[LLVM][AArch64InstrInfo] Prevent fill folding when DstReg is SP. (#148885)

We can remove subreg COPY instructions by filling directly into the
COPY's destination register. However, this is only valid when the copy
and fill have compatible register classes.

Fixes #148659

Author: paulwalker-arm

llvm/lib/Target/AArch64/AArch64InstrInfo.cpp

@@ -6288,13 +6288,13 @@ MachineInstr *AArch64InstrInfo::foldMemoryOperandImpl(
     //   LDRWui %0:sub_32<def,read-undef>, %stack.0
     //
     if (IsFill && SrcMO.getSubReg() == 0 && DstMO.isUndef()) {
-      const TargetRegisterClass *FillRC;
+      const TargetRegisterClass *FillRC = nullptr;
       switch (DstMO.getSubReg()) {
       default:
-        FillRC = nullptr;
         break;
       case AArch64::sub_32:
-        FillRC = &AArch64::GPR32RegClass;
+        if (AArch64::GPR64RegClass.hasSubClassEq(getRegClass(DstReg)))
+          FillRC = &AArch64::GPR32RegClass;
         break;
       case AArch64::ssub:
         FillRC = &AArch64::FPR32RegClass;

llvm/test/CodeGen/AArch64/spill-fold.mir

@@ -10,6 +10,7 @@
   define i64 @test_subreg_fill_fold() { ret i64 0 }
   define double @test_subreg_fill_fold2() { ret double 0.0 }
   define <4 x float> @test_subreg_fill_fold3() { ret <4 x float> undef }
+  define i64 @test_subreg_fill_fold4() { ret i64 0 }
   define i64 @test_nzcv_spill_fold() { ret i64 0 }
 ...
 ---
@@ -121,6 +122,24 @@ body:             |
     RET_ReallyLR implicit $s0
 ...
 ---
+# CHECK-LABEL: name: test_subreg_fill_fold4
+# Ensure the COPY is maintained when its result register class is not compatible
+# with the fill load's.
+name:            test_subreg_fill_fold4
+registers:
+  - { id: 0, class: gpr32 }
+  - { id: 1, class: gpr64sp }
+body:             |
+  bb.0:
+    %0 = COPY $wzr
+    INLINEASM &nop, 1, 12, implicit-def dead $x0, 12, implicit-def dead $x1, 12, implicit-def dead $x2, 12, implicit-def dead $x3, 12, implicit-def dead $x4, 12, implicit-def dead $x5, 12, implicit-def dead $x6, 12, implicit-def dead $x7, 12, implicit-def dead $x8, 12, implicit-def dead $x9, 12, implicit-def dead $x10, 12, implicit-def dead $x11, 12, implicit-def dead $x12, 12, implicit-def dead $x13, 12, implicit-def dead $x14, 12, implicit-def dead $x15, 12, implicit-def dead $x16, 12, implicit-def dead $x17, 12, implicit-def dead $x18, 12, implicit-def dead $x19, 12, implicit-def dead $x20, 12, implicit-def dead $x21, 12, implicit-def dead $x22, 12, implicit-def dead $x23, 12, implicit-def dead $x24, 12, implicit-def dead $x25, 12, implicit-def dead $x26, 12, implicit-def dead $x27, 12, implicit-def dead $x28, 12, implicit-def dead $fp, 12, implicit-def dead $lr, 12, implicit-def $sp
+    ; CHECK: %2:gpr32 = LDRWui %stack.0, 0 :: (load (s32) from %stack.0)
+    ; CHECK: undef %1.sub_32:gpr64sp = COPY %2
+    undef %1.sub_32:gpr64sp = COPY %0
+    $x0 = COPY %1
+    RET_ReallyLR implicit $x0
+...
+---
 # CHECK-LABEL: name: test_nzcv_spill_fold
 # Ensure that nzcv COPY cannot be folded.
 name:            test_nzcv_spill_fold