Re: [PATCH 1/14] cpumask: simplify sched_rt.c

["Gregory Haskins" <ghaskins@novell.com>]

>>> On 11/3/2009 at 11:41 AM, in message
<1257266509.26028.3343.camel@gandalf.stny.rr.com>, Steven Rostedt
<rostedt@goodmis.org> wrote: 
> Greg,
> 
> I believe this was mostly your code. Can you review this change.
> 
> Thanks,
> 
> -- Steve
> 
> 
> On Tue, 2009-11-03 at 14:53 +1030, Rusty Russell wrote:
>> find_lowest_rq() wants to call pick_optimal_cpu() on the intersection
>> of sched_domain_span(sd) and lowest_mask.  Rather than doing a cpus_and
>> into a temporary, we can open-code it.
>> 
>> This actually makes the code slightly clearer, IMHO.
>> 
>> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
>> To: Steven Rostedt <rostedt@goodmis.org>
>> Cc: Ingo Molnar <mingo@redhat.com>

I am not sure that I agree that its "clearer", but I do like the fact that the potentially nasty dynamic mask allocation is removed from the fast path.

I checked the logic, and it does appear to be functionally equivalent to the original.

Acked-by: Gregory Haskins <ghaskins@novell.com>

>> ---
>>  kernel/sched_rt.c |   59 +++++++++++++++++++++---------------------------------
>>  1 file changed, 23 insertions(+), 36 deletions(-)
>> 
>> diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c
>> --- a/kernel/sched_rt.c
>> +++ b/kernel/sched_rt.c
>> @@ -1115,29 +1115,12 @@ static struct task_struct *pick_next_hig
>>  
>>  static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask);
>>  
>> -static inline int pick_optimal_cpu(int this_cpu,
>> -				   const struct cpumask *mask)
>> -{
>> -	int first;
>> -
>> -	/* "this_cpu" is cheaper to preempt than a remote processor */
>> -	if ((this_cpu != -1) && cpumask_test_cpu(this_cpu, mask))
>> -		return this_cpu;
>> -
>> -	first = cpumask_first(mask);
>> -	if (first < nr_cpu_ids)
>> -		return first;
>> -
>> -	return -1;
>> -}
>> -
>>  static int find_lowest_rq(struct task_struct *task)
>>  {
>>  	struct sched_domain *sd;
>>  	struct cpumask *lowest_mask = __get_cpu_var(local_cpu_mask);
>>  	int this_cpu = smp_processor_id();
>>  	int cpu      = task_cpu(task);
>> -	cpumask_var_t domain_mask;
>>  
>>  	if (task->rt.nr_cpus_allowed == 1)
>>  		return -1; /* No other targets possible */
>> @@ -1167,28 +1150,26 @@ static int find_lowest_rq(struct task_st
>>  	 * Otherwise, we consult the sched_domains span maps to figure
>>  	 * out which cpu is logically closest to our hot cache data.
>>  	 */
>> -	if (this_cpu == cpu)
>> -		this_cpu = -1; /* Skip this_cpu opt if the same */
>> +	if (!cpumask_test_cpu(this_cpu, lowest_mask))
>> +		this_cpu = -1; /* Skip this_cpu opt if not among lowest */
>>  
>> -	if (alloc_cpumask_var(&domain_mask, GFP_ATOMIC)) {
>> -		for_each_domain(cpu, sd) {
>> -			if (sd->flags & SD_WAKE_AFFINE) {
>> -				int best_cpu;
>> +	for_each_domain(cpu, sd) {
>> +		if (sd->flags & SD_WAKE_AFFINE) {
>> +			int best_cpu;
>>  
>> -				cpumask_and(domain_mask,
>> -					    sched_domain_span(sd),
>> -					    lowest_mask);
>> +			/*
>> +			 * "this_cpu" is cheaper to preempt than a
>> +			 * remote processor.
>> +			 */
>> +			if (this_cpu != -1 &&
>> +			    cpumask_test_cpu(this_cpu, sched_domain_span(sd)))
>> +				return this_cpu;
>>  
>> -				best_cpu = pick_optimal_cpu(this_cpu,
>> -							    domain_mask);
>> -
>> -				if (best_cpu != -1) {
>> -					free_cpumask_var(domain_mask);
>> -					return best_cpu;
>> -				}
>> -			}
>> +			best_cpu = cpumask_first_and(lowest_mask,
>> +						     sched_domain_span(sd));
>> +			if (best_cpu < nr_cpu_ids)
>> +				return best_cpu;
>>  		}
>> -		free_cpumask_var(domain_mask);
>>  	}
>>  
>>  	/*
>> @@ -1196,7 +1177,13 @@ static int find_lowest_rq(struct task_st
>>  	 * just give the caller *something* to work with from the compatible
>>  	 * locations.
>>  	 */
>> -	return pick_optimal_cpu(this_cpu, lowest_mask);
>> +	if (this_cpu != -1)
>> +		return this_cpu;
>> +
>> +	cpu = cpumask_any(lowest_mask);
>> +	if (cpu < nr_cpu_ids)
>> +		return cpu;
>> +	return -1;
>>  }
>>  
>>  /* Will lock the rq it finds */
>>