Hi,
On Wed, Aug 07, 2024 at 09:36:59AM -0500, Nathan Bossart wrote:
> Also, do we need to worry about overflow here? It looks like the rest of
> instr_time.h is oblivious about overflow, so maybe this is better discussed
> in a separate thread...
Yeah, a separate thread would be better.
FWIW and just out of curiosity:
1. it seems to me that most of the time (always?) we are manipulating instr_time(s)
as interval(s) which (with int64) gives “space” for about 292 years interval time
measurement (if my maths are correct).
2. for "absolute" manipulation (if any) it would depend of the PG_INSTR_CLOCK.
A "man clock_gettime" mentions:
2.1 CLOCK_MONOTONIC: on Linux, time since the system was booted. Not sure what
the longest Linux uptime record is but can't be more than since the 90's.
2.2 CLOCK_REALTIME: Its time represents seconds and nanoseconds since the Epoch.
It means that we’re currently about 237 years away from the limit. So even,
if we were to say add 2 "recents" of them we are still about 183 years away from
the limit.
Regards,
--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com
