Problems getting target mode to work with isp driver

Mon Nov 1 21:19:00 UTC 2010

Hi Matthew,

I attached the Freebsd target and the linux initiator to a Brocade silkworm
3200. Additionally i have moved the 24xx card to the freebsd target. The
current setup on the target and initiator is
freebsd (8.1) -> 2340 and 2460 hba's
linux (RHEL) -> 2342 HBA

Device hints used are
hint.isp.0.fullduplex="1"
hint.isp.0.iid=75

With the switch configuration target mode for 24xx works nicely. No issues
so far. However point to point doesn't work with the 24xx target too.

Using the 23xx card for the target i have seen device timeouts on the
initiator side after which target mode stops working. I have yet to get a
scenario where  i can simulate the problem consistently. Attached
(iniitatorerrors.txt) contains the errors reported on the linux side. On the
FreeBSD side no errors were reported.

I did find the following warnings reported by the target (both with and with
the switch configuration) after every port logout (done by removing the
qla2xxx linux driver) by the initiator
Oct 31 20:00:13 scache kernel: isp0: isp_del_wwn_entry: Chan 0 IID
0x2100001b320f3212 N-Port Handle 0x0000 Port ID 0xffffff cannot be found to
be
cleared
Oct 31 20:00:13 scache kernel: isp0: Chan 0 [0]: hdl 0x0 PROB al0 tgt ---
INI
0x011000 => UNK 0x000000; WWNN 0x2000001b320f3212 WWPN 0x2100001b320f3212
Oct 31 20:00:13 scache kernel: isp0: Chan 0 [126]: hdl 0x7e PROB al0 tgt ---
UNK 0xfffffe => UNK 0xfffffe; WWNN 0x1000006069c0766f WWPN
0x2002006069c0766f
Oct 31 20:00:13 scache kernel: isp0: Chan 0 [511]: hdl 0x0 NIL  al0 tgt ---
UNK 0xffffff => UNK 0x000000; WWNN 0x0000000000000000 WWPN
0x2100001b320f3212
Oct 31 20:00:57 scache kernel: isp0: isp_del_wwn_entry: Chan 0 IID
0x2100001b320f3212 N-Port Handle 0x0000 Port ID 0xffffff cannot be found to
be
cleared
Oct 31 20:00:57 scache kernel: isp0: Chan 0 [0]: hdl 0x0 PROB al0 tgt ---
INI
0x011000 => UNK 0x000000; WWNN 0x2000001b320f3212 WWPN 0x2100001b320f3212
Oct 31 20:00:57 scache kernel: isp0: Chan 0 [126]: hdl 0x7e PROB al0 tgt ---
UNK 0xfffffe => UNK 0xfffffe; WWNN 0x1000006069c0766f WWPN
0x2002006069c0766f

Is there anything else i could try out ?

Thanks,
Penta

On Fri, Oct 29, 2010 at 1:52 AM, Penta Upa <bsdboot at gmail.com> wrote:

>
>
> On Thu, Oct 28, 2010 at 5:00 AM, Matthew Jacob <mj at feral.com> wrote:
>
>> Since this was all point to point, my guess is that it's a soft loop id
>> issue.
>>
>> You can try your original setup again (linux to FreeBSD), but for the
>> FreeBSD box add this to loader.conf:
>>
>>    set hint.isp.0.iid=75
>>
>>
> target mode doesn't work by setting the above hint in the loader.conf.
> (even with the solaris box) The earlier linux setup doesn't work. I also
> tried with another linux box (RHEL 5.5) which had both qla2340 and qla2460
> cards and with windows 2003 (which is the same as the solaris box). But they
> dont work with or without the hint. Once i removed the hint from the
> loader.conf, solaris works but not linux or windows
>
> I'll attach the the initiators and target to a switch over the weekend and
> narrow it down to a point to point issue. Thanks a lot for the help and many
> apologies for the delay in response. I'm trying out target mode as a
> personal project which i do after work.
>
> Thanks,
> Penta
>
-------------- next part --------------
Nov  2 02:27:16 scache1 kernel: qla2xxx 0000:06:01.1: LOOP UP detected (2 Gbps).
Nov  2 02:27:17 scache1 kernel:   Vendor: FreeBSD   Model: Emulated Disk     Rev: 0.1 
Nov  2 02:27:17 scache1 kernel:   Type:   Direct-Access                      ANSI SCSI revision: 03
Nov  2 02:27:17 scache1 kernel: SCSI device sdg: 204800 512-byte hdwr sectors (105 MB)
Nov  2 02:27:17 scache1 kernel: sdg: test WP failed, assume Write Enabled
Nov  2 02:27:17 scache1 kernel: sdg: cache data unavailable
Nov  2 02:27:17 scache1 kernel: sdg: assuming drive cache: write through
Nov  2 02:27:17 scache1 kernel: SCSI device sdg: 204800 512-byte hdwr sectors (105 MB)
Nov  2 02:27:17 scache1 kernel: sdg: test WP failed, assume Write Enabled
Nov  2 02:27:17 scache1 kernel: sdg: cache data unavailable
Nov  2 02:27:17 scache1 kernel: sdg: assuming drive cache: write through
Nov  2 02:27:17 scache1 kernel:  sdg: unknown partition table
Nov  2 02:27:17 scache1 kernel: sd 9:0:0:0: Attached scsi disk sdg
Nov  2 02:27:17 scache1 kernel: sd 9:0:0:0: Attached scsi generic sg6 type 0
Nov  2 02:28:19 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): Abort command issued -- 1 19 2002.
Nov  2 02:28:30 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): Abort command issued -- 1 19 2002.
Nov  2 02:28:30 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): DEVICE RESET ISSUED.
Nov  2 02:28:30 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): DEVICE RESET SUCCEEDED.
Nov  2 02:28:35 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): LOOP RESET ISSUED.
Nov  2 02:28:36 scache1 kernel: qla2xxx 0000:06:01.1: qla2xxx_eh_bus_reset: reset succeeded
Nov  2 02:28:46 scache1 kernel: qla2xxx 0000:06:01.1: scsi(9:0:0): ADAPTER RESET ISSUED.
Nov  2 02:28:46 scache1 kernel: qla2xxx 0000:06:01.1: Performing ISP error recovery - ha= ffff81006e1b04f8.
Nov  2 02:28:46 scache1 kernel: qla2xxx 0000:06:01.1: LIP reset occured (f7f7).
Nov  2 02:28:47 scache1 kernel: qla2xxx 0000:06:01.1: LOOP UP detected (2 Gbps).
Nov  2 02:28:47 scache1 kernel: qla2xxx 0000:06:01.1: qla2xxx_eh_host_reset: reset succeeded
Nov  2 02:28:57 scache1 kernel: sd 9:0:0:0: scsi: Device offlined - not ready after error recovery
Nov  2 02:28:57 scache1 kernel: sd 9:0:0:0: rejecting I/O to offline device
Nov  2 02:28:57 scache1 kernel: Buffer I/O error on device sdg, logical block 25592
Nov  2 02:28:57 scache1 kernel: sd 9:0:0:0: rejecting I/O to offline device