This is the mail archive of the
mailing list for the systemtap project.
Re: Tapset state & plans
- From: Li Guanglei <guanglei at cn dot ibm dot com>
- To: "Gui,Jian" <guij at cn dot ibm dot com>
- Cc: Mike Mason <mmlnx at us dot ibm dot com>, systemtap at sources dot redhat dot com
- Date: Fri, 22 Sep 2006 18:08:28 +0800
- Subject: Re: Tapset state & plans
- Organization: IBM CSTL
- References: <4512D828.email@example.com> <45139F63.firstname.lastname@example.org>
Mike Mason wrote:
I've been asked to do is help expand and enhance systemtap's set of
tapsets. As a starting point, I have some questions regarding the
current state of things:
- What analysis has been done to determine what tapsets should be
implemented? Has anyone done a comparison between the data that dtrace
provides and what systemtap does/should provide?
- Of the tapsets already in cvs, which are considered complete and
which are incomplete. I know things like this are never truly "done",
but some must be more done than others.
- What tapsets are people in the community currently working on?
- Does a prioritized list of tapsets "to be done" exist?
Current RPC tapset mainly picks probe points from high-level
RPC client, server and scheduler functions. It doesn't cover
other components such as XDR, credential, etc. Any contribution
to this/systemtap is welcome.
And the same rule also applies to other tapsets, such as NFS and
networking. And sometimes we may be a little slow at documenting the
implemented tapsets so the first place to find what probes available
in SystemTap is the /usr/share/systemtap/tapsets directory.
The tapsets we implemented as part of LKET come mostly from
requirement of our performance team. We have almost finished their
wish list, so at present there is not a clear plan of what tapsets we
will do in the next step.