View previous topic :: View next topic |
Author |
Message |
hab136 n00b
Joined: 12 Apr 2004 Posts: 18 Location: Charlotte, NC
|
Posted: Thu May 27, 2004 2:23 am Post subject: qlogic SCSI on sparc32 - scsi abort failure: 4005 |
|
|
Hi, I have a sparc 5, two Qlogic SCSI cards going to a bunch of SCA-connected disks in an enclosure. All equipment was at one time working under Solaris, but it's been moved and reassembled since then.
Code: |
# uname -a
Linux cancer 2.4.23-sparc-r1 #1 Sat Sep 24 06:03:56 EDT 2005 sparc sun4m Fujitsu MB86904 GNU/Linux
|
When I modprobe qlogicpti, it hangs, and I get this in dmesg:
Code: |
qpti0: IRQ 53 SCSI ID 7 (Firmware v1.31.32)(Firmware 1.21 95/05/18) [Fast Wide, using differential interface]
qpti1: IRQ 53 SCSI ID 7 (Firmware v1.31.32)(Firmware 1.21 95/05/18) [Fast Wide, using differential interface]
QPTI: Total of 2 PTI Qlogic/ISP hosts found, 2 actually in use.
scsi1 : PTI Qlogic,ISP SBUS SCSI irq 53 regs at fd01b000
scsi2 : PTI Qlogic,ISP SBUS SCSI irq 53 regs at fd01c000
scsi : aborting command due to timeout : pid 1645, scsi1, channel 0, id 0, lun 0 Inquiry 00 00 00 ff 00
qlogicpti : Aborting cmd for tgt[0] lun[0]
qlogicpti : scsi abort failure: 4005
SCSI host 1 abort (pid 1645) timed out - resetting
SCSI bus is being reset for host 1 channel 0.
qlogicpti : Resetting SCSI bus!
scsi : aborting command due to timeout : pid 1645, scsi1, channel 0, id 0, lun 0 Inquiry 00 00 00 ff 00
qlogicpti : Aborting cmd for tgt[0] lun[0]
qlogicpti : scsi abort failure: 4005
SCSI host 1 abort (pid 1645) timed out - resetting
SCSI bus is being reset for host 1 channel 0.
qlogicpti : Resetting SCSI bus!
scsi : aborting command due to timeout : pid 1645, scsi1, channel 0, id 0, lun 0 Inquiry 00 00 00 ff 00
qlogicpti : Aborting cmd for tgt[0] lun[0]
qlogicpti : scsi abort failure: 4005
SCSI host 1 abort (pid 1645) timed out - resetting
SCSI bus is being reset for host 1 channel 0.
qlogicpti : Resetting SCSI bus!
|
The "aborting command" stuff just repeats forever, and modprobe never returns.. the module just stays at "initializing"
Code: |
# lsmod
Module Size Used by Tainted: P
qlogicpti 30488 1 (initializing)
rtc 1892 0 (autoclean)
sunhme 19260 1
crc32 3264 0 [sunhme]
|
Any ideas? If it would just return from the init, I could poke at it more, but rebooting is getting old. Worse, the box is remote, and I only have sporadic physical access to it (though I do have remote serial console). |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Thu May 27, 2004 11:44 am Post subject: |
|
|
Hrm. How much do you like tracking down driver bugs and fixing them? |
|
Back to top |
|
|
hab136 n00b
Joined: 12 Apr 2004 Posts: 18 Location: Charlotte, NC
|
Posted: Thu May 27, 2004 12:43 pm Post subject: |
|
|
ciaranm wrote: | Hrm. How much do you like tracking down driver bugs and fixing them? |
*chuckle* not much.. I think first I'm going to try, in order:
physical changes (swapping disks, terminators, cabling, etc)
kernel 2.6 (probably 2.6.5-r1?)
then tracing down code bugs
It seems to me like it's trying to reset the bus, is unable to reset, so it aborts the reset command, is unable to abort, and so resets the bus.. repeat
Not sure if that's specific to qlogic or the SCSI subsystem, which is why I'll try 2.6 also.. compiling now on my little 170Mhz 64meg sparc 5.
I just figured if this was something totally obvious, someone would be able to say, "hey, have you checked $whatever?" |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Thu May 27, 2004 4:18 pm Post subject: |
|
|
It's a Qlogic driver bug, I'll tell you that much for free |
|
Back to top |
|
|
hab136 n00b
Joined: 12 Apr 2004 Posts: 18 Location: Charlotte, NC
|
Posted: Thu May 27, 2004 6:59 pm Post subject: |
|
|
ciaranm wrote: | It's a Qlogic driver bug, I'll tell you that much for free |
awesome.
Physically disconnecting one of the SCSI chains lets it modprobe just fine, although no disks showed up. (I think that particular chain has a bad disk, I'm not sure) Disconnecting both chains also works, unsurprisingly. I didn't have time to test disconnecting the non-suspect chain and leaving the suspect one in.
Code: |
qpti0: IRQ 53 SCSI ID 7 (Firmware v1.31.32)(Firmware 1.21 95/05/18) [Fast Wide, using differential interface]
qpti1: IRQ 53 SCSI ID 7 (Firmware v1.31.32)(Firmware 1.21 95/05/18) [Fast Wide, using differential interface]
QPTI: Total of 2 PTI Qlogic/ISP hosts found, 2 actually in use.
scsi1 : PTI Qlogic,ISP SBUS SCSI irq 53 regs at fd01b000
scsi2 : PTI Qlogic,ISP SBUS SCSI irq 53 regs at fd01c000
|
So now I have the controllers showing up, but not the disks. I tried rescan-scsi-bus.sh from http://www.garloff.de/kurt/linux/ but it didn't do anything. I'll play with it more tonight. |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|