mirror of
https://github.com/Druthulu/BFM-decomp
synced 2026-10-01 15:43:06 -04:00
feat(phase-13): T1a/T1b — SC01/077 onboarded + all-asm byte-match (d19c9580)
- ov_SC01_077 scaffolded: config/overlays.mk (OVERLAY_BINARIES + var block), config/splat.ov_SC01_077.yaml (from template), check.sha, empty symbols, + ov_SC01_077 entry & sentinel anchor in the 4 Python BINARIES dicts - GATE T1b: make check BINARY=ov_SC01_077 -> d19c9580 BYTE-IDENTICAL @ 100% INCLUDE_ASM; clean rebuild leaves main 143dbb89 + resident 8e17e02f unregressed - reusable non-word-aligned-overlay handling (~75% of fleet) added to the template + Makefile: (1) [word_floor, bin, trailing] carve for the final 1-3 bytes spimdisasm drops; (2) build/assets/%.o incbin rule (+ .data align=1) for splat bin assets (asset_path scoped per-alias); (3) objcopy end-align TRIM (shrink-only, <=3 B) removing the .ld's segment-end ALIGN(.,4) pad - .gitignore /assets/ (regenerable splat output); clean removes assets/
This commit is contained in:
Drew T
@@ -29,6 +29,7 @@
|
||||
# config/splat.us.exe.yaml + config/symbols.us.txt + the EXE. The durable RE record
|
||||
# is the symbol file, not these (build scaffolding, not committed work).
|
||||
/asm/
|
||||
/assets/
|
||||
/include/include_asm.h
|
||||
/include/macro.inc
|
||||
/include/labels.inc
|
||||
|
||||
@@ -366,6 +366,14 @@ ASM_SRCS := $(shell find $(ASM_DIR) -name '*.s' -not -path '$(ASM_DIR)/nonmatchi
|
||||
C_SRCS := $(shell find $(SRC_DIR) -name '*.c' $(SRC_PRUNE) 2>/dev/null)
|
||||
OBJS := $(ASM_SRCS:%.s=build/%.o) $(C_SRCS:%.c=build/%.o)
|
||||
|
||||
# splat `bin` subsegs (raw byte regions — e.g. an overlay's trailing non-word-aligned bytes that
|
||||
# spimdisasm's data path drops, since it won't emit a <4-byte partial word). splat extracts them to
|
||||
# assets/<alias>/*.bin and references build/assets/<alias>/*.o in the .ld; wrap each raw .bin into a
|
||||
# linkable object (bytes verbatim in .data). Per-binary: asset_path is scoped to assets/<alias> so
|
||||
# overlays' same-named `trailing.bin` never collide; main/resident have no assets -> empty.
|
||||
ASSET_BINS := $(shell find assets/$(BINARY) -name '*.bin' 2>/dev/null)
|
||||
ASSET_OBJS := $(ASSET_BINS:assets/%.bin=build/assets/%.o)
|
||||
|
||||
# extract: splat split -> asm/, the linker script, include/ macros, undefined_*_auto.txt.
|
||||
extract:
|
||||
@mkdir -p $(OUT_DIR)
|
||||
@@ -391,6 +399,17 @@ build/asm/%.o: asm/%.s
|
||||
@echo " AS $@"
|
||||
@$(AS) $(ASFLAGS) -o $@ $<
|
||||
|
||||
# Wrap a splat `bin` asset (raw bytes) into a linkable object: assemble a one-line stub that
|
||||
# .incbin's the raw file into .data (format-safe — same mipsel-as as everything else, no objcopy
|
||||
# -I binary arch guessing). The .ld pulls it by path. The --set-section-alignment forces .data to
|
||||
# 1-byte align (as defaults it to 16) so ld places the 1-3 trailing bytes at the exact word-floor
|
||||
# offset and does NOT pad the image up to a 16/8-byte boundary (that added a stray byte otherwise).
|
||||
build/assets/%.o: assets/%.bin
|
||||
@mkdir -p $(dir $@)
|
||||
@echo " INCBIN $@"
|
||||
@printf '.section .data\n.incbin "%s"\n' "$<" | $(AS) $(ASFLAGS) -o $@
|
||||
@$(OBJCOPY) --set-section-alignment .data=1 $@
|
||||
|
||||
# C path (Phase 6): modern cpp -> vintage cc1 -> maspsx -> modern as. Each src/*.c is
|
||||
# splat-generated INCLUDE_ASM stubs (file-scope __asm__ .include of the per-function
|
||||
# asm/nonmatchings/<seg>/<fn>.s); as we match, stubs are replaced by real C. The flags
|
||||
@@ -412,7 +431,7 @@ build/src/%.o: src/%.c
|
||||
build/src/boot.o: CC1FLAGS := -quiet -O0 -G0 -mips1 -mcpu=3000 -mgas -msoft-float -fgnu-linker
|
||||
|
||||
# link (the .ld pulls in the .o by path) + objcopy to the raw PS-X EXE image.
|
||||
$(OUT): $(OBJS) $(LD_SCRIPT)
|
||||
$(OUT): $(OBJS) $(ASSET_OBJS) $(LD_SCRIPT)
|
||||
@set -e
|
||||
mkdir -p $(dir $@)
|
||||
# PsyQ SDK library integrations are EXE-only (libgs/libgte/sound/apicard are
|
||||
@@ -473,6 +492,13 @@ endif
|
||||
$(LD) -T $(LD_SCRIPT) -T $(UNDEF_SYMS) -T $(UNDEF_FUNCS) $$SYMS --no-check-sections -Map $(MAPFILE) -o $(ELF)
|
||||
echo " OBJCOPY $@"
|
||||
$(OBJCOPY) -O binary $(ELF) $@
|
||||
# Trim the linker's end-of-segment 4-align pad: splat's .ld does `. = ALIGN(., 4)` after the
|
||||
# data section, over-emitting up to 3 zero bytes when the payload size isn't 4-aligned (most
|
||||
# overlays; the EXE + resident are 4-aligned so this never fires for them). Shrink-ONLY and
|
||||
# capped at 3 bytes -> it can never hide a real shortfall (build < target fails the SHA) nor
|
||||
# extend the image. The target's true byte length is the matched payload $(EXE).
|
||||
tsz=$$(stat -c%s "$(EXE)"); osz=$$(stat -c%s "$@"); d=$$((osz - tsz))
|
||||
if [ "$$d" -gt 0 ] && [ "$$d" -le 3 ]; then truncate -s "$$tsz" "$@"; echo " TRIM $@ (-$$d B linker end-align pad)"; fi
|
||||
|
||||
# build = produce $(OUT) and verify its SHA1 (check pulls in $(OUT)).
|
||||
build: check
|
||||
@@ -509,6 +535,6 @@ expected: build
|
||||
# clean: remove ALL regenerable outputs (build/ + the splat tree) so a config change
|
||||
# is followed by a stale-free `make clean && make extract && make build` (H3).
|
||||
clean:
|
||||
@rm -rf build expected asm undefined_syms_auto.txt undefined_funcs_auto.txt
|
||||
@rm -rf build expected asm assets undefined_syms_auto.txt undefined_funcs_auto.txt
|
||||
@rm -f include/include_asm.h include/macro.inc include/labels.inc include/gte_macros.inc
|
||||
@echo "clean: removed build/, expected/, and the regenerated splat tree (asm/, include macros, undefined_*_auto.txt)."
|
||||
@echo "clean: removed build/, expected/, and the regenerated splat tree (asm/, assets/, include macros, undefined_*_auto.txt)."
|
||||
|
||||
@@ -0,0 +1 @@
|
||||
d19c9580a02dc63ba1f0e7e0c770f3b10de35635 ov_SC01_077
|
||||
@@ -0,0 +1,34 @@
|
||||
# config/overlays.mk — GENERATED overlay-binary registry (Phase 13).
|
||||
# =============================================================================
|
||||
# Holds OVERLAY_BINARIES + the per-<ov> Makefile variable blocks for every onboarded
|
||||
# location overlay. `-include`d by the Makefile BEFORE its `BINARIES :=` line so the
|
||||
# aliases enter the build. Regenerated idempotently by tools/new_overlay.sh (T4); the
|
||||
# hand-maintained Makefile body is never edited to add an overlay. Committed (tiny
|
||||
# config, not ROM-derived). Each block mirrors the resident_* block (the flat-blob
|
||||
# recipe): EXE -> the extracted 0.4.dec, VRAM_BASE/TEXT_LO = the shared slot 0x80128158,
|
||||
# nested asm/<ov> + src/<ov> + build/<ov>, per-binary undefined_*_auto + GHIDRA_PROG.
|
||||
# =============================================================================
|
||||
OVERLAY_BINARIES := ov_SC01_077
|
||||
|
||||
# --- ov_SC01_077 (location overlay SC01/FILE_077, vram 0x80128158 — Phase 13 T1) ---
|
||||
# The milestone overlay (load byte-proven, Phase 3 T6b). 731,607 B (0xB29D7) -> end vram
|
||||
# 0x801DAB2F. Decompressed 0.4.dec payload (LZSS repack is Gen3, not in the verify path).
|
||||
ov_SC01_077_EXE := extracted/retail/SC01.CD.dir/FILE_077.dir/0.4.dec
|
||||
ov_SC01_077_NAME := ov_SC01_077
|
||||
ov_SC01_077_OUT_DIR := build/ov_SC01_077
|
||||
ov_SC01_077_OUT := $(ov_SC01_077_OUT_DIR)/$(ov_SC01_077_NAME)
|
||||
ov_SC01_077_ELF := $(ov_SC01_077_OUT).elf
|
||||
ov_SC01_077_MAPFILE := $(ov_SC01_077_OUT).map
|
||||
ov_SC01_077_LD_SCRIPT := $(ov_SC01_077_OUT).ld
|
||||
ov_SC01_077_SPLAT_YAML := config/splat.ov_SC01_077.yaml
|
||||
ov_SC01_077_CHECK_SHA := config/check.ov_SC01_077.sha
|
||||
ov_SC01_077_SYMBOLS := config/symbols.ov_SC01_077.txt
|
||||
ov_SC01_077_SIG := .run/sig.ov_SC01_077.jsonl
|
||||
ov_SC01_077_GHIDRA_PROG := ov_SC01_077
|
||||
ov_SC01_077_VRAM_BASE := 0x80128158
|
||||
ov_SC01_077_TEXT_LO := 0x80128158
|
||||
ov_SC01_077_TEXT_HI := 0x801DAB2F
|
||||
ov_SC01_077_ASM_DIR := asm/ov_SC01_077
|
||||
ov_SC01_077_SRC_DIR := src/ov_SC01_077
|
||||
ov_SC01_077_UNDEF_SYMS := build/ov_SC01_077/undefined_syms_auto.txt
|
||||
ov_SC01_077_UNDEF_FUNCS := build/ov_SC01_077/undefined_funcs_auto.txt
|
||||
@@ -0,0 +1,59 @@
|
||||
# splat config — Brave Fencer Musashi location overlay ov_SC01_077 (SC01.CD/FILE_077/0.4.dec)
|
||||
# Phase 13 T1: the milestone overlay (load byte-proven, Phase 3 T6b). Instantiated from
|
||||
# config/splat.us.overlay.template.yaml. Flat decompressed blob @ vram 0x80128158 (the shared
|
||||
# overlay slot); code at file 0x0; data tail from 0x5E978 (end of the last sig_image function,
|
||||
# vram 0x80186AD0) to EOF 0xB29D7 (731,607 B -> end vram 0x801DAB2F). The text/data split is
|
||||
# iterated vs `make check BINARY=ov_SC01_077` (Phase-5/10 precedent). Matches the DECOMPRESSED
|
||||
# 0.4.dec (sha1 d19c9580…); LZSS repack is Gen3 (not in the verify path).
|
||||
name: ov_SC01_077
|
||||
sha1: d19c9580a02dc63ba1f0e7e0c770f3b10de35635
|
||||
options:
|
||||
basename: ov_SC01_077
|
||||
target_path: extracted/retail/SC01.CD.dir/FILE_077.dir/0.4.dec
|
||||
elf_path: build/ov_SC01_077/ov_SC01_077.elf
|
||||
base_path: ..
|
||||
platform: psx
|
||||
compiler: PSYQ
|
||||
|
||||
asm_path: asm/ov_SC01_077
|
||||
src_path: src/ov_SC01_077
|
||||
asset_path: assets/ov_SC01_077 # `bin` subsegs (the trailing carve) -> assets/ov_SC01_077/*.bin
|
||||
build_path: build
|
||||
ld_script_path: build/ov_SC01_077/ov_SC01_077.ld
|
||||
ld_dependencies: True
|
||||
|
||||
undefined_syms_auto_path: build/ov_SC01_077/undefined_syms_auto.txt
|
||||
undefined_funcs_auto_path: build/ov_SC01_077/undefined_funcs_auto.txt
|
||||
|
||||
find_file_boundaries: False
|
||||
|
||||
o_as_suffix: True
|
||||
use_legacy_include_asm: False
|
||||
|
||||
section_order: [".rodata", ".text", ".data", ".bss"]
|
||||
|
||||
symbol_addrs_path:
|
||||
- config/symbols.us.txt
|
||||
- config/symbols.resident.txt
|
||||
- config/symbols.ov_SC01_077.txt
|
||||
|
||||
subalign: 2
|
||||
|
||||
string_encoding: ASCII
|
||||
data_string_encoding: ASCII
|
||||
rodata_string_guesser_level: 2
|
||||
data_string_guesser_level: 2
|
||||
|
||||
segments:
|
||||
- name: ov_SC01_077
|
||||
type: code
|
||||
start: 0x0
|
||||
vram: 0x80128158
|
||||
align: 4
|
||||
subsegments:
|
||||
- [0x0, c, ov_SC01_077] # code: vram 0x80128158..0x80186AD0 (file 0x0..0x5E978)
|
||||
- [0x5E978, data, tail] # data tail (word-aligned bulk): file 0x5E978..0xB29D4
|
||||
- [0xB29D4, bin, trailing] # final 3 bytes (EOF 0xB29D7 not word-aligned; spimdisasm drops
|
||||
# a trailing partial word and a <4-byte `data` carve emits nothing,
|
||||
# so use `bin` = raw .incbin, byte-exact). Word-aligned overlays omit this.
|
||||
- [0xB29D7] # EOF (731,607 B) -> end vram 0x801DAB2F
|
||||
@@ -38,6 +38,10 @@ options:
|
||||
|
||||
asm_path: asm/@ALIAS@
|
||||
src_path: src/@ALIAS@
|
||||
# asset_path scoped per-alias so a `bin` subseg (the trailing-partial-word carve, below) lands at
|
||||
# assets/@ALIAS@/*.bin and never collides with another overlay's same-named asset. The Makefile's
|
||||
# build/assets/%.o rule incbin's it into a linkable object.
|
||||
asset_path: assets/@ALIAS@
|
||||
# build_path = build (NOT build/<alias>): splat writes the .ld's object paths under
|
||||
# $(build_path), and the Makefile's pattern rules build them at build/asm/** + build/src/**.
|
||||
# Only the elf/ld/output live in build/<alias>/.
|
||||
|
||||
@@ -0,0 +1,6 @@
|
||||
// config/symbols.ov_SC01_077.txt — overlay SC01/FILE_077 local symbols (Phase 13)
|
||||
// R13/R15: overlay-REGION ONLY (vram 0x80128158..0x801DAB2F); NEVER merged into
|
||||
// config/symbols.us.txt. Seeded in T1c from sig_image (.run/sig.ov_SC01_077.jsonl) +
|
||||
// docs/duplicates.cross.md to fix conservative splat carves (jump-table-in-rodata, §8/§11)
|
||||
// and attach stable names. The all-asm byte-match (T1b) does not depend on this file.
|
||||
// count=0
|
||||
@@ -20,6 +20,10 @@ BINARIES = {
|
||||
"resident": dict(baseimg="expected/build/resident/resident.elf",
|
||||
myimg="build/resident/resident.elf",
|
||||
mapfile="build/resident/resident.map"),
|
||||
"ov_SC01_077": dict(baseimg="expected/build/ov_SC01_077/ov_SC01_077.elf",
|
||||
myimg="build/ov_SC01_077/ov_SC01_077.elf",
|
||||
mapfile="build/ov_SC01_077/ov_SC01_077.map"),
|
||||
# <<< overlays: tools/new_overlay.sh inserts ov_* entries above this line (Phase 13) >>>
|
||||
}
|
||||
|
||||
|
||||
|
||||
@@ -23,7 +23,8 @@ All load @ vram 0x80128158. SC01/077 starts with CODE at offset 0x0 (no leading
|
||||
|
||||
## Tasks
|
||||
- [x] **T0** — Overlay splat template (`config/splat.us.overlay.template.yaml`) + Makefile `-include config/overlays.mk` indirection. ✅ GATE: clean rebuild main `143dbb89…` + resident `8e17e02f…` BYTE-IDENTICAL (indirection proven a no-op; `overlays.mk` absent → silent `-include` → `BINARIES` unchanged).
|
||||
- [ ] **T1** — SC01/077 end-to-end. T1a scaffold · T1b all-asm byte-match (`d19c9580…`, milestone) · T1c seed boundaries + Ghidra import · T1d dedup-credit · T1e unique-remainder harvest (Ultracode §12 — prompt R26).
|
||||
- [~] **T1** — SC01/077 end-to-end. ✅ T1a scaffold (overlays.mk + splat.ov_SC01_077.yaml + check.sha + symbols + 4 Python dicts w/ sentinel) · ✅ T1b all-asm byte-match `d19c9580…` BYTE-IDENTICAL (milestone; main+resident unregressed) · ☐ T1c seed boundaries + Ghidra import · ☐ T1d dedup-credit · ☐ T1e unique-remainder harvest (Ultracode §12 — prompt R26).
|
||||
- **Reusable finding (→ cookbook §13 + PhaseEnd):** non-word-aligned overlays (≈75% of fleet) need 3 template/Makefile additions: (1) a `[<word_floor>, bin, trailing]` carve for the final 1-3 bytes spimdisasm's data path drops; (2) a `build/assets/%.o` incbin rule (+ `--set-section-alignment .data=1`) wiring splat `bin` assets (scoped `asset_path: assets/<alias>`); (3) an objcopy end-align **trim** (shrink-only, ≤3 B) removing the splat `.ld`'s `. = ALIGN(., 4)` segment-end pad. EXE/resident are 4-aligned so untouched.
|
||||
- [ ] **T2** — Dup-pair collapse SC01/005≡006 (both `56760dbe…` from one shared source).
|
||||
- [ ] **T3** — Free-roam SC03/001 (`f8fd92f5…`); recipe scales to a large overlay.
|
||||
- [ ] **T4** — `tools/new_overlay.sh` one-command onboarding (gen `overlays.mk`; sentinel-insert 4 dicts; idempotent). GATE: onboard a fresh overlay in one command.
|
||||
|
||||
@@ -18,6 +18,9 @@ BINARIES = {
|
||||
"main": dict(src="src", asm="asm/nonmatchings", md="docs/difficulty.md", csv=".run/difficulty.csv"),
|
||||
"resident": dict(src="src/resident", asm="asm/resident/nonmatchings",
|
||||
md="docs/difficulty.resident.md", csv=".run/difficulty.resident.csv"),
|
||||
"ov_SC01_077": dict(src="src/ov_SC01_077", asm="asm/ov_SC01_077/nonmatchings",
|
||||
md="docs/difficulty.ov_SC01_077.md", csv=".run/difficulty.ov_SC01_077.csv"),
|
||||
# <<< overlays: tools/new_overlay.sh inserts ov_* entries above this line (Phase 13) >>>
|
||||
}
|
||||
SRCS = ASM_ROOT = MD = CSV = None # set by main() from --binary
|
||||
|
||||
|
||||
@@ -26,6 +26,8 @@ ROOT = pathlib.Path(__file__).resolve().parent.parent
|
||||
BINARIES = {
|
||||
"main": dict(sig=".run/sig.SLUS_007.26.jsonl", md="docs/duplicates.md"),
|
||||
"resident": dict(sig=".run/sig.resident.jsonl", md="docs/duplicates.resident.md"),
|
||||
"ov_SC01_077": dict(sig=".run/sig.ov_SC01_077.jsonl", md="docs/duplicates.ov_SC01_077.md"),
|
||||
# <<< overlays: tools/new_overlay.sh inserts ov_* entries above this line (Phase 13) >>>
|
||||
}
|
||||
CROSS_MD = "docs/duplicates.cross.md"
|
||||
|
||||
|
||||
@@ -23,6 +23,9 @@ BINARIES = {
|
||||
src="src", asm="asm/nonmatchings", out="docs/progress.md"),
|
||||
"resident": dict(build="build/resident/resident", check="config/check.resident.sha",
|
||||
src="src/resident", asm="asm/resident/nonmatchings", out="docs/progress.resident.md"),
|
||||
"ov_SC01_077": dict(build="build/ov_SC01_077/ov_SC01_077", check="config/check.ov_SC01_077.sha",
|
||||
src="src/ov_SC01_077", asm="asm/ov_SC01_077/nonmatchings", out="docs/progress.ov_SC01_077.md"),
|
||||
# <<< overlays: tools/new_overlay.sh inserts ov_* entries above this line (Phase 13) >>>
|
||||
}
|
||||
BINARY = next((sys.argv[i + 1] for i, x in enumerate(sys.argv)
|
||||
if x == "--binary" and i + 1 < len(sys.argv)), "main")
|
||||
|
||||
Reference in New Issue
Block a user