// SPDX-License-Identifier: GPL-2.0 /* Copyright (c) 2019 HiSilicon Limited. */ #include #include #include #include #include #include #include #include #include #include "qm.h" /* eq/aeq irq enable */ #define QM_VF_AEQ_INT_SOURCE 0x0 #define QM_VF_AEQ_INT_MASK 0x4 #define QM_VF_EQ_INT_SOURCE 0x8 #define QM_VF_EQ_INT_MASK 0xc #define QM_IRQ_NUM_V1 1 #define QM_IRQ_NUM_PF_V2 4 #define QM_IRQ_NUM_VF_V2 2 #define QM_EQ_EVENT_IRQ_VECTOR 0 #define QM_AEQ_EVENT_IRQ_VECTOR 1 #define QM_ABNORMAL_EVENT_IRQ_VECTOR 3 /* mailbox */ #define QM_MB_CMD_SQC 0x0 #define QM_MB_CMD_CQC 0x1 #define QM_MB_CMD_EQC 0x2 #define QM_MB_CMD_AEQC 0x3 #define QM_MB_CMD_SQC_BT 0x4 #define QM_MB_CMD_CQC_BT 0x5 #define QM_MB_CMD_SQC_VFT_V2 0x6 #define QM_MB_CMD_SEND_BASE 0x300 #define QM_MB_EVENT_SHIFT 8 #define QM_MB_BUSY_SHIFT 13 #define QM_MB_OP_SHIFT 14 #define QM_MB_CMD_DATA_ADDR_L 0x304 #define QM_MB_CMD_DATA_ADDR_H 0x308 /* sqc shift */ #define QM_SQ_HOP_NUM_SHIFT 0 #define QM_SQ_PAGE_SIZE_SHIFT 4 #define QM_SQ_BUF_SIZE_SHIFT 8 #define QM_SQ_SQE_SIZE_SHIFT 12 #define QM_SQ_PRIORITY_SHIFT 0 #define QM_SQ_ORDERS_SHIFT 4 #define QM_SQ_TYPE_SHIFT 8 #define QM_SQ_TYPE_MASK GENMASK(3, 0) /* cqc shift */ #define QM_CQ_HOP_NUM_SHIFT 0 #define QM_CQ_PAGE_SIZE_SHIFT 4 #define QM_CQ_BUF_SIZE_SHIFT 8 #define QM_CQ_CQE_SIZE_SHIFT 12 #define QM_CQ_PHASE_SHIFT 0 #define QM_CQ_FLAG_SHIFT 1 #define QM_CQE_PHASE(cqe) (le16_to_cpu((cqe)->w7) & 0x1) #define QM_QC_CQE_SIZE 4 /* eqc shift */ #define QM_EQE_AEQE_SIZE (2UL << 12) #define QM_EQC_PHASE_SHIFT 16 #define QM_EQE_PHASE(eqe) ((le32_to_cpu((eqe)->dw0) >> 16) & 0x1) #define QM_EQE_CQN_MASK GENMASK(15, 0) #define QM_AEQE_PHASE(aeqe) ((le32_to_cpu((aeqe)->dw0) >> 16) & 0x1) #define QM_AEQE_TYPE_SHIFT 17 #define QM_DOORBELL_CMD_SQ 0 #define QM_DOORBELL_CMD_CQ 1 #define QM_DOORBELL_CMD_EQ 2 #define QM_DOORBELL_CMD_AEQ 3 #define QM_DOORBELL_BASE_V1 0x340 #define QM_DB_CMD_SHIFT_V1 16 #define QM_DB_INDEX_SHIFT_V1 32 #define QM_DB_PRIORITY_SHIFT_V1 48 #define QM_DOORBELL_SQ_CQ_BASE_V2 0x1000 #define QM_DOORBELL_EQ_AEQ_BASE_V2 0x2000 #define QM_DB_CMD_SHIFT_V2 12 #define QM_DB_RAND_SHIFT_V2 16 #define QM_DB_INDEX_SHIFT_V2 32 #define QM_DB_PRIORITY_SHIFT_V2 48 #define QM_MEM_START_INIT 0x100040 #define QM_MEM_INIT_DONE 0x100044 #define QM_VFT_CFG_RDY 0x10006c #define QM_VFT_CFG_OP_WR 0x100058 #define QM_VFT_CFG_TYPE 0x10005c #define QM_SQC_VFT 0x0 #define QM_CQC_VFT 0x1 #define QM_VFT_CFG 0x100060 #define QM_VFT_CFG_OP_ENABLE 0x100054 #define QM_VFT_CFG_DATA_L 0x100064 #define QM_VFT_CFG_DATA_H 0x100068 #define QM_SQC_VFT_BUF_SIZE (7ULL << 8) #define QM_SQC_VFT_SQC_SIZE (5ULL << 12) #define QM_SQC_VFT_INDEX_NUMBER (1ULL << 16) #define QM_SQC_VFT_START_SQN_SHIFT 28 #define QM_SQC_VFT_VALID (1ULL << 44) #define QM_SQC_VFT_SQN_SHIFT 45 #define QM_CQC_VFT_BUF_SIZE (7ULL << 8) #define QM_CQC_VFT_SQC_SIZE (5ULL << 12) #define QM_CQC_VFT_INDEX_NUMBER (1ULL << 16) #define QM_CQC_VFT_VALID (1ULL << 28) #define QM_SQC_VFT_BASE_SHIFT_V2 28 #define QM_SQC_VFT_BASE_MASK_V2 GENMASK(5, 0) #define QM_SQC_VFT_NUM_SHIFT_V2 45 #define QM_SQC_VFT_NUM_MASK_v2 GENMASK(9, 0) #define QM_DFX_CNT_CLR_CE 0x100118 #define QM_ABNORMAL_INT_SOURCE 0x100000 #define QM_ABNORMAL_INT_MASK 0x100004 #define QM_ABNORMAL_INT_MASK_VALUE 0x1fff #define QM_ABNORMAL_INT_STATUS 0x100008 #define QM_ABNORMAL_INF00 0x100010 #define QM_FIFO_OVERFLOW_TYPE 0xc0 #define QM_FIFO_OVERFLOW_TYPE_SHIFT 6 #define QM_FIFO_OVERFLOW_VF 0x3f #define QM_ABNORMAL_INF01 0x100014 #define QM_DB_TIMEOUT_TYPE 0xc0 #define QM_DB_TIMEOUT_TYPE_SHIFT 6 #define QM_DB_TIMEOUT_VF 0x3f #define QM_RAS_CE_ENABLE 0x1000ec #define QM_RAS_FE_ENABLE 0x1000f0 #define QM_RAS_NFE_ENABLE 0x1000f4 #define QM_RAS_CE_THRESHOLD 0x1000f8 #define QM_RAS_CE_TIMES_PER_IRQ 1 #define QM_RAS_MSI_INT_SEL 0x1040f4 #define QM_CACHE_WB_START 0x204 #define QM_CACHE_WB_DONE 0x208 #define PCI_BAR_2 2 #define QM_SQE_DATA_ALIGN_MASK GENMASK(6, 0) #define QMC_ALIGN(sz) ALIGN(sz, 32) #define QM_DBG_TMP_BUF_LEN 22 #define QM_MK_CQC_DW3_V1(hop_num, pg_sz, buf_sz, cqe_sz) \ (((hop_num) << QM_CQ_HOP_NUM_SHIFT) | \ ((pg_sz) << QM_CQ_PAGE_SIZE_SHIFT) | \ ((buf_sz) << QM_CQ_BUF_SIZE_SHIFT) | \ ((cqe_sz) << QM_CQ_CQE_SIZE_SHIFT)) #define QM_MK_CQC_DW3_V2(cqe_sz) \ ((QM_Q_DEPTH - 1) | ((cqe_sz) << QM_CQ_CQE_SIZE_SHIFT)) #define QM_MK_SQC_W13(priority, orders, alg_type) \ (((priority) << QM_SQ_PRIORITY_SHIFT) | \ ((orders) << QM_SQ_ORDERS_SHIFT) | \ (((alg_type) & QM_SQ_TYPE_MASK) << QM_SQ_TYPE_SHIFT)) #define QM_MK_SQC_DW3_V1(hop_num, pg_sz, buf_sz, sqe_sz) \ (((hop_num) << QM_SQ_HOP_NUM_SHIFT) | \ ((pg_sz) << QM_SQ_PAGE_SIZE_SHIFT) | \ ((buf_sz) << QM_SQ_BUF_SIZE_SHIFT) | \ ((u32)ilog2(sqe_sz) << QM_SQ_SQE_SIZE_SHIFT)) #define QM_MK_SQC_DW3_V2(sqe_sz) \ ((QM_Q_DEPTH - 1) | ((u32)ilog2(sqe_sz) << QM_SQ_SQE_SIZE_SHIFT)) #define INIT_QC_COMMON(qc, base, pasid) do { \ (qc)->head = 0; \ (qc)->tail = 0; \ (qc)->base_l = cpu_to_le32(lower_32_bits(base)); \ (qc)->base_h = cpu_to_le32(upper_32_bits(base)); \ (qc)->dw3 = 0; \ (qc)->w8 = 0; \ (qc)->rsvd0 = 0; \ (qc)->pasid = cpu_to_le16(pasid); \ (qc)->w11 = 0; \ (qc)->rsvd1 = 0; \ } while (0) enum vft_type { SQC_VFT = 0, CQC_VFT, }; struct qm_cqe { __le32 rsvd0; __le16 cmd_id; __le16 rsvd1; __le16 sq_head; __le16 sq_num; __le16 rsvd2; __le16 w7; }; struct qm_eqe { __le32 dw0; }; struct qm_aeqe { __le32 dw0; }; struct qm_sqc { __le16 head; __le16 tail; __le32 base_l; __le32 base_h; __le32 dw3; __le16 w8; __le16 rsvd0; __le16 pasid; __le16 w11; __le16 cq_num; __le16 w13; __le32 rsvd1; }; struct qm_cqc { __le16 head; __le16 tail; __le32 base_l; __le32 base_h; __le32 dw3; __le16 w8; __le16 rsvd0; __le16 pasid; __le16 w11; __le32 dw6; __le32 rsvd1; }; struct qm_eqc { __le16 head; __le16 tail; __le32 base_l; __le32 base_h; __le32 dw3; __le32 rsvd[2]; __le32 dw6; }; struct qm_aeqc { __le16 head; __le16 tail; __le32 base_l; __le32 base_h; __le32 dw3; __le32 rsvd[2]; __le32 dw6; }; struct qm_mailbox { __le16 w0; __le16 queue_num; __le32 base_l; __le32 base_h; __le32 rsvd; }; struct qm_doorbell { __le16 queue_num; __le16 cmd; __le16 index; __le16 priority; }; struct hisi_qm_hw_ops { int (*get_vft)(struct hisi_qm *qm, u32 *base, u32 *number); void (*qm_db)(struct hisi_qm *qm, u16 qn, u8 cmd, u16 index, u8 priority); u32 (*get_irq_num)(struct hisi_qm *qm); int (*debug_init)(struct hisi_qm *qm); void (*hw_error_init)(struct hisi_qm *qm, u32 ce, u32 nfe, u32 fe, u32 msi); pci_ers_result_t (*hw_error_handle)(struct hisi_qm *qm); }; static const char * const qm_debug_file_name[] = { [CURRENT_Q] = "current_q", [CLEAR_ENABLE] = "clear_enable", }; struct hisi_qm_hw_error { u32 int_msk; const char *msg; }; static const struct hisi_qm_hw_error qm_hw_error[] = { { .int_msk = BIT(0), .msg = "qm_axi_rresp" }, { .int_msk = BIT(1), .msg = "qm_axi_bresp" }, { .int_msk = BIT(2), .msg = "qm_ecc_mbit" }, { .int_msk = BIT(3), .msg = "qm_ecc_1bit" }, { .int_msk = BIT(4), .msg = "qm_acc_get_task_timeout" }, { .int_msk = BIT(5), .msg = "qm_acc_do_task_timeout" }, { .int_msk = BIT(6), .msg = "qm_acc_wb_not_ready_timeout" }, { .int_msk = BIT(7), .msg = "qm_sq_cq_vf_invalid" }, { .int_msk = BIT(8), .msg = "qm_cq_vf_invalid" }, { .int_msk = BIT(9), .msg = "qm_sq_vf_invalid" }, { .int_msk = BIT(10), .msg = "qm_db_timeout" }, { .int_msk = BIT(11), .msg = "qm_of_fifo_of" }, { .int_msk = BIT(12), .msg = "qm_db_random_invalid" }, { /* sentinel */ } }; static const char * const qm_db_timeout[] = { "sq", "cq", "eq", "aeq", }; static const char * const qm_fifo_overflow[] = { "cq", "eq", "aeq", }; /* return 0 mailbox ready, -ETIMEDOUT hardware timeout */ static int qm_wait_mb_ready(struct hisi_qm *qm) { u32 val; return readl_relaxed_poll_timeout(qm->io_base + QM_MB_CMD_SEND_BASE, val, !((val >> QM_MB_BUSY_SHIFT) & 0x1), 10, 1000); } /* 128 bit should be written to hardware at one time to trigger a mailbox */ static void qm_mb_write(struct hisi_qm *qm, const void *src) { void __iomem *fun_base = qm->io_base + QM_MB_CMD_SEND_BASE; unsigned long tmp0 = 0, tmp1 = 0; if (!IS_ENABLED(CONFIG_ARM64)) { memcpy_toio(fun_base, src, 16); wmb(); return; } asm volatile("ldp %0, %1, %3\n" "stp %0, %1, %2\n" "dsb sy\n" : "=&r" (tmp0), "=&r" (tmp1), "+Q" (*((char __iomem *)fun_base)) : "Q" (*((char *)src)) : "memory"); } static int qm_mb(struct hisi_qm *qm, u8 cmd, dma_addr_t dma_addr, u16 queue, bool op) { struct qm_mailbox mailbox; int ret = 0; dev_dbg(&qm->pdev->dev, "QM mailbox request to q%u: %u-%llx\n", queue, cmd, (unsigned long long)dma_addr); mailbox.w0 = cpu_to_le16(cmd | (op ? 0x1 << QM_MB_OP_SHIFT : 0) | (0x1 << QM_MB_BUSY_SHIFT)); mailbox.queue_num = cpu_to_le16(queue); mailbox.base_l = cpu_to_le32(lower_32_bits(dma_addr)); mailbox.base_h = cpu_to_le32(upper_32_bits(dma_addr)); mailbox.rsvd = 0; mutex_lock(&qm->mailbox_lock); if (unlikely(qm_wait_mb_ready(qm))) { ret = -EBUSY; dev_err(&qm->pdev->dev, "QM mailbox is busy to start!\n"); goto busy_unlock; } qm_mb_write(qm, &mailbox); if (unlikely(qm_wait_mb_ready(qm))) { ret = -EBUSY; dev_err(&qm->pdev->dev, "QM mailbox operation timeout!\n"); goto busy_unlock; } busy_unlock: mutex_unlock(&qm->mailbox_lock); return ret; } static void qm_db_v1(struct hisi_qm *qm, u16 qn, u8 cmd, u16 index, u8 priority) { u64 doorbell; doorbell = qn | ((u64)cmd << QM_DB_CMD_SHIFT_V1) | ((u64)index << QM_DB_INDEX_SHIFT_V1) | ((u64)priority << QM_DB_PRIORITY_SHIFT_V1); writeq(doorbell, qm->io_base + QM_DOORBELL_BASE_V1); } static void qm_db_v2(struct hisi_qm *qm, u16 qn, u8 cmd, u16 index, u8 priority) { u64 doorbell; u64 dbase; u16 randata = 0; if (cmd == QM_DOORBELL_CMD_SQ || cmd == QM_DOORBELL_CMD_CQ) dbase = QM_DOORBELL_SQ_CQ_BASE_V2; else dbase = QM_DOORBELL_EQ_AEQ_BASE_V2; doorbell = qn | ((u64)cmd << QM_DB_CMD_SHIFT_V2) | ((u64)randata << QM_DB_RAND_SHIFT_V2) | ((u64)index << QM_DB_INDEX_SHIFT_V2) | ((u64)priority << QM_DB_PRIORITY_SHIFT_V2); writeq(doorbell, qm->io_base + dbase); } static void qm_db(struct hisi_qm *qm, u16 qn, u8 cmd, u16 index, u8 priority) { dev_dbg(&qm->pdev->dev, "QM doorbell request: qn=%u, cmd=%u, index=%u\n", qn, cmd, index); qm->ops->qm_db(qm, qn, cmd, index, priority); } static int qm_dev_mem_reset(struct hisi_qm *qm) { u32 val; writel(0x1, qm->io_base + QM_MEM_START_INIT); return readl_relaxed_poll_timeout(qm->io_base + QM_MEM_INIT_DONE, val, val & BIT(0), 10, 1000); } static u32 qm_get_irq_num_v1(struct hisi_qm *qm) { return QM_IRQ_NUM_V1; } static u32 qm_get_irq_num_v2(struct hisi_qm *qm) { if (qm->fun_type == QM_HW_PF) return QM_IRQ_NUM_PF_V2; else return QM_IRQ_NUM_VF_V2; } static struct hisi_qp *qm_to_hisi_qp(struct hisi_qm *qm, struct qm_eqe *eqe) { u16 cqn = le32_to_cpu(eqe->dw0) & QM_EQE_CQN_MASK; return qm->qp_array[cqn]; } static void qm_cq_head_update(struct hisi_qp *qp) { if (qp->qp_status.cq_head == QM_Q_DEPTH - 1) { qp->qp_status.cqc_phase = !qp->qp_status.cqc_phase; qp->qp_status.cq_head = 0; } else { qp->qp_status.cq_head++; } } static void qm_poll_qp(struct hisi_qp *qp, struct hisi_qm *qm) { struct qm_cqe *cqe = qp->cqe + qp->qp_status.cq_head; if (qp->req_cb) { while (QM_CQE_PHASE(cqe) == qp->qp_status.cqc_phase) { dma_rmb(); qp->req_cb(qp, qp->sqe + qm->sqe_size * le16_to_cpu(cqe->sq_head)); qm_cq_head_update(qp); cqe = qp->cqe + qp->qp_status.cq_head; qm_db(qm, qp->qp_id, QM_DOORBELL_CMD_CQ, qp->qp_status.cq_head, 0); atomic_dec(&qp->qp_status.used); } /* set c_flag */ qm_db(qm, qp->qp_id, QM_DOORBELL_CMD_CQ, qp->qp_status.cq_head, 1); } } static void qm_qp_work_func(struct work_struct *work) { struct hisi_qp *qp; qp = container_of(work, struct hisi_qp, work); qm_poll_qp(qp, qp->qm); } static irqreturn_t qm_irq_handler(int irq, void *data) { struct hisi_qm *qm = data; struct qm_eqe *eqe = qm->eqe + qm->status.eq_head; struct hisi_qp *qp; int eqe_num = 0; while (QM_EQE_PHASE(eqe) == qm->status.eqc_phase) { eqe_num++; qp = qm_to_hisi_qp(qm, eqe); if (qp) queue_work(qp->wq, &qp->work); if (qm->status.eq_head == QM_Q_DEPTH - 1) { qm->status.eqc_phase = !qm->status.eqc_phase; eqe = qm->eqe; qm->status.eq_head = 0; } else { eqe++; qm->status.eq_head++; } if (eqe_num == QM_Q_DEPTH / 2 - 1) { eqe_num = 0; qm_db(qm, 0, QM_DOORBELL_CMD_EQ, qm->status.eq_head, 0); } } qm_db(qm, 0, QM_DOORBELL_CMD_EQ, qm->status.eq_head, 0); return IRQ_HANDLED; } static irqreturn_t qm_irq(int irq, void *data) { struct hisi_qm *qm = data; if (readl(qm->io_base + QM_VF_EQ_INT_SOURCE)) return qm_irq_handler(irq, data); dev_err(&qm->pdev->dev, "invalid int source\n"); qm_db(qm, 0, QM_DOORBELL_CMD_EQ, qm->status.eq_head, 0); return IRQ_NONE; } static irqreturn_t qm_aeq_irq(int irq, void *data) { struct hisi_qm *qm = data; struct qm_aeqe *aeqe = qm->aeqe + qm->status.aeq_head; u32 type; if (!readl(qm->io_base + QM_VF_AEQ_INT_SOURCE)) return IRQ_NONE; while (QM_AEQE_PHASE(aeqe) == qm->status.aeqc_phase) { type = le32_to_cpu(aeqe->dw0) >> QM_AEQE_TYPE_SHIFT; if (type < ARRAY_SIZE(qm_fifo_overflow)) dev_err(&qm->pdev->dev, "%s overflow\n", qm_fifo_overflow[type]); else dev_err(&qm->pdev->dev, "unknown error type %d\n", type); if (qm->status.aeq_head == QM_Q_DEPTH - 1) { qm->status.aeqc_phase = !qm->status.aeqc_phase; aeqe = qm->aeqe; qm->status.aeq_head = 0; } else { aeqe++; qm->status.aeq_head++; } qm_db(qm, 0, QM_DOORBELL_CMD_AEQ, qm->status.aeq_head, 0); } return IRQ_HANDLED; } static irqreturn_t qm_abnormal_irq(int irq, void *data) { const struct hisi_qm_hw_error *err = qm_hw_error; struct hisi_qm *qm = data; struct device *dev = &qm->pdev->dev; u32 error_status, tmp; /* read err sts */ tmp = readl(qm->io_base + QM_ABNORMAL_INT_STATUS); error_status = qm->msi_mask & tmp; while (err->msg) { if (err->int_msk & error_status) dev_err(dev, "%s [error status=0x%x] found\n", err->msg, err->int_msk); err++; } /* clear err sts */ writel(error_status, qm->io_base + QM_ABNORMAL_INT_SOURCE); return IRQ_HANDLED; } static int qm_irq_register(struct hisi_qm *qm) { struct pci_dev *pdev = qm->pdev; int ret; ret = request_irq(pci_irq_vector(pdev, QM_EQ_EVENT_IRQ_VECTOR), qm_irq, IRQF_SHARED, qm->dev_name, qm); if (ret) return ret; if (qm->ver == QM_HW_V2) { ret = request_irq(pci_irq_vector(pdev, QM_AEQ_EVENT_IRQ_VECTOR), qm_aeq_irq, IRQF_SHARED, qm->dev_name, qm); if (ret) goto err_aeq_irq; if (qm->fun_type == QM_HW_PF) { ret = request_irq(pci_irq_vector(pdev, QM_ABNORMAL_EVENT_IRQ_VECTOR), qm_abnormal_irq, IRQF_SHARED, qm->dev_name, qm); if (ret) goto err_abonormal_irq; } } return 0; err_abonormal_irq: free_irq(pci_irq_vector(pdev, QM_AEQ_EVENT_IRQ_VECTOR), qm); err_aeq_irq: free_irq(pci_irq_vector(pdev, QM_EQ_EVENT_IRQ_VECTOR), qm); return ret; } static void qm_irq_unregister(struct hisi_qm *qm) { struct pci_dev *pdev = qm->pdev; free_irq(pci_irq_vector(pdev, QM_EQ_EVENT_IRQ_VECTOR), qm); if (qm->ver == QM_HW_V2) { free_irq(pci_irq_vector(pdev, QM_AEQ_EVENT_IRQ_VECTOR), qm); if (qm->fun_type == QM_HW_PF) free_irq(pci_irq_vector(pdev, QM_ABNORMAL_EVENT_IRQ_VECTOR), qm); } } static void qm_init_qp_status(struct hisi_qp *qp) { struct hisi_qp_status *qp_status = &qp->qp_status; qp_status->sq_tail = 0; qp_status->cq_head = 0; qp_status->cqc_phase = true; qp_status->flags = 0; } static void qm_vft_data_cfg(struct hisi_qm *qm, enum vft_type type, u32 base, u32 number) { u64 tmp = 0; if (number > 0) { switch (type) { case SQC_VFT: switch (qm->ver) { case QM_HW_V1: tmp = QM_SQC_VFT_BUF_SIZE | QM_SQC_VFT_SQC_SIZE | QM_SQC_VFT_INDEX_NUMBER | QM_SQC_VFT_VALID | (u64)base << QM_SQC_VFT_START_SQN_SHIFT; break; case QM_HW_V2: tmp = (u64)base << QM_SQC_VFT_START_SQN_SHIFT | QM_SQC_VFT_VALID | (u64)(number - 1) << QM_SQC_VFT_SQN_SHIFT; break; case QM_HW_UNKNOWN: break; } break; case CQC_VFT: switch (qm->ver) { case QM_HW_V1: tmp = QM_CQC_VFT_BUF_SIZE | QM_CQC_VFT_SQC_SIZE | QM_CQC_VFT_INDEX_NUMBER | QM_CQC_VFT_VALID; break; case QM_HW_V2: tmp = QM_CQC_VFT_VALID; break; case QM_HW_UNKNOWN: break; } break; } } writel(lower_32_bits(tmp), qm->io_base + QM_VFT_CFG_DATA_L); writel(upper_32_bits(tmp), qm->io_base + QM_VFT_CFG_DATA_H); } static int qm_set_vft_common(struct hisi_qm *qm, enum vft_type type, u32 fun_num, u32 base, u32 number) { unsigned int val; int ret; ret = readl_relaxed_poll_timeout(qm->io_base + QM_VFT_CFG_RDY, val, val & BIT(0), 10, 1000); if (ret) return ret; writel(0x0, qm->io_base + QM_VFT_CFG_OP_WR); writel(type, qm->io_base + QM_VFT_CFG_TYPE); writel(fun_num, qm->io_base + QM_VFT_CFG); qm_vft_data_cfg(qm, type, base, number); writel(0x0, qm->io_base + QM_VFT_CFG_RDY); writel(0x1, qm->io_base + QM_VFT_CFG_OP_ENABLE); return readl_relaxed_poll_timeout(qm->io_base + QM_VFT_CFG_RDY, val, val & BIT(0), 10, 1000); } /* The config should be conducted after qm_dev_mem_reset() */ static int qm_set_sqc_cqc_vft(struct hisi_qm *qm, u32 fun_num, u32 base, u32 number) { int ret, i; for (i = SQC_VFT; i <= CQC_VFT; i++) { ret = qm_set_vft_common(qm, i, fun_num, base, number); if (ret) return ret; } return 0; } static int qm_get_vft_v2(struct hisi_qm *qm, u32 *base, u32 *number) { u64 sqc_vft; int ret; ret = qm_mb(qm, QM_MB_CMD_SQC_VFT_V2, 0, 0, 1); if (ret) return ret; sqc_vft = readl(qm->io_base + QM_MB_CMD_DATA_ADDR_L) | ((u64)readl(qm->io_base + QM_MB_CMD_DATA_ADDR_H) << 32); *base = QM_SQC_VFT_BASE_MASK_V2 & (sqc_vft >> QM_SQC_VFT_BASE_SHIFT_V2); *number = (QM_SQC_VFT_NUM_MASK_v2 & (sqc_vft >> QM_SQC_VFT_NUM_SHIFT_V2)) + 1; return 0; } static struct hisi_qm *file_to_qm(struct debugfs_file *file) { struct qm_debug *debug = file->debug; return container_of(debug, struct hisi_qm, debug); } static u32 current_q_read(struct debugfs_file *file) { struct hisi_qm *qm = file_to_qm(file); return readl(qm->io_base + QM_DFX_SQE_CNT_VF_SQN) >> QM_DFX_QN_SHIFT; } static int current_q_write(struct debugfs_file *file, u32 val) { struct hisi_qm *qm = file_to_qm(file); u32 tmp; if (val >= qm->debug.curr_qm_qp_num) return -EINVAL; tmp = val << QM_DFX_QN_SHIFT | (readl(qm->io_base + QM_DFX_SQE_CNT_VF_SQN) & CURRENT_FUN_MASK); writel(tmp, qm->io_base + QM_DFX_SQE_CNT_VF_SQN); tmp = val << QM_DFX_QN_SHIFT | (readl(qm->io_base + QM_DFX_CQE_CNT_VF_CQN) & CURRENT_FUN_MASK); writel(tmp, qm->io_base + QM_DFX_CQE_CNT_VF_CQN); return 0; } static u32 clear_enable_read(struct debugfs_file *file) { struct hisi_qm *qm = file_to_qm(file); return readl(qm->io_base + QM_DFX_CNT_CLR_CE); } /* rd_clr_ctrl 1 enable read clear, otherwise 0 disable it */ static int clear_enable_write(struct debugfs_file *file, u32 rd_clr_ctrl) { struct hisi_qm *qm = file_to_qm(file); if (rd_clr_ctrl > 1) return -EINVAL; writel(rd_clr_ctrl, qm->io_base + QM_DFX_CNT_CLR_CE); return 0; } static ssize_t qm_debug_read(struct file *filp, char __user *buf, size_t count, loff_t *pos) { struct debugfs_file *file = filp->private_data; enum qm_debug_file index = file->index; char tbuf[QM_DBG_TMP_BUF_LEN]; u32 val; int ret; mutex_lock(&file->lock); switch (index) { case CURRENT_Q: val = current_q_read(file); break; case CLEAR_ENABLE: val = clear_enable_read(file); break; default: mutex_unlock(&file->lock); return -EINVAL; } mutex_unlock(&file->lock); ret = sprintf(tbuf, "%u\n", val); return simple_read_from_buffer(buf, count, pos, tbuf, ret); } static ssize_t qm_debug_write(struct file *filp, const char __user *buf, size_t count, loff_t *pos) { struct debugfs_file *file = filp->private_data; enum qm_debug_file index = file->index; unsigned long val; char tbuf[QM_DBG_TMP_BUF_LEN]; int len, ret; if (*pos != 0) return 0; if (count >= QM_DBG_TMP_BUF_LEN) return -ENOSPC; len = simple_write_to_buffer(tbuf, QM_DBG_TMP_BUF_LEN - 1, pos, buf, count); if (len < 0) return len; tbuf[len] = '\0'; if (kstrtoul(tbuf, 0, &val)) return -EFAULT; mutex_lock(&file->lock); switch (index) { case CURRENT_Q: ret = current_q_write(file, val); if (ret) goto err_input; break; case CLEAR_ENABLE: ret = clear_enable_write(file, val); if (ret) goto err_input; break; default: ret = -EINVAL; goto err_input; } mutex_unlock(&file->lock); return count; err_input: mutex_unlock(&file->lock); return ret; } static const struct file_operations qm_debug_fops = { .owner = THIS_MODULE, .open = simple_open, .read = qm_debug_read, .write = qm_debug_write, }; struct qm_dfx_registers { char *reg_name; u64 reg_offset; }; #define CNT_CYC_REGS_NUM 10 static struct qm_dfx_registers qm_dfx_regs[] = { /* XXX_CNT are reading clear register */ {"QM_ECC_1BIT_CNT ", 0x104000ull}, {"QM_ECC_MBIT_CNT ", 0x104008ull}, {"QM_DFX_MB_CNT ", 0x104018ull}, {"QM_DFX_DB_CNT ", 0x104028ull}, {"QM_DFX_SQE_CNT ", 0x104038ull}, {"QM_DFX_CQE_CNT ", 0x104048ull}, {"QM_DFX_SEND_SQE_TO_ACC_CNT ", 0x104050ull}, {"QM_DFX_WB_SQE_FROM_ACC_CNT ", 0x104058ull}, {"QM_DFX_ACC_FINISH_CNT ", 0x104060ull}, {"QM_DFX_CQE_ERR_CNT ", 0x1040b4ull}, {"QM_DFX_FUNS_ACTIVE_ST ", 0x200ull}, {"QM_ECC_1BIT_INF ", 0x104004ull}, {"QM_ECC_MBIT_INF ", 0x10400cull}, {"QM_DFX_ACC_RDY_VLD0 ", 0x1040a0ull}, {"QM_DFX_ACC_RDY_VLD1 ", 0x1040a4ull}, {"QM_DFX_AXI_RDY_VLD ", 0x1040a8ull}, {"QM_DFX_FF_ST0 ", 0x1040c8ull}, {"QM_DFX_FF_ST1 ", 0x1040ccull}, {"QM_DFX_FF_ST2 ", 0x1040d0ull}, {"QM_DFX_FF_ST3 ", 0x1040d4ull}, {"QM_DFX_FF_ST4 ", 0x1040d8ull}, {"QM_DFX_FF_ST5 ", 0x1040dcull}, {"QM_DFX_FF_ST6 ", 0x1040e0ull}, {"QM_IN_IDLE_ST ", 0x1040e4ull}, { NULL, 0} }; static struct qm_dfx_registers qm_vf_dfx_regs[] = { {"QM_DFX_FUNS_ACTIVE_ST ", 0x200ull}, { NULL, 0} }; static int qm_regs_show(struct seq_file *s, void *unused) { struct hisi_qm *qm = s->private; struct qm_dfx_registers *regs; u32 val; if (qm->fun_type == QM_HW_PF) regs = qm_dfx_regs; else regs = qm_vf_dfx_regs; while (regs->reg_name) { val = readl(qm->io_base + regs->reg_offset); seq_printf(s, "%s= 0x%08x\n", regs->reg_name, val); regs++; } return 0; } static int qm_regs_open(struct inode *inode, struct file *file) { return single_open(file, qm_regs_show, inode->i_private); } static const struct file_operations qm_regs_fops = { .owner = THIS_MODULE, .open = qm_regs_open, .read = seq_read, .release = single_release, }; static int qm_create_debugfs_file(struct hisi_qm *qm, enum qm_debug_file index) { struct dentry *qm_d = qm->debug.qm_d, *tmp; struct debugfs_file *file = qm->debug.files + index; tmp = debugfs_create_file(qm_debug_file_name[index], 0600, qm_d, file, &qm_debug_fops); if (IS_ERR(tmp)) return -ENOENT; file->index = index; mutex_init(&file->lock); file->debug = &qm->debug; return 0; } static void qm_hw_error_init_v1(struct hisi_qm *qm, u32 ce, u32 nfe, u32 fe, u32 msi) { writel(QM_ABNORMAL_INT_MASK_VALUE, qm->io_base + QM_ABNORMAL_INT_MASK); } static void qm_hw_error_init_v2(struct hisi_qm *qm, u32 ce, u32 nfe, u32 fe, u32 msi) { u32 irq_enable = ce | nfe | fe | msi; u32 irq_unmask = ~irq_enable; qm->error_mask = ce | nfe | fe; qm->msi_mask = msi; /* configure error type */ writel(ce, qm->io_base + QM_RAS_CE_ENABLE); writel(QM_RAS_CE_TIMES_PER_IRQ, qm->io_base + QM_RAS_CE_THRESHOLD); writel(nfe, qm->io_base + QM_RAS_NFE_ENABLE); writel(fe, qm->io_base + QM_RAS_FE_ENABLE); /* use RAS irq default, so only set QM_RAS_MSI_INT_SEL for MSI */ writel(msi, qm->io_base + QM_RAS_MSI_INT_SEL); irq_unmask &= readl(qm->io_base + QM_ABNORMAL_INT_MASK); writel(irq_unmask, qm->io_base + QM_ABNORMAL_INT_MASK); } static void qm_log_hw_error(struct hisi_qm *qm, u32 error_status) { const struct hisi_qm_hw_error *err = qm_hw_error; struct device *dev = &qm->pdev->dev; u32 reg_val, type, vf_num; while (err->msg) { if (err->int_msk & error_status) { dev_err(dev, "%s [error status=0x%x] found\n", err->msg, err->int_msk); if (error_status & QM_DB_TIMEOUT) { reg_val = readl(qm->io_base + QM_ABNORMAL_INF01); type = (reg_val & QM_DB_TIMEOUT_TYPE) >> QM_DB_TIMEOUT_TYPE_SHIFT; vf_num = reg_val & QM_DB_TIMEOUT_VF; dev_err(dev, "qm %s doorbell timeout in function %u\n", qm_db_timeout[type], vf_num); } if (error_status & QM_OF_FIFO_OF) { reg_val = readl(qm->io_base + QM_ABNORMAL_INF00); type = (reg_val & QM_FIFO_OVERFLOW_TYPE) >> QM_FIFO_OVERFLOW_TYPE_SHIFT; vf_num = reg_val & QM_FIFO_OVERFLOW_VF; if (type < ARRAY_SIZE(qm_fifo_overflow)) dev_err(dev, "qm %s fifo overflow in function %u\n", qm_fifo_overflow[type], vf_num); else dev_err(dev, "unknown error type\n"); } } err++; } } static pci_ers_result_t qm_hw_error_handle_v2(struct hisi_qm *qm) { u32 error_status, tmp; /* read err sts */ tmp = readl(qm->io_base + QM_ABNORMAL_INT_STATUS); error_status = qm->error_mask & tmp; if (error_status) { qm_log_hw_error(qm, error_status); /* clear err sts */ writel(error_status, qm->io_base + QM_ABNORMAL_INT_SOURCE); return PCI_ERS_RESULT_NEED_RESET; } return PCI_ERS_RESULT_RECOVERED; } static const struct hisi_qm_hw_ops qm_hw_ops_v1 = { .qm_db = qm_db_v1, .get_irq_num = qm_get_irq_num_v1, .hw_error_init = qm_hw_error_init_v1, }; static const struct hisi_qm_hw_ops qm_hw_ops_v2 = { .get_vft = qm_get_vft_v2, .qm_db = qm_db_v2, .get_irq_num = qm_get_irq_num_v2, .hw_error_init = qm_hw_error_init_v2, .hw_error_handle = qm_hw_error_handle_v2, }; static void *qm_get_avail_sqe(struct hisi_qp *qp) { struct hisi_qp_status *qp_status = &qp->qp_status; u16 sq_tail = qp_status->sq_tail; if (unlikely(atomic_read(&qp->qp_status.used) == QM_Q_DEPTH)) return NULL; return qp->sqe + sq_tail * qp->qm->sqe_size; } /** * hisi_qm_create_qp() - Create a queue pair from qm. * @qm: The qm we create a qp from. * @alg_type: Accelerator specific algorithm type in sqc. * * return created qp, -EBUSY if all qps in qm allocated, -ENOMEM if allocating * qp memory fails. */ struct hisi_qp *hisi_qm_create_qp(struct hisi_qm *qm, u8 alg_type) { struct device *dev = &qm->pdev->dev; struct hisi_qp *qp; int qp_id, ret; qp = kzalloc(sizeof(*qp), GFP_KERNEL); if (!qp) return ERR_PTR(-ENOMEM); write_lock(&qm->qps_lock); qp_id = find_first_zero_bit(qm->qp_bitmap, qm->qp_num); if (qp_id >= qm->qp_num) { write_unlock(&qm->qps_lock); dev_info(&qm->pdev->dev, "QM all queues are busy!\n"); ret = -EBUSY; goto err_free_qp; } set_bit(qp_id, qm->qp_bitmap); qm->qp_array[qp_id] = qp; write_unlock(&qm->qps_lock); qp->qm = qm; if (qm->use_dma_api) { qp->qdma.size = qm->sqe_size * QM_Q_DEPTH + sizeof(struct qm_cqe) * QM_Q_DEPTH; qp->qdma.va = dma_alloc_coherent(dev, qp->qdma.size, &qp->qdma.dma, GFP_KERNEL); if (!qp->qdma.va) { ret = -ENOMEM; goto err_clear_bit; } dev_dbg(dev, "allocate qp dma buf(va=%pK, dma=%pad, size=%zx)\n", qp->qdma.va, &qp->qdma.dma, qp->qdma.size); } qp->qp_id = qp_id; qp->alg_type = alg_type; INIT_WORK(&qp->work, qm_qp_work_func); qp->wq = alloc_workqueue("hisi_qm", WQ_UNBOUND | WQ_HIGHPRI | WQ_CPU_INTENSIVE | WQ_MEM_RECLAIM, 0); if (!qp->wq) { ret = -EFAULT; goto err_free_qp_mem; } return qp; err_free_qp_mem: if (qm->use_dma_api) dma_free_coherent(dev, qp->qdma.size, qp->qdma.va, qp->qdma.dma); err_clear_bit: write_lock(&qm->qps_lock); qm->qp_array[qp_id] = NULL; clear_bit(qp_id, qm->qp_bitmap); write_unlock(&qm->qps_lock); err_free_qp: kfree(qp); return ERR_PTR(ret); } EXPORT_SYMBOL_GPL(hisi_qm_create_qp); /** * hisi_qm_release_qp() - Release a qp back to its qm. * @qp: The qp we want to release. * * This function releases the resource of a qp. */ void hisi_qm_release_qp(struct hisi_qp *qp) { struct hisi_qm *qm = qp->qm; struct qm_dma *qdma = &qp->qdma; struct device *dev = &qm->pdev->dev; if (qm->use_dma_api && qdma->va) dma_free_coherent(dev, qdma->size, qdma->va, qdma->dma); write_lock(&qm->qps_lock); qm->qp_array[qp->qp_id] = NULL; clear_bit(qp->qp_id, qm->qp_bitmap); write_unlock(&qm->qps_lock); kfree(qp); } EXPORT_SYMBOL_GPL(hisi_qm_release_qp); static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, int pasid) { struct hisi_qm *qm = qp->qm; struct device *dev = &qm->pdev->dev; enum qm_hw_ver ver = qm->ver; struct qm_sqc *sqc; struct qm_cqc *cqc; dma_addr_t sqc_dma; dma_addr_t cqc_dma; int ret; qm_init_qp_status(qp); sqc = kzalloc(sizeof(struct qm_sqc), GFP_KERNEL); if (!sqc) return -ENOMEM; sqc_dma = dma_map_single(dev, sqc, sizeof(struct qm_sqc), DMA_TO_DEVICE); if (dma_mapping_error(dev, sqc_dma)) { kfree(sqc); return -ENOMEM; } INIT_QC_COMMON(sqc, qp->sqe_dma, pasid); if (ver == QM_HW_V1) { sqc->dw3 = cpu_to_le32(QM_MK_SQC_DW3_V1(0, 0, 0, qm->sqe_size)); sqc->w8 = cpu_to_le16(QM_Q_DEPTH - 1); } else if (ver == QM_HW_V2) { sqc->dw3 = cpu_to_le32(QM_MK_SQC_DW3_V2(qm->sqe_size)); sqc->w8 = 0; /* rand_qc */ } sqc->cq_num = cpu_to_le16(qp_id); sqc->w13 = cpu_to_le16(QM_MK_SQC_W13(0, 1, qp->alg_type)); ret = qm_mb(qm, QM_MB_CMD_SQC, sqc_dma, qp_id, 0); dma_unmap_single(dev, sqc_dma, sizeof(struct qm_sqc), DMA_TO_DEVICE); kfree(sqc); if (ret) return ret; cqc = kzalloc(sizeof(struct qm_cqc), GFP_KERNEL); if (!cqc) return -ENOMEM; cqc_dma = dma_map_single(dev, cqc, sizeof(struct qm_cqc), DMA_TO_DEVICE); if (dma_mapping_error(dev, cqc_dma)) { kfree(cqc); return -ENOMEM; } INIT_QC_COMMON(cqc, qp->cqe_dma, pasid); if (ver == QM_HW_V1) { cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V1(0, 0, 0, 4)); cqc->w8 = cpu_to_le16(QM_Q_DEPTH - 1); } else if (ver == QM_HW_V2) { cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V2(4)); cqc->w8 = 0; } cqc->dw6 = cpu_to_le32(1 << QM_CQ_PHASE_SHIFT | 1 << QM_CQ_FLAG_SHIFT); ret = qm_mb(qm, QM_MB_CMD_CQC, cqc_dma, qp_id, 0); dma_unmap_single(dev, cqc_dma, sizeof(struct qm_cqc), DMA_TO_DEVICE); kfree(cqc); return ret; } /** * hisi_qm_start_qp() - Start a qp into running. * @qp: The qp we want to start to run. * @arg: Accelerator specific argument. * * After this function, qp can receive request from user. Return qp_id if * successful, Return -EBUSY if failed. */ int hisi_qm_start_qp(struct hisi_qp *qp, unsigned long arg) { struct hisi_qm *qm = qp->qm; struct device *dev = &qm->pdev->dev; enum qm_hw_ver ver = qm->ver; int qp_id = qp->qp_id; int pasid = arg; size_t off = 0; int ret; #define QP_INIT_BUF(qp, type, size) do { \ (qp)->type = ((qp)->qdma.va + (off)); \ (qp)->type##_dma = (qp)->qdma.dma + (off); \ off += (size); \ } while (0) if (!qp->qdma.dma) { dev_err(dev, "cannot get qm dma buffer\n"); return -EINVAL; } /* sq need 128 bytes alignment */ if (qp->qdma.dma & QM_SQE_DATA_ALIGN_MASK) { dev_err(dev, "qm sq is not aligned to 128 byte\n"); return -EINVAL; } QP_INIT_BUF(qp, sqe, qm->sqe_size * QM_Q_DEPTH); QP_INIT_BUF(qp, cqe, sizeof(struct qm_cqe) * QM_Q_DEPTH); dev_dbg(dev, "init qp buffer(v%d):\n" " sqe (%pK, %lx)\n" " cqe (%pK, %lx)\n", ver, qp->sqe, (unsigned long)qp->sqe_dma, qp->cqe, (unsigned long)qp->cqe_dma); ret = qm_qp_ctx_cfg(qp, qp_id, pasid); if (ret) return ret; dev_dbg(dev, "queue %d started\n", qp_id); return qp_id; } EXPORT_SYMBOL_GPL(hisi_qm_start_qp); /** * hisi_qm_stop_qp() - Stop a qp in qm. * @qp: The qp we want to stop. * * This function is reverse of hisi_qm_start_qp. Return 0 if successful. */ int hisi_qm_stop_qp(struct hisi_qp *qp) { struct device *dev = &qp->qm->pdev->dev; int i = 0; /* it is stopped */ if (test_bit(QP_STOP, &qp->qp_status.flags)) return 0; while (atomic_read(&qp->qp_status.used)) { i++; msleep(20); if (i == 10) { dev_err(dev, "Cannot drain out data for stopping, Force to stop!\n"); return 0; } } set_bit(QP_STOP, &qp->qp_status.flags); dev_dbg(dev, "stop queue %u!", qp->qp_id); return 0; } EXPORT_SYMBOL_GPL(hisi_qm_stop_qp); /** * hisi_qp_send() - Queue up a task in the hardware queue. * @qp: The qp in which to put the message. * @msg: The message. * * This function will return -EBUSY if qp is currently full, and -EAGAIN * if qp related qm is resetting. */ int hisi_qp_send(struct hisi_qp *qp, const void *msg) { struct hisi_qp_status *qp_status = &qp->qp_status; u16 sq_tail = qp_status->sq_tail; u16 sq_tail_next = (sq_tail + 1) % QM_Q_DEPTH; void *sqe = qm_get_avail_sqe(qp); if (unlikely(test_bit(QP_STOP, &qp->qp_status.flags))) { dev_info(&qp->qm->pdev->dev, "QP is stopped or resetting\n"); return -EAGAIN; } if (!sqe) return -EBUSY; memcpy(sqe, msg, qp->qm->sqe_size); qm_db(qp->qm, qp->qp_id, QM_DOORBELL_CMD_SQ, sq_tail_next, 0); atomic_inc(&qp->qp_status.used); qp_status->sq_tail = sq_tail_next; return 0; } EXPORT_SYMBOL_GPL(hisi_qp_send); static void hisi_qm_cache_wb(struct hisi_qm *qm) { unsigned int val; if (qm->ver == QM_HW_V2) { writel(0x1, qm->io_base + QM_CACHE_WB_START); if (readl_relaxed_poll_timeout(qm->io_base + QM_CACHE_WB_DONE, val, val & BIT(0), 10, 1000)) dev_err(&qm->pdev->dev, "QM writeback sqc cache fail!\n"); } } /** * hisi_qm_init() - Initialize configures about qm. * @qm: The qm needing init. * * This function init qm, then we can call hisi_qm_start to put qm into work. */ int hisi_qm_init(struct hisi_qm *qm) { struct pci_dev *pdev = qm->pdev; struct device *dev = &pdev->dev; unsigned int num_vec; int ret; switch (qm->ver) { case QM_HW_V1: qm->ops = &qm_hw_ops_v1; break; case QM_HW_V2: qm->ops = &qm_hw_ops_v2; break; default: return -EINVAL; } ret = pci_enable_device_mem(pdev); if (ret < 0) { dev_err(&pdev->dev, "Failed to enable device mem!\n"); return ret; } ret = pci_request_mem_regions(pdev, qm->dev_name); if (ret < 0) { dev_err(&pdev->dev, "Failed to request mem regions!\n"); goto err_disable_pcidev; } qm->io_base = ioremap(pci_resource_start(pdev, PCI_BAR_2), pci_resource_len(qm->pdev, PCI_BAR_2)); if (!qm->io_base) { ret = -EIO; goto err_release_mem_regions; } ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64)); if (ret < 0) goto err_iounmap; pci_set_master(pdev); if (!qm->ops->get_irq_num) { ret = -EOPNOTSUPP; goto err_iounmap; } num_vec = qm->ops->get_irq_num(qm); ret = pci_alloc_irq_vectors(pdev, num_vec, num_vec, PCI_IRQ_MSI); if (ret < 0) { dev_err(dev, "Failed to enable MSI vectors!\n"); goto err_iounmap; } ret = qm_irq_register(qm); if (ret) goto err_free_irq_vectors; mutex_init(&qm->mailbox_lock); rwlock_init(&qm->qps_lock); dev_dbg(dev, "init qm %s with %s\n", pdev->is_physfn ? "pf" : "vf", qm->use_dma_api ? "dma api" : "iommu api"); return 0; err_free_irq_vectors: pci_free_irq_vectors(pdev); err_iounmap: iounmap(qm->io_base); err_release_mem_regions: pci_release_mem_regions(pdev); err_disable_pcidev: pci_disable_device(pdev); return ret; } EXPORT_SYMBOL_GPL(hisi_qm_init); /** * hisi_qm_uninit() - Uninitialize qm. * @qm: The qm needed uninit. * * This function uninits qm related device resources. */ void hisi_qm_uninit(struct hisi_qm *qm) { struct pci_dev *pdev = qm->pdev; struct device *dev = &pdev->dev; if (qm->use_dma_api && qm->qdma.va) { hisi_qm_cache_wb(qm); dma_free_coherent(dev, qm->qdma.size, qm->qdma.va, qm->qdma.dma); memset(&qm->qdma, 0, sizeof(qm->qdma)); } qm_irq_unregister(qm); pci_free_irq_vectors(pdev); iounmap(qm->io_base); pci_release_mem_regions(pdev); pci_disable_device(pdev); } EXPORT_SYMBOL_GPL(hisi_qm_uninit); /** * hisi_qm_get_vft() - Get vft from a qm. * @qm: The qm we want to get its vft. * @base: The base number of queue in vft. * @number: The number of queues in vft. * * We can allocate multiple queues to a qm by configuring virtual function * table. We get related configures by this function. Normally, we call this * function in VF driver to get the queue information. * * qm hw v1 does not support this interface. */ int hisi_qm_get_vft(struct hisi_qm *qm, u32 *base, u32 *number) { if (!base || !number) return -EINVAL; if (!qm->ops->get_vft) { dev_err(&qm->pdev->dev, "Don't support vft read!\n"); return -EINVAL; } return qm->ops->get_vft(qm, base, number); } EXPORT_SYMBOL_GPL(hisi_qm_get_vft); /** * hisi_qm_set_vft() - Set "virtual function table" for a qm. * @fun_num: Number of operated function. * @qm: The qm in which to set vft, alway in a PF. * @base: The base number of queue in vft. * @number: The number of queues in vft. 0 means invalid vft. * * This function is alway called in PF driver, it is used to assign queues * among PF and VFs. * * Assign queues A~B to PF: hisi_qm_set_vft(qm, 0, A, B - A + 1) * Assign queues A~B to VF: hisi_qm_set_vft(qm, 2, A, B - A + 1) * (VF function number 0x2) */ int hisi_qm_set_vft(struct hisi_qm *qm, u32 fun_num, u32 base, u32 number) { u32 max_q_num = qm->ctrl_qp_num; if (base >= max_q_num || number > max_q_num || (base + number) > max_q_num) return -EINVAL; return qm_set_sqc_cqc_vft(qm, fun_num, base, number); } EXPORT_SYMBOL_GPL(hisi_qm_set_vft); static void qm_init_eq_aeq_status(struct hisi_qm *qm) { struct hisi_qm_status *status = &qm->status; status->eq_head = 0; status->aeq_head = 0; status->eqc_phase = true; status->aeqc_phase = true; } static int qm_eq_ctx_cfg(struct hisi_qm *qm) { struct device *dev = &qm->pdev->dev; struct qm_eqc *eqc; struct qm_aeqc *aeqc; dma_addr_t eqc_dma; dma_addr_t aeqc_dma; int ret; qm_init_eq_aeq_status(qm); eqc = kzalloc(sizeof(struct qm_eqc), GFP_KERNEL); if (!eqc) return -ENOMEM; eqc_dma = dma_map_single(dev, eqc, sizeof(struct qm_eqc), DMA_TO_DEVICE); if (dma_mapping_error(dev, eqc_dma)) { kfree(eqc); return -ENOMEM; } eqc->base_l = cpu_to_le32(lower_32_bits(qm->eqe_dma)); eqc->base_h = cpu_to_le32(upper_32_bits(qm->eqe_dma)); if (qm->ver == QM_HW_V1) eqc->dw3 = cpu_to_le32(QM_EQE_AEQE_SIZE); eqc->dw6 = cpu_to_le32((QM_Q_DEPTH - 1) | (1 << QM_EQC_PHASE_SHIFT)); ret = qm_mb(qm, QM_MB_CMD_EQC, eqc_dma, 0, 0); dma_unmap_single(dev, eqc_dma, sizeof(struct qm_eqc), DMA_TO_DEVICE); kfree(eqc); if (ret) return ret; aeqc = kzalloc(sizeof(struct qm_aeqc), GFP_KERNEL); if (!aeqc) return -ENOMEM; aeqc_dma = dma_map_single(dev, aeqc, sizeof(struct qm_aeqc), DMA_TO_DEVICE); if (dma_mapping_error(dev, aeqc_dma)) { kfree(aeqc); return -ENOMEM; } aeqc->base_l = cpu_to_le32(lower_32_bits(qm->aeqe_dma)); aeqc->base_h = cpu_to_le32(upper_32_bits(qm->aeqe_dma)); aeqc->dw6 = cpu_to_le32((QM_Q_DEPTH - 1) | (1 << QM_EQC_PHASE_SHIFT)); ret = qm_mb(qm, QM_MB_CMD_AEQC, aeqc_dma, 0, 0); dma_unmap_single(dev, aeqc_dma, sizeof(struct qm_aeqc), DMA_TO_DEVICE); kfree(aeqc); return ret; } static int __hisi_qm_start(struct hisi_qm *qm) { struct pci_dev *pdev = qm->pdev; struct device *dev = &pdev->dev; size_t off = 0; int ret; #define QM_INIT_BUF(qm, type, num) do { \ (qm)->type = ((qm)->qdma.va + (off)); \ (qm)->type##_dma = (qm)->qdma.dma + (off); \ off += QMC_ALIGN(sizeof(struct qm_##type) * (num)); \ } while (0) WARN_ON(!qm->qdma.dma); if (qm->qp_num == 0) return -EINVAL; if (qm->fun_type == QM_HW_PF) { ret = qm_dev_mem_reset(qm); if (ret) return ret; ret = hisi_qm_set_vft(qm, 0, qm->qp_base, qm->qp_num); if (ret) return ret; } QM_INIT_BUF(qm, eqe, QM_Q_DEPTH); QM_INIT_BUF(qm, aeqe, QM_Q_DEPTH); QM_INIT_BUF(qm, sqc, qm->qp_num); QM_INIT_BUF(qm, cqc, qm->qp_num); dev_dbg(dev, "init qm buffer:\n" " eqe (%pK, %lx)\n" " aeqe (%pK, %lx)\n" " sqc (%pK, %lx)\n" " cqc (%pK, %lx)\n", qm->eqe, (unsigned long)qm->eqe_dma, qm->aeqe, (unsigned long)qm->aeqe_dma, qm->sqc, (unsigned long)qm->sqc_dma, qm->cqc, (unsigned long)qm->cqc_dma); ret = qm_eq_ctx_cfg(qm); if (ret) return ret; ret = qm_mb(qm, QM_MB_CMD_SQC_BT, qm->sqc_dma, 0, 0); if (ret) return ret; ret = qm_mb(qm, QM_MB_CMD_CQC_BT, qm->cqc_dma, 0, 0); if (ret) return ret; writel(0x0, qm->io_base + QM_VF_EQ_INT_MASK); writel(0x0, qm->io_base + QM_VF_AEQ_INT_MASK); return 0; } /** * hisi_qm_start() - start qm * @qm: The qm to be started. * * This function starts a qm, then we can allocate qp from this qm. */ int hisi_qm_start(struct hisi_qm *qm) { struct device *dev = &qm->pdev->dev; dev_dbg(dev, "qm start with %d queue pairs\n", qm->qp_num); if (!qm->qp_num) { dev_err(dev, "qp_num should not be 0\n"); return -EINVAL; } if (!qm->qp_bitmap) { qm->qp_bitmap = devm_kcalloc(dev, BITS_TO_LONGS(qm->qp_num), sizeof(long), GFP_KERNEL); qm->qp_array = devm_kcalloc(dev, qm->qp_num, sizeof(struct hisi_qp *), GFP_KERNEL); if (!qm->qp_bitmap || !qm->qp_array) return -ENOMEM; } if (!qm->use_dma_api) { dev_dbg(&qm->pdev->dev, "qm delay start\n"); return 0; } else if (!qm->qdma.va) { qm->qdma.size = QMC_ALIGN(sizeof(struct qm_eqe) * QM_Q_DEPTH) + QMC_ALIGN(sizeof(struct qm_aeqe) * QM_Q_DEPTH) + QMC_ALIGN(sizeof(struct qm_sqc) * qm->qp_num) + QMC_ALIGN(sizeof(struct qm_cqc) * qm->qp_num); qm->qdma.va = dma_alloc_coherent(dev, qm->qdma.size, &qm->qdma.dma, GFP_KERNEL); dev_dbg(dev, "allocate qm dma buf(va=%pK, dma=%pad, size=%zx)\n", qm->qdma.va, &qm->qdma.dma, qm->qdma.size); if (!qm->qdma.va) return -ENOMEM; } return __hisi_qm_start(qm); } EXPORT_SYMBOL_GPL(hisi_qm_start); /** * hisi_qm_stop() - Stop a qm. * @qm: The qm which will be stopped. * * This function stops qm and its qps, then qm can not accept request. * Related resources are not released at this state, we can use hisi_qm_start * to let qm start again. */ int hisi_qm_stop(struct hisi_qm *qm) { struct device *dev; struct hisi_qp *qp; int ret = 0, i; if (!qm || !qm->pdev) { WARN_ON(1); return -EINVAL; } dev = &qm->pdev->dev; /* Mask eq and aeq irq */ writel(0x1, qm->io_base + QM_VF_EQ_INT_MASK); writel(0x1, qm->io_base + QM_VF_AEQ_INT_MASK); /* Stop all qps belong to this qm */ for (i = 0; i < qm->qp_num; i++) { qp = qm->qp_array[i]; if (qp) { ret = hisi_qm_stop_qp(qp); if (ret < 0) { dev_err(dev, "Failed to stop qp%d!\n", i); return -EBUSY; } } } if (qm->fun_type == QM_HW_PF) { ret = hisi_qm_set_vft(qm, 0, 0, 0); if (ret < 0) dev_err(dev, "Failed to set vft!\n"); } return ret; } EXPORT_SYMBOL_GPL(hisi_qm_stop); /** * hisi_qm_debug_init() - Initialize qm related debugfs files. * @qm: The qm for which we want to add debugfs files. * * Create qm related debugfs files. */ int hisi_qm_debug_init(struct hisi_qm *qm) { struct dentry *qm_d, *qm_regs; int i, ret; qm_d = debugfs_create_dir("qm", qm->debug.debug_root); if (IS_ERR(qm_d)) return -ENOENT; qm->debug.qm_d = qm_d; /* only show this in PF */ if (qm->fun_type == QM_HW_PF) for (i = CURRENT_Q; i < DEBUG_FILE_NUM; i++) if (qm_create_debugfs_file(qm, i)) { ret = -ENOENT; goto failed_to_create; } qm_regs = debugfs_create_file("qm_regs", 0444, qm->debug.qm_d, qm, &qm_regs_fops); if (IS_ERR(qm_regs)) { ret = -ENOENT; goto failed_to_create; } return 0; failed_to_create: debugfs_remove_recursive(qm_d); return ret; } EXPORT_SYMBOL_GPL(hisi_qm_debug_init); /** * hisi_qm_debug_regs_clear() - clear qm debug related registers. * @qm: The qm for which we want to clear its debug registers. */ void hisi_qm_debug_regs_clear(struct hisi_qm *qm) { struct qm_dfx_registers *regs; int i; /* clear current_q */ writel(0x0, qm->io_base + QM_DFX_SQE_CNT_VF_SQN); writel(0x0, qm->io_base + QM_DFX_CQE_CNT_VF_CQN); /* * these registers are reading and clearing, so clear them after * reading them. */ writel(0x1, qm->io_base + QM_DFX_CNT_CLR_CE); regs = qm_dfx_regs; for (i = 0; i < CNT_CYC_REGS_NUM; i++) { readl(qm->io_base + regs->reg_offset); regs++; } writel(0x0, qm->io_base + QM_DFX_CNT_CLR_CE); } EXPORT_SYMBOL_GPL(hisi_qm_debug_regs_clear); /** * hisi_qm_hw_error_init() - Configure qm hardware error report method. * @qm: The qm which we want to configure. * @ce: Bit mask of correctable error configure. * @nfe: Bit mask of non-fatal error configure. * @fe: Bit mask of fatal error configure. * @msi: Bit mask of error reported by message signal interrupt. * * Hardware errors of qm can be reported either by RAS interrupts which will * be handled by UEFI and then PCIe AER or by device MSI. User can configure * each error to use either of above two methods. For RAS interrupts, we can * configure an error as one of correctable error, non-fatal error or * fatal error. * * Bits indicating errors can be configured to ce, nfe, fe and msi to enable * related report methods. Error report will be masked if related error bit * does not configure. */ void hisi_qm_hw_error_init(struct hisi_qm *qm, u32 ce, u32 nfe, u32 fe, u32 msi) { if (!qm->ops->hw_error_init) { dev_err(&qm->pdev->dev, "QM doesn't support hw error handling!\n"); return; } qm->ops->hw_error_init(qm, ce, nfe, fe, msi); } EXPORT_SYMBOL_GPL(hisi_qm_hw_error_init); /** * hisi_qm_hw_error_handle() - Handle qm non-fatal hardware errors. * @qm: The qm which has non-fatal hardware errors. * * Accelerators use this function to handle qm non-fatal hardware errors. */ pci_ers_result_t hisi_qm_hw_error_handle(struct hisi_qm *qm) { if (!qm->ops->hw_error_handle) { dev_err(&qm->pdev->dev, "QM doesn't support hw error report!\n"); return PCI_ERS_RESULT_NONE; } return qm->ops->hw_error_handle(qm); } EXPORT_SYMBOL_GPL(hisi_qm_hw_error_handle); /** * hisi_qm_get_hw_version() - Get hardware version of a qm. * @pdev: The device which hardware version we want to get. * * This function gets the hardware version of a qm. Return QM_HW_UNKNOWN * if the hardware version is not supported. */ enum qm_hw_ver hisi_qm_get_hw_version(struct pci_dev *pdev) { switch (pdev->revision) { case QM_HW_V1: case QM_HW_V2: return pdev->revision; default: return QM_HW_UNKNOWN; } } EXPORT_SYMBOL_GPL(hisi_qm_get_hw_version); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Zhou Wang "); MODULE_DESCRIPTION("HiSilicon Accelerator queue manager driver");