freebsd-skq/contrib/ofed/libmlx5/verbs.c

2213 lines
50 KiB
C
Raw Normal View History

/*
* Copyright (c) 2012 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 <config.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <pthread.h>
#include <errno.h>
#include <limits.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/mman.h>
#include "mlx5.h"
#include "mlx5-abi.h"
#include "wqe.h"
int mlx5_single_threaded = 0;
static inline int is_xrc_tgt(int type)
{
return type == IBV_QPT_XRC_RECV;
}
int mlx5_query_device(struct ibv_context *context, struct ibv_device_attr *attr)
{
struct ibv_query_device cmd;
uint64_t raw_fw_ver;
unsigned major, minor, sub_minor;
int ret;
ret = ibv_cmd_query_device(context, attr, &raw_fw_ver, &cmd, sizeof cmd);
if (ret)
return ret;
major = (raw_fw_ver >> 32) & 0xffff;
minor = (raw_fw_ver >> 16) & 0xffff;
sub_minor = raw_fw_ver & 0xffff;
snprintf(attr->fw_ver, sizeof attr->fw_ver,
"%d.%d.%04d", major, minor, sub_minor);
return 0;
}
#define READL(ptr) (*((uint32_t *)(ptr)))
static int mlx5_read_clock(struct ibv_context *context, uint64_t *cycles)
{
unsigned int clockhi, clocklo, clockhi1;
int i;
struct mlx5_context *ctx = to_mctx(context);
if (!ctx->hca_core_clock)
return -EOPNOTSUPP;
/* Handle wraparound */
for (i = 0; i < 2; i++) {
clockhi = be32toh(READL(ctx->hca_core_clock));
clocklo = be32toh(READL(ctx->hca_core_clock + 4));
clockhi1 = be32toh(READL(ctx->hca_core_clock));
if (clockhi == clockhi1)
break;
}
*cycles = (uint64_t)clockhi << 32 | (uint64_t)clocklo;
return 0;
}
int mlx5_query_rt_values(struct ibv_context *context,
struct ibv_values_ex *values)
{
uint32_t comp_mask = 0;
int err = 0;
if (values->comp_mask & IBV_VALUES_MASK_RAW_CLOCK) {
uint64_t cycles;
err = mlx5_read_clock(context, &cycles);
if (!err) {
values->raw_clock.tv_sec = 0;
values->raw_clock.tv_nsec = cycles;
comp_mask |= IBV_VALUES_MASK_RAW_CLOCK;
}
}
values->comp_mask = comp_mask;
return err;
}
int mlx5_query_port(struct ibv_context *context, uint8_t port,
struct ibv_port_attr *attr)
{
struct ibv_query_port cmd;
return ibv_cmd_query_port(context, port, attr, &cmd, sizeof cmd);
}
struct ibv_pd *mlx5_alloc_pd(struct ibv_context *context)
{
struct ibv_alloc_pd cmd;
struct mlx5_alloc_pd_resp resp;
struct mlx5_pd *pd;
pd = calloc(1, sizeof *pd);
if (!pd)
return NULL;
if (ibv_cmd_alloc_pd(context, &pd->ibv_pd, &cmd, sizeof cmd,
&resp.ibv_resp, sizeof resp)) {
free(pd);
return NULL;
}
pd->pdn = resp.pdn;
return &pd->ibv_pd;
}
int mlx5_free_pd(struct ibv_pd *pd)
{
int ret;
ret = ibv_cmd_dealloc_pd(pd);
if (ret)
return ret;
free(to_mpd(pd));
return 0;
}
struct ibv_mr *mlx5_reg_mr(struct ibv_pd *pd, void *addr, size_t length,
int acc)
{
struct mlx5_mr *mr;
struct ibv_reg_mr cmd;
int ret;
enum ibv_access_flags access = (enum ibv_access_flags)acc;
struct ibv_reg_mr_resp resp;
mr = calloc(1, sizeof(*mr));
if (!mr)
return NULL;
ret = ibv_cmd_reg_mr(pd, addr, length, (uintptr_t)addr, access,
&(mr->ibv_mr), &cmd, sizeof(cmd), &resp,
sizeof resp);
if (ret) {
mlx5_free_buf(&(mr->buf));
free(mr);
return NULL;
}
mr->alloc_flags = acc;
return &mr->ibv_mr;
}
int mlx5_rereg_mr(struct ibv_mr *ibmr, int flags, struct ibv_pd *pd, void *addr,
size_t length, int access)
{
struct ibv_rereg_mr cmd;
struct ibv_rereg_mr_resp resp;
if (flags & IBV_REREG_MR_KEEP_VALID)
return ENOTSUP;
return ibv_cmd_rereg_mr(ibmr, flags, addr, length, (uintptr_t)addr,
access, pd, &cmd, sizeof(cmd), &resp,
sizeof(resp));
}
int mlx5_dereg_mr(struct ibv_mr *ibmr)
{
int ret;
struct mlx5_mr *mr = to_mmr(ibmr);
ret = ibv_cmd_dereg_mr(ibmr);
if (ret)
return ret;
free(mr);
return 0;
}
struct ibv_mw *mlx5_alloc_mw(struct ibv_pd *pd, enum ibv_mw_type type)
{
struct ibv_mw *mw;
struct ibv_alloc_mw cmd;
struct ibv_alloc_mw_resp resp;
int ret;
mw = malloc(sizeof(*mw));
if (!mw)
return NULL;
memset(mw, 0, sizeof(*mw));
ret = ibv_cmd_alloc_mw(pd, type, mw, &cmd, sizeof(cmd), &resp,
sizeof(resp));
if (ret) {
free(mw);
return NULL;
}
return mw;
}
int mlx5_dealloc_mw(struct ibv_mw *mw)
{
int ret;
struct ibv_dealloc_mw cmd;
ret = ibv_cmd_dealloc_mw(mw, &cmd, sizeof(cmd));
if (ret)
return ret;
free(mw);
return 0;
}
int mlx5_round_up_power_of_two(long long sz)
{
long long ret;
for (ret = 1; ret < sz; ret <<= 1)
; /* nothing */
if (ret > INT_MAX) {
fprintf(stderr, "%s: roundup overflow\n", __func__);
return -ENOMEM;
}
return (int)ret;
}
static int align_queue_size(long long req)
{
return mlx5_round_up_power_of_two(req);
}
static int get_cqe_size(void)
{
char *env;
int size = 64;
env = getenv("MLX5_CQE_SIZE");
if (env)
size = atoi(env);
switch (size) {
case 64:
case 128:
return size;
default:
return -EINVAL;
}
}
static int use_scatter_to_cqe(void)
{
char *env;
env = getenv("MLX5_SCATTER_TO_CQE");
if (env && !strcmp(env, "0"))
return 0;
return 1;
}
static int srq_sig_enabled(void)
{
char *env;
env = getenv("MLX5_SRQ_SIGNATURE");
if (env)
return 1;
return 0;
}
static int qp_sig_enabled(void)
{
char *env;
env = getenv("MLX5_QP_SIGNATURE");
if (env)
return 1;
return 0;
}
enum {
CREATE_CQ_SUPPORTED_WC_FLAGS = IBV_WC_STANDARD_FLAGS |
IBV_WC_EX_WITH_COMPLETION_TIMESTAMP |
IBV_WC_EX_WITH_CVLAN |
IBV_WC_EX_WITH_FLOW_TAG
};
enum {
CREATE_CQ_SUPPORTED_COMP_MASK = IBV_CQ_INIT_ATTR_MASK_FLAGS
};
enum {
CREATE_CQ_SUPPORTED_FLAGS = IBV_CREATE_CQ_ATTR_SINGLE_THREADED
};
static struct ibv_cq_ex *create_cq(struct ibv_context *context,
const struct ibv_cq_init_attr_ex *cq_attr,
int cq_alloc_flags,
struct mlx5dv_cq_init_attr *mlx5cq_attr)
{
struct mlx5_create_cq cmd;
struct mlx5_create_cq_resp resp;
struct mlx5_cq *cq;
int cqe_sz;
int ret;
int ncqe;
struct mlx5_context *mctx = to_mctx(context);
FILE *fp = to_mctx(context)->dbg_fp;
if (!cq_attr->cqe) {
mlx5_dbg(fp, MLX5_DBG_CQ, "CQE invalid\n");
errno = EINVAL;
return NULL;
}
if (cq_attr->comp_mask & ~CREATE_CQ_SUPPORTED_COMP_MASK) {
mlx5_dbg(fp, MLX5_DBG_CQ,
"Unsupported comp_mask for create_cq\n");
errno = EINVAL;
return NULL;
}
if (cq_attr->comp_mask & IBV_CQ_INIT_ATTR_MASK_FLAGS &&
cq_attr->flags & ~CREATE_CQ_SUPPORTED_FLAGS) {
mlx5_dbg(fp, MLX5_DBG_CQ,
"Unsupported creation flags requested for create_cq\n");
errno = EINVAL;
return NULL;
}
if (cq_attr->wc_flags & ~CREATE_CQ_SUPPORTED_WC_FLAGS) {
mlx5_dbg(fp, MLX5_DBG_CQ, "\n");
errno = ENOTSUP;
return NULL;
}
cq = calloc(1, sizeof *cq);
if (!cq) {
mlx5_dbg(fp, MLX5_DBG_CQ, "\n");
return NULL;
}
memset(&cmd, 0, sizeof cmd);
cq->cons_index = 0;
if (mlx5_spinlock_init(&cq->lock))
goto err;
ncqe = align_queue_size(cq_attr->cqe + 1);
if ((ncqe > (1 << 24)) || (ncqe < (cq_attr->cqe + 1))) {
mlx5_dbg(fp, MLX5_DBG_CQ, "ncqe %d\n", ncqe);
errno = EINVAL;
goto err_spl;
}
cqe_sz = get_cqe_size();
if (cqe_sz < 0) {
mlx5_dbg(fp, MLX5_DBG_CQ, "\n");
errno = -cqe_sz;
goto err_spl;
}
if (mlx5_alloc_cq_buf(to_mctx(context), cq, &cq->buf_a, ncqe, cqe_sz)) {
mlx5_dbg(fp, MLX5_DBG_CQ, "\n");
goto err_spl;
}
cq->dbrec = mlx5_alloc_dbrec(to_mctx(context));
if (!cq->dbrec) {
mlx5_dbg(fp, MLX5_DBG_CQ, "\n");
goto err_buf;
}
cq->dbrec[MLX5_CQ_SET_CI] = 0;
cq->dbrec[MLX5_CQ_ARM_DB] = 0;
cq->arm_sn = 0;
cq->cqe_sz = cqe_sz;
cq->flags = cq_alloc_flags;
if (cq_attr->comp_mask & IBV_CQ_INIT_ATTR_MASK_FLAGS &&
cq_attr->flags & IBV_CREATE_CQ_ATTR_SINGLE_THREADED)
cq->flags |= MLX5_CQ_FLAGS_SINGLE_THREADED;
cmd.buf_addr = (uintptr_t) cq->buf_a.buf;
cmd.db_addr = (uintptr_t) cq->dbrec;
cmd.cqe_size = cqe_sz;
if (mlx5cq_attr) {
if (mlx5cq_attr->comp_mask & ~(MLX5DV_CQ_INIT_ATTR_MASK_RESERVED - 1)) {
mlx5_dbg(fp, MLX5_DBG_CQ,
"Unsupported vendor comp_mask for create_cq\n");
errno = EINVAL;
goto err_db;
}
if (mlx5cq_attr->comp_mask & MLX5DV_CQ_INIT_ATTR_MASK_COMPRESSED_CQE) {
if (mctx->cqe_comp_caps.max_num &&
(mlx5cq_attr->cqe_comp_res_format &
mctx->cqe_comp_caps.supported_format)) {
cmd.cqe_comp_en = 1;
cmd.cqe_comp_res_format = mlx5cq_attr->cqe_comp_res_format;
} else {
mlx5_dbg(fp, MLX5_DBG_CQ, "CQE Compression is not supported\n");
errno = EINVAL;
goto err_db;
}
}
}
ret = ibv_cmd_create_cq(context, ncqe - 1, cq_attr->channel,
cq_attr->comp_vector,
ibv_cq_ex_to_cq(&cq->ibv_cq), &cmd.ibv_cmd,
sizeof(cmd), &resp.ibv_resp, sizeof(resp));
if (ret) {
mlx5_dbg(fp, MLX5_DBG_CQ, "ret %d\n", ret);
goto err_db;
}
cq->active_buf = &cq->buf_a;
cq->resize_buf = NULL;
cq->cqn = resp.cqn;
cq->stall_enable = to_mctx(context)->stall_enable;
cq->stall_adaptive_enable = to_mctx(context)->stall_adaptive_enable;
cq->stall_cycles = to_mctx(context)->stall_cycles;
if (cq_alloc_flags & MLX5_CQ_FLAGS_EXTENDED)
mlx5_cq_fill_pfns(cq, cq_attr);
return &cq->ibv_cq;
err_db:
mlx5_free_db(to_mctx(context), cq->dbrec);
err_buf:
mlx5_free_cq_buf(to_mctx(context), &cq->buf_a);
err_spl:
mlx5_spinlock_destroy(&cq->lock);
err:
free(cq);
return NULL;
}
struct ibv_cq *mlx5_create_cq(struct ibv_context *context, int cqe,
struct ibv_comp_channel *channel,
int comp_vector)
{
struct ibv_cq_ex *cq;
struct ibv_cq_init_attr_ex cq_attr = {.cqe = cqe, .channel = channel,
.comp_vector = comp_vector,
.wc_flags = IBV_WC_STANDARD_FLAGS};
if (cqe <= 0) {
errno = EINVAL;
return NULL;
}
cq = create_cq(context, &cq_attr, 0, NULL);
return cq ? ibv_cq_ex_to_cq(cq) : NULL;
}
struct ibv_cq_ex *mlx5_create_cq_ex(struct ibv_context *context,
struct ibv_cq_init_attr_ex *cq_attr)
{
return create_cq(context, cq_attr, MLX5_CQ_FLAGS_EXTENDED, NULL);
}
struct ibv_cq_ex *mlx5dv_create_cq(struct ibv_context *context,
struct ibv_cq_init_attr_ex *cq_attr,
struct mlx5dv_cq_init_attr *mlx5_cq_attr)
{
struct ibv_cq_ex *cq;
cq = create_cq(context, cq_attr, MLX5_CQ_FLAGS_EXTENDED, mlx5_cq_attr);
if (!cq)
return NULL;
verbs_init_cq(ibv_cq_ex_to_cq(cq), context,
cq_attr->channel, cq_attr->cq_context);
return cq;
}
int mlx5_resize_cq(struct ibv_cq *ibcq, int cqe)
{
struct mlx5_cq *cq = to_mcq(ibcq);
struct mlx5_resize_cq_resp resp;
struct mlx5_resize_cq cmd;
struct mlx5_context *mctx = to_mctx(ibcq->context);
int err;
if (cqe < 0) {
errno = EINVAL;
return errno;
}
memset(&cmd, 0, sizeof(cmd));
memset(&resp, 0, sizeof(resp));
if (((long long)cqe * 64) > INT_MAX)
return EINVAL;
mlx5_spin_lock(&cq->lock);
cq->active_cqes = cq->ibv_cq.cqe;
if (cq->active_buf == &cq->buf_a)
cq->resize_buf = &cq->buf_b;
else
cq->resize_buf = &cq->buf_a;
cqe = align_queue_size(cqe + 1);
if (cqe == ibcq->cqe + 1) {
cq->resize_buf = NULL;
err = 0;
goto out;
}
/* currently we don't change cqe size */
cq->resize_cqe_sz = cq->cqe_sz;
cq->resize_cqes = cqe;
err = mlx5_alloc_cq_buf(mctx, cq, cq->resize_buf, cq->resize_cqes, cq->resize_cqe_sz);
if (err) {
cq->resize_buf = NULL;
errno = ENOMEM;
goto out;
}
cmd.buf_addr = (uintptr_t)cq->resize_buf->buf;
cmd.cqe_size = cq->resize_cqe_sz;
err = ibv_cmd_resize_cq(ibcq, cqe - 1, &cmd.ibv_cmd, sizeof(cmd),
&resp.ibv_resp, sizeof(resp));
if (err)
goto out_buf;
mlx5_cq_resize_copy_cqes(cq);
mlx5_free_cq_buf(mctx, cq->active_buf);
cq->active_buf = cq->resize_buf;
cq->ibv_cq.cqe = cqe - 1;
mlx5_spin_unlock(&cq->lock);
cq->resize_buf = NULL;
return 0;
out_buf:
mlx5_free_cq_buf(mctx, cq->resize_buf);
cq->resize_buf = NULL;
out:
mlx5_spin_unlock(&cq->lock);
return err;
}
int mlx5_destroy_cq(struct ibv_cq *cq)
{
int ret;
ret = ibv_cmd_destroy_cq(cq);
if (ret)
return ret;
mlx5_free_db(to_mctx(cq->context), to_mcq(cq)->dbrec);
mlx5_free_cq_buf(to_mctx(cq->context), to_mcq(cq)->active_buf);
free(to_mcq(cq));
return 0;
}
struct ibv_srq *mlx5_create_srq(struct ibv_pd *pd,
struct ibv_srq_init_attr *attr)
{
struct mlx5_create_srq cmd;
struct mlx5_create_srq_resp resp;
struct mlx5_srq *srq;
int ret;
struct mlx5_context *ctx;
int max_sge;
struct ibv_srq *ibsrq;
ctx = to_mctx(pd->context);
srq = calloc(1, sizeof *srq);
if (!srq) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
return NULL;
}
ibsrq = &srq->vsrq.srq;
memset(&cmd, 0, sizeof cmd);
if (mlx5_spinlock_init(&srq->lock)) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err;
}
if (attr->attr.max_wr > ctx->max_srq_recv_wr) {
fprintf(stderr, "%s-%d:max_wr %d, max_srq_recv_wr %d\n", __func__, __LINE__,
attr->attr.max_wr, ctx->max_srq_recv_wr);
errno = EINVAL;
goto err;
}
/*
* this calculation does not consider required control segments. The
* final calculation is done again later. This is done so to avoid
* overflows of variables
*/
max_sge = ctx->max_rq_desc_sz / sizeof(struct mlx5_wqe_data_seg);
if (attr->attr.max_sge > max_sge) {
fprintf(stderr, "%s-%d:max_wr %d, max_srq_recv_wr %d\n", __func__, __LINE__,
attr->attr.max_wr, ctx->max_srq_recv_wr);
errno = EINVAL;
goto err;
}
srq->max = align_queue_size(attr->attr.max_wr + 1);
srq->max_gs = attr->attr.max_sge;
srq->counter = 0;
if (mlx5_alloc_srq_buf(pd->context, srq)) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err;
}
srq->db = mlx5_alloc_dbrec(to_mctx(pd->context));
if (!srq->db) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err_free;
}
*srq->db = 0;
cmd.buf_addr = (uintptr_t) srq->buf.buf;
cmd.db_addr = (uintptr_t) srq->db;
srq->wq_sig = srq_sig_enabled();
if (srq->wq_sig)
cmd.flags = MLX5_SRQ_FLAG_SIGNATURE;
attr->attr.max_sge = srq->max_gs;
pthread_mutex_lock(&ctx->srq_table_mutex);
ret = ibv_cmd_create_srq(pd, ibsrq, attr, &cmd.ibv_cmd, sizeof(cmd),
&resp.ibv_resp, sizeof(resp));
if (ret)
goto err_db;
ret = mlx5_store_srq(ctx, resp.srqn, srq);
if (ret)
goto err_destroy;
pthread_mutex_unlock(&ctx->srq_table_mutex);
srq->srqn = resp.srqn;
srq->rsc.rsn = resp.srqn;
srq->rsc.type = MLX5_RSC_TYPE_SRQ;
return ibsrq;
err_destroy:
ibv_cmd_destroy_srq(ibsrq);
err_db:
pthread_mutex_unlock(&ctx->srq_table_mutex);
mlx5_free_db(to_mctx(pd->context), srq->db);
err_free:
free(srq->wrid);
mlx5_free_buf(&srq->buf);
err:
free(srq);
return NULL;
}
int mlx5_modify_srq(struct ibv_srq *srq,
struct ibv_srq_attr *attr,
int attr_mask)
{
struct ibv_modify_srq cmd;
return ibv_cmd_modify_srq(srq, attr, attr_mask, &cmd, sizeof cmd);
}
int mlx5_query_srq(struct ibv_srq *srq,
struct ibv_srq_attr *attr)
{
struct ibv_query_srq cmd;
return ibv_cmd_query_srq(srq, attr, &cmd, sizeof cmd);
}
int mlx5_destroy_srq(struct ibv_srq *srq)
{
int ret;
struct mlx5_srq *msrq = to_msrq(srq);
struct mlx5_context *ctx = to_mctx(srq->context);
ret = ibv_cmd_destroy_srq(srq);
if (ret)
return ret;
if (ctx->cqe_version && msrq->rsc.type == MLX5_RSC_TYPE_XSRQ)
mlx5_clear_uidx(ctx, msrq->rsc.rsn);
else
mlx5_clear_srq(ctx, msrq->srqn);
mlx5_free_db(ctx, msrq->db);
mlx5_free_buf(&msrq->buf);
free(msrq->wrid);
free(msrq);
return 0;
}
static int sq_overhead(enum ibv_qp_type qp_type)
{
size_t size = 0;
size_t mw_bind_size =
sizeof(struct mlx5_wqe_umr_ctrl_seg) +
sizeof(struct mlx5_wqe_mkey_context_seg) +
max_t(size_t, sizeof(struct mlx5_wqe_umr_klm_seg), 64);
switch (qp_type) {
case IBV_QPT_RC:
size += sizeof(struct mlx5_wqe_ctrl_seg) +
max(sizeof(struct mlx5_wqe_atomic_seg) +
sizeof(struct mlx5_wqe_raddr_seg),
mw_bind_size);
break;
case IBV_QPT_UC:
size = sizeof(struct mlx5_wqe_ctrl_seg) +
max(sizeof(struct mlx5_wqe_raddr_seg),
mw_bind_size);
break;
case IBV_QPT_UD:
size = sizeof(struct mlx5_wqe_ctrl_seg) +
sizeof(struct mlx5_wqe_datagram_seg);
break;
case IBV_QPT_XRC_SEND:
size = sizeof(struct mlx5_wqe_ctrl_seg) + mw_bind_size;
SWITCH_FALLTHROUGH;
case IBV_QPT_XRC_RECV:
size = max(size, sizeof(struct mlx5_wqe_ctrl_seg) +
sizeof(struct mlx5_wqe_xrc_seg) +
sizeof(struct mlx5_wqe_raddr_seg));
break;
case IBV_QPT_RAW_PACKET:
size = sizeof(struct mlx5_wqe_ctrl_seg) +
sizeof(struct mlx5_wqe_eth_seg);
break;
default:
return -EINVAL;
}
return size;
}
static int mlx5_calc_send_wqe(struct mlx5_context *ctx,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
int size;
int inl_size = 0;
int max_gather;
int tot_size;
size = sq_overhead(attr->qp_type);
if (size < 0)
return size;
if (attr->cap.max_inline_data) {
inl_size = size + align(sizeof(struct mlx5_wqe_inl_data_seg) +
attr->cap.max_inline_data, 16);
}
if (attr->comp_mask & IBV_QP_INIT_ATTR_MAX_TSO_HEADER) {
size += align(attr->max_tso_header, 16);
qp->max_tso_header = attr->max_tso_header;
}
max_gather = (ctx->max_sq_desc_sz - size) /
sizeof(struct mlx5_wqe_data_seg);
if (attr->cap.max_send_sge > max_gather)
return -EINVAL;
size += attr->cap.max_send_sge * sizeof(struct mlx5_wqe_data_seg);
tot_size = max_int(size, inl_size);
if (tot_size > ctx->max_sq_desc_sz)
return -EINVAL;
return align(tot_size, MLX5_SEND_WQE_BB);
}
static int mlx5_calc_rcv_wqe(struct mlx5_context *ctx,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
uint32_t size;
int num_scatter;
if (attr->srq)
return 0;
num_scatter = max_t(uint32_t, attr->cap.max_recv_sge, 1);
size = sizeof(struct mlx5_wqe_data_seg) * num_scatter;
if (qp->wq_sig)
size += sizeof(struct mlx5_rwqe_sig);
if (size > ctx->max_rq_desc_sz)
return -EINVAL;
size = mlx5_round_up_power_of_two(size);
return size;
}
static int mlx5_calc_sq_size(struct mlx5_context *ctx,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
int wqe_size;
int wq_size;
FILE *fp = ctx->dbg_fp;
if (!attr->cap.max_send_wr)
return 0;
wqe_size = mlx5_calc_send_wqe(ctx, attr, qp);
if (wqe_size < 0) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return wqe_size;
}
if (wqe_size > ctx->max_sq_desc_sz) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return -EINVAL;
}
qp->max_inline_data = wqe_size - sq_overhead(attr->qp_type) -
sizeof(struct mlx5_wqe_inl_data_seg);
attr->cap.max_inline_data = qp->max_inline_data;
/*
* to avoid overflow, we limit max_send_wr so
* that the multiplication will fit in int
*/
if (attr->cap.max_send_wr > 0x7fffffff / ctx->max_sq_desc_sz) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return -EINVAL;
}
wq_size = mlx5_round_up_power_of_two(attr->cap.max_send_wr * wqe_size);
qp->sq.wqe_cnt = wq_size / MLX5_SEND_WQE_BB;
if (qp->sq.wqe_cnt > ctx->max_send_wqebb) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return -EINVAL;
}
qp->sq.wqe_shift = mlx5_ilog2(MLX5_SEND_WQE_BB);
qp->sq.max_gs = attr->cap.max_send_sge;
qp->sq.max_post = wq_size / wqe_size;
return wq_size;
}
static int mlx5_calc_rwq_size(struct mlx5_context *ctx,
struct mlx5_rwq *rwq,
struct ibv_wq_init_attr *attr)
{
size_t wqe_size;
int wq_size;
uint32_t num_scatter;
int scat_spc;
if (!attr->max_wr)
return -EINVAL;
/* TBD: check caps for RQ */
num_scatter = max_t(uint32_t, attr->max_sge, 1);
wqe_size = sizeof(struct mlx5_wqe_data_seg) * num_scatter;
if (rwq->wq_sig)
wqe_size += sizeof(struct mlx5_rwqe_sig);
if (wqe_size <= 0 || wqe_size > ctx->max_rq_desc_sz)
return -EINVAL;
wqe_size = mlx5_round_up_power_of_two(wqe_size);
wq_size = mlx5_round_up_power_of_two(attr->max_wr) * wqe_size;
wq_size = max(wq_size, MLX5_SEND_WQE_BB);
rwq->rq.wqe_cnt = wq_size / wqe_size;
rwq->rq.wqe_shift = mlx5_ilog2(wqe_size);
rwq->rq.max_post = 1 << mlx5_ilog2(wq_size / wqe_size);
scat_spc = wqe_size -
((rwq->wq_sig) ? sizeof(struct mlx5_rwqe_sig) : 0);
rwq->rq.max_gs = scat_spc / sizeof(struct mlx5_wqe_data_seg);
return wq_size;
}
static int mlx5_calc_rq_size(struct mlx5_context *ctx,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
int wqe_size;
int wq_size;
int scat_spc;
FILE *fp = ctx->dbg_fp;
if (!attr->cap.max_recv_wr)
return 0;
if (attr->cap.max_recv_wr > ctx->max_recv_wr) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return -EINVAL;
}
wqe_size = mlx5_calc_rcv_wqe(ctx, attr, qp);
if (wqe_size < 0 || wqe_size > ctx->max_rq_desc_sz) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return -EINVAL;
}
wq_size = mlx5_round_up_power_of_two(attr->cap.max_recv_wr) * wqe_size;
if (wqe_size) {
wq_size = max(wq_size, MLX5_SEND_WQE_BB);
qp->rq.wqe_cnt = wq_size / wqe_size;
qp->rq.wqe_shift = mlx5_ilog2(wqe_size);
qp->rq.max_post = 1 << mlx5_ilog2(wq_size / wqe_size);
scat_spc = wqe_size -
(qp->wq_sig ? sizeof(struct mlx5_rwqe_sig) : 0);
qp->rq.max_gs = scat_spc / sizeof(struct mlx5_wqe_data_seg);
} else {
qp->rq.wqe_cnt = 0;
qp->rq.wqe_shift = 0;
qp->rq.max_post = 0;
qp->rq.max_gs = 0;
}
return wq_size;
}
static int mlx5_calc_wq_size(struct mlx5_context *ctx,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
int ret;
int result;
ret = mlx5_calc_sq_size(ctx, attr, qp);
if (ret < 0)
return ret;
result = ret;
ret = mlx5_calc_rq_size(ctx, attr, qp);
if (ret < 0)
return ret;
result += ret;
qp->sq.offset = ret;
qp->rq.offset = 0;
return result;
}
static void map_uuar(struct ibv_context *context, struct mlx5_qp *qp,
int uuar_index)
{
struct mlx5_context *ctx = to_mctx(context);
qp->bf = &ctx->bfs[uuar_index];
}
static const char *qptype2key(enum ibv_qp_type type)
{
switch (type) {
case IBV_QPT_RC: return "HUGE_RC";
case IBV_QPT_UC: return "HUGE_UC";
case IBV_QPT_UD: return "HUGE_UD";
case IBV_QPT_RAW_PACKET: return "HUGE_RAW_ETH";
default: return "HUGE_NA";
}
}
static int mlx5_alloc_qp_buf(struct ibv_context *context,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp,
int size)
{
int err;
enum mlx5_alloc_type alloc_type;
enum mlx5_alloc_type default_alloc_type = MLX5_ALLOC_TYPE_ANON;
const char *qp_huge_key;
if (qp->sq.wqe_cnt) {
qp->sq.wrid = malloc(qp->sq.wqe_cnt * sizeof(*qp->sq.wrid));
if (!qp->sq.wrid) {
errno = ENOMEM;
err = -1;
return err;
}
qp->sq.wr_data = malloc(qp->sq.wqe_cnt * sizeof(*qp->sq.wr_data));
if (!qp->sq.wr_data) {
errno = ENOMEM;
err = -1;
goto ex_wrid;
}
}
qp->sq.wqe_head = malloc(qp->sq.wqe_cnt * sizeof(*qp->sq.wqe_head));
if (!qp->sq.wqe_head) {
errno = ENOMEM;
err = -1;
goto ex_wrid;
}
if (qp->rq.wqe_cnt) {
qp->rq.wrid = malloc(qp->rq.wqe_cnt * sizeof(uint64_t));
if (!qp->rq.wrid) {
errno = ENOMEM;
err = -1;
goto ex_wrid;
}
}
/* compatibility support */
qp_huge_key = qptype2key(qp->ibv_qp->qp_type);
if (mlx5_use_huge(qp_huge_key))
default_alloc_type = MLX5_ALLOC_TYPE_HUGE;
mlx5_get_alloc_type(MLX5_QP_PREFIX, &alloc_type,
default_alloc_type);
err = mlx5_alloc_prefered_buf(to_mctx(context), &qp->buf,
align(qp->buf_size, to_mdev
(context->device)->page_size),
to_mdev(context->device)->page_size,
alloc_type,
MLX5_QP_PREFIX);
if (err) {
err = -ENOMEM;
goto ex_wrid;
}
memset(qp->buf.buf, 0, qp->buf_size);
if (attr->qp_type == IBV_QPT_RAW_PACKET) {
size_t aligned_sq_buf_size = align(qp->sq_buf_size,
to_mdev(context->device)->page_size);
/* For Raw Packet QP, allocate a separate buffer for the SQ */
err = mlx5_alloc_prefered_buf(to_mctx(context), &qp->sq_buf,
aligned_sq_buf_size,
to_mdev(context->device)->page_size,
alloc_type,
MLX5_QP_PREFIX);
if (err) {
err = -ENOMEM;
goto rq_buf;
}
memset(qp->sq_buf.buf, 0, aligned_sq_buf_size);
}
return 0;
rq_buf:
mlx5_free_actual_buf(to_mctx(qp->verbs_qp.qp.context), &qp->buf);
ex_wrid:
if (qp->rq.wrid)
free(qp->rq.wrid);
if (qp->sq.wqe_head)
free(qp->sq.wqe_head);
if (qp->sq.wr_data)
free(qp->sq.wr_data);
if (qp->sq.wrid)
free(qp->sq.wrid);
return err;
}
static void mlx5_free_qp_buf(struct mlx5_qp *qp)
{
struct mlx5_context *ctx = to_mctx(qp->ibv_qp->context);
mlx5_free_actual_buf(ctx, &qp->buf);
if (qp->sq_buf.buf)
mlx5_free_actual_buf(ctx, &qp->sq_buf);
if (qp->rq.wrid)
free(qp->rq.wrid);
if (qp->sq.wqe_head)
free(qp->sq.wqe_head);
if (qp->sq.wrid)
free(qp->sq.wrid);
if (qp->sq.wr_data)
free(qp->sq.wr_data);
}
static int mlx5_cmd_create_rss_qp(struct ibv_context *context,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_qp *qp)
{
struct mlx5_create_qp_ex_rss cmd_ex_rss = {};
struct mlx5_create_qp_resp_ex resp = {};
int ret;
if (attr->rx_hash_conf.rx_hash_key_len > sizeof(cmd_ex_rss.rx_hash_key)) {
errno = EINVAL;
return errno;
}
cmd_ex_rss.rx_hash_fields_mask = attr->rx_hash_conf.rx_hash_fields_mask;
cmd_ex_rss.rx_hash_function = attr->rx_hash_conf.rx_hash_function;
cmd_ex_rss.rx_key_len = attr->rx_hash_conf.rx_hash_key_len;
memcpy(cmd_ex_rss.rx_hash_key, attr->rx_hash_conf.rx_hash_key,
attr->rx_hash_conf.rx_hash_key_len);
ret = ibv_cmd_create_qp_ex2(context, &qp->verbs_qp,
sizeof(qp->verbs_qp), attr,
&cmd_ex_rss.ibv_cmd, sizeof(cmd_ex_rss.ibv_cmd),
sizeof(cmd_ex_rss), &resp.ibv_resp,
sizeof(resp.ibv_resp), sizeof(resp));
if (ret)
return ret;
qp->rss_qp = 1;
return 0;
}
static int mlx5_cmd_create_qp_ex(struct ibv_context *context,
struct ibv_qp_init_attr_ex *attr,
struct mlx5_create_qp *cmd,
struct mlx5_qp *qp,
struct mlx5_create_qp_resp_ex *resp)
{
struct mlx5_create_qp_ex cmd_ex;
int ret;
memset(&cmd_ex, 0, sizeof(cmd_ex));
memcpy(&cmd_ex.ibv_cmd.base, &cmd->ibv_cmd.user_handle,
offsetof(typeof(cmd->ibv_cmd), is_srq) +
sizeof(cmd->ibv_cmd.is_srq) -
offsetof(typeof(cmd->ibv_cmd), user_handle));
memcpy(&cmd_ex.drv_ex, &cmd->buf_addr,
offsetof(typeof(*cmd), sq_buf_addr) +
sizeof(cmd->sq_buf_addr) - sizeof(cmd->ibv_cmd));
ret = ibv_cmd_create_qp_ex2(context, &qp->verbs_qp,
sizeof(qp->verbs_qp), attr,
&cmd_ex.ibv_cmd, sizeof(cmd_ex.ibv_cmd),
sizeof(cmd_ex), &resp->ibv_resp,
sizeof(resp->ibv_resp), sizeof(*resp));
return ret;
}
enum {
MLX5_CREATE_QP_SUP_COMP_MASK = (IBV_QP_INIT_ATTR_PD |
IBV_QP_INIT_ATTR_XRCD |
IBV_QP_INIT_ATTR_CREATE_FLAGS |
IBV_QP_INIT_ATTR_MAX_TSO_HEADER |
IBV_QP_INIT_ATTR_IND_TABLE |
IBV_QP_INIT_ATTR_RX_HASH),
};
enum {
MLX5_CREATE_QP_EX2_COMP_MASK = (IBV_QP_INIT_ATTR_CREATE_FLAGS |
IBV_QP_INIT_ATTR_MAX_TSO_HEADER |
IBV_QP_INIT_ATTR_IND_TABLE |
IBV_QP_INIT_ATTR_RX_HASH),
};
static struct ibv_qp *create_qp(struct ibv_context *context,
struct ibv_qp_init_attr_ex *attr)
{
struct mlx5_create_qp cmd;
struct mlx5_create_qp_resp resp;
struct mlx5_create_qp_resp_ex resp_ex;
struct mlx5_qp *qp;
int ret;
struct mlx5_context *ctx = to_mctx(context);
struct ibv_qp *ibqp;
int32_t usr_idx = 0;
uint32_t uuar_index;
FILE *fp = ctx->dbg_fp;
if (attr->comp_mask & ~MLX5_CREATE_QP_SUP_COMP_MASK)
return NULL;
if ((attr->comp_mask & IBV_QP_INIT_ATTR_MAX_TSO_HEADER) &&
(attr->qp_type != IBV_QPT_RAW_PACKET))
return NULL;
qp = calloc(1, sizeof(*qp));
if (!qp) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
return NULL;
}
ibqp = (struct ibv_qp *)&qp->verbs_qp;
qp->ibv_qp = ibqp;
memset(&cmd, 0, sizeof(cmd));
memset(&resp, 0, sizeof(resp));
memset(&resp_ex, 0, sizeof(resp_ex));
if (attr->comp_mask & IBV_QP_INIT_ATTR_RX_HASH) {
ret = mlx5_cmd_create_rss_qp(context, attr, qp);
if (ret)
goto err;
return ibqp;
}
qp->wq_sig = qp_sig_enabled();
if (qp->wq_sig)
cmd.flags |= MLX5_QP_FLAG_SIGNATURE;
if (use_scatter_to_cqe())
cmd.flags |= MLX5_QP_FLAG_SCATTER_CQE;
ret = mlx5_calc_wq_size(ctx, attr, qp);
if (ret < 0) {
errno = -ret;
goto err;
}
if (attr->qp_type == IBV_QPT_RAW_PACKET) {
qp->buf_size = qp->sq.offset;
qp->sq_buf_size = ret - qp->buf_size;
qp->sq.offset = 0;
} else {
qp->buf_size = ret;
qp->sq_buf_size = 0;
}
if (mlx5_alloc_qp_buf(context, attr, qp, ret)) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
goto err;
}
if (attr->qp_type == IBV_QPT_RAW_PACKET) {
qp->sq_start = qp->sq_buf.buf;
qp->sq.qend = qp->sq_buf.buf +
(qp->sq.wqe_cnt << qp->sq.wqe_shift);
} else {
qp->sq_start = qp->buf.buf + qp->sq.offset;
qp->sq.qend = qp->buf.buf + qp->sq.offset +
(qp->sq.wqe_cnt << qp->sq.wqe_shift);
}
mlx5_init_qp_indices(qp);
if (mlx5_spinlock_init(&qp->sq.lock) ||
mlx5_spinlock_init(&qp->rq.lock))
goto err_free_qp_buf;
qp->db = mlx5_alloc_dbrec(ctx);
if (!qp->db) {
mlx5_dbg(fp, MLX5_DBG_QP, "\n");
goto err_free_qp_buf;
}
qp->db[MLX5_RCV_DBR] = 0;
qp->db[MLX5_SND_DBR] = 0;
cmd.buf_addr = (uintptr_t) qp->buf.buf;
cmd.sq_buf_addr = (attr->qp_type == IBV_QPT_RAW_PACKET) ?
(uintptr_t) qp->sq_buf.buf : 0;
cmd.db_addr = (uintptr_t) qp->db;
cmd.sq_wqe_count = qp->sq.wqe_cnt;
cmd.rq_wqe_count = qp->rq.wqe_cnt;
cmd.rq_wqe_shift = qp->rq.wqe_shift;
if (ctx->atomic_cap == IBV_ATOMIC_HCA)
qp->atomics_enabled = 1;
if (!ctx->cqe_version) {
cmd.uidx = 0xffffff;
pthread_mutex_lock(&ctx->qp_table_mutex);
} else if (!is_xrc_tgt(attr->qp_type)) {
usr_idx = mlx5_store_uidx(ctx, qp);
if (usr_idx < 0) {
mlx5_dbg(fp, MLX5_DBG_QP, "Couldn't find free user index\n");
goto err_rq_db;
}
cmd.uidx = usr_idx;
}
if (attr->comp_mask & MLX5_CREATE_QP_EX2_COMP_MASK)
ret = mlx5_cmd_create_qp_ex(context, attr, &cmd, qp, &resp_ex);
else
ret = ibv_cmd_create_qp_ex(context, &qp->verbs_qp, sizeof(qp->verbs_qp),
attr, &cmd.ibv_cmd, sizeof(cmd),
&resp.ibv_resp, sizeof(resp));
if (ret) {
mlx5_dbg(fp, MLX5_DBG_QP, "ret %d\n", ret);
goto err_free_uidx;
}
uuar_index = (attr->comp_mask & MLX5_CREATE_QP_EX2_COMP_MASK) ?
resp_ex.uuar_index : resp.uuar_index;
if (!ctx->cqe_version) {
if (qp->sq.wqe_cnt || qp->rq.wqe_cnt) {
ret = mlx5_store_qp(ctx, ibqp->qp_num, qp);
if (ret) {
mlx5_dbg(fp, MLX5_DBG_QP, "ret %d\n", ret);
goto err_destroy;
}
}
pthread_mutex_unlock(&ctx->qp_table_mutex);
}
map_uuar(context, qp, uuar_index);
qp->rq.max_post = qp->rq.wqe_cnt;
if (attr->sq_sig_all)
qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
else
qp->sq_signal_bits = 0;
attr->cap.max_send_wr = qp->sq.max_post;
attr->cap.max_recv_wr = qp->rq.max_post;
attr->cap.max_recv_sge = qp->rq.max_gs;
qp->rsc.type = MLX5_RSC_TYPE_QP;
qp->rsc.rsn = (ctx->cqe_version && !is_xrc_tgt(attr->qp_type)) ?
usr_idx : ibqp->qp_num;
return ibqp;
err_destroy:
ibv_cmd_destroy_qp(ibqp);
err_free_uidx:
if (!ctx->cqe_version)
pthread_mutex_unlock(&to_mctx(context)->qp_table_mutex);
else if (!is_xrc_tgt(attr->qp_type))
mlx5_clear_uidx(ctx, usr_idx);
err_rq_db:
mlx5_free_db(to_mctx(context), qp->db);
err_free_qp_buf:
mlx5_free_qp_buf(qp);
err:
free(qp);
return NULL;
}
struct ibv_qp *mlx5_create_qp(struct ibv_pd *pd,
struct ibv_qp_init_attr *attr)
{
struct ibv_qp *qp;
struct ibv_qp_init_attr_ex attrx;
memset(&attrx, 0, sizeof(attrx));
memcpy(&attrx, attr, sizeof(*attr));
attrx.comp_mask = IBV_QP_INIT_ATTR_PD;
attrx.pd = pd;
qp = create_qp(pd->context, &attrx);
if (qp)
memcpy(attr, &attrx, sizeof(*attr));
return qp;
}
static void mlx5_lock_cqs(struct ibv_qp *qp)
{
struct mlx5_cq *send_cq = to_mcq(qp->send_cq);
struct mlx5_cq *recv_cq = to_mcq(qp->recv_cq);
if (send_cq && recv_cq) {
if (send_cq == recv_cq) {
mlx5_spin_lock(&send_cq->lock);
} else if (send_cq->cqn < recv_cq->cqn) {
mlx5_spin_lock(&send_cq->lock);
mlx5_spin_lock(&recv_cq->lock);
} else {
mlx5_spin_lock(&recv_cq->lock);
mlx5_spin_lock(&send_cq->lock);
}
} else if (send_cq) {
mlx5_spin_lock(&send_cq->lock);
} else if (recv_cq) {
mlx5_spin_lock(&recv_cq->lock);
}
}
static void mlx5_unlock_cqs(struct ibv_qp *qp)
{
struct mlx5_cq *send_cq = to_mcq(qp->send_cq);
struct mlx5_cq *recv_cq = to_mcq(qp->recv_cq);
if (send_cq && recv_cq) {
if (send_cq == recv_cq) {
mlx5_spin_unlock(&send_cq->lock);
} else if (send_cq->cqn < recv_cq->cqn) {
mlx5_spin_unlock(&recv_cq->lock);
mlx5_spin_unlock(&send_cq->lock);
} else {
mlx5_spin_unlock(&send_cq->lock);
mlx5_spin_unlock(&recv_cq->lock);
}
} else if (send_cq) {
mlx5_spin_unlock(&send_cq->lock);
} else if (recv_cq) {
mlx5_spin_unlock(&recv_cq->lock);
}
}
int mlx5_destroy_qp(struct ibv_qp *ibqp)
{
struct mlx5_qp *qp = to_mqp(ibqp);
struct mlx5_context *ctx = to_mctx(ibqp->context);
int ret;
if (qp->rss_qp) {
ret = ibv_cmd_destroy_qp(ibqp);
if (ret)
return ret;
goto free;
}
if (!ctx->cqe_version)
pthread_mutex_lock(&ctx->qp_table_mutex);
ret = ibv_cmd_destroy_qp(ibqp);
if (ret) {
if (!ctx->cqe_version)
pthread_mutex_unlock(&ctx->qp_table_mutex);
return ret;
}
mlx5_lock_cqs(ibqp);
__mlx5_cq_clean(to_mcq(ibqp->recv_cq), qp->rsc.rsn,
ibqp->srq ? to_msrq(ibqp->srq) : NULL);
if (ibqp->send_cq != ibqp->recv_cq)
__mlx5_cq_clean(to_mcq(ibqp->send_cq), qp->rsc.rsn, NULL);
if (!ctx->cqe_version) {
if (qp->sq.wqe_cnt || qp->rq.wqe_cnt)
mlx5_clear_qp(ctx, ibqp->qp_num);
}
mlx5_unlock_cqs(ibqp);
if (!ctx->cqe_version)
pthread_mutex_unlock(&ctx->qp_table_mutex);
else if (!is_xrc_tgt(ibqp->qp_type))
mlx5_clear_uidx(ctx, qp->rsc.rsn);
mlx5_free_db(ctx, qp->db);
mlx5_free_qp_buf(qp);
free:
free(qp);
return 0;
}
int mlx5_query_qp(struct ibv_qp *ibqp, struct ibv_qp_attr *attr,
int attr_mask, struct ibv_qp_init_attr *init_attr)
{
struct ibv_query_qp cmd;
struct mlx5_qp *qp = to_mqp(ibqp);
int ret;
if (qp->rss_qp)
return ENOSYS;
ret = ibv_cmd_query_qp(ibqp, attr, attr_mask, init_attr, &cmd, sizeof(cmd));
if (ret)
return ret;
init_attr->cap.max_send_wr = qp->sq.max_post;
init_attr->cap.max_send_sge = qp->sq.max_gs;
init_attr->cap.max_inline_data = qp->max_inline_data;
attr->cap = init_attr->cap;
return 0;
}
enum {
MLX5_MODIFY_QP_EX_ATTR_MASK = IBV_QP_RATE_LIMIT,
};
int mlx5_modify_qp(struct ibv_qp *qp, struct ibv_qp_attr *attr,
int attr_mask)
{
struct ibv_modify_qp cmd = {};
struct ibv_modify_qp_ex cmd_ex = {};
struct ibv_modify_qp_resp_ex resp = {};
struct mlx5_qp *mqp = to_mqp(qp);
struct mlx5_context *context = to_mctx(qp->context);
int ret;
uint32_t *db;
if (mqp->rss_qp)
return ENOSYS;
if (attr_mask & IBV_QP_PORT) {
switch (qp->qp_type) {
case IBV_QPT_RAW_PACKET:
if (context->cached_link_layer[attr->port_num - 1] ==
IBV_LINK_LAYER_ETHERNET) {
if (context->cached_device_cap_flags &
IBV_DEVICE_RAW_IP_CSUM)
mqp->qp_cap_cache |=
MLX5_CSUM_SUPPORT_RAW_OVER_ETH |
MLX5_RX_CSUM_VALID;
if (ibv_is_qpt_supported(
context->cached_tso_caps.supported_qpts,
IBV_QPT_RAW_PACKET))
mqp->max_tso =
context->cached_tso_caps.max_tso;
}
break;
default:
break;
}
}
if (attr_mask & MLX5_MODIFY_QP_EX_ATTR_MASK)
ret = ibv_cmd_modify_qp_ex(qp, attr, attr_mask,
&cmd_ex,
sizeof(cmd_ex), sizeof(cmd_ex),
&resp,
sizeof(resp), sizeof(resp));
else
ret = ibv_cmd_modify_qp(qp, attr, attr_mask,
&cmd, sizeof(cmd));
if (!ret &&
(attr_mask & IBV_QP_STATE) &&
attr->qp_state == IBV_QPS_RESET) {
if (qp->recv_cq) {
mlx5_cq_clean(to_mcq(qp->recv_cq), mqp->rsc.rsn,
qp->srq ? to_msrq(qp->srq) : NULL);
}
if (qp->send_cq != qp->recv_cq && qp->send_cq)
mlx5_cq_clean(to_mcq(qp->send_cq),
to_mqp(qp)->rsc.rsn, NULL);
mlx5_init_qp_indices(mqp);
db = mqp->db;
db[MLX5_RCV_DBR] = 0;
db[MLX5_SND_DBR] = 0;
}
/*
* When the Raw Packet QP is in INIT state, its RQ
* underneath is already in RDY, which means it can
* receive packets. According to the IB spec, a QP can't
* receive packets until moved to RTR state. To achieve this,
* for Raw Packet QPs, we update the doorbell record
* once the QP is moved to RTR.
*/
if (!ret &&
(attr_mask & IBV_QP_STATE) &&
attr->qp_state == IBV_QPS_RTR &&
qp->qp_type == IBV_QPT_RAW_PACKET) {
mlx5_spin_lock(&mqp->rq.lock);
mqp->db[MLX5_RCV_DBR] = htobe32(mqp->rq.head & 0xffff);
mlx5_spin_unlock(&mqp->rq.lock);
}
return ret;
}
#define RROCE_UDP_SPORT_MIN 0xC000
#define RROCE_UDP_SPORT_MAX 0xFFFF
struct ibv_ah *mlx5_create_ah(struct ibv_pd *pd, struct ibv_ah_attr *attr)
{
struct mlx5_context *ctx = to_mctx(pd->context);
struct ibv_port_attr port_attr;
struct mlx5_ah *ah;
uint32_t gid_type;
uint32_t tmp;
uint8_t grh;
int is_eth;
if (attr->port_num < 1 || attr->port_num > ctx->num_ports)
return NULL;
if (ctx->cached_link_layer[attr->port_num - 1]) {
is_eth = ctx->cached_link_layer[attr->port_num - 1] ==
IBV_LINK_LAYER_ETHERNET;
} else {
if (ibv_query_port(pd->context, attr->port_num, &port_attr))
return NULL;
is_eth = (port_attr.link_layer == IBV_LINK_LAYER_ETHERNET);
}
if (unlikely((!attr->is_global) && is_eth)) {
errno = EINVAL;
return NULL;
}
ah = calloc(1, sizeof *ah);
if (!ah)
return NULL;
if (is_eth) {
if (ibv_query_gid_type(pd->context, attr->port_num,
attr->grh.sgid_index, &gid_type))
goto err;
if (gid_type == IBV_GID_TYPE_ROCE_V2)
ah->av.rlid = htobe16(rand() % (RROCE_UDP_SPORT_MAX + 1
- RROCE_UDP_SPORT_MIN)
+ RROCE_UDP_SPORT_MIN);
/* Since RoCE packets must contain GRH, this bit is reserved
* for RoCE and shouldn't be set.
*/
grh = 0;
} else {
ah->av.fl_mlid = attr->src_path_bits & 0x7f;
ah->av.rlid = htobe16(attr->dlid);
grh = 1;
}
ah->av.stat_rate_sl = (attr->static_rate << 4) | attr->sl;
if (attr->is_global) {
ah->av.tclass = attr->grh.traffic_class;
ah->av.hop_limit = attr->grh.hop_limit;
tmp = htobe32((grh << 30) |
((attr->grh.sgid_index & 0xff) << 20) |
(attr->grh.flow_label & 0xfffff));
ah->av.grh_gid_fl = tmp;
memcpy(ah->av.rgid, attr->grh.dgid.raw, 16);
}
if (is_eth) {
if (ctx->cmds_supp_uhw & MLX5_USER_CMDS_SUPP_UHW_CREATE_AH) {
struct mlx5_create_ah_resp resp = {};
if (ibv_cmd_create_ah(pd, &ah->ibv_ah, attr, &resp.ibv_resp, sizeof(resp)))
goto err;
ah->kern_ah = true;
memcpy(ah->av.rmac, resp.dmac, ETHERNET_LL_SIZE);
} else {
uint16_t vid;
if (ibv_resolve_eth_l2_from_gid(pd->context, attr,
ah->av.rmac, &vid))
goto err;
}
}
return &ah->ibv_ah;
err:
free(ah);
return NULL;
}
int mlx5_destroy_ah(struct ibv_ah *ah)
{
struct mlx5_ah *mah = to_mah(ah);
int err;
if (mah->kern_ah) {
err = ibv_cmd_destroy_ah(ah);
if (err)
return err;
}
free(mah);
return 0;
}
int mlx5_attach_mcast(struct ibv_qp *qp, const union ibv_gid *gid, uint16_t lid)
{
return ibv_cmd_attach_mcast(qp, gid, lid);
}
int mlx5_detach_mcast(struct ibv_qp *qp, const union ibv_gid *gid, uint16_t lid)
{
return ibv_cmd_detach_mcast(qp, gid, lid);
}
struct ibv_qp *mlx5_create_qp_ex(struct ibv_context *context,
struct ibv_qp_init_attr_ex *attr)
{
return create_qp(context, attr);
}
int mlx5_get_srq_num(struct ibv_srq *srq, uint32_t *srq_num)
{
struct mlx5_srq *msrq = to_msrq(srq);
*srq_num = msrq->srqn;
return 0;
}
struct ibv_xrcd *
mlx5_open_xrcd(struct ibv_context *context,
struct ibv_xrcd_init_attr *xrcd_init_attr)
{
int err;
struct verbs_xrcd *xrcd;
struct ibv_open_xrcd cmd = {};
struct ibv_open_xrcd_resp resp = {};
xrcd = calloc(1, sizeof(*xrcd));
if (!xrcd)
return NULL;
err = ibv_cmd_open_xrcd(context, xrcd, sizeof(*xrcd), xrcd_init_attr,
&cmd, sizeof(cmd), &resp, sizeof(resp));
if (err) {
free(xrcd);
return NULL;
}
return &xrcd->xrcd;
}
int mlx5_close_xrcd(struct ibv_xrcd *ib_xrcd)
{
struct verbs_xrcd *xrcd = container_of(ib_xrcd, struct verbs_xrcd, xrcd);
int ret;
ret = ibv_cmd_close_xrcd(xrcd);
if (!ret)
free(xrcd);
return ret;
}
static struct ibv_srq *
mlx5_create_xrc_srq(struct ibv_context *context,
struct ibv_srq_init_attr_ex *attr)
{
int err;
struct mlx5_create_srq_ex cmd;
struct mlx5_create_srq_resp resp;
struct mlx5_srq *msrq;
struct mlx5_context *ctx = to_mctx(context);
int max_sge;
struct ibv_srq *ibsrq;
int uidx;
FILE *fp = ctx->dbg_fp;
msrq = calloc(1, sizeof(*msrq));
if (!msrq)
return NULL;
ibsrq = (struct ibv_srq *)&msrq->vsrq;
memset(&cmd, 0, sizeof(cmd));
memset(&resp, 0, sizeof(resp));
if (mlx5_spinlock_init(&msrq->lock)) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err;
}
if (attr->attr.max_wr > ctx->max_srq_recv_wr) {
fprintf(stderr, "%s-%d:max_wr %d, max_srq_recv_wr %d\n",
__func__, __LINE__, attr->attr.max_wr,
ctx->max_srq_recv_wr);
errno = EINVAL;
goto err;
}
/*
* this calculation does not consider required control segments. The
* final calculation is done again later. This is done so to avoid
* overflows of variables
*/
max_sge = ctx->max_recv_wr / sizeof(struct mlx5_wqe_data_seg);
if (attr->attr.max_sge > max_sge) {
fprintf(stderr, "%s-%d:max_wr %d, max_srq_recv_wr %d\n",
__func__, __LINE__, attr->attr.max_wr,
ctx->max_srq_recv_wr);
errno = EINVAL;
goto err;
}
msrq->max = align_queue_size(attr->attr.max_wr + 1);
msrq->max_gs = attr->attr.max_sge;
msrq->counter = 0;
if (mlx5_alloc_srq_buf(context, msrq)) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err;
}
msrq->db = mlx5_alloc_dbrec(ctx);
if (!msrq->db) {
fprintf(stderr, "%s-%d:\n", __func__, __LINE__);
goto err_free;
}
*msrq->db = 0;
cmd.buf_addr = (uintptr_t)msrq->buf.buf;
cmd.db_addr = (uintptr_t)msrq->db;
msrq->wq_sig = srq_sig_enabled();
if (msrq->wq_sig)
cmd.flags = MLX5_SRQ_FLAG_SIGNATURE;
attr->attr.max_sge = msrq->max_gs;
if (ctx->cqe_version) {
uidx = mlx5_store_uidx(ctx, msrq);
if (uidx < 0) {
mlx5_dbg(fp, MLX5_DBG_QP, "Couldn't find free user index\n");
goto err_free_db;
}
cmd.uidx = uidx;
} else {
cmd.uidx = 0xffffff;
pthread_mutex_lock(&ctx->srq_table_mutex);
}
err = ibv_cmd_create_srq_ex(context, &msrq->vsrq, sizeof(msrq->vsrq),
attr, &cmd.ibv_cmd, sizeof(cmd),
&resp.ibv_resp, sizeof(resp));
if (err)
goto err_free_uidx;
if (!ctx->cqe_version) {
err = mlx5_store_srq(to_mctx(context), resp.srqn, msrq);
if (err)
goto err_destroy;
pthread_mutex_unlock(&ctx->srq_table_mutex);
}
msrq->srqn = resp.srqn;
msrq->rsc.type = MLX5_RSC_TYPE_XSRQ;
msrq->rsc.rsn = ctx->cqe_version ? cmd.uidx : resp.srqn;
return ibsrq;
err_destroy:
ibv_cmd_destroy_srq(ibsrq);
err_free_uidx:
if (ctx->cqe_version)
mlx5_clear_uidx(ctx, cmd.uidx);
else
pthread_mutex_unlock(&ctx->srq_table_mutex);
err_free_db:
mlx5_free_db(ctx, msrq->db);
err_free:
free(msrq->wrid);
mlx5_free_buf(&msrq->buf);
err:
free(msrq);
return NULL;
}
struct ibv_srq *mlx5_create_srq_ex(struct ibv_context *context,
struct ibv_srq_init_attr_ex *attr)
{
if (!(attr->comp_mask & IBV_SRQ_INIT_ATTR_TYPE) ||
(attr->srq_type == IBV_SRQT_BASIC))
return mlx5_create_srq(attr->pd,
(struct ibv_srq_init_attr *)attr);
else if (attr->srq_type == IBV_SRQT_XRC)
return mlx5_create_xrc_srq(context, attr);
return NULL;
}
int mlx5_query_device_ex(struct ibv_context *context,
const struct ibv_query_device_ex_input *input,
struct ibv_device_attr_ex *attr,
size_t attr_size)
{
struct mlx5_context *mctx = to_mctx(context);
struct mlx5_query_device_ex_resp resp;
struct mlx5_query_device_ex cmd;
struct ibv_device_attr *a;
uint64_t raw_fw_ver;
unsigned sub_minor;
unsigned major;
unsigned minor;
int err;
int cmd_supp_uhw = mctx->cmds_supp_uhw &
MLX5_USER_CMDS_SUPP_UHW_QUERY_DEVICE;
memset(&cmd, 0, sizeof(cmd));
memset(&resp, 0, sizeof(resp));
err = ibv_cmd_query_device_ex(context, input, attr, attr_size,
&raw_fw_ver,
&cmd.ibv_cmd, sizeof(cmd.ibv_cmd), sizeof(cmd),
&resp.ibv_resp, sizeof(resp.ibv_resp),
cmd_supp_uhw ? sizeof(resp) : sizeof(resp.ibv_resp));
if (err)
return err;
attr->tso_caps = resp.tso_caps;
attr->rss_caps.rx_hash_fields_mask = resp.rss_caps.rx_hash_fields_mask;
attr->rss_caps.rx_hash_function = resp.rss_caps.rx_hash_function;
attr->packet_pacing_caps = resp.packet_pacing_caps.caps;
if (resp.support_multi_pkt_send_wqe)
mctx->vendor_cap_flags |= MLX5_VENDOR_CAP_FLAGS_MPW;
mctx->cqe_comp_caps = resp.cqe_comp_caps;
major = (raw_fw_ver >> 32) & 0xffff;
minor = (raw_fw_ver >> 16) & 0xffff;
sub_minor = raw_fw_ver & 0xffff;
a = &attr->orig_attr;
snprintf(a->fw_ver, sizeof(a->fw_ver), "%d.%d.%04d",
major, minor, sub_minor);
return 0;
}
static int rwq_sig_enabled(struct ibv_context *context)
{
char *env;
env = getenv("MLX5_RWQ_SIGNATURE");
if (env)
return 1;
return 0;
}
static void mlx5_free_rwq_buf(struct mlx5_rwq *rwq, struct ibv_context *context)
{
struct mlx5_context *ctx = to_mctx(context);
mlx5_free_actual_buf(ctx, &rwq->buf);
free(rwq->rq.wrid);
}
static int mlx5_alloc_rwq_buf(struct ibv_context *context,
struct mlx5_rwq *rwq,
int size)
{
int err;
enum mlx5_alloc_type default_alloc_type = MLX5_ALLOC_TYPE_PREFER_CONTIG;
rwq->rq.wrid = malloc(rwq->rq.wqe_cnt * sizeof(uint64_t));
if (!rwq->rq.wrid) {
errno = ENOMEM;
return -1;
}
err = mlx5_alloc_prefered_buf(to_mctx(context), &rwq->buf,
align(rwq->buf_size, to_mdev
(context->device)->page_size),
to_mdev(context->device)->page_size,
default_alloc_type,
MLX5_RWQ_PREFIX);
if (err) {
free(rwq->rq.wrid);
errno = ENOMEM;
return -1;
}
return 0;
}
struct ibv_wq *mlx5_create_wq(struct ibv_context *context,
struct ibv_wq_init_attr *attr)
{
struct mlx5_create_wq cmd;
struct mlx5_create_wq_resp resp;
int err;
struct mlx5_rwq *rwq;
struct mlx5_context *ctx = to_mctx(context);
int ret;
int32_t usr_idx = 0;
FILE *fp = ctx->dbg_fp;
if (attr->wq_type != IBV_WQT_RQ)
return NULL;
memset(&cmd, 0, sizeof(cmd));
memset(&resp, 0, sizeof(resp));
rwq = calloc(1, sizeof(*rwq));
if (!rwq)
return NULL;
rwq->wq_sig = rwq_sig_enabled(context);
if (rwq->wq_sig)
cmd.drv.flags = MLX5_RWQ_FLAG_SIGNATURE;
ret = mlx5_calc_rwq_size(ctx, rwq, attr);
if (ret < 0) {
errno = -ret;
goto err;
}
rwq->buf_size = ret;
if (mlx5_alloc_rwq_buf(context, rwq, ret))
goto err;
mlx5_init_rwq_indices(rwq);
if (mlx5_spinlock_init(&rwq->rq.lock))
goto err_free_rwq_buf;
rwq->db = mlx5_alloc_dbrec(ctx);
if (!rwq->db)
goto err_free_rwq_buf;
rwq->db[MLX5_RCV_DBR] = 0;
rwq->db[MLX5_SND_DBR] = 0;
rwq->pbuff = rwq->buf.buf + rwq->rq.offset;
rwq->recv_db = &rwq->db[MLX5_RCV_DBR];
cmd.drv.buf_addr = (uintptr_t)rwq->buf.buf;
cmd.drv.db_addr = (uintptr_t)rwq->db;
cmd.drv.rq_wqe_count = rwq->rq.wqe_cnt;
cmd.drv.rq_wqe_shift = rwq->rq.wqe_shift;
usr_idx = mlx5_store_uidx(ctx, rwq);
if (usr_idx < 0) {
mlx5_dbg(fp, MLX5_DBG_QP, "Couldn't find free user index\n");
goto err_free_db_rec;
}
cmd.drv.user_index = usr_idx;
err = ibv_cmd_create_wq(context, attr, &rwq->wq, &cmd.ibv_cmd,
sizeof(cmd.ibv_cmd),
sizeof(cmd),
&resp.ibv_resp, sizeof(resp.ibv_resp),
sizeof(resp));
if (err)
goto err_create;
rwq->rsc.type = MLX5_RSC_TYPE_RWQ;
rwq->rsc.rsn = cmd.drv.user_index;
rwq->wq.post_recv = mlx5_post_wq_recv;
return &rwq->wq;
err_create:
mlx5_clear_uidx(ctx, cmd.drv.user_index);
err_free_db_rec:
mlx5_free_db(to_mctx(context), rwq->db);
err_free_rwq_buf:
mlx5_free_rwq_buf(rwq, context);
err:
free(rwq);
return NULL;
}
int mlx5_modify_wq(struct ibv_wq *wq, struct ibv_wq_attr *attr)
{
struct mlx5_modify_wq cmd = {};
struct mlx5_rwq *rwq = to_mrwq(wq);
if ((attr->attr_mask & IBV_WQ_ATTR_STATE) &&
attr->wq_state == IBV_WQS_RDY) {
if ((attr->attr_mask & IBV_WQ_ATTR_CURR_STATE) &&
attr->curr_wq_state != wq->state)
return -EINVAL;
if (wq->state == IBV_WQS_RESET) {
mlx5_spin_lock(&to_mcq(wq->cq)->lock);
__mlx5_cq_clean(to_mcq(wq->cq),
rwq->rsc.rsn, NULL);
mlx5_spin_unlock(&to_mcq(wq->cq)->lock);
mlx5_init_rwq_indices(rwq);
rwq->db[MLX5_RCV_DBR] = 0;
rwq->db[MLX5_SND_DBR] = 0;
}
}
return ibv_cmd_modify_wq(wq, attr, &cmd.ibv_cmd, sizeof(cmd.ibv_cmd), sizeof(cmd));
}
int mlx5_destroy_wq(struct ibv_wq *wq)
{
struct mlx5_rwq *rwq = to_mrwq(wq);
int ret;
ret = ibv_cmd_destroy_wq(wq);
if (ret)
return ret;
mlx5_spin_lock(&to_mcq(wq->cq)->lock);
__mlx5_cq_clean(to_mcq(wq->cq), rwq->rsc.rsn, NULL);
mlx5_spin_unlock(&to_mcq(wq->cq)->lock);
mlx5_clear_uidx(to_mctx(wq->context), rwq->rsc.rsn);
mlx5_free_db(to_mctx(wq->context), rwq->db);
mlx5_free_rwq_buf(rwq, wq->context);
free(rwq);
return 0;
}
struct ibv_rwq_ind_table *mlx5_create_rwq_ind_table(struct ibv_context *context,
struct ibv_rwq_ind_table_init_attr *init_attr)
{
struct ibv_create_rwq_ind_table *cmd;
struct mlx5_create_rwq_ind_table_resp resp;
struct ibv_rwq_ind_table *ind_table;
uint32_t required_tbl_size;
int num_tbl_entries;
int cmd_size;
int err;
num_tbl_entries = 1 << init_attr->log_ind_tbl_size;
/* Data must be u64 aligned */
required_tbl_size = (num_tbl_entries * sizeof(uint32_t)) < sizeof(uint64_t) ?
sizeof(uint64_t) : (num_tbl_entries * sizeof(uint32_t));
cmd_size = required_tbl_size + sizeof(*cmd);
cmd = calloc(1, cmd_size);
if (!cmd)
return NULL;
memset(&resp, 0, sizeof(resp));
ind_table = calloc(1, sizeof(*ind_table));
if (!ind_table)
goto free_cmd;
err = ibv_cmd_create_rwq_ind_table(context, init_attr, ind_table, cmd,
cmd_size, cmd_size, &resp.ibv_resp, sizeof(resp.ibv_resp),
sizeof(resp));
if (err)
goto err;
free(cmd);
return ind_table;
err:
free(ind_table);
free_cmd:
free(cmd);
return NULL;
}
int mlx5_destroy_rwq_ind_table(struct ibv_rwq_ind_table *rwq_ind_table)
{
int ret;
ret = ibv_cmd_destroy_rwq_ind_table(rwq_ind_table);
if (ret)
return ret;
free(rwq_ind_table);
return 0;
}