[libdvdcss-devel] libdvdcss 1.2.9 and fatal error in vts css key

[Jorgen Lundman]

Having some issues with "fatal errors" when testing libdvdcss. I am
eventually hoping to add DVDCSS capabilities to an embedded device.

When reading a normal file (OsX, Linux and uClinux systems tested),
which ISN'T scrambled, I get the following situations:

# ./csstest aatcsdf-sample-ntx.vob 1
cool, I found libdvdcss version 1.2.9
libdvdcss debug: opening target `aatcsdf-sample-ntx.vob'
libdvdcss debug: using libc for access
libdvdcss debug: using CSS key cache dir:
/tmp//-bc1b0f9952e01b8d-0000000000/
requested sector: 1c00cb02...ca79061341080189a420d04c0032d1c8900600262180...
sector is not scrambled

# ./csstest aatcsdf-sample-ntx.vob 2
cool, I found libdvdcss version 1.2.9
libdvdcss debug: opening target `aatcsdf-sample-ntx.vob'
libdvdcss debug: using libc for access
libdvdcss debug: using CSS key cache dir:
/tmp//-bc1b0f9952e01b8d-0000000000/
requested sector: 00000000...00000000000000000000000000000000000000000000...
sector is not scrambled

# ./csstest aatcsdf-sample-ntx.vob 3
cool, I found libdvdcss version 1.2.9
libdvdcss debug: opening target `aatcsdf-sample-ntx.vob'
libdvdcss debug: using libc for access
libdvdcss debug: using CSS key cache dir:
/tmp//-bc1b0f9952e01b8d-0000000000/
requested sector: 50f654e1...fd4b6c8505b5e00ffb142960b328e8a9646e3f6467c2...
libdvdcss debug: cracking title key at block 3
libdvdcss debug: non MPEG block found at block 3 (end of title)
libdvdcss debug: end of title reached
libdvdcss debug: successful attempts 0/0, scrambled blocks 0/0
libdvdcss debug: title key is 00:00:00:00:00
libdvdcss error: fatal error in vts css key
seek failed (fatal error in vts css key)

Now if I ignore these errors and just save the buffers (from the
original read call) the file comes out identical, and indeed, plays. But
the errors seem annoying. I guess initially the function isscrambled()
is not supposed to return true for any packet in this .VOB. But why the
seek error?



Has anyone looked at the bigger picture of how one could use this
library transparently? I noticed there is a Fuse-fs using it, so that
could be an option, if I can get Fuse support to work.

Lund