From 8dbf5f6c17e3d217877119620e5bbfeba05abd7a Mon Sep 17 00:00:00 2001 From: David Guillen Fandos Date: Sun, 27 Jun 2021 01:16:28 +0200 Subject: Enable big-endian devices: gc/wii This patch adds big-endian compatibility in gpsp (in general but only for the interpreter). There's no performance hit for little-endian platforms (should be a no-op) and only add a small overhead in memory accesses for big-endian platforms. Most memory accesses are wrapped with a byteswap instruction and I/O reg accesses are also rewired for proper access (using macros). Video rendering has been fixed to also do byteswaps but there's a couple of games and rendering modes that still seem broken (but they amount to less than 20 games in my tests with 1K ROMs). This also adds build rules and CI for NGC/WII/WIIU (untested) --- .gitlab-ci.yml | 30 +++++++ Makefile | 8 ++ cheats.c | 2 +- common.h | 16 ++++ cpu.c | 26 +++--- gba_memory.c | 199 +++++++++++++++++++++--------------------- gba_memory.h | 13 +++ input.c | 4 +- main.c | 22 ++--- sound.c | 4 +- video.c | 267 ++++++++++++++++++++++++++++----------------------------- 11 files changed, 326 insertions(+), 265 deletions(-) diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index fdbedc4..62b38e9 100644 --- a/.gitlab-ci.yml +++ b/.gitlab-ci.yml @@ -63,6 +63,18 @@ include: - project: 'libretro-infrastructure/ci-templates' file: '/ctr-static.yml' + # Nintendo GameCube + - project: 'libretro-infrastructure/ci-templates' + file: '/ngc-static.yml' + + # Nintendo Wii + - project: 'libretro-infrastructure/ci-templates' + file: '/wii-static.yml' + + # Nintendo WiiU + - project: 'libretro-infrastructure/ci-templates' + file: '/wiiu-static.yml' + # Nintendo Switch - project: 'libretro-infrastructure/ci-templates' file: '/libnx-static.yml' @@ -174,6 +186,24 @@ libretro-build-ctr: - .libretro-ctr-static-retroarch-master - .core-defs +# Nintendo GameCube +libretro-build-ngc: + extends: + - .libretro-ngc-static-retroarch-master + - .core-defs + +# Nintendo Wii +libretro-build-wii: + extends: + - .libretro-wii-static-retroarch-master + - .core-defs + +# Nintendo WiiU +libretro-build-wiiu: + extends: + - .libretro-wiiu-static-retroarch-master + - .core-defs + # OpenDingux libretro-build-dingux-mips32: extends: diff --git a/Makefile b/Makefile index cf0bfd1..5f7afff 100644 --- a/Makefile +++ b/Makefile @@ -195,6 +195,14 @@ else ifeq ($(platform), switch) include $(LIBTRANSISTOR_HOME)/libtransistor.mk STATIC_LINKING=1 +# Nintendo Game Cube / Wii / WiiU +else ifneq (,$(filter $(platform), ngc wii wiiu)) + TARGET := $(TARGET_NAME)_libretro_$(platform).a + CC = $(DEVKITPPC)/bin/powerpc-eabi-gcc$(EXE_EXT) + AR = $(DEVKITPPC)/bin/powerpc-eabi-ar$(EXE_EXT) + CFLAGS += -DGEKKO -mcpu=750 -meabi -mhard-float -DHAVE_STRTOF_L + STATIC_LINKING = 1 + # PSP else ifeq ($(platform), psp1) TARGET := $(TARGET_NAME)_libretro_$(platform).a diff --git a/cheats.c b/cheats.c index 17556ca..3594730 100644 --- a/cheats.c +++ b/cheats.c @@ -190,7 +190,7 @@ void process_cheats(void) if(!cheats[i].cheat_active) continue; - process_cheat_codebreaker(&cheats[i], 0x3ff ^ io_registers[REG_P1]); + process_cheat_codebreaker(&cheats[i], 0x3ff ^ read_ioreg(REG_P1)); } } diff --git a/common.h b/common.h index ef6724a..ff4d522 100644 --- a/common.h +++ b/common.h @@ -124,6 +124,22 @@ typedef u32 fixed8_24; #define address32(base, offset) \ *((u32 *)((u8 *)base + (offset))) \ +#define eswap8(value) (value) +#if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ + #define eswap16(value) __builtin_bswap16(value) + #define eswap32(value) __builtin_bswap32(value) +#else + #define eswap16(value) (value) + #define eswap32(value) (value) +#endif + +#define readaddress8(base, offset) eswap8( address8( base, offset)) +#define readaddress16(base, offset) eswap16(address16(base, offset)) +#define readaddress32(base, offset) eswap32(address32(base, offset)) + +#define read_ioreg(regnum) (eswap16(io_registers[(regnum)])) +#define write_ioreg(regnum, val) io_registers[(regnum)] = eswap16(val) + #include #include #include diff --git a/cpu.c b/cpu.c index 6003849..873a5e3 100644 --- a/cpu.c +++ b/cpu.c @@ -721,8 +721,8 @@ u32 high_frequency_branch_targets = 0; // probably not worth optimizing for. #define check_for_interrupts() \ - if((io_registers[REG_IE] & io_registers[REG_IF]) && \ - io_registers[REG_IME] && ((reg[REG_CPSR] & 0x80) == 0)) \ + if((read_ioreg(REG_IE) & read_ioreg(REG_IF)) && \ + read_ioreg(REG_IME) && ((reg[REG_CPSR] & 0x80) == 0)) \ { \ reg_mode[MODE_IRQ][6] = reg[REG_PC] + 4; \ spsr[MODE_IRQ] = reg[REG_CPSR]; \ @@ -945,10 +945,10 @@ const u32 psr_masks[16] = #define aligned_address_mask16 0xF0000001 #define aligned_address_mask32 0xF0000003 -#define fast_read_memory(size, type, address, dest) \ +#define fast_read_memory(size, type, addr, dest) \ { \ u8 *map; \ - u32 _address = address; \ + u32 _address = addr; \ \ if(_address < 0x10000000) \ { \ @@ -964,7 +964,7 @@ const u32 psr_masks[16] = if(((_address & aligned_address_mask##size) == 0) && \ (map = memory_map_read[_address >> 15])) \ { \ - dest = *((type *)((u8 *)map + (_address & 0x7FFF))); \ + dest = (type)readaddress##size(map, (_address & 0x7FFF)); \ } \ else \ { \ @@ -984,7 +984,7 @@ const u32 psr_masks[16] = if(((_address & aligned_address_mask16) == 0) && \ (map = memory_map_read[_address >> 15])) \ { \ - dest = *((s16 *)((u8 *)map + (_address & 0x7FFF))); \ + dest = (s16)readaddress16(map, (_address & 0x7FFF)); \ } \ else \ { \ @@ -1018,7 +1018,7 @@ const u32 psr_masks[16] = } \ if(_address < 0x10000000 && map) \ { \ - dest = address32(map, _address & 0x7FFF); \ + dest = readaddress32(map, _address & 0x7FFF); \ } \ else \ { \ @@ -1123,10 +1123,10 @@ const u32 psr_masks[16] = #define arm_block_writeback_no(access_type) \ #define load_block_memory(address, dest) \ - dest = address32(address_region, (address + offset) & 0x7FFF) \ + dest = readaddress32(address_region, (address + offset) & 0x7FFF) \ #define store_block_memory(address, dest) \ - address32(address_region, (address + offset) & 0x7FFF) = dest \ + address32(address_region, (address + offset) & 0x7FFF) = eswap32(dest) \ #define arm_block_memory_offset_down_a() \ (base - (word_bit_count(reg_list) * 4) + 4) \ @@ -1601,9 +1601,9 @@ void raise_interrupt(irq_type irq_raised) { // The specific IRQ must be enabled in IE, master IRQ enable must be on, // and it must be on in the flags. - io_registers[REG_IF] |= irq_raised; + write_ioreg(REG_IF, read_ioreg(REG_IF) | irq_raised); - if((io_registers[REG_IE] & irq_raised) && io_registers[REG_IME] && + if((read_ioreg(REG_IE) & irq_raised) && read_ioreg(REG_IME) && ((reg[REG_CPSR] & 0x80) == 0)) { bios_read_protect = 0xe55ec002; @@ -1681,7 +1681,7 @@ arm_loop: using_instruction(arm); check_pc_region(); pc &= ~0x03; - opcode = address32(pc_address_block, (pc & 0x7FFF)); + opcode = readaddress32(pc_address_block, (pc & 0x7FFF)); condition = opcode >> 28; switch(condition) @@ -3256,7 +3256,7 @@ thumb_loop: using_instruction(thumb); check_pc_region(); pc &= ~0x01; - opcode = address16(pc_address_block, (pc & 0x7FFF)); + opcode = readaddress16(pc_address_block, (pc & 0x7FFF)); switch((opcode >> 8) & 0xFF) { diff --git a/gba_memory.c b/gba_memory.c index af94585..4376cf8 100644 --- a/gba_memory.c +++ b/gba_memory.c @@ -35,7 +35,7 @@ gbc_sound_channel[channel].envelope_status = (envelope_ticks != 0); \ gbc_sound_channel[channel].envelope_volume = initial_volume; \ gbc_sound_update = 1; \ - address16(io_registers, address) = value; \ + address16(io_registers, address) = eswap16(value); \ } \ #define gbc_sound_tone_control_high(channel, address) \ @@ -56,7 +56,7 @@ } \ \ gbc_sound_update = 1; \ - address16(io_registers, address) = value; \ + address16(io_registers, address) = eswap16(value); \ } \ #define gbc_sound_tone_control_sweep() \ @@ -68,7 +68,7 @@ gbc_sound_channel[0].sweep_ticks = sweep_ticks; \ gbc_sound_channel[0].sweep_initial_ticks = sweep_ticks; \ gbc_sound_update = 1; \ - address16(io_registers, 0x60) = value; \ + write_ioreg(REG_SOUND1CNT_L, value); \ } \ #define gbc_sound_wave_control() \ @@ -80,7 +80,7 @@ gbc_sound_channel[2].master_enable = 1; \ \ gbc_sound_update = 1; \ - address16(io_registers, 0x70) = value; \ + write_ioreg(REG_SOUND3CNT_L, value); \ } \ static u32 gbc_sound_wave_volume[4] = { 0, 16384, 8192, 4096 }; @@ -94,7 +94,7 @@ static u32 gbc_sound_wave_volume[4] = { 0, 16384, 8192, 4096 }; gbc_sound_channel[2].wave_volume = \ gbc_sound_wave_volume[(value >> 13) & 0x03]; \ gbc_sound_update = 1; \ - address16(io_registers, 0x72) = value; \ + write_ioreg(REG_SOUND3CNT_H, value); \ } \ #define gbc_sound_tone_control_high_wave() \ @@ -110,7 +110,7 @@ static u32 gbc_sound_wave_volume[4] = { 0, 16384, 8192, 4096 }; gbc_sound_channel[2].active_flag = 1; \ } \ gbc_sound_update = 1; \ - address16(io_registers, 0x74) = value; \ + write_ioreg(REG_SOUND3CNT_X, value); \ } \ #define gbc_sound_noise_control() \ @@ -141,7 +141,7 @@ static u32 gbc_sound_wave_volume[4] = { 0, 16384, 8192, 4096 }; gbc_sound_channel[3].envelope_initial_volume; \ } \ gbc_sound_update = 1; \ - address16(io_registers, 0x7C) = value; \ + write_ioreg(REG_SOUND4CNT_H, value); \ } \ static void gbc_trigger_sound(u32 value) @@ -156,7 +156,7 @@ static void gbc_trigger_sound(u32 value) gbc_sound_channel[channel].status = ((value >> (channel + 8)) & 0x01) | ((value >> (channel + 11)) & 0x03); } - address16(io_registers, 0x80) = value; + write_ioreg(REG_SOUNDCNT_L, value); } #define trigger_sound() \ @@ -175,7 +175,7 @@ static void gbc_trigger_sound(u32 value) sound_reset_fifo(0); \ if((value >> 15) & 0x01) \ sound_reset_fifo(1); \ - address16(io_registers, 0x82) = value; \ + write_ioreg(REG_SOUNDCNT_H, value); \ } \ static void sound_control_x(u32 value) @@ -193,8 +193,8 @@ static void sound_control_x(u32 value) sound_on = 0; } - address16(io_registers, 0x84) = - (address16(io_registers, 0x84) & 0x000F) | (value & 0xFFF0); + address16(io_registers, 0x84) = eswap16( + (readaddress16(io_registers, 0x84) & 0x000F) | (value & 0xFFF0)); } #define sound_update_frequency_step(timer_number) \ @@ -238,7 +238,7 @@ static void trigger_timer(u32 timer_number, u32 value) timer[timer_number].prescale = prescale; timer[timer_number].irq = (value >> 6) & 0x01; - address16(io_registers, 0x100 + (timer_number * 4)) = -timer_reload; + write_ioreg(REG_TM0D + (timer_number * 2), (u32)(-timer_reload)); timer_reload <<= prescale; timer[timer_number].count = timer_reload; @@ -266,7 +266,7 @@ static void trigger_timer(u32 timer_number, u32 value) timer[timer_number].stop_cpu_ticks = cpu_ticks; } } - address16(io_registers, 0x102 + (timer_number * 4)) = value; + write_ioreg(REG_TM0CNT + (timer_number * 2), value); } // This table is configured for sequential access on system defaults @@ -458,7 +458,7 @@ void function_cc write_eeprom(u32 unused_address, u32 value) if(eeprom_size == EEPROM_512_BYTE) { eeprom_address = - (address16(eeprom_buffer, 0) >> 2) * 8; + (readaddress16(eeprom_buffer, 0) >> 2) * 8; } else { @@ -508,10 +508,10 @@ void function_cc write_eeprom(u32 unused_address, u32 value) u32 gamepak_index = address >> 15; \ u8 *map = memory_map_read[gamepak_index]; \ \ - if(!map) \ + if(!map) \ map = load_gamepak_page(gamepak_index & 0x3FF); \ \ - value = address##type(map, address & 0x7FFF) \ + value = readaddress##type(map, address & 0x7FFF) \ #define read_open8() \ if(!(reg[REG_CPSR] & 0x20)) \ @@ -580,29 +580,29 @@ u32 function_cc read_eeprom(void) case 0x00: \ /* BIOS */ \ if(reg[REG_PC] >= 0x4000) \ - value = address##type(&bios_read_protect, address & 0x03); \ + value = readaddress##type(&bios_read_protect, address & 0x03); \ else \ - value = address##type(bios_rom, address & 0x3FFF); \ + value = readaddress##type(bios_rom, address & 0x3FFF); \ break; \ \ case 0x02: \ /* external work RAM */ \ - value = address##type(ewram, (address & 0x3FFFF)); \ + value = readaddress##type(ewram, (address & 0x3FFFF)); \ break; \ \ case 0x03: \ /* internal work RAM */ \ - value = address##type(iwram, (address & 0x7FFF) + 0x8000); \ + value = readaddress##type(iwram, (address & 0x7FFF) + 0x8000); \ break; \ \ case 0x04: \ /* I/O registers */ \ - value = address##type(io_registers, address & 0x3FF); \ + value = readaddress##type(io_registers, address & 0x3FF); \ break; \ \ case 0x05: \ /* palette RAM */ \ - value = address##type(palette_ram, address & 0x3FF); \ + value = readaddress##type(palette_ram, address & 0x3FF); \ break; \ \ case 0x06: \ @@ -611,12 +611,12 @@ u32 function_cc read_eeprom(void) if(address > 0x18000) \ address -= 0x8000; \ \ - value = address##type(vram, address); \ + value = readaddress##type(vram, address); \ break; \ \ case 0x07: \ /* OAM RAM */ \ - value = address##type(oam_ram, address & 0x3FF); \ + value = readaddress##type(oam_ram, address & 0x3FF); \ break; \ \ case 0x08: \ @@ -660,12 +660,12 @@ static cpu_alert_type trigger_dma(u32 dma_number, u32 value) if(dma[dma_number].start_type == DMA_INACTIVE) { u32 start_type = (value >> 12) & 0x03; - u32 dest_address = address32(io_registers, (dma_number * 12) + 0xB4) & + u32 dest_address = readaddress32(io_registers, (dma_number * 12) + 0xB4) & 0xFFFFFFF; dma[dma_number].dma_channel = dma_number; dma[dma_number].source_address = - address32(io_registers, (dma_number * 12) + 0xB0) & 0xFFFFFFF; + readaddress32(io_registers, (dma_number * 12) + 0xB0) & 0xFFFFFFF; dma[dma_number].dest_address = dest_address; dma[dma_number].source_direction = (value >> 7) & 0x03; dma[dma_number].repeat_type = (value >> 9) & 0x01; @@ -686,7 +686,7 @@ static cpu_alert_type trigger_dma(u32 dma_number, u32 value) } else { - u32 length = address16(io_registers, (dma_number * 12) + 0xB8); + u32 length = read_ioreg(REG_DMA0CNT_L + (dma_number * 6)); if((dma_number == 3) && ((dest_address >> 24) == 0x0D) && ((length & 0x1F) == 17)) @@ -708,7 +708,7 @@ static cpu_alert_type trigger_dma(u32 dma_number, u32 value) dma[dma_number].dest_direction = (value >> 5) & 0x03; } - address16(io_registers, (dma_number * 12) + 0xBA) = value; + write_ioreg(REG_DMA0CNT_H + (dma_number * 6), value); if(start_type == DMA_START_IMMEDIATELY) return dma_transfer(dma + dma_number); } @@ -717,7 +717,7 @@ static cpu_alert_type trigger_dma(u32 dma_number, u32 value) { dma[dma_number].start_type = DMA_INACTIVE; dma[dma_number].direct_sound_channel = DMA_NO_DIRECT_SOUND; - address16(io_registers, (dma_number * 12) + 0xBA) = value; + write_ioreg(REG_DMA0CNT_H + (dma_number * 6), value); } return CPU_ALERT_NONE; @@ -731,10 +731,10 @@ static cpu_alert_type trigger_dma(u32 dma_number, u32 value) value = ((address8(io_registers, address + 1)) << 8) | value \ #define access_register16_high(address) \ - value = (value << 16) | (address16(io_registers, address)) \ + value = (value << 16) | (readaddress16(io_registers, address)) \ #define access_register16_low(address) \ - value = ((address16(io_registers, address + 2)) << 16) | value \ + value = ((readaddress16(io_registers, address + 2)) << 16) | value \ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) { @@ -743,7 +743,7 @@ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) { case 0x00: { - u32 dispcnt = io_registers[REG_DISPCNT]; + u32 dispcnt = read_ioreg(REG_DISPCNT); if((value & 0x07) != (dispcnt & 0x07)) reg[OAM_UPDATED] = 1; @@ -768,28 +768,28 @@ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) access_register8_low(0x28); access_register16_low(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; case 0x29: access_register8_high(0x28); access_register16_low(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; case 0x2A: access_register8_low(0x2A); access_register16_high(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; case 0x2B: access_register8_high(0x2A); access_register16_high(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; // BG2 reference Y @@ -797,28 +797,28 @@ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) access_register8_low(0x2C); access_register16_low(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; case 0x2D: access_register8_high(0x2C); access_register16_low(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; case 0x2E: access_register8_low(0x2E); access_register16_high(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; case 0x2F: access_register8_high(0x2E); access_register16_high(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; // BG3 reference X @@ -826,28 +826,28 @@ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) access_register8_low(0x38); access_register16_low(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; case 0x39: access_register8_high(0x38); access_register16_low(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; case 0x3A: access_register8_low(0x3A); access_register16_high(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; case 0x3B: access_register8_high(0x3A); access_register16_high(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; // BG3 reference Y @@ -855,28 +855,28 @@ cpu_alert_type function_cc write_io_register8(u32 address, u32 value) access_register8_low(0x3C); access_register16_low(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; case 0x3D: access_register8_high(0x3C); access_register16_low(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; case 0x3E: access_register8_low(0x3E); access_register16_high(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; case 0x3F: access_register8_high(0x3E); access_register16_high(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; // Sound 1 control sweep @@ -1160,18 +1160,17 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) { case 0x00: { - u32 dispcnt = io_registers[REG_DISPCNT]; + u32 dispcnt = read_ioreg(REG_DISPCNT); if((value & 0x07) != (dispcnt & 0x07)) reg[OAM_UPDATED] = 1; - address16(io_registers, 0x00) = value; + write_ioreg(REG_DISPCNT, value); break; } // DISPSTAT case 0x04: - address16(io_registers, 0x04) = - (address16(io_registers, 0x04) & 0x07) | (value & ~0x07); + write_ioreg(REG_DISPSTAT, (read_ioreg(REG_DISPSTAT) & 0x07) | (value & ~0x07)); break; // VCOUNT @@ -1182,26 +1181,26 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) case 0x28: access_register16_low(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; case 0x2A: access_register16_high(0x28); affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; // BG2 reference Y case 0x2C: access_register16_low(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; case 0x2E: access_register16_high(0x2C); affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; // BG3 reference X @@ -1209,26 +1208,26 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) case 0x38: access_register16_low(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; case 0x3A: access_register16_high(0x38); affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; // BG3 reference Y case 0x3C: access_register16_low(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; case 0x3E: access_register16_high(0x3C); affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; // Sound 1 control sweep @@ -1313,7 +1312,7 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) case 0x9D: case 0x9E: gbc_sound_wave_update = 1; - address16(io_registers, address) = value; + address16(io_registers, address) = eswap16(value); break; // Sound FIFO A @@ -1382,7 +1381,7 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) // Interrupt flag case 0x202: - address16(io_registers, 0x202) &= ~value; + write_ioreg(REG_IF, read_ioreg(REG_IF) & (~value)); break; // WAITCNT @@ -1399,7 +1398,7 @@ cpu_alert_type function_cc write_io_register16(u32 address, u32 value) return CPU_ALERT_HALT; default: - address16(io_registers, address) = value; + address16(io_registers, address) = eswap16(value); break; } @@ -1414,25 +1413,25 @@ cpu_alert_type function_cc write_io_register32(u32 address, u32 value) // BG2 reference X case 0x28: affine_reference_x[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x28) = value; + address32(io_registers, 0x28) = eswap32(value); break; // BG2 reference Y case 0x2C: affine_reference_y[0] = (s32)(value << 4) >> 4; - address32(io_registers, 0x2C) = value; + address32(io_registers, 0x2C) = eswap32(value); break; // BG3 reference X case 0x38: affine_reference_x[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x38) = value; + address32(io_registers, 0x38) = eswap32(value); break; // BG3 reference Y case 0x3C: affine_reference_y[1] = (s32)(value << 4) >> 4; - address32(io_registers, 0x3C) = value; + address32(io_registers, 0x3C) = eswap32(value); break; // Sound FIFO A @@ -1468,7 +1467,7 @@ cpu_alert_type function_cc write_io_register32(u32 address, u32 value) #define write_palette16(address, value) \ { \ u32 palette_address = address; \ - address16(palette_ram, palette_address) = value; \ + address16(palette_ram, palette_address) = eswap16(value); \ convert_palette(value); \ address16(palette_ram_converted, palette_address) = value; \ } \ @@ -1478,11 +1477,11 @@ cpu_alert_type function_cc write_io_register32(u32 address, u32 value) u32 palette_address = address; \ u32 value_high = value >> 16; \ u32 value_low = value & 0xFFFF; \ - address32(palette_ram, palette_address) = value; \ + address32(palette_ram, palette_address) = eswap32(value); \ convert_palette(value_high); \ + address16(palette_ram_converted, palette_address + 2) = value_high; \ convert_palette(value_low); \ - value = (value_high << 16) | value_low; \ - address32(palette_ram_converted, palette_address) = value; \ + address16(palette_ram_converted, palette_address) = value_low; \ } \ @@ -1616,13 +1615,13 @@ void function_cc write_backup(u32 address, u32 value) #define write_vram8() \ address &= ~0x01; \ - address16(vram, address) = ((value << 8) | value) \ + address16(vram, address) = eswap16((value << 8) | value) \ #define write_vram16() \ - address16(vram, address) = value \ + address16(vram, address) = eswap16(value) \ #define write_vram32() \ - address32(vram, address) = value \ + address32(vram, address) = eswap32(value) \ // RTC code derived from VBA's (due to lack of any real publically available // documentation...) @@ -1682,7 +1681,7 @@ static u32 encode_bcd(u8 value) if(!map) \ map = load_gamepak_page(rtc_page_index & 0x3FF); \ \ - address16(map, update_address & 0x7FFF) = _value \ + address16(map, update_address & 0x7FFF) = eswap16(_value) \ void function_cc write_rtc(u32 address, u32 value) { @@ -1897,12 +1896,12 @@ void function_cc write_rtc(u32 address, u32 value) { \ case 0x02: \ /* external work RAM */ \ - address##type(ewram, (address & 0x3FFFF)) = value; \ + address##type(ewram, (address & 0x3FFFF)) = eswap##type(value); \ break; \ \ case 0x03: \ /* internal work RAM */ \ - address##type(iwram, (address & 0x7FFF) + 0x8000) = value; \ + address##type(iwram, (address & 0x7FFF) + 0x8000) = eswap##type(value); \ break; \ \ case 0x04: \ @@ -1926,7 +1925,7 @@ void function_cc write_rtc(u32 address, u32 value) case 0x07: \ /* OAM RAM */ \ reg[OAM_UPDATED] = 1; \ - address##type(oam_ram, address & 0x3FF) = value; \ + address##type(oam_ram, address & 0x3FF) = eswap##type(value); \ break; \ \ case 0x08: \ @@ -2544,26 +2543,26 @@ dma_region_type dma_region_map[16] = } \ #define dma_read_iwram(type, transfer_size) \ - read_value = address##transfer_size(iwram + 0x8000, type##_ptr & 0x7FFF) \ + read_value = readaddress##transfer_size(iwram + 0x8000, type##_ptr & 0x7FFF)\ #define dma_read_vram(type, transfer_size) \ - read_value = address##transfer_size(vram, type##_ptr & 0x1FFFF) \ + read_value = readaddress##transfer_size(vram, type##_ptr & 0x1FFFF) \ #define dma_read_io(type, transfer_size) \ - read_value = address##transfer_size(io_registers, type##_ptr & 0x7FFF) \ + read_value = readaddress##transfer_size(io_registers, type##_ptr & 0x7FFF) \ #define dma_read_oam_ram(type, transfer_size) \ - read_value = address##transfer_size(oam_ram, type##_ptr & 0x3FF) \ + read_value = readaddress##transfer_size(oam_ram, type##_ptr & 0x3FF) \ #define dma_read_palette_ram(type, transfer_size) \ - read_value = address##transfer_size(palette_ram, type##_ptr & 0x3FF) \ + read_value = readaddress##transfer_size(palette_ram, type##_ptr & 0x3FF) \ #define dma_read_ewram(type, transfer_size) \ - read_value = address##transfer_size(ewram, type##_ptr & 0x3FFFF) \ + read_value = readaddress##transfer_size(ewram, type##_ptr & 0x3FFFF) \ #define dma_read_gamepak(type, transfer_size) \ dma_gamepak_check_region(type); \ - read_value = address##transfer_size(type##_address_block, \ + read_value = readaddress##transfer_size(type##_address_block, \ type##_ptr & 0x7FFF) \ // DMAing from the BIOS is funny, just returns 0.. @@ -2575,17 +2574,20 @@ dma_region_type dma_region_map[16] = read_value = read_memory##transfer_size(type##_ptr) \ #define dma_write_iwram(type, transfer_size) \ - address##transfer_size(iwram + 0x8000, type##_ptr & 0x7FFF) = read_value; \ + address##transfer_size(iwram + 0x8000, type##_ptr & 0x7FFF) = \ + eswap##transfer_size(read_value); \ smc_trigger |= address##transfer_size(iwram, type##_ptr & 0x7FFF) \ #define dma_write_vram(type, transfer_size) \ - address##transfer_size(vram, type##_ptr & 0x1FFFF) = read_value \ + address##transfer_size(vram, type##_ptr & 0x1FFFF) = \ + eswap##transfer_size(read_value) \ #define dma_write_io(type, transfer_size) \ write_io_register##transfer_size(type##_ptr & 0x3FF, read_value) \ #define dma_write_oam_ram(type, transfer_size) \ - address##transfer_size(oam_ram, type##_ptr & 0x3FF) = read_value \ + address##transfer_size(oam_ram, type##_ptr & 0x3FF) = \ + eswap##transfer_size(read_value) \ #define dma_write_palette_ram(type, transfer_size) \ write_palette##transfer_size(type##_ptr & 0x3FF, read_value) \ @@ -2594,7 +2596,8 @@ dma_region_type dma_region_map[16] = write_memory##transfer_size(type##_ptr, read_value) \ #define dma_write_ewram(type, transfer_size) \ - address##transfer_size(ewram, type##_ptr & 0x3FFFF) = read_value; \ + address##transfer_size(ewram, type##_ptr & 0x3FFFF) = \ + eswap##transfer_size(read_value); \ smc_trigger |= address##transfer_size(ewram, \ (type##_ptr & 0x3FFFF) + 0x40000) \ @@ -3034,8 +3037,8 @@ cpu_alert_type dma_transfer(dma_transfer_type *dma) (dma->start_type == DMA_START_IMMEDIATELY)) { dma->start_type = DMA_INACTIVE; - address16(io_registers, (dma->dma_channel * 12) + 0xBA) &= - (~0x8000); + address16(io_registers, (dma->dma_channel * 12) + 0xBA) = + readaddress16(io_registers, (dma->dma_channel * 12) + 0xBA) & (~0x8000); } if(dma->irq) @@ -3208,13 +3211,13 @@ void init_memory(void) memset(ewram, 0, sizeof(ewram)); memset(vram, 0, sizeof(vram)); - io_registers[REG_DISPCNT] = 0x80; - io_registers[REG_P1] = 0x3FF; - io_registers[REG_BG2PA] = 0x100; - io_registers[REG_BG2PD] = 0x100; - io_registers[REG_BG3PA] = 0x100; - io_registers[REG_BG3PD] = 0x100; - io_registers[REG_RCNT] = 0x8000; + write_ioreg(REG_DISPCNT, 0x80); + write_ioreg(REG_P1, 0x3FF); + write_ioreg(REG_BG2PA, 0x100); + write_ioreg(REG_BG2PD, 0x100); + write_ioreg(REG_BG3PA, 0x100); + write_ioreg(REG_BG3PD, 0x100); + write_ioreg(REG_RCNT, 0x8000); backup_type = BACKUP_NONE; diff --git a/gba_memory.h b/gba_memory.h index 5319a30..0f9dc3a 100644 --- a/gba_memory.h +++ b/gba_memory.h @@ -122,6 +122,19 @@ typedef enum REG_BLDCNT = 0x28, REG_BLDALPHA = 0x29, REG_BLDY = 0x2A, + REG_SOUND1CNT_L = 0x30, + REG_SOUND1CNT_H = 0x31, + REG_SOUND3CNT_L = 0x38, + REG_SOUND3CNT_H = 0x39, + REG_SOUND3CNT_X = 0x3A, + REG_SOUND4CNT_H = 0x3E, + REG_SOUNDCNT_L = 0x40, + REG_SOUNDCNT_H = 0x41, + REG_SOUNDCNT_X = 0x42, + REG_DMA0SAD = 0x58, + REG_DMA0DAD = 0x5A, + REG_DMA0CNT_L = 0x5C, + REG_DMA0CNT_H = 0x5D, REG_TM0D = 0x80, REG_TM0CNT = 0x81, REG_TM1D = 0x82, diff --git a/input.c b/input.c index 9aef248..e030596 100644 --- a/input.c +++ b/input.c @@ -38,7 +38,7 @@ extern void set_fastforward_override(bool fastforward); static void trigger_key(u32 key) { - u32 p1_cnt = io_registers[REG_P1CNT]; + u32 p1_cnt = read_ioreg(REG_P1CNT); if((p1_cnt >> 14) & 0x01) { @@ -121,7 +121,7 @@ u32 update_input(void) trigger_key(new_key); old_key = new_key; - io_registers[REG_P1] = (~old_key) & 0x3FF; + write_ioreg(REG_P1, (~old_key) & 0x3FF); /* Handle fast forward button */ if (libretro_ff_enabled != libretro_ff_enabled_prev) diff --git a/main.c b/main.c index 260edd5..461dbcf 100644 --- a/main.c +++ b/main.c @@ -63,7 +63,7 @@ static void update_timers(irq_type *irq_raised) { timer[i].count -= execute_cycles; /* io_registers accessors range: REG_TM0D, REG_TM1D, REG_TM2D, REG_TM3D */ - io_registers[128 + (i * 2)] = -(timer[i].count > timer[i].prescale); + write_ioreg(REG_TM0D + (i * 2), -(timer[i].count > timer[i].prescale)); } if(timer[i].count > 0) @@ -76,7 +76,7 @@ static void update_timers(irq_type *irq_raised) if((i != 3) && (timer[i + 1].status == TIMER_CASCADE)) { timer[i + 1].count--; - io_registers[REG_TM0D + (i + 1) * 2] = -(timer[i + 1].count); + write_ioreg(REG_TM0D + (i + 1) * 2, -timer[i+1].count); } if(i < 2) @@ -146,8 +146,8 @@ u32 update_gba(void) if(video_count <= 0) { - u32 vcount = io_registers[REG_VCOUNT]; - u32 dispstat = io_registers[REG_DISPSTAT]; + u32 vcount = read_ioreg(REG_VCOUNT); + u32 dispstat = read_ioreg(REG_DISPSTAT); if((dispstat & 0x02) == 0) { @@ -162,7 +162,7 @@ u32 update_gba(void) oam_update_count++; if(no_alpha) - io_registers[REG_BLDCNT] = 0; + write_ioreg(REG_BLDCNT, 0); update_scanline(); // If in visible area also fire HDMA @@ -196,13 +196,13 @@ u32 update_gba(void) } affine_reference_x[0] = - (s32)(address32(io_registers, 0x28) << 4) >> 4; + (s32)(readaddress32(io_registers, 0x28) << 4) >> 4; affine_reference_y[0] = - (s32)(address32(io_registers, 0x2C) << 4) >> 4; + (s32)(readaddress32(io_registers, 0x2C) << 4) >> 4; affine_reference_x[1] = - (s32)(address32(io_registers, 0x38) << 4) >> 4; + (s32)(readaddress32(io_registers, 0x38) << 4) >> 4; affine_reference_y[1] = - (s32)(address32(io_registers, 0x3C) << 4) >> 4; + (s32)(readaddress32(io_registers, 0x3C) << 4) >> 4; for(i = 0; i < 4; i++) { @@ -251,9 +251,9 @@ u32 update_gba(void) else dispstat &= ~0x04; - io_registers[REG_VCOUNT] = vcount; + write_ioreg(REG_VCOUNT, vcount); } - io_registers[REG_DISPSTAT] = dispstat; + write_ioreg(REG_DISPSTAT, dispstat); } if(irq_raised) diff --git a/sound.c b/sound.c index 10a3261..bc88fff 100644 --- a/sound.c +++ b/sound.c @@ -437,7 +437,7 @@ void update_gbc_sound(u32 cpu_ticks) s32 volume_left, volume_right; u32 envelope_volume; s32 current_sample; - u32 sound_status = address16(io_registers, 0x84) & 0xFFF0; + u16 sound_status = read_ioreg(REG_SOUNDCNT_X) & 0xFFF0; s8 *sample_data; s8 *wave_bank; u8 *wave_ram = ((u8 *)io_registers) + 0x90; @@ -521,7 +521,7 @@ void update_gbc_sound(u32 cpu_ticks) } } - address16(io_registers, 0x84) = sound_status; + write_ioreg(REG_SOUNDCNT_X, sound_status); gbc_sound_last_cpu_ticks = cpu_ticks; gbc_sound_buffer_index = diff --git a/video.c b/video.c index 4221f25..274b39c 100644 --- a/video.c +++ b/video.c @@ -31,22 +31,8 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline u32 enable_flags, u32 dispcnt, u32 bldcnt, const bitmap_layer_render_struct *layer_renderers); -#define no_op \ - -// This old version is not necessary if the palette is either being converted -// transparently or the ABGR 1555 format is being used natively. The direct -// version (without conversion) is much faster. - -#define tile_lookup_palette_full(palette, source) \ - current_pixel = palette[source]; \ - convert_palette(current_pixel) \ - -#define tile_lookup_palette(palette, source) \ - current_pixel = palette[source]; \ - - #define tile_expand_base_normal(index) \ - tile_lookup_palette(palette, current_pixel); \ + current_pixel = palette[current_pixel]; \ dest_ptr[index] = current_pixel \ #define tile_expand_transparent_normal(index) \ @@ -68,14 +54,14 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define color_combine_mask_a(layer) \ - ((io_registers[REG_BLDCNT] >> layer) & 0x01) \ + ((read_ioreg(REG_BLDCNT) >> layer) & 0x01) \ // For color blending operations, will create a mask that has in bit // 10 if the layer is target B, and bit 9 if the layer is target A. #define color_combine_mask(layer) \ (color_combine_mask_a(layer) | \ - ((io_registers[REG_BLDCNT] >> (layer + 7)) & 0x02)) << 9 \ + ((read_ioreg(REG_BLDCNT) >> (layer + 7)) & 0x02)) << 9 \ // For alpha blending renderers, draw the palette index (9bpp) and // layer bits rather than the raw RGB. For the base this should write to @@ -183,7 +169,7 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline // Get the current tile from the map in 8bpp mode #define get_tile_8bpp() \ - current_tile = *map_ptr; \ + current_tile = eswap16(*map_ptr); \ tile_ptr = tile_base + ((current_tile & 0x3FF) * 64) \ @@ -243,16 +229,16 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define partial_tile_right_noflip_8bpp(combine_op, alpha_op) \ if(partial_tile_offset >= 4) \ { \ - current_pixels = *((u32 *)(tile_ptr + 4)) >> \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))) >> \ ((partial_tile_offset - 4) * 8); \ partial_tile_8bpp(combine_op, alpha_op); \ } \ else \ { \ partial_tile_run -= 4; \ - current_pixels = *((u32 *)tile_ptr) >> (partial_tile_offset * 8); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) >> (partial_tile_offset * 8);\ partial_tile_8bpp(combine_op, alpha_op); \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, noflip); \ advance_dest_ptr_##combine_op(4); \ } \ @@ -264,19 +250,19 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define partial_tile_mid_noflip_8bpp(combine_op, alpha_op) \ if(partial_tile_offset >= 4) \ { \ - current_pixels = *((u32 *)(tile_ptr + 4)) >> \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))) >> \ ((partial_tile_offset - 4) * 8); \ } \ else \ { \ - current_pixels = *((u32 *)tile_ptr) >> (partial_tile_offset * 8); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) >> (partial_tile_offset * 8);\ if((partial_tile_offset + partial_tile_run) > 4) \ { \ u32 old_run = partial_tile_run; \ partial_tile_run = 4 - partial_tile_offset; \ partial_tile_8bpp(combine_op, alpha_op); \ partial_tile_run = old_run - partial_tile_run; \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ } \ } \ partial_tile_8bpp(combine_op, alpha_op); \ @@ -288,23 +274,23 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define partial_tile_left_noflip_8bpp(combine_op, alpha_op) \ if(partial_tile_run >= 4) \ { \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, noflip); \ advance_dest_ptr_##combine_op(4); \ tile_ptr += 4; \ partial_tile_run -= 4; \ } \ \ - current_pixels = *((u32 *)(tile_ptr)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr))); \ partial_tile_8bpp(combine_op, alpha_op) \ // Draws a non-clipped (complete) 8bpp tile. #define tile_noflip_8bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, noflip); \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ tile_8bpp_draw_four_##combine_op(4, alpha_op, noflip) \ @@ -321,26 +307,28 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define partial_tile_right_flip_8bpp(combine_op, alpha_op) \ if(partial_tile_offset >= 4) \ { \ - current_pixels = *((u32 *)tile_ptr) << ((partial_tile_offset - 4) * 8); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) << \ + ((partial_tile_offset - 4) * 8); \ partial_tile_flip_8bpp(combine_op, alpha_op); \ } \ else \ { \ partial_tile_run -= 4; \ - current_pixels = *((u32 *)(tile_ptr + 4)) << \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))) << \ ((partial_tile_offset - 4) * 8); \ partial_tile_flip_8bpp(combine_op, alpha_op); \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, flip); \ advance_dest_ptr_##combine_op(4); \ } \ #define partial_tile_mid_flip_8bpp(combine_op, alpha_op) \ if(partial_tile_offset >= 4) \ - current_pixels = *((u32 *)tile_ptr) << ((partial_tile_offset - 4) * 8); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) << \ + ((partial_tile_offset - 4) * 8); \ else \ { \ - current_pixels = *((u32 *)(tile_ptr + 4)) << \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))) << \ ((partial_tile_offset - 4) * 8); \ \ if((partial_tile_offset + partial_tile_run) > 4) \ @@ -349,7 +337,7 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline partial_tile_run = 4 - partial_tile_offset; \ partial_tile_flip_8bpp(combine_op, alpha_op); \ partial_tile_run = old_run - partial_tile_run; \ - current_pixels = *((u32 *)(tile_ptr)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr))); \ } \ } \ partial_tile_flip_8bpp(combine_op, alpha_op); \ @@ -357,20 +345,20 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define partial_tile_left_flip_8bpp(combine_op, alpha_op) \ if(partial_tile_run >= 4) \ { \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, flip); \ advance_dest_ptr_##combine_op(4); \ tile_ptr -= 4; \ partial_tile_run -= 4; \ } \ \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ partial_tile_flip_8bpp(combine_op, alpha_op) \ #define tile_flip_8bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)(tile_ptr + 4)); \ + current_pixels = eswap32(*((u32 *)(tile_ptr + 4))); \ tile_8bpp_draw_four_##combine_op(0, alpha_op, flip); \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_8bpp_draw_four_##combine_op(4, alpha_op, flip) \ @@ -536,7 +524,7 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline // the pixel block. #define get_tile_4bpp() \ - current_tile = *map_ptr; \ + current_tile = eswap16(*map_ptr); \ current_palette = (current_tile >> 12) << 4; \ tile_ptr = tile_base + ((current_tile & 0x3FF) * 32); \ @@ -557,7 +545,7 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline // how many to draw. #define partial_tile_right_noflip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr) >> (partial_tile_offset * 4); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) >> (partial_tile_offset * 4); \ partial_tile_4bpp(combine_op, alpha_op) \ @@ -571,13 +559,13 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline // partial_tile_offset is how many to draw. #define partial_tile_left_noflip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ partial_tile_4bpp(combine_op, alpha_op) \ // Draws a complete 4bpp tile row (not clipped) #define tile_noflip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_4bpp_draw_eight_##combine_op(alpha_op, noflip) \ @@ -592,18 +580,18 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline } \ #define partial_tile_right_flip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr) << (partial_tile_offset * 4); \ + current_pixels = eswap32(*((u32 *)tile_ptr)) << (partial_tile_offset * 4); \ partial_tile_flip_4bpp(combine_op, alpha_op) \ #define partial_tile_mid_flip_4bpp(combine_op, alpha_op) \ partial_tile_right_flip_4bpp(combine_op, alpha_op) \ #define partial_tile_left_flip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ partial_tile_flip_4bpp(combine_op, alpha_op) \ #define tile_flip_4bpp(combine_op, alpha_op) \ - current_pixels = *((u32 *)tile_ptr); \ + current_pixels = eswap32(*((u32 *)tile_ptr)); \ tile_4bpp_draw_eight_##combine_op(alpha_op, flip) \ @@ -868,13 +856,13 @@ static void render_scanline_text_base_normal(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_base_normal(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1055,13 +1043,13 @@ static void render_scanline_text_transparent_normal(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_transparent_normal(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1241,13 +1229,13 @@ static void render_scanline_text_base_color16(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_base_color16(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1426,13 +1414,13 @@ static void render_scanline_text_transparent_color16(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_transparent_color16(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1611,13 +1599,13 @@ static void render_scanline_text_base_color32(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_base_color32(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1796,13 +1784,13 @@ static void render_scanline_text_transparent_color32(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_transparent_color32(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -1983,13 +1971,13 @@ static void render_scanline_text_base_alpha(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_base_alpha(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -2166,13 +2154,13 @@ static void render_scanline_text_transparent_alpha(u32 layer, u32 start, u32 end, void *scanline) { render_scanline_extra_variables_transparent_alpha(text); - u32 bg_control = io_registers[REG_BG0CNT + layer]; + u32 bg_control = read_ioreg(REG_BG0CNT + layer); u32 map_size = (bg_control >> 14) & 0x03; u32 map_width = map_widths[map_size]; u32 horizontal_offset = - (io_registers[REG_BG0HOFS + (layer * 2)] + start) % 512; - u32 vertical_offset = (io_registers[REG_VCOUNT] + - io_registers[REG_BG0VOFS + (layer * 2)]) % 512; + (read_ioreg(REG_BG0HOFS + (layer * 2)) + start) % 512; + u32 vertical_offset = (read_ioreg(REG_VCOUNT) + + read_ioreg(REG_BG0VOFS + (layer * 2))) % 512; u32 current_pixel; u32 current_pixels; u32 partial_tile_run = 0; @@ -2522,7 +2510,7 @@ void render_scanline_affine_##combine_op##_##alpha_op(u32 layer, \ u32 start, u32 end, void *scanline) \ { \ render_scanline_extra_variables_##combine_op##_##alpha_op(affine); \ - u32 bg_control = io_registers[REG_BG0CNT + layer]; \ + u32 bg_control = read_ioreg(REG_BG0CNT + layer); \ u32 current_pixel; \ s32 source_x, source_y; \ u32 pixel_x, pixel_y; \ @@ -2539,8 +2527,8 @@ void render_scanline_affine_##combine_op##_##alpha_op(u32 layer, \ render_scanline_dest_##alpha_op *dest_ptr = \ ((render_scanline_dest_##alpha_op *)scanline) + start; \ \ - dx = (s16)io_registers[REG_BG2PA + layer_offset]; \ - dy = (s16)io_registers[REG_BG2PC + layer_offset]; \ + dx = (s16)read_ioreg(REG_BG2PA + layer_offset); \ + dy = (s16)read_ioreg(REG_BG2PC + layer_offset); \ source_x = affine_reference_x[layer - 2] + (start * dx); \ source_y = affine_reference_y[layer - 2] + (start * dy); \ \ @@ -2609,7 +2597,7 @@ render_scanline_affine_builder(transparent, alpha); \ for(i = 0; (s32)i < (s32)end; i++) \ { \ - current_pixel = *src_ptr; \ + current_pixel = srcread_##type(*src_ptr); \ bitmap_render_pixel_##type(alpha_op); \ src_ptr++; \ dest_ptr++; \ @@ -2637,7 +2625,7 @@ render_scanline_affine_builder(transparent, alpha); if((u32)pixel_x >= (u32)width) \ break; \ \ - current_pixel = src_ptr[pixel_x]; \ + current_pixel = srcread_##type(src_ptr[pixel_x]); \ bitmap_render_pixel_##type(alpha_op); \ \ source_x += dx; \ @@ -2669,7 +2657,7 @@ render_scanline_affine_builder(transparent, alpha); if(((u32)pixel_x >= (u32)width) || ((u32)pixel_y >= (u32)height)) \ break; \ \ - current_pixel = src_ptr[pixel_x + (pixel_y * width)]; \ + current_pixel = srcread_##type(src_ptr[pixel_x + (pixel_y * width)]); \ bitmap_render_pixel_##type(alpha_op); \ \ source_x += dx; \ @@ -2683,16 +2671,19 @@ render_scanline_affine_builder(transparent, alpha); #define render_scanline_vram_setup_mode5() \ u16 *src_ptr = (u16 *)vram; \ - if(io_registers[REG_DISPCNT] & 0x10) \ + if(read_ioreg(REG_DISPCNT) & 0x10) \ src_ptr = (u16 *)(vram + 0xA000); \ #define render_scanline_vram_setup_mode4() \ u16 *palette = palette_ram_converted; \ u8 *src_ptr = vram; \ - if(io_registers[REG_DISPCNT] & 0x10) \ + if(read_ioreg(REG_DISPCNT) & 0x10) \ src_ptr = vram + 0xA000; \ +#define srcread_mode3(v) eswap16(v) +#define srcread_mode5(v) eswap16(v) +#define srcread_mode4(v) (v) // Build bitmap scanline rendering functions. @@ -2705,8 +2696,8 @@ static void render_scanline_bitmap_##type##_##alpha_op(u32 start, u32 end, \ s32 source_x, source_y; \ s32 pixel_x, pixel_y; \ \ - s32 dx = (s16)io_registers[REG_BG2PA]; \ - s32 dy = (s16)io_registers[REG_BG2PC]; \ + s32 dx = (s16)read_ioreg(REG_BG2PA); \ + s32 dy = (s16)read_ioreg(REG_BG2PC); \ \ u32 i; \ \ @@ -3058,11 +3049,11 @@ static const bitmap_layer_render_struct bitmap_mode_renderers[3] = #define obj_render_affine(combine_op, color_depth, alpha_op, map_space) \ { \ - s16 *params = (s16 *)oam_ram + (((obj_attribute_1 >> 9) & 0x1F) * 16); \ - s32 dx = params[3]; \ - s32 dmx = params[7]; \ - s32 dy = params[11]; \ - s32 dmy = params[15]; \ + u16 *params = (u16 *)oam_ram + (((obj_attribute_1 >> 9) & 0x1F) * 16); \ + s32 dx = (s16)eswap16(params[3]); \ + s32 dmx = (s16)eswap16(params[7]); \ + s32 dy = (s16)eswap16(params[11]); \ + s32 dmy = (s16)eswap16(params[15]); \ s32 source_x, source_y; \ s32 tile_x, tile_y; \ u32 tile_map_offset; \ @@ -3163,9 +3154,9 @@ static u32 obj_alpha_count[160]; u32 dest \ #define render_scanline_obj_extra_variables_copy(type) \ - u32 bldcnt = io_registers[REG_BLDCNT]; \ - u32 dispcnt = io_registers[REG_DISPCNT]; \ - u32 obj_enable = io_registers[REG_WINOUT] >> 8; \ + u32 bldcnt = read_ioreg(REG_BLDCNT); \ + u32 dispcnt = read_ioreg(REG_DISPCNT); \ + u32 obj_enable = read_ioreg(REG_WINOUT) >> 8; \ render_scanline_layer_functions_##type(); \ u32 copy_start, copy_end; \ u16 copy_buffer[240]; \ @@ -3278,7 +3269,7 @@ static void render_scanline_obj_##alpha_op##_##map_space(u32 priority, \ s32 obj_size; \ s32 obj_width, obj_height; \ u32 obj_attribute_0, obj_attribute_1, obj_attribute_2; \ - s32 vcount = io_registers[REG_VCOUNT]; \ + s32 vcount = read_ioreg(REG_VCOUNT); \ u32 tile_run; \ u32 current_pixels; \ u32 current_pixel; \ @@ -3296,9 +3287,9 @@ static void render_scanline_obj_##alpha_op##_##map_space(u32 priority, \ for(obj_num = 0; obj_num < obj_count; obj_num++) \ { \ oam_ptr = oam_ram + (obj_list[obj_num] * 4); \ - obj_attribute_0 = oam_ptr[0]; \ - obj_attribute_1 = oam_ptr[1]; \ - obj_attribute_2 = oam_ptr[2]; \ + obj_attribute_0 = eswap16(oam_ptr[0]); \ + obj_attribute_1 = eswap16(oam_ptr[1]); \ + obj_attribute_2 = eswap16(oam_ptr[2]); \ obj_size = ((obj_attribute_0 >> 12) & 0x0C) | (obj_attribute_1 >> 14); \ \ obj_x = (s32)(obj_attribute_1 << 23) >> 23; \ @@ -3355,8 +3346,8 @@ static void order_obj(u32 video_mode) for(obj_num = 127; obj_num >= 0; obj_num--, oam_ptr -= 4) { - obj_attribute_0 = oam_ptr[0]; - obj_attribute_2 = oam_ptr[2]; + obj_attribute_0 = eswap16(oam_ptr[0]); + obj_attribute_2 = eswap16(oam_ptr[2]); obj_size = obj_attribute_0 & 0xC000; obj_priority = (obj_attribute_2 >> 10) & 0x03; obj_mode = (obj_attribute_0 >> 10) & 0x03; @@ -3369,7 +3360,7 @@ static void order_obj(u32 video_mode) if(obj_y > 160) obj_y -= 256; - obj_attribute_1 = oam_ptr[1]; + obj_attribute_1 = eswap16(oam_ptr[1]); obj_size = ((obj_size >> 12) & 0x0C) | (obj_attribute_1 >> 14); obj_height = obj_height_table[obj_size]; obj_width = obj_width_table[obj_size]; @@ -3436,14 +3427,14 @@ static void order_layers(u32 layer_flags) for(layer_number = 3; layer_number >= 0; layer_number--) { if(((layer_flags >> layer_number) & 1) && - ((io_registers[REG_BG0CNT + layer_number] & 0x03) == priority)) + ((read_ioreg(REG_BG0CNT + layer_number) & 0x03) == priority)) { layer_order[layer_count] = layer_number; layer_count++; } } - if((obj_priority_count[priority][io_registers[REG_VCOUNT]] > 0) + if((obj_priority_count[priority][read_ioreg(REG_VCOUNT)] > 0) && (layer_flags & 0x10)) { layer_order[layer_count] = priority | 0x04; @@ -3614,7 +3605,7 @@ void expand_blend(u32 *screen_src_ptr, u16 *screen_dest_ptr, { u32 pixel_pair; u32 pixel_top, pixel_bottom; - u32 bldalpha = io_registers[REG_BLDALPHA]; + u32 bldalpha = read_ioreg(REG_BLDALPHA); u32 blend_a = bldalpha & 0x1F; u32 blend_b = (bldalpha >> 8) & 0x1F; u32 i; @@ -3645,7 +3636,7 @@ static void expand_darken(u16 *screen_src_ptr, u16 *screen_dest_ptr, u32 start, u32 end) { u32 pixel_top; - s32 blend = 16 - (io_registers[REG_BLDY] & 0x1F); + s32 blend = 16 - (read_ioreg(REG_BLDY) & 0x1F); u32 i; if(blend < 0) @@ -3661,7 +3652,7 @@ static void expand_brighten(u16 *screen_src_ptr, u16 *screen_dest_ptr, u32 start, u32 end) { u32 pixel_top; - u32 blend = io_registers[REG_BLDY] & 0x1F; + u32 blend = read_ioreg(REG_BLDY) & 0x1F; u32 upper; u32 i; @@ -3682,10 +3673,10 @@ static void expand_brighten(u16 *screen_src_ptr, u16 *screen_dest_ptr, static void expand_darken_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_ptr, u32 start, u32 end) { - s32 blend = 16 - (io_registers[REG_BLDY] & 0x1F); + s32 blend = 16 - (read_ioreg(REG_BLDY) & 0x1F); u32 pixel_pair; u32 pixel_top, pixel_bottom; - u32 bldalpha = io_registers[REG_BLDALPHA]; + u32 bldalpha = read_ioreg(REG_BLDALPHA); u32 blend_a = bldalpha & 0x1F; u32 blend_b = (bldalpha >> 8) & 0x1F; u32 i; @@ -3706,10 +3697,10 @@ static void expand_darken_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_pt static void expand_brighten_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_ptr, u32 start, u32 end) { - s32 blend = io_registers[REG_BLDY] & 0x1F; + s32 blend = read_ioreg(REG_BLDY) & 0x1F; u32 pixel_pair; u32 pixel_top, pixel_bottom; - u32 bldalpha = io_registers[REG_BLDALPHA]; + u32 bldalpha = read_ioreg(REG_BLDALPHA); u32 blend_a = bldalpha & 0x1F; u32 blend_b = (bldalpha >> 8) & 0x1F; u32 upper; @@ -3784,20 +3775,20 @@ static void expand_brighten_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_ } \ #define render_condition_alpha \ - (((io_registers[REG_BLDALPHA] & 0x1F1F) != 0x001F) && \ - ((io_registers[REG_BLDCNT] & 0x3F) != 0) && \ - ((io_registers[REG_BLDCNT] & 0x3F00) != 0)) \ + (((read_ioreg(REG_BLDALPHA) & 0x1F1F) != 0x001F) && \ + ((read_ioreg(REG_BLDCNT) & 0x3F) != 0) && \ + ((read_ioreg(REG_BLDCNT) & 0x3F00) != 0)) \ #define render_condition_fade \ - (((io_registers[REG_BLDY] & 0x1F) != 0) && \ - ((io_registers[REG_BLDCNT] & 0x3F) != 0)) \ + (((read_ioreg(REG_BLDY) & 0x1F) != 0) && \ + ((read_ioreg(REG_BLDCNT) & 0x3F) != 0)) \ #define render_layers_color_effect(renderer, layer_condition, \ alpha_condition, fade_condition, _start, _end) \ { \ if(layer_condition) \ { \ - if(obj_alpha_count[io_registers[REG_VCOUNT]] > 0) \ + if(obj_alpha_count[read_ioreg(REG_VCOUNT)] > 0) \ { \ /* Render based on special effects mode. */ \ u32 screen_buffer[240]; \ @@ -3902,7 +3893,7 @@ static void expand_brighten_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_ { \ if(color_combine_mask_a(5)) \ { \ - u32 blend = io_registers[REG_BLDY] & 0x1F; \ + u32 blend = read_ioreg(REG_BLDY) & 0x1F; \ u32 upper; \ \ if(blend > 16) \ @@ -3921,7 +3912,7 @@ static void expand_brighten_partial_alpha(u32 *screen_src_ptr, u16 *screen_dest_ { \ if(color_combine_mask_a(5)) \ { \ - s32 blend = 16 - (io_registers[REG_BLDY] & 0x1F); \ + s32 blend = 16 - (read_ioreg(REG_BLDY) & 0x1F); \ \ if(blend < 0) \ blend = 0; \ @@ -3942,7 +3933,7 @@ static void render_scanline_tile(u16 *scanline, u32 dispcnt) { u32 current_layer; u32 layer_order_pos; - u32 bldcnt = io_registers[REG_BLDCNT]; + u32 bldcnt = read_ioreg(REG_BLDCNT); render_scanline_layer_functions_tile(); render_layers_color_effect(render_layers, layer_count, @@ -4116,9 +4107,9 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define window_x_coords(window_number) \ window_##window_number##_x1 = \ - io_registers[REG_WIN##window_number##H] >> 8; \ + read_ioreg(REG_WIN##window_number##H) >> 8; \ window_##window_number##_x2 = \ - io_registers[REG_WIN##window_number##H] & 0xFF; \ + read_ioreg(REG_WIN##window_number##H) & 0xFF; \ window_##window_number##_enable = \ (winin >> (window_number * 8)) & 0x3F; \ \ @@ -4133,9 +4124,9 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline u32 window_##window_number##_y1, window_##window_number##_y2; \ u32 window_##window_number##_enable = 0; \ window_##window_number##_y1 = \ - io_registers[REG_WIN##window_number##V] >> 8; \ + read_ioreg(REG_WIN##window_number##V) >> 8; \ window_##window_number##_y2 = \ - io_registers[REG_WIN##window_number##V] & 0xFF; \ + read_ioreg(REG_WIN##window_number##V) & 0xFF; \ \ if(window_##window_number##_y1 > window_##window_number##_y2) \ { \ @@ -4306,7 +4297,7 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define render_window_single(type, window_number) \ - u32 winin = io_registers[REG_WININ]; \ + u32 winin = read_ioreg(REG_WININ); \ window_coords(window_number); \ if(window_##window_number##_x1 > window_##window_number##_x2) \ { \ @@ -4344,9 +4335,9 @@ static void render_scanline_conditional_bitmap(u32 start, u32 end, u16 *scanline #define render_scanline_window_builder(type) \ static void render_scanline_window_##type(u16 *scanline, u32 dispcnt) \ { \ - u32 vcount = io_registers[REG_VCOUNT]; \ - u32 winout = io_registers[REG_WINOUT]; \ - u32 bldcnt = io_registers[REG_BLDCNT]; \ + u32 vcount = read_ioreg(REG_VCOUNT); \ + u32 winout = read_ioreg(REG_WINOUT); \ + u32 bldcnt = read_ioreg(REG_BLDCNT); \ u32 window_out_enable = winout & 0x3F; \ \ render_scanline_layer_functions_##type(); \ @@ -4370,7 +4361,7 @@ static void render_scanline_window_##type(u16 *scanline, u32 dispcnt) \ /* Windows 1 and 2 */ \ case 0x03: \ { \ - u32 winin = io_registers[REG_WININ]; \ + u32 winin = read_ioreg(REG_WININ); \ window_coords(0); \ window_coords(1); \ render_window_multi(type, 0, 1); \ @@ -4387,7 +4378,7 @@ static void render_scanline_window_##type(u16 *scanline, u32 dispcnt) \ /* Window 0 and OBJ window */ \ case 0x05: \ { \ - u32 winin = io_registers[REG_WININ]; \ + u32 winin = read_ioreg(REG_WININ); \ window_coords(0); \ render_window_multi(type, 0, obj); \ break; \ @@ -4396,7 +4387,7 @@ static void render_scanline_window_##type(u16 *scanline, u32 dispcnt) \ /* Window 1 and OBJ window */ \ case 0x06: \ { \ - u32 winin = io_registers[REG_WININ]; \ + u32 winin = read_ioreg(REG_WININ); \ window_coords(1); \ render_window_multi(type, 1, obj); \ break; \ @@ -4405,7 +4396,7 @@ static void render_scanline_window_##type(u16 *scanline, u32 dispcnt) \ /* Window 0, 1, and OBJ window */ \ case 0x07: \ { \ - u32 winin = io_registers[REG_WININ]; \ + u32 winin = read_ioreg(REG_WININ); \ window_coords(0); \ window_coords(1); \ render_window_multi(type, 0, 1_obj); \ @@ -4422,8 +4413,8 @@ static const u32 active_layers[6] = { 0x1F, 0x17, 0x1C, 0x14, 0x14, 0x14 }; void update_scanline(void) { u32 pitch = get_screen_pitch(); - u32 dispcnt = io_registers[REG_DISPCNT]; - u32 vcount = io_registers[REG_VCOUNT]; + u32 dispcnt = read_ioreg(REG_DISPCNT); + u32 vcount = read_ioreg(REG_VCOUNT); u16 *screen_offset = get_screen_pixels() + (vcount * pitch); u32 video_mode = dispcnt & 0x07; @@ -4467,10 +4458,10 @@ void update_scanline(void) } } - affine_reference_x[0] += (s16)io_registers[REG_BG2PB]; - affine_reference_y[0] += (s16)io_registers[REG_BG2PD]; - affine_reference_x[1] += (s16)io_registers[REG_BG3PB]; - affine_reference_y[1] += (s16)io_registers[REG_BG3PD]; + affine_reference_x[0] += (s16)read_ioreg(REG_BG2PB); + affine_reference_y[0] += (s16)read_ioreg(REG_BG2PD); + affine_reference_x[1] += (s16)read_ioreg(REG_BG3PB); + affine_reference_y[1] += (s16)read_ioreg(REG_BG3PD); } #define video_savestate_builder(type) \ -- cgit v1.2.3