Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Wed, 8 Sep 2004 15:20:14 -0600

Hello Jinhu,

I created bug entry 130 in the mantis bug tracking system on the EPICS site.
Any of the following information that you can provide might speed up
reproducing the problem here.

o The version of EPICS that you are using
o The OS name, and OS version that you are using
o Stack trace information form the debugger, pstack, etc 
o Information about how your application is terminated (POSIX signal, ^C,
GUI, etc)

Please send this information via e-mail or else enter it directly into
Mantis as you see fit.

Thanks for your assistance.

Jeff

> -----Original Message-----
> From: Song, Jinhu [mailto:[email protected]]
> Sent: Wednesday, September 08, 2004 11:48 AM
> To: [email protected]
> Subject: Segmentation fault for CA client(EPICS 3.14.6)
> 
> Hi,
> 
> When the EPICS foreground thread is not the main thread(started by
> main(...)),
> it ends up with segmentation fault when I tried to shutdown my
> application gracefully (trace showed it happened at the last line:
> "exit(0)".
> 
> Quick look at the "ca" code, maybe it is caused by atexit( ).
> 
> The fault disappeared after I changed code to let the main thread
> create the context and let the EPICS foreground thread join it.
> 
> I am implementing a gateway to EPICS for a popular user interface
> of synchrotron x-ray (DCSS/Blu-Ice: http://smb.slac.stanford.edu/blu-
> ice/)
> It automatically configures the mapping from the DCSS server.
> Update messages from EPICS can be merged at a configurable rate
> to avoid saturate the DCSS server.
> 
> Jinhu

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System