diff options
Diffstat (limited to 'drivers/misc/habanalabs')
-rw-r--r-- | drivers/misc/habanalabs/command_submission.c | 5 | ||||
-rw-r--r-- | drivers/misc/habanalabs/context.c | 2 | ||||
-rw-r--r-- | drivers/misc/habanalabs/device.c | 1 | ||||
-rw-r--r-- | drivers/misc/habanalabs/goya/goya.c | 59 | ||||
-rw-r--r-- | drivers/misc/habanalabs/habanalabs.h | 2 | ||||
-rw-r--r-- | drivers/misc/habanalabs/memory.c | 30 |
6 files changed, 75 insertions, 24 deletions
diff --git a/drivers/misc/habanalabs/command_submission.c b/drivers/misc/habanalabs/command_submission.c index 6fe785e26859..52fd0229c4f7 100644 --- a/drivers/misc/habanalabs/command_submission.c +++ b/drivers/misc/habanalabs/command_submission.c @@ -764,8 +764,9 @@ int hl_cs_wait_ioctl(struct hl_fpriv *hpriv, void *data) memset(args, 0, sizeof(*args)); if (rc < 0) { - dev_err(hdev->dev, "Error %ld on waiting for CS handle %llu\n", - rc, seq); + dev_err_ratelimited(hdev->dev, + "Error %ld on waiting for CS handle %llu\n", + rc, seq); if (rc == -ERESTARTSYS) { args->out.status = HL_WAIT_CS_STATUS_INTERRUPTED; rc = -EINTR; diff --git a/drivers/misc/habanalabs/context.c b/drivers/misc/habanalabs/context.c index f4c92f110a72..4d5a0454208e 100644 --- a/drivers/misc/habanalabs/context.c +++ b/drivers/misc/habanalabs/context.c @@ -159,7 +159,7 @@ struct dma_fence *hl_ctx_get_fence(struct hl_ctx *ctx, u64 seq) spin_lock(&ctx->cs_lock); if (seq >= ctx->cs_sequence) { - dev_notice(hdev->dev, + dev_notice_ratelimited(hdev->dev, "Can't wait on seq %llu because current CS is at seq %llu\n", seq, ctx->cs_sequence); spin_unlock(&ctx->cs_lock); diff --git a/drivers/misc/habanalabs/device.c b/drivers/misc/habanalabs/device.c index 0b19d3eefb98..065cb7f127fb 100644 --- a/drivers/misc/habanalabs/device.c +++ b/drivers/misc/habanalabs/device.c @@ -934,6 +934,7 @@ int hl_device_init(struct hl_device *hdev, struct class *hclass) if (hdev->asic_funcs->get_hw_state(hdev) == HL_DEVICE_HW_STATE_DIRTY) { dev_info(hdev->dev, "H/W state is dirty, must reset before initializing\n"); + hdev->asic_funcs->halt_engines(hdev, true); hdev->asic_funcs->hw_fini(hdev, true); } diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c index ac6b252a1ddc..86f7b6dcfa76 100644 --- a/drivers/misc/habanalabs/goya/goya.c +++ b/drivers/misc/habanalabs/goya/goya.c @@ -826,6 +826,11 @@ void goya_init_dma_qmans(struct hl_device *hdev) */ static void goya_disable_external_queues(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_DMA)) + return; + WREG32(mmDMA_QM_0_GLBL_CFG0, 0); WREG32(mmDMA_QM_1_GLBL_CFG0, 0); WREG32(mmDMA_QM_2_GLBL_CFG0, 0); @@ -887,6 +892,11 @@ static int goya_stop_external_queues(struct hl_device *hdev) { int rc, retval = 0; + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_DMA)) + return retval; + rc = goya_stop_queue(hdev, mmDMA_QM_0_GLBL_CFG1, mmDMA_QM_0_CP_STS, @@ -1678,9 +1688,18 @@ void goya_init_tpc_qmans(struct hl_device *hdev) */ static void goya_disable_internal_queues(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_MME)) + goto disable_tpc; + WREG32(mmMME_QM_GLBL_CFG0, 0); WREG32(mmMME_CMDQ_GLBL_CFG0, 0); +disable_tpc: + if (!(goya->hw_cap_initialized & HW_CAP_TPC)) + return; + WREG32(mmTPC0_QM_GLBL_CFG0, 0); WREG32(mmTPC0_CMDQ_GLBL_CFG0, 0); @@ -1716,8 +1735,12 @@ static void goya_disable_internal_queues(struct hl_device *hdev) */ static int goya_stop_internal_queues(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; int rc, retval = 0; + if (!(goya->hw_cap_initialized & HW_CAP_MME)) + goto stop_tpc; + /* * Each queue (QMAN) is a separate H/W logic. That means that each * QMAN can be stopped independently and failure to stop one does NOT @@ -1744,6 +1767,10 @@ static int goya_stop_internal_queues(struct hl_device *hdev) retval = -EIO; } +stop_tpc: + if (!(goya->hw_cap_initialized & HW_CAP_TPC)) + return retval; + rc = goya_stop_queue(hdev, mmTPC0_QM_GLBL_CFG1, mmTPC0_QM_CP_STS, @@ -1909,6 +1936,11 @@ static int goya_stop_internal_queues(struct hl_device *hdev) static void goya_dma_stall(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_DMA)) + return; + WREG32(mmDMA_QM_0_GLBL_CFG1, 1 << DMA_QM_0_GLBL_CFG1_DMA_STOP_SHIFT); WREG32(mmDMA_QM_1_GLBL_CFG1, 1 << DMA_QM_1_GLBL_CFG1_DMA_STOP_SHIFT); WREG32(mmDMA_QM_2_GLBL_CFG1, 1 << DMA_QM_2_GLBL_CFG1_DMA_STOP_SHIFT); @@ -1918,6 +1950,11 @@ static void goya_dma_stall(struct hl_device *hdev) static void goya_tpc_stall(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_TPC)) + return; + WREG32(mmTPC0_CFG_TPC_STALL, 1 << TPC0_CFG_TPC_STALL_V_SHIFT); WREG32(mmTPC1_CFG_TPC_STALL, 1 << TPC1_CFG_TPC_STALL_V_SHIFT); WREG32(mmTPC2_CFG_TPC_STALL, 1 << TPC2_CFG_TPC_STALL_V_SHIFT); @@ -1930,6 +1967,11 @@ static void goya_tpc_stall(struct hl_device *hdev) static void goya_mme_stall(struct hl_device *hdev) { + struct goya_device *goya = hdev->asic_specific; + + if (!(goya->hw_cap_initialized & HW_CAP_MME)) + return; + WREG32(mmMME_STALL, 0xFFFFFFFF); } @@ -2107,7 +2149,7 @@ static int goya_push_linux_to_device(struct hl_device *hdev) static int goya_pldm_init_cpu(struct hl_device *hdev) { - u32 val, unit_rst_val; + u32 unit_rst_val; int rc; /* Must initialize SRAM scrambler before pushing u-boot to SRAM */ @@ -2115,14 +2157,14 @@ static int goya_pldm_init_cpu(struct hl_device *hdev) /* Put ARM cores into reset */ WREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL, CPU_RESET_ASSERT); - val = RREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL); + RREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL); /* Reset the CA53 MACRO */ unit_rst_val = RREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N); WREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N, CA53_RESET); - val = RREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N); + RREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N); WREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N, unit_rst_val); - val = RREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N); + RREG32(mmPSOC_GLOBAL_CONF_UNIT_RST_N); rc = goya_push_uboot_to_device(hdev); if (rc) @@ -2143,7 +2185,7 @@ static int goya_pldm_init_cpu(struct hl_device *hdev) /* Release ARM core 0 from reset */ WREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL, CPU_RESET_CORE0_DEASSERT); - val = RREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL); + RREG32(mmCPU_CA53_CFG_ARM_RST_CONTROL); return 0; } @@ -2411,13 +2453,12 @@ err: static int goya_hw_init(struct hl_device *hdev) { struct asic_fixed_properties *prop = &hdev->asic_prop; - u32 val; int rc; dev_info(hdev->dev, "Starting initialization of H/W\n"); /* Perform read from the device to make sure device is up */ - val = RREG32(mmPCIE_DBI_DEVICE_ID_VENDOR_ID_REG); + RREG32(mmPCIE_DBI_DEVICE_ID_VENDOR_ID_REG); /* * Let's mark in the H/W that we have reached this point. We check @@ -2486,7 +2527,7 @@ static int goya_hw_init(struct hl_device *hdev) } /* Perform read from the device to flush all MSI-X configuration */ - val = RREG32(mmPCIE_DBI_DEVICE_ID_VENDOR_ID_REG); + RREG32(mmPCIE_DBI_DEVICE_ID_VENDOR_ID_REG); return 0; @@ -4491,8 +4532,6 @@ static int goya_memset_device_memory(struct hl_device *hdev, u64 addr, u32 size, rc = goya_send_job_on_qman0(hdev, job); - hl_cb_put(job->patched_cb); - hl_debugfs_remove_job(hdev, job); kfree(job); cb->cs_cnt--; diff --git a/drivers/misc/habanalabs/habanalabs.h b/drivers/misc/habanalabs/habanalabs.h index d56ab65d5b2a..43fef59d1bce 100644 --- a/drivers/misc/habanalabs/habanalabs.h +++ b/drivers/misc/habanalabs/habanalabs.h @@ -23,7 +23,7 @@ #define HL_MMAP_CB_MASK (0x8000000000000000ull >> PAGE_SHIFT) -#define HL_PENDING_RESET_PER_SEC 5 +#define HL_PENDING_RESET_PER_SEC 30 #define HL_DEVICE_TIMEOUT_USEC 1000000 /* 1 s */ diff --git a/drivers/misc/habanalabs/memory.c b/drivers/misc/habanalabs/memory.c index 924a438ba973..77ad0bf15c15 100644 --- a/drivers/misc/habanalabs/memory.c +++ b/drivers/misc/habanalabs/memory.c @@ -965,17 +965,19 @@ init_page_pack_err: * * @ctx : current context * @vaddr : device virtual address to unmap + * @ctx_free : true if in context free flow, false otherwise. * * This function does the following: * - Unmap the physical pages related to the given virtual address * - return the device virtual block to the virtual block list */ -static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr) +static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr, bool ctx_free) { struct hl_device *hdev = ctx->hdev; struct hl_vm_phys_pg_pack *phys_pg_pack = NULL; struct hl_vm_hash_node *hnode = NULL; struct hl_userptr *userptr = NULL; + struct hl_va_range *va_range; enum vm_type_t *vm_type; u64 next_vaddr, i; u32 page_size; @@ -1003,6 +1005,7 @@ static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr) if (*vm_type == VM_TYPE_USERPTR) { is_userptr = true; + va_range = &ctx->host_va_range; userptr = hnode->ptr; rc = init_phys_pg_pack_from_userptr(ctx, userptr, &phys_pg_pack); @@ -1014,6 +1017,7 @@ static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr) } } else if (*vm_type == VM_TYPE_PHYS_PACK) { is_userptr = false; + va_range = &ctx->dram_va_range; phys_pg_pack = hnode->ptr; } else { dev_warn(hdev->dev, @@ -1052,12 +1056,18 @@ static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr) mutex_unlock(&ctx->mmu_lock); - if (add_va_block(hdev, - is_userptr ? &ctx->host_va_range : &ctx->dram_va_range, - vaddr, - vaddr + phys_pg_pack->total_size - 1)) - dev_warn(hdev->dev, "add va block failed for vaddr: 0x%llx\n", - vaddr); + /* + * No point in maintaining the free VA block list if the context is + * closing as the list will be freed anyway + */ + if (!ctx_free) { + rc = add_va_block(hdev, va_range, vaddr, + vaddr + phys_pg_pack->total_size - 1); + if (rc) + dev_warn(hdev->dev, + "add va block failed for vaddr: 0x%llx\n", + vaddr); + } atomic_dec(&phys_pg_pack->mapping_cnt); kfree(hnode); @@ -1189,8 +1199,8 @@ int hl_mem_ioctl(struct hl_fpriv *hpriv, void *data) break; case HL_MEM_OP_UNMAP: - rc = unmap_device_va(ctx, - args->in.unmap.device_virt_addr); + rc = unmap_device_va(ctx, args->in.unmap.device_virt_addr, + false); break; default: @@ -1620,7 +1630,7 @@ void hl_vm_ctx_fini(struct hl_ctx *ctx) dev_dbg(hdev->dev, "hl_mem_hash_node of vaddr 0x%llx of asid %d is still alive\n", hnode->vaddr, ctx->asid); - unmap_device_va(ctx, hnode->vaddr); + unmap_device_va(ctx, hnode->vaddr, true); } spin_lock(&vm->idr_lock); |