Xenomai creates segmentation fault at startup - not in gnulinux

Hi all, I have a problem with a xml schema that works in gnulinux but doesn't in xenomai. I attached to this post the output from gdb. I can say that I'm not using mutexes "intentionally" in my code, while I'm using ports and operations.

Do anybody of you have an idea about that issue?

Thank you in advance,

Best regards,

Mirko

AttachmentSize
error_gdb.txt3.51 KB

Xenomai creates segmentation fault at startup - not in gnulinux

Hi Mirko,

On Mon, Aug 26, 2013 at 11:43 AM, <mirkodaniele [dot] comparetti [..] ...> wrote:

> Hi all,
> I have a problem with a xml schema that works in gnulinux but doesn't in
> xenomai. I attached to this post the output from gdb.
> I can say that I'm not using mutexes "intentionally" in my code, while I'm
> using ports and operations.
>

Yes, these are Orocos mutexes that protect the internals of RTT against
concurrent access of user threads

>
> Do anybody of you have an idea about that issue?
>

Take a look at the output of 'dmesg' , this might explain why it crashes
your application. Orocos is quite heavy on resources, especially mutexes,
so you might have hit a limit there. Another option for crashes in Xenomai
is hitting the stack limit, but we dramatically increased that in the
latest versions.

Peter

Xenomai creates segmentation fault at startup - not in gnulinux

Hi Peter,
Sorry for being late... but I checked the output and I found the following.

output of gdb:
[...]
[New Thread 0x7fffd95bdb00 (LWP 11468)]
[Thread 0x7fffd95bdb00 (LWP 11468) exited]
12.113 [ Warning][Thread] Activity: an object with that name is already
existing in Xenomai.
[New Thread 0x7fffd95bdb00 (LWP 11469)]
[New Thread 0x7fffd959db00 (LWP 11470)]

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff5b84cf4 in rt_mutex_acquire_inner () from
/usr/xenomai/lib/libnative.so.3

output of dmesg:
[ 2414.241357] MainThread[10882]: segfault at 0 ip 00007f34cc945cf4 sp
00007fff751da810 error 6 in libnative.so.3.0.0[7f34cc942000+7000]
[ 2414.297423] Xenomai: native: cleaning up cond "" (ret=0).
[... repeated several times]
[ 2414.297466] Xenomai: native: cleaning up cond "" (ret=0).
[ 2414.297469] Xenomai: native: cleaning up mutex "" (ret=0).
[...repeated several times]
[ 2414.298249] Xenomai: native: cleaning up mutex "" (ret=0).
[ 2414.298254] Xenomai: native: cleaning up sem "" (ret=0).
[...repeated several times]
[ 2414.298387] Xenomai: native: cleaning up sem "" (ret=0).

Do these mean anything to you?

Thank you so much!

Best regards,

- - -
Mirko D. Comparetti

M.Sc., Eng'r, *PhD Student*
Politecnico di Milano - Department of Electronics, Information and
Bioengineering (DEIB)
Neuro Engineering and medicAl Robotics Lab,
Piazza Leonardo da Vinci, 32
20133 Milan, Italy
www.nearlab.polimi.it

Email: mirkodaniele [dot] comparetti [..] ...
Phone: +39 02 2399 9013
Fax: +39 02 2399 9003

- .... . / -- .. -. -.. / .. ... / .-.. .. -.- . / .- / .--. .- .-. .- -.-.
.... ..- - .
.. - / .-- --- .-. -.- ... / --- -. .-.. -.-- / .-- .... . -. / .. - / ..
... / --- .--. . -.

2013/9/5 Peter Soetens <peter [..] ...>

> Hi Mirko,
>
> On Mon, Aug 26, 2013 at 11:43 AM, <mirkodaniele [dot] comparetti [..] ...>wrote:
>
>> Hi all,
>> I have a problem with a xml schema that works in gnulinux but doesn't in
>> xenomai. I attached to this post the output from gdb.
>> I can say that I'm not using mutexes "intentionally" in my code, while
>> I'm using ports and operations.
>>
>
> Yes, these are Orocos mutexes that protect the internals of RTT against
> concurrent access of user threads
>
>
>>
>> Do anybody of you have an idea about that issue?
>>
>
> Take a look at the output of 'dmesg' , this might explain why it crashes
> your application. Orocos is quite heavy on resources, especially mutexes,
> so you might have hit a limit there. Another option for crashes in Xenomai
> is hitting the stack limit, but we dramatically increased that in the
> latest versions.
>
> Peter
>
>