Re: [PATCH] ipv4: mitigate an integer underflow when comparing tcptimestamps

[ZHANG, Le]

On 11:55 Sun 14 Nov     , David Miller wrote:
> From: Eric Dumazet <eric.dumazet@xxxxxxxxx>
> Date: Sun, 14 Nov 2010 09:52:25 +0100
> 
> > Really, if you have multiple clients behind a common NAT, you cannot use
> > this code at all, since NAT doesnt usually change TCP timestamps.
> 
> NAT is %100 incompatible with TW recycling, full stop.
> 
> There is no maybe, or maybe not.
> 
> If you are behind NAT you must not turn this feature on, ever.

Sorry, this question may be OT on this list, but I am just curious:

Is there any other OS has implemented this feature like Linux?

To be very specific, by this feature, I mean rejecting old duplicates based
on per-host cache of last timestamp received from any connections.
As suggested in RFC1323 Appendix B.2 (b).

Does anyone, by any chance, know the answer? Thanks in advance!

-- 
ZHANG, Le
http://zhangle.is-a-geek.org
0260 C902 B8F8 6506 6586 2B90 BC51 C808 1E4E 2973

Attachment: pgp00000.pgp
Description: PGP signature