> David Brown wrote:
> > Hi Arnd-Hendrik,
> >
> > I did not realise that msp430-gdbproxy accepted a parallel port like
this
> > (although I knew about the choice of tcp/ip port for communicating with
> > gdb).  That would certainly be a help if I can first get access to the
> > second parallel port.
>
> the argument is passed to the HIL.
>
> > Unfortunately, my PC (like most these days) has no ISA slots, so I can't
add
> > a second ISA parallel port.  That would probably have worked straight
away,
> > since HIL.c has the second ISA bus parallel port address hard-coded.
>
> you could use a parallel port switch and try switching while the CPU
> runs. you can also try to reconnect without downloading.  but of course
> that not as confortable as having both debuggers open at the same time.
>

My "parallel port switch" consists of having two debuggers connected to the
two boards, and moving the cable between them :-)  I generally seperate the
tasks of downloading and debugging (I make "downloading" a target of my
makefile) - I find that I get fastest downloads using msp430-jtag rather
than downloading via gdb.  An alternative setup would be to use serial BSL
downloading, so that I could at least reprogram the cards without swapping
over the debugger.

> (if you had a second PC, you could also run gdbproxy only on that one
> and connect from the other box over TCP/IP. that box does not need to be
> fast or have much memory as it does not need to run a full IDE or
debugger)
>

That's certainly an option, and a likely solution for the future.  At the
moment, I've got both programs working pretty way, so it's not worth the
effort for this project.

> > And unfortunately, I'm not running Linux (except as colinux) on this
> > machine.  It looks (from HIL.c) like it would be easier to get things
> > working with Linux, even with a PCI parallel port, since HIL.c passes
> > requests down to the /dev/parportX driver rather than trying to guess
the
> > direct hardware address of the parallel port.
>
> windows does not seem to have a API for parallel port bit-banging (just
>   limited printer access). that's the reason HIL.c does the lowlevel
> stuff  on itself. there is currently only a (Windows) implementation for
> the old ISA style IO mapped devices. to my knowledge, the PCI devices
> are not mapped to the IOs, but to the memory (which are different
> concepts on the x86 family processors)
>

I believe, but can't be sure, that my PCI parallel port card *is* mapped to
IO.  Windows (w2k) device manager reports it as using IO ports
0xb000..0xb007, which looks to me like it could be accessed as a standard
parallel port at a non-standard address.  I don't know of any way to get
that address information from windows.  For my own part, I'd be quite happy
to pass this on to the tools via the command line - it's easy enough for
users to find out manually.  I've had a look around the web for information,
and I could see no reliable way of getting the port addresses from a
program - only ones I saw involve searching the windows registery, and I
think manually passing the port number on the command line would be both
easier and more reliable.

It must somehow be possible to get this information - Altera's fpga software
identified my second parallel port, and I've run jtag programming using
their ByteBlaster parallel port adaptor on lpt2.

When I get the chance, I'll try some bit-banging on my lpt2 and report any
results back to you.

mvh.,

David




> if somebody knows of a free higher level API that allows bitbanging
> access to the parallel port, it would eb nice to know about it ;-)
>
> chris
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime
info,
> new features, or free trial, at: http://www.businessobjects.com/devxi/728
> _______________________________________________
> Mspgcc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/mspgcc-users
>
>
>



Reply via email to