Lpc2000

> > > Same experience here. Tried to do the interrupt thing, now I just 
> > > start the SPI transfer, do some other stuff and then poll 
> until the 
> > > transfer is ready. Is there anybody that has used the SPI 
> interrupt 
> > > successfully???
> > 
> > I use SPI interrupts for SPI transactions and haven't run into any 
> > issues as yet. Mind you it's on a lpc2124, not an lpc2138. My SPI 
> > module is part of the lpc newlib extensions.
> 
> Hey Bruce,
> 
> Care to enlighten me about that last comment.
> 
> > lpc newlib extensions
> 
> I've been meaning to go through the newlib re-entrancy 
> configuration etc to see what needs to be turned on for 
> uCOS-II thread-safe usage, and to create some sort of device 
> operations table interface (as per Bill Gatliff's docs from 
> Embedded Systems Programming).
> I saw a post the other day from someone at Aeolus regarding 
> their newlib port layer (perhaps that was you)
> 
> http://www.aeolusdevelopment.com/Articles/download.html
> 
> anyways, if you have a nice clean way of implementing a 
> device layer, let us know and people can contribute with a 
> common interface.

No it wasn't me. It was the Robert Ansett, the compiler of newlib_lpc.
I'm not sure if he has rolled my SPI module into his latest newlib_lpc
release as yet, but if necessary I can email it to you.
Yes, it is good to conform to a common interface, and due to other
newlib_lpc modules being available it significantly cut down my overall
effort, even though I coded the SPI, and added some zero-copy options to
the the uart modules.
The easiest way to code a new module is to copy an existing one and
modify the bottom end. In this way I have also created an IIC module,
but haven't had a chance to test as yet. Robert also has a copy of that,
but beware it is probably buggy.

Cheers,
Bruce