Re: [ovs-dev] [PATCH 1/5] netlink: extended ACK reporting

2017-04-18 Thread Joe Stringer 
On 18 April 2017 at 02:41, Johannes Berg  wrote:
> On Thu, 2017-04-13 at 14:44 -0700, Joe Stringer wrote
> (something that never made it to the list, due to HTML formatting)
>>
>> I think that OVS was doing some more elaborate validation than most
>> users, so over time we picked up a bunch of extra parsing code that
>> layers on top of nla_parse(). I took a look at trying to broaden this
>> and make it useful to other users a while ago, but when I posted
>> there wasn't much interest from others on it so I just moved on.
>> Maybe it's about time to pick that back up.
>
> Ah, ok. I didn't realize it was actually on top of nla_parse(). Some of
> this does seem rather useful though, and having more expressive policy
> would seem very useful too - I'd love to be able to express nesting
> better, for example.

Ah, correction - nla_parse() is used in some parts but not all of it.
More expressive policy sounds useful for OVS cases too.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Re: [ovs-dev] [PATCH 1/5] netlink: extended ACK reporting

2017-04-18 Thread Johannes Berg 
On Thu, 2017-04-13 at 14:44 -0700, Joe Stringer wrote
(something that never made it to the list, due to HTML formatting)
> 
> I think that OVS was doing some more elaborate validation than most
> users, so over time we picked up a bunch of extra parsing code that
> layers on top of nla_parse(). I took a look at trying to broaden this
> and make it useful to other users a while ago, but when I posted
> there wasn't much interest from others on it so I just moved on.
> Maybe it's about time to pick that back up.

Ah, ok. I didn't realize it was actually on top of nla_parse(). Some of
this does seem rather useful though, and having more expressive policy
would seem very useful too - I'd love to be able to express nesting
better, for example.

Also, I think we should - at least with the strict checking that Jiri
is proposing - think about checking the actual size, not just against a
minimum.

johannes
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Re: [ovs-dev] [PATCH 1/5] netlink: extended ACK reporting

2017-04-13 Thread Joe Stringer 
On 13/04/2017 12:24, "Johannes Berg"  wrote:

> On Thu, 2017-04-13 at 16:05 +0200, Nicolas Dichtel wrote:
>
> > Sure. It was just to mention that attribute 0 exists somewhere.
> > The other 0 attribute is OVS_TUNNEL_KEY_ATTR_ID.
>
> That looks like some really awkward hand-grown parsing - with all these
> "struct ovs_len_tbl" looking almost like a policy, but not using that
> code?
>
> Seems like something somebody should take a hard look at and see if it
> can't use more standard infrastructure.
>

I think that OVS was doing some more elaborate validation than most users,
so over time we picked up a bunch of extra parsing code that layers on top
of nla_parse(). I took a look at trying to broaden this and make it useful
to other users a while ago, but when I posted there wasn't much interest
from others on it so I just moved on. Maybe it's about time to pick that
back up.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev