summaryrefslogtreecommitdiffstats
path: root/libs/rs
diff options
context:
space:
mode:
Diffstat (limited to 'libs/rs')
-rw-r--r--libs/rs/java/tests/src/com/android/rs/test/math.rs140
-rw-r--r--libs/rs/rsContext.cpp27
-rw-r--r--libs/rs/rsLocklessFifo.cpp7
-rw-r--r--libs/rs/rsScriptC.cpp32
-rw-r--r--libs/rs/rsScriptC.h7
-rw-r--r--libs/rs/rsScriptC_Lib.cpp532
-rw-r--r--libs/rs/rsScriptC_LibCL.cpp4
-rw-r--r--libs/rs/scriptc/rs_cl.rsh759
-rw-r--r--libs/rs/scriptc/rs_core.rsh984
9 files changed, 975 insertions, 1517 deletions
diff --git a/libs/rs/java/tests/src/com/android/rs/test/math.rs b/libs/rs/java/tests/src/com/android/rs/test/math.rs
index 02993fe..8cad82b 100644
--- a/libs/rs/java/tests/src/com/android/rs/test/math.rs
+++ b/libs/rs/java/tests/src/com/android/rs/test/math.rs
@@ -12,6 +12,31 @@ volatile int2 i2;
volatile int3 i3;
volatile int4 i4;
+volatile uint ui1;
+volatile uint2 ui2;
+volatile uint3 ui3;
+volatile uint4 ui4;
+
+volatile short s1;
+volatile short2 s2;
+volatile short3 s3;
+volatile short4 s4;
+
+volatile ushort us1;
+volatile ushort2 us2;
+volatile ushort3 us3;
+volatile ushort4 us4;
+
+volatile char c1;
+volatile char2 c2;
+volatile char3 c3;
+volatile char4 c4;
+
+volatile uchar uc1;
+volatile uchar2 uc2;
+volatile uchar3 uc3;
+volatile uchar4 uc4;
+
#define TEST_FN_FUNC_FN(fnc) \
rsDebug("Testing " #fnc, 0); \
f1 = fnc(f1); \
@@ -168,9 +193,124 @@ static bool test_fp_math(uint32_t index) {
return failed;
}
+#define DECL_INT(prefix) \
+volatile char prefix##_c_1 = 1; \
+volatile char2 prefix##_c_2 = 1; \
+volatile char3 prefix##_c_3 = 1; \
+volatile char4 prefix##_c_4 = 1; \
+volatile uchar prefix##_uc_1 = 1; \
+volatile uchar2 prefix##_uc_2 = 1; \
+volatile uchar3 prefix##_uc_3 = 1; \
+volatile uchar4 prefix##_uc_4 = 1; \
+volatile short prefix##_s_1 = 1; \
+volatile short2 prefix##_s_2 = 1; \
+volatile short3 prefix##_s_3 = 1; \
+volatile short4 prefix##_s_4 = 1; \
+volatile ushort prefix##_us_1 = 1; \
+volatile ushort2 prefix##_us_2 = 1; \
+volatile ushort3 prefix##_us_3 = 1; \
+volatile ushort4 prefix##_us_4 = 1; \
+volatile int prefix##_i_1 = 1; \
+volatile int2 prefix##_i_2 = 1; \
+volatile int3 prefix##_i_3 = 1; \
+volatile int4 prefix##_i_4 = 1; \
+volatile uint prefix##_ui_1 = 1; \
+volatile uint2 prefix##_ui_2 = 1; \
+volatile uint3 prefix##_ui_3 = 1; \
+volatile uint4 prefix##_ui_4 = 1; \
+volatile long prefix##_l_1 = 1; \
+volatile ulong prefix##_ul_1 = 1;
+
+#define TEST_INT_OP_TYPE(op, type) \
+rsDebug("Testing " #op " for " #type "1", i++); \
+res_##type##_1 = src1_##type##_1 op src2_##type##_1; \
+rsDebug("Testing " #op " for " #type "2", i++); \
+res_##type##_2 = src1_##type##_2 op src2_##type##_2; \
+rsDebug("Testing " #op " for " #type "3", i++); \
+res_##type##_3 = src1_##type##_3 op src2_##type##_3; \
+rsDebug("Testing " #op " for " #type "4", i++); \
+res_##type##_4 = src1_##type##_4 op src2_##type##_4;
+
+#define TEST_INT_OP(op) \
+TEST_INT_OP_TYPE(op, c) \
+TEST_INT_OP_TYPE(op, uc) \
+TEST_INT_OP_TYPE(op, s) \
+TEST_INT_OP_TYPE(op, us) \
+TEST_INT_OP_TYPE(op, i) \
+TEST_INT_OP_TYPE(op, ui) \
+rsDebug("Testing " #op " for l1", i++); \
+res_l_1 = src1_l_1 op src2_l_1; \
+rsDebug("Testing " #op " for ul1", i++); \
+res_ul_1 = src1_ul_1 op src2_ul_1;
+
+DECL_INT(res)
+DECL_INT(src1)
+DECL_INT(src2)
+
+static bool test_basic_operators() {
+ bool failed = false;
+ int i = 0;
+
+ TEST_INT_OP(+);
+ TEST_INT_OP(-);
+ TEST_INT_OP(*);
+ TEST_INT_OP(/);
+ TEST_INT_OP(%);
+ TEST_INT_OP(<<);
+ TEST_INT_OP(>>);
+
+ if (failed) {
+ rsDebug("test_basic_operators FAILED", 0);
+ }
+ else {
+ rsDebug("test_basic_operators PASSED", 0);
+ }
+
+ return failed;
+}
+
+#define TEST_CVT(to, from, type) \
+rsDebug("Testing convert from " #from " to " #to, 0); \
+to##1 = from##1; \
+to##2 = convert_##type##2(from##2); \
+to##3 = convert_##type##3(from##3); \
+to##4 = convert_##type##4(from##4);
+
+#define TEST_CVT_MATRIX(to, type) \
+TEST_CVT(to, c, type); \
+TEST_CVT(to, uc, type); \
+TEST_CVT(to, s, type); \
+TEST_CVT(to, us, type); \
+TEST_CVT(to, i, type); \
+TEST_CVT(to, ui, type); \
+TEST_CVT(to, f, type); \
+
+static bool test_convert() {
+ bool failed = false;
+
+ TEST_CVT_MATRIX(c, char);
+ TEST_CVT_MATRIX(uc, uchar);
+ TEST_CVT_MATRIX(s, short);
+ TEST_CVT_MATRIX(us, ushort);
+ TEST_CVT_MATRIX(i, int);
+ TEST_CVT_MATRIX(ui, uint);
+ TEST_CVT_MATRIX(f, float);
+
+ if (failed) {
+ rsDebug("test_convert FAILED", 0);
+ }
+ else {
+ rsDebug("test_convert PASSED", 0);
+ }
+
+ return failed;
+}
+
void math_test(uint32_t index, int test_num) {
bool failed = false;
+ failed |= test_convert();
failed |= test_fp_math(index);
+ failed |= test_basic_operators();
if (failed) {
rsSendToClientBlocking(RS_MSG_TEST_FAILED);
diff --git a/libs/rs/rsContext.cpp b/libs/rs/rsContext.cpp
index 3acb624..40cb5c7 100644
--- a/libs/rs/rsContext.cpp
+++ b/libs/rs/rsContext.cpp
@@ -201,9 +201,9 @@ bool Context::initGLThread() {
mGL.mExtensions = glGetString(GL_EXTENSIONS);
//LOGV("EGL Version %i %i", mEGL.mMajorVersion, mEGL.mMinorVersion);
- LOGV("GL Version %s", mGL.mVersion);
+ //LOGV("GL Version %s", mGL.mVersion);
//LOGV("GL Vendor %s", mGL.mVendor);
- LOGV("GL Renderer %s", mGL.mRenderer);
+ //LOGV("GL Renderer %s", mGL.mRenderer);
//LOGV("GL Extensions %s", mGL.mExtensions);
const char *verptr = NULL;
@@ -468,7 +468,6 @@ void * Context::threadProc(void *vrsc) {
return NULL;
}
- rsc->mScriptC.init(rsc);
if (rsc->mIsGraphicsContext) {
rsc->mStateRaster.init(rsc);
rsc->setProgramRaster(NULL);
@@ -528,7 +527,7 @@ void * Context::threadProc(void *vrsc) {
}
void Context::destroyWorkerThreadResources() {
- LOGV("destroyWorkerThreadResources 1");
+ //LOGV("destroyWorkerThreadResources 1");
if (mIsGraphicsContext) {
mRaster.clear();
mFragment.clear();
@@ -544,7 +543,7 @@ void Context::destroyWorkerThreadResources() {
mShaderCache.cleanupAll();
}
ObjectBase::zeroAllUserRef(this);
- LOGV("destroyWorkerThreadResources 2");
+ //LOGV("destroyWorkerThreadResources 2");
mExit = true;
}
@@ -552,7 +551,7 @@ void * Context::helperThreadProc(void *vrsc) {
Context *rsc = static_cast<Context *>(vrsc);
uint32_t idx = (uint32_t)android_atomic_inc(&rsc->mWorkers.mLaunchCount);
- LOGV("RS helperThread starting %p idx=%i", rsc, idx);
+ //LOGV("RS helperThread starting %p idx=%i", rsc, idx);
rsc->mWorkers.mLaunchSignals[idx].init();
rsc->mWorkers.mNativeThreadId[idx] = gettid();
@@ -573,7 +572,7 @@ void * Context::helperThreadProc(void *vrsc) {
LOGE("pthread_setspecific %i", status);
}
- while (rsc->mRunning) {
+ while (!rsc->mExit) {
rsc->mWorkers.mLaunchSignals[idx].wait();
if (rsc->mWorkers.mLaunchCallback) {
rsc->mWorkers.mLaunchCallback(rsc->mWorkers.mLaunchData, idx);
@@ -582,7 +581,7 @@ void * Context::helperThreadProc(void *vrsc) {
rsc->mWorkers.mCompleteSignal.set();
}
- LOGV("RS helperThread exiting %p idx=%i", rsc, idx);
+ //LOGV("RS helperThread exited %p idx=%i", rsc, idx);
return NULL;
}
@@ -730,6 +729,18 @@ Context::~Context() {
mIO.shutdown();
int status = pthread_join(mThreadId, &res);
+ // Cleanup compute threads.
+ mWorkers.mLaunchData = NULL;
+ mWorkers.mLaunchCallback = NULL;
+ mWorkers.mRunningCount = (int)mWorkers.mCount;
+ for (uint32_t ct = 0; ct < mWorkers.mCount; ct++) {
+ mWorkers.mLaunchSignals[ct].set();
+ }
+ for (uint32_t ct = 0; ct < mWorkers.mCount; ct++) {
+ int status = pthread_join(mWorkers.mThreadId[ct], &res);
+ }
+ rsAssert(!mWorkers.mRunningCount);
+
// Global structure cleanup.
pthread_mutex_lock(&gInitMutex);
if (mDev) {
diff --git a/libs/rs/rsLocklessFifo.cpp b/libs/rs/rsLocklessFifo.cpp
index eb2af1c..3f88543 100644
--- a/libs/rs/rsLocklessFifo.cpp
+++ b/libs/rs/rsLocklessFifo.cpp
@@ -76,7 +76,8 @@ uint32_t LocklessCommandFifo::getFreeSpace() const {
}
bool LocklessCommandFifo::isEmpty() const {
- return mPut == mGet;
+ uint32_t p = android_atomic_acquire_load((int32_t *)&mPut);
+ return ((uint8_t *)p) == mGet;
}
@@ -155,7 +156,9 @@ const void * LocklessCommandFifo::get(uint32_t *command, uint32_t *bytesData) {
void LocklessCommandFifo::next() {
uint32_t bytes = reinterpret_cast<const uint16_t *>(mGet)[1];
- mGet += ((bytes + 3) & ~3) + 4;
+
+ android_atomic_add(((bytes + 3) & ~3) + 4, (int32_t *)&mGet);
+ //mGet += ((bytes + 3) & ~3) + 4;
if (isEmpty()) {
mSignalToControl.set();
}
diff --git a/libs/rs/rsScriptC.cpp b/libs/rs/rsScriptC.cpp
index eecfa16..3858e1c 100644
--- a/libs/rs/rsScriptC.cpp
+++ b/libs/rs/rsScriptC.cpp
@@ -421,21 +421,9 @@ void ScriptC::Invoke(Context *rsc, uint32_t slot, const void *data, uint32_t len
}
ScriptCState::ScriptCState() {
- mScript.clear();
}
ScriptCState::~ScriptCState() {
- mScript.clear();
-}
-
-void ScriptCState::init(Context *rsc) {
- clear(rsc);
-}
-
-void ScriptCState::clear(Context *rsc) {
- rsAssert(rsc);
- mScript.clear();
- mScript.set(new ScriptC(rsc));
}
static void* symbolLookup(void* pContext, char const* name) {
@@ -608,8 +596,6 @@ namespace android {
namespace renderscript {
void rsi_ScriptCBegin(Context * rsc) {
- ScriptCState *ss = &rsc->mScriptC;
- ss->clear(rsc);
}
void rsi_ScriptCSetText(Context *rsc, const char *text, uint32_t len) {
@@ -618,8 +604,8 @@ void rsi_ScriptCSetText(Context *rsc, const char *text, uint32_t len) {
char *t = (char *)malloc(len + 1);
memcpy(t, text, len);
t[len] = 0;
- ss->mScript->mEnviroment.mScriptText = t;
- ss->mScript->mEnviroment.mScriptTextLength = len;
+ ss->mScriptText = t;
+ ss->mScriptLen = len;
}
@@ -630,17 +616,19 @@ RsScript rsi_ScriptCCreate(Context *rsc,
{
ScriptCState *ss = &rsc->mScriptC;
- ObjectBaseRef<ScriptC> s(ss->mScript);
- ss->mScript.clear();
+ ScriptC *s = new ScriptC(rsc);
+ s->mEnviroment.mScriptText = ss->mScriptText;
+ s->mEnviroment.mScriptTextLength = ss->mScriptLen;
+ ss->mScriptText = NULL;
+ ss->mScriptLen = 0;
s->incUserRef();
- if (!ss->runCompiler(rsc, s.get(), resName, cacheDir)) {
+ if (!ss->runCompiler(rsc, s, resName, cacheDir)) {
// Error during compile, destroy s and return null.
- s->zeroUserRef();
+ delete s;
return NULL;
}
- ss->clear(rsc);
- return s.get();
+ return s;
}
}
diff --git a/libs/rs/rsScriptC.h b/libs/rs/rsScriptC.h
index 612e38a..7143c67 100644
--- a/libs/rs/rsScriptC.h
+++ b/libs/rs/rsScriptC.h
@@ -76,11 +76,9 @@ public:
ScriptCState();
~ScriptCState();
- ObjectBaseRef<ScriptC> mScript;
+ char * mScriptText;
+ size_t mScriptLen;
- void init(Context *rsc);
-
- void clear(Context *rsc);
bool runCompiler(Context *rsc, ScriptC *s, const char *resName, const char *cacheDir);
struct SymbolTable_t {
@@ -88,7 +86,6 @@ public:
void * mPtr;
bool threadable;
};
- //static SymbolTable_t gSyms[];
static const SymbolTable_t * lookupSymbol(const char *);
static const SymbolTable_t * lookupSymbolCL(const char *);
static const SymbolTable_t * lookupSymbolGL(const char *);
diff --git a/libs/rs/rsScriptC_Lib.cpp b/libs/rs/rsScriptC_Lib.cpp
index f550d98..8a85f6e 100644
--- a/libs/rs/rsScriptC_Lib.cpp
+++ b/libs/rs/rsScriptC_Lib.cpp
@@ -305,6 +305,14 @@ int SC_modsi3(int a, int b) {
return a % b;
}
+unsigned int SC_udivsi3(unsigned int a, unsigned int b) {
+ return a / b;
+}
+
+unsigned int SC_umodsi3(unsigned int a, unsigned int b) {
+ return a % b;
+}
+
int SC_getAllocation(const void *ptr) {
GET_TLS();
const Allocation *alloc = sc->ptrToAllocation(ptr);
@@ -339,6 +347,489 @@ void SC_ForEach2(RsScript vs,
s->runForEach(rsc, ain, aout, usr, call);
}
+
+//////////////////////////////////////////////////////////////////////////////
+// Heavy math functions
+//////////////////////////////////////////////////////////////////////////////
+
+typedef struct {
+ float m[16];
+} rs_matrix4x4;
+
+typedef struct {
+ float m[9];
+} rs_matrix3x3;
+
+typedef struct {
+ float m[4];
+} rs_matrix2x2;
+
+static inline void
+rsMatrixSet(rs_matrix4x4 *m, uint32_t row, uint32_t col, float v) {
+ m->m[row * 4 + col] = v;
+}
+
+static inline float
+rsMatrixGet(const rs_matrix4x4 *m, uint32_t row, uint32_t col) {
+ return m->m[row * 4 + col];
+}
+
+static inline void
+rsMatrixSet(rs_matrix3x3 *m, uint32_t row, uint32_t col, float v) {
+ m->m[row * 3 + col] = v;
+}
+
+static inline float
+rsMatrixGet(const rs_matrix3x3 *m, uint32_t row, uint32_t col) {
+ return m->m[row * 3 + col];
+}
+
+static inline void
+rsMatrixSet(rs_matrix2x2 *m, uint32_t row, uint32_t col, float v) {
+ m->m[row * 2 + col] = v;
+}
+
+static inline float
+rsMatrixGet(const rs_matrix2x2 *m, uint32_t row, uint32_t col) {
+ return m->m[row * 2 + col];
+}
+
+
+static void SC_MatrixLoadIdentity_4x4(rs_matrix4x4 *m) {
+ m->m[0] = 1.f;
+ m->m[1] = 0.f;
+ m->m[2] = 0.f;
+ m->m[3] = 0.f;
+ m->m[4] = 0.f;
+ m->m[5] = 1.f;
+ m->m[6] = 0.f;
+ m->m[7] = 0.f;
+ m->m[8] = 0.f;
+ m->m[9] = 0.f;
+ m->m[10] = 1.f;
+ m->m[11] = 0.f;
+ m->m[12] = 0.f;
+ m->m[13] = 0.f;
+ m->m[14] = 0.f;
+ m->m[15] = 1.f;
+}
+
+static void SC_MatrixLoadIdentity_3x3(rs_matrix3x3 *m) {
+ m->m[0] = 1.f;
+ m->m[1] = 0.f;
+ m->m[2] = 0.f;
+ m->m[3] = 0.f;
+ m->m[4] = 1.f;
+ m->m[5] = 0.f;
+ m->m[6] = 0.f;
+ m->m[7] = 0.f;
+ m->m[8] = 1.f;
+}
+
+static void SC_MatrixLoadIdentity_2x2(rs_matrix2x2 *m) {
+ m->m[0] = 1.f;
+ m->m[1] = 0.f;
+ m->m[2] = 0.f;
+ m->m[3] = 1.f;
+}
+
+static void SC_MatrixLoad_4x4_f(rs_matrix4x4 *m, const float *v) {
+ m->m[0] = v[0];
+ m->m[1] = v[1];
+ m->m[2] = v[2];
+ m->m[3] = v[3];
+ m->m[4] = v[4];
+ m->m[5] = v[5];
+ m->m[6] = v[6];
+ m->m[7] = v[7];
+ m->m[8] = v[8];
+ m->m[9] = v[9];
+ m->m[10] = v[10];
+ m->m[11] = v[11];
+ m->m[12] = v[12];
+ m->m[13] = v[13];
+ m->m[14] = v[14];
+ m->m[15] = v[15];
+}
+
+static void SC_MatrixLoad_3x3_f(rs_matrix3x3 *m, const float *v) {
+ m->m[0] = v[0];
+ m->m[1] = v[1];
+ m->m[2] = v[2];
+ m->m[3] = v[3];
+ m->m[4] = v[4];
+ m->m[5] = v[5];
+ m->m[6] = v[6];
+ m->m[7] = v[7];
+ m->m[8] = v[8];
+}
+
+static void SC_MatrixLoad_2x2_f(rs_matrix2x2 *m, const float *v) {
+ m->m[0] = v[0];
+ m->m[1] = v[1];
+ m->m[2] = v[2];
+ m->m[3] = v[3];
+}
+
+static void SC_MatrixLoad_4x4_4x4(rs_matrix4x4 *m, const rs_matrix4x4 *v) {
+ m->m[0] = v->m[0];
+ m->m[1] = v->m[1];
+ m->m[2] = v->m[2];
+ m->m[3] = v->m[3];
+ m->m[4] = v->m[4];
+ m->m[5] = v->m[5];
+ m->m[6] = v->m[6];
+ m->m[7] = v->m[7];
+ m->m[8] = v->m[8];
+ m->m[9] = v->m[9];
+ m->m[10] = v->m[10];
+ m->m[11] = v->m[11];
+ m->m[12] = v->m[12];
+ m->m[13] = v->m[13];
+ m->m[14] = v->m[14];
+ m->m[15] = v->m[15];
+}
+
+static void SC_MatrixLoad_4x4_3x3(rs_matrix4x4 *m, const rs_matrix3x3 *v) {
+ m->m[0] = v->m[0];
+ m->m[1] = v->m[1];
+ m->m[2] = v->m[2];
+ m->m[3] = 0.f;
+ m->m[4] = v->m[3];
+ m->m[5] = v->m[4];
+ m->m[6] = v->m[5];
+ m->m[7] = 0.f;
+ m->m[8] = v->m[6];
+ m->m[9] = v->m[7];
+ m->m[10] = v->m[8];
+ m->m[11] = 0.f;
+ m->m[12] = 0.f;
+ m->m[13] = 0.f;
+ m->m[14] = 0.f;
+ m->m[15] = 1.f;
+}
+
+static void SC_MatrixLoad_4x4_2x2(rs_matrix4x4 *m, const rs_matrix2x2 *v) {
+ m->m[0] = v->m[0];
+ m->m[1] = v->m[1];
+ m->m[2] = 0.f;
+ m->m[3] = 0.f;
+ m->m[4] = v->m[2];
+ m->m[5] = v->m[3];
+ m->m[6] = 0.f;
+ m->m[7] = 0.f;
+ m->m[8] = 0.f;
+ m->m[9] = 0.f;
+ m->m[10] = 1.f;
+ m->m[11] = 0.f;
+ m->m[12] = 0.f;
+ m->m[13] = 0.f;
+ m->m[14] = 0.f;
+ m->m[15] = 1.f;
+}
+
+static void SC_MatrixLoad_3x3_3x3(rs_matrix3x3 *m, const rs_matrix3x3 *v) {
+ m->m[0] = v->m[0];
+ m->m[1] = v->m[1];
+ m->m[2] = v->m[2];
+ m->m[3] = v->m[3];
+ m->m[4] = v->m[4];
+ m->m[5] = v->m[5];
+ m->m[6] = v->m[6];
+ m->m[7] = v->m[7];
+ m->m[8] = v->m[8];
+}
+
+static void SC_MatrixLoad_2x2_2x2(rs_matrix2x2 *m, const rs_matrix2x2 *v) {
+ m->m[0] = v->m[0];
+ m->m[1] = v->m[1];
+ m->m[2] = v->m[2];
+ m->m[3] = v->m[3];
+}
+
+static void SC_MatrixLoadRotate(rs_matrix4x4 *m, float rot, float x, float y, float z) {
+ float c, s;
+ m->m[3] = 0;
+ m->m[7] = 0;
+ m->m[11]= 0;
+ m->m[12]= 0;
+ m->m[13]= 0;
+ m->m[14]= 0;
+ m->m[15]= 1;
+ rot *= (float)(M_PI / 180.0f);
+ c = cos(rot);
+ s = sin(rot);
+
+ const float len = x*x + y*y + z*z;
+ if (len != 1) {
+ const float recipLen = 1.f / sqrt(len);
+ x *= recipLen;
+ y *= recipLen;
+ z *= recipLen;
+ }
+ const float nc = 1.0f - c;
+ const float xy = x * y;
+ const float yz = y * z;
+ const float zx = z * x;
+ const float xs = x * s;
+ const float ys = y * s;
+ const float zs = z * s;
+ m->m[ 0] = x*x*nc + c;
+ m->m[ 4] = xy*nc - zs;
+ m->m[ 8] = zx*nc + ys;
+ m->m[ 1] = xy*nc + zs;
+ m->m[ 5] = y*y*nc + c;
+ m->m[ 9] = yz*nc - xs;
+ m->m[ 2] = zx*nc - ys;
+ m->m[ 6] = yz*nc + xs;
+ m->m[10] = z*z*nc + c;
+}
+
+static void SC_MatrixLoadScale(rs_matrix4x4 *m, float x, float y, float z) {
+ SC_MatrixLoadIdentity_4x4(m);
+ m->m[0] = x;
+ m->m[5] = y;
+ m->m[10] = z;
+}
+
+static void SC_MatrixLoadTranslate(rs_matrix4x4 *m, float x, float y, float z) {
+ SC_MatrixLoadIdentity_4x4(m);
+ m->m[12] = x;
+ m->m[13] = y;
+ m->m[14] = z;
+}
+
+static void SC_MatrixLoadMultiply_4x4_4x4_4x4(rs_matrix4x4 *m, const rs_matrix4x4 *lhs, const rs_matrix4x4 *rhs) {
+ for (int i=0 ; i<4 ; i++) {
+ float ri0 = 0;
+ float ri1 = 0;
+ float ri2 = 0;
+ float ri3 = 0;
+ for (int j=0 ; j<4 ; j++) {
+ const float rhs_ij = rsMatrixGet(rhs, i,j);
+ ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
+ ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
+ ri2 += rsMatrixGet(lhs, j, 2) * rhs_ij;
+ ri3 += rsMatrixGet(lhs, j, 3) * rhs_ij;
+ }
+ rsMatrixSet(m, i, 0, ri0);
+ rsMatrixSet(m, i, 1, ri1);
+ rsMatrixSet(m, i, 2, ri2);
+ rsMatrixSet(m, i, 3, ri3);
+ }
+}
+
+static void SC_MatrixMultiply_4x4_4x4(rs_matrix4x4 *m, const rs_matrix4x4 *rhs) {
+ rs_matrix4x4 mt;
+ SC_MatrixLoadMultiply_4x4_4x4_4x4(&mt, m, rhs);
+ SC_MatrixLoad_4x4_4x4(m, &mt);
+}
+
+static void SC_MatrixLoadMultiply_3x3_3x3_3x3(rs_matrix3x3 *m, const rs_matrix3x3 *lhs, const rs_matrix3x3 *rhs) {
+ for (int i=0 ; i<3 ; i++) {
+ float ri0 = 0;
+ float ri1 = 0;
+ float ri2 = 0;
+ for (int j=0 ; j<3 ; j++) {
+ const float rhs_ij = rsMatrixGet(rhs, i,j);
+ ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
+ ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
+ ri2 += rsMatrixGet(lhs, j, 2) * rhs_ij;
+ }
+ rsMatrixSet(m, i, 0, ri0);
+ rsMatrixSet(m, i, 1, ri1);
+ rsMatrixSet(m, i, 2, ri2);
+ }
+}
+
+static void SC_MatrixMultiply_3x3_3x3(rs_matrix3x3 *m, const rs_matrix3x3 *rhs) {
+ rs_matrix3x3 mt;
+ SC_MatrixLoadMultiply_3x3_3x3_3x3(&mt, m, rhs);
+ SC_MatrixLoad_3x3_3x3(m, &mt);
+}
+
+static void SC_MatrixLoadMultiply_2x2_2x2_2x2(rs_matrix2x2 *m, const rs_matrix2x2 *lhs, const rs_matrix2x2 *rhs) {
+ for (int i=0 ; i<2 ; i++) {
+ float ri0 = 0;
+ float ri1 = 0;
+ for (int j=0 ; j<2 ; j++) {
+ const float rhs_ij = rsMatrixGet(rhs, i,j);
+ ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
+ ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
+ }
+ rsMatrixSet(m, i, 0, ri0);
+ rsMatrixSet(m, i, 1, ri1);
+ }
+}
+
+static void SC_MatrixMultiply_2x2_2x2(rs_matrix2x2 *m, const rs_matrix2x2 *rhs) {
+ rs_matrix2x2 mt;
+ SC_MatrixLoadMultiply_2x2_2x2_2x2(&mt, m, rhs);
+ SC_MatrixLoad_2x2_2x2(m, &mt);
+}
+
+static void SC_MatrixRotate(rs_matrix4x4 *m, float rot, float x, float y, float z) {
+ rs_matrix4x4 m1;
+ SC_MatrixLoadRotate(&m1, rot, x, y, z);
+ SC_MatrixMultiply_4x4_4x4(m, &m1);
+}
+
+static void SC_MatrixScale(rs_matrix4x4 *m, float x, float y, float z) {
+ rs_matrix4x4 m1;
+ SC_MatrixLoadScale(&m1, x, y, z);
+ SC_MatrixMultiply_4x4_4x4(m, &m1);
+}
+
+static void SC_MatrixTranslate(rs_matrix4x4 *m, float x, float y, float z) {
+ rs_matrix4x4 m1;
+ SC_MatrixLoadTranslate(&m1, x, y, z);
+ SC_MatrixMultiply_4x4_4x4(m, &m1);
+}
+
+static void SC_MatrixLoadOrtho(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far) {
+ SC_MatrixLoadIdentity_4x4(m);
+ m->m[0] = 2.f / (right - left);
+ m->m[5] = 2.f / (top - bottom);
+ m->m[10]= -2.f / (far - near);
+ m->m[12]= -(right + left) / (right - left);
+ m->m[13]= -(top + bottom) / (top - bottom);
+ m->m[14]= -(far + near) / (far - near);
+}
+
+static void SC_MatrixLoadFrustum(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far) {
+ SC_MatrixLoadIdentity_4x4(m);
+ m->m[0] = 2.f * near / (right - left);
+ m->m[5] = 2.f * near / (top - bottom);
+ m->m[8] = (right + left) / (right - left);
+ m->m[9] = (top + bottom) / (top - bottom);
+ m->m[10]= -(far + near) / (far - near);
+ m->m[11]= -1.f;
+ m->m[14]= -2.f * far * near / (far - near);
+ m->m[15]= 0.f;
+}
+
+static void SC_MatrixLoadPerspective(rs_matrix4x4* m, float fovy, float aspect, float near, float far) {
+ float top = near * tan((float) (fovy * M_PI / 360.0f));
+ float bottom = -top;
+ float left = bottom * aspect;
+ float right = top * aspect;
+ SC_MatrixLoadFrustum(m, left, right, bottom, top, near, far);
+}
+
+
+// Returns true if the matrix was successfully inversed
+static bool SC_MatrixInverse_4x4(rs_matrix4x4 *m) {
+ rs_matrix4x4 result;
+
+ int i, j;
+ for (i = 0; i < 4; ++i) {
+ for (j = 0; j < 4; ++j) {
+ // computeCofactor for int i, int j
+ int c0 = (i+1) % 4;
+ int c1 = (i+2) % 4;
+ int c2 = (i+3) % 4;
+ int r0 = (j+1) % 4;
+ int r1 = (j+2) % 4;
+ int r2 = (j+3) % 4;
+
+ float minor = (m->m[c0 + 4*r0] * (m->m[c1 + 4*r1] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r1]))
+ - (m->m[c0 + 4*r1] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r0]))
+ + (m->m[c0 + 4*r2] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r1] - m->m[c1 + 4*r1] * m->m[c2 + 4*r0]));
+
+ float cofactor = (i+j) & 1 ? -minor : minor;
+
+ result.m[4*i + j] = cofactor;
+ }
+ }
+
+ // Dot product of 0th column of source and 0th row of result
+ float det = m->m[0]*result.m[0] + m->m[4]*result.m[1] +
+ m->m[8]*result.m[2] + m->m[12]*result.m[3];
+
+ if (fabs(det) < 1e-6) {
+ return false;
+ }
+
+ det = 1.0f / det;
+ for (i = 0; i < 16; ++i) {
+ m->m[i] = result.m[i] * det;
+ }
+
+ return true;
+}
+
+// Returns true if the matrix was successfully inversed
+static bool SC_MatrixInverseTranspose_4x4(rs_matrix4x4 *m) {
+ rs_matrix4x4 result;
+
+ int i, j;
+ for (i = 0; i < 4; ++i) {
+ for (j = 0; j < 4; ++j) {
+ // computeCofactor for int i, int j
+ int c0 = (i+1) % 4;
+ int c1 = (i+2) % 4;
+ int c2 = (i+3) % 4;
+ int r0 = (j+1) % 4;
+ int r1 = (j+2) % 4;
+ int r2 = (j+3) % 4;
+
+ float minor = (m->m[c0 + 4*r0] * (m->m[c1 + 4*r1] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r1]))
+ - (m->m[c0 + 4*r1] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r0]))
+ + (m->m[c0 + 4*r2] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r1] - m->m[c1 + 4*r1] * m->m[c2 + 4*r0]));
+
+ float cofactor = (i+j) & 1 ? -minor : minor;
+
+ result.m[4*j + i] = cofactor;
+ }
+ }
+
+ // Dot product of 0th column of source and 0th column of result
+ float det = m->m[0]*result.m[0] + m->m[4]*result.m[4] +
+ m->m[8]*result.m[8] + m->m[12]*result.m[12];
+
+ if (fabs(det) < 1e-6) {
+ return false;
+ }
+
+ det = 1.0f / det;
+ for (i = 0; i < 16; ++i) {
+ m->m[i] = result.m[i] * det;
+ }
+
+ return true;
+}
+
+static void SC_MatrixTranspose_4x4(rs_matrix4x4 *m) {
+ int i, j;
+ float temp;
+ for (i = 0; i < 3; ++i) {
+ for (j = i + 1; j < 4; ++j) {
+ temp = m->m[i*4 + j];
+ m->m[i*4 + j] = m->m[j*4 + i];
+ m->m[j*4 + i] = temp;
+ }
+ }
+}
+
+static void SC_MatrixTranspose_3x3(rs_matrix3x3 *m) {
+ int i, j;
+ float temp;
+ for (i = 0; i < 2; ++i) {
+ for (j = i + 1; j < 3; ++j) {
+ temp = m->m[i*3 + j];
+ m->m[i*3 + j] = m->m[j*4 + i];
+ m->m[j*3 + i] = temp;
+ }
+ }
+}
+
+static void SC_MatrixTranspose_2x2(rs_matrix2x2 *m) {
+ float temp = m->m[1];
+ m->m[1] = m->m[2];
+ m->m[2] = temp;
+}
+
+
//////////////////////////////////////////////////////////////////////////////
// Class implementation
//////////////////////////////////////////////////////////////////////////////
@@ -363,6 +854,8 @@ void SC_ForEach2(RsScript vs,
static ScriptCState::SymbolTable_t gSyms[] = {
{ "__divsi3", (void *)&SC_divsi3, true },
{ "__modsi3", (void *)&SC_modsi3, true },
+ { "__udivsi3", (void *)&SC_udivsi3, true },
+ { "__umodsi3", (void *)&SC_umodsi3, true },
// allocation
{ "_Z19rsAllocationGetDimX13rs_allocation", (void *)&SC_allocGetDimX, true },
@@ -463,6 +956,45 @@ static ScriptCState::SymbolTable_t gSyms[] = {
{ "_Z22rsSendToClientBlockingi", (void *)&SC_toClientBlocking, false },
{ "_Z22rsSendToClientBlockingiPKvj", (void *)&SC_toClientBlocking2, false },
+ // matrix
+ { "_Z20rsMatrixLoadIdentityP12rs_matrix4x4", (void *)&SC_MatrixLoadIdentity_4x4, false },
+ { "_Z20rsMatrixLoadIdentityP12rs_matrix3x3", (void *)&SC_MatrixLoadIdentity_3x3, false },
+ { "_Z20rsMatrixLoadIdentityP12rs_matrix2x2", (void *)&SC_MatrixLoadIdentity_2x2, false },
+
+ { "_Z12rsMatrixLoadP12rs_matrix4x4PKf", (void *)&SC_MatrixLoad_4x4_f, false },
+ { "_Z12rsMatrixLoadP12rs_matrix3x3PKf", (void *)&SC_MatrixLoad_3x3_f, false },
+ { "_Z12rsMatrixLoadP12rs_matrix2x2PKf", (void *)&SC_MatrixLoad_2x2_f, false },
+
+ { "_Z12rsMatrixLoadP12rs_matrix4x4PKS_", (void *)&SC_MatrixLoad_4x4_4x4, false },
+ { "_Z12rsMatrixLoadP12rs_matrix4x4PK12rs_matrix3x3", (void *)&SC_MatrixLoad_4x4_3x3, false },
+ { "_Z12rsMatrixLoadP12rs_matrix4x4PK12rs_matrix2x2", (void *)&SC_MatrixLoad_4x4_2x2, false },
+ { "_Z12rsMatrixLoadP12rs_matrix3x3PKS_", (void *)&SC_MatrixLoad_3x3_3x3, false },
+ { "_Z12rsMatrixLoadP12rs_matrix2x2PKS_", (void *)&SC_MatrixLoad_2x2_2x2, false },
+
+ { "_Z18rsMatrixLoadRotateP12rs_matrix4x4ffff", (void *)&SC_MatrixLoadRotate, false },
+ { "_Z17rsMatrixLoadScaleP12rs_matrix4x4fff", (void *)&SC_MatrixLoadScale, false },
+ { "_Z21rsMatrixLoadTranslateP12rs_matrix4x4fff", (void *)&SC_MatrixLoadTranslate, false },
+ { "_Z14rsMatrixRotateP12rs_matrix4x4ffff", (void *)&SC_MatrixRotate, false },
+ { "_Z13rsMatrixScaleP12rs_matrix4x4fff", (void *)&SC_MatrixScale, false },
+ { "_Z17rsMatrixTranslateP12rs_matrix4x4fff", (void *)&SC_MatrixTranslate, false },
+
+ { "_Z20rsMatrixLoadMultiplyP12rs_matrix4x4PKS_S2_", (void *)&SC_MatrixLoadMultiply_4x4_4x4_4x4, false },
+ { "_Z16rsMatrixMultiplyP12rs_matrix4x4PKS_", (void *)&SC_MatrixMultiply_4x4_4x4, false },
+ { "_Z20rsMatrixLoadMultiplyP12rs_matrix3x3PKS_S2_", (void *)&SC_MatrixLoadMultiply_3x3_3x3_3x3, false },
+ { "_Z16rsMatrixMultiplyP12rs_matrix3x3PKS_", (void *)&SC_MatrixMultiply_3x3_3x3, false },
+ { "_Z20rsMatrixLoadMultiplyP12rs_matrix2x2PKS_S2_", (void *)&SC_MatrixLoadMultiply_2x2_2x2_2x2, false },
+ { "_Z16rsMatrixMultiplyP12rs_matrix2x2PKS_", (void *)&SC_MatrixMultiply_2x2_2x2, false },
+
+ { "_Z17rsMatrixLoadOrthoP12rs_matrix4x4ffffff", (void *)&SC_MatrixLoadOrtho, false },
+ { "_Z19rsMatrixLoadFrustumP12rs_matrix4x4ffffff", (void *)&SC_MatrixLoadFrustum, false },
+ { "_Z23rsMatrixLoadPerspectiveP12rs_matrix4x4ffff", (void *)&SC_MatrixLoadPerspective, false },
+
+ { "_Z15rsMatrixInverseP12rs_matrix4x4", (void *)&SC_MatrixInverse_4x4, false },
+ { "_Z24rsMatrixInverseTransposeP12rs_matrix4x4", (void *)&SC_MatrixInverseTranspose_4x4, false },
+ { "_Z17rsMatrixTransposeP12rs_matrix4x4", (void *)&SC_MatrixTranspose_4x4, false },
+ { "_Z17rsMatrixTransposeP12rs_matrix4x4", (void *)&SC_MatrixTranspose_3x3, false },
+ { "_Z17rsMatrixTransposeP12rs_matrix4x4", (void *)&SC_MatrixTranspose_2x2, false },
+
{ "_Z9rsForEach9rs_script13rs_allocationS0_PKv", (void *)&SC_ForEach, false },
//{ "_Z9rsForEach9rs_script13rs_allocationS0_PKv", (void *)&SC_ForEach2, true },
diff --git a/libs/rs/rsScriptC_LibCL.cpp b/libs/rs/rsScriptC_LibCL.cpp
index 02d33b7..57855db 100644
--- a/libs/rs/rsScriptC_LibCL.cpp
+++ b/libs/rs/rsScriptC_LibCL.cpp
@@ -195,7 +195,7 @@ static ScriptCState::SymbolTable_t gSyms[] = {
{ "_Z4logbf", (void *)&logbf, true },
{ "_Z3madfff", (void *)&SC_mad, true },
{ "_Z4modffPf", (void *)&modff, true },
- //{ "nan", (void *)&, true },
+ //{ "_Z3nanj", (void *)&SC_nan, true },
{ "_Z9nextafterff", (void *)&nextafterf, true },
{ "_Z3powff", (void *)&powf, true },
{ "_Z9remainderff", (void *)&remainderf, true },
@@ -210,7 +210,7 @@ static ScriptCState::SymbolTable_t gSyms[] = {
{ "_Z4sqrtf", (void *)&sqrtf, true },
{ "_Z3tanf", (void *)&tanf, true },
{ "_Z4tanhf", (void *)&tanhf, true },
- { "_Z6tgammaf", (void *)&lgammaf, true }, // FIXME!!! NEEDS TO USE tgammaf
+ { "_Z6tgammaf", (void *)&tgammaf, true },
{ "_Z5truncf", (void *)&truncf, true },
// OpenCL Int
diff --git a/libs/rs/scriptc/rs_cl.rsh b/libs/rs/scriptc/rs_cl.rsh
index 3c0496d..d78e62e 100644
--- a/libs/rs/scriptc/rs_cl.rsh
+++ b/libs/rs/scriptc/rs_cl.rsh
@@ -1,30 +1,17 @@
#ifndef __RS_CL_RSH__
#define __RS_CL_RSH__
-#ifdef BCC_PREPARE_BC
-#define _RS_STATIC extern
-#else
-#define _RS_STATIC static
-#endif
+#define _RS_RUNTIME extern
// Conversions
#define CVT_FUNC_2(typeout, typein) \
-_RS_STATIC typeout##2 __attribute__((overloadable)) \
- convert_##typeout##2(typein##2 v) { \
- typeout##2 r = {(typeout)v.x, (typeout)v.y}; \
- return r; \
-} \
-_RS_STATIC typeout##3 __attribute__((overloadable)) \
- convert_##typeout##3(typein##3 v) { \
- typeout##3 r = {(typeout)v.x, (typeout)v.y, (typeout)v.z}; \
- return r; \
-} \
-_RS_STATIC typeout##4 __attribute__((overloadable)) \
- convert_##typeout##4(typein##4 v) { \
- typeout##4 r = {(typeout)v.x, (typeout)v.y, (typeout)v.z, \
- (typeout)v.w}; \
- return r; \
-}
+_RS_RUNTIME typeout##2 __attribute__((overloadable)) \
+ convert_##typeout##2(typein##2 v); \
+_RS_RUNTIME typeout##3 __attribute__((overloadable)) \
+ convert_##typeout##3(typein##3 v); \
+_RS_RUNTIME typeout##4 __attribute__((overloadable)) \
+ convert_##typeout##4(typein##4 v);
+
#define CVT_FUNC(type) CVT_FUNC_2(type, uchar) \
CVT_FUNC_2(type, char) \
@@ -45,279 +32,63 @@ CVT_FUNC(float)
// Float ops, 6.11.2
#define FN_FUNC_FN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v) { \
- float2 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v) { \
- float3 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v) { \
- float4 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- r.w = fnc(v.w); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v);
#define IN_FUNC_FN(fnc) \
-_RS_STATIC int2 __attribute__((overloadable)) fnc(float2 v) { \
- int2 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- return r; \
-} \
-_RS_STATIC int3 __attribute__((overloadable)) fnc(float3 v) { \
- int3 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- return r; \
-} \
-_RS_STATIC int4 __attribute__((overloadable)) fnc(float4 v) { \
- int4 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- r.w = fnc(v.w); \
- return r; \
-}
+_RS_RUNTIME int2 __attribute__((overloadable)) fnc(float2 v); \
+_RS_RUNTIME int3 __attribute__((overloadable)) fnc(float3 v); \
+_RS_RUNTIME int4 __attribute__((overloadable)) fnc(float4 v);
#define FN_FUNC_FN_FN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v1, float2 v2) { \
- float2 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v1, float3 v2) { \
- float3 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v1, float4 v2) { \
- float4 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- r.w = fnc(v1.w, v2.w); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v1, float2 v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v1, float3 v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v1, float4 v2);
#define FN_FUNC_FN_F(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v1, float v2) { \
- float2 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v1, float v2) { \
- float3 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- r.z = fnc(v1.z, v2); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v1, float v2) { \
- float4 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- r.z = fnc(v1.z, v2); \
- r.w = fnc(v1.w, v2); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v1, float v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v1, float v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v1, float v2);
#define FN_FUNC_FN_IN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v1, int2 v2) { \
- float2 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v1, int3 v2) { \
- float3 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v1, int4 v2) { \
- float4 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- r.w = fnc(v1.w, v2.w); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v1, int2 v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v1, int3 v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v1, int4 v2); \
#define FN_FUNC_FN_I(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v1, int v2) { \
- float2 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v1, int v2) { \
- float3 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- r.z = fnc(v1.z, v2); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v1, int v2) { \
- float4 r; \
- r.x = fnc(v1.x, v2); \
- r.y = fnc(v1.y, v2); \
- r.z = fnc(v1.z, v2); \
- r.w = fnc(v1.w, v2); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v1, int v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v1, int v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v1, int v2);
#define FN_FUNC_FN_PFN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) \
- fnc(float2 v1, float2 *v2) { \
- float2 r; \
- float t[2]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) \
- fnc(float3 v1, float3 *v2) { \
- float3 r; \
- float t[3]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- r.z = fnc(v1.z, &t[2]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- v2->z = t[2]; \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) \
- fnc(float4 v1, float4 *v2) { \
- float4 r; \
- float t[4]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- r.z = fnc(v1.z, &t[2]); \
- r.w = fnc(v1.w, &t[3]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- v2->z = t[2]; \
- v2->w = t[3]; \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) \
+ fnc(float2 v1, float2 *v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) \
+ fnc(float3 v1, float3 *v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) \
+ fnc(float4 v1, float4 *v2);
#define FN_FUNC_FN_PIN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) fnc(float2 v1, int2 *v2) { \
- float2 r; \
- int t[2]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) fnc(float3 v1, int3 *v2) { \
- float3 r; \
- int t[3]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- r.z = fnc(v1.z, &t[2]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- v2->z = t[2]; \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) fnc(float4 v1, int4 *v2) { \
- float4 r; \
- int t[4]; \
- r.x = fnc(v1.x, &t[0]); \
- r.y = fnc(v1.y, &t[1]); \
- r.z = fnc(v1.z, &t[2]); \
- r.w = fnc(v1.w, &t[3]); \
- v2->x = t[0]; \
- v2->y = t[1]; \
- v2->z = t[2]; \
- v2->w = t[3]; \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) fnc(float2 v1, int2 *v2); \
+_RS_RUNTIME float3 __attribute__((overloadable)) fnc(float3 v1, int3 *v2); \
+_RS_RUNTIME float4 __attribute__((overloadable)) fnc(float4 v1, int4 *v2);
#define FN_FUNC_FN_FN_FN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) \
- fnc(float2 v1, float2 v2, float2 v3) { \
- float2 r; \
- r.x = fnc(v1.x, v2.x, v3.x); \
- r.y = fnc(v1.y, v2.y, v3.y); \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) \
- fnc(float3 v1, float3 v2, float3 v3) { \
- float3 r; \
- r.x = fnc(v1.x, v2.x, v3.x); \
- r.y = fnc(v1.y, v2.y, v3.y); \
- r.z = fnc(v1.z, v2.z, v3.z); \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) \
- fnc(float4 v1, float4 v2, float4 v3) { \
- float4 r; \
- r.x = fnc(v1.x, v2.x, v3.x); \
- r.y = fnc(v1.y, v2.y, v3.y); \
- r.z = fnc(v1.z, v2.z, v3.z); \
- r.w = fnc(v1.w, v2.w, v3.w); \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) \
+ fnc(float2 v1, float2 v2, float2 v3); \
+_RS_RUNTIME float3 __attribute__((overloadable)) \
+ fnc(float3 v1, float3 v2, float3 v3); \
+_RS_RUNTIME float4 __attribute__((overloadable)) \
+ fnc(float4 v1, float4 v2, float4 v3);
#define FN_FUNC_FN_FN_PIN(fnc) \
-_RS_STATIC float2 __attribute__((overloadable)) \
- fnc(float2 v1, float2 v2, int2 *v3) { \
- float2 r; \
- int t[2]; \
- r.x = fnc(v1.x, v2.x, &t[0]); \
- r.y = fnc(v1.y, v2.y, &t[1]); \
- v3->x = t[0]; \
- v3->y = t[1]; \
- return r; \
-} \
-_RS_STATIC float3 __attribute__((overloadable)) \
- fnc(float3 v1, float3 v2, int3 *v3) { \
- float3 r; \
- int t[3]; \
- r.x = fnc(v1.x, v2.x, &t[0]); \
- r.y = fnc(v1.y, v2.y, &t[1]); \
- r.z = fnc(v1.z, v2.z, &t[2]); \
- v3->x = t[0]; \
- v3->y = t[1]; \
- v3->z = t[2]; \
- return r; \
-} \
-_RS_STATIC float4 __attribute__((overloadable)) \
- fnc(float4 v1, float4 v2, int4 *v3) { \
- float4 r; \
- int t[4]; \
- r.x = fnc(v1.x, v2.x, &t[0]); \
- r.y = fnc(v1.y, v2.y, &t[1]); \
- r.z = fnc(v1.z, v2.z, &t[2]); \
- r.w = fnc(v1.w, v2.w, &t[3]); \
- v3->x = t[0]; \
- v3->y = t[1]; \
- v3->z = t[2]; \
- v3->w = t[3]; \
- return r; \
-}
+_RS_RUNTIME float2 __attribute__((overloadable)) \
+ fnc(float2 v1, float2 v2, int2 *v3); \
+_RS_RUNTIME float3 __attribute__((overloadable)) \
+ fnc(float3 v1, float3 v2, int3 *v3); \
+_RS_RUNTIME float4 __attribute__((overloadable)) \
+ fnc(float4 v1, float4 v2, int4 *v3);
extern float __attribute__((overloadable)) acos(float);
@@ -326,9 +97,9 @@ FN_FUNC_FN(acos)
extern float __attribute__((overloadable)) acosh(float);
FN_FUNC_FN(acosh)
-_RS_STATIC float __attribute__((overloadable)) acospi(float v) {
- return acos(v) / M_PI;
-}
+_RS_RUNTIME float __attribute__((overloadable)) acospi(float v);
+
+
FN_FUNC_FN(acospi)
extern float __attribute__((overloadable)) asin(float);
@@ -337,9 +108,8 @@ FN_FUNC_FN(asin)
extern float __attribute__((overloadable)) asinh(float);
FN_FUNC_FN(asinh)
-_RS_STATIC float __attribute__((overloadable)) asinpi(float v) {
- return asin(v) / M_PI;
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) asinpi(float v);
FN_FUNC_FN(asinpi)
extern float __attribute__((overloadable)) atan(float);
@@ -351,14 +121,12 @@ FN_FUNC_FN_FN(atan2)
extern float __attribute__((overloadable)) atanh(float);
FN_FUNC_FN(atanh)
-_RS_STATIC float __attribute__((overloadable)) atanpi(float v) {
- return atan(v) / M_PI;
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) atanpi(float v);
FN_FUNC_FN(atanpi)
-_RS_STATIC float __attribute__((overloadable)) atan2pi(float y, float x) {
- return atan2(y, x) / M_PI;
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) atan2pi(float y, float x);
FN_FUNC_FN_FN(atan2pi)
extern float __attribute__((overloadable)) cbrt(float);
@@ -376,9 +144,8 @@ FN_FUNC_FN(cos)
extern float __attribute__((overloadable)) cosh(float);
FN_FUNC_FN(cosh)
-_RS_STATIC float __attribute__((overloadable)) cospi(float v) {
- return cos(v * M_PI);
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) cospi(float v);
FN_FUNC_FN(cospi)
extern float __attribute__((overloadable)) erfc(float);
@@ -394,9 +161,8 @@ extern float __attribute__((overloadable)) exp2(float);
FN_FUNC_FN(exp2)
extern float __attribute__((overloadable)) pow(float, float);
-_RS_STATIC float __attribute__((overloadable)) exp10(float v) {
- return pow(10.f, v);
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) exp10(float v);
FN_FUNC_FN(exp10)
extern float __attribute__((overloadable)) expm1(float);
@@ -425,11 +191,8 @@ FN_FUNC_FN_F(fmin);
extern float __attribute__((overloadable)) fmod(float, float);
FN_FUNC_FN_FN(fmod)
-_RS_STATIC float __attribute__((overloadable)) fract(float v, float *iptr) {
- int i = (int)floor(v);
- iptr[0] = i;
- return fmin(v - i, 0x1.fffffep-1f);
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) fract(float v, float *iptr);
FN_FUNC_FN_PFN(fract)
extern float __attribute__((overloadable)) frexp(float, int *);
@@ -457,9 +220,8 @@ FN_FUNC_FN(log)
extern float __attribute__((overloadable)) log10(float);
FN_FUNC_FN(log10)
-_RS_STATIC float __attribute__((overloadable)) log2(float v) {
- return log10(v) / log10(2.f);
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) log2(float v);
FN_FUNC_FN(log2)
extern float __attribute__((overloadable)) log1p(float);
@@ -481,31 +243,15 @@ FN_FUNC_FN_FN(nextafter)
FN_FUNC_FN_FN(pow)
-_RS_STATIC float __attribute__((overloadable)) pown(float v, int p) {
- return pow(v, (float)p);
-}
-_RS_STATIC float2 __attribute__((overloadable)) pown(float2 v, int2 p) {
- return pow(v, (float2)p);
-}
-_RS_STATIC float3 __attribute__((overloadable)) pown(float3 v, int3 p) {
- return pow(v, (float3)p);
-}
-_RS_STATIC float4 __attribute__((overloadable)) pown(float4 v, int4 p) {
- return pow(v, (float4)p);
-}
-
-_RS_STATIC float __attribute__((overloadable)) powr(float v, float p) {
- return pow(v, p);
-}
-_RS_STATIC float2 __attribute__((overloadable)) powr(float2 v, float2 p) {
- return pow(v, p);
-}
-_RS_STATIC float3 __attribute__((overloadable)) powr(float3 v, float3 p) {
- return pow(v, p);
-}
-_RS_STATIC float4 __attribute__((overloadable)) powr(float4 v, float4 p) {
- return pow(v, p);
-}
+_RS_RUNTIME float __attribute__((overloadable)) pown(float v, int p);
+_RS_RUNTIME float2 __attribute__((overloadable)) pown(float2 v, int2 p);
+_RS_RUNTIME float3 __attribute__((overloadable)) pown(float3 v, int3 p);
+_RS_RUNTIME float4 __attribute__((overloadable)) pown(float4 v, int4 p);
+
+_RS_RUNTIME float __attribute__((overloadable)) powr(float v, float p);
+_RS_RUNTIME float2 __attribute__((overloadable)) powr(float2 v, float2 p);
+_RS_RUNTIME float3 __attribute__((overloadable)) powr(float3 v, float3 p);
+_RS_RUNTIME float4 __attribute__((overloadable)) powr(float4 v, float4 p);
extern float __attribute__((overloadable)) remainder(float, float);
FN_FUNC_FN_FN(remainder)
@@ -516,57 +262,33 @@ FN_FUNC_FN_FN_PIN(remquo)
extern float __attribute__((overloadable)) rint(float);
FN_FUNC_FN(rint)
-_RS_STATIC float __attribute__((overloadable)) rootn(float v, int r) {
- return pow(v, 1.f / r);
-}
-_RS_STATIC float2 __attribute__((overloadable)) rootn(float2 v, int2 r) {
- float2 t = {1.f / r.x, 1.f / r.y};
- return pow(v, t);
-}
-_RS_STATIC float3 __attribute__((overloadable)) rootn(float3 v, int3 r) {
- float3 t = {1.f / r.x, 1.f / r.y, 1.f / r.z};
- return pow(v, t);
-}
-_RS_STATIC float4 __attribute__((overloadable)) rootn(float4 v, int4 r) {
- float4 t = {1.f / r.x, 1.f / r.y, 1.f / r.z, 1.f / r.w};
- return pow(v, t);
-}
+
+_RS_RUNTIME float __attribute__((overloadable)) rootn(float v, int r);
+_RS_RUNTIME float2 __attribute__((overloadable)) rootn(float2 v, int2 r);
+_RS_RUNTIME float3 __attribute__((overloadable)) rootn(float3 v, int3 r);
+_RS_RUNTIME float4 __attribute__((overloadable)) rootn(float4 v, int4 r);
+
extern float __attribute__((overloadable)) round(float);
FN_FUNC_FN(round)
+
extern float __attribute__((overloadable)) sqrt(float);
-_RS_STATIC float __attribute__((overloadable)) rsqrt(float v) {
- return 1.f / sqrt(v);
-}
+_RS_RUNTIME float __attribute__((overloadable)) rsqrt(float v);
FN_FUNC_FN(rsqrt)
extern float __attribute__((overloadable)) sin(float);
FN_FUNC_FN(sin)
-_RS_STATIC float __attribute__((overloadable)) sincos(float v, float *cosptr) {
- *cosptr = cos(v);
- return sin(v);
-}
-_RS_STATIC float2 __attribute__((overloadable)) sincos(float2 v, float2 *cosptr) {
- *cosptr = cos(v);
- return sin(v);
-}
-_RS_STATIC float3 __attribute__((overloadable)) sincos(float3 v, float3 *cosptr) {
- *cosptr = cos(v);
- return sin(v);
-}
-_RS_STATIC float4 __attribute__((overloadable)) sincos(float4 v, float4 *cosptr) {
- *cosptr = cos(v);
- return sin(v);
-}
+_RS_RUNTIME float __attribute__((overloadable)) sincos(float v, float *cosptr);
+_RS_RUNTIME float2 __attribute__((overloadable)) sincos(float2 v, float2 *cosptr);
+_RS_RUNTIME float3 __attribute__((overloadable)) sincos(float3 v, float3 *cosptr);
+_RS_RUNTIME float4 __attribute__((overloadable)) sincos(float4 v, float4 *cosptr);
extern float __attribute__((overloadable)) sinh(float);
FN_FUNC_FN(sinh)
-_RS_STATIC float __attribute__((overloadable)) sinpi(float v) {
- return sin(v * M_PI);
-}
+_RS_RUNTIME float __attribute__((overloadable)) sinpi(float v);
FN_FUNC_FN(sinpi)
FN_FUNC_FN(sqrt)
@@ -577,11 +299,10 @@ FN_FUNC_FN(tan)
extern float __attribute__((overloadable)) tanh(float);
FN_FUNC_FN(tanh)
-_RS_STATIC float __attribute__((overloadable)) tanpi(float v) {
- return tan(v * M_PI);
-}
+_RS_RUNTIME float __attribute__((overloadable)) tanpi(float v);
FN_FUNC_FN(tanpi)
+
extern float __attribute__((overloadable)) tgamma(float);
FN_FUNC_FN(tgamma)
@@ -592,27 +313,9 @@ FN_FUNC_FN(trunc)
#define XN_FUNC_YN(typeout, fnc, typein) \
extern typeout __attribute__((overloadable)) fnc(typein); \
-_RS_STATIC typeout##2 __attribute__((overloadable)) fnc(typein##2 v) { \
- typeout##2 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- return r; \
-} \
-_RS_STATIC typeout##3 __attribute__((overloadable)) fnc(typein##3 v) { \
- typeout##3 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- return r; \
-} \
-_RS_STATIC typeout##4 __attribute__((overloadable)) fnc(typein##4 v) { \
- typeout##4 r; \
- r.x = fnc(v.x); \
- r.y = fnc(v.y); \
- r.z = fnc(v.z); \
- r.w = fnc(v.w); \
- return r; \
-}
+_RS_RUNTIME typeout##2 __attribute__((overloadable)) fnc(typein##2 v); \
+_RS_RUNTIME typeout##3 __attribute__((overloadable)) fnc(typein##3 v); \
+_RS_RUNTIME typeout##4 __attribute__((overloadable)) fnc(typein##4 v);
#define UIN_FUNC_IN(fnc) \
XN_FUNC_YN(uchar, fnc, char) \
@@ -627,35 +330,16 @@ XN_FUNC_YN(short, fnc, short) \
XN_FUNC_YN(uint, fnc, uint) \
XN_FUNC_YN(int, fnc, int)
+
#define XN_FUNC_XN_XN_BODY(type, fnc, body) \
-_RS_STATIC type __attribute__((overloadable)) \
- fnc(type v1, type v2) { \
- return body; \
-} \
-_RS_STATIC type##2 __attribute__((overloadable)) \
- fnc(type##2 v1, type##2 v2) { \
- type##2 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- return r; \
-} \
-_RS_STATIC type##3 __attribute__((overloadable)) \
- fnc(type##3 v1, type##3 v2) { \
- type##3 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- return r; \
-} \
-_RS_STATIC type##4 __attribute__((overloadable)) \
- fnc(type##4 v1, type##4 v2) { \
- type##4 r; \
- r.x = fnc(v1.x, v2.x); \
- r.y = fnc(v1.y, v2.y); \
- r.z = fnc(v1.z, v2.z); \
- r.w = fnc(v1.w, v2.w); \
- return r; \
-}
+_RS_RUNTIME type __attribute__((overloadable)) \
+ fnc(type v1, type v2); \
+_RS_RUNTIME type##2 __attribute__((overloadable)) \
+ fnc(type##2 v1, type##2 v2); \
+_RS_RUNTIME type##3 __attribute__((overloadable)) \
+ fnc(type##3 v1, type##3 v2); \
+_RS_RUNTIME type##4 __attribute__((overloadable)) \
+ fnc(type##4 v1, type##4 v2);
#define IN_FUNC_IN_IN_BODY(fnc, body) \
XN_FUNC_XN_XN_BODY(uchar, fnc, body) \
@@ -677,129 +361,35 @@ FN_FUNC_FN_F(max)
// 6.11.4
-_RS_STATIC float __attribute__((overloadable)) clamp(float amount, float low, float high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-_RS_STATIC float2 __attribute__((overloadable)) clamp(float2 amount, float2 low, float2 high) {
- float2 r;
- r.x = amount.x < low.x ? low.x : (amount.x > high.x ? high.x : amount.x);
- r.y = amount.y < low.y ? low.y : (amount.y > high.y ? high.y : amount.y);
- return r;
-}
-_RS_STATIC float3 __attribute__((overloadable)) clamp(float3 amount, float3 low, float3 high) {
- float3 r;
- r.x = amount.x < low.x ? low.x : (amount.x > high.x ? high.x : amount.x);
- r.y = amount.y < low.y ? low.y : (amount.y > high.y ? high.y : amount.y);
- r.z = amount.z < low.z ? low.z : (amount.z > high.z ? high.z : amount.z);
- return r;
-}
-_RS_STATIC float4 __attribute__((overloadable)) clamp(float4 amount, float4 low, float4 high) {
- float4 r;
- r.x = amount.x < low.x ? low.x : (amount.x > high.x ? high.x : amount.x);
- r.y = amount.y < low.y ? low.y : (amount.y > high.y ? high.y : amount.y);
- r.z = amount.z < low.z ? low.z : (amount.z > high.z ? high.z : amount.z);
- r.w = amount.w < low.w ? low.w : (amount.w > high.w ? high.w : amount.w);
- return r;
-}
-_RS_STATIC float2 __attribute__((overloadable)) clamp(float2 amount, float low, float high) {
- float2 r;
- r.x = amount.x < low ? low : (amount.x > high ? high : amount.x);
- r.y = amount.y < low ? low : (amount.y > high ? high : amount.y);
- return r;
-}
-_RS_STATIC float3 __attribute__((overloadable)) clamp(float3 amount, float low, float high) {
- float3 r;
- r.x = amount.x < low ? low : (amount.x > high ? high : amount.x);
- r.y = amount.y < low ? low : (amount.y > high ? high : amount.y);
- r.z = amount.z < low ? low : (amount.z > high ? high : amount.z);
- return r;
-}
-_RS_STATIC float4 __attribute__((overloadable)) clamp(float4 amount, float low, float high) {
- float4 r;
- r.x = amount.x < low ? low : (amount.x > high ? high : amount.x);
- r.y = amount.y < low ? low : (amount.y > high ? high : amount.y);
- r.z = amount.z < low ? low : (amount.z > high ? high : amount.z);
- r.w = amount.w < low ? low : (amount.w > high ? high : amount.w);
- return r;
-}
-
-_RS_STATIC float __attribute__((overloadable)) degrees(float radians) {
- return radians * (180.f / M_PI);
-}
+_RS_RUNTIME float __attribute__((overloadable)) clamp(float amount, float low, float high);
+_RS_RUNTIME float2 __attribute__((overloadable)) clamp(float2 amount, float2 low, float2 high);
+_RS_RUNTIME float3 __attribute__((overloadable)) clamp(float3 amount, float3 low, float3 high);
+_RS_RUNTIME float4 __attribute__((overloadable)) clamp(float4 amount, float4 low, float4 high);
+_RS_RUNTIME float2 __attribute__((overloadable)) clamp(float2 amount, float low, float high);
+_RS_RUNTIME float3 __attribute__((overloadable)) clamp(float3 amount, float low, float high);
+_RS_RUNTIME float4 __attribute__((overloadable)) clamp(float4 amount, float low, float high);
+
+_RS_RUNTIME float __attribute__((overloadable)) degrees(float radians);
FN_FUNC_FN(degrees)
-_RS_STATIC float __attribute__((overloadable)) mix(float start, float stop, float amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float2 amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float3 amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float4 amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float amount) {
- return start + (stop - start) * amount;
-}
-_RS_STATIC float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float amount) {
- return start + (stop - start) * amount;
-}
-
-_RS_STATIC float __attribute__((overloadable)) radians(float degrees) {
- return degrees * (M_PI / 180.f);
-}
+_RS_RUNTIME float __attribute__((overloadable)) mix(float start, float stop, float amount);
+_RS_RUNTIME float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float2 amount);
+_RS_RUNTIME float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float3 amount);
+_RS_RUNTIME float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float4 amount);
+_RS_RUNTIME float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float amount);
+_RS_RUNTIME float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float amount);
+_RS_RUNTIME float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float amount);
+
+_RS_RUNTIME float __attribute__((overloadable)) radians(float degrees);
FN_FUNC_FN(radians)
-_RS_STATIC float __attribute__((overloadable)) step(float edge, float v) {
- return (v < edge) ? 0.f : 1.f;
-}
-_RS_STATIC float2 __attribute__((overloadable)) step(float2 edge, float2 v) {
- float2 r;
- r.x = (v.x < edge.x) ? 0.f : 1.f;
- r.y = (v.y < edge.y) ? 0.f : 1.f;
- return r;
-}
-_RS_STATIC float3 __attribute__((overloadable)) step(float3 edge, float3 v) {
- float3 r;
- r.x = (v.x < edge.x) ? 0.f : 1.f;
- r.y = (v.y < edge.y) ? 0.f : 1.f;
- r.z = (v.z < edge.z) ? 0.f : 1.f;
- return r;
-}
-_RS_STATIC float4 __attribute__((overloadable)) step(float4 edge, float4 v) {
- float4 r;
- r.x = (v.x < edge.x) ? 0.f : 1.f;
- r.y = (v.y < edge.y) ? 0.f : 1.f;
- r.z = (v.z < edge.z) ? 0.f : 1.f;
- r.w = (v.w < edge.w) ? 0.f : 1.f;
- return r;
-}
-_RS_STATIC float2 __attribute__((overloadable)) step(float2 edge, float v) {
- float2 r;
- r.x = (v < edge.x) ? 0.f : 1.f;
- r.y = (v < edge.y) ? 0.f : 1.f;
- return r;
-}
-_RS_STATIC float3 __attribute__((overloadable)) step(float3 edge, float v) {
- float3 r;
- r.x = (v < edge.x) ? 0.f : 1.f;
- r.y = (v < edge.y) ? 0.f : 1.f;
- r.z = (v < edge.z) ? 0.f : 1.f;
- return r;
-}
-_RS_STATIC float4 __attribute__((overloadable)) step(float4 edge, float v) {
- float4 r;
- r.x = (v < edge.x) ? 0.f : 1.f;
- r.y = (v < edge.y) ? 0.f : 1.f;
- r.z = (v < edge.z) ? 0.f : 1.f;
- r.w = (v < edge.w) ? 0.f : 1.f;
- return r;
-}
+_RS_RUNTIME float __attribute__((overloadable)) step(float edge, float v);
+_RS_RUNTIME float2 __attribute__((overloadable)) step(float2 edge, float2 v);
+_RS_RUNTIME float3 __attribute__((overloadable)) step(float3 edge, float3 v);
+_RS_RUNTIME float4 __attribute__((overloadable)) step(float4 edge, float4 v);
+_RS_RUNTIME float2 __attribute__((overloadable)) step(float2 edge, float v);
+_RS_RUNTIME float3 __attribute__((overloadable)) step(float3 edge, float v);
+_RS_RUNTIME float4 __attribute__((overloadable)) step(float4 edge, float v);
extern float __attribute__((overloadable)) smoothstep(float, float, float);
extern float2 __attribute__((overloadable)) smoothstep(float2, float2, float2);
@@ -809,82 +399,33 @@ extern float2 __attribute__((overloadable)) smoothstep(float, float, float2);
extern float3 __attribute__((overloadable)) smoothstep(float, float, float3);
extern float4 __attribute__((overloadable)) smoothstep(float, float, float4);
-_RS_STATIC float __attribute__((overloadable)) sign(float v) {
- if (v > 0) return 1.f;
- if (v < 0) return -1.f;
- return v;
-}
+_RS_RUNTIME float __attribute__((overloadable)) sign(float v);
FN_FUNC_FN(sign)
// 6.11.5
-_RS_STATIC float3 __attribute__((overloadable)) cross(float3 lhs, float3 rhs) {
- float3 r;
- r.x = lhs.y * rhs.z - lhs.z * rhs.y;
- r.y = lhs.z * rhs.x - lhs.x * rhs.z;
- r.z = lhs.x * rhs.y - lhs.y * rhs.x;
- return r;
-}
-
-_RS_STATIC float4 __attribute__((overloadable)) cross(float4 lhs, float4 rhs) {
- float4 r;
- r.x = lhs.y * rhs.z - lhs.z * rhs.y;
- r.y = lhs.z * rhs.x - lhs.x * rhs.z;
- r.z = lhs.x * rhs.y - lhs.y * rhs.x;
- r.w = 0.f;
- return r;
-}
-
-_RS_STATIC float __attribute__((overloadable)) dot(float lhs, float rhs) {
- return lhs * rhs;
-}
-_RS_STATIC float __attribute__((overloadable)) dot(float2 lhs, float2 rhs) {
- return lhs.x*rhs.x + lhs.y*rhs.y;
-}
-_RS_STATIC float __attribute__((overloadable)) dot(float3 lhs, float3 rhs) {
- return lhs.x*rhs.x + lhs.y*rhs.y + lhs.z*rhs.z;
-}
-_RS_STATIC float __attribute__((overloadable)) dot(float4 lhs, float4 rhs) {
- return lhs.x*rhs.x + lhs.y*rhs.y + lhs.z*rhs.z + lhs.w*rhs.w;
-}
-
-_RS_STATIC float __attribute__((overloadable)) length(float v) {
- return v;
-}
-_RS_STATIC float __attribute__((overloadable)) length(float2 v) {
- return sqrt(v.x*v.x + v.y*v.y);
-}
-_RS_STATIC float __attribute__((overloadable)) length(float3 v) {
- return sqrt(v.x*v.x + v.y*v.y + v.z*v.z);
-}
-_RS_STATIC float __attribute__((overloadable)) length(float4 v) {
- return sqrt(v.x*v.x + v.y*v.y + v.z*v.z + v.w*v.w);
-}
-
-_RS_STATIC float __attribute__((overloadable)) distance(float lhs, float rhs) {
- return length(lhs - rhs);
-}
-_RS_STATIC float __attribute__((overloadable)) distance(float2 lhs, float2 rhs) {
- return length(lhs - rhs);
-}
-_RS_STATIC float __attribute__((overloadable)) distance(float3 lhs, float3 rhs) {
- return length(lhs - rhs);
-}
-_RS_STATIC float __attribute__((overloadable)) distance(float4 lhs, float4 rhs) {
- return length(lhs - rhs);
-}
-
-_RS_STATIC float __attribute__((overloadable)) normalize(float v) {
- return 1.f;
-}
-_RS_STATIC float2 __attribute__((overloadable)) normalize(float2 v) {
- return v / length(v);
-}
-_RS_STATIC float3 __attribute__((overloadable)) normalize(float3 v) {
- return v / length(v);
-}
-_RS_STATIC float4 __attribute__((overloadable)) normalize(float4 v) {
- return v / length(v);
-}
+_RS_RUNTIME float3 __attribute__((overloadable)) cross(float3 lhs, float3 rhs);
+
+_RS_RUNTIME float4 __attribute__((overloadable)) cross(float4 lhs, float4 rhs);
+
+_RS_RUNTIME float __attribute__((overloadable)) dot(float lhs, float rhs);
+_RS_RUNTIME float __attribute__((overloadable)) dot(float2 lhs, float2 rhs);
+_RS_RUNTIME float __attribute__((overloadable)) dot(float3 lhs, float3 rhs);
+_RS_RUNTIME float __attribute__((overloadable)) dot(float4 lhs, float4 rhs);
+
+_RS_RUNTIME float __attribute__((overloadable)) length(float v);
+_RS_RUNTIME float __attribute__((overloadable)) length(float2 v);
+_RS_RUNTIME float __attribute__((overloadable)) length(float3 v);
+_RS_RUNTIME float __attribute__((overloadable)) length(float4 v);
+
+_RS_RUNTIME float __attribute__((overloadable)) distance(float lhs, float rhs);
+_RS_RUNTIME float __attribute__((overloadable)) distance(float2 lhs, float2 rhs);
+_RS_RUNTIME float __attribute__((overloadable)) distance(float3 lhs, float3 rhs);
+_RS_RUNTIME float __attribute__((overloadable)) distance(float4 lhs, float4 rhs);
+
+_RS_RUNTIME float __attribute__((overloadable)) normalize(float v);
+_RS_RUNTIME float2 __attribute__((overloadable)) normalize(float2 v);
+_RS_RUNTIME float3 __attribute__((overloadable)) normalize(float3 v);
+_RS_RUNTIME float4 __attribute__((overloadable)) normalize(float4 v);
#undef CVT_FUNC
#undef CVT_FUNC_2
@@ -903,6 +444,6 @@ _RS_STATIC float4 __attribute__((overloadable)) normalize(float4 v) {
#undef IN_FUNC_IN
#undef XN_FUNC_XN_XN_BODY
#undef IN_FUNC_IN_IN_BODY
-#undef _RS_STATIC
+#undef _RS_RUNTIME
#endif
diff --git a/libs/rs/scriptc/rs_core.rsh b/libs/rs/scriptc/rs_core.rsh
index f3e0ab0..e32d435 100644
--- a/libs/rs/scriptc/rs_core.rsh
+++ b/libs/rs/scriptc/rs_core.rsh
@@ -1,11 +1,7 @@
#ifndef __RS_CORE_RSH__
#define __RS_CORE_RSH__
-#ifdef BCC_PREPARE_BC
-#define _RS_STATIC extern
-#else
-#define _RS_STATIC static
-#endif
+#define _RS_RUNTIME extern
// Debugging, print to the LOG a description string and a value.
extern void __attribute__((overloadable))
@@ -41,56 +37,19 @@ extern void __attribute__((overloadable))
#define RS_DEBUG(a) rsDebug(#a, a)
#define RS_DEBUG_MARKER rsDebug(__FILE__, __LINE__)
-_RS_STATIC void __attribute__((overloadable)) rsDebug(const char *s, float2 v) {
- rsDebug(s, v.x, v.y);
-}
-_RS_STATIC void __attribute__((overloadable)) rsDebug(const char *s, float3 v) {
- rsDebug(s, v.x, v.y, v.z);
-}
-_RS_STATIC void __attribute__((overloadable)) rsDebug(const char *s, float4 v) {
- rsDebug(s, v.x, v.y, v.z, v.w);
-}
-
-_RS_STATIC uchar4 __attribute__((overloadable)) rsPackColorTo8888(float r, float g, float b)
-{
- uchar4 c;
- c.x = (uchar)(r * 255.f);
- c.y = (uchar)(g * 255.f);
- c.z = (uchar)(b * 255.f);
- c.w = 255;
- return c;
-}
-
-_RS_STATIC uchar4 __attribute__((overloadable)) rsPackColorTo8888(float r, float g, float b, float a)
-{
- uchar4 c;
- c.x = (uchar)(r * 255.f);
- c.y = (uchar)(g * 255.f);
- c.z = (uchar)(b * 255.f);
- c.w = (uchar)(a * 255.f);
- return c;
-}
-
-_RS_STATIC uchar4 __attribute__((overloadable)) rsPackColorTo8888(float3 color)
-{
- color *= 255.f;
- uchar4 c = {color.x, color.y, color.z, 255};
- return c;
-}
-
-_RS_STATIC uchar4 __attribute__((overloadable)) rsPackColorTo8888(float4 color)
-{
- color *= 255.f;
- uchar4 c = {color.x, color.y, color.z, color.w};
- return c;
-}
-
-_RS_STATIC float4 rsUnpackColor8888(uchar4 c)
-{
- float4 ret = (float4)0.0039156862745f;
- ret *= convert_float4(c);
- return ret;
-}
+_RS_RUNTIME void __attribute__((overloadable)) rsDebug(const char *s, float2 v);
+_RS_RUNTIME void __attribute__((overloadable)) rsDebug(const char *s, float3 v);
+_RS_RUNTIME void __attribute__((overloadable)) rsDebug(const char *s, float4 v);
+
+_RS_RUNTIME uchar4 __attribute__((overloadable)) rsPackColorTo8888(float r, float g, float b);
+
+_RS_RUNTIME uchar4 __attribute__((overloadable)) rsPackColorTo8888(float r, float g, float b, float a);
+
+_RS_RUNTIME uchar4 __attribute__((overloadable)) rsPackColorTo8888(float3 color);
+
+_RS_RUNTIME uchar4 __attribute__((overloadable)) rsPackColorTo8888(float4 color);
+
+_RS_RUNTIME float4 rsUnpackColor8888(uchar4 c);
//extern uchar4 __attribute__((overloadable)) rsPackColorTo565(float r, float g, float b);
//extern uchar4 __attribute__((overloadable)) rsPackColorTo565(float3);
@@ -101,830 +60,117 @@ _RS_STATIC float4 rsUnpackColor8888(uchar4 c)
// Matrix ops
/////////////////////////////////////////////////////
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixSet(rs_matrix4x4 *m, uint32_t row, uint32_t col, float v) {
- m->m[row * 4 + col] = v;
-}
-
-_RS_STATIC float __attribute__((overloadable))
-rsMatrixGet(const rs_matrix4x4 *m, uint32_t row, uint32_t col) {
- return m->m[row * 4 + col];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixSet(rs_matrix3x3 *m, uint32_t row, uint32_t col, float v) {
- m->m[row * 3 + col] = v;
-}
-
-_RS_STATIC float __attribute__((overloadable))
-rsMatrixGet(const rs_matrix3x3 *m, uint32_t row, uint32_t col) {
- return m->m[row * 3 + col];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixSet(rs_matrix2x2 *m, uint32_t row, uint32_t col, float v) {
- m->m[row * 2 + col] = v;
-}
-
-_RS_STATIC float __attribute__((overloadable))
-rsMatrixGet(const rs_matrix2x2 *m, uint32_t row, uint32_t col) {
- return m->m[row * 2 + col];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadIdentity(rs_matrix4x4 *m) {
- m->m[0] = 1.f;
- m->m[1] = 0.f;
- m->m[2] = 0.f;
- m->m[3] = 0.f;
- m->m[4] = 0.f;
- m->m[5] = 1.f;
- m->m[6] = 0.f;
- m->m[7] = 0.f;
- m->m[8] = 0.f;
- m->m[9] = 0.f;
- m->m[10] = 1.f;
- m->m[11] = 0.f;
- m->m[12] = 0.f;
- m->m[13] = 0.f;
- m->m[14] = 0.f;
- m->m[15] = 1.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadIdentity(rs_matrix3x3 *m) {
- m->m[0] = 1.f;
- m->m[1] = 0.f;
- m->m[2] = 0.f;
- m->m[3] = 0.f;
- m->m[4] = 1.f;
- m->m[5] = 0.f;
- m->m[6] = 0.f;
- m->m[7] = 0.f;
- m->m[8] = 1.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadIdentity(rs_matrix2x2 *m) {
- m->m[0] = 1.f;
- m->m[1] = 0.f;
- m->m[2] = 0.f;
- m->m[3] = 1.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix4x4 *m, const float *v) {
- m->m[0] = v[0];
- m->m[1] = v[1];
- m->m[2] = v[2];
- m->m[3] = v[3];
- m->m[4] = v[4];
- m->m[5] = v[5];
- m->m[6] = v[6];
- m->m[7] = v[7];
- m->m[8] = v[8];
- m->m[9] = v[9];
- m->m[10] = v[10];
- m->m[11] = v[11];
- m->m[12] = v[12];
- m->m[13] = v[13];
- m->m[14] = v[14];
- m->m[15] = v[15];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix3x3 *m, const float *v) {
- m->m[0] = v[0];
- m->m[1] = v[1];
- m->m[2] = v[2];
- m->m[3] = v[3];
- m->m[4] = v[4];
- m->m[5] = v[5];
- m->m[6] = v[6];
- m->m[7] = v[7];
- m->m[8] = v[8];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix2x2 *m, const float *v) {
- m->m[0] = v[0];
- m->m[1] = v[1];
- m->m[2] = v[2];
- m->m[3] = v[3];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix4x4 *v) {
- m->m[0] = v->m[0];
- m->m[1] = v->m[1];
- m->m[2] = v->m[2];
- m->m[3] = v->m[3];
- m->m[4] = v->m[4];
- m->m[5] = v->m[5];
- m->m[6] = v->m[6];
- m->m[7] = v->m[7];
- m->m[8] = v->m[8];
- m->m[9] = v->m[9];
- m->m[10] = v->m[10];
- m->m[11] = v->m[11];
- m->m[12] = v->m[12];
- m->m[13] = v->m[13];
- m->m[14] = v->m[14];
- m->m[15] = v->m[15];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix3x3 *v) {
- m->m[0] = v->m[0];
- m->m[1] = v->m[1];
- m->m[2] = v->m[2];
- m->m[3] = 0.f;
- m->m[4] = v->m[3];
- m->m[5] = v->m[4];
- m->m[6] = v->m[5];
- m->m[7] = 0.f;
- m->m[8] = v->m[6];
- m->m[9] = v->m[7];
- m->m[10] = v->m[8];
- m->m[11] = 0.f;
- m->m[12] = 0.f;
- m->m[13] = 0.f;
- m->m[14] = 0.f;
- m->m[15] = 1.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix2x2 *v) {
- m->m[0] = v->m[0];
- m->m[1] = v->m[1];
- m->m[2] = 0.f;
- m->m[3] = 0.f;
- m->m[4] = v->m[3];
- m->m[5] = v->m[4];
- m->m[6] = 0.f;
- m->m[7] = 0.f;
- m->m[8] = v->m[6];
- m->m[9] = v->m[7];
- m->m[10] = 1.f;
- m->m[11] = 0.f;
- m->m[12] = 0.f;
- m->m[13] = 0.f;
- m->m[14] = 0.f;
- m->m[15] = 1.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix3x3 *m, const rs_matrix3x3 *v) {
- m->m[0] = v->m[0];
- m->m[1] = v->m[1];
- m->m[2] = v->m[2];
- m->m[3] = v->m[3];
- m->m[4] = v->m[4];
- m->m[5] = v->m[5];
- m->m[6] = v->m[6];
- m->m[7] = v->m[7];
- m->m[8] = v->m[8];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoad(rs_matrix2x2 *m, const rs_matrix2x2 *v) {
- m->m[0] = v->m[0];
- m->m[1] = v->m[1];
- m->m[2] = v->m[2];
- m->m[3] = v->m[3];
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadRotate(rs_matrix4x4 *m, float rot, float x, float y, float z) {
- float c, s;
- m->m[3] = 0;
- m->m[7] = 0;
- m->m[11]= 0;
- m->m[12]= 0;
- m->m[13]= 0;
- m->m[14]= 0;
- m->m[15]= 1;
- rot *= (float)(M_PI / 180.0f);
- c = cos(rot);
- s = sin(rot);
-
- const float len = x*x + y*y + z*z;
- if (len != 1) {
- const float recipLen = 1.f / sqrt(len);
- x *= recipLen;
- y *= recipLen;
- z *= recipLen;
- }
- const float nc = 1.0f - c;
- const float xy = x * y;
- const float yz = y * z;
- const float zx = z * x;
- const float xs = x * s;
- const float ys = y * s;
- const float zs = z * s;
- m->m[ 0] = x*x*nc + c;
- m->m[ 4] = xy*nc - zs;
- m->m[ 8] = zx*nc + ys;
- m->m[ 1] = xy*nc + zs;
- m->m[ 5] = y*y*nc + c;
- m->m[ 9] = yz*nc - xs;
- m->m[ 2] = zx*nc - ys;
- m->m[ 6] = yz*nc + xs;
- m->m[10] = z*z*nc + c;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadScale(rs_matrix4x4 *m, float x, float y, float z) {
- rsMatrixLoadIdentity(m);
- m->m[0] = x;
- m->m[5] = y;
- m->m[10] = z;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadTranslate(rs_matrix4x4 *m, float x, float y, float z) {
- rsMatrixLoadIdentity(m);
- m->m[12] = x;
- m->m[13] = y;
- m->m[14] = z;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadMultiply(rs_matrix4x4 *m, const rs_matrix4x4 *lhs, const rs_matrix4x4 *rhs) {
- for (int i=0 ; i<4 ; i++) {
- float ri0 = 0;
- float ri1 = 0;
- float ri2 = 0;
- float ri3 = 0;
- for (int j=0 ; j<4 ; j++) {
- const float rhs_ij = rsMatrixGet(rhs, i,j);
- ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
- ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
- ri2 += rsMatrixGet(lhs, j, 2) * rhs_ij;
- ri3 += rsMatrixGet(lhs, j, 3) * rhs_ij;
- }
- rsMatrixSet(m, i, 0, ri0);
- rsMatrixSet(m, i, 1, ri1);
- rsMatrixSet(m, i, 2, ri2);
- rsMatrixSet(m, i, 3, ri3);
- }
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix4x4 *m, const rs_matrix4x4 *rhs) {
- rs_matrix4x4 mt;
- rsMatrixLoadMultiply(&mt, m, rhs);
- rsMatrixLoad(m, &mt);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadMultiply(rs_matrix3x3 *m, const rs_matrix3x3 *lhs, const rs_matrix3x3 *rhs) {
- for (int i=0 ; i<3 ; i++) {
- float ri0 = 0;
- float ri1 = 0;
- float ri2 = 0;
- for (int j=0 ; j<3 ; j++) {
- const float rhs_ij = rsMatrixGet(rhs, i,j);
- ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
- ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
- ri2 += rsMatrixGet(lhs, j, 2) * rhs_ij;
- }
- rsMatrixSet(m, i, 0, ri0);
- rsMatrixSet(m, i, 1, ri1);
- rsMatrixSet(m, i, 2, ri2);
- }
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix3x3 *m, const rs_matrix3x3 *rhs) {
- rs_matrix3x3 mt;
- rsMatrixLoadMultiply(&mt, m, rhs);
- rsMatrixLoad(m, &mt);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadMultiply(rs_matrix2x2 *m, const rs_matrix2x2 *lhs, const rs_matrix2x2 *rhs) {
- for (int i=0 ; i<2 ; i++) {
- float ri0 = 0;
- float ri1 = 0;
- for (int j=0 ; j<2 ; j++) {
- const float rhs_ij = rsMatrixGet(rhs, i,j);
- ri0 += rsMatrixGet(lhs, j, 0) * rhs_ij;
- ri1 += rsMatrixGet(lhs, j, 1) * rhs_ij;
- }
- rsMatrixSet(m, i, 0, ri0);
- rsMatrixSet(m, i, 1, ri1);
- }
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix2x2 *m, const rs_matrix2x2 *rhs) {
- rs_matrix2x2 mt;
- rsMatrixLoadMultiply(&mt, m, rhs);
- rsMatrixLoad(m, &mt);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixRotate(rs_matrix4x4 *m, float rot, float x, float y, float z) {
- rs_matrix4x4 m1;
- rsMatrixLoadRotate(&m1, rot, x, y, z);
- rsMatrixMultiply(m, &m1);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixScale(rs_matrix4x4 *m, float x, float y, float z) {
- rs_matrix4x4 m1;
- rsMatrixLoadScale(&m1, x, y, z);
- rsMatrixMultiply(m, &m1);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixTranslate(rs_matrix4x4 *m, float x, float y, float z) {
- rs_matrix4x4 m1;
- rsMatrixLoadTranslate(&m1, x, y, z);
- rsMatrixMultiply(m, &m1);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadOrtho(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far) {
- rsMatrixLoadIdentity(m);
- m->m[0] = 2.f / (right - left);
- m->m[5] = 2.f / (top - bottom);
- m->m[10]= -2.f / (far - near);
- m->m[12]= -(right + left) / (right - left);
- m->m[13]= -(top + bottom) / (top - bottom);
- m->m[14]= -(far + near) / (far - near);
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadFrustum(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far) {
- rsMatrixLoadIdentity(m);
- m->m[0] = 2.f * near / (right - left);
- m->m[5] = 2.f * near / (top - bottom);
- m->m[8] = (right + left) / (right - left);
- m->m[9] = (top + bottom) / (top - bottom);
- m->m[10]= -(far + near) / (far - near);
- m->m[11]= -1.f;
- m->m[14]= -2.f * far * near / (far - near);
- m->m[15]= 0.f;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixLoadPerspective(rs_matrix4x4* m, float fovy, float aspect, float near, float far) {
- float top = near * tan((float) (fovy * M_PI / 360.0f));
- float bottom = -top;
- float left = bottom * aspect;
- float right = top * aspect;
- rsMatrixLoadFrustum(m, left, right, bottom, top, near, far);
-}
-
-_RS_STATIC float4 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix4x4 *m, float4 in) {
- float4 ret;
- ret.x = (m->m[0] * in.x) + (m->m[4] * in.y) + (m->m[8] * in.z) + (m->m[12] * in.w);
- ret.y = (m->m[1] * in.x) + (m->m[5] * in.y) + (m->m[9] * in.z) + (m->m[13] * in.w);
- ret.z = (m->m[2] * in.x) + (m->m[6] * in.y) + (m->m[10] * in.z) + (m->m[14] * in.w);
- ret.w = (m->m[3] * in.x) + (m->m[7] * in.y) + (m->m[11] * in.z) + (m->m[15] * in.w);
- return ret;
-}
-
-_RS_STATIC float4 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix4x4 *m, float3 in) {
- float4 ret;
- ret.x = (m->m[0] * in.x) + (m->m[4] * in.y) + (m->m[8] * in.z) + m->m[12];
- ret.y = (m->m[1] * in.x) + (m->m[5] * in.y) + (m->m[9] * in.z) + m->m[13];
- ret.z = (m->m[2] * in.x) + (m->m[6] * in.y) + (m->m[10] * in.z) + m->m[14];
- ret.w = (m->m[3] * in.x) + (m->m[7] * in.y) + (m->m[11] * in.z) + m->m[15];
- return ret;
-}
-
-_RS_STATIC float4 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix4x4 *m, float2 in) {
- float4 ret;
- ret.x = (m->m[0] * in.x) + (m->m[4] * in.y) + m->m[12];
- ret.y = (m->m[1] * in.x) + (m->m[5] * in.y) + m->m[13];
- ret.z = (m->m[2] * in.x) + (m->m[6] * in.y) + m->m[14];
- ret.w = (m->m[3] * in.x) + (m->m[7] * in.y) + m->m[15];
- return ret;
-}
-
-_RS_STATIC float3 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix3x3 *m, float3 in) {
- float3 ret;
- ret.x = (m->m[0] * in.x) + (m->m[3] * in.y) + (m->m[6] * in.z);
- ret.y = (m->m[1] * in.x) + (m->m[4] * in.y) + (m->m[7] * in.z);
- ret.z = (m->m[2] * in.x) + (m->m[5] * in.y) + (m->m[8] * in.z);
- return ret;
-}
-
-_RS_STATIC float3 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix3x3 *m, float2 in) {
- float3 ret;
- ret.x = (m->m[0] * in.x) + (m->m[3] * in.y);
- ret.y = (m->m[1] * in.x) + (m->m[4] * in.y);
- ret.z = (m->m[2] * in.x) + (m->m[5] * in.y);
- return ret;
-}
-
-_RS_STATIC float2 __attribute__((overloadable))
-rsMatrixMultiply(rs_matrix2x2 *m, float2 in) {
- float2 ret;
- ret.x = (m->m[0] * in.x) + (m->m[2] * in.y);
- ret.y = (m->m[1] * in.x) + (m->m[3] * in.y);
- return ret;
-}
+_RS_RUNTIME void __attribute__((overloadable))
+rsMatrixSet(rs_matrix4x4 *m, uint32_t row, uint32_t col, float v);
-// Returns true if the matrix was successfully inversed
-_RS_STATIC bool __attribute__((overloadable))
-rsMatrixInverse(rs_matrix4x4 *m) {
- rs_matrix4x4 result;
-
- int i, j;
- for (i = 0; i < 4; ++i) {
- for (j = 0; j < 4; ++j) {
- // computeCofactor for int i, int j
- int c0 = (i+1) % 4;
- int c1 = (i+2) % 4;
- int c2 = (i+3) % 4;
- int r0 = (j+1) % 4;
- int r1 = (j+2) % 4;
- int r2 = (j+3) % 4;
-
- float minor = (m->m[c0 + 4*r0] * (m->m[c1 + 4*r1] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r1]))
- - (m->m[c0 + 4*r1] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r0]))
- + (m->m[c0 + 4*r2] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r1] - m->m[c1 + 4*r1] * m->m[c2 + 4*r0]));
-
- float cofactor = (i+j) & 1 ? -minor : minor;
-
- result.m[4*i + j] = cofactor;
- }
- }
-
- // Dot product of 0th column of source and 0th row of result
- float det = m->m[0]*result.m[0] + m->m[4]*result.m[1] +
- m->m[8]*result.m[2] + m->m[12]*result.m[3];
-
- if (fabs(det) < 1e-6) {
- return false;
- }
-
- det = 1.0f / det;
- for (i = 0; i < 16; ++i) {
- m->m[i] = result.m[i] * det;
- }
-
- return true;
-}
+_RS_RUNTIME float __attribute__((overloadable))
+rsMatrixGet(const rs_matrix4x4 *m, uint32_t row, uint32_t col);
-// Returns true if the matrix was successfully inversed
-_RS_STATIC bool __attribute__((overloadable))
-rsMatrixInverseTranspose(rs_matrix4x4 *m) {
- rs_matrix4x4 result;
-
- int i, j;
- for (i = 0; i < 4; ++i) {
- for (j = 0; j < 4; ++j) {
- // computeCofactor for int i, int j
- int c0 = (i+1) % 4;
- int c1 = (i+2) % 4;
- int c2 = (i+3) % 4;
- int r0 = (j+1) % 4;
- int r1 = (j+2) % 4;
- int r2 = (j+3) % 4;
-
- float minor = (m->m[c0 + 4*r0] * (m->m[c1 + 4*r1] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r1]))
- - (m->m[c0 + 4*r1] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r2] - m->m[c1 + 4*r2] * m->m[c2 + 4*r0]))
- + (m->m[c0 + 4*r2] * (m->m[c1 + 4*r0] * m->m[c2 + 4*r1] - m->m[c1 + 4*r1] * m->m[c2 + 4*r0]));
-
- float cofactor = (i+j) & 1 ? -minor : minor;
-
- result.m[4*j + i] = cofactor;
- }
- }
-
- // Dot product of 0th column of source and 0th column of result
- float det = m->m[0]*result.m[0] + m->m[4]*result.m[4] +
- m->m[8]*result.m[8] + m->m[12]*result.m[12];
-
- if (fabs(det) < 1e-6) {
- return false;
- }
-
- det = 1.0f / det;
- for (i = 0; i < 16; ++i) {
- m->m[i] = result.m[i] * det;
- }
-
- return true;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixTranspose(rs_matrix4x4 *m) {
- int i, j;
- float temp;
- for (i = 0; i < 3; ++i) {
- for (j = i + 1; j < 4; ++j) {
- temp = m->m[i*4 + j];
- m->m[i*4 + j] = m->m[j*4 + i];
- m->m[j*4 + i] = temp;
- }
- }
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixTranspose(rs_matrix3x3 *m) {
- int i, j;
- float temp;
- for (i = 0; i < 2; ++i) {
- for (j = i + 1; j < 3; ++j) {
- temp = m->m[i*3 + j];
- m->m[i*3 + j] = m->m[j*4 + i];
- m->m[j*3 + i] = temp;
- }
- }
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsMatrixTranspose(rs_matrix2x2 *m) {
- float temp = m->m[1];
- m->m[1] = m->m[2];
- m->m[2] = temp;
-}
+_RS_RUNTIME void __attribute__((overloadable))
+rsMatrixSet(rs_matrix3x3 *m, uint32_t row, uint32_t col, float v);
-/////////////////////////////////////////////////////
-// quaternion ops
-/////////////////////////////////////////////////////
+_RS_RUNTIME float __attribute__((overloadable))
+rsMatrixGet(const rs_matrix3x3 *m, uint32_t row, uint32_t col);
-_RS_STATIC void __attribute__((overloadable))
-rsQuaternionSet(rs_quaternion *q, float w, float x, float y, float z) {
- q->w = w;
- q->x = x;
- q->y = y;
- q->z = z;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsQuaternionSet(rs_quaternion *q, const rs_quaternion *rhs) {
- q->w = rhs->w;
- q->x = rhs->x;
- q->y = rhs->y;
- q->z = rhs->z;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsQuaternionMultiply(rs_quaternion *q, float s) {
- q->w *= s;
- q->x *= s;
- q->y *= s;
- q->z *= s;
-}
-
-_RS_STATIC void __attribute__((overloadable))
-rsQuaternionMultiply(rs_quaternion *q, const rs_quaternion *rhs) {
- q->w = -q->x*rhs->x - q->y*rhs->y - q->z*rhs->z + q->w*rhs->w;
- q->x = q->x*rhs->w + q->y*rhs->z - q->z*rhs->y + q->w*rhs->x;
- q->y = -q->x*rhs->z + q->y*rhs->w + q->z*rhs->z + q->w*rhs->y;
- q->z = q->x*rhs->y - q->y*rhs->x + q->z*rhs->w + q->w*rhs->z;
-}
-
-_RS_STATIC void
-rsQuaternionAdd(rs_quaternion *q, const rs_quaternion *rhs) {
- q->w *= rhs->w;
- q->x *= rhs->x;
- q->y *= rhs->y;
- q->z *= rhs->z;
-}
-
-_RS_STATIC void
-rsQuaternionLoadRotateUnit(rs_quaternion *q, float rot, float x, float y, float z) {
- rot *= (float)(M_PI / 180.0f) * 0.5f;
- float c = cos(rot);
- float s = sin(rot);
-
- q->w = c;
- q->x = x * s;
- q->y = y * s;
- q->z = z * s;
-}
-
-_RS_STATIC void
-rsQuaternionLoadRotate(rs_quaternion *q, float rot, float x, float y, float z) {
- const float len = x*x + y*y + z*z;
- if (len != 1) {
- const float recipLen = 1.f / sqrt(len);
- x *= recipLen;
- y *= recipLen;
- z *= recipLen;
- }
- rsQuaternionLoadRotateUnit(q, rot, x, y, z);
-}
-
-_RS_STATIC void
-rsQuaternionConjugate(rs_quaternion *q) {
- q->x = -q->x;
- q->y = -q->y;
- q->z = -q->z;
-}
-
-_RS_STATIC float
-rsQuaternionDot(const rs_quaternion *q0, const rs_quaternion *q1) {
- return q0->w*q1->w + q0->x*q1->x + q0->y*q1->y + q0->z*q1->z;
-}
-
-_RS_STATIC void
-rsQuaternionNormalize(rs_quaternion *q) {
- const float len = rsQuaternionDot(q, q);
- if (len != 1) {
- const float recipLen = 1.f / sqrt(len);
- rsQuaternionMultiply(q, recipLen);
- }
-}
-
-_RS_STATIC void
-rsQuaternionSlerp(rs_quaternion *q, const rs_quaternion *q0, const rs_quaternion *q1, float t) {
- if (t <= 0.0f) {
- rsQuaternionSet(q, q0);
- return;
- }
- if (t >= 1.0f) {
- rsQuaternionSet(q, q1);
- return;
- }
-
- rs_quaternion tempq0, tempq1;
- rsQuaternionSet(&tempq0, q0);
- rsQuaternionSet(&tempq1, q1);
-
- float angle = rsQuaternionDot(q0, q1);
- if (angle < 0) {
- rsQuaternionMultiply(&tempq0, -1.0f);
- angle *= -1.0f;
- }
-
- float scale, invScale;
- if (angle + 1.0f > 0.05f) {
- if (1.0f - angle >= 0.05f) {
- float theta = acos(angle);
- float invSinTheta = 1.0f / sin(theta);
- scale = sin(theta * (1.0f - t)) * invSinTheta;
- invScale = sin(theta * t) * invSinTheta;
- } else {
- scale = 1.0f - t;
- invScale = t;
- }
- } else {
- rsQuaternionSet(&tempq1, tempq0.z, -tempq0.y, tempq0.x, -tempq0.w);
- scale = sin(M_PI * (0.5f - t));
- invScale = sin(M_PI * t);
- }
-
- rsQuaternionSet(q, tempq0.w*scale + tempq1.w*invScale, tempq0.x*scale + tempq1.x*invScale,
- tempq0.y*scale + tempq1.y*invScale, tempq0.z*scale + tempq1.z*invScale);
-}
-
-_RS_STATIC void rsQuaternionGetMatrixUnit(rs_matrix4x4 *m, const rs_quaternion *q) {
- float x2 = 2.0f * q->x * q->x;
- float y2 = 2.0f * q->y * q->y;
- float z2 = 2.0f * q->z * q->z;
- float xy = 2.0f * q->x * q->y;
- float wz = 2.0f * q->w * q->z;
- float xz = 2.0f * q->x * q->z;
- float wy = 2.0f * q->w * q->y;
- float wx = 2.0f * q->w * q->x;
- float yz = 2.0f * q->y * q->z;
-
- m->m[0] = 1.0f - y2 - z2;
- m->m[1] = xy - wz;
- m->m[2] = xz + wy;
- m->m[3] = 0.0f;
-
- m->m[4] = xy + wz;
- m->m[5] = 1.0f - x2 - z2;
- m->m[6] = yz - wx;
- m->m[7] = 0.0f;
-
- m->m[8] = xz - wy;
- m->m[9] = yz - wx;
- m->m[10] = 1.0f - x2 - y2;
- m->m[11] = 0.0f;
-
- m->m[12] = 0.0f;
- m->m[13] = 0.0f;
- m->m[14] = 0.0f;
- m->m[15] = 1.0f;
-}
+_RS_RUNTIME void __attribute__((overloadable))
+rsMatrixSet(rs_matrix2x2 *m, uint32_t row, uint32_t col, float v);
-/////////////////////////////////////////////////////
-// utility funcs
-/////////////////////////////////////////////////////
-__inline__ _RS_STATIC void __attribute__((overloadable, always_inline))
-rsExtractFrustumPlanes(const rs_matrix4x4 *modelViewProj,
- float4 *left, float4 *right,
- float4 *top, float4 *bottom,
- float4 *near, float4 *far) {
- // x y z w = a b c d in the plane equation
- left->x = modelViewProj->m[3] + modelViewProj->m[0];
- left->y = modelViewProj->m[7] + modelViewProj->m[4];
- left->z = modelViewProj->m[11] + modelViewProj->m[8];
- left->w = modelViewProj->m[15] + modelViewProj->m[12];
-
- right->x = modelViewProj->m[3] - modelViewProj->m[0];
- right->y = modelViewProj->m[7] - modelViewProj->m[4];
- right->z = modelViewProj->m[11] - modelViewProj->m[8];
- right->w = modelViewProj->m[15] - modelViewProj->m[12];
-
- top->x = modelViewProj->m[3] - modelViewProj->m[1];
- top->y = modelViewProj->m[7] - modelViewProj->m[5];
- top->z = modelViewProj->m[11] - modelViewProj->m[9];
- top->w = modelViewProj->m[15] - modelViewProj->m[13];
-
- bottom->x = modelViewProj->m[3] + modelViewProj->m[1];
- bottom->y = modelViewProj->m[7] + modelViewProj->m[5];
- bottom->z = modelViewProj->m[11] + modelViewProj->m[9];
- bottom->w = modelViewProj->m[15] + modelViewProj->m[13];
-
- near->x = modelViewProj->m[3] + modelViewProj->m[2];
- near->y = modelViewProj->m[7] + modelViewProj->m[6];
- near->z = modelViewProj->m[11] + modelViewProj->m[10];
- near->w = modelViewProj->m[15] + modelViewProj->m[14];
-
- far->x = modelViewProj->m[3] - modelViewProj->m[2];
- far->y = modelViewProj->m[7] - modelViewProj->m[6];
- far->z = modelViewProj->m[11] - modelViewProj->m[10];
- far->w = modelViewProj->m[15] - modelViewProj->m[14];
-
- float len = length(left->xyz);
- *left /= len;
- len = length(right->xyz);
- *right /= len;
- len = length(top->xyz);
- *top /= len;
- len = length(bottom->xyz);
- *bottom /= len;
- len = length(near->xyz);
- *near /= len;
- len = length(far->xyz);
- *far /= len;
-}
-
-__inline__ _RS_STATIC bool __attribute__((overloadable, always_inline))
-rsIsSphereInFrustum(float4 *sphere,
- float4 *left, float4 *right,
- float4 *top, float4 *bottom,
- float4 *near, float4 *far) {
-
- float distToCenter = dot(left->xyz, sphere->xyz) + left->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- distToCenter = dot(right->xyz, sphere->xyz) + right->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- distToCenter = dot(top->xyz, sphere->xyz) + top->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- distToCenter = dot(bottom->xyz, sphere->xyz) + bottom->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- distToCenter = dot(near->xyz, sphere->xyz) + near->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- distToCenter = dot(far->xyz, sphere->xyz) + far->w;
- if (distToCenter < -sphere->w) {
- return false;
- }
- return true;
-}
+_RS_RUNTIME float __attribute__((overloadable))
+rsMatrixGet(const rs_matrix2x2 *m, uint32_t row, uint32_t col);
+
+extern void __attribute__((overloadable)) rsMatrixLoadIdentity(rs_matrix4x4 *m);
+extern void __attribute__((overloadable)) rsMatrixLoadIdentity(rs_matrix3x3 *m);
+extern void __attribute__((overloadable)) rsMatrixLoadIdentity(rs_matrix2x2 *m);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix4x4 *m, const float *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix3x3 *m, const float *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix2x2 *m, const float *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix4x4 *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix3x3 *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix4x4 *m, const rs_matrix2x2 *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix3x3 *m, const rs_matrix3x3 *v);
+extern void __attribute__((overloadable)) rsMatrixLoad(rs_matrix2x2 *m, const rs_matrix2x2 *v);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadRotate(rs_matrix4x4 *m, float rot, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadScale(rs_matrix4x4 *m, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadTranslate(rs_matrix4x4 *m, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadMultiply(rs_matrix4x4 *m, const rs_matrix4x4 *lhs, const rs_matrix4x4 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix4x4 *m, const rs_matrix4x4 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadMultiply(rs_matrix3x3 *m, const rs_matrix3x3 *lhs, const rs_matrix3x3 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix3x3 *m, const rs_matrix3x3 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadMultiply(rs_matrix2x2 *m, const rs_matrix2x2 *lhs, const rs_matrix2x2 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix2x2 *m, const rs_matrix2x2 *rhs);
+
+extern void __attribute__((overloadable))
+rsMatrixRotate(rs_matrix4x4 *m, float rot, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixScale(rs_matrix4x4 *m, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixTranslate(rs_matrix4x4 *m, float x, float y, float z);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadOrtho(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far);
+extern void __attribute__((overloadable))
+rsMatrixLoadFrustum(rs_matrix4x4 *m, float left, float right, float bottom, float top, float near, float far);
+
+extern void __attribute__((overloadable))
+rsMatrixLoadPerspective(rs_matrix4x4* m, float fovy, float aspect, float near, float far);
+
+_RS_RUNTIME float4 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix4x4 *m, float4 in);
+
+_RS_RUNTIME float4 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix4x4 *m, float3 in);
+
+_RS_RUNTIME float4 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix4x4 *m, float2 in);
+
+_RS_RUNTIME float3 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix3x3 *m, float3 in);
+
+_RS_RUNTIME float3 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix3x3 *m, float2 in);
+
+_RS_RUNTIME float2 __attribute__((overloadable))
+rsMatrixMultiply(rs_matrix2x2 *m, float2 in);
+
+// Returns true if the matrix was successfully inversed
+extern bool __attribute__((overloadable)) rsMatrixInverse(rs_matrix4x4 *m);
+extern bool __attribute__((overloadable)) rsMatrixInverseTranspose(rs_matrix4x4 *m);
+extern void __attribute__((overloadable)) rsMatrixTranspose(rs_matrix4x4 *m);
+extern void __attribute__((overloadable)) rsMatrixTranspose(rs_matrix3x3 *m);
+extern void __attribute__((overloadable)) rsMatrixTranspose(rs_matrix2x2 *m);
/////////////////////////////////////////////////////
// int ops
/////////////////////////////////////////////////////
-__inline__ _RS_STATIC uint __attribute__((overloadable, always_inline)) rsClamp(uint amount, uint low, uint high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-__inline__ _RS_STATIC int __attribute__((overloadable, always_inline)) rsClamp(int amount, int low, int high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-__inline__ _RS_STATIC ushort __attribute__((overloadable, always_inline)) rsClamp(ushort amount, ushort low, ushort high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-__inline__ _RS_STATIC short __attribute__((overloadable, always_inline)) rsClamp(short amount, short low, short high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-__inline__ _RS_STATIC uchar __attribute__((overloadable, always_inline)) rsClamp(uchar amount, uchar low, uchar high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-__inline__ _RS_STATIC char __attribute__((overloadable, always_inline)) rsClamp(char amount, char low, char high) {
- return amount < low ? low : (amount > high ? high : amount);
-}
-
-#undef _RS_STATIC
+_RS_RUNTIME uint __attribute__((overloadable, always_inline)) rsClamp(uint amount, uint low, uint high);
+_RS_RUNTIME int __attribute__((overloadable, always_inline)) rsClamp(int amount, int low, int high);
+_RS_RUNTIME ushort __attribute__((overloadable, always_inline)) rsClamp(ushort amount, ushort low, ushort high);
+_RS_RUNTIME short __attribute__((overloadable, always_inline)) rsClamp(short amount, short low, short high);
+_RS_RUNTIME uchar __attribute__((overloadable, always_inline)) rsClamp(uchar amount, uchar low, uchar high);
+_RS_RUNTIME char __attribute__((overloadable, always_inline)) rsClamp(char amount, char low, char high);
-#endif
+#undef _RS_RUNTIME
+#endif
generated by cgit v1.1 at 2025-11-14 08:41:25 +0000