This is the mail archive of the systemtap@sourceware.org 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 8/18/05 meeting


Minutes of SystemTap Conference Call, August 18, 2005

Participants:
IBM:
	Vara Prasad (varap@us.ibm.com -- Beaverton)
	Jim Keniston (jkenisto@us.ibm.com -- Beaverton)
	Hien Nguyen (nguyhien@us.ibm.com -- Beaverton)
	Kevin Stafford (kevinrs@us.ibm.com -- Beaverton)
	Tom Zanussi (trz@us.ibm.com -- Chicago)
	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:
	Elena Zannoni (ezannoni@redhat.com -- Boston)
	Will Cohen (wcohen@redhat.com -- Raleigh)
	Martin Hunt (hunt@redhat.com -- Seattle)
	Frank Eigler (fche@redhat.com -- Toronto)
	Graydon Hoare (graydon@redhat.com)
Intel:
	Anil KeshavaMurthy (anil.s.keshavamurthy@intel.com -- Hillsboro)
	Charles Spirakis (charles.spirakis@intel.com)

Actions Required (ARs):
>From 5/19/05:
19. Vara: Collect sample kprobes modules and post them to the
SystemTap web site.

22. Rusty: Look into x86_64 limitation re: temporary disarming of
reentrant probes.

23. Jim, Martin: Ensure that the runtime's stack-trace feature works
appropriately when there are return probes in the call stack.
[Update: Martin is working on this.  Jim will prototype an improvement
whereby the saved IP is updated before the handler is called.]

>From 6/9/05:
26. Rusty/Brad: Forward email address of Intel test person to Jay
Turner.  [Update: As of 7/7/05, Brad hasn't done this.  Did Rusty?]

>From 6/23/05:
30. Rusty: Review Will's rendition of Rusty's test that exercises
probes on a wide variety of instruction types.  [Update: As of 7/7/05,
Will still needs to hear back from Rusty.]

>From 7/7/05:
39. Ananth: Summarize the status of the ppc64 runtime port for Martin.
[Update 8/17/05: Jose Santos of the IBM pSeries performance team will
complete the port.]

>From 7/28/05:
41. Anil: Post a summary of known djprobes implementation issues.

>From 8/4/05:
43. Prasanna: Post work on user-space probes.  [Update: Design is
under review at IBM.]

>From 8/11/05:
44. Tom: Automate relayfs tests and post them to mailing list.
[Update: In progress]

45. Will: Post Fedora-Core-4-compatible yum commands for updating
elfutils [DONE 8/11/05]

46. Will/Graydon: Help Anil resolve the gcc error he sees while
building SystemTap. [DONE]

>From 8/18/05:
47. Elena: Report on availability of tech writer(s) at Red Hat to
help with user documentation.

48. Elena, Vara: Manage planning of next face-to-face meeting.

49. Graydon, Frank: Publish an update to their translator feature
list of 8/4/05.

50. Jim: Characterize the problem(s) we've seen when probing
schedule().  [Update: So far, we haven't seen anything that really
indicates that we can't probe schedule().  One problem was a translator
bug regarding return probes; this is fixed.  Another shows up when we
probe do_IRQ() and schedule() concurrently on ppc64 or x86_64; this
is still being investigated.  Probes of schedule() were successful
in all other cases.]

51. Jim: Open bugzilla entry to document areas of the kernel that
aren't safe to probe.  [DONE: #1235]

52. Ananth: Post RCU-based SMP scalability patches to SystemTap
mailing list. [DONE 8/18/05]

53: Anil: Test RCU-based SMP scalability feature on ia64.

As the number of SystemTap users increases, so does the need for
user documentation.  Red Hat had identified a tech writer to support
this back in April.  Elena will see about this writer's availability.

We should do another face-to-face meeting like the 4/20/05 one
-- tentatively in late September.  Vara and Elena will discuss.
Send them suggestions about date, site, agenda.

The short-term schedule for upcoming SystemTap releases was discussed.

CVS includes a kernel-functionality test suite (tests/kernel), and
Will will soon check in a SystemTap test suite that uses DejaGnu.
Will can use some help writing stress tests.  If you have tests to
contribute, it's OK to post them to the SystemTap mailing list.

Tom added a read interface to relayfs in support of his request that
Linus include it.  This affected the API, with the result that the
relayfs API in the -mm tree no longer matches the relayfs API in
RHEL4 U2.  This has implications for people trying to run SystemTap
on recent -mm kernels.  Tom says that the relayfs test suite (AR #44)
will be smart enough to deal with either version of the API.

Work on the translator is progressing.  Support for statistics still
needs a lot of work.  Graydon and Frank will publish an update to
their feature list of 8/4/05.

There were some questions about the availability of kernel debug
info in RHEL4 U2.  These were addressed on the call or by Elena's
8/19/05 email.

Martin has eliminated netlink dependencies from the runtime, in favor
of a /proc interface.  relayfs will still be used as needed.

Vara mentioned that a couple of problems have been seen when we probe
schedule().  Jim will characterize the problem(s).  This brought up
the question of which areas of the kernel should be "blacklisted" --
i.e., documented as not safe to probe.  Jim will open a bugzilla
entry.

Ananth has his RCU-based SMP scalability patches working.  He has
tested it on i386 and ppc64, but needs more testing.  Ananth will
post his patches to the mailing list.  Anil will test them on ia64.

Prasanna has posted a user-space probes design document for internal
review at IBM, and will post it to the SystemTap list when internal
review is complete.

Prasanna has received some review comments on his implementation
of watchpoint probes (posted 7/20/05) and debug-register allocation
(posted 7/20/05, updated 8/2/05).  More review is welcome.

Anil has been giving the translator system a spin, as have Hien
and Kevin.

Charles has created a "poor man's oprofile" [using SystemTap?
Kprobes?].

Frank suggested that an IBMer get involved in the translator work
-- e.g., by adding support for timer-based probes.

Satoshi reports that the djprobes developers are exploring the
feasibility of porting djprobes to other architectures.  He
expects Masami Hiramatsu to post another patch in a few days.



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