Restore to factory default from diag menu (without using serial)

[datanoise]

Hi

I've access to a kindle that was stuck on the tree. Following knc1's advice on another thread I let it on the power for a couple of days, it ended up in diag mode. I have a hard time finding information about the diag mode and what can be done with it (maybe I missed a dedicated topic).

I suspect a corruption of the userspace filesystem. As I understand it the kindle now did a fresh format of this one.

What is the best way reset to factory defaults (or to a new/clean image I'd download from amazon servers) from the diag mode, without having to connect the serial?

edit: uploaded a new image on the filesystem, got out of diag mode it installed the upgrade, could now do a factory reset from the params. Problem solved. Would be curious to know what the best way is for cases where the kindle would have a broken system, unable to reboot on it.

for the charger + wait tip, worked great!

[hius07]

report here, please

[datanoise]

Seems like the system partition wasn't broken but user partition had a problem (the kindle wouldn't index new books before, partition was often mounted read only by my friend's OS because of errors...). He tried to reboot hoping it would index some books. The kindle got stuck on the tree of death, with lighting on. Sticking on a wall charger for a couple of day made it reboot into diag mode.

It was then possible to access the user storage (wiped clean by the kindle when it was stuck on the tree: that was not a boot loop, it was actually partitioning the buggy user partition) and an upgrade (usb mode / latest .bin from amazon / exit diag mode) was enough to get the device to reboot in a clean / stable mode.

Now I'm not sure what would be the best way to check if the nand is in good condition or not...

[knc1]

Originally Posted by datanoise

- - - - -
Now I'm not sure what would be the best way to check if the nand is in good condition or not...

I posted a utility that will read and report on the internal device records.
(Some nand flash has a table similar to a hard disk's SMART table).

It is an attachment to one of my posts, somewhere.

= = = =

Don't be concerned -
Since fsck has completed, the bad blocks have been replaced with spares.

It takes so long (longer than a battery would last - which is why the wall charger) because the way fsck works on a nand device, it has to start over every time a bad block gets remapped.

It will be fun when the people with 32Gbyte flash start having bad blocks.
What will that take to "fix&clean"? Weeks?

- - - -

With access to a Linux system, there is a way you can force a re-format of the user storage on the next boot.
It is also posted here somewhere (start with all the old, indexed, de-bricking threads to find it).

[datanoise]

Ok, good to know.