/* * Copyright (c) 2016 QLogic Corporation. * All rights reserved. * www.qlogic.com * * See LICENSE.qede_pmd for copyright and licensing details. */ #include "bcm_osal.h" #include "reg_addr.h" #include "ecore_gtt_reg_addr.h" #include "ecore.h" #include "ecore_chain.h" #include "ecore_status.h" #include "ecore_hw.h" #include "ecore_rt_defs.h" #include "ecore_init_ops.h" #include "ecore_int.h" #include "ecore_cxt.h" #include "ecore_spq.h" #include "ecore_init_fw_funcs.h" #include "ecore_sp_commands.h" #include "ecore_dev_api.h" #include "ecore_sriov.h" #include "ecore_vf.h" #include "ecore_mcp.h" #include "ecore_hw_defs.h" #include "mcp_public.h" #include "ecore_iro.h" #include "nvm_cfg.h" #include "ecore_dcbx.h" #include "ecore_l2.h" /* TODO - there's a bug in DCBx re-configuration flows in MF, as the QM * registers involved are not split and thus configuration is a race where * some of the PFs configuration might be lost. * Eventually, this needs to move into a MFW-covered HW-lock as arbitration * mechanism as this doesn't cover some cases [E.g., PDA or scenarios where * there's more than a single compiled ecore component in system]. */ static osal_spinlock_t qm_lock; static bool qm_lock_init; /******************** Doorbell Recovery *******************/ /* The doorbell recovery mechanism consists of a list of entries which represent * doorbelling entities (l2 queues, roce sq/rq/cqs, the slowpath spq, etc). Each * entity needs to register with the mechanism and provide the parameters * describing it's doorbell, including a location where last used doorbell data * can be found. The doorbell execute function will traverse the list and * doorbell all of the registered entries. */ struct ecore_db_recovery_entry { osal_list_entry_t list_entry; void OSAL_IOMEM *db_addr; void *db_data; enum ecore_db_rec_width db_width; enum ecore_db_rec_space db_space; u8 hwfn_idx; }; /* display a single doorbell recovery entry */ void ecore_db_recovery_dp_entry(struct ecore_hwfn *p_hwfn, struct ecore_db_recovery_entry *db_entry, const char *action) { DP_VERBOSE(p_hwfn, ECORE_MSG_SPQ, "(%s: db_entry %p, addr %p, data %p, width %s, %s space, hwfn %d)\n", action, db_entry, db_entry->db_addr, db_entry->db_data, db_entry->db_width == DB_REC_WIDTH_32B ? "32b" : "64b", db_entry->db_space == DB_REC_USER ? "user" : "kernel", db_entry->hwfn_idx); } /* doorbell address sanity (address within doorbell bar range) */ bool ecore_db_rec_sanity(struct ecore_dev *p_dev, void OSAL_IOMEM *db_addr, void *db_data) { /* make sure doorbell address is within the doorbell bar */ if (db_addr < p_dev->doorbells || (u8 *)db_addr > (u8 *)p_dev->doorbells + p_dev->db_size) { OSAL_WARN(true, "Illegal doorbell address: %p. Legal range for doorbell addresses is [%p..%p]\n", db_addr, p_dev->doorbells, (u8 *)p_dev->doorbells + p_dev->db_size); return false; } /* make sure doorbell data pointer is not null */ if (!db_data) { OSAL_WARN(true, "Illegal doorbell data pointer: %p", db_data); return false; } return true; } /* find hwfn according to the doorbell address */ struct ecore_hwfn *ecore_db_rec_find_hwfn(struct ecore_dev *p_dev, void OSAL_IOMEM *db_addr) { struct ecore_hwfn *p_hwfn; /* In CMT doorbell bar is split down the middle between engine 0 and * enigne 1 */ if (ECORE_IS_CMT(p_dev)) p_hwfn = db_addr < p_dev->hwfns[1].doorbells ? &p_dev->hwfns[0] : &p_dev->hwfns[1]; else p_hwfn = ECORE_LEADING_HWFN(p_dev); return p_hwfn; } /* add a new entry to the doorbell recovery mechanism */ enum _ecore_status_t ecore_db_recovery_add(struct ecore_dev *p_dev, void OSAL_IOMEM *db_addr, void *db_data, enum ecore_db_rec_width db_width, enum ecore_db_rec_space db_space) { struct ecore_db_recovery_entry *db_entry; struct ecore_hwfn *p_hwfn; /* shortcircuit VFs, for now */ if (IS_VF(p_dev)) { DP_VERBOSE(p_dev, ECORE_MSG_IOV, "db recovery - skipping VF doorbell\n"); return ECORE_SUCCESS; } /* sanitize doorbell address */ if (!ecore_db_rec_sanity(p_dev, db_addr, db_data)) return ECORE_INVAL; /* obtain hwfn from doorbell address */ p_hwfn = ecore_db_rec_find_hwfn(p_dev, db_addr); /* create entry */ db_entry = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(*db_entry)); if (!db_entry) { DP_NOTICE(p_dev, false, "Failed to allocate a db recovery entry\n"); return ECORE_NOMEM; } /* populate entry */ db_entry->db_addr = db_addr; db_entry->db_data = db_data; db_entry->db_width = db_width; db_entry->db_space = db_space; db_entry->hwfn_idx = p_hwfn->my_id; /* display */ ecore_db_recovery_dp_entry(p_hwfn, db_entry, "Adding"); /* protect the list */ OSAL_SPIN_LOCK(&p_hwfn->db_recovery_info.lock); OSAL_LIST_PUSH_TAIL(&db_entry->list_entry, &p_hwfn->db_recovery_info.list); OSAL_SPIN_UNLOCK(&p_hwfn->db_recovery_info.lock); return ECORE_SUCCESS; } /* remove an entry from the doorbell recovery mechanism */ enum _ecore_status_t ecore_db_recovery_del(struct ecore_dev *p_dev, void OSAL_IOMEM *db_addr, void *db_data) { struct ecore_db_recovery_entry *db_entry = OSAL_NULL; enum _ecore_status_t rc = ECORE_INVAL; struct ecore_hwfn *p_hwfn; /* shortcircuit VFs, for now */ if (IS_VF(p_dev)) { DP_VERBOSE(p_dev, ECORE_MSG_IOV, "db recovery - skipping VF doorbell\n"); return ECORE_SUCCESS; } /* sanitize doorbell address */ if (!ecore_db_rec_sanity(p_dev, db_addr, db_data)) return ECORE_INVAL; /* obtain hwfn from doorbell address */ p_hwfn = ecore_db_rec_find_hwfn(p_dev, db_addr); /* protect the list */ OSAL_SPIN_LOCK(&p_hwfn->db_recovery_info.lock); OSAL_LIST_FOR_EACH_ENTRY(db_entry, &p_hwfn->db_recovery_info.list, list_entry, struct ecore_db_recovery_entry) { /* search according to db_data addr since db_addr is not unique * (roce) */ if (db_entry->db_data == db_data) { ecore_db_recovery_dp_entry(p_hwfn, db_entry, "Deleting"); OSAL_LIST_REMOVE_ENTRY(&db_entry->list_entry, &p_hwfn->db_recovery_info.list); rc = ECORE_SUCCESS; break; } } OSAL_SPIN_UNLOCK(&p_hwfn->db_recovery_info.lock); if (rc == ECORE_INVAL) /*OSAL_WARN(true,*/ DP_NOTICE(p_hwfn, false, "Failed to find element in list. Key (db_data addr) was %p. db_addr was %p\n", db_data, db_addr); else OSAL_FREE(p_dev, db_entry); return rc; } /* initialize the doorbell recovery mechanism */ enum _ecore_status_t ecore_db_recovery_setup(struct ecore_hwfn *p_hwfn) { DP_VERBOSE(p_hwfn, ECORE_MSG_SPQ, "Setting up db recovery\n"); /* make sure db_size was set in p_dev */ if (!p_hwfn->p_dev->db_size) { DP_ERR(p_hwfn->p_dev, "db_size not set\n"); return ECORE_INVAL; } OSAL_LIST_INIT(&p_hwfn->db_recovery_info.list); #ifdef CONFIG_ECORE_LOCK_ALLOC OSAL_SPIN_LOCK_ALLOC(p_hwfn, &p_hwfn->db_recovery_info.lock); #endif OSAL_SPIN_LOCK_INIT(&p_hwfn->db_recovery_info.lock); p_hwfn->db_recovery_info.db_recovery_counter = 0; return ECORE_SUCCESS; } /* destroy the doorbell recovery mechanism */ void ecore_db_recovery_teardown(struct ecore_hwfn *p_hwfn) { struct ecore_db_recovery_entry *db_entry = OSAL_NULL; DP_VERBOSE(p_hwfn, ECORE_MSG_SPQ, "Tearing down db recovery\n"); if (!OSAL_LIST_IS_EMPTY(&p_hwfn->db_recovery_info.list)) { DP_VERBOSE(p_hwfn, false, "Doorbell Recovery teardown found the doorbell recovery list was not empty (Expected in disorderly driver unload (e.g. recovery) otherwise this probably means some flow forgot to db_recovery_del). Prepare to purge doorbell recovery list...\n"); while (!OSAL_LIST_IS_EMPTY(&p_hwfn->db_recovery_info.list)) { db_entry = OSAL_LIST_FIRST_ENTRY( &p_hwfn->db_recovery_info.list, struct ecore_db_recovery_entry, list_entry); ecore_db_recovery_dp_entry(p_hwfn, db_entry, "Purging"); OSAL_LIST_REMOVE_ENTRY(&db_entry->list_entry, &p_hwfn->db_recovery_info.list); OSAL_FREE(p_hwfn->p_dev, db_entry); } } #ifdef CONFIG_ECORE_LOCK_ALLOC OSAL_SPIN_LOCK_DEALLOC(&p_hwfn->db_recovery_info.lock); #endif p_hwfn->db_recovery_info.db_recovery_counter = 0; } /* print the content of the doorbell recovery mechanism */ void ecore_db_recovery_dp(struct ecore_hwfn *p_hwfn) { struct ecore_db_recovery_entry *db_entry = OSAL_NULL; DP_NOTICE(p_hwfn, false, "Dispalying doorbell recovery database. Counter was %d\n", p_hwfn->db_recovery_info.db_recovery_counter); /* protect the list */ OSAL_SPIN_LOCK(&p_hwfn->db_recovery_info.lock); OSAL_LIST_FOR_EACH_ENTRY(db_entry, &p_hwfn->db_recovery_info.list, list_entry, struct ecore_db_recovery_entry) { ecore_db_recovery_dp_entry(p_hwfn, db_entry, "Printing"); } OSAL_SPIN_UNLOCK(&p_hwfn->db_recovery_info.lock); } /* ring the doorbell of a single doorbell recovery entry */ void ecore_db_recovery_ring(struct ecore_hwfn *p_hwfn, struct ecore_db_recovery_entry *db_entry, enum ecore_db_rec_exec db_exec) { /* Print according to width */ if (db_entry->db_width == DB_REC_WIDTH_32B) DP_VERBOSE(p_hwfn, ECORE_MSG_SPQ, "%s doorbell address %p data %x\n", db_exec == DB_REC_DRY_RUN ? "would have rung" : "ringing", db_entry->db_addr, *(u32 *)db_entry->db_data); else DP_VERBOSE(p_hwfn, ECORE_MSG_SPQ, "%s doorbell address %p data %lx\n", db_exec == DB_REC_DRY_RUN ? "would have rung" : "ringing", db_entry->db_addr, *(unsigned long *)(db_entry->db_data)); /* Sanity */ if (!ecore_db_rec_sanity(p_hwfn->p_dev, db_entry->db_addr, db_entry->db_data)) return; /* Flush the write combined buffer. Since there are multiple doorbelling * entities using the same address, if we don't flush, a transaction * could be lost. */ OSAL_WMB(p_hwfn->p_dev); /* Ring the doorbell */ if (db_exec == DB_REC_REAL_DEAL || db_exec == DB_REC_ONCE) { if (db_entry->db_width == DB_REC_WIDTH_32B) DIRECT_REG_WR(p_hwfn, db_entry->db_addr, *(u32 *)(db_entry->db_data)); else DIRECT_REG_WR64(p_hwfn, db_entry->db_addr, *(u64 *)(db_entry->db_data)); } /* Flush the write combined buffer. Next doorbell may come from a * different entity to the same address... */ OSAL_WMB(p_hwfn->p_dev); } /* traverse the doorbell recovery entry list and ring all the doorbells */ void ecore_db_recovery_execute(struct ecore_hwfn *p_hwfn, enum ecore_db_rec_exec db_exec) { struct ecore_db_recovery_entry *db_entry = OSAL_NULL; if (db_exec != DB_REC_ONCE) { DP_NOTICE(p_hwfn, false, "Executing doorbell recovery. Counter was %d\n", p_hwfn->db_recovery_info.db_recovery_counter); /* track amount of times recovery was executed */ p_hwfn->db_recovery_info.db_recovery_counter++; } /* protect the list */ OSAL_SPIN_LOCK(&p_hwfn->db_recovery_info.lock); OSAL_LIST_FOR_EACH_ENTRY(db_entry, &p_hwfn->db_recovery_info.list, list_entry, struct ecore_db_recovery_entry) { ecore_db_recovery_ring(p_hwfn, db_entry, db_exec); if (db_exec == DB_REC_ONCE) break; } OSAL_SPIN_UNLOCK(&p_hwfn->db_recovery_info.lock); } /******************** Doorbell Recovery end ****************/ /* Configurable */ #define ECORE_MIN_DPIS (4) /* The minimal num of DPIs required to * load the driver. The number was * arbitrarily set. */ /* Derived */ #define ECORE_MIN_PWM_REGION (ECORE_WID_SIZE * ECORE_MIN_DPIS) static u32 ecore_hw_bar_size(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, enum BAR_ID bar_id) { u32 bar_reg = (bar_id == BAR_ID_0 ? PGLUE_B_REG_PF_BAR0_SIZE : PGLUE_B_REG_PF_BAR1_SIZE); u32 val; if (IS_VF(p_hwfn->p_dev)) return ecore_vf_hw_bar_size(p_hwfn, bar_id); val = ecore_rd(p_hwfn, p_ptt, bar_reg); if (val) return 1 << (val + 15); /* The above registers were updated in the past only in CMT mode. Since * they were found to be useful MFW started updating them from 8.7.7.0. * In older MFW versions they are set to 0 which means disabled. */ if (ECORE_IS_CMT(p_hwfn->p_dev)) { DP_INFO(p_hwfn, "BAR size not configured. Assuming BAR size of 256kB for GRC and 512kB for DB\n"); val = BAR_ID_0 ? 256 * 1024 : 512 * 1024; } else { DP_INFO(p_hwfn, "BAR size not configured. Assuming BAR size of 512kB for GRC and 512kB for DB\n"); val = 512 * 1024; } return val; } void ecore_init_dp(struct ecore_dev *p_dev, u32 dp_module, u8 dp_level, void *dp_ctx) { u32 i; p_dev->dp_level = dp_level; p_dev->dp_module = dp_module; p_dev->dp_ctx = dp_ctx; for (i = 0; i < MAX_HWFNS_PER_DEVICE; i++) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; p_hwfn->dp_level = dp_level; p_hwfn->dp_module = dp_module; p_hwfn->dp_ctx = dp_ctx; } } void ecore_init_struct(struct ecore_dev *p_dev) { u8 i; for (i = 0; i < MAX_HWFNS_PER_DEVICE; i++) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; p_hwfn->p_dev = p_dev; p_hwfn->my_id = i; p_hwfn->b_active = false; #ifdef CONFIG_ECORE_LOCK_ALLOC OSAL_MUTEX_ALLOC(p_hwfn, &p_hwfn->dmae_info.mutex); #endif OSAL_MUTEX_INIT(&p_hwfn->dmae_info.mutex); } /* hwfn 0 is always active */ p_dev->hwfns[0].b_active = true; /* set the default cache alignment to 128 (may be overridden later) */ p_dev->cache_shift = 7; } static void ecore_qm_info_free(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; OSAL_FREE(p_hwfn->p_dev, qm_info->qm_pq_params); OSAL_FREE(p_hwfn->p_dev, qm_info->qm_vport_params); OSAL_FREE(p_hwfn->p_dev, qm_info->qm_port_params); OSAL_FREE(p_hwfn->p_dev, qm_info->wfq_data); } void ecore_resc_free(struct ecore_dev *p_dev) { int i; if (IS_VF(p_dev)) { for_each_hwfn(p_dev, i) ecore_l2_free(&p_dev->hwfns[i]); return; } OSAL_FREE(p_dev, p_dev->fw_data); OSAL_FREE(p_dev, p_dev->reset_stats); for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; ecore_cxt_mngr_free(p_hwfn); ecore_qm_info_free(p_hwfn); ecore_spq_free(p_hwfn); ecore_eq_free(p_hwfn); ecore_consq_free(p_hwfn); ecore_int_free(p_hwfn); ecore_iov_free(p_hwfn); ecore_l2_free(p_hwfn); ecore_dmae_info_free(p_hwfn); ecore_dcbx_info_free(p_hwfn); /* @@@TBD Flush work-queue ? */ /* destroy doorbell recovery mechanism */ ecore_db_recovery_teardown(p_hwfn); } } /******************** QM initialization *******************/ /* bitmaps for indicating active traffic classes. * Special case for Arrowhead 4 port */ /* 0..3 actualy used, 4 serves OOO, 7 serves high priority stuff (e.g. DCQCN) */ #define ACTIVE_TCS_BMAP 0x9f /* 0..3 actually used, OOO and high priority stuff all use 3 */ #define ACTIVE_TCS_BMAP_4PORT_K2 0xf /* determines the physical queue flags for a given PF. */ static u32 ecore_get_pq_flags(struct ecore_hwfn *p_hwfn) { u32 flags; /* common flags */ flags = PQ_FLAGS_LB; /* feature flags */ if (IS_ECORE_SRIOV(p_hwfn->p_dev)) flags |= PQ_FLAGS_VFS; /* protocol flags */ switch (p_hwfn->hw_info.personality) { case ECORE_PCI_ETH: flags |= PQ_FLAGS_MCOS; break; case ECORE_PCI_FCOE: flags |= PQ_FLAGS_OFLD; break; case ECORE_PCI_ISCSI: flags |= PQ_FLAGS_ACK | PQ_FLAGS_OOO | PQ_FLAGS_OFLD; break; case ECORE_PCI_ETH_ROCE: flags |= PQ_FLAGS_MCOS | PQ_FLAGS_OFLD; break; case ECORE_PCI_ETH_IWARP: flags |= PQ_FLAGS_MCOS | PQ_FLAGS_ACK | PQ_FLAGS_OOO | PQ_FLAGS_OFLD; break; default: DP_ERR(p_hwfn, "unknown personality %d\n", p_hwfn->hw_info.personality); return 0; } return flags; } /* Getters for resource amounts necessary for qm initialization */ u8 ecore_init_qm_get_num_tcs(struct ecore_hwfn *p_hwfn) { return p_hwfn->hw_info.num_hw_tc; } u16 ecore_init_qm_get_num_vfs(struct ecore_hwfn *p_hwfn) { return IS_ECORE_SRIOV(p_hwfn->p_dev) ? p_hwfn->p_dev->p_iov_info->total_vfs : 0; } #define NUM_DEFAULT_RLS 1 u16 ecore_init_qm_get_num_pf_rls(struct ecore_hwfn *p_hwfn) { u16 num_pf_rls, num_vfs = ecore_init_qm_get_num_vfs(p_hwfn); /* @DPDK */ /* num RLs can't exceed resource amount of rls or vports or the * dcqcn qps */ num_pf_rls = (u16)OSAL_MIN_T(u32, RESC_NUM(p_hwfn, ECORE_RL), (u16)RESC_NUM(p_hwfn, ECORE_VPORT)); /* make sure after we reserve the default and VF rls we'll have * something left */ if (num_pf_rls < num_vfs + NUM_DEFAULT_RLS) { DP_NOTICE(p_hwfn, false, "no rate limiters left for PF rate limiting" " [num_pf_rls %d num_vfs %d]\n", num_pf_rls, num_vfs); return 0; } /* subtract rls necessary for VFs and one default one for the PF */ num_pf_rls -= num_vfs + NUM_DEFAULT_RLS; return num_pf_rls; } u16 ecore_init_qm_get_num_vports(struct ecore_hwfn *p_hwfn) { u32 pq_flags = ecore_get_pq_flags(p_hwfn); /* all pqs share the same vport (hence the 1 below), except for vfs * and pf_rl pqs */ return (!!(PQ_FLAGS_RLS & pq_flags)) * ecore_init_qm_get_num_pf_rls(p_hwfn) + (!!(PQ_FLAGS_VFS & pq_flags)) * ecore_init_qm_get_num_vfs(p_hwfn) + 1; } /* calc amount of PQs according to the requested flags */ u16 ecore_init_qm_get_num_pqs(struct ecore_hwfn *p_hwfn) { u32 pq_flags = ecore_get_pq_flags(p_hwfn); return (!!(PQ_FLAGS_RLS & pq_flags)) * ecore_init_qm_get_num_pf_rls(p_hwfn) + (!!(PQ_FLAGS_MCOS & pq_flags)) * ecore_init_qm_get_num_tcs(p_hwfn) + (!!(PQ_FLAGS_LB & pq_flags)) + (!!(PQ_FLAGS_OOO & pq_flags)) + (!!(PQ_FLAGS_ACK & pq_flags)) + (!!(PQ_FLAGS_OFLD & pq_flags)) + (!!(PQ_FLAGS_VFS & pq_flags)) * ecore_init_qm_get_num_vfs(p_hwfn); } /* initialize the top level QM params */ static void ecore_init_qm_params(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; bool four_port; /* pq and vport bases for this PF */ qm_info->start_pq = (u16)RESC_START(p_hwfn, ECORE_PQ); qm_info->start_vport = (u8)RESC_START(p_hwfn, ECORE_VPORT); /* rate limiting and weighted fair queueing are always enabled */ qm_info->vport_rl_en = 1; qm_info->vport_wfq_en = 1; /* TC config is different for AH 4 port */ four_port = p_hwfn->p_dev->num_ports_in_engine == MAX_NUM_PORTS_K2; /* in AH 4 port we have fewer TCs per port */ qm_info->max_phys_tcs_per_port = four_port ? NUM_PHYS_TCS_4PORT_K2 : NUM_OF_PHYS_TCS; /* unless MFW indicated otherwise, ooo_tc should be 3 for AH 4 port and * 4 otherwise */ if (!qm_info->ooo_tc) qm_info->ooo_tc = four_port ? DCBX_TCP_OOO_K2_4PORT_TC : DCBX_TCP_OOO_TC; } /* initialize qm vport params */ static void ecore_init_qm_vport_params(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; u8 i; /* all vports participate in weighted fair queueing */ for (i = 0; i < ecore_init_qm_get_num_vports(p_hwfn); i++) qm_info->qm_vport_params[i].vport_wfq = 1; } /* initialize qm port params */ static void ecore_init_qm_port_params(struct ecore_hwfn *p_hwfn) { /* Initialize qm port parameters */ u8 i, active_phys_tcs, num_ports = p_hwfn->p_dev->num_ports_in_engine; /* indicate how ooo and high pri traffic is dealt with */ active_phys_tcs = num_ports == MAX_NUM_PORTS_K2 ? ACTIVE_TCS_BMAP_4PORT_K2 : ACTIVE_TCS_BMAP; for (i = 0; i < num_ports; i++) { struct init_qm_port_params *p_qm_port = &p_hwfn->qm_info.qm_port_params[i]; p_qm_port->active = 1; p_qm_port->active_phys_tcs = active_phys_tcs; p_qm_port->num_pbf_cmd_lines = PBF_MAX_CMD_LINES_E4 / num_ports; p_qm_port->num_btb_blocks = BTB_MAX_BLOCKS / num_ports; } } /* Reset the params which must be reset for qm init. QM init may be called as * a result of flows other than driver load (e.g. dcbx renegotiation). Other * params may be affected by the init but would simply recalculate to the same * values. The allocations made for QM init, ports, vports, pqs and vfqs are not * affected as these amounts stay the same. */ static void ecore_init_qm_reset_params(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; qm_info->num_pqs = 0; qm_info->num_vports = 0; qm_info->num_pf_rls = 0; qm_info->num_vf_pqs = 0; qm_info->first_vf_pq = 0; qm_info->first_mcos_pq = 0; qm_info->first_rl_pq = 0; } static void ecore_init_qm_advance_vport(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; qm_info->num_vports++; if (qm_info->num_vports > ecore_init_qm_get_num_vports(p_hwfn)) DP_ERR(p_hwfn, "vport overflow! qm_info->num_vports %d," " qm_init_get_num_vports() %d\n", qm_info->num_vports, ecore_init_qm_get_num_vports(p_hwfn)); } /* initialize a single pq and manage qm_info resources accounting. * The pq_init_flags param determines whether the PQ is rate limited * (for VF or PF) * and whether a new vport is allocated to the pq or not (i.e. vport will be * shared) */ /* flags for pq init */ #define PQ_INIT_SHARE_VPORT (1 << 0) #define PQ_INIT_PF_RL (1 << 1) #define PQ_INIT_VF_RL (1 << 2) /* defines for pq init */ #define PQ_INIT_DEFAULT_WRR_GROUP 1 #define PQ_INIT_DEFAULT_TC 0 #define PQ_INIT_OFLD_TC (p_hwfn->hw_info.offload_tc) static void ecore_init_qm_pq(struct ecore_hwfn *p_hwfn, struct ecore_qm_info *qm_info, u8 tc, u32 pq_init_flags) { u16 pq_idx = qm_info->num_pqs, max_pq = ecore_init_qm_get_num_pqs(p_hwfn); if (pq_idx > max_pq) DP_ERR(p_hwfn, "pq overflow! pq %d, max pq %d\n", pq_idx, max_pq); /* init pq params */ qm_info->qm_pq_params[pq_idx].vport_id = qm_info->start_vport + qm_info->num_vports; qm_info->qm_pq_params[pq_idx].tc_id = tc; qm_info->qm_pq_params[pq_idx].wrr_group = PQ_INIT_DEFAULT_WRR_GROUP; qm_info->qm_pq_params[pq_idx].rl_valid = (pq_init_flags & PQ_INIT_PF_RL || pq_init_flags & PQ_INIT_VF_RL); /* qm params accounting */ qm_info->num_pqs++; if (!(pq_init_flags & PQ_INIT_SHARE_VPORT)) qm_info->num_vports++; if (pq_init_flags & PQ_INIT_PF_RL) qm_info->num_pf_rls++; if (qm_info->num_vports > ecore_init_qm_get_num_vports(p_hwfn)) DP_ERR(p_hwfn, "vport overflow! qm_info->num_vports %d," " qm_init_get_num_vports() %d\n", qm_info->num_vports, ecore_init_qm_get_num_vports(p_hwfn)); if (qm_info->num_pf_rls > ecore_init_qm_get_num_pf_rls(p_hwfn)) DP_ERR(p_hwfn, "rl overflow! qm_info->num_pf_rls %d," " qm_init_get_num_pf_rls() %d\n", qm_info->num_pf_rls, ecore_init_qm_get_num_pf_rls(p_hwfn)); } /* get pq index according to PQ_FLAGS */ static u16 *ecore_init_qm_get_idx_from_flags(struct ecore_hwfn *p_hwfn, u32 pq_flags) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; /* Can't have multiple flags set here */ if (OSAL_BITMAP_WEIGHT((unsigned long *)&pq_flags, sizeof(pq_flags)) > 1) goto err; switch (pq_flags) { case PQ_FLAGS_RLS: return &qm_info->first_rl_pq; case PQ_FLAGS_MCOS: return &qm_info->first_mcos_pq; case PQ_FLAGS_LB: return &qm_info->pure_lb_pq; case PQ_FLAGS_OOO: return &qm_info->ooo_pq; case PQ_FLAGS_ACK: return &qm_info->pure_ack_pq; case PQ_FLAGS_OFLD: return &qm_info->offload_pq; case PQ_FLAGS_VFS: return &qm_info->first_vf_pq; default: goto err; } err: DP_ERR(p_hwfn, "BAD pq flags %d\n", pq_flags); return OSAL_NULL; } /* save pq index in qm info */ static void ecore_init_qm_set_idx(struct ecore_hwfn *p_hwfn, u32 pq_flags, u16 pq_val) { u16 *base_pq_idx = ecore_init_qm_get_idx_from_flags(p_hwfn, pq_flags); *base_pq_idx = p_hwfn->qm_info.start_pq + pq_val; } /* get tx pq index, with the PQ TX base already set (ready for context init) */ u16 ecore_get_cm_pq_idx(struct ecore_hwfn *p_hwfn, u32 pq_flags) { u16 *base_pq_idx = ecore_init_qm_get_idx_from_flags(p_hwfn, pq_flags); return *base_pq_idx + CM_TX_PQ_BASE; } u16 ecore_get_cm_pq_idx_mcos(struct ecore_hwfn *p_hwfn, u8 tc) { u8 max_tc = ecore_init_qm_get_num_tcs(p_hwfn); if (tc > max_tc) DP_ERR(p_hwfn, "tc %d must be smaller than %d\n", tc, max_tc); return ecore_get_cm_pq_idx(p_hwfn, PQ_FLAGS_MCOS) + tc; } u16 ecore_get_cm_pq_idx_vf(struct ecore_hwfn *p_hwfn, u16 vf) { u16 max_vf = ecore_init_qm_get_num_vfs(p_hwfn); if (vf > max_vf) DP_ERR(p_hwfn, "vf %d must be smaller than %d\n", vf, max_vf); return ecore_get_cm_pq_idx(p_hwfn, PQ_FLAGS_VFS) + vf; } u16 ecore_get_cm_pq_idx_rl(struct ecore_hwfn *p_hwfn, u8 rl) { u16 max_rl = ecore_init_qm_get_num_pf_rls(p_hwfn); if (rl > max_rl) DP_ERR(p_hwfn, "rl %d must be smaller than %d\n", rl, max_rl); return ecore_get_cm_pq_idx(p_hwfn, PQ_FLAGS_RLS) + rl; } /* Functions for creating specific types of pqs */ static void ecore_init_qm_lb_pq(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_LB)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_LB, qm_info->num_pqs); ecore_init_qm_pq(p_hwfn, qm_info, PURE_LB_TC, PQ_INIT_SHARE_VPORT); } static void ecore_init_qm_ooo_pq(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_OOO)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_OOO, qm_info->num_pqs); ecore_init_qm_pq(p_hwfn, qm_info, qm_info->ooo_tc, PQ_INIT_SHARE_VPORT); } static void ecore_init_qm_pure_ack_pq(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_ACK)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_ACK, qm_info->num_pqs); ecore_init_qm_pq(p_hwfn, qm_info, PQ_INIT_OFLD_TC, PQ_INIT_SHARE_VPORT); } static void ecore_init_qm_offload_pq(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_OFLD)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_OFLD, qm_info->num_pqs); ecore_init_qm_pq(p_hwfn, qm_info, PQ_INIT_OFLD_TC, PQ_INIT_SHARE_VPORT); } static void ecore_init_qm_mcos_pqs(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; u8 tc_idx; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_MCOS)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_MCOS, qm_info->num_pqs); for (tc_idx = 0; tc_idx < ecore_init_qm_get_num_tcs(p_hwfn); tc_idx++) ecore_init_qm_pq(p_hwfn, qm_info, tc_idx, PQ_INIT_SHARE_VPORT); } static void ecore_init_qm_vf_pqs(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; u16 vf_idx, num_vfs = ecore_init_qm_get_num_vfs(p_hwfn); if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_VFS)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_VFS, qm_info->num_pqs); qm_info->num_vf_pqs = num_vfs; for (vf_idx = 0; vf_idx < num_vfs; vf_idx++) ecore_init_qm_pq(p_hwfn, qm_info, PQ_INIT_DEFAULT_TC, PQ_INIT_VF_RL); } static void ecore_init_qm_rl_pqs(struct ecore_hwfn *p_hwfn) { u16 pf_rls_idx, num_pf_rls = ecore_init_qm_get_num_pf_rls(p_hwfn); struct ecore_qm_info *qm_info = &p_hwfn->qm_info; if (!(ecore_get_pq_flags(p_hwfn) & PQ_FLAGS_RLS)) return; ecore_init_qm_set_idx(p_hwfn, PQ_FLAGS_RLS, qm_info->num_pqs); for (pf_rls_idx = 0; pf_rls_idx < num_pf_rls; pf_rls_idx++) ecore_init_qm_pq(p_hwfn, qm_info, PQ_INIT_OFLD_TC, PQ_INIT_PF_RL); } static void ecore_init_qm_pq_params(struct ecore_hwfn *p_hwfn) { /* rate limited pqs, must come first (FW assumption) */ ecore_init_qm_rl_pqs(p_hwfn); /* pqs for multi cos */ ecore_init_qm_mcos_pqs(p_hwfn); /* pure loopback pq */ ecore_init_qm_lb_pq(p_hwfn); /* out of order pq */ ecore_init_qm_ooo_pq(p_hwfn); /* pure ack pq */ ecore_init_qm_pure_ack_pq(p_hwfn); /* pq for offloaded protocol */ ecore_init_qm_offload_pq(p_hwfn); /* done sharing vports */ ecore_init_qm_advance_vport(p_hwfn); /* pqs for vfs */ ecore_init_qm_vf_pqs(p_hwfn); } /* compare values of getters against resources amounts */ static enum _ecore_status_t ecore_init_qm_sanity(struct ecore_hwfn *p_hwfn) { if (ecore_init_qm_get_num_vports(p_hwfn) > RESC_NUM(p_hwfn, ECORE_VPORT)) { DP_ERR(p_hwfn, "requested amount of vports exceeds resource\n"); return ECORE_INVAL; } if (ecore_init_qm_get_num_pqs(p_hwfn) > RESC_NUM(p_hwfn, ECORE_PQ)) { DP_ERR(p_hwfn, "requested amount of pqs exceeds resource\n"); return ECORE_INVAL; } return ECORE_SUCCESS; } /* * Function for verbose printing of the qm initialization results */ static void ecore_dp_init_qm_params(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; struct init_qm_vport_params *vport; struct init_qm_port_params *port; struct init_qm_pq_params *pq; int i, tc; /* top level params */ DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "qm init top level params: start_pq %d, start_vport %d," " pure_lb_pq %d, offload_pq %d, pure_ack_pq %d\n", qm_info->start_pq, qm_info->start_vport, qm_info->pure_lb_pq, qm_info->offload_pq, qm_info->pure_ack_pq); DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "ooo_pq %d, first_vf_pq %d, num_pqs %d, num_vf_pqs %d," " num_vports %d, max_phys_tcs_per_port %d\n", qm_info->ooo_pq, qm_info->first_vf_pq, qm_info->num_pqs, qm_info->num_vf_pqs, qm_info->num_vports, qm_info->max_phys_tcs_per_port); DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "pf_rl_en %d, pf_wfq_en %d, vport_rl_en %d, vport_wfq_en %d," " pf_wfq %d, pf_rl %d, num_pf_rls %d, pq_flags %x\n", qm_info->pf_rl_en, qm_info->pf_wfq_en, qm_info->vport_rl_en, qm_info->vport_wfq_en, qm_info->pf_wfq, qm_info->pf_rl, qm_info->num_pf_rls, ecore_get_pq_flags(p_hwfn)); /* port table */ for (i = 0; i < p_hwfn->p_dev->num_ports_in_engine; i++) { port = &qm_info->qm_port_params[i]; DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "port idx %d, active %d, active_phys_tcs %d," " num_pbf_cmd_lines %d, num_btb_blocks %d," " reserved %d\n", i, port->active, port->active_phys_tcs, port->num_pbf_cmd_lines, port->num_btb_blocks, port->reserved); } /* vport table */ for (i = 0; i < qm_info->num_vports; i++) { vport = &qm_info->qm_vport_params[i]; DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "vport idx %d, vport_rl %d, wfq %d," " first_tx_pq_id [ ", qm_info->start_vport + i, vport->vport_rl, vport->vport_wfq); for (tc = 0; tc < NUM_OF_TCS; tc++) DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "%d ", vport->first_tx_pq_id[tc]); DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "]\n"); } /* pq table */ for (i = 0; i < qm_info->num_pqs; i++) { pq = &qm_info->qm_pq_params[i]; DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "pq idx %d, vport_id %d, tc %d, wrr_grp %d," " rl_valid %d\n", qm_info->start_pq + i, pq->vport_id, pq->tc_id, pq->wrr_group, pq->rl_valid); } } static void ecore_init_qm_info(struct ecore_hwfn *p_hwfn) { /* reset params required for init run */ ecore_init_qm_reset_params(p_hwfn); /* init QM top level params */ ecore_init_qm_params(p_hwfn); /* init QM port params */ ecore_init_qm_port_params(p_hwfn); /* init QM vport params */ ecore_init_qm_vport_params(p_hwfn); /* init QM physical queue params */ ecore_init_qm_pq_params(p_hwfn); /* display all that init */ ecore_dp_init_qm_params(p_hwfn); } /* This function reconfigures the QM pf on the fly. * For this purpose we: * 1. reconfigure the QM database * 2. set new values to runtime array * 3. send an sdm_qm_cmd through the rbc interface to stop the QM * 4. activate init tool in QM_PF stage * 5. send an sdm_qm_cmd through rbc interface to release the QM */ enum _ecore_status_t ecore_qm_reconf(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; bool b_rc; enum _ecore_status_t rc; /* initialize ecore's qm data structure */ ecore_init_qm_info(p_hwfn); /* stop PF's qm queues */ OSAL_SPIN_LOCK(&qm_lock); b_rc = ecore_send_qm_stop_cmd(p_hwfn, p_ptt, false, true, qm_info->start_pq, qm_info->num_pqs); OSAL_SPIN_UNLOCK(&qm_lock); if (!b_rc) return ECORE_INVAL; /* clear the QM_PF runtime phase leftovers from previous init */ ecore_init_clear_rt_data(p_hwfn); /* prepare QM portion of runtime array */ ecore_qm_init_pf(p_hwfn, p_ptt); /* activate init tool on runtime array */ rc = ecore_init_run(p_hwfn, p_ptt, PHASE_QM_PF, p_hwfn->rel_pf_id, p_hwfn->hw_info.hw_mode); if (rc != ECORE_SUCCESS) return rc; /* start PF's qm queues */ OSAL_SPIN_LOCK(&qm_lock); b_rc = ecore_send_qm_stop_cmd(p_hwfn, p_ptt, true, true, qm_info->start_pq, qm_info->num_pqs); OSAL_SPIN_UNLOCK(&qm_lock); if (!b_rc) return ECORE_INVAL; return ECORE_SUCCESS; } static enum _ecore_status_t ecore_alloc_qm_data(struct ecore_hwfn *p_hwfn) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; enum _ecore_status_t rc; rc = ecore_init_qm_sanity(p_hwfn); if (rc != ECORE_SUCCESS) goto alloc_err; qm_info->qm_pq_params = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(struct init_qm_pq_params) * ecore_init_qm_get_num_pqs(p_hwfn)); if (!qm_info->qm_pq_params) goto alloc_err; qm_info->qm_vport_params = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(struct init_qm_vport_params) * ecore_init_qm_get_num_vports(p_hwfn)); if (!qm_info->qm_vport_params) goto alloc_err; qm_info->qm_port_params = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(struct init_qm_port_params) * p_hwfn->p_dev->num_ports_in_engine); if (!qm_info->qm_port_params) goto alloc_err; qm_info->wfq_data = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(struct ecore_wfq_data) * ecore_init_qm_get_num_vports(p_hwfn)); if (!qm_info->wfq_data) goto alloc_err; return ECORE_SUCCESS; alloc_err: DP_NOTICE(p_hwfn, false, "Failed to allocate memory for QM params\n"); ecore_qm_info_free(p_hwfn); return ECORE_NOMEM; } /******************** End QM initialization ***************/ enum _ecore_status_t ecore_resc_alloc(struct ecore_dev *p_dev) { enum _ecore_status_t rc = ECORE_SUCCESS; int i; if (IS_VF(p_dev)) { for_each_hwfn(p_dev, i) { rc = ecore_l2_alloc(&p_dev->hwfns[i]); if (rc != ECORE_SUCCESS) return rc; } return rc; } p_dev->fw_data = OSAL_ZALLOC(p_dev, GFP_KERNEL, sizeof(*p_dev->fw_data)); if (!p_dev->fw_data) return ECORE_NOMEM; for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; u32 n_eqes, num_cons; /* initialize the doorbell recovery mechanism */ rc = ecore_db_recovery_setup(p_hwfn); if (rc) goto alloc_err; /* First allocate the context manager structure */ rc = ecore_cxt_mngr_alloc(p_hwfn); if (rc) goto alloc_err; /* Set the HW cid/tid numbers (in the context manager) * Must be done prior to any further computations. */ rc = ecore_cxt_set_pf_params(p_hwfn); if (rc) goto alloc_err; rc = ecore_alloc_qm_data(p_hwfn); if (rc) goto alloc_err; /* init qm info */ ecore_init_qm_info(p_hwfn); /* Compute the ILT client partition */ rc = ecore_cxt_cfg_ilt_compute(p_hwfn); if (rc) goto alloc_err; /* CID map / ILT shadow table / T2 * The talbes sizes are determined by the computations above */ rc = ecore_cxt_tables_alloc(p_hwfn); if (rc) goto alloc_err; /* SPQ, must follow ILT because initializes SPQ context */ rc = ecore_spq_alloc(p_hwfn); if (rc) goto alloc_err; /* SP status block allocation */ p_hwfn->p_dpc_ptt = ecore_get_reserved_ptt(p_hwfn, RESERVED_PTT_DPC); rc = ecore_int_alloc(p_hwfn, p_hwfn->p_main_ptt); if (rc) goto alloc_err; rc = ecore_iov_alloc(p_hwfn); if (rc) goto alloc_err; /* EQ */ n_eqes = ecore_chain_get_capacity(&p_hwfn->p_spq->chain); if (ECORE_IS_RDMA_PERSONALITY(p_hwfn)) { /* Calculate the EQ size * --------------------- * Each ICID may generate up to one event at a time i.e. * the event must be handled/cleared before a new one * can be generated. We calculate the sum of events per * protocol and create an EQ deep enough to handle the * worst case: * - Core - according to SPQ. * - RoCE - per QP there are a couple of ICIDs, one * responder and one requester, each can * generate an EQE => n_eqes_qp = 2 * n_qp. * Each CQ can generate an EQE. There are 2 CQs * per QP => n_eqes_cq = 2 * n_qp. * Hence the RoCE total is 4 * n_qp or * 2 * num_cons. * - ENet - There can be up to two events per VF. One * for VF-PF channel and another for VF FLR * initial cleanup. The number of VFs is * bounded by MAX_NUM_VFS_BB, and is much * smaller than RoCE's so we avoid exact * calculation. */ if (ECORE_IS_ROCE_PERSONALITY(p_hwfn)) { num_cons = ecore_cxt_get_proto_cid_count( p_hwfn, PROTOCOLID_ROCE, OSAL_NULL); num_cons *= 2; } else { num_cons = ecore_cxt_get_proto_cid_count( p_hwfn, PROTOCOLID_IWARP, OSAL_NULL); } n_eqes += num_cons + 2 * MAX_NUM_VFS_BB; } else if (p_hwfn->hw_info.personality == ECORE_PCI_ISCSI) { num_cons = ecore_cxt_get_proto_cid_count(p_hwfn, PROTOCOLID_ISCSI, OSAL_NULL); n_eqes += 2 * num_cons; } if (n_eqes > 0xFFFF) { DP_ERR(p_hwfn, "Cannot allocate 0x%x EQ elements." "The maximum of a u16 chain is 0x%x\n", n_eqes, 0xFFFF); goto alloc_no_mem; } rc = ecore_eq_alloc(p_hwfn, (u16)n_eqes); if (rc) goto alloc_err; rc = ecore_consq_alloc(p_hwfn); if (rc) goto alloc_err; rc = ecore_l2_alloc(p_hwfn); if (rc != ECORE_SUCCESS) goto alloc_err; /* DMA info initialization */ rc = ecore_dmae_info_alloc(p_hwfn); if (rc) { DP_NOTICE(p_hwfn, true, "Failed to allocate memory for dmae_info" " structure\n"); goto alloc_err; } /* DCBX initialization */ rc = ecore_dcbx_info_alloc(p_hwfn); if (rc) { DP_NOTICE(p_hwfn, true, "Failed to allocate memory for dcbx structure\n"); goto alloc_err; } } p_dev->reset_stats = OSAL_ZALLOC(p_dev, GFP_KERNEL, sizeof(*p_dev->reset_stats)); if (!p_dev->reset_stats) { DP_NOTICE(p_dev, true, "Failed to allocate reset statistics\n"); goto alloc_no_mem; } return ECORE_SUCCESS; alloc_no_mem: rc = ECORE_NOMEM; alloc_err: ecore_resc_free(p_dev); return rc; } void ecore_resc_setup(struct ecore_dev *p_dev) { int i; if (IS_VF(p_dev)) { for_each_hwfn(p_dev, i) ecore_l2_setup(&p_dev->hwfns[i]); return; } for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; ecore_cxt_mngr_setup(p_hwfn); ecore_spq_setup(p_hwfn); ecore_eq_setup(p_hwfn); ecore_consq_setup(p_hwfn); /* Read shadow of current MFW mailbox */ ecore_mcp_read_mb(p_hwfn, p_hwfn->p_main_ptt); OSAL_MEMCPY(p_hwfn->mcp_info->mfw_mb_shadow, p_hwfn->mcp_info->mfw_mb_cur, p_hwfn->mcp_info->mfw_mb_length); ecore_int_setup(p_hwfn, p_hwfn->p_main_ptt); ecore_l2_setup(p_hwfn); ecore_iov_setup(p_hwfn); } } #define FINAL_CLEANUP_POLL_CNT (100) #define FINAL_CLEANUP_POLL_TIME (10) enum _ecore_status_t ecore_final_cleanup(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 id, bool is_vf) { u32 command = 0, addr, count = FINAL_CLEANUP_POLL_CNT; enum _ecore_status_t rc = ECORE_TIMEOUT; #ifndef ASIC_ONLY if (CHIP_REV_IS_TEDIBEAR(p_hwfn->p_dev) || CHIP_REV_IS_SLOW(p_hwfn->p_dev)) { DP_INFO(p_hwfn, "Skipping final cleanup for non-ASIC\n"); return ECORE_SUCCESS; } #endif addr = GTT_BAR0_MAP_REG_USDM_RAM + USTORM_FLR_FINAL_ACK_OFFSET(p_hwfn->rel_pf_id); if (is_vf) id += 0x10; command |= X_FINAL_CLEANUP_AGG_INT << SDM_AGG_INT_COMP_PARAMS_AGG_INT_INDEX_SHIFT; command |= 1 << SDM_AGG_INT_COMP_PARAMS_AGG_VECTOR_ENABLE_SHIFT; command |= id << SDM_AGG_INT_COMP_PARAMS_AGG_VECTOR_BIT_SHIFT; command |= SDM_COMP_TYPE_AGG_INT << SDM_OP_GEN_COMP_TYPE_SHIFT; /* Make sure notification is not set before initiating final cleanup */ if (REG_RD(p_hwfn, addr)) { DP_NOTICE(p_hwfn, false, "Unexpected; Found final cleanup notification"); DP_NOTICE(p_hwfn, false, " before initiating final cleanup\n"); REG_WR(p_hwfn, addr, 0); } DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, "Sending final cleanup for PFVF[%d] [Command %08x]\n", id, command); ecore_wr(p_hwfn, p_ptt, XSDM_REG_OPERATION_GEN, command); /* Poll until completion */ while (!REG_RD(p_hwfn, addr) && count--) OSAL_MSLEEP(FINAL_CLEANUP_POLL_TIME); if (REG_RD(p_hwfn, addr)) rc = ECORE_SUCCESS; else DP_NOTICE(p_hwfn, true, "Failed to receive FW final cleanup notification\n"); /* Cleanup afterwards */ REG_WR(p_hwfn, addr, 0); return rc; } static enum _ecore_status_t ecore_calc_hw_mode(struct ecore_hwfn *p_hwfn) { int hw_mode = 0; if (ECORE_IS_BB_B0(p_hwfn->p_dev)) { hw_mode |= 1 << MODE_BB; } else if (ECORE_IS_AH(p_hwfn->p_dev)) { hw_mode |= 1 << MODE_K2; } else { DP_NOTICE(p_hwfn, true, "Unknown chip type %#x\n", p_hwfn->p_dev->type); return ECORE_INVAL; } /* Ports per engine is based on the values in CNIG_REG_NW_PORT_MODE */ switch (p_hwfn->p_dev->num_ports_in_engine) { case 1: hw_mode |= 1 << MODE_PORTS_PER_ENG_1; break; case 2: hw_mode |= 1 << MODE_PORTS_PER_ENG_2; break; case 4: hw_mode |= 1 << MODE_PORTS_PER_ENG_4; break; default: DP_NOTICE(p_hwfn, true, "num_ports_in_engine = %d not supported\n", p_hwfn->p_dev->num_ports_in_engine); return ECORE_INVAL; } if (OSAL_TEST_BIT(ECORE_MF_OVLAN_CLSS, &p_hwfn->p_dev->mf_bits)) hw_mode |= 1 << MODE_MF_SD; else hw_mode |= 1 << MODE_MF_SI; #ifndef ASIC_ONLY if (CHIP_REV_IS_SLOW(p_hwfn->p_dev)) { if (CHIP_REV_IS_FPGA(p_hwfn->p_dev)) { hw_mode |= 1 << MODE_FPGA; } else { if (p_hwfn->p_dev->b_is_emul_full) hw_mode |= 1 << MODE_EMUL_FULL; else hw_mode |= 1 << MODE_EMUL_REDUCED; } } else #endif hw_mode |= 1 << MODE_ASIC; if (ECORE_IS_CMT(p_hwfn->p_dev)) hw_mode |= 1 << MODE_100G; p_hwfn->hw_info.hw_mode = hw_mode; DP_VERBOSE(p_hwfn, (ECORE_MSG_PROBE | ECORE_MSG_IFUP), "Configuring function for hw_mode: 0x%08x\n", p_hwfn->hw_info.hw_mode); return ECORE_SUCCESS; } #ifndef ASIC_ONLY /* MFW-replacement initializations for non-ASIC */ static enum _ecore_status_t ecore_hw_init_chip(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_dev *p_dev = p_hwfn->p_dev; u32 pl_hv = 1; int i; if (CHIP_REV_IS_EMUL(p_dev)) { if (ECORE_IS_AH(p_dev)) pl_hv |= 0x600; } ecore_wr(p_hwfn, p_ptt, MISCS_REG_RESET_PL_HV + 4, pl_hv); if (CHIP_REV_IS_EMUL(p_dev) && (ECORE_IS_AH(p_dev))) ecore_wr(p_hwfn, p_ptt, MISCS_REG_RESET_PL_HV_2_K2_E5, 0x3ffffff); /* initialize port mode to 4x10G_E (10G with 4x10 SERDES) */ /* CNIG_REG_NW_PORT_MODE is same for A0 and B0 */ if (!CHIP_REV_IS_EMUL(p_dev) || ECORE_IS_BB(p_dev)) ecore_wr(p_hwfn, p_ptt, CNIG_REG_NW_PORT_MODE_BB, 4); if (CHIP_REV_IS_EMUL(p_dev)) { if (ECORE_IS_AH(p_dev)) { /* 2 for 4-port, 1 for 2-port, 0 for 1-port */ ecore_wr(p_hwfn, p_ptt, MISC_REG_PORT_MODE, (p_dev->num_ports_in_engine >> 1)); ecore_wr(p_hwfn, p_ptt, MISC_REG_BLOCK_256B_EN, p_dev->num_ports_in_engine == 4 ? 0 : 3); } } /* Poll on RBC */ ecore_wr(p_hwfn, p_ptt, PSWRQ2_REG_RBC_DONE, 1); for (i = 0; i < 100; i++) { OSAL_UDELAY(50); if (ecore_rd(p_hwfn, p_ptt, PSWRQ2_REG_CFG_DONE) == 1) break; } if (i == 100) DP_NOTICE(p_hwfn, true, "RBC done failed to complete in PSWRQ2\n"); return ECORE_SUCCESS; } #endif /* Init run time data for all PFs and their VFs on an engine. * TBD - for VFs - Once we have parent PF info for each VF in * shmem available as CAU requires knowledge of parent PF for each VF. */ static void ecore_init_cau_rt_data(struct ecore_dev *p_dev) { u32 offset = CAU_REG_SB_VAR_MEMORY_RT_OFFSET; int i, igu_sb_id; for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; struct ecore_igu_info *p_igu_info; struct ecore_igu_block *p_block; struct cau_sb_entry sb_entry; p_igu_info = p_hwfn->hw_info.p_igu_info; for (igu_sb_id = 0; igu_sb_id < ECORE_MAPPING_MEMORY_SIZE(p_dev); igu_sb_id++) { p_block = &p_igu_info->entry[igu_sb_id]; if (!p_block->is_pf) continue; ecore_init_cau_sb_entry(p_hwfn, &sb_entry, p_block->function_id, 0, 0); STORE_RT_REG_AGG(p_hwfn, offset + igu_sb_id * 2, sb_entry); } } } static void ecore_init_cache_line_size(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { u32 val, wr_mbs, cache_line_size; val = ecore_rd(p_hwfn, p_ptt, PSWRQ2_REG_WR_MBS0); switch (val) { case 0: wr_mbs = 128; break; case 1: wr_mbs = 256; break; case 2: wr_mbs = 512; break; default: DP_INFO(p_hwfn, "Unexpected value of PSWRQ2_REG_WR_MBS0 [0x%x]. Avoid configuring PGLUE_B_REG_CACHE_LINE_SIZE.\n", val); return; } cache_line_size = OSAL_MIN_T(u32, OSAL_CACHE_LINE_SIZE, wr_mbs); switch (cache_line_size) { case 32: val = 0; break; case 64: val = 1; break; case 128: val = 2; break; case 256: val = 3; break; default: DP_INFO(p_hwfn, "Unexpected value of cache line size [0x%x]. Avoid configuring PGLUE_B_REG_CACHE_LINE_SIZE.\n", cache_line_size); } if (wr_mbs < OSAL_CACHE_LINE_SIZE) DP_INFO(p_hwfn, "The cache line size for padding is suboptimal for performance [OS cache line size 0x%x, wr mbs 0x%x]\n", OSAL_CACHE_LINE_SIZE, wr_mbs); STORE_RT_REG(p_hwfn, PGLUE_REG_B_CACHE_LINE_SIZE_RT_OFFSET, val); if (val > 0) { STORE_RT_REG(p_hwfn, PSWRQ2_REG_DRAM_ALIGN_WR_RT_OFFSET, val); STORE_RT_REG(p_hwfn, PSWRQ2_REG_DRAM_ALIGN_RD_RT_OFFSET, val); } } static enum _ecore_status_t ecore_hw_init_common(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, int hw_mode) { struct ecore_qm_info *qm_info = &p_hwfn->qm_info; struct ecore_dev *p_dev = p_hwfn->p_dev; u8 vf_id, max_num_vfs; u16 num_pfs, pf_id; u32 concrete_fid; enum _ecore_status_t rc = ECORE_SUCCESS; ecore_init_cau_rt_data(p_dev); /* Program GTT windows */ ecore_gtt_init(p_hwfn, p_ptt); #ifndef ASIC_ONLY if (CHIP_REV_IS_EMUL(p_dev)) { rc = ecore_hw_init_chip(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) return rc; } #endif if (p_hwfn->mcp_info) { if (p_hwfn->mcp_info->func_info.bandwidth_max) qm_info->pf_rl_en = 1; if (p_hwfn->mcp_info->func_info.bandwidth_min) qm_info->pf_wfq_en = 1; } ecore_qm_common_rt_init(p_hwfn, p_dev->num_ports_in_engine, qm_info->max_phys_tcs_per_port, qm_info->pf_rl_en, qm_info->pf_wfq_en, qm_info->vport_rl_en, qm_info->vport_wfq_en, qm_info->qm_port_params); ecore_cxt_hw_init_common(p_hwfn); ecore_init_cache_line_size(p_hwfn, p_ptt); rc = ecore_init_run(p_hwfn, p_ptt, PHASE_ENGINE, ANY_PHASE_ID, hw_mode); if (rc != ECORE_SUCCESS) return rc; /* @@TBD MichalK - should add VALIDATE_VFID to init tool... * need to decide with which value, maybe runtime */ ecore_wr(p_hwfn, p_ptt, PSWRQ2_REG_L2P_VALIDATE_VFID, 0); ecore_wr(p_hwfn, p_ptt, PGLUE_B_REG_USE_CLIENTID_IN_TAG, 1); if (ECORE_IS_BB(p_dev)) { /* Workaround clears ROCE search for all functions to prevent * involving non initialized function in processing ROCE packet. */ num_pfs = NUM_OF_ENG_PFS(p_dev); for (pf_id = 0; pf_id < num_pfs; pf_id++) { ecore_fid_pretend(p_hwfn, p_ptt, pf_id); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_ROCE, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_TCP, 0x0); } /* pretend to original PF */ ecore_fid_pretend(p_hwfn, p_ptt, p_hwfn->rel_pf_id); } /* Workaround for avoiding CCFC execution error when getting packets * with CRC errors, and allowing instead the invoking of the FW error * handler. * This is not done inside the init tool since it currently can't * perform a pretending to VFs. */ max_num_vfs = ECORE_IS_AH(p_dev) ? MAX_NUM_VFS_K2 : MAX_NUM_VFS_BB; for (vf_id = 0; vf_id < max_num_vfs; vf_id++) { concrete_fid = ecore_vfid_to_concrete(p_hwfn, vf_id); ecore_fid_pretend(p_hwfn, p_ptt, (u16)concrete_fid); ecore_wr(p_hwfn, p_ptt, CCFC_REG_STRONG_ENABLE_VF, 0x1); ecore_wr(p_hwfn, p_ptt, CCFC_REG_WEAK_ENABLE_VF, 0x0); ecore_wr(p_hwfn, p_ptt, TCFC_REG_STRONG_ENABLE_VF, 0x1); ecore_wr(p_hwfn, p_ptt, TCFC_REG_WEAK_ENABLE_VF, 0x0); } /* pretend to original PF */ ecore_fid_pretend(p_hwfn, p_ptt, p_hwfn->rel_pf_id); return rc; } #ifndef ASIC_ONLY #define MISC_REG_RESET_REG_2_XMAC_BIT (1 << 4) #define MISC_REG_RESET_REG_2_XMAC_SOFT_BIT (1 << 5) #define PMEG_IF_BYTE_COUNT 8 static void ecore_wr_nw_port(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 addr, u64 data, u8 reg_type, u8 port) { DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "CMD: %08x, ADDR: 0x%08x, DATA: %08x:%08x\n", ecore_rd(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_CMD_BB) | (8 << PMEG_IF_BYTE_COUNT), (reg_type << 25) | (addr << 8) | port, (u32)((data >> 32) & 0xffffffff), (u32)(data & 0xffffffff)); ecore_wr(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_CMD_BB, (ecore_rd(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_CMD_BB) & 0xffff00fe) | (8 << PMEG_IF_BYTE_COUNT)); ecore_wr(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_ADDR_BB, (reg_type << 25) | (addr << 8) | port); ecore_wr(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_WRDATA_BB, data & 0xffffffff); ecore_wr(p_hwfn, p_ptt, CNIG_REG_PMEG_IF_WRDATA_BB, (data >> 32) & 0xffffffff); } #define XLPORT_MODE_REG (0x20a) #define XLPORT_MAC_CONTROL (0x210) #define XLPORT_FLOW_CONTROL_CONFIG (0x207) #define XLPORT_ENABLE_REG (0x20b) #define XLMAC_CTRL (0x600) #define XLMAC_MODE (0x601) #define XLMAC_RX_MAX_SIZE (0x608) #define XLMAC_TX_CTRL (0x604) #define XLMAC_PAUSE_CTRL (0x60d) #define XLMAC_PFC_CTRL (0x60e) static void ecore_emul_link_init_bb(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { u8 loopback = 0, port = p_hwfn->port_id * 2; DP_INFO(p_hwfn->p_dev, "Configurating Emulation Link %02x\n", port); /* XLPORT MAC MODE *//* 0 Quad, 4 Single... */ ecore_wr_nw_port(p_hwfn, p_ptt, XLPORT_MODE_REG, (0x4 << 4) | 0x4, 1, port); ecore_wr_nw_port(p_hwfn, p_ptt, XLPORT_MAC_CONTROL, 0, 1, port); /* XLMAC: SOFT RESET */ ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_CTRL, 0x40, 0, port); /* XLMAC: Port Speed >= 10Gbps */ ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_MODE, 0x40, 0, port); /* XLMAC: Max Size */ ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_RX_MAX_SIZE, 0x3fff, 0, port); ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_TX_CTRL, 0x01000000800ULL | (0xa << 12) | ((u64)1 << 38), 0, port); ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_PAUSE_CTRL, 0x7c000, 0, port); ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_PFC_CTRL, 0x30ffffc000ULL, 0, port); ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_CTRL, 0x3 | (loopback << 2), 0, port); /* XLMAC: TX_EN, RX_EN */ /* XLMAC: TX_EN, RX_EN, SW_LINK_STATUS */ ecore_wr_nw_port(p_hwfn, p_ptt, XLMAC_CTRL, 0x1003 | (loopback << 2), 0, port); /* Enabled Parallel PFC interface */ ecore_wr_nw_port(p_hwfn, p_ptt, XLPORT_FLOW_CONTROL_CONFIG, 1, 0, port); /* XLPORT port enable */ ecore_wr_nw_port(p_hwfn, p_ptt, XLPORT_ENABLE_REG, 0xf, 1, port); } static void ecore_emul_link_init_ah_e5(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { u8 port = p_hwfn->port_id; u32 mac_base = NWM_REG_MAC0_K2_E5 + (port << 2) * NWM_REG_MAC0_SIZE; DP_INFO(p_hwfn->p_dev, "Configurating Emulation Link %02x\n", port); ecore_wr(p_hwfn, p_ptt, CNIG_REG_NIG_PORT0_CONF_K2_E5 + (port << 2), (1 << CNIG_REG_NIG_PORT0_CONF_NIG_PORT_ENABLE_0_K2_E5_SHIFT) | (port << CNIG_REG_NIG_PORT0_CONF_NIG_PORT_NWM_PORT_MAP_0_K2_E5_SHIFT) | (0 << CNIG_REG_NIG_PORT0_CONF_NIG_PORT_RATE_0_K2_E5_SHIFT)); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_XIF_MODE_K2_E5, 1 << ETH_MAC_REG_XIF_MODE_XGMII_K2_E5_SHIFT); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_FRM_LENGTH_K2_E5, 9018 << ETH_MAC_REG_FRM_LENGTH_FRM_LENGTH_K2_E5_SHIFT); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_TX_IPG_LENGTH_K2_E5, 0xc << ETH_MAC_REG_TX_IPG_LENGTH_TXIPG_K2_E5_SHIFT); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_RX_FIFO_SECTIONS_K2_E5, 8 << ETH_MAC_REG_RX_FIFO_SECTIONS_RX_SECTION_FULL_K2_E5_SHIFT); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_TX_FIFO_SECTIONS_K2_E5, (0xA << ETH_MAC_REG_TX_FIFO_SECTIONS_TX_SECTION_EMPTY_K2_E5_SHIFT) | (8 << ETH_MAC_REG_TX_FIFO_SECTIONS_TX_SECTION_FULL_K2_E5_SHIFT)); ecore_wr(p_hwfn, p_ptt, mac_base + ETH_MAC_REG_COMMAND_CONFIG_K2_E5, 0xa853); } static void ecore_emul_link_init(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { if (ECORE_IS_AH(p_hwfn->p_dev)) ecore_emul_link_init_ah_e5(p_hwfn, p_ptt); else /* BB */ ecore_emul_link_init_bb(p_hwfn, p_ptt); } static void ecore_link_init_bb(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u8 port) { int port_offset = port ? 0x800 : 0; u32 xmac_rxctrl = 0; /* Reset of XMAC */ /* FIXME: move to common start */ ecore_wr(p_hwfn, p_ptt, MISC_REG_RESET_PL_PDA_VAUX + 2 * sizeof(u32), MISC_REG_RESET_REG_2_XMAC_BIT); /* Clear */ OSAL_MSLEEP(1); ecore_wr(p_hwfn, p_ptt, MISC_REG_RESET_PL_PDA_VAUX + sizeof(u32), MISC_REG_RESET_REG_2_XMAC_BIT); /* Set */ ecore_wr(p_hwfn, p_ptt, MISC_REG_XMAC_CORE_PORT_MODE_BB, 1); /* Set the number of ports on the Warp Core to 10G */ ecore_wr(p_hwfn, p_ptt, MISC_REG_XMAC_PHY_PORT_MODE_BB, 3); /* Soft reset of XMAC */ ecore_wr(p_hwfn, p_ptt, MISC_REG_RESET_PL_PDA_VAUX + 2 * sizeof(u32), MISC_REG_RESET_REG_2_XMAC_SOFT_BIT); OSAL_MSLEEP(1); ecore_wr(p_hwfn, p_ptt, MISC_REG_RESET_PL_PDA_VAUX + sizeof(u32), MISC_REG_RESET_REG_2_XMAC_SOFT_BIT); /* FIXME: move to common end */ if (CHIP_REV_IS_FPGA(p_hwfn->p_dev)) ecore_wr(p_hwfn, p_ptt, XMAC_REG_MODE_BB + port_offset, 0x20); /* Set Max packet size: initialize XMAC block register for port 0 */ ecore_wr(p_hwfn, p_ptt, XMAC_REG_RX_MAX_SIZE_BB + port_offset, 0x2710); /* CRC append for Tx packets: init XMAC block register for port 1 */ ecore_wr(p_hwfn, p_ptt, XMAC_REG_TX_CTRL_LO_BB + port_offset, 0xC800); /* Enable TX and RX: initialize XMAC block register for port 1 */ ecore_wr(p_hwfn, p_ptt, XMAC_REG_CTRL_BB + port_offset, XMAC_REG_CTRL_TX_EN_BB | XMAC_REG_CTRL_RX_EN_BB); xmac_rxctrl = ecore_rd(p_hwfn, p_ptt, XMAC_REG_RX_CTRL_BB + port_offset); xmac_rxctrl |= XMAC_REG_RX_CTRL_PROCESS_VARIABLE_PREAMBLE_BB; ecore_wr(p_hwfn, p_ptt, XMAC_REG_RX_CTRL_BB + port_offset, xmac_rxctrl); } #endif static enum _ecore_status_t ecore_hw_init_dpi_size(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 pwm_region_size, u32 n_cpus) { u32 dpi_bit_shift, dpi_count, dpi_page_size; u32 min_dpis; u32 n_wids; /* Calculate DPI size * ------------------ * The PWM region contains Doorbell Pages. The first is reserverd for * the kernel for, e.g, L2. The others are free to be used by non- * trusted applications, typically from user space. Each page, called a * doorbell page is sectioned into windows that allow doorbells to be * issued in parallel by the kernel/application. The size of such a * window (a.k.a. WID) is 1kB. * Summary: * 1kB WID x N WIDS = DPI page size * DPI page size x N DPIs = PWM region size * Notes: * The size of the DPI page size must be in multiples of OSAL_PAGE_SIZE * in order to ensure that two applications won't share the same page. * It also must contain at least one WID per CPU to allow parallelism. * It also must be a power of 2, since it is stored as a bit shift. * * The DPI page size is stored in a register as 'dpi_bit_shift' so that * 0 is 4kB, 1 is 8kB and etc. Hence the minimum size is 4,096 * containing 4 WIDs. */ n_wids = OSAL_MAX_T(u32, ECORE_MIN_WIDS, n_cpus); dpi_page_size = ECORE_WID_SIZE * OSAL_ROUNDUP_POW_OF_TWO(n_wids); dpi_page_size = (dpi_page_size + OSAL_PAGE_SIZE - 1) & ~(OSAL_PAGE_SIZE - 1); dpi_bit_shift = OSAL_LOG2(dpi_page_size / 4096); dpi_count = pwm_region_size / dpi_page_size; min_dpis = p_hwfn->pf_params.rdma_pf_params.min_dpis; min_dpis = OSAL_MAX_T(u32, ECORE_MIN_DPIS, min_dpis); /* Update hwfn */ p_hwfn->dpi_size = dpi_page_size; p_hwfn->dpi_count = dpi_count; /* Update registers */ ecore_wr(p_hwfn, p_ptt, DORQ_REG_PF_DPI_BIT_SHIFT, dpi_bit_shift); if (dpi_count < min_dpis) return ECORE_NORESOURCES; return ECORE_SUCCESS; } enum ECORE_ROCE_EDPM_MODE { ECORE_ROCE_EDPM_MODE_ENABLE = 0, ECORE_ROCE_EDPM_MODE_FORCE_ON = 1, ECORE_ROCE_EDPM_MODE_DISABLE = 2, }; static enum _ecore_status_t ecore_hw_init_pf_doorbell_bar(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { u32 pwm_regsize, norm_regsize; u32 non_pwm_conn, min_addr_reg1; u32 db_bar_size, n_cpus; u32 roce_edpm_mode; u32 pf_dems_shift; enum _ecore_status_t rc = ECORE_SUCCESS; u8 cond; db_bar_size = ecore_hw_bar_size(p_hwfn, p_ptt, BAR_ID_1); if (ECORE_IS_CMT(p_hwfn->p_dev)) db_bar_size /= 2; /* Calculate doorbell regions * ----------------------------------- * The doorbell BAR is made of two regions. The first is called normal * region and the second is called PWM region. In the normal region * each ICID has its own set of addresses so that writing to that * specific address identifies the ICID. In the Process Window Mode * region the ICID is given in the data written to the doorbell. The * above per PF register denotes the offset in the doorbell BAR in which * the PWM region begins. * The normal region has ECORE_PF_DEMS_SIZE bytes per ICID, that is per * non-PWM connection. The calculation below computes the total non-PWM * connections. The DORQ_REG_PF_MIN_ADDR_REG1 register is * in units of 4,096 bytes. */ non_pwm_conn = ecore_cxt_get_proto_cid_start(p_hwfn, PROTOCOLID_CORE) + ecore_cxt_get_proto_cid_count(p_hwfn, PROTOCOLID_CORE, OSAL_NULL) + ecore_cxt_get_proto_cid_count(p_hwfn, PROTOCOLID_ETH, OSAL_NULL); norm_regsize = ROUNDUP(ECORE_PF_DEMS_SIZE * non_pwm_conn, OSAL_PAGE_SIZE); min_addr_reg1 = norm_regsize / 4096; pwm_regsize = db_bar_size - norm_regsize; /* Check that the normal and PWM sizes are valid */ if (db_bar_size < norm_regsize) { DP_ERR(p_hwfn->p_dev, "Doorbell BAR size 0x%x is too small (normal region is 0x%0x )\n", db_bar_size, norm_regsize); return ECORE_NORESOURCES; } if (pwm_regsize < ECORE_MIN_PWM_REGION) { DP_ERR(p_hwfn->p_dev, "PWM region size 0x%0x is too small. Should be at least 0x%0x (Doorbell BAR size is 0x%x and normal region size is 0x%0x)\n", pwm_regsize, ECORE_MIN_PWM_REGION, db_bar_size, norm_regsize); return ECORE_NORESOURCES; } /* Calculate number of DPIs */ roce_edpm_mode = p_hwfn->pf_params.rdma_pf_params.roce_edpm_mode; if ((roce_edpm_mode == ECORE_ROCE_EDPM_MODE_ENABLE) || ((roce_edpm_mode == ECORE_ROCE_EDPM_MODE_FORCE_ON))) { /* Either EDPM is mandatory, or we are attempting to allocate a * WID per CPU. */ n_cpus = OSAL_NUM_CPUS(); rc = ecore_hw_init_dpi_size(p_hwfn, p_ptt, pwm_regsize, n_cpus); } cond = ((rc != ECORE_SUCCESS) && (roce_edpm_mode == ECORE_ROCE_EDPM_MODE_ENABLE)) || (roce_edpm_mode == ECORE_ROCE_EDPM_MODE_DISABLE); if (cond || p_hwfn->dcbx_no_edpm) { /* Either EDPM is disabled from user configuration, or it is * disabled via DCBx, or it is not mandatory and we failed to * allocated a WID per CPU. */ n_cpus = 1; rc = ecore_hw_init_dpi_size(p_hwfn, p_ptt, pwm_regsize, n_cpus); /* If we entered this flow due to DCBX then the DPM register is * already configured. */ } DP_INFO(p_hwfn, "doorbell bar: normal_region_size=%d, pwm_region_size=%d", norm_regsize, pwm_regsize); DP_INFO(p_hwfn, " dpi_size=%d, dpi_count=%d, roce_edpm=%s\n", p_hwfn->dpi_size, p_hwfn->dpi_count, ((p_hwfn->dcbx_no_edpm) || (p_hwfn->db_bar_no_edpm)) ? "disabled" : "enabled"); /* Check return codes from above calls */ if (rc != ECORE_SUCCESS) { DP_ERR(p_hwfn, "Failed to allocate enough DPIs\n"); return ECORE_NORESOURCES; } /* Update hwfn */ p_hwfn->dpi_start_offset = norm_regsize; /* Update registers */ /* DEMS size is configured log2 of DWORDs, hence the division by 4 */ pf_dems_shift = OSAL_LOG2(ECORE_PF_DEMS_SIZE / 4); ecore_wr(p_hwfn, p_ptt, DORQ_REG_PF_ICID_BIT_SHIFT_NORM, pf_dems_shift); ecore_wr(p_hwfn, p_ptt, DORQ_REG_PF_MIN_ADDR_REG1, min_addr_reg1); return ECORE_SUCCESS; } static enum _ecore_status_t ecore_hw_init_port(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, int hw_mode) { u32 ppf_to_eng_sel[NIG_REG_PPF_TO_ENGINE_SEL_RT_SIZE]; u32 val; enum _ecore_status_t rc = ECORE_SUCCESS; u8 i; /* In CMT for non-RoCE packets - use connection based classification */ val = ECORE_IS_CMT(p_hwfn->p_dev) ? 0x8 : 0x0; for (i = 0; i < NIG_REG_PPF_TO_ENGINE_SEL_RT_SIZE; i++) ppf_to_eng_sel[i] = val; STORE_RT_REG_AGG(p_hwfn, NIG_REG_PPF_TO_ENGINE_SEL_RT_OFFSET, ppf_to_eng_sel); /* In CMT the gate should be cleared by the 2nd hwfn */ if (!ECORE_IS_CMT(p_hwfn->p_dev) || !IS_LEAD_HWFN(p_hwfn)) STORE_RT_REG(p_hwfn, NIG_REG_BRB_GATE_DNTFWD_PORT_RT_OFFSET, 0); rc = ecore_init_run(p_hwfn, p_ptt, PHASE_PORT, p_hwfn->port_id, hw_mode); if (rc != ECORE_SUCCESS) return rc; ecore_wr(p_hwfn, p_ptt, PGLUE_B_REG_MASTER_WRITE_PAD_ENABLE, 0); #ifndef ASIC_ONLY if (CHIP_REV_IS_ASIC(p_hwfn->p_dev)) return ECORE_SUCCESS; if (CHIP_REV_IS_FPGA(p_hwfn->p_dev)) { if (ECORE_IS_AH(p_hwfn->p_dev)) return ECORE_SUCCESS; else if (ECORE_IS_BB(p_hwfn->p_dev)) ecore_link_init_bb(p_hwfn, p_ptt, p_hwfn->port_id); } else if (CHIP_REV_IS_EMUL(p_hwfn->p_dev)) { if (ECORE_IS_CMT(p_hwfn->p_dev)) { /* Activate OPTE in CMT */ u32 val; val = ecore_rd(p_hwfn, p_ptt, MISCS_REG_RESET_PL_HV); val |= 0x10; ecore_wr(p_hwfn, p_ptt, MISCS_REG_RESET_PL_HV, val); ecore_wr(p_hwfn, p_ptt, MISC_REG_CLK_100G_MODE, 1); ecore_wr(p_hwfn, p_ptt, MISCS_REG_CLK_100G_MODE, 1); ecore_wr(p_hwfn, p_ptt, MISC_REG_OPTE_MODE, 1); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_ENG_CLS_TCP_4_TUPLE_SEARCH, 1); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_ENG_CLS_ENG_ID_TBL, 0x55555555); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_ENG_CLS_ENG_ID_TBL + 0x4, 0x55555555); } ecore_emul_link_init(p_hwfn, p_ptt); } else { DP_INFO(p_hwfn->p_dev, "link is not being configured\n"); } #endif return rc; } static enum _ecore_status_t ecore_hw_init_pf(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, struct ecore_tunnel_info *p_tunn, int hw_mode, bool b_hw_start, enum ecore_int_mode int_mode, bool allow_npar_tx_switch) { u8 rel_pf_id = p_hwfn->rel_pf_id; u32 prs_reg; enum _ecore_status_t rc = ECORE_SUCCESS; u16 ctrl; int pos; if (p_hwfn->mcp_info) { struct ecore_mcp_function_info *p_info; p_info = &p_hwfn->mcp_info->func_info; if (p_info->bandwidth_min) p_hwfn->qm_info.pf_wfq = p_info->bandwidth_min; /* Update rate limit once we'll actually have a link */ p_hwfn->qm_info.pf_rl = 100000; } ecore_cxt_hw_init_pf(p_hwfn, p_ptt); ecore_int_igu_init_rt(p_hwfn); /* Set VLAN in NIG if needed */ if (hw_mode & (1 << MODE_MF_SD)) { DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "Configuring LLH_FUNC_TAG\n"); STORE_RT_REG(p_hwfn, NIG_REG_LLH_FUNC_TAG_EN_RT_OFFSET, 1); STORE_RT_REG(p_hwfn, NIG_REG_LLH_FUNC_TAG_VALUE_RT_OFFSET, p_hwfn->hw_info.ovlan); DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "Configuring LLH_FUNC_FILTER_HDR_SEL\n"); STORE_RT_REG(p_hwfn, NIG_REG_LLH_FUNC_FILTER_HDR_SEL_RT_OFFSET, 1); } /* Enable classification by MAC if needed */ if (hw_mode & (1 << MODE_MF_SI)) { DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "Configuring TAGMAC_CLS_TYPE\n"); STORE_RT_REG(p_hwfn, NIG_REG_LLH_FUNC_TAGMAC_CLS_TYPE_RT_OFFSET, 1); } /* Protocl Configuration - @@@TBD - should we set 0 otherwise? */ STORE_RT_REG(p_hwfn, PRS_REG_SEARCH_TCP_RT_OFFSET, (p_hwfn->hw_info.personality == ECORE_PCI_ISCSI) ? 1 : 0); STORE_RT_REG(p_hwfn, PRS_REG_SEARCH_FCOE_RT_OFFSET, (p_hwfn->hw_info.personality == ECORE_PCI_FCOE) ? 1 : 0); STORE_RT_REG(p_hwfn, PRS_REG_SEARCH_ROCE_RT_OFFSET, 0); /* perform debug configuration when chip is out of reset */ OSAL_BEFORE_PF_START((void *)p_hwfn->p_dev, p_hwfn->my_id); /* PF Init sequence */ rc = ecore_init_run(p_hwfn, p_ptt, PHASE_PF, rel_pf_id, hw_mode); if (rc) return rc; /* QM_PF Init sequence (may be invoked separately e.g. for DCB) */ rc = ecore_init_run(p_hwfn, p_ptt, PHASE_QM_PF, rel_pf_id, hw_mode); if (rc) return rc; /* Pure runtime initializations - directly to the HW */ ecore_int_igu_init_pure_rt(p_hwfn, p_ptt, true, true); /* PCI relaxed ordering causes a decrease in the performance on some * systems. Till a root cause is found, disable this attribute in the * PCI config space. */ /* Not in use @DPDK * pos = OSAL_PCI_FIND_CAPABILITY(p_hwfn->p_dev, PCI_CAP_ID_EXP); * if (!pos) { * DP_NOTICE(p_hwfn, true, * "Failed to find the PCIe Cap\n"); * return ECORE_IO; * } * OSAL_PCI_READ_CONFIG_WORD(p_hwfn->p_dev, pos + PCI_EXP_DEVCTL, &ctrl); * ctrl &= ~PCI_EXP_DEVCTL_RELAX_EN; * OSAL_PCI_WRITE_CONFIG_WORD(p_hwfn->p_dev, pos + PCI_EXP_DEVCTL, ctrl); */ rc = ecore_hw_init_pf_doorbell_bar(p_hwfn, p_ptt); if (rc) return rc; if (b_hw_start) { /* enable interrupts */ rc = ecore_int_igu_enable(p_hwfn, p_ptt, int_mode); if (rc != ECORE_SUCCESS) return rc; /* send function start command */ rc = ecore_sp_pf_start(p_hwfn, p_ptt, p_tunn, allow_npar_tx_switch); if (rc) { DP_NOTICE(p_hwfn, true, "Function start ramrod failed\n"); } else { prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_TAG1); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_TAG1: %x\n", prs_reg); if (p_hwfn->hw_info.personality == ECORE_PCI_FCOE) { ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_TAG1, (1 << 2)); ecore_wr(p_hwfn, p_ptt, PRS_REG_PKT_LEN_STAT_TAGS_NOT_COUNTED_FIRST, 0x100); } DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH registers after start PFn\n"); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_TCP); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_TCP: %x\n", prs_reg); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_UDP); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_UDP: %x\n", prs_reg); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_FCOE); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_FCOE: %x\n", prs_reg); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_ROCE); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_ROCE: %x\n", prs_reg); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_TCP_FIRST_FRAG); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_TCP_FIRST_FRAG: %x\n", prs_reg); prs_reg = ecore_rd(p_hwfn, p_ptt, PRS_REG_SEARCH_TAG1); DP_VERBOSE(p_hwfn, ECORE_MSG_STORAGE, "PRS_REG_SEARCH_TAG1: %x\n", prs_reg); } } return rc; } enum _ecore_status_t ecore_pglueb_set_pfid_enable(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, bool b_enable) { u32 delay_idx = 0, val, set_val = b_enable ? 1 : 0; /* Configure the PF's internal FID_enable for master transactions */ ecore_wr(p_hwfn, p_ptt, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, set_val); /* Wait until value is set - try for 1 second every 50us */ for (delay_idx = 0; delay_idx < 20000; delay_idx++) { val = ecore_rd(p_hwfn, p_ptt, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER); if (val == set_val) break; OSAL_UDELAY(50); } if (val != set_val) { DP_NOTICE(p_hwfn, true, "PFID_ENABLE_MASTER wasn't changed after a second\n"); return ECORE_UNKNOWN_ERROR; } return ECORE_SUCCESS; } static void ecore_reset_mb_shadow(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_main_ptt) { /* Read shadow of current MFW mailbox */ ecore_mcp_read_mb(p_hwfn, p_main_ptt); OSAL_MEMCPY(p_hwfn->mcp_info->mfw_mb_shadow, p_hwfn->mcp_info->mfw_mb_cur, p_hwfn->mcp_info->mfw_mb_length); } static void ecore_pglueb_clear_err(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { ecore_wr(p_hwfn, p_ptt, PGLUE_B_REG_WAS_ERROR_PF_31_0_CLR, 1 << p_hwfn->abs_pf_id); } static void ecore_fill_load_req_params(struct ecore_load_req_params *p_load_req, struct ecore_drv_load_params *p_drv_load) { /* Make sure that if ecore-client didn't provide inputs, all the * expected defaults are indeed zero. */ OSAL_BUILD_BUG_ON(ECORE_DRV_ROLE_OS != 0); OSAL_BUILD_BUG_ON(ECORE_LOAD_REQ_LOCK_TO_DEFAULT != 0); OSAL_BUILD_BUG_ON(ECORE_OVERRIDE_FORCE_LOAD_NONE != 0); OSAL_MEM_ZERO(p_load_req, sizeof(*p_load_req)); if (p_drv_load != OSAL_NULL) { p_load_req->drv_role = p_drv_load->is_crash_kernel ? ECORE_DRV_ROLE_KDUMP : ECORE_DRV_ROLE_OS; p_load_req->timeout_val = p_drv_load->mfw_timeout_val; p_load_req->avoid_eng_reset = p_drv_load->avoid_eng_reset; p_load_req->override_force_load = p_drv_load->override_force_load; } } enum _ecore_status_t ecore_vf_start(struct ecore_hwfn *p_hwfn, struct ecore_hw_init_params *p_params) { if (p_params->p_tunn) { ecore_vf_set_vf_start_tunn_update_param(p_params->p_tunn); ecore_vf_pf_tunnel_param_update(p_hwfn, p_params->p_tunn); } p_hwfn->b_int_enabled = 1; return ECORE_SUCCESS; } enum _ecore_status_t ecore_hw_init(struct ecore_dev *p_dev, struct ecore_hw_init_params *p_params) { struct ecore_load_req_params load_req_params; u32 load_code, resp, param, drv_mb_param; bool b_default_mtu = true; struct ecore_hwfn *p_hwfn; enum _ecore_status_t rc = ECORE_SUCCESS; int i; if ((p_params->int_mode == ECORE_INT_MODE_MSI) && ECORE_IS_CMT(p_dev)) { DP_NOTICE(p_dev, false, "MSI mode is not supported for CMT devices\n"); return ECORE_INVAL; } if (IS_PF(p_dev)) { rc = ecore_init_fw_data(p_dev, p_params->bin_fw_data); if (rc != ECORE_SUCCESS) return rc; } for_each_hwfn(p_dev, i) { p_hwfn = &p_dev->hwfns[i]; /* If management didn't provide a default, set one of our own */ if (!p_hwfn->hw_info.mtu) { p_hwfn->hw_info.mtu = 1500; b_default_mtu = false; } if (IS_VF(p_dev)) { ecore_vf_start(p_hwfn, p_params); continue; } rc = ecore_calc_hw_mode(p_hwfn); if (rc != ECORE_SUCCESS) return rc; ecore_fill_load_req_params(&load_req_params, p_params->p_drv_load_params); rc = ecore_mcp_load_req(p_hwfn, p_hwfn->p_main_ptt, &load_req_params); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed sending a LOAD_REQ command\n"); return rc; } load_code = load_req_params.load_code; DP_VERBOSE(p_hwfn, ECORE_MSG_SP, "Load request was sent. Load code: 0x%x\n", load_code); ecore_mcp_set_capabilities(p_hwfn, p_hwfn->p_main_ptt); /* CQ75580: * When coming back from hiberbate state, the registers from * which shadow is read initially are not initialized. It turns * out that these registers get initialized during the call to * ecore_mcp_load_req request. So we need to reread them here * to get the proper shadow register value. * Note: This is a workaround for the missing MFW * initialization. It may be removed once the implementation * is done. */ ecore_reset_mb_shadow(p_hwfn, p_hwfn->p_main_ptt); /* Only relevant for recovery: * Clear the indication after the LOAD_REQ command is responded * by the MFW. */ p_dev->recov_in_prog = false; p_hwfn->first_on_engine = (load_code == FW_MSG_CODE_DRV_LOAD_ENGINE); if (!qm_lock_init) { OSAL_SPIN_LOCK_INIT(&qm_lock); qm_lock_init = true; } /* Clean up chip from previous driver if such remains exist. * This is not needed when the PF is the first one on the * engine, since afterwards we are going to init the FW. */ if (load_code != FW_MSG_CODE_DRV_LOAD_ENGINE) { rc = ecore_final_cleanup(p_hwfn, p_hwfn->p_main_ptt, p_hwfn->rel_pf_id, false); if (rc != ECORE_SUCCESS) { ecore_hw_err_notify(p_hwfn, ECORE_HW_ERR_RAMROD_FAIL); goto load_err; } } /* Log and clear previous pglue_b errors if such exist */ ecore_pglueb_rbc_attn_handler(p_hwfn, p_hwfn->p_main_ptt); /* Enable the PF's internal FID_enable in the PXP */ rc = ecore_pglueb_set_pfid_enable(p_hwfn, p_hwfn->p_main_ptt, true); if (rc != ECORE_SUCCESS) goto load_err; /* Clear the pglue_b was_error indication. * In E4 it must be done after the BME and the internal * FID_enable for the PF are set, since VDMs may cause the * indication to be set again. */ ecore_pglueb_clear_err(p_hwfn, p_hwfn->p_main_ptt); switch (load_code) { case FW_MSG_CODE_DRV_LOAD_ENGINE: rc = ecore_hw_init_common(p_hwfn, p_hwfn->p_main_ptt, p_hwfn->hw_info.hw_mode); if (rc != ECORE_SUCCESS) break; /* Fall into */ case FW_MSG_CODE_DRV_LOAD_PORT: rc = ecore_hw_init_port(p_hwfn, p_hwfn->p_main_ptt, p_hwfn->hw_info.hw_mode); if (rc != ECORE_SUCCESS) break; /* Fall into */ case FW_MSG_CODE_DRV_LOAD_FUNCTION: rc = ecore_hw_init_pf(p_hwfn, p_hwfn->p_main_ptt, p_params->p_tunn, p_hwfn->hw_info.hw_mode, p_params->b_hw_start, p_params->int_mode, p_params->allow_npar_tx_switch); break; default: DP_NOTICE(p_hwfn, false, "Unexpected load code [0x%08x]", load_code); rc = ECORE_NOTIMPL; break; } if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "init phase failed for loadcode 0x%x (rc %d)\n", load_code, rc); goto load_err; } rc = ecore_mcp_load_done(p_hwfn, p_hwfn->p_main_ptt); if (rc != ECORE_SUCCESS) return rc; /* send DCBX attention request command */ DP_VERBOSE(p_hwfn, ECORE_MSG_DCB, "sending phony dcbx set command to trigger DCBx attention handling\n"); rc = ecore_mcp_cmd(p_hwfn, p_hwfn->p_main_ptt, DRV_MSG_CODE_SET_DCBX, 1 << DRV_MB_PARAM_DCBX_NOTIFY_OFFSET, &resp, ¶m); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed to send DCBX attention request\n"); return rc; } p_hwfn->hw_init_done = true; } if (IS_PF(p_dev)) { p_hwfn = ECORE_LEADING_HWFN(p_dev); drv_mb_param = STORM_FW_VERSION; rc = ecore_mcp_cmd(p_hwfn, p_hwfn->p_main_ptt, DRV_MSG_CODE_OV_UPDATE_STORM_FW_VER, drv_mb_param, &resp, ¶m); if (rc != ECORE_SUCCESS) DP_INFO(p_hwfn, "Failed to update firmware version\n"); if (!b_default_mtu) rc = ecore_mcp_ov_update_mtu(p_hwfn, p_hwfn->p_main_ptt, p_hwfn->hw_info.mtu); if (rc != ECORE_SUCCESS) DP_INFO(p_hwfn, "Failed to update default mtu\n"); rc = ecore_mcp_ov_update_driver_state(p_hwfn, p_hwfn->p_main_ptt, ECORE_OV_DRIVER_STATE_DISABLED); if (rc != ECORE_SUCCESS) DP_INFO(p_hwfn, "Failed to update driver state\n"); } return rc; load_err: /* The MFW load lock should be released regardless of success or failure * of initialization. * TODO: replace this with an attempt to send cancel_load. */ ecore_mcp_load_done(p_hwfn, p_hwfn->p_main_ptt); return rc; } #define ECORE_HW_STOP_RETRY_LIMIT (10) static void ecore_hw_timers_stop(struct ecore_dev *p_dev, struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { int i; /* close timers */ ecore_wr(p_hwfn, p_ptt, TM_REG_PF_ENABLE_CONN, 0x0); ecore_wr(p_hwfn, p_ptt, TM_REG_PF_ENABLE_TASK, 0x0); for (i = 0; i < ECORE_HW_STOP_RETRY_LIMIT && !p_dev->recov_in_prog; i++) { if ((!ecore_rd(p_hwfn, p_ptt, TM_REG_PF_SCAN_ACTIVE_CONN)) && (!ecore_rd(p_hwfn, p_ptt, TM_REG_PF_SCAN_ACTIVE_TASK))) break; /* Dependent on number of connection/tasks, possibly * 1ms sleep is required between polls */ OSAL_MSLEEP(1); } if (i < ECORE_HW_STOP_RETRY_LIMIT) return; DP_NOTICE(p_hwfn, true, "Timers linear scans are not over" " [Connection %02x Tasks %02x]\n", (u8)ecore_rd(p_hwfn, p_ptt, TM_REG_PF_SCAN_ACTIVE_CONN), (u8)ecore_rd(p_hwfn, p_ptt, TM_REG_PF_SCAN_ACTIVE_TASK)); } void ecore_hw_timers_stop_all(struct ecore_dev *p_dev) { int j; for_each_hwfn(p_dev, j) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[j]; struct ecore_ptt *p_ptt = p_hwfn->p_main_ptt; ecore_hw_timers_stop(p_dev, p_hwfn, p_ptt); } } static enum _ecore_status_t ecore_verify_reg_val(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 addr, u32 expected_val) { u32 val = ecore_rd(p_hwfn, p_ptt, addr); if (val != expected_val) { DP_NOTICE(p_hwfn, true, "Value at address 0x%08x is 0x%08x while the expected value is 0x%08x\n", addr, val, expected_val); return ECORE_UNKNOWN_ERROR; } return ECORE_SUCCESS; } enum _ecore_status_t ecore_hw_stop(struct ecore_dev *p_dev) { struct ecore_hwfn *p_hwfn; struct ecore_ptt *p_ptt; enum _ecore_status_t rc, rc2 = ECORE_SUCCESS; int j; for_each_hwfn(p_dev, j) { p_hwfn = &p_dev->hwfns[j]; p_ptt = p_hwfn->p_main_ptt; DP_VERBOSE(p_hwfn, ECORE_MSG_IFDOWN, "Stopping hw/fw\n"); if (IS_VF(p_dev)) { ecore_vf_pf_int_cleanup(p_hwfn); rc = ecore_vf_pf_reset(p_hwfn); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "ecore_vf_pf_reset failed. rc = %d.\n", rc); rc2 = ECORE_UNKNOWN_ERROR; } continue; } /* mark the hw as uninitialized... */ p_hwfn->hw_init_done = false; /* Send unload command to MCP */ if (!p_dev->recov_in_prog) { rc = ecore_mcp_unload_req(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed sending a UNLOAD_REQ command. rc = %d.\n", rc); rc2 = ECORE_UNKNOWN_ERROR; } } OSAL_DPC_SYNC(p_hwfn); /* After this point no MFW attentions are expected, e.g. prevent * race between pf stop and dcbx pf update. */ rc = ecore_sp_pf_stop(p_hwfn); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed to close PF against FW [rc = %d]. Continue to stop HW to prevent illegal host access by the device.\n", rc); rc2 = ECORE_UNKNOWN_ERROR; } /* perform debug action after PF stop was sent */ OSAL_AFTER_PF_STOP((void *)p_dev, p_hwfn->my_id); /* close NIG to BRB gate */ ecore_wr(p_hwfn, p_ptt, NIG_REG_RX_LLH_BRB_GATE_DNTFWD_PERPF, 0x1); /* close parser */ ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_TCP, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_UDP, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_FCOE, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_ROCE, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_OPENFLOW, 0x0); /* @@@TBD - clean transmission queues (5.b) */ /* @@@TBD - clean BTB (5.c) */ ecore_hw_timers_stop(p_dev, p_hwfn, p_ptt); /* @@@TBD - verify DMAE requests are done (8) */ /* Disable Attention Generation */ ecore_int_igu_disable_int(p_hwfn, p_ptt); ecore_wr(p_hwfn, p_ptt, IGU_REG_LEADING_EDGE_LATCH, 0); ecore_wr(p_hwfn, p_ptt, IGU_REG_TRAILING_EDGE_LATCH, 0); ecore_int_igu_init_pure_rt(p_hwfn, p_ptt, false, true); rc = ecore_int_igu_reset_cam_default(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed to return IGU CAM to default\n"); rc2 = ECORE_UNKNOWN_ERROR; } /* Need to wait 1ms to guarantee SBs are cleared */ OSAL_MSLEEP(1); if (!p_dev->recov_in_prog) { ecore_verify_reg_val(p_hwfn, p_ptt, QM_REG_USG_CNT_PF_TX, 0); ecore_verify_reg_val(p_hwfn, p_ptt, QM_REG_USG_CNT_PF_OTHER, 0); /* @@@TBD - assert on incorrect xCFC values (10.b) */ } /* Disable PF in HW blocks */ ecore_wr(p_hwfn, p_ptt, DORQ_REG_PF_DB_ENABLE, 0); ecore_wr(p_hwfn, p_ptt, QM_REG_PF_EN, 0); if (!p_dev->recov_in_prog) { ecore_mcp_unload_done(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "Failed sending a UNLOAD_DONE command. rc = %d.\n", rc); rc2 = ECORE_UNKNOWN_ERROR; } } } /* hwfn loop */ if (IS_PF(p_dev) && !p_dev->recov_in_prog) { p_hwfn = ECORE_LEADING_HWFN(p_dev); p_ptt = ECORE_LEADING_HWFN(p_dev)->p_main_ptt; /* Clear the PF's internal FID_enable in the PXP. * In CMT this should only be done for first hw-function, and * only after all transactions have stopped for all active * hw-functions. */ rc = ecore_pglueb_set_pfid_enable(p_hwfn, p_hwfn->p_main_ptt, false); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "ecore_pglueb_set_pfid_enable() failed. rc = %d.\n", rc); rc2 = ECORE_UNKNOWN_ERROR; } } return rc2; } enum _ecore_status_t ecore_hw_stop_fastpath(struct ecore_dev *p_dev) { int j; for_each_hwfn(p_dev, j) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[j]; struct ecore_ptt *p_ptt; if (IS_VF(p_dev)) { ecore_vf_pf_int_cleanup(p_hwfn); continue; } p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_AGAIN; DP_VERBOSE(p_hwfn, ECORE_MSG_IFDOWN, "Shutting down the fastpath\n"); ecore_wr(p_hwfn, p_ptt, NIG_REG_RX_LLH_BRB_GATE_DNTFWD_PERPF, 0x1); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_TCP, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_UDP, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_FCOE, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_ROCE, 0x0); ecore_wr(p_hwfn, p_ptt, PRS_REG_SEARCH_OPENFLOW, 0x0); /* @@@TBD - clean transmission queues (5.b) */ /* @@@TBD - clean BTB (5.c) */ /* @@@TBD - verify DMAE requests are done (8) */ ecore_int_igu_init_pure_rt(p_hwfn, p_ptt, false, false); /* Need to wait 1ms to guarantee SBs are cleared */ OSAL_MSLEEP(1); ecore_ptt_release(p_hwfn, p_ptt); } return ECORE_SUCCESS; } enum _ecore_status_t ecore_hw_start_fastpath(struct ecore_hwfn *p_hwfn) { struct ecore_ptt *p_ptt; if (IS_VF(p_hwfn->p_dev)) return ECORE_SUCCESS; p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_AGAIN; /* If roce info is allocated it means roce is initialized and should * be enabled in searcher. */ if (p_hwfn->p_rdma_info) { if (p_hwfn->b_rdma_enabled_in_prs) ecore_wr(p_hwfn, p_ptt, p_hwfn->rdma_prs_search_reg, 0x1); ecore_wr(p_hwfn, p_ptt, TM_REG_PF_ENABLE_CONN, 0x1); } /* Re-open incoming traffic */ ecore_wr(p_hwfn, p_ptt, NIG_REG_RX_LLH_BRB_GATE_DNTFWD_PERPF, 0x0); ecore_ptt_release(p_hwfn, p_ptt); return ECORE_SUCCESS; } /* Free hwfn memory and resources acquired in hw_hwfn_prepare */ static void ecore_hw_hwfn_free(struct ecore_hwfn *p_hwfn) { ecore_ptt_pool_free(p_hwfn); OSAL_FREE(p_hwfn->p_dev, p_hwfn->hw_info.p_igu_info); } /* Setup bar access */ static void ecore_hw_hwfn_prepare(struct ecore_hwfn *p_hwfn) { /* clear indirect access */ if (ECORE_IS_AH(p_hwfn->p_dev)) { ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_E8_F0_K2_E5, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_EC_F0_K2_E5, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_F0_F0_K2_E5, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_F4_F0_K2_E5, 0); } else { ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_88_F0_BB, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_8C_F0_BB, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_90_F0_BB, 0); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_PGL_ADDR_94_F0_BB, 0); } /* Clean previous pglue_b errors if such exist */ ecore_pglueb_clear_err(p_hwfn, p_hwfn->p_main_ptt); /* enable internal target-read */ ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_INTERNAL_PFID_ENABLE_TARGET_READ, 1); } static void get_function_id(struct ecore_hwfn *p_hwfn) { /* ME Register */ p_hwfn->hw_info.opaque_fid = (u16)REG_RD(p_hwfn, PXP_PF_ME_OPAQUE_ADDR); p_hwfn->hw_info.concrete_fid = REG_RD(p_hwfn, PXP_PF_ME_CONCRETE_ADDR); /* Bits 16-19 from the ME registers are the pf_num */ p_hwfn->abs_pf_id = (p_hwfn->hw_info.concrete_fid >> 16) & 0xf; p_hwfn->rel_pf_id = GET_FIELD(p_hwfn->hw_info.concrete_fid, PXP_CONCRETE_FID_PFID); p_hwfn->port_id = GET_FIELD(p_hwfn->hw_info.concrete_fid, PXP_CONCRETE_FID_PORT); DP_VERBOSE(p_hwfn, ECORE_MSG_PROBE, "Read ME register: Concrete 0x%08x Opaque 0x%04x\n", p_hwfn->hw_info.concrete_fid, p_hwfn->hw_info.opaque_fid); } static void ecore_hw_set_feat(struct ecore_hwfn *p_hwfn) { u32 *feat_num = p_hwfn->hw_info.feat_num; struct ecore_sb_cnt_info sb_cnt; u32 non_l2_sbs = 0; OSAL_MEM_ZERO(&sb_cnt, sizeof(sb_cnt)); ecore_int_get_num_sbs(p_hwfn, &sb_cnt); /* L2 Queues require each: 1 status block. 1 L2 queue */ if (ECORE_IS_L2_PERSONALITY(p_hwfn)) { /* Start by allocating VF queues, then PF's */ feat_num[ECORE_VF_L2_QUE] = OSAL_MIN_T(u32, RESC_NUM(p_hwfn, ECORE_L2_QUEUE), sb_cnt.iov_cnt); feat_num[ECORE_PF_L2_QUE] = OSAL_MIN_T(u32, sb_cnt.cnt - non_l2_sbs, RESC_NUM(p_hwfn, ECORE_L2_QUEUE) - FEAT_NUM(p_hwfn, ECORE_VF_L2_QUE)); } if (ECORE_IS_FCOE_PERSONALITY(p_hwfn)) feat_num[ECORE_FCOE_CQ] = OSAL_MIN_T(u32, sb_cnt.cnt, RESC_NUM(p_hwfn, ECORE_CMDQS_CQS)); if (ECORE_IS_ISCSI_PERSONALITY(p_hwfn)) feat_num[ECORE_ISCSI_CQ] = OSAL_MIN_T(u32, sb_cnt.cnt, RESC_NUM(p_hwfn, ECORE_CMDQS_CQS)); DP_VERBOSE(p_hwfn, ECORE_MSG_PROBE, "#PF_L2_QUEUE=%d VF_L2_QUEUES=%d #ROCE_CNQ=%d #FCOE_CQ=%d #ISCSI_CQ=%d #SB=%d\n", (int)FEAT_NUM(p_hwfn, ECORE_PF_L2_QUE), (int)FEAT_NUM(p_hwfn, ECORE_VF_L2_QUE), (int)FEAT_NUM(p_hwfn, ECORE_RDMA_CNQ), (int)FEAT_NUM(p_hwfn, ECORE_FCOE_CQ), (int)FEAT_NUM(p_hwfn, ECORE_ISCSI_CQ), (int)sb_cnt.cnt); } const char *ecore_hw_get_resc_name(enum ecore_resources res_id) { switch (res_id) { case ECORE_L2_QUEUE: return "L2_QUEUE"; case ECORE_VPORT: return "VPORT"; case ECORE_RSS_ENG: return "RSS_ENG"; case ECORE_PQ: return "PQ"; case ECORE_RL: return "RL"; case ECORE_MAC: return "MAC"; case ECORE_VLAN: return "VLAN"; case ECORE_RDMA_CNQ_RAM: return "RDMA_CNQ_RAM"; case ECORE_ILT: return "ILT"; case ECORE_LL2_QUEUE: return "LL2_QUEUE"; case ECORE_CMDQS_CQS: return "CMDQS_CQS"; case ECORE_RDMA_STATS_QUEUE: return "RDMA_STATS_QUEUE"; case ECORE_BDQ: return "BDQ"; case ECORE_SB: return "SB"; default: return "UNKNOWN_RESOURCE"; } } static enum _ecore_status_t __ecore_hw_set_soft_resc_size(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, enum ecore_resources res_id, u32 resc_max_val, u32 *p_mcp_resp) { enum _ecore_status_t rc; rc = ecore_mcp_set_resc_max_val(p_hwfn, p_ptt, res_id, resc_max_val, p_mcp_resp); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "MFW response failure for a max value setting of resource %d [%s]\n", res_id, ecore_hw_get_resc_name(res_id)); return rc; } if (*p_mcp_resp != FW_MSG_CODE_RESOURCE_ALLOC_OK) DP_INFO(p_hwfn, "Failed to set the max value of resource %d [%s]. mcp_resp = 0x%08x.\n", res_id, ecore_hw_get_resc_name(res_id), *p_mcp_resp); return ECORE_SUCCESS; } static enum _ecore_status_t ecore_hw_set_soft_resc_size(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { bool b_ah = ECORE_IS_AH(p_hwfn->p_dev); u32 resc_max_val, mcp_resp; u8 res_id; enum _ecore_status_t rc; for (res_id = 0; res_id < ECORE_MAX_RESC; res_id++) { /* @DPDK */ switch (res_id) { case ECORE_LL2_QUEUE: case ECORE_RDMA_CNQ_RAM: case ECORE_RDMA_STATS_QUEUE: case ECORE_BDQ: resc_max_val = 0; break; default: continue; } rc = __ecore_hw_set_soft_resc_size(p_hwfn, p_ptt, res_id, resc_max_val, &mcp_resp); if (rc != ECORE_SUCCESS) return rc; /* There's no point to continue to the next resource if the * command is not supported by the MFW. * We do continue if the command is supported but the resource * is unknown to the MFW. Such a resource will be later * configured with the default allocation values. */ if (mcp_resp == FW_MSG_CODE_UNSUPPORTED) return ECORE_NOTIMPL; } return ECORE_SUCCESS; } static enum _ecore_status_t ecore_hw_get_dflt_resc(struct ecore_hwfn *p_hwfn, enum ecore_resources res_id, u32 *p_resc_num, u32 *p_resc_start) { u8 num_funcs = p_hwfn->num_funcs_on_engine; bool b_ah = ECORE_IS_AH(p_hwfn->p_dev); switch (res_id) { case ECORE_L2_QUEUE: *p_resc_num = (b_ah ? MAX_NUM_L2_QUEUES_K2 : MAX_NUM_L2_QUEUES_BB) / num_funcs; break; case ECORE_VPORT: *p_resc_num = (b_ah ? MAX_NUM_VPORTS_K2 : MAX_NUM_VPORTS_BB) / num_funcs; break; case ECORE_RSS_ENG: *p_resc_num = (b_ah ? ETH_RSS_ENGINE_NUM_K2 : ETH_RSS_ENGINE_NUM_BB) / num_funcs; break; case ECORE_PQ: *p_resc_num = (b_ah ? MAX_QM_TX_QUEUES_K2 : MAX_QM_TX_QUEUES_BB) / num_funcs; break; case ECORE_RL: *p_resc_num = MAX_QM_GLOBAL_RLS / num_funcs; break; case ECORE_MAC: case ECORE_VLAN: /* Each VFC resource can accommodate both a MAC and a VLAN */ *p_resc_num = ETH_NUM_MAC_FILTERS / num_funcs; break; case ECORE_ILT: *p_resc_num = (b_ah ? PXP_NUM_ILT_RECORDS_K2 : PXP_NUM_ILT_RECORDS_BB) / num_funcs; break; case ECORE_LL2_QUEUE: *p_resc_num = MAX_NUM_LL2_RX_QUEUES / num_funcs; break; case ECORE_RDMA_CNQ_RAM: case ECORE_CMDQS_CQS: /* CNQ/CMDQS are the same resource */ /* @DPDK */ *p_resc_num = (NUM_OF_GLOBAL_QUEUES / 2) / num_funcs; break; case ECORE_RDMA_STATS_QUEUE: /* @DPDK */ *p_resc_num = (b_ah ? MAX_NUM_VPORTS_K2 : MAX_NUM_VPORTS_BB) / num_funcs; break; case ECORE_BDQ: /* @DPDK */ *p_resc_num = 0; break; default: break; } switch (res_id) { case ECORE_BDQ: if (!*p_resc_num) *p_resc_start = 0; break; case ECORE_SB: /* Since we want its value to reflect whether MFW supports * the new scheme, have a default of 0. */ *p_resc_num = 0; break; default: *p_resc_start = *p_resc_num * p_hwfn->enabled_func_idx; break; } return ECORE_SUCCESS; } static enum _ecore_status_t __ecore_hw_set_resc_info(struct ecore_hwfn *p_hwfn, enum ecore_resources res_id, bool drv_resc_alloc) { u32 dflt_resc_num = 0, dflt_resc_start = 0; u32 mcp_resp, *p_resc_num, *p_resc_start; enum _ecore_status_t rc; p_resc_num = &RESC_NUM(p_hwfn, res_id); p_resc_start = &RESC_START(p_hwfn, res_id); rc = ecore_hw_get_dflt_resc(p_hwfn, res_id, &dflt_resc_num, &dflt_resc_start); if (rc != ECORE_SUCCESS) { DP_ERR(p_hwfn, "Failed to get default amount for resource %d [%s]\n", res_id, ecore_hw_get_resc_name(res_id)); return rc; } #ifndef ASIC_ONLY if (CHIP_REV_IS_SLOW(p_hwfn->p_dev)) { *p_resc_num = dflt_resc_num; *p_resc_start = dflt_resc_start; goto out; } #endif rc = ecore_mcp_get_resc_info(p_hwfn, p_hwfn->p_main_ptt, res_id, &mcp_resp, p_resc_num, p_resc_start); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, true, "MFW response failure for an allocation request for" " resource %d [%s]\n", res_id, ecore_hw_get_resc_name(res_id)); return rc; } /* Default driver values are applied in the following cases: * - The resource allocation MB command is not supported by the MFW * - There is an internal error in the MFW while processing the request * - The resource ID is unknown to the MFW */ if (mcp_resp != FW_MSG_CODE_RESOURCE_ALLOC_OK) { DP_INFO(p_hwfn, "Failed to receive allocation info for resource %d [%s]." " mcp_resp = 0x%x. Applying default values" " [%d,%d].\n", res_id, ecore_hw_get_resc_name(res_id), mcp_resp, dflt_resc_num, dflt_resc_start); *p_resc_num = dflt_resc_num; *p_resc_start = dflt_resc_start; goto out; } if ((*p_resc_num != dflt_resc_num || *p_resc_start != dflt_resc_start) && res_id != ECORE_SB) { DP_INFO(p_hwfn, "MFW allocation for resource %d [%s] differs from default values [%d,%d vs. %d,%d]%s\n", res_id, ecore_hw_get_resc_name(res_id), *p_resc_num, *p_resc_start, dflt_resc_num, dflt_resc_start, drv_resc_alloc ? " - Applying default values" : ""); if (drv_resc_alloc) { *p_resc_num = dflt_resc_num; *p_resc_start = dflt_resc_start; } } out: return ECORE_SUCCESS; } static enum _ecore_status_t ecore_hw_set_resc_info(struct ecore_hwfn *p_hwfn, bool drv_resc_alloc) { enum _ecore_status_t rc; u8 res_id; for (res_id = 0; res_id < ECORE_MAX_RESC; res_id++) { rc = __ecore_hw_set_resc_info(p_hwfn, res_id, drv_resc_alloc); if (rc != ECORE_SUCCESS) return rc; } return ECORE_SUCCESS; } static enum _ecore_status_t ecore_hw_get_resc(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, bool drv_resc_alloc) { struct ecore_resc_unlock_params resc_unlock_params; struct ecore_resc_lock_params resc_lock_params; bool b_ah = ECORE_IS_AH(p_hwfn->p_dev); u8 res_id; enum _ecore_status_t rc; #ifndef ASIC_ONLY u32 *resc_start = p_hwfn->hw_info.resc_start; u32 *resc_num = p_hwfn->hw_info.resc_num; /* For AH, an equal share of the ILT lines between the maximal number of * PFs is not enough for RoCE. This would be solved by the future * resource allocation scheme, but isn't currently present for * FPGA/emulation. For now we keep a number that is sufficient for RoCE * to work - the BB number of ILT lines divided by its max PFs number. */ u32 roce_min_ilt_lines = PXP_NUM_ILT_RECORDS_BB / MAX_NUM_PFS_BB; #endif /* Setting the max values of the soft resources and the following * resources allocation queries should be atomic. Since several PFs can * run in parallel - a resource lock is needed. * If either the resource lock or resource set value commands are not * supported - skip the the max values setting, release the lock if * needed, and proceed to the queries. Other failures, including a * failure to acquire the lock, will cause this function to fail. * Old drivers that don't acquire the lock can run in parallel, and * their allocation values won't be affected by the updated max values. */ ecore_mcp_resc_lock_default_init(&resc_lock_params, &resc_unlock_params, ECORE_RESC_LOCK_RESC_ALLOC, false); rc = ecore_mcp_resc_lock(p_hwfn, p_ptt, &resc_lock_params); if (rc != ECORE_SUCCESS && rc != ECORE_NOTIMPL) { return rc; } else if (rc == ECORE_NOTIMPL) { DP_INFO(p_hwfn, "Skip the max values setting of the soft resources since the resource lock is not supported by the MFW\n"); } else if (rc == ECORE_SUCCESS && !resc_lock_params.b_granted) { DP_NOTICE(p_hwfn, false, "Failed to acquire the resource lock for the resource allocation commands\n"); rc = ECORE_BUSY; goto unlock_and_exit; } else { rc = ecore_hw_set_soft_resc_size(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS && rc != ECORE_NOTIMPL) { DP_NOTICE(p_hwfn, false, "Failed to set the max values of the soft resources\n"); goto unlock_and_exit; } else if (rc == ECORE_NOTIMPL) { DP_INFO(p_hwfn, "Skip the max values setting of the soft resources since it is not supported by the MFW\n"); rc = ecore_mcp_resc_unlock(p_hwfn, p_ptt, &resc_unlock_params); if (rc != ECORE_SUCCESS) DP_INFO(p_hwfn, "Failed to release the resource lock for the resource allocation commands\n"); } } rc = ecore_hw_set_resc_info(p_hwfn, drv_resc_alloc); if (rc != ECORE_SUCCESS) goto unlock_and_exit; if (resc_lock_params.b_granted && !resc_unlock_params.b_released) { rc = ecore_mcp_resc_unlock(p_hwfn, p_ptt, &resc_unlock_params); if (rc != ECORE_SUCCESS) DP_INFO(p_hwfn, "Failed to release the resource lock for the resource allocation commands\n"); } #ifndef ASIC_ONLY if (CHIP_REV_IS_SLOW(p_hwfn->p_dev)) { /* Reduced build contains less PQs */ if (!(p_hwfn->p_dev->b_is_emul_full)) { resc_num[ECORE_PQ] = 32; resc_start[ECORE_PQ] = resc_num[ECORE_PQ] * p_hwfn->enabled_func_idx; } /* For AH emulation, since we have a possible maximal number of * 16 enabled PFs, in case there are not enough ILT lines - * allocate only first PF as RoCE and have all the other ETH * only with less ILT lines. */ if (!p_hwfn->rel_pf_id && p_hwfn->p_dev->b_is_emul_full) resc_num[ECORE_ILT] = OSAL_MAX_T(u32, resc_num[ECORE_ILT], roce_min_ilt_lines); } /* Correct the common ILT calculation if PF0 has more */ if (CHIP_REV_IS_SLOW(p_hwfn->p_dev) && p_hwfn->p_dev->b_is_emul_full && p_hwfn->rel_pf_id && resc_num[ECORE_ILT] < roce_min_ilt_lines) resc_start[ECORE_ILT] += roce_min_ilt_lines - resc_num[ECORE_ILT]; #endif /* Sanity for ILT */ if ((b_ah && (RESC_END(p_hwfn, ECORE_ILT) > PXP_NUM_ILT_RECORDS_K2)) || (!b_ah && (RESC_END(p_hwfn, ECORE_ILT) > PXP_NUM_ILT_RECORDS_BB))) { DP_NOTICE(p_hwfn, true, "Can't assign ILT pages [%08x,...,%08x]\n", RESC_START(p_hwfn, ECORE_ILT), RESC_END(p_hwfn, ECORE_ILT) - 1); return ECORE_INVAL; } /* This will also learn the number of SBs from MFW */ if (ecore_int_igu_reset_cam(p_hwfn, p_ptt)) return ECORE_INVAL; ecore_hw_set_feat(p_hwfn); DP_VERBOSE(p_hwfn, ECORE_MSG_PROBE, "The numbers for each resource are:\n"); for (res_id = 0; res_id < ECORE_MAX_RESC; res_id++) DP_VERBOSE(p_hwfn, ECORE_MSG_PROBE, "%s = %d start = %d\n", ecore_hw_get_resc_name(res_id), RESC_NUM(p_hwfn, res_id), RESC_START(p_hwfn, res_id)); return ECORE_SUCCESS; unlock_and_exit: if (resc_lock_params.b_granted && !resc_unlock_params.b_released) ecore_mcp_resc_unlock(p_hwfn, p_ptt, &resc_unlock_params); return rc; } static enum _ecore_status_t ecore_hw_get_nvm_info(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, struct ecore_hw_prepare_params *p_params) { u32 nvm_cfg1_offset, mf_mode, addr, generic_cont0, core_cfg, dcbx_mode; u32 port_cfg_addr, link_temp, nvm_cfg_addr, device_capabilities; struct ecore_mcp_link_capabilities *p_caps; struct ecore_mcp_link_params *link; enum _ecore_status_t rc; /* Read global nvm_cfg address */ nvm_cfg_addr = ecore_rd(p_hwfn, p_ptt, MISC_REG_GEN_PURP_CR0); /* Verify MCP has initialized it */ if (!nvm_cfg_addr) { DP_NOTICE(p_hwfn, false, "Shared memory not initialized\n"); if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_NVM; return ECORE_INVAL; } /* Read nvm_cfg1 (Notice this is just offset, and not offsize (TBD) */ nvm_cfg1_offset = ecore_rd(p_hwfn, p_ptt, nvm_cfg_addr + 4); addr = MCP_REG_SCRATCH + nvm_cfg1_offset + OFFSETOF(struct nvm_cfg1, glob) + OFFSETOF(struct nvm_cfg1_glob, core_cfg); core_cfg = ecore_rd(p_hwfn, p_ptt, addr); switch ((core_cfg & NVM_CFG1_GLOB_NETWORK_PORT_MODE_MASK) >> NVM_CFG1_GLOB_NETWORK_PORT_MODE_OFFSET) { case NVM_CFG1_GLOB_NETWORK_PORT_MODE_BB_2X40G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_2X40G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_2X50G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_2X50G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_BB_1X100G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_1X100G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_4X10G_F: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_4X10G_F; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_BB_4X10G_E: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_4X10G_E; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_BB_4X20G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_4X20G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_1X40G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_1X40G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_2X25G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_2X25G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_2X10G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_2X10G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_1X25G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_1X25G; break; case NVM_CFG1_GLOB_NETWORK_PORT_MODE_4X25G: p_hwfn->hw_info.port_mode = ECORE_PORT_MODE_DE_4X25G; break; default: DP_NOTICE(p_hwfn, true, "Unknown port mode in 0x%08x\n", core_cfg); break; } /* Read DCBX configuration */ port_cfg_addr = MCP_REG_SCRATCH + nvm_cfg1_offset + OFFSETOF(struct nvm_cfg1, port[MFW_PORT(p_hwfn)]); dcbx_mode = ecore_rd(p_hwfn, p_ptt, port_cfg_addr + OFFSETOF(struct nvm_cfg1_port, generic_cont0)); dcbx_mode = (dcbx_mode & NVM_CFG1_PORT_DCBX_MODE_MASK) >> NVM_CFG1_PORT_DCBX_MODE_OFFSET; switch (dcbx_mode) { case NVM_CFG1_PORT_DCBX_MODE_DYNAMIC: p_hwfn->hw_info.dcbx_mode = ECORE_DCBX_VERSION_DYNAMIC; break; case NVM_CFG1_PORT_DCBX_MODE_CEE: p_hwfn->hw_info.dcbx_mode = ECORE_DCBX_VERSION_CEE; break; case NVM_CFG1_PORT_DCBX_MODE_IEEE: p_hwfn->hw_info.dcbx_mode = ECORE_DCBX_VERSION_IEEE; break; default: p_hwfn->hw_info.dcbx_mode = ECORE_DCBX_VERSION_DISABLED; } /* Read default link configuration */ link = &p_hwfn->mcp_info->link_input; p_caps = &p_hwfn->mcp_info->link_capabilities; port_cfg_addr = MCP_REG_SCRATCH + nvm_cfg1_offset + OFFSETOF(struct nvm_cfg1, port[MFW_PORT(p_hwfn)]); link_temp = ecore_rd(p_hwfn, p_ptt, port_cfg_addr + OFFSETOF(struct nvm_cfg1_port, speed_cap_mask)); link_temp &= NVM_CFG1_PORT_DRV_SPEED_CAPABILITY_MASK_MASK; link->speed.advertised_speeds = link_temp; p_caps->speed_capabilities = link->speed.advertised_speeds; link_temp = ecore_rd(p_hwfn, p_ptt, port_cfg_addr + OFFSETOF(struct nvm_cfg1_port, link_settings)); switch ((link_temp & NVM_CFG1_PORT_DRV_LINK_SPEED_MASK) >> NVM_CFG1_PORT_DRV_LINK_SPEED_OFFSET) { case NVM_CFG1_PORT_DRV_LINK_SPEED_AUTONEG: link->speed.autoneg = true; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_1G: link->speed.forced_speed = 1000; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_10G: link->speed.forced_speed = 10000; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_25G: link->speed.forced_speed = 25000; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_40G: link->speed.forced_speed = 40000; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_50G: link->speed.forced_speed = 50000; break; case NVM_CFG1_PORT_DRV_LINK_SPEED_BB_100G: link->speed.forced_speed = 100000; break; default: DP_NOTICE(p_hwfn, true, "Unknown Speed in 0x%08x\n", link_temp); } p_caps->default_speed = link->speed.forced_speed; p_caps->default_speed_autoneg = link->speed.autoneg; link_temp &= NVM_CFG1_PORT_DRV_FLOW_CONTROL_MASK; link_temp >>= NVM_CFG1_PORT_DRV_FLOW_CONTROL_OFFSET; link->pause.autoneg = !!(link_temp & NVM_CFG1_PORT_DRV_FLOW_CONTROL_AUTONEG); link->pause.forced_rx = !!(link_temp & NVM_CFG1_PORT_DRV_FLOW_CONTROL_RX); link->pause.forced_tx = !!(link_temp & NVM_CFG1_PORT_DRV_FLOW_CONTROL_TX); link->loopback_mode = 0; if (p_hwfn->mcp_info->capabilities & FW_MB_PARAM_FEATURE_SUPPORT_EEE) { link_temp = ecore_rd(p_hwfn, p_ptt, port_cfg_addr + OFFSETOF(struct nvm_cfg1_port, ext_phy)); link_temp &= NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_MASK; link_temp >>= NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_OFFSET; p_caps->default_eee = ECORE_MCP_EEE_ENABLED; link->eee.enable = true; switch (link_temp) { case NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_DISABLED: p_caps->default_eee = ECORE_MCP_EEE_DISABLED; link->eee.enable = false; break; case NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_BALANCED: p_caps->eee_lpi_timer = EEE_TX_TIMER_USEC_BALANCED_TIME; break; case NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_AGGRESSIVE: p_caps->eee_lpi_timer = EEE_TX_TIMER_USEC_AGGRESSIVE_TIME; break; case NVM_CFG1_PORT_EEE_POWER_SAVING_MODE_LOW_LATENCY: p_caps->eee_lpi_timer = EEE_TX_TIMER_USEC_LATENCY_TIME; break; } link->eee.tx_lpi_timer = p_caps->eee_lpi_timer; link->eee.tx_lpi_enable = link->eee.enable; link->eee.adv_caps = ECORE_EEE_1G_ADV | ECORE_EEE_10G_ADV; } else { p_caps->default_eee = ECORE_MCP_EEE_UNSUPPORTED; } DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Read default link: Speed 0x%08x, Adv. Speed 0x%08x, AN: 0x%02x, PAUSE AN: 0x%02x\n EEE: %02x [%08x usec]", link->speed.forced_speed, link->speed.advertised_speeds, link->speed.autoneg, link->pause.autoneg, p_caps->default_eee, p_caps->eee_lpi_timer); /* Read Multi-function information from shmem */ addr = MCP_REG_SCRATCH + nvm_cfg1_offset + OFFSETOF(struct nvm_cfg1, glob) + OFFSETOF(struct nvm_cfg1_glob, generic_cont0); generic_cont0 = ecore_rd(p_hwfn, p_ptt, addr); mf_mode = (generic_cont0 & NVM_CFG1_GLOB_MF_MODE_MASK) >> NVM_CFG1_GLOB_MF_MODE_OFFSET; switch (mf_mode) { case NVM_CFG1_GLOB_MF_MODE_MF_ALLOWED: p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_OVLAN_CLSS; break; case NVM_CFG1_GLOB_MF_MODE_UFP: p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_OVLAN_CLSS | 1 << ECORE_MF_UFP_SPECIFIC | 1 << ECORE_MF_8021Q_TAGGING; break; case NVM_CFG1_GLOB_MF_MODE_BD: p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_OVLAN_CLSS | 1 << ECORE_MF_LLH_PROTO_CLSS | 1 << ECORE_MF_8021AD_TAGGING; break; case NVM_CFG1_GLOB_MF_MODE_NPAR1_0: p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_LLH_MAC_CLSS | 1 << ECORE_MF_LLH_PROTO_CLSS | 1 << ECORE_MF_LL2_NON_UNICAST | 1 << ECORE_MF_INTER_PF_SWITCH | 1 << ECORE_MF_DISABLE_ARFS; break; case NVM_CFG1_GLOB_MF_MODE_DEFAULT: p_hwfn->p_dev->mf_bits = 1 << ECORE_MF_LLH_MAC_CLSS | 1 << ECORE_MF_LLH_PROTO_CLSS | 1 << ECORE_MF_LL2_NON_UNICAST; if (ECORE_IS_BB(p_hwfn->p_dev)) p_hwfn->p_dev->mf_bits |= 1 << ECORE_MF_NEED_DEF_PF; break; } DP_INFO(p_hwfn, "Multi function mode is 0x%lx\n", p_hwfn->p_dev->mf_bits); if (ECORE_IS_CMT(p_hwfn->p_dev)) p_hwfn->p_dev->mf_bits |= (1 << ECORE_MF_DISABLE_ARFS); /* It's funny since we have another switch, but it's easier * to throw this away in linux this way. Long term, it might be * better to have have getters for needed ECORE_MF_* fields, * convert client code and eliminate this. */ switch (mf_mode) { case NVM_CFG1_GLOB_MF_MODE_MF_ALLOWED: case NVM_CFG1_GLOB_MF_MODE_BD: p_hwfn->p_dev->mf_mode = ECORE_MF_OVLAN; break; case NVM_CFG1_GLOB_MF_MODE_NPAR1_0: p_hwfn->p_dev->mf_mode = ECORE_MF_NPAR; break; case NVM_CFG1_GLOB_MF_MODE_DEFAULT: p_hwfn->p_dev->mf_mode = ECORE_MF_DEFAULT; break; case NVM_CFG1_GLOB_MF_MODE_UFP: p_hwfn->p_dev->mf_mode = ECORE_MF_UFP; break; } /* Read Multi-function information from shmem */ addr = MCP_REG_SCRATCH + nvm_cfg1_offset + OFFSETOF(struct nvm_cfg1, glob) + OFFSETOF(struct nvm_cfg1_glob, device_capabilities); device_capabilities = ecore_rd(p_hwfn, p_ptt, addr); if (device_capabilities & NVM_CFG1_GLOB_DEVICE_CAPABILITIES_ETHERNET) OSAL_SET_BIT(ECORE_DEV_CAP_ETH, &p_hwfn->hw_info.device_capabilities); if (device_capabilities & NVM_CFG1_GLOB_DEVICE_CAPABILITIES_FCOE) OSAL_SET_BIT(ECORE_DEV_CAP_FCOE, &p_hwfn->hw_info.device_capabilities); if (device_capabilities & NVM_CFG1_GLOB_DEVICE_CAPABILITIES_ISCSI) OSAL_SET_BIT(ECORE_DEV_CAP_ISCSI, &p_hwfn->hw_info.device_capabilities); if (device_capabilities & NVM_CFG1_GLOB_DEVICE_CAPABILITIES_ROCE) OSAL_SET_BIT(ECORE_DEV_CAP_ROCE, &p_hwfn->hw_info.device_capabilities); if (device_capabilities & NVM_CFG1_GLOB_DEVICE_CAPABILITIES_IWARP) OSAL_SET_BIT(ECORE_DEV_CAP_IWARP, &p_hwfn->hw_info.device_capabilities); rc = ecore_mcp_fill_shmem_func_info(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS && p_params->b_relaxed_probe) { rc = ECORE_SUCCESS; p_params->p_relaxed_res = ECORE_HW_PREPARE_BAD_MCP; } return rc; } static void ecore_get_num_funcs(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { u8 num_funcs, enabled_func_idx = p_hwfn->rel_pf_id; u32 reg_function_hide, tmp, eng_mask, low_pfs_mask; struct ecore_dev *p_dev = p_hwfn->p_dev; num_funcs = ECORE_IS_AH(p_dev) ? MAX_NUM_PFS_K2 : MAX_NUM_PFS_BB; /* Bit 0 of MISCS_REG_FUNCTION_HIDE indicates whether the bypass values * in the other bits are selected. * Bits 1-15 are for functions 1-15, respectively, and their value is * '0' only for enabled functions (function 0 always exists and * enabled). * In case of CMT in BB, only the "even" functions are enabled, and thus * the number of functions for both hwfns is learnt from the same bits. */ if (ECORE_IS_BB(p_dev) || ECORE_IS_AH(p_dev)) { reg_function_hide = ecore_rd(p_hwfn, p_ptt, MISCS_REG_FUNCTION_HIDE_BB_K2); } else { /* E5 */ reg_function_hide = 0; } if (reg_function_hide & 0x1) { if (ECORE_IS_BB(p_dev)) { if (ECORE_PATH_ID(p_hwfn) && !ECORE_IS_CMT(p_dev)) { num_funcs = 0; eng_mask = 0xaaaa; } else { num_funcs = 1; eng_mask = 0x5554; } } else { num_funcs = 1; eng_mask = 0xfffe; } /* Get the number of the enabled functions on the engine */ tmp = (reg_function_hide ^ 0xffffffff) & eng_mask; while (tmp) { if (tmp & 0x1) num_funcs++; tmp >>= 0x1; } /* Get the PF index within the enabled functions */ low_pfs_mask = (0x1 << p_hwfn->abs_pf_id) - 1; tmp = reg_function_hide & eng_mask & low_pfs_mask; while (tmp) { if (tmp & 0x1) enabled_func_idx--; tmp >>= 0x1; } } p_hwfn->num_funcs_on_engine = num_funcs; p_hwfn->enabled_func_idx = enabled_func_idx; #ifndef ASIC_ONLY if (CHIP_REV_IS_FPGA(p_dev)) { DP_NOTICE(p_hwfn, false, "FPGA: Limit number of PFs to 4 [would affect resource allocation, needed for IOV]\n"); p_hwfn->num_funcs_on_engine = 4; } #endif DP_VERBOSE(p_hwfn, ECORE_MSG_PROBE, "PF [rel_id %d, abs_id %d] occupies index %d within the %d enabled functions on the engine\n", p_hwfn->rel_pf_id, p_hwfn->abs_pf_id, p_hwfn->enabled_func_idx, p_hwfn->num_funcs_on_engine); } static void ecore_hw_info_port_num_bb(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_dev *p_dev = p_hwfn->p_dev; u32 port_mode; #ifndef ASIC_ONLY /* Read the port mode */ if (CHIP_REV_IS_FPGA(p_dev)) port_mode = 4; else if (CHIP_REV_IS_EMUL(p_dev) && ECORE_IS_CMT(p_dev)) /* In CMT on emulation, assume 1 port */ port_mode = 1; else #endif port_mode = ecore_rd(p_hwfn, p_ptt, CNIG_REG_NW_PORT_MODE_BB); if (port_mode < 3) { p_dev->num_ports_in_engine = 1; } else if (port_mode <= 5) { p_dev->num_ports_in_engine = 2; } else { DP_NOTICE(p_hwfn, true, "PORT MODE: %d not supported\n", p_dev->num_ports_in_engine); /* Default num_ports_in_engine to something */ p_dev->num_ports_in_engine = 1; } } static void ecore_hw_info_port_num_ah_e5(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_dev *p_dev = p_hwfn->p_dev; u32 port; int i; p_dev->num_ports_in_engine = 0; #ifndef ASIC_ONLY if (CHIP_REV_IS_EMUL(p_dev)) { port = ecore_rd(p_hwfn, p_ptt, MISCS_REG_ECO_RESERVED); switch ((port & 0xf000) >> 12) { case 1: p_dev->num_ports_in_engine = 1; break; case 3: p_dev->num_ports_in_engine = 2; break; case 0xf: p_dev->num_ports_in_engine = 4; break; default: DP_NOTICE(p_hwfn, false, "Unknown port mode in ECO_RESERVED %08x\n", port); } } else #endif for (i = 0; i < MAX_NUM_PORTS_K2; i++) { port = ecore_rd(p_hwfn, p_ptt, CNIG_REG_NIG_PORT0_CONF_K2_E5 + (i * 4)); if (port & 1) p_dev->num_ports_in_engine++; } if (!p_dev->num_ports_in_engine) { DP_NOTICE(p_hwfn, true, "All NIG ports are inactive\n"); /* Default num_ports_in_engine to something */ p_dev->num_ports_in_engine = 1; } } static void ecore_hw_info_port_num(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_dev *p_dev = p_hwfn->p_dev; /* Determine the number of ports per engine */ if (ECORE_IS_BB(p_dev)) ecore_hw_info_port_num_bb(p_hwfn, p_ptt); else ecore_hw_info_port_num_ah_e5(p_hwfn, p_ptt); /* Get the total number of ports of the device */ if (ECORE_IS_CMT(p_dev)) { /* In CMT there is always only one port */ p_dev->num_ports = 1; #ifndef ASIC_ONLY } else if (CHIP_REV_IS_EMUL(p_dev) || CHIP_REV_IS_TEDIBEAR(p_dev)) { p_dev->num_ports = p_dev->num_ports_in_engine * ecore_device_num_engines(p_dev); #endif } else { u32 addr, global_offsize, global_addr; addr = SECTION_OFFSIZE_ADDR(p_hwfn->mcp_info->public_base, PUBLIC_GLOBAL); global_offsize = ecore_rd(p_hwfn, p_ptt, addr); global_addr = SECTION_ADDR(global_offsize, 0); addr = global_addr + OFFSETOF(struct public_global, max_ports); p_dev->num_ports = (u8)ecore_rd(p_hwfn, p_ptt, addr); } } static void ecore_mcp_get_eee_caps(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_mcp_link_capabilities *p_caps; u32 eee_status; p_caps = &p_hwfn->mcp_info->link_capabilities; if (p_caps->default_eee == ECORE_MCP_EEE_UNSUPPORTED) return; p_caps->eee_speed_caps = 0; eee_status = ecore_rd(p_hwfn, p_ptt, p_hwfn->mcp_info->port_addr + OFFSETOF(struct public_port, eee_status)); eee_status = (eee_status & EEE_SUPPORTED_SPEED_MASK) >> EEE_SUPPORTED_SPEED_OFFSET; if (eee_status & EEE_1G_SUPPORTED) p_caps->eee_speed_caps |= ECORE_EEE_1G_ADV; if (eee_status & EEE_10G_ADV) p_caps->eee_speed_caps |= ECORE_EEE_10G_ADV; } static enum _ecore_status_t ecore_get_hw_info(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, enum ecore_pci_personality personality, struct ecore_hw_prepare_params *p_params) { bool drv_resc_alloc = p_params->drv_resc_alloc; enum _ecore_status_t rc; /* Since all information is common, only first hwfns should do this */ if (IS_LEAD_HWFN(p_hwfn)) { rc = ecore_iov_hw_info(p_hwfn); if (rc != ECORE_SUCCESS) { if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_BAD_IOV; else return rc; } } if (IS_LEAD_HWFN(p_hwfn)) ecore_hw_info_port_num(p_hwfn, p_ptt); ecore_mcp_get_capabilities(p_hwfn, p_ptt); #ifndef ASIC_ONLY if (CHIP_REV_IS_ASIC(p_hwfn->p_dev)) { #endif rc = ecore_hw_get_nvm_info(p_hwfn, p_ptt, p_params); if (rc != ECORE_SUCCESS) return rc; #ifndef ASIC_ONLY } #endif rc = ecore_int_igu_read_cam(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) { if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_BAD_IGU; else return rc; } #ifndef ASIC_ONLY if (CHIP_REV_IS_ASIC(p_hwfn->p_dev) && ecore_mcp_is_init(p_hwfn)) { #endif OSAL_MEMCPY(p_hwfn->hw_info.hw_mac_addr, p_hwfn->mcp_info->func_info.mac, ETH_ALEN); #ifndef ASIC_ONLY } else { static u8 mcp_hw_mac[6] = { 0, 2, 3, 4, 5, 6 }; OSAL_MEMCPY(p_hwfn->hw_info.hw_mac_addr, mcp_hw_mac, ETH_ALEN); p_hwfn->hw_info.hw_mac_addr[5] = p_hwfn->abs_pf_id; } #endif if (ecore_mcp_is_init(p_hwfn)) { if (p_hwfn->mcp_info->func_info.ovlan != ECORE_MCP_VLAN_UNSET) p_hwfn->hw_info.ovlan = p_hwfn->mcp_info->func_info.ovlan; ecore_mcp_cmd_port_init(p_hwfn, p_ptt); ecore_mcp_get_eee_caps(p_hwfn, p_ptt); ecore_mcp_read_ufp_config(p_hwfn, p_ptt); } if (personality != ECORE_PCI_DEFAULT) { p_hwfn->hw_info.personality = personality; } else if (ecore_mcp_is_init(p_hwfn)) { enum ecore_pci_personality protocol; protocol = p_hwfn->mcp_info->func_info.protocol; p_hwfn->hw_info.personality = protocol; } #ifndef ASIC_ONLY /* To overcome ILT lack for emulation, until at least until we'll have * a definite answer from system about it, allow only PF0 to be RoCE. */ if (CHIP_REV_IS_EMUL(p_hwfn->p_dev) && ECORE_IS_AH(p_hwfn->p_dev)) { if (!p_hwfn->rel_pf_id) p_hwfn->hw_info.personality = ECORE_PCI_ETH_ROCE; else p_hwfn->hw_info.personality = ECORE_PCI_ETH; } #endif /* although in BB some constellations may support more than 4 tcs, * that can result in performance penalty in some cases. 4 * represents a good tradeoff between performance and flexibility. */ p_hwfn->hw_info.num_hw_tc = NUM_PHYS_TCS_4PORT_K2; /* start out with a single active tc. This can be increased either * by dcbx negotiation or by upper layer driver */ p_hwfn->hw_info.num_active_tc = 1; ecore_get_num_funcs(p_hwfn, p_ptt); if (ecore_mcp_is_init(p_hwfn)) p_hwfn->hw_info.mtu = p_hwfn->mcp_info->func_info.mtu; /* In case of forcing the driver's default resource allocation, calling * ecore_hw_get_resc() should come after initializing the personality * and after getting the number of functions, since the calculation of * the resources/features depends on them. * This order is not harmful if not forcing. */ rc = ecore_hw_get_resc(p_hwfn, p_ptt, drv_resc_alloc); if (rc != ECORE_SUCCESS && p_params->b_relaxed_probe) { rc = ECORE_SUCCESS; p_params->p_relaxed_res = ECORE_HW_PREPARE_BAD_MCP; } return rc; } static enum _ecore_status_t ecore_get_dev_info(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_dev *p_dev = p_hwfn->p_dev; u16 device_id_mask; u32 tmp; /* Read Vendor Id / Device Id */ OSAL_PCI_READ_CONFIG_WORD(p_dev, PCICFG_VENDOR_ID_OFFSET, &p_dev->vendor_id); OSAL_PCI_READ_CONFIG_WORD(p_dev, PCICFG_DEVICE_ID_OFFSET, &p_dev->device_id); /* Determine type */ device_id_mask = p_dev->device_id & ECORE_DEV_ID_MASK; switch (device_id_mask) { case ECORE_DEV_ID_MASK_BB: p_dev->type = ECORE_DEV_TYPE_BB; break; case ECORE_DEV_ID_MASK_AH: p_dev->type = ECORE_DEV_TYPE_AH; break; default: DP_NOTICE(p_hwfn, true, "Unknown device id 0x%x\n", p_dev->device_id); return ECORE_ABORTED; } tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_CHIP_NUM); p_dev->chip_num = (u16)GET_FIELD(tmp, CHIP_NUM); tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_CHIP_REV); p_dev->chip_rev = (u8)GET_FIELD(tmp, CHIP_REV); /* Learn number of HW-functions */ tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_CMT_ENABLED_FOR_PAIR); if (tmp & (1 << p_hwfn->rel_pf_id)) { DP_NOTICE(p_dev->hwfns, false, "device in CMT mode\n"); p_dev->num_hwfns = 2; } else { p_dev->num_hwfns = 1; } #ifndef ASIC_ONLY if (CHIP_REV_IS_EMUL(p_dev)) { /* For some reason we have problems with this register * in B0 emulation; Simply assume no CMT */ DP_NOTICE(p_dev->hwfns, false, "device on emul - assume no CMT\n"); p_dev->num_hwfns = 1; } #endif tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_CHIP_TEST_REG); p_dev->chip_bond_id = (u8)GET_FIELD(tmp, CHIP_BOND_ID); tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_CHIP_METAL); p_dev->chip_metal = (u8)GET_FIELD(tmp, CHIP_METAL); DP_INFO(p_dev->hwfns, "Chip details - %s %c%d, Num: %04x Rev: %02x Bond id: %02x Metal: %02x\n", ECORE_IS_BB(p_dev) ? "BB" : "AH", 'A' + p_dev->chip_rev, (int)p_dev->chip_metal, p_dev->chip_num, p_dev->chip_rev, p_dev->chip_bond_id, p_dev->chip_metal); if (ECORE_IS_BB_A0(p_dev)) { DP_NOTICE(p_dev->hwfns, false, "The chip type/rev (BB A0) is not supported!\n"); return ECORE_ABORTED; } #ifndef ASIC_ONLY if (CHIP_REV_IS_EMUL(p_dev) && ECORE_IS_AH(p_dev)) ecore_wr(p_hwfn, p_ptt, MISCS_REG_PLL_MAIN_CTRL_4, 0x1); if (CHIP_REV_IS_EMUL(p_dev)) { tmp = ecore_rd(p_hwfn, p_ptt, MISCS_REG_ECO_RESERVED); if (tmp & (1 << 29)) { DP_NOTICE(p_hwfn, false, "Emulation: Running on a FULL build\n"); p_dev->b_is_emul_full = true; } else { DP_NOTICE(p_hwfn, false, "Emulation: Running on a REDUCED build\n"); } } #endif return ECORE_SUCCESS; } #ifndef LINUX_REMOVE void ecore_prepare_hibernate(struct ecore_dev *p_dev) { int j; if (IS_VF(p_dev)) return; for_each_hwfn(p_dev, j) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[j]; DP_VERBOSE(p_hwfn, ECORE_MSG_IFDOWN, "Mark hw/fw uninitialized\n"); p_hwfn->hw_init_done = false; ecore_ptt_invalidate(p_hwfn); } } #endif static enum _ecore_status_t ecore_hw_prepare_single(struct ecore_hwfn *p_hwfn, void OSAL_IOMEM * p_regview, void OSAL_IOMEM * p_doorbells, struct ecore_hw_prepare_params *p_params) { struct ecore_mdump_retain_data mdump_retain; struct ecore_dev *p_dev = p_hwfn->p_dev; struct ecore_mdump_info mdump_info; enum _ecore_status_t rc = ECORE_SUCCESS; /* Split PCI bars evenly between hwfns */ p_hwfn->regview = p_regview; p_hwfn->doorbells = p_doorbells; if (IS_VF(p_dev)) return ecore_vf_hw_prepare(p_hwfn); /* Validate that chip access is feasible */ if (REG_RD(p_hwfn, PXP_PF_ME_OPAQUE_ADDR) == 0xffffffff) { DP_ERR(p_hwfn, "Reading the ME register returns all Fs; Preventing further chip access\n"); if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_ME; return ECORE_INVAL; } get_function_id(p_hwfn); /* Allocate PTT pool */ rc = ecore_ptt_pool_alloc(p_hwfn); if (rc) { DP_NOTICE(p_hwfn, true, "Failed to prepare hwfn's hw\n"); if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_MEM; goto err0; } /* Allocate the main PTT */ p_hwfn->p_main_ptt = ecore_get_reserved_ptt(p_hwfn, RESERVED_PTT_MAIN); /* First hwfn learns basic information, e.g., number of hwfns */ if (!p_hwfn->my_id) { rc = ecore_get_dev_info(p_hwfn, p_hwfn->p_main_ptt); if (rc != ECORE_SUCCESS) { if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_DEV; goto err1; } } ecore_hw_hwfn_prepare(p_hwfn); /* Initialize MCP structure */ rc = ecore_mcp_cmd_init(p_hwfn, p_hwfn->p_main_ptt); if (rc) { DP_NOTICE(p_hwfn, true, "Failed initializing mcp command\n"); if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_MEM; goto err1; } /* Read the device configuration information from the HW and SHMEM */ rc = ecore_get_hw_info(p_hwfn, p_hwfn->p_main_ptt, p_params->personality, p_params); if (rc) { DP_NOTICE(p_hwfn, true, "Failed to get HW information\n"); goto err2; } /* Sending a mailbox to the MFW should be after ecore_get_hw_info() is * called, since among others it sets the ports number in an engine. */ if (p_params->initiate_pf_flr && IS_LEAD_HWFN(p_hwfn) && !p_dev->recov_in_prog) { rc = ecore_mcp_initiate_pf_flr(p_hwfn, p_hwfn->p_main_ptt); if (rc != ECORE_SUCCESS) DP_NOTICE(p_hwfn, false, "Failed to initiate PF FLR\n"); } /* Check if mdump logs/data are present and update the epoch value */ if (IS_LEAD_HWFN(p_hwfn)) { #ifndef ASIC_ONLY if (!CHIP_REV_IS_EMUL(p_dev)) { #endif rc = ecore_mcp_mdump_get_info(p_hwfn, p_hwfn->p_main_ptt, &mdump_info); if (rc == ECORE_SUCCESS && mdump_info.num_of_logs) DP_NOTICE(p_hwfn, false, "* * * IMPORTANT - HW ERROR register dump captured by device * * *\n"); rc = ecore_mcp_mdump_get_retain(p_hwfn, p_hwfn->p_main_ptt, &mdump_retain); if (rc == ECORE_SUCCESS && mdump_retain.valid) DP_NOTICE(p_hwfn, false, "mdump retained data: epoch 0x%08x, pf 0x%x, status 0x%08x\n", mdump_retain.epoch, mdump_retain.pf, mdump_retain.status); ecore_mcp_mdump_set_values(p_hwfn, p_hwfn->p_main_ptt, p_params->epoch); #ifndef ASIC_ONLY } #endif } /* Allocate the init RT array and initialize the init-ops engine */ rc = ecore_init_alloc(p_hwfn); if (rc) { DP_NOTICE(p_hwfn, true, "Failed to allocate the init array\n"); if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_MEM; goto err2; } #ifndef ASIC_ONLY if (CHIP_REV_IS_FPGA(p_dev)) { DP_NOTICE(p_hwfn, false, "FPGA: workaround; Prevent DMAE parities\n"); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PCIE_REG_PRTY_MASK_K2_E5, 7); DP_NOTICE(p_hwfn, false, "FPGA: workaround: Set VF bar0 size\n"); ecore_wr(p_hwfn, p_hwfn->p_main_ptt, PGLUE_B_REG_VF_BAR0_SIZE_K2_E5, 4); } #endif return rc; err2: if (IS_LEAD_HWFN(p_hwfn)) ecore_iov_free_hw_info(p_dev); ecore_mcp_free(p_hwfn); err1: ecore_hw_hwfn_free(p_hwfn); err0: return rc; } enum _ecore_status_t ecore_hw_prepare(struct ecore_dev *p_dev, struct ecore_hw_prepare_params *p_params) { struct ecore_hwfn *p_hwfn = ECORE_LEADING_HWFN(p_dev); enum _ecore_status_t rc; p_dev->chk_reg_fifo = p_params->chk_reg_fifo; p_dev->allow_mdump = p_params->allow_mdump; if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_SUCCESS; /* Store the precompiled init data ptrs */ if (IS_PF(p_dev)) ecore_init_iro_array(p_dev); /* Initialize the first hwfn - will learn number of hwfns */ rc = ecore_hw_prepare_single(p_hwfn, p_dev->regview, p_dev->doorbells, p_params); if (rc != ECORE_SUCCESS) return rc; p_params->personality = p_hwfn->hw_info.personality; /* initilalize 2nd hwfn if necessary */ if (ECORE_IS_CMT(p_dev)) { void OSAL_IOMEM *p_regview, *p_doorbell; u8 OSAL_IOMEM *addr; /* adjust bar offset for second engine */ addr = (u8 OSAL_IOMEM *)p_dev->regview + ecore_hw_bar_size(p_hwfn, p_hwfn->p_main_ptt, BAR_ID_0) / 2; p_regview = (void OSAL_IOMEM *)addr; addr = (u8 OSAL_IOMEM *)p_dev->doorbells + ecore_hw_bar_size(p_hwfn, p_hwfn->p_main_ptt, BAR_ID_1) / 2; p_doorbell = (void OSAL_IOMEM *)addr; /* prepare second hw function */ rc = ecore_hw_prepare_single(&p_dev->hwfns[1], p_regview, p_doorbell, p_params); /* in case of error, need to free the previously * initiliazed hwfn 0. */ if (rc != ECORE_SUCCESS) { if (p_params->b_relaxed_probe) p_params->p_relaxed_res = ECORE_HW_PREPARE_FAILED_ENG2; if (IS_PF(p_dev)) { ecore_init_free(p_hwfn); ecore_mcp_free(p_hwfn); ecore_hw_hwfn_free(p_hwfn); } else { DP_NOTICE(p_dev, true, "What do we need to free when VF hwfn1 init fails\n"); } return rc; } } return rc; } void ecore_hw_remove(struct ecore_dev *p_dev) { struct ecore_hwfn *p_hwfn = ECORE_LEADING_HWFN(p_dev); int i; if (IS_PF(p_dev)) ecore_mcp_ov_update_driver_state(p_hwfn, p_hwfn->p_main_ptt, ECORE_OV_DRIVER_STATE_NOT_LOADED); for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; if (IS_VF(p_dev)) { ecore_vf_pf_release(p_hwfn); continue; } ecore_init_free(p_hwfn); ecore_hw_hwfn_free(p_hwfn); ecore_mcp_free(p_hwfn); #ifdef CONFIG_ECORE_LOCK_ALLOC OSAL_MUTEX_DEALLOC(&p_hwfn->dmae_info.mutex); #endif } ecore_iov_free_hw_info(p_dev); } static void ecore_chain_free_next_ptr(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { void *p_virt = p_chain->p_virt_addr, *p_virt_next = OSAL_NULL; dma_addr_t p_phys = p_chain->p_phys_addr, p_phys_next = 0; struct ecore_chain_next *p_next; u32 size, i; if (!p_virt) return; size = p_chain->elem_size * p_chain->usable_per_page; for (i = 0; i < p_chain->page_cnt; i++) { if (!p_virt) break; p_next = (struct ecore_chain_next *)((u8 *)p_virt + size); p_virt_next = p_next->next_virt; p_phys_next = HILO_DMA_REGPAIR(p_next->next_phys); OSAL_DMA_FREE_COHERENT(p_dev, p_virt, p_phys, ECORE_CHAIN_PAGE_SIZE); p_virt = p_virt_next; p_phys = p_phys_next; } } static void ecore_chain_free_single(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { if (!p_chain->p_virt_addr) return; OSAL_DMA_FREE_COHERENT(p_dev, p_chain->p_virt_addr, p_chain->p_phys_addr, ECORE_CHAIN_PAGE_SIZE); } static void ecore_chain_free_pbl(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { void **pp_virt_addr_tbl = p_chain->pbl.pp_virt_addr_tbl; u8 *p_pbl_virt = (u8 *)p_chain->pbl_sp.p_virt_table; u32 page_cnt = p_chain->page_cnt, i, pbl_size; if (!pp_virt_addr_tbl) return; if (!p_pbl_virt) goto out; for (i = 0; i < page_cnt; i++) { if (!pp_virt_addr_tbl[i]) break; OSAL_DMA_FREE_COHERENT(p_dev, pp_virt_addr_tbl[i], *(dma_addr_t *)p_pbl_virt, ECORE_CHAIN_PAGE_SIZE); p_pbl_virt += ECORE_CHAIN_PBL_ENTRY_SIZE; } pbl_size = page_cnt * ECORE_CHAIN_PBL_ENTRY_SIZE; if (!p_chain->b_external_pbl) OSAL_DMA_FREE_COHERENT(p_dev, p_chain->pbl_sp.p_virt_table, p_chain->pbl_sp.p_phys_table, pbl_size); out: OSAL_VFREE(p_dev, p_chain->pbl.pp_virt_addr_tbl); } void ecore_chain_free(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { switch (p_chain->mode) { case ECORE_CHAIN_MODE_NEXT_PTR: ecore_chain_free_next_ptr(p_dev, p_chain); break; case ECORE_CHAIN_MODE_SINGLE: ecore_chain_free_single(p_dev, p_chain); break; case ECORE_CHAIN_MODE_PBL: ecore_chain_free_pbl(p_dev, p_chain); break; } } static enum _ecore_status_t ecore_chain_alloc_sanity_check(struct ecore_dev *p_dev, enum ecore_chain_cnt_type cnt_type, osal_size_t elem_size, u32 page_cnt) { u64 chain_size = ELEMS_PER_PAGE(elem_size) * page_cnt; /* The actual chain size can be larger than the maximal possible value * after rounding up the requested elements number to pages, and after * taking into acount the unusuable elements (next-ptr elements). * The size of a "u16" chain can be (U16_MAX + 1) since the chain * size/capacity fields are of a u32 type. */ if ((cnt_type == ECORE_CHAIN_CNT_TYPE_U16 && chain_size > ((u32)ECORE_U16_MAX + 1)) || (cnt_type == ECORE_CHAIN_CNT_TYPE_U32 && chain_size > ECORE_U32_MAX)) { DP_NOTICE(p_dev, true, "The actual chain size (0x%lx) is larger than the maximal possible value\n", (unsigned long)chain_size); return ECORE_INVAL; } return ECORE_SUCCESS; } static enum _ecore_status_t ecore_chain_alloc_next_ptr(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { void *p_virt = OSAL_NULL, *p_virt_prev = OSAL_NULL; dma_addr_t p_phys = 0; u32 i; for (i = 0; i < p_chain->page_cnt; i++) { p_virt = OSAL_DMA_ALLOC_COHERENT(p_dev, &p_phys, ECORE_CHAIN_PAGE_SIZE); if (!p_virt) { DP_NOTICE(p_dev, true, "Failed to allocate chain memory\n"); return ECORE_NOMEM; } if (i == 0) { ecore_chain_init_mem(p_chain, p_virt, p_phys); ecore_chain_reset(p_chain); } else { ecore_chain_init_next_ptr_elem(p_chain, p_virt_prev, p_virt, p_phys); } p_virt_prev = p_virt; } /* Last page's next element should point to the beginning of the * chain. */ ecore_chain_init_next_ptr_elem(p_chain, p_virt_prev, p_chain->p_virt_addr, p_chain->p_phys_addr); return ECORE_SUCCESS; } static enum _ecore_status_t ecore_chain_alloc_single(struct ecore_dev *p_dev, struct ecore_chain *p_chain) { dma_addr_t p_phys = 0; void *p_virt = OSAL_NULL; p_virt = OSAL_DMA_ALLOC_COHERENT(p_dev, &p_phys, ECORE_CHAIN_PAGE_SIZE); if (!p_virt) { DP_NOTICE(p_dev, true, "Failed to allocate chain memory\n"); return ECORE_NOMEM; } ecore_chain_init_mem(p_chain, p_virt, p_phys); ecore_chain_reset(p_chain); return ECORE_SUCCESS; } static enum _ecore_status_t ecore_chain_alloc_pbl(struct ecore_dev *p_dev, struct ecore_chain *p_chain, struct ecore_chain_ext_pbl *ext_pbl) { u32 page_cnt = p_chain->page_cnt, size, i; dma_addr_t p_phys = 0, p_pbl_phys = 0; void **pp_virt_addr_tbl = OSAL_NULL; u8 *p_pbl_virt = OSAL_NULL; void *p_virt = OSAL_NULL; size = page_cnt * sizeof(*pp_virt_addr_tbl); pp_virt_addr_tbl = (void **)OSAL_VZALLOC(p_dev, size); if (!pp_virt_addr_tbl) { DP_NOTICE(p_dev, true, "Failed to allocate memory for the chain virtual addresses table\n"); return ECORE_NOMEM; } /* The allocation of the PBL table is done with its full size, since it * is expected to be successive. * ecore_chain_init_pbl_mem() is called even in a case of an allocation * failure, since pp_virt_addr_tbl was previously allocated, and it * should be saved to allow its freeing during the error flow. */ size = page_cnt * ECORE_CHAIN_PBL_ENTRY_SIZE; if (ext_pbl == OSAL_NULL) { p_pbl_virt = OSAL_DMA_ALLOC_COHERENT(p_dev, &p_pbl_phys, size); } else { p_pbl_virt = ext_pbl->p_pbl_virt; p_pbl_phys = ext_pbl->p_pbl_phys; p_chain->b_external_pbl = true; } ecore_chain_init_pbl_mem(p_chain, p_pbl_virt, p_pbl_phys, pp_virt_addr_tbl); if (!p_pbl_virt) { DP_NOTICE(p_dev, true, "Failed to allocate chain pbl memory\n"); return ECORE_NOMEM; } for (i = 0; i < page_cnt; i++) { p_virt = OSAL_DMA_ALLOC_COHERENT(p_dev, &p_phys, ECORE_CHAIN_PAGE_SIZE); if (!p_virt) { DP_NOTICE(p_dev, true, "Failed to allocate chain memory\n"); return ECORE_NOMEM; } if (i == 0) { ecore_chain_init_mem(p_chain, p_virt, p_phys); ecore_chain_reset(p_chain); } /* Fill the PBL table with the physical address of the page */ *(dma_addr_t *)p_pbl_virt = p_phys; /* Keep the virtual address of the page */ p_chain->pbl.pp_virt_addr_tbl[i] = p_virt; p_pbl_virt += ECORE_CHAIN_PBL_ENTRY_SIZE; } return ECORE_SUCCESS; } enum _ecore_status_t ecore_chain_alloc(struct ecore_dev *p_dev, enum ecore_chain_use_mode intended_use, enum ecore_chain_mode mode, enum ecore_chain_cnt_type cnt_type, u32 num_elems, osal_size_t elem_size, struct ecore_chain *p_chain, struct ecore_chain_ext_pbl *ext_pbl) { u32 page_cnt; enum _ecore_status_t rc = ECORE_SUCCESS; if (mode == ECORE_CHAIN_MODE_SINGLE) page_cnt = 1; else page_cnt = ECORE_CHAIN_PAGE_CNT(num_elems, elem_size, mode); rc = ecore_chain_alloc_sanity_check(p_dev, cnt_type, elem_size, page_cnt); if (rc) { DP_NOTICE(p_dev, true, "Cannot allocate a chain with the given arguments:\n" "[use_mode %d, mode %d, cnt_type %d, num_elems %d, elem_size %zu]\n", intended_use, mode, cnt_type, num_elems, elem_size); return rc; } ecore_chain_init_params(p_chain, page_cnt, (u8)elem_size, intended_use, mode, cnt_type, p_dev->dp_ctx); switch (mode) { case ECORE_CHAIN_MODE_NEXT_PTR: rc = ecore_chain_alloc_next_ptr(p_dev, p_chain); break; case ECORE_CHAIN_MODE_SINGLE: rc = ecore_chain_alloc_single(p_dev, p_chain); break; case ECORE_CHAIN_MODE_PBL: rc = ecore_chain_alloc_pbl(p_dev, p_chain, ext_pbl); break; } if (rc) goto nomem; return ECORE_SUCCESS; nomem: ecore_chain_free(p_dev, p_chain); return rc; } enum _ecore_status_t ecore_fw_l2_queue(struct ecore_hwfn *p_hwfn, u16 src_id, u16 *dst_id) { if (src_id >= RESC_NUM(p_hwfn, ECORE_L2_QUEUE)) { u16 min, max; min = (u16)RESC_START(p_hwfn, ECORE_L2_QUEUE); max = min + RESC_NUM(p_hwfn, ECORE_L2_QUEUE); DP_NOTICE(p_hwfn, true, "l2_queue id [%d] is not valid, available indices [%d - %d]\n", src_id, min, max); return ECORE_INVAL; } *dst_id = RESC_START(p_hwfn, ECORE_L2_QUEUE) + src_id; return ECORE_SUCCESS; } enum _ecore_status_t ecore_fw_vport(struct ecore_hwfn *p_hwfn, u8 src_id, u8 *dst_id) { if (src_id >= RESC_NUM(p_hwfn, ECORE_VPORT)) { u8 min, max; min = (u8)RESC_START(p_hwfn, ECORE_VPORT); max = min + RESC_NUM(p_hwfn, ECORE_VPORT); DP_NOTICE(p_hwfn, true, "vport id [%d] is not valid, available indices [%d - %d]\n", src_id, min, max); return ECORE_INVAL; } *dst_id = RESC_START(p_hwfn, ECORE_VPORT) + src_id; return ECORE_SUCCESS; } enum _ecore_status_t ecore_fw_rss_eng(struct ecore_hwfn *p_hwfn, u8 src_id, u8 *dst_id) { if (src_id >= RESC_NUM(p_hwfn, ECORE_RSS_ENG)) { u8 min, max; min = (u8)RESC_START(p_hwfn, ECORE_RSS_ENG); max = min + RESC_NUM(p_hwfn, ECORE_RSS_ENG); DP_NOTICE(p_hwfn, true, "rss_eng id [%d] is not valid, available indices [%d - %d]\n", src_id, min, max); return ECORE_INVAL; } *dst_id = RESC_START(p_hwfn, ECORE_RSS_ENG) + src_id; return ECORE_SUCCESS; } static enum _ecore_status_t ecore_llh_add_mac_filter_bb_ah(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 high, u32 low, u32 *p_entry_num) { u32 en; int i; /* Find a free entry and utilize it */ for (i = 0; i < NIG_REG_LLH_FUNC_FILTER_EN_SIZE; i++) { en = ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32)); if (en) continue; ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32), low); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32), high); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_MODE_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_PROTOCOL_TYPE_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32), 1); break; } if (i >= NIG_REG_LLH_FUNC_FILTER_EN_SIZE) return ECORE_NORESOURCES; *p_entry_num = i; return ECORE_SUCCESS; } enum _ecore_status_t ecore_llh_add_mac_filter(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u8 *p_filter) { u32 high, low, entry_num; enum _ecore_status_t rc = ECORE_SUCCESS; if (!OSAL_TEST_BIT(ECORE_MF_LLH_MAC_CLSS, &p_hwfn->p_dev->mf_bits)) return ECORE_SUCCESS; high = p_filter[1] | (p_filter[0] << 8); low = p_filter[5] | (p_filter[4] << 8) | (p_filter[3] << 16) | (p_filter[2] << 24); if (ECORE_IS_BB(p_hwfn->p_dev) || ECORE_IS_AH(p_hwfn->p_dev)) rc = ecore_llh_add_mac_filter_bb_ah(p_hwfn, p_ptt, high, low, &entry_num); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, false, "Failed to find an empty LLH filter to utilize\n"); return rc; } DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "MAC: %02hhx:%02hhx:%02hhx:%02hhx:%02hhx:%02hhx is added at %d\n", p_filter[0], p_filter[1], p_filter[2], p_filter[3], p_filter[4], p_filter[5], entry_num); return rc; } static enum _ecore_status_t ecore_llh_remove_mac_filter_bb_ah(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 high, u32 low, u32 *p_entry_num) { int i; /* Find the entry and clean it */ for (i = 0; i < NIG_REG_LLH_FUNC_FILTER_EN_SIZE; i++) { if (ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32)) != low) continue; if (ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32)) != high) continue; ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32), 0); break; } if (i >= NIG_REG_LLH_FUNC_FILTER_EN_SIZE) return ECORE_INVAL; *p_entry_num = i; return ECORE_SUCCESS; } void ecore_llh_remove_mac_filter(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u8 *p_filter) { u32 high, low, entry_num; enum _ecore_status_t rc = ECORE_SUCCESS; if (!OSAL_TEST_BIT(ECORE_MF_LLH_MAC_CLSS, &p_hwfn->p_dev->mf_bits)) return; high = p_filter[1] | (p_filter[0] << 8); low = p_filter[5] | (p_filter[4] << 8) | (p_filter[3] << 16) | (p_filter[2] << 24); if (ECORE_IS_BB(p_hwfn->p_dev) || ECORE_IS_AH(p_hwfn->p_dev)) rc = ecore_llh_remove_mac_filter_bb_ah(p_hwfn, p_ptt, high, low, &entry_num); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, false, "Tried to remove a non-configured filter\n"); return; } DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "MAC: %02hhx:%02hhx:%02hhx:%02hhx:%02hhx:%02hhx was removed from %d\n", p_filter[0], p_filter[1], p_filter[2], p_filter[3], p_filter[4], p_filter[5], entry_num); } static enum _ecore_status_t ecore_llh_add_protocol_filter_bb_ah(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, enum ecore_llh_port_filter_type_t type, u32 high, u32 low, u32 *p_entry_num) { u32 en; int i; /* Find a free entry and utilize it */ for (i = 0; i < NIG_REG_LLH_FUNC_FILTER_EN_SIZE; i++) { en = ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32)); if (en) continue; ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32), low); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32), high); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_MODE_BB_K2 + i * sizeof(u32), 1); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_PROTOCOL_TYPE_BB_K2 + i * sizeof(u32), 1 << type); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32), 1); break; } if (i >= NIG_REG_LLH_FUNC_FILTER_EN_SIZE) return ECORE_NORESOURCES; *p_entry_num = i; return ECORE_SUCCESS; } enum _ecore_status_t ecore_llh_add_protocol_filter(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 source_port_or_eth_type, u16 dest_port, enum ecore_llh_port_filter_type_t type) { u32 high, low, entry_num; enum _ecore_status_t rc = ECORE_SUCCESS; if (!OSAL_TEST_BIT(ECORE_MF_LLH_PROTO_CLSS, &p_hwfn->p_dev->mf_bits)) return rc; high = 0; low = 0; switch (type) { case ECORE_LLH_FILTER_ETHERTYPE: high = source_port_or_eth_type; break; case ECORE_LLH_FILTER_TCP_SRC_PORT: case ECORE_LLH_FILTER_UDP_SRC_PORT: low = source_port_or_eth_type << 16; break; case ECORE_LLH_FILTER_TCP_DEST_PORT: case ECORE_LLH_FILTER_UDP_DEST_PORT: low = dest_port; break; case ECORE_LLH_FILTER_TCP_SRC_AND_DEST_PORT: case ECORE_LLH_FILTER_UDP_SRC_AND_DEST_PORT: low = (source_port_or_eth_type << 16) | dest_port; break; default: DP_NOTICE(p_hwfn, true, "Non valid LLH protocol filter type %d\n", type); return ECORE_INVAL; } if (ECORE_IS_BB(p_hwfn->p_dev) || ECORE_IS_AH(p_hwfn->p_dev)) rc = ecore_llh_add_protocol_filter_bb_ah(p_hwfn, p_ptt, type, high, low, &entry_num); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, false, "Failed to find an empty LLH filter to utilize\n"); return rc; } switch (type) { case ECORE_LLH_FILTER_ETHERTYPE: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "ETH type %x is added at %d\n", source_port_or_eth_type, entry_num); break; case ECORE_LLH_FILTER_TCP_SRC_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "TCP src port %x is added at %d\n", source_port_or_eth_type, entry_num); break; case ECORE_LLH_FILTER_UDP_SRC_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "UDP src port %x is added at %d\n", source_port_or_eth_type, entry_num); break; case ECORE_LLH_FILTER_TCP_DEST_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "TCP dst port %x is added at %d\n", dest_port, entry_num); break; case ECORE_LLH_FILTER_UDP_DEST_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "UDP dst port %x is added at %d\n", dest_port, entry_num); break; case ECORE_LLH_FILTER_TCP_SRC_AND_DEST_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "TCP src/dst ports %x/%x are added at %d\n", source_port_or_eth_type, dest_port, entry_num); break; case ECORE_LLH_FILTER_UDP_SRC_AND_DEST_PORT: DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "UDP src/dst ports %x/%x are added at %d\n", source_port_or_eth_type, dest_port, entry_num); break; } return rc; } static enum _ecore_status_t ecore_llh_remove_protocol_filter_bb_ah(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, enum ecore_llh_port_filter_type_t type, u32 high, u32 low, u32 *p_entry_num) { int i; /* Find the entry and clean it */ for (i = 0; i < NIG_REG_LLH_FUNC_FILTER_EN_SIZE; i++) { if (!ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32))) continue; if (!ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_MODE_BB_K2 + i * sizeof(u32))) continue; if (!(ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_PROTOCOL_TYPE_BB_K2 + i * sizeof(u32)) & (1 << type))) continue; if (ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32)) != low) continue; if (ecore_rd(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32)) != high) continue; ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_MODE_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_PROTOCOL_TYPE_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32), 0); break; } if (i >= NIG_REG_LLH_FUNC_FILTER_EN_SIZE) return ECORE_INVAL; *p_entry_num = i; return ECORE_SUCCESS; } void ecore_llh_remove_protocol_filter(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 source_port_or_eth_type, u16 dest_port, enum ecore_llh_port_filter_type_t type) { u32 high, low, entry_num; enum _ecore_status_t rc = ECORE_SUCCESS; if (!OSAL_TEST_BIT(ECORE_MF_LLH_PROTO_CLSS, &p_hwfn->p_dev->mf_bits)) return; high = 0; low = 0; switch (type) { case ECORE_LLH_FILTER_ETHERTYPE: high = source_port_or_eth_type; break; case ECORE_LLH_FILTER_TCP_SRC_PORT: case ECORE_LLH_FILTER_UDP_SRC_PORT: low = source_port_or_eth_type << 16; break; case ECORE_LLH_FILTER_TCP_DEST_PORT: case ECORE_LLH_FILTER_UDP_DEST_PORT: low = dest_port; break; case ECORE_LLH_FILTER_TCP_SRC_AND_DEST_PORT: case ECORE_LLH_FILTER_UDP_SRC_AND_DEST_PORT: low = (source_port_or_eth_type << 16) | dest_port; break; default: DP_NOTICE(p_hwfn, true, "Non valid LLH protocol filter type %d\n", type); return; } if (ECORE_IS_BB(p_hwfn->p_dev) || ECORE_IS_AH(p_hwfn->p_dev)) rc = ecore_llh_remove_protocol_filter_bb_ah(p_hwfn, p_ptt, type, high, low, &entry_num); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, false, "Tried to remove a non-configured filter [type %d, source_port_or_eth_type 0x%x, dest_port 0x%x]\n", type, source_port_or_eth_type, dest_port); return; } DP_VERBOSE(p_hwfn, ECORE_MSG_HW, "Protocol filter [type %d, source_port_or_eth_type 0x%x, dest_port 0x%x] was removed from %d\n", type, source_port_or_eth_type, dest_port, entry_num); } static void ecore_llh_clear_all_filters_bb_ah(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { int i; if (!(IS_MF_SI(p_hwfn) || IS_MF_DEFAULT(p_hwfn))) return; for (i = 0; i < NIG_REG_LLH_FUNC_FILTER_EN_SIZE; i++) { ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_EN_BB_K2 + i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + 2 * i * sizeof(u32), 0); ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_FUNC_FILTER_VALUE_BB_K2 + (2 * i + 1) * sizeof(u32), 0); } } void ecore_llh_clear_all_filters(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { if (!OSAL_TEST_BIT(ECORE_MF_LLH_PROTO_CLSS, &p_hwfn->p_dev->mf_bits) && !OSAL_TEST_BIT(ECORE_MF_LLH_MAC_CLSS, &p_hwfn->p_dev->mf_bits)) return; if (ECORE_IS_BB(p_hwfn->p_dev) || ECORE_IS_AH(p_hwfn->p_dev)) ecore_llh_clear_all_filters_bb_ah(p_hwfn, p_ptt); } enum _ecore_status_t ecore_llh_set_function_as_default(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { if (OSAL_TEST_BIT(ECORE_MF_NEED_DEF_PF, &p_hwfn->p_dev->mf_bits)) { ecore_wr(p_hwfn, p_ptt, NIG_REG_LLH_TAGMAC_DEF_PF_VECTOR, 1 << p_hwfn->abs_pf_id / 2); ecore_wr(p_hwfn, p_ptt, PRS_REG_MSG_INFO, 0); return ECORE_SUCCESS; } DP_NOTICE(p_hwfn, false, "This function can't be set as default\n"); return ECORE_INVAL; } static enum _ecore_status_t ecore_set_coalesce(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 hw_addr, void *p_eth_qzone, osal_size_t eth_qzone_size, u8 timeset) { struct coalescing_timeset *p_coal_timeset; if (p_hwfn->p_dev->int_coalescing_mode != ECORE_COAL_MODE_ENABLE) { DP_NOTICE(p_hwfn, true, "Coalescing configuration not enabled\n"); return ECORE_INVAL; } p_coal_timeset = p_eth_qzone; OSAL_MEMSET(p_eth_qzone, 0, eth_qzone_size); SET_FIELD(p_coal_timeset->value, COALESCING_TIMESET_TIMESET, timeset); SET_FIELD(p_coal_timeset->value, COALESCING_TIMESET_VALID, 1); ecore_memcpy_to(p_hwfn, p_ptt, hw_addr, p_eth_qzone, eth_qzone_size); return ECORE_SUCCESS; } enum _ecore_status_t ecore_set_queue_coalesce(struct ecore_hwfn *p_hwfn, u16 rx_coal, u16 tx_coal, void *p_handle) { struct ecore_queue_cid *p_cid = (struct ecore_queue_cid *)p_handle; enum _ecore_status_t rc = ECORE_SUCCESS; struct ecore_ptt *p_ptt; /* TODO - Configuring a single queue's coalescing but * claiming all queues are abiding same configuration * for PF and VF both. */ if (IS_VF(p_hwfn->p_dev)) return ecore_vf_pf_set_coalesce(p_hwfn, rx_coal, tx_coal, p_cid); p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_AGAIN; if (rx_coal) { rc = ecore_set_rxq_coalesce(p_hwfn, p_ptt, rx_coal, p_cid); if (rc) goto out; p_hwfn->p_dev->rx_coalesce_usecs = rx_coal; } if (tx_coal) { rc = ecore_set_txq_coalesce(p_hwfn, p_ptt, tx_coal, p_cid); if (rc) goto out; p_hwfn->p_dev->tx_coalesce_usecs = tx_coal; } out: ecore_ptt_release(p_hwfn, p_ptt); return rc; } enum _ecore_status_t ecore_set_rxq_coalesce(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 coalesce, struct ecore_queue_cid *p_cid) { struct ustorm_eth_queue_zone eth_qzone; u8 timeset, timer_res; u32 address; enum _ecore_status_t rc; /* Coalesce = (timeset << timer-resolution), timeset is 7bit wide */ if (coalesce <= 0x7F) { timer_res = 0; } else if (coalesce <= 0xFF) { timer_res = 1; } else if (coalesce <= 0x1FF) { timer_res = 2; } else { DP_ERR(p_hwfn, "Invalid coalesce value - %d\n", coalesce); return ECORE_INVAL; } timeset = (u8)(coalesce >> timer_res); rc = ecore_int_set_timer_res(p_hwfn, p_ptt, timer_res, p_cid->sb_igu_id, false); if (rc != ECORE_SUCCESS) goto out; address = BAR0_MAP_REG_USDM_RAM + USTORM_ETH_QUEUE_ZONE_OFFSET(p_cid->abs.queue_id); rc = ecore_set_coalesce(p_hwfn, p_ptt, address, ð_qzone, sizeof(struct ustorm_eth_queue_zone), timeset); if (rc != ECORE_SUCCESS) goto out; out: return rc; } enum _ecore_status_t ecore_set_txq_coalesce(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 coalesce, struct ecore_queue_cid *p_cid) { struct xstorm_eth_queue_zone eth_qzone; u8 timeset, timer_res; u32 address; enum _ecore_status_t rc; /* Coalesce = (timeset << timer-resolution), timeset is 7bit wide */ if (coalesce <= 0x7F) { timer_res = 0; } else if (coalesce <= 0xFF) { timer_res = 1; } else if (coalesce <= 0x1FF) { timer_res = 2; } else { DP_ERR(p_hwfn, "Invalid coalesce value - %d\n", coalesce); return ECORE_INVAL; } timeset = (u8)(coalesce >> timer_res); rc = ecore_int_set_timer_res(p_hwfn, p_ptt, timer_res, p_cid->sb_igu_id, true); if (rc != ECORE_SUCCESS) goto out; address = BAR0_MAP_REG_XSDM_RAM + XSTORM_ETH_QUEUE_ZONE_OFFSET(p_cid->abs.queue_id); rc = ecore_set_coalesce(p_hwfn, p_ptt, address, ð_qzone, sizeof(struct xstorm_eth_queue_zone), timeset); out: return rc; } /* Calculate final WFQ values for all vports and configure it. * After this configuration each vport must have * approx min rate = vport_wfq * min_pf_rate / ECORE_WFQ_UNIT */ static void ecore_configure_wfq_for_all_vports(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 min_pf_rate) { struct init_qm_vport_params *vport_params; int i; vport_params = p_hwfn->qm_info.qm_vport_params; for (i = 0; i < p_hwfn->qm_info.num_vports; i++) { u32 wfq_speed = p_hwfn->qm_info.wfq_data[i].min_speed; vport_params[i].vport_wfq = (wfq_speed * ECORE_WFQ_UNIT) / min_pf_rate; ecore_init_vport_wfq(p_hwfn, p_ptt, vport_params[i].first_tx_pq_id, vport_params[i].vport_wfq); } } static void ecore_init_wfq_default_param(struct ecore_hwfn *p_hwfn) { int i; for (i = 0; i < p_hwfn->qm_info.num_vports; i++) p_hwfn->qm_info.qm_vport_params[i].vport_wfq = 1; } static void ecore_disable_wfq_for_all_vports(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct init_qm_vport_params *vport_params; int i; vport_params = p_hwfn->qm_info.qm_vport_params; for (i = 0; i < p_hwfn->qm_info.num_vports; i++) { ecore_init_wfq_default_param(p_hwfn); ecore_init_vport_wfq(p_hwfn, p_ptt, vport_params[i].first_tx_pq_id, vport_params[i].vport_wfq); } } /* This function performs several validations for WFQ * configuration and required min rate for a given vport * 1. req_rate must be greater than one percent of min_pf_rate. * 2. req_rate should not cause other vports [not configured for WFQ explicitly] * rates to get less than one percent of min_pf_rate. * 3. total_req_min_rate [all vports min rate sum] shouldn't exceed min_pf_rate. */ static enum _ecore_status_t ecore_init_wfq_param(struct ecore_hwfn *p_hwfn, u16 vport_id, u32 req_rate, u32 min_pf_rate) { u32 total_req_min_rate = 0, total_left_rate = 0, left_rate_per_vp = 0; int non_requested_count = 0, req_count = 0, i, num_vports; num_vports = p_hwfn->qm_info.num_vports; /* Accounting for the vports which are configured for WFQ explicitly */ for (i = 0; i < num_vports; i++) { u32 tmp_speed; if ((i != vport_id) && p_hwfn->qm_info.wfq_data[i].configured) { req_count++; tmp_speed = p_hwfn->qm_info.wfq_data[i].min_speed; total_req_min_rate += tmp_speed; } } /* Include current vport data as well */ req_count++; total_req_min_rate += req_rate; non_requested_count = num_vports - req_count; /* validate possible error cases */ if (req_rate < min_pf_rate / ECORE_WFQ_UNIT) { DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Vport [%d] - Requested rate[%d Mbps] is less than one percent of configured PF min rate[%d Mbps]\n", vport_id, req_rate, min_pf_rate); return ECORE_INVAL; } /* TBD - for number of vports greater than 100 */ if (num_vports > ECORE_WFQ_UNIT) { DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Number of vports is greater than %d\n", ECORE_WFQ_UNIT); return ECORE_INVAL; } if (total_req_min_rate > min_pf_rate) { DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Total requested min rate for all vports[%d Mbps] is greater than configured PF min rate[%d Mbps]\n", total_req_min_rate, min_pf_rate); return ECORE_INVAL; } /* Data left for non requested vports */ total_left_rate = min_pf_rate - total_req_min_rate; left_rate_per_vp = total_left_rate / non_requested_count; /* validate if non requested get < 1% of min bw */ if (left_rate_per_vp < min_pf_rate / ECORE_WFQ_UNIT) { DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Non WFQ configured vports rate [%d Mbps] is less than one percent of configured PF min rate[%d Mbps]\n", left_rate_per_vp, min_pf_rate); return ECORE_INVAL; } /* now req_rate for given vport passes all scenarios. * assign final wfq rates to all vports. */ p_hwfn->qm_info.wfq_data[vport_id].min_speed = req_rate; p_hwfn->qm_info.wfq_data[vport_id].configured = true; for (i = 0; i < num_vports; i++) { if (p_hwfn->qm_info.wfq_data[i].configured) continue; p_hwfn->qm_info.wfq_data[i].min_speed = left_rate_per_vp; } return ECORE_SUCCESS; } static int __ecore_configure_vport_wfq(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u16 vp_id, u32 rate) { struct ecore_mcp_link_state *p_link; int rc = ECORE_SUCCESS; p_link = &p_hwfn->p_dev->hwfns[0].mcp_info->link_output; if (!p_link->min_pf_rate) { p_hwfn->qm_info.wfq_data[vp_id].min_speed = rate; p_hwfn->qm_info.wfq_data[vp_id].configured = true; return rc; } rc = ecore_init_wfq_param(p_hwfn, vp_id, rate, p_link->min_pf_rate); if (rc == ECORE_SUCCESS) ecore_configure_wfq_for_all_vports(p_hwfn, p_ptt, p_link->min_pf_rate); else DP_NOTICE(p_hwfn, false, "Validation failed while configuring min rate\n"); return rc; } static int __ecore_configure_vp_wfq_on_link_change(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, u32 min_pf_rate) { bool use_wfq = false; int rc = ECORE_SUCCESS; u16 i; /* Validate all pre configured vports for wfq */ for (i = 0; i < p_hwfn->qm_info.num_vports; i++) { u32 rate; if (!p_hwfn->qm_info.wfq_data[i].configured) continue; rate = p_hwfn->qm_info.wfq_data[i].min_speed; use_wfq = true; rc = ecore_init_wfq_param(p_hwfn, i, rate, min_pf_rate); if (rc != ECORE_SUCCESS) { DP_NOTICE(p_hwfn, false, "WFQ validation failed while configuring min rate\n"); break; } } if (rc == ECORE_SUCCESS && use_wfq) ecore_configure_wfq_for_all_vports(p_hwfn, p_ptt, min_pf_rate); else ecore_disable_wfq_for_all_vports(p_hwfn, p_ptt); return rc; } /* Main API for ecore clients to configure vport min rate. * vp_id - vport id in PF Range[0 - (total_num_vports_per_pf - 1)] * rate - Speed in Mbps needs to be assigned to a given vport. */ int ecore_configure_vport_wfq(struct ecore_dev *p_dev, u16 vp_id, u32 rate) { int i, rc = ECORE_INVAL; /* TBD - for multiple hardware functions - that is 100 gig */ if (ECORE_IS_CMT(p_dev)) { DP_NOTICE(p_dev, false, "WFQ configuration is not supported for this device\n"); return rc; } for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; struct ecore_ptt *p_ptt; p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_TIMEOUT; rc = __ecore_configure_vport_wfq(p_hwfn, p_ptt, vp_id, rate); if (rc != ECORE_SUCCESS) { ecore_ptt_release(p_hwfn, p_ptt); return rc; } ecore_ptt_release(p_hwfn, p_ptt); } return rc; } /* API to configure WFQ from mcp link change */ void ecore_configure_vp_wfq_on_link_change(struct ecore_dev *p_dev, struct ecore_ptt *p_ptt, u32 min_pf_rate) { int i; /* TBD - for multiple hardware functions - that is 100 gig */ if (ECORE_IS_CMT(p_dev)) { DP_VERBOSE(p_dev, ECORE_MSG_LINK, "WFQ configuration is not supported for this device\n"); return; } for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; __ecore_configure_vp_wfq_on_link_change(p_hwfn, p_ptt, min_pf_rate); } } int __ecore_configure_pf_max_bandwidth(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, struct ecore_mcp_link_state *p_link, u8 max_bw) { int rc = ECORE_SUCCESS; p_hwfn->mcp_info->func_info.bandwidth_max = max_bw; if (!p_link->line_speed && (max_bw != 100)) return rc; p_link->speed = (p_link->line_speed * max_bw) / 100; p_hwfn->qm_info.pf_rl = p_link->speed; /* Since the limiter also affects Tx-switched traffic, we don't want it * to limit such traffic in case there's no actual limit. * In that case, set limit to imaginary high boundary. */ if (max_bw == 100) p_hwfn->qm_info.pf_rl = 100000; rc = ecore_init_pf_rl(p_hwfn, p_ptt, p_hwfn->rel_pf_id, p_hwfn->qm_info.pf_rl); DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Configured MAX bandwidth to be %08x Mb/sec\n", p_link->speed); return rc; } /* Main API to configure PF max bandwidth where bw range is [1 - 100] */ int ecore_configure_pf_max_bandwidth(struct ecore_dev *p_dev, u8 max_bw) { int i, rc = ECORE_INVAL; if (max_bw < 1 || max_bw > 100) { DP_NOTICE(p_dev, false, "PF max bw valid range is [1-100]\n"); return rc; } for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; struct ecore_hwfn *p_lead = ECORE_LEADING_HWFN(p_dev); struct ecore_mcp_link_state *p_link; struct ecore_ptt *p_ptt; p_link = &p_lead->mcp_info->link_output; p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_TIMEOUT; rc = __ecore_configure_pf_max_bandwidth(p_hwfn, p_ptt, p_link, max_bw); ecore_ptt_release(p_hwfn, p_ptt); if (rc != ECORE_SUCCESS) break; } return rc; } int __ecore_configure_pf_min_bandwidth(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt, struct ecore_mcp_link_state *p_link, u8 min_bw) { int rc = ECORE_SUCCESS; p_hwfn->mcp_info->func_info.bandwidth_min = min_bw; p_hwfn->qm_info.pf_wfq = min_bw; if (!p_link->line_speed) return rc; p_link->min_pf_rate = (p_link->line_speed * min_bw) / 100; rc = ecore_init_pf_wfq(p_hwfn, p_ptt, p_hwfn->rel_pf_id, min_bw); DP_VERBOSE(p_hwfn, ECORE_MSG_LINK, "Configured MIN bandwidth to be %d Mb/sec\n", p_link->min_pf_rate); return rc; } /* Main API to configure PF min bandwidth where bw range is [1-100] */ int ecore_configure_pf_min_bandwidth(struct ecore_dev *p_dev, u8 min_bw) { int i, rc = ECORE_INVAL; if (min_bw < 1 || min_bw > 100) { DP_NOTICE(p_dev, false, "PF min bw valid range is [1-100]\n"); return rc; } for_each_hwfn(p_dev, i) { struct ecore_hwfn *p_hwfn = &p_dev->hwfns[i]; struct ecore_hwfn *p_lead = ECORE_LEADING_HWFN(p_dev); struct ecore_mcp_link_state *p_link; struct ecore_ptt *p_ptt; p_link = &p_lead->mcp_info->link_output; p_ptt = ecore_ptt_acquire(p_hwfn); if (!p_ptt) return ECORE_TIMEOUT; rc = __ecore_configure_pf_min_bandwidth(p_hwfn, p_ptt, p_link, min_bw); if (rc != ECORE_SUCCESS) { ecore_ptt_release(p_hwfn, p_ptt); return rc; } if (p_link->min_pf_rate) { u32 min_rate = p_link->min_pf_rate; rc = __ecore_configure_vp_wfq_on_link_change(p_hwfn, p_ptt, min_rate); } ecore_ptt_release(p_hwfn, p_ptt); } return rc; } void ecore_clean_wfq_db(struct ecore_hwfn *p_hwfn, struct ecore_ptt *p_ptt) { struct ecore_mcp_link_state *p_link; p_link = &p_hwfn->mcp_info->link_output; if (p_link->min_pf_rate) ecore_disable_wfq_for_all_vports(p_hwfn, p_ptt); OSAL_MEMSET(p_hwfn->qm_info.wfq_data, 0, sizeof(*p_hwfn->qm_info.wfq_data) * p_hwfn->qm_info.num_vports); } int ecore_device_num_engines(struct ecore_dev *p_dev) { return ECORE_IS_BB(p_dev) ? 2 : 1; } int ecore_device_num_ports(struct ecore_dev *p_dev) { return p_dev->num_ports; } void ecore_set_fw_mac_addr(__le16 *fw_msb, __le16 *fw_mid, __le16 *fw_lsb, u8 *mac) { ((u8 *)fw_msb)[0] = mac[1]; ((u8 *)fw_msb)[1] = mac[0]; ((u8 *)fw_mid)[0] = mac[3]; ((u8 *)fw_mid)[1] = mac[2]; ((u8 *)fw_lsb)[0] = mac[5]; ((u8 *)fw_lsb)[1] = mac[4]; }