Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 07 Mar 2000 16:35:12 -0600

[email protected] wrote:
> 
> 1) missing lib vix functions such as SetMODID

I can't help with that one, but...

> 2) missing or renamed unsolicited interrupt handlers
>    excStub and excIntStub.
>   (from  devlib.c defaultHandlerNames[])
> 3) Is anything else tobe used in place of intHandlerCreate()
>    for powerPC...or is this handled in some other way?

These two are known about, what version of EPICS are you using - I think
3.13.2 has sections of veclist.c #if...#endif'ed out for PPC CPUs.  Unless
your BSP provides some code for the vxWorks intVecGet() routine the EPICS
veclist output is of no use anyway (all zeros IIRC). I have code for the
mv2700 board that makes it somewhat more useful.

You also want to be careful about multiple interrupts on the same vector -
your BSP may allow several intConnect()s to the same vector number and
call the ISRs in turn.

> 4) sysBusTas() seems to be different for the PowerPC but
>    I can't find out what the difference is in the WindRiver
>    manuals. Does anyone know how this function differs for the
>    PPC?  When calling sysBusTas() from the allenBradley driver
>    the following occurs:

sysBusTas() on your PPC CPU may not be implemented in a way that is
compatible with the 68k TAS instruction, as the PPC can't generate
Read-Modify-Write cycles itself (it provides an alternative method for
multi-CPU atomic operations that involves bus snooping). Does the KSCorp
board use a Tundra Universe II chip to interface to the VME/VXI bus?  If
not then it may be impossible to control an AB scanner module from this
board; you have to be able to generate a true VMEbus Read-Modify-Write
cycle to talk to the AB cards.  I wrote some code which makes the Universe
II chip do this for the APS version of the mv2700 BSP.  If you don't have
a Universe chip you'll have to find out more about the VME interface
hardware.

- Andrew
-- 
Complexity comes for free, Simplicity you have to work for.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System