[llvm-branch-commits] [llvm] AMDGPU: Fix broken negative test from ancient times (PR #131106)

[Pravin Jagtap via llvm-branch-commits]

================
@@ -1,12 +1,22 @@
-; RUN: llc -mtriple=amdgcn -verify-machineinstrs %s -o - | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=amdgcn -mcpu=tahiti %s -o - | FileCheck %s
 
-; CHECK-LABEL: foo
-; CHECK-NOT: BUFFER_LOAD_DWORDX2_OFFSET
 ; After dead code elimination, that buffer load should be eliminated finally
 ; after dead lane detection.
 define amdgpu_kernel void @foo() {
+; CHECK-LABEL: foo:
+; CHECK:       ; %bb.0: ; %entry
+; CHECK-NEXT:    s_cbranch_execnz .LBB0_2
+; CHECK-NEXT:  ; %bb.1: ; %LeafBlock1
+; CHECK-NEXT:  .LBB0_2: ; %foo.exit
+; CHECK-NEXT:    s_mov_b32 s3, 0xf000
+; CHECK-NEXT:    s_mov_b32 s2, -1
+; CHECK-NEXT:    v_mov_b32_e32 v0, 0
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], 0
+; CHECK-NEXT:    s_endpgm
+; CHECK-NEXT:  ; %bb.3: ; %sw.bb10
 entry:
-  switch i8 undef, label %foo.exit [
+  switch i8 poison, label %foo.exit [
----------------
pravinjagtap wrote:

In one of the prev. PR github actions suggested following:

In tests, avoid using undef and having tests that trigger undefined behavior. If you need an operand with some unimportant value, you can add a new argument to the function and use that instead.


https://github.com/llvm/llvm-project/pull/131106