Skip to content

MON 505B (octal) - GetTrapReason (GERRC)

Reads the per-process ND-500 trap error code after a programmed trap, clears it, and returns it to the caller (GERRC = "get error code"): the reply message carries the code in field M505E with NUMPA=1.

Status: ND-500 level-12 call - the carved 51-word handler body is real SINTRAN L bytes (symbol boundary GERRC=141633B .. 5SIBM=141716B confirmed in L07); the level-12 GOSW[5] -> GERRC dispatch link is asserted, not tool-proven (see Honest caveats). All addresses/values are octal.

Dispatch path

MON 505B is an ND-500 level-12 call. It is not ND-100 GOTAB-dispatched (prove-mon.py 505: "NOT dispatched via the ND-100 GOTAB"). The dashed hop is the level-12 GOSW routing, which no tool run here proves.

flowchart TD
    A["ND-500 process<br/>MON 505B (stop-MOCALL)"] --> B["level-12 driver<br/>reads MCNO = 505B"]
    B --> C["5CMNO / L12MIN<br/>GOSW index 5"]
    C -.uncarved GOSW link.-> D["GERRC worker body<br/>026-S3IMPIT :141633B"]
    D --> E["read + clear trap ERREG<br/>build reply (M505E, NUMPA=1)"]
    class A blue
    class B,C teal
    class D,E green
    classDef blue fill:#E3F2FD,stroke:#0D47A1,color:#0D47A1
    classDef teal fill:#E0F7FA,stroke:#00838F,color:#00838F
    classDef green fill:#E8F5E9,stroke:#2E7D32,color:#2E7D32

The dashed hop (C ⇢ D) is the ND-500 level-12 GOSW entry into GERRC. prove-mon.py only walks the ND-100 GOTAB, so it can neither confirm nor deny this link - it is asserted by the folder attribution, not followed from a proven pointer.

Code location (dispatch path)

Every row is a real region you can open. Byte offset = (addr - loadbase) in octal words x 2. Segment 026-S3IMPIT load base = 32000B. Rows are in dispatch order.

Role Segment (full disasm) Addr range (octal) Byte offset (dec) Symbol Verdict
ND-100 GOTAB[505] slot (NOT this call's path) commoncode.asm · .hex 071740B (1 word) 59328 slot -> T105R=057455B MISATTRIBUTED - ND-500 call; the GOTAB slot is meaningless here (its T105R target lives in 025-S3IRPIT)
Level-12 GOSW dispatch (5CMNO/L12MIN, index 5) - (resident level-12 driver, not byte-located here) - - 5CMNO GOSW[5] UNVERIFIED - GOSW[5] -> GERRC is asserted by the folder, not proven by any tool run
GERRC worker body (read + clear trap error code) 026-S3IMPIT.asm · .hex 141633B-141715B (51 words) 73526 GERRC (next sym 5SIBM=141716B) real bytes; entry 051050=LDT I 50 matches on disk. VERIFIED (window); GOSW link MISATTRIBUTED
Indirect worker targets via pointer words 141711B/141712B/141713B 026-S3IMPIT + 023044B (below 026 base) 023044B, 145466B, 135067B not byte-confirmed (unnamed) UNVERIFIED - those words are indirect-pointer DATA; targets not confirmed as code

Verify by hand: grep '^141633 ' ../../segments-ref/026-S3IMPIT/026-S3IMPIT.hex -> byte offset 73526; then dd if=../../../segments/026-S3IMPIT.bin bs=1 skip=73526 count=8 status=none | od -An -tx1 -> 52 28 c6 e0 6a 27 f5 01 (= octal 051050 143340 065047 172401, the GERRC entry LDT I 50 / LDATX / ...).

Canonical layers for the handler segment: ../../segments-ref/026-S3IMPIT/ (026-S3IMPIT.asm, .hex, .symbols.txt, .meta.md); canonical bytes at ../../../segments/026-S3IMPIT.bin.

Instruction walkthrough

Full listing: 505B-GetTrapReason.ASM. Because this is overlay-recovered code and several words double as indirect-pointer data, the semantic reading is annotated but marked UNVERIFIED where the decode is ambiguous.

Entry prologue (141633-141641) - LDT I 50 / LDATX / SUB I 47 / AAA 1 / MPY 46 / ADD 46 / RADD. Loads a table base indirectly, indexes it, and does MPY/ADD index arithmetic - consistent with computing a per-process slot address (process_index * record_size + base). INFERRED which table.

Record fetch (141642-141653) - LDX 45 / LDA ,X 57 / ADD ,X 60 then a RADD register-shuffle chain. Reads two fields at fixed displacements (57, 60) off an X base - plausibly the per-proc trap/error record. INFERRED.

Clear-and-store (141654-141657) - LDDTX then STZTX / AAX 1 / STZTX: a double-word zero store to the indexed location. This is the strongest byte-level match to GERRC's documented "clear the error register after reading". VERIFIED as a zeroing store; INFERRED as to the target register.

Output setup (141660-141677) - a run of LDX I 30 / LDT I 22 / AAX .. / STDTX / STATX / STZTX staging words into an indexed block - plausibly building the reply message (M505E field + NUMPA=1). INFERRED mapping.

Indirect escapes (141700-141702) - every exit is indirect through a pointer word: 141700 JPL I 11 -> @141711 = 023044B; 141701 JPL I 11 -> @141712 = 145466B; 141702 JMP I 11 -> @141713 = 135067B (tail jump). Words 141711/141712/141713 are indirect-pointer DATA, so the disassembler's in-line decode of them (STD I ,X 44 / RORA ST DT / JPL I 67) is a data/code overlap artefact, not real instructions - and exact flow past 141702 is NOT byte-provable. Tail (141714-141715) is 000000 000000: alignment/padding.

There is no plain skip-return (EXIT/SKP) visible; every escape is JMP I/JPL I, which is why the validator reports ok.

Parameter / register contract

Item Dir Source Verdict
(no input parameter documented) in README/folder inferred (none decodable from bytes)
M505E (ND-500 message field) = trap error code out README inferred (field displacement not proven from bytes)
NUMPA = 1 out README inferred
double-word zero store at 141654-141657 internal .ASM VERIFIED as a zeroing store; INFERRED that the target is the trap ERREG
ND-100 registers ZAREG/ZXREG/ZTREG - - N/A - the caller ABI is the ND-500 message buffer, not the ND-100 MON register set; the code uses A/D/T/X internally only
skip return out .ASM none present; exit is via indirect JMP I/JPL I. VERIFIED (absence of direct skip)

The caller-visible convention lives in the ND-500 level-12 message buffer, not in this window, so the M505E/NUMPA field mapping is inferred, not byte-proven here.

Pseudo-code (for an emulator)

See 505B-GetTrapReason.pseudo.c - a pseudo-C model of the handler for emulator authors. Control flow inside the carved window is byte-verified; the semantic labels (which table, which field, which register is cleared) are inferred from the instruction shape and the folder description.

Every instruction in the .pseudo.c is translated against the canonical ND100-INSTRUCTION-SEMANTICS.md (bare LDx/LDT/AND disp = P-relative mem[P+disp], not literals; SKP/BSKP skip polarity; RADD CLD SD DA = A = D; T/X transfers LDATX/STATX/LDDTX/STDTX/STZTX = physical EL = ((T & 0377) << 16) | ((X + disp3) & 0177777)).

Honest caveats

What is byte-proven: the carved 51-word window is byte-exact, and its symbol boundary (GERRC=141633B .. 5SIBM=141716B) is confirmed in L07 (SYMBOL-2-LIST.SYMB.TXT). The entry word decodes cleanly (051050 = LDT I 50, on-disk bytes 52 28), the double-word zero store (141654-141657) matches the "read then clear" behaviour, and every direct branch stays in-file (validator ok).

What is NOT proven - one clear story: there are two independent dispatch facts and they do not agree. prove-mon.py's ND-100 GOTAB[505] points at T105R=057455B (a 4-word stub in overlay 025-S3IRPIT), not at GERRC. The folder's "MON 505B == GERRC" rests entirely on the ND-500 level-12 GOSW index-5 path (5CMNO/L12MIN), which is a separate dispatch chain that prove-mon.py does not walk - so it can neither confirm nor deny it. The reconciliation: MON 505B is an ND-500 call routed by the level-12 driver, so GOTAB[505] is simply the wrong table to read (it is MISATTRIBUTED for this call); but the GOSW[5] -> GERRC link itself is asserted by the folder, not followed from a proven pointer - hence UNVERIFIED. Additionally the pointer words 141711-141713 are indirect-pointer DATA, so control flow past 141702 and the worker targets (023044B/145466B/135067B) are not byte-provable, and the M505E/NUMPA field displacements are behavioural claims, not proven from the bytes. Confirming the dispatch needs the carved resident level-12 GOSW table decoded to show entry 5 -> GERRC=141633B.

Method: ../../../../../EXTRACTING-RESIDENT-CODE.md §7.6/7.7 · dispatch reality: ../../TASK-05-mismatches.md · master map: ../../MON-CALL-INDEX.md.