[llvm] AMDGPU: Avoid report_fatal_error in image intrinsic lowering (PR #145201)

[Sebastian Neubauer via llvm-commits]

================
@@ -8703,9 +8707,23 @@ SDValue SITargetLowering::lowerImage(SDValue Op,
     if (Subtarget->hasGFX90AInsts()) {
       Opcode = AMDGPU::getMIMGOpcode(IntrOpcode, AMDGPU::MIMGEncGfx90a,
                                      NumVDataDwords, NumVAddrDwords);
-      if (Opcode == -1)
-        report_fatal_error(
-            "requested image instruction is not supported on this GPU");
+      if (Opcode == -1) {
+        DAG.getContext()->diagnose(DiagnosticInfoUnsupported(
+            DAG.getMachineFunction().getFunction(),
+            "requested image instruction is not supported on this GPU",
+            DL.getDebugLoc()));
+
+        unsigned Idx = 0;
+        SmallVector<SDValue, 3> RetValues(OrigResultTypes.size());
+        for (EVT VT : OrigResultTypes) {
----------------
Flakebi wrote:

Could probably use enumerate?
```suggestion
        for (auto [Idx, VT] : llvm::enumerate(OrigResultTypes)) {
```

https://github.com/llvm/llvm-project/pull/145201