802.3z/D2.1 CLAUSE 35 COMMENTS With Resolutions as accepted by 802.3z 133 CommenterName: Bob Grow Subclause: General Page: Line: CommentType: E Comment: Review all references prior to publication of D3. SuggestedRemedy: Fix references as necessary. Editor's recommendation: Accept. 134 CommenterName: Bob Grow Subclause: 35.5.2.2 Page: 31 Line: 26, 27 CommentType: E Comment: Incomplete conversion of ISO decimal point style SuggestedRemedy: Change "," to ".". Editor's recommendation: Accept. 135 CommenterName: Don Wong Subclause: 35.2.1.2.2 Page: 5 Line: 27-28 CommentType: T Comment: Why was EXTEND_ERROR removed from the receive path? EXTEND_ERROR is still present in the transmit path. By removing EXTEND_ERROR makes it inconsistent with fig 35-11 Table 35-2 SuggestedRemedy: Please provide a reason for the removal of EXTEND_ERROR in the receive path and update inconsistent documentation Editor's recommendation: The change was made to be consistent with clause 4 (see clause 35/D2 comment #129). The referenced table and figure describe the GMII. 35.2.1.2.2 describes the PLS service interface. The Reconcilliation Sublayer accomodates the differences (see 35.2.1.5). 136 CommenterName: Don Wong Subclause: 35.2.2.9 Page: 35.13 Line: fig 35-11 CommentType: E Comment: Not clear what "E, X, T, 1f, N, D" is. SuggestedRemedy: Change entries in the RXD<7:0> from "E, X, T, 1F, N, D" to "0F, 0F, 0F, 1F, 0F, 0F" Editor's recommendation: Reject. The figure is illustrative. This can be carried to the extreme causing a loss of readability. With the same consistency arguement, "P, R, E, A, M, B, L, E" should change to the hex byte stream. Figure 35-7 then is little more than a string of octets in hex. It would be better to back out the D2 change that converted an "XX" to the "1F". That would leave the only octet value in figure 35-12 as is the case in clause 22. 137 CommenterName: Don Wong Subclause: 35.2.2.6 Page: 35.10, 35.11, 35.12, 35.13, 35.15 Line: fig 35-5, fig 35-6, fig 35-7, fig 35-9, fig 35-10, fig 35-14 CommentType: E Comment: 1F is used, but 0F is spelled out as EXTEND in TXD<7:0> and RXD<7:0> field SuggestedRemedy: Be consistent, insert "0F" in place of "E,X,T,E,N,D" in the TXD<7:0> or RXD<7:0> Editor's recommendation: Reject. See recommendation on comment #136. 138 CommenterName: Don Wong Subclause: 35.2.5.1 Page: 35.21 Line: 19 CommentType: E Comment: Typing Error SuggestedRemedy: Change "table 22-7 below" to "table 35-7 below" Editor's recommendation: Overtaken. 139 CommenterName: Don Wong Subclause: 35.2.1.5 Page: 35.6 Line: 40-43 CommentType: I Comment: Lines 40-43 suggest that the Reconciliation sublayer may corrupt the receive frame data, to insure a bad CRC results. This in not necessary since RX_DV & RX_ER already indicates this. SuggestedRemedy: I suggest that we remove line 40-43. Editor's recommendation: Reject. The MAC does not see GMII signals, only what is presented thought the PLS Service interface. The behavior described for the Reconciliation sublayer is consistent with clause 22 Fast Ethernet. An implementer may use other techniques. 140 CommenterName: Don Wong Subclause: 35.2.2.10 Page: 35.14 Line: 50 CommentType: E Comment: Typing Error SuggestedRemedy: Change "shall asserwhen" to "shall assert when" Editor's recommendation: Accept with modification, "shall assert CRS when". 141 CommenterName: Don Wong Subclause: 35.2.2.10 Page: 35.15 Line: 19 CommentType: E Comment: Add detail CommentEnd: SuggestedRemedy: Change "duplex mode bit 0.x" to "duplex mode bit 0.8" Editor's recommendation: Accept. 142 CommenterName: Don Wong Subclause: 35.2.2.10 Page: 35.15 Line: 19 CommentType: E Comment: Add detail SuggestedRemedy: Change "as described in 35.2.4.1.x" to as described in 35.2.5.1.8 Editor's recommendation: Overtaken. Referenced text will move to clause 22. All referrences to be revied prior to D3. 143 CommenterName: Don Wong Subclause: 35.2.5.1.3 Page: 35.22 Line: 18-21 CommentType: E Comment: "The combination of both bits set to logic one is reserved for future standardization" is inconsistent with the definition in Table 35-7. From table 35-7 is interpret as the following bit 0.6=1, bit 0.13=X : 1000 Mbps bit 0.6=0, bit 0.13=1 : 100 Mbps bit 0.6=0, bit 0.13=0 : 10 Mbps SuggestedRemedy: modifying in Table 35-7 1. Under the description for bit 0.13 from 1 = 100 Mb/s 0 = 10 Mb/s to if 0.6=0, 1=100 Mb/s if 0.6=0, 0=10 Mb/s 2. Under the description for bit 0.6 from 1 = 1000 Mb/s 0 = 10 Mb/s or 100 Mb/s to bit 0.6=0, bit 0.13=0 : 10 Mb/s bit 0.6=0, bit 0.13=1 : 100 Mb/s bit 0.6=1, bit 0.13=0 : 1000 Mb/s bit 0.6=1, bit 0.13=1 : reserved Editor's recommendation: Accept. 144 CommenterName: Don Wong Subclause: 35.2.5.1.11 Page: 35.23 Line: 37 CommentType: E Comment: typing error SuggestedRemedy: Change "Auto-negotiation" to "Link Configuration" Editor's recommendation: Overtaken. Link Configuration is to be deleted from draft. 145 CommenterName: Don Wong Subclause: 35.2.5.1.11 Page: 35.23 Line: 38 CommentType: E Comment: typing error SuggestedRemedy: Change "Auto-negotiation" to "Link Configuration" Editor's recommendation: Overtaken. Link Configuration is to be deleted from draft. 146 CommenterName: Don Wong Subclause: 35.3 Page: 35.27 Line: 40 CommentType: E Comment: typing error SuggestedRemedy: Change "Table 35-12" to Table 35-10" Editor's recommendation: Overtaken. Table will renumber because of move of information within 35.2.5. Editor will check all references prior to publication of next draft. 147 CommenterName: Don Wong Subclause: 35.5.2.2 Page: 35.31 Line: Table 35-14 CommentType: E Comment: Under parameter "Input High Voltage" and "Input Low Voltage", values are "1,8" and "0,8" SuggestedRemedy: Change "1,8" to "1.8" and "0,8" to "0.8" Editor's recommendation: Accept. 148 CommenterName: Ben Brown SubClause: 35.2.3.1 Page: 35.17 Line: 2 CommentType: E Comment: Since you are referencing Figure 35-10 for the receive case, shouldn't you reference Figure 35-7 for the transmit case? SuggestedRemedy: Reference Figure 35-7 for the transmit case Editor's recommendation: Accept. 149 CommenterName: Ben Brown SubClause: 35.2.5.3.2 Page: 35.27 Line: 3 CommentType: E Comment: Missing "half-duplex" SuggestedRemedy: Change the line to read "...indicates that the PHY has the ability to perform half-duplex link transmission and" Editor's recommendation: Accept. 150 CommenterName: Ben Brown SubClause: 35.2.5.3.1 Page: 35.26 Line: 49 CommentType: E Comment: Missing "full-duplex" SuggestedRemedy: Change the line to read "...indicates that the PHY has the ability to perform full-duplex link transmission and" Editor's recommendation: Accept. 151 CommenterName: Ben Brown SubClause: 35.2.5.3.2 Page: 35.27 Line: 5 CommentType: E Comment: "full-duplex" should be "half-duplex" SuggestedRemedy: Change the line to read "that the PHY lacks the ability to perform half-duplex link transmission and..." Editor's recommendation: Accept. 152 CommenterName: Steve Dreyer SubClause: 35.2.2.9 Page: 35.14 Line: 21 CommentType: E Comment: Missing reference. SuggestedRemedy: Editor's recommendation: Accept. References will be reviewed prior to publication of D3. 153 CommenterName: Steve Dreyer SubClause: 35.2.2.3 Page: 35.8 Line: 9 CommentType: E Comment: Change "+100ppm" to "+/-100ppm". SuggestedRemedy: Editor's recommendation: Accept. This may be a cross-platform font problem that will recreate the error in D3! 154 CommenterName: Steve Dreyer SubClause: 35.2.2.10 Page: 35.14 Line: 50 CommentType: E Comment: Change "asserwhen" to "assert CRS when". SuggestedRemedy: Editor's recommendation: Accept. 155 CommenterName: Steve Dreyer SubClause: 35.2.2.11 Page: 35.16 Line: 4-6 CommentType: E Comment: Remove Editor Note, left over from the deleted Note above it. SuggestedRemedy: Editor's recommendation: It is not in the source--FrameMaker oddity. 156 CommenterName: Steve Dreyer SubClause: 35.2.3.2.1 Page: 35.17 Line: 18, CommentType: E Comment: The section reference, 7.2.3.2, is actually the preamble and SFD for the AUI. A better reference would be the MAC frame format section in Clause 3 or 4. SuggestedRemedy: Editor's recommendation: Accept. 157 CommenterName: Steve Dreyer SubClause: 35.2.3.2.2 Page: 35.18 Line: 33 CommentType: E Comment: The phrase "... can result in shrinkage of the preamble during transmission." seems incorrect since the transmitters never shrink the preamble, but the receivers are allowed to lose preamble bits resulting in shrinkage on the receive GMII. SuggestedRemedy: Change "... shrinkage of the preamble during transmission." to ".shrinkage of the preamble during reception." Editor's recommendation: The clause 36 transmit PCS is allowed to shrink preamble one byte to produce ordered set alignment. Will improve text. Accepted resolution to a D2 comment did not make it into D2.1. 158 CommenterName: Steve Dreyer SubClause: 35.2.5.1 Page: 35.21 Line: 35-36 CommentType: E Comment: Bits 0.12 and 0.9 use the term "Auto-Negotiation" in the name and description, even though these bits are used for both Auto-Negotiation and Link Configuration. SuggestedRemedy: Change the term "Auto-Negotiation" to something more generic, like "Negotiation" or "Autonegotiation/Link Configuration". Editor's recommendation: Accept with modification. The subtask group in joint meeting with the PMA and Management subtask groups decided to eliminate the name Link Configuration, and to convert the register descriptions to clause 22 changes. 159 CommenterName: Steve Dreyer SubClause: 35.2.5.1 Page: 35.21 Line: 45 CommentType: E Comment: Note c on Bit 0.8 no longer applies since 802.3x is done. SuggestedRemedy: Editor's recommendation: Accept. 160 CommenterName: Steve Dreyer SubClause: 35.2.5.1 Page: 35.21 Line: 41 CommentType: E Comment: Note b on Bit 0.10 doesn't apply to GMII. It is a holdover from MII in cases where a MII connnector was used. SuggestedRemedy: Editor's recommendation: Reject. Bit definitions are maintained for 1000 Mb/s operation, though application may vary from 100 Mb/s. 161 CommenterName: Steve Dreyer SubClause: 35.2.5.2.11 Page: 35.25 Line: 35 CommentType: E Comment: Change "appies" to "applies". SuggestedRemedy: Editor's recommendation: Accept. 162 CommenterName: Steve Dreyer SubClause: 35.2.5.1.11 Page: 35.23 Line: 37-38 CommentType: T Comment: When bit 0.5=1, link configuration mode is selected. As such, the term "Auto-negotation" in lines 37-38 should be changed to "Link Configuration". SuggestedRemedy: Editor's recommendation: Accept with modification. The subtask group in joint meeting with the PMA and Management subtask groups decided to eliminate the name Link Configuration, and to convert the register descriptions to clause 22 changes. 163 CommenterName: Steve Dreyer SubClause: Page: 35.24 Line: 24 CommentType: E Comment: The term "Auto-Negotation" is used the bit 1.5 name and description even though it will refer to either Link Configuration or Auto-Negotaiton, depending on bit 0.5. Less confusing would be to change the term "Auto-Negotation" in this case to either "Negotiation" or "Auto-Negotation/Link Configuration" or something more generic or descriptive. SuggestedRemedy: Editor's recommendation: Accept with modification. The subtask group in joint meeting with the PMA and Management subtask groups decided to eliminate the name Link Configuration, and to convert the register descriptions to clause 22 changes. 164 CommenterName: Steve Dreyer SubClause: 35.2.5.2.13 Page: 35.25 Line: 43-44 CommentType: T Comment: Bit 1.3, Auto-Negotation ability, should be used for either Auto-Negotation or Link Configuration, depending on bit 0.5. SuggestedRemedy: Change description of Auto-Negotation ability bit to something like: " The operation of bit 1.3 is unchanged from 22.2.4.2.12 except that the status applies to either Auto-Negotation or Link Configuration as selected by bit 0.5. Editor's recommendation: Accept with modification. The subtask group in joint meeting with the PMA and Management subtask groups decided to eliminate the name Link Configuration, and to convert the register descriptions to clause 22 changes. 165 CommenterName: Steve Dreyer SubClause: 35.2.5.2.13 Page: 35.25 Line: 43-44 CommentType: T Comment: Bit 1.3, Auto-Negotation ability, should be used for either Auto-Negotation or Link Configuration, depending on bit 0.5. SuggestedRemedy: Change description of Auto-Negotation ability bit to something like: "The operation of bit 1.3 is unchanged from 22.2.4.2.12 except that the status applies to either Auto-Negotation or Link Configuration as selected by bit 0.5. Also name should be changed to be more generic. Editor's recommendation: Duplicate of 164. 166 CommenterName: Steve Dreyer SubClause: 35.5.2.1 Page: 35.30 Line: 22-28 CommentType: T Comment: Rise/fall issues: 1. For clock tr, Shouldn't there be a min, maybe 0.7? 2. For clock tf, total variation from min to max is 0.7-1.0. This may exceed normal IC process variations. Couldn't this be same as 10-bit interface, 0.-2.4? 3. For clock tr, 1.0ns seems fast, could this number be same as 10-bit interface, 0.7-2.4? 4. Shouldn't be a max on tr and tf for data? SuggestedRemedy: Clock tr, tf = 0.7-2.4 ns Data tr, tf = 0.7-2.4 ns Editor's recommendation: Reject. The subtask group reviewed the Clock Rise and Fall Time values and by motion adopted a set of values inconsistent with the SuggestedRemedy. 167 CommenterName: Steve Dreyer SubClause: 35.5.2.1 Page: 35.30-31 Line: 42,4 CommentType: T Comment: If setup and hold time is specified for both transmit and receive GMII, seems that tpd delay time spec is unnecessary and confusing in Tables 35-12 and 35-13. SuggestedRemedy: Delete tpd from Tables 35-12 and 35-13. Editor's recommendation: Accept. D3.0 will include new pictures, and tPD will be clarified. 168 CommenterName: Steve Dreyer SubClause: 35.5.2.2 Page: 35.31 Line: 23 CommentType: T Comment: Specifying VOH max=VCC with VCC=3.6v in Table 35-14 makes all 5v devices noncompliant with the standard. 5v process are capable of gigabit performance, standard should not be written to preclude a viable implementation. SuggestedRemedy: Have no spec for VOH max and leave up to implementors, or put note explaining overdrive considerations, or ? Editor's recommendation: Reject. Recommendation was to specify interface more for the future than for the past. A 5v device would not be judged noncompliant, since there is no way to test compliance to an internal interface. (The GMII is not required to be externally observable.) 169 CommenterName: Steve Dreyer SubClause: Figure Page: 35-28 Line: 33-44 CommentType: E Comment: "1 ns MIN" and "2 ns MIN" are really minimum values for tsetup and thold. It would be clearer of these labels were replaced with "tSETUP" and "tHOLD" as reference to in Tables 35-12 and 35-13. SuggestedRemedy: Editor's recommendation: Accept. Figure 35.17 will be edited in D3.0 to remove values and replace with signal names. 170 CommenterName: Steve Dreyer SubClause: Figure Page: 35-28 Line: 33-44 CommentType: E Comment: "1 ns MIN" and "2 ns MIN" are really minimum values for tsetup and thold. It would be clearer of these labels were replaced with "tSETUP" and "tHOLD" as reference to in Tables 35-12 and 35-13. SuggestedRemedy: Add tr/tf max = 2.4ns for data. Editor's recommendation: Rejected. This appears to be a duplicate in combination with a cut and paste problem. Commenter was notified and accepts comment as closed.