Experimental Physics and
Industrial Control System

"Ernest L. Williams Jr." <[email protected]> · Mon, 20 Nov 2006 11:23:03 -0500

On Mon, 2006-11-20 at 16:28 +0100, Dirk Kraemer wrote:
> Eric Norum schrieb:
> > On Nov 20, 2006, at 5:50 AM, Dirk Kraemer wrote:
> >
> >> Hello all,
> >>
> >> I have a problem with EDM. If I start EDM ,then
> >> I see for a moment the message  "loading fonts" and a moment later
> >> I get the message "segmentations fault".
> >>
> >>
> >> I am running EDM on :
> >>
> >> - Fedora 5 - gcc 4.1.0-3.i386
> >> - Epics 3.14.8.2
> >> - openMotif 2.3.0-0.1.9.i386
> >> - edm 1-11-0s
> >>
> >> I hope that someone can help me. It's the first time that I get this 
> >> fault.
> >>
> >> Regards
> >> Dirk Kraemer
> >>
> >>
> >>
> >
> > Sounds suspiciously like the problem I encountered on OS X.  Here's a 
> > copy of the letter I sent to Jeff about this.  Perhaps you're seeing 
> > the same sort of condition, but with some other pair of static 
> > constructors.
> >
> > ====================================================================================== 
> >
> >
> > I downloaded and built the latest version of edm with the current 
> > R3.14.2 CVS branch of EPICS.  Things build fine, but the executable 
> > suffers an access fault when run.  I tracked the problem down to a 
> > race condition in the C++ static constructors.  Code in 
> > EPICS_PV_Factory is invoking a guard class mutex that has yet to be 
> > set up.
> >
> > Program received signal EXC_BAD_ACCESS, Could not access memory.
> > Reason: KERN_PROTECTION_FAILURE at address: 0x00000008
> > 0x0008ae23 in epicsMutex::lock ()
> > (gdb) where
> > #0  0x0008ae23 in epicsMutex::lock ()
> > #1  0x00088e3a in ipAddrToAsciiEngine::allocate ()
> > #2  0x004aff0b in cac::cac ()
> > #3  0x004cfb5b in ca_client_context::ca_client_context ()
> > #4  0x004b25f4 in ca_context_create ()
> > #5  0x0041f4df in EPICS_PV_Factory::EPICS_PV_Factory ()
> > #6  0x0044cb2b in __static_initialization_and_destruction_0 ()
> > #7  0x8fe0fcb2 in 
> > __dyld__ZN16ImageLoaderMachO16doInitializationERKN11ImageLoader11LinkContextE 
> > ()
> > #8  0x8fe0a1c6 in 
> > __dyld__ZN11ImageLoader23recursiveInitializationERKNS_11LinkContextE ()
> > #9  0x8fe0a164 in 
> > __dyld__ZN11ImageLoader23recursiveInitializationERKNS_11LinkContextE ()
> > #10 0x8fe0a164 in 
> > __dyld__ZN11ImageLoader23recursiveInitializationERKNS_11LinkContextE ()
> > #11 0x8fe0c646 in 
> > __dyld__ZN11ImageLoader15runInitializersERKNS_11LinkContextE ()
> > #12 0x8fe03070 in __dyld__ZN4dyld24initializeMainExecutableEv ()
> > #13 0x0000245a in _start ()
> > #14 0x000023c5 in start ()
> >
> >
> >
> >
> >
> > The offending code in misc/ipAddrToAsciiAsynchronous.cpp is:
> > // for now its probably sufficent to allocate one
> > // DNS transaction thread for all codes sharing
> > // the same process that need DNS services but we
> > // leave our options open for the future
> > ipAddrToAsciiEngine & ipAddrToAsciiEngine::allocate ()
> > {
> >     epicsGuard < epicsMutex > guard ( 
> > ipAddrToAsciiEnginePrivate::globalMutex );
> >     if ( ! ipAddrToAsciiEnginePrivate::pEngine ) {
> >         ipAddrToAsciiEnginePrivate::pEngine = new 
> > ipAddrToAsciiEnginePrivate ();
> >     }
> >     ipAddrToAsciiEnginePrivate::numberOfReferences++;
> >     return * ipAddrToAsciiEnginePrivate::pEngine;
> > }
> >
> >
> >
> >
> > As a quick fix I just link libCom before the EDM libraries as well as 
> > after:
> > diff -ur /home/phoebus/NORUME/Desktop/edm/edmMain/Makefile 
> > /home/phoebus/NORUME/
> > src/EPICS/extensions/src/edm/edmMain/Makefile
> > --- /home/phoebus/NORUME/Desktop/edm/edmMain/Makefile   2006-03-06 
> > 13:25:32.0000
> > 00000 -0600
> > +++ 
> > /home/phoebus/NORUME/src/EPICS/extensions/src/edm/edmMain/Makefile  
> > 2006-11-
> > 14 13:12:30.000000000 -0600
> > @@ -10,7 +10,7 @@
> >     edm_LIBS_Linux += 114135a4-6f6c-11d3-95bc-00104b8742df
> >     edm_SYS_LIBS_Linux += Xm Xmu Xt Xp Xtst X11 pthread dl
> >
> > -   edm_LIBS_Darwin += EdmBase cfcaa62e-8199-11d3-a77f-00104b8742df
> > +   edm_LIBS_Darwin += Com EdmBase cfcaa62e-8199-11d3-a77f-00104b8742df
> >     edm_LIBS_Darwin += 114135a4-6f6c-11d3-95bc-00104b8742df
> >     edm_SYS_LIBS_Darwin += Xm Xmu Xt Xp Xtst X11 pthread dl
> >
> >
> >
> >
> >
> > I've sent mail to Jeff to see if he has a better fix, but for now I 
> > can at least work with EDM.
> >
> > --Eric Norum <[email protected]>
> > Advanced Photon Source
> > Argonne National Laboratory
> > (630) 252-4793
> >
> Hi
> 
> I have tried it with your changes, but the result is the same.
> 
> GNU gdb Red Hat Linux (6.3.0.0-1.122rh)
> Copyright 2004 Free Software Foundation, Inc.
> GDB is free software, covered by the GNU General Public License, and you are
> welcome to change it and/or distribute copies of it under certain
> conditions.
> Type "show copying" to see the conditions.
> There is absolutely no warranty for GDB.  Type "show warranty" for details.
> This GDB was configured as "i386-redhat-linux-gnu"...Using host
> libthread_db library "/lib/libthread_db.so.1".
> 
> (gdb) run
> Starting program: /home/Dirk/extensions/bin/linux-x86/edm
> Reading symbols from shared object read from target memory...done.
> Loaded system supplied DSO at 0x41cfd000
> [Thread debugging using libthread_db enabled]
> [New Thread -1208293152 (LWP 7736)]
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1208293152 (LWP 7736)]
> 0x00000000 in ?? ()
> (gdb)

Make sure that you are not having a problem with fonts.
Sheng Peng recently had run-time core dumps due to some font issue.
Hopefully, Sheng will reply to this thread.

I and not having any problems and have the following setup:
=====================================================================
OS: Fedora Core release 5 (Bordeaux)

KERNEL: Linux bengal 2.6.18-1.2200.fc5

Compiler: gcc version 4.1.1 20060525 (Red Hat 4.1.1-1)

MOTIF: openmotif-2.3.0-0.1.9.3

EPICS_BASE: R3.14.8.2

EDM: edm 1-11-0s
=======================================================================

Thanks,
Ernest

> 
> regards Dirk
> 
> 
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System