Xilinx IP core drivers for RTEMS- Keith Robertson still around?
gregory.menke at gsfc.nasa.gov
gregory.menke at gsfc.nasa.gov
Wed Nov 8 17:31:20 UTC 2006
I'm about ready with a bunch of pretty well tested gen405 changes
against the RTEMS tree;
- source & makefiles ported to cvs head
- rewritten uart driver, supports multiple uarts, exploits full
configured tx fifo depth buffering, polling mode only.
- considerably re-written soft core xilemac driver, supports up to 8
simultaneous instances of the mac with 32 or 64 bit packet fifos- no
DMA support yet. No hard core mac support in this driver.
- updated opc interrupt controller vectoring
- updated and simplified runtime memory map (workspace & heap
allocations are better defined and more simply controlled).
To bring up RTEMS on a Virtex 4 ppc, the only thing that needs doing is
ensuring the given base addresses for the uart & interrupt controllers
are specified. For the soft mac, the address can be given in the driver
control struct, others are for the moment hardcoded in bsp .h files.
I was thinking I could send out a provisional .diff since these are more
than simple changes, avoiding too much upheaval by commiting directly to
the cvs tree.
Keith Robertson writes:
> D. Peter Siddons wrote:
> > I'm interested in bringing up RTEMS on the PPC in a Memec XC4FX12 board.
> > Has anyone done that? I'd appreciate getting hold of anything close to
> > try and modify. Did anything get deposited anywhere after this email
> > thread? I couldn't see anything on CVS.
> > Thanks,
> > Pete.
> Hi Pete.
> I have a number of the exact same boards as yourself sitting on a desk
> waiting for a port, however, I haven't done it yet. (It's on the cards
> for my time in December.) As to cvs submissions, I've been waiting
> until I sorted the v4 out to submit a proper xilinx bsp.
> That said, I've used a Memec XC2VP50/VP20 extensively, along with a
> number of custom boards based on the V2 pro. They're all v2pro's, but
> there should be very little (if any) difference trying to get it up and
> running -- it's the same ppc405 inside. The code/diffs that have
> floated around this list are tested against this memec v2pro board.
> I don't anticipate a lot of problems or differences moving to the V4FX.
> It should be pretty straight forward. The only difference is the hard
> core emac that the V4FX has. The ethernet driver in my previous ports
> is for a soft core plb/opb xilinx mac. The V4FX has a hard core emac
> with a different (and as far as I can tell from my reading of the docs,
> much improved) register interface.
> If you're so inclined, a driver for the new hard core emac needs to be
> written -- it's not that bad if you follow the other examples. I can
> provide a fair bit of help with that if you get there before me. The
> other option is that you could ignore the hardcore one and continue to
> synthesise the softcore one in the V4. In this case, the driver from my
> diffs will continue to work.
> Hope this helps. Please stay in touch about your progress.
More information about the users