Re: mremap() proposal (was Re: malloc and joe)
Snow Cat (snowcat@math.csufresno.edu)
Sun, 24 Mar 1996 19:01:28 -0800 (PST)
Jeremy Fitzhardinge once wrote:
>
> Linus Torvalds wrote:
> >
> > Ok, here is a draft implementation of the "mremap()" system call, Wolfram could
> > you test?
> >
> > Essentially, it adds a new system call (#163 on x86) that looks like
> >
> > caddr_t mremap(caddr_t addr, size_t old, size_t new, int move);
>
> Urk. Surely a mmap of /proc/self/mem is a better way to go - why invent a new
> interface when you can generalise an existing one?
>
> I guess there would be interesting implications if you can mmap other process's
> address spaces; would that be hard (it would certainly be useful)?
>
> J
It might be convinient if malloc() doesn't have to use files. Imagine a user process
setting RLIMIT_NOFILE to 3 and then trying to allocate some memory.
--
Snow ^oo^ Cat <snowcat@math.CSUFresno.EDU>
_ -> <- aka Oleg Kibirev <oleg@math.CSUFresno.EDU>
___(_) _ _)_
/ _) finger oleg@math.math.CSUFresno.EDU for PGP public key
\_.-._
|___/ Purr!