Re: [Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-24 Thread Curtis Dutton
It has not happened to me again. It was a one time thing so far.

When I'm working with the machine, I copy down .ngc files to a folder on
the cnc controller machine. Then run through them. Often I delete these
files at the end of working with the cnc. I may have left the file open in
Axis, deleted them from the computer, ran my toolchange program, and when
it finished... bang! On the older 2.5 version this would just freeze axis
occasionally. I'm wondering if the beta branch has different code that
faults instead of freezes axis.


On Wed, Jan 22, 2014 at 9:28 PM, Jon Elson el...@pico-systems.com wrote:

 On 01/22/2014 02:04 PM, Sebastian Kuzminsky wrote:
 Oh, signal 11!  I know that one.  It can be caused by bad
 memory or
 cache.  The OP might want to run the memtest86 that at least
 used
 to be included on the live CD.

 Jon


 --
 CenturyLink Cloud: The Leader in Enterprise Cloud Services.
 Learn Why More Businesses Are Choosing CenturyLink Cloud For
 Critical Workloads, Development Environments  Everything In Between.
 Get a Quote or Start a Free Trial Today.

 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
 ___
 Emc-developers mailing list
 Emc-developers@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/emc-developers

--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


Re: [Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-24 Thread Robert Ellenberg
Hi Curtis,

Thanks for taking the plunge with the new TP!I think I've seen something
like this too. Typically it's when I move an ngc file, then try to open it
from the 'recent files' list. Predictably, the file isn't there and it
throws an error, but then segfaults in milltask.

Has anyone had experience using GDB to debug the task portion of
LinuxCNC? I've run rtapi_app through GDB in simulation to catch segfaults
in motion, but I'm not sure how to set this up for the task stuff or GUI's.

-Rob


On Fri, Jan 24, 2014 at 10:00 AM, Curtis Dutton curtd...@gmail.com wrote:

 It has not happened to me again. It was a one time thing so far.

 When I'm working with the machine, I copy down .ngc files to a folder on
 the cnc controller machine. Then run through them. Often I delete these
 files at the end of working with the cnc. I may have left the file open in
 Axis, deleted them from the computer, ran my toolchange program, and when
 it finished... bang! On the older 2.5 version this would just freeze axis
 occasionally. I'm wondering if the beta branch has different code that
 faults instead of freezes axis.


 On Wed, Jan 22, 2014 at 9:28 PM, Jon Elson el...@pico-systems.com wrote:

  On 01/22/2014 02:04 PM, Sebastian Kuzminsky wrote:
  Oh, signal 11!  I know that one.  It can be caused by bad
  memory or
  cache.  The OP might want to run the memtest86 that at least
  used
  to be included on the live CD.
 
  Jon
 
 
 
 --
  CenturyLink Cloud: The Leader in Enterprise Cloud Services.
  Learn Why More Businesses Are Choosing CenturyLink Cloud For
  Critical Workloads, Development Environments  Everything In Between.
  Get a Quote or Start a Free Trial Today.
 
 
 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
  ___
  Emc-developers mailing list
  Emc-developers@lists.sourceforge.net
  https://lists.sourceforge.net/lists/listinfo/emc-developers
 

 --
 CenturyLink Cloud: The Leader in Enterprise Cloud Services.
 Learn Why More Businesses Are Choosing CenturyLink Cloud For
 Critical Workloads, Development Environments  Everything In Between.
 Get a Quote or Start a Free Trial Today.

 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
 ___
 Emc-developers mailing list
 Emc-developers@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/emc-developers

--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


Re: [Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-24 Thread Sebastian Kuzminsky
On 1/24/14 11:43 , Robert Ellenberg wrote:
 Thanks for taking the plunge with the new TP!I think I've seen something
 like this too. Typically it's when I move an ngc file, then try to open it
 from the 'recent files' list. Predictably, the file isn't there and it
 throws an error, but then segfaults in milltask.

 Has anyone had experience using GDB to debug the task portion of
 LinuxCNC? I've run rtapi_app through GDB in simulation to catch segfaults
 in motion, but I'm not sure how to set this up for the task stuff or GUI's.

If you guys can come up with a simple way to reproduce the problem, 
preferably using a sim config that anyone can run without special 
hardware, that would be very valuable to whoever ends up debugging this. 
  Attach the steps to reproduce when you make the bug report at 
http://sf.net/p/emc/bugs


-- 
Sebastian Kuzminsky

--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


Re: [Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-24 Thread Kenneth Lerman
On 1/24/2014 1:43 PM, Robert Ellenberg wrote:
 Hi Curtis,

 Thanks for taking the plunge with the new TP!I think I've seen something
 like this too. Typically it's when I move an ngc file, then try to open it
 from the 'recent files' list. Predictably, the file isn't there and it
 throws an error, but then segfaults in milltask.

 Has anyone had experience using GDB to debug the task portion of
 LinuxCNC? I've run rtapi_app through GDB in simulation to catch segfaults
 in motion, but I'm not sure how to set this up for the task stuff or GUI's.

 -Rob
I've found that a quick way to attack problems of that nature are to:
1 -- ulimit -c unlimited  // enable core dumps -- I may have the command 
wrong.
2 -- run the application
3 -- find the core dump -- I think it will look like core.n where 
n is the process id of the process that crashed. It is sometimes 
hard to figure out which directory to look in for the core files. find 
is your friend.
4 -- then look at the core file using gdb -- the where command will give 
a traceback. Of course, if you didn't compile with -g (debug symbols), 
the traceback will be hard to read.

[Details to be provided by the user]

Ken



 On Fri, Jan 24, 2014 at 10:00 AM, Curtis Dutton curtd...@gmail.com wrote:

 It has not happened to me again. It was a one time thing so far.

 When I'm working with the machine, I copy down .ngc files to a folder on
 the cnc controller machine. Then run through them. Often I delete these
 files at the end of working with the cnc. I may have left the file open in
 Axis, deleted them from the computer, ran my toolchange program, and when
 it finished... bang! On the older 2.5 version this would just freeze axis
 occasionally. I'm wondering if the beta branch has different code that
 faults instead of freezes axis.


 On Wed, Jan 22, 2014 at 9:28 PM, Jon Elson el...@pico-systems.com wrote:

 On 01/22/2014 02:04 PM, Sebastian Kuzminsky wrote:
 Oh, signal 11!  I know that one.  It can be caused by bad
 memory or
 cache.  The OP might want to run the memtest86 that at least
 used
 to be included on the live CD.

 Jon



 --
 CenturyLink Cloud: The Leader in Enterprise Cloud Services.
 Learn Why More Businesses Are Choosing CenturyLink Cloud For
 Critical Workloads, Development Environments  Everything In Between.
 Get a Quote or Start a Free Trial Today.


 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
 ___
 Emc-developers mailing list
 Emc-developers@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/emc-developers

 --
 CenturyLink Cloud: The Leader in Enterprise Cloud Services.
 Learn Why More Businesses Are Choosing CenturyLink Cloud For
 Critical Workloads, Development Environments  Everything In Between.
 Get a Quote or Start a Free Trial Today.

 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
 ___
 Emc-developers mailing list
 Emc-developers@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/emc-developers

 --
 CenturyLink Cloud: The Leader in Enterprise Cloud Services.
 Learn Why More Businesses Are Choosing CenturyLink Cloud For
 Critical Workloads, Development Environments  Everything In Between.
 Get a Quote or Start a Free Trial Today.
 http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
 ___
 Emc-developers mailing list
 Emc-developers@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/emc-developers

--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


[Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-22 Thread Curtis Dutton
Does anyone know what may have caused this?

So I'm running the circular arc blend beta from Roberts github repository.
I'm running a build based on commit
8acb5e68a4e9c03b191555f912de53e639576dec


I'm not sure if this is related to circular arc blend or just latest
linuxcnc master code.

Attached is the backtrace. I was running a call (in the command window) o
tool_change call which is a simple program to touch a tool probe after a
manual change. It had just finished that subroutine when milltask died.

Attached is the backtrace.


Thanks,
Curtis


backtrace.6660
Description: Binary data
--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


Re: [Emc-developers] circular arc blend beta: milltask (pid 6660) died on signal 11.

2014-01-22 Thread Sebastian Kuzminsky
On 1/22/14 11:33 , Curtis Dutton wrote:
 Does anyone know what may have caused this?

 So I'm running the circular arc blend beta from Roberts github repository.
 I'm running a build based on commit
 8acb5e68a4e9c03b191555f912de53e639576dec


 I'm not sure if this is related to circular arc blend or just latest
 linuxcnc master code.

 Attached is the backtrace. I was running a call (in the command window) o
 tool_change call which is a simple program to touch a tool probe after a
 manual change. It had just finished that subroutine when milltask died.

 Attached is the backtrace.

Please open a bug against 2.6 in the tracker for this (include the 
backtrace and your tool_change.ngc file):  http://sf.net/p/emc/bugs

Is the problem repeatable?

Were you doing anything weird like moving or editing the tool_change.ngc 
file, or maybe a file it was calling, while running?

There's definitely some sketchy file handling code in master, in 
interp_o_word.c's execute_return() function.  Not sure if that's related 
to this segfault or just innocently located near where the problem happened.

Looks like Michael was there last, would you mind taking a look?  I'm 
thinking specifically of the unchecked file re-opening logic just before 
this segfault.


-- 
Sebastian Kuzminsky

--
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments  Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431iu=/4140/ostg.clktrk
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers