Messages in this thread |  | | Date | Sun, 26 Jan 97 02:46 EST | From | (Steven S. Dick) | Subject | AHA1542 SCSI reset code still broken??? |
| |
I have an AHA 1542CF scsi card with a HD, cdrom, and tape drive attached. Both the cdrom and tape drive are external, and the tape drive is usually off.
Back around 1.3.80 or so, I had a problem where the scsi system would stop working completely if I tried to copy more than a few hundred K from the cdrom at a time. The driver would try to reset the device, and then would whine that the scsi bus wasn't responding, and then everything touching the hard drive would hang.
Well, I'm now running 2.0.28, and the problem is back. I don't pull data from the cdrom much, so I haven't noticed. Also, I've upgraded my motherboard (but not the scsi card), and the cdrom data problems seem to have gone away with the upgrade, but are not gone. I've never had it get stuck while playing a music track until now...
I'm not sure exactly what happened below, but the cdrom was playing a music disk, and was instructed to go to a different track. I heard it repeatedly seek and then pause a while, and repeat, like it was having trouble finding it. It appeared that the command to play never returned from the syscall.
Jan 26 02:23:22 nevets kernel: scsi : aborting command due to timeout : pid 45339, scsi0, channel 0, id 6, lun 6 0x48 c0 00 00 07 01 00 08 01 00 Jan 26 02:23:23 nevets kernel: scsi : aborting command due to timeout : pid 45340, scsi0, channel 0, id 6, lun 5 0x42 a2 40 01 00 00 00 00 10 00 Jan 26 02:23:53 nevets kernel: scsi : aborting command due to timeout : pid 45339, scsi0, channel 0, id 6, lun 6 0x48 c0 00 00 07 01 00 08 01 00 Jan 26 02:23:53 nevets kernel: SCSI host 0 abort (pid 45339) timed out - resetting Jan 26 02:23:53 nevets kernel: SCSI bus is being reset for host 0 channel 0. Jan 26 02:23:53 nevets kernel: Sent BUS DEVICE RESET to target 6 Jan 26 02:23:53 nevets kernel: Sending DID_RESET for target 6 Jan 26 02:23:53 nevets kernel: Sending DID_RESET for target 6 Jan 26 02:24:00 nevets kernel: aha1542_intr_handle: Unexpected interrupt Jan 26 02:24:00 nevets kernel: tarstat=0, hastat=0 idlun=90 ccb#=5 Jan 26 02:24:26 nevets kernel: scsi : aborting command due to timeout : pid 45339, scsi0, channel 0, id 6, lun 6 0x48 c0 00 00 07 01 00 08 01 00 Jan 26 02:24:26 nevets kernel: SCSI host 0 abort (pid 45339) timed out - resetting Jan 26 02:24:26 nevets kernel: SCSI bus is being reset for host 0 channel 0. Jan 26 02:24:26 nevets kernel: Sent BUS DEVICE RESET to target 6 Jan 26 02:24:26 nevets kernel: Sending DID_RESET for target 6 Jan 26 02:24:26 nevets last message repeated 2 times Jan 26 02:24:26 nevets kernel: scsi : aborting command due to timeout : pid 45340, scsi0, channel 0, id 6, lun 5 0x42 a2 40 01 00 00 00 00 10 00 Jan 26 02:24:26 nevets kernel: SCSI host 0 abort (pid 45340) timed out - resetting Jan 26 02:24:26 nevets kernel: SCSI bus is being reset for host 0 channel 0. [[[[I power cycled the external cdrom at this point]]]]] Jan 26 02:24:28 nevets kernel: Disc change detected. Jan 26 02:24:57 nevets kernel: scsi : aborting command due to timeout : pid 45339, scsi0, channel 0, id 6, lun 6 0x48 c0 00 00 07 01 00 08 01 00 Jan 26 02:24:57 nevets kernel: SCSI host 0 abort (pid 45339) timed out - resetting [[[[the system got very slow here--a virtual console switch took 3 seconds]]]] Jan 26 02:24:57 nevets kernel: SCSI bus is being reset for host 0 channel 0. Jan 26 02:24:57 nevets kernel: Sent BUS DEVICE RESET to target 6 Jan 26 02:24:57 nevets kernel: Sending DID_RESET for target 6 Jan 26 02:24:57 nevets last message repeated 2 times [I rebooted successfully and cleanly with ctrl-alt-del]
The cdrom may actually be at fault for _causing_ the problem, but even when the problem was cleared, the scsi driver did not seem to recover. There were at least several successful disk transactions after the first scsi bus reset, as the disk was clean at boot...
In the past, I've let things go longer, only to have the machine lock up while trying to reset the card, with longer and longer pauses between shorter and shorter working moments until it gets reset.
An excerpt from the boot messages follows:
Configuring Adaptec (SCSI-ID 7) at IO:330, IRQ 11, DMA priority 5 scsi0 : Adaptec 1542 scsi : 1 host. Vendor: NEC Model: D3827 Rev: 0410 Type: Direct-Access ANSI SCSI revision: 02 Detected scsi disk sda at scsi0, channel 0, id 4, lun 0 Vendor: NRC Model: MBR-7 Rev: 110 Type: CD-ROM ANSI SCSI revision: 02 Detected scsi CD-ROM sr0 at scsi0, channel 0, id 6, lun 0 [this is a 7 disk changer, on id 6, lun 0-6, other entries deleted] scsi : detected 7 SCSI cdroms 1 SCSI disk total. [tape drive is currently off]
I would be willing to help debug this in any way I can. Unfortunately, this was a LOT more repeatable with my old [slower] motherboard, as it would die _every_ time I did a large copy. Now, it has problems only randomly; sooner if I set the DMA transfer rate over 5mps (where it is now).
Why would it never recover from the reset?
Steve ssd@nevets.oau.org
|  |