xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

mlx5: Add driver for Mellanox Connect-IB adapters

Browse Source 
The driver is comprised of two kernel modules: mlx5_ib and mlx5_core.
This partitioning resembles what we have for mlx4, except that mlx5_ib
is the pci device driver and not mlx5_core.

mlx5_core is essentially a library that provides general functionality
that is intended to be used by other Mellanox devices that will be
introduced in the future.  mlx5_ib has a similar role as any hardware
device under drivers/infiniband/hw.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>

[ Merge in coccinelle fixes from Fengguang Wu <fengguang.wu@intel.com>.
  - Roland ]

Signed-off-by: Roland Dreier <roland@purestorage.com>
This commit is contained in:
Eli Cohen
2013-07-07 17:25:49 +03:00
committed by Roland Dreier
parent 0134f16bc9
commit e126ba97db
45 changed files with 15779 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
drivers/infiniband/hw/mlx5/Kconfig Normal file
View File

@@ -0,0 +1,10 @@
config MLX5_INFINIBAND
tristate "Mellanox Connect-IB HCA support"
depends on NETDEVICES && ETHERNET && PCI && X86
select NET_VENDOR_MELLANOX
select MLX5_CORE
---help---
This driver provides low-level InfiniBand support for
Mellanox Connect-IB PCI Express host channel adapters (HCAs).
This is required to use InfiniBand protocols such as
IP-over-IB or SRP with these devices.

3
drivers/infiniband/hw/mlx5/Makefile Normal file
View File

@@ -0,0 +1,3 @@
obj-$(CONFIG_MLX5_INFINIBAND) += mlx5_ib.o
mlx5_ib-y := main.o cq.o doorbell.o qp.o mem.o srq.o mr.o ah.o mad.o

92
drivers/infiniband/hw/mlx5/ah.c Normal file
View File

@@ -0,0 +1,92 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include "mlx5_ib.h"
struct ib_ah *create_ib_ah(struct ib_ah_attr *ah_attr,
struct mlx5_ib_ah *ah)
{
if (ah_attr->ah_flags & IB_AH_GRH) {
memcpy(ah->av.rgid, &ah_attr->grh.dgid, 16);
ah->av.grh_gid_fl = cpu_to_be32(ah_attr->grh.flow_label |
(1 << 30) |
ah_attr->grh.sgid_index << 20);
ah->av.hop_limit = ah_attr->grh.hop_limit;
ah->av.tclass = ah_attr->grh.traffic_class;
}
ah->av.rlid = cpu_to_be16(ah_attr->dlid);
ah->av.fl_mlid = ah_attr->src_path_bits & 0x7f;
ah->av.stat_rate_sl = (ah_attr->static_rate << 4) | (ah_attr->sl & 0xf);
return &ah->ibah;
}
struct ib_ah *mlx5_ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr)
{
struct mlx5_ib_ah *ah;
ah = kzalloc(sizeof(*ah), GFP_ATOMIC);
if (!ah)
return ERR_PTR(-ENOMEM);
return create_ib_ah(ah_attr, ah); /* never fails */
}
int mlx5_ib_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr)
{
struct mlx5_ib_ah *ah = to_mah(ibah);
u32 tmp;
memset(ah_attr, 0, sizeof(*ah_attr));
tmp = be32_to_cpu(ah->av.grh_gid_fl);
if (tmp & (1 << 30)) {
ah_attr->ah_flags = IB_AH_GRH;
ah_attr->grh.sgid_index = (tmp >> 20) & 0xff;
ah_attr->grh.flow_label = tmp & 0xfffff;
memcpy(&ah_attr->grh.dgid, ah->av.rgid, 16);
ah_attr->grh.hop_limit = ah->av.hop_limit;
ah_attr->grh.traffic_class = ah->av.tclass;
}
ah_attr->dlid = be16_to_cpu(ah->av.rlid);
ah_attr->static_rate = ah->av.stat_rate_sl >> 4;
ah_attr->sl = ah->av.stat_rate_sl & 0xf;
return 0;
}
int mlx5_ib_destroy_ah(struct ib_ah *ah)
{
kfree(to_mah(ah));
return 0;
}

843
drivers/infiniband/hw/mlx5/cq.c Normal file
View File

@@ -0,0 +1,843 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/kref.h>
#include <rdma/ib_umem.h>
#include "mlx5_ib.h"
#include "user.h"
static void mlx5_ib_cq_comp(struct mlx5_core_cq *cq)
{
struct ib_cq *ibcq = &to_mibcq(cq)->ibcq;
ibcq->comp_handler(ibcq, ibcq->cq_context);
}
static void mlx5_ib_cq_event(struct mlx5_core_cq *mcq, enum mlx5_event type)
{
struct mlx5_ib_cq *cq = container_of(mcq, struct mlx5_ib_cq, mcq);
struct mlx5_ib_dev *dev = to_mdev(cq->ibcq.device);
struct ib_cq *ibcq = &cq->ibcq;
struct ib_event event;
if (type != MLX5_EVENT_TYPE_CQ_ERROR) {
mlx5_ib_warn(dev, "Unexpected event type %d on CQ %06x\n",
type, mcq->cqn);
return;
}
if (ibcq->event_handler) {
event.device = &dev->ib_dev;
event.event = IB_EVENT_CQ_ERR;
event.element.cq = ibcq;
ibcq->event_handler(&event, ibcq->cq_context);
}
}
static void *get_cqe_from_buf(struct mlx5_ib_cq_buf *buf, int n, int size)
{
return mlx5_buf_offset(&buf->buf, n * size);
}
static void *get_cqe(struct mlx5_ib_cq *cq, int n)
{
return get_cqe_from_buf(&cq->buf, n, cq->mcq.cqe_sz);
}
static void *get_sw_cqe(struct mlx5_ib_cq *cq, int n)
{
void *cqe = get_cqe(cq, n & cq->ibcq.cqe);
struct mlx5_cqe64 *cqe64;
cqe64 = (cq->mcq.cqe_sz == 64) ? cqe : cqe + 64;
return ((cqe64->op_own & MLX5_CQE_OWNER_MASK) ^
!!(n & (cq->ibcq.cqe + 1))) ? NULL : cqe;
}
static void *next_cqe_sw(struct mlx5_ib_cq *cq)
{
return get_sw_cqe(cq, cq->mcq.cons_index);
}
static enum ib_wc_opcode get_umr_comp(struct mlx5_ib_wq *wq, int idx)
{
switch (wq->wr_data[idx]) {
case MLX5_IB_WR_UMR:
return 0;
case IB_WR_LOCAL_INV:
return IB_WC_LOCAL_INV;
case IB_WR_FAST_REG_MR:
return IB_WC_FAST_REG_MR;
default:
pr_warn("unknown completion status\n");
return 0;
}
}
static void handle_good_req(struct ib_wc *wc, struct mlx5_cqe64 *cqe,
struct mlx5_ib_wq *wq, int idx)
{
wc->wc_flags = 0;
switch (be32_to_cpu(cqe->sop_drop_qpn) >> 24) {
case MLX5_OPCODE_RDMA_WRITE_IMM:
wc->wc_flags |= IB_WC_WITH_IMM;
case MLX5_OPCODE_RDMA_WRITE:
wc->opcode = IB_WC_RDMA_WRITE;
break;
case MLX5_OPCODE_SEND_IMM:
wc->wc_flags |= IB_WC_WITH_IMM;
case MLX5_OPCODE_SEND:
case MLX5_OPCODE_SEND_INVAL:
wc->opcode = IB_WC_SEND;
break;
case MLX5_OPCODE_RDMA_READ:
wc->opcode = IB_WC_RDMA_READ;
wc->byte_len = be32_to_cpu(cqe->byte_cnt);
break;
case MLX5_OPCODE_ATOMIC_CS:
wc->opcode = IB_WC_COMP_SWAP;
wc->byte_len = 8;
break;
case MLX5_OPCODE_ATOMIC_FA:
wc->opcode = IB_WC_FETCH_ADD;
wc->byte_len = 8;
break;
case MLX5_OPCODE_ATOMIC_MASKED_CS:
wc->opcode = IB_WC_MASKED_COMP_SWAP;
wc->byte_len = 8;
break;
case MLX5_OPCODE_ATOMIC_MASKED_FA:
wc->opcode = IB_WC_MASKED_FETCH_ADD;
wc->byte_len = 8;
break;
case MLX5_OPCODE_BIND_MW:
wc->opcode = IB_WC_BIND_MW;
break;
case MLX5_OPCODE_UMR:
wc->opcode = get_umr_comp(wq, idx);
break;
}
}
enum {
MLX5_GRH_IN_BUFFER = 1,
MLX5_GRH_IN_CQE = 2,
};
static void handle_responder(struct ib_wc *wc, struct mlx5_cqe64 *cqe,
struct mlx5_ib_qp *qp)
{
struct mlx5_ib_dev *dev = to_mdev(qp->ibqp.device);
struct mlx5_ib_srq *srq;
struct mlx5_ib_wq *wq;
u16 wqe_ctr;
u8 g;
if (qp->ibqp.srq || qp->ibqp.xrcd) {
struct mlx5_core_srq *msrq = NULL;
if (qp->ibqp.xrcd) {
msrq = mlx5_core_get_srq(&dev->mdev,
be32_to_cpu(cqe->srqn));
srq = to_mibsrq(msrq);
} else {
srq = to_msrq(qp->ibqp.srq);
}
if (srq) {
wqe_ctr = be16_to_cpu(cqe->wqe_counter);
wc->wr_id = srq->wrid[wqe_ctr];
mlx5_ib_free_srq_wqe(srq, wqe_ctr);
if (msrq && atomic_dec_and_test(&msrq->refcount))
complete(&msrq->free);
}
} else {
wq = &qp->rq;
wc->wr_id = wq->wrid[wq->tail & (wq->wqe_cnt - 1)];
++wq->tail;
}
wc->byte_len = be32_to_cpu(cqe->byte_cnt);
switch (cqe->op_own >> 4) {
case MLX5_CQE_RESP_WR_IMM:
wc->opcode = IB_WC_RECV_RDMA_WITH_IMM;
wc->wc_flags = IB_WC_WITH_IMM;
wc->ex.imm_data = cqe->imm_inval_pkey;
break;
case MLX5_CQE_RESP_SEND:
wc->opcode = IB_WC_RECV;
wc->wc_flags = 0;
break;
case MLX5_CQE_RESP_SEND_IMM:
wc->opcode = IB_WC_RECV;
wc->wc_flags = IB_WC_WITH_IMM;
wc->ex.imm_data = cqe->imm_inval_pkey;
break;
case MLX5_CQE_RESP_SEND_INV:
wc->opcode = IB_WC_RECV;
wc->wc_flags = IB_WC_WITH_INVALIDATE;
wc->ex.invalidate_rkey = be32_to_cpu(cqe->imm_inval_pkey);
break;
}
wc->slid = be16_to_cpu(cqe->slid);
wc->sl = (be32_to_cpu(cqe->flags_rqpn) >> 24) & 0xf;
wc->src_qp = be32_to_cpu(cqe->flags_rqpn) & 0xffffff;
wc->dlid_path_bits = cqe->ml_path;
g = (be32_to_cpu(cqe->flags_rqpn) >> 28) & 3;
wc->wc_flags |= g ? IB_WC_GRH : 0;
wc->pkey_index = be32_to_cpu(cqe->imm_inval_pkey) & 0xffff;
}
static void dump_cqe(struct mlx5_ib_dev *dev, struct mlx5_err_cqe *cqe)
{
__be32 *p = (__be32 *)cqe;
int i;
mlx5_ib_warn(dev, "dump error cqe\n");
for (i = 0; i < sizeof(*cqe) / 16; i++, p += 4)
pr_info("%08x %08x %08x %08x\n", be32_to_cpu(p[0]),
be32_to_cpu(p[1]), be32_to_cpu(p[2]),
be32_to_cpu(p[3]));
}
static void mlx5_handle_error_cqe(struct mlx5_ib_dev *dev,
struct mlx5_err_cqe *cqe,
struct ib_wc *wc)
{
int dump = 1;
switch (cqe->syndrome) {
case MLX5_CQE_SYNDROME_LOCAL_LENGTH_ERR:
wc->status = IB_WC_LOC_LEN_ERR;
break;
case MLX5_CQE_SYNDROME_LOCAL_QP_OP_ERR:
wc->status = IB_WC_LOC_QP_OP_ERR;
break;
case MLX5_CQE_SYNDROME_LOCAL_PROT_ERR:
wc->status = IB_WC_LOC_PROT_ERR;
break;
case MLX5_CQE_SYNDROME_WR_FLUSH_ERR:
dump = 0;
wc->status = IB_WC_WR_FLUSH_ERR;
break;
case MLX5_CQE_SYNDROME_MW_BIND_ERR:
wc->status = IB_WC_MW_BIND_ERR;
break;
case MLX5_CQE_SYNDROME_BAD_RESP_ERR:
wc->status = IB_WC_BAD_RESP_ERR;
break;
case MLX5_CQE_SYNDROME_LOCAL_ACCESS_ERR:
wc->status = IB_WC_LOC_ACCESS_ERR;
break;
case MLX5_CQE_SYNDROME_REMOTE_INVAL_REQ_ERR:
wc->status = IB_WC_REM_INV_REQ_ERR;
break;
case MLX5_CQE_SYNDROME_REMOTE_ACCESS_ERR:
wc->status = IB_WC_REM_ACCESS_ERR;
break;
case MLX5_CQE_SYNDROME_REMOTE_OP_ERR:
wc->status = IB_WC_REM_OP_ERR;
break;
case MLX5_CQE_SYNDROME_TRANSPORT_RETRY_EXC_ERR:
wc->status = IB_WC_RETRY_EXC_ERR;
dump = 0;
break;
case MLX5_CQE_SYNDROME_RNR_RETRY_EXC_ERR:
wc->status = IB_WC_RNR_RETRY_EXC_ERR;
dump = 0;
break;
case MLX5_CQE_SYNDROME_REMOTE_ABORTED_ERR:
wc->status = IB_WC_REM_ABORT_ERR;
break;
default:
wc->status = IB_WC_GENERAL_ERR;
break;
}
wc->vendor_err = cqe->vendor_err_synd;
if (dump)
dump_cqe(dev, cqe);
}
static int is_atomic_response(struct mlx5_ib_qp *qp, uint16_t idx)
{
/* TBD: waiting decision
*/
return 0;
}
static void *mlx5_get_atomic_laddr(struct mlx5_ib_qp *qp, uint16_t idx)
{
struct mlx5_wqe_data_seg *dpseg;
void *addr;
dpseg = mlx5_get_send_wqe(qp, idx) + sizeof(struct mlx5_wqe_ctrl_seg) +
sizeof(struct mlx5_wqe_raddr_seg) +
sizeof(struct mlx5_wqe_atomic_seg);
addr = (void *)(unsigned long)be64_to_cpu(dpseg->addr);
return addr;
}
static void handle_atomic(struct mlx5_ib_qp *qp, struct mlx5_cqe64 *cqe64,
uint16_t idx)
{
void *addr;
int byte_count;
int i;
if (!is_atomic_response(qp, idx))
return;
byte_count = be32_to_cpu(cqe64->byte_cnt);
addr = mlx5_get_atomic_laddr(qp, idx);
if (byte_count == 4) {
*(uint32_t *)addr = be32_to_cpu(*((__be32 *)addr));
} else {
for (i = 0; i < byte_count; i += 8) {
*(uint64_t *)addr = be64_to_cpu(*((__be64 *)addr));
addr += 8;
}
}
return;
}
static void handle_atomics(struct mlx5_ib_qp *qp, struct mlx5_cqe64 *cqe64,
u16 tail, u16 head)
{
int idx;
do {
idx = tail & (qp->sq.wqe_cnt - 1);
handle_atomic(qp, cqe64, idx);
if (idx == head)
break;
tail = qp->sq.w_list[idx].next;
} while (1);
tail = qp->sq.w_list[idx].next;
qp->sq.last_poll = tail;
}
static int mlx5_poll_one(struct mlx5_ib_cq *cq,
struct mlx5_ib_qp **cur_qp,
struct ib_wc *wc)
{
struct mlx5_ib_dev *dev = to_mdev(cq->ibcq.device);
struct mlx5_err_cqe *err_cqe;
struct mlx5_cqe64 *cqe64;
struct mlx5_core_qp *mqp;
struct mlx5_ib_wq *wq;
uint8_t opcode;
uint32_t qpn;
u16 wqe_ctr;
void *cqe;
int idx;
cqe = next_cqe_sw(cq);
if (!cqe)
return -EAGAIN;
cqe64 = (cq->mcq.cqe_sz == 64) ? cqe : cqe + 64;
++cq->mcq.cons_index;
/* Make sure we read CQ entry contents after we've checked the
* ownership bit.
*/
rmb();
/* TBD: resize CQ */
qpn = ntohl(cqe64->sop_drop_qpn) & 0xffffff;
if (!*cur_qp || (qpn != (*cur_qp)->ibqp.qp_num)) {
/* We do not have to take the QP table lock here,
* because CQs will be locked while QPs are removed
* from the table.
*/
mqp = __mlx5_qp_lookup(&dev->mdev, qpn);
if (unlikely(!mqp)) {
mlx5_ib_warn(dev, "CQE@CQ %06x for unknown QPN %6x\n",
cq->mcq.cqn, qpn);
return -EINVAL;
}
*cur_qp = to_mibqp(mqp);
}
wc->qp = &(*cur_qp)->ibqp;
opcode = cqe64->op_own >> 4;
switch (opcode) {
case MLX5_CQE_REQ:
wq = &(*cur_qp)->sq;
wqe_ctr = be16_to_cpu(cqe64->wqe_counter);
idx = wqe_ctr & (wq->wqe_cnt - 1);
handle_good_req(wc, cqe64, wq, idx);
handle_atomics(*cur_qp, cqe64, wq->last_poll, idx);
wc->wr_id = wq->wrid[idx];
wq->tail = wq->wqe_head[idx] + 1;
wc->status = IB_WC_SUCCESS;
break;
case MLX5_CQE_RESP_WR_IMM:
case MLX5_CQE_RESP_SEND:
case MLX5_CQE_RESP_SEND_IMM:
case MLX5_CQE_RESP_SEND_INV:
handle_responder(wc, cqe64, *cur_qp);
wc->status = IB_WC_SUCCESS;
break;
case MLX5_CQE_RESIZE_CQ:
break;
case MLX5_CQE_REQ_ERR:
case MLX5_CQE_RESP_ERR:
err_cqe = (struct mlx5_err_cqe *)cqe64;
mlx5_handle_error_cqe(dev, err_cqe, wc);
mlx5_ib_dbg(dev, "%s error cqe on cqn 0x%x:\n",
opcode == MLX5_CQE_REQ_ERR ?
"Requestor" : "Responder", cq->mcq.cqn);
mlx5_ib_dbg(dev, "syndrome 0x%x, vendor syndrome 0x%x\n",
err_cqe->syndrome, err_cqe->vendor_err_synd);
if (opcode == MLX5_CQE_REQ_ERR) {
wq = &(*cur_qp)->sq;
wqe_ctr = be16_to_cpu(cqe64->wqe_counter);
idx = wqe_ctr & (wq->wqe_cnt - 1);
wc->wr_id = wq->wrid[idx];
wq->tail = wq->wqe_head[idx] + 1;
} else {
struct mlx5_ib_srq *srq;
if ((*cur_qp)->ibqp.srq) {
srq = to_msrq((*cur_qp)->ibqp.srq);
wqe_ctr = be16_to_cpu(cqe64->wqe_counter);
wc->wr_id = srq->wrid[wqe_ctr];
mlx5_ib_free_srq_wqe(srq, wqe_ctr);
} else {
wq = &(*cur_qp)->rq;
wc->wr_id = wq->wrid[wq->tail & (wq->wqe_cnt - 1)];
++wq->tail;
}
}
break;
}
return 0;
}
int mlx5_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc)
{
struct mlx5_ib_cq *cq = to_mcq(ibcq);
struct mlx5_ib_qp *cur_qp = NULL;
unsigned long flags;
int npolled;
int err = 0;
spin_lock_irqsave(&cq->lock, flags);
for (npolled = 0; npolled < num_entries; npolled++) {
err = mlx5_poll_one(cq, &cur_qp, wc + npolled);
if (err)
break;
}
if (npolled)
mlx5_cq_set_ci(&cq->mcq);
spin_unlock_irqrestore(&cq->lock, flags);
if (err == 0 || err == -EAGAIN)
return npolled;
else
return err;
}
int mlx5_ib_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags flags)
{
mlx5_cq_arm(&to_mcq(ibcq)->mcq,
(flags & IB_CQ_SOLICITED_MASK) == IB_CQ_SOLICITED ?
MLX5_CQ_DB_REQ_NOT_SOL : MLX5_CQ_DB_REQ_NOT,
to_mdev(ibcq->device)->mdev.priv.uuari.uars[0].map,
MLX5_GET_DOORBELL_LOCK(&to_mdev(ibcq->device)->mdev.priv.cq_uar_lock));
return 0;
}
static int alloc_cq_buf(struct mlx5_ib_dev *dev, struct mlx5_ib_cq_buf *buf,
int nent, int cqe_size)
{
int err;
err = mlx5_buf_alloc(&dev->mdev, nent * cqe_size,
PAGE_SIZE * 2, &buf->buf);
if (err)
return err;
buf->cqe_size = cqe_size;
return 0;
}
static void free_cq_buf(struct mlx5_ib_dev *dev, struct mlx5_ib_cq_buf *buf)
{
mlx5_buf_free(&dev->mdev, &buf->buf);
}
static int create_cq_user(struct mlx5_ib_dev *dev, struct ib_udata *udata,
struct ib_ucontext *context, struct mlx5_ib_cq *cq,
int entries, struct mlx5_create_cq_mbox_in **cqb,
int *cqe_size, int *index, int *inlen)
{
struct mlx5_ib_create_cq ucmd;
int page_shift;
int npages;
int ncont;
int err;
if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd)))
return -EFAULT;
if (ucmd.cqe_size != 64 && ucmd.cqe_size != 128)
return -EINVAL;
*cqe_size = ucmd.cqe_size;
cq->buf.umem = ib_umem_get(context, ucmd.buf_addr,
entries * ucmd.cqe_size,
IB_ACCESS_LOCAL_WRITE, 1);
if (IS_ERR(cq->buf.umem)) {
err = PTR_ERR(cq->buf.umem);
return err;
}
err = mlx5_ib_db_map_user(to_mucontext(context), ucmd.db_addr,
&cq->db);
if (err)
goto err_umem;
mlx5_ib_cont_pages(cq->buf.umem, ucmd.buf_addr, &npages, &page_shift,
&ncont, NULL);
mlx5_ib_dbg(dev, "addr 0x%llx, size %u, npages %d, page_shift %d, ncont %d\n",
ucmd.buf_addr, entries * ucmd.cqe_size, npages, page_shift, ncont);
*inlen = sizeof(**cqb) + sizeof(*(*cqb)->pas) * ncont;
*cqb = mlx5_vzalloc(*inlen);
if (!*cqb) {
err = -ENOMEM;
goto err_db;
}
mlx5_ib_populate_pas(dev, cq->buf.umem, page_shift, (*cqb)->pas, 0);
(*cqb)->ctx.log_pg_sz = page_shift - PAGE_SHIFT;
*index = to_mucontext(context)->uuari.uars[0].index;
return 0;
err_db:
mlx5_ib_db_unmap_user(to_mucontext(context), &cq->db);
err_umem:
ib_umem_release(cq->buf.umem);
return err;
}
static void destroy_cq_user(struct mlx5_ib_cq *cq, struct ib_ucontext *context)
{
mlx5_ib_db_unmap_user(to_mucontext(context), &cq->db);
ib_umem_release(cq->buf.umem);
}
static void init_cq_buf(struct mlx5_ib_cq *cq, int nent)
{
int i;
void *cqe;
struct mlx5_cqe64 *cqe64;
for (i = 0; i < nent; i++) {
cqe = get_cqe(cq, i);
cqe64 = (cq->buf.cqe_size == 64) ? cqe : cqe + 64;
cqe64->op_own = 0xf1;
}
}
static int create_cq_kernel(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *cq,
int entries, int cqe_size,
struct mlx5_create_cq_mbox_in **cqb,
int *index, int *inlen)
{
int err;
err = mlx5_db_alloc(&dev->mdev, &cq->db);
if (err)
return err;
cq->mcq.set_ci_db = cq->db.db;
cq->mcq.arm_db = cq->db.db + 1;
*cq->mcq.set_ci_db = 0;
*cq->mcq.arm_db = 0;
cq->mcq.cqe_sz = cqe_size;
err = alloc_cq_buf(dev, &cq->buf, entries, cqe_size);
if (err)
goto err_db;
init_cq_buf(cq, entries);
*inlen = sizeof(**cqb) + sizeof(*(*cqb)->pas) * cq->buf.buf.npages;
*cqb = mlx5_vzalloc(*inlen);
if (!*cqb) {
err = -ENOMEM;
goto err_buf;
}
mlx5_fill_page_array(&cq->buf.buf, (*cqb)->pas);
(*cqb)->ctx.log_pg_sz = cq->buf.buf.page_shift - PAGE_SHIFT;
*index = dev->mdev.priv.uuari.uars[0].index;
return 0;
err_buf:
free_cq_buf(dev, &cq->buf);
err_db:
mlx5_db_free(&dev->mdev, &cq->db);
return err;
}
static void destroy_cq_kernel(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *cq)
{
free_cq_buf(dev, &cq->buf);
mlx5_db_free(&dev->mdev, &cq->db);
}
struct ib_cq *mlx5_ib_create_cq(struct ib_device *ibdev, int entries,
int vector, struct ib_ucontext *context,
struct ib_udata *udata)
{
struct mlx5_create_cq_mbox_in *cqb = NULL;
struct mlx5_ib_dev *dev = to_mdev(ibdev);
struct mlx5_ib_cq *cq;
int uninitialized_var(index);
int uninitialized_var(inlen);
int cqe_size;
int irqn;
int eqn;
int err;
entries = roundup_pow_of_two(entries + 1);
if (entries < 1 || entries > dev->mdev.caps.max_cqes)
return ERR_PTR(-EINVAL);
cq = kzalloc(sizeof(*cq), GFP_KERNEL);
if (!cq)
return ERR_PTR(-ENOMEM);
cq->ibcq.cqe = entries - 1;
mutex_init(&cq->resize_mutex);
spin_lock_init(&cq->lock);
cq->resize_buf = NULL;
cq->resize_umem = NULL;
if (context) {
err = create_cq_user(dev, udata, context, cq, entries,
&cqb, &cqe_size, &index, &inlen);
if (err)
goto err_create;
} else {
/* for now choose 64 bytes till we have a proper interface */
cqe_size = 64;
err = create_cq_kernel(dev, cq, entries, cqe_size, &cqb,
&index, &inlen);
if (err)
goto err_create;
}
cq->cqe_size = cqe_size;
cqb->ctx.cqe_sz_flags = cqe_sz_to_mlx_sz(cqe_size) << 5;
cqb->ctx.log_sz_usr_page = cpu_to_be32((ilog2(entries) << 24) | index);
err = mlx5_vector2eqn(dev, vector, &eqn, &irqn);
if (err)
goto err_cqb;
cqb->ctx.c_eqn = cpu_to_be16(eqn);
cqb->ctx.db_record_addr = cpu_to_be64(cq->db.dma);
err = mlx5_core_create_cq(&dev->mdev, &cq->mcq, cqb, inlen);
if (err)
goto err_cqb;
mlx5_ib_dbg(dev, "cqn 0x%x\n", cq->mcq.cqn);
cq->mcq.irqn = irqn;
cq->mcq.comp = mlx5_ib_cq_comp;
cq->mcq.event = mlx5_ib_cq_event;
if (context)
if (ib_copy_to_udata(udata, &cq->mcq.cqn, sizeof(__u32))) {
err = -EFAULT;
goto err_cmd;
}
mlx5_vfree(cqb);
return &cq->ibcq;
err_cmd:
mlx5_core_destroy_cq(&dev->mdev, &cq->mcq);
err_cqb:
mlx5_vfree(cqb);
if (context)
destroy_cq_user(cq, context);
else
destroy_cq_kernel(dev, cq);
err_create:
kfree(cq);
return ERR_PTR(err);
}
int mlx5_ib_destroy_cq(struct ib_cq *cq)
{
struct mlx5_ib_dev *dev = to_mdev(cq->device);
struct mlx5_ib_cq *mcq = to_mcq(cq);
struct ib_ucontext *context = NULL;
if (cq->uobject)
context = cq->uobject->context;
mlx5_core_destroy_cq(&dev->mdev, &mcq->mcq);
if (context)
destroy_cq_user(mcq, context);
else
destroy_cq_kernel(dev, mcq);
kfree(mcq);
return 0;
}
static int is_equal_rsn(struct mlx5_cqe64 *cqe64, struct mlx5_ib_srq *srq,
u32 rsn)
{
u32 lrsn;
if (srq)
lrsn = be32_to_cpu(cqe64->srqn) & 0xffffff;
else
lrsn = be32_to_cpu(cqe64->sop_drop_qpn) & 0xffffff;
return rsn == lrsn;
}
void __mlx5_ib_cq_clean(struct mlx5_ib_cq *cq, u32 rsn, struct mlx5_ib_srq *srq)
{
struct mlx5_cqe64 *cqe64, *dest64;
void *cqe, *dest;
u32 prod_index;
int nfreed = 0;
u8 owner_bit;
if (!cq)
return;
/* First we need to find the current producer index, so we
* know where to start cleaning from. It doesn't matter if HW
* adds new entries after this loop -- the QP we're worried
* about is already in RESET, so the new entries won't come
* from our QP and therefore don't need to be checked.
*/
for (prod_index = cq->mcq.cons_index; get_sw_cqe(cq, prod_index); prod_index++)
if (prod_index == cq->mcq.cons_index + cq->ibcq.cqe)
break;
/* Now sweep backwards through the CQ, removing CQ entries
* that match our QP by copying older entries on top of them.
*/
while ((int) --prod_index - (int) cq->mcq.cons_index >= 0) {
cqe = get_cqe(cq, prod_index & cq->ibcq.cqe);
cqe64 = (cq->mcq.cqe_sz == 64) ? cqe : cqe + 64;
if (is_equal_rsn(cqe64, srq, rsn)) {
if (srq)
mlx5_ib_free_srq_wqe(srq, be16_to_cpu(cqe64->wqe_counter));
++nfreed;
} else if (nfreed) {
dest = get_cqe(cq, (prod_index + nfreed) & cq->ibcq.cqe);
dest64 = (cq->mcq.cqe_sz == 64) ? dest : dest + 64;
owner_bit = dest64->op_own & MLX5_CQE_OWNER_MASK;
memcpy(dest, cqe, cq->mcq.cqe_sz);
dest64->op_own = owner_bit |
(dest64->op_own & ~MLX5_CQE_OWNER_MASK);
}
}
if (nfreed) {
cq->mcq.cons_index += nfreed;
/* Make sure update of buffer contents is done before
* updating consumer index.
*/
wmb();
mlx5_cq_set_ci(&cq->mcq);
}
}
void mlx5_ib_cq_clean(struct mlx5_ib_cq *cq, u32 qpn, struct mlx5_ib_srq *srq)
{
if (!cq)
return;
spin_lock_irq(&cq->lock);
__mlx5_ib_cq_clean(cq, qpn, srq);
spin_unlock_irq(&cq->lock);
}
int mlx5_ib_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period)
{
return -ENOSYS;
}
int mlx5_ib_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata)
{
return -ENOSYS;
}
int mlx5_ib_get_cqe_size(struct mlx5_ib_dev *dev, struct ib_cq *ibcq)
{
struct mlx5_ib_cq *cq;
if (!ibcq)
return 128;
cq = to_mcq(ibcq);
return cq->cqe_size;
}

100
drivers/infiniband/hw/mlx5/doorbell.c Normal file
View File

@@ -0,0 +1,100 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/kref.h>
#include <linux/slab.h>
#include <rdma/ib_umem.h>
#include "mlx5_ib.h"
struct mlx5_ib_user_db_page {
struct list_head list;
struct ib_umem *umem;
unsigned long user_virt;
int refcnt;
};
int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
struct mlx5_db *db)
{
struct mlx5_ib_user_db_page *page;
struct ib_umem_chunk *chunk;
int err = 0;
mutex_lock(&context->db_page_mutex);
list_for_each_entry(page, &context->db_page_list, list)
if (page->user_virt == (virt & PAGE_MASK))
goto found;
page = kmalloc(sizeof(*page), GFP_KERNEL);
if (!page) {
err = -ENOMEM;
goto out;
}
page->user_virt = (virt & PAGE_MASK);
page->refcnt = 0;
page->umem = ib_umem_get(&context->ibucontext, virt & PAGE_MASK,
PAGE_SIZE, 0, 0);
if (IS_ERR(page->umem)) {
err = PTR_ERR(page->umem);
kfree(page);
goto out;
}
list_add(&page->list, &context->db_page_list);
found:
chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
db->dma = sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
db->u.user_page = page;
++page->refcnt;
out:
mutex_unlock(&context->db_page_mutex);
return err;
}
void mlx5_ib_db_unmap_user(struct mlx5_ib_ucontext *context, struct mlx5_db *db)
{
mutex_lock(&context->db_page_mutex);
if (!--db->u.user_page->refcnt) {
list_del(&db->u.user_page->list);
ib_umem_release(db->u.user_page->umem);
kfree(db->u.user_page);
}
mutex_unlock(&context->db_page_mutex);
}

139
drivers/infiniband/hw/mlx5/mad.c Normal file
View File

@@ -0,0 +1,139 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/mlx5/cmd.h>
#include <rdma/ib_mad.h>
#include <rdma/ib_smi.h>
#include "mlx5_ib.h"
enum {
MLX5_IB_VENDOR_CLASS1 = 0x9,
MLX5_IB_VENDOR_CLASS2 = 0xa
};
int mlx5_MAD_IFC(struct mlx5_ib_dev *dev, int ignore_mkey, int ignore_bkey,
int port, struct ib_wc *in_wc, struct ib_grh *in_grh,
void *in_mad, void *response_mad)
{
u8 op_modifier = 0;
/* Key check traps can't be generated unless we have in_wc to
* tell us where to send the trap.
*/
if (ignore_mkey || !in_wc)
op_modifier |= 0x1;
if (ignore_bkey || !in_wc)
op_modifier |= 0x2;
return mlx5_core_mad_ifc(&dev->mdev, in_mad, response_mad, op_modifier, port);
}
int mlx5_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
struct ib_wc *in_wc, struct ib_grh *in_grh,
struct ib_mad *in_mad, struct ib_mad *out_mad)
{
u16 slid;
int err;
slid = in_wc ? in_wc->slid : be16_to_cpu(IB_LID_PERMISSIVE);
if (in_mad->mad_hdr.method == IB_MGMT_METHOD_TRAP && slid == 0)
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED;
if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED ||
in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
if (in_mad->mad_hdr.method != IB_MGMT_METHOD_GET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_SET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_TRAP_REPRESS)
return IB_MAD_RESULT_SUCCESS;
/* Don't process SMInfo queries -- the SMA can't handle them.
*/
if (in_mad->mad_hdr.attr_id == IB_SMP_ATTR_SM_INFO)
return IB_MAD_RESULT_SUCCESS;
} else if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_PERF_MGMT ||
in_mad->mad_hdr.mgmt_class == MLX5_IB_VENDOR_CLASS1 ||
in_mad->mad_hdr.mgmt_class == MLX5_IB_VENDOR_CLASS2 ||
in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_CONG_MGMT) {
if (in_mad->mad_hdr.method != IB_MGMT_METHOD_GET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_SET)
return IB_MAD_RESULT_SUCCESS;
} else {
return IB_MAD_RESULT_SUCCESS;
}
err = mlx5_MAD_IFC(to_mdev(ibdev),
mad_flags & IB_MAD_IGNORE_MKEY,
mad_flags & IB_MAD_IGNORE_BKEY,
port_num, in_wc, in_grh, in_mad, out_mad);
if (err)
return IB_MAD_RESULT_FAILURE;
/* set return bit in status of directed route responses */
if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)
out_mad->mad_hdr.status |= cpu_to_be16(1 << 15);
if (in_mad->mad_hdr.method == IB_MGMT_METHOD_TRAP_REPRESS)
/* no response for trap repress */
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED;
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_REPLY;
}
int mlx5_query_ext_port_caps(struct mlx5_ib_dev *dev, u8 port)
{
struct ib_smp *in_mad = NULL;
struct ib_smp *out_mad = NULL;
int err = -ENOMEM;
u16 packet_error;
in_mad = kzalloc(sizeof(*in_mad), GFP_KERNEL);
out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
if (!in_mad || !out_mad)
goto out;
init_query_mad(in_mad);
in_mad->attr_id = MLX5_ATTR_EXTENDED_PORT_INFO;
in_mad->attr_mod = cpu_to_be32(port);
err = mlx5_MAD_IFC(dev, 1, 1, 1, NULL, NULL, in_mad, out_mad);
packet_error = be16_to_cpu(out_mad->status);
dev->mdev.caps.ext_port_cap[port - 1] = (!err && !packet_error) ?
MLX_EXT_PORT_CAP_FLAG_EXTENDED_PORT_INFO : 0;
out:
kfree(in_mad);
kfree(out_mad);
return err;
}

1504
drivers/infiniband/hw/mlx5/main.c Normal file
View File

File diff suppressed because it is too large Load Diff

162
drivers/infiniband/hw/mlx5/mem.c Normal file
View File

@@ -0,0 +1,162 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/module.h>
#include <rdma/ib_umem.h>
#include "mlx5_ib.h"
/* @umem: umem object to scan
* @addr: ib virtual address requested by the user
* @count: number of PAGE_SIZE pages covered by umem
* @shift: page shift for the compound pages found in the region
* @ncont: number of compund pages
* @order: log2 of the number of compound pages
*/
void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
int *ncont, int *order)
{
struct ib_umem_chunk *chunk;
unsigned long tmp;
unsigned long m;
int i, j, k;
u64 base = 0;
int p = 0;
int skip;
int mask;
u64 len;
u64 pfn;
addr = addr >> PAGE_SHIFT;
tmp = (unsigned long)addr;
m = find_first_bit(&tmp, sizeof(tmp));
skip = 1 << m;
mask = skip - 1;
i = 0;
list_for_each_entry(chunk, &umem->chunk_list, list)
for (j = 0; j < chunk->nmap; j++) {
len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
pfn = sg_dma_address(&chunk->page_list[j]) >> PAGE_SHIFT;
for (k = 0; k < len; k++) {
if (!(i & mask)) {
tmp = (unsigned long)pfn;
m = min(m, find_first_bit(&tmp, sizeof(tmp)));
skip = 1 << m;
mask = skip - 1;
base = pfn;
p = 0;
} else {
if (base + p != pfn) {
tmp = (unsigned long)p;
m = find_first_bit(&tmp, sizeof(tmp));
skip = 1 << m;
mask = skip - 1;
base = pfn;
p = 0;
}
}
p++;
i++;
}
}
if (i) {
m = min_t(unsigned long, ilog2(roundup_pow_of_two(i)), m);
if (order)
*order = ilog2(roundup_pow_of_two(i) >> m);
*ncont = DIV_ROUND_UP(i, (1 << m));
} else {
m = 0;
if (order)
*order = 0;
*ncont = 0;
}
*shift = PAGE_SHIFT + m;
*count = i;
}
void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
int page_shift, __be64 *pas, int umr)
{
int shift = page_shift - PAGE_SHIFT;
int mask = (1 << shift) - 1;
struct ib_umem_chunk *chunk;
int i, j, k;
u64 cur = 0;
u64 base;
int len;
i = 0;
list_for_each_entry(chunk, &umem->chunk_list, list)
for (j = 0; j < chunk->nmap; j++) {
len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
base = sg_dma_address(&chunk->page_list[j]);
for (k = 0; k < len; k++) {
if (!(i & mask)) {
cur = base + (k << PAGE_SHIFT);
if (umr)
cur |= 3;
pas[i >> shift] = cpu_to_be64(cur);
mlx5_ib_dbg(dev, "pas[%d] 0x%llx\n",
i >> shift, be64_to_cpu(pas[i >> shift]));
} else
mlx5_ib_dbg(dev, "=====> 0x%llx\n",
base + (k << PAGE_SHIFT));
i++;
}
}
}
int mlx5_ib_get_buf_offset(u64 addr, int page_shift, u32 *offset)
{
u64 page_size;
u64 page_mask;
u64 off_size;
u64 off_mask;
u64 buf_off;
page_size = 1 << page_shift;
page_mask = page_size - 1;
buf_off = addr & page_mask;
off_size = page_size >> 6;
off_mask = off_size - 1;
if (buf_off & off_mask)
return -EINVAL;
*offset = buf_off >> ilog2(off_size);
return 0;
}

545
drivers/infiniband/hw/mlx5/mlx5_ib.h Normal file
View File

@@ -0,0 +1,545 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef MLX5_IB_H
#define MLX5_IB_H
#include <linux/kernel.h>
#include <linux/sched.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_smi.h>
#include <linux/mlx5/driver.h>
#include <linux/mlx5/cq.h>
#include <linux/mlx5/qp.h>
#include <linux/mlx5/srq.h>
#include <linux/types.h>
#define mlx5_ib_dbg(dev, format, arg...) \
pr_debug("%s:%s:%d:(pid %d): " format, (dev)->ib_dev.name, __func__, \
__LINE__, current->pid, ##arg)
#define mlx5_ib_err(dev, format, arg...) \
pr_err("%s:%s:%d:(pid %d): " format, (dev)->ib_dev.name, __func__, \
__LINE__, current->pid, ##arg)
#define mlx5_ib_warn(dev, format, arg...) \
pr_warn("%s:%s:%d:(pid %d): " format, (dev)->ib_dev.name, __func__, \
__LINE__, current->pid, ##arg)
enum {
MLX5_IB_MMAP_CMD_SHIFT = 8,
MLX5_IB_MMAP_CMD_MASK = 0xff,
};
enum mlx5_ib_mmap_cmd {
MLX5_IB_MMAP_REGULAR_PAGE = 0,
MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES = 1, /* always last */
};
enum {
MLX5_RES_SCAT_DATA32_CQE = 0x1,
MLX5_RES_SCAT_DATA64_CQE = 0x2,
MLX5_REQ_SCAT_DATA32_CQE = 0x11,
MLX5_REQ_SCAT_DATA64_CQE = 0x22,
};
enum mlx5_ib_latency_class {
MLX5_IB_LATENCY_CLASS_LOW,
MLX5_IB_LATENCY_CLASS_MEDIUM,
MLX5_IB_LATENCY_CLASS_HIGH,
MLX5_IB_LATENCY_CLASS_FAST_PATH
};
enum mlx5_ib_mad_ifc_flags {
MLX5_MAD_IFC_IGNORE_MKEY = 1,
MLX5_MAD_IFC_IGNORE_BKEY = 2,
MLX5_MAD_IFC_NET_VIEW = 4,
};
struct mlx5_ib_ucontext {
struct ib_ucontext ibucontext;
struct list_head db_page_list;
/* protect doorbell record alloc/free
*/
struct mutex db_page_mutex;
struct mlx5_uuar_info uuari;
};
static inline struct mlx5_ib_ucontext *to_mucontext(struct ib_ucontext *ibucontext)
{
return container_of(ibucontext, struct mlx5_ib_ucontext, ibucontext);
}
struct mlx5_ib_pd {
struct ib_pd ibpd;
u32 pdn;
u32 pa_lkey;
};
/* Use macros here so that don't have to duplicate
* enum ib_send_flags and enum ib_qp_type for low-level driver
*/
#define MLX5_IB_SEND_UMR_UNREG IB_SEND_RESERVED_START
#define MLX5_IB_QPT_REG_UMR IB_QPT_RESERVED1
#define MLX5_IB_WR_UMR IB_WR_RESERVED1
struct wr_list {
u16 opcode;
u16 next;
};
struct mlx5_ib_wq {
u64 *wrid;
u32 *wr_data;
struct wr_list *w_list;
unsigned *wqe_head;
u16 unsig_count;
/* serialize post to the work queue
*/
spinlock_t lock;
int wqe_cnt;
int max_post;
int max_gs;
int offset;
int wqe_shift;
unsigned head;
unsigned tail;
u16 cur_post;
u16 last_poll;
void *qend;
};
enum {
MLX5_QP_USER,
MLX5_QP_KERNEL,
MLX5_QP_EMPTY
};
struct mlx5_ib_qp {
struct ib_qp ibqp;
struct mlx5_core_qp mqp;
struct mlx5_buf buf;
struct mlx5_db db;
struct mlx5_ib_wq rq;
u32 doorbell_qpn;
u8 sq_signal_bits;
u8 fm_cache;
int sq_max_wqes_per_wr;
int sq_spare_wqes;
struct mlx5_ib_wq sq;
struct ib_umem *umem;
int buf_size;
/* serialize qp state modifications
*/
struct mutex mutex;
u16 xrcdn;
u32 flags;
u8 port;
u8 alt_port;
u8 atomic_rd_en;
u8 resp_depth;
u8 state;
int mlx_type;
int wq_sig;
int scat_cqe;
int max_inline_data;
struct mlx5_bf *bf;
int has_rq;
/* only for user space QPs. For kernel
* we have it from the bf object
*/
int uuarn;
int create_type;
u32 pa_lkey;
};
struct mlx5_ib_cq_buf {
struct mlx5_buf buf;
struct ib_umem *umem;
int cqe_size;
};
enum mlx5_ib_qp_flags {
MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK = 1 << 0,
MLX5_IB_QP_SIGNATURE_HANDLING = 1 << 1,
};
struct mlx5_shared_mr_info {
int mr_id;
struct ib_umem *umem;
};
struct mlx5_ib_cq {
struct ib_cq ibcq;
struct mlx5_core_cq mcq;
struct mlx5_ib_cq_buf buf;
struct mlx5_db db;
/* serialize access to the CQ
*/
spinlock_t lock;
/* protect resize cq
*/
struct mutex resize_mutex;
struct mlx5_ib_cq_resize *resize_buf;
struct ib_umem *resize_umem;
int cqe_size;
};
struct mlx5_ib_srq {
struct ib_srq ibsrq;
struct mlx5_core_srq msrq;
struct mlx5_buf buf;
struct mlx5_db db;
u64 *wrid;
/* protect SRQ hanlding
*/
spinlock_t lock;
int head;
int tail;
u16 wqe_ctr;
struct ib_umem *umem;
/* serialize arming a SRQ
*/
struct mutex mutex;
int wq_sig;
};
struct mlx5_ib_xrcd {
struct ib_xrcd ibxrcd;
u32 xrcdn;
};
struct mlx5_ib_mr {
struct ib_mr ibmr;
struct mlx5_core_mr mmr;
struct ib_umem *umem;
struct mlx5_shared_mr_info *smr_info;
struct list_head list;
int order;
int umred;
__be64 *pas;
dma_addr_t dma;
int npages;
struct completion done;
enum ib_wc_status status;
};
struct mlx5_ib_fast_reg_page_list {
struct ib_fast_reg_page_list ibfrpl;
__be64 *mapped_page_list;
dma_addr_t map;
};
struct umr_common {
struct ib_pd *pd;
struct ib_cq *cq;
struct ib_qp *qp;
struct ib_mr *mr;
/* control access to UMR QP
*/
struct semaphore sem;
};
enum {
MLX5_FMR_INVALID,
MLX5_FMR_VALID,
MLX5_FMR_BUSY,
};
struct mlx5_ib_fmr {
struct ib_fmr ibfmr;
struct mlx5_core_mr mr;
int access_flags;
int state;
/* protect fmr state
*/
spinlock_t lock;
u64 wrid;
struct ib_send_wr wr[2];
u8 page_shift;
struct ib_fast_reg_page_list page_list;
};
struct mlx5_cache_ent {
struct list_head head;
/* sync access to the cahce entry
*/
spinlock_t lock;
struct dentry *dir;
char name[4];
u32 order;
u32 size;
u32 cur;
u32 miss;
u32 limit;
struct dentry *fsize;
struct dentry *fcur;
struct dentry *fmiss;
struct dentry *flimit;
struct mlx5_ib_dev *dev;
struct work_struct work;
struct delayed_work dwork;
};
struct mlx5_mr_cache {
struct workqueue_struct *wq;
struct mlx5_cache_ent ent[MAX_MR_CACHE_ENTRIES];
int stopped;
struct dentry *root;
unsigned long last_add;
};
struct mlx5_ib_resources {
struct ib_cq *c0;
struct ib_xrcd *x0;
struct ib_xrcd *x1;
struct ib_pd *p0;
struct ib_srq *s0;
};
struct mlx5_ib_dev {
struct ib_device ib_dev;
struct mlx5_core_dev mdev;
MLX5_DECLARE_DOORBELL_LOCK(uar_lock);
struct list_head eqs_list;
int num_ports;
int num_comp_vectors;
/* serialize update of capability mask
*/
struct mutex cap_mask_mutex;
bool ib_active;
struct umr_common umrc;
/* sync used page count stats
*/
spinlock_t mr_lock;
struct mlx5_ib_resources devr;
struct mlx5_mr_cache cache;
};
static inline struct mlx5_ib_cq *to_mibcq(struct mlx5_core_cq *mcq)
{
return container_of(mcq, struct mlx5_ib_cq, mcq);
}
static inline struct mlx5_ib_xrcd *to_mxrcd(struct ib_xrcd *ibxrcd)
{
return container_of(ibxrcd, struct mlx5_ib_xrcd, ibxrcd);
}
static inline struct mlx5_ib_dev *to_mdev(struct ib_device *ibdev)
{
return container_of(ibdev, struct mlx5_ib_dev, ib_dev);
}
static inline struct mlx5_ib_fmr *to_mfmr(struct ib_fmr *ibfmr)
{
return container_of(ibfmr, struct mlx5_ib_fmr, ibfmr);
}
static inline struct mlx5_ib_cq *to_mcq(struct ib_cq *ibcq)
{
return container_of(ibcq, struct mlx5_ib_cq, ibcq);
}
static inline struct mlx5_ib_qp *to_mibqp(struct mlx5_core_qp *mqp)
{
return container_of(mqp, struct mlx5_ib_qp, mqp);
}
static inline struct mlx5_ib_pd *to_mpd(struct ib_pd *ibpd)
{
return container_of(ibpd, struct mlx5_ib_pd, ibpd);
}
static inline struct mlx5_ib_srq *to_msrq(struct ib_srq *ibsrq)
{
return container_of(ibsrq, struct mlx5_ib_srq, ibsrq);
}
static inline struct mlx5_ib_qp *to_mqp(struct ib_qp *ibqp)
{
return container_of(ibqp, struct mlx5_ib_qp, ibqp);
}
static inline struct mlx5_ib_srq *to_mibsrq(struct mlx5_core_srq *msrq)
{
return container_of(msrq, struct mlx5_ib_srq, msrq);
}
static inline struct mlx5_ib_mr *to_mmr(struct ib_mr *ibmr)
{
return container_of(ibmr, struct mlx5_ib_mr, ibmr);
}
static inline struct mlx5_ib_fast_reg_page_list *to_mfrpl(struct ib_fast_reg_page_list *ibfrpl)
{
return container_of(ibfrpl, struct mlx5_ib_fast_reg_page_list, ibfrpl);
}
struct mlx5_ib_ah {
struct ib_ah ibah;
struct mlx5_av av;
};
static inline struct mlx5_ib_ah *to_mah(struct ib_ah *ibah)
{
return container_of(ibah, struct mlx5_ib_ah, ibah);
}
static inline struct mlx5_ib_dev *mlx5_core2ibdev(struct mlx5_core_dev *dev)
{
return container_of(dev, struct mlx5_ib_dev, mdev);
}
static inline struct mlx5_ib_dev *mlx5_pci2ibdev(struct pci_dev *pdev)
{
return mlx5_core2ibdev(pci2mlx5_core_dev(pdev));
}
int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
struct mlx5_db *db);
void mlx5_ib_db_unmap_user(struct mlx5_ib_ucontext *context, struct mlx5_db *db);
void __mlx5_ib_cq_clean(struct mlx5_ib_cq *cq, u32 qpn, struct mlx5_ib_srq *srq);
void mlx5_ib_cq_clean(struct mlx5_ib_cq *cq, u32 qpn, struct mlx5_ib_srq *srq);
void mlx5_ib_free_srq_wqe(struct mlx5_ib_srq *srq, int wqe_index);
int mlx5_MAD_IFC(struct mlx5_ib_dev *dev, int ignore_mkey, int ignore_bkey,
int port, struct ib_wc *in_wc, struct ib_grh *in_grh,
void *in_mad, void *response_mad);
struct ib_ah *create_ib_ah(struct ib_ah_attr *ah_attr,
struct mlx5_ib_ah *ah);
struct ib_ah *mlx5_ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr);
int mlx5_ib_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr);
int mlx5_ib_destroy_ah(struct ib_ah *ah);
struct ib_srq *mlx5_ib_create_srq(struct ib_pd *pd,
struct ib_srq_init_attr *init_attr,
struct ib_udata *udata);
int mlx5_ib_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
enum ib_srq_attr_mask attr_mask, struct ib_udata *udata);
int mlx5_ib_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *srq_attr);
int mlx5_ib_destroy_srq(struct ib_srq *srq);
int mlx5_ib_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr);
struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
struct ib_qp_init_attr *init_attr,
struct ib_udata *udata);
int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
int attr_mask, struct ib_udata *udata);
int mlx5_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, int qp_attr_mask,
struct ib_qp_init_attr *qp_init_attr);
int mlx5_ib_destroy_qp(struct ib_qp *qp);
int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
struct ib_send_wr **bad_wr);
int mlx5_ib_post_recv(struct ib_qp *ibqp, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr);
void *mlx5_get_send_wqe(struct mlx5_ib_qp *qp, int n);
struct ib_cq *mlx5_ib_create_cq(struct ib_device *ibdev, int entries,
int vector, struct ib_ucontext *context,
struct ib_udata *udata);
int mlx5_ib_destroy_cq(struct ib_cq *cq);
int mlx5_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc);
int mlx5_ib_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags flags);
int mlx5_ib_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period);
int mlx5_ib_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata);
struct ib_mr *mlx5_ib_get_dma_mr(struct ib_pd *pd, int acc);
struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 virt_addr, int access_flags,
struct ib_udata *udata);
int mlx5_ib_dereg_mr(struct ib_mr *ibmr);
struct ib_mr *mlx5_ib_alloc_fast_reg_mr(struct ib_pd *pd,
int max_page_list_len);
struct ib_fast_reg_page_list *mlx5_ib_alloc_fast_reg_page_list(struct ib_device *ibdev,
int page_list_len);
void mlx5_ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list);
struct ib_fmr *mlx5_ib_fmr_alloc(struct ib_pd *pd, int acc,
struct ib_fmr_attr *fmr_attr);
int mlx5_ib_map_phys_fmr(struct ib_fmr *ibfmr, u64 *page_list,
int npages, u64 iova);
int mlx5_ib_unmap_fmr(struct list_head *fmr_list);
int mlx5_ib_fmr_dealloc(struct ib_fmr *ibfmr);
int mlx5_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
struct ib_wc *in_wc, struct ib_grh *in_grh,
struct ib_mad *in_mad, struct ib_mad *out_mad);
struct ib_xrcd *mlx5_ib_alloc_xrcd(struct ib_device *ibdev,
struct ib_ucontext *context,
struct ib_udata *udata);
int mlx5_ib_dealloc_xrcd(struct ib_xrcd *xrcd);
int mlx5_vector2eqn(struct mlx5_ib_dev *dev, int vector, int *eqn, int *irqn);
int mlx5_ib_get_buf_offset(u64 addr, int page_shift, u32 *offset);
int mlx5_query_ext_port_caps(struct mlx5_ib_dev *dev, u8 port);
int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
struct ib_port_attr *props);
int mlx5_ib_init_fmr(struct mlx5_ib_dev *dev);
void mlx5_ib_cleanup_fmr(struct mlx5_ib_dev *dev);
void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
int *ncont, int *order);
void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
int page_shift, __be64 *pas, int umr);
void mlx5_ib_copy_pas(u64 *old, u64 *new, int step, int num);
int mlx5_ib_get_cqe_size(struct mlx5_ib_dev *dev, struct ib_cq *ibcq);
int mlx5_mr_cache_init(struct mlx5_ib_dev *dev);
int mlx5_mr_cache_cleanup(struct mlx5_ib_dev *dev);
int mlx5_mr_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift);
void mlx5_umr_cq_handler(struct ib_cq *cq, void *cq_context);
static inline void init_query_mad(struct ib_smp *mad)
{
mad->base_version = 1;
mad->mgmt_class = IB_MGMT_CLASS_SUBN_LID_ROUTED;
mad->class_version = 1;
mad->method = IB_MGMT_METHOD_GET;
}
static inline u8 convert_access(int acc)
{
return (acc & IB_ACCESS_REMOTE_ATOMIC ? MLX5_PERM_ATOMIC : 0) |
(acc & IB_ACCESS_REMOTE_WRITE ? MLX5_PERM_REMOTE_WRITE : 0) |
(acc & IB_ACCESS_REMOTE_READ ? MLX5_PERM_REMOTE_READ : 0) |
(acc & IB_ACCESS_LOCAL_WRITE ? MLX5_PERM_LOCAL_WRITE : 0) |
MLX5_PERM_LOCAL_READ;
}
#endif /* MLX5_IB_H */

1007
drivers/infiniband/hw/mlx5/mr.c Normal file
View File

File diff suppressed because it is too large Load Diff

2524
drivers/infiniband/hw/mlx5/qp.c Normal file
View File

File diff suppressed because it is too large Load Diff

473
drivers/infiniband/hw/mlx5/srq.c Normal file
View File

@@ -0,0 +1,473 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/module.h>
#include <linux/mlx5/qp.h>
#include <linux/mlx5/srq.h>
#include <linux/slab.h>
#include <rdma/ib_umem.h>
#include "mlx5_ib.h"
#include "user.h"
/* not supported currently */
static int srq_signature;
static void *get_wqe(struct mlx5_ib_srq *srq, int n)
{
return mlx5_buf_offset(&srq->buf, n << srq->msrq.wqe_shift);
}
static void mlx5_ib_srq_event(struct mlx5_core_srq *srq, enum mlx5_event type)
{
struct ib_event event;
struct ib_srq *ibsrq = &to_mibsrq(srq)->ibsrq;
if (ibsrq->event_handler) {
event.device = ibsrq->device;
event.element.srq = ibsrq;
switch (type) {
case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT:
event.event = IB_EVENT_SRQ_LIMIT_REACHED;
break;
case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR:
event.event = IB_EVENT_SRQ_ERR;
break;
default:
pr_warn("mlx5_ib: Unexpected event type %d on SRQ %06x\n",
type, srq->srqn);
return;
}
ibsrq->event_handler(&event, ibsrq->srq_context);
}
}
static int create_srq_user(struct ib_pd *pd, struct mlx5_ib_srq *srq,
struct mlx5_create_srq_mbox_in **in,
struct ib_udata *udata, int buf_size, int *inlen)
{
struct mlx5_ib_dev *dev = to_mdev(pd->device);
struct mlx5_ib_create_srq ucmd;
int err;
int npages;
int page_shift;
int ncont;
u32 offset;
if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
mlx5_ib_dbg(dev, "failed copy udata\n");
return -EFAULT;
}
srq->wq_sig = !!(ucmd.flags & MLX5_SRQ_FLAG_SIGNATURE);
srq->umem = ib_umem_get(pd->uobject->context, ucmd.buf_addr, buf_size,
0, 0);
if (IS_ERR(srq->umem)) {
mlx5_ib_dbg(dev, "failed umem get, size %d\n", buf_size);
err = PTR_ERR(srq->umem);
return err;
}
mlx5_ib_cont_pages(srq->umem, ucmd.buf_addr, &npages,
&page_shift, &ncont, NULL);
err = mlx5_ib_get_buf_offset(ucmd.buf_addr, page_shift,
&offset);
if (err) {
mlx5_ib_warn(dev, "bad offset\n");
goto err_umem;
}
*inlen = sizeof(**in) + sizeof(*(*in)->pas) * ncont;
*in = mlx5_vzalloc(*inlen);
if (!(*in)) {
err = -ENOMEM;
goto err_umem;
}
mlx5_ib_populate_pas(dev, srq->umem, page_shift, (*in)->pas, 0);
err = mlx5_ib_db_map_user(to_mucontext(pd->uobject->context),
ucmd.db_addr, &srq->db);
if (err) {
mlx5_ib_dbg(dev, "map doorbell failed\n");
goto err_in;
}
(*in)->ctx.log_pg_sz = page_shift - PAGE_SHIFT;
(*in)->ctx.pgoff_cqn = cpu_to_be32(offset << 26);
return 0;
err_in:
mlx5_vfree(*in);
err_umem:
ib_umem_release(srq->umem);
return err;
}
static int create_srq_kernel(struct mlx5_ib_dev *dev, struct mlx5_ib_srq *srq,
struct mlx5_create_srq_mbox_in **in, int buf_size,
int *inlen)
{
int err;
int i;
struct mlx5_wqe_srq_next_seg *next;
int page_shift;
int npages;
err = mlx5_db_alloc(&dev->mdev, &srq->db);
if (err) {
mlx5_ib_warn(dev, "alloc dbell rec failed\n");
return err;
}
*srq->db.db = 0;
if (mlx5_buf_alloc(&dev->mdev, buf_size, PAGE_SIZE * 2, &srq->buf)) {
mlx5_ib_dbg(dev, "buf alloc failed\n");
err = -ENOMEM;
goto err_db;
}
page_shift = srq->buf.page_shift;
srq->head = 0;
srq->tail = srq->msrq.max - 1;
srq->wqe_ctr = 0;
for (i = 0; i < srq->msrq.max; i++) {
next = get_wqe(srq, i);
next->next_wqe_index =
cpu_to_be16((i + 1) & (srq->msrq.max - 1));
}
npages = DIV_ROUND_UP(srq->buf.npages, 1 << (page_shift - PAGE_SHIFT));
mlx5_ib_dbg(dev, "buf_size %d, page_shift %d, npages %d, calc npages %d\n",
buf_size, page_shift, srq->buf.npages, npages);
*inlen = sizeof(**in) + sizeof(*(*in)->pas) * npages;
*in = mlx5_vzalloc(*inlen);
if (!*in) {
err = -ENOMEM;
goto err_buf;
}
mlx5_fill_page_array(&srq->buf, (*in)->pas);
srq->wrid = kmalloc(srq->msrq.max * sizeof(u64), GFP_KERNEL);
if (!srq->wrid) {
mlx5_ib_dbg(dev, "kmalloc failed %lu\n",
(unsigned long)(srq->msrq.max * sizeof(u64)));
err = -ENOMEM;
goto err_in;
}
srq->wq_sig = !!srq_signature;
(*in)->ctx.log_pg_sz = page_shift - PAGE_SHIFT;
return 0;
err_in:
mlx5_vfree(*in);
err_buf:
mlx5_buf_free(&dev->mdev, &srq->buf);
err_db:
mlx5_db_free(&dev->mdev, &srq->db);
return err;
}
static void destroy_srq_user(struct ib_pd *pd, struct mlx5_ib_srq *srq)
{
mlx5_ib_db_unmap_user(to_mucontext(pd->uobject->context), &srq->db);
ib_umem_release(srq->umem);
}
static void destroy_srq_kernel(struct mlx5_ib_dev *dev, struct mlx5_ib_srq *srq)
{
kfree(srq->wrid);
mlx5_buf_free(&dev->mdev, &srq->buf);
mlx5_db_free(&dev->mdev, &srq->db);
}
struct ib_srq *mlx5_ib_create_srq(struct ib_pd *pd,
struct ib_srq_init_attr *init_attr,
struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(pd->device);
struct mlx5_ib_srq *srq;
int desc_size;
int buf_size;
int err;
struct mlx5_create_srq_mbox_in *uninitialized_var(in);
int uninitialized_var(inlen);
int is_xrc;
u32 flgs, xrcdn;
/* Sanity check SRQ size before proceeding */
if (init_attr->attr.max_wr >= dev->mdev.caps.max_srq_wqes) {
mlx5_ib_dbg(dev, "max_wr %d, cap %d\n",
init_attr->attr.max_wr,
dev->mdev.caps.max_srq_wqes);
return ERR_PTR(-EINVAL);
}
srq = kmalloc(sizeof(*srq), GFP_KERNEL);
if (!srq)
return ERR_PTR(-ENOMEM);
mutex_init(&srq->mutex);
spin_lock_init(&srq->lock);
srq->msrq.max = roundup_pow_of_two(init_attr->attr.max_wr + 1);
srq->msrq.max_gs = init_attr->attr.max_sge;
desc_size = sizeof(struct mlx5_wqe_srq_next_seg) +
srq->msrq.max_gs * sizeof(struct mlx5_wqe_data_seg);
desc_size = roundup_pow_of_two(desc_size);
desc_size = max_t(int, 32, desc_size);
srq->msrq.max_avail_gather = (desc_size - sizeof(struct mlx5_wqe_srq_next_seg)) /
sizeof(struct mlx5_wqe_data_seg);
srq->msrq.wqe_shift = ilog2(desc_size);
buf_size = srq->msrq.max * desc_size;
mlx5_ib_dbg(dev, "desc_size 0x%x, req wr 0x%x, srq size 0x%x, max_gs 0x%x, max_avail_gather 0x%x\n",
desc_size, init_attr->attr.max_wr, srq->msrq.max, srq->msrq.max_gs,
srq->msrq.max_avail_gather);
if (pd->uobject)
err = create_srq_user(pd, srq, &in, udata, buf_size, &inlen);
else
err = create_srq_kernel(dev, srq, &in, buf_size, &inlen);
if (err) {
mlx5_ib_warn(dev, "create srq %s failed, err %d\n",
pd->uobject ? "user" : "kernel", err);
goto err_srq;
}
is_xrc = (init_attr->srq_type == IB_SRQT_XRC);
in->ctx.state_log_sz = ilog2(srq->msrq.max);
flgs = ((srq->msrq.wqe_shift - 4) | (is_xrc << 5) | (srq->wq_sig << 7)) << 24;
xrcdn = 0;
if (is_xrc) {
xrcdn = to_mxrcd(init_attr->ext.xrc.xrcd)->xrcdn;
in->ctx.pgoff_cqn |= cpu_to_be32(to_mcq(init_attr->ext.xrc.cq)->mcq.cqn);
} else if (init_attr->srq_type == IB_SRQT_BASIC) {
xrcdn = to_mxrcd(dev->devr.x0)->xrcdn;
in->ctx.pgoff_cqn |= cpu_to_be32(to_mcq(dev->devr.c0)->mcq.cqn);
}
in->ctx.flags_xrcd = cpu_to_be32((flgs & 0xFF000000) | (xrcdn & 0xFFFFFF));
in->ctx.pd = cpu_to_be32(to_mpd(pd)->pdn);
in->ctx.db_record = cpu_to_be64(srq->db.dma);
err = mlx5_core_create_srq(&dev->mdev, &srq->msrq, in, inlen);
mlx5_vfree(in);
if (err) {
mlx5_ib_dbg(dev, "create SRQ failed, err %d\n", err);
goto err_srq;
}
mlx5_ib_dbg(dev, "create SRQ with srqn 0x%x\n", srq->msrq.srqn);
srq->msrq.event = mlx5_ib_srq_event;
srq->ibsrq.ext.xrc.srq_num = srq->msrq.srqn;
if (pd->uobject)
if (ib_copy_to_udata(udata, &srq->msrq.srqn, sizeof(__u32))) {
mlx5_ib_dbg(dev, "copy to user failed\n");
err = -EFAULT;
goto err_core;
}
init_attr->attr.max_wr = srq->msrq.max - 1;
return &srq->ibsrq;
err_core:
mlx5_core_destroy_srq(&dev->mdev, &srq->msrq);
if (pd->uobject)
destroy_srq_user(pd, srq);
else
destroy_srq_kernel(dev, srq);
err_srq:
kfree(srq);
return ERR_PTR(err);
}
int mlx5_ib_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
enum ib_srq_attr_mask attr_mask, struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(ibsrq->device);
struct mlx5_ib_srq *srq = to_msrq(ibsrq);
int ret;
/* We don't support resizing SRQs yet */
if (attr_mask & IB_SRQ_MAX_WR)
return -EINVAL;
if (attr_mask & IB_SRQ_LIMIT) {
if (attr->srq_limit >= srq->msrq.max)
return -EINVAL;
mutex_lock(&srq->mutex);
ret = mlx5_core_arm_srq(&dev->mdev, &srq->msrq, attr->srq_limit, 1);
mutex_unlock(&srq->mutex);
if (ret)
return ret;
}
return 0;
}
int mlx5_ib_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *srq_attr)
{
struct mlx5_ib_dev *dev = to_mdev(ibsrq->device);
struct mlx5_ib_srq *srq = to_msrq(ibsrq);
int ret;
struct mlx5_query_srq_mbox_out *out;
out = kzalloc(sizeof(*out), GFP_KERNEL);
if (!out)
return -ENOMEM;
ret = mlx5_core_query_srq(&dev->mdev, &srq->msrq, out);
if (ret)
goto out_box;
srq_attr->srq_limit = be16_to_cpu(out->ctx.lwm);
srq_attr->max_wr = srq->msrq.max - 1;
srq_attr->max_sge = srq->msrq.max_gs;
out_box:
kfree(out);
return ret;
}
int mlx5_ib_destroy_srq(struct ib_srq *srq)
{
struct mlx5_ib_dev *dev = to_mdev(srq->device);
struct mlx5_ib_srq *msrq = to_msrq(srq);
mlx5_core_destroy_srq(&dev->mdev, &msrq->msrq);
if (srq->uobject) {
mlx5_ib_db_unmap_user(to_mucontext(srq->uobject->context), &msrq->db);
ib_umem_release(msrq->umem);
} else {
kfree(msrq->wrid);
mlx5_buf_free(&dev->mdev, &msrq->buf);
mlx5_db_free(&dev->mdev, &msrq->db);
}
kfree(srq);
return 0;
}
void mlx5_ib_free_srq_wqe(struct mlx5_ib_srq *srq, int wqe_index)
{
struct mlx5_wqe_srq_next_seg *next;
/* always called with interrupts disabled. */
spin_lock(&srq->lock);
next = get_wqe(srq, srq->tail);
next->next_wqe_index = cpu_to_be16(wqe_index);
srq->tail = wqe_index;
spin_unlock(&srq->lock);
}
int mlx5_ib_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr)
{
struct mlx5_ib_srq *srq = to_msrq(ibsrq);
struct mlx5_wqe_srq_next_seg *next;
struct mlx5_wqe_data_seg *scat;
unsigned long flags;
int err = 0;
int nreq;
int i;
spin_lock_irqsave(&srq->lock, flags);
for (nreq = 0; wr; nreq++, wr = wr->next) {
if (unlikely(wr->num_sge > srq->msrq.max_gs)) {
err = -EINVAL;
*bad_wr = wr;
break;
}
if (unlikely(srq->head == srq->tail)) {
err = -ENOMEM;
*bad_wr = wr;
break;
}
srq->wrid[srq->head] = wr->wr_id;
next = get_wqe(srq, srq->head);
srq->head = be16_to_cpu(next->next_wqe_index);
scat = (struct mlx5_wqe_data_seg *)(next + 1);
for (i = 0; i < wr->num_sge; i++) {
scat[i].byte_count = cpu_to_be32(wr->sg_list[i].length);
scat[i].lkey = cpu_to_be32(wr->sg_list[i].lkey);
scat[i].addr = cpu_to_be64(wr->sg_list[i].addr);
}
if (i < srq->msrq.max_avail_gather) {
scat[i].byte_count = 0;
scat[i].lkey = cpu_to_be32(MLX5_INVALID_LKEY);
scat[i].addr = 0;
}
}
if (likely(nreq)) {
srq->wqe_ctr += nreq;
/* Make sure that descriptors are written before
* doorbell record.
*/
wmb();
*srq->db.db = cpu_to_be32(srq->wqe_ctr);
}
spin_unlock_irqrestore(&srq->lock, flags);
return err;
}

121
drivers/infiniband/hw/mlx5/user.h Normal file
View File

@@ -0,0 +1,121 @@
/*
* Copyright (c) 2013, Mellanox Technologies inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef MLX5_IB_USER_H
#define MLX5_IB_USER_H
#include <linux/types.h>
enum {
MLX5_QP_FLAG_SIGNATURE = 1 << 0,
MLX5_QP_FLAG_SCATTER_CQE = 1 << 1,
};
enum {
MLX5_SRQ_FLAG_SIGNATURE = 1 << 0,
};
/* Increment this value if any changes that break userspace ABI
* compatibility are made.
*/
#define MLX5_IB_UVERBS_ABI_VERSION 1
/* Make sure that all structs defined in this file remain laid out so
* that they pack the same way on 32-bit and 64-bit architectures (to
* avoid incompatibility between 32-bit userspace and 64-bit kernels).
* In particular do not use pointer types -- pass pointers in __u64
* instead.
*/
struct mlx5_ib_alloc_ucontext_req {
__u32 total_num_uuars;
__u32 num_low_latency_uuars;
};
struct mlx5_ib_alloc_ucontext_resp {
__u32 qp_tab_size;
__u32 bf_reg_size;
__u32 tot_uuars;
__u32 cache_line_size;
__u16 max_sq_desc_sz;
__u16 max_rq_desc_sz;
__u32 max_send_wqebb;
__u32 max_recv_wr;
__u32 max_srq_recv_wr;
__u16 num_ports;
__u16 reserved;
};
struct mlx5_ib_alloc_pd_resp {
__u32 pdn;
};
struct mlx5_ib_create_cq {
__u64 buf_addr;
__u64 db_addr;
__u32 cqe_size;
};
struct mlx5_ib_create_cq_resp {
__u32 cqn;
__u32 reserved;
};
struct mlx5_ib_resize_cq {
__u64 buf_addr;
};
struct mlx5_ib_create_srq {
__u64 buf_addr;
__u64 db_addr;
__u32 flags;
};
struct mlx5_ib_create_srq_resp {
__u32 srqn;
__u32 reserved;
};
struct mlx5_ib_create_qp {
__u64 buf_addr;
__u64 db_addr;
__u32 sq_wqe_count;
__u32 rq_wqe_count;
__u32 rq_wqe_shift;
__u32 flags;
};
struct mlx5_ib_create_qp_resp {
__u32 uuar_index;
};
#endif /* MLX5_IB_USER_H */