// Copyright (c) 2015- PPSSPP Project. // This program is free software: you can redistribute it and/or modify // it under the terms of the GNU General Public License as published by // the Free Software Foundation, version 2.0 or later versions. // This program is distributed in the hope that it will be useful, // but WITHOUT ANY WARRANTY; without even the implied warranty of // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the // GNU General Public License 2.0 for more details. // A copy of the GPL 2.0 should have been included with the program. // If not, see http://www.gnu.org/licenses/ // Official git repository and contact information can be found at // https://github.com/hrydgard/ppsspp and http://www.ppsspp.org/. #include #include #include "profiler/profiler.h" #include "base/timeutil.h" #include "math/lin/matrix4x4.h" #include "math/dataconv.h" #include "ext/native/thin3d/thin3d.h" #include "Common/Vulkan/VulkanContext.h" #include "Common/Vulkan/VulkanMemory.h" #include "Common/Vulkan/VulkanImage.h" #include "Common/ColorConv.h" #include "Core/Host.h" #include "Core/MemMap.h" #include "Core/Config.h" #include "Core/System.h" #include "Core/Reporting.h" #include "Core/HLE/sceDisplay.h" #include "GPU/ge_constants.h" #include "GPU/GPUState.h" #include "GPU/Common/PostShader.h" #include "GPU/Common/TextureDecoder.h" #include "GPU/Common/FramebufferCommon.h" #include "GPU/Debugger/Stepping.h" #include "GPU/GPUInterface.h" #include "GPU/GPUState.h" #include "Common/Vulkan/VulkanImage.h" #include "GPU/Vulkan/FramebufferVulkan.h" #include "GPU/Vulkan/DrawEngineVulkan.h" #include "GPU/Vulkan/TextureCacheVulkan.h" #include "GPU/Vulkan/ShaderManagerVulkan.h" #include "GPU/Vulkan/VulkanUtil.h" static const char tex_fs[] = R"(#version 400 #extension GL_ARB_separate_shader_objects : enable #extension GL_ARB_shading_language_420pack : enable layout (binding = 0) uniform sampler2D sampler0; layout (location = 0) in vec2 v_texcoord0; layout (location = 0) out vec4 fragColor; void main() { fragColor = texture(sampler0, v_texcoord0); } )"; static const char tex_vs[] = R"(#version 400 #extension GL_ARB_separate_shader_objects : enable #extension GL_ARB_shading_language_420pack : enable layout (location = 0) in vec3 a_position; layout (location = 1) in vec2 a_texcoord0; layout (location = 0) out vec2 v_texcoord0; out gl_PerVertex { vec4 gl_Position; }; void main() { v_texcoord0 = a_texcoord0; gl_Position = vec4(a_position, 1.0); } )"; void ConvertFromRGBA8888_Vulkan(u8 *dst, const u8 *src, u32 dstStride, u32 srcStride, u32 width, u32 height, GEBufferFormat format); FramebufferManagerVulkan::FramebufferManagerVulkan(Draw::DrawContext *draw, VulkanContext *vulkan) : FramebufferManagerCommon(draw), vulkan_(vulkan), drawPixelsTex_(nullptr), drawPixelsTexFormat_(GE_FORMAT_INVALID), convBuf_(nullptr), convBufSize_(0), textureCacheVulkan_(nullptr), shaderManagerVulkan_(nullptr), pixelBufObj_(nullptr), currentPBO_(0), curFrame_(0), pipelinePostShader_(VK_NULL_HANDLE), vulkan2D_(vulkan) { InitDeviceObjects(); } FramebufferManagerVulkan::~FramebufferManagerVulkan() { delete[] convBuf_; vulkan2D_.Shutdown(); DestroyDeviceObjects(); } void FramebufferManagerVulkan::SetTextureCache(TextureCacheVulkan *tc) { textureCacheVulkan_ = tc; textureCache_ = tc; } void FramebufferManagerVulkan::SetShaderManager(ShaderManagerVulkan *sm) { shaderManagerVulkan_ = sm; shaderManager_ = sm; } void FramebufferManagerVulkan::InitDeviceObjects() { // Initialize framedata for (int i = 0; i < 2; i++) { frameData_[i].push_ = new VulkanPushBuffer(vulkan_, 64 * 1024); } pipelineCache2D_ = vulkan_->CreatePipelineCache(); std::string fs_errors, vs_errors; fsBasicTex_ = CompileShaderModule(vulkan_, VK_SHADER_STAGE_FRAGMENT_BIT, tex_fs, &fs_errors); vsBasicTex_ = CompileShaderModule(vulkan_, VK_SHADER_STAGE_VERTEX_BIT, tex_vs, &vs_errors); assert(fsBasicTex_ != VK_NULL_HANDLE); assert(vsBasicTex_ != VK_NULL_HANDLE); // Prime the 2D pipeline cache. vulkan2D_.GetPipeline(pipelineCache2D_, vulkan_->GetSurfaceRenderPass(), vsBasicTex_, fsBasicTex_); vulkan2D_.GetPipeline(pipelineCache2D_, (VkRenderPass)draw_->GetNativeObject(Draw::NativeObject::COMPATIBLE_RENDERPASS), vsBasicTex_, fsBasicTex_); VkSamplerCreateInfo samp = { VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO }; samp.addressModeU = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE; samp.addressModeV = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE; samp.addressModeW = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_EDGE; samp.magFilter = VK_FILTER_NEAREST; samp.minFilter = VK_FILTER_NEAREST; VkResult res = vkCreateSampler(vulkan_->GetDevice(), &samp, nullptr, &nearestSampler_); assert(res == VK_SUCCESS); samp.magFilter = VK_FILTER_LINEAR; samp.minFilter = VK_FILTER_LINEAR; res = vkCreateSampler(vulkan_->GetDevice(), &samp, nullptr, &linearSampler_); assert(res == VK_SUCCESS); } void FramebufferManagerVulkan::DestroyDeviceObjects() { for (int i = 0; i < 2; i++) { if (frameData_[i].push_) { frameData_[i].push_->Destroy(vulkan_); delete frameData_[i].push_; frameData_[i].push_ = nullptr; } } delete drawPixelsTex_; drawPixelsTex_ = nullptr; if (fsBasicTex_ != VK_NULL_HANDLE) vulkan_->Delete().QueueDeleteShaderModule(fsBasicTex_); if (vsBasicTex_ != VK_NULL_HANDLE) vulkan_->Delete().QueueDeleteShaderModule(vsBasicTex_); if (linearSampler_ != VK_NULL_HANDLE) vulkan_->Delete().QueueDeleteSampler(linearSampler_); if (nearestSampler_ != VK_NULL_HANDLE) vulkan_->Delete().QueueDeleteSampler(nearestSampler_); // pipelineBasicTex_ and pipelineBasicTex_ come from vulkan2D_. if (pipelineCache2D_ != VK_NULL_HANDLE) vulkan_->Delete().QueueDeletePipelineCache(pipelineCache2D_); } void FramebufferManagerVulkan::NotifyClear(bool clearColor, bool clearAlpha, bool clearDepth, uint32_t color, float depth) { // if (!useBufferedRendering_) { float x, y, w, h; CenterDisplayOutputRect(&x, &y, &w, &h, 480.0f, 272.0f, (float)pixelWidth_, (float)pixelHeight_, ROTATION_LOCKED_HORIZONTAL); int mask = 0; // The Clear detection takes care of doing a regular draw instead if separate masking // of color and alpha is needed, so we can just treat them as the same. if (clearColor || clearAlpha) mask |= Draw::FBChannel::FB_COLOR_BIT; if (clearDepth) mask |= Draw::FBChannel::FB_DEPTH_BIT; if (clearAlpha) mask |= Draw::FBChannel::FB_STENCIL_BIT; draw_->Clear(mask, color, depth, 0); if (clearColor || clearAlpha) { SetColorUpdated(gstate_c.skipDrawReason); } if (clearDepth) { SetDepthUpdated(); } //} else { // TODO: Clever render pass magic. //} } void FramebufferManagerVulkan::DoNotifyDraw() { } void FramebufferManagerVulkan::UpdatePostShaderUniforms(int bufferWidth, int bufferHeight, int renderWidth, int renderHeight) { float u_delta = 1.0f / renderWidth; float v_delta = 1.0f / renderHeight; float u_pixel_delta = u_delta; float v_pixel_delta = v_delta; if (postShaderAtOutputResolution_) { float x, y, w, h; CenterDisplayOutputRect(&x, &y, &w, &h, 480.0f, 272.0f, (float)pixelWidth_, (float)pixelHeight_, ROTATION_LOCKED_HORIZONTAL); u_pixel_delta = (1.0f / w) * (480.0f / bufferWidth); v_pixel_delta = (1.0f / h) * (272.0f / bufferHeight); } postUniforms_.texelDelta[0] = u_delta; postUniforms_.texelDelta[1] = v_delta; postUniforms_.pixelDelta[0] = u_pixel_delta; postUniforms_.pixelDelta[1] = v_pixel_delta; int flipCount = __DisplayGetFlipCount(); int vCount = __DisplayGetVCount(); float time[4] = { time_now(), (vCount % 60) * 1.0f / 60.0f, (float)vCount, (float)(flipCount % 60) }; memcpy(postUniforms_.time, time, 4 * sizeof(float)); } void FramebufferManagerVulkan::Init() { FramebufferManagerCommon::Init(); // Workaround for upscaling shaders where we force x1 resolution without saving it Resized(); } void FramebufferManagerVulkan::MakePixelTexture(const u8 *srcPixels, GEBufferFormat srcPixelFormat, int srcStride, int width, int height, float &u1, float &v1) { if (drawPixelsTex_ && (drawPixelsTexFormat_ != srcPixelFormat || drawPixelsTex_->GetWidth() != width || drawPixelsTex_->GetHeight() != height)) { delete drawPixelsTex_; drawPixelsTex_ = nullptr; } if (!drawPixelsTex_) { drawPixelsTex_ = new VulkanTexture(vulkan_); drawPixelsTex_->CreateDirect(width, height, 1, VK_FORMAT_R8G8B8A8_UNORM, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT); // Initialize backbuffer texture for DrawPixels drawPixelsTexFormat_ = srcPixelFormat; } else { drawPixelsTex_->TransitionForUpload(); } // TODO: We can just change the texture format and flip some bits around instead of this. // Could share code with the texture cache perhaps. const uint8_t *data = srcPixels; if (srcPixelFormat != GE_FORMAT_8888 || srcStride != width) { u32 neededSize = width * height * 4; if (!convBuf_ || convBufSize_ < neededSize) { delete[] convBuf_; convBuf_ = new u8[neededSize]; convBufSize_ = neededSize; } data = convBuf_; for (int y = 0; y < height; y++) { switch (srcPixelFormat) { case GE_FORMAT_565: { const u16 *src = (const u16 *)srcPixels + srcStride * y; u8 *dst = convBuf_ + 4 * width * y; ConvertRGBA565ToRGBA8888((u32 *)dst, src, width); } break; case GE_FORMAT_5551: { const u16 *src = (const u16 *)srcPixels + srcStride * y; u8 *dst = convBuf_ + 4 * width * y; ConvertRGBA5551ToRGBA8888((u32 *)dst, src, width); } break; case GE_FORMAT_4444: { const u16 *src = (const u16 *)srcPixels + srcStride * y; u8 *dst = convBuf_ + 4 * width * y; ConvertRGBA4444ToRGBA8888((u32 *)dst, src, width); } break; case GE_FORMAT_8888: { const u8 *src = srcPixels + srcStride * 4 * y; u8 *dst = convBuf_ + 4 * width * y; memcpy(dst, src, 4 * width); } break; case GE_FORMAT_INVALID: _dbg_assert_msg_(G3D, false, "Invalid pixelFormat passed to DrawPixels()."); break; } } } VkBuffer buffer; size_t offset = frameData_[curFrame_].push_->Push(data, width * height * 4, &buffer); drawPixelsTex_->UploadMip(0, width, height, buffer, (uint32_t)offset, width); drawPixelsTex_->EndCreate(); overrideImageView_ = drawPixelsTex_->GetImageView(); } void FramebufferManagerVulkan::SetViewport2D(int x, int y, int w, int h) { VkViewport vp; vp.minDepth = 0.0; vp.maxDepth = 1.0; vp.x = (float)x; vp.y = (float)y; vp.width = (float)w; vp.height = (float)h; VkCommandBuffer cmd = (VkCommandBuffer)draw_->GetNativeObject(Draw::NativeObject::RENDERPASS_COMMANDBUFFER); vkCmdSetViewport(cmd, 0, 1, &vp); } void FramebufferManagerVulkan::DrawActiveTexture(float x, float y, float w, float h, float destW, float destH, float u0, float v0, float u1, float v1, int uvRotation, bool linearFilter) { float texCoords[8] = { u0,v0, u1,v0, u1,v1, u0,v1, }; if (uvRotation != ROTATION_LOCKED_HORIZONTAL) { float temp[8]; int rotation = 0; switch (uvRotation) { case ROTATION_LOCKED_HORIZONTAL180: rotation = 4; break; case ROTATION_LOCKED_VERTICAL: rotation = 2; break; case ROTATION_LOCKED_VERTICAL180: rotation = 6; break; } for (int i = 0; i < 8; i++) { temp[i] = texCoords[(i + rotation) & 7]; } memcpy(texCoords, temp, sizeof(temp)); } Vulkan2D::Vertex vtx[4] = { {x, y, 0, texCoords[0], texCoords[1]}, {x + w, y, 0, texCoords[2], texCoords[3]}, {x, y + h, 0, texCoords[6], texCoords[7]}, {x + w, y + h, 0, texCoords[4], texCoords[5]}, }; float invDestW = 1.0f / (destW * 0.5f); float invDestH = 1.0f / (destH * 0.5f); for (int i = 0; i < 4; i++) { vtx[i].x = vtx[i].x * invDestW - 1.0f; vtx[i].y = vtx[i].y * invDestH - 1.0f; } draw_->FlushState(); // TODO: Should probably use draw_ directly and not go low level VulkanPushBuffer *push = frameData_[curFrame_].push_; VkCommandBuffer cmd = (VkCommandBuffer)draw_->GetNativeObject(Draw::NativeObject::RENDERPASS_COMMANDBUFFER); VkImageView view = overrideImageView_ ? overrideImageView_ : (VkImageView)draw_->GetNativeObject(Draw::NativeObject::BOUND_TEXTURE_IMAGEVIEW); overrideImageView_ = VK_NULL_HANDLE; vulkan2D_.BindDescriptorSet(cmd, view, linearFilter ? linearSampler_ : nearestSampler_); vkCmdBindPipeline(cmd, VK_PIPELINE_BIND_POINT_GRAPHICS, cur2DPipeline_); VkBuffer vbuffer; VkDeviceSize offset = push->Push(vtx, sizeof(vtx), &vbuffer); vkCmdBindVertexBuffers(cmd, 0, 1, &vbuffer, &offset); vkCmdDraw(cmd, 4, 1, 0, 0); } void FramebufferManagerVulkan::Bind2DShader() { VkRenderPass rp = (VkRenderPass)draw_->GetNativeObject(Draw::NativeObject::COMPATIBLE_RENDERPASS); cur2DPipeline_ = vulkan2D_.GetPipeline(pipelineCache2D_, rp, vsBasicTex_, fsBasicTex_); } void FramebufferManagerVulkan::BindPostShader(const PostShaderUniforms &uniforms) { Bind2DShader(); } void FramebufferManagerVulkan::RebindFramebuffer() { if (currentRenderVfb_ && currentRenderVfb_->fbo) { draw_->BindFramebufferAsRenderTarget(currentRenderVfb_->fbo, { Draw::RPAction::KEEP, Draw::RPAction::KEEP }); } else { // Should this even happen? draw_->BindFramebufferAsRenderTarget(nullptr, { Draw::RPAction::KEEP, Draw::RPAction::KEEP }); } } bool FramebufferManagerVulkan::NotifyStencilUpload(u32 addr, int size, bool skipZero) { // In Vulkan we should be able to simply copy the stencil data directly to a stencil buffer without // messing about with bitplane textures and the like. return false; } int FramebufferManagerVulkan::GetLineWidth() { if (g_Config.iInternalResolution == 0) { return std::max(1, (int)(renderWidth_ / 480)); } else { return g_Config.iInternalResolution; } } // This also binds vfb as the current render target. void FramebufferManagerVulkan::ReformatFramebufferFrom(VirtualFramebuffer *vfb, GEBufferFormat old) { if (!useBufferedRendering_ || !vfb->fbo) { return; } // Technically, we should at this point re-interpret the bytes of the old format to the new. // That might get tricky, and could cause unnecessary slowness in some games. // For now, we just clear alpha/stencil from 565, which fixes shadow issues in Kingdom Hearts. // (it uses 565 to write zeros to the buffer, than 4444 to actually render the shadow.) // // The best way to do this may ultimately be to create a new FBO (combine with any resize?) // and blit with a shader to that, then replace the FBO on vfb. Stencil would still be complex // to exactly reproduce in 4444 and 8888 formats. if (old == GE_FORMAT_565) { draw_->BindFramebufferAsRenderTarget(vfb->fbo, { Draw::RPAction::CLEAR, Draw::RPAction::CLEAR }); } else { draw_->BindFramebufferAsRenderTarget(vfb->fbo, { Draw::RPAction::KEEP, Draw::RPAction::KEEP }); } } // Except for a missing rebind and silly scissor enables, identical copy of the same function in GPU_GLES - tricky parts are in thin3d. void FramebufferManagerVulkan::BlitFramebufferDepth(VirtualFramebuffer *src, VirtualFramebuffer *dst) { if (g_Config.bDisableSlowFramebufEffects) { return; } bool matchingDepthBuffer = src->z_address == dst->z_address && src->z_stride != 0 && dst->z_stride != 0; bool matchingSize = src->width == dst->width && src->height == dst->height; bool matchingRenderSize = src->renderWidth == dst->renderWidth && src->renderHeight == dst->renderHeight; if (gstate_c.Supports(GPU_SUPPORTS_ANY_COPY_IMAGE) && matchingDepthBuffer && matchingRenderSize && matchingSize) { draw_->CopyFramebufferImage(src->fbo, 0, 0, 0, 0, dst->fbo, 0, 0, 0, 0, src->renderWidth, src->renderHeight, 1, Draw::FB_DEPTH_BIT); } else if (matchingDepthBuffer && matchingSize) { int w = std::min(src->renderWidth, dst->renderWidth); int h = std::min(src->renderHeight, dst->renderHeight); if (gstate_c.Supports(GPU_SUPPORTS_ARB_FRAMEBUFFER_BLIT | GPU_SUPPORTS_NV_FRAMEBUFFER_BLIT)) { draw_->BlitFramebuffer(src->fbo, 0, 0, w, h, dst->fbo, 0, 0, w, h, Draw::FB_DEPTH_BIT, Draw::FB_BLIT_NEAREST); } } } VkImageView FramebufferManagerVulkan::BindFramebufferAsColorTexture(int stage, VirtualFramebuffer *framebuffer, int flags) { if (!framebuffer->fbo || !useBufferedRendering_) { gstate_c.skipDrawReason |= SKIPDRAW_BAD_FB_TEXTURE; return VK_NULL_HANDLE; } // currentRenderVfb_ will always be set when this is called, except from the GE debugger. // Let's just not bother with the copy in that case. bool skipCopy = (flags & BINDFBCOLOR_MAY_COPY) == 0; if (GPUStepping::IsStepping() || g_Config.bDisableSlowFramebufEffects) { skipCopy = true; } // Currently rendering to this framebuffer. Need to make a copy. if (!skipCopy && framebuffer == currentRenderVfb_) { // ignore this case for now, doesn't work // ILOG("Texturing from current render Vfb!"); return VK_NULL_HANDLE; // TODO: Maybe merge with bvfbs_? Not sure if those could be packing, and they're created at a different size. Draw::Framebuffer *renderCopy = GetTempFBO(framebuffer->renderWidth, framebuffer->renderHeight, (Draw::FBColorDepth)framebuffer->colorDepth); if (renderCopy) { VirtualFramebuffer copyInfo = *framebuffer; copyInfo.fbo = renderCopy; CopyFramebufferForColorTexture(©Info, framebuffer, flags); RebindFramebuffer(); draw_->BindFramebufferAsTexture(renderCopy, stage, Draw::FB_COLOR_BIT, 0); } else { draw_->BindFramebufferAsTexture(framebuffer->fbo, stage, Draw::FB_COLOR_BIT, 0); } return (VkImageView)draw_->GetNativeObject(Draw::NativeObject::BOUND_TEXTURE_IMAGEVIEW); } else if (framebuffer != currentRenderVfb_) { draw_->BindFramebufferAsTexture(framebuffer->fbo, stage, Draw::FB_COLOR_BIT, 0); return (VkImageView)draw_->GetNativeObject(Draw::NativeObject::BOUND_TEXTURE_IMAGEVIEW); } else { ERROR_LOG_REPORT_ONCE(vulkanSelfTexture, G3D, "Attempting to texture from target"); // To do this safely in Vulkan, we need to use input attachments. return VK_NULL_HANDLE; } } VulkanTexture *FramebufferManagerVulkan::GetFramebufferColor(u32 fbRawAddress, VirtualFramebuffer *framebuffer, int flags) { if (framebuffer == NULL) { framebuffer = currentRenderVfb_; } if (!framebuffer->fbo || !useBufferedRendering_) { gstate_c.skipDrawReason |= SKIPDRAW_BAD_FB_TEXTURE; return nullptr; } // currentRenderVfb_ will always be set when this is called, except from the GE debugger. // Let's just not bother with the copy in that case. bool skipCopy = (flags & BINDFBCOLOR_MAY_COPY) == 0; if (GPUStepping::IsStepping() || g_Config.bDisableSlowFramebufEffects) { skipCopy = true; } if (!skipCopy && currentRenderVfb_ && framebuffer->fb_address == fbRawAddress) { // TODO: Enable the below code return nullptr; // framebuffer->fbo->GetColor(); /* // TODO: Maybe merge with bvfbs_? Not sure if those could be packing, and they're created at a different size. VulkanFBO *renderCopy = GetTempFBO(framebuffer->renderWidth, framebuffer->renderHeight, (FBOColorDepth)framebuffer->colorDepth); if (renderCopy) { VirtualFramebuffer copyInfo = *framebuffer; copyInfo.fbo = renderCopy; int x = 0; int y = 0; int w = framebuffer->drawnWidth; int h = framebuffer->drawnHeight; // If max is not > min, we probably could not detect it. Skip. // See the vertex decoder, where this is updated. if ((flags & BINDFBCOLOR_MAY_COPY_WITH_UV) == BINDFBCOLOR_MAY_COPY_WITH_UV && gstate_c.vertBounds.maxU > gstate_c.vertBounds.minU) { x = gstate_c.vertBounds.minU; y = gstate_c.vertBounds.minV; w = gstate_c.vertBounds.maxU - x; h = gstate_c.vertBounds.maxV - y; // If we bound a framebuffer, apply the byte offset as pixels to the copy too. if (flags & BINDFBCOLOR_APPLY_TEX_OFFSET) { x += gstate_c.curTextureXOffset; y += gstate_c.curTextureYOffset; } } BlitFramebuffer(©Info, x, y, framebuffer, x, y, w, h, 0); return nullptr; // fbo_bind_color_as_texture(renderCopy, 0); } else { return framebuffer->fbo->GetColor(); } */ } else { return nullptr; // framebuffer->fbo->GetColor(); } } void FramebufferManagerVulkan::ReadFramebufferToMemory(VirtualFramebuffer *vfb, bool sync, int x, int y, int w, int h) { PROFILE_THIS_SCOPE("gpu-readback"); if (sync) { // flush async just in case when we go for synchronous update // Doesn't actually pack when sent a null argument. PackFramebufferAsync_(nullptr); } if (vfb) { // We'll pseudo-blit framebuffers here to get a resized version of vfb. VirtualFramebuffer *nvfb = FindDownloadTempBuffer(vfb); OptimizeDownloadRange(vfb, x, y, w, h); BlitFramebuffer(nvfb, x, y, vfb, x, y, w, h, 0); // PackFramebufferSync_() - Synchronous pixel data transfer using glReadPixels // PackFramebufferAsync_() - Asynchronous pixel data transfer using glReadPixels with PBOs // TODO: Can we fall back to sync without these? if (!sync) { PackFramebufferAsync_(nvfb); } else { PackFramebufferSync_(nvfb, x, y, w, h); } textureCacheVulkan_->ForgetLastTexture(); RebindFramebuffer(); } } void FramebufferManagerVulkan::DownloadFramebufferForClut(u32 fb_address, u32 loadBytes) { PROFILE_THIS_SCOPE("gpu-readback"); // Flush async just in case. PackFramebufferAsync_(nullptr); VirtualFramebuffer *vfb = GetVFBAt(fb_address); if (vfb && vfb->fb_stride != 0) { const u32 bpp = vfb->drawnFormat == GE_FORMAT_8888 ? 4 : 2; int x = 0; int y = 0; int pixels = loadBytes / bpp; // The height will be 1 for each stride or part thereof. int w = std::min(pixels % vfb->fb_stride, (int)vfb->width); int h = std::min((pixels + vfb->fb_stride - 1) / vfb->fb_stride, (int)vfb->height); // No need to download if we already have it. if (!vfb->memoryUpdated && vfb->clutUpdatedBytes < loadBytes) { // We intentionally don't call OptimizeDownloadRange() here - we don't want to over download. // CLUT framebuffers are often incorrectly estimated in size. if (x == 0 && y == 0 && w == vfb->width && h == vfb->height) { vfb->memoryUpdated = true; } vfb->clutUpdatedBytes = loadBytes; // We'll pseudo-blit framebuffers here to get a resized version of vfb. VirtualFramebuffer *nvfb = FindDownloadTempBuffer(vfb); BlitFramebuffer(nvfb, x, y, vfb, x, y, w, h, 0); PackFramebufferSync_(nvfb, x, y, w, h); textureCacheVulkan_->ForgetLastTexture(); RebindFramebuffer(); } } } bool FramebufferManagerVulkan::CreateDownloadTempBuffer(VirtualFramebuffer *nvfb) { // When updating VRAM, it need to be exact format. if (!gstate_c.Supports(GPU_PREFER_CPU_DOWNLOAD)) { switch (nvfb->format) { case GE_FORMAT_4444: nvfb->colorDepth = VK_FBO_4444; break; case GE_FORMAT_5551: nvfb->colorDepth = VK_FBO_5551; break; case GE_FORMAT_565: nvfb->colorDepth = VK_FBO_565; break; case GE_FORMAT_8888: default: nvfb->colorDepth = VK_FBO_8888; break; } } /* nvfb->fbo = CreateFramebuffer(nvfb->width, nvfb->height, 1, false, (FBOColorDepth)nvfb->colorDepth); if (!(nvfb->fbo)) { ERROR_LOG(FRAMEBUF, "Error creating FBO! %i x %i", nvfb->renderWidth, nvfb->renderHeight); return false; } BindFramebufferAsRenderTargetnvfb->fbo); ClearBuffer(); glDisable(GL_DITHER); */ return true; } void FramebufferManagerVulkan::UpdateDownloadTempBuffer(VirtualFramebuffer *nvfb) { // _assert_msg_(G3D, nvfb->fbo, "Expecting a valid nvfb in UpdateDownloadTempBuffer"); // Discard the previous contents of this buffer where possible. /* if (gl_extensions.GLES3 && glInvalidateFramebuffer != nullptr) { BindFramebufferAsRenderTargetnvfb->fbo); GLenum attachments[3] = { GL_COLOR_ATTACHMENT0, GL_STENCIL_ATTACHMENT, GL_DEPTH_ATTACHMENT }; glInvalidateFramebuffer(GL_FRAMEBUFFER, 3, attachments); } else if (gl_extensions.IsGLES) { BindFramebufferAsRenderTargetnvfb->fbo); ClearBuffer(); } */ } void FramebufferManagerVulkan::BlitFramebuffer(VirtualFramebuffer *dst, int dstX, int dstY, VirtualFramebuffer *src, int srcX, int srcY, int w, int h, int bpp) { if (!dst->fbo || !src->fbo || !useBufferedRendering_) { // This can happen if they recently switched from non-buffered. return; } // NOTE: There may be cases (like within a renderpass) where we want to // not use a blit. bool useBlit = true; float srcXFactor = useBlit ? (float)src->renderWidth / (float)src->bufferWidth : 1.0f; float srcYFactor = useBlit ? (float)src->renderHeight / (float)src->bufferHeight : 1.0f; const int srcBpp = src->format == GE_FORMAT_8888 ? 4 : 2; if (srcBpp != bpp && bpp != 0) { srcXFactor = (srcXFactor * bpp) / srcBpp; } int srcX1 = srcX * srcXFactor; int srcX2 = (srcX + w) * srcXFactor; int srcY1 = srcY * srcYFactor; int srcY2 = (srcY + h) * srcYFactor; float dstXFactor = useBlit ? (float)dst->renderWidth / (float)dst->bufferWidth : 1.0f; float dstYFactor = useBlit ? (float)dst->renderHeight / (float)dst->bufferHeight : 1.0f; const int dstBpp = dst->format == GE_FORMAT_8888 ? 4 : 2; if (dstBpp != bpp && bpp != 0) { dstXFactor = (dstXFactor * bpp) / dstBpp; } int dstX1 = dstX * dstXFactor; int dstX2 = (dstX + w) * dstXFactor; int dstY1 = dstY * dstYFactor; int dstY2 = (dstY + h) * dstYFactor; if (src == dst && srcX == dstX && srcY == dstY) { // Let's just skip a copy where the destination is equal to the source. WARN_LOG_REPORT_ONCE(blitSame, G3D, "Skipped blit with equal dst and src"); return; } // In case the src goes outside, we just skip the optimization in that case. const bool sameSize = dstX2 - dstX1 == srcX2 - srcX1 && dstY2 - dstY1 == srcY2 - srcY1; const bool sameDepth = dst->colorDepth == src->colorDepth; const bool srcInsideBounds = srcX2 <= src->renderWidth && srcY2 <= src->renderHeight; const bool dstInsideBounds = dstX2 <= dst->renderWidth && dstY2 <= dst->renderHeight; const bool xOverlap = src == dst && srcX2 > dstX1 && srcX1 < dstX2; const bool yOverlap = src == dst && srcY2 > dstY1 && srcY1 < dstY2; if (sameSize && sameDepth && srcInsideBounds && dstInsideBounds && !(xOverlap && yOverlap)) { VkImageCopy region = {}; region.extent = { (uint32_t)(dstX2 - dstX1), (uint32_t)(dstY2 - dstY1), 1 }; /* glCopyImageSubDataOES( fbo_get_color_texture(src->fbo), GL_TEXTURE_2D, 0, srcX1, srcY1, 0, fbo_get_color_texture(dst->fbo), GL_TEXTURE_2D, 0, dstX1, dstY1, 0, dstX2 - dstX1, dstY2 - dstY1, 1); */ return; } // BindFramebufferAsRenderTargetdst->fbo); if (useBlit) { // fbo_bind_for_read(src->fbo); //glBlitFramebuffer(srcX1, srcY1, srcX2, srcY2, dstX1, dstY1, dstX2, dstY2, GL_COLOR_BUFFER_BIT, GL_NEAREST); } else { // fbo_bind_color_as_texture(src->fbo, 0); // The first four coordinates are relative to the 6th and 7th arguments of DrawActiveTexture. // Should maybe revamp that interface. float srcW = src->bufferWidth; float srcH = src->bufferHeight; // DrawActiveTexture(0, dstX1, dstY1, w * dstXFactor, h, dst->bufferWidth, dst->bufferHeight, srcX1 / srcW, srcY1 / srcH, srcX2 / srcW, srcY2 / srcH, draw2dprogram_, ROTATION_LOCKED_HORIZONTAL); } } // TODO: SSE/NEON // Could also make C fake-simd for 64-bit, two 8888 pixels fit in a register :) void ConvertFromRGBA8888_Vulkan(u8 *dst, const u8 *src, u32 dstStride, u32 srcStride, u32 width, u32 height, GEBufferFormat format) { // Must skip stride in the cases below. Some games pack data into the cracks, like MotoGP. const u32 *src32 = (const u32 *)src; if (format == GE_FORMAT_8888) { u32 *dst32 = (u32 *)dst; if (src == dst) { return; } else { // Here let's assume they don't intersect for (u32 y = 0; y < height; ++y) { memcpy(dst32, src32, width * 4); src32 += srcStride; dst32 += dstStride; } } } else { // But here it shouldn't matter if they do intersect u16 *dst16 = (u16 *)dst; switch (format) { case GE_FORMAT_565: // BGR 565 for (u32 y = 0; y < height; ++y) { ConvertRGBA8888ToRGB565(dst16, src32, width); src32 += srcStride; dst16 += dstStride; } break; case GE_FORMAT_5551: // ABGR 1555 for (u32 y = 0; y < height; ++y) { ConvertBGRA8888ToRGBA5551(dst16, src32, width); src32 += srcStride; dst16 += dstStride; } break; case GE_FORMAT_4444: // ABGR 4444 for (u32 y = 0; y < height; ++y) { ConvertRGBA8888ToRGBA4444(dst16, src32, width); src32 += srcStride; dst16 += dstStride; } break; case GE_FORMAT_8888: case GE_FORMAT_INVALID: // Not possible. break; } } } #ifdef DEBUG_READ_PIXELS // TODO: Make more generic. static void LogReadPixelsError(GLenum error) { switch (error) { case GL_NO_ERROR: break; case GL_INVALID_ENUM: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_INVALID_ENUM"); break; case GL_INVALID_VALUE: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_INVALID_VALUE"); break; case GL_INVALID_OPERATION: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_INVALID_OPERATION"); break; case GL_INVALID_FRAMEBUFFER_OPERATION: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_INVALID_FRAMEBUFFER_OPERATION"); break; case GL_OUT_OF_MEMORY: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_OUT_OF_MEMORY"); break; #ifndef USING_GLES2 case GL_STACK_UNDERFLOW: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_STACK_UNDERFLOW"); break; case GL_STACK_OVERFLOW: ERROR_LOG(FRAMEBUF, "glReadPixels: GL_STACK_OVERFLOW"); break; #endif default: ERROR_LOG(FRAMEBUF, "glReadPixels: %08x", error); break; } } #endif // One frame behind, but no stalling. void FramebufferManagerVulkan::PackFramebufferAsync_(VirtualFramebuffer *vfb) { const int MAX_PBO = 2; uint8_t *packed = 0; const u8 nextPBO = (currentPBO_ + 1) % MAX_PBO; bool useCPU = false; // We'll prepare two PBOs to switch between readying and reading if (!pixelBufObj_) { if (!vfb) { // This call is just to flush the buffers. We don't have any yet, // so there's nothing to do. return; } // GLuint pbos[MAX_PBO]; // glGenBuffers(MAX_PBO, pbos); pixelBufObj_ = new AsyncPBOVulkan[MAX_PBO]; for (int i = 0; i < MAX_PBO; i++) { // TODO // pixelBufObj_[i].handle = pbos[i]; pixelBufObj_[i].maxSize = 0; pixelBufObj_[i].reading = false; } } // Receive previously requested data from a PBO AsyncPBOVulkan &pbo = pixelBufObj_[nextPBO]; if (pbo.reading) { // glBindBuffer(GL_PIXEL_PACK_BUFFER, pbo.handle); // packed = (GLubyte *)glMapBufferRange(GL_PIXEL_PACK_BUFFER, 0, pbo.size, GL_MAP_READ_BIT); if (packed) { DEBUG_LOG(FRAMEBUF, "Reading PBO to memory , bufSize = %u, packed = %p, fb_address = %08x, stride = %u, pbo = %u", pbo.size, packed, pbo.fb_address, pbo.stride, nextPBO); // We don't need to convert, GPU already did (or should have) // (vulkan: hopefully) Memory::MemcpyUnchecked(pbo.fb_address, packed, pbo.size); pbo.reading = false; } // glUnmapBuffer(GL_PIXEL_PACK_BUFFER); } // Order packing/readback of the framebuffer if (vfb) { // int pixelType, pixelFormat; int pixelSize, align; switch (vfb->format) { case GE_FORMAT_4444: // 16 bit RGBA // pixelType = GL_UNSIGNED_SHORT_4_4_4_4; // pixelFormat = GL_RGBA; pixelSize = 2; align = 2; break; case GE_FORMAT_5551: // 16 bit RGBA // pixelType = GL_UNSIGNED_SHORT_5_5_5_1; // pixelFormat = GL_RGBA; pixelSize = 2; align = 2; break; case GE_FORMAT_565: // 16 bit RGB // pixelType = GL_UNSIGNED_SHORT_5_6_5; // pixelFormat = GL_RGB; pixelSize = 2; align = 2; break; case GE_FORMAT_8888: // 32 bit RGBA default: // pixelType = GL_UNSIGNED_BYTE; // pixelFormat = UseBGRA8888() ? GL_BGRA_EXT : GL_RGBA; pixelSize = 4; align = 4; break; } // If using the CPU, we need 4 bytes per pixel always. u32 bufSize = vfb->fb_stride * vfb->height * 4 * (useCPU ? 4 : pixelSize); u32 fb_address = (0x04000000) | vfb->fb_address; if (vfb->fbo) { // fbo_bind_for_read(vfb->fbo); } else { ERROR_LOG_REPORT_ONCE(vfbfbozero, SCEGE, "PackFramebufferAsync_: vfb->fbo == 0"); // fbo_unbind_read(); return; } // glBindBuffer(GL_PIXEL_PACK_BUFFER, pixelBufObj_[currentPBO_].handle); if (pixelBufObj_[currentPBO_].maxSize < bufSize) { // We reserve a buffer big enough to fit all those pixels // glBufferData(GL_PIXEL_PACK_BUFFER, bufSize, NULL, GL_DYNAMIC_READ); pixelBufObj_[currentPBO_].maxSize = bufSize; } if (useCPU) { // If converting pixel formats on the CPU we'll always request RGBA8888 // SafeGLReadPixels(0, 0, vfb->fb_stride, vfb->height, UseBGRA8888() ? GL_BGRA_EXT : GL_RGBA, GL_UNSIGNED_BYTE, 0); } else { // Otherwise we'll directly request the format we need and let the GPU sort it out // SafeGLReadPixels(0, 0, vfb->fb_stride, vfb->height, pixelFormat, pixelType, 0); } pixelBufObj_[currentPBO_].fb_address = fb_address; pixelBufObj_[currentPBO_].size = bufSize; pixelBufObj_[currentPBO_].stride = vfb->fb_stride; pixelBufObj_[currentPBO_].height = vfb->height; pixelBufObj_[currentPBO_].format = vfb->format; pixelBufObj_[currentPBO_].reading = true; } currentPBO_ = nextPBO; } void FramebufferManagerVulkan::PackFramebufferSync_(VirtualFramebuffer *vfb, int x, int y, int w, int h) { } void FramebufferManagerVulkan::BeginFrameVulkan() { BeginFrame(); vulkan2D_.BeginFrame(); FrameData &frame = frameData_[curFrame_]; frame.push_->Reset(); frame.push_->Begin(vulkan_); if (!useBufferedRendering_) { // TODO: This hackery should not be necessary. Is it? Need to check. // We only use a single command buffer in this case. VkCommandBuffer cmd = vulkan_->GetSurfaceCommandBuffer(); VkRect2D scissor; scissor.offset = { 0, 0 }; scissor.extent = { (uint32_t)pixelWidth_, (uint32_t)pixelHeight_ }; vkCmdSetScissor(cmd, 0, 1, &scissor); } else { // Each render pass will set up scissor again. } } void FramebufferManagerVulkan::EndFrame() { // We flush to memory last requested framebuffer, if any. // Only do this in the read-framebuffer modes. if (updateVRAM_) PackFramebufferAsync_(nullptr); FrameData &frame = frameData_[curFrame_]; frame.push_->End(); vulkan2D_.EndFrame(); curFrame_++; curFrame_ &= 1; } void FramebufferManagerVulkan::DeviceLost() { vulkan2D_.DeviceLost(); DestroyAllFBOs(); DestroyDeviceObjects(); } void FramebufferManagerVulkan::DeviceRestore(VulkanContext *vulkan) { vulkan_ = vulkan; vulkan2D_.DeviceRestore(vulkan_); InitDeviceObjects(); } std::vector FramebufferManagerVulkan::GetFramebufferList() { std::vector list; for (size_t i = 0; i < vfbs_.size(); ++i) { VirtualFramebuffer *vfb = vfbs_[i]; FramebufferInfo info; info.fb_address = vfb->fb_address; info.z_address = vfb->z_address; info.format = vfb->format; info.width = vfb->width; info.height = vfb->height; info.fbo = vfb->fbo; list.push_back(info); } return list; } void FramebufferManagerVulkan::DestroyAllFBOs() { currentRenderVfb_ = 0; displayFramebuf_ = 0; prevDisplayFramebuf_ = 0; prevPrevDisplayFramebuf_ = 0; for (size_t i = 0; i < vfbs_.size(); ++i) { VirtualFramebuffer *vfb = vfbs_[i]; INFO_LOG(FRAMEBUF, "Destroying FBO for %08x : %i x %i x %i", vfb->fb_address, vfb->width, vfb->height, vfb->format); DestroyFramebuf(vfb); } vfbs_.clear(); for (size_t i = 0; i < bvfbs_.size(); ++i) { VirtualFramebuffer *vfb = bvfbs_[i]; DestroyFramebuf(vfb); } bvfbs_.clear(); } void FramebufferManagerVulkan::FlushBeforeCopy() { // Flush anything not yet drawn before blitting, downloading, or uploading. // This might be a stalled list, or unflushed before a block transfer, etc. // TODO: It's really bad that we are calling SetRenderFramebuffer here with // all the irrelevant state checking it'll use to decide what to do. Should // do something more focused here. SetRenderFrameBuffer(gstate_c.IsDirty(DIRTY_FRAMEBUF), gstate_c.skipDrawReason); if (!draw_->GetNativeObject(Draw::NativeObject::CURRENT_RENDERPASS)) Crash(); drawEngine_->Flush(); } void FramebufferManagerVulkan::Resized() { FramebufferManagerCommon::Resized(); if (UpdateSize()) { DestroyAllFBOs(); } } bool FramebufferManagerVulkan::GetFramebuffer(u32 fb_address, int fb_stride, GEBufferFormat format, GPUDebugBuffer &buffer, int maxStride) { // TODO: Doing this synchronously will require stalling the pipeline. Maybe better // to do it callback-style? /* VirtualFramebuffer *vfb = currentRenderVfb_; if (!vfb) { vfb = GetVFBAt(fb_address); } if (!vfb) { // If there's no vfb and we're drawing there, must be memory? buffer = GPUDebugBuffer(Memory::GetPointer(fb_address | 0x04000000), fb_stride, 512, format); return true; } buffer.Allocate(vfb->renderWidth, vfb->renderHeight, GE_FORMAT_8888, false, true); if (vfb->fbo) fbo_bind_for_read(vfb->fbo); if (gl_extensions.GLES3 || !gl_extensions.IsGLES) glReadBuffer(GL_COLOR_ATTACHMENT0); glPixelStorei(GL_PACK_ALIGNMENT, 4); SafeGLReadPixels(0, 0, vfb->renderWidth, vfb->renderHeight, GL_RGBA, GL_UNSIGNED_BYTE, buffer.GetData()); */ return false; } bool FramebufferManagerVulkan::GetOutputFramebuffer(GPUDebugBuffer &buffer) { // TODO: Doing this synchronously will require stalling the pipeline. Maybe better // to do it callback-style? /* fbo_unbind_read(); int pw = PSP_CoreParameter().pixelWidth; int ph = PSP_CoreParameter().pixelHeight; // The backbuffer is flipped. buffer.Allocate(pw, ph, GPU_DBG_FORMAT_888_RGB, true); glPixelStorei(GL_PACK_ALIGNMENT, 1); SafeGLReadPixels(0, 0, pw, ph, GL_RGB, GL_UNSIGNED_BYTE, buffer.GetData()); */ return false; } bool FramebufferManagerVulkan::GetDepthbuffer(u32 fb_address, int fb_stride, u32 z_address, int z_stride, GPUDebugBuffer &buffer) { // TODO: Doing this synchronously will require stalling the pipeline. Maybe better // to do it callback-style? VirtualFramebuffer *vfb = currentRenderVfb_; if (!vfb) { vfb = GetVFBAt(fb_address); } if (!vfb) { // If there's no vfb and we're drawing there, must be memory? buffer = GPUDebugBuffer(Memory::GetPointer(z_address | 0x04000000), z_stride, 512, GPU_DBG_FORMAT_16BIT); return true; } /* buffer.Allocate(vfb->renderWidth, vfb->renderHeight, GPU_DBG_FORMAT_FLOAT, false); SafeGLReadPixels(0, 0, vfb->renderWidth, vfb->renderHeight, GL_DEPTH_COMPONENT, GL_FLOAT, buffer.GetData()); */ return false; } bool FramebufferManagerVulkan::GetStencilbuffer(u32 fb_address, int fb_stride, GPUDebugBuffer &buffer) { // TODO: Doing this synchronously will require stalling the pipeline. Maybe better // to do it callback-style? VirtualFramebuffer *vfb = currentRenderVfb_; if (!vfb) { vfb = GetVFBAt(fb_address); } if (!vfb) { // If there's no vfb and we're drawing there, must be memory? // TODO: Actually get the stencil. buffer = GPUDebugBuffer(Memory::GetPointer(fb_address | 0x04000000), fb_stride, 512, GPU_DBG_FORMAT_8888); return true; } /* buffer.Allocate(vfb->renderWidth, vfb->renderHeight, GPU_DBG_FORMAT_8BIT, false); SafeGLReadPixels(0, 0, vfb->renderWidth, vfb->renderHeight, GL_STENCIL_INDEX, GL_UNSIGNED_BYTE, buffer.GetData()); return true; */ return false; } void FramebufferManagerVulkan::ClearBuffer(bool keepState) { // TODO: Ideally, this should never be called. // assert(false); }