zba/src/ppu.zig

305 lines
9.4 KiB
Zig

const std = @import("std");
const io = @import("bus/io.zig");
const EventKind = @import("scheduler.zig").EventKind;
const Scheduler = @import("scheduler.zig").Scheduler;
const Bit = @import("bitfield").Bit;
const Bitfield = @import("bitfield").Bitfield;
const Allocator = std.mem.Allocator;
pub const width = 240;
pub const height = 160;
pub const framebuf_pitch = width * @sizeOf(u16);
pub const Ppu = struct {
const Self = @This();
// Registers
bg0: Background,
bg1: Background,
bg2: Background,
bg3: Background,
dispcnt: io.DisplayControl,
dispstat: io.DisplayStatus,
vcount: io.VCount,
vram: Vram,
palette: Palette,
oam: Oam,
sched: *Scheduler,
framebuf: []u8,
alloc: Allocator,
pub fn init(alloc: Allocator, sched: *Scheduler) !Self {
// Queue first Hblank
sched.push(.Draw, sched.tick + (240 * 4));
const framebuf = try alloc.alloc(u8, framebuf_pitch * height);
std.mem.set(u8, framebuf, 0);
return Self{
.vram = try Vram.init(alloc),
.palette = try Palette.init(alloc),
.oam = try Oam.init(alloc),
.sched = sched,
.framebuf = framebuf,
.alloc = alloc,
// Registers
.bg0 = Background.init(),
.bg1 = Background.init(),
.bg2 = Background.init(),
.bg3 = Background.init(),
.dispcnt = .{ .raw = 0x0000 },
.dispstat = .{ .raw = 0x0000 },
.vcount = .{ .raw = 0x0000 },
};
}
pub fn deinit(self: Self) void {
self.alloc.free(self.framebuf);
self.vram.deinit();
self.palette.deinit();
}
pub fn drawScanline(self: *Self) void {
const bg_mode = self.dispcnt.bg_mode.read();
const scanline = self.vcount.scanline.read();
switch (bg_mode) {
0x0 => {
// A Tile is always 8x8 pixels
// Mode 0 Implementation Assuming:
// - Scrolling isn't a thing
// - Bill Gates said we'll never need more than BG0
// Write to this Scanline once we're done
const start = framebuf_pitch * @as(usize, scanline);
var scanline_buf = std.mem.zeroes([framebuf_pitch]u8);
// These we can probably move to top level?
const charblock_len: u32 = 0x4000;
const screenblock_len: u32 = 0x800;
const cbb: u2 = self.bg0.cnt.char_base.read(); // Char Block Base
const sbb: u5 = self.bg0.cnt.screen_base.read(); // Screen Block Base
const is_8bpp: bool = self.bg0.cnt.palette_type.read(); // Colour Mode
const size: u2 = self.bg0.cnt.screen_size.read(); // Background Size
// 0x0600_000 is implied because we can access VRAM without the Bus
const char_base: u32 = charblock_len * @as(u32, cbb);
const screen_base: u32 = screenblock_len * @as(u32, sbb);
const y = @as(u32, scanline);
var x: u32 = 0;
while (x < width) : (x += 1) {
const entry_addr = screen_base + tilemapIndex(size, x, y);
const entry = @bitCast(ScreenEntry, @as(u16, self.vram.buf[entry_addr + 1]) << 8 | @as(u16, self.vram.buf[entry_addr]));
const tile_id: u32 = entry.tile_id.read();
const px_y = if (entry.h_flip.read()) 7 - (y % 8) else y % 8;
const px_x = if (entry.v_flip.read()) 7 - (x % 8) else x % 8;
const tile_addr = char_base + if (is_8bpp) 0x40 * tile_id + 0x8 * px_y else 0x20 * tile_id + 0x4 * px_y;
var tile = self.vram.buf[tile_addr + if (is_8bpp) px_x else px_x >> 1];
tile = if (px_x & 1 == 1) tile >> 4 else tile & 0xF;
const pal_bank: u8 = @as(u8, entry.palette_bank.read()) << 4;
const colour = pal_bank | tile;
std.mem.copy(u8, scanline_buf[x * 2 ..][0..2], self.palette.buf[colour * 2 ..][0..2]);
}
std.mem.copy(u8, self.framebuf[start..][0..framebuf_pitch], &scanline_buf);
},
0x3 => {
const start = framebuf_pitch * @as(usize, scanline);
std.mem.copy(u8, self.framebuf[start..][0..framebuf_pitch], self.vram.buf[start..][0..framebuf_pitch]);
},
0x4 => {
const select = self.dispcnt.frame_select.read();
const vram_start = width * @as(usize, scanline);
const buf_start = vram_start * @sizeOf(u16);
const start = vram_start + if (select) 0xA000 else @as(usize, 0);
const end = start + width; // Each Entry is only a byte long
// Render Current Scanline
for (self.vram.buf[start..end]) |byte, i| {
const id = byte * 2;
const j = i * @sizeOf(u16);
std.mem.copy(u8, self.framebuf[(buf_start + j)..][0..2], self.palette.buf[id..][0..2]);
}
},
else => std.debug.panic("[PPU] TODO: Implement BG Mode {}", .{bg_mode}),
}
}
fn tilemapIndex(size: u2, x: u32, y: u32) u32 {
return switch (size) {
0 => (((y % 256) / 8) * 64) + (((x % 256) / 8) * 2),
1 => (((y % 256) / 8) * 64) + (((x % 256) / 8) * 2),
else => std.debug.panic("tile size {}", .{size}),
};
}
};
const Palette = struct {
const Self = @This();
buf: []u8,
alloc: Allocator,
fn init(alloc: Allocator) !Self {
const buf = try alloc.alloc(u8, 0x400);
std.mem.set(u8, buf, 0);
return Self{
.buf = buf,
.alloc = alloc,
};
}
fn deinit(self: Self) void {
self.alloc.free(self.buf);
}
pub fn get32(self: *const Self, idx: usize) u32 {
return (@as(u32, self.get16(idx + 2)) << 16) | @as(u32, self.get16(idx));
}
pub fn set32(self: *Self, idx: usize, word: u32) void {
self.set16(idx + 2, @truncate(u16, word >> 16));
self.set16(idx, @truncate(u16, word));
}
pub fn get16(self: *const Self, idx: usize) u16 {
return (@as(u16, self.buf[idx + 1]) << 8) | @as(u16, self.buf[idx]);
}
pub fn set16(self: *Self, idx: usize, halfword: u16) void {
self.buf[idx + 1] = @truncate(u8, halfword >> 8);
self.buf[idx] = @truncate(u8, halfword);
}
pub fn get8(self: *const Self, idx: usize) u8 {
return self.buf[idx];
}
};
const Vram = struct {
const Self = @This();
buf: []u8,
alloc: Allocator,
fn init(alloc: Allocator) !Self {
const buf = try alloc.alloc(u8, 0x18000);
std.mem.set(u8, buf, 0);
return Self{
.buf = buf,
.alloc = alloc,
};
}
fn deinit(self: Self) void {
self.alloc.free(self.buf);
}
pub fn get32(self: *const Self, idx: usize) u32 {
return (@as(u32, self.get16(idx + 2)) << 16) | @as(u32, self.get16(idx));
}
pub fn set32(self: *Self, idx: usize, word: u32) void {
self.set16(idx + 2, @truncate(u16, word >> 16));
self.set16(idx, @truncate(u16, word));
}
pub fn get16(self: *const Self, idx: usize) u16 {
return (@as(u16, self.buf[idx + 1]) << 8) | @as(u16, self.buf[idx]);
}
pub fn set16(self: *Self, idx: usize, halfword: u16) void {
self.buf[idx + 1] = @truncate(u8, halfword >> 8);
self.buf[idx] = @truncate(u8, halfword);
}
pub fn get8(self: *const Self, idx: usize) u8 {
return self.buf[idx];
}
};
const Oam = struct {
const Self = @This();
buf: []u8,
alloc: Allocator,
fn init(alloc: Allocator) !Self {
const buf = try alloc.alloc(u8, 0x400);
std.mem.set(u8, buf, 0);
return Self{
.buf = buf,
.alloc = alloc,
};
}
pub fn get32(self: *const Self, idx: usize) u32 {
return (@as(u32, self.buf[idx + 3]) << 24) | (@as(u32, self.buf[idx + 2]) << 16) | (@as(u32, self.buf[idx + 1]) << 8) | (@as(u32, self.buf[idx]));
}
pub fn set32(self: *Self, idx: usize, word: u32) void {
self.buf[idx + 3] = @truncate(u8, word >> 24);
self.buf[idx + 2] = @truncate(u8, word >> 16);
self.buf[idx + 1] = @truncate(u8, word >> 8);
self.buf[idx] = @truncate(u8, word);
}
pub fn get16(self: *const Self, idx: usize) u16 {
return (@as(u16, self.buf[idx + 1]) << 8) | @as(u16, self.buf[idx]);
}
pub fn set16(self: *Self, idx: usize, halfword: u16) void {
self.buf[idx + 1] = @truncate(u8, halfword >> 8);
self.buf[idx] = @truncate(u8, halfword);
}
pub fn get8(self: *const Self, idx: usize) u8 {
return self.buf[idx];
}
};
const Background = struct {
const Self = @This();
/// Read / Write
cnt: io.BackgroundControl,
/// Write Only
hofs: io.BackgroundOffset,
/// Write Only
vofs: io.BackgroundOffset,
fn init() Self {
return .{
.cnt = .{ .raw = 0x0000 },
.hofs = .{ .raw = 0x0000 },
.vofs = .{ .raw = 0x0000 },
};
}
};
const ScreenEntry = extern union {
tile_id: Bitfield(u16, 0, 10),
h_flip: Bit(u16, 10),
v_flip: Bit(u16, 11),
palette_bank: Bitfield(u16, 12, 4),
raw: u16,
};