Ingenic Jz4740/Jz4732 players

[mcuelenaere]

A quick status update:
currently the code is very stable, no unknown bugs are appearing anymore (except for the button/touchscreen handling code, but this isn't a real 'bug').

USB works, but the device can't seem to get recognized by a PC (the cause is unknown, my best guess is data isn't transmitted (correctly) from the DAP to the PC).

Audio works too, it only needs some tweaking in the sample rate, clock, etc.. Currently it outputs garbage (although I got it to play some slightly distorted music some weeks/months ago).

SD card is still unsupported, I'll need to find the GPIO signal to enable power to the SD card first for that..

Nothing has changed in the FTL area (it's still not figured out), and I'm trying to get USB to work so I can easily dump the whole NAND chip to my PC for analysing.

Do you, by chance, know if the GSL-2801A is a ChinaChip DAP?

http://gslltd.manufacturer.globalsources.com/si/6008827347504/pdtl/Flash-portable/1009251153/Flash-Portable-Media-Player.htm

No I don't, and a quick look on Google reveals this isn't a very popular product... So I wouldn't guess this is an Ingenic Jz4740 DAP...

[oiseau]

Hi mcuelenaere and all !

I do have a new Onda VX747+ since a week, which I acquired knowing I could start hack it out of the box, thanks to your work and to rockbox team !

I could run the bootloader build on the device like a charm, I only had to comment the storage_init function to build it. The lcd and touchscreen are working, but only the vol_down and off buttons where working.
I could easily find out with the bootloader the GPIO config, and I made a patch adding a model in rockbox, with buttons working.

I just added this patch to the tracker, but maybe adding a new model is overkill, could it be done only with #ifdefs ?

The GPIO config is:
#define BTN_HOLD     (1 << 22)  /* on GPIO 2 */
#define BTN_MENU     (1 << 20)
#define BTN_VOL_UP   (1 << 19)

should I just update the vx747 page on the wiki ?

[mcuelenaere]

Hi mcuelenaere and all !

I do have a new Onda VX747+ since a week, which I acquired knowing I could start hack it out of the box, thanks to your work and to rockbox team !

I could run the bootloader build on the device like a charm, I only had to comment the storage_init function to build it. The lcd and touchscreen are working, but only the vol_down and off buttons where working.
I could easily find out with the bootloader the GPIO config, and I made a patch adding a model in rockbox, with buttons working.

I just added this patch to the tracker, but maybe adding a new model is overkill, could it be done only with #ifdefs ?

The GPIO config is:
#define BTN_HOLD     (1 << 22)  /* on GPIO 2 */
#define BTN_MENU     (1 << 20)
#define BTN_VOL_UP   (1 << 19)

should I just update the vx747 page on the wiki ?

Hi!

It's always nice to see some new devices get supported

I know that current source code is a bit of a mess (especially the bootloader, who in current SVN probably isn't really working real great because the most accurate version is still in my local tree), but all the files in firmware/target/mips/ingenic-jz47xx/ should work with the HW.

I also prefer using #ifdef's for at least the sadc-onda_vx747.c file.

And you may edit the OndaVX747 wiki page, I think there's no need to make a new one as there's not much difference between the 2 devices?

edit:
I've now committed your (modified) patch.

[oiseau]

Hi !

Thanks for including it this fast, I just went to sleep and as I woke up you included, like in a dream And the code duplication is gone now, nice

I'll continue to test the ingenic code now, as I understand it the current only way is to modify the bootloader code, am I right ?

I just opened an account on the wiki but am still denied editing, when it's OK I'll add gpio config for vx747+.

[mcuelenaere]

Hi !

Thanks for including it this fast, I just went to sleep and as I woke up you included, like in a dream And the code duplication is gone now, nice

I'll continue to test the ingenic code now, as I understand it the current only way is to modify the bootloader code, am I right ?

Correct.
I think the main thing we should focus on currently is USB support (which is partly working).
With it we would be able to dump the full NAND data to a PC and then we can start analyzing it to RE the FTL (Flash Translation Layer, ie YAFFS2).

Or we could try getting microSD card to work, as this can also be a backend for the storage of Rockbox (I tried this, but didn't got far).

I just opened an account on the wiki but am still denied editing, when it's OK I'll add gpio config for vx747+.

Normally, you're supposed to ask in IRC for wiki write access; but you won't need to do that anymore as I just added you to the list

[oiseau]

Good

I just updated the wiki, expliciting it works on VX747+.

[jaining]

hello

am new the forum. am interested in help porting rockbox on vx747. i have hardware and knowledge on processors. i have player with me and able to upload bootloader.
has anyone able to upload rockbox yet? the stable release code base is not fully compile the code for Normal mode. on forum i found that someone was able to play audio, but garbage. where is that code base?

thanks

jaining

[mcuelenaere]

hello

am new the forum. am interested in help porting rockbox on vx747. i have hardware and knowledge on processors. i have player with me and able to upload bootloader.
has anyone able to upload rockbox yet? the stable release code base is not fully compile the code for Normal mode. on forum i found that someone was able to play audio, but garbage. where is that code base?

thanks

jaining

Hi,

the current focus is on USB and trying to get some storage to work (I still haven't committed the USB drivers, I'll certainly do it today).

I got audio out of the device several times, the first time it was 'recognizable' but after that it was just garbage and I haven't put any energy in it (as it's not a priority).

The codebase is in SVN and in my local tree (but not for long).

There's no point compiling Rockbox (non-bootloader) as there's no storage to put it on (except for the RAM-disk, but that's not really an option).

[jaining]

Hi,

the current focus is on USB and trying to get some storage to work (I still haven't committed the USB drivers, I'll certainly do it today).

I got audio out of the device several times, the first time it was 'recognizable' but after that it was just garbage and I haven't put any energy in it (as it's not a priority).

The codebase is in SVN and in my local tree (but not for long).

There's no point compiling Rockbox (non-bootloader) as there's no storage to put it on (except for the RAM-disk, but that's not really an option).

Thank you for quick reply. I hope to use the code after u comitt to the SVN.  please reply when u do.
For storage, i look at wiki page, shows that FLT not known yet. But that is needed only if we want to run both original FW of player as well as rockbox. what if we format the flash with known FS like FAT, and run only rockbox? please tell if i miss something here.

thanks

[mcuelenaere]

Thank you for quick reply. I hope to use the code after u comitt to the SVN.  please reply when u do.
For storage, i look at wiki page, shows that FLT not known yet. But that is needed only if we want to run both original FW of player as well as rockbox. what if we format the flash with known FS like FAT, and run only rockbox? please tell if i miss something here.

thanks

Yes, that's possible but how would you know which blocks are currently marked as a bad block?
If you would ignore that, you would get in trouble pretty quickly as on every NAND chip there are several bad blocks.

When a NAND chip is shipped, the manufacturer has marked these blocks in a way like they described in the datasheet; but as we currently don't know the FTL we also don't know which blocks are bad.

Also, there will always be a need to run the OF (especially in the beginning, when not all drivers are fully developed/working) so it would be better when we use the original FTL.

If you really want to get some storage working, I would say you should look after microSD.
I did a little bit of groundwork there, but got stuck at some point.

And just 'formatting' the NAND as FAT won't work either, there's really a *need* for a FTL (bad block management, Google for more information).

[jaining]

Thank you for quick reply. I hope to use the code after u comitt to the SVN.  please reply when u do.
For storage, i look at wiki page, shows that FLT not known yet. But that is needed only if we want to run both original FW of player as well as rockbox. what if we format the flash with known FS like FAT, and run only rockbox? please tell if i miss something here.

thanks

Yes, that's possible but how would you know which blocks are currently marked as a bad block?
If you would ignore that, you would get in trouble pretty quickly as on every NAND chip there are several bad blocks.

When a NAND chip is shipped, the manufacturer has marked these blocks in a way like they described in the datasheet; but as we currently don't know the FTL we also don't know which blocks are bad.

Also, there will always be a need to run the OF (especially in the beginning, when not all drivers are fully developed/working) so it would be better when we use the original FTL.

If you really want to get some storage working, I would say you should look after microSD.
I did a little bit of groundwork there, but got stuck at some point.

And just 'formatting' the NAND as FAT won't work either, there's really a *need* for a FTL (bad block management, Google for more information).

mcuelenaere,

thanks for replying. i will look up. pardon my ignorance on flash, i will study and see what i do about microSD.

[mcuelenaere]

I committed a working USB driver (receiving data on the device is a bit flaky) and an AFAICS working NAND driver.

Now we're able to dump the NAND, but I'm not sure whether the NAND driver outputs the data in the correct order (I'll need to look further into this issue).

If you want to test the code, you'll need to edit the bootloader and you can use this patch for it (you'll also need to define USB_STORAGE in usb_core.h).

[jaining]

thanks i will check out this week

[jaining]

I committed a working USB driver (receiving data on the device is a bit flaky) and an AFAICS working NAND driver.

hi mcuelenaere, do u have app.lds file for vx747? it is missing from tree.

[mcuelenaere]

hi mcuelenaere, do u have app.lds file for vx747? it is missing from tree.

No, I haven't added that yet because there's no point in adding support for the main Rockbox firmware due to the lack of (working) storage drivers..