From: Ullrich von Bassewitz (uz_at_musoftware.de)
Date: 2003-09-27 16:02:02
Hi!
On Thu, Sep 25, 2003 at 11:16:02PM +0200, Groepaz wrote:
> maybe you want to contact jens schönfeld (jens_at_ami.ga) and tell him you need
> hardware to create/test/support/whatnot rr/silversurfer/rr-net drivers...
> with some luck you might get stuff sponsored (i did :o))
I have a Swiftlink, which is enough to test the swiftlink driver. My plan was
to get a SilverSurfer, but this NMI thingie is somewhat disappointing, because
it means that the hardware cannot be used together with standard disk
routines. But the RR-Net is a must, of course:-)
> actually the use of NMI for this is more or less a "tradition".... also most
> software that uses the swithlink (or any other serial cart for that matter)
> uses NMI. i think there are other reasons aswell... the latency involved in a
> "normal" IRQ vs NMI (routed over $0314 etc vectors) may make a difference.
Especially for the 16550, the latency wouldn't matter because of the big
internal buffer. Using an interrupt after 8 chars in the FIFO, you have 8 free
slots in the FIFO plus the one in the shift register. This means that even
with 57600 or 115200 baud there's enough time to fetch the data from the chip
when using an IRQ. And, using the IRQ would remove all problems with the
standard disk routines. Using the NMI prevents using the disk in parallel
without rather complex countermeasures. This is unfortunate, because using the
IRQ was not really necessary.
Regards
Uz
--
Ullrich von Bassewitz uz_at_musoftware.de
----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo_at_musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.
This archive was generated by hypermail 2.1.3 : 2003-09-27 16:02:11 CEST