diff --git a/llvm/lib/Transforms/Utils/Local.cpp b/llvm/lib/Transforms/Utils/Local.cpp
--- a/llvm/lib/Transforms/Utils/Local.cpp
+++ b/llvm/lib/Transforms/Utils/Local.cpp
@@ -406,7 +406,13 @@
 
     // Lifetime intrinsics are dead when their right-hand is undef.
     if (II->isLifetimeStartOrEnd())
-      return isa<UndefValue>(II->getArgOperand(1));
+      return isa<UndefValue>(II->getArgOperand(1)) ||
+             llvm::all_of(II->getArgOperand(1)->uses(), [](Use &use) {
+               if (IntrinsicInst *intrinsicUse =
+                       dyn_cast<IntrinsicInst>(use.getUser()))
+                 return intrinsicUse->isLifetimeStartOrEnd();
+               return false;
+             });
 
     // Assumptions are dead if their condition is trivially true.  Guards on
     // true are operationally no-ops.  In the future we can consider more
diff --git a/llvm/test/Transforms/Attributor/memory_locations.ll b/llvm/test/Transforms/Attributor/memory_locations.ll
--- a/llvm/test/Transforms/Attributor/memory_locations.ll
+++ b/llvm/test/Transforms/Attributor/memory_locations.ll
@@ -388,14 +388,14 @@
   store i8 0, i8* %unknown
   ret void
 }
-define void @callerE(i8* %arg) {
+define i8* @callerE(i8* %arg) {
 ; CHECK: Function Attrs: argmemonly nounwind willreturn
 ; CHECK-LABEL: define {{[^@]+}}@callerE
-; CHECK-SAME: (i8* nocapture [[ARG:%.*]])
+; CHECK-SAME: (i8* returned [[ARG:%.*]])
 ; CHECK-NEXT:    call void @llvm.lifetime.start.p0i8(i64 4, i8* nocapture [[ARG]])
-; CHECK-NEXT:    ret void
+; CHECK-NEXT:    ret i8* [[ARG]]
 ;
   call void @llvm.lifetime.start.p0i8(i64 4, i8* %arg)
-  ret void
+  ret i8* %arg
 }
 
diff --git a/llvm/test/Transforms/Coroutines/coro-split-02.ll b/llvm/test/Transforms/Coroutines/coro-split-02.ll
--- a/llvm/test/Transforms/Coroutines/coro-split-02.ll
+++ b/llvm/test/Transforms/Coroutines/coro-split-02.ll
@@ -32,6 +32,7 @@
   %cast = bitcast i32* %testval to i8*
   call void @llvm.lifetime.start.p0i8(i64 4, i8* %cast)
   call void @llvm.lifetime.end.p0i8(i64 4, i8*  %cast)
+  %dummy = load i8, i8* %cast
   call void @print(i32 %val)
   br label %exit
 exit:
diff --git a/llvm/test/Transforms/DCE/basic.ll b/llvm/test/Transforms/DCE/basic.ll
--- a/llvm/test/Transforms/DCE/basic.ll
+++ b/llvm/test/Transforms/DCE/basic.ll
@@ -11,5 +11,23 @@
   ret void
 }
 
+declare void @llvm.lifetime.start.p0i8(i64, i8* nocapture) nounwind
+declare void @llvm.lifetime.end.p0i8(i64, i8* nocapture) nounwind
+
+; CHECK-LABEL: @test_lifetime
+define i32 @test_lifetime() {
+; Check that lifetime intrinsics are removed along with the pointer.
+; CHECK-NEXT: @llvm.dbg.value
+; CHECK-NEXT: @llvm.dbg.value
+; CHECK-NEXT: ret i32 0
+; CHECK-NOT: llvm.lifetime.start
+; CHECK-NOT: llvm.lifetime.end
+  %i = alloca i32, align 4
+  %i.ptr = bitcast i32* %i to i8*
+  call void @llvm.lifetime.start.p0i8(i64 -1, i8* %i.ptr)
+  call void @llvm.lifetime.end.p0i8(i64 -1, i8* %i.ptr)
+  ret i32 0
+}
+
 ; CHECK: [[add]] = !DILocalVariable
 ; CHECK: [[sub]] = !DILocalVariable
diff --git a/llvm/test/Transforms/DeadStoreElimination/lifetime.ll b/llvm/test/Transforms/DeadStoreElimination/lifetime.ll
--- a/llvm/test/Transforms/DeadStoreElimination/lifetime.ll
+++ b/llvm/test/Transforms/DeadStoreElimination/lifetime.ll
@@ -12,7 +12,6 @@
 
   store i8 0, i8* %A  ;; Written to by memset
   call void @llvm.lifetime.end.p0i8(i64 1, i8* %A)
-; CHECK: lifetime.end
 
   call void @llvm.memset.p0i8.i8(i8* %A, i8 0, i8 -1, i1 false)
 ; CHECK-NOT: memset
@@ -26,11 +25,9 @@
   %Q = getelementptr i32, i32* %P, i32 1
   %R = bitcast i32* %Q to i8*
   call void @llvm.lifetime.start.p0i8(i64 4, i8* %R)
-; CHECK: lifetime.start
   store i32 0, i32* %Q  ;; This store is dead.
 ; CHECK-NOT: store
   call void @llvm.lifetime.end.p0i8(i64 4, i8* %R)
-; CHECK: lifetime.end
   ret void
 }
 
diff --git a/llvm/test/Transforms/Inline/byval-tail-call.ll b/llvm/test/Transforms/Inline/byval-tail-call.ll
--- a/llvm/test/Transforms/Inline/byval-tail-call.ll
+++ b/llvm/test/Transforms/Inline/byval-tail-call.ll
@@ -17,8 +17,7 @@
 
 define void @foo(i32* %x) {
 ; CHECK-LABEL: define void @foo(
-; CHECK: llvm.lifetime.start
-; CHECK: store i32 %2, i32* %x
+; CHECK: store i32 %1, i32* %x
   call void @bar(i32* byval %x)
   ret void
 }
diff --git a/llvm/test/Transforms/InstCombine/lower-dbg-declare.ll b/llvm/test/Transforms/InstCombine/lower-dbg-declare.ll
--- a/llvm/test/Transforms/InstCombine/lower-dbg-declare.ll
+++ b/llvm/test/Transforms/InstCombine/lower-dbg-declare.ll
@@ -35,7 +35,7 @@
   br label %while.cond, !dbg !22
 
 while.cond:                                       ; preds = %while.body, %entry
-; CHECK: dbg.value(metadata i32 %1, metadata [[METADATA_IDX1]], metadata !DIExpression())
+; CHECK: dbg.value(metadata i32 %0, metadata [[METADATA_IDX1]], metadata !DIExpression())
 ; CHECK-NEXT: call zeroext i1 @_ZL5emptyi
   %1 = load i32, i32* %d1, align 4, !dbg !22
   %call = call zeroext i1 @_ZL5emptyi(i32 %1), !dbg !22
diff --git a/llvm/test/Transforms/InstCombine/vararg.ll b/llvm/test/Transforms/InstCombine/vararg.ll
--- a/llvm/test/Transforms/InstCombine/vararg.ll
+++ b/llvm/test/Transforms/InstCombine/vararg.ll
@@ -1,4 +1,4 @@
-; RUN: opt < %s -instcombine -instcombine-infinite-loop-threshold=2 -S | FileCheck %s
+; RUN: opt < %s -instcombine -instcombine-infinite-loop-threshold=3 -S | FileCheck %s
 
 %struct.__va_list = type { i8*, i8*, i8*, i32, i32 }
 
diff --git a/llvm/test/Transforms/OpenMP/parallel_deletion.ll b/llvm/test/Transforms/OpenMP/parallel_deletion.ll
--- a/llvm/test/Transforms/OpenMP/parallel_deletion.ll
+++ b/llvm/test/Transforms/OpenMP/parallel_deletion.ll
@@ -141,15 +141,11 @@
 ; CHECK-LABEL: define {{[^@]+}}@delete_parallel_2()
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[A:%.*]] = alloca i32, align 4
-; CHECK-NEXT:    [[TMP:%.*]] = bitcast i32* [[A]] to i8*
-; CHECK-NEXT:    call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull align 4 dereferenceable(4) [[TMP]]) #0
 ; CHECK-NEXT:    store i32 0, i32* [[A]], align 4
 ; CHECK-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* nonnull align 8 dereferenceable(24) @0, i32 1, void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*)* @.omp_outlined..3 to void (i32*, i32*, ...)*), i32* nocapture nofree nonnull align 4 dereferenceable(4) [[A]])
 ; CHECK-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* nonnull align 8 dereferenceable(24) @0, i32 1, void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*)* @.omp_outlined..4 to void (i32*, i32*, ...)*), i32* nocapture nonnull align 4 dereferenceable(4) [[A]])
 ; CHECK-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* nonnull align 8 dereferenceable(24) @0, i32 1, void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*)* @.omp_outlined..5 to void (i32*, i32*, ...)*), i32* nocapture nonnull align 4 dereferenceable(4) [[A]])
 ; CHECK-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* nonnull align 8 dereferenceable(24) @0, i32 1, void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*)* @.omp_outlined..6 to void (i32*, i32*, ...)*), i32* nocapture nonnull align 4 dereferenceable(4) [[A]])
-; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[A]] to i8*
-; CHECK-NEXT:    call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[TMP1]])
 ; CHECK-NEXT:    ret void
 ;
 entry: