Re: [Cake] Possible conntrack lookup improvements

[Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>]

> On 3 May 2019, at 15:22, Toke Høiland-Jørgensen <toke@redhat.com> wrote:
> 
> Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk> writes:
> 
>> Hi Everyone,
>> 
>> I’ve been working on act_ctinfo toward getting that upstream and it is
>> getting closer. Since that module along with act_connmark does its own
>> conntrack lookups I’ve been looking at what they do and what we do in
>> cake.
>> 
>> Two patches attached - one is a simple variable elimination with no
>> functional change. The second changes/simplifies the conntrack tuple
>> lookup & usage. I’ve had a play and I don’t think I’ve broken any of
>> the host fairness BUT it could do with some more testing, that’s where
>> you come in… probably Pete & George :-)
> 
> Seems reasonable. But please fold these two patches into one; changing
> everything, then immediately changing it again does not help
> readability... And the explanation makes a lot more sense if you just
> change the whole thing in one patch :)
> 
> -Toke

Yeah, when I do the PR after testing confirms I haven’t totally screwed up host fairness in the process I’ll of course squash them together :-) The 1st patch is a no brainer, the second should be a no brainer but it needs more testing than I have given it.

I went down this path as a result of my act_ctinfo work which in the latest version is able to restore DSCP & skb->marks from conntrack.  I had an idea to restore the ct info as well, so CAKE didn’t have to do its ‘look harder’ lookup.  Then I noticed how cake sort of does the harder lookup backwards.


Cheers,

Kevin D-B

gpg: 012C ACB2 28C6 C53E 9775  9123 B3A2 389B 9DE2 334A