Martin Kunze [EMAIL PROTECTED] wrote:
Digital logic told me that they can not tell me a abs. address
for this device and they dont have a device driver for it.
No, they indeed can't tell you an absolute address as the ISA IO address
range is mapped into some address range through PnP services at system
startup and can vary from PC to PC and even between startups. That they
do not have a device driver for their board is a much more serious problem.
Somebody today selling PC hardware without at least Windows NT support
drivers is definitely in the wrong business. But hey maybe they support
Linux much better.
How can I find out the base address of a ISA device connected via a PCI bridge from
LV?
You can try to check in the Device Manager. Usually you see the assigned resources
in there but that doesn't give you a guarantee, that this address couldn't change
after a reboot. With Linux this would be in fact easier.
How can I talk to the device from LV? Is it possible to use InPort an OutPort from LV?
Well, here we get into Plug Play OS services with resource enumeration. The ISA
bus on the other side of the PCI bridge is typically mapped into memory space at
startup by the OS. It could also be mapped into the upper IO address space sometimes
but that is rather tricky. An ISA bus can theoretically have 64kB IO address range
and the 80x86 architecture has only 64kB IO address range so trying to map secondary
PCI bus addresses into the 80x86 IO address range has the possibility of conflicts
with other IO addresses. Therefore PCI bridges sometimes map IO addresses to memory
addresses to avoid such problems. The particularities are not that simple and in fact
that is usually the task of the device driver to use kernel PnP services to map the
addresses internally.
I have no specific experience with this device nor in fact with any ISA IO port
access through a PCI bridge and so can't really help you here but I'm afraid it
is not going to be very simple. Maybe you can influence something on BIOS level
and try to force your OS to accept the BIOS settings but in XP this chance is
also quite small as Microsoft rather wants to manage this stuff themself than
trying to work around buggy PnP BIOS implementations.
Does anybody have a device driver for things like that in connection with Win XP?
No sorry, but you could check out the portIO device driver on OpenG. As it is, it
does not care about PCI bus mapping in the function HalTranslateBusAddress() but
you could extend it to do so, eventhough Microsoft says this function is depreciated
in favor of newer PnP enumeration kernel services, which however are not available
before Win2000. I didn't want to use newer APIs as the driver should run on as much
Windows versions as possible. As to the limited mapping, I didn't feel that this would
be a huge limitation and in lack of hardware to test anything on didn't really care,
but in your case it would be necessary to actually allow for that as well.
The portIO driver is part of the OppenG Toolkit but not currently included in the
binary distribution of the Toolkit. You can get it however from the sourceforge CVS
repository of the OpenG Toolkit or directly though CVS web access.
http://cvs.sourceforge.net/viewcvs.py/opengtoolkit/portIO/
Rolf Kalbermatter
CIT Engineering Nederland BVtel: +31 (070) 415 9190
Treubstraat 7H fax: +31 (070) 415 9191
2288 EG Rijswijkhttp://www.citengineering.com
Netherlands mailto:[EMAIL PROTECTED]
