linux-kernel - Re: [PATCH V4 for-6.10/block] loop: Fix a race between loop detach and loop open

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: <20240612051940.GA27294@lst.de>
Date: Wed, 12 Jun 2024 07:19:40 +0200
From: Christoph Hellwig <hch@....de>
To: Gulam Mohamed <gulam.mohamed@...cle.com>
Cc: linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
	yukuai1@...weicloud.com, hch@....de, axboe@...nel.dk
Subject: Re: [PATCH V4 for-6.10/block] loop: Fix a race between loop detach
 and loop open

On Fri, Jun 07, 2024 at 07:06:07PM +0000, Gulam Mohamed wrote:
> Setting the lo_state to Lo_rundown in loop_clr_fd() may not help in
> stopping the incoming open(), when the loop is being detached, as the
> open() could invoke the lo_open() before the lo_state is set to Lo_rundown
> and increment the disk_openers refcnt later.
> As the actual cleanup is deferred to last close, in release, there is no
> chance for the open() to kick in to take the reference. Because both open()
> and release() are protected by open_mutex and hence they cannot run in
> parallel.
> So, lo_open() and setting lo_state to Lo_rundown is not needed. Removing
> the loop state Lo_rundown as its not used anymore.

Looks like LTP still expects Lo_rundown to be set.