Re: PR/46896 CVS commit: src/sys/dev/iscsi

["Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>]

The following reply was made to PR kern/46896; it has been noted by GNATS.

From: "Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: PR/46896 CVS commit: src/sys/dev/iscsi
Date: Sun, 9 Sep 2012 01:37:30 -0600 (MDT)

 On Sun, 9 Sep 2012, Martin Husemann wrote:
 
 > With this patch applied to -6 I still see pretty much repeatable locks up
 > (endless loop) here:
 
    After I applied that to -6, I started having trouble as well, although 
 in my case it seems to be loosing the session.  I went back to my -current
 kernel, and it also seems to have the same problem.
 
    My -6 has ISCSI_DEBUG set to 0 so I can easily patch the debug level, 
 and when I changed the level to 1, I got this:
 
 S2C1: CheckCmdSN resending - CmdSN=578093, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578094, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578095, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578096, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578097, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578098, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578099, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578100, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578101, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578102, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578103, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578104, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578105, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578106, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578107, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578108, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578109, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578110, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578111, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578112, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578113, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578114, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578115, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578116, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578117, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578118, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578119, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578120, ExpCmdSn=578092
 S2C1: CheckCmdSN resending - CmdSN=578121, ExpCmdSn=578092
 Received Reject PDU, reason = 4, data_len = 48
 Received Asynch PDU, Event 1
 ISCSI: kill_session 2, status 44, logout 0, recover 0
 S2C1: Write failed sock 0xffffa00002514f00 (ret: 32, req: 48, resid: 48)
 S2C1: *** Connection Error, status=18, logout=-1, state=3
 S2C1: Kill_connection: terminating=0, status=18, logout=-1, state=3
 Session 2 not found
 
    [This is with tagged queueing enabled, so I've got lots more 
 transactions going on.]
 
    I need to look at the more;  I'm not sure if I should back out my last 
 change, although setting the ccb->disp is not the correct thing at that 
 point (later code relies on ccb->disp to be <= CCBDISP_NOWAIT to add the 
 ccb to the ccbs_waiting queue).
 
 --
 Michael L. Hitch                       mhitch%montana.edu@localhost
 Computer Consultant
 Information Technology Center
 Montana State University       Bozeman, MT     USA