Software only cloning tool that can do soft resets?

[maximus]

Are there any software only cloning tools that can perform a software reset after a set time instead of waiting for the drive to finish with error each recovery attempt and report a bad sector/read?

A more specific example: software performs read command, after 250ms there is no success so a soft reset is done, then after reset the next read is attempted.

[fzabkar]

Just FYI ...

http://www.deepspar.com/pdf/DeepSparDis ... paper3.pdf

[fzabkar]

This is interesting :

For simplicity, the ECC bytes can be thought of as achecksum to validate data integrity. Most current hard disk drives
are Advanced Technology Attachment (ATA) compliant devices. The system software reads the drive sector-by-sector,
using the standard ATA read sector command. If a sector has a wrong ECC checksum, it returns no data, just an error,
despite the fact that most of the bytes in the sector are likely correct and all are actually readable, even if some of the
data is corrupt. The system software is unable to access data byte-by-byte, using other ATA read commands that read
data regardless of ECC status.

On our real/write long experiment when both of you damaged the sector with wrong ECC on write long you couldn't retrieve data from the sector with "Read" using the BIOS/MHDD Script, while i was using HRT PCI card and the "reading" using the dre would allow to see the data on the sector even with bad/wrong ECC.

Maybe DRE transparently uses READ LONG or some other means. In any case, the ATA standards committee have retired this command (and SCT READ LONG), stating that "recent changes in recording technology have made the idea of a host being able to actually force its own error correction/detection data to the device meaningless". I have also seen statements to the effect that, on modern drives, a good sector that is read using READ LONG or SCT READ LONG wil return scrambled, meaningless data. Perhaps you should experiment with these commands on a more recent drive.

ACS-2 Obsolete SCT Read and Write Long - T13 (February 19, 2009):
http://www.t13.org/documents/UploadedDo ... e_Long.pdf

[data-medics]

The best software option you have is ddrescue in linux. I'm not sure if it actually sends soft resets, but it manages to image on most drives without completely hanging on bad sectors. Nothing in Windows is capable of doing this, the OS prevents it unless you have hardware like a PC-3000 or MRT card which is directly controlled by the software instead of the OS.

[fzabkar]

The was some research done by "xsoliman" and "Enoch" that attempted to understand the ECC system on Seagate's F3 drives:

http://hddoracle.com/viewtopic.php?t=713
http://hddoracle.com/viewtopic.php?f=59&t=665

[maximus]

The best software option you have is ddrescue in linux. I'm not sure if it actually sends soft resets, but it manages to image on most drives without completely hanging on bad sectors. Nothing in Windows is capable of doing this, the OS prevents it unless you have hardware like a PC-3000 or MRT card which is directly controlled by the software instead of the OS.

As stated by Spildit in a post above, ddrescue is very much unable to do what I ask. And I am VERY familiar with what ddrescue can do, as I have done some benchmarking tests with it and have had an influence on the latest releases in regards to the algorithm. Not to mention my passthrough patch for ddresuce.

If ddrescue is currently the best software option (and I do believe it is unless someone can state otherwise, which is why I am asking the question), then that would mean that there is not a software solution that can perform as I asked.

[maximus]

This question is not about read long. If anyone wants to test how and if read long works, look up my ddrutility software, the utility ddru_diskutility can perform read long commands. It can even do an average of the reads. My hddscripttool can also perform read long commands. Both can do the newer SCT long sector access.

But long commands are obsolete (twice over). And my experience is that even when they do work they are not very productive. Until someone can produce real results that we can see, I don't consider read long as a usable option for accurate data recovery.

[data-medics]

I would think it's theoretically possible to make software to do it, however I'd bet it would have to be a bootable custom build. In any OS there's going to be drivers already managing the drives which I'd tend to doubt allow software to send such direct commands to the drive. Possibly you could build a custom driver to plug into an existing OS such as Linux which would allow this, but you'd need someone with extensive knowledge of these drivers. This might be a better question to ask on a Linux development forum. The ATA commands themselves should be pretty straight forward if the driver allows direct commands to be given.

[guru]

What's the cheapest HW/SW imaging tool?

[data-medics]

That actually works, probably MRT..

[data-medics]

I don't know if it'll help, as I'm not a programmer and can't make much sense of the code. But here's some ATA scripting written for Linux which seems to include code to call a soft reset: http://gaztek.sourceforge.net/osdev/hardware/ata.c

Just search for the term "soft reset" in the code.

[guru]

An that thing. What's the best, is it Deepspar? Must admit I haven't looked at current market hardware for many years

[guru]

Support individual head imaging?

[guru]

Why is it better than DDI4?

[fzabkar]

The ATA commands themselves should be pretty straight forward if the driver allows direct commands to be given.

The OP has written "pass-through" patch for ddrescue that does exactly that.

[fzabkar]

It doesn't address the OP's quest for a software solution, but I've always thought that it might be worth exploring the possibility of placing the PCB's "POK" signal, or equivalent, under control of the imaging tool. Hopefully, inactivating this signal will not cause the drive to unload its heads and spin down the motor. With any luck it'll behave like a watchdog timeout and cause the code to reset itself without any "brutal" physical activity. ISTM that a Maxtor DiamondMax Plus 9 drive would be a suitable candidate for experimentation since it uses the well documented SMOOTH L7250 motor controller, as well as identifying the test points on the PCB. A single open-collector transistor driven from a COM, LPT, or USB port would then suffice to control this signal.

[guru]

@spildit no DDI4

[maximus]

I would think it's theoretically possible to make software to do it, however I'd bet it would have to be a bootable custom build. In any OS there's going to be drivers already managing the drives which I'd tend to doubt allow software to send such direct commands to the drive. Possibly you could build a custom driver to plug into an existing OS such as Linux which would allow this, but you'd need someone with extensive knowledge of these drivers. This might be a better question to ask on a Linux development forum. The ATA commands themselves should be pretty straight forward if the driver allows direct commands to be given.

Can be done directly in Linux without a driver (the program can be the driver). The BIOS has to be in IDE mode for SATA drives (AHCI will not work as there would definitely be a conflict with the driver). And there is the possibility that the hardware (computer) does not properly emulate the IDE mode and Linux ends up using the AHCI driver anyway (have seen that already on one of my computers). Those computers would simply not work with SATA. But computers are cheap compared to data recovery hardware. The up side is that it can run on a computer with Linux installed, or run from a live CD (you have to provide the program on something like a USB stick as it would not be on the live CD).

The trick is to plug the drive in AFTER Linux boots up, and the drive should be the only one on that "channel". I can see a drive when it is plugged in after boot, but Linux does nothing to even give it a listing as a drive. And if it doesn't recognize there is a drive there and there is no other drive on the same channel, there should (in theory) be no driver conflict issue. Still has to be fully tested, but initial tests are proving to be good.

I just figured out how to do the direct I/O last week, and am wondering what the market would be for something like this if I can make it work and be safe and dependable. The downside is I don't do graphical interfaces, so it would be command line. But I can still do simple menus. And I may just have to learn ncurses so I could make a command window interface like testdisk uses.

[maximus]

Hard (power cycle) resets could be done easily with a USBmicro U451 USB Relay Interface. Programming examples are already provided. And it is affordable. More info found here:
http://www.usbmicro.com/

Edit: Just ordered one

[data-medics]

I'm not sure how much it costs but EPOS makes a BadDrive Adapter which is a hardware device capable of setting the timeouts and I'm pretty sure it'll handle the soft/hard resets as needed, while still allowing the system to access the drive as normal: http://www.epos.ua/view.php/products_ep ... ve_adapter

Might be a nice option for you if the price isn't bad.