You know what ..?
We resolved all our MMC problems with ... a capacitor !
Fact is that we supply gumstix with 3.6V, but writing on MMC is 'extremely' greedy, so it lowers our voltage to a non return point .. and it reboots ! So we just increased our capacitor to 1 mF, and expect to increase voltage too.

So gumstix seems rugged back again !)

PS : current consumption vary with MMC : some of them didn't work before, some of them worked with little write access (<100ko)


2005/7/23, Jérôme Multrier <jerome.multrier@gmail.com>:
We finally succedded in repairing my filesystem.
On a MMC card reader :
fsck.vfat /dev/sdd1
dosfsck 2.11, 12 Mar 2005, FAT32, LFN
/RobotacticData/config/trames.txt
  Contains a free cluster (57585). Assuming EOF.
/RobotacticData/config/trames.txt
  File size is 26346 bytes, cluster chain length is 0 bytes.
  Truncating file to 0 bytes.
Reclaimed 18216 unused clusters (37306368 bytes).
Leaving file system unchanged.
/dev/sdd1: 187 files, 29545/62412 clusters

And fsck.vfat -a /dev/sdd1 repaired it, and it was running again !
But we still don't know when happened the problem, and how we could pass though thoses types of errors ...

Thanks anyway !

2005/7/23, Jérôme Multrier <jerome.multrier@gmail.com>:
Hello guys !
During working hard on our gumstix software (at 5:00), we had a really big problem :

TRAMES : ecriture dans le fichier /root/.Robotactic/config/traFAT: Filesystem panic (dev mmcblk0p1)
mes.txt
    fat_get_cluster: invalid cluster chain (i_pos 3689299)
    File system has been set read-only
FAT: Filesystem panic (dev mmcblk0p1)
    fat_get_cluster: invalid cluster chain (i_pos 3689299)
FAT: Filesystem panic (dev mmcblk0p1)
    fat_get_cluster: invalid cluster chain (i_pos 3689299)
FAT: Filesystem panic (dev mmcblk0p1)
    fat_get_cluster: invalid cluster chain (i_pos 3689299)

It seems to be a problem when we append some data to the file trames.txt on our MMC.
Did anyone had a similar problem, or have an idea to solve that ...?


--
Jerome



--
Jerome



--
Jerome