In some mail from itojun@iijlab.net, sie said:
> 
> >> >Looking at ALTQ from KAME (I just did a cvs update from kame), there
> >> >is currently no use of m_tag that is found in the OpenBSD source.
> >> >Is this because changes are still pending being pulled up into KAME
> >> >from OpenBSD to support this ?
> >> 	you have to look at kame/openbsd tar.gz, not under kame/sys/altq.
> >> 	the latter is obsolete.
> >Do you want to expand on this change in how KAME development is
> >proceeding or do we have to guess about what's going on ?
> >
> >kame/kame/sys/altq shows changes as recent as June 27 but this is
> >obsolete, you say ?  Why ?
> 
> 	these months ALTQ development has been done on openbsd primarily
> 	(openbsd hackerthon, kjc-PF collaboration), and openbsd/sys/altq has
> 	the latest piece of software.  kame/sys/altq is still there
> 	(and distributed in tarballs) but is obsolete.

While we are nutting out the proper interface for moving packet
classification from KAME to other code, is there any reason why
an update, in NetBSD, of ALTQ to something just prior to the
merging with PF would be bad ?

> 	our plan is:
> 	- make kame/sys/net/pf* portable (netbsd done, freebsd still to go)
> 	- move content of openbsd/sys/altq to kame/sys/altq, put #ifdef for
> 	  portability
> 	so, to provide PF+ALTQ on all platforms (free/net/open) is what we're
> 	targetting.

OK, but I think there is more work required on the ALTQ side of things -
see my comments to Kenjiro - before it should be merged with NetBSD.

Darren