Hi Martin, > There is a similar mechanism for pointer-to-member-functions > used by C++. Is this correct on aarch64?
/* By default, the C++ compiler will use the lowest bit of the pointer to function to indicate a pointer-to-member-function points to a virtual member function. However, if FUNCTION_BOUNDARY indicates function addresses aren't always even, the lowest bit of the delta field will be used. */ #ifndef TARGET_PTRMEMFUNC_VBIT_LOCATION #define TARGET_PTRMEMFUNC_VBIT_LOCATION \ (FUNCTION_BOUNDARY >= 2 * BITS_PER_UNIT \ ? ptrmemfunc_vbit_in_pfn : ptrmemfunc_vbit_in_delta) #endif AArch64 overrides this like Arm to ptrmemfunc_vbit_in_delta. But yes that's yet another example of GCC using the wrong default. The assumption that FUNCTION_BOUNDARY implies no use of low bits is simply false - a generic default should be safe for all targets (and if that's not possible, there should not be a default). Cheers, Wilco
