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: <20201127024052.GB126383@T590>
Date:   Fri, 27 Nov 2020 10:40:52 +0800
From:   Ming Lei <ming.lei@...hat.com>
To:     Daniel Wagner <dwagner@...e.de>
Cc:     linux-block@...r.kernel.org, Jens Axboe <axboe@...nel.dk>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] blk-mq: Make running from the wrong CPU less scary

On Thu, Nov 26, 2020 at 10:51:52AM +0100, Daniel Wagner wrote:
> The current warning looks aweful like a proper crash. This is
> confusing. There is not much information to gained from the stack
> trace anyway, let's drop it.
> 
> While at it print the cpumask as there might be additial helpful
> information when debugging the sitation.
> 
> Signed-off-by: Daniel Wagner <dwagner@...e.de>
> ---
> Hi,
> 
> We got a report from a customer because he was concerned about the log
> entries. As it turns out, it fooled me too to be honest. What do you
> think about making it a bit less look-a-like a kernel oops?
> 
> 
>  smpboot: Booting Node 0 Processor 12 APIC 0x26                                                
> WARNING, didn't collect load info for all cpus, balancing is broken                            
>  run queue from wrong CPU 0, hctx active                                                       
>  CPU: 0 PID: 42300 Comm: kworker/13:2H Kdump: loaded Tainted: G           OE  X    5.3.18-109.$
>  Hardware name: IBM System x3650 M5 -[5462AC1]-/00KG915, BIOS -[TCE144J-3.11]- 12/03/2019      
>  Workqueue: kblockd blk_mq_run_work_fn                                                         
>  Call Trace:                                                                                   
>   dump_stack+0x66/0x8b                                                                         
>   __blk_mq_run_hw_queue+0xee/0x100                                                             
>   process_one_work+0x1f4/0x3e0                                                                 
>   worker_thread+0x2d/0x3e0                                                                     
>   ? process_one_work+0x3e0/0x3e0                                                               
>   kthread+0x10d/0x130                                                                          
>   ? kthread_park+0xa0/0xa0                                                                     
>   ret_from_fork+0x35/0x40                                                                      
>  run queue from wrong CPU 0, hctx active                                                       
>  CPU: 0 PID: 42300 Comm: kworker/13:2H Kdump: loaded Tainted: G           OE  X    5.3.18-109.$
>  Hardware name: IBM System x3650 M5 -[5462AC1]-/00KG915, BIOS -[TCE144J-3.11]- 12/03/2019      
>  Workqueue: kblockd blk_mq_run_work_fn    
> 
> 
> Thanks,
> Daniel
> 
>  block/blk-mq.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 55bcee5dc032..0427b719d9c4 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1514,10 +1514,8 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
>  	 */
>  	if (!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask) &&
>  		cpu_online(hctx->next_cpu)) {
> -		printk(KERN_WARNING "run queue from wrong CPU %d, hctx %s\n",
> -			raw_smp_processor_id(),
> -			cpumask_empty(hctx->cpumask) ? "inactive": "active");
> -		dump_stack();
> +		printk(KERN_WARNING "run queue from wrong CPU %d, hctx %*pbl\n",
> +			raw_smp_processor_id(), cpumask_pr_args(hctx->cpumask));
>  	}

Now we have guaranteed that no any requests originated from one hctx exists
when this hctx is going to offline, which is strong enough for killing the check.

The reason why such warning is triggered is that wq's cpu hot unplug is
handled before blk-mq's handling.

I'd suggest to kill the whole branch in the fast path.

Thanks,
Ming

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ