[OpenCL] Fix atomic_fetch_add/sub half overloads

Some of the atomic_fetch_add and atomic_fetch_sub overloads intended
for atomic_half types accidentally had an atomic_float parameter.

Author: svenvh

clang/lib/Sema/OpenCLBuiltins.td

@@ -1185,7 +1185,7 @@ let MinVersion = CL20 in {
       defvar extension_fp64 = !cast<FunctionExtension>("FuncExtFloatAtomicsFp64" # addrspace # "Add");
 
       defm : BuiltinAtomicExplicit<"atomic_fetch_" # ModOp,
-          [Half, PointerType<VolatileType<AtomicFloat>, addrspace>, Half], extension_fp16>;
+          [Half, PointerType<VolatileType<AtomicHalf>, addrspace>, Half], extension_fp16>;
       defm : BuiltinAtomicExplicit<"atomic_fetch_" # ModOp,
           [Float, PointerType<VolatileType<AtomicFloat>, addrspace>, Float], extension_fp32>;
       defm : BuiltinAtomicExplicit<"atomic_fetch_" # ModOp,