netdev - Re: [PATCH net v2] devlink: Delay health recover notification until devlink registered

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: <20230810103300.186b42c8@kernel.org>
Date: Thu, 10 Aug 2023 10:33:00 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Shay Drory <shayd@...dia.com>
Cc: <netdev@...r.kernel.org>, <pabeni@...hat.com>, <davem@...emloft.net>,
 <edumazet@...gle.com>, Jiri Pirko <jiri@...dia.com>
Subject: Re: [PATCH net v2] devlink: Delay health recover notification until
 devlink registered

On Wed, 9 Aug 2023 23:35:21 +0300 Shay Drory wrote:
> From one side, devl_register() is done last in device initialization
> phase, in order to expose devlink to the user only when device is
> ready. From second side, it is valid to create health reporters
> during device initialization, in order to recover and/or notify the
> user.
> As a result, a health recover can be invoked before devl_register().
> However, invoking health recover before devl_register() triggers a
> WARN_ON.

My comment on v1 wasn't clear enough, I guess.

What I was trying to get across is that because drivers can take
devl_lock(), devl_register() does not have to be last.

AFAIU your driver does:

  devlink_port_health_reporter_create()
  ...
  devlink_register()

why not change it to do:

  devl_lock()
  devl_register()
  devl_port_health_reporter_create()
  ...
  devl_unlock() # until unlock user space can't access the instance

?