timeouts and failure

[Joel Sherrill]

On Tue, Mar 21, 2017 at 1:38 PM, Tanu Hari Dixit <tokencolour at gmail.com>
wrote:

> Hello Chris, All,
>
> Regarding the ticket https://devel.rtems.org/ticket/2946
> for the test spcontext01 do we need to add a tag in a new .tcfg file  ?
> If so what would be the name of the tcfg file?
>
>
This is Chris' to answer.


> Also, can you please help me investigate why 5 more test fail?
>
>
I would run one of them directly and make sure they do in fact complete
successfully on your machine. On the gdb simulators, I don't think there
is an explicit way to say "stop after N simulated seconds" so it is an
approximation of killing it after N host CPU seconds which should be
enough on any machine.

Anyway, pick one of the failing tests and do this:

time sparc-rtems4.12-run XXX.exe

and see (1) if it completes successfully and (2) how long it actually took.

Using sparc-rtems4.12-sis you can "go" and "perf" to see a different view
of similar information.

Maybe 20 seconds isn't enough time to complete. Worth a try.

--joel


> Thanks,
> Tanu Hari Dixit.
>
> On Tue, Mar 21, 2017 at 12:36 PM, Tanu Hari Dixit <tokencolour at gmail.com>
> wrote:
> >> I noticed you have 17 tests more than I do enabled. How did you
> configure?
> >
> > ../../rtems/configure --target=sparc-rtems4.12 --enable-tests
> > --enable-rtemsbsp=erc32
> > --prefix=/home/thd/development/rtems/install_path_erc32_alltest_new
> >
> > I used Joel's configuration, that is:
> > ../../rtems/configure --target=sparc-rtems4.12 --enable-rtemsbsp=erc32
> > --prefix=/home/thd/development/rtems/install_path_erc32_js
> > --disable-networking --enable-posix --disable-multiprocessing
> > --disable-smp --enable-tests --enable-cxx --enable-maintainer-mode
> >
> > I got the following results:
> > Passed:   540
> > Failed:     6
> > Timeouts:   9
> > Invalid:    0
> > -------------
> > Total:    555
> >
> > Failures:
> >  spcontext01.exe
> >  mdosfs_fsscandir01.exe
> >  jffs2_fsscandir01.exe
> >  imfs_fsscandir01.exe
> >  mimfs_fsscandir01.exe
> >  mrfs_fsscandir01.exe
> > Timeouts:
> >  fsdosfsname01.exe
> >  fsdosfswrite01.exe
> >  fsimfsconfig02.exe
> >  fsimfsgeneric01.exe
> >  fsjffs2gc01.exe
> >  fsdosfsformat01.exe
> >  fsrfsbitmap01.exe
> >  imfs_fserror.exe
> >  crypt01.exe
> > Average test time: 0:00:01.434556
> > Testing time     : 0:13:16.178701
> >
> >
> > This is still one more test than Joel's total. I understand
> > spcontext01.exe is expected to fail. But I can't figure out why the
> > rest fail on my host. (My rtems source tree is updated.)
> >
> > Also, how can we change the number of cores operational? I found the
> > --timeout option but none regarding the cores.
> >
> > Can the devs please help me with this?
> >
> > Thanking you,
> > Tanu Hari Dixit.
> >
> > On Tue, Mar 21, 2017 at 5:55 AM, Chris Johns <chrisj at rtems.org> wrote:
> >> On 20/03/2017 17:20, Sebastian Huber wrote:
> >>>
> >>>
> >>>
> >>> On 18/03/17 21:52, Joel Sherrill wrote:
> >>>>
> >>>>
> >>>> FWIW here are my results for sparc/erc32:
> >>>>
> >>>> Passed:   548
> >>>> Failed:     1
> >>>> Timeouts:   5
> >>>> Invalid:    0
> >>>> -------------
> >>>> Total:    554
> >>>>
> >>>> Failures:
> >>>>  spcontext01.exe
> >>>
> >>>
> >>> Yes, this is expected.
> >>>
> >>>> Timeouts:
> >>>>  fileio.exe
> >>>>  top.exe
> >>>>  termios.exe
> >>>>  capture.exe
> >>>>  monitor.exe
> >>>
> >>>
> >>> These tests should terminate after 20 seconds or so without input. A
> >>> timeout should therefore not occur.
> >>
> >>
> >> That is 20 seconds of target time and not real-time if running on a
> heavily
> >> loaded machine running simulators in parallel. The test tool's timeout
> is
> >> real time.
> >>
> >> I have created https://devel.rtems.org/ticket/2946 to look into this.
> >>
> >> Chris
> >>
> >>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20170321/80e53925/attachment-0002.html>