I've just upragded from 2.425-vrs2-tcl1 to the 2.6.8.1-tcl1 kernel to see
if it would help with my CompactFlash card woes.

When booted with a 512MB memory card in the CF slot, the ide_cs driver now
gets a lot further than it ever did before, and the scripts in
/etc/pcmcia even get as for as trying to mount it for me - trying various
fs types in turn:

  hda: SajDisk SDCFH-112, ATA DISK drive
  ide0 at 0xc4800000-0xc4800007,0xc480000e on irq 42
  ide-cs: hda: Vcc = 5.0, Vpp = 0.0
  hda: max request size: 128KiB
  hda: 1000944 sectors (512 MB) w/1KiB Cache, CHS=993/16/63
  hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
  hda: task_no_data_intr: error=0x00 { }
   /dev/ide/host0/bus0/target0/lun0: p1
   /dev/ide/host0/bus0/target0/lun0: p1
  FAT: invalid first entry of FAT (0xfff8 != 0xfbf8)
  VFS: Can't find a valid FAT filesystem on dev hda1.
   /dev/ide/host0/bus0/target0/lun0: p1
  FAT: invalid first entry of FAT (0xfff8 != 0xfbf8)
  VFS: Can't find a valid FAT filesystem on dev hda1.
   /dev/ide/host0/bus0/target0/lun0: p1
  VFS: Can't find ext3 filesystem on dev hda1.
   /dev/ide/host0/bus0/target0/lun0: p1
  VFS: Can't find ext2 filesystem on dev hda1.
   /dev/ide/host0/bus0/target0/lun0: p1
  cramfs: wrong magic
   /dev/ide/host0/bus0/target0/lun0: p1

I am most concerned about the first line:

  hda: SajDisk SDCFH-112, ATA DISK drive

which I think should actually say:

  hda: SanDisk SDCFH-512, ATA DISK drive

The 'n' and the '5' appear as 'j' and '1' respectively, which I suspect is
due to a single bit error:

  'n' = 0x6e = 01101110
  'j' = 0x6a = 01101010
                    ^
  '5' = 0x35 = 00110101
  '1' = 0x31 = 00110001
                    ^

However it's clearly not affecting every byte as (for example)  'D' is
unaffected:

  'D' = 0x44 = 01000100
                    ^
although it has the problem bit set.

The problem is repeatable, and I presume it is the reason that none of the
attempts to mount filesystems worked:

  FAT: invalid first entry of FAT (0xfff8 != 0xfbf8)

0xfff8 = 11111111 11111000
0xfbf8 = 11111011 11111000
              ^

Have I got a dodgy connection somewhere on my board?



Paul Fidler
-- 
Cambridge University Engineering Department | Tel: +44 1223 332816
Trumpington Street, Cambridge, CB2 1PZ, UK  | Fax: +44 1223 332662