| View previous topic :: View next topic |
| Author |
Message |
caramboo n00b


Joined: 01 Sep 2003 Posts: 12
|
Posted: Sun Jan 01, 2006 3:44 pm Post subject: SCSI DVD Drive Errors |
|
|
Hi
I recently moved from 2.6.12-gentoo-r7 to 2.6.14-gentoo-r6. At the same time I got automounting working with pmount/dbus and emerged KDE 3.5.
In the confusion of getting used to pmount etc I found that my Pioneer SCSI DVD Drive seemed to have developed a fault which I initially put down to the new automounting enviroment. After spending most of an afternoon trying to find the fault and almost giving up on the drive as dead I thought about booting back into 2.6.12. It worked, the seemingly dead drive sprang back into life and works as happily as before.
So, what's changed SCSI wise 2.6.12 --> 2.6.14?
Here's what dmesg shows for both kernels:-
| Code: |
ACPI: PCI Interrupt 0000:00:0d.0[A] -> GSI 16 (level, low) -> IRQ 19
scsi1 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
<Adaptec aic7899 Ultra160 SCSI adapter>
aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
Vendor: PIONEER Model: DVD-ROM DVD-305 Rev: 1.06
Type: CD-ROM ANSI SCSI revision: 02
target1:0:2: Beginning Domain Validation
target1:0:2: Domain Validation skipping write tests
target1:0:2: FAST-20 SCSI 20.0 MB/s ST (50 ns, offset 16)
target1:0:2: Ending Domain Validation
|
Then, for the 2.6.14 kernel, this is what it says after bunging in a DVD:-
| Code: |
(scsi1:A:2:0): No or incomplete CDB sent to device.
scsi1: Issued Channel A Bus Reset. 1 SCBs aborted
sr 1:0:2:0: SCSI error: return code = 0x10000
(scsi1:A:2:0): No or incomplete CDB sent to device.
scsi1: Issued Channel A Bus Reset. 1 SCBs aborted
sr 1:0:2:0: SCSI error: return code = 0x10000
|
..ad infinitum.
I tried compiling the "old" adaptec driver in the kernel sources but that didn't work either.
Anyone any ideas? |
|
| Back to top |
|
 |
Jaglover Watchman


Joined: 29 May 2005 Posts: 8291 Location: Saint Amant, Acadiana
|
Posted: Mon Jan 02, 2006 3:16 am Post subject: |
|
|
| I have 2.6.14-r2 and everything works with aic7xxx. (I have DVD and CDR). Could it be some termination or cabling problem? Could you post output of cat /proc/scsi/aic7xxx/0 please? |
|
| Back to top |
|
 |
caramboo n00b


Joined: 01 Sep 2003 Posts: 12
|
Posted: Mon Jan 02, 2006 8:47 am Post subject: |
|
|
| Jaglover wrote: | | I have 2.6.14-r2 and everything works with aic7xxx. (I have DVD and CDR). Could it be some termination or cabling problem? Could you post output of cat /proc/scsi/aic7xxx/0 please? |
Ok dokey:-
edited to show cat /proc/scsi/aic7xxx/1
| Code: |
# cat /proc/scsi/aic7xxx/1
Adaptec AIC7xxx driver version: 7.0
Adaptec aic7899 Ultra160 SCSI adapter
aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
Allocated SCBs: 4, SG List Length: 128
Serial EEPROM:
0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a
0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a 0xc33a
0x08f4 0x585e 0x2807 0x0010 0x0300 0xffff 0xffff 0xffff
0xffff 0xffff 0xffff 0xffff 0xffff 0xffff 0x0250 0xc250
Target 0 Negotiation Settings
User: 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
Target 1 Negotiation Settings
User: 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
Target 2 Negotiation Settings
User: 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
Goal: 20.000MB/s transfers (20.000MHz, offset 16)
Curr: 20.000MB/s transfers (20.000MHz, offset 16)
Channel A Target 2 Lun 0 Settings
Commands Queued 273935
Commands Active 0
Command Openings 1
Max Tagged Openings 0
Device Queue Frozen Count 0
|
I have an active terminator on both ends of the SCSI cable and I've tried adding/removing the termination jumper on the back of the drive. It's really puzzling because the drive has worked flawlessly from 2.4.* through to 2.6.12. |
|
| Back to top |
|
 |
Jaglover Watchman


Joined: 29 May 2005 Posts: 8291 Location: Saint Amant, Acadiana
|
Posted: Mon Jan 02, 2006 1:28 pm Post subject: |
|
|
OK. My (very wild) guess is: some setting with newer kernel is a little more agressive and this is why your misconfigured(?) h/w fails to follow. Termination thing. You should not use terminator on controller side, controller terminates it's end automatically. Active terminators require external power. If your DVD has no jumper marked TPwr or something like this and this is not activated then with current setup the device side is probably not terminated.
My advice. Remove both external terminators. Enable internal terminator in DVD-ROM. Connect one end of cable to controller and the other end to the DVD. Do not use sockets in the middle. Open ends cause serious problems. Cut your cable shorter if necessary. |
|
| Back to top |
|
 |
caramboo n00b


Joined: 01 Sep 2003 Posts: 12
|
Posted: Tue Jan 03, 2006 6:39 am Post subject: |
|
|
Thanks for your comments. I tried all sorts yesterday after reading your reply, removing terminators, shortening the cable, changing IDs, swapping chains and all gave the same result. A duff drive under 2.6.14 and a working one under 2.6.12.
Ah well, I'll persevere. It won't defeat me!  |
|
| Back to top |
|
 |
Jaglover Watchman


Joined: 29 May 2005 Posts: 8291 Location: Saint Amant, Acadiana
|
Posted: Tue Jan 03, 2006 12:59 pm Post subject: |
|
|
Sorry to hear it.
One last thing. Go into SCSI BIOS and reduce transfer rate for this device. (Ctrl+A from startup screen) |
|
| Back to top |
|
 |
spiralvoice Apprentice

Joined: 09 Mar 2004 Posts: 238
|
Posted: Thu Jan 05, 2006 1:39 am Post subject: |
|
|
Hi,
had the same problem: http://bugzilla.kernel.org/show_bug.cgi?id=5659
Using linux-2.6.13-gentoo-r5 + dbus/hald works without problems.
I stress-tested the drive, it also works on other OS as well, such
as latest Kanotix 2005-04/kernel 2.6.14.5.
Using vanilla-2.6.15/gentoo-2.6.14_all_versions + dbus/hald
produces the logged errors and renders the drive practically unusable.
Turning of dbus/hald solves the problem. |
|
| Back to top |
|
 |
caramboo n00b


Joined: 01 Sep 2003 Posts: 12
|
Posted: Fri Jan 06, 2006 5:50 pm Post subject: |
|
|
| Yep - same here. Simply stopping hald here brings the drive back to life (still 2.6.14-gentoo-r6). Going to build 2.6.15(gentoo) this weekend to see what happens. |
|
| Back to top |
|
 |
spiralvoice Apprentice

Joined: 09 Mar 2004 Posts: 238
|
Posted: Fri Jan 06, 2006 8:21 pm Post subject: |
|
|
| caramboo wrote: | | Going to build 2.6.15(gentoo) this weekend to see what happens. |
The bug is still present there. |
|
| Back to top |
|
 |
Polynomial-C Retired Dev


Joined: 01 Jun 2003 Posts: 1432 Location: Germany
|
Posted: Mon Jan 09, 2006 1:16 am Post subject: |
|
|
Hi,
I just reopened this bug @ bugzilla.kernel.org as I neither have dbus nor hal installed and I still suffer from this error when starting k3b or kscd with any kernel higher than 2.6.13.x.
Cheers
Poly _________________ The manual said "Requires Windows10 or better" so I installed GNU/Linux...
my portage overlay
Need a stage1 tarball? (Unofficial builds) |
|
| Back to top |
|
 |
spiralvoice Apprentice

Joined: 09 Mar 2004 Posts: 238
|
Posted: Mon Jan 09, 2006 11:56 am Post subject: |
|
|
| Polynomial-C wrote: | | I just reopened this bug @ bugzilla.kernel.org as I neither have dbus nor hal installed and I still suffer from this error when starting k3b or kscd with any kernel higher than 2.6.13.x. |
Please try Kanotix or any other LiveCD with kernel > 2.6.14 |
|
| Back to top |
|
 |
Polynomial-C Retired Dev


Joined: 01 Jun 2003 Posts: 1432 Location: Germany
|
Posted: Mon Jan 09, 2006 2:05 pm Post subject: |
|
|
Hi,
| spiralvoice wrote: | | Please try Kanotix or any other LiveCD with kernel > 2.6.14 |
tried Kanotix-2005-04 with Kernel-2.6.14.5-kanotix-9. Same problem. As soon as I run k3b the whole SCSI-bus gets resetted over and over again until the whole system is unuseable because it cannot access the Kanotix-CD anymore...
The errormessages are the same like those I posted @ bugzilla.kernel.org
Cheers
Polynomial-C _________________ The manual said "Requires Windows10 or better" so I installed GNU/Linux...
my portage overlay
Need a stage1 tarball? (Unofficial builds) |
|
| Back to top |
|
 |
Gon0S n00b


Joined: 15 Mar 2004 Posts: 27 Location: France
|
Posted: Wed Feb 15, 2006 2:07 pm Post subject: |
|
|
Hello
I encounter the almost same error with my rig.
The SCSI burner ID3 lead to error similar to this :
Nov 25 21:07:15 gentoo (scsi0:A:4:0): No or incomplete CDB sent to device.
Nov 25 21:07:15 gentoo scsi0: Issued Channel A Bus Reset. 1 SCBs aborted
Nov 25 21:07:15 gentoo sr 0:0:4:0: SCSI error: return code = 0x10000
Just the return code is = 0x70000
So, I just wanted to bump this topic to know if anyone managed to get rid of this problem |
|
| Back to top |
|
 |
spiralvoice Apprentice

Joined: 09 Mar 2004 Posts: 238
|
Posted: Wed Feb 15, 2006 5:53 pm Post subject: |
|
|
| Gon0S wrote: | | So, I just wanted to bump this topic to know if anyone managed to get rid of this problem |
Same problem here, with gentoo-sources-2.6.15-r5 |
|
| Back to top |
|
 |
Polynomial-C Retired Dev


Joined: 01 Jun 2003 Posts: 1432 Location: Germany
|
|
| Back to top |
|
 |
hamletmun Tux's lil' helper


Joined: 13 Jun 2002 Posts: 111 Location: Buenos Aires, Argentina
|
Posted: Tue Mar 28, 2006 2:53 am Post subject: Gentoo-sources-2.6.16 |
|
|
Same problem with Gentoo-sources-2.6.16
Has anyone found a solution for this???? |
|
| Back to top |
|
 |
hamletmun Tux's lil' helper


Joined: 13 Jun 2002 Posts: 111 Location: Buenos Aires, Argentina
|
Posted: Fri Apr 14, 2006 4:19 am Post subject: This could be one causing the error? |
|
|
Well... after git bisecting... I found this.
It has been a painful process: rebooting from 1966 revisions to 1 revision.
That means 1966->983->492->....->8->4->2->1 and, at least, two rebooting between
the numbers
Now someone has to find the solution for this problem
Author: James Bottomley <James.Bottomley@steeleye.com>
Date: Sun Aug 14 17:09:01 2005 -0500
[SCSI] correct transport class abstraction to work outside SCSI
I recently tried to construct a totally generic transport class and
found there were certain features missing from the current abstract
transport class. Most notable is that you have to hang the data on the
class_device but most of the API is framed in terms of the generic
device, not the class_device.
These changes are two fold
- Provide the class_device to all of the setup and configure APIs
- Provide and extra API to take the device and the attribute class and
return the corresponding class_device
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
:040000 040000 d9c2ca4cf51784a18a93b6ee7cabda19f7adcbd01567b7b65e5e34b7746f9bd78947eb7e26156b40 M drivers
:040000 040000 f38e13cdf96c1fb7c189409bb09fdda2c112d9848dff007873523a292c4841be1912fde22116993a M include |
|
| Back to top |
|
 |
spiralvoice Apprentice

Joined: 09 Mar 2004 Posts: 238
|
Posted: Mon Jun 19, 2006 7:37 pm Post subject: |
|
|
| I could be wrong, but gentoo-sources-2.6.17 with up-to-date ~x86 stuff does not have this problem anymore. I did not try to reproduce this bug for weeks so I do not know what really fixed it. |
|
| Back to top |
|
 |
|