netdev - Re: [Patch net v5 4/9] net_sched: Prevent using netem duplication in non-initial user namespace

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aS/Nk8ujLJttzKNo@pop-os.localdomain>
Date: Tue, 2 Dec 2025 21:41:39 -0800
From: Cong Wang <xiyou.wangcong@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: netdev@...r.kernel.org, stephen@...workplumber.org,
	Cong Wang <cwang@...tikernel.io>
Subject: Re: [Patch net v5 4/9] net_sched: Prevent using netem duplication in
 non-initial user namespace

On Mon, Dec 01, 2025 at 04:25:24PM -0800, Jakub Kicinski wrote:
> On Wed, 26 Nov 2025 11:52:39 -0800 Cong Wang wrote:
> > The netem qdisc has a known security issue with packet duplication
> > that makes it unsafe to use in unprivileged contexts. While netem
> > typically requires CAP_NET_ADMIN to load, users with "root" privileges
> > inside a user namespace also have CAP_NET_ADMIN within that namespace,
> > allowing them to potentially exploit this feature.
> > 
> > To address this, we need to restrict the netem duplication to only the
> > initial user namespace.
> 
> What gives us the confidence that this won't break existing setups?
> Pretty sure we use user ns at Meta, tho not sure if any of our
> workloads uses both those and netem dup.

All the reports (https://bugzilla.kernel.org/show_bug.cgi?id=220774) we
had so far didn't mention user namespace. This is the only data point I
have.

I can drop this patch, but I am not sure if patch 3/9 is sufficient to
convince Will on user namespace security.

Regards,
Cong