Re: Libburn reported an error SCSI error on write(544, 16): [5 21 02] Illegal request. Invalid address for write.

[Thomas Schmitt]

Hi,

Sid T wrote:
> Please refer to https://gitlab.gnome.org/GNOME/brasero/-/issues/357
> [gitlab.gnome.org] for more information.

Well, the only tangible info is in brasero-session.log which you would
be invited to show here.
And Brasero does not tell much ...

> BraseroLibburn Drive (/dev/sr0) init result = 1

At least it tells the drive's device file address.

> BraseroLibburn dvd/bd Profile= 11h , obs= 32768 , obs_pad= 1

So the medium is a DVD-R.

> BraseroLibburn DVD pre-track 01 : get_nwa(0), ret= 1 , d->nwa= 0

Obviously it is blank.

> BraseroLibburn SCSI error condition on command 2Ah WRITE(10): [5 21 02] 
> Illegal request. Invalid address for write.

The drive does not like this write command.
Probably because libburn saw a reason to repeat it several times.
At the third retry the drive answered by a different error, which invites
for re-trying:

> BraseroLibburn CDB= WRITE(10) #3,[3 02 00] : 2a 00 00 00 02 20 00 00 10 00  : 
> dxfer_len= 32768

This means according to MMC-5:
  3     = Medium error
  02 00 = No seek complete

Although we don't see the error messages of the first and second try,
this looks like the drive has a problem with positioning its laser.
Finally it replied by 5 21 02 which says that the write attempt goes to
a wrong address.

When writing DVD-R there is at any time only one block number where
writing is allowed. 33 such write commands must have succeed until this
one failed. It might be, though, that already the first write to block 0
has problems, but the drive's buffer accepted more until the burn unit
gave up with writing earlier blocks.

Is the medium still reported as blank afterwards ?
Inquire it by

  xorriso -outdev /dev/sr0 -toc

If the DVD-R did not record any data the output of the xorriso run should
contain the line

  Media status : is blank

Also interesting to me iare the lines which tell the drive maker and model
and the maker of the medium:

  Drive type   : vendor 'ASUS' product 'BW-16D1HT' revision '3.10'
  ...
  Media product: RITEKF1 , Ritek Corp


------------------------------------------------------------------------
Test request:

To get more info about the dialog between libburn and drive, do with a
blank DVD-R:

  path_to_payload_file=...path.to.some.file.of.at.least.4.MB...

  xorriso -scsi_log on -outdev /dev/sr0 -dummy on \
          -map "$path_to_payload_file" /test_file \
    2>&1 | tee -i /tmp/xorriso.log

For me this run lasts 64 seconds and produces 1737 lines of diagnostic
messages which get recorded in file  /tmp/xorriso.log .
Send this file directly to me
  scdbackup@gmx.net
I will then report to the mailing list what i read from it.

Don't get impatient if it lasts longer than 64 seconds. Depending on the
drive it might last as long a writing of 1 GB.

Because of "-dummy on", the DVD-R will not be really written.
After ejecting and re-loading it, it should still be reported as blank.

If all works well you will see near the end of the xorriso.log file

  ISO image produced: 4224 sectors
  Written to medium : 4384 sectors at LBA 0
  Writing to '/dev/sr0' completed successfully.

------------------------------------------------------------------------

In this case you may try to do a real burn with your intended payload
(which Brasero hid in /tmp/brasero_tmp_P4R280). I.e. without the xorriso
command "-dummy on".
Like

  xorriso -scsi_log on -outdev /dev/sr0 \
          -map /...path.on.hard.disk... /...desired.path.in.ISO... \
          ...maybe.more.map.commands... \
          -find / -exec mkisofs_r -- \
          -joliet on \
    2>&1 | tee -i /tmp/xorriso_real_burn.log

You may use as many -map commands as needed to describe source and
destination of the files which you want to put on the DVD.
The source paths may point to a directory, which will cause its whole file
tree to be put on the DVD, or a single data file (i.e no wildcards).

There will be many message lines. No need to send me the file
xorriso_real_burn.log if it succeeds.
But i don't think that xorriso will have more success than Brasero.
The problem seems to be in the drive and its relation to the medium.
If the run fails, then send me xorriso_real_burn.log .


Have a nice day :)

Thomas