NIvisaic crashes with "libnipalu.so failed to initialize"

[acb763se]

Hi, my object is trying to read an Agilent scope using LabView 2012 on a Scientific Linux 6.7 32-bit machine.

I installed latest VISA (15) and scope driver, but all examples found in

/usr/local/natinst/LabVIEW-2012/instr.lib/Agilent\ 2000\ 3000\ X-Series/Examples/

immediately crash LabView upon loading.

As a first step in finding what the problem could be I tried to run NIvisaic, that fails as indicated.

I ran NIUpdateDrivers as indicated and rebooted, but NIvisaic still does not work, and neither the examples.

I would appreciate any help, and attach the result of running niSystemRepor.

Thanks!

[acb763se]

OK, some more information.

1st, I checked that NIvisaic works on another machine, same OS (SL6) but 64-bit, and it does. So the problem is with 32-bitness.

Next, I noticed file nipalk-load.log, made during installation of NIVISA 15:

FATAL: Error inserting nipalk (/lib/modules/2.6.32-573.12.1.el6.i686/kernel/natinst/nipal/nipalk.ko): Kernel does not have module support

GOOGling for this error concerning nipalk (was NIvisaic lying in its teeth denouncing nipalu.so?) I found this page:

https://decibel.ni.com/content/message/23793

seeming to assert that 32-bit machines may work only if they have 4GB, OR if you add mem=4096M to the GRUB boot line

But the page is 4 years old, and my 32-bit machine happens to have exactly 4 GB, so I do not know.

In any case I tried to add the superfluous mem=4096M to the grub line, but then it won't boot at all. SO funny......

Any suggestion? Thanks everybody willing.

[Stefano.Caiola]

Have you installed the 64 bit version or the 32 bit version of drivers?

Is the machine you used where everything works a 64 bit or 32 bit machine?

[acb763se]

Hi, thank you for the answer.

The machine where "everything" works (but is not connected to my scope, being the desktop

in my office) is 64 bit.

As for NI-VISA15 being 32- or 64-bits, I do not know: it is the only version for SL6 available on the National site,

so the site does not give me a choice of bitness. The README.txt (in attachment) does not declare

its own bitness either, and I do notk know how to check. I only know that it seems to compile everything.

I also tried NI-VISA14 but it does not compile because of libstdc++ incompatibility: it requires

a version of libstdc PREVIOUS than the one I have

[Stefano.Caiola]

From this KB:

http://digital.ni.com/public.nsf/allkb/20789C3E4A3DCAFB8625708B00693CF9?OpenDocument

"However, in some cases, mem=4096M may allocate 4GBs outside of the addressable range. This is why it is encouraged to use the memmap method as opposed to the mem=XX method. "

Check if with this something changes!

Cheers 

[acb763se]

OK, the new switch memmap DID make the trick, and I have been able to boot and install everything (thanks Scaiola88!).

The Agilent 3000x-series examples now do load without choking LabView, and so does NIvisaic. Whoopee.....

Problem is, the scope _doesn't_ show up in the VISA sources: neither from a common user process

nor when running as root. And, yes, I did put the bus in write mode for everybody.

On the other side, my _other_ USB interface (to CANbus if you are interested) _does_ show up no problem.

Could the scope USB bus be broken? I don't think so, because it shows up w

---

Scaiola88 ha scritto:

From this KB:

http://digital.ni.com/public.nsf/allkb/20789C3E4A3DCAFB8625708B00693CF9?OpenDocument

"However, in some cases, mem=4096M may allocate 4GBs outside of the addressable range. This is why it is encouraged to use the memmap method as opposed to the mem=XX method. "

Check if with this something changes!

Cheers 

---

hen running lsusb,

disappearing when I pull the USB cable, and reappearing directly I plug it back.

Could the scope firmware be only minimally supporting USB? Maybe, but in any case

I think the problem is getting now outside this forum: too tricky, and I give up.

Thank you all!