[Coco] SuperIDE backup/restore
Gustavo Ranaur Schoenaker
ranaur at ranaur.net
Mon Jul 29 13:39:44 EDT 2013
Just a short update.
I got two 256MB cards and copied the Luis's image on them (dd ing them).
They work perfectly. Only the 512MB card is broken.
I really don't think my sIDE is broken. It is working very well on the
256MB cards. I got 3 cards, so I don't need to make the 512MB anymore.
Actually, I'm still testing it just for the fun. :-)
I made a short test, "dding" /dev/zero through the entire card. Plugged the
card on sIDE and DSKINI the first disk. OK.
Then I wrote a small basic program to DSKO$ every sector of every disk with
a string showing the DISK/TRACK/SECTOR count.
I extracted the image with DD and analyzed it (with hexdump).
The fist HDB-DOS disk starts at 0xA5BB800 (ok, that is the magic Cloud-9
number 52DDC * 512 full sector size) and ends 630 * 256 sectors after. This
makes sense and shows that I'm not messing with some hidden partition. For
the records I'm using /dev/rdisk4 on a mac. (never use /dev/rdisk4s1: this
is the partition, /dev/rdisk4 is the entire disk) You can use /dev/disk4
aswell.
I was changing the program do DSKI$ every sector and compare the contents
with the "ideal" position. I did it, but I'm getting ?OS errors on the
DSKI$ when I read the sectors to the strings, well ... I got to fix it
somehow.
Then it was 3:20 in the morning and I needed to go to work next day. :-)
My pending list is:
- Run the same program to write sectors into a 256MB card and compare
the images. The hypothesis is that due some bizarre CF IDE emulation it may
be writting the sectors in a different order: instead CHS it is writing
CSH. I got to figure out this. This would explain a lot of phenomena I get,
but the odds are small, I know.
- Copy the 256MB image into the 512MB and compare it using the program.
- Try to copy the Luis's image only from 0xA5BB800 onwards. I'll loose
OS-9,but this may shine a light on what is happening.
I'll post the results.
Ed,
I have an idea you may try. Put one "broken" CF card into sIDE as
initialize the disk 0 with DSKINI. Save some dummy basic program to fill
the disk. Make the image of the card, and write it to another "broken"
card. See if it works.
Good luck,
On Mon, Jul 29, 2013 at 11:48 AM, Luis Antoniosi (CoCoDemus) <
retrocanada76 at gmail.com> wrote:
> it's not the sIDE. It's something about the CF cards.
>
>
>
>
> On Mon, Jul 29, 2013 at 10:04 AM, Mark Marlette
> <mmarlette at frontiernet.net>wrote:
>
> > sIDE user's,
> >
> > I really don't know what to say here. Since 2004 when the device was
> > released, I have found only one generic CF card, no name brand, that
> > wouldn't work properly on the sIDE. This had to do with some internal
> > caching of the data WRITTEN to the CF. The cache had to be flushed
> >
> > before powering down. READS were fine of course, so it was noted and not
> > used.
> >
> > I did look at Ed's cards, he sent four in, one didn't have enough
> clusters
> > so, the image wouldn't fit on it. All devices are not created equal.
> When I
> > picked the top sector count, I analyzed all the 256MB CFs and picked a
> > number down from the last sector so that I would run into this problem.
> > This one of Ed's is really short clusters. So I didn't image that one.
> >
> > The rest I imaged and functioned no problem.
> >
> > Gustavo's example was flawed in his copying of the CF at the time. As
> > IIRC, he just picked a sector to stop copying at. You have to do the
> math.
> > He stopped copying too early. Not sure where his games where or if this
> was
> > related or not. NOS starts at secotr 0 and works up. HDB-DOS starts at a
> > sector I picked at the top and works it's way down. There is a partition
> in
> > the middle that is unused.
> >
> >
> > Since I was able to make image Ed's CF here and he can't. I would
> normally
> > say then the sIDE is defective(only two ever and those where
> > bumped/pulled/inserted with power applied, blew CPLD), BUT the image
> that I
> > made works on his device. ????
> >
> >
> > So that leads me then to the process/tool. Ed is using Winhex, just as I
> > am. This is a power tool and has all kinds of options. Not sure if that
> is
> > an issue as well.
> >
> > Don't have any free time to work on this to figure the cause...already
> > taking vacation from dayjob to work on orders.
> >
> > I have no more CFs but the last 100 or so I bought from Ebay from a guy
> in
> > Chicago. SanDisk. He sold single pieces as well. As my $16 charge is
> higher
> > it does pay for imagining the CF and some limited support.
> >
> > Since NOS, SuperDriver and HDB-DOS are now free and were Boisy's
> products.
> > I did the hardware design and production builds, it is REALLY hard to
> > devote much time to this to figure this issue out.
> >
> > If you think your hardware is defective, please contact me via our
> support
> > web page contact information and I'll setup a RMA and we can take a look
> at
> > it. No charge but you will have to pay for the round trip
> > postage....Gustavo's case....OUCH! :(
> >
> >
> > Open to ideas, but I am out ATM.
> >
> > Regards,
> >
> > Mark
> >
> >
> >
> >
> >
> >
> >
> > ________________________________
> > From: Gustavo Ranaur Schoenaker <ranaur at ranaur.net>
> > To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com>
> > Sent: Sunday, July 28, 2013 10:48 PM
> > Subject: Re: [Coco] SuperIDE backup/restore
> >
> >
> > Ed,
> >
> > I'm having a very similar problem with my SuperIDE. I have the 256GB
> > original Cloud 9 card, and I would like to backup it an work from a copy.
> >
> > Initially I used one of 4GB Transcend Card. I DD the content from the
> > original Cloud-9 to it. I could boot it, but when I run som problems, the
> > computer hangs.
> >
> > I tried both on a CoCo 2 and a CoCo 3 with the very same results.
> Original
> > OK, copy bad.
> >
> > For example, everytime I ran Calixto Island, I got some vertical black
> bars
> > on the credit screen (I sent the pictures from the list two weeks ago),
> and
> > the games went crazy when I start.
> >
> > Then I bought a 512MB and a 256MB card. They're not easy to find theese
> > days ...
> >
> > When I copied to the 512MB I got the very same vertical bars and those
> > strange behaviour. With the 256MB card everything went fine. I`m using
> it!
> >
> > I tried everything (with some help from Retrocanada) to use the 4GB and
> the
> > 512MB card with no luck. I also have a Powerpak, a 8 bit Nintendo
> cartridge
> > that reads the ROM from a CF card. The 512MB card is not recognized on
> > powerpak either.
> >
> > BTW, the card works perfectly on a Linux, a Mac or a windows machine.
> >
> > There is something very, very strange in the way CF cards works. Who is
> the
> > culprit? The CF card's IDE emulation? Maybe ... but why does it works on
> > Intel Systems? SuperIDE? Maybe but why it hasn't worked on Powerpak
> either?
> > HDB-DOS way of accessing the information? Maybe ... but this doesn't
> makes
> > sense, since it only uses sector count. Everything together? Maybe ...
> > Ghosts from 80`s haunting your old CoCo? Maybe ...
> >
> > I just bought two more 256 CF cards. I'll post as soon as I have more
> > information.
> >
> > Up to now the best way to backup your data is saving your data on a
> > Drivewire disks using wired. But it takes time *and* patience.
> >
> > Keep us informed. I`ll keep you informed as soon as I have more info.
> >
> > Good luck!
> >
> > On Sun, Jul 28, 2013 at 6:04 PM, Ed Orbea <ed.orbea at gmail.com> wrote:
> >
> > > List Members
> > >
> > > My goal is to be able to backup and restore my SuperIDE CF Card.I could
> > > copy each HDB-DOS virtual drive to physical floppies (and then copy
> them
> > to
> > > DSK image files).
> > >
> > > I could use dsave and make copies of the contents of the various
> NitrOS-9
> > > directories and sub directories to physical floppies (and then copy
> them
> > to
> > > DSK image files).
> > >
> > > But why do I need to do that?
> > >
> > > I "should" be able to use (via WinXP) the 'dd' utility from Cygwin or
> the
> > > Win32DI application or WinHex (from X-Ways software) or Disk Imager (by
> > > Roadkil.net) to extract an image of the entire CF drive and then
> restore
> > it.
> > >
> > > In order to try to do this, I purchased several 256mg CF cards (which I
> > > have referenced in previous posts).
> > >
> > > I successfully used the above listed utilities to extract an image from
> > my
> > > Cloud9 distribution card.
> > >
> > > I generated as MD5 hash for each image (after it was extracted from the
> > CF
> > > card) and all hash codes were identical.
> > >
> > > I tried several different CF card (that Mark was kind enough to test
> for
> > > me, as I have been struggling with this for several weeks).
> > >
> > > The 'target' CF cards are all identical to the Cloud9 distribution CF
> > card
> > > according to that is shown via the IDEPROBE utility contained on the
> > > Clloud9 distribution disk
> > >
> > > Logical cylinders: 980
> > >
> > > Logical heads: 16
> > >
> > > Logical sectors: 32
> > >
> > > Total sectors: 501760
> > >
> > > LBA: Is Supported
> > >
> > > PIO Timing: 2
> > >
> > > The only differences are the Serial Number (obviously this should be
> > > different) and the # sectors per R/W multiple(Cloud9 distribution disk
> =
> > 1,
> > > the other CF cards = 4).
> > >
> > > So, I should be able to "extract" the image from the Cloud9
> distribution
> > > CF card and "restore" it to any of the 'matching' CF cards.
> > >
> > > The problem is that I can't. And let me explain further.
> > >
> > > I can extract the image (without any failures or errors)
> > >
> > > I can generate a MD5 has of the actual CF card and the extracted
> > > image, and they match. That tells me that I do not have to do a
> > > byte-by-byre comparison, as the image is an exact copy of the card.
> > >
> > > I can then restore the image to one of the other CF cards without
> > > encountering an error or failure
> > >
> > > I can then generate a MD5 hash of the new CF cards and that hash
> > > matches what has previously generated.
> > >
> > > But when I put the new CF card into the SuperIDE device and power on
> > > the system (CoCo3), I see the Disk Basic banner, the HBD-DOS LBA
> > > banner, and then the system either: starts showing 'sparklies';
> > > starts showing random colors and letters; or just hangs.
> > >
> > > The LED on the SuperIDE remains a solid color (no blinking).
> > >
> > > I have tried using only the SuperIDE cartridge (no MPI)
> > >
> > > I have tried various MPI configurations/combinations.
> > >
> > > Everything works correctly, when I am using the Cloud9 distribution
> card
> > > and doesn't work when using the new CF card
> > >
> > > My CoCo system (as mentioned in previous posts) is:
> > >
> > > CoCo3 (6809)
> > >
> > > Cloud9 Triad memory upgrade
> > >
> > > Roy's VGA converted
> > >
> > > MPI with slot 4 (either Disto SCII FDCC with DW3 in Eprom OR TRS-80
> > > 26-3129 'long' FDCC); Slot 3 Orch-90; Slot 2 Modified SSC; Slot 1
> > > SuperIDE.
> > >
> > > Your help is appreciated
> > >
> > >
> > > --
> > > Coco mailing list
> > > Coco at maltedmedia.com
> > > http://five.pairlist.net/**mailman/listinfo/coco<
> > http://five.pairlist.net/mailman/listinfo/coco>
> > >
> >
> > --
> > Coco mailing list
> > Coco at maltedmedia.com
> > http://five.pairlist.net/mailman/listinfo/coco
> >
> > --
> > Coco mailing list
> > Coco at maltedmedia.com
> > http://five.pairlist.net/mailman/listinfo/coco
> >
>
>
>
> --
> Long live the CoCo
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
>
More information about the Coco
mailing list