Linux Netfilter development
 help / color / mirror / Atom feed
From: Duncan Roe <duncan_roe@optusnet.com.au>
To: Florian Westphal <fw@strlen.de>
Cc: Pablo Neira Ayuso <pablo@netfilter.org>,
	Netfilter Development <netfilter-devel@vger.kernel.org>
Subject: Re: Documentation question (verdicts)
Date: Sat, 7 Dec 2019 22:49:09 +1100	[thread overview]
Message-ID: <20191207114909.GA928@dimstar.local.net> (raw)
In-Reply-To: <20191207111645.GB795@breakpoint.cc>

On Sat, Dec 07, 2019 at 12:16:45PM +0100, Florian Westphal wrote:
> Duncan Roe <duncan_roe@optusnet.com.au> wrote:
> > Hi Pablo,
> >
> > On Mon, Dec 02, 2019 at 09:26:23PM +1100, Duncan Roe wrote:
> > > Hi Pablo,
> > >
> > > Queue handling [DEPRECATED] in libnetfilter_queue.c documents these 3:
> > >
> > > > 278  *   - NF_ACCEPT the packet passes, continue iterations
> > > > 281  *   - NF_REPEAT iterate the same cycle once more
> > > > 282  *   - NF_STOP accept, but don't continue iterations
> > >
> > > In my tests, NF_REPEAT works as documented - the input hook presents the packet
> > > a second time. But, contrary to the above, the packet does not show again
> > > after NF_ACCEPT.
> > >
> > > Is that expected behaviour nowadays?
> > >
> > > And if so, does that make NF_STOP redundant?
> > >
> > > BTW if you'd like to try it, my test program nfq6 is a subdirectory at
> > > https://github.com/duncan-roe/nfq (nfq itself is an ad blocker).
> > >
> > > Cheers ... Duncan.
> >
> > Specifically I need to know whether to document NF_STOP as obsolete (same as
> > NF_ACCEPT).
>
> They are not the same.  STOP calls the okfn directly so packet will be
> done with the hook location.  NF_ACCEPT moves on to the next hook.
>
> table ip raw {
>         chain p1 {
>                 type filter hook prerouting priority -1000; policy accept;
>                 ip protocol icmp queue num 0 bypass
>         }
>
>         chain p2 {
>                 type filter hook prerouting priority filter; policy accept;
>                 ip protocol icmp meta mark 0x0000002a counter
>         }
> }
>
> If nfqueue tool sets mark 42 and ACCEPT, the counter will increment.
> If it uses STOP, the prerouting hook processing ends immediately
> and the packet will continue stack traversal, all further prerouting
> base chains are skipped.
>
> It will eventually appear in forward or input.

Thanks Florian.

So are you saying that the existing documentation:
> NF_STOP accept, but don't continue iterations
can be more clearly expressed by something like:
> NF_STOP accept, but skip any further base chains using the current hook
?

Cheers ... Duncan.

  reply	other threads:[~2019-12-07 11:49 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-02 10:26 Documentation question (verdicts) Duncan Roe
2019-12-07  1:28 ` Duncan Roe
2019-12-07 11:16   ` Florian Westphal
2019-12-07 11:49     ` Duncan Roe [this message]
2019-12-07 11:52       ` Florian Westphal

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191207114909.GA928@dimstar.local.net \
    --to=duncan_roe@optusnet.com.au \
    --cc=fw@strlen.de \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox