From 8b25d1ad5d2264bdfc2818c7bda74ee2697df6db Mon Sep 17 00:00:00 2001 From: Christian Ehrhardt Date: Wed, 6 Jul 2016 09:22:35 +0200 Subject: Imported Upstream version 16.07-rc1 Change-Id: I40a523e52f12e8496fdd69e902824b0226c303de Signed-off-by: Christian Ehrhardt --- drivers/net/qede/base/ecore_vf.c | 1332 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 1332 insertions(+) create mode 100644 drivers/net/qede/base/ecore_vf.c (limited to 'drivers/net/qede/base/ecore_vf.c') diff --git a/drivers/net/qede/base/ecore_vf.c b/drivers/net/qede/base/ecore_vf.c new file mode 100644 index 00000000..d32fb35a --- /dev/null +++ b/drivers/net/qede/base/ecore_vf.c @@ -0,0 +1,1332 @@ +/* + * 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 "ecore.h" +#include "ecore_hsi_eth.h" +#include "ecore_sriov.h" +#include "ecore_l2_api.h" +#include "ecore_vf.h" +#include "ecore_vfpf_if.h" +#include "ecore_status.h" +#include "reg_addr.h" +#include "ecore_int.h" +#include "ecore_l2.h" +#include "ecore_mcp_api.h" +#include "ecore_vf_api.h" + +static void *ecore_vf_pf_prep(struct ecore_hwfn *p_hwfn, u16 type, u16 length) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + void *p_tlv; + + /* This lock is released when we receive PF's response + * in ecore_send_msg2pf(). + * So, ecore_vf_pf_prep() and ecore_send_msg2pf() + * must come in sequence. + */ + OSAL_MUTEX_ACQUIRE(&p_iov->mutex); + + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "preparing to send %s tlv over vf pf channel\n", + ecore_channel_tlvs_string[type]); + + /* Reset Request offset */ + p_iov->offset = (u8 *)(p_iov->vf2pf_request); + + /* Clear mailbox - both request and reply */ + OSAL_MEMSET(p_iov->vf2pf_request, 0, sizeof(union vfpf_tlvs)); + OSAL_MEMSET(p_iov->pf2vf_reply, 0, sizeof(union pfvf_tlvs)); + + /* Init type and length */ + p_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, type, length); + + /* Init first tlv header */ + ((struct vfpf_first_tlv *)p_tlv)->reply_address = + (u64)p_iov->pf2vf_reply_phys; + + return p_tlv; +} + +static int ecore_send_msg2pf(struct ecore_hwfn *p_hwfn, + u8 *done, u32 resp_size) +{ + struct ustorm_vf_zone *zone_data = (struct ustorm_vf_zone *) + ((u8 *)PXP_VF_BAR0_START_USDM_ZONE_B); + union vfpf_tlvs *p_req = p_hwfn->vf_iov_info->vf2pf_request; + struct ustorm_trigger_vf_zone trigger; + int rc = ECORE_SUCCESS, time = 100; + u8 pf_id; + + /* output tlvs list */ + ecore_dp_tlv_list(p_hwfn, p_req); + + /* need to add the END TLV to the message size */ + resp_size += sizeof(struct channel_list_end_tlv); + + if (!p_hwfn->p_dev->sriov_info.b_hw_channel) { + rc = OSAL_VF_SEND_MSG2PF(p_hwfn->p_dev, + done, + p_req, + p_hwfn->vf_iov_info->pf2vf_reply, + sizeof(union vfpf_tlvs), resp_size); + /* TODO - no prints about message ? */ + goto exit; + } + + /* Send TLVs over HW channel */ + OSAL_MEMSET(&trigger, 0, sizeof(struct ustorm_trigger_vf_zone)); + trigger.vf_pf_msg_valid = 1; + /* TODO - FW should remove this requirement */ + pf_id = GET_FIELD(p_hwfn->hw_info.concrete_fid, PXP_CONCRETE_FID_PFID); + + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VF -> PF [%02x] message: [%08x, %08x] --> %p, %08x --> %p\n", + pf_id, + U64_HI(p_hwfn->vf_iov_info->vf2pf_request_phys), + U64_LO(p_hwfn->vf_iov_info->vf2pf_request_phys), + &zone_data->non_trigger.vf_pf_msg_addr, + *((u32 *)&trigger), &zone_data->trigger); + + REG_WR(p_hwfn, + (osal_uintptr_t)&zone_data->non_trigger.vf_pf_msg_addr.lo, + U64_LO(p_hwfn->vf_iov_info->vf2pf_request_phys)); + + REG_WR(p_hwfn, + (osal_uintptr_t)&zone_data->non_trigger.vf_pf_msg_addr.hi, + U64_HI(p_hwfn->vf_iov_info->vf2pf_request_phys)); + + /* The message data must be written first, to prevent trigger before + * data is written. + */ + OSAL_WMB(p_hwfn->p_dev); + + REG_WR(p_hwfn, (osal_uintptr_t)&zone_data->trigger, + *((u32 *)&trigger)); + + while ((!*done) && time) { + OSAL_MSLEEP(25); + time--; + } + + if (!*done) { + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VF <-- PF Timeout [Type %d]\n", + p_req->first_tlv.tl.type); + rc = ECORE_TIMEOUT; + goto exit; + } else { + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "PF response: %d [Type %d]\n", + *done, p_req->first_tlv.tl.type); + } + +exit: + OSAL_MUTEX_RELEASE(&p_hwfn->vf_iov_info->mutex); + + return rc; +} + +#define VF_ACQUIRE_THRESH 3 +#define VF_ACQUIRE_MAC_FILTERS 1 +#define VF_ACQUIRE_MC_FILTERS 10 + +static enum _ecore_status_t ecore_vf_pf_acquire(struct ecore_hwfn *p_hwfn) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct pfvf_acquire_resp_tlv *resp = &p_iov->pf2vf_reply->acquire_resp; + struct pf_vf_pfdev_info *pfdev_info = &resp->pfdev_info; + struct ecore_vf_acquire_sw_info vf_sw_info; + struct vfpf_acquire_tlv *req; + int rc = 0, attempts = 0; + bool resources_acquired = false; + + /* @@@ TBD: MichalK take this from somewhere else... */ + u8 rx_count = 1, tx_count = 1, num_sbs = 1; + u8 num_mac = VF_ACQUIRE_MAC_FILTERS, num_mc = VF_ACQUIRE_MC_FILTERS; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_ACQUIRE, sizeof(*req)); + + /* @@@ TBD: PF may not be ready bnx2x_get_vf_id... */ + req->vfdev_info.opaque_fid = p_hwfn->hw_info.opaque_fid; + + req->resc_request.num_rxqs = rx_count; + req->resc_request.num_txqs = tx_count; + req->resc_request.num_sbs = num_sbs; + req->resc_request.num_mac_filters = num_mac; + req->resc_request.num_mc_filters = num_mc; + req->resc_request.num_vlan_filters = ECORE_ETH_VF_NUM_VLAN_FILTERS; + + OSAL_MEMSET(&vf_sw_info, 0, sizeof(vf_sw_info)); + OSAL_VF_FILL_ACQUIRE_RESC_REQ(p_hwfn, &req->resc_request, &vf_sw_info); + + req->vfdev_info.os_type = vf_sw_info.os_type; + req->vfdev_info.driver_version = vf_sw_info.driver_version; + req->vfdev_info.fw_major = FW_MAJOR_VERSION; + req->vfdev_info.fw_minor = FW_MINOR_VERSION; + req->vfdev_info.fw_revision = FW_REVISION_VERSION; + req->vfdev_info.fw_engineering = FW_ENGINEERING_VERSION; + + if (vf_sw_info.override_fw_version) + req->vfdev_info.capabilties |= VFPF_ACQUIRE_CAP_OVERRIDE_FW_VER; + + /* pf 2 vf bulletin board address */ + req->bulletin_addr = p_iov->bulletin.phys; + req->bulletin_size = p_iov->bulletin.size; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + while (!resources_acquired) { + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "attempting to acquire resources\n"); + + /* send acquire request */ + rc = ecore_send_msg2pf(p_hwfn, + &resp->hdr.status, sizeof(*resp)); + + /* PF timeout */ + if (rc) + return rc; + + /* copy acquire response from buffer to p_hwfn */ + OSAL_MEMCPY(&p_iov->acquire_resp, + resp, sizeof(p_iov->acquire_resp)); + + attempts++; + + /* PF agrees to allocate our resources */ + if (resp->hdr.status == PFVF_STATUS_SUCCESS) { + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "resources acquired\n"); + resources_acquired = true; + } /* PF refuses to allocate our resources */ + else if (resp->hdr.status == + PFVF_STATUS_NO_RESOURCE && + attempts < VF_ACQUIRE_THRESH) { + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "PF unwilling to fullfill resource request. Try PF recommended amount\n"); + + /* humble our request */ + req->resc_request.num_txqs = resp->resc.num_txqs; + req->resc_request.num_rxqs = resp->resc.num_rxqs; + req->resc_request.num_sbs = resp->resc.num_sbs; + req->resc_request.num_mac_filters = + resp->resc.num_mac_filters; + req->resc_request.num_vlan_filters = + resp->resc.num_vlan_filters; + req->resc_request.num_mc_filters = + resp->resc.num_mc_filters; + + /* Clear response buffer */ + OSAL_MEMSET(p_iov->pf2vf_reply, 0, + sizeof(union pfvf_tlvs)); + } else { + DP_ERR(p_hwfn, + "PF returned error %d to VF acquisition request\n", + resp->hdr.status); + return ECORE_AGAIN; + } + } + + rc = OSAL_VF_UPDATE_ACQUIRE_RESC_RESP(p_hwfn, &resp->resc); + if (rc) { + DP_NOTICE(p_hwfn, true, + "VF_UPDATE_ACQUIRE_RESC_RESP Failed: status = 0x%x.\n", + rc); + return ECORE_AGAIN; + } + + /* Update bulletin board size with response from PF */ + p_iov->bulletin.size = resp->bulletin_size; + + /* get HW info */ + p_hwfn->p_dev->type = resp->pfdev_info.dev_type; + p_hwfn->p_dev->chip_rev = resp->pfdev_info.chip_rev; + + DP_INFO(p_hwfn, "Chip details - %s%d\n", + ECORE_IS_BB(p_hwfn->p_dev) ? "BB" : "AH", + CHIP_REV_IS_A0(p_hwfn->p_dev) ? 0 : 1); + + /* @@@TBD MichalK: Fw ver... */ + /* strlcpy(p_hwfn->fw_ver, p_hwfn->acquire_resp.pfdev_info.fw_ver, + * sizeof(p_hwfn->fw_ver)); + */ + + p_hwfn->p_dev->chip_num = pfdev_info->chip_num & 0xffff; + + return 0; +} + +enum _ecore_status_t ecore_vf_hw_prepare(struct ecore_dev *p_dev) +{ + enum _ecore_status_t rc = ECORE_NOMEM; + struct ecore_vf_iov *p_sriov; + struct ecore_hwfn *p_hwfn = &p_dev->hwfns[0]; /* @@@TBD CMT */ + + p_dev->num_hwfns = 1; /* @@@TBD CMT must be fixed... */ + + p_hwfn->regview = p_dev->regview; + if (p_hwfn->regview == OSAL_NULL) { + DP_ERR(p_hwfn, + "regview should be initialized before" + " ecore_vf_hw_prepare is called\n"); + return ECORE_INVAL; + } + + /* Set the doorbell bar. Assumption: regview is set */ + p_hwfn->doorbells = (u8 OSAL_IOMEM *)p_hwfn->regview + + PXP_VF_BAR0_START_DQ; + + p_hwfn->hw_info.opaque_fid = (u16)REG_RD(p_hwfn, + PXP_VF_BAR0_ME_OPAQUE_ADDRESS); + + p_hwfn->hw_info.concrete_fid = REG_RD(p_hwfn, + PXP_VF_BAR0_ME_CONCRETE_ADDRESS); + + /* Allocate vf sriov info */ + p_sriov = OSAL_ZALLOC(p_hwfn->p_dev, GFP_KERNEL, sizeof(*p_sriov)); + if (!p_sriov) { + DP_NOTICE(p_hwfn, true, + "Failed to allocate `struct ecore_sriov'\n"); + return ECORE_NOMEM; + } + + OSAL_MEMSET(p_sriov, 0, sizeof(*p_sriov)); + + /* Allocate vf2pf msg */ + p_sriov->vf2pf_request = OSAL_DMA_ALLOC_COHERENT(p_hwfn->p_dev, + &p_sriov-> + vf2pf_request_phys, + sizeof(union + vfpf_tlvs)); + if (!p_sriov->vf2pf_request) { + DP_NOTICE(p_hwfn, true, + "Failed to allocate `vf2pf_request' DMA memory\n"); + goto free_p_sriov; + } + + p_sriov->pf2vf_reply = OSAL_DMA_ALLOC_COHERENT(p_hwfn->p_dev, + &p_sriov-> + pf2vf_reply_phys, + sizeof(union pfvf_tlvs)); + if (!p_sriov->pf2vf_reply) { + DP_NOTICE(p_hwfn, true, + "Failed to allocate `pf2vf_reply' DMA memory\n"); + goto free_vf2pf_request; + } + + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VF's Request mailbox [%p virt 0x%" PRIx64 " phys], " + "Response mailbox [%p virt 0x%" PRIx64 " phys]\n", + p_sriov->vf2pf_request, + (u64)p_sriov->vf2pf_request_phys, + p_sriov->pf2vf_reply, (u64)p_sriov->pf2vf_reply_phys); + + /* Allocate Bulletin board */ + p_sriov->bulletin.size = sizeof(struct ecore_bulletin_content); + p_sriov->bulletin.p_virt = OSAL_DMA_ALLOC_COHERENT(p_hwfn->p_dev, + &p_sriov->bulletin. + phys, + p_sriov->bulletin. + size); + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VF's bulletin Board [%p virt 0x%" PRIx64 " phys 0x%08x bytes]\n", + p_sriov->bulletin.p_virt, (u64)p_sriov->bulletin.phys, + p_sriov->bulletin.size); + + OSAL_MUTEX_ALLOC(p_hwfn, &p_sriov->mutex); + OSAL_MUTEX_INIT(&p_sriov->mutex); + + p_hwfn->vf_iov_info = p_sriov; + + p_hwfn->hw_info.personality = ECORE_PCI_ETH; + + /* First VF needs to query for information from PF */ + if (!p_hwfn->my_id) + rc = ecore_vf_pf_acquire(p_hwfn); + + return rc; + +free_vf2pf_request: + OSAL_DMA_FREE_COHERENT(p_hwfn->p_dev, p_sriov->vf2pf_request, + p_sriov->vf2pf_request_phys, + sizeof(union vfpf_tlvs)); +free_p_sriov: + OSAL_FREE(p_hwfn->p_dev, p_sriov); + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_init(struct ecore_hwfn *p_hwfn) +{ + p_hwfn->b_int_enabled = 1; + + return 0; +} + +/* TEMP TEMP until in HSI */ +#define TSTORM_QZONE_START PXP_VF_BAR0_START_SDM_ZONE_A +#define MSTORM_QZONE_START(dev) (TSTORM_QZONE_START + \ + (TSTORM_QZONE_SIZE * NUM_OF_L2_QUEUES(dev))) +#define USTORM_QZONE_START(dev) (MSTORM_QZONE_START + \ + (MSTORM_QZONE_SIZE * NUM_OF_L2_QUEUES(dev))) + +enum _ecore_status_t ecore_vf_pf_rxq_start(struct ecore_hwfn *p_hwfn, + u8 rx_qid, + u16 sb, + u8 sb_index, + u16 bd_max_bytes, + dma_addr_t bd_chain_phys_addr, + dma_addr_t cqe_pbl_addr, + u16 cqe_pbl_size, + void OSAL_IOMEM * *pp_prod) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_start_rxq_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + u8 hw_qid; + u64 init_prod_val = 0; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_START_RXQ, sizeof(*req)); + + /* @@@TBD MichalK TPA */ + + req->rx_qid = rx_qid; + req->cqe_pbl_addr = cqe_pbl_addr; + req->cqe_pbl_size = cqe_pbl_size; + req->rxq_addr = bd_chain_phys_addr; + req->hw_sb = sb; + req->sb_index = sb_index; + req->hc_rate = 0; /* @@@TBD MichalK -> host coalescing! */ + req->bd_max_bytes = bd_max_bytes; + req->stat_id = -1; /* No stats at the moment */ + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + if (pp_prod) { + hw_qid = p_iov->acquire_resp.resc.hw_qid[rx_qid]; + + *pp_prod = (u8 OSAL_IOMEM *)p_hwfn->regview + + MSTORM_QZONE_START(p_hwfn->p_dev) + + (hw_qid) * MSTORM_QZONE_SIZE + + OFFSETOF(struct mstorm_eth_queue_zone, rx_producers); + + /* Init the rcq, rx bd and rx sge (if valid) producers to 0 */ + __internal_ram_wr(p_hwfn, *pp_prod, sizeof(u64), + (u32 *)(&init_prod_val)); + } + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_rxq_stop(struct ecore_hwfn *p_hwfn, + u16 rx_qid, bool cqe_completion) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_stop_rxqs_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_STOP_RXQS, sizeof(*req)); + + /* @@@TBD MichalK TPA */ + + /* @@@TBD MichalK - relevant ??? + * flags VFPF_QUEUE_FLG_OV VFPF_QUEUE_FLG_VLAN + */ + req->rx_qid = rx_qid; + req->num_rxqs = 1; + req->cqe_completion = cqe_completion; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_txq_start(struct ecore_hwfn *p_hwfn, + u16 tx_queue_id, + u16 sb, + u8 sb_index, + dma_addr_t pbl_addr, + u16 pbl_size, + void OSAL_IOMEM * *pp_doorbell) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_start_txq_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_START_TXQ, sizeof(*req)); + + /* @@@TBD MichalK TPA */ + + req->tx_qid = tx_queue_id; + + /* Tx */ + req->pbl_addr = pbl_addr; + req->pbl_size = pbl_size; + req->hw_sb = sb; + req->sb_index = sb_index; + req->hc_rate = 0; /* @@@TBD MichalK -> host coalescing! */ + req->flags = 0; /* @@@TBD MichalK -> flags... */ + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + if (pp_doorbell) { + u8 cid = p_iov->acquire_resp.resc.cid[tx_queue_id]; + + *pp_doorbell = (u8 OSAL_IOMEM *)p_hwfn->doorbells + + DB_ADDR_VF(cid, DQ_DEMS_LEGACY); + } + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_txq_stop(struct ecore_hwfn *p_hwfn, u16 tx_qid) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_stop_txqs_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_STOP_TXQS, sizeof(*req)); + + /* @@@TBD MichalK TPA */ + + /* @@@TBD MichalK - relevant ??? flags + * VFPF_QUEUE_FLG_OV VFPF_QUEUE_FLG_VLAN + */ + req->tx_qid = tx_qid; + req->num_txqs = 1; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_rxqs_update(struct ecore_hwfn *p_hwfn, + u16 rx_queue_id, + u8 num_rxqs, + u8 comp_cqe_flg, u8 comp_event_flg) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + struct vfpf_update_rxq_tlv *req; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_UPDATE_RXQ, sizeof(*req)); + + req->rx_qid = rx_queue_id; + req->num_rxqs = num_rxqs; + + if (comp_cqe_flg) + req->flags |= VFPF_RXQ_UPD_COMPLETE_CQE_FLAG; + if (comp_event_flg) + req->flags |= VFPF_RXQ_UPD_COMPLETE_EVENT_FLAG; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +enum _ecore_status_t +ecore_vf_pf_vport_start(struct ecore_hwfn *p_hwfn, u8 vport_id, + u16 mtu, u8 inner_vlan_removal, + enum ecore_tpa_mode tpa_mode, u8 max_buffers_per_cqe, + u8 only_untagged) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_vport_start_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc, i; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_VPORT_START, sizeof(*req)); + + req->mtu = mtu; + req->vport_id = vport_id; + req->inner_vlan_removal = inner_vlan_removal; + req->tpa_mode = tpa_mode; + req->max_buffers_per_cqe = max_buffers_per_cqe; + req->only_untagged = only_untagged; + + /* status blocks */ + for (i = 0; i < p_hwfn->vf_iov_info->acquire_resp.resc.num_sbs; i++) + if (p_hwfn->sbs_info[i]) + req->sb_addr[i] = p_hwfn->sbs_info[i]->sb_phys; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_vport_stop(struct ecore_hwfn *p_hwfn) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_VPORT_TEARDOWN, + sizeof(struct vfpf_first_tlv)); + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return rc; +} + +static void +ecore_vf_handle_vp_update_tlvs_resp(struct ecore_hwfn *p_hwfn, + struct ecore_sp_vport_update_params *p_data) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct pfvf_def_resp_tlv *p_resp; + u16 tlv; + + if (p_data->update_vport_active_rx_flg || + p_data->update_vport_active_tx_flg) { + tlv = CHANNEL_TLV_VPORT_UPDATE_ACTIVATE; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update activate tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update activate tlv config failed\n"); + } + + if (p_data->update_tx_switching_flg) { + tlv = CHANNEL_TLV_VPORT_UPDATE_TX_SWITCH; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update tx switch tlv configured\n"); +#ifndef ASIC_ONLY + else if (CHIP_REV_IS_FPGA(p_hwfn->p_dev)) + DP_NOTICE(p_hwfn, false, + "FPGA: Skip checking whether PF" + " replied to Tx-switching request\n"); +#endif + else + DP_NOTICE(p_hwfn, true, + "VP update tx switch tlv config failed\n"); + } + + if (p_data->update_inner_vlan_removal_flg) { + tlv = CHANNEL_TLV_VPORT_UPDATE_VLAN_STRIP; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update vlan strip tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update vlan strip tlv config failed\n"); + } + + if (p_data->update_approx_mcast_flg) { + tlv = CHANNEL_TLV_VPORT_UPDATE_MCAST; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update mcast tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update mcast tlv config failed\n"); + } + + if (p_data->accept_flags.update_rx_mode_config || + p_data->accept_flags.update_tx_mode_config) { + tlv = CHANNEL_TLV_VPORT_UPDATE_ACCEPT_PARAM; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update accept_mode tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update accept_mode tlv config failed\n"); + } + + if (p_data->rss_params) { + tlv = CHANNEL_TLV_VPORT_UPDATE_RSS; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update rss tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update rss tlv config failed\n"); + } + + if (p_data->sge_tpa_params) { + tlv = CHANNEL_TLV_VPORT_UPDATE_SGE_TPA; + p_resp = (struct pfvf_def_resp_tlv *) + ecore_iov_search_list_tlvs(p_hwfn, p_iov->pf2vf_reply, tlv); + if (p_resp && p_resp->hdr.status) + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "VP update sge tpa tlv configured\n"); + else + DP_NOTICE(p_hwfn, true, + "VP update sge tpa tlv config failed\n"); + } +} + +enum _ecore_status_t +ecore_vf_pf_vport_update(struct ecore_hwfn *p_hwfn, + struct ecore_sp_vport_update_params *p_params) +{ + struct vfpf_vport_update_accept_any_vlan_tlv *p_any_vlan_tlv; + struct vfpf_vport_update_accept_param_tlv *p_accept_tlv; + struct vfpf_vport_update_tx_switch_tlv *p_tx_switch_tlv; + struct vfpf_vport_update_mcast_bin_tlv *p_mcast_tlv; + struct vfpf_vport_update_vlan_strip_tlv *p_vlan_tlv; + struct vfpf_vport_update_sge_tpa_tlv *p_sge_tpa_tlv; + struct vfpf_vport_update_activate_tlv *p_act_tlv; + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_vport_update_rss_tlv *p_rss_tlv; + struct vfpf_vport_update_tlv *req; + struct pfvf_def_resp_tlv *resp; + u8 update_rx, update_tx; + u32 resp_size = 0; + u16 size, tlv; + int rc; + + resp = &p_iov->pf2vf_reply->default_resp; + resp_size = sizeof(*resp); + + update_rx = p_params->update_vport_active_rx_flg; + update_tx = p_params->update_vport_active_tx_flg; + + /* clear mailbox and prep header tlv */ + ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_VPORT_UPDATE, sizeof(*req)); + + /* Prepare extended tlvs */ + if (update_rx || update_tx) { + size = sizeof(struct vfpf_vport_update_activate_tlv); + p_act_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_VPORT_UPDATE_ACTIVATE, + size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + if (update_rx) { + p_act_tlv->update_rx = update_rx; + p_act_tlv->active_rx = p_params->vport_active_rx_flg; + } + + if (update_tx) { + p_act_tlv->update_tx = update_tx; + p_act_tlv->active_tx = p_params->vport_active_tx_flg; + } + } + + if (p_params->update_inner_vlan_removal_flg) { + size = sizeof(struct vfpf_vport_update_vlan_strip_tlv); + p_vlan_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_VPORT_UPDATE_VLAN_STRIP, + size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + p_vlan_tlv->remove_vlan = p_params->inner_vlan_removal_flg; + } + + if (p_params->update_tx_switching_flg) { + size = sizeof(struct vfpf_vport_update_tx_switch_tlv); + tlv = CHANNEL_TLV_VPORT_UPDATE_TX_SWITCH; + p_tx_switch_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + tlv, size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + p_tx_switch_tlv->tx_switching = p_params->tx_switching_flg; + } + + if (p_params->update_approx_mcast_flg) { + size = sizeof(struct vfpf_vport_update_mcast_bin_tlv); + p_mcast_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_VPORT_UPDATE_MCAST, + size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + OSAL_MEMCPY(p_mcast_tlv->bins, p_params->bins, + sizeof(unsigned long) * + ETH_MULTICAST_MAC_BINS_IN_REGS); + } + + update_rx = p_params->accept_flags.update_rx_mode_config; + update_tx = p_params->accept_flags.update_tx_mode_config; + + if (update_rx || update_tx) { + tlv = CHANNEL_TLV_VPORT_UPDATE_ACCEPT_PARAM; + size = sizeof(struct vfpf_vport_update_accept_param_tlv); + p_accept_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, tlv, size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + if (update_rx) { + p_accept_tlv->update_rx_mode = update_rx; + p_accept_tlv->rx_accept_filter = + p_params->accept_flags.rx_accept_filter; + } + + if (update_tx) { + p_accept_tlv->update_tx_mode = update_tx; + p_accept_tlv->tx_accept_filter = + p_params->accept_flags.tx_accept_filter; + } + } + + if (p_params->rss_params) { + struct ecore_rss_params *rss_params = p_params->rss_params; + + size = sizeof(struct vfpf_vport_update_rss_tlv); + p_rss_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_VPORT_UPDATE_RSS, size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + if (rss_params->update_rss_config) + p_rss_tlv->update_rss_flags |= + VFPF_UPDATE_RSS_CONFIG_FLAG; + if (rss_params->update_rss_capabilities) + p_rss_tlv->update_rss_flags |= + VFPF_UPDATE_RSS_CAPS_FLAG; + if (rss_params->update_rss_ind_table) + p_rss_tlv->update_rss_flags |= + VFPF_UPDATE_RSS_IND_TABLE_FLAG; + if (rss_params->update_rss_key) + p_rss_tlv->update_rss_flags |= VFPF_UPDATE_RSS_KEY_FLAG; + + p_rss_tlv->rss_enable = rss_params->rss_enable; + p_rss_tlv->rss_caps = rss_params->rss_caps; + p_rss_tlv->rss_table_size_log = rss_params->rss_table_size_log; + OSAL_MEMCPY(p_rss_tlv->rss_ind_table, rss_params->rss_ind_table, + sizeof(rss_params->rss_ind_table)); + OSAL_MEMCPY(p_rss_tlv->rss_key, rss_params->rss_key, + sizeof(rss_params->rss_key)); + } + + if (p_params->update_accept_any_vlan_flg) { + size = sizeof(struct vfpf_vport_update_accept_any_vlan_tlv); + tlv = CHANNEL_TLV_VPORT_UPDATE_ACCEPT_ANY_VLAN; + p_any_vlan_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + tlv, size); + + resp_size += sizeof(struct pfvf_def_resp_tlv); + p_any_vlan_tlv->accept_any_vlan = p_params->accept_any_vlan; + p_any_vlan_tlv->update_accept_any_vlan_flg = + p_params->update_accept_any_vlan_flg; + } + + if (p_params->sge_tpa_params) { + struct ecore_sge_tpa_params *sge_tpa_params = + p_params->sge_tpa_params; + + size = sizeof(struct vfpf_vport_update_sge_tpa_tlv); + p_sge_tpa_tlv = ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_VPORT_UPDATE_SGE_TPA, + size); + resp_size += sizeof(struct pfvf_def_resp_tlv); + + if (sge_tpa_params->update_tpa_en_flg) + p_sge_tpa_tlv->update_sge_tpa_flags |= + VFPF_UPDATE_TPA_EN_FLAG; + if (sge_tpa_params->update_tpa_param_flg) + p_sge_tpa_tlv->update_sge_tpa_flags |= + VFPF_UPDATE_TPA_PARAM_FLAG; + + if (sge_tpa_params->tpa_ipv4_en_flg) + p_sge_tpa_tlv->sge_tpa_flags |= VFPF_TPA_IPV4_EN_FLAG; + if (sge_tpa_params->tpa_ipv6_en_flg) + p_sge_tpa_tlv->sge_tpa_flags |= VFPF_TPA_IPV6_EN_FLAG; + if (sge_tpa_params->tpa_pkt_split_flg) + p_sge_tpa_tlv->sge_tpa_flags |= VFPF_TPA_PKT_SPLIT_FLAG; + if (sge_tpa_params->tpa_hdr_data_split_flg) + p_sge_tpa_tlv->sge_tpa_flags |= + VFPF_TPA_HDR_DATA_SPLIT_FLAG; + if (sge_tpa_params->tpa_gro_consistent_flg) + p_sge_tpa_tlv->sge_tpa_flags |= + VFPF_TPA_GRO_CONSIST_FLAG; + + p_sge_tpa_tlv->tpa_max_aggs_num = + sge_tpa_params->tpa_max_aggs_num; + p_sge_tpa_tlv->tpa_max_size = sge_tpa_params->tpa_max_size; + p_sge_tpa_tlv->tpa_min_size_to_start = + sge_tpa_params->tpa_min_size_to_start; + p_sge_tpa_tlv->tpa_min_size_to_cont = + sge_tpa_params->tpa_min_size_to_cont; + + p_sge_tpa_tlv->max_buffers_per_cqe = + sge_tpa_params->max_buffers_per_cqe; + } + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, resp_size); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + ecore_vf_handle_vp_update_tlvs_resp(p_hwfn, p_params); + + return rc; +} + +enum _ecore_status_t ecore_vf_pf_reset(struct ecore_hwfn *p_hwfn) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_first_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_CLOSE, sizeof(*req)); + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_AGAIN; + + p_hwfn->b_int_enabled = 0; + + return ECORE_SUCCESS; +} + +enum _ecore_status_t ecore_vf_pf_release(struct ecore_hwfn *p_hwfn) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_first_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + u32 size; + int rc; + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_RELEASE, sizeof(*req)); + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + + if (rc == ECORE_SUCCESS && resp->hdr.status != PFVF_STATUS_SUCCESS) + rc = ECORE_AGAIN; + + p_hwfn->b_int_enabled = 0; + + /* TODO - might need to revise this for 100g */ + if (IS_LEAD_HWFN(p_hwfn)) + OSAL_MUTEX_DEALLOC(&p_iov->mutex); + + if (p_iov->vf2pf_request) + OSAL_DMA_FREE_COHERENT(p_hwfn->p_dev, + p_iov->vf2pf_request, + p_iov->vf2pf_request_phys, + sizeof(union vfpf_tlvs)); + if (p_iov->pf2vf_reply) + OSAL_DMA_FREE_COHERENT(p_hwfn->p_dev, + p_iov->pf2vf_reply, + p_iov->pf2vf_reply_phys, + sizeof(union pfvf_tlvs)); + + if (p_iov->bulletin.p_virt) { + size = sizeof(struct ecore_bulletin_content); + OSAL_DMA_FREE_COHERENT(p_hwfn->p_dev, + p_iov->bulletin.p_virt, + p_iov->bulletin.phys, size); + } + + OSAL_FREE(p_hwfn->p_dev, p_hwfn->vf_iov_info); + p_hwfn->vf_iov_info = OSAL_NULL; + + return rc; +} + +void ecore_vf_pf_filter_mcast(struct ecore_hwfn *p_hwfn, + struct ecore_filter_mcast *p_filter_cmd) +{ + struct ecore_sp_vport_update_params sp_params; + int i; + + OSAL_MEMSET(&sp_params, 0, sizeof(sp_params)); + sp_params.update_approx_mcast_flg = 1; + + if (p_filter_cmd->opcode == ECORE_FILTER_ADD) { + for (i = 0; i < p_filter_cmd->num_mc_addrs; i++) { + u32 bit; + + bit = ecore_mcast_bin_from_mac(p_filter_cmd->mac[i]); + OSAL_SET_BIT(bit, sp_params.bins); + } + } + + ecore_vf_pf_vport_update(p_hwfn, &sp_params); +} + +enum _ecore_status_t ecore_vf_pf_filter_ucast(struct ecore_hwfn *p_hwfn, + struct ecore_filter_ucast + *p_ucast) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct vfpf_ucast_filter_tlv *req; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* Sanitize */ + if (p_ucast->opcode == ECORE_FILTER_MOVE) { + DP_NOTICE(p_hwfn, true, + "VFs don't support Moving of filters\n"); + return ECORE_INVAL; + } + + /* clear mailbox and prep first tlv */ + req = ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_UCAST_FILTER, sizeof(*req)); + req->opcode = (u8)p_ucast->opcode; + req->type = (u8)p_ucast->type; + OSAL_MEMCPY(req->mac, p_ucast->mac, ETH_ALEN); + req->vlan = p_ucast->vlan; + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_AGAIN; + + return ECORE_SUCCESS; +} + +enum _ecore_status_t ecore_vf_pf_int_cleanup(struct ecore_hwfn *p_hwfn) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + struct pfvf_def_resp_tlv *resp = &p_iov->pf2vf_reply->default_resp; + int rc; + + /* clear mailbox and prep first tlv */ + ecore_vf_pf_prep(p_hwfn, CHANNEL_TLV_INT_CLEANUP, + sizeof(struct vfpf_first_tlv)); + + /* add list termination tlv */ + ecore_add_tlv(p_hwfn, &p_iov->offset, + CHANNEL_TLV_LIST_END, + sizeof(struct channel_list_end_tlv)); + + rc = ecore_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + if (rc) + return rc; + + if (resp->hdr.status != PFVF_STATUS_SUCCESS) + return ECORE_INVAL; + + return ECORE_SUCCESS; +} + +enum _ecore_status_t ecore_vf_read_bulletin(struct ecore_hwfn *p_hwfn, + u8 *p_change) +{ + struct ecore_bulletin_content shadow; + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + u32 crc, crc_size = sizeof(p_iov->bulletin.p_virt->crc); + + *p_change = 0; + + /* Need to guarantee PF is not in the middle of writing it */ + OSAL_MEMCPY(&shadow, p_iov->bulletin.p_virt, p_iov->bulletin.size); + + /* If version did not update, no need to do anything */ + if (shadow.version == p_iov->bulletin_shadow.version) + return ECORE_SUCCESS; + + /* Verify the bulletin we see is valid */ + crc = ecore_crc32(0, (u8 *)&shadow + crc_size, + p_iov->bulletin.size - crc_size); + if (crc != shadow.crc) + return ECORE_AGAIN; + + /* Set the shadow bulletin and process it */ + OSAL_MEMCPY(&p_iov->bulletin_shadow, &shadow, p_iov->bulletin.size); + + DP_VERBOSE(p_hwfn, ECORE_MSG_IOV, + "Read a bulletin update %08x\n", shadow.version); + + *p_change = 1; + + return ECORE_SUCCESS; +} + +u16 ecore_vf_get_igu_sb_id(struct ecore_hwfn *p_hwfn, u16 sb_id) +{ + struct ecore_vf_iov *p_iov = p_hwfn->vf_iov_info; + + if (!p_iov) { + DP_NOTICE(p_hwfn, true, "vf_sriov_info isn't initialized\n"); + return 0; + } + + return p_iov->acquire_resp.resc.hw_sbs[sb_id].hw_sb_id; +} + +void __ecore_vf_get_link_params(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_params *p_params, + struct ecore_bulletin_content *p_bulletin) +{ + OSAL_MEMSET(p_params, 0, sizeof(*p_params)); + + p_params->speed.autoneg = p_bulletin->req_autoneg; + p_params->speed.advertised_speeds = p_bulletin->req_adv_speed; + p_params->speed.forced_speed = p_bulletin->req_forced_speed; + p_params->pause.autoneg = p_bulletin->req_autoneg_pause; + p_params->pause.forced_rx = p_bulletin->req_forced_rx; + p_params->pause.forced_tx = p_bulletin->req_forced_tx; + p_params->loopback_mode = p_bulletin->req_loopback; +} + +void ecore_vf_get_link_params(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_params *params) +{ + __ecore_vf_get_link_params(p_hwfn, params, + &p_hwfn->vf_iov_info->bulletin_shadow); +} + +void __ecore_vf_get_link_state(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_state *p_link, + struct ecore_bulletin_content *p_bulletin) +{ + OSAL_MEMSET(p_link, 0, sizeof(*p_link)); + + p_link->link_up = p_bulletin->link_up; + p_link->speed = p_bulletin->speed; + p_link->full_duplex = p_bulletin->full_duplex; + p_link->an = p_bulletin->autoneg; + p_link->an_complete = p_bulletin->autoneg_complete; + p_link->parallel_detection = p_bulletin->parallel_detection; + p_link->pfc_enabled = p_bulletin->pfc_enabled; + p_link->partner_adv_speed = p_bulletin->partner_adv_speed; + p_link->partner_tx_flow_ctrl_en = p_bulletin->partner_tx_flow_ctrl_en; + p_link->partner_rx_flow_ctrl_en = p_bulletin->partner_rx_flow_ctrl_en; + p_link->partner_adv_pause = p_bulletin->partner_adv_pause; + p_link->sfp_tx_fault = p_bulletin->sfp_tx_fault; +} + +void ecore_vf_get_link_state(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_state *link) +{ + __ecore_vf_get_link_state(p_hwfn, link, + &p_hwfn->vf_iov_info->bulletin_shadow); +} + +void __ecore_vf_get_link_caps(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_capabilities *p_link_caps, + struct ecore_bulletin_content *p_bulletin) +{ + OSAL_MEMSET(p_link_caps, 0, sizeof(*p_link_caps)); + p_link_caps->speed_capabilities = p_bulletin->capability_speed; +} + +void ecore_vf_get_link_caps(struct ecore_hwfn *p_hwfn, + struct ecore_mcp_link_capabilities *p_link_caps) +{ + __ecore_vf_get_link_caps(p_hwfn, p_link_caps, + &p_hwfn->vf_iov_info->bulletin_shadow); +} + +void ecore_vf_get_num_rxqs(struct ecore_hwfn *p_hwfn, u8 *num_rxqs) +{ + *num_rxqs = p_hwfn->vf_iov_info->acquire_resp.resc.num_rxqs; +} + +void ecore_vf_get_port_mac(struct ecore_hwfn *p_hwfn, u8 *port_mac) +{ + OSAL_MEMCPY(port_mac, + p_hwfn->vf_iov_info->acquire_resp.pfdev_info.port_mac, + ETH_ALEN); +} + +void ecore_vf_get_num_vlan_filters(struct ecore_hwfn *p_hwfn, + u8 *num_vlan_filters) +{ + struct ecore_vf_iov *p_vf; + + p_vf = p_hwfn->vf_iov_info; + *num_vlan_filters = p_vf->acquire_resp.resc.num_vlan_filters; +} + +/* @DPDK */ +void ecore_vf_get_num_mac_filters(struct ecore_hwfn *p_hwfn, + u32 *num_mac) +{ + struct ecore_vf_iov *p_vf; + + p_vf = p_hwfn->vf_iov_info; + *num_mac = p_vf->acquire_resp.resc.num_mac_filters; +} + +bool ecore_vf_check_mac(struct ecore_hwfn *p_hwfn, u8 *mac) +{ + struct ecore_bulletin_content *bulletin; + + bulletin = &p_hwfn->vf_iov_info->bulletin_shadow; + if (!(bulletin->valid_bitmap & (1 << MAC_ADDR_FORCED))) + return true; + + /* Forbid VF from changing a MAC enforced by PF */ + if (OSAL_MEMCMP(bulletin->mac, mac, ETH_ALEN)) + return false; + + return false; +} + +bool ecore_vf_bulletin_get_forced_mac(struct ecore_hwfn *hwfn, u8 *dst_mac, + u8 *p_is_forced) +{ + struct ecore_bulletin_content *bulletin; + + bulletin = &hwfn->vf_iov_info->bulletin_shadow; + + if (bulletin->valid_bitmap & (1 << MAC_ADDR_FORCED)) { + if (p_is_forced) + *p_is_forced = 1; + } else if (bulletin->valid_bitmap & (1 << VFPF_BULLETIN_MAC_ADDR)) { + if (p_is_forced) + *p_is_forced = 0; + } else { + return false; + } + + OSAL_MEMCPY(dst_mac, bulletin->mac, ETH_ALEN); + + return true; +} + +bool ecore_vf_bulletin_get_forced_vlan(struct ecore_hwfn *hwfn, u16 *dst_pvid) +{ + struct ecore_bulletin_content *bulletin; + + bulletin = &hwfn->vf_iov_info->bulletin_shadow; + + if (!(bulletin->valid_bitmap & (1 << VLAN_ADDR_FORCED))) + return false; + + if (dst_pvid) + *dst_pvid = bulletin->pvid; + + return true; +} + +void ecore_vf_get_fw_version(struct ecore_hwfn *p_hwfn, + u16 *fw_major, u16 *fw_minor, u16 *fw_rev, + u16 *fw_eng) +{ + struct pf_vf_pfdev_info *info; + + info = &p_hwfn->vf_iov_info->acquire_resp.pfdev_info; + + *fw_major = info->fw_major; + *fw_minor = info->fw_minor; + *fw_rev = info->fw_rev; + *fw_eng = info->fw_eng; +} -- cgit 1.2.3-korg