old-www/LDP/nag/node45.html

145 lines
5.9 KiB
HTML

<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<!--Converted with LaTeX2HTML 96.1-c (Feb 29, 1996) by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds -->
<HTML>
<HEAD>
<TITLE>Kernel Options in 1.1.14 and Higher</TITLE>
</HEAD>
<BODY LANG="EN">
<A HREF="node1.html"><IMG WIDTH=65 HEIGHT=24 ALIGN=BOTTOM ALT="contents" SRC="contents_motif.gif"></A> <BR>
<B> Next:</B> <A HREF="node46.html">A Tour of Network </A>
<B>Up:</B> <A HREF="node43.html">Kernel Configuration</A>
<B> Previous:</B> <A HREF="node44.html">Kernel Options in 1.0 </A>
<BR> <P>
<H2><A NAME="SECTION005220000">Kernel Options in 1.1.14 and Higher</A></H2>
<P>
Starting with 1.1.14, which added alpha support for IPX,
the configuration procedure changed slightly. The general options section
now asks whether you want networking support in general. It is
immediately followed by a couple of question on miscellaneous
networking options.
<P>
<P><P>
<P>
To use TCP/IP networking, you must answer this question with
y. If you answer with n, however, you will still
be able to compile the kernel with IPX support.
<P>
<P><P>
<P>
<A NAME="1955"></A>
<A NAME="1956"></A>
You have to enable this option if your system acts as a
gateway between two Ethernets, or between and Ethernet and a
SLIP link, etc. Although it doesn't hurt to enable this by
default, you may want to disable this to configure a host as a
so-called firewall. Firewalls are hosts that are connected to
two or more networks, but don't route traffic between them.
They are commonly used to provide users from a company network
with Internet access at a minimal risk to the internal
network. Users will be allowed to log into the firewall and
use Internet services, but the company's machines will be
protected from outside attacks because any incoming
connections can't cross the firewall.
<P>
<P><P>
<P>
<A NAME="1959"></A>
This option works around an incompatibility with some versions
of PC/TCP, a commercial TCP/IP implementation for DOS-based
PCs. If you enable this option, you will still be able to
communicate with normal machines, but performance may
be hurt over slow links.
<P>
<P><P>
<P>
<A NAME="1963"></A>
This function enables RARP, the Reverse Address Resolution
Protocol. RARP is used by diskless clients and X-terminals to
inquire their IP-address when booting. You should enable RARP
only when you plan to serve this sort of clients. The latest
package of network utilities (net-0.32d) contains a small
utility named rarp that allows you to add systems to the
RARP cache.
<P>
<P><P>
<P>
<A NAME="1968"></A>
<A NAME="1969"></A>
<A NAME="1970"></A>
<A NAME="1971"></A>
<A NAME="2207"></A>
When sending data over TCP, the kernel has to break up the
stream into several packets before giving it to IP. For hosts
that can be reached over a local network such as an Ethernet,
larger packets will be used than for hosts where data has to
go through long-distance links.<A HREF="footnode.html#1973"><IMG ALIGN=BOTTOM ALT="gif" SRC="foot_motif.gif"></A> If you don't enable SNARL, the kernel will assume
only those networks are local that it actually has an interface
to. However, if you look at the class B network at Groucho
Marx University, the whole class B network is local, but most
hosts interface to only one or two subnets. If you enable
SNARL, the kernel will assume <em>all</em> subnets
are local and use large packets when talking to all hosts on
campus.
<P>
If you do want to use smaller packet sizes for data sent to
specific hosts (because, for instance, the data goes through
a SLIP link), you can do so using the mtu option of
route, which is briefly discussed at the end of this
chapter.
<P>
<P><P>
<P>
<A NAME="1981"></A>
<A NAME="1982"></A>
<A NAME="1983"></A>
<A NAME="1984"></A>
Nagle's rule is a heuristic to avoid sending particularly
small IP packets, also called tinygrams. Tinygrams are usually
created by interactive networking tools that transmit single
keystrokes, such as telnet or rsh. Tinygrams can
become particularly wasteful on low-bandwidth links like SLIP.
The Nagle algorithm attempts to avoid them by holding back
transmission of TCP data briefly under some circumstances. You
might only want to disable Nagle's algorithm if you have
severe problems with packets getting dropped.
<P>
<P><P>
<P>
<A NAME="1989"></A>
<A NAME="1990"></A>
<A NAME="1991"></A>
This enables support for IPX, the transport protocol used by
Novell Networking. It is still under development, and isn't
really useful yet. One benefit of this will be that you can
exchange data with IPX-based DOS utilities one day, and route
traffic between your Novell-based networks through a PPP link.
Support for the high-level protocols of Novell
networking is however not in sight, as the specifications for
these are available only at horrendous cost and under a
non-disclosure agreement.
<P>
Starting in the 1.1.16 kernel, supports another
driver type, the dummy driver. The following question appears
toward the start of the device driver section.
<P>
<P><P>
<P>
The dummy driver doesn't really do much, but is quite useful
on standalone or SLIP hosts. It is basically a masqueraded
loopback interface. The reason to have this sort of interface
is that on hosts that do SLIP but have no Ethernet, you want to
have an interface that bears your IP address all the time.
This is discussed in a little more detail in
section-<A HREF="#ifaceinterfacedummy"><IMG ALIGN=BOTTOM ALT="gif" SRC="cross_ref_motif.gif"></A> in chapter-<A HREF="node58.html#iface"><IMG ALIGN=BOTTOM ALT="gif" SRC="cross_ref_motif.gif"></A>.
<P>
<HR><A HREF="node1.html"><IMG WIDTH=65 HEIGHT=24 ALIGN=BOTTOM ALT="contents" SRC="contents_motif.gif"></A> <BR>
<B> Next:</B> <A HREF="node46.html">A Tour of Network </A>
<B>Up:</B> <A HREF="node43.html">Kernel Configuration</A>
<B> Previous:</B> <A HREF="node44.html">Kernel Options in 1.0 </A>
<P><ADDRESS>
<I>Andrew Anderson <BR>
Thu Mar 7 23:22:06 EST 1996</I>
</ADDRESS>
</BODY>
</HTML>