This is the mail archive of the
systemtap@sources.redhat.com
mailing list for the systemtap project.
Re: Minutes of 6/2/05 meeting
- From: Karim Yaghmour <karim at opersys dot com>
- To: Jim Keniston <jkenisto at us dot ibm dot com>
- Cc: SystemTAP <systemtap at sources dot redhat dot com>
- Date: Mon, 06 Jun 2005 20:23:00 -0400
- Subject: Re: Minutes of 6/2/05 meeting
- Organization: Opersys inc.
- References: <1118098057.2784.139.camel@dyn9047018078.beaverton.ibm.com>
- Reply-to: karim at opersys dot com
Jim Keniston wrote:
> Martin has been working on runtime support for aggregations and
> statistics. He has not yet done any relayfs-vs.-netlink performance
> testing.
Note that there are authors of the netlink RFC that readily acknowledge
that netlink and relayfs serve very different purposes. Such a
comparison entertains a confusion which has been dissipated quite some
time ago.
Here's a direct quote from Andi Kleen back when Andrew first put relayfs
in 2.6.11-rc1-mm1:
> imho relayfs and netlink are for completely problem spaces.
> relayfs is for relaying a lot of data quickly (e.g. for kernel
> instrumentation). There it fills a niche that printk doesn't fill
> (since it's too slow). netlink is quite slow (allocates data for each
> event, does lots of other gunk), but an useful extensible format
> for low frequency events.
Andi then goes on to list a couple of things he thinks need to be
fixed with relayfs. Since then, these things have been fixed thanks
to Tom Zanussi's continued work.
Original found here:
http://marc.theaimsgroup.com/?l=linux-kernel&m=110569261401528&w=2
At present, ltt is using relayfs, but it will also use netlink in
order to replace all the kernel/user-space communication which is
currently implemented in all sorts of ways which are not very
elegant (/proc, ioctls, etc.)
Best regards,
Karim
--
Author, Speaker, Developer, Consultant
Pushing Embedded and Real-Time Linux Systems Beyond the Limits
http://www.opersys.com || karim@opersys.com || 1-866-677-4546