summaryrefslogtreecommitdiff
path: root/drivers/nvme
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/nvme')
-rw-r--r--drivers/nvme/host/Kconfig17
-rw-r--r--drivers/nvme/host/Makefile3
-rw-r--r--drivers/nvme/host/core.c131
-rw-r--r--drivers/nvme/host/fabrics.c33
-rw-r--r--drivers/nvme/host/fc.c2586
-rw-r--r--drivers/nvme/host/lightnvm.c223
-rw-r--r--drivers/nvme/host/nvme.h51
-rw-r--r--drivers/nvme/host/pci.c90
-rw-r--r--drivers/nvme/host/rdma.c85
-rw-r--r--drivers/nvme/host/scsi.c11
-rw-r--r--drivers/nvme/target/Kconfig24
-rw-r--r--drivers/nvme/target/Makefile4
-rw-r--r--drivers/nvme/target/admin-cmd.c3
-rw-r--r--drivers/nvme/target/configfs.c30
-rw-r--r--drivers/nvme/target/core.c32
-rw-r--r--drivers/nvme/target/fabrics-cmd.c14
-rw-r--r--drivers/nvme/target/fc.c2288
-rw-r--r--drivers/nvme/target/fcloop.c1148
-rw-r--r--drivers/nvme/target/io-cmd.c39
-rw-r--r--drivers/nvme/target/loop.c26
-rw-r--r--drivers/nvme/target/nvmet.h8
-rw-r--r--drivers/nvme/target/rdma.c26
22 files changed, 6588 insertions, 284 deletions
diff --git a/drivers/nvme/host/Kconfig b/drivers/nvme/host/Kconfig
index f7d37a62f874..90745a616df7 100644
--- a/drivers/nvme/host/Kconfig
+++ b/drivers/nvme/host/Kconfig
@@ -43,3 +43,20 @@ config NVME_RDMA
from https://github.com/linux-nvme/nvme-cli.
If unsure, say N.
+
+config NVME_FC
+ tristate "NVM Express over Fabrics FC host driver"
+ depends on BLOCK
+ depends on HAS_DMA
+ select NVME_CORE
+ select NVME_FABRICS
+ select SG_POOL
+ help
+ This provides support for the NVMe over Fabrics protocol using
+ the FC transport. This allows you to use remote block devices
+ exported using the NVMe protocol set.
+
+ To configure a NVMe over Fabrics controller use the nvme-cli tool
+ from https://github.com/linux-nvme/nvme-cli.
+
+ If unsure, say N.
diff --git a/drivers/nvme/host/Makefile b/drivers/nvme/host/Makefile
index 47abcec23514..f1a7d945fbb6 100644
--- a/drivers/nvme/host/Makefile
+++ b/drivers/nvme/host/Makefile
@@ -2,6 +2,7 @@ obj-$(CONFIG_NVME_CORE) += nvme-core.o
obj-$(CONFIG_BLK_DEV_NVME) += nvme.o
obj-$(CONFIG_NVME_FABRICS) += nvme-fabrics.o
obj-$(CONFIG_NVME_RDMA) += nvme-rdma.o
+obj-$(CONFIG_NVME_FC) += nvme-fc.o
nvme-core-y := core.o
nvme-core-$(CONFIG_BLK_DEV_NVME_SCSI) += scsi.o
@@ -12,3 +13,5 @@ nvme-y += pci.o
nvme-fabrics-y += fabrics.o
nvme-rdma-y += rdma.o
+
+nvme-fc-y += fc.o
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 79e679d12f3b..b40cfb076f02 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -201,13 +201,7 @@ fail:
void nvme_requeue_req(struct request *req)
{
- unsigned long flags;
-
- blk_mq_requeue_request(req);
- spin_lock_irqsave(req->q->queue_lock, flags);
- if (!blk_queue_stopped(req->q))
- blk_mq_kick_requeue_list(req->q);
- spin_unlock_irqrestore(req->q->queue_lock, flags);
+ blk_mq_requeue_request(req, !blk_mq_queue_stopped(req->q));
}
EXPORT_SYMBOL_GPL(nvme_requeue_req);
@@ -227,8 +221,7 @@ struct request *nvme_alloc_request(struct request_queue *q,
req->cmd_type = REQ_TYPE_DRV_PRIV;
req->cmd_flags |= REQ_FAILFAST_DRIVER;
- req->cmd = (unsigned char *)cmd;
- req->cmd_len = sizeof(struct nvme_command);
+ nvme_req(req)->cmd = cmd;
return req;
}
@@ -246,8 +239,6 @@ static inline int nvme_setup_discard(struct nvme_ns *ns, struct request *req,
struct nvme_command *cmnd)
{
struct nvme_dsm_range *range;
- struct page *page;
- int offset;
unsigned int nr_bytes = blk_rq_bytes(req);
range = kmalloc(sizeof(*range), GFP_ATOMIC);
@@ -264,19 +255,12 @@ static inline int nvme_setup_discard(struct nvme_ns *ns, struct request *req,
cmnd->dsm.nr = 0;
cmnd->dsm.attributes = cpu_to_le32(NVME_DSMGMT_AD);
- req->completion_data = range;
- page = virt_to_page(range);
- offset = offset_in_page(range);
- blk_add_request_payload(req, page, offset, sizeof(*range));
-
- /*
- * we set __data_len back to the size of the area to be discarded
- * on disk. This allows us to report completion on the full amount
- * of blocks described by the request.
- */
- req->__data_len = nr_bytes;
+ req->special_vec.bv_page = virt_to_page(range);
+ req->special_vec.bv_offset = offset_in_page(range);
+ req->special_vec.bv_len = sizeof(*range);
+ req->rq_flags |= RQF_SPECIAL_PAYLOAD;
- return 0;
+ return BLK_MQ_RQ_QUEUE_OK;
}
static inline void nvme_setup_rw(struct nvme_ns *ns, struct request *req,
@@ -295,7 +279,6 @@ static inline void nvme_setup_rw(struct nvme_ns *ns, struct request *req,
memset(cmnd, 0, sizeof(*cmnd));
cmnd->rw.opcode = (rq_data_dir(req) ? nvme_cmd_write : nvme_cmd_read);
- cmnd->rw.command_id = req->tag;
cmnd->rw.nsid = cpu_to_le32(ns->ns_id);
cmnd->rw.slba = cpu_to_le64(nvme_block_nr(ns, blk_rq_pos(req)));
cmnd->rw.length = cpu_to_le16((blk_rq_bytes(req) >> ns->lba_shift) - 1);
@@ -324,10 +307,10 @@ static inline void nvme_setup_rw(struct nvme_ns *ns, struct request *req,
int nvme_setup_cmd(struct nvme_ns *ns, struct request *req,
struct nvme_command *cmd)
{
- int ret = 0;
+ int ret = BLK_MQ_RQ_QUEUE_OK;
if (req->cmd_type == REQ_TYPE_DRV_PRIV)
- memcpy(cmd, req->cmd, sizeof(*cmd));
+ memcpy(cmd, nvme_req(req)->cmd, sizeof(*cmd));
else if (req_op(req) == REQ_OP_FLUSH)
nvme_setup_flush(ns, cmd);
else if (req_op(req) == REQ_OP_DISCARD)
@@ -335,6 +318,8 @@ int nvme_setup_cmd(struct nvme_ns *ns, struct request *req,
else
nvme_setup_rw(ns, req, cmd);
+ cmd->common.command_id = req->tag;
+
return ret;
}
EXPORT_SYMBOL_GPL(nvme_setup_cmd);
@@ -344,7 +329,7 @@ EXPORT_SYMBOL_GPL(nvme_setup_cmd);
* if the result is positive, it's an NVM Express status code
*/
int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
- struct nvme_completion *cqe, void *buffer, unsigned bufflen,
+ union nvme_result *result, void *buffer, unsigned bufflen,
unsigned timeout, int qid, int at_head, int flags)
{
struct request *req;
@@ -355,7 +340,6 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
return PTR_ERR(req);
req->timeout = timeout ? timeout : ADMIN_TIMEOUT;
- req->special = cqe;
if (buffer && bufflen) {
ret = blk_rq_map_kern(q, req, buffer, bufflen, GFP_KERNEL);
@@ -364,6 +348,8 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
}
blk_execute_rq(req->q, NULL, req, at_head);
+ if (result)
+ *result = nvme_req(req)->result;
ret = req->errors;
out:
blk_mq_free_request(req);
@@ -385,7 +371,6 @@ int __nvme_submit_user_cmd(struct request_queue *q, struct nvme_command *cmd,
u32 *result, unsigned timeout)
{
bool write = nvme_is_write(cmd);
- struct nvme_completion cqe;
struct nvme_ns *ns = q->queuedata;
struct gendisk *disk = ns ? ns->disk : NULL;
struct request *req;
@@ -398,7 +383,6 @@ int __nvme_submit_user_cmd(struct request_queue *q, struct nvme_command *cmd,
return PTR_ERR(req);
req->timeout = timeout ? timeout : ADMIN_TIMEOUT;
- req->special = &cqe;
if (ubuffer && bufflen) {
ret = blk_rq_map_user(q, req, NULL, ubuffer, bufflen,
@@ -453,7 +437,7 @@ int __nvme_submit_user_cmd(struct request_queue *q, struct nvme_command *cmd,
blk_execute_rq(req->q, disk, req, 0);
ret = req->errors;
if (result)
- *result = le32_to_cpu(cqe.result);
+ *result = le32_to_cpu(nvme_req(req)->result.u32);
if (meta && !ret && !write) {
if (copy_to_user(meta_buffer, meta, meta_len))
ret = -EFAULT;
@@ -602,7 +586,7 @@ int nvme_get_features(struct nvme_ctrl *dev, unsigned fid, unsigned nsid,
void *buffer, size_t buflen, u32 *result)
{
struct nvme_command c;
- struct nvme_completion cqe;
+ union nvme_result res;
int ret;
memset(&c, 0, sizeof(c));
@@ -610,10 +594,10 @@ int nvme_get_features(struct nvme_ctrl *dev, unsigned fid, unsigned nsid,
c.features.nsid = cpu_to_le32(nsid);
c.features.fid = cpu_to_le32(fid);
- ret = __nvme_submit_sync_cmd(dev->admin_q, &c, &cqe, buffer, buflen, 0,
+ ret = __nvme_submit_sync_cmd(dev->admin_q, &c, &res, buffer, buflen, 0,
NVME_QID_ANY, 0, 0);
if (ret >= 0 && result)
- *result = le32_to_cpu(cqe.result);
+ *result = le32_to_cpu(res.u32);
return ret;
}
@@ -621,7 +605,7 @@ int nvme_set_features(struct nvme_ctrl *dev, unsigned fid, unsigned dword11,
void *buffer, size_t buflen, u32 *result)
{
struct nvme_command c;
- struct nvme_completion cqe;
+ union nvme_result res;
int ret;
memset(&c, 0, sizeof(c));
@@ -629,10 +613,10 @@ int nvme_set_features(struct nvme_ctrl *dev, unsigned fid, unsigned dword11,
c.features.fid = cpu_to_le32(fid);
c.features.dword11 = cpu_to_le32(dword11);
- ret = __nvme_submit_sync_cmd(dev->admin_q, &c, &cqe,
+ ret = __nvme_submit_sync_cmd(dev->admin_q, &c, &res,
buffer, buflen, 0, NVME_QID_ANY, 0, 0);
if (ret >= 0 && result)
- *result = le32_to_cpu(cqe.result);
+ *result = le32_to_cpu(res.u32);
return ret;
}
@@ -1683,27 +1667,24 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid)
if (nvme_revalidate_ns(ns, &id))
goto out_free_queue;
- if (nvme_nvm_ns_supported(ns, id)) {
- if (nvme_nvm_register(ns, disk_name, node,
- &nvme_ns_attr_group)) {
- dev_warn(ctrl->dev, "%s: LightNVM init failure\n",
- __func__);
- goto out_free_id;
- }
- } else {
- disk = alloc_disk_node(0, node);
- if (!disk)
- goto out_free_id;
+ if (nvme_nvm_ns_supported(ns, id) &&
+ nvme_nvm_register(ns, disk_name, node)) {
+ dev_warn(ctrl->dev, "%s: LightNVM init failure\n", __func__);
+ goto out_free_id;
+ }
- disk->fops = &nvme_fops;
- disk->private_data = ns;
- disk->queue = ns->queue;
- disk->flags = GENHD_FL_EXT_DEVT;
- memcpy(disk->disk_name, disk_name, DISK_NAME_LEN);
- ns->disk = disk;
+ disk = alloc_disk_node(0, node);
+ if (!disk)
+ goto out_free_id;
- __nvme_revalidate_disk(disk, id);
- }
+ disk->fops = &nvme_fops;
+ disk->private_data = ns;
+ disk->queue = ns->queue;
+ disk->flags = GENHD_FL_EXT_DEVT;
+ memcpy(disk->disk_name, disk_name, DISK_NAME_LEN);
+ ns->disk = disk;
+
+ __nvme_revalidate_disk(disk, id);
mutex_lock(&ctrl->namespaces_mutex);
list_add_tail(&ns->list, &ctrl->namespaces);
@@ -1713,14 +1694,14 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid)
kfree(id);
- if (ns->ndev)
- return;
-
device_add_disk(ctrl->device, ns->disk);
if (sysfs_create_group(&disk_to_dev(ns->disk)->kobj,
&nvme_ns_attr_group))
pr_warn("%s: failed to create sysfs group for identification\n",
ns->disk->disk_name);
+ if (ns->ndev && nvme_nvm_register_sysfs(ns))
+ pr_warn("%s: failed to register lightnvm sysfs group for identification\n",
+ ns->disk->disk_name);
return;
out_free_id:
kfree(id);
@@ -1742,6 +1723,8 @@ static void nvme_ns_remove(struct nvme_ns *ns)
blk_integrity_unregister(ns->disk);
sysfs_remove_group(&disk_to_dev(ns->disk)->kobj,
&nvme_ns_attr_group);
+ if (ns->ndev)
+ nvme_nvm_unregister_sysfs(ns);
del_gendisk(ns->disk);
blk_mq_abort_requeue_list(ns->queue);
blk_cleanup_queue(ns->queue);
@@ -1905,18 +1888,25 @@ static void nvme_async_event_work(struct work_struct *work)
spin_unlock_irq(&ctrl->lock);
}
-void nvme_complete_async_event(struct nvme_ctrl *ctrl,
- struct nvme_completion *cqe)
+void nvme_complete_async_event(struct nvme_ctrl *ctrl, __le16 status,
+ union nvme_result *res)
{
- u16 status = le16_to_cpu(cqe->status) >> 1;
- u32 result = le32_to_cpu(cqe->result);
+ u32 result = le32_to_cpu(res->u32);
+ bool done = true;
- if (status == NVME_SC_SUCCESS || status == NVME_SC_ABORT_REQ) {
+ switch (le16_to_cpu(status) >> 1) {
+ case NVME_SC_SUCCESS:
+ done = false;
+ /*FALLTHRU*/
+ case NVME_SC_ABORT_REQ:
++ctrl->event_limit;
schedule_work(&ctrl->async_event_work);
+ break;
+ default:
+ break;
}
- if (status != NVME_SC_SUCCESS)
+ if (done)
return;
switch (result & 0xff07) {
@@ -2078,14 +2068,8 @@ void nvme_stop_queues(struct nvme_ctrl *ctrl)
struct nvme_ns *ns;
mutex_lock(&ctrl->namespaces_mutex);
- list_for_each_entry(ns, &ctrl->namespaces, list) {
- spin_lock_irq(ns->queue->queue_lock);
- queue_flag_set(QUEUE_FLAG_STOPPED, ns->queue);
- spin_unlock_irq(ns->queue->queue_lock);
-
- blk_mq_cancel_requeue_work(ns->queue);
- blk_mq_stop_hw_queues(ns->queue);
- }
+ list_for_each_entry(ns, &ctrl->namespaces, list)
+ blk_mq_quiesce_queue(ns->queue);
mutex_unlock(&ctrl->namespaces_mutex);
}
EXPORT_SYMBOL_GPL(nvme_stop_queues);
@@ -2096,7 +2080,6 @@ void nvme_start_queues(struct nvme_ctrl *ctrl)
mutex_lock(&ctrl->namespaces_mutex);
list_for_each_entry(ns, &ctrl->namespaces, list) {
- queue_flag_clear_unlocked(QUEUE_FLAG_STOPPED, ns->queue);
blk_mq_start_stopped_hw_queues(ns->queue, true);
blk_mq_kick_requeue_list(ns->queue);
}
diff --git a/drivers/nvme/host/fabrics.c b/drivers/nvme/host/fabrics.c
index 5a3f008d3480..916d13608059 100644
--- a/drivers/nvme/host/fabrics.c
+++ b/drivers/nvme/host/fabrics.c
@@ -161,7 +161,7 @@ EXPORT_SYMBOL_GPL(nvmf_get_subsysnqn);
int nvmf_reg_read32(struct nvme_ctrl *ctrl, u32 off, u32 *val)
{
struct nvme_command cmd;
- struct nvme_completion cqe;
+ union nvme_result res;
int ret;
memset(&cmd, 0, sizeof(cmd));
@@ -169,11 +169,11 @@ int nvmf_reg_read32(struct nvme_ctrl *ctrl, u32 off, u32 *val)
cmd.prop_get.fctype = nvme_fabrics_type_property_get;
cmd.prop_get.offset = cpu_to_le32(off);
- ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &cqe, NULL, 0, 0,
+ ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &res, NULL, 0, 0,
NVME_QID_ANY, 0, 0);
if (ret >= 0)
- *val = le64_to_cpu(cqe.result64);
+ *val = le64_to_cpu(res.u64);
if (unlikely(ret != 0))
dev_err(ctrl->device,
"Property Get error: %d, offset %#x\n",
@@ -207,7 +207,7 @@ EXPORT_SYMBOL_GPL(nvmf_reg_read32);
int nvmf_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val)
{
struct nvme_command cmd;
- struct nvme_completion cqe;
+ union nvme_result res;
int ret;
memset(&cmd, 0, sizeof(cmd));
@@ -216,11 +216,11 @@ int nvmf_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val)
cmd.prop_get.attrib = 1;
cmd.prop_get.offset = cpu_to_le32(off);
- ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &cqe, NULL, 0, 0,
+ ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &res, NULL, 0, 0,
NVME_QID_ANY, 0, 0);
if (ret >= 0)
- *val = le64_to_cpu(cqe.result64);
+ *val = le64_to_cpu(res.u64);
if (unlikely(ret != 0))
dev_err(ctrl->device,
"Property Get error: %d, offset %#x\n",
@@ -368,7 +368,7 @@ static void nvmf_log_connect_error(struct nvme_ctrl *ctrl,
int nvmf_connect_admin_queue(struct nvme_ctrl *ctrl)
{
struct nvme_command cmd;
- struct nvme_completion cqe;
+ union nvme_result res;
struct nvmf_connect_data *data;
int ret;
@@ -400,16 +400,16 @@ int nvmf_connect_admin_queue(struct nvme_ctrl *ctrl)
strncpy(data->subsysnqn, ctrl->opts->subsysnqn, NVMF_NQN_SIZE);
strncpy(data->hostnqn, ctrl->opts->host->nqn, NVMF_NQN_SIZE);
- ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &cqe,
+ ret = __nvme_submit_sync_cmd(ctrl->admin_q, &cmd, &res,
data, sizeof(*data), 0, NVME_QID_ANY, 1,
BLK_MQ_REQ_RESERVED | BLK_MQ_REQ_NOWAIT);
if (ret) {
- nvmf_log_connect_error(ctrl, ret, le32_to_cpu(cqe.result),
+ nvmf_log_connect_error(ctrl, ret, le32_to_cpu(res.u32),
&cmd, data);
goto out_free_data;
}
- ctrl->cntlid = le16_to_cpu(cqe.result16);
+ ctrl->cntlid = le16_to_cpu(res.u16);
out_free_data:
kfree(data);
@@ -441,7 +441,7 @@ int nvmf_connect_io_queue(struct nvme_ctrl *ctrl, u16 qid)
{
struct nvme_command cmd;
struct nvmf_connect_data *data;
- struct nvme_completion cqe;
+ union nvme_result res;
int ret;
memset(&cmd, 0, sizeof(cmd));
@@ -459,11 +459,11 @@ int nvmf_connect_io_queue(struct nvme_ctrl *ctrl, u16 qid)
strncpy(data->subsysnqn, ctrl->opts->subsysnqn, NVMF_NQN_SIZE);
strncpy(data->hostnqn, ctrl->opts->host->nqn, NVMF_NQN_SIZE);
- ret = __nvme_submit_sync_cmd(ctrl->connect_q, &cmd, &cqe,
+ ret = __nvme_submit_sync_cmd(ctrl->connect_q, &cmd, &res,
data, sizeof(*data), 0, qid, 1,
BLK_MQ_REQ_RESERVED | BLK_MQ_REQ_NOWAIT);
if (ret) {
- nvmf_log_connect_error(ctrl, ret, le32_to_cpu(cqe.result),
+ nvmf_log_connect_error(ctrl, ret, le32_to_cpu(res.u32),
&cmd, data);
}
kfree(data);
@@ -576,7 +576,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts,
nqnlen = strlen(opts->subsysnqn);
if (nqnlen >= NVMF_NQN_SIZE) {
pr_err("%s needs to be < %d bytes\n",
- opts->subsysnqn, NVMF_NQN_SIZE);
+ opts->subsysnqn, NVMF_NQN_SIZE);
ret = -EINVAL;
goto out;
}
@@ -666,10 +666,12 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts,
if (nqnlen >= NVMF_NQN_SIZE) {
pr_err("%s needs to be < %d bytes\n",
p, NVMF_NQN_SIZE);
+ kfree(p);
ret = -EINVAL;
goto out;
}
opts->host = nvmf_host_add(p);
+ kfree(p);
if (!opts->host) {
ret = -ENOMEM;
goto out;
@@ -825,8 +827,7 @@ nvmf_create_ctrl(struct device *dev, const char *buf, size_t count)
out_unlock:
mutex_unlock(&nvmf_transports_mutex);
out_free_opts:
- nvmf_host_put(opts->host);
- kfree(opts);
+ nvmf_free_options(opts);
return ERR_PTR(ret);
}
diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
new file mode 100644
index 000000000000..771e2e761872
--- /dev/null
+++ b/drivers/nvme/host/fc.c
@@ -0,0 +1,2586 @@
+/*
+ * Copyright (c) 2016 Avago Technologies. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful.
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND WARRANTIES,
+ * INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A
+ * PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE DISCLAIMED, EXCEPT TO
+ * THE EXTENT THAT SUCH DISCLAIMERS ARE HELD TO BE LEGALLY INVALID.
+ * See the GNU General Public License for more details, a copy of which
+ * can be found in the file COPYING included with this package
+ *
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/module.h>
+#include <linux/parser.h>
+#include <uapi/scsi/fc/fc_fs.h>
+#include <uapi/scsi/fc/fc_els.h>
+
+#include "nvme.h"
+#include "fabrics.h"
+#include <linux/nvme-fc-driver.h>
+#include <linux/nvme-fc.h>
+
+
+/* *************************** Data Structures/Defines ****************** */
+
+
+/*
+ * We handle AEN commands ourselves and don't even let the
+ * block layer know about them.
+ */
+#define NVME_FC_NR_AEN_COMMANDS 1
+#define NVME_FC_AQ_BLKMQ_DEPTH \
+ (NVMF_AQ_DEPTH - NVME_FC_NR_AEN_COMMANDS)
+#define AEN_CMDID_BASE (NVME_FC_AQ_BLKMQ_DEPTH + 1)
+
+enum nvme_fc_queue_flags {
+ NVME_FC_Q_CONNECTED = (1 << 0),
+};
+
+#define NVMEFC_QUEUE_DELAY 3 /* ms units */
+
+struct nvme_fc_queue {
+ struct nvme_fc_ctrl *ctrl;
+ struct device *dev;
+ struct blk_mq_hw_ctx *hctx;
+ void *lldd_handle;
+ int queue_size;
+ size_t cmnd_capsule_len;
+ u32 qnum;
+ u32 rqcnt;
+ u32 seqno;
+
+ u64 connection_id;
+ atomic_t csn;
+
+ unsigned long flags;
+} __aligned(sizeof(u64)); /* alignment for other things alloc'd with */
+
+struct nvmefc_ls_req_op {
+ struct nvmefc_ls_req ls_req;
+
+ struct nvme_fc_ctrl *ctrl;
+ struct nvme_fc_queue *queue;
+ struct request *rq;
+
+ int ls_error;
+ struct completion ls_done;
+ struct list_head lsreq_list; /* ctrl->ls_req_list */
+ bool req_queued;
+};
+
+enum nvme_fcpop_state {
+ FCPOP_STATE_UNINIT = 0,
+ FCPOP_STATE_IDLE = 1,
+ FCPOP_STATE_ACTIVE = 2,
+ FCPOP_STATE_ABORTED = 3,
+};
+
+struct nvme_fc_fcp_op {
+ struct nvme_request nreq; /*
+ * nvme/host/core.c
+ * requires this to be
+ * the 1st element in the
+ * private structure
+ * associated with the
+ * request.
+ */
+ struct nvmefc_fcp_req fcp_req;
+
+ struct nvme_fc_ctrl *ctrl;
+ struct nvme_fc_queue *queue;
+ struct request *rq;
+
+ atomic_t state;
+ u32 rqno;
+ u32 nents;
+
+ struct nvme_fc_cmd_iu cmd_iu;
+ struct nvme_fc_ersp_iu rsp_iu;
+};
+
+struct nvme_fc_lport {
+ struct nvme_fc_local_port localport;
+
+ struct ida endp_cnt;
+ struct list_head port_list; /* nvme_fc_port_list */
+ struct list_head endp_list;
+ struct device *dev; /* physical device for dma */
+ struct nvme_fc_port_template *ops;
+ struct kref ref;
+} __aligned(sizeof(u64)); /* alignment for other things alloc'd with */
+
+struct nvme_fc_rport {
+ struct nvme_fc_remote_port remoteport;
+
+ struct list_head endp_list; /* for lport->endp_list */
+ struct list_head ctrl_list;
+ spinlock_t lock;
+ struct kref ref;
+} __aligned(sizeof(u64)); /* alignment for other things alloc'd with */
+
+enum nvme_fcctrl_state {
+ FCCTRL_INIT = 0,
+ FCCTRL_ACTIVE = 1,
+};
+
+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 cnum;
+
+ u64 association_id;
+
+ u64 cap;
+
+ struct list_head ctrl_list; /* rport->ctrl_list */
+ struct list_head ls_req_list;
+
+ struct blk_mq_tag_set admin_tag_set;
+ struct blk_mq_tag_set tag_set;
+
+ struct work_struct delete_work;
+ struct kref ref;
+ int state;
+
+ struct nvme_fc_fcp_op aen_ops[NVME_FC_NR_AEN_COMMANDS];
+
+ struct nvme_ctrl ctrl;
+};
+
+static inline struct nvme_fc_ctrl *
+to_fc_ctrl(struct nvme_ctrl *ctrl)
+{
+ return container_of(ctrl, struct nvme_fc_ctrl, ctrl);
+}
+
+static inline struct nvme_fc_lport *
+localport_to_lport(struct nvme_fc_local_port *portptr)
+{
+ return container_of(portptr, struct nvme_fc_lport, localport);
+}
+
+static inline struct nvme_fc_rport *
+remoteport_to_rport(struct nvme_fc_remote_port *portptr)
+{
+ return container_of(portptr, struct nvme_fc_rport, remoteport);
+}
+
+static inline struct nvmefc_ls_req_op *
+ls_req_to_lsop(struct nvmefc_ls_req *lsreq)
+{
+ return container_of(lsreq, struct nvmefc_ls_req_op, ls_req);
+}
+
+static inline struct nvme_fc_fcp_op *
+fcp_req_to_fcp_op(struct nvmefc_fcp_req *fcpreq)
+{
+ return container_of(fcpreq, struct nvme_fc_fcp_op, fcp_req);
+}
+
+
+
+/* *************************** Globals **************************** */
+
+
+static DEFINE_SPINLOCK(nvme_fc_lock);
+
+static LIST_HEAD(nvme_fc_lport_list);
+static DEFINE_IDA(nvme_fc_local_port_cnt);
+static DEFINE_IDA(nvme_fc_ctrl_cnt);
+
+static struct workqueue_struct *nvme_fc_wq;
+
+
+
+/* *********************** FC-NVME Port Management ************************ */
+
+static int __nvme_fc_del_ctrl(struct nvme_fc_ctrl *);
+static void __nvme_fc_delete_hw_queue(struct nvme_fc_ctrl *,
+ struct nvme_fc_queue *, unsigned int);
+
+
+/**
+ * nvme_fc_register_localport - transport entry point called by an
+ * LLDD to register the existence of a NVME
+ * host FC port.
+ * @pinfo: pointer to information about the port to be registered
+ * @template: LLDD entrypoints and operational parameters for the port
+ * @dev: physical hardware device node port corresponds to. Will be
+ * used for DMA mappings
+ * @lport_p: pointer to a local port pointer. Upon success, the routine
+ * will allocate a nvme_fc_local_port structure and place its
+ * address in the local port pointer. Upon failure, local port
+ * pointer will be set to 0.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvme_fc_register_localport(struct nvme_fc_port_info *pinfo,
+ struct nvme_fc_port_template *template,
+ struct device *dev,
+ struct nvme_fc_local_port **portptr)
+{
+ struct nvme_fc_lport *newrec;
+ unsigned long flags;
+ int ret, idx;
+
+ if (!template->localport_delete || !template->remoteport_delete ||
+ !template->ls_req || !template->fcp_io ||
+ !template->ls_abort || !template->fcp_abort ||
+ !template->max_hw_queues || !template->max_sgl_segments ||
+ !template->max_dif_sgl_segments || !template->dma_boundary) {
+ ret = -EINVAL;
+ goto out_reghost_failed;
+ }
+
+ newrec = kmalloc((sizeof(*newrec) + template->local_priv_sz),
+ GFP_KERNEL);
+ if (!newrec) {
+ ret = -ENOMEM;
+ goto out_reghost_failed;
+ }
+
+ idx = ida_simple_get(&nvme_fc_local_port_cnt, 0, 0, GFP_KERNEL);
+ if (idx < 0) {
+ ret = -ENOSPC;
+ goto out_fail_kfree;
+ }
+
+ if (!get_device(dev) && dev) {
+ ret = -ENODEV;
+ goto out_ida_put;
+ }
+
+ INIT_LIST_HEAD(&newrec->port_list);
+ INIT_LIST_HEAD(&newrec->endp_list);
+ kref_init(&newrec->ref);
+ newrec->ops = template;
+ newrec->dev = dev;
+ ida_init(&newrec->endp_cnt);
+ newrec->localport.private = &newrec[1];
+ newrec->localport.node_name = pinfo->node_name;
+ newrec->localport.port_name = pinfo->port_name;
+ newrec->localport.port_role = pinfo->port_role;
+ newrec->localport.port_id = pinfo->port_id;
+ newrec->localport.port_state = FC_OBJSTATE_ONLINE;
+ newrec->localport.port_num = idx;
+
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+ list_add_tail(&newrec->port_list, &nvme_fc_lport_list);
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ if (dev)
+ dma_set_seg_boundary(dev, template->dma_boundary);
+
+ *portptr = &newrec->localport;
+ return 0;
+
+out_ida_put:
+ ida_simple_remove(&nvme_fc_local_port_cnt, idx);
+out_fail_kfree:
+ kfree(newrec);
+out_reghost_failed:
+ *portptr = NULL;
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(nvme_fc_register_localport);
+
+static void
+nvme_fc_free_lport(struct kref *ref)
+{
+ struct nvme_fc_lport *lport =
+ container_of(ref, struct nvme_fc_lport, ref);
+ unsigned long flags;
+
+ WARN_ON(lport->localport.port_state != FC_OBJSTATE_DELETED);
+ WARN_ON(!list_empty(&lport->endp_list));
+
+ /* remove from transport list */
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+ list_del(&lport->port_list);
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ /* let the LLDD know we've finished tearing it down */
+ lport->ops->localport_delete(&lport->localport);
+
+ ida_simple_remove(&nvme_fc_local_port_cnt, lport->localport.port_num);
+ ida_destroy(&lport->endp_cnt);
+
+ put_device(lport->dev);
+
+ kfree(lport);
+}
+
+static void
+nvme_fc_lport_put(struct nvme_fc_lport *lport)
+{
+ kref_put(&lport->ref, nvme_fc_free_lport);
+}
+
+static int
+nvme_fc_lport_get(struct nvme_fc_lport *lport)
+{
+ return kref_get_unless_zero(&lport->ref);
+}
+
+/**
+ * nvme_fc_unregister_localport - transport entry point called by an
+ * LLDD to deregister/remove a previously
+ * registered a NVME host FC port.
+ * @localport: pointer to the (registered) local port that is to be
+ * deregistered.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvme_fc_unregister_localport(struct nvme_fc_local_port *portptr)
+{
+ struct nvme_fc_lport *lport = localport_to_lport(portptr);
+ unsigned long flags;
+
+ if (!portptr)
+ return -EINVAL;
+
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+
+ if (portptr->port_state != FC_OBJSTATE_ONLINE) {
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+ return -EINVAL;
+ }
+ portptr->port_state = FC_OBJSTATE_DELETED;
+
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ nvme_fc_lport_put(lport);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(nvme_fc_unregister_localport);
+
+/**
+ * nvme_fc_register_remoteport - transport entry point called by an
+ * LLDD to register the existence of a NVME
+ * subsystem FC port on its fabric.
+ * @localport: pointer to the (registered) local port that the remote
+ * subsystem port is connected to.
+ * @pinfo: pointer to information about the port to be registered
+ * @rport_p: pointer to a remote port pointer. Upon success, the routine
+ * will allocate a nvme_fc_remote_port structure and place its
+ * address in the remote port pointer. Upon failure, remote port
+ * pointer will be set to 0.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvme_fc_register_remoteport(struct nvme_fc_local_port *localport,
+ struct nvme_fc_port_info *pinfo,
+ struct nvme_fc_remote_port **portptr)
+{
+ struct nvme_fc_lport *lport = localport_to_lport(localport);
+ struct nvme_fc_rport *newrec;
+ unsigned long flags;
+ int ret, idx;
+
+ newrec = kmalloc((sizeof(*newrec) + lport->ops->remote_priv_sz),
+ GFP_KERNEL);
+ if (!newrec) {
+ ret = -ENOMEM;
+ goto out_reghost_failed;
+ }
+
+ if (!nvme_fc_lport_get(lport)) {
+ ret = -ESHUTDOWN;
+ goto out_kfree_rport;
+ }
+
+ idx = ida_simple_get(&lport->endp_cnt, 0, 0, GFP_KERNEL);
+ if (idx < 0) {
+ ret = -ENOSPC;
+ goto out_lport_put;
+ }
+
+ INIT_LIST_HEAD(&newrec->endp_list);
+ INIT_LIST_HEAD(&newrec->ctrl_list);
+ kref_init(&newrec->ref);
+ spin_lock_init(&newrec->lock);
+ newrec->remoteport.localport = &lport->localport;
+ newrec->remoteport.private = &newrec[1];
+ newrec->remoteport.port_role = pinfo->port_role;
+ newrec->remoteport.node_name = pinfo->node_name;
+ newrec->remoteport.port_name = pinfo->port_name;
+ newrec->remoteport.port_id = pinfo->port_id;
+ newrec->remoteport.port_state = FC_OBJSTATE_ONLINE;
+ newrec->remoteport.port_num = idx;
+
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+ list_add_tail(&newrec->endp_list, &lport->endp_list);
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ *portptr = &newrec->remoteport;
+ return 0;
+
+out_lport_put:
+ nvme_fc_lport_put(lport);
+out_kfree_rport:
+ kfree(newrec);
+out_reghost_failed:
+ *portptr = NULL;
+ return ret;
+
+}
+EXPORT_SYMBOL_GPL(nvme_fc_register_remoteport);
+
+static void
+nvme_fc_free_rport(struct kref *ref)
+{
+ struct nvme_fc_rport *rport =
+ container_of(ref, struct nvme_fc_rport, ref);
+ struct nvme_fc_lport *lport =
+ localport_to_lport(rport->remoteport.localport);
+ unsigned long flags;
+
+ WARN_ON(rport->remoteport.port_state != FC_OBJSTATE_DELETED);
+ WARN_ON(!list_empty(&rport->ctrl_list));
+
+ /* remove from lport list */
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+ list_del(&rport->endp_list);
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ /* let the LLDD know we've finished tearing it down */
+ lport->ops->remoteport_delete(&rport->remoteport);
+
+ ida_simple_remove(&lport->endp_cnt, rport->remoteport.port_num);
+
+ kfree(rport);
+
+ nvme_fc_lport_put(lport);
+}
+
+static void
+nvme_fc_rport_put(struct nvme_fc_rport *rport)
+{
+ kref_put(&rport->ref, nvme_fc_free_rport);
+}
+
+static int
+nvme_fc_rport_get(struct nvme_fc_rport *rport)
+{
+ return kref_get_unless_zero(&rport->ref);
+}
+
+/**
+ * nvme_fc_unregister_remoteport - transport entry point called by an
+ * LLDD to deregister/remove a previously
+ * registered a NVME subsystem FC port.
+ * @remoteport: pointer to the (registered) remote port that is to be
+ * deregistered.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvme_fc_unregister_remoteport(struct nvme_fc_remote_port *portptr)
+{
+ struct nvme_fc_rport *rport = remoteport_to_rport(portptr);
+ struct nvme_fc_ctrl *ctrl;
+ unsigned long flags;
+
+ if (!portptr)
+ return -EINVAL;
+
+ spin_lock_irqsave(&rport->lock, flags);
+
+ if (portptr->port_state != FC_OBJSTATE_ONLINE) {
+ spin_unlock_irqrestore(&rport->lock, flags);
+ return -EINVAL;
+ }
+ portptr->port_state = FC_OBJSTATE_DELETED;
+
+ /* tear down all associations to the remote port */
+ list_for_each_entry(ctrl, &rport->ctrl_list, ctrl_list)
+ __nvme_fc_del_ctrl(ctrl);
+
+ spin_unlock_irqrestore(&rport->lock, flags);
+
+ nvme_fc_rport_put(rport);
+ return 0;
+}
+EXPORT_SYMBOL_GPL(nvme_fc_unregister_remoteport);
+
+
+/* *********************** FC-NVME DMA Handling **************************** */
+
+/*
+ * The fcloop device passes in a NULL device pointer. Real LLD's will
+ * pass in a valid device pointer. If NULL is passed to the dma mapping
+ * routines, depending on the platform, it may or may not succeed, and
+ * may crash.
+ *
+ * As such:
+ * Wrapper all the dma routines and check the dev pointer.
+ *
+ * If simple mappings (return just a dma address, we'll noop them,
+ * returning a dma address of 0.
+ *
+ * On more complex mappings (dma_map_sg), a pseudo routine fills
+ * in the scatter list, setting all dma addresses to 0.
+ */
+
+static inline dma_addr_t
+fc_dma_map_single(struct device *dev, void *ptr, size_t size,
+ enum dma_data_direction dir)
+{
+ return dev ? dma_map_single(dev, ptr, size, dir) : (dma_addr_t)0L;
+}
+
+static inline int
+fc_dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
+{
+ return dev ? dma_mapping_error(dev, dma_addr) : 0;
+}
+
+static inline void
+fc_dma_unmap_single(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_unmap_single(dev, addr, size, dir);
+}
+
+static inline void
+fc_dma_sync_single_for_cpu(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_sync_single_for_cpu(dev, addr, size, dir);
+}
+
+static inline void
+fc_dma_sync_single_for_device(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_sync_single_for_device(dev, addr, size, dir);
+}
+
+/* pseudo dma_map_sg call */
+static int
+fc_map_sg(struct scatterlist *sg, int nents)
+{
+ struct scatterlist *s;
+ int i;
+
+ WARN_ON(nents == 0 || sg[0].length == 0);
+
+ for_each_sg(sg, s, nents, i) {
+ s->dma_address = 0L;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+ s->dma_length = s->length;
+#endif
+ }
+ return nents;
+}
+
+static inline int
+fc_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction dir)
+{
+ return dev ? dma_map_sg(dev, sg, nents, dir) : fc_map_sg(sg, nents);
+}
+
+static inline void
+fc_dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_unmap_sg(dev, sg, nents, dir);
+}
+
+
+/* *********************** FC-NVME LS Handling **************************** */
+
+static void nvme_fc_ctrl_put(struct nvme_fc_ctrl *);
+static int nvme_fc_ctrl_get(struct nvme_fc_ctrl *);
+
+
+static void
+__nvme_fc_finish_ls_req(struct nvme_fc_ctrl *ctrl,
+ struct nvmefc_ls_req_op *lsop)
+{
+ struct nvmefc_ls_req *lsreq = &lsop->ls_req;
+ unsigned long flags;
+
+ spin_lock_irqsave(&ctrl->lock, flags);
+
+ if (!lsop->req_queued) {
+ spin_unlock_irqrestore(&ctrl->lock, flags);
+ return;
+ }
+
+ list_del(&lsop->lsreq_list);
+
+ lsop->req_queued = false;
+
+ spin_unlock_irqrestore(&ctrl->lock, flags);
+
+ fc_dma_unmap_single(ctrl->dev, lsreq->rqstdma,
+ (lsreq->rqstlen + lsreq->rsplen),
+ DMA_BIDIRECTIONAL);
+
+ nvme_fc_ctrl_put(ctrl);
+}
+
+static int
+__nvme_fc_send_ls_req(struct nvme_fc_ctrl *ctrl,
+ struct nvmefc_ls_req_op *lsop,
+ void (*done)(struct nvmefc_ls_req *req, int status))
+{
+ struct nvmefc_ls_req *lsreq = &lsop->ls_req;
+ unsigned long flags;
+ int ret;
+
+ if (!nvme_fc_ctrl_get(ctrl))
+ return -ESHUTDOWN;
+
+ lsreq->done = done;
+ lsop->ctrl = ctrl;
+ lsop->req_queued = false;
+ INIT_LIST_HEAD(&lsop->lsreq_list);
+ init_completion(&lsop->ls_done);
+
+ lsreq->rqstdma = fc_dma_map_single(ctrl->dev, lsreq->rqstaddr,
+ lsreq->rqstlen + lsreq->rsplen,
+ DMA_BIDIRECTIONAL);
+ if (fc_dma_mapping_error(ctrl->dev, lsreq->rqstdma)) {
+ nvme_fc_ctrl_put(ctrl);
+ dev_err(ctrl->dev,
+ "els request command failed EFAULT.\n");
+ return -EFAULT;
+ }
+ lsreq->rspdma = lsreq->rqstdma + lsreq->rqstlen;
+
+ spin_lock_irqsave(&ctrl->lock, flags);
+
+ list_add_tail(&lsop->lsreq_list, &ctrl->ls_req_list);
+
+ lsop->req_queued = true;
+
+ spin_unlock_irqrestore(&ctrl->lock, flags);
+
+ ret = ctrl->lport->ops->ls_req(&ctrl->lport->localport,
+ &ctrl->rport->remoteport, lsreq);
+ if (ret)
+ lsop->ls_error = ret;
+
+ return ret;
+}
+
+static void
+nvme_fc_send_ls_req_done(struct nvmefc_ls_req *lsreq, int status)
+{
+ struct nvmefc_ls_req_op *lsop = ls_req_to_lsop(lsreq);
+
+ lsop->ls_error = status;
+ complete(&lsop->ls_done);
+}
+
+static int
+nvme_fc_send_ls_req(struct nvme_fc_ctrl *ctrl, struct nvmefc_ls_req_op *lsop)
+{
+ struct nvmefc_ls_req *lsreq = &lsop->ls_req;
+ struct fcnvme_ls_rjt *rjt = lsreq->rspaddr;
+ int ret;
+
+ ret = __nvme_fc_send_ls_req(ctrl, lsop, nvme_fc_send_ls_req_done);
+
+ if (!ret)
+ /*
+ * No timeout/not interruptible as we need the struct
+ * to exist until the lldd calls us back. Thus mandate
+ * wait until driver calls back. lldd responsible for
+ * the timeout action
+ */
+ wait_for_completion(&lsop->ls_done);
+
+ __nvme_fc_finish_ls_req(ctrl, lsop);
+
+ if (ret) {
+ dev_err(ctrl->dev,
+ "ls request command failed (%d).\n", ret);
+ return ret;
+ }
+
+ /* ACC or RJT payload ? */
+ if (rjt->w0.ls_cmd == FCNVME_LS_RJT)
+ return -ENXIO;
+
+ return 0;
+}
+
+static void
+nvme_fc_send_ls_req_async(struct nvme_fc_ctrl *ctrl,
+ struct nvmefc_ls_req_op *lsop,
+ void (*done)(struct nvmefc_ls_req *req, int status))
+{
+ int ret;
+
+ ret = __nvme_fc_send_ls_req(ctrl, lsop, done);
+
+ /* don't wait for completion */
+
+ if (ret)
+ done(&lsop->ls_req, ret);
+}
+
+/* Validation Error indexes into the string table below */
+enum {
+ VERR_NO_ERROR = 0,
+ VERR_LSACC = 1,
+ VERR_LSDESC_RQST = 2,
+ VERR_LSDESC_RQST_LEN = 3,
+ VERR_ASSOC_ID = 4,
+ VERR_ASSOC_ID_LEN = 5,
+ VERR_CONN_ID = 6,
+ VERR_CONN_ID_LEN = 7,
+ VERR_CR_ASSOC = 8,
+ VERR_CR_ASSOC_ACC_LEN = 9,
+ VERR_CR_CONN = 10,
+ VERR_CR_CONN_ACC_LEN = 11,
+ VERR_DISCONN = 12,
+ VERR_DISCONN_ACC_LEN = 13,
+};
+
+static char *validation_errors[] = {
+ "OK",
+ "Not LS_ACC",
+ "Not LSDESC_RQST",
+ "Bad LSDESC_RQST Length",
+ "Not Association ID",
+ "Bad Association ID Length",
+ "Not Connection ID",
+ "Bad Connection ID Length",
+ "Not CR_ASSOC Rqst",
+ "Bad CR_ASSOC ACC Length",
+ "Not CR_CONN Rqst",
+ "Bad CR_CONN ACC Length",
+ "Not Disconnect Rqst",
+ "Bad Disconnect ACC Length",
+};
+
+static int
+nvme_fc_connect_admin_queue(struct nvme_fc_ctrl *ctrl,
+ struct nvme_fc_queue *queue, u16 qsize, u16 ersp_ratio)
+{
+ struct nvmefc_ls_req_op *lsop;
+ struct nvmefc_ls_req *lsreq;
+ struct fcnvme_ls_cr_assoc_rqst *assoc_rqst;
+ struct fcnvme_ls_cr_assoc_acc *assoc_acc;
+ int ret, fcret = 0;
+
+ lsop = kzalloc((sizeof(*lsop) +
+ ctrl->lport->ops->lsrqst_priv_sz +
+ sizeof(*assoc_rqst) + sizeof(*assoc_acc)), GFP_KERNEL);
+ if (!lsop) {
+ ret = -ENOMEM;
+ goto out_no_memory;
+ }
+ lsreq = &lsop->ls_req;
+
+ lsreq->private = (void *)&lsop[1];
+ assoc_rqst = (struct fcnvme_ls_cr_assoc_rqst *)
+ (lsreq->private + ctrl->lport->ops->lsrqst_priv_sz);
+ assoc_acc = (struct fcnvme_ls_cr_assoc_acc *)&assoc_rqst[1];
+
+ assoc_rqst->w0.ls_cmd = FCNVME_LS_CREATE_ASSOCIATION;
+ assoc_rqst->desc_list_len =
+ cpu_to_be32(sizeof(struct fcnvme_lsdesc_cr_assoc_cmd));
+
+ assoc_rqst->assoc_cmd.desc_tag =
+ cpu_to_be32(FCNVME_LSDESC_CREATE_ASSOC_CMD);
+ assoc_rqst->assoc_cmd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_cr_assoc_cmd));
+
+ assoc_rqst->assoc_cmd.ersp_ratio = cpu_to_be16(ersp_ratio);
+ assoc_rqst->assoc_cmd.sqsize = cpu_to_be16(qsize);
+ /* Linux supports only Dynamic controllers */
+ assoc_rqst->assoc_cmd.cntlid = cpu_to_be16(0xffff);
+ memcpy(&assoc_rqst->assoc_cmd.hostid, &ctrl->ctrl.opts->host->id,
+ min_t(size_t, FCNVME_ASSOC_HOSTID_LEN, sizeof(uuid_be)));
+ strncpy(assoc_rqst->assoc_cmd.hostnqn, ctrl->ctrl.opts->host->nqn,
+ min(FCNVME_ASSOC_HOSTNQN_LEN, NVMF_NQN_SIZE));
+ strncpy(assoc_rqst->assoc_cmd.subnqn, ctrl->ctrl.opts->subsysnqn,
+ min(FCNVME_ASSOC_SUBNQN_LEN, NVMF_NQN_SIZE));
+
+ lsop->queue = queue;
+ lsreq->rqstaddr = assoc_rqst;
+ lsreq->rqstlen = sizeof(*assoc_rqst);
+ lsreq->rspaddr = assoc_acc;
+ lsreq->rsplen = sizeof(*assoc_acc);
+ lsreq->timeout = NVME_FC_CONNECT_TIMEOUT_SEC;
+
+ ret = nvme_fc_send_ls_req(ctrl, lsop);
+ if (ret)
+ goto out_free_buffer;
+
+ /* process connect LS completion */
+
+ /* validate the ACC response */
+ if (assoc_acc->hdr.w0.ls_cmd != FCNVME_LS_ACC)
+ fcret = VERR_LSACC;
+ if (assoc_acc->hdr.desc_list_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_cr_assoc_acc)))
+ fcret = VERR_CR_ASSOC_ACC_LEN;
+ if (assoc_acc->hdr.rqst.desc_tag != cpu_to_be32(FCNVME_LSDESC_RQST))
+ fcret = VERR_LSDESC_RQST;
+ else if (assoc_acc->hdr.rqst.desc_len !=
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_rqst)))
+ fcret = VERR_LSDESC_RQST_LEN;
+ else if (assoc_acc->hdr.rqst.w0.ls_cmd != FCNVME_LS_CREATE_ASSOCIATION)
+ fcret = VERR_CR_ASSOC;
+ else if (assoc_acc->associd.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_ASSOC_ID))
+ fcret = VERR_ASSOC_ID;
+ else if (assoc_acc->associd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id)))
+ fcret = VERR_ASSOC_ID_LEN;
+ else if (assoc_acc->connectid.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_CONN_ID))
+ fcret = VERR_CONN_ID;
+ else if (assoc_acc->connectid.desc_len !=
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_conn_id)))
+ fcret = VERR_CONN_ID_LEN;
+
+ if (fcret) {
+ ret = -EBADF;
+ dev_err(ctrl->dev,
+ "q %d connect failed: %s\n",
+ queue->qnum, validation_errors[fcret]);
+ } else {
+ ctrl->association_id =
+ be64_to_cpu(assoc_acc->associd.association_id);
+ queue->connection_id =
+ be64_to_cpu(assoc_acc->connectid.connection_id);
+ set_bit(NVME_FC_Q_CONNECTED, &queue->flags);
+ }
+
+out_free_buffer:
+ kfree(lsop);
+out_no_memory:
+ if (ret)
+ dev_err(ctrl->dev,
+ "queue %d connect admin queue failed (%d).\n",
+ queue->qnum, ret);
+ return ret;
+}
+
+static int
+nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
+ u16 qsize, u16 ersp_ratio)
+{
+ struct nvmefc_ls_req_op *lsop;
+ struct nvmefc_ls_req *lsreq;
+ struct fcnvme_ls_cr_conn_rqst *conn_rqst;
+ struct fcnvme_ls_cr_conn_acc *conn_acc;
+ int ret, fcret = 0;
+
+ lsop = kzalloc((sizeof(*lsop) +
+ ctrl->lport->ops->lsrqst_priv_sz +
+ sizeof(*conn_rqst) + sizeof(*conn_acc)), GFP_KERNEL);
+ if (!lsop) {
+ ret = -ENOMEM;
+ goto out_no_memory;
+ }
+ lsreq = &lsop->ls_req;
+
+ lsreq->private = (void *)&lsop[1];
+ conn_rqst = (struct fcnvme_ls_cr_conn_rqst *)
+ (lsreq->private + ctrl->lport->ops->lsrqst_priv_sz);
+ conn_acc = (struct fcnvme_ls_cr_conn_acc *)&conn_rqst[1];
+
+ conn_rqst->w0.ls_cmd = FCNVME_LS_CREATE_CONNECTION;
+ conn_rqst->desc_list_len = cpu_to_be32(
+ sizeof(struct fcnvme_lsdesc_assoc_id) +
+ sizeof(struct fcnvme_lsdesc_cr_conn_cmd));
+
+ conn_rqst->associd.desc_tag = cpu_to_be32(FCNVME_LSDESC_ASSOC_ID);
+ conn_rqst->associd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id));
+ conn_rqst->associd.association_id = cpu_to_be64(ctrl->association_id);
+ conn_rqst->connect_cmd.desc_tag =
+ cpu_to_be32(FCNVME_LSDESC_CREATE_CONN_CMD);
+ conn_rqst->connect_cmd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_cr_conn_cmd));
+ conn_rqst->connect_cmd.ersp_ratio = cpu_to_be16(ersp_ratio);
+ conn_rqst->connect_cmd.qid = cpu_to_be16(queue->qnum);
+ conn_rqst->connect_cmd.sqsize = cpu_to_be16(qsize);
+
+ lsop->queue = queue;
+ lsreq->rqstaddr = conn_rqst;
+ lsreq->rqstlen = sizeof(*conn_rqst);
+ lsreq->rspaddr = conn_acc;
+ lsreq->rsplen = sizeof(*conn_acc);
+ lsreq->timeout = NVME_FC_CONNECT_TIMEOUT_SEC;
+
+ ret = nvme_fc_send_ls_req(ctrl, lsop);
+ if (ret)
+ goto out_free_buffer;
+
+ /* process connect LS completion */
+
+ /* validate the ACC response */
+ if (conn_acc->hdr.w0.ls_cmd != FCNVME_LS_ACC)
+ fcret = VERR_LSACC;
+ if (conn_acc->hdr.desc_list_len !=
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_ls_cr_conn_acc)))
+ fcret = VERR_CR_CONN_ACC_LEN;
+ if (conn_acc->hdr.rqst.desc_tag != cpu_to_be32(FCNVME_LSDESC_RQST))
+ fcret = VERR_LSDESC_RQST;
+ else if (conn_acc->hdr.rqst.desc_len !=
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_rqst)))
+ fcret = VERR_LSDESC_RQST_LEN;
+ else if (conn_acc->hdr.rqst.w0.ls_cmd != FCNVME_LS_CREATE_CONNECTION)
+ fcret = VERR_CR_CONN;
+ else if (conn_acc->connectid.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_CONN_ID))
+ fcret = VERR_CONN_ID;
+ else if (conn_acc->connectid.desc_len !=
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_conn_id)))
+ fcret = VERR_CONN_ID_LEN;
+
+ if (fcret) {
+ ret = -EBADF;
+ dev_err(ctrl->dev,
+ "q %d connect failed: %s\n",
+ queue->qnum, validation_errors[fcret]);
+ } else {
+ queue->connection_id =
+ be64_to_cpu(conn_acc->connectid.connection_id);
+ set_bit(NVME_FC_Q_CONNECTED, &queue->flags);
+ }
+
+out_free_buffer:
+ kfree(lsop);
+out_no_memory:
+ if (ret)
+ dev_err(ctrl->dev,
+ "queue %d connect command failed (%d).\n",
+ queue->qnum, ret);
+ return ret;
+}
+
+static void
+nvme_fc_disconnect_assoc_done(struct nvmefc_ls_req *lsreq, int status)
+{
+ struct nvmefc_ls_req_op *lsop = ls_req_to_lsop(lsreq);
+ struct nvme_fc_ctrl *ctrl = lsop->ctrl;
+
+ __nvme_fc_finish_ls_req(ctrl, lsop);
+
+ if (status)
+ dev_err(ctrl->dev,
+ "disconnect assoc ls request command failed (%d).\n",
+ status);
+
+ /* fc-nvme iniator doesn't care about success or failure of cmd */
+
+ kfree(lsop);
+}
+
+/*
+ * This routine sends a FC-NVME LS to disconnect (aka terminate)
+ * the FC-NVME Association. Terminating the association also
+ * terminates the FC-NVME connections (per queue, both admin and io
+ * queues) that are part of the association. E.g. things are torn
+ * down, and the related FC-NVME Association ID and Connection IDs
+ * become invalid.
+ *
+ * The behavior of the fc-nvme initiator is such that it's
+ * understanding of the association and connections will implicitly
+ * be torn down. The action is implicit as it may be due to a loss of
+ * connectivity with the fc-nvme target, so you may never get a
+ * response even if you tried. As such, the action of this routine
+ * is to asynchronously send the LS, ignore any results of the LS, and
+ * continue on with terminating the association. If the fc-nvme target
+ * is present and receives the LS, it too can tear down.
+ */
+static void
+nvme_fc_xmt_disconnect_assoc(struct nvme_fc_ctrl *ctrl)
+{
+ struct fcnvme_ls_disconnect_rqst *discon_rqst;
+ struct fcnvme_ls_disconnect_acc *discon_acc;
+ struct nvmefc_ls_req_op *lsop;
+ struct nvmefc_ls_req *lsreq;
+
+ lsop = kzalloc((sizeof(*lsop) +
+ ctrl->lport->ops->lsrqst_priv_sz +
+ sizeof(*discon_rqst) + sizeof(*discon_acc)),
+ GFP_KERNEL);
+ if (!lsop)
+ /* couldn't sent it... too bad */
+ return;
+
+ lsreq = &lsop->ls_req;
+
+ lsreq->private = (void *)&lsop[1];
+ discon_rqst = (struct fcnvme_ls_disconnect_rqst *)
+ (lsreq->private + ctrl->lport->ops->lsrqst_priv_sz);
+ discon_acc = (struct fcnvme_ls_disconnect_acc *)&discon_rqst[1];
+
+ discon_rqst->w0.ls_cmd = FCNVME_LS_DISCONNECT;
+ discon_rqst->desc_list_len = cpu_to_be32(
+ sizeof(struct fcnvme_lsdesc_assoc_id) +
+ sizeof(struct fcnvme_lsdesc_disconn_cmd));
+
+ discon_rqst->associd.desc_tag = cpu_to_be32(FCNVME_LSDESC_ASSOC_ID);
+ discon_rqst->associd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id));
+
+ discon_rqst->associd.association_id = cpu_to_be64(ctrl->association_id);
+
+ discon_rqst->discon_cmd.desc_tag = cpu_to_be32(
+ FCNVME_LSDESC_DISCONN_CMD);
+ discon_rqst->discon_cmd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_disconn_cmd));
+ discon_rqst->discon_cmd.scope = FCNVME_DISCONN_ASSOCIATION;
+ discon_rqst->discon_cmd.id = cpu_to_be64(ctrl->association_id);
+
+ lsreq->rqstaddr = discon_rqst;
+ lsreq->rqstlen = sizeof(*discon_rqst);
+ lsreq->rspaddr = discon_acc;
+ lsreq->rsplen = sizeof(*discon_acc);
+ lsreq->timeout = NVME_FC_CONNECT_TIMEOUT_SEC;
+
+ nvme_fc_send_ls_req_async(ctrl, lsop, nvme_fc_disconnect_assoc_done);
+
+ /* only meaningful part to terminating the association */
+ ctrl->association_id = 0;
+}
+
+
+/* *********************** NVME Ctrl Routines **************************** */
+
+
+static int
+nvme_fc_reinit_request(void *data, struct request *rq)
+{
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+ struct nvme_fc_cmd_iu *cmdiu = &op->cmd_iu;
+
+ memset(cmdiu, 0, sizeof(*cmdiu));
+ cmdiu->scsi_id = NVME_CMD_SCSI_ID;
+ cmdiu->fc_id = NVME_CMD_FC_ID;
+ cmdiu->iu_len = cpu_to_be16(sizeof(*cmdiu) / sizeof(u32));
+ memset(&op->rsp_iu, 0, sizeof(op->rsp_iu));
+
+ return 0;
+}
+
+static void
+__nvme_fc_exit_request(struct nvme_fc_ctrl *ctrl,
+ struct nvme_fc_fcp_op *op)
+{
+ fc_dma_unmap_single(ctrl->lport->dev, op->fcp_req.rspdma,
+ sizeof(op->rsp_iu), DMA_FROM_DEVICE);
+ fc_dma_unmap_single(ctrl->lport->dev, op->fcp_req.cmddma,
+ sizeof(op->cmd_iu), DMA_TO_DEVICE);
+
+ atomic_set(&op->state, FCPOP_STATE_UNINIT);
+}
+
+static void
+nvme_fc_exit_request(void *data, struct request *rq,
+ unsigned int hctx_idx, unsigned int rq_idx)
+{
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+
+ return __nvme_fc_exit_request(data, op);
+}
+
+static void
+nvme_fc_exit_aen_ops(struct nvme_fc_ctrl *ctrl)
+{
+ struct nvme_fc_fcp_op *aen_op = ctrl->aen_ops;
+ int i;
+
+ for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) {
+ if (atomic_read(&aen_op->state) == FCPOP_STATE_UNINIT)
+ continue;
+ __nvme_fc_exit_request(ctrl, aen_op);
+ nvme_fc_ctrl_put(ctrl);
+ }
+}
+
+void
+nvme_fc_fcpio_done(struct nvmefc_fcp_req *req)
+{
+ struct nvme_fc_fcp_op *op = fcp_req_to_fcp_op(req);
+ struct request *rq = op->rq;
+ struct nvmefc_fcp_req *freq = &op->fcp_req;
+ struct nvme_fc_ctrl *ctrl = op->ctrl;
+ struct nvme_fc_queue *queue = op->queue;
+ struct nvme_completion *cqe = &op->rsp_iu.cqe;
+ u16 status;
+
+ /*
+ * WARNING:
+ * The current linux implementation of a nvme controller
+ * allocates a single tag set for all io queues and sizes
+ * the io queues to fully hold all possible tags. Thus, the
+ * implementation does not reference or care about the sqhd
+ * value as it never needs to use the sqhd/sqtail pointers
+ * for submission pacing.
+ *
+ * This affects the FC-NVME implementation in two ways:
+ * 1) As the value doesn't matter, we don't need to waste
+ * cycles extracting it from ERSPs and stamping it in the
+ * cases where the transport fabricates CQEs on successful
+ * completions.
+ * 2) The FC-NVME implementation requires that delivery of
+ * ERSP completions are to go back to the nvme layer in order
+ * relative to the rsn, such that the sqhd value will always
+ * be "in order" for the nvme layer. As the nvme layer in
+ * linux doesn't care about sqhd, there's no need to return
+ * them in order.
+ *
+ * Additionally:
+ * As the core nvme layer in linux currently does not look at
+ * every field in the cqe - in cases where the FC transport must
+ * fabricate a CQE, the following fields will not be set as they
+ * are not referenced:
+ * cqe.sqid, cqe.sqhd, cqe.command_id
+ */
+
+ fc_dma_sync_single_for_cpu(ctrl->lport->dev, op->fcp_req.rspdma,
+ sizeof(op->rsp_iu), DMA_FROM_DEVICE);
+
+ if (atomic_read(&op->state) == FCPOP_STATE_ABORTED)
+ status = NVME_SC_ABORT_REQ | NVME_SC_DNR;
+ else
+ status = freq->status;
+
+ /*
+ * For the linux implementation, if we have an unsuccesful
+ * status, they blk-mq layer can typically be called with the
+ * non-zero status and the content of the cqe isn't important.
+ */
+ if (status)
+ goto done;
+
+ /*
+ * command completed successfully relative to the wire
+ * protocol. However, validate anything received and
+ * extract the status and result from the cqe (create it
+ * where necessary).
+ */
+
+ switch (freq->rcv_rsplen) {
+
+ case 0:
+ case NVME_FC_SIZEOF_ZEROS_RSP:
+ /*
+ * No response payload or 12 bytes of payload (which
+ * should all be zeros) are considered successful and
+ * no payload in the CQE by the transport.
+ */
+ if (freq->transferred_length !=
+ be32_to_cpu(op->cmd_iu.data_len)) {
+ status = -EIO;
+ goto done;
+ }
+ op->nreq.result.u64 = 0;
+ break;
+
+ case sizeof(struct nvme_fc_ersp_iu):
+ /*
+ * The ERSP IU contains a full completion with CQE.
+ * Validate ERSP IU and look at cqe.
+ */
+ if (unlikely(be16_to_cpu(op->rsp_iu.iu_len) !=
+ (freq->rcv_rsplen / 4) ||
+ be32_to_cpu(op->rsp_iu.xfrd_len) !=
+ freq->transferred_length ||
+ op->rqno != le16_to_cpu(cqe->command_id))) {
+ status = -EIO;
+ goto done;
+ }
+ op->nreq.result = cqe->result;
+ status = le16_to_cpu(cqe->status) >> 1;
+ break;
+
+ default:
+ status = -EIO;
+ goto done;
+ }
+
+done:
+ if (!queue->qnum && op->rqno >= AEN_CMDID_BASE) {
+ nvme_complete_async_event(&queue->ctrl->ctrl, status,
+ &op->nreq.result);
+ nvme_fc_ctrl_put(ctrl);
+ return;
+ }
+
+ blk_mq_complete_request(rq, status);
+}
+
+static int
+__nvme_fc_init_request(struct nvme_fc_ctrl *ctrl,
+ struct nvme_fc_queue *queue, struct nvme_fc_fcp_op *op,
+ struct request *rq, u32 rqno)
+{
+ struct nvme_fc_cmd_iu *cmdiu = &op->cmd_iu;
+ int ret = 0;
+
+ memset(op, 0, sizeof(*op));
+ op->fcp_req.cmdaddr = &op->cmd_iu;
+ op->fcp_req.cmdlen = sizeof(op->cmd_iu);
+ op->fcp_req.rspaddr = &op->rsp_iu;
+ op->fcp_req.rsplen = sizeof(op->rsp_iu);
+ op->fcp_req.done = nvme_fc_fcpio_done;
+ op->fcp_req.first_sgl = (struct scatterlist *)&op[1];
+ op->fcp_req.private = &op->fcp_req.first_sgl[SG_CHUNK_SIZE];
+ op->ctrl = ctrl;
+ op->queue = queue;
+ op->rq = rq;
+ op->rqno = rqno;
+
+ cmdiu->scsi_id = NVME_CMD_SCSI_ID;
+ cmdiu->fc_id = NVME_CMD_FC_ID;
+ cmdiu->iu_len = cpu_to_be16(sizeof(*cmdiu) / sizeof(u32));
+
+ op->fcp_req.cmddma = fc_dma_map_single(ctrl->lport->dev,
+ &op->cmd_iu, sizeof(op->cmd_iu), DMA_TO_DEVICE);
+ if (fc_dma_mapping_error(ctrl->lport->dev, op->fcp_req.cmddma)) {
+ dev_err(ctrl->dev,
+ "FCP Op failed - cmdiu dma mapping failed.\n");
+ ret = EFAULT;
+ goto out_on_error;
+ }
+
+ op->fcp_req.rspdma = fc_dma_map_single(ctrl->lport->dev,
+ &op->rsp_iu, sizeof(op->rsp_iu),
+ DMA_FROM_DEVICE);
+ if (fc_dma_mapping_error(ctrl->lport->dev, op->fcp_req.rspdma)) {
+ dev_err(ctrl->dev,
+ "FCP Op failed - rspiu dma mapping failed.\n");
+ ret = EFAULT;
+ }
+
+ atomic_set(&op->state, FCPOP_STATE_IDLE);
+out_on_error:
+ return ret;
+}
+
+static int
+nvme_fc_init_request(void *data, struct request *rq,
+ unsigned int hctx_idx, unsigned int rq_idx,
+ unsigned int numa_node)
+{
+ struct nvme_fc_ctrl *ctrl = data;
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+ struct nvme_fc_queue *queue = &ctrl->queues[hctx_idx+1];
+
+ return __nvme_fc_init_request(ctrl, queue, op, rq, queue->rqcnt++);
+}
+
+static int
+nvme_fc_init_admin_request(void *data, struct request *rq,
+ unsigned int hctx_idx, unsigned int rq_idx,
+ unsigned int numa_node)
+{
+ struct nvme_fc_ctrl *ctrl = data;
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+ struct nvme_fc_queue *queue = &ctrl->queues[0];
+
+ return __nvme_fc_init_request(ctrl, queue, op, rq, queue->rqcnt++);
+}
+
+static int
+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;
+ int i, ret;
+
+ aen_op = ctrl->aen_ops;
+ for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) {
+ 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)
+ return ret;
+
+ memset(sqe, 0, sizeof(*sqe));
+ sqe->common.opcode = nvme_admin_async_event;
+ sqe->common.command_id = AEN_CMDID_BASE + i;
+ }
+ return 0;
+}
+
+
+static inline void
+__nvme_fc_init_hctx(struct blk_mq_hw_ctx *hctx, struct nvme_fc_ctrl *ctrl,
+ unsigned int qidx)
+{
+ struct nvme_fc_queue *queue = &ctrl->queues[qidx];
+
+ hctx->driver_data = queue;
+ queue->hctx = hctx;
+}
+
+static int
+nvme_fc_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
+ unsigned int hctx_idx)
+{
+ struct nvme_fc_ctrl *ctrl = data;
+
+ __nvme_fc_init_hctx(hctx, ctrl, hctx_idx + 1);
+
+ return 0;
+}
+
+static int
+nvme_fc_init_admin_hctx(struct blk_mq_hw_ctx *hctx, void *data,
+ unsigned int hctx_idx)
+{
+ struct nvme_fc_ctrl *ctrl = data;
+
+ __nvme_fc_init_hctx(hctx, ctrl, hctx_idx);
+
+ return 0;
+}
+
+static void
+nvme_fc_init_queue(struct nvme_fc_ctrl *ctrl, int idx, size_t queue_size)
+{
+ struct nvme_fc_queue *queue;
+
+ queue = &ctrl->queues[idx];
+ memset(queue, 0, sizeof(*queue));
+ queue->ctrl = ctrl;
+ queue->qnum = idx;
+ atomic_set(&queue->csn, 1);
+ queue->dev = ctrl->dev;
+
+ if (idx > 0)
+ queue->cmnd_capsule_len = ctrl->ctrl.ioccsz * 16;
+ else
+ queue->cmnd_capsule_len = sizeof(struct nvme_command);
+
+ queue->queue_size = queue_size;
+
+ /*
+ * Considered whether we should allocate buffers for all SQEs
+ * and CQEs and dma map them - mapping their respective entries
+ * into the request structures (kernel vm addr and dma address)
+ * thus the driver could use the buffers/mappings directly.
+ * It only makes sense if the LLDD would use them for its
+ * messaging api. It's very unlikely most adapter api's would use
+ * a native NVME sqe/cqe. More reasonable if FC-NVME IU payload
+ * structures were used instead.
+ */
+}
+
+/*
+ * This routine terminates a queue at the transport level.
+ * The transport has already ensured that all outstanding ios on
+ * the queue have been terminated.
+ * The transport will send a Disconnect LS request to terminate
+ * the queue's connection. Termination of the admin queue will also
+ * terminate the association at the target.
+ */
+static void
+nvme_fc_free_queue(struct nvme_fc_queue *queue)
+{
+ if (!test_and_clear_bit(NVME_FC_Q_CONNECTED, &queue->flags))
+ return;
+
+ /*
+ * Current implementation never disconnects a single queue.
+ * It always terminates a whole association. So there is never
+ * a disconnect(queue) LS sent to the target.
+ */
+
+ queue->connection_id = 0;
+ clear_bit(NVME_FC_Q_CONNECTED, &queue->flags);
+}
+
+static void
+__nvme_fc_delete_hw_queue(struct nvme_fc_ctrl *ctrl,
+ struct nvme_fc_queue *queue, unsigned int qidx)
+{
+ if (ctrl->lport->ops->delete_queue)
+ ctrl->lport->ops->delete_queue(&ctrl->lport->localport, qidx,
+ queue->lldd_handle);
+ 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)
+{
+ int i;
+
+ for (i = 1; i < ctrl->queue_count; i++)
+ nvme_fc_free_queue(&ctrl->queues[i]);
+}
+
+static int
+__nvme_fc_create_hw_queue(struct nvme_fc_ctrl *ctrl,
+ struct nvme_fc_queue *queue, unsigned int qidx, u16 qsize)
+{
+ int ret = 0;
+
+ queue->lldd_handle = NULL;
+ if (ctrl->lport->ops->create_queue)
+ ret = ctrl->lport->ops->create_queue(&ctrl->lport->localport,
+ qidx, qsize, &queue->lldd_handle);
+
+ return ret;
+}
+
+static void
+nvme_fc_delete_hw_io_queues(struct nvme_fc_ctrl *ctrl)
+{
+ struct nvme_fc_queue *queue = &ctrl->queues[ctrl->queue_count - 1];
+ int i;
+
+ for (i = ctrl->queue_count - 1; i >= 1; i--, queue--)
+ __nvme_fc_delete_hw_queue(ctrl, queue, i);
+}
+
+static int
+nvme_fc_create_hw_io_queues(struct nvme_fc_ctrl *ctrl, u16 qsize)
+{
+ struct nvme_fc_queue *queue = &ctrl->queues[1];
+ int i, j, ret;
+
+ for (i = 1; i < ctrl->queue_count; i++, queue++) {
+ ret = __nvme_fc_create_hw_queue(ctrl, queue, i, qsize);
+ if (ret) {
+ for (j = i-1; j >= 0; j--)
+ __nvme_fc_delete_hw_queue(ctrl,
+ &ctrl->queues[j], j);
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
+static int
+nvme_fc_connect_io_queues(struct nvme_fc_ctrl *ctrl, u16 qsize)
+{
+ int i, ret = 0;
+
+ for (i = 1; i < ctrl->queue_count; i++) {
+ ret = nvme_fc_connect_queue(ctrl, &ctrl->queues[i], qsize,
+ (qsize / 5));
+ if (ret)
+ break;
+ ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
+ if (ret)
+ break;
+ }
+
+ return ret;
+}
+
+static void
+nvme_fc_init_io_queues(struct nvme_fc_ctrl *ctrl)
+{
+ int i;
+
+ for (i = 1; i < ctrl->queue_count; i++)
+ nvme_fc_init_queue(ctrl, i, ctrl->ctrl.sqsize);
+}
+
+static void
+nvme_fc_ctrl_free(struct kref *ref)
+{
+ struct nvme_fc_ctrl *ctrl =
+ 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);
+ }
+
+ 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);
+ kfree(ctrl);
+}
+
+static void
+nvme_fc_ctrl_put(struct nvme_fc_ctrl *ctrl)
+{
+ kref_put(&ctrl->ref, nvme_fc_ctrl_free);
+}
+
+static int
+nvme_fc_ctrl_get(struct nvme_fc_ctrl *ctrl)
+{
+ return kref_get_unless_zero(&ctrl->ref);
+}
+
+/*
+ * All accesses from nvme core layer done - can now free the
+ * controller. Called after last nvme_put_ctrl() call
+ */
+static void
+nvme_fc_free_nvme_ctrl(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)
+{
+ int state;
+
+ state = atomic_xchg(&op->state, FCPOP_STATE_ABORTED);
+ if (state != FCPOP_STATE_ACTIVE) {
+ atomic_set(&op->state, state);
+ return -ECANCELED; /* fail */
+ }
+
+ ctrl->lport->ops->fcp_abort(&ctrl->lport->localport,
+ &ctrl->rport->remoteport,
+ op->queue->lldd_handle,
+ &op->fcp_req);
+
+ return 0;
+}
+
+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);
+ struct nvme_fc_ctrl *ctrl = op->ctrl;
+ int ret;
+
+ if (reserved)
+ return BLK_EH_RESET_TIMER;
+
+ ret = __nvme_fc_abort_op(ctrl, op);
+ if (ret)
+ /* io wasn't active to abort consider it done */
+ 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.
+ */
+
+ return BLK_EH_HANDLED;
+}
+
+static int
+nvme_fc_map_data(struct nvme_fc_ctrl *ctrl, struct request *rq,
+ struct nvme_fc_fcp_op *op)
+{
+ struct nvmefc_fcp_req *freq = &op->fcp_req;
+ u32 map_len = nvme_map_len(rq);
+ enum dma_data_direction dir;
+ int ret;
+
+ freq->sg_cnt = 0;
+
+ if (!map_len)
+ return 0;
+
+ freq->sg_table.sgl = freq->first_sgl;
+ ret = sg_alloc_table_chained(&freq->sg_table, rq->nr_phys_segments,
+ freq->sg_table.sgl);
+ if (ret)
+ return -ENOMEM;
+
+ op->nents = blk_rq_map_sg(rq->q, rq, freq->sg_table.sgl);
+ WARN_ON(op->nents > rq->nr_phys_segments);
+ dir = (rq_data_dir(rq) == WRITE) ? DMA_TO_DEVICE : DMA_FROM_DEVICE;
+ freq->sg_cnt = fc_dma_map_sg(ctrl->lport->dev, freq->sg_table.sgl,
+ op->nents, dir);
+ if (unlikely(freq->sg_cnt <= 0)) {
+ sg_free_table_chained(&freq->sg_table, true);
+ freq->sg_cnt = 0;
+ return -EFAULT;
+ }
+
+ /*
+ * TODO: blk_integrity_rq(rq) for DIF
+ */
+ return 0;
+}
+
+static void
+nvme_fc_unmap_data(struct nvme_fc_ctrl *ctrl, struct request *rq,
+ struct nvme_fc_fcp_op *op)
+{
+ struct nvmefc_fcp_req *freq = &op->fcp_req;
+
+ if (!freq->sg_cnt)
+ return;
+
+ fc_dma_unmap_sg(ctrl->lport->dev, freq->sg_table.sgl, op->nents,
+ ((rq_data_dir(rq) == WRITE) ?
+ DMA_TO_DEVICE : DMA_FROM_DEVICE));
+
+ nvme_cleanup_cmd(rq);
+
+ sg_free_table_chained(&freq->sg_table, true);
+
+ freq->sg_cnt = 0;
+}
+
+/*
+ * In FC, the queue is a logical thing. At transport connect, the target
+ * creates its "queue" and returns a handle that is to be given to the
+ * target whenever it posts something to the corresponding SQ. When an
+ * SQE is sent on a SQ, FC effectively considers the SQE, or rather the
+ * command contained within the SQE, an io, and assigns a FC exchange
+ * to it. The SQE and the associated SQ handle are sent in the initial
+ * CMD IU sents on the exchange. All transfers relative to the io occur
+ * as part of the exchange. The CQE is the last thing for the io,
+ * which is transferred (explicitly or implicitly) with the RSP IU
+ * sent on the exchange. After the CQE is received, the FC exchange is
+ * terminaed and the Exchange may be used on a different io.
+ *
+ * The transport to LLDD api has the transport making a request for a
+ * new fcp io request to the LLDD. The LLDD then allocates a FC exchange
+ * resource and transfers the command. The LLDD will then process all
+ * steps to complete the io. Upon completion, the transport done routine
+ * is called.
+ *
+ * So - while the operation is outstanding to the LLDD, there is a link
+ * level FC exchange resource that is also outstanding. This must be
+ * considered in all cleanup operations.
+ */
+static int
+nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
+ struct nvme_fc_fcp_op *op, u32 data_len,
+ enum nvmefc_fcp_datadir io_dir)
+{
+ struct nvme_fc_cmd_iu *cmdiu = &op->cmd_iu;
+ struct nvme_command *sqe = &cmdiu->sqe;
+ u32 csn;
+ int ret;
+
+ if (!nvme_fc_ctrl_get(ctrl))
+ return BLK_MQ_RQ_QUEUE_ERROR;
+
+ /* format the FC-NVME CMD IU and fcp_req */
+ cmdiu->connection_id = cpu_to_be64(queue->connection_id);
+ csn = atomic_inc_return(&queue->csn);
+ cmdiu->csn = cpu_to_be32(csn);
+ cmdiu->data_len = cpu_to_be32(data_len);
+ switch (io_dir) {
+ case NVMEFC_FCP_WRITE:
+ cmdiu->flags = FCNVME_CMD_FLAGS_WRITE;
+ break;
+ case NVMEFC_FCP_READ:
+ cmdiu->flags = FCNVME_CMD_FLAGS_READ;
+ break;
+ case NVMEFC_FCP_NODATA:
+ cmdiu->flags = 0;
+ break;
+ }
+ op->fcp_req.payload_length = data_len;
+ op->fcp_req.io_dir = io_dir;
+ op->fcp_req.transferred_length = 0;
+ op->fcp_req.rcv_rsplen = 0;
+ op->fcp_req.status = 0;
+ op->fcp_req.sqid = cpu_to_le16(queue->qnum);
+
+ /*
+ * validate per fabric rules, set fields mandated by fabric spec
+ * as well as those by FC-NVME spec.
+ */
+ WARN_ON_ONCE(sqe->common.metadata);
+ WARN_ON_ONCE(sqe->common.dptr.prp1);
+ WARN_ON_ONCE(sqe->common.dptr.prp2);
+ sqe->common.flags |= NVME_CMD_SGL_METABUF;
+
+ /*
+ * format SQE DPTR field per FC-NVME rules
+ * type=data block descr; subtype=offset;
+ * offset is currently 0.
+ */
+ sqe->rw.dptr.sgl.type = NVME_SGL_FMT_OFFSET;
+ 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 */
+ 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) ?
+ BLK_MQ_RQ_QUEUE_BUSY : BLK_MQ_RQ_QUEUE_ERROR;
+ }
+ }
+
+ fc_dma_sync_single_for_device(ctrl->lport->dev, op->fcp_req.cmddma,
+ sizeof(op->cmd_iu), DMA_TO_DEVICE);
+
+ atomic_set(&op->state, FCPOP_STATE_ACTIVE);
+
+ if (op->rq)
+ blk_mq_start_request(op->rq);
+
+ ret = ctrl->lport->ops->fcp_io(&ctrl->lport->localport,
+ &ctrl->rport->remoteport,
+ 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);
+ }
+ /* else - aen. no cleanup needed */
+
+ nvme_fc_ctrl_put(ctrl);
+
+ if (ret != -EBUSY)
+ return BLK_MQ_RQ_QUEUE_ERROR;
+
+ if (op->rq) {
+ blk_mq_stop_hw_queues(op->rq->q);
+ blk_mq_delay_queue(queue->hctx, NVMEFC_QUEUE_DELAY);
+ }
+ return BLK_MQ_RQ_QUEUE_BUSY;
+ }
+
+ return BLK_MQ_RQ_QUEUE_OK;
+}
+
+static int
+nvme_fc_queue_rq(struct blk_mq_hw_ctx *hctx,
+ const struct blk_mq_queue_data *bd)
+{
+ struct nvme_ns *ns = hctx->queue->queuedata;
+ struct nvme_fc_queue *queue = hctx->driver_data;
+ struct nvme_fc_ctrl *ctrl = queue->ctrl;
+ struct request *rq = bd->rq;
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+ struct nvme_fc_cmd_iu *cmdiu = &op->cmd_iu;
+ struct nvme_command *sqe = &cmdiu->sqe;
+ enum nvmefc_fcp_datadir io_dir;
+ u32 data_len;
+ int ret;
+
+ ret = nvme_setup_cmd(ns, rq, sqe);
+ if (ret)
+ return ret;
+
+ data_len = nvme_map_len(rq);
+ if (data_len)
+ io_dir = ((rq_data_dir(rq) == WRITE) ?
+ NVMEFC_FCP_WRITE : NVMEFC_FCP_READ);
+ else
+ io_dir = NVMEFC_FCP_NODATA;
+
+ return nvme_fc_start_fcp_op(ctrl, queue, op, data_len, io_dir);
+}
+
+static struct blk_mq_tags *
+nvme_fc_tagset(struct nvme_fc_queue *queue)
+{
+ if (queue->qnum == 0)
+ return queue->ctrl->admin_tag_set.tags[queue->qnum];
+
+ return queue->ctrl->tag_set.tags[queue->qnum - 1];
+}
+
+static int
+nvme_fc_poll(struct blk_mq_hw_ctx *hctx, unsigned int tag)
+
+{
+ struct nvme_fc_queue *queue = hctx->driver_data;
+ struct nvme_fc_ctrl *ctrl = queue->ctrl;
+ struct request *req;
+ 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);
+ return 0;
+ }
+
+ op = blk_mq_rq_to_pdu(req);
+
+ if ((atomic_read(&op->state) == FCPOP_STATE_ACTIVE) &&
+ (ctrl->lport->ops->poll_queue))
+ ctrl->lport->ops->poll_queue(&ctrl->lport->localport,
+ queue->lldd_handle);
+
+ return ((atomic_read(&op->state) != FCPOP_STATE_ACTIVE));
+}
+
+static void
+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;
+ int ret;
+
+ if (aer_idx > NVME_FC_NR_AEN_COMMANDS)
+ return;
+
+ aen_op = &ctrl->aen_ops[aer_idx];
+
+ ret = nvme_fc_start_fcp_op(ctrl, aen_op->queue, aen_op, 0,
+ NVMEFC_FCP_NODATA);
+ if (ret)
+ dev_err(ctrl->ctrl.device,
+ "failed async event work [%d]\n", aer_idx);
+}
+
+static void
+nvme_fc_complete_rq(struct request *rq)
+{
+ struct nvme_fc_fcp_op *op = blk_mq_rq_to_pdu(rq);
+ struct nvme_fc_ctrl *ctrl = op->ctrl;
+ int error = 0, state;
+
+ state = atomic_xchg(&op->state, FCPOP_STATE_IDLE);
+
+ nvme_cleanup_cmd(rq);
+
+ nvme_fc_unmap_data(ctrl, rq, op);
+
+ if (unlikely(rq->errors)) {
+ if (nvme_req_needs_retry(rq, rq->errors)) {
+ nvme_requeue_req(rq);
+ return;
+ }
+
+ if (rq->cmd_type == REQ_TYPE_DRV_PRIV)
+ error = rq->errors;
+ else
+ error = nvme_error_status(rq->errors);
+ }
+
+ nvme_fc_ctrl_put(ctrl);
+
+ blk_mq_end_request(rq, error);
+}
+
+static 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 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)
+{
+ u32 segs;
+ int error;
+
+ nvme_fc_init_queue(ctrl, 0, NVME_FC_AQ_BLKMQ_DEPTH);
+
+ 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) + 1, 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;
+}
+
+/*
+ * This routine is used by the transport when it needs to find active
+ * io on a queue that is to be terminated. The transport uses
+ * blk_mq_tagset_busy_itr() to find the busy requests, which then invoke
+ * this routine to kill them on a 1 by 1 basis.
+ *
+ * As FC allocates FC exchange for each io, the transport must contact
+ * the LLDD to terminate the exchange, thus releasing the FC exchange.
+ * After terminating the exchange the LLDD will call the transport's
+ * normal io done path for the request, but it will have an aborted
+ * status. The done path will return the io request back to the block
+ * layer with an error status.
+ */
+static void
+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;
+
+ 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);
+ }
+
+ if (ctrl->ctrl.state == NVME_CTRL_LIVE)
+ nvme_shutdown_ctrl(&ctrl->ctrl);
+
+ /*
+ * 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);
+}
+
+/*
+ * 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 int
+nvme_fc_create_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;
+ }
+
+ ctrl->queue_count = opts->nr_io_queues + 1;
+ if (!opts->nr_io_queues)
+ return 0;
+
+ dev_info(ctrl->ctrl.device, "creating %d I/O queues.\n",
+ opts->nr_io_queues);
+
+ nvme_fc_init_io_queues(ctrl);
+
+ memset(&ctrl->tag_set, 0, sizeof(ctrl->tag_set));
+ ctrl->tag_set.ops = &nvme_fc_mq_ops;
+ ctrl->tag_set.queue_depth = ctrl->ctrl.opts->queue_size;
+ ctrl->tag_set.reserved_tags = 1; /* fabric connect */
+ ctrl->tag_set.numa_node = NUMA_NO_NODE;
+ ctrl->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
+ ctrl->tag_set.cmd_size = sizeof(struct nvme_fc_fcp_op) +
+ (SG_CHUNK_SIZE *
+ sizeof(struct scatterlist)) +
+ ctrl->lport->ops->fcprqst_priv_sz;
+ ctrl->tag_set.driver_data = ctrl;
+ ctrl->tag_set.nr_hw_queues = ctrl->queue_count - 1;
+ ctrl->tag_set.timeout = NVME_IO_TIMEOUT;
+
+ ret = blk_mq_alloc_tag_set(&ctrl->tag_set);
+ if (ret)
+ return ret;
+
+ ctrl->ctrl.tagset = &ctrl->tag_set;
+
+ ctrl->ctrl.connect_q = blk_mq_init_queue(&ctrl->tag_set);
+ if (IS_ERR(ctrl->ctrl.connect_q)) {
+ ret = PTR_ERR(ctrl->ctrl.connect_q);
+ goto out_free_tag_set;
+ }
+
+ ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.opts->queue_size);
+ if (ret)
+ goto out_cleanup_blk_queue;
+
+ 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_cleanup_blk_queue:
+ nvme_stop_keep_alive(&ctrl->ctrl);
+ blk_cleanup_queue(ctrl->ctrl.connect_q);
+out_free_tag_set:
+ blk_mq_free_tag_set(&ctrl->tag_set);
+ nvme_fc_free_io_queues(ctrl);
+
+ /* force put free routine to ignore io queues */
+ ctrl->ctrl.tagset = NULL;
+
+ return ret;
+}
+
+
+static struct nvme_ctrl *
+__nvme_fc_create_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) {
+ ret = -ENOMEM;
+ goto out_fail;
+ }
+
+ idx = ida_simple_get(&nvme_fc_ctrl_cnt, 0, 0, GFP_KERNEL);
+ if (idx < 0) {
+ ret = -ENOSPC;
+ goto out_free_ctrl;
+ }
+
+ ctrl->ctrl.opts = opts;
+ INIT_LIST_HEAD(&ctrl->ctrl_list);
+ INIT_LIST_HEAD(&ctrl->ls_req_list);
+ 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);
+ spin_lock_init(&ctrl->lock);
+
+ /* io queue count */
+ ctrl->queue_count = min_t(unsigned int,
+ opts->nr_io_queues,
+ lport->ops->max_hw_queues);
+ opts->nr_io_queues = ctrl->queue_count; /* so opts has valid value */
+ ctrl->queue_count++; /* +1 for admin queue */
+
+ ctrl->ctrl.sqsize = opts->queue_size - 1;
+ ctrl->ctrl.kato = opts->kato;
+
+ ret = -ENOMEM;
+ ctrl->queues = kcalloc(ctrl->queue_count, sizeof(struct nvme_fc_queue),
+ GFP_KERNEL);
+ if (!ctrl->queues)
+ goto out_uninit_ctrl;
+
+ ret = nvme_fc_configure_admin_queue(ctrl);
+ if (ret)
+ goto out_uninit_ctrl;
+
+ /* sanity checks */
+
+ /* FC-NVME supports 64-byte SQE only */
+ if (ctrl->ctrl.ioccsz != 4) {
+ dev_err(ctrl->ctrl.device, "ioccsz %d is not supported!\n",
+ ctrl->ctrl.ioccsz);
+ goto out_remove_admin_queue;
+ }
+ /* FC-NVME supports 16-byte CQE only */
+ if (ctrl->ctrl.iorcsz != 1) {
+ dev_err(ctrl->ctrl.device, "iorcsz %d is not supported!\n",
+ ctrl->ctrl.iorcsz);
+ goto out_remove_admin_queue;
+ }
+ /* 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;
+ }
+
+ /* 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_exit_aen_ops;
+
+ 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\" (%p)\n",
+ ctrl->cnum, ctrl->ctrl.opts->subsysnqn, &ctrl);
+
+ kref_get(&ctrl->ctrl.kref);
+
+ 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);
+ }
+
+ 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_free_ida:
+ 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);
+}
+
+enum {
+ FCT_TRADDR_ERR = 0,
+ FCT_TRADDR_WWNN = 1 << 0,
+ FCT_TRADDR_WWPN = 1 << 1,
+};
+
+struct nvmet_fc_traddr {
+ u64 nn;
+ u64 pn;
+};
+
+static const match_table_t traddr_opt_tokens = {
+ { FCT_TRADDR_WWNN, "nn-%s" },
+ { FCT_TRADDR_WWPN, "pn-%s" },
+ { FCT_TRADDR_ERR, NULL }
+};
+
+static int
+nvme_fc_parse_address(struct nvmet_fc_traddr *traddr, char *buf)
+{
+ substring_t args[MAX_OPT_ARGS];
+ char *options, *o, *p;
+ int token, ret = 0;
+ u64 token64;
+
+ options = o = kstrdup(buf, GFP_KERNEL);
+ if (!options)
+ return -ENOMEM;
+
+ while ((p = strsep(&o, ":\n")) != NULL) {
+ if (!*p)
+ continue;
+
+ token = match_token(p, traddr_opt_tokens, args);
+ switch (token) {
+ case FCT_TRADDR_WWNN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out;
+ }
+ traddr->nn = token64;
+ break;
+ case FCT_TRADDR_WWPN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out;
+ }
+ traddr->pn = token64;
+ break;
+ default:
+ pr_warn("unknown traddr token or missing value '%s'\n",
+ p);
+ ret = -EINVAL;
+ goto out;
+ }
+ }
+
+out:
+ kfree(options);
+ return ret;
+}
+
+static struct nvme_ctrl *
+nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts)
+{
+ struct nvme_fc_lport *lport;
+ struct nvme_fc_rport *rport;
+ struct nvmet_fc_traddr laddr = { 0L, 0L };
+ struct nvmet_fc_traddr raddr = { 0L, 0L };
+ unsigned long flags;
+ int ret;
+
+ ret = nvme_fc_parse_address(&raddr, opts->traddr);
+ if (ret || !raddr.nn || !raddr.pn)
+ return ERR_PTR(-EINVAL);
+
+ ret = nvme_fc_parse_address(&laddr, opts->host_traddr);
+ if (ret || !laddr.nn || !laddr.pn)
+ return ERR_PTR(-EINVAL);
+
+ /* find the host and remote ports to connect together */
+ spin_lock_irqsave(&nvme_fc_lock, flags);
+ list_for_each_entry(lport, &nvme_fc_lport_list, port_list) {
+ if (lport->localport.node_name != laddr.nn ||
+ lport->localport.port_name != laddr.pn)
+ continue;
+
+ list_for_each_entry(rport, &lport->endp_list, endp_list) {
+ if (rport->remoteport.node_name != raddr.nn ||
+ rport->remoteport.port_name != raddr.pn)
+ continue;
+
+ /* if fail to get reference fall through. Will error */
+ if (!nvme_fc_rport_get(rport))
+ break;
+
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ return __nvme_fc_create_ctrl(dev, opts, lport, rport);
+ }
+ }
+ spin_unlock_irqrestore(&nvme_fc_lock, flags);
+
+ return ERR_PTR(-ENOENT);
+}
+
+
+static struct nvmf_transport_ops nvme_fc_transport = {
+ .name = "fc",
+ .required_opts = NVMF_OPT_TRADDR | NVMF_OPT_HOST_TRADDR,
+ .allowed_opts = NVMF_OPT_RECONNECT_DELAY,
+ .create_ctrl = nvme_fc_create_ctrl,
+};
+
+static int __init nvme_fc_init_module(void)
+{
+ nvme_fc_wq = create_workqueue("nvme_fc_wq");
+ if (!nvme_fc_wq)
+ return -ENOMEM;
+
+ nvmf_register_transport(&nvme_fc_transport);
+ return 0;
+}
+
+static void __exit nvme_fc_exit_module(void)
+{
+ /* sanity check - all lports should be removed */
+ if (!list_empty(&nvme_fc_lport_list))
+ pr_warn("%s: localport list not empty\n", __func__);
+
+ nvmf_unregister_transport(&nvme_fc_transport);
+
+ destroy_workqueue(nvme_fc_wq);
+
+ ida_destroy(&nvme_fc_local_port_cnt);
+ ida_destroy(&nvme_fc_ctrl_cnt);
+}
+
+module_init(nvme_fc_init_module);
+module_exit(nvme_fc_exit_module);
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index f5e3011e31fc..588d4a34c083 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -146,14 +146,6 @@ struct nvme_nvm_command {
};
};
-struct nvme_nvm_completion {
- __le64 result; /* Used by LightNVM to return ppa completions */
- __le16 sq_head; /* how much of this queue may be reclaimed */
- __le16 sq_id; /* submission queue that generated this entry */
- __u16 command_id; /* of the command which completed */
- __le16 status; /* did the command fail, and if so, why? */
-};
-
#define NVME_NVM_LP_MLC_PAIRS 886
struct nvme_nvm_lp_mlc {
__le16 num_pairs;
@@ -360,6 +352,7 @@ static int nvme_nvm_get_l2p_tbl(struct nvm_dev *nvmdev, u64 slba, u32 nlb,
while (nlb) {
u32 cmd_nlb = min(nlb_pr_rq, nlb);
+ u64 elba = slba + cmd_nlb;
c.l2p.slba = cpu_to_le64(cmd_slba);
c.l2p.nlb = cpu_to_le32(cmd_nlb);
@@ -373,6 +366,14 @@ static int nvme_nvm_get_l2p_tbl(struct nvm_dev *nvmdev, u64 slba, u32 nlb,
goto out;
}
+ if (unlikely(elba > nvmdev->total_secs)) {
+ pr_err("nvm: L2P data from device is out of bounds!\n");
+ return -EINVAL;
+ }
+
+ /* Transform physical address to target address space */
+ nvmdev->mt->part_to_tgt(nvmdev, entries, cmd_nlb);
+
if (update_l2p(cmd_slba, cmd_nlb, entries, priv)) {
ret = -EINTR;
goto out;
@@ -391,11 +392,12 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
u8 *blks)
{
struct request_queue *q = nvmdev->q;
+ struct nvm_geo *geo = &nvmdev->geo;
struct nvme_ns *ns = q->queuedata;
struct nvme_ctrl *ctrl = ns->ctrl;
struct nvme_nvm_command c = {};
struct nvme_nvm_bb_tbl *bb_tbl;
- int nr_blks = nvmdev->blks_per_lun * nvmdev->plane_mode;
+ int nr_blks = geo->blks_per_lun * geo->plane_mode;
int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_blks;
int ret = 0;
@@ -436,7 +438,7 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
goto out;
}
- memcpy(blks, bb_tbl->blk, nvmdev->blks_per_lun * nvmdev->plane_mode);
+ memcpy(blks, bb_tbl->blk, geo->blks_per_lun * geo->plane_mode);
out:
kfree(bb_tbl);
return ret;
@@ -481,14 +483,11 @@ static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
static void nvme_nvm_end_io(struct request *rq, int error)
{
struct nvm_rq *rqd = rq->end_io_data;
- struct nvme_nvm_completion *cqe = rq->special;
-
- if (cqe)
- rqd->ppa_status = le64_to_cpu(cqe->result);
+ rqd->ppa_status = nvme_req(rq)->result.u64;
nvm_end_io(rqd, error);
- kfree(rq->cmd);
+ kfree(nvme_req(rq)->cmd);
blk_mq_free_request(rq);
}
@@ -500,20 +499,18 @@ static int nvme_nvm_submit_io(struct nvm_dev *dev, struct nvm_rq *rqd)
struct bio *bio = rqd->bio;
struct nvme_nvm_command *cmd;
- rq = blk_mq_alloc_request(q, bio_data_dir(bio), 0);
- if (IS_ERR(rq))
+ cmd = kzalloc(sizeof(struct nvme_nvm_command), GFP_KERNEL);
+ if (!cmd)
return -ENOMEM;
- cmd = kzalloc(sizeof(struct nvme_nvm_command) +
- sizeof(struct nvme_nvm_completion), GFP_KERNEL);
- if (!cmd) {
- blk_mq_free_request(rq);
+ rq = nvme_alloc_request(q, (struct nvme_command *)cmd, 0, NVME_QID_ANY);
+ if (IS_ERR(rq)) {
+ kfree(cmd);
return -ENOMEM;
}
+ rq->cmd_flags &= ~REQ_FAILFAST_DRIVER;
- rq->cmd_type = REQ_TYPE_DRV_PRIV;
rq->ioprio = bio_prio(bio);
-
if (bio_has_data(bio))
rq->nr_phys_segments = bio_phys_segments(q, bio);
@@ -522,10 +519,6 @@ static int nvme_nvm_submit_io(struct nvm_dev *dev, struct nvm_rq *rqd)
nvme_nvm_rqtocmd(rq, rqd, ns, cmd);
- rq->cmd = (unsigned char *)cmd;
- rq->cmd_len = sizeof(struct nvme_nvm_command);
- rq->special = cmd + 1;
-
rq->end_io_data = rqd;
blk_execute_rq_nowait(q, NULL, rq, 0, nvme_nvm_end_io);
@@ -543,6 +536,7 @@ static int nvme_nvm_erase_block(struct nvm_dev *dev, struct nvm_rq *rqd)
c.erase.nsid = cpu_to_le32(ns->ns_id);
c.erase.spba = cpu_to_le64(rqd->ppa_addr.ppa);
c.erase.length = cpu_to_le16(rqd->nr_ppas - 1);
+ c.erase.control = cpu_to_le16(rqd->flags);
return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
}
@@ -592,12 +586,10 @@ static struct nvm_dev_ops nvme_nvm_dev_ops = {
.max_phys_sect = 64,
};
-int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node,
- const struct attribute_group *attrs)
+int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node)
{
struct request_queue *q = ns->queue;
struct nvm_dev *dev;
- int ret;
dev = nvm_alloc_dev(node);
if (!dev)
@@ -606,18 +598,10 @@ int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node,
dev->q = q;
memcpy(dev->name, disk_name, DISK_NAME_LEN);
dev->ops = &nvme_nvm_dev_ops;
- dev->parent_dev = ns->ctrl->device;
dev->private_data = ns;
ns->ndev = dev;
- ret = nvm_register(dev);
-
- ns->lba_shift = ilog2(dev->sec_size) - 9;
-
- if (sysfs_create_group(&dev->dev.kobj, attrs))
- pr_warn("%s: failed to create sysfs group for identification\n",
- disk_name);
- return ret;
+ return nvm_register(dev);
}
void nvme_nvm_unregister(struct nvme_ns *ns)
@@ -625,6 +609,167 @@ void nvme_nvm_unregister(struct nvme_ns *ns)
nvm_unregister(ns->ndev);
}
+static ssize_t nvm_dev_attr_show(struct device *dev,
+ struct device_attribute *dattr, char *page)
+{
+ struct nvme_ns *ns = nvme_get_ns_from_dev(dev);
+ struct nvm_dev *ndev = ns->ndev;
+ struct nvm_id *id;
+ struct nvm_id_group *grp;
+ struct attribute *attr;
+
+ if (!ndev)
+ return 0;
+
+ id = &ndev->identity;
+ grp = &id->groups[0];
+ attr = &dattr->attr;
+
+ if (strcmp(attr->name, "version") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", id->ver_id);
+ } else if (strcmp(attr->name, "vendor_opcode") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", id->vmnt);
+ } else if (strcmp(attr->name, "capabilities") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", id->cap);
+ } else if (strcmp(attr->name, "device_mode") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", id->dom);
+ } else if (strcmp(attr->name, "media_manager") == 0) {
+ if (!ndev->mt)
+ return scnprintf(page, PAGE_SIZE, "%s\n", "none");
+ return scnprintf(page, PAGE_SIZE, "%s\n", ndev->mt->name);
+ } else if (strcmp(attr->name, "ppa_format") == 0) {
+ return scnprintf(page, PAGE_SIZE,
+ "0x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x\n",
+ id->ppaf.ch_offset, id->ppaf.ch_len,
+ id->ppaf.lun_offset, id->ppaf.lun_len,
+ id->ppaf.pln_offset, id->ppaf.pln_len,
+ id->ppaf.blk_offset, id->ppaf.blk_len,
+ id->ppaf.pg_offset, id->ppaf.pg_len,
+ id->ppaf.sect_offset, id->ppaf.sect_len);
+ } else if (strcmp(attr->name, "media_type") == 0) { /* u8 */
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->mtype);
+ } else if (strcmp(attr->name, "flash_media_type") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->fmtype);
+ } else if (strcmp(attr->name, "num_channels") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_ch);
+ } else if (strcmp(attr->name, "num_luns") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_lun);
+ } else if (strcmp(attr->name, "num_planes") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_pln);
+ } else if (strcmp(attr->name, "num_blocks") == 0) { /* u16 */
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_blk);
+ } else if (strcmp(attr->name, "num_pages") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_pg);
+ } else if (strcmp(attr->name, "page_size") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->fpg_sz);
+ } else if (strcmp(attr->name, "hw_sector_size") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->csecs);
+ } else if (strcmp(attr->name, "oob_sector_size") == 0) {/* u32 */
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->sos);
+ } else if (strcmp(attr->name, "read_typ") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->trdt);
+ } else if (strcmp(attr->name, "read_max") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->trdm);
+ } else if (strcmp(attr->name, "prog_typ") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->tprt);
+ } else if (strcmp(attr->name, "prog_max") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->tprm);
+ } else if (strcmp(attr->name, "erase_typ") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->tbet);
+ } else if (strcmp(attr->name, "erase_max") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", grp->tbem);
+ } else if (strcmp(attr->name, "multiplane_modes") == 0) {
+ return scnprintf(page, PAGE_SIZE, "0x%08x\n", grp->mpos);
+ } else if (strcmp(attr->name, "media_capabilities") == 0) {
+ return scnprintf(page, PAGE_SIZE, "0x%08x\n", grp->mccap);
+ } else if (strcmp(attr->name, "max_phys_secs") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n",
+ ndev->ops->max_phys_sect);
+ } else {
+ return scnprintf(page,
+ PAGE_SIZE,
+ "Unhandled attr(%s) in `nvm_dev_attr_show`\n",
+ attr->name);
+ }
+}
+
+#define NVM_DEV_ATTR_RO(_name) \
+ DEVICE_ATTR(_name, S_IRUGO, nvm_dev_attr_show, NULL)
+
+static NVM_DEV_ATTR_RO(version);
+static NVM_DEV_ATTR_RO(vendor_opcode);
+static NVM_DEV_ATTR_RO(capabilities);
+static NVM_DEV_ATTR_RO(device_mode);
+static NVM_DEV_ATTR_RO(ppa_format);
+static NVM_DEV_ATTR_RO(media_manager);
+
+static NVM_DEV_ATTR_RO(media_type);
+static NVM_DEV_ATTR_RO(flash_media_type);
+static NVM_DEV_ATTR_RO(num_channels);
+static NVM_DEV_ATTR_RO(num_luns);
+static NVM_DEV_ATTR_RO(num_planes);
+static NVM_DEV_ATTR_RO(num_blocks);
+static NVM_DEV_ATTR_RO(num_pages);
+static NVM_DEV_ATTR_RO(page_size);
+static NVM_DEV_ATTR_RO(hw_sector_size);
+static NVM_DEV_ATTR_RO(oob_sector_size);
+static NVM_DEV_ATTR_RO(read_typ);
+static NVM_DEV_ATTR_RO(read_max);
+static NVM_DEV_ATTR_RO(prog_typ);
+static NVM_DEV_ATTR_RO(prog_max);
+static NVM_DEV_ATTR_RO(erase_typ);
+static NVM_DEV_ATTR_RO(erase_max);
+static NVM_DEV_ATTR_RO(multiplane_modes);
+static NVM_DEV_ATTR_RO(media_capabilities);
+static NVM_DEV_ATTR_RO(max_phys_secs);
+
+static struct attribute *nvm_dev_attrs[] = {
+ &dev_attr_version.attr,
+ &dev_attr_vendor_opcode.attr,
+ &dev_attr_capabilities.attr,
+ &dev_attr_device_mode.attr,
+ &dev_attr_media_manager.attr,
+
+ &dev_attr_ppa_format.attr,
+ &dev_attr_media_type.attr,
+ &dev_attr_flash_media_type.attr,
+ &dev_attr_num_channels.attr,
+ &dev_attr_num_luns.attr,
+ &dev_attr_num_planes.attr,
+ &dev_attr_num_blocks.attr,
+ &dev_attr_num_pages.attr,
+ &dev_attr_page_size.attr,
+ &dev_attr_hw_sector_size.attr,
+ &dev_attr_oob_sector_size.attr,
+ &dev_attr_read_typ.attr,
+ &dev_attr_read_max.attr,
+ &dev_attr_prog_typ.attr,
+ &dev_attr_prog_max.attr,
+ &dev_attr_erase_typ.attr,
+ &dev_attr_erase_max.attr,
+ &dev_attr_multiplane_modes.attr,
+ &dev_attr_media_capabilities.attr,
+ &dev_attr_max_phys_secs.attr,
+ NULL,
+};
+
+static const struct attribute_group nvm_dev_attr_group = {
+ .name = "lightnvm",
+ .attrs = nvm_dev_attrs,
+};
+
+int nvme_nvm_register_sysfs(struct nvme_ns *ns)
+{
+ return sysfs_create_group(&disk_to_dev(ns->disk)->kobj,
+ &nvm_dev_attr_group);
+}
+
+void nvme_nvm_unregister_sysfs(struct nvme_ns *ns)
+{
+ sysfs_remove_group(&disk_to_dev(ns->disk)->kobj,
+ &nvm_dev_attr_group);
+}
+
/* move to shared place when used in multiple places. */
#define PCI_VENDOR_ID_CNEX 0x1d1d
#define PCI_DEVICE_ID_CNEX_WL 0x2807
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index d47f5a5d18c7..bd5321441d12 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -79,6 +79,20 @@ enum nvme_quirks {
NVME_QUIRK_DELAY_BEFORE_CHK_RDY = (1 << 3),
};
+/*
+ * Common request structure for NVMe passthrough. All drivers must have
+ * this structure as the first member of their request-private data.
+ */
+struct nvme_request {
+ struct nvme_command *cmd;
+ union nvme_result result;
+};
+
+static inline struct nvme_request *nvme_req(struct request *req)
+{
+ return blk_mq_rq_to_pdu(req);
+}
+
/* The below value is the specific amount of delay needed before checking
* readiness in case of the PCI_DEVICE(0x1c58, 0x0003), which needs the
* NVME_QUIRK_DELAY_BEFORE_CHK_RDY quirk enabled. The value (in ms) was
@@ -222,8 +236,10 @@ static inline unsigned nvme_map_len(struct request *rq)
static inline void nvme_cleanup_cmd(struct request *req)
{
- if (req_op(req) == REQ_OP_DISCARD)
- kfree(req->completion_data);
+ if (req->rq_flags & RQF_SPECIAL_PAYLOAD) {
+ kfree(page_address(req->special_vec.bv_page) +
+ req->special_vec.bv_offset);
+ }
}
static inline int nvme_error_status(u16 status)
@@ -261,8 +277,8 @@ void nvme_queue_scan(struct nvme_ctrl *ctrl);
void nvme_remove_namespaces(struct nvme_ctrl *ctrl);
#define NVME_NR_AERS 1
-void nvme_complete_async_event(struct nvme_ctrl *ctrl,
- struct nvme_completion *cqe);
+void nvme_complete_async_event(struct nvme_ctrl *ctrl, __le16 status,
+ union nvme_result *res);
void nvme_queue_async_events(struct nvme_ctrl *ctrl);
void nvme_stop_queues(struct nvme_ctrl *ctrl);
@@ -278,7 +294,7 @@ int nvme_setup_cmd(struct nvme_ns *ns, struct request *req,
int nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
void *buf, unsigned bufflen);
int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
- struct nvme_completion *cqe, void *buffer, unsigned bufflen,
+ union nvme_result *result, void *buffer, unsigned bufflen,
unsigned timeout, int qid, int at_head, int flags);
int nvme_submit_user_cmd(struct request_queue *q, struct nvme_command *cmd,
void __user *ubuffer, unsigned bufflen, u32 *result,
@@ -307,36 +323,33 @@ int nvme_sg_get_version_num(int __user *ip);
#ifdef CONFIG_NVM
int nvme_nvm_ns_supported(struct nvme_ns *ns, struct nvme_id_ns *id);
-int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node,
- const struct attribute_group *attrs);
+int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node);
void nvme_nvm_unregister(struct nvme_ns *ns);
-
-static inline struct nvme_ns *nvme_get_ns_from_dev(struct device *dev)
-{
- if (dev->type->devnode)
- return dev_to_disk(dev)->private_data;
-
- return (container_of(dev, struct nvm_dev, dev))->private_data;
-}
+int nvme_nvm_register_sysfs(struct nvme_ns *ns);
+void nvme_nvm_unregister_sysfs(struct nvme_ns *ns);
#else
static inline int nvme_nvm_register(struct nvme_ns *ns, char *disk_name,
- int node,
- const struct attribute_group *attrs)
+ int node)
{
return 0;
}
static inline void nvme_nvm_unregister(struct nvme_ns *ns) {};
-
+static inline int nvme_nvm_register_sysfs(struct nvme_ns *ns)
+{
+ return 0;
+}
+static inline void nvme_nvm_unregister_sysfs(struct nvme_ns *ns) {};
static inline int nvme_nvm_ns_supported(struct nvme_ns *ns, struct nvme_id_ns *id)
{
return 0;
}
+#endif /* CONFIG_NVM */
+
static inline struct nvme_ns *nvme_get_ns_from_dev(struct device *dev)
{
return dev_to_disk(dev)->private_data;
}
-#endif /* CONFIG_NVM */
int __init nvme_core_init(void);
void nvme_core_exit(void);
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index 0248d0e21fee..2fd7dc2e8fc4 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -141,6 +141,7 @@ struct nvme_queue {
* allocated to store the PRP list.
*/
struct nvme_iod {
+ struct nvme_request req;
struct nvme_queue *nvmeq;
int aborted;
int npages; /* In the PRP list. 0 means small pool in use */
@@ -302,14 +303,14 @@ static void __nvme_submit_cmd(struct nvme_queue *nvmeq,
static __le64 **iod_list(struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
- return (__le64 **)(iod->sg + req->nr_phys_segments);
+ return (__le64 **)(iod->sg + blk_rq_nr_phys_segments(req));
}
static int nvme_init_iod(struct request *rq, unsigned size,
struct nvme_dev *dev)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(rq);
- int nseg = rq->nr_phys_segments;
+ int nseg = blk_rq_nr_phys_segments(rq);
if (nseg > NVME_INT_PAGES || size > NVME_INT_BYTES(dev)) {
iod->sg = kmalloc(nvme_iod_alloc_size(dev, size, nseg), GFP_ATOMIC);
@@ -324,11 +325,11 @@ static int nvme_init_iod(struct request *rq, unsigned size,
iod->nents = 0;
iod->length = size;
- if (!(rq->cmd_flags & REQ_DONTPREP)) {
+ if (!(rq->rq_flags & RQF_DONTPREP)) {
rq->retries = 0;
- rq->cmd_flags |= REQ_DONTPREP;
+ rq->rq_flags |= RQF_DONTPREP;
}
- return 0;
+ return BLK_MQ_RQ_QUEUE_OK;
}
static void nvme_free_iod(struct nvme_dev *dev, struct request *req)
@@ -339,8 +340,6 @@ static void nvme_free_iod(struct nvme_dev *dev, struct request *req)
__le64 **list = iod_list(req);
dma_addr_t prp_dma = iod->first_dma;
- nvme_cleanup_cmd(req);
-
if (iod->npages == 0)
dma_pool_free(dev->prp_small_pool, list[0], prp_dma);
for (i = 0; i < iod->npages; i++) {
@@ -510,7 +509,7 @@ static int nvme_map_data(struct nvme_dev *dev, struct request *req,
DMA_TO_DEVICE : DMA_FROM_DEVICE;
int ret = BLK_MQ_RQ_QUEUE_ERROR;
- sg_init_table(iod->sg, req->nr_phys_segments);
+ sg_init_table(iod->sg, blk_rq_nr_phys_segments(req));
iod->nents = blk_rq_map_sg(q, req, iod->sg);
if (!iod->nents)
goto out;
@@ -566,6 +565,7 @@ static void nvme_unmap_data(struct nvme_dev *dev, struct request *req)
}
}
+ nvme_cleanup_cmd(req);
nvme_free_iod(dev, req);
}
@@ -596,22 +596,21 @@ static int nvme_queue_rq(struct blk_mq_hw_ctx *hctx,
}
}
- map_len = nvme_map_len(req);
- ret = nvme_init_iod(req, map_len, dev);
- if (ret)
+ ret = nvme_setup_cmd(ns, req, &cmnd);
+ if (ret != BLK_MQ_RQ_QUEUE_OK)
return ret;
- ret = nvme_setup_cmd(ns, req, &cmnd);
- if (ret)
- goto out;
+ map_len = nvme_map_len(req);
+ ret = nvme_init_iod(req, map_len, dev);
+ if (ret != BLK_MQ_RQ_QUEUE_OK)
+ goto out_free_cmd;
- if (req->nr_phys_segments)
+ if (blk_rq_nr_phys_segments(req))
ret = nvme_map_data(dev, req, map_len, &cmnd);
- if (ret)
- goto out;
+ if (ret != BLK_MQ_RQ_QUEUE_OK)
+ goto out_cleanup_iod;
- cmnd.common.command_id = req->tag;
blk_mq_start_request(req);
spin_lock_irq(&nvmeq->q_lock);
@@ -621,14 +620,16 @@ static int nvme_queue_rq(struct blk_mq_hw_ctx *hctx,
else
ret = BLK_MQ_RQ_QUEUE_ERROR;
spin_unlock_irq(&nvmeq->q_lock);
- goto out;
+ goto out_cleanup_iod;
}
__nvme_submit_cmd(nvmeq, &cmnd);
nvme_process_cq(nvmeq);
spin_unlock_irq(&nvmeq->q_lock);
return BLK_MQ_RQ_QUEUE_OK;
-out:
+out_cleanup_iod:
nvme_free_iod(dev, req);
+out_free_cmd:
+ nvme_cleanup_cmd(req);
return ret;
}
@@ -703,13 +704,13 @@ static void __nvme_process_cq(struct nvme_queue *nvmeq, unsigned int *tag)
*/
if (unlikely(nvmeq->qid == 0 &&
cqe.command_id >= NVME_AQ_BLKMQ_DEPTH)) {
- nvme_complete_async_event(&nvmeq->dev->ctrl, &cqe);
+ nvme_complete_async_event(&nvmeq->dev->ctrl,
+ cqe.status, &cqe.result);
continue;
}
req = blk_mq_tag_to_rq(*nvmeq->tags, cqe.command_id);
- if (req->cmd_type == REQ_TYPE_DRV_PRIV && req->special)
- memcpy(req->special, &cqe, sizeof(cqe));
+ nvme_req(req)->result = cqe.result;
blk_mq_complete_request(req, le16_to_cpu(cqe.status) >> 1);
}
@@ -1242,20 +1243,16 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev)
result = nvme_enable_ctrl(&dev->ctrl, cap);
if (result)
- goto free_nvmeq;
+ return result;
nvmeq->cq_vector = 0;
result = queue_request_irq(nvmeq);
if (result) {
nvmeq->cq_vector = -1;
- goto free_nvmeq;
+ return result;
}
return result;
-
- free_nvmeq:
- nvme_free_queues(dev, 0);
- return result;
}
static bool nvme_should_reset(struct nvme_dev *dev, u32 csts)
@@ -1285,6 +1282,24 @@ static bool nvme_should_reset(struct nvme_dev *dev, u32 csts)
return true;
}
+static void nvme_warn_reset(struct nvme_dev *dev, u32 csts)
+{
+ /* Read a config register to help see what died. */
+ u16 pci_status;
+ int result;
+
+ result = pci_read_config_word(to_pci_dev(dev->dev), PCI_STATUS,
+ &pci_status);
+ if (result == PCIBIOS_SUCCESSFUL)
+ dev_warn(dev->dev,
+ "controller is down; will reset: CSTS=0x%x, PCI_STATUS=0x%hx\n",
+ csts, pci_status);
+ else
+ dev_warn(dev->dev,
+ "controller is down; will reset: CSTS=0x%x, PCI_STATUS read failed (%d)\n",
+ csts, result);
+}
+
static void nvme_watchdog_timer(unsigned long data)
{
struct nvme_dev *dev = (struct nvme_dev *)data;
@@ -1293,9 +1308,7 @@ static void nvme_watchdog_timer(unsigned long data)
/* Skip controllers under certain specific conditions. */
if (nvme_should_reset(dev, csts)) {
if (!nvme_reset(dev))
- dev_warn(dev->dev,
- "Failed status: 0x%x, reset controller.\n",
- csts);
+ nvme_warn_reset(dev, csts);
return;
}
@@ -1317,10 +1330,8 @@ static int nvme_create_io_queues(struct nvme_dev *dev)
max = min(dev->max_qid, dev->queue_count - 1);
for (i = dev->online_queues; i <= max; i++) {
ret = nvme_create_queue(dev->queues[i], i);
- if (ret) {
- nvme_free_queues(dev, i);
+ if (ret)
break;
- }
}
/*
@@ -1460,13 +1471,9 @@ static int nvme_setup_io_queues(struct nvme_dev *dev)
result = queue_request_irq(adminq);
if (result) {
adminq->cq_vector = -1;
- goto free_queues;
+ return result;
}
return nvme_create_io_queues(dev);
-
- free_queues:
- nvme_free_queues(dev, 1);
- return result;
}
static void nvme_del_queue_end(struct request *req, int error)
@@ -2095,9 +2102,6 @@ static const struct pci_error_handlers nvme_err_handler = {
.reset_notify = nvme_reset_notify,
};
-/* Move to pci_ids.h later */
-#define PCI_CLASS_STORAGE_EXPRESS 0x010802
-
static const struct pci_device_id nvme_id_table[] = {
{ PCI_VDEVICE(INTEL, 0x0953),
.driver_data = NVME_QUIRK_STRIPE_SIZE |
diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index accbe8edb6c4..f587af345889 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -28,7 +28,6 @@
#include <rdma/ib_verbs.h>
#include <rdma/rdma_cm.h>
-#include <rdma/ib_cm.h>
#include <linux/nvme-rdma.h>
#include "nvme.h"
@@ -88,6 +87,7 @@ struct nvme_rdma_qe {
struct nvme_rdma_queue;
struct nvme_rdma_request {
+ struct nvme_request req;
struct ib_mr *mr;
struct nvme_rdma_qe sqe;
struct ib_sge sge[1 + NVME_RDMA_MAX_INLINE_SEGMENTS];
@@ -105,6 +105,7 @@ enum nvme_rdma_queue_flags {
NVME_RDMA_Q_CONNECTED = (1 << 0),
NVME_RDMA_IB_QUEUE_ALLOCATED = (1 << 1),
NVME_RDMA_Q_DELETING = (1 << 2),
+ NVME_RDMA_Q_LIVE = (1 << 3),
};
struct nvme_rdma_queue {
@@ -262,7 +263,9 @@ out_free_ring:
static void nvme_rdma_qp_event(struct ib_event *event, void *context)
{
- pr_debug("QP event %d\n", event->event);
+ pr_debug("QP event %s (%d)\n",
+ ib_event_msg(event->event), event->event);
+
}
static int nvme_rdma_wait_for_cm(struct nvme_rdma_queue *queue)
@@ -646,10 +649,18 @@ static int nvme_rdma_connect_io_queues(struct nvme_rdma_ctrl *ctrl)
for (i = 1; i < ctrl->queue_count; i++) {
ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
- if (ret)
- break;
+ if (ret) {
+ dev_info(ctrl->ctrl.device,
+ "failed to connect i/o queue: %d\n", ret);
+ goto out_free_queues;
+ }
+ set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[i].flags);
}
+ return 0;
+
+out_free_queues:
+ nvme_rdma_free_io_queues(ctrl);
return ret;
}
@@ -734,6 +745,8 @@ static void nvme_rdma_reconnect_ctrl_work(struct work_struct *work)
if (ret)
goto stop_admin_q;
+ set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
+
ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
if (ret)
goto stop_admin_q;
@@ -783,8 +796,10 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
nvme_stop_keep_alive(&ctrl->ctrl);
- for (i = 0; i < ctrl->queue_count; i++)
+ for (i = 0; i < ctrl->queue_count; i++) {
clear_bit(NVME_RDMA_Q_CONNECTED, &ctrl->queues[i].flags);
+ clear_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[i].flags);
+ }
if (ctrl->queue_count > 1)
nvme_stop_queues(&ctrl->ctrl);
@@ -972,8 +987,7 @@ static int nvme_rdma_map_data(struct nvme_rdma_queue *queue,
struct nvme_rdma_request *req = blk_mq_rq_to_pdu(rq);
struct nvme_rdma_device *dev = queue->device;
struct ib_device *ibdev = dev->dev;
- int nents, count;
- int ret;
+ int count, ret;
req->num_sge = 1;
req->inline_data = false;
@@ -985,16 +999,14 @@ static int nvme_rdma_map_data(struct nvme_rdma_queue *queue,
return nvme_rdma_set_sg_null(c);
req->sg_table.sgl = req->first_sgl;
- ret = sg_alloc_table_chained(&req->sg_table, rq->nr_phys_segments,
- req->sg_table.sgl);
+ ret = sg_alloc_table_chained(&req->sg_table,
+ blk_rq_nr_phys_segments(rq), req->sg_table.sgl);
if (ret)
return -ENOMEM;
- nents = blk_rq_map_sg(rq->q, rq, req->sg_table.sgl);
- BUG_ON(nents > rq->nr_phys_segments);
- req->nents = nents;
+ req->nents = blk_rq_map_sg(rq->q, rq, req->sg_table.sgl);
- count = ib_dma_map_sg(ibdev, req->sg_table.sgl, nents,
+ count = ib_dma_map_sg(ibdev, req->sg_table.sgl, req->nents,
rq_data_dir(rq) == WRITE ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
if (unlikely(count <= 0)) {
sg_free_table_chained(&req->sg_table, true);
@@ -1139,13 +1151,10 @@ static void nvme_rdma_submit_async_event(struct nvme_ctrl *arg, int aer_idx)
static int nvme_rdma_process_nvme_rsp(struct nvme_rdma_queue *queue,
struct nvme_completion *cqe, struct ib_wc *wc, int tag)
{
- u16 status = le16_to_cpu(cqe->status);
struct request *rq;
struct nvme_rdma_request *req;
int ret = 0;
- status >>= 1;
-
rq = blk_mq_tag_to_rq(nvme_rdma_tagset(queue), cqe->command_id);
if (!rq) {
dev_err(queue->ctrl->ctrl.device,
@@ -1156,9 +1165,6 @@ static int nvme_rdma_process_nvme_rsp(struct nvme_rdma_queue *queue,
}
req = blk_mq_rq_to_pdu(rq);
- if (rq->cmd_type == REQ_TYPE_DRV_PRIV && rq->special)
- memcpy(rq->special, cqe, sizeof(*cqe));
-
if (rq->tag == tag)
ret = 1;
@@ -1166,8 +1172,8 @@ static int nvme_rdma_process_nvme_rsp(struct nvme_rdma_queue *queue,
wc->ex.invalidate_rkey == req->mr->rkey)
req->mr->need_inval = false;
- blk_mq_complete_request(rq, status);
-
+ req->req.result = cqe->result;
+ blk_mq_complete_request(rq, le16_to_cpu(cqe->status) >> 1);
return ret;
}
@@ -1195,7 +1201,8 @@ static int __nvme_rdma_recv_done(struct ib_cq *cq, struct ib_wc *wc, int tag)
*/
if (unlikely(nvme_rdma_queue_idx(queue) == 0 &&
cqe->command_id >= NVME_RDMA_AQ_BLKMQ_DEPTH))
- nvme_complete_async_event(&queue->ctrl->ctrl, cqe);
+ nvme_complete_async_event(&queue->ctrl->ctrl, cqe->status,
+ &cqe->result);
else
ret = nvme_rdma_process_nvme_rsp(queue, cqe, wc, tag);
ib_dma_sync_single_for_device(ibdev, qe->dma, len, DMA_FROM_DEVICE);
@@ -1408,6 +1415,24 @@ nvme_rdma_timeout(struct request *rq, bool reserved)
return BLK_EH_HANDLED;
}
+/*
+ * We cannot accept any other command until the Connect command has completed.
+ */
+static inline bool nvme_rdma_queue_is_ready(struct nvme_rdma_queue *queue,
+ struct request *rq)
+{
+ if (unlikely(!test_bit(NVME_RDMA_Q_LIVE, &queue->flags))) {
+ struct nvme_command *cmd = (struct nvme_command *)rq->cmd;
+
+ if (rq->cmd_type != REQ_TYPE_DRV_PRIV ||
+ cmd->common.opcode != nvme_fabrics_command ||
+ cmd->fabrics.fctype != nvme_fabrics_type_connect)
+ return false;
+ }
+
+ return true;
+}
+
static int nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
const struct blk_mq_queue_data *bd)
{
@@ -1424,15 +1449,17 @@ static int nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
WARN_ON_ONCE(rq->tag < 0);
+ if (!nvme_rdma_queue_is_ready(queue, rq))
+ return BLK_MQ_RQ_QUEUE_BUSY;
+
dev = queue->device->dev;
ib_dma_sync_single_for_cpu(dev, sqe->dma,
sizeof(struct nvme_command), DMA_TO_DEVICE);
ret = nvme_setup_cmd(ns, rq, c);
- if (ret)
+ if (ret != BLK_MQ_RQ_QUEUE_OK)
return ret;
- c->common.command_id = rq->tag;
blk_mq_start_request(rq);
map_len = nvme_map_len(rq);
@@ -1574,6 +1601,8 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl)
if (error)
goto out_cleanup_queue;
+ set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
+
error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap);
if (error) {
dev_err(ctrl->ctrl.device,
@@ -1938,6 +1967,14 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
opts->queue_size = ctrl->ctrl.maxcmd;
}
+ if (opts->queue_size > ctrl->ctrl.sqsize + 1) {
+ /* warn if sqsize is lower than queue_size */
+ dev_warn(ctrl->ctrl.device,
+ "queue_size %zu > ctrl sqsize %u, clamping down\n",
+ opts->queue_size, ctrl->ctrl.sqsize + 1);
+ opts->queue_size = ctrl->ctrl.sqsize + 1;
+ }
+
if (opts->nr_io_queues) {
ret = nvme_rdma_create_io_queues(ctrl);
if (ret)
diff --git a/drivers/nvme/host/scsi.c b/drivers/nvme/host/scsi.c
index 3eaa4d27801e..b71e95044b43 100644
--- a/drivers/nvme/host/scsi.c
+++ b/drivers/nvme/host/scsi.c
@@ -1280,10 +1280,6 @@ static inline void nvme_trans_modesel_get_bd_len(u8 *parm_list, u8 cdb10,
static void nvme_trans_modesel_save_bd(struct nvme_ns *ns, u8 *parm_list,
u16 idx, u16 bd_len, u8 llbaa)
{
- u16 bd_num;
-
- bd_num = bd_len / ((llbaa == 0) ?
- SHORT_DESC_BLOCK : LONG_DESC_BLOCK);
/* Store block descriptor info if a FORMAT UNIT comes later */
/* TODO Saving 1st BD info; what to do if multiple BD received? */
if (llbaa == 0) {
@@ -1528,7 +1524,7 @@ static int nvme_trans_fmt_send_cmd(struct nvme_ns *ns, struct sg_io_hdr *hdr,
int nvme_sc;
struct nvme_id_ns *id_ns;
u8 i;
- u8 flbas, nlbaf;
+ u8 nlbaf;
u8 selected_lbaf = 0xFF;
u32 cdw10 = 0;
struct nvme_command c;
@@ -1539,7 +1535,6 @@ static int nvme_trans_fmt_send_cmd(struct nvme_ns *ns, struct sg_io_hdr *hdr,
if (res)
return res;
- flbas = (id_ns->flbas) & 0x0F;
nlbaf = id_ns->nlbaf;
for (i = 0; i < nlbaf; i++) {
@@ -2168,12 +2163,10 @@ static int nvme_trans_synchronize_cache(struct nvme_ns *ns,
static int nvme_trans_start_stop(struct nvme_ns *ns, struct sg_io_hdr *hdr,
u8 *cmd)
{
- u8 immed, pcmod, no_flush, start;
+ u8 immed, no_flush;
immed = cmd[1] & 0x01;
- pcmod = cmd[3] & 0x0f;
no_flush = cmd[4] & 0x04;
- start = cmd[4] & 0x01;
if (immed != 0) {
return nvme_trans_completion(hdr, SAM_STAT_CHECK_CONDITION,
diff --git a/drivers/nvme/target/Kconfig b/drivers/nvme/target/Kconfig
index 3a5b9d0576cb..03e4ab65fe77 100644
--- a/drivers/nvme/target/Kconfig
+++ b/drivers/nvme/target/Kconfig
@@ -34,3 +34,27 @@ config NVME_TARGET_RDMA
devices over RDMA.
If unsure, say N.
+
+config NVME_TARGET_FC
+ tristate "NVMe over Fabrics FC target driver"
+ depends on NVME_TARGET
+ depends on HAS_DMA
+ help
+ This enables the NVMe FC target support, which allows exporting NVMe
+ devices over FC.
+
+ If unsure, say N.
+
+config NVME_TARGET_FCLOOP
+ tristate "NVMe over Fabrics FC Transport Loopback Test driver"
+ depends on NVME_TARGET
+ select NVME_CORE
+ select NVME_FABRICS
+ select SG_POOL
+ depends on NVME_FC
+ depends on NVME_TARGET_FC
+ help
+ This enables the NVMe FC loopback test support, which can be useful
+ to test NVMe-FC transport interfaces.
+
+ If unsure, say N.
diff --git a/drivers/nvme/target/Makefile b/drivers/nvme/target/Makefile
index b7a06232c9da..fecc14f535b2 100644
--- a/drivers/nvme/target/Makefile
+++ b/drivers/nvme/target/Makefile
@@ -2,8 +2,12 @@
obj-$(CONFIG_NVME_TARGET) += nvmet.o
obj-$(CONFIG_NVME_TARGET_LOOP) += nvme-loop.o
obj-$(CONFIG_NVME_TARGET_RDMA) += nvmet-rdma.o
+obj-$(CONFIG_NVME_TARGET_FC) += nvmet-fc.o
+obj-$(CONFIG_NVME_TARGET_FCLOOP) += nvme-fcloop.o
nvmet-y += core.o configfs.o admin-cmd.o io-cmd.o fabrics-cmd.o \
discovery.o
nvme-loop-y += loop.o
nvmet-rdma-y += rdma.o
+nvmet-fc-y += fc.o
+nvme-fcloop-y += fcloop.o
diff --git a/drivers/nvme/target/admin-cmd.c b/drivers/nvme/target/admin-cmd.c
index 6fe4c48a21e4..ec1ad2aa0a4c 100644
--- a/drivers/nvme/target/admin-cmd.c
+++ b/drivers/nvme/target/admin-cmd.c
@@ -237,7 +237,8 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
id->maxcmd = cpu_to_le16(NVMET_MAX_CMD);
id->nn = cpu_to_le32(ctrl->subsys->max_nsid);
- id->oncs = cpu_to_le16(NVME_CTRL_ONCS_DSM);
+ id->oncs = cpu_to_le16(NVME_CTRL_ONCS_DSM |
+ NVME_CTRL_ONCS_WRITE_ZEROES);
/* XXX: don't report vwc if the underlying device is write through */
id->vwc = NVME_CTRL_VWC_PRESENT;
diff --git a/drivers/nvme/target/configfs.c b/drivers/nvme/target/configfs.c
index af5e2dc4a3d5..6f5074153dcd 100644
--- a/drivers/nvme/target/configfs.c
+++ b/drivers/nvme/target/configfs.c
@@ -37,6 +37,8 @@ static ssize_t nvmet_addr_adrfam_show(struct config_item *item,
return sprintf(page, "ipv6\n");
case NVMF_ADDR_FAMILY_IB:
return sprintf(page, "ib\n");
+ case NVMF_ADDR_FAMILY_FC:
+ return sprintf(page, "fc\n");
default:
return sprintf(page, "\n");
}
@@ -59,6 +61,8 @@ static ssize_t nvmet_addr_adrfam_store(struct config_item *item,
port->disc_addr.adrfam = NVMF_ADDR_FAMILY_IP6;
} else if (sysfs_streq(page, "ib")) {
port->disc_addr.adrfam = NVMF_ADDR_FAMILY_IB;
+ } else if (sysfs_streq(page, "fc")) {
+ port->disc_addr.adrfam = NVMF_ADDR_FAMILY_FC;
} else {
pr_err("Invalid value '%s' for adrfam\n", page);
return -EINVAL;
@@ -209,6 +213,8 @@ static ssize_t nvmet_addr_trtype_show(struct config_item *item,
return sprintf(page, "rdma\n");
case NVMF_TRTYPE_LOOP:
return sprintf(page, "loop\n");
+ case NVMF_TRTYPE_FC:
+ return sprintf(page, "fc\n");
default:
return sprintf(page, "\n");
}
@@ -229,6 +235,12 @@ static void nvmet_port_init_tsas_loop(struct nvmet_port *port)
memset(&port->disc_addr.tsas, 0, NVMF_TSAS_SIZE);
}
+static void nvmet_port_init_tsas_fc(struct nvmet_port *port)
+{
+ port->disc_addr.trtype = NVMF_TRTYPE_FC;
+ memset(&port->disc_addr.tsas, 0, NVMF_TSAS_SIZE);
+}
+
static ssize_t nvmet_addr_trtype_store(struct config_item *item,
const char *page, size_t count)
{
@@ -244,6 +256,8 @@ static ssize_t nvmet_addr_trtype_store(struct config_item *item,
nvmet_port_init_tsas_rdma(port);
} else if (sysfs_streq(page, "loop")) {
nvmet_port_init_tsas_loop(port);
+ } else if (sysfs_streq(page, "fc")) {
+ nvmet_port_init_tsas_fc(port);
} else {
pr_err("Invalid value '%s' for trtype\n", page);
return -EINVAL;
@@ -271,7 +285,7 @@ static ssize_t nvmet_ns_device_path_store(struct config_item *item,
mutex_lock(&subsys->lock);
ret = -EBUSY;
- if (nvmet_ns_enabled(ns))
+ if (ns->enabled)
goto out_unlock;
kfree(ns->device_path);
@@ -307,7 +321,7 @@ static ssize_t nvmet_ns_device_nguid_store(struct config_item *item,
int ret = 0;
mutex_lock(&subsys->lock);
- if (nvmet_ns_enabled(ns)) {
+ if (ns->enabled) {
ret = -EBUSY;
goto out_unlock;
}
@@ -339,7 +353,7 @@ CONFIGFS_ATTR(nvmet_ns_, device_nguid);
static ssize_t nvmet_ns_enable_show(struct config_item *item, char *page)
{
- return sprintf(page, "%d\n", nvmet_ns_enabled(to_nvmet_ns(item)));
+ return sprintf(page, "%d\n", to_nvmet_ns(item)->enabled);
}
static ssize_t nvmet_ns_enable_store(struct config_item *item,
@@ -466,7 +480,7 @@ out_free_link:
return ret;
}
-static int nvmet_port_subsys_drop_link(struct config_item *parent,
+static void nvmet_port_subsys_drop_link(struct config_item *parent,
struct config_item *target)
{
struct nvmet_port *port = to_nvmet_port(parent->ci_parent);
@@ -479,7 +493,7 @@ static int nvmet_port_subsys_drop_link(struct config_item *parent,
goto found;
}
up_write(&nvmet_config_sem);
- return -EINVAL;
+ return;
found:
list_del(&p->entry);
@@ -488,7 +502,6 @@ found:
nvmet_disable_port(port);
up_write(&nvmet_config_sem);
kfree(p);
- return 0;
}
static struct configfs_item_operations nvmet_port_subsys_item_ops = {
@@ -542,7 +555,7 @@ out_free_link:
return ret;
}
-static int nvmet_allowed_hosts_drop_link(struct config_item *parent,
+static void nvmet_allowed_hosts_drop_link(struct config_item *parent,
struct config_item *target)
{
struct nvmet_subsys *subsys = to_subsys(parent->ci_parent);
@@ -555,14 +568,13 @@ static int nvmet_allowed_hosts_drop_link(struct config_item *parent,
goto found;
}
up_write(&nvmet_config_sem);
- return -EINVAL;
+ return;
found:
list_del(&p->entry);
nvmet_genctr++;
up_write(&nvmet_config_sem);
kfree(p);
- return 0;
}
static struct configfs_item_operations nvmet_allowed_hosts_item_ops = {
diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c
index b4cacb6f0258..b1d66ed655c9 100644
--- a/drivers/nvme/target/core.c
+++ b/drivers/nvme/target/core.c
@@ -264,7 +264,7 @@ int nvmet_ns_enable(struct nvmet_ns *ns)
int ret = 0;
mutex_lock(&subsys->lock);
- if (!list_empty(&ns->dev_link))
+ if (ns->enabled)
goto out_unlock;
ns->bdev = blkdev_get_by_path(ns->device_path, FMODE_READ | FMODE_WRITE,
@@ -309,6 +309,7 @@ int nvmet_ns_enable(struct nvmet_ns *ns)
list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry)
nvmet_add_async_event(ctrl, NVME_AER_TYPE_NOTICE, 0, 0);
+ ns->enabled = true;
ret = 0;
out_unlock:
mutex_unlock(&subsys->lock);
@@ -325,11 +326,11 @@ void nvmet_ns_disable(struct nvmet_ns *ns)
struct nvmet_ctrl *ctrl;
mutex_lock(&subsys->lock);
- if (list_empty(&ns->dev_link)) {
- mutex_unlock(&subsys->lock);
- return;
- }
- list_del_init(&ns->dev_link);
+ if (!ns->enabled)
+ goto out_unlock;
+
+ ns->enabled = false;
+ list_del_rcu(&ns->dev_link);
mutex_unlock(&subsys->lock);
/*
@@ -351,6 +352,7 @@ void nvmet_ns_disable(struct nvmet_ns *ns)
if (ns->bdev)
blkdev_put(ns->bdev, FMODE_WRITE|FMODE_READ);
+out_unlock:
mutex_unlock(&subsys->lock);
}
@@ -617,7 +619,7 @@ u16 nvmet_ctrl_find_get(const char *subsysnqn, const char *hostnqn, u16 cntlid,
if (!subsys) {
pr_warn("connect request for invalid subsystem %s!\n",
subsysnqn);
- req->rsp->result = IPO_IATTR_CONNECT_DATA(subsysnqn);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_DATA(subsysnqn);
return NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
}
@@ -638,7 +640,7 @@ u16 nvmet_ctrl_find_get(const char *subsysnqn, const char *hostnqn, u16 cntlid,
pr_warn("could not find controller %d for subsys %s / host %s\n",
cntlid, subsysnqn, hostnqn);
- req->rsp->result = IPO_IATTR_CONNECT_DATA(cntlid);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_DATA(cntlid);
status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
out:
@@ -700,7 +702,7 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
if (!subsys) {
pr_warn("connect request for invalid subsystem %s!\n",
subsysnqn);
- req->rsp->result = IPO_IATTR_CONNECT_DATA(subsysnqn);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_DATA(subsysnqn);
goto out;
}
@@ -709,7 +711,7 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
if (!nvmet_host_allowed(req, subsys, hostnqn)) {
pr_info("connect by host %s for subsystem %s not allowed\n",
hostnqn, subsysnqn);
- req->rsp->result = IPO_IATTR_CONNECT_DATA(hostnqn);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_DATA(hostnqn);
up_read(&nvmet_config_sem);
goto out_put_subsystem;
}
@@ -838,9 +840,13 @@ static void nvmet_fatal_error_handler(struct work_struct *work)
void nvmet_ctrl_fatal_error(struct nvmet_ctrl *ctrl)
{
- ctrl->csts |= NVME_CSTS_CFS;
- INIT_WORK(&ctrl->fatal_err_work, nvmet_fatal_error_handler);
- schedule_work(&ctrl->fatal_err_work);
+ mutex_lock(&ctrl->lock);
+ if (!(ctrl->csts & NVME_CSTS_CFS)) {
+ ctrl->csts |= NVME_CSTS_CFS;
+ INIT_WORK(&ctrl->fatal_err_work, nvmet_fatal_error_handler);
+ schedule_work(&ctrl->fatal_err_work);
+ }
+ mutex_unlock(&ctrl->lock);
}
EXPORT_SYMBOL_GPL(nvmet_ctrl_fatal_error);
diff --git a/drivers/nvme/target/fabrics-cmd.c b/drivers/nvme/target/fabrics-cmd.c
index 9a97ae67e656..f4088198cd0d 100644
--- a/drivers/nvme/target/fabrics-cmd.c
+++ b/drivers/nvme/target/fabrics-cmd.c
@@ -69,7 +69,7 @@ static void nvmet_execute_prop_get(struct nvmet_req *req)
}
}
- req->rsp->result64 = cpu_to_le64(val);
+ req->rsp->result.u64 = cpu_to_le64(val);
nvmet_req_complete(req, status);
}
@@ -125,7 +125,7 @@ static void nvmet_execute_admin_connect(struct nvmet_req *req)
d = kmap(sg_page(req->sg)) + req->sg->offset;
/* zero out initial completion result, assign values as needed */
- req->rsp->result = 0;
+ req->rsp->result.u32 = 0;
if (c->recfmt != 0) {
pr_warn("invalid connect version (%d).\n",
@@ -138,7 +138,7 @@ static void nvmet_execute_admin_connect(struct nvmet_req *req)
pr_warn("connect attempt for invalid controller ID %#x\n",
d->cntlid);
status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
- req->rsp->result = IPO_IATTR_CONNECT_DATA(cntlid);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_DATA(cntlid);
goto out;
}
@@ -155,7 +155,7 @@ static void nvmet_execute_admin_connect(struct nvmet_req *req)
pr_info("creating controller %d for NQN %s.\n",
ctrl->cntlid, ctrl->hostnqn);
- req->rsp->result16 = cpu_to_le16(ctrl->cntlid);
+ req->rsp->result.u16 = cpu_to_le16(ctrl->cntlid);
out:
kunmap(sg_page(req->sg));
@@ -173,7 +173,7 @@ static void nvmet_execute_io_connect(struct nvmet_req *req)
d = kmap(sg_page(req->sg)) + req->sg->offset;
/* zero out initial completion result, assign values as needed */
- req->rsp->result = 0;
+ req->rsp->result.u32 = 0;
if (c->recfmt != 0) {
pr_warn("invalid connect version (%d).\n",
@@ -191,14 +191,14 @@ static void nvmet_execute_io_connect(struct nvmet_req *req)
if (unlikely(qid > ctrl->subsys->max_qid)) {
pr_warn("invalid queue id (%d)\n", qid);
status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
- req->rsp->result = IPO_IATTR_CONNECT_SQE(qid);
+ req->rsp->result.u32 = IPO_IATTR_CONNECT_SQE(qid);
goto out_ctrl_put;
}
status = nvmet_install_queue(ctrl, req);
if (status) {
/* pass back cntlid that had the issue of installing queue */
- req->rsp->result16 = cpu_to_le16(ctrl->cntlid);
+ req->rsp->result.u16 = cpu_to_le16(ctrl->cntlid);
goto out_ctrl_put;
}
diff --git a/drivers/nvme/target/fc.c b/drivers/nvme/target/fc.c
new file mode 100644
index 000000000000..173e842f19c9
--- /dev/null
+++ b/drivers/nvme/target/fc.c
@@ -0,0 +1,2288 @@
+/*
+ * Copyright (c) 2016 Avago Technologies. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful.
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND WARRANTIES,
+ * INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A
+ * PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE DISCLAIMED, EXCEPT TO
+ * THE EXTENT THAT SUCH DISCLAIMERS ARE HELD TO BE LEGALLY INVALID.
+ * See the GNU General Public License for more details, a copy of which
+ * can be found in the file COPYING included with this package
+ *
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/blk-mq.h>
+#include <linux/parser.h>
+#include <linux/random.h>
+#include <uapi/scsi/fc/fc_fs.h>
+#include <uapi/scsi/fc/fc_els.h>
+
+#include "nvmet.h"
+#include <linux/nvme-fc-driver.h>
+#include <linux/nvme-fc.h>
+
+
+/* *************************** Data Structures/Defines ****************** */
+
+
+#define NVMET_LS_CTX_COUNT 4
+
+/* for this implementation, assume small single frame rqst/rsp */
+#define NVME_FC_MAX_LS_BUFFER_SIZE 2048
+
+struct nvmet_fc_tgtport;
+struct nvmet_fc_tgt_assoc;
+
+struct nvmet_fc_ls_iod {
+ struct nvmefc_tgt_ls_req *lsreq;
+ struct nvmefc_tgt_fcp_req *fcpreq; /* only if RS */
+
+ struct list_head ls_list; /* tgtport->ls_list */
+
+ struct nvmet_fc_tgtport *tgtport;
+ struct nvmet_fc_tgt_assoc *assoc;
+
+ u8 *rqstbuf;
+ u8 *rspbuf;
+ u16 rqstdatalen;
+ dma_addr_t rspdma;
+
+ struct scatterlist sg[2];
+
+ struct work_struct work;
+} __aligned(sizeof(unsigned long long));
+
+#define NVMET_FC_MAX_KB_PER_XFR 256
+
+enum nvmet_fcp_datadir {
+ NVMET_FCP_NODATA,
+ NVMET_FCP_WRITE,
+ NVMET_FCP_READ,
+ NVMET_FCP_ABORTED,
+};
+
+struct nvmet_fc_fcp_iod {
+ struct nvmefc_tgt_fcp_req *fcpreq;
+
+ struct nvme_fc_cmd_iu cmdiubuf;
+ struct nvme_fc_ersp_iu rspiubuf;
+ dma_addr_t rspdma;
+ struct scatterlist *data_sg;
+ struct scatterlist *next_sg;
+ int data_sg_cnt;
+ u32 next_sg_offset;
+ u32 total_length;
+ u32 offset;
+ enum nvmet_fcp_datadir io_dir;
+ bool active;
+ bool abort;
+ spinlock_t flock;
+
+ struct nvmet_req req;
+ struct work_struct work;
+
+ struct nvmet_fc_tgtport *tgtport;
+ struct nvmet_fc_tgt_queue *queue;
+
+ struct list_head fcp_list; /* tgtport->fcp_list */
+};
+
+struct nvmet_fc_tgtport {
+
+ struct nvmet_fc_target_port fc_target_port;
+
+ struct list_head tgt_list; /* nvmet_fc_target_list */
+ struct device *dev; /* dev for dma mapping */
+ struct nvmet_fc_target_template *ops;
+
+ struct nvmet_fc_ls_iod *iod;
+ spinlock_t lock;
+ struct list_head ls_list;
+ struct list_head ls_busylist;
+ struct list_head assoc_list;
+ struct ida assoc_cnt;
+ struct nvmet_port *port;
+ struct kref ref;
+};
+
+struct nvmet_fc_tgt_queue {
+ bool ninetypercent;
+ u16 qid;
+ u16 sqsize;
+ u16 ersp_ratio;
+ u16 sqhd;
+ int cpu;
+ atomic_t connected;
+ atomic_t sqtail;
+ atomic_t zrspcnt;
+ atomic_t rsn;
+ spinlock_t qlock;
+ struct nvmet_port *port;
+ struct nvmet_cq nvme_cq;
+ struct nvmet_sq nvme_sq;
+ struct nvmet_fc_tgt_assoc *assoc;
+ struct nvmet_fc_fcp_iod *fod; /* array of fcp_iods */
+ struct list_head fod_list;
+ struct workqueue_struct *work_q;
+ struct kref ref;
+} __aligned(sizeof(unsigned long long));
+
+struct nvmet_fc_tgt_assoc {
+ u64 association_id;
+ u32 a_id;
+ struct nvmet_fc_tgtport *tgtport;
+ struct list_head a_list;
+ struct nvmet_fc_tgt_queue *queues[NVMET_NR_QUEUES];
+ struct kref ref;
+};
+
+
+static inline int
+nvmet_fc_iodnum(struct nvmet_fc_ls_iod *iodptr)
+{
+ return (iodptr - iodptr->tgtport->iod);
+}
+
+static inline int
+nvmet_fc_fodnum(struct nvmet_fc_fcp_iod *fodptr)
+{
+ return (fodptr - fodptr->queue->fod);
+}
+
+
+/*
+ * Association and Connection IDs:
+ *
+ * Association ID will have random number in upper 6 bytes and zero
+ * in lower 2 bytes
+ *
+ * Connection IDs will be Association ID with QID or'd in lower 2 bytes
+ *
+ * note: Association ID = Connection ID for queue 0
+ */
+#define BYTES_FOR_QID sizeof(u16)
+#define BYTES_FOR_QID_SHIFT (BYTES_FOR_QID * 8)
+#define NVMET_FC_QUEUEID_MASK ((u64)((1 << BYTES_FOR_QID_SHIFT) - 1))
+
+static inline u64
+nvmet_fc_makeconnid(struct nvmet_fc_tgt_assoc *assoc, u16 qid)
+{
+ return (assoc->association_id | qid);
+}
+
+static inline u64
+nvmet_fc_getassociationid(u64 connectionid)
+{
+ return connectionid & ~NVMET_FC_QUEUEID_MASK;
+}
+
+static inline u16
+nvmet_fc_getqueueid(u64 connectionid)
+{
+ return (u16)(connectionid & NVMET_FC_QUEUEID_MASK);
+}
+
+static inline struct nvmet_fc_tgtport *
+targetport_to_tgtport(struct nvmet_fc_target_port *targetport)
+{
+ return container_of(targetport, struct nvmet_fc_tgtport,
+ fc_target_port);
+}
+
+static inline struct nvmet_fc_fcp_iod *
+nvmet_req_to_fod(struct nvmet_req *nvme_req)
+{
+ return container_of(nvme_req, struct nvmet_fc_fcp_iod, req);
+}
+
+
+/* *************************** Globals **************************** */
+
+
+static DEFINE_SPINLOCK(nvmet_fc_tgtlock);
+
+static LIST_HEAD(nvmet_fc_target_list);
+static DEFINE_IDA(nvmet_fc_tgtport_cnt);
+
+
+static void nvmet_fc_handle_ls_rqst_work(struct work_struct *work);
+static void nvmet_fc_handle_fcp_rqst_work(struct work_struct *work);
+static void nvmet_fc_tgt_a_put(struct nvmet_fc_tgt_assoc *assoc);
+static int nvmet_fc_tgt_a_get(struct nvmet_fc_tgt_assoc *assoc);
+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);
+
+
+/* *********************** FC-NVME DMA Handling **************************** */
+
+/*
+ * The fcloop device passes in a NULL device pointer. Real LLD's will
+ * pass in a valid device pointer. If NULL is passed to the dma mapping
+ * routines, depending on the platform, it may or may not succeed, and
+ * may crash.
+ *
+ * As such:
+ * Wrapper all the dma routines and check the dev pointer.
+ *
+ * If simple mappings (return just a dma address, we'll noop them,
+ * returning a dma address of 0.
+ *
+ * On more complex mappings (dma_map_sg), a pseudo routine fills
+ * in the scatter list, setting all dma addresses to 0.
+ */
+
+static inline dma_addr_t
+fc_dma_map_single(struct device *dev, void *ptr, size_t size,
+ enum dma_data_direction dir)
+{
+ return dev ? dma_map_single(dev, ptr, size, dir) : (dma_addr_t)0L;
+}
+
+static inline int
+fc_dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
+{
+ return dev ? dma_mapping_error(dev, dma_addr) : 0;
+}
+
+static inline void
+fc_dma_unmap_single(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_unmap_single(dev, addr, size, dir);
+}
+
+static inline void
+fc_dma_sync_single_for_cpu(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_sync_single_for_cpu(dev, addr, size, dir);
+}
+
+static inline void
+fc_dma_sync_single_for_device(struct device *dev, dma_addr_t addr, size_t size,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_sync_single_for_device(dev, addr, size, dir);
+}
+
+/* pseudo dma_map_sg call */
+static int
+fc_map_sg(struct scatterlist *sg, int nents)
+{
+ struct scatterlist *s;
+ int i;
+
+ WARN_ON(nents == 0 || sg[0].length == 0);
+
+ for_each_sg(sg, s, nents, i) {
+ s->dma_address = 0L;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+ s->dma_length = s->length;
+#endif
+ }
+ return nents;
+}
+
+static inline int
+fc_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction dir)
+{
+ return dev ? dma_map_sg(dev, sg, nents, dir) : fc_map_sg(sg, nents);
+}
+
+static inline void
+fc_dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction dir)
+{
+ if (dev)
+ dma_unmap_sg(dev, sg, nents, dir);
+}
+
+
+/* *********************** FC-NVME Port Management ************************ */
+
+
+static int
+nvmet_fc_alloc_ls_iodlist(struct nvmet_fc_tgtport *tgtport)
+{
+ struct nvmet_fc_ls_iod *iod;
+ int i;
+
+ iod = kcalloc(NVMET_LS_CTX_COUNT, sizeof(struct nvmet_fc_ls_iod),
+ GFP_KERNEL);
+ if (!iod)
+ return -ENOMEM;
+
+ tgtport->iod = iod;
+
+ for (i = 0; i < NVMET_LS_CTX_COUNT; iod++, i++) {
+ INIT_WORK(&iod->work, nvmet_fc_handle_ls_rqst_work);
+ iod->tgtport = tgtport;
+ list_add_tail(&iod->ls_list, &tgtport->ls_list);
+
+ iod->rqstbuf = kcalloc(2, NVME_FC_MAX_LS_BUFFER_SIZE,
+ GFP_KERNEL);
+ if (!iod->rqstbuf)
+ goto out_fail;
+
+ iod->rspbuf = iod->rqstbuf + NVME_FC_MAX_LS_BUFFER_SIZE;
+
+ iod->rspdma = fc_dma_map_single(tgtport->dev, iod->rspbuf,
+ NVME_FC_MAX_LS_BUFFER_SIZE,
+ DMA_TO_DEVICE);
+ if (fc_dma_mapping_error(tgtport->dev, iod->rspdma))
+ goto out_fail;
+ }
+
+ return 0;
+
+out_fail:
+ kfree(iod->rqstbuf);
+ list_del(&iod->ls_list);
+ for (iod--, i--; i >= 0; iod--, i--) {
+ fc_dma_unmap_single(tgtport->dev, iod->rspdma,
+ NVME_FC_MAX_LS_BUFFER_SIZE, DMA_TO_DEVICE);
+ kfree(iod->rqstbuf);
+ list_del(&iod->ls_list);
+ }
+
+ kfree(iod);
+
+ return -EFAULT;
+}
+
+static void
+nvmet_fc_free_ls_iodlist(struct nvmet_fc_tgtport *tgtport)
+{
+ struct nvmet_fc_ls_iod *iod = tgtport->iod;
+ int i;
+
+ for (i = 0; i < NVMET_LS_CTX_COUNT; iod++, i++) {
+ fc_dma_unmap_single(tgtport->dev,
+ iod->rspdma, NVME_FC_MAX_LS_BUFFER_SIZE,
+ DMA_TO_DEVICE);
+ kfree(iod->rqstbuf);
+ list_del(&iod->ls_list);
+ }
+ kfree(tgtport->iod);
+}
+
+static struct nvmet_fc_ls_iod *
+nvmet_fc_alloc_ls_iod(struct nvmet_fc_tgtport *tgtport)
+{
+ static struct nvmet_fc_ls_iod *iod;
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ iod = list_first_entry_or_null(&tgtport->ls_list,
+ struct nvmet_fc_ls_iod, ls_list);
+ if (iod)
+ list_move_tail(&iod->ls_list, &tgtport->ls_busylist);
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ return iod;
+}
+
+
+static void
+nvmet_fc_free_ls_iod(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_move(&iod->ls_list, &tgtport->ls_list);
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+}
+
+static void
+nvmet_fc_prep_fcp_iodlist(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_tgt_queue *queue)
+{
+ struct nvmet_fc_fcp_iod *fod = queue->fod;
+ int i;
+
+ for (i = 0; i < queue->sqsize; fod++, i++) {
+ INIT_WORK(&fod->work, nvmet_fc_handle_fcp_rqst_work);
+ fod->tgtport = tgtport;
+ fod->queue = queue;
+ fod->active = false;
+ list_add_tail(&fod->fcp_list, &queue->fod_list);
+ spin_lock_init(&fod->flock);
+
+ fod->rspdma = fc_dma_map_single(tgtport->dev, &fod->rspiubuf,
+ sizeof(fod->rspiubuf), DMA_TO_DEVICE);
+ if (fc_dma_mapping_error(tgtport->dev, fod->rspdma)) {
+ list_del(&fod->fcp_list);
+ for (fod--, i--; i >= 0; fod--, i--) {
+ fc_dma_unmap_single(tgtport->dev, fod->rspdma,
+ sizeof(fod->rspiubuf),
+ DMA_TO_DEVICE);
+ fod->rspdma = 0L;
+ list_del(&fod->fcp_list);
+ }
+
+ return;
+ }
+ }
+}
+
+static void
+nvmet_fc_destroy_fcp_iodlist(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_tgt_queue *queue)
+{
+ struct nvmet_fc_fcp_iod *fod = queue->fod;
+ int i;
+
+ for (i = 0; i < queue->sqsize; fod++, i++) {
+ if (fod->rspdma)
+ fc_dma_unmap_single(tgtport->dev, fod->rspdma,
+ sizeof(fod->rspiubuf), DMA_TO_DEVICE);
+ }
+}
+
+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);
+ fod = list_first_entry_or_null(&queue->fod_list,
+ struct nvmet_fc_fcp_iod, fcp_list);
+ if (fod) {
+ list_del(&fod->fcp_list);
+ fod->active = true;
+ fod->abort = false;
+ /*
+ * no queue reference is taken, as it was taken by the
+ * queue lookup just prior to the allocation. The iod
+ * will "inherit" that reference.
+ */
+ }
+ spin_unlock_irqrestore(&queue->qlock, flags);
+ return fod;
+}
+
+
+static void
+nvmet_fc_free_fcp_iod(struct nvmet_fc_tgt_queue *queue,
+ struct nvmet_fc_fcp_iod *fod)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&queue->qlock, flags);
+ list_add_tail(&fod->fcp_list, &fod->queue->fod_list);
+ fod->active = false;
+ spin_unlock_irqrestore(&queue->qlock, flags);
+
+ /*
+ * release the reference taken at queue lookup and fod allocation
+ */
+ nvmet_fc_tgt_q_put(queue);
+}
+
+static int
+nvmet_fc_queue_to_cpu(struct nvmet_fc_tgtport *tgtport, int qid)
+{
+ int cpu, idx, cnt;
+
+ if (!(tgtport->ops->target_features &
+ NVMET_FCTGTFEAT_NEEDS_CMD_CPUSCHED) ||
+ tgtport->ops->max_hw_queues == 1)
+ return WORK_CPU_UNBOUND;
+
+ /* Simple cpu selection based on qid modulo active cpu count */
+ idx = !qid ? 0 : (qid - 1) % num_active_cpus();
+
+ /* find the n'th active cpu */
+ for (cpu = 0, cnt = 0; ; ) {
+ if (cpu_active(cpu)) {
+ if (cnt == idx)
+ break;
+ cnt++;
+ }
+ cpu = (cpu + 1) % num_possible_cpus();
+ }
+
+ return cpu;
+}
+
+static struct nvmet_fc_tgt_queue *
+nvmet_fc_alloc_target_queue(struct nvmet_fc_tgt_assoc *assoc,
+ u16 qid, u16 sqsize)
+{
+ struct nvmet_fc_tgt_queue *queue;
+ unsigned long flags;
+ int ret;
+
+ if (qid >= NVMET_NR_QUEUES)
+ return NULL;
+
+ queue = kzalloc((sizeof(*queue) +
+ (sizeof(struct nvmet_fc_fcp_iod) * sqsize)),
+ GFP_KERNEL);
+ if (!queue)
+ return NULL;
+
+ if (!nvmet_fc_tgt_a_get(assoc))
+ goto out_free_queue;
+
+ queue->work_q = alloc_workqueue("ntfc%d.%d.%d", 0, 0,
+ assoc->tgtport->fc_target_port.port_num,
+ assoc->a_id, qid);
+ if (!queue->work_q)
+ goto out_a_put;
+
+ queue->fod = (struct nvmet_fc_fcp_iod *)&queue[1];
+ queue->qid = qid;
+ queue->sqsize = sqsize;
+ queue->assoc = assoc;
+ queue->port = assoc->tgtport->port;
+ queue->cpu = nvmet_fc_queue_to_cpu(assoc->tgtport, qid);
+ INIT_LIST_HEAD(&queue->fod_list);
+ atomic_set(&queue->connected, 0);
+ atomic_set(&queue->sqtail, 0);
+ atomic_set(&queue->rsn, 1);
+ atomic_set(&queue->zrspcnt, 0);
+ spin_lock_init(&queue->qlock);
+ kref_init(&queue->ref);
+
+ nvmet_fc_prep_fcp_iodlist(assoc->tgtport, queue);
+
+ ret = nvmet_sq_init(&queue->nvme_sq);
+ if (ret)
+ goto out_fail_iodlist;
+
+ WARN_ON(assoc->queues[qid]);
+ spin_lock_irqsave(&assoc->tgtport->lock, flags);
+ assoc->queues[qid] = queue;
+ spin_unlock_irqrestore(&assoc->tgtport->lock, flags);
+
+ return queue;
+
+out_fail_iodlist:
+ nvmet_fc_destroy_fcp_iodlist(assoc->tgtport, queue);
+ destroy_workqueue(queue->work_q);
+out_a_put:
+ nvmet_fc_tgt_a_put(assoc);
+out_free_queue:
+ kfree(queue);
+ return NULL;
+}
+
+
+static void
+nvmet_fc_tgt_queue_free(struct kref *ref)
+{
+ struct nvmet_fc_tgt_queue *queue =
+ container_of(ref, struct nvmet_fc_tgt_queue, ref);
+ unsigned long flags;
+
+ spin_lock_irqsave(&queue->assoc->tgtport->lock, flags);
+ queue->assoc->queues[queue->qid] = NULL;
+ spin_unlock_irqrestore(&queue->assoc->tgtport->lock, flags);
+
+ nvmet_fc_destroy_fcp_iodlist(queue->assoc->tgtport, queue);
+
+ nvmet_fc_tgt_a_put(queue->assoc);
+
+ destroy_workqueue(queue->work_q);
+
+ kfree(queue);
+}
+
+static void
+nvmet_fc_tgt_q_put(struct nvmet_fc_tgt_queue *queue)
+{
+ kref_put(&queue->ref, nvmet_fc_tgt_queue_free);
+}
+
+static int
+nvmet_fc_tgt_q_get(struct nvmet_fc_tgt_queue *queue)
+{
+ return kref_get_unless_zero(&queue->ref);
+}
+
+
+static void
+nvmet_fc_abort_op(struct nvmet_fc_tgtport *tgtport,
+ struct nvmefc_tgt_fcp_req *fcpreq)
+{
+ int ret;
+
+ fcpreq->op = NVMET_FCOP_ABORT;
+ fcpreq->offset = 0;
+ fcpreq->timeout = 0;
+ fcpreq->transfer_length = 0;
+ fcpreq->transferred_length = 0;
+ fcpreq->fcp_error = 0;
+ fcpreq->sg_cnt = 0;
+
+ ret = tgtport->ops->fcp_op(&tgtport->fc_target_port, fcpreq);
+ if (ret)
+ /* should never reach here !! */
+ WARN_ON(1);
+}
+
+
+static void
+nvmet_fc_delete_target_queue(struct nvmet_fc_tgt_queue *queue)
+{
+ struct nvmet_fc_fcp_iod *fod = queue->fod;
+ unsigned long flags;
+ int i;
+ bool disconnect;
+
+ disconnect = atomic_xchg(&queue->connected, 0);
+
+ spin_lock_irqsave(&queue->qlock, flags);
+ /* about outstanding io's */
+ for (i = 0; i < queue->sqsize; fod++, i++) {
+ if (fod->active) {
+ spin_lock(&fod->flock);
+ fod->abort = true;
+ spin_unlock(&fod->flock);
+ }
+ }
+ spin_unlock_irqrestore(&queue->qlock, flags);
+
+ flush_workqueue(queue->work_q);
+
+ if (disconnect)
+ nvmet_sq_destroy(&queue->nvme_sq);
+
+ nvmet_fc_tgt_q_put(queue);
+}
+
+static struct nvmet_fc_tgt_queue *
+nvmet_fc_find_target_queue(struct nvmet_fc_tgtport *tgtport,
+ u64 connection_id)
+{
+ struct nvmet_fc_tgt_assoc *assoc;
+ struct nvmet_fc_tgt_queue *queue;
+ u64 association_id = nvmet_fc_getassociationid(connection_id);
+ u16 qid = nvmet_fc_getqueueid(connection_id);
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_for_each_entry(assoc, &tgtport->assoc_list, a_list) {
+ if (association_id == assoc->association_id) {
+ queue = assoc->queues[qid];
+ if (queue &&
+ (!atomic_read(&queue->connected) ||
+ !nvmet_fc_tgt_q_get(queue)))
+ queue = NULL;
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ return queue;
+ }
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ return NULL;
+}
+
+static struct nvmet_fc_tgt_assoc *
+nvmet_fc_alloc_target_assoc(struct nvmet_fc_tgtport *tgtport)
+{
+ struct nvmet_fc_tgt_assoc *assoc, *tmpassoc;
+ unsigned long flags;
+ u64 ran;
+ int idx;
+ bool needrandom = true;
+
+ assoc = kzalloc(sizeof(*assoc), GFP_KERNEL);
+ if (!assoc)
+ return NULL;
+
+ idx = ida_simple_get(&tgtport->assoc_cnt, 0, 0, GFP_KERNEL);
+ if (idx < 0)
+ goto out_free_assoc;
+
+ if (!nvmet_fc_tgtport_get(tgtport))
+ goto out_ida_put;
+
+ assoc->tgtport = tgtport;
+ assoc->a_id = idx;
+ INIT_LIST_HEAD(&assoc->a_list);
+ kref_init(&assoc->ref);
+
+ while (needrandom) {
+ get_random_bytes(&ran, sizeof(ran) - BYTES_FOR_QID);
+ ran = ran << BYTES_FOR_QID_SHIFT;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ needrandom = false;
+ list_for_each_entry(tmpassoc, &tgtport->assoc_list, a_list)
+ if (ran == tmpassoc->association_id) {
+ needrandom = true;
+ break;
+ }
+ if (!needrandom) {
+ assoc->association_id = ran;
+ list_add_tail(&assoc->a_list, &tgtport->assoc_list);
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ }
+
+ return assoc;
+
+out_ida_put:
+ ida_simple_remove(&tgtport->assoc_cnt, idx);
+out_free_assoc:
+ kfree(assoc);
+ return NULL;
+}
+
+static void
+nvmet_fc_target_assoc_free(struct kref *ref)
+{
+ struct nvmet_fc_tgt_assoc *assoc =
+ container_of(ref, struct nvmet_fc_tgt_assoc, ref);
+ struct nvmet_fc_tgtport *tgtport = assoc->tgtport;
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_del(&assoc->a_list);
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ ida_simple_remove(&tgtport->assoc_cnt, assoc->a_id);
+ kfree(assoc);
+ nvmet_fc_tgtport_put(tgtport);
+}
+
+static void
+nvmet_fc_tgt_a_put(struct nvmet_fc_tgt_assoc *assoc)
+{
+ kref_put(&assoc->ref, nvmet_fc_target_assoc_free);
+}
+
+static int
+nvmet_fc_tgt_a_get(struct nvmet_fc_tgt_assoc *assoc)
+{
+ return kref_get_unless_zero(&assoc->ref);
+}
+
+static void
+nvmet_fc_delete_target_assoc(struct nvmet_fc_tgt_assoc *assoc)
+{
+ struct nvmet_fc_tgtport *tgtport = assoc->tgtport;
+ struct nvmet_fc_tgt_queue *queue;
+ unsigned long flags;
+ int i;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ for (i = NVMET_NR_QUEUES - 1; i >= 0; i--) {
+ queue = assoc->queues[i];
+ if (queue) {
+ if (!nvmet_fc_tgt_q_get(queue))
+ continue;
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ nvmet_fc_delete_target_queue(queue);
+ nvmet_fc_tgt_q_put(queue);
+ spin_lock_irqsave(&tgtport->lock, flags);
+ }
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+
+ nvmet_fc_tgt_a_put(assoc);
+}
+
+static struct nvmet_fc_tgt_assoc *
+nvmet_fc_find_target_assoc(struct nvmet_fc_tgtport *tgtport,
+ u64 association_id)
+{
+ struct nvmet_fc_tgt_assoc *assoc;
+ struct nvmet_fc_tgt_assoc *ret = NULL;
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_for_each_entry(assoc, &tgtport->assoc_list, a_list) {
+ if (association_id == assoc->association_id) {
+ ret = assoc;
+ nvmet_fc_tgt_a_get(assoc);
+ break;
+ }
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+
+ return ret;
+}
+
+
+/**
+ * nvme_fc_register_targetport - transport entry point called by an
+ * LLDD to register the existence of a local
+ * NVME subystem FC port.
+ * @pinfo: pointer to information about the port to be registered
+ * @template: LLDD entrypoints and operational parameters for the port
+ * @dev: physical hardware device node port corresponds to. Will be
+ * used for DMA mappings
+ * @portptr: pointer to a local port pointer. Upon success, the routine
+ * will allocate a nvme_fc_local_port structure and place its
+ * address in the local port pointer. Upon failure, local port
+ * pointer will be set to NULL.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvmet_fc_register_targetport(struct nvmet_fc_port_info *pinfo,
+ struct nvmet_fc_target_template *template,
+ struct device *dev,
+ struct nvmet_fc_target_port **portptr)
+{
+ struct nvmet_fc_tgtport *newrec;
+ unsigned long flags;
+ int ret, idx;
+
+ if (!template->xmt_ls_rsp || !template->fcp_op ||
+ !template->targetport_delete ||
+ !template->max_hw_queues || !template->max_sgl_segments ||
+ !template->max_dif_sgl_segments || !template->dma_boundary) {
+ ret = -EINVAL;
+ goto out_regtgt_failed;
+ }
+
+ newrec = kzalloc((sizeof(*newrec) + template->target_priv_sz),
+ GFP_KERNEL);
+ if (!newrec) {
+ ret = -ENOMEM;
+ goto out_regtgt_failed;
+ }
+
+ idx = ida_simple_get(&nvmet_fc_tgtport_cnt, 0, 0, GFP_KERNEL);
+ if (idx < 0) {
+ ret = -ENOSPC;
+ goto out_fail_kfree;
+ }
+
+ if (!get_device(dev) && dev) {
+ ret = -ENODEV;
+ goto out_ida_put;
+ }
+
+ newrec->fc_target_port.node_name = pinfo->node_name;
+ newrec->fc_target_port.port_name = pinfo->port_name;
+ newrec->fc_target_port.private = &newrec[1];
+ newrec->fc_target_port.port_id = pinfo->port_id;
+ newrec->fc_target_port.port_num = idx;
+ INIT_LIST_HEAD(&newrec->tgt_list);
+ newrec->dev = dev;
+ newrec->ops = template;
+ spin_lock_init(&newrec->lock);
+ INIT_LIST_HEAD(&newrec->ls_list);
+ INIT_LIST_HEAD(&newrec->ls_busylist);
+ INIT_LIST_HEAD(&newrec->assoc_list);
+ kref_init(&newrec->ref);
+ ida_init(&newrec->assoc_cnt);
+
+ ret = nvmet_fc_alloc_ls_iodlist(newrec);
+ if (ret) {
+ ret = -ENOMEM;
+ goto out_free_newrec;
+ }
+
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ list_add_tail(&newrec->tgt_list, &nvmet_fc_target_list);
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+
+ *portptr = &newrec->fc_target_port;
+ return 0;
+
+out_free_newrec:
+ put_device(dev);
+out_ida_put:
+ ida_simple_remove(&nvmet_fc_tgtport_cnt, idx);
+out_fail_kfree:
+ kfree(newrec);
+out_regtgt_failed:
+ *portptr = NULL;
+ return ret;
+}
+EXPORT_SYMBOL_GPL(nvmet_fc_register_targetport);
+
+
+static void
+nvmet_fc_free_tgtport(struct kref *ref)
+{
+ struct nvmet_fc_tgtport *tgtport =
+ container_of(ref, struct nvmet_fc_tgtport, ref);
+ struct device *dev = tgtport->dev;
+ unsigned long flags;
+
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ list_del(&tgtport->tgt_list);
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+
+ nvmet_fc_free_ls_iodlist(tgtport);
+
+ /* let the LLDD know we've finished tearing it down */
+ tgtport->ops->targetport_delete(&tgtport->fc_target_port);
+
+ ida_simple_remove(&nvmet_fc_tgtport_cnt,
+ tgtport->fc_target_port.port_num);
+
+ ida_destroy(&tgtport->assoc_cnt);
+
+ kfree(tgtport);
+
+ put_device(dev);
+}
+
+static void
+nvmet_fc_tgtport_put(struct nvmet_fc_tgtport *tgtport)
+{
+ kref_put(&tgtport->ref, nvmet_fc_free_tgtport);
+}
+
+static int
+nvmet_fc_tgtport_get(struct nvmet_fc_tgtport *tgtport)
+{
+ return kref_get_unless_zero(&tgtport->ref);
+}
+
+static void
+__nvmet_fc_free_assocs(struct nvmet_fc_tgtport *tgtport)
+{
+ struct nvmet_fc_tgt_assoc *assoc, *next;
+ unsigned long flags;
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_for_each_entry_safe(assoc, next,
+ &tgtport->assoc_list, a_list) {
+ if (!nvmet_fc_tgt_a_get(assoc))
+ continue;
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+ nvmet_fc_delete_target_assoc(assoc);
+ nvmet_fc_tgt_a_put(assoc);
+ spin_lock_irqsave(&tgtport->lock, flags);
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+}
+
+/*
+ * nvmet layer has called to terminate an association
+ */
+static void
+nvmet_fc_delete_ctrl(struct nvmet_ctrl *ctrl)
+{
+ struct nvmet_fc_tgtport *tgtport, *next;
+ struct nvmet_fc_tgt_assoc *assoc;
+ struct nvmet_fc_tgt_queue *queue;
+ unsigned long flags;
+ bool found_ctrl = false;
+
+ /* this is a bit ugly, but don't want to make locks layered */
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ list_for_each_entry_safe(tgtport, next, &nvmet_fc_target_list,
+ tgt_list) {
+ if (!nvmet_fc_tgtport_get(tgtport))
+ continue;
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+
+ spin_lock_irqsave(&tgtport->lock, flags);
+ list_for_each_entry(assoc, &tgtport->assoc_list, a_list) {
+ queue = assoc->queues[0];
+ if (queue && queue->nvme_sq.ctrl == ctrl) {
+ if (nvmet_fc_tgt_a_get(assoc))
+ found_ctrl = true;
+ break;
+ }
+ }
+ spin_unlock_irqrestore(&tgtport->lock, flags);
+
+ nvmet_fc_tgtport_put(tgtport);
+
+ if (found_ctrl) {
+ nvmet_fc_delete_target_assoc(assoc);
+ nvmet_fc_tgt_a_put(assoc);
+ return;
+ }
+
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ }
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+}
+
+/**
+ * nvme_fc_unregister_targetport - transport entry point called by an
+ * LLDD to deregister/remove a previously
+ * registered a local NVME subsystem FC port.
+ * @tgtport: pointer to the (registered) target port that is to be
+ * deregistered.
+ *
+ * Returns:
+ * a completion status. Must be 0 upon success; a negative errno
+ * (ex: -ENXIO) upon failure.
+ */
+int
+nvmet_fc_unregister_targetport(struct nvmet_fc_target_port *target_port)
+{
+ struct nvmet_fc_tgtport *tgtport = targetport_to_tgtport(target_port);
+
+ /* terminate any outstanding associations */
+ __nvmet_fc_free_assocs(tgtport);
+
+ nvmet_fc_tgtport_put(tgtport);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(nvmet_fc_unregister_targetport);
+
+
+/* *********************** FC-NVME LS Handling **************************** */
+
+
+static void
+nvmet_fc_format_rsp_hdr(void *buf, u8 ls_cmd, u32 desc_len, u8 rqst_ls_cmd)
+{
+ struct fcnvme_ls_acc_hdr *acc = buf;
+
+ acc->w0.ls_cmd = ls_cmd;
+ acc->desc_list_len = desc_len;
+ acc->rqst.desc_tag = cpu_to_be32(FCNVME_LSDESC_RQST);
+ acc->rqst.desc_len =
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_rqst));
+ acc->rqst.w0.ls_cmd = rqst_ls_cmd;
+}
+
+static int
+nvmet_fc_format_rjt(void *buf, u16 buflen, u8 ls_cmd,
+ u8 reason, u8 explanation, u8 vendor)
+{
+ struct fcnvme_ls_rjt *rjt = buf;
+
+ nvmet_fc_format_rsp_hdr(buf, FCNVME_LSDESC_RQST,
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_ls_rjt)),
+ ls_cmd);
+ rjt->rjt.desc_tag = cpu_to_be32(FCNVME_LSDESC_RJT);
+ rjt->rjt.desc_len = fcnvme_lsdesc_len(sizeof(struct fcnvme_lsdesc_rjt));
+ rjt->rjt.reason_code = reason;
+ rjt->rjt.reason_explanation = explanation;
+ rjt->rjt.vendor = vendor;
+
+ return sizeof(struct fcnvme_ls_rjt);
+}
+
+/* Validation Error indexes into the string table below */
+enum {
+ VERR_NO_ERROR = 0,
+ VERR_CR_ASSOC_LEN = 1,
+ VERR_CR_ASSOC_RQST_LEN = 2,
+ VERR_CR_ASSOC_CMD = 3,
+ VERR_CR_ASSOC_CMD_LEN = 4,
+ VERR_ERSP_RATIO = 5,
+ VERR_ASSOC_ALLOC_FAIL = 6,
+ VERR_QUEUE_ALLOC_FAIL = 7,
+ VERR_CR_CONN_LEN = 8,
+ VERR_CR_CONN_RQST_LEN = 9,
+ VERR_ASSOC_ID = 10,
+ VERR_ASSOC_ID_LEN = 11,
+ VERR_NO_ASSOC = 12,
+ VERR_CONN_ID = 13,
+ VERR_CONN_ID_LEN = 14,
+ VERR_NO_CONN = 15,
+ VERR_CR_CONN_CMD = 16,
+ VERR_CR_CONN_CMD_LEN = 17,
+ VERR_DISCONN_LEN = 18,
+ VERR_DISCONN_RQST_LEN = 19,
+ VERR_DISCONN_CMD = 20,
+ VERR_DISCONN_CMD_LEN = 21,
+ VERR_DISCONN_SCOPE = 22,
+ VERR_RS_LEN = 23,
+ VERR_RS_RQST_LEN = 24,
+ VERR_RS_CMD = 25,
+ VERR_RS_CMD_LEN = 26,
+ VERR_RS_RCTL = 27,
+ VERR_RS_RO = 28,
+};
+
+static char *validation_errors[] = {
+ "OK",
+ "Bad CR_ASSOC Length",
+ "Bad CR_ASSOC Rqst Length",
+ "Not CR_ASSOC Cmd",
+ "Bad CR_ASSOC Cmd Length",
+ "Bad Ersp Ratio",
+ "Association Allocation Failed",
+ "Queue Allocation Failed",
+ "Bad CR_CONN Length",
+ "Bad CR_CONN Rqst Length",
+ "Not Association ID",
+ "Bad Association ID Length",
+ "No Association",
+ "Not Connection ID",
+ "Bad Connection ID Length",
+ "No Connection",
+ "Not CR_CONN Cmd",
+ "Bad CR_CONN Cmd Length",
+ "Bad DISCONN Length",
+ "Bad DISCONN Rqst Length",
+ "Not DISCONN Cmd",
+ "Bad DISCONN Cmd Length",
+ "Bad Disconnect Scope",
+ "Bad RS Length",
+ "Bad RS Rqst Length",
+ "Not RS Cmd",
+ "Bad RS Cmd Length",
+ "Bad RS R_CTL",
+ "Bad RS Relative Offset",
+};
+
+static void
+nvmet_fc_ls_create_association(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ struct fcnvme_ls_cr_assoc_rqst *rqst =
+ (struct fcnvme_ls_cr_assoc_rqst *)iod->rqstbuf;
+ struct fcnvme_ls_cr_assoc_acc *acc =
+ (struct fcnvme_ls_cr_assoc_acc *)iod->rspbuf;
+ struct nvmet_fc_tgt_queue *queue;
+ int ret = 0;
+
+ memset(acc, 0, sizeof(*acc));
+
+ if (iod->rqstdatalen < sizeof(struct fcnvme_ls_cr_assoc_rqst))
+ ret = VERR_CR_ASSOC_LEN;
+ else if (rqst->desc_list_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_cr_assoc_rqst)))
+ ret = VERR_CR_ASSOC_RQST_LEN;
+ else if (rqst->assoc_cmd.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_CREATE_ASSOC_CMD))
+ ret = VERR_CR_ASSOC_CMD;
+ else if (rqst->assoc_cmd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_cr_assoc_cmd)))
+ ret = VERR_CR_ASSOC_CMD_LEN;
+ else if (!rqst->assoc_cmd.ersp_ratio ||
+ (be16_to_cpu(rqst->assoc_cmd.ersp_ratio) >=
+ be16_to_cpu(rqst->assoc_cmd.sqsize)))
+ ret = VERR_ERSP_RATIO;
+
+ else {
+ /* new association w/ admin queue */
+ iod->assoc = nvmet_fc_alloc_target_assoc(tgtport);
+ if (!iod->assoc)
+ ret = VERR_ASSOC_ALLOC_FAIL;
+ else {
+ queue = nvmet_fc_alloc_target_queue(iod->assoc, 0,
+ be16_to_cpu(rqst->assoc_cmd.sqsize));
+ if (!queue)
+ ret = VERR_QUEUE_ALLOC_FAIL;
+ }
+ }
+
+ if (ret) {
+ dev_err(tgtport->dev,
+ "Create Association LS failed: %s\n",
+ validation_errors[ret]);
+ iod->lsreq->rsplen = nvmet_fc_format_rjt(acc,
+ NVME_FC_MAX_LS_BUFFER_SIZE, rqst->w0.ls_cmd,
+ ELS_RJT_LOGIC,
+ ELS_EXPL_NONE, 0);
+ return;
+ }
+
+ queue->ersp_ratio = be16_to_cpu(rqst->assoc_cmd.ersp_ratio);
+ atomic_set(&queue->connected, 1);
+ queue->sqhd = 0; /* best place to init value */
+
+ /* format a response */
+
+ iod->lsreq->rsplen = sizeof(*acc);
+
+ nvmet_fc_format_rsp_hdr(acc, FCNVME_LS_ACC,
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_cr_assoc_acc)),
+ FCNVME_LS_CREATE_ASSOCIATION);
+ acc->associd.desc_tag = cpu_to_be32(FCNVME_LSDESC_ASSOC_ID);
+ acc->associd.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id));
+ acc->associd.association_id =
+ cpu_to_be64(nvmet_fc_makeconnid(iod->assoc, 0));
+ acc->connectid.desc_tag = cpu_to_be32(FCNVME_LSDESC_CONN_ID);
+ acc->connectid.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_conn_id));
+ acc->connectid.connection_id = acc->associd.association_id;
+}
+
+static void
+nvmet_fc_ls_create_connection(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ struct fcnvme_ls_cr_conn_rqst *rqst =
+ (struct fcnvme_ls_cr_conn_rqst *)iod->rqstbuf;
+ struct fcnvme_ls_cr_conn_acc *acc =
+ (struct fcnvme_ls_cr_conn_acc *)iod->rspbuf;
+ struct nvmet_fc_tgt_queue *queue;
+ int ret = 0;
+
+ memset(acc, 0, sizeof(*acc));
+
+ if (iod->rqstdatalen < sizeof(struct fcnvme_ls_cr_conn_rqst))
+ ret = VERR_CR_CONN_LEN;
+ else if (rqst->desc_list_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_cr_conn_rqst)))
+ ret = VERR_CR_CONN_RQST_LEN;
+ else if (rqst->associd.desc_tag != cpu_to_be32(FCNVME_LSDESC_ASSOC_ID))
+ ret = VERR_ASSOC_ID;
+ else if (rqst->associd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id)))
+ ret = VERR_ASSOC_ID_LEN;
+ else if (rqst->connect_cmd.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_CREATE_CONN_CMD))
+ ret = VERR_CR_CONN_CMD;
+ else if (rqst->connect_cmd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_cr_conn_cmd)))
+ ret = VERR_CR_CONN_CMD_LEN;
+ else if (!rqst->connect_cmd.ersp_ratio ||
+ (be16_to_cpu(rqst->connect_cmd.ersp_ratio) >=
+ be16_to_cpu(rqst->connect_cmd.sqsize)))
+ ret = VERR_ERSP_RATIO;
+
+ else {
+ /* new io queue */
+ iod->assoc = nvmet_fc_find_target_assoc(tgtport,
+ be64_to_cpu(rqst->associd.association_id));
+ if (!iod->assoc)
+ ret = VERR_NO_ASSOC;
+ else {
+ queue = nvmet_fc_alloc_target_queue(iod->assoc,
+ be16_to_cpu(rqst->connect_cmd.qid),
+ be16_to_cpu(rqst->connect_cmd.sqsize));
+ if (!queue)
+ ret = VERR_QUEUE_ALLOC_FAIL;
+
+ /* release get taken in nvmet_fc_find_target_assoc */
+ nvmet_fc_tgt_a_put(iod->assoc);
+ }
+ }
+
+ if (ret) {
+ dev_err(tgtport->dev,
+ "Create Connection LS failed: %s\n",
+ validation_errors[ret]);
+ iod->lsreq->rsplen = nvmet_fc_format_rjt(acc,
+ NVME_FC_MAX_LS_BUFFER_SIZE, rqst->w0.ls_cmd,
+ (ret == VERR_NO_ASSOC) ?
+ ELS_RJT_PROT : ELS_RJT_LOGIC,
+ ELS_EXPL_NONE, 0);
+ return;
+ }
+
+ queue->ersp_ratio = be16_to_cpu(rqst->connect_cmd.ersp_ratio);
+ atomic_set(&queue->connected, 1);
+ queue->sqhd = 0; /* best place to init value */
+
+ /* format a response */
+
+ iod->lsreq->rsplen = sizeof(*acc);
+
+ nvmet_fc_format_rsp_hdr(acc, FCNVME_LS_ACC,
+ fcnvme_lsdesc_len(sizeof(struct fcnvme_ls_cr_conn_acc)),
+ FCNVME_LS_CREATE_CONNECTION);
+ acc->connectid.desc_tag = cpu_to_be32(FCNVME_LSDESC_CONN_ID);
+ acc->connectid.desc_len =
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_conn_id));
+ acc->connectid.connection_id =
+ cpu_to_be64(nvmet_fc_makeconnid(iod->assoc,
+ be16_to_cpu(rqst->connect_cmd.qid)));
+}
+
+static void
+nvmet_fc_ls_disconnect(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ struct fcnvme_ls_disconnect_rqst *rqst =
+ (struct fcnvme_ls_disconnect_rqst *)iod->rqstbuf;
+ struct fcnvme_ls_disconnect_acc *acc =
+ (struct fcnvme_ls_disconnect_acc *)iod->rspbuf;
+ struct nvmet_fc_tgt_queue *queue;
+ struct nvmet_fc_tgt_assoc *assoc;
+ int ret = 0;
+ bool del_assoc = false;
+
+ memset(acc, 0, sizeof(*acc));
+
+ if (iod->rqstdatalen < sizeof(struct fcnvme_ls_disconnect_rqst))
+ ret = VERR_DISCONN_LEN;
+ else if (rqst->desc_list_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_disconnect_rqst)))
+ ret = VERR_DISCONN_RQST_LEN;
+ else if (rqst->associd.desc_tag != cpu_to_be32(FCNVME_LSDESC_ASSOC_ID))
+ ret = VERR_ASSOC_ID;
+ else if (rqst->associd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_assoc_id)))
+ ret = VERR_ASSOC_ID_LEN;
+ else if (rqst->discon_cmd.desc_tag !=
+ cpu_to_be32(FCNVME_LSDESC_DISCONN_CMD))
+ ret = VERR_DISCONN_CMD;
+ else if (rqst->discon_cmd.desc_len !=
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_lsdesc_disconn_cmd)))
+ ret = VERR_DISCONN_CMD_LEN;
+ else if ((rqst->discon_cmd.scope != FCNVME_DISCONN_ASSOCIATION) &&
+ (rqst->discon_cmd.scope != FCNVME_DISCONN_CONNECTION))
+ ret = VERR_DISCONN_SCOPE;
+ else {
+ /* match an active association */
+ assoc = nvmet_fc_find_target_assoc(tgtport,
+ be64_to_cpu(rqst->associd.association_id));
+ iod->assoc = assoc;
+ if (!assoc)
+ ret = VERR_NO_ASSOC;
+ }
+
+ if (ret) {
+ dev_err(tgtport->dev,
+ "Disconnect LS failed: %s\n",
+ validation_errors[ret]);
+ iod->lsreq->rsplen = nvmet_fc_format_rjt(acc,
+ NVME_FC_MAX_LS_BUFFER_SIZE, rqst->w0.ls_cmd,
+ (ret == 8) ? ELS_RJT_PROT : ELS_RJT_LOGIC,
+ ELS_EXPL_NONE, 0);
+ return;
+ }
+
+ /* format a response */
+
+ iod->lsreq->rsplen = sizeof(*acc);
+
+ nvmet_fc_format_rsp_hdr(acc, FCNVME_LS_ACC,
+ fcnvme_lsdesc_len(
+ sizeof(struct fcnvme_ls_disconnect_acc)),
+ FCNVME_LS_DISCONNECT);
+
+
+ if (rqst->discon_cmd.scope == FCNVME_DISCONN_CONNECTION) {
+ queue = nvmet_fc_find_target_queue(tgtport,
+ be64_to_cpu(rqst->discon_cmd.id));
+ if (queue) {
+ int qid = queue->qid;
+
+ nvmet_fc_delete_target_queue(queue);
+
+ /* release the get taken by find_target_queue */
+ nvmet_fc_tgt_q_put(queue);
+
+ /* tear association down if io queue terminated */
+ if (!qid)
+ del_assoc = true;
+ }
+ }
+
+ /* release get taken in nvmet_fc_find_target_assoc */
+ nvmet_fc_tgt_a_put(iod->assoc);
+
+ if (del_assoc)
+ nvmet_fc_delete_target_assoc(iod->assoc);
+}
+
+
+/* *********************** NVME Ctrl Routines **************************** */
+
+
+static void nvmet_fc_fcp_nvme_cmd_done(struct nvmet_req *nvme_req);
+
+static struct nvmet_fabrics_ops nvmet_fc_tgt_fcp_ops;
+
+static void
+nvmet_fc_xmt_ls_rsp_done(struct nvmefc_tgt_ls_req *lsreq)
+{
+ struct nvmet_fc_ls_iod *iod = lsreq->nvmet_fc_private;
+ struct nvmet_fc_tgtport *tgtport = iod->tgtport;
+
+ fc_dma_sync_single_for_cpu(tgtport->dev, iod->rspdma,
+ NVME_FC_MAX_LS_BUFFER_SIZE, DMA_TO_DEVICE);
+ nvmet_fc_free_ls_iod(tgtport, iod);
+ nvmet_fc_tgtport_put(tgtport);
+}
+
+static void
+nvmet_fc_xmt_ls_rsp(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ int ret;
+
+ fc_dma_sync_single_for_device(tgtport->dev, iod->rspdma,
+ NVME_FC_MAX_LS_BUFFER_SIZE, DMA_TO_DEVICE);
+
+ ret = tgtport->ops->xmt_ls_rsp(&tgtport->fc_target_port, iod->lsreq);
+ if (ret)
+ nvmet_fc_xmt_ls_rsp_done(iod->lsreq);
+}
+
+/*
+ * Actual processing routine for received FC-NVME LS Requests from the LLD
+ */
+static void
+nvmet_fc_handle_ls_rqst(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_ls_iod *iod)
+{
+ struct fcnvme_ls_rqst_w0 *w0 =
+ (struct fcnvme_ls_rqst_w0 *)iod->rqstbuf;
+
+ iod->lsreq->nvmet_fc_private = iod;
+ iod->lsreq->rspbuf = iod->rspbuf;
+ iod->lsreq->rspdma = iod->rspdma;
+ iod->lsreq->done = nvmet_fc_xmt_ls_rsp_done;
+ /* Be preventative. handlers will later set to valid length */
+ iod->lsreq->rsplen = 0;
+
+ iod->assoc = NULL;
+
+ /*
+ * handlers:
+ * parse request input, execute the request, and format the
+ * LS response
+ */
+ switch (w0->ls_cmd) {
+ case FCNVME_LS_CREATE_ASSOCIATION:
+ /* Creates Association and initial Admin Queue/Connection */
+ nvmet_fc_ls_create_association(tgtport, iod);
+ break;
+ case FCNVME_LS_CREATE_CONNECTION:
+ /* Creates an IO Queue/Connection */
+ nvmet_fc_ls_create_connection(tgtport, iod);
+ break;
+ case FCNVME_LS_DISCONNECT:
+ /* Terminate a Queue/Connection or the Association */
+ nvmet_fc_ls_disconnect(tgtport, iod);
+ break;
+ default:
+ iod->lsreq->rsplen = nvmet_fc_format_rjt(iod->rspbuf,
+ NVME_FC_MAX_LS_BUFFER_SIZE, w0->ls_cmd,
+ ELS_RJT_INVAL, ELS_EXPL_NONE, 0);
+ }
+
+ nvmet_fc_xmt_ls_rsp(tgtport, iod);
+}
+
+/*
+ * Actual processing routine for received FC-NVME LS Requests from the LLD
+ */
+static void
+nvmet_fc_handle_ls_rqst_work(struct work_struct *work)
+{
+ struct nvmet_fc_ls_iod *iod =
+ container_of(work, struct nvmet_fc_ls_iod, work);
+ struct nvmet_fc_tgtport *tgtport = iod->tgtport;
+
+ nvmet_fc_handle_ls_rqst(tgtport, iod);
+}
+
+
+/**
+ * nvmet_fc_rcv_ls_req - transport entry point called by an LLDD
+ * upon the reception of a NVME LS request.
+ *
+ * The nvmet-fc layer will copy payload to an internal structure for
+ * processing. As such, upon completion of the routine, the LLDD may
+ * immediately free/reuse the LS request buffer passed in the call.
+ *
+ * If this routine returns error, the LLDD should abort the exchange.
+ *
+ * @tgtport: pointer to the (registered) target port the LS was
+ * received on.
+ * @lsreq: pointer to a lsreq request structure to be used to reference
+ * the exchange corresponding to the LS.
+ * @lsreqbuf: pointer to the buffer containing the LS Request
+ * @lsreqbuf_len: length, in bytes, of the received LS request
+ */
+int
+nvmet_fc_rcv_ls_req(struct nvmet_fc_target_port *target_port,
+ struct nvmefc_tgt_ls_req *lsreq,
+ void *lsreqbuf, u32 lsreqbuf_len)
+{
+ struct nvmet_fc_tgtport *tgtport = targetport_to_tgtport(target_port);
+ struct nvmet_fc_ls_iod *iod;
+
+ if (lsreqbuf_len > NVME_FC_MAX_LS_BUFFER_SIZE)
+ return -E2BIG;
+
+ if (!nvmet_fc_tgtport_get(tgtport))
+ return -ESHUTDOWN;
+
+ iod = nvmet_fc_alloc_ls_iod(tgtport);
+ if (!iod) {
+ nvmet_fc_tgtport_put(tgtport);
+ return -ENOENT;
+ }
+
+ iod->lsreq = lsreq;
+ iod->fcpreq = NULL;
+ memcpy(iod->rqstbuf, lsreqbuf, lsreqbuf_len);
+ iod->rqstdatalen = lsreqbuf_len;
+
+ schedule_work(&iod->work);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(nvmet_fc_rcv_ls_req);
+
+
+/*
+ * **********************
+ * Start of FCP handling
+ * **********************
+ */
+
+static int
+nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
+{
+ struct scatterlist *sg;
+ struct page *page;
+ unsigned int nent;
+ u32 page_len, length;
+ int i = 0;
+
+ length = fod->total_length;
+ nent = DIV_ROUND_UP(length, PAGE_SIZE);
+ sg = kmalloc_array(nent, sizeof(struct scatterlist), GFP_KERNEL);
+ if (!sg)
+ goto out;
+
+ sg_init_table(sg, nent);
+
+ while (length) {
+ page_len = min_t(u32, length, PAGE_SIZE);
+
+ page = alloc_page(GFP_KERNEL);
+ if (!page)
+ goto out_free_pages;
+
+ sg_set_page(&sg[i], page, page_len, 0);
+ length -= page_len;
+ i++;
+ }
+
+ fod->data_sg = sg;
+ fod->data_sg_cnt = nent;
+ fod->data_sg_cnt = fc_dma_map_sg(fod->tgtport->dev, sg, nent,
+ ((fod->io_dir == NVMET_FCP_WRITE) ?
+ DMA_FROM_DEVICE : DMA_TO_DEVICE));
+ /* note: write from initiator perspective */
+
+ return 0;
+
+out_free_pages:
+ while (i > 0) {
+ i--;
+ __free_page(sg_page(&sg[i]));
+ }
+ kfree(sg);
+ fod->data_sg = NULL;
+ fod->data_sg_cnt = 0;
+out:
+ return NVME_SC_INTERNAL;
+}
+
+static void
+nvmet_fc_free_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
+{
+ struct scatterlist *sg;
+ int count;
+
+ if (!fod->data_sg || !fod->data_sg_cnt)
+ return;
+
+ fc_dma_unmap_sg(fod->tgtport->dev, fod->data_sg, fod->data_sg_cnt,
+ ((fod->io_dir == NVMET_FCP_WRITE) ?
+ DMA_FROM_DEVICE : DMA_TO_DEVICE));
+ for_each_sg(fod->data_sg, sg, fod->data_sg_cnt, count)
+ __free_page(sg_page(sg));
+ kfree(fod->data_sg);
+}
+
+
+static bool
+queue_90percent_full(struct nvmet_fc_tgt_queue *q, u32 sqhd)
+{
+ u32 sqtail, used;
+
+ /* egad, this is ugly. And sqtail is just a best guess */
+ sqtail = atomic_read(&q->sqtail) % q->sqsize;
+
+ used = (sqtail < sqhd) ? (sqtail + q->sqsize - sqhd) : (sqtail - sqhd);
+ return ((used * 10) >= (((u32)(q->sqsize - 1) * 9)));
+}
+
+/*
+ * Prep RSP payload.
+ * May be a NVMET_FCOP_RSP or NVMET_FCOP_READDATA_RSP op
+ */
+static void
+nvmet_fc_prep_fcp_rsp(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_fcp_iod *fod)
+{
+ struct nvme_fc_ersp_iu *ersp = &fod->rspiubuf;
+ struct nvme_common_command *sqe = &fod->cmdiubuf.sqe.common;
+ struct nvme_completion *cqe = &ersp->cqe;
+ u32 *cqewd = (u32 *)cqe;
+ bool send_ersp = false;
+ u32 rsn, rspcnt, xfr_length;
+
+ if (fod->fcpreq->op == NVMET_FCOP_READDATA_RSP)
+ xfr_length = fod->total_length;
+ else
+ xfr_length = fod->offset;
+
+ /*
+ * check to see if we can send a 0's rsp.
+ * Note: to send a 0's response, the NVME-FC host transport will
+ * recreate the CQE. The host transport knows: sq id, SQHD (last
+ * seen in an ersp), and command_id. Thus it will create a
+ * zero-filled CQE with those known fields filled in. Transport
+ * must send an ersp for any condition where the cqe won't match
+ * this.
+ *
+ * Here are the FC-NVME mandated cases where we must send an ersp:
+ * every N responses, where N=ersp_ratio
+ * force fabric commands to send ersp's (not in FC-NVME but good
+ * practice)
+ * normal cmds: any time status is non-zero, or status is zero
+ * but words 0 or 1 are non-zero.
+ * the SQ is 90% or more full
+ * the cmd is a fused command
+ * transferred data length not equal to cmd iu length
+ */
+ rspcnt = atomic_inc_return(&fod->queue->zrspcnt);
+ if (!(rspcnt % fod->queue->ersp_ratio) ||
+ sqe->opcode == nvme_fabrics_command ||
+ 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))
+ send_ersp = true;
+
+ /* re-set the fields */
+ fod->fcpreq->rspaddr = ersp;
+ fod->fcpreq->rspdma = fod->rspdma;
+
+ if (!send_ersp) {
+ memset(ersp, 0, NVME_FC_SIZEOF_ZEROS_RSP);
+ fod->fcpreq->rsplen = NVME_FC_SIZEOF_ZEROS_RSP;
+ } else {
+ ersp->iu_len = cpu_to_be16(sizeof(*ersp)/sizeof(u32));
+ rsn = atomic_inc_return(&fod->queue->rsn);
+ ersp->rsn = cpu_to_be32(rsn);
+ ersp->xfrd_len = cpu_to_be32(xfr_length);
+ fod->fcpreq->rsplen = sizeof(*ersp);
+ }
+
+ fc_dma_sync_single_for_device(tgtport->dev, fod->rspdma,
+ sizeof(fod->rspiubuf), DMA_TO_DEVICE);
+}
+
+static void nvmet_fc_xmt_fcp_op_done(struct nvmefc_tgt_fcp_req *fcpreq);
+
+static void
+nvmet_fc_xmt_fcp_rsp(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_fcp_iod *fod)
+{
+ int ret;
+
+ fod->fcpreq->op = NVMET_FCOP_RSP;
+ fod->fcpreq->timeout = 0;
+
+ nvmet_fc_prep_fcp_rsp(tgtport, fod);
+
+ ret = tgtport->ops->fcp_op(&tgtport->fc_target_port, fod->fcpreq);
+ if (ret)
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+}
+
+static void
+nvmet_fc_transfer_fcp_data(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_fcp_iod *fod, u8 op)
+{
+ struct nvmefc_tgt_fcp_req *fcpreq = fod->fcpreq;
+ struct scatterlist *sg, *datasg;
+ u32 tlen, sg_off;
+ int ret;
+
+ fcpreq->op = op;
+ fcpreq->offset = fod->offset;
+ fcpreq->timeout = NVME_FC_TGTOP_TIMEOUT_SEC;
+ tlen = min_t(u32, (NVMET_FC_MAX_KB_PER_XFR * 1024),
+ (fod->total_length - fod->offset));
+ tlen = min_t(u32, tlen, NVME_FC_MAX_SEGMENTS * PAGE_SIZE);
+ tlen = min_t(u32, tlen, fod->tgtport->ops->max_sgl_segments
+ * PAGE_SIZE);
+ fcpreq->transfer_length = tlen;
+ fcpreq->transferred_length = 0;
+ fcpreq->fcp_error = 0;
+ fcpreq->rsplen = 0;
+
+ fcpreq->sg_cnt = 0;
+
+ datasg = fod->next_sg;
+ sg_off = fod->next_sg_offset;
+
+ for (sg = fcpreq->sg ; tlen; sg++) {
+ *sg = *datasg;
+ if (sg_off) {
+ sg->offset += sg_off;
+ sg->length -= sg_off;
+ sg->dma_address += sg_off;
+ sg_off = 0;
+ }
+ if (tlen < sg->length) {
+ sg->length = tlen;
+ fod->next_sg = datasg;
+ fod->next_sg_offset += tlen;
+ } else if (tlen == sg->length) {
+ fod->next_sg_offset = 0;
+ fod->next_sg = sg_next(datasg);
+ } else {
+ fod->next_sg_offset = 0;
+ datasg = sg_next(datasg);
+ }
+ tlen -= sg->length;
+ fcpreq->sg_cnt++;
+ }
+
+ /*
+ * If the last READDATA request: check if LLDD supports
+ * combined xfr with response.
+ */
+ if ((op == NVMET_FCOP_READDATA) &&
+ ((fod->offset + fcpreq->transfer_length) == fod->total_length) &&
+ (tgtport->ops->target_features & NVMET_FCTGTFEAT_READDATA_RSP)) {
+ fcpreq->op = NVMET_FCOP_READDATA_RSP;
+ nvmet_fc_prep_fcp_rsp(tgtport, fod);
+ }
+
+ ret = tgtport->ops->fcp_op(&tgtport->fc_target_port, fod->fcpreq);
+ if (ret) {
+ /*
+ * should be ok to set w/o lock as its in the thread of
+ * execution (not an async timer routine) and doesn't
+ * contend with any clearing action
+ */
+ fod->abort = true;
+
+ if (op == NVMET_FCOP_WRITEDATA)
+ nvmet_req_complete(&fod->req,
+ NVME_SC_FC_TRANSPORT_ERROR);
+ else /* NVMET_FCOP_READDATA or NVMET_FCOP_READDATA_RSP */ {
+ fcpreq->fcp_error = ret;
+ fcpreq->transferred_length = 0;
+ nvmet_fc_xmt_fcp_op_done(fod->fcpreq);
+ }
+ }
+}
+
+static void
+nvmet_fc_xmt_fcp_op_done(struct nvmefc_tgt_fcp_req *fcpreq)
+{
+ struct nvmet_fc_fcp_iod *fod = fcpreq->nvmet_fc_private;
+ struct nvmet_fc_tgtport *tgtport = fod->tgtport;
+ unsigned long flags;
+ bool abort;
+
+ spin_lock_irqsave(&fod->flock, flags);
+ abort = fod->abort;
+ spin_unlock_irqrestore(&fod->flock, flags);
+
+ /* if in the middle of an io and we need to tear down */
+ if (abort && fcpreq->op != NVMET_FCOP_ABORT) {
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+
+ if (fcpreq->fcp_error || abort)
+ nvmet_req_complete(&fod->req, fcpreq->fcp_error);
+
+ return;
+ }
+
+ switch (fcpreq->op) {
+
+ case NVMET_FCOP_WRITEDATA:
+ if (abort || fcpreq->fcp_error ||
+ fcpreq->transferred_length != fcpreq->transfer_length) {
+ nvmet_req_complete(&fod->req,
+ NVME_SC_FC_TRANSPORT_ERROR);
+ return;
+ }
+
+ fod->offset += fcpreq->transferred_length;
+ if (fod->offset != fod->total_length) {
+ /* transfer the next chunk */
+ nvmet_fc_transfer_fcp_data(tgtport, fod,
+ NVMET_FCOP_WRITEDATA);
+ return;
+ }
+
+ /* data transfer complete, resume with nvmet layer */
+
+ fod->req.execute(&fod->req);
+
+ break;
+
+ case NVMET_FCOP_READDATA:
+ case NVMET_FCOP_READDATA_RSP:
+ if (abort || fcpreq->fcp_error ||
+ fcpreq->transferred_length != fcpreq->transfer_length) {
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+ return;
+ }
+
+ /* success */
+
+ if (fcpreq->op == NVMET_FCOP_READDATA_RSP) {
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+ fc_dma_sync_single_for_cpu(tgtport->dev, fod->rspdma,
+ sizeof(fod->rspiubuf), DMA_TO_DEVICE);
+ nvmet_fc_free_fcp_iod(fod->queue, fod);
+ return;
+ }
+
+ fod->offset += fcpreq->transferred_length;
+ if (fod->offset != fod->total_length) {
+ /* transfer the next chunk */
+ nvmet_fc_transfer_fcp_data(tgtport, fod,
+ NVMET_FCOP_READDATA);
+ return;
+ }
+
+ /* data transfer complete, send response */
+
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+
+ nvmet_fc_xmt_fcp_rsp(tgtport, fod);
+
+ break;
+
+ case NVMET_FCOP_RSP:
+ case NVMET_FCOP_ABORT:
+ fc_dma_sync_single_for_cpu(tgtport->dev, fod->rspdma,
+ sizeof(fod->rspiubuf), DMA_TO_DEVICE);
+ nvmet_fc_free_fcp_iod(fod->queue, fod);
+ break;
+
+ default:
+ nvmet_fc_free_tgt_pgs(fod);
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+ break;
+ }
+}
+
+/*
+ * actual completion handler after execution by the nvmet layer
+ */
+static void
+__nvmet_fc_fcp_nvme_cmd_done(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_fcp_iod *fod, int status)
+{
+ struct nvme_common_command *sqe = &fod->cmdiubuf.sqe.common;
+ struct nvme_completion *cqe = &fod->rspiubuf.cqe;
+ unsigned long flags;
+ bool abort;
+
+ spin_lock_irqsave(&fod->flock, flags);
+ abort = fod->abort;
+ spin_unlock_irqrestore(&fod->flock, flags);
+
+ /* if we have a CQE, snoop the last sq_head value */
+ if (!status)
+ fod->queue->sqhd = cqe->sq_head;
+
+ if (abort) {
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+ return;
+ }
+
+ /* if an error handling the cmd post initial parsing */
+ if (status) {
+ /* fudge up a failed CQE status for our transport error */
+ memset(cqe, 0, sizeof(*cqe));
+ cqe->sq_head = fod->queue->sqhd; /* echo last cqe sqhd */
+ cqe->sq_id = cpu_to_le16(fod->queue->qid);
+ cqe->command_id = sqe->command_id;
+ cqe->status = cpu_to_le16(status);
+ } else {
+
+ /*
+ * try to push the data even if the SQE status is non-zero.
+ * There may be a status where data still was intended to
+ * be moved
+ */
+ if ((fod->io_dir == NVMET_FCP_READ) && (fod->data_sg_cnt)) {
+ /* push the data over before sending rsp */
+ nvmet_fc_transfer_fcp_data(tgtport, fod,
+ NVMET_FCOP_READDATA);
+ return;
+ }
+
+ /* writes & no data - fall thru */
+ }
+
+ /* data no longer needed */
+ nvmet_fc_free_tgt_pgs(fod);
+
+ nvmet_fc_xmt_fcp_rsp(tgtport, fod);
+}
+
+
+static void
+nvmet_fc_fcp_nvme_cmd_done(struct nvmet_req *nvme_req)
+{
+ struct nvmet_fc_fcp_iod *fod = nvmet_req_to_fod(nvme_req);
+ struct nvmet_fc_tgtport *tgtport = fod->tgtport;
+
+ __nvmet_fc_fcp_nvme_cmd_done(tgtport, fod, 0);
+}
+
+
+/*
+ * Actual processing routine for received FC-NVME LS Requests from the LLD
+ */
+void
+nvmet_fc_handle_fcp_rqst(struct nvmet_fc_tgtport *tgtport,
+ struct nvmet_fc_fcp_iod *fod)
+{
+ struct nvme_fc_cmd_iu *cmdiu = &fod->cmdiubuf;
+ int ret;
+
+ /*
+ * Fused commands are currently not supported in the linux
+ * implementation.
+ *
+ * As such, the implementation of the FC transport does not
+ * look at the fused commands and order delivery to the upper
+ * layer until we have both based on csn.
+ */
+
+ fod->fcpreq->done = nvmet_fc_xmt_fcp_op_done;
+
+ fod->total_length = be32_to_cpu(cmdiu->data_len);
+ if (cmdiu->flags & FCNVME_CMD_FLAGS_WRITE) {
+ fod->io_dir = NVMET_FCP_WRITE;
+ if (!nvme_is_write(&cmdiu->sqe))
+ goto transport_error;
+ } else if (cmdiu->flags & FCNVME_CMD_FLAGS_READ) {
+ fod->io_dir = NVMET_FCP_READ;
+ if (nvme_is_write(&cmdiu->sqe))
+ goto transport_error;
+ } else {
+ fod->io_dir = NVMET_FCP_NODATA;
+ if (fod->total_length)
+ goto transport_error;
+ }
+
+ fod->req.cmd = &fod->cmdiubuf.sqe;
+ fod->req.rsp = &fod->rspiubuf.cqe;
+ fod->req.port = fod->queue->port;
+
+ /* ensure nvmet handlers will set cmd handler callback */
+ fod->req.execute = NULL;
+
+ /* clear any response payload */
+ memset(&fod->rspiubuf, 0, sizeof(fod->rspiubuf));
+
+ ret = nvmet_req_init(&fod->req,
+ &fod->queue->nvme_cq,
+ &fod->queue->nvme_sq,
+ &nvmet_fc_tgt_fcp_ops);
+ if (!ret) { /* bad SQE content */
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+ return;
+ }
+
+ /* keep a running counter of tail position */
+ atomic_inc(&fod->queue->sqtail);
+
+ fod->data_sg = NULL;
+ fod->data_sg_cnt = 0;
+ if (fod->total_length) {
+ ret = nvmet_fc_alloc_tgt_pgs(fod);
+ if (ret) {
+ nvmet_req_complete(&fod->req, ret);
+ return;
+ }
+ }
+ fod->req.sg = fod->data_sg;
+ fod->req.sg_cnt = fod->data_sg_cnt;
+ fod->offset = 0;
+ fod->next_sg = fod->data_sg;
+ fod->next_sg_offset = 0;
+
+ if (fod->io_dir == NVMET_FCP_WRITE) {
+ /* pull the data over before invoking nvmet layer */
+ nvmet_fc_transfer_fcp_data(tgtport, fod, NVMET_FCOP_WRITEDATA);
+ return;
+ }
+
+ /*
+ * Reads or no data:
+ *
+ * can invoke the nvmet_layer now. If read data, cmd completion will
+ * push the data
+ */
+
+ fod->req.execute(&fod->req);
+
+ return;
+
+transport_error:
+ nvmet_fc_abort_op(tgtport, fod->fcpreq);
+}
+
+/*
+ * Actual processing routine for received FC-NVME LS Requests from the LLD
+ */
+static void
+nvmet_fc_handle_fcp_rqst_work(struct work_struct *work)
+{
+ struct nvmet_fc_fcp_iod *fod =
+ container_of(work, struct nvmet_fc_fcp_iod, work);
+ struct nvmet_fc_tgtport *tgtport = fod->tgtport;
+
+ nvmet_fc_handle_fcp_rqst(tgtport, fod);
+}
+
+/**
+ * nvmet_fc_rcv_fcp_req - transport entry point called by an LLDD
+ * upon the reception of a NVME FCP CMD IU.
+ *
+ * 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.
+ *
+ * If this routine returns error, the lldd should abort the exchange.
+ *
+ * @target_port: pointer to the (registered) target port the FCP CMD IU
+ * was receive on.
+ * @fcpreq: pointer to a fcpreq request structure to be used to reference
+ * the exchange corresponding to the FCP Exchange.
+ * @cmdiubuf: pointer to the buffer containing the FCP CMD IU
+ * @cmdiubuf_len: length, in bytes, of the received FCP CMD IU
+ */
+int
+nvmet_fc_rcv_fcp_req(struct nvmet_fc_target_port *target_port,
+ struct nvmefc_tgt_fcp_req *fcpreq,
+ void *cmdiubuf, u32 cmdiubuf_len)
+{
+ struct nvmet_fc_tgtport *tgtport = targetport_to_tgtport(target_port);
+ struct nvme_fc_cmd_iu *cmdiu = cmdiubuf;
+ struct nvmet_fc_tgt_queue *queue;
+ struct nvmet_fc_fcp_iod *fod;
+
+ /* validate iu, so the connection id can be used to find the queue */
+ if ((cmdiubuf_len != sizeof(*cmdiu)) ||
+ (cmdiu->scsi_id != NVME_CMD_SCSI_ID) ||
+ (cmdiu->fc_id != NVME_CMD_FC_ID) ||
+ (be16_to_cpu(cmdiu->iu_len) != (sizeof(*cmdiu)/4)))
+ return -EIO;
+
+
+ queue = nvmet_fc_find_target_queue(tgtport,
+ be64_to_cpu(cmdiu->connection_id));
+ if (!queue)
+ return -ENOTCONN;
+
+ /*
+ * note: reference taken by find_target_queue
+ * After successful fod allocation, the fod will inherit the
+ * ownership of that reference and will remove the reference
+ * when the fod is freed.
+ */
+
+ fod = nvmet_fc_alloc_fcp_iod(queue);
+ if (!fod) {
+ /* 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);
+
+ queue_work_on(queue->cpu, queue->work_q, &fod->work);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(nvmet_fc_rcv_fcp_req);
+
+enum {
+ FCT_TRADDR_ERR = 0,
+ FCT_TRADDR_WWNN = 1 << 0,
+ FCT_TRADDR_WWPN = 1 << 1,
+};
+
+struct nvmet_fc_traddr {
+ u64 nn;
+ u64 pn;
+};
+
+static const match_table_t traddr_opt_tokens = {
+ { FCT_TRADDR_WWNN, "nn-%s" },
+ { FCT_TRADDR_WWPN, "pn-%s" },
+ { FCT_TRADDR_ERR, NULL }
+};
+
+static int
+nvmet_fc_parse_traddr(struct nvmet_fc_traddr *traddr, char *buf)
+{
+ substring_t args[MAX_OPT_ARGS];
+ char *options, *o, *p;
+ int token, ret = 0;
+ u64 token64;
+
+ options = o = kstrdup(buf, GFP_KERNEL);
+ if (!options)
+ return -ENOMEM;
+
+ while ((p = strsep(&o, ",\n")) != NULL) {
+ if (!*p)
+ continue;
+
+ token = match_token(p, traddr_opt_tokens, args);
+ switch (token) {
+ case FCT_TRADDR_WWNN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out;
+ }
+ traddr->nn = token64;
+ break;
+ case FCT_TRADDR_WWPN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out;
+ }
+ traddr->pn = token64;
+ break;
+ default:
+ pr_warn("unknown traddr token or missing value '%s'\n",
+ p);
+ ret = -EINVAL;
+ goto out;
+ }
+ }
+
+out:
+ kfree(options);
+ return ret;
+}
+
+static int
+nvmet_fc_add_port(struct nvmet_port *port)
+{
+ struct nvmet_fc_tgtport *tgtport;
+ struct nvmet_fc_traddr traddr = { 0L, 0L };
+ unsigned long flags;
+ int ret;
+
+ /* validate the address info */
+ if ((port->disc_addr.trtype != NVMF_TRTYPE_FC) ||
+ (port->disc_addr.adrfam != NVMF_ADDR_FAMILY_FC))
+ return -EINVAL;
+
+ /* map the traddr address info to a target port */
+
+ ret = nvmet_fc_parse_traddr(&traddr, port->disc_addr.traddr);
+ if (ret)
+ return ret;
+
+ ret = -ENXIO;
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ list_for_each_entry(tgtport, &nvmet_fc_target_list, tgt_list) {
+ if ((tgtport->fc_target_port.node_name == traddr.nn) &&
+ (tgtport->fc_target_port.port_name == traddr.pn)) {
+ /* a FC port can only be 1 nvmet port id */
+ if (!tgtport->port) {
+ tgtport->port = port;
+ port->priv = tgtport;
+ ret = 0;
+ } else
+ ret = -EALREADY;
+ break;
+ }
+ }
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+ return ret;
+}
+
+static void
+nvmet_fc_remove_port(struct nvmet_port *port)
+{
+ struct nvmet_fc_tgtport *tgtport = port->priv;
+ unsigned long flags;
+
+ spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
+ if (tgtport->port == port) {
+ nvmet_fc_tgtport_put(tgtport);
+ tgtport->port = NULL;
+ }
+ spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
+}
+
+static struct nvmet_fabrics_ops nvmet_fc_tgt_fcp_ops = {
+ .owner = THIS_MODULE,
+ .type = NVMF_TRTYPE_FC,
+ .msdbd = 1,
+ .add_port = nvmet_fc_add_port,
+ .remove_port = nvmet_fc_remove_port,
+ .queue_response = nvmet_fc_fcp_nvme_cmd_done,
+ .delete_ctrl = nvmet_fc_delete_ctrl,
+};
+
+static int __init nvmet_fc_init_module(void)
+{
+ return nvmet_register_transport(&nvmet_fc_tgt_fcp_ops);
+}
+
+static void __exit nvmet_fc_exit_module(void)
+{
+ /* sanity check - all lports should be removed */
+ if (!list_empty(&nvmet_fc_target_list))
+ pr_warn("%s: targetport list not empty\n", __func__);
+
+ nvmet_unregister_transport(&nvmet_fc_tgt_fcp_ops);
+
+ ida_destroy(&nvmet_fc_tgtport_cnt);
+}
+
+module_init(nvmet_fc_init_module);
+module_exit(nvmet_fc_exit_module);
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/nvme/target/fcloop.c b/drivers/nvme/target/fcloop.c
new file mode 100644
index 000000000000..bcb8ebeb01c5
--- /dev/null
+++ b/drivers/nvme/target/fcloop.c
@@ -0,0 +1,1148 @@
+/*
+ * Copyright (c) 2016 Avago Technologies. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful.
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND WARRANTIES,
+ * INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A
+ * PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE DISCLAIMED, EXCEPT TO
+ * THE EXTENT THAT SUCH DISCLAIMERS ARE HELD TO BE LEGALLY INVALID.
+ * See the GNU General Public License for more details, a copy of which
+ * can be found in the file COPYING included with this package
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/module.h>
+#include <linux/parser.h>
+#include <uapi/scsi/fc/fc_fs.h>
+
+#include "../host/nvme.h"
+#include "../target/nvmet.h"
+#include <linux/nvme-fc-driver.h>
+#include <linux/nvme-fc.h>
+
+
+enum {
+ NVMF_OPT_ERR = 0,
+ NVMF_OPT_WWNN = 1 << 0,
+ NVMF_OPT_WWPN = 1 << 1,
+ NVMF_OPT_ROLES = 1 << 2,
+ NVMF_OPT_FCADDR = 1 << 3,
+ NVMF_OPT_LPWWNN = 1 << 4,
+ NVMF_OPT_LPWWPN = 1 << 5,
+};
+
+struct fcloop_ctrl_options {
+ int mask;
+ u64 wwnn;
+ u64 wwpn;
+ u32 roles;
+ u32 fcaddr;
+ u64 lpwwnn;
+ u64 lpwwpn;
+};
+
+static const match_table_t opt_tokens = {
+ { NVMF_OPT_WWNN, "wwnn=%s" },
+ { NVMF_OPT_WWPN, "wwpn=%s" },
+ { NVMF_OPT_ROLES, "roles=%d" },
+ { NVMF_OPT_FCADDR, "fcaddr=%x" },
+ { NVMF_OPT_LPWWNN, "lpwwnn=%s" },
+ { NVMF_OPT_LPWWPN, "lpwwpn=%s" },
+ { NVMF_OPT_ERR, NULL }
+};
+
+static int
+fcloop_parse_options(struct fcloop_ctrl_options *opts,
+ const char *buf)
+{
+ substring_t args[MAX_OPT_ARGS];
+ char *options, *o, *p;
+ int token, ret = 0;
+ u64 token64;
+
+ options = o = kstrdup(buf, GFP_KERNEL);
+ if (!options)
+ return -ENOMEM;
+
+ while ((p = strsep(&o, ",\n")) != NULL) {
+ if (!*p)
+ continue;
+
+ token = match_token(p, opt_tokens, args);
+ opts->mask |= token;
+ switch (token) {
+ case NVMF_OPT_WWNN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->wwnn = token64;
+ break;
+ case NVMF_OPT_WWPN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->wwpn = token64;
+ break;
+ case NVMF_OPT_ROLES:
+ if (match_int(args, &token)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->roles = token;
+ break;
+ case NVMF_OPT_FCADDR:
+ if (match_hex(args, &token)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->fcaddr = token;
+ break;
+ case NVMF_OPT_LPWWNN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->lpwwnn = token64;
+ break;
+ case NVMF_OPT_LPWWPN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ opts->lpwwpn = token64;
+ break;
+ default:
+ pr_warn("unknown parameter or missing value '%s'\n", p);
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ }
+
+out_free_options:
+ kfree(options);
+ return ret;
+}
+
+
+static int
+fcloop_parse_nm_options(struct device *dev, u64 *nname, u64 *pname,
+ const char *buf)
+{
+ substring_t args[MAX_OPT_ARGS];
+ char *options, *o, *p;
+ int token, ret = 0;
+ u64 token64;
+
+ *nname = -1;
+ *pname = -1;
+
+ options = o = kstrdup(buf, GFP_KERNEL);
+ if (!options)
+ return -ENOMEM;
+
+ while ((p = strsep(&o, ",\n")) != NULL) {
+ if (!*p)
+ continue;
+
+ token = match_token(p, opt_tokens, args);
+ switch (token) {
+ case NVMF_OPT_WWNN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ *nname = token64;
+ break;
+ case NVMF_OPT_WWPN:
+ if (match_u64(args, &token64)) {
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ *pname = token64;
+ break;
+ default:
+ pr_warn("unknown parameter or missing value '%s'\n", p);
+ ret = -EINVAL;
+ goto out_free_options;
+ }
+ }
+
+out_free_options:
+ kfree(options);
+
+ if (!ret) {
+ if (*nname == -1)
+ return -EINVAL;
+ if (*pname == -1)
+ return -EINVAL;
+ }
+
+ return ret;
+}
+
+
+#define LPORT_OPTS (NVMF_OPT_WWNN | NVMF_OPT_WWPN)
+
+#define RPORT_OPTS (NVMF_OPT_WWNN | NVMF_OPT_WWPN | \
+ NVMF_OPT_LPWWNN | NVMF_OPT_LPWWPN)
+
+#define TGTPORT_OPTS (NVMF_OPT_WWNN | NVMF_OPT_WWPN)
+
+#define ALL_OPTS (NVMF_OPT_WWNN | NVMF_OPT_WWPN | NVMF_OPT_ROLES | \
+ NVMF_OPT_FCADDR | NVMF_OPT_LPWWNN | NVMF_OPT_LPWWPN)
+
+
+static DEFINE_SPINLOCK(fcloop_lock);
+static LIST_HEAD(fcloop_lports);
+static LIST_HEAD(fcloop_nports);
+
+struct fcloop_lport {
+ struct nvme_fc_local_port *localport;
+ struct list_head lport_list;
+ struct completion unreg_done;
+};
+
+struct fcloop_rport {
+ struct nvme_fc_remote_port *remoteport;
+ struct nvmet_fc_target_port *targetport;
+ struct fcloop_nport *nport;
+ struct fcloop_lport *lport;
+};
+
+struct fcloop_tport {
+ struct nvmet_fc_target_port *targetport;
+ struct nvme_fc_remote_port *remoteport;
+ struct fcloop_nport *nport;
+ struct fcloop_lport *lport;
+};
+
+struct fcloop_nport {
+ struct fcloop_rport *rport;
+ struct fcloop_tport *tport;
+ struct fcloop_lport *lport;
+ struct list_head nport_list;
+ struct kref ref;
+ struct completion rport_unreg_done;
+ struct completion tport_unreg_done;
+ u64 node_name;
+ u64 port_name;
+ u32 port_role;
+ u32 port_id;
+};
+
+struct fcloop_lsreq {
+ struct fcloop_tport *tport;
+ struct nvmefc_ls_req *lsreq;
+ struct work_struct work;
+ struct nvmefc_tgt_ls_req tgt_ls_req;
+ int status;
+};
+
+struct fcloop_fcpreq {
+ struct fcloop_tport *tport;
+ struct nvmefc_fcp_req *fcpreq;
+ u16 status;
+ struct work_struct work;
+ struct nvmefc_tgt_fcp_req tgt_fcp_req;
+};
+
+
+static inline struct fcloop_lsreq *
+tgt_ls_req_to_lsreq(struct nvmefc_tgt_ls_req *tgt_lsreq)
+{
+ return container_of(tgt_lsreq, struct fcloop_lsreq, tgt_ls_req);
+}
+
+static inline struct fcloop_fcpreq *
+tgt_fcp_req_to_fcpreq(struct nvmefc_tgt_fcp_req *tgt_fcpreq)
+{
+ return container_of(tgt_fcpreq, struct fcloop_fcpreq, tgt_fcp_req);
+}
+
+
+static int
+fcloop_create_queue(struct nvme_fc_local_port *localport,
+ unsigned int qidx, u16 qsize,
+ void **handle)
+{
+ *handle = localport;
+ return 0;
+}
+
+static void
+fcloop_delete_queue(struct nvme_fc_local_port *localport,
+ unsigned int idx, void *handle)
+{
+}
+
+
+/*
+ * Transmit of LS RSP done (e.g. buffers all set). call back up
+ * initiator "done" flows.
+ */
+static void
+fcloop_tgt_lsrqst_done_work(struct work_struct *work)
+{
+ struct fcloop_lsreq *tls_req =
+ container_of(work, struct fcloop_lsreq, work);
+ struct fcloop_tport *tport = tls_req->tport;
+ struct nvmefc_ls_req *lsreq = tls_req->lsreq;
+
+ if (tport->remoteport)
+ lsreq->done(lsreq, tls_req->status);
+}
+
+static int
+fcloop_ls_req(struct nvme_fc_local_port *localport,
+ struct nvme_fc_remote_port *remoteport,
+ struct nvmefc_ls_req *lsreq)
+{
+ struct fcloop_lsreq *tls_req = lsreq->private;
+ struct fcloop_rport *rport = remoteport->private;
+ int ret = 0;
+
+ tls_req->lsreq = lsreq;
+ INIT_WORK(&tls_req->work, fcloop_tgt_lsrqst_done_work);
+
+ if (!rport->targetport) {
+ tls_req->status = -ECONNREFUSED;
+ schedule_work(&tls_req->work);
+ return ret;
+ }
+
+ tls_req->status = 0;
+ tls_req->tport = rport->targetport->private;
+ ret = nvmet_fc_rcv_ls_req(rport->targetport, &tls_req->tgt_ls_req,
+ lsreq->rqstaddr, lsreq->rqstlen);
+
+ return ret;
+}
+
+static int
+fcloop_xmt_ls_rsp(struct nvmet_fc_target_port *tport,
+ struct nvmefc_tgt_ls_req *tgt_lsreq)
+{
+ struct fcloop_lsreq *tls_req = tgt_ls_req_to_lsreq(tgt_lsreq);
+ struct nvmefc_ls_req *lsreq = tls_req->lsreq;
+
+ memcpy(lsreq->rspaddr, tgt_lsreq->rspbuf,
+ ((lsreq->rsplen < tgt_lsreq->rsplen) ?
+ lsreq->rsplen : tgt_lsreq->rsplen));
+ tgt_lsreq->done(tgt_lsreq);
+
+ schedule_work(&tls_req->work);
+
+ return 0;
+}
+
+/*
+ * FCP IO operation done. call back up initiator "done" flows.
+ */
+static void
+fcloop_tgt_fcprqst_done_work(struct work_struct *work)
+{
+ struct fcloop_fcpreq *tfcp_req =
+ container_of(work, struct fcloop_fcpreq, work);
+ struct fcloop_tport *tport = tfcp_req->tport;
+ struct nvmefc_fcp_req *fcpreq = tfcp_req->fcpreq;
+
+ if (tport->remoteport) {
+ fcpreq->status = tfcp_req->status;
+ fcpreq->done(fcpreq);
+ }
+}
+
+
+static int
+fcloop_fcp_req(struct nvme_fc_local_port *localport,
+ struct nvme_fc_remote_port *remoteport,
+ void *hw_queue_handle,
+ struct nvmefc_fcp_req *fcpreq)
+{
+ struct fcloop_fcpreq *tfcp_req = fcpreq->private;
+ struct fcloop_rport *rport = remoteport->private;
+ int ret = 0;
+
+ INIT_WORK(&tfcp_req->work, fcloop_tgt_fcprqst_done_work);
+
+ if (!rport->targetport) {
+ tfcp_req->status = NVME_SC_FC_TRANSPORT_ERROR;
+ schedule_work(&tfcp_req->work);
+ return ret;
+ }
+
+ tfcp_req->fcpreq = fcpreq;
+ tfcp_req->tport = rport->targetport->private;
+
+ ret = nvmet_fc_rcv_fcp_req(rport->targetport, &tfcp_req->tgt_fcp_req,
+ fcpreq->cmdaddr, fcpreq->cmdlen);
+
+ return ret;
+}
+
+static void
+fcloop_fcp_copy_data(u8 op, struct scatterlist *data_sg,
+ struct scatterlist *io_sg, u32 offset, u32 length)
+{
+ void *data_p, *io_p;
+ u32 data_len, io_len, tlen;
+
+ io_p = sg_virt(io_sg);
+ io_len = io_sg->length;
+
+ for ( ; offset; ) {
+ tlen = min_t(u32, offset, io_len);
+ offset -= tlen;
+ io_len -= tlen;
+ if (!io_len) {
+ io_sg = sg_next(io_sg);
+ io_p = sg_virt(io_sg);
+ io_len = io_sg->length;
+ } else
+ io_p += tlen;
+ }
+
+ data_p = sg_virt(data_sg);
+ data_len = data_sg->length;
+
+ for ( ; length; ) {
+ tlen = min_t(u32, io_len, data_len);
+ tlen = min_t(u32, tlen, length);
+
+ if (op == NVMET_FCOP_WRITEDATA)
+ memcpy(data_p, io_p, tlen);
+ else
+ memcpy(io_p, data_p, tlen);
+
+ length -= tlen;
+
+ io_len -= tlen;
+ if ((!io_len) && (length)) {
+ io_sg = sg_next(io_sg);
+ io_p = sg_virt(io_sg);
+ io_len = io_sg->length;
+ } else
+ io_p += tlen;
+
+ data_len -= tlen;
+ if ((!data_len) && (length)) {
+ data_sg = sg_next(data_sg);
+ data_p = sg_virt(data_sg);
+ data_len = data_sg->length;
+ } else
+ data_p += tlen;
+ }
+}
+
+static int
+fcloop_fcp_op(struct nvmet_fc_target_port *tgtport,
+ struct nvmefc_tgt_fcp_req *tgt_fcpreq)
+{
+ struct fcloop_fcpreq *tfcp_req = tgt_fcp_req_to_fcpreq(tgt_fcpreq);
+ struct nvmefc_fcp_req *fcpreq = tfcp_req->fcpreq;
+ u32 rsplen = 0, xfrlen = 0;
+ int fcp_err = 0;
+ u8 op = tgt_fcpreq->op;
+
+ switch (op) {
+ case NVMET_FCOP_WRITEDATA:
+ xfrlen = tgt_fcpreq->transfer_length;
+ fcloop_fcp_copy_data(op, tgt_fcpreq->sg, fcpreq->first_sgl,
+ tgt_fcpreq->offset, xfrlen);
+ fcpreq->transferred_length += xfrlen;
+ break;
+
+ case NVMET_FCOP_READDATA:
+ case NVMET_FCOP_READDATA_RSP:
+ xfrlen = tgt_fcpreq->transfer_length;
+ fcloop_fcp_copy_data(op, tgt_fcpreq->sg, fcpreq->first_sgl,
+ tgt_fcpreq->offset, xfrlen);
+ fcpreq->transferred_length += xfrlen;
+ if (op == NVMET_FCOP_READDATA)
+ break;
+
+ /* Fall-Thru to RSP handling */
+
+ case NVMET_FCOP_RSP:
+ rsplen = ((fcpreq->rsplen < tgt_fcpreq->rsplen) ?
+ fcpreq->rsplen : tgt_fcpreq->rsplen);
+ memcpy(fcpreq->rspaddr, tgt_fcpreq->rspaddr, rsplen);
+ if (rsplen < tgt_fcpreq->rsplen)
+ fcp_err = -E2BIG;
+ fcpreq->rcv_rsplen = rsplen;
+ fcpreq->status = 0;
+ tfcp_req->status = 0;
+ break;
+
+ case NVMET_FCOP_ABORT:
+ tfcp_req->status = NVME_SC_FC_TRANSPORT_ABORTED;
+ break;
+
+ default:
+ fcp_err = -EINVAL;
+ break;
+ }
+
+ tgt_fcpreq->transferred_length = xfrlen;
+ tgt_fcpreq->fcp_error = fcp_err;
+ tgt_fcpreq->done(tgt_fcpreq);
+
+ if ((!fcp_err) && (op == NVMET_FCOP_RSP ||
+ op == NVMET_FCOP_READDATA_RSP ||
+ op == NVMET_FCOP_ABORT))
+ schedule_work(&tfcp_req->work);
+
+ return 0;
+}
+
+static void
+fcloop_ls_abort(struct nvme_fc_local_port *localport,
+ struct nvme_fc_remote_port *remoteport,
+ struct nvmefc_ls_req *lsreq)
+{
+}
+
+static void
+fcloop_fcp_abort(struct nvme_fc_local_port *localport,
+ struct nvme_fc_remote_port *remoteport,
+ void *hw_queue_handle,
+ struct nvmefc_fcp_req *fcpreq)
+{
+}
+
+static void
+fcloop_localport_delete(struct nvme_fc_local_port *localport)
+{
+ struct fcloop_lport *lport = localport->private;
+
+ /* release any threads waiting for the unreg to complete */
+ complete(&lport->unreg_done);
+}
+
+static void
+fcloop_remoteport_delete(struct nvme_fc_remote_port *remoteport)
+{
+ struct fcloop_rport *rport = remoteport->private;
+
+ /* release any threads waiting for the unreg to complete */
+ complete(&rport->nport->rport_unreg_done);
+}
+
+static void
+fcloop_targetport_delete(struct nvmet_fc_target_port *targetport)
+{
+ struct fcloop_tport *tport = targetport->private;
+
+ /* release any threads waiting for the unreg to complete */
+ complete(&tport->nport->tport_unreg_done);
+}
+
+#define FCLOOP_HW_QUEUES 4
+#define FCLOOP_SGL_SEGS 256
+#define FCLOOP_DMABOUND_4G 0xFFFFFFFF
+
+struct nvme_fc_port_template fctemplate = {
+ .localport_delete = fcloop_localport_delete,
+ .remoteport_delete = fcloop_remoteport_delete,
+ .create_queue = fcloop_create_queue,
+ .delete_queue = fcloop_delete_queue,
+ .ls_req = fcloop_ls_req,
+ .fcp_io = fcloop_fcp_req,
+ .ls_abort = fcloop_ls_abort,
+ .fcp_abort = fcloop_fcp_abort,
+ .max_hw_queues = FCLOOP_HW_QUEUES,
+ .max_sgl_segments = FCLOOP_SGL_SEGS,
+ .max_dif_sgl_segments = FCLOOP_SGL_SEGS,
+ .dma_boundary = FCLOOP_DMABOUND_4G,
+ /* sizes of additional private data for data structures */
+ .local_priv_sz = sizeof(struct fcloop_lport),
+ .remote_priv_sz = sizeof(struct fcloop_rport),
+ .lsrqst_priv_sz = sizeof(struct fcloop_lsreq),
+ .fcprqst_priv_sz = sizeof(struct fcloop_fcpreq),
+};
+
+struct nvmet_fc_target_template tgttemplate = {
+ .targetport_delete = fcloop_targetport_delete,
+ .xmt_ls_rsp = fcloop_xmt_ls_rsp,
+ .fcp_op = fcloop_fcp_op,
+ .max_hw_queues = FCLOOP_HW_QUEUES,
+ .max_sgl_segments = FCLOOP_SGL_SEGS,
+ .max_dif_sgl_segments = FCLOOP_SGL_SEGS,
+ .dma_boundary = FCLOOP_DMABOUND_4G,
+ /* optional features */
+ .target_features = NVMET_FCTGTFEAT_READDATA_RSP |
+ NVMET_FCTGTFEAT_NEEDS_CMD_CPUSCHED,
+ /* sizes of additional private data for data structures */
+ .target_priv_sz = sizeof(struct fcloop_tport),
+};
+
+static ssize_t
+fcloop_create_local_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct nvme_fc_port_info pinfo;
+ struct fcloop_ctrl_options *opts;
+ struct nvme_fc_local_port *localport;
+ struct fcloop_lport *lport;
+ int ret;
+
+ opts = kzalloc(sizeof(*opts), GFP_KERNEL);
+ if (!opts)
+ return -ENOMEM;
+
+ ret = fcloop_parse_options(opts, buf);
+ if (ret)
+ goto out_free_opts;
+
+ /* everything there ? */
+ if ((opts->mask & LPORT_OPTS) != LPORT_OPTS) {
+ ret = -EINVAL;
+ goto out_free_opts;
+ }
+
+ pinfo.node_name = opts->wwnn;
+ pinfo.port_name = opts->wwpn;
+ pinfo.port_role = opts->roles;
+ pinfo.port_id = opts->fcaddr;
+
+ ret = nvme_fc_register_localport(&pinfo, &fctemplate, NULL, &localport);
+ if (!ret) {
+ unsigned long flags;
+
+ /* success */
+ lport = localport->private;
+ lport->localport = localport;
+ INIT_LIST_HEAD(&lport->lport_list);
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+ list_add_tail(&lport->lport_list, &fcloop_lports);
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ /* mark all of the input buffer consumed */
+ ret = count;
+ }
+
+out_free_opts:
+ kfree(opts);
+ return ret ? ret : count;
+}
+
+
+static void
+__unlink_local_port(struct fcloop_lport *lport)
+{
+ list_del(&lport->lport_list);
+}
+
+static int
+__wait_localport_unreg(struct fcloop_lport *lport)
+{
+ int ret;
+
+ init_completion(&lport->unreg_done);
+
+ ret = nvme_fc_unregister_localport(lport->localport);
+
+ wait_for_completion(&lport->unreg_done);
+
+ return ret;
+}
+
+
+static ssize_t
+fcloop_delete_local_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct fcloop_lport *tlport, *lport = NULL;
+ u64 nodename, portname;
+ unsigned long flags;
+ int ret;
+
+ ret = fcloop_parse_nm_options(dev, &nodename, &portname, buf);
+ if (ret)
+ return ret;
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+
+ list_for_each_entry(tlport, &fcloop_lports, lport_list) {
+ if (tlport->localport->node_name == nodename &&
+ tlport->localport->port_name == portname) {
+ lport = tlport;
+ __unlink_local_port(lport);
+ break;
+ }
+ }
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ if (!lport)
+ return -ENOENT;
+
+ ret = __wait_localport_unreg(lport);
+
+ return ret ? ret : count;
+}
+
+static void
+fcloop_nport_free(struct kref *ref)
+{
+ struct fcloop_nport *nport =
+ container_of(ref, struct fcloop_nport, ref);
+ unsigned long flags;
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+ list_del(&nport->nport_list);
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ kfree(nport);
+}
+
+static void
+fcloop_nport_put(struct fcloop_nport *nport)
+{
+ kref_put(&nport->ref, fcloop_nport_free);
+}
+
+static int
+fcloop_nport_get(struct fcloop_nport *nport)
+{
+ return kref_get_unless_zero(&nport->ref);
+}
+
+static struct fcloop_nport *
+fcloop_alloc_nport(const char *buf, size_t count, bool remoteport)
+{
+ struct fcloop_nport *newnport, *nport = NULL;
+ struct fcloop_lport *tmplport, *lport = NULL;
+ struct fcloop_ctrl_options *opts;
+ unsigned long flags;
+ u32 opts_mask = (remoteport) ? RPORT_OPTS : TGTPORT_OPTS;
+ int ret;
+
+ opts = kzalloc(sizeof(*opts), GFP_KERNEL);
+ if (!opts)
+ return NULL;
+
+ ret = fcloop_parse_options(opts, buf);
+ if (ret)
+ goto out_free_opts;
+
+ /* everything there ? */
+ if ((opts->mask & opts_mask) != opts_mask) {
+ ret = -EINVAL;
+ goto out_free_opts;
+ }
+
+ newnport = kzalloc(sizeof(*newnport), GFP_KERNEL);
+ if (!newnport)
+ goto out_free_opts;
+
+ INIT_LIST_HEAD(&newnport->nport_list);
+ newnport->node_name = opts->wwnn;
+ newnport->port_name = opts->wwpn;
+ if (opts->mask & NVMF_OPT_ROLES)
+ newnport->port_role = opts->roles;
+ if (opts->mask & NVMF_OPT_FCADDR)
+ newnport->port_id = opts->fcaddr;
+ kref_init(&newnport->ref);
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+
+ list_for_each_entry(tmplport, &fcloop_lports, lport_list) {
+ if (tmplport->localport->node_name == opts->wwnn &&
+ tmplport->localport->port_name == opts->wwpn)
+ goto out_invalid_opts;
+
+ if (tmplport->localport->node_name == opts->lpwwnn &&
+ tmplport->localport->port_name == opts->lpwwpn)
+ lport = tmplport;
+ }
+
+ if (remoteport) {
+ if (!lport)
+ goto out_invalid_opts;
+ newnport->lport = lport;
+ }
+
+ list_for_each_entry(nport, &fcloop_nports, nport_list) {
+ if (nport->node_name == opts->wwnn &&
+ nport->port_name == opts->wwpn) {
+ if ((remoteport && nport->rport) ||
+ (!remoteport && nport->tport)) {
+ nport = NULL;
+ goto out_invalid_opts;
+ }
+
+ fcloop_nport_get(nport);
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ if (remoteport)
+ nport->lport = lport;
+ if (opts->mask & NVMF_OPT_ROLES)
+ nport->port_role = opts->roles;
+ if (opts->mask & NVMF_OPT_FCADDR)
+ nport->port_id = opts->fcaddr;
+ goto out_free_newnport;
+ }
+ }
+
+ list_add_tail(&newnport->nport_list, &fcloop_nports);
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ kfree(opts);
+ return newnport;
+
+out_invalid_opts:
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+out_free_newnport:
+ kfree(newnport);
+out_free_opts:
+ kfree(opts);
+ return nport;
+}
+
+static ssize_t
+fcloop_create_remote_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct nvme_fc_remote_port *remoteport;
+ struct fcloop_nport *nport;
+ struct fcloop_rport *rport;
+ struct nvme_fc_port_info pinfo;
+ int ret;
+
+ nport = fcloop_alloc_nport(buf, count, true);
+ if (!nport)
+ return -EIO;
+
+ pinfo.node_name = nport->node_name;
+ pinfo.port_name = nport->port_name;
+ pinfo.port_role = nport->port_role;
+ pinfo.port_id = nport->port_id;
+
+ ret = nvme_fc_register_remoteport(nport->lport->localport,
+ &pinfo, &remoteport);
+ if (ret || !remoteport) {
+ fcloop_nport_put(nport);
+ return ret;
+ }
+
+ /* success */
+ rport = remoteport->private;
+ rport->remoteport = remoteport;
+ rport->targetport = (nport->tport) ? nport->tport->targetport : NULL;
+ if (nport->tport) {
+ nport->tport->remoteport = remoteport;
+ nport->tport->lport = nport->lport;
+ }
+ rport->nport = nport;
+ rport->lport = nport->lport;
+ nport->rport = rport;
+
+ return ret ? ret : count;
+}
+
+
+static struct fcloop_rport *
+__unlink_remote_port(struct fcloop_nport *nport)
+{
+ struct fcloop_rport *rport = nport->rport;
+
+ if (rport && nport->tport)
+ nport->tport->remoteport = NULL;
+ nport->rport = NULL;
+
+ return rport;
+}
+
+static int
+__wait_remoteport_unreg(struct fcloop_nport *nport, struct fcloop_rport *rport)
+{
+ int ret;
+
+ if (!rport)
+ return -EALREADY;
+
+ init_completion(&nport->rport_unreg_done);
+
+ ret = nvme_fc_unregister_remoteport(rport->remoteport);
+ if (ret)
+ return ret;
+
+ wait_for_completion(&nport->rport_unreg_done);
+
+ fcloop_nport_put(nport);
+
+ return ret;
+}
+
+static ssize_t
+fcloop_delete_remote_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct fcloop_nport *nport = NULL, *tmpport;
+ static struct fcloop_rport *rport;
+ u64 nodename, portname;
+ unsigned long flags;
+ int ret;
+
+ ret = fcloop_parse_nm_options(dev, &nodename, &portname, buf);
+ if (ret)
+ return ret;
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+
+ list_for_each_entry(tmpport, &fcloop_nports, nport_list) {
+ if (tmpport->node_name == nodename &&
+ tmpport->port_name == portname && tmpport->rport) {
+ nport = tmpport;
+ rport = __unlink_remote_port(nport);
+ break;
+ }
+ }
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ if (!nport)
+ return -ENOENT;
+
+ ret = __wait_remoteport_unreg(nport, rport);
+
+ return ret ? ret : count;
+}
+
+static ssize_t
+fcloop_create_target_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct nvmet_fc_target_port *targetport;
+ struct fcloop_nport *nport;
+ struct fcloop_tport *tport;
+ struct nvmet_fc_port_info tinfo;
+ int ret;
+
+ nport = fcloop_alloc_nport(buf, count, false);
+ if (!nport)
+ return -EIO;
+
+ tinfo.node_name = nport->node_name;
+ tinfo.port_name = nport->port_name;
+ tinfo.port_id = nport->port_id;
+
+ ret = nvmet_fc_register_targetport(&tinfo, &tgttemplate, NULL,
+ &targetport);
+ if (ret) {
+ fcloop_nport_put(nport);
+ return ret;
+ }
+
+ /* success */
+ tport = targetport->private;
+ tport->targetport = targetport;
+ tport->remoteport = (nport->rport) ? nport->rport->remoteport : NULL;
+ if (nport->rport)
+ nport->rport->targetport = targetport;
+ tport->nport = nport;
+ tport->lport = nport->lport;
+ nport->tport = tport;
+
+ return ret ? ret : count;
+}
+
+
+static struct fcloop_tport *
+__unlink_target_port(struct fcloop_nport *nport)
+{
+ struct fcloop_tport *tport = nport->tport;
+
+ if (tport && nport->rport)
+ nport->rport->targetport = NULL;
+ nport->tport = NULL;
+
+ return tport;
+}
+
+static int
+__wait_targetport_unreg(struct fcloop_nport *nport, struct fcloop_tport *tport)
+{
+ int ret;
+
+ if (!tport)
+ return -EALREADY;
+
+ init_completion(&nport->tport_unreg_done);
+
+ ret = nvmet_fc_unregister_targetport(tport->targetport);
+ if (ret)
+ return ret;
+
+ wait_for_completion(&nport->tport_unreg_done);
+
+ fcloop_nport_put(nport);
+
+ return ret;
+}
+
+static ssize_t
+fcloop_delete_target_port(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct fcloop_nport *nport = NULL, *tmpport;
+ struct fcloop_tport *tport;
+ u64 nodename, portname;
+ unsigned long flags;
+ int ret;
+
+ ret = fcloop_parse_nm_options(dev, &nodename, &portname, buf);
+ if (ret)
+ return ret;
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+
+ list_for_each_entry(tmpport, &fcloop_nports, nport_list) {
+ if (tmpport->node_name == nodename &&
+ tmpport->port_name == portname && tmpport->tport) {
+ nport = tmpport;
+ tport = __unlink_target_port(nport);
+ break;
+ }
+ }
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ if (!nport)
+ return -ENOENT;
+
+ ret = __wait_targetport_unreg(nport, tport);
+
+ return ret ? ret : count;
+}
+
+
+static DEVICE_ATTR(add_local_port, 0200, NULL, fcloop_create_local_port);
+static DEVICE_ATTR(del_local_port, 0200, NULL, fcloop_delete_local_port);
+static DEVICE_ATTR(add_remote_port, 0200, NULL, fcloop_create_remote_port);
+static DEVICE_ATTR(del_remote_port, 0200, NULL, fcloop_delete_remote_port);
+static DEVICE_ATTR(add_target_port, 0200, NULL, fcloop_create_target_port);
+static DEVICE_ATTR(del_target_port, 0200, NULL, fcloop_delete_target_port);
+
+static struct attribute *fcloop_dev_attrs[] = {
+ &dev_attr_add_local_port.attr,
+ &dev_attr_del_local_port.attr,
+ &dev_attr_add_remote_port.attr,
+ &dev_attr_del_remote_port.attr,
+ &dev_attr_add_target_port.attr,
+ &dev_attr_del_target_port.attr,
+ NULL
+};
+
+static struct attribute_group fclopp_dev_attrs_group = {
+ .attrs = fcloop_dev_attrs,
+};
+
+static const struct attribute_group *fcloop_dev_attr_groups[] = {
+ &fclopp_dev_attrs_group,
+ NULL,
+};
+
+static struct class *fcloop_class;
+static struct device *fcloop_device;
+
+
+static int __init fcloop_init(void)
+{
+ int ret;
+
+ fcloop_class = class_create(THIS_MODULE, "fcloop");
+ if (IS_ERR(fcloop_class)) {
+ pr_err("couldn't register class fcloop\n");
+ ret = PTR_ERR(fcloop_class);
+ return ret;
+ }
+
+ fcloop_device = device_create_with_groups(
+ fcloop_class, NULL, MKDEV(0, 0), NULL,
+ fcloop_dev_attr_groups, "ctl");
+ if (IS_ERR(fcloop_device)) {
+ pr_err("couldn't create ctl device!\n");
+ ret = PTR_ERR(fcloop_device);
+ goto out_destroy_class;
+ }
+
+ get_device(fcloop_device);
+
+ return 0;
+
+out_destroy_class:
+ class_destroy(fcloop_class);
+ return ret;
+}
+
+static void __exit fcloop_exit(void)
+{
+ struct fcloop_lport *lport;
+ struct fcloop_nport *nport;
+ struct fcloop_tport *tport;
+ struct fcloop_rport *rport;
+ unsigned long flags;
+ int ret;
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+
+ for (;;) {
+ nport = list_first_entry_or_null(&fcloop_nports,
+ typeof(*nport), nport_list);
+ if (!nport)
+ break;
+
+ tport = __unlink_target_port(nport);
+ rport = __unlink_remote_port(nport);
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ ret = __wait_targetport_unreg(nport, tport);
+ if (ret)
+ pr_warn("%s: Failed deleting target port\n", __func__);
+
+ ret = __wait_remoteport_unreg(nport, rport);
+ if (ret)
+ pr_warn("%s: Failed deleting remote port\n", __func__);
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+ }
+
+ for (;;) {
+ lport = list_first_entry_or_null(&fcloop_lports,
+ typeof(*lport), lport_list);
+ if (!lport)
+ break;
+
+ __unlink_local_port(lport);
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ ret = __wait_localport_unreg(lport);
+ if (ret)
+ pr_warn("%s: Failed deleting local port\n", __func__);
+
+ spin_lock_irqsave(&fcloop_lock, flags);
+ }
+
+ spin_unlock_irqrestore(&fcloop_lock, flags);
+
+ put_device(fcloop_device);
+
+ device_destroy(fcloop_class, MKDEV(0, 0));
+ class_destroy(fcloop_class);
+}
+
+module_init(fcloop_init);
+module_exit(fcloop_exit);
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/nvme/target/io-cmd.c b/drivers/nvme/target/io-cmd.c
index 4a96c2049b7b..4195115c7e54 100644
--- a/drivers/nvme/target/io-cmd.c
+++ b/drivers/nvme/target/io-cmd.c
@@ -37,9 +37,7 @@ static void nvmet_inline_bio_init(struct nvmet_req *req)
{
struct bio *bio = &req->inline_bio;
- bio_init(bio);
- bio->bi_max_vecs = NVMET_MAX_INLINE_BIOVEC;
- bio->bi_io_vec = req->inline_bvec;
+ bio_init(bio, req->inline_bvec, NVMET_MAX_INLINE_BIOVEC);
}
static void nvmet_execute_rw(struct nvmet_req *req)
@@ -58,7 +56,7 @@ static void nvmet_execute_rw(struct nvmet_req *req)
if (req->cmd->rw.opcode == nvme_cmd_write) {
op = REQ_OP_WRITE;
- op_flags = WRITE_ODIRECT;
+ op_flags = REQ_SYNC | REQ_IDLE;
if (req->cmd->rw.control & cpu_to_le16(NVME_RW_FUA))
op_flags |= REQ_FUA;
} else {
@@ -96,7 +94,7 @@ static void nvmet_execute_rw(struct nvmet_req *req)
cookie = submit_bio(bio);
- blk_poll(bdev_get_queue(req->ns->bdev), cookie);
+ blk_mq_poll(bdev_get_queue(req->ns->bdev), cookie);
}
static void nvmet_execute_flush(struct nvmet_req *req)
@@ -109,7 +107,7 @@ static void nvmet_execute_flush(struct nvmet_req *req)
bio->bi_bdev = req->ns->bdev;
bio->bi_private = req;
bio->bi_end_io = nvmet_bio_done;
- bio_set_op_attrs(bio, REQ_OP_WRITE, WRITE_FLUSH);
+ bio->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH;
submit_bio(bio);
}
@@ -172,6 +170,32 @@ static void nvmet_execute_dsm(struct nvmet_req *req)
}
}
+static void nvmet_execute_write_zeroes(struct nvmet_req *req)
+{
+ struct nvme_write_zeroes_cmd *write_zeroes = &req->cmd->write_zeroes;
+ struct bio *bio = NULL;
+ u16 status = NVME_SC_SUCCESS;
+ sector_t sector;
+ sector_t nr_sector;
+
+ sector = le64_to_cpu(write_zeroes->slba) <<
+ (req->ns->blksize_shift - 9);
+ nr_sector = (((sector_t)le32_to_cpu(write_zeroes->length)) <<
+ (req->ns->blksize_shift - 9)) + 1;
+
+ if (__blkdev_issue_zeroout(req->ns->bdev, sector, nr_sector,
+ GFP_KERNEL, &bio, true))
+ status = NVME_SC_INTERNAL | NVME_SC_DNR;
+
+ if (bio) {
+ bio->bi_private = req;
+ bio->bi_end_io = nvmet_bio_done;
+ submit_bio(bio);
+ } else {
+ nvmet_req_complete(req, status);
+ }
+}
+
int nvmet_parse_io_cmd(struct nvmet_req *req)
{
struct nvme_command *cmd = req->cmd;
@@ -209,6 +233,9 @@ int nvmet_parse_io_cmd(struct nvmet_req *req)
req->data_len = le32_to_cpu(cmd->dsm.nr + 1) *
sizeof(struct nvme_dsm_range);
return 0;
+ case nvme_cmd_write_zeroes:
+ req->execute = nvmet_execute_write_zeroes;
+ return 0;
default:
pr_err("nvmet: unhandled cmd %d\n", cmd->common.opcode);
return NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
diff --git a/drivers/nvme/target/loop.c b/drivers/nvme/target/loop.c
index d5df77d686b2..9aaa70071ae5 100644
--- a/drivers/nvme/target/loop.c
+++ b/drivers/nvme/target/loop.c
@@ -36,6 +36,7 @@
(NVME_LOOP_AQ_DEPTH - NVME_LOOP_NR_AEN_COMMANDS)
struct nvme_loop_iod {
+ struct nvme_request nvme_req;
struct nvme_command cmd;
struct nvme_completion rsp;
struct nvmet_req req;
@@ -112,10 +113,10 @@ static void nvme_loop_complete_rq(struct request *req)
blk_mq_end_request(req, error);
}
-static void nvme_loop_queue_response(struct nvmet_req *nvme_req)
+static void nvme_loop_queue_response(struct nvmet_req *req)
{
struct nvme_loop_iod *iod =
- container_of(nvme_req, struct nvme_loop_iod, req);
+ container_of(req, struct nvme_loop_iod, req);
struct nvme_completion *cqe = &iod->rsp;
/*
@@ -126,13 +127,13 @@ static void nvme_loop_queue_response(struct nvmet_req *nvme_req)
*/
if (unlikely(nvme_loop_queue_idx(iod->queue) == 0 &&
cqe->command_id >= NVME_LOOP_AQ_BLKMQ_DEPTH)) {
- nvme_complete_async_event(&iod->queue->ctrl->ctrl, cqe);
+ nvme_complete_async_event(&iod->queue->ctrl->ctrl, cqe->status,
+ &cqe->result);
} else {
- struct request *req = blk_mq_rq_from_pdu(iod);
+ struct request *rq = blk_mq_rq_from_pdu(iod);
- if (req->cmd_type == REQ_TYPE_DRV_PRIV && req->special)
- memcpy(req->special, cqe, sizeof(*cqe));
- blk_mq_complete_request(req, le16_to_cpu(cqe->status) >> 1);
+ iod->nvme_req.result = cqe->result;
+ blk_mq_complete_request(rq, le16_to_cpu(cqe->status) >> 1);
}
}
@@ -168,7 +169,7 @@ static int nvme_loop_queue_rq(struct blk_mq_hw_ctx *hctx,
int ret;
ret = nvme_setup_cmd(ns, req, &iod->cmd);
- if (ret)
+ if (ret != BLK_MQ_RQ_QUEUE_OK)
return ret;
iod->cmd.common.flags |= NVME_CMD_SGL_METABUF;
@@ -178,26 +179,25 @@ static int nvme_loop_queue_rq(struct blk_mq_hw_ctx *hctx,
nvme_cleanup_cmd(req);
blk_mq_start_request(req);
nvme_loop_queue_response(&iod->req);
- return 0;
+ return BLK_MQ_RQ_QUEUE_OK;
}
if (blk_rq_bytes(req)) {
iod->sg_table.sgl = iod->first_sgl;
ret = sg_alloc_table_chained(&iod->sg_table,
- req->nr_phys_segments, iod->sg_table.sgl);
+ blk_rq_nr_phys_segments(req),
+ iod->sg_table.sgl);
if (ret)
return BLK_MQ_RQ_QUEUE_BUSY;
iod->req.sg = iod->sg_table.sgl;
iod->req.sg_cnt = blk_rq_map_sg(req->q, req, iod->sg_table.sgl);
- BUG_ON(iod->req.sg_cnt > req->nr_phys_segments);
}
- iod->cmd.common.command_id = req->tag;
blk_mq_start_request(req);
schedule_work(&iod->work);
- return 0;
+ return BLK_MQ_RQ_QUEUE_OK;
}
static void nvme_loop_submit_async_event(struct nvme_ctrl *arg, int aer_idx)
diff --git a/drivers/nvme/target/nvmet.h b/drivers/nvme/target/nvmet.h
index 76b6eedccaf9..23d5eb1c944f 100644
--- a/drivers/nvme/target/nvmet.h
+++ b/drivers/nvme/target/nvmet.h
@@ -47,6 +47,7 @@ struct nvmet_ns {
loff_t size;
u8 nguid[16];
+ bool enabled;
struct nvmet_subsys *subsys;
const char *device_path;
@@ -61,11 +62,6 @@ static inline struct nvmet_ns *to_nvmet_ns(struct config_item *item)
return container_of(to_config_group(item), struct nvmet_ns, group);
}
-static inline bool nvmet_ns_enabled(struct nvmet_ns *ns)
-{
- return !list_empty_careful(&ns->dev_link);
-}
-
struct nvmet_cq {
u16 qid;
u16 size;
@@ -238,7 +234,7 @@ static inline void nvmet_set_status(struct nvmet_req *req, u16 status)
static inline void nvmet_set_result(struct nvmet_req *req, u32 result)
{
- req->rsp->result = cpu_to_le32(result);
+ req->rsp->result.u32 = cpu_to_le32(result);
}
/*
diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c
index 50f237a46ea1..8c3760a78ac0 100644
--- a/drivers/nvme/target/rdma.c
+++ b/drivers/nvme/target/rdma.c
@@ -951,6 +951,7 @@ err_destroy_cq:
static void nvmet_rdma_destroy_queue_ib(struct nvmet_rdma_queue *queue)
{
+ ib_drain_qp(queue->cm_id->qp);
rdma_destroy_qp(queue->cm_id);
ib_free_cq(queue->cq);
}
@@ -1044,8 +1045,10 @@ nvmet_rdma_alloc_queue(struct nvmet_rdma_device *ndev,
}
ret = nvmet_sq_init(&queue->nvme_sq);
- if (ret)
+ if (ret) {
+ ret = NVME_RDMA_CM_NO_RSC;
goto out_free_queue;
+ }
ret = nvmet_rdma_parse_cm_connect_req(&event->param.conn, queue);
if (ret)
@@ -1066,6 +1069,7 @@ nvmet_rdma_alloc_queue(struct nvmet_rdma_device *ndev,
spin_lock_init(&queue->rsp_wr_wait_lock);
INIT_LIST_HEAD(&queue->free_rsps);
spin_lock_init(&queue->rsps_lock);
+ INIT_LIST_HEAD(&queue->queue_list);
queue->idx = ida_simple_get(&nvmet_rdma_queue_ida, 0, 0, GFP_KERNEL);
if (queue->idx < 0) {
@@ -1114,6 +1118,7 @@ out_destroy_sq:
out_free_queue:
kfree(queue);
out_reject:
+ pr_debug("rejecting connect request with status code %d\n", ret);
nvmet_rdma_cm_reject(cm_id, ret);
return NULL;
}
@@ -1127,7 +1132,8 @@ static void nvmet_rdma_qp_event(struct ib_event *event, void *priv)
rdma_notify(queue->cm_id, event->event);
break;
default:
- pr_err("received unrecognized IB QP event %d\n", event->event);
+ pr_err("received IB QP event: %s (%d)\n",
+ ib_event_msg(event->event), event->event);
break;
}
}
@@ -1244,7 +1250,6 @@ static void __nvmet_rdma_queue_disconnect(struct nvmet_rdma_queue *queue)
if (disconnect) {
rdma_disconnect(queue->cm_id);
- ib_drain_qp(queue->cm_id->qp);
schedule_work(&queue->release_work);
}
}
@@ -1269,7 +1274,12 @@ static void nvmet_rdma_queue_connect_fail(struct rdma_cm_id *cm_id,
{
WARN_ON_ONCE(queue->state != NVMET_RDMA_Q_CONNECTING);
- pr_err("failed to connect queue\n");
+ mutex_lock(&nvmet_rdma_queue_mutex);
+ if (!list_empty(&queue->queue_list))
+ list_del_init(&queue->queue_list);
+ mutex_unlock(&nvmet_rdma_queue_mutex);
+
+ pr_err("failed to connect queue %d\n", queue->idx);
schedule_work(&queue->release_work);
}
@@ -1352,7 +1362,13 @@ static int nvmet_rdma_cm_handler(struct rdma_cm_id *cm_id,
case RDMA_CM_EVENT_ADDR_CHANGE:
case RDMA_CM_EVENT_DISCONNECTED:
case RDMA_CM_EVENT_TIMEWAIT_EXIT:
- nvmet_rdma_queue_disconnect(queue);
+ /*
+ * We might end up here when we already freed the qp
+ * which means queue release sequence is in progress,
+ * so don't get in the way...
+ */
+ if (queue)
+ nvmet_rdma_queue_disconnect(queue);
break;
case RDMA_CM_EVENT_DEVICE_REMOVAL:
ret = nvmet_rdma_device_removal(cm_id, queue);