AArch64/GlobalISel: Narrow stack passed argument access size

This fixes a verifier error in the testcase from bug 47619.

The stack passed s3 value was widened to 4-bytes, and producing a
4-byte memory access with a < 1 byte result type. We need to either
widen the result type or narrow the access size. This copies the code
directly from the AMDGPU handling, which narrows the load size. I
don't like that every target has to handle this, but this is currently
broken on the 11 release branch and this is the simplest fix.

This reverts commit 42bfa7c.

(cherry picked from commit 6cb0d23)

Author: arsenm

llvm/lib/Target/AArch64/GISel/AArch64CallLowering.cpp

@@ -84,11 +84,16 @@ struct IncomingArgHandler : public CallLowering::ValueHandler {
     }
   }
 
-  void assignValueToAddress(Register ValVReg, Register Addr, uint64_t Size,
+  void assignValueToAddress(Register ValVReg, Register Addr, uint64_t MemSize,
                             MachinePointerInfo &MPO, CCValAssign &VA) override {
     MachineFunction &MF = MIRBuilder.getMF();
+
+    // The reported memory location may be wider than the value.
+    const LLT RegTy = MRI.getType(ValVReg);
+    MemSize = std::min(static_cast<uint64_t>(RegTy.getSizeInBytes()), MemSize);
+
     auto MMO = MF.getMachineMemOperand(
-        MPO, MachineMemOperand::MOLoad | MachineMemOperand::MOInvariant, Size,
+        MPO, MachineMemOperand::MOLoad | MachineMemOperand::MOInvariant, MemSize,
         inferAlignFromPtrInfo(MF, MPO));
     MIRBuilder.buildLoad(ValVReg, Addr, *MMO);
   }

llvm/test/CodeGen/AArch64/GlobalISel/irtranslator-stack-evt-bug47619.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
-; RUN: llc -global-isel -mtriple=aarch64-unknown-unknown -stop-after=irtranslator %s -o - | FileCheck %s
+; RUN: llc -global-isel -mtriple=aarch64-unknown-unknown -stop-after=irtranslator -verify-machineinstrs %s -o - | FileCheck %s
 
 ; Make sure the i3 %arg8 value is correctly handled. This was trying
 ; to use MVT for EVT values passed on the stack and asserting before
@@ -17,7 +17,7 @@ define i3 @bug47619(i64 %arg, i64 %arg1, i64 %arg2, i64 %arg3, i64 %arg4, i64 %a
   ; CHECK:   [[COPY6:%[0-9]+]]:_(s64) = COPY $x6
   ; CHECK:   [[COPY7:%[0-9]+]]:_(s64) = COPY $x7
   ; CHECK:   [[FRAME_INDEX:%[0-9]+]]:_(p0) = G_FRAME_INDEX %fixed-stack.0
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s3) = G_LOAD [[FRAME_INDEX]](p0) :: (invariant load 4 from %fixed-stack.0, align 16)
+  ; CHECK:   [[LOAD:%[0-9]+]]:_(s3) = G_LOAD [[FRAME_INDEX]](p0) :: (invariant load 1 from %fixed-stack.0, align 16)
   ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s3)
   ; CHECK:   $w0 = COPY [[ANYEXT]](s32)
   ; CHECK:   RET_ReallyLR implicit $w0