http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59600
--- Comment #7 from Yury Gribov <y.gribov at samsung dot com> --- (In reply to Andrew Pinski from comment #6) > Move the check to before the check of the target > attribute table. My bad. > And second you should compare the current function > attribute to fndecl attribute. You mean compare caller's and callee's attributes? Makes a lot of sense but I don't think this info is available at the time function_attribute_inlinable_p is called: (gdb) b function_attribute_inlinable_p (gdb) r (gdb) call debug_generic_stmt(fndecl) mark_maybe_pointer (gdb) call debug_generic_stmt(current_function_decl) mark_maybe_pointer Perhaps this check should be moved to e.g. expand_call_inline?
