On 13/01/2016, Erwan Velu wrote: [snip] > You consider the MONOTONIC clock drifting a lot ? > I mean an NTP adjustement can make a serious jump forward or even worse backward. So, CLOCK_MONOTONIC will never jump backward (thus the name). And there is a system-defined limit on how much it will ever jump forward at any given time, though I don't know what that value is offhand. On a well-synchronized system, at least it should give more accurate durations than CLOCK_MONOTONIC_RAW. That said, if you're coming up from a cold boot and you're out of sync, it could be jumpier. As I said, I am not a Timelord, this is just what I've gathered as received wisdom. [snip] > You mean we could switch some of ceph_clock_now() calls to ceph_time ? > We could use coarse_mono_clock to perform that. > > If you agree that switching code like the one I'm speaking is valuable, I can work on it. I think that would be very valuable, thank you. -- Senior Software Engineer Red Hat Storage, Ann Arbor, MI, US IRC: Aemerson@{RedHat, OFTC, Freenode} 0x80F7544B90EDBFB9 E707 86BA 0C1B 62CC 152C 7C12 80F7 544B 90ED BFB9 -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html