linux-kernel - Re: [PATCH net-next v7 0/5] netconsole: support automatic target recovery

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: <20251127164748.4db1e1fe@kernel.org>
Date: Thu, 27 Nov 2025 16:47:48 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: Andre Carvalho <asantostc@...il.com>
Cc: Breno Leitao <leitao@...ian.org>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet
 <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>, Shuah Khan
 <shuah@...nel.org>, Simon Horman <horms@...nel.org>,
 netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-kselftest@...r.kernel.org
Subject: Re: [PATCH net-next v7 0/5] netconsole: support automatic target
 recovery

On Thu, 27 Nov 2025 23:07:02 +0000 Andre Carvalho wrote:
> > Should we not be listening for the REGISTER event then? On boot
> > we force UP the device if we find it, theoretically there may
> > be a case where user space is not configured to UP the device,
> > and then we'd never resume the target?  
> 
> This is indeed a limitation on the current implementation. Based on
> your feedback, I'm working on a new version of this series handling REGISTER
> instead of UP and ensuring we force UP the device.
> This will make it consistent with the boot behavior you described.
> 
> Based on my tests, I can't force the device UP while handling the REGISTER event.
> I believe this is due to dev_open attempting to lock the device which is already held.
> For this reason, I'm resorting to defering this to a workqueue, similar to my approach 
> on v1 [1] (but correctly handling target_list lock).
> 
> Let me know if this approach makes sense or if I'm missing something.

SG, that's probably the most resilient solution.