(In reply to Miroslav Beneš from comment #12) > Having said that the support for noreturn functions in objtool is not great. > There have been talks about improving it but it needs to happen in > cooperation with GCC because there is currently no way to know that GCC > decided to place unreachable somewhere (decided that a function is > noreturn). I do not know if there is a progress on this front behind the > scenes. Once we start building our enterprise kernels with GCC 13+, asking our performance team to evaluate the option -funreachable-traps might be worthwhile.