linux-kernel - Re: [PATCH v6 1/3] revocable: Add fops replacement

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: <20251117153301.GD10864@nvidia.com>
Date: Mon, 17 Nov 2025 11:33:01 -0400
From: Jason Gunthorpe <jgg@...dia.com>
To: Tzung-Bi Shih <tzungbi@...nel.org>
Cc: Benson Leung <bleung@...omium.org>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	"Rafael J . Wysocki" <rafael@...nel.org>,
	Danilo Krummrich <dakr@...nel.org>,
	Jonathan Corbet <corbet@....net>, Shuah Khan <shuah@...nel.org>,
	linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
	chrome-platform@...ts.linux.dev, linux-kselftest@...r.kernel.org,
	Laurent Pinchart <laurent.pinchart@...asonboard.com>,
	Bartosz Golaszewski <brgl@...ev.pl>,
	Wolfram Sang <wsa+renesas@...g-engineering.com>,
	Simona Vetter <simona.vetter@...ll.ch>,
	Dan Williams <dan.j.williams@...el.com>
Subject: Re: [PATCH v6 1/3] revocable: Add fops replacement

On Mon, Nov 10, 2025 at 06:28:17AM +0000, Tzung-Bi Shih wrote:
> > It just means the user of this needs to understand there are
> > limitations on what release can do. Usually release just frees memory,
> > that is fine.
> > 
> > I think it would be strange for a release to touch revocable data,
> > that might suggest some larger problem.
> 
> I think it'd be inevitable for accessing some devm memory in ->release(),
> e.g. [1].
> 
> [1] https://elixir.bootlin.com/linux/v6.17/source/drivers/platform/chrome/cros_ec_chardev.c#L260

Again, that's symptomatic of a "larger problem" :\

The blocking notifier registration to receive events to relay out to
the file descriptors should be part this drivers probe/remove
lifecycle so it has the proper lifetime. Trying to use a revocable
idea here will create a weirdly unbalanced/dangling blocking notifier
registration :(

Inside cros_ec_chardev it should have a simple list of open file
descriptors and the single blocking notifier callback can iterate over
the list and memdup the events.  Then your release is just a simple
locked list del on a global list which doesn't have a lifecycle
problem.

Jason