m68k SPI or QSPI drivers?

[Aaron J. Grier]

On Wed, Mar 21, 2001 at 05:28:46PM -0700, "Michael P. Collins" on korat wrote:
> I have an application for which I must operate the SPI on an MC68331
> in slave mode.  I've used this SPI in master mode quite a lot, simply
> polling the status register, but it looks as though I'll need to
> enable the receiver interrupt to avoid overruns in slave mode.  Does
> anyone know of an RTEMS driver which might exist that I can use?  One
> for the QSPI would be ideal, but I'd expect any working SPI driver
> would probably be good as a starting point.

I have a very simplistic master-only QSM driver internal to my BSP
implemented as a library, as well as a rather general GPT driver.  I'd
move them into libcpu, but they are woefully incomplete and depend on a
set of headers entirely separate from those in exec/score/cpu/m68k.

the GPT and TPU don't have headers distributed with RTEMS, and as far as
I know, there aren't any "official" motorola C headers for the 683xx
family, so everything's kind of ad-hoc.  I wrote my headers for the
68331 before I even started with RTEMS, and have put them in libcpu, (I
can provide them to anybody who's interested,) but as the '331 is the
only member of the 683xx family I'm familiar with, I'm not sure if
they'd integrate well with other family members, or if I'd be better off
putting them in a 68331-specific directory.  How general do we want to
make the include files?  should 68331, 68360, and 68302 share header
files, for instance?

Joel?  what's the vision for libcpu?

-- 
  Aaron J. Grier   |    Frye Electronics, Tigard, OR   |   aaron at frye.com