$OpenBSD: README,v 1.2 2018/09/04 12:46:23 espie Exp $
+-----------------------------------------------------------------------
| Running ${PKGSTEM} on OpenBSD
+-----------------------------------------------------------------------
Utilities
=========
cd_disect Display SCSI / ATAPI device information
cd_sessions Optical device sessions report
mmc_format Format optical media
newfs_udf Format a filesystem
udfclient Read or write to a filesystem
udfdump Display contents of a filesystem
Issuing any command without options will produce a usage summary.
Considerations
==============
Special files and links
-----------------------
* All links will be followed and files replicated when copied to
a UDF filesystem, whether hard or symbolic links.
* Block, character, FIFO, and socket special files are not
supported.
UDF on disk drives
------------------
Disklabels are not used, except for virtual partition "a" with
2048-byte blocksize devices. MBRs are not used, the filesystem
begins at LBA 0. Partition "c" should be used if writing directly
to a drive instead of an image file. If the drive contains prior
data, zero the leading sectors before use as shown in the first
command below.
# dd if=/dev/zero of=/dev/rsd9c count=63
# newfs_udf -L shared.drive /dev/rsd9c
# ufsclient -W /dev/rsd9c
While it is possible to direct a filesystem image to a disklabel
partition, there would be no interoperability with other OSes.
Interoperability with mount_udf(8)
----------------------------------
Kernel recognition (virtual partition "a") and mounting r/o with
mount_udf() currently requires a 2048-byte logical block size, to
map a filesystem image to an optical drive's sector size.
This precludes direct use with sd(4) devices, and may prevent
interoperability with other OSes. The OSTA specification
requires a logical blocksize equal to the media sector size, and
the filesystem does not include a Virtual Allocation Table
required for -R media.
If you wish to use mount_udf(), include '-b 2048' with both
the newfs_udf and udfclient utilities.
UDF root directory
------------------
The UDF root directory is named with colon separated strings.
"<volume set>:<primary volume>:<logical volume>:<additional tag>"
The udfclient requires a 'cd' into this root directory before you can
read or write files and directories. Spaces will require quoting the
string between " " characters.
udfclient interactive commands
------------------------------
These are ftp-like, with some differences. Use the 'help' command at
the "UDF>" prompt to display a summary.
Of particular value is a 'get' or 'put' of a directory. This will copy
the entire structure in or out of the filesystem.
Example
=======
$ cd /tmp
$ mkdir test.directory
$ touch test.directory/{a,b,c,d}
$ dd if=/dev/zero of=test.iso bs=10m count=10
$ newfs_udf -S test -P test -L test test.iso
$ udfclient -W test.iso
.
.
.
UDF> ls
Directory listing of /
drwxrwxrwx -1 -1 40 test:test:test:fileset
UDF> cd test:test:test:fileset
UDF working directory is /test:test:test:fileset
Current FS working directory /tmp
UDF> put test.directory
A total of 0 kb transfered at an overal average of 0 kb/sec
UDF> get test.directory new.directory
new.directory/a ...
new.directory/b ...
new.directory/c ...
new.directory/d ...
A total of 0 kb transfered at an overal average of 115 kb/sec
UDF> quit
.
.
.
$ ls new.directory
a b c d
$
