These changes sync QA up with the event records stuff I just committed.
Changes committed to git://oss.sgi.com/pcp/pcpqa.git
002 | 2
003 | 7
006.out | 2
010 | 4
010.out | 83 ++++++++
029.out | 22 ++
108.out | 12 +
139.out | 15 +
149 | 2
156 | 10
156.out | 64 ++++--
162.out.4 | 11 +
256.out.2 | 15 +
274.out | 2
284 | 2
290 | 2
329 | 2
330 | 2
332 | 2
333 | 2
379.out | 58 +----
385 | 39 +++
385.out | 21 ++
386 | 28 ++
386.out | 25 ++
387 | 27 ++
387.out | 31 ++
388 | 50 ++++
388.out | 67 +++++-
391 | 88 +++++++-
391.out | 157 +++++++++++++++
393 | 43 ++++
393.out | 31 ++
394 | 56 +++++
394.out | 51 ++++
395 | 43 ++++
395.out | 31 ++
397 | 59 +++++
397.out | 39 +++
400 | 67 ++++++
400.out | 507 +++++++++++++++++++++++++++++++++++--------------
401 | 41 +++
401.out | 57 +++++
404 | 41 +++
404.out | 128 +++++++++++-
405 | 37 +++
405.out | 8
406 | 26 ++
406.out | 30 ++
407 | 44 ++++
407.out | 22 ++
409 | 39 +++
409.out | 56 +++++
410 | 33 +++
410.out | 60 +++++
413 | 32 +++
413.out | 9
415 | 35 +++
415.out | 15 +
569 | 6
569.out | 77 ++++++-
587 | 11 -
643 | 2
common.check | 48 ++++
group | 75 ++++---
owner | 65 ++++--
src-oss/.gitignore | 1
src-oss/GNUmakefile | 4
src-oss/eventrec.0 |binary
src-oss/eventrec.index |binary
src-oss/eventrec.meta |binary
src-oss/interp0.c | 5
src-oss/mkeventrec | 15 +
src-oss/unpack.c | 334 ++++++++++++++++++++++++++++++++
src-oss/xval.c | 51 +++-
valgrind-suppress | 41 +++
76 files changed, 2899 insertions(+), 330 deletions(-)
commit 1cbf865341869d91431794236397a8a99f7946b4
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 19:35:43 2010 +1100
413, 415 PM_TYPE_EVENT and assorted tools
Check pmie (413), pmstore (415) and pmval (415).
commit d3a5b0a3c8704a937e83b8c71b3bceebf6e4c6b7
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 16:31:56 2010 +1100
410 - pmdumplog with event records
commit c0447dab5c299a691e189ab7100610886fe85a85
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 13:45:28 2010 +1100
404 - more verbose diags from pmUnpackEventRecords
The second iteration of pmUnpackEventRecords uses __pmCheckEventRecords
to validate the integrity of the array of event records, and if a
problem is found, calls __pmDumpEventRecords ... this produces more
verbose output for the error cases of this QA test.
commit a040ceab0ee93dd2a4ae9dbc828a42f1c8d0e5a9
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 13:45:08 2010 +1100
400 - cosmetic change
commit 8fbcf1605dcd38c66eda5133578d4a97b41e91aa
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 13:42:59 2010 +1100
409 - PM_EVENT_TYPE data in an archive
commit 0cba408df33a1f9d02c81cb40e2c883ede332c9c
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Nov 23 13:41:54 2010 +1100
unpack.c - track API change for pmUnpackEventRecords
commit 459122270b8d94bb7fe9188b531c8d544f8add8f
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 08:55:26 2010 +1100
407 - PM_TYPE_EVENT and derived metrics exercises
commit d3b1d5ee6b2f99a62ad9bcb67fe0cd56dea03d4e
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 07:10:33 2010 +1100
406 - no interpolation for event records
Check that using pmFetch and PM_MODE_INTERP returns PM_ERR_TYPE
as expected for event records.
commit d592a118a2603b23ad7bee63bf5497f19c833f2e
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 07:05:31 2010 +1100
587, 643 - make sure trivial PMDA survives QA
Better engineered solution than previous attempt to prevent
removal of pmdatrivial as a side-effect of a QA run.
commit 91b83b061f62d51b140cf8cb1b16e4777721747b
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 06:35:44 2010 +1100
405 - new valgrind test for pmUnpackEventRecords and pmFreeEventResult
Just 404 run under valgrind.
commit af10cd62539ec1b9f247091f5212993b8474d105
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 06:30:52 2010 +1100
404 - extend coverage to error cases
pmUnpackEventRecords has a lot of internal checks for the consistency
of the packed array of event records ... exercise this logic.
Also the valgrind suppression fingerprint for the one (so far) known
false hit is slightly different for binaries built and installed from a
workarea versus those installed from a package (at least on Ubuntu 10.10).
commit 249e0eaa529a9979982325610fad5d0ac487f6ba
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Nov 22 06:29:59 2010 +1100
391, 397 - improve output filtering
commit 646f9e5b27a88b678e9155e76f16b288ed5d4d6b
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun Nov 21 14:22:46 2010 +1100
388 - reduce iterations for valgrind slowdown
commit 573db87f2a34703f799f2658733b8b8009c5e33f
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat Nov 20 08:04:54 2010 +1100
404 exercise pmUnpackEventRecords and pmFreeEventResult
commit 86949c8d4a997a63428260ac67670195b88a2ea9
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat Nov 20 07:57:11 2010 +1100
400 (filter tweak), 401 (400 without valgrind)
Completing the coverage of pminfo -x, especially for platforms
that don't have valgrind installed.
commit 711541f22e976c9b3fa0ea3470c50e67868ca777
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat Nov 20 07:41:06 2010 +1100
400 pminfo -x and valgrind
All manner of things being exercised here ...
- most of the PM_TYPE_EVENT changes
- new -x option to pminfo
- new libc routines pmUnpackEventRecords and pmFreeEventResult
- PM_CONTEXT_LOCAL with all of the above
- check for memory leaks
commit bc984f9d76a8ef14dcb12ea6952fd84fdb7b7526
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat Nov 20 07:19:31 2010 +1100
397 Check PM_TYPE_EVENT endian conversions
Use dbpmda to avoid double conversion (pmda -> pmcd -> client)
hiding any problems.
commit 1062bfffac492980755d74ba2e7a173be0c1d5bb
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat Nov 20 07:17:39 2010 +1100
assorted - second round of sample.event.* changes
More metrics, changed some of the pmid assignments.
commit 4ebaeb1bb4eb713abcb9ba9cca3ab2255715a28d
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:17:47 2010 +1100
587 fix long-standing itch
This test removed the trivial pmda binary as an unintended by
product of its cleanup ... if trivial pmda is installed it now
still works after running QA.
commit 81b5d8ba3083d953be7002f9e158772b335f6c3e
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:17:04 2010 +1100
379 better diagnostics from cisco pmda parser
commit 4c5c695a132357e07be9aff35a05b04a42b76f19
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:15:52 2010 +1100
Add valgrind tests
Add infrastructure and valgrind versions of all the old Purify
tests.
commit e4be626d7710e94c55e4131fdf7f173e6a9d2255
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:11:03 2010 +1100
006 274 - changes for new PM_ERR_TYPE
commit 4613b464fa273a163cb28e8cd487e7edbb31b3b8
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:09:38 2010 +1100
assorted changes for new sample.event.* metrics
commit e468b655e721fdf3aaed218075759878a76e8367
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Nov 19 07:06:17 2010 +1100
385 check PM_TYPE_EVENT using sample.event.*
commit b16a2198c85e50768014a133c43df3bcc11c4465
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Nov 18 17:40:54 2010 +1100
010 actually make xval to some work!
Not sure how long this has been broken, but the answer is hidden
someplace inside BugWorks at sgi.
commit 5ea75660ac77da30df394cd431bc20b974ab0826
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Nov 18 17:39:17 2010 +1100
002 filter out sample.event metrics
Avoids errors and non-determinism from the state of event.records
commit 135abb8eb24e30fbe1b2ae3753ae60c128a6dc52
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Nov 18 17:37:50 2010 +1100
Fix mode to be 644 for miscellaneous *.out files
|