On 29.08.2017 19:27, Greg Rose wrote:
> On 08/28/2017 12:08 AM, Ilya Maximets wrote:
>> TIME_UPDATE_INTERVAL was removed long time ago.
>> Now each call leads to time update via syscall and it's
>> granularity is system dependent.
>>
>> Fixes: 31ef9f5178de ("timeval: Remove CACHE_TIME scheme.")
>> Signed-off-by: Ilya Maximets <[email protected]>
>> ---
>>   lib/timeval.c | 10 ++++------
>>   1 file changed, 4 insertions(+), 6 deletions(-)
>>
>> diff --git a/lib/timeval.c b/lib/timeval.c
>> index dd63f03..2c7f43a 100644
>> --- a/lib/timeval.c
>> +++ b/lib/timeval.c
>> @@ -171,16 +171,14 @@ time_timespec__(struct clock *c, struct timespec *ts)
>>       }
>>   }
>>
>> -/* Stores a monotonic timer, accurate within TIME_UPDATE_INTERVAL ms, into
>> - * '*ts'. */
>> +/* Stores a monotonic timer into '*ts'. */
>>   void
>>   time_timespec(struct timespec *ts)
>>   {
>>       time_timespec__(&monotonic_clock, ts);
>>   }
>>
>> -/* Stores the current time, accurate within TIME_UPDATE_INTERVAL ms, into
>> - * '*ts'. */
>> +/* Stores the current time into '*ts'. */
>>   void
>>   time_wall_timespec(struct timespec *ts)
>>   {
>> @@ -219,14 +217,14 @@ time_msec__(struct clock *c)
>>       return timespec_to_msec(&ts);
>>   }
>>
>> -/* Returns a monotonic timer, in ms (within TIME_UPDATE_INTERVAL ms). */
>> +/* Returns a monotonic timer, in ms. */
>>   long long int
>>   time_msec(void)
>>   {
>>       return time_msec__(&monotonic_clock);
>>   }
>>
>> -/* Returns the current time, in ms (within TIME_UPDATE_INTERVAL ms). */
>> +/* Returns the current time, in ms. */
>>   long long int
>>   time_wall_msec(void)
>>   {
>>
> How about replaciing "accurate within TIME_UPDATE_INTERVAL ms" with
> "accurate within the system dependent timer interval"?
> 
> I'm concerned about leaving the wrong impression with the change in comments.
> 
> Thanks,
> 
> - Greg

Hi Greg,

Thanks for looking at this, but I don't think that such a comment is useful
because I believe that any system nowadays is able to provide millisecond
granularity.

Best regards, Ilya Maximets.
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to