https://github.com/torvalds/linux
Tip revision: 3650b228f83adda7e5ee532e2b90429c03f7b9ec authored by Linus Torvalds on 25 October 2020, 22:14:11 UTC
Linux 5.10-rc1
Linux 5.10-rc1
Tip revision: 3650b22
smc_ib.c
// SPDX-License-Identifier: GPL-2.0
/*
* Shared Memory Communications over RDMA (SMC-R) and RoCE
*
* IB infrastructure:
* Establish SMC-R as an Infiniband Client to be notified about added and
* removed IB devices of type RDMA.
* Determine device and port characteristics for these IB devices.
*
* Copyright IBM Corp. 2016
*
* Author(s): Ursula Braun <ubraun@linux.vnet.ibm.com>
*/
#include <linux/random.h>
#include <linux/workqueue.h>
#include <linux/scatterlist.h>
#include <linux/wait.h>
#include <linux/mutex.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_cache.h>
#include "smc_pnet.h"
#include "smc_ib.h"
#include "smc_core.h"
#include "smc_wr.h"
#include "smc.h"
#define SMC_MAX_CQE 32766 /* max. # of completion queue elements */
#define SMC_QP_MIN_RNR_TIMER 5
#define SMC_QP_TIMEOUT 15 /* 4096 * 2 ** timeout usec */
#define SMC_QP_RETRY_CNT 7 /* 7: infinite */
#define SMC_QP_RNR_RETRY 7 /* 7: infinite */
struct smc_ib_devices smc_ib_devices = { /* smc-registered ib devices */
.mutex = __MUTEX_INITIALIZER(smc_ib_devices.mutex),
.list = LIST_HEAD_INIT(smc_ib_devices.list),
};
u8 local_systemid[SMC_SYSTEMID_LEN]; /* unique system identifier */
static int smc_ib_modify_qp_init(struct smc_link *lnk)
{
struct ib_qp_attr qp_attr;
memset(&qp_attr, 0, sizeof(qp_attr));
qp_attr.qp_state = IB_QPS_INIT;
qp_attr.pkey_index = 0;
qp_attr.port_num = lnk->ibport;
qp_attr.qp_access_flags = IB_ACCESS_LOCAL_WRITE
| IB_ACCESS_REMOTE_WRITE;
return ib_modify_qp(lnk->roce_qp, &qp_attr,
IB_QP_STATE | IB_QP_PKEY_INDEX |
IB_QP_ACCESS_FLAGS | IB_QP_PORT);
}
static int smc_ib_modify_qp_rtr(struct smc_link *lnk)
{
enum ib_qp_attr_mask qp_attr_mask =
IB_QP_STATE | IB_QP_AV | IB_QP_PATH_MTU | IB_QP_DEST_QPN |
IB_QP_RQ_PSN | IB_QP_MAX_DEST_RD_ATOMIC | IB_QP_MIN_RNR_TIMER;
struct ib_qp_attr qp_attr;
memset(&qp_attr, 0, sizeof(qp_attr));
qp_attr.qp_state = IB_QPS_RTR;
qp_attr.path_mtu = min(lnk->path_mtu, lnk->peer_mtu);
qp_attr.ah_attr.type = RDMA_AH_ATTR_TYPE_ROCE;
rdma_ah_set_port_num(&qp_attr.ah_attr, lnk->ibport);
rdma_ah_set_grh(&qp_attr.ah_attr, NULL, 0, lnk->sgid_index, 1, 0);
rdma_ah_set_dgid_raw(&qp_attr.ah_attr, lnk->peer_gid);
memcpy(&qp_attr.ah_attr.roce.dmac, lnk->peer_mac,
sizeof(lnk->peer_mac));
qp_attr.dest_qp_num = lnk->peer_qpn;
qp_attr.rq_psn = lnk->peer_psn; /* starting receive packet seq # */
qp_attr.max_dest_rd_atomic = 1; /* max # of resources for incoming
* requests
*/
qp_attr.min_rnr_timer = SMC_QP_MIN_RNR_TIMER;
return ib_modify_qp(lnk->roce_qp, &qp_attr, qp_attr_mask);
}
int smc_ib_modify_qp_rts(struct smc_link *lnk)
{
struct ib_qp_attr qp_attr;
memset(&qp_attr, 0, sizeof(qp_attr));
qp_attr.qp_state = IB_QPS_RTS;
qp_attr.timeout = SMC_QP_TIMEOUT; /* local ack timeout */
qp_attr.retry_cnt = SMC_QP_RETRY_CNT; /* retry count */
qp_attr.rnr_retry = SMC_QP_RNR_RETRY; /* RNR retries, 7=infinite */
qp_attr.sq_psn = lnk->psn_initial; /* starting send packet seq # */
qp_attr.max_rd_atomic = 1; /* # of outstanding RDMA reads and
* atomic ops allowed
*/
return ib_modify_qp(lnk->roce_qp, &qp_attr,
IB_QP_STATE | IB_QP_TIMEOUT | IB_QP_RETRY_CNT |
IB_QP_SQ_PSN | IB_QP_RNR_RETRY |
IB_QP_MAX_QP_RD_ATOMIC);
}
int smc_ib_modify_qp_reset(struct smc_link *lnk)
{
struct ib_qp_attr qp_attr;
memset(&qp_attr, 0, sizeof(qp_attr));
qp_attr.qp_state = IB_QPS_RESET;
return ib_modify_qp(lnk->roce_qp, &qp_attr, IB_QP_STATE);
}
int smc_ib_ready_link(struct smc_link *lnk)
{
struct smc_link_group *lgr = smc_get_lgr(lnk);
int rc = 0;
rc = smc_ib_modify_qp_init(lnk);
if (rc)
goto out;
rc = smc_ib_modify_qp_rtr(lnk);
if (rc)
goto out;
smc_wr_remember_qp_attr(lnk);
rc = ib_req_notify_cq(lnk->smcibdev->roce_cq_recv,
IB_CQ_SOLICITED_MASK);
if (rc)
goto out;
rc = smc_wr_rx_post_init(lnk);
if (rc)
goto out;
smc_wr_remember_qp_attr(lnk);
if (lgr->role == SMC_SERV) {
rc = smc_ib_modify_qp_rts(lnk);
if (rc)
goto out;
smc_wr_remember_qp_attr(lnk);
}
out:
return rc;
}
static int smc_ib_fill_mac(struct smc_ib_device *smcibdev, u8 ibport)
{
const struct ib_gid_attr *attr;
int rc;
attr = rdma_get_gid_attr(smcibdev->ibdev, ibport, 0);
if (IS_ERR(attr))
return -ENODEV;
rc = rdma_read_gid_l2_fields(attr, NULL, smcibdev->mac[ibport - 1]);
rdma_put_gid_attr(attr);
return rc;
}
/* Create an identifier unique for this instance of SMC-R.
* The MAC-address of the first active registered IB device
* plus a random 2-byte number is used to create this identifier.
* This name is delivered to the peer during connection initialization.
*/
static inline void smc_ib_define_local_systemid(struct smc_ib_device *smcibdev,
u8 ibport)
{
memcpy(&local_systemid[2], &smcibdev->mac[ibport - 1],
sizeof(smcibdev->mac[ibport - 1]));
}
bool smc_ib_is_valid_local_systemid(void)
{
return !is_zero_ether_addr(&local_systemid[2]);
}
static void smc_ib_init_local_systemid(void)
{
get_random_bytes(&local_systemid[0], 2);
}
bool smc_ib_port_active(struct smc_ib_device *smcibdev, u8 ibport)
{
return smcibdev->pattr[ibport - 1].state == IB_PORT_ACTIVE;
}
/* determine the gid for an ib-device port and vlan id */
int smc_ib_determine_gid(struct smc_ib_device *smcibdev, u8 ibport,
unsigned short vlan_id, u8 gid[], u8 *sgid_index)
{
const struct ib_gid_attr *attr;
const struct net_device *ndev;
int i;
for (i = 0; i < smcibdev->pattr[ibport - 1].gid_tbl_len; i++) {
attr = rdma_get_gid_attr(smcibdev->ibdev, ibport, i);
if (IS_ERR(attr))
continue;
rcu_read_lock();
ndev = rdma_read_gid_attr_ndev_rcu(attr);
if (!IS_ERR(ndev) &&
((!vlan_id && !is_vlan_dev(attr->ndev)) ||
(vlan_id && is_vlan_dev(attr->ndev) &&
vlan_dev_vlan_id(attr->ndev) == vlan_id)) &&
attr->gid_type == IB_GID_TYPE_ROCE) {
rcu_read_unlock();
if (gid)
memcpy(gid, &attr->gid, SMC_GID_SIZE);
if (sgid_index)
*sgid_index = attr->index;
rdma_put_gid_attr(attr);
return 0;
}
rcu_read_unlock();
rdma_put_gid_attr(attr);
}
return -ENODEV;
}
static int smc_ib_remember_port_attr(struct smc_ib_device *smcibdev, u8 ibport)
{
int rc;
memset(&smcibdev->pattr[ibport - 1], 0,
sizeof(smcibdev->pattr[ibport - 1]));
rc = ib_query_port(smcibdev->ibdev, ibport,
&smcibdev->pattr[ibport - 1]);
if (rc)
goto out;
/* the SMC protocol requires specification of the RoCE MAC address */
rc = smc_ib_fill_mac(smcibdev, ibport);
if (rc)
goto out;
if (!smc_ib_is_valid_local_systemid() &&
smc_ib_port_active(smcibdev, ibport))
/* create unique system identifier */
smc_ib_define_local_systemid(smcibdev, ibport);
out:
return rc;
}
/* process context wrapper for might_sleep smc_ib_remember_port_attr */
static void smc_ib_port_event_work(struct work_struct *work)
{
struct smc_ib_device *smcibdev = container_of(
work, struct smc_ib_device, port_event_work);
u8 port_idx;
for_each_set_bit(port_idx, &smcibdev->port_event_mask, SMC_MAX_PORTS) {
smc_ib_remember_port_attr(smcibdev, port_idx + 1);
clear_bit(port_idx, &smcibdev->port_event_mask);
if (!smc_ib_port_active(smcibdev, port_idx + 1)) {
set_bit(port_idx, smcibdev->ports_going_away);
smcr_port_err(smcibdev, port_idx + 1);
} else {
clear_bit(port_idx, smcibdev->ports_going_away);
smcr_port_add(smcibdev, port_idx + 1);
}
}
}
/* can be called in IRQ context */
static void smc_ib_global_event_handler(struct ib_event_handler *handler,
struct ib_event *ibevent)
{
struct smc_ib_device *smcibdev;
bool schedule = false;
u8 port_idx;
smcibdev = container_of(handler, struct smc_ib_device, event_handler);
switch (ibevent->event) {
case IB_EVENT_DEVICE_FATAL:
/* terminate all ports on device */
for (port_idx = 0; port_idx < SMC_MAX_PORTS; port_idx++) {
set_bit(port_idx, &smcibdev->port_event_mask);
if (!test_and_set_bit(port_idx,
smcibdev->ports_going_away))
schedule = true;
}
if (schedule)
schedule_work(&smcibdev->port_event_work);
break;
case IB_EVENT_PORT_ACTIVE:
port_idx = ibevent->element.port_num - 1;
if (port_idx >= SMC_MAX_PORTS)
break;
set_bit(port_idx, &smcibdev->port_event_mask);
if (test_and_clear_bit(port_idx, smcibdev->ports_going_away))
schedule_work(&smcibdev->port_event_work);
break;
case IB_EVENT_PORT_ERR:
port_idx = ibevent->element.port_num - 1;
if (port_idx >= SMC_MAX_PORTS)
break;
set_bit(port_idx, &smcibdev->port_event_mask);
if (!test_and_set_bit(port_idx, smcibdev->ports_going_away))
schedule_work(&smcibdev->port_event_work);
break;
case IB_EVENT_GID_CHANGE:
port_idx = ibevent->element.port_num - 1;
if (port_idx >= SMC_MAX_PORTS)
break;
set_bit(port_idx, &smcibdev->port_event_mask);
schedule_work(&smcibdev->port_event_work);
break;
default:
break;
}
}
void smc_ib_dealloc_protection_domain(struct smc_link *lnk)
{
if (lnk->roce_pd)
ib_dealloc_pd(lnk->roce_pd);
lnk->roce_pd = NULL;
}
int smc_ib_create_protection_domain(struct smc_link *lnk)
{
int rc;
lnk->roce_pd = ib_alloc_pd(lnk->smcibdev->ibdev, 0);
rc = PTR_ERR_OR_ZERO(lnk->roce_pd);
if (IS_ERR(lnk->roce_pd))
lnk->roce_pd = NULL;
return rc;
}
static void smc_ib_qp_event_handler(struct ib_event *ibevent, void *priv)
{
struct smc_link *lnk = (struct smc_link *)priv;
struct smc_ib_device *smcibdev = lnk->smcibdev;
u8 port_idx;
switch (ibevent->event) {
case IB_EVENT_QP_FATAL:
case IB_EVENT_QP_ACCESS_ERR:
port_idx = ibevent->element.qp->port - 1;
if (port_idx >= SMC_MAX_PORTS)
break;
set_bit(port_idx, &smcibdev->port_event_mask);
if (!test_and_set_bit(port_idx, smcibdev->ports_going_away))
schedule_work(&smcibdev->port_event_work);
break;
default:
break;
}
}
void smc_ib_destroy_queue_pair(struct smc_link *lnk)
{
if (lnk->roce_qp)
ib_destroy_qp(lnk->roce_qp);
lnk->roce_qp = NULL;
}
/* create a queue pair within the protection domain for a link */
int smc_ib_create_queue_pair(struct smc_link *lnk)
{
struct ib_qp_init_attr qp_attr = {
.event_handler = smc_ib_qp_event_handler,
.qp_context = lnk,
.send_cq = lnk->smcibdev->roce_cq_send,
.recv_cq = lnk->smcibdev->roce_cq_recv,
.srq = NULL,
.cap = {
/* include unsolicited rdma_writes as well,
* there are max. 2 RDMA_WRITE per 1 WR_SEND
*/
.max_send_wr = SMC_WR_BUF_CNT * 3,
.max_recv_wr = SMC_WR_BUF_CNT * 3,
.max_send_sge = SMC_IB_MAX_SEND_SGE,
.max_recv_sge = 1,
},
.sq_sig_type = IB_SIGNAL_REQ_WR,
.qp_type = IB_QPT_RC,
};
int rc;
lnk->roce_qp = ib_create_qp(lnk->roce_pd, &qp_attr);
rc = PTR_ERR_OR_ZERO(lnk->roce_qp);
if (IS_ERR(lnk->roce_qp))
lnk->roce_qp = NULL;
else
smc_wr_remember_qp_attr(lnk);
return rc;
}
void smc_ib_put_memory_region(struct ib_mr *mr)
{
ib_dereg_mr(mr);
}
static int smc_ib_map_mr_sg(struct smc_buf_desc *buf_slot, u8 link_idx)
{
unsigned int offset = 0;
int sg_num;
/* map the largest prefix of a dma mapped SG list */
sg_num = ib_map_mr_sg(buf_slot->mr_rx[link_idx],
buf_slot->sgt[link_idx].sgl,
buf_slot->sgt[link_idx].orig_nents,
&offset, PAGE_SIZE);
return sg_num;
}
/* Allocate a memory region and map the dma mapped SG list of buf_slot */
int smc_ib_get_memory_region(struct ib_pd *pd, int access_flags,
struct smc_buf_desc *buf_slot, u8 link_idx)
{
if (buf_slot->mr_rx[link_idx])
return 0; /* already done */
buf_slot->mr_rx[link_idx] =
ib_alloc_mr(pd, IB_MR_TYPE_MEM_REG, 1 << buf_slot->order);
if (IS_ERR(buf_slot->mr_rx[link_idx])) {
int rc;
rc = PTR_ERR(buf_slot->mr_rx[link_idx]);
buf_slot->mr_rx[link_idx] = NULL;
return rc;
}
if (smc_ib_map_mr_sg(buf_slot, link_idx) != 1)
return -EINVAL;
return 0;
}
/* synchronize buffer usage for cpu access */
void smc_ib_sync_sg_for_cpu(struct smc_link *lnk,
struct smc_buf_desc *buf_slot,
enum dma_data_direction data_direction)
{
struct scatterlist *sg;
unsigned int i;
/* for now there is just one DMA address */
for_each_sg(buf_slot->sgt[lnk->link_idx].sgl, sg,
buf_slot->sgt[lnk->link_idx].nents, i) {
if (!sg_dma_len(sg))
break;
ib_dma_sync_single_for_cpu(lnk->smcibdev->ibdev,
sg_dma_address(sg),
sg_dma_len(sg),
data_direction);
}
}
/* synchronize buffer usage for device access */
void smc_ib_sync_sg_for_device(struct smc_link *lnk,
struct smc_buf_desc *buf_slot,
enum dma_data_direction data_direction)
{
struct scatterlist *sg;
unsigned int i;
/* for now there is just one DMA address */
for_each_sg(buf_slot->sgt[lnk->link_idx].sgl, sg,
buf_slot->sgt[lnk->link_idx].nents, i) {
if (!sg_dma_len(sg))
break;
ib_dma_sync_single_for_device(lnk->smcibdev->ibdev,
sg_dma_address(sg),
sg_dma_len(sg),
data_direction);
}
}
/* Map a new TX or RX buffer SG-table to DMA */
int smc_ib_buf_map_sg(struct smc_link *lnk,
struct smc_buf_desc *buf_slot,
enum dma_data_direction data_direction)
{
int mapped_nents;
mapped_nents = ib_dma_map_sg(lnk->smcibdev->ibdev,
buf_slot->sgt[lnk->link_idx].sgl,
buf_slot->sgt[lnk->link_idx].orig_nents,
data_direction);
if (!mapped_nents)
return -ENOMEM;
return mapped_nents;
}
void smc_ib_buf_unmap_sg(struct smc_link *lnk,
struct smc_buf_desc *buf_slot,
enum dma_data_direction data_direction)
{
if (!buf_slot->sgt[lnk->link_idx].sgl->dma_address)
return; /* already unmapped */
ib_dma_unmap_sg(lnk->smcibdev->ibdev,
buf_slot->sgt[lnk->link_idx].sgl,
buf_slot->sgt[lnk->link_idx].orig_nents,
data_direction);
buf_slot->sgt[lnk->link_idx].sgl->dma_address = 0;
}
long smc_ib_setup_per_ibdev(struct smc_ib_device *smcibdev)
{
struct ib_cq_init_attr cqattr = {
.cqe = SMC_MAX_CQE, .comp_vector = 0 };
int cqe_size_order, smc_order;
long rc;
mutex_lock(&smcibdev->mutex);
rc = 0;
if (smcibdev->initialized)
goto out;
/* the calculated number of cq entries fits to mlx5 cq allocation */
cqe_size_order = cache_line_size() == 128 ? 7 : 6;
smc_order = MAX_ORDER - cqe_size_order - 1;
if (SMC_MAX_CQE + 2 > (0x00000001 << smc_order) * PAGE_SIZE)
cqattr.cqe = (0x00000001 << smc_order) * PAGE_SIZE - 2;
smcibdev->roce_cq_send = ib_create_cq(smcibdev->ibdev,
smc_wr_tx_cq_handler, NULL,
smcibdev, &cqattr);
rc = PTR_ERR_OR_ZERO(smcibdev->roce_cq_send);
if (IS_ERR(smcibdev->roce_cq_send)) {
smcibdev->roce_cq_send = NULL;
goto out;
}
smcibdev->roce_cq_recv = ib_create_cq(smcibdev->ibdev,
smc_wr_rx_cq_handler, NULL,
smcibdev, &cqattr);
rc = PTR_ERR_OR_ZERO(smcibdev->roce_cq_recv);
if (IS_ERR(smcibdev->roce_cq_recv)) {
smcibdev->roce_cq_recv = NULL;
goto err;
}
smc_wr_add_dev(smcibdev);
smcibdev->initialized = 1;
goto out;
err:
ib_destroy_cq(smcibdev->roce_cq_send);
out:
mutex_unlock(&smcibdev->mutex);
return rc;
}
static void smc_ib_cleanup_per_ibdev(struct smc_ib_device *smcibdev)
{
mutex_lock(&smcibdev->mutex);
if (!smcibdev->initialized)
goto out;
smcibdev->initialized = 0;
ib_destroy_cq(smcibdev->roce_cq_recv);
ib_destroy_cq(smcibdev->roce_cq_send);
smc_wr_remove_dev(smcibdev);
out:
mutex_unlock(&smcibdev->mutex);
}
static struct ib_client smc_ib_client;
/* callback function for ib_register_client() */
static int smc_ib_add_dev(struct ib_device *ibdev)
{
struct smc_ib_device *smcibdev;
u8 port_cnt;
int i;
if (ibdev->node_type != RDMA_NODE_IB_CA)
return -EOPNOTSUPP;
smcibdev = kzalloc(sizeof(*smcibdev), GFP_KERNEL);
if (!smcibdev)
return -ENOMEM;
smcibdev->ibdev = ibdev;
INIT_WORK(&smcibdev->port_event_work, smc_ib_port_event_work);
atomic_set(&smcibdev->lnk_cnt, 0);
init_waitqueue_head(&smcibdev->lnks_deleted);
mutex_init(&smcibdev->mutex);
mutex_lock(&smc_ib_devices.mutex);
list_add_tail(&smcibdev->list, &smc_ib_devices.list);
mutex_unlock(&smc_ib_devices.mutex);
ib_set_client_data(ibdev, &smc_ib_client, smcibdev);
INIT_IB_EVENT_HANDLER(&smcibdev->event_handler, smcibdev->ibdev,
smc_ib_global_event_handler);
ib_register_event_handler(&smcibdev->event_handler);
/* trigger reading of the port attributes */
port_cnt = smcibdev->ibdev->phys_port_cnt;
pr_warn_ratelimited("smc: adding ib device %s with port count %d\n",
smcibdev->ibdev->name, port_cnt);
for (i = 0;
i < min_t(size_t, port_cnt, SMC_MAX_PORTS);
i++) {
set_bit(i, &smcibdev->port_event_mask);
/* determine pnetids of the port */
if (smc_pnetid_by_dev_port(ibdev->dev.parent, i,
smcibdev->pnetid[i]))
smc_pnetid_by_table_ib(smcibdev, i + 1);
pr_warn_ratelimited("smc: ib device %s port %d has pnetid "
"%.16s%s\n",
smcibdev->ibdev->name, i + 1,
smcibdev->pnetid[i],
smcibdev->pnetid_by_user[i] ?
" (user defined)" :
"");
}
schedule_work(&smcibdev->port_event_work);
return 0;
}
/* callback function for ib_unregister_client() */
static void smc_ib_remove_dev(struct ib_device *ibdev, void *client_data)
{
struct smc_ib_device *smcibdev = client_data;
mutex_lock(&smc_ib_devices.mutex);
list_del_init(&smcibdev->list); /* remove from smc_ib_devices */
mutex_unlock(&smc_ib_devices.mutex);
pr_warn_ratelimited("smc: removing ib device %s\n",
smcibdev->ibdev->name);
smc_smcr_terminate_all(smcibdev);
smc_ib_cleanup_per_ibdev(smcibdev);
ib_unregister_event_handler(&smcibdev->event_handler);
cancel_work_sync(&smcibdev->port_event_work);
kfree(smcibdev);
}
static struct ib_client smc_ib_client = {
.name = "smc_ib",
.add = smc_ib_add_dev,
.remove = smc_ib_remove_dev,
};
int __init smc_ib_register_client(void)
{
smc_ib_init_local_systemid();
return ib_register_client(&smc_ib_client);
}
void smc_ib_unregister_client(void)
{
ib_unregister_client(&smc_ib_client);
}