[syslinux] rpcinfo *very* slow on diskless client

[Adam Nielsen]

> Unfortunately during system boot up to test whether the RPC portmapper
> is loaded, the boot scripts run "rpcinfo -p" which causes the PC to
> sit there for about ten minutes and then eventually time out and
> continue with the boot process.
> Does anyone have any idea why this would happen?

Well, in case anyone has a similar problem in the future, I worked out
what the problem was.  It seems that the 'lo' interface wasn't being
assigned the IP address 127.0.0.1, so whenever any program tried to
connect to 'localhost' it would just sit there and eventually time out
instead of either connecting or returning a 'connection refused'
response.

It seems that my distro's boot scripts (Slackware 9.1) were checking to
see whether the 'lo' interface was up, and if not they enabled it and
assigned it an IP.  But it seems that the kernel's IP autoconfiguration
process while correctly setting up eth0, also enabled 'lo' but didn't
assign it an IP - which tricked the init scripts into thinking 'lo' was
already enabled and so they left it alone, when in actual fact it still
needed an IP assigned to it.

Anyway, I've changed the init script to check whether 'lo' is lacking an
IP instead, and it's now assigned an IP during boot, fixing all the
problems.  Yay! ;-)

Cheers,
Adam.