Bug 399986

Summary: LSB3.2 tests failed on opensuse-11-beta2-ppc64-Power6.
Product: [openSUSE] openSUSE 11.0 Reporter: LTC BugProxy <bugproxy>
Component: OtherAssignee: Jiri Dluhos <jdluhos>
Status: RESOLVED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: matz
Version: Beta 2   
Target Milestone: ---   
Hardware: PowerPC-64   
OS: All   
URL: http://
See Also: https://bugzilla.linux.ibm.com/show_bug.cgi?id=44924
Whiteboard:
Found By: Third Party Developer/Partner Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: tar-gzip ball of the results-3.2
tar ball of the results-3.1
A simple test case.
Source of the offending test.
Desktop test result-3.2.0.8
libstdc++ test result -3.2.0.8
desktop-t2c test results-3.2.0.8

Description LTC BugProxy 2008-06-13 10:09:03 UTC
=Comment: #0=================================================
Indhu Durai <IndhuDurai@in.ibm.com> - 2008-05-20 03:07 EDT
Problem description:
====================
While running LSB3.2(ppc64 bit)tests on OpenSuse-11-Beta2 ppc64 
( 2.6.25-26-ppc64) using latest
lsb-dtk-manager(lsb-dist-testkit-3.2.0-4.ppc64.tar.gz), noticed results as below;

Certification Tests:
   Runtime Interface Tests
    Core Tests 	  v 3.2.0-2 	Failures: 32 	Passed: 5315
   Desktop Tests: v 3.2.0-2	
        Fontconfig Tests 	Success 	Passed: 3
        GTK Tests 		Failures: 8 	Passed: 8072
        Libpng Tests 		Success 	Passed: 137
        Libxml Tests 		No Result  [?] 	Passed: 0
        Qt3 Tests 		Failures: 1 	Passed: 0
   Libstdc++ Tests v 3.4.3-7 	Failures: 4 	Passed: 1270
   Perl Tests 	   v 3.2.0-1 	Failures: 1 	Passed: 0
  Python Runtime Tests v 3.2.0-1Failures: 1 	Passed: 107
 Unsorted
   
    freetype_test 	Success 	Passed: 4
    qt4_test 		Failures: 1 	Passed: 0
    xft_test 		Success 	Passed: 4
    xrender_test 	Success 	Passed: 1

----------------------------------------------------------------------------------

uname -a:
=========
Linux linux 2.6.25-26-ppc64 #1 SMP 2008-04-30 07:56:05 +0200 ppc64 ppc64 ppc64
GNU/Linux

----------------------------------------------------------------------------------

Hardware Environment:
=====================
    Machine type (p650, x235, SF2, etc.):llm62
    Cpu type (Power4, Power5, IA-64, etc.):Power6

----------------------------------------------------------------------------------

Reproduce the issue:
====================
Is this reproducible?
    If so, how long does it (did it) take to reproduce it? 5 hours

    Describe the steps:  
# $ tar -xzvf lsb-dist-testkit-3.2.0.4.ppc64.tar.gz
# $ cd lsb-dist-testkit
# $ ./install.sh
# run lsb tests:
    * $ /opt/lsb/tests/manager/autotest-ext/lsb-tef.pl all
=Comment: #2=================================================
Sridhar Vinay <vinaysridhar@in.ibm.com> - 2008-05-21 04:45 EDT
Indhu,
Could you confirm if opensuse11 is certified upto 3.2 version of LSB?
Else, what version of lsb is opensuse11 certified under?

=Comment: #3=================================================
Indhu Durai <IndhuDurai@in.ibm.com> - 2008-05-22 04:37 EDT
(In reply to comment #2)
> Indhu,
> Could you confirm if opensuse11 is certified upto 3.2 version of LSB?
> Else, what version of lsb is opensuse11 certified under?
> 

Hi George,

Any comments on this??

Regards,
Indhu D
=Comment: #4=================================================
Pavan Naregundi <pnaregun@in.ibm.com> - 2008-05-23 06:28 EDT
(In reply to comment #2)
> Indhu,
> Could you confirm if opensuse11 is certified upto 3.2 version of LSB?
> Else, what version of lsb is opensuse11 certified under?
> 
Vinay,
openSUSE11 is certified upto 3.1, however looking forward SLES 10.x would be
LSB4( next LSB release) certified. So its better to do openSUSE11 testing with
latest LSB 3.2. Same will apply for fedora9.
=Comment: #5=================================================
Sridhar Vinay <vinaysridhar@in.ibm.com> - 2008-05-23 06:54 EDT
Indhu,
Could you post the results of the test here?
Also, could you run the lsb-3.1 test suite and post the results please?
=Comment: #6=================================================
Indhu Durai <IndhuDurai@in.ibm.com> - 2008-05-28 08:18 EDT

tar-gzip ball of the results-3.2

tar-gzip ball of the results-3.2
=Comment: #7=================================================
Indhu Durai <IndhuDurai@in.ibm.com> - 2008-05-28 08:20 EDT
Hi Vinay,

Will upload the results of lsb3.1 once complete.

Cheers,
Indhu D
=Comment: #9=================================================
Indhu Durai <IndhuDurai@in.ibm.com> - 2008-06-05 09:25 EDT

tar ball of the results-3.1

tar zip ball of the results-3.1

=Comment:  =================================================
Response from George Kraft
Regarding OpenSuse-11, it just recently GAed and I don't see any LSB certification records for it;
therefore, I would recommend using LSB-3.2 tests and report any issues to Novell.
Comment 1 LTC BugProxy 2008-06-13 10:09:14 UTC
Created attachment 222000 [details]
tar-gzip ball of the results-3.2
Comment 2 LTC BugProxy 2008-06-13 10:09:17 UTC
Created attachment 222001 [details]
tar ball of the results-3.1
Comment 3 Jiri Dluhos 2008-06-13 11:15:55 UTC
Thank you for the logs; I am now analysing them.

Many of the failures are known to us, are in process of being fixed. Here is a quick summary for the Core test:

/tset/LI18NUX2K.L1/base/strptime/T.strptime 32:

- a glibc bug
- please see https://bugzilla.novell.com/show_bug.cgi?id=355887

/tset/LSB.os/procprim/execl/T.exec* 3,
/tset/LSB.os/procenv/nice_X/T.nice_X 1,
/tset/LSB.os/procprim/fork/T.fork 8:

- bug in kernel privilege check for SCHED_FIFO
- please see https://bugzilla.novell.com/show_bug.cgi?id=394548

/tset/POSIX.os/procprim/exec/T.exec* 15,
/tset/POSIX.os/procenv/sysconf/T.sysconf 1:

- wrong value of ARG_MAX in libc headers
- please see https://bugzilla.novell.com/show_bug.cgi?id=389791

/tset/lsb_release/testcases/lsb_release/lsb_release-tc 2

- this failure is harmless, caused by the fact that lsb_release does not report LSB 3.2 (as we did not yet reached full LSB 3.2 compatibility)

/tset/LSB.os/ioprim/readv_L/T.readv_L 24,
/tset/LSB.os/ioprim/writev_L/T.writev_L 30:

- these are test suite deficiencies:
- please see https://www.linuxfoundation.org/lsbprs/pr/116, https://www.linuxfoundation.org/lsbprs/pr/117

/tset/LSB.pam/testcases/*:

- all these failures seem to be caused by a miscompiled test suite
Comment 4 Jiri Dluhos 2008-06-13 12:02:22 UTC
Could you please re-run the tests with the newest OpenSUSE pre-release *and* with the newest beta build of the LSB tests?

(I am generally using the most recent beta of the LSB tests when testing our own betas; it is useful because the betas typically have some additional fixes.)
Comment 5 Jiri Dluhos 2008-06-16 11:25:16 UTC
To the other tests in the report:

At least part of the Gtk failures look to me like test suite deficiencies; results should be better when using the most recent beta of the LSB tests.

Results in the Qt3 and libstc++ test are strange; I suspect that you had accidentally downloaded a miscompiled version of the test suite. Again, please inform us how the result changes with the most recent beta.
Comment 6 LTC BugProxy 2008-06-24 08:41:45 UTC
------- Comment From IndhuDurai@in.ibm.com 2008-06-24 04:32 EDT-------
Vinay,

As opensuse11.0 public release is over, is it ok to run on that or you want me
to run on the prior release?

regards,
Indhu Durai
Comment 7 LTC BugProxy 2008-07-11 06:41:34 UTC
------- Comment From IndhuDurai@in.ibm.com 2008-07-11 02:34 EDT-------
Hi Vinay,

The size of the results tar ball is exceeding the limits(2M) of the
attachment(4M) size. I have uploaded the results in the
repo->9.124.111.85:/dump/IMAGES/opensuse/11/lsb/results.tgz.

Regards,
Indhu Durai
Comment 8 LTC BugProxy 2008-07-11 10:21:49 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-11 06:16 EDT-------
Novell,

You can obtain the results file from
"testcase.software.ibm.com/fromibm/linux/results.tgz" using the anonymous login
Comment 9 Jiri Dluhos 2008-07-11 17:41:04 UTC
Thank you for the info; I am now analyzing it. So far I can tell:

Results from the library check seem fine to me; all failures are caused by missing libraries, which is normal for a server system (unless you know that these libraries were installed and should be found).

Results of the runtime test are reflecting two known bugs where fixing is underway (#355887, #389791), and two known test suite deficiencies:

https://www.linuxfoundation.org/lsbprs/pr/116
https://www.linuxfoundation.org/lsbprs/pr/117

The only really interesting failure is

/tset/LI18NUX2K.L1/base/iswctype/T.iswctype.2

I don't recall seeing this before; it seems to only occur on ppc64.


Comment 10 Jiri Dluhos 2008-07-11 17:59:45 UTC
Created attachment 227370 [details]
A simple test case.

I have created a small test case which, hopefully, might reproduce the bug. Could you please test it on the same machine? I will try to test it on-site as soon as I get my paws on a free PPC machine.

(It is necessary to install the locale first, then run the test.)
Comment 11 LTC BugProxy 2008-07-21 12:27:19 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-21 08:14 EDT-------
Novell,

Could you please specify how the locale is to be set up?
Comment 12 Jiri Dluhos 2008-07-21 13:53:30 UTC
If everything works as expected, it should be enough to enter the directory with the two testing locale source files (LTP_1 and UTF-8) and enter (as root):

localedef -c -f UTF-8 -i LTP_1 "LTP_1.utf8"

This should install the testing locale to the system.

Comment 13 LTC BugProxy 2008-07-22 05:31:47 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-22 01:23 EDT-------
Novell,

On running the above command I get the following message :

"LC_MONETARY: value of field `int_curr_symbol' does not correspond to a valid
name in ISO 4217"
Comment 14 Jiri Dluhos 2008-07-22 09:37:38 UTC
Do not worry, the message is only a warning. The installation should succeed; the "-c" flag is there for that purpose.
Comment 15 LTC BugProxy 2008-07-22 10:12:18 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-22 06:08 EDT-------
Ok, ran the test after setting the locale. Test ran silently with the "return"
variable = 0

------- Comment From vinaysridhar@in.ibm.com 2008-07-22 06:09 EDT-------
Sorry, I meant the "result" variable = 0
Comment 16 LTC BugProxy 2008-07-24 10:01:58 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-24 05:52 EDT-------
Novell,

Did the above test result reveal anything?
Comment 17 Jiri Dluhos 2008-07-24 13:04:51 UTC
The result is interesting because the test did not detect the iswctype() problem the LSB test suite has detected. Did you use the same computer and OS as the one where the last LSB test suite was run?

(Maybe I have a bug in the small test program, or maybe the LSB test suite failure is caused by a problem in a different test.)
Comment 18 Jiri Dluhos 2008-07-25 16:28:41 UTC
Finally I have a working ppc machine! :-)

So far, it seems that the problem does _not_ appear in openSUSE 11 if a ppc32 version is used; the /tset/LI18NUX2K.L1/base/iswctype/T.iswctype.2 test normally passes.

I will try the ppc64 version now...
Comment 19 Jiri Dluhos 2008-07-28 11:30:56 UTC
Well, it seems I have misunderstood something :-(

The ppc64 version of LSB tests, which I assume is the version you used, refuses to even install on our testing machine.

After some thinking I have realized that this is probably normal because in openSUSE for ppc, the userspace is mostly 32-bit; only the kernel is 64-bit, plus some libraries. The primary cause why the package does not install is that it cannot find the lsb-core-ppc64 symbol, which is not present because the 'lsb' package in openSUSE is built in 32-bit mode. I am able to bypass this obstacle by building a custom package, but it is a hack adnd I'm not sure whether the result can be called a clean distro. :-(

Please, how did you manage to install the 64-bit test suite at all?
Comment 20 LTC BugProxy 2008-07-28 11:52:33 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-28 07:44 EDT-------
Indhu,

Could you please reply to the queries in comments 37 and 39?

Thanks,
Vinay
Comment 21 LTC BugProxy 2008-07-29 10:34:57 UTC
------- Comment From pnaregun@in.ibm.com 2008-07-29 06:28 EDT-------
(In reply to comment #39)
>
> Please, how did you manage to install the 64-bit test suite at all?

Created a soft link of /lib64/ld64.so.1 to /lib64/ld-lsb-ppc64.so.3 and
/lib64/ld-lsb-ppc64.so.2. This managed us to run it on 64-bit test suite.
Comment 22 Jiri Dluhos 2008-07-30 15:50:04 UTC
Finally I have managed to reproduce the problem; it requires not only the symlinks but also install a specially compiled LSB package that provides the lsb-core-ppc64 atom.

The results look very similar to yours, with the single problem that worries me:

/tset/LI18NUX2K.L1/base/iswctype/T.iswctype 2

CC'ing to our libc gurus.
Comment 23 Jiri Dluhos 2008-07-30 17:04:33 UTC
*** Bug 384903 has been marked as a duplicate of this bug. ***
Comment 24 Jiri Dluhos 2008-07-30 17:07:34 UTC
Michael, Pasky, do you have any idea about the iswctype problem? It seems that a test that does iswctype() segfaults on ppc64; as far as I can tell, the test is straightforward, I don't know why it crashes.

Attaching the source code of the test.

Comment 25 Jiri Dluhos 2008-07-30 17:12:37 UTC
Created attachment 230860 [details]
Source of the offending test.
Comment 26 LTC BugProxy 2008-07-31 10:51:56 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-07-31 06:44 EDT-------
Novell,

Is this a libc issue or a test-suite issue? Because running the test you posted
earlier did not cause any libc issue. The iswctype problem comes only from lsb...
Comment 27 Michael Matz 2008-07-31 14:12:53 UTC
Jiri, please give me exact instructions how I should reproduce this.
I.e. commands and files I need (it seems I need some exact locale files?).
Best if you pack every necessary file together in a tarball and attach it here.
Comment 28 Jiri Dluhos 2008-07-31 14:52:40 UTC
Vinay: So far, I still can't tell whether it is a libc bug or a problem of the test suite. It is true that the test program I produced did not reproduce it, but that may be because I failed to mimic the behavior of the test suite close enough. We are still working on it - please stay tuned.
Comment 29 Jiri Dluhos 2008-07-31 16:55:53 UTC
Ah well. I am an idiot.

http://bugs.linuxbase.org/show_bug.cgi?id=2146

The iswctype problem is already known for 14 days; it is a bug in the test suite (uninitialized variable), and fixed in the bazaar version. That's why my test program did not crash - I extracted it from the bazaar version, where the problem is already fixed. When I un-fixed the code, it exhibits the crash perfectly.

Morale: always google first before getting entangled in the code.

Michael, Vinay, Pasky, Jan: I'm sorry for bothering you. Alarm canceled.
Comment 30 Jiri Dluhos 2008-07-31 17:04:36 UTC
If I count correctly, this was the last problem, at least regarding the core test. However, we need to wait until the LSB team builds a new beta of the test suite with these fixes included, so that it can be tested.
Comment 31 LTC BugProxy 2008-08-05 11:32:15 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-08-05 07:29 EDT-------
Novell,

The core failure issues seem to be settled. Could you share information on the
other failures : gtk, qt3, libstdc++, etc?
Comment 32 Jiri Dluhos 2008-08-07 13:04:56 UTC
Hmmm... The set of failures in libstdc++ is interesting. When I have run the test on our ppc64 machine, it passed (which is also strange - LSB tests are so strict that it is a sign of special luck when some of them passes).

I will try the rest and attach the resulting report; maybe we will find the causes by comparison.
Comment 33 Jiri Dluhos 2008-08-07 16:19:33 UTC
The tests are still running, but so far it seems that the number of failures is significantly lower than in your tests. I think it might be caused by old tests;
I would recommend downloading a fresh new Desktop Test Toolkit from linuxfoundation.org:

http://www.linuxfoundation.org/en/Downloads

and re-running the desktop and libstdc++ tests, allowing the DTK to download new versions of tests from the web. With all other tests disabled, the run should be quite fast - let's say two hours.

I will attach my results as soon as they are completed.
Comment 34 Jiri Dluhos 2008-08-07 17:45:22 UTC
Created attachment 232334 [details]
Results of the desktop test run on a ppc64 machine in Novell

I have attached the results of the test run on our testing machine. From a quick view, the results look quite good; most of the remaining failures are known test suite problems.
Comment 35 Jiri Dluhos 2008-08-08 13:10:15 UTC
Can you please re-run the desktop and libstc++ tests (plus any other you need) with the newest version of the Desktop Test Kit and new test packages? I believe the results will be much better.
Comment 36 Jiri Dluhos 2008-08-08 14:22:58 UTC
For better orientation, here is a short summary of failures I have seen in the test run on our machine:

=== libchk ===

(Only uninstalled libraries, plus known test defects.)

=== desktop T2C test ===

* /glib-t2c/tests/glib_unicode/glib_unicode 18
* /glib-t2c/tests/glib_character_set_conversion/glib_character_set_conversion 32
* /gmodule-t2c/tests/gmodule/gmodule 23

(These three look like test suite problems, but deserve investigation.)

=== gtk test ===

* /tests/functions/GtkToolbar/GtkToolbar 15

(Test suite bug, http://bugs.linuxbase.org/show_bug.cgi?id=1895)

* /tests/functions/GdkImages/GdkImages 2

(Test suite bug, http://bugs.linuxbase.org/show_bug.cgi?id=1893)

* /tests/functions/Drag_and_Drop/Drag_and_Drop 41

(Looks like http://bugs.linuxbase.org/show_bug.cgi?id=1260)
Comment 37 LTC BugProxy 2008-08-12 10:12:57 UTC
------- Comment From pavan.naregundi@in.ibm.com 2008-08-12 06:03 EDT-------
(In reply to comment #54)
> ------- Comment From jdluhos@novell.com 2008-08-08 07:10:15 MDT-------
Result of the libstdc++(4 failures), desktop(10 failures) and desktop-t2c(16
failures) tests on opensuse11.1 Alpha1 is attached below. Some of the failures
are same as mentioned in comment #55. I have run these tests with
lsb-dist-testkit-3.2.0.8.
Comment 38 LTC BugProxy 2008-08-12 10:12:59 UTC
Created attachment 232913 [details]
Desktop test result-3.2.0.8
Comment 39 LTC BugProxy 2008-08-12 10:13:01 UTC
Created attachment 232914 [details]
libstdc++ test result -3.2.0.8
Comment 40 LTC BugProxy 2008-08-12 10:13:02 UTC
Created attachment 232915 [details]
desktop-t2c test results-3.2.0.8
Comment 41 Jiri Dluhos 2008-08-12 15:26:57 UTC
I'm sorry but results from openSUSE 11.1 alpha1 cannot help us in this case. We need results from openSUSE 11.0 final. The new openSUSE 11.1 is too different.

Therefore, results #38, #39, #40 are flawed - sorry for that.

Also, the desktop test is still old; the most recent version is currently 3.2.1-2; this affects at least the GTK test (I think the Drag-and-Drop failures are all already fixed bugs in the test suite, probably even the Resource_Files.34 failure).

Comment 42 LTC BugProxy 2008-08-13 08:32:23 UTC
------- Comment From pavan.naregundi@in.ibm.com 2008-08-13 04:23 EDT-------
(In reply to comment #60)
> ------- Comment From jdluhos@novell.com 2008-08-12 09:26:57 MDT-------
I have rerun the tests on opensuse11. Results are similar to one mentioned in
comment #55, except that following failure did not occur.
/glib-t2c/tests/glib_character_set_conversion/glib_character_set_conversion 32
Comment 43 Jiri Dluhos 2008-08-13 11:47:26 UTC
This is a bit strange because my results are still different; please, did you use the newest version of the desktop test? (3.2.1-2)
Comment 44 LTC BugProxy 2008-08-19 04:52:26 UTC
------- Comment From pavan.naregundi@in.ibm.com 2008-08-19 00:43 EDT-------
(In reply to comment #62)
> ------- Comment From jdluhos@novell.com 2008-08-13 05:47:26 MDT-------
> This is a bit strange because my results are still different; please, did you
use the newest version of the desktop test? (3.2.1-2)

Yes.
Comment 45 LTC BugProxy 2008-09-02 06:14:54 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-09-02 02:04 EDT-------
Pavan, Novell, how do we proceed on this? If we could agree on a common list of
test failures, we could see what could be worked on. Could you please update
with your views?

Thanks,
Vinay
Comment 46 Jiri Dluhos 2008-09-02 16:45:39 UTC
I'm sorry for the delay, I somehow managed to miss your first answer :-(

I had a more thorough look at the results of the desktop-t2c tests; it seems that at least part of the fails are test bugs that behave semi-randomly, which might explain why you have got a different set of fails than we have.

So far, I have found these failures to be test bugs, and reported them to the LSB committee as such:

/glib-t2c/tests/glib_unicode/glib_unicode 18
   - reported as http://bugs.linuxbase.org/show_bug.cgi?id=2276

/glib-t2c/tests/glib_trash_stack/glib_trash_stack 1
   - reported as http://bugs.linuxbase.org/show_bug.cgi?id=2277

/glib-t2c/tests/glib_threads/glib_threads 2
   - reported as http://bugs.linuxbase.org/show_bug.cgi?id=2278

I am still pondering over the remaining failures - please stay tuned.
Comment 47 Jiri Dluhos 2008-09-03 13:02:34 UTC
desktop-t2c tests, continued:

/glib-t2c/tests/glib_string_utility/glib_string_utility 55-60
   - reported as http://bugs.linuxbase.org/show_bug.cgi?id=2279
Comment 48 Jiri Dluhos 2008-09-04 13:36:42 UTC
desktop-t2c tests, continued:

/glib-t2c/tests/glib_messagelog/glib_messagelog 9,10
/glib-t2c/tests/glib_commandline_option_parser/glib_commandline_option_parser 24,25,46,47
   - with these, I have no idea; I have not found any specific bug in the tests, and these failures do not occur at our testing machine.
Comment 49 LTC BugProxy 2008-09-05 12:02:14 UTC
------- Comment From pavan.naregundi@in.ibm.com 2008-09-05 07:50 EDT-------
(In reply to comment #67)
> ------- Comment From jdluhos@novell.com 2008-09-04 07:36:42 MDT-------
> desktop-t2c tests, continued:
>
> /glib-t2c/tests/glib_messagelog/glib_messagelog 9,10
> /glib-t2c/tests/glib_commandline_option_parser/glib_commandline_option_parser
24,25,46,47
> - with these, I have no idea; I have not found any specific bug in the tests,
and these failures do not occur at our testing machine.

Novell,
I think you are not able to see my post of the result of latest run using
desktop test(3.2.1-2), as I have mentioned comment 55, but it does not point to
anything. Sorry for that.

Here is the summary of result I got in latest run using desktop test(3.2.1-2)

=== desktop T2C test ===

* /glib-t2c/tests/glib_unicode/glib_unicode 18
* /gmodule-t2c/tests/gmodule/gmodule 23

These are only errors.

=== gtk test ===

* /tests/functions/GtkToolbar/GtkToolbar 15

(Test suite bug, http://bugs.linuxbase.org/show_bug.cgi?id=1895)

* /tests/functions/GdkImages/GdkImages 2

(Test suite bug, http://bugs.linuxbase.org/show_bug.cgi?id=1893)

* /tests/functions/Drag_and_Drop/Drag_and_Drop 41

(Looks like http://bugs.linuxbase.org/show_bug.cgi?id=1260)

These are known failures.
Comment 50 Jiri Dluhos 2008-09-05 12:13:26 UTC
Excellent! It seems that many test suite failures are fixed in the newest test suite version, which is good.

Were the other tests (desktop and libstdc++) also better? Can we consider this issue closed?
Comment 51 LTC BugProxy 2008-09-10 11:52:17 UTC
------- Comment From pavan.naregundi@in.ibm.com 2008-09-10 07:47 EDT-------
(In reply to comment #69)
> ------- Comment From jdluhos@novell.com 2008-09-05 06:13:26 MDT-------
> Were the other tests (desktop and libstdc++) also better? Can we consider this
issue closed?

There where only known failures in libstdc++ and desktop tests. Yes, we can
close this bug now.

Thanks
Pavan
Comment 52 LTC BugProxy 2008-09-11 09:53:21 UTC
------- Comment From vinaysridhar@in.ibm.com 2008-09-11 05:48 EDT-------
Closing as per the above comments
Comment 53 LTC BugProxy 2008-09-17 17:14:27 UTC
Rodrigo Sampaio Vaz <rsampaio@br.ibm.com> - 2008-04-04 16:51 EDT
---Problem Description---
While running LSB 3.1 runtime-tests on a power machine I noticed 9 failures on
SLES-10sp2-RC1 (2.6.16.60-0.9-ppc64 #1 SMP) using
lsb-dist-testkit-3.1.1-5.ppc64.tar.gz
Following are the 9 errors collected from html report generated by runtime-tests

1) /tset/LI18NUX2K.L1/base/iswctype/T.iswctype 2 	Unresolved
Messages from the test

If wc does not belong character class which shows desc, verify this function
returns 0.
Test Locale is LTP_1.UTF-8
unexpected signal 11 (SIGSEGV) received

2) /tset/LI18NUX2K.L1/base/swscanf/T.swscanf 12 	Failed
Messages from the test

If a conversion wide character is `p', verify this function stores the pointer
value.
Test Locale is LTP_1.UTF-8
Conversion wide character p is not interpreted correctly
path tracing error: path counter 0, expected 1

3) /tset/LI18NUX2K.L1/base/__wcstold_internal/T.__wcstold_internal 8 	Failed
Messages from the test

When the correct value is outside of representable values,
verify that plus or minus HUGE_VALL is returned and ERANGE
is stored in errno.
Test Locale is C
The return value is not correct.
The errno is not correct.
path tracing error: path counter 0, expected 2

4) /tset/LI18NUX2K.L1/base/wcstold/T.wcstold 8 	Failed
Messages from the test

When the correct value is outside of representable values,
verify that plus or minus HUGE_VALL is returned and ERANGE
is stored in errno.
Test Locale is C
The return value is not correct.
The errno is not correct.
path tracing error: path counter 0, expected 2

5) /tset/LSB.fhs/root/bin/bin-tc 11 	Failed
Messages from the test

Reference 3.4-11 (A)
The implementation provides an exec-able version of the dmesg
utility in the /bin directory.
Expected output to stdout, but none written

6) /tset/LSB.os/ioprim/readv_L/T.readv_L 24 	Failed
Messages from the test

iov[0].iov_base = 0x8f4003f4010, iov[0].iov_len = 1024
iov[1].iov_base = 0x8f4003f4010, iov[1].iov_len = 1024
iov[2].iov_base = 0x8f4003f4010, iov[2].iov_len = 1024
....
iov[1021].iov_base = 0x8f4003f4010, iov[1021].iov_len = 1024
iov[1022].iov_base = 0x8f4003f4010, iov[1022].iov_len = 1024
iov[1023].iov_base = 0x8f4003f4010, iov[1023].iov_len = 1024
readv(5, iov, 1024) did not behave as expected
ERRNO VALUES: expected: 22 (EINVAL), observed: 14 (EFAULT)

7) /tset/LSB.os/ioprim/writev_L/T.writev_L 30 	Failed
Messages from the test

Could not allocate a large write buffer (wanted 1024 bytes)
Will attempt to test using a buffer of 8589934592 bytes
This may result in a SIGSEGV being raised
iov[0].iov_base = 0x8f4003f4010, iov[0].iov_len = 1024
iov[1].iov_base = 0x8f4003f4010, iov[1].iov_len = 1024
iov[2].iov_base = 0x8f4003f4010, iov[2].iov_len = 1024
...
iov[1022].iov_base = 0x8f4003f4010, iov[1022].iov_len = 1024
iov[1023].iov_base = 0x8f4003f4010, iov[1023].iov_len = 1024
writev(7, iov, 1024) did not behave as expected:
ERRNO VALUES: expected: 22 (EINVAL), observed: 14 (EFAULT)

8) /tset/PTHR.os/procenv/getlogin_r/T.getlogin_r 1 	Failed 	Known problem
Messages from the test

If the feature test macro _POSIX_THREAD_SAFE_FUNCTIONS is defined:
concurrent calls by multiple threads to getlogin_r(name, namesize)
store the name associated by the login activity with the
controlling terminal of the current process in the character
array referenced by name and return 0 if successful.
Posix Ref: Component GETLOGIN_R
Assertion 9945-1:1996 4.2.4.2-1(C)
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
The system does not support the PTHREAD_SCOPE_PROCESS scope, using
PTHREAD_SCOPE_SYSTEM
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
Unexpected login name root, should be vsx0
LSB Report Status:
Reported to the LSB Committee as PR 0137.

---Additional Hardware Info---
keechi-lp1:~ # cat /proc/cpuinfo
processor       : 0
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 1
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 2
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 3
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 4
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 5
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 6
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

processor       : 7
cpu             : POWER6 (architected), altivec supported
clock           : 3504.000000MHz
revision        : 3.1 (pvr 003e 0301)

timebase        : 512000000
machine         : CHRP IBM,9125-F2A
keechi-lp1:~ # free -m
total       used       free     shared    buffers     cached
Mem:          8037       1596       6441          0         94       1287
-/+ buffers/cache:        213       7823
Swap:         2055          0       2055

---uname output---
Linux keechi-lp1 2.6.16.60-0.9-ppc64 #1 SMP Mon Mar 17 17:16:31 UTC 2008 ppc64
ppc64 ppc64 GNU/Linux

Machine Type = P6/IH - p 575

---Steps to Reproduce---
1 - fetch the latest autotest from test.kernel.org/autotest (svn checkout
svn://test.kernel.org/autotest/trunk/client /usr/local/autotest )
2 - now cd /usr/local/autotest
3 - run bin/autotest tests/lsb_dtk/control
4 - check results in /usr/local/autotest/results/default
(Takes about 6 hrs to reproduce)

---Other Component Data---
Userspace tool common name: lsb-dist-testkit-3.1.1-5.ppc64.tar.gz

The userspace tool has the following bit modes: 64-bit

Userspace rpm: lsb-runtime-test-3.1.1-4.ppc64.rpm

---------------------------
Mirroring this bug to your for your awareness of 8 LSB 3.1 failures on SLES10 SP2.
-thanks.

Jiri, can you have a look at this LSB issue?

*** This bug has been marked as a duplicate of bug 399986 ***
Comment 54 Jiri Dluhos 2008-09-18 15:39:03 UTC
I have looked at the problems; most are known test suite deficiencies. Only #2, #5 and #8 are interesting. I would recommend running newer tests as many test suite bugs were already fixed, and see whether some of the failures persist.

1) Test suite bug (uninitialized variable), please see
   http://bugs.linuxbase.org/show_bug.cgi?id=2146

2) Further info would be needed; this might, but does not have to, be related to
   http://bugs.linuxbase.org/show_bug.cgi?id=1958

3) Test suite bug, see http://bugs.linuxbase.org/show_bug.cgi?id=1771

4) Looks like the same problem as 3)

5) More data would be needed; looks like some simple problem, like missing
   /bin/dmesg.

6) Known problem, https://www.linuxfoundation.org/lsbprs/pr/117

7) Same as 6), only a different call: https://www.linuxfoundation.org/lsbprs/pr/116

8) The automated test analyzer reports marks this as a test suite problem
   (see https://www.linuxfoundation.org/lsbprs/pr/137), which I'm not sure
   if it is the same thing, although I believe it is a test suite problem.
Comment 55 Jiri Dluhos 2008-09-18 16:38:42 UTC
As I believe these bugs are all test suite problems, closing this bug; if the failures persist with newer tests, please re-open.
Comment 56 Jiri Dluhos 2008-09-18 16:39:05 UTC
Closed :-)