Issues installing Rockbox on HIFI Walker H2 v1.3

[Arrefur]

Hello! I've been trying to install rockbox on my HIFI Walker H2 v1.3. I've tried every possible installation method to no avail. Here are all the errors and issues I've been having:

Installing through jztool, I get these errors in the command prompt:
[  ERROR] cannot open file bootloader.erosq (tar error: -2)
[  ERROR] Boot failed: -2

When using the jztooljs website installation, I manage to get to the rockbox recovery menu screen, but backing up and installing result in this error:
NAND open error (5)

Lastly when installing through the update files, the players system reboots, but never launches rockbox.

Any help would be much appreciated! <3

[dconrad]

When you use the update files method, do you get a Success message? You have the special version of the HifiWalker which needs the unique update file, https://download.rockbox.org/bootloader/aigo/native/erosqnative-hw1hw2-eros_h2.upt.

If you could load up the bootloader with jztool, could you send the results of the bootloader flash info options "Flash Info" and "Flash ONFI Info", both under Debug tools? It almost sounds like your player has a flash chip that the bootloader doesn't know about.

[Arrefur]

When I update files, I do not get a success message, the device simply reboots, with no changes.

I can get the flash_info without issues, but when getting the flash_onfi_info file, I get the "NAND open failed" prompt, and it dumps an empty txt file.
I have attached both to this message.
the flash_info.txt:

readID opcode  = FF E5 71 E5
readID address = E5 71 E5 71
readID dummy   = E5 71 E5 71
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 6E 02 02
sfc params3  = A4 2F 00 00

[dconrad]

That NAND chip appears to be one of the ones in the list, but I'll try to look into this when I get a chance - I happen to have a V1.3 device at home to compare against. Hopefully it has the same chip installed.

Running the original firmware's updater should give either a "Success v_v" or "failed v_v" message, so it sounds like you're not even getting to the updater, which is weird. I'll have a think about it.

[dconrad]

Just an update: My device has a different chip (ato25d1ga) from your (apparent) DS35X1GAXXX. The ds35x1gaxxx is apparently identical for our purposes to the gd5f1gq4xexx. I'm not sure if that's not true and our definition is wrong, or it's not identifying it correctly, or if something is physically wrong. Onwards and downwards.

[dconrad]

More information: Here's my nand info on my v1.3:

Code: [Select]

readID opcode  = FF 9B 12 9B
readID address = 9B 12 9B 12
readID dummy   = 9B 12 9B 12
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 5E 02 02
sfc params3  = A8 24 00 00

I notice there's 3 bytes different in the sfc information...

And from my v1.6 unit:

Code: [Select]

readID opcode  = FF EF AA 21
readID address = EF AA 21 00
readID dummy   = EF AA 21 00
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 5E 02 02
sfc params3  = A8 24 00 00


[dconrad]

Ok, I would like to run a sanity check. Can you please use jztool to temporarily load this bootloader and then attempt to dump the flash info and the flash ONFI info?

https://drive.google.com/file/d/1wkVcvOU_fufoXPuUNrhI6dPEuuxGFrWv/view?usp=sharing

Edit:

Here is also a shot-in-the-dark change to add more delay between resetting the nand chip and trying to identify it. Can you please do the same procedure with this as well?

https://drive.google.com/file/d/1bHqN6USYKDhLneUl6dXAEskuEjzmSi69/view?usp=sharing

[Arrefur]

I have ran the sanity_check_bootloader and got the following flash_info and again got the "NAND open failed" prompt, with an empty flash_onfi_info file.

Code: [Select]

readID opcode  = FF E5 71 E5
readID address = E5 71 E5 71
readID dummy   = E5 71 E5 71
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 6E 02 02
sfc params3  = A4 2F 00 00

Same thing with the long_reset_delay_bootloader, got the following flash_info and the exact same problem with the flash_onfi_info file.

Code: [Select]

readID opcode  = FF E5 71 E5
readID address = E5 71 E5 71
readID dummy   = E5 71 E5 71
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 6E 02 02
sfc params3  = A4 2F 00 00


[dconrad]

Ok, no big surprise there but I had to try it. It appears that we can successfully read the flash when dumping the info, but we fail to correctly ID the chip for some reason. There are a couple of discrepancies between the two, maybe they are significant.

Let's do the same test with these two files:

second_read_bootloader.erosq: https://drive.google.com/file/d/1a5lfRcoPRjX4OtIImpP5uOjQCPLk64SP/view?usp=sharing
tsh0_bootloader.erosq: https://drive.google.com/file/d/1xbKolnWmISUpHl-XZVESQnx4_W-z3OuR/view?usp=sharing

[Arrefur]

The flash_info files are the exact same, and the exact same problem with the flash_onfi_info file, with the "NAND open failed" prompt.

[dconrad]

Hmm, ok, let's try another - https://drive.google.com/file/d/13E60WlTTVf5s43Dt69YyeeA2GL_dBzVa/view?usp=sharing

This time I just copy/pasted parts from the identify function into the flash_info function, so let's see if our flash_info file is any different. No need to run the flash_onfi_info.

The mystery continues.

[Arrefur]

The flash_info file seems to be the same.

Code: [Select]

readID opcode  = FF E5 71 E5
readID address = E5 71 E5 71
readID dummy   = E5 71 E5 71
sfc params0  = 06 05 04 03
sfc params1  = 02 55 AA 55
sfc params2  = 00 6E 02 02
sfc params3  = A4 2F 00 00
The mystery that keeps on giving.

[dconrad]

Hold up, I commented out the part where it tries to read the readID via the opcode address. I think you might not be trying the new file when you download it. Can you double check your command you're running (each file has a different filename)?

[Arrefur]

I've been replacing the bootloader files each time, they all have different names (eg. sanity_check_bootloader.erosq; long_reset_delay_bootloader.erosq; second_read_bootloader.erosq; tsh0_bootloader.erosq; read_info_with_identify_bootloader.erosq). Did I misunderstand on how to run them?

[dconrad]

What jztool command are you running? It will need to change to match the new filename each time. Are you putting these test files on the SD card, or in the same place as where jztool is?

The wiki might not be clear on this - when you run something like "jztool erosq load bootloader.erosq", it's looking for the file "bootloader.erosq" in the same folder as jztool and sending that over USB. So you'll need the new files in the same folder and change that filename to send the desired one.

EDIT:

Wait. Are you using the jztooljs webpage? If so, that would explain it. It is loading a very old bootloader over USB, about 2 months before the definition for your NAND chip was added.

It's hard to explain how the system works unfortunately - this is why we're trying to get away from jztool. The jztool program only sends a file over USB from the computer to the player to run temporarily. Then the temporarily running bootloader reads a file from the SD card and writes it to flash. It's very confusing if you're not used to it.

If all this is true, please try the official release bootloader by putting jztool and the official bootloader bootloader.erosq file in the same folder and running the correct command. Then do the flash info and flash ONFI info and see if it works.

If that works, I would like to understand why the update file doesn't work, because I have a suspicion we may also have a miscommunication on that too, it may work fine.