Re: [Xen-devel] [PATCH 2/2] xen: fill topology info for online cpus only

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 30.08.18 at 10:31, <jgross@xxxxxxxx> wrote:
> On 30/08/18 10:16, Jan Beulich wrote:
>>>>> On 29.08.18 at 20:23, <jgross@xxxxxxxx> wrote:
>>> The topology information obtainable via XEN_SYSCTL_cputopoinfo is
>>> filled rather weird: the size of the array is derived from the highest
>>> online cpu number, while the data is set to "invalid" for not present
>>> cpus only.
>>>
>>> With smt=0 the information for parked threads is all zero, so it should
>>> be best to return "invalid" information for offline cpus.
>>>
>>> On a dual core system without this patch xl info -n will print:
>>>
>>> cpu_topology           :
>>> cpu:    core    socket     node
>>>   0:       0        0        0
>>>   1:       0        0        0
>>>   2:       1        0        0
>> 
>> But there's nothing wrong here. The interesting part is what would be
>> printed for CPU 3 (perhaps on a more than two cores system). After
>> all topology is valid irrespective of whether a CPU is online - it all
>> depends on whether the hypervisor still has the information available.
>> It is for a reason that cpu_smpboot_free() invalidates certain fields
>> only upon CPU removal:
>> 
>>     if ( remove )
>>     {
>>         c[cpu].phys_proc_id = XEN_INVALID_SOCKET_ID;
>>         c[cpu].cpu_core_id = XEN_INVALID_CORE_ID;
>>         c[cpu].compute_unit_id = INVALID_CUID;
>> 
>> On a 6-core system I see
>> 
>> cpu:    core    socket     node
>>   0:       0        0        0
>>   1:       0        0        0
>>   2:       1        0        0
>>   3:       1        0        0
>>   4:       2        0        0
>>   5:       2        0        0
>>   6:       8        0        0
>>   7:       8        0        0
>>   8:       9        0        0
>>   9:       9        0        0
>>  10:      10        0        0
>> 
>> which looks fine to me, apart from the missing info on CPU 11.
> 
> I can change the patch to print the information for the offline cpus
> (including the now missing ones), too.
> 
> What is the preferred solution?

Well, by implication from my earlier reply I think adding the missing
CPU's info would be better. Let's see what others think.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel