[Xenomai] Fw:some API about RT_TASK occur Segmentation fault

[Xenomai] Fw:some API about RT_TASK occur Segmentation fault

[dreamtale90]

System configuration:
debian-8.6, 3.18.46-rt50
Xenomai/mercury v3.0.3
Build args: --with-core=mercury --enable-smp --enable-pshared


Now I am need help. Thanks.





-------- Forwarding messages --------
From: "dreamtale90@163.com" <dreamtale90@163.com>
Date: 2017-02-04 12:04:15
To: xenomai@xenomai.org
Subject: some API about RT_TASK occur Segmentation fault

Hi
    During using API about RT_TASK I found some doubt.
1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is NULL will occur Segmentation fault, but in API reference this is valid.
http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8


2) After main task called rt_task_shadow(), program will create other tasks depend on request, but when create second task will  occur Segmentation fault.
Now main task don't call  rt_task_shadow(), program will OK. I don't know why.


3) I write a demo which create the same task three times as follows:
#include <stdio.h>
#include <unistd.h>
#include <alchemy/task.h>
RT_TASK task1;
void func(void *arg)
{
 printf("enter func ...\n");
}
int main()
{
 rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
 sleep(1);
 rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
 sleep(1);
 rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
 //rt_task_delete(&task1);
 return 0;
}


When the third time create , program will  occur Segmentation fault.




 

Re: [Xenomai] Fw:some API about RT_TASK occur Segmentation fault

[Philippe Gerum]

On 02/06/2017 02:30 AM, dreamtale90@163.com wrote:
> System configuration:
> debian-8.6, 3.18.46-rt50
> Xenomai/mercury v3.0.3
> Build args: --with-core=mercury --enable-smp --enable-pshared
> 
> Now I am need help. Thanks.
> 
> 
> -------- Forwarding messages --------
> From: "dreamtale90@163.com" <dreamtale90@163.com>
> Date: 2017-02-04 12:04:15
> To: xenomai@xenomai.org
> Subject: some API about RT_TASK occur Segmentation fault
> Hi
>     During using API about RT_TASK I found some doubt.
> 1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is
> NULL will occur Segmentation fault, but in API reference this is valid.
> http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8
> 
> 2) After main task called rt_task_shadow(), program will create other
> tasks depend on request, but when create second task will  occur
> Segmentation fault.
> Now main task don't call  rt_task_shadow(), program will OK. I don't
> know why.
> 
> 3) I write a demo which create the same task three times as follows:
> #include <stdio.h>
> #include <unistd.h>
> #include <alchemy/task.h>
> RT_TASK task1;
> void func(void *arg)
> {
>  printf("enter func ...\n");
> }
> int main()
> {
>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>  sleep(1);
>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>  sleep(1);
>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>  //rt_task_delete(&task1);
>  return 0;
> }
> 
> When the third time create , program will  occur Segmentation fault.
> 

This bug was fixed in the stable branch some time ago:
git://git.xenomai.org/xenomai-3.git, branch stable-3.0.x

-- 
Philippe.

Re: [Xenomai] Fw:some API about RT_TASK occur Segmentation fault

[dreamtale90]

Thank you for help.
I had git clone git://git.xenomai.org/xenomai-3.git, compile and install it with：
    ./scripts/bootstrap
    ../xenomai-3/configure --with-core=mercury  --enable-smp --enable-pshared --enable-debug=full

    make install
Then run my demo, but the result is the same as before ：
----------------------------------------------------------------------------------------------------------
   0"002.570| WARNING: [main] Xenomai compiled with full debug enabled,
                              very high latencies expected [--enable-debug=full]
enter func ...
enter func ...
rt_task_create: ../../../xenomai-3/lib/copperplate/heapobj-pshared.c:452: alloc_block: Assertion `0' failed.
Aborted
----------------------------------------------------------------------------------------------------------
Yesterday, I add some printf to alloc_block() of heapobj-pshared.c. I found  that segmentaion fault occur when run to the
line 455 "++extent->pagemap[pgnum].bcount;" . At this time the value of "heap->buckets[ilog].freelist" is very big.


Now I think whether my xenomai configuration or system environment have something wrong.
Besides, my system(x86_64bit) run in a 8 cores Oracle virtual machine, the host cpu is Inter(R) xeon(R) CPU E5-2630 v4 @ 2.20GHz.
I don't know whether these information is useful.

Thanks in advance, best regards.





At 2017-02-06 21:50:13, "Philippe Gerum" <rpm@xenomai.org> wrote:
>On 02/06/2017 02:30 AM, dreamtale90@163.com wrote:
>> System configuration:
>> debian-8.6, 3.18.46-rt50
>> Xenomai/mercury v3.0.3
>> Build args: --with-core=mercury --enable-smp --enable-pshared
>> 
>> Now I am need help. Thanks.
>> 
>> 
>> -------- Forwarding messages --------
>> From: "dreamtale90@163.com" <dreamtale90@163.com>
>> Date: 2017-02-04 12:04:15
>> To: xenomai@xenomai.org
>> Subject: some API about RT_TASK occur Segmentation fault
>> Hi
>>     During using API about RT_TASK I found some doubt.
>> 1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is
>> NULL will occur Segmentation fault, but in API reference this is valid.
>> http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8
>> 
>> 2) After main task called rt_task_shadow(), program will create other
>> tasks depend on request, but when create second task will  occur
>> Segmentation fault.
>> Now main task don't call  rt_task_shadow(), program will OK. I don't
>> know why.
>> 
>> 3) I write a demo which create the same task three times as follows:
>> #include <stdio.h>
>> #include <unistd.h>
>> #include <alchemy/task.h>
>> RT_TASK task1;
>> void func(void *arg)
>> {
>>  printf("enter func ...\n");
>> }
>> int main()
>> {
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  sleep(1);
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  sleep(1);
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  //rt_task_delete(&task1);
>>  return 0;
>> }
>> 
>> When the third time create , program will  occur Segmentation fault.
>> 
>
>This bug was fixed in the stable branch some time ago:
>git://git.xenomai.org/xenomai-3.git, branch stable-3.0.x
>
>-- 
>Philippe.

Re: [Xenomai] some API about RT_TASK occur Segmentation fault

[dreamtale90]

Just now when I compiling xenomai, removed "--enable-pshared" configuration option. My demo is normal finally, and print out third sentence "enter func ..." without segmentaion fault.
Although the question is solved temporarily , but still don't know the reason. So I'm still worried about my system environment have some problems.

Thanks again, best regards.


发自网易邮箱大师


在2017年02月07日 14:18，dreamtale90@163.com 写道：
Thank you for help.
I had git clone git://git.xenomai.org/xenomai-3.git, compile and install it with：
    ./scripts/bootstrap
    ../xenomai-3/configure --with-core=mercury  --enable-smp --enable-pshared --enable-debug=full

    make install
Then run my demo, but the result is the same as before ：
----------------------------------------------------------------------------------------------------------
   0"002.570| WARNING: [main] Xenomai compiled with full debug enabled,
                              very high latencies expected [--enable-debug=full]
enter func ...
enter func ...
rt_task_create: ../../../xenomai-3/lib/copperplate/heapobj-pshared.c:452: alloc_block: Assertion `0' failed.
Aborted
----------------------------------------------------------------------------------------------------------
Yesterday, I add some printf to alloc_block() of heapobj-pshared.c. I found  that segmentaion fault occur when run to the
line 455 "++extent->pagemap[pgnum].bcount;" . At this time the value of "heap->buckets[ilog].freelist" is very big.


Now I think whether my xenomai configuration or system environment have something wrong.
Besides, my system(x86_64bit) run in a 8 cores Oracle virtual machine, the host cpu is Inter(R) xeon(R) CPU E5-2630 v4 @ 2.20GHz.
I don't know whether these information is useful.

Thanks in advance, best regards.





At 2017-02-06 21:50:13, "Philippe Gerum" <rpm@xenomai.org> wrote:
>On 02/06/2017 02:30 AM, dreamtale90@163.com wrote:
>> System configuration:
>> debian-8.6, 3.18.46-rt50
>> Xenomai/mercury v3.0.3
>> Build args: --with-core=mercury --enable-smp --enable-pshared
>> 
>> Now I am need help. Thanks.
>> 
>> 
>> -------- Forwarding messages --------
>> From: "dreamtale90@163.com" <dreamtale90@163.com>
>> Date: 2017-02-04 12:04:15
>> To: xenomai@xenomai.org
>> Subject: some API about RT_TASK occur Segmentation fault
>> Hi
>>     During using API about RT_TASK I found some doubt.
>> 1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is
>> NULL will occur Segmentation fault, but in API reference this is valid.
>> http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8
>> 
>> 2) After main task called rt_task_shadow(), program will create other
>> tasks depend on request, but when create second task will  occur
>> Segmentation fault.
>> Now main task don't call  rt_task_shadow(), program will OK. I don't
>> know why.
>> 
>> 3) I write a demo which create the same task three times as follows:
>> #include <stdio.h>
>> #include <unistd.h>
>> #include <alchemy/task.h>
>> RT_TASK task1;
>> void func(void *arg)
>> {
>>  printf("enter func ...\n");
>> }
>> int main()
>> {
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  sleep(1);
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  sleep(1);
>>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>  //rt_task_delete(&task1);
>>  return 0;
>> }
>> 
>> When the third time create , program will  occur Segmentation fault.
>> 
>
>This bug was fixed in the stable branch some time ago:
>git://git.xenomai.org/xenomai-3.git, branch stable-3.0.x
>
>-- 
>Philippe.





 

Re: [Xenomai] some API about RT_TASK occur Segmentation fault

[Philippe Gerum]

On 02/07/2017 03:19 PM, dreamtale90@163.com wrote:
> Just now when I compiling xenomai, removed "--enable-pshared"
> configuration option. My demo is normal finally, and print out third
> sentence "enter func ..." without segmentaion fault.

I believe that you are not using the right code. Did you pull the code
from the stable-3.0.x branch as mentioned earlier?

If you did so, you should see this output when passing the --version
switch to your program:

based on Xenomai/mercury v3.0.3 -- #00e3d01 (2017-01-27 10:16:11 +0100)

> Although the question is solved temporarily , but still don't know the
> reason. So I'm still worried about my system environment have some problems.
> 
> Thanks again, best regards.
> 
> 
> 发自网易邮箱大师
> 
>     在2017年02月07日 14:18，dreamtale90@163.com
>     <mailto:dreamtale90@163.com> 写道：
>     Thank you for help.
>     I had git clone git://git.xenomai.org/xenomai-3.git, compile and
>     install it with：
>         ./scripts/bootstrap
>         ../xenomai-3/configure --with-core=mercury  --enable-smp
>     --enable-pshared --enable-debug=full
>         make install
>     Then run my demo, but the result is the same as before ：
>     ----------------------------------------------------------------------------------------------------------
>        0"002.570| WARNING: [main] Xenomai compiled with full debug enabled,
>                                   very high latencies expected
>     [--enable-debug=full]
>     enter func ...
>     enter func ...
>     rt_task_create:
>     ../../../xenomai-3/lib/copperplate/heapobj-pshared.c:452:
>     alloc_block: Assertion `0' failed.
>     Aborted
>     ----------------------------------------------------------------------------------------------------------
>     Yesterday, I add some printf to alloc_block() of heapobj-pshared.c.
>     I found  that segmentaion fault occur when run to the
>     line 455 "++extent->pagemap[pgnum].bcount;" . At this time the value
>     of "heap->buckets[ilog].freelist" is very big.
> 
>     Now I think whether my xenomai
>     configuration or system environment have something wrong.
>     Besides, my system(x86_64bit) run in a 8 cores Oracle virtual
>     machine, the host cpu is Inter(R) xeon(R) CPU E5-2630 v4 @ 2.20GHz.
>     I don't know whether these information is useful.
> 
>     Thanks in advance, best regards.
> 
> 
> 
>     At 2017-02-06 21:50:13, "Philippe Gerum" <rpm@xenomai.org <mailto:rpm@xenomai.org>> wrote:
>     >On 02/06/2017 02:30 AM, dreamtale90@163.com <mailto:dreamtale90@163.com> wrote:
>     >> System configuration:
>     >> debian-8.6, 3.18.46-rt50
>     >> Xenomai/mercury v3.0.3
>     >> Build args: --with-core=mercury --enable-smp --enable-pshared
>     >> 
>     >> Now I am need help. Thanks.
>     >> 
>     >> 
>     >> -------- Forwarding messages --------
>     >> From: "dreamtale90@163.com <mailto:dreamtale90@163.com>" <dreamtale90@163.com <mailto:dreamtale90@163.com>>
>     >> Date: 2017-02-04 12:04:15
>     >> To: xenomai@xenomai.org <mailto:xenomai@xenomai.org>
>     >> Subject: some API about RT_TASK occur Segmentation fault
>     >> Hi
>     >>     During using API about RT_TASK I found some doubt.
>     >> 1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is
>     >> NULL will occur Segmentation fault, but in API reference this is valid.
>     >> http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8
>     >> 
>     >> 2) After main task called rt_task_shadow(), program will create other
>     >> tasks depend on request, but when create second task will  occur
>     >> Segmentation fault.
>     >> Now main task don't call  rt_task_shadow(), program will OK. I don't
>     >> know why.
>     >> 
>     >> 3) I write a demo which create the same task three times as follows:
>     >> #include <stdio.h>
>     >> #include <unistd.h>
>     >> #include <alchemy/task.h>
>     >> RT_TASK task1;
>     >> void func(void *arg)
>     >> {
>     >>  printf("enter func ...\n");
>     >> }
>     >> int main()
>     >> {
>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>     >>  sleep(1);
>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>     >>  sleep(1);
>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>     >>  //rt_task_delete(&task1);
>     >>  return 0;
>     >> }
>     >> 
>     >> When the third time create , program will  occur Segmentation fault.
>     >> 
>     >
>     >This bug was fixed in the stable branch some time ago:
>     >git://git.xenomai.org/xenomai-3.git, branch stable-3.0.x
>     >
>     >-- 
>     >Philippe.
> 
> 
> 
>      
> 


-- 
Philippe.

Re: [Xenomai] some API about RT_TASK occur Segmentation fault

[dreamtale90]

Philippe,


Thank you very much.Exactly as you say, the reason is that my code isn't lastest.

Best regards.





At 2017-02-07 23:17:28, "Philippe Gerum" <rpm@xenomai.org> wrote:
>On 02/07/2017 03:19 PM, dreamtale90@163.com wrote:
>> Just now when I compiling xenomai, removed "--enable-pshared"
>> configuration option. My demo is normal finally, and print out third
>> sentence "enter func ..." without segmentaion fault.
>
>I believe that you are not using the right code. Did you pull the code
>from the stable-3.0.x branch as mentioned earlier?
>
>If you did so, you should see this output when passing the --version
>switch to your program:
>
>based on Xenomai/mercury v3.0.3 -- #00e3d01 (2017-01-27 10:16:11 +0100)
>
>> Although the question is solved temporarily , but still don't know the
>> reason. So I'm still worried about my system environment have some problems.
>> 
>> Thanks again, best regards.
>> 
>> 
>> 发自网易邮箱大师
>> 
>>     在2017年02月07日 14:18，dreamtale90@163.com
>>     <mailto:dreamtale90@163.com> 写道：
>>     Thank you for help.
>>     I had git clone git://git.xenomai.org/xenomai-3.git, compile and
>>     install it with：
>>         ./scripts/bootstrap
>>         ../xenomai-3/configure --with-core=mercury  --enable-smp
>>     --enable-pshared --enable-debug=full
>>         make install
>>     Then run my demo, but the result is the same as before ：
>>     ----------------------------------------------------------------------------------------------------------
>>        0"002.570| WARNING: [main] Xenomai compiled with full debug enabled,
>>                                   very high latencies expected
>>     [--enable-debug=full]
>>     enter func ...
>>     enter func ...
>>     rt_task_create:
>>     ../../../xenomai-3/lib/copperplate/heapobj-pshared.c:452:
>>     alloc_block: Assertion `0' failed.
>>     Aborted
>>     ----------------------------------------------------------------------------------------------------------
>>     Yesterday, I add some printf to alloc_block() of heapobj-pshared.c.
>>     I found  that segmentaion fault occur when run to the
>>     line 455 "++extent->pagemap[pgnum].bcount;" . At this time the value
>>     of "heap->buckets[ilog].freelist" is very big.
>> 
>>     Now I think whether my xenomai
>>     configuration or system environment have something wrong.
>>     Besides, my system(x86_64bit) run in a 8 cores Oracle virtual
>>     machine, the host cpu is Inter(R) xeon(R) CPU E5-2630 v4 @ 2.20GHz.
>>     I don't know whether these information is useful.
>> 
>>     Thanks in advance, best regards.
>> 
>> 
>> 
>>     At 2017-02-06 21:50:13, "Philippe Gerum" <rpm@xenomai.org <mailto:rpm@xenomai.org>> wrote:
>>     >On 02/06/2017 02:30 AM, dreamtale90@163.com <mailto:dreamtale90@163.com> wrote:
>>     >> System configuration:
>>     >> debian-8.6, 3.18.46-rt50
>>     >> Xenomai/mercury v3.0.3
>>     >> Build args: --with-core=mercury --enable-smp --enable-pshared
>>     >> 
>>     >> Now I am need help. Thanks.
>>     >> 
>>     >> 
>>     >> -------- Forwarding messages --------
>>     >> From: "dreamtale90@163.com <mailto:dreamtale90@163.com>" <dreamtale90@163.com <mailto:dreamtale90@163.com>>
>>     >> Date: 2017-02-04 12:04:15
>>     >> To: xenomai@xenomai.org <mailto:xenomai@xenomai.org>
>>     >> Subject: some API about RT_TASK occur Segmentation fault
>>     >> Hi
>>     >>     During using API about RT_TASK I found some doubt.
>>     >> 1) rt_task_inquire(RT_TASK * task, RT_TASK_INFO * info ), if info is
>>     >> NULL will occur Segmentation fault, but in API reference this is valid.
>>     >> http://www.xenomai.org/documentation/xenomai-3/html/xeno3prm/group__alchemy__task.html#ga53bad6acf5496f8008e594954507dce8
>>     >> 
>>     >> 2) After main task called rt_task_shadow(), program will create other
>>     >> tasks depend on request, but when create second task will  occur
>>     >> Segmentation fault.
>>     >> Now main task don't call  rt_task_shadow(), program will OK. I don't
>>     >> know why.
>>     >> 
>>     >> 3) I write a demo which create the same task three times as follows:
>>     >> #include <stdio.h>
>>     >> #include <unistd.h>
>>     >> #include <alchemy/task.h>
>>     >> RT_TASK task1;
>>     >> void func(void *arg)
>>     >> {
>>     >>  printf("enter func ...\n");
>>     >> }
>>     >> int main()
>>     >> {
>>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>     >>  sleep(1);
>>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>     >>  sleep(1);
>>     >>  rt_task_spawn(&task1, "task1", 0, 80, 0, func, NULL);
>>     >>  //rt_task_delete(&task1);
>>     >>  return 0;
>>     >> }
>>     >> 
>>     >> When the third time create , program will  occur Segmentation fault.
>>     >> 
>>     >
>>     >This bug was fixed in the stable branch some time ago:
>>     >git://git.xenomai.org/xenomai-3.git, branch stable-3.0.x
>>     >
>>     >-- 
>>     >Philippe.
>> 
>> 
>> 
>>      
>> 
>
>
>-- 
>Philippe.