SAMSUNG YH-920

[wolly-p]

Bootloader is attached

[pyro_maniac]

Bootloader is attached

Ok thanks, but how can i get this onto the device? Copy that into the System directory?

[wolly-p]

It's already flashed into the ROM; at present, I don't see a need to change it. If you really want to do that, you would first have to debug it, next to modify it, finally to flash it back to the ROM, but the flashing method is hardware-dependent. Copying will not work.

However, I do not understand the way you want to proceed. Like other PP-based players, the (fixed) bootloader does the necessary steps to initialise the device (e.g. CPU, ATA-interface and a bit more),  and then reads its main firmware from the encrypted mi4-file. This mi4 file is what you have to build completely new, based partly on the available Rockbox sources, and partly with new modules you have to write on your own, dependent on the YH-920 hardware. In order to understand how the hardware works, it will indeed be helpful to debug the bootloader.

[pyro_maniac]

I don't know if you have watched into the bootloader but there are some interesting strings in it. I attachted what i have found. Maybe you will take a look and tell me, what you think.

[wolly-p]

Looks like strings of diagnostic tests / error messages. But I don't see how it could be utilised for your intended Rockbox build. BTW, about 2 years ago, I sent various emails to the Samsung guy whose e-mail address can be found in the strings, and the address was existing, but I never got any reply. Seems to be company policy ...

[pyro_maniac]

I was wonderung how much this bootloader can do and i thougt it functions could be helpful for developing.

[wolly-p]

Not to get me wrong: for sure it would be helpful to disassemble / debug the bootloader, because it could serve as the valuable source for hardware addresses, initialisations of hardware components etc. But as mentioned before, I expect proper disassembling / debugging the firmware (i.e. mi4-file) to deliver quicker results. Just my guess, since I haven't done it myself. You could browse the other build forums, this could give you an indication how to proceed.

[pyro_maniac]

I played around with the h10 build again and i found some interessting. the screen is still empty after booting but i connected the device to my windows mashine and that noticed a usb device with the name "Rockbox media player". gevaerts gave me the hint with logf() so i will try this tomorrow.

I hope i can sleep tonight! :-D

[wolly-p]

That's really interesting. Maybe logf() can also give you hints concerning the display (just a guess).

[pyro_maniac]

So now there is a serial connection to the device. But I have to stop here and need to learn what i can do with it and how it works.

I am open for any hints and advices.

[wolly-p]

Can you attach the results of logf (logf.log, logf.txt or whatever it's called) here?

[pyro_maniac]

Until now there is nothing that came from the device. Thats why i have to leran more about it and i think i have to build in more debug. I try to find some time on the weekend.

/EDIT: The only thing i got was: "Connected,"

/EDIT: I recived a "test" out of the connection from the usb_serial_transfer_complete function. But there seems to be still some issues on in the serial connection. i can't read out the logf buffer, where i wrote my init logs into.

[wolly-p]

You still don't have any other log or debug info, i.e. files on the harddisk? It not, it could be worthwile disassembling the bootloader in order to understand why the current ATA routine does not seem to work. Maybe Rockbox stumbles over something which causes it to stop before writing something to the HDD.

[pyro_maniac]

You still don't have any other log or debug info, i.e. files on the harddisk? It not, it could be worthwile disassembling the bootloader in order to understand why the current ATA routine does not seem to work. Maybe Rockbox stumbles over something which causes it to stop before writing something to the HDD.

I don't write to the hdd. I try to recive the data of logf() over the serial usb connection.

I didn't understand the whole functionality yet. I recived a test message via usb serial connection but i wasn't able to read out the logf buffer. To disassemble the bootloader would be a bigger thing to me than to understand and get to work the usb serial thing.

/EDIT: now i have redone my work and i got a connection again. as gevaerts adviced me, i am trying to rearange the init of the components in apps/main.c and to let them wait for a serial usb connection.

/EDIT: i manualy red out old logf entrys but they are very cryptic and to rearange the init seems to be very tricky. i got very often a system freeze on connecting.

/EDIT: i am actually testing the patch FS#9253 to write logf() results into a file. the file was created but nothing in yet. i am working on it. And with the new usb stack i even have access to the hdd after booting rockbox. :-)

[pyro_maniac]

I need some help in testing my patch in FS#9253 because i have no device rockbox running on.