This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: kernel function probe overhead
- From: Cody Santing <csanting at redhat dot com>
- To: Aubrey Li <aubreylee at gmail dot com>, "Frank Ch. Eigler" <fche at redhat dot com>, Josh Stone <jistone at redhat dot com>
- Cc: systemtap at sourceware dot org
- Date: Thu, 1 Dec 2016 13:47:00 -0500
- Subject: Re: kernel function probe overhead
- Authentication-results: sourceware.org; auth=none
- References: <CAGPKeUKWAvBv=SL_ebVkR3ZwCeGZoCQ1MEJ5xQi-VXuF=kt3Wg@mail.gmail.com> <y0m1sxzpj9i.fsf@fche.csb> <CAGPKeUJjf5qqxO2Gqo-_LzzaFbwQ=-wZ=sq0fwTzm=c+MEAtSw@mail.gmail.com> <y0mmvgjo1jl.fsf@fche.csb> <CAGPKeULiGUPMQ6e-YjSY3MTW+r=rjxYHYmoKEb1sU4DN+SL1Lg@mail.gmail.com>
Hello Aubrey,
>> Ah, I note that we don't track occurrences of lock contention,
>> even though we could. Opened https://sourceware.org/bugzilla/show_bug.cgi?id=20879
>
> okay, thanks, I'll keep an eye on this ticket.
>
I have yet to update the ticket but I have added tracking of lock
contention. The number of lock contention occurrences on global
variables will be printed following the probe hit report from the '-t'
option.
>> That inconsistency dramatic contention or caching or some other source
>> of jitter. Having the variance figure in there would be useful though.
>> Would anyone like to give that a try? (translate.cxx line 2283ish and
>> runtime/linux/common_session_state.h _stp_stat_init area).
>>
I have also added variance to the '-t' option. It is immediately
following min/avg/max of each probe.
>
> If there is a patch delivered against git head, I can apply it and
> paste the output here.
>
You should find the patch there. Give it a shot, if you have any
feedback I would like to hear it.
Thank you,
Cody Santing