From: Nikita Popov Date: Tue, 27 Dec 2022 15:22:10 +0000 (+0100) Subject: [InstCombine] Convert some tests to opaque pointers (NFC) X-Git-Tag: upstream/17.0.6~22568 X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=c93e7dec7e8075e717528555e05991d2e75a0648;p=platform%2Fupstream%2Fllvm.git [InstCombine] Convert some tests to opaque pointers (NFC) Check lines for these were regenerated, but without any significant changes (mostly different GEP source element types). --- diff --git a/llvm/test/Transforms/InstCombine/2008-04-22-ByValBitcast.ll b/llvm/test/Transforms/InstCombine/2008-04-22-ByValBitcast.ll index 15ccce5..8433761 100644 --- a/llvm/test/Transforms/InstCombine/2008-04-22-ByValBitcast.ll +++ b/llvm/test/Transforms/InstCombine/2008-04-22-ByValBitcast.ll @@ -7,16 +7,14 @@ target triple = "i386-apple-darwin9" %struct.NSRect = type { [4 x float] } -define void @foo(i8* %context) nounwind { +define void @foo(ptr %context) nounwind { ; CHECK-LABEL: @foo( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[TMP1:%.*]] = bitcast i8* [[CONTEXT:%.*]] to %struct.NSRect* -; CHECK-NEXT: call void (i32, ...) @bar(i32 3, %struct.NSRect* byval([[STRUCT_NSRECT:%.*]]) align 4 [[TMP1]]) #[[ATTR0:[0-9]+]] +; CHECK-NEXT: call void (i32, ...) @bar(i32 3, ptr byval([[STRUCT_NSRECT:%.*]]) align 4 [[CONTEXT:%.*]]) #[[ATTR0:[0-9]+]] ; CHECK-NEXT: ret void ; entry: - %tmp1 = bitcast i8* %context to %struct.NSRect* ; <%struct.NSRect*> [#uses=1] - call void (i32, ...) @bar( i32 3, %struct.NSRect* byval(%struct.NSRect) align 4 %tmp1 ) nounwind + call void (i32, ...) @bar( i32 3, ptr byval(%struct.NSRect) align 4 %context ) nounwind ret void } diff --git a/llvm/test/Transforms/InstCombine/compare-alloca.ll b/llvm/test/Transforms/InstCombine/compare-alloca.ll index 6201b34..6bea1ad 100644 --- a/llvm/test/Transforms/InstCombine/compare-alloca.ll +++ b/llvm/test/Transforms/InstCombine/compare-alloca.ll @@ -3,119 +3,118 @@ target datalayout = "p:32:32" -define i1 @alloca_argument_compare(i64* %arg) { +define i1 @alloca_argument_compare(ptr %arg) { ; CHECK-LABEL: @alloca_argument_compare( ; CHECK-NEXT: ret i1 false ; %alloc = alloca i64 - %cmp = icmp eq i64* %arg, %alloc + %cmp = icmp eq ptr %arg, %alloc ret i1 %cmp } -define i1 @alloca_argument_compare_swapped(i64* %arg) { +define i1 @alloca_argument_compare_swapped(ptr %arg) { ; CHECK-LABEL: @alloca_argument_compare_swapped( ; CHECK-NEXT: ret i1 false ; %alloc = alloca i64 - %cmp = icmp eq i64* %alloc, %arg + %cmp = icmp eq ptr %alloc, %arg ret i1 %cmp } -define i1 @alloca_argument_compare_ne(i64* %arg) { +define i1 @alloca_argument_compare_ne(ptr %arg) { ; CHECK-LABEL: @alloca_argument_compare_ne( ; CHECK-NEXT: ret i1 true ; %alloc = alloca i64 - %cmp = icmp ne i64* %arg, %alloc + %cmp = icmp ne ptr %arg, %alloc ret i1 %cmp } -define i1 @alloca_argument_compare_derived_ptrs(i64* %arg, i64 %x) { +define i1 @alloca_argument_compare_derived_ptrs(ptr %arg, i64 %x) { ; CHECK-LABEL: @alloca_argument_compare_derived_ptrs( ; CHECK-NEXT: ret i1 false ; %alloc = alloca i64, i64 8 - %p = getelementptr i64, i64* %arg, i64 %x - %q = getelementptr i64, i64* %alloc, i64 3 - %cmp = icmp eq i64* %p, %q + %p = getelementptr i64, ptr %arg, i64 %x + %q = getelementptr i64, ptr %alloc, i64 3 + %cmp = icmp eq ptr %p, %q ret i1 %cmp } -declare void @escape(i64*) -define i1 @alloca_argument_compare_escaped_alloca(i64* %arg) { +declare void @escape(ptr) +define i1 @alloca_argument_compare_escaped_alloca(ptr %arg) { ; CHECK-LABEL: @alloca_argument_compare_escaped_alloca( ; CHECK-NEXT: [[ALLOC:%.*]] = alloca i64, align 8 -; CHECK-NEXT: call void @escape(i64* nonnull [[ALLOC]]) -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i64* [[ALLOC]], [[ARG:%.*]] +; CHECK-NEXT: call void @escape(ptr nonnull [[ALLOC]]) +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[ALLOC]], [[ARG:%.*]] ; CHECK-NEXT: ret i1 [[CMP]] ; %alloc = alloca i64 - call void @escape(i64* %alloc) - %cmp = icmp eq i64* %alloc, %arg + call void @escape(ptr %alloc) + %cmp = icmp eq ptr %alloc, %arg ret i1 %cmp } declare void @check_compares(i1, i1) -define void @alloca_argument_compare_two_compares(i64* %p) { +define void @alloca_argument_compare_two_compares(ptr %p) { ; CHECK-LABEL: @alloca_argument_compare_two_compares( ; CHECK-NEXT: [[Q1:%.*]] = alloca [8 x i64], align 8 -; CHECK-NEXT: [[Q1_SUB:%.*]] = getelementptr inbounds [8 x i64], [8 x i64]* [[Q1]], i32 0, i32 0 -; CHECK-NEXT: [[R:%.*]] = getelementptr i64, i64* [[P:%.*]], i32 1 -; CHECK-NEXT: [[S:%.*]] = getelementptr inbounds [8 x i64], [8 x i64]* [[Q1]], i32 0, i32 2 -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i64* [[Q1_SUB]], [[P]] -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i64* [[R]], [[S]] +; CHECK-NEXT: [[R:%.*]] = getelementptr i64, ptr [[P:%.*]], i32 1 +; CHECK-NEXT: [[S:%.*]] = getelementptr inbounds i64, ptr [[Q1]], i32 2 +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[Q1]], [[P]] +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[R]], [[S]] ; CHECK-NEXT: call void @check_compares(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; %q = alloca i64, i64 8 - %r = getelementptr i64, i64* %p, i64 1 - %s = getelementptr i64, i64* %q, i64 2 - %cmp1 = icmp eq i64* %p, %q - %cmp2 = icmp eq i64* %r, %s + %r = getelementptr i64, ptr %p, i64 1 + %s = getelementptr i64, ptr %q, i64 2 + %cmp1 = icmp eq ptr %p, %q + %cmp2 = icmp eq ptr %r, %s call void @check_compares(i1 %cmp1, i1 %cmp2) ret void ; We will only fold if there is a single cmp. } -define i1 @alloca_argument_compare_escaped_through_store(i64* %arg, i64** %ptr) { +define i1 @alloca_argument_compare_escaped_through_store(ptr %arg, ptr %ptr) { ; CHECK-LABEL: @alloca_argument_compare_escaped_through_store( ; CHECK-NEXT: [[ALLOC:%.*]] = alloca i64, align 8 -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i64* [[ALLOC]], [[ARG:%.*]] -; CHECK-NEXT: [[P:%.*]] = getelementptr inbounds i64, i64* [[ALLOC]], i32 1 -; CHECK-NEXT: store i64* [[P]], i64** [[PTR:%.*]], align 4 +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[ALLOC]], [[ARG:%.*]] +; CHECK-NEXT: [[P:%.*]] = getelementptr inbounds i64, ptr [[ALLOC]], i32 1 +; CHECK-NEXT: store ptr [[P]], ptr [[PTR:%.*]], align 4 ; CHECK-NEXT: ret i1 [[CMP]] ; %alloc = alloca i64 - %cmp = icmp eq i64* %alloc, %arg - %p = getelementptr i64, i64* %alloc, i64 1 - store i64* %p, i64** %ptr + %cmp = icmp eq ptr %alloc, %arg + %p = getelementptr i64, ptr %alloc, i64 1 + store ptr %p, ptr %ptr ret i1 %cmp } -declare void @llvm.lifetime.start.p0i8(i64, i8* nocapture) -declare void @llvm.lifetime.end.p0i8(i64, i8* nocapture) -define i1 @alloca_argument_compare_benign_instrs(i8* %arg) { +declare void @llvm.lifetime.start.p0(i64, ptr nocapture) +declare void @llvm.lifetime.end.p0(i64, ptr nocapture) +define i1 @alloca_argument_compare_benign_instrs(ptr %arg) { ; CHECK-LABEL: @alloca_argument_compare_benign_instrs( ; CHECK-NEXT: ret i1 false ; %alloc = alloca i8 - call void @llvm.lifetime.start.p0i8(i64 1, i8* %alloc) - %cmp = icmp eq i8* %arg, %alloc - %x = load i8, i8* %arg - store i8 %x, i8* %alloc - call void @llvm.lifetime.end.p0i8(i64 1, i8* %alloc) + call void @llvm.lifetime.start.p0(i64 1, ptr %alloc) + %cmp = icmp eq ptr %arg, %alloc + %x = load i8, ptr %arg + store i8 %x, ptr %alloc + call void @llvm.lifetime.end.p0(i64 1, ptr %alloc) ret i1 %cmp } -declare i64* @allocator() +declare ptr @allocator() define i1 @alloca_call_compare() { ; CHECK-LABEL: @alloca_call_compare( -; CHECK-NEXT: [[Q:%.*]] = call i64* @allocator() +; CHECK-NEXT: [[Q:%.*]] = call ptr @allocator() ; CHECK-NEXT: ret i1 false ; %p = alloca i64 - %q = call i64* @allocator() - %cmp = icmp eq i64* %p, %q + %q = call ptr @allocator() + %cmp = icmp eq ptr %p, %q ret i1 %cmp } @@ -129,16 +128,16 @@ define i1 @alloca_call_compare() { ; These two functions represents either a) forging a pointer via inttoptr or ; b) indexing off an adjacent allocation. In either case, the operation is ; obscured by an uninlined helper and not visible to instcombine. -declare i8* @hidden_inttoptr() -declare i8* @hidden_offset(i8* %other) +declare ptr @hidden_inttoptr() +declare ptr @hidden_offset(ptr %other) define i1 @ptrtoint_single_cmp() { ; CHECK-LABEL: @ptrtoint_single_cmp( ; CHECK-NEXT: ret i1 false ; %m = alloca i8, i32 4 - %rhs = inttoptr i64 2048 to i8* - %cmp = icmp eq i8* %m, %rhs + %rhs = inttoptr i64 2048 to ptr + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } @@ -148,8 +147,8 @@ define i1 @offset_single_cmp() { ; %m = alloca i8, i32 4 %n = alloca i8, i32 4 - %rhs = getelementptr i8, i8* %n, i32 4 - %cmp = icmp eq i8* %m, %rhs + %rhs = getelementptr i8, ptr %n, i32 4 + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } @@ -158,18 +157,17 @@ declare void @witness(i1, i1) define void @neg_consistent_fold1() { ; CHECK-LABEL: @neg_consistent_fold1( ; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 -; CHECK-NEXT: [[M1_SUB:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[M1]], i32 0, i32 0 -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_inttoptr() -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i8* [[M1_SUB]], inttoptr (i64 2048 to i8*) -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[M1_SUB]], [[RHS2]] +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_inttoptr() +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M1]], inttoptr (i64 2048 to ptr) +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M1]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; %m = alloca i8, i32 4 - %rhs = inttoptr i64 2048 to i8* - %rhs2 = call i8* @hidden_inttoptr() - %cmp1 = icmp eq i8* %m, %rhs - %cmp2 = icmp eq i8* %m, %rhs2 + %rhs = inttoptr i64 2048 to ptr + %rhs2 = call ptr @hidden_inttoptr() + %cmp1 = icmp eq ptr %m, %rhs + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } @@ -178,42 +176,38 @@ define void @neg_consistent_fold2() { ; CHECK-LABEL: @neg_consistent_fold2( ; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 ; CHECK-NEXT: [[N2:%.*]] = alloca [4 x i8], align 1 -; CHECK-NEXT: [[N2_SUB:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[N2]], i32 0, i32 0 -; CHECK-NEXT: [[M1_SUB:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[M1]], i32 0, i32 0 -; CHECK-NEXT: [[RHS:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[N2]], i32 0, i32 4 -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_offset(i8* nonnull [[N2_SUB]]) -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i8* [[M1_SUB]], [[RHS]] -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[M1_SUB]], [[RHS2]] +; CHECK-NEXT: [[RHS:%.*]] = getelementptr inbounds i8, ptr [[N2]], i32 4 +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_offset(ptr nonnull [[N2]]) +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M1]], [[RHS]] +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M1]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; %m = alloca i8, i32 4 %n = alloca i8, i32 4 - %rhs = getelementptr i8, i8* %n, i32 4 - %rhs2 = call i8* @hidden_offset(i8* %n) - %cmp1 = icmp eq i8* %m, %rhs - %cmp2 = icmp eq i8* %m, %rhs2 + %rhs = getelementptr i8, ptr %n, i32 4 + %rhs2 = call ptr @hidden_offset(ptr %n) + %cmp1 = icmp eq ptr %m, %rhs + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } define void @neg_consistent_fold3() { ; CHECK-LABEL: @neg_consistent_fold3( -; CHECK-NEXT: [[M1:%.*]] = alloca i32, align 1 -; CHECK-NEXT: [[M1_SUB:%.*]] = bitcast i32* [[M1]] to i8* -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8 -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_inttoptr() -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i32* [[M1]], [[LGP]] -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[RHS2]], [[M1_SUB]] +; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8 +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_inttoptr() +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M1]], [[LGP]] +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M1]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; %m = alloca i8, i32 4 - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %rhs2 = call i8* @hidden_inttoptr() - %cmp1 = icmp eq i32* %bc, %lgp - %cmp2 = icmp eq i8* %m, %rhs2 + %lgp = load ptr, ptr @gp, align 8 + %rhs2 = call ptr @hidden_inttoptr() + %cmp1 = icmp eq ptr %m, %lgp + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } @@ -224,10 +218,9 @@ define void @neg_consistent_fold4() { ; CHECK-NEXT: ret void ; %m = alloca i8, i32 4 - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %cmp1 = icmp eq i32* %bc, %lgp - %cmp2 = icmp eq i32* %bc, %lgp + %lgp = load ptr, ptr @gp, align 8 + %cmp1 = icmp eq ptr %m, %lgp + %cmp2 = icmp eq ptr %m, %lgp call void @witness(i1 %cmp1, i1 %cmp2) ret void } @@ -236,54 +229,50 @@ define void @neg_consistent_fold4() { ; assumption) not able to contain a comparison which might capture the ; address. -declare void @unknown(i8*) +declare void @unknown(ptr) ; TODO: Missing optimization define i1 @consistent_nocapture_inttoptr() { ; CHECK-LABEL: @consistent_nocapture_inttoptr( ; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 -; CHECK-NEXT: [[M1_SUB:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[M1]], i32 0, i32 0 -; CHECK-NEXT: call void @unknown(i8* nocapture nonnull [[M1_SUB]]) -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i8* [[M1_SUB]], inttoptr (i64 2048 to i8*) +; CHECK-NEXT: call void @unknown(ptr nocapture nonnull [[M1]]) +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[M1]], inttoptr (i64 2048 to ptr) ; CHECK-NEXT: ret i1 [[CMP]] ; %m = alloca i8, i32 4 - call void @unknown(i8* nocapture %m) - %rhs = inttoptr i64 2048 to i8* - %cmp = icmp eq i8* %m, %rhs + call void @unknown(ptr nocapture %m) + %rhs = inttoptr i64 2048 to ptr + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } define i1 @consistent_nocapture_offset() { ; CHECK-LABEL: @consistent_nocapture_offset( ; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 -; CHECK-NEXT: [[M1_SUB:%.*]] = getelementptr inbounds [4 x i8], [4 x i8]* [[M1]], i32 0, i32 0 -; CHECK-NEXT: call void @unknown(i8* nocapture nonnull [[M1_SUB]]) +; CHECK-NEXT: call void @unknown(ptr nocapture nonnull [[M1]]) ; CHECK-NEXT: ret i1 false ; %m = alloca i8, i32 4 - call void @unknown(i8* nocapture %m) + call void @unknown(ptr nocapture %m) %n = alloca i8, i32 4 - %rhs = getelementptr i8, i8* %n, i32 4 - %cmp = icmp eq i8* %m, %rhs + %rhs = getelementptr i8, ptr %n, i32 4 + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } -@gp = global i32* null, align 8 +@gp = global ptr null, align 8 ; TODO: Missing optimization define i1 @consistent_nocapture_through_global() { ; CHECK-LABEL: @consistent_nocapture_through_global( -; CHECK-NEXT: [[M1:%.*]] = alloca i32, align 1 -; CHECK-NEXT: [[M1_SUB:%.*]] = bitcast i32* [[M1]] to i8* -; CHECK-NEXT: call void @unknown(i8* nocapture nonnull [[M1_SUB]]) -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8, !nonnull !0 -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32* [[M1]], [[LGP]] +; CHECK-NEXT: [[M1:%.*]] = alloca [4 x i8], align 1 +; CHECK-NEXT: call void @unknown(ptr nocapture nonnull [[M1]]) +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8, !nonnull !0 +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[M1]], [[LGP]] ; CHECK-NEXT: ret i1 [[CMP]] ; %m = alloca i8, i32 4 - call void @unknown(i8* nocapture %m) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !{} - %cmp = icmp eq i32* %bc, %lgp + call void @unknown(ptr nocapture %m) + %lgp = load ptr, ptr @gp, align 8, !nonnull !{} + %cmp = icmp eq ptr %m, %lgp ret i1 %cmp } diff --git a/llvm/test/Transforms/InstCombine/compare-unescaped.ll b/llvm/test/Transforms/InstCombine/compare-unescaped.ll index 2d45904..e661cf9 100644 --- a/llvm/test/Transforms/InstCombine/compare-unescaped.ll +++ b/llvm/test/Transforms/InstCombine/compare-unescaped.ll @@ -1,18 +1,17 @@ ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py ; RUN: opt -passes=instcombine -S < %s | FileCheck %s -@gp = global i32* null, align 8 +@gp = global ptr null, align 8 -declare noalias i8* @malloc(i64) allockind("alloc,uninitialized") allocsize(0) +declare noalias ptr @malloc(i64) allockind("alloc,uninitialized") allocsize(0) define i1 @compare_global_trivialeq() { ; CHECK-LABEL: @compare_global_trivialeq( ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %cmp = icmp eq i32* %bc, %lgp + %m = call ptr @malloc(i64 4) + %lgp = load ptr, ptr @gp, align 8 + %cmp = icmp eq ptr %m, %lgp ret i1 %cmp } @@ -20,10 +19,9 @@ define i1 @compare_global_trivialne() { ; CHECK-LABEL: @compare_global_trivialne( ; CHECK-NEXT: ret i1 true ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %cmp = icmp ne i32* %bc, %lgp + %m = call ptr @malloc(i64 4) + %lgp = load ptr, ptr @gp, align 8 + %cmp = icmp ne ptr %m, %lgp ret i1 %cmp } @@ -34,75 +32,69 @@ define i1 @compare_global_trivialne() { declare void @f() define i1 @compare_and_call_with_deopt() { ; CHECK-LABEL: @compare_and_call_with_deopt( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) i8* @malloc(i64 24) -; CHECK-NEXT: tail call void @f() [ "deopt"(i8* [[M]]) ] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) ptr @malloc(i64 24) +; CHECK-NEXT: tail call void @f() [ "deopt"(ptr [[M]]) ] ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 24) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !0 - %cmp = icmp eq i32* %lgp, %bc - tail call void @f() [ "deopt"(i8* %m) ] + %m = call ptr @malloc(i64 24) + %lgp = load ptr, ptr @gp, align 8, !nonnull !0 + %cmp = icmp eq ptr %lgp, %m + tail call void @f() [ "deopt"(ptr %m) ] ret i1 %cmp } ; Same functon as above with deopt operand in function f, but comparison is NE define i1 @compare_ne_and_call_with_deopt() { ; CHECK-LABEL: @compare_ne_and_call_with_deopt( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) i8* @malloc(i64 24) -; CHECK-NEXT: tail call void @f() [ "deopt"(i8* [[M]]) ] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) ptr @malloc(i64 24) +; CHECK-NEXT: tail call void @f() [ "deopt"(ptr [[M]]) ] ; CHECK-NEXT: ret i1 true ; - %m = call i8* @malloc(i64 24) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !0 - %cmp = icmp ne i32* %lgp, %bc - tail call void @f() [ "deopt"(i8* %m) ] + %m = call ptr @malloc(i64 24) + %lgp = load ptr, ptr @gp, align 8, !nonnull !0 + %cmp = icmp ne ptr %lgp, %m + tail call void @f() [ "deopt"(ptr %m) ] ret i1 %cmp } ; Same function as above, but global not marked nonnull, and we cannot fold the comparison define i1 @compare_ne_global_maybe_null() { ; CHECK-LABEL: @compare_ne_global_maybe_null( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) i8* @malloc(i64 24) -; CHECK-NEXT: [[BC:%.*]] = bitcast i8* [[M]] to i32* -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8 -; CHECK-NEXT: [[CMP:%.*]] = icmp ne i32* [[LGP]], [[BC]] -; CHECK-NEXT: tail call void @f() [ "deopt"(i8* [[M]]) ] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) ptr @malloc(i64 24) +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8 +; CHECK-NEXT: [[CMP:%.*]] = icmp ne ptr [[LGP]], [[M]] +; CHECK-NEXT: tail call void @f() [ "deopt"(ptr [[M]]) ] ; CHECK-NEXT: ret i1 [[CMP]] ; - %m = call i8* @malloc(i64 24) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp - %cmp = icmp ne i32* %lgp, %bc - tail call void @f() [ "deopt"(i8* %m) ] + %m = call ptr @malloc(i64 24) + %lgp = load ptr, ptr @gp + %cmp = icmp ne ptr %lgp, %m + tail call void @f() [ "deopt"(ptr %m) ] ret i1 %cmp } ; FIXME: The comparison should fold to false since %m escapes (call to function escape) ; after the comparison. -declare void @escape(i8*) +declare void @escape(ptr) define i1 @compare_and_call_after() { ; CHECK-LABEL: @compare_and_call_after( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) i8* @malloc(i64 24) -; CHECK-NEXT: [[BC:%.*]] = bitcast i8* [[M]] to i32* -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8, !nonnull !0 -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32* [[LGP]], [[BC]] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(24) ptr @malloc(i64 24) +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8, !nonnull !0 +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[M]], [[LGP]] ; CHECK-NEXT: br i1 [[CMP]], label [[ESCAPE_CALL:%.*]], label [[JUST_RETURN:%.*]] ; CHECK: escape_call: -; CHECK-NEXT: call void @escape(i8* [[M]]) +; CHECK-NEXT: call void @escape(ptr [[M]]) ; CHECK-NEXT: ret i1 true ; CHECK: just_return: ; CHECK-NEXT: ret i1 [[CMP]] ; - %m = call i8* @malloc(i64 24) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !0 - %cmp = icmp eq i32* %bc, %lgp + %m = call ptr @malloc(i64 24) + %lgp = load ptr, ptr @gp, align 8, !nonnull !0 + %cmp = icmp eq ptr %m, %lgp br i1 %cmp, label %escape_call, label %just_return escape_call: - call void @escape(i8* %m) + call void @escape(ptr %m) ret i1 true just_return: @@ -113,9 +105,9 @@ define i1 @compare_distinct_mallocs() { ; CHECK-LABEL: @compare_distinct_mallocs( ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 4) - %n = call i8* @malloc(i64 4) - %cmp = icmp eq i8* %m, %n + %m = call ptr @malloc(i64 4) + %n = call ptr @malloc(i64 4) + %cmp = icmp eq ptr %m, %n ret i1 %cmp } @@ -125,10 +117,8 @@ define i1 @compare_samepointer_under_bitcast() { ; CHECK-LABEL: @compare_samepointer_under_bitcast( ; CHECK-NEXT: ret i1 true ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %bcback = bitcast i32* %bc to i8* - %cmp = icmp eq i8* %m, %bcback + %m = call ptr @malloc(i64 4) + %cmp = icmp eq ptr %m, %m ret i1 %cmp } @@ -136,15 +126,13 @@ define i1 @compare_samepointer_under_bitcast() { ; The malloc call for %m cannot be elided since it is used in the call to function f. define i1 @compare_samepointer_escaped() { ; CHECK-LABEL: @compare_samepointer_escaped( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: call void @f() [ "deopt"(i8* [[M]]) ] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: call void @f() [ "deopt"(ptr [[M]]) ] ; CHECK-NEXT: ret i1 true ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %bcback = bitcast i32* %bc to i8* - %cmp = icmp eq i8* %m, %bcback - call void @f() [ "deopt"(i8* %m) ] + %m = call ptr @malloc(i64 4) + %cmp = icmp eq ptr %m, %m + call void @f() [ "deopt"(ptr %m) ] ret i1 %cmp } @@ -154,52 +142,50 @@ define i1 @compare_samepointer_escaped() { ; FIXME: Folding this %cmp2 when %m escapes through ret could be an issue with ; cross-threading data dependencies since we do not make the distinction between ; atomic and non-atomic loads in capture tracking. -define i8* @compare_ret_escape(i8* %c) { +define ptr @compare_ret_escape(ptr %c) { ; CHECK-LABEL: @compare_ret_escape( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[N:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i8* [[N]], [[C:%.*]] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[N:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[N]], [[C:%.*]] ; CHECK-NEXT: br i1 [[CMP]], label [[RETST:%.*]], label [[CHK:%.*]] ; CHECK: retst: -; CHECK-NEXT: ret i8* [[M]] +; CHECK-NEXT: ret ptr [[M]] ; CHECK: chk: -; CHECK-NEXT: [[BC:%.*]] = bitcast i8* [[M]] to i32* -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8, !nonnull !0 -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i32* [[LGP]], [[BC]] +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8, !nonnull !0 +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M]], [[LGP]] ; CHECK-NEXT: br i1 [[CMP2]], label [[RETST]], label [[CHK2:%.*]] ; CHECK: chk2: -; CHECK-NEXT: ret i8* [[N]] +; CHECK-NEXT: ret ptr [[N]] ; - %m = call i8* @malloc(i64 4) - %n = call i8* @malloc(i64 4) - %cmp = icmp eq i8* %n, %c + %m = call ptr @malloc(i64 4) + %n = call ptr @malloc(i64 4) + %cmp = icmp eq ptr %n, %c br i1 %cmp, label %retst, label %chk retst: - ret i8* %m + ret ptr %m chk: - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !0 - %cmp2 = icmp eq i32* %bc, %lgp + %lgp = load ptr, ptr @gp, align 8, !nonnull !0 + %cmp2 = icmp eq ptr %m, %lgp br i1 %cmp2, label %retst, label %chk2 chk2: - ret i8* %n + ret ptr %n } ; The malloc call for %m cannot be elided since it is used in the call to function f. ; However, the cmp can be folded to true as %n doesnt escape and %m, %n are distinct allocations define i1 @compare_distinct_pointer_escape() { ; CHECK-LABEL: @compare_distinct_pointer_escape( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: tail call void @f() [ "deopt"(i8* [[M]]) ] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: tail call void @f() [ "deopt"(ptr [[M]]) ] ; CHECK-NEXT: ret i1 true ; - %m = call i8* @malloc(i64 4) - %n = call i8* @malloc(i64 4) - tail call void @f() [ "deopt"(i8* %m) ] - %cmp = icmp ne i8* %m, %n + %m = call ptr @malloc(i64 4) + %n = call ptr @malloc(i64 4) + tail call void @f() [ "deopt"(ptr %m) ] + %cmp = icmp ne ptr %m, %n ret i1 %cmp } @@ -213,19 +199,19 @@ define i1 @compare_distinct_pointer_escape() { ; These two functions represents either a) forging a pointer via inttoptr or ; b) indexing off an adjacent allocation. In either case, the operation is ; obscured by an uninlined helper and not visible to instcombine. -declare i8* @hidden_inttoptr() -declare i8* @hidden_offset(i8* %other) +declare ptr @hidden_inttoptr() +declare ptr @hidden_offset(ptr %other) ; FIXME: Missed oppurtunity define i1 @ptrtoint_single_cmp() { ; CHECK-LABEL: @ptrtoint_single_cmp( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i8* [[M]], inttoptr (i64 2048 to i8*) +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[M]], inttoptr (i64 2048 to ptr) ; CHECK-NEXT: ret i1 [[CMP]] ; - %m = call i8* @malloc(i64 4) - %rhs = inttoptr i64 2048 to i8* - %cmp = icmp eq i8* %m, %rhs + %m = call ptr @malloc(i64 4) + %rhs = inttoptr i64 2048 to ptr + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } @@ -233,10 +219,10 @@ define i1 @offset_single_cmp() { ; CHECK-LABEL: @offset_single_cmp( ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 4) - %n = call i8* @malloc(i64 4) - %rhs = getelementptr i8, i8* %n, i32 4 - %cmp = icmp eq i8* %m, %rhs + %m = call ptr @malloc(i64 4) + %n = call ptr @malloc(i64 4) + %rhs = getelementptr i8, ptr %n, i32 4 + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } @@ -244,60 +230,58 @@ declare void @witness(i1, i1) define void @neg_consistent_fold1() { ; CHECK-LABEL: @neg_consistent_fold1( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_inttoptr() -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i8* [[M]], inttoptr (i64 2048 to i8*) -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[M]], [[RHS2]] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_inttoptr() +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M]], inttoptr (i64 2048 to ptr) +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; - %m = call i8* @malloc(i64 4) - %rhs = inttoptr i64 2048 to i8* - %rhs2 = call i8* @hidden_inttoptr() - %cmp1 = icmp eq i8* %m, %rhs - %cmp2 = icmp eq i8* %m, %rhs2 + %m = call ptr @malloc(i64 4) + %rhs = inttoptr i64 2048 to ptr + %rhs2 = call ptr @hidden_inttoptr() + %cmp1 = icmp eq ptr %m, %rhs + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } define void @neg_consistent_fold2() { ; CHECK-LABEL: @neg_consistent_fold2( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[N:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[RHS:%.*]] = getelementptr i8, i8* [[N]], i64 4 -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_offset(i8* [[N]]) -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i8* [[M]], [[RHS]] -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[M]], [[RHS2]] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[N:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[RHS:%.*]] = getelementptr i8, ptr [[N]], i64 4 +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_offset(ptr [[N]]) +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M]], [[RHS]] +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; - %m = call i8* @malloc(i64 4) - %n = call i8* @malloc(i64 4) - %rhs = getelementptr i8, i8* %n, i32 4 - %rhs2 = call i8* @hidden_offset(i8* %n) - %cmp1 = icmp eq i8* %m, %rhs - %cmp2 = icmp eq i8* %m, %rhs2 + %m = call ptr @malloc(i64 4) + %n = call ptr @malloc(i64 4) + %rhs = getelementptr i8, ptr %n, i32 4 + %rhs2 = call ptr @hidden_offset(ptr %n) + %cmp1 = icmp eq ptr %m, %rhs + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } define void @neg_consistent_fold3() { ; CHECK-LABEL: @neg_consistent_fold3( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[BC:%.*]] = bitcast i8* [[M]] to i32* -; CHECK-NEXT: [[LGP:%.*]] = load i32*, i32** @gp, align 8 -; CHECK-NEXT: [[RHS2:%.*]] = call i8* @hidden_inttoptr() -; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i32* [[LGP]], [[BC]] -; CHECK-NEXT: [[CMP2:%.*]] = icmp eq i8* [[M]], [[RHS2]] +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[LGP:%.*]] = load ptr, ptr @gp, align 8 +; CHECK-NEXT: [[RHS2:%.*]] = call ptr @hidden_inttoptr() +; CHECK-NEXT: [[CMP1:%.*]] = icmp eq ptr [[M]], [[LGP]] +; CHECK-NEXT: [[CMP2:%.*]] = icmp eq ptr [[M]], [[RHS2]] ; CHECK-NEXT: call void @witness(i1 [[CMP1]], i1 [[CMP2]]) ; CHECK-NEXT: ret void ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %rhs2 = call i8* @hidden_inttoptr() - %cmp1 = icmp eq i32* %bc, %lgp - %cmp2 = icmp eq i8* %m, %rhs2 + %m = call ptr @malloc(i64 4) + %lgp = load ptr, ptr @gp, align 8 + %rhs2 = call ptr @hidden_inttoptr() + %cmp1 = icmp eq ptr %m, %lgp + %cmp2 = icmp eq ptr %m, %rhs2 call void @witness(i1 %cmp1, i1 %cmp2) ret void } @@ -310,16 +294,15 @@ define void @neg_consistent_fold4() { ; CHECK-NEXT: call void @witness(i1 false, i1 false) ; CHECK-NEXT: ret void ; - %m = call i8* @malloc(i64 4) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8 - %cmp1 = icmp eq i32* %bc, %lgp - %cmp2 = icmp eq i32* %bc, %lgp + %m = call ptr @malloc(i64 4) + %lgp = load ptr, ptr @gp, align 8 + %cmp1 = icmp eq ptr %m, %lgp + %cmp2 = icmp eq ptr %m, %lgp call void @witness(i1 %cmp1, i1 %cmp2) ret void } -declare void @unknown(i8*) +declare void @unknown(ptr) ; Points out that a nocapture call can't cause a consistent result issue ; as it is (by assumption) not able to contain a comparison which might @@ -327,43 +310,42 @@ declare void @unknown(i8*) define i1 @consistent_nocapture_inttoptr() { ; CHECK-LABEL: @consistent_nocapture_inttoptr( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: call void @unknown(i8* nocapture [[M]]) -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i8* [[M]], inttoptr (i64 2048 to i8*) +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: call void @unknown(ptr nocapture [[M]]) +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[M]], inttoptr (i64 2048 to ptr) ; CHECK-NEXT: ret i1 [[CMP]] ; - %m = call i8* @malloc(i64 4) - call void @unknown(i8* nocapture %m) - %rhs = inttoptr i64 2048 to i8* - %cmp = icmp eq i8* %m, %rhs + %m = call ptr @malloc(i64 4) + call void @unknown(ptr nocapture %m) + %rhs = inttoptr i64 2048 to ptr + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } define i1 @consistent_nocapture_offset() { ; CHECK-LABEL: @consistent_nocapture_offset( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: call void @unknown(i8* nocapture [[M]]) +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: call void @unknown(ptr nocapture [[M]]) ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 4) - call void @unknown(i8* nocapture %m) - %n = call i8* @malloc(i64 4) - %rhs = getelementptr i8, i8* %n, i32 4 - %cmp = icmp eq i8* %m, %rhs + %m = call ptr @malloc(i64 4) + call void @unknown(ptr nocapture %m) + %n = call ptr @malloc(i64 4) + %rhs = getelementptr i8, ptr %n, i32 4 + %cmp = icmp eq ptr %m, %rhs ret i1 %cmp } define i1 @consistent_nocapture_through_global() { ; CHECK-LABEL: @consistent_nocapture_through_global( -; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) i8* @malloc(i64 4) -; CHECK-NEXT: call void @unknown(i8* nocapture [[M]]) +; CHECK-NEXT: [[M:%.*]] = call dereferenceable_or_null(4) ptr @malloc(i64 4) +; CHECK-NEXT: call void @unknown(ptr nocapture [[M]]) ; CHECK-NEXT: ret i1 false ; - %m = call i8* @malloc(i64 4) - call void @unknown(i8* nocapture %m) - %bc = bitcast i8* %m to i32* - %lgp = load i32*, i32** @gp, align 8, !nonnull !0 - %cmp = icmp eq i32* %bc, %lgp + %m = call ptr @malloc(i64 4) + call void @unknown(ptr nocapture %m) + %lgp = load ptr, ptr @gp, align 8, !nonnull !0 + %cmp = icmp eq ptr %m, %lgp ret i1 %cmp } @@ -374,9 +356,9 @@ define i1 @two_nonnull_mallocs() { ; CHECK-LABEL: @two_nonnull_mallocs( ; CHECK-NEXT: ret i1 false ; - %m = call nonnull i8* @malloc(i64 4) - %n = call nonnull i8* @malloc(i64 4) - %cmp = icmp eq i8* %m, %n + %m = call nonnull ptr @malloc(i64 4) + %n = call nonnull ptr @malloc(i64 4) + %cmp = icmp eq ptr %m, %n ret i1 %cmp } @@ -384,32 +366,32 @@ define i1 @two_nonnull_mallocs() { ; the comparison non-equal. define i1 @two_nonnull_mallocs2() { ; CHECK-LABEL: @two_nonnull_mallocs2( -; CHECK-NEXT: [[N:%.*]] = call nonnull dereferenceable(4) i8* @malloc(i64 4) -; CHECK-NEXT: call void @unknown(i8* nonnull [[N]]) +; CHECK-NEXT: [[N:%.*]] = call nonnull dereferenceable(4) ptr @malloc(i64 4) +; CHECK-NEXT: call void @unknown(ptr nonnull [[N]]) ; CHECK-NEXT: ret i1 false ; - %m = call nonnull i8* @malloc(i64 4) - %n = call nonnull i8* @malloc(i64 4) - call void @unknown(i8* %n) - %cmp = icmp eq i8* %m, %n + %m = call nonnull ptr @malloc(i64 4) + %n = call nonnull ptr @malloc(i64 4) + call void @unknown(ptr %n) + %cmp = icmp eq ptr %m, %n ret i1 %cmp } ; TODO: We can fold this, but don't with the current scheme. define i1 @two_nonnull_mallocs_hidden() { ; CHECK-LABEL: @two_nonnull_mallocs_hidden( -; CHECK-NEXT: [[M:%.*]] = call nonnull dereferenceable(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[N:%.*]] = call nonnull dereferenceable(4) i8* @malloc(i64 4) -; CHECK-NEXT: [[GEP1:%.*]] = getelementptr i8, i8* [[M]], i64 1 -; CHECK-NEXT: [[GEP2:%.*]] = getelementptr i8, i8* [[N]], i64 2 -; CHECK-NEXT: [[CMP:%.*]] = icmp eq i8* [[GEP1]], [[GEP2]] +; CHECK-NEXT: [[M:%.*]] = call nonnull dereferenceable(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[N:%.*]] = call nonnull dereferenceable(4) ptr @malloc(i64 4) +; CHECK-NEXT: [[GEP1:%.*]] = getelementptr i8, ptr [[M]], i64 1 +; CHECK-NEXT: [[GEP2:%.*]] = getelementptr i8, ptr [[N]], i64 2 +; CHECK-NEXT: [[CMP:%.*]] = icmp eq ptr [[GEP1]], [[GEP2]] ; CHECK-NEXT: ret i1 [[CMP]] ; - %m = call nonnull i8* @malloc(i64 4) - %n = call nonnull i8* @malloc(i64 4) - %gep1 = getelementptr i8, i8* %m, i32 1 - %gep2 = getelementptr i8, i8* %n, i32 2 - %cmp = icmp eq i8* %gep1, %gep2 + %m = call nonnull ptr @malloc(i64 4) + %n = call nonnull ptr @malloc(i64 4) + %gep1 = getelementptr i8, ptr %m, i32 1 + %gep2 = getelementptr i8, ptr %n, i32 2 + %cmp = icmp eq ptr %gep1, %gep2 ret i1 %cmp } diff --git a/llvm/test/Transforms/InstCombine/gepphigep.ll b/llvm/test/Transforms/InstCombine/gepphigep.ll index b4a5660..22c0e5a 100644 --- a/llvm/test/Transforms/InstCombine/gepphigep.ll +++ b/llvm/test/Transforms/InstCombine/gepphigep.ll @@ -1,95 +1,87 @@ ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py ; RUN: opt -passes=instcombine -instcombine-infinite-loop-threshold=3 -S < %s | FileCheck %s -%struct1 = type { %struct2*, i32, i32, i32 } +%struct1 = type { ptr, i32, i32, i32 } %struct2 = type { i32, i32 } %struct3 = type { i32, %struct4, %struct4 } %struct4 = type { %struct2, %struct2 } -define i32 @test1(%struct1* %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19) { +define i32 @test1(ptr %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19) { ; CHECK-LABEL: @test1( ; CHECK-NEXT: bb: -; CHECK-NEXT: [[TMP:%.*]] = getelementptr inbounds [[STRUCT1:%.*]], %struct1* [[DM:%.*]], i64 0, i32 0 -; CHECK-NEXT: [[TMP1:%.*]] = load %struct2*, %struct2** [[TMP]], align 8 +; CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DM:%.*]], align 8 ; CHECK-NEXT: br i1 [[TMP4:%.*]], label [[BB1:%.*]], label [[BB2:%.*]] ; CHECK: bb1: -; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT2:%.*]], %struct2* [[TMP1]], i64 [[TMP9:%.*]], i32 0 -; CHECK-NEXT: store i32 0, i32* [[TMP11]], align 4 +; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT2:%.*]], ptr [[TMP1]], i64 [[TMP9:%.*]] +; CHECK-NEXT: store i32 0, ptr [[TMP10]], align 4 ; CHECK-NEXT: br label [[BB3:%.*]] ; CHECK: bb2: -; CHECK-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[STRUCT2]], %struct2* [[TMP1]], i64 [[TMP19:%.*]], i32 0 -; CHECK-NEXT: store i32 0, i32* [[TMP21]], align 4 +; CHECK-NEXT: [[TMP20:%.*]] = getelementptr inbounds [[STRUCT2]], ptr [[TMP1]], i64 [[TMP19:%.*]] +; CHECK-NEXT: store i32 0, ptr [[TMP20]], align 4 ; CHECK-NEXT: br label [[BB3]] ; CHECK: bb3: ; CHECK-NEXT: [[TMP0:%.*]] = phi i64 [ [[TMP9]], [[BB1]] ], [ [[TMP19]], [[BB2]] ] -; CHECK-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT2]], %struct2* [[TMP1]], i64 [[TMP0]], i32 1 -; CHECK-NEXT: [[TMP25:%.*]] = load i32, i32* [[TMP24]], align 4 +; CHECK-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT2]], ptr [[TMP1]], i64 [[TMP0]], i32 1 +; CHECK-NEXT: [[TMP25:%.*]] = load i32, ptr [[TMP24]], align 4 ; CHECK-NEXT: ret i32 [[TMP25]] ; bb: - %tmp = getelementptr inbounds %struct1, %struct1* %dm, i64 0, i32 0 - %tmp1 = load %struct2*, %struct2** %tmp, align 8 + %tmp1 = load ptr, ptr %dm, align 8 br i1 %tmp4, label %bb1, label %bb2 bb1: - %tmp10 = getelementptr inbounds %struct2, %struct2* %tmp1, i64 %tmp9 - %tmp11 = getelementptr inbounds %struct2, %struct2* %tmp10, i64 0, i32 0 - store i32 0, i32* %tmp11, align 4 + %tmp10 = getelementptr inbounds %struct2, ptr %tmp1, i64 %tmp9 + store i32 0, ptr %tmp10, align 4 br label %bb3 bb2: - %tmp20 = getelementptr inbounds %struct2, %struct2* %tmp1, i64 %tmp19 - %tmp21 = getelementptr inbounds %struct2, %struct2* %tmp20, i64 0, i32 0 - store i32 0, i32* %tmp21, align 4 + %tmp20 = getelementptr inbounds %struct2, ptr %tmp1, i64 %tmp19 + store i32 0, ptr %tmp20, align 4 br label %bb3 bb3: - %phi = phi %struct2* [ %tmp10, %bb1 ], [ %tmp20, %bb2 ] - %tmp24 = getelementptr inbounds %struct2, %struct2* %phi, i64 0, i32 1 - %tmp25 = load i32, i32* %tmp24, align 4 + %phi = phi ptr [ %tmp10, %bb1 ], [ %tmp20, %bb2 ] + %tmp24 = getelementptr inbounds %struct2, ptr %phi, i64 0, i32 1 + %tmp25 = load i32, ptr %tmp24, align 4 ret i32 %tmp25 } -define i32 @test2(%struct1* %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19) { +define i32 @test2(ptr %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19) { ; CHECK-LABEL: @test2( ; CHECK-NEXT: bb: -; CHECK-NEXT: [[TMP:%.*]] = getelementptr inbounds [[STRUCT1:%.*]], %struct1* [[DM:%.*]], i64 0, i32 0 -; CHECK-NEXT: [[TMP1:%.*]] = load %struct2*, %struct2** [[TMP]], align 8 -; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT2:%.*]], %struct2* [[TMP1]], i64 [[TMP9:%.*]], i32 0 -; CHECK-NEXT: store i32 0, i32* [[TMP11]], align 4 -; CHECK-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[STRUCT2]], %struct2* [[TMP1]], i64 [[TMP19:%.*]], i32 0 -; CHECK-NEXT: store i32 0, i32* [[TMP21]], align 4 -; CHECK-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT2]], %struct2* [[TMP1]], i64 [[TMP9]], i32 1 -; CHECK-NEXT: [[TMP25:%.*]] = load i32, i32* [[TMP24]], align 4 +; CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DM:%.*]], align 8 +; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT2:%.*]], ptr [[TMP1]], i64 [[TMP9:%.*]] +; CHECK-NEXT: store i32 0, ptr [[TMP10]], align 4 +; CHECK-NEXT: [[TMP20:%.*]] = getelementptr inbounds [[STRUCT2]], ptr [[TMP1]], i64 [[TMP19:%.*]] +; CHECK-NEXT: store i32 0, ptr [[TMP20]], align 4 +; CHECK-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT2]], ptr [[TMP1]], i64 [[TMP9]], i32 1 +; CHECK-NEXT: [[TMP25:%.*]] = load i32, ptr [[TMP24]], align 4 ; CHECK-NEXT: ret i32 [[TMP25]] ; bb: - %tmp = getelementptr inbounds %struct1, %struct1* %dm, i64 0, i32 0 - %tmp1 = load %struct2*, %struct2** %tmp, align 8 - %tmp10 = getelementptr inbounds %struct2, %struct2* %tmp1, i64 %tmp9 - %tmp11 = getelementptr inbounds %struct2, %struct2* %tmp10, i64 0, i32 0 - store i32 0, i32* %tmp11, align 4 - %tmp20 = getelementptr inbounds %struct2, %struct2* %tmp1, i64 %tmp19 - %tmp21 = getelementptr inbounds %struct2, %struct2* %tmp20, i64 0, i32 0 - store i32 0, i32* %tmp21, align 4 - %tmp24 = getelementptr inbounds %struct2, %struct2* %tmp10, i64 0, i32 1 - %tmp25 = load i32, i32* %tmp24, align 4 + %tmp1 = load ptr, ptr %dm, align 8 + %tmp10 = getelementptr inbounds %struct2, ptr %tmp1, i64 %tmp9 + store i32 0, ptr %tmp10, align 4 + %tmp20 = getelementptr inbounds %struct2, ptr %tmp1, i64 %tmp19 + store i32 0, ptr %tmp20, align 4 + %tmp24 = getelementptr inbounds %struct2, ptr %tmp10, i64 0, i32 1 + %tmp25 = load i32, ptr %tmp24, align 4 ret i32 %tmp25 } ; Check that instcombine doesn't insert GEPs before landingpad. -define i32 @test3(%struct3* %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19, i64 %tmp20, i64 %tmp21) personality i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*) { +define i32 @test3(ptr %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19, i64 %tmp20, i64 %tmp21) personality ptr @__gxx_personality_v0 { ; CHECK-LABEL: @test3( ; CHECK-NEXT: bb: ; CHECK-NEXT: br i1 [[TMP4:%.*]], label [[BB1:%.*]], label [[BB2:%.*]] ; CHECK: bb1: -; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT3:%.*]], %struct3* [[DM:%.*]], i64 [[TMP19:%.*]], i32 1, i32 0, i32 0 -; CHECK-NEXT: store i32 0, i32* [[TMP11]], align 4 +; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT3:%.*]], ptr [[DM:%.*]], i64 [[TMP19:%.*]], i32 1 +; CHECK-NEXT: store i32 0, ptr [[TMP1]], align 4 ; CHECK-NEXT: br label [[BB3:%.*]] ; CHECK: bb2: -; CHECK-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT3]], %struct3* [[DM]], i64 [[TMP20:%.*]], i32 1, i32 0, i32 1 -; CHECK-NEXT: store i32 0, i32* [[TMP12]], align 4 +; CHECK-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT3]], ptr [[DM]], i64 [[TMP20:%.*]], i32 1, i32 0, i32 1 +; CHECK-NEXT: store i32 0, ptr [[TMP12]], align 4 ; CHECK-NEXT: br label [[BB3]] ; CHECK: bb3: ; CHECK-NEXT: [[TMP0:%.*]] = phi i64 [ [[TMP19]], [[BB1]] ], [ [[TMP20]], [[BB2]] ] @@ -98,45 +90,43 @@ define i32 @test3(%struct3* %dm, i1 %tmp4, i64 %tmp9, i64 %tmp19, i64 %tmp20, i6 ; CHECK: bb4: ; CHECK-NEXT: ret i32 0 ; CHECK: bb5: -; CHECK-NEXT: [[TMP27:%.*]] = landingpad { i8*, i32 } -; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*) -; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT3]], %struct3* [[DM]], i64 [[TMP0]], i32 1 -; CHECK-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT4:%.*]], %struct4* [[TMP1]], i64 [[TMP21:%.*]], i32 1, i32 1 -; CHECK-NEXT: [[TMP25:%.*]] = load i32, i32* [[TMP35]], align 4 +; CHECK-NEXT: [[TMP27:%.*]] = landingpad { ptr, i32 } +; CHECK-NEXT: catch ptr @_ZTIi +; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT3]], ptr [[DM]], i64 [[TMP0]], i32 1 +; CHECK-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT4:%.*]], ptr [[TMP1]], i64 [[TMP21:%.*]], i32 1, i32 1 +; CHECK-NEXT: [[TMP25:%.*]] = load i32, ptr [[TMP35]], align 4 ; CHECK-NEXT: ret i32 [[TMP25]] ; bb: - %tmp = getelementptr inbounds %struct3, %struct3* %dm, i64 0 br i1 %tmp4, label %bb1, label %bb2 bb1: - %tmp1 = getelementptr inbounds %struct3, %struct3* %tmp, i64 %tmp19, i32 1 - %tmp11 = getelementptr inbounds %struct4, %struct4* %tmp1, i64 0, i32 0, i32 0 - store i32 0, i32* %tmp11, align 4 + %tmp1 = getelementptr inbounds %struct3, ptr %dm, i64 %tmp19, i32 1 + store i32 0, ptr %tmp1, align 4 br label %bb3 bb2: - %tmp2 = getelementptr inbounds %struct3, %struct3* %tmp, i64 %tmp20, i32 1 - %tmp12 = getelementptr inbounds %struct4, %struct4* %tmp2, i64 0, i32 0, i32 1 - store i32 0, i32* %tmp12, align 4 + %tmp2 = getelementptr inbounds %struct3, ptr %dm, i64 %tmp20, i32 1 + %tmp12 = getelementptr inbounds %struct4, ptr %tmp2, i64 0, i32 0, i32 1 + store i32 0, ptr %tmp12, align 4 br label %bb3 bb3: - %phi = phi %struct4* [ %tmp1, %bb1 ], [ %tmp2, %bb2 ] + %phi = phi ptr [ %tmp1, %bb1 ], [ %tmp2, %bb2 ] %tmp22 = invoke i32 @foo1(i32 11) to label %bb4 unwind label %bb5 bb4: ret i32 0 bb5: - %tmp27 = landingpad { i8*, i32 } catch i8* bitcast (i8** @_ZTIi to i8*) - %tmp34 = getelementptr inbounds %struct4, %struct4* %phi, i64 %tmp21, i32 1 - %tmp35 = getelementptr inbounds %struct2, %struct2* %tmp34, i64 0, i32 1 - %tmp25 = load i32, i32* %tmp35, align 4 + %tmp27 = landingpad { ptr, i32 } catch ptr @_ZTIi + %tmp34 = getelementptr inbounds %struct4, ptr %phi, i64 %tmp21, i32 1 + %tmp35 = getelementptr inbounds %struct2, ptr %tmp34, i64 0, i32 1 + %tmp25 = load i32, ptr %tmp35, align 4 ret i32 %tmp25 } -@_ZTIi = external constant i8* +@_ZTIi = external constant ptr declare i32 @__gxx_personality_v0(...) declare i32 @foo1(i32) @@ -144,7 +134,7 @@ declare i32 @foo1(i32) ; Check that instcombine doesn't fold GEPs into themselves through a loop ; back-edge. -define i8* @test4(i32 %value, i8* %buffer) { +define ptr @test4(i32 %value, ptr %buffer) { ; CHECK-LABEL: @test4( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[CMP:%.*]] = icmp ugt i32 [[VALUE:%.*]], 127 @@ -152,21 +142,21 @@ define i8* @test4(i32 %value, i8* %buffer) { ; CHECK: loop.header: ; CHECK-NEXT: br label [[LOOP_BODY:%.*]] ; CHECK: loop.body: -; CHECK-NEXT: [[BUFFER_PN:%.*]] = phi i8* [ [[BUFFER:%.*]], [[LOOP_HEADER]] ], [ [[LOOPPTR:%.*]], [[LOOP_BODY]] ] +; CHECK-NEXT: [[BUFFER_PN:%.*]] = phi ptr [ [[BUFFER:%.*]], [[LOOP_HEADER]] ], [ [[LOOPPTR:%.*]], [[LOOP_BODY]] ] ; CHECK-NEXT: [[NEWVAL:%.*]] = phi i32 [ [[VALUE]], [[LOOP_HEADER]] ], [ [[SHR:%.*]], [[LOOP_BODY]] ] -; CHECK-NEXT: [[LOOPPTR]] = getelementptr inbounds i8, i8* [[BUFFER_PN]], i64 1 +; CHECK-NEXT: [[LOOPPTR]] = getelementptr inbounds i8, ptr [[BUFFER_PN]], i64 1 ; CHECK-NEXT: [[SHR]] = lshr i32 [[NEWVAL]], 7 ; CHECK-NEXT: [[CMP2:%.*]] = icmp ugt i32 [[NEWVAL]], 16383 ; CHECK-NEXT: br i1 [[CMP2]], label [[LOOP_BODY]], label [[LOOP_EXIT:%.*]] ; CHECK: loop.exit: ; CHECK-NEXT: br label [[EXIT]] ; CHECK: exit: -; CHECK-NEXT: [[TMP0:%.*]] = phi i8* [ [[LOOPPTR]], [[LOOP_EXIT]] ], [ [[BUFFER]], [[ENTRY:%.*]] ] -; CHECK-NEXT: [[INCPTR3:%.*]] = getelementptr inbounds i8, i8* [[TMP0]], i64 2 -; CHECK-NEXT: ret i8* [[INCPTR3]] +; CHECK-NEXT: [[TMP0:%.*]] = phi ptr [ [[LOOPPTR]], [[LOOP_EXIT]] ], [ [[BUFFER]], [[ENTRY:%.*]] ] +; CHECK-NEXT: [[INCPTR3:%.*]] = getelementptr inbounds i8, ptr [[TMP0]], i64 2 +; CHECK-NEXT: ret ptr [[INCPTR3]] ; entry: - %incptr = getelementptr inbounds i8, i8* %buffer, i64 1 + %incptr = getelementptr inbounds i8, ptr %buffer, i64 1 %cmp = icmp ugt i32 %value, 127 br i1 %cmp, label %loop.header, label %exit @@ -174,55 +164,55 @@ loop.header: br label %loop.body loop.body: - %loopptr = phi i8* [ %incptr, %loop.header ], [ %incptr2, %loop.body ] + %loopptr = phi ptr [ %incptr, %loop.header ], [ %incptr2, %loop.body ] %newval = phi i32 [ %value, %loop.header ], [ %shr, %loop.body ] %shr = lshr i32 %newval, 7 - %incptr2 = getelementptr inbounds i8, i8* %loopptr, i64 1 + %incptr2 = getelementptr inbounds i8, ptr %loopptr, i64 1 %cmp2 = icmp ugt i32 %shr, 127 br i1 %cmp2, label %loop.body, label %loop.exit loop.exit: - %exitptr = phi i8* [ %incptr2, %loop.body ] + %exitptr = phi ptr [ %incptr2, %loop.body ] br label %exit exit: - %ptr2 = phi i8* [ %exitptr, %loop.exit ], [ %incptr, %entry ] - %incptr3 = getelementptr inbounds i8, i8* %ptr2, i64 1 - ret i8* %incptr3 + %ptr2 = phi ptr [ %exitptr, %loop.exit ], [ %incptr, %entry ] + %incptr3 = getelementptr inbounds i8, ptr %ptr2, i64 1 + ret ptr %incptr3 } @.str.4 = external unnamed_addr constant [100 x i8], align 1 ; Instcombine shouldn't add new PHI nodes while folding GEPs if that will leave ; old PHI nodes behind as this is not clearly beneficial. -define void @test5(i16 *%idx, i8 **%in) #0 { +define void @test5(ptr %idx, ptr %in) #0 { ; CHECK-LABEL: @test5( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[TMP0:%.*]] = load i8*, i8** [[IN:%.*]], align 8 -; CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds i8, i8* [[TMP0]], i64 1 -; CHECK-NEXT: [[TMP1:%.*]] = load i8, i8* [[INCDEC_PTR]], align 1 +; CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[IN:%.*]], align 8 +; CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds i8, ptr [[TMP0]], i64 1 +; CHECK-NEXT: [[TMP1:%.*]] = load i8, ptr [[INCDEC_PTR]], align 1 ; CHECK-NEXT: [[CMP23:%.*]] = icmp eq i8 [[TMP1]], 54 ; CHECK-NEXT: br i1 [[CMP23]], label [[WHILE_COND:%.*]], label [[IF_THEN_25:%.*]] ; CHECK: if.then.25: -; CHECK-NEXT: call void @g(i8* nonnull getelementptr inbounds ([100 x i8], [100 x i8]* @.str.4, i64 0, i64 0)) +; CHECK-NEXT: call void @g(ptr nonnull @.str.4) ; CHECK-NEXT: br label [[WHILE_COND]] ; CHECK: while.cond: -; CHECK-NEXT: [[PTR:%.*]] = phi i8* [ [[INCDEC_PTR]], [[ENTRY:%.*]] ], [ [[INCDEC_PTR32:%.*]], [[WHILE_BODY:%.*]] ], [ [[INCDEC_PTR]], [[IF_THEN_25]] ] -; CHECK-NEXT: [[TMP2:%.*]] = load i8, i8* [[PTR]], align 1 +; CHECK-NEXT: [[PTR:%.*]] = phi ptr [ [[INCDEC_PTR]], [[ENTRY:%.*]] ], [ [[INCDEC_PTR32:%.*]], [[WHILE_BODY:%.*]] ], [ [[INCDEC_PTR]], [[IF_THEN_25]] ] +; CHECK-NEXT: [[TMP2:%.*]] = load i8, ptr [[PTR]], align 1 ; CHECK-NEXT: [[AND:%.*]] = and i8 [[TMP2]], 64 ; CHECK-NEXT: [[LNOT:%.*]] = icmp eq i8 [[AND]], 0 ; CHECK-NEXT: br i1 [[LNOT]], label [[WHILE_BODY]], label [[WHILE_COND_33:%.*]] ; CHECK: while.body: -; CHECK-NEXT: [[INCDEC_PTR32]] = getelementptr inbounds i8, i8* [[PTR]], i64 1 +; CHECK-NEXT: [[INCDEC_PTR32]] = getelementptr inbounds i8, ptr [[PTR]], i64 1 ; CHECK-NEXT: br label [[WHILE_COND]] ; CHECK: while.cond.33: -; CHECK-NEXT: [[INCDEC_PTR34:%.*]] = getelementptr inbounds i8, i8* [[PTR]], i64 1 +; CHECK-NEXT: [[INCDEC_PTR34:%.*]] = getelementptr inbounds i8, ptr [[PTR]], i64 1 ; CHECK-NEXT: br label [[WHILE_COND_57:%.*]] ; CHECK: while.cond.57: -; CHECK-NEXT: [[TMP3:%.*]] = load i8, i8* [[INCDEC_PTR34]], align 1 +; CHECK-NEXT: [[TMP3:%.*]] = load i8, ptr [[INCDEC_PTR34]], align 1 ; CHECK-NEXT: [[TMP4:%.*]] = zext i8 [[TMP3]] to i64 -; CHECK-NEXT: [[ARRAYIDX61:%.*]] = getelementptr inbounds i16, i16* [[IDX:%.*]], i64 [[TMP4]] -; CHECK-NEXT: [[TMP5:%.*]] = load i16, i16* [[ARRAYIDX61]], align 2 +; CHECK-NEXT: [[ARRAYIDX61:%.*]] = getelementptr inbounds i16, ptr [[IDX:%.*]], i64 [[TMP4]] +; CHECK-NEXT: [[TMP5:%.*]] = load i16, ptr [[ARRAYIDX61]], align 2 ; CHECK-NEXT: [[AND63:%.*]] = and i16 [[TMP5]], 2048 ; CHECK-NEXT: [[TOBOOL64:%.*]] = icmp eq i16 [[AND63]], 0 ; CHECK-NEXT: br i1 [[TOBOOL64]], label [[WHILE_COND_73:%.*]], label [[WHILE_COND_57]] @@ -230,36 +220,36 @@ define void @test5(i16 *%idx, i8 **%in) #0 { ; CHECK-NEXT: br label [[WHILE_COND_73]] ; entry: - %0 = load i8*, i8** %in - %incdec.ptr = getelementptr inbounds i8, i8* %0, i32 1 - %1 = load i8, i8* %incdec.ptr, align 1 + %0 = load ptr, ptr %in + %incdec.ptr = getelementptr inbounds i8, ptr %0, i32 1 + %1 = load i8, ptr %incdec.ptr, align 1 %cmp23 = icmp eq i8 %1, 54 br i1 %cmp23, label %while.cond, label %if.then.25 if.then.25: - call void @g(i8* getelementptr inbounds ([100 x i8], [100 x i8]* @.str.4, i32 0, i32 0)) + call void @g(ptr @.str.4) br label %while.cond while.cond: - %Ptr = phi i8* [ %incdec.ptr, %entry ], [ %incdec.ptr32, %while.body], [%incdec.ptr, %if.then.25 ] - %2 = load i8, i8* %Ptr + %Ptr = phi ptr [ %incdec.ptr, %entry ], [ %incdec.ptr32, %while.body], [%incdec.ptr, %if.then.25 ] + %2 = load i8, ptr %Ptr %and = and i8 %2, 64 %lnot = icmp eq i8 %and, 0 br i1 %lnot, label %while.body, label %while.cond.33 while.body: - %incdec.ptr32 = getelementptr inbounds i8, i8* %Ptr, i32 1 + %incdec.ptr32 = getelementptr inbounds i8, ptr %Ptr, i32 1 br label %while.cond while.cond.33: - %incdec.ptr34 = getelementptr inbounds i8, i8* %Ptr, i32 1 + %incdec.ptr34 = getelementptr inbounds i8, ptr %Ptr, i32 1 br label %while.cond.57 while.cond.57: - %3 = load i8, i8* %incdec.ptr34, align 1 + %3 = load i8, ptr %incdec.ptr34, align 1 %conv59 = zext i8 %3 to i32 - %arrayidx61 = getelementptr inbounds i16, i16* %idx, i32 %conv59 - %4 = load i16, i16* %arrayidx61, align 2 + %arrayidx61 = getelementptr inbounds i16, ptr %idx, i32 %conv59 + %4 = load i16, ptr %arrayidx61, align 2 %and63 = and i16 %4, 2048 %tobool64 = icmp eq i16 %and63, 0 br i1 %tobool64, label %while.cond.73, label %while.cond.57 @@ -268,4 +258,4 @@ while.cond.73: br label %while.cond.73 } -declare void @g(i8*) +declare void @g(ptr) diff --git a/llvm/test/Transforms/InstCombine/memchr-8.ll b/llvm/test/Transforms/InstCombine/memchr-8.ll index 7f0228d..aefd518 100644 --- a/llvm/test/Transforms/InstCombine/memchr-8.ll +++ b/llvm/test/Transforms/InstCombine/memchr-8.ll @@ -6,58 +6,58 @@ ; at 64K for the largest supported zeroinitiazer. If the limit changes ; the test might need to be adjusted. -declare i8* @memrchr(i8*, i32, i64) +declare ptr @memrchr(ptr, i32, i64) @a = constant <{ i8, [4294967295 x i8] }> <{ i8 1, [4294967295 x i8] zeroinitializer }> ; Verify reading an initializer INT32_MAX + 1 bytes large (starting at ; offset 2147483647 into a which is UINT32_MAX bytes in size). -define i8* @call_a_pi32max_p1() { +define ptr @call_a_pi32max_p1() { ; CHECK-LABEL: @call_a_pi32max_p1( -; CHECK-NEXT: [[CHR:%.*]] = tail call i8* @memrchr(i8* noundef nonnull dereferenceable(2147483647) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i64 0, i32 1, i64 2147483647), i32 0, i64 2147483647) -; CHECK-NEXT: ret i8* [[CHR]] +; CHECK-NEXT: [[CHR:%.*]] = tail call ptr @memrchr(ptr noundef nonnull dereferenceable(2147483647) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, ptr @a, i64 0, i32 1, i64 2147483647), i32 0, i64 2147483647) +; CHECK-NEXT: ret ptr [[CHR]] ; %ptr = getelementptr <{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i32 0, i32 1, i32 2147483647 - %chr = tail call i8* @memrchr(i8* %ptr, i32 0, i64 2147483647) - ret i8* %chr + %chr = tail call ptr @memrchr(ptr %ptr, i32 0, i64 2147483647) + ret ptr %chr } ; Verify reading an initializer INT32_MAX bytes large (starting at offset ; 2147483648 into a which is UINT32_MAX bytes in size). -define i8* @call_a_pi32max() { +define ptr @call_a_pi32max() { ; CHECK-LABEL: @call_a_pi32max( -; CHECK-NEXT: [[CHR:%.*]] = tail call i8* @memrchr(i8* noundef nonnull dereferenceable(2147483647) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i64 0, i32 1, i64 2147483648), i32 0, i64 2147483647) -; CHECK-NEXT: ret i8* [[CHR]] +; CHECK-NEXT: [[CHR:%.*]] = tail call ptr @memrchr(ptr noundef nonnull dereferenceable(2147483647) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, ptr @a, i64 0, i32 1, i64 2147483648), i32 0, i64 2147483647) +; CHECK-NEXT: ret ptr [[CHR]] ; %ptr = getelementptr <{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i32 0, i32 1, i64 2147483648 - %chr = tail call i8* @memrchr(i8* %ptr, i32 0, i64 2147483647) - ret i8* %chr + %chr = tail call ptr @memrchr(ptr %ptr, i32 0, i64 2147483647) + ret ptr %chr } ; Verify reading an initializer UINT32_MAX bytes large (starting at offset ; 1 into a). -define i8* @call_a_pui32max() { +define ptr @call_a_pui32max() { ; CHECK-LABEL: @call_a_pui32max( -; CHECK-NEXT: [[CHR:%.*]] = tail call i8* @memrchr(i8* noundef nonnull dereferenceable(4294967295) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i64 0, i32 1, i64 0), i32 0, i64 4294967295) -; CHECK-NEXT: ret i8* [[CHR]] +; CHECK-NEXT: [[CHR:%.*]] = tail call ptr @memrchr(ptr noundef nonnull dereferenceable(4294967295) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, ptr @a, i64 0, i32 1, i64 0), i32 0, i64 4294967295) +; CHECK-NEXT: ret ptr [[CHR]] ; %ptr = getelementptr <{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i32 0, i32 1, i32 0 - %chr = tail call i8* @memrchr(i8* %ptr, i32 0, i64 4294967295) - ret i8* %chr + %chr = tail call ptr @memrchr(ptr %ptr, i32 0, i64 4294967295) + ret ptr %chr } ; Verify reading an initializer UINT32_MAX + 1 bytes large (all of a). -define i8* @call_a_puimax_p1() { +define ptr @call_a_puimax_p1() { ; CHECK-LABEL: @call_a_puimax_p1( -; CHECK-NEXT: [[CHR:%.*]] = tail call i8* @memrchr(i8* noundef nonnull dereferenceable(4294967296) getelementptr inbounds (<{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i64 0, i32 0), i32 0, i64 4294967296) -; CHECK-NEXT: ret i8* [[CHR]] +; CHECK-NEXT: [[CHR:%.*]] = tail call ptr @memrchr(ptr noundef nonnull dereferenceable(4294967296) @a, i32 0, i64 4294967296) +; CHECK-NEXT: ret ptr [[CHR]] ; %ptr = getelementptr <{ i8, [4294967295 x i8] }>, <{ i8, [4294967295 x i8] }>* @a, i32 0, i32 0 - %chr = tail call i8* @memrchr(i8* %ptr, i32 0, i64 4294967296) - ret i8* %chr + %chr = tail call ptr @memrchr(ptr %ptr, i32 0, i64 4294967296) + ret ptr %chr } diff --git a/llvm/test/Transforms/InstCombine/memcpy-addrspace.ll b/llvm/test/Transforms/InstCombine/memcpy-addrspace.ll index c0f5e6f..d154369 100644 --- a/llvm/test/Transforms/InstCombine/memcpy-addrspace.ll +++ b/llvm/test/Transforms/InstCombine/memcpy-addrspace.ll @@ -3,152 +3,141 @@ @test.data = private unnamed_addr addrspace(2) constant [8 x i32] [i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7], align 4 -define void @test_load(i32 addrspace(1)* %out, i64 %x) { +define void @test_load(ptr addrspace(1) %out, i64 %x) { ; CHECK-LABEL: @test_load( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr [8 x i32], [8 x i32] addrspace(2)* @test.data, i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32 addrspace(2)* [[ARRAYIDX]], align 4 -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP0]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr [8 x i32], ptr addrspace(2) @test.data, i64 0, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(2) [[ARRAYIDX]], align 4 +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* %data, i64 0, i64 %x - %1 = load i32, i32* %arrayidx, align 4 - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %1, i32 addrspace(1)* %arrayidx1, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds [8 x i32], ptr %data, i64 0, i64 %x + %0 = load i32, ptr %arrayidx, align 4 + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 ret void } -define void @test_load_bitcast_chain(i32 addrspace(1)* %out, i64 %x) { +define void @test_load_bitcast_chain(ptr addrspace(1) %out, i64 %x) { ; CHECK-LABEL: @test_load_bitcast_chain( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr [8 x i32], [8 x i32] addrspace(2)* @test.data, i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32 addrspace(2)* [[ARRAYIDX2]], align 4 -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP0]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr i32, ptr addrspace(2) @test.data, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(2) [[ARRAYIDX]], align 4 +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %1 = bitcast i8* %0 to i32* - %arrayidx = getelementptr inbounds i32, i32* %1, i64 %x - %2 = load i32, i32* %arrayidx, align 4 - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %2, i32 addrspace(1)* %arrayidx1, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds i32, ptr %data, i64 %x + %0 = load i32, ptr %arrayidx, align 4 + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 ret void } -define void @test_call(i32 addrspace(1)* %out, i64 %x) { +define void @test_call(ptr addrspace(1) %out, i64 %x) { ; CHECK-LABEL: @test_call( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[DATA:%.*]] = alloca [8 x i32], align 4 -; CHECK-NEXT: [[TMP0:%.*]] = bitcast [8 x i32]* [[DATA]] to i8* -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* noundef nonnull align 4 dereferenceable(32) [[TMP0]], i8 addrspace(2)* noundef align 4 dereferenceable(32) bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) -; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[DATA]], i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP1:%.*]] = call i32 @foo(i32* nonnull [[ARRAYIDX]]) -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP1]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 +; CHECK-NEXT: call void @llvm.memcpy.p0.p2.i64(ptr noundef nonnull align 4 dereferenceable(32) [[DATA]], ptr addrspace(2) noundef align 4 dereferenceable(32) @test.data, i64 32, i1 false) +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], ptr [[DATA]], i64 0, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @foo(ptr nonnull [[ARRAYIDX]]) +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* %data, i64 0, i64 %x - %1 = call i32 @foo(i32* %arrayidx) - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %1, i32 addrspace(1)* %arrayidx1, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds [8 x i32], ptr %data, i64 0, i64 %x + %0 = call i32 @foo(ptr %arrayidx) + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 ret void } -define void @test_call_no_null_opt(i32 addrspace(1)* %out, i64 %x) #0 { +define void @test_call_no_null_opt(ptr addrspace(1) %out, i64 %x) #0 { ; CHECK-LABEL: @test_call_no_null_opt( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[DATA:%.*]] = alloca [8 x i32], align 4 -; CHECK-NEXT: [[TMP0:%.*]] = bitcast [8 x i32]* [[DATA]] to i8* -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* noundef nonnull align 4 dereferenceable(32) [[TMP0]], i8 addrspace(2)* noundef align 4 dereferenceable(32) bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) -; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[DATA]], i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP1:%.*]] = call i32 @foo(i32* [[ARRAYIDX]]) -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP1]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 +; CHECK-NEXT: call void @llvm.memcpy.p0.p2.i64(ptr noundef nonnull align 4 dereferenceable(32) [[DATA]], ptr addrspace(2) noundef align 4 dereferenceable(32) @test.data, i64 32, i1 false) +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], ptr [[DATA]], i64 0, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @foo(ptr [[ARRAYIDX]]) +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* %data, i64 0, i64 %x - %1 = call i32 @foo(i32* %arrayidx) - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %1, i32 addrspace(1)* %arrayidx1, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds [8 x i32], ptr %data, i64 0, i64 %x + %0 = call i32 @foo(ptr %arrayidx) + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 ret void } -define void @test_load_and_call(i32 addrspace(1)* %out, i64 %x, i64 %y) { +define void @test_load_and_call(ptr addrspace(1) %out, i64 %x, i64 %y) { ; CHECK-LABEL: @test_load_and_call( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[DATA:%.*]] = alloca [8 x i32], align 4 -; CHECK-NEXT: [[TMP0:%.*]] = bitcast [8 x i32]* [[DATA]] to i8* -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* noundef nonnull align 4 dereferenceable(32) [[TMP0]], i8 addrspace(2)* noundef align 4 dereferenceable(32) bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) -; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[DATA]], i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[ARRAYIDX]], align 4 -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP1]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 -; CHECK-NEXT: [[TMP2:%.*]] = call i32 @foo(i32* nonnull [[ARRAYIDX]]) -; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT]], i64 [[Y:%.*]] -; CHECK-NEXT: store i32 [[TMP2]], i32 addrspace(1)* [[ARRAYIDX2]], align 4 +; CHECK-NEXT: call void @llvm.memcpy.p0.p2.i64(ptr noundef nonnull align 4 dereferenceable(32) [[DATA]], ptr addrspace(2) noundef align 4 dereferenceable(32) @test.data, i64 32, i1 false) +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], ptr [[DATA]], i64 0, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[ARRAYIDX]], align 4 +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 +; CHECK-NEXT: [[TMP1:%.*]] = call i32 @foo(ptr nonnull [[ARRAYIDX]]) +; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT]], i64 [[Y:%.*]] +; CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(1) [[ARRAYIDX2]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* %data, i64 0, i64 %x - %1 = load i32, i32* %arrayidx, align 4 - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %1, i32 addrspace(1)* %arrayidx1, align 4 - %2 = call i32 @foo(i32* %arrayidx) - %arrayidx2 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %y - store i32 %2, i32 addrspace(1)* %arrayidx2, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds [8 x i32], ptr %data, i64 0, i64 %x + %0 = load i32, ptr %arrayidx, align 4 + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 + %1 = call i32 @foo(ptr %arrayidx) + %arrayidx2 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %y + store i32 %1, ptr addrspace(1) %arrayidx2, align 4 ret void } -define void @test_load_and_call_no_null_opt(i32 addrspace(1)* %out, i64 %x, i64 %y) #0 { +define void @test_load_and_call_no_null_opt(ptr addrspace(1) %out, i64 %x, i64 %y) #0 { ; CHECK-LABEL: @test_load_and_call_no_null_opt( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[DATA:%.*]] = alloca [8 x i32], align 4 -; CHECK-NEXT: [[TMP0:%.*]] = bitcast [8 x i32]* [[DATA]] to i8* -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* noundef nonnull align 4 dereferenceable(32) [[TMP0]], i8 addrspace(2)* noundef align 4 dereferenceable(32) bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) -; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[DATA]], i64 0, i64 [[X:%.*]] -; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[ARRAYIDX]], align 4 -; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT:%.*]], i64 [[X]] -; CHECK-NEXT: store i32 [[TMP1]], i32 addrspace(1)* [[ARRAYIDX1]], align 4 -; CHECK-NEXT: [[TMP2:%.*]] = call i32 @foo(i32* [[ARRAYIDX]]) -; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[OUT]], i64 [[Y:%.*]] -; CHECK-NEXT: store i32 [[TMP2]], i32 addrspace(1)* [[ARRAYIDX2]], align 4 +; CHECK-NEXT: call void @llvm.memcpy.p0.p2.i64(ptr noundef nonnull align 4 dereferenceable(32) [[DATA]], ptr addrspace(2) noundef align 4 dereferenceable(32) @test.data, i64 32, i1 false) +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], ptr [[DATA]], i64 0, i64 [[X:%.*]] +; CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[ARRAYIDX]], align 4 +; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT:%.*]], i64 [[X]] +; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[ARRAYIDX1]], align 4 +; CHECK-NEXT: [[TMP1:%.*]] = call i32 @foo(ptr [[ARRAYIDX]]) +; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT]], i64 [[Y:%.*]] +; CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(1) [[ARRAYIDX2]], align 4 ; CHECK-NEXT: ret void ; entry: %data = alloca [8 x i32], align 4 - %0 = bitcast [8 x i32]* %data to i8* - call void @llvm.memcpy.p0i8.p2i8.i64(i8* align 4 %0, i8 addrspace(2)* align 4 bitcast ([8 x i32] addrspace(2)* @test.data to i8 addrspace(2)*), i64 32, i1 false) - %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* %data, i64 0, i64 %x - %1 = load i32, i32* %arrayidx, align 4 - %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %x - store i32 %1, i32 addrspace(1)* %arrayidx1, align 4 - %2 = call i32 @foo(i32* %arrayidx) - %arrayidx2 = getelementptr inbounds i32, i32 addrspace(1)* %out, i64 %y - store i32 %2, i32 addrspace(1)* %arrayidx2, align 4 + call void @llvm.memcpy.p0.p2.i64(ptr align 4 %data, ptr addrspace(2) align 4 @test.data, i64 32, i1 false) + %arrayidx = getelementptr inbounds [8 x i32], ptr %data, i64 0, i64 %x + %0 = load i32, ptr %arrayidx, align 4 + %arrayidx1 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %x + store i32 %0, ptr addrspace(1) %arrayidx1, align 4 + %1 = call i32 @foo(ptr %arrayidx) + %arrayidx2 = getelementptr inbounds i32, ptr addrspace(1) %out, i64 %y + store i32 %1, ptr addrspace(1) %arrayidx2, align 4 ret void } -declare void @llvm.memcpy.p0i8.p2i8.i64(i8* nocapture writeonly, i8 addrspace(2)* nocapture readonly, i64, i1) -declare i32 @foo(i32* %x) +declare void @llvm.memcpy.p0.p2.i64(ptr nocapture writeonly, ptr addrspace(2) nocapture readonly, i64, i1) +declare i32 @foo(ptr %x) attributes #0 = { null_pointer_is_valid } diff --git a/llvm/test/Transforms/InstCombine/memrchr-4.ll b/llvm/test/Transforms/InstCombine/memrchr-4.ll index c383d43..54dde6b 100644 --- a/llvm/test/Transforms/InstCombine/memrchr-4.ll +++ b/llvm/test/Transforms/InstCombine/memrchr-4.ll @@ -4,7 +4,7 @@ ; Verify that memrchr calls with a string consisting of all the same ; characters are folded and those with mixed strings are not. -declare i8* @memrchr(i8*, i32, i64) +declare ptr @memrchr(ptr, i32, i64) @a11111 = constant [5 x i8] c"\01\01\01\01\01" @a1110111 = constant [7 x i8] c"\01\01\01\00\01\01\01" @@ -12,94 +12,88 @@ declare i8* @memrchr(i8*, i32, i64) ; Fold memrchr(a11111, C, 5) to *a11111 == C ? a11111 + 5 - 1 : null. -define i8* @fold_memrchr_a11111_c_5(i32 %C) { +define ptr @fold_memrchr_a11111_c_5(i32 %C) { ; CHECK-LABEL: @fold_memrchr_a11111_c_5( ; CHECK-NEXT: [[TMP1:%.*]] = trunc i32 [[C:%.*]] to i8 ; CHECK-NEXT: [[TMP2:%.*]] = icmp eq i8 [[TMP1]], 1 -; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP2]], i8* getelementptr inbounds ([5 x i8], [5 x i8]* @a11111, i64 0, i64 4), i8* null -; CHECK-NEXT: ret i8* [[MEMRCHR_SEL]] +; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP2]], ptr getelementptr inbounds ([5 x i8], ptr @a11111, i64 0, i64 4), ptr null +; CHECK-NEXT: ret ptr [[MEMRCHR_SEL]] ; - %ptr = getelementptr [5 x i8], [5 x i8]* @a11111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 5) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a11111, i32 %C, i64 5) + ret ptr %ret } ; Fold memrchr(a11111, C, N) to N && *a11111 == C ? a11111 + N - 1 : null, ; on the assumption that N is in bounds. -define i8* @fold_memrchr_a11111_c_n(i32 %C, i64 %N) { +define ptr @fold_memrchr_a11111_c_n(i32 %C, i64 %N) { ; CHECK-LABEL: @fold_memrchr_a11111_c_n( ; CHECK-NEXT: [[TMP1:%.*]] = icmp ne i64 [[N:%.*]], 0 ; CHECK-NEXT: [[TMP2:%.*]] = trunc i32 [[C:%.*]] to i8 ; CHECK-NEXT: [[TMP3:%.*]] = icmp eq i8 [[TMP2]], 1 ; CHECK-NEXT: [[TMP4:%.*]] = select i1 [[TMP1]], i1 [[TMP3]], i1 false ; CHECK-NEXT: [[TMP5:%.*]] = add i64 [[N]], -1 -; CHECK-NEXT: [[MEMRCHR_PTR_PLUS:%.*]] = getelementptr inbounds [5 x i8], [5 x i8]* @a11111, i64 0, i64 [[TMP5]] -; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP4]], i8* [[MEMRCHR_PTR_PLUS]], i8* null -; CHECK-NEXT: ret i8* [[MEMRCHR_SEL]] +; CHECK-NEXT: [[MEMRCHR_PTR_PLUS:%.*]] = getelementptr inbounds i8, ptr @a11111, i64 [[TMP5]] +; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP4]], ptr [[MEMRCHR_PTR_PLUS]], ptr null +; CHECK-NEXT: ret ptr [[MEMRCHR_SEL]] ; - %ptr = getelementptr [5 x i8], [5 x i8]* @a11111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 %N) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a11111, i32 %C, i64 %N) + ret ptr %ret } ; Fold memrchr(a1110111, C, 3) to a1110111[2] == C ? a1110111 + 2 : null. -define i8* @fold_memrchr_a1110111_c_3(i32 %C) { +define ptr @fold_memrchr_a1110111_c_3(i32 %C) { ; CHECK-LABEL: @fold_memrchr_a1110111_c_3( ; CHECK-NEXT: [[TMP1:%.*]] = trunc i32 [[C:%.*]] to i8 ; CHECK-NEXT: [[TMP2:%.*]] = icmp eq i8 [[TMP1]], 1 -; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP2]], i8* getelementptr inbounds ([7 x i8], [7 x i8]* @a1110111, i64 0, i64 2), i8* null -; CHECK-NEXT: ret i8* [[MEMRCHR_SEL]] +; CHECK-NEXT: [[MEMRCHR_SEL:%.*]] = select i1 [[TMP2]], ptr getelementptr inbounds ([7 x i8], ptr @a1110111, i64 0, i64 2), ptr null +; CHECK-NEXT: ret ptr [[MEMRCHR_SEL]] ; - %ptr = getelementptr [7 x i8], [7 x i8]* @a1110111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 3) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a1110111, i32 %C, i64 3) + ret ptr %ret } ; Don't fold memrchr(a1110111, C, 4). -define i8* @call_memrchr_a1110111_c_4(i32 %C) { +define ptr @call_memrchr_a1110111_c_4(i32 %C) { ; CHECK-LABEL: @call_memrchr_a1110111_c_4( -; CHECK-NEXT: [[RET:%.*]] = call i8* @memrchr(i8* noundef nonnull dereferenceable(4) getelementptr inbounds ([7 x i8], [7 x i8]* @a1110111, i64 0, i64 0), i32 [[C:%.*]], i64 4) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = call ptr @memrchr(ptr noundef nonnull dereferenceable(4) @a1110111, i32 [[C:%.*]], i64 4) +; CHECK-NEXT: ret ptr [[RET]] ; - %ptr = getelementptr [7 x i8], [7 x i8]* @a1110111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 4) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a1110111, i32 %C, i64 4) + ret ptr %ret } ; Don't fold memrchr(a1110111, C, 7). -define i8* @call_memrchr_a1110111_c_7(i32 %C) { +define ptr @call_memrchr_a1110111_c_7(i32 %C) { ; CHECK-LABEL: @call_memrchr_a1110111_c_7( -; CHECK-NEXT: [[RET:%.*]] = call i8* @memrchr(i8* noundef nonnull dereferenceable(7) getelementptr inbounds ([7 x i8], [7 x i8]* @a1110111, i64 0, i64 0), i32 [[C:%.*]], i64 7) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = call ptr @memrchr(ptr noundef nonnull dereferenceable(7) @a1110111, i32 [[C:%.*]], i64 7) +; CHECK-NEXT: ret ptr [[RET]] ; - %ptr = getelementptr [7 x i8], [7 x i8]* @a1110111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 7) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a1110111, i32 %C, i64 7) + ret ptr %ret } ; Don't fold memrchr(a1110111, C, N). -define i8* @call_memrchr_a1110111_c_n(i32 %C, i64 %N) { +define ptr @call_memrchr_a1110111_c_n(i32 %C, i64 %N) { ; CHECK-LABEL: @call_memrchr_a1110111_c_n( -; CHECK-NEXT: [[RET:%.*]] = call i8* @memrchr(i8* nonnull getelementptr inbounds ([7 x i8], [7 x i8]* @a1110111, i64 0, i64 0), i32 [[C:%.*]], i64 [[N:%.*]]) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = call ptr @memrchr(ptr nonnull @a1110111, i32 [[C:%.*]], i64 [[N:%.*]]) +; CHECK-NEXT: ret ptr [[RET]] ; - %ptr = getelementptr [7 x i8], [7 x i8]* @a1110111, i64 0, i64 0 - %ret = call i8* @memrchr(i8* %ptr, i32 %C, i64 %N) - ret i8* %ret + %ret = call ptr @memrchr(ptr @a1110111, i32 %C, i64 %N) + ret ptr %ret } diff --git a/llvm/test/Transforms/InstCombine/select.ll b/llvm/test/Transforms/InstCombine/select.ll index 7388f64b..2aca9fc 100644 --- a/llvm/test/Transforms/InstCombine/select.ll +++ b/llvm/test/Transforms/InstCombine/select.ll @@ -338,73 +338,73 @@ define i1 @test14b(i1 %C, i32 %X) { ret i1 %R } -define i32 @test16(i1 %C, i32* %P) { +define i32 @test16(i1 %C, ptr %P) { ; CHECK-LABEL: @test16( -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P:%.*]], align 4 +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P:%.*]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %P2 = select i1 %C, i32* %P, i32* null - %V = load i32, i32* %P2 + %P2 = select i1 %C, ptr %P, ptr null + %V = load i32, ptr %P2 ret i32 %V } ;; It may be legal to load from a null address in a non-zero address space -define i32 @test16_neg(i1 %C, i32 addrspace(1)* %P) { +define i32 @test16_neg(i1 %C, ptr addrspace(1) %P) { ; CHECK-LABEL: @test16_neg( -; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], i32 addrspace(1)* [[P:%.*]], i32 addrspace(1)* null -; CHECK-NEXT: [[V:%.*]] = load i32, i32 addrspace(1)* [[P2]], align 4 +; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], ptr addrspace(1) [[P:%.*]], ptr addrspace(1) null +; CHECK-NEXT: [[V:%.*]] = load i32, ptr addrspace(1) [[P2]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %P2 = select i1 %C, i32 addrspace(1)* %P, i32 addrspace(1)* null - %V = load i32, i32 addrspace(1)* %P2 + %P2 = select i1 %C, ptr addrspace(1) %P, ptr addrspace(1) null + %V = load i32, ptr addrspace(1) %P2 ret i32 %V } -define i32 @test16_neg2(i1 %C, i32 addrspace(1)* %P) { +define i32 @test16_neg2(i1 %C, ptr addrspace(1) %P) { ; CHECK-LABEL: @test16_neg2( -; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], i32 addrspace(1)* null, i32 addrspace(1)* [[P:%.*]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32 addrspace(1)* [[P2]], align 4 +; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], ptr addrspace(1) null, ptr addrspace(1) [[P:%.*]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr addrspace(1) [[P2]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %P2 = select i1 %C, i32 addrspace(1)* null, i32 addrspace(1)* %P - %V = load i32, i32 addrspace(1)* %P2 + %P2 = select i1 %C, ptr addrspace(1) null, ptr addrspace(1) %P + %V = load i32, ptr addrspace(1) %P2 ret i32 %V } ;; It may be legal to load from a null address with null pointer valid attribute. -define i32 @test16_no_null_opt(i1 %C, i32* %P) #0 { +define i32 @test16_no_null_opt(i1 %C, ptr %P) #0 { ; CHECK-LABEL: @test16_no_null_opt( -; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], i32* [[P:%.*]], i32* null -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P2]], align 4 +; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], ptr [[P:%.*]], ptr null +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P2]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %P2 = select i1 %C, i32* %P, i32* null - %V = load i32, i32* %P2 + %P2 = select i1 %C, ptr %P, ptr null + %V = load i32, ptr %P2 ret i32 %V } -define i32 @test16_no_null_opt_2(i1 %C, i32* %P) #0 { +define i32 @test16_no_null_opt_2(i1 %C, ptr %P) #0 { ; CHECK-LABEL: @test16_no_null_opt_2( -; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], i32* null, i32* [[P:%.*]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P2]], align 4 +; CHECK-NEXT: [[P2:%.*]] = select i1 [[C:%.*]], ptr null, ptr [[P:%.*]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P2]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %P2 = select i1 %C, i32* null, i32* %P - %V = load i32, i32* %P2 + %P2 = select i1 %C, ptr null, ptr %P + %V = load i32, ptr %P2 ret i32 %V } attributes #0 = { null_pointer_is_valid } -define i1 @test17(i32* %X, i1 %C) { +define i1 @test17(ptr %X, i1 %C) { ; CHECK-LABEL: @test17( -; CHECK-NEXT: [[RV1:%.*]] = icmp eq i32* [[X:%.*]], null +; CHECK-NEXT: [[RV1:%.*]] = icmp eq ptr [[X:%.*]], null ; CHECK-NEXT: [[NOT_C:%.*]] = xor i1 [[C:%.*]], true ; CHECK-NEXT: [[RV:%.*]] = select i1 [[NOT_C]], i1 true, i1 [[RV1]] ; CHECK-NEXT: ret i1 [[RV]] ; - %R = select i1 %C, i32* %X, i32* null - %RV = icmp eq i32* %R, null + %R = select i1 %C, ptr %X, ptr null + %RV = icmp eq ptr %R, null ret i1 %RV } @@ -658,8 +658,8 @@ define i1 @test38(i1 %cond) { ; %zero = alloca i32 %one = alloca i32 - %ptr = select i1 %cond, i32* %zero, i32* %one - %isnull = icmp eq i32* %ptr, null + %ptr = select i1 %cond, ptr %zero, ptr %one + %isnull = icmp eq ptr %ptr, null ret i1 %isnull } @@ -679,8 +679,8 @@ define i1 @test40(i1 %cond) { %a = alloca i32 %b = alloca i32 %c = alloca i32 - %s = select i1 %cond, i32* %a, i32* %b - %r = icmp eq i32* %s, %c + %s = select i1 %cond, ptr %a, ptr %b + %r = icmp eq ptr %s, %c ret i1 %r } @@ -748,34 +748,31 @@ define @bitcast_select_bitcast( %icmp, %bc2 } -define void @select_oneuse_bitcast( %a, %b, %c, %d, * %ptr1) { +define void @select_oneuse_bitcast( %a, %b, %c, %d, ptr %ptr1) { ; CHECK-LABEL: @select_oneuse_bitcast( ; CHECK-NEXT: [[CMP:%.*]] = icmp ult [[C:%.*]], [[D:%.*]] ; CHECK-NEXT: [[SEL1_V:%.*]] = select [[CMP]], [[A:%.*]], [[B:%.*]] -; CHECK-NEXT: [[TMP1:%.*]] = bitcast * [[PTR1:%.*]] to * -; CHECK-NEXT: store [[SEL1_V]], * [[TMP1]], align 16 +; CHECK-NEXT: store [[SEL1_V]], ptr [[PTR1:%.*]], align 16 ; CHECK-NEXT: ret void ; %cmp = icmp ult %c, %d %bc1 = bitcast %a to %bc2 = bitcast %b to %sel1 = select %cmp, %bc1, %bc2 - store %sel1, * %ptr1 + store %sel1, ptr %ptr1 ret void } ; Allow select promotion even if there are multiple uses of bitcasted ops. ; Hoisting the selects allows later pattern matching to see that these are min/max ops. -define void @min_max_bitcast(<4 x float> %a, <4 x float> %b, <4 x i32>* %ptr1, <4 x i32>* %ptr2) { +define void @min_max_bitcast(<4 x float> %a, <4 x float> %b, ptr %ptr1, ptr %ptr2) { ; CHECK-LABEL: @min_max_bitcast( ; CHECK-NEXT: [[CMP:%.*]] = fcmp olt <4 x float> [[A:%.*]], [[B:%.*]] ; CHECK-NEXT: [[SEL1_V:%.*]] = select <4 x i1> [[CMP]], <4 x float> [[A]], <4 x float> [[B]] ; CHECK-NEXT: [[SEL2_V:%.*]] = select <4 x i1> [[CMP]], <4 x float> [[B]], <4 x float> [[A]] -; CHECK-NEXT: [[TMP1:%.*]] = bitcast <4 x i32>* [[PTR1:%.*]] to <4 x float>* -; CHECK-NEXT: store <4 x float> [[SEL1_V]], <4 x float>* [[TMP1]], align 16 -; CHECK-NEXT: [[TMP2:%.*]] = bitcast <4 x i32>* [[PTR2:%.*]] to <4 x float>* -; CHECK-NEXT: store <4 x float> [[SEL2_V]], <4 x float>* [[TMP2]], align 16 +; CHECK-NEXT: store <4 x float> [[SEL1_V]], ptr [[PTR1:%.*]], align 16 +; CHECK-NEXT: store <4 x float> [[SEL2_V]], ptr [[PTR2:%.*]], align 16 ; CHECK-NEXT: ret void ; %cmp = fcmp olt <4 x float> %a, %b @@ -783,20 +780,18 @@ define void @min_max_bitcast(<4 x float> %a, <4 x float> %b, <4 x i32>* %ptr1, < %bc2 = bitcast <4 x float> %b to <4 x i32> %sel1 = select <4 x i1> %cmp, <4 x i32> %bc1, <4 x i32> %bc2 %sel2 = select <4 x i1> %cmp, <4 x i32> %bc2, <4 x i32> %bc1 - store <4 x i32> %sel1, <4 x i32>* %ptr1 - store <4 x i32> %sel2, <4 x i32>* %ptr2 + store <4 x i32> %sel1, ptr %ptr1 + store <4 x i32> %sel2, ptr %ptr2 ret void } -define void @min_max_bitcast1( %a, %b, * %ptr1, * %ptr2) { +define void @min_max_bitcast1( %a, %b, ptr %ptr1, ptr %ptr2) { ; CHECK-LABEL: @min_max_bitcast1( ; CHECK-NEXT: [[CMP:%.*]] = fcmp olt [[A:%.*]], [[B:%.*]] ; CHECK-NEXT: [[SEL1_V:%.*]] = select [[CMP]], [[A]], [[B]] ; CHECK-NEXT: [[SEL2_V:%.*]] = select [[CMP]], [[B]], [[A]] -; CHECK-NEXT: [[TMP1:%.*]] = bitcast * [[PTR1:%.*]] to * -; CHECK-NEXT: store [[SEL1_V]], * [[TMP1]], align 16 -; CHECK-NEXT: [[TMP2:%.*]] = bitcast * [[PTR2:%.*]] to * -; CHECK-NEXT: store [[SEL2_V]], * [[TMP2]], align 16 +; CHECK-NEXT: store [[SEL1_V]], ptr [[PTR1:%.*]], align 16 +; CHECK-NEXT: store [[SEL2_V]], ptr [[PTR2:%.*]], align 16 ; CHECK-NEXT: ret void ; %cmp = fcmp olt %a, %b @@ -804,22 +799,22 @@ define void @min_max_bitcast1( %a, %b, %bc2 = bitcast %b to %sel1 = select %cmp, %bc1, %bc2 %sel2 = select %cmp, %bc2, %bc1 - store %sel1, * %ptr1 - store %sel2, * %ptr2 + store %sel1, ptr %ptr1 + store %sel2, ptr %ptr2 ret void } ; To avoid potential backend problems, we don't do the same transform for other casts. -define void @truncs_before_selects(<4 x float> %f1, <4 x float> %f2, <4 x i64> %a, <4 x i64> %b, <4 x i32>* %ptr1, <4 x i32>* %ptr2) { +define void @truncs_before_selects(<4 x float> %f1, <4 x float> %f2, <4 x i64> %a, <4 x i64> %b, ptr %ptr1, ptr %ptr2) { ; CHECK-LABEL: @truncs_before_selects( ; CHECK-NEXT: [[CMP:%.*]] = fcmp olt <4 x float> [[F1:%.*]], [[F2:%.*]] ; CHECK-NEXT: [[BC1:%.*]] = trunc <4 x i64> [[A:%.*]] to <4 x i32> ; CHECK-NEXT: [[BC2:%.*]] = trunc <4 x i64> [[B:%.*]] to <4 x i32> ; CHECK-NEXT: [[SEL1:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[BC1]], <4 x i32> [[BC2]] ; CHECK-NEXT: [[SEL2:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[BC2]], <4 x i32> [[BC1]] -; CHECK-NEXT: store <4 x i32> [[SEL1]], <4 x i32>* [[PTR1:%.*]], align 16 -; CHECK-NEXT: store <4 x i32> [[SEL2]], <4 x i32>* [[PTR2:%.*]], align 16 +; CHECK-NEXT: store <4 x i32> [[SEL1]], ptr [[PTR1:%.*]], align 16 +; CHECK-NEXT: store <4 x i32> [[SEL2]], ptr [[PTR2:%.*]], align 16 ; CHECK-NEXT: ret void ; %cmp = fcmp olt <4 x float> %f1, %f2 @@ -827,8 +822,8 @@ define void @truncs_before_selects(<4 x float> %f1, <4 x float> %f2, <4 x i64> % %bc2 = trunc <4 x i64> %b to <4 x i32> %sel1 = select <4 x i1> %cmp, <4 x i32> %bc1, <4 x i32> %bc2 %sel2 = select <4 x i1> %cmp, <4 x i32> %bc2, <4 x i32> %bc1 - store <4 x i32> %sel1, <4 x i32>* %ptr1, align 16 - store <4 x i32> %sel2, <4 x i32>* %ptr2, align 16 + store <4 x i32> %sel1, ptr %ptr1, align 16 + store <4 x i32> %sel2, ptr %ptr2, align 16 ret void } @@ -931,28 +926,28 @@ define i32 @test59(i32 %x, i32 %y) { ret i32 %.and } -define i1 @test60(i32 %x, i1* %y) { +define i1 @test60(i32 %x, ptr %y) { ; CHECK-LABEL: @test60( ; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[X:%.*]], 0 -; CHECK-NEXT: [[LOAD:%.*]] = load i1, i1* [[Y:%.*]], align 1 +; CHECK-NEXT: [[LOAD:%.*]] = load i1, ptr [[Y:%.*]], align 1 ; CHECK-NEXT: [[CMP1:%.*]] = icmp slt i32 [[X]], 1 ; CHECK-NEXT: [[SEL:%.*]] = select i1 [[CMP]], i1 [[LOAD]], i1 [[CMP1]] ; CHECK-NEXT: ret i1 [[SEL]] ; %cmp = icmp eq i32 %x, 0 - %load = load i1, i1* %y, align 1 + %load = load i1, ptr %y, align 1 %cmp1 = icmp slt i32 %x, 1 %sel = select i1 %cmp, i1 %load, i1 %cmp1 ret i1 %sel } @glbl = constant i32 10 -define i32 @test61(i32* %ptr) { +define i32 @test61(ptr %ptr) { ; CHECK-LABEL: @test61( ; CHECK-NEXT: ret i32 10 ; - %A = load i32, i32* %ptr - %B = icmp eq i32* %ptr, @glbl + %A = load i32, ptr %ptr + %B = icmp eq ptr %ptr, @glbl %C = select i1 %B, i32 %A, i32 10 ret i32 %C } @@ -1008,134 +1003,130 @@ while.body: ; The load here must not be speculated around the select. One side of the ; select is trivially dereferenceable but may have a lower alignment than the ; load does. -define i32 @test76(i1 %flag, i32* %x) { +define i32 @test76(i1 %flag, ptr %x) { ; CHECK-LABEL: @test76( -; CHECK-NEXT: store i32 0, i32* [[X:%.*]], align 4 -; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], i32* @under_aligned, i32* [[X]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P]], align 4 +; CHECK-NEXT: store i32 0, ptr [[X:%.*]], align 4 +; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], ptr @under_aligned, ptr [[X]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - store i32 0, i32* %x - %p = select i1 %flag, i32* @under_aligned, i32* %x - %v = load i32, i32* %p + store i32 0, ptr %x + %p = select i1 %flag, ptr @under_aligned, ptr %x + %v = load i32, ptr %p ret i32 %v } -declare void @scribble_on_i32(i32*) +declare void @scribble_on_i32(ptr) ; The load here must not be speculated around the select. One side of the ; select is trivially dereferenceable but may have a lower alignment than the ; load does. -define i32 @test77(i1 %flag, i32* %x) { +define i32 @test77(i1 %flag, ptr %x) { ; CHECK-LABEL: @test77( ; CHECK-NEXT: [[UNDER_ALIGNED:%.*]] = alloca i32, align 1 -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[UNDER_ALIGNED]]) -; CHECK-NEXT: store i32 0, i32* [[X:%.*]], align 4 -; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], i32* [[UNDER_ALIGNED]], i32* [[X]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P]], align 4 +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[UNDER_ALIGNED]]) +; CHECK-NEXT: store i32 0, ptr [[X:%.*]], align 4 +; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], ptr [[UNDER_ALIGNED]], ptr [[X]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; %under_aligned = alloca i32, align 1 - call void @scribble_on_i32(i32* %under_aligned) - store i32 0, i32* %x - %p = select i1 %flag, i32* %under_aligned, i32* %x - %v = load i32, i32* %p + call void @scribble_on_i32(ptr %under_aligned) + store i32 0, ptr %x + %p = select i1 %flag, ptr %under_aligned, ptr %x + %v = load i32, ptr %p ret i32 %v } -define i32 @test78(i1 %flag, i32* %x, i32* %y, i32* %z) { +define i32 @test78(i1 %flag, ptr %x, ptr %y, ptr %z) { ; Test that we can speculate the loads around the select even when we can't ; fold the load completely away. ; CHECK-LABEL: @test78( ; CHECK-NEXT: entry: -; CHECK-NEXT: store i32 0, i32* [[X:%.*]], align 4 -; CHECK-NEXT: store i32 0, i32* [[Y:%.*]], align 4 -; CHECK-NEXT: store i32 42, i32* [[Z:%.*]], align 4 -; CHECK-NEXT: [[X_VAL:%.*]] = load i32, i32* [[X]], align 4 -; CHECK-NEXT: [[Y_VAL:%.*]] = load i32, i32* [[Y]], align 4 +; CHECK-NEXT: store i32 0, ptr [[X:%.*]], align 4 +; CHECK-NEXT: store i32 0, ptr [[Y:%.*]], align 4 +; CHECK-NEXT: store i32 42, ptr [[Z:%.*]], align 4 +; CHECK-NEXT: [[X_VAL:%.*]] = load i32, ptr [[X]], align 4 +; CHECK-NEXT: [[Y_VAL:%.*]] = load i32, ptr [[Y]], align 4 ; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], i32 [[X_VAL]], i32 [[Y_VAL]] ; CHECK-NEXT: ret i32 [[V]] ; entry: - store i32 0, i32* %x - store i32 0, i32* %y + store i32 0, ptr %x + store i32 0, ptr %y ; Block forwarding by storing to %z which could alias either %x or %y. - store i32 42, i32* %z - %p = select i1 %flag, i32* %x, i32* %y - %v = load i32, i32* %p + store i32 42, ptr %z + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p ret i32 %v } ; Test that we can speculate the loads around the select even when we can't ; fold the load completely away. -define i32 @test78_deref(i1 %flag, i32* dereferenceable(4) align 4 %x, i32* dereferenceable(4) align 4 %y, i32* %z) nofree nosync { +define i32 @test78_deref(i1 %flag, ptr dereferenceable(4) align 4 %x, ptr dereferenceable(4) align 4 %y, ptr %z) nofree nosync { ; CHECK-LABEL: @test78_deref( -; CHECK-NEXT: [[X_VAL:%.*]] = load i32, i32* [[X:%.*]], align 4 -; CHECK-NEXT: [[Y_VAL:%.*]] = load i32, i32* [[Y:%.*]], align 4 +; CHECK-NEXT: [[X_VAL:%.*]] = load i32, ptr [[X:%.*]], align 4 +; CHECK-NEXT: [[Y_VAL:%.*]] = load i32, ptr [[Y:%.*]], align 4 ; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], i32 [[X_VAL]], i32 [[Y_VAL]] ; CHECK-NEXT: ret i32 [[V]] ; - %p = select i1 %flag, i32* %x, i32* %y - %v = load i32, i32* %p + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p ret i32 %v } ; The same as @test78 but we can't speculate the load because it can trap ; if under-aligned. -define i32 @test78_neg(i1 %flag, i32* %x, i32* %y, i32* %z) { +define i32 @test78_neg(i1 %flag, ptr %x, ptr %y, ptr %z) { ; CHECK-LABEL: @test78_neg( -; CHECK-NEXT: store i32 0, i32* [[X:%.*]], align 4 -; CHECK-NEXT: store i32 0, i32* [[Y:%.*]], align 4 -; CHECK-NEXT: store i32 42, i32* [[Z:%.*]], align 4 -; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], i32* [[X]], i32* [[Y]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P]], align 16 +; CHECK-NEXT: store i32 0, ptr [[X:%.*]], align 4 +; CHECK-NEXT: store i32 0, ptr [[Y:%.*]], align 4 +; CHECK-NEXT: store i32 42, ptr [[Z:%.*]], align 4 +; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], ptr [[X]], ptr [[Y]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P]], align 16 ; CHECK-NEXT: ret i32 [[V]] ; - store i32 0, i32* %x - store i32 0, i32* %y + store i32 0, ptr %x + store i32 0, ptr %y ; Block forwarding by storing to %z which could alias either %x or %y. - store i32 42, i32* %z - %p = select i1 %flag, i32* %x, i32* %y - %v = load i32, i32* %p, align 16 + store i32 42, ptr %z + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p, align 16 ret i32 %v } ; The same as @test78_deref but we can't speculate the load because ; one of the arguments is not sufficiently dereferenceable. -define i32 @test78_deref_neg(i1 %flag, i32* dereferenceable(2) %x, i32* dereferenceable(4) %y, i32* %z) nofree nosync { +define i32 @test78_deref_neg(i1 %flag, ptr dereferenceable(2) %x, ptr dereferenceable(4) %y, ptr %z) nofree nosync { ; CHECK-LABEL: @test78_deref_neg( -; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], i32* [[X:%.*]], i32* [[Y:%.*]] -; CHECK-NEXT: [[V:%.*]] = load i32, i32* [[P]], align 4 +; CHECK-NEXT: [[P:%.*]] = select i1 [[FLAG:%.*]], ptr [[X:%.*]], ptr [[Y:%.*]] +; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[P]], align 4 ; CHECK-NEXT: ret i32 [[V]] ; - %p = select i1 %flag, i32* %x, i32* %y - %v = load i32, i32* %p + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p ret i32 %v } ; Test that we can speculate the loads around the select even when we can't ; fold the load completely away. -define float @test79(i1 %flag, float* %x, i32* %y, i32* %z) { +define float @test79(i1 %flag, ptr %x, ptr %y, ptr %z) { ; CHECK-LABEL: @test79( -; CHECK-NEXT: [[X1:%.*]] = bitcast float* [[X:%.*]] to i32* -; CHECK-NEXT: [[Y1:%.*]] = bitcast i32* [[Y:%.*]] to float* -; CHECK-NEXT: store i32 0, i32* [[X1]], align 4 -; CHECK-NEXT: store i32 0, i32* [[Y]], align 4 -; CHECK-NEXT: store i32 42, i32* [[Z:%.*]], align 4 -; CHECK-NEXT: [[X_VAL:%.*]] = load float, float* [[X]], align 4 -; CHECK-NEXT: [[Y1_VAL:%.*]] = load float, float* [[Y1]], align 4 -; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], float [[X_VAL]], float [[Y1_VAL]] +; CHECK-NEXT: store i32 0, ptr [[X:%.*]], align 4 +; CHECK-NEXT: store i32 0, ptr [[Y:%.*]], align 4 +; CHECK-NEXT: store i32 42, ptr [[Z:%.*]], align 4 +; CHECK-NEXT: [[X_VAL:%.*]] = load float, ptr [[X]], align 4 +; CHECK-NEXT: [[Y_VAL:%.*]] = load float, ptr [[Y]], align 4 +; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], float [[X_VAL]], float [[Y_VAL]] ; CHECK-NEXT: ret float [[V]] ; - %x1 = bitcast float* %x to i32* - %y1 = bitcast i32* %y to float* - store i32 0, i32* %x1 - store i32 0, i32* %y + store i32 0, ptr %x + store i32 0, ptr %y ; Block forwarding by storing to %z which could alias either %x or %y. - store i32 42, i32* %z - %p = select i1 %flag, float* %x, float* %y1 - %v = load float, float* %p + store i32 42, ptr %z + %p = select i1 %flag, ptr %x, ptr %y + %v = load float, ptr %p ret float %v } @@ -1145,20 +1136,20 @@ define i32 @test80(i1 %flag) { ; CHECK-LABEL: @test80( ; CHECK-NEXT: [[X:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[Y:%.*]] = alloca i32, align 4 -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[X]]) -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[Y]]) -; CHECK-NEXT: [[T:%.*]] = load i32, i32* [[X]], align 4 -; CHECK-NEXT: store i32 [[T]], i32* [[Y]], align 4 +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load i32, ptr [[X]], align 4 +; CHECK-NEXT: store i32 [[T]], ptr [[Y]], align 4 ; CHECK-NEXT: ret i32 [[T]] ; %x = alloca i32 %y = alloca i32 - call void @scribble_on_i32(i32* %x) - call void @scribble_on_i32(i32* %y) - %t = load i32, i32* %x - store i32 %t, i32* %y - %p = select i1 %flag, i32* %x, i32* %y - %v = load i32, i32* %p + call void @scribble_on_i32(ptr %x) + call void @scribble_on_i32(ptr %y) + %t = load i32, ptr %x + store i32 %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p ret i32 %v } @@ -1166,25 +1157,23 @@ define i32 @test80(i1 %flag) { ; differently typed pointers. define float @test81(i1 %flag) { ; CHECK-LABEL: @test81( -; CHECK-NEXT: [[X:%.*]] = alloca i32, align 4 +; CHECK-NEXT: [[X:%.*]] = alloca float, align 4 ; CHECK-NEXT: [[Y:%.*]] = alloca i32, align 4 -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[X]]) -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[Y]]) -; CHECK-NEXT: [[T:%.*]] = load i32, i32* [[X]], align 4 -; CHECK-NEXT: store i32 [[T]], i32* [[Y]], align 4 +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load i32, ptr [[X]], align 4 +; CHECK-NEXT: store i32 [[T]], ptr [[Y]], align 4 ; CHECK-NEXT: [[V:%.*]] = bitcast i32 [[T]] to float ; CHECK-NEXT: ret float [[V]] ; %x = alloca float %y = alloca i32 - %x1 = bitcast float* %x to i32* - %y1 = bitcast i32* %y to float* - call void @scribble_on_i32(i32* %x1) - call void @scribble_on_i32(i32* %y) - %t = load i32, i32* %x1 - store i32 %t, i32* %y - %p = select i1 %flag, float* %x, float* %y1 - %v = load float, float* %p + call void @scribble_on_i32(ptr %x) + call void @scribble_on_i32(ptr %y) + %t = load i32, ptr %x + store i32 %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load float, ptr %p ret float %v } @@ -1194,118 +1183,100 @@ define i32 @test82(i1 %flag) { ; CHECK-LABEL: @test82( ; CHECK-NEXT: [[X:%.*]] = alloca float, align 4 ; CHECK-NEXT: [[Y:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[X1:%.*]] = bitcast float* [[X]] to i32* -; CHECK-NEXT: [[Y1:%.*]] = bitcast i32* [[Y]] to float* -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[X1]]) -; CHECK-NEXT: call void @scribble_on_i32(i32* nonnull [[Y]]) -; CHECK-NEXT: [[T:%.*]] = load float, float* [[X]], align 4 -; CHECK-NEXT: store float [[T]], float* [[Y1]], align 4 +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i32(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load float, ptr [[X]], align 4 +; CHECK-NEXT: store float [[T]], ptr [[Y]], align 4 ; CHECK-NEXT: [[V:%.*]] = bitcast float [[T]] to i32 ; CHECK-NEXT: ret i32 [[V]] ; %x = alloca float %y = alloca i32 - %x1 = bitcast float* %x to i32* - %y1 = bitcast i32* %y to float* - call void @scribble_on_i32(i32* %x1) - call void @scribble_on_i32(i32* %y) - %t = load float, float* %x - store float %t, float* %y1 - %p = select i1 %flag, i32* %x1, i32* %y - %v = load i32, i32* %p + call void @scribble_on_i32(ptr %x) + call void @scribble_on_i32(ptr %y) + %t = load float, ptr %x + store float %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load i32, ptr %p ret i32 %v } -declare void @scribble_on_i64(i64*) -declare void @scribble_on_i128(i128*) +declare void @scribble_on_i64(ptr) +declare void @scribble_on_i128(ptr) ; Test that we can speculate the load around the select even though they use ; differently typed pointers and requires inttoptr casts. -define i8* @test83(i1 %flag) { +define ptr @test83(i1 %flag) { ; CHECK-LABEL: @test83( -; CHECK-NEXT: [[X:%.*]] = alloca i8*, align 8 -; CHECK-NEXT: [[Y:%.*]] = alloca i8*, align 8 -; CHECK-NEXT: [[TMPCAST:%.*]] = bitcast i8** [[Y]] to i64* -; CHECK-NEXT: [[X1:%.*]] = bitcast i8** [[X]] to i64* -; CHECK-NEXT: call void @scribble_on_i64(i64* nonnull [[X1]]) -; CHECK-NEXT: call void @scribble_on_i64(i64* nonnull [[TMPCAST]]) -; CHECK-NEXT: [[T:%.*]] = load i64, i64* [[X1]], align 8 -; CHECK-NEXT: store i64 [[T]], i64* [[TMPCAST]], align 8 -; CHECK-NEXT: [[V:%.*]] = inttoptr i64 [[T]] to i8* -; CHECK-NEXT: ret i8* [[V]] -; - %x = alloca i8* +; CHECK-NEXT: [[X:%.*]] = alloca ptr, align 8 +; CHECK-NEXT: [[Y:%.*]] = alloca i64, align 8 +; CHECK-NEXT: call void @scribble_on_i64(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i64(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load i64, ptr [[X]], align 8 +; CHECK-NEXT: store i64 [[T]], ptr [[Y]], align 8 +; CHECK-NEXT: [[V:%.*]] = inttoptr i64 [[T]] to ptr +; CHECK-NEXT: ret ptr [[V]] +; + %x = alloca ptr %y = alloca i64 - %x1 = bitcast i8** %x to i64* - %y1 = bitcast i64* %y to i8** - call void @scribble_on_i64(i64* %x1) - call void @scribble_on_i64(i64* %y) - %t = load i64, i64* %x1 - store i64 %t, i64* %y - %p = select i1 %flag, i8** %x, i8** %y1 - %v = load i8*, i8** %p - ret i8* %v + call void @scribble_on_i64(ptr %x) + call void @scribble_on_i64(ptr %y) + %t = load i64, ptr %x + store i64 %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load ptr, ptr %p + ret ptr %v } ; Test that we can speculate the load around the select even though they use ; differently typed pointers and requires a ptrtoint cast. define i64 @test84(i1 %flag) { ; CHECK-LABEL: @test84( -; CHECK-NEXT: [[X:%.*]] = alloca i8*, align 8 -; CHECK-NEXT: [[Y:%.*]] = alloca i8*, align 8 -; CHECK-NEXT: [[TMPCAST:%.*]] = bitcast i8** [[Y]] to i64* -; CHECK-NEXT: [[X1:%.*]] = bitcast i8** [[X]] to i64* -; CHECK-NEXT: call void @scribble_on_i64(i64* nonnull [[X1]]) -; CHECK-NEXT: call void @scribble_on_i64(i64* nonnull [[TMPCAST]]) -; CHECK-NEXT: [[T:%.*]] = load i8*, i8** [[X]], align 8 -; CHECK-NEXT: store i8* [[T]], i8** [[Y]], align 8 -; CHECK-NEXT: [[V:%.*]] = ptrtoint i8* [[T]] to i64 +; CHECK-NEXT: [[X:%.*]] = alloca ptr, align 8 +; CHECK-NEXT: [[Y:%.*]] = alloca i64, align 8 +; CHECK-NEXT: call void @scribble_on_i64(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i64(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load ptr, ptr [[X]], align 8 +; CHECK-NEXT: store ptr [[T]], ptr [[Y]], align 8 +; CHECK-NEXT: [[V:%.*]] = ptrtoint ptr [[T]] to i64 ; CHECK-NEXT: ret i64 [[V]] ; - %x = alloca i8* + %x = alloca ptr %y = alloca i64 - %x1 = bitcast i8** %x to i64* - %y1 = bitcast i64* %y to i8** - call void @scribble_on_i64(i64* %x1) - call void @scribble_on_i64(i64* %y) - %t = load i8*, i8** %x - store i8* %t, i8** %y1 - %p = select i1 %flag, i64* %x1, i64* %y - %v = load i64, i64* %p + call void @scribble_on_i64(ptr %x) + call void @scribble_on_i64(ptr %y) + %t = load ptr, ptr %x + store ptr %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load i64, ptr %p ret i64 %v } ; Test that we can't speculate the load around the select. The load of the ; pointer doesn't load all of the stored integer bits. We could fix this, but it ; would require endianness checks and other nastiness. -define i8* @test85(i1 %flag) { +define ptr @test85(i1 %flag) { ; CHECK-LABEL: @test85( -; CHECK-NEXT: [[X1:%.*]] = alloca [2 x i8*], align 8 +; CHECK-NEXT: [[X:%.*]] = alloca [2 x ptr], align 8 ; CHECK-NEXT: [[Y:%.*]] = alloca i128, align 8 -; CHECK-NEXT: [[X1_SUB:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[X1]], i64 0, i64 0 -; CHECK-NEXT: [[X2:%.*]] = bitcast [2 x i8*]* [[X1]] to i128* -; CHECK-NEXT: [[Y1:%.*]] = bitcast i128* [[Y]] to i8** -; CHECK-NEXT: call void @scribble_on_i128(i128* nonnull [[X2]]) -; CHECK-NEXT: call void @scribble_on_i128(i128* nonnull [[Y]]) -; CHECK-NEXT: [[T:%.*]] = load i128, i128* [[X2]], align 8 -; CHECK-NEXT: store i128 [[T]], i128* [[Y]], align 8 -; CHECK-NEXT: [[X1_SUB_VAL:%.*]] = load i8*, i8** [[X1_SUB]], align 8 -; CHECK-NEXT: [[Y1_VAL:%.*]] = load i8*, i8** [[Y1]], align 8 -; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], i8* [[X1_SUB_VAL]], i8* [[Y1_VAL]] -; CHECK-NEXT: ret i8* [[V]] -; - %x = alloca [2 x i8*] +; CHECK-NEXT: call void @scribble_on_i128(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i128(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load i128, ptr [[X]], align 8 +; CHECK-NEXT: store i128 [[T]], ptr [[Y]], align 8 +; CHECK-NEXT: [[X_VAL:%.*]] = load ptr, ptr [[X]], align 8 +; CHECK-NEXT: [[Y_VAL:%.*]] = load ptr, ptr [[Y]], align 8 +; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], ptr [[X_VAL]], ptr [[Y_VAL]] +; CHECK-NEXT: ret ptr [[V]] +; + %x = alloca [2 x ptr] %y = alloca i128 - %x1 = bitcast [2 x i8*]* %x to i8** - %x2 = bitcast i8** %x1 to i128* - %y1 = bitcast i128* %y to i8** - call void @scribble_on_i128(i128* %x2) - call void @scribble_on_i128(i128* %y) - %t = load i128, i128* %x2 - store i128 %t, i128* %y - %p = select i1 %flag, i8** %x1, i8** %y1 - %v = load i8*, i8** %p - ret i8* %v + call void @scribble_on_i128(ptr %x) + call void @scribble_on_i128(ptr %y) + %t = load i128, ptr %x + store i128 %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load ptr, ptr %p + ret ptr %v } ; Test that we can't speculate the load around the select when the integer size @@ -1313,31 +1284,25 @@ define i8* @test85(i1 %flag) { ; the bits of the integer. define i128 @test86(i1 %flag) { ; CHECK-LABEL: @test86( -; CHECK-NEXT: [[X1:%.*]] = alloca [2 x i8*], align 8 +; CHECK-NEXT: [[X:%.*]] = alloca [2 x ptr], align 8 ; CHECK-NEXT: [[Y:%.*]] = alloca i128, align 8 -; CHECK-NEXT: [[X1_SUB:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[X1]], i64 0, i64 0 -; CHECK-NEXT: [[X2:%.*]] = bitcast [2 x i8*]* [[X1]] to i128* -; CHECK-NEXT: [[Y1:%.*]] = bitcast i128* [[Y]] to i8** -; CHECK-NEXT: call void @scribble_on_i128(i128* nonnull [[X2]]) -; CHECK-NEXT: call void @scribble_on_i128(i128* nonnull [[Y]]) -; CHECK-NEXT: [[T:%.*]] = load i8*, i8** [[X1_SUB]], align 8 -; CHECK-NEXT: store i8* [[T]], i8** [[Y1]], align 8 -; CHECK-NEXT: [[X2_VAL:%.*]] = load i128, i128* [[X2]], align 8 -; CHECK-NEXT: [[Y_VAL:%.*]] = load i128, i128* [[Y]], align 8 -; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], i128 [[X2_VAL]], i128 [[Y_VAL]] +; CHECK-NEXT: call void @scribble_on_i128(ptr nonnull [[X]]) +; CHECK-NEXT: call void @scribble_on_i128(ptr nonnull [[Y]]) +; CHECK-NEXT: [[T:%.*]] = load ptr, ptr [[X]], align 8 +; CHECK-NEXT: store ptr [[T]], ptr [[Y]], align 8 +; CHECK-NEXT: [[X_VAL:%.*]] = load i128, ptr [[X]], align 8 +; CHECK-NEXT: [[Y_VAL:%.*]] = load i128, ptr [[Y]], align 8 +; CHECK-NEXT: [[V:%.*]] = select i1 [[FLAG:%.*]], i128 [[X_VAL]], i128 [[Y_VAL]] ; CHECK-NEXT: ret i128 [[V]] ; - %x = alloca [2 x i8*] + %x = alloca [2 x ptr] %y = alloca i128 - %x1 = bitcast [2 x i8*]* %x to i8** - %x2 = bitcast i8** %x1 to i128* - %y1 = bitcast i128* %y to i8** - call void @scribble_on_i128(i128* %x2) - call void @scribble_on_i128(i128* %y) - %t = load i8*, i8** %x1 - store i8* %t, i8** %y1 - %p = select i1 %flag, i128* %x2, i128* %y - %v = load i128, i128* %p + call void @scribble_on_i128(ptr %x) + call void @scribble_on_i128(ptr %y) + %t = load ptr, ptr %x + store ptr %t, ptr %y + %p = select i1 %flag, ptr %x, ptr %y + %v = load i128, ptr %p ret i128 %v } @@ -1392,28 +1357,28 @@ define i32 @PR23757_swapped(i32 %x) { ret i32 %sel } -define i32 @PR23757_ne(i32 %x, i1* %p) { +define i32 @PR23757_ne(i32 %x, ptr %p) { ; CHECK-LABEL: @PR23757_ne( ; CHECK-NEXT: [[CMP:%.*]] = icmp ne i32 [[X:%.*]], 2147483647 -; CHECK-NEXT: store i1 [[CMP]], i1* [[P:%.*]], align 1 +; CHECK-NEXT: store i1 [[CMP]], ptr [[P:%.*]], align 1 ; CHECK-NEXT: ret i32 -2147483648 ; %cmp = icmp ne i32 %x, 2147483647 - store i1 %cmp, i1* %p ; thwart predicate canonicalization + store i1 %cmp, ptr %p ; thwart predicate canonicalization %add = add nsw i32 %x, 1 %sel = select i1 %cmp, i32 -2147483648, i32 %add ret i32 %sel } -define i32 @PR23757_ne_swapped(i32 %x, i1* %p) { +define i32 @PR23757_ne_swapped(i32 %x, ptr %p) { ; CHECK-LABEL: @PR23757_ne_swapped( ; CHECK-NEXT: [[CMP:%.*]] = icmp ne i32 [[X:%.*]], 2147483647 -; CHECK-NEXT: store i1 [[CMP]], i1* [[P:%.*]], align 1 +; CHECK-NEXT: store i1 [[CMP]], ptr [[P:%.*]], align 1 ; CHECK-NEXT: [[ADD:%.*]] = add i32 [[X]], 1 ; CHECK-NEXT: ret i32 [[ADD]] ; %cmp = icmp ne i32 %x, 2147483647 - store i1 %cmp, i1* %p ; thwart predicate canonicalization + store i1 %cmp, ptr %p ; thwart predicate canonicalization %add = add nsw i32 %x, 1 %sel = select i1 %cmp, i32 %add, i32 -2147483648 ret i32 %sel @@ -1519,19 +1484,19 @@ define <4 x i32> @undef_elts_in_condition(<4 x i32> %a, <4 x i32> %b) { define <4 x i32> @cannot_canonicalize_to_shuffle1(<4 x i32> %a, <4 x i32> %b) { ; CHECK-LABEL: @cannot_canonicalize_to_shuffle1( -; CHECK-NEXT: [[SEL:%.*]] = select <4 x i1> bitcast (i4 ptrtoint (i32* @g to i4) to <4 x i1>), <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]] +; CHECK-NEXT: [[SEL:%.*]] = select <4 x i1> bitcast (i4 ptrtoint (ptr @g to i4) to <4 x i1>), <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]] ; CHECK-NEXT: ret <4 x i32> [[SEL]] ; - %sel = select <4 x i1> bitcast (i4 ptrtoint (i32* @g to i4) to <4 x i1>), <4 x i32> %a, <4 x i32> %b + %sel = select <4 x i1> bitcast (i4 ptrtoint (ptr @g to i4) to <4 x i1>), <4 x i32> %a, <4 x i32> %b ret <4 x i32> %sel } define <4 x i32> @cannot_canonicalize_to_shuffle2(<4 x i32> %a, <4 x i32> %b) { ; CHECK-LABEL: @cannot_canonicalize_to_shuffle2( -; CHECK-NEXT: [[SEL:%.*]] = select <4 x i1> , <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]] +; CHECK-NEXT: [[SEL:%.*]] = select <4 x i1> , <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]] ; CHECK-NEXT: ret <4 x i32> [[SEL]] ; - %sel = select <4 x i1> , <4 x i32> %a, <4 x i32> %b + %sel = select <4 x i1> , <4 x i32> %a, <4 x i32> %b ret <4 x i32> %sel } @@ -2203,7 +2168,7 @@ merge: declare i32 @__gxx_personality_v0(...) declare i1 @foo() -define i32 @test_invoke_neg(i32 %x, i32 %y) nounwind uwtable ssp personality i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*) { +define i32 @test_invoke_neg(i32 %x, i32 %y) nounwind uwtable ssp personality ptr @__gxx_personality_v0 { ; CHECK-LABEL: @test_invoke_neg( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[COND:%.*]] = invoke i1 @foo() @@ -2232,7 +2197,7 @@ lpad: declare i32 @bar() -define i32 @test_invoke_2_neg(i1 %cond, i32 %x, i32 %y) nounwind uwtable ssp personality i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*) { +define i32 @test_invoke_2_neg(i1 %cond, i32 %x, i32 %y) nounwind uwtable ssp personality ptr @__gxx_personality_v0 { ; CHECK-LABEL: @test_invoke_2_neg( ; CHECK-NEXT: entry: ; CHECK-NEXT: br i1 [[COND:%.*]], label [[IF_TRUE:%.*]], label [[IF_FALSE:%.*]] @@ -2508,12 +2473,12 @@ exit: ret i32 %sel } -define i32 @test_select_into_phi_not_idom_no_dom_input_1(i1 %cond, i32 %A, i32 %B, i32 *%p) { +define i32 @test_select_into_phi_not_idom_no_dom_input_1(i1 %cond, i32 %A, i32 %B, ptr %p) { ; CHECK-LABEL: @test_select_into_phi_not_idom_no_dom_input_1( ; CHECK-NEXT: entry: ; CHECK-NEXT: br i1 [[COND:%.*]], label [[IF_TRUE:%.*]], label [[IF_FALSE:%.*]] ; CHECK: if.true: -; CHECK-NEXT: [[C:%.*]] = load i32, i32* [[P:%.*]], align 4 +; CHECK-NEXT: [[C:%.*]] = load i32, ptr [[P:%.*]], align 4 ; CHECK-NEXT: br label [[MERGE:%.*]] ; CHECK: if.false: ; CHECK-NEXT: br label [[MERGE]] @@ -2527,7 +2492,7 @@ entry: br i1 %cond, label %if.true, label %if.false if.true: - %C = load i32, i32* %p + %C = load i32, ptr %p br label %merge if.false: @@ -2542,14 +2507,14 @@ exit: ret i32 %sel } -define i32 @test_select_into_phi_not_idom_no_dom_input_2(i1 %cond, i32 %A, i32 %B, i32 *%p) { +define i32 @test_select_into_phi_not_idom_no_dom_input_2(i1 %cond, i32 %A, i32 %B, ptr %p) { ; CHECK-LABEL: @test_select_into_phi_not_idom_no_dom_input_2( ; CHECK-NEXT: entry: ; CHECK-NEXT: br i1 [[COND:%.*]], label [[IF_TRUE:%.*]], label [[IF_FALSE:%.*]] ; CHECK: if.true: ; CHECK-NEXT: br label [[MERGE:%.*]] ; CHECK: if.false: -; CHECK-NEXT: [[C:%.*]] = load i32, i32* [[P:%.*]], align 4 +; CHECK-NEXT: [[C:%.*]] = load i32, ptr [[P:%.*]], align 4 ; CHECK-NEXT: br label [[MERGE]] ; CHECK: merge: ; CHECK-NEXT: [[SEL:%.*]] = phi i32 [ [[C]], [[IF_FALSE]] ], [ [[B:%.*]], [[IF_TRUE]] ] @@ -2564,7 +2529,7 @@ if.true: br label %merge if.false: - %C = load i32, i32* %p + %C = load i32, ptr %p br label %merge merge: @@ -2928,15 +2893,15 @@ define i32 @select_replacement_loop2(i32 %arg, i32 %arg2) { } ; TODO: Dropping the inbounds flag should not be necessary for this fold. -define i8* @select_replacement_gep_inbounds(i8* %base, i64 %offset) { +define ptr @select_replacement_gep_inbounds(ptr %base, i64 %offset) { ; CHECK-LABEL: @select_replacement_gep_inbounds( -; CHECK-NEXT: [[GEP:%.*]] = getelementptr i8, i8* [[BASE:%.*]], i64 [[OFFSET:%.*]] -; CHECK-NEXT: ret i8* [[GEP]] +; CHECK-NEXT: [[GEP:%.*]] = getelementptr i8, ptr [[BASE:%.*]], i64 [[OFFSET:%.*]] +; CHECK-NEXT: ret ptr [[GEP]] ; %cmp = icmp eq i64 %offset, 0 - %gep = getelementptr inbounds i8, i8* %base, i64 %offset - %sel = select i1 %cmp, i8* %base, i8* %gep - ret i8* %sel + %gep = getelementptr inbounds i8, ptr %base, i64 %offset + %sel = select i1 %cmp, ptr %base, ptr %gep + ret ptr %sel } define <2 x i1> @partial_true_undef_condval(<2 x i1> %x) { @@ -3234,22 +3199,22 @@ define <2 x i8> @ne0_is_all_ones_swap_vec_poison(<2 x i8> %x) { define i64 @udiv_of_select_constexpr(i1 %c, i64 %x) { ; CHECK-LABEL: @udiv_of_select_constexpr( -; CHECK-NEXT: [[SEL:%.*]] = select i1 [[C:%.*]], i64 [[X:%.*]], i64 ptrtoint (i32* @glbl to i64) +; CHECK-NEXT: [[SEL:%.*]] = select i1 [[C:%.*]], i64 [[X:%.*]], i64 ptrtoint (ptr @glbl to i64) ; CHECK-NEXT: [[OP:%.*]] = udiv i64 [[SEL]], 3 ; CHECK-NEXT: ret i64 [[OP]] ; - %sel = select i1 %c, i64 %x, i64 ptrtoint (i32* @glbl to i64) + %sel = select i1 %c, i64 %x, i64 ptrtoint (ptr @glbl to i64) %op = udiv i64 %sel, 3 ret i64 %op } define i64 @udiv_of_select_constexpr_commuted(i1 %c, i64 %x) { ; CHECK-LABEL: @udiv_of_select_constexpr_commuted( -; CHECK-NEXT: [[SEL:%.*]] = select i1 [[C:%.*]], i64 ptrtoint (i32* @glbl to i64), i64 [[X:%.*]] +; CHECK-NEXT: [[SEL:%.*]] = select i1 [[C:%.*]], i64 ptrtoint (ptr @glbl to i64), i64 [[X:%.*]] ; CHECK-NEXT: [[OP:%.*]] = udiv i64 [[SEL]], 3 ; CHECK-NEXT: ret i64 [[OP]] ; - %sel = select i1 %c, i64 ptrtoint (i32* @glbl to i64), i64 %x + %sel = select i1 %c, i64 ptrtoint (ptr @glbl to i64), i64 %x %op = udiv i64 %sel, 3 ret i64 %op } diff --git a/llvm/test/Transforms/InstCombine/sink_sideeffecting_instruction.ll b/llvm/test/Transforms/InstCombine/sink_sideeffecting_instruction.ll index 9a3ee2e..f6262fd 100644 --- a/llvm/test/Transforms/InstCombine/sink_sideeffecting_instruction.ll +++ b/llvm/test/Transforms/InstCombine/sink_sideeffecting_instruction.ll @@ -2,17 +2,17 @@ ; RUN: opt -passes=instcombine -S < %s | FileCheck %s ; Function Attrs: noinline uwtable -define i32 @foo(i32* nocapture writeonly %arg) { +define i32 @foo(ptr nocapture writeonly %arg) { ; CHECK-LABEL: @foo( ; CHECK-NEXT: bb: ; CHECK-NEXT: [[VAR:%.*]] = call i32 @baz() -; CHECK-NEXT: store i32 [[VAR]], i32* [[ARG:%.*]], align 4 +; CHECK-NEXT: store i32 [[VAR]], ptr [[ARG:%.*]], align 4 ; CHECK-NEXT: [[VAR1:%.*]] = call i32 @baz() ; CHECK-NEXT: ret i32 [[VAR1]] ; bb: %var = call i32 @baz() - store i32 %var, i32* %arg, align 4 + store i32 %var, ptr %arg, align 4 %var1 = call i32 @baz() ret i32 %var1 } @@ -49,17 +49,15 @@ define i32 @test() { ; CHECK-NEXT: bb: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[VAR1:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[VAR2:%.*]] = bitcast i32* [[VAR]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[VAR2]]) -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @foo(i32* nonnull writeonly [[VAR]]) +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR]]) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @foo(ptr nonnull writeonly [[VAR]]) ; CHECK-NEXT: [[VAR4:%.*]] = icmp eq i32 [[VAR3]], 0 ; CHECK-NEXT: br i1 [[VAR4]], label [[BB5:%.*]], label [[BB14:%.*]] ; CHECK: bb5: -; CHECK-NEXT: [[VAR6:%.*]] = bitcast i32* [[VAR1]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[VAR6]]) -; CHECK-NEXT: [[VAR8:%.*]] = load i32, i32* [[VAR]], align 4 +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR1]]) +; CHECK-NEXT: [[VAR8:%.*]] = load i32, ptr [[VAR]], align 4 ; CHECK-NEXT: [[VAR9:%.*]] = icmp eq i32 [[VAR8]], 0 -; CHECK-NEXT: [[VAR7:%.*]] = call i32 @foo(i32* nonnull writeonly [[VAR1]]) +; CHECK-NEXT: [[VAR7:%.*]] = call i32 @foo(ptr nonnull writeonly [[VAR1]]) ; CHECK-NEXT: br i1 [[VAR9]], label [[BB10:%.*]], label [[BB_CRIT_EDGE:%.*]] ; CHECK: bb10: ; CHECK-NEXT: [[VAR11:%.*]] = call i32 @bar() @@ -68,28 +66,26 @@ define i32 @test() { ; CHECK-NEXT: br label [[BB12]] ; CHECK: bb12: ; CHECK-NEXT: [[VAR13:%.*]] = phi i32 [ [[VAR11]], [[BB10]] ], [ [[VAR7]], [[BB_CRIT_EDGE]] ] -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[VAR6]]) +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR1]]) ; CHECK-NEXT: br label [[BB14]] ; CHECK: bb14: ; CHECK-NEXT: [[VAR15:%.*]] = phi i32 [ [[VAR13]], [[BB12]] ], [ 0, [[BB:%.*]] ] -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[VAR2]]) +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR]]) ; CHECK-NEXT: ret i32 [[VAR15]] ; bb: %var = alloca i32, align 4 %var1 = alloca i32, align 4 - %var2 = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %var2) #4 - %var3 = call i32 @foo(i32* nonnull writeonly %var) + call void @llvm.lifetime.start.p0(i64 4, ptr nonnull %var) #4 + %var3 = call i32 @foo(ptr nonnull writeonly %var) %var4 = icmp eq i32 %var3, 0 br i1 %var4, label %bb5, label %bb14 bb5: ; preds = %bb - %var6 = bitcast i32* %var1 to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %var6) #4 - %var8 = load i32, i32* %var, align 4 + call void @llvm.lifetime.start.p0(i64 4, ptr nonnull %var1) #4 + %var8 = load i32, ptr %var, align 4 %var9 = icmp eq i32 %var8, 0 - %var7 = call i32 @foo(i32* nonnull writeonly %var1) + %var7 = call i32 @foo(ptr nonnull writeonly %var1) br i1 %var9, label %bb10, label %bb_crit_edge bb10: ; preds = %bb5 @@ -101,17 +97,17 @@ bb_crit_edge: bb12: ; preds = %bb10, %bb5 %var13 = phi i32 [ %var11, %bb10 ], [ %var7, %bb_crit_edge ] - call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %var6) #4 + call void @llvm.lifetime.end.p0(i64 4, ptr nonnull %var1) #4 br label %bb14 bb14: ; preds = %bb12, %bb %var15 = phi i32 [ %var13, %bb12 ], [ 0, %bb ] - call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %var2) + call void @llvm.lifetime.end.p0(i64 4, ptr nonnull %var) ret i32 %var15 } -declare i32 @unknown(i32* %dest) -declare i32 @unknown.as2(i32 addrspace(2)* %dest) +declare i32 @unknown(ptr %dest) +declare i32 @unknown.as2(ptr addrspace(2) %dest) define i32 @sink_write_to_use(i1 %c) { ; CHECK-LABEL: @sink_write_to_use( @@ -121,12 +117,12 @@ define i32 @sink_write_to_use(i1 %c) { ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull writeonly [[VAR]]) #[[ATTR1:[0-9]+]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull writeonly [[VAR]]) #[[ATTR1:[0-9]+]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i32, align 4 - %var3 = call i32 @unknown(i32* writeonly %var) argmemonly nounwind willreturn + %var3 = call i32 @unknown(ptr writeonly %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -144,12 +140,12 @@ define i32 @sink_readwrite_to_use(i1 %c) { ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i32, align 4 - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind willreturn + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -167,14 +163,12 @@ define i32 @sink_bitcast(i1 %c) { ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[BITCAST:%.*]] = bitcast i8* [[VAR]] to i32* -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[BITCAST]]) #[[ATTR1]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i8, align 8 - %bitcast = bitcast i8* %var to i32* - %var3 = call i32 @unknown(i32* %bitcast) argmemonly nounwind willreturn + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -188,20 +182,19 @@ use_block: define i32 @sink_gep1(i1 %c) { ; CHECK-LABEL: @sink_gep1( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[VAR1:%.*]] = alloca [2 x i32], align 8 +; CHECK-NEXT: [[VAR:%.*]] = alloca i64, align 8 ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[VAR1]], i64 0, i64 1 -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[GEP]]) #[[ATTR1]] +; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[VAR]], i64 1 +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[GEP]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i64, align 8 - %bitcast = bitcast i64* %var to i32* - %gep = getelementptr i32, i32* %bitcast, i32 1 - %var3 = call i32 @unknown(i32* %gep) argmemonly nounwind willreturn + %gep = getelementptr i32, ptr %var, i32 1 + %var3 = call i32 @unknown(ptr %gep) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -214,19 +207,17 @@ use_block: define i32 @sink_gep2(i1 %c) { ; CHECK-LABEL: @sink_gep2( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[VAR1:%.*]] = alloca [2 x i32], align 8 +; CHECK-NEXT: [[VAR:%.*]] = alloca i64, align 8 ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[VAR1_SUB:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[VAR1]], i64 0, i64 0 -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR1_SUB]]) #[[ATTR1]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i64, align 8 - %bitcast = bitcast i64* %var to i32* - %var3 = call i32 @unknown(i32* %bitcast) argmemonly nounwind willreturn + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -244,14 +235,14 @@ define i32 @sink_addrspacecast(i1 %c) { ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[CAST:%.*]] = addrspacecast i32* [[VAR]] to i32 addrspace(2)* -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown.as2(i32 addrspace(2)* [[CAST]]) #[[ATTR1]] +; CHECK-NEXT: [[CAST:%.*]] = addrspacecast ptr [[VAR]] to ptr addrspace(2) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown.as2(ptr addrspace(2) [[CAST]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i32, align 8 - %cast = addrspacecast i32* %var to i32 addrspace(2)* - %var3 = call i32 @unknown.as2(i32 addrspace(2)* %cast) argmemonly nounwind willreturn + %cast = addrspacecast ptr %var to ptr addrspace(2) + %var3 = call i32 @unknown.as2(ptr addrspace(2) %cast) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -265,7 +256,7 @@ define i32 @neg_infinite_loop(i1 %c) { ; CHECK-LABEL: @neg_infinite_loop( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR2:[0-9]+]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR2:[0-9]+]] ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -274,7 +265,7 @@ define i32 @neg_infinite_loop(i1 %c) { ; entry: %var = alloca i32, align 4 - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind br i1 %c, label %early_return, label %use_block early_return: @@ -288,7 +279,7 @@ define i32 @neg_throw(i1 %c) { ; CHECK-LABEL: @neg_throw( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR3:[0-9]+]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR3:[0-9]+]] ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -297,7 +288,7 @@ define i32 @neg_throw(i1 %c) { ; entry: %var = alloca i32, align 4 - %var3 = call i32 @unknown(i32* %var) argmemonly willreturn + %var3 = call i32 @unknown(ptr %var) argmemonly willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -311,7 +302,7 @@ define i32 @neg_unknown_write(i1 %c) { ; CHECK-LABEL: @neg_unknown_write( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR4:[0-9]+]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR4:[0-9]+]] ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -320,7 +311,7 @@ define i32 @neg_unknown_write(i1 %c) { ; entry: %var = alloca i32, align 4 - %var3 = call i32 @unknown(i32* %var) nounwind willreturn + %var3 = call i32 @unknown(ptr %var) nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -334,28 +325,26 @@ define i32 @sink_lifetime1(i1 %c) { ; CHECK-LABEL: @sink_lifetime1( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[BITCAST:%.*]] = bitcast i32* [[VAR]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[BITCAST]]) -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR]]) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[BITCAST]]) +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR]]) ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i32, align 4 - %bitcast = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* %bitcast) - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind willreturn + call void @llvm.lifetime.start.p0(i64 4, ptr %var) + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: ret i32 0 use_block: - call void @llvm.lifetime.end.p0i8(i64 4, i8* %bitcast) + call void @llvm.lifetime.end.p0(i64 4, ptr %var) ret i32 %var3 } @@ -363,27 +352,25 @@ define i32 @sink_lifetime2(i1 %c) { ; CHECK-LABEL: @sink_lifetime2( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[BITCAST:%.*]] = bitcast i32* [[VAR]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[BITCAST]]) -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR]]) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: br i1 [[C:%.*]], label [[MERGE:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: merge: ; CHECK-NEXT: [[RET:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[VAR3]], [[USE_BLOCK]] ] -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[BITCAST]]) +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR]]) ; CHECK-NEXT: ret i32 [[RET]] ; CHECK: use_block: ; CHECK-NEXT: br label [[MERGE]] ; entry: %var = alloca i32, align 4 - %bitcast = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* %bitcast) - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind willreturn + call void @llvm.lifetime.start.p0(i64 4, ptr %var) + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %merge, label %use_block merge: %ret = phi i32 [0, %entry], [%var3, %use_block] - call void @llvm.lifetime.end.p0i8(i64 4, i8* %bitcast) + call void @llvm.lifetime.end.p0(i64 4, ptr %var) ret i32 %ret use_block: @@ -398,16 +385,15 @@ define i32 @sink_lifetime3(i1 %c) { ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 ; CHECK: use_block: -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] ; CHECK-NEXT: ret i32 [[VAR3]] ; entry: %var = alloca i32, align 4 - %bitcast = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* %bitcast) - call void @llvm.lifetime.end.p0i8(i64 4, i8* %bitcast) + call void @llvm.lifetime.start.p0(i64 4, ptr %var) + call void @llvm.lifetime.end.p0(i64 4, ptr %var) ; If unknown accesses %var, that's UB - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind willreturn + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn br i1 %c, label %early_return, label %use_block early_return: @@ -421,10 +407,9 @@ define i32 @sink_lifetime4a(i1 %c) { ; CHECK-LABEL: @sink_lifetime4a( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[BITCAST:%.*]] = bitcast i32* [[VAR]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[BITCAST]]) -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull [[VAR]]) #[[ATTR1]] -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[BITCAST]]) +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR]]) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR]]) ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -433,10 +418,9 @@ define i32 @sink_lifetime4a(i1 %c) { ; entry: %var = alloca i32, align 4 - %bitcast = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* %bitcast) - %var3 = call i32 @unknown(i32* %var) argmemonly nounwind willreturn - call void @llvm.lifetime.end.p0i8(i64 4, i8* %bitcast) + call void @llvm.lifetime.start.p0(i64 4, ptr %var) + %var3 = call i32 @unknown(ptr %var) argmemonly nounwind willreturn + call void @llvm.lifetime.end.p0(i64 4, ptr %var) br i1 %c, label %early_return, label %use_block early_return: @@ -452,10 +436,9 @@ define i32 @sink_lifetime4b(i1 %c) { ; CHECK-LABEL: @sink_lifetime4b( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: [[BITCAST:%.*]] = bitcast i32* [[VAR]] to i8* -; CHECK-NEXT: call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull [[BITCAST]]) -; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(i32* nonnull writeonly [[VAR]]) #[[ATTR1]] -; CHECK-NEXT: call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull [[BITCAST]]) +; CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[VAR]]) +; CHECK-NEXT: [[VAR3:%.*]] = call i32 @unknown(ptr nonnull writeonly [[VAR]]) #[[ATTR1]] +; CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[VAR]]) ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -464,10 +447,9 @@ define i32 @sink_lifetime4b(i1 %c) { ; entry: %var = alloca i32, align 4 - %bitcast = bitcast i32* %var to i8* - call void @llvm.lifetime.start.p0i8(i64 4, i8* %bitcast) - %var3 = call i32 @unknown(i32* writeonly %var) argmemonly nounwind willreturn - call void @llvm.lifetime.end.p0i8(i64 4, i8* %bitcast) + call void @llvm.lifetime.start.p0(i64 4, ptr %var) + %var3 = call i32 @unknown(ptr writeonly %var) argmemonly nounwind willreturn + call void @llvm.lifetime.end.p0(i64 4, ptr %var) br i1 %c, label %early_return, label %use_block early_return: @@ -481,8 +463,8 @@ define i32 @sink_atomicrmw_to_use(i1 %c) { ; CHECK-LABEL: @sink_atomicrmw_to_use( ; CHECK-NEXT: entry: ; CHECK-NEXT: [[VAR:%.*]] = alloca i32, align 4 -; CHECK-NEXT: store i32 0, i32* [[VAR]], align 4 -; CHECK-NEXT: [[VAR3:%.*]] = atomicrmw add i32* [[VAR]], i32 1 seq_cst, align 4 +; CHECK-NEXT: store i32 0, ptr [[VAR]], align 4 +; CHECK-NEXT: [[VAR3:%.*]] = atomicrmw add ptr [[VAR]], i32 1 seq_cst, align 4 ; CHECK-NEXT: br i1 [[C:%.*]], label [[EARLY_RETURN:%.*]], label [[USE_BLOCK:%.*]] ; CHECK: early_return: ; CHECK-NEXT: ret i32 0 @@ -491,8 +473,8 @@ define i32 @sink_atomicrmw_to_use(i1 %c) { ; entry: %var = alloca i32, align 4 - store i32 0, i32* %var - %var3 = atomicrmw add i32* %var, i32 1 seq_cst, align 4 + store i32 0, ptr %var + %var3 = atomicrmw add ptr %var, i32 1 seq_cst, align 4 br i1 %c, label %early_return, label %use_block early_return: @@ -504,6 +486,6 @@ use_block: declare i32 @bar() -declare void @llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) -declare void @llvm.lifetime.end.p0i8(i64 immarg, i8* nocapture) +declare void @llvm.lifetime.start.p0(i64 immarg, ptr nocapture) +declare void @llvm.lifetime.end.p0(i64 immarg, ptr nocapture) diff --git a/llvm/test/Transforms/InstCombine/stpcpy-1.ll b/llvm/test/Transforms/InstCombine/stpcpy-1.ll index cb178f3..86691a0 100644 --- a/llvm/test/Transforms/InstCombine/stpcpy-1.ll +++ b/llvm/test/Transforms/InstCombine/stpcpy-1.ll @@ -11,67 +11,59 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f3 @b = common global [32 x i8] zeroinitializer, align 1 @percent_s = constant [3 x i8] c"%s\00" -declare i8* @stpcpy(i8*, i8*) +declare ptr @stpcpy(ptr, ptr) -define i8* @test_simplify1() { +define ptr @test_simplify1() { ; CHECK-LABEL: @test_simplify1( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(6) getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(6) getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i32 6, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 5) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(6) @a, ptr noundef nonnull align 1 dereferenceable(6) @hello, i32 6, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([32 x i8], ptr @a, i32 0, i32 5) ; - %dst = getelementptr [32 x i8], [32 x i8]* @a, i32 0, i32 0 - %src = getelementptr [6 x i8], [6 x i8]* @hello, i32 0, i32 0 - %ret = call i8* @stpcpy(i8* %dst, i8* %src) - ret i8* %ret + %ret = call ptr @stpcpy(ptr @a, ptr @hello) + ret ptr %ret } -define i8* @test_simplify2() { +define ptr @test_simplify2() { ; CHECK-LABEL: @test_simplify2( -; CHECK-NEXT: [[STRLEN:%.*]] = call i32 @strlen(i8* noundef nonnull dereferenceable(1) getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 0)) -; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [32 x i8], [32 x i8]* @a, i32 0, i32 [[STRLEN]] -; CHECK-NEXT: ret i8* [[TMP1]] +; CHECK-NEXT: [[STRLEN:%.*]] = call i32 @strlen(ptr noundef nonnull dereferenceable(1) @a) +; CHECK-NEXT: [[RET:%.*]] = getelementptr inbounds i8, ptr @a, i32 [[STRLEN]] +; CHECK-NEXT: ret ptr [[RET]] ; - %dst = getelementptr [32 x i8], [32 x i8]* @a, i32 0, i32 0 - %ret = call i8* @stpcpy(i8* %dst, i8* %dst) - ret i8* %ret + %ret = call ptr @stpcpy(ptr @a, ptr @a) + ret ptr %ret } -define void @test_simplify3(i8* %dst) { +define void @test_simplify3(ptr %dst) { ; CHECK-LABEL: @test_simplify3( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(6) [[DST:%.*]], i8* noundef nonnull align 1 dereferenceable(6) getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i32 6, i1 false) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(6) [[DST:%.*]], ptr noundef nonnull align 1 dereferenceable(6) @hello, i32 6, i1 false) ; CHECK-NEXT: ret void ; - %src = getelementptr [6 x i8], [6 x i8]* @hello, i32 0, i32 0 - call i8* @stpcpy(i8* dereferenceable(80) %dst, i8* %src) + call ptr @stpcpy(ptr dereferenceable(80) %dst, ptr @hello) ret void } -define i8* @test_no_simplify1() { +define ptr @test_no_simplify1() { ; CHECK-LABEL: @test_no_simplify1( -; CHECK-NEXT: [[RET:%.*]] = call i8* @stpcpy(i8* nonnull getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([32 x i8], [32 x i8]* @b, i32 0, i32 0)) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = call ptr @stpcpy(ptr nonnull @a, ptr nonnull @b) +; CHECK-NEXT: ret ptr [[RET]] ; - %dst = getelementptr [32 x i8], [32 x i8]* @a, i32 0, i32 0 - %src = getelementptr [32 x i8], [32 x i8]* @b, i32 0, i32 0 - %ret = call i8* @stpcpy(i8* %dst, i8* %src) - ret i8* %ret + %ret = call ptr @stpcpy(ptr @a, ptr @b) + ret ptr %ret } -define i8* @test_no_simplify2(i8* %dst, i8* %src) { +define ptr @test_no_simplify2(ptr %dst, ptr %src) { ; CHECK-LABEL: @test_no_simplify2( -; CHECK-NEXT: [[RET:%.*]] = musttail call i8* @stpcpy(i8* [[DST:%.*]], i8* [[SRC:%.*]]) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = musttail call ptr @stpcpy(ptr [[DST:%.*]], ptr [[SRC:%.*]]) +; CHECK-NEXT: ret ptr [[RET]] ; - %ret = musttail call i8* @stpcpy(i8* %dst, i8* %src) - ret i8* %ret + %ret = musttail call ptr @stpcpy(ptr %dst, ptr %src) + ret ptr %ret } -define i8* @test_no_incompatible_attr() { +define ptr @test_no_incompatible_attr() { ; CHECK-LABEL: @test_no_incompatible_attr( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(6) getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(6) getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i32 6, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([32 x i8], [32 x i8]* @a, i32 0, i32 5) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(6) @a, ptr noundef nonnull align 1 dereferenceable(6) @hello, i32 6, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([32 x i8], ptr @a, i32 0, i32 5) ; - %dst = getelementptr [32 x i8], [32 x i8]* @a, i32 0, i32 0 - %src = getelementptr [6 x i8], [6 x i8]* @hello, i32 0, i32 0 - %ret = call dereferenceable(1) i8* @stpcpy(i8* %dst, i8* %src) - ret i8* %ret + %ret = call dereferenceable(1) ptr @stpcpy(ptr @a, ptr @hello) + ret ptr %ret } diff --git a/llvm/test/Transforms/InstCombine/stpcpy_chk-1.ll b/llvm/test/Transforms/InstCombine/stpcpy_chk-1.ll index 73a354d..5ebd9fa 100644 --- a/llvm/test/Transforms/InstCombine/stpcpy_chk-1.ll +++ b/llvm/test/Transforms/InstCombine/stpcpy_chk-1.ll @@ -12,138 +12,119 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f3 ; Check cases where slen >= strlen (src). -define i8* @test_simplify1() { +define ptr @test_simplify1() { ; CHECK-LABEL: @test_simplify1( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(12) @a, ptr noundef nonnull align 1 dereferenceable(12) @.str, i32 12, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([60 x i8], ptr @a, i32 0, i32 11) ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0 - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 60) - ret i8* %ret + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @.str, i32 60) + ret ptr %ret } -define i8* @test_simplify2() { +define ptr @test_simplify2() { ; CHECK-LABEL: @test_simplify2( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(12) @a, ptr noundef nonnull align 1 dereferenceable(12) @.str, i32 12, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([60 x i8], ptr @a, i32 0, i32 11) ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0 - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 12) - ret i8* %ret + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @.str, i32 12) + ret ptr %ret } -define i8* @test_simplify3() { +define ptr @test_simplify3() { ; CHECK-LABEL: @test_simplify3( -; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11) +; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(12) @a, ptr noundef nonnull align 1 dereferenceable(12) @.str, i32 12, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([60 x i8], ptr @a, i32 0, i32 11) ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0 - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) - ret i8* %ret + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @.str, i32 -1) + ret ptr %ret } -define i8* @test_simplify1_tail() { +define ptr @test_simplify1_tail() { ; CHECK-LABEL: @test_simplify1_tail( -; CHECK-NEXT: tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* noundef nonnull align 1 dereferenceable(12) getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i1 false) -; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11) +; CHECK-NEXT: tail call void @llvm.memcpy.p0.p0.i32(ptr noundef nonnull align 1 dereferenceable(12) @a, ptr noundef nonnull align 1 dereferenceable(12) @.str, i32 12, i1 false) +; CHECK-NEXT: ret ptr getelementptr inbounds ([60 x i8], ptr @a, i32 0, i32 11) ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0 - %ret = tail call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 60) - ret i8* %ret + %ret = tail call ptr @__stpcpy_chk(ptr @a, ptr @.str, i32 60) + ret ptr %ret } ; Check cases where there are no string constants. -define i8* @test_simplify4() { +define ptr @test_simplify4() { ; CHECK-LABEL: @test_simplify4( -; CHECK-NEXT: [[STPCPY:%.*]] = call i8* @stpcpy(i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0)) -; CHECK-NEXT: ret i8* [[STPCPY]] +; CHECK-NEXT: [[STPCPY:%.*]] = call ptr @stpcpy(ptr nonnull @a, ptr nonnull @b) +; CHECK-NEXT: ret ptr [[STPCPY]] ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0 - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) - ret i8* %ret + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @b, i32 -1) + ret ptr %ret } -define i8* @test_simplify4_tail() { +define ptr @test_simplify4_tail() { ; CHECK-LABEL: @test_simplify4_tail( -; CHECK-NEXT: [[STPCPY:%.*]] = tail call i8* @stpcpy(i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0)) -; CHECK-NEXT: ret i8* [[STPCPY]] +; CHECK-NEXT: [[STPCPY:%.*]] = tail call ptr @stpcpy(ptr nonnull @a, ptr nonnull @b) +; CHECK-NEXT: ret ptr [[STPCPY]] ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0 - %ret = tail call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) - ret i8* %ret + %ret = tail call ptr @__stpcpy_chk(ptr @a, ptr @b, i32 -1) + ret ptr %ret } ; Check case where the string length is not constant. -define i8* @test_simplify5() { +define ptr @test_simplify5() { ; CHECK-LABEL: @test_simplify5( -; CHECK-NEXT: [[LEN:%.*]] = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false, i1 false) -; CHECK-NEXT: [[TMP1:%.*]] = call i8* @__memcpy_chk(i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i32 [[LEN]]) -; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11) +; CHECK-NEXT: [[LEN:%.*]] = call i32 @llvm.objectsize.i32.p0(ptr @a, i1 false, i1 false, i1 false) +; CHECK-NEXT: [[TMP1:%.*]] = call ptr @__memcpy_chk(ptr nonnull @a, ptr nonnull @.str, i32 12, i32 [[LEN]]) +; CHECK-NEXT: ret ptr getelementptr inbounds ([60 x i8], ptr @a, i32 0, i32 11) ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0 - %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false, i1 false) - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 %len) - ret i8* %ret + %len = call i32 @llvm.objectsize.i32.p0(ptr @a, i1 false, i1 false, i1 false) + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @.str, i32 %len) + ret ptr %ret } ; Check case where the source and destination are the same. -define i8* @test_simplify6() { +define ptr @test_simplify6() { ; CHECK-LABEL: @test_simplify6( -; CHECK-NEXT: [[STRLEN:%.*]] = call i32 @strlen(i8* noundef nonnull dereferenceable(1) getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0)) -; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 [[STRLEN]] -; CHECK-NEXT: ret i8* [[TMP1]] +; CHECK-NEXT: [[STRLEN:%.*]] = call i32 @strlen(ptr noundef nonnull dereferenceable(1) @a) +; CHECK-NEXT: [[RET:%.*]] = getelementptr inbounds i8, ptr @a, i32 [[STRLEN]] +; CHECK-NEXT: ret ptr [[RET]] ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false, i1 false) - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %dst, i32 %len) - ret i8* %ret + %len = call i32 @llvm.objectsize.i32.p0(ptr @a, i1 false, i1 false, i1 false) + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @a, i32 %len) + ret ptr %ret } ; Check cases where there are no string constants, and is a tail call. -define i8* @test_simplify7() { +define ptr @test_simplify7() { ; CHECK-LABEL: @test_simplify7( -; CHECK-NEXT: [[STPCPY:%.*]] = tail call i8* @stpcpy(i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0)) -; CHECK-NEXT: ret i8* [[STPCPY]] +; CHECK-NEXT: [[STPCPY:%.*]] = tail call ptr @stpcpy(ptr nonnull @a, ptr nonnull @b) +; CHECK-NEXT: ret ptr [[STPCPY]] ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0 - %ret = tail call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) - ret i8* %ret + %ret = tail call ptr @__stpcpy_chk(ptr @a, ptr @b, i32 -1) + ret ptr %ret } ; Check case where slen < strlen (src). -define i8* @test_no_simplify1() { +define ptr @test_no_simplify1() { ; CHECK-LABEL: @test_no_simplify1( -; CHECK-NEXT: [[RET:%.*]] = call i8* @__stpcpy_chk(i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* nonnull getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0), i32 8) -; CHECK-NEXT: ret i8* [[RET]] +; CHECK-NEXT: [[RET:%.*]] = call ptr @__stpcpy_chk(ptr nonnull @a, ptr nonnull @b, i32 8) +; CHECK-NEXT: ret ptr [[RET]] ; - %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0 - %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 8) - ret i8* %ret + %ret = call ptr @__stpcpy_chk(ptr @a, ptr @b, i32 8) + ret ptr %ret } -declare i8* @__stpcpy_chk(i8*, i8*, i32) nounwind -declare i32 @llvm.objectsize.i32.p0i8(i8*, i1, i1, i1) nounwind readonly +declare ptr @__stpcpy_chk(ptr, ptr, i32) nounwind +declare i32 @llvm.objectsize.i32.p0(ptr, i1, i1, i1) nounwind readonly