ps2_ascii_table¶
Source: Verilog/Terminals/rtl/ps2_ascii_table.v
Hierarchy: not instantiated by any of the 9 build tops (elaborated by yosys).
Module hierarchy - All modules
Schematic¶
Drawn from the Verilog: no build top uses this module, so it was elaborated from its own file with no defines and default parameters. Sub-modules are boxes (click the picture to open it full size; there every sub-module box links to its page, and every wire shows its Verilog name).
Ports¶
| Direction | Width | Name | Description |
|---|---|---|---|
| input | [7:0] |
code |
|
| input | 1 |
layout_no |
|
| output | [7:0] |
unshifted |
|
| output | [7:0] |
shifted |
|
| output | [7:0] |
extended |
Verilog source¶
Verilog/Terminals/rtl/ps2_ascii_table.v on GitHub.
Show the Verilog of ps2_ascii_table (253 lines)
//============================================================================
//! PS/2 scancode set 2 -> ASCII, unshifted and shifted
//!
//! Part of the board-independent terminal core (Verilog/Terminals/).
//! Pure combinational lookup; 0x00 out means "this key has no character".
//!
//! SCANCODES CROSS-CHECKED 30-AUG-2026 against two independent published
//! references, which agree with every entry below including the F7=0x83
//! oddity: the OSDev wiki's set-2 table (wiki.osdev.org/PS/2_Keyboard) and
//! Vetra Systems' scan-code translation table (vetra.com/scancodes.html).
//! Scan code set 2 is a fixed standard every PS/2-compatible keyboard and
//! the Nexys board's USB-HID bridge implement, so the codes are facts, not
//! transcription hopes. What documentation CANNOT settle - and typing on
//! the real board still can (fpga/nexys4ddr/PLAN-vga-console.md phase 3):
//! whether the board's bridge really emits set 2 unsurprisingly, and the
//! NORWEGIAN LAYOUT POSITIONS below, which come from RetroTerm's
//! KBD-ND-246.md rather than from a scancode standard.
//!
//! Extended (E0-prefixed) keys ARE here, in their own `extended` output and
//! their own case block below. Since the 30-AUG-2026 VT100 decision the
//! arrows and HOME emit SEQUENCE MARKERS (0x80 | final byte) that
//! key_vt100.v expands into the real ESC [ x bytes on the way to the UART -
//! see the block comment on the case below.
//!
//! Written 27-AUG-2026.
//============================================================================
`default_nettype none
module ps2_ascii_table (
input wire [7:0] code,
//! 0 = US ANSI, 1 = Norwegian (NS 4551-1). See the block comment below.
input wire layout_no,
output wire [7:0] unshifted,
output wire [7:0] shifted,
//! Byte for the same scancode when it arrived with the E0 (extended)
//! prefix - the arrows, HOME and DELETE. 0x00 = this extended key sends
//! nothing; 0x80|f = key_vt100.v expands it to ESC [ f (see below).
output reg [7:0] extended
);
//--------------------------------------------------------------------------
// Extended (E0-prefixed) keys -> VT100 (30-AUG-2026)
//
// A VT100 sends multi-byte escape sequences for the special keys - three
// to five bytes per keypress. One lookup cannot produce that, so the table
// emits a MARKER byte that key_vt100.v expands. Bit 7 is the marker flag
// (safe: the ND-120 is a 7-bit machine, no typed character has it, and a
// marker that leaks unexpanded is visible instead of vanishing); bits 6:5
// select the sequence FAMILY, bits 4:0 the payload
// (SPEC-vt100-keys.md has the full story):
//
// 100f_ffff ESC [ <0x40+f> arrows: f 1..4 = A B C D
// 101f_ffff ESC O <0x40+f> DEC PF keys: f 16..19 = P Q R S
// 110n_nnnn ESC [ <n> ~ editing keys n=1..6, F5-F12 n=15..24
// 111x_xxxx reserved - key_vt100 drops it
//
// (Until 30-AUG-2026 these were the TDV bare C0 bytes - UP=0x1C, DOWN=0x0B,
// LEFT=0x08, RIGHT=0x18, HOME=0x1D - correct for terminal type 53, whose
// evidence trail is in docs/SPEC-tdv2200.md and the git history. The
// terminal is a VT100 now.)
//
// Scancodes cross-checked against OSDev + Vetra 30-AUG-2026 - see the
// module header. Both agree on every code below.
//--------------------------------------------------------------------------
always @(*) begin
case (code)
8'h75: extended = 8'h81; // Up -> ESC [ A
8'h72: extended = 8'h82; // Down -> ESC [ B
8'h74: extended = 8'h83; // Right -> ESC [ C
8'h6B: extended = 8'h84; // Left -> ESC [ D
// Home is ESC [ H, MEASURED (30-AUG-2026, live PED on a type-6
// SINTRAN line): ESC[H toggles to/from the PED: command line, while
// ESC[1~ (DEC FIND, which the spec first chose) does NOTHING in PED.
// The doc-derived guess pointed both ways; the live editor settled it.
8'h6C: extended = 8'h88; // Home -> ESC [ H
8'h70: extended = 8'hC2; // Insert -> ESC [ 2 ~ (DEC INSERT HERE)
8'h71: extended = 8'hC3; // Delete -> ESC [ 3 ~ (DEC REMOVE)
8'h69: extended = 8'hC4; // End -> ESC [ 4 ~ (DEC SELECT)
8'h7D: extended = 8'hC5; // PgUp -> ESC [ 5 ~ (DEC PREV SCREEN)
8'h7A: extended = 8'hC6; // PgDn -> ESC [ 6 ~ (DEC NEXT SCREEN)
8'h5A: extended = 8'h0D; // keypad Enter -> CR
default: extended = 8'h00; // no VT100 equivalent (GUI keys, menu...):
// send NOTHING rather than invent bytes.
endcase
end
//--------------------------------------------------------------------------
// Norwegian (NS 4551-1) - the layout_no overrides
//
// TWO THINGS DIFFER, and getting only one right produces nonsense:
//
// 1. WHICH KEY carries the character. Positions taken from RetroTerm's
// docs/KBD-ND-246.md (the ND 246 keyboard grid): D11, where US has [,
// is AA; C10 (US ;) is OE; C11 (US ') is AE; B8 (US ,) is , / ; ;
// B9 (US .) is . / : ; B10 (US /) is - / _ ; E11 (US -) is + / ?.
// Those are the same positions a modern PC Norwegian keyboard uses,
// which is the corroboration - two independent sources agreeing beats
// one source trusted.
//
// 2. WHICH BYTE it sends. NOT Latin-1. The ND-120 is a 7-BIT machine
// speaking ISO 646, where AE OE AA REPLACE [ \ ] and ae oe aa replace
// { | }. RetroTerm's FontTDV2200.cs says it outright: "The host sends
// 7-bit ASCII position bytes (e.g. '[' = 0x5B for AE)". Send 0xC6 for
// AE and SINTRAN sees a byte with no meaning; send 0x5B and it is right.
//
// The SCREEN has to agree, which is why font_rom carries a second page (see
// font/make_font.py). Keyboard and font are selected by the SAME bit -
// letting them disagree would draw '[' for a typed AE.
//
// DEAD KEYS ARE NOT IMPLEMENTED. On a PC Norwegian layout the ¨/^ and ´/`
// keys compose with the next keystroke, which needs state this table does
// not have. They send their shifted character where that is plain ISO 646,
// and nothing unshifted. Sending a plausible wrong byte would be worse than
// sending none.
//--------------------------------------------------------------------------
reg [7:0] no_unshifted, no_shifted;
reg no_override;
always @(*) begin
no_override = 1'b1;
no_unshifted = 8'h00;
no_shifted = 8'h00;
case (code)
// number row - only the SHIFTED symbols move
8'h1E: begin no_unshifted = "2"; no_shifted = 8'h22; end // " where US has @
8'h25: begin no_unshifted = "4"; no_shifted = 8'h24; end // 0x24 is the currency sign in ISO 646 NO
8'h36: begin no_unshifted = "6"; no_shifted = "&"; end
8'h3D: begin no_unshifted = "7"; no_shifted = "/"; end
8'h3E: begin no_unshifted = "8"; no_shifted = "("; end
8'h46: begin no_unshifted = "9"; no_shifted = ")"; end
8'h45: begin no_unshifted = "0"; no_shifted = "="; end
// the three Norwegian letters, in their ISO 646 positions
8'h54: begin no_unshifted = 8'h7D; no_shifted = 8'h5D; end // D11 aa / AA
8'h4C: begin no_unshifted = 8'h7C; no_shifted = 8'h5C; end // C10 oe / OE
8'h52: begin no_unshifted = 8'h7B; no_shifted = 8'h5B; end // C11 ae / AE
// the punctuation that moves
8'h4E: begin no_unshifted = "+"; no_shifted = "?"; end // E11
8'h55: begin no_unshifted = 8'h5C; no_shifted = 8'h60; end // \ and `
8'h5D: begin no_unshifted = 8'h27; no_shifted = "*"; end // C12 ' / *
8'h41: begin no_unshifted = ","; no_shifted = ";"; end // B8
8'h49: begin no_unshifted = "."; no_shifted = ":"; end // B9
8'h4A: begin no_unshifted = "-"; no_shifted = 8'h5F; end // B10 - / _
8'h61: begin no_unshifted = "<"; no_shifted = ">"; end // the extra ISO key US ANSI does not have
// dead keys - shifted only, see the note above
8'h5B: begin no_unshifted = 8'h00; no_shifted = 8'h5E; end // D12 dead diaeresis / ^
default: no_override = 1'b0; // this key is the same in both layouts
endcase
end
reg [7:0] us_unshifted, us_shifted;
assign unshifted = (layout_no && no_override) ? no_unshifted : us_unshifted;
assign shifted = (layout_no && no_override) ? no_shifted : us_shifted;
always @(*) begin
case (code)
// --- letters ---------------------------------------------------------
8'h1C: begin us_unshifted = "a"; us_shifted = "A"; end
8'h32: begin us_unshifted = "b"; us_shifted = "B"; end
8'h21: begin us_unshifted = "c"; us_shifted = "C"; end
8'h23: begin us_unshifted = "d"; us_shifted = "D"; end
8'h24: begin us_unshifted = "e"; us_shifted = "E"; end
8'h2B: begin us_unshifted = "f"; us_shifted = "F"; end
8'h34: begin us_unshifted = "g"; us_shifted = "G"; end
8'h33: begin us_unshifted = "h"; us_shifted = "H"; end
8'h43: begin us_unshifted = "i"; us_shifted = "I"; end
8'h3B: begin us_unshifted = "j"; us_shifted = "J"; end
8'h42: begin us_unshifted = "k"; us_shifted = "K"; end
8'h4B: begin us_unshifted = "l"; us_shifted = "L"; end
8'h3A: begin us_unshifted = "m"; us_shifted = "M"; end
8'h31: begin us_unshifted = "n"; us_shifted = "N"; end
8'h44: begin us_unshifted = "o"; us_shifted = "O"; end
8'h4D: begin us_unshifted = "p"; us_shifted = "P"; end
8'h15: begin us_unshifted = "q"; us_shifted = "Q"; end
8'h2D: begin us_unshifted = "r"; us_shifted = "R"; end
8'h1B: begin us_unshifted = "s"; us_shifted = "S"; end
8'h2C: begin us_unshifted = "t"; us_shifted = "T"; end
8'h3C: begin us_unshifted = "u"; us_shifted = "U"; end
8'h2A: begin us_unshifted = "v"; us_shifted = "V"; end
8'h1D: begin us_unshifted = "w"; us_shifted = "W"; end
8'h22: begin us_unshifted = "x"; us_shifted = "X"; end
8'h35: begin us_unshifted = "y"; us_shifted = "Y"; end
8'h1A: begin us_unshifted = "z"; us_shifted = "Z"; end
// --- digits and their shifted symbols (US layout) --------------------
8'h45: begin us_unshifted = "0"; us_shifted = ")"; end
8'h16: begin us_unshifted = "1"; us_shifted = "!"; end
8'h1E: begin us_unshifted = "2"; us_shifted = "@"; end
8'h26: begin us_unshifted = "3"; us_shifted = "#"; end
8'h25: begin us_unshifted = "4"; us_shifted = "$"; end
8'h2E: begin us_unshifted = "5"; us_shifted = "%"; end
8'h36: begin us_unshifted = "6"; us_shifted = "^"; end
8'h3D: begin us_unshifted = "7"; us_shifted = "&"; end
8'h3E: begin us_unshifted = "8"; us_shifted = "*"; end
8'h46: begin us_unshifted = "9"; us_shifted = "("; end
// --- punctuation -----------------------------------------------------
8'h0E: begin us_unshifted = "`"; us_shifted = "~"; end
8'h4E: begin us_unshifted = "-"; us_shifted = "_"; end
8'h55: begin us_unshifted = "="; us_shifted = "+"; end
8'h54: begin us_unshifted = "["; us_shifted = "{"; end
8'h5B: begin us_unshifted = "]"; us_shifted = "}"; end
8'h5D: begin us_unshifted = 8'h5C; us_shifted = "|"; end // 0x5C = backslash
8'h4C: begin us_unshifted = ";"; us_shifted = ":"; end
8'h52: begin us_unshifted = "'"; us_shifted = 8'h22; end // 0x22 = double quote
8'h41: begin us_unshifted = ","; us_shifted = "<"; end
8'h49: begin us_unshifted = "."; us_shifted = ">"; end
8'h4A: begin us_unshifted = "/"; us_shifted = "?"; end
// --- the control keys a console needs --------------------------------
8'h29: begin us_unshifted = 8'h20; us_shifted = 8'h20; end // space
8'h5A: begin us_unshifted = 8'h0D; us_shifted = 8'h0D; end // Enter -> CR
8'h66: begin us_unshifted = 8'h08; us_shifted = 8'h08; end // Backspace -> BS
8'h0D: begin us_unshifted = 8'h09; us_shifted = 8'h09; end // Tab -> HT
8'h76: begin us_unshifted = 8'h1B; us_shifted = 8'h1B; end // Esc
// 0x00 = no character for this key. Modifiers land here too, which is
// correct: the decoder handles them before it ever consults this table.
// --- function keys - marker in BOTH columns: a VT220 sends the same
// code shifted or not, and there is no Shift-F encoding in this era
// (SPEC-vt100-keys.md, "Modifiers"). F1-F4 are DEC's keypad PF1-PF4
// by universal emulator convention; F5-F12 the VT220 top-row codes
// 15,17,18,19,20,21,23,24 - the gaps at 16 and 22 are DEC's.
// NOTE F7 = scancode 0x83, the one printable-range oddity in set 2.
8'h05: begin us_unshifted = 8'hB0; us_shifted = 8'hB0; end // F1 ESC O P
8'h06: begin us_unshifted = 8'hB1; us_shifted = 8'hB1; end // F2 ESC O Q
8'h04: begin us_unshifted = 8'hB2; us_shifted = 8'hB2; end // F3 ESC O R
8'h0C: begin us_unshifted = 8'hB3; us_shifted = 8'hB3; end // F4 ESC O S
8'h03: begin us_unshifted = 8'hCF; us_shifted = 8'hCF; end // F5 ESC [ 15 ~
8'h0B: begin us_unshifted = 8'hD1; us_shifted = 8'hD1; end // F6 ESC [ 17 ~
8'h83: begin us_unshifted = 8'hD2; us_shifted = 8'hD2; end // F7 ESC [ 18 ~
8'h0A: begin us_unshifted = 8'hD3; us_shifted = 8'hD3; end // F8 ESC [ 19 ~
8'h01: begin us_unshifted = 8'hD4; us_shifted = 8'hD4; end // F9 ESC [ 20 ~
8'h09: begin us_unshifted = 8'hD5; us_shifted = 8'hD5; end // F10 ESC [ 21 ~
8'h78: begin us_unshifted = 8'hD7; us_shifted = 8'hD7; end // F11 ESC [ 23 ~
8'h07: begin us_unshifted = 8'hD8; us_shifted = 8'hD8; end // F12 ESC [ 24 ~
default: begin us_unshifted = 8'h00; us_shifted = 8'h00; end
endcase
end
endmodule
`default_nettype wire