[llvm] [AMDGPU] Always Inline preserved analyses (PR #91198)

[via llvm-commits]

================
@@ -98,6 +99,7 @@ static bool alwaysInlineImpl(Module &M, bool GlobalOpt) {
       if (TT.getArch() == Triple::amdgcn &&
           A.getLinkage() != GlobalValue::InternalLinkage)
         continue;
+      Preserved = false;
----------------
jofrn wrote:

ah, thank you!

https://github.com/llvm/llvm-project/pull/91198