bunch of math functions, permuter settings

Makefile

@@ -68,7 +68,7 @@ ENDLINE := \n'
 ### Compiler Options ###
 
 ASFLAGS      := -G 0 -I include -mips3 -mabi=32
-CFLAGS       := -O1 -G0 -mips3 -mgp32 -mfp32
+CFLAGS       := -O1 -G0 -mips3 -mgp32 -mfp32 -Wa,--vr4300mul-off
 CPPFLAGS     := -I include -I $(BUILD_DIR)/include -I src -DF3DEX_GBI_2
 LDFLAGS      := -T undefined_syms.txt -T undefined_funcs.txt -T undefined_funcs_auto.txt -T undefined_syms_auto.txt -T $(LD_SCRIPT) -Map $(LD_MAP) --no-check-sections
 CFLAGS_CHECK := -fsyntax-only -fsigned-char -nostdinc -fno-builtin -D CC_CHECK\
@@ -86,9 +86,7 @@ DEPENDS := $(OBJECTS:=.d)
 
 ### Targets ###
 
-#build/src/libultra/os/%.o: CFLAGS := -O2 $(CFLAGSCOMMON)
-#build/src/libultra/libc/%.o: CFLAGS := -O2 $(CFLAGSCOMMON)
-#build/src/lib/%.o: CFLAGS := -O2 $(CFLAGSCOMMON)
+build/src/libhmath/%.o: CFLAGS := -O2 -G0 -mips3 -mgp32 -mfp32 -ffast-math
 
 all: $(ROM)
 
@@ -112,8 +110,6 @@ split:
 
 test: $(ROM)
 	$(V)$(EMULATOR) $<
-	
-export VR4300MUL := OFF
 # Compile .c files with kmc gcc (use strip to fix objects so that they can be linked with modern gnu ld) 
 $(BUILD_DIR)/src/%.c.o: src/%.c
 	@$(PRINT)$(GREEN)Compiling C file: $(ENDGREEN)$(BLUE)$<$(ENDBLUE)$(ENDLINE)

include/common_structs.h

@@ -3,18 +3,13 @@
 
 #include "common.h"
 #include "process.h"
+#include "math.h"
 
 #define MAX_PLAYERS 4
 
 #define    OS_K0_TO_PHYSICAL(x)    (u32)(((char *)(x)-0x80000000))
 #define    OS_PHYSICAL_TO_K0(x)    (void *)(((u32)(x)+0x80000000))
 
-typedef struct {
-    f32 x;
-    f32 y;
-    f32 z;
-} HuVec3F;
-
 typedef struct objectt {
     /*0x00*/ struct objectt *prev;
     /*0x04*/ struct objectt *next;

include/math.h

@@ -1,8 +1,22 @@
 #ifndef _MATH_H_
 #define _MATH_H_
 
+#include "PR/ultratypes.h"
+
 #define M_PI 3.14159265358979323846
 
+typedef struct {
+    f32 x;
+    f32 y;
+    f32 z;
+} HuVec3F;
+
+typedef struct {
+    s32 x;
+    s32 y;
+    s32 z;
+} HuVec3I;
+
 // TODO: Where are these in libultra, if anywhere?
 float sinf(float);
 //double sin(double);
@@ -10,5 +24,11 @@ float cosf(float);
 //double cos(double);
 
 float sqrtf(float);
+float HuSqrtf(float x);
+
+void HuVecCopyXYZ(HuVec3F * out, f32 x, f32 y, f32 z);
+void HuVecCopy3F(HuVec3F * out, HuVec3F * a);
+f32 HuVecGetLengthSqr3F(HuVec3F * vec);
+f32 HuVecGetLength3F(HuVec3F * vec);
 
 #endif

marioparty3.yaml

@@ -72,6 +72,7 @@ segments:
     - [0x47D60, asm]
     - [0x4B120, asm]
     - [0x4BF40, asm]
+    - [0x4E010, asm, profiler]
     - [0x4E5A0, c, rom]
     - [0x4E800, asm]
     - [0x4F720, c, "process"]
@@ -310,10 +311,12 @@ segments:
     - [0x88440, asm]
     - [0x88460, asm]
     - [0x88710, asm]
-    - [0x88830, asm]
-    - [0x88CE0, asm]
-    - [0x88CF0, asm]
-    - [0x88D50, asm]
+    - [0x88830, asm] # last libultra scope
+    - [0x88CE0, c, libhmath/sqrtf]
+    - [0x88CF0, c, libhmath/vecmag]
+    - [0x88D50, c, libhmath/vecnorm]
+    - [0x88DE0, c, libhmath/vecdot]
+    - [0x88E10, asm]
     - [0x890B0, asm]
     - [0x89120, asm]
     - [0x89510, asm]
@@ -325,15 +328,20 @@ segments:
     - [0x8A4A0, asm]
     - [0x8A4D0, asm]
     - [0x8A580, asm]
-    - [0x8A610, asm]
-    - [0x8A640, asm]
-    - [0x8A670, asm]
-    - [0x8A6B0, asm]
-    - [0x8A6F0, asm]
-    - [0x8A770, asm]
-    - [0x8A830, asm]
-    - [0x8A860, asm]
-    - [0x8A890, asm]
+    - [0x8A610, c, libhmath/veccopy]
+    - [0x8A640, c, libhmath/vecneg]
+    - [0x8A670, c, libhmath/vecadd]
+    - [0x8A6B0, c, libhmath/vecsub]
+    - [0x8A6F0, c, libhmath/vecmuls]
+    - [0x8A720, c, libhmath/vecadds]
+    - [0x8A770, c, libhmath/veclinc]
+    - [0x8A830, c, libhmath/vecmags3f]
+    - [0x8A860, c, libhmath/vecmag3f]
+    - [0x8A890, c, libhmath/vecnorm3f]
+    - [0x8A920, c, libhmath/vecdist]
+    - [0x8A9B0, c, libhmath/vecdot3f]
+    - [0x8A9E0, c, libhmath/veccross]
+    - [0x8AA30, c, libhmath/veceq]
     - [0x8AA90, asm]
     - [0x8AC30, asm]
     - [0x8AC70, asm]

permuter_settings.toml

@@ -0,0 +1,13 @@
+compiler_command = "tools/gcc_2.7.2/gcc -w -Iinclude -Isrc  -D_LANGUAGE_C -D_FINALROM -DVERSION=us -DF3DEX_GBI_2 -D_MIPS_SZLONG=32 -nostdinc -c -G0 -O3 -x c -B tools/gcc_2.7.2/  -Wmissing-braces -Wimplicit -Wredundant-decls -Wstrict-prototypes -DPERMUTER -fforce-addr"
+
+assembler_command = "mips-linux-gnu-as -EB -march=vr4300 -mtune=vr4300 -Iinclude"
+
+compiler_type = "gcc"
+
+[preserve_macros]
+"gs?[DS]P.*" = "void"
+OVERRIDE_FLAG_CHECK = "int"
+OS_K0_TO_PHYSICAL = "int"
+"G_.*" = "int"
+"TEXEL.*" = "int"
+PRIMITIVE = "int"

src/libhmath/sqrtf.c

@@ -0,0 +1,5 @@
+#include "common.h"
+
+f32 HuSqrtf(f32 x) {
+    return sqrtf(x);
+}

src/libhmath/vecadd.c

@@ -0,0 +1,14 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecAdd(HuVec3F * out, HuVec3F * a, HuVec3F * b)
+{
+    float x = a->x + b->x;
+    float y = a->y + b->y;
+    float z = a->z + b->z;
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}
+

src/libhmath/vecadds.c

@@ -0,0 +1,14 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecAddScaled(HuVec3F * out, HuVec3F * a, f32 scalar, HuVec3F * b)
+{
+    float x = a->x + (scalar * b->x);
+    float y = a->y + (scalar * b->y);
+    float z = a->z + (scalar * b->z);
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}
+

src/libhmath/veccopy.c

@@ -0,0 +1,15 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecCopyXYZ(HuVec3F * out, f32 x, f32 y, f32 z)
+{
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}
+
+void HuVecCopy3F(HuVec3F * out, HuVec3F * a)
+{
+    *out = *a;
+}
+

src/libhmath/veccross.c

@@ -0,0 +1,15 @@
+#include "math.h"
+
+void HuVecCross(HuVec3F * out, HuVec3F * a, HuVec3F * b)
+{
+    f32 vector1Y = a->y;
+    f32 vector2Z = b->z;
+    f32 vector1Z = a->z;
+    f32 vector2Y = b->y;
+    f32 vector2X = b->x;
+    f32 vector1X = a->x;
+
+    out->x = (vector1Y * vector2Z) - (vector1Z * vector2Y);
+    out->y = (vector1Z * vector2X) - (vector1X * vector2Z);
+    out->z = (vector1X * vector2Y) - (vector1Y * vector2X);
+}

src/libhmath/vecdist.c

@@ -0,0 +1,37 @@
+#include "common.h"
+#include "math.h"
+
+// this matches, but the entire file needs to have inlined assembly for the mulmul fix to work
+#ifdef NONMATCHING
+f32 _HuVecDistanceSqr(HuVec3F * a, HuVec3F * b)
+{
+    f32 x;
+    f32 y;
+    f32 z;
+
+    x = a->x - b->x;
+    y = a->y - b->y;
+    z = a->z - b->z;
+
+    return (x * x) + (y * y) + (z * z);
+}
+#else
+INCLUDE_ASM(f32, "libhmath/vecdist", HuVecDistanceSqr, HuVec3F * a, HuVec3F * b);
+#endif
+
+// https://decomp.me/scratch/02P6S
+// // the original output has the function prologue near the end rather than the start
+#ifdef NONMATCHING
+f32 _HuVecDistance(HuVec3F * arg0, HuVec3F * arg1) {
+    f32 temp_f0;
+    f32 temp_f12;
+    f32 temp_f2;
+
+    temp_f0 = arg0->x - arg1->x;
+    temp_f2 = arg0->y - arg1->y;
+    temp_f12 = arg0->z - arg1->z;
+    return HuSqrtf((temp_f0 * temp_f0) + (temp_f2 * temp_f2) + (temp_f12 * temp_f12));
+}
+#else
+INCLUDE_ASM(f32, "libhmath/vecdist", HuVecDistance, HuVec3F * a, HuVec3F * b);
+#endif

src/libhmath/vecdot.c

@@ -0,0 +1,7 @@
+#include "common.h"
+
+f32 HuVecDotProduct(f32 aX, f32 aY, f32 aZ, f32 bX, f32 bY, f32 bZ)
+{
+    return (aX * bX) + (aY * bY) + (aZ * bZ);
+}
+

src/libhmath/vecdot3f.c

@@ -0,0 +1,6 @@
+#include "math.h"
+
+f32 HuVecDot3F(HuVec3F * a, HuVec3F * b)
+{
+    return (a->x * b->x) + (a->y * b->y) + (a->z * b->z);
+}

src/libhmath/veceq.c

@@ -0,0 +1,9 @@
+#include "common.h"
+#include "math.h"
+
+s32 HuVecIsEqual(HuVec3F * a, HuVec3F * b) {
+    if ((a->x == b->x) && (a->y == b->y) && (a->z == b->z)) {
+        return TRUE;
+    } 
+    return FALSE;
+}

src/libhmath/veclinc.c

@@ -0,0 +1,33 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecLinearComb(HuVec3F * out, f32 scale, HuVec3F * a, f32 arg3, HuVec3F * b)
+{
+    float x = (scale * a->x) + (arg3 * b->x);
+    float y = (scale * a->y) + (arg3 * b->y);
+    float z = (scale * a->z) + (arg3 * b->z);
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}
+
+void HuVecLinearComb2D(HuVec3F * out, HuVec3F * scale, HuVec3F * a, HuVec3F * b)
+{
+    f32 scaleX;
+    f32 scaleY;
+    f32 x;
+    f32 y;
+    f32 z;
+
+    scaleX = scale->x;
+    scaleY = scale->y;
+
+    x = (scaleX * a->x) + (scaleY * b->x);
+    y = (scaleX * a->y) + (scaleY * b->y);
+    z = (scaleX * a->z) + (scaleY * b->z);
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}

src/libhmath/vecmag.c

@@ -0,0 +1,14 @@
+#include <PR/ultratypes.h>
+#include "math.h"
+
+f32 HuSqrtf(f32);
+f32 HuVecGetLengthSqr(f32 arg0, f32 arg1, f32 arg2);
+f32 HuVecGetLength(f32 arg0, f32 arg1, f32 arg2);
+
+f32 HuVecGetLengthSqr(f32 arg0, f32 arg1, f32 arg2) {
+    return (arg0 * arg0) + (arg1 * arg1) + (arg2 * arg2);
+}
+
+f32 mulHuVecGetLength(f32 arg0, f32 arg1, f32 arg2) {
+    return HuSqrtf(HuVecGetLengthSqr(arg0, arg1, arg2));
+}

src/libhmath/vecmag3f.c

@@ -0,0 +1,7 @@
+#include "common.h"
+#include "math.h"
+
+float HuVecGetLength3F(HuVec3F * vec)
+{
+    return HuSqrtf(HuVecGetLengthSqr3F(vec));
+}

src/libhmath/vecmags3f.c

@@ -0,0 +1,16 @@
+#include "common.h"
+#include "math.h"
+
+f32 HuVecGetLengthSqr3F(HuVec3F * vec)
+{
+    f32 x;
+    f32 y;
+    f32 z;
+
+    x = vec->x;
+    y = vec->y;
+    z = vec->z;
+
+    return (x * x) + (y * y) + (z * z);
+}
+

src/libhmath/vecmuls.c

@@ -0,0 +1,13 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecMulScalar(HuVec3F * out, f32 scalar, HuVec3F * a)
+{
+    float x = scalar * a->x;
+    float y = scalar * a->y;
+    float z = scalar * a->z;
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}

src/libhmath/vecneg.c

@@ -0,0 +1,14 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecNeg(HuVec3F * out, HuVec3F * a) {
+
+    float x = -a->x;
+    float y = -a->y;
+    float z = -a->z;
+
+    out->x = x;
+    out->y = y;
+    out->z = z;
+}
+

src/libhmath/vecnorm.c

@@ -0,0 +1,19 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecNormalize(f32 * x, f32 * y, f32 * z)
+{
+    f32 oldX;
+    f32 oldY;
+    f32 oldZ;
+    f32 invMag;
+
+    oldX = *x;
+    oldY = *y;
+    oldZ = *z;
+    invMag = 1.0f / HuSqrtf((oldX * oldX) + (oldY * oldY) + (oldZ * oldZ));
+
+    *x *= invMag;
+    *y *= invMag;
+    *z *= invMag;
+}

src/libhmath/vecnorm3f.c

@@ -0,0 +1,25 @@
+#include "common.h"
+#include "math.h"
+
+// https://decomp.me/scratch/6UMIQ
+// the original output has the function prologue near the end rather than the start
+#ifdef NONMATCHING
+void _HuVecNormalize3F(HuVec3F * vec)
+{
+    f32 x;
+    f32 y;
+    f32 z;
+    f32 invMag;
+
+    x = vec->x;
+    y = vec->y;
+    z = vec->z;
+    invMag = 1.0f / HuSqrtf((x * x) + (y * y) + (z * z));
+
+    vec->x = invMag * x;
+    vec->y = invMag * y;
+    vec->z = invMag * z;
+}
+#else
+INCLUDE_ASM(void, "libhmath/vecnorm3f", HuVecNormalize3F, HuVec3F * vec);
+#endif

src/libhmath/vecsub.c

@@ -0,0 +1,14 @@
+#include "common.h"
+#include "math.h"
+
+void HuVecSubtract(HuVec3F * dest, HuVec3F * a, HuVec3F * b) {
+
+
+    float x = a->x - b->x;
+    float y = a->y - b->y;
+    float z = a->z - b->z;
+
+    dest->x = x;
+    dest->y = y;
+    dest->z = z;
+}

src/overlays/board_gate_guy/3D8AC0.c

@@ -1,4 +1,5 @@
 #include "GateGuy.h"
+#include "math.h"
 
 void func_801059D0_3D8AC0(void) {
     D_800CDD58 = 1;
@@ -142,10 +143,10 @@ void* func_80106740_3D9830(s32 arg0) {
     return temp_v0;
 }
 
-void func_801067D4_3D98C4(object* arg0, s32 arg1) {
+void func_801067D4_3D98C4(object* arg0, HuVec3F * arg1) {
     func_800D8E88();
     func_8001C814_1D414(*arg0->unk3C->unk40, 3, 0);
-    func_80089A20_8A620(&arg0->coords, arg1);
+    HuVecCopy3F(&arg0->coords, arg1);
 }
 
 INCLUDE_ASM(s32, "overlays/board_gate_guy/3D8AC0", func_80106828_3D9918);

src/overlays/gate_guy_intro1/3D8930.c

@@ -12,4 +12,4 @@ extern board_overlay_entrypoint D_80105AE0_3D8A70[];
 void func_801059A0_3D8930(void) {
     //func_800359E0_365E0 is ExecBoardScene from mp1
     func_800359E0_365E0(D_80105AE0_3D8A70, D_800A1764_A2364);
-}
+}

symbol_addrs.txt

@@ -450,4 +450,27 @@ guPerspective = 0x8007B010;
 guNormalize = 0x80082B00;
 gCurrentPlayerIndex = 0x800CD067;
 guLookAtF = 0x80088210;
-guLookAt = 0x80079800;
+guLookAt = 0x80079800;
+HuSqrtf = 0x800880E0;
+
+HuVecGetLengthSqr = 0x800880F4;
+HuVecGetLength = 0x80088120;
+HuVecNormalize = 0x80088150;
+HuVecDotProduct = 0x800881E0;
+HuVecCopyXYZ = 0x80089A10;
+HuVecCopy3F = 0x80089A20;
+HuVecNeg = 0x80089A40;
+HuVecAdd = 0x80089A70;
+HuVecSubtract = 0x80089AB0;
+HuVecMulScalar = 0x80089AF0;
+HuVecAddScaled = 0x80089B20;
+HuVecLinearComb = 0x80089B70;
+HuVecLinearComb2D = 0x80089BD0;
+HuVecGetLengthSqr3F = 0x80089C30;
+HuVecGetLength3F = 0x80089C60;
+HuVecNormalize3F = 0x80089C90;
+HuVecDistanceSqr = 0x80089D20;
+HuVecDistance = 0x80089D60;
+HuVecDot3F = 0x80089DB0;
+HuVecCross = 0x80089DE0;
+HuVecIsEqual = 0x80089E30;