diff options
author | Linus Torvalds <torvalds@g5.osdl.org> | 2005-11-04 16:32:36 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@g5.osdl.org> | 2005-11-04 16:32:36 -0800 |
commit | 127f2fa31ac624c744f3767363c4919209980956 (patch) | |
tree | 972b7ffbcd2c846bbb34879efb91e324dc0677be | |
parent | ba77df570c6710c9c19b31e0e48e4bcdf31cefe8 (diff) | |
parent | aef9ec39c47f0cece886ddd6b53c440321e0b2a6 (diff) |
Merge branch 'srp' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
-rw-r--r-- | drivers/infiniband/Kconfig | 2 | ||||
-rw-r--r-- | drivers/infiniband/Makefile | 1 | ||||
-rw-r--r-- | drivers/infiniband/ulp/srp/Kbuild | 1 | ||||
-rw-r--r-- | drivers/infiniband/ulp/srp/Kconfig | 11 | ||||
-rw-r--r-- | drivers/infiniband/ulp/srp/ib_srp.c | 1700 | ||||
-rw-r--r-- | drivers/infiniband/ulp/srp/ib_srp.h | 150 | ||||
-rw-r--r-- | include/scsi/srp.h | 226 |
7 files changed, 2091 insertions, 0 deletions
diff --git a/drivers/infiniband/Kconfig b/drivers/infiniband/Kconfig index 325d502e25cd..bdf0891a92dd 100644 --- a/drivers/infiniband/Kconfig +++ b/drivers/infiniband/Kconfig @@ -33,4 +33,6 @@ source "drivers/infiniband/hw/mthca/Kconfig" source "drivers/infiniband/ulp/ipoib/Kconfig" +source "drivers/infiniband/ulp/srp/Kconfig" + endmenu diff --git a/drivers/infiniband/Makefile b/drivers/infiniband/Makefile index d256cf798218..a43fb34cca94 100644 --- a/drivers/infiniband/Makefile +++ b/drivers/infiniband/Makefile @@ -1,3 +1,4 @@ obj-$(CONFIG_INFINIBAND) += core/ obj-$(CONFIG_INFINIBAND_MTHCA) += hw/mthca/ obj-$(CONFIG_INFINIBAND_IPOIB) += ulp/ipoib/ +obj-$(CONFIG_INFINIBAND_SRP) += ulp/srp/ diff --git a/drivers/infiniband/ulp/srp/Kbuild b/drivers/infiniband/ulp/srp/Kbuild new file mode 100644 index 000000000000..a16c73c667cb --- /dev/null +++ b/drivers/infiniband/ulp/srp/Kbuild @@ -0,0 +1 @@ +obj-$(CONFIG_INFINIBAND_SRP) += ib_srp.o diff --git a/drivers/infiniband/ulp/srp/Kconfig b/drivers/infiniband/ulp/srp/Kconfig new file mode 100644 index 000000000000..8fe3be4e9910 --- /dev/null +++ b/drivers/infiniband/ulp/srp/Kconfig @@ -0,0 +1,11 @@ +config INFINIBAND_SRP + tristate "InfiniBand SCSI RDMA Protocol" + depends on INFINIBAND && SCSI + ---help--- + Support for the SCSI RDMA Protocol over InfiniBand. This + allows you to access storage devices that speak SRP over + InfiniBand. + + The SRP protocol is defined by the INCITS T10 technical + committee. See <http://www.t10.org/>. + diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c new file mode 100644 index 000000000000..2687e34aa5bc --- /dev/null +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -0,0 +1,1700 @@ +/* + * Copyright (c) 2005 Cisco Systems. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + * $Id: ib_srp.c 3932 2005-11-01 17:19:29Z roland $ + */ + +#include <linux/version.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/err.h> +#include <linux/string.h> +#include <linux/parser.h> +#include <linux/random.h> + +#include <asm/atomic.h> + +#include <scsi/scsi.h> +#include <scsi/scsi_device.h> +#include <scsi/scsi_dbg.h> +#include <scsi/srp.h> + +#include <rdma/ib_cache.h> + +#include "ib_srp.h" + +#define DRV_NAME "ib_srp" +#define PFX DRV_NAME ": " +#define DRV_VERSION "0.2" +#define DRV_RELDATE "November 1, 2005" + +MODULE_AUTHOR("Roland Dreier"); +MODULE_DESCRIPTION("InfiniBand SCSI RDMA Protocol initiator " + "v" DRV_VERSION " (" DRV_RELDATE ")"); +MODULE_LICENSE("Dual BSD/GPL"); + +static int topspin_workarounds = 1; + +module_param(topspin_workarounds, int, 0444); +MODULE_PARM_DESC(topspin_workarounds, + "Enable workarounds for Topspin/Cisco SRP target bugs if != 0"); + +static const u8 topspin_oui[3] = { 0x00, 0x05, 0xad }; + +static void srp_add_one(struct ib_device *device); +static void srp_remove_one(struct ib_device *device); +static void srp_completion(struct ib_cq *cq, void *target_ptr); +static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event); + +static struct ib_client srp_client = { + .name = "srp", + .add = srp_add_one, + .remove = srp_remove_one +}; + +static inline struct srp_target_port *host_to_target(struct Scsi_Host *host) +{ + return (struct srp_target_port *) host->hostdata; +} + +static const char *srp_target_info(struct Scsi_Host *host) +{ + return host_to_target(host)->target_name; +} + +static struct srp_iu *srp_alloc_iu(struct srp_host *host, size_t size, + gfp_t gfp_mask, + enum dma_data_direction direction) +{ + struct srp_iu *iu; + + iu = kmalloc(sizeof *iu, gfp_mask); + if (!iu) + goto out; + + iu->buf = kzalloc(size, gfp_mask); + if (!iu->buf) + goto out_free_iu; + + iu->dma = dma_map_single(host->dev->dma_device, iu->buf, size, direction); + if (dma_mapping_error(iu->dma)) + goto out_free_buf; + + iu->size = size; + iu->direction = direction; + + return iu; + +out_free_buf: + kfree(iu->buf); +out_free_iu: + kfree(iu); +out: + return NULL; +} + +static void srp_free_iu(struct srp_host *host, struct srp_iu *iu) +{ + if (!iu) + return; + + dma_unmap_single(host->dev->dma_device, iu->dma, iu->size, iu->direction); + kfree(iu->buf); + kfree(iu); +} + +static void srp_qp_event(struct ib_event *event, void *context) +{ + printk(KERN_ERR PFX "QP event %d\n", event->event); +} + +static int srp_init_qp(struct srp_target_port *target, + struct ib_qp *qp) +{ + struct ib_qp_attr *attr; + int ret; + + attr = kmalloc(sizeof *attr, GFP_KERNEL); + if (!attr) + return -ENOMEM; + + ret = ib_find_cached_pkey(target->srp_host->dev, + target->srp_host->port, + be16_to_cpu(target->path.pkey), + &attr->pkey_index); + if (ret) + goto out; + + attr->qp_state = IB_QPS_INIT; + attr->qp_access_flags = (IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE); + attr->port_num = target->srp_host->port; + + ret = ib_modify_qp(qp, attr, + IB_QP_STATE | + IB_QP_PKEY_INDEX | + IB_QP_ACCESS_FLAGS | + IB_QP_PORT); + +out: + kfree(attr); + return ret; +} + +static int srp_create_target_ib(struct srp_target_port *target) +{ + struct ib_qp_init_attr *init_attr; + int ret; + + init_attr = kzalloc(sizeof *init_attr, GFP_KERNEL); + if (!init_attr) + return -ENOMEM; + + target->cq = ib_create_cq(target->srp_host->dev, srp_completion, + NULL, target, SRP_CQ_SIZE); + if (IS_ERR(target->cq)) { + ret = PTR_ERR(target->cq); + goto out; + } + + ib_req_notify_cq(target->cq, IB_CQ_NEXT_COMP); + + init_attr->event_handler = srp_qp_event; + init_attr->cap.max_send_wr = SRP_SQ_SIZE; + init_attr->cap.max_recv_wr = SRP_RQ_SIZE; + init_attr->cap.max_recv_sge = 1; + init_attr->cap.max_send_sge = 1; + init_attr->sq_sig_type = IB_SIGNAL_ALL_WR; + init_attr->qp_type = IB_QPT_RC; + init_attr->send_cq = target->cq; + init_attr->recv_cq = target->cq; + + target->qp = ib_create_qp(target->srp_host->pd, init_attr); + if (IS_ERR(target->qp)) { + ret = PTR_ERR(target->qp); + ib_destroy_cq(target->cq); + goto out; + } + + ret = srp_init_qp(target, target->qp); + if (ret) { + ib_destroy_qp(target->qp); + ib_destroy_cq(target->cq); + goto out; + } + +out: + kfree(init_attr); + return ret; +} + +static void srp_free_target_ib(struct srp_target_port *target) +{ + int i; + + ib_destroy_qp(target->qp); + ib_destroy_cq(target->cq); + + for (i = 0; i < SRP_RQ_SIZE; ++i) + srp_free_iu(target->srp_host, target->rx_ring[i]); + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) + srp_free_iu(target->srp_host, target->tx_ring[i]); +} + +static void srp_path_rec_completion(int status, + struct ib_sa_path_rec *pathrec, + void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + target->status = status; + if (status) + printk(KERN_ERR PFX "Got failed path rec status %d\n", status); + else + target->path = *pathrec; + complete(&target->done); +} + +static int srp_lookup_path(struct srp_target_port *target) +{ + target->path.numb_path = 1; + + init_completion(&target->done); + + target->path_query_id = ib_sa_path_rec_get(target->srp_host->dev, + target->srp_host->port, + &target->path, + IB_SA_PATH_REC_DGID | + IB_SA_PATH_REC_SGID | + IB_SA_PATH_REC_NUMB_PATH | + IB_SA_PATH_REC_PKEY, + SRP_PATH_REC_TIMEOUT_MS, + GFP_KERNEL, + srp_path_rec_completion, + target, &target->path_query); + if (target->path_query_id < 0) + return target->path_query_id; + + wait_for_completion(&target->done); + + if (target->status < 0) + printk(KERN_WARNING PFX "Path record query failed\n"); + + return target->status; +} + +static int srp_send_req(struct srp_target_port *target) +{ + struct { + struct ib_cm_req_param param; + struct srp_login_req priv; + } *req = NULL; + int status; + + req = kzalloc(sizeof *req, GFP_KERNEL); + if (!req) + return -ENOMEM; + + req->param.primary_path = &target->path; + req->param.alternate_path = NULL; + req->param.service_id = target->service_id; + req->param.qp_num = target->qp->qp_num; + req->param.qp_type = target->qp->qp_type; + req->param.private_data = &req->priv; + req->param.private_data_len = sizeof req->priv; + req->param.flow_control = 1; + + get_random_bytes(&req->param.starting_psn, 4); + req->param.starting_psn &= 0xffffff; + + /* + * Pick some arbitrary defaults here; we could make these + * module parameters if anyone cared about setting them. + */ + req->param.responder_resources = 4; + req->param.remote_cm_response_timeout = 20; + req->param.local_cm_response_timeout = 20; + req->param.retry_count = 7; + req->param.rnr_retry_count = 7; + req->param.max_cm_retries = 15; + + req->priv.opcode = SRP_LOGIN_REQ; + req->priv.tag = 0; + req->priv.req_it_iu_len = cpu_to_be32(SRP_MAX_IU_LEN); + req->priv.req_buf_fmt = cpu_to_be16(SRP_BUF_FORMAT_DIRECT | + SRP_BUF_FORMAT_INDIRECT); + memcpy(req->priv.initiator_port_id, target->srp_host->initiator_port_id, 16); + /* + * Topspin/Cisco SRP targets will reject our login unless we + * zero out the first 8 bytes of our initiator port ID. The + * second 8 bytes must be our local node GUID, but we always + * use that anyway. + */ + if (topspin_workarounds && !memcmp(&target->ioc_guid, topspin_oui, 3)) { + printk(KERN_DEBUG PFX "Topspin/Cisco initiator port ID workaround " + "activated for target GUID %016llx\n", + (unsigned long long) be64_to_cpu(target->ioc_guid)); + memset(req->priv.initiator_port_id, 0, 8); + } + memcpy(req->priv.target_port_id, &target->id_ext, 8); + memcpy(req->priv.target_port_id + 8, &target->ioc_guid, 8); + + status = ib_send_cm_req(target->cm_id, &req->param); + + kfree(req); + + return status; +} + +static void srp_disconnect_target(struct srp_target_port *target) +{ + /* XXX should send SRP_I_LOGOUT request */ + + init_completion(&target->done); + ib_send_cm_dreq(target->cm_id, NULL, 0); + wait_for_completion(&target->done); +} + +static void srp_remove_work(void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state != SRP_TARGET_DEAD) { + spin_unlock_irq(target->scsi_host->host_lock); + scsi_host_put(target->scsi_host); + return; + } + target->state = SRP_TARGET_REMOVED; + spin_unlock_irq(target->scsi_host->host_lock); + + down(&target->srp_host->target_mutex); + list_del(&target->list); + up(&target->srp_host->target_mutex); + + scsi_remove_host(target->scsi_host); + ib_destroy_cm_id(target->cm_id); + srp_free_target_ib(target); + scsi_host_put(target->scsi_host); + /* And another put to really free the target port... */ + scsi_host_put(target->scsi_host); +} + +static int srp_connect_target(struct srp_target_port *target) +{ + int ret; + + ret = srp_lookup_path(target); + if (ret) + return ret; + + while (1) { + init_completion(&target->done); + ret = srp_send_req(target); + if (ret) + return ret; + wait_for_completion(&target->done); + + /* + * The CM event handling code will set status to + * SRP_PORT_REDIRECT if we get a port redirect REJ + * back, or SRP_DLID_REDIRECT if we get a lid/qp + * redirect REJ back. + */ + switch (target->status) { + case 0: + return 0; + + case SRP_PORT_REDIRECT: + ret = srp_lookup_path(target); + if (ret) + return ret; + break; + + case SRP_DLID_REDIRECT: + break; + + default: + return target->status; + } + } +} + +static int srp_reconnect_target(struct srp_target_port *target) +{ + struct ib_cm_id *new_cm_id; + struct ib_qp_attr qp_attr; + struct srp_request *req; + struct ib_wc wc; + int ret; + int i; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state != SRP_TARGET_LIVE) { + spin_unlock_irq(target->scsi_host->host_lock); + return -EAGAIN; + } + target->state = SRP_TARGET_CONNECTING; + spin_unlock_irq(target->scsi_host->host_lock); + + srp_disconnect_target(target); + /* + * Now get a new local CM ID so that we avoid confusing the + * target in case things are really fouled up. + */ + new_cm_id = ib_create_cm_id(target->srp_host->dev, + srp_cm_handler, target); + if (IS_ERR(new_cm_id)) { + ret = PTR_ERR(new_cm_id); + goto err; + } + ib_destroy_cm_id(target->cm_id); + target->cm_id = new_cm_id; + + qp_attr.qp_state = IB_QPS_RESET; + ret = ib_modify_qp(target->qp, &qp_attr, IB_QP_STATE); + if (ret) + goto err; + + ret = srp_init_qp(target, target->qp); + if (ret) + goto err; + + while (ib_poll_cq(target->cq, 1, &wc) > 0) + ; /* nothing */ + + list_for_each_entry(req, &target->req_queue, list) { + req->scmnd->result = DID_RESET << 16; + req->scmnd->scsi_done(req->scmnd); + } + + target->rx_head = 0; + target->tx_head = 0; + target->tx_tail = 0; + target->req_head = 0; + for (i = 0; i < SRP_SQ_SIZE - 1; ++i) + target->req_ring[i].next = i + 1; + target->req_ring[SRP_SQ_SIZE - 1].next = -1; + INIT_LIST_HEAD(&target->req_queue); + + ret = srp_connect_target(target); + if (ret) + goto err; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state == SRP_TARGET_CONNECTING) { + ret = 0; + target->state = SRP_TARGET_LIVE; + } else + ret = -EAGAIN; + spin_unlock_irq(target->scsi_host->host_lock); + + return ret; + +err: + printk(KERN_ERR PFX "reconnect failed (%d), removing target port.\n", ret); + + /* + * We couldn't reconnect, so kill our target port off. + * However, we have to defer the real removal because we might + * be in the context of the SCSI error handler now, which + * would deadlock if we call scsi_remove_host(). + */ + spin_lock_irq(target->scsi_host->host_lock); + if (target->state == SRP_TARGET_CONNECTING) { + target->state = SRP_TARGET_DEAD; + INIT_WORK(&target->work, srp_remove_work, target); + schedule_work(&target->work); + } + spin_unlock_irq(target->scsi_host->host_lock); + + return ret; +} + +static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, + struct srp_request *req) +{ + struct srp_cmd *cmd = req->cmd->buf; + int len; + u8 fmt; + + if (!scmnd->request_buffer || scmnd->sc_data_direction == DMA_NONE) + return sizeof (struct srp_cmd); + + if (scmnd->sc_data_direction != DMA_FROM_DEVICE && + scmnd->sc_data_direction != DMA_TO_DEVICE) { + printk(KERN_WARNING PFX "Unhandled data direction %d\n", + scmnd->sc_data_direction); + return -EINVAL; + } + + if (scmnd->use_sg) { + struct scatterlist *scat = scmnd->request_buffer; + int n; + int i; + + n = dma_map_sg(target->srp_host->dev->dma_device, + scat, scmnd->use_sg, scmnd->sc_data_direction); + + if (n == 1) { + struct srp_direct_buf *buf = (void *) cmd->add_data; + + fmt = SRP_DATA_DESC_DIRECT; + + buf->va = cpu_to_be64(sg_dma_address(scat)); + buf->key = cpu_to_be32(target->srp_host->mr->rkey); + buf->len = cpu_to_be32(sg_dma_len(scat)); + + len = sizeof (struct srp_cmd) + + sizeof (struct srp_direct_buf); + } else { + struct srp_indirect_buf *buf = (void *) cmd->add_data; + u32 datalen = 0; + + fmt = SRP_DATA_DESC_INDIRECT; + + if (scmnd->sc_data_direction == DMA_TO_DEVICE) + cmd->data_out_desc_cnt = n; + else + cmd->data_in_desc_cnt = n; + + buf->table_desc.va = cpu_to_be64(req->cmd->dma + + sizeof *cmd + + sizeof *buf); + buf->table_desc.key = + cpu_to_be32(target->srp_host->mr->rkey); + buf->table_desc.len = + cpu_to_be32(n * sizeof (struct srp_direct_buf)); + + for (i = 0; i < n; ++i) { + buf->desc_list[i].va = cpu_to_be64(sg_dma_address(&scat[i])); + buf->desc_list[i].key = + cpu_to_be32(target->srp_host->mr->rkey); + buf->desc_list[i].len = cpu_to_be32(sg_dma_len(&scat[i])); + + datalen += sg_dma_len(&scat[i]); + } + + buf->len = cpu_to_be32(datalen); + + len = sizeof (struct srp_cmd) + + sizeof (struct srp_indirect_buf) + + n * sizeof (struct srp_direct_buf); + } + } else { + struct srp_direct_buf *buf = (void *) cmd->add_data; + dma_addr_t dma; + + dma = dma_map_single(target->srp_host->dev->dma_device, + scmnd->request_buffer, scmnd->request_bufflen, + scmnd->sc_data_direction); + if (dma_mapping_error(dma)) { + printk(KERN_WARNING PFX "unable to map %p/%d (dir %d)\n", + scmnd->request_buffer, (int) scmnd->request_bufflen, + scmnd->sc_data_direction); + return -EINVAL; + } + + pci_unmap_addr_set(req, direct_mapping, dma); + + buf->va = cpu_to_be64(dma); + buf->key = cpu_to_be32(target->srp_host->mr->rkey); + buf->len = cpu_to_be32(scmnd->request_bufflen); + + fmt = SRP_DATA_DESC_DIRECT; + + len = sizeof (struct srp_cmd) + sizeof (struct srp_direct_buf); + } + + if (scmnd->sc_data_direction == DMA_TO_DEVICE) + cmd->buf_fmt = fmt << 4; + else + cmd->buf_fmt = fmt; + + + return len; +} + +static void srp_unmap_data(struct scsi_cmnd *scmnd, + struct srp_target_port *target, + struct srp_request *req) +{ + if (!scmnd->request_buffer || + (scmnd->sc_data_direction != DMA_TO_DEVICE && + scmnd->sc_data_direction != DMA_FROM_DEVICE)) + return; + + if (scmnd->use_sg) + dma_unmap_sg(target->srp_host->dev->dma_device, + (struct scatterlist *) scmnd->request_buffer, + scmnd->use_sg, scmnd->sc_data_direction); + else + dma_unmap_single(target->srp_host->dev->dma_device, + pci_unmap_addr(req, direct_mapping), + scmnd->request_bufflen, + scmnd->sc_data_direction); +} + +static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp) +{ + struct srp_request *req; + struct scsi_cmnd *scmnd; + unsigned long flags; + s32 delta; + + delta = (s32) be32_to_cpu(rsp->req_lim_delta); + + spin_lock_irqsave(target->scsi_host->host_lock, flags); + + target->req_lim += delta; + + req = &target->req_ring[rsp->tag & ~SRP_TAG_TSK_MGMT]; + + if (unlikely(rsp->tag & SRP_TAG_TSK_MGMT)) { + if (be32_to_cpu(rsp->resp_data_len) < 4) + req->tsk_status = -1; + else + req->tsk_status = rsp->data[3]; + complete(&req->done); + } else { + scmnd = req->scmnd; + if (!scmnd) + printk(KERN_ERR "Null scmnd for RSP w/tag %016llx\n", + (unsigned long long) rsp->tag); + scmnd->result = rsp->status; + + if (rsp->flags & SRP_RSP_FLAG_SNSVALID) { + memcpy(scmnd->sense_buffer, rsp->data + + be32_to_cpu(rsp->resp_data_len), + min_t(int, be32_to_cpu(rsp->sense_data_len), + SCSI_SENSE_BUFFERSIZE)); + } + + if (rsp->flags & (SRP_RSP_FLAG_DOOVER | SRP_RSP_FLAG_DOUNDER)) + scmnd->resid = be32_to_cpu(rsp->data_out_res_cnt); + else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER)) + scmnd->resid = be32_to_cpu(rsp->data_in_res_cnt); + + srp_unmap_data(scmnd, target, req); + + if (!req->tsk_mgmt) { + req->scmnd = NULL; + scmnd->host_scribble = (void *) -1L; + scmnd->scsi_done(scmnd); + + list_del(&req->list); + req->next = target->req_head; + target->req_head = rsp->tag & ~SRP_TAG_TSK_MGMT; + } else + req->cmd_done = 1; + } + + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); +} + +static void srp_reconnect_work(void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + srp_reconnect_target(target); +} + +static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc) +{ + struct srp_iu *iu; + u8 opcode; + + iu = target->rx_ring[wc->wr_id & ~SRP_OP_RECV]; + + dma_sync_single_for_cpu(target->srp_host->dev->dma_device, iu->dma, + target->max_ti_iu_len, DMA_FROM_DEVICE); + + opcode = *(u8 *) iu->buf; + + if (0) { + int i; + + printk(KERN_ERR PFX "recv completion, opcode 0x%02x\n", opcode); + + for (i = 0; i < wc->byte_len; ++i) { + if (i % 8 == 0) + printk(KERN_ERR " [%02x] ", i); + printk(" %02x", ((u8 *) iu->buf)[i]); + if ((i + 1) % 8 == 0) + printk("\n"); + } + + if (wc->byte_len % 8) + printk("\n"); + } + + switch (opcode) { + case SRP_RSP: + srp_process_rsp(target, iu->buf); + break; + + case SRP_T_LOGOUT: + /* XXX Handle target logout */ + printk(KERN_WARNING PFX "Got target logout request\n"); + break; + + default: + printk(KERN_WARNING PFX "Unhandled SRP opcode 0x%02x\n", opcode); + break; + } + + dma_sync_single_for_device(target->srp_host->dev->dma_device, iu->dma, + target->max_ti_iu_len, DMA_FROM_DEVICE); +} + +static void srp_completion(struct ib_cq *cq, void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + struct ib_wc wc; + unsigned long flags; + + ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); + while (ib_poll_cq(cq, 1, &wc) > 0) { + if (wc.status) { + printk(KERN_ERR PFX "failed %s status %d\n", + wc.wr_id & SRP_OP_RECV ? "receive" : "send", + wc.status); + spin_lock_irqsave(target->scsi_host->host_lock, flags); + if (target->state == SRP_TARGET_LIVE) + schedule_work(&target->work); + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); + break; + } + + if (wc.wr_id & SRP_OP_RECV) + srp_handle_recv(target, &wc); + else + ++target->tx_tail; + } +} + +static int __srp_post_recv(struct srp_target_port *target) +{ + struct srp_iu *iu; + struct ib_sge list; + struct ib_recv_wr wr, *bad_wr; + unsigned int next; + int ret; + + next = target->rx_head & (SRP_RQ_SIZE - 1); + wr.wr_id = next | SRP_OP_RECV; + iu = target->rx_ring[next]; + + list.addr = iu->dma; + list.length = iu->size; + list.lkey = target->srp_host->mr->lkey; + + wr.next = NULL; + wr.sg_list = &list; + wr.num_sge = 1; + + ret = ib_post_recv(target->qp, &wr, &bad_wr); + if (!ret) + ++target->rx_head; + + return ret; +} + +static int srp_post_recv(struct srp_target_port *target) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(target->scsi_host->host_lock, flags); + ret = __srp_post_recv(target); + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); + + return ret; +} + +/* + * Must be called with target->scsi_host->host_lock held to protect + * req_lim and tx_head. + */ +static struct srp_iu *__srp_get_tx_iu(struct srp_target_port *target) +{ + if (target->tx_head - target->tx_tail >= SRP_SQ_SIZE) + return NULL; + + return target->tx_ring[target->tx_head & SRP_SQ_SIZE]; +} + +/* + * Must be called with target->scsi_host->host_lock held to protect + * req_lim and tx_head. + */ +static int __srp_post_send(struct srp_target_port *target, + struct srp_iu *iu, int len) +{ + struct ib_sge list; + struct ib_send_wr wr, *bad_wr; + int ret = 0; + + if (target->req_lim < 1) { + printk(KERN_ERR PFX "Target has req_lim %d\n", target->req_lim); + return -EAGAIN; + } + + list.addr = iu->dma; + list.length = len; + list.lkey = target->srp_host->mr->lkey; + + wr.next = NULL; + wr.wr_id = target->tx_head & SRP_SQ_SIZE; + wr.sg_list = &list; + wr.num_sge = 1; + wr.opcode = IB_WR_SEND; + wr.send_flags = IB_SEND_SIGNALED; + + ret = ib_post_send(target->qp, &wr, &bad_wr); + + if (!ret) { + ++target->tx_head; + --target->req_lim; + } + + return ret; +} + +static int srp_queuecommand(struct scsi_cmnd *scmnd, + void (*done)(struct scsi_cmnd *)) +{ + struct srp_target_port *target = host_to_target(scmnd->device->host); + struct srp_request *req; + struct srp_iu *iu; + struct srp_cmd *cmd; + long req_index; + int len; + + if (target->state == SRP_TARGET_CONNECTING) + goto err; + + if (target->state == SRP_TARGET_DEAD || + target->state == SRP_TARGET_REMOVED) { + scmnd->result = DID_BAD_TARGET << 16; + done(scmnd); + return 0; + } + + iu = __srp_get_tx_iu(target); + if (!iu) + goto err; + + dma_sync_single_for_cpu(target->srp_host->dev->dma_device, iu->dma, + SRP_MAX_IU_LEN, DMA_TO_DEVICE); + + req_index = target->req_head; + + scmnd->scsi_done = done; + scmnd->result = 0; + scmnd->host_scribble = (void *) req_index; + + cmd = iu->buf; + memset(cmd, 0, sizeof *cmd); + + cmd->opcode = SRP_CMD; + cmd->lun = cpu_to_be64((u64) scmnd->device->lun << 48); + cmd->tag = req_index; + memcpy(cmd->cdb, scmnd->cmnd, scmnd->cmd_len); + + req = &target->req_ring[req_index]; + + req->scmnd = scmnd; + req->cmd = iu; + req->cmd_done = 0; + req->tsk_mgmt = NULL; + + len = srp_map_data(scmnd, target, req); + if (len < 0) { + printk(KERN_ERR PFX "Failed to map data\n"); + goto err; + } + + if (__srp_post_recv(target)) { + printk(KERN_ERR PFX "Recv failed\n"); + goto err_unmap; + } + + dma_sync_single_for_device(target->srp_host->dev->dma_device, iu->dma, + SRP_MAX_IU_LEN, DMA_TO_DEVICE); + + if (__srp_post_send(target, iu, len)) { + printk(KERN_ERR PFX "Send failed\n"); + goto err_unmap; + } + + target->req_head = req->next; + list_add_tail(&req->list, &target->req_queue); + + return 0; + +err_unmap: + srp_unmap_data(scmnd, target, req); + +err: + return SCSI_MLQUEUE_HOST_BUSY; +} + +static int srp_alloc_iu_bufs(struct srp_target_port *target) +{ + int i; + + for (i = 0; i < SRP_RQ_SIZE; ++i) { + target->rx_ring[i] = srp_alloc_iu(target->srp_host, + target->max_ti_iu_len, + GFP_KERNEL, DMA_FROM_DEVICE); + if (!target->rx_ring[i]) + goto err; + } + + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) { + target->tx_ring[i] = srp_alloc_iu(target->srp_host, + SRP_MAX_IU_LEN, + GFP_KERNEL, DMA_TO_DEVICE); + if (!target->tx_ring[i]) + goto err; + } + + return 0; + +err: + for (i = 0; i < SRP_RQ_SIZE; ++i) { + srp_free_iu(target->srp_host, target->rx_ring[i]); + target->rx_ring[i] = NULL; + } + + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) { + srp_free_iu(target->srp_host, target->tx_ring[i]); + target->tx_ring[i] = NULL; + } + + return -ENOMEM; +} + +static void srp_cm_rej_handler(struct ib_cm_id *cm_id, + struct ib_cm_event *event, + struct srp_target_port *target) +{ + struct ib_class_port_info *cpi; + int opcode; + + switch (event->param.rej_rcvd.reason) { + case IB_CM_REJ_PORT_CM_REDIRECT: + cpi = event->param.rej_rcvd.ari; + target->path.dlid = cpi->redirect_lid; + target->path.pkey = cpi->redirect_pkey; + cm_id->remote_cm_qpn = be32_to_cpu(cpi->redirect_qp) & 0x00ffffff; + memcpy(target->path.dgid.raw, cpi->redirect_gid, 16); + + target->status = target->path.dlid ? + SRP_DLID_REDIRECT : SRP_PORT_REDIRECT; + break; + + case IB_CM_REJ_PORT_REDIRECT: + if (topspin_workarounds && + !memcmp(&target->ioc_guid, topspin_oui, 3)) { + /* + * Topspin/Cisco SRP gateways incorrectly send + * reject reason code 25 when they mean 24 + * (port redirect). + */ + memcpy(target->path.dgid.raw, + event->param.rej_rcvd.ari, 16); + + printk(KERN_DEBUG PFX "Topspin/Cisco redirect to target port GID %016llx%016llx\n", + (unsigned long long) be64_to_cpu(target->path.dgid.global.subnet_prefix), + (unsigned long long) be64_to_cpu(target->path.dgid.global.interface_id)); + + target->status = SRP_PORT_REDIRECT; + } else { + printk(KERN_WARNING " REJ reason: IB_CM_REJ_PORT_REDIRECT\n"); + target->status = -ECONNRESET; + } + break; + + case IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID: + printk(KERN_WARNING " REJ reason: IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID\n"); + target->status = -ECONNRESET; + break; + + case IB_CM_REJ_CONSUMER_DEFINED: + opcode = *(u8 *) event->private_data; + if (opcode == SRP_LOGIN_REJ) { + struct srp_login_rej *rej = event->private_data; + u32 reason = be32_to_cpu(rej->reason); + + if (reason == SRP_LOGIN_REJ_REQ_IT_IU_LENGTH_TOO_LARGE) + printk(KERN_WARNING PFX + "SRP_LOGIN_REJ: requested max_it_iu_len too large\n"); + else + printk(KERN_WARNING PFX + "SRP LOGIN REJECTED, reason 0x%08x\n", reason); + } else + printk(KERN_WARNING " REJ reason: IB_CM_REJ_CONSUMER_DEFINED," + " opcode 0x%02x\n", opcode); + target->status = -ECONNRESET; + break; + + default: + printk(KERN_WARNING " REJ reason 0x%x\n", + event->param.rej_rcvd.reason); + target->status = -ECONNRESET; + } +} + +static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) +{ + struct srp_target_port *target = cm_id->context; + struct ib_qp_attr *qp_attr = NULL; + int attr_mask = 0; + int comp = 0; + int opcode = 0; + + switch (event->event) { + case IB_CM_REQ_ERROR: + printk(KERN_DEBUG PFX "Sending CM REQ failed\n"); + comp = 1; + target->status = -ECONNRESET; + break; + + case IB_CM_REP_RECEIVED: + comp = 1; + opcode = *(u8 *) event->private_data; + + if (opcode == SRP_LOGIN_RSP) { + struct srp_login_rsp *rsp = event->private_data; + + target->max_ti_iu_len = be32_to_cpu(rsp->max_ti_iu_len); + target->req_lim = be32_to_cpu(rsp->req_lim_delta); + + target->scsi_host->can_queue = min(target->req_lim, + target->scsi_host->can_queue); + } else { + printk(KERN_WARNING PFX "Unhandled RSP opcode %#x\n", opcode); |