Conner 2.5" CHS (non-LBA) hard drive archival

[Zucca]

So I have this old Conner laptop hard drive. I try to archive its contents.
It's from Intel 386 era. I use USB adapter to attach it to my, much modern, box.
The drive advertises its name to kernel as "Conner P eripherals 84MB". So that's working. In the name we can also get somewhat accurate capacity of the drive.

However this happens when I plug in the drive:

Code: Select all

[1480665.965916] usb 1-4: new high-speed USB device number 10 using xhci_hcd
[1480671.127372] usb 1-4: New USB device found, idVendor=152d, idProduct=2338, bcdDevice= 1.00
[1480671.127412] usb 1-4: New USB device strings: Mfr=1, Product=2, SerialNumber=5
[1480671.127432] usb 1-4: Product: USB to ATA/ATAPI bridge
[1480671.127446] usb 1-4: Manufacturer: JMicron
[1480671.127460] usb 1-4: SerialNumber: 142014524147
[1480671.132692] usb-storage 1-4:1.0: USB Mass Storage device detected
[1480671.133596] scsi host2: usb-storage 1-4:1.0
[1480672.160244] scsi 2:0:0:0: Direct-Access     Conner P eripherals 84MB       PQ: 0 ANSI: 5
[1480672.160904] scsi 2:0:0:1: Direct-Access     Conner P eripherals 84MB       PQ: 0 ANSI: 5
[1480672.163426] sd 2:0:0:0: Attached scsi generic sg2 type 0
[1480672.163875] sd 2:0:0:0: [sdc] Very big device. Trying to use READ CAPACITY(16).
[1480672.169696] sd 2:0:0:1: Attached scsi generic sg3 type 0
[1480672.170497] sd 2:0:0:0: [sdc] Using 0xffffffff as device size
[1480672.171080] sd 2:0:0:1: [sdd] Very big device. Trying to use READ CAPACITY(16).
[1480672.172043] sd 2:0:0:0: [sdc] 4294967296 512-byte logical blocks: (2.20 TB/2.00 TiB)
[1480672.173025] sd 2:0:0:1: [sdd] Using 0xffffffff as device size
[1480672.173653] sd 2:0:0:0: [sdc] Write Protect is off
[1480672.173679] sd 2:0:0:0: [sdc] Mode Sense: 28 00 00 00
[1480672.174530] sd 2:0:0:1: [sdd] 4294967296 512-byte logical blocks: (2.20 TB/2.00 TiB)
[1480672.175103] sd 2:0:0:0: [sdc] No Caching mode page found
[1480672.175127] sd 2:0:0:0: [sdc] Assuming drive cache: write through
[1480672.175655] sd 2:0:0:1: [sdd] Write Protect is off
[1480672.175684] sd 2:0:0:1: [sdd] Mode Sense: 28 00 00 00
[1480672.176338] sd 2:0:0:1: [sdd] No Caching mode page found
[1480672.176363] sd 2:0:0:1: [sdd] Assuming drive cache: write through
[1480672.190617] sd 2:0:0:0: [sdc] Very big device. Trying to use READ CAPACITY(16).
[1480672.191522] sd 2:0:0:0: [sdc] Using 0xffffffff as device size
[1480672.191973] sd 2:0:0:1: [sdd] Very big device. Trying to use READ CAPACITY(16).
[1480672.192859] sd 2:0:0:1: [sdd] Using 0xffffffff as device size
[1480672.196442] scsi_io_completion_action: 8 callbacks suppressed
[1480672.196468] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.196494] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.196510] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.196529] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.196539] blk_print_req_error: 8 callbacks suppressed
[1480672.196546] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.196567] buffer_io_error: 6 callbacks suppressed
[1480672.196572] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.200190] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.200217] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.200231] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.200245] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.200254] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.200272] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.201639] sd 2:0:0:0: [sdc] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.201659] sd 2:0:0:0: [sdc] tag#0 Sense Key : Illegal Request [current] 
[1480672.201672] sd 2:0:0:0: [sdc] tag#0 Add. Sense: Invalid command operation code
[1480672.201684] sd 2:0:0:0: [sdc] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.201693] critical target error, dev sdc, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.201709] Buffer I/O error on dev sdc, logical block 0, async page read
[1480672.202496] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.202515] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.202527] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.202540] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.202548] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.202563] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.207200] ldm_validate_partition_table(): Disk read failed.
[1480672.211853] sd 2:0:0:0: [sdc] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.211874] sd 2:0:0:0: [sdc] tag#0 Sense Key : Illegal Request [current] 
[1480672.211886] sd 2:0:0:0: [sdc] tag#0 Add. Sense: Invalid command operation code
[1480672.211899] sd 2:0:0:0: [sdc] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.211907] critical target error, dev sdc, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.211923] Buffer I/O error on dev sdc, logical block 0, async page read
[1480672.212708] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.212725] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.212737] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.212748] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.212756] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.212770] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.213525] sd 2:0:0:0: [sdc] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.213540] sd 2:0:0:0: [sdc] tag#0 Sense Key : Illegal Request [current] 
[1480672.213549] sd 2:0:0:0: [sdc] tag#0 Add. Sense: Invalid command operation code
[1480672.213559] sd 2:0:0:0: [sdc] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.213565] critical target error, dev sdc, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.213577] Buffer I/O error on dev sdc, logical block 0, async page read
[1480672.214232] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.214247] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.214256] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.214266] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.214272] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.214283] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.214491] ldm_validate_partition_table(): Disk read failed.
[1480672.215209] sd 2:0:0:1: [sdd] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.215225] sd 2:0:0:1: [sdd] tag#0 Sense Key : Illegal Request [current] 
[1480672.215235] sd 2:0:0:1: [sdd] tag#0 Add. Sense: Invalid command operation code
[1480672.215245] sd 2:0:0:1: [sdd] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.215253] critical target error, dev sdd, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.215265] Buffer I/O error on dev sdd, logical block 0, async page read
[1480672.215996] sd 2:0:0:0: [sdc] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[1480672.216010] sd 2:0:0:0: [sdc] tag#0 Sense Key : Illegal Request [current] 
[1480672.216020] sd 2:0:0:0: [sdc] tag#0 Add. Sense: Invalid command operation code
[1480672.216030] sd 2:0:0:0: [sdc] tag#0 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 08 00 00
[1480672.216036] critical target error, dev sdc, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
[1480672.216048] Buffer I/O error on dev sdc, logical block 0, async page read
[1480672.216944]  sdd: unable to read partition table
[1480672.217310] sd 2:0:0:1: [sdd] Attached SCSI disk
[1480672.228359]  sdc: unable to read partition table
[1480672.228459] sd 2:0:0:0: [sdc] Attached SCSI disk
[1480672.386306] usb 1-4: reset high-speed USB device number 10 using xhci_hcd
[1480691.009892] usb 1-4: USB disconnect, device number 10

... so clearly kernel tries to guess the capacity. That's first thing that is wrong.

What physically happens is that the drive starts to spin. Sounds ok. Then continues to spin until spins down. Then spins up again... This repeats about five times before it seemingly gives up. There's no new messages on dmesg after the first spin up. Also I can't hear rw head moving at any point. Since it's still IDE (Integrated Drive Electronics) I'd assume it should do some head movements at startup to seek and determine start and end positions.

So my guesses are:

• read write head is stuck
• kernel doesn't wait long enough for the drive to be ready (timeout issue)
• I need to manually tell kernel the number of sectors etc.
• I'm missing the correct driver from kernel
• the problem lies in the adapter itself (dmesg shows two drives)

The drive itself doesn't have much writings on it. There are barcodes... I could try to scan them if needed.

So... any old timers here who have experience on old hard drives and preferably connecting them to modern machines?
Point me to a some direction?

EDIT01: Changed the title to reflect better what kind of Hard drive is in question.

[steve_v]

Likely of no real help in your situation, but I have archived many similar drives, going back as far as 1989... And the answer to all uncertainty is "use a real IDE controller, preferably a period-correct one."
In my case that's a K6-2 or 486 machine, which I keep around largely because USB IDE controllers are a complete pain in the arse, and some software only runs properly in real-mode DOS with an ISA bus... Also DooM sounds better with a hardware wavetable

One thing I would suggest is messing with the master/slave/auto jumpers if the drive has them. I've seen plenty of USB adapters that make incorrect assumptions, and sometimes the "wrong" jumper setting is the one that will have it identify the drive.

I've seen drives that don't do anything besides spin the platter until they are selected (though I can't speak of that particular unit), and all manner of funky behaviour and initialisation-order jank from USB adapters.
As much as old 2.5" drives were never particularly reliable, personally, my bet is #5.

[b11n]

Your USB adapter may not be helping. I always avoid them when dealing with flaky disks, because they often respond to drive errors in unpredictable ways. Errors you see come from the usb-storage protocol, and not from the IDE controller, so troubleshooting is a guessing game.

Laptop drives this old draw a lot of power. This is probably manifesting as an undervoltage condition in the USB adapter itself. It could be drive is spinning up but the drive electronics simply fail to operate in the undervoltage condition. This could also explain the device size being reported as 0xffffffff blocks.

Of course, all of this could also be explained by the r/w head being stuck, or the drive electronics being rotted away. There are many unknowns, and the USB adapter is adding to them.

I know your options are probably limited, but if I were in your situation I'd be looking for a machine with a native IDE host controller and a 3.5" to 2.5" adapter cable before going any further.

edit: Oh, snap. Hi Steve!

[Zucca]

One thing I would suggest is messing with the master/slave/auto jumpers if the drive has them. I've seen plenty of USB adapters that make incorrect assumptions, and sometimes the "wrong" jumper setting is the one that will have it identify the drive.

I've seen drives that don't do anything besides spin the platter until they are selected (though I can't speak of that particular unit), and all manner of funky behaviour and initialisation-order jank from USB adapters.
As much as old 2.5" drives were never particularly reliable, personally, my bet is #5.

Thanks for the reply.
I'll try to get more info from the drive itself to determine how to set master/slave settings for that drive. There are four extra pins on the back in a 2x2 formation, just beside the small IDE interface.

If software tricks don't work, then I'll start blaming the adapter. Unfortunately I lack the perioid correct hardware... I do have few laptops, but with one hard drive slot Iäll face "some" obstacles.

EDIT: Oh hi b11n! Thanks for your input too.
The usb provided power could be a limit indeed. Although the drive should be able to run on USB power, but there are losses on the adapter itself too.

Oh well. I try to read and scan every bit of text and barcodes on the drive to have more information about the drive first.

[b11n]

If you're desperate, you could whip up a custom bootable USB / CD with dd_rescue built in, to help with that (if the old Connor even fits, height-wise).

[Zucca]

if the old Connor even fits, height-wise

Well... you guessed it. It's, like yongsters tend to say, thicc boi.
Is also poses some challenges.

[NeddySeagoon]

Zucca,

It looks wrong here.

Code: Select all

[1480672.160244] scsi 2:0:0:0: Direct-Access     Conner P eripherals 84MB       PQ: 0 ANSI: 5 

Conner would not have added the space after the P in Peripherals. It was their company name :)

Look at the drive and tell us its model number. The specs are probably on the web.

You may have more luck with an PATA to SATA converter.

[b11n]

Conner would not have added the space after the P in Peripherals. It was their company name :)

Are you ready for a surprise?

Code: Select all

[ 2226.679088] scsi 6:0:0:0: Direct-Access     Conner P eripherals 120MB  - C PQ: 0 ANSI: 4

[ 3909.336782] scsi 9:0:0:0: Direct-Access Conner P eripherals 340MB PQ: 0 ANSI: 5

(and many more around the web)

[steve_v]

Laptop drives this old draw a lot of power. This is probably manifesting as an undervoltage condition in the USB adapter itself.

I was actually typing a sentence regarding power, then I thought "nah, every USB adapter I have takes external power input, (because 12v rail for 3.5" devices), surely Zucca's is the same" and stopped.

This, apparently, was a mistake. Power is a very good point if it's coming from the host USB port.

[NeddySeagoon]

b11n,

Eewww. What more can I say?

[steve_v]

Are you ready for a surprise?

Huh, my Conner CP-344 (incidentally, oldest fully-operational IDE drive on hand) is the same.
Would'na thunk it, but there it is.
Also incidentally, I didn't try particularly hard but I never managed to get that drive working with a modern host - not with 2 different USB adapters or through an IDE-SATA bridge. Runs perfectly with no bad sectors on a real IDE controller though.

[b11n]

There's probably some interesting tech-lore behind it, lost to the passage of time. If you told me that some space-strapped firmware developer discovered that by adding a space. "r P er" was also an x86 subroutine that did something useful, I'd believe you. OG machine language programmers did the craziest stuff.

[Zucca]

Oh.

https://retrocomputing.stackexchange.co ... how-do-i-c

I think my options just got limited to zero. I need an IDE-card.

[b11n]

Look on the bright side: your (immediate) issue is known, we have a better explanation than controller bit-rot.

IMO want need an IDE controller anyway. That disk is probably going to throw at least some read errors, and you don't want to have doubts about how the USB converter will deal with those. You'll be wanting to use dd-rescue to get as much image as you can, and quickly defer and move to other areas when encountering problems, saving the troublesome sectors for last. The USB controller might thwart that process.

[Zucca]

you don't want to have doubts about how the USB converter will deal with those. You'll be wanting to use dd-rescue to get as much image as you can, and quickly defer and move to other areas when encountering problems, saving the troublesome sectors for last. The USB controller might thwart that process.

That's true.

I need to start digging out my old hardware... I might as well setup one as a some kind of backups server at the same time... one which can read IDE hard drives.

[Zucca]

Ok.
The drive has type plate, but it's been "smudged" by some QA stampings. *sigh*
I think the model is CP2084.

Then there is the sticker which seem to have clear text marking along with their barcode euivalents:

Code: Select all

AF9G8ZA
CP2084
9317
P72113 PAN4NT1.62 ML2
TBB04

... the fist line is serial number (most probably), which is also on another sticker which has SN at the beginning.

About the power requirement... The drive type plate/sticker states 5V 450mA. So it should be able to be powered though USB. I assume that the adapter just sends the USB power trough to the hard drive... I doubt the adapter electronics take too much current... But that really doesn't help me as the problem lies in USB -> SCSI -> ATA chain... I believe.

There are some cheapish PCIe SATA+IDE cards on ebay. I have the feeling those have almost the same problem.
Best bet would be to buy some regular PCI, perioid correct, used card from ebay.
Then I also would need 3.5" to 2.5" IDE adapter cable.

Oh this is going to be a long jorney. I'll shelve this as long as I gather required parts... be it from ebay or from my piles of relics.

EDIT: I guess I could try the adapter one more time: https://www.computerhope.com/hdd/hdd0031.htm
EDIT2: Nope. That site reports spin up takes 1.1A...

[NeddySeagoon]

Zucca,

I suspect a drive that small predates LBA, so you are going to need something that can do C/H/S addressing. It certainly predates 48 bit LBA.
You may even need to know the actual geometry of the drive to get started.
A long time ago, I recall entering the geometries of HDD into the BIOS to be able to use drives.

It's a CP2084. I'll check the web when I'm on a PC, not my phone.

[Goverp]

It might be worth checking your motherboard for IDE connectors. IIRC my relatively modern MB, full of PATA connectors, also had a couple of ribbon cable connectors gathering dust. And some old-style cables on my power supply.

[Zucca]

I recall entering the geometries of HDD into the BIOS to be able to use drives.

I vaguely remember doing this few times.

It might be worth checking your motherboard for IDE connectors. IIRC my relatively modern MB, full of PATA connectors, also had a couple of ribbon cable connectors gathering dust. And some old-style cables on my power supply.

Mine is Dell T7810. I don't think it'll have any IDE capability. Although it has PCI port...

I'd rather dig out my old desktop... eventually.

[NeddySeagoon]

Zucca,

In case you are still looking https://www.computerhope.com/hdd/hdd0031.htm and https://stason.org/TULARC/pc/hard-drive ... DE-AT.html for two.

That it needs (or may need) CMOS settings is scary.

I suspect that USB to IDE dodas assume LBA so you are doomed going that way.

[Zucca]

Neddy, yes. I think I'd actually need some special driver or the PCI IDE card and maybe pass it to some old PC emulator.
... or some way to specify cylinder, head and sector counts...

I guess this drive is from late 80s. So predating Linux. My best bet is to find a PC from around the same time period.

[Zucca]

Interesting writing on Wikipedia:

In 2002 the ATA-6 specification introduced an optional 48 bits Logical Block Addressing and declared CHS addressing as obsolete, but still allowed to implement the ATA-5 translations. Unsurprisingly the CHS to LBA translation formula given below also matches the last ATA-5 CHS translation.

[steve_v]

some way to specify cylinder, head and sector counts

Usually that was set up by the host BIOS or an opROM on the card, but that only matters if the OS uses BIOS routines to access the disk... Which Linux does not and never has.
Long ago, with the original IDE driver stack, we had 'hd=cyls,heads,secs' as kernel command-line parameters, but I have no idea how the modern libATA drivers handle that.

My best bet is to find a PC from around the same time period.

FWIW, the setup I imaged the aforementioned 42MB conner drive on is an IBM DX4-100 running Slackware 7.1 (kernel 2.2) specifically for disk geometry translation support (ontrack and the like).
I expect you won't need to go back quite that far though, Linux bypasses the BIOS entirely so anything that has a real PATA controller should work, even if it lacks C/H/S setup options (unless you want to boot from it of course).
That kind of hardware is still pretty easy to find at your local ewaste processor, and common enough that it goes dirt cheap even in the vintage/retro scene.
Ditto drivers, IIRC the old ide stack wasn't fully removed until somewhere around kernel 5.x, so if libata's autodetection is a problem you'll just need an old-ish (as opposed to prehistoric) distro.

[NeddySeagoon]

Zucca,

I have a Gentoo install from 2003 if you want to play.
You can even play the Gentoo way by installing it yourself.

[Zucca]

Thanks Neddy. That might come useful later on.

Meanwhile, I found some reading: https://www.kernel.org/doc/Documentation/ide/ide.txt