This is the mail archive of the systemtap@sources.redhat.com mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Minutes of 5/12/05 meeting


Minutes of SystemTAP Conference Call, May 12, 2005

Participants:
IBM:
	Larry Kessler (lkessler@us.ibm.com -- Beaverton, OR)
	Vara Prasad (varap@us.ibm.com -- Beaverton)
	Jim Keniston (jkenisto@us.ibm.com -- Beaverton)
	Hien Nguyen (nguyhien@us.ibm.com -- Beaverton)
	Ananth Mavinakayanahalli (amavin@redhat.com, ananth@in.ibm.com --
Boston)
	Prasanna Panchamukhi (prasanna@in.ibm.com -- Bangalore)
	Maneesh Soni (maneesh@in.ibm.com -- Bangalore)
Red Hat:
	Will Cohen (wcohen@redhat.com -- Raleigh)
	Martin Hunt (hunt@redhat.com -- Seattle)
	Elena Zannoni (ezannoni@redhat.com -- Boston)
	Andrew Cagney (cagney@redhat.com -- Toronto)
Intel:
	Brad Chen (brad.chen@intel.com -- Santa Clara)
	Rusty Lynch (rusty.lynch@intel.com -- Hillsboro, OR)
	Anil KeshavaMurthy (anil.s.keshavamurthy@intel.com -- Hillsboro)
	K. Sridharan (k.sridharan@intel.com -- Hillsboro)
	Charles Spirakis (charles.spirakis@intel.com)

Actions Required (ARs):
>From 4/28/05:
1. All: Review recent changes to archpaper/systemtap.tex.

2. All: Give Brad suggestions re: the test plan.

>From 5/5/05:
10. Will, etc: Chime in on LKML in support of return probes.
[Update: Hien's return-probes patch is in v2.6.12-rc4-mm1.]

11. Hien: Notify Rusty when we feel return probes is stable enough
to start the x86_64 port. [DONE]

13. kprobes authors: Check in patches and accompanying test scripts.
[Some done.  Will remove this AR next week.]

14. Jim: Check the assertion that kprobes handlers currently run
with interrupts enabled.  [DONE for i386.  They're disabled.]

15. Rohit/all: If you want to have interrupts enabled while kprobes
handlers are run, make a case for it.

16. Frank: Propose/implement a mechanism for keeping the website
link pointing at a fresh copy of the architecture paper.
[DONE.  No easy magic.  Just update manually after major changes.]

>From 5/12/05:
17. Vara: Publish new call-in number via private email.

18. All: Verify that your email address, as published in these
minutes, is correct.

Elena has added bugs for the various required kernel patches to
bugzilla.

Since RHEL U2 will be based on U1, it's not sufficient to get our
kprobes patches accepted into Linus's kernel; they must also be
backported to U1.  Ananth will help here.

The rest of the meeting was a roundtable status update.

In the patches directory in CVS, Rusty has checked in a patch to move
arming/disarming of kprobes out of the architecture-independent code.

The IA64 port of kprobes (including jprobes) needs some work on
handling probepoints set on branch instructions, but it seems to be
working well otherwise.  More testing is needed.

The x86_64 return-probes port is progressing.

It was suggested that we need to move away from ad hoc testing
(into something broader and more turn-key?).

There was a request once again for documentation on known kprobes
bugs (any not being addressed?) and other limitations.

All kprobes patches that were in the -mm kernel when v2.6.12-rc4
was created were included in v2.6.12-rc4.

Among other things, Prasanna is porting the
don't-permanently-disarm-probe-on-collision patch to architectures
other than i386.

Ananth's ppc64 port of return probes seems to be working well.
More testing is needed.  He hopes to post this by Monday.

Brad is working on the test plan.  Progress is being made on an
event-based sampling tapset.

Martin has been enhancing the runtime.  Histogram support has been
added, and associative arrays can handle up to 5 keys.  He'll check
in this work soon.  He still needs to look at the enhancements that
Tom provided on 5/6/05.

Will has been working on the OLS paper and bringing some testers
(?) up to speed on SystemTap.

Hien posted return probes on LKML; it has been pulled into the
-mm kernel.

Jim's patch to run kprobes handlers on a separate stack (i386
only) has been tested and posted to the patches directory in CVS.
Speculation notwithstanding, handlers do indeed run with interrupts
disabled on i386.

Vara is reworking his C-tapset ideas.

Tom is on vacation.  Relayfs seems to work, says Martin, who will
soon undertake netlink-vs.-relayfs performance testing.

Did Frank automate keeping the website link to the architecture paper
fresh?  Nobody knew.  Jim agreed to check in changes to the kprobes
sections and see what happens.

The call-in number we have been using will go away in a couple of
weeks.  Vara will publish the new number via private email.  Please
verify that your email, as published in these minutes, is correct.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]