16:22:21 <tsoome> huh, I was wondering why this patch does not receive any feedback from advocates.... So, I found out, I haven't done RTI build and therefore also haven't mailed the request... sigh.
17:52:07 <gwr> Any dtrace experts around to advise me?  I'm looking into a performance issue where I'm measuring the run time of a function with dtrace FBT entry/return probes, collecting an aggregation of run times base on return timestamp minus entry timestamp.
17:52:59 <gwr> The resulting distribution is bi-modal, with the majority of samples in an expected range, and a smaller group taking much, much longer than the majority group.
17:53:38 <gwr> What's a good way to get details on the calls below this function, just for the minority cases that take longer?
17:55:16 <jclulow> You might try speculative tracing
17:55:35 <jclulow> https://illumos.org/books/dtrace/chp-spec.html#chp-spec
17:56:04 <jclulow> gwr: ^^
18:01:55 <gwr> thanks
19:09:04 <gitomat> [illumos-gate] 15089 WPTS BVT_TreeMgmt_SMB311_Disconnect_NoSignedNoEncryptedTreeConnect -- Gordon Ross <gwr⊙rc>