[Coco] Assembler question
gheskett at wdtv.com
Thu Sep 15 11:39:17 EDT 2011
On Thursday, September 15, 2011 10:26:14 AM Helio Emery Jr. did opine:
> Hi Gene, thanks for the tip.And sorry for taking so long for me to
> respond, I was stucked with some things in my job and wasn't looking
> personal e-mails for a while. I really liked your tip, but didn't know
> where to find krnp3 and krnp4 for the OS-9 Level 1, v2.00 I'm using.Can
> you direct me? Thanks,Helio > From: gheskett at wdtv.com
First, the level 1 version of os9 is now called Nitros9 level 1, and the
tools to build boot disks are both more complex, and more versatile than
the level 1 version 2.00 was, and has some optimizations that make it
faster to boot. Those 2 kernel extensions were written, then rewritten
after level 2 shipped by at least a couple years so they never were on a
std Tandy shipped diskette.
If you can write to an os9 formatted diskette (windows, see rawwrite, linux
see 'dd') (some modern floppy hardware no longer does 256 byte sector disk
formats) then bootable images of the level 1 disks can be downloaded from
using the other downloads link to select which level, and which cpu (some
of us have replaced our 6809 cpu's with hitachi 6309 cpu's, an enhanced
instruction set which is 100% compatible with normal 6809 code but about
10% faster. Running code optimized for it, it can be as much as 50%
I have such broken floppy hardware here, so as I run linux on this machine,
I download the .dsk image I want to this box, then since I have an rs-232
pack in my coco3's MPI, I run a shell on /t2 on the coco, then open a
session of minicom in a shell here, and send the disk image (.dsk) to the
coco's hard drive, useing rz on the coco and sz here, with a -W 256 option
to sz to keep it from upsizing the transmission window beyond the Sacia's
i/o buffer. I run both ends at 9600 baud, but miss-matched handshaking
problems might call for 2400 baud as that's slow enough that no xon/xoff vs
7 wire should come up unless you are writing to a floppy with a halting
controller, but that should be handled with the -W256 option to sz.
Then I can 'install' it to a floppy formatted SSDD with this basic09
INPUT #1,"destination device:",oname
REM now reset lsn0 of target device
REM that should stop the disk change errors
REM by resetting os9's disk id data
ON ERROR GOTO 10
WHILE NOT(EOF(#ipath)) DO
10 CLOSE #opath
PRINT "bytes written:"
Or, I have been known take a spare descriptor and treat that .dsk image as
a 'partition' of a bigger disk, so I can cd to it, and execute the install
procedure or one of the mb files in the scripts directory directly from the
disk image just as if it was on a real disk. It is, but it happens to be
one of a pair of 1Gb seacrate hard drives, but the descriptor, if done
right, hides this from os9.
However, the drive SW in your install will not do that 'partition' trick
If your system is floppies only, try the basic09 program above.
If your system has only SSDD drives, then maybe one of us can send you a
disk set if we knew where to send it. That may even be the best option,
but my 5.25" floppy is dying, a worn out head I think. With a cable adapter
plugged into most any 3.5" drive, it will become /d1, giving you a 720k
drive for almost nothing but a little bin diving for the drive. Xmode /d1
sid=2 cyl=50. voila! But HD disks will probably forget overnight, so find
some DD disks while bin diving or yard sale touring.
These disk images are re-made fresh every night, so they are current as of
the day of the download. Believe it or not, development and bug fixing
continues with the latest commits being disk size limit fixes and cluster
size fixes for both cobbler and os9gen earlier this week, with many thanks
to Robert Gould who did _all_ the heavy lifting.
Good luck & Cheers Helio, gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Didn't I buy a 1951 Packard from you last March in Cairo?
More information about the Coco