Added libmario, renamed roundf to lroundf to conform to C standard naming, wrapped inline asm functions in TARGET_N64 ifdef

Makefile

@@ -630,8 +630,7 @@ $(BUILD_DIR)/src/usb/usb.o: CFLAGS += -Wno-unused-variable -Wno-sign-compare -Wn
 $(BUILD_DIR)/src/usb/debug.o: OPT_FLAGS := -O0
 $(BUILD_DIR)/src/usb/debug.o: CFLAGS += -Wno-unused-parameter -Wno-maybe-uninitialized
 # File specific opt flags
-$(BUILD_DIR)/src/audio/heap.o:          OPT_FLAGS := -Os -fno-jump-tables
-$(BUILD_DIR)/src/audio/synthesis.o:     OPT_FLAGS := -Os -fno-jump-tables
+$(BUILD_DIR)/src/audio/*.o:                   OPT_FLAGS := -Os -fno-jump-tables
 
 $(BUILD_DIR)/src/engine/surface_collision.o:  OPT_FLAGS := $(COLLISION_OPT_FLAGS)
 $(BUILD_DIR)/src/engine/math_util.o:          OPT_FLAGS := $(MATH_UTIL_OPT_FLAGS)

README.md

@@ -34,7 +34,7 @@ This is a fork of the ultrasm64 repo by CrashOveride which includes the followin
 Thanks to Frame#5375 and AloXado320 for also helping with silhouette stuff
 
 **Lighting Engine by Wiseguy**
-- Lighting Engine is available on a separate branch ([base/lighting-engine](https://github.com/Reonu/HackerSM64/tree/base/lighting-engine)). Instructions on how to use it are in the readme of that branch.
+- Lighting Engine is available on a separate branch ([base/lighting-engine-wip](https://github.com/Reonu/HackerSM64/tree/base/lighting-engine-wip)). Instructions on how to use it are in the readme of that branch.
 - Alternatively, the main repo has `Puppylights` available, which is a more lightweight, but limited lighting library intended to be used to modify existing light properties. You can look at `puppylights.c` to find out how to use it.
 
 **Puppycam**

VERSION.txt

@@ -1 +1 @@
-v2.0.3
+v2.0.1

include/config/config_debug.h

@@ -10,10 +10,6 @@
 // Disables all debug options (except PUPPYPRINT).
 #define DISABLE_ALL
 
-// Enables a comprehensive standalone profiler. Automatically enabled by PUPPYPRINT_DEBUG.
-// If not using PUPPYPRINT_DEBUG, press L to toggle the profiler.
-// #define USE_PROFILER
-
 // TEST LEVEL
 // Uncomment this define and set a test level in order to boot straight into said level.
 // This allows you to quickly test the level you're working on.

include/config/config_objects.h

@@ -40,8 +40,9 @@
 
 // -- WATER RING --
 
-// Fix the water rings in DDD by checking for interaction rather than normals.
-#define FIX_WATER_RINGS
+// Fix DDD water rings by checking for interaction rather than normals.
+//! NOT IMPLEMENTED
+// #define FIX_WATER_RINGS
 
 // -- POWER STAR --
 

include/config/config_safeguards.h

@@ -33,21 +33,15 @@
  * config_debug.h
  */
 
-#ifdef DISABLE_ALL
-    #undef DEBUG_ALL
-    #undef TEST_LEVEL
-    #undef DEBUG_LEVEL_SELECT
-    #undef ENABLE_DEBUG_FREE_MOVE
-    #undef VANILLA_DEBUG
-    #undef VANILLA_STYLE_CUSTOM_DEBUG
-    #undef PUPPYPRINT_DEBUG
-    #undef PUPPYPRINT_DEBUG_CYCLES
-    #undef VISUAL_DEBUG
+#ifdef PUPPYPRINT_DEBUG
+    #undef PUPPYPRINT
+    #define PUPPYPRINT
+#endif // PUPPYPRINT_DEBUG
+
+#ifdef COMPLETE_SAVE_FILE
     #undef UNLOCK_ALL
-    #undef COMPLETE_SAVE_FILE
-    #undef DEBUG_FORCE_CRASH_ON_BOOT
-    #undef USE_PROFILER
-#endif // DISABLE_ALL
+    #define UNLOCK_ALL
+#endif // COMPLETE_SAVE_FILE
 
 #ifdef DEBUG_ALL
     #undef DEBUG_LEVEL_SELECT
@@ -72,17 +66,20 @@
     #define COMPLETE_SAVE_FILE
 #endif // DEBUG_ALL
 
-#ifdef PUPPYPRINT_DEBUG
-    #undef PUPPYPRINT
-    #define PUPPYPRINT
-    #undef USE_PROFILER
-    #define USE_PROFILER
-#endif // PUPPYPRINT_DEBUG
-
-#ifdef COMPLETE_SAVE_FILE
+#ifdef DISABLE_ALL
+    #undef DEBUG_ALL
+    #undef TEST_LEVEL
+    #undef DEBUG_LEVEL_SELECT
+    #undef ENABLE_DEBUG_FREE_MOVE
+    #undef VANILLA_DEBUG
+    #undef VANILLA_STYLE_CUSTOM_DEBUG
+    #undef PUPPYPRINT_DEBUG
+    #undef PUPPYPRINT_DEBUG_CYCLES
+    #undef VISUAL_DEBUG
     #undef UNLOCK_ALL
-    #define UNLOCK_ALL
-#endif // COMPLETE_SAVE_FILE
+    #undef COMPLETE_SAVE_FILE
+    #undef DEBUG_FORCE_CRASH_ON_BOOT
+#endif // DISABLE_ALL
 
 
 /*****************

include/object_fields.h

@@ -973,7 +973,11 @@
 #define /*0x104*/ oSnufitBodyScalePeriod OBJECT_FIELD_S32(0x1F)
 #define /*0x108*/ oSnufitBodyBaseScale   OBJECT_FIELD_S32(0x20)
 #define /*0x10C*/ oSnufitBullets         OBJECT_FIELD_S32(0x21)
-#define /*0x1AC*/ oSnufitBodyScale       OBJECT_FIELD_S16(0x49, 0)
+#define /*0x1AC*/ oSnufitOffsetVec       OBJECT_FIELD_S16(0x49, 0) // start pos of s16 vec
+#define /*0x1AC*/ oSnufitXOffset         OBJECT_FIELD_S16(0x49, 0)
+#define /*0x1AE*/ oSnufitYOffset         OBJECT_FIELD_S16(0x49, 1)
+#define /*0x1B0*/ oSnufitZOffset         OBJECT_FIELD_S16(0x4A, 0)
+#define /*0x1B2*/ oSnufitBodyScale       OBJECT_FIELD_S16(0x4A, 1)
 
 /* Spindel */
 #define /*0x0F4*/ oSpindelMoveTimer     OBJECT_FIELD_S32(0x1B)
@@ -1043,6 +1047,7 @@
 #define /*0x0F4*/ oTiltingPyramidNormalX         OBJECT_FIELD_F32(O_TILTING_PYRAMID_NORMAL_X_INDEX)
 #define /*0x0F8*/ oTiltingPyramidNormalY         OBJECT_FIELD_F32(O_TILTING_PYRAMID_NORMAL_Y_INDEX)
 #define /*0x0FC*/ oTiltingPyramidNormalZ         OBJECT_FIELD_F32(O_TILTING_PYRAMID_NORMAL_Z_INDEX)
+#define /*0x10C*/ oTiltingPyramidMarioOnPlatform OBJECT_FIELD_S32(0x21)
 
 /* Toad Message */
 #define /*0x108*/ oToadMessageDialogId       OBJECT_FIELD_U32(0x20)

levels/ending/leveldata.c

@@ -9,7 +9,6 @@
 #ifdef VERSION_EU
 #include "levels/ending/cake_eu.inc.c"
 
-#ifndef EU_CUSTOM_CAKE_FIX
 // 0x07023000 - 0x07023FFF
 ALIGNED8 static const Texture cake_end_texture_eu_35_thank_you[] = {
 #include "levels/ending/eu_023000.rgba16.inc.c"
@@ -39,7 +38,6 @@ ALIGNED8 static const Texture cake_end_texture_eu_39_fin[] = {
 ALIGNED8 static const Texture cake_end_texture_eu_40_ende[] = {
 #include "levels/ending/eu_028000.rgba16.inc.c"
 };
-#endif // !EU_CUSTOM_CAKE_FIX
 
 // 0x07029000 - 0x070296D8
 const Gfx dl_cake_end_screen[] = {

levels/intro/geo.c

@@ -72,11 +72,11 @@ const GeoLayout intro_geo_mario_head_regular[] = {
          GEO_CLOSE_NODE(),
 #endif
       GEO_CLOSE_NODE(),
-#if defined(ENABLE_RUMBLE)
-      GEO_ZBUFFER(0),
-      GEO_OPEN_NODE(),
-         GEO_ASM(INTRO_CONTEXT_NORMAL, geo_intro_rumble_pak_graphic),
-      GEO_CLOSE_NODE(),
+#if defined(VERSION_SH)
+   GEO_ZBUFFER(0),
+   GEO_OPEN_NODE(),
+      GEO_ASM(INTRO_CONTEXT_NORMAL, geo_intro_rumble_pak_graphic),
+   GEO_CLOSE_NODE(),
 #endif
    GEO_CLOSE_NODE(),
    GEO_END(),
@@ -105,7 +105,7 @@ const GeoLayout intro_geo_mario_head_dizzy[] = {
          GEO_CLOSE_NODE(),
 #endif
       GEO_CLOSE_NODE(),
-#if defined(ENABLE_RUMBLE)
+#if defined(VERSION_SH)
       GEO_ZBUFFER(0),
       GEO_OPEN_NODE(),
          GEO_ASM(INTRO_CONTEXT_GAME_OVER, geo_intro_rumble_pak_graphic),

libmario/Makefile

@@ -0,0 +1,11 @@
+DECOMP_INPUT_FILES := ../src/game/mario_actions_airborne.c ../src/game/mario_actions_automatic.c ../src/game/mario_actions_cutscene.c \
+    ../src/game/mario_actions_moving.c ../src/game/mario_actions_object.c ../src/game/mario_actions_stationary.c ../src/game/mario_actions_submerged.c \
+	../src/game/mario_misc.c ../src/game/mario_step.c ../src/game/mario.c ../src/engine/math_util.c ../src/game/object_helpers.c
+
+all: libmario.dll libmario.so
+
+libmario.dll: $(wildcard mariolib/*.c) $(DECOMP_INPUT_FILES)
+	x86_64-w64-mingw32-gcc $^ -o $@ -shared -DNON_MATCHING -DAVOID_UB -D_LANGUAGE_C -ggdb -I.. -I../include -I../include/n64 -I../src/engine -I../src/game -I../src -Ism64 -lm -Wl,--subsystem,windows
+
+libmario.so: $(wildcard mariolib/*.c) $(DECOMP_INPUT_FILES)
+	gcc $^ -o $@ -shared -DNON_MATCHING -DAVOID_UB -D_LANGUAGE_C -ggdb -I.. -I../include -I../include/n64 -I../src/engine -I../src/game -I../src -Ism64 -lm -fPIC

libmario/test.py

@@ -0,0 +1,205 @@
+#!/usr/bin/env python3
+import inputs
+import time
+import threading
+import sys
+import fpstimer
+import os
+from ctypes import *
+
+pluginPath = os.path.dirname(os.path.realpath(__file__))
+
+library = 'libmario.so'
+if sys.platform == 'win32':
+    library = 'libmario.dll'
+if sys.platform == 'darwin':
+    library = 'libmario.dynlib'
+
+libmario = CDLL(os.path.join(pluginPath, library))
+
+    # /*0x00*/ s16 type;
+    # /*0x02*/ s16 force;
+    # /*0x04*/ s8 flags;
+    # /*0x05*/ s8 room;
+    # /*0x06*/ s16 lowerY;
+    # /*0x08*/ s16 upperY;
+    # /*0x0A*/ Vec3s vertex1;
+    # /*0x10*/ Vec3s vertex2;
+    # /*0x16*/ Vec3s vertex3;
+    # /*0x1C*/ struct {
+    #     f32 x;
+    #     f32 y;
+    #     f32 z;
+    # } normal;
+    # /*0x28*/ f32 originOffset;
+    # /*0x2C*/ struct Object *object;
+
+Vec3f = (c_float * 3)
+Vec3s = (c_int16 * 3)
+class Surface(Structure):
+    _fields_ = [
+        ('type', c_int16),
+        ('force', c_int16),
+        ('flags', c_int8),
+        ('room', c_int8),
+        ('lowerY', c_int16),
+        ('upperY', c_int16),
+        ('vertex1', Vec3s),
+        ('vertex2', Vec3s),
+        ('vertex3', Vec3s),
+        ('normal', Vec3f),
+        ('origin_offset', c_float),
+        ('object', c_void_p)
+    ]
+
+FindFloorHandlerType = CFUNCTYPE(c_float, c_float, c_float, c_float, POINTER(Surface), POINTER(c_int32))
+FindCeilHandlerType = CFUNCTYPE(c_float, c_float, c_float, c_float, POINTER(Surface), POINTER(c_int32))
+FindWallsHandlerType = CFUNCTYPE(c_int32, c_float, c_float, c_float, c_float, c_float, POINTER(Surface), POINTER(c_float))
+FindWaterLevelHandlerType = CFUNCTYPE(c_float, c_float, c_float)
+
+libmario.init.restype = None
+libmario.init.artypes = [FindFloorHandlerType, FindCeilHandlerType, FindWallsHandlerType, FindWaterLevelHandlerType]
+
+libmario.step.restype = None
+libmario.step.artypes = [c_int32, c_float, c_float]
+
+libmario.getMarioPosition.restype = None
+libmario.getMarioPosition.artypes = [Vec3f]
+
+libmario.getMarioVelocity.restype = None
+libmario.getMarioVelocity.artypes = [Vec3f]
+
+libmario.getMarioAnimFrame.restype = c_int32
+libmario.getMarioAnimFrame.artypes = []
+
+libmario.getMarioAnimIndex.restype = c_int32
+libmario.getMarioAnimIndex.artypes = []
+
+CONT_A      = 0x8000
+CONT_B      = 0x4000
+CONT_G      = 0x2000
+CONT_START  = 0x1000
+CONT_UP     = 0x0800
+CONT_DOWN   = 0x0400
+CONT_LEFT   = 0x0200
+CONT_RIGHT  = 0x0100
+CONT_L      = 0x0020
+CONT_R      = 0x0010
+CONT_E      = 0x0008
+CONT_D      = 0x0004
+CONT_C      = 0x0002
+CONT_F      = 0x0001
+
+A_BUTTON     = CONT_A
+B_BUTTON     = CONT_B
+L_TRIG       = CONT_L
+R_TRIG       = CONT_R
+Z_TRIG       = CONT_G
+START_BUTTON = CONT_START
+U_JPAD       = CONT_UP
+L_JPAD       = CONT_LEFT
+R_JPAD       = CONT_RIGHT
+D_JPAD       = CONT_DOWN
+U_CBUTTONS   = CONT_E
+L_CBUTTONS   = CONT_C
+R_CBUTTONS   = CONT_F
+D_CBUTTONS   = CONT_D
+
+events = []
+_t = None
+_handler = None
+
+def find_floor(x, y, z, surface_out, found_out):
+    found_out[0] = 1
+    surface_out[0].vertex1[0] = -100
+    surface_out[0].vertex1[1] = 0
+    surface_out[0].vertex1[2] = -100
+
+    surface_out[0].vertex2[0] = 100
+    surface_out[0].vertex2[1] = 0
+    surface_out[0].vertex2[2] = -100
+
+    surface_out[0].vertex3[0] = 100
+    surface_out[0].vertex3[1] = 0
+    surface_out[0].vertex3[2] = 100
+
+    surface_out[0].normal[0] = 0.0
+    surface_out[0].normal[1] = 1.0
+    surface_out[0].normal[2] = 0.0
+
+    surface_out[0].origin_offset = 0.0
+
+    return -1000.0
+
+def find_water_level(x, z):
+    return -100.0
+
+# Needs to be global to avoid getting garbage collected during execution
+find_floor_handler = FindFloorHandlerType(find_floor)
+find_water_level_handler = FindWaterLevelHandlerType(find_water_level)
+
+def worker():
+    global events
+    while True:
+        events.append(inputs.get_gamepad())
+        
+
+def main():
+    global _t
+    global _handler
+    global events
+    if not _t :
+        _t = threading.Thread(target=worker)
+        _t.daemon = True
+        _t.start()
+    stick_x = 0.0
+    stick_y = 0.0
+    buttons = 0
+    libmario.init(find_floor_handler, None, None, find_water_level_handler)
+    timer = fpstimer.FPSTimer(30)
+    try:
+        while True:
+            while len(events) > 0 :
+                for event in events[0]:
+                    if event.code == "ABS_X":
+                        stick_x = float(event.state) / 32768.0
+                    elif event.code == "ABS_Y":
+                        stick_y = float(event.state) / 32768.0
+                    elif event.code == "ABS_RX":
+                        gpd_input = "Right Stick X"
+                    elif event.code == "ABS_RY":
+                        gpd_input = "Right Stick Y"
+                    elif event.code == "BTN_SOUTH":
+                        if event.state == 1:
+                            buttons |= A_BUTTON
+                        else:
+                            buttons &= ~A_BUTTON
+                    elif event.code == "BTN_WEST":
+                        if event.state == 1:
+                            buttons |= B_BUTTON
+                        else:
+                            buttons &= ~B_BUTTON
+                    elif event.code == "ABS_Z":
+                        if event.state == 255:
+                            buttons |= Z_TRIG
+                        else:
+                            buttons &= ~Z_TRIG
+                    elif event.code != "SYN_REPORT":
+                        print(event.code + ':' + str(event.state))
+                events.pop(0)
+            
+            libmario.step(buttons, c_float(stick_x), c_float(stick_y))
+            pos = Vec3f()
+            vel = Vec3f()
+            libmario.getMarioPosition(pos)
+            libmario.getMarioVelocity(vel)
+            
+            print('Position: %8.2f %8.2f %8.2f  Velocity: %8.2f %8.2f %8.2f Buttons: 0x%08X Anim: 0x%02X AnimFrame: %d' % (pos[0], pos[1], pos[2], vel[0], vel[1], vel[2], buttons, libmario.getMarioAnimIndex(), libmario.getMarioAnimFrame()))
+
+            timer.sleep()
+    except KeyboardInterrupt:
+        print("Ctrl+C pressed...")
+        sys.exit(0)
+
+if __name__ == '__main__':
+    main()

src/audio/effects.c

@@ -343,12 +343,16 @@ f32 adsr_update(struct AdsrState *adsr) {
 s32 adsr_update(struct AdsrState *adsr) {
 #endif
     u8 action = adsr->action;
+#if defined(VERSION_EU) || defined(VERSION_SH)
     u8 state = adsr->state;
     switch (state) {
+#else
+    switch (adsr->state) {
+#endif
         case ADSR_STATE_DISABLED:
             return 0;
 
-        case ADSR_STATE_INITIAL:
+        case ADSR_STATE_INITIAL: {
 #if defined(VERSION_JP) || defined(VERSION_US)
             adsr->current = adsr->initial;
             adsr->target = adsr->initial;
@@ -357,7 +361,8 @@ s32 adsr_update(struct AdsrState *adsr) {
                 adsr->state = ADSR_STATE_HANG;
                 break;
             }
-            // fallthrough
+        }
+        // fallthrough
 
         case ADSR_STATE_START_LOOP:
             adsr->envIndex = 0;
@@ -365,12 +370,11 @@ s32 adsr_update(struct AdsrState *adsr) {
             adsr->currentHiRes = (adsr->current << 0x10);
 #endif
             adsr->state = ADSR_STATE_LOOP;
+
 #ifdef VERSION_SH
-            // fallthrough
             restart:
 #endif
             // fallthrough
-
         case ADSR_STATE_LOOP:
             adsr->delay = BSWAP16(adsr->envelope[adsr->envIndex].delay);
             switch (adsr->delay) {
@@ -395,23 +399,25 @@ s32 adsr_update(struct AdsrState *adsr) {
 #if defined(VERSION_EU) || defined(VERSION_SH)
                     if (adsr->delay >= 4) {
                         adsr->delay = adsr->delay * gAudioBufferParameters.updatesPerFrame
- #ifdef VERSION_SH
+#ifdef VERSION_SH
                         / gAudioBufferParameters.presetUnk4
- #endif
+#endif
                         / 4;
                     }
- #ifdef VERSION_SH
+#if defined(VERSION_SH)
                     if (adsr->delay == 0) {
                         adsr->delay = 1;
                     }
- #endif
                     adsr->target = (f32) BSWAP16(adsr->envelope[adsr->envIndex].arg) / 32767.0f;
+#elif defined(VERSION_EU)
+                    adsr->target = (f32) BSWAP16(adsr->envelope[adsr->envIndex].arg) / 32767.0;
+#endif
                     adsr->target = adsr->target * adsr->target;
                     adsr->velocity = (adsr->target - adsr->current) / adsr->delay;
-#else // !(VERSION_EU || VERSION_SH)
+#else
                     adsr->target = BSWAP16(adsr->envelope[adsr->envIndex].arg);
                     adsr->velocity = ((adsr->target - adsr->current) << 0x10) / adsr->delay;
-#endif // !(VERSION_EU || VERSION_SH)
+#endif
                     adsr->state = ADSR_STATE_FADE;
                     adsr->envIndex++;
                     break;

src/audio/external.c

@@ -1262,7 +1262,7 @@ static void update_game_sound(void) {
 #endif
                                 } else {
 #if defined(VERSION_EU) || defined(VERSION_SH)
-                                    func_802ad728(0x04020000 | ((channelIndex & 0xff) << 8),
+                                    func_802ad728((x04020000 | ((channelIndex & 0xff) << 8),
                                         get_sound_freq_scale(bank, soundIndex) + ((f32) sSoundMovingSpeed[bank] / 400.0f));
 #else
                                     value = get_sound_freq_scale(bank, soundIndex);

src/audio/heap.c

@@ -169,7 +169,7 @@ void reset_bank_and_seq_load_status(void) {
 #ifdef VERSION_SH
     bzero(&gBankLoadStatus, sizeof(gBankLoadStatus));
     bzero(&gUnkLoadStatus,  sizeof(gUnkLoadStatus));
-    bzero(&gSeqLoadStatus,  sizeof(gSeqLoadStatus));
+    bzero(&gSeqLoadStatus,  sizeof(gBankLoadStatus));
 #else
     bzero(&gBankLoadStatus, sizeof(gBankLoadStatus)); // Setting this array to zero is equivilent to SOUND_LOAD_STATUS_NOT_LOADED
     bzero(&gSeqLoadStatus,  sizeof(gSeqLoadStatus));  // Same dealio
@@ -274,7 +274,7 @@ void sound_alloc_pool_init(struct SoundAllocPool *pool, void *memAddr, u32 size)
 #ifdef VERSION_SH
     pool->size = size - ((uintptr_t) memAddr & 0xf);
 #else
-    pool->size = ALIGN16(size);
+    pool->size = size;
 #endif
     pool->numAllocatedEntries = 0;
 }
@@ -380,7 +380,7 @@ void *alloc_bank_or_seq(struct SoundMultiPool *arg0, s32 arg1, s32 size, s32 arg
     // arg3 = 0, 1 or 2?
 
 #ifdef VERSION_SH
-    struct SoundMultiPool *arg0 = NULL;
+    struct SoundMultiPool *arg0;
 #define isSound poolIdx
 #endif
     struct TemporaryPool *tp;
@@ -408,8 +408,6 @@ void *alloc_bank_or_seq(struct SoundMultiPool *arg0, s32 arg1, s32 size, s32 arg
     u32 leftAvail, rightAvail;
 #endif
 
-size = ALIGN16(size);
-
 #ifdef VERSION_SH
     switch (poolIdx) {
         case 0:
@@ -658,7 +656,7 @@ size = ALIGN16(size);
 #if defined(VERSION_SH)
                 tp->entries[1].ptr = (u8 *) ((uintptr_t) (pool->start + pool->size - size) & ~0x0f);
 #else
-                tp->entries[1].ptr = pool->start + pool->size - size;
+                tp->entries[1].ptr = pool->start + pool->size - size - 0x10;
 #endif
                 tp->entries[1].id = id;
                 tp->entries[1].size = size;
@@ -760,7 +758,7 @@ void *get_bank_or_seq(s32 poolIdx, s32 arg1, s32 id) {
 }
 void *get_bank_or_seq_inner(s32 poolIdx, s32 arg1, s32 bankId) {
     u32 i;
-    struct SoundMultiPool* loadedPool = NULL;
+    struct SoundMultiPool* loadedPool;
     struct TemporaryPool* temporary;
     struct PersistentPool* persistent;
 
@@ -933,7 +931,7 @@ void decrease_reverb_gain(void) {
 
 #if defined(VERSION_EU) || defined(VERSION_SH)
 s32 audio_shut_down_and_reset_step(void) {
-    s32 i;
+    s32 i, j;
 
     switch (gAudioResetStatus) {
         case 5:
@@ -1009,7 +1007,7 @@ void init_reverb_eu(void) {
     gNumSynthesisReverbs = preset->numReverbs;
     for (j = 0; j < gNumSynthesisReverbs; j++) {
         reverb = &gSynthesisReverbs[j];
-        reverbSettings = &sReverbSettings[MIN((u32)(gAudioResetPresetIdToLoad + j), (sizeof(sReverbSettings) / sizeof(struct ReverbSettingsEU)) - 1)];
+        reverbSettings = &sReverbSettings[MIN((gAudioResetPresetIdToLoad + j), (sizeof(sReverbSettings) / sizeof(struct ReverbSettingsEU)) - 1)];
         reverb->windowSize = (reverbSettings->windowSize * 0x40);
         reverb->downsampleRate = reverbSettings->downsampleRate;
         reverb->reverbGain = reverbSettings->gain;

src/audio/heap.h

@@ -38,14 +38,8 @@ struct SeqOrBankEntry {
 struct PersistentPool {
     /*0x00*/ u32 numEntries;
     /*0x04*/ struct SoundAllocPool pool;
-#ifdef EXPAND_AUDIO_HEAP // TODO: Make this a configurable define rather than using static values
-    /*0x14*/ struct SeqOrBankEntry entries[64];
-    // size = 0x314
-#else
     /*0x14*/ struct SeqOrBankEntry entries[32];
-    // size = 0x194
-#endif
-};
+}; // size = 0x194
 
 struct TemporaryPool {
     /*EU,   SH*/
@@ -70,14 +64,9 @@ struct SoundMultiPool {
     /*     */ u32 pad2[4];
 }; // size = 0x1D0
 
-#ifdef VERSION_SH
 struct Unk1Pool {
     struct SoundAllocPool pool;
-#ifdef EXPAND_AUDIO_HEAP
-    struct SeqOrBankEntry entries[64];
-#else
     struct SeqOrBankEntry entries[32];
-#endif
 };
 
 struct UnkEntry {
@@ -96,7 +85,6 @@ struct UnkPool {
     /*0x510*/ s32 numEntries;
     /*0x514*/ u32 unk514;
 };
-#endif
 
 extern u8 gAudioHeap[];
 extern s16 gVolume;

src/audio/load.c

@@ -24,8 +24,6 @@ struct SharedDma {
 // EU only
 void port_eu_init(void);
 
-ALIGNED16 u32 dmaTempBuffer[4];
-
 struct Note *gNotes;
 
 struct SequencePlayer gSequencePlayers[SEQUENCE_PLAYERS];
@@ -139,8 +137,14 @@ u8 audioString49[] = "BANK LOAD MISS! FOR %d\n";
  * Performs an asynchronus (normal priority) DMA copy
  */
 void audio_dma_copy_async(uintptr_t devAddr, void *vAddr, size_t nbytes, OSMesgQueue *queue, OSIoMesg *mesg) {
+#if PUPPYPRINT_DEBUG
+    OSTime first = osGetTime();
+#endif
     osInvalDCache(vAddr, nbytes);
     osPiStartDma(mesg, OS_MESG_PRI_NORMAL, OS_READ, devAddr, vAddr, nbytes, queue);
+#if PUPPYPRINT_DEBUG
+    dmaAudioTime[perfIteration] += (osGetTime() - first);
+#endif
 }
 
 /**
@@ -148,12 +152,18 @@ void audio_dma_copy_async(uintptr_t devAddr, void *vAddr, size_t nbytes, OSMesgQ
  * to 0x1000 bytes transfer at once.
  */
 void audio_dma_partial_copy_async(uintptr_t *devAddr, u8 **vAddr, ssize_t *remaining, OSMesgQueue *queue, OSIoMesg *mesg) {
+#if PUPPYPRINT_DEBUG
+    OSTime first = osGetTime();
+#endif
     ssize_t transfer = MIN(*remaining, 0x1000);
     *remaining -= transfer;
     osInvalDCache(*vAddr, transfer);
     osPiStartDma(mesg, OS_MESG_PRI_NORMAL, OS_READ, *devAddr, *vAddr, transfer, queue);
     *devAddr += transfer;
     *vAddr += transfer;
+#if PUPPYPRINT_DEBUG
+    dmaAudioTime[perfIteration] += (osGetTime() - first);
+#endif
 }
 
 void decrease_sample_dma_ttls() {
@@ -198,10 +208,17 @@ void *dma_sample_data(uintptr_t devAddr, u32 size, s32 arg2, u8 *dmaIndexRef) {
     u32 i;
     u32 dmaIndex;
     ssize_t bufferPos;
+#if PUPPYPRINT_DEBUG
+    OSTime first = osGetTime();
+#endif
 
     if (arg2 != 0 || *dmaIndexRef >= sSampleDmaListSize1) {
         for (i = sSampleDmaListSize1; i < gSampleDmaNumListItems; i++) {
+#if defined(VERSION_EU)
             dma = &sSampleDmas[i];
+#else
+            dma = sSampleDmas + i;
+#endif
             bufferPos = devAddr - dma->source;
             if (0 <= bufferPos && (size_t) bufferPos <= dma->bufSize - size) {
                 // We already have a DMA request for this memory range.
@@ -218,7 +235,14 @@ void *dma_sample_data(uintptr_t devAddr, u32 size, s32 arg2, u8 *dmaIndexRef) {
                 }
                 dma->ttl = 60;
                 *dmaIndexRef = (u8) i;
+#if PUPPYPRINT_DEBUG
+                dmaAudioTime[perfIteration] += (osGetTime() - first);
+#endif
+#if defined(VERSION_EU)
+                return &dma->buffer[(devAddr - dma->source)];
+#else
                 return (devAddr - dma->source) + dma->buffer;
+#endif
             }
         }
 
@@ -231,7 +255,12 @@ void *dma_sample_data(uintptr_t devAddr, u32 size, s32 arg2, u8 *dmaIndexRef) {
             hasDma = TRUE;
         }
     } else {
+#if defined(VERSION_EU)
+        dma = sSampleDmas;
+        dma += *dmaIndexRef;
+#else
         dma = sSampleDmas + *dmaIndexRef;
+#endif
         bufferPos = devAddr - dma->source;
         if (0 <= bufferPos && (size_t) bufferPos <= dma->bufSize - size) {
             // We already have DMA for this memory range.
@@ -247,7 +276,17 @@ void *dma_sample_data(uintptr_t devAddr, u32 size, s32 arg2, u8 *dmaIndexRef) {
                 sSampleDmaReuseQueueTail1++;
             }
             dma->ttl = 2;
+#if defined(VERSION_EU)
+    #if PUPPYPRINT_DEBUG
+            dmaAudioTime[perfIteration] += (osGetTime() - first);
+    #endif
             return dma->buffer + (devAddr - dma->source);
+#else
+    #if PUPPYPRINT_DEBUG
+            dmaAudioTime[perfIteration] += (osGetTime() - first);
+    #endif
+            return (devAddr - dma->source) + dma->buffer;
+#endif
         }
     }
 
@@ -264,13 +303,27 @@ void *dma_sample_data(uintptr_t devAddr, u32 size, s32 arg2, u8 *dmaIndexRef) {
     dma->ttl = 2;
     dma->source = dmaDevAddr;
     dma->sizeUnused = transfer;
-#ifdef VERSION_US // TODO: Is there a reason this only exists in US?
+#ifdef VERSION_US
     osInvalDCache(dma->buffer, transfer);
 #endif
+#if defined(VERSION_EU)
     osPiStartDma(&gCurrAudioFrameDmaIoMesgBufs[gCurrAudioFrameDmaCount++], OS_MESG_PRI_NORMAL,
                      OS_READ, dmaDevAddr, dma->buffer, transfer, &gCurrAudioFrameDmaQueue);
     *dmaIndexRef = dmaIndex;
+#if PUPPYPRINT_DEBUG
+    dmaAudioTime[perfIteration] += (osGetTime() - first);
+#endif
     return (devAddr - dmaDevAddr) + dma->buffer;
+#else
+    gCurrAudioFrameDmaCount++;
+    osPiStartDma(&gCurrAudioFrameDmaIoMesgBufs[gCurrAudioFrameDmaCount - 1], OS_MESG_PRI_NORMAL,
+                 OS_READ, dmaDevAddr, dma->buffer, transfer, &gCurrAudioFrameDmaQueue);
+    *dmaIndexRef = dmaIndex;
+#if PUPPYPRINT_DEBUG
+    dmaAudioTime[perfIteration] += (osGetTime() - first);
+#endif
+    return dma->buffer + (devAddr - dmaDevAddr);
+#endif
 }
 
 
@@ -503,6 +556,8 @@ void patch_audio_bank(struct AudioBank *mem, u8 *offset, u32 numInstruments, u32
 }
 
 struct AudioBank *bank_load_immediate(s32 bankId, s32 arg1) {
+    u32 buf[4];
+
     // (This is broken if the length is 1 (mod 16), but that never happens --
     // it's always divisible by 4.)
     s32 alloc = gAlCtlHeader->seqArray[bankId].len + 0xf;
@@ -514,9 +569,9 @@ struct AudioBank *bank_load_immediate(s32 bankId, s32 arg1) {
         return NULL;
     }
 
-    audio_dma_copy_immediate((uintptr_t) ctlData, dmaTempBuffer, 0x10);
-    u32 numInstruments = dmaTempBuffer[0];
-    u32 numDrums = dmaTempBuffer[1];
+    audio_dma_copy_immediate((uintptr_t) ctlData, buf, 0x10);
+    u32 numInstruments = buf[0];
+    u32 numDrums = buf[1];
     audio_dma_copy_immediate((uintptr_t)(ctlData + 0x10), ret, alloc);
     patch_audio_bank(ret, gAlTbl->seqArray[bankId].offset, numInstruments, numDrums);
     gCtlEntries[bankId].numInstruments = (u8) numInstruments;
@@ -528,6 +583,8 @@ struct AudioBank *bank_load_immediate(s32 bankId, s32 arg1) {
 }
 
 struct AudioBank *bank_load_async(s32 bankId, s32 arg1, struct SequencePlayer *seqPlayer) {
+    u32 buf[4];
+
     size_t alloc = gAlCtlHeader->seqArray[bankId].len + 0xf;
     alloc = ALIGN16(alloc);
     alloc -= 0x10;
@@ -537,9 +594,9 @@ struct AudioBank *bank_load_async(s32 bankId, s32 arg1, struct SequencePlayer *s
         return NULL;
     }
 
-    audio_dma_copy_immediate((uintptr_t) ctlData, dmaTempBuffer, 0x10);
-    u32 numInstruments = dmaTempBuffer[0];
-    u32 numDrums = dmaTempBuffer[1];
+    audio_dma_copy_immediate((uintptr_t) ctlData, buf, 0x10);
+    u32 numInstruments = buf[0];
+    u32 numDrums = buf[1];
     seqPlayer->loadingBankId = (u8) bankId;
 #if defined(VERSION_EU)
     gCtlEntries[bankId].numInstruments = numInstruments;
@@ -803,8 +860,11 @@ void load_sequence_internal(u32 player, u32 seqId, s32 loadAsync) {
 
 // (void) must be omitted from parameters to fix stack with -framepointer
 void audio_init() {
+#if defined(VERSION_JP) || defined(VERSION_US) || defined(VERSION_EU)
+    u8 buf[0x10];
+#endif
     s32 i, /*j,*/ k;
-    u32 size;
+    UNUSED u32 size;
     void *data;
 
     gAudioLoadLock = AUDIO_LOCK_UNINITIALIZED;
@@ -879,6 +939,10 @@ void audio_init() {
     bzero(&gAiBuffers, sizeof(gAiBuffers));
     for (i = 0; i < NUMAIBUFFERS; i++) {
         gAiBuffers[i] = soundAlloc(&gAudioInitPool, AIBUFFER_LEN);
+
+        /*for (j = 0; j < (s32) (AIBUFFER_LEN / sizeof(s16)); j++) {
+            gAiBuffers[i][j] = 0;
+        }*/
     }
 
 #if defined(VERSION_EU)
@@ -895,7 +959,7 @@ void audio_init() {
     eu_stubbed_printf_0("Main Heap Initialize.\n");
 
     // Load headers for sounds and sequences
-    gSeqFileHeader = (ALSeqFile *) dmaTempBuffer;
+    gSeqFileHeader = (ALSeqFile *) buf;
     data = gMusicData;
     audio_dma_copy_immediate((uintptr_t) data, gSeqFileHeader, 0x10);
     gSequenceCount = gSeqFileHeader->seqCount;
@@ -906,7 +970,7 @@ void audio_init() {
     alSeqFileNew(gSeqFileHeader, data);
 
     // Load header for CTL (instrument metadata)
-    gAlCtlHeader = (ALSeqFile *) dmaTempBuffer;
+    gAlCtlHeader = (ALSeqFile *) buf;
     data = gSoundDataADSR;
     audio_dma_copy_immediate((uintptr_t) data, gAlCtlHeader, 0x10);
     size = gAlCtlHeader->seqCount * sizeof(ALSeqData) + 4;
@@ -917,7 +981,7 @@ void audio_init() {
     alSeqFileNew(gAlCtlHeader, data);
 
     // Load header for TBL (raw sound data)
-    gAlTbl = (ALSeqFile *) dmaTempBuffer;
+    gAlTbl = (ALSeqFile *) buf;
     audio_dma_copy_immediate((uintptr_t) data, gAlTbl, 0x10);
     size = gAlTbl->seqCount * sizeof(ALSeqData) + 4;
     size = ALIGN16(size);