netdev - Re: [patch net-next] devlink: don't take instance lock for nested handle put

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: <ZSQeNxmoual7ewcl@nanopsycho>
Date: Mon, 9 Oct 2023 17:37:27 +0200
From: Jiri Pirko <jiri@...nulli.us>
To: Jakub Kicinski <kuba@...nel.org>
Cc: netdev@...r.kernel.org, pabeni@...hat.com, davem@...emloft.net,
	edumazet@...gle.com, gal@...dia.com
Subject: Re: [patch net-next] devlink: don't take instance lock for nested
 handle put

Mon, Oct 09, 2023 at 05:15:32PM CEST, kuba@...nel.org wrote:
>On Sat, 7 Oct 2023 12:17:31 +0200 Jiri Pirko wrote:
>>> Isn't the PF driver processing the "FW events"? A is PF here, and B 
>>> is SF, are you saying that the PF devlink instance can be completely
>>> removed (not just unregistered, freed) before the SF instance is
>>> unregistered?  
>> 
>> Kernel-wise, yes. The FW probably holds necessary resource until SF goes
>> away.
>
>I think kernel assuming that this should not happen and requiring 
>the PF driver to work around potentially stupid FW designs should
>be entirely without our rights.

But why is it stupid? The SF may be spawned on the same host, but it
could be spawned on another one. The FW creates SF internally and shows
that to the kernel. Symetrically, the FW is asked to remove SF and it
tells to the host that the SF is going away. Flows have to go
through FW.