Florian Obser <[email protected]> wrote:
> That's my gut feeling as well. But I don't have any hard evidence that
> that is actually correct. However, since we learned that router-alert is
> an actual problem it is also a good incremental step.
How many years did it take to notice this bug? Surely this was affecting
other people, but noone sniffed to discover the reason.
If we only make pf forgiving for a few options, but other options cause
problems in the future, it may also take a long time to fix them.
Can we turn it upside down, and consider blocking only options which are
believed to be dangerous, or at least make such a list to think it through?
Like maybe;
known option, known to be dangerous
known option, but good or required
known option, but no dangerous impact seen
currently undefined options (assume bad)
And then use such a list to guide the switch () statement in this icmp case?