Replace utils::cnttz{32,64} with std::countr_{zero,one}

Make #include <bit> mandatory.

Utilities/Thread.cpp

@@ -2406,7 +2406,7 @@ void thread_ctrl::set_thread_affinity_mask(u64 mask)
 	SetThreadAffinityMask(_this_thread, mask);
 #elif __APPLE__
 	// Supports only one core
-	thread_affinity_policy_data_t policy = { static_cast<integer_t>(utils::cnttz64(mask)) };
+	thread_affinity_policy_data_t policy = { static_cast<integer_t>(std::countr_zero(mask)) };
 	thread_port_t mach_thread = pthread_mach_thread_np(pthread_self());
 	thread_policy_set(mach_thread, THREAD_AFFINITY_POLICY, reinterpret_cast<thread_policy_t>(&policy), 1);
 #elif defined(__linux__) || defined(__DragonFly__) || defined(__FreeBSD__)

Utilities/asm.h

@@ -28,30 +28,6 @@ namespace utils
 #endif
 	}
 
-	inline u32 cnttz32(u32 arg, bool nonzero = false)
-	{
-#ifdef _MSC_VER
-		ulong res;
-		return _BitScanForward(&res, arg) || nonzero ? res : 32;
-#elif __BMI__
-		return _tzcnt_u32(arg);
-#else
-		return arg || nonzero ? __builtin_ctz(arg) : 32;
-#endif
-	}
-
-	inline u64 cnttz64(u64 arg, bool nonzero = false)
-	{
-#ifdef _MSC_VER
-		ulong res;
-		return _BitScanForward64(&res, arg) || nonzero ? res : 64;
-#elif __BMI__
-		return _tzcnt_u64(arg);
-#else
-		return arg || nonzero ? __builtin_ctzll(arg) : 64;
-#endif
-	}
-
 	inline u8 popcnt32(u32 arg)
 	{
 #ifdef _MSC_VER

Utilities/types.h

@@ -17,9 +17,12 @@
 #include <limits>
 #include <array>
 
-#if __has_include(<bit>)
-#include <bit>
+#ifdef _MSC_VER
+#ifndef __cpp_lib_bitops
+#define __cpp_lib_bitops
 #endif
+#endif
+#include <bit>
 
 #ifndef __has_builtin
 	#define __has_builtin(x) 0

rpcs3/Emu/CPU/CPUThread.cpp

@@ -280,7 +280,7 @@ struct cpu_counter
 			if (ok) [[likely]]
 			{
 				// Get actual slot number
-				array_slot = i * 64 + utils::cnttz64(~bits, false);
+				array_slot = i * 64 + std::countr_one(bits);
 				break;
 			}
 		}
@@ -314,7 +314,7 @@ void for_all_cpu(F&& func) noexcept
 	{
 		for (u64 bits = ctr->cpu_array_bits[i]; bits; bits &= bits - 1)
 		{
-			const u64 index = i * 64 + utils::cnttz64(bits, true);
+			const u64 index = i * 64 + std::countr_zero(bits);
 
 			if (cpu_thread* cpu = ctr->cpu_array[index].load())
 			{

rpcs3/Emu/Cell/PPUAnalyser.cpp

@@ -2144,7 +2144,7 @@ void ppu_acontext::MULLI(ppu_opcode_t op)
 		}
 	}
 
-	gpr[op.rd] = spec_gpr::range(min, max, gpr[op.ra].tz() + utils::cnttz64(op.simm16));
+	gpr[op.rd] = spec_gpr::range(min, max, gpr[op.ra].tz() + std::countr_zero<u64>(op.simm16));
 }
 
 void ppu_acontext::SUBFIC(ppu_opcode_t op)

rpcs3/Emu/Cell/PPUAnalyser.h

@@ -989,7 +989,7 @@ struct ppu_acontext
 		// Return number of trailing zero bits
 		u64 tz() const
 		{
-			return utils::cnttz64(mask());
+			return std::countr_zero(mask());
 		}
 
 		// Range NOT

rpcs3/Emu/RSX/Overlays/overlays.cpp

@@ -32,7 +32,7 @@ namespace rsx
 				return 0;
 			}
 
-			const u64 r = u64{1} << utils::cnttz64(~_old, false);
+			const u64 r = u64{1} << std::countr_one(_old);
 			::overlays.trace("Bit allocated (%u)", r);
 			return r;
 		}

rpcs3/Emu/RSX/VK/VKGSRender.cpp

@@ -819,7 +819,7 @@ void VKGSRender::check_heap_status(u32 flags)
 	else if (flags)
 	{
 		heap_critical = false;
-		u32 test = 1 << utils::cnttz32(flags, true);
+		u32 test = 1u << std::countr_zero(flags);
 
 		do
 		{

rpcs3/util/atomic.cpp

@@ -107,7 +107,7 @@ static u64 slot_alloc()
 		if (ok)
 		{
 			// Find lowest clear bit
-			return group * 64 + utils::cnttz64(~bits, false);
+			return group * 64 + std::countr_one(bits);
 		}
 	}
 
@@ -269,7 +269,7 @@ static u32 sema_alloc()
 		if (ok)
 		{
 			// Find lowest clear bit
-			const u32 id = group * 64 + static_cast<u32>(utils::cnttz64(~bits, false));
+			const u32 id = group * 64 + static_cast<u32>(std::countr_one(bits));
 
 #ifdef USE_POSIX
 			// Initialize semaphore (should be very fast)

rpcs3/util/atomic2.cpp

@@ -1,6 +1,5 @@
 #include "atomic2.hpp"
 #include "Utilities/JIT.h"
-#include "Utilities/asm.h"
 #include "Utilities/sysinfo.h"
 
 //
@@ -265,7 +264,7 @@ static u64 rec_alloc()
 		if (ok)
 		{
 			// Find lowest clear bit
-			return group * 64 + utils::cnttz64(~bits, false);
+			return group * 64 + std::countr_one(bits);
 		}
 	}