LTTng 0.6.36 for 2.6.18 (now with markers) [LWN.net]

From:		Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
To:		linux-kernel@vger.kernel.org, Christoph Hellwig <hch@infradead.org>, Andrew Morton <akpm@osdl.org>, Ingo Molnar <mingo@redhat.com>, Greg Kroah-Hartman <gregkh@suse.de>, Thomas Gleixner <tglx@linutronix.de>, Tom Zanussi <zanussi@us.ibm.com>, Karim Yaghmour <karim@opersys.com>, Paul Mundt <lethal@linux-sh.org>, Jes Sorensen <jes@sgi.com>, Richard J Moore <richardj_moore@uk.ibm.com>, "Martin J. Bligh" <mbligh@mbligh.org>, Michel Dagenais <michel.dagenais@polymtl.ca>, Douglas Niehaus <niehaus@eecs.ku.edu>, ltt-dev@shafik.org, systemtap@sources.redhat.com
Subject:		[PATCH] LTTng 0.6.36 for 2.6.18 (now with markers)
Date:		Fri, 24 Nov 2006 16:48:31 -0500
Archive-link:		Article, Thread

Hi,

I have, since a few weeks, moved LTTng to the markers infrastructure. I
left the ltt-dev users and contributors test the markers and LTTng on various
architectures (i386, x86_64, PowerPC, -ppc, ARM, MIPS) before posting it on
LKML.

The most important new features since the my post :

- Use DebugFS
- Use the "Markers" infrastructure (updated since the last post on LKML).
- Dynamically loadable "probes" that connects to the markers.
- CPU Hotplug support (this piece seemed necessary for the Xen port I am
  currently working on)
- Use of per-CPU atomic operations even on SMP machines (no lock prefix, no
  memory barriers) to update the per-cpu counters. An explicit smp_wmb()
  is used at the one place where the subbuffers are tagged "full" and
  smp_rmb() is used in the buffer consumer just after it reads this counter
  indicating that the subbuffer is full.

I am not submitting the probes themselves, as they can be provided as separate
kernel modules.

Comments and constructive criticism are, as always, welcome.

The patches follow.

Mathieu


OpenPGP public key:              http://krystal.dyndns.org:8080/key/compudj.gpg
Key fingerprint:     8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68