Commit Graph
955 Commits
Author SHA1 Message Date
FX192 dc57f367e1 itch.io browse: don't re-fire the load-more trigger when a page is appended 2026-10-24 13:07:16 +03:00
FX192 5838587639 itch.io browse: let a failed page be retried instead of ending the list 2026-10-24 13:02:17 +03:00
FX192 7e87cdacb5 Merge remote-tracking branch 'upstream/main' into feat/itch-browse 2026-10-24 12:46:26 +03:00
FX192 dbcbf0b6fa itch.io browse: use itch's own sort, price and platform paths
The sort and price chips now cover everything itch's browse page offers, with the
paths copied from its links: new-and-popular, top-sellers, top-rated and newest
for sorts; free, on-sale, store, 5-dollars-or-less and 15-dollars-or-less for
prices; platform-windows, -osx, -linux, -android, -ios, -web and -mobile-web.
The guessed alternative spelling of the free filter is gone, so a query is one
address. The tests assert the addresses itch itself uses.

New and reworded strings in all 33 locales.
2026-10-24 09:27:49 +03:00
maxjivi05 476642be71 chore: drop the added comments 2026-10-24 01:42:20 -04:00
Nikolay Nadorichev 8c86b718a0 fix: find the signed-in Steam account when loginusers.vdf has no MostRecent
droiddeck-steam-shortcuts and AddedGames.accountConfig took the account
marked "MostRecent" in loginusers.vdf, else the only userdata folder.
The arm64 client here writes no MostRecent, so with two accounts in one
install nothing was picked: no added or store game reached
shortcuts.vdf ("no signed-in account yet; nothing written") and Steam
answered their launch with "App configuration unavailable".

Without MostRecent, take the account marked AutoLogin, then the one with
the latest timestamp. Both sides use the same order so the app reads the
config the guest script writes.
2026-10-24 01:42:20 -04:00
maxjivi05 473ac666b3 Take over Steam's Default startup movie pick 2026-10-24 01:41:08 -04:00
FX192 cc0a4f5aa9 itch.io: browse the Store tab by sort and price, with paging and a disk cache
The Store tab only had four fixed shelves and a title search. A row of sort
(Popular, New, Top rated, Top sellers) and price (Any, Free, On sale) chips now
sets a BrowseQuery; anything but the defaults swaps the shelves for that query's
games, loaded page by page as the grid scrolls.

ItchBrowse builds itch's listing-feed URL (platform and tag segments are in the
builder for the next change), tries each spelling of a price filter until one
answers, spaces requests 1.2 s apart, backs off once on a 429, and keeps pages
in the app cache for three hours. Every request is logged with its status and
game count so a wrong URL guess can be fixed from a device. Owned games replace
their browse copy so Install and the owned markers are right, and the detail
page finds browse results.

Strings for all 33 locales; URL-builder tests; docs/development/itch-browse.md.
2026-10-24 07:49:05 +03:00
maxjivi05 f89dc04072 fix: Proton's EasyAntiCheat and BattlEye runtimes are not listed as games
Steam installs Proton EasyAntiCheat Runtime (1826330) and Proton BattlEye Runtime (1161040) as tool depots. The launcher's Games tab and the desktop's game entries now skip them like the other runtimes.
2026-10-24 00:47:45 -04:00
maxjivi05 dd79f44b3a fix: the Steam download monitor no longer stalls the main thread on a long content log
The content log was read with RandomAccessFile.readLine, one byte per read, on the main thread. A 4 MB log took about 2 s on the first poll, enough for an ANR on slower devices while a download runs. It is now read through a buffered reader with the same Latin-1 lines and offset, and the per-line regexes only run on lines that can match.
2026-10-24 00:27:22 -04:00
maxjivi05 b0e405ebf3 fix: Epic client imports get the launch command, versioned installs need a matching manifest build
Games the Epic client installed now carry their LaunchCommand arguments too, split by a new launch.split op. A doubled quote inside a quoted argument is one literal quote, and an install with a version ignores a manifest without the same build.
2026-10-23 23:59:37 -04:00
maxjivi05 42026b382f fix: keep manifest launch arguments to the installed build and their own additions
The manifest's arguments apply only when its build matches the installed version, only arguments the merge inserted are tracked as its own, escaped quotes split like a Windows command line, and the REDprelauncher rule needs Cyberpunk2077.exe beside it.
2026-10-23 23:59:37 -04:00
maxjivi05 46cd8e1777 fix: Epic games get their manifest launch command
Epic manifests carry a launch command next to the launch executable, which the Epic launcher passes to the game. It was read and dropped, so Caravan SandWitch never got -PCStore=EGS and quit with "started outside of any known PC store". The manifest's launch command now joins the game's arguments, including for games already installed.
2026-10-23 23:59:37 -04:00
XxJSONDeruloxX 546574e982 fix(xbox): enable proot fast path in Xodus 2026-10-23 23:56:24 -04:00
XxJSONDeruloxX b1ab55856e fix(ui): clear search focus and use store accents 2026-10-23 23:56:24 -04:00
XxJSONDeruloxX 4a0eb14d1f fix(xbox): mount selected SD library in Xodus 2026-10-23 23:56:24 -04:00
XxJSONDeruloxX 2c8ae3753c fix(xbox): mount external storage for downloads 2026-10-23 23:56:24 -04:00
Kurt Himebauch fcfbb74068 fix: one refused thermal or power query no longer drops the others from the session report (#991) 2026-10-23 23:43:01 -04:00
MaxsTechReview 0340be5a7d Steam client download: resumes, retries, shows megabytes, and isn't paused out of sight (#587)
A user's first Steam setup never finished on an Honor Pad (Snapdragon 6
Gen 1, metered Wi-Fi with a 14 Mbit/s link). Their log stops at
`downloading Steam: bins_linuxarm64_linuxarm64 (2/17)`. The log was
shared 31 s into the session, so nothing had failed yet. The way the
download works explains why it never got further.

## Why

- **The session pauses the download.** With the default background
behavior (Manual), a session pauses as soon as DroidDeck is hidden or
the screen turns off. In the user's log that happened one second after
start, before the download had begun.
- "Keep downloads running" doesn't help. It's off by default, and it
only recognizes Steam's own game downloads, not the runtime fetching the
client.
- A first run shows nothing for minutes, so leaving the app or letting
the screen time out is exactly what people do.
- **The download is bigger than the screen says.** The client is about
360 MB in 17 parts, and part 2 alone is 110 MB. The loading screen said
"~1-2 min" and counted parts, not bytes, so on a slow link it looks
frozen at about 5%.
- **Nothing handles a bad connection.** Each part was a plain `curl
-fsSL` with no timeout, no retry and no progress output.
  - A stalled connection could wait forever.
- A failed one ended the session and threw away everything fetched, so
the next try started from part 1.

## What changes

- **`droiddeck-steam-install`**:
- Parts go to `package/droiddeck-download` and stay there until the
client is installed.
  - Each part resumes from the bytes on disk (`curl -C -`).
  - A transfer under 4 KB/s for a minute is dropped and resumed.
- Attempts that make no progress back off 5–30 s, then give up with a
clear error. Whatever arrived is kept.
  - A finished part is marked unpacked, so a later session skips it.
- Bytes that fail their checksum are fetched whole again. A server that
won't resume starts the part over. Parts left over from another client
version are dropped.
- A repair redoes every part once, and keeps what an interrupted repair
already redid.
- Every 5 s the step line also carries megabytes overall: `… (2/17), 45
of 360 MB`.
- **Loading screen** (`LoadingState`): shows `Downloading the Steam
client · 45 / 360 MB`, with the bar following the bytes. Older runtimes
still show the part name.
- **`SessionService`**: until the session's first frame, the download
watch also checks whether the client is still to be fetched (no install
stamp or client binary, or a repair requested). While it is:
- the session stays awake with its wake and Wi-Fi locks, whatever the
downloads setting says;
  - the notification shows "Steam download active" with **Suspend now**;
  - a `steam.client_download` event marks start and end.
- The watch is set up before the first suspend decision, so a session
started with the screen off isn't paused first. Once the client is
there, the chosen behavior applies again.
- **`droiddeck-session`**: the step no longer promises 1–2 minutes.

## Testing

- `python3 -m unittest discover -s tools/tests -p 'test_*.py'`: 497 OK,
1 skipped.
- `test_steam_install.py` gets a fake curl that drops connections,
refuses to resume or fails on cue.
- New cases cover: megabyte progress; resuming after a dropped
connection; keeping finished and partial parts across a failed session;
restarting a part the server won't resume; re-fetching bad bytes;
dropping another version's parts; a repair resumed once.
- Kotlin: `:app:testDebugUnitTest --tests
'com.droiddeck.launcher.session.*'` passes. That includes the new
`LoadingStateTest` and the `clientInstallPending` cases in
`SteamDownloadMonitorTest`.
- RedMagic NP06J, with Background behavior set to Manual and "Keep
downloads running" off (the user's settings), and a client repair
requested:
- Started Steam and sent DroidDeck to the background straight away. The
client download stayed active for 30 s with no pause, and all 360 MB
arrived. Then the session paused as Manual should.
- Stopped a second repair during part 2. Part 1 was marked unpacked, the
110 MB part 2 was kept, and the repair marker stayed.
- Started again: it resumed at 110 of 360 MB, unpacked part 2 without
fetching it again, finished, and Steam started normally.
- The loading screen showed `Downloading the Steam client · 248 / 360
MB`.

Retested 2026-10-23 on the REDMAGIC tablet after merging main, which
brought in the stable-channel fallback. Background behavior was set to
Manual with "Keep downloads running" off. A client repair was forced,
then Home was pressed right after Start:

| Build | Session while hidden | Client download |
|---|---|---|
| main | suspended 1 s after Home | stuck at part 3/17 after 90 s |
| this PR | kept awake until the client was in | all 17 parts, 362 MB in
46 s with MB progress; suspended once done |

`tools.tests.test_steam_install` (22), `SteamDownloadMonitorTest` and
`LoadingStateTest` pass.

## Not changed

- **The download is the plain zips.** Valve's manifest also lists `.vz`
(LZMA) versions, 228 MB instead of 360 MB. Using them needs a VZ decoder
in the guest; that could be a follow-up.
- **The stall and retry path** is covered by the tests but wasn't forced
on the device.
2026-10-23 23:28:05 -04:00
maxjivi05 132f40ac97 fix: install step progress cannot overflow 2026-10-23 23:11:34 -04:00
maxjivi05 5981bbb1e2 fix: an oversized install step count falls back to the step text 2026-10-23 22:51:28 -04:00
maxjivi05 7419ba114f fix: the client install stays awake until Steam's first frame and its end is recorded 2026-10-23 22:22:14 -04:00
MaxsTechReview 116d0e4f1a fix: a game started from another launcher returns to that launcher (#1438)
A game started from another launcher (ES-DE, a `droiddeck://game/` link,
a pinned shortcut) should send the player back to that launcher when it
closes. Instead DroidDeck always opened the Android home screen, which
is a different place whenever that launcher isn't the default Home app.

## Cause
The link is handled by `FrontEndActivity`, which starts the session in
its own task and stays behind it in a second DroidDeck task. When the
session ended, DroidDeck started `CATEGORY_HOME` to cover that leftover
task, so the player always landed on the system Home.

## Fix
When a link-launched session ends, `SessionActivity` removes the task
that handled the link (a `FrontEndActivity` task whose base intent is a
game link), then removes its own task. Android then shows whatever was
underneath: ES-DE, any other app that launched the game, or Home when
Home launched it.

Sessions started from DroidDeck's own Games tab are unchanged.

## Testing
On a REDMAGIC NP06J, with the `droiddeck://game/945360` link started
using ES-DE's flags (`NEW_TASK | CLEAR_TASK`). Among Us was closed and
"Return to launcher after a game" was on.

| Launched from | main | this branch |
|---|---|---|
| Calculator (stands in for a non-Home frontend) | Android home |
Calculator |
| Home | Android home | Android home |
| DroidDeck's Games tab | DroidDeck | DroidDeck |

## Credits
Reported on Discord by whatsupbros.
2026-10-23 20:45:32 -04:00
maxjivi05 4a57c6f6f1 chore: drop the comments this change added 2026-10-23 20:45:14 -04:00
maxjivi05 2d9d4dbeb0 Merge branch 'main' into fix/steam-client-download
# Conflicts:
#	tools/linuxfs/overlay/usr/local/bin/droiddeck-steam-install
#	tools/tests/test_steam_install.py
2026-10-23 20:23:44 -04:00
maxjivi05 a0c6545e77 fix: a link that joins DroidDeck's open task still returns to the caller 2026-10-23 20:22:05 -04:00
maxjivi05 c7bcd82782 fix: Snapdragon 8 Gen 2 no longer pins Steam to one core by default 2026-10-23 19:47:41 -04:00
maxjivi05 caaf6b3a01 fix: keep the link task id when an in-app resume reaches the session 2026-10-23 19:47:32 -04:00
maxjivi05 69b2b8fcdf fix: only close the game-link task that started the session 2026-10-23 19:17:08 -04:00
maxjivi05 7261b26264 fix: a game started from another launcher returns to that launcher 2026-10-23 18:56:44 -04:00
maxjivi05 db9aae7b87 fix: delete the package before saving, and match Protons the same way 2026-10-23 18:56:24 -04:00
maxjivi05 94b530ecb2 fix: stored components can be deleted after a Proton update 2026-10-23 18:28:35 -04:00
maxjivi05 992912fdf8 fix: other apps can record game audio 2026-10-23 17:43:44 -04:00
MaxsTechReview 4ff4be2371 fix: keep a slightly early vsync tick for frame pacing (#1405)
Gamescope paces from the presentation timestamp. The Choreographer tick
is often still slightly ahead of the render thread, and that tick was
thrown away, so the draw time was reported instead.

A tick within one refresh of the draw is now kept, whether it is
slightly ahead or slightly behind. A missing tick, or one more than a
refresh away, still falls back to the draw time.
2026-10-23 17:00:32 -04:00
Kurt Himebauch 61482ffee7 perf(ci): reuse verified build inputs
## Summary

Follow-up to #1412, focused on repeat work observed in completed CI
logs.

- Give APK compilation and Android tests the same build-label and
API-key inputs. Previously `generateReleaseBuildConfig` changed between
invocations and caused Kotlin/Java recompilation before tests.
- Continue executing the Android test suite on every CI run: disable
reuse of the test task's results in CI while retaining compilation
caching. Keep the existing heap and process isolation settings.
- Cache the four remote checksum-pinned model/AAR downloads used by
add-on packs. Hash-check every cache hit, atomically replace corrupt
entries only after a verified download, and verify the destination
bytes. Packs are still regenerated from current prepared components and
their checksums remain part of the APK.
- Bound download time and retry connection resets as well as HTTP
errors; keep checksum mismatches fatal.
- Include the pack recipes in runtime source freshness checks.
- Restrict PR report updates to the workflow bot's own marker comment,
addressing the minor extraction review finding from #1412.

## Evidence and validation

The previous successful run explicitly recorded changed
`BuildConfig.java` as the reason for repeated Kotlin/Java compilation
between assembly and tests. The final #1412 normal run spent 38 seconds
building/downloading packs, 5m59s building the APK, and 3m03s in the
Android test step. A forced run also hit a connection reset while
fetching pack inputs and passed on retry.

Focused tests cover cold/warm input reuse, corrupt-cache repair,
changed-pin invalidation, rejected download bytes, actual local HTTP
connection failure/retry, identical build metadata and bot-only comment
selection. Actionlint passed. The clean Linux full suite passed 1,149
tests (18 skipped) before the final Gradle test-execution policy block.
Host, APK/ABI and Android checks are green on the final head, including
the ordinary PR run against the updated main branch and both controlled
branch attempts below.

Validation will compare cold and warm runs on the same commit, confirm
cache-hit hash verification, confirm app compilation remains up-to-date
in the test invocation, and confirm the Android suite actually executes
on both runs. Source-tree, preload, APK/ABI, all-test, signing and
publishing gates remain.

Existing Gradle, CMake and Cargo caching remains. Native compilation was
tens of seconds in the inspected baseline; no speculative native cache
changes or unmeasured speed claim are included. Android test execution
itself was about 2m30s; this PR removes avoidable setup/recompilation
rather than weakening coverage or changing concurrency.

## Follow-up correction

The first CI run failed two artwork-key tests that assumed the
compiled-in API key was empty. Keeping build inputs consistent exposed
that assumption. Tests now check the configured build key and its
trimmed fallback after clearing a user key; existing pure-function
empty-key and precedence coverage remains. No key values are included in
assertion messages. The failed run confirmed BuildConfig generation and
Kotlin/Java compilation were UP-TO-DATE in the test invocation;
subsequent full runs passed. The merged #1412 main run40007656633 passed
signing, pack publishing, preview announcement and catalog updates.

## Gradle cache audit fixes

Completed-run logs showed setup-gradle was read-only on PRs.
Same-repository PRs and branch runs can now persist their own cache
entries; fork PRs remain read-only. GitHub cache ref isolation is
retained. See [Gradle cache
guidance](https://github.com/gradle/actions/blob/main/docs/setup-gradle.md)
and [GitHub cache
scopes](https://docs.github.com/en/actions/reference/workflows-and-actions/dependency-caching).
PR caches consume additional shared quota; no claim of reuse between
different PR refs is made.

COMMIT_TIME now comes from SOURCE_COMMIT, matching the provenance field
and avoiding timestamp changes from temporary PR merge commits. The
source tree and pack validation remain unchanged. This enables a
meaningful same-commit warm check. Per-commit generated Java metadata
may still cause cross-commit recompilation; moving metadata into
packaged data is a separate candidate, not part of this patch. Head
9835c4e3 passed full CI and the same-commit warm rerun below.

## Confirmed final results

Both controlled runs use exactly
9835c4e3e1 and workflow_dispatch, avoiding
PR merge-tree changes between attempts.

| Measurement | First attempt | Warm attempt | Reduction |
| --- | ---: | ---: | ---: |
| Request to required gate green | 13m09s | 7m11s | 5m58s (45%) |
| Gradle APK build step | 6m11s | 2m08s | 4m03s (66%) |
| Android test step | 2m57s | 2m08s | 49s |
| Add-on pack build step | 40s | 31s | 9s |

[First
attempt](https://github.com/Droid-Deck/DroidDeck/actions/runs/40015059221/attempts/1)
and [warm
attempt](https://github.com/Droid-Deck/DroidDeck/actions/runs/40015059221/attempts/2)
both passed. Initial queue delay was 3 seconds in each. This is one
observed pair, not a guaranteed latency or isolated attribution of every
timing difference to cache changes. The first attempt already had some
dependency/build cache reuse.

Warm logs confirm the exact head's Gradle-home cache restored; release
Kotlin and Java compilation were FROM-CACHE during assembly, then
UP-TO-DATE during tests. Assembly reported 22 tasks FROM-CACHE. The
Android test executor actually ran; test results were neither cached nor
skipped. All four remote pack input hits were hash-verified; packs were
regenerated. Runtime inputs and APK native exports/dependencies
verified. The ordinary [PR
workflow](https://github.com/Droid-Deck/DroidDeck/actions/runs/40013127993/attempts/2)
also passed against the updated main merge tree.

A fresh Gradle daemon and "no cached configuration" remain expected on a
new runner: configuration/task-graph persistence is distinct from
build-output caching and currently has no configured encryption key. The
measured Kotlin/Java FROM-CACHE results demonstrate output reuse despite
that message.

This proves same-commit warm reuse. New pushes can still invalidate
compilation through changed source, toolchain, generated provenance or
pack metadata, and cache eviction/ref scope can cause misses.
Cross-commit compilation reuse is not established by this pair. Branch
validation intentionally skips main-only publishing; merged #1412's main
run40007656633 separately passed signing and publication. The user
authorized merging after the final cache-recovery correction passed.

## Final cache-recovery correction

Commit c95e1147 addresses the remaining minor review finding: after a
download passes its SHA check, an unexpected directory at the exact
hash-named cache entry is removed before the verified file is promoted.
Invalid downloads leave that directory untouched. A regression test
covers failed verification, successful repair and subsequent warm reuse.
All five focused download tests passed in a clean Linux export. All
final-head PR checks passed in
[run40018498031](https://github.com/Droid-Deck/DroidDeck/actions/runs/40018498031):
host suite ran 1,164 tests (49 platform skips), APK build and native
validation passed, Android tests executed successfully, and the
source/build validation status is green. Benchmark results above remain
measurements of 9835c4e3.
2026-10-23 16:34:50 -04:00
MaxsTechReview 92ef850999 fix: Steam starts on ROMs that refuse the Android ID read (#1421)
On a Redmi K70 (Android 14), tapping Steam drops straight back to the
main screen. The session thread dies while writing the guest network
link:

```
java.lang.IllegalArgumentException: Illegal char: - at offset 8
at android.provider.Settings$Secure.getStringForUser(Settings.java:6486)
at com.droiddeck.launcher.runtime.LinuxNetworkLinkComponent.describe(LinuxNetworkLinkComponent.kt:281)
```

The guest MAC is seeded from `ANDROID_ID`. That ROM throws from the
read, which had no guard. A failed read now uses the same fallback seed
as a missing ID. Devices that can read it keep the MAC they have now.

- `SeededMacTest`: stable MAC from an ID, and a read that throws the
ROM's exception falls back instead of crashing.
- REDMAGIC NP06J: the guest MAC is the same before and after the change
(`16:4e:5b:53:e9:ef`).

## Credits
Reported by lc666 on Discord.
2026-10-23 16:08:53 -04:00
MaxsTechReview 64b94f7cdb fix: the app follows automatic landscape rotation again unless an orientation was picked (#1420)
Fixes #1348, fixes #1372, fixes #1388.

Since 0.6.0 the app ignores the sensor and stays in one landscape side,
even with auto-rotate on. #1287 changed the default orientation from
sensor landscape to fixed landscape, so everyone who never picked an
orientation in settings got locked.

The default is sensor landscape again. A picked orientation is still
saved and used as before.

| | Before | After |
| --- | --- | --- |
| No orientation picked | `SCREEN_ORIENTATION_LANDSCAPE`, no flip |
`SCREEN_ORIENTATION_SENSOR_LANDSCAPE`, follows the sensor |
| Orientation picked in settings | that orientation | that orientation |

Tested on REDMAGIC NP06J (Android 16) with no saved orientation:
`dumpsys activity` reports the requested orientation, and the app flips
with the device.
2026-10-23 16:08:02 -04:00
maxjivi05 06946fe8ae chore: drop comments 2026-10-23 16:01:41 -04:00
maxjivi05 b6ca673d6e fix: Steam starts on ROMs that refuse the Android ID read 2026-10-23 15:50:16 -04:00
maxjivi05 6378c3f3d2 fix: the app follows automatic landscape rotation again unless an orientation was picked 2026-10-23 15:48:03 -04:00
Kurt Himebauch 2d59f03e7b fix: preserve GPU stats while pinning clocks (#1416)
Enabling GPU clock pinning froze Thor's kernel utilization counters,
leaving the HUD stuck at an arbitrary value (10–11% in the reproduced
sessions). KGSL_PROP_PWRCTRL disables both the governor and its
utilization sampling.

Use a dedicated KGSL context with a maximum-power constraint, renewed by
a small NOP submission every 25 ms. Keep devfreq running, bound the
queue to one submission, and release the context, mapped command buffer
and GPU allocation on stop or failure. Preserve recovery of legacy
global pins and report native setup failures instead of claiming
success.

Validation:
- AYN Thor: reproduced frozen counters with the old pin. The replacement
native implementation, run as a standalone ARM64 harness alongside the
installed Steam/Bulletstorm session, kept 680 MHz while utilization
changed from 17% in Steam to about 4% loading and 44% at the animated
title. A final-source smoke run also returned changing 60–66% samples at
680 MHz.
- Native lifecycle regression passed on Linux and as an ARM64 executable
on Thor: setup failure cleanup, renewal, bounded queue, runtime failure,
restart, and type-3/type-7 NOP selection for older/newer Adreno cores.
- Android native build and Kotlin compilation passed locally (NDK
27.0.12077973).
- CI build, Android tests, native regressions, and signing passed.
Installed the release-signed PR APK on Thor with app data preserved.
With GPU pin enabled, the normal Steam/Bulletstorm session held 680 MHz
while the HUD changed from 3% loading to 53% in the animated game scene
at 60 FPS; raw gpubusy samples also continued changing. This checks
utilization reporting, not long-term gameplay stability.

The earlier session exit was recorded by Android as LOW_MEMORY; this PR
addresses the utilization freeze.
2026-10-23 15:41:12 -04:00
XxJSONDeruloxX a7d6279ee3 test: check artwork key normalization 2026-10-23 15:22:12 -04:00
XxJSONDeruloxX 5a806b35f1 test: support configured artwork key 2026-10-23 15:22:01 -04:00
MaxsTechReview 8f0bd51fd8 Merge pull request #1401 from Droid-Deck/perf/ntsync-fast-paths
## Summary

droiddeck-ntsync (the userspace stand-in for the Linux ntsync driver,
Wine sync = ntsync) was slower than droiddeck-esync on every cheap
operation. On the tablet a `SetEvent` cost 0.94 µs against esync's 0.21
µs, and a wait on a signaled event 0.45 µs against 0.08 µs. Four things
made it slow:

- **Signal masking:** every operation blocked and restored the signal
mask, so each call made 2 `rt_sigprocmask` syscalls, and a blocking wait
made 4. The masking itself is needed: if Wine's suspend signal stopped a
thread while it held an object lock, a GC thread that suspended it and
then touched the same object would deadlock.
- **Queueing for objects that were already free:** a wait linked itself
onto every object and unlinked again even when the first object could
simply be taken.
- **Wakes nobody needed:** a futex wake syscall went out for every
handoff, even when the waiter was still spinning and would have seen the
change anyway.
- **One shared counter:** every call took and dropped a reference on a
single process-wide device counter, so threads working on unrelated
objects bounced one cache line between cores.

### Changes

- **Signals are deferred instead of masked:** handlers installed by
Wine's `ntdll.so` run behind a small trampoline. If such a signal
arrives inside a critical section (a per-thread depth count), it is
saved with its `siginfo` and sent again with `rt_tgsigqueueinfo`
(falling back to `tgkill`) as soon as the section ends.
  - `sigaction` queries still return Wine's own handler.
- Synchronous signals (SEGV, BUS, ILL, FPE, TRAP, SYS) are never
wrapped, as before.
  - Handlers installed by anything other than `ntdll.so` are left alone.
- **Wait-any fast path:** an object that is signaled and has nobody
queued on it is taken directly, in index order. A wait whose timeout has
already passed returns `ETIMEDOUT` without building a queue.
- **Short spin before sleeping:** a waiter spins for up to 4 µs, bounded
by time rather than by iteration count, before it sleeps on the futex.
- **Wake only sleeping waiters:** the waiter publishes a `sleeping`
flag, and the waker only calls `FUTEX_WAKE` when it is set. Both sides
use seq_cst ordering, so a wake can't be lost. The queue layout changed,
so `NS_VERSION` is now 2.
- **Per-thread device reference:** a thread takes one reference on the
device the first time it uses it and keeps it until it exits or moves to
another device, instead of two atomics on the shared counter per call.
- **More than four signals in one section:** a fifth distinct signal is
held pending with the thread's signal mask, then released once the
section ends. Its handler never runs while an object lock is held.
- **Replay:** deferred signals are replayed with every signal blocked,
so each one is sent exactly once. A held signal is released at the same
signal-nesting level that held it. Handlers installed with
`SA_RESETHAND` are not wrapped, so a one-shot handler keeps its kernel
behaviour.
- **Waits from a handler during another wait's sleep:** they get their
own queue and, for a second device, a reference for that call only. The
outer wait keeps its queue and can't pick up the inner wait's result.

### ntsync becomes the default

- **New installs:** start on ntsync.
- **Existing installs:** move to ntsync once, on the first start of this
version, whatever sync they were on before (a one-time `ntsyncDefaultV1`
marker). Anything picked after that is kept, including esync, fsync and
wineserver.
- **Wording:** the ntsync switch in Performance no longer says
"experimental" or "for testing", in all 33 languages.
- **Unchanged:** per-game sync choices on a game's settings page still
win over the app-wide choice. Proton builds without ntsync still fall
back to droiddeck-esync, then droiddeck-fsync.
- **Tests:** `SyncBackendTest` covers a fresh install, each earlier
choice moving to ntsync once, and a later choice surviving the next
start. All 2,160 unit tests pass.

## Performance (REDMAGIC tablet)

Windows microbenchmark (`syncbench.exe`, x86-64 under FEX), µs per
operation:

| | esync | ntsync before | ntsync after |
|---|---|---|---|
| SetEvent | 0.57 / 0.70 | 0.51 / 0.65 | 0.38 / 0.35 |
| wait on signaled event | 0.13 / 0.13 | 0.44 / 0.43 | 0.27 / 0.20 |
| wait, timeout 0 | 0.27 / 0.25 | 0.52 / 0.51 | 0.19 / 0.18 |
| semaphore release + wait | 0.72 / 0.68 | 0.66 / 0.63 | 0.12 / 0.16 |
| mutex wait + release | 0.77 / 0.73 | 0.66 / 0.64 | 0.12 / 0.12 |
| wait on 8 objects | 0.10 / 0.09 | 0.59 / 0.57 | 0.15 / 0.15 |
| alertable wait | 0.06 / 0.05 | 0.43 / 0.42 | 0.07 / 0.07 |
| event ping-pong, 2 threads | 25.6 / 24.5 | 24.4 / 29.3 | 1.97 / 1.66 |
| 4 threads on one semaphore | 9.4 / 4.6 | 20.3 / 19.1 | 7.3 / 7.6 |

The backend each run actually used was checked from its file
descriptors:
- **esync:** wineserver held eventfds.
- **ntsync:** wineserver held the `droiddeck-ntsync` device and its
timerfd objects.

Native arm64 benchmark against the device directly, µs (before / after):
- set: 0.31–0.36 → 0.06–0.08
- signaled wait: 0.37–0.45 → 0.04–0.08
- semaphore pair: 0.56–0.58 → 0.06–0.07
- mutex pair: 0.56–0.61 → 0.04
- ping-pong: 15–30 → 1.7–2.5

Removing the shared device counter on its own takes 4 threads working on
their own objects from 0.21 to 0.04 µs per operation (x86-64 host).

Games, default Steam launch, The Forest and PEAK at the same scene each
run, 20 s pacing. Columns are esync, ntsync before and ntsync after;
each cell gives two runs.

| | esync | ntsync before | ntsync after |
|---|---|---|---|
| The Forest Low, fps | 64.5 / 61.9 | 70.1 / 74.8 | 72.1 / 72.1 |
| The Forest Low, CPU | 489 / 511% | 434 / 419% | 436 / 426% |
| The Forest High, fps | 58.3 / 57.8 | 72.0 / 69.8 | 72.5 / 69.9 |
| The Forest High, CPU | 506 / 504% | 417 / 429% | 425 / 435% |
| PEAK, fps | 50.4 / 48.6 | — | 47.0 |
| RDR2, launch to signed in / game process | — | — | 29.2 / 42.0 s and
28.3 / 41.1 s |

Two more games, esync against ntsync after, three runs each, default
settings:

| | esync | ntsync after |
|---|---|---|
| Among Us, menu fps (59 cap) | 59.0 / 58.9 / 59.3 | 59.0 / 59.0 / 58.7
|
| Among Us, launch to main menu | 6.65 / 6.84 / 6.85 s | 6.39 / 6.52 /
6.64 s |
| Among Us, CPU | 123 / 132 / 133% | 133 / 133 / 124% |
| Mortal Kombat X attract fight, fps | 56.7 / 54.1 / 56.3 | 61.1 / 56.9
/ 56.2 |
| Mortal Kombat X, 1% low | 20.0 / 30.3 / 30.3 | 30.3 / 30.3 / 30.3 |
| Mortal Kombat X, launch to window | 7.75 / 7.79 / 7.75 s | 7.81 / 7.71
/ 7.63 s |

Mortal Kombat X's CPU moves between 140% and 275% in both modes because
the attract mode changes fighters and stage, so it isn't compared.

The Forest's 1% low was 41.7 fps in every run. PEAK is GPU-bound and
stays in run-to-run noise. Its second ntsync run stopped at the title
menu because the rig's menu input didn't register; the game was still
running at 87 fps, and the same rig miss happened on esync before. RDR2
signs in and reaches the game on the new code.

## Tests

- **`tools/tests/test_ntsync_preload.py` (new):** builds the preload and
checks the emulation against the semantics documented for the kernel
driver:
  - semaphore counts and `EOVERFLOW`;
- mutex owners, recursion, `EPERM`, kill/abandon with `EOWNERDEAD` on
read and on the next wait;
  - auto and manual events, set/reset/pulse previous state;
- wait-any index order, duplicates, the alert object, `EINVAL` limits,
timeouts;
  - wait-all atomicity, duplicates and alert;
- wakes across threads and across processes (inherited fds and fds sent
with `SCM_RIGHTS`);
- 20,000-round ping-pong and 4-thread semaphore stress with infinite
timeouts;
- two signal cases. In one, a handler that re-enters the same object is
fired 2,000 times at a busy thread. In the other, a handler parks the
thread while another thread uses the same object, as Wine's suspend does
under a GC.
- eight real-time signals fired in bursts at a thread busy on one
object, each handler touching that object, so more signals arrive inside
a section than can be deferred;
- a handler that waits on the same device and on a second device while
its thread sleeps in another wait;
- a one-shot (`SA_RESETHAND`) handler installed and fired 300 times at a
busy thread, every one of which must run.
- **The signal tests catch a broken implementation:** with signal
protection removed, the program hangs and the test fails. Before the
last fix, the burst test hung and the outer wait returned the nested
wait's result.
- **The suite passes:** 8 of 8 runs natively on arm64 under proot on the
tablet, and 40 of 40 on the host.
- **On the final build:** The Forest Low ran at 77.7 fps, PEAK at 47.5
fps, and Among Us and two Mortal Kombat X runs played through.
- **The new paths stay out of the way in games:** a build that logged
them saw about 1,400 Wine signals land during ntsync sleeps over two
Mortal Kombat X runs and one The Forest run. None of those handlers
waited on ntsync and no signal was held back, so games run the same code
as before the last fix. The Forest Low ran at 71.6 fps and PEAK at 47.2
fps on the final build.
- **Existing preload tests still pass:** fsync, openat2, usr2 (also with
`BL_SYNC=1`), tracer, dircache, noexec, binfmt, faultreport and esync
packs.

## Credits

- [Linux](https://github.com/torvalds/linux)
`Documentation/userspace-api/ntsync.rst` (GPL-2.0): the object, wait and
error semantics the new test checks. Also added to the README credits.
2026-10-23 15:10:28 -04:00
XxJSONDeruloxX 9c66a1ac0a fix: select NOP packets for older Adreno cores 2026-10-23 15:04:50 -04:00
XxJSONDeruloxX 4d86839910 fix: preserve GPU stats while pinning clocks 2026-10-23 15:00:11 -04:00
maxjivi05 293e84a64c feat: droiddeck-ntsync is the default Wine sync
- New installs start on ntsync.
- Existing installs move to ntsync once on the first start of this version, whatever sync they used before. A choice made after that is kept.
- The ntsync switch is no longer labelled experimental or for testing, in every language.
2026-10-23 13:57:38 -04:00
Kurt Himebauch 16bedf3635 fix: keep the UBWC hint consistent after custom guest environment 2026-10-23 13:54:01 -04:00
Kurt Himebauch e06b942e67 fix: warn about UBWC restart only after a hint is active 2026-10-23 13:54:01 -04:00
Kurt Himebauch 2e1156038e test: cover the inclusive A740 blob 700 boundary 2026-10-23 13:54:01 -04:00