[Mtools] Mtools 3.9.11 on VMS (Alpha) v. Nikon + SD card + USB card reader

[Steven M. Schweda]

   I'm having problems using "MTOOLS version 3.9.11a, dated May 31st,
2007" (that's modified for use on VMS) with a 1GB SD flash card in an
Inland 08310 USB "Multi Plus Card Reader", which is connected to an
XP1000 Alpha system running VMS V8.3.

      http://www.inland-products.com/singleproduct.asp?partnum=08310

I'm wondering if this is anything which is known to work (or not)
anywhere else, or if I should be worrying about the VMS interface code.

   The SD card was formatted in a Nikon COOLPIX L11 camera.  If I load
the SD card into the USB card reader, and attach it to a Mac running Mac
OS 10.4.10, it seems to work as expected.

   If I look at a dump of the data on the Alpha, I see what appears to
be a partition table (MBR) in sector zero, showing one partition which
starts at sector 245 with length 1999627 sectors (read the hex section
right to left):

Logical block number 0 (00000000), 512 (0200) bytes

[...]
 03 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 0001B0
 00 00 00 1E 83 0B 00 00 00 F5 DF FF 1F 06 00 39 9....ßõ......... 0001C0
[...]
 AA 55 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ..............Uª 0001F0

That is, as I see it, non-bootable (0x00), first sector at C|H|S =
0x03|0x39|0x00, partition type 0x06, last sector at C|H|S =
0x1F|0xFF|0xDF (???), logical block address of first sector = 0x000000F5
(= 245), sectors in partition = 0x001E830B = (1999627).

   At sector 245 (counting from zero), I see what looks like the start
of a FAT16 file system partition:

Logical block number 245 (000000F5), 512 (0200) bytes

 00 01 20 02 00 64 74 4C 6E 61 72 6F 5A 00 00 E9 é..ZoranLtd.. .. 000000
 00 00 00 F5 00 20 00 3F 00 F5 F8 00 00 02 00 02 .....øõ.?. .õ... 000010
 41 4E 5F 4F 4E 00 00 00 00 29 00 80 00 1E 83 0B ......)....NO_NA 000020
 00 00 20 20 20 36 31 54 41 46 20 20 20 20 45 4D ME    FAT16   .. 000030
 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 000040
[...]

where I seem to see reasonable data, like 0x003f (= 63) sectors per
track, 0x0020 (= 32) heads, 0x000000F5 (= 245) hidden sectors,
0x001E830B (= 1999627) total sectors, and so on.

   So, I specified the drive this way in the _mtoolsrc (".mtoolsrc")
file:
      drive w: file="MTOOLS_W" vms_block mformat_only partition=1
where the logical name MTOOLS_W points to the actual USB disk device.

   On the Alpha with mtools, I get complaints like:

ALP2 $ mtools -c minfo -v w:
 MTOOLS_W (_ALP2$DNA4:) mounted.
Partition 1 is not aligned
Warning: inconsistent partition table
Possibly unpartitioned device

*** Maybe try without partition=1 in device definition ***

If this is a PCMCIA card, or a disk partitioned on another computer, this messag
e may be in error: add mtools_skip_check=1 to your .mtoolsrc file to suppress th
is warning
 _ALP2$DNA4: dismounted.
init W: non DOS media

   If I change this test in mpartition.c: consistencyCheck():

      sector(partTable[i].start) != 1
to:
      BEGIN(partTable[i])% used_dev->sectors+ 1 != sector(partTable[i].start)

then I can get past the "inconsistent partition table" complaint, but
still remaining is:

init W: non DOS media

   At this point, knowing precious little about FATxx or Mtools
structure, I seem to be getting in deeper than I ought to, so I haven't
looked much further, but it looks possible that init.c: get_media_type()
is looking at sector zero instead of sector 245, and things go downhill
from there.

   So, is this anything which is known to work (or not) anywhere else, or
should I be worrying about the VMS interface code, or what?  (I'm hoping
that someone who knows something will look at the code, see that a case
like this is not handled correctly, but that it's trivial to fix, and so
on.)

ALP2 $ mtools --version
MTOOLS version 3.9.11a, dated May 31st, 2007
configured with the following options: disable-xdf disable-vold disable-new-vold
 disable-debug disable-raw-term

   If more info from me would be helpful, feel free to ask.

   In case you're wondering, the camera itself seems to present itself
as a big floppy disk instead of a SCSI disk, and the VMS USB device
support code doesn't see that as a disk, so the SD card + USB adapter
seems to be the closest thing to a solution, if only Mtools could deal
with the data it finds there.

------------------------------------------------------------------------

   Steven M. Schweda               address@hidden
   382 South Warwick Street        (+1) 651-699-9818
   Saint Paul  MN  55105-2547