[2991] in linux-scsi channel archive

home help back first fref pref prev next nref lref last post

Re: CD Writing problems

daemon@ATHENA.MIT.EDU (Leonard N. Zubkoff)
Tue Dec 23 12:03:06 1997

Date: 	Tue, 23 Dec 1997 08:51:04 -0800
From: "Leonard N. Zubkoff" <lnz@dandelion.com>
To: rstagg@csc.com
CC: linux-scsi@vger.rutgers.edu
In-reply-to: <80256576.003CC159.00@csc.com> (rstagg@csc.com)

  From: rstagg@csc.com
  Date: 	Tue, 23 Dec 1997 11:09:54 +0000

  I'd appreciate any advice that you folks can give about tuning a system for
  CD writing. I'm currently writing most of my discs successfully, but a too
  significant number are failing.

  My system: 2.0.32 kernel, Buslogic BT958, 2xQuantum Fireball ST, Matshita
  CD-ROM and the Yamaha CDR-100.

  Software: cdwrite 2.0

  Problem: During writes, sometimes, the writing stops and I get hundreds of
  errors along the lines of "SENSE_ERROR iter 771: pipe_to_cd result 0,
  pack_id 27 sense ......" and then a pile of bytes. Followed by "18 of 18
  pipe_to_cd reply bytes:" and another pile of bytes. Once this happens, the
  disk has to be thrown away.

  I've not spotted any event which makes this happen. I write with the
  network interface turned off and all users slung off the machine. I can do,
  for example, three test writes perfectly well, and the fourth will come up
  with the above error. Sometimes it gets nearly to the end of the disk and
  then fails. Sometimes it does it almost immediately.

  Occasionally, this problem seems to upset the CDR, and I have to halt the
  system and power cycle before it starts behaving again.

  Any ideas or suggestions?

Are there any kernel error messages logged by the BusLogic driver or SCSI
Subsystem?  Without the sense data you omitted above, there's no way of
determining why the drive is unhappy.  The very fact that you are receiving
SENSE_ERROR indicates the problem is more likely with the CD writer itself than
the SCSI code.

You may be running into a problem that I ran into and reported on the LMP list
and may also have been reported on the kernel list too now: some read/write
calls are taking far longer than they should for no apparent reason.  If that's 
the case, you may be seeing an underrun error from the CD writer.

		Leonard

home help back first fref pref prev next nref lref last post