netdev - Re: [Bug 220774] netem is broken in 6.18

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] [day] [month] [year] [list]

Message-ID: <CAM0EoMmpuCdD6LpSOPuy0cXEvr7aJhSRDOhqzQ5mrq8-zjJaxQ@mail.gmail.com>
Date: Mon, 24 Nov 2025 15:54:44 -0500
From: Jamal Hadi Salim <jhs@...atatu.com>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: William Liu <will@...lsroot.io>, 
	Linux Kernel Network Developers <netdev@...r.kernel.org>, Jakub Kicinski <kuba@...nel.org>, lrGerlinde@...lfence.com
Subject: Re: [Bug 220774] netem is broken in 6.18

Hi lrGerlinde@...lfence.com (sorry, no name on that email!),

On Sun, Nov 23, 2025 at 8:33 PM Cong Wang <xiyou.wangcong@...il.com> wrote:
>
> ---------- Forwarded message ---------
> From: <bugzilla-daemon@...nel.org>
> Date: Fri, Nov 21, 2025 at 3:39 PM
> Subject: [Bug 220774] netem is broken in 6.18
> To: <xiyou.wangcong@...il.com>
>
>
> https://bugzilla.kernel.org/show_bug.cgi?id=220774
>
> Gerlinde (lrGerlinde@...lfence.com) changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |lrGerlinde@...lfence.com
>
> --- Comment #2 from Gerlinde (lrGerlinde@...lfence.com) ---
> You know, in our HA and ECMP setup we sometimes get this strange thing: packets
> are not lost, but some flows get duplicated for a few hundert milliseconds. Not
> all flows, only the unlucky ones that get hashed through the broken path. And
> this makes debugging really a pain, because all metrics say “no loss”, but the
> application still behave stupid.
>
> So we use netem duplicate on a single mq-queue to copy exactly this situation.


I may be misunderstanding: You seem to be using a single mq-queue with
netem and your requirement seems to be replicating the packet on that
single mq-queue? You mention "flow" - how do you map these to the
queues?
BTW, you may be aware of this, but you should also be able to mirror
arbitrary flows of choice without resorting to using netem.
Can you share your config?

cheers,
jamal


> --
> You may reply to this email to add a comment.
>
> You are receiving this mail because:
> You are on the CC list for the bug.