b3Solver -> pass pointer to source instead of 0 (was left over from a debugging session), thanks to David for the report

Break up clipHullHullConcaveConvexKernel into multiple stages, so it might 'fit' in Apple's OpenCL implementation
Implemented bvhTraversalKernel and findConcaveSeparatingAxis on CPU (debugging, possible future CPU version)

src/Bullet3OpenCL/NarrowphaseCollision/kernels/primitiveContacts.h

@@ -13,6 +13,7 @@ static const char* primitiveContactsKernelsCL= \
 "#ifdef __cplusplus\n"
 "#else\n"
 "#define b3AtomicInc atomic_inc\n"
+"#define b3AtomicAdd atomic_add\n"
 "#define b3Fabs fabs\n"
 "#define b3Sqrt native_sqrt\n"
 "#define b3Sin native_sin\n"
@@ -36,6 +37,9 @@ static const char* primitiveContactsKernelsCL= \
 "		float4 b1 = b3MakeFloat4(v1.xyz,0.f);\n"
 "		return cross(a1, b1);\n"
 "	}\n"
+"	#define b3MinFloat4 min\n"
+"	#define b3MaxFloat4 max\n"
+"	#define b3Normalized(a) normalize(a)\n"
 "#endif \n"
 "		\n"
 "inline bool b3IsAlmostZero(b3Float4ConstArg v)\n"