haze: use gpu console for rendering

This commit is contained in:
Liam 2023-10-13 21:44:52 -04:00 committed by SciresM
parent c866c15856
commit d9fff85bc4
6 changed files with 592 additions and 9 deletions

View File

@ -42,7 +42,10 @@ BUILD := build
SOURCES := source
DATA := data
INCLUDES := include ../../libraries/libvapours/include
#ROMFS := romfs
ROMFS := romfs
# Output folders for autogenerated files in romfs
OUT_SHADERS := shaders
APP_TITLE := USB File Transfer
APP_AUTHOR := Atmosphere-NX
@ -63,7 +66,7 @@ CXXFLAGS := $(CFLAGS) -fno-rtti -fno-exceptions -std=gnu++20
ASFLAGS := -g $(ARCH)
LDFLAGS = -specs=$(DEVKITPRO)/libnx/switch.specs -g $(ARCH) -Wl,-Map,$(notdir $*.map)
LIBS := -lnx
LIBS := -ldeko3d -lnx -lm
#---------------------------------------------------------------------------------
# list of directories containing libraries, this must be the top level containing
@ -90,6 +93,7 @@ export DEPSDIR := $(CURDIR)/$(BUILD)
CFILES := $(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.c)))
CPPFILES := $(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.cpp)))
SFILES := $(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.s)))
GLSLFILES := $(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.glsl)))
BINFILES := $(foreach dir,$(DATA),$(notdir $(wildcard $(dir)/*.*)))
#---------------------------------------------------------------------------------
@ -155,19 +159,61 @@ ifneq ($(APP_TITLEID),)
export NACPFLAGS += --titleid=$(APP_TITLEID)
endif
ifneq ($(ROMFS),)
ifneq ($(strip $(ROMFS)),)
ROMFS_TARGETS :=
ROMFS_FOLDERS :=
ifneq ($(strip $(OUT_SHADERS)),)
ROMFS_SHADERS := $(ROMFS)/$(OUT_SHADERS)
ROMFS_TARGETS += $(patsubst %.glsl, $(ROMFS_SHADERS)/%.dksh, $(GLSLFILES))
ROMFS_FOLDERS += $(ROMFS_SHADERS)
endif
export ROMFS_DEPS := $(foreach file,$(ROMFS_TARGETS),$(CURDIR)/$(file))
export NROFLAGS += --romfsdir=$(CURDIR)/$(ROMFS)
endif
.PHONY: $(BUILD) clean all
#---------------------------------------------------------------------------------
all: $(BUILD)
all: $(ROMFS_TARGETS) | $(BUILD)
$(BUILD):
@[ -d $@ ] || mkdir -p $@
@$(MAKE) --no-print-directory -C $(BUILD) -f $(CURDIR)/Makefile
ifneq ($(strip $(ROMFS_TARGETS)),)
$(ROMFS_TARGETS): | $(ROMFS_FOLDERS)
$(ROMFS_FOLDERS):
@mkdir -p $@
$(ROMFS_SHADERS)/%_vsh.dksh: %_vsh.glsl
@echo {vert} $(notdir $<)
@uam -s vert -o $@ $<
$(ROMFS_SHADERS)/%_tcsh.dksh: %_tcsh.glsl
@echo {tess_ctrl} $(notdir $<)
@uam -s tess_ctrl -o $@ $<
$(ROMFS_SHADERS)/%_tesh.dksh: %_tesh.glsl
@echo {tess_eval} $(notdir $<)
@uam -s tess_eval -o $@ $<
$(ROMFS_SHADERS)/%_gsh.dksh: %_gsh.glsl
@echo {geom} $(notdir $<)
@uam -s geom -o $@ $<
$(ROMFS_SHADERS)/%_fsh.dksh: %_fsh.glsl
@echo {frag} $(notdir $<)
@uam -s frag -o $@ $<
$(ROMFS_SHADERS)/%.dksh: %.glsl
@echo {comp} $(notdir $<)
@uam -s comp -o $@ $<
endif
#---------------------------------------------------------------------------------
clean:
@echo clean ...
@ -192,9 +238,9 @@ ifeq ($(strip $(APP_JSON)),)
all : $(OUTPUT).nro
ifeq ($(strip $(NO_NACP)),)
$(OUTPUT).nro : $(OUTPUT).elf $(OUTPUT).nacp
$(OUTPUT).nro : $(OUTPUT).elf $(OUTPUT).nacp $(ROMFS_DEPS)
else
$(OUTPUT).nro : $(OUTPUT).elf
$(OUTPUT).nro : $(OUTPUT).elf $(ROMFS_DEPS)
endif
else

View File

@ -143,7 +143,7 @@ namespace haze {
if (m_is_applet_mode) {
/* Print "Applet Mode" in red text. */
printf("\n" CONSOLE_ESC(38;5;196m) "Applet Mode" CONSOLE_ESC(0m) "\n");
printf("\n" CONSOLE_ESC(31;1m) "Applet Mode" CONSOLE_ESC(0m) "\n");
}
consoleUpdate(nullptr);

View File

@ -0,0 +1,15 @@
#version 460
layout (location = 0) noperspective in vec3 inTexCoord;
layout (location = 1) flat in vec4 inFrontPal;
layout (location = 2) flat in vec4 inBackPal;
layout (location = 0) out vec4 outColor;
layout (binding = 0) uniform sampler2DArray tileset;
void main()
{
float value = texture(tileset, inTexCoord).r;
outColor = mix(inBackPal, inFrontPal, value);
}

View File

@ -0,0 +1,35 @@
#version 460
layout (location = 0) in float inTileId;
layout (location = 1) in uvec2 inColorId;
layout (location = 0) out vec3 outTexCoord;
layout (location = 1) out vec4 outFrontPal;
layout (location = 2) out vec4 outBackPal;
layout (std140, binding = 0) uniform Config
{
vec4 dimensions;
vec4 vertices[3];
vec4 palettes[24];
} u;
void main()
{
float id = float(gl_InstanceID);
float tileRow = floor(id / u.dimensions.z);
float tileCol = id - tileRow * u.dimensions.z;
vec2 basePos;
basePos.x = 2.0 * (tileCol + 0.5) / u.dimensions.z - 1.0;
basePos.y = 2.0 * (1.0 - (tileRow + 0.5) / u.dimensions.w) - 1.0;
vec2 vtxData = u.vertices[gl_VertexID].xy;
vec2 scale = vec2(1.0) / u.dimensions.zw;
gl_Position.xy = vtxData * scale + basePos;
gl_Position.zw = vec2(0.5, 1.0);
outTexCoord = vec3(u.vertices[gl_VertexID].zw, inTileId);
outFrontPal = u.palettes[inColorId.x];
outBackPal = u.palettes[inColorId.y];
}

View File

@ -0,0 +1,487 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/iosupport.h>
#include <switch.h>
#include <deko3d.h>
// Define the desired number of framebuffers
#define FB_NUM 2
// Define the size of the memory block that will hold code
#define CODEMEMSIZE (64*1024)
// Define the size of the memory block that will hold command lists
#define CMDMEMSIZE (64*1024)
#define NUM_IMAGE_SLOTS 1
#define NUM_SAMPLER_SLOTS 1
typedef struct {
float pos[2];
float tex[2];
} VertexDef;
typedef struct {
float red;
float green;
float blue;
float alpha;
} PaletteColor;
typedef struct {
float dimensions[4];
VertexDef vertices[3];
PaletteColor palettes[24];
} ConsoleConfig;
static const VertexDef g_vertexData[3] = {
{ { 0.0f, +1.0f }, { 0.5f, 0.0f, } },
{ { -1.0f, -1.0f }, { 0.0f, 1.0f, } },
{ { +1.0f, -1.0f }, { 1.0f, 1.0f, } },
};
static const PaletteColor g_paletteData[24] = {
{ 0.0f, 0.0f, 0.0f, 0.0f }, // black
{ 0.5f, 0.0f, 0.0f, 1.0f }, // red
{ 0.0f, 0.5f, 0.0f, 1.0f }, // green
{ 0.5f, 0.5f, 0.0f, 1.0f }, // yellow
{ 0.0f, 0.0f, 0.5f, 1.0f }, // blue
{ 0.5f, 0.0f, 0.5f, 1.0f }, // magenta
{ 0.0f, 0.5f, 0.5f, 1.0f }, // cyan
{ 0.75f, 0.75f, 0.75f, 1.0f }, // white
{ 0.5f, 0.5f, 0.5f, 1.0f }, // bright black
{ 1.0f, 0.0f, 0.0f, 1.0f }, // bright red
{ 0.0f, 1.0f, 0.0f, 1.0f }, // bright green
{ 1.0f, 1.0f, 0.0f, 1.0f }, // bright yellow
{ 0.0f, 0.0f, 1.0f, 1.0f }, // bright blue
{ 1.0f, 0.0f, 1.0f, 1.0f }, // bright magenta
{ 0.0f, 1.0f, 1.0f, 1.0f }, // bright cyan
{ 1.0f, 1.0f, 1.0f, 1.0f }, // bright white
{ 0.0f, 0.0f, 0.0f, 0.0f }, // faint black
{ 0.25f, 0.0f, 0.0f, 1.0f }, // faint red
{ 0.0f, 0.25f, 0.0f, 1.0f }, // faint green
{ 0.25f, 0.25f, 0.0f, 1.0f }, // faint yellow
{ 0.0f, 0.0f, 0.25f, 1.0f }, // faint blue
{ 0.25f, 0.0f, 0.25f, 1.0f }, // faint magenta
{ 0.0f, 0.25f, 0.25f, 1.0f }, // faint cyan
{ 0.375f, 0.375f, 0.375f, 1.0f }, // faint white
};
typedef struct {
uint16_t tileId;
uint8_t frontPal;
uint8_t backPal;
} ConsoleChar;
static const DkVtxAttribState g_attribState[] = {
{ .bufferId=0, .isFixed=0, .offset=offsetof(ConsoleChar,tileId), .size=DkVtxAttribSize_1x16, .type=DkVtxAttribType_Uscaled, .isBgra=0 },
{ .bufferId=0, .isFixed=0, .offset=offsetof(ConsoleChar,frontPal), .size=DkVtxAttribSize_2x8, .type=DkVtxAttribType_Uint, .isBgra=0 },
};
static const DkVtxBufferState g_vtxbufState[] = {
{ .stride=sizeof(ConsoleChar), .divisor=1 },
};
struct GpuRenderer {
ConsoleRenderer base;
bool initialized;
DkDevice device;
DkQueue queue;
DkMemBlock imageMemBlock;
DkMemBlock codeMemBlock;
DkMemBlock dataMemBlock;
DkSwapchain swapchain;
DkImage framebuffers[FB_NUM];
DkImage tileset;
ConsoleChar* charBuf;
uint32_t codeMemOffset;
DkShader vertexShader;
DkShader fragmentShader;
DkCmdBuf cmdbuf;
DkCmdList cmdsBindFramebuffer[FB_NUM];
DkCmdList cmdsRender;
DkFence lastRenderFence;
};
static struct GpuRenderer* GpuRenderer(PrintConsole* con)
{
return (struct GpuRenderer*)con->renderer;
}
static void GpuRenderer_destroy(struct GpuRenderer* r)
{
// Make sure the queue is idle before destroying anything
dkQueueWaitIdle(r->queue);
// Destroy all the resources we've created
dkQueueDestroy(r->queue);
dkCmdBufDestroy(r->cmdbuf);
dkSwapchainDestroy(r->swapchain);
dkMemBlockDestroy(r->dataMemBlock);
dkMemBlockDestroy(r->codeMemBlock);
dkMemBlockDestroy(r->imageMemBlock);
dkDeviceDestroy(r->device);
// Clear out all state
memset(&r->initialized, 0, sizeof(*r) - offsetof(struct GpuRenderer, initialized));
}
// Simple function for loading a shader from the filesystem
static void GpuRenderer_loadShader(struct GpuRenderer* r, DkShader* pShader, const char* path)
{
// Open the file, and retrieve its size
FILE* f = fopen(path, "rb");
fseek(f, 0, SEEK_END);
uint32_t size = ftell(f);
rewind(f);
// Look for a spot in the code memory block for loading this shader. Note that
// we are just using a simple incremental offset; this isn't a general purpose
// allocation algorithm.
uint32_t codeOffset = r->codeMemOffset;
r->codeMemOffset += (size + DK_SHADER_CODE_ALIGNMENT - 1) &~ (DK_SHADER_CODE_ALIGNMENT - 1);
// Read the file into memory, and close the file
fread((uint8_t*)dkMemBlockGetCpuAddr(r->codeMemBlock) + codeOffset, size, 1, f);
fclose(f);
// Initialize the user provided shader object with the code we've just loaded
DkShaderMaker shaderMaker;
dkShaderMakerDefaults(&shaderMaker, r->codeMemBlock, codeOffset);
dkShaderInitialize(pShader, &shaderMaker);
}
static bool GpuRenderer_init(PrintConsole* con)
{
struct GpuRenderer* r = GpuRenderer(con);
if (r->initialized) {
// We're already initialized
return true;
}
// Create the deko3d device, which is the root object
DkDeviceMaker deviceMaker;
dkDeviceMakerDefaults(&deviceMaker);
r->device = dkDeviceCreate(&deviceMaker);
// Create the queue
DkQueueMaker queueMaker;
dkQueueMakerDefaults(&queueMaker, r->device);
queueMaker.flags = DkQueueFlags_Graphics;
r->queue = dkQueueCreate(&queueMaker);
// Calculate required width/height for the framebuffers
u32 width = con->font.tileWidth * con->consoleWidth;
u32 height = con->font.tileHeight * con->consoleHeight;
u32 totalConSize = con->consoleWidth * con->consoleHeight;
// Calculate layout for the framebuffers
DkImageLayoutMaker imageLayoutMaker;
dkImageLayoutMakerDefaults(&imageLayoutMaker, r->device);
imageLayoutMaker.flags = DkImageFlags_UsageRender | DkImageFlags_UsagePresent | DkImageFlags_HwCompression;
imageLayoutMaker.format = DkImageFormat_RGBA8_Unorm;
imageLayoutMaker.dimensions[0] = width;
imageLayoutMaker.dimensions[1] = height;
// Calculate layout for the framebuffers
DkImageLayout framebufferLayout;
dkImageLayoutInitialize(&framebufferLayout, &imageLayoutMaker);
// Calculate layout for the tileset
dkImageLayoutMakerDefaults(&imageLayoutMaker, r->device);
imageLayoutMaker.type = DkImageType_2DArray;
imageLayoutMaker.format = DkImageFormat_R32_Float;
imageLayoutMaker.dimensions[0] = con->font.tileWidth;
imageLayoutMaker.dimensions[1] = con->font.tileHeight;
imageLayoutMaker.dimensions[2] = con->font.numChars;
// Calculate layout for the tileset
DkImageLayout tilesetLayout;
dkImageLayoutInitialize(&tilesetLayout, &imageLayoutMaker);
// Retrieve necessary size and alignment for the framebuffers
uint32_t framebufferSize = dkImageLayoutGetSize(&framebufferLayout);
uint32_t framebufferAlign = dkImageLayoutGetAlignment(&framebufferLayout);
framebufferSize = (framebufferSize + framebufferAlign - 1) &~ (framebufferAlign - 1);
// Retrieve necessary size and alignment for the tileset
uint32_t tilesetSize = dkImageLayoutGetSize(&tilesetLayout);
uint32_t tilesetAlign = dkImageLayoutGetAlignment(&tilesetLayout);
tilesetSize = (tilesetSize + tilesetAlign - 1) &~ (tilesetAlign - 1);
// Create a memory block that will host the framebuffers and the tileset
DkMemBlockMaker memBlockMaker;
dkMemBlockMakerDefaults(&memBlockMaker, r->device, FB_NUM*framebufferSize + tilesetSize);
memBlockMaker.flags = DkMemBlockFlags_GpuCached | DkMemBlockFlags_Image;
r->imageMemBlock = dkMemBlockCreate(&memBlockMaker);
// Initialize the framebuffers with the layout and backing memory we've just created
DkImage const* swapchainImages[FB_NUM];
for (unsigned i = 0; i < FB_NUM; i ++) {
swapchainImages[i] = &r->framebuffers[i];
dkImageInitialize(&r->framebuffers[i], &framebufferLayout, r->imageMemBlock, i*framebufferSize);
}
// Create a swapchain out of the framebuffers we've just initialized
DkSwapchainMaker swapchainMaker;
dkSwapchainMakerDefaults(&swapchainMaker, r->device, nwindowGetDefault(), swapchainImages, FB_NUM);
r->swapchain = dkSwapchainCreate(&swapchainMaker);
// Initialize the tileset
dkImageInitialize(&r->tileset, &tilesetLayout, r->imageMemBlock, FB_NUM*framebufferSize);
// Create a memory block onto which we will load shader code
dkMemBlockMakerDefaults(&memBlockMaker, r->device, CODEMEMSIZE);
memBlockMaker.flags = DkMemBlockFlags_CpuUncached | DkMemBlockFlags_GpuCached | DkMemBlockFlags_Code;
r->codeMemBlock = dkMemBlockCreate(&memBlockMaker);
r->codeMemOffset = 0;
// Load our shaders (both vertex and fragment)
romfsInit();
GpuRenderer_loadShader(r, &r->vertexShader, "romfs:/shaders/console_vsh.dksh");
GpuRenderer_loadShader(r, &r->fragmentShader, "romfs:/shaders/console_fsh.dksh");
// Generate the descriptors
struct {
DkImageDescriptor images[NUM_IMAGE_SLOTS];
DkSamplerDescriptor samplers[NUM_SAMPLER_SLOTS];
} descriptors;
// Generate a image descriptor for the tileset
DkImageView tilesetView;
dkImageViewDefaults(&tilesetView, &r->tileset);
dkImageDescriptorInitialize(&descriptors.images[0], &tilesetView, false, false);
// Generate a sampler descriptor for the tileset
DkSampler sampler;
dkSamplerDefaults(&sampler);
sampler.wrapMode[0] = DkWrapMode_ClampToEdge;
sampler.wrapMode[1] = DkWrapMode_ClampToEdge;
sampler.minFilter = DkFilter_Nearest;
sampler.magFilter = DkFilter_Nearest;
dkSamplerDescriptorInitialize(&descriptors.samplers[0], &sampler);
uint32_t descriptorsOffset = CMDMEMSIZE;
uint32_t configOffset = (descriptorsOffset + sizeof(descriptors) + DK_UNIFORM_BUF_ALIGNMENT - 1) &~ (DK_UNIFORM_BUF_ALIGNMENT - 1);
uint32_t configSize = (sizeof(ConsoleConfig) + DK_UNIFORM_BUF_ALIGNMENT - 1) &~ (DK_UNIFORM_BUF_ALIGNMENT - 1);
uint32_t charBufOffset = configOffset + configSize;
uint32_t charBufSize = totalConSize * sizeof(ConsoleChar);
// Create a memory block which will be used for recording command lists using a command buffer
dkMemBlockMakerDefaults(&memBlockMaker, r->device,
(charBufOffset + charBufSize + DK_MEMBLOCK_ALIGNMENT - 1) &~ (DK_MEMBLOCK_ALIGNMENT - 1)
);
memBlockMaker.flags = DkMemBlockFlags_CpuUncached | DkMemBlockFlags_GpuCached;
r->dataMemBlock = dkMemBlockCreate(&memBlockMaker);
// Create a command buffer object
DkCmdBufMaker cmdbufMaker;
dkCmdBufMakerDefaults(&cmdbufMaker, r->device);
r->cmdbuf = dkCmdBufCreate(&cmdbufMaker);
// Feed our memory to the command buffer so that we can start recording commands
dkCmdBufAddMemory(r->cmdbuf, r->dataMemBlock, 0, CMDMEMSIZE);
// Create a temporary buffer that will hold the tileset
dkMemBlockMakerDefaults(&memBlockMaker, r->device,
(sizeof(float)*con->font.tileWidth*con->font.tileHeight*con->font.numChars + DK_MEMBLOCK_ALIGNMENT - 1) &~ (DK_MEMBLOCK_ALIGNMENT - 1)
);
memBlockMaker.flags = DkMemBlockFlags_CpuUncached | DkMemBlockFlags_GpuCached;
DkMemBlock scratchMemBlock = dkMemBlockCreate(&memBlockMaker);
float* scratchMem = (float*)dkMemBlockGetCpuAddr(scratchMemBlock);
// Unpack 1bpp tileset into a texture image the GPU can read
unsigned packedTileWidth = (con->font.tileWidth+7)/8;
for (unsigned tile = 0; tile < con->font.numChars; tile ++) {
const uint8_t* data = (const uint8_t*)con->font.gfx + con->font.tileHeight*packedTileWidth*tile;
for (unsigned y = 0; y < con->font.tileHeight; y ++) {
const uint8_t* row = &data[packedTileWidth*(y+1)];
uint8_t c = 0;
for (unsigned x = 0; x < con->font.tileWidth; x ++) {
if (!(x & 7))
c = *--row;
*scratchMem++ = (c & 0x80) ? 1.0f : 0.0f;
c <<= 1;
}
}
}
// Set up configuration
DkGpuAddr configAddr = dkMemBlockGetGpuAddr(r->dataMemBlock) + configOffset;
ConsoleConfig consoleConfig = {};
consoleConfig.dimensions[0] = width;
consoleConfig.dimensions[1] = height;
consoleConfig.dimensions[2] = con->consoleWidth;
consoleConfig.dimensions[3] = con->consoleHeight;
memcpy(consoleConfig.vertices, g_vertexData, sizeof(g_vertexData));
memcpy(consoleConfig.palettes, g_paletteData, sizeof(g_paletteData));
// Generate a temporary command list for uploading stuff and run it
DkGpuAddr descriptorSet = dkMemBlockGetGpuAddr(r->dataMemBlock) + descriptorsOffset;
DkCopyBuf copySrc = { dkMemBlockGetGpuAddr(scratchMemBlock), 0, 0 };
DkImageRect copyDst = { 0, 0, 0, con->font.tileWidth, con->font.tileHeight, con->font.numChars };
dkCmdBufPushData(r->cmdbuf, descriptorSet, &descriptors, sizeof(descriptors));
dkCmdBufPushConstants(r->cmdbuf, configAddr, configSize, 0, sizeof(consoleConfig), &consoleConfig);
dkCmdBufBindImageDescriptorSet(r->cmdbuf, descriptorSet, NUM_IMAGE_SLOTS);
dkCmdBufBindSamplerDescriptorSet(r->cmdbuf, descriptorSet + NUM_IMAGE_SLOTS*sizeof(DkImageDescriptor), NUM_SAMPLER_SLOTS);
dkCmdBufCopyBufferToImage(r->cmdbuf, &copySrc, &tilesetView, &copyDst, 0);
dkQueueSubmitCommands(r->queue, dkCmdBufFinishList(r->cmdbuf));
dkQueueFlush(r->queue);
dkQueueWaitIdle(r->queue);
dkCmdBufClear(r->cmdbuf);
// Destroy the scratch memory block since we don't need it anymore
dkMemBlockDestroy(scratchMemBlock);
// Retrieve the address of the character buffer
DkGpuAddr charBufAddr = dkMemBlockGetGpuAddr(r->dataMemBlock) + charBufOffset;
r->charBuf = (ConsoleChar*)((uint8_t*)dkMemBlockGetCpuAddr(r->dataMemBlock) + charBufOffset);
memset(r->charBuf, 0, charBufSize);
// Generate a command list for each framebuffer, which will bind each of them as a render target
for (unsigned i = 0; i < FB_NUM; i ++) {
DkImageView imageView;
dkImageViewDefaults(&imageView, &r->framebuffers[i]);
dkCmdBufBindRenderTarget(r->cmdbuf, &imageView, NULL);
r->cmdsBindFramebuffer[i] = dkCmdBufFinishList(r->cmdbuf);
}
// Declare structs that will be used for binding state
DkViewport viewport = { 0.0f, 0.0f, (float)width, (float)height, 0.0f, 1.0f };
DkScissor scissor = { 0, 0, width, height };
DkShader const* shaders[] = { &r->vertexShader, &r->fragmentShader };
DkRasterizerState rasterizerState;
DkColorState colorState;
DkColorWriteState colorWriteState;
// Initialize state structs with the deko3d defaults
dkRasterizerStateDefaults(&rasterizerState);
dkColorStateDefaults(&colorState);
dkColorWriteStateDefaults(&colorWriteState);
rasterizerState.fillRectangleEnable = true;
colorState.alphaCompareOp = DkCompareOp_Greater;
// Generate the main rendering command list
dkCmdBufSetViewports(r->cmdbuf, 0, &viewport, 1);
dkCmdBufSetScissors(r->cmdbuf, 0, &scissor, 1);
//dkCmdBufClearColorFloat(r->cmdbuf, 0, DkColorMask_RGBA, 0.125f, 0.294f, 0.478f, 0.0f);
dkCmdBufClearColorFloat(r->cmdbuf, 0, DkColorMask_RGBA, 0.0f, 0.0f, 0.0f, 0.0f);
dkCmdBufBindShaders(r->cmdbuf, DkStageFlag_GraphicsMask, shaders, sizeof(shaders)/sizeof(shaders[0]));
dkCmdBufBindRasterizerState(r->cmdbuf, &rasterizerState);
dkCmdBufBindColorState(r->cmdbuf, &colorState);
dkCmdBufBindColorWriteState(r->cmdbuf, &colorWriteState);
dkCmdBufBindUniformBuffer(r->cmdbuf, DkStage_Vertex, 0, configAddr, configSize);
dkCmdBufBindTexture(r->cmdbuf, DkStage_Fragment, 0, dkMakeTextureHandle(0, 0));
dkCmdBufBindVtxAttribState(r->cmdbuf, g_attribState, sizeof(g_attribState)/sizeof(g_attribState[0]));
dkCmdBufBindVtxBufferState(r->cmdbuf, g_vtxbufState, sizeof(g_vtxbufState)/sizeof(g_vtxbufState[0]));
dkCmdBufBindVtxBuffer(r->cmdbuf, 0, charBufAddr, charBufSize);
dkCmdBufSetAlphaRef(r->cmdbuf, 0.0f);
dkCmdBufDraw(r->cmdbuf, DkPrimitive_Triangles, 3, totalConSize, 0, 0);
r->cmdsRender = dkCmdBufFinishList(r->cmdbuf);
r->initialized = true;
return true;
}
static void GpuRenderer_deinit(PrintConsole* con)
{
struct GpuRenderer* r = GpuRenderer(con);
if (r->initialized) {
GpuRenderer_destroy(r);
}
}
static void GpuRenderer_drawChar(PrintConsole* con, int x, int y, int c)
{
struct GpuRenderer* r = GpuRenderer(con);
int writingColor = con->fg;
int screenColor = con->bg;
if (con->flags & CONSOLE_COLOR_BOLD) {
writingColor += 8;
} else if (con->flags & CONSOLE_COLOR_FAINT) {
writingColor += 16;
}
if (con->flags & CONSOLE_COLOR_REVERSE) {
int tmp = writingColor;
writingColor = screenColor;
screenColor = tmp;
}
// Wait for the fence
dkFenceWait(&r->lastRenderFence, UINT64_MAX);
ConsoleChar* pos = &r->charBuf[y*con->consoleWidth+x];
pos->tileId = c;
pos->frontPal = writingColor;
pos->backPal = screenColor;
}
static void GpuRenderer_scrollWindow(PrintConsole* con)
{
struct GpuRenderer* r = GpuRenderer(con);
// Wait for the fence
dkFenceWait(&r->lastRenderFence, UINT64_MAX);
// Perform the scrolling
for (int y = 0; y < con->windowHeight-1; y ++) {
memcpy(
&r->charBuf[(con->windowY+y+0)*con->consoleWidth + con->windowX],
&r->charBuf[(con->windowY+y+1)*con->consoleWidth + con->windowX],
sizeof(ConsoleChar)*con->windowWidth);
}
}
static void GpuRenderer_flushAndSwap(PrintConsole* con)
{
struct GpuRenderer* r = GpuRenderer(con);
// Acquire a framebuffer from the swapchain (and wait for it to be available)
int slot = dkQueueAcquireImage(r->queue, r->swapchain);
// Run the command list that binds said framebuffer as a render target
dkQueueSubmitCommands(r->queue, r->cmdsBindFramebuffer[slot]);
// Run the main rendering command list
dkQueueSubmitCommands(r->queue, r->cmdsRender);
// Signal the fence
dkQueueSignalFence(r->queue, &r->lastRenderFence, false);
// Now that we are done rendering, present it to the screen
dkQueuePresentImage(r->queue, r->swapchain, slot);
}
static struct GpuRenderer s_gpuRenderer =
{
{
GpuRenderer_init,
GpuRenderer_deinit,
GpuRenderer_drawChar,
GpuRenderer_scrollWindow,
GpuRenderer_flushAndSwap,
}
};
ConsoleRenderer* getDefaultConsoleRenderer(void)
{
return &s_gpuRenderer.base;
}

View File

@ -19,8 +19,8 @@ namespace haze {
namespace {
/* Allow 20MiB for use by libnx. */
static constexpr size_t LibnxReservedMemorySize = 20_MB;
/* Allow 30MiB for use by libnx. */
static constexpr size_t LibnxReservedMemorySize = 30_MB;
}