This patch adds new functions to map/unmap a usermode queue into the FW, using the MES ring. As soon as this mapping is done, the queue would be considered ready to accept the workload. V1: Addressed review comments from Alex on the RFC patch series - Map/Unmap should be IP specific. V2: Addressed review comments from Christian: - Fix the wptr_mc_addr calculation (moved into another patch) Addressed review comments from Alex: - Do not add fptrs for map/unmap V3: Integration with doorbell manager V4: Rebase V5: Use gfx_v11_0 for function names (Alex) V6: Removed queue->proc/gang/fw_ctx_address variables and doing the address calculations locally to keep the queue structure GEN independent (Alex) V7: Added R-B from Alex V8: Rebase V9: Rebase V10: Rebase Cc: Alex Deucher <alexander.deucher@xxxxxxx> Cc: Christian Koenig <christian.koenig@xxxxxxx> Reviewed-by: Alex Deucher <alexander.deucher@xxxxxxx> Signed-off-by: Shashank Sharma <shashank.sharma@xxxxxxx> --- .../gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c | 74 +++++++++++++++++++ 1 file changed, 74 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c index 2486ea2d72fe..a1bc6f488928 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c @@ -30,6 +30,69 @@ #define AMDGPU_USERQ_PROC_CTX_SZ PAGE_SIZE #define AMDGPU_USERQ_GANG_CTX_SZ PAGE_SIZE +static int mes_v11_0_userq_map(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue, + struct amdgpu_mqd_prop *userq_props) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct amdgpu_userq_obj *ctx = &queue->fw_obj; + struct mes_add_queue_input queue_input; + int r; + + memset(&queue_input, 0x0, sizeof(struct mes_add_queue_input)); + + queue_input.process_va_start = 0; + queue_input.process_va_end = (adev->vm_manager.max_pfn - 1) << AMDGPU_GPU_PAGE_SHIFT; + + /* set process quantum to 10 ms and gang quantum to 1 ms as default */ + queue_input.process_quantum = 100000; + queue_input.gang_quantum = 10000; + queue_input.paging = false; + + queue_input.process_context_addr = ctx->gpu_addr; + queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ; + queue_input.inprocess_gang_priority = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; + queue_input.gang_global_priority_level = AMDGPU_MES_PRIORITY_LEVEL_NORMAL; + + queue_input.process_id = queue->vm->pasid; + queue_input.queue_type = queue->queue_type; + queue_input.mqd_addr = queue->mqd.gpu_addr; + queue_input.wptr_addr = userq_props->wptr_gpu_addr; + queue_input.queue_size = userq_props->queue_size >> 2; + queue_input.doorbell_offset = userq_props->doorbell_index; + queue_input.page_table_base_addr = amdgpu_gmc_pd_addr(queue->vm->root.bo); + + amdgpu_mes_lock(&adev->mes); + r = adev->mes.funcs->add_hw_queue(&adev->mes, &queue_input); + amdgpu_mes_unlock(&adev->mes); + if (r) { + DRM_ERROR("Failed to map queue in HW, err (%d)\n", r); + return r; + } + + DRM_DEBUG_DRIVER("Queue (doorbell:%d) mapped successfully\n", userq_props->doorbell_index); + return 0; +} + +static void mes_v11_0_userq_unmap(struct amdgpu_userq_mgr *uq_mgr, + struct amdgpu_usermode_queue *queue) +{ + struct amdgpu_device *adev = uq_mgr->adev; + struct mes_remove_queue_input queue_input; + struct amdgpu_userq_obj *ctx = &queue->fw_obj; + int r; + + memset(&queue_input, 0x0, sizeof(struct mes_remove_queue_input)); + queue_input.doorbell_offset = queue->doorbell_index; + queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ; + + amdgpu_mes_lock(&adev->mes); + r = adev->mes.funcs->remove_hw_queue(&adev->mes, &queue_input); + amdgpu_mes_unlock(&adev->mes); + if (r) + DRM_ERROR("Failed to unmap queue in HW, err (%d)\n", r); +} + static int mes_v11_0_userq_create_ctx_space(struct amdgpu_userq_mgr *uq_mgr, struct amdgpu_usermode_queue *queue, struct drm_amdgpu_userq_in *mqd_user) @@ -105,8 +168,18 @@ static int mes_v11_0_userq_mqd_create(struct amdgpu_userq_mgr *uq_mgr, goto free_mqd; } + /* Map userqueue into FW using MES */ + r = mes_v11_0_userq_map(uq_mgr, queue, userq_props); + if (r) { + DRM_ERROR("Failed to init MQD\n"); + goto free_ctx; + } + return 0; +free_ctx: + amdgpu_userqueue_destroy_object(uq_mgr, &queue->fw_obj); + free_mqd: amdgpu_userqueue_destroy_object(uq_mgr, &queue->mqd); @@ -120,6 +193,7 @@ static void mes_v11_0_userq_mqd_destroy(struct amdgpu_userq_mgr *uq_mgr, struct amdgpu_usermode_queue *queue) { + mes_v11_0_userq_unmap(uq_mgr, queue); amdgpu_userqueue_destroy_object(uq_mgr, &queue->fw_obj); kfree(queue->userq_prop); amdgpu_userqueue_destroy_object(uq_mgr, &queue->mqd); -- 2.45.1