- Windows (`_WIN32`): replace full-duration waitable-timer sleep in `PreciseSleepHundredNanos` and the repeated `PreciseSleepUs` call loop in `PreciseSleepUntil` with a hybrid sleep-then-spin helper (`PreciseSleepFromCounterWithTimer`) that sleeps on the thread-local high-resolution waitable timer up to a 250 us spin tail, falls back to `std::this_thread::sleep_for` if timer setup/wait fails, and finishes the remaining window using `QueryPerformanceCounter` + `YieldProcessor()` (thresholded to 200 ns and ceiling-rounded via `PreciseHundredNanosToQpcTicks` for non-10MHz QPC frequencies).
- Linux & POSIX (`!_WIN32 && !__EMSCRIPTEN__`): replace repeated `nanosleep` loop in `PreciseSleepUs` and `PreciseSleepUntil` with `PreciseSleepFromNanos` using `clock_gettime(CLOCK_MONOTONIC)`. On Linux (`__linux__`), cache `PR_SET_TIMERSLACK` (`1 ns`) once per thread (`thread_local`) and sleep via `clock_nanosleep(CLOCK_MONOTONIC, TIMER_ABSTIME)` up to a 15 us spin tail (falling back to a conservative 100 us tail if `prctl` fails; 100 us `nanosleep` tail on other POSIX platforms) before finishing with `_mm_pause()` on x86 or `std::atomic_signal_fence` on non-x86.
- Tests: expand `core/libs/tests/PreciseWait.cpp` to assert that the 31-sample median of `PreciseSleepUs(50)` stays below 90 us (which fails on `main` at ~118 us on Linux and ~518 us on Windows) and verify ceiling QPC tick conversion and timer fallbacks.
* fix(agc): keep the driver's per-thread scratch out of thread_local teardown
With MinGW's emulated TLS, a thread_local with a destructor in a DLL is
destroyed after its storage is gone: when a std::thread ends, winpthread
runs libgcc's emutls key destructor, which frees the thread's thread_local
storage, and only then does the loader's DLL_THREAD_DETACH run the DLL's
tls_atexit list, so the destructors run on freed memory. The queue worker
touches four of them in libSceAgcDriver (the dispatch-ahead map, the
pending waits, the draw-snapshot compare vectors and the image decode
memo), and Driver::stop's join of the worker then corrupts the heap.
Under cdb's checking heap agc_driver_draw_cache_reuse stopped at the
worker's exit with heap corruption in all 127 runs, with graphics
pipeline libraries, with APS5_NO_PIPELINE_LTO=1 and with APS5_NO_GPL=1.
Without it the damage shows intermittently, as exit 0xc0000374 (2 of 143
runs with a cold NVIDIA shader cache) or as a fault in whatever allocates
next, such as NVIDIA's compiler on the optimized-link thread (#2980).
These objects now live in HostThreadLocal, fiber-local storage that owns
them, as the recorder's buffers do since e424b6b7b.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(libs): keep thread_local objects with destructors out of module teardown
The teardown order behind the driver's heap corruption holds in every
module: with MinGW's emulated TLS a thread's thread_local storage is freed
before the module's DLL_THREAD_DETACH runs its thread_local destructors,
so they run on freed objects. PreciseSleepUs's timer (used, among others,
by the video-out vblank thread) then closes a stale handle; libkernel's
directory path, socket waker and UUID random device, libSceRandom's random
device and libScePosixForWebKit's random device and host lookup free or
close stale objects. They move to HostThreadLocal. The other thread_locals
in the modules are trivially destructible or only built for Linux.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(libs): fail when a module registers a thread_local destructor
prx_thread_local_destructors lists the symbols of every built module and,
where MinGW's thread_local destructor support is linked in, disassembles
the module for calls to __cxa_thread_atexit outside libc's registration of
guest destructors. A module that keeps a thread_local with a destructor
fails it with the function that registers one: with main's libSceRandom
it reports sceRandomGetRandomNumber, and with this branch all 122 modules
pass, libc being the only one that links the support.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(agc): keep the image memo behind its pointer
The image decode memo went into HostThreadLocal as the whole array. On
Linux, where HostThreadLocal is a plain thread_local, that put 1024 memo
entries into every thread's TLS block instead of one pointer. It keeps
its unique_ptr as before, now owned by HostThreadLocal, so Linux builds
the same thread_local as main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Dean Galvin <dean.galvin@live.com>
On Windows, nanosleep and this_thread::sleep_for/sleep_until go
through winpthreads, which ends a wait on a 15.6 ms GetTickCount64
step. Blocking sceAudioOutOutput and sceAudioOut2ContextPush calls
returned in bursts about every 16 ms, a blocking sceAudio3dPortPush
ran at about 68 % of real time, and sceAudioInInput without a device
woke anywhere from 1 ms to 31 ms apart. They now sleep with
PreciseSleepUs or the new PreciseSleepUntil.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Comments in this codebase may only mark technical debt (per
docs/dev/CONVENTIONS.md), plus the #endif and namespace-close
structural exceptions. Removes the descriptive comments added on
this branch across PreciseWait.hpp, FormattingWide.cpp, the
direct-memory pool query/free-run plumbing, and the module loader
wiring, keeping the two allowed structural markers intact.
- Add FormattingWide.cpp implementing vswprintf_nid_postfix and
wprintf_nid_postfix, the only wide-char printf-family functions
upstream had left as unimplemented stubs (Formatting.cpp's checked
and unchecked narrow printf/scanf families are unaffected). Removes
the dead vswprintf_nid_postfix stub from Export.cpp.
- Add PreciseWait.hpp: std::condition_variable::wait_for cannot return
in under about one system tick on this MinGW toolchain regardless of
the requested budget, even with NtSetTimerResolution; PreciseSleepUs
uses a high-resolution waitable timer instead (~0.5ms vs ~15.6ms).
Header-only for now: its guest-side callers (Equeue/EventFlag/
Semaphore/AgcDriver wait loops) are out of scope for this change, so
it is not wired into them here.
hamb3r