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

[OldOne]

Hello,

Oldone here.  I only have an E260v2. I also had some problems with corruption on my internal card while trying the new builds with the wide bus update mentioned a few posts up.

I have been experimenting with all of the changes in the SD card module and haven't noticed any problems until the widebus change.  After the file corruptions,  I switched back to 23935.  I have also found that by changing the lowered voltage from 1.10 to 1.15, my troublesome PQI class 6 8gb cards work fine.  This has worked since about 23830. I build one with 1.10 voltage for my standard use and one with the voltage reduction set at 1.15.  I name the 115 version rock115.sansa and copy into the .rockbox directory.  When I want to listen to something on my troublesome cards, I just run the modified version.

I will wait on some more testing before using the widebus changes again.  The change was made at version 23977 I believe.  I have tried every change or patch for the SD cards Only changing the voltage scaling from 1.10 to 1.15 makes my PQI cards work perfectly.  Is there someway to switch up the speed or voltage sooner in the card access process(Does that make sense?) Is it possible to set the lower voltage to something between  1.10 and 1.15?  

By the way my SD cards seemed to work OK with widebus.  I tried copying from the card to internal card when the corruption on internal card occurred.

I have ordered some refurb 1 gb clips from woot.  I hope they are v1's so I can play with rockbox on them.

[epithetless]

I am also experiencing some newfound data corruption following my update to a post-"4 bit widebus mode" build (r24003, to be exact). Here's how it happened: After editing a WPS file on my Fuze and shutting the player off, any attempts to start the Fuze now result in the splashscreen coming up, a "scanning disk.." dialog box appearing for a second, and then "Data abort at 30800968" displaying. Any press of a button reboots the player and repeats the cycle, whether or not my microSDHC card is installed.

Looking at the contents of the .rockbox folder from my computer, I'm now seeing several new gibberish files and folders, the folders named "è;Å;" and the files named "è;ï;", ".", "c", "i", and "x".

[FlynDice]

Ok, I'm getting it now also.  For some reason the /.rockbox/playlist_control file is getting corrupted,  for me at least.  But test_disk write & verify passes and I can copy files from the uSd to the internal with no corruption at all.

EDIT:  I have changed the small write delay for non-HS uSD cards to a small write delay for all non-HS sd cards which includes the internal SD cards.  This seems to solve the problem for me so I committed the change.  For those experiencing problems would you please try it out now and let me know your results good or bad. Thanks!

[epithetless]

Interestingly, at this point, the latest build I have which doesn't result in the "Data abort at 30800968" message during bootup is r23903 -- if directory cache is enabled. When directory cache is off, Rockbox boots without error. This is true of the most recent build, r24020, as well.

After my initial "Data abort at 30800968" encounter, I ran Check Disk on the internal memory, formatted it using the OF, formatted it with the Panasonic SDFormatter v2.0, flashed the Fuze with an unpatched SanDisk firmware file, and reflashed it with a patched firmware file -- all in an effort to weed out any lingering issues -- but the problem persists (as I said, as long as directory cache is on). Any thoughts on this, FlynDice?

[ChiefAlex]

Since today, i get following after some seconds booted Rockbox:
*PANIC*
SD Xfer read err:0x8 Disk0
Anyone knows whats the reason?

regards,ChiefAlex

[FlynDice]

@ChiefAlex:

Thanks for the report and the error code, I haven't seen the read errors before.  That's a data timeout on the internal disk.

Which player are you using?


EDIT:

@ all:  For what it's worth my e280V2 is not displaying any of these problems right now.  I don't say that to discount the problems being experienced but to explain why this is becoming tricky to fix...

EDIT2:

@ epithetless:

"Data abort at 30800968"  seems to fall in the memcpy function as I look at rockbox.map.  It's always 30800968 correct?

[epithetless]

@ epithetless:

"Data abort at 30800968"  seems to fall in the memcpy function as I look at rockbox.map.  It's always 30800968 correct?

Yes, it's always 30800968.

Are any other Fuze users seeing this data abort once directory cache is enabled? The strange thing is I'd actually had directory cache enabled for a long time (as in many months) without any issue, but once the data corruption occurred yesterday, even builds which accepted directory caching before now fail with the abort message. Running Check Disk, thoroughly reformatting the device, and reflashing the firmware haven't changed this.

[ChiefAlex]

Hi FlynDice,
I use the Sandisk Fuze V1.
Some more ino: When i connect player to PC some "Files" are visible in the root folder. They have no-displayable names and cant get deleted. only way to fix it (for me) format the player via Windows. FAT or FAT32 seems to make no difference?

With a very old build from xx.09.09 this not occur. Thanks god i had this zip on my PC.
If you need more info about settings/whatever feel free to say, ill do my best to fix this.
Just DLed Rev.24039, lets see if it still occur

[samstar]

Hi all,
I have a sansa e260v2 w/ a kingston 2gb uSD card, and I've been following all of your work for quite some time.  This is my first time posting.  First let me say I'm very impressed at the progress being made here.  It's brought my music player back to life, as the OF was too frustrating to work with.  Much thanks to you all.

I wanted to confirm that updating my sansa from revision 2373X (give or take a few numbers) to 20429 yesterday also gave me corruption problems.  I spent the time formatting the internal memory after the update, and uploading the songs to it.  Whenever I tried to initialize the database, it would always respond after a reboot that the database needed to be initialized.  Cleaning the drive with chkdsk only temporarily solved the problem, and let me access the database for about ten minutes before it became corrupted.  I tried creating playlists from the file browser then, only to find that all the playlists I saved became corrupted as well.  It came to the point where linux would no longer recognize the partition as a usb storage drive when plugged in with the OF running.  I'd have to use the OF to reformat the drive first.  Also, I would often get panic errors, but I didn't have the presence of mind to write them down

After reverting my player to rev. 23929 (before the 4-bit widebus update) I have been able to initialize the database and create and move files without any corruption issues at all.  For the record, I've not had any corruption issues with the uSD card for any of the revisions.

If I can be of any other assistance, please let me know!
Thanks again!
Sam

[FlynDice]

Well, after leaving my player on all night I came back to find several strange files in my .rockbox folder this morning  ...

I have reverted the 4 bit widebus for now until this is figured out.  I'm sorry for the inconvenience, I'll test this better before it comes back.

Please let me know if you are still getting problems now!

[pwhodges]

I loaded Rockbox on my e280v2 for the first time yesterday (r24036).  It played ogg files happily for a while, and then I paused it and left it to turn off.  When I returned an hour later and turned it back on, every attempt to play a file gave "codec error" - for both ogg and mp3.  I looked at the .rockbox directory and found it full of corrupt filenames - single letters, "É;É;", and so on; some had mad sizes, GB more than the Sansa's memory. 

Paul

[samstar]

Updated to rev24060, put memory through its paces, copying files, creating databases, etc.  All work flawlessly so far.

[mc2739]

I loaded Rockbox on my e280v2 for the first time yesterday (r24036).  It played ogg files happily for a while, and then I paused it and left it to turn off.  When I returned an hour later and turned it back on, every attempt to play a file gave "codec error" - for both ogg and mp3.  I looked at the .rockbox directory and found it full of corrupt filenames - single letters, "É;É;", and so on; some had mad sizes, GB more than the Sansa's memory. 

Paul

Please update to the latest build.  This problem was corrected in builds r24054 and r24055.

[FlynDice]

For those who found corrupt file names in their .rockbox directory I would suggest formatting the internal drive with the OF.  It seems the damage is all contained within .rockbox but I have been unable to delete just the .rockbox directory.  Simply reinstalling an updated revision will not remove the damaged files.

Once again, my apologies for letting this slip through.  I used both test_disk and various size file/directory copies from disk to disk to test and found no problems.  I'm a bit busy in the role of Santa's elf at the moment but I'll get 4 bit widebus up on flyspray shortly and see if we can figure it out.

[mc2739]

The "Data abort at 30800968" looks like it might be another problem. Check FS#10860 for more details.