summaryrefslogtreecommitdiff
path: root/drivers/infiniband/hw/mlx5/qp.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/infiniband/hw/mlx5/qp.c')
-rw-r--r--drivers/infiniband/hw/mlx5/qp.c5269
1 files changed, 2486 insertions, 2783 deletions
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index dd2ae640bc84..69af20790481 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -30,17 +30,19 @@
* SOFTWARE.
*/
-#include <linux/module.h>
+#include <linux/etherdevice.h>
#include <rdma/ib_umem.h>
#include <rdma/ib_cache.h>
#include <rdma/ib_user_verbs.h>
+#include <rdma/rdma_counter.h>
#include <linux/mlx5/fs.h>
#include "mlx5_ib.h"
#include "ib_rep.h"
+#include "counters.h"
#include "cmd.h"
-
-/* not supported currently */
-static int wq_signature;
+#include "umr.h"
+#include "qp.h"
+#include "wr.h"
enum {
MLX5_IB_ACK_REQ_FREQ = 8,
@@ -53,37 +55,15 @@ enum {
MLX5_IB_LINK_TYPE_ETH = 1
};
-enum {
- MLX5_IB_SQ_STRIDE = 6,
- MLX5_IB_SQ_UMR_INLINE_THRESHOLD = 64,
-};
-
-static const u32 mlx5_ib_opcode[] = {
- [IB_WR_SEND] = MLX5_OPCODE_SEND,
- [IB_WR_LSO] = MLX5_OPCODE_LSO,
- [IB_WR_SEND_WITH_IMM] = MLX5_OPCODE_SEND_IMM,
- [IB_WR_RDMA_WRITE] = MLX5_OPCODE_RDMA_WRITE,
- [IB_WR_RDMA_WRITE_WITH_IMM] = MLX5_OPCODE_RDMA_WRITE_IMM,
- [IB_WR_RDMA_READ] = MLX5_OPCODE_RDMA_READ,
- [IB_WR_ATOMIC_CMP_AND_SWP] = MLX5_OPCODE_ATOMIC_CS,
- [IB_WR_ATOMIC_FETCH_AND_ADD] = MLX5_OPCODE_ATOMIC_FA,
- [IB_WR_SEND_WITH_INV] = MLX5_OPCODE_SEND_INVAL,
- [IB_WR_LOCAL_INV] = MLX5_OPCODE_UMR,
- [IB_WR_REG_MR] = MLX5_OPCODE_UMR,
- [IB_WR_MASKED_ATOMIC_CMP_AND_SWP] = MLX5_OPCODE_ATOMIC_MASKED_CS,
- [IB_WR_MASKED_ATOMIC_FETCH_AND_ADD] = MLX5_OPCODE_ATOMIC_MASKED_FA,
- [MLX5_IB_WR_UMR] = MLX5_OPCODE_UMR,
-};
-
-struct mlx5_wqe_eth_pad {
- u8 rsvd0[16];
-};
-
enum raw_qp_set_mask_map {
MLX5_RAW_QP_MOD_SET_RQ_Q_CTR_ID = 1UL << 0,
MLX5_RAW_QP_RATE_LIMIT = 1UL << 1,
};
+enum {
+ MLX5_QP_RM_GO_BACK_N = 0x1,
+};
+
struct mlx5_modify_raw_qp_param {
u16 operation;
@@ -92,8 +72,17 @@ struct mlx5_modify_raw_qp_param {
struct mlx5_rate_limit rl;
u8 rq_q_ctr_id;
+ u32 port;
};
+struct mlx5_ib_qp_event_work {
+ struct work_struct work;
+ struct mlx5_core_qp *qp;
+ int type;
+};
+
+static struct workqueue_struct *mlx5_ib_qp_event_wq;
+
static void get_cqs(enum ib_qp_type qp_type,
struct ib_cq *ib_send_cq, struct ib_cq *ib_recv_cq,
struct mlx5_ib_cq **send_cq, struct mlx5_ib_cq **recv_cq);
@@ -109,122 +98,336 @@ static int is_sqp(enum ib_qp_type qp_type)
}
/**
- * mlx5_ib_read_user_wqe() - Copy a user-space WQE to kernel space.
+ * mlx5_ib_read_user_wqe_common() - Copy a WQE (or part of) from user WQ
+ * to kernel buffer
*
- * @qp: QP to copy from.
- * @send: copy from the send queue when non-zero, use the receive queue
- * otherwise.
- * @wqe_index: index to start copying from. For send work queues, the
- * wqe_index is in units of MLX5_SEND_WQE_BB.
- * For receive work queue, it is the number of work queue
- * element in the queue.
- * @buffer: destination buffer.
- * @length: maximum number of bytes to copy.
+ * @umem: User space memory where the WQ is
+ * @buffer: buffer to copy to
+ * @buflen: buffer length
+ * @wqe_index: index of WQE to copy from
+ * @wq_offset: offset to start of WQ
+ * @wq_wqe_cnt: number of WQEs in WQ
+ * @wq_wqe_shift: log2 of WQE size
+ * @bcnt: number of bytes to copy
+ * @bytes_copied: number of bytes to copy (return value)
*
- * Copies at least a single WQE, but may copy more data.
+ * Copies from start of WQE bcnt or less bytes.
+ * Does not gurantee to copy the entire WQE.
*
- * Return: the number of bytes copied, or an error code.
+ * Return: zero on success, or an error code.
*/
-int mlx5_ib_read_user_wqe(struct mlx5_ib_qp *qp, int send, int wqe_index,
- void *buffer, u32 length,
- struct mlx5_ib_qp_base *base)
-{
- struct ib_device *ibdev = qp->ibqp.device;
- struct mlx5_ib_dev *dev = to_mdev(ibdev);
- struct mlx5_ib_wq *wq = send ? &qp->sq : &qp->rq;
- size_t offset;
- size_t wq_end;
+static int mlx5_ib_read_user_wqe_common(struct ib_umem *umem, void *buffer,
+ size_t buflen, int wqe_index,
+ int wq_offset, int wq_wqe_cnt,
+ int wq_wqe_shift, int bcnt,
+ size_t *bytes_copied)
+{
+ size_t offset = wq_offset + ((wqe_index % wq_wqe_cnt) << wq_wqe_shift);
+ size_t wq_end = wq_offset + (wq_wqe_cnt << wq_wqe_shift);
+ size_t copy_length;
+ int ret;
+
+ /* don't copy more than requested, more than buffer length or
+ * beyond WQ end
+ */
+ copy_length = min_t(u32, buflen, wq_end - offset);
+ copy_length = min_t(u32, copy_length, bcnt);
+
+ ret = ib_umem_copy_from(buffer, umem, offset, copy_length);
+ if (ret)
+ return ret;
+
+ if (!ret && bytes_copied)
+ *bytes_copied = copy_length;
+
+ return 0;
+}
+
+static int mlx5_ib_read_kernel_wqe_sq(struct mlx5_ib_qp *qp, int wqe_index,
+ void *buffer, size_t buflen, size_t *bc)
+{
+ struct mlx5_wqe_ctrl_seg *ctrl;
+ size_t bytes_copied = 0;
+ size_t wqe_length;
+ void *p;
+ int ds;
+
+ wqe_index = wqe_index & qp->sq.fbc.sz_m1;
+
+ /* read the control segment first */
+ p = mlx5_frag_buf_get_wqe(&qp->sq.fbc, wqe_index);
+ ctrl = p;
+ ds = be32_to_cpu(ctrl->qpn_ds) & MLX5_WQE_CTRL_DS_MASK;
+ wqe_length = ds * MLX5_WQE_DS_UNITS;
+
+ /* read rest of WQE if it spreads over more than one stride */
+ while (bytes_copied < wqe_length) {
+ size_t copy_length =
+ min_t(size_t, buflen - bytes_copied, MLX5_SEND_WQE_BB);
+
+ if (!copy_length)
+ break;
+
+ memcpy(buffer + bytes_copied, p, copy_length);
+ bytes_copied += copy_length;
+
+ wqe_index = (wqe_index + 1) & qp->sq.fbc.sz_m1;
+ p = mlx5_frag_buf_get_wqe(&qp->sq.fbc, wqe_index);
+ }
+ *bc = bytes_copied;
+ return 0;
+}
+
+static int mlx5_ib_read_user_wqe_sq(struct mlx5_ib_qp *qp, int wqe_index,
+ void *buffer, size_t buflen, size_t *bc)
+{
+ struct mlx5_ib_qp_base *base = &qp->trans_qp.base;
struct ib_umem *umem = base->ubuffer.umem;
- u32 first_copy_length;
- int wqe_length;
+ struct mlx5_ib_wq *wq = &qp->sq;
+ struct mlx5_wqe_ctrl_seg *ctrl;
+ size_t bytes_copied;
+ size_t bytes_copied2;
+ size_t wqe_length;
int ret;
+ int ds;
+
+ /* at first read as much as possible */
+ ret = mlx5_ib_read_user_wqe_common(umem, buffer, buflen, wqe_index,
+ wq->offset, wq->wqe_cnt,
+ wq->wqe_shift, buflen,
+ &bytes_copied);
+ if (ret)
+ return ret;
- if (wq->wqe_cnt == 0) {
- mlx5_ib_dbg(dev, "mlx5_ib_read_user_wqe for a QP with wqe_cnt == 0. qp_type: 0x%x\n",
- qp->ibqp.qp_type);
+ /* we need at least control segment size to proceed */
+ if (bytes_copied < sizeof(*ctrl))
return -EINVAL;
+
+ ctrl = buffer;
+ ds = be32_to_cpu(ctrl->qpn_ds) & MLX5_WQE_CTRL_DS_MASK;
+ wqe_length = ds * MLX5_WQE_DS_UNITS;
+
+ /* if we copied enough then we are done */
+ if (bytes_copied >= wqe_length) {
+ *bc = bytes_copied;
+ return 0;
}
- offset = wq->offset + ((wqe_index % wq->wqe_cnt) << wq->wqe_shift);
- wq_end = wq->offset + (wq->wqe_cnt << wq->wqe_shift);
+ /* otherwise this a wrapped around wqe
+ * so read the remaining bytes starting
+ * from wqe_index 0
+ */
+ ret = mlx5_ib_read_user_wqe_common(umem, buffer + bytes_copied,
+ buflen - bytes_copied, 0, wq->offset,
+ wq->wqe_cnt, wq->wqe_shift,
+ wqe_length - bytes_copied,
+ &bytes_copied2);
- if (send && length < sizeof(struct mlx5_wqe_ctrl_seg))
- return -EINVAL;
+ if (ret)
+ return ret;
+ *bc = bytes_copied + bytes_copied2;
+ return 0;
+}
+
+int mlx5_ib_read_wqe_sq(struct mlx5_ib_qp *qp, int wqe_index, void *buffer,
+ size_t buflen, size_t *bc)
+{
+ struct mlx5_ib_qp_base *base = &qp->trans_qp.base;
+ struct ib_umem *umem = base->ubuffer.umem;
- if (offset > umem->length ||
- (send && offset + sizeof(struct mlx5_wqe_ctrl_seg) > umem->length))
+ if (buflen < sizeof(struct mlx5_wqe_ctrl_seg))
return -EINVAL;
- first_copy_length = min_t(u32, offset + length, wq_end) - offset;
- ret = ib_umem_copy_from(buffer, umem, offset, first_copy_length);
+ if (!umem)
+ return mlx5_ib_read_kernel_wqe_sq(qp, wqe_index, buffer,
+ buflen, bc);
+
+ return mlx5_ib_read_user_wqe_sq(qp, wqe_index, buffer, buflen, bc);
+}
+
+static int mlx5_ib_read_user_wqe_rq(struct mlx5_ib_qp *qp, int wqe_index,
+ void *buffer, size_t buflen, size_t *bc)
+{
+ struct mlx5_ib_qp_base *base = &qp->trans_qp.base;
+ struct ib_umem *umem = base->ubuffer.umem;
+ struct mlx5_ib_wq *wq = &qp->rq;
+ size_t bytes_copied;
+ int ret;
+
+ ret = mlx5_ib_read_user_wqe_common(umem, buffer, buflen, wqe_index,
+ wq->offset, wq->wqe_cnt,
+ wq->wqe_shift, buflen,
+ &bytes_copied);
+
if (ret)
return ret;
+ *bc = bytes_copied;
+ return 0;
+}
- if (send) {
- struct mlx5_wqe_ctrl_seg *ctrl = buffer;
- int ds = be32_to_cpu(ctrl->qpn_ds) & MLX5_WQE_CTRL_DS_MASK;
+int mlx5_ib_read_wqe_rq(struct mlx5_ib_qp *qp, int wqe_index, void *buffer,
+ size_t buflen, size_t *bc)
+{
+ struct mlx5_ib_qp_base *base = &qp->trans_qp.base;
+ struct ib_umem *umem = base->ubuffer.umem;
+ struct mlx5_ib_wq *wq = &qp->rq;
+ size_t wqe_size = 1 << wq->wqe_shift;
- wqe_length = ds * MLX5_WQE_DS_UNITS;
- } else {
- wqe_length = 1 << wq->wqe_shift;
- }
+ if (buflen < wqe_size)
+ return -EINVAL;
- if (wqe_length <= first_copy_length)
- return first_copy_length;
+ if (!umem)
+ return -EOPNOTSUPP;
+
+ return mlx5_ib_read_user_wqe_rq(qp, wqe_index, buffer, buflen, bc);
+}
+
+static int mlx5_ib_read_user_wqe_srq(struct mlx5_ib_srq *srq, int wqe_index,
+ void *buffer, size_t buflen, size_t *bc)
+{
+ struct ib_umem *umem = srq->umem;
+ size_t bytes_copied;
+ int ret;
+
+ ret = mlx5_ib_read_user_wqe_common(umem, buffer, buflen, wqe_index, 0,
+ srq->msrq.max, srq->msrq.wqe_shift,
+ buflen, &bytes_copied);
- ret = ib_umem_copy_from(buffer + first_copy_length, umem, wq->offset,
- wqe_length - first_copy_length);
if (ret)
return ret;
+ *bc = bytes_copied;
+ return 0;
+}
+
+int mlx5_ib_read_wqe_srq(struct mlx5_ib_srq *srq, int wqe_index, void *buffer,
+ size_t buflen, size_t *bc)
+{
+ struct ib_umem *umem = srq->umem;
+ size_t wqe_size = 1 << srq->msrq.wqe_shift;
+
+ if (buflen < wqe_size)
+ return -EINVAL;
- return wqe_length;
+ if (!umem)
+ return -EOPNOTSUPP;
+
+ return mlx5_ib_read_user_wqe_srq(srq, wqe_index, buffer, buflen, bc);
+}
+
+static void mlx5_ib_qp_err_syndrome(struct ib_qp *ibqp)
+{
+ struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
+ int outlen = MLX5_ST_SZ_BYTES(query_qp_out);
+ struct mlx5_ib_qp *qp = to_mqp(ibqp);
+ void *pas_ext_union, *err_syn;
+ u32 *outb;
+ int err;
+
+ if (!MLX5_CAP_GEN(dev->mdev, qpc_extension) ||
+ !MLX5_CAP_GEN(dev->mdev, qp_error_syndrome))
+ return;
+
+ outb = kzalloc(outlen, GFP_KERNEL);
+ if (!outb)
+ return;
+
+ err = mlx5_core_qp_query(dev, &qp->trans_qp.base.mqp, outb, outlen,
+ true);
+ if (err)
+ goto out;
+
+ pas_ext_union =
+ MLX5_ADDR_OF(query_qp_out, outb, qp_pas_or_qpc_ext_and_pas);
+ err_syn = MLX5_ADDR_OF(qpc_extension_and_pas_list_in, pas_ext_union,
+ qpc_data_extension.error_syndrome);
+
+ pr_err("%s/%d: QP %d error: %s (0x%x 0x%x 0x%x)\n",
+ ibqp->device->name, ibqp->port, ibqp->qp_num,
+ ib_wc_status_msg(
+ MLX5_GET(cqe_error_syndrome, err_syn, syndrome)),
+ MLX5_GET(cqe_error_syndrome, err_syn, vendor_error_syndrome),
+ MLX5_GET(cqe_error_syndrome, err_syn, hw_syndrome_type),
+ MLX5_GET(cqe_error_syndrome, err_syn, hw_error_syndrome));
+out:
+ kfree(outb);
+}
+
+static void mlx5_ib_handle_qp_event(struct work_struct *_work)
+{
+ struct mlx5_ib_qp_event_work *qpe_work =
+ container_of(_work, struct mlx5_ib_qp_event_work, work);
+ struct ib_qp *ibqp = &to_mibqp(qpe_work->qp)->ibqp;
+ struct ib_event event = {};
+
+ event.device = ibqp->device;
+ event.element.qp = ibqp;
+ switch (qpe_work->type) {
+ case MLX5_EVENT_TYPE_PATH_MIG:
+ event.event = IB_EVENT_PATH_MIG;
+ break;
+ case MLX5_EVENT_TYPE_COMM_EST:
+ event.event = IB_EVENT_COMM_EST;
+ break;
+ case MLX5_EVENT_TYPE_SQ_DRAINED:
+ event.event = IB_EVENT_SQ_DRAINED;
+ break;
+ case MLX5_EVENT_TYPE_SRQ_LAST_WQE:
+ event.event = IB_EVENT_QP_LAST_WQE_REACHED;
+ break;
+ case MLX5_EVENT_TYPE_WQ_CATAS_ERROR:
+ event.event = IB_EVENT_QP_FATAL;
+ break;
+ case MLX5_EVENT_TYPE_PATH_MIG_FAILED:
+ event.event = IB_EVENT_PATH_MIG_ERR;
+ break;
+ case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
+ event.event = IB_EVENT_QP_REQ_ERR;
+ break;
+ case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR:
+ event.event = IB_EVENT_QP_ACCESS_ERR;
+ break;
+ default:
+ pr_warn("mlx5_ib: Unexpected event type %d on QP %06x\n",
+ qpe_work->type, qpe_work->qp->qpn);
+ goto out;
+ }
+
+ if ((event.event == IB_EVENT_QP_FATAL) ||
+ (event.event == IB_EVENT_QP_ACCESS_ERR))
+ mlx5_ib_qp_err_syndrome(ibqp);
+
+ ibqp->event_handler(&event, ibqp->qp_context);
+
+out:
+ mlx5_core_res_put(&qpe_work->qp->common);
+ kfree(qpe_work);
}
static void mlx5_ib_qp_event(struct mlx5_core_qp *qp, int type)
{
struct ib_qp *ibqp = &to_mibqp(qp)->ibqp;
- struct ib_event event;
+ struct mlx5_ib_qp_event_work *qpe_work;
if (type == MLX5_EVENT_TYPE_PATH_MIG) {
/* This event is only valid for trans_qps */
to_mibqp(qp)->port = to_mibqp(qp)->trans_qp.alt_port;
}
- if (ibqp->event_handler) {
- event.device = ibqp->device;
- event.element.qp = ibqp;
- switch (type) {
- case MLX5_EVENT_TYPE_PATH_MIG:
- event.event = IB_EVENT_PATH_MIG;
- break;
- case MLX5_EVENT_TYPE_COMM_EST:
- event.event = IB_EVENT_COMM_EST;
- break;
- case MLX5_EVENT_TYPE_SQ_DRAINED:
- event.event = IB_EVENT_SQ_DRAINED;
- break;
- case MLX5_EVENT_TYPE_SRQ_LAST_WQE:
- event.event = IB_EVENT_QP_LAST_WQE_REACHED;
- break;
- case MLX5_EVENT_TYPE_WQ_CATAS_ERROR:
- event.event = IB_EVENT_QP_FATAL;
- break;
- case MLX5_EVENT_TYPE_PATH_MIG_FAILED:
- event.event = IB_EVENT_PATH_MIG_ERR;
- break;
- case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
- event.event = IB_EVENT_QP_REQ_ERR;
- break;
- case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR:
- event.event = IB_EVENT_QP_ACCESS_ERR;
- break;
- default:
- pr_warn("mlx5_ib: Unexpected event type %d on QP %06x\n", type, qp->qpn);
- return;
- }
+ if (!ibqp->event_handler)
+ goto out_no_handler;
- ibqp->event_handler(&event, ibqp->qp_context);
- }
+ qpe_work = kzalloc(sizeof(*qpe_work), GFP_ATOMIC);
+ if (!qpe_work)
+ goto out_no_handler;
+
+ qpe_work->qp = qp;
+ qpe_work->type = type;
+ INIT_WORK(&qpe_work->work, mlx5_ib_handle_qp_event);
+ queue_work(mlx5_ib_qp_event_wq, &qpe_work->work);
+ return;
+
+out_no_handler:
+ mlx5_core_res_put(&qp->common);
}
static int set_rq_size(struct mlx5_ib_dev *dev, struct ib_qp_cap *cap,
@@ -244,17 +447,26 @@ static int set_rq_size(struct mlx5_ib_dev *dev, struct ib_qp_cap *cap,
cap->max_recv_wr = 0;
cap->max_recv_sge = 0;
} else {
+ int wq_sig = !!(qp->flags_en & MLX5_QP_FLAG_SIGNATURE);
+
if (ucmd) {
qp->rq.wqe_cnt = ucmd->rq_wqe_count;
if (ucmd->rq_wqe_shift > BITS_PER_BYTE * sizeof(ucmd->rq_wqe_shift))
return -EINVAL;
qp->rq.wqe_shift = ucmd->rq_wqe_shift;
- if ((1 << qp->rq.wqe_shift) / sizeof(struct mlx5_wqe_data_seg) < qp->wq_sig)
+ if ((1 << qp->rq.wqe_shift) /
+ sizeof(struct mlx5_wqe_data_seg) <
+ wq_sig)
return -EINVAL;
- qp->rq.max_gs = (1 << qp->rq.wqe_shift) / sizeof(struct mlx5_wqe_data_seg) - qp->wq_sig;
+ qp->rq.max_gs =
+ (1 << qp->rq.wqe_shift) /
+ sizeof(struct mlx5_wqe_data_seg) -
+ wq_sig;
qp->rq.max_post = qp->rq.wqe_cnt;
} else {
- wqe_size = qp->wq_sig ? sizeof(struct mlx5_wqe_signature_seg) : 0;
+ wqe_size =
+ wq_sig ? sizeof(struct mlx5_wqe_signature_seg) :
+ 0;
wqe_size += cap->max_recv_sge * sizeof(struct mlx5_wqe_data_seg);
wqe_size = roundup_pow_of_two(wqe_size);
wq_size = roundup_pow_of_two(cap->max_recv_wr) * wqe_size;
@@ -268,7 +480,10 @@ static int set_rq_size(struct mlx5_ib_dev *dev, struct ib_qp_cap *cap,
return -EINVAL;
}
qp->rq.wqe_shift = ilog2(wqe_size);
- qp->rq.max_gs = (1 << qp->rq.wqe_shift) / sizeof(struct mlx5_wqe_data_seg) - qp->wq_sig;
+ qp->rq.max_gs =
+ (1 << qp->rq.wqe_shift) /
+ sizeof(struct mlx5_wqe_data_seg) -
+ wq_sig;
qp->rq.max_post = qp->rq.wqe_cnt;
}
}
@@ -283,7 +498,7 @@ static int sq_overhead(struct ib_qp_init_attr *attr)
switch (attr->qp_type) {
case IB_QPT_XRC_INI:
size += sizeof(struct mlx5_wqe_xrc_seg);
- /* fall through */
+ fallthrough;
case IB_QPT_RC:
size += sizeof(struct mlx5_wqe_ctrl_seg) +
max(sizeof(struct mlx5_wqe_atomic_seg) +
@@ -308,7 +523,7 @@ static int sq_overhead(struct ib_qp_init_attr *attr)
if (attr->create_flags & IB_QP_CREATE_IPOIB_UD_LSO)
size += sizeof(struct mlx5_wqe_eth_pad) +
sizeof(struct mlx5_wqe_eth_seg);
- /* fall through */
+ fallthrough;
case IB_QPT_SMI:
case MLX5_IB_QPT_HW_GSI:
size += sizeof(struct mlx5_wqe_ctrl_seg) +
@@ -343,9 +558,9 @@ static int calc_send_wqe(struct ib_qp_init_attr *attr)
}
size += attr->cap.max_send_sge * sizeof(struct mlx5_wqe_data_seg);
- if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN &&
+ if (attr->create_flags & IB_QP_CREATE_INTEGRITY_EN &&
ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB) < MLX5_SIG_WQE_SIZE)
- return MLX5_SIG_WQE_SIZE;
+ return MLX5_SIG_WQE_SIZE;
else
return ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB);
}
@@ -397,9 +612,6 @@ static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
sizeof(struct mlx5_wqe_inline_seg);
attr->cap.max_inline_data = qp->max_inline_data;
- if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN)
- qp->signature_en = true;
-
wq_size = roundup_pow_of_two(attr->cap.max_send_wr * wqe_size);
qp->sq.wqe_cnt = wq_size / MLX5_SEND_WQE_BB;
if (qp->sq.wqe_cnt > (1 << MLX5_CAP_GEN(dev->mdev, log_max_qp_sz))) {
@@ -435,9 +647,9 @@ static int set_user_buf_size(struct mlx5_ib_dev *dev,
return -EINVAL;
}
- if (ucmd->sq_wqe_count && ((1 << ilog2(ucmd->sq_wqe_count)) != ucmd->sq_wqe_count)) {
- mlx5_ib_warn(dev, "sq_wqe_count %d, sq_wqe_count %d\n",
- ucmd->sq_wqe_count, ucmd->sq_wqe_count);
+ if (ucmd->sq_wqe_count && !is_power_of_2(ucmd->sq_wqe_count)) {
+ mlx5_ib_warn(dev, "sq_wqe_count %d is not a power of two\n",
+ ucmd->sq_wqe_count);
return -EINVAL;
}
@@ -451,7 +663,7 @@ static int set_user_buf_size(struct mlx5_ib_dev *dev,
}
if (attr->qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) {
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) {
base->ubuffer.buf_size = qp->rq.wqe_cnt << qp->rq.wqe_shift;
qp->raw_packet_qp.sq.ubuffer.buf_size = qp->sq.wqe_cnt << 6;
} else {
@@ -484,7 +696,8 @@ enum {
static int max_bfregs(struct mlx5_ib_dev *dev, struct mlx5_bfreg_info *bfregi)
{
- return get_num_static_uars(dev, bfregi) * MLX5_NON_FP_BFREGS_PER_UAR;
+ return get_uars_per_sys_page(dev, bfregi->lib_uar_4k) *
+ bfregi->num_static_sys_pages * MLX5_NON_FP_BFREGS_PER_UAR;
}
static int num_med_bfreg(struct mlx5_ib_dev *dev,
@@ -553,6 +766,9 @@ static int alloc_bfreg(struct mlx5_ib_dev *dev,
{
int bfregn = -ENOMEM;
+ if (bfregi->lib_uar_dyn)
+ return -EINVAL;
+
mutex_lock(&bfregi->lock);
if (bfregi->ver >= 2) {
bfregn = alloc_high_class_bfreg(dev, bfregi);
@@ -603,10 +819,7 @@ static int to_mlx5_st(enum ib_qp_type type)
case IB_QPT_SMI: return MLX5_QP_ST_QP0;
case MLX5_IB_QPT_HW_GSI: return MLX5_QP_ST_QP1;
case MLX5_IB_QPT_DCI: return MLX5_QP_ST_DCI;
- case IB_QPT_RAW_IPV6: return MLX5_QP_ST_RAW_IPV6;
- case IB_QPT_RAW_PACKET:
- case IB_QPT_RAW_ETHERTYPE: return MLX5_QP_ST_RAW_ETHERTYPE;
- case IB_QPT_MAX:
+ case IB_QPT_RAW_PACKET: return MLX5_QP_ST_RAW_ETHERTYPE;
default: return -EINVAL;
}
}
@@ -624,6 +837,9 @@ int bfregn_to_uar_index(struct mlx5_ib_dev *dev,
u32 index_of_sys_page;
u32 offset;
+ if (bfregi->lib_uar_dyn)
+ return -EINVAL;
+
bfregs_per_sys_page = get_uars_per_sys_page(dev, bfregi->lib_uar_4k) *
MLX5_NON_FP_BFREGS_PER_UAR;
index_of_sys_page = bfregn / bfregs_per_sys_page;
@@ -645,103 +861,69 @@ int bfregn_to_uar_index(struct mlx5_ib_dev *dev,
return bfregi->sys_pages[index_of_sys_page] + offset;
}
-static int mlx5_ib_umem_get(struct mlx5_ib_dev *dev,
- struct ib_pd *pd,
- unsigned long addr, size_t size,
- struct ib_umem **umem,
- int *npages, int *page_shift, int *ncont,
- u32 *offset)
-{
- int err;
-
- *umem = ib_umem_get(pd->uobject->context, addr, size, 0, 0);
- if (IS_ERR(*umem)) {
- mlx5_ib_dbg(dev, "umem_get failed\n");
- return PTR_ERR(*umem);
- }
-
- mlx5_ib_cont_pages(*umem, addr, 0, npages, page_shift, ncont, NULL);
-
- err = mlx5_ib_get_buf_offset(addr, *page_shift, offset);
- if (err) {
- mlx5_ib_warn(dev, "bad offset\n");
- goto err_umem;
- }
-
- mlx5_ib_dbg(dev, "addr 0x%lx, size %zu, npages %d, page_shift %d, ncont %d, offset %d\n",
- addr, size, *npages, *page_shift, *ncont, *offset);
-
- return 0;
-
-err_umem:
- ib_umem_release(*umem);
- *umem = NULL;
-
- return err;
-}
-
static void destroy_user_rq(struct mlx5_ib_dev *dev, struct ib_pd *pd,
- struct mlx5_ib_rwq *rwq)
+ struct mlx5_ib_rwq *rwq, struct ib_udata *udata)
{
- struct mlx5_ib_ucontext *context;
+ struct mlx5_ib_ucontext *context =
+ rdma_udata_to_drv_context(
+ udata,
+ struct mlx5_ib_ucontext,
+ ibucontext);
if (rwq->create_flags & MLX5_IB_WQ_FLAGS_DELAY_DROP)
atomic_dec(&dev->delay_drop.rqs_cnt);
- context = to_mucontext(pd->uobject->context);
mlx5_ib_db_unmap_user(context, &rwq->db);
- if (rwq->umem)
- ib_umem_release(rwq->umem);
+ ib_umem_release(rwq->umem);
}
static int create_user_rq(struct mlx5_ib_dev *dev, struct ib_pd *pd,
- struct mlx5_ib_rwq *rwq,
+ struct ib_udata *udata, struct mlx5_ib_rwq *rwq,
struct mlx5_ib_create_wq *ucmd)
{
- struct mlx5_ib_ucontext *context;
- int page_shift = 0;
- int npages;
+ struct mlx5_ib_ucontext *ucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
+ unsigned long page_size = 0;
u32 offset = 0;
- int ncont = 0;
int err;
if (!ucmd->buf_addr)
return -EINVAL;
- context = to_mucontext(pd->uobject->context);
- rwq->umem = ib_umem_get(pd->uobject->context, ucmd->buf_addr,
- rwq->buf_size, 0, 0);
+ rwq->umem = ib_umem_get(&dev->ib_dev, ucmd->buf_addr, rwq->buf_size, 0);
if (IS_ERR(rwq->umem)) {
mlx5_ib_dbg(dev, "umem_get failed\n");
err = PTR_ERR(rwq->umem);
return err;
}
- mlx5_ib_cont_pages(rwq->umem, ucmd->buf_addr, 0, &npages, &page_shift,
- &ncont, NULL);
- err = mlx5_ib_get_buf_offset(ucmd->buf_addr, page_shift,
- &rwq->rq_page_offset);
- if (err) {
+ page_size = mlx5_umem_find_best_quantized_pgoff(
+ rwq->umem, wq, log_wq_pg_sz, MLX5_ADAPTER_PAGE_SHIFT,
+ page_offset, 64, &rwq->rq_page_offset);
+ if (!page_size) {
mlx5_ib_warn(dev, "bad offset\n");
+ err = -EINVAL;
goto err_umem;
}
- rwq->rq_num_pas = ncont;
- rwq->page_shift = page_shift;
- rwq->log_page_size = page_shift - MLX5_ADAPTER_PAGE_SHIFT;
+ rwq->rq_num_pas = ib_umem_num_dma_blocks(rwq->umem, page_size);
+ rwq->page_shift = order_base_2(page_size);
+ rwq->log_page_size = rwq->page_shift - MLX5_ADAPTER_PAGE_SHIFT;
rwq->wq_sig = !!(ucmd->flags & MLX5_WQ_FLAG_SIGNATURE);
- mlx5_ib_dbg(dev, "addr 0x%llx, size %zd, npages %d, page_shift %d, ncont %d, offset %d\n",
- (unsigned long long)ucmd->buf_addr, rwq->buf_size,
- npages, page_shift, ncont, offset);
+ mlx5_ib_dbg(
+ dev,
+ "addr 0x%llx, size %zd, npages %zu, page_size %ld, ncont %d, offset %d\n",
+ (unsigned long long)ucmd->buf_addr, rwq->buf_size,
+ ib_umem_num_pages(rwq->umem), page_size, rwq->rq_num_pas,
+ offset);
- err = mlx5_ib_db_map_user(context, ucmd->db_addr, &rwq->db);
+ err = mlx5_ib_db_map_user(ucontext, ucmd->db_addr, &rwq->db);
if (err) {
mlx5_ib_dbg(dev, "map failed\n");
goto err_umem;
}
- rwq->create_type = MLX5_WQ_USER;
return 0;
err_umem:
@@ -756,52 +938,51 @@ static int adjust_bfregn(struct mlx5_ib_dev *dev,
bfregn % MLX5_NON_FP_BFREGS_PER_UAR;
}
-static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
- struct mlx5_ib_qp *qp, struct ib_udata *udata,
- struct ib_qp_init_attr *attr,
- u32 **in,
- struct mlx5_ib_create_qp_resp *resp, int *inlen,
- struct mlx5_ib_qp_base *base)
+static int _create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp, struct ib_udata *udata,
+ struct ib_qp_init_attr *attr, u32 **in,
+ struct mlx5_ib_create_qp_resp *resp, int *inlen,
+ struct mlx5_ib_qp_base *base,
+ struct mlx5_ib_create_qp *ucmd)
{
struct mlx5_ib_ucontext *context;
- struct mlx5_ib_create_qp ucmd;
struct mlx5_ib_ubuffer *ubuffer = &base->ubuffer;
- int page_shift = 0;
+ unsigned int page_offset_quantized = 0;
+ unsigned long page_size = 0;
int uar_index = 0;
- int npages;
- u32 offset = 0;
int bfregn;
int ncont = 0;
__be64 *pas;
void *qpc;
int err;
u16 uid;
-
- err = ib_copy_from_udata(&ucmd, udata, sizeof(ucmd));
- if (err) {
- mlx5_ib_dbg(dev, "copy failed\n");
- return err;
- }
-
- context = to_mucontext(pd->uobject->context);
- if (ucmd.flags & MLX5_QP_FLAG_BFREG_INDEX) {
+ u32 uar_flags;
+
+ context = rdma_udata_to_drv_context(udata, struct mlx5_ib_ucontext,
+ ibucontext);
+ uar_flags = qp->flags_en &
+ (MLX5_QP_FLAG_UAR_PAGE_INDEX | MLX5_QP_FLAG_BFREG_INDEX);
+ switch (uar_flags) {
+ case MLX5_QP_FLAG_UAR_PAGE_INDEX:
+ uar_index = ucmd->bfreg_index;
+ bfregn = MLX5_IB_INVALID_BFREG;
+ break;
+ case MLX5_QP_FLAG_BFREG_INDEX:
uar_index = bfregn_to_uar_index(dev, &context->bfregi,
- ucmd.bfreg_index, true);
+ ucmd->bfreg_index, true);
if (uar_index < 0)
return uar_index;
-
bfregn = MLX5_IB_INVALID_BFREG;
- } else if (qp->flags & MLX5_IB_QP_CROSS_CHANNEL) {
- /*
- * TBD: should come from the verbs when we have the API
- */
- /* In CROSS_CHANNEL CQ and QP must use the same UAR */
- bfregn = MLX5_CROSS_CHANNEL_BFREG;
- }
- else {
+ break;
+ case 0:
+ if (qp->flags & IB_QP_CREATE_CROSS_CHANNEL)
+ return -EINVAL;
bfregn = alloc_bfreg(dev, &context->bfregi);
if (bfregn < 0)
return bfregn;
+ break;
+ default:
+ return -EINVAL;
}
mlx5_ib_dbg(dev, "bfregn 0x%x, uar_index 0x%x\n", bfregn, uar_index);
@@ -813,18 +994,27 @@ static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
qp->sq.wqe_shift = ilog2(MLX5_SEND_WQE_BB);
qp->sq.offset = qp->rq.wqe_cnt << qp->rq.wqe_shift;
- err = set_user_buf_size(dev, qp, &ucmd, base, attr);
+ err = set_user_buf_size(dev, qp, ucmd, base, attr);
if (err)
goto err_bfreg;
- if (ucmd.buf_addr && ubuffer->buf_size) {
- ubuffer->buf_addr = ucmd.buf_addr;
- err = mlx5_ib_umem_get(dev, pd, ubuffer->buf_addr,
- ubuffer->buf_size,
- &ubuffer->umem, &npages, &page_shift,
- &ncont, &offset);
- if (err)
+ if (ucmd->buf_addr && ubuffer->buf_size) {
+ ubuffer->buf_addr = ucmd->buf_addr;
+ ubuffer->umem = ib_umem_get(&dev->ib_dev, ubuffer->buf_addr,
+ ubuffer->buf_size, 0);
+ if (IS_ERR(ubuffer->umem)) {
+ err = PTR_ERR(ubuffer->umem);
goto err_bfreg;
+ }
+ page_size = mlx5_umem_find_best_quantized_pgoff(
+ ubuffer->umem, qpc, log_page_size,
+ MLX5_ADAPTER_PAGE_SHIFT, page_offset, 64,
+ &page_offset_quantized);
+ if (!page_size) {
+ err = -EINVAL;
+ goto err_umem;
+ }
+ ncont = ib_umem_num_dma_blocks(ubuffer->umem, page_size);
} else {
ubuffer->umem = NULL;
}
@@ -837,18 +1027,16 @@ static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
goto err_umem;
}
- uid = (attr->qp_type != IB_QPT_XRC_TGT &&
- attr->qp_type != IB_QPT_XRC_INI) ? to_mpd(pd)->uid : 0;
+ uid = (attr->qp_type != IB_QPT_XRC_INI) ? to_mpd(pd)->uid : 0;
MLX5_SET(create_qp_in, *in, uid, uid);
- pas = (__be64 *)MLX5_ADDR_OF(create_qp_in, *in, pas);
- if (ubuffer->umem)
- mlx5_ib_populate_pas(dev, ubuffer->umem, page_shift, pas, 0);
-
qpc = MLX5_ADDR_OF(create_qp_in, *in, qpc);
-
- MLX5_SET(qpc, qpc, log_page_size, page_shift - MLX5_ADAPTER_PAGE_SHIFT);
- MLX5_SET(qpc, qpc, page_offset, offset);
-
+ pas = (__be64 *)MLX5_ADDR_OF(create_qp_in, *in, pas);
+ if (ubuffer->umem) {
+ mlx5_ib_populate_pas(ubuffer->umem, page_size, pas, 0);
+ MLX5_SET(qpc, qpc, log_page_size,
+ order_base_2(page_size) - MLX5_ADAPTER_PAGE_SHIFT);
+ MLX5_SET(qpc, qpc, page_offset, page_offset_quantized);
+ }
MLX5_SET(qpc, qpc, uar_page, uar_index);
if (bfregn != MLX5_IB_INVALID_BFREG)
resp->bfreg_index = adjust_bfregn(dev, &context->bfregi, bfregn);
@@ -856,30 +1044,19 @@ static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
resp->bfreg_index = MLX5_IB_INVALID_BFREG;
qp->bfregn = bfregn;
- err = mlx5_ib_db_map_user(context, ucmd.db_addr, &qp->db);
+ err = mlx5_ib_db_map_user(context, ucmd->db_addr, &qp->db);
if (err) {
mlx5_ib_dbg(dev, "map failed\n");
goto err_free;
}
- err = ib_copy_to_udata(udata, resp, min(udata->outlen, sizeof(*resp)));
- if (err) {
- mlx5_ib_dbg(dev, "copy failed\n");
- goto err_unmap;
- }
- qp->create_type = MLX5_QP_USER;
-
return 0;
-err_unmap:
- mlx5_ib_db_unmap_user(context, &qp->db);
-
err_free:
kvfree(*in);
err_umem:
- if (ubuffer->umem)
- ib_umem_release(ubuffer->umem);
+ ib_umem_release(ubuffer->umem);
err_bfreg:
if (bfregn != MLX5_IB_INVALID_BFREG)
@@ -887,65 +1064,47 @@ err_bfreg:
return err;
}
-static void destroy_qp_user(struct mlx5_ib_dev *dev, struct ib_pd *pd,
- struct mlx5_ib_qp *qp, struct mlx5_ib_qp_base *base)
+static void destroy_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ struct mlx5_ib_qp_base *base, struct ib_udata *udata)
{
- struct mlx5_ib_ucontext *context;
+ struct mlx5_ib_ucontext *context = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
- context = to_mucontext(pd->uobject->context);
- mlx5_ib_db_unmap_user(context, &qp->db);
- if (base->ubuffer.umem)
+ if (udata) {
+ /* User QP */
+ mlx5_ib_db_unmap_user(context, &qp->db);
ib_umem_release(base->ubuffer.umem);
- /*
- * Free only the BFREGs which are handled by the kernel.
- * BFREGs of UARs allocated dynamically are handled by user.
- */
- if (qp->bfregn != MLX5_IB_INVALID_BFREG)
- mlx5_ib_free_bfreg(dev, &context->bfregi, qp->bfregn);
-}
-
-/* get_sq_edge - Get the next nearby edge.
- *
- * An 'edge' is defined as the first following address after the end
- * of the fragment or the SQ. Accordingly, during the WQE construction
- * which repetitively increases the pointer to write the next data, it
- * simply should check if it gets to an edge.
- *
- * @sq - SQ buffer.
- * @idx - Stride index in the SQ buffer.
- *
- * Return:
- * The new edge.
- */
-static void *get_sq_edge(struct mlx5_ib_wq *sq, u32 idx)
-{
- void *fragment_end;
-
- fragment_end = mlx5_frag_buf_get_wqe
- (&sq->fbc,
- mlx5_frag_buf_get_idx_last_contig_stride(&sq->fbc, idx));
+ /*
+ * Free only the BFREGs which are handled by the kernel.
+ * BFREGs of UARs allocated dynamically are handled by user.
+ */
+ if (qp->bfregn != MLX5_IB_INVALID_BFREG)
+ mlx5_ib_free_bfreg(dev, &context->bfregi, qp->bfregn);
+ return;
+ }
- return fragment_end + MLX5_SEND_WQE_BB;
+ /* Kernel QP */
+ kvfree(qp->sq.wqe_head);
+ kvfree(qp->sq.w_list);
+ kvfree(qp->sq.wrid);
+ kvfree(qp->sq.wr_data);
+ kvfree(qp->rq.wrid);
+ if (qp->db.db)
+ mlx5_db_free(dev->mdev, &qp->db);
+ if (qp->buf.frags)
+ mlx5_frag_buf_free(dev->mdev, &qp->buf);
}
-static int create_kernel_qp(struct mlx5_ib_dev *dev,
- struct ib_qp_init_attr *init_attr,
- struct mlx5_ib_qp *qp,
- u32 **in, int *inlen,
- struct mlx5_ib_qp_base *base)
+static int _create_kernel_qp(struct mlx5_ib_dev *dev,
+ struct ib_qp_init_attr *init_attr,
+ struct mlx5_ib_qp *qp, u32 **in, int *inlen,
+ struct mlx5_ib_qp_base *base)
{
int uar_index;
void *qpc;
int err;
- if (init_attr->create_flags & ~(IB_QP_CREATE_SIGNATURE_EN |
- IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK |
- IB_QP_CREATE_IPOIB_UD_LSO |
- IB_QP_CREATE_NETIF_QP |
- mlx5_ib_create_qp_sqpn_qp1()))
- return -EINVAL;
-
if (init_attr->qp_type == MLX5_IB_QPT_REG_UMR)
qp->bf.bfreg = &dev->fp_bfreg;
else
@@ -1000,16 +1159,15 @@ static int create_kernel_qp(struct mlx5_ib_dev *dev,
qpc = MLX5_ADDR_OF(create_qp_in, *in, qpc);
MLX5_SET(qpc, qpc, uar_page, uar_index);
+ MLX5_SET(qpc, qpc, ts_format, mlx5_get_qp_default_ts(dev->mdev));
MLX5_SET(qpc, qpc, log_page_size, qp->buf.page_shift - MLX5_ADAPTER_PAGE_SHIFT);
/* Set "fast registration enabled" for all kernel QPs */
MLX5_SET(qpc, qpc, fre, 1);
MLX5_SET(qpc, qpc, rlky, 1);
- if (init_attr->create_flags & mlx5_ib_create_qp_sqpn_qp1()) {
+ if (qp->flags & MLX5_IB_QP_CREATE_SQPN_QP1)
MLX5_SET(qpc, qpc, deth_sqpn, 1);
- qp->flags |= MLX5_IB_QP_SQPN_QP1;
- }
mlx5_fill_page_frag_array(&qp->buf,
(__be64 *)MLX5_ADDR_OF(create_qp_in,
@@ -1037,7 +1195,6 @@ static int create_kernel_qp(struct mlx5_ib_dev *dev,
err = -ENOMEM;
goto err_wrid;
}
- qp->create_type = MLX5_QP_KERNEL;
return 0;
@@ -1057,36 +1214,15 @@ err_buf:
return err;
}
-static void destroy_qp_kernel(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp)
-{
- kvfree(qp->sq.wqe_head);
- kvfree(qp->sq.w_list);
- kvfree(qp->sq.wrid);
- kvfree(qp->sq.wr_data);
- kvfree(qp->rq.wrid);
- mlx5_db_free(dev->mdev, &qp->db);
- mlx5_frag_buf_free(dev->mdev, &qp->buf);
-}
-
static u32 get_rx_type(struct mlx5_ib_qp *qp, struct ib_qp_init_attr *attr)
{
- if (attr->srq || (attr->qp_type == IB_QPT_XRC_TGT) ||
- (attr->qp_type == MLX5_IB_QPT_DCI) ||
- (attr->qp_type == IB_QPT_XRC_INI))
+ if (attr->srq || (qp->type == IB_QPT_XRC_TGT) ||
+ (qp->type == MLX5_IB_QPT_DCI) || (qp->type == IB_QPT_XRC_INI))
return MLX5_SRQ_RQ;
else if (!qp->has_rq)
return MLX5_ZERO_LEN_RQ;
- else
- return MLX5_NON_ZERO_RQ;
-}
-
-static int is_connected(enum ib_qp_type qp_type)
-{
- if (qp_type == IB_QPT_RC || qp_type == IB_QPT_UC ||
- qp_type == MLX5_IB_QPT_DCI)
- return 1;
- return 0;
+ return MLX5_NON_ZERO_RQ;
}
static int create_raw_packet_qp_tis(struct mlx5_ib_dev *dev,
@@ -1094,15 +1230,18 @@ static int create_raw_packet_qp_tis(struct mlx5_ib_dev *dev,
struct mlx5_ib_sq *sq, u32 tdn,
struct ib_pd *pd)
{
- u32 in[MLX5_ST_SZ_DW(create_tis_in)] = {0};
+ u32 in[MLX5_ST_SZ_DW(create_tis_in)] = {};
void *tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
MLX5_SET(create_tis_in, in, uid, to_mpd(pd)->uid);
MLX5_SET(tisc, tisc, transport_domain, tdn);
- if (qp->flags & MLX5_IB_QP_UNDERLAY)
+ if (!mlx5_ib_lag_should_assign_affinity(dev) &&
+ mlx5_lag_is_lacp_owner(dev->mdev))
+ MLX5_SET(tisc, tisc, strict_lag_tx_port_affinity, 1);
+ if (qp->flags & IB_QP_CREATE_SOURCE_QPN)
MLX5_SET(tisc, tisc, underlay_qpn, qp->underlay_qpn);
- return mlx5_core_create_tis(dev->mdev, in, sizeof(in), &sq->tisn);
+ return mlx5_core_create_tis(dev->mdev, in, &sq->tisn);
}
static void destroy_raw_packet_qp_tis(struct mlx5_ib_dev *dev,
@@ -1111,16 +1250,92 @@ static void destroy_raw_packet_qp_tis(struct mlx5_ib_dev *dev,
mlx5_cmd_destroy_tis(dev->mdev, sq->tisn, to_mpd(pd)->uid);
}
-static void destroy_flow_rule_vport_sq(struct mlx5_ib_dev *dev,
- struct mlx5_ib_sq *sq)
+static void destroy_flow_rule_vport_sq(struct mlx5_ib_sq *sq)
{
if (sq->flow_rule)
mlx5_del_flow_rules(sq->flow_rule);
+ sq->flow_rule = NULL;
+}
+
+static bool fr_supported(int ts_cap)
+{
+ return ts_cap == MLX5_TIMESTAMP_FORMAT_CAP_FREE_RUNNING ||
+ ts_cap == MLX5_TIMESTAMP_FORMAT_CAP_FREE_RUNNING_AND_REAL_TIME;
+}
+
+static int get_ts_format(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *cq,
+ bool fr_sup, bool rt_sup)
+{
+ if (cq->private_flags & MLX5_IB_CQ_PR_FLAGS_REAL_TIME_TS) {
+ if (!rt_sup) {
+ mlx5_ib_dbg(dev,
+ "Real time TS format is not supported\n");
+ return -EOPNOTSUPP;
+ }
+ return MLX5_TIMESTAMP_FORMAT_REAL_TIME;
+ }
+ if (cq->create_flags & IB_UVERBS_CQ_FLAGS_TIMESTAMP_COMPLETION) {
+ if (!fr_sup) {
+ mlx5_ib_dbg(dev,
+ "Free running TS format is not supported\n");
+ return -EOPNOTSUPP;
+ }
+ return MLX5_TIMESTAMP_FORMAT_FREE_RUNNING;
+ }
+ return fr_sup ? MLX5_TIMESTAMP_FORMAT_FREE_RUNNING :
+ MLX5_TIMESTAMP_FORMAT_DEFAULT;
+}
+
+static int get_rq_ts_format(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *recv_cq)
+{
+ u8 ts_cap = MLX5_CAP_GEN(dev->mdev, rq_ts_format);
+
+ return get_ts_format(dev, recv_cq, fr_supported(ts_cap),
+ rt_supported(ts_cap));
+}
+
+static int get_sq_ts_format(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *send_cq)
+{
+ u8 ts_cap = MLX5_CAP_GEN(dev->mdev, sq_ts_format);
+
+ return get_ts_format(dev, send_cq, fr_supported(ts_cap),
+ rt_supported(ts_cap));
+}
+
+static int get_qp_ts_format(struct mlx5_ib_dev *dev, struct mlx5_ib_cq *send_cq,
+ struct mlx5_ib_cq *recv_cq)
+{
+ u8 ts_cap = MLX5_CAP_ROCE(dev->mdev, qp_ts_format);
+ bool fr_sup = fr_supported(ts_cap);
+ bool rt_sup = rt_supported(ts_cap);
+ u8 default_ts = fr_sup ? MLX5_TIMESTAMP_FORMAT_FREE_RUNNING :
+ MLX5_TIMESTAMP_FORMAT_DEFAULT;
+ int send_ts_format =
+ send_cq ? get_ts_format(dev, send_cq, fr_sup, rt_sup) :
+ default_ts;
+ int recv_ts_format =
+ recv_cq ? get_ts_format(dev, recv_cq, fr_sup, rt_sup) :
+ default_ts;
+
+ if (send_ts_format < 0 || recv_ts_format < 0)
+ return -EOPNOTSUPP;
+
+ if (send_ts_format != MLX5_TIMESTAMP_FORMAT_DEFAULT &&
+ recv_ts_format != MLX5_TIMESTAMP_FORMAT_DEFAULT &&
+ send_ts_format != recv_ts_format) {
+ mlx5_ib_dbg(
+ dev,
+ "The send ts_format does not match the receive ts_format\n");
+ return -EOPNOTSUPP;
+ }
+
+ return send_ts_format == default_ts ? recv_ts_format : send_ts_format;
}
static int create_raw_packet_qp_sq(struct mlx5_ib_dev *dev,
+ struct ib_udata *udata,
struct mlx5_ib_sq *sq, void *qpin,
- struct ib_pd *pd)
+ struct ib_pd *pd, struct mlx5_ib_cq *cq)
{
struct mlx5_ib_ubuffer *ubuffer = &sq->ubuffer;
__be64 *pas;
@@ -1130,18 +1345,29 @@ static int create_raw_packet_qp_sq(struct mlx5_ib_dev *dev,
void *wq;
int inlen;
int err;
- int page_shift = 0;
- int npages;
- int ncont = 0;
- u32 offset = 0;
-
- err = mlx5_ib_umem_get(dev, pd, ubuffer->buf_addr, ubuffer->buf_size,
- &sq->ubuffer.umem, &npages, &page_shift,
- &ncont, &offset);
- if (err)
- return err;
+ unsigned int page_offset_quantized;
+ unsigned long page_size;
+ int ts_format;
+
+ ts_format = get_sq_ts_format(dev, cq);
+ if (ts_format < 0)
+ return ts_format;
+
+ sq->ubuffer.umem = ib_umem_get(&dev->ib_dev, ubuffer->buf_addr,
+ ubuffer->buf_size, 0);
+ if (IS_ERR(sq->ubuffer.umem))
+ return PTR_ERR(sq->ubuffer.umem);
+ page_size = mlx5_umem_find_best_quantized_pgoff(
+ ubuffer->umem, wq, log_wq_pg_sz, MLX5_ADAPTER_PAGE_SHIFT,
+ page_offset, 64, &page_offset_quantized);
+ if (!page_size) {
+ err = -EINVAL;
+ goto err_umem;
+ }
- inlen = MLX5_ST_SZ_BYTES(create_sq_in) + sizeof(u64) * ncont;
+ inlen = MLX5_ST_SZ_BYTES(create_sq_in) +
+ sizeof(u64) *
+ ib_umem_num_dma_blocks(sq->ubuffer.umem, page_size);
in = kvzalloc(inlen, GFP_KERNEL);
if (!in) {
err = -ENOMEM;
@@ -1154,6 +1380,7 @@ static int create_raw_packet_qp_sq(struct mlx5_ib_dev *dev,
if (MLX5_CAP_ETH(dev->mdev, multi_pkt_send_wqe))
MLX5_SET(sqc, sqc, allow_multi_pkt_send_wqe, 1);
MLX5_SET(sqc, sqc, state, MLX5_SQC_STATE_RST);
+ MLX5_SET(sqc, sqc, ts_format, ts_format);
MLX5_SET(sqc, sqc, user_index, MLX5_GET(qpc, qpc, user_index));
MLX5_SET(sqc, sqc, cqn, MLX5_GET(qpc, qpc, cqn_snd));
MLX5_SET(sqc, sqc, tis_lst_sz, 1);
@@ -1169,28 +1396,22 @@ static int create_raw_packet_qp_sq(struct mlx5_ib_dev *dev,
MLX5_SET64(wq, wq, dbr_addr, MLX5_GET64(qpc, qpc, dbr_addr));
MLX5_SET(wq, wq, log_wq_stride, ilog2(MLX5_SEND_WQE_BB));
MLX5_SET(wq, wq, log_wq_sz, MLX5_GET(qpc, qpc, log_sq_size));
- MLX5_SET(wq, wq, log_wq_pg_sz, page_shift - MLX5_ADAPTER_PAGE_SHIFT);
- MLX5_SET(wq, wq, page_offset, offset);
+ MLX5_SET(wq, wq, log_wq_pg_sz,
+ order_base_2(page_size) - MLX5_ADAPTER_PAGE_SHIFT);
+ MLX5_SET(wq, wq, page_offset, page_offset_quantized);
pas = (__be64 *)MLX5_ADDR_OF(wq, wq, pas);
- mlx5_ib_populate_pas(dev, sq->ubuffer.umem, page_shift, pas, 0);
+ mlx5_ib_populate_pas(sq->ubuffer.umem, page_size, pas, 0);
- err = mlx5_core_create_sq_tracked(dev->mdev, in, inlen, &sq->base.mqp);
+ err = mlx5_core_create_sq_tracked(dev, in, inlen, &sq->base.mqp);
kvfree(in);
if (err)
goto err_umem;
- err = create_flow_rule_vport_sq(dev, sq);
- if (err)
- goto err_flow;
-
return 0;
-err_flow:
- mlx5_core_destroy_sq_tracked(dev->mdev, &sq->base.mqp);
-
err_umem:
ib_umem_release(sq->ubuffer.umem);
sq->ubuffer.umem = NULL;
@@ -1201,45 +1422,41 @@ err_umem:
static void destroy_raw_packet_qp_sq(struct mlx5_ib_dev *dev,
struct mlx5_ib_sq *sq)
{
- destroy_flow_rule_vport_sq(dev, sq);
- mlx5_core_destroy_sq_tracked(dev->mdev, &sq->base.mqp);
+ destroy_flow_rule_vport_sq(sq);
+ mlx5_core_destroy_sq_tracked(dev, &sq->base.mqp);
ib_umem_release(sq->ubuffer.umem);
}
-static size_t get_rq_pas_size(void *qpc)
-{
- u32 log_page_size = MLX5_GET(qpc, qpc, log_page_size) + 12;
- u32 log_rq_stride = MLX5_GET(qpc, qpc, log_rq_stride);
- u32 log_rq_size = MLX5_GET(qpc, qpc, log_rq_size);
- u32 page_offset = MLX5_GET(qpc, qpc, page_offset);
- u32 po_quanta = 1 << (log_page_size - 6);
- u32 rq_sz = 1 << (log_rq_size + 4 + log_rq_stride);
- u32 page_size = 1 << log_page_size;
- u32 rq_sz_po = rq_sz + (page_offset * po_quanta);
- u32 rq_num_pas = (rq_sz_po + page_size - 1) / page_size;
-
- return rq_num_pas * sizeof(u64);
-}
-
static int create_raw_packet_qp_rq(struct mlx5_ib_dev *dev,
struct mlx5_ib_rq *rq, void *qpin,
- size_t qpinlen, struct ib_pd *pd)
+ struct ib_pd *pd, struct mlx5_ib_cq *cq)
{
struct mlx5_ib_qp *mqp = rq->base.container_mibqp;
__be64 *pas;
- __be64 *qp_pas;
void *in;
void *rqc;
void *wq;
void *qpc = MLX5_ADDR_OF(create_qp_in, qpin, qpc);
- size_t rq_pas_size = get_rq_pas_size(qpc);
+ struct ib_umem *umem = rq->base.ubuffer.umem;
+ unsigned int page_offset_quantized;
+ unsigned long page_size = 0;
+ int ts_format;
size_t inlen;
int err;
- if (qpinlen < rq_pas_size + MLX5_BYTE_OFF(create_qp_in, pas))
+ ts_format = get_rq_ts_format(dev, cq);
+ if (ts_format < 0)
+ return ts_format;
+
+ page_size = mlx5_umem_find_best_quantized_pgoff(umem, wq, log_wq_pg_sz,
+ MLX5_ADAPTER_PAGE_SHIFT,
+ page_offset, 64,
+ &page_offset_quantized);
+ if (!page_size)
return -EINVAL;
- inlen = MLX5_ST_SZ_BYTES(create_rq_in) + rq_pas_size;
+ inlen = MLX5_ST_SZ_BYTES(create_rq_in) +
+ sizeof(u64) * ib_umem_num_dma_blocks(umem, page_size);
in = kvzalloc(inlen, GFP_KERNEL);
if (!in)
return -ENOMEM;
@@ -1250,29 +1467,30 @@ static int create_raw_packet_qp_rq(struct mlx5_ib_dev *dev,
MLX5_SET(rqc, rqc, vsd, 1);
MLX5_SET(rqc, rqc, mem_rq_type, MLX5_RQC_MEM_RQ_TYPE_MEMORY_RQ_INLINE);
MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RST);
+ MLX5_SET(rqc, rqc, ts_format, ts_format);
MLX5_SET(rqc, rqc, flush_in_error_en, 1);
MLX5_SET(rqc, rqc, user_index, MLX5_GET(qpc, qpc, user_index));
MLX5_SET(rqc, rqc, cqn, MLX5_GET(qpc, qpc, cqn_rcv));
- if (mqp->flags & MLX5_IB_QP_CAP_SCATTER_FCS)
+ if (mqp->flags & IB_QP_CREATE_SCATTER_FCS)
MLX5_SET(rqc, rqc, scatter_fcs, 1);
wq = MLX5_ADDR_OF(rqc, rqc, wq);
MLX5_SET(wq, wq, wq_type, MLX5_WQ_TYPE_CYCLIC);
if (rq->flags & MLX5_IB_RQ_PCI_WRITE_END_PADDING)
MLX5_SET(wq, wq, end_padding_mode, MLX5_WQ_END_PAD_MODE_ALIGN);
- MLX5_SET(wq, wq, page_offset, MLX5_GET(qpc, qpc, page_offset));
+ MLX5_SET(wq, wq, page_offset, page_offset_quantized);
MLX5_SET(wq, wq, pd, MLX5_GET(qpc, qpc, pd));
MLX5_SET64(wq, wq, dbr_addr, MLX5_GET64(qpc, qpc, dbr_addr));
MLX5_SET(wq, wq, log_wq_stride, MLX5_GET(qpc, qpc, log_rq_stride) + 4);
- MLX5_SET(wq, wq, log_wq_pg_sz, MLX5_GET(qpc, qpc, log_page_size));
+ MLX5_SET(wq, wq, log_wq_pg_sz,
+ order_base_2(page_size) - MLX5_ADAPTER_PAGE_SHIFT);
MLX5_SET(wq, wq, log_wq_sz, MLX5_GET(qpc, qpc, log_rq_size));
pas = (__be64 *)MLX5_ADDR_OF(wq, wq, pas);
- qp_pas = (__be64 *)MLX5_ADDR_OF(create_qp_in, qpin, pas);
- memcpy(pas, qp_pas, rq_pas_size);
+ mlx5_ib_populate_pas(umem, page_size, pas, 0);
- err = mlx5_core_create_rq_tracked(dev->mdev, in, inlen, &rq->base.mqp);
+ err = mlx5_core_create_rq_tracked(dev, in, inlen, &rq->base.mqp);
kvfree(in);
@@ -1282,14 +1500,7 @@ static int create_raw_packet_qp_rq(struct mlx5_ib_dev *dev,
static void destroy_raw_packet_qp_rq(struct mlx5_ib_dev *dev,
struct mlx5_ib_rq *rq)
{
- mlx5_core_destroy_rq_tracked(dev->mdev, &rq->base.mqp);
-}
-
-static bool tunnel_offload_supported(struct mlx5_core_dev *dev)
-{
- return (MLX5_CAP_ETH(dev, tunnel_stateless_vxlan) ||
- MLX5_CAP_ETH(dev, tunnel_stateless_gre) ||
- MLX5_CAP_ETH(dev, tunnel_stateless_geneve_rx));
+ mlx5_core_destroy_rq_tracked(dev, &rq->base.mqp);
}
static void destroy_raw_packet_qp_tir(struct mlx5_ib_dev *dev,
@@ -1305,8 +1516,8 @@ static void destroy_raw_packet_qp_tir(struct mlx5_ib_dev *dev,
static int create_raw_packet_qp_tir(struct mlx5_ib_dev *dev,
struct mlx5_ib_rq *rq, u32 tdn,
- u32 *qp_flags_en,
- struct ib_pd *pd)
+ u32 *qp_flags_en, struct ib_pd *pd,
+ u32 *out)
{
u8 lb_flag = 0;
u32 *in;
@@ -1333,15 +1544,15 @@ static int create_raw_packet_qp_tir(struct mlx5_ib_dev *dev,
if (*qp_flags_en & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC)
lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_MULTICAST;
- if (dev->rep) {
+ if (dev->is_rep) {
lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_UNICAST;
*qp_flags_en |= MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC;
}
MLX5_SET(tirc, tirc, self_lb_block, lb_flag);
-
- err = mlx5_core_create_tir(dev->mdev, in, inlen, &rq->tirn);
-
+ MLX5_SET(create_tir_in, in, opcode, MLX5_CMD_OP_CREATE_TIR);
+ err = mlx5_cmd_exec_inout(dev->mdev, create_tir, in, out);
+ rq->tirn = MLX5_GET(create_tir_out, out, tirn);
if (!err && MLX5_GET(tirc, tirc, self_lb_block)) {
err = mlx5_ib_enable_lb(dev, false, true);
@@ -1354,27 +1565,30 @@ static int create_raw_packet_qp_tir(struct mlx5_ib_dev *dev,
}
static int create_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
- u32 *in, size_t inlen,
- struct ib_pd *pd,
+ u32 *in, size_t inlen, struct ib_pd *pd,
struct ib_udata *udata,
- struct mlx5_ib_create_qp_resp *resp)
+ struct mlx5_ib_create_qp_resp *resp,
+ struct ib_qp_init_attr *init_attr)
{
struct mlx5_ib_raw_packet_qp *raw_packet_qp = &qp->raw_packet_qp;
struct mlx5_ib_sq *sq = &raw_packet_qp->sq;
struct mlx5_ib_rq *rq = &raw_packet_qp->rq;
- struct ib_uobject *uobj = pd->uobject;
- struct ib_ucontext *ucontext = uobj->context;
- struct mlx5_ib_ucontext *mucontext = to_mucontext(ucontext);
+ struct mlx5_ib_ucontext *mucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
int err;
u32 tdn = mucontext->tdn;
u16 uid = to_mpd(pd)->uid;
+ u32 out[MLX5_ST_SZ_DW(create_tir_out)] = {};
+ if (!qp->sq.wqe_cnt && !qp->rq.wqe_cnt)
+ return -EINVAL;
if (qp->sq.wqe_cnt) {
err = create_raw_packet_qp_tis(dev, qp, sq, tdn, pd);
if (err)
return err;
- err = create_raw_packet_qp_sq(dev, sq, in, pd);
+ err = create_raw_packet_qp_sq(dev, udata, sq, in, pd,
+ to_mcq(init_attr->send_cq));
if (err)
goto err_destroy_tis;
@@ -1392,15 +1606,17 @@ static int create_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
if (qp->rq.wqe_cnt) {
rq->base.container_mibqp = qp;
- if (qp->flags & MLX5_IB_QP_CVLAN_STRIPPING)
+ if (qp->flags & IB_QP_CREATE_CVLAN_STRIPPING)
rq->flags |= MLX5_IB_RQ_CVLAN_STRIPPING;
- if (qp->flags & MLX5_IB_QP_PCI_WRITE_END_PADDING)
+ if (qp->flags & IB_QP_CREATE_PCI_WRITE_END_PADDING)
rq->flags |= MLX5_IB_RQ_PCI_WRITE_END_PADDING;
- err = create_raw_packet_qp_rq(dev, rq, in, inlen, pd);
+ err = create_raw_packet_qp_rq(dev, rq, in, pd,
+ to_mcq(init_attr->recv_cq));
if (err)
goto err_destroy_sq;
- err = create_raw_packet_qp_tir(dev, rq, tdn, &qp->flags_en, pd);
+ err = create_raw_packet_qp_tir(dev, rq, tdn, &qp->flags_en, pd,
+ out);
if (err)
goto err_destroy_rq;
@@ -1409,19 +1625,28 @@ static int create_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
resp->comp_mask |= MLX5_IB_CREATE_QP_RESP_MASK_RQN;
resp->tirn = rq->tirn;
resp->comp_mask |= MLX5_IB_CREATE_QP_RESP_MASK_TIRN;
+ if (MLX5_CAP_FLOWTABLE_NIC_RX(dev->mdev, sw_owner) ||
+ MLX5_CAP_FLOWTABLE_NIC_RX(dev->mdev, sw_owner_v2)) {
+ resp->tir_icm_addr = MLX5_GET(
+ create_tir_out, out, icm_address_31_0);
+ resp->tir_icm_addr |=
+ (u64)MLX5_GET(create_tir_out, out,
+ icm_address_39_32)
+ << 32;
+ resp->tir_icm_addr |=
+ (u64)MLX5_GET(create_tir_out, out,
+ icm_address_63_40)
+ << 40;
+ resp->comp_mask |=
+ MLX5_IB_CREATE_QP_RESP_MASK_TIR_ICM_ADDR;
+ }
}
}
qp->trans_qp.base.mqp.qpn = qp->sq.wqe_cnt ? sq->base.mqp.qpn :
rq->base.mqp.qpn;
- err = ib_copy_to_udata(udata, resp, min(udata->outlen, sizeof(*resp)));
- if (err)
- goto err_destroy_tir;
-
return 0;
-err_destroy_tir:
- destroy_raw_packet_qp_tir(dev, rq, qp->flags_en, pd);
err_destroy_rq:
destroy_raw_packet_qp_rq(dev, rq);
err_destroy_sq:
@@ -1473,101 +1698,66 @@ static void destroy_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *q
to_mpd(qp->ibqp.pd)->uid);
}
-static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
- struct ib_pd *pd,
- struct ib_qp_init_attr *init_attr,
- struct ib_udata *udata)
+struct mlx5_create_qp_params {
+ struct ib_udata *udata;
+ size_t inlen;
+ size_t outlen;
+ size_t ucmd_size;
+ void *ucmd;
+ u8 is_rss_raw : 1;
+ struct ib_qp_init_attr *attr;
+ u32 uidx;
+ struct mlx5_ib_create_qp_resp resp;
+};
+
+static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
{
- struct ib_uobject *uobj = pd->uobject;
- struct ib_ucontext *ucontext = uobj->context;
- struct mlx5_ib_ucontext *mucontext = to_mucontext(ucontext);
- struct mlx5_ib_create_qp_resp resp = {};
+ struct ib_qp_init_attr *init_attr = params->attr;
+ struct mlx5_ib_create_qp_rss *ucmd = params->ucmd;
+ struct ib_udata *udata = params->udata;
+ struct mlx5_ib_ucontext *mucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
int inlen;
+ int outlen;
int err;
u32 *in;
+ u32 *out;
void *tirc;
void *hfso;
u32 selected_fields = 0;
u32 outer_l4;
- size_t min_resp_len;
u32 tdn = mucontext->tdn;
- struct mlx5_ib_create_qp_rss ucmd = {};
- size_t required_cmd_sz;
u8 lb_flag = 0;
- if (init_attr->qp_type != IB_QPT_RAW_PACKET)
- return -EOPNOTSUPP;
-
- if (init_attr->create_flags || init_attr->send_cq)
- return -EINVAL;
-
- min_resp_len = offsetof(typeof(resp), bfreg_index) + sizeof(resp.bfreg_index);
- if (udata->outlen < min_resp_len)
- return -EINVAL;
-
- required_cmd_sz = offsetof(typeof(ucmd), flags) + sizeof(ucmd.flags);
- if (udata->inlen < required_cmd_sz) {
- mlx5_ib_dbg(dev, "invalid inlen\n");
- return -EINVAL;
- }
-
- if (udata->inlen > sizeof(ucmd) &&
- !ib_is_udata_cleared(udata, sizeof(ucmd),
- udata->inlen - sizeof(ucmd))) {
- mlx5_ib_dbg(dev, "inlen is not supported\n");
- return -EOPNOTSUPP;
- }
-
- if (ib_copy_from_udata(&ucmd, udata, min(sizeof(ucmd), udata->inlen))) {
- mlx5_ib_dbg(dev, "copy failed\n");
- return -EFAULT;
- }
-
- if (ucmd.comp_mask) {
+ if (ucmd->comp_mask) {
mlx5_ib_dbg(dev, "invalid comp mask\n");
return -EOPNOTSUPP;
}
- if (ucmd.flags & ~(MLX5_QP_FLAG_TUNNEL_OFFLOADS |
- MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC |
- MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC)) {
- mlx5_ib_dbg(dev, "invalid flags\n");
- return -EOPNOTSUPP;
- }
-
- if (ucmd.flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS &&
- !tunnel_offload_supported(dev->mdev)) {
- mlx5_ib_dbg(dev, "tunnel offloads isn't supported\n");
- return -EOPNOTSUPP;
- }
-
- if (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_INNER &&
- !(ucmd.flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS)) {
+ if (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_INNER &&
+ !(ucmd->flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS)) {
mlx5_ib_dbg(dev, "Tunnel offloads must be set for inner RSS\n");
return -EOPNOTSUPP;
}
- if (ucmd.flags & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC || dev->rep) {
- lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_UNICAST;
+ if (dev->is_rep)
qp->flags_en |= MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC;
- }
- if (ucmd.flags & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC) {
- lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_MULTICAST;
- qp->flags_en |= MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC;
- }
+ if (qp->flags_en & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC)
+ lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_UNICAST;
- err = ib_copy_to_udata(udata, &resp, min(udata->outlen, sizeof(resp)));
- if (err) {
- mlx5_ib_dbg(dev, "copy failed\n");
- return -EINVAL;
- }
+ if (qp->flags_en & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC)
+ lb_flag |= MLX5_TIRC_SELF_LB_BLOCK_BLOCK_MULTICAST;
inlen = MLX5_ST_SZ_BYTES(create_tir_in);
- in = kvzalloc(inlen, GFP_KERNEL);
+ outlen = MLX5_ST_SZ_BYTES(create_tir_out);
+ in = kvzalloc(inlen + outlen, GFP_KERNEL);
if (!in)
return -ENOMEM;
+ out = in + MLX5_ST_SZ_DW(create_tir_in);
MLX5_SET(create_tir_in, in, uid, to_mpd(pd)->uid);
tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
MLX5_SET(tirc, tirc, disp_type,
@@ -1578,30 +1768,29 @@ static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
hfso = MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer);
- if (ucmd.flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS)
+ if (ucmd->flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS)
MLX5_SET(tirc, tirc, tunneled_offload_en, 1);
MLX5_SET(tirc, tirc, self_lb_block, lb_flag);
- if (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_INNER)
+ if (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_INNER)
hfso = MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_inner);
else
hfso = MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer);
- switch (ucmd.rx_hash_function) {
+ switch (ucmd->rx_hash_function) {
case MLX5_RX_HASH_FUNC_TOEPLITZ:
{
void *rss_key = MLX5_ADDR_OF(tirc, tirc, rx_hash_toeplitz_key);
size_t len = MLX5_FLD_SZ_BYTES(tirc, rx_hash_toeplitz_key);
- if (len != ucmd.rx_key_len) {
+ if (len != ucmd->rx_key_len) {
err = -EINVAL;
goto err;
}
MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FN_TOEPLITZ);
- MLX5_SET(tirc, tirc, rx_hash_symmetric, 1);
- memcpy(rss_key, ucmd.rx_hash_key, len);
+ memcpy(rss_key, ucmd->rx_hash_key, len);
break;
}
default:
@@ -1609,7 +1798,7 @@ static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
goto err;
}
- if (!ucmd.rx_hash_fields_mask) {
+ if (!ucmd->rx_hash_fields_mask) {
/* special case when this TIR serves as steering entry without hashing */
if (!init_attr->rwq_ind_tbl->log_ind_tbl_size)
goto create_tir;
@@ -1617,29 +1806,31 @@ static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
goto err;
}
- if (((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4)) &&
- ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))) {
+ if (((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4)) &&
+ ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))) {
err = -EINVAL;
goto err;
}
/* If none of IPV4 & IPV6 SRC/DST was set - this bit field is ignored */
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4))
MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
MLX5_L3_PROT_TYPE_IPV4);
- else if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))
+ else if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))
MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
MLX5_L3_PROT_TYPE_IPV6);
- outer_l4 = ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP)) << 0 |
- ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP)) << 1 |
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_IPSEC_SPI) << 2;
+ outer_l4 = ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP))
+ << 0 |
+ ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP))
+ << 1 |
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_IPSEC_SPI) << 2;
/* Check that only one l4 protocol is set */
if (outer_l4 & (outer_l4 - 1)) {
@@ -1648,39 +1839,41 @@ static int create_rss_raw_qp_tir(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
}
/* If none of TCP & UDP SRC/DST was set - this bit field is ignored */
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP))
MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
MLX5_L4_PROT_TYPE_TCP);
- else if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP))
+ else if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP))
MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
MLX5_L4_PROT_TYPE_UDP);
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV4) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_IPV6))
selected_fields |= MLX5_HASH_FIELD_SEL_SRC_IP;
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV4) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_IPV6))
selected_fields |= MLX5_HASH_FIELD_SEL_DST_IP;
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_TCP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_SRC_PORT_UDP))
selected_fields |= MLX5_HASH_FIELD_SEL_L4_SPORT;
- if ((ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP) ||
- (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP))
+ if ((ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_TCP) ||
+ (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_DST_PORT_UDP))
selected_fields |= MLX5_HASH_FIELD_SEL_L4_DPORT;
- if (ucmd.rx_hash_fields_mask & MLX5_RX_HASH_IPSEC_SPI)
+ if (ucmd->rx_hash_fields_mask & MLX5_RX_HASH_IPSEC_SPI)
selected_fields |= MLX5_HASH_FIELD_SEL_IPSEC_SPI;
MLX5_SET(rx_hash_field_select, hfso, selected_fields, selected_fields);
create_tir:
- err = mlx5_core_create_tir(dev->mdev, in, inlen, &qp->rss_qp.tirn);
+ MLX5_SET(create_tir_in, in, opcode, MLX5_CMD_OP_CREATE_TIR);
+ err = mlx5_cmd_exec_inout(dev->mdev, create_tir, in, out);
+ qp->rss_qp.tirn = MLX5_GET(create_tir_out, out, tirn);
if (!err && MLX5_GET(tirc, tirc, self_lb_block)) {
err = mlx5_ib_enable_lb(dev, false, true);
@@ -1693,60 +1886,45 @@ create_tir:
goto err;
if (mucontext->devx_uid) {
- resp.comp_mask |= MLX5_IB_CREATE_QP_RESP_MASK_TIRN;
- resp.tirn = qp->rss_qp.tirn;
+ params->resp.comp_mask |= MLX5_IB_CREATE_QP_RESP_MASK_TIRN;
+ params->resp.tirn = qp->rss_qp.tirn;
+ if (MLX5_CAP_FLOWTABLE_NIC_RX(dev->mdev, sw_owner) ||
+ MLX5_CAP_FLOWTABLE_NIC_RX(dev->mdev, sw_owner_v2)) {
+ params->resp.tir_icm_addr =
+ MLX5_GET(create_tir_out, out, icm_address_31_0);
+ params->resp.tir_icm_addr |=
+ (u64)MLX5_GET(create_tir_out, out,
+ icm_address_39_32)
+ << 32;
+ params->resp.tir_icm_addr |=
+ (u64)MLX5_GET(create_tir_out, out,
+ icm_address_63_40)
+ << 40;
+ params->resp.comp_mask |=
+ MLX5_IB_CREATE_QP_RESP_MASK_TIR_ICM_ADDR;
+ }
}
- err = ib_copy_to_udata(udata, &resp, min(udata->outlen, sizeof(resp)));
- if (err)
- goto err_copy;
-
kvfree(in);
/* qpn is reserved for that QP */
qp->trans_qp.base.mqp.qpn = 0;
- qp->flags |= MLX5_IB_QP_RSS;
+ qp->is_rss = true;
return 0;
-err_copy:
- mlx5_cmd_destroy_tir(dev->mdev, qp->rss_qp.tirn, mucontext->devx_uid);
err:
kvfree(in);
return err;
}
-static void configure_responder_scat_cqe(struct ib_qp_init_attr *init_attr,
- void *qpc)
-{
- int rcqe_sz;
-
- if (init_attr->qp_type == MLX5_IB_QPT_DCI)
- return;
-
- rcqe_sz = mlx5_ib_get_cqe_size(init_attr->recv_cq);
-
- if (rcqe_sz == 128) {
- MLX5_SET(qpc, qpc, cs_res, MLX5_RES_SCAT_DATA64_CQE);
- return;
- }
-
- if (init_attr->qp_type != MLX5_IB_QPT_DCT)
- MLX5_SET(qpc, qpc, cs_res, MLX5_RES_SCAT_DATA32_CQE);
-}
-
static void configure_requester_scat_cqe(struct mlx5_ib_dev *dev,
+ struct mlx5_ib_qp *qp,
struct ib_qp_init_attr *init_attr,
- struct mlx5_ib_create_qp *ucmd,
void *qpc)
{
- enum ib_qp_type qpt = init_attr->qp_type;
int scqe_sz;
- bool allow_scat_cqe = 0;
+ bool allow_scat_cqe = false;
- if (qpt == IB_QPT_UC || qpt == IB_QPT_UD)
- return;
-
- if (ucmd)
- allow_scat_cqe = ucmd->flags & MLX5_QP_FLAG_ALLOW_SCATTER_CQE;
+ allow_scat_cqe = qp->flags_en & MLX5_QP_FLAG_ALLOW_SCATTER_CQE;
if (!allow_scat_cqe && init_attr->sq_sig_type != IB_SIGNAL_ALL_WR)
return;
@@ -1782,7 +1960,7 @@ static int atomic_size_to_mode(int size_mask)
}
static int get_atomic_mode(struct mlx5_ib_dev *dev,
- enum ib_qp_type qp_type)
+ struct mlx5_ib_qp *qp)
{
u8 atomic_operations = MLX5_CAP_ATOMIC(dev->mdev, atomic_operations);
u8 atomic = MLX5_CAP_GEN(dev->mdev, atomic);
@@ -1792,7 +1970,7 @@ static int get_atomic_mode(struct mlx5_ib_dev *dev,
if (!atomic)
return -EOPNOTSUPP;
- if (qp_type == MLX5_IB_QPT_DCT)
+ if (qp->type == MLX5_IB_QPT_DCT)
atomic_size_mask = MLX5_CAP_ATOMIC(dev->mdev, atomic_size_dc);
else
atomic_size_mask = MLX5_CAP_ATOMIC(dev->mdev, atomic_size_qp);
@@ -1806,274 +1984,362 @@ static int get_atomic_mode(struct mlx5_ib_dev *dev,
atomic_operations & MLX5_ATOMIC_OPS_FETCH_ADD))
atomic_mode = MLX5_ATOMIC_MODE_IB_COMP;
+ /* OOO DP QPs do not support larger than 8-Bytes atomic operations */
+ if (atomic_mode > MLX5_ATOMIC_MODE_8B && qp->is_ooo_rq)
+ atomic_mode = MLX5_ATOMIC_MODE_8B;
+
return atomic_mode;
}
-static inline bool check_flags_mask(uint64_t input, uint64_t supported)
+static int create_xrc_tgt_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
{
- return (input & ~supported) == 0;
+ struct ib_qp_init_attr *attr = params->attr;
+ u32 uidx = params->uidx;
+ struct mlx5_ib_resources *devr = &dev->devr;
+ u32 out[MLX5_ST_SZ_DW(create_qp_out)] = {};
+ int inlen = MLX5_ST_SZ_BYTES(create_qp_in);
+ struct mlx5_core_dev *mdev = dev->mdev;
+ struct mlx5_ib_qp_base *base;
+ unsigned long flags;
+ void *qpc;
+ u32 *in;
+ int err;
+
+ if (attr->sq_sig_type == IB_SIGNAL_ALL_WR)
+ qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
+
+ in = kvzalloc(inlen, GFP_KERNEL);
+ if (!in)
+ return -ENOMEM;
+
+ qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+
+ MLX5_SET(qpc, qpc, st, MLX5_QP_ST_XRC);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+ MLX5_SET(qpc, qpc, pd, to_mpd(devr->p0)->pdn);
+
+ if (qp->flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK)
+ MLX5_SET(qpc, qpc, block_lb_mc, 1);
+ if (qp->flags & IB_QP_CREATE_CROSS_CHANNEL)
+ MLX5_SET(qpc, qpc, cd_master, 1);
+ if (qp->flags & IB_QP_CREATE_MANAGED_SEND)
+ MLX5_SET(qpc, qpc, cd_slave_send, 1);
+ if (qp->flags & IB_QP_CREATE_MANAGED_RECV)
+ MLX5_SET(qpc, qpc, cd_slave_receive, 1);
+
+ MLX5_SET(qpc, qpc, ts_format, mlx5_get_qp_default_ts(dev->mdev));
+ MLX5_SET(qpc, qpc, rq_type, MLX5_SRQ_RQ);
+ MLX5_SET(qpc, qpc, no_sq, 1);
+ MLX5_SET(qpc, qpc, cqn_rcv, to_mcq(devr->c0)->mcq.cqn);
+ MLX5_SET(qpc, qpc, cqn_snd, to_mcq(devr->c0)->mcq.cqn);
+ MLX5_SET(qpc, qpc, srqn_rmpn_xrqn, to_msrq(devr->s0)->msrq.srqn);
+ MLX5_SET(qpc, qpc, xrcd, to_mxrcd(attr->xrcd)->xrcdn);
+ MLX5_SET64(qpc, qpc, dbr_addr, qp->db.dma);
+
+ /* 0xffffff means we ask to work with cqe version 0 */
+ if (MLX5_CAP_GEN(mdev, cqe_version) == MLX5_CQE_VERSION_V1)
+ MLX5_SET(qpc, qpc, user_index, uidx);
+
+ if (qp->flags & IB_QP_CREATE_PCI_WRITE_END_PADDING) {
+ MLX5_SET(qpc, qpc, end_padding_mode,
+ MLX5_WQ_END_PAD_MODE_ALIGN);
+ /* Special case to clean flag */
+ qp->flags &= ~IB_QP_CREATE_PCI_WRITE_END_PADDING;
+ }
+
+ base = &qp->trans_qp.base;
+ err = mlx5_qpc_create_qp(dev, &base->mqp, in, inlen, out);
+ kvfree(in);
+ if (err)
+ return err;
+
+ base->container_mibqp = qp;
+ base->mqp.event = mlx5_ib_qp_event;
+ if (MLX5_CAP_GEN(mdev, ece_support))
+ params->resp.ece_options = MLX5_GET(create_qp_out, out, ece);
+
+ spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
+ list_add_tail(&qp->qps_list, &dev->qp_list);
+ spin_unlock_irqrestore(&dev->reset_flow_resource_lock, flags);
+
+ qp->trans_qp.xrcdn = to_mxrcd(attr->xrcd)->xrcdn;
+ return 0;
}
-static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
- struct ib_qp_init_attr *init_attr,
- struct ib_udata *udata, struct mlx5_ib_qp *qp)
+static int create_dci(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
{
+ struct ib_qp_init_attr *init_attr = params->attr;
+ struct mlx5_ib_create_qp *ucmd = params->ucmd;
+ u32 out[MLX5_ST_SZ_DW(create_qp_out)] = {};
+ struct ib_udata *udata = params->udata;
+ u32 uidx = params->uidx;
struct mlx5_ib_resources *devr = &dev->devr;
int inlen = MLX5_ST_SZ_BYTES(create_qp_in);
struct mlx5_core_dev *mdev = dev->mdev;
- struct mlx5_ib_create_qp_resp resp = {};
struct mlx5_ib_cq *send_cq;
struct mlx5_ib_cq *recv_cq;
unsigned long flags;
- u32 uidx = MLX5_IB_DEFAULT_UIDX;
- struct mlx5_ib_create_qp ucmd;
struct mlx5_ib_qp_base *base;
+ int ts_format;
int mlx5_st;
void *qpc;
u32 *in;
int err;
- mutex_init(&qp->mutex);
spin_lock_init(&qp->sq.lock);
spin_lock_init(&qp->rq.lock);
- mlx5_st = to_mlx5_st(init_attr->qp_type);
+ mlx5_st = to_mlx5_st(qp->type);
if (mlx5_st < 0)
return -EINVAL;
- if (init_attr->rwq_ind_tbl) {
- if (!udata)
- return -ENOSYS;
+ if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
+ qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
+
+ base = &qp->trans_qp.base;
- err = create_rss_raw_qp_tir(dev, qp, pd, init_attr, udata);
+ qp->has_rq = qp_has_rq(init_attr);
+ err = set_rq_size(dev, &init_attr->cap, qp->has_rq, qp, ucmd);
+ if (err) {
+ mlx5_ib_dbg(dev, "err %d\n", err);
return err;
}
- if (init_attr->create_flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK) {
- if (!MLX5_CAP_GEN(mdev, block_lb_mc)) {
- mlx5_ib_dbg(dev, "block multicast loopback isn't supported\n");
- return -EINVAL;
- } else {
- qp->flags |= MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK;
- }
+ if (ucmd->rq_wqe_shift != qp->rq.wqe_shift ||
+ ucmd->rq_wqe_count != qp->rq.wqe_cnt)
+ return -EINVAL;
+
+ if (ucmd->sq_wqe_count > (1 << MLX5_CAP_GEN(mdev, log_max_qp_sz)))
+ return -EINVAL;
+
+ ts_format = get_qp_ts_format(dev, to_mcq(init_attr->send_cq),
+ to_mcq(init_attr->recv_cq));
+
+ if (ts_format < 0)
+ return ts_format;
+
+ err = _create_user_qp(dev, pd, qp, udata, init_attr, &in, &params->resp,
+ &inlen, base, ucmd);
+ if (err)
+ return err;
+
+ if (MLX5_CAP_GEN(mdev, ece_support))
+ MLX5_SET(create_qp_in, in, ece, ucmd->ece_options);
+ qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+
+ MLX5_SET(qpc, qpc, st, mlx5_st);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+ MLX5_SET(qpc, qpc, pd, to_mpd(pd)->pdn);
+
+ if (qp->flags_en & MLX5_QP_FLAG_SIGNATURE)
+ MLX5_SET(qpc, qpc, wq_signature, 1);
+
+ if (qp->flags & IB_QP_CREATE_CROSS_CHANNEL)
+ MLX5_SET(qpc, qpc, cd_master, 1);
+ if (qp->flags & IB_QP_CREATE_MANAGED_SEND)
+ MLX5_SET(qpc, qpc, cd_slave_send, 1);
+ if (qp->flags_en & MLX5_QP_FLAG_SCATTER_CQE)
+ configure_requester_scat_cqe(dev, qp, init_attr, qpc);
+
+ if (qp->rq.wqe_cnt) {
+ MLX5_SET(qpc, qpc, log_rq_stride, qp->rq.wqe_shift - 4);
+ MLX5_SET(qpc, qpc, log_rq_size, ilog2(qp->rq.wqe_cnt));
}
- if (init_attr->create_flags &
- (IB_QP_CREATE_CROSS_CHANNEL |
- IB_QP_CREATE_MANAGED_SEND |
- IB_QP_CREATE_MANAGED_RECV)) {
- if (!MLX5_CAP_GEN(mdev, cd)) {
- mlx5_ib_dbg(dev, "cross-channel isn't supported\n");
- return -EINVAL;
- }
- if (init_attr->create_flags & IB_QP_CREATE_CROSS_CHANNEL)
- qp->flags |= MLX5_IB_QP_CROSS_CHANNEL;
- if (init_attr->create_flags & IB_QP_CREATE_MANAGED_SEND)
- qp->flags |= MLX5_IB_QP_MANAGED_SEND;
- if (init_attr->create_flags & IB_QP_CREATE_MANAGED_RECV)
- qp->flags |= MLX5_IB_QP_MANAGED_RECV;
- }
-
- if (init_attr->qp_type == IB_QPT_UD &&
- (init_attr->create_flags & IB_QP_CREATE_IPOIB_UD_LSO))
- if (!MLX5_CAP_GEN(mdev, ipoib_basic_offloads)) {
- mlx5_ib_dbg(dev, "ipoib UD lso qp isn't supported\n");
- return -EOPNOTSUPP;
- }
+ if (qp->flags_en & MLX5_QP_FLAG_DCI_STREAM) {
+ MLX5_SET(qpc, qpc, log_num_dci_stream_channels,
+ ucmd->dci_streams.log_num_concurent);
+ MLX5_SET(qpc, qpc, log_num_dci_errored_streams,
+ ucmd->dci_streams.log_num_errored);
+ }
- if (init_attr->create_flags & IB_QP_CREATE_SCATTER_FCS) {
- if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
- mlx5_ib_dbg(dev, "Scatter FCS is supported only for Raw Packet QPs");
- return -EOPNOTSUPP;
- }
- if (!MLX5_CAP_GEN(dev->mdev, eth_net_offloads) ||
- !MLX5_CAP_ETH(dev->mdev, scatter_fcs)) {
- mlx5_ib_dbg(dev, "Scatter FCS isn't supported\n");
- return -EOPNOTSUPP;
- }
- qp->flags |= MLX5_IB_QP_CAP_SCATTER_FCS;
+ MLX5_SET(qpc, qpc, ts_format, ts_format);
+ MLX5_SET(qpc, qpc, rq_type, get_rx_type(qp, init_attr));
+
+ MLX5_SET(qpc, qpc, log_sq_size, ilog2(qp->sq.wqe_cnt));
+
+ /* Set default resources */
+ if (init_attr->srq) {
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn0);
+ MLX5_SET(qpc, qpc, srqn_rmpn_xrqn,
+ to_msrq(init_attr->srq)->msrq.srqn);
+ } else {
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn1);
+ MLX5_SET(qpc, qpc, srqn_rmpn_xrqn,
+ to_msrq(devr->s1)->msrq.srqn);
}
- if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
- qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
+ if (init_attr->send_cq)
+ MLX5_SET(qpc, qpc, cqn_snd,
+ to_mcq(init_attr->send_cq)->mcq.cqn);
- if (init_attr->create_flags & IB_QP_CREATE_CVLAN_STRIPPING) {
- if (!(MLX5_CAP_GEN(dev->mdev, eth_net_offloads) &&
- MLX5_CAP_ETH(dev->mdev, vlan_cap)) ||
- (init_attr->qp_type != IB_QPT_RAW_PACKET))
- return -EOPNOTSUPP;
- qp->flags |= MLX5_IB_QP_CVLAN_STRIPPING;
+ if (init_attr->recv_cq)
+ MLX5_SET(qpc, qpc, cqn_rcv,
+ to_mcq(init_attr->recv_cq)->mcq.cqn);
+
+ MLX5_SET64(qpc, qpc, dbr_addr, qp->db.dma);
+
+ /* 0xffffff means we ask to work with cqe version 0 */
+ if (MLX5_CAP_GEN(mdev, cqe_version) == MLX5_CQE_VERSION_V1)
+ MLX5_SET(qpc, qpc, user_index, uidx);
+
+ if (qp->flags & IB_QP_CREATE_PCI_WRITE_END_PADDING) {
+ MLX5_SET(qpc, qpc, end_padding_mode,
+ MLX5_WQ_END_PAD_MODE_ALIGN);
+ /* Special case to clean flag */
+ qp->flags &= ~IB_QP_CREATE_PCI_WRITE_END_PADDING;
}
- if (udata) {
- if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
- mlx5_ib_dbg(dev, "copy failed\n");
- return -EFAULT;
- }
+ err = mlx5_qpc_create_qp(dev, &base->mqp, in, inlen, out);
- if (!check_flags_mask(ucmd.flags,
- MLX5_QP_FLAG_SIGNATURE |
- MLX5_QP_FLAG_SCATTER_CQE |
- MLX5_QP_FLAG_TUNNEL_OFFLOADS |
- MLX5_QP_FLAG_BFREG_INDEX |
- MLX5_QP_FLAG_TYPE_DCT |
- MLX5_QP_FLAG_TYPE_DCI |
- MLX5_QP_FLAG_ALLOW_SCATTER_CQE |
- MLX5_QP_FLAG_PACKET_BASED_CREDIT_MODE))
- return -EINVAL;
+ kvfree(in);
+ if (err)
+ goto err_create;
- err = get_qp_user_index(to_mucontext(pd->uobject->context),
- &ucmd, udata->inlen, &uidx);
- if (err)
- return err;
+ base->container_mibqp = qp;
+ base->mqp.event = mlx5_ib_qp_event;
+ if (MLX5_CAP_GEN(mdev, ece_support))
+ params->resp.ece_options = MLX5_GET(create_qp_out, out, ece);
- qp->wq_sig = !!(ucmd.flags & MLX5_QP_FLAG_SIGNATURE);
- if (MLX5_CAP_GEN(dev->mdev, sctr_data_cqe))
- qp->scat_cqe = !!(ucmd.flags & MLX5_QP_FLAG_SCATTER_CQE);
- if (ucmd.flags & MLX5_QP_FLAG_TUNNEL_OFFLOADS) {
- if (init_attr->qp_type != IB_QPT_RAW_PACKET ||
- !tunnel_offload_supported(mdev)) {
- mlx5_ib_dbg(dev, "Tunnel offload isn't supported\n");
- return -EOPNOTSUPP;
- }
- qp->flags_en |= MLX5_QP_FLAG_TUNNEL_OFFLOADS;
- }
+ get_cqs(qp->type, init_attr->send_cq, init_attr->recv_cq,
+ &send_cq, &recv_cq);
+ spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
+ mlx5_ib_lock_cqs(send_cq, recv_cq);
+ /* Maintain device to QPs access, needed for further handling via reset
+ * flow
+ */
+ list_add_tail(&qp->qps_list, &dev->qp_list);
+ /* Maintain CQ to QPs access, needed for further handling via reset flow
+ */
+ if (send_cq)
+ list_add_tail(&qp->cq_send_list, &send_cq->list_send_qp);
+ if (recv_cq)
+ list_add_tail(&qp->cq_recv_list, &recv_cq->list_recv_qp);
+ mlx5_ib_unlock_cqs(send_cq, recv_cq);
+ spin_unlock_irqrestore(&dev->reset_flow_resource_lock, flags);
- if (ucmd.flags & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC) {
- if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
- mlx5_ib_dbg(dev, "Self-LB UC isn't supported\n");
- return -EOPNOTSUPP;
- }
- qp->flags_en |= MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC;
- }
+ return 0;
- if (ucmd.flags & MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC) {
- if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
- mlx5_ib_dbg(dev, "Self-LB UM isn't supported\n");
- return -EOPNOTSUPP;
- }
- qp->flags_en |= MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC;
- }
+err_create:
+ destroy_qp(dev, qp, base, udata);
+ return err;
+}
- if (ucmd.flags & MLX5_QP_FLAG_PACKET_BASED_CREDIT_MODE) {
- if (init_attr->qp_type != IB_QPT_RC ||
- !MLX5_CAP_GEN(dev->mdev, qp_packet_based)) {
- mlx5_ib_dbg(dev, "packet based credit mode isn't supported\n");
- return -EOPNOTSUPP;
- }
- qp->flags |= MLX5_IB_QP_PACKET_BASED_CREDIT;
- }
+static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
+{
+ struct ib_qp_init_attr *init_attr = params->attr;
+ struct mlx5_ib_create_qp *ucmd = params->ucmd;
+ u32 out[MLX5_ST_SZ_DW(create_qp_out)] = {};
+ struct ib_udata *udata = params->udata;
+ u32 uidx = params->uidx;
+ struct mlx5_ib_resources *devr = &dev->devr;
+ int inlen = MLX5_ST_SZ_BYTES(create_qp_in);
+ struct mlx5_core_dev *mdev = dev->mdev;
+ struct mlx5_ib_cq *send_cq;
+ struct mlx5_ib_cq *recv_cq;
+ unsigned long flags;
+ struct mlx5_ib_qp_base *base;
+ int ts_format;
+ int mlx5_st;
+ void *qpc;
+ u32 *in;
+ int err;
- if (init_attr->create_flags & IB_QP_CREATE_SOURCE_QPN) {
- if (init_attr->qp_type != IB_QPT_UD ||
- (MLX5_CAP_GEN(dev->mdev, port_type) !=
- MLX5_CAP_PORT_TYPE_IB) ||
- !mlx5_get_flow_namespace(dev->mdev, MLX5_FLOW_NAMESPACE_BYPASS)) {
- mlx5_ib_dbg(dev, "Source QP option isn't supported\n");
- return -EOPNOTSUPP;
- }
+ spin_lock_init(&qp->sq.lock);
+ spin_lock_init(&qp->rq.lock);
- qp->flags |= MLX5_IB_QP_UNDERLAY;
- qp->underlay_qpn = init_attr->source_qpn;
- }
- } else {
- qp->wq_sig = !!wq_signature;
- }
+ mlx5_st = to_mlx5_st(qp->type);
+ if (mlx5_st < 0)
+ return -EINVAL;
+
+ if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
+ qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
+
+ if (qp->flags & IB_QP_CREATE_SOURCE_QPN)
+ qp->underlay_qpn = init_attr->source_qpn;
base = (init_attr->qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) ?
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) ?
&qp->raw_packet_qp.rq.base :
&qp->trans_qp.base;
qp->has_rq = qp_has_rq(init_attr);
- err = set_rq_size(dev, &init_attr->cap, qp->has_rq,
- qp, udata ? &ucmd : NULL);
+ err = set_rq_size(dev, &init_attr->cap, qp->has_rq, qp, ucmd);
if (err) {
mlx5_ib_dbg(dev, "err %d\n", err);
return err;
}
- if (pd) {
- if (udata) {
- __u32 max_wqes =
- 1 << MLX5_CAP_GEN(mdev, log_max_qp_sz);
- mlx5_ib_dbg(dev, "requested sq_wqe_count (%d)\n", ucmd.sq_wqe_count);
- if (ucmd.rq_wqe_shift != qp->rq.wqe_shift ||
- ucmd.rq_wqe_count != qp->rq.wqe_cnt) {
- mlx5_ib_dbg(dev, "invalid rq params\n");
- return -EINVAL;
- }
- if (ucmd.sq_wqe_count > max_wqes) {
- mlx5_ib_dbg(dev, "requested sq_wqe_count (%d) > max allowed (%d)\n",
- ucmd.sq_wqe_count, max_wqes);
- return -EINVAL;
- }
- if (init_attr->create_flags &
- mlx5_ib_create_qp_sqpn_qp1()) {
- mlx5_ib_dbg(dev, "user-space is not allowed to create UD QPs spoofing as QP1\n");
- return -EINVAL;
- }
- err = create_user_qp(dev, pd, qp, udata, init_attr, &in,
- &resp, &inlen, base);
- if (err)
- mlx5_ib_dbg(dev, "err %d\n", err);
- } else {
- err = create_kernel_qp(dev, init_attr, qp, &in, &inlen,
- base);
- if (err)
- mlx5_ib_dbg(dev, "err %d\n", err);
- }
+ if (ucmd->rq_wqe_shift != qp->rq.wqe_shift ||
+ ucmd->rq_wqe_count != qp->rq.wqe_cnt)
+ return -EINVAL;
- if (err)
- return err;
- } else {
- in = kvzalloc(inlen, GFP_KERNEL);
- if (!in)
- return -ENOMEM;
+ if (ucmd->sq_wqe_count > (1 << MLX5_CAP_GEN(mdev, log_max_qp_sz)))
+ return -EINVAL;
- qp->create_type = MLX5_QP_EMPTY;
+ if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
+ ts_format = get_qp_ts_format(dev, to_mcq(init_attr->send_cq),
+ to_mcq(init_attr->recv_cq));
+ if (ts_format < 0)
+ return ts_format;
}
+ err = _create_user_qp(dev, pd, qp, udata, init_attr, &in, &params->resp,
+ &inlen, base, ucmd);
+ if (err)
+ return err;
+
if (is_sqp(init_attr->qp_type))
qp->port = init_attr->port_num;
+ if (MLX5_CAP_GEN(mdev, ece_support))
+ MLX5_SET(create_qp_in, in, ece, ucmd->ece_options);
qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
MLX5_SET(qpc, qpc, st, mlx5_st);
MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+ MLX5_SET(qpc, qpc, pd, to_mpd(pd)->pdn);
- if (init_attr->qp_type != MLX5_IB_QPT_REG_UMR)
- MLX5_SET(qpc, qpc, pd, to_mpd(pd ? pd : devr->p0)->pdn);
- else
- MLX5_SET(qpc, qpc, latency_sensitive, 1);
-
-
- if (qp->wq_sig)
+ if (qp->flags_en & MLX5_QP_FLAG_SIGNATURE)
MLX5_SET(qpc, qpc, wq_signature, 1);
- if (qp->flags & MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK)
+ if (qp->flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK)
MLX5_SET(qpc, qpc, block_lb_mc, 1);
- if (qp->flags & MLX5_IB_QP_CROSS_CHANNEL)
+ if (qp->flags & IB_QP_CREATE_CROSS_CHANNEL)
MLX5_SET(qpc, qpc, cd_master, 1);
- if (qp->flags & MLX5_IB_QP_MANAGED_SEND)
+ if (qp->flags & IB_QP_CREATE_MANAGED_SEND)
MLX5_SET(qpc, qpc, cd_slave_send, 1);
- if (qp->flags & MLX5_IB_QP_MANAGED_RECV)
+ if (qp->flags & IB_QP_CREATE_MANAGED_RECV)
MLX5_SET(qpc, qpc, cd_slave_receive, 1);
- if (qp->flags & MLX5_IB_QP_PACKET_BASED_CREDIT)
+ if (qp->flags_en & MLX5_QP_FLAG_PACKET_BASED_CREDIT_MODE)
MLX5_SET(qpc, qpc, req_e2e_credit_mode, 1);
- if (qp->scat_cqe && is_connected(init_attr->qp_type)) {
- configure_responder_scat_cqe(init_attr, qpc);
- configure_requester_scat_cqe(dev, init_attr,
- udata ? &ucmd : NULL,
- qpc);
+ if ((qp->flags_en & MLX5_QP_FLAG_SCATTER_CQE) &&
+ (init_attr->qp_type == IB_QPT_RC ||
+ init_attr->qp_type == IB_QPT_UC)) {
+ int rcqe_sz = mlx5_ib_get_cqe_size(init_attr->recv_cq);
+
+ MLX5_SET(qpc, qpc, cs_res,
+ rcqe_sz == 128 ? MLX5_RES_SCAT_DATA64_CQE :
+ MLX5_RES_SCAT_DATA32_CQE);
}
+ if ((qp->flags_en & MLX5_QP_FLAG_SCATTER_CQE) &&
+ (qp->type == MLX5_IB_QPT_DCI || qp->type == IB_QPT_RC))
+ configure_requester_scat_cqe(dev, qp, init_attr, qpc);
if (qp->rq.wqe_cnt) {
MLX5_SET(qpc, qpc, log_rq_stride, qp->rq.wqe_shift - 4);
MLX5_SET(qpc, qpc, log_rq_size, ilog2(qp->rq.wqe_cnt));
}
+ if (init_attr->qp_type != IB_QPT_RAW_PACKET)
+ MLX5_SET(qpc, qpc, ts_format, ts_format);
+
MLX5_SET(qpc, qpc, rq_type, get_rx_type(qp, init_attr));
if (qp->sq.wqe_cnt) {
@@ -2088,23 +2354,17 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
/* Set default resources */
switch (init_attr->qp_type) {
- case IB_QPT_XRC_TGT:
- MLX5_SET(qpc, qpc, cqn_rcv, to_mcq(devr->c0)->mcq.cqn);
- MLX5_SET(qpc, qpc, cqn_snd, to_mcq(devr->c0)->mcq.cqn);
- MLX5_SET(qpc, qpc, srqn_rmpn_xrqn, to_msrq(devr->s0)->msrq.srqn);
- MLX5_SET(qpc, qpc, xrcd, to_mxrcd(init_attr->xrcd)->xrcdn);
- break;
case IB_QPT_XRC_INI:
MLX5_SET(qpc, qpc, cqn_rcv, to_mcq(devr->c0)->mcq.cqn);
- MLX5_SET(qpc, qpc, xrcd, to_mxrcd(devr->x1)->xrcdn);
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn1);
MLX5_SET(qpc, qpc, srqn_rmpn_xrqn, to_msrq(devr->s0)->msrq.srqn);
break;
default:
if (init_attr->srq) {
- MLX5_SET(qpc, qpc, xrcd, to_mxrcd(devr->x0)->xrcdn);
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn0);
MLX5_SET(qpc, qpc, srqn_rmpn_xrqn, to_msrq(init_attr->srq)->msrq.srqn);
} else {
- MLX5_SET(qpc, qpc, xrcd, to_mxrcd(devr->x1)->xrcdn);
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn1);
MLX5_SET(qpc, qpc, srqn_rmpn_xrqn, to_msrq(devr->s1)->msrq.srqn);
}
}
@@ -2121,52 +2381,166 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
if (MLX5_CAP_GEN(mdev, cqe_version) == MLX5_CQE_VERSION_V1)
MLX5_SET(qpc, qpc, user_index, uidx);
- /* we use IB_QP_CREATE_IPOIB_UD_LSO to indicates ipoib qp */
- if (init_attr->qp_type == IB_QPT_UD &&
- (init_attr->create_flags & IB_QP_CREATE_IPOIB_UD_LSO)) {
- MLX5_SET(qpc, qpc, ulp_stateless_offload_mode, 1);
- qp->flags |= MLX5_IB_QP_LSO;
+ if (qp->flags & IB_QP_CREATE_PCI_WRITE_END_PADDING &&
+ init_attr->qp_type != IB_QPT_RAW_PACKET) {
+ MLX5_SET(qpc, qpc, end_padding_mode,
+ MLX5_WQ_END_PAD_MODE_ALIGN);
+ /* Special case to clean flag */
+ qp->flags &= ~IB_QP_CREATE_PCI_WRITE_END_PADDING;
}
- if (init_attr->create_flags & IB_QP_CREATE_PCI_WRITE_END_PADDING) {
- if (!MLX5_CAP_GEN(dev->mdev, end_pad)) {
- mlx5_ib_dbg(dev, "scatter end padding is not supported\n");
- err = -EOPNOTSUPP;
- goto err;
- } else if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
- MLX5_SET(qpc, qpc, end_padding_mode,
- MLX5_WQ_END_PAD_MODE_ALIGN);
- } else {
- qp->flags |= MLX5_IB_QP_PCI_WRITE_END_PADDING;
- }
+ if (init_attr->qp_type == IB_QPT_RAW_PACKET ||
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) {
+ qp->raw_packet_qp.sq.ubuffer.buf_addr = ucmd->sq_buf_addr;
+ raw_packet_qp_copy_info(qp, &qp->raw_packet_qp);
+ err = create_raw_packet_qp(dev, qp, in, inlen, pd, udata,
+ &params->resp, init_attr);
+ } else
+ err = mlx5_qpc_create_qp(dev, &base->mqp, in, inlen, out);
+
+ kvfree(in);
+ if (err)
+ goto err_create;
+
+ base->container_mibqp = qp;
+ base->mqp.event = mlx5_ib_qp_event;
+ if (MLX5_CAP_GEN(mdev, ece_support))
+ params->resp.ece_options = MLX5_GET(create_qp_out, out, ece);
+
+ get_cqs(qp->type, init_attr->send_cq, init_attr->recv_cq,
+ &send_cq, &recv_cq);
+ spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
+ mlx5_ib_lock_cqs(send_cq, recv_cq);
+ /* Maintain device to QPs access, needed for further handling via reset
+ * flow
+ */
+ list_add_tail(&qp->qps_list, &dev->qp_list);
+ /* Maintain CQ to QPs access, needed for further handling via reset flow
+ */
+ if (send_cq)
+ list_add_tail(&qp->cq_send_list, &send_cq->list_send_qp);
+ if (recv_cq)
+ list_add_tail(&qp->cq_recv_list, &recv_cq->list_recv_qp);
+ mlx5_ib_unlock_cqs(send_cq, recv_cq);
+ spin_unlock_irqrestore(&dev->reset_flow_resource_lock, flags);
+
+ return 0;
+
+err_create:
+ destroy_qp(dev, qp, base, udata);
+ return err;
+}
+
+static int create_kernel_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
+{
+ struct ib_qp_init_attr *attr = params->attr;
+ u32 uidx = params->uidx;
+ struct mlx5_ib_resources *devr = &dev->devr;
+ u32 out[MLX5_ST_SZ_DW(create_qp_out)] = {};
+ int inlen = MLX5_ST_SZ_BYTES(create_qp_in);
+ struct mlx5_core_dev *mdev = dev->mdev;
+ struct mlx5_ib_cq *send_cq;
+ struct mlx5_ib_cq *recv_cq;
+ unsigned long flags;
+ struct mlx5_ib_qp_base *base;
+ int mlx5_st;
+ void *qpc;
+ u32 *in;
+ int err;
+
+ spin_lock_init(&qp->sq.lock);
+ spin_lock_init(&qp->rq.lock);
+
+ mlx5_st = to_mlx5_st(qp->type);
+ if (mlx5_st < 0)
+ return -EINVAL;
+
+ if (attr->sq_sig_type == IB_SIGNAL_ALL_WR)
+ qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
+
+ base = &qp->trans_qp.base;
+
+ qp->has_rq = qp_has_rq(attr);
+ err = set_rq_size(dev, &attr->cap, qp->has_rq, qp, NULL);
+ if (err) {
+ mlx5_ib_dbg(dev, "err %d\n", err);
+ return err;
}
- if (inlen < 0) {
- err = -EINVAL;
- goto err;
+ err = _create_kernel_qp(dev, attr, qp, &in, &inlen, base);
+ if (err)
+ return err;
+
+ if (is_sqp(attr->qp_type))
+ qp->port = attr->port_num;
+
+ qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+
+ MLX5_SET(qpc, qpc, st, mlx5_st);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+
+ if (attr->qp_type != MLX5_IB_QPT_REG_UMR)
+ MLX5_SET(qpc, qpc, pd, to_mpd(pd ? pd : devr->p0)->pdn);
+ else
+ MLX5_SET(qpc, qpc, latency_sensitive, 1);
+
+
+ if (qp->flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK)
+ MLX5_SET(qpc, qpc, block_lb_mc, 1);
+
+ if (qp->rq.wqe_cnt) {
+ MLX5_SET(qpc, qpc, log_rq_stride, qp->rq.wqe_shift - 4);
+ MLX5_SET(qpc, qpc, log_rq_size, ilog2(qp->rq.wqe_cnt));
}
- if (init_attr->qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) {
- qp->raw_packet_qp.sq.ubuffer.buf_addr = ucmd.sq_buf_addr;
- raw_packet_qp_copy_info(qp, &qp->raw_packet_qp);
- err = create_raw_packet_qp(dev, qp, in, inlen, pd, udata,
- &resp);
+ MLX5_SET(qpc, qpc, rq_type, get_rx_type(qp, attr));
+
+ if (qp->sq.wqe_cnt)
+ MLX5_SET(qpc, qpc, log_sq_size, ilog2(qp->sq.wqe_cnt));
+ else
+ MLX5_SET(qpc, qpc, no_sq, 1);
+
+ if (attr->srq) {
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn0);
+ MLX5_SET(qpc, qpc, srqn_rmpn_xrqn,
+ to_msrq(attr->srq)->msrq.srqn);
} else {
- err = mlx5_core_create_qp(dev->mdev, &base->mqp, in, inlen);
+ MLX5_SET(qpc, qpc, xrcd, devr->xrcdn1);
+ MLX5_SET(qpc, qpc, srqn_rmpn_xrqn,
+ to_msrq(devr->s1)->msrq.srqn);
}
- if (err) {
- mlx5_ib_dbg(dev, "create qp failed\n");
- goto err_create;
- }
+ if (attr->send_cq)
+ MLX5_SET(qpc, qpc, cqn_snd, to_mcq(attr->send_cq)->mcq.cqn);
+
+ if (attr->recv_cq)
+ MLX5_SET(qpc, qpc, cqn_rcv, to_mcq(attr->recv_cq)->mcq.cqn);
+
+ MLX5_SET64(qpc, qpc, dbr_addr, qp->db.dma);
+
+ /* 0xffffff means we ask to work with cqe version 0 */
+ if (MLX5_CAP_GEN(mdev, cqe_version) == MLX5_CQE_VERSION_V1)
+ MLX5_SET(qpc, qpc, user_index, uidx);
+
+ /* we use IB_QP_CREATE_IPOIB_UD_LSO to indicates ipoib qp */
+ if (qp->flags & IB_QP_CREATE_IPOIB_UD_LSO)
+ MLX5_SET(qpc, qpc, ulp_stateless_offload_mode, 1);
+
+ if (qp->flags & IB_QP_CREATE_INTEGRITY_EN &&
+ MLX5_CAP_GEN(mdev, go_back_n))
+ MLX5_SET(qpc, qpc, retry_mode, MLX5_QP_RM_GO_BACK_N);
+ err = mlx5_qpc_create_qp(dev, &base->mqp, in, inlen, out);
kvfree(in);
+ if (err)
+ goto err_create;
base->container_mibqp = qp;
base->mqp.event = mlx5_ib_qp_event;
- get_cqs(init_attr->qp_type, init_attr->send_cq, init_attr->recv_cq,
+ get_cqs(qp->type, attr->send_cq, attr->recv_cq,
&send_cq, &recv_cq);
spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
mlx5_ib_lock_cqs(send_cq, recv_cq);
@@ -2186,13 +2560,7 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
return 0;
err_create:
- if (qp->create_type == MLX5_QP_USER)
- destroy_qp_user(dev, pd, qp, base);
- else if (qp->create_type == MLX5_QP_KERNEL)
- destroy_qp_kernel(dev, qp);
-
-err:
- kvfree(in);
+ destroy_qp(dev, qp, base, NULL);
return err;
}
@@ -2254,11 +2622,6 @@ static void mlx5_ib_unlock_cqs(struct mlx5_ib_cq *send_cq, struct mlx5_ib_cq *re
}
}
-static struct mlx5_ib_pd *get_pd(struct mlx5_ib_qp *qp)
-{
- return to_mpd(qp->ibqp.pd);
-}
-
static void get_cqs(enum ib_qp_type qp_type,
struct ib_cq *ib_send_cq, struct ib_cq *ib_recv_cq,
struct mlx5_ib_cq **send_cq, struct mlx5_ib_cq **recv_cq)
@@ -2279,14 +2642,10 @@ static void get_cqs(enum ib_qp_type qp_type,
case IB_QPT_RC:
case IB_QPT_UC:
case IB_QPT_UD:
- case IB_QPT_RAW_IPV6:
- case IB_QPT_RAW_ETHERTYPE:
case IB_QPT_RAW_PACKET:
*send_cq = ib_send_cq ? to_mcq(ib_send_cq) : NULL;
*recv_cq = ib_recv_cq ? to_mcq(ib_recv_cq) : NULL;
break;
-
- case IB_QPT_MAX:
default:
*send_cq = NULL;
*recv_cq = NULL;
@@ -2298,29 +2657,29 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
const struct mlx5_modify_raw_qp_param *raw_qp_param,
u8 lag_tx_affinity);
-static void destroy_qp_common(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp)
+static void destroy_qp_common(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ struct ib_udata *udata)
{
struct mlx5_ib_cq *send_cq, *recv_cq;
struct mlx5_ib_qp_base *base;
unsigned long flags;
int err;
- if (qp->ibqp.rwq_ind_tbl) {
+ if (qp->is_rss) {
destroy_rss_raw_qp_tir(dev, qp);
return;
}
- base = (qp->ibqp.qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) ?
- &qp->raw_packet_qp.rq.base :
- &qp->trans_qp.base;
+ base = (qp->type == IB_QPT_RAW_PACKET ||
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) ?
+ &qp->raw_packet_qp.rq.base :
+ &qp->trans_qp.base;
if (qp->state != IB_QPS_RESET) {
- if (qp->ibqp.qp_type != IB_QPT_RAW_PACKET &&
- !(qp->flags & MLX5_IB_QP_UNDERLAY)) {
- err = mlx5_core_qp_modify(dev->mdev,
- MLX5_CMD_OP_2RST_QP, 0,
- NULL, &base->mqp);
+ if (qp->type != IB_QPT_RAW_PACKET &&
+ !(qp->flags & IB_QP_CREATE_SOURCE_QPN)) {
+ err = mlx5_core_qp_modify(dev, MLX5_CMD_OP_2RST_QP, 0,
+ NULL, &base->mqp, NULL);
} else {
struct mlx5_modify_raw_qp_param raw_qp_param = {
.operation = MLX5_CMD_OP_2RST_QP
@@ -2333,8 +2692,8 @@ static void destroy_qp_common(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp)
base->mqp.qpn);
}
- get_cqs(qp->ibqp.qp_type, qp->ibqp.send_cq, qp->ibqp.recv_cq,
- &send_cq, &recv_cq);
+ get_cqs(qp->type, qp->ibqp.send_cq, qp->ibqp.recv_cq, &send_cq,
+ &recv_cq);
spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
mlx5_ib_lock_cqs(send_cq, recv_cq);
@@ -2346,7 +2705,7 @@ static void destroy_qp_common(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp)
if (recv_cq)
list_del(&qp->cq_recv_list);
- if (qp->create_type == MLX5_QP_KERNEL) {
+ if (!udata) {
__mlx5_ib_cq_clean(recv_cq, base->mqp.qpn,
qp->ibqp.srq ? to_msrq(qp->ibqp.srq) : NULL);
if (send_cq != recv_cq)
@@ -2356,260 +2715,488 @@ static void destroy_qp_common(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp)
mlx5_ib_unlock_cqs(send_cq, recv_cq);
spin_unlock_irqrestore(&dev->reset_flow_resource_lock, flags);
- if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) {
+ if (qp->type == IB_QPT_RAW_PACKET ||
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) {
destroy_raw_packet_qp(dev, qp);
} else {
- err = mlx5_core_destroy_qp(dev->mdev, &base->mqp);
+ err = mlx5_core_destroy_qp(dev, &base->mqp);
if (err)
mlx5_ib_warn(dev, "failed to destroy QP 0x%x\n",
base->mqp.qpn);
}
- if (qp->create_type == MLX5_QP_KERNEL)
- destroy_qp_kernel(dev, qp);
- else if (qp->create_type == MLX5_QP_USER)
- destroy_qp_user(dev, &get_pd(qp)->ibpd, qp, base);
+ destroy_qp(dev, qp, base, udata);
}
-static const char *ib_qp_type_str(enum ib_qp_type type)
+static int create_dct(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
{
- switch (type) {
- case IB_QPT_SMI:
- return "IB_QPT_SMI";
- case IB_QPT_GSI:
- return "IB_QPT_GSI";
+ struct ib_qp_init_attr *attr = params->attr;
+ struct mlx5_ib_create_qp *ucmd = params->ucmd;
+ u32 uidx = params->uidx;
+ void *dctc;
+
+ if (mlx5_lag_is_active(dev->mdev) && !MLX5_CAP_GEN(dev->mdev, lag_dct))
+ return -EOPNOTSUPP;
+
+ qp->dct.in = kzalloc(MLX5_ST_SZ_BYTES(create_dct_in), GFP_KERNEL);
+ if (!qp->dct.in)
+ return -ENOMEM;
+
+ MLX5_SET(create_dct_in, qp->dct.in, uid, to_mpd(pd)->uid);
+ dctc = MLX5_ADDR_OF(create_dct_in, qp->dct.in, dct_context_entry);
+ MLX5_SET(dctc, dctc, pd, to_mpd(pd)->pdn);
+ MLX5_SET(dctc, dctc, srqn_xrqn, to_msrq(attr->srq)->msrq.srqn);
+ MLX5_SET(dctc, dctc, cqn, to_mcq(attr->recv_cq)->mcq.cqn);
+ MLX5_SET64(dctc, dctc, dc_access_key, ucmd->access_key);
+ MLX5_SET(dctc, dctc, user_index, uidx);
+ if (MLX5_CAP_GEN(dev->mdev, ece_support))
+ MLX5_SET(dctc, dctc, ece, ucmd->ece_options);
+
+ if (qp->flags_en & MLX5_QP_FLAG_SCATTER_CQE) {
+ int rcqe_sz = mlx5_ib_get_cqe_size(attr->recv_cq);
+
+ if (rcqe_sz == 128)
+ MLX5_SET(dctc, dctc, cs_res, MLX5_RES_SCAT_DATA64_CQE);
+ }
+
+ qp->state = IB_QPS_RESET;
+ return 0;
+}
+
+static int check_qp_type(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
+ enum ib_qp_type *type)
+{
+ if (attr->qp_type == IB_QPT_DRIVER && !MLX5_CAP_GEN(dev->mdev, dct))
+ goto out;
+
+ switch (attr->qp_type) {
+ case IB_QPT_XRC_TGT:
+ case IB_QPT_XRC_INI:
+ if (!MLX5_CAP_GEN(dev->mdev, xrc))
+ goto out;
+ fallthrough;
case IB_QPT_RC:
- return "IB_QPT_RC";
case IB_QPT_UC:
- return "IB_QPT_UC";
- case IB_QPT_UD:
- return "IB_QPT_UD";
- case IB_QPT_RAW_IPV6:
- return "IB_QPT_RAW_IPV6";
- case IB_QPT_RAW_ETHERTYPE:
- return "IB_QPT_RAW_ETHERTYPE";
- case IB_QPT_XRC_INI:
- return "IB_QPT_XRC_INI";
- case IB_QPT_XRC_TGT:
- return "IB_QPT_XRC_TGT";
+ case IB_QPT_SMI:
+ case MLX5_IB_QPT_HW_GSI:
+ case IB_QPT_DRIVER:
+ case IB_QPT_GSI:
case IB_QPT_RAW_PACKET:
- return "IB_QPT_RAW_PACKET";
+ case IB_QPT_UD:
case MLX5_IB_QPT_REG_UMR:
- return "MLX5_IB_QPT_REG_UMR";
- case IB_QPT_DRIVER:
- return "IB_QPT_DRIVER";
- case IB_QPT_MAX:
+ break;
default:
- return "Invalid QP type";
+ goto out;
}
+
+ *type = attr->qp_type;
+ return 0;
+
+out:
+ mlx5_ib_dbg(dev, "Unsupported QP type %d\n", attr->qp_type);
+ return -EOPNOTSUPP;
}
-static struct ib_qp *mlx5_ib_create_dct(struct ib_pd *pd,
- struct ib_qp_init_attr *attr,
- struct mlx5_ib_create_qp *ucmd)
+static int check_valid_flow(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct ib_qp_init_attr *attr,
+ struct ib_udata *udata)
{
- struct mlx5_ib_qp *qp;
- int err = 0;
- u32 uidx = MLX5_IB_DEFAULT_UIDX;
- void *dctc;
+ struct mlx5_ib_ucontext *ucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
- if (!attr->srq || !attr->recv_cq)
- return ERR_PTR(-EINVAL);
+ if (!udata) {
+ /* Kernel create_qp callers */
+ if (attr->rwq_ind_tbl)
+ return -EOPNOTSUPP;
- err = get_qp_user_index(to_mucontext(pd->uobject->context),
- ucmd, sizeof(*ucmd), &uidx);
- if (err)
- return ERR_PTR(err);
+ switch (attr->qp_type) {
+ case IB_QPT_RAW_PACKET:
+ case IB_QPT_DRIVER:
+ return -EOPNOTSUPP;
+ default:
+ return 0;
+ }
+ }
- qp = kzalloc(sizeof(*qp), GFP_KERNEL);
- if (!qp)
- return ERR_PTR(-ENOMEM);
+ /* Userspace create_qp callers */
+ if (attr->qp_type == IB_QPT_RAW_PACKET && !ucontext->cqe_version) {
+ mlx5_ib_dbg(dev,
+ "Raw Packet QP is only supported for CQE version > 0\n");
+ return -EINVAL;
+ }
- qp->dct.in = kzalloc(MLX5_ST_SZ_BYTES(create_dct_in), GFP_KERNEL);
- if (!qp->dct.in) {
- err = -ENOMEM;
- goto err_free;
+ if (attr->qp_type != IB_QPT_RAW_PACKET && attr->rwq_ind_tbl) {
+ mlx5_ib_dbg(dev,
+ "Wrong QP type %d for the RWQ indirect table\n",
+ attr->qp_type);
+ return -EINVAL;
}
- MLX5_SET(create_dct_in, qp->dct.in, uid, to_mpd(pd)->uid);
- dctc = MLX5_ADDR_OF(create_dct_in, qp->dct.in, dct_context_entry);
- qp->qp_sub_type = MLX5_IB_QPT_DCT;
- MLX5_SET(dctc, dctc, pd, to_mpd(pd)->pdn);
- MLX5_SET(dctc, dctc, srqn_xrqn, to_msrq(attr->srq)->msrq.srqn);
- MLX5_SET(dctc, dctc, cqn, to_mcq(attr->recv_cq)->mcq.cqn);
- MLX5_SET64(dctc, dctc, dc_access_key, ucmd->access_key);
- MLX5_SET(dctc, dctc, user_index, uidx);
+ /*
+ * We don't need to see this warning, it means that kernel code
+ * missing ib_pd. Placed here to catch developer's mistakes.
+ */
+ WARN_ONCE(!pd && attr->qp_type != IB_QPT_XRC_TGT,
+ "There is a missing PD pointer assignment\n");
+ return 0;
+}
- if (ucmd->flags & MLX5_QP_FLAG_SCATTER_CQE)
- configure_responder_scat_cqe(attr, dctc);
+static bool get_dp_ooo_cap(struct mlx5_core_dev *mdev, enum ib_qp_type qp_type)
+{
+ if (!MLX5_CAP_GEN_2(mdev, dp_ordering_force))
+ return false;
- qp->state = IB_QPS_RESET;
+ switch (qp_type) {
+ case IB_QPT_RC:
+ return MLX5_CAP_GEN(mdev, dp_ordering_ooo_all_rc);
+ case IB_QPT_XRC_INI:
+ case IB_QPT_XRC_TGT:
+ return MLX5_CAP_GEN(mdev, dp_ordering_ooo_all_xrc);
+ case IB_QPT_UC:
+ return MLX5_CAP_GEN(mdev, dp_ordering_ooo_all_uc);
+ case IB_QPT_UD:
+ return MLX5_CAP_GEN(mdev, dp_ordering_ooo_all_ud);
+ case MLX5_IB_QPT_DCI:
+ case MLX5_IB_QPT_DCT:
+ return MLX5_CAP_GEN(mdev, dp_ordering_ooo_all_dc);
+ default:
+ return false;
+ }
+}
- return &qp->ibqp;
-err_free:
- kfree(qp);
- return ERR_PTR(err);
+static void process_vendor_flag(struct mlx5_ib_dev *dev, int *flags, int flag,
+ bool cond, struct mlx5_ib_qp *qp)
+{
+ if (!(*flags & flag))
+ return;
+
+ if (cond) {
+ qp->flags_en |= flag;
+ *flags &= ~flag;
+ return;
+ }
+
+ switch (flag) {
+ case MLX5_QP_FLAG_SCATTER_CQE:
+ case MLX5_QP_FLAG_ALLOW_SCATTER_CQE:
+ /*
+ * We don't return error if these flags were provided,
+ * and mlx5 doesn't have right capability.
+ */
+ *flags &= ~(MLX5_QP_FLAG_SCATTER_CQE |
+ MLX5_QP_FLAG_ALLOW_SCATTER_CQE);
+ return;
+ default:
+ break;
+ }
+ mlx5_ib_dbg(dev, "Vendor create QP flag 0x%X is not supported\n", flag);
}
-static int set_mlx_qp_type(struct mlx5_ib_dev *dev,
- struct ib_qp_init_attr *init_attr,
- struct mlx5_ib_create_qp *ucmd,
- struct ib_udata *udata)
+static int process_vendor_flags(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ void *ucmd, struct ib_qp_init_attr *attr)
{
- enum { MLX_QP_FLAGS = MLX5_QP_FLAG_TYPE_DCT | MLX5_QP_FLAG_TYPE_DCI };
- int err;
+ struct mlx5_core_dev *mdev = dev->mdev;
+ bool cond;
+ int flags;
- if (!udata)
+ if (attr->rwq_ind_tbl)
+ flags = ((struct mlx5_ib_create_qp_rss *)ucmd)->flags;
+ else
+ flags = ((struct mlx5_ib_create_qp *)ucmd)->flags;
+
+ switch (flags & (MLX5_QP_FLAG_TYPE_DCT | MLX5_QP_FLAG_TYPE_DCI)) {
+ case MLX5_QP_FLAG_TYPE_DCI:
+ qp->type = MLX5_IB_QPT_DCI;
+ break;
+ case MLX5_QP_FLAG_TYPE_DCT:
+ qp->type = MLX5_IB_QPT_DCT;
+ break;
+ default:
+ if (qp->type != IB_QPT_DRIVER)
+ break;
+ /*
+ * It is IB_QPT_DRIVER and or no subtype or
+ * wrong subtype were provided.
+ */
return -EINVAL;
+ }
- if (udata->inlen < sizeof(*ucmd)) {
- mlx5_ib_dbg(dev, "create_qp user command is smaller than expected\n");
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_TYPE_DCI, true, qp);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_TYPE_DCT, true, qp);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_DCI_STREAM,
+ MLX5_CAP_GEN(mdev, log_max_dci_stream_channels),
+ qp);
+
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_SIGNATURE, true, qp);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_SCATTER_CQE,
+ MLX5_CAP_GEN(mdev, sctr_data_cqe), qp);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_ALLOW_SCATTER_CQE,
+ MLX5_CAP_GEN(mdev, sctr_data_cqe), qp);
+
+ if (qp->type == IB_QPT_RAW_PACKET) {
+ cond = MLX5_CAP_ETH(mdev, tunnel_stateless_vxlan) ||
+ MLX5_CAP_ETH(mdev, tunnel_stateless_gre) ||
+ MLX5_CAP_ETH(mdev, tunnel_stateless_geneve_rx);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_TUNNEL_OFFLOADS,
+ cond, qp);
+ process_vendor_flag(dev, &flags,
+ MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC, true,
+ qp);
+ process_vendor_flag(dev, &flags,
+ MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC, true,
+ qp);
+ }
+
+ if (qp->type == IB_QPT_RC)
+ process_vendor_flag(dev, &flags,
+ MLX5_QP_FLAG_PACKET_BASED_CREDIT_MODE,
+ MLX5_CAP_GEN(mdev, qp_packet_based), qp);
+
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_BFREG_INDEX, true, qp);
+ process_vendor_flag(dev, &flags, MLX5_QP_FLAG_UAR_PAGE_INDEX, true, qp);
+
+ cond = qp->flags_en & ~(MLX5_QP_FLAG_TUNNEL_OFFLOADS |
+ MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_UC |
+ MLX5_QP_FLAG_TIR_ALLOW_SELF_LB_MC);
+ if (attr->rwq_ind_tbl && cond) {
+ mlx5_ib_dbg(dev, "RSS RAW QP has unsupported flags 0x%X\n",
+ cond);
return -EINVAL;
}
- err = ib_copy_from_udata(ucmd, udata, sizeof(*ucmd));
- if (err)
- return err;
- if ((ucmd->flags & MLX_QP_FLAGS) == MLX5_QP_FLAG_TYPE_DCI) {
- init_attr->qp_type = MLX5_IB_QPT_DCI;
- } else {
- if ((ucmd->flags & MLX_QP_FLAGS) == MLX5_QP_FLAG_TYPE_DCT) {
- init_attr->qp_type = MLX5_IB_QPT_DCT;
- } else {
- mlx5_ib_dbg(dev, "Invalid QP flags\n");
- return -EINVAL;
- }
+ if (flags)
+ mlx5_ib_dbg(dev, "udata has unsupported flags 0x%X\n", flags);
+
+ return (flags) ? -EINVAL : 0;
}
- if (!MLX5_CAP_GEN(dev->mdev, dct)) {
- mlx5_ib_dbg(dev, "DC transport is not supported\n");
- return -EOPNOTSUPP;
+static void process_create_flag(struct mlx5_ib_dev *dev, int *flags, int flag,
+ bool cond, struct mlx5_ib_qp *qp)
+{
+ if (!(*flags & flag))
+ return;
+
+ if (cond) {
+ qp->flags |= flag;
+ *flags &= ~flag;
+ return;
}
+ mlx5_ib_dbg(dev, "Verbs create QP flag 0x%X is not supported\n", flag);
+}
+
+static int process_create_flags(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ struct ib_qp_init_attr *attr)
+{
+ enum ib_qp_type qp_type = qp->type;
+ struct mlx5_core_dev *mdev = dev->mdev;
+ int create_flags = attr->create_flags;
+ bool cond;
+
+ if (qp_type == MLX5_IB_QPT_DCT)
+ return (create_flags) ? -EINVAL : 0;
+
+ if (qp_type == IB_QPT_RAW_PACKET && attr->rwq_ind_tbl)
+ return (create_flags) ? -EINVAL : 0;
+
+ process_create_flag(dev, &create_flags, IB_QP_CREATE_NETIF_QP,
+ mlx5_get_flow_namespace(dev->mdev,
+ MLX5_FLOW_NAMESPACE_BYPASS),
+ qp);
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_INTEGRITY_EN,
+ MLX5_CAP_GEN(mdev, sho), qp);
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK,
+ MLX5_CAP_GEN(mdev, block_lb_mc), qp);
+ process_create_flag(dev, &create_flags, IB_QP_CREATE_CROSS_CHANNEL,
+ MLX5_CAP_GEN(mdev, cd), qp);
+ process_create_flag(dev, &create_flags, IB_QP_CREATE_MANAGED_SEND,
+ MLX5_CAP_GEN(mdev, cd), qp);
+ process_create_flag(dev, &create_flags, IB_QP_CREATE_MANAGED_RECV,
+ MLX5_CAP_GEN(mdev, cd), qp);
+
+ if (qp_type == IB_QPT_UD) {
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_IPOIB_UD_LSO,
+ MLX5_CAP_GEN(mdev, ipoib_basic_offloads),
+ qp);
+ cond = MLX5_CAP_GEN(mdev, port_type) == MLX5_CAP_PORT_TYPE_IB;
+ process_create_flag(dev, &create_flags, IB_QP_CREATE_SOURCE_QPN,
+ cond, qp);
+ }
+
+ if (qp_type == IB_QPT_RAW_PACKET) {
+ cond = MLX5_CAP_GEN(mdev, eth_net_offloads) &&
+ MLX5_CAP_ETH(mdev, scatter_fcs);
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_SCATTER_FCS, cond, qp);
+
+ cond = MLX5_CAP_GEN(mdev, eth_net_offloads) &&
+ MLX5_CAP_ETH(mdev, vlan_cap);
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_CVLAN_STRIPPING, cond, qp);
+ }
+
+ process_create_flag(dev, &create_flags,
+ IB_QP_CREATE_PCI_WRITE_END_PADDING,
+ MLX5_CAP_GEN(mdev, end_pad), qp);
+
+ process_create_flag(dev, &create_flags, MLX5_IB_QP_CREATE_SQPN_QP1,
+ true, qp);
+
+ if (create_flags) {
+ mlx5_ib_dbg(dev, "Create QP has unsupported flags 0x%X\n",
+ create_flags);
+ return -EOPNOTSUPP;
+ }
return 0;
}
-struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
- struct ib_qp_init_attr *verbs_init_attr,
- struct ib_udata *udata)
+static int process_udata_size(struct mlx5_ib_dev *dev,
+ struct mlx5_create_qp_params *params)
{
- struct mlx5_ib_dev *dev;
- struct mlx5_ib_qp *qp;
- u16 xrcdn = 0;
- int err;
- struct ib_qp_init_attr mlx_init_attr;
- struct ib_qp_init_attr *init_attr = verbs_init_attr;
-
- if (pd) {
- dev = to_mdev(pd->device);
-
- if (init_attr->qp_type == IB_QPT_RAW_PACKET) {
- if (!udata) {
- mlx5_ib_dbg(dev, "Raw Packet QP is not supported for kernel consumers\n");
- return ERR_PTR(-EINVAL);
- } else if (!to_mucontext(pd->uobject->context)->cqe_version) {
- mlx5_ib_dbg(dev, "Raw Packet QP is only supported for CQE version > 0\n");
- return ERR_PTR(-EINVAL);
- }
- }
- } else {
- /* being cautious here */
- if (init_attr->qp_type != IB_QPT_XRC_TGT &&
- init_attr->qp_type != MLX5_IB_QPT_REG_UMR) {
- pr_warn("%s: no PD for transport %s\n", __func__,
- ib_qp_type_str(init_attr->qp_type));
- return ERR_PTR(-EINVAL);
- }
- dev = to_mdev(to_mxrcd(init_attr->xrcd)->ibxrcd.device);
+ size_t ucmd = sizeof(struct mlx5_ib_create_qp);
+ struct ib_udata *udata = params->udata;
+ size_t outlen = udata->outlen;
+ size_t inlen = udata->inlen;
+
+ params->outlen = min(outlen, sizeof(struct mlx5_ib_create_qp_resp));
+ params->ucmd_size = ucmd;
+ if (!params->is_rss_raw) {
+ /* User has old rdma-core, which doesn't support ECE */
+ size_t min_inlen =
+ offsetof(struct mlx5_ib_create_qp, ece_options);
+
+ /*
+ * We will check in check_ucmd_data() that user
+ * cleared everything after inlen.
+ */
+ params->inlen = (inlen < min_inlen) ? 0 : min(inlen, ucmd);
+ goto out;
}
- if (init_attr->qp_type == IB_QPT_DRIVER) {
- struct mlx5_ib_create_qp ucmd;
+ /* RSS RAW QP */
+ if (inlen < offsetofend(struct mlx5_ib_create_qp_rss, flags))
+ return -EINVAL;
- init_attr = &mlx_init_attr;
- memcpy(init_attr, verbs_init_attr, sizeof(*verbs_init_attr));
- err = set_mlx_qp_type(dev, init_attr, &ucmd, udata);
- if (err)
- return ERR_PTR(err);
+ if (outlen < offsetofend(struct mlx5_ib_create_qp_resp, bfreg_index))
+ return -EINVAL;
- if (init_attr->qp_type == MLX5_IB_QPT_DCI) {
- if (init_attr->cap.max_recv_wr ||
- init_attr->cap.max_recv_sge) {
- mlx5_ib_dbg(dev, "DCI QP requires zero size receive queue\n");
- return ERR_PTR(-EINVAL);
- }
- } else {
- return mlx5_ib_create_dct(pd, init_attr, &ucmd);
- }
+ ucmd = sizeof(struct mlx5_ib_create_qp_rss);
+ params->ucmd_size = ucmd;
+ if (inlen > ucmd && !ib_is_udata_cleared(udata, ucmd, inlen - ucmd))
+ return -EINVAL;
+
+ params->inlen = min(ucmd, inlen);
+out:
+ if (!params->inlen)
+ mlx5_ib_dbg(dev, "udata is too small\n");
+
+ return (params->inlen) ? 0 : -EINVAL;
+}
+
+static int create_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
+ struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
+{
+ int err;
+
+ if (params->is_rss_raw) {
+ err = create_rss_raw_qp_tir(dev, pd, qp, params);
+ goto out;
}
- switch (init_attr->qp_type) {
+ switch (qp->type) {
+ case MLX5_IB_QPT_DCT:
+ err = create_dct(dev, pd, qp, params);
+ break;
+ case MLX5_IB_QPT_DCI:
+ err = create_dci(dev, pd, qp, params);
+ break;
case IB_QPT_XRC_TGT:
- case IB_QPT_XRC_INI:
- if (!MLX5_CAP_GEN(dev->mdev, xrc)) {
- mlx5_ib_dbg(dev, "XRC not supported\n");
- return ERR_PTR(-ENOSYS);
- }
- init_attr->recv_cq = NULL;
- if (init_attr->qp_type == IB_QPT_XRC_TGT) {
- xrcdn = to_mxrcd(init_attr->xrcd)->xrcdn;
- init_attr->send_cq = NULL;
- }
-
- /* fall through */
- case IB_QPT_RAW_PACKET:
- case IB_QPT_RC:
- case IB_QPT_UC:
- case IB_QPT_UD:
- case IB_QPT_SMI:
+ err = create_xrc_tgt_qp(dev, qp, params);
+ break;
+ case IB_QPT_GSI:
+ err = mlx5_ib_create_gsi(pd, qp, params->attr);
+ break;
case MLX5_IB_QPT_HW_GSI:
+ rdma_restrack_no_track(&qp->ibqp.res);
+ fallthrough;
case MLX5_IB_QPT_REG_UMR:
- case MLX5_IB_QPT_DCI:
- qp = kzalloc(sizeof(*qp), GFP_KERNEL);
- if (!qp)
- return ERR_PTR(-ENOMEM);
-
- err = create_qp_common(dev, pd, init_attr, udata, qp);
- if (err) {
- mlx5_ib_dbg(dev, "create_qp_common failed\n");
- kfree(qp);
- return ERR_PTR(err);
- }
-
- if (is_qp0(init_attr->qp_type))
- qp->ibqp.qp_num = 0;
- else if (is_qp1(init_attr->qp_type))
- qp->ibqp.qp_num = 1;
+ default:
+ if (params->udata)
+ err = create_user_qp(dev, pd, qp, params);
else
- qp->ibqp.qp_num = qp->trans_qp.base.mqp.qpn;
+ err = create_kernel_qp(dev, pd, qp, params);
+ }
- mlx5_ib_dbg(dev, "ib qpnum 0x%x, mlx qpn 0x%x, rcqn 0x%x, scqn 0x%x\n",
- qp->ibqp.qp_num, qp->trans_qp.base.mqp.qpn,
- init_attr->recv_cq ? to_mcq(init_attr->recv_cq)->mcq.cqn : -1,
- init_attr->send_cq ? to_mcq(init_attr->send_cq)->mcq.cqn : -1);
+out:
+ if (err) {
+ mlx5_ib_err(dev, "Create QP type %d failed\n", qp->type);
+ return err;
+ }
- qp->trans_qp.xrcdn = xrcdn;
+ if (is_qp0(qp->type))
+ qp->ibqp.qp_num = 0;
+ else if (is_qp1(qp->type))
+ qp->ibqp.qp_num = 1;
+ else
+ qp->ibqp.qp_num = qp->trans_qp.base.mqp.qpn;
- break;
+ mlx5_ib_dbg(dev,
+ "QP type %d, ib qpn 0x%X, mlx qpn 0x%x, rcqn 0x%x, scqn 0x%x, ece 0x%x\n",
+ qp->type, qp->ibqp.qp_num, qp->trans_qp.base.mqp.qpn,
+ params->attr->recv_cq ? to_mcq(params->attr->recv_cq)->mcq.cqn :
+ -1,
+ params->attr->send_cq ? to_mcq(params->attr->send_cq)->mcq.cqn :
+ -1,
+ params->resp.ece_options);
- case IB_QPT_GSI:
- return mlx5_ib_gsi_create_qp(pd, init_attr);
+ return 0;
+}
- case IB_QPT_RAW_IPV6:
- case IB_QPT_RAW_ETHERTYPE:
- case IB_QPT_MAX:
+static int check_qp_attr(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
+ struct ib_qp_init_attr *attr)
+{
+ int ret = 0;
+
+ switch (qp->type) {
+ case MLX5_IB_QPT_DCT:
+ ret = (!attr->srq || !attr->recv_cq) ? -EINVAL : 0;
+ break;
+ case MLX5_IB_QPT_DCI:
+ ret = (attr->cap.max_recv_wr || attr->cap.max_recv_sge) ?
+ -EINVAL :
+ 0;
+ break;
+ case IB_QPT_RAW_PACKET:
+ ret = (attr->rwq_ind_tbl && attr->send_cq) ? -EINVAL : 0;
+ break;
default:
- mlx5_ib_dbg(dev, "unsupported qp type %d\n",
- init_attr->qp_type);
- /* Don't support raw QPs */
- return ERR_PTR(-EINVAL);
+ break;
}
- if (verbs_init_attr->qp_type == IB_QPT_DRIVER)
- qp->qp_sub_type = init_attr->qp_type;
+ if (ret)
+ mlx5_ib_dbg(dev, "QP type %d has wrong attributes\n", qp->type);
- return &qp->ibqp;
+ return ret;
+}
+
+static int get_qp_uidx(struct mlx5_ib_qp *qp,
+ struct mlx5_create_qp_params *params)
+{
+ struct mlx5_ib_create_qp *ucmd = params->ucmd;
+ struct ib_udata *udata = params->udata;
+ struct mlx5_ib_ucontext *ucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
+
+ if (params->is_rss_raw)
+ return 0;
+
+ return get_qp_user_index(ucontext, ucmd, sizeof(*ucmd), &params->uidx);
}
static int mlx5_ib_destroy_dct(struct mlx5_ib_qp *mqp)
@@ -2619,7 +3206,7 @@ static int mlx5_ib_destroy_dct(struct mlx5_ib_qp *mqp)
if (mqp->state == IB_QPS_RTR) {
int err;
- err = mlx5_core_destroy_dct(dev->mdev, &mqp->dct.mdct);
+ err = mlx5_core_destroy_dct(dev, &mqp->dct.mdct);
if (err) {
mlx5_ib_warn(dev, "failed to destroy DCT %d\n", err);
return err;
@@ -2627,37 +3214,166 @@ static int mlx5_ib_destroy_dct(struct mlx5_ib_qp *mqp)
}
kfree(mqp->dct.in);
- kfree(mqp);
return 0;
}
-int mlx5_ib_destroy_qp(struct ib_qp *qp)
+static int check_ucmd_data(struct mlx5_ib_dev *dev,
+ struct mlx5_create_qp_params *params)
+{
+ struct ib_udata *udata = params->udata;
+ size_t size, last;
+ int ret;
+
+ if (params->is_rss_raw)
+ /*
+ * These QPs don't have "reserved" field in their
+ * create_qp input struct, so their data is always valid.
+ */
+ last = sizeof(struct mlx5_ib_create_qp_rss);
+ else
+ last = offsetof(struct mlx5_ib_create_qp, reserved);
+
+ if (udata->inlen <= last)
+ return 0;
+
+ /*
+ * User provides different create_qp structures based on the
+ * flow and we need to know if he cleared memory after our
+ * struct create_qp ends.
+ */
+ size = udata->inlen - last;
+ ret = ib_is_udata_cleared(params->udata, last, size);
+ if (!ret)
+ mlx5_ib_dbg(
+ dev,
+ "udata is not cleared, inlen = %zu, ucmd = %zu, last = %zu, size = %zu\n",
+ udata->inlen, params->ucmd_size, last, size);
+ return ret ? 0 : -EINVAL;
+}
+
+int mlx5_ib_create_qp(struct ib_qp *ibqp, struct ib_qp_init_attr *attr,
+ struct ib_udata *udata)
+{
+ struct mlx5_create_qp_params params = {};
+ struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
+ struct mlx5_ib_qp *qp = to_mqp(ibqp);
+ struct ib_pd *pd = ibqp->pd;
+ enum ib_qp_type type;
+ int err;
+
+ err = mlx5_ib_dev_res_srq_init(dev);
+ if (err)
+ return err;
+
+ err = check_qp_type(dev, attr, &type);
+ if (err)
+ return err;
+
+ err = check_valid_flow(dev, pd, attr, udata);
+ if (err)
+ return err;
+
+ params.udata = udata;
+ params.uidx = MLX5_IB_DEFAULT_UIDX;
+ params.attr = attr;
+ params.is_rss_raw = !!attr->rwq_ind_tbl;
+
+ if (udata) {
+ err = process_udata_size(dev, &params);
+ if (err)
+ return err;
+
+ err = check_ucmd_data(dev, &params);
+ if (err)
+ return err;
+
+ params.ucmd = kzalloc(params.ucmd_size, GFP_KERNEL);
+ if (!params.ucmd)
+ return -ENOMEM;
+
+ err = ib_copy_from_udata(params.ucmd, udata, params.inlen);
+ if (err)
+ goto free_ucmd;
+ }
+
+ mutex_init(&qp->mutex);
+ qp->type = type;
+ if (udata) {
+ err = process_vendor_flags(dev, qp, params.ucmd, attr);
+ if (err)
+ goto free_ucmd;
+
+ err = get_qp_uidx(qp, &params);
+ if (err)
+ goto free_ucmd;
+ }
+ err = process_create_flags(dev, qp, attr);
+ if (err)
+ goto free_ucmd;
+
+ err = check_qp_attr(dev, qp, attr);
+ if (err)
+ goto free_ucmd;
+
+ err = create_qp(dev, pd, qp, &params);
+ if (err)
+ goto free_ucmd;
+
+ kfree(params.ucmd);
+ params.ucmd = NULL;
+
+ if (udata)
+ /*
+ * It is safe to copy response for all user create QP flows,
+ * including MLX5_IB_QPT_DCT, which doesn't need it.
+ * In that case, resp will be filled with zeros.
+ */
+ err = ib_copy_to_udata(udata, &params.resp, params.outlen);
+ if (err)
+ goto destroy_qp;
+
+ return 0;
+
+destroy_qp:
+ switch (qp->type) {
+ case MLX5_IB_QPT_DCT:
+ mlx5_ib_destroy_dct(qp);
+ break;
+ case IB_QPT_GSI:
+ mlx5_ib_destroy_gsi(qp);
+ break;
+ default:
+ destroy_qp_common(dev, qp, udata);
+ }
+
+free_ucmd:
+ kfree(params.ucmd);
+ return err;
+}
+
+int mlx5_ib_destroy_qp(struct ib_qp *qp, struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(qp->device);
struct mlx5_ib_qp *mqp = to_mqp(qp);
- if (unlikely(qp->qp_type == IB_QPT_GSI))
- return mlx5_ib_gsi_destroy_qp(qp);
+ if (mqp->type == IB_QPT_GSI)
+ return mlx5_ib_destroy_gsi(mqp);
- if (mqp->qp_sub_type == MLX5_IB_QPT_DCT)
+ if (mqp->type == MLX5_IB_QPT_DCT)
return mlx5_ib_destroy_dct(mqp);
- destroy_qp_common(dev, mqp);
-
- kfree(mqp);
-
+ destroy_qp_common(dev, mqp, udata);
return 0;
}
-static int to_mlx5_access_flags(struct mlx5_ib_qp *qp,
- const struct ib_qp_attr *attr,
- int attr_mask, __be32 *hw_access_flags)
+static int set_qpc_atomic_flags(struct mlx5_ib_qp *qp,
+ const struct ib_qp_attr *attr, int attr_mask,
+ void *qpc)
{
+ struct mlx5_ib_dev *dev = to_mdev(qp->ibqp.device);
u8 dest_rd_atomic;
u32 access_flags;
- struct mlx5_ib_dev *dev = to_mdev(qp->ibqp.device);
-
if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC)
dest_rd_atomic = attr->max_dest_rd_atomic;
else
@@ -2671,24 +3387,20 @@ static int to_mlx5_access_flags(struct mlx5_ib_qp *qp,
if (!dest_rd_atomic)
access_flags &= IB_ACCESS_REMOTE_WRITE;
- if (access_flags & IB_ACCESS_REMOTE_READ)
- *hw_access_flags |= MLX5_QP_BIT_RRE;
+ MLX5_SET(qpc, qpc, rre, !!(access_flags & IB_ACCESS_REMOTE_READ));
+
if (access_flags & IB_ACCESS_REMOTE_ATOMIC) {
int atomic_mode;
- atomic_mode = get_atomic_mode(dev, qp->ibqp.qp_type);
+ atomic_mode = get_atomic_mode(dev, qp);
if (atomic_mode < 0)
return -EOPNOTSUPP;
- *hw_access_flags |= MLX5_QP_BIT_RAE;
- *hw_access_flags |= atomic_mode << MLX5_ATOMIC_MODE_OFFSET;
+ MLX5_SET(qpc, qpc, rae, 1);
+ MLX5_SET(qpc, qpc, atomic_mode, atomic_mode);
}
- if (access_flags & IB_ACCESS_REMOTE_WRITE)
- *hw_access_flags |= MLX5_QP_BIT_RWE;
-
- *hw_access_flags = cpu_to_be32(*hw_access_flags);
-
+ MLX5_SET(qpc, qpc, rwe, !!(access_flags & IB_ACCESS_REMOTE_WRITE));
return 0;
}
@@ -2698,20 +3410,60 @@ enum {
MLX5_PATH_FLAG_COUNTER = 1 << 2,
};
-static int ib_rate_to_mlx5(struct mlx5_ib_dev *dev, u8 rate)
+static int mlx5_to_ib_rate_map(u8 rate)
{
- if (rate == IB_RATE_PORT_CURRENT)
+ static const int rates[] = { IB_RATE_PORT_CURRENT, IB_RATE_56_GBPS,
+ IB_RATE_25_GBPS, IB_RATE_100_GBPS,
+ IB_RATE_200_GBPS, IB_RATE_50_GBPS,
+ IB_RATE_400_GBPS };
+
+ if (rate < ARRAY_SIZE(rates))
+ return rates[rate];
+
+ return rate - MLX5_STAT_RATE_OFFSET;
+}
+
+static int ib_to_mlx5_rate_map(u8 rate)
+{
+ switch (rate) {
+ case IB_RATE_PORT_CURRENT:
return 0;
+ case IB_RATE_56_GBPS:
+ return 1;
+ case IB_RATE_25_GBPS:
+ return 2;
+ case IB_RATE_100_GBPS:
+ return 3;
+ case IB_RATE_200_GBPS:
+ return 4;
+ case IB_RATE_50_GBPS:
+ return 5;
+ case IB_RATE_400_GBPS:
+ return 6;
+ default:
+ return rate + MLX5_STAT_RATE_OFFSET;
+ }
- if (rate < IB_RATE_2_5_GBPS || rate > IB_RATE_600_GBPS)
+ return 0;
+}
+
+int mlx5r_ib_rate(struct mlx5_ib_dev *dev, u8 rate)
+{
+ u32 stat_rate_support;
+
+ if (rate == IB_RATE_PORT_CURRENT || rate == IB_RATE_800_GBPS ||
+ rate == IB_RATE_1600_GBPS)
+ return 0;
+
+ if (rate < IB_RATE_2_5_GBPS || rate > IB_RATE_1600_GBPS)
return -EINVAL;
+ stat_rate_support = MLX5_CAP_GEN(dev->mdev, stat_rate_support);
while (rate != IB_RATE_PORT_CURRENT &&
- !(1 << (rate + MLX5_STAT_RATE_OFFSET) &
- MLX5_CAP_GEN(dev->mdev, stat_rate_support)))
+ !(1 << ib_to_mlx5_rate_map(rate) & stat_rate_support))
--rate;
- return rate ? rate + MLX5_STAT_RATE_OFFSET : rate;
+ return ib_to_mlx5_rate_map(rate);
}
static int modify_raw_packet_eth_prio(struct mlx5_core_dev *dev,
@@ -2734,7 +3486,7 @@ static int modify_raw_packet_eth_prio(struct mlx5_core_dev *dev,
tisc = MLX5_ADDR_OF(modify_tis_in, in, ctx);
MLX5_SET(tisc, tisc, prio, ((sl & 0x7) << 1));
- err = mlx5_core_modify_tis(dev, sq->tisn, in, inlen);
+ err = mlx5_core_modify_tis(dev, sq->tisn, in);
kvfree(in);
@@ -2761,18 +3513,29 @@ static int modify_raw_packet_tx_affinity(struct mlx5_core_dev *dev,
tisc = MLX5_ADDR_OF(modify_tis_in, in, ctx);
MLX5_SET(tisc, tisc, lag_tx_port_affinity, tx_affinity);
- err = mlx5_core_modify_tis(dev, sq->tisn, in, inlen);
+ err = mlx5_core_modify_tis(dev, sq->tisn, in);
kvfree(in);
return err;
}
+static void mlx5_set_path_udp_sport(void *path, const struct rdma_ah_attr *ah,
+ u32 lqpn, u32 rqpn)
+
+{
+ u32 fl = ah->grh.flow_label;
+
+ if (!fl)
+ fl = rdma_calc_flow_label(lqpn, rqpn);
+
+ MLX5_SET(ads, path, udp_sport, rdma_flow_label_to_udp_sport(fl));
+}
+
static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
- const struct rdma_ah_attr *ah,
- struct mlx5_qp_path *path, u8 port, int attr_mask,
- u32 path_flags, const struct ib_qp_attr *attr,
- bool alt)
+ const struct rdma_ah_attr *ah, void *path, u8 port,
+ int attr_mask, u32 path_flags,
+ const struct ib_qp_attr *attr, bool alt)
{
const struct ib_global_route *grh = rdma_ah_read_grh(ah);
int err;
@@ -2781,15 +3544,17 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
u8 sl = rdma_ah_get_sl(ah);
if (attr_mask & IB_QP_PKEY_INDEX)
- path->pkey_index = cpu_to_be16(alt ? attr->alt_pkey_index :
- attr->pkey_index);
+ MLX5_SET(ads, path, pkey_index,
+ alt ? attr->alt_pkey_index : attr->pkey_index);
if (ah_flags & IB_AH_GRH) {
- if (grh->sgid_index >=
- dev->mdev->port_caps[port - 1].gid_table_len) {
+ const struct ib_port_immutable *immutable;
+
+ immutable = ib_port_immutable_read(&dev->ib_dev, port);
+ if (grh->sgid_index >= immutable->gid_tbl_len) {
pr_err("sgid_index (%u) too large. max is %d\n",
grh->sgid_index,
- dev->mdev->port_caps[port - 1].gid_table_len);
+ immutable->gid_tbl_len);
return -EINVAL;
}
}
@@ -2798,47 +3563,51 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
if (!(ah_flags & IB_AH_GRH))
return -EINVAL;
- memcpy(path->rmac, ah->roce.dmac, sizeof(ah->roce.dmac));
- if (qp->ibqp.qp_type == IB_QPT_RC ||
- qp->ibqp.qp_type == IB_QPT_UC ||
- qp->ibqp.qp_type == IB_QPT_XRC_INI ||
- qp->ibqp.qp_type == IB_QPT_XRC_TGT)
- path->udp_sport =
- mlx5_get_roce_udp_sport(dev, ah->grh.sgid_attr);
- path->dci_cfi_prio_sl = (sl & 0x7) << 4;
+ ether_addr_copy(MLX5_ADDR_OF(ads, path, rmac_47_32),
+ ah->roce.dmac);
+ if ((qp->type == IB_QPT_RC ||
+ qp->type == IB_QPT_UC ||
+ qp->type == IB_QPT_XRC_INI ||
+ qp->type == IB_QPT_XRC_TGT) &&
+ (grh->sgid_attr->gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP) &&
+ (attr_mask & IB_QP_DEST_QPN))
+ mlx5_set_path_udp_sport(path, ah,
+ qp->ibqp.qp_num,
+ attr->dest_qp_num);
+ MLX5_SET(ads, path, eth_prio, sl & 0x7);
gid_type = ah->grh.sgid_attr->gid_type;
if (gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP)
- path->ecn_dscp = (grh->traffic_class >> 2) & 0x3f;
+ MLX5_SET(ads, path, dscp, grh->traffic_class >> 2);
} else {
- path->fl_free_ar = (path_flags & MLX5_PATH_FLAG_FL) ? 0x80 : 0;
- path->fl_free_ar |=
- (path_flags & MLX5_PATH_FLAG_FREE_AR) ? 0x40 : 0;
- path->rlid = cpu_to_be16(rdma_ah_get_dlid(ah));
- path->grh_mlid = rdma_ah_get_path_bits(ah) & 0x7f;
- if (ah_flags & IB_AH_GRH)
- path->grh_mlid |= 1 << 7;
- path->dci_cfi_prio_sl = sl & 0xf;
+ MLX5_SET(ads, path, fl, !!(path_flags & MLX5_PATH_FLAG_FL));
+ MLX5_SET(ads, path, free_ar,
+ !!(path_flags & MLX5_PATH_FLAG_FREE_AR));
+ MLX5_SET(ads, path, rlid, rdma_ah_get_dlid(ah));
+ MLX5_SET(ads, path, mlid, rdma_ah_get_path_bits(ah));
+ MLX5_SET(ads, path, grh, !!(ah_flags & IB_AH_GRH));
+ MLX5_SET(ads, path, sl, sl);
}
if (ah_flags & IB_AH_GRH) {
- path->mgid_index = grh->sgid_index;
- path->hop_limit = grh->hop_limit;
- path->tclass_flowlabel =
- cpu_to_be32((grh->traffic_class << 20) |
- (grh->flow_label));
- memcpy(path->rgid, grh->dgid.raw, 16);
+ MLX5_SET(ads, path, src_addr_index, grh->sgid_index);
+ MLX5_SET(ads, path, hop_limit, grh->hop_limit);
+ MLX5_SET(ads, path, tclass, grh->traffic_class);
+ MLX5_SET(ads, path, flow_label, grh->flow_label);
+ memcpy(MLX5_ADDR_OF(ads, path, rgid_rip), grh->dgid.raw,
+ sizeof(grh->dgid.raw));
}
- err = ib_rate_to_mlx5(dev, rdma_ah_get_static_rate(ah));
+ err = mlx5r_ib_rate(dev, rdma_ah_get_static_rate(ah));
if (err < 0)
return err;
- path->static_rate = err;
- path->port = port;
+ MLX5_SET(ads, path, stat_rate, err);
+ MLX5_SET(ads, path, vhca_port_num, port);
if (attr_mask & IB_QP_TIMEOUT)
- path->ackto_lt = (alt ? attr->alt_timeout : attr->timeout) << 3;
+ MLX5_SET(ads, path, ack_timeout,
+ alt ? attr->alt_timeout : attr->timeout);
- if ((qp->ibqp.qp_type == IB_QPT_RAW_PACKET) && qp->sq.wqe_cnt)
+ if ((qp->type == IB_QPT_RAW_PACKET) && qp->sq.wqe_cnt)
return modify_raw_packet_eth_prio(dev->mdev,
&qp->raw_packet_qp.sq,
sl & 0xf, qp->ibqp.pd);
@@ -2853,23 +3622,33 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
MLX5_QP_OPTPAR_RAE |
MLX5_QP_OPTPAR_RWE |
MLX5_QP_OPTPAR_PKEY_INDEX |
- MLX5_QP_OPTPAR_PRI_PORT,
+ MLX5_QP_OPTPAR_PRI_PORT |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
[MLX5_QP_ST_UC] = MLX5_QP_OPTPAR_RWE |
MLX5_QP_OPTPAR_PKEY_INDEX |
- MLX5_QP_OPTPAR_PRI_PORT,
+ MLX5_QP_OPTPAR_PRI_PORT |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_PKEY_INDEX |
MLX5_QP_OPTPAR_Q_KEY |
MLX5_QP_OPTPAR_PRI_PORT,
+ [MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_RRE |
+ MLX5_QP_OPTPAR_RAE |
+ MLX5_QP_OPTPAR_RWE |
+ MLX5_QP_OPTPAR_PKEY_INDEX |
+ MLX5_QP_OPTPAR_PRI_PORT |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
},
[MLX5_QP_STATE_RTR] = {
[MLX5_QP_ST_RC] = MLX5_QP_OPTPAR_ALT_ADDR_PATH |
MLX5_QP_OPTPAR_RRE |
MLX5_QP_OPTPAR_RAE |
MLX5_QP_OPTPAR_RWE |
- MLX5_QP_OPTPAR_PKEY_INDEX,
+ MLX5_QP_OPTPAR_PKEY_INDEX |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
[MLX5_QP_ST_UC] = MLX5_QP_OPTPAR_ALT_ADDR_PATH |
MLX5_QP_OPTPAR_RWE |
- MLX5_QP_OPTPAR_PKEY_INDEX,
+ MLX5_QP_OPTPAR_PKEY_INDEX |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_PKEY_INDEX |
MLX5_QP_OPTPAR_Q_KEY,
[MLX5_QP_ST_MLX] = MLX5_QP_OPTPAR_PKEY_INDEX |
@@ -2878,7 +3657,8 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
MLX5_QP_OPTPAR_RRE |
MLX5_QP_OPTPAR_RAE |
MLX5_QP_OPTPAR_RWE |
- MLX5_QP_OPTPAR_PKEY_INDEX,
+ MLX5_QP_OPTPAR_PKEY_INDEX |
+ MLX5_QP_OPTPAR_LAG_TX_AFF,
},
},
[MLX5_QP_STATE_RTR] = {
@@ -2893,6 +3673,12 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
MLX5_QP_OPTPAR_RWE |
MLX5_QP_OPTPAR_PM_STATE,
[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_Q_KEY,
+ [MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_ALT_ADDR_PATH |
+ MLX5_QP_OPTPAR_RRE |
+ MLX5_QP_OPTPAR_RAE |
+ MLX5_QP_OPTPAR_RWE |
+ MLX5_QP_OPTPAR_PM_STATE |
+ MLX5_QP_OPTPAR_RNR_TIMEOUT,
},
},
[MLX5_QP_STATE_RTS] = {
@@ -2909,6 +3695,12 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_Q_KEY |
MLX5_QP_OPTPAR_SRQN |
MLX5_QP_OPTPAR_CQN_RCV,
+ [MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_RRE |
+ MLX5_QP_OPTPAR_RAE |
+ MLX5_QP_OPTPAR_RWE |
+ MLX5_QP_OPTPAR_RNR_TIMEOUT |
+ MLX5_QP_OPTPAR_PM_STATE |
+ MLX5_QP_OPTPAR_ALT_ADDR_PATH,
},
},
[MLX5_QP_STATE_SQER] = {
@@ -2920,6 +3712,21 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
MLX5_QP_OPTPAR_RWE |
MLX5_QP_OPTPAR_RAE |
MLX5_QP_OPTPAR_RRE,
+ [MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_RNR_TIMEOUT |
+ MLX5_QP_OPTPAR_RWE |
+ MLX5_QP_OPTPAR_RAE |
+ MLX5_QP_OPTPAR_RRE,
+ },
+ },
+ [MLX5_QP_STATE_SQD] = {
+ [MLX5_QP_STATE_RTS] = {
+ [MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_Q_KEY,
+ [MLX5_QP_ST_MLX] = MLX5_QP_OPTPAR_Q_KEY,
+ [MLX5_QP_ST_UC] = MLX5_QP_OPTPAR_RWE,
+ [MLX5_QP_ST_RC] = MLX5_QP_OPTPAR_RNR_TIMEOUT |
+ MLX5_QP_OPTPAR_RWE |
+ MLX5_QP_OPTPAR_RAE |
+ MLX5_QP_OPTPAR_RRE,
},
},
};
@@ -3020,7 +3827,7 @@ static int modify_raw_packet_qp_rq(
"RAW PACKET QP counters are not supported on current FW\n");
}
- err = mlx5_core_modify_rq(dev->mdev, rq->base.mqp.qpn, in, inlen);
+ err = mlx5_core_modify_rq(dev->mdev, rq->base.mqp.qpn, in);
if (err)
goto out;
@@ -3083,7 +3890,7 @@ static int modify_raw_packet_qp_sq(
MLX5_SET(sqc, sqc, packet_pacing_rate_limit_index, rl_index);
}
- err = mlx5_core_modify_sq(dev, sq->base.mqp.qpn, in, inlen);
+ err = mlx5_core_modify_sq(dev, sq->base.mqp.qpn, in);
if (err) {
/* Remove new rate from table if failed */
if (new_rate_added)
@@ -3092,10 +3899,12 @@ static int modify_raw_packet_qp_sq(
}
/* Only remove the old rate after new rate was set */
- if ((old_rl.rate &&
- !mlx5_rl_are_equal(&old_rl, &new_rl)) ||
- (new_state != MLX5_SQC_STATE_RDY))
+ if ((old_rl.rate && !mlx5_rl_are_equal(&old_rl, &new_rl)) ||
+ (new_state != MLX5_SQC_STATE_RDY)) {
mlx5_rl_remove_rate(dev, &old_rl);
+ if (new_state != MLX5_SQC_STATE_RDY)
+ memset(&new_rl, 0, sizeof(new_rl));
+ }
ibqp->rl = new_rl;
sq->state = new_state;
@@ -3121,7 +3930,7 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
switch (raw_qp_param->operation) {
case MLX5_CMD_OP_RST2INIT_QP:
rq_state = MLX5_RQC_STATE_RDY;
- sq_state = MLX5_SQC_STATE_RDY;
+ sq_state = MLX5_SQC_STATE_RST;
break;
case MLX5_CMD_OP_2ERR_QP:
rq_state = MLX5_RQC_STATE_ERR;
@@ -3133,13 +3942,11 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
break;
case MLX5_CMD_OP_RTR2RTS_QP:
case MLX5_CMD_OP_RTS2RTS_QP:
- if (raw_qp_param->set_mask ==
- MLX5_RAW_QP_RATE_LIMIT) {
- modify_rq = 0;
- sq_state = sq->state;
- } else {
- return raw_qp_param->set_mask ? -EINVAL : 0;
- }
+ if (raw_qp_param->set_mask & ~MLX5_RAW_QP_RATE_LIMIT)
+ return -EINVAL;
+
+ modify_rq = 0;
+ sq_state = MLX5_SQC_STATE_RDY;
break;
case MLX5_CMD_OP_INIT2INIT_QP:
case MLX5_CMD_OP_INIT2RTR_QP:
@@ -3160,6 +3967,8 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
}
if (modify_sq) {
+ struct mlx5_flow_handle *flow_rule;
+
if (tx_affinity) {
err = modify_raw_packet_tx_affinity(dev->mdev, sq,
tx_affinity,
@@ -3168,48 +3977,154 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
return err;
}
- return modify_raw_packet_qp_sq(dev->mdev, sq, sq_state,
- raw_qp_param, qp->ibqp.pd);
+ flow_rule = create_flow_rule_vport_sq(dev, sq,
+ raw_qp_param->port);
+ if (IS_ERR(flow_rule))
+ return PTR_ERR(flow_rule);
+
+ err = modify_raw_packet_qp_sq(dev->mdev, sq, sq_state,
+ raw_qp_param, qp->ibqp.pd);
+ if (err) {
+ if (flow_rule)
+ mlx5_del_flow_rules(flow_rule);
+ return err;
+ }
+
+ if (flow_rule) {
+ destroy_flow_rule_vport_sq(sq);
+ sq->flow_rule = flow_rule;
+ }
+
+ return err;
}
return 0;
}
-static unsigned int get_tx_affinity(struct mlx5_ib_dev *dev,
- struct mlx5_ib_pd *pd,
- struct mlx5_ib_qp_base *qp_base,
- u8 port_num)
+static unsigned int get_tx_affinity_rr(struct mlx5_ib_dev *dev,
+ struct ib_udata *udata)
+{
+ struct mlx5_ib_ucontext *ucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
+ u8 port_num = mlx5_core_native_port_num(dev->mdev) - 1;
+ atomic_t *tx_port_affinity;
+
+ if (ucontext)
+ tx_port_affinity = &ucontext->tx_port_affinity;
+ else
+ tx_port_affinity = &dev->port[port_num].roce.tx_port_affinity;
+
+ return (unsigned int)atomic_add_return(1, tx_port_affinity) %
+ (dev->lag_active ? dev->lag_ports : MLX5_CAP_GEN(dev->mdev, num_lag_ports)) + 1;
+}
+
+static bool qp_supports_affinity(struct mlx5_ib_qp *qp)
+{
+ if ((qp->type == IB_QPT_RC) || (qp->type == IB_QPT_UD) ||
+ (qp->type == IB_QPT_UC) || (qp->type == IB_QPT_RAW_PACKET) ||
+ (qp->type == IB_QPT_XRC_INI) || (qp->type == IB_QPT_XRC_TGT) ||
+ (qp->type == MLX5_IB_QPT_DCI))
+ return true;
+ return false;
+}
+
+static unsigned int get_tx_affinity(struct ib_qp *qp,
+ const struct ib_qp_attr *attr,
+ int attr_mask, u8 init,
+ struct ib_udata *udata)
{
- struct mlx5_ib_ucontext *ucontext = NULL;
- unsigned int tx_port_affinity;
+ struct mlx5_ib_ucontext *ucontext = rdma_udata_to_drv_context(
+ udata, struct mlx5_ib_ucontext, ibucontext);
+ struct mlx5_ib_dev *dev = to_mdev(qp->device);
+ struct mlx5_ib_qp *mqp = to_mqp(qp);
+ struct mlx5_ib_qp_base *qp_base;
+ unsigned int tx_affinity;
- if (pd && pd->ibpd.uobject && pd->ibpd.uobject->context)
- ucontext = to_mucontext(pd->ibpd.uobject->context);
+ if (!(mlx5_ib_lag_should_assign_affinity(dev) &&
+ qp_supports_affinity(mqp)))
+ return 0;
+
+ if (mqp->flags & MLX5_IB_QP_CREATE_SQPN_QP1)
+ tx_affinity = mqp->gsi_lag_port;
+ else if (init)
+ tx_affinity = get_tx_affinity_rr(dev, udata);
+ else if ((attr_mask & IB_QP_AV) && attr->xmit_slave)
+ tx_affinity =
+ mlx5_lag_get_slave_port(dev->mdev, attr->xmit_slave);
+ else
+ return 0;
- if (ucontext) {
- tx_port_affinity = (unsigned int)atomic_add_return(
- 1, &ucontext->tx_port_affinity) %
- MLX5_MAX_PORTS +
- 1;
+ qp_base = &mqp->trans_qp.base;
+ if (ucontext)
mlx5_ib_dbg(dev, "Set tx affinity 0x%x to qpn 0x%x ucontext %p\n",
- tx_port_affinity, qp_base->mqp.qpn, ucontext);
- } else {
- tx_port_affinity =
- (unsigned int)atomic_add_return(
- 1, &dev->roce[port_num].tx_port_affinity) %
- MLX5_MAX_PORTS +
- 1;
+ tx_affinity, qp_base->mqp.qpn, ucontext);
+ else
mlx5_ib_dbg(dev, "Set tx affinity 0x%x to qpn 0x%x\n",
- tx_port_affinity, qp_base->mqp.qpn);
- }
+ tx_affinity, qp_base->mqp.qpn);
+ return tx_affinity;
+}
+
+static int __mlx5_ib_qp_set_raw_qp_counter(struct mlx5_ib_qp *qp, u32 set_id,
+ struct mlx5_core_dev *mdev)
+{
+ struct mlx5_ib_raw_packet_qp *raw_packet_qp = &qp->raw_packet_qp;
+ struct mlx5_ib_rq *rq = &raw_packet_qp->rq;
+ u32 in[MLX5_ST_SZ_DW(modify_rq_in)] = {};
+ void *rqc;
+
+ if (!qp->rq.wqe_cnt)
+ return 0;
+
+ MLX5_SET(modify_rq_in, in, rq_state, rq->state);
+ MLX5_SET(modify_rq_in, in, uid, to_mpd(qp->ibqp.pd)->uid);
- return tx_port_affinity;
+ rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+ MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RDY);
+
+ MLX5_SET64(modify_rq_in, in, modify_bitmask,
+ MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_RQ_COUNTER_SET_ID);
+ MLX5_SET(rqc, rqc, counter_set_id, set_id);
+
+ return mlx5_core_modify_rq(mdev, rq->base.mqp.qpn, in);
+}
+
+static int __mlx5_ib_qp_set_counter(struct ib_qp *qp,
+ struct rdma_counter *counter)
+{
+ struct mlx5_ib_dev *dev = to_mdev(qp->device);
+ u32 in[MLX5_ST_SZ_DW(rts2rts_qp_in)] = {};
+ struct mlx5_ib_qp *mqp = to_mqp(qp);
+ struct mlx5_ib_qp_base *base;
+ u32 set_id;
+ u32 *qpc;
+
+ if (counter)
+ set_id = counter->id;
+ else
+ set_id = mlx5_ib_get_counters_id(dev, mqp->port - 1);
+
+ if (mqp->type == IB_QPT_RAW_PACKET)
+ return __mlx5_ib_qp_set_raw_qp_counter(mqp, set_id, dev->mdev);
+
+ base = &mqp->trans_qp.base;
+ MLX5_SET(rts2rts_qp_in, in, opcode, MLX5_CMD_OP_RTS2RTS_QP);
+ MLX5_SET(rts2rts_qp_in, in, qpn, base->mqp.qpn);
+ MLX5_SET(rts2rts_qp_in, in, uid, base->mqp.uid);
+ MLX5_SET(rts2rts_qp_in, in, opt_param_mask,
+ MLX5_QP_OPTPAR_COUNTER_SET_ID);
+
+ qpc = MLX5_ADDR_OF(rts2rts_qp_in, in, qpc);
+ MLX5_SET(qpc, qpc, counter_set_id, set_id);
+ return mlx5_cmd_exec_in(dev->mdev, rts2rts_qp, in);
}
static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
const struct ib_qp_attr *attr, int attr_mask,
- enum ib_qp_state cur_state, enum ib_qp_state new_state,
- const struct mlx5_ib_modify_qp *ucmd)
+ enum ib_qp_state cur_state,
+ enum ib_qp_state new_state,
+ const struct mlx5_ib_modify_qp *ucmd,
+ struct mlx5_ib_modify_qp_resp *resp,
+ struct ib_udata *udata)
{
static const u16 optab[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE] = {
[MLX5_QP_STATE_RST] = {
@@ -3236,6 +4151,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
[MLX5_QP_STATE_SQD] = {
[MLX5_QP_STATE_RST] = MLX5_CMD_OP_2RST_QP,
[MLX5_QP_STATE_ERR] = MLX5_CMD_OP_2ERR_QP,
+ [MLX5_QP_STATE_RTS] = MLX5_CMD_OP_SQD_RTS_QP,
},
[MLX5_QP_STATE_SQER] = {
[MLX5_QP_STATE_RST] = MLX5_CMD_OP_2RST_QP,
@@ -3252,66 +4168,60 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
struct mlx5_ib_qp *qp = to_mqp(ibqp);
struct mlx5_ib_qp_base *base = &qp->trans_qp.base;
struct mlx5_ib_cq *send_cq, *recv_cq;
- struct mlx5_qp_context *context;
struct mlx5_ib_pd *pd;
- struct mlx5_ib_port *mibport = NULL;
enum mlx5_qp_state mlx5_cur, mlx5_new;
- enum mlx5_qp_optpar optpar;
+ void *qpc, *pri_path, *alt_path;
+ enum mlx5_qp_optpar optpar = 0;
+ u32 set_id = 0;
int mlx5_st;
int err;
u16 op;
u8 tx_affinity = 0;
- mlx5_st = to_mlx5_st(ibqp->qp_type == IB_QPT_DRIVER ?
- qp->qp_sub_type : ibqp->qp_type);
+ mlx5_st = to_mlx5_st(qp->type);
if (mlx5_st < 0)
return -EINVAL;
- context = kzalloc(sizeof(*context), GFP_KERNEL);
- if (!context)
+ qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL);
+ if (!qpc)
return -ENOMEM;
- pd = get_pd(qp);
- context->flags = cpu_to_be32(mlx5_st << 16);
+ pd = to_mpd(qp->ibqp.pd);
+ MLX5_SET(qpc, qpc, st, mlx5_st);
if (!(attr_mask & IB_QP_PATH_MIG_STATE)) {
- context->flags |= cpu_to_be32(MLX5_QP_PM_MIGRATED << 11);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
} else {
switch (attr->path_mig_state) {
case IB_MIG_MIGRATED:
- context->flags |= cpu_to_be32(MLX5_QP_PM_MIGRATED << 11);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
break;
case IB_MIG_REARM:
- context->flags |= cpu_to_be32(MLX5_QP_PM_REARM << 11);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_REARM);
break;
case IB_MIG_ARMED:
- context->flags |= cpu_to_be32(MLX5_QP_PM_ARMED << 11);
+ MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_ARMED);
break;
}
}
- if ((cur_state == IB_QPS_RESET) && (new_state == IB_QPS_INIT)) {
- if ((ibqp->qp_type == IB_QPT_RC) ||
- (ibqp->qp_type == IB_QPT_UD &&
- !(qp->flags & MLX5_IB_QP_SQPN_QP1)) ||
- (ibqp->qp_type == IB_QPT_UC) ||
- (ibqp->qp_type == IB_QPT_RAW_PACKET) ||
- (ibqp->qp_type == IB_QPT_XRC_INI) ||
- (ibqp->qp_type == IB_QPT_XRC_TGT)) {
- if (dev->lag_active) {
- u8 p = mlx5_core_native_port_num(dev->mdev);
- tx_affinity = get_tx_affinity(dev, pd, base, p);
- context->flags |= cpu_to_be32(tx_affinity << 24);
- }
- }
- }
+ tx_affinity = get_tx_affinity(ibqp, attr, attr_mask,
+ cur_state == IB_QPS_RESET &&
+ new_state == IB_QPS_INIT, udata);
- if (is_sqp(ibqp->qp_type)) {
- context->mtu_msgmax = (IB_MTU_256 << 5) | 8;
- } else if ((ibqp->qp_type == IB_QPT_UD &&
- !(qp->flags & MLX5_IB_QP_UNDERLAY)) ||
- ibqp->qp_type == MLX5_IB_QPT_REG_UMR) {
- context->mtu_msgmax = (IB_MTU_4096 << 5) | 12;
+ MLX5_SET(qpc, qpc, lag_tx_port_affinity, tx_affinity);
+ if (tx_affinity && new_state == IB_QPS_RTR &&
+ MLX5_CAP_GEN(dev->mdev, init2_lag_tx_port_affinity))
+ optpar |= MLX5_QP_OPTPAR_LAG_TX_AFF;
+
+ if (is_sqp(qp->type)) {
+ MLX5_SET(qpc, qpc, mtu, IB_MTU_256);
+ MLX5_SET(qpc, qpc, log_msg_max, 8);
+ } else if ((qp->type == IB_QPT_UD &&
+ !(qp->flags & IB_QP_CREATE_SOURCE_QPN)) ||
+ qp->type == MLX5_IB_QPT_REG_UMR) {
+ MLX5_SET(qpc, qpc, mtu, IB_MTU_4096);
+ MLX5_SET(qpc, qpc, log_msg_max, 12);
} else if (attr_mask & IB_QP_PATH_MTU) {
if (attr->path_mtu < IB_MTU_256 ||
attr->path_mtu > IB_MTU_4096) {
@@ -3319,114 +4229,125 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
err = -EINVAL;
goto out;
}
- context->mtu_msgmax = (attr->path_mtu << 5) |
- (u8)MLX5_CAP_GEN(dev->mdev, log_max_msg);
+ MLX5_SET(qpc, qpc, mtu, attr->path_mtu);
+ MLX5_SET(qpc, qpc, log_msg_max,
+ MLX5_CAP_GEN(dev->mdev, log_max_msg));
}
if (attr_mask & IB_QP_DEST_QPN)
- context->log_pg_sz_remote_qpn = cpu_to_be32(attr->dest_qp_num);
+ MLX5_SET(qpc, qpc, remote_qpn, attr->dest_qp_num);
+
+ pri_path = MLX5_ADDR_OF(qpc, qpc, primary_address_path);
+ alt_path = MLX5_ADDR_OF(qpc, qpc, secondary_address_path);
if (attr_mask & IB_QP_PKEY_INDEX)
- context->pri_path.pkey_index = cpu_to_be16(attr->pkey_index);
+ MLX5_SET(ads, pri_path, pkey_index, attr->pkey_index);
/* todo implement counter_index functionality */
- if (is_sqp(ibqp->qp_type))
- context->pri_path.port = qp->port;
+ if (dev->ib_dev.type == RDMA_DEVICE_TYPE_SMI && is_qp0(qp->type)) {
+ MLX5_SET(ads, pri_path, vhca_port_num,
+ smi_to_native_portnum(dev, qp->port));
+ if (cur_state == IB_QPS_INIT && new_state == IB_QPS_RTR)
+ MLX5_SET(ads, pri_path, plane_index, qp->port);
+ } else if (is_sqp(qp->type))
+ MLX5_SET(ads, pri_path, vhca_port_num, qp->port);
if (attr_mask & IB_QP_PORT)
- context->pri_path.port = attr->port_num;
+ MLX5_SET(ads, pri_path, vhca_port_num, attr->port_num);
if (attr_mask & IB_QP_AV) {
- err = mlx5_set_path(dev, qp, &attr->ah_attr, &context->pri_path,
- attr_mask & IB_QP_PORT ? attr->port_num : qp->port,
+ err = mlx5_set_path(dev, qp, &attr->ah_attr, pri_path,
+ attr_mask & IB_QP_PORT ? attr->port_num :
+ qp->port,
attr_mask, 0, attr, false);
if (err)
goto out;
}
if (attr_mask & IB_QP_TIMEOUT)
- context->pri_path.ackto_lt |= attr->timeout << 3;
+ MLX5_SET(ads, pri_path, ack_timeout, attr->timeout);
if (attr_mask & IB_QP_ALT_PATH) {
- err = mlx5_set_path(dev, qp, &attr->alt_ah_attr,
- &context->alt_path,
+ err = mlx5_set_path(dev, qp, &attr->alt_ah_attr, alt_path,
attr->alt_port_num,
- attr_mask | IB_QP_PKEY_INDEX | IB_QP_TIMEOUT,
+ attr_mask | IB_QP_PKEY_INDEX |
+ IB_QP_TIMEOUT,
0, attr, true);
if (err)
goto out;
}
- get_cqs(qp->ibqp.qp_type, qp->ibqp.send_cq, qp->ibqp.recv_cq,
+ get_cqs(qp->type, qp->ibqp.send_cq, qp->ibqp.recv_cq,
&send_cq, &recv_cq);
- context->flags_pd = cpu_to_be32(pd ? pd->pdn : to_mpd(dev->devr.p0)->pdn);
- context->cqn_send = send_cq ? cpu_to_be32(send_cq->mcq.cqn) : 0;
- context->cqn_recv = recv_cq ? cpu_to_be32(recv_cq->mcq.cqn) : 0;
- context->params1 = cpu_to_be32(MLX5_IB_ACK_REQ_FREQ << 28);
+ MLX5_SET(qpc, qpc, pd, pd ? pd->pdn : to_mpd(dev->devr.p0)->pdn);
+ if (send_cq)
+ MLX5_SET(qpc, qpc, cqn_snd, send_cq->mcq.cqn);
+ if (recv_cq)
+ MLX5_SET(qpc, qpc, cqn_rcv, recv_cq->mcq.cqn);
+
+ MLX5_SET(qpc, qpc, log_ack_req_freq, MLX5_IB_ACK_REQ_FREQ);
if (attr_mask & IB_QP_RNR_RETRY)
- context->params1 |= cpu_to_be32(attr->rnr_retry << 13);
+ MLX5_SET(qpc, qpc, rnr_retry, attr->rnr_retry);
if (attr_mask & IB_QP_RETRY_CNT)
- context->params1 |= cpu_to_be32(attr->retry_cnt << 16);
+ MLX5_SET(qpc, qpc, retry_count, attr->retry_cnt);
- if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC) {
- if (attr->max_rd_atomic)
- context->params1 |=
- cpu_to_be32(fls(attr->max_rd_atomic - 1) << 21);
- }
+ if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC && attr->max_rd_atomic)
+ MLX5_SET(qpc, qpc, log_sra_max, fls(attr->max_rd_atomic - 1));
if (attr_mask & IB_QP_SQ_PSN)
- context->next_send_psn = cpu_to_be32(attr->sq_psn);
+ MLX5_SET(qpc, qpc, next_send_psn, attr->sq_psn);
- if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC) {
- if (attr->max_dest_rd_atomic)
- context->params2 |=
- cpu_to_be32(fls(attr->max_dest_rd_atomic - 1) << 21);
- }
+ if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC && attr->max_dest_rd_atomic)
+ MLX5_SET(qpc, qpc, log_rra_max,
+ fls(attr->max_dest_rd_atomic - 1));
if (attr_mask & (IB_QP_ACCESS_FLAGS | IB_QP_MAX_DEST_RD_ATOMIC)) {
- __be32 access_flags = 0;
-
- err = to_mlx5_access_flags(qp, attr, attr_mask, &access_flags);
+ err = set_qpc_atomic_flags(qp, attr, attr_mask, qpc);
if (err)
goto out;
-
- context->params2 |= access_flags;
}
if (attr_mask & IB_QP_MIN_RNR_TIMER)
- context->rnr_nextrecvpsn |= cpu_to_be32(attr->min_rnr_timer << 24);
+ MLX5_SET(qpc, qpc, min_rnr_nak, attr->min_rnr_timer);
if (attr_mask & IB_QP_RQ_PSN)
- context->rnr_nextrecvpsn |= cpu_to_be32(attr->rq_psn);
+ MLX5_SET(qpc, qpc, next_rcv_psn, attr->rq_psn);
if (attr_mask & IB_QP_QKEY)
- context->qkey = cpu_to_be32(attr->qkey);
+ MLX5_SET(qpc, qpc, q_key, attr->qkey);
if (qp->rq.wqe_cnt && cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT)
- context->db_rec_addr = cpu_to_be64(qp->db.dma);
+ MLX5_SET64(qpc, qpc, dbr_addr, qp->db.dma);
if (cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT) {
u8 port_num = (attr_mask & IB_QP_PORT ? attr->port_num :
qp->port) - 1;
/* Underlay port should be used - index 0 function per port */
- if (qp->flags & MLX5_IB_QP_UNDERLAY)
+ if (qp->flags & IB_QP_CREATE_SOURCE_QPN)
port_num = 0;
- mibport = &dev->port[port_num];
- context->qp_counter_set_usr_page |=
- cpu_to_be32((u32)(mibport->cnts.set_id) << 24);
+ if (ibqp->counter)
+ set_id = ibqp->counter->id;
+ else
+ set_id = mlx5_ib_get_counters_id(dev, port_num);
+ MLX5_SET(qpc, qpc, counter_set_id, set_id);
}
if (!ibqp->uobject && cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT)
- context->sq_crq_size |= cpu_to_be16(1 << 4);
+ MLX5_SET(qpc, qpc, rlky, 1);
+
+ if (qp->flags & MLX5_IB_QP_CREATE_SQPN_QP1)
+ MLX5_SET(qpc, qpc, deth_sqpn, 1);
- if (qp->flags & MLX5_IB_QP_SQPN_QP1)
- context->deth_sqpn = cpu_to_be32(1);
+ if (qp->is_ooo_rq && cur_state == IB_QPS_INIT && new_state == IB_QPS_RTR) {
+ MLX5_SET(qpc, qpc, dp_ordering_1, 1);
+ MLX5_SET(qpc, qpc, dp_ordering_force, 1);
+ }
mlx5_cur = to_mlx5_state(cur_state);
mlx5_new = to_mlx5_state(new_state);
@@ -3438,19 +4359,22 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
}
op = optab[mlx5_cur][mlx5_new];
- optpar = ib_mask_to_mlx5_opt(attr_mask);
+ optpar |= ib_mask_to_mlx5_opt(attr_mask);
optpar &= opt_mask[mlx5_cur][mlx5_new][mlx5_st];
- if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) {
+ if (qp->type == IB_QPT_RAW_PACKET ||
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) {
struct mlx5_modify_raw_qp_param raw_qp_param = {};
raw_qp_param.operation = op;
if (cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT) {
- raw_qp_param.rq_q_ctr_id = mibport->cnts.set_id;
+ raw_qp_param.rq_q_ctr_id = set_id;
raw_qp_param.set_mask |= MLX5_RAW_QP_MOD_SET_RQ_Q_CTR_ID;
}
+ if (attr_mask & IB_QP_PORT)
+ raw_qp_param.port = attr->port_num;
+
if (attr_mask & IB_QP_RATE_LIMIT) {
raw_qp_param.rl.rate = attr->rate_limit;
@@ -3481,8 +4405,15 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
err = modify_raw_packet_qp(dev, qp, &raw_qp_param, tx_affinity);
} else {
- err = mlx5_core_qp_modify(dev->mdev, op, optpar, context,
- &base->mqp);
+ if (udata) {
+ /* For the kernel flows, the resp will stay zero */
+ resp->ece_options =
+ MLX5_CAP_GEN(dev->mdev, ece_support) ?
+ ucmd->ece_options : 0;
+ resp->response_length = sizeof(*resp);
+ }
+ err = mlx5_core_qp_modify(dev, op, optpar, qpc, &base->mqp,
+ &resp->ece_options);
}
if (err)
@@ -3504,7 +4435,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
* entries and reinitialize the QP.
*/
if (new_state == IB_QPS_RESET &&
- !ibqp->uobject && ibqp->qp_type != IB_QPT_XRC_TGT) {
+ !ibqp->uobject && qp->type != IB_QPT_XRC_TGT) {
mlx5_ib_cq_clean(recv_cq, base->mqp.qpn,
ibqp->srq ? to_msrq(ibqp->srq) : NULL);
if (send_cq != recv_cq)
@@ -3517,12 +4448,19 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
qp->sq.cur_post = 0;
if (qp->sq.wqe_cnt)
qp->sq.cur_edge = get_sq_edge(&qp->sq, 0);
+ qp->sq.last_poll = 0;
qp->db.db[MLX5_RCV_DBR] = 0;
qp->db.db[MLX5_SND_DBR] = 0;
}
+ if ((new_state == IB_QPS_RTS) && qp->counter_pending) {
+ err = __mlx5_ib_qp_set_counter(ibqp, ibqp->counter);
+ if (!err)
+ qp->counter_pending = 0;
+ }
+
out:
- kfree(context);
+ kfree(qpc);
return err;
}
@@ -3580,14 +4518,15 @@ static bool modify_dci_qp_is_ok(enum ib_qp_state cur_state, enum ib_qp_state new
* Other transitions and attributes are illegal
*/
static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
- int attr_mask, struct ib_udata *udata)
+ int attr_mask, struct mlx5_ib_modify_qp *ucmd,
+ struct ib_udata *udata)
{
struct mlx5_ib_qp *qp = to_mqp(ibqp);
struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
enum ib_qp_state cur_state, new_state;
- int err = 0;
int required = IB_QP_STATE;
void *dctc;
+ int err;
if (!(attr_mask & IB_QP_STATE))
return -EINVAL;
@@ -3596,13 +4535,24 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
new_state = attr->qp_state;
dctc = MLX5_ADDR_OF(create_dct_in, qp->dct.in, dct_context_entry);
+ if (MLX5_CAP_GEN(dev->mdev, ece_support) && ucmd->ece_options)
+ /*
+ * DCT doesn't initialize QP till modify command is executed,
+ * so we need to overwrite previously set ECE field if user
+ * provided any value except zero, which means not set/not
+ * valid.
+ */
+ MLX5_SET(dctc, dctc, ece, ucmd->ece_options);
+
if (cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT) {
+ u16 set_id;
+
required |= IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT;
if (!is_valid_mask(attr_mask, required, 0))
return -EINVAL;
if (attr->port_num == 0 ||
- attr->port_num > MLX5_CAP_GEN(dev->mdev, num_ports)) {
+ attr->port_num > dev->num_ports) {
mlx5_ib_dbg(dev, "invalid port number %d. number of ports is %d\n",
attr->port_num, dev->num_ports);
return -EINVAL;
@@ -3614,7 +4564,7 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
if (attr->qp_access_flags & IB_ACCESS_REMOTE_ATOMIC) {
int atomic_mode;
- atomic_mode = get_atomic_mode(dev, MLX5_IB_QPT_DCT);
+ atomic_mode = get_atomic_mode(dev, qp);
if (atomic_mode < 0)
return -EOPNOTSUPP;
@@ -3622,17 +4572,30 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
MLX5_SET(dctc, dctc, rae, 1);
}
MLX5_SET(dctc, dctc, pkey_index, attr->pkey_index);
- MLX5_SET(dctc, dctc, port, attr->port_num);
- MLX5_SET(dctc, dctc, counter_set_id, dev->port[attr->port_num - 1].cnts.set_id);
+ if (mlx5_lag_is_active(dev->mdev))
+ MLX5_SET(dctc, dctc, port,
+ get_tx_affinity_rr(dev, udata));
+ else
+ MLX5_SET(dctc, dctc, port, attr->port_num);
+ set_id = mlx5_ib_get_counters_id(dev, attr->port_num - 1);
+ MLX5_SET(dctc, dctc, counter_set_id, set_id);
+
+ qp->port = attr->port_num;
} else if (cur_state == IB_QPS_INIT && new_state == IB_QPS_RTR) {
struct mlx5_ib_modify_qp_resp resp = {};
- u32 min_resp_len = offsetof(typeof(resp), dctn) +
- sizeof(resp.dctn);
+ u32 out[MLX5_ST_SZ_DW(create_dct_out)] = {};
+ u32 min_resp_len = offsetofend(typeof(resp), dctn);
if (udata->outlen < min_resp_len)
return -EINVAL;
- resp.response_length = min_resp_len;
+ /*
+ * If we don't have enough space for the ECE options,
+ * simply indicate it with resp.response_length.
+ */
+ resp.response_length = (udata->outlen < sizeof(resp)) ?
+ min_resp_len :
+ sizeof(resp);
required |= IB_QP_MIN_RNR_TIMER | IB_QP_AV | IB_QP_PATH_MTU;
if (!is_valid_mask(attr_mask, required, 0))
@@ -3643,47 +4606,104 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
MLX5_SET(dctc, dctc, mtu, attr->path_mtu);
MLX5_SET(dctc, dctc, my_addr_index, attr->ah_attr.grh.sgid_index);
MLX5_SET(dctc, dctc, hop_limit, attr->ah_attr.grh.hop_limit);
+ if (attr->ah_attr.type == RDMA_AH_ATTR_TYPE_ROCE)
+ MLX5_SET(dctc, dctc, eth_prio, attr->ah_attr.sl & 0x7);
+ if (qp->is_ooo_rq) {
+ MLX5_SET(dctc, dctc, dp_ordering_1, 1);
+ MLX5_SET(dctc, dctc, dp_ordering_force, 1);
+ }
- err = mlx5_core_create_dct(dev->mdev, &qp->dct.mdct, qp->dct.in,
- MLX5_ST_SZ_BYTES(create_dct_in));
+ err = mlx5_core_create_dct(dev, &qp->dct.mdct, qp->dct.in,
+ MLX5_ST_SZ_BYTES(create_dct_in), out,
+ sizeof(out));
+ err = mlx5_cmd_check(dev->mdev, err, qp->dct.in, out);
if (err)
return err;
resp.dctn = qp->dct.mdct.mqp.qpn;
+ if (MLX5_CAP_GEN(dev->mdev, ece_support))
+ resp.ece_options = MLX5_GET(create_dct_out, out, ece);
err = ib_copy_to_udata(udata, &resp, resp.response_length);
if (err) {
- mlx5_core_destroy_dct(dev->mdev, &qp->dct.mdct);
+ mlx5_core_destroy_dct(dev, &qp->dct.mdct);
return err;
}
} else {
mlx5_ib_warn(dev, "Modify DCT: Invalid transition from %d to %d\n", cur_state, new_state);
return -EINVAL;
}
- if (err)
- qp->state = IB_QPS_ERR;
- else
- qp->state = new_state;
- return err;
+
+ qp->state = new_state;
+ return 0;
+}
+
+static bool mlx5_ib_modify_qp_allowed(struct mlx5_ib_dev *dev,
+ struct mlx5_ib_qp *qp)
+{
+ if (dev->profile != &raw_eth_profile)
+ return true;
+
+ if (qp->type == IB_QPT_RAW_PACKET || qp->type == MLX5_IB_QPT_REG_UMR)
+ return true;
+
+ return false;
+}
+
+static int validate_rd_atomic(struct mlx5_ib_dev *dev, struct ib_qp_attr *attr,
+ int attr_mask, enum ib_qp_type qp_type)
+{
+ int log_max_ra_res;
+ int log_max_ra_req;
+
+ if (qp_type == MLX5_IB_QPT_DCI) {
+ log_max_ra_res = 1 << MLX5_CAP_GEN(dev->mdev,
+ log_max_ra_res_dc);
+ log_max_ra_req = 1 << MLX5_CAP_GEN(dev->mdev,
+ log_max_ra_req_dc);
+ } else {
+ log_max_ra_res = 1 << MLX5_CAP_GEN(dev->mdev,
+ log_max_ra_res_qp);
+ log_max_ra_req = 1 << MLX5_CAP_GEN(dev->mdev,
+ log_max_ra_req_qp);
+ }
+
+ if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
+ attr->max_rd_atomic > log_max_ra_res) {
+ mlx5_ib_dbg(dev, "invalid max_rd_atomic value %d\n",
+ attr->max_rd_atomic);
+ return false;
+ }
+
+ if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
+ attr->max_dest_rd_atomic > log_max_ra_req) {
+ mlx5_ib_dbg(dev, "invalid max_dest_rd_atomic value %d\n",
+ attr->max_dest_rd_atomic);
+ return false;
+ }
+ return true;
}
int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
int attr_mask, struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
+ struct mlx5_ib_modify_qp_resp resp = {};
struct mlx5_ib_qp *qp = to_mqp(ibqp);
struct mlx5_ib_modify_qp ucmd = {};
enum ib_qp_type qp_type;
enum ib_qp_state cur_state, new_state;
- size_t required_cmd_sz;
int err = -EINVAL;
- int port;
+
+ if (!mlx5_ib_modify_qp_allowed(dev, qp))
+ return -EOPNOTSUPP;
+
+ if (attr_mask & ~(IB_QP_ATTR_STANDARD_BITS | IB_QP_RATE_LIMIT))
+ return -EOPNOTSUPP;
if (ibqp->rwq_ind_tbl)
return -ENOSYS;
if (udata && udata->inlen) {
- required_cmd_sz = offsetof(typeof(ucmd), reserved) +
- sizeof(ucmd.reserved);
- if (udata->inlen < required_cmd_sz)
+ if (udata->inlen < offsetofend(typeof(ucmd), ece_options))
return -EINVAL;
if (udata->inlen > sizeof(ucmd) &&
@@ -3695,35 +4715,32 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
min(udata->inlen, sizeof(ucmd))))
return -EFAULT;
- if (ucmd.comp_mask ||
- memchr_inv(&ucmd.reserved, 0, sizeof(ucmd.reserved)) ||
+ if (ucmd.comp_mask & ~MLX5_IB_MODIFY_QP_OOO_DP ||
memchr_inv(&ucmd.burst_info.reserved, 0,
sizeof(ucmd.burst_info.reserved)))
return -EOPNOTSUPP;
+
+ if (ucmd.comp_mask & MLX5_IB_MODIFY_QP_OOO_DP) {
+ if (!get_dp_ooo_cap(dev->mdev, qp->type))
+ return -EOPNOTSUPP;
+ qp->is_ooo_rq = 1;
+ }
}
- if (unlikely(ibqp->qp_type == IB_QPT_GSI))
+ if (qp->type == IB_QPT_GSI)
return mlx5_ib_gsi_modify_qp(ibqp, attr, attr_mask);
- if (ibqp->qp_type == IB_QPT_DRIVER)
- qp_type = qp->qp_sub_type;
- else
- qp_type = (unlikely(ibqp->qp_type == MLX5_IB_QPT_HW_GSI)) ?
- IB_QPT_GSI : ibqp->qp_type;
+ qp_type = (qp->type == MLX5_IB_QPT_HW_GSI) ? IB_QPT_GSI : qp->type;
if (qp_type == MLX5_IB_QPT_DCT)
- return mlx5_ib_modify_dct(ibqp, attr, attr_mask, udata);
+ return mlx5_ib_modify_dct(ibqp, attr, attr_mask, &ucmd, udata);
mutex_lock(&qp->mutex);
cur_state = attr_mask & IB_QP_CUR_STATE ? attr->cur_qp_state : qp->state;
new_state = attr_mask & IB_QP_STATE ? attr->qp_state : cur_state;
- if (!(cur_state == new_state && cur_state == IB_QPS_RESET)) {
- port = attr_mask & IB_QP_PORT ? attr->port_num : qp->port;
- }
-
- if (qp->flags & MLX5_IB_QP_UNDERLAY) {
+ if (qp->flags & IB_QP_CREATE_SOURCE_QPN) {
if (attr_mask & ~(IB_QP_STATE | IB_QP_CUR_STATE)) {
mlx5_ib_dbg(dev, "invalid attr_mask 0x%x when underlay QP is used\n",
attr_mask);
@@ -3734,7 +4751,7 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
!ib_modify_qp_is_ok(cur_state, new_state, qp_type,
attr_mask)) {
mlx5_ib_dbg(dev, "invalid QP state transition from %d to %d, qp_type %d, attr_mask 0x%x\n",
- cur_state, new_state, ibqp->qp_type, attr_mask);
+ cur_state, new_state, qp->type, attr_mask);
goto out;
} else if (qp_type == MLX5_IB_QPT_DCI &&
!modify_dci_qp_is_ok(cur_state, new_state, attr_mask)) {
@@ -3751,31 +4768,14 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
goto out;
}
- if (attr_mask & IB_QP_PKEY_INDEX) {
- port = attr_mask & IB_QP_PORT ? attr->port_num : qp->port;
- if (attr->pkey_index >=
- dev->mdev->port_caps[port - 1].pkey_table_len) {
- mlx5_ib_dbg(dev, "invalid pkey index %d\n",
- attr->pkey_index);
- goto out;
- }
- }
-
- if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
- attr->max_rd_atomic >
- (1 << MLX5_CAP_GEN(dev->mdev, log_max_ra_res_qp))) {
- mlx5_ib_dbg(dev, "invalid max_rd_atomic value %d\n",
- attr->max_rd_atomic);
+ if ((attr_mask & IB_QP_PKEY_INDEX) &&
+ attr->pkey_index >= dev->pkey_table_len) {
+ mlx5_ib_dbg(dev, "invalid pkey index %d\n", attr->pkey_index);
goto out;
}
- if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
- attr->max_dest_rd_atomic >
- (1 << MLX5_CAP_GEN(dev->mdev, log_max_ra_req_qp))) {
- mlx5_ib_dbg(dev, "invalid max_dest_rd_atomic value %d\n",
- attr->max_dest_rd_atomic);
+ if (!validate_rd_atomic(dev, attr, attr_mask, qp_type))
goto out;
- }
if (cur_state == new_state && cur_state == IB_QPS_RESET) {
err = 0;
@@ -3783,1348 +4783,19 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
}
err = __mlx5_ib_modify_qp(ibqp, attr, attr_mask, cur_state,
- new_state, &ucmd);
-
-out:
- mutex_unlock(&qp->mutex);
- return err;
-}
-
-static void _handle_post_send_edge(struct mlx5_ib_wq *sq, void **seg,
- u32 wqe_sz, void **cur_edge)
-{
- u32 idx;
-
- idx = (sq->cur_post + (wqe_sz >> 2)) & (sq->wqe_cnt - 1);
- *cur_edge = get_sq_edge(sq, idx);
-
- *seg = mlx5_frag_buf_get_wqe(&sq->fbc, idx);
-}
-
-/* handle_post_send_edge - Check if we get to SQ edge. If yes, update to the
- * next nearby edge and get new address translation for current WQE position.
- * @sq - SQ buffer.
- * @seg: Current WQE position (16B aligned).
- * @wqe_sz: Total current WQE size [16B].
- * @cur_edge: Updated current edge.
- */
-static inline void handle_post_send_edge(struct mlx5_ib_wq *sq, void **seg,
- u32 wqe_sz, void **cur_edge)
-{
- if (likely(*seg != *cur_edge))
- return;
-
- _handle_post_send_edge(sq, seg, wqe_sz, cur_edge);
-}
-
-/* memcpy_send_wqe - copy data from src to WQE and update the relevant WQ's
- * pointers. At the end @seg is aligned to 16B regardless the copied size.
- * @sq - SQ buffer.
- * @cur_edge: Updated current edge.
- * @seg: Current WQE position (16B aligned).
- * @wqe_sz: Total current WQE size [16B].
- * @src: Pointer to copy from.
- * @n: Number of bytes to copy.
- */
-static inline void memcpy_send_wqe(struct mlx5_ib_wq *sq, void **cur_edge,
- void **seg, u32 *wqe_sz, const void *src,
- size_t n)
-{
- while (likely(n)) {
- size_t leftlen = *cur_edge - *seg;
- size_t copysz = min_t(size_t, leftlen, n);
- size_t stride;
-
- memcpy(*seg, src, copysz);
-
- n -= copysz;
- src += copysz;
- stride = !n ? ALIGN(copysz, 16) : copysz;
- *seg += stride;
- *wqe_sz += stride >> 4;
- handle_post_send_edge(sq, seg, *wqe_sz, cur_edge);
- }
-}
-
-static int mlx5_wq_overflow(struct mlx5_ib_wq *wq, int nreq, struct ib_cq *ib_cq)
-{
- struct mlx5_ib_cq *cq;
- unsigned cur;
-
- cur = wq->head - wq->tail;
- if (likely(cur + nreq < wq->max_post))
- return 0;
-
- cq = to_mcq(ib_cq);
- spin_lock(&cq->lock);
- cur = wq->head - wq->tail;
- spin_unlock(&cq->lock);
-
- return cur + nreq >= wq->max_post;
-}
-
-static __always_inline void set_raddr_seg(struct mlx5_wqe_raddr_seg *rseg,
- u64 remote_addr, u32 rkey)
-{
- rseg->raddr = cpu_to_be64(remote_addr);
- rseg->rkey = cpu_to_be32(rkey);
- rseg->reserved = 0;
-}
-
-static void set_eth_seg(const struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
- void **seg, int *size, void **cur_edge)
-{
- struct mlx5_wqe_eth_seg *eseg = *seg;
-
- memset(eseg, 0, sizeof(struct mlx5_wqe_eth_seg));
-
- if (wr->send_flags & IB_SEND_IP_CSUM)
- eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM |
- MLX5_ETH_WQE_L4_CSUM;
-
- if (wr->opcode == IB_WR_LSO) {
- struct ib_ud_wr *ud_wr = container_of(wr, struct ib_ud_wr, wr);
- size_t left, copysz;
- void *pdata = ud_wr->header;
- size_t stride;
-
- left = ud_wr->hlen;
- eseg->mss = cpu_to_be16(ud_wr->mss);
- eseg->inline_hdr.sz = cpu_to_be16(left);
-
- /* memcpy_send_wqe should get a 16B align address. Hence, we
- * first copy up to the current edge and then, if needed,
- * fall-through to memcpy_send_wqe.
- */
- copysz = min_t(u64, *cur_edge - (void *)eseg->inline_hdr.start,
- left);
- memcpy(eseg->inline_hdr.start, pdata, copysz);
- stride = ALIGN(sizeof(struct mlx5_wqe_eth_seg) -
- sizeof(eseg->inline_hdr.start) + copysz, 16);
- *size += stride / 16;
- *seg += stride;
-
- if (copysz < left) {
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
- left -= copysz;
- pdata += copysz;
- memcpy_send_wqe(&qp->sq, cur_edge, seg, size, pdata,
- left);
- }
-
- return;
- }
-
- *seg += sizeof(struct mlx5_wqe_eth_seg);
- *size += sizeof(struct mlx5_wqe_eth_seg) / 16;
-}
-
-static void set_datagram_seg(struct mlx5_wqe_datagram_seg *dseg,
- const struct ib_send_wr *wr)
-{
- memcpy(&dseg->av, &to_mah(ud_wr(wr)->ah)->av, sizeof(struct mlx5_av));
- dseg->av.dqp_dct = cpu_to_be32(ud_wr(wr)->remote_qpn | MLX5_EXTENDED_UD_AV);
- dseg->av.key.qkey.qkey = cpu_to_be32(ud_wr(wr)->remote_qkey);
-}
-
-static void set_data_ptr_seg(struct mlx5_wqe_data_seg *dseg, struct ib_sge *sg)
-{
- dseg->byte_count = cpu_to_be32(sg->length);
- dseg->lkey = cpu_to_be32(sg->lkey);
- dseg->addr = cpu_to_be64(sg->addr);
-}
-
-static u64 get_xlt_octo(u64 bytes)
-{
- return ALIGN(bytes, MLX5_IB_UMR_XLT_ALIGNMENT) /
- MLX5_IB_UMR_OCTOWORD;
-}
-
-static __be64 frwr_mkey_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_LEN |
- MLX5_MKEY_MASK_PAGE_SIZE |
- MLX5_MKEY_MASK_START_ADDR |
- MLX5_MKEY_MASK_EN_RINVAL |
- MLX5_MKEY_MASK_KEY |
- MLX5_MKEY_MASK_LR |
- MLX5_MKEY_MASK_LW |
- MLX5_MKEY_MASK_RR |
- MLX5_MKEY_MASK_RW |
- MLX5_MKEY_MASK_A |
- MLX5_MKEY_MASK_SMALL_FENCE |
- MLX5_MKEY_MASK_FREE;
-
- return cpu_to_be64(result);
-}
-
-static __be64 sig_mkey_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_LEN |
- MLX5_MKEY_MASK_PAGE_SIZE |
- MLX5_MKEY_MASK_START_ADDR |
- MLX5_MKEY_MASK_EN_SIGERR |
- MLX5_MKEY_MASK_EN_RINVAL |
- MLX5_MKEY_MASK_KEY |
- MLX5_MKEY_MASK_LR |
- MLX5_MKEY_MASK_LW |
- MLX5_MKEY_MASK_RR |
- MLX5_MKEY_MASK_RW |
- MLX5_MKEY_MASK_SMALL_FENCE |
- MLX5_MKEY_MASK_FREE |
- MLX5_MKEY_MASK_BSF_EN;
-
- return cpu_to_be64(result);
-}
-
-static void set_reg_umr_seg(struct mlx5_wqe_umr_ctrl_seg *umr,
- struct mlx5_ib_mr *mr, bool umr_inline)
-{
- int size = mr->ndescs * mr->desc_size;
-
- memset(umr, 0, sizeof(*umr));
-
- umr->flags = MLX5_UMR_CHECK_NOT_FREE;
- if (umr_inline)
- umr->flags |= MLX5_UMR_INLINE;
- umr->xlt_octowords = cpu_to_be16(get_xlt_octo(size));
- umr->mkey_mask = frwr_mkey_mask();
-}
-
-static void set_linv_umr_seg(struct mlx5_wqe_umr_ctrl_seg *umr)
-{
- memset(umr, 0, sizeof(*umr));
- umr->mkey_mask = cpu_to_be64(MLX5_MKEY_MASK_FREE);
- umr->flags = MLX5_UMR_INLINE;
-}
-
-static __be64 get_umr_enable_mr_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_KEY |
- MLX5_MKEY_MASK_FREE;
-
- return cpu_to_be64(result);
-}
-
-static __be64 get_umr_disable_mr_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_FREE;
-
- return cpu_to_be64(result);
-}
-
-static __be64 get_umr_update_translation_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_LEN |
- MLX5_MKEY_MASK_PAGE_SIZE |
- MLX5_MKEY_MASK_START_ADDR;
-
- return cpu_to_be64(result);
-}
-
-static __be64 get_umr_update_access_mask(int atomic)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_LR |
- MLX5_MKEY_MASK_LW |
- MLX5_MKEY_MASK_RR |
- MLX5_MKEY_MASK_RW;
-
- if (atomic)
- result |= MLX5_MKEY_MASK_A;
-
- return cpu_to_be64(result);
-}
-
-static __be64 get_umr_update_pd_mask(void)
-{
- u64 result;
-
- result = MLX5_MKEY_MASK_PD;
-
- return cpu_to_be64(result);
-}
-
-static int umr_check_mkey_mask(struct mlx5_ib_dev *dev, u64 mask)
-{
- if ((mask & MLX5_MKEY_MASK_PAGE_SIZE &&
- MLX5_CAP_GEN(dev->mdev, umr_modify_entity_size_disabled)) ||
- (mask & MLX5_MKEY_MASK_A &&
- MLX5_CAP_GEN(dev->mdev, umr_modify_atomic_disabled)))
- return -EPERM;
- return 0;
-}
-
-static int set_reg_umr_segment(struct mlx5_ib_dev *dev,
- struct mlx5_wqe_umr_ctrl_seg *umr,
- const struct ib_send_wr *wr, int atomic)
-{
- const struct mlx5_umr_wr *umrwr = umr_wr(wr);
-
- memset(umr, 0, sizeof(*umr));
-
- if (wr->send_flags & MLX5_IB_SEND_UMR_FAIL_IF_FREE)
- umr->flags = MLX5_UMR_CHECK_FREE; /* fail if free */
- else
- umr->flags = MLX5_UMR_CHECK_NOT_FREE; /* fail if not free */
-
- umr->xlt_octowords = cpu_to_be16(get_xlt_octo(umrwr->xlt_size));
- if (wr->send_flags & MLX5_IB_SEND_UMR_UPDATE_XLT) {
- u64 offset = get_xlt_octo(umrwr->offset);
-
- umr->xlt_offset = cpu_to_be16(offset & 0xffff);
- umr->xlt_offset_47_16 = cpu_to_be32(offset >> 16);
- umr->flags |= MLX5_UMR_TRANSLATION_OFFSET_EN;
- }
- if (wr->send_flags & MLX5_IB_SEND_UMR_UPDATE_TRANSLATION)
- umr->mkey_mask |= get_umr_update_translation_mask();
- if (wr->send_flags & MLX5_IB_SEND_UMR_UPDATE_PD_ACCESS) {
- umr->mkey_mask |= get_umr_update_access_mask(atomic);
- umr->mkey_mask |= get_umr_update_pd_mask();
- }
- if (wr->send_flags & MLX5_IB_SEND_UMR_ENABLE_MR)
- umr->mkey_mask |= get_umr_enable_mr_mask();
- if (wr->send_flags & MLX5_IB_SEND_UMR_DISABLE_MR)
- umr->mkey_mask |= get_umr_disable_mr_mask();
-
- if (!wr->num_sge)
- umr->flags |= MLX5_UMR_INLINE;
-
- return umr_check_mkey_mask(dev, be64_to_cpu(umr->mkey_mask));
-}
-
-static u8 get_umr_flags(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 | MLX5_PERM_UMR_EN;
-}
-
-static void set_reg_mkey_seg(struct mlx5_mkey_seg *seg,
- struct mlx5_ib_mr *mr,
- u32 key, int access)
-{
- int ndescs = ALIGN(mr->ndescs, 8) >> 1;
-
- memset(seg, 0, sizeof(*seg));
-
- if (mr->access_mode == MLX5_MKC_ACCESS_MODE_MTT)
- seg->log2_page_size = ilog2(mr->ibmr.page_size);
- else if (mr->access_mode == MLX5_MKC_ACCESS_MODE_KLMS)
- /* KLMs take twice the size of MTTs */
- ndescs *= 2;
-
- seg->flags = get_umr_flags(access) | mr->access_mode;
- seg->qpn_mkey7_0 = cpu_to_be32((key & 0xff) | 0xffffff00);
- seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL);
- seg->start_addr = cpu_to_be64(mr->ibmr.iova);
- seg->len = cpu_to_be64(mr->ibmr.length);
- seg->xlt_oct_size = cpu_to_be32(ndescs);
-}
-
-static void set_linv_mkey_seg(struct mlx5_mkey_seg *seg)
-{
- memset(seg, 0, sizeof(*seg));
- seg->status = MLX5_MKEY_STATUS_FREE;
-}
-
-static void set_reg_mkey_segment(struct mlx5_mkey_seg *seg,
- const struct ib_send_wr *wr)
-{
- const struct mlx5_umr_wr *umrwr = umr_wr(wr);
-
- memset(seg, 0, sizeof(*seg));
- if (wr->send_flags & MLX5_IB_SEND_UMR_DISABLE_MR)
- seg->status = MLX5_MKEY_STATUS_FREE;
-
- seg->flags = convert_access(umrwr->access_flags);
- if (umrwr->pd)
- seg->flags_pd = cpu_to_be32(to_mpd(umrwr->pd)->pdn);
- if (wr->send_flags & MLX5_IB_SEND_UMR_UPDATE_TRANSLATION &&
- !umrwr->length)
- seg->flags_pd |= cpu_to_be32(MLX5_MKEY_LEN64);
-
- seg->start_addr = cpu_to_be64(umrwr->virt_addr);
- seg->len = cpu_to_be64(umrwr->length);
- seg->log2_page_size = umrwr->page_shift;
- seg->qpn_mkey7_0 = cpu_to_be32(0xffffff00 |
- mlx5_mkey_variant(umrwr->mkey));
-}
-
-static void set_reg_data_seg(struct mlx5_wqe_data_seg *dseg,
- struct mlx5_ib_mr *mr,
- struct mlx5_ib_pd *pd)
-{
- int bcount = mr->desc_size * mr->ndescs;
-
- dseg->addr = cpu_to_be64(mr->desc_map);
- dseg->byte_count = cpu_to_be32(ALIGN(bcount, 64));
- dseg->lkey = cpu_to_be32(pd->ibpd.local_dma_lkey);
-}
-
-static __be32 send_ieth(const struct ib_send_wr *wr)
-{
- switch (wr->opcode) {
- case IB_WR_SEND_WITH_IMM:
- case IB_WR_RDMA_WRITE_WITH_IMM:
- return wr->ex.imm_data;
-
- case IB_WR_SEND_WITH_INV:
- return cpu_to_be32(wr->ex.invalidate_rkey);
-
- default:
- return 0;
- }
-}
-
-static u8 calc_sig(void *wqe, int size)
-{
- u8 *p = wqe;
- u8 res = 0;
- int i;
-
- for (i = 0; i < size; i++)
- res ^= p[i];
-
- return ~res;
-}
-
-static u8 wq_sig(void *wqe)
-{
- return calc_sig(wqe, (*((u8 *)wqe + 8) & 0x3f) << 4);
-}
-
-static int set_data_inl_seg(struct mlx5_ib_qp *qp, const struct ib_send_wr *wr,
- void **wqe, int *wqe_sz, void **cur_edge)
-{
- struct mlx5_wqe_inline_seg *seg;
- size_t offset;
- int inl = 0;
- int i;
-
- seg = *wqe;
- *wqe += sizeof(*seg);
- offset = sizeof(*seg);
-
- for (i = 0; i < wr->num_sge; i++) {
- size_t len = wr->sg_list[i].length;
- void *addr = (void *)(unsigned long)(wr->sg_list[i].addr);
-
- inl += len;
-
- if (unlikely(inl > qp->max_inline_data))
- return -ENOMEM;
-
- while (likely(len)) {
- size_t leftlen;
- size_t copysz;
-
- handle_post_send_edge(&qp->sq, wqe,
- *wqe_sz + (offset >> 4),
- cur_edge);
-
- leftlen = *cur_edge - *wqe;
- copysz = min_t(size_t, leftlen, len);
-
- memcpy(*wqe, addr, copysz);
- len -= copysz;
- addr += copysz;
- *wqe += copysz;
- offset += copysz;
- }
- }
-
- seg->byte_count = cpu_to_be32(inl | MLX5_INLINE_SEG);
-
- *wqe_sz += ALIGN(inl + sizeof(seg->byte_count), 16) / 16;
-
- return 0;
-}
-
-static u16 prot_field_size(enum ib_signature_type type)
-{
- switch (type) {
- case IB_SIG_TYPE_T10_DIF:
- return MLX5_DIF_SIZE;
- default:
- return 0;
- }
-}
-
-static u8 bs_selector(int block_size)
-{
- switch (block_size) {
- case 512: return 0x1;
- case 520: return 0x2;
- case 4096: return 0x3;
- case 4160: return 0x4;
- case 1073741824: return 0x5;
- default: return 0;
- }
-}
-
-static void mlx5_fill_inl_bsf(struct ib_sig_domain *domain,
- struct mlx5_bsf_inl *inl)
-{
- /* Valid inline section and allow BSF refresh */
- inl->vld_refresh = cpu_to_be16(MLX5_BSF_INL_VALID |
- MLX5_BSF_REFRESH_DIF);
- inl->dif_apptag = cpu_to_be16(domain->sig.dif.app_tag);
- inl->dif_reftag = cpu_to_be32(domain->sig.dif.ref_tag);
- /* repeating block */
- inl->rp_inv_seed = MLX5_BSF_REPEAT_BLOCK;
- inl->sig_type = domain->sig.dif.bg_type == IB_T10DIF_CRC ?
- MLX5_DIF_CRC : MLX5_DIF_IPCS;
-
- if (domain->sig.dif.ref_remap)
- inl->dif_inc_ref_guard_check |= MLX5_BSF_INC_REFTAG;
-
- if (domain->sig.dif.app_escape) {
- if (domain->sig.dif.ref_escape)
- inl->dif_inc_ref_guard_check |= MLX5_BSF_APPREF_ESCAPE;
- else
- inl->dif_inc_ref_guard_check |= MLX5_BSF_APPTAG_ESCAPE;
- }
-
- inl->dif_app_bitmask_check =
- cpu_to_be16(domain->sig.dif.apptag_check_mask);
-}
-
-static int mlx5_set_bsf(struct ib_mr *sig_mr,
- struct ib_sig_attrs *sig_attrs,
- struct mlx5_bsf *bsf, u32 data_size)
-{
- struct mlx5_core_sig_ctx *msig = to_mmr(sig_mr)->sig;
- struct mlx5_bsf_basic *basic = &bsf->basic;
- struct ib_sig_domain *mem = &sig_attrs->mem;
- struct ib_sig_domain *wire = &sig_attrs->wire;
-
- memset(bsf, 0, sizeof(*bsf));
-
- /* Basic + Extended + Inline */
- basic->bsf_size_sbs = 1 << 7;
- /* Input domain check byte mask */
- basic->check_byte_mask = sig_attrs->check_mask;
- basic->raw_data_size = cpu_to_be32(data_size);
-
- /* Memory domain */
- switch (sig_attrs->mem.sig_type) {
- case IB_SIG_TYPE_NONE:
- break;
- case IB_SIG_TYPE_T10_DIF:
- basic->mem.bs_selector = bs_selector(mem->sig.dif.pi_interval);
- basic->m_bfs_psv = cpu_to_be32(msig->psv_memory.psv_idx);
- mlx5_fill_inl_bsf(mem, &bsf->m_inl);
- break;
- default:
- return -EINVAL;
- }
-
- /* Wire domain */
- switch (sig_attrs->wire.sig_type) {
- case IB_SIG_TYPE_NONE:
- break;
- case IB_SIG_TYPE_T10_DIF:
- if (mem->sig.dif.pi_interval == wire->sig.dif.pi_interval &&
- mem->sig_type == wire->sig_type) {
- /* Same block structure */
- basic->bsf_size_sbs |= 1 << 4;
- if (mem->sig.dif.bg_type == wire->sig.dif.bg_type)
- basic->wire.copy_byte_mask |= MLX5_CPY_GRD_MASK;
- if (mem->sig.dif.app_tag == wire->sig.dif.app_tag)
- basic->wire.copy_byte_mask |= MLX5_CPY_APP_MASK;
- if (mem->sig.dif.ref_tag == wire->sig.dif.ref_tag)
- basic->wire.copy_byte_mask |= MLX5_CPY_REF_MASK;
- } else
- basic->wire.bs_selector = bs_selector(wire->sig.dif.pi_interval);
-
- basic->w_bfs_psv = cpu_to_be32(msig->psv_wire.psv_idx);
- mlx5_fill_inl_bsf(wire, &bsf->w_inl);
- break;
- default:
- return -EINVAL;
- }
-
- return 0;
-}
-
-static int set_sig_data_segment(const struct ib_sig_handover_wr *wr,
- struct mlx5_ib_qp *qp, void **seg,
- int *size, void **cur_edge)
-{
- struct ib_sig_attrs *sig_attrs = wr->sig_attrs;
- struct ib_mr *sig_mr = wr->sig_mr;
- struct mlx5_bsf *bsf;
- u32 data_len = wr->wr.sg_list->length;
- u32 data_key = wr->wr.sg_list->lkey;
- u64 data_va = wr->wr.sg_list->addr;
- int ret;
- int wqe_size;
-
- if (!wr->prot ||
- (data_key == wr->prot->lkey &&
- data_va == wr->prot->addr &&
- data_len == wr->prot->length)) {
- /**
- * Source domain doesn't contain signature information
- * or data and protection are interleaved in memory.
- * So need construct:
- * ------------------
- * | data_klm |
- * ------------------
- * | BSF |
- * ------------------
- **/
- struct mlx5_klm *data_klm = *seg;
-
- data_klm->bcount = cpu_to_be32(data_len);
- data_klm->key = cpu_to_be32(data_key);
- data_klm->va = cpu_to_be64(data_va);
- wqe_size = ALIGN(sizeof(*data_klm), 64);
- } else {
- /**
- * Source domain contains signature information
- * So need construct a strided block format:
- * ---------------------------
- * | stride_block_ctrl |
- * ---------------------------
- * | data_klm |
- * ---------------------------
- * | prot_klm |
- * ---------------------------
- * | BSF |
- * ---------------------------
- **/
- struct mlx5_stride_block_ctrl_seg *sblock_ctrl;
- struct mlx5_stride_block_entry *data_sentry;
- struct mlx5_stride_block_entry *prot_sentry;
- u32 prot_key = wr->prot->lkey;
- u64 prot_va = wr->prot->addr;
- u16 block_size = sig_attrs->mem.sig.dif.pi_interval;
- int prot_size;
-
- sblock_ctrl = *seg;
- data_sentry = (void *)sblock_ctrl + sizeof(*sblock_ctrl);
- prot_sentry = (void *)data_sentry + sizeof(*data_sentry);
-
- prot_size = prot_field_size(sig_attrs->mem.sig_type);
- if (!prot_size) {
- pr_err("Bad block size given: %u\n", block_size);
- return -EINVAL;
- }
- sblock_ctrl->bcount_per_cycle = cpu_to_be32(block_size +
- prot_size);
- sblock_ctrl->op = cpu_to_be32(MLX5_STRIDE_BLOCK_OP);
- sblock_ctrl->repeat_count = cpu_to_be32(data_len / block_size);
- sblock_ctrl->num_entries = cpu_to_be16(2);
-
- data_sentry->bcount = cpu_to_be16(block_size);
- data_sentry->key = cpu_to_be32(data_key);
- data_sentry->va = cpu_to_be64(data_va);
- data_sentry->stride = cpu_to_be16(block_size);
-
- prot_sentry->bcount = cpu_to_be16(prot_size);
- prot_sentry->key = cpu_to_be32(prot_key);
- prot_sentry->va = cpu_to_be64(prot_va);
- prot_sentry->stride = cpu_to_be16(prot_size);
-
- wqe_size = ALIGN(sizeof(*sblock_ctrl) + sizeof(*data_sentry) +
- sizeof(*prot_sentry), 64);
- }
-
- *seg += wqe_size;
- *size += wqe_size / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- bsf = *seg;
- ret = mlx5_set_bsf(sig_mr, sig_attrs, bsf, data_len);
- if (ret)
- return -EINVAL;
-
- *seg += sizeof(*bsf);
- *size += sizeof(*bsf) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- return 0;
-}
-
-static void set_sig_mkey_segment(struct mlx5_mkey_seg *seg,
- const struct ib_sig_handover_wr *wr, u32 size,
- u32 length, u32 pdn)
-{
- struct ib_mr *sig_mr = wr->sig_mr;
- u32 sig_key = sig_mr->rkey;
- u8 sigerr = to_mmr(sig_mr)->sig->sigerr_count & 1;
-
- memset(seg, 0, sizeof(*seg));
-
- seg->flags = get_umr_flags(wr->access_flags) |
- MLX5_MKC_ACCESS_MODE_KLMS;
- seg->qpn_mkey7_0 = cpu_to_be32((sig_key & 0xff) | 0xffffff00);
- seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL | sigerr << 26 |
- MLX5_MKEY_BSF_EN | pdn);
- seg->len = cpu_to_be64(length);
- seg->xlt_oct_size = cpu_to_be32(get_xlt_octo(size));
- seg->bsfs_octo_size = cpu_to_be32(MLX5_MKEY_BSF_OCTO_SIZE);
-}
-
-static void set_sig_umr_segment(struct mlx5_wqe_umr_ctrl_seg *umr,
- u32 size)
-{
- memset(umr, 0, sizeof(*umr));
-
- umr->flags = MLX5_FLAGS_INLINE | MLX5_FLAGS_CHECK_FREE;
- umr->xlt_octowords = cpu_to_be16(get_xlt_octo(size));
- umr->bsf_octowords = cpu_to_be16(MLX5_MKEY_BSF_OCTO_SIZE);
- umr->mkey_mask = sig_mkey_mask();
-}
-
-
-static int set_sig_umr_wr(const struct ib_send_wr *send_wr,
- struct mlx5_ib_qp *qp, void **seg, int *size,
- void **cur_edge)
-{
- const struct ib_sig_handover_wr *wr = sig_handover_wr(send_wr);
- struct mlx5_ib_mr *sig_mr = to_mmr(wr->sig_mr);
- u32 pdn = get_pd(qp)->pdn;
- u32 xlt_size;
- int region_len, ret;
-
- if (unlikely(wr->wr.num_sge != 1) ||
- unlikely(wr->access_flags & IB_ACCESS_REMOTE_ATOMIC) ||
- unlikely(!sig_mr->sig) || unlikely(!qp->signature_en) ||
- unlikely(!sig_mr->sig->sig_status_checked))
- return -EINVAL;
-
- /* length of the protected region, data + protection */
- region_len = wr->wr.sg_list->length;
- if (wr->prot &&
- (wr->prot->lkey != wr->wr.sg_list->lkey ||
- wr->prot->addr != wr->wr.sg_list->addr ||
- wr->prot->length != wr->wr.sg_list->length))
- region_len += wr->prot->length;
-
- /**
- * KLM octoword size - if protection was provided
- * then we use strided block format (3 octowords),
- * else we use single KLM (1 octoword)
- **/
- xlt_size = wr->prot ? 0x30 : sizeof(struct mlx5_klm);
-
- set_sig_umr_segment(*seg, xlt_size);
- *seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
- *size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- set_sig_mkey_segment(*seg, wr, xlt_size, region_len, pdn);
- *seg += sizeof(struct mlx5_mkey_seg);
- *size += sizeof(struct mlx5_mkey_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- ret = set_sig_data_segment(wr, qp, seg, size, cur_edge);
- if (ret)
- return ret;
-
- sig_mr->sig->sig_status_checked = false;
- return 0;
-}
-
-static int set_psv_wr(struct ib_sig_domain *domain,
- u32 psv_idx, void **seg, int *size)
-{
- struct mlx5_seg_set_psv *psv_seg = *seg;
-
- memset(psv_seg, 0, sizeof(*psv_seg));
- psv_seg->psv_num = cpu_to_be32(psv_idx);
- switch (domain->sig_type) {
- case IB_SIG_TYPE_NONE:
- break;
- case IB_SIG_TYPE_T10_DIF:
- psv_seg->transient_sig = cpu_to_be32(domain->sig.dif.bg << 16 |
- domain->sig.dif.app_tag);
- psv_seg->ref_tag = cpu_to_be32(domain->sig.dif.ref_tag);
- break;
- default:
- pr_err("Bad signature type (%d) is given.\n",
- domain->sig_type);
- return -EINVAL;
- }
-
- *seg += sizeof(*psv_seg);
- *size += sizeof(*psv_seg) / 16;
-
- return 0;
-}
-
-static int set_reg_wr(struct mlx5_ib_qp *qp,
- const struct ib_reg_wr *wr,
- void **seg, int *size, void **cur_edge)
-{
- struct mlx5_ib_mr *mr = to_mmr(wr->mr);
- struct mlx5_ib_pd *pd = to_mpd(qp->ibqp.pd);
- size_t mr_list_size = mr->ndescs * mr->desc_size;
- bool umr_inline = mr_list_size <= MLX5_IB_SQ_UMR_INLINE_THRESHOLD;
-
- if (unlikely(wr->wr.send_flags & IB_SEND_INLINE)) {
- mlx5_ib_warn(to_mdev(qp->ibqp.device),
- "Invalid IB_SEND_INLINE send flag\n");
- return -EINVAL;
- }
-
- set_reg_umr_seg(*seg, mr, umr_inline);
- *seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
- *size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- set_reg_mkey_seg(*seg, mr, wr->key, wr->access);
- *seg += sizeof(struct mlx5_mkey_seg);
- *size += sizeof(struct mlx5_mkey_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-
- if (umr_inline) {
- memcpy_send_wqe(&qp->sq, cur_edge, seg, size, mr->descs,
- mr_list_size);
- *size = ALIGN(*size, MLX5_SEND_WQE_BB >> 4);
- } else {
- set_reg_data_seg(*seg, mr, pd);
- *seg += sizeof(struct mlx5_wqe_data_seg);
- *size += (sizeof(struct mlx5_wqe_data_seg) / 16);
- }
- return 0;
-}
-
-static void set_linv_wr(struct mlx5_ib_qp *qp, void **seg, int *size,
- void **cur_edge)
-{
- set_linv_umr_seg(*seg);
- *seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
- *size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
- set_linv_mkey_seg(*seg);
- *seg += sizeof(struct mlx5_mkey_seg);
- *size += sizeof(struct mlx5_mkey_seg) / 16;
- handle_post_send_edge(&qp->sq, seg, *size, cur_edge);
-}
-
-static void dump_wqe(struct mlx5_ib_qp *qp, u32 idx, int size_16)
-{
- __be32 *p = NULL;
- u32 tidx = idx;
- int i, j;
-
- pr_debug("dump WQE index %u:\n", idx);
- for (i = 0, j = 0; i < size_16 * 4; i += 4, j += 4) {
- if ((i & 0xf) == 0) {
- tidx = (tidx + 1) & (qp->sq.wqe_cnt - 1);
- p = mlx5_frag_buf_get_wqe(&qp->sq.fbc, tidx);
- pr_debug("WQBB at %p:\n", (void *)p);
- j = 0;
- }
- pr_debug("%08x %08x %08x %08x\n", be32_to_cpu(p[j]),
- be32_to_cpu(p[j + 1]), be32_to_cpu(p[j + 2]),
- be32_to_cpu(p[j + 3]));
- }
-}
-
-static int __begin_wqe(struct mlx5_ib_qp *qp, void **seg,
- struct mlx5_wqe_ctrl_seg **ctrl,
- const struct ib_send_wr *wr, unsigned int *idx,
- int *size, void **cur_edge, int nreq,
- bool send_signaled, bool solicited)
-{
- if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq)))
- return -ENOMEM;
-
- *idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1);
- *seg = mlx5_frag_buf_get_wqe(&qp->sq.fbc, *idx);
- *ctrl = *seg;
- *(uint32_t *)(*seg + 8) = 0;
- (*ctrl)->imm = send_ieth(wr);
- (*ctrl)->fm_ce_se = qp->sq_signal_bits |
- (send_signaled ? MLX5_WQE_CTRL_CQ_UPDATE : 0) |
- (solicited ? MLX5_WQE_CTRL_SOLICITED : 0);
-
- *seg += sizeof(**ctrl);
- *size = sizeof(**ctrl) / 16;
- *cur_edge = qp->sq.cur_edge;
-
- return 0;
-}
-
-static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
- struct mlx5_wqe_ctrl_seg **ctrl,
- const struct ib_send_wr *wr, unsigned *idx,
- int *size, void **cur_edge, int nreq)
-{
- return __begin_wqe(qp, seg, ctrl, wr, idx, size, cur_edge, nreq,
- wr->send_flags & IB_SEND_SIGNALED,
- wr->send_flags & IB_SEND_SOLICITED);
-}
-
-static void finish_wqe(struct mlx5_ib_qp *qp,
- struct mlx5_wqe_ctrl_seg *ctrl,
- void *seg, u8 size, void *cur_edge,
- unsigned int idx, u64 wr_id, int nreq, u8 fence,
- u32 mlx5_opcode)
-{
- u8 opmod = 0;
-
- ctrl->opmod_idx_opcode = cpu_to_be32(((u32)(qp->sq.cur_post) << 8) |
- mlx5_opcode | ((u32)opmod << 24));
- ctrl->qpn_ds = cpu_to_be32(size | (qp->trans_qp.base.mqp.qpn << 8));
- ctrl->fm_ce_se |= fence;
- if (unlikely(qp->wq_sig))
- ctrl->signature = wq_sig(ctrl);
-
- qp->sq.wrid[idx] = wr_id;
- qp->sq.w_list[idx].opcode = mlx5_opcode;
- qp->sq.wqe_head[idx] = qp->sq.head + nreq;
- qp->sq.cur_post += DIV_ROUND_UP(size * 16, MLX5_SEND_WQE_BB);
- qp->sq.w_list[idx].next = qp->sq.cur_post;
-
- /* We save the edge which was possibly updated during the WQE
- * construction, into SQ's cache.
- */
- seg = PTR_ALIGN(seg, MLX5_SEND_WQE_BB);
- qp->sq.cur_edge = (unlikely(seg == cur_edge)) ?
- get_sq_edge(&qp->sq, qp->sq.cur_post &
- (qp->sq.wqe_cnt - 1)) :
- cur_edge;
-}
-
-static int _mlx5_ib_post_send(struct ib_qp *ibqp, const struct ib_send_wr *wr,
- const struct ib_send_wr **bad_wr, bool drain)
-{
- struct mlx5_wqe_ctrl_seg *ctrl = NULL; /* compiler warning */
- struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
- struct mlx5_core_dev *mdev = dev->mdev;
- struct mlx5_ib_qp *qp;
- struct mlx5_ib_mr *mr;
- struct mlx5_wqe_xrc_seg *xrc;
- struct mlx5_bf *bf;
- void *cur_edge;
- int uninitialized_var(size);
- unsigned long flags;
- unsigned idx;
- int err = 0;
- int num_sge;
- void *seg;
- int nreq;
- int i;
- u8 next_fence = 0;
- u8 fence;
-
- if (unlikely(mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR &&
- !drain)) {
- *bad_wr = wr;
- return -EIO;
- }
-
- if (unlikely(ibqp->qp_type == IB_QPT_GSI))
- return mlx5_ib_gsi_post_send(ibqp, wr, bad_wr);
-
- qp = to_mqp(ibqp);
- bf = &qp->bf;
-
- spin_lock_irqsave(&qp->sq.lock, flags);
-
- for (nreq = 0; wr; nreq++, wr = wr->next) {
- if (unlikely(wr->opcode >= ARRAY_SIZE(mlx5_ib_opcode))) {
- mlx5_ib_warn(dev, "\n");
- err = -EINVAL;
- *bad_wr = wr;
- goto out;
- }
-
- num_sge = wr->num_sge;
- if (unlikely(num_sge > qp->sq.max_gs)) {
- mlx5_ib_warn(dev, "\n");
- err = -EINVAL;
- *bad_wr = wr;
- goto out;
- }
-
- err = begin_wqe(qp, &seg, &ctrl, wr, &idx, &size, &cur_edge,
- nreq);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- err = -ENOMEM;
- *bad_wr = wr;
- goto out;
- }
-
- if (wr->opcode == IB_WR_REG_MR) {
- fence = dev->umr_fence;
- next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
- } else {
- if (wr->send_flags & IB_SEND_FENCE) {
- if (qp->next_fence)
- fence = MLX5_FENCE_MODE_SMALL_AND_FENCE;
- else
- fence = MLX5_FENCE_MODE_FENCE;
- } else {
- fence = qp->next_fence;
- }
- }
-
- switch (ibqp->qp_type) {
- case IB_QPT_XRC_INI:
- xrc = seg;
- seg += sizeof(*xrc);
- size += sizeof(*xrc) / 16;
- /* fall through */
- case IB_QPT_RC:
- switch (wr->opcode) {
- case IB_WR_RDMA_READ:
- case IB_WR_RDMA_WRITE:
- case IB_WR_RDMA_WRITE_WITH_IMM:
- set_raddr_seg(seg, rdma_wr(wr)->remote_addr,
- rdma_wr(wr)->rkey);
- seg += sizeof(struct mlx5_wqe_raddr_seg);
- size += sizeof(struct mlx5_wqe_raddr_seg) / 16;
- break;
-
- case IB_WR_ATOMIC_CMP_AND_SWP:
- case IB_WR_ATOMIC_FETCH_AND_ADD:
- case IB_WR_MASKED_ATOMIC_CMP_AND_SWP:
- mlx5_ib_warn(dev, "Atomic operations are not supported yet\n");
- err = -ENOSYS;
- *bad_wr = wr;
- goto out;
-
- case IB_WR_LOCAL_INV:
- qp->sq.wr_data[idx] = IB_WR_LOCAL_INV;
- ctrl->imm = cpu_to_be32(wr->ex.invalidate_rkey);
- set_linv_wr(qp, &seg, &size, &cur_edge);
- num_sge = 0;
- break;
-
- case IB_WR_REG_MR:
- qp->sq.wr_data[idx] = IB_WR_REG_MR;
- ctrl->imm = cpu_to_be32(reg_wr(wr)->key);
- err = set_reg_wr(qp, reg_wr(wr), &seg, &size,
- &cur_edge);
- if (err) {
- *bad_wr = wr;
- goto out;
- }
- num_sge = 0;
- break;
-
- case IB_WR_REG_SIG_MR:
- qp->sq.wr_data[idx] = IB_WR_REG_SIG_MR;
- mr = to_mmr(sig_handover_wr(wr)->sig_mr);
-
- ctrl->imm = cpu_to_be32(mr->ibmr.rkey);
- err = set_sig_umr_wr(wr, qp, &seg, &size,
- &cur_edge);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- *bad_wr = wr;
- goto out;
- }
-
- finish_wqe(qp, ctrl, seg, size, cur_edge, idx,
- wr->wr_id, nreq, fence,
- MLX5_OPCODE_UMR);
- /*
- * SET_PSV WQEs are not signaled and solicited
- * on error
- */
- err = __begin_wqe(qp, &seg, &ctrl, wr, &idx,
- &size, &cur_edge, nreq, false,
- true);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- err = -ENOMEM;
- *bad_wr = wr;
- goto out;
- }
-
- err = set_psv_wr(&sig_handover_wr(wr)->sig_attrs->mem,
- mr->sig->psv_memory.psv_idx, &seg,
- &size);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- *bad_wr = wr;
- goto out;
- }
-
- finish_wqe(qp, ctrl, seg, size, cur_edge, idx,
- wr->wr_id, nreq, fence,
- MLX5_OPCODE_SET_PSV);
- err = __begin_wqe(qp, &seg, &ctrl, wr, &idx,
- &size, &cur_edge, nreq, false,
- true);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- err = -ENOMEM;
- *bad_wr = wr;
- goto out;
- }
-
- err = set_psv_wr(&sig_handover_wr(wr)->sig_attrs->wire,
- mr->sig->psv_wire.psv_idx, &seg,
- &size);
- if (err) {
- mlx5_ib_warn(dev, "\n");
- *bad_wr = wr;
- goto out;
- }
-
- finish_wqe(qp, ctrl, seg, size, cur_edge, idx,
- wr->wr_id, nreq, fence,
- MLX5_OPCODE_SET_PSV);
- qp->next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
- num_sge = 0;
- goto skip_psv;
-
- default:
- break;
- }
- break;
-
- case IB_QPT_UC:
- switch (wr->opcode) {
- case IB_WR_RDMA_WRITE:
- case IB_WR_RDMA_WRITE_WITH_IMM:
- set_raddr_seg(seg, rdma_wr(wr)->remote_addr,
- rdma_wr(wr)->rkey);
- seg += sizeof(struct mlx5_wqe_raddr_seg);
- size += sizeof(struct mlx5_wqe_raddr_seg) / 16;
- break;
-
- default:
- break;
- }
- break;
-
- case IB_QPT_SMI:
- if (unlikely(!mdev->port_caps[qp->port - 1].has_smi)) {
- mlx5_ib_warn(dev, "Send SMP MADs is not allowed\n");
- err = -EPERM;
- *bad_wr = wr;
- goto out;
- }
- /* fall through */
- case MLX5_IB_QPT_HW_GSI:
- set_datagram_seg(seg, wr);
- seg += sizeof(struct mlx5_wqe_datagram_seg);
- size += sizeof(struct mlx5_wqe_datagram_seg) / 16;
- handle_post_send_edge(&qp->sq, &seg, size, &cur_edge);
-
- break;
- case IB_QPT_UD:
- set_datagram_seg(seg, wr);
- seg += sizeof(struct mlx5_wqe_datagram_seg);
- size += sizeof(struct mlx5_wqe_datagram_seg) / 16;
- handle_post_send_edge(&qp->sq, &seg, size, &cur_edge);
-
- /* handle qp that supports ud offload */
- if (qp->flags & IB_QP_CREATE_IPOIB_UD_LSO) {
- struct mlx5_wqe_eth_pad *pad;
-
- pad = seg;
- memset(pad, 0, sizeof(struct mlx5_wqe_eth_pad));
- seg += sizeof(struct mlx5_wqe_eth_pad);
- size += sizeof(struct mlx5_wqe_eth_pad) / 16;
- set_eth_seg(wr, qp, &seg, &size, &cur_edge);
- handle_post_send_edge(&qp->sq, &seg, size,
- &cur_edge);
- }
- break;
- case MLX5_IB_QPT_REG_UMR:
- if (wr->opcode != MLX5_IB_WR_UMR) {
- err = -EINVAL;
- mlx5_ib_warn(dev, "bad opcode\n");
- goto out;
- }
- qp->sq.wr_data[idx] = MLX5_IB_WR_UMR;
- ctrl->imm = cpu_to_be32(umr_wr(wr)->mkey);
- err = set_reg_umr_segment(dev, seg, wr, !!(MLX5_CAP_GEN(mdev, atomic)));
- if (unlikely(err))
- goto out;
- seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
- size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
- handle_post_send_edge(&qp->sq, &seg, size, &cur_edge);
- set_reg_mkey_segment(seg, wr);
- seg += sizeof(struct mlx5_mkey_seg);
- size += sizeof(struct mlx5_mkey_seg) / 16;
- handle_post_send_edge(&qp->sq, &seg, size, &cur_edge);
- break;
-
- default:
- break;
- }
-
- if (wr->send_flags & IB_SEND_INLINE && num_sge) {
- err = set_data_inl_seg(qp, wr, &seg, &size, &cur_edge);
- if (unlikely(err)) {
- mlx5_ib_warn(dev, "\n");
- *bad_wr = wr;
- goto out;
- }
- } else {
- for (i = 0; i < num_sge; i++) {
- handle_post_send_edge(&qp->sq, &seg, size,
- &cur_edge);
- if (likely(wr->sg_list[i].length)) {
- set_data_ptr_seg
- ((struct mlx5_wqe_data_seg *)seg,
- wr->sg_list + i);
- size += sizeof(struct mlx5_wqe_data_seg) / 16;
- seg += sizeof(struct mlx5_wqe_data_seg);
- }
- }
- }
-
- qp->next_fence = next_fence;
- finish_wqe(qp, ctrl, seg, size, cur_edge, idx, wr->wr_id, nreq,
- fence, mlx5_ib_opcode[wr->opcode]);
-skip_psv:
- if (0)
- dump_wqe(qp, idx, size);
- }
-
-out:
- if (likely(nreq)) {
- qp->sq.head += nreq;
-
- /* Make sure that descriptors are written before
- * updating doorbell record and ringing the doorbell
- */
- wmb();
-
- qp->db.db[MLX5_SND_DBR] = cpu_to_be32(qp->sq.cur_post);
-
- /* Make sure doorbell record is visible to the HCA before
- * we hit doorbell */
- wmb();
-
- /* currently we support only regular doorbells */
- mlx5_write64((__be32 *)ctrl, bf->bfreg->map + bf->offset, NULL);
- /* Make sure doorbells don't leak out of SQ spinlock
- * and reach the HCA out of order.
- */
- mmiowb();
- bf->offset ^= bf->buf_size;
- }
-
- spin_unlock_irqrestore(&qp->sq.lock, flags);
-
- return err;
-}
-
-int mlx5_ib_post_send(struct ib_qp *ibqp, const struct ib_send_wr *wr,
- const struct ib_send_wr **bad_wr)
-{
- return _mlx5_ib_post_send(ibqp, wr, bad_wr, false);
-}
-
-static void set_sig_seg(struct mlx5_rwqe_sig *sig, int size)
-{
- sig->signature = calc_sig(sig, size);
-}
-
-static int _mlx5_ib_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *wr,
- const struct ib_recv_wr **bad_wr, bool drain)
-{
- struct mlx5_ib_qp *qp = to_mqp(ibqp);
- struct mlx5_wqe_data_seg *scat;
- struct mlx5_rwqe_sig *sig;
- struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
- struct mlx5_core_dev *mdev = dev->mdev;
- unsigned long flags;
- int err = 0;
- int nreq;
- int ind;
- int i;
-
- if (unlikely(mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR &&
- !drain)) {
- *bad_wr = wr;
- return -EIO;
- }
-
- if (unlikely(ibqp->qp_type == IB_QPT_GSI))
- return mlx5_ib_gsi_post_recv(ibqp, wr, bad_wr);
-
- spin_lock_irqsave(&qp->rq.lock, flags);
-
- ind = qp->rq.head & (qp->rq.wqe_cnt - 1);
-
- for (nreq = 0; wr; nreq++, wr = wr->next) {
- if (mlx5_wq_overflow(&qp->rq, nreq, qp->ibqp.recv_cq)) {
- err = -ENOMEM;
- *bad_wr = wr;
- goto out;
- }
-
- if (unlikely(wr->num_sge > qp->rq.max_gs)) {
- err = -EINVAL;
- *bad_wr = wr;
- goto out;
- }
-
- scat = mlx5_frag_buf_get_wqe(&qp->rq.fbc, ind);
- if (qp->wq_sig)
- scat++;
-
- for (i = 0; i < wr->num_sge; i++)
- set_data_ptr_seg(scat + i, wr->sg_list + i);
-
- if (i < qp->rq.max_gs) {
- scat[i].byte_count = 0;
- scat[i].lkey = cpu_to_be32(MLX5_INVALID_LKEY);
- scat[i].addr = 0;
- }
+ new_state, &ucmd, &resp, udata);
- if (qp->wq_sig) {
- sig = (struct mlx5_rwqe_sig *)scat;
- set_sig_seg(sig, (qp->rq.max_gs + 1) << 2);
- }
-
- qp->rq.wrid[ind] = wr->wr_id;
-
- ind = (ind + 1) & (qp->rq.wqe_cnt - 1);
- }
+ /* resp.response_length is set in ECE supported flows only */
+ if (!err && resp.response_length &&
+ udata->outlen >= resp.response_length)
+ /* Return -EFAULT to the user and expect him to destroy QP. */
+ err = ib_copy_to_udata(udata, &resp, resp.response_length);
out:
- if (likely(nreq)) {
- qp->rq.head += nreq;
-
- /* Make sure that descriptors are written before
- * doorbell record.
- */
- wmb();
-
- *qp->db.db = cpu_to_be32(qp->rq.head & 0xffff);
- }
-
- spin_unlock_irqrestore(&qp->rq.lock, flags);
-
+ mutex_unlock(&qp->mutex);
return err;
}
-int mlx5_ib_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *wr,
- const struct ib_recv_wr **bad_wr)
-{
- return _mlx5_ib_post_recv(ibqp, wr, bad_wr, false);
-}
-
static inline enum ib_qp_state to_ib_qp_state(enum mlx5_qp_state mlx5_state)
{
switch (mlx5_state) {
@@ -5150,48 +4821,34 @@ static inline enum ib_mig_state to_ib_mig_state(int mlx5_mig_state)
}
}
-static int to_ib_qp_access_flags(int mlx5_flags)
-{
- int ib_flags = 0;
-
- if (mlx5_flags & MLX5_QP_BIT_RRE)
- ib_flags |= IB_ACCESS_REMOTE_READ;
- if (mlx5_flags & MLX5_QP_BIT_RWE)
- ib_flags |= IB_ACCESS_REMOTE_WRITE;
- if (mlx5_flags & MLX5_QP_BIT_RAE)
- ib_flags |= IB_ACCESS_REMOTE_ATOMIC;
-
- return ib_flags;
-}
-
static void to_rdma_ah_attr(struct mlx5_ib_dev *ibdev,
- struct rdma_ah_attr *ah_attr,
- struct mlx5_qp_path *path)
+ struct rdma_ah_attr *ah_attr, void *path)
{
+ int port = MLX5_GET(ads, path, vhca_port_num);
+ int static_rate;
memset(ah_attr, 0, sizeof(*ah_attr));
- if (!path->port || path->port > ibdev->num_ports)
+ if (!port || port > ibdev->num_ports)
return;
- ah_attr->type = rdma_ah_find_type(&ibdev->ib_dev, path->port);
+ ah_attr->type = rdma_ah_find_type(&ibdev->ib_dev, port);
- rdma_ah_set_port_num(ah_attr, path->port);
- rdma_ah_set_sl(ah_attr, path->dci_cfi_prio_sl & 0xf);
+ rdma_ah_set_port_num(ah_attr, port);
+ rdma_ah_set_sl(ah_attr, MLX5_GET(ads, path, sl));
- rdma_ah_set_dlid(ah_attr, be16_to_cpu(path->rlid));
- rdma_ah_set_path_bits(ah_attr, path->grh_mlid & 0x7f);
- rdma_ah_set_static_rate(ah_attr,
- path->static_rate ? path->static_rate - 5 : 0);
- if (path->grh_mlid & (1 << 7)) {
- u32 tc_fl = be32_to_cpu(path->tclass_flowlabel);
+ rdma_ah_set_dlid(ah_attr, MLX5_GET(ads, path, rlid));
+ rdma_ah_set_path_bits(ah_attr, MLX5_GET(ads, path, mlid));
- rdma_ah_set_grh(ah_attr, NULL,
- tc_fl & 0xfffff,
- path->mgid_index,
- path->hop_limit,
- (tc_fl >> 20) & 0xff);
- rdma_ah_set_dgid_raw(ah_attr, path->rgid);
+ static_rate = MLX5_GET(ads, path, stat_rate);
+ rdma_ah_set_static_rate(ah_attr, mlx5_to_ib_rate_map(static_rate));
+ if (MLX5_GET(ads, path, grh) ||
+ ah_attr->type == RDMA_AH_ATTR_TYPE_ROCE) {
+ rdma_ah_set_grh(ah_attr, NULL, MLX5_GET(ads, path, flow_label),
+ MLX5_GET(ads, path, src_addr_index),
+ MLX5_GET(ads, path, hop_limit),
+ MLX5_GET(ads, path, tclass));
+ rdma_ah_set_dgid_raw(ah_attr, MLX5_ADDR_OF(ads, path, rgid_rip));
}
}
@@ -5248,7 +4905,7 @@ static int sqrq_state_to_qp_state(u8 sq_state, u8 rq_state,
[MLX5_SQ_STATE_NA] = IB_QPS_RESET,
},
[MLX5_RQC_STATE_RDY] = {
- [MLX5_SQC_STATE_RST] = MLX5_QP_STATE_BAD,
+ [MLX5_SQC_STATE_RST] = MLX5_QP_STATE,
[MLX5_SQC_STATE_RDY] = MLX5_QP_STATE,
[MLX5_SQC_STATE_ERR] = IB_QPS_SQE,
[MLX5_SQ_STATE_NA] = MLX5_QP_STATE,
@@ -5260,7 +4917,7 @@ static int sqrq_state_to_qp_state(u8 sq_state, u8 rq_state,
[MLX5_SQ_STATE_NA] = IB_QPS_ERR,
},
[MLX5_RQ_STATE_NA] = {
- [MLX5_SQC_STATE_RST] = IB_QPS_RESET,
+ [MLX5_SQC_STATE_RST] = MLX5_QP_STATE,
[MLX5_SQC_STATE_RDY] = MLX5_QP_STATE,
[MLX5_SQC_STATE_ERR] = MLX5_QP_STATE,
[MLX5_SQ_STATE_NA] = MLX5_QP_STATE_BAD,
@@ -5313,61 +4970,60 @@ static int query_qp_attr(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
struct ib_qp_attr *qp_attr)
{
int outlen = MLX5_ST_SZ_BYTES(query_qp_out);
- struct mlx5_qp_context *context;
- int mlx5_state;
+ void *qpc, *pri_path, *alt_path;
u32 *outb;
- int err = 0;
+ int err;
outb = kzalloc(outlen, GFP_KERNEL);
if (!outb)
return -ENOMEM;
- err = mlx5_core_qp_query(dev->mdev, &qp->trans_qp.base.mqp, outb,
- outlen);
+ err = mlx5_core_qp_query(dev, &qp->trans_qp.base.mqp, outb, outlen,
+ false);
if (err)
goto out;
- /* FIXME: use MLX5_GET rather than mlx5_qp_context manual struct */
- context = (struct mlx5_qp_context *)MLX5_ADDR_OF(query_qp_out, outb, qpc);
+ qpc = MLX5_ADDR_OF(query_qp_out, outb, qpc);
- mlx5_state = be32_to_cpu(context->flags) >> 28;
+ qp->state = to_ib_qp_state(MLX5_GET(qpc, qpc, state));
+ if (MLX5_GET(qpc, qpc, state) == MLX5_QP_STATE_SQ_DRAINING)
+ qp_attr->sq_draining = 1;
- qp->state = to_ib_qp_state(mlx5_state);
- qp_attr->path_mtu = context->mtu_msgmax >> 5;
- qp_attr->path_mig_state =
- to_ib_mig_state((be32_to_cpu(context->flags) >> 11) & 0x3);
- qp_attr->qkey = be32_to_cpu(context->qkey);
- qp_attr->rq_psn = be32_to_cpu(context->rnr_nextrecvpsn) & 0xffffff;
- qp_attr->sq_psn = be32_to_cpu(context->next_send_psn) & 0xffffff;
- qp_attr->dest_qp_num = be32_to_cpu(context->log_pg_sz_remote_qpn) & 0xffffff;
- qp_attr->qp_access_flags =
- to_ib_qp_access_flags(be32_to_cpu(context->params2));
+ qp_attr->path_mtu = MLX5_GET(qpc, qpc, mtu);
+ qp_attr->path_mig_state = to_ib_mig_state(MLX5_GET(qpc, qpc, pm_state));
+ qp_attr->qkey = MLX5_GET(qpc, qpc, q_key);
+ qp_attr->rq_psn = MLX5_GET(qpc, qpc, next_rcv_psn);
+ qp_attr->sq_psn = MLX5_GET(qpc, qpc, next_send_psn);
+ qp_attr->dest_qp_num = MLX5_GET(qpc, qpc, remote_qpn);
- if (qp->ibqp.qp_type == IB_QPT_RC || qp->ibqp.qp_type == IB_QPT_UC) {
- to_rdma_ah_attr(dev, &qp_attr->ah_attr, &context->pri_path);
- to_rdma_ah_attr(dev, &qp_attr->alt_ah_attr, &context->alt_path);
- qp_attr->alt_pkey_index =
- be16_to_cpu(context->alt_path.pkey_index);
- qp_attr->alt_port_num =
- rdma_ah_get_port_num(&qp_attr->alt_ah_attr);
- }
+ if (MLX5_GET(qpc, qpc, rre))
+ qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ;
+ if (MLX5_GET(qpc, qpc, rwe))
+ qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_WRITE;
+ if (MLX5_GET(qpc, qpc, rae))
+ qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_ATOMIC;
- qp_attr->pkey_index = be16_to_cpu(context->pri_path.pkey_index);
- qp_attr->port_num = context->pri_path.port;
+ qp_attr->max_rd_atomic = 1 << MLX5_GET(qpc, qpc, log_sra_max);
+ qp_attr->max_dest_rd_atomic = 1 << MLX5_GET(qpc, qpc, log_rra_max);
+ qp_attr->min_rnr_timer = MLX5_GET(qpc, qpc, min_rnr_nak);
+ qp_attr->retry_cnt = MLX5_GET(qpc, qpc, retry_count);
+ qp_attr->rnr_retry = MLX5_GET(qpc, qpc, rnr_retry);
- /* qp_attr->en_sqd_async_notify is only applicable in modify qp */
- qp_attr->sq_draining = mlx5_state == MLX5_QP_STATE_SQ_DRAINING;
+ pri_path = MLX5_ADDR_OF(qpc, qpc, primary_address_path);
+ alt_path = MLX5_ADDR_OF(qpc, qpc, secondary_address_path);
- qp_attr->max_rd_atomic = 1 << ((be32_to_cpu(context->params1) >> 21) & 0x7);
+ if (qp->type == IB_QPT_RC || qp->type == IB_QPT_UC ||
+ qp->type == IB_QPT_XRC_INI || qp->type == IB_QPT_XRC_TGT) {
+ to_rdma_ah_attr(dev, &qp_attr->ah_attr, pri_path);
+ to_rdma_ah_attr(dev, &qp_attr->alt_ah_attr, alt_path);
+ qp_attr->alt_pkey_index = MLX5_GET(ads, alt_path, pkey_index);
+ qp_attr->alt_port_num = MLX5_GET(ads, alt_path, vhca_port_num);
+ }
- qp_attr->max_dest_rd_atomic =
- 1 << ((be32_to_cpu(context->params2) >> 21) & 0x7);
- qp_attr->min_rnr_timer =
- (be32_to_cpu(context->rnr_nextrecvpsn) >> 24) & 0x1f;
- qp_attr->timeout = context->pri_path.ackto_lt >> 3;
- qp_attr->retry_cnt = (be32_to_cpu(context->params1) >> 16) & 0x7;
- qp_attr->rnr_retry = (be32_to_cpu(context->params1) >> 13) & 0x7;
- qp_attr->alt_timeout = context->alt_path.ackto_lt >> 3;
+ qp_attr->pkey_index = MLX5_GET(ads, pri_path, pkey_index);
+ qp_attr->port_num = MLX5_GET(ads, pri_path, vhca_port_num);
+ qp_attr->timeout = MLX5_GET(ads, pri_path, ack_timeout);
+ qp_attr->alt_timeout = MLX5_GET(ads, alt_path, ack_timeout);
out:
kfree(outb);
@@ -5401,7 +5057,7 @@ static int mlx5_ib_dct_query_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *mqp,
if (!out)
return -ENOMEM;
- err = mlx5_core_dct_query(dev->mdev, dct, out, outlen);
+ err = mlx5_core_dct_query(dev, dct, out, outlen);
if (err)
goto out;
@@ -5421,7 +5077,7 @@ static int mlx5_ib_dct_query_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *mqp,
}
if (qp_attr_mask & IB_QP_PORT)
- qp_attr->port_num = MLX5_GET(dctc, dctc, port);
+ qp_attr->port_num = mqp->port;
if (qp_attr_mask & IB_QP_MIN_RNR_TIMER)
qp_attr->min_rnr_timer = MLX5_GET(dctc, dctc, min_rnr_nak);
if (qp_attr_mask & IB_QP_AV) {
@@ -5450,7 +5106,7 @@ int mlx5_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
if (ibqp->rwq_ind_tbl)
return -ENOSYS;
- if (unlikely(ibqp->qp_type == IB_QPT_GSI))
+ if (qp->type == IB_QPT_GSI)
return mlx5_ib_gsi_query_qp(ibqp, qp_attr, qp_attr_mask,
qp_init_attr);
@@ -5458,14 +5114,14 @@ int mlx5_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
memset(qp_init_attr, 0, sizeof(*qp_init_attr));
memset(qp_attr, 0, sizeof(*qp_attr));
- if (unlikely(qp->qp_sub_type == MLX5_IB_QPT_DCT))
+ if (unlikely(qp->type == MLX5_IB_QPT_DCT))
return mlx5_ib_dct_query_qp(dev, qp, qp_attr,
qp_attr_mask, qp_init_attr);
mutex_lock(&qp->mutex);
- if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET ||
- qp->flags & MLX5_IB_QP_UNDERLAY) {
+ if (qp->type == IB_QPT_RAW_PACKET ||
+ qp->flags & IB_QP_CREATE_SOURCE_QPN) {
err = query_raw_packet_qp_state(dev, qp, &raw_packet_qp_state);
if (err)
goto out;
@@ -5491,7 +5147,7 @@ int mlx5_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
qp_attr->cap.max_send_sge = 0;
}
- qp_init_attr->qp_type = ibqp->qp_type;
+ qp_init_attr->qp_type = qp->type;
qp_init_attr->recv_cq = ibqp->recv_cq;
qp_init_attr->send_cq = ibqp->send_cq;
qp_init_attr->srq = ibqp->srq;
@@ -5499,18 +5155,7 @@ int mlx5_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
qp_init_attr->cap = qp_attr->cap;
- qp_init_attr->create_flags = 0;
- if (qp->flags & MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK)
- qp_init_attr->create_flags |= IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK;
-
- if (qp->flags & MLX5_IB_QP_CROSS_CHANNEL)
- qp_init_attr->create_flags |= IB_QP_CREATE_CROSS_CHANNEL;
- if (qp->flags & MLX5_IB_QP_MANAGED_SEND)
- qp_init_attr->create_flags |= IB_QP_CREATE_MANAGED_SEND;
- if (qp->flags & MLX5_IB_QP_MANAGED_RECV)
- qp_init_attr->create_flags |= IB_QP_CREATE_MANAGED_RECV;
- if (qp->flags & MLX5_IB_QP_SQPN_QP1)
- qp_init_attr->create_flags |= mlx5_ib_create_qp_sqpn_qp1();
+ qp_init_attr->create_flags = qp->flags;
qp_init_attr->sq_sig_type = qp->sq_signal_bits & MLX5_WQE_CTRL_CQ_UPDATE ?
IB_SIGNAL_ALL_WR : IB_SIGNAL_REQ_WR;
@@ -5520,42 +5165,23 @@ out:
return err;
}
-struct ib_xrcd *mlx5_ib_alloc_xrcd(struct ib_device *ibdev,
- struct ib_ucontext *context,
- struct ib_udata *udata)
+int mlx5_ib_alloc_xrcd(struct ib_xrcd *ibxrcd, struct ib_udata *udata)
{
- struct mlx5_ib_dev *dev = to_mdev(ibdev);
- struct mlx5_ib_xrcd *xrcd;
- int err;
+ struct mlx5_ib_dev *dev = to_mdev(ibxrcd->device);
+ struct mlx5_ib_xrcd *xrcd = to_mxrcd(ibxrcd);
if (!MLX5_CAP_GEN(dev->mdev, xrc))
- return ERR_PTR(-ENOSYS);
-
- xrcd = kmalloc(sizeof(*xrcd), GFP_KERNEL);
- if (!xrcd)
- return ERR_PTR(-ENOMEM);
-
- err = mlx5_cmd_xrcd_alloc(dev->mdev, &xrcd->xrcdn, 0);
- if (err) {
- kfree(xrcd);
- return ERR_PTR(-ENOMEM);
- }
+ return -EOPNOTSUPP;
- return &xrcd->ibxrcd;
+ return mlx5_cmd_xrcd_alloc(dev->mdev, &xrcd->xrcdn, 0);
}
-int mlx5_ib_dealloc_xrcd(struct ib_xrcd *xrcd)
+int mlx5_ib_dealloc_xrcd(struct ib_xrcd *xrcd, struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(xrcd->device);
u32 xrcdn = to_mxrcd(xrcd)->xrcdn;
- int err;
-
- err = mlx5_cmd_xrcd_dealloc(dev->mdev, xrcdn, 0);
- if (err)
- mlx5_ib_warn(dev, "failed to dealloc xrcdn 0x%x\n", xrcdn);
- kfree(xrcd);
- return 0;
+ return mlx5_cmd_xrcd_dealloc(dev->mdev, xrcdn, 0);
}
static void mlx5_ib_wq_event(struct mlx5_core_qp *core_qp, int type)
@@ -5588,7 +5214,7 @@ static int set_delay_drop(struct mlx5_ib_dev *dev)
if (dev->delay_drop.activate)
goto out;
- err = mlx5_core_set_delay_drop(dev->mdev, dev->delay_drop.timeout);
+ err = mlx5_core_set_delay_drop(dev, dev->delay_drop.timeout);
if (err)
goto out;
@@ -5607,6 +5233,7 @@ static int create_rq(struct mlx5_ib_rwq *rwq, struct ib_pd *pd,
struct mlx5_ib_dev *dev;
int has_net_offloads;
__be64 *rq_pas0;
+ int ts_format;
void *in;
void *rqc;
void *wq;
@@ -5615,6 +5242,10 @@ static int create_rq(struct mlx5_ib_rwq *rwq, struct ib_pd *pd,
dev = to_mdev(pd->device);
+ ts_format = get_rq_ts_format(dev, to_mcq(init_attr->cq));
+ if (ts_format < 0)
+ return ts_format;
+
inlen = MLX5_ST_SZ_BYTES(create_rq_in) + sizeof(u64) * rwq->rq_num_pas;
in = kvzalloc(inlen, GFP_KERNEL);
if (!in)
@@ -5624,6 +5255,7 @@ static int create_rq(struct mlx5_ib_rwq *rwq, struct ib_pd *pd,
rqc = MLX5_ADDR_OF(create_rq_in, in, ctx);
MLX5_SET(rqc, rqc, mem_rq_type,
MLX5_RQC_MEM_RQ_TYPE_MEMORY_RQ_INLINE);
+ MLX5_SET(rqc, rqc, ts_format, ts_format);
MLX5_SET(rqc, rqc, user_index, rwq->user_index);
MLX5_SET(rqc, rqc, cqn, to_mcq(init_attr->cq)->mcq.cqn);
MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RST);
@@ -5643,12 +5275,21 @@ static int create_rq(struct mlx5_ib_rwq *rwq, struct ib_pd *pd,
}
MLX5_SET(wq, wq, log_wq_stride, rwq->log_rq_stride);
if (rwq->create_flags & MLX5_IB_WQ_FLAGS_STRIDING_RQ) {
+ /*
+ * In Firmware number of strides in each WQE is:
+ * "512 * 2^single_wqe_log_num_of_strides"
+ * Values 3 to 8 are accepted as 10 to 15, 9 to 18 are
+ * accepted as 0 to 9
+ */
+ static const u8 fw_map[] = { 10, 11, 12, 13, 14, 15, 0, 1,
+ 2, 3, 4, 5, 6, 7, 8, 9 };
MLX5_SET(wq, wq, two_byte_shift_en, rwq->two_byte_shift_en);
MLX5_SET(wq, wq, log_wqe_stride_size,
rwq->single_stride_log_num_of_bytes -
MLX5_MIN_SINGLE_STRIDE_LOG_NUM_BYTES);
- MLX5_SET(wq, wq, log_wqe_num_of_strides, rwq->log_num_strides -
- MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES);
+ MLX5_SET(wq, wq, log_wqe_num_of_strides,
+ fw_map[rwq->log_num_strides -
+ MLX5_EXT_MIN_SINGLE_WQE_LOG_NUM_STRIDES]);
}
MLX5_SET(wq, wq, log_wq_sz, rwq->log_rq_size);
MLX5_SET(wq, wq, pd, to_mpd(pd)->pdn);
@@ -5684,14 +5325,14 @@ static int create_rq(struct mlx5_ib_rwq *rwq, struct ib_pd *pd,
MLX5_SET(rqc, rqc, delay_drop_en, 1);
}
rq_pas0 = (__be64 *)MLX5_ADDR_OF(wq, wq, pas);
- mlx5_ib_populate_pas(dev, rwq->umem, rwq->page_shift, rq_pas0, 0);
- err = mlx5_core_create_rq_tracked(dev->mdev, in, inlen, &rwq->core_qp);
+ mlx5_ib_populate_pas(rwq->umem, 1UL << rwq->page_shift, rq_pas0, 0);
+ err = mlx5_core_create_rq_tracked(dev, in, inlen, &rwq->core_qp);
if (!err && init_attr->create_flags & IB_WQ_FLAGS_DELAY_DROP) {
err = set_delay_drop(dev);
if (err) {
mlx5_ib_warn(dev, "Failed to enable delay drop err=%d\n",
err);
- mlx5_core_destroy_rq_tracked(dev->mdev, &rwq->core_qp);
+ mlx5_core_destroy_rq_tracked(dev, &rwq->core_qp);
} else {
rwq->create_flags |= MLX5_IB_WQ_FLAGS_DELAY_DROP;
}
@@ -5723,6 +5364,19 @@ static int set_user_rq_size(struct mlx5_ib_dev *dev,
return 0;
}
+static bool log_of_strides_valid(struct mlx5_ib_dev *dev, u32 log_num_strides)
+{
+ if ((log_num_strides > MLX5_MAX_SINGLE_WQE_LOG_NUM_STRIDES) ||
+ (log_num_strides < MLX5_EXT_MIN_SINGLE_WQE_LOG_NUM_STRIDES))
+ return false;
+
+ if (!MLX5_CAP_GEN(dev->mdev, ext_stride_num_range) &&
+ (log_num_strides < MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES))
+ return false;
+
+ return true;
+}
+
static int prepare_user_rq(struct ib_pd *pd,
struct ib_wq_init_attr *init_attr,
struct ib_udata *udata,
@@ -5733,8 +5387,8 @@ static int prepare_user_rq(struct ib_pd *pd,
int err;
size_t required_cmd_sz;
- required_cmd_sz = offsetof(typeof(ucmd), single_stride_log_num_of_bytes)
- + sizeof(ucmd.single_stride_log_num_of_bytes);
+ required_cmd_sz = offsetofend(struct mlx5_ib_create_wq,
+ single_stride_log_num_of_bytes);
if (udata->inlen < required_cmd_sz) {
mlx5_ib_dbg(dev, "invalid inlen\n");
return -EINVAL;
@@ -5770,14 +5424,16 @@ static int prepare_user_rq(struct ib_pd *pd,
MLX5_MAX_SINGLE_STRIDE_LOG_NUM_BYTES);
return -EINVAL;
}
- if ((ucmd.single_wqe_log_num_of_strides >
- MLX5_MAX_SINGLE_WQE_LOG_NUM_STRIDES) ||
- (ucmd.single_wqe_log_num_of_strides <
- MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES)) {
- mlx5_ib_dbg(dev, "Invalid log num strides (%u. Range is %u - %u)\n",
- ucmd.single_wqe_log_num_of_strides,
- MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES,
- MLX5_MAX_SINGLE_WQE_LOG_NUM_STRIDES);
+ if (!log_of_strides_valid(dev,
+ ucmd.single_wqe_log_num_of_strides)) {
+ mlx5_ib_dbg(
+ dev,
+ "Invalid log num strides (%u. Range is %u - %u)\n",
+ ucmd.single_wqe_log_num_of_strides,
+ MLX5_CAP_GEN(dev->mdev, ext_stride_num_range) ?
+ MLX5_EXT_MIN_SINGLE_WQE_LOG_NUM_STRIDES :
+ MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES,
+ MLX5_MAX_SINGLE_WQE_LOG_NUM_STRIDES);
return -EINVAL;
}
rwq->single_stride_log_num_of_bytes =
@@ -5793,7 +5449,7 @@ static int prepare_user_rq(struct ib_pd *pd,
return err;
}
- err = create_user_rq(dev, pd, rwq, &ucmd);
+ err = create_user_rq(dev, pd, udata, rwq, &ucmd);
if (err) {
mlx5_ib_dbg(dev, "err %d\n", err);
return err;
@@ -5816,10 +5472,14 @@ struct ib_wq *mlx5_ib_create_wq(struct ib_pd *pd,
if (!udata)
return ERR_PTR(-ENOSYS);
- min_resp_len = offsetof(typeof(resp), reserved) + sizeof(resp.reserved);
+ min_resp_len = offsetofend(struct mlx5_ib_create_wq_resp, reserved);
if (udata->outlen && udata->outlen < min_resp_len)
return ERR_PTR(-EINVAL);
+ if (!capable(CAP_SYS_RAWIO) &&
+ init_attr->create_flags & IB_WQ_FLAGS_DELAY_DROP)
+ return ERR_PTR(-EPERM);
+
dev = to_mdev(pd->device);
switch (init_attr->wq_type) {
case IB_WQT_RQ:
@@ -5842,8 +5502,8 @@ struct ib_wq *mlx5_ib_create_wq(struct ib_pd *pd,
rwq->ibwq.wq_num = rwq->core_qp.qpn;
rwq->ibwq.state = IB_WQS_RESET;
if (udata->outlen) {
- resp.response_length = offsetof(typeof(resp), response_length) +
- sizeof(resp.response_length);
+ resp.response_length = offsetofend(
+ struct mlx5_ib_create_wq_resp, response_length);
err = ib_copy_to_udata(udata, &resp, resp.response_length);
if (err)
goto err_copy;
@@ -5854,32 +5514,35 @@ struct ib_wq *mlx5_ib_create_wq(struct ib_pd *pd,
return &rwq->ibwq;
err_copy:
- mlx5_core_destroy_rq_tracked(dev->mdev, &rwq->core_qp);
+ mlx5_core_destroy_rq_tracked(dev, &rwq->core_qp);
err_user_rq:
- destroy_user_rq(dev, pd, rwq);
+ destroy_user_rq(dev, pd, rwq, udata);
err:
kfree(rwq);
return ERR_PTR(err);
}
-int mlx5_ib_destroy_wq(struct ib_wq *wq)
+int mlx5_ib_destroy_wq(struct ib_wq *wq, struct ib_udata *udata)
{
struct mlx5_ib_dev *dev = to_mdev(wq->device);
struct mlx5_ib_rwq *rwq = to_mrwq(wq);
+ int ret;
- mlx5_core_destroy_rq_tracked(dev->mdev, &rwq->core_qp);
- destroy_user_rq(dev, wq->pd, rwq);
+ ret = mlx5_core_destroy_rq_tracked(dev, &rwq->core_qp);
+ if (ret)
+ return ret;
+ destroy_user_rq(dev, wq->pd, rwq, udata);
kfree(rwq);
-
return 0;
}
-struct ib_rwq_ind_table *mlx5_ib_create_rwq_ind_table(struct ib_device *device,
- struct ib_rwq_ind_table_init_attr *init_attr,
- struct ib_udata *udata)
+int mlx5_ib_create_rwq_ind_table(struct ib_rwq_ind_table *ib_rwq_ind_table,
+ struct ib_rwq_ind_table_init_attr *init_attr,
+ struct ib_udata *udata)
{
- struct mlx5_ib_dev *dev = to_mdev(device);
- struct mlx5_ib_rwq_ind_table *rwq_ind_tbl;
+ struct mlx5_ib_rwq_ind_table *rwq_ind_tbl =
+ to_mrwq_ind_table(ib_rwq_ind_table);
+ struct mlx5_ib_dev *dev = to_mdev(ib_rwq_ind_table->device);
int sz = 1 << init_attr->log_ind_tbl_size;
struct mlx5_ib_create_rwq_ind_tbl_resp resp = {};
size_t min_resp_len;
@@ -5892,30 +5555,25 @@ struct ib_rwq_ind_table *mlx5_ib_create_rwq_ind_table(struct ib_device *device,
if (udata->inlen > 0 &&
!ib_is_udata_cleared(udata, 0,
udata->inlen))
- return ERR_PTR(-EOPNOTSUPP);
+ return -EOPNOTSUPP;
if (init_attr->log_ind_tbl_size >
MLX5_CAP_GEN(dev->mdev, log_max_rqt_size)) {
mlx5_ib_dbg(dev, "log_ind_tbl_size = %d is bigger than supported = %d\n",
init_attr->log_ind_tbl_size,
MLX5_CAP_GEN(dev->mdev, log_max_rqt_size));
- return ERR_PTR(-EINVAL);
+ return -EINVAL;
}
- min_resp_len = offsetof(typeof(resp), reserved) + sizeof(resp.reserved);
+ min_resp_len =
+ offsetofend(struct mlx5_ib_create_rwq_ind_tbl_resp, reserved);
if (udata->outlen && udata->outlen < min_resp_len)
- return ERR_PTR(-EINVAL);
-
- rwq_ind_tbl = kzalloc(sizeof(*rwq_ind_tbl), GFP_KERNEL);
- if (!rwq_ind_tbl)
- return ERR_PTR(-ENOMEM);
+ return -EINVAL;
inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz;
in = kvzalloc(inlen, GFP_KERNEL);
- if (!in) {
- err = -ENOMEM;
- goto err;
- }
+ if (!in)
+ return -ENOMEM;
rqtc = MLX5_ADDR_OF(create_rqt_in, in, rqt_context);
@@ -5930,26 +5588,24 @@ struct ib_rwq_ind_table *mlx5_ib_create_rwq_ind_table(struct ib_device *device,
err = mlx5_core_create_rqt(dev->mdev, in, inlen, &rwq_ind_tbl->rqtn);
kvfree(in);
-
if (err)
- goto err;
+ return err;
rwq_ind_tbl->ib_rwq_ind_tbl.ind_tbl_num = rwq_ind_tbl->rqtn;
if (udata->outlen) {
- resp.response_length = offsetof(typeof(resp), response_length) +
- sizeof(resp.response_length);
+ resp.response_length =
+ offsetofend(struct mlx5_ib_create_rwq_ind_tbl_resp,
+ response_length);
err = ib_copy_to_udata(udata, &resp, resp.response_length);
if (err)
goto err_copy;
}
- return &rwq_ind_tbl->ib_rwq_ind_tbl;
+ return 0;
err_copy:
mlx5_cmd_destroy_rqt(dev->mdev, rwq_ind_tbl->rqtn, rwq_ind_tbl->uid);
-err:
- kfree(rwq_ind_tbl);
- return ERR_PTR(err);
+ return err;
}
int mlx5_ib_destroy_rwq_ind_table(struct ib_rwq_ind_table *ib_rwq_ind_tbl)
@@ -5957,10 +5613,7 @@ int mlx5_ib_destroy_rwq_ind_table(struct ib_rwq_ind_table *ib_rwq_ind_tbl)
struct mlx5_ib_rwq_ind_table *rwq_ind_tbl = to_mrwq_ind_table(ib_rwq_ind_tbl);
struct mlx5_ib_dev *dev = to_mdev(ib_rwq_ind_tbl->device);
- mlx5_cmd_destroy_rqt(dev->mdev, rwq_ind_tbl->rqtn, rwq_ind_tbl->uid);
-
- kfree(rwq_ind_tbl);
- return 0;
+ return mlx5_cmd_destroy_rqt(dev->mdev, rwq_ind_tbl->rqtn, rwq_ind_tbl->uid);
}
int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
@@ -5977,7 +5630,7 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
void *rqc;
void *in;
- required_cmd_sz = offsetof(typeof(ucmd), reserved) + sizeof(ucmd.reserved);
+ required_cmd_sz = offsetofend(struct mlx5_ib_modify_wq, reserved);
if (udata->inlen < required_cmd_sz)
return -EINVAL;
@@ -5999,10 +5652,8 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
- curr_wq_state = (wq_attr_mask & IB_WQ_CUR_STATE) ?
- wq_attr->curr_wq_state : wq->state;
- wq_state = (wq_attr_mask & IB_WQ_STATE) ?
- wq_attr->wq_state : curr_wq_state;
+ curr_wq_state = wq_attr->curr_wq_state;
+ wq_state = wq_attr->wq_state;
if (curr_wq_state == IB_WQS_ERR)
curr_wq_state = MLX5_RQC_STATE_ERR;
if (wq_state == IB_WQS_ERR)
@@ -6015,8 +5666,7 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
if (wq_attr->flags_mask & IB_WQ_FLAGS_CVLAN_STRIPPING) {
if (!(MLX5_CAP_GEN(dev->mdev, eth_net_offloads) &&
MLX5_CAP_ETH(dev->mdev, vlan_cap))) {
- mlx5_ib_dbg(dev, "VLAN offloads are not "
- "supported\n");
+ mlx5_ib_dbg(dev, "VLAN offloads are not supported\n");
err = -EOPNOTSUPP;
goto out;
}
@@ -6034,18 +5684,20 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
}
if (curr_wq_state == IB_WQS_RESET && wq_state == IB_WQS_RDY) {
+ u16 set_id;
+
+ set_id = mlx5_ib_get_counters_id(dev, 0);
if (MLX5_CAP_GEN(dev->mdev, modify_rq_counter_set_id)) {
MLX5_SET64(modify_rq_in, in, modify_bitmask,
MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_RQ_COUNTER_SET_ID);
- MLX5_SET(rqc, rqc, counter_set_id,
- dev->port->cnts.set_id);
+ MLX5_SET(rqc, rqc, counter_set_id, set_id);
} else
dev_info_once(
&dev->ib_dev.dev,
"Receive WQ counters are not supported on current FW\n");
}
- err = mlx5_core_modify_rq(dev->mdev, rwq->core_qp.qpn, in, inlen);
+ err = mlx5_core_modify_rq(dev->mdev, rwq->core_qp.qpn, in);
if (!err)
rwq->ibwq.state = (wq_state == MLX5_RQC_STATE_ERR) ? IB_WQS_ERR : wq_state;
@@ -6112,7 +5764,7 @@ static void handle_drain_completion(struct ib_cq *cq,
/* Run the CQ handler - this makes sure that the drain WR will
* be processed if wasn't processed yet.
*/
- mcq->mcq.comp(&mcq->mcq);
+ mcq->mcq.comp(&mcq->mcq, NULL);
}
wait_for_completion(&sdrain->done);
@@ -6144,7 +5796,7 @@ void mlx5_ib_drain_sq(struct ib_qp *qp)
sdrain.cqe.done = mlx5_ib_drain_qp_done;
init_completion(&sdrain.done);
- ret = _mlx5_ib_post_send(qp, &swr.wr, &bad_swr, true);
+ ret = mlx5_ib_post_send_drain(qp, &swr.wr, &bad_swr);
if (ret) {
WARN_ONCE(ret, "failed to drain send queue: %d\n", ret);
return;
@@ -6174,7 +5826,7 @@ void mlx5_ib_drain_rq(struct ib_qp *qp)
rdrain.cqe.done = mlx5_ib_drain_qp_done;
init_completion(&rdrain.done);
- ret = _mlx5_ib_post_recv(qp, &rwr, &bad_rwr, true);
+ ret = mlx5_ib_post_recv_drain(qp, &rwr, &bad_rwr);
if (ret) {
WARN_ONCE(ret, "failed to drain recv queue: %d\n", ret);
return;
@@ -6182,3 +5834,54 @@ void mlx5_ib_drain_rq(struct ib_qp *qp)
handle_drain_completion(cq, &rdrain, dev);
}
+
+/*
+ * Bind a qp to a counter. If @counter is NULL then bind the qp to
+ * the default counter
+ */
+int mlx5_ib_qp_set_counter(struct ib_qp *qp, struct rdma_counter *counter)
+{
+ struct mlx5_ib_dev *dev = to_mdev(qp->device);
+ struct mlx5_ib_qp *mqp = to_mqp(qp);
+ int err = 0;
+
+ mutex_lock(&mqp->mutex);
+ if (mqp->state == IB_QPS_RESET) {
+ qp->counter = counter;
+ goto out;
+ }
+
+ if (!MLX5_CAP_GEN(dev->mdev, rts2rts_qp_counters_set_id)) {
+ err = -EOPNOTSUPP;
+ goto out;
+ }
+
+ if (mqp->state == IB_QPS_RTS) {
+ err = __mlx5_ib_qp_set_counter(qp, counter);
+ if (!err)
+ qp->counter = counter;
+
+ goto out;
+ }
+
+ mqp->counter_pending = 1;
+ qp->counter = counter;
+
+out:
+ mutex_unlock(&mqp->mutex);
+ return err;
+}
+
+int mlx5_ib_qp_event_init(void)
+{
+ mlx5_ib_qp_event_wq = alloc_ordered_workqueue("mlx5_ib_qp_event_wq", 0);
+ if (!mlx5_ib_qp_event_wq)
+ return -ENOMEM;
+
+ return 0;
+}
+
+void mlx5_ib_qp_event_cleanup(void)
+{
+ destroy_workqueue(mlx5_ib_qp_event_wq);
+}
generated by cgit (git 2.34.1) at 2025-12-24 23:33:51 +0000