Re: kerneld/request-route

Bo Johansson (bo.johansson@mbox2.swipnet.se)
04 Jul 1996 01:26:57 +0200


>>>>> "Eric" == Eric Schenk <schenk@cs.toronto.edu> writes:

Eric> Alan Cox <alan@cymru.net> writes:
>> Bo Johansson <bo.johansson@mbox2.swipnet.se> writes:
>>> But to use diald, if you are using PPP then you have to add SLIP
>>> to your kernel and run yet another daemon and still run kerneld
>>> (if you want to load modules automatically), am I right?.
>>> That doesn't seem like a good solution to me.
>>
>> So teach diald to also know how to watch for PPP over a tty/pty pair

Eric> As the sentiment expressed by Bo Johansson: I'm curious, are you objecting
Eric> more to the idea of loading an extra modul into the kernel (SLIP),
Eric> or to having an extra user space process?
Eric> If it's the former then note that the proposed fixes to make kerneld
Eric> actually work properly amount to requiring you to have the dummy
Eric> module loaded.
Eric> If it's the later, then why aren't you also objecting that kerneld
Eric> doesn't perform, say, the functions of inetd, which also starts up
Eric> user space processes on demand?

To say that I object is a bit much, I just think that the fewer daemons and
the less memory you can get it to run with the better. I certainly don't have
the knowledge to suggest a different solution, it's just that I think it was
a valid argument well worth consideration. As for inetd, I don't think
that all daemons are necessarily evil, it's just that every new daemon you start
up chews up another bit of your precious memory (well perhaps not that precious
any more, since the price of memory has dropped quite a lot recently).

Bosse