[Open-FCoE] Dump seen on target when initiater is enabled

Mithlesh mithlesh at linsyssoft.com
Tue Apr 8 10:48:22 UTC 2008


On Tuesday 08 April 2008 04:15:00 pm Mithlesh wrote:
> This problem was previously seen on 2.6.23 kernel and related FCoE setup
> which is described in the wikipage.
Sorry about bungle up. Forgot to mention that this time i am seeing this 
problem on the latest git on target as well as initiator side.


>
> Could it be the case that CBD which is copied from the incoming FC frame
> have incorrect data ? The "cmd" parameter which is being passed between the
> functions and also ends up in scst_parse_cmd() is one which is copied from
> FC frame.
>
> Also i am attaching the wireshark packets captured on initiator end.
>
> Regards,
> Mithlesh Thukral
>
> On Friday 04 April 2008 09:20:21 pm Zou, Yi wrote:
> > >fc_local_port_recv_flogi_req: FLOGI from port WWPN 20000007e90acfce
> > >[2835]: scst: scst_parse_cmd:541:***ERROR*** Expected data direction 0
> >
> > for
> >
> > >opcode 0x12 (handler dev_disk, target openfct) doesn't match decoded
> >
> > value
> >
> > >2
> > >Failed CDB:
> > > (h)___0__1__2__3__4__5__6__7__8__9__A__B__C__D__E__F
> > >   0: 12 00 00 00 24 00   ....$.
> >
> > I myself haven't seen this, wonder if anyone else has seen the similar
> > error? I also wonder what the last few packets on wireshark were when
> > this happened.
> >
> > >[root at oort ~]#
> > >
> > >On the initiator side the fcconf shows following output:
> > >
> > >[root at hogwarts ~]# fcconf target
> > >eth0 target ports:
> > >FC ID     WWPN                     State      Role
> > >010102    20:00:00:07:e9:0b:48:65  Online     FCP Target
> > >[root at hogwarts ~]# fcconf eth0
> > >Name          WWPN                      FC-ID   FC         Network
> > >eth0          20:00:00:07:e9:0a:cf:ce    10101  Online     UP
> > >[root at hogwarts ~]#
> > >
> > >But inspite of this i cant see the disks which are exported on target
> >
> > on
> >
> > >initiator.
> >
> > I assume you can still r/w to the target?
> >
> > >Is this some problem with the SCST or the target side code of FCoE ?
> >
> > I might be wrong but I feel it's in the target code. Still, we'll need
> > some further investigation on this.
> >
> > Thanks.
> > yi





More information about the devel mailing list