Cowon D2

[ChristianGmeiner]

Yes I can run code on the target.. a nice way to develop

This is what I get when starting:

ATA
2 banks
* 1048576 pages
* 2048 bytes per page
707 sequential segments found (707Mb)
KEYPRESS
mount
Hello World!
...

[shotofadds]

I'm not convinced that the power-off code is working as intended.

As I understand it, the PCF50606 automatically powers the unit off when a) the power key has been pressed for >1sec and b) its ONKEY1S interrupt has gone unacknowledged for 8 seconds. I think that is the behaviour currently being seen (hence the delay).

Yeeze: I got your PM, thanks for the screen caps. I'm currently digesting the information in there, should be useful. Now I need sleep

[phi]

Hi all,

I just submitted a patch to enable ECC in the NAND driver. The error rate on my D2 seems to be fairly low, but I figured we'd need this eventually so here it is.

FS#8669

You won't actually notice anything different in the bootloader unless you've written code to test file reading.

[shotofadds]

Much appreciated, that was on the to-do list.

I'm currently in the process of re-working the NAND driver to make use of the write cache (block type 0x15) segments. When that is done we should have a reliable read-only filesystem and the 'real' work can begin to boot the main Rockbox image. 8-)

/me kicks himself to update the TelechipsNAND wiki page...

[shotofadds]

There's not much progress to report this week I'm afraid (I've been spending too long killing off my brain cells with beer ) but I thought those following this thread might be interested in a quick photo:



This is what I failed to show at 'DevConPub' this week: it shows the D2 booting as far as the Rockbox logo and version number, and then freezing soon afterwards.  But it's good progress all the same... 8-)

[Michael66]

Thanks for the info, shotofadds! I'm following this thread for a while now and really appreciate all the work you guys put into making Rockbox work on the D2.
The official firmware releases seem to degenerate by the minute (according to the discussions at iaudiophile.net), so what you do here is probably the only hope left for the D2 community.

[cybergrind]

I'v compiled file with "B" option, copied .bin to device. load:
see :

Code: [Select]

ATA
1 banks
....

when I pressed any-key:

Code: [Select]

*PANIC*
Stkov scroll

I try to comment

Code: [Select]

    if(((unsigned int *)old->stack)[0] != DEADBEEF)
        thread_stkov(old); 

in thread.c, to view pre-panic messages, and have:

Code: [Select]

mount
No partition found

2 Gb version.

+P.S. when I removed check in telechips.c:

Code: [Select]

//    if (rc<=0)
//    {
//        error(EDISK,rc);
//    }

All seems work... what i doing wrong?

[shotofadds]

I don't think you're doing anything wrong. Yeeze reported a similar problem on his 2Gb player, but I've not yet had time to investigate the problem. I will look into it, but I may need you to run some tests for me.

In other news, this afternoon I got the Rockbox menu to boot, but nothing is working yet

[andiator]

Added button definitions for plugins, wps and settings menu. The simulator is now usable.
FS#8708 http://www.rockbox.org/tracker/task/8708

@shotofadds:  Coooool

[shotofadds]

Thanks for that, andiator - I'll get those merged in soon. The keymaps will probably have to evolve over time, but it's certainly a good starting point.

Your real name (for the credits) is listed on the tracker, right?

[andiator]

@shotofadds
Yes it's me.
For the WPS and menu are the four direction buttons (up/down...) enough. Some plugins need more. And yes, the mappings have to be altered to be usefull.

[cybergrind]

I tryed add some printf in disk.c and fat.c, and get log:

Code: [Select]

mount
start mounting disks. fat_init. NUM_VOLUMES : 1

start disk_mount(0)

start fat_mount()
fat_mount() - BPB is not sane

start fat_mount()
fat_mount() - Couldn't read BPB (error code -1)

start fat_mount()
fat_mount() - BPB is not sane

start fat_mount()
This is not FAT32. Go away!

No partition foudn, trying to mount sector 0.
start fat_mount()

after this I comment, near 360 line in fat.c:

Code: [Select]

#else /* #ifdef HAVE_FAT16SUPPORT */
//        DEBUGF("This is not FAT32. Go away!\n");
//        printf("This is not FAT32. Go away!");
//        return -2;

above this fragment we have such code:

Code: [Select]

#ifdef HAVE_FAT16SUPPORT
        fat_bpb->is_fat16 = true;
        if (fat_bpb->dataclusters < 4085)
        { /* FAT12 */
            DEBUGF("This is FAT12. Go away!\n");
            printf("This is FAT12. Go away!");
            return -2;
        }

maybe some defines is missed (2 GB model)?

[shotofadds]

The first 4 "start fat_mount()" lines are because Rockbox attempts to mount each partition in the partition table in turn - but the D2 has no valid partition table, since it is formatted in "superfloppy" mode. The code in disk.c and fat.c is fine, don't worry about this.

The last line "No partition found, trying to mount sector 0. start fat_mount()" is the important one - what happens after this?

Are you familiar with the linux 'dd' utility (or it's Windows equivalent)? A dump of the first 512 bytes from your D2 would be useful at this point.

[cybergrind]

Oh I'm sorry, not all log, last message

Code: [Select]

This is not FAT32. Go away!

When I added in config-cowond2.h:

Code: [Select]

#define HAVE_FAT16SUPPORT

All seems ok. Maybe this trouble appeared only in 2 Gb version, because defaulf system is fat16 =)

dump (attached) shows that filesystem is Fat16.

Another problem: something strange with screen during shutdown,
maybe we should add

Code: [Select]

#define HAVE_LCD_SHUTDOWN

into config-cowond2.h

and rename:

Code: [Select]

static void lcd_display_off(void)

to:

Code: [Select]

void lcd_shutdown(void)

in lcd-cowond2.c

and use it during shutdown.

P.S. I didn't familiar with ROCKBOX rules, maybe I should create and commit patches for this issues?

[shotofadds]

I thought that might be the case: the missing log message was the important one . I'll add in HAVE_FAT16SUPPORT next time I update the code.

Shutdown/poweroff does not currently work - this requires some more thought (more than just the LCD driver - powering off the CPU & hardware needs to be done too).

If you make patches, it's best to submit them to the tracker (Flyspray). But if they're just 1-liners like the above, post them here.