Date: Mon, 1 Jul 1996 13:58:41 +0300 (EET DST) From: Jukka Antero Ukkonen <jau@aphrodite.funet.fi> To: bde@zeta.org.au (Bruce Evans) Cc: hackers@freebsd.org Subject: Re: POSIX.4 signals + other POSIX.4 stuff to FreeBSD... Message-ID: <199607011058.NAA24561@aphrodite.funet.fi> In-Reply-To: <199607010840.SAA18029@godzilla.zeta.org.au> from "Bruce Evans" at Jul 1, 96 06:40:26 pm
next in thread | previous in thread | raw e-mail | index | archive | help
Quoting Bruce Evans:
>
> >> >struct sigaction {
> >> > union {
> >> >....
> >> > void (*sa_handler) _P((int, ...));
>
> > then some" implementation. Those dots were put there just to keep
> > the compiler quiet when I needed the BSD style extended argument
> > list. They don't break anything really, but for reasons of strict
>
> They break the possibility of compiling (all) applications with -mrtd on
> i386's, and they break implementations where the equivalent of -mrtd or
> some other complicated (efficient) arg passing convention is standard.
How do you think one could compile a signal handler to assume
there is the size of the argument list saved in the stack if the
system does not use the same convention?
Or how would one do the exact opposite thing using the simpler ret
instruction when the size of the argument list has been saved in
the stack? (The stack would be trashed and -fcaller-saves would
certainly become a serious trouble.)
If both the system and the signal handlers have been compiled to
use the same calling convention the stack pointer will be modified
correctly anyway independent of whether the calling method expects
ret or rtd to be used.
It is the fact that the system and the signal handler both use the
same call/return method that counts, not whether the argument list
is variable or not. The whole business is about the caller saving
the correct size of the real parameter list in the stack when it
is needed, not about the return from the called function becoming
somehow different when the size of the argument list has been
claimed variable in C. The prototype affects the caller.
A system using rtd return convention would be expected to save the
correct size of the argument list in the stack anyway. Right?
BTW - As far as I know -mrtd has no meaning on iX86 machines. It is
a valid option only for 680X0 family of processors.
> The kernel shouldn't use the user macros because it would become slower
> if the the user macros were fixed to actually meet their specification.
Kernel could use it's own variations. The only thing that would
be affected were that it were easier to change the real data type
that is used represent sigset_t.
Cheers,
// jau
------
/ Jukka A. Ukkonen, FUNET / Centre for Scientific Computing
/__ M.Sc. (sw-eng & cs) Tel: (Home) +358-0-6215280
/ Internet: ukkonen@csc.fi (Work) +358-0-4573208
v Internet: jau@funet.fi (Mobile) +358-400-606671
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199607011058.NAA24561>