nx2: run the guest counter at 31.25 MHz for Switch 2 titles

The Switch 2 counter runs at 31.25 MHz and its SDK converts ticks with
that constant: Mario Kart World's sdk builds 31,250,000 in
nn::os::GetSystemTickFrequency, TickManager::ConvertToTimeSpan and
TickManager::ConvertToTick (19.2 MHz appears nowhere). Eglantine counted
at the Switch 1 rate for every title, so every tick-based time in a
Switch 2 game was off by 1.63x: elapsed times read short, and a deadline
of one 60 Hz frame converted to ticks waited ~27 ms.

The wall clock now has a guest counter frequency, set at boot: 31.25 MHz
for Switch 2 programs, 19.2 MHz otherwise, EDEN_NX2_CNTFRQ=<Hz> to force
one. CNTPCT reads, CNTFRQ_EL0, svcGetSystemTick, thread tick counts and
the time services' tick conversions follow it, and the CNTPCT spin
detection keeps its ~1.7 us threshold.

Mario Kart World, title screen and attract demo, 180 s, same build:
frame times unchanged (17-19 ms once warm), process CPU time 848 s
instead of 930 s.

2 files changed, 45 insertions, 8 deletions

diff --git a/src/common/cpu_features.cpp b/src/common/cpu_features.cpp
index 52fd003c44..93eecb1dc5 100644
--- a/src/common/cpu_features.cpp
+++ b/src/common/cpu_features.cpp
@@ -278,8 +278,16 @@ std::chrono::milliseconds WallClock::GetTimeMS() const {
 
 s64 WallClock::GetCNTPCT() const {
     if (!invariant)
-        return GetUptime() * NsToCNTPCTRatio::num / NsToCNTPCTRatio::den;
-    return MultiplyHigh(GetUptime(), cntpct_rdtsc_factor);
+        return MultiplyAndDivide64(GetUptime(), GetGuestCNTFRQ(), NsRatio::den);
+    return MultiplyHigh(GetUptime(), cntpct_rdtsc_factor.load(std::memory_order_relaxed));
+}
+
+void WallClock::SetGuestCNTFRQ(u64 frequency) {
+    guest_cntfrq.store(frequency, std::memory_order_relaxed);
+    if (invariant) {
+        cntpct_rdtsc_factor.store(GetFixedPoint64Factor(frequency, rdtsc_frequency),
+                                  std::memory_order_relaxed);
+    }
 }
 
 s64 WallClock::GetGPUTick() const {
@@ -336,7 +344,7 @@ namespace {
 } // namespace
 
 WallClock::WallClock(bool invariant_, u64 rdtsc_frequency_) noexcept {
-    const u64 host_cntfrq = std::max<u64>(GetHostCNTFRQ(), 1);
+    host_cntfrq = std::max<u64>(GetHostCNTFRQ(), 1);
     ns_cntfrq_factor = GetFixedPointFactor(NsRatio::den, host_cntfrq);
     us_cntfrq_factor = GetFixedPointFactor(UsRatio::den, host_cntfrq);
     ms_cntfrq_factor = GetFixedPointFactor(MsRatio::den, host_cntfrq);
@@ -361,6 +369,11 @@ s64 WallClock::GetCNTPCT() const {
     return MultiplyHigh(GetUptime(), guest_cntfrq_factor);
 }
 
+void WallClock::SetGuestCNTFRQ(u64 frequency) {
+    guest_cntfrq.store(frequency, std::memory_order_relaxed);
+    guest_cntfrq_factor = GetFixedPointFactor(frequency, host_cntfrq);
+}
+
 s64 WallClock::GetGPUTick() const {
     return MultiplyHigh(GetUptime(), gputick_cntfrq_factor);
 }
@@ -399,7 +412,11 @@ std::chrono::milliseconds WallClock::GetTimeMS() const {
 }
 
 s64 WallClock::GetCNTPCT() const {
-    return GetUptime() * NsToCNTPCTRatio::num / NsToCNTPCTRatio::den;
+    return MultiplyAndDivide64(GetUptime(), GetGuestCNTFRQ(), NsRatio::den);
+}
+
+void WallClock::SetGuestCNTFRQ(u64 frequency) {
+    guest_cntfrq.store(frequency, std::memory_order_relaxed);
 }
 
 s64 WallClock::GetGPUTick() const {
@@ -419,9 +436,13 @@ u64 WallClock::NsToTicks(std::chrono::nanoseconds ns) const {
 }
 #endif
 
+u64 WallClock::CPUTickToGuestCNTPCT(u64 cpu_tick) const {
+    return MultiplyAndDivide64(cpu_tick, GetGuestCNTFRQ(), CPUTickFreq);
+}
+
 // Wall clock MUST be initialized AFTER g_cpu_caps
 // C++ only guarantees ctor init in the order they appear in TU
-const WallClock g_wall_clock = [] {
+WallClock g_wall_clock = [] {
 #if defined(ARCHITECTURE_x86_64)
     auto const& caps = Common::g_cpu_caps;
     return WallClock(caps.invariant_tsc && caps.tsc_frequency > std::nano::den, caps.tsc_frequency);
diff --git a/src/common/cpu_features.h b/src/common/cpu_features.h
index 61aa2767ab..31be2dc956 100644
--- a/src/common/cpu_features.h
+++ b/src/common/cpu_features.h
@@ -7,6 +7,7 @@
 
 #pragma once
 
+#include <atomic>
 #include <optional>
 #include <string_view>
 #include <chrono>
@@ -20,11 +21,24 @@ namespace Common {
 class WallClock {
 public:
     static constexpr u64 CNTFRQ = 19'200'000;         // CNTPCT_EL0 Frequency = 19.2 MHz
+    static constexpr u64 Switch2CNTFRQ = 31'250'000;  // Switch 2 CNTPCT_EL0 Frequency = 31.25 MHz
     static constexpr u64 GPUTickFreq = 614'400'000;   // GM20B GPU Tick Frequency = 614.4 MHz
     static constexpr u64 CPUTickFreq = 1'020'000'000; // T210/4 A57 CPU Tick Frequency = 1020.0 MHz
 
     explicit WallClock(bool invariant, u64 rdtsc_frequency_) noexcept;
 
+    /// Sets the guest counter's frequency: CNTFRQ, or Switch2CNTFRQ for Switch 2 titles (their SDK
+    /// converts ticks with a constant 31.25 MHz). Called at boot, before guest code runs.
+    void SetGuestCNTFRQ(u64 frequency);
+
+    /// @returns The guest counter's frequency (CNTFRQ_EL0).
+    [[nodiscard]] u64 GetGuestCNTFRQ() const {
+        return guest_cntfrq.load(std::memory_order_relaxed);
+    }
+
+    /// @returns Emulated CPU ticks (single core timing) as guest CNTPCT ticks.
+    [[nodiscard]] u64 CPUTickToGuestCNTPCT(u64 cpu_tick) const;
+
     /// @returns The time in nanoseconds since the construction of this clock.
     std::chrono::nanoseconds GetTimeNS() const;
 
@@ -34,7 +48,7 @@ public:
     /// @returns The time in milliseconds since the construction of this clock.
     std::chrono::milliseconds GetTimeMS() const;
 
-    /// @returns The guest CNTPCT ticks since the construction of this clock.
+    /// @returns The guest CNTPCT ticks (at GetGuestCNTFRQ()) since the construction of this clock.
     s64 GetCNTPCT() const;
 
     /// @returns The guest GPU ticks since the construction of this clock.
@@ -91,6 +105,7 @@ public:
     using CPUTickToCNTPCTRatio = std::ratio<CNTFRQ, CPUTickFreq>;
     using CPUTickToGPUTickRatio = std::ratio<GPUTickFreq, CPUTickFreq>;
 
+    std::atomic<u64> guest_cntfrq{CNTFRQ};
 #if defined(ARCHITECTURE_x86_64)
     u64 rdtsc_frequency;
     u64 ns_rdtsc_factor;
@@ -98,7 +113,7 @@ public:
     u64 ms_rdtsc_factor;
     u64 rdtsc_ns_integer;
     u64 rdtsc_ns_factor;
-    u64 cntpct_rdtsc_factor;
+    std::atomic<u64> cntpct_rdtsc_factor;
     u64 gputick_rdtsc_factor;
     bool invariant;
 #elif defined(HAS_NCE)
@@ -106,6 +121,7 @@ public:
     [[nodiscard]] inline FactorType GetGuestCNTFRQFactor() const {
         return guest_cntfrq_factor;
     }
+    u64 host_cntfrq;
     FactorType ns_cntfrq_factor;
     FactorType us_cntfrq_factor;
     FactorType ms_cntfrq_factor;
@@ -183,6 +199,6 @@ std::optional<int> GetProcessorCount();
 /// @brief Global cpu caps
 extern const CPUCaps g_cpu_caps;
 /// @brief Global wall clock
-extern const WallClock g_wall_clock;
+extern WallClock g_wall_clock;
 
 } // namespace Common