Hello,

I am using VFS Manager APIs to work with raw binary files on the SD card.  I 
have recently had an alarming portion of my users experience a corruption of 
their filesystem, however.

It tends to manifest as a corrupted directory table.  A snippit of a "dir /s " 
on one such user's corrupted directory:

08/12/2006  03:08 AM            16,384 58.DAT
08/12/2006  03:08 AM            16,384 59.DAT
08/12/2006  03:08 AM            16,384 60.DAT
08/12/2006  03:08 AM            16,384 61.DAT
08/12/2006  03:08 AM       504,022,537 N
11/18/2035  10:01 AM     1,869,181,811 urveilla.nce
01/08/2005  01:50 PM    <DIR>          nicians:.\n8
01/13/2006  06:49 AM       959,788,845 nprofess.ion
11/18/2029  02:43 PM     1,181,637,725 rp.\n[Ma.nuf
10/16/2032  01:50 PM    <DIR>          inicians.: 8
03/15/1996  02:33 PM     1,767,255,671 mond-sha.ped
11/05/2037  10:17 AM     1,936,024,434 oration\.n[M
03/01/2037  01:34 PM     1,663,989,861 ient Inf.orm
03/24/2003  12:11 PM    <DIR>          866-448-.759
03/15/2037  02:03 PM       943,208,558 s:\nPage.t's

Here, you can see the tail end of the first 61 of my uniformly-sized data files 
show up just fine (there happen to be 118 in this directory); then, the 
contents of the data files themselves appear to be interpreted, in err, as 
FAT16 directory entries!  Yow.

Again, I'm not doing anything low-level with slot drivers or hacksih expansion 
manager stuff.  I'm simply using VFS Manager APIs to create, read, write, and 
seek into files.  This code has been working fine for about a year for a very 
large number of Palm users, but something appears to have changed (new 
hardware? poorly-engineered SD cards? new OS bug?) that is now causing some 
grief.

Has anybody else ever seen this kind of corruption?

Thanks,
-Jeff Ishaq
Palm OS(R) Certified Developer
-- 
For information on using the PalmSource Developer Forums, or to unsubscribe, 
please see http://www.palmos.com/dev/support/forums/

Reply via email to