From 635d479322b9240af586e36acdcbf56db42778c2 Mon Sep 17 00:00:00 2001
From: Matt Arsenault <Matthew.Arsenault@amd.com>
Date: Wed, 3 Oct 2018 02:47:25 +0000
Subject: [PATCH] AMDGPU: Always run AMDGPUAlwaysInline

Even if calls are enabled, it still needs to be run
for forcing inline of functions that use LDS.

llvm-svn: 343657
---
 llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp      | 21 +++++++++------------
 ...force-alwaysinline-lds-global-address-codegen.ll | 21 +++++++++++++++++++++
 2 files changed, 30 insertions(+), 12 deletions(-)
 create mode 100644 llvm/test/CodeGen/AMDGPU/force-alwaysinline-lds-global-address-codegen.ll

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index f6ebc3d..ccefdf3 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -605,18 +605,15 @@ void AMDGPUPassConfig::addIRPasses() {
   addPass(createAtomicExpandPass());
   addPass(createAMDGPULowerIntrinsicsPass());
 
-  if (TM.getTargetTriple().getArch() == Triple::r600 ||
-      !EnableAMDGPUFunctionCalls) {
-    // Function calls are not supported, so make sure we inline everything.
-    addPass(createAMDGPUAlwaysInlinePass());
-    addPass(createAlwaysInlinerLegacyPass());
-    // We need to add the barrier noop pass, otherwise adding the function
-    // inlining pass will cause all of the PassConfigs passes to be run
-    // one function at a time, which means if we have a nodule with two
-    // functions, then we will generate code for the first function
-    // without ever running any passes on the second.
-    addPass(createBarrierNoopPass());
-  }
+  // Function calls are not supported, so make sure we inline everything.
+  addPass(createAMDGPUAlwaysInlinePass());
+  addPass(createAlwaysInlinerLegacyPass());
+  // We need to add the barrier noop pass, otherwise adding the function
+  // inlining pass will cause all of the PassConfigs passes to be run
+  // one function at a time, which means if we have a nodule with two
+  // functions, then we will generate code for the first function
+  // without ever running any passes on the second.
+  addPass(createBarrierNoopPass());
 
   if (TM.getTargetTriple().getArch() == Triple::amdgcn) {
     // TODO: May want to move later or split into an early and late one.
diff --git a/llvm/test/CodeGen/AMDGPU/force-alwaysinline-lds-global-address-codegen.ll b/llvm/test/CodeGen/AMDGPU/force-alwaysinline-lds-global-address-codegen.ll
new file mode 100644
index 0000000..6d90cf9
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/force-alwaysinline-lds-global-address-codegen.ll
@@ -0,0 +1,21 @@
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -amdgpu-function-calls -amdgpu-stress-function-calls < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -amdgpu-stress-function-calls < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa < %s | FileCheck -check-prefix=GCN %s
+
+@lds0 = addrspace(3) global i32 undef, align 4
+
+; GCN-NOT: load_lds_simple
+
+define internal i32 @load_lds_simple() {
+  %load = load i32, i32 addrspace(3)* @lds0, align 4
+  ret i32 %load
+}
+
+; GCN-LABEL: {{^}}kernel:
+; GCN: v_mov_b32_e32 [[ADDR:v[0-9]+]], 0
+; GCN: ds_read_b32 v{{[0-9]+}}, [[ADDR]]
+define amdgpu_kernel void @kernel(i32 addrspace(1)* %out) {
+  %call = call i32 @load_lds_simple()
+  store i32 %call, i32 addrspace(1)* %out
+  ret void
+}
-- 
2.7.4