On Tue, Jul 28, 2009 at 4:30 AM, Christopher
Morrow<morrowc.li...@gmail.com> wrote:
> (hopefully this time gmail selects the right outbound from addr grr)
>
> On Tue, Jul 28, 2009 at 4:24 AM, <sth...@nethelp.no> wrote:
>>> Out of curiosity, what's the signal back to the sender that his/her
>>> packet was dropped?? NFS (in some implementations) doesn't checksum
>>> UDP packets, DNS doesn't, there are quite a few things that don't
>>> checksum UDP packets.
>>
>> I believe this is normally a function of the operating system, not
>> NFS, DNS etc. Hopefully old SunOS versions that had UDP checksums
>
> sure, but still how does the application and user know that their
> 'thing they are doing' is broken, why it's broken and how to fix it?
> Keep in mind that today this all works and users don't have to
> understand 'checksum' or 'udp' or 'ip version'... tomorrow something
> (hopefully) seemless will happen and their 'thing they are doing' will
> just keep on working.
>
> Spurious failures as would be happen in this proposal will be
> problematic, and will (again) cause hesitance in deployments.
>
>> off by default are dead by now.
>
> linux nfs doesn't do (nor does netapp apparently) checksums on udp nfs
> today (current 2.6 linux kernel, ubuntu 8.x)

apologies, I had a tcpdump expr fail :( I do see DNS though with out
checksums, I'll go dig for some more NFS or other UDP on my test
host(s).
-chris
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to