Unknown HPA status

[gorbiWTF]

I've got a lot of drives that report HS? ???. Running hdparm -N /dev/sda gives me this result:
SG_IO: bad/missing sense data, sb[]: 70 00 05 00 00 00 00 0a 04 51 40 00 21 04 00 00 80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
SG_IO: bad/missing sense data, sb[]: 70 00 05 00 00 00 00 0a 04 51 40 01 21 04 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
max sectors = 156301488/1(1?), HPA setting seems invalid (buggy kernel device driver?)

Affected drives are, for example:

• Toshiba MK8037GSX
• WDC WD5000AAKX-603CA0
• WDC WD2500BEKT-60A25T1

Other drives work correctly, either reporting no HPA or successfully responding to changing sector size.
Also tried the same thing in Ubuntu 22.04.3 with the same result.
Using rtcwake -s 5 -m mem to make the drives not frozen doesn't change anything.

Does anyone have ideas what I could do?

[PartialVolume]

Looks like you may already know this stuff, so apologies if I'm telling you things you already know.

Not all drives support HPA or maybe DCO.

In the case of hdparm and nwipe, both programs issue a ioctl command { 0x85, 0x08, 0x0e, 0x00, 0xc2, 0, 0x01, 0, 0, 0, 0, 0, 0, 0x40, 0xb1, 0 }; If the ioctl command returns an error the sense buffer will be read to determine what the error was. That's what is happening above. In the case of nwipe it not only checks DCO/HPA using low level commands like hdparm it also runs hdparm and looks at it's output and looks at the max disc size provided by libparted. It then makes an assessment based on these three results. It looks like nwipe comes to the same conclusion as hdparm, i.e the three question marks in [HS? ???]. i.e the drive is not responding to either HPA or DCO ATA commands. Running nwipe with the -V option you will see all the sense data.

Regarding the max sectors 156301488/1(1?) returned by hdparm. The slash /1 means no valid data was returned. The hdparm code adds a 1 for some reason that I don't know (but suspect it's a bug), but the data returned is not valid.

This could be caused by

• A drive that does not support HPA. I can't tell with this drive as I can't find a comprehensive technical manual that lists all the supported ATA commands for this model/firmware. A plus for Seagate drives as they do tend to provide comprehensive drive programming manuals.
• Iffy firmware on the drive that doesn't respond unless you jump up and down on one leg three times.
• A interface such as raid or USB that does not support the ATA commands that are required to interrogate the drive.

So, regarding iffy firmware.

Some drives, I don't remember which, won't respond to DCO HPA command requests if you just changed, or attempted to change the HPA value. You need to power cycle after every change. Some aren't so finicky.

So assuming you are using a SATA interface and not raid or USB unless the interface supports full ATA pass through I would try seeing if the DCO can be reset, remembering to power cycle then try setting the HPA and power cycle and then check HPA status. So

What does the DCO make of the hard disk size, find the line that starts Real max sectors: xxxx

hdparm --dco-identify /dev/sdx
            Real Max sectors: 976773168

If the above command returns SG_IO: bad/missing sense data then check you are connected to a proper SATA interface, not raid or USB.

Reset the DCO

sudo hdparm --dco-restore /dev/sda

Power cycle the drive after making any change

Check the HPA

hdparm -N /dev/sda

If it now works, great ! if you still get SG_IO: bad/missing sense data try setting the HPA assuming you know what the max number of sectors should be for this drive.

Reset HPA

hdparm -Np976773168 /dev/sda

If you are still getting SG_IO: bad/missing sense data errors, i would probably put it down to the model of drive, especially if you have lots of these drives and they all do the same thing.

It would be nice if it could be confirmed by the manufacturers technical documentation if anybody knows whether Western Digital provide that. I can't find it.

Can you list all the drives make/models that return SG_IO: bad/missing sense data or [HS? ???] status. It would be really nice to have a database of drives that don't work for DCO/HPA.

Correct me if I'm wrong but I've got a feeling Seagate EXOS 10TB ST10000NM001G doesn't support HPA ??

[gorbiWTF]

Thanks for the information so far!

I have all the drives hooked up directly to the mainboard with SATA. Set to AHCI in the BIOS.

I get weird results with hdparm --dco-identify: All of three mentioned drives report the exact same as one of my "working" (reporting HS? NO) drives (ST500DM002-1BD14)

SG_IO: bad/missing sense data, sb[]: 70 00 05 00 00 00 00 0a 04 51 40 01 21 04 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
DCO Checksum verified.
DCO Revision; 0x0000 -- unknown, treating as 0002
The following features can be selectively disabled via DCO:
    Transfer modes:

    Real max sectors: 1
    ATA command/feature sets:

Running hdparm --dco-restore produces again a SG_IO: bad/missing sense data error without changing anything after a reboot. Also trying to change it by just using the current number of sectors doesn't do anything.

Could this be related to the computer hardware I'm using? I'll try a different PC later to see if the results are the same.

Yes, I would happily report all the drives I have that don't support DCO/HPA.

[PartialVolume]

anything after a reboot.

Just to clarify, not just a computer reboot but the drive needs to be powered down and up and not just go into standby either. Also I think I read somewhere Seagate have moved away from supporting HPA/DCO on the newer Enterprise drives.

I get weird results with hdparm --dco-identify: All of three mentioned drives report the exact same as one of my "working" (reporting HS? NO) drives (ST500DM002-1BD14)

I wonder if the firmware is different? Does smartctl -a show the firmware rev on these drives?

[PartialVolume]

This might be a long shot, but also consider plugging the non working drives onto the same SATA port as the one that works. Sometimes there is one SATA port that is faster than the others as it uses a different 'faster' chipset and maybe doesn't support full ATA pass through. Like I say though, that may be a real long shot.. but you never know until you try it.

[gorbiWTF]

anything after a reboot.

Just to clarify, not just a computer reboot but the drive needs to be powered down and up and not just go into standby either. Also I think I read somewhere Seagate have moved away from supporting HPA/DCO on the newer Enterprise drives.

I turned off the power supply and held the power button for a couple of seconds until all lights turned off on the mainboard. Same results. Is that enough?

I get weird results with hdparm --dco-identify: All of three mentioned drives report the exact same as one of my "working" (reporting HS? NO) drives (ST500DM002-1BD14)

I wonder if the firmware is different? Does smartctl -a show the firmware rev on these drives?

Yes, firmware version shows up, but I can't find two drives of the same model where one works and one doesn't. So I guess it doesn't matter.

This might be a long shot, but also consider plugging the non working drives onto the same SATA port as the one that works. Sometimes there is one SATA port that is faster than the others as it uses a different 'faster' chipset and maybe doesn't support full ATA pass through. Like I say though, that may be a real long shot.. but you never know until you try it.

Nothing changes, drives either work or don't, independent of the used SATA port.

Also tried it on a different PC, same results. hdparm --dco-identify always just says "Real max sectors: 1".

---

I tried running HDAT2 (v7.5) and it says there is no HPA on the drives I tested. When I unplug the power cable to the drives and plug it in again it also says there is no DCO...
Weird, but I guess there is no feasible way for me to figure out if there really is no HPA and DCO and/or if there is a bug in nwipe/HDAT2/drive firmware. So I'll just physically destroy the drives where nwipe doesn't report HS? NO. Just to be safe.