diff --git a/src/ARMInterpreter_Branch.cpp b/src/ARMInterpreter_Branch.cpp
index 88b14ab7..5731a0b6 100644
--- a/src/ARMInterpreter_Branch.cpp
+++ b/src/ARMInterpreter_Branch.cpp
@@ -104,7 +104,7 @@ void T_BL_LONG_1(ARM* cpu)
 
 void T_BL_LONG_2(ARM* cpu)
 {
-    if ((cpu->CurInstr & 0x1801) == 0x0801) // "BLX" with bit 0 set is an unvalid instruction.
+    if ((cpu->CurInstr & 0x1801) == 0x0801) // "BLX" with bit 0 set is an undefined instruction.
         return T_UNK(cpu); // TODO: Check ARM7 for exceptions
     
     s32 offset = (cpu->CurInstr & 0x7FF) << 1;
diff --git a/src/ARMInterpreter_LoadStore.cpp b/src/ARMInterpreter_LoadStore.cpp
index 159fc86f..a2c9d7cc 100644
--- a/src/ARMInterpreter_LoadStore.cpp
+++ b/src/ARMInterpreter_LoadStore.cpp
@@ -67,7 +67,7 @@ enum class Writeback
 };
 
 template<bool signextend, int size, Writeback writeback>
-void LoadSingle(ARM* cpu, u8 rd, u8 rn, s32 offset)
+void LoadSingle(ARM* cpu, const u8 rd, const u8 rn, const s32 offset)
 {
     static_assert((size == 8) || (size == 16) || (size == 32), "dummy this function only takes 8/16/32 for size!!!");
 
@@ -116,7 +116,19 @@ void LoadSingle(ARM* cpu, u8 rd, u8 rn, s32 offset)
 
 
     if constexpr (writeback >= Writeback::Post) addr += offset;
-    if constexpr (writeback != Writeback::None) cpu->R[rn] = addr;
+    if constexpr (writeback != Writeback::None) 
+    {
+        if (rn != 15) [[likely]] // r15 writeback fails on arm9
+        {
+            cpu->R[rn] = addr;
+        }
+        else if (cpu->Num == 1) // arm 7
+        {
+            // note that at no point does it actually write the value it loaded to a register...
+            cpu->JumpTo((addr+4) & ~1);
+            return;
+        }
+    }
 
     if (rd == 15)
     {
@@ -127,7 +139,7 @@ void LoadSingle(ARM* cpu, u8 rd, u8 rn, s32 offset)
 }
 
 template<int size, Writeback writeback>
-void StoreSingle(ARM* cpu, u8 rd, u8 rn, s32 offset)
+void StoreSingle(ARM* cpu, const u8 rd, const u8 rn, const s32 offset)
 {
     static_assert((size == 8) || (size == 16) || (size == 32), "dummy this function only takes 8/16/32 for size!!!");
 
@@ -163,7 +175,17 @@ void StoreSingle(ARM* cpu, u8 rd, u8 rn, s32 offset)
     }
     
     if constexpr (writeback >= Writeback::Post) addr += offset;
-    if constexpr (writeback != Writeback::None) cpu->R[rn] = addr;
+    if constexpr (writeback != Writeback::None) 
+    {
+        if (rn != 15) [[likely]] // r15 writeback fails on arm9
+        {
+            cpu->R[rn] = addr;
+        }
+        else if (cpu->Num == 1) // arm 7
+        {
+            cpu->JumpTo(addr & ~1);
+        }
+    }
 }