Re: UDP packet handling weird behaviour of various operating systems

[Keith Warno <keith.warno () valaran com>]

"Juergen P. Meier" wrote:
> > http://rootshell.com/archive-j457nxiqi3gq59dv/199803/biffit.c
> >
> > 1. Linux 2.4.7 UP (pristine source, waiting for a new shiny Alan Cox patch)
> >       - system gets frozen after 3 seconds of flood on a gigabit link.
> > Same result at a 100Mbps. The top utility shows (at least as long as it can)
> > that system(kernel) gets 100% of the CPU in its march to death. Same for
> > Linux kernel 2.2.19.
>
> 2.4.6 (modular, unpatched, selfcompiled) on an old P133:
>
> biffit against loopback: 99% cpu(system), no slowdown, system
> responds normaly. (no slowdown)
> biffit against eth0: same effect. (doh, cause linux sends it over loopback)

Confirmed, with kernel 2.4.7 (unpatched, selfcompiled, modular).  ~99%
CPU usage, but no slowdown.  Although the hardware is quite different --
proc: Thunderbird 1000
ram: 384 MB
mobo: MSI MS6340 micro atx, VIA KT 133
NIC: LinkSys LNE100TX v4.1 (using kernel-distributed tulip driver, not
that from Scyld)
(eth0: ADMtek Comet rev 17) -- altho this is irrelevent isn't it.

I've been using 2.4.whatever for only the past couple of weeks.  Maybe
I'm missing something.  I don't run any UDP services whatsoever and
would never run comsat under any circumstance.  So, why does the
sendto() in biffit.c not fail when sending to localhost?  When I boot
back to 2.2.19 and try the same thing, I get what I expect: Connection
refused.  Hrmm.  Unless I missed something in the kernel docs regarding
loopback behavior, the displayed 2.4.7 behavior seems like a Bad Thing.

kw
--                           
| Keith Warno                       cell: +1 609-209-5800
| http://www.valaran.com/           work: +1 609-716-7200 x243
| Valaran Corporation Penguin Guy   SMS : kw-mobile () valaran com
+--------------------------------------------------------------//