To: None <port-atari@NetBSD.ORG>
From: Leo Weppelman <email@example.com>
Date: 01/24/1996 09:28:00
I'm going on holiday tomorrow. As I don't have an internet connection
during that time (this is intentional ;-) ), don't expect answers to
mail directed to me in the oncoming 1.5 weeks.
As I am writing to you anyway, I can give a quick status report on the
Atari-port just as well.....
- A lot of work is being done to create a bootloader for NetBSD/Atari.
It allows most people to (re)boot NetBSD without starting GEM. I think
that the bootloader stuff is nearly in the beta-stage. Waldi (who has
done most of the job) and I are currently running the alpha-version.
I said 'most people' on purpose :-( As we don't have enough docs on
the FX-card and are not able to get it, it seems that the loader is
not going to work for it. So if you have tech-info about it or when
you are able to get it, drop Waldi or me a mail.
- It seems that the cause for the 'kernel-hang' when closing one of the
serial ports (ser02/mdm02), as reported by Thomas and Waldi, has been
found and fixed. Fetch a new atari/dev/zs.c and recompile.
- A new device-driver was made for the nvram. This driver was needed
for the bootloader. Currently the only info I have about the nvram
are the locations and values for the boot-preference and the scsi-
hostadapter id. If you know more, drop me a mail.
- Work is being done to fix linked-commands in the scsi-driver. I
introduced linked-commands to speed things up. Some drives however
go completely bananas when a linked command is send to them :-(
So if you experience lockups/resets on the scsi-bus please say so.
Currently I'm debugging this with Allen Briggs & Paul Goyette. There
are still some loose ends.
- When the previous point is finished, I am planning to work on the
Falcon scsi again. I still have some patches lying around that were
tested by Thomas Gerner to allow DMA on interrupt basis on the
- I have made a patch to increase transfer-rates on the serial port.
The first results look positive. On the TT transfer-rates of
2250-2300cps (gzipped files) seem possible. Also depends on your
modem of course ;-)
- The -current kernel has support for a separate statistics clock. So
when you are interested in (kernel) profiling, you should give it
The above list is probably incomplete, but long enough ;-)