linux-kernel - Re: [PATCH] selftests: rtnetlink: Fix bridge_parent

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: <Z_5E-wuE6b3HyHRU@shredder>
Date: Tue, 15 Apr 2025 14:37:31 +0300
From: Ido Schimmel <idosch@...sch.org>
To: Alok Tiwari <alok.a.tiwari@...cle.com>
Cc: davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
	pabeni@...hat.com, horms@...nel.org, shuah@...nel.org,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	linux-kselftest@...r.kernel.org, darren.kenny@...cle.com
Subject: Re: [PATCH] selftests: rtnetlink: Fix bridge_parent_id failure on
 interface state

On Mon, Apr 14, 2025 at 10:25:33AM -0700, Alok Tiwari wrote:
> The selftest "kci_test_bridge_parent_id" fails with the error:
> "Device can not be enslaved while up" when trying to attach interfaces
> (`eni10np1`, `eni20np1`) to a bonding device (`test-bond0`) while the
> interfaces are in the UP state.

Why are they up? The test creates the interfaces and never brings them
up.

It's most likely caused by some interface manager in your user space. I
suggest fixing that instead.

> 
> Failure log:
> COMMAND: ip link set dev eni10np1 master test-bond0
>     Error: Device can not be enslaved while up.
> COMMAND: ip link set dev eni20np1 master test-bond0
>     Error: Device can not be enslaved while up.
> FAIL: bridge_parent_id
> 
> This behavior aligns with bonding driver requirements, where a slave
> interface must be in the DOWN state before being enslaved. This was
> reinforced in upstream commit: 'ec4ffd100ffb ("Revert 'net: rtnetlink:
> Enslave device before bringing it up'")'.
> 
> This patch updates the test to bring interfaces down explicitly before
> adding them to the bonding device:

I don't see why the test needs to bring them down when it never brought
them up to begin with.