diff options
Diffstat (limited to 'drivers/infiniband/hw')
37 files changed, 395 insertions, 202 deletions
diff --git a/drivers/infiniband/hw/bnxt_re/bnxt_re.h b/drivers/infiniband/hw/bnxt_re/bnxt_re.h index 08772836fded..85527532c49d 100644 --- a/drivers/infiniband/hw/bnxt_re/bnxt_re.h +++ b/drivers/infiniband/hw/bnxt_re/bnxt_re.h @@ -51,6 +51,8 @@ #define BNXT_RE_PAGE_SIZE_8M BIT(23) #define BNXT_RE_PAGE_SIZE_1G BIT(30) +#define BNXT_RE_MAX_MR_SIZE BIT(30) + #define BNXT_RE_MAX_QPC_COUNT (64 * 1024) #define BNXT_RE_MAX_MRW_COUNT (64 * 1024) #define BNXT_RE_MAX_SRQC_COUNT (64 * 1024) @@ -60,6 +62,13 @@ #define BNXT_RE_RQ_WQE_THRESHOLD 32 +/* + * Setting the default ack delay value to 16, which means + * the default timeout is approx. 260ms(4 usec * 2 ^(timeout)) + */ + +#define BNXT_RE_DEFAULT_ACK_DELAY 16 + struct bnxt_re_work { struct work_struct work; unsigned long event; diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index c7bd68311d0c..f0e01b3ac711 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -145,10 +145,8 @@ int bnxt_re_query_device(struct ib_device *ibdev, ib_attr->fw_ver = (u64)(unsigned long)(dev_attr->fw_ver); bnxt_qplib_get_guid(rdev->netdev->dev_addr, (u8 *)&ib_attr->sys_image_guid); - ib_attr->max_mr_size = ~0ull; - ib_attr->page_size_cap = BNXT_RE_PAGE_SIZE_4K | BNXT_RE_PAGE_SIZE_8K | - BNXT_RE_PAGE_SIZE_64K | BNXT_RE_PAGE_SIZE_2M | - BNXT_RE_PAGE_SIZE_8M | BNXT_RE_PAGE_SIZE_1G; + ib_attr->max_mr_size = BNXT_RE_MAX_MR_SIZE; + ib_attr->page_size_cap = BNXT_RE_PAGE_SIZE_4K; ib_attr->vendor_id = rdev->en_dev->pdev->vendor; ib_attr->vendor_part_id = rdev->en_dev->pdev->device; @@ -174,9 +172,11 @@ int bnxt_re_query_device(struct ib_device *ibdev, ib_attr->max_mr = dev_attr->max_mr; ib_attr->max_pd = dev_attr->max_pd; ib_attr->max_qp_rd_atom = dev_attr->max_qp_rd_atom; - ib_attr->max_qp_init_rd_atom = dev_attr->max_qp_rd_atom; - ib_attr->atomic_cap = IB_ATOMIC_HCA; - ib_attr->masked_atomic_cap = IB_ATOMIC_HCA; + ib_attr->max_qp_init_rd_atom = dev_attr->max_qp_init_rd_atom; + if (dev_attr->is_atomic) { + ib_attr->atomic_cap = IB_ATOMIC_HCA; + ib_attr->masked_atomic_cap = IB_ATOMIC_HCA; + } ib_attr->max_ee_rd_atom = 0; ib_attr->max_res_rd_atom = 0; @@ -201,7 +201,7 @@ int bnxt_re_query_device(struct ib_device *ibdev, ib_attr->max_fast_reg_page_list_len = MAX_PBL_LVL_1_PGS; ib_attr->max_pkeys = 1; - ib_attr->local_ca_ack_delay = 0; + ib_attr->local_ca_ack_delay = BNXT_RE_DEFAULT_ACK_DELAY; return 0; } @@ -390,15 +390,17 @@ int bnxt_re_del_gid(struct ib_device *ibdev, u8 port_num, return -EINVAL; ctx->refcnt--; if (!ctx->refcnt) { - rc = bnxt_qplib_del_sgid - (sgid_tbl, - &sgid_tbl->tbl[ctx->idx], true); - if (rc) + rc = bnxt_qplib_del_sgid(sgid_tbl, + &sgid_tbl->tbl[ctx->idx], + true); + if (rc) { dev_err(rdev_to_dev(rdev), "Failed to remove GID: %#x", rc); - ctx_tbl = sgid_tbl->ctx; - ctx_tbl[ctx->idx] = NULL; - kfree(ctx); + } else { + ctx_tbl = sgid_tbl->ctx; + ctx_tbl[ctx->idx] = NULL; + kfree(ctx); + } } } else { return -EINVAL; @@ -588,10 +590,10 @@ static int bnxt_re_create_fence_mr(struct bnxt_re_pd *pd) /* Create a fence MW only for kernel consumers */ mw = bnxt_re_alloc_mw(&pd->ib_pd, IB_MW_TYPE_1, NULL); - if (!mw) { + if (IS_ERR(mw)) { dev_err(rdev_to_dev(rdev), "Failed to create fence-MW for PD: %p\n", pd); - rc = -EINVAL; + rc = PTR_ERR(mw); goto fail; } fence->mw = mw; @@ -612,30 +614,13 @@ int bnxt_re_dealloc_pd(struct ib_pd *ib_pd) int rc; bnxt_re_destroy_fence_mr(pd); - if (ib_pd->uobject && pd->dpi.dbr) { - struct ib_ucontext *ib_uctx = ib_pd->uobject->context; - struct bnxt_re_ucontext *ucntx; - /* Free DPI only if this is the first PD allocated by the - * application and mark the context dpi as NULL - */ - ucntx = container_of(ib_uctx, struct bnxt_re_ucontext, ib_uctx); - - rc = bnxt_qplib_dealloc_dpi(&rdev->qplib_res, - &rdev->qplib_res.dpi_tbl, - &pd->dpi); + if (pd->qplib_pd.id) { + rc = bnxt_qplib_dealloc_pd(&rdev->qplib_res, + &rdev->qplib_res.pd_tbl, + &pd->qplib_pd); if (rc) - dev_err(rdev_to_dev(rdev), "Failed to deallocate HW DPI"); - /* Don't fail, continue*/ - ucntx->dpi = NULL; - } - - rc = bnxt_qplib_dealloc_pd(&rdev->qplib_res, - &rdev->qplib_res.pd_tbl, - &pd->qplib_pd); - if (rc) { - dev_err(rdev_to_dev(rdev), "Failed to deallocate HW PD"); - return rc; + dev_err(rdev_to_dev(rdev), "Failed to deallocate HW PD"); } kfree(pd); @@ -667,23 +652,22 @@ struct ib_pd *bnxt_re_alloc_pd(struct ib_device *ibdev, if (udata) { struct bnxt_re_pd_resp resp; - if (!ucntx->dpi) { + if (!ucntx->dpi.dbr) { /* Allocate DPI in alloc_pd to avoid failing of * ibv_devinfo and family of application when DPIs * are depleted. */ if (bnxt_qplib_alloc_dpi(&rdev->qplib_res.dpi_tbl, - &pd->dpi, ucntx)) { + &ucntx->dpi, ucntx)) { rc = -ENOMEM; goto dbfail; } - ucntx->dpi = &pd->dpi; } resp.pdid = pd->qplib_pd.id; /* Still allow mapping this DBR to the new user PD. */ - resp.dpi = ucntx->dpi->dpi; - resp.dbr = (u64)ucntx->dpi->umdbr; + resp.dpi = ucntx->dpi.dpi; + resp.dbr = (u64)ucntx->dpi.umdbr; rc = ib_copy_to_udata(udata, &resp, sizeof(resp)); if (rc) { @@ -960,7 +944,7 @@ static int bnxt_re_init_user_qp(struct bnxt_re_dev *rdev, struct bnxt_re_pd *pd, qplib_qp->rq.nmap = umem->nmap; } - qplib_qp->dpi = cntx->dpi; + qplib_qp->dpi = &cntx->dpi; return 0; rqfail: ib_umem_release(qp->sumem); @@ -1530,13 +1514,24 @@ int bnxt_re_modify_qp(struct ib_qp *ib_qp, struct ib_qp_attr *qp_attr, if (qp_attr_mask & IB_QP_MAX_QP_RD_ATOMIC) { qp->qplib_qp.modify_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_MAX_RD_ATOMIC; - qp->qplib_qp.max_rd_atomic = qp_attr->max_rd_atomic; + /* Cap the max_rd_atomic to device max */ + qp->qplib_qp.max_rd_atomic = min_t(u32, qp_attr->max_rd_atomic, + dev_attr->max_qp_rd_atom); } if (qp_attr_mask & IB_QP_SQ_PSN) { qp->qplib_qp.modify_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_SQ_PSN; qp->qplib_qp.sq.psn = qp_attr->sq_psn; } if (qp_attr_mask & IB_QP_MAX_DEST_RD_ATOMIC) { + if (qp_attr->max_dest_rd_atomic > + dev_attr->max_qp_init_rd_atom) { + dev_err(rdev_to_dev(rdev), + "max_dest_rd_atomic requested%d is > dev_max%d", + qp_attr->max_dest_rd_atomic, + dev_attr->max_qp_init_rd_atom); + return -EINVAL; + } + qp->qplib_qp.modify_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_MAX_DEST_RD_ATOMIC; qp->qplib_qp.max_dest_rd_atomic = qp_attr->max_dest_rd_atomic; @@ -2403,7 +2398,7 @@ struct ib_cq *bnxt_re_create_cq(struct ib_device *ibdev, } cq->qplib_cq.sghead = cq->umem->sg_head.sgl; cq->qplib_cq.nmap = cq->umem->nmap; - cq->qplib_cq.dpi = uctx->dpi; + cq->qplib_cq.dpi = &uctx->dpi; } else { cq->max_cql = min_t(u32, entries, MAX_CQL_PER_POLL); cq->cql = kcalloc(cq->max_cql, sizeof(struct bnxt_qplib_cqe), @@ -2905,6 +2900,7 @@ int bnxt_re_poll_cq(struct ib_cq *ib_cq, int num_entries, struct ib_wc *wc) spin_lock_irqsave(&cq->cq_lock, flags); budget = min_t(u32, num_entries, cq->max_cql); + num_entries = budget; if (!cq->cql) { dev_err(rdev_to_dev(cq->rdev), "POLL CQ : no CQL to use"); goto exit; @@ -3031,6 +3027,11 @@ int bnxt_re_req_notify_cq(struct ib_cq *ib_cq, else if (ib_cqn_flags & IB_CQ_SOLICITED) type = DBR_DBR_TYPE_CQ_ARMSE; + /* Poll to see if there are missed events */ + if ((ib_cqn_flags & IB_CQ_REPORT_MISSED_EVENTS) && + !(bnxt_qplib_is_cq_empty(&cq->qplib_cq))) + return 1; + bnxt_qplib_req_notify_cq(&cq->qplib_cq, type); return 0; @@ -3245,6 +3246,12 @@ struct ib_mr *bnxt_re_reg_user_mr(struct ib_pd *ib_pd, u64 start, u64 length, struct scatterlist *sg; int entry; + if (length > BNXT_RE_MAX_MR_SIZE) { + dev_err(rdev_to_dev(rdev), "MR Size: %lld > Max supported:%ld\n", + length, BNXT_RE_MAX_MR_SIZE); + return ERR_PTR(-ENOMEM); + } + mr = kzalloc(sizeof(*mr), GFP_KERNEL); if (!mr) return ERR_PTR(-ENOMEM); @@ -3388,8 +3395,26 @@ int bnxt_re_dealloc_ucontext(struct ib_ucontext *ib_uctx) struct bnxt_re_ucontext *uctx = container_of(ib_uctx, struct bnxt_re_ucontext, ib_uctx); + + struct bnxt_re_dev *rdev = uctx->rdev; + int rc = 0; + if (uctx->shpg) free_page((unsigned long)uctx->shpg); + + if (uctx->dpi.dbr) { + /* Free DPI only if this is the first PD allocated by the + * application and mark the context dpi as NULL + */ + rc = bnxt_qplib_dealloc_dpi(&rdev->qplib_res, + &rdev->qplib_res.dpi_tbl, + &uctx->dpi); + if (rc) + dev_err(rdev_to_dev(rdev), "Deallocte HW DPI failed!"); + /* Don't fail, continue*/ + uctx->dpi.dbr = NULL; + } + kfree(uctx); return 0; } diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.h b/drivers/infiniband/hw/bnxt_re/ib_verbs.h index 6c160f6a5398..a0bb7e33d7ca 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.h +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.h @@ -59,7 +59,6 @@ struct bnxt_re_pd { struct bnxt_re_dev *rdev; struct ib_pd ib_pd; struct bnxt_qplib_pd qplib_pd; - struct bnxt_qplib_dpi dpi; struct bnxt_re_fence_data fence; }; @@ -127,7 +126,7 @@ struct bnxt_re_mw { struct bnxt_re_ucontext { struct bnxt_re_dev *rdev; struct ib_ucontext ib_uctx; - struct bnxt_qplib_dpi *dpi; + struct bnxt_qplib_dpi dpi; void *shpg; spinlock_t sh_lock; /* protect shpg */ }; diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c index 1fce5e73216b..ceae2d92fb08 100644 --- a/drivers/infiniband/hw/bnxt_re/main.c +++ b/drivers/infiniband/hw/bnxt_re/main.c @@ -333,6 +333,7 @@ static int bnxt_re_net_stats_ctx_alloc(struct bnxt_re_dev *rdev, bnxt_re_init_hwrm_hdr(rdev, (void *)&req, HWRM_STAT_CTX_ALLOC, -1, -1); req.update_period_ms = cpu_to_le32(1000); req.stats_dma_addr = cpu_to_le64(dma_map); + req.stat_ctx_flags = STAT_CTX_ALLOC_REQ_STAT_CTX_FLAGS_ROCE; bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp, sizeof(resp), DFLT_HWRM_CMD_TIMEOUT); rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index f05500bcdcf1..9af1514e5944 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1128,6 +1128,11 @@ int bnxt_qplib_post_send(struct bnxt_qplib_qp *qp, } /* Each SGE entry = 1 WQE size16 */ wqe_size16 = wqe->num_sge; + /* HW requires wqe size has room for atleast one SGE even if + * none was supplied by ULP + */ + if (!wqe->num_sge) + wqe_size16++; } /* Specifics */ @@ -1364,6 +1369,11 @@ int bnxt_qplib_post_recv(struct bnxt_qplib_qp *qp, rqe->flags = wqe->flags; rqe->wqe_size = wqe->num_sge + ((offsetof(typeof(*rqe), data) + 15) >> 4); + /* HW requires wqe size has room for atleast one SGE even if none + * was supplied by ULP + */ + if (!wqe->num_sge) + rqe->wqe_size++; /* Supply the rqe->wr_id index to the wr_id_tbl for now */ rqe->wr_id[0] = cpu_to_le32(sw_prod); @@ -1885,6 +1895,25 @@ flush_rq: return rc; } +bool bnxt_qplib_is_cq_empty(struct bnxt_qplib_cq *cq) +{ + struct cq_base *hw_cqe, **hw_cqe_ptr; + unsigned long flags; + u32 sw_cons, raw_cons; + bool rc = true; + + spin_lock_irqsave(&cq->hwq.lock, flags); + raw_cons = cq->hwq.cons; + sw_cons = HWQ_CMP(raw_cons, &cq->hwq); + hw_cqe_ptr = (struct cq_base **)cq->hwq.pbl_ptr; + hw_cqe = &hw_cqe_ptr[CQE_PG(sw_cons)][CQE_IDX(sw_cons)]; + + /* Check for Valid bit. If the CQE is valid, return false */ + rc = !CQE_CMP_VALID(hw_cqe, raw_cons, cq->hwq.max_elements); + spin_unlock_irqrestore(&cq->hwq.lock, flags); + return rc; +} + static int bnxt_qplib_cq_process_res_raweth_qp1(struct bnxt_qplib_cq *cq, struct cq_res_raweth_qp1 *hwcqe, struct bnxt_qplib_cqe **pcqe, diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.h b/drivers/infiniband/hw/bnxt_re/qplib_fp.h index 36b7b7db0e3f..19176e06c98a 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.h @@ -449,6 +449,7 @@ int bnxt_qplib_create_cq(struct bnxt_qplib_res *res, struct bnxt_qplib_cq *cq); int bnxt_qplib_destroy_cq(struct bnxt_qplib_res *res, struct bnxt_qplib_cq *cq); int bnxt_qplib_poll_cq(struct bnxt_qplib_cq *cq, struct bnxt_qplib_cqe *cqe, int num, struct bnxt_qplib_qp **qp); +bool bnxt_qplib_is_cq_empty(struct bnxt_qplib_cq *cq); void bnxt_qplib_req_notify_cq(struct bnxt_qplib_cq *cq, u32 arm_type); void bnxt_qplib_free_nq(struct bnxt_qplib_nq *nq); int bnxt_qplib_alloc_nq(struct pci_dev *pdev, struct bnxt_qplib_nq *nq); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.c b/drivers/infiniband/hw/bnxt_re/qplib_sp.c index fde18cf0e406..ef91ab786dd4 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.c @@ -51,6 +51,19 @@ const struct bnxt_qplib_gid bnxt_qplib_gid_zero = {{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } }; /* Device */ + +static bool bnxt_qplib_is_atomic_cap(struct bnxt_qplib_rcfw *rcfw) +{ + int rc; + u16 pcie_ctl2; + + rc = pcie_capability_read_word(rcfw->pdev, PCI_EXP_DEVCTL2, + &pcie_ctl2); + if (rc) + return false; + return !!(pcie_ctl2 & PCI_EXP_DEVCTL2_ATOMIC_REQ); +} + int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, struct bnxt_qplib_dev_attr *attr) { @@ -81,6 +94,8 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, /* Extract the context from the side buffer */ attr->max_qp = le32_to_cpu(sb->max_qp); + /* max_qp value reported by FW for PF doesn't include the QP1 for PF */ + attr->max_qp += 1; attr->max_qp_rd_atom = sb->max_qp_rd_atom > BNXT_QPLIB_MAX_OUT_RD_ATOM ? BNXT_QPLIB_MAX_OUT_RD_ATOM : sb->max_qp_rd_atom; @@ -129,6 +144,7 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, attr->tqm_alloc_reqs[i * 4 + 3] = *(++tqm_alloc); } + attr->is_atomic = bnxt_qplib_is_atomic_cap(rcfw); bail: bnxt_qplib_rcfw_free_sbuf(rcfw, sbuf); return rc; diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.h b/drivers/infiniband/hw/bnxt_re/qplib_sp.h index a543f959098b..2ce7e2a32cf0 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.h @@ -42,6 +42,8 @@ #define BNXT_QPLIB_RESERVED_QP_WRS 128 +#define PCI_EXP_DEVCTL2_ATOMIC_REQ 0x0040 + struct bnxt_qplib_dev_attr { char fw_ver[32]; u16 max_sgid; @@ -70,6 +72,7 @@ struct bnxt_qplib_dev_attr { u32 max_inline_data; u32 l2_db_size; u8 tqm_alloc_reqs[MAX_TQM_ALLOC_REQ]; + bool is_atomic; }; struct bnxt_qplib_pd { diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c index 29d30744d6c9..0cd0c1fa27d4 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_provider.c +++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c @@ -718,7 +718,7 @@ static struct ib_mr *iwch_alloc_mr(struct ib_pd *pd, struct iwch_mr *mhp; u32 mmid; u32 stag = 0; - int ret = 0; + int ret = -ENOMEM; if (mr_type != IB_MR_TYPE_MEM_REG || max_num_sg > T3_MAX_FASTREG_DEPTH) @@ -731,10 +731,8 @@ static struct ib_mr *iwch_alloc_mr(struct ib_pd *pd, goto err; mhp->pages = kcalloc(max_num_sg, sizeof(u64), GFP_KERNEL); - if (!mhp->pages) { - ret = -ENOMEM; + if (!mhp->pages) goto pl_err; - } mhp->rhp = rhp; ret = iwch_alloc_pbl(mhp, max_num_sg); @@ -751,7 +749,8 @@ static struct ib_mr *iwch_alloc_mr(struct ib_pd *pd, mhp->attr.state = 1; mmid = (stag) >> 8; mhp->ibmr.rkey = mhp->ibmr.lkey = stag; - if (insert_handle(rhp, &rhp->mmidr, mhp, mmid)) + ret = insert_handle(rhp, &rhp->mmidr, mhp, mmid); + if (ret) goto err3; pr_debug("%s mmid 0x%x mhp %p stag 0x%x\n", __func__, mmid, mhp, stag); diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c index e16fcaf6b5a3..be07da1997e6 100644 --- a/drivers/infiniband/hw/cxgb4/cq.c +++ b/drivers/infiniband/hw/cxgb4/cq.c @@ -963,6 +963,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, goto err3; if (ucontext) { + ret = -ENOMEM; mm = kmalloc(sizeof *mm, GFP_KERNEL); if (!mm) goto err4; diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c index bfc77596acbe..cb7fc0d35d1d 100644 --- a/drivers/infiniband/hw/cxgb4/qp.c +++ b/drivers/infiniband/hw/cxgb4/qp.c @@ -569,7 +569,7 @@ static int build_rdma_read(union t4_wr *wqe, struct ib_send_wr *wr, u8 *len16) { if (wr->num_sge > 1) return -EINVAL; - if (wr->num_sge) { + if (wr->num_sge && wr->sg_list[0].length) { wqe->read.stag_src = cpu_to_be32(rdma_wr(wr)->rkey); wqe->read.to_src_hi = cpu_to_be32((u32)(rdma_wr(wr)->remote_addr >> 32)); diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c index 2ba00b89df6a..94b54850ec75 100644 --- a/drivers/infiniband/hw/hfi1/chip.c +++ b/drivers/infiniband/hw/hfi1/chip.c @@ -12847,7 +12847,12 @@ static void remap_intr(struct hfi1_devdata *dd, int isrc, int msix_intr) /* clear from the handled mask of the general interrupt */ m = isrc / 64; n = isrc % 64; - dd->gi_mask[m] &= ~((u64)1 << n); + if (likely(m < CCE_NUM_INT_CSRS)) { + dd->gi_mask[m] &= ~((u64)1 << n); + } else { + dd_dev_err(dd, "remap interrupt err\n"); + return; + } /* direct the chip source to the given MSI-X interrupt */ m = isrc / 8; diff --git a/drivers/infiniband/hw/hfi1/qp.c b/drivers/infiniband/hw/hfi1/qp.c index 650305cc0373..1a7af9f60c13 100644 --- a/drivers/infiniband/hw/hfi1/qp.c +++ b/drivers/infiniband/hw/hfi1/qp.c @@ -647,18 +647,17 @@ void qp_iter_print(struct seq_file *s, struct qp_iter *iter) qp->pid); } -void *qp_priv_alloc(struct rvt_dev_info *rdi, struct rvt_qp *qp, - gfp_t gfp) +void *qp_priv_alloc(struct rvt_dev_info *rdi, struct rvt_qp *qp) { struct hfi1_qp_priv *priv; - priv = kzalloc_node(sizeof(*priv), gfp, rdi->dparms.node); + priv = kzalloc_node(sizeof(*priv), GFP_KERNEL, rdi->dparms.node); if (!priv) return ERR_PTR(-ENOMEM); priv->owner = qp; - priv->s_ahg = kzalloc_node(sizeof(*priv->s_ahg), gfp, + priv->s_ahg = kzalloc_node(sizeof(*priv->s_ahg), GFP_KERNEL, rdi->dparms.node); if (!priv->s_ahg) { kfree(priv); diff --git a/drivers/infiniband/hw/hfi1/qp.h b/drivers/infiniband/hw/hfi1/qp.h index 1eb9cd7b8c19..6fe542b6a927 100644 --- a/drivers/infiniband/hw/hfi1/qp.h +++ b/drivers/infiniband/hw/hfi1/qp.h @@ -123,8 +123,7 @@ void hfi1_migrate_qp(struct rvt_qp *qp); /* * Functions provided by hfi1 driver for rdmavt to use */ -void *qp_priv_alloc(struct rvt_dev_info *rdi, struct rvt_qp *qp, - gfp_t gfp); +void *qp_priv_alloc(struct rvt_dev_info *rdi, struct rvt_qp *qp); void qp_priv_free(struct rvt_dev_info *rdi, struct rvt_qp *qp); unsigned free_all_qps(struct rvt_dev_info *rdi); void notify_qp_reset(struct rvt_qp *qp); diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c index 37d5d29597a4..23fad6d96944 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c @@ -228,14 +228,14 @@ int hns_roce_v1_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, switch (wr->opcode) { case IB_WR_RDMA_READ: ps_opcode = HNS_ROCE_WQE_OPCODE_RDMA_READ; - set_raddr_seg(wqe, atomic_wr(wr)->remote_addr, - atomic_wr(wr)->rkey); + set_raddr_seg(wqe, rdma_wr(wr)->remote_addr, + rdma_wr(wr)->rkey); break; case IB_WR_RDMA_WRITE: case IB_WR_RDMA_WRITE_WITH_IMM: ps_opcode = HNS_ROCE_WQE_OPCODE_RDMA_WRITE; - set_raddr_seg(wqe, atomic_wr(wr)->remote_addr, - atomic_wr(wr)->rkey); + set_raddr_seg(wqe, rdma_wr(wr)->remote_addr, + rdma_wr(wr)->rkey); break; case IB_WR_SEND: case IB_WR_SEND_WITH_INV: @@ -661,9 +661,11 @@ static int hns_roce_v1_rsv_lp_qp(struct hns_roce_dev *hr_dev) union ib_gid dgid; u64 subnet_prefix; int attr_mask = 0; - int i; + int i, j; int ret; + u8 queue_en[HNS_ROCE_V1_RESV_QP] = { 0 }; u8 phy_port; + u8 port = 0; u8 sl; priv = (struct hns_roce_v1_priv *)hr_dev->hw->priv; @@ -709,11 +711,27 @@ static int hns_roce_v1_rsv_lp_qp(struct hns_roce_dev *hr_dev) attr.rnr_retry = 7; attr.timeout = 0x12; attr.path_mtu = IB_MTU_256; + attr.ah_attr.type = RDMA_AH_ATTR_TYPE_ROCE; rdma_ah_set_grh(&attr.ah_attr, NULL, 0, 0, 1, 0); rdma_ah_set_static_rate(&attr.ah_attr, 3); subnet_prefix = cpu_to_be64(0xfe80000000000000LL); for (i = 0; i < HNS_ROCE_V1_RESV_QP; i++) { + phy_port = (i >= HNS_ROCE_MAX_PORTS) ? (i - 2) : + (i % HNS_ROCE_MAX_PORTS); + sl = i / HNS_ROCE_MAX_PORTS; + + for (j = 0; j < caps->num_ports; j++) { + if (hr_dev->iboe.phy_port[j] == phy_port) { + queue_en[i] = 1; + port = j; + break; + } + } + + if (!queue_en[i]) + continue; + free_mr->mr_free_qp[i] = hns_roce_v1_create_lp_qp(hr_dev, pd); if (IS_ERR(free_mr->mr_free_qp[i])) { dev_err(dev, "Create loop qp failed!\n"); @@ -721,15 +739,7 @@ static int hns_roce_v1_rsv_lp_qp(struct hns_roce_dev *hr_dev) } hr_qp = free_mr->mr_free_qp[i]; - sl = i / caps->num_ports; - - if (caps->num_ports == HNS_ROCE_MAX_PORTS) - phy_port = (i >= HNS_ROCE_MAX_PORTS) ? (i - 2) : - (i % caps->num_ports); - else - phy_port = i % caps->num_ports; - - hr_qp->port = phy_port + 1; + hr_qp->port = port; hr_qp->phy_port = phy_port; hr_qp->ibqp.qp_type = IB_QPT_RC; hr_qp->ibqp.device = &hr_dev->ib_dev; @@ -739,23 +749,22 @@ static int hns_roce_v1_rsv_lp_qp(struct hns_roce_dev *hr_dev) hr_qp->ibqp.recv_cq = cq; hr_qp->ibqp.send_cq = cq; - rdma_ah_set_port_num(&attr.ah_attr, phy_port + 1); - rdma_ah_set_sl(&attr.ah_attr, phy_port + 1); - attr.port_num = phy_port + 1; + rdma_ah_set_port_num(&attr.ah_attr, port + 1); + rdma_ah_set_sl(&attr.ah_attr, sl); + attr.port_num = port + 1; attr.dest_qp_num = hr_qp->qpn; memcpy(rdma_ah_retrieve_dmac(&attr.ah_attr), - hr_dev->dev_addr[phy_port], + hr_dev->dev_addr[port], MAC_ADDR_OCTET_NUM); memcpy(&dgid.raw, &subnet_prefix, sizeof(u64)); - memcpy(&dgid.raw[8], hr_dev->dev_addr[phy_port], 3); - memcpy(&dgid.raw[13], hr_dev->dev_addr[phy_port] + 3, 3); + memcpy(&dgid.raw[8], hr_dev->dev_addr[port], 3); + memcpy(&dgid.raw[13], hr_dev->dev_addr[port] + 3, 3); dgid.raw[11] = 0xff; dgid.raw[12] = 0xfe; dgid.raw[8] ^= 2; rdma_ah_set_dgid_raw(&attr.ah_attr, dgid.raw); - attr_mask |= IB_QP_PORT; ret = hr_dev->hw->modify_qp(&hr_qp->ibqp, &attr, attr_mask, IB_QPS_RESET, IB_QPS_INIT); @@ -812,6 +821,9 @@ static void hns_roce_v1_release_lp_qp(struct hns_roce_dev *hr_dev) for (i = 0; i < HNS_ROCE_V1_RESV_QP; i++) { hr_qp = free_mr->mr_free_qp[i]; + if (!hr_qp) + continue; + ret = hns_roce_v1_destroy_qp(&hr_qp->ibqp); if (ret) dev_err(dev, "Destroy qp %d for mr free failed(%d)!\n", @@ -963,7 +975,7 @@ static void hns_roce_v1_mr_free_work_fn(struct work_struct *work) msecs_to_jiffies(HNS_ROCE_V1_FREE_MR_TIMEOUT_MSECS) + jiffies; int i; int ret; - int ne; + int ne = 0; mr_work = container_of(work, struct hns_roce_mr_free_work, work); hr_mr = (struct hns_roce_mr *)mr_work->mr; @@ -976,6 +988,10 @@ static void hns_roce_v1_mr_free_work_fn(struct work_struct *work) for (i = 0; i < HNS_ROCE_V1_RESV_QP; i++) { hr_qp = free_mr->mr_free_qp[i]; + if (!hr_qp) + continue; + ne++; + ret = hns_roce_v1_send_lp_wqe(hr_qp); if (ret) { dev_err(dev, @@ -985,7 +1001,6 @@ static void hns_roce_v1_mr_free_work_fn(struct work_struct *work) } } - ne = HNS_ROCE_V1_RESV_QP; do { ret = hns_roce_v1_poll_cq(&mr_free_cq->ib_cq, ne, wc); if (ret < 0) { @@ -995,7 +1010,8 @@ static void hns_roce_v1_mr_free_work_fn(struct work_struct *work) goto free_work; } ne -= ret; - msleep(HNS_ROCE_V1_FREE_MR_WAIT_VALUE); + usleep_range(HNS_ROCE_V1_FREE_MR_WAIT_VALUE * 1000, + (1 + HNS_ROCE_V1_FREE_MR_WAIT_ |