Re: [Patch 3] thread/core siblings info for guests

[Avi Kivity <avi@redhat.com>]

Kamble, Nitin A wrote:
>> The capability itself can return the count; for example
>>
>>  case KVM_CAP_NR_CPUID_LEAVES:
>>      return KVM_MAX_CPUID_ENTRIES;
>>
>> which is simpler to use and shorter.
>>     
>
> Avi,
>    Yes, it is simpler and shorter, but is returning a constant. It will be wasting space for the unused entries. Also it put's a restriction that the list count can not go more than that.
>    The dynamic size finding, the patch I sent, is not complicated or long to be an issue. Would you like to give another consideration for the same patch (patch-3) I sent earlier?
>
>   

Oh, I confused the sizing code with something else.

We can add this, but older kernels will still miss the code so we have 
to work around it.  I guess userspace can start with a large number, and 
double it each time the ioctl fails.

-- 
error compiling committee.c: too many arguments to function