Unify 128bit EE reads/writes

* Initially the LQ/SQ instructions were implemented to perform two
sequential 64bit operations to emulate 128bit reads/writes. However
this won't work well for us, especially when writing to the GIF FIFO.
To mitigate this we can use the __int128 gcc extension (yay for switching
to clang once again!), which provides us with an optimized way of storing
128bit data.

src/common/emulator.h

 #pragma once
 #include <common/component.h>
 #include <cpu/iop/dma.h>
-#include <fmt/core.h>
+#include <fmt/format.h>
 #include <memory>
 #include <fstream>
+#include <type_traits>
 
 namespace ee 
 {
@@ -121,6 +122,12 @@ namespace common
         {
             return (*handler)(paddr, data);
         }
+        
+        if constexpr (std::is_same<T, unsigned __int128>::value)
+        {
+            uint64_t upper = (data >> 64);
+            fmt::print("[{}] {:d}bit write {:#x}{:016x} to unknown address {:#x}\n", component_name[id], sizeof(T) * 8, upper, (uint64_t)data, paddr);
+        }
         else
         {
             fmt::print("[{}] {:d}bit write {:#x} to unknown address {:#x}\n", component_name[id], sizeof(T) * 8, data, paddr);

src/cpu/ee/ee.cc

@@ -350,8 +350,7 @@ namespace ee
         int16_t imm = (int16_t)instr.i_type.immediate;
 
         uint32_t vaddr = gpr[base].word[0] + imm;
-        gpr[rt].dword[0] = read<uint64_t>(vaddr);
-        gpr[rt].dword[1] = read<uint64_t>(vaddr + 8);
+        gpr[rt].qword = read<uint128_t>(vaddr);
 
         log("LQ: GPR[{:d}] = {:#x} from address {:#x} = GPR[{:d}] ({:#x} + {:#x}\n", rt, gpr[rt].dword[0], vaddr, base, gpr[base].word[0], imm);
     }
@@ -761,8 +760,7 @@ namespace ee
         }
         else
         {
-            write<uint64_t>(vaddr, data1);
-            write<uint64_t>(vaddr + 8, data2);
+            write<uint128_t>(vaddr, gpr[rt].qword);
         }
     }
 

src/cpu/ee/ee.h

@@ -52,9 +52,12 @@ namespace ee
         }
     };
 
+    using uint128_t = unsigned __int128;
+
     union Register
     {
-        uint64_t dword[2] = {};
+        uint128_t qword;
+        uint64_t dword[2];
         uint32_t word[4];
     };
 
@@ -133,7 +136,7 @@ namespace ee
         common::Emulator* emulator;
 
         /* Registers. */
-        Register gpr[32];
+        Register gpr[32] = {};
         uint32_t pc;
         uint64_t hi0, hi1, lo0, lo1;
         uint32_t sa;