[Libcdio-devel] Re: DVD problems.

[Thomas Schmitt]

Hi,

> My current thinking right now is to create a MMC object for all of the
> cached information

Is libcdio always aware when the media state
changes ? So that it can discard invald cache
parts and re-inquire them from the drive when
the application asks for them.

If not, then there should be an opportinity
for the application to force re-inquiry.
Maybe it knows better when not to trust the
old cache info any more.


> > [Media recognition]
> > Give me a stub where to examine "DVD-ROM" and
> > "BD-ROM". Map all other valid non-CD profiles
> > to the appropriate CDIO_DISC_MODE_*.
> > Invent new ones which for BD profiles.

> So if I have the above correct, I think this has to wait until we have a way
> to get and cache current profile information. Right?

The planned call for profile inquiry would be
used if it is available.
But there is not much penalty in issueing a GET
CONFIGURATION command to obtain the Current
Profile directly from the drive.

Sketch:

- Call GET CONFIGURATION for Current Profile in
  byte 6 and 7 of its reply

- All profiles map directly to media types,
  except:

  - Invalid profile or CD profiles 0x08 - 0x0a
    get handled by traditional libcdio code.

  - Profiles 0x10 "DVD-ROM" and 0x40 "BD-ROM"
    get examined by READ DISC INFORMATION Format
    code 00h. The reply differs between DVD and
    BD, but both have a 4-byte header before the
    actually interesting data:

    - DVD:
      Disk Category in bits 4 to 7 of byte 0+4
      gives the "Book Type". See table 400/401
      in previous mail.
      The book types and the DVD profiles are
      not exactly the same thing. But if a DVD
      says it is book type "DVD+R" then we can
      expect that it appears as profile DVD+R in
      a burner drive.

    - BD:
      Disc Type Identifier in bytes 8+4 to 10+4
      is clear text :
      BDO for BD-ROM, BDW for BD-RE, BDR for BD-R
      (mmc5r03c.pdf,
       Table 448 - General DI Unit Format.
       I wrote a wrong statement about this being
       in 108+4 to 110+4.)
      This does not distinguish the two possible
      BD-R profiles. Assume SRM, if distinction
      of BD-R SRM and BD-R RRM is desired at
      all.

So the MMC part would be quite clear.
What i cannot do yet, is to identify the places
in the code where this examination shall happen.

Which parts of libcdio are currently determining
media types ?
How to do the sidestep into the new recognizer
and how to decide whether that suffices or
whether to dive into the old CD recognizer ?


You could assume a magic recognizer and define
its prototype and reply values.
Then you write a recognizer dummy which always
replies "don't recognize" and place its calls
where needed.
I would then put more brain into the dummy, so
that it can say: "this is a DVD+R, regardless
what the Current Profile states".


Have a nice day :)

Thomas