Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 16 May 2002 17:26:32 +1000 (EST)
From:      Cristan Szmajda <cls@cse.unsw.edu.au>
To:        Giorgos Keramidas <keramida@FreeBSD.ORG>
Cc:        current@FreeBSD.ORG
Subject:   Re: bad tcp cksum fffe!
Message-ID:  <Pine.GSO.4.44.0205161719480.15479-100000@mozart.orchestra.cse.unsw.EDU.AU>
In-Reply-To: <20020514173036.GF5715@hades.hell.gr>

next in thread | previous in thread | raw e-mail | index | archive | help
Dear Giorgos,

Thanks for your suggestion!  The problem has gone away by
backing off to -O1 in kernel and -O2 in userland.  Strange,
because I have been using -O3 -fomit-frame-pointer for a
while now (/usr/share/examples/etc/make.conf notwithstanding)
with ill effects.  My bad.

Sorry, but I can't be more specific on which particular flag
or which particular CVS update caused the problem.  CPU too
slow.

Thanks again,
-- Chris


On Tuesday, Giorgos Keramidas said:

> On 2002-05-14 17:39, Cristan Szmajda wrote:
> > Dear freebsd-current,
> >
> > Any suggestions you have regarding this problem would be
> > much appreciated.
> >
> > My laptop running -CURRENT is suddenly generating bad TCP
> > checksums when talking to some IPs but not others.  For
> > example, 129.94.209.220 is a problem,
> >
> >     17:08:47.026823 129.94.233.200.1032 > 129.94.209.220.22: S
> >     [bad tcp cksum fffe!] 3868790363:3868790363(0) win 65535
> >     <mss 1460,nop,wscale 1,nop,nop,timestamp 542813 0,nop,nop,ccnew 10>
> >     (DF) (ttl 64, id 347, len 68)
>
> I've seen this only twice.  The first time I was using too funky
> optimizations to the CFLAGS of my kernel, and GCC was obviously doing
> something wrong with the checksum generation code.  If this is the
> case, then rebuild your kernel and userland with the suggested
> optimizations from /usr/share/examples/etc/make.conf and see if this
> fixes things for you.
>
> The second one was when my tcpdump (and the rest of the userland) was
> really VERY out of sync with the running kernel (I had managed to
> upgrade only the kernel and tcpdump was probably reading random data
> from bpf).  But there were other more important things wrong with this
> case.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.44.0205161719480.15479-100000>