mirror of
https://git.eden-emu.dev/eden-emu/eden
synced 2026-05-13 08:48:38 +02:00
[windows] Return x86 microsleep for Windows (#3563)
Microsleep removal may have regressed AOC in msvc Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3563 Reviewed-by: CamilleLaVey <camillelavey99@gmail.com> Reviewed-by: Lizzie <lizzie@eden-emu.dev> Co-authored-by: John <john@eden-emu.dev> Co-committed-by: John <john@eden-emu.dev>
This commit is contained in:
parent
86e02aceba
commit
5fb3ae487c
4 changed files with 98 additions and 28 deletions
|
|
@ -186,6 +186,8 @@ if(ARCHITECTURE_x86_64)
|
||||||
common
|
common
|
||||||
PRIVATE x64/cpu_detect.cpp
|
PRIVATE x64/cpu_detect.cpp
|
||||||
x64/cpu_detect.h
|
x64/cpu_detect.h
|
||||||
|
x64/cpu_wait.cpp
|
||||||
|
x64/cpu_wait.h
|
||||||
x64/native_clock.cpp
|
x64/native_clock.cpp
|
||||||
x64/native_clock.h
|
x64/native_clock.h
|
||||||
x64/rdtsc.cpp
|
x64/rdtsc.cpp
|
||||||
|
|
|
||||||
75
src/common/x64/cpu_wait.cpp
Normal file
75
src/common/x64/cpu_wait.cpp
Normal file
|
|
@ -0,0 +1,75 @@
|
||||||
|
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
|
||||||
|
// SPDX-License-Identifier: GPL-3.0-or-later
|
||||||
|
|
||||||
|
#include <thread>
|
||||||
|
|
||||||
|
#ifdef _MSC_VER
|
||||||
|
#include <intrin.h>
|
||||||
|
#endif
|
||||||
|
|
||||||
|
#include "common/x64/cpu_detect.h"
|
||||||
|
#include "common/x64/cpu_wait.h"
|
||||||
|
#include "common/x64/rdtsc.h"
|
||||||
|
|
||||||
|
namespace Common::X64 {
|
||||||
|
|
||||||
|
namespace {
|
||||||
|
|
||||||
|
// 100,000 cycles is a reasonable amount of time to wait to save on CPU resources.
|
||||||
|
// For reference:
|
||||||
|
// At 1 GHz, 100K cycles is 100us
|
||||||
|
// At 2 GHz, 100K cycles is 50us
|
||||||
|
// At 4 GHz, 100K cycles is 25us
|
||||||
|
constexpr auto PauseCycles = 100'000U;
|
||||||
|
|
||||||
|
} // Anonymous namespace
|
||||||
|
|
||||||
|
#if defined(_MSC_VER) && !defined(__clang__)
|
||||||
|
__forceinline static void TPAUSE() {
|
||||||
|
static constexpr auto RequestC02State = 0U;
|
||||||
|
_tpause(RequestC02State, FencedRDTSC() + PauseCycles);
|
||||||
|
}
|
||||||
|
|
||||||
|
__forceinline static void MWAITX() {
|
||||||
|
static constexpr auto EnableWaitTimeFlag = 1U << 1;
|
||||||
|
static constexpr auto RequestC1State = 0U;
|
||||||
|
|
||||||
|
// monitor_var should be aligned to a cache line.
|
||||||
|
alignas(64) u64 monitor_var{};
|
||||||
|
_mm_monitorx(&monitor_var, 0, 0);
|
||||||
|
_mm_mwaitx(EnableWaitTimeFlag, RequestC1State, PauseCycles);
|
||||||
|
}
|
||||||
|
#else
|
||||||
|
static void TPAUSE() {
|
||||||
|
static constexpr auto RequestC02State = 0U;
|
||||||
|
const auto tsc = FencedRDTSC() + PauseCycles;
|
||||||
|
const auto eax = static_cast<u32>(tsc & 0xFFFFFFFF);
|
||||||
|
const auto edx = static_cast<u32>(tsc >> 32);
|
||||||
|
asm volatile("tpause %0" : : "r"(RequestC02State), "d"(edx), "a"(eax));
|
||||||
|
}
|
||||||
|
|
||||||
|
static void MWAITX() {
|
||||||
|
static constexpr auto EnableWaitTimeFlag = 1U << 1;
|
||||||
|
static constexpr auto RequestC1State = 0U;
|
||||||
|
|
||||||
|
// monitor_var should be aligned to a cache line.
|
||||||
|
alignas(64) u64 monitor_var{};
|
||||||
|
asm volatile("monitorx" : : "a"(&monitor_var), "c"(0), "d"(0));
|
||||||
|
asm volatile("mwaitx" : : "a"(RequestC1State), "b"(PauseCycles), "c"(EnableWaitTimeFlag));
|
||||||
|
}
|
||||||
|
#endif
|
||||||
|
|
||||||
|
void MicroSleep() {
|
||||||
|
static const bool has_waitpkg = GetCPUCaps().waitpkg;
|
||||||
|
static const bool has_monitorx = GetCPUCaps().monitorx;
|
||||||
|
|
||||||
|
if (has_waitpkg) {
|
||||||
|
TPAUSE();
|
||||||
|
} else if (has_monitorx) {
|
||||||
|
MWAITX();
|
||||||
|
} else {
|
||||||
|
std::this_thread::yield();
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
} // namespace Common::X64
|
||||||
10
src/common/x64/cpu_wait.h
Normal file
10
src/common/x64/cpu_wait.h
Normal file
|
|
@ -0,0 +1,10 @@
|
||||||
|
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
|
||||||
|
// SPDX-License-Identifier: GPL-3.0-or-later
|
||||||
|
|
||||||
|
#pragma once
|
||||||
|
|
||||||
|
namespace Common::X64 {
|
||||||
|
|
||||||
|
void MicroSleep();
|
||||||
|
|
||||||
|
} // namespace Common::X64
|
||||||
|
|
@ -13,9 +13,8 @@
|
||||||
#include "common/windows/timer_resolution.h"
|
#include "common/windows/timer_resolution.h"
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
#if defined(_WIN32) && defined(ARCHITECTURE_x86_64) && defined(__MINGW64__)
|
#ifdef ARCHITECTURE_x86_64
|
||||||
#include "common/x64/cpu_detect.h"
|
#include "common/x64/cpu_wait.h"
|
||||||
#include "common/x64/rdtsc.h"
|
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
#include "common/settings.h"
|
#include "common/settings.h"
|
||||||
|
|
@ -287,38 +286,22 @@ void CoreTiming::ThreadLoop() {
|
||||||
const auto next_time = Advance();
|
const auto next_time = Advance();
|
||||||
if (next_time) {
|
if (next_time) {
|
||||||
// There are more events left in the queue, wait until the next event.
|
// There are more events left in the queue, wait until the next event.
|
||||||
if (auto wait_time = *next_time - GetGlobalTimeNs().count(); wait_time > 0) {
|
auto wait_time = *next_time - GetGlobalTimeNs().count();
|
||||||
#if defined(_WIN32) && defined(ARCHITECTURE_x86_64) && defined(__MINGW64__)
|
if (wait_time > 0) {
|
||||||
|
#ifdef _WIN32
|
||||||
while (!paused && !event.IsSet() && wait_time > 0) {
|
while (!paused && !event.IsSet() && wait_time > 0) {
|
||||||
wait_time = *next_time - GetGlobalTimeNs().count();
|
wait_time = *next_time - GetGlobalTimeNs().count();
|
||||||
if (wait_time >= timer_resolution_ns) {
|
if (wait_time >= timer_resolution_ns) {
|
||||||
Common::Windows::SleepForOneTick();
|
Common::Windows::SleepForOneTick();
|
||||||
} else {
|
} else {
|
||||||
// 100,000 cycles is a reasonable amount of time to wait to save on CPU resources.
|
#ifdef ARCHITECTURE_x86_64
|
||||||
// For reference:
|
Common::X64::MicroSleep();
|
||||||
// At 1 GHz, 100K cycles is 100us
|
#else
|
||||||
// At 2 GHz, 100K cycles is 50us
|
std::this_thread::yield();
|
||||||
// At 4 GHz, 100K cycles is 25us
|
#endif
|
||||||
constexpr auto PauseCycles = 100'000U;
|
|
||||||
auto const& caps = Common::GetCPUCaps();
|
|
||||||
if (caps.waitpkg) {
|
|
||||||
static constexpr auto RequestC02State = 0U;
|
|
||||||
const auto tsc = Common::X64::FencedRDTSC() + PauseCycles;
|
|
||||||
const auto eax = u32(tsc & 0xFFFFFFFF);
|
|
||||||
const auto edx = u32(tsc >> 32);
|
|
||||||
asm volatile("tpause %0" : : "r"(RequestC02State), "d"(edx), "a"(eax));
|
|
||||||
} else if (caps.monitorx) {
|
|
||||||
static constexpr auto EnableWaitTimeFlag = 1U << 1;
|
|
||||||
static constexpr auto RequestC1State = 0U;
|
|
||||||
// monitor_var should be aligned to a cache line.
|
|
||||||
alignas(64) u64 monitor_var{};
|
|
||||||
asm volatile("monitorx" : : "a"(&monitor_var), "c"(0), "d"(0));
|
|
||||||
asm volatile("mwaitx" : : "a"(RequestC1State), "b"(PauseCycles), "c"(EnableWaitTimeFlag));
|
|
||||||
} else {
|
|
||||||
std::this_thread::yield();
|
|
||||||
}
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
if (event.IsSet()) {
|
if (event.IsSet()) {
|
||||||
event.Reset();
|
event.Reset();
|
||||||
}
|
}
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue