[Debburn-devel] DATACE 8X DVD Burns at 0.1x speed

Thomas Schmitt scdbackup at gmx.net
Sat Nov 17 06:32:00 UTC 2007


Hi,

> wodim -force dev=/dev/scd0 fs=14M speed=8 -dao -eject -overburn -v blah.iso

Did you try it without option -force ?


> Errno: 5 (Input/output error), send opc scsi sendcmd: no error
> CDB:  54 01 00 00 00 00 00 00 00 00
> status: 0x2 (CHECK CONDITION)
> Sense Bytes: 70 00 00 00 00 00 00 0E 00 00 00 00 00 00 00 00
> Sense Key: 0x0 No Additional Sense, Segment 0
> Sense Code: 0x00 Qual 0x00 (no additional sense information) Fru 0x0

The failed command was
54h SEND OPC INFORMATION (see also MMC-5 6.36)
and the failure was not handled properly between
drive and operating system.
There should be "additional sense information"
but either the drive did not tell or the OS did
not listen to it.

Whatever, OPC means "Optimum Power Calibration"
and is the procedure by which the drive explores
the media and tries to accomodate to it.

CDB:  54 01 tells that automatic OPC was demanded
by wodim. A drive would be allowed to skip OPC
if no such command was sent. (I would not send it
and leave media accomodation to the drive alone.)

Command 54h SEND OPC INFORMATION would also allow
to override this procedure by own parameter settings.
The specs sparsely state:
"The OPC Values are device, media, and speed specific."
A program would be very daring to go this path.
wodim was not so daring.


> static char m_shenzen[] = "SHENZEN SG&GAST DIGITAL OPTICAL DISCS"
> [next mail]
> Use growisofs instead.  Surprisingly, growisofs uses a more up to
> date media manufacturer entry list and knows about DATACE dvd's and
> the write speed is optimal.

Such stuff is surely not present in the
dvd+rw-tools source package which contains growisofs.
I cannot tell much about wodim sources but growisofs
i have studied.

It does a normal MMC compliant question and answer game
with the drive.
E.g. it asks by command ACh GET PERFORMANCE first for
the size of the performance descriptor list, then by
a second call requests exactly that amount of data,
then examines the offered alternatives and chooses one.
This choice is sent to the drive by command
B6h SET STREAMING.
See  dvd+rw-tools-7.0/growisofs_mmc.cpp  function
set_speed_B6h().

See ialso the output of command dvd+rw-mediainfo which
tells the MMC commands as headlines of the info
blocks. (One can directly hop into the specs from
there. Thanks a lot to Andy Polyakov. Specs available
e.g. as http://www.t10.org/ftp/t10/drafts/mmc5/mmc5r03c.pdf )


> Question:  Why is it that the media manufacturer entry list is not
> as up-to-date in wodim?  Shouldn't they be using the same source base?

The question should rather be:
What voodoo is stored in that piece of code you found
in the wodim sources ?
Does it affect drive settings at all ?

Another question:
Are you sure growisofs performs better than
wodim in a significant number of tries ?

To my opinion, drive and media are at the
edge of mutual incompatibility. This could
mean that the success of growisofs was just
luck and that wodim was not so lucky.

Nevertheless, option -force plus cdrecord
genes gives wodim a clear theoretical disadvantage
when it comes to a competition with growisofs.


Have a nice day :)

Thomas




More information about the Debburn-devel mailing list