Re: which stack direction?

Michael Meissner (meissner@cygnus.com)
Fri, 12 Mar 1999 22:55:18 -0500

On Fri, Mar 12, 1999 at 11:45:36AM -0800, Larry McVoy wrote:
> I know of a new high end machine, under gentleman's NDA, which might be
> running Linux some day. The compiler people for this machine want to
> know if the Linux folks care which way the stack grows. The exact question
> is:
>
> "Does it matter to Linux whether the stacks in user codes run up
> (pointer starts at a low address and increments to get more memory)
> or down (pointer starts at a high address and decrements to get
> more memory)? If it does matter, which is preferred?"
>
> More details (some stuff removed to protect the guilty :-)

Given that HP has at least announced plans for Linux on their boxes, if there
is code in Linux that cares about the stack growth (other than the usual kernal
machine dependent files and compiler varargs support), it probably soon will be
identified (PA-Riscs have stacks that go in the opposite direction to most
machines these days). 10 years ago this was a big problem due to the fact that
varargs/stdarg wasn't established practice, and 'real men' wrote their own
variable argument handling (I had been worked on the Data General MV/Eclipse C
compiler for a couple of years, and it was another machine with the stack going
the oppisite direction).

Note, having the stack grow in the oppisite direction, means that you can't
have the stack and the heap grow towards each other. Instead one or both areas
must be fixed in size.

-- 
Michael Meissner, Cygnus Solutions
PO Box 98, Ayer Massachusetts, USA 01432-0098
meissner@cygnus.com


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/