#![allow(dead_code)] use memory::{VMem, PPUMemory, CPUBus}; use core::mem::{size_of, transmute}; use utils::{Read, Write, load_prefix, save_prefix}; pub trait Screen { fn put(&mut self, x: u8, y: u8, color: u8); fn render(&mut self); fn frame(&mut self); } #[repr(C, packed)] #[derive(Copy, Clone)] struct Sprite { y: u8, /* is the (actualy y) - 1 */ tile: u8, attr: u8, x: u8 } #[repr(C)] pub struct PPU<'a> { /*-- begin state --*/ pub scanline: u16, pub cycle: u16, /* cycle in the current scanline */ /* registers */ ppuctl: u8, ppumask: u8, ppustatus: u8, oamaddr: u8, reg: u8, x: u8, /* fine x scroll */ v: u16, /* current vram addr */ t: u16, /* temporary vram addr */ w: bool, /* first/second write toggle */ f: bool, /* if it is an odd frame */ /* rendering regs & latches */ /* background register (current two tiles) */ bg_pixel: u64, /* background latches for next tile */ bg_nt: u8, bg_attr: u8, bg_bit_low: u8, bg_bit_high: u8, /* sprites */ oam: [Sprite; 64], oam2: [usize; 8], sp_pixel: [u32; 8], sp_idx: [usize; 8], sp_cnt: [u8; 8], vblank: bool, pub vblank_lines: bool, buffered_read: u8, early_read: bool, /*-- end state --*/ /*-- begin sub-state --*/ mem: PPUMemory<'a>, /*-- end sub-state --*/ pub scr: &'a mut Screen, } macro_rules! PPU_IGNORED_SIZE { () => (size_of::() + size_of::<&mut Screen>()) } impl<'a> PPU<'a> { #[inline] pub fn write_ctl(&mut self, data: u8) { self.reg = data; self.ppuctl = data; self.t = (self.t & 0x73ff) | ((data as u16 & 3) << 10); } #[inline] pub fn write_mask(&mut self, data: u8) { self.reg = data; self.ppumask = data; } #[inline] pub fn read_status(&mut self) -> u8 { let res = (self.ppustatus & !0x1fu8) | (self.reg & 0x1f); self.ppustatus &= !PPU::FLAG_VBLANK; self.w = false; if self.scanline == 241 && self.cycle == 1 { self.early_read = true /* read before cycle 1 */ } res } #[inline] pub fn write_oamaddr(&mut self, data: u8) { self.reg = data; self.oamaddr = data; } #[inline] pub fn write_oamdata(&mut self, data: u8) { self.reg = data; self.get_oam_raw_mut()[self.oamaddr as usize] = data; self.oamaddr = self.oamaddr.wrapping_add(1); } #[inline] pub fn read_oamdata(&self) -> u8 { self.get_oam_raw()[self.oamaddr as usize] } #[inline] pub fn write_scroll(&mut self, data: u8) { self.reg = data; let data = data as u16; match self.w { false => { self.t = (self.t & 0x7fe0) | (data >> 3); self.x = (data & 0x07) as u8; self.w = true; }, true => { self.t = (self.t & 0x0c1f) | ((data & 0xf8) << 2) | ((data & 0x07) << 12); self.w = false; } } } #[inline] pub fn write_addr(&mut self, data: u8) { self.reg = data; let data = data as u16; match self.w { false => { self.t = (self.t & 0x00ff) | ((data & 0x3f) << 8); self.w = true; }, true => { self.t = (self.t & 0xff00) | data; self.v = self.t; self.w = false; } } } #[inline] pub fn read_data(&mut self) -> u8 { let data = self.mem.read(self.v); let res = if self.v & 0x3fff < 0x3f00 { let prev = self.buffered_read; self.buffered_read = data; prev } else { self.buffered_read = self.mem.read(self.v - 0x1000); data }; self.v = self.v.wrapping_add(match self.get_vram_inc() { 0 => 1, _ => 32 }); res } #[inline] pub fn write_data(&mut self, data: u8) { self.reg = data; self.mem.write(self.v, data); self.v = self.v.wrapping_add(match self.get_vram_inc() { 0 => 1, _ => 32 }); } #[inline] pub fn write_oamdma(&mut self, data: u8, bus: &CPUBus) { let cpu = bus.get_cpu(); self.reg = data; let mut addr = (data as u16) << 8; let stall = 1 + (cpu.cycle & 1) + 512; bus.cpu_stall(stall); let mut oamaddr = self.oamaddr; for _ in 0..stall - 0x100 { cpu.mem.bus.tick() } { let oam_raw = self.get_oam_raw_mut(); for _ in 0..0x100 { oam_raw[oamaddr as usize] = cpu.mem.read_without_tick(addr); addr = addr.wrapping_add(1); oamaddr = oamaddr.wrapping_add(1); } } self.oamaddr = oamaddr; } #[inline(always)] fn get_spritesize(&self) -> u8 {(self.ppuctl >> 5) & 1} #[inline(always)] pub fn get_flag_nmi(&self) -> bool { (self.ppuctl >> 7) == 1 } #[inline(always)] fn get_vram_inc(&self) -> u8 { (self.ppuctl >> 2) & 1} #[inline(always)] fn get_show_leftmost_bg(&self) -> bool { (self.ppumask >> 1) & 1 == 1} #[inline(always)] fn get_show_leftmost_sp(&self) -> bool { (self.ppumask >> 2) & 1 == 1} #[inline(always)] pub fn get_show_bg(&self) -> bool { (self.ppumask >> 3) & 1 == 1} #[inline(always)] pub fn get_show_sp(&self) -> bool { (self.ppumask >> 4) & 1 == 1} #[inline(always)] fn get_flag_vblank(&self) -> bool { (self.ppustatus >> 7) & 1 == 1 } #[inline(always)] fn get_oam_arr(&self) -> &[[u8; 4]; 64] { unsafe {transmute::<&[Sprite; 64], &[[u8; 4]; 64]>(&self.oam)} } #[inline(always)] fn get_oam_raw_mut(&mut self) -> &mut[u8; 256] { unsafe {transmute::<&mut[Sprite; 64], &mut[u8; 256]>(&mut self.oam)} } #[inline(always)] fn get_oam_raw(&self) -> &[u8; 256] { unsafe {transmute::<&[Sprite; 64], &[u8; 256]>(&self.oam)} } const FLAG_OVERFLOW: u8 = 1 << 5; const FLAG_SPRITE_ZERO: u8 = 1 << 6; const FLAG_VBLANK: u8 = 1 << 7; #[inline(always)] fn fetch_nametable_byte(&mut self) { self.bg_nt = self.mem.read_nametable(self.v & 0x0fff); } #[inline(always)] fn fetch_attrtable_byte(&mut self) { let v = self.v; /* the byte representing 4x4 tiles */ let b = self.mem.read_nametable(0x03c0 | (v & 0x0c00) | ((v >> 4) & 0x38) | ((v >> 2) & 0x07)); self.bg_attr = (b >> ((v & 2) | ((v & 0x40) >> 4))) & 3; } #[inline(always)] fn fetch_low_bgtile_byte(&mut self) { /* 0x?000 */ self.bg_bit_low = self.mem.read_mapper(((self.ppuctl as u16 & 0x10) << 8) | /* 0x-??0 */ ((self.bg_nt as u16) << 4) | /* 0x---? (0 - 7) */ ((self.v >> 12) & 7) | 0x0); } #[inline(always)] fn fetch_high_bgtile_byte(&mut self) { /* 0x?000 */ self.bg_bit_high = self.mem.read_mapper(((self.ppuctl as u16 & 0x10) << 8) | /* 0x-??0 */ ((self.bg_nt as u16) << 4) | /* 0x---? (8 - f) */ ((self.v >> 12) & 7) | 0x8); } #[inline(always)] fn load_bgtile(&mut self) { /* load the tile bitmap to high 8 bits of bitmap, * assume the high 8 bits are zeros */ debug_assert!(self.bg_pixel >> 32 == 0); let mut t: u64 = 0; let mut bl = self.bg_bit_low; let mut bh = self.bg_bit_high; for _ in 0..8 { t = (t << 4) | ((self.bg_attr << 2) | (bl & 1) | ((bh & 1) << 1)) as u64; bl >>= 1; bh >>= 1; } self.bg_pixel |= t << 32; } #[inline(always)] fn shift_sprites(&mut self) { for (i, c) in self.sp_cnt.iter_mut().enumerate() { if self.sp_idx[i] > 0xff { break } let c0 = *c; match c0 { 0 => self.sp_pixel[i] >>= 4, _ => *c = c0 - 1 } } } #[inline(always)] fn shift_bgtile(&mut self) { self.bg_pixel >>= 4; } #[inline(always)] fn wrapping_inc_cx(&mut self) { match self.v & 0x001f { 31 => { self.v &= !0x001fu16; /* reset coarse x */ self.v ^= 0x0400; /* switch horizontal nametable */ } _ => self.v += 1 } } #[inline(always)] fn wrapping_inc_y(&mut self) { match (self.v & 0x7000) == 0x7000 { false => self.v += 0x1000, /* fine y < 7 */ true => { self.v &= !0x7000u16; /* fine y <- 0 */ let y = match (self.v & 0x03e0) >> 5 { 29 => {self.v ^= 0x0800; 0}, /* at bottom of scanline */ 31 => 0, /* do not switch nt */ y => y + 1 }; self.v = (self.v & !0x03e0u16) | (y << 5); } } } #[inline(always)] fn reset_cx(&mut self) { self.v = (self.v & !0x041fu16) | (self.t & 0x041f); } #[inline(always)] fn reset_y(&mut self) { self.v = (self.v & !0x7be0u16) | (self.t & 0x7be0); } #[inline(always)] fn clear_sprite(&mut self) { debug_assert!(self.scanline != 261); self.oam2 = [0x100; 8]; } fn eval_sprite(&mut self) { debug_assert!(self.scanline != 261); /* we use scanline here because s.y is the (actual y) - 1 */ let mut nidx = 0; let mut n = 0; let scanline = self.scanline; let h = match self.get_spritesize() { 0 => 8, _ => 16 }; for (i, s) in self.oam.iter().enumerate() { let y = s.y as u16; if y <= scanline && scanline < y + h { self.oam2[nidx] = i; nidx += 1; if nidx == 8 { n = i + 1; break; } } } if nidx == 8 { let mut m = 0; let mut ppustatus = self.ppustatus; { let oam_raw = self.get_oam_arr(); while n < 64 { let y = oam_raw[n][m] as u16; if y <= scanline && scanline < y + h { ppustatus |= PPU::FLAG_OVERFLOW; /* set overflow */ } else { m = (m + 1) & 3; /* emulates hardware bug */ } n += 1; } } self.ppustatus = ppustatus; } } #[inline(always)] fn reverse_byte(mut x: u8) -> u8 { x = ((x & 0xaa) >> 1) | ((x & 0x55) << 1); x = ((x & 0xcc) >> 2) | ((x & 0x33) << 2); x = ((x & 0xf0) >> 4) | ((x & 0x0f) << 4); x } fn fetch_sprite(&mut self) { if self.scanline == 261 { return } /* we use scanline here because s.y is the (actual y) - 1 */ self.sp_idx = [0x100; 8]; for (i, v) in self.oam2.iter().enumerate() { let j = *v; if j > 0xff { break } let s = &self.oam[j]; let vflip = (s.attr & 0x80) == 0x80; let y0 = self.scanline - s.y as u16; let (ptable, tidx, y) = match self.get_spritesize() { 0 => { let y = if vflip {7 - y0 as u8} else {y0 as u8}; ((self.ppuctl as u16 & 0x08) << 9, s.tile, y) }, _ => { let y = if vflip {15 - y0 as u8} else {y0 as u8}; ((s.tile as u16 & 1) << 12, (s.tile & !1u8) | (y >> 3), y & 0x7) } }; self.sp_idx[i] = j; self.sp_cnt[i] = s.x; let mut low = self.mem.read_mapper(ptable | ((tidx as u16) << 4) | 0x0 | y as u16); let mut high = self.mem.read_mapper(ptable | ((tidx as u16) << 4) | 0x8 | y as u16); if (s.attr & 0x40) == 0x40 { low = PPU::reverse_byte(low); high = PPU::reverse_byte(high); } let attr = s.attr & 3; let mut t = 0u32; for _ in 0..8 { t = (t << 4) | ((attr << 2) | ((high & 1) << 1) | (low & 1)) as u32; high >>= 1; low >>= 1; } self.sp_pixel[i] = t; } } fn render_pixel(&mut self) { let x = self.cycle - 1; let bg = ((self.bg_pixel >> (self.x << 2)) & 0xf) as u16; let bg_pidx = if x >= 8 || self.get_show_leftmost_bg() { if self.get_show_bg() {bg & 3} else {0} } else {0}; let mut sp_pidx = 0x0; let mut pri = 0x1; let mut sp = 0; let show_sp = self.get_show_sp(); if x >= 8 || self.get_show_leftmost_sp() { for i in 0..8 { if self.sp_idx[i] > 0xff { break } if self.sp_cnt[i] != 0 { continue; } /* not active */ let s = &self.oam[self.sp_idx[i]]; sp = if show_sp {(self.sp_pixel[i] & 0xf) as u16} else { 0 }; match sp & 3 { 0x0 => (), pidx => { if bg_pidx != 0 && self.sp_idx[i] == 0 && x != 0xff && s.y != 0xff { self.ppustatus |= PPU::FLAG_SPRITE_ZERO; /* set sprite zero hit */ } sp_pidx = pidx; pri = (s.attr >> 5) & 1; break; } } } } debug_assert!(0 < self.cycle && self.cycle < 257); debug_assert!(self.scanline < 240); self.scr.put((self.cycle - 1) as u8, self.scanline as u8, self.mem.read_palette(if (pri == 0 || bg_pidx == 0) && sp_pidx != 0 { 0x0010 | sp } else { 0x0000 | match bg_pidx { 0 => 0, _ => bg } }) & 0x3f); } pub fn new(mem: PPUMemory<'a>, scr: &'a mut Screen) -> Self { let ppuctl = 0x00; let ppumask = 0x00; let ppustatus = 0xa0; let oamaddr = 0x00; let buffered_read = 0x00; let cycle = 0; let scanline = 241; PPU { scanline, ppuctl, ppumask, ppustatus, oamaddr, reg: 0, x: 0, v: 0, t: 0, w: false, f: true, cycle, bg_pixel: 0, bg_nt: 0, bg_attr: 0, bg_bit_low: 0, bg_bit_high: 0, oam: [Sprite{y: 0, tile: 0, attr: 0, x: 0}; 64], oam2: [0x100; 8], sp_idx: [0x100; 8], sp_pixel: [0; 8], sp_cnt: [0; 8], vblank: false, vblank_lines: true, buffered_read, early_read: false, mem, scr, } } pub fn load(&mut self, reader: &mut Read) -> bool { load_prefix(self, PPU_IGNORED_SIZE!(), reader) && self.mem.load(reader) } pub fn save(&self, writer: &mut Write) -> bool { save_prefix(self, PPU_IGNORED_SIZE!(), writer) && self.mem.save(writer) } pub fn reset(&mut self) { self.ppuctl = 0x00; self.ppumask = 0x00; self.ppustatus = self.ppustatus & 0x80; self.w = false; self.buffered_read = 0x00; self.cycle = 0; self.scanline = 241; self.vblank_lines = true; } #[inline(always)] pub fn try_nmi(&mut self) -> bool { self.get_flag_vblank() && self.get_flag_nmi() } pub fn tick(&mut self, bus: &CPUBus) -> bool { let res = self._tick(); self.mem.tick(bus); res } fn _tick(&mut self) -> bool { let cycle = self.cycle; if cycle == 0 { self.cycle = 1; if self.scanline == 240 { self.vblank_lines = true } else if self.scanline == 261 { self.vblank_lines = false } return false; } let rendering = self.get_show_bg() || self.get_show_sp(); let visible_line = self.scanline < 240; let pre_line = self.scanline == 261; if (pre_line || visible_line) && rendering { if pre_line && 279 < cycle && cycle < 305 { self.reset_y(); } else { let visible_cycle = 0 < cycle && cycle < 257; /* 1..256 */ let prefetch_cycle = 320 < cycle && cycle < 337; let fetch_cycle = visible_cycle || prefetch_cycle; if (visible_line && fetch_cycle) || (pre_line && prefetch_cycle) { match cycle & 0x7 { 1 => { self.load_bgtile(); self.fetch_nametable_byte(); }, 3 => self.fetch_attrtable_byte(), 5 => self.fetch_low_bgtile_byte(), 7 => self.fetch_high_bgtile_byte(), 0 => self.wrapping_inc_cx(), _ => () } match cycle { 1 => self.clear_sprite(), /* clear secondary OAM */ 65 => self.eval_sprite(), /* sprite evaluation */ 256 => self.wrapping_inc_y(), _ => () } if visible_cycle { self.render_pixel(); self.shift_sprites(); } self.shift_bgtile(); } else if cycle == 257 { /* we don't emulate fetch to per cycle precision because all data are fetched * from the secondary OAM which is not subject to any change during this * scanline */ self.reset_cx(); self.fetch_sprite(); self.cycle = 258; return false } /* skip at 338 because of 10-even_odd_timing test indicates an undocumented * behavior of NES */ if pre_line && cycle == 338 && self.f { self.cycle = 340; return false; } } } else { if !rendering { self.bg_pixel = 0 } if self.scanline == 241 && self.cycle == 1 { if !self.early_read { self.ppustatus |= PPU::FLAG_VBLANK } self.early_read = false; self.vblank = true; self.scr.render(); self.cycle = 2; return self.try_nmi() } } if pre_line && cycle == 1 { /* clear vblank, sprite zero hit & overflow */ self.vblank = false; self.ppustatus &= !(PPU::FLAG_VBLANK | PPU::FLAG_SPRITE_ZERO | PPU::FLAG_OVERFLOW); self.bg_pixel = 0; self.cycle = 2; return false } self.cycle += 1; if self.cycle > 340 { self.cycle = 0; self.scanline += 1; if self.scanline > 261 { self.scanline = 0; self.f = !self.f; } } false } }