[llvm] [AMDGPU] Use table strategy for LowerModuleLDSPass at O0 (PR #160181)

[via llvm-commits]

================
@@ -0,0 +1,264 @@
+; RUN: opt -S -mtriple=amdgcn-- -mcpu=gfx942 -amdgpu-lower-module-lds < %s 2>&1 | FileCheck %s
+; RUN: opt -S -mtriple=amdgcn-- -mcpu=gfx942 -passes=amdgpu-lower-module-lds < %s 2>&1 | FileCheck %s
+
----------------
hjagasiaAMD wrote:

Pls note, the godbolt IR does not show the issue. The reproducer in the squashed patch (which Jon has also pasted in his comment does)

https://github.com/llvm/llvm-project/pull/160181