Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <cd119ce943d9ec62ef1bff237ebb49e35a337c3b.1589407872.git.lance.digby@gmail.com>
 <93c437ce-f881-9f54-5e39-afa8afd96141@redhat.com>
In-Reply-To: <93c437ce-f881-9f54-5e39-afa8afd96141@redhat.com>
From:   Lance Digby <lance.digby@gmail.com>
Date:   Sun, 17 May 2020 09:29:38 +1000
Message-ID: <CAHh6B+poeCFeZjCt04QT2c12bgvCf9UN+5vdx5EeK6PcgD2y3g@mail.gmail.com>
Subject: Re: [PATCH target] target: Add initiatorname to NON_EXISTENT_LUN error
To:     Mike Christie <mchristi@redhat.com>
Cc:     martin.petersen@oracle.com, linux-scsi@vger.kernel.org,
        target-devel@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Mike,  Thanks for the review!
  The pr_err  Detected NON_EXISTENT_LUN is the error messages issued
for the TCM_NON_EXISTENT_LUN retcode so I believe they are the same.
  Simply scanning for the wrong lun on an initiator will generate this
error on the target but not generate an error on the initiator. And I
have seen installs, with a lot of initiators, automate the scanning of
such luns incorrectly deemed missing.
  While this looks like a simple problem it can take days to get
access or the tcp traces to sort it out.

   Within the same routine there is another pr_err for
TCM_WRITE_PROTECTED that I did not add the initiatorname to as I
thought this would leave a heavy footprint on the initiator. If you
believe this should be changed for consistency please let me know and
I will add this and change to nacl->initiatorname.


On Sat, May 16, 2020 at 9:50 AM Mike Christie <mchristi@redhat.com> wrote:
>
> On 5/13/20 11:01 PM, Lance Digby wrote:
> > The NON_EXISTENT_LUN error can be written without an error condition
> >  on the initiator responsible. Adding the initiatorname to this message
> >  will reduce the effort required to fix this when many initiators are
> > supported by a target.
> >
> > Signed-off-by: Lance Digby <lance.digby@gmail.com>
> > ---
> >  drivers/target/target_core_device.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c
> > index 4cee113..604dea0 100644
> > --- a/drivers/target/target_core_device.c
> > +++ b/drivers/target/target_core_device.c
> > @@ -100,9 +100,10 @@
> >                */
> >               if (unpacked_lun != 0) {
> >                       pr_err("TARGET_CORE[%s]: Detected NON_EXISTENT_LUN"
> > -                             " Access for 0x%08llx\n",
> > +                             " Access for 0x%08llx from %s\n",
> >                               se_cmd->se_tfo->fabric_name,
> > -                             unpacked_lun);
> > +                             unpacked_lun,
> > +                             se_sess->se_node_acl->initiatorname);
>
> You can do nacl->initiatorname.
>
> Do you also want add the name to the tmr case? It's probably not common,
> but the error message would be consistent.
>
> >                       return TCM_NON_EXISTENT_LUN;
> >               }
> >
>