Cleanup 64bit and remove deprecated functions

Change-Id: I8f4570af8ab7da05831258dd264118b28ce5d352

12 files changed, 36 insertions, 244 deletions

diff --git a/driver/runtime/Android.mk b/driver/runtime/Android.mk
index c16b1f2b..5f5a7081 100755
--- a/driver/runtime/Android.mk
+++ b/driver/runtime/Android.mk
@@ -28,8 +28,7 @@ clcore_base_files := \
     rs_sample.c \
     rs_sampler.c \
     rs_convert.c \
-    allocation.ll \
-    rsClamp.ll
+    allocation.ll
 
 clcore_files := \
     $(clcore_base_files) \
diff --git a/driver/runtime/matrix.ll b/driver/runtime/matrix.ll
deleted file mode 100644
index c56405d1..00000000
--- a/driver/runtime/matrix.ll
+++ /dev/null
@@ -1,176 +0,0 @@
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:64:128-a0:0:64-n32-S64"
-target triple = "armv7-none-linux-gnueabi"
-
-
-%struct.rs_matrix4x4 = type { [16 x float] }
-%struct.rs_matrix3x3 = type { [9 x float] }
-%struct.rs_matrix2x2 = type { [4 x float] }
-
-define internal <4 x float> @smear_f(float %in) nounwind readnone alwaysinline {
-  %1 = insertelement <4 x float> undef, float %in, i32 0
-  %2 = insertelement <4 x float> %1, float %in, i32 1
-  %3 = insertelement <4 x float> %2, float %in, i32 2
-  %4 = insertelement <4 x float> %3, float %in, i32 3
-  ret <4 x float> %4
-}
-
-
-define <3 x float> @_Z16rsMatrixMultiplyPK12rs_matrix3x3Dv3_f(%struct.rs_matrix3x3* nocapture %m, <3 x float> %in) nounwind readonly {
-  %x0 = extractelement <3 x float> %in, i32 0
-  %x = tail call <4 x float> @smear_f(float %x0) nounwind readnone
-  %y0 = extractelement <3 x float> %in, i32 1
-  %y = tail call <4 x float> @smear_f(float %y0) nounwind readnone
-  %z0 = extractelement <3 x float> %in, i32 2
-  %z = tail call <4 x float> @smear_f(float %z0) nounwind readnone
-
-  %px = getelementptr inbounds %struct.rs_matrix3x3* %m, i32 0, i32 0, i32 0
-  %px2 = bitcast float* %px to <4 x float>*
-  %xm = load <4 x float>* %px2, align 4
-  %py = getelementptr inbounds %struct.rs_matrix3x3* %m, i32 0, i32 0, i32 3
-  %py2 = bitcast float* %py to <4 x float>*
-  %ym = load <4 x float>* %py2, align 4
-  %pz = getelementptr inbounds %struct.rs_matrix3x3* %m, i32 0, i32 0, i32 6
-  %pz2 = bitcast float* %pz to <3 x float>*
-  %zm2 = load <3 x float>* %pz2, align 4
-  %zm = shufflevector <3 x float> %zm2, <3 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-
-  %a1 = fmul <4 x float> %x, %xm
-  %a2 = fmul <4 x float> %y, %ym
-  %a3 = fadd <4 x float> %a1, %a2
-  %a4 = fmul <4 x float> %z, %zm
-  %a5 = fadd <4 x float> %a4, %a3
-  %a6 = shufflevector <4 x float> %a5, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
-  ret <3 x float> %a6
-}
-
-define <3 x float> @_Z16rsMatrixMultiplyP12rs_matrix3x3Dv3_f(%struct.rs_matrix3x3* nocapture %m, <3 x float> %in) nounwind readonly {
-  %r = tail call <3 x float> @_Z16rsMatrixMultiplyPK12rs_matrix3x3Dv3_f(%struct.rs_matrix3x3* nocapture %m, <3 x float> %in) nounwind
-  ret <3 x float> %r
-}
-
-define <3 x float> @_Z16rsMatrixMultiplyPK12rs_matrix3x3Dv2_f(%struct.rs_matrix3x3* nocapture %m, <2 x float> %in) nounwind readonly {
-  %x0 = extractelement <2 x float> %in, i32 0
-  %x = tail call <4 x float> @smear_f(float %x0) nounwind readnone
-  %y0 = extractelement <2 x float> %in, i32 1
-  %y = tail call <4 x float> @smear_f(float %y0) nounwind readnone
-
-  %px = getelementptr inbounds %struct.rs_matrix3x3* %m, i32 0, i32 0, i32 0
-  %px2 = bitcast float* %px to <4 x float>*
-  %xm = load <4 x float>* %px2, align 4
-  %py = getelementptr inbounds %struct.rs_matrix3x3* %m, i32 0, i32 0, i32 3
-  %py2 = bitcast float* %py to <4 x float>*
-  %ym = load <4 x float>* %py2, align 4
-
-  %a1 = fmul <4 x float> %x, %xm
-  %a2 = fmul <4 x float> %y, %ym
-  %a3 = fadd <4 x float> %a1, %a2
-  %a4 = shufflevector <4 x float> %a3, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
-  ret <3 x float> %a4
-}
-
-define <3 x float> @_Z16rsMatrixMultiplyP12rs_matrix3x3Dv2_f(%struct.rs_matrix3x3* nocapture %m, <2 x float> %in) nounwind readonly {
-  %r = tail call <3 x float> @_Z16rsMatrixMultiplyPK12rs_matrix3x3Dv2_f(%struct.rs_matrix3x3* nocapture %m, <2 x float> %in) nounwind
-  ret <3 x float> %r
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv4_f(%struct.rs_matrix4x4* nocapture %m, <4 x float> %in) nounwind readonly {
-  %x0 = extractelement <4 x float> %in, i32 0
-  %x = tail call <4 x float> @smear_f(float %x0) nounwind readnone
-  %y0 = extractelement <4 x float> %in, i32 1
-  %y = tail call <4 x float> @smear_f(float %y0) nounwind readnone
-  %z0 = extractelement <4 x float> %in, i32 2
-  %z = tail call <4 x float> @smear_f(float %z0) nounwind readnone
-  %w0 = extractelement <4 x float> %in, i32 3
-  %w = tail call <4 x float> @smear_f(float %w0) nounwind readnone
-
-  %px = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 0
-  %px2 = bitcast float* %px to <4 x float>*
-  %xm = load <4 x float>* %px2, align 4
-  %py = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 4
-  %py2 = bitcast float* %py to <4 x float>*
-  %ym = load <4 x float>* %py2, align 4
-  %pz = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 8
-  %pz2 = bitcast float* %pz to <4 x float>*
-  %zm = load <4 x float>* %pz2, align 4
-  %pw = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 12
-  %pw2 = bitcast float* %pw to <4 x float>*
-  %wm = load <4 x float>* %pw2, align 4
-
-  %a1 = fmul <4 x float> %x, %xm
-  %a2 = fmul <4 x float> %y, %ym
-  %a3 = fadd <4 x float> %a1, %a2
-  %a4 = fmul <4 x float> %z, %zm
-  %a5 = fadd <4 x float> %a3, %a4
-  %a6 = fmul <4 x float> %w, %wm
-  %a7 = fadd <4 x float> %a5, %a6
-  ret <4 x float> %a7
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyP12rs_matrix4x4Dv4_f(%struct.rs_matrix4x4* nocapture %m, <4 x float> %in) nounwind readonly {
-  %r = tail call <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv4_f(%struct.rs_matrix4x4* nocapture %m, <4 x float> %in) nounwind
-  ret <4 x float> %r
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv3_f(%struct.rs_matrix4x4* nocapture %m, <3 x float> %in) nounwind readonly {
-  %x0 = extractelement <3 x float> %in, i32 0
-  %x = tail call <4 x float> @smear_f(float %x0) nounwind readnone
-  %y0 = extractelement <3 x float> %in, i32 1
-  %y = tail call <4 x float> @smear_f(float %y0) nounwind readnone
-  %z0 = extractelement <3 x float> %in, i32 2
-  %z = tail call <4 x float> @smear_f(float %z0) nounwind readnone
-
-  %px = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 0
-  %px2 = bitcast float* %px to <4 x float>*
-  %xm = load <4 x float>* %px2, align 4
-  %py = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 4
-  %py2 = bitcast float* %py to <4 x float>*
-  %ym = load <4 x float>* %py2, align 4
-  %pz = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 8
-  %pz2 = bitcast float* %pz to <4 x float>*
-  %zm = load <4 x float>* %pz2, align 4
-  %pw = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 12
-  %pw2 = bitcast float* %pw to <4 x float>*
-  %wm = load <4 x float>* %pw2, align 4
-
-  %a1 = fmul <4 x float> %x, %xm
-  %a2 = fadd <4 x float> %wm, %a1
-  %a3 = fmul <4 x float> %y, %ym
-  %a4 = fadd <4 x float> %a2, %a3
-  %a5 = fmul <4 x float> %z, %zm
-  %a6 = fadd <4 x float> %a4, %a5
-  ret <4 x float> %a6
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyP12rs_matrix4x4Dv3_f(%struct.rs_matrix4x4* nocapture %m, <3 x float> %in) nounwind readonly {
-  %r = tail call <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv3_f(%struct.rs_matrix4x4* nocapture %m, <3 x float> %in) nounwind
-  ret <4 x float> %r
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv2_f(%struct.rs_matrix4x4* nocapture %m, <2 x float> %in) nounwind readonly {
-  %x0 = extractelement <2 x float> %in, i32 0
-  %x = tail call <4 x float> @smear_f(float %x0) nounwind readnone
-  %y0 = extractelement <2 x float> %in, i32 1
-  %y = tail call <4 x float> @smear_f(float %y0) nounwind readnone
-
-  %px = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 0
-  %px2 = bitcast float* %px to <4 x float>*
-  %xm = load <4 x float>* %px2, align 4
-  %py = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 4
-  %py2 = bitcast float* %py to <4 x float>*
-  %ym = load <4 x float>* %py2, align 4
-  %pw = getelementptr inbounds %struct.rs_matrix4x4* %m, i32 0, i32 0, i32 12
-  %pw2 = bitcast float* %pw to <4 x float>*
-  %wm = load <4 x float>* %pw2, align 4
-
-  %a1 = fmul <4 x float> %x, %xm
-  %a2 = fadd <4 x float> %wm, %a1
-  %a3 = fmul <4 x float> %y, %ym
-  %a4 = fadd <4 x float> %a2, %a3
-  ret <4 x float> %a4
-}
-
-define <4 x float> @_Z16rsMatrixMultiplyP12rs_matrix4x4Dv2_f(%struct.rs_matrix4x4* nocapture %m, <2 x float> %in) nounwind readonly {
-  %r = tail call <4 x float> @_Z16rsMatrixMultiplyPK12rs_matrix4x4Dv2_f(%struct.rs_matrix4x4* nocapture %m, <2 x float> %in) nounwind
-  ret <4 x float> %r
-}
-
diff --git a/driver/runtime/rsClamp.ll b/driver/runtime/rsClamp.ll
deleted file mode 100644
index eba678a0..00000000
--- a/driver/runtime/rsClamp.ll
+++ /dev/null
@@ -1,60 +0,0 @@
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:64:128-a0:0:64-n32-S64"
-target triple = "armv7-none-linux-gnueabi"
-
-
-define float @_Z7rsClampfff(float %value, float %low, float %high) nounwind readonly {
-  %1 = fcmp olt float %value, %high
-  %2 = select i1 %1, float %value, float %high
-  %3 = fcmp ogt float %2, %low
-  %4 = select i1 %3, float %2, float %low
-  ret float %4
-}
-
-define signext i8 @_Z7rsClampccc(i8 signext %value, i8 signext %low, i8 signext %high) nounwind readonly {
-  %1 = icmp slt i8 %value, %high
-  %2 = select i1 %1, i8 %value, i8 %high
-  %3 = icmp sgt i8 %2, %low
-  %4 = select i1 %3, i8 %2, i8 %low
-  ret i8 %4
-}
-
-define zeroext i8 @_Z7rsClamphhh(i8 zeroext %value, i8 zeroext %low, i8 zeroext %high) nounwind readonly {
-  %1 = icmp ult i8 %value, %high
-  %2 = select i1 %1, i8 %value, i8 %high
-  %3 = icmp ugt i8 %2, %low
-  %4 = select i1 %3, i8 %2, i8 %low
-  ret i8 %4
-}
-
-define signext i16 @_Z7rsClampsss(i16 signext %value, i16 signext %low, i16 signext %high) nounwind readonly {
-  %1 = icmp slt i16 %value, %high
-  %2 = select i1 %1, i16 %value, i16 %high
-  %3 = icmp sgt i16 %2, %low
-  %4 = select i1 %3, i16 %2, i16 %low
-  ret i16 %4
-}
-
-define zeroext i16 @_Z7rsClampttt(i16 zeroext %value, i16 zeroext %low, i16 zeroext %high) nounwind readonly {
-  %1 = icmp ult i16 %value, %high
-  %2 = select i1 %1, i16 %value, i16 %high
-  %3 = icmp ugt i16 %2, %low
-  %4 = select i1 %3, i16 %2, i16 %low
-  ret i16 %4
-}
-
-define i32 @_Z7rsClampiii(i32 %value, i32 %low, i32 %high) nounwind readonly {
-  %1 = icmp slt i32 %value, %high
-  %2 = select i1 %1, i32 %value, i32 %high
-  %3 = icmp sgt i32 %2, %low
-  %4 = select i1 %3, i32 %2, i32 %low
-  ret i32 %4
-}
-
-define i32 @_Z7rsClampjjj(i32 %value, i32 %low, i32 %high) nounwind readonly {
-  %1 = icmp ult i32 %value, %high
-  %2 = select i1 %1, i32 %value, i32 %high
-  %3 = icmp ugt i32 %2, %low
-  %4 = select i1 %3, i32 %2, i32 %low
-  ret i32 %4
-}
-
diff --git a/driver/runtime/rs_allocation.c b/driver/runtime/rs_allocation.c
index 0722680d..bf377138 100644
--- a/driver/runtime/rs_allocation.c
+++ b/driver/runtime/rs_allocation.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 // Opaque Allocation type operations
diff --git a/driver/runtime/rs_core.c b/driver/runtime/rs_core.c
index b1c2af1e..f908e9ea 100644
--- a/driver/runtime/rs_core.c
+++ b/driver/runtime/rs_core.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 /* Function declarations from libRS */
@@ -13,6 +12,28 @@ extern float4 rsUnpackColor8888(uchar4 c)
 }
 
 
+extern float __attribute__((overloadable)) rsClamp(float v, float l, float h) {
+    return clamp(v, l, h);
+}
+extern char __attribute__((overloadable)) rsClamp(char v, char l, char h) {
+    return clamp(v, l, h);
+}
+extern uchar __attribute__((overloadable)) rsClamp(uchar v, uchar l, uchar h) {
+    return clamp(v, l, h);
+}
+extern short __attribute__((overloadable)) rsClamp(short v, short l, short h) {
+    return clamp(v, l, h);
+}
+extern ushort __attribute__((overloadable)) rsClamp(ushort v, ushort l, ushort h) {
+    return clamp(v, l, h);
+}
+extern int __attribute__((overloadable)) rsClamp(int v, int l, int h) {
+    return clamp(v, l, h);
+}
+extern uint __attribute__((overloadable)) rsClamp(uint v, uint l, uint h) {
+    return clamp(v, l, h);
+}
+
 extern int32_t __attribute__((overloadable)) rsAtomicCas(volatile int32_t *ptr, int32_t expectedValue, int32_t newValue) {
     return __sync_val_compare_and_swap(ptr, expectedValue, newValue);
 }
diff --git a/driver/runtime/rs_element.c b/driver/runtime/rs_element.c
index 4db5883c..a67bdd5c 100644
--- a/driver/runtime/rs_element.c
+++ b/driver/runtime/rs_element.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 /**
diff --git a/driver/runtime/rs_matrix.c b/driver/runtime/rs_matrix.c
index 3afccc18..052e829a 100644
--- a/driver/runtime/rs_matrix.c
+++ b/driver/runtime/rs_matrix.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 /* Function declarations from libRS */
diff --git a/driver/runtime/rs_mesh.c b/driver/runtime/rs_mesh.c
index bb533bc4..283b448d 100644
--- a/driver/runtime/rs_mesh.c
+++ b/driver/runtime/rs_mesh.c
@@ -1,3 +1,5 @@
+#ifndef __LP64__
+
 #include "rs_core.rsh"
 #include "rs_graphics.rsh"
 #include "rs_structs.h"
@@ -53,3 +55,6 @@ extern rs_primitive __attribute__((overloadable))
     }
     return mesh->mHal.state.primitives[index];
 }
+
+#endif
+
diff --git a/driver/runtime/rs_program.c b/driver/runtime/rs_program.c
index 64c656f8..320c5a4e 100644
--- a/driver/runtime/rs_program.c
+++ b/driver/runtime/rs_program.c
@@ -1,3 +1,5 @@
+#ifndef __LP64__
+
 #include "rs_core.rsh"
 #include "rs_graphics.rsh"
 #include "rs_structs.h"
@@ -106,3 +108,5 @@ extern rs_cull_mode __attribute__((overloadable))
     }
     return prog->mHal.state.cull;
 }
+
+#endif
diff --git a/driver/runtime/rs_sample.c b/driver/runtime/rs_sample.c
index 2cd5bdc0..f4d52f72 100644
--- a/driver/runtime/rs_sample.c
+++ b/driver/runtime/rs_sample.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 
diff --git a/driver/runtime/rs_sampler.c b/driver/runtime/rs_sampler.c
index 39782de6..6c3e7ffe 100644
--- a/driver/runtime/rs_sampler.c
+++ b/driver/runtime/rs_sampler.c
@@ -1,5 +1,4 @@
 #include "rs_core.rsh"
-#include "rs_graphics.rsh"
 #include "rs_structs.h"
 
 /**
diff --git a/driver/runtime/rs_structs.h b/driver/runtime/rs_structs.h
index 26a41eed..80c129cc 100644
--- a/driver/runtime/rs_structs.h
+++ b/driver/runtime/rs_structs.h
@@ -63,6 +63,7 @@ typedef struct Allocation {
     } mHal;
 } Allocation_t;
 
+#ifndef __LP64__
 /*****************************************************************************
  * CAUTION
  *
@@ -124,6 +125,7 @@ typedef struct ProgramRaster {
         } state;
     } mHal;
 } ProgramRaster_t;
+#endif //__LP64__
 
 /*****************************************************************************
  * CAUTION
@@ -231,6 +233,7 @@ typedef struct Type {
     } mHal;
 } Type_t;
 
+#ifndef __LP64__
 /*****************************************************************************
  * CAUTION
  *
@@ -264,4 +267,5 @@ typedef struct Mesh {
         } state;
     } mHal;
 } Mesh_t;
+#endif //__LP64__
 #endif // _RS_CORE_H_