D3D11渲染函数调用device_draw

渲染的必要过程

• gs_draw(GS_TRISTRIP, 0, 0);
• void device_draw

device_draw流程如下：

void device_draw(gs_device_t *device, enum gs_draw_mode draw_mode,
         uint32_t start_vert, uint32_t num_verts)
{
    device->FlushOutputViews();//设置渲染目标OMSetRenderTargets

    gs_effect_t *effect = gs_get_effect();
    if (effect)
        gs_effect_update_params(effect);

    device->LoadVertexBufferData();
    device->UpdateBlendState();
    device->UpdateRasterState();
    device->UpdateZStencilState();
    device->UpdateViewProjMatrix();
    device->curVertexShader->UploadParams();
    device->curPixelShader->UploadParams();


    D3D11_PRIMITIVE_TOPOLOGY newTopology = ConvertGSTopology(draw_mode);
    if (device->curToplogy != newTopology)
    {
        device->context->IASetPrimitiveTopology(newTopology);//    设置顶点元的组合模式
        device->curToplogy = newTopology;
    }

    //顶点绘制（0，4）
    num_verts = (uint32_t)device->curVertexBuffer->numVerts;
    device->context->Draw(num_verts, start_vert);

}

设置渲染目标

设置渲染目标功能是由函数device->FlushOutputViews()实现的，其核心是调用OMSetRenderTargets。这里看只有一个渲染视图和深度视图

void gs_device::FlushOutputViews()
{
    if (curFramebufferInvalidate)
    {
        ID3D11RenderTargetView *rtv = nullptr;
        if (curRenderTarget) 
        {
            const int i = curRenderSide;
            rtv = curFramebufferSrgb ? curRenderTarget->renderTargetLinear[i].Get()
            :curRenderTarget->renderTarget[i].Get();
            if (!rtv) 
            {
                blog(LOG_ERROR,"device_draw (D3D11): texture is not a render target");
                return;
            }
        }
        ID3D11DepthStencilView *dsv = nullptr;
        if (curZStencilBuffer)
            dsv = curZStencilBuffer->view;
        context->OMSetRenderTargets(1, &rtv, dsv);
        curFramebufferInvalidate = false;
    }
}

而当前的渲染目标是通过函数device_set_render_target实现的。
这里一个比较重要的是会根据curFramebufferSrgb来选择到底是使用那个当前渲染视图中的那一个。该标识是通过函数
gs_enable_framebuffer_srgb(bool enable)来决定的。