Merge pull request #80 from chrisps/canary_experimental

Reduce size of final exe by about 1.6mb

premake5.lua

@@ -62,6 +62,12 @@ filter({"configurations:Checked", "platforms:Linux"})
   defines({
     "_GLIBCXX_DEBUG",   -- libstdc++ debug mode
   })
+filter({"configurations:Release", "platforms:Windows"})
+	buildoptions({
+		"/Gw", 
+		"/GS-", 
+		"/Oy"
+	})
 
 filter("configurations:Debug")
   runtime("Release")

src/xenia/app/premake5.lua

@@ -119,6 +119,11 @@ project("xenia-app")
       "xenia-gpu-d3d12-trace-viewer",
       "xenia-ui-window-d3d12-demo",
     })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
 
   filter("platforms:Windows")
     -- Only create the .user file if it doesn't already exist.

src/xenia/apu/conversion.h

@@ -21,29 +21,7 @@ namespace conversion {
 
 #if XE_ARCH_AMD64
 
-#if 0
+
-inline void sequential_6_BE_to_interleaved_6_LE(float* output,
-                                                const float* input,
-                                                size_t ch_sample_count) {
-  const uint32_t* in = reinterpret_cast<const uint32_t*>(input);
-  uint32_t* out = reinterpret_cast<uint32_t*>(output);
-  const __m128i byte_swap_shuffle =
-      _mm_set_epi8(12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);
-  for (size_t sample = 0; sample < ch_sample_count; sample++) {
-    __m128i sample0 = _mm_set_epi32(
-        in[3 * ch_sample_count + sample], in[2 * ch_sample_count + sample],
-        in[1 * ch_sample_count + sample], in[0 * ch_sample_count + sample]);
-    uint32_t sample1 = in[4 * ch_sample_count + sample];
-    uint32_t sample2 = in[5 * ch_sample_count + sample];
-    sample0 = _mm_shuffle_epi8(sample0, byte_swap_shuffle);
-    _mm_storeu_si128(reinterpret_cast<__m128i*>(&out[sample * 6]), sample0);
-    sample1 = xe::byte_swap(sample1);
-    out[sample * 6 + 4] = sample1;
-    sample2 = xe::byte_swap(sample2);
-    out[sample * 6 + 5] = sample2;
-  }
-}
-#else
 XE_NOINLINE
 static void _generic_sequential_6_BE_to_interleaved_6_LE(
     float* XE_RESTRICT output, const float* XE_RESTRICT input,
@@ -58,6 +36,8 @@ static void _generic_sequential_6_BE_to_interleaved_6_LE(
     }
   }
 }
+#if XE_COMPILER_CLANG_CL != 1
+// load_be_u32 unavailable on clang-cl
 XE_NOINLINE
 static void _movbe_sequential_6_BE_to_interleaved_6_LE(
     float* XE_RESTRICT output, const float* XE_RESTRICT input,
@@ -80,7 +60,13 @@ inline static void sequential_6_BE_to_interleaved_6_LE(
                                                  ch_sample_count);
   }
 }
+#else
+inline static void sequential_6_BE_to_interleaved_6_LE(
+    float* output, const float* input, unsigned ch_sample_count) {
+  _generic_sequential_6_BE_to_interleaved_6_LE(output, input, ch_sample_count);
+}
 #endif
+
 inline void sequential_6_BE_to_interleaved_2_LE(float* output,
                                                 const float* input,
                                                 size_t ch_sample_count) {

src/xenia/base/cvar.cc

@@ -8,7 +8,7 @@
  */
 
 #include "xenia/base/cvar.h"
-
+#include <iostream>
 #define UTF_CPP_CPLUSPLUS 201703L
 #include "third_party/utfcpp/source/utf8.h"
 
@@ -87,7 +87,7 @@ void ParseLaunchArguments(int& argc, char**& argv,
         configVar->LoadFromLaunchOptions(&result);
       }
     }
-  } catch (const cxxopts::OptionException& e) {
+  } catch (const cxxopts::exceptions::exception& e) {
     xe::AttachConsole();
     if (xe::has_console_attached()) {
       std::cout << e.what() << std::endl;

src/xenia/base/dma.cc

@@ -2,7 +2,6 @@
 #include "logging.h"
 #include "mutex.h"
 #include "platform_win.h"
-#include "xbyak/xbyak/xbyak_util.h"
 
 XE_NTDLL_IMPORT(NtDelayExecution, cls_NtDelayExecution,
                 NtDelayExecutionPointer);
@@ -22,7 +21,13 @@ static void xedmaloghelper(const char (&fmt)[N], Ts... args) {
 #define XEDMALOG(...) static_cast<void>(0)
 using xe::swcache::CacheLine;
 static constexpr unsigned NUM_CACHELINES_IN_PAGE = 4096 / sizeof(CacheLine);
-
+#if defined(__clang__)
+XE_FORCEINLINE
+static void mvdir64b(void* to, const void* from) {
+  __asm__("movdir64b %1, %0" : : "r"(to), "m"(*(char*)from) : "memory");
+}
+#define _movdir64b mvdir64b
+#endif
 XE_FORCEINLINE
 static void XeCopy16384StreamingAVX(CacheLine* XE_RESTRICT to,
                                     CacheLine* XE_RESTRICT from) {
@@ -140,6 +145,7 @@ static void vastcpy_impl_avx(CacheLine* XE_RESTRICT physaddr,
     xe::swcache::WriteLineNT(physaddr + i, &line0);
   }
 }
+
 static void vastcpy_impl_movdir64m(CacheLine* XE_RESTRICT physaddr,
                                    CacheLine* XE_RESTRICT rdmapping,
                                    uint32_t written_length) {
@@ -171,24 +177,6 @@ static void vastcpy_impl_movdir64m(CacheLine* XE_RESTRICT physaddr,
   }
 }
 
-static class DMAFeatures {
- public:
-  uint32_t has_fast_rep_movsb : 1;
-  uint32_t has_movdir64b : 1;
-
-  DMAFeatures() {
-    unsigned int data[4];
-    memset(data, 0, sizeof(data));
-    // intel extended features
-    Xbyak::util::Cpu::getCpuidEx(7, 0, data);
-    if (data[2] & (1 << 28)) {
-      has_movdir64b = 1;
-    }
-    if (data[1] & (1 << 9)) {
-      has_fast_rep_movsb = 1;
-    }
-  }
-} dma_x86_features;
 XE_COLD
 static void first_vastcpy(CacheLine* XE_RESTRICT physaddr,
                           CacheLine* XE_RESTRICT rdmapping,
@@ -201,7 +189,7 @@ static void first_vastcpy(CacheLine* XE_RESTRICT physaddr,
                           CacheLine* XE_RESTRICT rdmapping,
                           uint32_t written_length) {
   VastCpyDispatch dispatch_to_use = nullptr;
-  if (dma_x86_features.has_movdir64b) {
+  if (amd64::GetFeatureFlags() & amd64::kX64EmitMovdir64M) {
     XELOGI("Selecting MOVDIR64M vastcpy.");
     dispatch_to_use = vastcpy_impl_movdir64m;
   } else {
@@ -271,10 +259,10 @@ class XeDMACGeneric : public XeDMAC {
   virtual void WaitJobDone(DMACJobHandle handle) override {
     while (WaitForSingleObject((HANDLE)handle, 2) == WAIT_TIMEOUT) {
       // NtAlertThreadByThreadId.invoke<void>(thrd_->system_id());
-    //  while (SignalObjectAndWait(gotjob_event, (HANDLE)handle, 2, false) ==
+      //  while (SignalObjectAndWait(gotjob_event, (HANDLE)handle, 2, false) ==
-  //           WAIT_TIMEOUT) {
+      //           WAIT_TIMEOUT) {
-   //    ;
+      //    ;
-      }
+    }
     //}
 
     // SignalObjectAndWait(gotjob_event, (HANDLE)handle, INFINITE, false);

src/xenia/base/memory.h

@@ -616,23 +616,23 @@ static void Prefetch(const void* addr) {
 }
 
 template <>
-static void Prefetch<PrefetchTag::Write>(const void* addr) {
+void Prefetch<PrefetchTag::Write>(const void* addr) {
   PrefetchW(addr);
 }
 template <>
-static void Prefetch<PrefetchTag::Nontemporal>(const void* addr) {
+void Prefetch<PrefetchTag::Nontemporal>(const void* addr) {
   PrefetchNTA(addr);
 }
 template <>
-static void Prefetch<PrefetchTag::Level3>(const void* addr) {
+void Prefetch<PrefetchTag::Level3>(const void* addr) {
   PrefetchL3(addr);
 }
 template <>
-static void Prefetch<PrefetchTag::Level2>(const void* addr) {
+void Prefetch<PrefetchTag::Level2>(const void* addr) {
   PrefetchL2(addr);
 }
 template <>
-static void Prefetch<PrefetchTag::Level1>(const void* addr) {
+void Prefetch<PrefetchTag::Level1>(const void* addr) {
   PrefetchL1(addr);
 }
 // todo: does aarch64 have streaming stores/loads?

src/xenia/base/platform_amd64.cc

@@ -9,7 +9,7 @@
 
 #include "xenia/base/cvar.h"
 #include "xenia/base/platform.h"
-
+#define XBYAK_NO_OP_NAMES
 #include "third_party/xbyak/xbyak/xbyak.h"
 #include "third_party/xbyak/xbyak/xbyak_util.h"
 DEFINE_int64(x64_extension_mask, -1LL,

src/xenia/cpu/export_resolver.cc

@@ -81,7 +81,7 @@ void ExportResolver::SetFunctionMapping(const std::string_view module_name,
   auto export_entry = GetExportByOrdinal(module_name, ordinal);
   assert_not_null(export_entry);
   export_entry->tags |= ExportTag::kImplemented;
-  export_entry->function_data.shim = shim;
+  export_entry->function_data.trampoline = (ExportTrampoline)(void*)shim;
 }
 
 void ExportResolver::SetFunctionMapping(const std::string_view module_name,

src/xenia/cpu/export_resolver.h

@@ -44,57 +44,50 @@ struct ExportTag {
   // packed like so:
   // ll...... cccccccc ........ ..bihssi
 
-  static const int CategoryShift = 16;
+  static constexpr int CategoryShift = 16;
 
   // Export is implemented in some form and can be used.
-  static const type kImplemented = 1u << 0;
+  static constexpr type kImplemented = 1u << 0;
   // Export is a stub and is probably bad.
-  static const type kStub = 1u << 1;
+  static constexpr type kStub = 1u << 1;
   // Export is known to cause problems, or may not be complete.
-  static const type kSketchy = 1u << 2;
+  static constexpr type kSketchy = 1u << 2;
   // Export is called *a lot*.
-  static const type kHighFrequency = 1u << 3;
+  static constexpr type kHighFrequency = 1u << 3;
   // Export is important and should always be logged.
-  static const type kImportant = 1u << 4;
+  static constexpr type kImportant = 1u << 4;
   // Export blocks the calling thread
-  static const type kBlocking = 1u << 5;
+  static constexpr type kBlocking = 1u << 5;
-
+  static constexpr type kIsVariable = 1u << 6;
   // Export will be logged on each call.
-  static const type kLog = 1u << 30;
+  static constexpr type kLog = 1u << 30;
   // Export's result will be logged on each call.
-  static const type kLogResult = 1u << 31;
+  static constexpr type kLogResult = 1u << 31;
 };
 
 // DEPRECATED
 typedef void (*xe_kernel_export_shim_fn)(void*, void*);
 
 typedef void (*ExportTrampoline)(ppc::PPCContext* ppc_context);
-
+#pragma pack(push, 1)
 class Export {
  public:
   enum class Type {
     kFunction = 0,
     kVariable = 1,
   };
-
+  constexpr Export(uint16_t ordinal, Type type, const char* name,
-  Export(uint16_t ordinal, Type type, const char* name,
+                   ExportTag::type tags = 0)
-         ExportTag::type tags = 0)
+      : function_data({nullptr}),
-      : ordinal(ordinal),
+        name(name ? name : ""),
-        type(type),
         tags(tags),
-        function_data({nullptr, nullptr, 0}) {
+        ordinal(ordinal)
-    std::strncpy(this->name, name, xe::countof(this->name));
+
+  {
+    if (type == Type::kVariable) {
+      this->tags |= ExportTag::kIsVariable;
+    }
   }
-
-  uint16_t ordinal;
-  Type type;
-  char name[96];
-  ExportTag::type tags;
-
-  bool is_implemented() const {
-    return (tags & ExportTag::kImplemented) == ExportTag::kImplemented;
-  }
-
   union {
     // Variable data. Only valid when kXEKernelExportFlagVariable is set.
     // This is an address in the client memory space that the variable can
@@ -102,17 +95,26 @@ class Export {
     uint32_t variable_ptr;
 
     struct {
-      // DEPRECATED
-      xe_kernel_export_shim_fn shim;
 
       // Trampoline that is called from the guest-to-host thunk.
       // Expects only PPC context as first arg.
       ExportTrampoline trampoline;
-      uint64_t call_count;
     } function_data;
   };
-};
+  const char* const name;
+  ExportTag::type tags;
+  uint16_t ordinal;
+  // Type type;
 
+  constexpr bool is_implemented() const {
+    return (tags & ExportTag::kImplemented) == ExportTag::kImplemented;
+  }
+  constexpr Type get_type() const {
+    return (this->tags & ExportTag::kIsVariable) ? Type::kVariable
+                                                 : Type::kFunction;
+  }
+};
+#pragma pack(pop)
 class ExportResolver {
  public:
   class Table {

src/xenia/cpu/ppc/ppc_opcode_disasm_gen.cc

@@ -4948,8 +4948,8 @@ void PrintDisasm_xorx(const PPCDecodeData& d, StringBuffer* str) {
 }
 #define INIT_LIST(...) {__VA_ARGS__}
 #define INSTRUCTION(opcode, mnem, form, group, type, desc, reads, writes, fn) \
-    {PPCOpcodeGroup::group, PPCOpcodeFormat::form, opcode, mnem, desc, INIT_LIST reads, INIT_LIST writes, fn}
+    {PPCOpcodeGroup::group, PPCOpcodeFormat::form, opcode, mnem, desc,  fn}
-PPCOpcodeDisasmInfo ppc_opcode_disasm_table[] = {
+static constexpr PPCOpcodeDisasmInfo ppc_opcode_disasm_table[] = {
   INSTRUCTION(0x7c000014, "addcx"       , kXO     , kI, kGeneral, "Add Carrying"                                                               , (PPCOpcodeField::kRA,PPCOpcodeField::kRB), (PPCOpcodeField::kRD,PPCOpcodeField::kCA,PPCOpcodeField::kOEcond,PPCOpcodeField::kCRcond), PrintDisasm_addcx),
   INSTRUCTION(0x7c000114, "addex"       , kXO     , kI, kGeneral, "Add Extended"                                                               , (PPCOpcodeField::kRA,PPCOpcodeField::kRB,PPCOpcodeField::kCA), (PPCOpcodeField::kRD,PPCOpcodeField::kOEcond,PPCOpcodeField::kCRcond), PrintDisasm_addex),
   INSTRUCTION(0x38000000, "addi"        , kD      , kI, kGeneral, "Add Immediate"                                                              , (PPCOpcodeField::kRA0,PPCOpcodeField::kSIMM), (PPCOpcodeField::kRD), PrintDisasm_addi),
@@ -5414,7 +5414,7 @@ const PPCOpcodeDisasmInfo& GetOpcodeDisasmInfo(PPCOpcode opcode) {
 }
 void RegisterOpcodeDisasm(PPCOpcode opcode, InstrDisasmFn fn) {
   assert_null(ppc_opcode_disasm_table[static_cast<int>(opcode)].disasm);
-  ppc_opcode_disasm_table[static_cast<int>(opcode)].disasm = fn;
+  const_cast<PPCOpcodeDisasmInfo*>( &ppc_opcode_disasm_table[static_cast<int>(opcode)])->disasm = fn;
 }
 
 }  // namespace ppc

src/xenia/cpu/ppc/ppc_opcode_info.h

@@ -133,18 +133,18 @@ enum class PPCOpcodeField : uint32_t {
   kTO,
   kLEV,
 };
-
+#pragma pack(push, 1)
 struct PPCOpcodeDisasmInfo {
   PPCOpcodeGroup group;
   PPCOpcodeFormat format;
   uint32_t opcode;
   const char* name;
   const char* description;
-  std::vector<PPCOpcodeField> reads;
+  // std::vector<PPCOpcodeField> reads;
-  std::vector<PPCOpcodeField> writes;
+  // std::vector<PPCOpcodeField> writes;
   InstrDisasmFn disasm;
 };
-
+#pragma pack(pop)
 PPCOpcode LookupOpcode(uint32_t code);
 
 const PPCOpcodeInfo& GetOpcodeInfo(PPCOpcode opcode);

src/xenia/cpu/premake5.lua

@@ -10,6 +10,11 @@ project("xenia-cpu")
     "xenia-base",
     "mspack",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   includedirs({
     project_root.."/third_party/llvm/include",
   })

src/xenia/cpu/xex_module.cc

@@ -1194,11 +1194,11 @@ bool XexModule::SetupLibraryImports(const std::string_view name,
       }
 
       if (kernel_export) {
-        if (kernel_export->type == Export::Type::kFunction) {
+        if (kernel_export->get_type() == Export::Type::kFunction) {
           // Not exactly sure what this should be...
           // Appears to be ignored.
           *record_slot = 0xDEADC0DE;
-        } else if (kernel_export->type == Export::Type::kVariable) {
+        } else if (kernel_export->get_type() == Export::Type::kVariable) {
           // Kernel import variable
           if (kernel_export->is_implemented()) {
             // Implemented - replace with pointer.
@@ -1287,8 +1287,9 @@ bool XexModule::SetupLibraryImports(const std::string_view name,
             handler = (GuestFunction::ExternHandler)
                           kernel_export->function_data.trampoline;
           } else {
-            handler =
+            //__debugbreak();
-                (GuestFunction::ExternHandler)kernel_export->function_data.shim;
+            // handler =
+            //     (GuestFunction::ExternHandler)kernel_export->function_data.shim;
           }
         } else {
           XELOGW("WARNING: Imported kernel function {} is unimplemented!",

src/xenia/debug/ui/debug_window.cc

@@ -258,7 +258,6 @@ void DebugWindow::DrawFrame(ImGuiIO& io) {
   ImGui::PopStyleVar();
 
   if (cvars::imgui_debug) {
-    ImGui::ShowDemoWindow();
     ImGui::ShowMetricsWindow();
   }
 }
@@ -1254,7 +1253,7 @@ void DebugWindow::DrawBreakpointsPane() {
         continue;
       }
       auto export_entry = all_exports[call_rankings[i].first];
-      if (export_entry->type != cpu::Export::Type::kFunction ||
+      if (export_entry->get_type() != cpu::Export::Type::kFunction ||
           !export_entry->is_implemented()) {
         continue;
       }

src/xenia/debug/ui/premake5.lua

@@ -12,6 +12,11 @@ project("xenia-debug-ui")
     "xenia-cpu",
     "xenia-ui",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   defines({
   })
   includedirs({

src/xenia/gpu/command_processor.cc

@@ -527,56 +527,48 @@ void CommandProcessor::WriteFetchRangeFromRing(xe::RingBuffer* ring,
   WriteRegisterRangeFromRing(ring, base + 0x4800, num_times);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteBoolRangeFromRing(xe::RingBuffer* ring,
                                               uint32_t base,
                                               uint32_t num_times) {
   WriteRegisterRangeFromRing(ring, base + 0x4900, num_times);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteLoopRangeFromRing(xe::RingBuffer* ring,
                                               uint32_t base,
                                               uint32_t num_times) {
   WriteRegisterRangeFromRing(ring, base + 0x4908, num_times);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteREGISTERSRangeFromRing(xe::RingBuffer* ring,
                                                    uint32_t base,
                                                    uint32_t num_times) {
   WriteRegisterRangeFromRing(ring, base + 0x2000, num_times);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteALURangeFromMem(uint32_t start_index,
                                             uint32_t* base,
                                             uint32_t num_registers) {
   WriteRegistersFromMem(start_index + 0x4000, base, num_registers);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteFetchRangeFromMem(uint32_t start_index,
                                               uint32_t* base,
                                               uint32_t num_registers) {
   WriteRegistersFromMem(start_index + 0x4800, base, num_registers);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteBoolRangeFromMem(uint32_t start_index,
                                              uint32_t* base,
                                              uint32_t num_registers) {
   WriteRegistersFromMem(start_index + 0x4900, base, num_registers);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteLoopRangeFromMem(uint32_t start_index,
                                              uint32_t* base,
                                              uint32_t num_registers) {
   WriteRegistersFromMem(start_index + 0x4908, base, num_registers);
 }
 
-XE_FORCEINLINE
 void CommandProcessor::WriteREGISTERSRangeFromMem(uint32_t start_index,
                                                   uint32_t* base,
                                                   uint32_t num_registers) {

src/xenia/gpu/command_processor.h

@@ -178,43 +178,33 @@ class CommandProcessor {
           num_times);  // repeatedly write a value to one register, presumably a
                        // register with special handling for writes
 
-  XE_FORCEINLINE
   void WriteALURangeFromRing(xe::RingBuffer* ring, uint32_t base,
                              uint32_t num_times);
 
-  XE_FORCEINLINE
   void WriteFetchRangeFromRing(xe::RingBuffer* ring, uint32_t base,
                                uint32_t num_times);
 
-  XE_FORCEINLINE
   void WriteBoolRangeFromRing(xe::RingBuffer* ring, uint32_t base,
                               uint32_t num_times);
 
-  XE_FORCEINLINE
   void WriteLoopRangeFromRing(xe::RingBuffer* ring, uint32_t base,
                               uint32_t num_times);
 
-  XE_FORCEINLINE
   void WriteREGISTERSRangeFromRing(xe::RingBuffer* ring, uint32_t base,
                                    uint32_t num_times);
 
-  XE_FORCEINLINE
   void WriteALURangeFromMem(uint32_t start_index, uint32_t* base,
                             uint32_t num_registers);
 
-  XE_FORCEINLINE
   void WriteFetchRangeFromMem(uint32_t start_index, uint32_t* base,
                               uint32_t num_registers);
 
-  XE_FORCEINLINE
   void WriteBoolRangeFromMem(uint32_t start_index, uint32_t* base,
                              uint32_t num_registers);
 
-  XE_FORCEINLINE
   void WriteLoopRangeFromMem(uint32_t start_index, uint32_t* base,
                              uint32_t num_registers);
 
-  XE_FORCEINLINE
   void WriteREGISTERSRangeFromMem(uint32_t start_index, uint32_t* base,
                                   uint32_t num_registers);
 

src/xenia/gpu/pm4_command_processor_implement.h

@@ -9,8 +9,6 @@ void COMMAND_PROCESSOR::ExecuteIndirectBuffer(uint32_t ptr,
   RingBuffer old_reader = reader_;
 
   // Execute commands!
-  // RingBuffer reader(memory_->TranslatePhysical(ptr), count *
-  // sizeof(uint32_t)); reader.set_write_offset(count * sizeof(uint32_t));
   new (&reader_)
       RingBuffer(memory_->TranslatePhysical(ptr), count * sizeof(uint32_t));
   reader_.set_write_offset(count * sizeof(uint32_t));
@@ -429,6 +427,38 @@ bool COMMAND_PROCESSOR::ExecutePacketType3_INDIRECT_BUFFER(
   COMMAND_PROCESSOR::ExecuteIndirectBuffer(GpuToCpu(list_ptr), list_length);
   return true;
 }
+
+XE_NOINLINE
+static bool MatchValueAndRef(uint32_t value, uint32_t ref, uint32_t wait_info) {
+  bool matched = false;
+  switch (wait_info & 0x7) {
+    case 0x0:  // Never.
+      matched = false;
+      break;
+    case 0x1:  // Less than reference.
+      matched = value < ref;
+      break;
+    case 0x2:  // Less than or equal to reference.
+      matched = value <= ref;
+      break;
+    case 0x3:  // Equal to reference.
+      matched = value == ref;
+      break;
+    case 0x4:  // Not equal to reference.
+      matched = value != ref;
+      break;
+    case 0x5:  // Greater than or equal to reference.
+      matched = value >= ref;
+      break;
+    case 0x6:  // Greater than reference.
+      matched = value > ref;
+      break;
+    case 0x7:  // Always
+      matched = true;
+      break;
+  }
+  return matched;
+}
 XE_NOINLINE
 bool COMMAND_PROCESSOR::ExecutePacketType3_WAIT_REG_MEM(
     uint32_t packet, uint32_t count) XE_RESTRICT {
@@ -459,32 +489,8 @@ bool COMMAND_PROCESSOR::ExecutePacketType3_WAIT_REG_MEM(
         value = register_file_->values[poll_reg_addr].u32;
       }
     }
-    switch (wait_info & 0x7) {
+    matched = MatchValueAndRef(value & mask, ref, wait_info);
-      case 0x0:  // Never.
+
-        matched = false;
-        break;
-      case 0x1:  // Less than reference.
-        matched = (value & mask) < ref;
-        break;
-      case 0x2:  // Less than or equal to reference.
-        matched = (value & mask) <= ref;
-        break;
-      case 0x3:  // Equal to reference.
-        matched = (value & mask) == ref;
-        break;
-      case 0x4:  // Not equal to reference.
-        matched = (value & mask) != ref;
-        break;
-      case 0x5:  // Greater than or equal to reference.
-        matched = (value & mask) >= ref;
-        break;
-      case 0x6:  // Greater than reference.
-        matched = (value & mask) > ref;
-        break;
-      case 0x7:  // Always
-        matched = true;
-        break;
-    }
     if (!matched) {
       // Wait.
       if (wait >= 0x100) {
@@ -598,34 +604,8 @@ bool COMMAND_PROCESSOR::ExecutePacketType3_COND_WRITE(
     assert_true(poll_reg_addr < RegisterFile::kRegisterCount);
     value = register_file_->values[poll_reg_addr].u32;
   }
-  bool matched = false;
+  bool matched = MatchValueAndRef(value & mask, ref, wait_info);
-  value &= mask;
+
-  switch (wait_info & 0x7) {
-    case 0x0:  // Never.
-      matched = false;
-      break;
-    case 0x1:  // Less than reference.
-      matched = value < ref;
-      break;
-    case 0x2:  // Less than or equal to reference.
-      matched = value <= ref;
-      break;
-    case 0x3:  // Equal to reference.
-      matched = value == ref;
-      break;
-    case 0x4:  // Not equal to reference.
-      matched = value != ref;
-      break;
-    case 0x5:  // Greater than or equal to reference.
-      matched = value >= ref;
-      break;
-    case 0x6:  // Greater than reference.
-      matched = value > ref;
-      break;
-    case 0x7:  // Always
-      matched = true;
-      break;
-  }
   if (matched) {
     // Write.
     if (wait_info & 0x100) {
@@ -718,9 +698,6 @@ bool COMMAND_PROCESSOR::ExecutePacketType3_EVENT_WRITE_EXT(
   for (unsigned i = 0; i < 6; ++i) {
     destination[i] = extents[i];
   }
-  // xe::copy_and_swap_16_unaligned(memory_->TranslatePhysical(address),
-  // extents,
-  //                                xe::countof(extents));
 
   trace_writer_.WriteMemoryWrite(CpuToGpu(address), sizeof(extents));
   return true;

src/xenia/hid/premake5.lua

@@ -38,7 +38,11 @@ project("xenia-hid-demo")
   resincludedirs({
     project_root,
   })
-
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   filter("platforms:not Android-*")
     links({
       "xenia-helper-sdl",

src/xenia/kernel/kernel_module.cc

@@ -94,7 +94,7 @@ uint32_t KernelModule::GetProcAddressByOrdinal(uint16_t ordinal) {
     // Export (or its parent library) not found.
     return 0;
   }
-  if (export_entry->type == cpu::Export::Type::kVariable) {
+  if (export_entry->get_type() == cpu::Export::Type::kVariable) {
     if (export_entry->variable_ptr) {
       return export_entry->variable_ptr;
     } else {
@@ -105,8 +105,7 @@ uint32_t KernelModule::GetProcAddressByOrdinal(uint16_t ordinal) {
       return 0;
     }
   } else {
-    if (export_entry->function_data.trampoline ||
+    if (export_entry->function_data.trampoline) {
-        export_entry->function_data.shim) {
       auto global_lock = global_critical_region_.Acquire();
 
       // See if the function has been generated already.
@@ -119,9 +118,6 @@ uint32_t KernelModule::GetProcAddressByOrdinal(uint16_t ordinal) {
       if (export_entry->function_data.trampoline) {
         handler = (cpu::GuestFunction::ExternHandler)
                       export_entry->function_data.trampoline;
-      } else {
-        handler =
-            (cpu::GuestFunction::ExternHandler)export_entry->function_data.shim;
       }
 
       uint32_t guest_addr =

src/xenia/kernel/user_module.cc

@@ -783,7 +783,7 @@ void UserModule::Dump() {
           }
         }
         if (kernel_export &&
-            kernel_export->type == cpu::Export::Type::kVariable) {
+            kernel_export->get_type() == cpu::Export::Type::kVariable) {
           sb.AppendFormat("   V {:08X}          {:03X} ({:4}) {} {}\n",
                           info->value_address, info->ordinal, info->ordinal,
                           implemented ? "  " : "!!", name);

src/xenia/kernel/util/shim_utils.h

@@ -554,7 +554,6 @@ struct ExportRegistrerHelper {
         new cpu::Export(ORDINAL, xe::cpu::Export::Type::kFunction, name, TAGS);
     struct X {
       static void Trampoline(PPCContext* ppc_context) {
-        ++export_entry->function_data.call_count;
         Param::Init init = {
             ppc_context,
             0,

src/xenia/kernel/xam/xam_input.cc

@@ -91,7 +91,7 @@ DECLARE_XAM_EXPORT1(XamInputGetCapabilitiesEx, kInput, kSketchy);
 dword_result_t XamInputGetState_entry(dword_t user_index, dword_t flags,
                                       pointer_t<X_INPUT_STATE> input_state) {
   if (input_state) {
-    memset((void*)input_state.host_address(), 0, sizeof X_INPUT_STATE);
+    memset((void*)input_state.host_address(), 0, sizeof(X_INPUT_STATE));
   }
   if (user_index >= 4) {
     return X_ERROR_DEVICE_NOT_CONNECTED;

src/xenia/kernel/xam/xam_module.cc

@@ -41,21 +41,21 @@ xe::cpu::Export* RegisterExport_xam(xe::cpu::Export* export_entry) {
   xam_exports[export_entry->ordinal] = export_entry;
   return export_entry;
 }
-
+// Build the export table used for resolution.
+#include "xenia/kernel/util/export_table_pre.inc"
+static constexpr xe::cpu::Export xam_export_table[] = {
+#include "xenia/kernel/xam/xam_table.inc"
+};
+#include "xenia/kernel/util/export_table_post.inc"
 void XamModule::RegisterExportTable(xe::cpu::ExportResolver* export_resolver) {
   assert_not_null(export_resolver);
 
-// Build the export table used for resolution.
-#include "xenia/kernel/util/export_table_pre.inc"
-  static xe::cpu::Export xam_export_table[] = {
-#include "xenia/kernel/xam/xam_table.inc"
-  };
-#include "xenia/kernel/util/export_table_post.inc"
   for (size_t i = 0; i < xe::countof(xam_export_table); ++i) {
     auto& export_entry = xam_export_table[i];
     assert_true(export_entry.ordinal < xam_exports.size());
     if (!xam_exports[export_entry.ordinal]) {
-      xam_exports[export_entry.ordinal] = &export_entry;
+      xam_exports[export_entry.ordinal] =
+          const_cast<xe::cpu::Export*>(&export_entry);
     }
   }
   export_resolver->RegisterTable("xam.xex", &xam_exports);

src/xenia/kernel/xboxkrnl/xboxkrnl_strings.cc

@@ -824,7 +824,7 @@ class WideCountFormatData : public FormatData {
   int32_t count_;
 };
 
-SHIM_CALL DbgPrint_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL DbgPrint_entry(PPCContext* ppc_context) {
   uint32_t format_ptr = SHIM_GET_ARG_32(0);
   if (!format_ptr) {
     SHIM_SET_RETURN_32(X_STATUS_INVALID_PARAMETER);
@@ -854,7 +854,7 @@ SHIM_CALL DbgPrint_entry(PPCContext* ppc_context, KernelState* kernel_state) {
 }
 
 // https://msdn.microsoft.com/en-us/library/2ts7cx93.aspx
-SHIM_CALL _snprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL _snprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   int32_t buffer_count = SHIM_GET_ARG_32(1);
   uint32_t format_ptr = SHIM_GET_ARG_32(2);
@@ -894,7 +894,7 @@ SHIM_CALL _snprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
 }
 
 // https://msdn.microsoft.com/en-us/library/ybk95axf.aspx
-SHIM_CALL sprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL sprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   uint32_t format_ptr = SHIM_GET_ARG_32(1);
 
@@ -925,7 +925,7 @@ SHIM_CALL sprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
 }
 
 // https://msdn.microsoft.com/en-us/library/2ts7cx93.aspx
-SHIM_CALL _snwprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL _snwprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   int32_t buffer_count = SHIM_GET_ARG_32(1);
   uint32_t format_ptr = SHIM_GET_ARG_32(2);
@@ -966,7 +966,7 @@ SHIM_CALL _snwprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
 }
 
 // https://msdn.microsoft.com/en-us/library/ybk95axf.aspx
-SHIM_CALL swprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL swprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   uint32_t format_ptr = SHIM_GET_ARG_32(1);
 
@@ -998,7 +998,7 @@ SHIM_CALL swprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
 }
 
 // https://msdn.microsoft.com/en-us/library/1kt27hek.aspx
-SHIM_CALL _vsnprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL _vsnprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   int32_t buffer_count = SHIM_GET_ARG_32(1);
   uint32_t format_ptr = SHIM_GET_ARG_32(2);
@@ -1087,7 +1087,7 @@ SHIM_CALL _vsnwprintf_entry(PPCContext* ppc_context,
 }
 
 // https://msdn.microsoft.com/en-us/library/28d5ce15.aspx
-SHIM_CALL vsprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL vsprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   uint32_t format_ptr = SHIM_GET_ARG_32(1);
   uint32_t arg_ptr = SHIM_GET_ARG_32(2);
@@ -1147,7 +1147,7 @@ SHIM_CALL _vscwprintf_entry(PPCContext* ppc_context,
 }
 
 // https://msdn.microsoft.com/en-us/library/28d5ce15.aspx
-SHIM_CALL vswprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
+SHIM_CALL vswprintf_entry(PPCContext* ppc_context) {
   uint32_t buffer_ptr = SHIM_GET_ARG_32(0);
   uint32_t format_ptr = SHIM_GET_ARG_32(1);
   uint32_t arg_ptr = SHIM_GET_ARG_32(2);
@@ -1179,7 +1179,7 @@ SHIM_CALL vswprintf_entry(PPCContext* ppc_context, KernelState* kernel_state) {
   }
   SHIM_SET_RETURN_32(count);
 }
-
+#if 1
 void RegisterStringExports(xe::cpu::ExportResolver* export_resolver,
                            KernelState* state) {
   SHIM_SET_MAPPING("xboxkrnl.exe", DbgPrint, state);
@@ -1193,7 +1193,7 @@ void RegisterStringExports(xe::cpu::ExportResolver* export_resolver,
   SHIM_SET_MAPPING("xboxkrnl.exe", vswprintf, state);
   SHIM_SET_MAPPING("xboxkrnl.exe", _vsnwprintf, state);
 }
-
+#endif
 }  // namespace xboxkrnl
 }  // namespace kernel
 }  // namespace xe

src/xenia/ui/d3d12/premake5.lua

@@ -10,6 +10,11 @@ project("xenia-ui-d3d12")
     "xenia-base",
     "xenia-ui",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   local_platform_files()
   files({
     "../shaders/bytecode/d3d12_5_1/*.h",

src/xenia/ui/premake5.lua

@@ -14,7 +14,11 @@ project("xenia-ui")
   local_platform_files()
   removefiles({"*_demo.cc"})
   removefiles({"windowed_app_main_*.cc"})
-
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   filter("platforms:Android-*")
     -- Exports JNI functions.
     wholelib("On")

src/xenia/ui/vulkan/premake5.lua

@@ -10,6 +10,11 @@ project("xenia-ui-vulkan")
     "xenia-base",
     "xenia-ui",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   includedirs({
     project_root.."/third_party/Vulkan-Headers/include",
   })

src/xenia/vfs/premake5.lua

@@ -11,6 +11,11 @@ project("xenia-vfs")
   })
   defines({
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   recursive_platform_files()
   removefiles({"vfs_dump.cc"})
 

third_party/SDL2-static.lua

@@ -7,6 +7,8 @@ project("SDL2")
 
   defines({
     "HAVE_LIBC",
+	"SDL_LEAN_AND_MEAN=1",
+	"SDL_RENDER_DISABLED=1"
   })
   links({
     "setupapi.lib",

third_party/capstone.lua

@@ -11,6 +11,11 @@ project("capstone")
     "CAPSTONE_USE_SYS_DYN_MEM",
     "_LIB",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+	buildoptions({
+		"/Os",
+		"/O1"
+	})
   includedirs({
     "capstone",
     "capstone/include",

third_party/cxxopts

@@ -1 +1 @@
-Subproject commit 48e265dc4b347d3178e4458de81704e0d88047b2
+Subproject commit 2e3c6991d33811878ebcc0839d3815850d129b3a

third_party/fmt.lua

@@ -8,6 +8,11 @@ project("fmt")
   defines({
     "_LIB",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   includedirs({
     "fmt/include",
   })

third_party/glslang-spirv.lua

@@ -10,6 +10,11 @@ project("glslang-spirv")
   })
   includedirs({
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   files({
     "glslang/SPIRV/bitutils.h",
     "glslang/SPIRV/disassemble.cpp",

third_party/imgui.lua

@@ -11,6 +11,11 @@ project("imgui")
   includedirs({
     "imgui",
   })
+  filter({"configurations:Release", "platforms:Windows"})
+    buildoptions({
+      "/Os",
+      "/O1"
+    })
   files({
     "imgui/imconfig.h",
     "imgui/imgui.cpp",