|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] RE: iSCSI - first working group Last Call round issues list and resolutionsHi all, While the attachment probably went thru for most people, David and I wanted to resend as inline text, to make sure everyone sees this. This is the list of comment resolutions for comments from the first iSCSI WG last call. It is up to date thru changes made on July 18. People should review this list, especially those who made iSCSI WG last call comments. The working version of the next version of the draft is available at Julian's web site at http://www.haifa.il.ibm.com/satran/ips People should be reviewing the deltas to this document, and bringing up issues with any modifications to this document. There are still a few issues that are still being addressed, and these are noted in this document. The list of issues and resolutions follow. Thanks Elizabeth iSCSI - WG - Last Call - Issues and Resolution # Description Resolution ---+----------------------------------------------------------+--------- -------------------- 1 | Wording in 4.2 for empty Text Request/Response | Changed wording ---+----------------------------------------------------------+--------- -------------------- 2 | wording for waiting for data in 9.17 | Added words ---+----------------------------------------------------------+--------- -------------------- 3 | wording for connection clearing after retry in 2.2.2.1 | Changed wording ---+----------------------------------------------------------+--------- -------------------- 4 | link command complete mappes also to 0x00 in 9.4.3 | added it to mapping ---+----------------------------------------------------------+--------- -------------------- 5 | Stray reference to IPV6 dotted decimal on page 250 | removed ---+----------------------------------------------------------+--------- -------------------- 6 | Text in 4.3 and 4.4 not allowing SendTargets | fixed text ---+----------------------------------------------------------+--------- -------------------- 7 | Clarification on COLD RESET - required by SAM | fixed text ---+----------------------------------------------------------+--------- -------------------- 8 | 9.5.4 Initiator (Task Tag) to be replaced by Referenced | fixed text ---+----------------------------------------------------------+--------- -------------------- 9 | 9.5 recommendation on empty data inconsistent with R2T | fixed text ---+----------------------------------------------------------+--------- -------------------- 10 | 2.2.2.3 and 9.8.1 and 9.7 text numbering data/r2t | fixed text ---+----------------------------------------------------------+--------- -------------------- 11 | 5.2 Text not clear about connection logout mandated | added text ---+----------------------------------------------------------+--------- -------------------- 12 | 9.4.6.2 text reffers only to firstburstsize | changed text to "incorrect | | amount of data" ---+----------------------------------------------------------+--------- -------------------- 13 | Both Length and Size used in text - requested use of | Changed Size in text | of only one of them | to Length ---+----------------------------------------------------------+--------- -------------------- 14 | Not clear how to number retransmitted Data-In | Added text to 9.16 ---+----------------------------------------------------------+--------- -------------------- 15 | Concluding text in 4.3.2 consideered unclear | Slightly changed text ---+----------------------------------------------------------+--------- -------------------- 16 | Say Reject PDU when the intent is Reject PDU | Added PDU when Reject PDU was | | meant and lower case reject ---+----------------------------------------------------------+--------- -------------------- 17 | Reinstate I bit in text request (typo) | Fixed figure ---+----------------------------------------------------------+--------- -------------------- 18 | C bit relation to T and F bit "MUST" not "must" | Fixed text ---+----------------------------------------------------------+--------- -------------------- 19 | I bit on a Text negotiation the same | Added to text ---+----------------------------------------------------------+--------- -------------------- 20 | Lingering reference to referenced task tag in 2.5.1.4 | Removed ---+----------------------------------------------------------+--------- -------------------- 21 | StatSN is retransmitted R2T should be the new value | fixed text in 9.16 ---+----------------------------------------------------------+--------- -------------------- 22 | REf. Task Tag to replace ITT in 9.5.4 and 9.6.1 | fixed text ---+----------------------------------------------------------+--------- -------------------- 23 | Old reference to ACL | Removed ---+----------------------------------------------------------+--------- -------------------- 24 | Confusion about digest position | Spell-out digest names ---+----------------------------------------------------------+--------- -------------------- 25 | Async Message - Logout Request timer handling | Fixed in appendix E ---+----------------------------------------------------------+--------- -------------------- 26 | Numeric instead of Numerical in Appendix A | Fixed text ---+----------------------------------------------------------+--------- -------------------- 27 | Error in DefaultTime2Wait spec | Fixed text and changed format ---+----------------------------------------------------------+--------- -------------------- 28 | Correction of DefaultTime2Wait and reformat | Fiexed and reformated ---+----------------------------------------------------------+--------- -------------------- 29 | Security aligned with security draft | Aligned ---+----------------------------------------------------------+--------- -------------------- 30 | Command queue missunderstanding | Fixed text in 6.1.2 ---+----------------------------------------------------------+--------- -------------------- ======================================================================== ==================== Draft 13 to 14 - Watershed ======================================================================== ==================== ---+----------------------------------------------------------+--------- -------------------- 30 | Requirement to support 16kB total key=value text | Agreed to 8k limit | considered excessive - reduction (to 2k?) proposed | ---+----------------------------------------------------------+--------- -------------------- 31 | 64 bit decimals considered difficult | leave as is ---+----------------------------------------------------------+--------- -------------------- 32 | Decimal encoded binary are harmfull | an entrenched evil | | restrict them to given length ---+----------------------------------------------------------+--------- -------------------- 33 | Combined device (target and IPsec) made explicit | Added text to 7.3 ---+----------------------------------------------------------+--------- -------------------- 34 | TargePortalGroupTag is a 16 bit binary - not numeric | Fixed text in text to 11.9 ---+----------------------------------------------------------+--------- -------------------- 35 | IPsec implementation in a "combined-device" has to be | Spelled out in 7.3 | spelled out | ---+----------------------------------------------------------+--------- -------------------- 36 | Boolean functions not spelled out | Spelled out in 11 ---+----------------------------------------------------------+--------- -------------------- 37 | Boolean functions not spelled out | Spelled out in 11 ---+----------------------------------------------------------+--------- -------------------- 38 | Remove BiDiR2T key | --- ---+----------------------------------------------------------+--------- -------------------- 39 | Offset of Immediate data | spelled out in 2.2.4 ---+----------------------------------------------------------+--------- -------------------- 40 | 11.8 Portal Group Tag ommited/required | made it MUST ---+----------------------------------------------------------+--------- -------------------- ======================================================================== ==================== DLB's issues with their own numbering ======================================================================== ==================== ---+----------------------------------------------------------+--------- -------------------- T1 | 2.2.2.2 recovery MUST be undertaken | fixed text ---+----------------------------------------------------------+--------- -------------------- ---+----------------------------------------------------------+--------- -------------------- T2 | 2.2.6.1 target name may be ignored- replace with | fixed text | MAY be omitted | ---+----------------------------------------------------------+--------- -------------------- T3 | 2.2.6.1 upper case MUST in name requirements | fixed text | and in encoding | ---+----------------------------------------------------------+--------- -------------------- T4 | 2.2.6.3.1 .iqn - needs date unammbiguous | fixed based on DLB ---+----------------------------------------------------------+--------- -------------------- T5 | Synch and Steering bloated - for what we choose | fixed text | WHAT a WASTE of EFFORT! | ---+----------------------------------------------------------+--------- -------------------- T6 | Clarify text in 2.3 about discovery | Changed MAY accept to | | equivalent MUST ONLY and | | spelled-out rejects | | MUST stays - nothing fancy | | in discovery ---+----------------------------------------------------------+--------- -------------------- T7 | Portal defined by IP address does work through NAPT | It is not a protocol | | element - rather a model | | entity. In case of NAPT the | | model entity name may change ---+----------------------------------------------------------+--------- -------------------- T8 |2-concerns - the underlying issue stems from an obsolete | Removed port specific mode | port specific mode page notion - iSCSI has none | replaced text with session | parameters | params and removed 2.4.3.2 ---+----------------------------------------------------------+--------- -------------------- T9 |SCSI port name inapropriate | The encoding is meant for SCSI | | and it has no padding in the | | middle - CHANGED ENCODING ---+----------------------------------------------------------+--------- -------------------- T10|2.4.3.2 - should not uppercased | 2.4.3.2 is obsolete ---+----------------------------------------------------------+--------- -------------------- T11| Decimal to binary | limited to numbers that are | | allowed values less than 2**64 | | and bitstring with defined | | length less than 2**64 ---+----------------------------------------------------------+--------- -------------------- T12| large-numerical-value does not cover lower than 2**64 | large are on purpose | | different in order to restrict | | them ---+----------------------------------------------------------+--------- -------------------- T13|=30 and clarify when 64k has to be supported | fixed text for 64k support ---+----------------------------------------------------------+--------- -------------------- T14| Declaration not explained | added a statement to 4.2 ---+----------------------------------------------------------+--------- -------------------- T15| Make TPGT return on login MUST always | MUST sticks ---+----------------------------------------------------------+--------- -------------------- T16| Second connection issue | I did not use SHOULD as | | there are recovery scenarios | | that do not need a second | | connection. | | MUST in 4.3.4 | | Rest of text is consistent ---+----------------------------------------------------------+--------- -------------------- T17| Format Error - 6.4 | Examples are out of place | | and have been removed | | legal values are specified | | in chapter 9 ---+----------------------------------------------------------+--------- -------------------- T18| Text on abort after an ULP timeout | Fixed text. There is only | | one specific Abort Task ---+----------------------------------------------------------+--------- -------------------- T19| Conservative reuse in 8.1.1 SHOULd is appropriate | changed to SHOULD | | SCSI reference already i ---+----------------------------------------------------------+--------- -------------------- T20| MUST support and use autosense in 8.2 | fixed text ---+----------------------------------------------------------+--------- -------------------- T21| Guidance for timeouts in 8.3 set too low | raised ---+----------------------------------------------------------+--------- -------------------- T22| Padding replace SHOULD with MUST be sent as 0 | why is that better? | | receiver always ignores pad | | testing is implicit in CRC ---+----------------------------------------------------------+--------- -------------------- T23|o, u etc not exclusive is a protocol error | made it explicit ---+----------------------------------------------------------+--------- -------------------- T24| Residual Counts should be reserved when not valid | fixed text ---+----------------------------------------------------------+--------- -------------------- T25| Task reassign may need LUN | No - we dont want checks if | | we can avoid them ---+----------------------------------------------------------+--------- -------------------- T26| Add Task Reassign to list of responses | Fixed ---+----------------------------------------------------------+--------- -------------------- T27| ExpStatSN spell out for non-first | spelled out | Additional concern CmdSN | Login is immediate ---+----------------------------------------------------------+--------- -------------------- T28| Concern about discarding | Discarding refers only to | | REORDERING QUEUE | | changed wording ---+----------------------------------------------------------+--------- -------------------- T29| Logout request request refers to implict login | made the text clearer and | | added a cross reference from | | 9.12.8 ---+----------------------------------------------------------+--------- -------------------- T30| Resegmenting may come as a surprize - suggested a | New text for resegmentation | different request | 9.4 and 9.16 ---+----------------------------------------------------------+--------- -------------------- T31| Operational Error Level instead of supported | Fixed text ---+----------------------------------------------------------+--------- -------------------- T32| Vendor Specific Authentication contradictory statement | Fixed text in 10 ---+----------------------------------------------------------+--------- -------------------- T33| MD5 SHOULD be offered change to MUST for interop. | made MUST ---+----------------------------------------------------------+--------- -------------------- T34| Digests CRC MUST be offered | fixed text ---+----------------------------------------------------------+--------- -------------------- T35| Target and Initiator name not changed | There is a general restriction | | about restating a key except | | when allowed and those are not | | allowed - I added text anyhow ---+----------------------------------------------------------+--------- -------------------- T36| TargetAlias/InitiatorAlias warning | Only a weak one possible as | | it can't be enforced ---+----------------------------------------------------------+--------- -------------------- T37| Unsolicited data inclarity | Already speced in 2.2.4 ---+----------------------------------------------------------+--------- -------------------- T38| IANA text | Fixed - why is that technical ---+----------------------------------------------------------+--------- -------------------- T39| IANA registry for keys | Proposed formulation for | | vendor keys and options | | Read carefully ---+----------------------------------------------------------+--------- -------------------- ======================================================================== ==================== ER issues with their own numbering ======================================================================== ==================== ---+----------------------------------------------------------+--------- -------------------- E1 | Version number out of place | There was a complaint | | it is hidden and it should | | the next place may the title ---+----------------------------------------------------------+--------- -------------------- E2 | use of may in the text on pag 36 3rd paragraph | even after reareading author | | would not use MAY and MUST ---+----------------------------------------------------------+--------- -------------------- E3 | 2.2.2.2 replace SHOULD not exceed 2**31-1 with MUST | fixed text ---+----------------------------------------------------------+--------- -------------------- E4 | wordsmithing the second paragraph on pg. 41 | may is used here as "can do" | | and the explanation is | | accurate ---+----------------------------------------------------------+--------- -------------------- E5 | EUI64 is not used by FC directly | fixed text in 2.2.6 ---+----------------------------------------------------------+--------- -------------------- E6 | Caution on EUI64 used as iSCSI names are not tied to | Added text to 8.1.2 | hardware | ---+----------------------------------------------------------+--------- -------------------- E7 | T and F used in 4 without being defined | defined in introduction to 4 ---+----------------------------------------------------------+--------- -------------------- E8 | Number figure and tables | no before last call - tool ---+----------------------------------------------------------+--------- -------------------- E9 | move 5.1.3 & to fron (5.1.1 & 2) | will give it a try! | | it is not consistent though! ---+----------------------------------------------------------+--------- -------------------- E10| 6.1.2 SHOULD and MAY issue | the intent is to avoid | |data and better avoid dropping! | |removed MAY; wording MAY | |contradicts SHOULD ---+----------------------------------------------------------+--------- -------------------- E11| Optional to OPTIONAL in 6.1.2 |fixed text ---+----------------------------------------------------------+--------- -------------------- E12| MUST etc. should refer to case when used | Added text at the start | | of 6.12 ---+----------------------------------------------------------+--------- -------------------- E13| MUST, MAY in 7.2 | fixed text although | | unequivocal even before ---+----------------------------------------------------------+--------- -------------------- E14| 8.1.1 should -> SHOULD | already there ---+----------------------------------------------------------+--------- -------------------- E15| Conservative Reuse - change to SHOULD | it was agreed as recommended | | no RECOMMENDED ---+----------------------------------------------------------+--------- -------------------- E14| Configurable SHOULD in 8.1.2 | fixed text ---+----------------------------------------------------------+--------- -------------------- E15| Vendor MUST allow ISID coordination | fixed text ---+----------------------------------------------------------+--------- -------------------- E16| Why padding is SHOULD be 0 nad not MUST? | not strictly required | | some strictly secure things | | may want them random! ---+----------------------------------------------------------+--------- -------------------- E17| Initiator and Target note? | In iSCSI they must be 2 | | entities ---+----------------------------------------------------------+--------- -------------------- E18| 9.4 MUST contain sense data | fixed text ---+----------------------------------------------------------+--------- -------------------- E19| Add wording to stress the fact that reset is on all LUs | Added wording on 9.5 that | | reset is accross all LUs | | known to the initiator ---+----------------------------------------------------------+--------- -------------------- E20| Wording that the A bit MUST not be set for ErrorRecovery| Why? The alternative could | Level is 0 | be bit is set to 0 by target | | and ignored by I (no test!) ---+----------------------------------------------------------+--------- -------------------- E21| 9.11.1 why MUST NOT ... that may | text is correct - other | | wording perhaps? ---+----------------------------------------------------------+--------- -------------------- E22| Make a single section for version | well if it looks better for | | at least one reader :-) ---+----------------------------------------------------------+--------- -------------------- E23| 9.14.5 make internal check must not MUST | changed text to indicate | | command ordering is aimed ---+----------------------------------------------------------+--------- -------------------- E24| 9.16.1 SNACK relation to ErrorRecoveryLevel in 9.16 | State that all SNACK have | | to be supported if | | if ErrorRecoveryLevel>0 ---+----------------------------------------------------------+--------- -------------------- E25| Reject Status text in 9.16.1 confusing | added text for ErrorRecoVery | | Level ---+----------------------------------------------------------+--------- -------------------- E26| MUST in 10.1 | fixed text ---+----------------------------------------------------------+--------- -------------------- E27| MUST for CHAP | fixed text (as in T33) ---+----------------------------------------------------------+--------- -------------------- E28| SHOULD/may in Initiator and Target Alias in 11 | fixed see also T36 ---+----------------------------------------------------------+--------- -------------------- E29| 11.8 replace port with reference to IANA cons. | ---+----------------------------------------------------------+--------- -------------------- E30| Phrasing for immediate data - MUST? | might be confusing - suggest | | leave as it is ---+----------------------------------------------------------+--------- -------------------- E31| Immediate & R2T - more explanation | there is a table ---+----------------------------------------------------------+--------- -------------------- E32| Addition to Copyright! IP Rights Notices | Added ---+----------------------------------------------------------+--------- -------------------- -----Original Message----- From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu] On Behalf Of Julian Satran Sent: Thursday, July 18, 2002 7:21 AM To: ips@ece.cmu.edu Subject: iSCSI - first working group Last Call round issues list and resolutions Dear colleagues, The list of issues raised and their resolution is attached. The working version of draft 15 is still up for review on my site. Julo
Home Last updated: Tue Jul 30 10:39:13 2002 11481 messages in chronological order |