> Perhaps, but it would defeat the whole purpose of switching to > init_array (see my previous mail about why any call in .init/.fini > causes incorrect unwind info). Since this part of the section is generated normally along with the prologue, I would think that its unwind info would come out normally too.