Cowon D2

[Bkd11]

I just tried the latest build (r17774) and I keep getting an error "undefined instruction at 100C760" and the player freezes when selecting a .mp3 file. The same .mp3 file played fine with (r17755).

I did notice the hold key is working now.

Thanks for all your work shotofadds! Now get some sleep!!

[shotofadds]

I just tried the latest build (r17774) and I keep getting an error "undefined instruction at 100C760" and the player freezes when selecting a .mp3 file. The same .mp3 file played fine with (r17755).

100xxxxx is IRAM, which I disabled in the lastest build. Have you done a full make clean / make / make zip?

EDIT: I had to make a further change to fully disable IRAM in plugins/codecs. This should fix the crash and also give a further speed boost.

Another edit: Wow. I've just done a few test_codec runs (see CodecPerformanceComparison) and it seems the 192MHz ARM926EJ-S in the D2 is nearly as fast in tests as the 300MHz ARM920 in the Gigabeat F. I guess the v5 architecture really does make a good bit of difference.

[ThibG]

Yeah! Sound is working fine, except it skip milliseconds from time to time. (when playing music, at least)

[shotofadds]

Yeah! Sound is working fine, except it skip milliseconds from time to time. (when playing music, at least)

Are you sure you're using the very latest version? I don't hear any skips/clicks/pops/etc with the current revision..

If you're still seeing this problem with the latest revision (r17779 or later), can you describe the problem in a bit more detail?

[ThibG]

I'm using the latest svn revision, and it sometimes skip milliseconds.
It doesn't sound like the "clicks/pops" there were before, and it happens less often (I don't know how often, but less than one in 20s).
I was playing ogg (q6?) on a 4GB D2.

Edit: I noticed 3 of these "skips" in 11min of OGG/Vorbis playback

[shotofadds]

I've been listening to a collection of various vorbis files (gaplessly ) for the last 30mins or so and haven't noticed any skips.

A possible explanation is that currently we don't implement error-correction when reading from the flash drive, so the occasional data bit may be read incorrectly leading to errors in decoding.

I wouldn't recommend Rockbox for serious music listening at this stage - with the current NAND driver it's also possible that completely the wrong data could be read instead of the intended file. That can lead to a complete crash, which I've seen once or twice now.

@cybergrind: yeah, the Rockboy keymap needs some work. I'll take your ideas on board, thanks.

[Bkd11]

EDIT: I had to make a further change to fully disable IRAM in plugins/codecs. This should fix the crash and also give a further speed boost.

After updating to (r17781) my .mp3s now play flawlessly and gaplessly (is that a word). It did crash while testing a .flac file though. Tried it again and it crashed at the same spot. Tried another .flac file and it also crashed about 45 seconds in.

[shotofadds]

Read the last paragraph of my previous post...

[dasimmet]

Hi

i upgraded my d2 on r16807 and it was stuck on startup
i turned it off and tried it again, the same thing happend and now it cant be turned off any more 
thats what the console shows:

Code: [Select]

Rockbox boot loader
Version r16807:17603M-080522
ATA
mount
Lengt: 74478
checksum: 2EA7479
Model name: d2
Loading rockbox.iaudio
Sum: 2EC7660
Bad checksum

i dont know what to do, hopefully after some time it will run out of battery...

[andiator]

i upgraded my d2 on r16807 and it was stuck on startup

It's somewhat old, try newer one  .
You can reset your D2 - close to the usb is a little hole - below is the reset button, you can use the touchpen.
Or you hold the power off button in on/off position for some seconds, till it goes off

[dasimmet]

thanks for help
holding the power swich was what i tried firt time...
i dont know how this old version could get on my d2, i downloaded it from the info page.
now its running fine

[Dicky Dave]

Hi Guys,

I just tried installing Andiator's pre-built patched binary for v3.57 (rev17784) and I am getting a similar error to that reported by talex and couple of weeks ago:

Code: [Select]

*PANIC*
BMPM total_banks m
ismatch

However, this is on a 4GB D2 not a 16GB. I had my 16GB SDHC card in at first and thought this might be the problem. So I tried again with it removed but kept getting the same problem. Any ideas?

[shotofadds]

Interesting. Could you try uploading this binary with tcctool? It's one that I built for debugging the 16Gb support, so it should show some diagnostic information when you run it.

Let me know what shows up...

[Dicky Dave]

Hi shotofadds,

Here is the output:

Code: [Select]

Bank 0: Manuf, 0xEC Id 0xD5
Bank 1: Manuf, 0xEC Id 0xD5
Bank 2: Manuf, 0xEC Id 0xD5
Bank 3: Manuf, 0xEC Id 0xD5
Banks 0,1 have same manuf. id
Banks 0,2 have same manuf. id
Banks 0,2 have same UID
Total banks: 2
BMP tag: 0x42 0x4D 0x50 0xFF 0xFF

The other thing that might be worth noting is that my 4GB D2 is one of the earliest revisions (I think it is 1.0, but the revision is on the box and I cannot find it at the moment). It was bought in Feb 07.

[shotofadds]

Code: [Select]

BMP tag: 0x42 0x4D 0x50 0xFF 0xFF

The other thing that might be worth noting is that my 4GB D2 is one of the earliest revisions (I think it is 1.0, but the revision is on the box and I cannot find it at the moment). It was bought in Feb 07.

Ah, I've seen that before on a 2Gb D2, running firmware 2.20. On most D2's the first 0xFF shows up as 0x4D (the ASCII code for 'M') and the second is the number of NAND banks present. It's easy to fix - I'll to update the code to ignore those last two values if they're 0xFF.

I guess it's dependant on which firmware version originally formatted the drive - newer versions fill in those two values, older ones don't.

EDIT: This version should work for you, it's v3.57 patched with a modified bootloader.