->cfg_nvmet_mrq; - - for (idx = 0; idx < max; idx++) - if (lpfc_alloc_nvme_wq_cq(phba, idx)) - goto out_error; - if (phba->nvmet_support) { for (idx = 0; idx < phba->cfg_nvmet_mrq; idx++) { qdesc = lpfc_sli4_queue_alloc(phba, @@ -8221,11 +8264,11 @@ lpfc_sli4_queue_destroy(struct lpfc_hba *phba) /* Release FCP cqs */ lpfc_sli4_release_queues(&phba->sli4_hba.fcp_cq, - phba->cfg_fcp_io_channel); + phba->cfg_fcp_io_channel); /* Release FCP wqs */ lpfc_sli4_release_queues(&phba->sli4_hba.fcp_wq, - phba->cfg_fcp_io_channel); + phba->cfg_fcp_io_channel); /* Release FCP CQ mapping array */ lpfc_sli4_release_queue_map(&phba->sli4_hba.fcp_cq_map); @@ -8571,15 +8614,15 @@ lpfc_sli4_queue_setup(struct lpfc_hba *phba) lpfc_printf_log(phba, KERN_ERR, LOG_INIT, "0528 %s not allocated\n", phba->sli4_hba.mbx_cq ? - "Mailbox WQ" : "Mailbox CQ"); + "Mailbox WQ" : "Mailbox CQ"); rc = -ENOMEM; goto out_destroy; } rc = lpfc_create_wq_cq(phba, phba->sli4_hba.hba_eq[0], - phba->sli4_hba.mbx_cq, - phba->sli4_hba.mbx_wq, - NULL, 0, LPFC_MBOX); + phba->sli4_hba.mbx_cq, + phba->sli4_hba.mbx_wq, + NULL, 0, LPFC_MBOX); if (rc) { lpfc_printf_log(phba, KERN_ERR, LOG_INIT, "0529 Failed setup of mailbox WQ/CQ: rc = 0x%x\n", @@ -9934,17 +9977,19 @@ lpfc_sli4_xri_exchange_busy_wait(struct lpfc_hba *phba) { int wait_time = 0; int nvme_xri_cmpl = 1; + int nvmet_xri_cmpl = 1; int fcp_xri_cmpl = 1; int els_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_els_sgl_list); - int nvmet_xri_cmpl = - list_empty(&phba->sli4_hba.lpfc_abts_nvmet_sgl_list); if (phba->cfg_enable_fc4_type & LPFC_ENABLE_FCP) fcp_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_scsi_buf_list); - if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) + if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) { nvme_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_nvme_buf_list); + nvmet_xri_cmpl = + list_empty(&phba->sli4_hba.lpfc_abts_nvmet_ctx_list); + } while (!fcp_xri_cmpl || !els_xri_cmpl || !nvme_xri_cmpl || !nvmet_xri_cmpl) { @@ -9970,9 +10015,12 @@ lpfc_sli4_xri_exchange_busy_wait(struct lpfc_hba *phba) msleep(LPFC_XRI_EXCH_BUSY_WAIT_T1); wait_time += LPFC_XRI_EXCH_BUSY_WAIT_T1; } - if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) + if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME) { nvme_xri_cmpl = list_empty( &phba->sli4_hba.lpfc_abts_nvme_buf_list); + nvmet_xri_cmpl = list_empty( + &phba->sli4_hba.lpfc_abts_nvmet_ctx_list); + } if (phba->cfg_enable_fc4_type & LPFC_ENABLE_FCP) fcp_xri_cmpl = list_empty( @@ -9981,8 +10029,6 @@ lpfc_sli4_xri_exchange_busy_wait(struct lpfc_hba *phba) els_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_els_sgl_list); - nvmet_xri_cmpl = - list_empty(&phba->sli4_hba.lpfc_abts_nvmet_sgl_list); } } @@ -10048,9 +10094,14 @@ lpfc_sli4_hba_unset(struct lpfc_hba *phba) /* Stop kthread signal shall trigger work_done one more time */ kthread_stop(phba->worker_thread); + /* Unset the queues shared with the hardware then release all + * allocated resources. + */ + lpfc_sli4_queue_unset(phba); + lpfc_sli4_queue_destroy(phba); + /* Reset SLI4 HBA FCoE function */ lpfc_pci_function_reset(phba); - lpfc_sli4_queue_destroy(phba); /* Stop the SLI4 device port */ phba->pport->work_port_events = 0; @@ -10306,6 +10357,7 @@ lpfc_pci_probe_one_s3(struct pci_dev *pdev, const struct pci_device_id *pid) } /* Initialize and populate the iocb list per host */ + error = lpfc_init_iocb_list(phba, LPFC_IOCB_LIST_CNT); if (error) { lpfc_printf_log(phba, KERN_ERR, LOG_INIT, @@ -11051,7 +11103,7 @@ lpfc_pci_probe_one_s4(struct pci_dev *pdev, const struct pci_device_id *pid) struct lpfc_hba *phba; struct lpfc_vport *vport = NULL; struct Scsi_Host *shost = NULL; - int error; + int error, cnt; uint32_t cfg_mode, intr_mode; /* Allocate memory for HBA structure */ @@ -11085,12 +11137,15 @@ lpfc_pci_probe_one_s4(struct pci_dev *pdev, const struct pci_device_id *pid) goto out_unset_pci_mem_s4; } - /* Initialize and populate the iocb list per host */ + cnt = phba->cfg_iocb_cnt * 1024; + if (phba->nvmet_support) + cnt += phba->cfg_nvmet_mrq_post * phba->cfg_nvmet_mrq; + /* Initialize and populate the iocb list per host */ lpfc_printf_log(phba, KERN_INFO, LOG_INIT, - "2821 initialize iocb list %d.\n", - phba->cfg_iocb_cnt*1024); - error = lpfc_init_iocb_list(phba, phba->cfg_iocb_cnt*1024); + "2821 initialize iocb list %d total %d\n", + phba->cfg_iocb_cnt, cnt); + error = lpfc_init_iocb_list(phba, cnt); if (error) { lpfc_printf_log(phba, KERN_ERR, LOG_INIT, @@ -11177,7 +11232,9 @@ lpfc_pci_probe_one_s4(struct pci_dev *pdev, const struct pci_device_id *pid) if ((phba->nvmet_support == 0) && (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME)) { /* Create NVME binding with nvme_fc_transport. This - * ensures the vport is initialized. + * ensures the vport is initialized. If the localport + * create fails, it should not unload the driver to + * support field issues. */ error = lpfc_nvme_create_localport(vport); if (error) { @@ -11185,7 +11242,6 @@ lpfc_pci_probe_one_s4(struct pci_dev *pdev, const struct pci_device_id *pid) "6004 NVME registration failed, " "error x%x\n", error); - goto out_disable_intr; } } @@ -11984,6 +12040,7 @@ int lpfc_fof_queue_create(struct lpfc_hba *phba) { struct lpfc_queue *qdesc; + uint32_t wqesize; /* Create FOF EQ */ qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.eq_esize, @@ -12004,8 +12061,11 @@ lpfc_fof_queue_create(struct lpfc_hba *phba) phba->sli4_hba.oas_cq = qdesc; /* Create OAS WQ */ - qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.wq_esize, + wqesize = (phba->fcp_embed_io) ? + LPFC_WQE128_SIZE : phba->sli4_hba.wq_esize; + qdesc = lpfc_sli4_queue_alloc(phba, wqesize, phba->sli4_hba.wq_ecount); + if (!qdesc) goto out_error; diff --git a/drivers/scsi/lpfc/lpfc_mbox.c b/drivers/scsi/lpfc/lpfc_mbox.c index a928f5187fa4..ce25a18367b5 100644 --- a/drivers/scsi/lpfc/lpfc_mbox.c +++ b/drivers/scsi/lpfc/lpfc_mbox.c @@ -2083,9 +2083,12 @@ lpfc_request_features(struct lpfc_hba *phba, struct lpfcMboxq *mboxq) if (phba->max_vpi && phba->cfg_enable_npiv) bf_set(lpfc_mbx_rq_ftr_rq_npiv, &mboxq->u.mqe.un.req_ftrs, 1); - if (phba->nvmet_support) + if (phba->nvmet_support) { bf_set(lpfc_mbx_rq_ftr_rq_mrqp, &mboxq->u.mqe.un.req_ftrs, 1); - + /* iaab/iaar NOT set for now */ + bf_set(lpfc_mbx_rq_ftr_rq_iaab, &mboxq->u.mqe.un.req_ftrs, 0); + bf_set(lpfc_mbx_rq_ftr_rq_iaar, &mboxq->u.mqe.un.req_ftrs, 0); + } return; } diff --git a/drivers/scsi/lpfc/lpfc_nportdisc.c b/drivers/scsi/lpfc/lpfc_nportdisc.c index 061626bdf701..8777c2d5f50d 100644 --- a/drivers/scsi/lpfc/lpfc_nportdisc.c +++ b/drivers/scsi/lpfc/lpfc_nportdisc.c @@ -361,8 +361,12 @@ lpfc_rcv_plogi(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, case NLP_STE_PRLI_ISSUE: case NLP_STE_UNMAPPED_NODE: case NLP_STE_MAPPED_NODE: - /* lpfc_plogi_confirm_nport skips fabric did, handle it here */ - if (!(ndlp->nlp_type & NLP_FABRIC)) { + /* For initiators, lpfc_plogi_confirm_nport skips fabric did. + * For target mode, execute implicit logo. + * Fabric nodes go into NPR. + */ + if (!(ndlp->nlp_type & NLP_FABRIC) && + !(phba->nvmet_support)) { lpfc_els_rsp_acc(vport, ELS_CMD_PLOGI, cmdiocb, ndlp, NULL); return 1; diff --git a/drivers/scsi/lpfc/lpfc_nvme.c b/drivers/scsi/lpfc/lpfc_nvme.c index 0024de1c6c1f..8008c8205fb6 100644 --- a/drivers/scsi/lpfc/lpfc_nvme.c +++ b/drivers/scsi/lpfc/lpfc_nvme.c @@ -401,6 +401,7 @@ lpfc_nvme_ls_req(struct nvme_fc_local_port *pnvme_lport, struct lpfc_nodelist *ndlp; struct ulp_bde64 *bpl; struct lpfc_dmabuf *bmp; + uint16_t ntype, nstate; /* there are two dma buf in the request, actually there is one and * the second one is just the start address + cmd size. @@ -417,11 +418,26 @@ lpfc_nvme_ls_req(struct nvme_fc_local_port *pnvme_lport, vport = lport->vport; ndlp = lpfc_findnode_did(vport, pnvme_rport->port_id); - if (!ndlp) { - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_DISC, - "6043 Could not find node for DID %x\n", + if (!ndlp || !NLP_CHK_NODE_ACT(ndlp)) { + lpfc_printf_vlog(vport, KERN_ERR, LOG_NODE | LOG_NVME_IOERR, + "6051 DID x%06x not an active rport.\n", pnvme_rport->port_id); - return 1; + return -ENODEV; + } + + /* The remote node has to be a mapped nvme target or an + * unmapped nvme initiator or it's an error. + */ + ntype = ndlp->nlp_type; + nstate = ndlp->nlp_state; + if ((ntype & NLP_NVME_TARGET && nstate != NLP_STE_MAPPED_NODE) || + (ntype & NLP_NVME_INITIATOR && nstate != NLP_STE_UNMAPPED_NODE)) { + lpfc_printf_vlog(vport, KERN_ERR, LOG_NODE | LOG_NVME_IOERR, + "6088 DID x%06x not ready for " + "IO. State x%x, Type x%x\n", + pnvme_rport->port_id, + ndlp->nlp_state, ndlp->nlp_type); + return -ENODEV; } bmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); if (!bmp) { @@ -456,7 +472,7 @@ lpfc_nvme_ls_req(struct nvme_fc_local_port *pnvme_lport, /* Expand print to include key fields. */ lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME_DISC, - "6051 ENTER. lport %p, rport %p lsreq%p rqstlen:%d " + "6149 ENTER. lport %p, rport %p lsreq%p rqstlen:%d " "rsplen:%d %pad %pad\n", pnvme_lport, pnvme_rport, pnvme_lsreq, pnvme_lsreq->rqstlen, @@ -745,6 +761,7 @@ lpfc_nvme_io_cmd_wqe_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pwqeIn, struct nvme_fc_cmd_iu *cp; struct lpfc_nvme_rport *rport; struct lpfc_nodelist *ndlp; + struct lpfc_nvme_fcpreq_priv *freqpriv; unsigned long flags; uint32_t code; uint16_t cid, sqhd, data; @@ -772,9 +789,8 @@ lpfc_nvme_io_cmd_wqe_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pwqeIn, ndlp = rport->ndlp; if (!ndlp || !NLP_CHK_NODE_ACT(ndlp)) { lpfc_printf_vlog(vport, KERN_ERR, LOG_NODE | LOG_NVME_IOERR, - "6061 rport %p, ndlp %p, DID x%06x ndlp " - "not ready.\n", - rport, ndlp, rport->remoteport->port_id); + "6061 rport %p, DID x%06x node not ready.\n", + rport, rport->remoteport->port_id); ndlp = lpfc_findnode_did(vport, rport->remoteport->port_id); if (!ndlp) { @@ -853,15 +869,18 @@ lpfc_nvme_io_cmd_wqe_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pwqeIn, break; lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_IOERR, "6081 NVME Completion Protocol Error: " - "status x%x result x%x placed x%x\n", + "xri %x status x%x result x%x " + "placed x%x\n", + lpfc_ncmd->cur_iocbq.sli4_xritag, lpfc_ncmd->status, lpfc_ncmd->result, wcqe->total_data_placed); break; default: out_err: lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_IOERR, - "6072 NVME Completion Error: " + "6072 NVME Completion Error: xri %x " "status x%x result x%x placed x%x\n", + lpfc_ncmd->cur_iocbq.sli4_xritag, lpfc_ncmd->status, lpfc_ncmd->result, wcqe->total_data_placed); nCmd->transferred_length = 0; @@ -900,6 +919,8 @@ out_err: phba->cpucheck_cmpl_io[lpfc_ncmd->cpu]++; } #endif + freqpriv = nCmd->private; + freqpriv->nvme_buf = NULL; nCmd->done(nCmd); spin_lock_irqsave(&phba->hbalock, flags); @@ -1099,12 +1120,12 @@ lpfc_nvme_prep_io_dma(struct lpfc_vport *vport, first_data_sgl = sgl; lpfc_ncmd->seg_cnt = nCmd->sg_cnt; - if (lpfc_ncmd->seg_cnt > phba->cfg_sg_seg_cnt) { + if (lpfc_ncmd->seg_cnt > phba->cfg_nvme_seg_cnt) { lpfc_printf_log(phba, KERN_ERR, LOG_NVME_IOERR, "6058 Too many sg segments from " "NVME Transport. Max %d, " "nvmeIO sg_cnt %d\n", - phba->cfg_sg_seg_cnt, + phba->cfg_nvme_seg_cnt, lpfc_ncmd->seg_cnt); lpfc_ncmd->seg_cnt = 0; return 1; @@ -1196,6 +1217,7 @@ lpfc_nvme_fcp_io_submit(struct nvme_fc_local_port *pnvme_lport, struct lpfc_nvme_buf *lpfc_ncmd; struct lpfc_nvme_rport *rport; struct lpfc_nvme_qhandle *lpfc_queue_info; + struct lpfc_nvme_fcpreq_priv *freqpriv = pnvme_fcreq->private; #ifdef CONFIG_SCSI_LPFC_DEBUG_FS uint64_t start = 0; #endif @@ -1274,7 +1296,7 @@ lpfc_nvme_fcp_io_submit(struct nvme_fc_local_port *pnvme_lport, * Do not let the IO hang out forever. There is no midlayer issuing * an abort so inform the FW of the maximum IO pending time. */ - pnvme_fcreq->private = (void *)lpfc_ncmd; + freqpriv->nvme_buf = lpfc_ncmd; lpfc_ncmd->nvmeCmd = pnvme_fcreq; lpfc_ncmd->nrport = rport; lpfc_ncmd->ndlp = ndlp; @@ -1404,6 +1426,7 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, struct lpfc_nvme_buf *lpfc_nbuf; struct lpfc_iocbq *abts_buf; struct lpfc_iocbq *nvmereq_wqe; + struct lpfc_nvme_fcpreq_priv *freqpriv = pnvme_fcreq->private; union lpfc_wqe *abts_wqe; unsigned long flags; int ret_val; @@ -1414,7 +1437,7 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, phba = vport->phba; /* Announce entry to new IO submit field. */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, + lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME_ABTS, "6002 Abort Request to rport DID x%06x " "for nvme_fc_req %p\n", pnvme_rport->port_id, @@ -1444,7 +1467,7 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, /* The remote node has to be ready to send an abort. */ if ((ndlp->nlp_state != NLP_STE_MAPPED_NODE) && !(ndlp->nlp_type & NLP_NVME_TARGET)) { - lpfc_printf_vlog(vport, KERN_ERR, LOG_NODE | LOG_NVME_ABTS, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6048 rport %p, DID x%06x not ready for " "IO. State x%x, Type x%x\n", rport, pnvme_rport->port_id, @@ -1459,27 +1482,28 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, /* driver queued commands are in process of being flushed */ if (phba->hba_flag & HBA_NVME_IOQ_FLUSH) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6139 Driver in reset cleanup - flushing " "NVME Req now. hba_flag x%x\n", phba->hba_flag); return; } - lpfc_nbuf = (struct lpfc_nvme_buf *)pnvme_fcreq->private; + lpfc_nbuf = freqpriv->nvme_buf; if (!lpfc_nbuf) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6140 NVME IO req has no matching lpfc nvme " "io buffer. Skipping abort req.\n"); return; } else if (!lpfc_nbuf->nvmeCmd) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6141 lpfc NVME IO req has no nvme_fcreq " "io buffer. Skipping abort req.\n"); return; } + nvmereq_wqe = &lpfc_nbuf->cur_iocbq; /* * The lpfc_nbuf and the mapped nvme_fcreq in the driver's @@ -1490,23 +1514,22 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, */ if (lpfc_nbuf->nvmeCmd != pnvme_fcreq) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6143 NVME req mismatch: " "lpfc_nbuf %p nvmeCmd %p, " - "pnvme_fcreq %p. Skipping Abort\n", + "pnvme_fcreq %p. Skipping Abort xri x%x\n", lpfc_nbuf, lpfc_nbuf->nvmeCmd, - pnvme_fcreq); + pnvme_fcreq, nvmereq_wqe->sli4_xritag); return; } /* Don't abort IOs no longer on the pending queue. */ - nvmereq_wqe = &lpfc_nbuf->cur_iocbq; if (!(nvmereq_wqe->iocb_flag & LPFC_IO_ON_TXCMPLQ)) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6142 NVME IO req %p not queued - skipping " - "abort req\n", - pnvme_fcreq); + "abort req xri x%x\n", + pnvme_fcreq, nvmereq_wqe->sli4_xritag); return; } @@ -1517,21 +1540,22 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, /* Outstanding abort is in progress */ if (nvmereq_wqe->iocb_flag & LPFC_DRIVER_ABORTED) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6144 Outstanding NVME I/O Abort Request " "still pending on nvme_fcreq %p, " - "lpfc_ncmd %p\n", - pnvme_fcreq, lpfc_nbuf); + "lpfc_ncmd %p xri x%x\n", + pnvme_fcreq, lpfc_nbuf, + nvmereq_wqe->sli4_xritag); return; } abts_buf = __lpfc_sli_get_iocbq(phba); if (!abts_buf) { spin_unlock_irqrestore(&phba->hbalock, flags); - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6136 No available abort wqes. Skipping " - "Abts req for nvme_fcreq %p.\n", - pnvme_fcreq); + "Abts req for nvme_fcreq %p xri x%x\n", + pnvme_fcreq, nvmereq_wqe->sli4_xritag); return; } @@ -1580,7 +1604,7 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, ret_val = lpfc_sli4_issue_wqe(phba, LPFC_FCP_RING, abts_buf); spin_unlock_irqrestore(&phba->hbalock, flags); if (ret_val == IOCB_ERROR) { - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, + lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_ABTS, "6137 Failed abts issue_wqe with status x%x " "for nvme_fcreq %p.\n", ret_val, pnvme_fcreq); @@ -1588,8 +1612,8 @@ lpfc_nvme_fcp_abort(struct nvme_fc_local_port *pnvme_lport, return; } - lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME, - "6138 Transport Abort NVME Request Issued for\n" + lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME_ABTS, + "6138 Transport Abort NVME Request Issued for " "ox_id x%x on reqtag x%x\n", nvmereq_wqe->sli4_xritag, abts_buf->iotag); @@ -1618,7 +1642,7 @@ static struct nvme_fc_port_template lpfc_nvme_template = { .local_priv_sz = sizeof(struct lpfc_nvme_lport), .remote_priv_sz = sizeof(struct lpfc_nvme_rport), .lsrqst_priv_sz = 0, - .fcprqst_priv_sz = 0, + .fcprqst_priv_sz = sizeof(struct lpfc_nvme_fcpreq_priv), }; /** @@ -2049,7 +2073,7 @@ lpfc_get_nvme_buf(struct lpfc_hba *phba, struct lpfc_nodelist *ndlp) if (lpfc_test_rrq_active(phba, ndlp, lpfc_ncmd->cur_iocbq.sli4_lxritag)) continue; - list_del(&lpfc_ncmd->list); + list_del_init(&lpfc_ncmd->list); found = 1; break; } @@ -2064,7 +2088,7 @@ lpfc_get_nvme_buf(struct lpfc_hba *phba, struct lpfc_nodelist *ndlp) if (lpfc_test_rrq_active( phba, ndlp, lpfc_ncmd->cur_iocbq.sli4_lxritag)) continue; - list_del(&lpfc_ncmd->list); + list_del_init(&lpfc_ncmd->list); found = 1; break; } @@ -2092,6 +2116,12 @@ lpfc_release_nvme_buf(struct lpfc_hba *phba, struct lpfc_nvme_buf *lpfc_ncmd) lpfc_ncmd->nonsg_phys = 0; if (lpfc_ncmd->flags & LPFC_SBUF_XBUSY) { + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6310 XB release deferred for " + "ox_id x%x on reqtag x%x\n", + lpfc_ncmd->cur_iocbq.sli4_xritag, + lpfc_ncmd->cur_iocbq.iotag); + spin_lock_irqsave(&phba->sli4_hba.abts_nvme_buf_list_lock, iflag); lpfc_ncmd->nvmeCmd = NULL; @@ -2142,8 +2172,18 @@ lpfc_nvme_create_localport(struct lpfc_vport *vport) nfcp_info.node_name = wwn_to_u64(vport->fc_nodename.u.wwn); nfcp_info.port_name = wwn_to_u64(vport->fc_portname.u.wwn); - /* For now need + 1 to get around NVME transport logic */ - lpfc_nvme_template.max_sgl_segments = phba->cfg_sg_seg_cnt + 1; + /* Limit to LPFC_MAX_NVME_SEG_CNT. + * For now need + 1 to get around NVME transport logic. + */ + if (phba->cfg_sg_seg_cnt > LPFC_MAX_NVME_SEG_CNT) { + lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME | LOG_INIT, + "6300 Reducing sg segment cnt to %d\n", + LPFC_MAX_NVME_SEG_CNT); + phba->cfg_nvme_seg_cnt = LPFC_MAX_NVME_SEG_CNT; + } else { + phba->cfg_nvme_seg_cnt = phba->cfg_sg_seg_cnt; + } + lpfc_nvme_template.max_sgl_segments = phba->cfg_nvme_seg_cnt + 1; lpfc_nvme_template.max_hw_queues = phba->cfg_nvme_io_channel; /* localport is allocated from the stack, but the registration @@ -2249,12 +2289,23 @@ lpfc_nvme_destroy_localport(struct lpfc_vport *vport) void lpfc_nvme_update_localport(struct lpfc_vport *vport) { +#if (IS_ENABLED(CONFIG_NVME_FC)) struct nvme_fc_local_port *localport; struct lpfc_nvme_lport *lport; localport = vport->localport; + if (!localport) { + lpfc_printf_vlog(vport, KERN_WARNING, LOG_NVME, + "6710 Update NVME fail. No localport\n"); + return; + } lport = (struct lpfc_nvme_lport *)localport->private; - + if (!lport) { + lpfc_printf_vlog(vport, KERN_WARNING, LOG_NVME, + "6171 Update NVME fail. localP %p, No lport\n", + localport); + return; + } lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME, "6012 Update NVME lport %p did x%x\n", localport, vport->fc_myDID); @@ -2268,7 +2319,7 @@ lpfc_nvme_update_localport(struct lpfc_vport *vport) lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME_DISC, "6030 bound lport %p to DID x%06x\n", lport, localport->port_id); - +#endif } int @@ -2409,6 +2460,7 @@ lpfc_nvme_unregister_port(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp) struct lpfc_nvme_lport *lport; struct lpfc_nvme_rport *rport; struct nvme_fc_remote_port *remoteport; + unsigned long wait_tmo; localport = vport->localport; @@ -2451,11 +2503,12 @@ lpfc_nvme_unregister_port(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp) * before proceeding. This guarantees the transport and driver * have completed the unreg process. */ - ret = wait_for_completion_timeout(&rport->rport_unreg_done, 5); + wait_tmo = msecs_to_jiffies(5000); + ret = wait_for_completion_timeout(&rport->rport_unreg_done, + wait_tmo); if (ret == 0) { lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_DISC, - "6169 Unreg nvme wait failed %d\n", - ret); + "6169 Unreg nvme wait timeout\n"); } } return; @@ -2463,7 +2516,7 @@ lpfc_nvme_unregister_port(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp) input_err: #endif lpfc_printf_vlog(vport, KERN_ERR, LOG_NVME_DISC, - "6168: State error: lport %p, rport%p FCID x%06x\n", + "6168 State error: lport %p, rport%p FCID x%06x\n", vport->localport, ndlp->rport, ndlp->nlp_DID); } @@ -2494,7 +2547,7 @@ lpfc_sli4_nvme_xri_aborted(struct lpfc_hba *phba, &phba->sli4_hba.lpfc_abts_nvme_buf_list, list) { if (lpfc_ncmd->cur_iocbq.sli4_xritag == xri) { - list_del(&lpfc_ncmd->list); + list_del_init(&lpfc_ncmd->list); lpfc_ncmd->flags &= ~LPFC_SBUF_XBUSY; lpfc_ncmd->status = IOSTAT_SUCCESS; spin_unlock( @@ -2510,6 +2563,12 @@ lpfc_sli4_nvme_xri_aborted(struct lpfc_hba *phba, rxid, 1); lpfc_sli4_abts_err_handler(phba, ndlp, axri); } + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6311 XRI Aborted xri x%x tag x%x " + "released\n", + xri, lpfc_ncmd->cur_iocbq.iotag); + lpfc_release_nvme_buf(phba, lpfc_ncmd); if (rrq_empty) lpfc_worker_wake_up(phba); @@ -2518,4 +2577,8 @@ lpfc_sli4_nvme_xri_aborted(struct lpfc_hba *phba, } spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); spin_unlock_irqrestore(&phba->hbalock, iflag); + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6312 XRI Aborted xri x%x not found\n", xri); + } diff --git a/drivers/scsi/lpfc/lpfc_nvme.h b/drivers/scsi/lpfc/lpfc_nvme.h index 1347deb8dd6c..ec32f45daa66 100644 --- a/drivers/scsi/lpfc/lpfc_nvme.h +++ b/drivers/scsi/lpfc/lpfc_nvme.h @@ -21,12 +21,7 @@ * included with this package. * ********************************************************************/ -#define LPFC_NVME_MIN_SEGS 16 -#define LPFC_NVME_DEFAULT_SEGS 66 /* 256K IOs - 64 + 2 */ -#define LPFC_NVME_MAX_SEGS 510 -#define LPFC_NVMET_MIN_POSTBUF 16 -#define LPFC_NVMET_DEFAULT_POSTBUF 1024 -#define LPFC_NVMET_MAX_POSTBUF 4096 +#define LPFC_NVME_DEFAULT_SEGS (64 + 1) /* 256K IOs */ #define LPFC_NVME_WQSIZE 256 #define LPFC_NVME_ERSP_LEN 0x20 @@ -102,3 +97,7 @@ struct lpfc_nvme_buf { uint64_t ts_data_nvme; #endif }; + +struct lpfc_nvme_fcpreq_priv { + struct lpfc_nvme_buf *nvme_buf; +}; diff --git a/drivers/scsi/lpfc/lpfc_nvmet.c b/drivers/scsi/lpfc/lpfc_nvmet.c index b2333b3889c7..94434e621c33 100644 --- a/drivers/scsi/lpfc/lpfc_nvmet.c +++ b/drivers/scsi/lpfc/lpfc_nvmet.c @@ -71,6 +71,26 @@ static int lpfc_nvmet_unsol_ls_issue_abort(struct lpfc_hba *, struct lpfc_nvmet_rcv_ctx *, uint32_t, uint16_t); +void +lpfc_nvmet_defer_release(struct lpfc_hba *phba, struct lpfc_nvmet_rcv_ctx *ctxp) +{ + unsigned long iflag; + + lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, + "6313 NVMET Defer ctx release xri x%x flg x%x\n", + ctxp->oxid, ctxp->flag); + + spin_lock_irqsave(&phba->sli4_hba.abts_nvme_buf_list_lock, iflag); + if (ctxp->flag & LPFC_NVMET_CTX_RLS) { + spin_unlock_irqrestore(&phba->sli4_hba.abts_nvme_buf_list_lock, + iflag); + return; + } + ctxp->flag |= LPFC_NVMET_CTX_RLS; + list_add_tail(&ctxp->list, &phba->sli4_hba.lpfc_abts_nvmet_ctx_list); + spin_unlock_irqrestore(&phba->sli4_hba.abts_nvme_buf_list_lock, iflag); +} + /** * lpfc_nvmet_xmt_ls_rsp_cmp - Completion handler for LS Response * @phba: Pointer to HBA context object. @@ -139,6 +159,11 @@ lpfc_nvmet_rq_post(struct lpfc_hba *phba, struct lpfc_nvmet_rcv_ctx *ctxp, struct lpfc_dmabuf *mp) { if (ctxp) { + if (ctxp->flag) + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6314 rq_post ctx xri x%x flag x%x\n", + ctxp->oxid, ctxp->flag); + if (ctxp->txrdy) { pci_pool_free(phba->txrdy_payload_pool, ctxp->txrdy, ctxp->txrdy_phys); @@ -337,39 +362,55 @@ lpfc_nvmet_xmt_fcp_op_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, #endif ctxp = cmdwqe->context2; + ctxp->flag &= ~LPFC_NVMET_IO_INP; + rsp = &ctxp->ctx.fcp_req; op = rsp->op; - ctxp->flag &= ~LPFC_NVMET_IO_INP; status = bf_get(lpfc_wcqe_c_status, wcqe); result = wcqe->parameter; - if (!phba->targetport) - goto out; + if (phba->targetport) + tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; + else + tgtp = NULL; lpfc_nvmeio_data(phba, "NVMET FCP CMPL: xri x%x op x%x status x%x\n", ctxp->oxid, op, status); - tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; if (status) { rsp->fcp_error = NVME_SC_DATA_XFER_ERROR; rsp->transferred_length = 0; - atomic_inc(&tgtp->xmt_fcp_rsp_error); + if (tgtp) + atomic_inc(&tgtp->xmt_fcp_rsp_error); + + /* pick up SLI4 exhange busy condition */ + if (bf_get(lpfc_wcqe_c_xb, wcqe)) { + ctxp->flag |= LPFC_NVMET_XBUSY; + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6315 IO Cmpl XBUSY: xri x%x: %x/%x\n", + ctxp->oxid, status, result); + } else { + ctxp->flag &= ~LPFC_NVMET_XBUSY; + } + } else { rsp->fcp_error = NVME_SC_SUCCESS; if (op == NVMET_FCOP_RSP) rsp->transferred_length = rsp->rsplen; else rsp->transferred_length = rsp->transfer_length; - atomic_inc(&tgtp->xmt_fcp_rsp_cmpl); + if (tgtp) + atomic_inc(&tgtp->xmt_fcp_rsp_cmpl); } -out: if ((op == NVMET_FCOP_READDATA_RSP) || (op == NVMET_FCOP_RSP)) { /* Sanity check */ ctxp->state = LPFC_NVMET_STE_DONE; ctxp->entry_cnt++; + #ifdef CONFIG_SCSI_LPFC_DEBUG_FS if (phba->ktime_on) { if (rsp->op == NVMET_FCOP_READDATA_RSP) { @@ -517,7 +558,6 @@ lpfc_nvmet_xmt_fcp_op(struct nvmet_fc_target_port *tgtport, container_of(rsp, struct lpfc_nvmet_rcv_ctx, ctx.fcp_req); struct lpfc_hba *phba = ctxp->phba; struct lpfc_iocbq *nvmewqeq; - unsigned long iflags; int rc; #ifdef CONFIG_SCSI_LPFC_DEBUG_FS @@ -543,10 +583,11 @@ lpfc_nvmet_xmt_fcp_op(struct nvmet_fc_target_port *tgtport, #endif /* Sanity check */ - if (ctxp->state == LPFC_NVMET_STE_ABORT) { + if ((ctxp->flag & LPFC_NVMET_ABTS_RCV) || + (ctxp->state == LPFC_NVMET_STE_ABORT)) { atomic_inc(&lpfc_nvmep->xmt_fcp_drop); lpfc_printf_log(phba, KERN_ERR, LOG_NVME_IOERR, - "6102 Bad state IO x%x aborted\n", + "6102 IO xri x%x aborted\n", ctxp->oxid); rc = -ENXIO; goto aerr; @@ -571,10 +612,7 @@ lpfc_nvmet_xmt_fcp_op(struct nvmet_fc_target_port *tgtport, lpfc_nvmeio_data(phba, "NVMET FCP CMND: xri x%x op x%x len x%x\n", ctxp->oxid, rsp->op, rsp->rsplen); - /* For now we take hbalock */ - spin_lock_irqsave(&phba->hbalock, iflags); rc = lpfc_sli4_issue_wqe(phba, LPFC_FCP_RING, nvmewqeq); - spin_unlock_irqrestore(&phba->hbalock, iflags); if (rc == WQE_SUCCESS) { ctxp->flag |= LPFC_NVMET_IO_INP; #ifdef CONFIG_SCSI_LPFC_DEBUG_FS @@ -619,16 +657,27 @@ lpfc_nvmet_xmt_fcp_abort(struct nvmet_fc_target_port *tgtport, struct lpfc_nvmet_rcv_ctx *ctxp = container_of(req, struct lpfc_nvmet_rcv_ctx, ctx.fcp_req); struct lpfc_hba *phba = ctxp->phba; + unsigned long flags; lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, - "6103 Abort op: oxri x%x %d cnt %d\n", - ctxp->oxid, ctxp->state, ctxp->entry_cnt); + "6103 Abort op: oxri x%x flg x%x cnt %d\n", + ctxp->oxid, ctxp->flag, ctxp->entry_cnt); - lpfc_nvmeio_data(phba, "NVMET FCP ABRT: xri x%x state x%x cnt x%x\n", - ctxp->oxid, ctxp->state, ctxp->entry_cnt); + lpfc_nvmeio_data(phba, "NVMET FCP ABRT: " + "xri x%x flg x%x cnt x%x\n", + ctxp->oxid, ctxp->flag, ctxp->entry_cnt); atomic_inc(&lpfc_nvmep->xmt_fcp_abort); ctxp->entry_cnt++; + spin_lock_irqsave(&ctxp->ctxlock, flags); + + /* Since iaab/iaar are NOT set, we need to check + * if the firmware is in process of aborting IO + */ + if (ctxp->flag & LPFC_NVMET_XBUSY) { + spin_unlock_irqrestore(&ctxp->ctxlock, flags); + return; + } ctxp->flag |= LPFC_NVMET_ABORT_OP; if (ctxp->flag & LPFC_NVMET_IO_INP) lpfc_nvmet_sol_fcp_issue_abort(phba, ctxp, ctxp->sid, @@ -636,13 +685,13 @@ lpfc_nvmet_xmt_fcp_abort(struct nvmet_fc_target_port *tgtport, else lpfc_nvmet_unsol_fcp_issue_abort(phba, ctxp, ctxp->sid, ctxp->oxid); + spin_unlock_irqrestore(&ctxp->ctxlock, flags); } static void lpfc_nvmet_xmt_fcp_release(struct nvmet_fc_target_port *tgtport, struct nvmefc_tgt_fcp_req *rsp) { - struct lpfc_nvmet_tgtport *lpfc_nvmep = tgtport->private; struct lpfc_nvmet_rcv_ctx *ctxp = container_of(rsp, struct lpfc_nvmet_rcv_ctx, ctx.fcp_req); struct lpfc_hba *phba = ctxp->phba; @@ -650,27 +699,20 @@ lpfc_nvmet_xmt_fcp_release(struct nvmet_fc_target_port *tgtport, bool aborting = false; spin_lock_irqsave(&ctxp->ctxlock, flags); - if (ctxp->flag & LPFC_NVMET_ABORT_OP) { + if ((ctxp->flag & LPFC_NVMET_ABORT_OP) || + (ctxp->flag & LPFC_NVMET_XBUSY)) { aborting = true; - ctxp->flag |= LPFC_NVMET_CTX_RLS; + /* let the abort path do the real release */ + lpfc_nvmet_defer_release(phba, ctxp); } spin_unlock_irqrestore(&ctxp->ctxlock, flags); - if (aborting) - /* let the abort path do the real release */ - return; - - /* Sanity check */ - if (ctxp->state != LPFC_NVMET_STE_DONE) { - atomic_inc(&lpfc_nvmep->xmt_fcp_drop); - lpfc_printf_log(phba, KERN_ERR, LOG_NVME_IOERR, - "6117 Bad state IO x%x aborted\n", - ctxp->oxid); - } - lpfc_nvmeio_data(phba, "NVMET FCP FREE: xri x%x ste %d\n", ctxp->oxid, ctxp->state, 0); + if (aborting) + return; + lpfc_nvmet_rq_post(phba, ctxp, &ctxp->rqb_buffer->hbuf); } @@ -708,8 +750,19 @@ lpfc_nvmet_create_targetport(struct lpfc_hba *phba) pinfo.port_name = wwn_to_u64(vport->fc_portname.u.wwn); pinfo.port_id = vport->fc_myDID; + /* Limit to LPFC_MAX_NVME_SEG_CNT. + * For now need + 1 to get around NVME transport logic. + */ + if (phba->cfg_sg_seg_cnt > LPFC_MAX_NVME_SEG_CNT) { + lpfc_printf_log(phba, KERN_INFO, LOG_NVME | LOG_INIT, + "6400 Reducing sg segment cnt to %d\n", + LPFC_MAX_NVME_SEG_CNT); + phba->cfg_nvme_seg_cnt = LPFC_MAX_NVME_SEG_CNT; + } else { + phba->cfg_nvme_seg_cnt = phba->cfg_sg_seg_cnt; + } + lpfc_tgttemplate.max_sgl_segments = phba->cfg_nvme_seg_cnt + 1; lpfc_tgttemplate.max_hw_queues = phba->cfg_nvme_io_channel; - lpfc_tgttemplate.max_sgl_segments = phba->cfg_sg_seg_cnt; lpfc_tgttemplate.target_features = NVMET_FCTGTFEAT_READDATA_RSP | NVMET_FCTGTFEAT_NEEDS_CMD_CPUSCHED | NVMET_FCTGTFEAT_CMD_IN_ISR | @@ -794,7 +847,120 @@ void lpfc_sli4_nvmet_xri_aborted(struct lpfc_hba *phba, struct sli4_wcqe_xri_aborted *axri) { - /* TODO: work in progress */ + uint16_t xri = bf_get(lpfc_wcqe_xa_xri, axri); + uint16_t rxid = bf_get(lpfc_wcqe_xa_remote_xid, axri); + struct lpfc_nvmet_rcv_ctx *ctxp, *next_ctxp; + struct lpfc_nodelist *ndlp; + unsigned long iflag = 0; + int rrq_empty = 0; + bool released = false; + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6317 XB aborted xri x%x rxid x%x\n", xri, rxid); + + if (!(phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME)) + return; + spin_lock_irqsave(&phba->hbalock, iflag); + spin_lock(&phba->sli4_hba.abts_nvme_buf_list_lock); + list_for_each_entry_safe(ctxp, next_ctxp, + &phba->sli4_hba.lpfc_abts_nvmet_ctx_list, + list) { + if (ctxp->rqb_buffer->sglq->sli4_xritag != xri) + continue; + + /* Check if we already received a free context call + * and we have completed processing an abort situation. + */ + if (ctxp->flag & LPFC_NVMET_CTX_RLS && + !(ctxp->flag & LPFC_NVMET_ABORT_OP)) { + list_del(&ctxp->list); + released = true; + } + ctxp->flag &= ~LPFC_NVMET_XBUSY; + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + + rrq_empty = list_empty(&phba->active_rrq_list); + spin_unlock_irqrestore(&phba->hbalock, iflag); + ndlp = lpfc_findnode_did(phba->pport, ctxp->sid); + if (ndlp && NLP_CHK_NODE_ACT(ndlp) && + (ndlp->nlp_state == NLP_STE_UNMAPPED_NODE || + ndlp->nlp_state == NLP_STE_MAPPED_NODE)) { + lpfc_set_rrq_active(phba, ndlp, + ctxp->rqb_buffer->sglq->sli4_lxritag, + rxid, 1); + lpfc_sli4_abts_err_handler(phba, ndlp, axri); + } + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6318 XB aborted %x flg x%x (%x)\n", + ctxp->oxid, ctxp->flag, released); + if (released) + lpfc_nvmet_rq_post(phba, ctxp, + &ctxp->rqb_buffer->hbuf); + if (rrq_empty) + lpfc_worker_wake_up(phba); + return; + } + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + spin_unlock_irqrestore(&phba->hbalock, iflag); +} + +int +lpfc_nvmet_rcv_unsol_abort(struct lpfc_vport *vport, + struct fc_frame_header *fc_hdr) + +{ +#if (IS_ENABLED(CONFIG_NVME_TARGET_FC)) + struct lpfc_hba *phba = vport->phba; + struct lpfc_nvmet_rcv_ctx *ctxp, *next_ctxp; + struct nvmefc_tgt_fcp_req *rsp; + uint16_t xri; + unsigned long iflag = 0; + + xri = be16_to_cpu(fc_hdr->fh_ox_id); + + spin_lock_irqsave(&phba->hbalock, iflag); + spin_lock(&phba->sli4_hba.abts_nvme_buf_list_lock); + list_for_each_entry_safe(ctxp, next_ctxp, + &phba->sli4_hba.lpfc_abts_nvmet_ctx_list, + list) { + if (ctxp->rqb_buffer->sglq->sli4_xritag != xri) + continue; + + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + spin_unlock_irqrestore(&phba->hbalock, iflag); + + spin_lock_irqsave(&ctxp->ctxlock, iflag); + ctxp->flag |= LPFC_NVMET_ABTS_RCV; + spin_unlock_irqrestore(&ctxp->ctxlock, iflag); + + lpfc_nvmeio_data(phba, + "NVMET ABTS RCV: xri x%x CPU %02x rjt %d\n", + xri, smp_processor_id(), 0); + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6319 NVMET Rcv ABTS:acc xri x%x\n", xri); + + rsp = &ctxp->ctx.fcp_req; + nvmet_fc_rcv_fcp_abort(phba->targetport, rsp); + + /* Respond with BA_ACC accordingly */ + lpfc_sli4_seq_abort_rsp(vport, fc_hdr, 1); + return 0; + } + spin_unlock(&phba->sli4_hba.abts_nvme_buf_list_lock); + spin_unlock_irqrestore(&phba->hbalock, iflag); + + lpfc_nvmeio_data(phba, "NVMET ABTS RCV: xri x%x CPU %02x rjt %d\n", + xri, smp_processor_id(), 1); + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6320 NVMET Rcv ABTS:rjt xri x%x\n", xri); + + /* Respond with BA_RJT accordingly */ + lpfc_sli4_seq_abort_rsp(vport, fc_hdr, 0); +#endif + return 0; } void @@ -876,7 +1042,6 @@ dropit: ctxp->wqeq = NULL; ctxp->state = LPFC_NVMET_STE_RCV; ctxp->rqb_buffer = (void *)nvmebuf; - spin_lock_init(&ctxp->ctxlock); lpfc_nvmeio_data(phba, "NVMET LS RCV: xri x%x sz %d from %06x\n", oxid, size, sid); @@ -985,6 +1150,7 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba, ctxp->rqb_buffer = nvmebuf; ctxp->entry_cnt = 1; ctxp->flag = 0; + spin_lock_init(&ctxp->ctxlock); #ifdef CONFIG_SCSI_LPFC_DEBUG_FS if (phba->ktime_on) { @@ -1007,8 +1173,8 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba, } #endif - lpfc_nvmeio_data(phba, "NVMET FCP RCV: xri x%x sz %d from %06x\n", - oxid, size, sid); + lpfc_nvmeio_data(phba, "NVMET FCP RCV: xri x%x sz %d CPU %02x\n", + oxid, size, smp_processor_id()); atomic_inc(&tgtp->rcv_fcp_cmd_in); /* @@ -1282,11 +1448,11 @@ lpfc_nvmet_prep_fcp_wqe(struct lpfc_hba *phba, return NULL; } - if (rsp->sg_cnt > phba->cfg_sg_seg_cnt) { + if (rsp->sg_cnt > phba->cfg_nvme_seg_cnt) { lpfc_printf_log(phba, KERN_ERR, LOG_NVME_IOERR, "6109 lpfc_nvmet_prep_fcp_wqe: seg cnt err: " - "NPORT x%x oxid:x%x\n", - ctxp->sid, ctxp->oxid); + "NPORT x%x oxid:x%x cnt %d\n", + ctxp->sid, ctxp->oxid, phba->cfg_nvme_seg_cnt); return NULL; } @@ -1648,18 +1814,27 @@ lpfc_nvmet_sol_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; atomic_inc(&tgtp->xmt_abort_cmpl); - lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, - "6165 Abort cmpl: xri x%x WCQE: %08x %08x %08x %08x\n", - ctxp->oxid, wcqe->word0, wcqe->total_data_placed, - result, wcqe->word3); - ctxp->state = LPFC_NVMET_STE_DONE; + + /* Check if we already received a free context call + * and we have completed processing an abort situation. + */ spin_lock_irqsave(&ctxp->ctxlock, flags); - if (ctxp->flag & LPFC_NVMET_CTX_RLS) + if ((ctxp->flag & LPFC_NVMET_CTX_RLS) && + !(ctxp->flag & LPFC_NVMET_XBUSY)) { + list_del(&ctxp->list); released = true; + } ctxp->flag &= ~LPFC_NVMET_ABORT_OP; spin_unlock_irqrestore(&ctxp->ctxlock, flags); + lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, + "6165 ABORT cmpl: xri x%x flg x%x (%d) " + "WCQE: %08x %08x %08x %08x\n", + ctxp->oxid, ctxp->flag, released, + wcqe->word0, wcqe->total_data_placed, + result, wcqe->word3); + /* * if transport has released ctx, then can reuse it. Otherwise, * will be recycled by transport release call. @@ -1670,10 +1845,15 @@ lpfc_nvmet_sol_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, cmdwqe->context2 = NULL; cmdwqe->context3 = NULL; lpfc_sli_release_iocbq(phba, cmdwqe); + + /* Since iaab/iaar are NOT set, there is no work left. + * For LPFC_NVMET_XBUSY, lpfc_sli4_nvmet_xri_aborted + * should have been called already. + */ } /** - * lpfc_nvmet_xmt_fcp_abort_cmp - Completion handler for ABTS + * lpfc_nvmet_unsol_fcp_abort_cmp - Completion handler for ABTS * @phba: Pointer to HBA context object. * @cmdwqe: Pointer to driver command WQE object. * @wcqe: Pointer to driver response CQE object. @@ -1683,8 +1863,8 @@ lpfc_nvmet_sol_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, * The function frees memory resources used for the NVME commands. **/ static void -lpfc_nvmet_xmt_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, - struct lpfc_wcqe_complete *wcqe) +lpfc_nvmet_unsol_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, + struct lpfc_wcqe_complete *wcqe) { struct lpfc_nvmet_rcv_ctx *ctxp; struct lpfc_nvmet_tgtport *tgtp; @@ -1699,35 +1879,55 @@ lpfc_nvmet_xmt_fcp_abort_cmp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdwqe, tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; atomic_inc(&tgtp->xmt_abort_cmpl); - lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, - "6070 Abort cmpl: ctx %p WCQE: %08x %08x %08x %08x\n", - ctxp, wcqe->word0, wcqe->total_data_placed, - result, wcqe->word3); - - if (ctxp) { - /* Sanity check */ - if (ctxp->state != LPFC_NVMET_STE_ABORT) { - lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, - "6112 ABORT Wrong state:%d oxid x%x\n", - ctxp->state, ctxp->oxid); - } - ctxp->state = LPFC_NVMET_STE_DONE; - spin_lock_irqsave(&ctxp->ctxlock, flags); - if (ctxp->flag & LPFC_NVMET_CTX_RLS) - released = true; - ctxp->flag &= ~LPFC_NVMET_ABORT_OP; - spin_unlock_irqrestore(&ctxp->ctxlock, flags); - - /* - * if transport has released ctx, then can reuse it. Otherwise, - * will be recycled by transport release call. - */ - if (released) - lpfc_nvmet_rq_post(phba, ctxp, &ctxp->rqb_buffer->hbuf); - - cmdwqe->context2 = NULL; - cmdwqe->context3 = NULL; + if (!ctxp) { + /* if context is clear, related io alrady complete */ + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6070 ABTS cmpl: WCQE: %08x %08x %08x %08x\n", + wcqe->word0, wcqe->total_data_placed, + result, wcqe->word3); + return; } + + /* Sanity check */ + if (ctxp->state != LPFC_NVMET_STE_ABORT) { + lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, + "6112 ABTS Wrong state:%d oxid x%x\n", + ctxp->state, ctxp->oxid); + } + + /* Check if we already received a free context call + * and we have completed processing an abort situation. + */ + ctxp->state = LPFC_NVMET_STE_DONE; + spin_lock_irqsave(&ctxp->ctxlock, flags); + if ((ctxp->flag & LPFC_NVMET_CTX_RLS) && + !(ctxp->flag & LPFC_NVMET_XBUSY)) { + list_del(&ctxp->list); + released = true; + } + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; + spin_unlock_irqrestore(&ctxp->ctxlock, flags); + + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6316 ABTS cmpl xri x%x flg x%x (%x) " + "WCQE: %08x %08x %08x %08x\n", + ctxp->oxid, ctxp->flag, released, + wcqe->word0, wcqe->total_data_placed, + result, wcqe->word3); + /* + * if transport has released ctx, then can reuse it. Otherwise, + * will be recycled by transport release call. + */ + if (released) + lpfc_nvmet_rq_post(phba, ctxp, &ctxp->rqb_buffer->hbuf); + + cmdwqe->context2 = NULL; + cmdwqe->context3 = NULL; + + /* Since iaab/iaar are NOT set, there is no work left. + * For LPFC_NVMET_XBUSY, lpfc_sli4_nvmet_xri_aborted + * should have been called already. + */ } /** @@ -1780,10 +1980,14 @@ lpfc_nvmet_unsol_issue_abort(struct lpfc_hba *phba, struct lpfc_nodelist *ndlp; lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, - "6067 Abort: sid %x xri x%x/x%x\n", + "6067 ABTS: sid %x xri x%x/x%x\n", sid, xri, ctxp->wqeq->sli4_xritag); tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private; + if (!ctxp->wqeq) { + ctxp->wqeq = ctxp->rqb_buffer->iocbq; + ctxp->wqeq->hba_wqidx = 0; + } ndlp = lpfc_findnode_did(phba->pport, sid); if (!ndlp || !NLP_CHK_NODE_ACT(ndlp) || @@ -1889,10 +2093,11 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, (ndlp->nlp_state != NLP_STE_MAPPED_NODE))) { atomic_inc(&tgtp->xmt_abort_rsp_error); lpfc_printf_log(phba, KERN_WARNING, LOG_NVME_ABTS, - "6160 Drop ABTS - wrong NDLP state x%x.\n", + "6160 Drop ABORT - wrong NDLP state x%x.\n", (ndlp) ? ndlp->nlp_state : NLP_STE_MAX_STATE); /* No failure to an ABTS request. */ + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; return 0; } @@ -1900,9 +2105,10 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, ctxp->abort_wqeq = lpfc_sli_get_iocbq(phba); if (!ctxp->abort_wqeq) { lpfc_printf_log(phba, KERN_WARNING, LOG_NVME_ABTS, - "6161 Abort failed: No wqeqs: " + "6161 ABORT failed: No wqeqs: " "xri: x%x\n", ctxp->oxid); /* No failure to an ABTS request. */ + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; return 0; } abts_wqeq = ctxp->abort_wqeq; @@ -1910,8 +2116,8 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, ctxp->state = LPFC_NVMET_STE_ABORT; /* Announce entry to new IO submit field. */ - lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, - "6162 Abort Request to rport DID x%06x " + lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS, + "6162 ABORT Request to rport DID x%06x " "for xri x%x x%x\n", ctxp->sid, ctxp->oxid, ctxp->wqeq->sli4_xritag); @@ -1927,6 +2133,7 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, "NVME Req now. hba_flag x%x oxid x%x\n", phba->hba_flag, ctxp->oxid); lpfc_sli_release_iocbq(phba, abts_wqeq); + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; return 0; } @@ -1938,6 +2145,7 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, "still pending on oxid x%x\n", ctxp->oxid); lpfc_sli_release_iocbq(phba, abts_wqeq); + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; return 0; } @@ -1985,9 +2193,10 @@ lpfc_nvmet_sol_fcp_issue_abort(struct lpfc_hba *phba, if (rc == WQE_SUCCESS) return 0; + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; lpfc_sli_release_iocbq(phba, abts_wqeq); - lpfc_printf_log(phba, KERN_ERR, LOG_NVME, - "6166 Failed abts issue_wqe with status x%x " + lpfc_printf_log(phba, KERN_ERR, LOG_NVME_ABTS, + "6166 Failed ABORT issue_wqe with status x%x " "for oxid x%x.\n", rc, ctxp->oxid); return 1; @@ -2016,8 +2225,8 @@ lpfc_nvmet_unsol_fcp_issue_abort(struct lpfc_hba *phba, spin_lock_irqsave(&phba->hbalock, flags); abts_wqeq = ctxp->wqeq; - abts_wqeq->wqe_cmpl = lpfc_nvmet_xmt_fcp_abort_cmp; - abts_wqeq->iocb_cmpl = 0; + abts_wqeq->wqe_cmpl = lpfc_nvmet_unsol_fcp_abort_cmp; + abts_wqeq->iocb_cmpl = NULL; abts_wqeq->iocb_flag |= LPFC_IO_NVMET; rc = lpfc_sli4_issue_wqe(phba, LPFC_FCP_RING, abts_wqeq); spin_unlock_irqrestore(&phba->hbalock, flags); @@ -2027,7 +2236,7 @@ lpfc_nvmet_unsol_fcp_issue_abort(struct lpfc_hba *phba, } aerr: - lpfc_nvmet_rq_post(phba, ctxp, &ctxp->rqb_buffer->hbuf); + ctxp->flag &= ~LPFC_NVMET_ABORT_OP; atomic_inc(&tgtp->xmt_abort_rsp_error); lpfc_printf_log(phba, KERN_WARNING, LOG_NVME_ABTS, "6135 Failed to Issue ABTS for oxid x%x. Status x%x\n", diff --git a/drivers/scsi/lpfc/lpfc_nvmet.h b/drivers/scsi/lpfc/lpfc_nvmet.h index 02735fc6fd41..128759fe6650 100644 --- a/drivers/scsi/lpfc/lpfc_nvmet.h +++ b/drivers/scsi/lpfc/lpfc_nvmet.h @@ -21,9 +21,7 @@ * included with this package. * ********************************************************************/ -#define LPFC_NVMET_MIN_SEGS 16 -#define LPFC_NVMET_DEFAULT_SEGS 64 /* 256K IOs */ -#define LPFC_NVMET_MAX_SEGS 510 +#define LPFC_NVMET_DEFAULT_SEGS (64 + 1) /* 256K IOs */ #define LPFC_NVMET_SUCCESS_LEN 12 /* Used for NVME Target */ @@ -77,6 +75,7 @@ struct lpfc_nvmet_rcv_ctx { struct nvmefc_tgt_ls_req ls_req; struct nvmefc_tgt_fcp_req fcp_req; } ctx; + struct list_head list; struct lpfc_hba *phba; struct lpfc_iocbq *wqeq; struct lpfc_iocbq *abort_wqeq; @@ -98,10 +97,11 @@ struct lpfc_nvmet_rcv_ctx { #define LPFC_NVMET_STE_RSP 4 #define LPFC_NVMET_STE_DONE 5 uint16_t flag; -#define LPFC_NVMET_IO_INP 0x1 -#define LPFC_NVMET_ABORT_OP 0x2 -#define LPFC_NVMET_CTX_RLS 0x4 - +#define LPFC_NVMET_IO_INP 0x1 /* IO is in progress on exchange */ +#define LPFC_NVMET_ABORT_OP 0x2 /* Abort WQE issued on exchange */ +#define LPFC_NVMET_XBUSY 0x4 /* XB bit set on IO cmpl */ +#define LPFC_NVMET_CTX_RLS 0x8 /* ctx free requested */ +#define LPFC_NVMET_ABTS_RCV 0x10 /* ABTS received on exchange */ struct rqb_dmabuf *rqb_buffer; #ifdef CONFIG_SCSI_LPFC_DEBUG_FS diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index 1c9fa45df7eb..cf19f4976f5f 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -6338,7 +6338,7 @@ lpfc_sli4_get_allocated_extnts(struct lpfc_hba *phba, uint16_t type, } /** - * lpfc_sli4_repost_sgl_list - Repsot the buffers sgl pages as block + * lpfc_sli4_repost_sgl_list - Repost the buffers sgl pages as block * @phba: pointer to lpfc hba data structure. * @pring: Pointer to driver SLI ring object. * @sgl_list: linked link of sgl buffers to post @@ -13758,7 +13758,10 @@ lpfc_sli4_queue_free(struct lpfc_queue *queue) lpfc_free_rq_buffer(queue->phba, queue); kfree(queue->rqbp); } - kfree(queue->pring); + + if (!list_empty(&queue->wq_list)) + list_del(&queue->wq_list); + kfree(queue); return; } @@ -14738,6 +14741,9 @@ lpfc_wq_create(struct lpfc_hba *phba, struct lpfc_queue *wq, case LPFC_Q_CREATE_VERSION_1: bf_set(lpfc_mbx_wq_create_wqe_count, &wq_create->u.request_1, wq->entry_count); + bf_set(lpfc_mbox_hdr_version, &shdr->request, + LPFC_Q_CREATE_VERSION_1); + switch (wq->entry_size) { default: case 64: @@ -15561,6 +15567,8 @@ lpfc_wq_destroy(struct lpfc_hba *phba, struct lpfc_queue *wq) } /* Remove wq from any list */ list_del_init(&wq->list); + kfree(wq->pring); + wq->pring = NULL; mempool_free(mbox, wq->phba->mbox_mem_pool); return status; } @@ -16513,7 +16521,7 @@ lpfc_sli4_xri_inrange(struct lpfc_hba *phba, * This function sends a basic response to a previous unsol sequence abort * event after aborting the sequence handling. **/ -static void +void lpfc_sli4_seq_abort_rsp(struct lpfc_vport *vport, struct fc_frame_header *fc_hdr, bool aborted) { @@ -16534,14 +16542,13 @@ lpfc_sli4_seq_abort_rsp(struct lpfc_vport *vport, ndlp = lpfc_findnode_did(vport, sid); if (!ndlp) { - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, sid); if (!ndlp) { lpfc_printf_vlog(vport, KERN_WARNING, LOG_ELS, "1268 Failed to allocate ndlp for " "oxid:x%x SID:x%x\n", oxid, sid); return; } - lpfc_nlp_init(vport, ndlp, sid); /* Put ndlp onto pport node list */ lpfc_enqueue_node(vport, ndlp); } else if (!NLP_CHK_NODE_ACT(ndlp)) { @@ -16690,6 +16697,11 @@ lpfc_sli4_handle_unsol_abort(struct lpfc_vport *vport, } lpfc_in_buf_free(phba, &dmabuf->dbuf); + if (phba->nvmet_support) { + lpfc_nvmet_rcv_unsol_abort(vport, &fc_hdr); + return; + } + /* Respond with BA_ACC or BA_RJT accordingly */ lpfc_sli4_seq_abort_rsp(vport, &fc_hdr, aborted); } diff --git a/drivers/scsi/lpfc/lpfc_sli4.h b/drivers/scsi/lpfc/lpfc_sli4.h index 710458cf11d6..da46471337c8 100644 --- a/drivers/scsi/lpfc/lpfc_sli4.h +++ b/drivers/scsi/lpfc/lpfc_sli4.h @@ -620,7 +620,7 @@ struct lpfc_sli4_hba { struct list_head lpfc_els_sgl_list; struct list_head lpfc_abts_els_sgl_list; struct list_head lpfc_nvmet_sgl_list; - struct list_head lpfc_abts_nvmet_sgl_list; + struct list_head lpfc_abts_nvmet_ctx_list; struct list_head lpfc_abts_scsi_buf_list; struct list_head lpfc_abts_nvme_buf_list; struct lpfc_sglq **lpfc_sglq_active_list; diff --git a/drivers/scsi/lpfc/lpfc_version.h b/drivers/scsi/lpfc/lpfc_version.h index d4e95e28f4e3..1c26dc67151b 100644 --- a/drivers/scsi/lpfc/lpfc_version.h +++ b/drivers/scsi/lpfc/lpfc_version.h @@ -20,7 +20,7 @@ * included with this package. * *******************************************************************/ -#define LPFC_DRIVER_VERSION "11.2.0.10" +#define LPFC_DRIVER_VERSION "11.2.0.12" #define LPFC_DRIVER_NAME "lpfc" /* Used for SLI 2/3 */ diff --git a/drivers/scsi/lpfc/lpfc_vport.c b/drivers/scsi/lpfc/lpfc_vport.c index 9a0339dbc024..c714482bf4c5 100644 --- a/drivers/scsi/lpfc/lpfc_vport.c +++ b/drivers/scsi/lpfc/lpfc_vport.c @@ -738,10 +738,9 @@ lpfc_vport_delete(struct fc_vport *fc_vport) ndlp = lpfc_findnode_did(vport, Fabric_DID); if (!ndlp) { /* Cannot find existing Fabric ndlp, allocate one */ - ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); + ndlp = lpfc_nlp_init(vport, Fabric_DID); if (!ndlp) goto skip_logo; - lpfc_nlp_init(vport, ndlp, Fabric_DID); /* Indicate free memory when release */ NLP_SET_FREE_REQ(ndlp); } else {