Re: cdrecord problems

["Thomas Schmitt" <scdbackup@gmx.net>]

Hi,

> I have tried all these modes such as tao, sao, dao, raw, etc. but none of
> these would work , I think may be linux kernel support to my new drive is
> not very good 

The Linux kernel does not provide support for a
particular drive. It rather forwards the MMC
commands of our burn programs (cdrecord, wodim,
growisofs, cdrskin) to the controller (SCSI,
USB, ATA, SATA, ...) and forwards the reply from
controller to burn program.
Any drive specific precautions would need to be
done by the burn program.


Your drive 1,0,0 (do you know its Linux device address ?)
does not confess to support any write mode.

It comes to me that i examined such a problem before.
  http://lists.alioth.debian.org/pipermail/debburn-devel/2007-February/000301.html
Interestingly it is with the same type of drive,
a DVD-RAM GSA-H30N.

Eduard Martinescu and me learned that wodim tries to find out
wether a write mode is supported by sending the preparation
commands for a drive. Quite the same is done by cdrskin's
foundation libburn.
The preliminary end points of our examination were
these two statements:

> Any mode page 05h sent via 55h MODE SELCT causes an
> error reply with Sense Key 5 ASC 1A ASCQ 00.
> That's "PARAMETER LIST LENGTH ERROR".
> tested with Page Length 0x32 and 0x36 as MMC-5
> would allow a 4 byte vendor specific add-on.

> Google finds me for "GSA-H30N" "mode page" a statement
> from the libata guys:
> commit 2e5704f63ed56b040a3189f6b7eb17f6f849ea22
> Author: Tejun Heo <htejun at gmail.com>
> Date:   Sun Dec 17 10:46:33 2006 +0900
> "[...] and HL-DT-ST DVD-RAM GSA-H30N malfunctions on
> initialization commands issued from SCSI."


Eduard had no time for further experiments. I don't know what
he finally did. Probably he just bought another burner.

My last proposals were:
- Try with original cdrecord as provided by Joerg Schilling
- Find out more about the reputation of GSA-H30N
- Try cdrskin -force (version >= 0.3.3) with very few hope
  for success

Remote ideas:
- Try to send the problematic command with a wider
  range of Page Length values. (That would need an
  experimental version of cdrskin.)


Have a nice day :)

Thomas