Request for comments: keydb.cfg convention

[awhitehead]

In keydb.cfg there are fields for the day, month and year of the manufacture of the disc.

For example, in a situation with "Relentless Enemies" timestamp is clear:

Volume in drive A is RELENTLESSENEMIES
Volume Serial Number is A3F7-6471

Directory of A:\

01/16/2007 01:25 AM <DIR> AACS
01/16/2007 01:25 AM <DIR> HVDVD_TS
01/16/2007 01:25 AM <DIR> ADV_OBJ
01/16/2007 01:25 AM <DIR> AACS_BAK
0 File(s) 0 bytes
4 Dir(s) 0 bytes free



That's why "Relentless Enemies" is still decryptable using the old processing key - it was mastered when only the first AACS attacks were starting to surface, and before the January 24th, when AACS LA confirmed that the first VUKs that were leaked were indeed legitimate.

However there are situations when on a single disc there are multiple timestamps:


Volume in drive A is PLANET_EARTH_D2
Volume Serial Number is 0829-4EE7

Directory of A:\

12/04/2006 03:49 AM <DIR> AACS
12/04/2006 03:49 AM <DIR> AACS_BAK
02/21/2007 02:20 PM <DIR> ADV_OBJ
02/13/2007 09:04 PM <DIR> HVDVD_TS
0 File(s) 0 bytes
4 Dir(s) 0 bytes free



I am leaning towards putting the latest time stamp as the timestamp in keydb.cfg, as it's likely the last date that the disc was modified. In this example, however, it seems like AACS keys were generated well in advance of the actual content, and the spirit of the time stamps (at least that's how I read it) was to track of the AACS changes.

So what time stamp should we be putting in? Latest timestamp on the disc, that would indicate mastering time, or the timestamp on \AACS?

Comments?

[JeffAlso]

Since \AACS\Unit_Key_RO.inf (for bluray) and \AACS\VTKF.AACS (for hddvd) are the files being hashed to produce the unique disc identifier used in keydb.cfg, it would make sense to me to use the timestamp of those files for the date.

Just my $0.02