Hi, On Thu, Jul 11, 2024 at 10:15:41AM -0500, Sami Imseih wrote: > > > I did a few tests with the patch and did not see any "large" drifts like the > > ones observed above. > > Thanks for testing. > > > I think it could be "simplified" by making use of instr_time instead of > > timespec > > for current and absolute. Then I think it would be enough to compare their > > ticks. > > Correct I attached a v2 of this patch that uses instr_time to check the > elapsed > time and break out of the loop. It needs some more benchmarking.
Thanks! Outside of Nathan's comment: 1 === + * However, since nanosleep is susceptible to time drift when interrupted + * frequently, we add a safeguard to break out of the nanosleep whenever the I'm not sure that "nanosleep is susceptible to time drift when interrupted frequently" is a correct wording. What about? " However, since the time between frequent interruptions and restarts of the nanosleep calls can substantially lead to drift in the time when the sleep finally completes, we add...." 2 === +static void vacuum_sleep(double msec); What about a more generic name that could be used outside of vacuum? Regards, -- Bertrand Drouvot PostgreSQL Contributors Team RDS Open Source Databases Amazon Web Services: https://aws.amazon.com
