mirror/BFM-decomp
1
0
Fork 0
mirror of https://github.com/Druthulu/BFM-decomp synced 2026-10-06 17:11:04 -04:00
Code Issues Packages Projects Releases Wiki Activity

feat(phase-31): S79 #5 — the libpad 4.2.1 + libapi 4.2 band and the apicard region LINKED from real objects: 13 stubs + 4 TUs + the reorder island gone; main 16 stubs, fleet 38

Browse Source 
800c3 (0x8005CE18-0x8005FC68, one contiguous run of 33 interleaved Sony objects) is now four
stub rows — libapi1 (21 BIOS trampolines + COUNTER), libpad1 (PADENTRY + PADMAIN 760), libapi2
(L02/L03), libpad2 (PADCMD PADIF PADPORTD PADSEQD WAITRC2) — fed by two WINDOWED psyq_integrate
calls from the raw .run/obj42/{libapi42,libpad421} dirs (integrate tiles each stub with one
library; every boundary checked against .text SECTION sizes). The apicard region's three
"game code" rows were libapi 4.2's C objects to the byte: 800c2 = FIRST.o (firstfile + the
"no jump table wall" stub func_80062144), 800c2_2 = PAD.o, 800c2_3 = PATCH.o + CHCLRPAD.o ->
apicard5/6/7; make_apicard_used.py sources libapi from 4.2 (the EXE's real libapi; libcard
stays 4.0) into .run/obj42/apicard_used, 26 objects / 7 blocks, no game code left in
0x80061F38-0x80062888. src/800c3.c (129 hand-matched "C", 62 verbatim bodies, 19 stubs incl.
the four §332 %lo-in-a-delay-slot "walls"), src/800c2.c, src/800c2_2.c, src/800c2_3.c removed;
REORDER_TUS is empty (mechanism kept). Cookbook §490.

Two stale instruments fixed: exclude_audit let a pinned WALL outrank LINKED (PopMatrix/
PushMatrix had sat as walls since S68 while living in libgte3, linked since Phase 8) — LINKED
dominates now, config/wave_exclude.txt 13 -> 3; frontier_classify carried a hard-coded 49-name
LINKED set (R51) and reported 337 "stubs" — derived from the Makefile now.

Verified: main 143dbb89f34491258bbc27810d0a12ec8b43a8dd WITH all SDK dirs and WITHOUT them from
a fresh extract; make tools-health OK; R22 fleet extract-all 212/212 + check-all 213/213.
Metrics: main REAL 839->773, LINKED 1,150->1,256, VERBATIM 29->3, stubs 29->16, byte-identical
2,075/2,091 = 99.2%; game-code weighted 93.3% (38,748/41,534), remainder 2,786 = the open-stub
sum; fleet stubs 51->38 (frontier_classify: 39 rows incl. the data word). Verbatim manifest
33 -> 6. Docs: worklist rows + "S79 task #5", SETUP (fresh-clone obj42 commands, Makefile
blocks, exclude_audit), decision-log "S79 addendum 2", accelerators "S79 (2)", CURRENT_PHASE
S79 FINAL refreshed (census, metrics, the task #6 brief).
This commit is contained in:
Drew T
2026-09-04 17:53:02 -06:00
parent 58996ca4f7
commit 02f060f607
28 changed files with 609 additions and 2654 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Makefile
+39 -8
View File
@@ -607,13 +607,31 @@ SND_OBJDIR := build/psyq/snd
SND_SYMS := build/psyq/snd_externals.ld
SND_STUBS := snd1,snd2,snd3,snd4,snd5,snd6,snd7,snd8,snd9,snd10,snd11,snd12
# Combined libapi+libcard 800c2 region (Phase 8): 22 objects in 4 blocks (apicard1..4). Curated dir
# .run/obj40/apicard_used (tools/make_apicard_used.py). Window 0x61F38..0x62888. (libapi's ~22 objects
# in the 800c3 region are DEFERRED — lowest value.)
APICARD_ELF := .run/obj40/apicard_used
# Combined libapi+libcard 800c2 region (Phase 8; S79 #5): 26 objects in 7 blocks (apicard1..7) tiling
# 0x80061F38..0x80062888 with no game code left between them. Curated dir .run/obj42/apicard_used
# (tools/make_apicard_used.py: libapi from tools/psyq/lib421 = 4.2, the EXE's real libapi; libcard 4.0).
# Window 0x61F38..0x62888. The former 800c2/800c2_2/800c2_3 "game code" rows were FIRST.o / PAD.o /
# PATCH.o+CHCLRPAD.o (libapi 4.2 C objects) — apicard5/6/7.
APICARD_ELF := .run/obj42/apicard_used
APICARD_OBJDIR := build/psyq/apicard
APICARD_SYMS := build/psyq/apicard_externals.ld
APICARD_STUBS := apicard1,apicard2,apicard3,apicard4
APICARD_STUBS := apicard1,apicard2,apicard3,apicard4,apicard5,apicard6,apicard7
# The libapi 4.2 + libpad 4.2.1 band (S78 #12 named it, S79 #13 found the archive, S79 #5 wired it):
# 0x8005CE18..0x8005FC68 = 33 interleaved Sony objects, all byte-identical from tools/psyq/lib421
# (SCE's 1998-02-26 "libpad.lib 4.2.1 for the DUAL SHOCK" patch + libapi.lib 4.2; ELF regenerated into
# .run/obj42/{libapi42,libpad421} by psyq_lib_split.py + psyq-obj-parser — see docs/SETUP.md). Two raw
# dirs, two calls, each windowed to its own objects and tiling its own stubs (libapi1/libapi2,
# libpad1/libpad2). This was the src/800c3.c "REORDER_TUS island" — 129 hand-matched "C", 62 verbatim
# bodies, 19 stubs incl. the four §332 "%lo-in-a-delay-slot walls": Sony code assembled in reorder mode.
LIBAPI42_ELF := .run/obj42/libapi42
LIBAPI42_OBJDIR := build/psyq/libapi42
LIBAPI42_SYMS := build/psyq/libapi42_externals.ld
LIBAPI42_STUBS := libapi1,libapi2
LIBPAD_ELF := .run/obj42/libpad421
LIBPAD_OBJDIR := build/psyq/libpad
LIBPAD_SYMS := build/psyq/libpad_externals.ld
LIBPAD_STUBS := libpad1,libpad2
# Assembler flags (docs/SETUP.md §6.2). -G0 is confirmed by the disassembly
# (ledger #8: zero $gp-relative addressing). -no-pad-sections keeps section ends
@@ -763,7 +781,10 @@ CC1FLAGS := -quiet -O2 -G0 -mips1 -mcpu=3000 -mgas -msoft-float -fgnu-linker
# `lui at / jr ra / sw a0,lo(at)` read as COMPILER-INEXPRESSIBLE (P31 S75): a probe showed cc1 +
# reorder_passthrough + `as -O2` emits exactly that sequence. It was a build-config gap, not a
# gcc limit (cookbook §452 corrected).
REORDER_TUS := 800c2 800c2_2 800c2_3 800c3
# S79 #5: the island is EMPTY — all four TUs (800c2, 800c2_2, 800c2_3, 800c3) were libapi 4.2 / libpad 4.2.1
# objects and are LINKED now (apicard5-7, libapi1/2, libpad1/2). The mechanism stays for any future
# reorder-assembled TU (cookbook §332b); match_one/rtu_match read this list and tolerate it empty.
REORDER_TUS :=
ASFLAGS_REORDER := -Iinclude -march=r3000 -mtune=r3000 -no-pad-sections -O2 -G0
build/src/%.o: src/%.c
@@ -880,10 +901,20 @@ ifeq ($(BINARY),main)
if [ -d "$(APICARD_ELF)" ]; then
$(PYTHON) tools/psyq_integrate.py --vram-base $(main_VRAM_BASE) --exe $(main_EXE) --symbols $(main_SYMBOLS) --yaml $(main_SPLAT_YAML) $(APICARD_ELF) $(LD_SCRIPT) $(APICARD_OBJDIR) $(APICARD_SYMS) $(APICARD_STUBS) 0x80061F38 0x80062888
else
echo " (no $(APICARD_ELF) — apicard region stays asm stubs; run tools/psyq_build_libs.sh LIBAPI LIBCARD + tools/make_apicard_used.py)"
echo " (no $(APICARD_ELF) — apicard region stays asm stubs; run tools/psyq_build_libs.sh LIBCARD + the lib421 ELF step in docs/SETUP.md + tools/make_apicard_used.py)"
fi
if [ -d "$(LIBAPI42_ELF)" ]; then
$(PYTHON) tools/psyq_integrate.py --vram-base $(main_VRAM_BASE) --exe $(main_EXE) --symbols $(main_SYMBOLS) --yaml $(main_SPLAT_YAML) $(LIBAPI42_ELF) $(LD_SCRIPT) $(LIBAPI42_OBJDIR) $(LIBAPI42_SYMS) $(LIBAPI42_STUBS) 0x8005CE18 0x8005E188
else
echo " (no $(LIBAPI42_ELF) — libapi band blocks stay asm stubs; convert tools/psyq/lib421/LIBAPI.LIB per docs/SETUP.md)"
fi
if [ -d "$(LIBPAD_ELF)" ]; then
$(PYTHON) tools/psyq_integrate.py --vram-base $(main_VRAM_BASE) --exe $(main_EXE) --symbols $(main_SYMBOLS) --yaml $(main_SPLAT_YAML) $(LIBPAD_ELF) $(LD_SCRIPT) $(LIBPAD_OBJDIR) $(LIBPAD_SYMS) $(LIBPAD_STUBS) 0x8005D0D8 0x8005FC68
else
echo " (no $(LIBPAD_ELF) — libpad band blocks stay asm stubs; convert tools/psyq/lib421/LIBPAD.LIB per docs/SETUP.md)"
fi
endif
SYMS=""; [ -f "$(LIBCD_SYMS)" ] && SYMS="-T $(LIBCD_SYMS)"; [ -f "$(LIBGS_SYMS)" ] && SYMS="$$SYMS -T $(LIBGS_SYMS)"; [ -f "$(LIBETC_SYMS)" ] && SYMS="$$SYMS -T $(LIBETC_SYMS)"; [ -f "$(LIBGPU_SYMS)" ] && SYMS="$$SYMS -T $(LIBGPU_SYMS)"; [ -f "$(LIBMCRD_SYMS)" ] && SYMS="$$SYMS -T $(LIBMCRD_SYMS)"; [ -f "$(LIBC2_SYMS)" ] && SYMS="$$SYMS -T $(LIBC2_SYMS)"; [ -f "$(LIBGTE_SYMS)" ] && SYMS="$$SYMS -T $(LIBGTE_SYMS)"; [ -f "$(SND_SYMS)" ] && SYMS="$$SYMS -T $(SND_SYMS)"; [ -f "$(APICARD_SYMS)" ] && SYMS="$$SYMS -T $(APICARD_SYMS)"
SYMS=""; [ -f "$(LIBCD_SYMS)" ] && SYMS="-T $(LIBCD_SYMS)"; [ -f "$(LIBGS_SYMS)" ] && SYMS="$$SYMS -T $(LIBGS_SYMS)"; [ -f "$(LIBETC_SYMS)" ] && SYMS="$$SYMS -T $(LIBETC_SYMS)"; [ -f "$(LIBGPU_SYMS)" ] && SYMS="$$SYMS -T $(LIBGPU_SYMS)"; [ -f "$(LIBMCRD_SYMS)" ] && SYMS="$$SYMS -T $(LIBMCRD_SYMS)"; [ -f "$(LIBC2_SYMS)" ] && SYMS="$$SYMS -T $(LIBC2_SYMS)"; [ -f "$(LIBGTE_SYMS)" ] && SYMS="$$SYMS -T $(LIBGTE_SYMS)"; [ -f "$(SND_SYMS)" ] && SYMS="$$SYMS -T $(SND_SYMS)"; [ -f "$(APICARD_SYMS)" ] && SYMS="$$SYMS -T $(APICARD_SYMS)"; [ -f "$(LIBAPI42_SYMS)" ] && SYMS="$$SYMS -T $(LIBAPI42_SYMS)"; [ -f "$(LIBPAD_SYMS)" ] && SYMS="$$SYMS -T $(LIBPAD_SYMS)"
echo " LD $(ELF)"
$(LD) -T $(LD_SCRIPT) -T $(UNDEF_SYMS) -T $(UNDEF_FUNCS) $$SYMS --no-check-sections -Map $(MAPFILE) -o $(ELF)
echo " OBJCOPY $@"
config/splat.us.exe.yaml
+21 -7
View File
@@ -201,21 +201,35 @@ segments:
- [0x4CAC8, c, libc2_1] # libc2 main block (16 objs) -> src/libc2_1.c (vram 0x8005C2C8-0x8005CE18);
# end = SETJMP.o's .text SECTION size (0x80, 8-aligned), NOT its 30-ins
# count (0x78) — the trailing 8-byte align pad belongs to the object.
- [0x4D618, c, 800c3] # -O2 game code -> src/800c3.c (vram 0x8005CE18-0x8005FC68)
# PsyQ LIBAPI 4.2 + LIBPAD 4.2.1 band (P31 S78 #12 named it, S79 #13 found the archive, S79 #5 wired it):
# 0x8005CE18-0x8005FC68 is ONE contiguous run of 33 Sony objects, interleaved libapi/libpad, every one
# byte-identical from tools/psyq/lib421 (SCE's 1998-02-26 "libpad.lib 4.2.1 for the DUAL SHOCK" patch,
# shipped with libapi.lib 4.2 — the EXE's 4.2 / 4.2.1x `Ps` stamps). It used to be src/800c3.c: 129
# hand-matched "C", 62 verbatim bodies and 19 stubs incl. the four §332 "%lo-in-a-delay-slot walls" —
# all of it Sony code assembled in reorder mode (§332b), none of it game code. Split into 4 rows because
# psyq_integrate tiles each stub with ONE library's objects (windows: libapi 0x8005CE18..0x8005E188,
# libpad 0x8005D0D8..0x8005FC68). Object sizes are .text SECTION sizes (exact tile, verified S79).
- [0x4D618, c, libapi1] # libapi 4.2 block 1: C57 C68 C73 C114 A07-A13 A23-A25 A36 A37 A52 A53 A91 L10 (21 trampolines) + COUNTER (92) -> src/libapi1.c (vram 0x8005CE18-0x8005D0D8)
- [0x4D8D8, c, libpad1] # libpad 4.2.1 block 1: PADENTRY (300) + PADMAIN (760) -> src/libpad1.c (vram 0x8005D0D8-0x8005E168)
- [0x4E968, c, libapi2] # libapi 4.2 block 2: L02 + L03 (SysEnqIntRP/SysDeqIntRP) -> src/libapi2.c (vram 0x8005E168-0x8005E188)
- [0x4E988, c, libpad2] # libpad 4.2.1 block 2: PADCMD (600) PADIF (376) PADPORTD (408) PADSEQD (288) WAITRC2 (48) -> src/libpad2.c (vram 0x8005E188-0x8005FC68)
# PsyQ libmcrd (Phase 8): 2 objects — LIBMCRD.o (2186 ins; the 55 LIBMCRD_OBJ_* + _card_* memcard
# I/O) and USERFUNC.o (68 ins, at the very end), 2 non-adjacent blocks split by game/libc2/libapi/
# libcard code (800c2). psyq_integrate libmcrd1,libmcrd2. (The GAME save logic / Q#5 is Phase 12.)
- [0x50468, c, libmcrd1] # libmcrd block 1: LIBMCRD.o -> src/libmcrd1.c (vram 0x8005FC68-0x80061E90)
- [0x52690, c, libc2_2] # libc2 block 2: STRCAT.o -> src/libc2_2.c (vram 0x80061E90-0x80061F38)
# PsyQ libapi+libcard COMBINED 800c2 region (Phase 8): BIOS syscall trampolines + card I/O,
# 22 objects in 4 blocks (curated by tools/make_apicard_used.py; C112.o shared, no exclusions).
# integrate window 0x61F38..0x62888. (libapi's ~22 objects in the 800c3 region are DEFERRED.)
# PsyQ libapi+libcard COMBINED 800c2 region (Phase 8; S79 #5): BIOS syscall trampolines + card I/O,
# now 26 objects in 7 blocks tiling 0x80061F38-0x80062888 with NO game code between them (curated by
# tools/make_apicard_used.py: libapi from tools/psyq/lib421 = 4.2, the EXE's real libapi; libcard 4.0,
# byte-identical; C112.o shared, no exclusions). integrate window 0x61F38..0x62888. The three former
# "game code" rows were libapi 4.2's C objects: 800c2 = FIRST.o (168 ins, `firstfile` + the "wall"
# stub func_80062144), 800c2_2 = PAD.o (192), 800c2_3 = PATCH.o (40) + CHCLRPAD.o (28) — exact tiles.
- [0x52738, c, apicard1] # apicard block 1: 7 obj (C112.o..A69.o) vram 0x80061F38-0x80061FA8
- [0x527A8, c, 800c2] # game code (vram 0x80061FA8-0x80062248)
- [0x527A8, c, apicard5] # apicard block 5 (S79 #5): FIRST.o (168 ins, exact tile) vram 0x80061FA8-0x80062248 — was '800c2'
- [0x52A48, c, apicard2] # apicard block 2: 7 obj (A66.o..INIT.o) vram 0x80062248-0x80062388
- [0x52B88, c, 800c2_2] # game code (vram 0x80062388-0x80062688)
- [0x52B88, c, apicard6] # apicard block 6 (S79 #5): PAD.o (192 ins, exact tile) vram 0x80062388-0x80062688 — was '800c2_2'
- [0x52E88, c, apicard3] # apicard block 3: 4 obj (A18.o..A21.o) vram 0x80062688-0x800626C8
- [0x52EC8, c, 800c2_3] # game code (vram 0x800626C8-0x800627D8)
- [0x52EC8, c, apicard7] # apicard block 7 (S79 #5): PATCH.o (40) + CHCLRPAD.o (28), exact tile, vram 0x800626C8-0x800627D8 — was '800c2_3'
- [0x52FD8, c, apicard4] # apicard block 4: 4 obj (A74.o..END.o) vram 0x800627D8-0x80062888
- [0x53088, c, libmcrd2] # libmcrd block 2: USERFUNC.o -> src/libmcrd2.c (vram 0x80062888-0x80062998)
# Phase 7 (Task 2' / LZSS) — SURGICAL rodata carve for the LZSS switch only.
config/verbatim_manifest.json
-324
View File
@@ -12,246 +12,6 @@
],
"generated": "P31 S75 from .run/S75/triage/taxonomy.json",
"rows": [
{
"binary": "main",
"fn": "ChangeClearPAD",
"addr": "0x8005cf48",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "ChangeClearPAD",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "ChangeClearRCnt",
"addr": "0x8005cf58",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "ChangeClearRCnt",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "CloseEvent",
"addr": "0x8005ce88",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "CloseEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "DeliverEvent",
"addr": "0x8005ce68",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "DeliverEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "DisableEvent",
"addr": "0x8005cec8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "DisableEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "DisablePAD",
"addr": "0x800626dc",
"nins": 5,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_800626dc",
"unit_nins": 5,
"path": "src/800c2_3.c",
"why": "lui/lw/nop/jr $t1 through jtbl_80078D24 \u2014 vector hook"
},
{
"binary": "main",
"fn": "EnableEvent",
"addr": "0x8005ceb8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "EnableEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "EnablePAD",
"addr": "0x800626c8",
"nins": 5,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_800626c8",
"unit_nins": 5,
"path": "src/800c2_3.c",
"why": "lui/lw/nop/jr $t1 through jtbl_80078D20 \u2014 vector hook"
},
{
"binary": "main",
"fn": "FlushCache",
"addr": "0x8005ce28",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "FlushCache",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "HookEntryInt",
"addr": "0x8005cef8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "HookEntryInt",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "InitHeap",
"addr": "0x8005ce18",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "InitHeap",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "OpenEvent",
"addr": "0x8005ce78",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "OpenEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "ResetEntryInt",
"addr": "0x8005cee8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "ResetEntryInt",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "ReturnFromException",
"addr": "0x8005ced8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "ReturnFromException",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "SetInitPadFlag",
"addr": "0x80062388",
"nins": 3,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "SetInitPadFlag",
"unit_nins": 3,
"path": "src/800c2_2.c",
"why": "lui $at; jr $ra; sw $a0,lo($at) \u2014 hand-written setter (libcard region, 800c2 reorder island)"
},
{
"binary": "main",
"fn": "SysDeqIntRP",
"addr": "0x8005e178",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "SysDeqIntRP",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "SysEnqIntRP",
"addr": "0x8005e168",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "SysEnqIntRP",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "TestEvent",
"addr": "0x8005cea8",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "TestEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "WaitEvent",
"addr": "0x8005ce98",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "WaitEvent",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "_96_remove",
"addr": "0x8005ce50",
"nins": 6,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "_96_remove",
"unit_nins": 6,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "__do_global_dtors",
@@ -276,78 +36,6 @@
"path": "src/boot.c",
"why": "libgcc2.c C compiled -O2 inside the -O0 boot TU (per src comment); gcc-2.7.2 has no per-function opt level"
},
{
"binary": "main",
"fn": "_padInitDirSeq",
"addr": "0x8005f728",
"nins": 13,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_8005f728",
"unit_nins": 13,
"path": "src/800c3.c",
"why": "jr-slot symbolic store shape (as -O2 reorder island form)"
},
{
"binary": "main",
"fn": "_patch_pad",
"addr": "0x800626f0",
"nins": 30,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_800626f0",
"unit_nins": 30,
"path": "src/800c2_3.c",
"why": "hand-written (sibling of _remove_ChgclrPAD)"
},
{
"binary": "main",
"fn": "_remove_ChgclrPAD",
"addr": "0x80062768",
"nins": 28,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "_remove_ChgclrPAD",
"unit_nins": 28,
"path": "src/800c2_3.c",
"why": "splat Handwritten; B0-table call with $ra stashed in D_80078D28"
},
{
"binary": "main",
"fn": "func_8005CE38",
"addr": "0x8005ce38",
"nins": 0,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_8005ce38",
"unit_nins": 0,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "read",
"addr": "0x8005cf28",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "read",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "main",
"fn": "setRC2wait",
"addr": "0x8005fba8",
"nins": 8,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "func_8005fba8",
"unit_nins": 8,
"path": "src/800c3.c",
"why": "jr-slot symbolic store; reads 0x1F801120 (RCNT) \u2014 hand-written"
},
{
"binary": "main",
"fn": "start",
@@ -360,18 +48,6 @@
"path": "src/boot.c",
"why": "crt0: `addi` (trapping add), break 1, $gp load; splat Handwritten"
},
{
"binary": "main",
"fn": "write",
"addr": "0x8005cf38",
"nins": 4,
"cls": "SDK-ASM",
"disposition": "PERMANENT-VERBATIM",
"unit_entry": "write",
"unit_nins": 4,
"path": "src/800c3.c",
"why": "BIOS A/B/C-table trampoline: li $t2,0xA0/B0/C0; jr $t2; li $t1,N; pad nop (or li $a0; syscall)"
},
{
"binary": "md_MAIN_003",
"fn": "func_800D3204",
config/wave_exclude.txt
+1 -18
View File
@@ -1,24 +1,7 @@
# regenerated by tools/exclude_audit.py from config/wave_exclude.txt
# 11 still-valid of 27; 16 dropped as stale (banked / linked / blocker-since-fixed).
# 3 still-valid of 11; 8 dropped as stale (banked / linked / blocker-since-fixed).
# An exclude list records what the TOOLING could not do — regenerate it as
# part of every tool fix, or it becomes a list of work you decided not to do.
main:PopMatrix # WALL: compiler wall (merged from S68_walls_332.txt)
main:PushMatrix # WALL: compiler wall (merged from S68_walls_332.txt)
main:_padInitSioMode # WALL: no jump table — curated compiler fact, not a tooling limit
main:func_8005ECC0 # NEAR-2/35 (DELAY-SLOT) — NOT a wall. Same wrong-oracle provenance as
# PadInitDirect; re-measured S77 at 2 of 35. permuter_ils 4 cycles reached 5, not 0.
# Excluded as UNSOLVED, not as impossible.
main:PadInitDirect # NEAR-3/88 (ADDRESSING/move!=lui) — NOT a wall. The S68 'compiler wall'
# verdict was measured under the WRONG oracle (800c3 is a REORDER_TUS island; S76 fixed
# match_one/rtu_match to use reorder_passthrough + as -O2). Re-measured S77: 3 of 88.
# permuter_ils 4 cycles reached 3, not 0. Excluded as UNSOLVED, not as impossible.
main:func_80062144 # WALL: no jump table — curated compiler fact, not a tooling limit
md_MAIN_034:func_800CB00C # WALL: compiler wall (merged from S68_walls_332.txt)
ov_SC03_105:func_801834A4 # WALL: loop.c movable ordering, closeness 6 (S71)
ov_SC06_022:func_8017DF28 # WALL: expand_block_move copy_addr_to_reg pseudo + cse reuse, closeness 2 (S71)
# S77: wall_sweep named these and the list did not carry them — an agent handed one
# always returns a NEAR with an unexplainable tail, indistinguishable from a hard function.
main:func_8005D734 # WALL: §332 %lo in a delay slot (wall_sweep, 1 site) — no C can place it
main:_padStartCom # WALL: §332 %lo in a delay slot (wall_sweep, 1 site) — no C can place it
main:func_8005ED4C # WALL: §332 %lo in a delay slot (wall_sweep, 2 sites) — no C can place it
main:func_8005F450 # WALL: §332 %lo in a delay slot (wall_sweep, 2 sites) — no C can place it
docs/SETUP.md
+4 -2
View File
@@ -739,7 +739,7 @@ Every script under `tools/` (plus the two report make-targets), grouped by purpo
| | `tools/make_libgs.sh` | Build/link the `libgs` block (cookbook §9). |
| | `tools/gen_lib_subsegs.py` | **(Phase 8)** Generate splat subseg lines + integrate stub list for a multi-block library (section-size-correct boundaries; cookbook §9.6). |
| | `tools/make_snd_used.py` | **(Phase 8)** Build the combined libspu+libsnd curated dir (alias dedup by byte-match; 3 address exclusions since S78 #4 — VM_F rejoined via the `.bss` split; §9.6). |
| | `tools/make_apicard_used.py` | **(Phase 8)** Build the combined libapi+libcard curated dir (§9.6). |
| | `tools/make_apicard_used.py` | **(Phase 8; S79 #5)** Build the combined libapi+libcard curated dir `.run/obj42/apicard_used` — libapi **4.2** (`.run/obj42/libapi42`, the EXE's real libapi, incl. FIRST/PAD/PATCH/CHCLRPAD) + libcard 4.0 (§9.6, §490). |
| | `tools/ld_interleave.py` | Interleave linker inputs to match the original section ordering. Three forms: `--front/--tail` (the original data->rodata->data sandwich), **`--order`** (P31 S72 — an address-ordered leaf list, needed once a binary owns SEVERAL `.rodata` carves: a `*.data.o` leaf contributes its `(.data)`, a code-object leaf its `(.rodata)`; main's island is a 7-piece sandwich `--front/--tail` cannot express), and **`--pre`** (P31 S74, cookbook §440 — a piece that must land BEFORE the text). `--pre` exists for the **resident**, which opens with `- [0x0, rodata, hdr]`, a 1-word `.rodata` header ahead of its code: every `--order` piece is emitted after TEXT_START, so `hdr.rodata.o(.rodata)` would fall into the unchecked `empties` bucket, be parked after the text, and move every byte in the binary. |
| | `tools/jtbl_rodata_pads.py` | **(Phase 29, cookbook §8e)** Post-maspsx filter for multi-table `.rodata` carve spans: REPLACES each cc1 `.align 3` (which is section-relative and would mis-pad a merged/4-mod-8 span) with the ORIGINAL's exact pad bytes per the object's `JTBL_PADS` spec (written by `jtbl_carve` into `config/overlays.mk`; armed via `$(if $(JTBL_PADS),…)` in the Makefile `build/src/%.o` recipe). Fail-loud on table-count drift / non-`.align 3` / non-jtbl rodata content. Unset var ⇒ pipeline byte-identical. |
| | `tools/split_src_region.py` | Split a `src/` region file at object boundaries. |
@@ -999,7 +999,9 @@ fills fast). Nothing is leaking — but the host does not get the memory back on
| `tools/psyq_link.py` / `psyq_link_region.py` / `psyq_integrate.py` | the NOBITS predicate is `^\.s?bss\d*$` (was the literal pair `.bss`/`.sbss`), so split pieces are NOLOAD-placed from their own section symbols; `link_object` links with `--no-check-sections` like the build (piece extents tile the PACKED section, so unreferenced commons overlap other pieces harmlessly) | automatic |
| `LIBGPU_ELF := .run/obj40/libgpu` | the `libgpu_used` curated dir is retired (it existed only to exclude SYS.o); `psyq_identify` drops the 8 objects the EXE never links | fresh clone: `tools/psyq_build_libs.sh LIBGPU` is now the whole libgpu step |
| `tools/make_libgs.sh` (+GS_001) / `tools/make_snd_used.py` (VM_F rejoined) | curated dirs now carry the two objects the split unlocked: libgs 34 objects / 8 blocks (`libgs8`), sound 63 / 12 (`snd12`) | re-run after `psyq_build_libs.sh` on a fresh clone |
| `tools/psyq/lib42/` + `lib421/` (task #13) | PsyQ RTL 4.2 + the 1998-02-26 libpad 4.2.1 patch (see the archive table above); `.run/obj42/libpad421` (7 placed) + `.run/obj42/libapi42` (39 placed), all byte-identical | task #5 wires them (band + apicard region); regenerate the ELF dirs with `psyq_lib_split.py <LIB> .run/obj42/<name>` + `psyq-obj-parser` per object |
| `tools/psyq/lib42/` + `lib421/` (task #13) | PsyQ RTL 4.2 + the 1998-02-26 libpad 4.2.1 patch (see the archive table above); `.run/obj42/libpad421` (7 placed) + `.run/obj42/libapi42` (39 placed), all byte-identical | **wired in task #5** (below); fresh clone: `7z x tools/psyq/PlayStation_Programmer_Tool_-_Runtime_Library_Version_4.2.7z -o.run/psyq_hunt/rtl42 && unzip .run/psyq_hunt/rtl42/LIB/42PATCH/J421PD.ZIP -d tools/psyq/lib421`, then for each of `lib421/LIBPAD.LIB`→`.run/obj42/libpad421` and `lib421/LIBAPI.LIB`→`.run/obj42/libapi42`: `python3 tools/psyq_lib_split.py <LIB> <dir>` and `for o in <dir>/*.obj; do tools/psyq/psyq-obj-parser "$o" -o "${o%.obj}.o"; done` |
| Makefile `LIBAPI42_*` / `LIBPAD_*` blocks + `APICARD_ELF := .run/obj42/apicard_used` (task #5) | the band 0x8005CE18–0x8005FC68 = 4 stub rows (`libapi1`, `libpad1`, `libapi2`, `libpad2`) fed by two WINDOWED `psyq_integrate` calls from the raw obj42 dirs; the apicard region = 7 rows (`apicard1..7`, the former `800c2`/`800c2_2`/`800c2_3` game-code rows were FIRST/PAD/PATCH+CHCLRPAD) from `tools/make_apicard_used.py`, which now takes libapi from `.run/obj42/libapi42` (4.2) and libcard from `.run/obj40/libcard` (4.0). `REORDER_TUS` is EMPTY (all four island TUs were Sony objects; the mechanism stays) | automatic in `make build BINARY=main`; the build is byte-identical WITHOUT any of the obj dirs (stub fallback), as always |
| `tools/exclude_audit.py` | LINKED now outranks a pinned `# WALL` (a function in a linked PsyQ subseg is never a target, whatever compiler fact was recorded about it): dropped the four §332 walls + `func_80062144` + `PopMatrix`/`PushMatrix` (libgte3 since Phase 8) | every regeneration of `config/wave_exclude.txt` |
**Oracle corrections — re-read any verdict recorded before these:**
* `match_one` and `rtu_match` now route the Makefile's `REORDER_TUS` (`800c2 800c2_2 800c2_3 800c3`)
docs/accelerators.md
+9
View File
@@ -677,3 +677,12 @@ independent oracle (here the other objects' by-name recoveries agreed on all sev
the transformation in the shared prepare path, not in a curated artifact, so it is re-derived from the
bytes on every build and negative-controlled over the whole placed population for free.
## S79 (2) — provenance → archive → link → compiler, in that order
Twelve stubs in one band carried compiler-wall verdicts from three sessions of wall-sweeping. The band was
Sony's controller library; the exact library was a 383 KB archive.org download away; every object linked
byte-identical on the first try. Accelerator: **before any wall verdict on a function in a mixed binary,
ask whose code it is and whether the shipped object exists** — the loader's signature sets name the
library and version for free (§487), and the SDK archives are indexed on archive.org by version. A wall
verdict on code you were never going to compile is correct and worthless.
docs/cookbook-index.md
+5 -2
View File
@@ -2,7 +2,7 @@
> **Generated by `tools/cookbook_index.py` — do not hand-edit** (R33). Regenerate after adding a cookbook section.
>
> `docs/matching-cookbook.md` is ~716 KB / 1157 sections. Grepping it blind is how three P30 wave-1 agents each "discovered" an idiom that was already written down. **Start here, then read the section.** A section appears under every symptom it addresses.
> `docs/matching-cookbook.md` is ~716 KB / 1158 sections. Grepping it blind is how three P30 wave-1 agents each "discovered" an idiom that was already written down. **Start here, then read the section.** A section appears under every symptom it addresses.
**How to use:** name what you SEE in the diff (a stolen delay slot, an extra `la`, a swapped register pair, a `conflicting types` error), find that symptom below, read those sections first. If nothing fits, THEN grind — and add a section when you win.
@@ -1080,7 +1080,7 @@
- **§442** — ★★★ — A RECOVERY RUNG THAT REWRITES A LEGAL CONSTRUCT INTO AN ILLEGAL ONE READS EXACTLY LIKE A CODEGEN WALL (P31 S75; `reconcile_tu`, 344 ins unblocked) <sub>L34647</sub>
- **§477** — ★★★ — THE `self_decl_tu` CLASS IS A SOLVED, MECHANICAL LANE: 16 DRAFTS, 16 BANKS (P31 S77) <sub>L35713</sub>
### build graph, splat & the harness (209)
### build graph, splat & the harness (210)
- **§4** — Flag/toolchain gotchas <sub>L190</sub>
- **Build** — mechanism — per-file opt override (splat resegmentation) <sub>L307</sub>
@@ -1291,6 +1291,7 @@
- **§487** — ★★★ — THE PSX LOADER'S PER-VERSION SIGNATURE SETS ARE A FREE PROVENANCE ORACLE: main's "WALL" BAND IS LIBPAD 4.2.1 (P31 S78) <sub>L36150</sub>
- **§488** — ★★ — THE "GAME CODE" GAPS BETWEEN LIBRARY BLOCKS WERE LIBRARY OBJECTS: 13 SUBSEGS → LINKED, EXACT-TILED, ZERO TOKENS (P31 S78) <sub>L36202</sub>
- **§489** — ★★★ — SCATTERED `.bss` IS A PARTITION PROBLEM: SPLIT THE SECTION INTO PER-BASE PIECES AT LINK-PREPARE (P31 S78 #4; closes the §9.1 exclusion class 3/3) <sub>L36242</sub>
- **§490** — ★★★ — THE "WALL" BAND WAS A LIBRARY VERSION AWAY: LIBPAD 4.2.1 + LIBAPI 4.2 FOUND, THE WHOLE 0x8005CE18–0x8005FC68 BAND + THE APICARD REGION LINKED FROM REAL OBJECTS (P31 S79 #13/#5) <sub>L36300</sub>
### process, measurement & doctrine (141)
@@ -2936,6 +2937,7 @@
- **§487** — ★★★ — THE PSX LOADER'S PER-VERSION SIGNATURE SETS ARE A FREE PROVENANCE ORACLE: main's "WALL" BAND IS LIBPAD 4.2.1 (P31 S78) <sub>L36150</sub>
- **§488** — ★★ — THE "GAME CODE" GAPS BETWEEN LIBRARY BLOCKS WERE LIBRARY OBJECTS: 13 SUBSEGS → LINKED, EXACT-TILED, ZERO TOKENS (P31 S78) <sub>L36202</sub>
- **§489** — ★★★ — SCATTERED `.bss` IS A PARTITION PROBLEM: SPLIT THE SECTION INTO PER-BASE PIECES AT LINK-PREPARE (P31 S78 #4; closes the §9.1 exclusion class 3/3) <sub>L36242</sub>
- **§490** — ★★★ — THE "WALL" BAND WAS A LIBRARY VERSION AWAY: LIBPAD 4.2.1 + LIBAPI 4.2 FOUND, THE WHOLE 0x8005CE18–0x8005FC68 BAND + THE APICARD REGION LINKED FROM REAL OBJECTS (P31 S79 #13/#5) <sub>L36300</sub>
---
@@ -4105,3 +4107,4 @@ Notes routinely quote that as a section id. This table resolves it. Grep bait: `
| L36150 | §487 | ★★★ — THE PSX LOADER'S PER-VERSION SIGNATURE SETS ARE A FREE PROVENANCE ORACLE: main's "WA |
| L36202 | §488 | ★★ — THE "GAME CODE" GAPS BETWEEN LIBRARY BLOCKS WERE LIBRARY OBJECTS: 13 SUBSEGS → LINKED |
| L36242 | §489 | ★★★ — SCATTERED `.bss` IS A PARTITION PROBLEM: SPLIT THE SECTION INTO PER-BASE PIECES AT L |
| L36300 | §490 | ★★★ — THE "WALL" BAND WAS A LIBRARY VERSION AWAY: LIBPAD 4.2.1 + LIBAPI 4.2 FOUND, THE WHO |
docs/decision-log.md
+29
View File
@@ -3141,3 +3141,32 @@ probe's stricter test then ratified one of them as a wall. General form: when a
X reproduces it", the next question is "can X be partitioned", and the partition should follow the
structure the ORIGINAL producer used (here: symbols), not the structure the measurement happened to
group by.
### S79 addendum 2 (2026-09-04, tasks #13/#5) — the band's twelve "walls" were a library version away; the §332 wall verdicts are closed
**Belief (S68 → S78).** Twelve of main's open stubs sat in `800c3`/`800c2`, four of them curated as §332
"%lo-in-a-delay-slot" compiler walls, one as a "no jump table" wall, all excluded from waves as facts about
gcc. S78 named the band (libpad 4.2.1 + libapi 4.2, §487) but expected to finish it as C under the
reorder island because no archive we held could link it; the S78 brief budgeted a bounded hunt (#13) with
three leads and a fallback to C.
**What happened.** The first lead — archive.org — held a 383 KB 7z of the Runtime Library 4.2 with SCE's
February-1998 libpad 4.2.1 patch inside. Every one of the 46 objects it places in main links byte-identical
(cookbook §490). Task #5 then linked the whole band and the apicard region's C objects (FIRST, PAD, PATCH,
CHCLRPAD) in one pass: four TUs deleted, the `REORDER_TUS` island empty, main byte-identical with and
without the SDK dirs, fleet 213/213.
**Why the walls were wrong as WORK while right as FACTS.** §332 is a true statement about cc1 + `as -O2`
reorder output: no C source reproduces a `%lo` in a delay slot through our pipeline. But Sony never
compiled that C through our pipeline; they shipped the object. A wall verdict answers "can we match this
function?"; it never asks "is this function ours to match?" — and that provenance question, once asked
(§487), was answered by a 383 KB download. The exclude audit had the same blind spot in code: a pinned
`# WALL` outranked LINKED, so `PopMatrix`/`PushMatrix` sat as walls for eleven sessions while living in
libgte3, linked since Phase 8. Fixed: LINKED dominates.
**Hindsight.** The order of questions for any stubborn function in a mixed binary is provenance → archive
→ link → (only then) compiler. We ran it backwards for the band: wall_sweep in S68/S77, names in S78,
archive in S79. The §487 sentence — "a wall inside bytes no archive you hold can place is a provenance
question before it is a compiler question" — now has its second half: **and a provenance answer is an
archive to go and find, not a label to match under.**
docs/difficulty.md
+11 -11
View File
@@ -1,9 +1,9 @@
# Unmatched difficulty inventory (generated by tools/difficulty.py — harvest queue)
unmatched functions : 1179
trivial (<=5 ins) : 238
non-jtbl leaves : 797 (best harvest targets)
jump-table funcs : 36 (deferred — need the rodata-island workflow, Task 2')
unmatched functions : 1272
trivial (<=5 ins) : 267
non-jtbl leaves : 842 (best harvest targets)
jump-table funcs : 58 (deferred — need the rodata-island workflow, Task 2')
## Easiest 120 unmatched (score asc) — the work queue
| score | name | nins | br | calls | jtbl | leaf |
@@ -56,7 +56,9 @@ jump-table funcs : 36 (deferred — need the rodata-island workflow, Task
| 3 | SetDQB | 3 | 0 | 0 | - | Y |
| 3 | SetData32 | 3 | 0 | 0 | - | Y |
| 3 | SetIR0 | 3 | 0 | 0 | - | Y |
| 3 | SetInitPadFlag | 3 | 0 | 0 | - | Y |
| 3 | UT_KEYV_OBJ_174 | 3 | 0 | 0 | - | Y |
| 3 | _padSetAct | 3 | 0 | 0 | - | Y |
| 4 | CD_set_test_parmnum | 4 | 0 | 0 | - | Y |
| 4 | C_011_OBJ_960 | 4 | 0 | 0 | - | Y |
| 4 | CdLastCom | 4 | 0 | 0 | - | Y |
@@ -71,6 +73,7 @@ jump-table funcs : 36 (deferred — need the rodata-island workflow, Task
| 4 | LIBMCRD_OBJ_84C | 4 | 0 | 0 | - | Y |
| 4 | LIBMCRD_OBJ_A5C | 4 | 0 | 0 | - | Y |
| 4 | MIDIREAD_OBJ_6AC | 4 | 0 | 0 | - | Y |
| 4 | ReadInitPadFlag | 4 | 0 | 0 | - | Y |
| 4 | SPU_OBJ_510 | 4 | 0 | 0 | - | Y |
| 4 | SPU_OBJ_AC4 | 4 | 0 | 0 | - | Y |
| 4 | SSSTART_OBJ_240 | 2 | 0 | 1 | - | - |
@@ -92,6 +95,7 @@ jump-table funcs : 36 (deferred — need the rodata-island workflow, Task
| 4 | UT_KEYV_OBJ_3FC | 4 | 0 | 0 | - | Y |
| 4 | VSYNC_OBJ_1D4 | 4 | 0 | 0 | - | Y |
| 4 | VS_VH_OBJ_180 | 4 | 0 | 0 | - | Y |
| 4 | _padSetCmd | 4 | 0 | 0 | - | Y |
| 4 | func_800426D4 | 4 | 0 | 0 | - | Y |
| 4 | func_80043078 | 4 | 0 | 0 | - | Y |
| 4 | func_80043410 | 4 | 0 | 0 | - | Y |
@@ -100,6 +104,9 @@ jump-table funcs : 36 (deferred — need the rodata-island workflow, Task
| 4 | func_800491EC | 4 | 0 | 0 | - | Y |
| 4 | func_8004923C | 4 | 0 | 0 | - | Y |
| 4 | func_8005C4CC | 4 | 0 | 0 | - | Y |
| 4 | func_8005CF08 | 4 | 0 | 0 | - | Y |
| 4 | func_8005CF18 | 4 | 0 | 0 | - | Y |
| 4 | func_8005F384 | 4 | 0 | 0 | - | Y |
| 4 | func_80062988 | 4 | 0 | 0 | - | Y |
| 4 | gfx2D_COM0_OBJ_30 | 4 | 0 | 0 | - | Y |
| 5 | C_004_OBJ_DC | 5 | 0 | 0 | - | Y |
@@ -121,10 +128,3 @@ jump-table funcs : 36 (deferred — need the rodata-island workflow, Task
| 5 | PRNT_OBJ_1E8 | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_1F0 | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_1F8 | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_340 | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_64C | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_C8 | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_EC | 2 | 1 | 0 | - | Y |
| 5 | PRNT_OBJ_F4 | 2 | 1 | 0 | - | Y |
| 5 | RATAN_OBJ_148 | 5 | 0 | 0 | - | Y |
| 5 | SPU_OBJ_848 | 5 | 0 | 0 | - | Y |
docs/matching-cookbook.md
+46
View File
@@ -36297,3 +36297,49 @@ retired (`LIBGPU_ELF` = the raw dir; identify drops the 8 objects the EXE never
with and without the SDK objects. The class that remains in the sound region — `S_R`/`S_W`, `S_GRMDT*` —
has NO `.bss` of its own (cross-object commons at a minority address): a different wall, not this one.
#### §490 ★★★ — THE "WALL" BAND WAS A LIBRARY VERSION AWAY: LIBPAD 4.2.1 + LIBAPI 4.2 FOUND, THE WHOLE 0x8005CE18–0x8005FC68 BAND + THE APICARD REGION LINKED FROM REAL OBJECTS (P31 S79 #13/#5)
**Where §487 left it.** The psx loader's per-version signature sets named the band (libpad 4.2.1 + libapi
4.2) but no archive we held could LINK it: 4.0 has no libpad, 4.6/4.7 differ, and the loader's own 4.2
signatures showed PADMAIN drifting +4/+12. The plan was "C under the reorder island with real names" and
twelve stubs — the four §332 "%lo-in-a-delay-slot walls" among them — sat in `config/wave_exclude.txt` as
curated compiler facts.
**The hunt took one lead.** archive.org's `play-station-programmer-tool-runtime-library-version-4.2.7z`
(383 KB) is the PsyQ Runtime Library 4.2 (1998-01-21) plus `LIB/42PATCH/J421PD.ZIP`: SCE R&D's 1998-02-26
notice "Libpad.lib version 4.2.1 for the Analog Controller (DUAL SHOCK)" with LIBPAD.LIB 4.2.1, LIBAPI.LIB
4.2 and their headers. Converted with `psyq_lib_split.py` + `psyq-obj-parser`, `psyq_identify` places
libpad 4.2.1 7/11 and libapi 4.2 39/88 over 0x8005CE18–0x800629DC, and `psyq_link.py` PASSES all 46 —
PADMAIN at 760 ins exactly. Neighbours for the record: plain libpad 4.2 (the loader's signature source) and
the 4.3 disc (DTL-S2340, 1998-05-18: PADMAIN 832, PADIF 380, PADSEQD 292) each place only 4. 4.2.1 is the
unique exact match, which also dates the build to between February and May 1998.
**The wiring is §488's procedure, twice.** The band is ONE contiguous run of 33 interleaved objects
(21 libapi trampolines · COUNTER · PADENTRY · PADMAIN · L02/L03 · PADCMD · PADIF · PADPORTD · PADSEQD ·
WAITRC2) and every yaml-relevant boundary sits on an object edge (checked against `.text` SECTION sizes —
§9.6's align-pad gotcha). Because `psyq_integrate` tiles each stub with ONE library's objects, `800c3`
became four rows — `libapi1`, `libpad1`, `libapi2`, `libpad2` — fed by two windowed calls from the RAW
dirs (`.run/obj42/libapi42` 0x8005CE18..0x8005E188, `.run/obj42/libpad421` 0x8005D0D8..0x8005FC68). The
apicard region's three "game code" rows were libapi 4.2's C objects to the byte — `800c2` = FIRST.o
(`firstfile` + the "wall" stub `func_80062144`), `800c2_2` = PAD.o, `800c2_3` = PATCH.o+CHCLRPAD.o — so
they became `apicard5/6/7`, `make_apicard_used.py` now sources libapi from 4.2 (the EXE's real libapi;
libcard stays 4.0, byte-identical) and the region tiles 0x80061F38–0x80062888 with no game code left.
Four TUs deleted (`800c3.c`: 129 hand-matched "C" + 62 verbatim bodies + 19 stubs; the three `800c2*`),
and the `REORDER_TUS` island (§332b) is EMPTY — the variable stays for any future reorder-assembled TU.
main `143dbb89` with all SDK dirs and, from a fresh extract, with none.
**What the four "walls" were.** `_padInitSioMode`, `_padStartCom`, `func_8005ED4C`, `func_8005F450`
(§332 "%lo in a delay slot — no C can place it") were TRUE as compiler facts and irrelevant as work: Sony
assembled libpad in reorder mode and shipped the object. The same for `func_80062144` ("no jump table")
inside FIRST.o and for `PopMatrix`/`PushMatrix`, which had sat in the exclude list since S68 while living
in libgte3 — LINKED since Phase 8. The exclude audit kept all seven because its pinned-WALL class was
decided BEFORE its LINKED class; fixed S79 (LINKED dominates: a function that is not a target has no
wall). Law: **a wall verdict is a statement about a compiler and a function; whether the function is
OURS to match is a provenance question that precedes it** (§487 → §490: name the band, find the archive,
link it — in that order, before any C).
**Fresh clone.** `tools/psyq/PlayStation_Programmer_Tool_-_Runtime_Library_Version_4.2.7z` is tracked;
`7z x` it, `psyq_lib_split.py` + `psyq-obj-parser` the two lib421 LIBs into `.run/obj42/{libpad421,libapi42}`,
`tools/psyq_build_libs.sh LIBCARD`, `tools/make_apicard_used.py` — docs/SETUP.md carries the exact commands.
The build is byte-identical without any of it (the seven stub TUs), as always.
docs/progress.fleet.md
+9 -9
View File
@@ -4,21 +4,21 @@
# cross-binary collapsible-byte leverage: docs/duplicates.cross.md.
# THREE progress metrics (all matter — see the labels):
FLEET fn-count byte-ident: 363105 / 363156 = 99.99% (REAL+LINKED+empties; FUNCTION-count, ×134-inflated — one crack counts per overlay)
FLEET instr-weighted : 13481631 / 13489225 = 99.9% (shipped .text across main + resident + 211 overlays; the decomp.dev-DISPLAY number)
FLEET distinct-code(uniq): 5809877 / 5817332 = 99.9% (90909/90929 unique fns; the DISTINCT-RE number)
MAIN game-code weighted : 40895 / 44870 = 91.1% (INCLUDED in the fleet numbers above since 2026-07-22 — roadmap §1 metrics contract; LINKED-excluding Ghidra sig dated 2026-08-05; boundaries INDEPENDENTLY VERIFIED since P31 S77 — `make sig-main-oracle` + `make audit-corpus`: 0 phantom, 0 truncated, 1 explained pad-tail)
FLEET fn-count byte-ident: 363119 / 363157 = 99.99% (REAL+LINKED+empties; FUNCTION-count, ×134-inflated — one crack counts per overlay)
FLEET instr-weighted : 13479484 / 13485889 = 100.0% (shipped .text across main + resident + 211 overlays; the decomp.dev-DISPLAY number)
FLEET distinct-code(uniq): 5807730 / 5813996 = 99.9% (90909/90929 unique fns; the DISTINCT-RE number)
MAIN game-code weighted : 38748 / 41534 = 93.3% (INCLUDED in the fleet numbers above since 2026-07-22 — roadmap §1 metrics contract; LINKED-excluding Ghidra sig dated 2026-08-05; boundaries INDEPENDENTLY VERIFIED since P31 S77 — `make sig-main-oracle` + `make audit-corpus`: 0 phantom, 0 truncated, 1 explained pad-tail)
(fleet EXCLUDING main, for continuity with pre-2026-07-22 readings: 13440736 / 13444355 = 100.0%)
FLEET REAL substantive : 360708 (of which dedup-shared 255632 via 2220 groups / 255708 instances)
FLEET LINKED PsyQ objs : 1150
FLEET REAL substantive : 360642 (of which dedup-shared 255632 via 2220 groups / 255708 instances)
FLEET LINKED PsyQ objs : 1256
FLEET NON_MATCHING : 0 (0 in any default build — G4)
FLEET INCLUDE_ASM stubs : 51
FLEET matchable : 363156
FLEET INCLUDE_ASM stubs : 38
FLEET matchable : 363157
| binary | REAL | shared | LINKED | byte-ident | matchable | byte-ident % |
|---|---:|---:|---:|---:|---:|---:|
| main | 839 | 2 | 1150 | 2061 | 2090 | 98.6% |
| main | 773 | 2 | 1256 | 2075 | 2091 | 99.2% |
| resident | 141 | 0 | 0 | 143 | 145 | 98.6% |
| md_MAIN_001 | 11 | 0 | 0 | 11 | 11 | 100.0% |
| md_MAIN_003 | 53 | 0 | 0 | 59 | 64 | 92.2% |
docs/progress.md
+10 -10
View File
File diff suppressed because one or more lines are too long
docs/psyq-worklist.md
+26 -2
View File
@@ -17,7 +17,8 @@
| 2 | **libetc** | 5 / 7 | 800 tail (`0x8004239C`) | ~3.3k | VSYNC/INTR/INTR_VB/INTR_DMA/VMODE **contiguous**, ends exactly at libcd1 (`0x80043088`). Clean. |
| 3 | **libmcrd** | 2 / 2 | 800b2 (`0x8005FC68`,`0x80062888`) | ~9.0k | LIBMCRD.o (2186 ins, huge, holds the 55 `LIBMCRD_OBJ_*`) + USERFUNC.o — **2 blocks**. `_card_*` h_norm dups collapse inside LIBMCRD. |
| 4 | **libc2** | 17 / 46 | 800b2 (`0x8005C2C8`–`0x5CD98` + STRCAT `0x80061E90`) | ~3.1k | C stdlib (BZERO/MEMCPY/STRCMP/PRINTF/PRNT…); contiguous run + 1 outlier. **PRNT.o (418 ins) has an internal jtbl** — verify NOLOAD `.rodata` placement (the `PRNT_OBJ_24C` rodata note). |
| 5+6 | **libapi+libcard** | 22 (800c2) | 800c2 (`0x80061F38`–`0x80062888`) | ~0.7k | ✅ **DONE (combined apicard region): 22 objs / 4 blocks / +24 fns** (`tools/make_apicard_used.py`, C112 dedup, 0 exclusions). **libapi's ~22 objects in the 800c3 region (`0x5CE18`..) DEFERRED** — lowest value, separate resegmentation. |
| 5+6 | **libapi+libcard** | 26 (800c2) + 23 (band) | 800c2 (`0x80061F38`–`0x80062888`) + the band (`0x8005CE18`–`0x8005E188`) | ~2.6k | ✅ **DONE (S79 #5): apicard region 26 objs / 7 blocks tiling the whole range** (`tools/make_apicard_used.py`, libapi **4.2** from `tools/psyq/lib421` + libcard 4.0; FIRST/PAD/PATCH/CHCLRPAD were the "game code" rows 800c2/800c2_2/800c2_3) **+ the band's 23 libapi 4.2 objects (libapi1/libapi2)** from the raw `.run/obj42/libapi42` dir, windowed. |
| 9 | **libpad 4.2.1** | 7 / 11 | the band (`0x8005D0D8`–`0x8005FC68`) | ~11.5k | ✅ **DONE (S79 #13 found it, #5 wired it): PADENTRY PADMAIN PADCMD PADIF PADPORTD PADSEQD WAITRC2 as libpad1/libpad2** from the raw `.run/obj42/libpad421` dir (SCE's 1998-02-26 J421PD patch, `tools/psyq/lib421`). Absent: PADGUN/PADPORTM/PADSEQM/GUNHOOK (no gun, no multitap). |
| 7+8 | **libspu+libsnd** | 63 / (38+32) | 800 sound (`0x8003A444`–`0x8004239C`) | ~24k | ✅ **DONE (combined region; S78 #3/#4): 63 objs / 12 blocks.** The two libs interleave, so linked as ONE region (`tools/make_snd_used.py` dedups + excludes 3 addresses; `gen_lib_subsegs.py` + window). **Excluded:** S_R/S_W `0x3C438`, S_GRMDT* `0x3D424` (cross-object commons — these objects have NO `.bss`, so the S78 split cannot apply), S_IH/UT_RON `0x3D94C` (false-positive, inside SSSTART). VM_F `0x3FA64` (237 ins) rejoined in S78 #4 (`snd12`, `.bss` split at `_svm_sreg_buf`). SSGM.o `0x1BD80` deferred (isolated in matched-C, 8 ins). |
| 9 | **libgte** | 58 / 381 | 800b (`0x8004787C`–`0x5082C`) + **libgs gaps** | ~48k | ✅ **DONE (T11): 53 objs / 22 blocks linked** in 800b (subsegs via `gen_lib_subsegs.py`; integrate window 0x4787C..0x51804). **5 libgs-gap objects DEFERRED** (MTX_05/07/11/REG03/REG11 → gsgap1/2/4/5 stay stubs; gsgap2≠MTX_07 exactly so needs sub-split). |
@@ -69,7 +70,7 @@ references resolve to >1 base in the EXE. Curate the library's `_used` dir to dr
| `MTX_05/07/11`,`REG03`,`REG11` | libgte | sit in libgs gaps gsgap1/2/4/5; gsgap2(48B)≠MTX_07(36B) so the gap stub needs a sub-split | **deferred (T11)**; small GTE fns; link byte-identical, just need the gsgap region resegmented (low priority) |
| `0x3C438`,`0x3D424`,`0x3D94C` (S_R/S_GRMDT/S_IH) | libspu/snd | cross-object commons referenced at a minority address (S_R/S_GRMDT — these objects have NO `.bss` of their own, so the S78 split does not apply) + a false placement inside SSSTART (S_IH) | **excluded (sound region)**; stay stubs (24+4+24 ins). VM_F `0x3FA64` (237 ins) left this row in S78 #4 → `snd12` |
| `SSGM.o` | libsnd | isolated @0x1BD80, inside the matched-C region (near func_8001Bxxx) | **deferred**; 8 ins; would need a 1-object carve amid matched C |
| libapi 800c3 cluster (~22 objs) | libapi | C57..L10/L02/L03 @0x5CE18.. in the 800c3 region (separate from the 800c2 apicard region) | **deferred**; ~22 4-ins BIOS syscall stubs; lowest value; another region resegmentation |
| libapi 800c3 cluster (~22 objs) | libapi | C57..L10/L02/L03 @0x5CE18.. in the 800c3 region (separate from the 800c2 apicard region) | **LINKED S79 #5** (`libapi1`/`libapi2`, from libapi **4.2** — the EXE's real libapi, cookbook §490); the row's "lowest value" verdict missed that the same region held libpad 4.2.1's 2,780 ins |
*~~If scattered-`.bss` proves prevalent across libgte/libspu/libsnd, escalate to a Max general fix (split each object's `.bss` into per-common NOLOAD sections at their EXE-resolved addresses); otherwise excluding the few affected objects is the GS_001-precedent decision.~~ **Done in P31 S78 #4 — exactly that fix, twenty-three phases later: `tools/psyq_bss_split.py` runs inside the link-prepare step of `psyq_link` / `psyq_link_region` / `psyq_integrate` and tiles any such section into per-base NOBITS pieces from the bytes (cookbook §489). All three excluded objects link byte-identical; 235 placed objects across the 9 curated dirs, 0 refusals (R39 negative control).***
@@ -248,3 +249,26 @@ places the same 4 objects as plain 4.2 (PADENTRY, PADCMD, PADPORTD, WAITRC2) —
PADSEQD 292 against the EXE's 760 / 376 / 288 — and its LIBAPI.LIB places 38 (no 4.2 `C114`). So 4.2.1 is the
unique exact match: the game was built between the February 1998 patch and the May 1998 4.3 disc. Not banked
(>100 MB, R20 exception; the item id is the pointer; the two LIBs sit in `.run/psyq_hunt/rtl43/lib43/`).
### S79 task #5 — the band and the apicard region LINKED from libapi 4.2 + libpad 4.2.1 (0 tokens; cookbook §490)
| new block | was | objects | ins | note |
|---|---|---|---|---|
| `libapi1` | 800c3 head | C57 C68 C73 C114 A07–A13 A23–A25 A36 A37 A52 A53 A91 L10 + COUNTER | 176 | 21 BIOS trampolines + the RCnt object; window 0x8005CE18–0x8005E188 |
| `libpad1` | 800c3 | PADENTRY + PADMAIN | 1,060 | PADMAIN 760 = the 4.2.1 build |
| `libapi2` | 800c3 | L02 + L03 | 8 | SysEnqIntRP / SysDeqIntRP |
| `libpad2` | 800c3 tail | PADCMD PADIF PADPORTD PADSEQD WAITRC2 | 1,720 | window 0x8005D0D8–0x8005FC68 |
| `apicard5` | 800c2 | FIRST | 168 | `firstfile` (REAL C until now) + the "no jump table wall" stub `func_80062144` |
| `apicard6` | 800c2_2 | PAD | 192 | SetInitPadFlag … StopPAD + 4 statics (hand-matched C + 4 verbatims until now) |
| `apicard7` | 800c2_3 | PATCH + CHCLRPAD | 68 | `_patch_pad`, `_remove_ChgclrPAD` (4 verbatims until now) |
Two windowed integrate calls from the raw `.run/obj42/{libapi42,libpad421}` dirs (the band is one interleaved
run and `psyq_integrate` tiles each stub with one library) + the apicard call from the re-sourced
`.run/obj42/apicard_used` (libapi 4.2 + libcard 4.0, 26 objects / 7 blocks, no game code left in
0x80061F38–0x80062888). TUs gone: `src/800c3.c` (129 hand-matched "C", 62 verbatim bodies, 19 stubs — Sony's
reorder-assembled code all along, §332b), `src/800c2.c`, `src/800c2_2.c`, `src/800c2_3.c`; the `REORDER_TUS`
island is EMPTY (the Makefile mechanism stays). main `143dbb89` with every SDK dir and, from a fresh extract,
with none. `config/wave_exclude.txt` regenerated (`exclude_audit` now lets LINKED outrank a pinned WALL: the
four §332 walls, `func_80062144`, and `PopMatrix`/`PushMatrix` — in libgte3 since Phase 8 — dropped);
verbatim manifest 33 → 6 rows. Located-but-unwired SDK code left in main: `SSGM.o` (8 ins, inside matched C)
and the sound region's two cross-object-common walls (S_R/S_W, S_GRMDT*).
phase-ends/CURRENT_PHASE.md
+138 -101
View File
@@ -105,6 +105,22 @@ Instead of roadmap-v2 P31's per-function grind, Phase 31 organizes the 12,059 re
LINKABLE from real objects → #5 is re-scoped to a carve + wire (S79 FINAL block §6). The 4.3 disc is not
banked (>100 MB; the item id is the pointer; its two LIBs kept in `.run/psyq_hunt/rtl43/lib43/`).
Records: worklist "S79 task #13", SETUP archive table + §5.1 + S79 tool table.
**#5 DONE (S79, 2026-09-04, `commit:orphan-24`):** the whole band + the apicard region LINKED from libapi 4.2 +
libpad 4.2.1 (cookbook **§490**). `800c3` → 4 rows `libapi1` (21 trampolines + COUNTER, 176 ins) /
`libpad1` (PADENTRY+PADMAIN, 1,060) / `libapi2` (L02/L03, 8) / `libpad2` (PADCMD…WAITRC2, 1,720) fed by
two WINDOWED integrate calls from the raw `.run/obj42/{libapi42,libpad421}` dirs (integrate tiles each
stub with one library); `800c2`/`800c2_2`/`800c2_3` were FIRST.o / PAD.o / PATCH+CHCLRPAD → `apicard5/6/7`,
and `make_apicard_used.py` now sources libapi from 4.2 (`.run/obj42/apicard_used`, 26 objs / 7 blocks,
no game code left in 0x80061F38–0x80062888). Four TUs deleted; `REORDER_TUS` EMPTY (mechanism kept).
main `143dbb89` WITH all SDK dirs and, from a fresh extract, WITHOUT (`.run/S79_build_band_*.log`).
`exclude_audit` fixed (LINKED outranks a pinned WALL): `config/wave_exclude.txt` 13 → 3 (the four §332
walls, `func_80062144`, and `PopMatrix`/`PushMatrix` — libgte3 since Phase 8 — dropped); verbatim
manifest 33 → 6 (`--update`). Metrics: main REAL 839→**773** · LINKED 1,150→**1,256** · VERBATIM 29→**3** ·
stubs 29→**16** · byte-identical 2,075/2,091 = 99.2% · **game-code 93.3% (38,748 / 41,534)**, remainder
2,786 = the open-stub sum; fleet stubs 51→**38**. Docs: cookbook §490 (+index), worklist rows + "S79 task
#5", SETUP (fresh-clone obj42 commands, Makefile blocks, exclude_audit), decision-log "S79 addendum 2",
accelerators "S79 (2)". `frontier_classify` had a HARD-CODED LINKED set (R51) — derived from the Makefile
now. tools-health OK; R22 fleet 213/213 (`.run/S79_check_all_band.log`).
- [ ] **Tclose — PhaseEnd** (gate 2). (Max)
## Standing verification (every task)
@@ -8022,19 +8038,23 @@ for free. Three of our own measuring tools were lying (one hid a broken build, o
saved names, one under-counted the main program by 32 points); all fixed, and the main program is now
honestly 91.8% real C. Next: a small piece of linker surgery so two more Sony objects can be linked.
## 🛑 SESSION CHECKPOINT — S79 FINAL (2026-09-04). SUPERSEDES every earlier block in this file. Phase 31 T10 CONTINUES — the COMPLETION SPRINT, tasks #5 → #11 remain (#4 and #13 DONE this session).
## 🛑 SESSION CHECKPOINT — S79 FINAL (2026-09-04). SUPERSEDES every earlier block in this file. Phase 31 T10 CONTINUES — the COMPLETION SPRINT, tasks #6 → #11 remain (#4, #13 and #5 DONE this session).
Written for a FRESH SESSION with none of this context. Read it in full before doing anything.
**HEAD = the S79 task-#13 commit, on top of the task-#4 commit `commit:3865` (Drew pushes, R6); no
`Claude-Session:`/`Co-Authored-By` trailers (R5 + the S78 decision).** Model Fable 5.1; effort **xHigh** default
for #5–#10, **Max** for #11 (R27: prompt Drew and WAIT for the actual `/effort` line). No Ultracode / no waves in this sprint. The Ghidra MCP server
**HEAD = the S79 task-#5 commit `commit:orphan-24` (amended with this block), on top of #13 `commit:3866` and #4
`commit:3865` (Drew pushes, R6); no `Claude-Session:`/`Co-Authored-By` trailers (R5 + the S78 decision).** Model
Fable 5.1; effort **xHigh** default for #6–#10, **Max** for #11 (R27: prompt Drew and WAIT for the actual
`/effort` line). No Ultracode / no waves in this sprint. The Ghidra MCP server
restarts via the SessionStart hook → Drew runs `/mcp` (R29); no RE task is next, so no G2 ping needed
until one is. Ghidra DB churn (`db.*.gbf`, `~index`) is R23 restart-noise: do NOT stage it.
**Verified at close (R22/R58):** main `143dbb89f34491258bbc27810d0a12ec8b43a8dd` byte-identical WITH the SDK
objects (`.run/S79_build_main_sdk.log`) and WITHOUT them from a FRESH extract (`.run/S79_build_main_fallback.log`);
`make tools-health` (`.run/S79_tools_health.log`) and the fleet `make clean && make extract-all && make
check-all` (`.run/S79_check_all.log`): extract-all **212/212** (+ main), check-all **213 passed, 0 failed of 213** (rc=0, 2m42s).
**Verified at close (R22/R58):** main `143dbb89f34491258bbc27810d0a12ec8b43a8dd` byte-identical WITH every SDK
dir (`.run/obj40` + `.run/obj42`; `.run/S79_build_band_sdk.log`) and WITHOUT them from a FRESH extract
(`.run/S79_build_band_fallback.log`); `make tools-health` (`.run/S79_tools_health_band.log`, rc=0); the fleet
`make clean && make extract-all && make check-all` (`.run/S79_check_all_band.log`): extract-all **212/212**
(+ main), check-all **213 passed, 0 failed of 213** (rc=0, 2m32s); `frontier_classify` regenerated after it:
**39 rows = 38 stubs + the data word**, main 16. (The task-#4 state was verified the same way earlier in the
session: `.run/S79_check_all.log` 213/213.)
# 1. THE SESSION IN ONE PARAGRAPH
@@ -8049,44 +8069,55 @@ the SDK objects; the game-code metric moved 91.8% → 91.1% only because 3,667 S
numerator and its denominator (the open remainder is unchanged at 3,975 ins).
Then task #13, the bounded hunt, succeeded on its first lead: an archive.org 7z of the PsyQ Runtime Library 4.2
carries SCE's February-1998 "libpad.lib 4.2.1 for the DUAL SHOCK" patch — the EXE's exact libpad and libapi.
All 46 objects it places in main link byte-identical, so the whole 0x8005CE18–0x8005FC68 band (the twelve
"wall"/plumbing stubs, the eight reorder verbatims and the band's hand-matched C) is Sony code that task #5
now links instead of matching (§5/§6 below).
All 46 objects it places in main link byte-identical, so task #5 linked the whole 0x8005CE18–0x8005FC68 band
(the twelve "wall"/plumbing stubs, the reorder verbatims, the band's hand-matched C) and the apicard region's
three C objects from the real objects: four TUs gone, the `REORDER_TUS` island empty, main's open stubs 29 → 16,
fleet 51 → 38, game-code 93.3% with the remainder still equal to the open-stub sum. Two instruments were
found stale on the way and fixed (the exclude audit's WALL-before-LINKED order; the frontier census's
hard-coded LINKED set).
# 2. THE CENSUS (S79 close; every number with its denominator, R41)
* **Stubs: 51 of 363,151 matchable = 7,710 ins — UNCHANGED by #4** (`.run/frontier_s78.json`,
`tools/frontier_classify.py`; the three objects were REAL/VERBATIM, not stubs). Classes and names exactly as
the S78 FINAL block §2 listed them: A-TWIN-REMAP 2 · B-CARVE jtbl 6 · C-PLUMBING 7 · D-NEAR 14 · F-FAR 10 ·
G-DRAFTED-UNKNOWN 12 · H-VIRGIN 1 (`D_800D3200`, a data word). Per binary: **main 29 (3,975 ins)** ·
resident 2 · md 7 · overlays 14 across 11 binaries.
main's 29 by name: PadInfoAct (jtbl carve) · `_padSetMainMode` (plumbing) · `func_80011380` 192 (boot -O0
floor §474), `_padInitSioMode` 133, `func_80015760` 106, `func_8005D734` 91, `PadInitDirect` 88,
`func_80015608` 86, `func_80039B20` 79, `func_80038698` 74, `func_80062144` 65, `func_8005F290` 61,
`_dirFailAuto` 55, `_padStartCom` 51, `func_8005ECC0` 35 (D-NEAR) · `func_80032A74` 422, `func_8005F830`
153, `func_80015B6C` 120, `func_8002AC98` 114, `func_80020DA4` 100, `func_800391D4` 75, `func_80039DEC` 74,
`func_8002FDE8` 73, `func_8001BC6C` 69 (F-FAR) · `func_80039308` 518, `func_8001EFE0` 468, `func_80023BF0`
281, `func_8005ED4C` 223, `func_8005F450` 159 (G).
* **Verbatim `__asm__` bodies: 33 in tree** (main 29 + 4 elsewhere); manifest `config/verbatim_manifest.json`
now **33 rows** (S79 `--update`, subtractive: 200 → 33; 32 PERMANENT-VERBATIM + 1 DECOMPILE-NOW). #10
ratifies the PERMANENT set (S78 estimated ≈ 9 genuinely permanent: crt0 `start`/`__main`/`__do_global_dtors`,
GAME-ASM `md_MAIN_003:func_800D3204/3234`, GAME-GTE `ov_SC05_005:func_80181828` + `ov_SC06_032:func_8017D810`,
2 SDK frags) and decompiles `main()` (509 ins) + the md_MAIN_003 -O0 cluster; the GAME-GTE UNCERTAIN ×3
(`ov_SC01_001:func_80181E04` 269, `ov_SC03_105:func_80185810` 489, `ov_SC07_002:func_8017DC80` 346) need a
verbatim-vs-C verdict first (#9).
* **Located-but-unwired SDK objects** (the build PRINTS them per library as `~~ N located object(s) …`): after
#4 there is NO `~~` line for libgpu / libgs / snd. Still open: the libpad/libapi band pieces (#5), `SSGM.o`
8 ins @0x8001BD80 inside matched C (`800`), and the sound region's two cross-object-common walls `S_R`/`S_W`
@0x8003C438 + `S_GRMDT*` @0x8003D424 (these objects have NO `.bss` — a DIFFERENT class from §489; 24+4 ins).
* **Stubs: 38 of 363,157 matchable (`.run/frontier_s79.json`, `tools/frontier_classify.py` — regenerated after
the fleet; #5 removed the 13 band stubs; #4 removed none, its objects were REAL/VERBATIM).** Classes:
A-TWIN-REMAP 2 (ov_SC04_018 `func_80181804` 77 / `func_80181CB8` 67, proven twin ov_SC04_019) ·
B-CARVE jtbl 5 (resident `func_800D06E8` 344 + `func_800D128C` 243, ov_SC02_017 `func_80186C64` 209,
md_MAIN_034 `func_800CB00C` 152, ov_SC05_010 `func_8017FFA8` 88) · C-PLUMBING close=0 6 (ov_SC05_005
`func_80181828` 87, ov_SC05_018 `func_80180BE0` 65, ov_SC06_010 `func_801809E4` 33, md_MAIN_003
`func_800D06BC` 33 + `func_800D0100` 29, md_MAIN_020 `func_800CB17C` 30) · D-NEAR ≤25 6 (main
`func_80011380` 192 boot -O0 floor §474, `func_80015760` 106, `func_80015608` 86, `func_80039B20` 79,
`func_80038698` 74; ov_SC03_105 `func_801834A4` 106) · F-FAR 9 (main `func_80032A74` 422, `func_80015B6C`
120, `func_8002AC98` 114, `func_80020DA4` 100, `func_800391D4` 75, `func_80039DEC` 74, `func_8002FDE8` 73,
`func_8001BC6C` 69; ov_SC02_027 `func_80180B3C` 297) · G-DRAFTED-UNKNOWN 10 (main `func_80039308` 518,
`func_8001EFE0` 468, `func_80023BF0` 281; ov_SC03_105 `func_80185810` 489; md_MAIN_003 `func_800CF3E8`
469 + `func_800D1D14` 65 + `func_800D0174` 36; ov_SC07_002 `func_8017DC80` 346; ov_SC01_001 `func_80181E04`
269; ov_SC06_022 `func_8017DF28` 119) · H-VIRGIN 1 (md_MAIN_003 `D_800D3200`, a data word → splat fix).
Per binary: **main 16 (2,786 ins)** · resident 2 · md 7 · overlays 13 across 10 binaries. The 13 that
left with #5: PadInfoAct, `_padSetMainMode`, `_padInitSioMode`, `func_8005D734`, `PadInitDirect`,
`func_80062144`, `func_8005F290`, `_dirFailAuto`, `_padStartCom`, `func_8005ECC0`, `func_8005F830`,
`func_8005ED4C`, `func_8005F450` — all libpad 4.2.1 / libapi 4.2 objects.
* **Verbatim `__asm__` bodies: 6 in tree** (main 3 + md_MAIN_003 2 + ov_SC03_107 1); manifest
`config/verbatim_manifest.json` = **6 rows** after two subtractive `--update`s this session (200 → 33 → 6):
crt0 `start`/`__main`/`__do_global_dtors` (CRT, PERMANENT), GAME-ASM `md_MAIN_003:func_800D3204/3234`
(PERMANENT), and `ov_SC03_107:func_8017D878` (GAME-C, DECOMPILE-NOW). #10 ratifies the PERMANENT five in the
manifest `_README`, decompiles the one GAME-C, `main()` (509 ins) and the md_MAIN_003 -O0 cluster; the
GAME-GTE UNCERTAIN ×3 (`ov_SC01_001:func_80181E04` 269, `ov_SC03_105:func_80185810` 489,
`ov_SC07_002:func_8017DC80` 346) are G-class STUBS now, not verbatims — verdict first (#9).
* **Located-but-unwired SDK objects** (the build PRINTS them per library as `~~ N located object(s) …`): NO
`~~` line remains for any library. Still unlinked by design: `SSGM.o` 8 ins @0x8001BD80 inside matched C
(`800`), and the sound region's two cross-object-common walls `S_R`/`S_W` @0x8003C438 + `S_GRMDT*`
@0x8003D424 (these objects have NO `.bss` — a DIFFERENT class from §489; 24+4 ins, byte-identical stubs).
* **Disc: 5 unclaimed code payloads of 220** (MAIN/7, MAIN/9, SC03/53/54/56) — the explicit exclusion the
100% claim cites (`docs/disc-completeness.md`).
* **Metrics at close** (`docs/progress.md` / `progress.fleet.md`, regenerated by tools-health): main REAL
**839** · LINKED **1,150** · VERBATIM **29** · stubs **29** · matchable 2,090 · byte-identical 2,061/2,090 =
98.6%; **MAIN game-code weighted 91.1% (40,895 / 44,870)**, remainder 3,975 = the open-stub sum exactly;
fleet instr 99.9% (13,481,631 / 13,489,225), distinct 99.9%, stubs 51. Linked libraries: libcd 18/2,
libgs 34/8, libetc 5/1, libgpu 3/2, libmcrd 2/2, libc2 17/2, libgte 70/30, snd 63/12, apicard 22/4.
**773** · LINKED **1,256** · VERBATIM **3** · stubs **16** · matchable 2,091 · byte-identical 2,075/2,091 =
**99.2%**; **MAIN game-code weighted 93.3% (38,748 / 41,534)**, remainder 2,786 = the open-stub sum exactly
(S78 close: 91.8% / 44,562 / 48,537 — both terms have since lost ~7,000 Sony instructions that were counted
as matched game code); fleet instr 99.9%, distinct 99.9%, stubs **38** of 363,157. Linked libraries:
libcd 18/2, libgs 34/8, libetc 5/1, libgpu 3/2, libmcrd 2/2, libc2 17/2, libgte 70/30, snd 63/12,
apicard 26/7 (libapi 4.2 + libcard 4.0), libapi42 23/2, libpad421 7/2 — 268 objects / 70 blocks.
# 3. THE TASK LIST (harness tasks; Drew-confirmed order 2026-09-04) — DONE: #1 #2 #12 #3 #4 #13
# 3. THE TASK LIST (harness tasks; Drew-confirmed order 2026-09-04) — DONE: #1 #2 #12 #3 #4 #13 #5
| # | task | status | effort |
|---|---|---|---|
@@ -8095,9 +8126,9 @@ now links instead of matching (§5/§6 below).
| 12 | name the band (46 names, Ghidra mirrored, provenance docs) | DONE `commit:3861` | — |
| 3 | wire the exact-tile library subsegs (13 subsegs → LINKED) | DONE `commit:3863` | — |
| 4 | SYS.o + VM_F (+ GS_001) `.bss` split → LINKED (`libgpu2`, `snd12`, `libgs8`) | DONE S79 `commit:3865` | — |
| 13 | bounded hunt for LIBPAD.LIB 4.2.1 / 4.3 — **FOUND** (RTL 4.2 + the J421PD 4.2.1 patch; 46/46 byte-identical) | DONE S79 (this commit) | — |
| **5** | **the band: LINK ALL OF IT from libpad 4.2.1 + libapi 4.2 (carve `800c3`/`800c2*` at object boundaries; re-source the apicard region's libapi from 4.2)** | **NEXT** | **xHigh** |
| 6 | mechanical: A-TWIN-REMAP 2 + B-CARVE 6 + `D_800D3200` | pending | xHigh |
| 13 | bounded hunt for LIBPAD.LIB 4.2.1 / 4.3 — **FOUND** (RTL 4.2 + the J421PD 4.2.1 patch; 46/46 byte-identical) | DONE S79 `commit:3866` | — |
| 5 | the band + the apicard region LINKED from libpad 4.2.1 + libapi 4.2 (`libapi1/2`, `libpad1/2`, `apicard5-7`; 13 stubs, 4 TUs, the reorder island gone) | DONE S79 `commit:orphan-24` | — |
| **6** | **mechanical: A-TWIN-REMAP 2 + B-CARVE 5 + `D_800D3200`** | **NEXT** | **xHigh** |
| 7 | C-PLUMBING 7 via `recover_route` / §376–§378 | pending | xHigh |
| 8 | D-NEAR non-band 6: permuter/ILS/§31 or §474 wall-proof each | pending | xHigh |
| 9 | F-FAR 10 + G-UNKNOWN 12: one journal-noted agent per function | pending | xHigh |
@@ -8144,65 +8175,62 @@ mv back; make extract BINARY=main`) after ANY change to `psyq_identify`/`psyq_in
yaml; `lint_symbol_refs` — read its WHOLE output; a splat symbol comment must not contain `name:`; verify a
build from its EXIT CODE (R53).
# 5. TASK #13 — DONE: THE ARCHIVE EXISTS (facts, not to re-derive)
# 5. WHAT #13 AND #5 CHANGED (facts, not to re-derive)
`tools/psyq/PlayStation_Programmer_Tool_-_Runtime_Library_Version_4.2.7z` (tracked; sha256 `e4f5a678…`) →
`tools/psyq/lib42/*.LIB` (RTL 4.2, 1998-01-21; 23 libs + `42PATCH/` + `MC42PTCH/`) and `tools/psyq/lib421/`
(J421PD payload: `LIBPAD.LIB` 4.2.1 sha `eb41454b…`, `LIBAPI.LIB` 4.2 sha `88294cd8…`, `LIBPAD.H`, `LIBAPI.H`,
`KERNEL.H`, `J421PD.txt`). ELF: `.run/obj42/libpad421/` (11 objects) and `.run/obj42/libapi42/` (88); regenerate
with `python3 tools/psyq_lib_split.py <LIB> <dir>` then `tools/psyq/psyq-obj-parser <obj> -o <o>` per object.
Placement over 0x8005CE18–0x800629DC (`psyq_identify … 0x8005CE18 0x800629DC`): libpad421 7/11, libapi42 39/88,
every one `psyq_link.py` PASS. Neighbours for the record: plain libpad 4.2 and libpad 4.3 (disc DTL-S2340,
1998-05-18, `.run/psyq_hunt/rtl43/lib43/`, `.run/obj43/`) each place only 4 — 4.2.1 is the unique exact match.
The 4.2.1 headers, not 4.7's, are the band's prototype oracle from now on.
**(a) The archive.** `tools/psyq/PlayStation_Programmer_Tool_-_Runtime_Library_Version_4.2.7z` (tracked; sha256
`e4f5a678…`) → `tools/psyq/lib42/*.LIB` (RTL 4.2, 1998-01-21) and `tools/psyq/lib421/` (J421PD: `LIBPAD.LIB`
4.2.1 sha `eb41454b…`, `LIBAPI.LIB` 4.2 sha `88294cd8…`, `LIBPAD.H`, `LIBAPI.H`, `KERNEL.H` = the band's
prototype oracle). ELF: `.run/obj42/libpad421/` (11 objects) and `.run/obj42/libapi42/` (88); regenerate with
`python3 tools/psyq_lib_split.py <LIB> <dir>` then `tools/psyq/psyq-obj-parser <obj> -o <o>` (SETUP has the
commands). Placement 0x8005CE18–0x800629DC: libpad421 7/11, libapi42 39/88, every one `psyq_link.py` PASS.
Neighbours: plain libpad 4.2 and the 4.3 disc (DTL-S2340, 1998-05-18; `.run/psyq_hunt/rtl43/lib43/`) place 4
each — 4.2.1 is the unique exact match; the game was built between Feb and May 1998.
**(b) The wiring** (cookbook §490). Yaml: `800c3` → `libapi1` [0x4D618] · `libpad1` [0x4D8D8] · `libapi2`
[0x4E968] · `libpad2` [0x4E988]; `800c2` → `apicard5`, `800c2_2` → `apicard6`, `800c2_3` → `apicard7`.
Makefile: `LIBAPI42_*` (raw `.run/obj42/libapi42`, window 0x8005CE18 0x8005E188, stubs libapi1,libapi2),
`LIBPAD_*` (raw `.run/obj42/libpad421`, window 0x8005D0D8 0x8005FC68, stubs libpad1,libpad2),
`APICARD_ELF := .run/obj42/apicard_used` with `APICARD_STUBS` apicard1..7; the SYMS line carries the two new
`*_externals.ld`; `REORDER_TUS :=` (empty, mechanism kept). `tools/make_apicard_used.py`: `SOURCES =
(.run/obj42/libapi42, .run/obj40/libcard)` → 26 objects. TUs: `git rm src/800c3.c src/800c2.c src/800c2_2.c
src/800c2_3.c`; splat emitted the 7 stub TUs (`libapi1.c` 25 stubs, `libpad1.c` 24, `libapi2.c` 2, `libpad2.c`
40, `apicard5.c` 2, `apicard6.c` 10, `apicard7.c` 4) for the no-SDK fallback. `psyq_integrate` printed 26/7 ·
23/2 · 7/2 blocks, no `~~`, no `!!`.
**(c) Instruments fixed.** `tools/exclude_audit.py`: LINKED now outranks a pinned `# WALL` (`config/wave_exclude.txt`
13 → 3: the four §332 walls, `func_80062144`, `PopMatrix`/`PushMatrix` — the last two in libgte3 since Phase 8).
`tools/frontier_classify.py`: its LINKED set was a hard-coded 49-name literal (R51) and reported 337 "stubs" after
#5; derived from the Makefile now (`progress._main_linked_segs_from_makefile`).
**(d) Still true from S78/S79 #4:** §4 above (the `.bss` split, the band identity, ApplySymbols for renames,
main's LINKED path only exercised in-tree with the obj dirs present, R53 exit codes, `lint_symbol_refs` whole
output, no `name:` in splat symbol comments).
# 6. TASK #5 — WIRE THE WHOLE BAND (start here; xHigh; design before code, X1)
# 6. TASK #6 — MECHANICAL: A-TWIN-REMAP 2 + B-CARVE 5 + `D_800D3200` (start here; xHigh; design before code, X1)
**Target:** every byte of 0x8005CE18–0x8005FC68 becomes LINKED, plus `FIRST.o`/`PAD.o`/`PATCH.o`/`CHCLRPAD.o`
and the apicard trampolines re-sourced from libapi 4.2. Object tiling of the band, in vram order (all
verified byte-identical S79): 21 libapi trampolines C57 C68 C73 C114 A07 A08 A09 A10 A11 A12 A13 A23 A24 A25
A36 A37 A52 A53 A91 L10 at 0x8005CE18 + 0x10·k (4 ins each, 0x8005CE18–0x8005CF68) · COUNTER 0x8005CF68 (92)
· PADENTRY 0x8005D0D8 (300) · PADMAIN 0x8005D588 (760, ends 0x8005E168) · L02 0x8005E168 · L03 0x8005E178 ·
PADCMD 0x8005E188 (600) · PADIF 0x8005EAE8 (376) · PADPORTD 0x8005F0C8 (408) · PADSEQD 0x8005F728 (288) ·
WAITRC2 0x8005FBA8 (48, ends 0x8005FC68). Check the tiling has NO gap (0x8005CE18 + 21·0x10 = 0x8005CF68 ✓;
PADMAIN 760·4 = 0xBE0 → 0x8005D588+0xBE0 = 0x8005E168 ✓; PADENTRY ends 0x8005D588 ✓ …) before carving —
integrate's exact-tiling check will refuse otherwise (R43).
**The carve.** Today the band lives in the C TUs of the `REORDER_TUS` island (`800c3` = 0x8005CE18…, plus
`800c2`, `800c2_2`, `800c2_3` for the apicard-side pieces; §332b). Follow the §486/§488 procedure: yaml rows
at each object-block boundary (block names e.g. `libapi42_1` = the 21 trampolines + COUNTER 0x8005CE18–0x8005D0D8,
`libpad1` = PADENTRY+PADMAIN 0x8005D0D8–0x8005E168, `libapi42_2` = L02/L03 0x8005E168–0x8005E188, `libpad2` =
PADCMD…WAITRC2 0x8005E188–0x8005FC68 — or one block per object; pick what tiles the yaml cleanly), each
row commented with objects + ins + "exact tile", then `git rm` the TUs that become 100% SDK (`800c3` is
entirely band? CHECK the row range vs 0x8005FC68 — anything after the band inside `800c3` stays game C in a
carved-off TU: split the TU at the object end, keep its game C, per §488). Curated dir: build
`.run/obj42/band_used` = the 7 libpad421 objects + the band's 23 libapi42 objects (the apicard-region 16 go to
the apicard call), or pass the raw dirs with WINDOWS (`lo hi`) so each call places only its region — two
integrate calls: libpad421 window 0x8005D0D8–0x8005FC68, libapi42 window 0x8005CE18–0x8005E188 (+ a third
libapi42 call for the apicard window 0x80061F38–0x80062888 REPLACING the 4.0 `apicard_used` libapi objects;
`make_apicard_used.py` merges libapi+libcard — point its libapi source at obj42 and keep libcard 4.0, then
`FIRST.o` needs a carved stub for the REAL-C `firstfile` in `800c2` [0x80061FA8,0x80062248) and `PATCH`+
`CHCLRPAD` = all of `800c2_3` after A18–A21). Makefile: new `*_ELF/_OBJDIR/_SYMS` triples + stub lists;
`$(main_SPLAT_YAML)` on every call; `progress.py` picks the blocks up from the Makefile automatically.
**Names.** `config/symbols.us.txt` already carries the 46 band names (S78 #12); the objects export them
(`--redefine-sym` covers any curated-vs-object spelling, R15). `PadInfoAct`'s jtbl (B-CARVE) disappears with
the object — remove it from `config/wave_exclude.txt` via `exclude_audit.py --write` (a stale list is refused).
Ghidra: no renames needed (names already mirrored S78); if any curated name changes, `tools/ghidra_apply_symbols.sh` + R9.
**Gate exactly as #3/#4:** in-tree `make extract BINARY=main` → `make build BINARY=main -j` with `.run/obj40`+`obj42`
present (watch the `integrate` lines: block counts, `~~` residue must be EMPTY for the band, `!!` unresolved
none) → the fresh-extract fallback WITHOUT the SDK dirs → `make tools-health` → R22 fleet (background) → docs
(cookbook § for "the band links", worklist rows incl. the "5+6 libapi+libcard" footprint row and the §332
wall verdicts now VOID, SETUP, decision-log R31: the §332 "walls" were reorder-assembled SDK code — recorded
S78, close the loop) → `verbatim_check --update` (the 8 SDK-C-REORDER verbatims go) → commit. Expect: main's
open stubs 29 → 17 (the 12 band stubs), VERBATIM −8, REAL −(the band's hand-matched C incl. `firstfile`),
LINKED +~60, game-code denominator −~3,200 ins with the remainder = the new open-stub sum (reconcile it, R41).
Fresh-clone note (SETUP): `.run/obj42` is regenerable from the tracked 7z — document the two commands.
**A-TWIN-REMAP (2, ov_SC04_018):** `func_80181804` (77) and `func_80181CB8` (67) have a PROVEN twin in
ov_SC04_019 (banked). Run `tools/twin_rescan.py` first (a bank changes the twin graph; §397) to confirm the
pairing, then `family_remap` / the structural-family mechanical remap (memory `structural-family-mechanical-remap`:
h_exact remaps bank at 88.5%). Gate ov_SC04_018 with the parallel gate in a worktree (memory
`parallel-gate-via-worktrees`; never xargs -P over gate_stage). Then `twin_rescan` again.
**B-CARVE jtbl (5):** resident `func_800D06E8` (344) + `func_800D128C` (243) need the RESIDENT carve path
(`jtbl_carve` / `jtbl_family_bank`; `o0_subsplit` refuses main only — check what it says for resident);
ov_SC02_017 `func_80186C64` (209), ov_SC05_010 `func_8017FFA8` (88) via `jtbl_carve`; md_MAIN_034
`func_800CB00C` (152) sits in `config/wave_exclude.txt` as "WALL: compiler wall (S68)" — it is B-CARVE in the
census, so re-probe it under the CURRENT carve tooling before believing the S68 label (memory
`reprobe-exclude-lists-after-tool-fixes`; `exclude_audit` keeps it only because it is pinned). `split_indicator.py`
is the carve oracle (tools-health runs it as a hard gate): a subseg owning raw tables in >1 non-adjacent span
must be split before any switch function in it can bank. Read `docs/matching-cookbook.md` §486 (the 5-piece
manual carve) and §53 before carving.
**H-VIRGIN `D_800D3200` (md_MAIN_003):** a data word mis-sliced as a stub — a splat symbol type fix in
`config/splat.md_MAIN_003.yaml` / its symbols file (declare the word as data), then re-extract md_MAIN_003; it
should vanish from `corpus.stubs` without any C.
**Gate/verify per binary:** overlays/modules through `gate_stage`/`parallel_gate` (each carve edits that
binary's yaml — R60: a gate commits only its own binary's carve-state lines; `interleave_check` + `pads_audit`
after any yaml touch); resident through its own path; then `twin_rescan`, R22 fleet, `frontier_classify --json`,
docs (cookbook idioms if any new), checkpoint, one commit per binary or per class (R42: commit before the next
command that can touch src/).
# 7. TASKS #6–#11 — CARRIED CONTEXT
* #6: `tools/twin_rescan.py` / `family_remap` for the ov_SC04_018 pair (`func_80181804`/`func_80181CB8`,
twin ov_SC04_019); jtbl carves via `jtbl_carve`/`jtbl_family_bank` (resident `func_800D06E8` 344 +
`func_800D128C` 243 need the resident carve path; `o0_subsplit` refuses main → the §486 manual 5-piece
procedure for `PadInfoAct`); `D_800D3200` = splat symbol type fix.
* #6: see §6 above (`PadInfoAct` is gone with the band — no main carve left in this class).
* #7: bodies proven (close=0); the TU's spelling refuses — `recover_route.py` (routes a DROP to the applicable
tool), `cast_self_callers`, `sync_tu_decls`, `--sync-decls`; prove plumbing byte-neutral BEFORE gating and
commit the plumbing before the gate (S77 laws). `gate_main.py` is the ONLY main gate.
@@ -8212,10 +8240,11 @@ Fresh-clone note (SETUP): `.run/obj42` is regenerable from the tracked 7z — do
* #9: single agents with `journal_notes.py` + `neighbor_ref` (via `claude_wave_packs`); the 3 GAME-GTE
UNCERTAIN bodies get a verbatim-vs-C verdict first. Route models per the ladder (Haiku ≤50 ins → Sonnet →
Opus; Fable only for a NEW wall class).
* #10: ratify PERMANENT in the manifest `_README` (33 rows now; expected ≈ 9 truly permanent); decompile
`main()` (509 ins, GAME-C) + the md_MAIN_003 -O0 cluster.
* #11: PhaseEnd per the format; the 5 unclaimed payloads as the explicit exclusion; wall ledger with proofs;
the dashboard on the corrected denominators (main 91.1%-class numbers, LINKED 1,150).
* #10: ratify the 5 PERMANENT rows in the manifest `_README` (6 rows now); decompile `ov_SC03_107:func_8017D878`
(GAME-C, the one DECOMPILE-NOW), `main()` (509 ins) + the md_MAIN_003 -O0 cluster.
* #11: PhaseEnd per the format; the 5 unclaimed payloads as the explicit exclusion; wall ledger with proofs
(3 pinned walls left: md_MAIN_034 `func_800CB00C` (re-probe in #6), ov_SC03_105 `func_801834A4`, ov_SC06_022
`func_8017DF28`); the dashboard on the corrected denominators (main 93.3% / 99.2% byte-identical, LINKED 1,256).
# 8. HABITS THIS SESSION PAID FOR (keep)
* Design before code, and probe the design's PREDICTION before trusting it: the raw objects failed by exactly
@@ -8225,6 +8254,11 @@ Fresh-clone note (SETUP): `.run/obj42` is regenerable from the tracked 7z — do
* Put a transformation in the shared prepare path, then negative-control it over the whole population it now
touches (235 objects) — the R39 control caught the end-pointer refusal that the three targets never showed.
* Confirm every derived cut against an oracle that did not derive it (the other objects' by-name recoveries).
* Provenance → archive → link → compiler, in that order (§490): twelve "walls" and a 383 KB download.
* Every "linked" list must be DERIVED from the Makefile (R51): two hard-coded copies (exclude pins, the census)
each lied the day the link state changed.
* Edit the LIVE checkpoint block through a slice of the file that starts at its own header: the S78 and S79
blocks share every section heading, and `str.index` on a heading duplicated a region twice this session.
# 9. PLAIN ENGLISH
Three chunks of the main program had been treated for months as "Sony library code we can't link because
@@ -8237,5 +8271,8 @@ changed; the accounting is just more honest: 47 fewer "hand-matched" functions t
110 more linked library functions, and the remaining work is exactly what it was — 51 functions with named
reasons. Then the bounded search for the exact Sony controller-library version the game shipped with found it
in under an hour: a small archive on archive.org holds Sony's February-1998 controller-library patch, and every
one of its pieces matches the game byte for byte. The hardest-looking stretch of the main program can now be
linked straight from Sony's library instead of rewritten by hand. Next: wire it in.
one of its pieces matches the game byte for byte. We then wired it in: the hardest-looking stretch of the main
program — twelve functions carried as "compiler walls", plus the controller and BIOS-call code around it —
now links straight from Sony's library instead of being rewritten by hand. Main's open functions went from 29
to 16 and the whole game's from 51 to 38; nothing about the game changed. Next: the mechanical leftovers (two
twin functions, five jump-table carves, one mis-typed data word).
src/800c2.c
-61
View File
@@ -1,61 +0,0 @@
#include "common.h"
#include "common.h"
struct Rec {
char *name; /* 0x00 */
u8 pad04[0x30]; /* 0x04 */
s32 (*callback)(); /* 0x34 */
u8 pad38[0x18]; /* 0x38 */
};
extern u8 D_80078CE0[];
extern s32 D_80078CD8;
extern s32 func_8005C4DC();
extern s32 firstfile2();
extern void func_80062144();
s32 firstfile(char *name, s32 mode)
{
u8 *src;
u8 *dst;
struct Rec *base;
struct Rec *rec;
u32 n;
s32 found;
src = (u8 *)name;
dst = D_80078CE0;
while ((s8)*src > ':') {
*dst++ = *src++;
}
*dst = 0;
n = *(u32 *)0x154 / 80;
base = *(struct Rec **)0x150;
for (rec = base; rec < base + n; rec++) {
if (rec->name != NULL && func_8005C4DC(rec->name, D_80078CE0) == 0) {
D_80078CD8 = (s32)rec->callback;
found = 1;
goto done;
}
}
found = 0;
done:
if (found == 0) {
return 0;
}
n = *(u32 *)0x154 / 80;
base = *(struct Rec **)0x150;
for (rec = base; rec < base + n; rec++) {
if (rec->name != NULL && func_8005C4DC(rec->name, D_80078CE0) == 0) {
rec->callback = (s32 (*)())func_80062144;
break;
}
}
return firstfile2(name, mode);
}
INCLUDE_ASM("asm/nonmatchings/800c2", func_80062144);
src/800c2_2.c
-171
View File
@@ -1,171 +0,0 @@
#include "common.h"
__asm__(
".text\n"
".align 2\n"
".globl SetInitPadFlag\n"
".ent\tSetInitPadFlag\n"
"SetInitPadFlag:\n"
".frame $sp,0,$31\n"
".mask 0x00000000,0\n"
".fmask 0x00000000,0\n"
".set\tnoreorder\n"
"lui $at, %hi(D_80072A24)\n"
"jr $ra\n"
"sw $a0, %lo(D_80072A24)($at)\n"
".set\treorder\n"
".end\tSetInitPadFlag\n"
);
extern s32 D_80072A24;
s32 ReadInitPadFlag(void)
{
return D_80072A24;
}
extern void _remove_ChgclrPAD(void);
extern void func_8005CF08(void);
extern void _patch_pad(void);
extern void func_8005CF18(void);
extern s32 ChangeClearPAD(s32 mode);
extern s32 func_8006252C(void);
extern s32 PAD_init2(s32, s32, s32, s32);
extern s32 D_80072A24;
void PAD_init(s32 a0, s32 a1, s32 a2, s32 a3)
{
_remove_ChgclrPAD();
func_8005CF08();
_patch_pad();
func_8005CF18();
ChangeClearPAD(0);
((s32 (*)())func_8006252C)();
PAD_init2(a0, a1, a2, a3);
D_80072A24 = 1;
}
extern void _remove_ChgclrPAD(void);
extern void func_8005CF08(void);
extern void _patch_pad(void);
extern void func_8005CF18(void);
extern s32 ChangeClearPAD(s32 mode);
extern s32 func_8006252C(void);
extern void InitPAD2(s32 a0, s32 a1, s32 a2, s32 a3);
extern s32 D_80072A24;
void InitPAD(s32 a0, s32 a1, s32 a2, s32 a3)
{
_remove_ChgclrPAD();
func_8005CF08();
_patch_pad();
func_8005CF18();
ChangeClearPAD(0);
((s32 (*)())func_8006252C)();
InitPAD2(a0, a1, a2, a3);
D_80072A24 = 1;
}
extern void StartPAD2(void);
extern s32 ChangeClearPAD(s32 mode);
extern void EnablePAD(void);
s32 StartPAD(void)
{
StartPAD2();
ChangeClearPAD(0);
EnablePAD();
return 1;
}
extern void DisablePAD(void);
extern void StopPAD2(void);
extern s32 func_800625A4(void);
extern s32 D_80072A24;
void StopPAD(void)
{
DisablePAD();
StopPAD2();
func_800625A4();
D_80072A24 = 0;
}
extern void func_8005CF08(void);
extern void func_8005CF18(void);
extern s32 SysDeqIntRP(s32 priority, void *intr);
extern s32 SysEnqIntRP(s32 priority, void *intr);
extern s32 func_800625DC(void);
extern s32 func_80062644(void);
extern s32 D_80078D0C;
extern u8 D_80078D08[];
extern s32 D_80078D14;
s32 func_8006252C(void) {
s32 *p;
s32 *q;
s32 pri;
func_8005CF08();
pri = 1;
__asm__ __volatile__("" : "=r"(pri) : "0"(pri));
p = &D_80078D0C;
__asm__ __volatile__("" : "=r"(p) : "0"(p));
q = p - 1;
p[0] = (s32)func_800625DC;
p[1] = (s32)func_80062644;
*(s32 *)D_80078D08 = 0;
D_80078D14 = 0;
SysDeqIntRP(pri, q);
SysEnqIntRP(1, q);
func_8005CF18();
return 1;
}
extern void func_8005CF08(void);
extern void func_8005CF18(void);
extern s32 SysDeqIntRP(s32 priority, void *intr);
extern unsigned char D_80078D08[];
s32 func_800625A4(void) {
func_8005CF08();
SysDeqIntRP(1, &D_80078D08);
func_8005CF18();
return 1;
}
extern u8 *D_80072A28;
s32 func_800625DC(void) {
volatile s32 i;
s32 frame_pad[2];
(void)&frame_pad;
*(u16 *)(D_80072A28 + 10) = 0;
i = 10;
if (--i != -1) {
do {
} while (--i != -1);
}
return 0;
}
s32 func_80062644(void) {
extern s32 *D_80072A2C;
s32 *p = D_80072A2C;
if ((p[1] & 1) == 0) {
return 0;
}
if ((p[0] & 1) != 0) {
return 1;
}
if ((p[0] & 1) != 0) {
return 1;
}
return 0;
}
__asm__(".word 0x00000000\n");
src/800c2_3.c
-137
View File
@@ -1,137 +0,0 @@
#include "common.h"
__asm__(
".text\n"
".align 2\n"
".globl EnablePAD\n"
".ent\tEnablePAD\n"
"EnablePAD:\n"
".frame $sp, 0, $31\n"
".mask 0x00000000, 0\n"
".fmask 0x00000000, 0\n"
".set\tnoreorder\n"
"lui $t1, %hi(jtbl_80078D20_main)\n"
"lw $t1, %lo(jtbl_80078D20_main)($t1)\n"
"nop\n"
"jr $t1\n"
"nop\n"
".set\treorder\n"
".end\tEnablePAD\n"
);
__asm__(
".text\n"
".align 2\n"
".globl DisablePAD\n"
".ent\tDisablePAD\n"
"DisablePAD:\n"
".frame $sp, 0, $31\n"
".mask 0x00000000, 0\n"
".fmask 0x00000000, 0\n"
".set\tnoreorder\n"
"lui $t1, %hi(jtbl_80078D24_main)\n"
"lw $t1, %lo(jtbl_80078D24_main)($t1)\n"
"nop\n"
"jr $t1\n"
"nop\n"
".set\treorder\n"
".end\tDisablePAD\n"
);
__asm__(
".text\n"
".align 2\n"
".globl _patch_pad\n"
".ent\t_patch_pad\n"
"_patch_pad:\n"
".frame $sp, 0, $31\n"
".mask 0x00000000, 0\n"
".fmask 0x00000000, 0\n"
".set\tnoreorder\n"
"lui $at, %hi(D_80078D18)\n"
"sw $ra, %lo(D_80078D18)($at)\n"
"jal func_8005CF08\n"
"nop\n"
"addiu $t1, $0, 87\n"
"addiu $t2, $0, 176\n"
"jalr $t2\n"
"nop\n"
"lw $v0, 364($v0)\n"
"addiu $t1, $0, 11\n"
"addi $v1, $v0, 2180\n"
"lui $at, %hi(jtbl_80078D20_main)\n"
"sw $v1, %lo(jtbl_80078D20_main)($at)\n"
"addi $v1, $v0, 2196\n"
"lui $at, %hi(jtbl_80078D24_main)\n"
"sw $v1, %lo(jtbl_80078D24_main)($at)\n"
".L80062730:\n"
"sw $0, 1428($v0)\n"
"addiu $v0, $v0, 4\n"
"addiu $t1, $t1, -1\n"
"bnez $t1, .L80062730\n"
"nop\n"
"jal FlushCache\n"
"nop\n"
"lui $ra, %hi(D_80078D18)\n"
"lw $ra, %lo(D_80078D18)($ra)\n"
"nop\n"
"jr $ra\n"
"nop\n"
".set\treorder\n"
".end\t_patch_pad\n"
"nop\n"
"nop\n"
);
/* Handwritten function (splat header) — banked verbatim per §265 form 1.
* Recovered semantics:
* D_80078D28 = return address; func_8005CF08(); call vector 0xB0 ($t2=176);
* p = (*(ret + 0x16C)) + 0x62C; for (i = 9; i; --i) *p++ = 0;
* FlushCache(); func_8005CF18(); return via saved $ra.
* NOTE: maspsx did NOT auto-insert load-delay nops here (after
* "lw $v0,364($v0)" and "lw $ra,%lo(D_80078D28)($ra)") — both written by
* hand, matching the banked sibling _patch_pad in this same TU; this
* contradicts §179-B rule 4 as observed on this build.
*/
__asm__(
".text\n"
".align 2\n"
".globl _remove_ChgclrPAD\n"
".ent\t_remove_ChgclrPAD\n"
"_remove_ChgclrPAD:\n"
".frame $sp, 0, $31\n"
".mask 0x00000000, 0\n"
".fmask 0x00000000, 0\n"
".set\tnoreorder\n"
"lui $at, %hi(D_80078D28)\n"
"sw $ra, %lo(D_80078D28)($at)\n"
"jal func_8005CF08\n"
"nop\n"
"addiu $t1, $0, 87\n"
"addiu $t2, $0, 176\n"
"jalr $t2\n"
"nop\n"
"addiu $t2, $0, 9\n"
"lw $v0, 364($v0)\n"
"nop\n"
"addi $v1, $v0, 1580\n"
".L80062798:\n"
"sw $0, 0($v1)\n"
"addiu $v1, $v1, 4\n"
"addiu $t2, $t2, -1\n"
"bnez $t2, .L80062798\n"
"nop\n"
"jal FlushCache\n"
"nop\n"
"jal func_8005CF18\n"
"nop\n"
"lui $ra, %hi(D_80078D28)\n"
"lw $ra, %lo(D_80078D28)($ra)\n"
"nop\n"
"jr $ra\n"
"nop\n"
".set\treorder\n"
".end\t_remove_ChgclrPAD\n"
"nop\n"
"nop\n"
);
src/800c3.c
-1771
View File
File diff suppressed because it is too large Load Diff
src/apicard5.c
+5
View File
@@ -0,0 +1,5 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/apicard5", firstfile);
INCLUDE_ASM("asm/nonmatchings/apicard5", func_80062144);
src/apicard6.c
+21
View File
@@ -0,0 +1,21 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/apicard6", SetInitPadFlag);
INCLUDE_ASM("asm/nonmatchings/apicard6", ReadInitPadFlag);
INCLUDE_ASM("asm/nonmatchings/apicard6", PAD_init);
INCLUDE_ASM("asm/nonmatchings/apicard6", InitPAD);
INCLUDE_ASM("asm/nonmatchings/apicard6", StartPAD);
INCLUDE_ASM("asm/nonmatchings/apicard6", StopPAD);
INCLUDE_ASM("asm/nonmatchings/apicard6", func_8006252C);
INCLUDE_ASM("asm/nonmatchings/apicard6", func_800625A4);
INCLUDE_ASM("asm/nonmatchings/apicard6", func_800625DC);
INCLUDE_ASM("asm/nonmatchings/apicard6", func_80062644);
src/apicard7.c
+9
View File
@@ -0,0 +1,9 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/apicard7", EnablePAD);
INCLUDE_ASM("asm/nonmatchings/apicard7", DisablePAD);
INCLUDE_ASM("asm/nonmatchings/apicard7", _patch_pad);
INCLUDE_ASM("asm/nonmatchings/apicard7", _remove_ChgclrPAD);
src/libapi1.c
+51
View File
@@ -0,0 +1,51 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/libapi1", InitHeap);
INCLUDE_ASM("asm/nonmatchings/libapi1", FlushCache);
INCLUDE_ASM("asm/nonmatchings/libapi1", func_8005CE38);
INCLUDE_ASM("asm/nonmatchings/libapi1", _96_remove);
INCLUDE_ASM("asm/nonmatchings/libapi1", DeliverEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", OpenEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", CloseEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", WaitEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", TestEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", EnableEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", DisableEvent);
INCLUDE_ASM("asm/nonmatchings/libapi1", ReturnFromException);
INCLUDE_ASM("asm/nonmatchings/libapi1", ResetEntryInt);
INCLUDE_ASM("asm/nonmatchings/libapi1", HookEntryInt);
INCLUDE_ASM("asm/nonmatchings/libapi1", func_8005CF08);
INCLUDE_ASM("asm/nonmatchings/libapi1", func_8005CF18);
INCLUDE_ASM("asm/nonmatchings/libapi1", read);
INCLUDE_ASM("asm/nonmatchings/libapi1", write);
INCLUDE_ASM("asm/nonmatchings/libapi1", ChangeClearPAD);
INCLUDE_ASM("asm/nonmatchings/libapi1", ChangeClearRCnt);
INCLUDE_ASM("asm/nonmatchings/libapi1", SetRCnt);
INCLUDE_ASM("asm/nonmatchings/libapi1", GetRCnt);
INCLUDE_ASM("asm/nonmatchings/libapi1", StartRCnt);
INCLUDE_ASM("asm/nonmatchings/libapi1", StopRCnt);
INCLUDE_ASM("asm/nonmatchings/libapi1", ResetRCnt);
src/libapi2.c
+5
View File
@@ -0,0 +1,5 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/libapi2", SysEnqIntRP);
INCLUDE_ASM("asm/nonmatchings/libapi2", SysDeqIntRP);
src/libpad1.c
+49
View File
@@ -0,0 +1,49 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/libpad1", PadChkVsync);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadStartCom);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadStopCom);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadChkMtap);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadGetState);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadInfoMode);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadInfoAct);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadInfoComb);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadSetActAlign);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadSetMainMode);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadSetAct);
INCLUDE_ASM("asm/nonmatchings/libpad1", PadEnableCom);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padSetVsyncParam);
INCLUDE_ASM("asm/nonmatchings/libpad1", func_8005D6CC);
INCLUDE_ASM("asm/nonmatchings/libpad1", func_8005D734);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padChkVsync);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padStartCom);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padStopCom);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padInitSioMode);
INCLUDE_ASM("asm/nonmatchings/libpad1", func_8005DBD8);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padSioRW);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padSioRW2);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padClrIntSio0);
INCLUDE_ASM("asm/nonmatchings/libpad1", _padWaitRXready);
src/libpad2.c
+81
View File
@@ -0,0 +1,81 @@
#include "common.h"
INCLUDE_ASM("asm/nonmatchings/libpad2", _padSetAct);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padSetCmd);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padSendAtLoadInfo);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padRecvAtLoadInfo);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padGetActSize);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padLoadActInfo);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E480);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E528);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padSetActAlign);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E804);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E820);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padSetMainMode);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E980);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005E9D4);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padCmdParaMode);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EA54);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EA68);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EA88);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EAA8);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EAC8);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EAE8);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EB28);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005EC00);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005ECC0);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005ED4C);
INCLUDE_ASM("asm/nonmatchings/libpad2", PadInitDirect);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F228);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F290);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F384);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F394);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F450);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F6CC);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F704);
INCLUDE_ASM("asm/nonmatchings/libpad2", _padInitDirSeq);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F75C);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005F830);
INCLUDE_ASM("asm/nonmatchings/libpad2", _dirFailAuto);
INCLUDE_ASM("asm/nonmatchings/libpad2", func_8005FB70);
INCLUDE_ASM("asm/nonmatchings/libpad2", setRC2wait);
INCLUDE_ASM("asm/nonmatchings/libpad2", chkRC2wait);
tools/exclude_audit.py
+9 -4
View File
@@ -116,14 +116,19 @@ def classify(path=None, rows=None):
st = stubs.get(f)
if st is None:
out.append((b, f, 'BANKED', 'no INCLUDE_ASM stub in the source')); continue
sub = os.path.basename(os.path.dirname(st.asm_path))
if sub in linked_of.get(b, ()):
# LINKED outranks a pinned WALL (P31 S79 #5). A wall verdict is a fact about a compiler and
# a function; whether the function is OURS to match is a provenance question that precedes
# it. With the pin checked first, PopMatrix/PushMatrix sat here as "compiler walls" for
# eleven sessions while living in libgte3 (LINKED since Phase 8), and the four §332
# "%lo-in-a-delay-slot" walls survived the day libpad 4.2.1 linked their objects (§490).
out.append((b, f, 'LINKED', f'subseg {sub} is linked PsyQ — never a target')); continue
if pinned == 'WALL':
# BANKED still wins above -- a wall that got matched is simply no longer a wall.
# BANKED and LINKED win above -- a wall that got matched or linked is simply no longer a wall.
out.append((b, f, 'WALL',
why or 'curated compiler fact (pinned in the list; not re-derivable)'))
continue
sub = os.path.basename(os.path.dirname(st.asm_path))
if sub in linked_of.get(b, ()):
out.append((b, f, 'LINKED', f'subseg {sub} is linked PsyQ — never a target')); continue
txt = open(st.asm_path, errors='replace').read()
# §401: a jtbl function jumps through a register that is NOT $ra; `jr $ra` ends every function
if re.search(r'\bjr\s+\$(?!ra\b)\w+', txt):
tools/frontier_classify.py
+11 -5
View File
@@ -22,11 +22,17 @@ def _load(m, rel):
s = importlib.util.spec_from_file_location(m, os.path.join(REPO, rel))
x = importlib.util.module_from_spec(s); sys.modules[m] = x; s.loader.exec_module(x); return x
LINKED = set('apicard1 apicard2 apicard3 apicard4 libc2_1 libc2_2 libcd1 libcd2 libetc libgpu '
'libgs1 libgs2 libgs3 libgs4 libgs5 libgs6 libgte1 libgte10 libgte11 libgte12 libgte13 '
'libgte14 libgte15 libgte16 libgte17 libgte18 libgte19 libgte2 libgte20 libgte21 '
'libgte22 libgte3 libgte4 libgte5 libgte6 libgte7 libgte8 libgte9 libmcrd1 libmcrd2 '
'snd1 snd2 snd3 snd4 snd5 snd6 snd7 snd8 snd9'.split())
def _linked_subsegs():
"""main's LINKED subsegs, DERIVED from the Makefile's psyq_integrate stub lists (progress.py's parser).
P31 S79: this was a hard-coded 49-name set (R51 — a derived property stored as config goes stale and
takes a metric with it). It missed every block wired after it was written (libgte23-30, libgs7/8,
snd10-12, libgpu2, libapi1/2, libpad1/2, apicard5-7) and reported 337 "open stubs" — 300 of them the
INCLUDE_ASM records of linked Sony objects — while progress.py, reading the same Makefile, said 16."""
progress = _load('progress', 'tools/progress.py')
return set(progress._main_linked_segs_from_makefile())
LINKED = _linked_subsegs()
INS = re.compile(r'\s*/\* [0-9A-F]+ [0-9A-F]{8} ')
tools/make_apicard_used.py
+20 -11
View File
@@ -1,10 +1,15 @@
#!/usr/bin/env python3
"""Build .run/obj40/apicard_used = the combined libapi+libcard objects in the 800c2 region (Phase 8).
"""Build .run/obj42/apicard_used = the combined libapi+libcard objects in the 800c2 region (Phase 8; S79 #5).
libapi (BIOS syscall trampolines) and libcard interleave in 0x80061F38..0x80062888 with one shared
object (C112.o, identical in both libs). Linked as one 22-object combined region. No exclusions — every
object byte-matches. (libapi ALSO has ~22 objects in the 800c3 region 0x5CE18.., DEFERRED — lowest
value, separate resegmentation.) Regenerate: tools/psyq_build_libs.sh LIBAPI LIBCARD first.
libapi (BIOS syscall trampolines + the C objects FIRST/PAD/PATCH/CHCLRPAD) and libcard interleave in
0x80061F38..0x80062888 with one shared object (C112.o, identical in both libs). Linked as one 26-object
combined region tiling the whole range (apicard1..7). No exclusions — every object byte-matches.
P31 S79 #5: libapi comes from `.run/obj42/libapi42` — PsyQ LIBAPI.LIB **4.2**, the EXE's real libapi (its 4.2
`Ps` stamp is C114 at 0x8005CE48; tools/psyq/lib421, cookbook §490) — which also supplies FIRST.o (168 ins,
`firstfile`, formerly REAL C in 800c2), PAD.o (192, formerly 800c2_2) and PATCH.o + CHCLRPAD.o (formerly
800c2_3). libcard stays 4.0 (`.run/obj40/libcard`): byte-identical in this region. libapi 4.2's band objects
(0x8005CE18..0x8005E188) are wired separately by the Makefile's LIBAPI42 call from the raw dir.
Regenerate: tools/psyq_build_libs.sh LIBCARD + the lib421 ELF step (docs/SETUP.md) first.
"""
import os, re, shutil, subprocess, sys
from collections import defaultdict
@@ -18,8 +23,12 @@ VRAM_BASE = 0x8000F800
RLO, RHI = 0x80061F38, 0x80062888
# object sources: libapi 4.2 (S79 #5) + libcard 4.0 — see the module docstring
SOURCES = (".run/obj42/libapi42", ".run/obj40/libcard")
def place(lib, exe, vram_base):
out = subprocess.check_output(["python3", "tools/psyq_identify.py", f".run/obj40/{lib}",
out = subprocess.check_output(["python3", "tools/psyq_identify.py", lib,
"--vram-base", hex(vram_base), "--exe", exe], text=True)
return {m.group(2): int(m.group(1), 16)
for ln in out.splitlines()
@@ -35,10 +44,10 @@ def main():
exe_path, vram_base = args.exe, int(args.vram_base, 0)
exe = open(exe_path, "rb").read()
byaddr = defaultdict(list)
for lib in ("libapi", "libcard"):
for lib in SOURCES:
for nm, a in place(lib, exe_path, vram_base).items():
byaddr[a].append((lib, nm))
dst = ".run/obj40/apicard_used"
dst = ".run/obj42/apicard_used"
shutil.rmtree(dst, ignore_errors=True)
os.makedirs(dst)
n = 0
@@ -46,12 +55,12 @@ def main():
if not (RLO <= a < RHI):
continue
lib, nm = next(((l, m) for l, m in byaddr[a]
if link_object(f".run/obj40/{l}/{m}", a, name=m, exe_bytes=exe,
if link_object(f"{l}/{m}", a, name=m, exe_bytes=exe,
vram_base=vram_base)["ok"]),
byaddr[a][0])
shutil.copy(f".run/obj40/{lib}/{nm}", f"{dst}/{nm}")
shutil.copy(f"{lib}/{nm}", f"{dst}/{nm}")
n += 1
print(f"apicard_used: {n} objects (libapi+libcard combined, 800c2 region)")
print(f"apicard_used: {n} objects (libapi 4.2 + libcard 4.0 combined, 800c2 region) -> {dst}")
if __name__ == "__main__":