Minor issue with shell utility "dd". Stopping writing at 65535 blocks using SanDisk Extreme 128GB CF Card on Amiga 600 with MantiCore v4 accelerator.

[Tallguy58]

Hi Again,
I have a minor issue with your shell utility dd.
Wiping my SanDisk Extreme 128GB CF Card with 0s using the shell command dd FILL: scs2.device 0 freezes when the number hits 65535.
Exiting using CTRL-C does not work. The only way to exit is to reboot.
However, if I run the shell command dd FILL: scs2.device 0 0 1048576 512, works as it should.
I plugged the same card into the Amiga's IDE port and run the command dd FILL: scsi.device 0 and works perfectly.
Using other CF Cards, like the Lexar 32GB, in the Amiga's IDE port or the MantiCore IDE port and executing same shell commands works fine.

This is a minor issue, but I thought it was something that should be addressed.

FYI, I should explain what scs2.device is.
This is my system...

The SanDisk Extreme 128GB CF Card plugged into the adapter on the accelerator card is scs2.device.
Thank you,
Alexander.

[pulchart]

Yes, that's possible. In this fork of fat95 the included dd tool hasn't been updated yet, only recompiled from original source (I hadn't really looked at it yet, till now). What I can see is that it's missing NSD/TD64 (64-bit) support, so it can't handle >4 GB ... and maybe other problems with "bigger numbers".

[pulchart]

I prepared a new major update of dd for testing. There is a breaking change in usage, please read doc and adopt usage before use.

• it support >=4GiB IO
• command line parameters parser refactored

Download (wip, not fully tested yet): dd 2.0-dev (27.05.2026)
Documentation: docs/dd.md

[Tallguy58]

Hi,
Loving the new options and the help menu.
The dd FILL: scs2.device 0 still fails at 65535 on the SanDisk Extreme 128GB CF Card.
Plugging the same card into the Amiga's IDE port and running the command dd FILL: scsi.device 0 works fine.
Strange?

Thank you.

[pulchart]

It looks like my assumption that NSD support was needed for scs2.device driver was wrong. It's a useful addition anyway, so no harm done.

I realised IceDrake accelerator has the same scs2.device driver and I could test it locally too (my own blindness for not checking sooner). I can't reproduce the reported problem on my setup as everything works as expected (even with older dd). I suspect you may have a different (probably older) version of the driver?

I added a new dd INSPECT command that probes the IO commands a device exposes. On my Amiga it shows:

  $ dd INSPECT scs2.device
  scs2.device unit 0:
    sector size:    512
    total sectors:  7831152
    cylinders:      7769
    heads:          16
    sec/track:      63
    NSD:            not supported
    for <=4 GiB:    CMD_READ / CMD_WRITE
    for  >4 GiB:    HD_SCSICMD / HD_SCSICMD

my driver version is

$ version scs2.device
scs2.device 43.45

Could you run both commands on your setup and paste the output?

Download: dd 2.0-dev (28.05.2026)
Documentation: docs/dd.md

[Tallguy58]

HI,
Information as requested...

$dd INSPECT scs2.device
scs2.device unit 0:
  sector size:    512
  total sectors:  250085376
  cylinders:      65535
  heads:          16
  sec/track:      238
  NSD:            not supported
  for <=4 GiB:    CMD_READ / CMD_WRITE
  for  >4 GiB:    HD_SCSICMD / HD_SCSICMD

My driver version is...

$ version scs2.device
scs2.device 43.45

Thank you...

[pulchart]

You mentioned the same card works fine when used via the IDE port (scsi.device), right? Could you also plug that 128 GB CF into the Amiga's internal IDE port and run:

dd INSPECT scsi.device

Then paste both outputs side by side so we can compare what the two drivers see for the exact same card. I suspect the scs2.device driver is saturating the cylinder count to the 16-bit (65535 = 0xFFFF) when the disk's geometry exceeds that.

[Tallguy58]

Hi,
Here are the results...

$ dd INSPECT scsi.device
scsi.device unit 0:
  sector size:    512
  total sectors:  250085376
  cylinders:      65535
  heads:          16
  sec/track:      238
  NSD:            not supported
  for <=4 GiB:    CMD_READ / CMD_WRITE
  for  >4 GiB:    HD_SCSICMD / HD_SCSICMD

$dd INSPECT scs2.device
scs2.device unit 0:
  sector size:    512
  total sectors:  250085376
  cylinders:      65535
  heads:          16
  sec/track:      238
  NSD:            not supported
  for <=4 GiB:    CMD_READ / CMD_WRITE
  for  >4 GiB:    HD_SCSICMD / HD_SCSICMD

Here's some more detailed information using the scsiquery command.
I hope this is useful?

$ scsiquery scsi.device 0
SCSIQuery 1.34 (15.12.2018) © THOR
Generic sense data : 

Good.

Device inquiry data : 

Device status           : connected.
Device class            : direct access device.
Removable medium device.
Device type modifier    : 00.
ISO version             : 00.
ECMA version            : 00.
ANSI version            : SCSI-2.
Response data format    : SCSI-2.
Additional data length 1F.
Flags:
	Relative addressing not supported.
	32 bit wide SCSI not supported.
	16 bit wide SCSI supported.
	Synchron data transfer not supported.
	Linked commands not supported.
	Tagged queues not supported.
	Device reacts with hard reset.
Vendor                  : SDCFXS-1
Product                 : 28G             
Version                 : HDX1
Date                    : 
Comment                 : 

Extended inquiry data : 

	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00     ....... 



Control page information:

Mode Data Length : 5
Medium Type      : 0
Device specific  : 0
Block des. len   : 0



Device capacity : 
Capacity in blocks : 250085376.
Block length       : 512.

Blocks from start up to a substiantial delay : 
Number of blocks   : 1008.



Primary defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.


Grown defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.

$ scsiquery scs2.device 0
SCSIQuery 1.34 (15.12.2018) © THOR
Generic sense data : 

Good.

Device inquiry data : 

Device status           : connected.
Device class            : optical memory device.
Removable medium device.
Device type modifier    : 00.
ISO version             : 00.
ECMA version            : 00.
ANSI version            : SCSI-2.
Response data format    : SCSI-2.
Additional data length 1F.
Flags:
	Relative addressing not supported.
	32 bit wide SCSI not supported.
	16 bit wide SCSI supported.
	Synchron data transfer not supported.
	Linked commands not supported.
	Tagged queues not supported.
	Device reacts with hard reset.
Vendor                  : SDCFXS-1
Product                 : 28G             
Version                 : HDX1
Date                    : 
Comment                 : 

Extended inquiry data : 

	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00     ....... 



Control page information:

Mode Data Length : 5
Medium Type      : 0
Device specific  : 0
Block des. len   : 0



Device capacity : 
Capacity in blocks : 250085376.
Block length       : 512.

Blocks from start up to a substiantial delay : 
Number of blocks   : 1008.



Primary defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.


Grown defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.

Thank you...

[pulchart]

At this point I believe this isn't something dd can fix on its own end. The behaviour looks specific to scs2.device when a 128 GB card is attached (I don't own one, so I can't reproduce it locally), and the same card works fine through scsi.device on the IDE port, which suggests the issue lives inside the scs2.device driver rather than in dd. Have you tried contacting the scs2.device maintainers?

[Tallguy58]

I spoke to the Apollo-Core team and they weren't much help.
However I solved the problem, thanks to you.
Instead of using the command...

dd FILL: scs2.device 0

Get the info from the INSPECT option and use the command and this works fine...

dd FILL: scs2.device 0 0 250085376 512

Thank you...