🚀 go-pugleaf

In article <2j7tpj$98h@lsupoz.apana.org.au> smilie@lsupoz.apana.org.au (Anthony Rumble) writes:
>Bah.. ALl these problems with the standard networking in Linux..
>
>Beta-4 is sitting there running REALLY REALLY well and stable..
>I haven't had a crash AT ALL.. and ive been running beta-4 since
>last week..
>
Oh yes and this is a bug in the code that fixes the AT1500 driver properly
I don't believe net2e even supports it. Do you have Nagle working properly
yet.

For the bored try binding a tcp socket in 2eb4 and connecting it to itself
or maybe using TCP options before TCP_MSS or at least on 2E BETA3 you
could remotely crash any 2E beta3 machine by sending it a TCP_MSS with
an option length of 0... All of those happen in the real world.


>Hmm.. I wish someone would get the two camps to talk and get the
>fixed code into the kernel.
>

Me too. After 1.0 I hope we can find some kind of accomodation or at least
pull out some of the structural improvements from NET2E, and the little
niceties like SIOCGIFSTATS. Just at the moment I'm more concerned in
chasing the possibly network possibly tty bug in SLIP dropping carrier
and the small buffer locking problem with lance cards.


In the longer term neither 2Debugged or 2E will be up to the demands
being placed on them (trying variable length device headers one day,
or a hardware address length different to the arp length, or having
a socket layer open another socket in the kernel and work cleanly
with it. For that matter try NFS swapping) - all of these need 
addressing and with luck we can all get together and pool ideas.

Alan
iiitac@pyr.swan.ac.uk