Commit Graph
256 Commits
Author SHA1 Message Date
Dean c4446bbb09 Merge pull request #9332 from rooklz/fix/windows-console-base
fix(relinker): prefer 0x400000 for Windows executables
2026-10-25 12:57:28 +13:00
SlugZero 58a1bcc5a3 feat(pad): support up to 4 controllers with dynamic user discovery
- Track up to 4 pad slots in libScePad with per-slot inputs, outputs, sensor fusion, and deadbands
- Support dynamic user login/logout events and multi-user registration in libSceUserService
- Map SDL controllers 0-3 to Players 1-4 in libSceVideoOut while preserving Player 1 keyboard and mouse composition
- Add test coverage for multi-controller pads and multi-user services
2026-10-24 19:21:15 +02:00
Adrià d01070d6c3 Merge pull request #5640 from EagleBearerYS/feat/relinker-frame-frameless-leaves
feat(relinker): frame red-zone leaves without a frame pointer by moving their rsp operands
2026-10-24 18:14:53 +02:00
Rookie 39df8e081c fix(relinker): prefer 0x400000 for Windows executables 2026-10-24 15:58:35 +01:00
Dean a8c8b8117b Merge pull request #8900 from WhiteAssassins/feat/patch-advisor
feat(tools): recommend verified local game patches
2026-10-25 03:51:47 +13:00
Adrià b84322da6a Merge pull request #7918 from Kernel-Hunter/fix/rosetta-xsaveopt
fix(relinker): rewrite XSAVEOPT to XSAVE under --to-rosetta
2026-10-24 14:14:59 +02:00
Adrià 0210d3b878 Merge pull request #8909 from dantdmdarealguy/docs/usage-mimalloc-dlls
docs(usage): name the mimalloc DLLs a self-built Windows libs/ needs
2026-10-24 11:29:00 +02:00
Karim Masmoudi ce6c2a0c66 fix(relinker): rewrite XSAVEOPT to XSAVE under --to-rosetta
XSAVEOPT raises an illegal instruction under Rosetta on an M5 running
macOS 27.0.1, as XSAVEC does, while XSAVE and XRSTOR save and restore
x87, SSE and AVX state in the standard format. The PS5's Zen 2 has
XSAVEOPT, so code built for it can use the instruction without a CPUID
check.

XSAVEOPT differs from XSAVE only in the init and modified optimizations:
it may skip writing state that is in its initial configuration or
unchanged since the last XRSTOR from the same area. XRSTOR restores the
same state from the area XSAVE writes, so --to-rosetta rewrites
XSAVEOPT (NP 0F AE /6, memory operand) to XSAVE (/4) in place, keeping
the prefixes and the operand, as RDSEED becomes RDRAND. CLWB (66 0F AE
/6), MFENCE, UMONITOR, UMWAIT and the Intel target are left alone.
XSAVEC is not converted: it writes the compacted format.

The converter tests check the rewrite for the plain, REX.W,
RSP-relative, extended-base, RIP-relative and FS-relative forms, check
that the neighbouring 0F AE encodings are not taken, and execute the
rewritten XSAVEOPT64 followed by XRSTOR on YMM3 in the x86-64 build.
The --to-rosetta integration test now covers XSAVEOPT as well.
2026-10-24 11:01:44 +02:00
Dean ec3ea2f558 Merge pull request #6586 from JJenkx/fix/linux-runtime-module-start
fix(loader): defer Linux runtime module starts until requested
2026-10-24 19:54:29 +13:00
Dean 5540852c54 Merge pull request #9241 from veosett/codex/audit-sdk-symbol-names
fix(tools): flag qualified imports exported only under a NID
2026-10-24 14:21:32 +13:00
JJenkx adf0176d3a fix(loader): preserve SDK startup in browser conversions 2026-10-23 11:15:40 -04:00
Santhosh Kumar 3bcd89e10e fix(agc): reclaim idle images safely during cache creation
Forward new-image recovery permission through AllocateImageMemory while preserving host-memory fallback and allocated-type reporting. Reuse held cache locks only at safe construction points, retain caller-held sampled textures and pending storage results, and try resource-cache removal without waiting.

Count released Vulkan allocations and saturate the reclaim reserve. Cover sampled/storage allocation failures, host fallback after image recovery, bounded attempts and overflow. Remove stale conflict markers and unrelated debt, and refresh the recorded validation.
2026-10-23 16:37:26 +02:00
Dean 5e9f4b8966 Merge pull request #9832 from johandavidcm/fix/libkernel-dlopen-unrelinked
fix(libkernel): fail dlopen of an unrelinked module without the Windows Bad Image dialog
2026-10-23 23:18:00 +13:00
Kovalev Vladislav 46ff145113 fix(tools): flag qualified imports exported only under a NID 2026-10-23 12:18:57 +06:00
Yunus Shaikh a376e37b88 fix(relinker): decode each red-zone leaf over its whole FDE range 2026-10-23 05:22:26 +00:00
Yunus Shaikh 542cb30f70 docs(relinker): describe the operand stubs and refused patches of --frame-red-zone
The debt entry no longer says that a leaf with an rsp operand that
cannot move in place is skipped. A new entry says that such an operand
moves to a stub, when the leaf is still skipped, and that a game patch
over a framed site or a moved displacement is refused. The USAGE row
says the same in short.
2026-10-23 04:50:13 +00:00
Yunus Shaikh a665ef2dfb docs(relinker): describe the per-leaf shift and sub allocations of --frame-red-zone 2026-10-23 03:45:27 +00:00
Yunus Shaikh ff16118b63 Merge branch 'main' into feat/relinker-frame-frameless-leaves-next 2026-10-23 03:13:53 +00:00
raghav314andClaude Opus 5.5 c0dbd1871c perf(agc): draw finisher pipeline (APS5_DRAW_PIPELINE)
The second half of a pipelined draw (stage materialization, cache insert,
enqueue) runs on a finisher thread, overlapping the next draw's packet
work; draws whose stages could write memory are held back. Commits skip
work whose inputs did not change since the last commit, color metadata
passes are ordered through the pipeline, resource and storage-image proofs
are reused within one collect epoch until a write breaks them, and a shape
entry is refreshed in place at the second sight of the same read user words.
APS5_VERIFY_DRAW_PIPELINE checks reused proofs against a full check.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-23 03:46:57 +05:30
Dean d2b1c3edd5 Merge pull request #9079 from JJenkx/feat/pad-input-chords
feat(tools): press several buttons together from pad_input.py
2026-10-23 11:09:00 +13:00
Dean 44ce348a74 Merge pull request #9291 from veosett/codex/self-alignment-padding
fix(tools): bound total SELF alignment padding
2026-10-23 09:40:16 +13:00
Adrià da750e65c0 Merge pull request #9354 from xXPBSXx/docs/input-mapping-overlay-default
docs(input): list the F10 performance overlay in the default bindings table
2026-10-22 20:22:54 +02:00
Johan David dc84bec7b9 fix(libkernel): fail dlopen of an unrelinked module without the Windows Bad Image dialog
On Windows, dlopen passed a module that was never relinked to LoadLibraryExW, which raised the
modal Bad Image hard-error dialog and blocked the calling thread until someone answered it. The
load now runs with SEM_FAILCRITICALERRORS set for that thread only, so it fails at once with the
same result as before. Every failed load is reported once per file on stderr with a [module]
prefix; a file that exists with no relinked .guest.prx names the relinker's --module-dir option.
2026-10-22 13:05:34 -05:00
Dean 830ccd121f Merge pull request #9303 from oneandonlydean/fix/getdents-app0-symlink-type
fix(libkernel): report app0 symlinks as their target type in getdirentries
2026-10-23 02:19:24 +13:00
xXPBSXxandClaude Sonnet 5.5 949737625b docs(input): list the F10 performance overlay in the default bindings table
Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
2026-10-21 21:00:50 +02:00
Rookie 123ba70839 fix(relinker): load Linux executables at 0x400000 by default 2026-10-21 19:25:33 +01:00
Kovalev Vladislav da1d9407ed fix(tools): bound total SELF alignment padding 2026-10-21 23:23:52 +06:00
Dean Galvin 58ad4eda77 fix(libkernel): report app0 symlinks as their target type in getdirentries
On Linux sceKernelGetdirentries copied the host d_type, so a directory under app0 that is a host symlink was listed as a link, and a title that only descends into directory entries skipped it. Inside the app0 tree, report a symlink (or an entry of unknown type) as the type of its target; links elsewhere, and dangling links, stay links.
2026-10-22 06:20:04 +13:00
Dean 27bcd6a29e Merge pull request #8162 from savvaniss/fix/windows-runtime-module-start
fix(relinker): start Windows runtime modules the executable needs when the title loads them
2026-10-22 02:10:11 +13:00
JJenkxandClaude Opus 5.5 505943c4ef feat(tools): press several buttons together from pad_input.py
Some titles wait for a chord on their title screen: RollerCoaster Tycoon
Adventures Deluxe (PPSA17330) asks for L1 + R1. The command file already
carries a button bit field, but the helper accepted one name. Names joined
by '+' are now combined into one command; unknown names, duplicates and
release in a chord are rejected without touching the command file.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-21 07:55:36 -04:00
WhiteAssassins 087cf11045 fix(tools): continue after relinker candidate errors 2026-10-21 05:16:06 -04:00
WhiteAssassins efab492f4f fix(tools): verify local relinker and isolate bad catalog entries 2026-10-21 05:05:36 -04:00
WhiteAssassins c9fe159e6d docs: clarify local patch advisor usage 2026-10-21 04:51:59 -04:00
WhiteAssassins 21481b3cd4 test(tools): verify patch advice with real relinker 2026-10-21 04:37:58 -04:00
dantdmdarealguy e2b5a3e9d4 docs(usage): name the mimalloc DLLs a self-built Windows libs/ needs 2026-10-21 00:40:55 -05:00
WhiteAssassins 05f005e341 feat(tools): verify additional executables in patch advisor 2026-10-21 01:35:06 -04:00
WhiteAssassins 04ec954284 feat(tools): recommend verified local game patches 2026-10-21 01:05:51 -04:00
JJenkxandClaude Opus 5.5 3316ca3b24 feat(libc): list the reporting thread's own calls in the call trace report
The crash and abort reports printed the last 64 traced calls of the whole
process, so a thread that loops on a traced import (a mutex lock/unlock
pair, a profiler marker) pushes the calls of the faulting thread out of the
report. The ring already keeps the last 65536 calls with their thread.

CopyThreadCallTrace_nid_no_patch returns the most recent calls of one
thread from that ring, oldest first, and both reports (Windows crash and
abort, Linux abort) list up to 32 of them for the reporting thread after
the process-wide list.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-20 20:25:31 -04:00
Dean Galvin 143ed02750 Merge branch 'main' into #7417
Resolve docs/dev/debt/libc.md: keep main's _Throw_Cpp_error entry and this branch's call tracing entry.
2026-10-21 06:33:59 +13:00
Dean 083a3be61c Merge pull request #8058 from HQasper/claude/project-thread-md0z38
feat(libSceFont): open the other script files of a system font set
2026-10-21 06:26:24 +13:00
Dean 1a3a1b2e0e Merge pull request #3998 from MrMha/feat/controller-timelines
feat(pad): queue timed controller button sequences
2026-10-21 04:08:30 +13:00
Adrià 4cfcb003de Merge pull request #7862 from parl0v/fix/storage-budget-sampled-reserve
fix(agc): reserve only the sampled textures alive from the storage image budget
2026-10-20 16:45:56 +02:00
Adrià 576b4b94b2 Merge pull request #7873 from whamre/feat/hang-dump-stack-candidates
feat(libc): list stack return address candidates in the hang dump
2026-10-20 16:20:04 +02:00
Savvanis Spyros 079e6ca316 Merge remote-tracking branch 'origin/main' into fix/windows-runtime-module-start 2026-10-20 15:27:06 +03:00
Josip Parlov 321460a4ed fix(agc): reserve only the sampled textures alive from the storage image budget
StorageTextureBudget reserved the larger of the sampled textures alive
and their 2 GiB floor. The sampled budget already reserves
TextureCacheBudget, at least the storage cap, for the storage cache, so
on an 8 GB card the two caches each kept the other's maximum: with
3936 MiB used outside the caches on a 7189 MiB budget and 49 MiB of
sampled textures, the storage budget came out at 306 MiB and sat at its
512 MiB floor while about 2.7 GB of the budget stayed free.

The storage budget now reserves the sampled textures alive and gives
room back as they grow, at its next read. Where they hold 2 GiB or more
the budget is unchanged.
2026-10-20 13:09:20 +02:00
Adrià 5b2654a28b Merge pull request #6771 from XairRrR/fix/pipeline-cache-version
fix(agc): keep the Vulkan pipeline cache file with its shader cache version
2026-10-20 11:07:49 +02:00
Adrià 7faab0aea9 Merge pull request #4891 from dantdmdarealguy/feat/screenshot-hotkey
feat(agc,videoout): save the displayed frame as a PNG with a screenshot key
2026-10-20 10:40:02 +02:00
Ethan Barcus a8e06fb008 Merge branch 'main' into feat/trace-bench-fields 2026-10-20 02:57:23 -05:00
dantdmdarealguy e3125e43af feat(agc,videoout): save the displayed frame as a PNG with a screenshot key
F12 (the new Screenshot input action) requests a capture of the next
presented display buffer. The presenter records it with the existing
APS5_DUMP_FRAMES GPU readback at full resolution, and the frame dump
writer thread encodes it with decoder_png into anyps5-screenshots/
beside the executable, or ANYPS5_SCREENSHOTS.
2026-10-20 00:28:39 -05:00
Savvanis SpyrosandClaude Opus 5.5 45cc02c49c fix(relinker): start Windows runtime modules the executable needs when the title loads them
A --module-dir module that only the executable needs was initialized by
the Windows entry stub with no start arguments, so a later
sceKernelLoadStartModule with arguments found it started and its
module_start never saw them. Unity's mono-ps4.sprx keeps those
arguments for the plugins it loads. Such a module is now deferred like
any other runtime module: mapped at startup and started, with its
arguments, when the title loads it. Modules another guest module needs
are still started at startup, and Linux and macOS are unchanged.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-20 03:57:24 +03:00