But why does Solaris ipfilter not handle dohwcksum=1 correctly? After
all,
this parameter _is_ tested in ip_nat.c, ip_proxy.c, and solaris.c.
Most likely because they've changed the way it works.
I haven't gone looking at header files or STREAMS messages yet to see what
is different or wrong in IPFilter's handling of this. The problem for me
with IPFilter is that this is effectively undocumented (except for the
header files) so they're at liberty to change how it works without needing
to worry about compatibility.
I could rant here quite extensively about this issue.
So basically, it's a private interface? If that's the case, yeah, you don't
stand a chance, since that's how it's meant to work then. Some of the stuff
is kept private so that changes could be made without too much fuss.
Did you file an RFE to have a public interface designed and exported for
that?
If you did file an RFE, what is the status?
