libnftables set element printing, DATA_CHAIN

libnftables set element printing, DATA_CHAIN

[Patrick McHardy]

[resending because of incorrect list address]

I was looking at some incorrect output in netlink debugging for set elements
and noticed a few things that seem odd:

First, the regular netlink debugging seems very incomplete, it doesn't
print verdict types, it prints [end] whether the set contains intervals
and the set is an interval end or not and its messing up the output.

Next I'm wondering about what DATA_CHAIN is supposed to be. I guess its
the chain for a jump or goto verdict, This is even encoded in the
XML and JSON output. This seems wrong to me, there is no DATA_CHAIN,
there are JUMP or GOTO verdicts that include a chain. The specific
verdict is also missing, so its not possible to distinguish these
two cases.

I could fix up the debugging output, but this looks like someone more
familiar with the XML and JSON stuff should have a look at this and
fix all of this consistently before we release it as wire format.

Re: libnftables set element printing, DATA_CHAIN

[Arturo Borrero Gonzalez]

On 15 January 2014 23:50, Patrick McHardy <kaber@trash.net> wrote:
> [resending because of incorrect list address]
>
> I was looking at some incorrect output in netlink debugging for set elements
> and noticed a few things that seem odd:
>
> First, the regular netlink debugging seems very incomplete, it doesn't
> print verdict types, it prints [end] whether the set contains intervals
> and the set is an interval end or not and its messing up the output.
>

It's true, most of the development went to the JSON/XML formats, not
the default one.

> Next I'm wondering about what DATA_CHAIN is supposed to be. I guess its
> the chain for a jump or goto verdict, This is even encoded in the
> XML and JSON output. This seems wrong to me, there is no DATA_CHAIN,
> there are JUMP or GOTO verdicts that include a chain. The specific
> verdict is also missing, so its not possible to distinguish these
> two cases.
>

Ok, let's fix that. Do you have a proposal?

This is what I think you want:

* a <data_reg type=value> with raw data.
* a <data_reg type=verdict> with a concrete verdict (eg drop accept)
* a <data_reg type=jump> with a chain to jump to.
* a <data_reg type=goto> with a chain to go to.

Or maybe:

* a <data_reg type=value> with raw data.
* a <data_reg type=verdict> with a concrete verdict (eg drop accept)
and if verdict == jump or goto, then a <chain> with the destination.

> I could fix up the debugging output, but this looks like someone more
> familiar with the XML and JSON stuff should have a look at this and
> fix all of this consistently before we release it as wire format.

I can handle it.

regards
-- 
Arturo Borrero González
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: libnftables set element printing, DATA_CHAIN

[Patrick McHardy]

On Thu, Jan 16, 2014 at 11:30:26AM +0100, Arturo Borrero Gonzalez wrote:
> On 15 January 2014 23:50, Patrick McHardy <kaber@trash.net> wrote:
> > [resending because of incorrect list address]
> >
> > I was looking at some incorrect output in netlink debugging for set elements
> > and noticed a few things that seem odd:
> >
> > First, the regular netlink debugging seems very incomplete, it doesn't
> > print verdict types, it prints [end] whether the set contains intervals
> > and the set is an interval end or not and its messing up the output.
> >
> 
> It's true, most of the development went to the JSON/XML formats, not
> the default one.
> 
> > Next I'm wondering about what DATA_CHAIN is supposed to be. I guess its
> > the chain for a jump or goto verdict, This is even encoded in the
> > XML and JSON output. This seems wrong to me, there is no DATA_CHAIN,
> > there are JUMP or GOTO verdicts that include a chain. The specific
> > verdict is also missing, so its not possible to distinguish these
> > two cases.
> >
> 
> Ok, let's fix that. Do you have a proposal?
> 
> This is what I think you want:
> 
> * a <data_reg type=value> with raw data.
> * a <data_reg type=verdict> with a concrete verdict (eg drop accept)
> * a <data_reg type=jump> with a chain to jump to.
> * a <data_reg type=goto> with a chain to go to.
> 
> Or maybe:
> 
> * a <data_reg type=value> with raw data.
> * a <data_reg type=verdict> with a concrete verdict (eg drop accept)
> and if verdict == jump or goto, then a <chain> with the destination.

Yes, this is what I think we should do.

> > I could fix up the debugging output, but this looks like someone more
> > familiar with the XML and JSON stuff should have a look at this and
> > fix all of this consistently before we release it as wire format.
> 
> I can handle it.

Great, thanks a lot.