[Xenomai] Assertion current-magic 0 failed

[Xenomai] Assertion current-magic 0 failed

[Stéphane Reichert]

Hi,

I am testing the new version 3.0.7 on our system (iMX 6 on 4.1.46 kernel), 
and I would like to know
if anybody else encountered this error :

threadobj.c:1344: threadobj_prologue: Assertion `current->magic == 0' 
failed.

I compile the kernel with --enable-pshared --enable 
registry --enable-debug=full.

Here is a backtrace I get on the debug signal callback:

# Mode switch (reason: triggered fault), aborting. Backtrace:
Retrieve backtrace...
./tests_integration_lib_ipc(Unable to handle kernel NULL pointer dereference 
at virtual address 00000004
4x180)[0[x01020206000]08l0e0r0+
]]/ l*ipbd=39180831/libc.so.,6 (*+p0txe2=50105000)0[00, 
*ppte=000000000x76ca3150

/lib/libcopperplate.so.0(syncobj_lIonctke+r0nxa8l4 error: Oops: 17 [#1] 
PREEMPT SMP ARM
l0x76Meo1dau7lee8s]
 i/nlkiebd/in:libcopper pcltarte.so.0(syncluster_fin 
dcocbmj+0x34)[0x76e18eec]
/lib/li buaslbc_hfe_mayc.mso.0(al u_serialchemy_bind 
_uosbbj_efc_te+c0mx98)[0x76e 3ge_1ccd0c]
nlib/libalchemy.so. 0u(_rett_hheerap_bind+0x30 libcomposite)[0x76e411c 4c]o
 f/irgofost/libs/lib_ iwplc1.2sxox(init_local_data+0x6 0w)l[c0oxr7e6f1085c]
/roo mac80211t/libs/li bc_figp8c0.2s1o1(rt_shm_open+ 0rxf7k0i)l[l0x76f10bfc]
 /tests _filnetxecgarnation_lib_ip wlcore_sdioc[0x12d68]
dc/il_ihbd/rlci_biamlchemy. suos.b0m(i+s0cx_ai4m4x8)[0x76e43 4c4i8_]h
]r/clib/libcoppe udc_corerplate.so. 0c(a+n0_xd8e3v14)[0x76e1a314

Ulib/libcobalt.so.2(+0x1058c)[0x76df558cC]P
 : 1 PID: 296 Comm: Reception task  Not tainted 4.1.46-ipipe #6
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
task: a885dd80 ti: a97fa000 task.ti: a97fa000
PC is at cobalt_signal_send+0xf8/0x1f8
LR is at 0x18
pc : [<800acedc>]    lr : [<00000018>]    psr: a0080093
sp : a97fbf00  ip : a94ff948  fp : 8095f680
r10: ab70e6e0  r9 : 00000124  r8 : 8088c880
r7 : 00000001  r6 : 00000000  r5 : 00800000  r4 : a94ff4d8
r3 : 00000017  r2 : 00000000  r1 : c08f9000  r0 : c09e5800
Flags: NzCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 10c53c7d  Table: 38ec804a  DAC: 00000015
Process Reception task  (pid: 296, stack limit = 0xa97fa220)
Stack: (0xa97fbf00 to 0xa97fc000)
bf00: 00000018 c08f9000 c09e5800 800ad534 c09e5800 00000018 00000001 
00000000
bf20: 8088c880 808945fc 00000018 800ad788 c09e9008 00000022 a97fbfb0 
800ad71c
bf40: 808945fc 00000001 ab70e6e0 800ae248 762bce38 a97fbfb0 808945fc 
20080013
bf60: 00000001 ab70d6e0 8095f680 8088b6e0 ab70e6e0 8008a14c 8095f680 
a97fbfb0
bf80: 8088b6e0 8088da20 04000004 762bce14 762bce38 00000000 000f0042 
8000f028
bfa0: a97fa000 00000002 762718cc 8000ef84 10000040 00000124 00000018 
00000124
bfc0: 762bce14 762bce38 00000000 000f0042 76f23594 0001e030 76271c74 
762718cc
bfe0: 00000018 76271870 000122ac 76df5148 20080010 10000040 8473a46f 
c307ae3a
[<800acedc>] (cobalt_signal_send) from [<800ad534>] 
(__cobalt_kill+0x190/0x1b0)
[<800ad534>] (__cobalt_kill) from [<800ad788>] (CoBaLt_kill+0x6c/0xcc)
[<800ad788>] (CoBaLt_kill) from [<800ae248>] (ipipe_syscall_hook+0xfc/0x270)
[<800ae248>] (ipipe_syscall_hook) from [<8008a14c>] 
(__ipipe_notify_syscall+0xa8/0x1a4)
[<8008a14c>] (__ipipe_notify_syscall) from [<8000ef84>] 
(pipeline_syscall+0x8/0x24)
Code: e5813084 e583c000 e8bd8070 e59421b0 (e592c004)
---[ end trace fcd211fc12a8e500 ]---
tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue: Assertion 
`current->magic == 0' failed.

[2]+  Aborted                    ./tests_integration_lib_ipc --session=rct 
cha
#

Best regards
Stephane

Re: [Xenomai] Assertion current-magic 0 failed

[Philippe Gerum]

On 06/28/2018 05:28 PM, Stéphane Reichert wrote:
> Hi,
> 
> I am testing the new version 3.0.7 on our system (iMX 6 on 4.1.46 kernel), 
> and I would like to know
> if anybody else encountered this error :
> 
> threadobj.c:1344: threadobj_prologue: Assertion `current->magic == 0' 
> failed.
> 
> I compile the kernel with --enable-pshared --enable 
> registry --enable-debug=full.
> 
> Here is a backtrace I get on the debug signal callback:
> 
> # Mode switch (reason: triggered fault), aborting. Backtrace:
> Retrieve backtrace...
> ./tests_integration_lib_ipc(Unable to handle kernel NULL pointer dereference 
> at virtual address 00000004
> 4x180)[0[x01020206000]08l0e0r0+
> ]]/ l*ipbd=39180831/libc.so.,6 (*+p0txe2=50105000)0[00, 
> *ppte=000000000x76ca3150
> 
> /lib/libcopperplate.so.0(syncobj_lIonctke+r0nxa8l4 error: Oops: 17 [#1] 
> PREEMPT SMP ARM
> l0x76Meo1dau7lee8s]
>  i/nlkiebd/in:libcopper pcltarte.so.0(syncluster_fin 
> dcocbmj+0x34)[0x76e18eec]
> /lib/li buaslbc_hfe_mayc.mso.0(al u_serialchemy_bind 
> _uosbbj_efc_te+c0mx98)[0x76e 3ge_1ccd0c]
> nlib/libalchemy.so. 0u(_rett_hheerap_bind+0x30 libcomposite)[0x76e411c 4c]o
>  f/irgofost/libs/lib_ iwplc1.2sxox(init_local_data+0x6 0w)l[c0oxr7e6f1085c]
> /roo mac80211t/libs/li bc_figp8c0.2s1o1(rt_shm_open+ 0rxf7k0i)l[l0x76f10bfc]
>  /tests _filnetxecgarnation_lib_ip wlcore_sdioc[0x12d68]
> dc/il_ihbd/rlci_biamlchemy. suos.b0m(i+s0cx_ai4m4x8)[0x76e43 4c4i8_]h
> ]r/clib/libcoppe udc_corerplate.so. 0c(a+n0_xd8e3v14)[0x76e1a314
> 

It looks like there are two backtraces, one generated from the app in
userland, a second one by the kernel, none of them is readable as they
intermingle.

-- 
Philippe.

Re: [Xenomai] Assertion current-magic 0 failed

[Stéphane Reichert]

Sorry about that.

Here is a better trace:
Mode switch (reason: triggered fault), aborting. Backtrace:
[Xenomai] switching Reception task shmem to secondary mode after exception 
#0 in kernel-space at 0x800aced8 (pid 327)
Unable to handle kernel paging request at virtual address ffeccffe
pgd = a96dc000
[ffeccffe] *pgd=3bf5e821, *pte=00000000, *ppte=00000000
Internal error: Oops: 17 [#2] PREEMPT SMP ARM
Modules linked in: ctr ccm usb_f_acm u_serial usb_f_ecm g_cdc u_ether 
libcomposite configfs wl12xx wlcore mac80211 cfg80211 rfkill flexcan 
wlcore_sdio ci_h                             drc_imx usbmisc_imx ci_hdrc 
udc_core can_dev
CPU: 2 PID: 327 Comm: Reception task  Tainted: G      D         4.1.46-ipipe 
#6
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
task: a8bfb740 ti: a9e0c000 task.ti: a9e0c000
/lib/libc.so.6(+0x25150) [0x7P6Cdis at cobalt_signal_send+0xf8/0x1f8
07150]
/lib/libcopperplate.so.0(syncobj_lock+0x84) [0x76e7e7e8]
/lib/libcopperplate.so.0(syncluster_findobj+0x34) [0x76e7ceec]
/lib/libalchemy.so.0(alchemy_bind_object+0x98) [0x76ea21c0]
/lib/libalchemy.so.0(rt_heap_bind+0x30) [0x76ea51c4]
/root/libs/lib_ipc.so(init_local_data+0x60) [0x76f7485c]
/root/libs/lib_ipc.so(rt_shm_open+0x70) [0x76f74bfc]
./tests_integration_lib_ipc() [0x12de0]
/lib/libalchemy.so.0(+0xa448) [0x76ea7448]
/lib/libcopperplate.so.0(+0x8314) [0x76e7e314]
/lib/libcobalt.so.2(+0x1058c) [0x76e5958c]
LR is at 0x18
pc : [<800acedc>]    lr : [<00000018>]    psr: 80070093
sp : a9e0df00  ip : a94fd948  fp : 8095f680
r10: ab71c6e0  r9 : 00000142  r8 : 8088c880
r7 : 00000001  r6 : 00000000  r5 : 00800000  r4 : a94fd4d8
r3 : 00000017  r2 : ffeccffa  r1 : c08f9088  r0 : c09e5800
Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 10c53c7d  Table: 396dc04a  DAC: 00000015
Process Reception task  (pid: 327, stack limit = 0xa9e0c220)
Stack: (0xa9e0df00 to 0xa9e0e000)
df00: 00000018 c08f9088 c09e5800 800ad534 c09e5800 00000018 00000001 
00000000
df20: 8088c880 808945fc 00000018 800ad788 c09e9008 00000022 a9e0dfb0 
800ad71c
df40: 808945fc 00000001 ab71c6e0 800ae248 76320e38 a9e0dfb0 808945fc 
20070013
df60: 00000001 ab71b6e0 8095f680 8088b6e0 ab71c6e0 8008a14c 8095f680 
a9e0dfb0
df80: 8088b6e0 8088da20 00000007 76320e14 76320e38 00000000 000f0042 
8000f028
dfa0: a9e0c000 00000002 761cc8cc 8000ef84 10000040 00000142 00000018 
00000142
dfc0: 76320e14 76320e38 00000000 000f0042 76f87594 0001e030 761ccc74 
761cc8cc
dfe0: 00000018 761cc868 00012318 76e59148 20070010 10000040 00000000 
00000000
[<800acedc>] (cobalt_signal_send) from [<800ad534>] 
(__cobalt_kill+0x190/0x1b0)
[<800ad534>] (__cobalt_kill) from [<800ad788>] (CoBaLt_kill+0x6c/0xcc)
[<800ad788>] (CoBaLt_kill) from [<800ae248>] (ipipe_syscall_hook+0xfc/0x270)
[<800ae248>] (ipipe_syscall_hook) from [<8008a14c>] 
(__ipipe_notify_syscall+0xa8/0x1a4)
[<8008a14c>] (__ipipe_notify_syscall) from [<8000ef84>] 
(pipeline_syscall+0x8/0x24)
Code: e5813084 e583c000 e8bd8070 e59421b0 (e592c004)
---[ end trace fcd211fc12a8e501 ]---
tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue: Assertion 
`current->magic == 0' failed.

Stéphane

-----Message d'origine-----
De : Philippe Gerum [mailto:rpm@xenomai.org]
Envoyé : jeudi 28 juin 2018 17:38
À : Stéphane Reichert; xenomai@xenomai.org
Objet : Re: [Xenomai] Assertion current-magic 0 failed

On 06/28/2018 05:28 PM, Stéphane Reichert wrote:
> Hi,
>
> I am testing the new version 3.0.7 on our system (iMX 6 on 4.1.46
> kernel), and I would like to know if anybody else encountered this
> error :
>
> threadobj.c:1344: threadobj_prologue: Assertion `current->magic == 0'
> failed.
>
> I compile the kernel with --enable-pshared --enable registry
> --enable-debug=full.
>
> Here is a backtrace I get on the debug signal callback:
>
> # Mode switch (reason: triggered fault), aborting. Backtrace:
> Retrieve backtrace...
> ./tests_integration_lib_ipc(Unable to handle kernel NULL pointer
> dereference at virtual address 00000004
> 4x180)[0[x01020206000]08l0e0r0+ ]]/ l*ipbd=39180831/libc.so.,6
> (*+p0txe2=50105000)0[00,
> *ppte=000000000x76ca3150
>
> /lib/libcopperplate.so.0(syncobj_lIonctke+r0nxa8l4 error: Oops: 17
> [#1] PREEMPT SMP ARM l0x76Meo1dau7lee8s]  i/nlkiebd/in:libcopper
> pcltarte.so.0(syncluster_fin
> dcocbmj+0x34)[0x76e18eec]
> /lib/li buaslbc_hfe_mayc.mso.0(al u_serialchemy_bind
> _uosbbj_efc_te+c0mx98)[0x76e 3ge_1ccd0c] nlib/libalchemy.so.
> 0u(_rett_hheerap_bind+0x30 libcomposite)[0x76e411c 4c]o
> f/irgofost/libs/lib_ iwplc1.2sxox(init_local_data+0x6
> 0w)l[c0oxr7e6f1085c] /roo mac80211t/libs/li
> bc_figp8c0.2s1o1(rt_shm_open+ 0rxf7k0i)l[l0x76f10bfc]  /tests
> _filnetxecgarnation_lib_ip wlcore_sdioc[0x12d68]
> dc/il_ihbd/rlci_biamlchemy. suos.b0m(i+s0cx_ai4m4x8)[0x76e43 4c4i8_]h
> ]r/clib/libcoppe udc_corerplate.so. 0c(a+n0_xd8e3v14)[0x76e1a314
>

It looks like there are two backtraces, one generated from the app in 
userland, a second one by the kernel, none of them is readable as they 
intermingle.

--
Philippe.

Re: [Xenomai] Assertion current-magic 0 failed

[Philippe Gerum]

On 06/29/2018 08:57 AM, Stéphane Reichert wrote:
> Sorry about that.
> 
> Here is a better trace:
> Mode switch (reason: triggered fault), aborting. Backtrace:
> [Xenomai] switching Reception task shmem to secondary mode after exception 
> #0 in kernel-space at 0x800aced8 (pid 327)
> Unable to handle kernel paging request at virtual address ffeccffe
> pgd = a96dc000
> [ffeccffe] *pgd=3bf5e821, *pte=00000000, *ppte=00000000
> Internal error: Oops: 17 [#2] PREEMPT SMP ARM
> Modules linked in: ctr ccm usb_f_acm u_serial usb_f_ecm g_cdc u_ether 
> libcomposite configfs wl12xx wlcore mac80211 cfg80211 rfkill flexcan 
> wlcore_sdio ci_h                             drc_imx usbmisc_imx ci_hdrc 
> udc_core can_dev
> CPU: 2 PID: 327 Comm: Reception task  Tainted: G      D         4.1.46-ipipe 
> #6
> Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
> task: a8bfb740 ti: a9e0c000 task.ti: a9e0c000
> /lib/libc.so.6(+0x25150) [0x7P6Cdis at cobalt_signal_send+0xf8/0x1f8
> 07150]
> /lib/libcopperplate.so.0(syncobj_lock+0x84) [0x76e7e7e8]
> /lib/libcopperplate.so.0(syncluster_findobj+0x34) [0x76e7ceec]
> /lib/libalchemy.so.0(alchemy_bind_object+0x98) [0x76ea21c0]
> /lib/libalchemy.so.0(rt_heap_bind+0x30) [0x76ea51c4]
> /root/libs/lib_ipc.so(init_local_data+0x60) [0x76f7485c]
> /root/libs/lib_ipc.so(rt_shm_open+0x70) [0x76f74bfc]
> ./tests_integration_lib_ipc() [0x12de0]
> /lib/libalchemy.so.0(+0xa448) [0x76ea7448]
> /lib/libcopperplate.so.0(+0x8314) [0x76e7e314]
> /lib/libcobalt.so.2(+0x1058c) [0x76e5958c]
> LR is at 0x18
> pc : [<800acedc>]    lr : [<00000018>]    psr: 80070093
> sp : a9e0df00  ip : a94fd948  fp : 8095f680
> r10: ab71c6e0  r9 : 00000142  r8 : 8088c880
> r7 : 00000001  r6 : 00000000  r5 : 00800000  r4 : a94fd4d8
> r3 : 00000017  r2 : ffeccffa  r1 : c08f9088  r0 : c09e5800
> Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
> Control: 10c53c7d  Table: 396dc04a  DAC: 00000015
> Process Reception task  (pid: 327, stack limit = 0xa9e0c220)
> Stack: (0xa9e0df00 to 0xa9e0e000)
> df00: 00000018 c08f9088 c09e5800 800ad534 c09e5800 00000018 00000001 
> 00000000
> df20: 8088c880 808945fc 00000018 800ad788 c09e9008 00000022 a9e0dfb0 
> 800ad71c
> df40: 808945fc 00000001 ab71c6e0 800ae248 76320e38 a9e0dfb0 808945fc 
> 20070013
> df60: 00000001 ab71b6e0 8095f680 8088b6e0 ab71c6e0 8008a14c 8095f680 
> a9e0dfb0
> df80: 8088b6e0 8088da20 00000007 76320e14 76320e38 00000000 000f0042 
> 8000f028
> dfa0: a9e0c000 00000002 761cc8cc 8000ef84 10000040 00000142 00000018 
> 00000142
> dfc0: 76320e14 76320e38 00000000 000f0042 76f87594 0001e030 761ccc74 
> 761cc8cc
> dfe0: 00000018 761cc868 00012318 76e59148 20070010 10000040 00000000 
> 00000000
> [<800acedc>] (cobalt_signal_send) from [<800ad534>] 
> (__cobalt_kill+0x190/0x1b0)
> [<800ad534>] (__cobalt_kill) from [<800ad788>] (CoBaLt_kill+0x6c/0xcc)
> [<800ad788>] (CoBaLt_kill) from [<800ae248>] (ipipe_syscall_hook+0xfc/0x270)
> [<800ae248>] (ipipe_syscall_hook) from [<8008a14c>] 
> (__ipipe_notify_syscall+0xa8/0x1a4)
> [<8008a14c>] (__ipipe_notify_syscall) from [<8000ef84>] 
> (pipeline_syscall+0x8/0x24)
> Code: e5813084 e583c000 e8bd8070 e59421b0 (e592c004)
> ---[ end trace fcd211fc12a8e501 ]---
> tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue: Assertion 
> `current->magic == 0' failed.
> 

I can't make any sense of the user backtrace, which does not seem to
match any possible code path which could be leading to the assertion
above. The kernel panic may require a separate investigation.

1. the assertion from libcopperplate. Please run your application over
GDB, then paste the backtrace reported by the debugger when the
assertion trips ("bt" command). You may want to use a non-console
terminal for running GDB, so that we can receive a clean trace unspoiled
by any kernel warning.

2. the kernel panic.

* which I-pipe patch have you been using?
* is there any specifics we should know from your kernel setup?
* is your application using kill(2), and if so, passing which signal number?
* building your kernel with CONFIG_DEBUG_INFO enabled and triggering the
bug again, please paste the output of the following command:

<your-toolchain>/bin/arm-linux-gnueabihf-addr2line -e
<path-to-kernel>/vmlinux $pc

$pc is the panic trace above would be 800acedc.

-- 
Philippe.

Re: [Xenomai] Assertion current-magic 0 failed

[Stéphane Reichert]

Hi Philippe:

Thanks for your answer. In fact, the assertion occurs when I call these two 
statements at the end of sigdebug_handler:
   signal (sig, SIG_DFL);
   kill (getpid(), sig);

You can find below the answers you asked for:

Backtrace reported by the debugger:
------------------------------------------------
Mode switch (reason: triggered fault), aborting. Backtrace:
/root/tests_integration_lib_ipc(sigdebug_handler+0x120)[0x12200]
/lib/libc.so.6(+0x25150)[0x76d56150]
/lib/libcopperplate.so.0(syncobj_lock+0x84)[0x76ecd7e8]
/lib/libcopperplate.so.0(syncluster_findobj+0x34)[0x76ecbeec]
/lib/libalchemy.so.0(alchemy_bind_object+0x98)[0x76ef11c0]
/lib/libalchemy.so.0(rt_heap_bind+0x30)[0x76ef41c4]
/root/libs/lib_ipc.so(init_local_data+0x60)[0x76fc385c]
/root/libs/lib_ipc.so(rt_shm_open+0x70)[0x76fc3bfc]
/root/tests_integration_lib_ipc[0x12d08]
/lib/libalchemy.so.0(+0xa448)[0x76ef6448]
/lib/libcopperplate.so.0(+0x8314)[0x76ecd314]
/lib/libcobalt.so.2(+0x1058c)[0x76ea858c]

tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue: Assertion 
`current->magic == 0' failed.

This log is displayed in the sigdebug_handler when I call rt_heap_bind 
(switch into secondary mode).

Kernel Panic:
-----------------
. We have adapted the ipipe-core-patch-4.1.18 -arm-10.patch to our kernel 
(4.1.46).
. The kernel is compiled with these options for xenomai : full-debug, 
registry on, relax traces on.
. The problem occurs when we call the kill at the end of the 
sigdebug_handler (same function as in the "Finding
spurious relaxes" example).
. The result of the addr2line gives me :
..../kernel/xenomai/posix/signal.c:75
This line code is (line 75 - function cobalt_signal_deliver):
    if (sigismember(swc->set, sig))

Best regards
Stéphane

-----Message d'origine-----
De : Philippe Gerum [mailto:rpm@xenomai.org]
Envoyé : vendredi 29 juin 2018 15:57
À : Stéphane Reichert; xenomai@xenomai.org
Objet : Re: [Xenomai] Assertion current-magic 0 failed

On 06/29/2018 08:57 AM, Stéphane Reichert wrote:
> Sorry about that.
>
> Here is a better trace:
> Mode switch (reason: triggered fault), aborting. Backtrace:
> [Xenomai] switching Reception task shmem to secondary mode after
> exception
> #0 in kernel-space at 0x800aced8 (pid 327) Unable to handle kernel
> paging request at virtual address ffeccffe pgd = a96dc000 [ffeccffe]
> *pgd=3bf5e821, *pte=00000000, *ppte=00000000 Internal error: Oops: 17
> [#2] PREEMPT SMP ARM Modules linked in: ctr ccm usb_f_acm u_serial
> usb_f_ecm g_cdc u_ether libcomposite configfs wl12xx wlcore mac80211
> cfg80211 rfkill flexcan
> wlcore_sdio ci_h                             drc_imx usbmisc_imx ci_hdrc
> udc_core can_dev
> CPU: 2 PID: 327 Comm: Reception task  Tainted: G      D 
> 4.1.46-ipipe
> #6
> Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
> task: a8bfb740 ti: a9e0c000 task.ti: a9e0c000
> /lib/libc.so.6(+0x25150) [0x7P6Cdis at cobalt_signal_send+0xf8/0x1f8
> 07150]
> /lib/libcopperplate.so.0(syncobj_lock+0x84) [0x76e7e7e8]
> /lib/libcopperplate.so.0(syncluster_findobj+0x34) [0x76e7ceec]
> /lib/libalchemy.so.0(alchemy_bind_object+0x98) [0x76ea21c0]
> /lib/libalchemy.so.0(rt_heap_bind+0x30) [0x76ea51c4]
> /root/libs/lib_ipc.so(init_local_data+0x60) [0x76f7485c]
> /root/libs/lib_ipc.so(rt_shm_open+0x70) [0x76f74bfc]
> ./tests_integration_lib_ipc() [0x12de0]
> /lib/libalchemy.so.0(+0xa448) [0x76ea7448]
> /lib/libcopperplate.so.0(+0x8314) [0x76e7e314]
> /lib/libcobalt.so.2(+0x1058c) [0x76e5958c] LR is at 0x18
> pc : [<800acedc>]    lr : [<00000018>]    psr: 80070093
> sp : a9e0df00  ip : a94fd948  fp : 8095f680
> r10: ab71c6e0  r9 : 00000142  r8 : 8088c880
> r7 : 00000001  r6 : 00000000  r5 : 00800000  r4 : a94fd4d8
> r3 : 00000017  r2 : ffeccffa  r1 : c08f9088  r0 : c09e5800
> Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
> Control: 10c53c7d  Table: 396dc04a  DAC: 00000015 Process Reception
> task  (pid: 327, stack limit = 0xa9e0c220)
> Stack: (0xa9e0df00 to 0xa9e0e000)
> df00: 00000018 c08f9088 c09e5800 800ad534 c09e5800 00000018 00000001
> 00000000
> df20: 8088c880 808945fc 00000018 800ad788 c09e9008 00000022 a9e0dfb0
> 800ad71c
> df40: 808945fc 00000001 ab71c6e0 800ae248 76320e38 a9e0dfb0 808945fc
> 20070013
> df60: 00000001 ab71b6e0 8095f680 8088b6e0 ab71c6e0 8008a14c 8095f680
> a9e0dfb0
> df80: 8088b6e0 8088da20 00000007 76320e14 76320e38 00000000 000f0042
> 8000f028
> dfa0: a9e0c000 00000002 761cc8cc 8000ef84 10000040 00000142 00000018
> 00000142
> dfc0: 76320e14 76320e38 00000000 000f0042 76f87594 0001e030 761ccc74
> 761cc8cc
> dfe0: 00000018 761cc868 00012318 76e59148 20070010 10000040 00000000
> 00000000
> [<800acedc>] (cobalt_signal_send) from [<800ad534>]
> (__cobalt_kill+0x190/0x1b0)
> [<800ad534>] (__cobalt_kill) from [<800ad788>] (CoBaLt_kill+0x6c/0xcc)
> [<800ad788>] (CoBaLt_kill) from [<800ae248>]
> (ipipe_syscall_hook+0xfc/0x270) [<800ae248>] (ipipe_syscall_hook) from
> [<8008a14c>]
> (__ipipe_notify_syscall+0xa8/0x1a4)
> [<8008a14c>] (__ipipe_notify_syscall) from [<8000ef84>]
> (pipeline_syscall+0x8/0x24)
> Code: e5813084 e583c000 e8bd8070 e59421b0 (e592c004) ---[ end trace
> fcd211fc12a8e501 ]---
> tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue:
> Assertion `current->magic == 0' failed.
>

I can't make any sense of the user backtrace, which does not seem to match 
any possible code path which could be leading to the assertion above. The 
kernel panic may require a separate investigation.

1. the assertion from libcopperplate. Please run your application over GDB, 
then paste the backtrace reported by the debugger when the assertion trips 
("bt" command). You may want to use a non-console terminal for running GDB, 
so that we can receive a clean trace unspoiled by any kernel warning.

2. the kernel panic.

* which I-pipe patch have you been using?
* is there any specifics we should know from your kernel setup?
* is your application using kill(2), and if so, passing which signal number?
* building your kernel with CONFIG_DEBUG_INFO enabled and triggering the bug 
again, please paste the output of the following command:

<your-toolchain>/bin/arm-linux-gnueabihf-addr2line -e 
<path-to-kernel>/vmlinux $pc

$pc is the panic trace above would be 800acedc.

--
Philippe.

Re: [Xenomai] Assertion current-magic 0 failed

[Philippe Gerum]

On 07/04/2018 09:54 AM, Stéphane Reichert wrote:
> Hi Philippe:
> 
> Thanks for your answer. In fact, the assertion occurs when I call these two 
> statements at the end of sigdebug_handler:
>    signal (sig, SIG_DFL);
>    kill (getpid(), sig);
> 
> You can find below the answers you asked for:
> 
> Backtrace reported by the debugger:
> ------------------------------------------------
> Mode switch (reason: triggered fault), aborting. Backtrace:

The mode switch is due to a memory access fault, and looking at the
implementation of rt_heap_bind(), the only reason for this to happen in
such situation would be that alchemy_init() did not run prior to calling
rt_heap_bind().

In other words, I suspect that auto-init was disabled via xeno-config,
but xenomai_init() was not yet called when rt_heap_bind() is issued.
Would that be a plausible scenario?

> /root/tests_integration_lib_ipc(sigdebug_handler+0x120)[0x12200]
> /lib/libc.so.6(+0x25150)[0x76d56150]
> /lib/libcopperplate.so.0(syncobj_lock+0x84)[0x76ecd7e8]
> /lib/libcopperplate.so.0(syncluster_findobj+0x34)[0x76ecbeec]
> /lib/libalchemy.so.0(alchemy_bind_object+0x98)[0x76ef11c0]
> /lib/libalchemy.so.0(rt_heap_bind+0x30)[0x76ef41c4]
> /root/libs/lib_ipc.so(init_local_data+0x60)[0x76fc385c]
> /root/libs/lib_ipc.so(rt_shm_open+0x70)[0x76fc3bfc]
> /root/tests_integration_lib_ipc[0x12d08]
> /lib/libalchemy.so.0(+0xa448)[0x76ef6448]
> /lib/libcopperplate.so.0(+0x8314)[0x76ecd314]
> /lib/libcobalt.so.2(+0x1058c)[0x76ea858c]
> 
> tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue: Assertion 
> `current->magic == 0' failed.
> 
> This log is displayed in the sigdebug_handler when I call rt_heap_bind 
> (switch into secondary mode).
> 
> Kernel Panic:
> -----------------
> . We have adapted the ipipe-core-patch-4.1.18 -arm-10.patch to our kernel 
> (4.1.46).
> . The kernel is compiled with these options for xenomai : full-debug, 
> registry on, relax traces on.
> . The problem occurs when we call the kill at the end of the 
> sigdebug_handler (same function as in the "Finding
> spurious relaxes" example).
> . The result of the addr2line gives me :
> ..../kernel/xenomai/posix/signal.c:75
> This line code is (line 75 - function cobalt_signal_deliver):
>     if (sigismember(swc->set, sig))
> 

That kernel panic has to be fixed, however I have not been able to
reproduce it yet. If you could either send me a basic test code
triggering the issue, or the executable causing the fault (no source
needed), that would help.

-- 
Philippe.

Re: [Xenomai] Assertion current-magic 0 failed

[Stéphane Reichert]

Hi Philippe,

Unfortunately, we did not disable auto-init.

Maybe the fact that rt_heap_bind is called from a shared library could be 
the problem ?

I joined the 2 executables we are using. In order to reproduce the problem, 
you have to run these commands:
./start_lib_ipc --session=rct &
./tests_integration_lib_ipc --session=rct CH1 &

The shared library lib_ipc.so contains the call to rt_heap_bind.

The compiler options for the shared library are:
...../bin/arm-linux-gnueabihf-gcc  -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -D_LARGEFILE64_SOURCE 
 -D_LARGEFILE_SOURCE -D_REENTRANT -D__COBALT_WRAP__ -D__COBALT__ -D__XENO__  
-Dlib_ipc_EXPORTS -fmessage-length=0 -g -fPIC -Wall -Werror -fasynchronous-unwind-tables 
....
The linker options for the shared library are:
...... 
/bin/arm-linux-gnueabihf-gcc  -fPIC  -fmessage-length=0 -g   -shared -Wl,-soname,lib_ipc.so 
 -o lib_ipc.so CMakeFiles/lib_ipc.dir/SRC/messages.c.o 
CMakeFiles/lib_ipc.dir/SRC/shared_memory.c.o 
CMakeFiles/lib_ipc.dir/SRC/fifoplus.c.o 
MakeFiles/lib_ipc.dir/SRC/init.c.o  -L/home2/sreichert/workspace/lib_ipc/.   
-L/home2/sreichert/workspace/lib_ipc/../robot/lib  -L/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomai 
 -lm -lalchemy -lcopperplate -lcobalt -lmodechk -lpthread -lrt 
/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomai/bootstrap-pic.o 
 -Wl,-rpath,/home2/sreichert/workspace/lib_ipc/.:/home2/sreichert/workspace/lib_ipc/../robot/lib:/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomaiThe compiler options for the executables are:...../bin/arm-linux-gnueabihf-gcc  -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -D_REENTRANT -D__COBALT_WRAP__ -D__COBALT__ -D__XENO__ -g -Wall -Werror -fasynchronous-unwind-tables ....The linker options for the executables are:...../bin/arm-linux-gnueabihf-gcc  -gCMakeFiles/tests_integration_lib_ipc.dir/SRC/test_lib_ipc.c.oCMakeFiles/tests_integration_lib_ipc.dir/SRC/sighandler.c.oCMakeFiles/tests_integration_lib_ipc.dir/SRC/test_fifo.c.oCMakeFiles/tests_integration_lib_ipc.dir/SRC/test_shared_memory.c.o  -otests_integration_lib_ipc  -L/home2/sreichert/workspace/tests_integration_lib_ipc/.  -L/home2/sreichert/workspace/tests_integration_lib_ipc/../robot/lib  -L/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomai -rdynamic -l_ipc -lm -lfuse -lalchemy -lcopperplate -lcobalt -lmodechk -lpthread -lrt -Wl,--wrap=main,@/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/cobalt.wrappers/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomai/bootstrap.o -Wl,-rpath,/home2/sreichert/workspace/tests_integration_lib_ipc/.:/home2/sreichert/workspace/tests_integration_lib_ipc/../robot/lib:/opt/toolchain/arm-linux-gnueabihf/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/lib/xenomaiI did another try and this time the pc was stuck oninclude/linux/signal.h:79This line is :   return 1 & (set->sig[sig / _NSIG_BPW] >> (sig % _NSIG_BPW));I hope this could help.Best regardsStéphane-----Message d'origine-----De : Philippe Gerum [mailto:rpm@xenomai.org]Envoyé : mercredi 4 juillet 2018 11:30À : Stéphane Reichert; xenomai@xenomai.orgObjet : Re: [Xenomai] Assertion current-magic 0 failedOn 07/04/2018 09:54 AM, Stéphane Reichert wrote:> Hi Philippe:>> Thanks for your answer. In fact, the assertion occurs when I call> these two statements at the end of sigdebug_handler:>    signal (sig, SIG_DFL);>    kill (getpid(), sig);>> You can find below the answers you asked for:>> Backtrace reported by the debugger:> ------------------------------------------------> Mode switch (reason: triggered fault), aborting. Backtrace:The mode switch is due to a memory access fault, and looking at theimplementation of rt_heap_bind(), the only reason for this to happen in suchsituation would be that alchemy_init() did not run prior to callingrt_heap_bind().In other words, I suspect that auto-init was disabled via xeno-config, butxenomai_init() was not yet called when rt_heap_bind() is issued.Would that be a plausible scenario?> /root/tests_integration_lib_ipc(sigdebug_handler+0x120)[0x12200]> /lib/libc.so.6(+0x25150)[0x76d56150]> /lib/libcopperplate.so.0(syncobj_lock+0x84)[0x76ecd7e8]> /lib/libcopperplate.so.0(syncluster_findobj+0x34)[0x76ecbeec]> /lib/libalchemy.so.0(alchemy_bind_object+0x98)[0x76ef11c0]> /lib/libalchemy.so.0(rt_heap_bind+0x30)[0x76ef41c4]> /root/libs/lib_ipc.so(init_local_data+0x60)[0x76fc385c]> /root/libs/lib_ipc.so(rt_shm_open+0x70)[0x76fc3bfc]> /root/tests_integration_lib_ipc[0x12d08]> /lib/libalchemy.so.0(+0xa448)[0x76ef6448]> /lib/libcopperplate.so.0(+0x8314)[0x76ecd314]> /lib/libcobalt.so.2(+0x1058c)[0x76ea858c]>> tests_integration_lib_ipc: threadobj.c:1344: threadobj_prologue:> Assertion `current->magic == 0' failed.>> This log is displayed in the sigdebug_handler when I call rt_heap_bind> (switch into secondary mode).>> Kernel Panic:> -----------------> . We have adapted the ipipe-core-patch-4.1.18 -arm-10.patch to our> kernel (4.1.46).> . The kernel is compiled with these options for xenomai : full-debug,> registry on, relax traces on.> . The problem occurs when we call the kill at the end of the> sigdebug_handler (same function as in the "Finding spurious relaxes"> example).> . The result of the addr2line gives me :> ..../kernel/xenomai/posix/signal.c:75> This line code is (line 75 - function cobalt_signal_deliver):>     if (sigismember(swc->set, sig))>That kernel panic has to be fixed, however I have not been able to reproduceit yet. If you could either send me a basic test code triggering the issue,or the executable causing the fault (no source needed), that would help.--Philippe.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: start_lib_ipc
Type: application/octet-stream
Size: 14712 bytes
Desc: not available
URL: <http://xenomai.org/pipermail/xenomai/attachments/20180704/bf6c01f7/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tests_integration_lib_ipc
Type: application/octet-stream
Size: 50588 bytes
Desc: not available
URL: <http://xenomai.org/pipermail/xenomai/attachments/20180704/bf6c01f7/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lib_ipc.so
Type: application/octet-stream
Size: 62708 bytes
Desc: not available
URL: <http://xenomai.org/pipermail/xenomai/attachments/20180704/bf6c01f7/attachment.so>

Re: [Xenomai] Assertion current-magic 0 failed

[Philippe Gerum]

On 07/04/2018 03:26 PM, Stéphane Reichert wrote:
> Hi Philippe,
> 
> Unfortunately, we did not disable auto-init.
> 
> Maybe the fact that rt_heap_bind is called from a shared library could be 
> the problem ?

Does that code run over a constructor routine (i.e. tagged with
__attribute__((constructor))?

> 
> I joined the 2 executables we are using. In order to reproduce the problem, 
> you have to run these commands:
> ./start_lib_ipc --session=rct &
> ./tests_integration_lib_ipc --session=rct CH1 &
> 
> The shared library lib_ipc.so contains the call to rt_heap_bind.
> 

Ok, I'll have a look. Thanks.

-- 
Philippe.