From 0fb228d30b8d72bfee51f57e638d412324d44a11 Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 1 Aug 2017 15:12:39 -0700 Subject: [PATCH] nvmet_fc: add defer_req callback for deferment of cmd buffer return At queue creation, the transport allocates a local job struct (struct nvmet_fc_fcp_iod) for each possible element of the queue. When a new CMD is received from the wire, a jobs struct is allocated from the queue and then used for the duration of the command. The job struct contains buffer space for the wire command iu. Thus, upon allocation of the job struct, the cmd iu buffer is copied to the job struct and the LLDD may immediately free/reuse the CMD IU buffer passed in the call. However, in some circumstances, due to the packetized nature of FC and the api of the FC LLDD which may issue a hw command to send the wire response, but the LLDD may not get the hw completion for the command and upcall the nvmet_fc layer before a new command may be asynchronously received on the wire. In other words, its possible for the initiator to get the response from the wire, thus believe a command slot free, and send a new command iu. The new command iu may be received by the LLDD and passed to the transport before the LLDD had serviced the hw completion and made the teardown calls for the original job struct. As such, there is no available job struct available for the new io. E.g. it appears like the host sent more queue elements than the queue size. It didn't based on it's understanding. Rather than treat this as a hard connection failure queue the new request until the job struct does free up. As the buffer isn't copied as there's no job struct, a special return value must be returned to the LLDD to signify to hold off on recycling the cmd iu buffer. And later, when a job struct is allocated and the buffer copied, a new LLDD callback is introduced to notify the LLDD and allow it to recycle it's command iu buffer. Signed-off-by: James Smart Reviewed-by: Johannes Thumshirn Signed-off-by: Christoph Hellwig --- drivers/nvme/target/fc.c | 212 ++++++++++++++++++++++++++++----- include/linux/nvme-fc-driver.h | 7 ++ 2 files changed, 191 insertions(+), 28 deletions(-) diff --git a/drivers/nvme/target/fc.c b/drivers/nvme/target/fc.c index 31ca55dfcb1d..1b7f2520a20d 100644 --- a/drivers/nvme/target/fc.c +++ b/drivers/nvme/target/fc.c @@ -114,6 +114,11 @@ struct nvmet_fc_tgtport { struct kref ref; }; +struct nvmet_fc_defer_fcp_req { + struct list_head req_list; + struct nvmefc_tgt_fcp_req *fcp_req; +}; + struct nvmet_fc_tgt_queue { bool ninetypercent; u16 qid; @@ -132,6 +137,8 @@ struct nvmet_fc_tgt_queue { struct nvmet_fc_tgt_assoc *assoc; struct nvmet_fc_fcp_iod *fod; /* array of fcp_iods */ struct list_head fod_list; + struct list_head pending_cmd_list; + struct list_head avail_defer_list; struct workqueue_struct *work_q; struct kref ref; } __aligned(sizeof(unsigned long long)); @@ -223,6 +230,8 @@ static void nvmet_fc_tgt_q_put(struct nvmet_fc_tgt_queue *queue); static int nvmet_fc_tgt_q_get(struct nvmet_fc_tgt_queue *queue); static void nvmet_fc_tgtport_put(struct nvmet_fc_tgtport *tgtport); static int nvmet_fc_tgtport_get(struct nvmet_fc_tgtport *tgtport); +static void nvmet_fc_handle_fcp_rqst(struct nvmet_fc_tgtport *tgtport, + struct nvmet_fc_fcp_iod *fod); /* *********************** FC-NVME DMA Handling **************************** */ @@ -463,9 +472,9 @@ static struct nvmet_fc_fcp_iod * nvmet_fc_alloc_fcp_iod(struct nvmet_fc_tgt_queue *queue) { static struct nvmet_fc_fcp_iod *fod; - unsigned long flags; - spin_lock_irqsave(&queue->qlock, flags); + lockdep_assert_held(&queue->qlock); + fod = list_first_entry_or_null(&queue->fod_list, struct nvmet_fc_fcp_iod, fcp_list); if (fod) { @@ -477,17 +486,37 @@ nvmet_fc_alloc_fcp_iod(struct nvmet_fc_tgt_queue *queue) * will "inherit" that reference. */ } - spin_unlock_irqrestore(&queue->qlock, flags); return fod; } +static void +nvmet_fc_queue_fcp_req(struct nvmet_fc_tgtport *tgtport, + struct nvmet_fc_tgt_queue *queue, + struct nvmefc_tgt_fcp_req *fcpreq) +{ + struct nvmet_fc_fcp_iod *fod = fcpreq->nvmet_fc_private; + + /* + * put all admin cmds on hw queue id 0. All io commands go to + * the respective hw queue based on a modulo basis + */ + fcpreq->hwqid = queue->qid ? + ((queue->qid - 1) % tgtport->ops->max_hw_queues) : 0; + + if (tgtport->ops->target_features & NVMET_FCTGTFEAT_CMD_IN_ISR) + queue_work_on(queue->cpu, queue->work_q, &fod->work); + else + nvmet_fc_handle_fcp_rqst(tgtport, fod); +} + static void nvmet_fc_free_fcp_iod(struct nvmet_fc_tgt_queue *queue, struct nvmet_fc_fcp_iod *fod) { struct nvmefc_tgt_fcp_req *fcpreq = fod->fcpreq; struct nvmet_fc_tgtport *tgtport = fod->tgtport; + struct nvmet_fc_defer_fcp_req *deferfcp; unsigned long flags; fc_dma_sync_single_for_cpu(tgtport->dev, fod->rspdma, @@ -495,21 +524,56 @@ nvmet_fc_free_fcp_iod(struct nvmet_fc_tgt_queue *queue, fcpreq->nvmet_fc_private = NULL; - spin_lock_irqsave(&queue->qlock, flags); - list_add_tail(&fod->fcp_list, &fod->queue->fod_list); fod->active = false; fod->abort = false; fod->aborted = false; fod->writedataactive = false; fod->fcpreq = NULL; + + tgtport->ops->fcp_req_release(&tgtport->fc_target_port, fcpreq); + + spin_lock_irqsave(&queue->qlock, flags); + deferfcp = list_first_entry_or_null(&queue->pending_cmd_list, + struct nvmet_fc_defer_fcp_req, req_list); + if (!deferfcp) { + list_add_tail(&fod->fcp_list, &fod->queue->fod_list); + spin_unlock_irqrestore(&queue->qlock, flags); + + /* Release reference taken at queue lookup and fod allocation */ + nvmet_fc_tgt_q_put(queue); + return; + } + + /* Re-use the fod for the next pending cmd that was deferred */ + list_del(&deferfcp->req_list); + + fcpreq = deferfcp->fcp_req; + + /* deferfcp can be reused for another IO at a later date */ + list_add_tail(&deferfcp->req_list, &queue->avail_defer_list); + spin_unlock_irqrestore(&queue->qlock, flags); + /* Save NVME CMD IO in fod */ + memcpy(&fod->cmdiubuf, fcpreq->rspaddr, fcpreq->rsplen); + + /* Setup new fcpreq to be processed */ + fcpreq->rspaddr = NULL; + fcpreq->rsplen = 0; + fcpreq->nvmet_fc_private = fod; + fod->fcpreq = fcpreq; + fod->active = true; + + /* inform LLDD IO is now being processed */ + tgtport->ops->defer_rcv(&tgtport->fc_target_port, fcpreq); + + /* Submit deferred IO for processing */ + nvmet_fc_queue_fcp_req(tgtport, queue, fcpreq); + /* - * release the reference taken at queue lookup and fod allocation + * Leave the queue lookup get reference taken when + * fod was originally allocated. */ - nvmet_fc_tgt_q_put(queue); - - tgtport->ops->fcp_req_release(&tgtport->fc_target_port, fcpreq); } static int @@ -569,6 +633,8 @@ nvmet_fc_alloc_target_queue(struct nvmet_fc_tgt_assoc *assoc, queue->port = assoc->tgtport->port; queue->cpu = nvmet_fc_queue_to_cpu(assoc->tgtport, qid); INIT_LIST_HEAD(&queue->fod_list); + INIT_LIST_HEAD(&queue->avail_defer_list); + INIT_LIST_HEAD(&queue->pending_cmd_list); atomic_set(&queue->connected, 0); atomic_set(&queue->sqtail, 0); atomic_set(&queue->rsn, 1); @@ -638,6 +704,7 @@ nvmet_fc_delete_target_queue(struct nvmet_fc_tgt_queue *queue) { struct nvmet_fc_tgtport *tgtport = queue->assoc->tgtport; struct nvmet_fc_fcp_iod *fod = queue->fod; + struct nvmet_fc_defer_fcp_req *deferfcp; unsigned long flags; int i, writedataactive; bool disconnect; @@ -666,6 +733,35 @@ nvmet_fc_delete_target_queue(struct nvmet_fc_tgt_queue *queue) } } } + + /* Cleanup defer'ed IOs in queue */ + list_for_each_entry(deferfcp, &queue->avail_defer_list, req_list) { + list_del(&deferfcp->req_list); + kfree(deferfcp); + } + + for (;;) { + deferfcp = list_first_entry_or_null(&queue->pending_cmd_list, + struct nvmet_fc_defer_fcp_req, req_list); + if (!deferfcp) + break; + + list_del(&deferfcp->req_list); + spin_unlock_irqrestore(&queue->qlock, flags); + + tgtport->ops->defer_rcv(&tgtport->fc_target_port, + deferfcp->fcp_req); + + tgtport->ops->fcp_abort(&tgtport->fc_target_port, + deferfcp->fcp_req); + + tgtport->ops->fcp_req_release(&tgtport->fc_target_port, + deferfcp->fcp_req); + + kfree(deferfcp); + + spin_lock_irqsave(&queue->qlock, flags); + } spin_unlock_irqrestore(&queue->qlock, flags); flush_workqueue(queue->work_q); @@ -2172,11 +2268,38 @@ nvmet_fc_handle_fcp_rqst_work(struct work_struct *work) * Pass a FC-NVME FCP CMD IU received from the FC link to the nvmet-fc * layer for processing. * - * The nvmet-fc layer will copy cmd payload to an internal structure for - * processing. As such, upon completion of the routine, the LLDD may - * immediately free/reuse the CMD IU buffer passed in the call. + * The nvmet_fc layer allocates a local job structure (struct + * nvmet_fc_fcp_iod) from the queue for the io and copies the + * CMD IU buffer to the job structure. As such, on a successful + * completion (returns 0), the LLDD may immediately free/reuse + * the CMD IU buffer passed in the call. + * + * However, in some circumstances, due to the packetized nature of FC + * and the api of the FC LLDD which may issue a hw command to send the + * response, but the LLDD may not get the hw completion for that command + * and upcall the nvmet_fc layer before a new command may be + * asynchronously received - its possible for a command to be received + * before the LLDD and nvmet_fc have recycled the job structure. It gives + * the appearance of more commands received than fits in the sq. + * To alleviate this scenario, a temporary queue is maintained in the + * transport for pending LLDD requests waiting for a queue job structure. + * In these "overrun" cases, a temporary queue element is allocated + * the LLDD request and CMD iu buffer information remembered, and the + * routine returns a -EOVERFLOW status. Subsequently, when a queue job + * structure is freed, it is immediately reallocated for anything on the + * pending request list. The LLDDs defer_rcv() callback is called, + * informing the LLDD that it may reuse the CMD IU buffer, and the io + * is then started normally with the transport. * - * If this routine returns error, the lldd should abort the exchange. + * The LLDD, when receiving an -EOVERFLOW completion status, is to treat + * the completion as successful but must not reuse the CMD IU buffer + * until the LLDD's defer_rcv() callback has been called for the + * corresponding struct nvmefc_tgt_fcp_req pointer. + * + * If there is any other condition in which an error occurs, the + * transport will return a non-zero status indicating the error. + * In all cases other than -EOVERFLOW, the transport has not accepted the + * request and the LLDD should abort the exchange. * * @target_port: pointer to the (registered) target port the FCP CMD IU * was received on. @@ -2194,6 +2317,8 @@ nvmet_fc_rcv_fcp_req(struct nvmet_fc_target_port *target_port, struct nvme_fc_cmd_iu *cmdiu = cmdiubuf; struct nvmet_fc_tgt_queue *queue; struct nvmet_fc_fcp_iod *fod; + struct nvmet_fc_defer_fcp_req *deferfcp; + unsigned long flags; /* validate iu, so the connection id can be used to find the queue */ if ((cmdiubuf_len != sizeof(*cmdiu)) || @@ -2214,29 +2339,60 @@ nvmet_fc_rcv_fcp_req(struct nvmet_fc_target_port *target_port, * when the fod is freed. */ + spin_lock_irqsave(&queue->qlock, flags); + fod = nvmet_fc_alloc_fcp_iod(queue); - if (!fod) { + if (fod) { + spin_unlock_irqrestore(&queue->qlock, flags); + + fcpreq->nvmet_fc_private = fod; + fod->fcpreq = fcpreq; + + memcpy(&fod->cmdiubuf, cmdiubuf, cmdiubuf_len); + + nvmet_fc_queue_fcp_req(tgtport, queue, fcpreq); + + return 0; + } + + if (!tgtport->ops->defer_rcv) { + spin_unlock_irqrestore(&queue->qlock, flags); /* release the queue lookup reference */ nvmet_fc_tgt_q_put(queue); return -ENOENT; } - fcpreq->nvmet_fc_private = fod; - fod->fcpreq = fcpreq; - /* - * put all admin cmds on hw queue id 0. All io commands go to - * the respective hw queue based on a modulo basis - */ - fcpreq->hwqid = queue->qid ? - ((queue->qid - 1) % tgtport->ops->max_hw_queues) : 0; - memcpy(&fod->cmdiubuf, cmdiubuf, cmdiubuf_len); + deferfcp = list_first_entry_or_null(&queue->avail_defer_list, + struct nvmet_fc_defer_fcp_req, req_list); + if (deferfcp) { + /* Just re-use one that was previously allocated */ + list_del(&deferfcp->req_list); + } else { + spin_unlock_irqrestore(&queue->qlock, flags); - if (tgtport->ops->target_features & NVMET_FCTGTFEAT_CMD_IN_ISR) - queue_work_on(queue->cpu, queue->work_q, &fod->work); - else - nvmet_fc_handle_fcp_rqst(tgtport, fod); + /* Now we need to dynamically allocate one */ + deferfcp = kmalloc(sizeof(*deferfcp), GFP_KERNEL); + if (!deferfcp) { + /* release the queue lookup reference */ + nvmet_fc_tgt_q_put(queue); + return -ENOMEM; + } + spin_lock_irqsave(&queue->qlock, flags); + } - return 0; + /* For now, use rspaddr / rsplen to save payload information */ + fcpreq->rspaddr = cmdiubuf; + fcpreq->rsplen = cmdiubuf_len; + deferfcp->fcp_req = fcpreq; + + /* defer processing till a fod becomes available */ + list_add_tail(&deferfcp->req_list, &queue->pending_cmd_list); + + /* NOTE: the queue lookup reference is still valid */ + + spin_unlock_irqrestore(&queue->qlock, flags); + + return -EOVERFLOW; } EXPORT_SYMBOL_GPL(nvmet_fc_rcv_fcp_req); diff --git a/include/linux/nvme-fc-driver.h b/include/linux/nvme-fc-driver.h index 6c8c5d8041b7..2591878c1d48 100644 --- a/include/linux/nvme-fc-driver.h +++ b/include/linux/nvme-fc-driver.h @@ -346,6 +346,11 @@ struct nvme_fc_remote_port { * indicating an FC transport Aborted status. * Entrypoint is Mandatory. * + * @defer_rcv: Called by the transport to signal the LLLD that it has + * begun processing of a previously received NVME CMD IU. The LLDD + * is now free to re-use the rcv buffer associated with the + * nvmefc_tgt_fcp_req. + * * @max_hw_queues: indicates the maximum number of hw queues the LLDD * supports for cpu affinitization. * Value is Mandatory. Must be at least 1. @@ -846,6 +851,8 @@ struct nvmet_fc_target_template { struct nvmefc_tgt_fcp_req *fcpreq); void (*fcp_req_release)(struct nvmet_fc_target_port *tgtport, struct nvmefc_tgt_fcp_req *fcpreq); + void (*defer_rcv)(struct nvmet_fc_target_port *tgtport, + struct nvmefc_tgt_fcp_req *fcpreq); u32 max_hw_queues; u16 max_sgl_segments; -- 2.20.1