[DSE] Support eliminating memcpy.inline.

MemoryLocation has been taught about memcpy.inline, which means we can
get the memory locations read and written by it. This means DSE can
handle memcpy.inline

diff --git a/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp b/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
index 12514be..d703f13 100644 (file)
--- a/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
+++ b/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
@@ -229,6 +229,7 @@ static bool hasAnalyzableMemoryWrite(Instruction *I,
     case Intrinsic::memset:
     case Intrinsic::memmove:
     case Intrinsic::memcpy:
+    case Intrinsic::memcpy_inline:
     case Intrinsic::memcpy_element_unordered_atomic:
     case Intrinsic::memmove_element_unordered_atomic:
     case Intrinsic::memset_element_unordered_atomic:
@@ -323,6 +324,7 @@ static bool isRemovable(Instruction *I) {
     case Intrinsic::memset:
     case Intrinsic::memmove:
     case Intrinsic::memcpy:
+    case Intrinsic::memcpy_inline:
       // Don't remove volatile memory intrinsics.
       return !cast<MemIntrinsic>(II)->isVolatile();
     case Intrinsic::memcpy_element_unordered_atomic:

diff --git a/llvm/test/Transforms/DeadStoreElimination/MSSA/memset-and-memcpy.ll b/llvm/test/Transforms/DeadStoreElimination/MSSA/memset-and-memcpy.ll
index 5aeda18..02fc8f2 100644 (file)
--- a/llvm/test/Transforms/DeadStoreElimination/MSSA/memset-and-memcpy.ll
+++ b/llvm/test/Transforms/DeadStoreElimination/MSSA/memset-and-memcpy.ll
@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: opt < %s -basic-aa -dse -enable-dse-memoryssa -S | FileCheck %s
+; RUN: opt < %s -basic-aa -dse -enable-dse-memoryssa=false -S | FileCheck %s
 ; RUN: opt < %s -aa-pipeline=basic-aa -passes=dse -enable-dse-memoryssa -S | FileCheck %s
 target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-v64:64:64-v128:128:128"
 
@@ -91,3 +92,21 @@ define void @test18_atomic(i8* %P, i8* %Q, i8* %R) nounwind ssp {
   tail call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i64(i8* align 1 %P, i8* align 1 %R, i64 12, i32 1)
   ret void
 }
+
+define void @test_memset_memcpy_inline(i8* noalias %P, i8* noalias %Q) {
+  tail call void @llvm.memset.p0i8.i64(i8* %P, i8 42, i64 8, i1 false)
+  tail call void @llvm.memcpy.inline.p0i8.p0i8.i64(i8* align 1 %P, i8* align 1 %Q, i64 12, i1 false)
+ ret void
+}
+
+define void @test_store_memcpy_inline(i8* noalias %P, i8* noalias %Q) {
+  store i8 0, i8* %P
+  %P.1 = getelementptr i8, i8* %P, i64 1
+  store i8 1, i8* %P.1
+  %P.4 = getelementptr i8, i8* %P, i64 4
+  store i8 4, i8* %P.4
+  tail call void @llvm.memcpy.inline.p0i8.p0i8.i64(i8* align 1 %P, i8* align 1 %Q, i64 4, i1 false)
+ ret void
+}
+
+declare void @llvm.memcpy.inline.p0i8.p0i8.i64(i8* noalias nocapture writeonly, i8* noalias nocapture readonly, i64 immarg, i1 immarg)