ide-scsi problems with multiple CD-Rs on same channel (2.2.14)

From: Tkil (tkil-lk@scrye.com)
Date: Sat Jan 22 2000 - 17:42:12 EST


Greetings!

Summary:

  I am trying to burn to multiple IDE CD-Rs simultaneously. I can use
  the ide-scsi to burn a single drive, and I can simultaneously burn
  to two drives if they are on different IDE channels. If I try to
  burn to the third drive (slave on second channel), it seems that the
  commands are actually sent to the master, disrupting the "cdrecord"
  process there.

Hardware:

  CPU: K6-2 @ 262MHz (3.5x 75MHz bus), 96MB RAM
  Mobo: Asus P/I-P55T2P4 (Intel 430HX chipset)
  HD: 1x Quantum Bigfoot (hda)
  CD-Rs: 3x Mitsumi 4802TE CDRW (hdb, hdc, hdd)

  I'm not sure if it matters, but hdb has firmware revision 1.8, while
  hdc and hdd both have revision 1.9. Also, the rest of this hardware
  (the first three items) have been static and rock-solid for a long
  time, so I don't yet believe it's an intrinsic hardware issue.

Software:

  RedHat 6.1 + errata [1]
  Linux 2.2.14 (stock) with ide-cd and ide-scsi as modules
  cdrecord 1.8a37

Behavior:

  After loading the ide-scsi module, I have the following mapping:
  
    hdb -> scsi 0 id 0 lun 0
    hdc -> scsi 0 id 1 lun 0
    hdd -> scsi 0 id 2 lun 0
  
  I can record to any single CD-R successfully. I can successfully
  record to hdb and hdc simultaneously, as well. Even when burning to
  both, the machine seems to have less than 10% load, so I don't think
  that the problem is performance/bandwidth related (3 4x burns should
  be ~2 MB/s, which shouldn't be a problem even for poor old slow-ass
  quantum drive...)
  
  The problem starts when I try to burn to all three drives at the
  same time. Starting the burns in order (from same ISO image, if it
  matters), both hdb and hdc start up normally. When I try to start
  the burn to hdd, it seems to instead talk to hdc. This, quite
  reasonably, confuses the second cdrecord (the one that is
  "legitimately" talking to hdc). So the second cdrecord errors out
  (I can transcribe some sense data if it would help; I don't have any
  handy right now). The third cdrecord goes on as though nothing
  happened, but I would doubt that it would produce a functional
  platter (I've tested mostly with "-dummy" so far).
  
  The same thing happens if I am only burning to hdc (dev=0,1,0), then
  try to start a burn to hdd (0,2,0).

  Finally, I can successfully burn to hdd by itself. I'm not sure if
  I ever tried a burn to hdb and hdd simultaneously; I will probably
  try that a bit later today.

Plea:

  Does anyone have any ideas? I have looked at the ide-scsi source,
  but I didn't get very far. Has anyone else successfully used this
  sort of configuration? Is there more information I could provide
  that would help others with this issue?

Thanks very much,
t.

[1] It is running the 2000-01-01 6.1 release of KRUD:

       http://www.tummy.com/krud/

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:28 EST