https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93492
--- Comment #22 from Iain Sandoe <iains at gcc dot gnu.org> --- I looked yesterday and, as of Xcode 11.5, the patchable_function_entry attribute is not recognised but the system tools. That means, for now, that there is no de-facto ABI for patchable functions on Mach-O/Darwin platforms (regardless of the fact that upstream clang accepts the attribute for Darwin targets). The "old" patchable function option (-mfix-and-continue) is recognised but is a NOP for modern Darwin. My plan is to skip the test(s) on Darwin - if there are enough test(s) then we should have a target-supports for the facility. I note in passing thet the target hook is not terribly friendly - the syntax error on Darwin is only because of the hard-wired assumption that named sections are ELF-named-sections. Which means that we will (when the ABI is settled) have to re-write the whole hook potentially for one line of code difference. In general, for a compiler like GCC that is supposed to support many targets, it would be a Good Idea is folks would not assume that everything is ELF.
