[llvm-branch-commits] [llvm] [BPF] expand cttz, ctlz for i32, i64 (PR #73668)

[Yingchi Long via llvm-branch-commits]

================
@@ -0,0 +1,304 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
+; RUN: llc < %s -march=bpf | FileCheck %s
+
+; test that we can expand CTTZ & CTLZ
+
+declare i32 @llvm.cttz.i32(i32, i1)
+
+define i32 @cttz_i32_zdef(i32 %a) {
+; CHECK-LABEL: cttz_i32_zdef:
----------------
inclyc wrote:

> (As we don't really test the expansion logic, we just test that some expansion is applied)

Hmm, I'm worried about if we do at tests like

```
; CHECK-LABEL: cttz_i32_zdef:
; CHECK:         r0 =
; CHECK:         exit
```

how can this catch BPF backend changes like `setOperationAction()` to `Custom` and do some custom lowering, as if custom lowering might be incorrect though?

I think we should test implementation details about "how to expand / custom" actually. For example if someone really changes the logic of expansion, I think these tests shall be regenerated and we can see such patches (touched our backend tests)

https://github.com/llvm/llvm-project/pull/73668