linux-kernel - Re: [PATCH 1/2] stop_machine: check work->done while handling enqueued works

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: <20130209190826.GC2875@htj.dyndns.org>
Date:	Sat, 9 Feb 2013 11:08:26 -0800
From:	Tejun Heo <tj@...nel.org>
To:	Hillf Danton <dhillf@...il.com>
Cc:	Rusty Russell <rusty@...tcorp.com.au>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Ingo Molnar <mingo@...e.hu>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] stop_machine: check work->done while handling
 enqueued works

Hello, Hillf.

On Fri, Feb 08, 2013 at 11:39:56AM +0800, Hillf Danton wrote:
> The comment just above cpu_stop_signal_done() says it is uncertain that
> the input @done is valid, and the works enqueued through the function
> stop_one_cpu_nowait() do carry no done, thus we have to check if it is
> valid when updating work result.

How about something like the following?

In cpu_stopper_thread(), @work->done may be NULL if the cpu stop work
is queued from stop_one_cpu_nowait(); however, cpu_stopper_thread()
updates @done->ret without checking whether @done exists or not when
the work function fails.

While this can lead to oops, the only current user of
stop_one_cpu_nowait() - active_load_balance_cpu_stop() - always
returns 0 and thus there's no in-kernel user which triggers this bug.

Fix it by checking whether @done exists before updating @done->ret.

> Signed-off-by: Hillf Danton <dhillf@...il.com>
> ---
> 
> --- a/kernel/stop_machine.c	Thu Feb  7 20:03:10 2013
> +++ b/kernel/stop_machine.c	Fri Feb  8 11:07:40 2013
> @@ -279,7 +279,7 @@ repeat:
>  		preempt_disable();
> 
>  		ret = fn(arg);
> -		if (ret)
> +		if (ret && done != NULL)

It's a nitpick and probalby is just a preference but I've never liked
!= NULL or != 0.  Can we just do if (ret && done)?

Thanks.

-- 
tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/