Conner CFN170A imaging

[iampaulh]

I'm trying to image this disk using WinUAE 4.3.0

It's a CHS only disk, so I get the prompt to enter the geometry when I try to create a hardfile. The defaults (or detected?) are correct according to the Conner PDF at ftp://ftp.seagate.com/techsuppt/seag...s/allconnr.pdf

and also here: https://stason.org/TULARC/pc/hard-dr...SL-IDE-AT.html

Unfortunately, there is no information on the label of the hard disk.

However, when I start the creation, WinUAE gives an error

Hard drive image file creation failed.
Error 0:-14 ((null)).

However if I change the number of Heads to 15 or 8 or 4, WinUAE will create an image, however, it is unreadable by the Amiga.

Is this a limitation of the IDE to USB interface I have, or something to do with WinUAE

[Toni Wilen]

Hmm.. IDE protocol max heads is 15 but when using SCSI read hack, it probably will accept larger (logical) head values..

Does http://www.winuae.net/files/b/winuae.7z read it correctly? (Don't use it for anything else, it is current development version)

[iampaulh]

Same error as before, unfortunately

[Toni Wilen]

Right, default geometry of 16 heads is a bit weird..

Lets see what log shows, run with -log command line parameter and do imaging attempt with 16 and 15 heads. Attach logs (winuaebootlog.txt, you can strip beginning of log files, there is nothing important)

[mark_k]

Did you try the native geometry? That's 1339 cyls, 4 heads, 63 sectors/track. Using the translated geometry will omit sectors near the end of the drive. (Edit: I was wrong about that.)

Also, can you post the drive's IDENTIFY DEVICE data? (The Read Identity button should read that.)

[iampaulh]

I've uploaded logfiles here for 16 heads, 15 heads and native geometry.
https://drive.google.com/drive/folde...K5?usp=sharing

16 fails
15 creates an HDF but the Amiga sees it as non dos
Native create and HDF, but also not a dos disk in device DH0

I noticed the native HDF shows a different geometry than specified when creating the HDF

[mark_k]

My suggestion about "native geometry" above was wrong. The drive might have 1339 physical heads and tracks, but it uses a variable number of physical sectors per track, so that geometry isn't used or relevant for accessing the drive.

I think the 892/6/63 is just made up by WinUAE, based on the size of the .hdf file.

What make/model of IDE-USB converter are you using?

[mark_k]

Originally Posted by Toni Wilen

Hmm.. IDE protocol max heads is 15 but when using SCSI read hack, it probably will accept larger (logical) head values..

??? IDE supports 16 heads (0 to 15), as used by the default geometry of any large drive (16 heads, 63 sectors/track).

[Toni Wilen]

Originally Posted by mark_k

??? IDE supports 16 heads (0 to 15), as used by the default geometry of any large drive (16 heads, 63 sectors/track).

I counted it from 1 like CHS mode sectors but I forgot that heads are 0 to 15...

winuae.7z updated. It really should work now. Heads=16 was considered out of range.

[iampaulh]

No errors using 16 heads this time, though the log file (updated version in google drive) is showing a lot of ATA errors.

IOCTL_ATA_PASS_THROUGH_DIRECT ID failed 00000032
IOCTL_SCSI_PASS_THROUGH_DIRECT: 28.00.0B.00.27.01.00.00.3F.00. OK (32256 bytes)

I'm guessing the problem is now my USB to IDE converter, a Hiyatek USB2 to IDE https://www.komplett.no/product/350789

Does anyone know of a USB or PCI(E) to IDE that supports CHS?

[Toni Wilen]

Originally Posted by iampaulh

No errors using 16 heads this time, though the log file (updated version in google drive) is showing a lot of ATA errors.

IOCTL_ATA_PASS_THROUGH_DIRECT ID failed 00000032
IOCTL_SCSI_PASS_THROUGH_DIRECT: 28.00.0B.00.27.01.00.00.3F.00. OK (32256 bytes)

This is normal. It first tries SAT (which is almost always unsupported), then it tries SCSI passthrough with LBA field set as fake CHS field which seemed to work, at least it returned expect number of bytes. (Can't say if returned data is correct..)

Did you try mounting the image? or checking it in hex editor if it starts with normal looking RDSK block?

I'm guessing the problem is now my USB to IDE converter, a Hiyatek USB2 to IDE https://www.komplett.no/product/350789

Does anyone know of a USB or PCI(E) to IDE that supports CHS?

So far it seems there is none..

[mark_k]

I haven't tried, but I would think a cheap PCI-E IDE card is likely to support CHS accesses (which would be done automatically by the Windows driver).

Does IOCTL_ATA_PASS_THROUGH_DIRECT cause Windows to send either ATA PASS THROUGH (12) or ATA PASS THROUGH (16) SCSI commands? Or does it only apply to native IDE/SATA controllers?

Also, I seem to remember WinUAE only uses ATA PASS THROUGH (16) for CD/DVD devices. But perhaps some USB converters only support that and not ATA PASS THROUGH (12)? The 16-byte command, apart from not clashing with the SCSI opcode for some CD/DVD function, supports 48-byte ATA commands.

[iampaulh]

I tried a cheap PCI-E IDE card, and it didn't even detect the hard disk. I have another old amiga hard disk, which it did detect.

Looking at the HDF file from the CHS disk obtained with the latest WinUAE beta, the hex view does not show a normal RDSK block at the header,


but a little further down at 600, it does appear.


Then even further down, you can see some text from an email on the disk

[Tomislav]

First is MBR. Some Amiga HD partition tools like in AROS can make that combination of MBR + RDB.

Also MICROBTX is for Microbotics.

[mark_k]

Have you tried booting from the new image file in WinUAE?

If you post a hex dump of the other PART sectors I can tell you which offsets/lengths the individual partitions are at.

[Toni Wilen]

RDB screenshot looks valid. RDSK PART block offset matches PART block and next offset points to following 512 byte block (which is not visible in screenshot but probably is another PART block). RDSK block does not need to be located in first block.

[iampaulh]

I tried mounting the image, but it just comes up with "not a dos disk in DH0"

I also tried mounting the hard disk directly via the USB adapter, and it also gives the same error.

I connected the disk to my A1200 again just to make sure, and it works fine.

[mark_k]

Are you able to upload the first (say) couple of megabytes of the image file? (You might want to view it in a hex editor before uploading to make sure there's no personal data in it.)

Maybe there could be some duplicated sectors?