search

grodansparadis / vscp

VSCP (Very Simple Control Protocol) IoT/m2m framework
http://www.vscp.org
Other
52 stars 20 forks source link

Segmentation faults in vscphelper #86

Closed BlueAndi closed 9 years ago

BlueAndi commented 9 years ago

From time to time I get segmentation faults in the vscphelper lib (last official version) on linux and windows.

gdb output on linux: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb6affb40 (LWP 3157)] ns_out (nc=0x0, buf=0xb4826ff0, len=156130) at ../../../../../common/net_skeleton.c:87 87 if (nc->flags & NSF_UDP) {

grodansparadis commented 9 years ago

Something strange there, net_skeleton.c is not used in vscp_helper anymore. Or should at least not be.

Andreas Merkle skrev den 2015-02-17 20:17:

From time to time I get segmentation faults in the vscphelper lib (last official version) on linux and windows.

gdb output on linux: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb6affb40 (LWP 3157)] ns_out (nc=0x0, buf=0xb4826ff0, len=156130) at ../../../../../common/net_skeleton.c:87 87 if (nc->flags & NSF_UDP) {

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

grodansparadis commented 9 years ago

Try ./configure to generate correct makefile and build

Andreas Merkle skrev den 2015-02-17 20:17:

From time to time I get segmentation faults in the vscphelper lib (last official version) on linux and windows.

gdb output on linux: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb6affb40 (LWP 3157)] ns_out (nc=0x0, buf=0xb4826ff0, len=156130) at ../../../../../common/net_skeleton.c:87 87 if (nc->flags & NSF_UDP) {

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

BlueAndi commented 9 years ago

I don't use the git head, I use the official released version ... so the problem is gone by removing net_skeleton ... I will build a new one for windows and give it a try there ... after I will do it for linux.

grodansparadis commented 9 years ago

The release version have a couple of errors so I recommend going to the head.

btw my fix upload is on the way. Apparently some one else here is using all bandwidth at the moment.

Andreas Merkle skrev den 2015-02-17 20:30:

I don't use the git head, I use the official released version ... so the problem is gone by removing net_skeleton ... I will build a new one for windows and give it a try there ... after I will do it for linux.

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86#issuecomment-74735290.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

BlueAndi commented 9 years ago

Ok, so lets close the issue and I will try later with a new lib.

BlueAndi commented 9 years ago

I tested yesterday evening again with the current head and the segmentation fault was still there, but less often.

Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb57ffb40 (LWP 9833)] __memmove_ia32 () at ../sysdeps/i386/i686/multiarch/../memmove.S:103 103 ../sysdeps/i386/i686/multiarch/../memmove.S: Datei oder Verzeichnis nicht gefunden. (gdb) where

0 __memmove_ia32 () at ../sysdeps/i386/i686/multiarch/../memmove.S:103

1 0xb7f86c59 in iobuf_insert () from /home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so

2 0xb7f86a9f in iobuf_append () from /home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so

3 0xb7f86f1d in ns_out () from /home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so

4 0xb7f87d9b in ns_send () from /home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so

5 0xb7f6cfab in VscpRemoteTcpIf::doCmdSendEx (this=0x8084770, pEvent=pEvent@entry=0xb57feed4)

at ../../common/vscpremotetcpif.cpp:711

6 0xb7f4d710 in vscphlp_sendEventEx (handle=handle@entry=1681, pEvent=pEvent@entry=0xb57feed4) at ../vscphelperlib.cpp:346

7 0x0804bcfe in vscp_tp_adapter_writeMessage (msg=msg@entry=0xb57ff13a) at vscp_user/vscp_tp_adapter.c:286

8 0x0804f5a7 in vscp_transport_writeMessage (msg=msg@entry=0xb57ff13a) at ../../vscp/vscp_transport.c:205

9 0x0804ecb7 in vscp_core_stateActive () at ../../vscp/vscp_core.c:1048

10 vscp_core_process () at ../../vscp/vscp_core.c:467

11 0x0804ad95 in vscp_thread_frameworkThread (par=0x80553e0 ) at vscp_thread.c:269

12 0xb7fadf70 in start_thread (arg=0xb57ffb40) at pthread_create.c:312

13 0xb7e4b50e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:129

(gdb) list 98 in ../sysdeps/i386/i686/multiarch/../memmove.S (gdb) frame 5

5 0xb7f6cfab in VscpRemoteTcpIf::doCmdSendEx (this=0x8084770, pEvent=pEvent@entry=0xb57feed4)

at ../../common/vscpremotetcpif.cpp:711

711 strBuf.length() ); (gdb) print strBuf $1 = {static npos = 4294967295, m_impl = {static npos = 4294967295, _M_dataplus = {std::allocator = {<__gnu_cxx::new_allocator> = {}, }, _M_p = 0xb3a00014 L"SEND 0,0,0,0,350224384,00:00:00:00:00:00:00:00:00:00:00:00:01:03:00:00,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0"...}}, m_convertedToChar = { m_str = 0xb464df88 "SEND 0,0,0,0,350224384,00:00:00:00:00:00:00:00:00:00:00:00:01:03:00:00,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0"..., m_len = 153789}}

BlueAndi commented 9 years ago

The event in the strBuf is bad. Its destroyed during the way through the helper lib.

Note, that this problem seems to exists only on linux, but not on windows. Or in other words, I can't reproduce it on windows.

BlueAndi commented 9 years ago

Hmm ... after running over an hour I got now a out of bounce error on windows:

warning: c:\wxwidgets-3.0.2\include\wx\arrstr.h(178): assert "nIndex < m_nCount" failed in wxArrayString::Item(): wxArra yString: index out of bounds [in thread 1460]

But this seems to be a different bug.

Sorry, no callstack this time.

grodansparadis commented 9 years ago

Strange. I haven't seen this a single time since the fix. Do you have a way to repeat it?

Andreas Merkle skrev den 2015-03-01 11:22:

I tested yesterday evening again with the current head and the segmentation fault was still there, but less often.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb57ffb40 (LWP 9833)]
__memmove_ia32 () at ../sysdeps/i386/i686/multiarch/../memmove.S:103
103 ../sysdeps/i386/i686/multiarch/../memmove.S: Datei oder
Verzeichnis nicht gefunden.
(gdb) where
#0 __memmove_ia32 () at
../sysdeps/i386/i686/multiarch/../memmove.S:103
#1 <https://github.com/grodansparadis/vscp_software/issues/1>
0xb7f86c59 in iobuf_insert () from
/home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so
#2 <https://github.com/grodansparadis/vscp_software/issues/2>
0xb7f86a9f in iobuf_append () from
/home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so
#3 <https://github.com/grodansparadis/vscp_software/issues/3>
0xb7f86f1d in ns_out () from
/home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so
#4 <https://github.com/grodansparadis/vscp_software/issues/4>
0xb7f87d9b in ns_send () from
/home/amerkle/Projekte/github/vscp-framework/examples/pc/bin/../lib/linux/libvscphelper.so
#5 <https://github.com/grodansparadis/vscp_software/issues/5>
0xb7f6cfab in VscpRemoteTcpIf::doCmdSendEx (this=0x8084770,
pEvent=pEvent@entry=0xb57feed4)
at ../../common/vscpremotetcpif.cpp:711
#6 <https://github.com/grodansparadis/vscp_software/issues/6>
0xb7f4d710 in vscphlp_sendEventEx (handle=handle@entry=1681,
pEvent=pEvent@entry=0xb57feed4) at ../vscphelperlib.cpp:346
#7 <https://github.com/grodansparadis/vscp_software/issues/7>
0x0804bcfe in vscp_tp_adapter_writeMessage
(msg=msg@entry=0xb57ff13a) at vscp_user/vscp_tp_adapter.c:286
#8 <https://github.com/grodansparadis/vscp_software/issues/8>
0x0804f5a7 in vscp_transport_writeMessage
(msg=msg@entry=0xb57ff13a) at ../../vscp/vscp_transport.c:205
#9 <https://github.com/grodansparadis/vscp_software/issues/9>
0x0804ecb7 in vscp_core_stateActive () at ../../vscp/vscp_core.c:1048
#10 <https://github.com/grodansparadis/vscp_software/issues/10>
vscp_core_process () at ../../vscp/vscp_core.c:467
#11 <https://github.com/grodansparadis/vscp_software/issues/11>
0x0804ad95 in vscp_thread_frameworkThread (par=0x80553e0 ) at
vscp_thread.c:269
#12 <https://github.com/grodansparadis/vscp_software/issues/12>
0xb7fadf70 in start_thread (arg=0xb57ffb40) at pthread_create.c:312
#13 <https://github.com/grodansparadis/vscp_software/issues/13>
0xb7e4b50e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:129
(gdb) list
98 in ../sysdeps/i386/i686/multiarch/../memmove.S
(gdb) frame 5
#5 <https://github.com/grodansparadis/vscp_software/issues/5>
0xb7f6cfab in VscpRemoteTcpIf::doCmdSendEx (this=0x8084770,
pEvent=pEvent@entry=0xb57feed4)
at ../../common/vscpremotetcpif.cpp:711
711 strBuf.length() );
(gdb) print strBuf
$1 = {static npos = 4294967295, m_impl = {static npos = 4294967295,
_M_dataplus = {std::allocator<wchar_t> =
{<__gnu_cxx::new_allocator> = {}, },
_M_p = 0xb3a00014 L"SEND
0,0,0,0,350224384,00:00:00:00:00:00:00:00:00:00:00:00:01:03:00:00,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0"...}},
m_convertedToChar = {
m_str = 0xb464df88 "SEND
0,0,0,0,350224384,00:00:00:00:00:00:00:00:00:00:00:00:01:03:00:00,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0"...,
m_len = 153789}}

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86#issuecomment-76590910.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

BlueAndi commented 9 years ago

Not really a way ... only running the pc node long enough. I try to hunt this the whole day now and since I use a gdb frontend (a hour ago) it never happen again.

The event is always bad at dll level, one step back to the application where vscphlp_sendEventEx() is called, the event looks perfect. Absolute no glue at the moment how this can happen.

grodansparadis commented 9 years ago

Hi,

You are on Linux or Windows? I don't think this is related with the last problem. I will try to set up a test for it if you give me your environment but can't really spend much time on it until in a couple of weeks.

/Ake

Andreas Merkle skrev den 2015-03-01 16:50:

Not really a way ... only running the pc node long enough. I try to hunt this the whole day now and since I use a gdb frontend (a hour ago) it never happen again.

The event is always bad at dll level, one step back to the application where vscphlp_sendEventEx() is called, the event looks perfect. Absolute no glue at the moment how this can happen.

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86#issuecomment-76603114.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

grodansparadis commented 9 years ago

In libvscphelper, vespworks, the daemon?

Andreas Merkle skrev den 2015-03-01 14:50:

Hmm ... after running over an hour I got now a out of bounce error on windows:

warning: c:\wxwidgets-3.0.2\include\wx\arrstr.h(178): assert "nIndex < m_nCount" failed in wxArrayString::Item(): wxArra yString: index out of bounds [in thread 1460]

But this seems to be a different bug.

Sorry, no callstack this time.

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86#issuecomment-76597436.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org

Detta e-postmeddelande har sökts igenom efter virus med antivirusprogram från Avast. http://www.avast.com

BlueAndi commented 9 years ago

I am at both sides, but with different bugs. Both bugs happen in the libvscphelper. Don't spend too much time for it. Maybe we keep it open and test it just a minute before the release again. I will write here any further info I get during debugging.

grodansparadis commented 9 years ago

Still a problem?

BlueAndi commented 9 years ago

Will check it ...

BlueAndi commented 9 years ago

Can't reproduce it right now, except #128 ... lets close it, until it is back or hopefully never will come again.

grodansparadis commented 9 years ago

Thumbs held...

Den 2015-09-02 kl. 21:46, skrev Andreas Merkle:

Can't reproduce it right now, except #128 https://github.com/grodansparadis/vscp_software/issues/128 ... lets close it, until it is back or hopefully never will come again.

— Reply to this email directly or view it on GitHub https://github.com/grodansparadis/vscp_software/issues/86#issuecomment-137224437.

Be Hungry - Stay Foolish! Åke Hedman, Grodans Paradis AB / Paradise of the Frog Brattbergavägen 17, 82050 LOS, SWEDEN Phone: +46 (0)840011835 CellPhone: +46 (0)730-533146 http://www.grodansparadis.com http://www.vscp.org