Message 142856 - Python tracker

Message142856

Author	neologix
Recipients	methane, neologix, vstinner
Date	2011-08-23.18:54:28
SpamBayes Score	3.5602568e-07
Marked as misclassified	No
Message-id	<1314125669.3.0.498159085449.issue12822@psf.upfronthosting.co.za>
In-reply-to

Content
> Using CLOCK_MONOTONIC is better than CLOCK_REALTIME (default) for GIL > because settimeofday() may break the pthread_cond_timedwait(). Indeed. A couple remarks: - regular locks and conditions variables exposed by the threading module suffer from the same problem - POSIX semaphores are also affected, but you can't select an alternative clock source - actually, CLOCK_MONOTONIC is affected by NTP adjustements: while it's guaranteed not to go backward, its rate can be affected Did you really encounter this problem, or is this just a theoretical concern?

Content

> Using CLOCK_MONOTONIC is better than CLOCK_REALTIME (default) for GIL
> because settimeofday() may break the pthread_cond_timedwait().

Indeed.
A couple remarks:
- regular locks and conditions variables exposed by the threading module suffer from the same problem
- POSIX semaphores are also affected, but you can't select an alternative clock source
- actually, CLOCK_MONOTONIC is affected by NTP adjustements: while it's guaranteed not to go backward, its rate can be affected

Did you really encounter this problem, or is this just a theoretical concern?

History
Date	User	Action	Args
2011-08-23 18:54:29	neologix	set	recipients: + neologix, vstinner, methane
2011-08-23 18:54:29	neologix	set	messageid: <1314125669.3.0.498159085449.issue12822@psf.upfronthosting.co.za>
2011-08-23 18:54:28	neologix	link	issue12822 messages
2011-08-23 18:54:28	neologix	create