RE: iSCSI 05.txt is out

[julian_satran@il.ibm.com]

I was afraid that that will be the answer -:).

If nobody objects we will make the Target Task Tag use (alone or with LUN)
an option at Login (LO)
with the default being "alone".

Regards,
Julo

"Venkat Rangan" <venkat@rhapsodynetworks.com> on 06/03/2001 18:56:36

Please respond to "Venkat Rangan" <venkat@rhapsodynetworks.com>

To:   Julian Satran/Haifa/IBM@IBMIL
cc:
Subject:  RE: iSCSI 05.txt is out




Julian,

Thanks for your responses.

I've isolated a couple of your responses for clarity.

+++ No need - length does not include padding +++

This is problematic because now digests will have to be run over a
byte-stream and not a word-stream. My ASIC designer tells me this is a "bid
deal". I assume digests cover only the length and not the padding. Again,
Fibre Channel does it correctly, by having the length include the pad and
the digest include the pad and the fill-bytes indicate to ULP any padding.

+++We can make checking that optional although I don't understand what is
all
the fuss about it+++

The issue is that in some implementations, the target exchanges are
maintained by context tables. Once the Command is processed, a context
table
entry is set up. There is no need to record the LUN in the context table.
But the iSCSI requirement that the LUN in WRITE should match the LUN in
Command requires the LUN also to be recorded. LUN is a large field, and in
our implementation, we manage some 10,000 active exchanges. Adding 8-byte
overhead just for catching an errant initiator seems wasteful. If the
checking and reporting is Optional, i.e., change MUST to MAY, will ease
this
burden, and still keeps the mechanism available for targets that may make
use of it.

+++ even when received in order the initiator may decide to fulfill them
out of order - this statement prevents it+++

I absolutely agree with your statement. But what I'm objecting to is the
unnecessary phrase "Within a connection, " It should be removed.

Regards,

-venkat


-----Original Message-----
From: julian_satran@il.ibm.com [mailto:julian_satran@il.ibm.com]
Sent: Monday, March 05, 2001 8:03 PM
To: Venkat Rangan
Subject: RE: iSCSI 05.txt is out




Answers in text - +++

Thanks for the careful reading,
Julo

"Venkat Rangan" <venkat@rhapsodynetworks.com> on 06/03/2001 00:40:18

Please respond to "Venkat Rangan" <venkat@rhapsodynetworks.com>

To:   Julian Satran/Haifa/IBM@IBMIL, ips@ece.cmu.edu
cc:
Subject:  RE: iSCSI  05.txt is out




Julian,

Here are a few technical and editorial comments on the latest draft.

Technical:

Page 12:

- CmdSN - the current Command Sequence Number advanced by 1 on each command
shipped

Add: skipping zero, which is reserved for immediate delivery.
+++ will fix +++
Page 12:
Remove the sentence:
"A large difference between StatSN and ExpStatSN may indicate a failed
connection."

This may have made sense when StatSN and ExpStatSN are session-wide. Now
that it is per-connection, this may not be valid. As was pointed out in the
"Holes in StatSN" thread, ExpStatSN may be "stuck" at a low value when a
single PDU encountered a Digest Error, followed by several well-formed
PDUs.
+++ if it cant't be recovered with a SACK it is stll a sign of error+++
Page 24 - Section 2.1
Add: The length of the PDU SHALL include any padding bytes added.
+++ No the length does not include padding - padding is implicit +++
This raises a question as to whether it may be useful to include a two-bit
"Fill Bytes" field in the header (BHS and AHS) which indicates the number
of
bytes that were added. Without this, it may be harder for the receiver to
know how many bytes are part of the data. Fibre Channel has the Fill Byte
specifier in its F_CTL part of the header.
+++ No need - length does not include padding +++
Page 25:
This may be covered in another thread.

What does the first Next-Qualifier (at byte 0) describe? Is it the BHS or
the header following the BHS? If a PDU has a single BHS and a single AHS,
is
the sequence WN-BHS-AHS or is it WN-BHS-WN-AHS?
+++ in your example the second answer is correct and the second WN descibes
the data payload +++
Page 25 and 26:
The text next to the bit description and the headers in the text for
sections 2.2.2.1 to 2.2.2.3 appear to be inconsistent.  An example is
"read-data transfer" vs. "Read Data" "Long Data" vs. "Long Data Transfer"

Also, 2.2.2.3 appears to have text about Long Data Header that probably
belongs in section 2.2.2.2 (or I am confused about how WN is supposed to
work.)
+++ will make consistent +++
Page 43
Section 2.7
The PDU diagrams should not show "Digests if any..." etc. It should be
covered by NQ and corresponding AHS.
+++ correct by tastes vary - and many people prefer a complete picture - I
have a hard time now trying to find a way to reconcile all +++

Page 44
Section 2.7.2
The requirement of having to specify a valid LUN as part of WRITE Data (and
NOP-Out) may pose unnecessary overhead for Target processing. The fact that
Targets are now REQUIRED to reject errant initiators who may place a LUN
inconsistent with the one sent in the CommandPdu means that targets should
save the LUN against each context entry for the task. This was discussed
earlier, and I think we said that unless the folks who originally wanted it
spoke up, we will remove it. Note that Fibre Channel does not have this
feature, and its original purpose can be accomplished by using Target
Transfer Tag.

If we still want it, targets should be given the option of negotiating this
so that it can operate in a mode where a LUN specified as part of WRITE
data
will be ignored (as opposed to rejected).
+++ Several people on this list reuested the freedom to build the tags at
the device-server (part of the LU according to SAM) without regard to what
other LUs the target has
The simplest way to accomodate them is to have the initiator direct the dat
through a LUN
We can make checking that optional although I don't understand what is all
the fuss about it+++
Section 2.17
The following text:
"Within a connection, outstanding R2Ts MUST be fulfilled by the initiator
in
the order in which they were received."

implies that R2Ts for the same WRITE command may be sent by the target on
multiple connections? Since this is not possible, all R2Ts for a command
are
always on a single connection, so I am not sure how the above sentence
should be interpreted.

+++ even when received in order the initiator may decide to fulfill them
out of order - this statement prevents it+++

Page 75:
Add: "i.e., from Most Preferred to Least Preferred" to "reverse order of
preference."
+++ why? is there something ambiguous in the statement? ++
------------
Editorial - typos and such

1. Page 11:

Not covered are he means

Change "he" to "the"

+++ will fix ++
2. Page 20:

Change "later" to "latter" in "the later can be used in subsequent
commands."
+++ will fix ++
3. Page 23:

Change "isa"  to "is a"
+++ will fix ++
Page 42:
Section 2.6.1
Change to: Initiator Task Tag of the task that was not found; used in
conjuction with Response value 1. It MUST be set to zero in other cases.

+++ will fix +++
Page 51:
Change "CID does not change and this command is performs first" to

"CID does not change and this command first performs"
+++ will fix +++
Page 55:
There is nothing that indicates the status codes in the table are sent as
part of Status-Detail.
+++ it says above the table +++
Page 66:
Change "iSSCSI Data-out PDU" to "iSCSI Data (WRITE) PDU"

The Data-out PDU is a new concept not introduced anywhere.
+++ will fix +++
Page 75:
The term "Security Context Complete" is referred to prior to its
definition.

Page 81:
Change "later" to "latter"
++ will fix +++
------------

Venkat Rangan
Rhapsody Networks Inc.
http://www.rhapsodynetworks.com