RE: EPICS R3.13.4 Build question???

["Jeff Hill" <johill@lanl.gov>]

Len,

Perhaps the problem is that R3.13.4 will not work correctly with
Tornado I if the SENS IP kernel is installed. I believe that
R3.13.4 is used routinely with Tornado II (SENS IP kernel is always 
used there) and with Tornado I without the SENS IP kernel.

This would point to my ifdef checking the version of the
BSD macro not being a good indication of whether SENS is 
installed or not. This might be testing the version of the
vxWorks kernel only and not whether SENS is installed.
Open to suggestions for a better test.

The problem with this theory is that I dont see that the
message "CAC: SIOCGIFCONF ioctl failed because 6" would be 
occurring either in Tornado I or in Tornado I with the SENS
IP kernel unless this early version of the SENS IP kernel
has a problem with the SIOCGIFCONF ioctl. Problems with
the BSD macro based version macro should exhibit different
symptoms.

Jeff

> -----Original Message-----
> From: Leonard J. Reder [mailto:reder@huey.jpl.nasa.gov]
> Sent: Friday, October 19, 2001 11:33 AM
> To: Jeff Hill
> Cc: EPICS Tech-Talk
> Subject: Re: EPICS R3.13.4 Build question???
> 
> 
> Thanks very much Jeff....
> 
> I think the problem is with our new Tornado 1 installation.
> It seems that EPICS R.3.13.3 when rebuilt gives the same
> errors.  When built with our old install it works fine.
> Since we are upgrading to Tornado 2 in January I think
> I will shelve this for now but may need to trace it in
> the future.
> 
> Thanks for you response it was very very helpful.
> 
> Len
> 
> Jeff Hill wrote:
> > 
> > Leonard,
> > 
> > It would appear that ioctl(SIOCGIFCONF) returns errno six on your
> > system. This indicates "No such device or address". There are some
> > subtle difference in this io control between the IP kernels in Tornado
> > I and in Tornado II, but the call should be initially successful on
> > both OS. So I am confused. What do you see when you type ifShow on
> > your system? Is there any possibility that EPICS was built against
> > Tornado II header files and then loaded into a Tornado I system?
> > I have not tried this. Perhaps there would be undefined symbols.
> > 
> > A quick look at the source does not indicate why your shell is hanging.
> > The "exit" in the output below is probably a clue, but I don't see any
> > code that prints this message in R3.13. If you typed "exit" at a shell
> > prompt a hung ( non-existent ) shell would certainly result.
> > 
> > Otherwise to help us track down the hang please use "sp iocInit" instead
> > of "iocInit", and then use tt <task id> on the resulting task to
> > see where it is hanging. Usually the first task spawned with "sp" has
> > name t1 and so you can alternatively type "tt t1". Please e-mail the output
> > from tt.
> > 
> > Thanks
> > 
> > Jeff
> > 
> > > -----Original Message-----
> > > From: Leonard J. Reder [mailto:reder@huey.jpl.nasa.gov]
> > > Sent: Thursday, October 18, 2001 4:15 PM
> > > To: EPICS Tech-Talk
> > > Subject: EPICS R3.13.4 Build question???
> > >
> > >
> > > Hello,
> > >
> > > I was using R3.13.3 but with our systems upgrade to Solaris 2.8
> > > I decided to build R3.13.4.  It seemed build fine! with Tornado 1
> > > tools.  I have to use Tornado 1 for now.  I am seeing the following
> > > error messages after the iocInit is executed and then the VxWorks
> > > shell hangs.  Does anyone have any ideas about this problem or how
> > > I can fix it?
> > >
> > > Thanks,
> > >
> > > Len
> > >
> > > SCREEN DUMP STARTS:
> > >
> > > sysClkRateSet( 200 )
> > > value = 0 = 0x0
> > >
> > > iocInit
> > > ############################################################################
> > > St###  arting iocInit
> > > @(#)EPICS IOC CORE built on Oct 15 2001
> > > ###  @(#)Version R3.13.4 $2000/12/15 15:52:33$
> > > ############################################################################
> > > Failed to set clock from Unix server
> > > CAC: SIOCGIFCONF ioctl failed because 6
> > > value = 0 = 0x0
> > > exit
> > > iocInit: All initialization complete
> > > osiSockDiscoverInterfaces(): unable to fetch network interface
> > > configuration
> > >
> > > Done executing startup script
> > > /home/keck/reder/kroot/kss/if/to/startup/O.mv2304/dsmTest
> > >
> > > --
> > > ____________________________________________________
> > > Leonard J. Reder
> > > Jet Propulsion Laboratory
> > > Interferometry Systems and Technology Section 383
> > > Email: reder@huey.jpl.nasa.gov
> > > Phone (Voice): 818-354-3639
> > > ____________________________________________________
> 
> -- 
> ____________________________________________________
> Leonard J. Reder
> Jet Propulsion Laboratory
> Interferometry Systems and Technology Section 383
> Email: reder@huey.jpl.nasa.gov
> Phone (Voice): 818-354-3639
> ____________________________________________________