[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: CC112 VME board



On Mon, Jun 02, 2003 at 11:13:06PM +0200, Kars de Jong wrote:
> On Sun, 2003-06-01 at 15:58, Frank Gevaerts wrote:
> > I have a compcontrol cc112 VME board. Specs are:
> > 
> > - 20 MHz 68030/68882
> 
> Should be no problem.
> 
> > - 4 MB local RAM
> 
> Hmm. That's not much RAM. But if you don't need networking it might work
> (painfully slow though). Do you have VME RAM? Or a VME Ethernet board?

Both. I have a VME RAM board that seems to be 4 MB (36 TC511000P-85
chips), and an ethernet board (seeq 8005 based).

> > - 2 serial ports (SCN2681)
> 
> I am writing a driver for this chip at the moment. I already have a
> basic one based on drivers/char/sb1250_duart.c (which looks a lot like
> the SCN2681) but I'm now writing one using the generic_serial framework.
> 
> Early startup and serial console works fine. In the current version of
> arch/m68k/kernel/head.S there is already support for the SCN2681 since
> this chip is also present in the Apollo port.
> 
> > - SCSI (wd33c93)
> 
> Probably shouldn't be a problem if you get the DMA working.
> 
> > - an OS/9 boot eeprom with debugger
> 
> If you can download code this greatly simplifies things ;-)

I can download it from a serial line. Not very fast.

> It also helps if there are entry points into the firmware so you can use
> it to do boring stuff like read from the harddisk or write to the serial
> port. I recently ported Linux/m68k to a Plessey CPU board which has just
> about NO firmware, which wasn't too much fun.

The firmware can load OS9 from disk, so I might try to masquerade as an
OS9 kernel. I think I have all necessary documentation for that.

> > I have full documentation, but not enough time/experience to get linux
> > running on this. I have so far managed to load some code from a serial
> > port that printed some messages to the other serial port, but I'm lost
> > as soon as I look at the kernel boot and memory mapping code.
> 
> Well, most of the memory mapping code is generic these days. All you
> basically need to do is add some mappings for the I/O regions if they're
> mapped in low memory normally, moving them up. Or, if they are already
> 'up there' you can just use transparent translation. All this can be
> done using some simple macros, plenty examples in the file.
> 
> You might also want to add support for printing some characters on one
> of your serial ports there. Shouldn't be too hard.
> 
> I only had to add about 40 lines of code to head.S for the Plessey port.
> 
> Getting a boot loader working is IMO a lot harder than getting the basic
> kernel to work. It took me about 2 weeks to add a new port to VMELILO
> and only 2 days or so to get my first kernel panic ;-) 

Right now I'm thinking about cheating : I just bought an mvme147 from
ebay. If I manage to get that one booting (should be easy, just build a
new P2 connector module), I might use it to put a kernel in the CC112
RAM...

> If you have callable firmware it makes things a lot easier. I had to
> write a small SCSI driver...

I'm afraid I'll have to do the same. I don't think I can make an OS9
lookalike big enough to contain a full kernel.

> > I guess that once I have a booting kernel I can manage the individual drivers.
> > 
> > I would like to get this running so I can use my VME serial and digital
> > IO boards, but I do not really know how to proceed. Another alternative
> > would be trading it for a supported VME board with someone who likes a
> > challenge.
> 
> If you need more help, please contact me directly. I'd be happy to help
> you out.

Thanks

Frank

> Kind regards,
> 
> Kars.



Reply to: