AOSCOS: Remove invalid 'd' suffix for double literals

They are causing compilation errors on my Clang setup.
It seems to work with GCC as a GCC extension. No matter how compilers
behave, it is still a GCC extension, not a standard.
According to C++ 14 specification, "2.14.4 Floating literals", 'd' is
not a floating-suffix. Having no suffix
Having no suffix makes a constant of double, which is expected here, so
it can be removed safely.

Signed-off-by: Bingwu Zhang <[email protected]>

Author: xtexx

src/hotspot/cpu/loongarch/stubRoutines_loongarch_64.cpp

@@ -191,5 +191,5 @@ ATTRIBUTE_ALIGNED(64) jfloat StubRoutines::la::_round_float_imm[] = {
 };
 
 ATTRIBUTE_ALIGNED(64) jdouble StubRoutines::la::_round_double_imm[] = {
-  -0.5d, 0.49999999999999994d // magic number for ties
+  -0.5, 0.49999999999999994 // magic number for ties
 };