[LTP] More observations on HEAD

[LTP] More observations on HEAD

[Joseph Beckenbach]

Hi, all!

Feel free to ignore this -- it's an experience report of a cluster of runs I've just done against 3.12.24 kernels.

The default test set runs move_pages* twice.

setfsuid04_16 and setregid{02,04,05,07}_16 TFAIL rather than TCONF on no-16-bit-syscalls systems.

run_cgroup_test_fj.sh hangs on heavy-load systems with $LTPROOT/output/ not present.

Some test under container_test.sh (sourcing childns.sh) attempts to create veth pairs without verifying veth functionality supported.
Other tests under container_test.sh fail to source files (childns.sh cannot source cmdlib.sh;  child_1.sh and childipv6.sh cannot source initialize.sh).

Many tests under cgroup_regression_test.sh blindly log to hard-coded $LTPROOT/output/ directory without making sure it's there.
Also, subtest #7 "fails to mount hugetlb" because it seemingly did not first check whether the functionality was enabled.

Under cgroup_fj also seems to have some tests which do latency tests and other tests, outputting "TINFO" and "FAIL" and "PASS" not in standard format.
I will see if I can bisect it down and work out which tests those are.

dma_thread_diotest invocations all claim to be "dma_thread_diotest07" in the runltp outputfile.

Thanks!

Joseph
Joseph Beckenbach, senior QA engineer : Automation
x97166
M 678-602-3192


------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] More observations on HEAD

[Stanislav Kholmanskikh]

On 07/19/2014 01:35 AM, Joseph Beckenbach wrote:
> Hi, all!

Hi!

>
> Feel free to ignore this -- it's an experience report of a cluster of runs I've just done against 3.12.24 kernels.
>
> The default test set runs move_pages* twice.
>
> setfsuid04_16 and setregid{02,04,05,07}_16 TFAIL rather than TCONF on no-16-bit-syscalls systems.

Hmm. Indeed, in my environment (OL6 + 3.8.13-26.2.3.el6uek.x86_64) 
setfsuid04_16 fails, but setregid{01,02,03,04}_16 don't. And there are 
no setregid{05, 07}_16

I'll look at setfsuid04_16.


>
> run_cgroup_test_fj.sh hangs on heavy-load systems with $LTPROOT/output/ not present.
>
> Some test under container_test.sh (sourcing childns.sh) attempts to create veth pairs without verifying veth functionality supported.
> Other tests under container_test.sh fail to source files (childns.sh cannot source cmdlib.sh;  child_1.sh and childipv6.sh cannot source initialize.sh).
>
> Many tests under cgroup_regression_test.sh blindly log to hard-coded $LTPROOT/output/ directory without making sure it's there.
> Also, subtest #7 "fails to mount hugetlb" because it seemingly did not first check whether the functionality was enabled.
>
> Under cgroup_fj also seems to have some tests which do latency tests and other tests, outputting "TINFO" and "FAIL" and "PASS" not in standard format.
> I will see if I can bisect it down and work out which tests those are.
>
> dma_thread_diotest invocations all claim to be "dma_thread_diotest07" in the runltp outputfile.
>
> Thanks!
>
> Joseph
> Joseph Beckenbach, senior QA engineer : Automation
> x97166
> M 678-602-3192
>
>
> ------------------------------------------------------------------------------
> Want fast and easy access to all the code in your enterprise? Index and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ltp-list
>

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] More observations on HEAD

[Stanislav Kholmanskikh]

On 07/25/2014 12:03 PM, Stanislav Kholmanskikh wrote:
>
>
> On 07/19/2014 01:35 AM, Joseph Beckenbach wrote:
>> Hi, all!
>
> Hi!
>
>>
>> Feel free to ignore this -- it's an experience report of a cluster of runs I've just done against 3.12.24 kernels.
>>
>> The default test set runs move_pages* twice.
>>
>> setfsuid04_16 and setregid{02,04,05,07}_16 TFAIL rather than TCONF on no-16-bit-syscalls systems.
>
> Hmm. Indeed, in my environment (OL6 + 3.8.13-26.2.3.el6uek.x86_64)
> setfsuid04_16 fails, but setregid{01,02,03,04}_16 don't. And there are
> no setregid{05, 07}_16
>
> I'll look at setfsuid04_16.
>

Sorry, I overlooked that thread "[PATCH v2 1/2] lib/tst_res.c: introduce 
tst_record_childstatus()" already addresses these issues.

>
>>
>> run_cgroup_test_fj.sh hangs on heavy-load systems with $LTPROOT/output/ not present.
>>
>> Some test under container_test.sh (sourcing childns.sh) attempts to create veth pairs without verifying veth functionality supported.
>> Other tests under container_test.sh fail to source files (childns.sh cannot source cmdlib.sh;  child_1.sh and childipv6.sh cannot source initialize.sh).
>>
>> Many tests under cgroup_regression_test.sh blindly log to hard-coded $LTPROOT/output/ directory without making sure it's there.
>> Also, subtest #7 "fails to mount hugetlb" because it seemingly did not first check whether the functionality was enabled.
>>
>> Under cgroup_fj also seems to have some tests which do latency tests and other tests, outputting "TINFO" and "FAIL" and "PASS" not in standard format.
>> I will see if I can bisect it down and work out which tests those are.
>>
>> dma_thread_diotest invocations all claim to be "dma_thread_diotest07" in the runltp outputfile.
>>
>> Thanks!
>>
>> Joseph
>> Joseph Beckenbach, senior QA engineer : Automation
>> x97166
>> M 678-602-3192
>>
>>
>> ------------------------------------------------------------------------------
>> Want fast and easy access to all the code in your enterprise? Index and
>> search up to 200,000 lines of code with a free copy of Black Duck
>> Code Sight - the same software that powers the world's largest code
>> search on Ohloh, the Black Duck Open Hub! Try it now.
>> http://p.sf.net/sfu/bds
>> _______________________________________________
>> Ltp-list mailing list
>> Ltp-list@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/ltp-list
>>
>
> ------------------------------------------------------------------------------
> Want fast and easy access to all the code in your enterprise? Index and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ltp-list
>

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] More observations on HEAD

[chrubis]

Hi!
> Sorry, I overlooked that thread "[PATCH v2 1/2] lib/tst_res.c: introduce 
> tst_record_childstatus()" already addresses these issues.

And it took longer than it should because it require changes in the test
library.

I've just commited a fix, please retest. :)

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls. 
Build a bridge from your legacy apps to the future.
http://pubads.g.doubleclick.net/gampad/clk?id=153845071&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] More observations on HEAD

[chrubis]

Hi!
> The default test set runs move_pages* twice.

True, these are part of two different runtest files and both runtest
files are part of default scenario. I do not think that this is wrong as
it is.

It's rather limitation of the current system to run testcases. What I
would rather is is some label based system, where testcases can be
labeled with different scenarios.

I.e. in this cases move_pages would be labeled as syscalls, numa and
default. Then we can run testcases by labels rather than runtest
files...

> run_cgroup_test_fj.sh hangs on heavy-load systems with $LTPROOT/output/ not present.
> 
> Some test under container_test.sh (sourcing childns.sh) attempts to create veth pairs without verifying veth functionality supported.
> Other tests under container_test.sh fail to source files (childns.sh cannot source cmdlib.sh;  child_1.sh and childipv6.sh cannot source initialize.sh).
> 
> Many tests under cgroup_regression_test.sh blindly log to hard-coded $LTPROOT/output/ directory without making sure it's there.
> Also, subtest #7 "fails to mount hugetlb" because it seemingly did not first check whether the functionality was enabled.
> 
> Under cgroup_fj also seems to have some tests which do latency tests and other tests, outputting "TINFO" and "FAIL" and "PASS" not in standard format.
> I will see if I can bisect it down and work out which tests those are.

Adding checks for missing functionality should be pretty
straightforward and patches are welcome :).

The runltp script contains code to create the $LTPROOT/output directory
so there are a few testcases that expect that directory to exist,
nevertheless they should create test temporary directory with tst_tmpdir
and use it instead.

> dma_thread_diotest invocations all claim to be "dma_thread_diotest07" in the runltp outputfile.

We forget to rename the TCID after we renamed the file. Fixed.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list