SanDisk Sansa c200v2, m200v4, clipv1, clipv2, clip+, and fuzev2

[Xanikseo]

Wow quite a successful day !

Can confirm that the hold switch is instantly detected. However (yes, unfortunately there is a but) sometimes the backlight flashes on and off quickly for about a second or two before the backlight finally turns off. It doesn't always happen, but I think it mainly happens during heavy cpu load. If you boost cpu through debug menu, it is easily reproducible. Unfortunately another case of extra delays for the boost ?..

EDIT: WTF? Sometimes backlight randomly flashes while in hold??? (After unboosting cpu). Ok this is weird, it seems that when the cpu is boosted, it randomly makes the backlight flash (but only when in hold I think)...

[bilditup1]

Hey all-
Just now, I crashed after selecting an mp3, with the following error:

data abort
at 30027024
FSR 0x8
(domain 0, fault
address 0xBB0F879F

I'm not sure if this was already addressed; if so, sorry. This is from the latest nightly (25733).
-e

UPDATE: Another thing- I'm not sure if this is standard behavior or not, but after shutting down from rockbox, if I connect a USB chord in the Fuze, the original firmware is loaded up and a connection is established, without having to slide the power button...is it staying after shutdown?

[funman]

Can confirm that the hold switch is instantly detected. However (yes, unfortunately there is a but) sometimes the backlight flashes on and off quickly for about a second or two before the backlight finally turns off. If you boost cpu through debug menu, it is easily reproducible. Unfortunately another case of extra delays for the boost ?..

Yes hold doesn't work when boosted, so possibly the delay needs some tweaking.
EDIT: fixed
EDIT2: fixed properly :p

Hey all-
Just now, I crashed after selecting an mp3, with the following error:

As I just wrote today:

Make sure you boosted the CPU before reporting crashes (read last pages of this thread)

You can do so by increasing the counter in debug menu -> cpu frequency

Another thing- I'm not sure if this is standard behavior or not, but after shutting down from rockbox, if I connect a USB chord in the Fuze, the original firmware is loaded up and a connection is established, without having to slide the power button...is it staying after shutdown?

Plugging USB powers the device, I think it's not documented by rockbox as it's a feature of the hardware (nothing to do with rockbox).

Fuzev2 automatically boots to OF if USB is detected at boot time (we use this as a recovery method).

This is documented in the Fuzev1 manual where most of the content should apply to fuzev2 as well (although not guaranteed).

[bilditup1]

Aight, so I guess I'll boost the freqs and try to reproduce the error. Sorry for not paying more attention.
-e

Update:

I began to try the newest build today, but it seems microSD support is broken; my card is not recognized. Reverting back to 25733 from 25747 fixes this. Again, apologies if this is not helpful.

[Xanikseo]

OK some updates...

Hold definitely totally works now, which is great!
The button light has started flashing again while switching tracks though . EDIT: Only when unboosted though...
While boosted in debug, and when switching track, I got this error (2 times so far):

Code: [Select]

Data abort
at 30027024
FSR 0x8
(domain 0, fault 8)
address 0xBB0F89FF

In r25750, I got this error while doing the same:

Code: [Select]

Data abort
at 30026FD8
FSR 0x8
(domain 0, fault 8)
address 0xBB0F89BF

I notice that your most recent svn submission seems to have solved all plugin problems, funman .

EDIT: Switching tracks seems to be generally very slow when done from the now playing screen. While viewing the playlist and selecting files to play, switching is fast. I think it just takes some time to gather the various data in the tags of the current file playing (while in now playing screen), before it even considers playing the next track?

Yes, it seems that when selecting a new file, and then if you jump to the now playing screen, you can't do anything for about 2 seconds. If you don't go to the now playing screen, there's no problem.

[funman]

it seems microSD support is broken; my card is not recognized. Reverting back to 25733 from 25747 fixes this.

can you compile your own build ? If so try to change the udelay(1000); into udelay(10000); in firmware/target/arm/as3525/sd-as3525v2.c
EDIT: 10000 might be too high for the function so no need to try : it won't work

Again, apologies if this is not helpful.

You gave a clear description with 2 revisions : working and not working. Sounds like an helpful report to me


- pasting what I posted on ABI -

I've put a Clip+ custom build on my ISP homepage, can you give it a try ?

I had one crash with it but it seems much more stable than current builds, so I'd be interested to know how it works for other people.

It's based on current SVN + a patch which could give a hint on how to fix the problems properly. And it should work with any bootloader.

No need to force CPU boosting with it (the patch affects the function which changes CPU frequency so if you force boosting then the patch won't run at all).

EDIT: better use a current build, i've pushed some modifications.

I could play a full album on Clip+ / Fuzev2, got a single crash on Clipv2 but different from what I used to see.

[Timar]

I had one crash with it but it seems much more stable than current builds, so I'd be interested to know how it works for other people.

That's great news, thanks! Runs stable so far (a half an hour)... all other recent builds crashed after 20 seconds of playback at the latest if unboosted

EDIT: Still running stable. There's another nice side effect of the new build. On all other builds as well as on the OF there was a buzzing sound at the end of the fade-out as well as at the begin of the fade-in of playback (it was much more distracting on the rockbox though). This sound is completely gone now

[bilditup1]

Hmm. In 25753, microSD(HC?) support still seems to be screwy.
On start up, an 8GB microSDHC card I've put in doesn't show up. If I insert, remove it a few times, eventually, I get the following screen (black text on white background):

Code: [Select]

*PANIC*
microSD init failed : -8

A regular 2GB microSD card was recognized with no problems.
Any suggestions, guys?

UPDATE: Good pt - this was a Fuze v2.

[Timar]

On my 8gb Clip+, an 8gb MicroSD card works flawless...

[FlynDice]

@Timar

Can you startup with your uSD inserted?   I have been unable to for some time now and I think I have tracked it down to INT_GPIOA not being triggered for some reason on startup.  I can get it to init normally if I wait until rockbox is running but if I try to boot with the card inserted I hang at the rockbox logo.  Anyone else seeing this on clip+?

[mitk]

I've no such problem with clip+ 8G and Sandisk 16G uSDHC card inserted. svn:25754

[Timar]

Can you startup with your uSD inserted?

Yes, on all recent rockbox builds, the MicroSD support worked for me, no hangups at startup. Strange... maybe it's dependant of the SD card's parameters? (I'm using an 8gb Sandisc card)

[bilditup1]

On my Fuze v2, I can startup with a usDHC inserted, but the 8gb card will not be recognized (the 2gb one will). Reverted back to 25733 for the meantime.
-e

[Timar]

Got a playback crash on r25761:

Code: [Select]

Data Abort @ 30026228
FSR 0x8 (0, 8)
0xB48ACDA0

EDIT: On Clip+. This crash occured a few times since upgrading from funman's custom build to the latest SVN. I also get "Can't load codec" error msgs.

[funman]

That would be in apps/gui/skin_engine/skin_display.c::evaluate_conditional()

If I compare to my rockbox.elf it crashes on line 589 when accessing *value.

I have no idea if it could come from a bug in this code or from mem corruption.