Re: [PATCH] /proc/sysvipc

Riccardo Facchetti (fizban@tin.it)
Sun, 26 Jul 1998 00:53:22 +0200 (MET DST)


On Sat, 25 Jul 1998, Brandon S. Allbery KF8NH wrote:

> Let's put it this way: the documented SYSVIPC interface provides
> SYSV-compatible access to IPC information. /proc is the Linux-native way to
> get at such things, so /proc access should be supported on Linux.
>
> I'll also note that we have both sysctl and /proc/sys; is /proc/sys, then,
> redundant and "not very useful"?

Let's suppose that we have an user space program that knows _every_ and
_all_ sysctl parameters. In this case (and _only_ in this case), the
/proc/sys interface to sysctl IMHO is not very useful. But we live in a
real world and it is _very_ likely that an user space program will be
outdated in few future kernels: /proc/sys is useful to have access to new
parameters that an old user space application don't even know. And more:
some parameters in /proc/sys are read/write and control some vital kernel
parameters so it is very useful to have a simple interface like /proc/sys
to tune the kernel behaviour with a 'cat > /proc/sys/foobar'.

On the other hand IPC have a well defined interface that is not likely to
be changed in the future. A readonly /proc/ipc interface just to read ipc
information is IMHO not very useful.

But I don't want to start a thread on this subject. The first post was
just to tell Dragos my opinion. In fact he have implemented the /proc
interface to SYSVIPC and he can ask Linus to include it into the main
kernel source tree: Linus have the final control on what will make its way
into the kernel and what will stay out.

Ciao,
Riccardo.

-
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.altern.org/andrebalsa/doc/lkml-faq.html