On Tue, 06 Jan 2004 20:53:11 +0100
Olaf Titz <olaf,AT,bigred,DOT,inka,DOT,de> wrote:
> > As you can see the ident-code relies on setting the frag_off as well as
> > sk->daddr. If you do not do that, everything runs well under low load. But
> > as
> Thanks. I'm checking that in to the CVS head (cf. next mail).
> The other one was already fixed, but the fix was wrong too (just
> omitted those two lines, so under 2.2 it wouldn't count at all, or so
> it looks to me.)
Did you incorporate the ignoredf/forcemtu patch (from long ago :-)? It really
does work and and all mtu-related issues are gone afterwards.
It would be nice to release some 1.5.X stable for 2.2/2.4 kernel use, just to
have something like a clear cut. 1.5.4 is not really up-to-date any more.
Thanks for your continuing efforts, especially regarding 2.6. I feel very
comfortable with your maintainership.