Re: Why shm fs (Was [2.3.51pre2] wrong shm_statfs in ipc/shm.c)

From: Jamie Lokier (lk@tantalophile.demon.co.uk)
Date: Fri Mar 10 2000 - 17:09:52 EST


Christoph Rohland wrote:
> Yes it is like tmpfs. E.g. Solaris implements posix shm with tmpfs
> like we do with shm fs.
>
> But shm fs is also missing many features like directories, read/write
> support for files, etc. So in the current state it is only usable for
> shared memory. But it could be extended later. I am not sure if it
> really makes sense. We have a very efficient cache and I once used a
> RAM disk for /tmp and did not see a performance increase.

I'm not sure it is worth doing either. If it is done, it should
probably a different fs that uses the same shared memory primitives as
shmfs -- so you don't have to include the directory etc. code when
you're only interested in shm.

The most important thing to come from this work is surely the cleanup of
shared memory in general.

-- Jamie

-
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/



This archive was generated by hypermail 2b29 : Wed Mar 15 2000 - 21:00:18 EST