Subject: Re: wm0: wm_tx_offload: packet headers did not fit in first mbuf,
To: Bill Studenmund <wrstuden@NetBSD.org>
From: Markus W Kilbinger <kilbi@rad.rwth-aachen.de>
List: current-users
Date: 12/02/2004 20:18:37
>>>>> "Bill" == Bill Studenmund <wrstuden@NetBSD.org> writes:

    >> [...]
    >> Though the machines is still accessible from one side (for
    >> login e. g.), it no longer routes network packets ...
    >> 
    >> A kernel compiled from Nov 13th -current sources does not show
    >> this misbehavior.

Some news: As Allen Briggs pointed out in a personal mail: Turning off
all the *4csum hardware features of these nics 'solved' the problem.

    Bill> My undetstanding is that this change reflects cathing an
    Bill> error condition that went undetected (and thus unreported)
    Bill> before. So it's not that the kernel started misbehaving,
    Bill> it's that it started telling you about issues.

Hmm, why did routing work before? Is the newly detected 'issue'
preventing a formerly (and falsely) functioning non misbehavior?

Markus.