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

[FlynDice]

This is an eMMC document but with all the similarities between SD and MMC it fills in a lot of the holes that are missing in the simplified SD 2.0 spec.

http://www.jedec.org/download/search/JESD84-A44.pdf

[breno]

Just installed Rockbox on my sansa c240v2 (r22231) and the system overall is working fine but with some major problems:
- When playing any music it plays for 1 second them crashes
- Volume buttons dont seem to work

Besides that, it's working:
- Radio is working (change volume in the menu)
- Microsd (it recognizes the files at least)
- Plugins (games work fine except for doom that gives me some z_malloc errors and them restarts the system)
- Display (seems to bem ok)
- USB Charge works (changes to OFW)
- Dual boot works fine (holding the right key when turning on)

I compiled it using cygwin.
Has someone got rockbox to work on c200v2 yet? which release was it?

thx guys!

[funman]

Hi, some things are still to be fixed on c200v2: see this page.

All these are known: only the button cross work (not volume -/+, hold, and rec), doom doesn't and will never work, playing music crashes.

What does the radio debug screen say ? (system -> debug -> radio), FM radio worked fine on my c200v2 when i tried it

[breno]

Hi, some things are still to be fixed on c200v2: see this page.

All these are known: only the button cross work (not volume -/+, hold, and rec), doom doesn't and will never work, playing music crashes.

What does the radio debug screen say ? (system -> debug -> radio), FM radio worked fine on my c200v2 when i tried it

Well actually the radio is now working, dont know why it wasnt working yesterday but somehow is working now.
Playing music still crashes. So why on the SansaAMS page it shows that Playback is working?

[funman]

PCM driver works fine, but there is a problem in rockbox generic buffering code, when there is only a few memory available (this is the case for clip, m200v4 and c200v2).

I added a red message in the comments for those 3 players.

[iudex]

Is it possible to disable Crossfade to earn more memory?
(Did you understood this? )

[funman]

Crossfading is already disabled, removing it completely from the compiled code would earn a bit of memory and prevent people enabling it from seeing crashes.

I would prefer to see if crossfading can be enabled at runtime by using audiobuffer size checks, perhaps someone can work on this when looking at the playback crashes on c200v2/clip/m200v4

[saratoga]

Making it a compile time option controlled by the config-XXX.h files might be even better, since it would maximize the saved memory on low mem targets.

[breno]

If you guys need someone to test something in the c200v2 build, I can compile a new build and see how it is working. And send the debug info for you guys.

[matsch]

MP3 playing seems to be stable with PCLK 31 Mhz, FCLK 62 MHz and IDE_FREQ 31 MHz
for CLIP V1. At least I did not experience a runtime error while playing for some hours in
contrast to the 248/62/90 MHz setting.

[funman]

The playback stopping is a bug in buffering, not in clocking.

The clock speeds are the same on all other Sansa AMS models.

You can trigger the bug by using an even smaller audio buffer (using static unsigned char[A_LOT]; in firmware code for example)

[breno]

I've noticed you have made some changes regarding low ram devices like mine in the latest releases. I've compiled the r22249.

Music playback is still the same, mp3 plays for a second them stops and flac dont even plays anything. After trying to play something i still can go back to the menu but the activity icon stays on and i cant do anything else (like going to fm radio) - this was already like this in previous builds, im just explaining how it is working on my c240v2.

[FlynDice]

I posted FS#10507 to flyspray tonight which lets me run my uSD card at 31 MHz which is within the SD specs.  I also ran it at 15.5 which would be the non-highspeed setting and it ran fine.

EDIT: Nevermind, this does not run at the lower freqs, the bypass bit is still set...

[matsch]

The playback stopping is a bug in buffering, not in clocking.

If there is a bug in buffering, the frequency of runtime error caused by this bug depends on
clocking and audio buffer size. Difficult for me to understand why clocking has an effect.
 
 

[kugel.]

If there is a bug in buffering, the frequency of runtime error caused by this bug depends on
clocking and audio buffer size. Difficult for me to understand why clocking has an effect.

He said clocking isn't related to it.

Did anyone try flyndice's patch with the problematic microsds, or impact to transfer rate?