Surfans F20 "NAND open error"

[SelectOkay]

I'm new to the forums and Rockbox, but I've installed it on a Sandisk Clip Zip without much issue. The problem is  I bought a Surfans F20 and I can't install Rockbox on it.
I've tried using the Rockbox Utility but after using the firmware from the company itself, the utility constantly crashes.
SO I tried to use the native port and got farther, but now it's telling me "NAND open error (5)" I tried using two different mircoSD cards and even changing the partition from GPT to MBR but it's not working. Does anyone have any advice or help? I'm kinda stumped here.

[SelectOkay]

Update: I manually installed the bootloader and that fixed all my problems. I guess I'm newer to all this than I thought. Whoops!

[amachronic]

I'm not sure if the native port has been tested on the Surfans F20. The NAND open error is a bit odd. Usually this would mean your flash chip is not supported, but since it worked eventually... meh. It's good you got it working, though!

[dconrad]

Nevermind, ignore

[kaas]

Update: I manually installed the bootloader and that fixed all my problems. I guess I'm newer to all this than I thought. Whoops!

Did you manage to install the native port, or are you running the hosted port now? If you installed the native port, I'd love to hear how you did it, as I'm having the exact same issue.

[amachronic]

kaas - The F20 may have a different flash chip that isn't recognized. I'm happy to assist, but we need to figure out is what flash chip is used. If you are willing and able to take apart the player, that's the fastest way to find out. The other option is probing the flash in software, but that'll require a modified bootloader.

[kaas]

kaas - The F20 may have a different flash chip that isn't recognized. I'm happy to assist, but we need to figure out is what flash chip is used. If you are willing and able to take apart the player, that's the fastest way to find out. The other option is probing the flash in software, but that'll require a modified bootloader.

Thanks very much, I would prefer to go the software route as I'm not very comfortable taking it apart (don't even see any screws so it would probably involve at least some amount of breakage..). I assume in that case you can send me the required bootloader, and I'll run it with jztool?

Edit: I've since resolved my issue with the hosted port (see this bug), which was why I wanted to try the native port in the first place. So from my end, there's no particular hurry to get the native port working on my specific player. I'd still be glad to assist, though.

[amachronic]

kaas - The F20 may have a different flash chip that isn't recognized. I'm happy to assist, but we need to figure out is what flash chip is used. If you are willing and able to take apart the player, that's the fastest way to find out. The other option is probing the flash in software, but that'll require a modified bootloader.

Thanks very much, I would prefer to go the software route as I'm not very comfortable taking it apart (don't even see any screws so it would probably involve at least some amount of breakage..). I assume in that case you can send me the required bootloader, and I'll run it with jztool?

Edit: I've since resolved my issue with the hosted port (see this bug), which was why I wanted to try the native port in the first place. So from my end, there's no particular hurry to get the native port working on my specific player. I'd still be glad to assist, though.

Okay, I added a flash info command to the bootloader, download link here: https://drive.google.com/file/d/1BgvDvGBKBSnWTDBHfE_uii3fjpSARqxt/view
(which is patch g#4566)

If you load that with jztool and select Debug tools > Flash info from the menu, it will dump some info to a file called flash_info.txt on the SD card. Copy & paste the file contents here. It'll show the ID bytes and some additional info - whether the chip is NAND or NOR and the page and block sizes. The probing method is basically the same as what the CPU does on boot to identify the flash, and it's read-only, so it should be safe. I also tested it on my players as a sanity check and it's working as expected.

[kaas]

Here you go

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 63 02 02
sfc params3  = A4 2F 00 00

[amachronic]

Thanks, it appears to be a Winbond W25N01GV based on the ID bytes.

Here's an updated bootloader which supports that chip: https://drive.google.com/file/d/1zj8gq1G8DK-9581uVWq-9FmKXLlPuw5w/view?usp=sharing

This one has a "Flash ONFI info" option which you can run, paste the flash_onfi_info.txt file here. If the info matches up with the datasheet (8.2.27 Parameter Page Data Definitions) then we can be pretty confident the chip is correct. If you also take a normal bootloader backup and upload it somewhere, I can check that out to make sure it looks sane.

[kaas]

signature = 2222AAAA
revision = AAAA
manufacturer = "ÿÿÿÿÿÿÿÿÿÿÿÿ"
device model = "ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ"
JEDEC mf. id = FF
data bytes per page = 4294967295
spare bytes per page = 65535
pages per block = 4294967295
blocks per lun = 4294967295
number of luns = 255
bits per cell = 255
max bad blocks = 65535
block endurance = 65535
programs per page = 255
page program time = 65535
block erase time = 65535
page read time = 65535

And here is the bootloader backup. Although using the bootloader went fine I did not go ahead flashing it, since the datasheet values don't seem to match up (e.g. spare bytes per page is 4x higher in the txt than in the datasheet), I may be misinterpreting it however.

[amachronic]

It's not reading data off the chip properly, the data is mostly 0xff, so yeah flashing it would be a really bad idea. Not sure what's wrong. I'll try reducing the clock speed & disable quad I/O, and double check for any mistakes I might've made.

[amachronic]

Here's another build, try dumping the ONFI info and take a bootloader backup again. I wouldn't advise flashing it.
https://drive.google.com/file/d/1IJjG_zZxdSxNZK3XM-nAeaJzVnR2oHcK/view?usp=sharing

[kaas]

probed chip ID = EF AA 21
signature = FFFFFFFF
revision = FFFF
manufacturer = "ÿÿÿÿÿÿÿÿÿÿÿÿ"
device model = "ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ"
JEDEC mf. id = FF
data bytes per page = 4294967295
spare bytes per page = 65535
pages per block = 4294967295
blocks per lun = 4294967295
number of luns = 255
bits per cell = 255
max bad blocks = 65535
block endurance = 65535
programs per page = 255
page program time = 65535
block erase time = 65535
page read time = 65535

Bootloader backup. Thanks again for your continued efforts.

[amachronic]

That last dump was almost good but shifted by one byte. This build should fix it.
https://drive.google.com/file/d/1LQtakDKCNTGvXU-u9QVYc056nN78CkDl/view?usp=sharing

Try another bootloader backup. No need to bother with the ONFI info this time.

Also, could you unzip a daily build to your SD and run rockbox directly from the bootloader? You don't need to install the bootloader, just select the Rockbox boot option while you're in the bootloader. EDIT: I forgot to mention, this is so you can check if rockbox actually works. If they changed the flash there's no guarantee the rest of the hardware is the same as older units.