scan error after FC link recovery when device was in use

Hi,

With 2.6.17.2, if a FC device is in use (mounted FS), a scan error is reported when recovering after a 'long' link down period (>30s, 'no device timeout'). Then, devices aren't detected again (and oops are generated when reloading HBA drivers, but this may be a side effect).

We have this issue with point to point configuration using:
- Emulex + DDN
- Emulex + NEC
- QLogic + DDN
- QLogic + NEC

In all cases, we got the message "Unexpected response from lun 0 while scanning, scan aborted" in the scsi_report_lun_scan function (scsi_scan.c) and I'm thinking there's something wrong when calling scsi_probe_and_add_lun. We don't have such error if devices aren't used (not mounted) when link failure occurs.

Do you know if such issue has already been encountered ?
May be an already existing patch ? (I look for scsi_probe_and_add_lun in the mailling list but didn't find any revelant message)

Here are the extracts from syslog about these errors:

lpfc + DDN:
===========
...

Jul  3 09:15:30 s_kernel@iotiger2 kernel: lpfc 0000:06:02.0: 0:1305 Link Down Event x2 received Data: x2 x20 x0
Jul  3 09:16:00 s_kernel@iotiger2 kernel: lpfc 0000:06:02.0: 0:0203 Nodev timeout on WWPN 23:0:0:1:ff:3:2:a8 NPort xef Data: x8 x7 x0
Jul  3 09:16:05 s_kernel@iotiger2 kernel: rport-2:0-0: blocked FC remote port time out: removing target and saving binding
Jul  3 09:17:01 s_kernel@iotiger2 kernel: lpfc 0000:06:02.0: 0:1303 Link Up Event x3 received Data: x3 x1 x8 x2
Jul  3 09:17:01 s_kernel@iotiger2 kernel: Vendor: DDN       Model: S2A 8500 
     Rev: 5.22
Jul  3 09:17:01 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 03
Jul  3 09:17:01 s_kernel@iotiger2 kernel: SCSI device sdd: 571416576 512-byte hdwr sectors (292565 MB)
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sdd: Write Protect is off
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sdd: Mode Sense: a7 00 10 08
Jul  3 09:17:01 s_kernel@iotiger2 kernel: SCSI device sdd: drive cache: write back w/ FUA
Jul  3 09:17:01 s_kernel@iotiger2 kernel: SCSI device sdd: 571416576 512-byte hdwr sectors (292565 MB)
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sdd: Write Protect is off
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sdd: Mode Sense: a7 00 10 08
Jul  3 09:17:01 s_kernel@iotiger2 kernel: SCSI device sdd: drive cache: write back w/ FUA
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sdd: unknown partition table
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sd 2:0:0:0: Attached scsi disk sdd
Jul  3 09:17:01 s_kernel@iotiger2 kernel: sd 2:0:0:0: Attached scsi generic sg4 type 0
Jul  3 09:17:01 s_kernel@iotiger2 kernel: Vendor: DDN       Model: S2A 8500 
     Rev: 5.22
Jul  3 09:17:01 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 03
Jul  3 09:17:01 s_kernel@iotiger2 kernel: error 1
Jul  3 09:17:01 s_kernel@iotiger2 kernel: 2:0:0:0: Unexpected response from lun 0 while scanning, scan aborted

...

lpfc + NEC:
===========
...

Jul  4 06:11:44 s_kernel@iotiger2 kernel: EXT3 FS on sdg, internal journal
Jul  4 06:11:44 s_kernel@iotiger2 kernel: EXT3-fs: recovery complete.
Jul  4 06:11:44 s_kernel@iotiger2 kernel: EXT3-fs: mounted filesystem with ordered data mode.
Jul  4 06:12:03 s_kernel@iotiger2 kernel: lpfc 0000:06:02.1: 1:1305 Link Down Event x2 received Data: x2 x20 x0
Jul  4 06:12:33 s_kernel@iotiger2 kernel: lpfc 0000:06:02.1: 1:0203 Nodev timeout on WWPN 20:6:0:0:13:84:0:35 NPort xdc Data: x8 x7 x0
Jul  4 06:12:38 s_kernel@iotiger2 kernel: rport-3:0-0: blocked FC remote port time out: removing target and saving binding
Jul  4 06:12:42 s_kernel@iotiger2 kernel: lpfc 0000:06:02.1: 1:1303 Link Up Event x3 received Data: x3 x1 x8 x2
Jul  4 06:12:42 s_kernel@iotiger2 kernel: Vendor: NEC       Model: iStorage 2000 
    Rev: 2800
Jul  4 06:12:42 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 04
Jul  4 06:12:42 s_kernel@iotiger2 kernel: SCSI device sdh: 138412032 512-byte hdwr sectors (70867 MB)
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sdh: Write Protect is off
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sdh: Mode Sense: 97 00 00 08
Jul  4 06:12:42 s_kernel@iotiger2 kernel: SCSI device sdh: drive cache: write back
Jul  4 06:12:42 s_kernel@iotiger2 kernel: SCSI device sdh: 138412032 512-byte hdwr sectors (70867 MB)
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sdh: Write Protect is off
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sdh: Mode Sense: 97 00 00 08
Jul  4 06:12:42 s_kernel@iotiger2 kernel: SCSI device sdh: drive cache: write back
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sdh: unknown partition table
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sd 3:0:0:0: Attached scsi disk sdh
Jul  4 06:12:42 s_kernel@iotiger2 kernel: sd 3:0:0:0: Attached scsi generic sg12 type 0
Jul  4 06:12:42 s_kernel@iotiger2 kernel: Vendor: NEC       Model: iStorage 2000 
    Rev: 2800
Jul  4 06:12:42 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 04
Jul  4 06:12:42 s_kernel@iotiger2 kernel: error 1
Jul  4 06:12:42 s_kernel@iotiger2 kernel: 3:0:0:0: Unexpected response from lun 0 while scanning, scan aborted

...

QLogic+NEC
==========
...

Jul  4 09:47:10 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LIP reset occured (f8ef).
Jul  4 09:47:10 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LOOP DOWN detected (2).
Jul  4 09:47:14 s_kernel@iotiger2 kernel: qla2400 0000:07:01.0: LOOP DOWN detected (2).
Jul  4 09:47:45 s_kernel@iotiger2 kernel: rport-9:0-0: blocked FC remote port time out: removing target and saving binding
Jul  4 09:47:50 s_kernel@iotiger2 kernel: rport-8:0-0: blocked FC remote port time out: removing target and saving binding
Jul  4 09:48:11 s_kernel@iotiger2 kernel: qla2400 0000:07:01.0: LIP reset occured (f7f7).
Jul  4 09:48:11 s_kernel@iotiger2 kernel: qla2400 0000:07:01.0: LIP occured (f7f7).
Jul  4 09:48:11 s_kernel@iotiger2 kernel: qla2400 0000:07:01.0: LOOP UP detected (2 Gbps).
Jul  4 09:48:11 s_kernel@iotiger2 kernel: Vendor: NEC       Model: iStorage 2000 
    Rev: 2800
Jul  4 09:48:11 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 04
Jul  4 09:48:11 s_kernel@iotiger2 kernel: SCSI device sdi: 138412032 512-byte hdwr sectors (70867 MB)
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sdi: Write Protect is off
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sdi: Mode Sense: 97 00 00 08
Jul  4 09:48:11 s_kernel@iotiger2 kernel: SCSI device sdi: drive cache: write back
Jul  4 09:48:11 s_kernel@iotiger2 kernel: SCSI device sdi: 138412032 512-byte hdwr sectors (70867 MB)
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sdi: Write Protect is off
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sdi: Mode Sense: 97 00 00 08
Jul  4 09:48:11 s_kernel@iotiger2 kernel: SCSI device sdi: drive cache: write back
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sdi: unknown partition table
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sd 8:0:0:0: Attached scsi disk sdi
Jul  4 09:48:11 s_kernel@iotiger2 kernel: sd 8:0:0:0: Attached scsi generic sg14 type 0
Jul  4 09:48:11 s_kernel@iotiger2 kernel: Vendor: NEC       Model: iStorage 2000 
    Rev: 2800
Jul  4 09:48:11 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 04
Jul  4 09:48:11 s_kernel@iotiger2 kernel: error 1
Jul  4 09:48:11 s_kernel@iotiger2 kernel: 8:0:0:0: Unexpected response from lun 0 while scanning, scan aborted

...

QLogic+DDN
==========
...

Jul  4 10:01:31 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LOOP DOWN detected (2).
Jul  4 10:02:06 s_kernel@iotiger2 kernel: rport-9:0-0: blocked FC remote port time out: removing target and saving binding
Jul  4 10:03:05 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LIP reset occured (f7f7).
Jul  4 10:03:05 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LIP occured (f7f7).
Jul  4 10:03:05 s_kernel@iotiger2 kernel: qla2400 0000:07:01.1: LOOP UP detected (2 Gbps).
Jul  4 10:03:05 s_kernel@iotiger2 kernel: Vendor: DDN       Model: S2A 8500 
     Rev: 5.22
Jul  4 10:03:05 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 03
Jul  4 10:03:05 s_kernel@iotiger2 kernel: SCSI device sdl: 571416576 512-byte hdwr sectors (292565 MB)
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sdl: Write Protect is off
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sdl: Mode Sense: a7 00 10 08
Jul  4 10:03:05 s_kernel@iotiger2 kernel: SCSI device sdl: drive cache: write back w/ FUA
Jul  4 10:03:05 s_kernel@iotiger2 kernel: SCSI device sdl: 571416576 512-byte hdwr sectors (292565 MB)
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sdl: Write Protect is off
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sdl: Mode Sense: a7 00 10 08
Jul  4 10:03:05 s_kernel@iotiger2 kernel: SCSI device sdl: drive cache: write back w/ FUA
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sdl: unknown partition table
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sd 9:0:0:0: Attached scsi disk sdl
Jul  4 10:03:05 s_kernel@iotiger2 kernel: sd 9:0:0:0: Attached scsi generic sg15 type 0
Jul  4 10:03:05 s_kernel@iotiger2 kernel: Vendor: DDN       Model: S2A 8500 
     Rev: 5.22
Jul  4 10:03:05 s_kernel@iotiger2 kernel: Type:   Direct-Access 
     ANSI SCSI revision: 03
Jul  4 10:03:05 s_kernel@iotiger2 kernel: error 1
Jul  4 10:03:05 s_kernel@iotiger2 kernel: 9:0:0:0: Unexpected response from lun 0 while scanning, scan aborted

...

Frederic TEMPORELLI

This is a known side effect of removing a fibre target which has references. There is a crude workaround available. Essentially, it modifies the fc transport to not delete the target when the timer expires.

http://marc.theaimsgroup.com/?l=linux-scsi&m=115015423722568&w=2

Root cause is being investigated by various interested parties.

Be sure to read the entire thread. This patch is not going to be accepted into the kernel.

More interesting threads:

http://marc.theaimsgroup.com/?l=linux-scsi&m=114979410315677&w=2
http://marc.theaimsgroup.com/?l=linux-scsi&m=115039021604020&w=2
http://marc.theaimsgroup.com/?l=linux-scsi&m=115039057504409&w=2
http://marc.theaimsgroup.com/?l=linux-scsi&m=115134812215106&w=2

Be sure to read the complete thread before applying any of the patches....

Mike