On Fri, Apr 14, 2023 at 8:33 PM Samuel Thibault <samuel.thiba...@gnu.org> wrote:
> Applied with fixing the __mig_dealloc_reply_port(NULL) cases, thanks!
By the way: that __mig_dealloc_reply_port () inside
_dl_sysdep_start_cleanup () is not doing what the author of that code
wanted it to do. It deallocates the current reply port, but while
doing that, it creates a fresh one in its place. So we just waste an
RPC (and a syscall). We should just not do that.
It would be nice to not deallocate __mach_{task,host}_self_ too, and
instead migrate them into the new libc.so slots. Again, we'd save some
RPCs and syscalls without really losing anything.
Besides, are we 100% sure that after initially entering libc.so, ld.so
will never do any Mach things (RPCs) anymore? We override __open &
friends, but it might still be using some smaller Mach things, no? If
yes, imagine what would happen when it keeps using its
__mach_{task,host}_self_, which it has deallocated, but not even reset
to MACH_PORT_NULL.
Sergey
