Re: [PATCH 0/3] Fix regression in auto port forwarding

[Stefano Brivio <sbrivio@redhat.com>]

On Wed, 19 Nov 2025 15:26:31 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> My recently merged series cleaning up the auto-port-forward scanning
> contained a series bug: automatic forwards would only appear on
> alternating seconds.
> 
> This turned out to be due to a fundamentally broken premise in my
> thinking.  I was thinking that for consistency we wanted the most
> recent port map information throughout the process.  But that's not
> really true: for the purposes of exclusion what we really need to know
> is which of the listening sockets we scan are ours.  That's given by
> the *prior* state of the forward maps, not the updated one based on a
> new scan.
> 
> The series also had a number of worthwhile changes though.  This
> series fixes it up, by reverting the most misguided of the patches and
> correcting behaviour of another one while preserving the accompanying
> code re-orgs.
> 
> Link: https://bugs.passt.top/show_bug.cgi?id=176
> 
> David Gibson (3):
>   Revert "fwd: Update all port maps before applying exclusions"
>   fwd: Exclude ports based on prior mapping state
>   fwd: Don't explicitly exclude reverse-direction TCP ports for UDP

Applied.

-- 
Stefano