On 07/17/2018 08:25 PM, Huang, Ying wrote:
>> Seriously, though, does it hurt us to add a comment or two to say
>> something like:
>>
>>      /*
>>       * Should not even be attempting cluster allocations when
>>       * huge page swap is disabled.  Warn and fail the allocation.
>>       */
>>      if (!IS_ENABLED(CONFIG_THP_SWAP)) {
>>              VM_WARN_ON_ONCE(1);
>>              return 0;
>>      }
> I totally agree with you that we should add more comments for THP swap
> to improve the code readability.  As for this specific case,
> VM_WARN_ON_ONCE() here is just to capture some programming error during
> development.  Do we really need comments here?

If it's code in mainline, we need to know what it is doing.

If it's not useful to have in mainline, then let's remove it.

Reply via email to