[llvm] [PowerPC] Do not rely on MachineSink threshold to optimize CR spilling (PR #128745)

[Min-Yih Hsu via llvm-commits]

================
@@ -0,0 +1,27 @@
+# RUN: llc -run-pass=machine-sink -mtriple=powerpc64-unknown-linux-gnu -o - %s | FileCheck %s
+
+---
+name:            sink-crset-crunset
+body:             |
+  ; CHECK-LABEL: bb.0.entry
+  bb.0.entry:
+    successors: %bb.1(0x40000000), %bb.2(0x40000000)
+    liveins: $r3
+    ; CHECK-NOT: CRSET
----------------
mshockwave wrote:

could you elaborate why MIR test is a better choice over normal codegen test? Is is because CRSET and CRUNSET being pseudo instructions that will be expanded into immediate as you mentioned?

https://github.com/llvm/llvm-project/pull/128745