linux-kernel - Re: RFC: SysRq nice-all-RT-tasks is broken

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: <0478c913-d93a-73a0-9d52-5bf60fe54ee4@linux.vnet.ibm.com>
Date:   Thu, 9 Mar 2017 10:02:04 +0100
From:   Laurent Dufour <ldufour@...ux.vnet.ibm.com>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Andrew Morton <akpm@...l.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...e.hu>,
        Peter Zijlstra <peterz@...radead.org>
Subject: Re: RFC: SysRq nice-all-RT-tasks is broken

On 08/03/2017 18:46, Steven Rostedt wrote:
> On Wed, 8 Mar 2017 12:40:12 -0500
> Steven Rostedt <rostedt@...dmis.org> wrote:
> 
>> I wonder if we should just have a special flag sent by that sysrq
>> trigger. Since it is causing all tasks to go "nice" there's no need to
>> do the pi chain walk in __sched_setscheduler().
> 
> Hah, there already is a flag!
> 
> Laurent, can you test this patch:

Tested on ppc64, no more panic \o/

FWIW,
Tested-by: Laurent Dufour <ldufour@...ux.vnet.ibm.com>

Thanks,
Laurent.

> 
> -- Steve
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 3b31fc0..7292fa9 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4129,8 +4129,8 @@ static int __sched_setscheduler(struct task_struct *p,
>  	int queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE;
>  	struct rq *rq;
> 
> -	/* May grab non-irq protected spin_locks: */
> -	BUG_ON(in_interrupt());
> +	/* The pi code expects interrupts enabled */
> +	BUG_ON(pi && in_interrupt());
>  recheck:
>  	/* Double check policy once rq lock held: */
>  	if (policy < 0) {
>