From: firstname.lastname@example.org (Tommy Thorn)
Subject: mmap & shared memory
Date: 14 Feb 92 14:43:02 GMT
Organization: DAIMI: Computer Science Department, Aarhus University, Denmark
What's status for shared memory? I'd suspect that
about the same interface in the kernel is needed
to provide both. I'm willing to dive into it,
but suspect others must be doing something similar
I need shared memory for Postgres. mmap is not
as needed, but shouldn't be that hard compared to
Comments anyone? Linus?
/Tommy Thorn, another LPF member. Terminate software patents.
Join The League for Programming Freedom, mail email@example.com.
Are *you* sure the *your* program isn't already covered by patents?
From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)
Subject: Re: mmap & shared memory
Date: 16 Feb 92 11:32:31 GMT
Organization: University of Helsinki
In article <1992Feb14.firstname.lastname@example.org> email@example.com
(Tommy Thorn) writes:
> What's status for shared memory? I'd suspect that
> about the same interface in the kernel is needed
> to provide both. I'm willing to dive into it,
> but suspect others must be doing something similar
> (shared libs.)
Well, shared libs aren't using shared memory (oh, they share pages if
they can, but basically the shared libs are just a "secondary
executable" loaded at the 60M mark in the process space). To my
knowledge nobody else is working on shared memory.
Shared memory should probably be relatively easy (from the page-sharing
point of view: you'd have to keep track of it all, of course): I did
think about it when I wrote the mm. The problem with this (and so many
other things) is that I've never used the shared memory syscalls, so I
don't feel I'm the right person to implement them: I wouldn't see a bug
if it hit me on the head with a sledgehammer, as I wouldn't know the