diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2022-12-14 09:27:13 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2022-12-14 09:27:13 -0800 |
commit | ab425febda94c7d287ea3433cbd0971771d6aeb4 (patch) | |
tree | 66240fae1e9720214afda604635c3f0da0b9ebfa /drivers/infiniband/hw | |
parent | 08cdc2157966c07d3f986a097ddaa74cee312751 (diff) | |
parent | dbc94a0fb81771a38733c0e8f2ea8c4fa6934dc1 (diff) | |
download | linux-ab425febda94c7d287ea3433cbd0971771d6aeb4.tar.gz linux-ab425febda94c7d287ea3433cbd0971771d6aeb4.tar.bz2 linux-ab425febda94c7d287ea3433cbd0971771d6aeb4.zip |
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma updates from Jason Gunthorpe:
"Usual size of updates, a new driver, and most of the bulk focusing on
rxe:
- Usual typos, style, and language updates
- Driver updates for mlx5, irdma, siw, rts, srp, hfi1, hns, erdma,
mlx4, srp
- Lots of RXE updates:
* Improve reply error handling for bad MR operations
* Code tidying
* Debug printing uses common loggers
* Remove half implemented RD related stuff
* Support IBA's recently defined Atomic Write and Flush operations
- erdma support for atomic operations
- New driver 'mana' for Ethernet HW available in Azure VMs. This
driver only supports DPDK"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (122 commits)
IB/IPoIB: Fix queue count inconsistency for PKEY child interfaces
RDMA: Add missed netdev_put() for the netdevice_tracker
RDMA/rxe: Enable RDMA FLUSH capability for rxe device
RDMA/cm: Make QP FLUSHABLE for supported device
RDMA/rxe: Implement flush completion
RDMA/rxe: Implement flush execution in responder side
RDMA/rxe: Implement RC RDMA FLUSH service in requester side
RDMA/rxe: Extend rxe packet format to support flush
RDMA/rxe: Allow registering persistent flag for pmem MR only
RDMA/rxe: Extend rxe user ABI to support flush
RDMA: Extend RDMA kernel verbs ABI to support flush
RDMA: Extend RDMA user ABI to support flush
RDMA/rxe: Fix incorrect responder length checking
RDMA/rxe: Fix oops with zero length reads
RDMA/mlx5: Remove not-used IB_FLOW_SPEC_IB define
RDMA/hns: Fix XRC caps on HIP08
RDMA/hns: Fix error code of CMD
RDMA/hns: Fix page size cap from firmware
RDMA/hns: Fix PBL page MTR find
RDMA/hns: Fix AH attr queried by query_qp
...
Diffstat (limited to 'drivers/infiniband/hw')
42 files changed, 2393 insertions, 346 deletions
diff --git a/drivers/infiniband/hw/Makefile b/drivers/infiniband/hw/Makefile index 6b3a88046125..1211f4317a9f 100644 --- a/drivers/infiniband/hw/Makefile +++ b/drivers/infiniband/hw/Makefile @@ -4,6 +4,7 @@ obj-$(CONFIG_INFINIBAND_QIB) += qib/ obj-$(CONFIG_INFINIBAND_CXGB4) += cxgb4/ obj-$(CONFIG_INFINIBAND_EFA) += efa/ obj-$(CONFIG_INFINIBAND_IRDMA) += irdma/ +obj-$(CONFIG_MANA_INFINIBAND) += mana/ obj-$(CONFIG_MLX4_INFINIBAND) += mlx4/ obj-$(CONFIG_MLX5_INFINIBAND) += mlx5/ obj-$(CONFIG_INFINIBAND_OCRDMA) += ocrdma/ diff --git a/drivers/infiniband/hw/erdma/erdma.h b/drivers/infiniband/hw/erdma/erdma.h index 730783fbc894..3d8c11aa23a2 100644 --- a/drivers/infiniband/hw/erdma/erdma.h +++ b/drivers/infiniband/hw/erdma/erdma.h @@ -124,6 +124,7 @@ struct erdma_devattr { u32 fw_version; unsigned char peer_addr[ETH_ALEN]; + unsigned long cap_flags; int numa_node; enum erdma_cc_alg cc; @@ -189,6 +190,7 @@ struct erdma_dev { struct net_device *netdev; struct pci_dev *pdev; struct notifier_block netdev_nb; + struct workqueue_struct *reflush_wq; resource_size_t func_bar_addr; resource_size_t func_bar_len; @@ -218,7 +220,7 @@ struct erdma_dev { DECLARE_BITMAP(sdb_page, ERDMA_DWQE_TYPE0_CNT); /* * We provide max 496 uContexts that each has one SQ normal Db, - * and one directWQE db。 + * and one directWQE db. */ DECLARE_BITMAP(sdb_entry, ERDMA_DWQE_TYPE1_CNT); diff --git a/drivers/infiniband/hw/erdma/erdma_cq.c b/drivers/infiniband/hw/erdma/erdma_cq.c index 58e0dc5c75d1..cabd8678b355 100644 --- a/drivers/infiniband/hw/erdma/erdma_cq.c +++ b/drivers/infiniband/hw/erdma/erdma_cq.c @@ -64,6 +64,8 @@ static const enum ib_wc_opcode wc_mapping_table[ERDMA_NUM_OPCODES] = { [ERDMA_OP_REG_MR] = IB_WC_REG_MR, [ERDMA_OP_LOCAL_INV] = IB_WC_LOCAL_INV, [ERDMA_OP_READ_WITH_INV] = IB_WC_RDMA_READ, + [ERDMA_OP_ATOMIC_CAS] = IB_WC_COMP_SWAP, + [ERDMA_OP_ATOMIC_FAD] = IB_WC_FETCH_ADD, }; static const struct { diff --git a/drivers/infiniband/hw/erdma/erdma_hw.h b/drivers/infiniband/hw/erdma/erdma_hw.h index e788887732e1..ab371fec610c 100644 --- a/drivers/infiniband/hw/erdma/erdma_hw.h +++ b/drivers/infiniband/hw/erdma/erdma_hw.h @@ -145,6 +145,7 @@ enum CMDQ_RDMA_OPCODE { CMDQ_OPCODE_MODIFY_QP = 3, CMDQ_OPCODE_CREATE_CQ = 4, CMDQ_OPCODE_DESTROY_CQ = 5, + CMDQ_OPCODE_REFLUSH = 6, CMDQ_OPCODE_REG_MR = 8, CMDQ_OPCODE_DEREG_MR = 9 }; @@ -224,8 +225,7 @@ struct erdma_cmdq_create_cq_req { /* regmr cfg1 */ #define ERDMA_CMD_REGMR_PD_MASK GENMASK(31, 12) #define ERDMA_CMD_REGMR_TYPE_MASK GENMASK(7, 6) -#define ERDMA_CMD_REGMR_RIGHT_MASK GENMASK(5, 2) -#define ERDMA_CMD_REGMR_ACC_MODE_MASK GENMASK(1, 0) +#define ERDMA_CMD_REGMR_RIGHT_MASK GENMASK(5, 1) /* regmr cfg2 */ #define ERDMA_CMD_REGMR_PAGESIZE_MASK GENMASK(31, 27) @@ -302,8 +302,16 @@ struct erdma_cmdq_destroy_qp_req { u32 qpn; }; +struct erdma_cmdq_reflush_req { + u64 hdr; + u32 qpn; + u32 sq_pi; + u32 rq_pi; +}; + /* cap qword 0 definition */ #define ERDMA_CMD_DEV_CAP_MAX_CQE_MASK GENMASK_ULL(47, 40) +#define ERDMA_CMD_DEV_CAP_FLAGS_MASK GENMASK_ULL(31, 24) #define ERDMA_CMD_DEV_CAP_MAX_RECV_WR_MASK GENMASK_ULL(23, 16) #define ERDMA_CMD_DEV_CAP_MAX_MR_SIZE_MASK GENMASK_ULL(7, 0) @@ -315,6 +323,10 @@ struct erdma_cmdq_destroy_qp_req { #define ERDMA_NQP_PER_QBLOCK 1024 +enum { + ERDMA_DEV_CAP_FLAGS_ATOMIC = 1 << 7, +}; + #define ERDMA_CMD_INFO0_FW_VER_MASK GENMASK_ULL(31, 0) /* CQE hdr */ @@ -340,9 +352,9 @@ struct erdma_cqe { }; struct erdma_sge { - __aligned_le64 laddr; + __aligned_le64 addr; __le32 length; - __le32 lkey; + __le32 key; }; /* Receive Queue Element */ @@ -370,8 +382,7 @@ struct erdma_rqe { #define ERDMA_SQE_HDR_WQEBB_INDEX_MASK GENMASK_ULL(15, 0) /* REG MR attrs */ -#define ERDMA_SQE_MR_MODE_MASK GENMASK(1, 0) -#define ERDMA_SQE_MR_ACCESS_MASK GENMASK(5, 2) +#define ERDMA_SQE_MR_ACCESS_MASK GENMASK(5, 1) #define ERDMA_SQE_MR_MTT_TYPE_MASK GENMASK(7, 6) #define ERDMA_SQE_MR_MTT_CNT_MASK GENMASK(31, 12) @@ -410,6 +421,16 @@ struct erdma_readreq_sqe { __le32 rsvd; }; +struct erdma_atomic_sqe { + __le64 hdr; + __le64 rsvd; + __le64 fetchadd_swap_data; + __le64 cmp_data; + + struct erdma_sge remote; + struct erdma_sge sgl; +}; + struct erdma_reg_mr_sqe { __le64 hdr; __le64 addr; @@ -469,7 +490,9 @@ enum erdma_opcode { ERDMA_OP_REG_MR = 14, ERDMA_OP_LOCAL_INV = 15, ERDMA_OP_READ_WITH_INV = 16, - ERDMA_NUM_OPCODES = 17, + ERDMA_OP_ATOMIC_CAS = 17, + ERDMA_OP_ATOMIC_FAD = 18, + ERDMA_NUM_OPCODES = 19, ERDMA_OP_INVALID = ERDMA_NUM_OPCODES + 1 }; diff --git a/drivers/infiniband/hw/erdma/erdma_main.c b/drivers/infiniband/hw/erdma/erdma_main.c index 49778bb294ae..5dc31e5df5cb 100644 --- a/drivers/infiniband/hw/erdma/erdma_main.c +++ b/drivers/infiniband/hw/erdma/erdma_main.c @@ -374,6 +374,7 @@ static int erdma_dev_attrs_init(struct erdma_dev *dev) dev->attrs.max_qp = ERDMA_NQP_PER_QBLOCK * ERDMA_GET_CAP(QBLOCK, cap1); dev->attrs.max_mr = dev->attrs.max_qp << 1; dev->attrs.max_cq = dev->attrs.max_qp << 1; + dev->attrs.cap_flags = ERDMA_GET_CAP(FLAGS, cap0); dev->attrs.max_send_wr = ERDMA_MAX_SEND_WR; dev->attrs.max_ord = ERDMA_MAX_ORD; @@ -520,13 +521,22 @@ static int erdma_ib_device_add(struct pci_dev *pdev) u64_to_ether_addr(mac, dev->attrs.peer_addr); + dev->reflush_wq = alloc_workqueue("erdma-reflush-wq", WQ_UNBOUND, + WQ_UNBOUND_MAX_ACTIVE); + if (!dev->reflush_wq) { + ret = -ENOMEM; + goto err_alloc_workqueue; + } + ret = erdma_device_register(dev); if (ret) - goto err_out; + goto err_register; return 0; -err_out: +err_register: + destroy_workqueue(dev->reflush_wq); +err_alloc_workqueue: xa_destroy(&dev->qp_xa); xa_destroy(&dev->cq_xa); @@ -542,6 +552,7 @@ static void erdma_ib_device_remove(struct pci_dev *pdev) unregister_netdevice_notifier(&dev->netdev_nb); ib_unregister_device(&dev->ibdev); + destroy_workqueue(dev->reflush_wq); erdma_res_cb_free(dev); xa_destroy(&dev->qp_xa); xa_destroy(&dev->cq_xa); diff --git a/drivers/infiniband/hw/erdma/erdma_qp.c b/drivers/infiniband/hw/erdma/erdma_qp.c index 5fe1a339a435..d088d6bef431 100644 --- a/drivers/infiniband/hw/erdma/erdma_qp.c +++ b/drivers/infiniband/hw/erdma/erdma_qp.c @@ -120,6 +120,7 @@ static int erdma_modify_qp_state_to_stop(struct erdma_qp *qp, int erdma_modify_qp_internal(struct erdma_qp *qp, struct erdma_qp_attrs *attrs, enum erdma_qp_attr_mask mask) { + bool need_reflush = false; int drop_conn, ret = 0; if (!mask) @@ -135,6 +136,7 @@ int erdma_modify_qp_internal(struct erdma_qp *qp, struct erdma_qp_attrs *attrs, ret = erdma_modify_qp_state_to_rts(qp, attrs, mask); } else if (attrs->state == ERDMA_QP_STATE_ERROR) { qp->attrs.state = ERDMA_QP_STATE_ERROR; + need_reflush = true; if (qp->cep) { erdma_cep_put(qp->cep); qp->cep = NULL; @@ -145,17 +147,12 @@ int erdma_modify_qp_internal(struct erdma_qp *qp, struct erdma_qp_attrs *attrs, case ERDMA_QP_STATE_RTS: drop_conn = 0; - if (attrs->state == ERDMA_QP_STATE_CLOSING) { + if (attrs->state == ERDMA_QP_STATE_CLOSING || + attrs->state == ERDMA_QP_STATE_TERMINATE || + attrs->state == ERDMA_QP_STATE_ERROR) { ret = erdma_modify_qp_state_to_stop(qp, attrs, mask); drop_conn = 1; - } else if (attrs->state == ERDMA_QP_STATE_TERMINATE) { - qp->attrs.state = ERDMA_QP_STATE_TERMINATE; - ret = erdma_modify_qp_state_to_stop(qp, attrs, mask); - drop_conn = 1; - } else if (attrs->state == ERDMA_QP_STATE_ERROR) { - ret = erdma_modify_qp_state_to_stop(qp, attrs, mask); - qp->attrs.state = ERDMA_QP_STATE_ERROR; - drop_conn = 1; + need_reflush = true; } if (drop_conn) @@ -180,6 +177,12 @@ int erdma_modify_qp_internal(struct erdma_qp *qp, struct erdma_qp_attrs *attrs, break; } + if (need_reflush && !ret && rdma_is_kernel_res(&qp->ibqp.res)) { + qp->flags |= ERDMA_QP_IN_FLUSHING; + mod_delayed_work(qp->dev->reflush_wq, &qp->reflush_dwork, + usecs_to_jiffies(100)); + } + return ret; } @@ -285,15 +288,16 @@ static int erdma_push_one_sqe(struct erdma_qp *qp, u16 *pi, u32 wqe_size, wqebb_cnt, hw_op, flags, sgl_offset; u32 idx = *pi & (qp->attrs.sq_size - 1); enum ib_wr_opcode op = send_wr->opcode; + struct erdma_atomic_sqe *atomic_sqe; struct erdma_readreq_sqe *read_sqe; struct erdma_reg_mr_sqe *regmr_sge; struct erdma_write_sqe *write_sqe; struct erdma_send_sqe *send_sqe; struct ib_rdma_wr *rdma_wr; - struct erdma_mr *mr; + struct erdma_sge *sge; __le32 *length_field; + struct erdma_mr *mr; u64 wqe_hdr, *entry; - struct ib_sge *sge; u32 attrs; int ret; @@ -360,9 +364,9 @@ static int erdma_push_one_sqe(struct erdma_qp *qp, u16 *pi, sge = get_queue_entry(qp->kern_qp.sq_buf, idx + 1, qp->attrs.sq_size, SQEBB_SHIFT); - sge->addr = rdma_wr->remote_addr; - sge->lkey = rdma_wr->rkey; - sge->length = send_wr->sg_list[0].length; + sge->addr = cpu_to_le64(rdma_wr->remote_addr); + sge->key = cpu_to_le32(rdma_wr->rkey); + sge->length = cpu_to_le32(send_wr->sg_list[0].length); wqe_size = sizeof(struct erdma_readreq_sqe) + send_wr->num_sge * sizeof(struct ib_sge); @@ -397,8 +401,7 @@ static int erdma_push_one_sqe(struct erdma_qp *qp, u16 *pi, regmr_sge->addr = cpu_to_le64(mr->ibmr.iova); regmr_sge->length = cpu_to_le32(mr->ibmr.length); regmr_sge->stag = cpu_to_le32(reg_wr(send_wr)->key); - attrs = FIELD_PREP(ERDMA_SQE_MR_MODE_MASK, 0) | - FIELD_PREP(ERDMA_SQE_MR_ACCESS_MASK, mr->access) | + attrs = FIELD_PREP(ERDMA_SQE_MR_ACCESS_MASK, mr->access) | FIELD_PREP(ERDMA_SQE_MR_MTT_CNT_MASK, mr->mem.mtt_nents); @@ -424,6 +427,35 @@ static int erdma_push_one_sqe(struct erdma_qp *qp, u16 *pi, regmr_sge->stag = cpu_to_le32(send_wr->ex.invalidate_rkey); wqe_size = sizeof(struct erdma_reg_mr_sqe); goto out; + case IB_WR_ATOMIC_CMP_AND_SWP: + case IB_WR_ATOMIC_FETCH_AND_ADD: + atomic_sqe = (struct erdma_atomic_sqe *)entry; + if (op == IB_WR_ATOMIC_CMP_AND_SWP) { + wqe_hdr |= FIELD_PREP(ERDMA_SQE_HDR_OPCODE_MASK, + ERDMA_OP_ATOMIC_CAS); + atomic_sqe->fetchadd_swap_data = + cpu_to_le64(atomic_wr(send_wr)->swap); + atomic_sqe->cmp_data = + cpu_to_le64(atomic_wr(send_wr)->compare_add); + } else { + wqe_hdr |= FIELD_PREP(ERDMA_SQE_HDR_OPCODE_MASK, + ERDMA_OP_ATOMIC_FAD); + atomic_sqe->fetchadd_swap_data = + cpu_to_le64(atomic_wr(send_wr)->compare_add); + } + + sge = get_queue_entry(qp->kern_qp.sq_buf, idx + 1, + qp->attrs.sq_size, SQEBB_SHIFT); + sge->addr = cpu_to_le64(atomic_wr(send_wr)->remote_addr); + sge->key = cpu_to_le32(atomic_wr(send_wr)->rkey); + sge++; + + sge->addr = cpu_to_le64(send_wr->sg_list[0].addr); + sge->key = cpu_to_le32(send_wr->sg_list[0].lkey); + sge->length = cpu_to_le32(send_wr->sg_list[0].length); + + wqe_size = sizeof(*atomic_sqe); + goto out; default: return -EOPNOTSUPP; } @@ -498,6 +530,10 @@ int erdma_post_send(struct ib_qp *ibqp, const struct ib_send_wr *send_wr, } spin_unlock_irqrestore(&qp->lock, flags); + if (unlikely(qp->flags & ERDMA_QP_IN_FLUSHING)) + mod_delayed_work(qp->dev->reflush_wq, &qp->reflush_dwork, + usecs_to_jiffies(100)); + return ret; } @@ -551,5 +587,10 @@ int erdma_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *recv_wr, } spin_unlock_irqrestore(&qp->lock, flags); + + if (unlikely(qp->flags & ERDMA_QP_IN_FLUSHING)) + mod_delayed_work(qp->dev->reflush_wq, &qp->reflush_dwork, + usecs_to_jiffies(100)); + return ret; } diff --git a/drivers/infiniband/hw/erdma/erdma_verbs.c b/drivers/infiniband/hw/erdma/erdma_verbs.c index 62be98e2b941..5dab1e87975b 100644 --- a/drivers/infiniband/hw/erdma/erdma_verbs.c +++ b/drivers/infiniband/hw/erdma/erdma_verbs.c @@ -118,8 +118,7 @@ static int regmr_cmd(struct erdma_dev *dev, struct erdma_mr *mr) FIELD_PREP(ERDMA_CMD_MR_MPT_IDX_MASK, mr->ibmr.lkey >> 8); req.cfg1 = FIELD_PREP(ERDMA_CMD_REGMR_PD_MASK, pd->pdn) | FIELD_PREP(ERDMA_CMD_REGMR_TYPE_MASK, mr->type) | - FIELD_PREP(ERDMA_CMD_REGMR_RIGHT_MASK, mr->access) | - FIELD_PREP(ERDMA_CMD_REGMR_ACC_MODE_MASK, 0); + FIELD_PREP(ERDMA_CMD_REGMR_RIGHT_MASK, mr->access); req.cfg2 = FIELD_PREP(ERDMA_CMD_REGMR_PAGESIZE_MASK, ilog2(mr->mem.page_size)) | FIELD_PREP(ERDMA_CMD_REGMR_MTT_TYPE_MASK, mr->mem.mtt_type) | @@ -289,6 +288,10 @@ int erdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr, attr->max_mw = dev->attrs.max_mw; attr->max_fast_reg_page_list_len = ERDMA_MAX_FRMR_PA; attr->page_size_cap = ERDMA_PAGE_SIZE_SUPPORT; + + if (dev->attrs.cap_flags & ERDMA_DEV_CAP_FLAGS_ATOMIC) + attr->atomic_cap = IB_ATOMIC_GLOB; + attr->fw_ver = dev->attrs.fw_version; if (dev->netdev) @@ -376,6 +379,21 @@ int erdma_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata) return 0; } +static void erdma_flush_worker(struct work_struct *work) +{ + struct delayed_work *dwork = to_delayed_work(work); + struct erdma_qp *qp = + container_of(dwork, struct erdma_qp, reflush_dwork); + struct erdma_cmdq_reflush_req req; + + erdma_cmdq_build_reqhdr(&req.hdr, CMDQ_SUBMOD_RDMA, + CMDQ_OPCODE_REFLUSH); + req.qpn = QP_ID(qp); + req.sq_pi = qp->kern_qp.sq_pi; + req.rq_pi = qp->kern_qp.rq_pi; + erdma_post_cmd_wait(&qp->dev->cmdq, &req, sizeof(req), NULL, NULL); +} + static int erdma_qp_validate_cap(struct erdma_dev *dev, struct ib_qp_init_attr *attrs) { @@ -732,6 +750,7 @@ int erdma_create_qp(struct ib_qp *ibqp, struct ib_qp_init_attr *attrs, qp->attrs.max_send_sge = attrs->cap.max_send_sge; qp->attrs.max_recv_sge = attrs->cap.max_recv_sge; qp->attrs.state = ERDMA_QP_STATE_IDLE; + INIT_DELAYED_WORK(&qp->reflush_dwork, erdma_flush_worker); ret = create_qp_cmd(dev, qp); if (ret) @@ -1025,6 +1044,8 @@ int erdma_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata) erdma_modify_qp_internal(qp, &qp_attrs, ERDMA_QP_ATTR_STATE); up_write(&qp->state_lock); + cancel_delayed_work_sync(&qp->reflush_dwork); + erdma_cmdq_build_reqhdr(&req.hdr, CMDQ_SUBMOD_RDMA, CMDQ_OPCODE_DESTROY_QP); req.qpn = QP_ID(qp); diff --git a/drivers/infiniband/hw/erdma/erdma_verbs.h b/drivers/infiniband/hw/erdma/erdma_verbs.h index ab6380635e9e..e0a993bc032a 100644 --- a/drivers/infiniband/hw/erdma/erdma_verbs.h +++ b/drivers/infiniband/hw/erdma/erdma_verbs.h @@ -71,16 +71,18 @@ struct erdma_pd { #define ERDMA_MR_INLINE_MTT 0 #define ERDMA_MR_INDIRECT_MTT 1 -#define ERDMA_MR_ACC_LR BIT(0) -#define ERDMA_MR_ACC_LW BIT(1) -#define ERDMA_MR_ACC_RR BIT(2) -#define ERDMA_MR_ACC_RW BIT(3) +#define ERDMA_MR_ACC_RA BIT(0) +#define ERDMA_MR_ACC_LR BIT(1) +#define ERDMA_MR_ACC_LW BIT(2) +#define ERDMA_MR_ACC_RR BIT(3) +#define ERDMA_MR_ACC_RW BIT(4) static inline u8 to_erdma_access_flags(int access) { return (access & IB_ACCESS_REMOTE_READ ? ERDMA_MR_ACC_RR : 0) | (access & IB_ACCESS_LOCAL_WRITE ? ERDMA_MR_ACC_LW : 0) | - (access & IB_ACCESS_REMOTE_WRITE ? ERDMA_MR_ACC_RW : 0); + (access & IB_ACCESS_REMOTE_WRITE ? ERDMA_MR_ACC_RW : 0) | + (access & IB_ACCESS_REMOTE_ATOMIC ? ERDMA_MR_ACC_RA : 0); } struct erdma_mem { @@ -171,6 +173,10 @@ enum erdma_qp_attr_mask { ERDMA_QP_ATTR_MPA = (1 << 7) }; +enum erdma_qp_flags { + ERDMA_QP_IN_FLUSHING = (1 << 0), +}; + struct erdma_qp_attrs { enum erdma_qp_state state; enum erdma_cc_alg cc; /* Congestion control algorithm */ @@ -195,6 +201,9 @@ struct erdma_qp { struct erdma_cep *cep; struct rw_semaphore state_lock; + unsigned long flags; + struct delayed_work reflush_dwork; + union { struct erdma_kqp kern_qp; struct erdma_uqp user_qp; diff --git a/drivers/infiniband/hw/hfi1/affinity.c b/drivers/infiniband/hw/hfi1/affinity.c index 877f8e84a672..77ee77d4000f 100644 --- a/drivers/infiniband/hw/hfi1/affinity.c +++ b/drivers/infiniband/hw/hfi1/affinity.c @@ -177,6 +177,8 @@ out: for (node = 0; node < node_affinity.num_possible_nodes; node++) hfi1_per_node_cntr[node] = 1; + pci_dev_put(dev); + return 0; } diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c index 8e71bef9d982..bcc6bc0540f0 100644 --- a/drivers/infiniband/hw/hfi1/driver.c +++ b/drivers/infiniband/hw/hfi1/driver.c @@ -112,7 +112,7 @@ static int hfi1_caps_get(char *buffer, const struct kernel_param *kp) cap_mask &= ~HFI1_CAP_LOCKED_SMASK; cap_mask |= ((cap_mask & HFI1_CAP_K2U) << HFI1_CAP_USER_SHIFT); - return scnprintf(buffer, PAGE_SIZE, "0x%lx", cap_mask); + return sysfs_emit(buffer, "0x%lx\n", cap_mask); } struct pci_dev *get_pci_dev(struct rvt_dev_info *rdi) diff --git a/drivers/infiniband/hw/hfi1/firmware.c b/drivers/infiniband/hw/hfi1/firmware.c index 1d77514ebbee..0c0cef5b1e0e 100644 --- a/drivers/infiniband/hw/hfi1/firmware.c +++ b/drivers/infiniband/hw/hfi1/firmware.c @@ -1743,6 +1743,7 @@ int parse_platform_config(struct hfi1_devdata *dd) if (!dd->platform_config.data) { dd_dev_err(dd, "%s: Missing config file\n", __func__); + ret = -EINVAL; goto bail; } ptr = (u32 *)dd->platform_config.data; @@ -1751,6 +1752,7 @@ int parse_platform_config(struct hfi1_devdata *dd) ptr++; if (magic_num != PLATFORM_CONFIG_MAGIC_NUM) { dd_dev_err(dd, "%s: Bad config file\n", __func__); + ret = -EINVAL; goto bail; } @@ -1774,6 +1776,7 @@ int parse_platform_config(struct hfi1_devdata *dd) if (file_length > dd->platform_config.size) { dd_dev_info(dd, "%s:File claims to be larger than read size\n", __func__); + ret = -EINVAL; goto bail; } else if (file_length < dd->platform_config.size) { dd_dev_info(dd, @@ -1794,6 +1797,7 @@ int parse_platform_config(struct hfi1_devdata *dd) dd_dev_err(dd, "%s: Failed validation at offset %ld\n", __func__, (ptr - (u32 *) dd->platform_config.data)); + ret = -EINVAL; goto bail; } @@ -1837,6 +1841,7 @@ int parse_platform_config(struct hfi1_devdata *dd) __func__, table_type, (ptr - (u32 *) dd->platform_config.data)); + ret = -EINVAL; goto bail; /* We don't trust this file now */ } pcfgcache->config_tables[table_type].table = ptr; @@ -1856,6 +1861,7 @@ int parse_platform_config(struct hfi1_devdata *dd) __func__, table_type, (ptr - (u32 *)dd->platform_config.data)); + ret = -EINVAL; goto bail; /* We don't trust this file now */ } pcfgcache->config_tables[table_type].table_metadata = diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c index 4146a2113a95..e5e783c45810 100644 --- a/drivers/infiniband/hw/hfi1/mad.c +++ b/drivers/infiniband/hw/hfi1/mad.c @@ -2437,9 +2437,9 @@ struct opa_port_data_counters_msg { __be64 port_vl_xmit_wait_data; __be64 port_vl_rcv_bubble; __be64 port_vl_mark_fecn; - } vls[0]; + } vls[]; /* array size defined by #bits set in vl_select_mask*/ - } port[1]; /* array size defined by #ports in attribute modifier */ + } port; }; struct opa_port_error_counters64_msg { @@ -2470,9 +2470,9 @@ struct opa_port_error_counters64_msg { u8 reserved3[7]; struct _vls_ectrs { __be64 port_vl_xmit_discards; - } vls[0]; + } vls[]; /* array size defined by #bits set in vl_select_mask */ - } port[1]; /* array size defined by #ports in attribute modifier */ + } port; }; struct opa_port_error_info_msg { @@ -2543,7 +2543,7 @@ struct opa_port_error_info_msg { u8 error_info; } __packed fm_config_ei; __u32 reserved9; - } port[1]; /* actual array size defined by #ports in attr modifier */ + } port; }; /* opa_port_error_info_msg error_info_select_mask bit definitions */ @@ -2966,7 +2966,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp, } /* Sanity check */ - response_data_size = struct_size(req, port[0].vls, num_vls); + response_data_size = struct_size(req, port.vls, num_vls); if (response_data_size > sizeof(pmp->data)) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; @@ -2986,7 +2986,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp, return reply((struct ib_mad_hdr *)pmp); } - rsp = &req->port[0]; + rsp = &req->port; memset(rsp, 0, sizeof(*rsp)); rsp->port_number = port; @@ -3182,7 +3182,7 @@ static int pma_get_opa_porterrors(struct opa_pma_mad *pmp, return reply((struct ib_mad_hdr *)pmp); } - response_data_size = struct_size(req, port[0].vls, num_vls); + response_data_size = struct_size(req, port.vls, num_vls); if (response_data_size > sizeof(pmp->data)) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; @@ -3201,7 +3201,7 @@ static int pma_get_opa_porterrors(struct opa_pma_mad *pmp, return reply((struct ib_mad_hdr *)pmp); } - rsp = &req->port[0]; + rsp = &req->port; ibp = to_iport(ibdev, port_num); ppd = ppd_from_ibp(ibp); @@ -3340,7 +3340,7 @@ static int pma_get_opa_errorinfo(struct opa_pma_mad *pmp, u64 reg; req = (struct opa_port_error_info_msg *)pmp->data; - rsp = &req->port[0]; + rsp = &req->port; num_ports = OPA_AM_NPORT(be32_to_cpu(pmp->mad_hdr.attr_mod)); num_pslm = hweight64(be64_to_cpu(req->port_select_mask[3])); @@ -3590,7 +3590,7 @@ static int pma_set_opa_errorinfo(struct opa_pma_mad *pmp, u32 error_info_select; req = (struct opa_port_error_info_msg *)pmp->data; - rsp = &req->port[0]; + rsp = &req->port; num_ports = OPA_AM_NPORT(be32_to_cpu(pmp->mad_hdr.attr_mod)); num_pslm = hweight64(be64_to_cpu(req->port_select_mask[3])); diff --git a/drivers/infiniband/hw/hfi1/netdev_rx.c b/drivers/infiniband/hw/hfi1/netdev_rx.c index 3dfa5aff2512..720d4c85c9c9 100644 --- a/drivers/infiniband/hw/hfi1/netdev_rx.c +++ b/drivers/infiniband/hw/hfi1/netdev_rx.c @@ -216,7 +216,7 @@ static int hfi1_netdev_rxq_init(struct hfi1_netdev_rx *rx) * right now. */ set_bit(NAPI_STATE_NO_BUSY_POLL, &rxq->napi.state); - netif_napi_add_weight(dev, &rxq->napi, hfi1_netdev_rx_napi, 64); + netif_napi_add(dev, &rxq->napi, hfi1_netdev_rx_napi); rc = msix_netdev_request_rcd_irq(rxq->rcd); if (rc) goto bail_context_irq_failure; diff --git a/drivers/infiniband/hw/hns/hns_roce_device.h b/drivers/infiniband/hw/hns/hns_roce_device.h index 723e55a7de8d..f701cc86896b 100644 --- a/drivers/infiniband/hw/hns/hns_roce_device.h +++ b/drivers/infiniband/hw/hns/hns_roce_device.h @@ -202,6 +202,7 @@ struct hns_roce_ucontext { struct list_head page_list; struct mutex page_mutex; struct hns_user_mmap_entry *db_mmap_entry; + u32 config; }; struct hns_roce_pd { @@ -334,6 +335,7 @@ struct hns_roce_wq { u32 head; u32 tail; void __iomem *db_reg; + u32 ext_sge_cnt; }; struct hns_roce_sge { @@ -635,6 +637,7 @@ struct hns_roce_qp { struct list_head rq_node; /* all recv qps are on a list */ struct list_head sq_node; /* all send qps are on a list */ struct hns_user_mmap_entry *dwqe_mmap_entry; + u32 config; }; struct hns_roce_ib_iboe { diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c index 1435fe2ea176..b2421883993b 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c @@ -192,7 +192,6 @@ static int fill_ext_sge_inl_data(struct hns_roce_qp *qp, unsigned int *sge_idx, u32 msg_len) { struct ib_device *ibdev = &(to_hr_dev(qp->ibqp.device))->ib_dev; - unsigned int ext_sge_sz = qp->sq.max_gs * HNS_ROCE_SGE_SIZE; unsigned int left_len_in_pg; unsigned int idx = *sge_idx; unsigned int i = 0; @@ -200,7 +199,7 @@ static int fill_ext_sge_inl_data(struct hns_roce_qp *qp, void *addr; void *dseg; - if (msg_len > ext_sge_sz) { + if (msg_len > qp->sq.ext_sge_cnt * HNS_ROCE_SGE_SIZE) { ibdev_err(ibdev, "no enough extended sge space for inline data.\n"); return -EINVAL; @@ -1274,6 +1273,30 @@ static void update_cmdq_status(struct hns_roce_dev *hr_dev) hr_dev->cmd.state = HNS_ROCE_CMDQ_STATE_FATAL_ERR; } +static int hns_roce_cmd_err_convert_errno(u16 desc_ret) +{ + struct hns_roce_cmd_errcode errcode_table[] = { + {CMD_EXEC_SUCCESS, 0}, + {CMD_NO_AUTH, -EPERM}, + {CMD_NOT_EXIST, -EOPNOTSUPP}, + {CMD_CRQ_FULL, -EXFULL}, + {CMD_NEXT_ERR, -ENOSR}, + {CMD_NOT_EXEC, -ENOTBLK}, + {CMD_PARA_ERR, -EINVAL}, + {CMD_RESULT_ERR, -ERANGE}, + {CMD_TIMEOUT, -ETIME}, + {CMD_HILINK_ERR, -ENOLINK}, + {CMD_INFO_ILLEGAL, -ENXIO}, + {CMD_INVALID, -EBADR}, + }; + u16 i; + + for (i = 0; i < ARRAY_SIZE(errcode_table); i++) + if (desc_ret == errcode_table[i].return_status) + return errcode_table[i].errno; + return -EIO; +} + static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev, struct hns_roce_cmq_desc *desc, int num) { @@ -1319,7 +1342,7 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev, dev_err_ratelimited(hr_dev->dev, "Cmdq IO error, opcode = 0x%x, return = 0x%x.\n", desc->opcode, desc_ret); - ret = -EIO; + ret = hns_roce_cmd_err_convert_errno(desc_ret); } } else { /* FW/HW reset or incorrect number of desc */ @@ -2024,13 +2047,14 @@ static void set_default_caps(struct hns_roce_dev *hr_dev) caps->flags |= HNS_ROCE_CAP_FLAG_ATOMIC | HNS_ROCE_CAP_FLAG_MW | HNS_ROCE_CAP_FLAG_SRQ | HNS_ROCE_CAP_FLAG_FRMR | - HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL | HNS_ROCE_CAP_FLAG_XRC; + HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL; caps->gid_table_len[0] = HNS_ROCE_V2_GID_INDEX_NUM; if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09) { caps->flags |= HNS_ROCE_CAP_FLAG_STASH | - HNS_ROCE_CAP_FLAG_DIRECT_WQE; + HNS_ROCE_CAP_FLAG_DIRECT_WQE | + HNS_ROCE_CAP_FLAG_XRC; caps->max_sq_inline = HNS_ROCE_V3_MAX_SQ_INLINE; } else { caps->max_sq_inline = HNS_ROCE_V2_MAX_SQ_INLINE; @@ -2342,6 +2366,9 @@ static int hns_roce_query_pf_caps(struct hns_roce_dev *hr_dev) caps->wqe_sge_hop_num = hr_reg_read(resp_d, PF_CAPS_D_EX_SGE_HOP_NUM); caps->wqe_rq_hop_num = hr_reg_read(resp_d, PF_CAPS_D_RQWQE_HOP_NUM); + if (!(caps->page_size_cap & PAGE_SIZE)) + caps->page_size_cap = HNS_ROCE_V2_PAGE_SIZE_SUPPORTED; + return 0; } @@ -2631,31 +2658,124 @@ static void free_dip_list(struct hns_roce_dev *hr_dev) spin_unlock_irqrestore(&hr_dev->dip_list_lock, flags); } -static void free_mr_exit(struct hns_roce_dev *hr_dev) +static struct ib_pd *free_mr_init_pd(struct hns_roce_dev *hr_dev) +{ + struct hns_roce_v2_priv *priv = hr_dev->priv; + struct hns_roce_v2_free_mr *free_mr = &priv->free_mr; + struct ib_device *ibdev = &hr_dev->ib_dev; + struct hns_roce_pd *hr_pd; + struct ib_pd *pd; + + hr_pd = kzalloc(sizeof(*hr_pd), GFP_KERNEL); + if (ZERO_OR_NULL_PTR(hr_pd)) + return NULL; + pd = &hr_pd->ibpd; + pd->device = ibdev; + + if (hns_roce_alloc_pd(pd, NULL)) { + ibdev_err(ibdev, "failed to create pd for free mr.\n"); + kfree(hr_pd); + return NULL; + } + free_mr->rsv_pd = to_hr_pd(pd); + free_mr->rsv_pd->ibpd.device = &hr_dev->ib_dev; + free_mr->rsv_pd->ibpd.uobject = NULL; + free_mr->rsv_pd->ibpd.__internal_mr = NULL; + atomic_set(&free_mr->rsv_pd->ibpd.usecnt, 0); + + return pd; +} + +static struct ib_cq *free_mr_init_cq(struct hns_roce_dev *hr_dev) { struct hns_roce_v2_priv *priv = hr_dev->priv; struct hns_roce_v2_free_mr *free_mr = &priv->free_mr; + struct ib_device *ibdev = &hr_dev->ib_dev; + struct ib_cq_init_attr cq_init_attr = {}; + struct hns_roce_cq *hr_cq; + struct ib_cq *cq; + + cq_init_attr.cqe = HNS_ROCE_FREE_MR_USED_CQE_NUM; + + hr_cq = kzalloc(sizeof(*hr_cq), GFP_KERNEL); + if (ZERO_OR_NULL_PTR(hr_cq)) + return NULL; + + cq = &hr_cq->ib_cq; + cq->device = ibdev; + + if (hns_roce_create_cq(cq, &cq_init_attr, NULL)) { + ibdev_err(ibdev, "failed to create cq for free mr.\n"); + kfree(hr_cq); + return NULL; + } + free_mr->rsv_cq = to_hr_cq(cq); + free_mr->rsv_cq->ib_cq.device = &hr_dev->ib_dev; + free_mr->rsv_cq->ib_cq.uobject = NULL; + free_mr->rsv_cq->ib_cq.comp_handler = NULL; + free_mr->rsv_cq->ib_cq.event_handler = NULL; + free_mr->rsv_cq->ib_cq.cq_context = NULL; + atomic_set(&free_mr->rsv_cq->ib_cq.usecnt, 0); + + return cq; +} + +static int free_mr_init_qp(struct hns_roce_dev *hr_dev, struct ib_cq *cq, + struct ib_qp_init_attr *init_attr, int i) +{ + struct hns_roce_v2_priv *priv = hr_dev->priv; + struct hns_roce_v2_free_mr *free_mr = &priv->free_mr; + struct ib_device *ibdev = &hr_dev->ib_dev; + struct hns_roce_qp *hr_qp; + struct ib_qp *qp; int ret; + + hr_qp = kzalloc(sizeof(*hr_qp), GFP_KERNEL); + if (ZERO_OR_NULL_PTR(hr_qp)) + return -ENOMEM; + + qp = &hr_qp->ibqp; + qp->device = ibdev; + + ret = hns_roce_create_qp(qp, init_attr, NULL); + if (ret) { + ibdev_err(ibdev, "failed to create qp for free mr.\n"); + kfree(hr_qp); + return ret; + } + + free_mr->rsv_qp[i] = hr_qp; + free_mr->rsv_qp[i]->ibqp.recv_cq = cq; + free_mr->rsv_qp[i]->ibqp.send_cq = cq; + + return 0; +} + +static void free_mr_exit(struct hns_roce_dev *hr_dev) +{ + struct hns_roce_v2_priv *priv = hr_dev->priv; + struct hns_roce_v2_free_mr *free_mr = &priv->free_mr; + struct ib_qp *qp; int i; for (i = 0; i < ARRAY_SIZE(free_mr->rsv_qp); i++) { if (free_mr->rsv_qp[i]) { - ret = ib_destroy_qp(free_mr->rsv_qp[i]); - if (ret) - ibdev_err(&hr_dev->ib_dev, - "failed to destroy qp in free mr.\n"); - + qp = &free_mr->rsv_qp[i]->ibqp; + hns_roce_v2_destroy_qp(qp, NULL); + kfree(free_mr->rsv_qp[i]); free_mr->rsv_qp[i] = NULL; } } if (free_mr->rsv_cq) { - ib_destroy_cq(free_mr->rsv_cq); + hns_roce_destroy_cq(&free_mr->rsv_cq->ib_cq, NULL); + kfree(free_mr->rsv_cq); free_mr->rsv_cq = NULL; } if (free_mr->rsv_pd) { - ib_dealloc_pd(free_mr->rsv_pd); + hns_roce_dealloc_pd(&free_mr->rsv_pd->ibpd, NULL); + kfree(free_mr->rsv_pd); free_mr->rsv_pd = NULL; } } @@ -2664,55 +2784,46 @@ static int free_mr_alloc_res(struct hns_roce_dev *hr_dev) { struct hns_roce_v2_priv *priv = hr_dev->priv; struct hns_roce_v2_free_mr *free_mr = &priv->free_mr; - struct ib_device *ibdev = &hr_dev->ib_dev; - struct ib_cq_init_attr cq_init_attr = {}; struct ib_qp_init_attr qp_init_attr = {}; struct ib_pd *pd; struct ib_cq *cq; - struct ib_qp *qp; int ret; int i; - pd = ib_alloc_pd(ibdev, 0); - if (IS_ERR(pd)) { - ibdev_err(ibdev, "failed to create pd for free mr.\n"); - return PTR_ERR(pd); - } - free_mr->rsv_pd = pd; + pd = free_mr_init_pd(hr_dev); + if (!pd) + return -ENOMEM; - cq_init_attr.cqe = HNS_ROCE_FREE_MR_USED_CQE_NUM; - cq = ib_create_cq(ibdev, NULL, NULL, NULL, &cq_init_attr); - if (IS_ERR(cq)) { - ibdev_err(ibdev, "failed to create cq for free mr.\n"); - ret = PTR_ERR(cq); - goto create_failed; + cq = free_mr_init_cq(hr_dev); + if (!cq) { + ret = -ENOMEM; + goto create_failed_cq; } - free_mr->rsv_cq = cq; qp_init_attr.qp_type = IB_QPT_RC; qp_init_attr.sq_sig_type = IB_SIGNAL_ALL_WR; - qp_init_attr.send_cq = free_mr->rsv_cq; - qp_init_attr.recv_cq = free_mr->rsv_cq; + qp_init_attr.send_cq = cq; + qp_init_attr.recv_cq = cq; for (i = 0; i < ARRAY_SIZE(free_mr->rsv_qp); i++) { qp_init_attr.cap.max_send_wr = HNS_ROCE_FREE_MR_USED_SQWQE_NUM; qp_init_attr.cap.max_send_sge = HNS_ROCE_FREE_MR_USED_SQSGE_NUM; qp_init_attr.cap.max_recv_wr = HNS_ROCE_FREE_MR_USED_RQWQE_NUM; qp_init_attr.cap.max_recv_sge = HNS_ROCE_FREE_MR_USED_RQSGE_NUM; - qp = ib_create_qp(free_mr->rsv_pd, &qp_init_attr); - if (IS_ERR(qp)) { - ibdev_err(ibdev, "failed to create qp for free mr.\n"); - ret = PTR_ERR(qp); - goto create_failed; - } - - free_mr->rsv_qp[i] = qp; + ret = free_mr_init_qp(hr_dev, cq, &qp_init_attr, i); + if (ret) + goto create_failed_qp; } return 0; -create_failed: - free_mr_exit(hr_dev); +create_failed_qp: + hns_roce_destroy_cq(cq, NULL); + kfree(cq); + +create_failed_cq: + hns_roce_dealloc_pd(pd, NULL); + kfree(pd); return ret; } @@ -2728,14 +2839,17 @@ static int free_mr_modify_rsv_qp(struct hns_roce_dev *hr_dev, int mask; int ret; - hr_qp = to_hr_qp(free_mr->rsv_qp[sl_num]); + hr_qp = to_hr_qp(&free_mr->rsv_qp[sl_num]->ibqp); hr_qp->free_mr_en = 1; + hr_qp->ibqp.device = ibdev; + hr_qp->ibqp.qp_type = IB_QPT_RC; mask = IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_PORT | IB_QP_ACCESS_FLAGS; attr->qp_state = IB_QPS_INIT; attr->port_num = 1; attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE; - ret = ib_modify_qp(&hr_qp->ibqp, attr, mask); + ret = hr_dev->hw->modify_qp(&hr_qp->ibqp, attr, mask, IB_QPS_INIT, + IB_QPS_INIT); if (ret) { ibdev_err(ibdev, "failed to modify qp to init, ret = %d.\n", ret); @@ -2756,7 +2870,8 @@ static int free_mr_modify_rsv_qp(struct hns_roce_dev *hr_dev, rdma_ah_set_sl(&attr->ah_attr, (u8)sl_num); - ret = ib_modify_qp(&hr_qp->ibqp, attr, mask); + ret = hr_dev->hw->modify_qp(&hr_qp->ibqp, attr, mask, IB_QPS_INIT, + IB_QPS_RTR); hr_dev->loop_idc = loopback; if (ret) { ibdev_err(ibdev, "failed to modify qp to rtr, ret = %d.\n", @@ -2770,7 +2885,8 @@ static int free_mr_modify_rsv_qp(struct hns_roce_dev *hr_dev, attr->sq_psn = HNS_ROCE_FREE_MR_USED_PSN; attr->retry_cnt = HNS_ROCE_FREE_MR_USED_QP_RETRY_CNT; attr->timeout = HNS_ROCE_FREE_MR_USED_QP_TIMEOUT; - ret = ib_modify_qp(&hr_qp->ibqp, attr, mask); + ret = hr_dev->hw->modify_qp(&hr_qp->ibqp, attr, mask, IB_QPS_RTR, + IB_QPS_RTS); if (ret) ibdev_err(ibdev, "failed to modify qp to rts, ret = %d.\n", ret); @@ -3186,7 +3302,8 @@ static int set_mtpt_pbl(struct hns_roce_dev *hr_dev, int i, count; count = hns_roce_mtr_find(hr_dev, &mr->pbl_mtr, 0, pages, - ARRAY_SIZE(pages), &pbl_ba); + min_t(int, ARRAY_SIZE(pages), mr->npages), + &pbl_ba); if (count < 1) { ibdev_err(ibdev, "failed to find PBL mtr, count = %d.\n", count); @@ -3414,7 +3531,7 @@ static void free_mr_send_cmd_to_hw(struct hns_roce_dev *hr_dev) mutex_lock(&free_mr->mutex); for (i = 0; i < ARRAY_SIZE(free_mr->rsv_qp); i++) { - hr_qp = to_hr_qp(free_mr->rsv_qp[i]); + hr_qp = free_mr->rsv_qp[i]; ret = free_mr_post_send_lp_wqe(hr_qp); if (ret) { @@ -3429,7 +3546,7 @@ static void free_mr_send_cmd_to_hw(struct hns_roce_dev *hr_dev) end = msecs_to_jiffies(HNS_ROCE_V2_FREE_MR_TIMEOUT) + jiffies; while (cqe_cnt) { - npolled = hns_roce_v2_poll_cq(free_mr->rsv_cq, cqe_cnt, wc); + npolled = hns_roce_v2_poll_cq(&free_mr->rsv_cq->ib_cq, cqe_cnt, wc); if (npolled < 0) { ibdev_err(ibdev, "failed to poll cqe for free mr, remain %d cqe.\n", @@ -5375,6 +5492,8 @@ static int hns_roce_v2_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, rdma_ah_set_sl(&qp_attr->ah_attr, hr_reg_read(&context, QPC_SL)); + rdma_ah_set_port_num(&qp_attr->ah_attr, hr_qp->port + 1); + rdma_ah_set_ah_flags(&qp_attr->ah_attr, IB_AH_GRH); grh->flow_label = hr_reg_read(&context, QPC_FL); grh->sgid_index = hr_reg_read(&context, QPC_GMV_IDX); grh->hop_limit = hr_reg_read(&context, QPC_HOPLIMIT); @@ -5468,7 +5587,7 @@ static int hns_roce_v2_destroy_qp_common(struct hns_roce_dev *hr_dev, return ret; } -static int hns_roce_v2_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata) +int hns_roce_v2_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata) { struct hns_roce_dev *hr_dev = to_hr_dev(ibqp->device); struct hns_roce_qp *hr_qp = to_hr_qp(ibqp); diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h index c7bf2d52c1cd..b1b3e1e0b84e 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h @@ -272,6 +272,11 @@ enum hns_roce_cmd_return_status { CMD_OTHER_ERR = 0xff }; +struct hns_roce_cmd_errcode { + enum hns_roce_cmd_return_status return_status; + int errno; +}; + enum hns_roce_sgid_type { GID_TYPE_FLAG_ROCE_V1 = 0, GID_TYPE_FLAG_ROCE_V2_IPV4, @@ -1327,9 +1332,9 @@ struct hns_roce_link_table { #define HNS_ROCE_EXT_LLM_MIN_PAGES(que_num) ((que_num) * 4 + 2) struct hns_roce_v2_free_mr { - struct ib_qp *rsv_qp[HNS_ROCE_FREE_MR_USED_QP_NUM]; - struct ib_cq *rsv_cq; - struct ib_pd *rsv_pd; + struct hns_roce_qp *rsv_qp[HNS_ROCE_FREE_MR_USED_QP_NUM]; + struct hns_roce_cq *rsv_cq; + struct hns_roce_pd *rsv_pd; struct mutex mutex; }; @@ -1459,6 +1464,8 @@ struct hns_roce_sccc_clr_done { __le32 rsv[5]; }; +int hns_roce_v2_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata); + static inline void hns_roce_write64(struct hns_roce_dev *hr_dev, __le32 val[2], void __iomem *dest) { diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c index dcf89689a4c6..8ba68ac12388 100644 --- a/drivers/infiniband/hw/hns/hns_roce_main.c +++ b/drivers/infiniband/hw/hns/hns_roce_main.c @@ -354,10 +354,11 @@ static int hns_roce_alloc_uar_entry(struct ib_ucontext *uctx) static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx, struct ib_udata *udata) { - int ret; struct hns_roce_ucontext *context = to_hr_ucontext(uctx); - struct hns_roce_ib_alloc_ucontext_resp resp = {}; struct hns_roce_dev *hr_dev = to_hr_dev(uctx->device); + struct hns_roce_ib_alloc_ucontext_resp resp = {}; + struct hns_roce_ib_alloc_ucontext ucmd = {}; + int ret; if (!hr_dev->active) return -EAGAIN; @@ -365,6 +366,19 @@ static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx, resp.qp_tab_size = hr_dev->caps.num_qps; resp.srq_tab_size = hr_dev->caps.num_srqs; + ret = ib_copy_from_udata(&ucmd, udata, + min(udata->inlen, sizeof(ucmd))); + if (ret) + return ret; + + if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09) + context->config = ucmd.config & HNS_ROCE_EXSGE_FLAGS; + + if (context->config & HNS_ROCE_EXSGE_FLAGS) { + resp.config |= HNS_ROCE_RSP_EXSGE_FLAGS; + resp.max_inline_data = hr_dev->caps.max_sq_inline; + } + ret = hns_roce_uar_alloc(hr_dev, &context->uar); if (ret) goto error_fail_uar_alloc; diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c index 845ac7d3831f..37a5cf62f88b 100644 --- a/drivers/infiniband/hw/hns/hns_roce_mr.c +++ b/drivers/infiniband/hw/hns/hns_roce_mr.c @@ -392,10 +392,10 @@ struct ib_mr *hns_roce_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type, return &mr->ibmr; -err_key: - free_mr_key(hr_dev, mr); err_pbl: free_mr_pbl(hr_dev, mr); +err_key: + free_mr_key(hr_dev, mr); err_free: kfree(mr); return ERR_PTR(ret); diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c index f0bd82a18069..0ae335fb205c 100644 --- a/drivers/infiniband/hw/hns/hns_roce_qp.c +++ b/drivers/infiniband/hw/hns/hns_roce_qp.c @@ -476,38 +476,109 @@ static int set_rq_size(struct hns_roce_dev *hr_dev, struct ib_qp_cap *cap, return 0; } -static u32 get_wqe_ext_sge_cnt(struct hns_roce_qp *qp) +static u32 get_max_inline_data(struct hns_roce_dev *hr_dev, + struct ib_qp_cap *cap) { - /* GSI/UD QP only has extended sge */ - if (qp->ibqp.qp_type == IB_QPT_GSI || qp->ibqp.qp_type == IB_QPT_UD) - return qp->sq.max_gs; - - if (qp->sq.max_gs > HNS_ROCE_SGE_IN_WQE) - return qp->sq.max_gs - HNS_ROCE_SGE_IN_WQE; + if (cap->max_inline_data) { + cap->max_inline_data = roundup_pow_of_two(cap->max_inline_data); + return min(cap->max_inline_data, + hr_dev->caps.max_sq_inline); + } return 0; } +static void update_inline_data(struct hns_roce_qp *hr_qp, + struct ib_qp_cap *cap) +{ + u32 sge_num = hr_qp->sq.ext_sge_cnt; + + if (hr_qp->config & HNS_ROCE_EXSGE_FLAGS) { + if (!(hr_qp->ibqp.qp_type == IB_QPT_GSI || + hr_qp->ibqp.qp_type == IB_QPT_UD)) + sge_num = max((u32)HNS_ROCE_SGE_IN_WQE, sge_num); + + cap->max_inline_data = max(cap->max_inline_data, + sge_num * HNS_ROCE_SGE_SIZE); + } + + hr_qp->max_inline_data = cap->max_inline_data; +} + +static u32 get_sge_num_from_max_send_sge(bool is_ud_or_gsi, + u32 max_send_sge) +{ + unsigned int std_sge_num; + unsigned int min_sge; + + std_sge_num = is_ud_or_gsi ? 0 : HNS_ROCE_SGE_IN_WQE; + min_sge = is_ud_or_gsi ? 1 : 0; + return max_send_sge > std_sge_num ? (max_send_sge - std_sge_num) : + min_sge; +} + +static unsigned int get_sge_num_from_max_inl_data(bool is_ud_or_gsi, + u32 max_inline_data) +{ + unsigned int inline_sge; + + inline_sge = roundup_pow_of_two(max_inline_data) / HNS_ROCE_SGE_SIZE; + + /* + * if max_inline_data less than + * HNS_ROCE_SGE_IN_WQE * HNS_ROCE_SGE_SIZE, + * In addition to ud's mode, no need to extend sge. + */ + if (!is_ud_or_gsi && inline_sge <= HNS_ROCE_SGE_IN_WQE) + inline_sge = 0; + + return inline_sge; +} + static void set_ext_sge_param(struct hns_roce_dev *hr_dev, u32 sq_wqe_cnt, struct hns_roce_qp *hr_qp, struct ib_qp_cap *cap) { + bool is_ud_or_gsi = (hr_qp->ibqp.qp_type == IB_QPT_GSI || + hr_qp->ibqp.qp_type == IB_QPT_UD); + unsigned int std_sge_num; + u32 inline_ext_sge = 0; + u32 ext_wqe_sge_cnt; u32 total_sge_cnt; - u32 wqe_sge_cnt; + + cap->max_inline_data = get_max_inline_data(hr_dev, cap); hr_qp->sge.sge_shift = HNS_ROCE_SGE_SHIFT; + std_sge_num = is_ud_or_gsi ? 0 : HNS_ROCE_SGE_IN_WQE; + ext_wqe_sge_cnt = get_sge_num_from_max_send_sge(is_ud_or_gsi, + cap->max_send_sge); - hr_qp->sq.max_gs = max(1U, cap->max_send_sge); + if (hr_qp->config & HNS_ROCE_EXSGE_FLAGS) { + inline_ext_sge = max(ext_wqe_sge_cnt, + get_sge_num_from_max_inl_data(is_ud_or_gsi, + cap->max_inline_data)); + hr_qp->sq.ext_sge_cnt = inline_ext_sge ? + roundup_pow_of_two(inline_ext_sge) : 0; - wqe_sge_cnt = get_wqe_ext_sge_cnt(hr_qp); + hr_qp->sq.max_gs = max(1U, (hr_qp->sq.ext_sge_cnt + std_sge_num)); + hr_qp->sq.max_gs = min(hr_qp->sq.max_gs, hr_dev->caps.max_sq_sg); + + ext_wqe_sge_cnt = hr_qp->sq.ext_sge_cnt; + } else { + hr_qp->sq.max_gs = max(1U, cap->max_send_sge); + hr_qp->sq.max_gs = min(hr_qp->sq.max_gs, hr_dev->caps.max_sq_sg); + hr_qp->sq.ext_sge_cnt = hr_qp->sq.max_gs; + } /* If the number of extended sge is not zero, they MUST use the * space of HNS_HW_PAGE_SIZE at least. */ - if (wqe_sge_cnt) { - total_sge_cnt = roundup_pow_of_two(sq_wqe_cnt * wqe_sge_cnt); + if (ext_wqe_sge_cnt) { + total_sge_cnt = roundup_pow_of_two(sq_wqe_cnt * ext_wqe_sge_cnt); hr_qp->sge.sge_cnt = max(total_sge_cnt, (u32)HNS_HW_PAGE_SIZE / HNS_ROCE_SGE_SIZE); } + + update_inline_data(hr_qp, cap); } static int check_sq_size_with_integrity(struct hns_roce_dev *hr_dev, @@ -556,6 +627,7 @@ static int set_user_sq_size(struct hns_roce_dev *hr_dev, hr_qp->sq.wqe_shift = ucmd->log_sq_stride; hr_qp->sq.wqe_cnt = cnt; + cap->max_send_sge = hr_qp->sq.max_gs; return 0; } @@ -986,13 +1058,9 @@ static int set_qp_param(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp, struct hns_roce_ib_create_qp *ucmd) { struct ib_device *ibdev = &hr_dev->ib_dev; + struct hns_roce_ucontext *uctx; int ret; - if (init_attr->cap.max_inline_data > hr_dev->caps.max_sq_inline) - init_attr->cap.max_inline_data = hr_dev->caps.max_sq_inline; - - hr_qp->max_inline_data = init_attr->cap.max_inline_data; - if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR) hr_qp->sq_signal_bits = IB_SIGNAL_ALL_WR; else @@ -1015,12 +1083,17 @@ static int set_qp_param(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp, return ret; } + uctx = rdma_udata_to_drv_context(udata, struct hns_roce_ucontext, + ibucontext); + hr_qp->config = uctx->config; ret = set_user_sq_size(hr_dev, &init_attr->cap, hr_qp, ucmd); if (ret) ibdev_err(ibdev, "failed to set user SQ size, ret = %d.\n", ret); } else { + if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09) + hr_qp->config = HNS_ROCE_EXSGE_FLAGS; ret = set_kernel_sq_size(hr_dev, &init_attr->cap, hr_qp); if (ret) ibdev_err(ibdev, diff --git a/drivers/infiniband/hw/irdma/uk.c b/drivers/infiniband/hw/irdma/uk.c index a6e5d350a94c..16183e894da7 100644 --- a/drivers/infiniband/hw/irdma/uk.c +++ b/drivers/infiniband/hw/irdma/uk.c @@ -566,21 +566,37 @@ static void irdma_set_mw_bind_wqe_gen_1(__le64 *wqe, /** * irdma_copy_inline_data_gen_1 - Copy inline data to wqe - * @dest: pointer to wqe - * @src: pointer to inline data - * @len: length of inline data to copy + * @wqe: pointer to wqe + * @sge_list: table of pointers to inline data + * @num_sges: Total inline data length * @polarity: compatibility parameter */ -static void irdma_copy_inline_data_gen_1(u8 *dest, u8 *src, u32 len, - u8 polarity) +static void irdma_copy_inline_data_gen_1(u8 *wqe, struct ib_sge *sge_list, + u32 num_sges, u8 polarity) { - if (len <= 16) { - memcpy(dest, src, len); - } else { - memcpy(dest, src, 16); - src += 16; - dest = dest + 32; - memcpy(dest, src, len - 16); + u32 quanta_bytes_remaining = 16; + int i; + + for (i = 0; i < num_sges; i++) { + u8 *cur_sge = (u8 *)(uintptr_t)sge_list[i].addr; + u32 sge_len = sge_list[i].length; + + while (sge_len) { + u32 bytes_copied; + + bytes_copied = min(sge_len, quanta_bytes_remaining); + memcpy(wqe, cur_sge, bytes_copied); + wqe += bytes_copied; + cur_sge += bytes_copied; + quanta_bytes_remaining -= bytes_copied; + sge_len -= bytes_copied; + + if (!quanta_bytes_remaining) { + /* Remaining inline bytes reside after hdr */ + wqe += 16; + quanta_bytes_remaining = 32; + } + } } } @@ -612,35 +628,51 @@ static void irdma_set_mw_bind_wqe(__le64 *wqe, /** * irdma_copy_inline_data - Copy inline data to wqe - * @dest: pointer to wqe - * @src: pointer to inline data - * @len: length of inline data to copy + * @wqe: pointer to wqe + * @sge_list: table of pointers to inline data + * @num_sges: number of SGE's * @polarity: polarity of wqe valid bit */ -static void irdma_copy_inline_data(u8 *dest, u8 *src, u32 len, u8 polarity) +static void irdma_copy_inline_data(u8 *wqe, struct ib_sge *sge_list, + u32 num_sges, u8 polarity) { u8 inline_valid = polarity << IRDMA_INLINE_VALID_S; - u32 copy_size; - - dest += 8; - if (len <= 8) { - memcpy(dest, src, len); - return; - } - - *((u64 *)dest) = *((u64 *)src); - len -= 8; - src += 8; - dest += 24; /* point to additional 32 byte quanta */ - - while (len) { - copy_size = len < 31 ? len : 31; - memcpy(dest, src, copy_size); - *(dest + 31) = inline_valid; - len -= copy_size; - dest += 32; - src += copy_size; + u32 quanta_bytes_remaining = 8; + bool first_quanta = true; + int i; + + wqe += 8; + + for (i = 0; i < num_sges; i++) { + u8 *cur_sge = (u8 *)(uintptr_t)sge_list[i].addr; + u32 sge_len = sge_list[i].length; + + while (sge_len) { + u32 bytes_copied; + + bytes_copied = min(sge_len, quanta_bytes_remaining); + memcpy(wqe, cur_sge, bytes_copied); + wqe += bytes_copied; + cur_sge += bytes_copied; + quanta_bytes_remaining -= bytes_copied; + sge_len -= bytes_copied; + + if (!quanta_bytes_remaining) { + quanta_bytes_remaining = 31; + + /* Remaining inline bytes reside after hdr */ + if (first_quanta) { + first_quanta = false; + wqe += 16; + } else { + *wqe = inline_valid; + wqe++; + } + } + } } + if (!first_quanta && quanta_bytes_remaining < 31) + *(wqe + quanta_bytes_remaining) = inline_valid; } /** @@ -679,20 +711,27 @@ int irdma_uk_inline_rdma_write(struct irdma_qp_uk *qp, struct irdma_post_sq_info *info, bool post_sq) { __le64 *wqe; - struct irdma_inline_rdma_write *op_info; + struct irdma_rdma_write *op_info; u64 hdr = 0; u32 wqe_idx; bool read_fence = false; + u32 i, total_size = 0; u16 quanta; info->push_wqe = qp->push_db ? true : false; - op_info = &info->op.inline_rdma_write; + op_info = &info->op.rdma_write; + + if (unlikely(qp->max_sq_frag_cnt < op_info->num_lo_sges)) + return -EINVAL; + + for (i = 0; i < op_info->num_lo_sges; i++) + total_size += op_info->lo_sg_list[i].length; - if (op_info->len > qp->max_inline_data) + if (unlikely(total_size > qp->max_inline_data)) return -EINVAL; - quanta = qp->wqe_ops.iw_inline_data_size_to_quanta(op_info->len); - wqe = irdma_qp_get_next_send_wqe(qp, &wqe_idx, quanta, op_info->len, + quanta = qp->wqe_ops.iw_inline_data_size_to_quanta(total_size); + wqe = irdma_qp_get_next_send_wqe(qp, &wqe_idx, quanta, total_size, info); if (!wqe) return -ENOMEM; @@ -705,7 +744,7 @@ int irdma_uk_inline_rdma_write(struct irdma_qp_uk *qp, hdr = FIELD_PREP(IRDMAQPSQ_REMSTAG, op_info->rem_addr.lkey) | FIELD_PREP(IRDMAQPSQ_OPCODE, info->op_type) | - FIELD_PREP(IRDMAQPSQ_INLINEDATALEN, op_info->len) | + FIELD_PREP(IRDMAQPSQ_INLINEDATALEN, total_size) | FIELD_PREP(IRDMAQPSQ_REPORTRTT, info->report_rtt ? 1 : 0) | FIELD_PREP(IRDMAQPSQ_INLINEDATAFLAG, 1) | FIELD_PREP(IRDMAQPSQ_IMMDATAFLAG, info->imm_data_valid ? 1 : 0) | @@ -719,7 +758,8 @@ int irdma_uk_inline_rdma_write(struct irdma_qp_uk *qp, set_64bit_val(wqe, 0, FIELD_PREP(IRDMAQPSQ_IMMDATA, info->imm_data)); - qp->wqe_ops.iw_copy_inline_data((u8 *)wqe, op_info->data, op_info->len, + qp->wqe_ops.iw_copy_inline_data((u8 *)wqe, op_info->lo_sg_list, + op_info->num_lo_sges, qp->swqe_polarity); dma_wmb(); /* make sure WQE is populated before valid bit is set */ @@ -745,20 +785,27 @@ int irdma_uk_inline_send(struct irdma_qp_uk *qp, struct irdma_post_sq_info *info, bool post_sq) { __le64 *wqe; - struct irdma_post_inline_send *op_info; + struct irdma_post_send *op_info; u64 hdr; u32 wqe_idx; bool read_fence = false; + u32 i, total_size = 0; u16 quanta; info->push_wqe = qp->push_db ? true : false; - op_info = &info->op.inline_send; + op_info = &info->op.send; + + if (unlikely(qp->max_sq_frag_cnt < op_info->num_sges)) + return -EINVAL; - if (op_info->len > qp->max_inline_data) + for (i = 0; i < op_info->num_sges; i++) + total_size += op_info->sg_list[i].length; + + if (unlikely(total_size > qp->max_inline_data)) return -EINVAL; - quanta = qp->wqe_ops.iw_inline_data_size_to_quanta(op_info->len); - wqe = irdma_qp_get_next_send_wqe(qp, &wqe_idx, quanta, op_info->len, + quanta = qp->wqe_ops.iw_inline_data_size_to_quanta(total_size); + wqe = irdma_qp_get_next_send_wqe(qp, &wqe_idx, quanta, total_size, info); if (!wqe) return -ENOMEM; @@ -773,7 +820,7 @@ int irdma_uk_inline_send(struct irdma_qp_uk *qp, hdr = FIELD_PREP(IRDMAQPSQ_REMSTAG, info->stag_to_inv) | FIELD_PREP(IRDMAQPSQ_AHID, op_info->ah_id) | FIELD_PREP(IRDMAQPSQ_OPCODE, info->op_type) | - FIELD_PREP(IRDMAQPSQ_INLINEDATALEN, op_info->len) | + FIELD_PREP(IRDMAQPSQ_INLINEDATALEN, total_size) | FIELD_PREP(IRDMAQPSQ_IMMDATAFLAG, (info->imm_data_valid ? 1 : 0)) | FIELD_PREP(IRDMAQPSQ_REPORTRTT, (info->report_rtt ? 1 : 0)) | @@ -789,8 +836,8 @@ int irdma_uk_inline_send(struct irdma_qp_uk *qp, if (info->imm_data_valid) set_64bit_val(wqe, 0, FIELD_PREP(IRDMAQPSQ_IMMDATA, info->imm_data)); - qp->wqe_ops.iw_copy_inline_data((u8 *)wqe, op_info->data, op_info->len, - qp->swqe_polarity); + qp->wqe_ops.iw_copy_inline_data((u8 *)wqe, op_info->sg_list, + op_info->num_sges, qp->swqe_polarity); dma_wmb(); /* make sure WQE is populated before valid bit is set */ @@ -1002,11 +1049,10 @@ int irdma_uk_cq_poll_cmpl(struct irdma_cq_uk *cq, __le64 *cqe; struct irdma_qp_uk *qp; struct irdma_ring *pring = NULL; - u32 wqe_idx, q_type; + u32 wqe_idx; int ret_code; bool move_cq_head = true; u8 polarity; - u8 op_type; bool ext_valid; __le64 *ext_cqe; @@ -1074,7 +1120,7 @@ int irdma_uk_cq_poll_cmpl(struct irdma_cq_uk *cq, info->ud_vlan_valid = false; } - q_type = (u8)FIELD_GET(IRDMA_CQ_SQ, qword3); + info->q_type = (u8)FIELD_GET(IRDMA_CQ_SQ, qword3); info->error = (bool)FIELD_GET(IRDMA_CQ_ERROR, qword3); info->push_dropped = (bool)FIELD_GET(IRDMACQ_PSHDROP, qword3); info->ipv4 = (bool)FIELD_GET(IRDMACQ_IPV4, qword3); @@ -1113,8 +1159,9 @@ int irdma_uk_cq_poll_cmpl(struct irdma_cq_uk *cq, } wqe_idx = (u32)FIELD_GET(IRDMA_CQ_WQEIDX, qword3); info->qp_handle = (irdma_qp_handle)(unsigned long)qp; + info->op_type = (u8)FIELD_GET(IRDMA_CQ_SQ, qword3); - if (q_type == IRDMA_CQE_QTYPE_RQ) { + if (info->q_type == IRDMA_CQE_QTYPE_RQ) { u32 array_idx; array_idx = wqe_idx / qp->rq_wqe_size_multiplier; @@ -1134,10 +1181,6 @@ int irdma_uk_cq_poll_cmpl(struct irdma_cq_uk *cq, info->bytes_xfered = (u32)FIELD_GET(IRDMACQ_PAYLDLEN, qword0); - if (info->imm_valid) - info->op_type = IRDMA_OP_TYPE_REC_IMM; - else - info->op_type = IRDMA_OP_TYPE_REC; if (qword3 & IRDMACQ_STAG) { info->stag_invalid_set = true; info->inv_stag = (u32)FIELD_GET(IRDMACQ_INVSTAG, qword2); @@ -1195,17 +1238,18 @@ int irdma_uk_cq_poll_cmpl(struct irdma_cq_uk *cq, sw_wqe = qp->sq_base[tail].elem; get_64bit_val(sw_wqe, 24, &wqe_qword); - op_type = (u8)FIELD_GET(IRDMAQPSQ_OPCODE, wqe_qword); - info->op_type = op_type; + info->op_type = (u8)FIELD_GET(IRDMAQPSQ_OPCODE, + wqe_qword); IRDMA_RING_SET_TAIL(qp->sq_ring, tail + qp->sq_wrtrk_array[tail].quanta); - if (op_type != IRDMAQP_OP_NOP) { + if (info->op_type != IRDMAQP_OP_NOP) { info->wr_id = qp->sq_wrtrk_array[tail].wrid; info->bytes_xfered = qp->sq_wrtrk_array[tail].wr_len; break; } } while (1); - if (op_type == IRDMA_OP_TYPE_BIND_MW && info->minor_err == FLUSH_PROT_ERR) + if (info->op_type == IRDMA_OP_TYPE_BIND_MW && + info->minor_err == FLUSH_PROT_ERR) info->minor_err = FLUSH_MW_BIND_ERR; qp->sq_flush_seen = true; if (!IRDMA_RING_MORE_WORK(qp->sq_ring)) diff --git a/drivers/infiniband/hw/irdma/user.h b/drivers/infiniband/hw/irdma/user.h index 2ef61923c926..d0cdf609f5e0 100644 --- a/drivers/infiniband/hw/irdma/user.h +++ b/drivers/infiniband/hw/irdma/user.h @@ -173,14 +173,6 @@ struct irdma_post_send { u32 ah_id; }; -struct irdma_post_inline_send { - void *data; - u32 len; - u32 qkey; - u32 dest_qp; - u32 ah_id; -}; - struct irdma_post_rq_info { u64 wr_id; struct ib_sge *sg_list; @@ -193,12 +185,6 @@ struct irdma_rdma_write { struct ib_sge rem_addr; }; -struct irdma_inline_rdma_write { - void *data; - u32 len; - struct ib_sge rem_addr; -}; - struct irdma_rdma_read { struct ib_sge *lo_sg_list; u32 num_lo_sges; @@ -241,8 +227,6 @@ struct irdma_post_sq_info { struct irdma_rdma_read rdma_read; struct irdma_bind_window bind_window; struct irdma_inv_local_stag inv_local_stag; - struct irdma_inline_rdma_write inline_rdma_write; - struct irdma_post_inline_send inline_send; } op; }; @@ -261,6 +245,7 @@ struct irdma_cq_poll_info { u16 ud_vlan; u8 ud_smac[6]; u8 op_type; + u8 q_type; bool stag_invalid_set:1; /* or L_R_Key set */ bool push_dropped:1; bool error:1; @@ -291,7 +276,8 @@ int irdma_uk_stag_local_invalidate(struct irdma_qp_uk *qp, bool post_sq); struct irdma_wqe_uk_ops { - void (*iw_copy_inline_data)(u8 *dest, u8 *src, u32 len, u8 polarity); + void (*iw_copy_inline_data)(u8 *dest, struct ib_sge *sge_list, + u32 num_sges, u8 polarity); u16 (*iw_inline_data_size_to_quanta)(u32 data_size); void (*iw_set_fragment)(__le64 *wqe, u32 offset, struct ib_sge *sge, u8 valid); diff --git a/drivers/infiniband/hw/irdma/utils.c b/drivers/infiniband/hw/irdma/utils.c index 8dfc9e154d73..445e69e86409 100644 --- a/drivers/infiniband/hw/irdma/utils.c +++ b/drivers/infiniband/hw/irdma/utils.c @@ -2591,6 +2591,7 @@ void irdma_generate_flush_completions(struct irdma_qp *iwqp) sw_wqe = qp->sq_base[wqe_idx].elem; get_64bit_val(sw_wqe, 24, &wqe_qword); cmpl->cpi.op_type = (u8)FIELD_GET(IRDMAQPSQ_OPCODE, IRDMAQPSQ_OPCODE); + cmpl->cpi.q_type = IRDMA_CQE_QTYPE_SQ; /* remove the SQ WR by moving SQ tail*/ IRDMA_RING_SET_TAIL(*sq_ring, sq_ring->tail + qp->sq_wrtrk_array[sq_ring->tail].quanta); @@ -2629,6 +2630,7 @@ void irdma_generate_flush_completions(struct irdma_qp *iwqp) cmpl->cpi.wr_id = qp->rq_wrid_array[wqe_idx]; cmpl->cpi.op_type = IRDMA_OP_TYPE_REC; + cmpl->cpi.q_type = IRDMA_CQE_QTYPE_RQ; /* remove the RQ WR by moving RQ tail */ IRDMA_RING_SET_TAIL(*rq_ring, rq_ring->tail + 1); ibdev_dbg(iwqp->iwrcq->ibcq.device, diff --git a/drivers/infiniband/hw/irdma/verbs.c b/drivers/infiniband/hw/irdma/verbs.c index a22afbb25bc5..f6973ea55eda 100644 --- a/drivers/infiniband/hw/irdma/verbs.c +++ b/drivers/infiniband/hw/irdma/verbs.c @@ -64,36 +64,6 @@ static int irdma_query_device(struct ib_device *ibdev, } /** - * irdma_get_eth_speed_and_width - Get IB port speed and width from netdev speed - * @link_speed: netdev phy link speed - * @active_speed: IB port speed - * @active_width: IB port width - */ -static void irdma_get_eth_speed_and_width(u32 link_speed, u16 *active_speed, - u8 *active_width) -{ - if (link_speed <= SPEED_1000) { - *active_width = IB_WIDTH_1X; - *active_speed = IB_SPEED_SDR; - } else if (link_speed <= SPEED_10000) { - *active_width = IB_WIDTH_1X; - *active_speed = IB_SPEED_FDR10; - } else if (link_speed <= SPEED_20000) { - *active_width = IB_WIDTH_4X; - *active_speed = IB_SPEED_DDR; - } else if (link_speed <= SPEED_25000) { - *active_width = IB_WIDTH_1X; - *active_speed = IB_SPEED_EDR; - } else if (link_speed <= SPEED_40000) { - *active_width = IB_WIDTH_4X; - *active_speed = IB_SPEED_FDR10; - } else { - *active_width = IB_WIDTH_4X; - *active_speed = IB_SPEED_EDR; - } -} - -/** * irdma_query_port - get port attributes * @ibdev: device pointer from stack * @port: port number for query @@ -120,8 +90,9 @@ static int irdma_query_port(struct ib_device *ibdev, u32 port, props->state = IB_PORT_DOWN; props->phys_state = IB_PORT_PHYS_STATE_DISABLED; } - irdma_get_eth_speed_and_width(SPEED_100000, &props->active_speed, - &props->active_width); + + ib_get_eth_speed(ibdev, port, &props->active_speed, + &props->active_width); if (rdma_protocol_roce(ibdev, 1)) { props->gid_tbl_len = 32; @@ -1242,6 +1213,7 @@ int irdma_modify_qp_roce(struct ib_qp *ibqp, struct ib_qp_attr *attr, av->attrs = attr->ah_attr; rdma_gid2ip((struct sockaddr *)&av->sgid_addr, &sgid_attr->gid); rdma_gid2ip((struct sockaddr *)&av->dgid_addr, &attr->ah_attr.grh.dgid); + av->net_type = rdma_gid_attr_network_type(sgid_attr); if (av->net_type == RDMA_NETWORK_IPV6) { __be32 *daddr = av->dgid_addr.saddr_in6.sin6_addr.in6_u.u6_addr32; @@ -2358,9 +2330,10 @@ static bool irdma_check_mr_contiguous(struct irdma_pble_alloc *palloc, * @rf: RDMA PCI function * @iwmr: mr pointer for this memory registration * @use_pbles: flag if to use pble's + * @lvl_1_only: request only level 1 pble if true */ static int irdma_setup_pbles(struct irdma_pci_f *rf, struct irdma_mr *iwmr, - bool use_pbles) + bool use_pbles, bool lvl_1_only) { struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; @@ -2371,7 +2344,7 @@ static int irdma_setup_pbles(struct irdma_pci_f *rf, struct irdma_mr *iwmr, if (use_pbles) { status = irdma_get_pble(rf->pble_rsrc, palloc, iwmr->page_cnt, - false); + lvl_1_only); if (status) return status; @@ -2414,16 +2387,10 @@ static int irdma_handle_q_mem(struct irdma_device *iwdev, bool ret = true; pg_size = iwmr->page_size; - err = irdma_setup_pbles(iwdev->rf, iwmr, use_pbles); + err = irdma_setup_pbles(iwdev->rf, iwmr, use_pbles, true); if (err) return err; - if (use_pbles && palloc->level != PBLE_LEVEL_1) { - irdma_free_pble(iwdev->rf->pble_rsrc, palloc); - iwpbl->pbl_allocated = false; - return -ENOMEM; - } - if (use_pbles) arr = palloc->level1.addr; @@ -2899,7 +2866,7 @@ static struct ib_mr *irdma_reg_user_mr(struct ib_pd *pd, u64 start, u64 len, case IRDMA_MEMREG_TYPE_MEM: use_pbles = (iwmr->page_cnt != 1); - err = irdma_setup_pbles(iwdev->rf, iwmr, use_pbles); + err = irdma_setup_pbles(iwdev->rf, iwmr, use_pbles, false); if (err) goto error; @@ -3165,30 +3132,20 @@ static int irdma_post_send(struct ib_qp *ibqp, info.stag_to_inv = ib_wr->ex.invalidate_rkey; } - if (ib_wr->send_flags & IB_SEND_INLINE) { - info.op.inline_send.data = (void *)(unsigned long) - ib_wr->sg_list[0].addr; - info.op.inline_send.len = ib_wr->sg_list[0].length; - if (iwqp->ibqp.qp_type == IB_QPT_UD || - iwqp->ibqp.qp_type == IB_QPT_GSI) { - ah = to_iwah(ud_wr(ib_wr)->ah); - info.op.inline_send.ah_id = ah->sc_ah.ah_info.ah_idx; - info.op.inline_send.qkey = ud_wr(ib_wr)->remote_qkey; - info.op.inline_send.dest_qp = ud_wr(ib_wr)->remote_qpn; - } + info.op.send.num_sges = ib_wr->num_sge; + info.op.send.sg_list = ib_wr->sg_list; + if (iwqp->ibqp.qp_type == IB_QPT_UD || + iwqp->ibqp.qp_type == IB_QPT_GSI) { + ah = to_iwah(ud_wr(ib_wr)->ah); + info.op.send.ah_id = ah->sc_ah.ah_info.ah_idx; + info.op.send.qkey = ud_wr(ib_wr)->remote_qkey; + info.op.send.dest_qp = ud_wr(ib_wr)->remote_qpn; + } + + if (ib_wr->send_flags & IB_SEND_INLINE) err = irdma_uk_inline_send(ukqp, &info, false); - } else { - info.op.send.num_sges = ib_wr->num_sge; - info.op.send.sg_list = ib_wr->sg_list; - if (iwqp->ibqp.qp_type == IB_QPT_UD || - iwqp->ibqp.qp_type == IB_QPT_GSI) { - ah = to_iwah(ud_wr(ib_wr)->ah); - info.op.send.ah_id = ah->sc_ah.ah_info.ah_idx; - info.op.send.qkey = ud_wr(ib_wr)->remote_qkey; - info.op.send.dest_qp = ud_wr(ib_wr)->remote_qpn; - } + else err = irdma_uk_send(ukqp, &info, false); - } break; case IB_WR_RDMA_WRITE_WITH_IMM: if (ukqp->qp_caps & IRDMA_WRITE_WITH_IMM) { @@ -3205,22 +3162,15 @@ static int irdma_post_send(struct ib_qp *ibqp, else info.op_type = IRDMA_OP_TYPE_RDMA_WRITE; - if (ib_wr->send_flags & IB_SEND_INLINE) { - info.op.inline_rdma_write.data = (void *)(uintptr_t)ib_wr->sg_list[0].addr; - info.op.inline_rdma_write.len = - ib_wr->sg_list[0].length; - info.op.inline_rdma_write.rem_addr.addr = - rdma_wr(ib_wr)->remote_addr; - info.op.inline_rdma_write.rem_addr.lkey = - rdma_wr(ib_wr)->rkey; + info.op.rdma_write.num_lo_sges = ib_wr->num_sge; + info.op.rdma_write.lo_sg_list = ib_wr->sg_list; + info.op.rdma_write.rem_addr.addr = + rdma_wr(ib_wr)->remote_addr; + info.op.rdma_write.rem_addr.lkey = rdma_wr(ib_wr)->rkey; + if (ib_wr->send_flags & IB_SEND_INLINE) err = irdma_uk_inline_rdma_write(ukqp, &info, false); - } else { - info.op.rdma_write.lo_sg_list = (void *)ib_wr->sg_list; - info.op.rdma_write.num_lo_sges = ib_wr->num_sge; - info.op.rdma_write.rem_addr.addr = rdma_wr(ib_wr)->remote_addr; - info.op.rdma_write.rem_addr.lkey = rdma_wr(ib_wr)->rkey; + else err = irdma_uk_rdma_write(ukqp, &info, false); - } break; case IB_WR_RDMA_READ_WITH_INV: inv_stag = true; @@ -3380,7 +3330,6 @@ static enum ib_wc_status irdma_flush_err_to_ib_wc_status(enum irdma_flush_opcode static void irdma_process_cqe(struct ib_wc *entry, struct irdma_cq_poll_info *cq_poll_info) { - struct irdma_qp *iwqp; struct irdma_sc_qp *qp; entry->wc_flags = 0; @@ -3388,7 +3337,6 @@ static void irdma_process_cqe(struct ib_wc *entry, entry->wr_id = cq_poll_info->wr_id; qp = cq_poll_info->qp_handle; - iwqp = qp->qp_uk.back_qp; entry->qp = qp->qp_uk.back_qp; if (cq_poll_info->error) { @@ -3421,42 +3369,17 @@ static void irdma_process_cqe(struct ib_wc *entry, } } - switch (cq_poll_info->op_type) { - case IRDMA_OP_TYPE_RDMA_WRITE: - case IRDMA_OP_TYPE_RDMA_WRITE_SOL: - entry->opcode = IB_WC_RDMA_WRITE; - break; - case IRDMA_OP_TYPE_RDMA_READ_INV_STAG: - case IRDMA_OP_TYPE_RDMA_READ: - entry->opcode = IB_WC_RDMA_READ; - break; - case IRDMA_OP_TYPE_SEND_INV: - case IRDMA_OP_TYPE_SEND_SOL: - case IRDMA_OP_TYPE_SEND_SOL_INV: - case IRDMA_OP_TYPE_SEND: - entry->opcode = IB_WC_SEND; - break; - case IRDMA_OP_TYPE_FAST_REG_NSMR: - entry->opcode = IB_WC_REG_MR; - break; - case IRDMA_OP_TYPE_INV_STAG: - entry->opcode = IB_WC_LOCAL_INV; - break; - case IRDMA_OP_TYPE_REC_IMM: - case IRDMA_OP_TYPE_REC: - entry->opcode = cq_poll_info->op_type == IRDMA_OP_TYPE_REC_IMM ? - IB_WC_RECV_RDMA_WITH_IMM : IB_WC_RECV; + if (cq_poll_info->q_type == IRDMA_CQE_QTYPE_SQ) { + set_ib_wc_op_sq(cq_poll_info, entry); + } else { + set_ib_wc_op_rq(cq_poll_info, entry, + qp->qp_uk.qp_caps & IRDMA_SEND_WITH_IMM ? + true : false); if (qp->qp_uk.qp_type != IRDMA_QP_TYPE_ROCE_UD && cq_poll_info->stag_invalid_set) { entry->ex.invalidate_rkey = cq_poll_info->inv_stag; entry->wc_flags |= IB_WC_WITH_INVALIDATE; } - break; - default: - ibdev_err(&iwqp->iwdev->ibdev, - "Invalid opcode = %d in CQE\n", cq_poll_info->op_type); - entry->status = IB_WC_GENERAL_ERR; - return; } if (qp->qp_uk.qp_type == IRDMA_QP_TYPE_ROCE_UD) { diff --git a/drivers/infiniband/hw/irdma/verbs.h b/drivers/infiniband/hw/irdma/verbs.h index 4309b7159f42..a536e9fa85eb 100644 --- a/drivers/infiniband/hw/irdma/verbs.h +++ b/drivers/infiniband/hw/irdma/verbs.h @@ -232,6 +232,59 @@ static inline u16 irdma_fw_minor_ver(struct irdma_sc_dev *dev) return (u16)FIELD_GET(IRDMA_FW_VER_MINOR, dev->feature_info[IRDMA_FEATURE_FW_INFO]); } +static inline void set_ib_wc_op_sq(struct irdma_cq_poll_info *cq_poll_info, + struct ib_wc *entry) +{ + switch (cq_poll_info->op_type) { + case IRDMA_OP_TYPE_RDMA_WRITE: + case IRDMA_OP_TYPE_RDMA_WRITE_SOL: + entry->opcode = IB_WC_RDMA_WRITE; + break; + case IRDMA_OP_TYPE_RDMA_READ_INV_STAG: + case IRDMA_OP_TYPE_RDMA_READ: + entry->opcode = IB_WC_RDMA_READ; + break; + case IRDMA_OP_TYPE_SEND_SOL: + case IRDMA_OP_TYPE_SEND_SOL_INV: + case IRDMA_OP_TYPE_SEND_INV: + case IRDMA_OP_TYPE_SEND: + entry->opcode = IB_WC_SEND; + break; + case IRDMA_OP_TYPE_FAST_REG_NSMR: + entry->opcode = IB_WC_REG_MR; + break; + case IRDMA_OP_TYPE_INV_STAG: + entry->opcode = IB_WC_LOCAL_INV; + break; + default: + entry->status = IB_WC_GENERAL_ERR; + } +} + +static inline void set_ib_wc_op_rq(struct irdma_cq_poll_info *cq_poll_info, + struct ib_wc *entry, bool send_imm_support) +{ + /** + * iWARP does not support sendImm, so the presence of Imm data + * must be WriteImm. + */ + if (!send_imm_support) { + entry->opcode = cq_poll_info->imm_valid ? + IB_WC_RECV_RDMA_WITH_IMM : + IB_WC_RECV; + return; + } + + switch (cq_poll_info->op_type) { + case IB_OPCODE_RDMA_WRITE_ONLY_WITH_IMMEDIATE: + case IB_OPCODE_RDMA_WRITE_LAST_WITH_IMMEDIATE: + entry->opcode = IB_WC_RECV_RDMA_WITH_IMM; + break; + default: + entry->opcode = IB_WC_RECV; + } +} + void irdma_mcast_mac(u32 *ip_addr, u8 *mac, bool ipv4); int irdma_ib_register_device(struct irdma_device *iwdev); void irdma_ib_unregister_device(struct irdma_device *iwdev); diff --git a/drivers/infiniband/hw/mana/Kconfig b/drivers/infiniband/hw/mana/Kconfig new file mode 100644 index 000000000000..546640657bac --- /dev/null +++ b/drivers/infiniband/hw/mana/Kconfig @@ -0,0 +1,10 @@ +# SPDX-License-Identifier: GPL-2.0-only +config MANA_INFINIBAND + tristate "Microsoft Azure Network Adapter support" + depends on NETDEVICES && ETHERNET && PCI && MICROSOFT_MANA + help + This driver provides low-level RDMA support for Microsoft Azure + Network Adapter (MANA). MANA supports RDMA features that can be used + for workloads (e.g. DPDK, MPI etc) that uses RDMA verbs to directly + access hardware from user-mode processes in Microsoft Azure cloud + environment. diff --git a/drivers/infiniband/hw/mana/Makefile b/drivers/infiniband/hw/mana/Makefile new file mode 100644 index 000000000000..88655fe5e398 --- /dev/null +++ b/drivers/infiniband/hw/mana/Makefile @@ -0,0 +1,4 @@ +# SPDX-License-Identifier: GPL-2.0-only +obj-$(CONFIG_MANA_INFINIBAND) += mana_ib.o + +mana_ib-y := device.o main.o wq.o qp.o cq.o mr.o diff --git a/drivers/infiniband/hw/mana/cq.c b/drivers/infiniband/hw/mana/cq.c new file mode 100644 index 000000000000..d141cab8a1e6 --- /dev/null +++ b/drivers/infiniband/hw/mana/cq.c @@ -0,0 +1,79 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" + +int mana_ib_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr, + struct ib_udata *udata) +{ + struct mana_ib_cq *cq = container_of(ibcq, struct mana_ib_cq, ibcq); + struct ib_device *ibdev = ibcq->device; + struct mana_ib_create_cq ucmd = {}; + struct mana_ib_dev *mdev; + int err; + + mdev = container_of(ibdev, struct mana_ib_dev, ib_dev); + + if (udata->inlen < sizeof(ucmd)) + return -EINVAL; + + err = ib_copy_from_udata(&ucmd, udata, min(sizeof(ucmd), udata->inlen)); + if (err) { + ibdev_dbg(ibdev, + "Failed to copy from udata for create cq, %d\n", err); + return err; + } + + if (attr->cqe > MAX_SEND_BUFFERS_PER_QUEUE) { + ibdev_dbg(ibdev, "CQE %d exceeding limit\n", attr->cqe); + return -EINVAL; + } + + cq->cqe = attr->cqe; + cq->umem = ib_umem_get(ibdev, ucmd.buf_addr, cq->cqe * COMP_ENTRY_SIZE, + IB_ACCESS_LOCAL_WRITE); + if (IS_ERR(cq->umem)) { + err = PTR_ERR(cq->umem); + ibdev_dbg(ibdev, "Failed to get umem for create cq, err %d\n", + err); + return err; + } + + err = mana_ib_gd_create_dma_region(mdev, cq->umem, &cq->gdma_region); + if (err) { + ibdev_dbg(ibdev, + "Failed to create dma region for create cq, %d\n", + err); + goto err_release_umem; + } + + ibdev_dbg(ibdev, + "mana_ib_gd_create_dma_region ret %d gdma_region 0x%llx\n", + err, cq->gdma_region); + + /* + * The CQ ID is not known at this time. The ID is generated at create_qp + */ + + return 0; + +err_release_umem: + ib_umem_release(cq->umem); + return err; +} + +int mana_ib_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata) +{ + struct mana_ib_cq *cq = container_of(ibcq, struct mana_ib_cq, ibcq); + struct ib_device *ibdev = ibcq->device; + struct mana_ib_dev *mdev; + + mdev = container_of(ibdev, struct mana_ib_dev, ib_dev); + + mana_ib_gd_destroy_dma_region(mdev, cq->gdma_region); + ib_umem_release(cq->umem); + + return 0; +} diff --git a/drivers/infiniband/hw/mana/device.c b/drivers/infiniband/hw/mana/device.c new file mode 100644 index 000000000000..d4541b8707e4 --- /dev/null +++ b/drivers/infiniband/hw/mana/device.c @@ -0,0 +1,117 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" +#include <net/mana/mana_auxiliary.h> + +MODULE_DESCRIPTION("Microsoft Azure Network Adapter IB driver"); +MODULE_LICENSE("GPL"); +MODULE_IMPORT_NS(NET_MANA); + +static const struct ib_device_ops mana_ib_dev_ops = { + .owner = THIS_MODULE, + .driver_id = RDMA_DRIVER_MANA, + .uverbs_abi_ver = MANA_IB_UVERBS_ABI_VERSION, + + .alloc_pd = mana_ib_alloc_pd, + .alloc_ucontext = mana_ib_alloc_ucontext, + .create_cq = mana_ib_create_cq, + .create_qp = mana_ib_create_qp, + .create_rwq_ind_table = mana_ib_create_rwq_ind_table, + .create_wq = mana_ib_create_wq, + .dealloc_pd = mana_ib_dealloc_pd, + .dealloc_ucontext = mana_ib_dealloc_ucontext, + .dereg_mr = mana_ib_dereg_mr, + .destroy_cq = mana_ib_destroy_cq, + .destroy_qp = mana_ib_destroy_qp, + .destroy_rwq_ind_table = mana_ib_destroy_rwq_ind_table, + .destroy_wq = mana_ib_destroy_wq, + .disassociate_ucontext = mana_ib_disassociate_ucontext, + .get_port_immutable = mana_ib_get_port_immutable, + .mmap = mana_ib_mmap, + .modify_qp = mana_ib_modify_qp, + .modify_wq = mana_ib_modify_wq, + .query_device = mana_ib_query_device, + .query_gid = mana_ib_query_gid, + .query_port = mana_ib_query_port, + .reg_user_mr = mana_ib_reg_user_mr, + + INIT_RDMA_OBJ_SIZE(ib_cq, mana_ib_cq, ibcq), + INIT_RDMA_OBJ_SIZE(ib_pd, mana_ib_pd, ibpd), + INIT_RDMA_OBJ_SIZE(ib_qp, mana_ib_qp, ibqp), + INIT_RDMA_OBJ_SIZE(ib_ucontext, mana_ib_ucontext, ibucontext), + INIT_RDMA_OBJ_SIZE(ib_rwq_ind_table, mana_ib_rwq_ind_table, + ib_ind_table), +}; + +static int mana_ib_probe(struct auxiliary_device *adev, + const struct auxiliary_device_id *id) +{ + struct mana_adev *madev = container_of(adev, struct mana_adev, adev); + struct gdma_dev *mdev = madev->mdev; + struct mana_context *mc; + struct mana_ib_dev *dev; + int ret; + + mc = mdev->driver_data; + + dev = ib_alloc_device(mana_ib_dev, ib_dev); + if (!dev) + return -ENOMEM; + + ib_set_device_ops(&dev->ib_dev, &mana_ib_dev_ops); + + dev->ib_dev.phys_port_cnt = mc->num_ports; + + ibdev_dbg(&dev->ib_dev, "mdev=%p id=%d num_ports=%d\n", mdev, + mdev->dev_id.as_uint32, dev->ib_dev.phys_port_cnt); + + dev->gdma_dev = mdev; + dev->ib_dev.node_type = RDMA_NODE_IB_CA; + + /* + * num_comp_vectors needs to set to the max MSIX index + * when interrupts and event queues are implemented + */ + dev->ib_dev.num_comp_vectors = 1; + dev->ib_dev.dev.parent = mdev->gdma_context->dev; + + ret = ib_register_device(&dev->ib_dev, "mana_%d", + mdev->gdma_context->dev); + if (ret) { + ib_dealloc_device(&dev->ib_dev); + return ret; + } + + dev_set_drvdata(&adev->dev, dev); + + return 0; +} + +static void mana_ib_remove(struct auxiliary_device *adev) +{ + struct mana_ib_dev *dev = dev_get_drvdata(&adev->dev); + + ib_unregister_device(&dev->ib_dev); + ib_dealloc_device(&dev->ib_dev); +} + +static const struct auxiliary_device_id mana_id_table[] = { + { + .name = "mana.rdma", + }, + {}, +}; + +MODULE_DEVICE_TABLE(auxiliary, mana_id_table); + +static struct auxiliary_driver mana_driver = { + .name = "rdma", + .probe = mana_ib_probe, + .remove = mana_ib_remove, + .id_table = mana_id_table, +}; + +module_auxiliary_driver(mana_driver); diff --git a/drivers/infiniband/hw/mana/main.c b/drivers/infiniband/hw/mana/main.c new file mode 100644 index 000000000000..8b3bc302d6f3 --- /dev/null +++ b/drivers/infiniband/hw/mana/main.c @@ -0,0 +1,521 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" + +void mana_ib_uncfg_vport(struct mana_ib_dev *dev, struct mana_ib_pd *pd, + u32 port) +{ + struct gdma_dev *gd = dev->gdma_dev; + struct mana_port_context *mpc; + struct net_device *ndev; + struct mana_context *mc; + + mc = gd->driver_data; + ndev = mc->ports[port]; + mpc = netdev_priv(ndev); + + mutex_lock(&pd->vport_mutex); + + pd->vport_use_count--; + WARN_ON(pd->vport_use_count < 0); + + if (!pd->vport_use_count) + mana_uncfg_vport(mpc); + + mutex_unlock(&pd->vport_mutex); +} + +int mana_ib_cfg_vport(struct mana_ib_dev *dev, u32 port, struct mana_ib_pd *pd, + u32 doorbell_id) +{ + struct gdma_dev *mdev = dev->gdma_dev; + struct mana_port_context *mpc; + struct mana_context *mc; + struct net_device *ndev; + int err; + + mc = mdev->driver_data; + ndev = mc->ports[port]; + mpc = netdev_priv(ndev); + + mutex_lock(&pd->vport_mutex); + + pd->vport_use_count++; + if (pd->vport_use_count > 1) { + ibdev_dbg(&dev->ib_dev, + "Skip as this PD is already configured vport\n"); + mutex_unlock(&pd->vport_mutex); + return 0; + } + + err = mana_cfg_vport(mpc, pd->pdn, doorbell_id); + if (err) { + pd->vport_use_count--; + mutex_unlock(&pd->vport_mutex); + + ibdev_dbg(&dev->ib_dev, "Failed to configure vPort %d\n", err); + return err; + } + + mutex_unlock(&pd->vport_mutex); + + pd->tx_shortform_allowed = mpc->tx_shortform_allowed; + pd->tx_vp_offset = mpc->tx_vp_offset; + + ibdev_dbg(&dev->ib_dev, "vport handle %llx pdid %x doorbell_id %x\n", + mpc->port_handle, pd->pdn, doorbell_id); + + return 0; +} + +int mana_ib_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata) +{ + struct mana_ib_pd *pd = container_of(ibpd, struct mana_ib_pd, ibpd); + struct ib_device *ibdev = ibpd->device; + struct gdma_create_pd_resp resp = {}; + struct gdma_create_pd_req req = {}; + enum gdma_pd_flags flags = 0; + struct mana_ib_dev *dev; + struct gdma_dev *mdev; + int err; + + dev = container_of(ibdev, struct mana_ib_dev, ib_dev); + mdev = dev->gdma_dev; + + mana_gd_init_req_hdr(&req.hdr, GDMA_CREATE_PD, sizeof(req), + sizeof(resp)); + + req.flags = flags; + err = mana_gd_send_request(mdev->gdma_context, sizeof(req), &req, + sizeof(resp), &resp); + + if (err || resp.hdr.status) { + ibdev_dbg(&dev->ib_dev, + "Failed to get pd_id err %d status %u\n", err, + resp.hdr.status); + if (!err) + err = -EPROTO; + + return err; + } + + pd->pd_handle = resp.pd_handle; + pd->pdn = resp.pd_id; + ibdev_dbg(&dev->ib_dev, "pd_handle 0x%llx pd_id %d\n", + pd->pd_handle, pd->pdn); + + mutex_init(&pd->vport_mutex); + pd->vport_use_count = 0; + return 0; +} + +int mana_ib_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata) +{ + struct mana_ib_pd *pd = container_of(ibpd, struct mana_ib_pd, ibpd); + struct ib_device *ibdev = ibpd->device; + struct gdma_destory_pd_resp resp = {}; + struct gdma_destroy_pd_req req = {}; + struct mana_ib_dev *dev; + struct gdma_dev *mdev; + int err; + + dev = container_of(ibdev, struct mana_ib_dev, ib_dev); + mdev = dev->gdma_dev; + + mana_gd_init_req_hdr(&req.hdr, GDMA_DESTROY_PD, sizeof(req), + sizeof(resp)); + + req.pd_handle = pd->pd_handle; + err = mana_gd_send_request(mdev->gdma_context, sizeof(req), &req, + sizeof(resp), &resp); + + if (err || resp.hdr.status) { + ibdev_dbg(&dev->ib_dev, + "Failed to destroy pd_handle 0x%llx err %d status %u", + pd->pd_handle, err, resp.hdr.status); + if (!err) + err = -EPROTO; + } + + return err; +} + +static int mana_gd_destroy_doorbell_page(struct gdma_context *gc, + int doorbell_page) +{ + struct gdma_destroy_resource_range_req req = {}; + struct gdma_resp_hdr resp = {}; + int err; + + mana_gd_init_req_hdr(&req.hdr, GDMA_DESTROY_RESOURCE_RANGE, + sizeof(req), sizeof(resp)); + + req.resource_type = GDMA_RESOURCE_DOORBELL_PAGE; + req.num_resources = 1; + req.allocated_resources = doorbell_page; + + err = mana_gd_send_request(gc, sizeof(req), &req, sizeof(resp), &resp); + if (err || resp.status) { + dev_err(gc->dev, + "Failed to destroy doorbell page: ret %d, 0x%x\n", + err, resp.status); + return err ?: -EPROTO; + } + + return 0; +} + +static int mana_gd_allocate_doorbell_page(struct gdma_context *gc, + int *doorbell_page) +{ + struct gdma_allocate_resource_range_req req = {}; + struct gdma_allocate_resource_range_resp resp = {}; + int err; + + mana_gd_init_req_hdr(&req.hdr, GDMA_ALLOCATE_RESOURCE_RANGE, + sizeof(req), sizeof(resp)); + + req.resource_type = GDMA_RESOURCE_DOORBELL_PAGE; + req.num_resources = 1; + req.alignment = 1; + + /* Have GDMA start searching from 0 */ + req.allocated_resources = 0; + + err = mana_gd_send_request(gc, sizeof(req), &req, sizeof(resp), &resp); + if (err || resp.hdr.status) { + dev_err(gc->dev, + "Failed to allocate doorbell page: ret %d, 0x%x\n", + err, resp.hdr.status); + return err ?: -EPROTO; + } + + *doorbell_page = resp.allocated_resources; + + return 0; +} + +int mana_ib_alloc_ucontext(struct ib_ucontext *ibcontext, + struct ib_udata *udata) +{ + struct mana_ib_ucontext *ucontext = + container_of(ibcontext, struct mana_ib_ucontext, ibucontext); + struct ib_device *ibdev = ibcontext->device; + struct mana_ib_dev *mdev; + struct gdma_context *gc; + struct gdma_dev *dev; + int doorbell_page; + int ret; + + mdev = container_of(ibdev, struct mana_ib_dev, ib_dev); + dev = mdev->gdma_dev; + gc = dev->gdma_context; + + /* Allocate a doorbell page index */ + ret = mana_gd_allocate_doorbell_page(gc, &doorbell_page); + if (ret) { + ibdev_dbg(ibdev, "Failed to allocate doorbell page %d\n", ret); + return ret; + } + + ibdev_dbg(ibdev, "Doorbell page allocated %d\n", doorbell_page); + + ucontext->doorbell = doorbell_page; + + return 0; +} + +void mana_ib_dealloc_ucontext(struct ib_ucontext *ibcontext) +{ + struct mana_ib_ucontext *mana_ucontext = + container_of(ibcontext, struct mana_ib_ucontext, ibucontext); + struct ib_device *ibdev = ibcontext->device; + struct mana_ib_dev *mdev; + struct gdma_context *gc; + int ret; + + mdev = container_of(ibdev, struct mana_ib_dev, ib_dev); + gc = mdev->gdma_dev->gdma_context; + + ret = mana_gd_destroy_doorbell_page(gc, mana_ucontext->doorbell); + if (ret) + ibdev_dbg(ibdev, "Failed to destroy doorbell page %d\n", ret); +} + +static int +mana_ib_gd_first_dma_region(struct mana_ib_dev *dev, + struct gdma_context *gc, + struct gdma_create_dma_region_req *create_req, + size_t num_pages, mana_handle_t *gdma_region) +{ + struct gdma_create_dma_region_resp create_resp = {}; + unsigned int create_req_msg_size; + int err; + + create_req_msg_size = + struct_size(create_req, page_addr_list, num_pages); + create_req->page_addr_list_len = num_pages; + + err = mana_gd_send_request(gc, create_req_msg_size, create_req, + sizeof(create_resp), &create_resp); + if (err || create_resp.hdr.status) { + ibdev_dbg(&dev->ib_dev, + "Failed to create DMA region: %d, 0x%x\n", + err, create_resp.hdr.status); + if (!err) + err = -EPROTO; + + return err; + } + + *gdma_region = create_resp.dma_region_handle; + ibdev_dbg(&dev->ib_dev, "Created DMA region handle 0x%llx\n", + *gdma_region); + + return 0; +} + +static int +mana_ib_gd_add_dma_region(struct mana_ib_dev *dev, struct gdma_context *gc, + struct gdma_dma_region_add_pages_req *add_req, + unsigned int num_pages, u32 expected_status) +{ + unsigned int add_req_msg_size = + struct_size(add_req, page_addr_list, num_pages); + struct gdma_general_resp add_resp = {}; + int err; + + mana_gd_init_req_hdr(&add_req->hdr, GDMA_DMA_REGION_ADD_PAGES, + add_req_msg_size, sizeof(add_resp)); + add_req->page_addr_list_len = num_pages; + + err = mana_gd_send_request(gc, add_req_msg_size, add_req, + sizeof(add_resp), &add_resp); + if (err || add_resp.hdr.status != expected_status) { + ibdev_dbg(&dev->ib_dev, + "Failed to create DMA region: %d, 0x%x\n", + err, add_resp.hdr.status); + + if (!err) + err = -EPROTO; + + return err; + } + + return 0; +} + +int mana_ib_gd_create_dma_region(struct mana_ib_dev *dev, struct ib_umem *umem, + mana_handle_t *gdma_region) +{ + struct gdma_dma_region_add_pages_req *add_req = NULL; + size_t num_pages_processed = 0, num_pages_to_handle; + struct gdma_create_dma_region_req *create_req; + unsigned int create_req_msg_size; + struct hw_channel_context *hwc; + struct ib_block_iter biter; + size_t max_pgs_add_cmd = 0; + size_t max_pgs_create_cmd; + struct gdma_context *gc; + size_t num_pages_total; + struct gdma_dev *mdev; + unsigned long page_sz; + unsigned int tail = 0; + u64 *page_addr_list; + void *request_buf; + int err; + + mdev = dev->gdma_dev; + gc = mdev->gdma_context; + hwc = gc->hwc.driver_data; + + /* Hardware requires dma region to align to chosen page size */ + page_sz = ib_umem_find_best_pgsz(umem, PAGE_SZ_BM, 0); + if (!page_sz) { + ibdev_dbg(&dev->ib_dev, "failed to find page size.\n"); + return -ENOMEM; + } + num_pages_total = ib_umem_num_dma_blocks(umem, page_sz); + + max_pgs_create_cmd = + (hwc->max_req_msg_size - sizeof(*create_req)) / sizeof(u64); + num_pages_to_handle = + min_t(size_t, num_pages_total, max_pgs_create_cmd); + create_req_msg_size = + struct_size(create_req, page_addr_list, num_pages_to_handle); + + request_buf = kzalloc(hwc->max_req_msg_size, GFP_KERNEL); + if (!request_buf) + return -ENOMEM; + + create_req = request_buf; + mana_gd_init_req_hdr(&create_req->hdr, GDMA_CREATE_DMA_REGION, + create_req_msg_size, + sizeof(struct gdma_create_dma_region_resp)); + + create_req->length = umem->length; + create_req->offset_in_page = umem->address & (page_sz - 1); + create_req->gdma_page_type = order_base_2(page_sz) - PAGE_SHIFT; + create_req->page_count = num_pages_total; + + ibdev_dbg(&dev->ib_dev, "size_dma_region %lu num_pages_total %lu\n", + umem->length, num_pages_total); + + ibdev_dbg(&dev->ib_dev, "page_sz %lu offset_in_page %u\n", + page_sz, create_req->offset_in_page); + + ibdev_dbg(&dev->ib_dev, "num_pages_to_handle %lu, gdma_page_type %u", + num_pages_to_handle, create_req->gdma_page_type); + + page_addr_list = create_req->page_addr_list; + rdma_umem_for_each_dma_block(umem, &biter, page_sz) { + page_addr_list[tail++] = rdma_block_iter_dma_address(&biter); + if (tail < num_pages_to_handle) + continue; + + if (!num_pages_processed) { + /* First create message */ + err = mana_ib_gd_first_dma_region(dev, gc, create_req, + tail, gdma_region); + if (err) + goto out; + + max_pgs_add_cmd = (hwc->max_req_msg_size - + sizeof(*add_req)) / sizeof(u64); + + add_req = request_buf; + add_req->dma_region_handle = *gdma_region; + add_req->reserved3 = 0; + page_addr_list = add_req->page_addr_list; + } else { + /* Subsequent create messages */ + u32 expected_s = 0; + + if (num_pages_processed + num_pages_to_handle < + num_pages_total) + expected_s = GDMA_STATUS_MORE_ENTRIES; + + err = mana_ib_gd_add_dma_region(dev, gc, add_req, tail, + expected_s); + if (err) + break; + } + + num_pages_processed += tail; + tail = 0; + + /* The remaining pages to create */ + num_pages_to_handle = + min_t(size_t, + num_pages_total - num_pages_processed, + max_pgs_add_cmd); + } + + if (err) + mana_ib_gd_destroy_dma_region(dev, *gdma_region); + +out: + kfree(request_buf); + return err; +} + +int mana_ib_gd_destroy_dma_region(struct mana_ib_dev *dev, u64 gdma_region) +{ + struct gdma_dev *mdev = dev->gdma_dev; + struct gdma_context *gc; + + gc = mdev->gdma_context; + ibdev_dbg(&dev->ib_dev, "destroy dma region 0x%llx\n", gdma_region); + + return mana_gd_destroy_dma_region(gc, gdma_region); +} + +int mana_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma) +{ + struct mana_ib_ucontext *mana_ucontext = + container_of(ibcontext, struct mana_ib_ucontext, ibucontext); + struct ib_device *ibdev = ibcontext->device; + struct mana_ib_dev *mdev; + struct gdma_context *gc; + phys_addr_t pfn; + pgprot_t prot; + int ret; + + mdev = container_of(ibdev, struct mana_ib_dev, ib_dev); + gc = mdev->gdma_dev->gdma_context; + + if (vma->vm_pgoff != 0) { + ibdev_dbg(ibdev, "Unexpected vm_pgoff %lu\n", vma->vm_pgoff); + return -EINVAL; + } + + /* Map to the page indexed by ucontext->doorbell */ + pfn = (gc->phys_db_page_base + + gc->db_page_size * mana_ucontext->doorbell) >> + PAGE_SHIFT; + prot = pgprot_writecombine(vma->vm_page_prot); + + ret = rdma_user_mmap_io(ibcontext, vma, pfn, gc->db_page_size, prot, + NULL); + if (ret) + ibdev_dbg(ibdev, "can't rdma_user_mmap_io ret %d\n", ret); + else + ibdev_dbg(ibdev, "mapped I/O pfn 0x%llx page_size %u, ret %d\n", + pfn, gc->db_page_size, ret); + + return ret; +} + +int mana_ib_get_port_immutable(struct ib_device *ibdev, u32 port_num, + struct ib_port_immutable *immutable) +{ + /* + * This version only support RAW_PACKET + * other values need to be filled for other types + */ + immutable->core_cap_flags = RDMA_CORE_PORT_RAW_PACKET; + + return 0; +} + +int mana_ib_query_device(struct ib_device *ibdev, struct ib_device_attr *props, + struct ib_udata *uhw) +{ + props->max_qp = MANA_MAX_NUM_QUEUES; + props->max_qp_wr = MAX_SEND_BUFFERS_PER_QUEUE; + + /* + * max_cqe could be potentially much bigger. + * As this version of driver only support RAW QP, set it to the same + * value as max_qp_wr + */ + props->max_cqe = MAX_SEND_BUFFERS_PER_QUEUE; + + props->max_mr_size = MANA_IB_MAX_MR_SIZE; + props->max_mr = MANA_IB_MAX_MR; + props->max_send_sge = MAX_TX_WQE_SGL_ENTRIES; + props->max_recv_sge = MAX_RX_WQE_SGL_ENTRIES; + + return 0; +} + +int mana_ib_query_port(struct ib_device *ibdev, u32 port, + struct ib_port_attr *props) +{ + /* This version doesn't return port properties */ + return 0; +} + +int mana_ib_query_gid(struct ib_device *ibdev, u32 port, int index, + union ib_gid *gid) +{ + /* This version doesn't return GID properties */ + return 0; +} + +void mana_ib_disassociate_ucontext(struct ib_ucontext *ibcontext) +{ +} diff --git a/drivers/infiniband/hw/mana/mana_ib.h b/drivers/infiniband/hw/mana/mana_ib.h new file mode 100644 index 000000000000..502cc8672eef --- /dev/null +++ b/drivers/infiniband/hw/mana/mana_ib.h @@ -0,0 +1,162 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 Microsoft Corporation. All rights reserved. + */ + +#ifndef _MANA_IB_H_ +#define _MANA_IB_H_ + +#include <rdma/ib_verbs.h> +#include <rdma/ib_mad.h> +#include <rdma/ib_umem.h> +#include <rdma/mana-abi.h> +#include <rdma/uverbs_ioctl.h> + +#include <net/mana/mana.h> + +#define PAGE_SZ_BM \ + (SZ_4K | SZ_8K | SZ_16K | SZ_32K | SZ_64K | SZ_128K | SZ_256K | \ + SZ_512K | SZ_1M | SZ_2M) + +/* MANA doesn't have any limit for MR size */ +#define MANA_IB_MAX_MR_SIZE U64_MAX + +/* + * The hardware limit of number of MRs is greater than maximum number of MRs + * that can possibly represent in 24 bits + */ +#define MANA_IB_MAX_MR 0xFFFFFFu + +struct mana_ib_dev { + struct ib_device ib_dev; + struct gdma_dev *gdma_dev; +}; + +struct mana_ib_wq { + struct ib_wq ibwq; + struct ib_umem *umem; + int wqe; + u32 wq_buf_size; + u64 gdma_region; + u64 id; + mana_handle_t rx_object; +}; + +struct mana_ib_pd { + struct ib_pd ibpd; + u32 pdn; + mana_handle_t pd_handle; + + /* Mutex for sharing access to vport_use_count */ + struct mutex vport_mutex; + int vport_use_count; + + bool tx_shortform_allowed; + u32 tx_vp_offset; +}; + +struct mana_ib_mr { + struct ib_mr ibmr; + struct ib_umem *umem; + mana_handle_t mr_handle; +}; + +struct mana_ib_cq { + struct ib_cq ibcq; + struct ib_umem *umem; + int cqe; + u64 gdma_region; + u64 id; +}; + +struct mana_ib_qp { + struct ib_qp ibqp; + + /* Work queue info */ + struct ib_umem *sq_umem; + int sqe; + u64 sq_gdma_region; + u64 sq_id; + mana_handle_t tx_object; + + /* The port on the IB device, starting with 1 */ + u32 port; +}; + +struct mana_ib_ucontext { + struct ib_ucontext ibucontext; + u32 doorbell; +}; + +struct mana_ib_rwq_ind_table { + struct ib_rwq_ind_table ib_ind_table; +}; + +int mana_ib_gd_create_dma_region(struct mana_ib_dev *dev, struct ib_umem *umem, + mana_handle_t *gdma_region); + +int mana_ib_gd_destroy_dma_region(struct mana_ib_dev *dev, + mana_handle_t gdma_region); + +struct ib_wq *mana_ib_create_wq(struct ib_pd *pd, + struct ib_wq_init_attr *init_attr, + struct ib_udata *udata); + +int mana_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr, + u32 wq_attr_mask, struct ib_udata *udata); + +int mana_ib_destroy_wq(struct ib_wq *ibwq, struct ib_udata *udata); + +int mana_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); + +int mana_ib_destroy_rwq_ind_table(struct ib_rwq_ind_table *ib_rwq_ind_tbl); + +struct ib_mr *mana_ib_get_dma_mr(struct ib_pd *ibpd, int access_flags); + +struct ib_mr *mana_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, + u64 iova, int access_flags, + struct ib_udata *udata); + +int mana_ib_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata); + +int mana_ib_create_qp(struct ib_qp *qp, struct ib_qp_init_attr *qp_init_attr, + struct ib_udata *udata); + +int mana_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, + int attr_mask, struct ib_udata *udata); + +int mana_ib_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata); + +int mana_ib_cfg_vport(struct mana_ib_dev *dev, u32 port_id, + struct mana_ib_pd *pd, u32 doorbell_id); +void mana_ib_uncfg_vport(struct mana_ib_dev *dev, struct mana_ib_pd *pd, + u32 port); + +int mana_ib_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr, + struct ib_udata *udata); + +int mana_ib_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata); + +int mana_ib_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata); +int mana_ib_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata); + +int mana_ib_alloc_ucontext(struct ib_ucontext *ibcontext, + struct ib_udata *udata); +void mana_ib_dealloc_ucontext(struct ib_ucontext *ibcontext); + +int mana_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma); + +int mana_ib_get_port_immutable(struct ib_device *ibdev, u32 port_num, + struct ib_port_immutable *immutable); +int mana_ib_query_device(struct ib_device *ibdev, struct ib_device_attr *props, + struct ib_udata *uhw); +int mana_ib_query_port(struct ib_device *ibdev, u32 port, + struct ib_port_attr *props); +int mana_ib_query_gid(struct ib_device *ibdev, u32 port, int index, + union ib_gid *gid); + +void mana_ib_disassociate_ucontext(struct ib_ucontext *ibcontext); + +#endif diff --git a/drivers/infiniband/hw/mana/mr.c b/drivers/infiniband/hw/mana/mr.c new file mode 100644 index 000000000000..351207c60eb6 --- /dev/null +++ b/drivers/infiniband/hw/mana/mr.c @@ -0,0 +1,197 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" + +#define VALID_MR_FLAGS \ + (IB_ACCESS_LOCAL_WRITE | IB_ACCESS_REMOTE_WRITE | IB_ACCESS_REMOTE_READ) + +static enum gdma_mr_access_flags +mana_ib_verbs_to_gdma_access_flags(int access_flags) +{ + enum gdma_mr_access_flags flags = GDMA_ACCESS_FLAG_LOCAL_READ; + + if (access_flags & IB_ACCESS_LOCAL_WRITE) + flags |= GDMA_ACCESS_FLAG_LOCAL_WRITE; + + if (access_flags & IB_ACCESS_REMOTE_WRITE) + flags |= GDMA_ACCESS_FLAG_REMOTE_WRITE; + + if (access_flags & IB_ACCESS_REMOTE_READ) + flags |= GDMA_ACCESS_FLAG_REMOTE_READ; + + return flags; +} + +static int mana_ib_gd_create_mr(struct mana_ib_dev *dev, struct mana_ib_mr *mr, + struct gdma_create_mr_params *mr_params) +{ + struct gdma_create_mr_response resp = {}; + struct gdma_create_mr_request req = {}; + struct gdma_dev *mdev = dev->gdma_dev; + struct gdma_context *gc; + int err; + + gc = mdev->gdma_context; + + mana_gd_init_req_hdr(&req.hdr, GDMA_CREATE_MR, sizeof(req), + sizeof(resp)); + req.pd_handle = mr_params->pd_handle; + req.mr_type = mr_params->mr_type; + + switch (mr_params->mr_type) { + case GDMA_MR_TYPE_GVA: + req.gva.dma_region_handle = mr_params->gva.dma_region_handle; + req.gva.virtual_address = mr_params->gva.virtual_address; + req.gva.access_flags = mr_params->gva.access_flags; + break; + + default: + ibdev_dbg(&dev->ib_dev, + "invalid param (GDMA_MR_TYPE) passed, type %d\n", + req.mr_type); + return -EINVAL; + } + + err = mana_gd_send_request(gc, sizeof(req), &req, sizeof(resp), &resp); + + if (err || resp.hdr.status) { + ibdev_dbg(&dev->ib_dev, "Failed to create mr %d, %u", err, + resp.hdr.status); + if (!err) + err = -EPROTO; + + return err; + } + + mr->ibmr.lkey = resp.lkey; + mr->ibmr.rkey = resp.rkey; + mr->mr_handle = resp.mr_handle; + + return 0; +} + +static int mana_ib_gd_destroy_mr(struct mana_ib_dev *dev, u64 mr_handle) +{ + struct gdma_destroy_mr_response resp = {}; + struct gdma_destroy_mr_request req = {}; + struct gdma_dev *mdev = dev->gdma_dev; + struct gdma_context *gc; + int err; + + gc = mdev->gdma_context; + + mana_gd_init_req_hdr(&req.hdr, GDMA_DESTROY_MR, sizeof(req), + sizeof(resp)); + + req.mr_handle = mr_handle; + + err = mana_gd_send_request(gc, sizeof(req), &req, sizeof(resp), &resp); + if (err || resp.hdr.status) { + dev_err(gc->dev, "Failed to destroy MR: %d, 0x%x\n", err, + resp.hdr.status); + if (!err) + err = -EPROTO; + return err; + } + + return 0; +} + +struct ib_mr *mana_ib_reg_user_mr(struct ib_pd *ibpd, u64 start, u64 length, + u64 iova, int access_flags, + struct ib_udata *udata) +{ + struct mana_ib_pd *pd = container_of(ibpd, struct mana_ib_pd, ibpd); + struct gdma_create_mr_params mr_params = {}; + struct ib_device *ibdev = ibpd->device; + struct mana_ib_dev *dev; + struct mana_ib_mr *mr; + u64 dma_region_handle; + int err; + + dev = container_of(ibdev, struct mana_ib_dev, ib_dev); + + ibdev_dbg(ibdev, + "start 0x%llx, iova 0x%llx length 0x%llx access_flags 0x%x", + start, iova, length, access_flags); + + if (access_flags & ~VALID_MR_FLAGS) + return ERR_PTR(-EINVAL); + + mr = kzalloc(sizeof(*mr), GFP_KERNEL); + if (!mr) + return ERR_PTR(-ENOMEM); + + mr->umem = ib_umem_get(ibdev, start, length, access_flags); + if (IS_ERR(mr->umem)) { + err = PTR_ERR(mr->umem); + ibdev_dbg(ibdev, + "Failed to get umem for register user-mr, %d\n", err); + goto err_free; + } + + err = mana_ib_gd_create_dma_region(dev, mr->umem, &dma_region_handle); + if (err) { + ibdev_dbg(ibdev, "Failed create dma region for user-mr, %d\n", + err); + goto err_umem; + } + + ibdev_dbg(ibdev, + "mana_ib_gd_create_dma_region ret %d gdma_region %llx\n", err, + dma_region_handle); + + mr_params.pd_handle = pd->pd_handle; + mr_params.mr_type = GDMA_MR_TYPE_GVA; + mr_params.gva.dma_region_handle = dma_region_handle; + mr_params.gva.virtual_address = iova; + mr_params.gva.access_flags = + mana_ib_verbs_to_gdma_access_flags(access_flags); + + err = mana_ib_gd_create_mr(dev, mr, &mr_params); + if (err) + goto err_dma_region; + + /* + * There is no need to keep track of dma_region_handle after MR is + * successfully created. The dma_region_handle is tracked in the PF + * as part of the lifecycle of this MR. + */ + + return &mr->ibmr; + +err_dma_region: + mana_gd_destroy_dma_region(dev->gdma_dev->gdma_context, + dma_region_handle); + +err_umem: + ib_umem_release(mr->umem); + +err_free: + kfree(mr); + return ERR_PTR(err); +} + +int mana_ib_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) +{ + struct mana_ib_mr *mr = container_of(ibmr, struct mana_ib_mr, ibmr); + struct ib_device *ibdev = ibmr->device; + struct mana_ib_dev *dev; + int err; + + dev = container_of(ibdev, struct mana_ib_dev, ib_dev); + + err = mana_ib_gd_destroy_mr(dev, mr->mr_handle); + if (err) + return err; + + if (mr->umem) + ib_umem_release(mr->umem); + + kfree(mr); + + return 0; +} diff --git a/drivers/infiniband/hw/mana/qp.c b/drivers/infiniband/hw/mana/qp.c new file mode 100644 index 000000000000..ea15ec77e321 --- /dev/null +++ b/drivers/infiniband/hw/mana/qp.c @@ -0,0 +1,506 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" + +static int mana_ib_cfg_vport_steering(struct mana_ib_dev *dev, + struct net_device *ndev, + mana_handle_t default_rxobj, + mana_handle_t ind_table[], + u32 log_ind_tbl_size, u32 rx_hash_key_len, + u8 *rx_hash_key) +{ + struct mana_port_context *mpc = netdev_priv(ndev); + struct mana_cfg_rx_steer_req *req = NULL; + struct mana_cfg_rx_steer_resp resp = {}; + mana_handle_t *req_indir_tab; + struct gdma_context *gc; + struct gdma_dev *mdev; + u32 req_buf_size; + int i, err; + + mdev = dev->gdma_dev; + gc = mdev->gdma_context; + + req_buf_size = + sizeof(*req) + sizeof(mana_handle_t) * MANA_INDIRECT_TABLE_SIZE; + req = kzalloc(req_buf_size, GFP_KERNEL); + if (!req) + return -ENOMEM; + + mana_gd_init_req_hdr(&req->hdr, MANA_CONFIG_VPORT_RX, req_buf_size, + sizeof(resp)); + + req->vport = mpc->port_handle; + req->rx_enable = 1; + req->update_default_rxobj = 1; + req->default_rxobj = default_rxobj; + req->hdr.dev_id = mdev->dev_id; + + /* If there are more than 1 entries in indirection table, enable RSS */ + if (log_ind_tbl_size) + req->rss_enable = true; + + req->num_indir_entries = MANA_INDIRECT_TABLE_SIZE; + req->indir_tab_offset = sizeof(*req); + req->update_indir_tab = true; + + req_indir_tab = (mana_handle_t *)(req + 1); + /* The ind table passed to the hardware must have + * MANA_INDIRECT_TABLE_SIZE entries. Adjust the verb + * ind_table to MANA_INDIRECT_TABLE_SIZE if required + */ + ibdev_dbg(&dev->ib_dev, "ind table size %u\n", 1 << log_ind_tbl_size); + for (i = 0; i < MANA_INDIRECT_TABLE_SIZE; i++) { + req_indir_tab[i] = ind_table[i % (1 << log_ind_tbl_size)]; + ibdev_dbg(&dev->ib_dev, "index %u handle 0x%llx\n", i, + req_indir_tab[i]); + } + + req->update_hashkey = true; + if (rx_hash_key_len) + memcpy(req->hashkey, rx_hash_key, rx_hash_key_len); + else + netdev_rss_key_fill(req->hashkey, MANA_HASH_KEY_SIZE); + + ibdev_dbg(&dev->ib_dev, "vport handle %llu default_rxobj 0x%llx\n", + req->vport, default_rxobj); + + err = mana_gd_send_request(gc, req_buf_size, req, sizeof(resp), &resp); + if (err) { + netdev_err(ndev, "Failed to configure vPort RX: %d\n", err); + goto out; + } + + if (resp.hdr.status) { + netdev_err(ndev, "vPort RX configuration failed: 0x%x\n", + resp.hdr.status); + err = -EPROTO; + goto out; + } + + netdev_info(ndev, "Configured steering vPort %llu log_entries %u\n", + mpc->port_handle, log_ind_tbl_size); + +out: + kfree(req); + return err; +} + +static int mana_ib_create_qp_rss(struct ib_qp *ibqp, struct ib_pd *pd, + struct ib_qp_init_attr *attr, + struct ib_udata *udata) +{ + struct mana_ib_qp *qp = container_of(ibqp, struct mana_ib_qp, ibqp); + struct mana_ib_dev *mdev = + container_of(pd->device, struct mana_ib_dev, ib_dev); + struct ib_rwq_ind_table *ind_tbl = attr->rwq_ind_tbl; + struct mana_ib_create_qp_rss_resp resp = {}; + struct mana_ib_create_qp_rss ucmd = {}; + struct gdma_dev *gd = mdev->gdma_dev; + mana_handle_t *mana_ind_table; + struct mana_port_context *mpc; + struct mana_context *mc; + struct net_device *ndev; + struct mana_ib_cq *cq; + struct mana_ib_wq *wq; + unsigned int ind_tbl_size; + struct ib_cq *ibcq; + struct ib_wq *ibwq; + int i = 0; + u32 port; + int ret; + + mc = gd->driver_data; + + if (!udata || udata->inlen < sizeof(ucmd)) + return -EINVAL; + + ret = ib_copy_from_udata(&ucmd, udata, min(sizeof(ucmd), udata->inlen)); + if (ret) { + ibdev_dbg(&mdev->ib_dev, + "Failed copy from udata for create rss-qp, err %d\n", + ret); + return ret; + } + + if (attr->cap.max_recv_wr > MAX_SEND_BUFFERS_PER_QUEUE) { + ibdev_dbg(&mdev->ib_dev, + "Requested max_recv_wr %d exceeding limit\n", + attr->cap.max_recv_wr); + return -EINVAL; + } + + if (attr->cap.max_recv_sge > MAX_RX_WQE_SGL_ENTRIES) { + ibdev_dbg(&mdev->ib_dev, + "Requested max_recv_sge %d exceeding limit\n", + attr->cap.max_recv_sge); + return -EINVAL; + } + + ind_tbl_size = 1 << ind_tbl->log_ind_tbl_size; + if (ind_tbl_size > MANA_INDIRECT_TABLE_SIZE) { + ibdev_dbg(&mdev->ib_dev, + "Indirect table size %d exceeding limit\n", + ind_tbl_size); + return -EINVAL; + } + + if (ucmd.rx_hash_function != MANA_IB_RX_HASH_FUNC_TOEPLITZ) { + ibdev_dbg(&mdev->ib_dev, + "RX Hash function is not supported, %d\n", + ucmd.rx_hash_function); + return -EINVAL; + } + + /* IB ports start with 1, MANA start with 0 */ + port = ucmd.port; + if (port < 1 || port > mc->num_ports) { + ibdev_dbg(&mdev->ib_dev, "Invalid port %u in creating qp\n", + port); + return -EINVAL; + } + ndev = mc->ports[port - 1]; + mpc = netdev_priv(ndev); + + ibdev_dbg(&mdev->ib_dev, "rx_hash_function %d port %d\n", + ucmd.rx_hash_function, port); + + mana_ind_table = kcalloc(ind_tbl_size, sizeof(mana_handle_t), + GFP_KERNEL); + if (!mana_ind_table) { + ret = -ENOMEM; + goto fail; + } + + qp->port = port; + + for (i = 0; i < ind_tbl_size; i++) { + struct mana_obj_spec wq_spec = {}; + struct mana_obj_spec cq_spec = {}; + + ibwq = ind_tbl->ind_tbl[i]; + wq = container_of(ibwq, struct mana_ib_wq, ibwq); + + ibcq = ibwq->cq; + cq = container_of(ibcq, struct mana_ib_cq, ibcq); + + wq_spec.gdma_region = wq->gdma_region; + wq_spec.queue_size = wq->wq_buf_size; + + cq_spec.gdma_region = cq->gdma_region; + cq_spec.queue_size = cq->cqe * COMP_ENTRY_SIZE; + cq_spec.modr_ctx_id = 0; + cq_spec.attached_eq = GDMA_CQ_NO_EQ; + + ret = mana_create_wq_obj(mpc, mpc->port_handle, GDMA_RQ, + &wq_spec, &cq_spec, &wq->rx_object); + if (ret) + goto fail; + + /* The GDMA regions are now owned by the WQ object */ + wq->gdma_region = GDMA_INVALID_DMA_REGION; + cq->gdma_region = GDMA_INVALID_DMA_REGION; + + wq->id = wq_spec.queue_index; + cq->id = cq_spec.queue_index; + + ibdev_dbg(&mdev->ib_dev, + "ret %d rx_object 0x%llx wq id %llu cq id %llu\n", + ret, wq->rx_object, wq->id, cq->id); + + resp.entries[i].cqid = cq->id; + resp.entries[i].wqid = wq->id; + + mana_ind_table[i] = wq->rx_object; + } + resp.num_entries = i; + + ret = mana_ib_cfg_vport_steering(mdev, ndev, wq->rx_object, + mana_ind_table, + ind_tbl->log_ind_tbl_size, + ucmd.rx_hash_key_len, + ucmd.rx_hash_key); + if (ret) + goto fail; + + ret = ib_copy_to_udata(udata, &resp, sizeof(resp)); + if (ret) { + ibdev_dbg(&mdev->ib_dev, + "Failed to copy to udata create rss-qp, %d\n", + ret); + goto fail; + } + + kfree(mana_ind_table); + + return 0; + +fail: + while (i-- > 0) { + ibwq = ind_tbl->ind_tbl[i]; + wq = container_of(ibwq, struct mana_ib_wq, ibwq); + mana_destroy_wq_obj(mpc, GDMA_RQ, wq->rx_object); + } + + kfree(mana_ind_table); + + return ret; +} + +static int mana_ib_create_qp_raw(struct ib_qp *ibqp, struct ib_pd *ibpd, + struct ib_qp_init_attr *attr, + struct ib_udata *udata) +{ + struct mana_ib_pd *pd = container_of(ibpd, struct mana_ib_pd, ibpd); + struct mana_ib_qp *qp = container_of(ibqp, struct mana_ib_qp, ibqp); + struct mana_ib_dev *mdev = + container_of(ibpd->device, struct mana_ib_dev, ib_dev); + struct mana_ib_cq *send_cq = + container_of(attr->send_cq, struct mana_ib_cq, ibcq); + struct mana_ib_ucontext *mana_ucontext = + rdma_udata_to_drv_context(udata, struct mana_ib_ucontext, + ibucontext); + struct mana_ib_create_qp_resp resp = {}; + struct gdma_dev *gd = mdev->gdma_dev; + struct mana_ib_create_qp ucmd = {}; + struct mana_obj_spec wq_spec = {}; + struct mana_obj_spec cq_spec = {}; + struct mana_port_context *mpc; + struct mana_context *mc; + struct net_device *ndev; + struct ib_umem *umem; + int err; + u32 port; + + mc = gd->driver_data; + + if (!mana_ucontext || udata->inlen < sizeof(ucmd)) + return -EINVAL; + + err = ib_copy_from_udata(&ucmd, udata, min(sizeof(ucmd), udata->inlen)); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed to copy from udata create qp-raw, %d\n", err); + return err; + } + + /* IB ports start with 1, MANA Ethernet ports start with 0 */ + port = ucmd.port; + if (ucmd.port > mc->num_ports) + return -EINVAL; + + if (attr->cap.max_send_wr > MAX_SEND_BUFFERS_PER_QUEUE) { + ibdev_dbg(&mdev->ib_dev, + "Requested max_send_wr %d exceeding limit\n", + attr->cap.max_send_wr); + return -EINVAL; + } + + if (attr->cap.max_send_sge > MAX_TX_WQE_SGL_ENTRIES) { + ibdev_dbg(&mdev->ib_dev, + "Requested max_send_sge %d exceeding limit\n", + attr->cap.max_send_sge); + return -EINVAL; + } + + ndev = mc->ports[port - 1]; + mpc = netdev_priv(ndev); + ibdev_dbg(&mdev->ib_dev, "port %u ndev %p mpc %p\n", port, ndev, mpc); + + err = mana_ib_cfg_vport(mdev, port - 1, pd, mana_ucontext->doorbell); + if (err) + return -ENODEV; + + qp->port = port; + + ibdev_dbg(&mdev->ib_dev, "ucmd sq_buf_addr 0x%llx port %u\n", + ucmd.sq_buf_addr, ucmd.port); + + umem = ib_umem_get(ibpd->device, ucmd.sq_buf_addr, ucmd.sq_buf_size, + IB_ACCESS_LOCAL_WRITE); + if (IS_ERR(umem)) { + err = PTR_ERR(umem); + ibdev_dbg(&mdev->ib_dev, + "Failed to get umem for create qp-raw, err %d\n", + err); + goto err_free_vport; + } + qp->sq_umem = umem; + + err = mana_ib_gd_create_dma_region(mdev, qp->sq_umem, + &qp->sq_gdma_region); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed to create dma region for create qp-raw, %d\n", + err); + goto err_release_umem; + } + + ibdev_dbg(&mdev->ib_dev, + "mana_ib_gd_create_dma_region ret %d gdma_region 0x%llx\n", + err, qp->sq_gdma_region); + + /* Create a WQ on the same port handle used by the Ethernet */ + wq_spec.gdma_region = qp->sq_gdma_region; + wq_spec.queue_size = ucmd.sq_buf_size; + + cq_spec.gdma_region = send_cq->gdma_region; + cq_spec.queue_size = send_cq->cqe * COMP_ENTRY_SIZE; + cq_spec.modr_ctx_id = 0; + cq_spec.attached_eq = GDMA_CQ_NO_EQ; + + err = mana_create_wq_obj(mpc, mpc->port_handle, GDMA_SQ, &wq_spec, + &cq_spec, &qp->tx_object); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed to create wq for create raw-qp, err %d\n", + err); + goto err_destroy_dma_region; + } + + /* The GDMA regions are now owned by the WQ object */ + qp->sq_gdma_region = GDMA_INVALID_DMA_REGION; + send_cq->gdma_region = GDMA_INVALID_DMA_REGION; + + qp->sq_id = wq_spec.queue_index; + send_cq->id = cq_spec.queue_index; + + ibdev_dbg(&mdev->ib_dev, + "ret %d qp->tx_object 0x%llx sq id %llu cq id %llu\n", err, + qp->tx_object, qp->sq_id, send_cq->id); + + resp.sqid = qp->sq_id; + resp.cqid = send_cq->id; + resp.tx_vp_offset = pd->tx_vp_offset; + + err = ib_copy_to_udata(udata, &resp, sizeof(resp)); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed copy udata for create qp-raw, %d\n", + err); + goto err_destroy_wq_obj; + } + + return 0; + +err_destroy_wq_obj: + mana_destroy_wq_obj(mpc, GDMA_SQ, qp->tx_object); + +err_destroy_dma_region: + mana_ib_gd_destroy_dma_region(mdev, qp->sq_gdma_region); + +err_release_umem: + ib_umem_release(umem); + +err_free_vport: + mana_ib_uncfg_vport(mdev, pd, port - 1); + + return err; +} + +int mana_ib_create_qp(struct ib_qp *ibqp, struct ib_qp_init_attr *attr, + struct ib_udata *udata) +{ + switch (attr->qp_type) { + case IB_QPT_RAW_PACKET: + /* When rwq_ind_tbl is used, it's for creating WQs for RSS */ + if (attr->rwq_ind_tbl) + return mana_ib_create_qp_rss(ibqp, ibqp->pd, attr, + udata); + + return mana_ib_create_qp_raw(ibqp, ibqp->pd, attr, udata); + default: + /* Creating QP other than IB_QPT_RAW_PACKET is not supported */ + ibdev_dbg(ibqp->device, "Creating QP type %u not supported\n", + attr->qp_type); + } + + return -EINVAL; +} + +int mana_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, + int attr_mask, struct ib_udata *udata) +{ + /* modify_qp is not supported by this version of the driver */ + return -EOPNOTSUPP; +} + +static int mana_ib_destroy_qp_rss(struct mana_ib_qp *qp, + struct ib_rwq_ind_table *ind_tbl, + struct ib_udata *udata) +{ + struct mana_ib_dev *mdev = + container_of(qp->ibqp.device, struct mana_ib_dev, ib_dev); + struct gdma_dev *gd = mdev->gdma_dev; + struct mana_port_context *mpc; + struct mana_context *mc; + struct net_device *ndev; + struct mana_ib_wq *wq; + struct ib_wq *ibwq; + int i; + + mc = gd->driver_data; + ndev = mc->ports[qp->port - 1]; + mpc = netdev_priv(ndev); + + for (i = 0; i < (1 << ind_tbl->log_ind_tbl_size); i++) { + ibwq = ind_tbl->ind_tbl[i]; + wq = container_of(ibwq, struct mana_ib_wq, ibwq); + ibdev_dbg(&mdev->ib_dev, "destroying wq->rx_object %llu\n", + wq->rx_object); + mana_destroy_wq_obj(mpc, GDMA_RQ, wq->rx_object); + } + + return 0; +} + +static int mana_ib_destroy_qp_raw(struct mana_ib_qp *qp, struct ib_udata *udata) +{ + struct mana_ib_dev *mdev = + container_of(qp->ibqp.device, struct mana_ib_dev, ib_dev); + struct gdma_dev *gd = mdev->gdma_dev; + struct ib_pd *ibpd = qp->ibqp.pd; + struct mana_port_context *mpc; + struct mana_context *mc; + struct net_device *ndev; + struct mana_ib_pd *pd; + + mc = gd->driver_data; + ndev = mc->ports[qp->port - 1]; + mpc = netdev_priv(ndev); + pd = container_of(ibpd, struct mana_ib_pd, ibpd); + + mana_destroy_wq_obj(mpc, GDMA_SQ, qp->tx_object); + + if (qp->sq_umem) { + mana_ib_gd_destroy_dma_region(mdev, qp->sq_gdma_region); + ib_umem_release(qp->sq_umem); + } + + mana_ib_uncfg_vport(mdev, pd, qp->port - 1); + + return 0; +} + +int mana_ib_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata) +{ + struct mana_ib_qp *qp = container_of(ibqp, struct mana_ib_qp, ibqp); + + switch (ibqp->qp_type) { + case IB_QPT_RAW_PACKET: + if (ibqp->rwq_ind_tbl) + return mana_ib_destroy_qp_rss(qp, ibqp->rwq_ind_tbl, + udata); + + return mana_ib_destroy_qp_raw(qp, udata); + + default: + ibdev_dbg(ibqp->device, "Unexpected QP type %u\n", + ibqp->qp_type); + } + + return -ENOENT; +} diff --git a/drivers/infiniband/hw/mana/wq.c b/drivers/infiniband/hw/mana/wq.c new file mode 100644 index 000000000000..372d361510e0 --- /dev/null +++ b/drivers/infiniband/hw/mana/wq.c @@ -0,0 +1,115 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022, Microsoft Corporation. All rights reserved. + */ + +#include "mana_ib.h" + +struct ib_wq *mana_ib_create_wq(struct ib_pd *pd, + struct ib_wq_init_attr *init_attr, + struct ib_udata *udata) +{ + struct mana_ib_dev *mdev = + container_of(pd->device, struct mana_ib_dev, ib_dev); + struct mana_ib_create_wq ucmd = {}; + struct mana_ib_wq *wq; + struct ib_umem *umem; + int err; + + if (udata->inlen < sizeof(ucmd)) + return ERR_PTR(-EINVAL); + + err = ib_copy_from_udata(&ucmd, udata, min(sizeof(ucmd), udata->inlen)); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed to copy from udata for create wq, %d\n", err); + return ERR_PTR(err); + } + + wq = kzalloc(sizeof(*wq), GFP_KERNEL); + if (!wq) + return ERR_PTR(-ENOMEM); + + ibdev_dbg(&mdev->ib_dev, "ucmd wq_buf_addr 0x%llx\n", ucmd.wq_buf_addr); + + umem = ib_umem_get(pd->device, ucmd.wq_buf_addr, ucmd.wq_buf_size, + IB_ACCESS_LOCAL_WRITE); + if (IS_ERR(umem)) { + err = PTR_ERR(umem); + ibdev_dbg(&mdev->ib_dev, + "Failed to get umem for create wq, err %d\n", err); + goto err_free_wq; + } + + wq->umem = umem; + wq->wqe = init_attr->max_wr; + wq->wq_buf_size = ucmd.wq_buf_size; + wq->rx_object = INVALID_MANA_HANDLE; + + err = mana_ib_gd_create_dma_region(mdev, wq->umem, &wq->gdma_region); + if (err) { + ibdev_dbg(&mdev->ib_dev, + "Failed to create dma region for create wq, %d\n", + err); + goto err_release_umem; + } + + ibdev_dbg(&mdev->ib_dev, + "mana_ib_gd_create_dma_region ret %d gdma_region 0x%llx\n", + err, wq->gdma_region); + + /* WQ ID is returned at wq_create time, doesn't know the value yet */ + + return &wq->ibwq; + +err_release_umem: + ib_umem_release(umem); + +err_free_wq: + kfree(wq); + + return ERR_PTR(err); +} + +int mana_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr, + u32 wq_attr_mask, struct ib_udata *udata) +{ + /* modify_wq is not supported by this version of the driver */ + return -EOPNOTSUPP; +} + +int mana_ib_destroy_wq(struct ib_wq *ibwq, struct ib_udata *udata) +{ + struct mana_ib_wq *wq = container_of(ibwq, struct mana_ib_wq, ibwq); + struct ib_device *ib_dev = ibwq->device; + struct mana_ib_dev *mdev; + + mdev = container_of(ib_dev, struct mana_ib_dev, ib_dev); + + mana_ib_gd_destroy_dma_region(mdev, wq->gdma_region); + ib_umem_release(wq->umem); + + kfree(wq); + + return 0; +} + +int mana_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) +{ + /* + * There is no additional data in ind_table to be maintained by this + * driver, do nothing + */ + return 0; +} + +int mana_ib_destroy_rwq_ind_table(struct ib_rwq_ind_table *ib_rwq_ind_tbl) +{ + /* + * There is no additional data in ind_table to be maintained by this + * driver, do nothing + */ + return 0; +} diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c index ba47874f90d3..dceebcd885bb 100644 --- a/drivers/infiniband/hw/mlx4/main.c +++ b/drivers/infiniband/hw/mlx4/main.c @@ -144,8 +144,7 @@ static struct net_device *mlx4_ib_get_netdev(struct ib_device *device, } } } - if (dev) - dev_hold(dev); + dev_hold(dev); rcu_read_unlock(); return dev; @@ -1307,8 +1306,7 @@ int mlx4_ib_add_mc(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp, spin_lock_bh(&mdev->iboe.lock); ndev = mdev->iboe.netdevs[mqp->port - 1]; - if (ndev) - dev_hold(ndev); + dev_hold(ndev); spin_unlock_bh(&mdev->iboe.lock); if (ndev) { @@ -1955,11 +1953,9 @@ static int mlx4_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid) if (ge) { spin_lock_bh(&mdev->iboe.lock); ndev = ge->added ? mdev->iboe.netdevs[ge->port - 1] : NULL; - if (ndev) - dev_hold(ndev); + dev_hold(ndev); spin_unlock_bh(&mdev->iboe.lock); - if (ndev) - dev_put(ndev); + dev_put(ndev); list_del(&ge->list); kfree(ge); } else diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c index be189e0525de..efc9e4a6df04 100644 --- a/drivers/infiniband/hw/mlx5/cq.c +++ b/drivers/infiniband/hw/mlx5/cq.c @@ -267,17 +267,20 @@ static void handle_responder(struct ib_wc *wc, struct mlx5_cqe64 *cqe, wc->wc_flags |= IB_WC_WITH_NETWORK_HDR_TYPE; } -static void dump_cqe(struct mlx5_ib_dev *dev, struct mlx5_err_cqe *cqe) +static void dump_cqe(struct mlx5_ib_dev *dev, struct mlx5_err_cqe *cqe, + struct ib_wc *wc, const char *level) { - mlx5_ib_warn(dev, "dump error cqe\n"); - mlx5_dump_err_cqe(dev->mdev, cqe); + mlx5_ib_log(level, dev, "WC error: %d, Message: %s\n", wc->status, + ib_wc_status_msg(wc->status)); + print_hex_dump(level, "cqe_dump: ", DUMP_PREFIX_OFFSET, 16, 1, + cqe, sizeof(*cqe), false); } static void mlx5_handle_error_cqe(struct mlx5_ib_dev *dev, struct mlx5_err_cqe *cqe, struct ib_wc *wc) { - int dump = 1; + const char *dump = KERN_WARNING; switch (cqe->syndrome) { case MLX5_CQE_SYNDROME_LOCAL_LENGTH_ERR: @@ -287,10 +290,11 @@ static void mlx5_handle_error_cqe(struct mlx5_ib_dev *dev, wc->status = IB_WC_LOC_QP_OP_ERR; break; case MLX5_CQE_SYNDROME_LOCAL_PROT_ERR: + dump = KERN_DEBUG; wc->status = IB_WC_LOC_PROT_ERR; break; case MLX5_CQE_SYNDROME_WR_FLUSH_ERR: - dump = 0; + dump = NULL; wc->status = IB_WC_WR_FLUSH_ERR; break; case MLX5_CQE_SYNDROME_MW_BIND_ERR: @@ -306,18 +310,20 @@ static void mlx5_handle_error_cqe(struct mlx5_ib_dev *dev, wc->status = IB_WC_REM_INV_REQ_ERR; break; case MLX5_CQE_SYNDROME_REMOTE_ACCESS_ERR: + dump = KERN_DEBUG; wc->status = IB_WC_REM_ACCESS_ERR; break; case MLX5_CQE_SYNDROME_REMOTE_OP_ERR: + dump = KERN_DEBUG; wc->status = IB_WC_REM_OP_ERR; break; case MLX5_CQE_SYNDROME_TRANSPORT_RETRY_EXC_ERR: + dump = NULL; wc->status = IB_WC_RETRY_EXC_ERR; - dump = 0; break; case MLX5_CQE_SYNDROME_RNR_RETRY_EXC_ERR: + dump = NULL; wc->status = IB_WC_RNR_RETRY_EXC_ERR; - dump = 0; break; case MLX5_CQE_SYNDROME_REMOTE_ABORTED_ERR: wc->status = IB_WC_REM_ABORT_ERR; @@ -328,11 +334,8 @@ static void mlx5_handle_error_cqe(struct mlx5_ib_dev *dev, } wc->vendor_err = cqe->vendor_err_synd; - if (dump) { - mlx5_ib_warn(dev, "WC error: %d, Message: %s\n", wc->status, - ib_wc_status_msg(wc->status)); - dump_cqe(dev, cqe); - } + if (dump) + dump_cqe(dev, cqe, wc, dump); } static void handle_atomics(struct mlx5_ib_qp *qp, struct mlx5_cqe64 *cqe64, diff --git a/drivers/infiniband/hw/mlx5/fs.c b/drivers/infiniband/hw/mlx5/fs.c index 490ec308e309..3008632a6c20 100644 --- a/drivers/infiniband/hw/mlx5/fs.c +++ b/drivers/infiniband/hw/mlx5/fs.c @@ -127,7 +127,6 @@ static int check_mpls_supp_fields(u32 field_support, const __be32 *set_mask) } #define LAST_ETH_FIELD vlan_tag -#define LAST_IB_FIELD sl #define LAST_IPV4_FIELD tos #define LAST_IPV6_FIELD traffic_class #define LAST_TCP_UDP_FIELD src_port diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index 4a7f7064bd0e..8b91babdd4c0 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -38,6 +38,10 @@ dev_warn(&(_dev)->ib_dev.dev, "%s:%d:(pid %d): " format, __func__, \ __LINE__, current->pid, ##arg) +#define mlx5_ib_log(lvl, _dev, format, arg...) \ + dev_printk(lvl, &(_dev)->ib_dev.dev, "%s:%d:(pid %d): " format, \ + __func__, __LINE__, current->pid, ##arg) + #define MLX5_IB_DEFAULT_UIDX 0xffffff #define MLX5_USER_ASSIGNED_UIDX_MASK __mlx5_mask(qpc, user_index) diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c index 410cc5fd2523..053fe946e45a 100644 --- a/drivers/infiniband/hw/mlx5/mr.c +++ b/drivers/infiniband/hw/mlx5/mr.c @@ -1929,10 +1929,8 @@ int mlx5_ib_alloc_mw(struct ib_mw *ibmw, struct ib_udata *udata) ndescs = req.num_klms ? roundup(req.num_klms, 4) : roundup(1, 4); in = kzalloc(inlen, GFP_KERNEL); - if (!in) { - err = -ENOMEM; - goto free; - } + if (!in) + return -ENOMEM; mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry); diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c index ba0c3e4c07d8..ecdfeff3d44f 100644 --- a/drivers/infiniband/hw/qedr/main.c +++ b/drivers/infiniband/hw/qedr/main.c @@ -479,7 +479,7 @@ static irqreturn_t qedr_irq_handler(int irq, void *handle) /* The CQ's CNQ notification counter is checked before * destroying the CQ in a busy-wait loop that waits for all of * the CQ's CNQ interrupts to be processed. It is increased - * here, only after the completion handler, to ensure that the + * here, only after the completion handler, to ensure that * the handler is not running when the CQ is destroyed. */ cq->cnq_notif++; diff --git a/drivers/infiniband/hw/qib/qib_iba6120.c b/drivers/infiniband/hw/qib/qib_iba6120.c index 07386117f21a..277769fa9745 100644 --- a/drivers/infiniband/hw/qib/qib_iba6120.c +++ b/drivers/infiniband/hw/qib/qib_iba6120.c @@ -799,12 +799,9 @@ static void qib_handle_6120_hwerrors(struct qib_devdata *dd, char *msg, hwerrs &= ~TXE_PIO_PARITY; } - if (!hwerrs) { - static u32 freeze_cnt; - - freeze_cnt++; + if (!hwerrs) qib_6120_clear_freeze(dd); - } else + else isfatal = 1; } diff --git a/drivers/infiniband/hw/qib/qib_tx.c b/drivers/infiniband/hw/qib/qib_tx.c index 6a8148851f21..1325110237cd 100644 --- a/drivers/infiniband/hw/qib/qib_tx.c +++ b/drivers/infiniband/hw/qib/qib_tx.c @@ -82,7 +82,6 @@ int qib_disarm_piobufs_ifneeded(struct qib_ctxtdata *rcd) struct qib_devdata *dd = rcd->dd; unsigned i; unsigned last; - unsigned n = 0; last = rcd->pio_base + rcd->piocnt; /* @@ -102,10 +101,8 @@ int qib_disarm_piobufs_ifneeded(struct qib_ctxtdata *rcd) } spin_lock_irq(&dd->pioavail_lock); for (i = rcd->pio_base; i < last; i++) { - if (__test_and_clear_bit(i, dd->pio_need_disarm)) { - n++; + if (__test_and_clear_bit(i, dd->pio_need_disarm)) dd->f_sendctrl(rcd->ppd, QIB_SENDCTRL_DISARM_BUF(i)); - } } spin_unlock_irq(&dd->pioavail_lock); return 0; diff --git a/drivers/infiniband/hw/qib/qib_user_sdma.c b/drivers/infiniband/hw/qib/qib_user_sdma.c index bf2f30d67949..9fe03d6ffac1 100644 --- a/drivers/infiniband/hw/qib/qib_user_sdma.c +++ b/drivers/infiniband/hw/qib/qib_user_sdma.c @@ -851,7 +851,7 @@ static int qib_user_sdma_queue_pkts(const struct qib_devdata *dd, } /* - * This assignment is a bit strange. it's because the + * This assignment is a bit strange. it's because * the pbc counts the number of 32 bit words in the full * packet _except_ the first word of the pbc itself... */ |