[Xenomai] [xenomai-forge] Task blocked in t_delete

[Xenomai] [xenomai-forge] Task blocked in t_delete

[Ronny Meeus]

Hello

we are using the xenomai-forge mercury core together with the pSOS interface.

We observe following issue: we have 3 tasks at different priorities
(see test code attached).
The task with the middle priority is running forever (while (1)).

The higest prio task is responsible cleaning up the 2 other tasks.
If the middle prio task is deleted first, followed by the low prio, it works.
If the low prio task is deleted first, the high prio task just blocks
on the delete and the system hangs in that phase.

My assumption is that this is a bug in the delete functionality.
Is this a known issue?


Ronny
-------------- next part --------------
A non-text attachment was scrubbed...
Name: task_hang.c
Type: text/x-csrc
Size: 1645 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150430/aa0c3e0f/attachment.c>

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Philippe Gerum]

On 04/30/2015 10:14 AM, Ronny Meeus wrote:
> Hello
> 
> we are using the xenomai-forge mercury core together with the pSOS interface.
> 
> We observe following issue: we have 3 tasks at different priorities
> (see test code attached).
> The task with the middle priority is running forever (while (1)).
> 
> The higest prio task is responsible cleaning up the 2 other tasks.
> If the middle prio task is deleted first, followed by the low prio, it works.
> If the low prio task is deleted first, the high prio task just blocks
> on the delete and the system hangs in that phase.
> 

I could not reproduce this bug with your test case unmodified after an
hour running this program in loop. Could you paste the output of
--dump-config so that I can align on your settings?

> My assumption is that this is a bug in the delete functionality.
> Is this a known issue?
> 

No report from such issue.

-- 
Philippe.

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Ronny Meeus]

On Thu, Apr 30, 2015 at 12:01 PM, Philippe Gerum <rpm@xenomai.org> wrote:
> On 04/30/2015 10:14 AM, Ronny Meeus wrote:
>> Hello
>>
>> we are using the xenomai-forge mercury core together with the pSOS interface.
>>
>> We observe following issue: we have 3 tasks at different priorities
>> (see test code attached).
>> The task with the middle priority is running forever (while (1)).
>>
>> The higest prio task is responsible cleaning up the 2 other tasks.
>> If the middle prio task is deleted first, followed by the low prio, it works.
>> If the low prio task is deleted first, the high prio task just blocks
>> on the delete and the system hangs in that phase.
>>
>
> I could not reproduce this bug with your test case unmodified after an
> hour running this program in loop. Could you paste the output of
> --dump-config so that I can align on your settings?
>

CONFIG_MMU=1
CONFIG_SMP=1
CONFIG_XENO_ASYNC_CANCEL=1
CONFIG_XENO_BUILD_ARGS=" '--target=armeb-buildroot-linux-gnueabi'
'--host=armeb-buildroot-linux-gnueabi'
'--build=x86_64-unknown-linux-gnu' '--prefix=/usr'
'--exec-prefix=/usr' '--sysconfdir=/etc' '--program-prefix='
'--disable-gtk-doc' '--disable-doc' '--disable-docs'
'--disable-documentation' '--with-xmlto=no' '--with-fop=no'
'--enable-ipv6' '--enable-static' '--enable-shared'
'--with-core=mercury' '--includedir=/usr/include/xenomai-forge'
'--disable-doc-install' '--enable-lores-clock' '--enable-smp'
'--enable-async-cancel' '--enable-registry'
'build_alias=x86_64-unknown-linux-gnu'
'host_alias=armeb-buildroot-linux-gnueabi'
'target_alias=armeb-buildroot-linux-gnueabi'
'CC=/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/host/usr/bin/armeb-linux-gcc'
'CFLAGS=-D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE
-D_FILE_OFFSET_BITS=64  -pipe -mcpu=arm1176jz-s -Os ' 'LDFLAGS='
'CPPFLAGS=-D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE
-D_FILE_OFFSET_BITS=64'
'CPP=/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/host/usr/bin/armeb-linux-cpp'
'PKG_CONFIG=/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/host/usr/bin/pkg-config'"
CONFIG_XENO_BUILD_STRING="x86_64-unknown-linux-gnu"
CONFIG_XENO_COMPILER="gcc version 4.4.3 (crosstool-NG 1.19.0) "
CONFIG_XENO_DEFAULT_PERIOD=1000000
CONFIG_XENO_FORTIFY=1
CONFIG_XENO_HOST_STRING="armeb-buildroot-linux-gnueabi"
CONFIG_XENO_MERCURY=1
CONFIG_XENO_PREFIX="/usr"
CONFIG_XENO_REGISTRY=1
CONFIG_XENO_REVISION_LEVEL=7
CONFIG_XENO_SANITY=1
CONFIG_XENO_TLSF=1
CONFIG_XENO_TLS_MODEL="initial-exec"
CONFIG_XENO_UAPI_LEVEL=8
CONFIG_XENO_VERSION_MAJOR=2
CONFIG_XENO_VERSION_MINOR=99
CONFIG_XENO_X86_VSYSCALL=1
---
CONFIG_XENO_ARM_SA1100 is OFF
CONFIG_XENO_ARM_TSC_TYPE is OFF
CONFIG_XENO_COBALT is OFF
CONFIG_XENO_COPPERPLATE_CLOCK_RESTRICTED is OFF
CONFIG_XENO_CPU_XSC3 is OFF
CONFIG_XENO_DEBUG is OFF
CONFIG_XENO_DEBUG_FULL is OFF
CONFIG_XENO_LIBS_DLOPEN is OFF
CONFIG_XENO_LORES_CLOCK_DISABLED is OFF
CONFIG_XENO_PSHARED is OFF
CONFIG_XENO_RAW_CLOCK_ENABLED is OFF
CONFIG_XENO_VALGRIND_API is OFF

>> My assumption is that this is a bug in the delete functionality.
>> Is this a known issue?
>>
>
> No report from such issue.
>

If I run the test I see:

/tmp # ./task_delete
Start of test
Create and start main task:Create task M:
Create task L:
Start task M:
Start task L:
M: entry
L: entry
M: Dead loop starts
Cleaning ..
End of test

Here is the piece of the code that goes wrong:

  printf("Cleaning ..\r\n");
  /* Remove tasks */
  t_delete(L_id);
  /* Followed code is not executed because t_delete is somehow blocked */
  printf("Removed L task id = 0x%lx\r\n", L_id);
  t_delete(M_id);
  printf("Removed M task id = 0x%lx\r\n", M_id);
}

Since the trace "Removed L task id..." is not generated, my assumption
it that the t_delete(L_id) blocks.


Ronny

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Philippe Gerum]

On 04/30/2015 02:39 PM, Ronny Meeus wrote:
> If I run the test I see:
> 
> /tmp # ./task_delete
> Start of test
> Create and start main task:Create task M:
> Create task L:
> Start task M:
> Start task L:
> M: entry
> L: entry
> M: Dead loop starts
> Cleaning ..
> End of test
> 
> Here is the piece of the code that goes wrong:
> 
>   printf("Cleaning ..\r\n");
>   /* Remove tasks */
>   t_delete(L_id);
>   /* Followed code is not executed because t_delete is somehow blocked */
>   printf("Removed L task id = 0x%lx\r\n", L_id);
>   t_delete(M_id);
>   printf("Removed M task id = 0x%lx\r\n", M_id);
> }
> 
> Since the trace "Removed L task id..." is not generated, my assumption
> it that the t_delete(L_id) blocks.
> 
> 

Could you attach GDB to the hung process, then dump a backtrace for all
the remaining threads?

-- 
Philippe.

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Ronny Meeus]

On Thu, Apr 30, 2015 at 3:09 PM, Philippe Gerum <rpm@xenomai.org> wrote:
> On 04/30/2015 02:39 PM, Ronny Meeus wrote:
>> If I run the test I see:
>>
>> /tmp # ./task_delete
>> Start of test
>> Create and start main task:Create task M:
>> Create task L:
>> Start task M:
>> Start task L:
>> M: entry
>> L: entry
>> M: Dead loop starts
>> Cleaning ..
>> End of test
>>
>> Here is the piece of the code that goes wrong:
>>
>>   printf("Cleaning ..\r\n");
>>   /* Remove tasks */
>>   t_delete(L_id);
>>   /* Followed code is not executed because t_delete is somehow blocked */
>>   printf("Removed L task id = 0x%lx\r\n", L_id);
>>   t_delete(M_id);
>>   printf("Removed M task id = 0x%lx\r\n", M_id);
>> }
>>
>> Since the trace "Removed L task id..." is not generated, my assumption
>> it that the t_delete(L_id) blocks.
>>
>>
>
> Could you attach GDB to the hung process, then dump a backtrace for all
> the remaining threads?
>
> --
> Philippe.

Philippe

I have put a breakpoint in main @ line 62:

57        printf("Create and start main task:");
58        t_create("MaIn",10, 16000, 16000,0,&tid);
59        ret = t_start(tid, 0, main_task, NULL);
60        if (0 != ret) {printf("Failed %lu\r\n",ret);return -1;}
61        sleep(10);
62        t_delete(tid);
63        printf("End of test\n\r");
64        return 0;

and I generated a backtrace for all threads, see below.
It looks like the "MaIn" task is wating on a semaphore.

(gdb) set pagination off
(gdb) info threads
  Id   Target Id         Frame
  6    Thread 16395      0xb6eb6c34 in nanosleep () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
  5    Thread 16394      m_task () at task_delete.c:12
  4    Thread 16374      0xb6f9fcf4 in do_sigwait () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
* 3    Thread 16372      0xb6f9e900 in read () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
  2    Thread 16393      0xb6f9d700 in sem_wait@@GLIBC_2.4 () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
  1    Thread 16325      main (argc=1, argv=0x11110, envp=0xbefffc08)
at task_delete.c:62

   0    16325    16322   20 N    0     t (main)
   0    16372    16322   20 N    0     t (task_delete)
   0    16374    16322  -95 F    0     t (timer-internal)
   0    16393    16322  -11 F    0     t (MaIn)
   0    16394    16322   -6 F    0     t (M)
   0    16395    16322   -5 F    0     t (L)

(gdb) thread apply all bt full

Thread 6 (Thread 16395):
#0  0xb6eb6c34 in nanosleep () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#1  0xb6eb69cc in sleep () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#2  0x00008718 in l_task () at task_delete.c:18
No locals.
#3  0xb6fd2400 in task_trampoline (arg=<optimized out>) at task.c:225
        task = 0x23218
        param_ex = {__sched_priority = -1224959724, sched_u = {rr =
{__sched_rr_quantum = {tv_sec = 0, tv_nsec = -1227485752}}}}
        svc = {cancel_type = 1}
        __FUNCTION__ = "task_trampoline"
#4  0xb6fb6a78 in thread_trampoline (arg=<optimized out>) at
internal.c:234
        cta = <optimized out>
        _cta = {stacksize = 131072, detachstate = 1, policy = 1,
param_ex = {__sched_priority = 4, sched_u = {rr = {__sched_rr_quantum
= {tv_sec = -1227287440, tv_nsec = 16}}}}, prologue = 0xb6fd2424
<task_prologue>, run = 0xb6fd22f4 <task_trampoline>, arg = 0x23218,
__reserved = {status = -38, warm = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\266\364\265\300",
__align = 0}, released = 0xb6f4c220}}
        released = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\000\000\000",
__align = 0}
        ret = 0
        __FUNCTION__ = "thread_trampoline"
#5  0xb6f956fc in start_thread () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#6  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#7  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt
stack?)

Thread 5 (Thread 16394):
#0  m_task () at task_delete.c:12
No locals.
#1  0xb6fd2400 in task_trampoline (arg=<optimized out>) at task.c:225
        task = 0x22ea0
        param_ex = {__sched_priority = -1224959724, sched_u = {rr =
{__sched_rr_quantum = {tv_sec = 0, tv_nsec = -1227354680}}}}
        svc = {cancel_type = 1}
        __FUNCTION__ = "task_trampoline"
#2  0xb6fb6a78 in thread_trampoline (arg=<optimized out>) at
internal.c:234
        cta = <optimized out>
        _cta = {stacksize = 131072, detachstate = 1, policy = 1,
param_ex = {__sched_priority = 5, sched_u = {rr = {__sched_rr_quantum
= {tv_sec = -1227287440, tv_nsec = 44}}}}, prologue = 0xb6fd2424
<task_prologue>, run = 0xb6fd22f4 <task_trampoline>, arg = 0x22ea0,
__reserved = {status = -38, warm = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\266\364\265\300",
__align = 0}, released = 0xb6f4c220}}
        released = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\000\000\000",
__align = 0}
        ret = 0
        __FUNCTION__ = "thread_trampoline"
#3  0xb6f956fc in start_thread () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#4  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#5  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt
stack?)

Thread 4 (Thread 16374):
#0  0xb6f9fcf4 in do_sigwait () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#1  0xb6f9fd90 in sigwait () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#2  0xb6fb9c00 in timerobj_server (arg=<optimized out>) at
timerobj.c:141
        now = {tv_sec = 199495, tv_nsec = 940506114}
        value = {tv_sec = 0, tv_nsec = -1225166640}
        interval = {tv_sec = -1227027076, tv_nsec = 0}
        tmobj = 0xb6fc9ce8
        tmp = 0xb6fc9ce8
        set = {__val = {8192, 0 <repeats 31 times>}}
        sig = 14
        sev = {sigev_value = {sival_int = 0, sival_ptr = 0x0},
sigev_signo = 14, sigev_notify = 4, _sigev_un = {_pad = {16374, 0
<repeats 12 times>}, _tid = 16374, _sigev_thread = {_function =
0x3ff6, _attribute = 0x0}}}
        periodic_timer = 0x30b47
        it = {it_interval = {tv_sec = 0, tv_nsec = 10000000}, it_value
= {tv_sec = 0, tv_nsec = 10000000}}
        ticks = 983
        __FUNCTION__ = "timerobj_server"
#3  0xb6fb6a78 in thread_trampoline (arg=<optimized out>) at
internal.c:234
        cta = <optimized out>
        _cta = {stacksize = 262144, detachstate = 1, policy = 1,
param_ex = {__sched_priority = 94, sched_u = {rr = {__sched_rr_quantum
= {tv_sec = 1, tv_nsec = 5}}}}, prologue = 0xb6fb9714
<server_prologue>, run = 0xb6fb9b3c <timerobj_server>, arg = 0x0,
__reserved = {status = -38, warm = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\001\266\374\231\024",
__align = 0}, released = 0xbefffb94}}
        released = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\000\000\000",
__align = 0}
        ret = 0
        __FUNCTION__ = "thread_trampoline"
#4  0xb6f956fc in start_thread () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#5  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#6  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt
stack?)

Thread 3 (Thread 16372):
#0  0xb6f9e900 in read () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#1  0xb6f9e1bc in __pthread_enable_asynccancel () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#2  0x00021000 in ?? ()
No symbol table info available.
#3  0x00021000 in ?? ()
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 2 (Thread 16393):
#0  0xb6f9d700 in sem_wait@@GLIBC_2.4 () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#1  0xb6fb87f8 in cancel_sync (thobj=<optimized out>) at threadobj.c:1274
        oldstate = 0
        ret = -1227484048
        sem = 0x235c0
#2  0xb6fb88d8 in threadobj_cancel (thobj=0x232a0) at threadobj.c:1310
No locals.
#3  0xb6fd2214 in t_delete (tid=143896) at task.c:713
        task = <optimized out>
        svc = {cancel_type = 1}
        ret = <optimized out>
#4  0x00008894 in main_task () at task_delete.c:45
        M_id = 143008
        L_id = 143896
        ret = 0
#5  0xb6fd2400 in task_trampoline (arg=<optimized out>) at task.c:225
        task = 0x22b10
        param_ex = {__sched_priority = -1224959724, sched_u = {rr =
{__sched_rr_quantum = {tv_sec = 0, tv_nsec = -1227289144}}}}
        svc = {cancel_type = 1}
        __FUNCTION__ = "task_trampoline"
#6  0xb6fb6a78 in thread_trampoline (arg=<optimized out>) at internal.c:234
        cta = <optimized out>
        _cta = {stacksize = 32000, detachstate = 1, policy = 1,
param_ex = {__sched_priority = 10, sched_u = {rr = {__sched_rr_quantum
= {tv_sec = -1224757248, tv_nsec = -1224842860}}}}, prologue =
0xb6fd2424 <task_prologue>, run = 0xb6fd22f4 <task_trampoline>, arg =
0x22b10, __reserved = {status = -38, warm = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\001\000\000\000",
__align = 0}, released = 0x0}}
        released = {__size =
"\000\000\000\000\000\000\000\200\000\000\000\000\000\000\000",
__align = 0}
        ret = 0
        __FUNCTION__ = "thread_trampoline"
#7  0xb6f956fc in start_thread () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libpthread.so.0
No symbol table info available.
#8  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
#9  0xb6ee8c08 in clone () from
/repo/meeusr/sw-memmngt/vobs/esam/build/reborn/buildroot-isam-reborn-arm-nrnti/output/staging/lib/libc.so.6
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 1 (Thread 16325):
#0  main (argc=1, argv=0x11110, envp=0xbefffc08) at task_delete.c:62
        tid = 142096
        ret = 0
(gdb)

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Philippe Gerum]

On 04/30/2015 03:55 PM, Ronny Meeus wrote:
> On Thu, Apr 30, 2015 at 3:09 PM, Philippe Gerum <rpm@xenomai.org> wrote:
>> On 04/30/2015 02:39 PM, Ronny Meeus wrote:
>>> If I run the test I see:
>>>
>>> /tmp # ./task_delete
>>> Start of test
>>> Create and start main task:Create task M:
>>> Create task L:
>>> Start task M:
>>> Start task L:
>>> M: entry
>>> L: entry
>>> M: Dead loop starts
>>> Cleaning ..
>>> End of test
>>>
>>> Here is the piece of the code that goes wrong:
>>>
>>>   printf("Cleaning ..\r\n");
>>>   /* Remove tasks */
>>>   t_delete(L_id);
>>>   /* Followed code is not executed because t_delete is somehow blocked */
>>>   printf("Removed L task id = 0x%lx\r\n", L_id);
>>>   t_delete(M_id);
>>>   printf("Removed M task id = 0x%lx\r\n", M_id);
>>> }
>>>
>>> Since the trace "Removed L task id..." is not generated, my assumption
>>> it that the t_delete(L_id) blocks.
>>>
>>>
>>
>> Could you attach GDB to the hung process, then dump a backtrace for all
>> the remaining threads?
>>
>> --
>> Philippe.
> 
> Philippe
> 
> I have put a breakpoint in main @ line 62:
> 
> 57        printf("Create and start main task:");
> 58        t_create("MaIn",10, 16000, 16000,0,&tid);
> 59        ret = t_start(tid, 0, main_task, NULL);
> 60        if (0 != ret) {printf("Failed %lu\r\n",ret);return -1;}
> 61        sleep(10);
> 62        t_delete(tid);
> 63        printf("End of test\n\r");
> 64        return 0;
> 
> and I generated a backtrace for all threads, see below.
> It looks like the "MaIn" task is wating on a semaphore.

[snip]

This is a priority inversion issue caused by the handshake mechanism
between the caller of t_delete() and the killed task finalizer. I
eventually reproduced it after I figured out that pinning all tasks on
the same CPU was required to trigger such bug. This patch should help:

http://git.xenomai.org/xenomai-3.git/commit/?h=next&id=8431d7149405235fb0aebb53ed32a537ad968c7e

-- 
Philippe.

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Ronny Meeus]

> This is a priority inversion issue caused by the handshake mechanism
> between the caller of t_delete() and the killed task finalizer. I
> eventually reproduced it after I figured out that pinning all tasks on
> the same CPU was required to trigger such bug. This patch should help:
>
> http://git.xenomai.org/xenomai-3.git/commit/?h=next&id=8431d7149405235fb0aebb53ed32a537ad968c7e
>

I will try the patch on Monday.
(I did run the example on a single core CPU).

Thanks for the quick feedback.

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Ronny Meeus]

On Sat, May 2, 2015 at 8:48 PM, Ronny Meeus <ronny.meeus@gmail.com> wrote:
>> This is a priority inversion issue caused by the handshake mechanism
>> between the caller of t_delete() and the killed task finalizer. I
>> eventually reproduced it after I figured out that pinning all tasks on
>> the same CPU was required to trigger such bug. This patch should help:
>>
>> http://git.xenomai.org/xenomai-3.git/commit/?h=next&id=8431d7149405235fb0aebb53ed32a537ad968c7e
>>
>
> I will try the patch on Monday.
> (I did run the example on a single core CPU).
>
> Thanks for the quick feedback.

Philippe,
the patch seems to solve the issue.
Thanks

Just for my information, is there somewhere documentation about the
remote agent?

Ronny

Re: [Xenomai] [xenomai-forge] Task blocked in t_delete

[Philippe Gerum]

On 05/05/2015 08:04 AM, Ronny Meeus wrote:
> On Sat, May 2, 2015 at 8:48 PM, Ronny Meeus <ronny.meeus@gmail.com> wrote:
>>> This is a priority inversion issue caused by the handshake mechanism
>>> between the caller of t_delete() and the killed task finalizer. I
>>> eventually reproduced it after I figured out that pinning all tasks on
>>> the same CPU was required to trigger such bug. This patch should help:
>>>
>>> http://git.xenomai.org/xenomai-3.git/commit/?h=next&id=8431d7149405235fb0aebb53ed32a537ad968c7e
>>>
>>
>> I will try the patch on Monday.
>> (I did run the example on a single core CPU).
>>
>> Thanks for the quick feedback.
> 
> Philippe,
> the patch seems to solve the issue.
> Thanks
> 
> Just for my information, is there somewhere documentation about the
> remote agent?
> 

No documentation so far. Basically it's an internal proxy thread for
carrying out remote POSIX requests that are otherwise restricted to the
local process. This only applies when --enable-pshared is in effect, and
a multi-process application does remote thread control. So far we only
need to proxy pthread_setschedparam() and pthread_cancel().

-- 
Philippe.