ChromeOS Recovery [ Chromebook Pixel -- LINK ]

[Mario]

Hi!

I stuck during recovery of chromeos on 5 chromebook pixels. Those units where equipped with an r/w legacy boot stub to start linux for ssd secure erase. After this step the original firmware was restored with the mrchromebox script [ FULL RESTORE ]. The recovery runs without any errors, but after the first reboot/shutdown i got on all 5 units "chrome os missing or damaged" screen. Tried it several times in dev mode and also trusted mode. Every time the same issue, first shutdown, chrome os damaged.

[Bernie Thompson]

If you hit tab at the 'os missing or damaged screen' does it give a failure reason in the upper left corner?

On Thu, Nov 30, 2017 at 4:47 AM, Mario <testa...@gmail.com> wrote:

Hi!

I stuck during recovery of chromeos on 5 chromebook pixels. Those units where equipped with an r/w legacy boot stub to start linux for ssd secure erase. After this step the original firmware was restored with the mrchromebox script [ FULL RESTORE ]. The recovery runs without any errors, but after the first reboot/shutdown i got on all 5 units "chrome os missing or damaged" screen. Tried it several times in dev mode and also trusted mode. Every time the same issue, first shutdown, chrome os damaged.

--
--
Chromium OS Developers mailing list: chromiu...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-os-dev?hl=en

[Mario]

Yes, please see the screenshot attached. No bootable disk found after successfull restore and first boot?!

[Mario]

It is possible to get the recovery log in any way during recovery? I have read something about a defect FPROM.

--> https://bugs.chromium.org/p/chromium/issues/detail?id=468977

[Vincent Palatin]

On Fri, Dec 1, 2017 at 9:36 AM, Mario <testa...@gmail.com> wrote:

It is possible to get the recovery log in any way during recovery?

The recovery process is writing a log  back to the USB stick. 

After the recovery process, you can plug the USB stick in another Linux machine, it has a 'recovery.log' file (and hardware_diagnostics.log, dmesg too) in the partition 1 (labeled as 'STATE').

During the recovery process, you can also switch to VT2/VT3 (ctrl+alt+-> or refresh), it displays more info there.

 

I have read something about a defect FPROM.

--> https://bugs.chromium.org/p/chromium/issues/detail?id=468977

Something might have definitely failed in your hardware.

The whole 'FPROM' theory is far-fetched though (and to be honest there is no even a 'FPROM', all firmwares are stored on flash memories) 

On those machines, the SSD has been the most likely point of failure.

[Mario]

Hi!

Thank you for this info. I have attached the recovery log, looks ok for me. Maybe a hardware problem in that case :(

[Mario]

Is it possible to deactivate the soldered emmc and equipp this one with a msata? There is a msata slot.

[Vincent Palatin]

On Fri, Dec 1, 2017 at 12:02 PM, Mario <testa...@gmail.com> wrote:

Is it possible to deactivate the soldered emmc and equipp this one with a msata? There is a msata slot.

As per your kernel traces, what is soldered on the mainboard is a SATA SSD not an eMMC.

The miniPCIe slot is likely not wired as mSATA, I imagine you are referring to the modem one, It has USB2.0 connected not SATA

(maybe PCIe too but I don't think so).

[Gwendal Grignou]

On Fri, Dec 1, 2017 at 3:12 AM, Vincent Palatin <vpal...@chromium.org> wrote:

On Fri, Dec 1, 2017 at 12:02 PM, Mario <testa...@gmail.com> wrote:

Is it possible to deactivate the soldered emmc and equipp this one with a msata? There is a msata slot.

As per your kernel traces, what is soldered on the mainboard is a SATA SSD not an eMMC.

The miniPCIe slot is likely not wired as mSATA, I imagine you are referring to the modem one, It has USB2.0 connected not SATA

(maybe PCIe too but I don't think so).

The storage device is a Sata based Sandisk i100, the slot is a M.2 E keyed connector (USB, UART, PCIe, not sata).

From the log, the recovery went well, but coreboot is unable to initialize the SSD.

You mentioned that you upgraded the firmware to perform a secure erase. Did you just write 0s or did you sanitize the drive with a ATA Secure Erase command set?

In dmesg, I see 

[    1.401030] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

[    1.402046]  sda: unknown partition table

[    1.402483] sd 0:0:0:0: [sda] Attached SCSI disk

I assume it was your first try to recover chromeos on that machine after the SSD was fully erased.

I am interested to know if the SSD after secure erase was put in a specific state coreboot is unable to handle. 

If you boot a test image on a USB stick, append /var/log/storage_info.txt, A coreboot log [/sys/firmware/log] and check if install has properly wrote the partitions [with 'cgpt show /dev/sda' for instance.]

Thanks,

Gwendal.

[Vincent Palatin]

On Fri, Dec 1, 2017 at 6:51 PM, Gwendal Grignou <gwe...@chromium.org> wrote:

On Fri, Dec 1, 2017 at 3:12 AM, Vincent Palatin <vpal...@chromium.org> wrote:

On Fri, Dec 1, 2017 at 12:02 PM, Mario <testa...@gmail.com> wrote:

Is it possible to deactivate the soldered emmc and equipp this one with a msata? There is a msata slot.

As per your kernel traces, what is soldered on the mainboard is a SATA SSD not an eMMC.

The miniPCIe slot is likely not wired as mSATA, I imagine you are referring to the modem one, It has USB2.0 connected not SATA

(maybe PCIe too but I don't think so).

The storage device is a Sata based Sandisk i100, 

the slot is a M.2 E keyed connector (USB, UART, PCIe, not sata).

No, this is Samus ?

I have checked against the schematics, on Link, the full size slot is mini-PCIe with only USB2.0 lanes connected.

[Mario]

Hi Gwendeal!

The ssd was sanitized with secure erase command trough blancco wiping software. This was done on all 5 units. I have booted a linux and the ssd has 0 bad blocks but no parition table.
I have tried to use coreboot with tiano, but same thing after i have installed xubuntu. Ssd is shown in the installation menu, but not in the tiano bios itself. Strange!

I am interested to know if the SSD after secure erase was put in a specific state coreboot is unable to handle.

If
 you boot a test image on a USB stick, append /var/log/storage_info.txt,
 A coreboot log [/sys/firmware/log] and check if install has properly
wrote the partitions [with 'cgpt show /dev/sda' for instance.]

Can i do this in any linux?