svn commit: r277570 - head/usr.bin/vmstat

[Alan Somers]

On Fri, Jan 23, 2015 at 2:18 PM, Davide Italiano <davide at freebsd.org> wrote:
> On Fri, Jan 23, 2015 at 12:56 PM, Will Andrews <will at freebsd.org> wrote:
>> On Fri, Jan 23, 2015 at 11:23:57AM -0800, Davide Italiano wrote:
>>> What's the rationale behind this change?
>>
>> The rationale was to match the clock used to the name (and implied intent)
>> of the variables used to store/manipulate the data.
>>
>> It's possible that 'uptime' is entirely the wrong term in vmstat, in which
>> case the naming used elsewhere in vmstat needs to be changed too.
>>
>> --
>> wca
>
> I don't think this is a reason strong enough, we should rather go the
> other way around as already suggested.
> I would like to see this reverted.
>
> Thanks,
>
> --
> Davide
>

I agree with Ian that vmstat ought to be using a monotonic clock, and
that the ideal change would be to rename all of its "uptime" variables
accordingly.  But reverting this change would be the worst possible
outcome because we'd be back in the original situation of wrong
variable names that are wrongly set.  As is, the change has no
functional effect because CLOCK_MONOTONIC and CLOCK_UPTIME are
identical (see kern_clock_gettime).

-Alan