I'm seeing it with 3.4. Regarding your other question (Do you also see
messages from IPFilter on B about interfaces "detaching" and then "attaching"
?) - I can reproduce the problem simply by running "ipf -y" which would lead me
to conclude that the ipfsync() call is indeed the culprit.
Erick.
--- Darren Reed <[EMAIL PROTECTED]> wrote:
> In some email I received from [EMAIL PROTECTED], sie wrote:
> >
> > Well, I just witnessed the exact same happening when logged in to home.
> >
> > I suppose it has something to do with:
> >
> > int _info(modinfop)
> > struct modinfo *modinfop;
> > {
> > int ipfinst;
> >
> > ipfinst = mod_info(&modlink1, modinfop);
> > #ifdef IPFDEBUG
> > if (ipf_debug)
> > cmn_err(CE_NOTE, "IP Filter: _info(%x) = %x",
> > modinfop, ipfinst);
> > #endif
> > if (fr_running > 0)
> > ipfsync();
> > return ipfinst;
> > }
> >
> >
> > Darren can possibly comment on what the code does and what may cause the
> > state information to be lost?
>
> I believe this is a hangover from very early on and can probably
> disappear now. What it is trying to do is use the kernel walking
> through all of the modules as an indication that a device has
> appeared or disappeared.
>
> If I look at the current IPFilter code, the two lines:
>
> if (fr_running > 0)
> ipfsync();
>
> are no longer present.
>
> Are you seeing this with 3.4 or 4.1 ?
>
> Darren
>
