Re: Summary lightweight tunnel discussion at NFWS

[roopa <roopa@cumulusnetworks.com>]

On 7/6/15, 6:03 AM, Thomas Graf wrote:
> The integration into the routing code looks much better now. Any
> chance you can rebase your tree and squash it into logical
> commits?
Will do today.
>   I'm rebasing again onto yours and I think we should submit
> all of this in a single series so everything can be reviewed in a
> single thread.
agreed.
>
> All of this is only useful if distros enable this by default which
> minimize the value of a config option quite a bit. I think what you
> have your latest tree is a good balance.
ok
>
>> I have been thinking of moving lwtstate from rtable to struct dst_entry.
>> I will also look at the dst_metadata.
> I thought about the same. The introduction of dst_metadata makes
> that a bit pointless because we don't want to store a pointer to the
> tunnel metadata but allocate the metadata as a dst directly to avoid
> the indirection. We need the separate code paths anyway until v6 and
> v4 routing are merged so we would save very little while penalizing
> everyone except rtable and fib6.
agreed,

thanks,
Roopa