[llvm] [AMDGPU] Folding imm offset in more cases for scratch access (PR #70634)

[Fangrui Song via llvm-commits]

================
@@ -4494,14 +4502,78 @@ bool AMDGPUInstructionSelector::isDSOffset2Legal(Register Base, int64_t Offset0,
   return KB->signBitIsZero(Base);
 }
 
-bool AMDGPUInstructionSelector::isFlatScratchBaseLegal(
-    Register Base, uint64_t FlatVariant) const {
-  if (FlatVariant != SIInstrFlags::FlatScratch)
+// Return whether the operation has NoUnsignedWrap property.
+bool isNoUnsignedWrap(MachineInstr *Addr) {
----------------
MaskRay wrote:

I added `static` to make this internal linkage and changed a `};`to `}` in commit 9535e011c7b25ece5e51c0b9e0f2c43a2912a417

https://github.com/llvm/llvm-project/pull/70634