RFC 466
Telnet logger/server for host LL-67
RFC 466: TELNET LOGGER/SERVER For Host LL-67 ✨
Okay bestie, the title block is giving old-school 🦋. It's from the Network Working Group 💜 and it's Request for Comments: 466 (NIC 14740). J. Winett of LL-67 wrote it on 27 February 1973, and the category is TELNET 🌸. It doesn't list anything it obsoletes or updates.
The writeup covers the TELNET LOGGER/SERVER for the CP/CMS system on the Lincoln Laboratory 360>67 (host 10) 🖥️. It works with both half duplex and full duplex TELNET users, and the data can be in either ASCII or EBCDIC codes. Obviously this is so cute 🥹
The hide-your-input and noecho TELNET controls are used for the EBCDIC print suppress (bypass) and print restore features during the login procedure. Half duplex terminals get the TELNET control break (reverse break) sent as an input prompt when input is wanted. That same code can also say that a previous line sent without an end of line sequence (CR-LF) should be printed. Iconic 💅
This material has not been reviewed for public release and is intended only for use with the ARPA network. It should not be quoted or cited in any publication not related to the ARPA network. 🔒
Operation of the Lincoln Laboratory CP/CMS TELNET LOGGER/SERVER ✨
ICP Connection
The TELNET LOGGER/SERVER follows the ICP protocol for making a pair of connections. The LOGGER starts out enabled for a connection on socket X'00000001'. When an RFC is received for this socket, a pair of sockets gets picked for the TELNET connections. If the maximum number of TELNET users who can be served are already active, the initial connection gets refused. Currently, three TELNET users can be served 🧚
TELNET LOGGER
Once the ICP connections are set up, the LOGGER waits for a TELNET data type code, a string of network ASCII characters, or a null line (just CR-LF) to figure out whether it should run in ASCII or EBCDIC character codes. ASCII is assumed unless the first byte received is the TELNET EBCDIC data type code (X'A2'). Once something arrives, the LOGGER sends this welcome message:
Lincoln Laboratory CP/CMS OnlineFor example, if you want ASCII, send a null line (just CR-LF) on the send socket and the welcome message comes back in ASCII. Then the CP login procedure can start 🌈. If you want EBCDIC instead, the very first byte you send should be the TELNET data type code for EBCDIC (X'A2'). After that, everything stays in whichever code you started with. Main character energy 👑
CP expects the user to type:
LOGIN useridHere, the userid picks the virtual machine. CP then replies with:
ENTER PASSWORD:Right after that, CP sends the EBCDIC code for bypass (x'24'), which gets mapped to the TELNET code hide-your-input 🙈.
Now the user types in a password. Network passwords can be different from the ones typed on a local terminal, and the LOGGER maps network passwords to a matching CP password. So you can only get into an account from the network if a network password, a CP password, and a userid are all listed in a file that the LOGGER reads. If the userid from the network isn't in the LOGGER FILE (or the network password doesn't match the one in the file for that userid), the LOGGER passes an invalid userid (or password) to CP. Then CP's response for an invalid userid or password goes back to the network user. Not me getting locked out 💀
After CP receives a password, CP sends the EBCDIC code for restore (X'14'), which gets mapped to the TELNET control noecho 🍓
TELNET SERVER
CP/CMS runs on EBCDIC codes, so ASCII codes have to be translated into EBCDIC before they're sent to a virtual machine. The ASCII codes and their EBCDIC mappings are supposed to be in Figure 1 🎀
Heads up: Figure 1 isn't included in this copy of the text, so the table isn't here.
When the ASCII sequence CR-LF comes in, it gets mapped into the EBCDIC code NL. Whenever the TELNET control NOP shows up in an input string, it gets mapped into an EBCDIC idle (X'17') and then taken out of the string. So if TELNET NOP codes sit between a CR and an LF, they're removed first, and then the CR-LF becomes the EBCDIC NL. Lowkey genius 🧠
The TELNET control hide-your-input maps to the EBCDIC code for bypass (X'24'), and the TELNET control echo maps to the EBCDIC control for restore (X'14'). If the TELNET control echo is received, the SERVER should send the control noecho, but this feature hasn't been implemented yet 😭. For now, the TELNET control echo gets mapped to the EBCDIC code X'23'. If the TELNET break comes in, it's read as an attention signal and CP or CMS takes the matching action 🔥.
CP/CMS works one line at a time and expects all input to be lines that end with an NL code. Characters get buffered until the newline code shows up 📝.
CP/CMS is also half duplex, so it only looks at characters when it's expecting input. If it isn't expecting input, a network interrupt is needed to make the SERVER process the characters that came in. When a network interrupt happens, the characters received before the TELNET data mark are examined and then thrown away. The one exception: if a TELNET break code is found, the right CP/CMS interrupt action gets triggered ✨
On output, EBCDIC codes get mapped into network ASCII if a mapping exists. If there isn't one, they get mapped into the TELNET control NOP. An NL code becomes CR-LF. The EBCDIC code for bypass becomes the TELNET control hide-your-input, and the EBCDIC code for restore becomes the TELNET control noecho. Also, X'23' becomes the TELNET control echo, and X'38' becomes the TELNET control break 💖.
Since CP/CMS is a one-line-at-a-time, half duplex system, the TELNET control break gets sent as an end of message signal and also as an input prompt code. If characters were output without an NL, the break works as an end of message code. For a user TELNET running in one-line-at-a-time mode, that tells it to print the characters that were already sent without waiting for the end of line sequence. If the user TELNET is also in half duplex mode, the break works as an input prompt and tells the user the system is ready for input 🦋.
If a full duplex user TELNET sends input early, the TELNET SERVER will already have that input ready for immediate processing. So for a full duplex user TELNET, the break used as a prompt should be ignored 🫶.
CP/CMS runs half duplex, but it still supports half duplex terminals through the reverse break feature. That lets the system cancel an input mode so it can send a priority output message. In that case, the TELNET SERVER sends a TELNET SYNC. A half duplex user TELNET should read that as a cue to cancel input mode, which means it revokes a previous TELNET break that was treated as an input prompt 🌸.
No codes in the output character stream can make the TELNET data mark get sent. 💅
Part 2 is coming 🫶💜
LOGOUT ✨
okay so bestie, logging out is actually simple 💅 When a user logs out from their virtual machine, CP passes the equivalent of a line disconnect to the LOGGER 📞 Then the LOGGER closes the TELNET send and receive sockets 🔌 Done and dusted, no drama 🫶
Now for the big ASCII/EBCDIC chart 🧚 Every character gets its code in both systems, so you can translate between them without losing your mind 🤯 Here's the whole thing:
ASCII ASCII ASCII SYMBOLS EBCDIC EBCDIC
DEC OCT HEX HEX DEC
0 0 (00) NUL (00) 00
1 1 (01) SOH (01) 01
2 2 (02) STX (02) 02
3 3 (03) ETX (03) 03
4 4 (04) EOT (37) 55
5 5 (05) ENQ (2D) 45
6 6 (06) ACK (2E) 46
7 7 (07) BEL (2F) 47
8 10 (08) BS (16) 22
9 11 (09) HT (05) 05
10 12 (0A) LF (25) 37
11 13 (0B) VT (0B) 11
12 14 (0C) FF (0C) 12
13 15 (0D) CR (0D) 13
14 16 (0E) SO (0E) 14
15 17 (0F) SI (0F) 15
16 20 (10) DLE (10) 16
17 21 (11) DC1 (11) 17
18 22 (12) DC2 (12) 18
19 23 (13) DC3 (13) 19
20 24 (14) DC4 (3C) 60
21 25 (15) NAK (3D) 61
22 26 (16) SYN (32) 50
23 27 (17) ETB (26) 38
24 30 (18) CAN (18) 24
25 31 (19) EM (19) 25
26 32 (1A) SUB (3F) 63
27 33 (1B) CTL (27) 39
28 34 (1C) FS (1C) 28
29 35 (1D) GS (1D) 29
30 36 (1E) RS (1E) 30
31 37 (1F) US (1F) 31
ASCII/EBCDIC Code Mappings
FIGURE 1 ASCII ASCII ASCII SYMBOLS EBCDIC EBCDIC
DEC OCT HEX HEX DEC
32 40 (20) SP (40) 64
33 41 (21) ! (5A) 90
34 42 (22) " (7F) 127
35 43 (23) # (7B) 123
36 44 (24) $ (5B) 91
37 45 (25) % (6C) 108
38 46 (26) & (50) 80
39 47 (27) ' (7D) 124
40 50 (28) ( (4D) 77
41 51 (29) ) (5D) 93
42 52 (2A) * (5C) 92
43 53 (2B) + (4E) 78
44 54 (2C) , (6D) 109
45 55 (2D) - (60) 96
46 56 (2E) . (4B) 75
47 57 (2F) / (61) 97
48 60 (30) 0 (F0) 240
49 61 (31) 1 (F1) 241
50 62 (32) 2 (F2) 242
51 63 (33) 3 (F3) 243
52 64 (34) 4 (F4) 244
53 65 (35) 5 (F5) 245
54 66 (36) 6 (F6) 246
55 67 (37) 7 (F7) 247
56 70 (38) 8 (F8) 248
57 71 (39) 9 (F9) 249
58 72 (3A) : (7A) 122
59 73 (3B) ; (5E) 94
60 74 (3C) < (4C) 76
61 75 (3D) = (7E) 126
62 76 (3E) > (6E) 110
63 77 (3F) ? (6F) 111
ASCII/EBCDIC Code Mappings
FIGURE 1 (CONTINUED) ASCII ASCII ASCII SYMBOLS EBCDIC EBCDIC
DEC OCT HEX HEX DEC
64 100 (40) @ (7C) 124
65 101 (41) A (C1) 193
66 102 (42) B (C2) 194
67 103 (43) C (C3) 195
68 104 (44) D (C4) 196
69 105 (45) E (C5) 197
70 106 (46) F (C6) 198
71 107 (47) G (C7) 199
72 110 (48) H (C8) 200
73 111 (49) I (C9) 201
74 112 (4A) J (D1) 209
75 113 (4B) K (D2) 210
76 114 (4C) L (D3) 211
77 115 (4D) M (D4) 212
78 116 (4E) N (D5) 213
79 117 (4F) O (D6) 214
80 120 (50) P (D7) 215
81 121 (51) Q (D8) 216
82 122 (52) R (D9) 217
83 123 (53) S (E2) 226
84 124 (54) T (E3) 227
85 125 (55) U (E4) 228
86 126 (56) V (E5) 229
87 127 (57) W (E6) 230
88 130 (58) X (E7) 231
89 131 (59) Y (E8) 232
90 132 (5A) Z (E9) 233
91 133 (5B) [ (AD) 173
92 134 (5C) <cent> (4A) 74 (BACK-SLASH)
93 135 (5D) ] (BD) 189
94 136 (5E) ^ (71) 113 (CARAT)
95 137 (5F) _ (6D) 109
ASCII/EBCDIC Code Mappings
FIGURE 1 (CONTINUED) ASCII ASCII ASCII SYMBOLS EBCDIC EBCDIC
DEC OCT HEX HEX DEC
96 140 (60) ` (79) 121 (GRAVE)
97 141 (61) a (81) 129
98 142 (62) b (82) 130
99 143 (63) c (83) 131
100 144 (64) d (84) 132
101 145 (65) e (85) 133
102 146 (66) f (86) 134
103 147 (67) g (87) 135
104 150 (68) h (88) 136
105 151 (69) i (89) 137
106 152 (6A) j (91) 145
107 153 (6B) k (92) 146
108 154 (6C) l (93) 147
109 155 (6D) m (94) 148
110 156 (6E) n (95) 149
111 157 (6F) o (96) 150
112 160 (70) p (97) 151
113 161 (71) q (98) 152
114 162 (72) r (99) 153
115 163 (73) s (A2) 162
116 164 (74) t (A3) 163
117 165 (75) u (A4) 164
118 166 (76) v (A5) 165
119 167 (77) w (A6) 166
120 170 (78) x (A7) 167
121 171 (79) y (A8) 168
122 172 (7A) z (A9) 169
123 173 (7B) { (8B) 139
124 174 (7C) | (4F) 79 (BAR/OR)
125 175 (7D) } (9B) 155
126 176 (7E) <bent bar> (5F) 95 (TILDE/NOT)
127 177 (7F) DEL (07) 7 ASCII ASCII ASCII TELNET EBCDIC EBCDIC
DEC OCT HEX CONTROLS HEX DEC
128 100 (80) DATA-MARK (80) 128
129 101 (81) BREAK (38) 56
130 102 (82) NOP (17) 23 IDLE
131 103 (83) NOECHO (14) 20 RESTORE
132 104 (84) ECHO (23) 35
133 105 (85) HIDE-YOUR INPUT (24) 36 BYPASS
ASCII/EBCDIC Code Mappings
FIGURE 1 (CONTINUED) [This RFC was put into machine readable form for entry]
[into the online RFC archives by Helene Morin, Via Genie, 12/99]📦 where this page comes from
- ietf: rfc466 1973-02
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms