From: Stefano Brivio <sbrivio@redhat.com>
To: David Gibson <david@gibson.dropbear.id.au>
Cc: Jon Maloy <jmaloy@redhat.com>, passt-dev@passt.top
Subject: Re: [PATCH v5] udp: support traceroute
Date: Fri, 4 Apr 2025 13:01:45 +0200 [thread overview]
Message-ID: <20250404130145.57b91f6c@elisabeth> (raw)
In-Reply-To: <Z-8uliWIs7YYWCdF@zatzit>
On Fri, 4 Apr 2025 11:57:58 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:
> On Fri, Apr 04, 2025 at 10:40:27AM +1100, David Gibson wrote:
> > On Thu, Apr 03, 2025 at 06:27:06PM -0400, Jon Maloy wrote:
> > > Now that ICMP pass-through from socket-to-tap is in place, it is
> > > easy to support UDP based traceroute functionality in direction
> > > tap-to-socket.
> > >
> > > We fix that in this commit.
> > >
> > > Link: https://bugs.passt.top/show_bug.cgi?id=64
> > > Signed-off-by: Jon Maloy <jmaloy@redhat.com>
> >
> > Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
> >
> > One commont below.
>
> Oh.. wait. I just realised this patch has a weird side effect, for
> flows which are initiated from outside, rather than from the guest.
>
> If the flow is initiated from outside, it's maybe a bit unlikely, but
> we could still get a non-default TTL from the guest on reply
> datagrams. That will trigger this code and alter the socket's TTL.
> But in this case the socket is not a flow specific socket, but the
> listening socket which initiated the flow, which could be handling
> packets on many flows.
>
> The "cached" TTL is stored per-flow, not per-socket, so we won't look
> at the right value when we process datagrams from other flows, so they
> may go out with the wrong TTL.
>
> I think the obvious way to address this is to stop using the
> "listening" socket to send datagrams for a flow, using a connect()ed
> socket instead. We have other reasons to do that too, and I'm working
> on implementing that right now.
>
> The question is whether this is a serious enough problem to delay this
> until the "both sides connect()ed sockets" change is merged.
On the other hand, this patch applies cleanly on top of your "Use
connect()ed sockets for both sides of UDP flows" series, and also
semantically I couldn't spot any particular change or integration
that we would need as a result in this patch.
I haven't reviewed the series yet, but I don't think that an eventual
re-spin would change this, so... problem solved I guess? I can just
wait and apply them together.
--
Stefano
next prev parent reply other threads:[~2025-04-04 11:01 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-03 22:27 [PATCH v5] udp: support traceroute Jon Maloy
2025-04-03 23:40 ` David Gibson
2025-04-04 0:57 ` David Gibson
2025-04-04 11:01 ` Stefano Brivio [this message]
2025-04-04 11:11 ` David Gibson
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=20250404130145.57b91f6c@elisabeth \
--to=sbrivio@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=jmaloy@redhat.com \
--cc=passt-dev@passt.top \
/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.
Code repositories for project(s) associated with this public inbox
https://passt.top/passt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).