[PATCH v2] selftests: powerpc: Fix online CPU selection

[Sandipan Das]

Hi Srikar, Michael,

On 29/07/20 9:33 pm, Srikar Dronamraju wrote:
> * Sandipan Das <sandipan at linux.ibm.com> [2020-06-09 13:07:33]:
> 
>> The size of the CPU affinity mask must be large enough for
>> systems with a very large number of CPUs. Otherwise, tests
>> which try to determine the first online CPU by calling
>> sched_getaffinity() will fail. This makes sure that the size
>> of the allocated affinity mask is dependent on the number of
>> CPUs as reported by get_nprocs().
>>
>> Fixes: 3752e453f6ba ("selftests/powerpc: Add tests of PMU EBBs")
>> Reported-by: Shirisha Ganta <shiganta at in.ibm.com>
>> Signed-off-by: Sandipan Das <sandipan at linux.ibm.com>
>> Reviewed-by: Kamalesh Babulal <kamalesh at linux.vnet.ibm.com>
>> ---
>> Previous versions can be found at:
>> v1: https://lore.kernel.org/linuxppc-dev/20200608144212.985144-1-sandipan@linux.ibm.com/
>>
>> @@ -88,28 +89,40 @@ void *get_auxv_entry(int type)
>>
>>  int pick_online_cpu(void)
>>  {
>> -	cpu_set_t mask;
>> -	int cpu;
>> +	int ncpus, cpu = -1;
>> +	cpu_set_t *mask;
>> +	size_t size;
>> +
>> +	ncpus = get_nprocs();
> 
> Please use get_nprocs_conf or sysconf(_SC_NPROCESSORS_CONF). The manpage
> seems to suggest the latter. Not sure how accurate the manpage is.
> 
> get_nprocs is returning online cpus and when smt is off, the cpu numbers
> would be sparse and hence the result from get_nprocs wouldn't be ideal for
> allocating cpumask. However get_nprocs_conf would return the max configured
> cpus and would be able to handle it. 
> 
> I think this was the same situation hit by Michael Ellerman.
> 

Yes, that seems to be the case. Thanks for testing this.
Will fix this in v3.

- Sandipan