#include "common.h" #include "kernel/vc4_packet.h" /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkCreateImageView */ VkResult rpi_vkCreateImageView(VkDevice device, const VkImageViewCreateInfo* pCreateInfo, const VkAllocationCallbacks* pAllocator, VkImageView* pView) { assert(device); assert(pCreateInfo); assert(pView); _imageView* view = ALLOCATE(sizeof(_imageView), 1, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT); if(!view) { return VK_ERROR_OUT_OF_HOST_MEMORY; } view->image = pCreateInfo->image; view->viewType = pCreateInfo->viewType; view->interpretedFormat = pCreateInfo->format; view->swizzle = pCreateInfo->components; view->subresourceRange = pCreateInfo->subresourceRange; *pView = view; return VK_SUCCESS; } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkCreateBuffer */ VkResult rpi_vkCreateBuffer(VkDevice device, const VkBufferCreateInfo* pCreateInfo, const VkAllocationCallbacks* pAllocator, VkBuffer* pBuffer) { assert(device); assert(pCreateInfo); assert(pBuffer); _buffer* buf = ALLOCATE(sizeof(_buffer), 1, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT); if(!buf) { return VK_ERROR_OUT_OF_HOST_MEMORY; } buf->size = pCreateInfo->size; buf->usage = pCreateInfo->usage; buf->boundMem = 0; buf->alignment = ARM_PAGE_SIZE; buf->alignedSize = getBOAlignedSize(buf->size, ARM_PAGE_SIZE); *pBuffer = buf; return VK_SUCCESS; } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkGetBufferMemoryRequirements */ void rpi_vkGetBufferMemoryRequirements(VkDevice device, VkBuffer buffer, VkMemoryRequirements* pMemoryRequirements) { assert(device); assert(buffer); assert(pMemoryRequirements); pMemoryRequirements->alignment = ((_buffer*)buffer)->alignment; pMemoryRequirements->size = getBOAlignedSize(((_buffer*)buffer)->size, ARM_PAGE_SIZE); //there's only one memory type so that's gonna be it... pMemoryRequirements->memoryTypeBits = VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT | VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT; } VKAPI_ATTR void VKAPI_CALL rpi_vkGetBufferMemoryRequirements2( VkDevice device, const VkBufferMemoryRequirementsInfo2* pInfo, VkMemoryRequirements2* pMemoryRequirements) { assert(device); assert(pInfo); assert(pMemoryRequirements); rpi_vkGetBufferMemoryRequirements(device, pInfo->buffer, &pMemoryRequirements->memoryRequirements); } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkBindBufferMemory */ VkResult rpi_vkBindBufferMemory(VkDevice device, VkBuffer buffer, VkDeviceMemory memory, VkDeviceSize memoryOffset) { assert(device); assert(buffer); assert(memory); _buffer* buf = buffer; _deviceMemory* mem = memory; assert(!buf->boundMem); assert(memoryOffset < mem->size); assert(memoryOffset % buf->alignment == 0); assert(buf->alignedSize <= mem->size - memoryOffset); buf->boundMem = mem; buf->boundOffset = memoryOffset; return VK_SUCCESS; } void rpi_vkDestroyBuffer(VkDevice device, VkBuffer buffer, const VkAllocationCallbacks* pAllocator) { assert(device); _buffer* buf = buffer; if(buf) { FREE(buf); } } void rpi_vkDestroyImageView(VkDevice device, VkImageView imageView, const VkAllocationCallbacks* pAllocator) { assert(device); _imageView* view = imageView; if(view) { FREE(view); } } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkCreateBufferView */ VKAPI_ATTR VkResult VKAPI_CALL rpi_vkCreateBufferView( VkDevice device, const VkBufferViewCreateInfo* pCreateInfo, const VkAllocationCallbacks* pAllocator, VkBufferView* pView) { assert(device); assert(pCreateInfo); assert(pView); _bufferView* bv = ALLOCATE(sizeof(_bufferView), 1, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT); if(!bv) { return VK_ERROR_OUT_OF_HOST_MEMORY; } bv->buffer = pCreateInfo->buffer; bv->format = pCreateInfo->format; bv->offset = pCreateInfo->offset; bv->range = pCreateInfo->range; *pView = bv; return VK_SUCCESS; } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkDestroyBufferView */ VKAPI_ATTR void VKAPI_CALL rpi_vkDestroyBufferView( VkDevice device, VkBufferView bufferView, const VkAllocationCallbacks* pAllocator) { assert(device); _bufferView* bv = bufferView; if(bv) { FREE(bv); } } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkCreateImage */ VKAPI_ATTR VkResult VKAPI_CALL rpi_vkCreateImage( VkDevice device, const VkImageCreateInfo* pCreateInfo, const VkAllocationCallbacks* pAllocator, VkImage* pImage) { assert(device); assert(pCreateInfo); assert(pImage); _image* i = ALLOCATE(sizeof(_image), 1, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT); if(!i) { return VK_ERROR_OUT_OF_HOST_MEMORY; } i->flags = pCreateInfo->flags; i->type = pCreateInfo->imageType; i->fb = 0; //needed for modeset i->width = pCreateInfo->extent.width; i->height = pCreateInfo->extent.height; i->depth = pCreateInfo->extent.depth; i->paddedWidth = 0; //when format is T i->paddedHeight = 0; i->miplevels = pCreateInfo->mipLevels; memset(i->levelOffsets, 0, sizeof(uint32_t) * 11); i->samples = pCreateInfo->samples; i->layers = pCreateInfo->arrayLayers; i->size = 0; i->stride = 0; i->usageBits = pCreateInfo->usage; i->format = pCreateInfo->format; i->imageSpace = 0; uint32_t nonPaddedSize = (i->width * i->height * getFormatBpp(i->format)) >> 3; i->tiling = 0; if(pCreateInfo->tiling == VK_IMAGE_TILING_LINEAR) { i->tiling = VC4_TILING_FORMAT_LINEAR; } else { if(nonPaddedSize >= 4096) { i->tiling = VC4_TILING_FORMAT_T; } else { i->tiling = VC4_TILING_FORMAT_LT; } } i->layout = pCreateInfo->initialLayout; i->boundMem = 0; i->boundOffset = 0; i->alignment = ARM_PAGE_SIZE; i->concurrentAccess = pCreateInfo->sharingMode; //TODO? i->numQueueFamiliesWithAccess = pCreateInfo->queueFamilyIndexCount; if(i->numQueueFamiliesWithAccess > 0) { i->queueFamiliesWithAccess = ALLOCATE(sizeof(uint32_t) * i->numQueueFamiliesWithAccess, 1, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT); if(!i->queueFamiliesWithAccess) { return VK_ERROR_OUT_OF_HOST_MEMORY; } memcpy(i->queueFamiliesWithAccess, pCreateInfo->pQueueFamilyIndices, sizeof(uint32_t) * i->numQueueFamiliesWithAccess); } i->preTransformMode = 0; i->compositeAlpha = 0; i->presentMode = 0; i->clipped = 0; *pImage = i; return VK_SUCCESS; } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkDestroyImage */ VKAPI_ATTR void VKAPI_CALL rpi_vkDestroyImage( VkDevice device, VkImage image, const VkAllocationCallbacks* pAllocator) { assert(device); _image* i = image; if(i) { if(i->numQueueFamiliesWithAccess > 0) { FREE(i->queueFamiliesWithAccess); } FREE(i); } } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkGetImageMemoryRequirements */ VKAPI_ATTR void VKAPI_CALL rpi_vkGetImageMemoryRequirements( VkDevice device, VkImage image, VkMemoryRequirements* pMemoryRequirements) { assert(device); assert(image); assert(pMemoryRequirements); _image* i = image; uint32_t bpp = getFormatBpp(i->format); uint32_t utileW, utileH; getUTileDimensions(bpp, &utileW, &utileH); switch(i->tiling) { case VC4_TILING_FORMAT_T: { //need to pad to T format, as HW automatically chooses that i->paddedWidth = roundUp(i->width, utileW * 8); i->paddedHeight = roundUp(i->height, utileH * 8); break; } case VC4_TILING_FORMAT_LT: { //LT format i->paddedWidth = roundUp(i->width, utileW); i->paddedHeight = roundUp(i->height, utileH); break; } case VC4_TILING_FORMAT_LINEAR: { //linear format i->paddedWidth = roundUp(i->width, utileW); i->paddedHeight = i->height; break; } } i->stride = (i->paddedWidth * bpp) >> 3; //mip levels are laid out in memory the following way: //0x0.................................................0xffffff //smallest mip level ... largest mip level - 1, base mip level //base mip level offset must be a multiple of 4KB //mip levels other than the base must be aligned to power of two sizes //mip levels must be padded to either T or LT format depending on size uint32_t prevMipPaddedSize = 0; uint32_t mipSizes[11]; for(uint32_t c = i->miplevels - 1; c >= 1; --c) { uint32_t mipWidth = max(i->width >> c, 1); uint32_t mipHeight = max(i->height >> c, 1); uint32_t nonPaddedSize = (mipWidth * mipHeight * bpp) >> 3; uint32_t mipPaddedWidth, mipPaddedHeight; // uint32_t tiling = i->tiling; // if(i->tiling == VC4_TILING_FORMAT_T && nonPaddedSize <= 4096) // { // tiling = VC4_TILING_FORMAT_LT; // } // switch(tiling) // { // case VC4_TILING_FORMAT_T: // { // //T format // mipPaddedWidth = roundUp(mipWidth, utileW * 8); // mipPaddedHeight = roundUp(mipHeight, utileH * 8); // break; // } // case VC4_TILING_FORMAT_LT: // { // //LT format // mipPaddedWidth = roundUp(mipWidth, utileW); // mipPaddedHeight = roundUp(mipHeight, utileH); // break; // } // case VC4_TILING_FORMAT_LINEAR: // { // //linear format // mipPaddedWidth = roundUp(mipWidth, utileW); // mipPaddedHeight = mipHeight; // break; // } // } mipPaddedWidth = getPow2Pad(mipWidth); mipPaddedHeight = getPow2Pad(mipHeight); uint32_t greater = mipPaddedWidth > mipPaddedHeight ? mipPaddedWidth : mipPaddedHeight; greater = greater < 4 ? 4 : greater; mipPaddedWidth = mipPaddedHeight = greater; uint32_t mipPaddedSize = (mipPaddedWidth * mipPaddedHeight * bpp) >> 3; mipSizes[c] = mipPaddedSize; i->levelOffsets[c] = prevMipPaddedSize; prevMipPaddedSize += mipPaddedSize; // fprintf(stderr, "mipPaddedWidth: %u\n", mipPaddedWidth); // fprintf(stderr, "mipPaddedHeight: %u\n", mipPaddedHeight); // fprintf(stderr, "i->levelOffsets[%u]: %u\n", c, i->levelOffsets[c]); } //must be a multiple of 4096 bytes i->levelOffsets[0] = getBOAlignedSize(prevMipPaddedSize, 4096); for(uint32_t c = 1; c < i->miplevels; c++) { i->levelOffsets[c] = i->levelOffsets[c - 1] - mipSizes[c]; } i->size = getBOAlignedSize(((i->paddedWidth * i->paddedHeight * bpp) >> 3) + i->levelOffsets[0], ARM_PAGE_SIZE) * i->layers; // fprintf(stderr, "i->tiling %u\n", i->tiling); // fprintf(stderr, "i->levelOffsets[0] %u\n", i->levelOffsets[0]); // fprintf(stderr, "i->size %u\n", i->size); // fprintf(stderr, "i->paddedWidth %u\n", i->paddedWidth); // fprintf(stderr, "i->paddedHeight %u\n", i->paddedHeight); // fprintf(stderr, "mipSize %u\n", mipSize); // fprintf(stderr, "bpp %u\n", bpp); pMemoryRequirements->alignment = ARM_PAGE_SIZE; pMemoryRequirements->memoryTypeBits = memoryTypes[0].propertyFlags; //TODO pMemoryRequirements->size = i->size; } /* * https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#vkBindImageMemory */ VKAPI_ATTR VkResult VKAPI_CALL rpi_vkBindImageMemory( VkDevice device, VkImage image, VkDeviceMemory memory, VkDeviceSize memoryOffset) { assert(device); assert(image); assert(memory); _image* i = image; _deviceMemory* m = memory; assert(!i->boundMem); assert(memoryOffset < m->size); assert(memoryOffset % i->alignment == 0); assert(i->size <= m->size - memoryOffset); //image memory must be aligned to 4096 bytes //assert((memoryOffset % 4096) == 0); i->boundMem = m; i->boundOffset = memoryOffset; //TODO this is necessary, but maybe don't do it here? if(i->tiling == VC4_TILING_FORMAT_LINEAR) { int ret = vc4_bo_set_tiling(controlFd, i->boundMem->bo, DRM_FORMAT_MOD_LINEAR); assert(ret); } else if(i->tiling == VC4_TILING_FORMAT_T || i->tiling == VC4_TILING_FORMAT_LT) { int ret = vc4_bo_set_tiling(controlFd, i->boundMem->bo, DRM_FORMAT_MOD_BROADCOM_VC4_T_TILED); assert(ret); } return VK_SUCCESS; } VKAPI_ATTR VkResult VKAPI_CALL rpi_vkBindBufferMemory2( VkDevice device, uint32_t bindInfoCount, const VkBindBufferMemoryInfo* pBindInfos) { assert(device); assert(pBindInfos); VkResult ret = VK_SUCCESS; for(uint32_t c = 0; c < bindInfoCount; ++c) { VkResult res = rpi_vkBindBufferMemory(device, pBindInfos[c].buffer, pBindInfos[c].memory, pBindInfos[c].memoryOffset); if(res != VK_SUCCESS) { ret = res; } } return ret; } VKAPI_ATTR void VKAPI_CALL rpi_vkGetImageMemoryRequirements2( VkDevice device, const VkImageMemoryRequirementsInfo2* pInfo, VkMemoryRequirements2* pMemoryRequirements) { assert(device); assert(pInfo); assert(pMemoryRequirements); rpi_vkGetImageMemoryRequirements(device, pInfo->image, pMemoryRequirements); } VKAPI_ATTR VkResult VKAPI_CALL rpi_vkBindImageMemory2( VkDevice device, uint32_t bindInfoCount, const VkBindImageMemoryInfo* pBindInfos) { assert(device); assert(pBindInfos); VkResult ret = VK_SUCCESS; for(uint32_t c = 0; c < bindInfoCount; ++c) { VkResult res = rpi_vkBindImageMemory(device, pBindInfos[c].image, pBindInfos[c].memory, pBindInfos[c].memoryOffset); if(res != VK_SUCCESS) { ret = res; } } return ret; } VKAPI_ATTR void VKAPI_CALL rpi_vkCmdPushConstants( VkCommandBuffer commandBuffer, VkPipelineLayout layout, VkShaderStageFlags stageFlags, uint32_t offset, uint32_t size, const void* pValues) { assert(commandBuffer); assert(layout); _pipelineLayout* pl = layout; _commandBuffer* cb = commandBuffer; if(stageFlags & VK_SHADER_STAGE_VERTEX_BIT) { memcpy(cb->pushConstantBufferVertex + offset, pValues, size); } if(stageFlags & VK_SHADER_STAGE_FRAGMENT_BIT) { memcpy(cb->pushConstantBufferPixel + offset, pValues, size); } cb->pushConstantDirty = 1; } VKAPI_ATTR void VKAPI_CALL rpi_vkGetImageSubresourceLayout( VkDevice device, VkImage image, const VkImageSubresource* pSubresource, VkSubresourceLayout* pLayout) { //TODO }