Commit Graph
10 Commits
Author SHA1 Message Date
Valiantsin DzerakhandClaude Opus 5.5 c2943970cc Merge the spin wait gates: they also end the fight hangs
The generator emits a guest poll such as `ldr w9, [x19]; cmp w9, #1; b.eq` as a plain C load, and clang hoists it out
of the loop: the lock wait in f_7001fa0e2c compiled to `cmp; jne; jmp .`, so a thread that found the lock held spun
forever after its holder released it. A fight run hung this way in 3 of 13 launches on an RX 9070, with the UBO
snapshot on and off; gdb showed the thread on that `jmp` in every sample. The gate is an opaque call on every
back-edge of the 868 loops tools/spin_waits.py lists, so each pass reloads memory: the exe goes from 182 self-jumps to
148, and 144 of those are the game's own `b .` instructions. 8 of 8 fight runs then passed; tests/diff/func.py with
--patches passed 5470 cases on 878 functions, and the four probes passed on 1.4.3.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 20:05:44 +03:00
Valiantsin DzerakhandClaude Opus 5.5 1f4c6fcdb5 Decode the game's zstd frames with the host's libzstd: the patch set zstd (TOTK_ZSTD_NATIVE, on by default) sends the calls of ZSTD_decompressMultiFrame and ZSTD_decompress_usingDDict in main to runtime/ov_zstd.c, which makes host dictionaries from the guest DDicts, gives the virtual core up for inputs of 64 KiB and more, and leaves in-place decodes and every error to the guest decoder
The lifted decoder ran its copy loops as chains of small functions. On the island and fight routes TOTK_ZSTD_CHECK=1 decoded every call both ways with no difference; the host took 0.72 s where the guest decoder took 1.99 s. Load time and fps on an RX 9070 and a 7900X did not change: the gain is CPU time on the loading threads, for machines with fewer cores. The caller at main+0x547a04 stays on the guest decoder: it resets a static DCtx with which the game's zstd rejects some frames libzstd decodes, and the game relies on that result.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 18:27:12 +03:00
Valiantsin DzerakhandClaude Opus 5.5 2fe6a6d4c1 Keep MXCSR in line with FPCR where FPCR changes instead of checking it in every FP op
The FP helpers compared c->fpcr with a cached copy before every operation to resync MXCSR lazily. MXCSR now
follows FPCR.FZ at once: MSR FPCR calls rt_fpcr_sync, and so do rt_call (thread, fiber and host entry into guest
code), fiber resume (FIBER_FLAG_FLOAT_SWITCH fibers keep their own MXCSR) and the differential test harness per
case. op and vop only test FPCR.RMode for the slow path, and their redo check is one ordered-not-equal compare of
|r| with the smallest normal, the same NaN-or-smallest-normal set as before.

Hot guest code shrinks by about a quarter (132 functions holding half the guest cycles, 1.4.3); fight frame time
moves 1-3%, inside the run-to-run spread. lift, simd, func, atomic and flags_check pass on 1.4.3; simd.py fails
without the harness sync.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 12:25:25 +03:00
Valiantsin DzerakhandClaude Opus 5.5 0ca11dfba2 Keep the six most used general registers of each generated function in C locals
Every lifted instruction read and wrote c->x[n] and c->sp, so each register passed from one guest instruction to the
next went through a store and a load, and clang could not keep it in a host register across a guest store.
recomp/regs.py now copies the six most used X registers and SP of a function into locals at its entry. A forward pass
over the function's flow tracks which locals changed since the last sync, and only those are stored back before a
call, an SVC, a trap, a return or a tail exit. After an SVC every local reloads; after a call only X0-X18 and X30,
since AAPCS64 has the callee preserve X19-X29 and SP. TotK does not break that: main and subsdk0 import neither
longjmp nor the unwinder, sdk's have no direct callers, and the runtime writes no register above X3 except X30.
gen.py --no-callee-saved-trust reloads them too, --reg-locals sets the count. Functions with an unresolved computed br
keep their registers in the Cpu, as they keep NZCV.

All registers in locals was slower than none: clang spills them to the stack and the syncs stay. Over 300 hot
functions, 4-6 locals gave the fewest instructions and memory operands.

1.4.3 .text 299 -> 266 MiB. Fight bench (bench-fight2-2026-10-27b, frames 5400-7500, uncapped): 11.30 and 11.92 ms vs
13.5-13.9 on main; frames 2400-5300 11.24 and 11.36 vs 13.25. Without the callee-saved trust 12.58 ms.
tests/diff/func.py: 0 failing on 1.4.3 (889 functions, 583 with calls) and 1.2.1 (810); a plan with no spills fails
at once. Probes kernel, vkexec, nvn-tex and shader pass in both versions.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 22:21:18 +03:00
Valiantsin DzerakhandClaude Opus 5.5 27056b212a Keep NZCV in a local of each generated function instead of c->nzcv
Every flag-setting instruction packed N, Z, C and V into c->nzcv and every reader unpacked them. recomp/flags.py
now runs a reaching-definitions pass per function: a compare stores its operands in a local Flags F, and a reader
whose definitions are all of one kind computes its condition from them, so cmp and b.cond become one host compare.
CCMP, FCMP, MSR NZCV and ADCS write the packed form, which readers reached by several kinds also use. Calls and
returns publish nothing (AAPCS64 neither passes nor preserves NZCV); tail branches, falls into the next function and
br through rt_dispatch publish c->nzcv, since discovery can start a function inside another one. Functions with an
unresolved computed br keep flags in c->nzcv.

1.4.3 main: 492k direct and 89k packed flag reads, .text 342 -> 299 MiB. Fight bench (bench-fight2-2026-10-27b,
frames 5400-7500, uncapped) 13.15 and 13.63 ms vs 13.9 before. tests/flags_check.py checks the new helpers against
the packed ones; tests/diff/func.py no longer compares NZCV, which generated code leaves untouched.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:07:58 +03:00
Valiantsin DzerakhandClaude Opus 5.5 3984344635 Spin wait gates: tools/spin_waits.py (the nxnative search) lists the loops that only another thread can end, the patch tables carry them as spin lines (857 in 1.4.3, 864 in 1.2.1), the generator sends their back-edges through rt_spin_gate, and the gate pauses the core once per pass and hands the virtual core to a waiter after 50 us of polling (TOTK_SPIN, TOTK_SPIN_HANDOFF_US)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 15:51:47 +03:00
Valiantsin DzerakhandClaude Opus 5.5 7abdbc90ac Render the game above 1080p and scale it to the window with FSR 1
The output and scene patch sets make the game read its output, swapchain and scene sizes from the patch page
(MacNativeGames' sites for 1.2.1 and 1.4.3); resolution.c picks them from TOTK_OUTPUT_RES and TOTK_RENDER_SCALE and
grows the HeapDef texture heap, its parent and the application pool to match. vk_fsr.c runs AMD FSR 1 EASU and RCAS
on the present when the window is larger than the game's output (TOTK_UPSCALER, TOTK_SHARPNESS).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 13:40:34 +03:00
Valiantsin DzerakhandClaude Opus 5.5 cd60d20a7f Boot 1.4.3 to its first NVN call with the 1.4.3 patch table, loader values and SDK 20.5 swkbd layout
recomp/patches-v143.txt ports MacNativeGames' 1.4.3 fps60, real-time step, cloth and bounce sites; the real-time
producer reads the nominal interval from patch page +8, which frame_rate.c now writes as seconds * 60.
make_romfs_patches.py takes --version (Common.Product.140 on 1.4.3).

SDK 20.5 grew KeyboardConfig to 0x518 bytes, so ShowKeyboardArg's work buffer is at +0x518; swkbd_win.c and the
services probe pick the offset by version. 1.4.3's movie::Decoder vtable names the complete-object destructor D1,
which gets the D2 override. The loader probe has the measured 1.4.3 import and relocation counts, and probe-ipc
expects REV15 from the 1.4.3 SDK and skips the 1.2.1 sf heap constructors (1.4.3 creates the audio heap lazily).
A non-1.2.1 build defaults to its own profile, build/profile-<version>: 1.4.3 rewrites the saves it loads.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 17:50:40 +03:00
Valiantsin DzerakhandClaude Opus 5.5 776d547169 Load 1.4.3 modules: DT_RELR, symbol count from DT_GNU_HASH, stg as a no-op, version-aware boot anchors and loader probe values
The jump-table check now explains oracle targets for case indices peeled off by cmp #k; b.ne before the table.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 17:27:36 +03:00
Valiantsin DzerakhandClaude Opus 5.5 19c9b648d1 Add the recompiler, Windows runtime, tools and docs
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 22:27:54 +03:00