[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <BYUIZiIrbb9DuEGCzlZalUF5QXjBrQOyJOBOS6fyWQWUJkUN29LhrahP0603mdAW3y5Y7pgSJc0i8Y2O5yJ6ythPAknipI3h4N-aceF6jos=@1g4.org>
Date: Mon, 19 Jan 2026 18:50:08 +0000
From: Paul Moses <p@....org>
To: Jakub Kicinski <kuba@...nel.org>
Cc: netdev@...r.kernel.org
Subject: Re: [PATCH net v1 0/3] act_gate fixes and gate selftest update
Hi Jakub,
I originally submitted to security@...nel.org on Dec 27th and this submission was a direct request from the maintainers. So if your concern is whether or not they are tracking, I can confirm that they are. If it's just a matter of adherence to process, than I will go ahead and do it, but wanted to avoid the churn if unnecessary.
The big question currently is what is appropriate for stable. I do not believe it's possible to fix this issue without negatively impacting the performance of the gate without either using a unlock → cancel → relock pattern or conversion to RCU. Neither of which is ideal for stable, so more constructive input is needed.
Thanks
Paul
On Monday, January 19th, 2026 at 12:32 PM, Jakub Kicinski <kuba@...nel.org> wrote:
>
>
> On Fri, 16 Jan 2026 11:25:50 +0000 Paul Moses wrote:
>
> > This series fixes act_gate schedule update races by switching to an RCU
> > prepare-then-swap update pattern and ensures netlink dump structs are
> > zeroed to avoid leaking padding to userspace. It also updates the
> > tc-testing gate replace test to include the mandatory schedule entries
> > so the test suite reflects the action's strict semantics.
>
>
> Please repost and CC appropriate maintainers and authors.
> ./scripts/get_maintainer.pl $patch
> --
> pw-bot: cr
Powered by blists - more mailing lists