sd/mmc drive

[phill]

HEllo,

I have a sd/mmc drive in my laptop. When I put a card in it I see the following in dmesg:
scsibus1 at sdmmc0: 2 targets, initiator 0
sd1 at scsibus1 targ 1 lun 0: <SD/MMC, Drive #01, > SCSI2 0/direct fixed
sd1: 1962MB, 1024 bytes/sec, 2009600 sec total

I try to mount as follows:
$ sudo mount /dev/sd1i /mnt/pen/
mount_ffs: /dev/sd1i on /mnt/pen: Device not configured

Am I doing something wrong or is the drive simply not supported?

Thanks

[shep]

I think you are missing the file type specification in your mount command.
Typically flash devices use the msdos file type (aka 'vfat' in Linux).

mount -t msdos /dev/sd1i /mnt/pen

See 14.17 in the FAQ for more info
http://www.openbsd.org/faq/faq14.html#flashmem

[phill]

It does not work either...

$ sudo mount -t msdos /dev/sd1i /mnt/pen/                                                                                                                    
mount_msdos: /dev/sd1i on /mnt/pen: Device not configured

[jggimi]

"sd1i" will only exist on the if there is a FAT or EXT2/EXT3 MBR partition on the device. Don't assume it, check it.

Look at the output of:

sudo fdisk sd1

This will show you what your MBR partition table has within it.
Look at the output of:

sudo disklabel sd1

This will show you what real BSD partitions are on the drive, or, what virtual BSD partitions are on the drive, having been created by the kernel from FAT or EXT2/3 MBR partitions.

----

If you have a valid MBR partition, but it is an NTFS filesystem, then you will need a custom kernel to mount it on OpenBSD. NTFS has been considered "experimental, read-only" for many years on OpenBSD.

[backrow]

Originally Posted by jggimi

If you have a valid MBR partition, but it is an NTFS filesystem, then you will need a custom kernel to mount it on OpenBSD. NTFS has been considered "experimental, read-only" for many years on OpenBSD.

NTFS is enabled in GENERIC now (4.9 will be the first release to have it). That said, I tried it a few weeks ago and couldn’t get it to work.

[phill]

I performed the commands:

$ sudo fdisk sd1 

Disk: sd1	geometry: 125/255/63 [2009600 1024-byte Sectors]
Offset: 0	Signature: 0xAA55
            Starting         Ending         LBA Info:
 #: id      C   H   S -      C   H   S [       start:        size ]
-------------------------------------------------------------------------------
 0: 06      0   3  55 -    250  46  52 [         243:     4018957 ] DOS > 32MB  
 1: 00      0   0   0 -      0   0   0 [           0:           0 ] unused      
 2: 00      0   0   0 -      0   0   0 [           0:           0 ] unused      
 3: 00      0   0   0 -      0   0   0 [           0:           0 ] unused      

$ sudo disklabel sd1

# /dev/rsd1c:
type: SCSI
disk: SCSI disk
label: Drive #01
uid: 0000000000000000
flags:
bytes/sector: 1024
sectors/track: 63
tracks/cylinder: 255
sectors/cylinder: 16065
cylinders: 125
total sectors: 2009600
boundstart: 0
boundend: 2009600
drivedata: 0 

16 partitions:
#                size           offset  fstype [fsize bsize  cpg]
  c:          2009600                0  unused

So the fstype is "unused"?

[BSDfan666]

No, "c" always represents the entire disk.

The problem here is that OpenBSD isn't spoofing a disklabel entry so that you can mount the msdos (..FAT) formatted partition.

The OpenBSD kernel creates a pseudo-disklabel for disks that have foreign filesystems, but sometimes, as is in your case.. it has not.

You should send a bug report to the mailing lists.

[jggimi]

It is not recognized, as far as I can tell, because of an overarching problem: 1024 bytes/sector. This is neither the standard 512-byte sector, nor the newfangled SSD 4096-byte sector. Nor even the 2048-byte sectors of optical disks.

[jggimi]

I should add that as a circumvention, a BSD disklabel can be manually added. The FAT MBR partition starts at sector 243 for a size of 4018957 sectors. Manually add this using disklabel(8) and its -E option or -e option. Here's a quick guess at admin inputs using disklabel -E, to create an "i" BSD partition to align with the MBR partition, just as an example:

$ sudo disklabel -E sd1
a
i
243
4018957
MSDOS
q
y

I have not actually tested this, these responses to questions from the disklabel -E editing tool are from memory. I have no OpenBSD platform available to me at this instant. If you do this, read the prompts from the disklabel program and answer appropriately.