linux-kernel - Re: docs: bug in "Krefs and RCU" example

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]

Date: Tue, 19 Mar 2024 03:36:10 +0000
From: Matthew Wilcox <willy@...radead.org>
To: Dmitry Mastykin <dmastykin@...ralinux.ru>
Cc: corbet@....net, linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
	thellstrom@...are.com, mastichi@...il.com
Subject: Re: docs: bug in "Krefs and RCU" example

On Mon, Mar 18, 2024 at 02:27:44PM +0300, Dmitry Mastykin wrote:
> Hello all,
> It seems there is a problem in "Krefs and RCU" example, that may cause
> a crash.
> I marked the place between two problem lines with "problem is here" comment.
> If list_del_rcu() will be called between these lines, and list will become
> empty, then q.next will not point to a valid struct my_data.
> entry->refcount will also be invalid.
> Instead, q.next must be read first, and then compared with q to check
> list's emptiness (for example like in list_for_each_entry_rcu macro).

I agree you've identified a problem, but there's no way we should apply
this patch that adds "problem is here"!  Instead we should show the
proper usage, which is a call to list_first_or_null_rcu().