[llvm] [AMDGPU] IGLP: Fixes for VMEM load detection and unsigned int handling (PR #135090)

[Robert Imschweiler via llvm-commits]

================
@@ -80,6 +80,10 @@ enum class SchedGroupMask {
   LLVM_MARK_AS_BITMASK_ENUM(/* LargestFlag = */ ALL)
 };
 
+static bool handleAsVMEMInstr(const MachineInstr &MI, const SIInstrInfo *TII) {
+  return TII->isVMEM(MI) || (TII->isFLAT(MI) && !TII->isDS(MI));
----------------
ro-i wrote:

Apparently not:
```
(gdb) p MI.dump()
  GLOBAL_STORE_DWORDX4_SADDR %14:vgpr_32, %135:vreg_128_align2, %143.sub2_sub3:sgpr_128, 0, 0, implicit $exec :: (store (s128) into %ir.gep2, align 128, addrspace 1)
$1 = void
(gdb) p TII->isVMEM(MI)
$2 = false
(gdb) p TII->isFLAT(MI)
$3 = true
(gdb) p TII->isDS(MI)
$4 = false
```

https://github.com/llvm/llvm-project/pull/135090