On Fri, 2003-02-21 at 15:04, Nathan Ryan Milford wrote:
PF does not break Linux NFS.
Using the scrub directive will probably detect and drop packets from Linux
NFS as they decided to not follow standards.
I'm not a pf developer, but I'd doubt they'd waste thier time on something
that isn't thier fault. I'd wait for Linux NFS to follow protocol.
I'm not suggesting it's PF's fault (hence the quotes around broken).
If you've followed recent developments, you'd understand the reason
Linux NFS doesn't work through normalized PF (scrub) is that the PF
developers refused to respect the DF bit on fragmented Linux NFS traffic
without understanding WHY in the hell they wanted to do it that way.
Although the RFC's dictate otherwise, the Linux maintainers stand by the
notion that the RFC's are _not_LAW_, and are only there as a guideline.
Now that this has come to light
(http://kerneltrap.org/node.php?id=579cid=2386pid=2382#2386), I expect
that Daniel plans on integrating this in due course.
Just curious when that may be.
-J.
- Original Message -
From: Jason Dixon [EMAIL PROTECTED]
To: PF Mailing List [EMAIL PROTECTED]
Sent: Friday, February 21, 2003 2:40 PM
Subject: Linux NFS no-DF status
I don't want anyone's hair to stand on end, but I was just curious...
with the clarification recently given by the Linux camp on the
NFS/DF-bit issue, is there an effort currently under way to recognize
and support their implementation for 3.3 -release? Meaning, PF won't
break it anymore?
I don't need it, I was just sitting around thinking about it and got
curious.
-J.
