diff options
Diffstat (limited to 'drivers/infiniband')
31 files changed, 705 insertions, 169 deletions
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c index 33f9d02f9b60..7acc0f936dad 100644 --- a/drivers/infiniband/core/cache.c +++ b/drivers/infiniband/core/cache.c @@ -400,6 +400,9 @@ static void del_gid(struct ib_device *ib_dev, u32 port, table->data_vec[ix] = NULL; write_unlock_irq(&table->rwlock); + if (rdma_cap_roce_gid_table(ib_dev, port)) + ib_dev->ops.del_gid(&entry->attr, &entry->context); + ndev_storage = entry->ndev_storage; if (ndev_storage) { entry->ndev_storage = NULL; @@ -407,9 +410,6 @@ static void del_gid(struct ib_device *ib_dev, u32 port, call_rcu(&ndev_storage->rcu_head, put_gid_ndev); } - if (rdma_cap_roce_gid_table(ib_dev, port)) - ib_dev->ops.del_gid(&entry->attr, &entry->context); - put_gid_entry_locked(entry); } diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 8bd6cb867381..c343edf2f664 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -4084,6 +4084,8 @@ static int resolve_prepare_src(struct rdma_id_private *id_priv, RDMA_CM_ADDR_QUERY))) return -EINVAL; + } else { + memcpy(cma_dst_addr(id_priv), dst_addr, rdma_addr_size(dst_addr)); } if (cma_family(id_priv) != dst_addr->sa_family) { diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c index 755a9c57db6f..f9ab671c8eda 100644 --- a/drivers/infiniband/core/umem.c +++ b/drivers/infiniband/core/umem.c @@ -85,6 +85,8 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, dma_addr_t mask; int i; + umem->iova = va = virt; + if (umem->is_odp) { unsigned int page_size = BIT(to_ib_umem_odp(umem)->page_shift); @@ -100,7 +102,6 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, */ pgsz_bitmap &= GENMASK(BITS_PER_LONG - 1, PAGE_SHIFT); - umem->iova = va = virt; /* The best result is the smallest page size that results in the minimum * number of required pages. Compute the largest page size that could * work based on VA address bits that don't change. diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index 34c806a62269..0848c2c2ffcf 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -877,7 +877,10 @@ fail: int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) { struct bnxt_re_qp *qp = container_of(ib_qp, struct bnxt_re_qp, ib_qp); + struct bnxt_qplib_qp *qplib_qp = &qp->qplib_qp; struct bnxt_re_dev *rdev = qp->rdev; + struct bnxt_qplib_nq *scq_nq = NULL; + struct bnxt_qplib_nq *rcq_nq = NULL; unsigned int flags; int rc; @@ -911,6 +914,15 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) ib_umem_release(qp->rumem); ib_umem_release(qp->sumem); + /* Flush all the entries of notification queue associated with + * given qp. + */ + scq_nq = qplib_qp->scq->nq; + rcq_nq = qplib_qp->rcq->nq; + bnxt_re_synchronize_nq(scq_nq); + if (scq_nq != rcq_nq) + bnxt_re_synchronize_nq(rcq_nq); + return 0; } diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c index 8654ff9744a1..c9066aade412 100644 --- a/drivers/infiniband/hw/bnxt_re/main.c +++ b/drivers/infiniband/hw/bnxt_re/main.c @@ -1462,6 +1462,8 @@ static int bnxt_re_dev_init(struct bnxt_re_dev *rdev, u8 wqe_mode) rc = bnxt_re_setup_chip_ctx(rdev, wqe_mode); if (rc) { + bnxt_unregister_dev(rdev->en_dev); + clear_bit(BNXT_RE_FLAG_NETDEV_REGISTERED, &rdev->flags); ibdev_err(&rdev->ibdev, "Failed to get chip context\n"); return -EINVAL; } @@ -1743,8 +1745,8 @@ static void bnxt_re_remove(struct auxiliary_device *adev) } bnxt_re_setup_cc(rdev, false); ib_unregister_device(&rdev->ibdev); - ib_dealloc_device(&rdev->ibdev); bnxt_re_dev_uninit(rdev); + ib_dealloc_device(&rdev->ibdev); skip_remove: mutex_unlock(&bnxt_re_mutex); } diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index db9890e14ae9..abbabea7f5fa 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -381,6 +381,24 @@ static void bnxt_qplib_service_nq(struct tasklet_struct *t) spin_unlock_bh(&hwq->lock); } +/* bnxt_re_synchronize_nq - self polling notification queue. + * @nq - notification queue pointer + * + * This function will start polling entries of a given notification queue + * for all pending entries. + * This function is useful to synchronize notification entries while resources + * are going away. + */ + +void bnxt_re_synchronize_nq(struct bnxt_qplib_nq *nq) +{ + int budget = nq->budget; + + nq->budget = nq->hwq.max_elements; + bnxt_qplib_service_nq(&nq->nq_tasklet); + nq->budget = budget; +} + static irqreturn_t bnxt_qplib_nq_irq(int irq, void *dev_instance) { struct bnxt_qplib_nq *nq = dev_instance; @@ -402,19 +420,19 @@ void bnxt_qplib_nq_stop_irq(struct bnxt_qplib_nq *nq, bool kill) if (!nq->requested) return; - tasklet_disable(&nq->nq_tasklet); + nq->requested = false; /* Mask h/w interrupt */ bnxt_qplib_ring_nq_db(&nq->nq_db.dbinfo, nq->res->cctx, false); /* Sync with last running IRQ handler */ synchronize_irq(nq->msix_vec); - if (kill) - tasklet_kill(&nq->nq_tasklet); - irq_set_affinity_hint(nq->msix_vec, NULL); free_irq(nq->msix_vec, nq); kfree(nq->name); nq->name = NULL; - nq->requested = false; + + if (kill) + tasklet_kill(&nq->nq_tasklet); + tasklet_disable(&nq->nq_tasklet); } void bnxt_qplib_disable_nq(struct bnxt_qplib_nq *nq) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.h b/drivers/infiniband/hw/bnxt_re/qplib_fp.h index a42820821c47..404b851091ca 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.h @@ -553,6 +553,7 @@ int bnxt_qplib_process_flush_list(struct bnxt_qplib_cq *cq, struct bnxt_qplib_cqe *cqe, int num_cqes); void bnxt_qplib_flush_cqn_wq(struct bnxt_qplib_qp *qp); +void bnxt_re_synchronize_nq(struct bnxt_qplib_nq *nq); static inline void *bnxt_qplib_get_swqe(struct bnxt_qplib_q *que, u32 *swq_idx) { diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c index 524a5ff58872..c8c4017fe405 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c @@ -990,19 +990,18 @@ void bnxt_qplib_rcfw_stop_irq(struct bnxt_qplib_rcfw *rcfw, bool kill) if (!creq->requested) return; - tasklet_disable(&creq->creq_tasklet); + creq->requested = false; /* Mask h/w interrupts */ bnxt_qplib_ring_nq_db(&creq->creq_db.dbinfo, rcfw->res->cctx, false); /* Sync with last running IRQ-handler */ synchronize_irq(creq->msix_vec); - if (kill) - tasklet_kill(&creq->creq_tasklet); - free_irq(creq->msix_vec, rcfw); kfree(creq->irq_name); creq->irq_name = NULL; - creq->requested = false; atomic_set(&rcfw->rcfw_intr_enabled, 0); + if (kill) + tasklet_kill(&creq->creq_tasklet); + tasklet_disable(&creq->creq_tasklet); } void bnxt_qplib_disable_rcfw_channel(struct bnxt_qplib_rcfw *rcfw) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.c b/drivers/infiniband/hw/bnxt_re/qplib_res.c index 79c43c26d511..157db6b7e119 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_res.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_res.c @@ -833,6 +833,7 @@ static int bnxt_qplib_alloc_dpi_tbl(struct bnxt_qplib_res *res, } memset((u8 *)dpit->tbl, 0xFF, bytes); + mutex_init(&res->dpi_tbl_lock); dpit->priv_db = dpit->ucreg.bar_reg + dpit->ucreg.offset; return 0; diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c index e954af824f37..0814291a0412 100644 --- a/drivers/infiniband/hw/hfi1/chip.c +++ b/drivers/infiniband/hw/hfi1/chip.c @@ -12309,6 +12309,7 @@ static void free_cntrs(struct hfi1_devdata *dd) if (dd->synth_stats_timer.function) del_timer_sync(&dd->synth_stats_timer); + cancel_work_sync(&dd->update_cntr_work); ppd = (struct hfi1_pportdata *)(dd + 1); for (i = 0; i < dd->num_pports; i++, ppd++) { kfree(ppd->cntrs); diff --git a/drivers/infiniband/hw/irdma/ctrl.c b/drivers/infiniband/hw/irdma/ctrl.c index 55421a92882c..8a6200e55c54 100644 --- a/drivers/infiniband/hw/irdma/ctrl.c +++ b/drivers/infiniband/hw/irdma/ctrl.c @@ -2712,13 +2712,13 @@ static int irdma_sc_cq_modify(struct irdma_sc_cq *cq, */ void irdma_check_cqp_progress(struct irdma_cqp_timeout *timeout, struct irdma_sc_dev *dev) { - if (timeout->compl_cqp_cmds != dev->cqp_cmd_stats[IRDMA_OP_CMPL_CMDS]) { - timeout->compl_cqp_cmds = dev->cqp_cmd_stats[IRDMA_OP_CMPL_CMDS]; + u64 completed_ops = atomic64_read(&dev->cqp->completed_ops); + + if (timeout->compl_cqp_cmds != completed_ops) { + timeout->compl_cqp_cmds = completed_ops; timeout->count = 0; - } else { - if (dev->cqp_cmd_stats[IRDMA_OP_REQ_CMDS] != - timeout->compl_cqp_cmds) - timeout->count++; + } else if (timeout->compl_cqp_cmds != dev->cqp->requested_ops) { + timeout->count++; } } @@ -2761,7 +2761,7 @@ static int irdma_cqp_poll_registers(struct irdma_sc_cqp *cqp, u32 tail, if (newtail != tail) { /* SUCCESS */ IRDMA_RING_MOVE_TAIL(cqp->sq_ring); - cqp->dev->cqp_cmd_stats[IRDMA_OP_CMPL_CMDS]++; + atomic64_inc(&cqp->completed_ops); return 0; } udelay(cqp->dev->hw_attrs.max_sleep_count); @@ -3121,8 +3121,8 @@ int irdma_sc_cqp_init(struct irdma_sc_cqp *cqp, info->dev->cqp = cqp; IRDMA_RING_INIT(cqp->sq_ring, cqp->sq_size); - cqp->dev->cqp_cmd_stats[IRDMA_OP_REQ_CMDS] = 0; - cqp->dev->cqp_cmd_stats[IRDMA_OP_CMPL_CMDS] = 0; + cqp->requested_ops = 0; + atomic64_set(&cqp->completed_ops, 0); /* for the cqp commands backlog. */ INIT_LIST_HEAD(&cqp->dev->cqp_cmd_head); @@ -3274,7 +3274,7 @@ __le64 *irdma_sc_cqp_get_next_send_wqe_idx(struct irdma_sc_cqp *cqp, u64 scratch if (ret_code) return NULL; - cqp->dev->cqp_cmd_stats[IRDMA_OP_REQ_CMDS]++; + cqp->requested_ops++; if (!*wqe_idx) cqp->polarity = !cqp->polarity; wqe = cqp->sq_base[*wqe_idx].elem; @@ -3363,6 +3363,9 @@ int irdma_sc_ccq_get_cqe_info(struct irdma_sc_cq *ccq, if (polarity != ccq->cq_uk.polarity) return -ENOENT; + /* Ensure CEQE contents are read after valid bit is checked */ + dma_rmb(); + get_64bit_val(cqe, 8, &qp_ctx); cqp = (struct irdma_sc_cqp *)(unsigned long)qp_ctx; info->error = (bool)FIELD_GET(IRDMA_CQ_ERROR, temp); @@ -3397,7 +3400,7 @@ int irdma_sc_ccq_get_cqe_info(struct irdma_sc_cq *ccq, dma_wmb(); /* make sure shadow area is updated before moving tail */ IRDMA_RING_MOVE_TAIL(cqp->sq_ring); - ccq->dev->cqp_cmd_stats[IRDMA_OP_CMPL_CMDS]++; + atomic64_inc(&cqp->completed_ops); return ret_code; } @@ -4008,13 +4011,17 @@ int irdma_sc_get_next_aeqe(struct irdma_sc_aeq *aeq, u8 polarity; aeqe = IRDMA_GET_CURRENT_AEQ_ELEM(aeq); - get_64bit_val(aeqe, 0, &compl_ctx); get_64bit_val(aeqe, 8, &temp); polarity = (u8)FIELD_GET(IRDMA_AEQE_VALID, temp); if (aeq->polarity != polarity) return -ENOENT; + /* Ensure AEQE contents are read after valid bit is checked */ + dma_rmb(); + + get_64bit_val(aeqe, 0, &compl_ctx); + print_hex_dump_debug("WQE: AEQ_ENTRY WQE", DUMP_PREFIX_OFFSET, 16, 8, aeqe, 16, false); diff --git a/drivers/infiniband/hw/irdma/defs.h b/drivers/infiniband/hw/irdma/defs.h index 6014b9d06a9b..d06e45d2c23f 100644 --- a/drivers/infiniband/hw/irdma/defs.h +++ b/drivers/infiniband/hw/irdma/defs.h @@ -191,32 +191,30 @@ enum irdma_cqp_op_type { IRDMA_OP_MANAGE_VF_PBLE_BP = 25, IRDMA_OP_QUERY_FPM_VAL = 26, IRDMA_OP_COMMIT_FPM_VAL = 27, - IRDMA_OP_REQ_CMDS = 28, - IRDMA_OP_CMPL_CMDS = 29, - IRDMA_OP_AH_CREATE = 30, - IRDMA_OP_AH_MODIFY = 31, - IRDMA_OP_AH_DESTROY = 32, - IRDMA_OP_MC_CREATE = 33, - IRDMA_OP_MC_DESTROY = 34, - IRDMA_OP_MC_MODIFY = 35, - IRDMA_OP_STATS_ALLOCATE = 36, - IRDMA_OP_STATS_FREE = 37, - IRDMA_OP_STATS_GATHER = 38, - IRDMA_OP_WS_ADD_NODE = 39, - IRDMA_OP_WS_MODIFY_NODE = 40, - IRDMA_OP_WS_DELETE_NODE = 41, - IRDMA_OP_WS_FAILOVER_START = 42, - IRDMA_OP_WS_FAILOVER_COMPLETE = 43, - IRDMA_OP_SET_UP_MAP = 44, - IRDMA_OP_GEN_AE = 45, - IRDMA_OP_QUERY_RDMA_FEATURES = 46, - IRDMA_OP_ALLOC_LOCAL_MAC_ENTRY = 47, - IRDMA_OP_ADD_LOCAL_MAC_ENTRY = 48, - IRDMA_OP_DELETE_LOCAL_MAC_ENTRY = 49, - IRDMA_OP_CQ_MODIFY = 50, + IRDMA_OP_AH_CREATE = 28, + IRDMA_OP_AH_MODIFY = 29, + IRDMA_OP_AH_DESTROY = 30, + IRDMA_OP_MC_CREATE = 31, + IRDMA_OP_MC_DESTROY = 32, + IRDMA_OP_MC_MODIFY = 33, + IRDMA_OP_STATS_ALLOCATE = 34, + IRDMA_OP_STATS_FREE = 35, + IRDMA_OP_STATS_GATHER = 36, + IRDMA_OP_WS_ADD_NODE = 37, + IRDMA_OP_WS_MODIFY_NODE = 38, + IRDMA_OP_WS_DELETE_NODE = 39, + IRDMA_OP_WS_FAILOVER_START = 40, + IRDMA_OP_WS_FAILOVER_COMPLETE = 41, + IRDMA_OP_SET_UP_MAP = 42, + IRDMA_OP_GEN_AE = 43, + IRDMA_OP_QUERY_RDMA_FEATURES = 44, + IRDMA_OP_ALLOC_LOCAL_MAC_ENTRY = 45, + IRDMA_OP_ADD_LOCAL_MAC_ENTRY = 46, + IRDMA_OP_DELETE_LOCAL_MAC_ENTRY = 47, + IRDMA_OP_CQ_MODIFY = 48, /* Must be last entry*/ - IRDMA_MAX_CQP_OPS = 51, + IRDMA_MAX_CQP_OPS = 49, }; /* CQP SQ WQES */ diff --git a/drivers/infiniband/hw/irdma/hw.c b/drivers/infiniband/hw/irdma/hw.c index 9f5d418c6e4a..7cbdd5433dba 100644 --- a/drivers/infiniband/hw/irdma/hw.c +++ b/drivers/infiniband/hw/irdma/hw.c @@ -191,6 +191,7 @@ static void irdma_set_flush_fields(struct irdma_sc_qp *qp, case IRDMA_AE_AMP_MWBIND_INVALID_RIGHTS: case IRDMA_AE_AMP_MWBIND_BIND_DISABLED: case IRDMA_AE_AMP_MWBIND_INVALID_BOUNDS: + case IRDMA_AE_AMP_MWBIND_VALID_STAG: qp->flush_code = FLUSH_MW_BIND_ERR; qp->event_type = IRDMA_QP_EVENT_ACCESS_ERR; break; @@ -2092,7 +2093,7 @@ void irdma_cqp_ce_handler(struct irdma_pci_f *rf, struct irdma_sc_cq *cq) cqp_request->compl_info.error = info.error; if (cqp_request->waiting) { - cqp_request->request_done = true; + WRITE_ONCE(cqp_request->request_done, true); wake_up(&cqp_request->waitq); irdma_put_cqp_request(&rf->cqp, cqp_request); } else { diff --git a/drivers/infiniband/hw/irdma/main.h b/drivers/infiniband/hw/irdma/main.h index 282cd492fe9b..82fc5f5b002c 100644 --- a/drivers/infiniband/hw/irdma/main.h +++ b/drivers/infiniband/hw/irdma/main.h @@ -161,8 +161,8 @@ struct irdma_cqp_request { void (*callback_fcn)(struct irdma_cqp_request *cqp_request); void *param; struct irdma_cqp_compl_info compl_info; + bool request_done; /* READ/WRITE_ONCE macros operate on it */ bool waiting:1; - bool request_done:1; bool dynamic:1; }; diff --git a/drivers/infiniband/hw/irdma/puda.c b/drivers/infiniband/hw/irdma/puda.c index 4ec9639f1bdb..562531712ea4 100644 --- a/drivers/infiniband/hw/irdma/puda.c +++ b/drivers/infiniband/hw/irdma/puda.c @@ -230,6 +230,9 @@ static int irdma_puda_poll_info(struct irdma_sc_cq *cq, if (valid_bit != cq_uk->polarity) return -ENOENT; + /* Ensure CQE contents are read after valid bit is checked */ + dma_rmb(); + if (cq->dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) ext_valid = (bool)FIELD_GET(IRDMA_CQ_EXTCQE, qword3); @@ -243,6 +246,9 @@ static int irdma_puda_poll_info(struct irdma_sc_cq *cq, if (polarity != cq_uk->polarity) return -ENOENT; + /* Ensure ext CQE contents are read after ext valid bit is checked */ + dma_rmb(); + IRDMA_RING_MOVE_HEAD_NOCHECK(cq_uk->cq_ring); if (!IRDMA_RING_CURRENT_HEAD(cq_uk->cq_ring)) cq_uk->polarity = !cq_uk->polarity; diff --git a/drivers/infiniband/hw/irdma/type.h b/drivers/infiniband/hw/irdma/type.h index bee9609f4be7..c84ec4dd8536 100644 --- a/drivers/infiniband/hw/irdma/type.h +++ b/drivers/infiniband/hw/irdma/type.h @@ -365,6 +365,8 @@ struct irdma_sc_cqp { struct irdma_dcqcn_cc_params dcqcn_params; __le64 *host_ctx; u64 *scratch_array; + u64 requested_ops; + atomic64_t completed_ops; u32 cqp_id; u32 sq_size; u32 hw_sq_size; diff --git a/drivers/infiniband/hw/irdma/uk.c b/drivers/infiniband/hw/irdma/uk.c index 9f84d997e3ab..d8285ca16293 100644 --- a/drivers/infiniband/hw/irdma/uk.c +++ b/drivers/infiniband/hw/irdma/uk.c @@ -1083,7 +1083,7 @@ 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); + info->op_type = (u8)FIELD_GET(IRDMACQ_OP, qword3); if (info->q_type == IRDMA_CQE_QTYPE_RQ) { u32 array_idx; @@ -1508,6 +1508,9 @@ void irdma_uk_clean_cq(void *q, struct irdma_cq_uk *cq) if (polarity != temp) break; + /* Ensure CQE contents are read after valid bit is checked */ + dma_rmb(); + get_64bit_val(cqe, 8, &comp_ctx); if ((void *)(unsigned long)comp_ctx == q) set_64bit_val(cqe, 8, 0); diff --git a/drivers/infiniband/hw/irdma/utils.c b/drivers/infiniband/hw/irdma/utils.c index 1008e158bba2..6cd5cb85dafe 100644 --- a/drivers/infiniband/hw/irdma/utils.c +++ b/drivers/infiniband/hw/irdma/utils.c @@ -481,7 +481,7 @@ void irdma_free_cqp_request(struct irdma_cqp *cqp, if (cqp_request->dynamic) { kfree(cqp_request); } else { - cqp_request->request_done = false; + WRITE_ONCE(cqp_request->request_done, false); cqp_request->callback_fcn = NULL; cqp_request->waiting = false; @@ -515,7 +515,7 @@ irdma_free_pending_cqp_request(struct irdma_cqp *cqp, { if (cqp_request->waiting) { cqp_request->compl_info.error = true; - cqp_request->request_done = true; + WRITE_ONCE(cqp_request->request_done, true); wake_up(&cqp_request->waitq); } wait_event_timeout(cqp->remove_wq, @@ -567,11 +567,11 @@ static int irdma_wait_event(struct irdma_pci_f *rf, bool cqp_error = false; int err_code = 0; - cqp_timeout.compl_cqp_cmds = rf->sc_dev.cqp_cmd_stats[IRDMA_OP_CMPL_CMDS]; + cqp_timeout.compl_cqp_cmds = atomic64_read(&rf->sc_dev.cqp->completed_ops); do { irdma_cqp_ce_handler(rf, &rf->ccq.sc_cq); if (wait_event_timeout(cqp_request->waitq, - cqp_request->request_done, + READ_ONCE(cqp_request->request_done), msecs_to_jiffies(CQP_COMPL_WAIT_TIME_MS))) break; diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c index 2d2cd17e02e6..529db874d67c 100644 --- a/drivers/infiniband/hw/mlx4/main.c +++ b/drivers/infiniband/hw/mlx4/main.c @@ -82,6 +82,8 @@ static const char mlx4_ib_version[] = static void do_slave_init(struct mlx4_ib_dev *ibdev, int slave, int do_init); static enum rdma_link_layer mlx4_ib_port_link_layer(struct ib_device *device, u32 port_num); +static int mlx4_ib_event(struct notifier_block *this, unsigned long event, + void *param); static struct workqueue_struct *wq; @@ -125,12 +127,14 @@ static struct net_device *mlx4_ib_get_netdev(struct ib_device *device, u32 port_num) { struct mlx4_ib_dev *ibdev = to_mdev(device); - struct net_device *dev; + struct net_device *dev, *ret = NULL; rcu_read_lock(); - dev = mlx4_get_protocol_dev(ibdev->dev, MLX4_PROT_ETH, port_num); + for_each_netdev_rcu(&init_net, dev) { + if (dev->dev.parent != ibdev->ib_dev.dev.parent || + dev->dev_port + 1 != port_num) + continue; - if (dev) { if (mlx4_is_bonded(ibdev->dev)) { struct net_device *upper; @@ -143,11 +147,14 @@ static struct net_device *mlx4_ib_get_netdev(struct ib_device *device, dev = active; } } + + dev_hold(dev); + ret = dev; + break; } - dev_hold(dev); rcu_read_unlock(); - return dev; + return ret; } static int mlx4_ib_update_gids_v1(struct gid_entry *gids, @@ -2318,61 +2325,53 @@ unlock: mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]); } -static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev, - struct net_device *dev, - unsigned long event) +static void mlx4_ib_scan_netdev(struct mlx4_ib_dev *ibdev, + struct net_device *dev, + unsigned long event) { - struct mlx4_ib_iboe *iboe; - int update_qps_port = -1; - int port; + struct mlx4_ib_iboe *iboe = &ibdev->iboe; ASSERT_RTNL(); - iboe = &ibdev->iboe; + if (dev->dev.parent != ibdev->ib_dev.dev.parent) + return; spin_lock_bh(&iboe->lock); - mlx4_foreach_ib_transport_port(port, ibdev->dev) { - - iboe->netdevs[port - 1] = - mlx4_get_protocol_dev(ibdev->dev, MLX4_PROT_ETH, port); - if (dev == iboe->netdevs[port - 1] && - (event == NETDEV_CHANGEADDR || event == NETDEV_REGISTER || - event == NETDEV_UP || event == NETDEV_CHANGE)) - update_qps_port = port; + iboe->netdevs[dev->dev_port] = event != NETDEV_UNREGISTER ? dev : NULL; - if (dev == iboe->netdevs[port - 1] && - (event == NETDEV_UP || event == NETDEV_DOWN)) { - enum ib_port_state port_state; - struct ib_event ibev = { }; + if (event == NETDEV_UP || event == NETDEV_DOWN) { + enum ib_port_state port_state; + struct ib_event ibev = { }; - if (ib_get_cached_port_state(&ibdev->ib_dev, port, - &port_state)) - continue; + if (ib_get_cached_port_state(&ibdev->ib_dev, dev->dev_port + 1, + &port_state)) + goto iboe_out; - if (event == NETDEV_UP && - (port_state != IB_PORT_ACTIVE || - iboe->last_port_state[port - 1] != IB_PORT_DOWN)) - continue; - if (event == NETDEV_DOWN && - (port_state != IB_PORT_DOWN || - iboe->last_port_state[port - 1] != IB_PORT_ACTIVE)) - continue; - iboe->last_port_state[port - 1] = port_state; - - ibev.device = &ibdev->ib_dev; - ibev.element.port_num = port; - ibev.event = event == NETDEV_UP ? IB_EVENT_PORT_ACTIVE : - IB_EVENT_PORT_ERR; - ib_dispatch_event(&ibev); - } + if (event == NETDEV_UP && + (port_state != IB_PORT_ACTIVE || + iboe->last_port_state[dev->dev_port] != IB_PORT_DOWN)) + goto iboe_out; + if (event == NETDEV_DOWN && + (port_state != IB_PORT_DOWN || + iboe->last_port_state[dev->dev_port] != IB_PORT_ACTIVE)) + goto iboe_out; + iboe->last_port_state[dev->dev_port] = port_state; + ibev.device = &ibdev->ib_dev; + ibev.element.port_num = dev->dev_port + 1; + ibev.event = event == NETDEV_UP ? IB_EVENT_PORT_ACTIVE : + IB_EVENT_PORT_ERR; + ib_dispatch_event(&ibev); } + +iboe_out: spin_unlock_bh(&iboe->lock); - if (update_qps_port > 0) - mlx4_ib_update_qps(ibdev, dev, update_qps_port); + if (event == NETDEV_CHANGEADDR || event == NETDEV_REGISTER || + event == NETDEV_UP || event == NETDEV_CHANGE) + mlx4_ib_update_qps(ibdev, dev, dev->dev_port + 1); } static int mlx4_ib_netdev_event(struct notifier_block *this, @@ -2385,7 +2384,7 @@ static int mlx4_ib_netdev_event(struct notifier_block *this, return NOTIFY_DONE; ibdev = container_of(this, struct mlx4_ib_dev, iboe.nb); - mlx4_ib_scan_netdevs(ibdev, dev, event); + mlx4_ib_scan_netdev(ibdev, dev, event); return NOTIFY_DONE; } @@ -2609,8 +2608,11 @@ static const struct ib_device_ops mlx4_ib_dev_fs_ops = { .destroy_flow = mlx4_ib_destroy_flow, }; -static void *mlx4_ib_add(struct mlx4_dev *dev) +static int mlx4_ib_probe(struct auxiliary_device *adev, + const struct auxiliary_device_id *id) { + struct mlx4_adev *madev = container_of(adev, struct mlx4_adev, adev); + struct mlx4_dev *dev = madev->mdev; struct mlx4_ib_dev *ibdev; int num_ports = 0; int i, j; @@ -2630,27 +2632,31 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) /* No point in registering a device with no ports... */ if (num_ports == 0) - return NULL; + return -ENODEV; ibdev = ib_alloc_device(mlx4_ib_dev, ib_dev); if (!ibdev) { dev_err(&dev->persist->pdev->dev, "Device struct alloc failed\n"); - return NULL; + return -ENOMEM; } iboe = &ibdev->iboe; - if (mlx4_pd_alloc(dev, &ibdev->priv_pdn)) + err = mlx4_pd_alloc(dev, &ibdev->priv_pdn); + if (err) goto err_dealloc; - if (mlx4_uar_alloc(dev, &ibdev->priv_uar)) + err = mlx4_uar_alloc(dev, &ibdev->priv_uar); + if (err) goto err_pd; ibdev->uar_map = ioremap((phys_addr_t) ibdev->priv_uar.pfn << PAGE_SHIFT, PAGE_SIZE); - if (!ibdev->uar_map) + if (!ibdev->uar_map) { + err = -ENOMEM; goto err_uar; + } MLX4_INIT_DOORBELL_LOCK(&ibdev->uar_lock); ibdev->dev = dev; @@ -2694,7 +2700,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) spin_lock_init(&iboe->lock); - if (init_node_data(ibdev)) + err = init_node_data(ibdev); + if (err) goto err_map; mlx4_init_sl2vl_tbl(ibdev); @@ -2726,6 +2733,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) new_counter_index = kmalloc(sizeof(*new_counter_index), GFP_KERNEL); if (!new_counter_index) { + err = -ENOMEM; if (allocated) mlx4_counter_free(ibdev->dev, counter_index); goto err_counter; @@ -2743,8 +2751,10 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) new_counter_index = kmalloc(sizeof(struct counter_index), GFP_KERNEL); - if (!new_counter_index) + if (!new_counter_index) { + err = -ENOMEM; goto err_counter; + } new_counter_index->index = counter_index; new_counter_index->allocated = 0; list_add_tail(&new_counter_index->list, @@ -2773,8 +2783,10 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) ibdev->ib_uc_qpns_bitmap = bitmap_alloc(ibdev->steer_qpn_count, GFP_KERNEL); - if (!ibdev->ib_uc_qpns_bitmap) + if (!ibdev->ib_uc_qpns_bitmap) { + err = -ENOMEM; goto err_steer_qp_release; + } if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_DMFS_IPOIB) { bitmap_zero(ibdev->ib_uc_qpns_bitmap, @@ -2794,17 +2806,21 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) for (j = 1; j <= ibdev->dev->caps.num_ports; j++) atomic64_set(&iboe->mac[j - 1], ibdev->dev->caps.def_mac[j]); - if (mlx4_ib_alloc_diag_counters(ibdev)) + err = mlx4_ib_alloc_diag_counters(ibdev); + if (err) goto err_steer_free_bitmap; - if (ib_register_device(&ibdev->ib_dev, "mlx4_%d", - &dev->persist->pdev->dev)) + err = ib_register_device(&ibdev->ib_dev, "mlx4_%d", + &dev->persist->pdev->dev); + if (err) goto err_diag_counters; - if (mlx4_ib_mad_init(ibdev)) + err = mlx4_ib_mad_init(ibdev); + if (err) goto err_reg; - if (mlx4_ib_init_sriov(ibdev)) + err = mlx4_ib_init_sriov(ibdev); + if (err) goto err_mad; if (!iboe->nb.notifier_call) { @@ -2838,7 +2854,14 @@ static void *mlx4_ib_add(struct mlx4_dev *dev) do_slave_init(ibdev, j, 1); } } - return ibdev; + + /* register mlx4 core notifier */ + ibdev->mlx_nb.notifier_call = mlx4_ib_event; + err = mlx4_register_event_notifier(dev, &ibdev->mlx_nb); + WARN(err, "failed to register mlx4 event notifier (%d)", err); + + auxiliary_set_drvdata(adev, ibdev); + return 0; err_notif: if (ibdev->iboe.nb.notifier_call) { @@ -2882,7 +2905,7 @@ err_pd: err_dealloc: ib_dealloc_device(&ibdev->ib_dev); - return NULL; + return err; } int mlx4_ib_steer_qp_alloc(struct mlx4_ib_dev *dev, int count, int *qpn) @@ -2949,12 +2972,16 @@ int mlx4_ib_steer_qp_reg(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp, return __mlx4_ib_destroy_flow(mdev->dev, mqp->reg_id); } -static void mlx4_ib_remove(struct mlx4_dev *dev, void *ibdev_ptr) +static void mlx4_ib_remove(struct auxiliary_device *adev) { - struct mlx4_ib_dev *ibdev = ibdev_ptr; + struct mlx4_adev *madev = container_of(adev, struct mlx4_adev, adev); + struct mlx4_dev *dev = madev->mdev; + struct mlx4_ib_dev *ibdev = auxiliary_get_drvdata(adev); int p; int i; + mlx4_unregister_event_notifier(dev, &ibdev->mlx_nb); + mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB) devlink_port_type_clear(mlx4_get_devlink_port(dev, i)); ibdev->ib_active = false; @@ -3175,11 +3202,13 @@ void mlx4_sched_ib_sl2vl_update_work(struct mlx4_ib_dev *ibdev, } } -static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, - enum mlx4_dev_event event, unsigned long param) +static int mlx4_ib_event(struct notifier_block *this, unsigned long event, + void *param) { + struct mlx4_ib_dev *ibdev = + container_of(this, struct mlx4_ib_dev, mlx_nb); + struct mlx4_dev *dev = ibdev->dev; struct ib_event ibev; - struct mlx4_ib_dev *ibdev = to_mdev((struct ib_device *) ibdev_ptr); struct mlx4_eqe *eqe = NULL; struct ib_event_work *ew; int p = 0; @@ -3189,22 +3218,28 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, (event == MLX4_DEV_EVENT_PORT_DOWN))) { ew = kmalloc(sizeof(*ew), GFP_ATOMIC); if (!ew) - return; + return NOTIFY_DONE; INIT_WORK(&ew->work, handle_bonded_port_state_event); ew->ib_dev = ibdev; queue_work(wq, &ew->work); - return; + return NOTIFY_DONE; } - if (event == MLX4_DEV_EVENT_PORT_MGMT_CHANGE) + switch (event) { + case MLX4_DEV_EVENT_CATASTROPHIC_ERROR: + break; + case MLX4_DEV_EVENT_PORT_MGMT_CHANGE: eqe = (struct mlx4_eqe *)param; - else - p = (int) param; + break; + default: + p = *(int *)param; + break; + } switch (event) { case MLX4_DEV_EVENT_PORT_UP: if (p > ibdev->num_ports) - return; + return NOTIFY_DONE; if (!mlx4_is_slave(dev) && rdma_port_get_link_layer(&ibdev->ib_dev, p) == IB_LINK_LAYER_INFINIBAND) { @@ -3219,7 +3254,7 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, case MLX4_DEV_EVENT_PORT_DOWN: if (p > ibdev->num_ports) - return; + return NOTIFY_DONE; ibev.event = IB_EVENT_PORT_ERR; break; @@ -3232,7 +3267,7 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, case MLX4_DEV_EVENT_PORT_MGMT_CHANGE: ew = kmalloc(sizeof *ew, GFP_ATOMIC); if (!ew) - return; + return NOTIFY_DONE; INIT_WORK(&ew->work, handle_port_mgmt_change_event); memcpy(&ew->ib_eqe, eqe, sizeof *eqe); @@ -3242,7 +3277,7 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, queue_work(wq, &ew->work); else handle_port_mgmt_change_event(&ew->work); - return; + return NOTIFY_DONE; case MLX4_DEV_EVENT_SLAVE_INIT: /* here, p is the slave id */ @@ -3258,7 +3293,7 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, 1); } } - return; + return NOTIFY_DONE; case MLX4_DEV_EVENT_SLAVE_SHUTDOWN: if (mlx4_is_master(dev)) { @@ -3274,22 +3309,33 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr, } /* here, p is the slave id */ do_slave_init(ibdev, p, 0); - return; + return NOTIFY_DONE; default: - return; + return NOTIFY_DONE; } - ibev.device = ibdev_ptr; + ibev.device = &ibdev->ib_dev; ibev.element.port_num = mlx4_is_bonded(ibdev->dev) ? 1 : (u8)p; ib_dispatch_event(&ibev); + return NOTIFY_DONE; } -static struct mlx4_interface mlx4_ib_interface = { - .add = mlx4_ib_add, - .remove = mlx4_ib_remove, - .event = mlx4_ib_event, +static const struct auxiliary_device_id mlx4_ib_id_table[] = { + { .name = MLX4_ADEV_NAME ".ib" }, + {}, +}; + +MODULE_DEVICE_TABLE(auxiliary, mlx4_ib_id_table); + +static struct mlx4_adrv mlx4_ib_adrv = { + .adrv = { + .name = "ib", + .probe = mlx4_ib_probe, + .remove = mlx4_ib_remove, + .id_table = mlx4_ib_id_table, + }, .protocol = MLX4_PROT_IB_IPV6, .flags = MLX4_INTFF_BONDING }; @@ -3314,7 +3360,7 @@ static int __init mlx4_ib_init(void) if (err) goto clean_cm; - err = mlx4_register_interface(&mlx4_ib_interface); + err = mlx4_register_auxiliary_driver(&mlx4_ib_adrv); if (err) goto clean_mcg; @@ -3336,7 +3382,7 @@ clean_qp_event: static void __exit mlx4_ib_cleanup(void) { - mlx4_unregister_interface(&mlx4_ib_interface); + mlx4_unregister_auxiliary_driver(&mlx4_ib_adrv); mlx4_ib_mcg_destroy(); mlx4_ib_cm_destroy(); mlx4_ib_qp_event_cleanup(); diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h index 17fee1e73a45..41ca1114a995 100644 --- a/drivers/infiniband/hw/mlx4/mlx4_ib.h +++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h @@ -38,6 +38,7 @@ #include <linux/list.h> #include <linux/mutex.h> #include <linux/idr.h> +#include <linux/notifier.h> #include <rdma/ib_verbs.h> #include <rdma/ib_umem.h> @@ -644,6 +645,7 @@ struct mlx4_ib_dev { spinlock_t reset_flow_resource_lock; struct list_head qp_list; struct mlx4_ib_diag_counters diag_counters[MLX4_DIAG_COUNTERS_TYPES]; + struct notifier_block mlx_nb; }; struct ib_event_work { diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c index 456656617c33..9d08aa99f3cb 100644 --- a/drivers/infiniband/hw/mlx4/qp.c +++ b/drivers/infiniband/hw/mlx4/qp.c @@ -565,15 +565,15 @@ static int set_qp_rss(struct mlx4_ib_dev *dev, struct mlx4_ib_rss *rss_ctx, return (-EOPNOTSUPP); } - if (ucmd->rx_hash_fields_mask & ~(MLX4_IB_RX_HASH_SRC_IPV4 | - MLX4_IB_RX_HASH_DST_IPV4 | - MLX4_IB_RX_HASH_SRC_IPV6 | - MLX4_IB_RX_HASH_DST_IPV6 | - MLX4_IB_RX_HASH_SRC_PORT_TCP | - MLX4_IB_RX_HASH_DST_PORT_TCP | - MLX4_IB_RX_HASH_SRC_PORT_UDP | - MLX4_IB_RX_HASH_DST_PORT_UDP | - MLX4_IB_RX_HASH_INNER)) { + if (ucmd->rx_hash_fields_mask & ~(u64)(MLX4_IB_RX_HASH_SRC_IPV4 | + MLX4_IB_RX_HASH_DST_IPV4 | + MLX4_IB_RX_HASH_SRC_IPV6 | + MLX4_IB_RX_HASH_DST_IPV6 | + MLX4_IB_RX_HASH_SRC_PORT_TCP | + MLX4_IB_RX_HASH_DST_PORT_TCP | + MLX4_IB_RX_HASH_SRC_PORT_UDP | + MLX4_IB_RX_HASH_DST_PORT_UDP | + MLX4_IB_RX_HASH_INNER)) { pr_debug("RX Hash fields_mask has unsupported mask (0x%llx)\n", ucmd->rx_hash_fields_mask); return (-EOPNOTSUPP); diff --git a/drivers/infiniband/hw/mlx5/Makefile b/drivers/infiniband/hw/mlx5/Makefile index 612ee8190a2d..72a526236c2e 100644 --- a/drivers/infiniband/hw/mlx5/Makefile +++ b/drivers/infiniband/hw/mlx5/Makefile @@ -28,3 +28,4 @@ mlx5_ib-$(CONFIG_INFINIBAND_USER_ACCESS) += devx.o \ fs.o \ qos.o \ std_types.o +mlx5_ib-$(CONFIG_MLX5_MACSEC) += macsec.o diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c index efc9e4a6df04..9773d2a3d97f 100644 --- a/drivers/infiniband/hw/mlx5/cq.c +++ b/drivers/infiniband/hw/mlx5/cq.c @@ -993,7 +993,7 @@ int mlx5_ib_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr, INIT_WORK(&cq->notify_work, notify_soft_wc_handler); } - err = mlx5_vector2eqn(dev->mdev, vector, &eqn); + err = mlx5_comp_eqn_get(dev->mdev, vector, &eqn); if (err) goto err_cqb; diff --git a/drivers/infiniband/hw/mlx5/devx.c b/drivers/infiniband/hw/mlx5/devx.c index db5fb196c728..8ba53edf2311 100644 --- a/drivers/infiniband/hw/mlx5/devx.c +++ b/drivers/infiniband/hw/mlx5/devx.c @@ -1002,7 +1002,7 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_DEVX_QUERY_EQN)( return PTR_ERR(c); dev = to_mdev(c->ibucontext.device); - err = mlx5_vector2eqn(dev->mdev, user_vector, &dev_eqn); + err = mlx5_comp_eqn_get(dev->mdev, user_vector, &dev_eqn); if (err < 0) return err; diff --git a/drivers/infiniband/hw/mlx5/macsec.c b/drivers/infiniband/hw/mlx5/macsec.c new file mode 100644 index 000000000000..3c56eb5eddf3 --- /dev/null +++ b/drivers/infiniband/hw/mlx5/macsec.c @@ -0,0 +1,364 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* Copyright (c) 2022, NVIDIA CORPORATION & AFFILIATES. */ + +#include "macsec.h" +#include <linux/mlx5/macsec.h> + +struct mlx5_reserved_gids { + int macsec_index; + const struct ib_gid_attr *physical_gid; +}; + +struct mlx5_roce_gids { + struct list_head roce_gid_list_entry; + u16 gid_idx; + union { + struct sockaddr_in sockaddr_in; + struct sockaddr_in6 sockaddr_in6; + } addr; +}; + +struct mlx5_macsec_device { + struct list_head macsec_devices_list_entry; + void *macdev; + struct list_head macsec_roce_gids; + struct list_head tx_rules_list; + struct list_head rx_rules_list; +}; + +static void cleanup_macsec_device(struct mlx5_macsec_device *macsec_device) +{ + if (!list_empty(&macsec_device->tx_rules_list) || + !list_empty(&macsec_device->rx_rules_list) || + !list_empty(&macsec_device->macsec_roce_gids)) + return; + + list_del(&macsec_device->macsec_devices_list_entry); + kfree(macsec_device); +} + +static struct mlx5_macsec_device *get_macsec_device(void *macdev, + struct list_head *macsec_devices_list) +{ + struct mlx5_macsec_device *iter, *macsec_device = NULL; + + list_for_each_entry(iter, macsec_devices_list, macsec_devices_list_entry) { + if (iter->macdev == macdev) { + macsec_device = iter; + break; + } + } + + if (macsec_device) + return macsec_device; + + macsec_device = kzalloc(sizeof(*macsec_device), GFP_KERNEL); + if (!macsec_device) + return NULL; + + macsec_device->macdev = macdev; + INIT_LIST_HEAD(&macsec_device->tx_rules_list); + INIT_LIST_HEAD(&macsec_device->rx_rules_list); + INIT_LIST_HEAD(&macsec_device->macsec_roce_gids); + list_add(&macsec_device->macsec_devices_list_entry, macsec_devices_list); + + return macsec_device; +} + +static void mlx5_macsec_del_roce_gid(struct mlx5_macsec_device *macsec_device, u16 gid_idx) +{ + struct mlx5_roce_gids *current_gid, *next_gid; + + list_for_each_entry_safe(current_gid, next_gid, &macsec_device->macsec_roce_gids, + roce_gid_list_entry) + if (current_gid->gid_idx == gid_idx) { + list_del(¤t_gid->roce_gid_list_entry); + kfree(current_gid); + } +} + +static void mlx5_macsec_save_roce_gid(struct mlx5_macsec_device *macsec_device, + const struct sockaddr *addr, u16 gid_idx) +{ + struct mlx5_roce_gids *roce_gids; + + roce_gids = kzalloc(sizeof(*roce_gids), GFP_KERNEL); + if (!roce_gids) + return; + + roce_gids->gid_idx = gid_idx; + if (addr->sa_family == AF_INET) + memcpy(&roce_gids->addr.sockaddr_in, addr, sizeof(roce_gids->addr.sockaddr_in)); + else + memcpy(&roce_gids->addr.sockaddr_in6, addr, sizeof(roce_gids->addr.sockaddr_in6)); + + list_add_tail(&roce_gids->roce_gid_list_entry, &macsec_device->macsec_roce_gids); +} + +static void handle_macsec_gids(struct list_head *macsec_devices_list, + struct mlx5_macsec_event_data *data) +{ + struct mlx5_macsec_device *macsec_device; + struct mlx5_roce_gids *gid; + + macsec_device = get_macsec_device(data->macdev, macsec_devices_list); + if (!macsec_device) + return; + + list_for_each_entry(gid, &macsec_device->macsec_roce_gids, roce_gid_list_entry) { + mlx5_macsec_add_roce_sa_rules(data->fs_id, (struct sockaddr *)&gid->addr, + gid->gid_idx, &macsec_device->tx_rules_list, + &macsec_device->rx_rules_list, data->macsec_fs, + data->is_tx); + } +} + +static void del_sa_roce_rule(struct list_head *macsec_devices_list, + struct mlx5_macsec_event_data *data) +{ + struct mlx5_macsec_device *macsec_device; + + macsec_device = get_macsec_device(data->macdev, macsec_devices_list); + WARN_ON(!macsec_device); + + mlx5_macsec_del_roce_sa_rules(data->fs_id, data->macsec_fs, + &macsec_device->tx_rules_list, + &macsec_device->rx_rules_list, data->is_tx); +} + +static int macsec_event(struct notifier_block *nb, unsigned long event, void *data) +{ + struct mlx5_macsec *macsec = container_of(nb, struct mlx5_macsec, blocking_events_nb); + + mutex_lock(&macsec->lock); + switch (event) { + case MLX5_DRIVER_EVENT_MACSEC_SA_ADDED: + handle_macsec_gids(&macsec->macsec_devices_list, data); + break; + case MLX5_DRIVER_EVENT_MACSEC_SA_DELETED: + del_sa_roce_rule(&macsec->macsec_devices_list, data); + break; + default: + mutex_unlock(&macsec->lock); + return NOTIFY_DONE; + } + mutex_unlock(&macsec->lock); + return NOTIFY_OK; +} + +void mlx5r_macsec_event_register(struct mlx5_ib_dev *dev) +{ + if (!mlx5_is_macsec_roce_supported(dev->mdev)) { + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + return; + } + + dev->macsec.blocking_events_nb.notifier_call = macsec_event; + blocking_notifier_chain_register(&dev->mdev->macsec_nh, + &dev->macsec.blocking_events_nb); +} + +void mlx5r_macsec_event_unregister(struct mlx5_ib_dev *dev) +{ + if (!mlx5_is_macsec_roce_supported(dev->mdev)) { + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + return; + } + + blocking_notifier_chain_unregister(&dev->mdev->macsec_nh, + &dev->macsec.blocking_events_nb); +} + +int mlx5r_macsec_init_gids_and_devlist(struct mlx5_ib_dev *dev) +{ + int i, j, max_gids; + + if (!mlx5_is_macsec_roce_supported(dev->mdev)) { + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + return 0; + } + + max_gids = MLX5_CAP_ROCE(dev->mdev, roce_address_table_size); + for (i = 0; i < dev->num_ports; i++) { + dev->port[i].reserved_gids = kcalloc(max_gids, + sizeof(*dev->port[i].reserved_gids), + GFP_KERNEL); + if (!dev->port[i].reserved_gids) + goto err; + + for (j = 0; j < max_gids; j++) + dev->port[i].reserved_gids[j].macsec_index = -1; + } + + INIT_LIST_HEAD(&dev->macsec.macsec_devices_list); + mutex_init(&dev->macsec.lock); + + return 0; +err: + while (i >= 0) { + kfree(dev->port[i].reserved_gids); + i--; + } + return -ENOMEM; +} + +void mlx5r_macsec_dealloc_gids(struct mlx5_ib_dev *dev) +{ + int i; + + if (!mlx5_is_macsec_roce_supported(dev->mdev)) + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + + for (i = 0; i < dev->num_ports; i++) + kfree(dev->port[i].reserved_gids); + + mutex_destroy(&dev->macsec.lock); +} + +int mlx5r_add_gid_macsec_operations(const struct ib_gid_attr *attr) +{ + struct mlx5_ib_dev *dev = to_mdev(attr->device); + struct mlx5_macsec_device *macsec_device; + const struct ib_gid_attr *physical_gid; + struct mlx5_reserved_gids *mgids; + struct net_device *ndev; + int ret = 0; + union { + struct sockaddr_in sockaddr_in; + struct sockaddr_in6 sockaddr_in6; + } addr; + + if (attr->gid_type != IB_GID_TYPE_ROCE_UDP_ENCAP) + return 0; + + if (!mlx5_is_macsec_roce_supported(dev->mdev)) { + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + return 0; + } + + rcu_read_lock(); + ndev = rcu_dereference(attr->ndev); + if (!ndev) { + rcu_read_unlock(); + return -ENODEV; + } + + if (!netif_is_macsec(ndev) || !macsec_netdev_is_offloaded(ndev)) { + rcu_read_unlock(); + return 0; + } + dev_hold(ndev); + rcu_read_unlock(); + + mutex_lock(&dev->macsec.lock); + macsec_device = get_macsec_device(ndev, &dev->macsec.macsec_devices_list); + if (!macsec_device) { + ret = -ENOMEM; + goto dev_err; + } + + physical_gid = rdma_find_gid(attr->device, &attr->gid, + attr->gid_type, NULL); + if (!IS_ERR(physical_gid)) { + ret = set_roce_addr(to_mdev(physical_gid->device), + physical_gid->port_num, + physical_gid->index, NULL, + physical_gid); + if (ret) + goto gid_err; + + mgids = &dev->port[attr->port_num - 1].reserved_gids[physical_gid->index]; + mgids->macsec_index = attr->index; + mgids->physical_gid = physical_gid; + } + + /* Proceed with adding steering rules, regardless if there was gid ambiguity or not.*/ + rdma_gid2ip((struct sockaddr *)&addr, &attr->gid); + ret = mlx5_macsec_add_roce_rule(ndev, (struct sockaddr *)&addr, attr->index, + &macsec_device->tx_rules_list, + &macsec_device->rx_rules_list, dev->mdev->macsec_fs); + if (ret && !IS_ERR(physical_gid)) + goto rule_err; + + mlx5_macsec_save_roce_gid(macsec_device, (struct sockaddr *)&addr, attr->index); + + dev_put(ndev); + mutex_unlock(&dev->macsec.lock); + return ret; + +rule_err: + set_roce_addr(to_mdev(physical_gid->device), physical_gid->port_num, + physical_gid->index, &physical_gid->gid, physical_gid); + mgids->macsec_index = -1; +gid_err: + rdma_put_gid_attr(physical_gid); + cleanup_macsec_device(macsec_device); +dev_err: + dev_put(ndev); + mutex_unlock(&dev->macsec.lock); + return ret; +} + +void mlx5r_del_gid_macsec_operations(const struct ib_gid_attr *attr) +{ + struct mlx5_ib_dev *dev = to_mdev(attr->device); + struct mlx5_macsec_device *macsec_device; + struct mlx5_reserved_gids *mgids; + struct net_device *ndev; + int i, max_gids; + + if (attr->gid_type != IB_GID_TYPE_ROCE_UDP_ENCAP) + return; + + if (!mlx5_is_macsec_roce_supported(dev->mdev)) { + mlx5_ib_dbg(dev, "RoCE MACsec not supported due to capabilities\n"); + return; + } + + mgids = &dev->port[attr->port_num - 1].reserved_gids[attr->index]; + if (mgids->macsec_index != -1) { /* Checking if physical gid has ambiguous IP */ + rdma_put_gid_attr(mgids->physical_gid); + mgids->macsec_index = -1; + return; + } + + rcu_read_lock(); + ndev = rcu_dereference(attr->ndev); + if (!ndev) { + rcu_read_unlock(); + return; + } + + if (!netif_is_macsec(ndev) || !macsec_netdev_is_offloaded(ndev)) { + rcu_read_unlock(); + return; + } + dev_hold(ndev); + rcu_read_unlock(); + + mutex_lock(&dev->macsec.lock); + max_gids = MLX5_CAP_ROCE(dev->mdev, roce_address_table_size); + for (i = 0; i < max_gids; i++) { /* Checking if macsec gid has ambiguous IP */ + mgids = &dev->port[attr->port_num - 1].reserved_gids[i]; + if (mgids->macsec_index == attr->index) { + const struct ib_gid_attr *physical_gid = mgids->physical_gid; + + set_roce_addr(to_mdev(physical_gid->device), + physical_gid->port_num, + physical_gid->index, + &physical_gid->gid, physical_gid); + + rdma_put_gid_attr(physical_gid); + mgids->macsec_index = -1; + break; + } + } + macsec_device = get_macsec_device(ndev, &dev->macsec.macsec_devices_list); + mlx5_macsec_del_roce_rule(attr->index, dev->mdev->macsec_fs, + &macsec_device->tx_rules_list, &macsec_device->rx_rules_list); + mlx5_macsec_del_roce_gid(macsec_device, attr->index); + cleanup_macsec_device(macsec_device); + + dev_put(ndev); + mutex_unlock(&dev->macsec.lock); +} diff --git a/drivers/infiniband/hw/mlx5/macsec.h b/drivers/infiniband/hw/mlx5/macsec.h new file mode 100644 index 000000000000..9b77ba90f0f4 --- /dev/null +++ b/drivers/infiniband/hw/mlx5/macsec.h @@ -0,0 +1,29 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2022, NVIDIA CORPORATION & AFFILIATES. */ + +#ifndef __MLX5_MACSEC_H__ +#define __MLX5_MACSEC_H__ + +#include <net/macsec.h> +#include <rdma/ib_cache.h> +#include <rdma/ib_addr.h> +#include "mlx5_ib.h" + +#ifdef CONFIG_MLX5_MACSEC +struct mlx5_reserved_gids; + +int mlx5r_add_gid_macsec_operations(const struct ib_gid_attr *attr); +void mlx5r_del_gid_macsec_operations(const struct ib_gid_attr *attr); +int mlx5r_macsec_init_gids_and_devlist(struct mlx5_ib_dev *dev); +void mlx5r_macsec_dealloc_gids(struct mlx5_ib_dev *dev); +void mlx5r_macsec_event_register(struct mlx5_ib_dev *dev); +void mlx5r_macsec_event_unregister(struct mlx5_ib_dev *dev); +#else +static inline int mlx5r_add_gid_macsec_operations(const struct ib_gid_attr *attr) { return 0; } +static inline void mlx5r_del_gid_macsec_operations(const struct ib_gid_attr *attr) {} +static inline int mlx5r_macsec_init_gids_and_devlist(struct mlx5_ib_dev *dev) { return 0; } +static inline void mlx5r_macsec_dealloc_gids(struct mlx5_ib_dev *dev) {} +static inline void mlx5r_macsec_event_register(struct mlx5_ib_dev *dev) {} +static inline void mlx5r_macsec_event_unregister(struct mlx5_ib_dev *dev) {} +#endif +#endif /* __MLX5_MACSEC_H__ */ diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index f0b394ed7452..aed5cdea50e6 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -46,6 +46,7 @@ #include <rdma/uverbs_ioctl.h> #include <rdma/mlx5_user_ioctl_verbs.h> #include <rdma/mlx5_user_ioctl_cmds.h> +#include "macsec.h" #define UVERBS_MODULE_NAME mlx5_ib #include <rdma/uverbs_named_ioctl.h> @@ -564,9 +565,9 @@ out: return err; } -static int set_roce_addr(struct mlx5_ib_dev *dev, u32 port_num, - unsigned int index, const union ib_gid *gid, - const struct ib_gid_attr *attr) +int set_roce_addr(struct mlx5_ib_dev *dev, u32 port_num, + unsigned int index, const union ib_gid *gid, + const struct ib_gid_attr *attr) { enum ib_gid_type gid_type; u16 vlan_id = 0xffff; @@ -607,6 +608,12 @@ static int set_roce_addr(struct mlx5_ib_dev *dev, u32 port_num, static int mlx5_ib_add_gid(const struct ib_gid_attr *attr, __always_unused void **context) { + int ret; + + ret = mlx5r_add_gid_macsec_operations(attr); + if (ret) + return ret; + return set_roce_addr(to_mdev(attr->device), attr->port_num, attr->index, &attr->gid, attr); } @@ -614,8 +621,15 @@ static int mlx5_ib_add_gid(const struct ib_gid_attr *attr, static int mlx5_ib_del_gid(const struct ib_gid_attr *attr, __always_unused void **context) { - return set_roce_addr(to_mdev(attr->device), attr->port_num, - attr->index, NULL, attr); + int ret; + + ret = set_roce_addr(to_mdev(attr->device), attr->port_num, + attr->index, NULL, attr); + if (ret) + return ret; + + mlx5r_del_gid_macsec_operations(attr); + return 0; } __be16 mlx5_get_roce_udp_sport_min(const struct mlx5_ib_dev *dev, @@ -3644,13 +3658,13 @@ static void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev) mutex_destroy(&dev->cap_mask_mutex); WARN_ON(!xa_empty(&dev->sig_mrs)); WARN_ON(!bitmap_empty(dev->dm.memic_alloc_pages, MLX5_MAX_MEMIC_PAGES)); + mlx5r_macsec_dealloc_gids(dev); } static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev) { struct mlx5_core_dev *mdev = dev->mdev; - int err; - int i; + int err, i; dev->ib_dev.node_type = RDMA_NODE_IB_CA; dev->ib_dev.local_dma_lkey = 0 /* not supported for now */; @@ -3670,10 +3684,14 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev) if (err) return err; - err = mlx5_ib_init_multiport_master(dev); + err = mlx5r_macsec_init_gids_and_devlist(dev); if (err) return err; + err = mlx5_ib_init_multiport_master(dev); + if (err) + goto err; + err = set_has_smi_cap(dev); if (err) goto err_mp; @@ -3685,7 +3703,7 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev) if (mlx5_use_mad_ifc(dev)) get_ext_port_caps(dev); - dev->ib_dev.num_comp_vectors = mlx5_comp_vectors_count(mdev); + dev->ib_dev.num_comp_vectors = mlx5_comp_vectors_max(mdev); mutex_init(&dev->cap_mask_mutex); INIT_LIST_HEAD(&dev->qp_list); @@ -3697,7 +3715,8 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev) spin_lock_init(&dev->dm.lock); dev->dm.dev = mdev; return 0; - +err: + mlx5r_macsec_dealloc_gids(dev); err_mp: mlx5_ib_cleanup_multiport_master(dev); return err; @@ -4106,11 +4125,15 @@ static int mlx5_ib_stage_dev_notifier_init(struct mlx5_ib_dev *dev) { dev->mdev_events.notifier_call = mlx5_ib_event; mlx5_notifier_register(dev->mdev, &dev->mdev_events); + + mlx5r_macsec_event_register(dev); + return 0; } static void mlx5_ib_stage_dev_notifier_cleanup(struct mlx5_ib_dev *dev) { + mlx5r_macsec_event_unregister(dev); mlx5_notifier_unregister(dev->mdev, &dev->mdev_events); } diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index 9c33d960af3c..16713baf0d06 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -26,6 +26,7 @@ #include "srq.h" #include "qp.h" +#include "macsec.h" #define mlx5_ib_dbg(_dev, format, arg...) \ dev_dbg(&(_dev)->ib_dev.dev, "%s:%d:(pid %d): " format, __func__, \ @@ -870,6 +871,9 @@ struct mlx5_ib_port { struct mlx5_ib_dbg_cc_params *dbg_cc_params; struct mlx5_roce roce; struct mlx5_eswitch_rep *rep; +#ifdef CONFIG_MLX5_MACSEC + struct mlx5_reserved_gids *reserved_gids; +#endif }; struct mlx5_ib_dbg_param { @@ -1086,6 +1090,12 @@ struct mlx5_special_mkeys { __be32 terminate_scatter_list_mkey; }; +struct mlx5_macsec { + struct mutex lock; /* Protects mlx5_macsec internal contexts */ + struct list_head macsec_devices_list; + struct notifier_block blocking_events_nb; +}; + struct mlx5_ib_dev { struct ib_device ib_dev; struct mlx5_core_dev *mdev; @@ -1145,6 +1155,10 @@ struct mlx5_ib_dev { u16 pkey_table_len; u8 lag_ports; struct mlx5_special_mkeys mkeys; + +#ifdef CONFIG_MLX5_MACSEC + struct mlx5_macsec macsec; +#endif }; static inline struct mlx5_ib_cq *to_mibcq(struct mlx5_core_cq *mcq) @@ -1648,4 +1662,7 @@ static inline bool mlx5_umem_needs_ats(struct mlx5_ib_dev *dev, return access_flags & IB_ACCESS_RELAXED_ORDERING; } +int set_roce_addr(struct mlx5_ib_dev *dev, u32 port_num, + unsigned int index, const union ib_gid *gid, + const struct ib_gid_attr *attr); #endif /* MLX5_IB_H */ diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c index 69bba0ef4a5d..53f43649f7d0 100644 --- a/drivers/infiniband/hw/mthca/mthca_qp.c +++ b/drivers/infiniband/hw/mthca/mthca_qp.c @@ -1393,7 +1393,7 @@ int mthca_alloc_sqp(struct mthca_dev *dev, if (mthca_array_get(&dev->qp_table.qp, mqpn)) err = -EBUSY; else - mthca_array_set(&dev->qp_table.qp, mqpn, qp->sqp); + mthca_array_set(&dev->qp_table.qp, mqpn, qp); spin_unlock_irq(&dev->qp_table.lock); if (err) diff --git a/drivers/infiniband/hw/qib/qib_fs.c b/drivers/infiniband/hw/qib/qib_fs.c index a973905afd13..ed7d4b02f45a 100644 --- a/drivers/infiniband/hw/qib/qib_fs.c +++ b/drivers/infiniband/hw/qib/qib_fs.c @@ -64,9 +64,8 @@ static int qibfs_mknod(struct inode *dir, struct dentry *dentry, inode->i_uid = GLOBAL_ROOT_UID; inode->i_gid = GLOBAL_ROOT_GID; inode->i_blocks = 0; - inode->i_atime = current_time(inode); + inode->i_atime = inode_set_ctime_current(inode); inode->i_mtime = inode->i_atime; - inode->i_ctime = inode->i_atime; inode->i_private = data; if (S_ISDIR(mode)) { inode->i_op = &simple_dir_inode_operations; diff --git a/drivers/infiniband/sw/rxe/rxe_mw.c b/drivers/infiniband/sw/rxe/rxe_mw.c index d8a43d87de93..d9312b5c9d20 100644 --- a/drivers/infiniband/sw/rxe/rxe_mw.c +++ b/drivers/infiniband/sw/rxe/rxe_mw.c @@ -199,7 +199,8 @@ int rxe_bind_mw(struct rxe_qp *qp, struct rxe_send_wqe *wqe) if (access & ~RXE_ACCESS_SUPPORTED_MW) { rxe_err_mw(mw, "access %#x not supported", access); - return -EOPNOTSUPP; + ret = -EOPNOTSUPP; + goto err_drop_mr; } spin_lock_bh(&mw->lock); |