14 files changed, 544 insertions, 54 deletions
diff --git a/test/Transforms/GVN/cond_br2.ll b/test/Transforms/GVN/cond_br2.ll
index 27e6f75..a7ca219 100644
--- a/test/Transforms/GVN/cond_br2.ll
+++ b/test/Transforms/GVN/cond_br2.ll
@@ -132,9 +132,9 @@ attributes #1 = { nounwind }
 attributes #2 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
 attributes #3 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
 
-!0 = metadata !{metadata !"any pointer", metadata !1}
-!1 = metadata !{metadata !"omnipotent char", metadata !2}
-!2 = metadata !{metadata !"Simple C/C++ TBAA"}
-!3 = metadata !{metadata !"int", metadata !1}
-!4 = metadata !{metadata !0, metadata !0, i64 0}
-!5 = metadata !{metadata !3, metadata !3, i64 0}
+!0 = !{!"any pointer", !1}
+!1 = !{!"omnipotent char", !2}
+!2 = !{!"Simple C/C++ TBAA"}
+!3 = !{!"int", !1}
+!4 = !{!0, !0, i64 0}
+!5 = !{!3, !3, i64 0}
diff --git a/test/Transforms/GVN/condprop.ll b/test/Transforms/GVN/condprop.ll
index 708e4b2..845f88e 100644
--- a/test/Transforms/GVN/condprop.ll
+++ b/test/Transforms/GVN/condprop.ll
@@ -144,6 +144,22 @@ different:
   ret i1 %cmp3
 }
 
+; CHECK-LABEL: @test6_fp(
+define i1 @test6_fp(float %x, float %y) {
+  %cmp2 = fcmp une float %x, %y
+  %cmp = fcmp oeq float %x, %y
+  %cmp3 = fcmp oeq float  %x, %y
+  br i1 %cmp, label %same, label %different
+
+same:
+; CHECK: ret i1 false
+  ret i1 %cmp2
+
+different:
+; CHECK: ret i1 false
+  ret i1 %cmp3
+}
+
 ; CHECK-LABEL: @test7(
 define i1 @test7(i32 %x, i32 %y) {
   %cmp = icmp sgt i32 %x, %y
@@ -160,6 +176,22 @@ different:
   ret i1 %cmp3
 }
 
+; CHECK-LABEL: @test7_fp(
+define i1 @test7_fp(float %x, float %y) {
+  %cmp = fcmp ogt float %x, %y
+  br i1 %cmp, label %same, label %different
+
+same:
+  %cmp2 = fcmp ule float %x, %y
+; CHECK: ret i1 false
+  ret i1 %cmp2
+
+different:
+  %cmp3 = fcmp ogt float %x, %y
+; CHECK: ret i1 false
+  ret i1 %cmp3
+}
+
 ; CHECK-LABEL: @test8(
 define i1 @test8(i32 %x, i32 %y) {
   %cmp2 = icmp sle i32 %x, %y
@@ -176,6 +208,22 @@ different:
   ret i1 %cmp3
 }
 
+; CHECK-LABEL: @test8_fp(
+define i1 @test8_fp(float %x, float %y) {
+  %cmp2 = fcmp ule float %x, %y
+  %cmp = fcmp ogt float %x, %y
+  %cmp3 = fcmp ogt float %x, %y
+  br i1 %cmp, label %same, label %different
+
+same:
+; CHECK: ret i1 false
+  ret i1 %cmp2
+
+different:
+; CHECK: ret i1 false
+  ret i1 %cmp3
+}
+
 ; PR1768
 ; CHECK-LABEL: @test9(
 define i32 @test9(i32 %i, i32 %j) {
diff --git a/test/Transforms/GVN/edge.ll b/test/Transforms/GVN/edge.ll
index 646e10c..0c1a3fb 100644
--- a/test/Transforms/GVN/edge.ll
+++ b/test/Transforms/GVN/edge.ll
@@ -58,3 +58,113 @@ bb2:
 ; CHECK: call void @g(i1 %y)
   ret void
 }
+
+define double @fcmp_oeq_not_zero(double %x, double %y) {
+entry:
+  %cmp = fcmp oeq double %y, 2.0
+  br i1 %cmp, label %if, label %return
+
+if:
+  %div = fdiv double %x, %y
+  br label %return
+
+return:
+  %retval = phi double [ %div, %if ], [ %x, %entry ]
+  ret double %retval
+
+; CHECK-LABEL: define double @fcmp_oeq_not_zero(
+; CHECK: %div = fdiv double %x, 2.0
+}
+
+define double @fcmp_une_not_zero(double %x, double %y) {
+entry:
+  %cmp = fcmp une double %y, 2.0
+  br i1 %cmp, label %return, label %else
+
+else:
+  %div = fdiv double %x, %y
+  br label %return
+
+return:
+  %retval = phi double [ %div, %else ], [ %x, %entry ]
+  ret double %retval
+
+; CHECK-LABEL: define double @fcmp_une_not_zero(
+; CHECK: %div = fdiv double %x, 2.0
+}
+
+; PR22376 - We can't propagate zero constants because -0.0 
+; compares equal to 0.0. If %y is -0.0 in this test case,
+; we would produce the wrong sign on the infinity return value.
+define double @fcmp_oeq_zero(double %x, double %y) {
+entry:
+  %cmp = fcmp oeq double %y, 0.0
+  br i1 %cmp, label %if, label %return
+
+if:
+  %div = fdiv double %x, %y
+  br label %return
+
+return:
+  %retval = phi double [ %div, %if ], [ %x, %entry ]
+  ret double %retval
+
+; CHECK-LABEL: define double @fcmp_oeq_zero(
+; CHECK: %div = fdiv double %x, %y
+}
+
+define double @fcmp_une_zero(double %x, double %y) {
+entry:
+  %cmp = fcmp une double %y, -0.0
+  br i1 %cmp, label %return, label %else
+
+else:
+  %div = fdiv double %x, %y
+  br label %return
+
+return:
+  %retval = phi double [ %div, %else ], [ %x, %entry ]
+  ret double %retval
+
+; CHECK-LABEL: define double @fcmp_une_zero(
+; CHECK: %div = fdiv double %x, %y
+}
+
+; We also cannot propagate a value if it's not a constant.
+; This is because the value could be 0.0 or -0.0.
+
+define double @fcmp_oeq_maybe_zero(double %x, double %y, double %z1, double %z2) {
+entry:
+ %z = fadd double %z1, %z2
+ %cmp = fcmp oeq double %y, %z
+ br i1 %cmp, label %if, label %return
+
+if:
+ %div = fdiv double %x, %z
+ br label %return
+
+return:
+ %retval = phi double [ %div, %if ], [ %x, %entry ]
+ ret double %retval
+
+; CHECK-LABEL: define double @fcmp_oeq_maybe_zero(
+; CHECK: %div = fdiv double %x, %z
+}
+
+define double @fcmp_une_maybe_zero(double %x, double %y, double %z1, double %z2) {
+entry:
+ %z = fadd double %z1, %z2
+ %cmp = fcmp une double %y, %z
+ br i1 %cmp, label %return, label %else
+
+else:
+ %div = fdiv double %x, %z
+ br label %return
+
+return:
+ %retval = phi double [ %div, %else ], [ %x, %entry ]
+ ret double %retval
+
+; CHECK-LABEL: define double @fcmp_une_maybe_zero(
+; CHECK: %div = fdiv double %x, %z
+}
diff --git a/test/Transforms/GVN/fpmath.ll b/test/Transforms/GVN/fpmath.ll
index 403df5c..d164fb5 100644
--- a/test/Transforms/GVN/fpmath.ll
+++ b/test/Transforms/GVN/fpmath.ll
@@ -41,5 +41,5 @@ define double @test4(double %x, double %y) {
   ret double %foo
 }
 
-!0 = metadata !{ float 5.0 }
-!1 = metadata !{ float 2.5 }
+!0 = !{ float 5.0 }
+!1 = !{ float 2.5 }
diff --git a/test/Transforms/GVN/invariant-load.ll b/test/Transforms/GVN/invariant-load.ll
index 80e2226..2a83c45 100644
--- a/test/Transforms/GVN/invariant-load.ll
+++ b/test/Transforms/GVN/invariant-load.ll
@@ -27,5 +27,43 @@ entry:
   ret i32 %add
 }
 
-!0 = metadata !{ }
+; With the invariant.load metadata, what would otherwise
+; be a case for PRE becomes a full redundancy.
+define i32 @test3(i1 %cnd, i32* %p, i32* %q) {
+; CHECK-LABEL: test3
+; CHECK-NOT: load
+entry:
+  %v1 = load i32* %p
+  br i1 %cnd, label %bb1, label %bb2
+
+bb1:
+  store i32 5, i32* %q
+  br label %bb2
+
+bb2:
+  %v2 = load i32* %p, !invariant.load !0
+  %res = sub i32 %v1, %v2
+  ret i32 %res
+}
+
+; This test is here to document a case which doesn't optimize
+; as well as it could.  
+define i32 @test4(i1 %cnd, i32* %p, i32* %q) {
+; CHECK-LABEL: test4
+; %v2 is redundant, but GVN currently doesn't catch that
+entry:
+  %v1 = load i32* %p, !invariant.load !0
+  br i1 %cnd, label %bb1, label %bb2
+
+bb1:
+  store i32 5, i32* %q
+  br label %bb2
+
+bb2:
+  %v2 = load i32* %p
+  %res = sub i32 %v1, %v2
+  ret i32 %res
+}
+
+!0 = !{ }
 
diff --git a/test/Transforms/GVN/load-from-unreachable-predecessor.ll b/test/Transforms/GVN/load-from-unreachable-predecessor.ll
new file mode 100644
index 0000000..b676d95
--- /dev/null
+++ b/test/Transforms/GVN/load-from-unreachable-predecessor.ll
@@ -0,0 +1,20 @@
+; RUN: opt -gvn -S < %s | FileCheck %s
+
+; Check that an unreachable predecessor to a PHI node doesn't cause a crash.
+; PR21625.
+
+define i32 @f(i32** %f) {
+; CHECK: bb0:
+; Load should be removed, since it's ignored.
+; CHECK-NEXT: br label
+bb0:
+  %bar = load i32** %f
+  br label %bb2
+bb1:
+  %zed = load i32** %f
+  br i1 false, label %bb1, label %bb2
+bb2:
+  %foo = phi i32* [ null, %bb0 ], [ %zed, %bb1 ]
+  %storemerge = load i32* %foo
+  ret i32 %storemerge
+}
diff --git a/test/Transforms/GVN/load-pre-nonlocal.ll b/test/Transforms/GVN/load-pre-nonlocal.ll
index 7bac1b7..ae508b9 100644
--- a/test/Transforms/GVN/load-pre-nonlocal.ll
+++ b/test/Transforms/GVN/load-pre-nonlocal.ll
@@ -79,9 +79,9 @@ if.end:
   ret i32 %add1
 }
 
-!1 = metadata !{metadata !2, metadata !2, i64 0}
-!2 = metadata !{metadata !"any pointer", metadata !3, i64 0}
-!3 = metadata !{metadata !"omnipotent char", metadata !4, i64 0}
-!4 = metadata !{metadata !"Simple C/C++ TBAA"}
-!5 = metadata !{metadata !6, metadata !6, i64 0}
-!6 = metadata !{metadata !"int", metadata !3, i64 0}
+!1 = !{!2, !2, i64 0}
+!2 = !{!"any pointer", !3, i64 0}
+!3 = !{!"omnipotent char", !4, i64 0}
+!4 = !{!"Simple C/C++ TBAA"}
+!5 = !{!6, !6, i64 0}
+!6 = !{!"int", !3, i64 0}
diff --git a/test/Transforms/GVN/noalias.ll b/test/Transforms/GVN/noalias.ll
index a774f38..6c310fa 100644
--- a/test/Transforms/GVN/noalias.ll
+++ b/test/Transforms/GVN/noalias.ll
@@ -37,7 +37,7 @@ define i32 @test3(i32* %p, i32* %q) {
 
 declare i32 @foo(i32*) readonly
 
-!0 = metadata !{metadata !0}
-!1 = metadata !{metadata !1}
-!2 = metadata !{metadata !0, metadata !1}
+!0 = !{!0}
+!1 = !{!1}
+!2 = !{!0, !1}
 
diff --git a/test/Transforms/GVN/pre-gep-load.ll b/test/Transforms/GVN/pre-gep-load.ll
new file mode 100644
index 0000000..3ee3a37
--- /dev/null
+++ b/test/Transforms/GVN/pre-gep-load.ll
@@ -0,0 +1,49 @@
+; RUN: opt < %s -basicaa -gvn -enable-load-pre -S | FileCheck %s
+target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
+target triple = "aarch64--linux-gnu"
+
+define double @foo(i32 %stat, i32 %i, double** %p) {
+; CHECK-LABEL: @foo(
+entry:
+  switch i32 %stat, label %sw.default [
+    i32 0, label %sw.bb
+    i32 1, label %sw.bb
+    i32 2, label %sw.bb2
+  ]
+
+sw.bb:                                            ; preds = %entry, %entry
+  %idxprom = sext i32 %i to i64
+  %arrayidx = getelementptr inbounds double** %p, i64 0
+  %0 = load double** %arrayidx, align 8
+  %arrayidx1 = getelementptr inbounds double* %0, i64 %idxprom
+  %1 = load double* %arrayidx1, align 8
+  %sub = fsub double %1, 1.000000e+00
+  %cmp = fcmp olt double %sub, 0.000000e+00
+  br i1 %cmp, label %if.then, label %if.end
+
+if.then:                                          ; preds = %sw.bb
+  br label %return
+
+if.end:                                           ; preds = %sw.bb
+  br label %sw.bb2
+
+sw.bb2:                                           ; preds = %if.end, %entry
+  %idxprom3 = sext i32 %i to i64
+  %arrayidx4 = getelementptr inbounds double** %p, i64 0
+  %2 = load double** %arrayidx4, align 8
+  %arrayidx5 = getelementptr inbounds double* %2, i64 %idxprom3
+  %3 = load double* %arrayidx5, align 8
+; CHECK: sw.bb2:
+; CHECK-NEXT-NOT: sext
+; CHECK-NEXT: phi double [
+; CHECK-NOT: load
+  %sub6 = fsub double 3.000000e+00, %3
+  br label %return
+
+sw.default:                                       ; preds = %entry
+  br label %return
+
+return:                                           ; preds = %sw.default, %sw.bb2, %if.then
+  %retval.0 = phi double [ 0.000000e+00, %sw.default ], [ %sub6, %sw.bb2 ], [ %sub, %if.then ]
+  ret double %retval.0
+}
diff --git a/test/Transforms/GVN/pre-no-cost-phi.ll b/test/Transforms/GVN/pre-no-cost-phi.ll
new file mode 100644
index 0000000..4c5afa1
--- /dev/null
+++ b/test/Transforms/GVN/pre-no-cost-phi.ll
@@ -0,0 +1,31 @@
+; RUN: opt < %s -gvn -S | FileCheck %s
+; This testcase tests insertion of no-cost phis.  That is,
+; when the value is already available in every predecessor,
+; and we just need to insert a phi node to merge the available values.
+
+@c = global i32 0, align 4
+@d = global i32 0, align 4
+
+
+define i32 @mai(i32 %foo, i32 %a, i32 %b) {
+  %1 = icmp ne i32 %foo, 0
+  br i1 %1, label %bb1, label %bb2
+
+bb1:
+  %2 = add nsw i32 %a, %b
+  store i32 %2, i32* @c, align 4
+  br label %mergeblock
+
+bb2:
+  %3 = add nsw i32 %a, %b
+  store i32 %3, i32* @d, align 4
+  br label %mergeblock
+
+mergeblock:
+; CHECK: pre-phi = phi i32 [ %3, %bb2 ], [ %2, %bb1 ]
+; CHECK-NEXT: ret i32 %.pre-phi
+  %4 = add nsw i32 %a, %b
+  ret i32 %4
+}
+
+
diff --git a/test/Transforms/GVN/preserve-tbaa.ll b/test/Transforms/GVN/preserve-tbaa.ll
index c52ed96..587d463 100644
--- a/test/Transforms/GVN/preserve-tbaa.ll
+++ b/test/Transforms/GVN/preserve-tbaa.ll
@@ -25,7 +25,7 @@ for.end:                                          ; preds = %for.body, %entry
   ret void
 }
 
-!0 = metadata !{metadata !3, metadata !3, i64 0}
-!1 = metadata !{metadata !"omnipotent char", metadata !2}
-!2 = metadata !{metadata !"Simple C/C++ TBAA", null}
-!3 = metadata !{metadata !"short", metadata !1}
+!0 = !{!3, !3, i64 0}
+!1 = !{!"omnipotent char", !2}
+!2 = !{!"Simple C/C++ TBAA", null}
+!3 = !{!"short", !1}
diff --git a/test/Transforms/GVN/range.ll b/test/Transforms/GVN/range.ll
index 2115fe8..3720232 100644
--- a/test/Transforms/GVN/range.ll
+++ b/test/Transforms/GVN/range.ll
@@ -82,20 +82,20 @@ define i32 @test8(i32* %p) {
   ret i32 %c
 }
 
-; CHECK: ![[DISJOINT_RANGE]] = metadata !{i32 0, i32 2, i32 3, i32 5}
-; CHECK: ![[MERGED_RANGE]] = metadata !{i32 0, i32 5}
-; CHECK: ![[MERGED_SIGNED_RANGE]] = metadata !{i32 -3, i32 -2, i32 1, i32 2}
-; CHECK: ![[MERGED_TEST6]] = metadata !{i32 10, i32 1}
-; CHECK: ![[MERGED_TEST7]] = metadata !{i32 3, i32 4, i32 5, i32 2}
+; CHECK: ![[DISJOINT_RANGE]] = !{i32 0, i32 2, i32 3, i32 5}
+; CHECK: ![[MERGED_RANGE]] = !{i32 0, i32 5}
+; CHECK: ![[MERGED_SIGNED_RANGE]] = !{i32 -3, i32 -2, i32 1, i32 2}
+; CHECK: ![[MERGED_TEST6]] = !{i32 10, i32 1}
+; CHECK: ![[MERGED_TEST7]] = !{i32 3, i32 4, i32 5, i32 2}
 
-!0 = metadata !{i32 0, i32 2}
-!1 = metadata !{i32 3, i32 5}
-!2 = metadata !{i32 2, i32 5}
-!3 = metadata !{i32 -3, i32 -2}
-!4 = metadata !{i32 1, i32 2}
-!5 = metadata !{i32 10, i32 1}
-!6 = metadata !{i32 12, i32 13}
-!7 = metadata !{i32 1, i32 2, i32 3, i32 4}
-!8 = metadata !{i32 5, i32 1}
-!9 = metadata !{i32 1, i32 5}
-!10 = metadata !{i32 5, i32 1}
+!0 = !{i32 0, i32 2}
+!1 = !{i32 3, i32 5}
+!2 = !{i32 2, i32 5}
+!3 = !{i32 -3, i32 -2}
+!4 = !{i32 1, i32 2}
+!5 = !{i32 10, i32 1}
+!6 = !{i32 12, i32 13}
+!7 = !{i32 1, i32 2, i32 3, i32 4}
+!8 = !{i32 5, i32 1}
+!9 = !{i32 1, i32 5}
+!10 = !{i32 5, i32 1}
diff --git a/test/Transforms/GVN/tbaa.ll b/test/Transforms/GVN/tbaa.ll
index d6412fc..71fbed41 100644
--- a/test/Transforms/GVN/tbaa.ll
+++ b/test/Transforms/GVN/tbaa.ll
@@ -1,4 +1,4 @@
-; RUN: opt -basicaa -gvn -S < %s | FileCheck %s
+; RUN: opt -tbaa -basicaa -gvn -S < %s | FileCheck %s
 
 define i32 @test1(i8* %p, i8* %q) {
 ; CHECK: @test1(i8* %p, i8* %q)
@@ -72,20 +72,57 @@ define i32 @test7(i8* %p, i8* %q) {
   ret i32 %c
 }
 
+
+
+define i32 @test8(i32* %p, i32* %q) {
+; CHECK-LABEL: test8
+; CHECK-NEXT: store i32 15, i32* %p
+; CHECK-NEXT: ret i32 0
+; Since we know the location is invariant, we can forward the
+; load across the potentially aliasing store.
+
+  %a = load i32* %q, !tbaa !10
+  store i32 15, i32* %p
+  %b = load i32* %q, !tbaa !10
+  %c = sub i32 %a, %b
+  ret i32 %c
+}
+define i32 @test9(i32* %p, i32* %q) {
+; CHECK-LABEL: test9
+; CHECK-NEXT: call void @clobber()
+; CHECK-NEXT: ret i32 0
+; Since we know the location is invariant, we can forward the
+; load across the potentially aliasing store (within the call).
+
+  %a = load i32* %q, !tbaa !10
+  call void @clobber()
+  %b = load i32* %q, !tbaa !10
+  %c = sub i32 %a, %b
+  ret i32 %c
+}
+
+
+declare void @clobber()
 declare i32 @foo(i8*) readonly
 
-; CHECK: [[TAGC]] = metadata !{metadata [[TYPEC:!.*]], metadata [[TYPEC]], i64 0}
-; CHECK: [[TYPEC]] = metadata !{metadata !"C", metadata [[TYPEA:!.*]]}
-; CHECK: [[TYPEA]] = metadata !{metadata !"A", metadata !{{.*}}}
-; CHECK: [[TAGB]] = metadata !{metadata [[TYPEB:!.*]], metadata [[TYPEB]], i64 0}
-; CHECK: [[TYPEB]] = metadata !{metadata !"B", metadata [[TYPEA]]}
-; CHECK: [[TAGA]] = metadata !{metadata [[TYPEA]], metadata [[TYPEA]], i64 0}
-!0 = metadata !{metadata !5, metadata !5, i64 0}
-!1 = metadata !{metadata !6, metadata !6, i64 0}
-!2 = metadata !{metadata !"tbaa root", null}
-!3 = metadata !{metadata !7, metadata !7, i64 0}
-!4 = metadata !{metadata !8, metadata !8, i64 0}
-!5 = metadata !{metadata !"C", metadata !6}
-!6 = metadata !{metadata !"A", metadata !2}
-!7 = metadata !{metadata !"B", metadata !6}
-!8 = metadata !{metadata !"another root", null}
+; CHECK: [[TAGC]] = !{[[TYPEC:!.*]], [[TYPEC]], i64 0}
+; CHECK: [[TYPEC]] = !{!"C", [[TYPEA:!.*]]}
+; CHECK: [[TYPEA]] = !{!"A", !{{.*}}}
+; CHECK: [[TAGB]] = !{[[TYPEB:!.*]], [[TYPEB]], i64 0}
+; CHECK: [[TYPEB]] = !{!"B", [[TYPEA]]}
+; CHECK: [[TAGA]] = !{[[TYPEA]], [[TYPEA]], i64 0}
+!0 = !{!5, !5, i64 0}
+!1 = !{!6, !6, i64 0}
+!2 = !{!"tbaa root", null}
+!3 = !{!7, !7, i64 0}
+!4 = !{!8, !8, i64 0}
+!5 = !{!"C", !6}
+!6 = !{!"A", !2}
+!7 = !{!"B", !6}
+!8 = !{!"another root", null}
+
+
+;; A TBAA structure who's only point is to have a constant location
+!9 = !{!"yet another root"}
+!10 = !{!"node", !9, i64 1}
+
diff --git a/test/Transforms/GVN/volatile.ll b/test/Transforms/GVN/volatile.ll
new file mode 100644
index 0000000..5ba03d9
--- /dev/null
+++ b/test/Transforms/GVN/volatile.ll
@@ -0,0 +1,157 @@
+; Tests that check our handling of volatile instructions encountered
+; when scanning for dependencies
+; RUN: opt -basicaa -gvn -S < %s | FileCheck %s
+
+; Check that we can bypass a volatile load when searching
+; for dependencies of a non-volatile load
+define i32 @test1(i32* nocapture %p, i32* nocapture %q) {
+; CHECK-LABEL: test1
+; CHECK:      %0 = load volatile i32* %q
+; CHECK-NEXT: ret i32 0
+entry:
+  %x = load i32* %p
+  load volatile i32* %q
+  %y = load i32* %p
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+; We can not value forward if the query instruction is 
+; volatile, this would be (in effect) removing the volatile load
+define i32 @test2(i32* nocapture %p, i32* nocapture %q) {
+; CHECK-LABEL: test2
+; CHECK:      %x = load i32* %p
+; CHECK-NEXT: %y = load volatile i32* %p
+; CHECK-NEXT: %add = sub i32 %y, %x
+entry:
+  %x = load i32* %p
+  %y = load volatile i32* %p
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+; If the query instruction is itself volatile, we *cannot*
+; reorder it even if p and q are noalias
+define i32 @test3(i32* noalias nocapture %p, i32* noalias nocapture %q) {
+; CHECK-LABEL: test3
+; CHECK:      %x = load i32* %p
+; CHECK-NEXT: %0 = load volatile i32* %q
+; CHECK-NEXT: %y = load volatile i32* %p
+entry:
+  %x = load i32* %p
+  load volatile i32* %q
+  %y = load volatile i32* %p
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+; If an encountered instruction is both volatile and ordered, 
+; we need to use the strictest ordering of either.  In this 
+; case, the ordering prevents forwarding.
+define i32 @test4(i32* noalias nocapture %p, i32* noalias nocapture %q) {
+; CHECK-LABEL: test4
+; CHECK:      %x = load i32* %p
+; CHECK-NEXT: %0 = load atomic volatile i32* %q seq_cst 
+; CHECK-NEXT: %y = load atomic i32* %p seq_cst
+entry:
+  %x = load i32* %p
+  load atomic volatile i32* %q seq_cst, align 4
+  %y = load atomic i32* %p seq_cst, align 4
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+; Value forwarding from a volatile load is perfectly legal
+define i32 @test5(i32* nocapture %p, i32* nocapture %q) {
+; CHECK-LABEL: test5
+; CHECK:      %x = load volatile i32* %p
+; CHECK-NEXT: ret i32 0
+entry:
+  %x = load volatile i32* %p
+  %y = load i32* %p
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+; Does cross block redundancy elimination work with volatiles?
+define i32 @test6(i32* noalias nocapture %p, i32* noalias nocapture %q) {
+; CHECK-LABEL: test6
+; CHECK:      %y1 = load i32* %p
+; CHECK-LABEL: header
+; CHECK:      %x = load volatile i32* %q
+; CHECK-NEXT: %add = sub i32 %y1, %x
+entry:
+  %y1 = load i32* %p
+  call void @use(i32 %y1)
+  br label %header
+header:
+  %x = load volatile i32* %q
+  %y = load i32* %p
+  %add = sub i32 %y, %x
+  %cnd = icmp eq i32 %add, 0
+  br i1 %cnd, label %exit, label %header
+exit:
+  ret i32 %add
+}
+
+; Does cross block PRE work with volatiles?
+define i32 @test7(i1 %c, i32* noalias nocapture %p, i32* noalias nocapture %q) {
+; CHECK-LABEL: test7
+; CHECK-LABEL: entry.header_crit_edge:
+; CHECK:       %y.pre = load i32* %p
+; CHECK-LABEL: skip:
+; CHECK:       %y1 = load i32* %p
+; CHECK-LABEL: header:
+; CHECK:      %y = phi i32
+; CHECK-NEXT: %x = load volatile i32* %q
+; CHECK-NEXT: %add = sub i32 %y, %x
+entry:
+  br i1 %c, label %header, label %skip
+skip:
+  %y1 = load i32* %p
+  call void @use(i32 %y1)
+  br label %header
+header:
+  %x = load volatile i32* %q
+  %y = load i32* %p
+  %add = sub i32 %y, %x
+  %cnd = icmp eq i32 %add, 0
+  br i1 %cnd, label %exit, label %header
+exit:
+  ret i32 %add
+}
+
+; Another volatile PRE case - two paths through a loop
+; load in preheader, one path read only, one not
+define i32 @test8(i1 %b, i1 %c, i32* noalias %p, i32* noalias %q) {
+; CHECK-LABEL: test8
+; CHECK-LABEL: entry
+; CHECK:       %y1 = load i32* %p
+; CHECK-LABEL: header:
+; CHECK:      %y = phi i32
+; CHECK-NEXT: %x = load volatile i32* %q
+; CHECK-NOT:  load
+; CHECK-LABEL: skip.header_crit_edge:
+; CHECK:       %y.pre = load i32* %p
+entry:
+  %y1 = load i32* %p
+  call void @use(i32 %y1)
+  br label %header
+header:
+  %x = load volatile i32* %q
+  %y = load i32* %p
+  call void @use(i32 %y)
+  br i1 %b, label %skip, label %header
+skip:
+  ; escaping the arguments is explicitly required since we marked 
+  ; them noalias
+  call void @clobber(i32* %p, i32* %q)
+  br i1 %c, label %header, label %exit
+exit:
+  %add = sub i32 %y, %x
+  ret i32 %add
+}
+
+declare void @use(i32) readonly
+declare void @clobber(i32* %p, i32* %q)
+