EPICS Home

Experimental Physics and Industrial Control System


 
1994  1995  1996  1997  1998  1999  2000  2001  <20022003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  <20022003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: GreenSpring memory
From: Andrew Johnson <[email protected]>
To: "Burkhard W. Kolb" <[email protected]>
Cc: John Faucett <[email protected]>, [email protected]
Date: Mon, 01 Jul 2002 10:04:52 -0500
"Burkhard W. Kolb" wrote:
> 
> I had a similar problem with the ipOctal card when I have alse an TEWS
> Tip810 canbus card on a 162. The one address which is used to map the
> interrupt vector of the ipOctal card does not seem to be setup in the
> mapping registers to point to the internal address on the IP card, but
> rather seems to be mapped to VME space. I put an external memory card
> in the crate and the problem goes away.
> Looks like the mapping register setup is done in a separate thread and
> not finished yet when the drivers try to access the registers.

That sounds like a very different problem.  The interrupt vector on the
ipOctal card is usually only accessible from the IP memory space because
the IP register space allocated to the card is too small - the UART
registers use it all up.  GreenSpring's VIPC carriers actually provide
more register space than the IPAC spec requires and the GreenSpring
ipOctal module maps its interrupt vector register into that extra area,
but the mv162 sticks to the IP spec so the only way to get to the vector
is through the register's memory space mapping.  This means that you have
to modify the sysPhysMemDesc[] table of the mv162 BSP to provide an MMU
memory area for the IP memory space to appear in.

I don't understand what you're doing with the external memory card, but
there's nothing in the drvIpac code that does mapping like that.

I can't explain John Faucet's problem, although we have found that the
VIPC cards don't produce correct signals on their IACKOUT daisy-chain and
usually have to be placed to the right of all other interrupter cards in
the chassis.

- Andrew
-- 
Larry McVoy: Learn how to think in C++ but don't ever program in it.

References:
GreenSpring memory John Faucett
Re: GreenSpring memory Burkhard W. Kolb

Navigate by Date:
Prev: Re: GreenSpring memory Burkhard W. Kolb
Next: tnetDrv behavior Peregrine McGehee
Index: 1994  1995  1996  1997  1998  1999  2000  2001  <20022003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: GreenSpring memory Burkhard W. Kolb
Next: waveform calculations Rolf Keitel
Index: 1994  1995  1996  1997  1998  1999  2000  2001  <20022003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024