[openbeosnetteam] Re: Kernel move...

OK, you've still failed to convince me. why would we have to re release the
kernel?

Let's just restate what we're talking about here.

The call to socket creates a socket. That's it. the code to do this has NO
dependency on any protocols. Calling socket with no protocol modules simply
gives an error. Adding a new protocol module gives us the ability to call
socket for that protocol. Again - the code for socket never get changed.

I guess this is where I'm not understanding your argument. The socket call
has been part of the kernel on unix for almost as long as sockets have
existed and it's inclusion makes perfect sense when you think of the low
level tinkering it has access to. Look at the (no offence meant here) nasty
code that Philippe had to add for accept in libnet. Having accept in the
kernel removes that mess and makes life much more straightforward.

/dev/net/socket|stack was a hack that got us working, nothing more. We'll
make the move as it makes sense. Now we have our own kernel code the kernel
should no longer be viewed as a "back box" (yes I know only too well they're
orange) but as a set of code we can modify to our needs. Let's take
advantage of that.

david

----- Original Message -----
From: "Nathan Whitehorn" <nathan.whitehorn@xxxxxxxxxxx>
To: <openbeosnetteam@xxxxxxxxxxxxx>
Sent: Saturday, June 01, 2002 1:50 AM
Subject: [openbeosnetteam] Re: Kernel move...


> > Why do you think they don't need a syscall? Why is the driver more
> > extensible?
>
> From the point of view of the kernel, adding special cases for anything
> is usually a bad idea. Also, if we were to want to update that, and it
> was compiled into the kernel, we would have to re-release the kernel
> too, which would be unfortunate. Finally, there really isn't any
> advantage to a syscall: open() for socket() and ioctl() for the rest
> really work just as well, without any real speed problems.
> -Nathan
>
> --
> Fortune Cookie Says:
>
> One seldom sees a monument to a committee.
>
>
>

Other related posts: