Subject: Re: [PATCH] workqueue: Ensure that cpumask set for pools created
 after boot
To: Tejun Heo <tj@kernel.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>, linux-kernel@vger.kernel.org,
        Nathan Fontenot <nfont@linux.vnet.ibm.com>
References: <20170525150353.GE23493@htj.duckdns.org>
 <20170525150752.GF23493@htj.duckdns.org>
 <f0723547-1383-efa2-a691-eb91ff0620c4@linux.vnet.ibm.com>
 <d4eaef3a-7371-a9e6-d371-2b2dd5703df8@linux.vnet.ibm.com>
 <20170606180913.GA32062@htj.duckdns.org>
 <736f7f6e-8d47-eaea-acc6-8ed75014a287@linux.vnet.ibm.com>
 <20170612161433.GB19206@htj.duckdns.org>
 <69c4bbad-5d40-d054-0004-38ac81377b0b@linux.vnet.ibm.com>
 <20170612173225.GD19206@htj.duckdns.org>
 <0b95fc96-a481-4439-de65-ffdffa207f47@linux.vnet.ibm.com>
 <20170613201029.GA7359@htj.duckdns.org>
From: Michael Bringmann <mwb@linux.vnet.ibm.com>
Organization: IBM Linux Technology Center
Date: Wed, 28 Jun 2017 16:15:09 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.1.0
MIME-Version: 1.0
In-Reply-To: <20170613201029.GA7359@htj.duckdns.org>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Message-Id: <acd1d7b8-074b-6a60-bae0-43293461b135@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1821
Lines: 56

I will try that patch tomorrow.  My only concern about that is the use of WARN_ON().
As I may have mentioned in my note of 6/27, I saw about 600 instances of the warning
message just during boot of the PowerPC kernel.  I doubt that we want to see that on
an ongoing basis.

Michael

On 06/13/2017 03:10 PM, Tejun Heo wrote:
> Hello,
> 
> On Tue, Jun 13, 2017 at 03:04:30PM -0500, Michael Bringmann wrote:
>> @@ -3564,19 +3564,28 @@ static struct pool_workqueue *alloc_unbound_pwq(struct workqueue_struct *wq,
>>  static bool wq_calc_node_cpumask(const struct workqueue_attrs *attrs, int node,
>>  				 int cpu_going_down, cpumask_t *cpumask)
>>  {
>> +	cpumask_t	onl_targ_cm;
>> +
>>  	if (!wq_numa_enabled || attrs->no_numa)
>>  		goto use_dfl;
>>  
>>  	/* does @node have any online CPUs @attrs wants? */
>> -	cpumask_and(cpumask, cpumask_of_node(node), attrs->cpumask);
>> +	cpumask_and(&onl_targ_cm, cpumask_of_node(node), attrs->cpumask);
>>  	if (cpu_going_down >= 0)
>> -		cpumask_clear_cpu(cpu_going_down, cpumask);
>> +		cpumask_clear_cpu(cpu_going_down, &onl_targ_cm);
>>  
>> -	if (cpumask_empty(cpumask))
>> +	if (cpumask_empty(&onl_targ_cm))
>>  		goto use_dfl;
>>  
>>  	/* yeap, return possible CPUs in @node that @attrs wants */
>>  	cpumask_and(cpumask, attrs->cpumask, wq_numa_possible_cpumask[node]);
>> +
>> +	if (cpumask_weight(&onl_targ_cm) > cpumask_weight(cpumask)) {
>> +		printk(KERN_INFO "WARNING: WQ cpumask: onl intersect > "
>> +						"possible intersect\n");
>> +		return false;
>> +	}
> 
> Yeah, alternatively, we can just add right before returning,
> 
> 	if (WARN_ON(cpumask_empty(cpumask)))
> 		return false;
> 
> Thanks.
> 

-- 
Michael W. Bringmann
Linux Technology Center
IBM Corporation
Tie-Line  363-5196
External: (512) 286-5196
Cell:       (512) 466-0650
mwb@linux.vnet.ibm.com