> setuid(2) differs from the OpenBSD setuid(2)
> Either I am non compos or the thing is very wrong.
> The docs (man-pages-1.35) say ...

Yes, setuid() has a behaviour that varies a bit from system to system.
Moreover, it has varied in the history of Linux. The manpage may have
been correct when it was written, but it is not today, and I just
fixed it.

ERRORS
       EPERM  The  user  is  not the super-user, and uid does not
              match the real or saved user ID of the calling pro­
              cess.

Our norm is the coming POSIX standard, roughly the Austin 7 draft,
which again is based on the SUSv2. According to this, the current
kernel code is correct here.

Andries