diff options
Diffstat (limited to 'test/CodeGen/X86')
75 files changed, 1407 insertions, 1106 deletions
diff --git a/test/CodeGen/X86/2006-10-19-SwitchUnnecessaryBranching.ll b/test/CodeGen/X86/2006-10-19-SwitchUnnecessaryBranching.ll index 88e8b4a..d583e59 100644 --- a/test/CodeGen/X86/2006-10-19-SwitchUnnecessaryBranching.ll +++ b/test/CodeGen/X86/2006-10-19-SwitchUnnecessaryBranching.ll @@ -6,8 +6,8 @@ define i32 @test(i32 %argc, i8** %argv) nounwind { entry: ; CHECK: cmpl $2 -; CHECK-NEXT: je -; CHECK-NEXT: %entry +; CHECK-NEXT: jne +; CHECK-NEXT: %bb2 switch i32 %argc, label %UnifiedReturnBlock [ i32 1, label %bb diff --git a/test/CodeGen/X86/2008-05-01-InvalidOrdCompare.ll b/test/CodeGen/X86/2008-05-01-InvalidOrdCompare.ll index a708224..4160b20 100644 --- a/test/CodeGen/X86/2008-05-01-InvalidOrdCompare.ll +++ b/test/CodeGen/X86/2008-05-01-InvalidOrdCompare.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -enable-unsafe-fp-math -march=x86 | grep jnp +; RUN: llc < %s -enable-unsafe-fp-math -march=x86 | grep jp ; rdar://5902801 declare void @test2() diff --git a/test/CodeGen/X86/2010-06-28-DbgEntryPC.ll b/test/CodeGen/X86/2010-06-28-DbgEntryPC.ll deleted file mode 100644 index 628790c..0000000 --- a/test/CodeGen/X86/2010-06-28-DbgEntryPC.ll +++ /dev/null @@ -1,108 +0,0 @@ -; RUN: llc -O2 -mtriple=i386-apple-darwin <%s | FileCheck %s -; Use DW_FORM_addr for DW_AT_entry_pc. -; Radar 8094785 - -; CHECK: .byte 17 ## DW_TAG_compile_unit -; CHECK-NEXT: .byte 1 ## DW_CHILDREN_yes -; CHECK-NEXT: .byte 37 ## DW_AT_producer -; CHECK-NEXT: .byte 14 ## DW_FORM_strp -; CHECK-NEXT: .byte 19 ## DW_AT_language -; CHECK-NEXT: .byte 5 ## DW_FORM_data2 -; CHECK-NEXT: .byte 3 ## DW_AT_name -; CHECK-NEXT: .byte 14 ## DW_FORM_strp -; CHECK-NEXT: .byte 82 ## DW_AT_entry_pc -; CHECK-NEXT: .byte 1 ## DW_FORM_addr -; CHECK-NEXT: .byte 16 ## DW_AT_stmt_list -; CHECK-NEXT: .byte 6 ## DW_FORM_data4 -; CHECK-NEXT: .byte 27 ## DW_AT_comp_dir -; CHECK-NEXT: .byte 14 ## DW_FORM_strp -; CHECK-NEXT: .ascii "\341\177" ## DW_AT_APPLE_optimized - -%struct.a = type { i32, %struct.a* } - -@ret = common global i32 0 ; <i32*> [#uses=2] - -define void @foo(i32 %x) nounwind noinline ssp { -entry: - tail call void @llvm.dbg.value(metadata !{i32 %x}, i64 0, metadata !21), !dbg !28 - store i32 %x, i32* @ret, align 4, !dbg !29 - ret void, !dbg !31 -} - -declare void @llvm.dbg.declare(metadata, metadata) nounwind readnone - -declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone - -define i8* @bar(%struct.a* %b) nounwind noinline ssp { -entry: - tail call void @llvm.dbg.value(metadata !{%struct.a* %b}, i64 0, metadata !22), !dbg !32 - %0 = getelementptr inbounds %struct.a* %b, i64 0, i32 0, !dbg !33 ; <i32*> [#uses=1] - %1 = load i32* %0, align 8, !dbg !33 ; <i32> [#uses=1] - tail call void @foo(i32 %1) nounwind noinline ssp, !dbg !33 - %2 = bitcast %struct.a* %b to i8*, !dbg !35 ; <i8*> [#uses=1] - ret i8* %2, !dbg !35 -} - -define i32 @main(i32 %argc, i8** nocapture %argv) nounwind ssp { -entry: - %e = alloca %struct.a, align 8 ; <%struct.a*> [#uses=4] - call void @llvm.dbg.value(metadata !{i32 %argc}, i64 0, metadata !23), !dbg !36 - call void @llvm.dbg.value(metadata !{i8** %argv}, i64 0, metadata !24), !dbg !36 - call void @llvm.dbg.declare(metadata !{%struct.a* %e}, metadata !25), !dbg !37 - %0 = getelementptr inbounds %struct.a* %e, i64 0, i32 0, !dbg !38 ; <i32*> [#uses=1] - store i32 4, i32* %0, align 8, !dbg !38 - %1 = getelementptr inbounds %struct.a* %e, i64 0, i32 1, !dbg !39 ; <%struct.a**> [#uses=1] - store %struct.a* %e, %struct.a** %1, align 8, !dbg !39 - %2 = call i8* @bar(%struct.a* %e) nounwind noinline ssp, !dbg !40 ; <i8*> [#uses=0] - %3 = load i32* @ret, align 4, !dbg !41 ; <i32> [#uses=1] - ret i32 %3, !dbg !41 -} - -!llvm.dbg.sp = !{!0, !6, !15} -!llvm.dbg.lv.foo = !{!21} -!llvm.dbg.lv.bar = !{!22} -!llvm.dbg.lv.main = !{!23, !24, !25} -!llvm.dbg.gv = !{!27} - -!0 = metadata !{i32 524334, i32 0, metadata !1, metadata !"foo", metadata !"foo", metadata !"foo", metadata !1, i32 34, metadata !3, i1 false, i1 true, i32 0, i32 0, null, i1 false, i1 true, void (i32)* @foo} ; [ DW_TAG_subprogram ] -!1 = metadata !{i32 524329, metadata !"2010-06-28-DbgEntryPC.c", metadata !"/Users/yash/clean/llvm/test/FrontendC", metadata !2} ; [ DW_TAG_file_type ] -!2 = metadata !{i32 524305, i32 0, i32 1, metadata !"2010-06-28-DbgEntryPC.c", metadata !"/Users/yash/clean/llvm/test/FrontendC", metadata !"4.2.1 (Based on Apple Inc. build 5658) (LLVM build)", i1 true, i1 true, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] -!3 = metadata !{i32 524309, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !4, i32 0, null} ; [ DW_TAG_subroutine_type ] -!4 = metadata !{null, metadata !5} -!5 = metadata !{i32 524324, metadata !1, metadata !"int", metadata !1, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] -!6 = metadata !{i32 524334, i32 0, metadata !1, metadata !"bar", metadata !"bar", metadata !"bar", metadata !1, i32 38, metadata !7, i1 false, i1 true, i32 0, i32 0, null, i1 false, i1 true, i8* (%struct.a*)* @bar} ; [ DW_TAG_subprogram ] -!7 = metadata !{i32 524309, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !8, i32 0, null} ; [ DW_TAG_subroutine_type ] -!8 = metadata !{metadata !9, metadata !10} -!9 = metadata !{i32 524303, metadata !1, metadata !"", metadata !1, i32 0, i64 64, i64 64, i64 0, i32 0, null} ; [ DW_TAG_pointer_type ] -!10 = metadata !{i32 524303, metadata !1, metadata !"", metadata !1, i32 0, i64 64, i64 64, i64 0, i32 0, metadata !11} ; [ DW_TAG_pointer_type ] -!11 = metadata !{i32 524307, metadata !1, metadata !"a", metadata !1, i32 23, i64 128, i64 64, i64 0, i32 0, null, metadata !12, i32 0, null} ; [ DW_TAG_structure_type ] -!12 = metadata !{metadata !13, metadata !14} -!13 = metadata !{i32 524301, metadata !11, metadata !"c", metadata !1, i32 24, i64 32, i64 32, i64 0, i32 0, metadata !5} ; [ DW_TAG_member ] -!14 = metadata !{i32 524301, metadata !11, metadata !"d", metadata !1, i32 25, i64 64, i64 64, i64 64, i32 0, metadata !10} ; [ DW_TAG_member ] -!15 = metadata !{i32 524334, i32 0, metadata !1, metadata !"main", metadata !"main", metadata !"main", metadata !1, i32 43, metadata !16, i1 false, i1 true, i32 0, i32 0, null, i1 false, i1 true, i32 (i32, i8**)* @main} ; [ DW_TAG_subprogram ] -!16 = metadata !{i32 524309, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !17, i32 0, null} ; [ DW_TAG_subroutine_type ] -!17 = metadata !{metadata !5, metadata !5, metadata !18} -!18 = metadata !{i32 524303, metadata !1, metadata !"", metadata !1, i32 0, i64 64, i64 64, i64 0, i32 0, metadata !19} ; [ DW_TAG_pointer_type ] -!19 = metadata !{i32 524303, metadata !1, metadata !"", metadata !1, i32 0, i64 64, i64 64, i64 0, i32 0, metadata !20} ; [ DW_TAG_pointer_type ] -!20 = metadata !{i32 524324, metadata !1, metadata !"char", metadata !1, i32 0, i64 8, i64 8, i64 0, i32 0, i32 6} ; [ DW_TAG_base_type ] -!21 = metadata !{i32 524545, metadata !0, metadata !"x", metadata !1, i32 33, metadata !5} ; [ DW_TAG_arg_variable ] -!22 = metadata !{i32 524545, metadata !6, metadata !"b", metadata !1, i32 38, metadata !10} ; [ DW_TAG_arg_variable ] -!23 = metadata !{i32 524545, metadata !15, metadata !"argc", metadata !1, i32 43, metadata !5} ; [ DW_TAG_arg_variable ] -!24 = metadata !{i32 524545, metadata !15, metadata !"argv", metadata !1, i32 43, metadata !18} ; [ DW_TAG_arg_variable ] -!25 = metadata !{i32 524544, metadata !26, metadata !"e", metadata !1, i32 44, metadata !11} ; [ DW_TAG_auto_variable ] -!26 = metadata !{i32 524299, metadata !15, i32 43, i32 0} ; [ DW_TAG_lexical_block ] -!27 = metadata !{i32 524340, i32 0, metadata !1, metadata !"ret", metadata !"ret", metadata !"", metadata !1, i32 28, metadata !5, i1 false, i1 true, i32* @ret} ; [ DW_TAG_variable ] -!28 = metadata !{i32 33, i32 0, metadata !0, null} -!29 = metadata !{i32 35, i32 0, metadata !30, null} -!30 = metadata !{i32 524299, metadata !0, i32 34, i32 0} ; [ DW_TAG_lexical_block ] -!31 = metadata !{i32 36, i32 0, metadata !30, null} -!32 = metadata !{i32 38, i32 0, metadata !6, null} -!33 = metadata !{i32 39, i32 0, metadata !34, null} -!34 = metadata !{i32 524299, metadata !6, i32 38, i32 0} ; [ DW_TAG_lexical_block ] -!35 = metadata !{i32 40, i32 0, metadata !34, null} -!36 = metadata !{i32 43, i32 0, metadata !15, null} -!37 = metadata !{i32 44, i32 0, metadata !26, null} -!38 = metadata !{i32 45, i32 0, metadata !26, null} -!39 = metadata !{i32 46, i32 0, metadata !26, null} -!40 = metadata !{i32 48, i32 0, metadata !26, null} -!41 = metadata !{i32 49, i32 0, metadata !26, null} diff --git a/test/CodeGen/X86/2010-08-04-MaskedSignedCompare.ll b/test/CodeGen/X86/2010-08-04-MaskedSignedCompare.ll index 1919d2e..12a8274 100644 --- a/test/CodeGen/X86/2010-08-04-MaskedSignedCompare.ll +++ b/test/CodeGen/X86/2010-08-04-MaskedSignedCompare.ll @@ -17,7 +17,7 @@ entry: ; CHECK: andl $150 ; CHECK-NEXT: testb -; CHECK-NEXT: jg +; CHECK-NEXT: jle entry.if.end_crit_edge: ; preds = %entry %tmp4.pre = load i32* @g_38 ; <i32> [#uses=1] diff --git a/test/CodeGen/X86/2010-11-18-SelectOfExtload.ll b/test/CodeGen/X86/2010-11-18-SelectOfExtload.ll index a1074b6..6d54c7e 100644 --- a/test/CodeGen/X86/2010-11-18-SelectOfExtload.ll +++ b/test/CodeGen/X86/2010-11-18-SelectOfExtload.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86 | FileCheck %s +; RUN: llc < %s -march=x86 -mattr=+cmov | FileCheck %s ; Both values were being zero extended. @u = external global i8 @s = external global i8 diff --git a/test/CodeGen/X86/2011-09-14-valcoalesce.ll b/test/CodeGen/X86/2011-09-14-valcoalesce.ll index 1068d1b..a5ec614 100644 --- a/test/CodeGen/X86/2011-09-14-valcoalesce.ll +++ b/test/CodeGen/X86/2011-09-14-valcoalesce.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86 | FileCheck %s +; RUN: llc < %s -march=x86 -disable-code-place | FileCheck %s ; ; Test RegistersDefinedFromSameValue. We have multiple copies of the same vreg: ; while.body85.i: diff --git a/test/CodeGen/X86/2011-10-27-tstore.ll b/test/CodeGen/X86/2011-10-27-tstore.ll index 016e02c..6e83f67 100644 --- a/test/CodeGen/X86/2011-10-27-tstore.ll +++ b/test/CodeGen/X86/2011-10-27-tstore.ll @@ -3,14 +3,14 @@ target triple = "x86_64-unknown-linux-gnu" ;CHECK: ltstore -;CHECK: pshufd -;CHECK: pshufd +;CHECK: movq +;CHECK: movq ;CHECK: ret -define void @ltstore() { +define void @ltstore(<4 x i32>* %pA, <2 x i32>* %pB) { entry: - %in = load <4 x i32>* undef + %in = load <4 x i32>* %pA %j = shufflevector <4 x i32> %in, <4 x i32> undef, <2 x i32> <i32 0, i32 1> - store <2 x i32> %j, <2 x i32>* undef + store <2 x i32> %j, <2 x i32>* %pB ret void } diff --git a/test/CodeGen/X86/2012-02-20-MachineCPBug.ll b/test/CodeGen/X86/2012-02-20-MachineCPBug.ll index 3013f16..557d49d 100644 --- a/test/CodeGen/X86/2012-02-20-MachineCPBug.ll +++ b/test/CodeGen/X86/2012-02-20-MachineCPBug.ll @@ -52,7 +52,8 @@ for.body9.lr.ph.i: ; preds = %if.end.i, %for.cond for.body9.i: ; preds = %for.body9.i, %for.body9.lr.ph.i ; CHECK: %for.body9.i -; CHECK: movb %al, %ch +; CHECK: movb +; CHECK: shrdl %i6.02.i = phi i32 [ 0, %for.body9.lr.ph.i ], [ %inc21.i, %for.body9.i ] %add10.i = add i32 %i6.02.i, %div.i %arrayidx12.i = getelementptr inbounds i64* %.pre.i, i32 %add10.i diff --git a/test/CodeGen/X86/2012-03-26-PostRALICMBug.ll b/test/CodeGen/X86/2012-03-26-PostRALICMBug.ll new file mode 100644 index 0000000..101ecca --- /dev/null +++ b/test/CodeGen/X86/2012-03-26-PostRALICMBug.ll @@ -0,0 +1,59 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin10 -stats |& \ +; RUN: not grep {Number of machine instructions hoisted out of loops post regalloc} + +; rdar://11095580 + +%struct.ref_s = type { %union.color_sample, i16, i16 } +%union.color_sample = type { i64 } + +@table = external global [3891 x i64] + +declare i32 @foo() + +define i32 @zarray(%struct.ref_s* nocapture %op) nounwind ssp { +entry: + %call = tail call i32 @foo() + %tmp = ashr i32 %call, 31 + %0 = and i32 %tmp, 1396 + %index9 = add i32 %0, 2397 + indirectbr i8* undef, [label %return, label %if.end] + +if.end: ; preds = %entry + %size5 = getelementptr inbounds %struct.ref_s* %op, i64 0, i32 2 + %tmp6 = load i16* %size5, align 2 + %tobool1 = icmp eq i16 %tmp6, 0 + %1 = select i1 %tobool1, i32 1396, i32 -1910 + %index10 = add i32 %index9, %1 + indirectbr i8* undef, [label %return, label %while.body.lr.ph] + +while.body.lr.ph: ; preds = %if.end + %refs = bitcast %struct.ref_s* %op to %struct.ref_s** + %tmp9 = load %struct.ref_s** %refs, align 8 + %tmp4 = zext i16 %tmp6 to i64 + %index13 = add i32 %index10, 1658 + %2 = sext i32 %index13 to i64 + %3 = getelementptr [3891 x i64]* @table, i64 0, i64 %2 + %blockaddress14 = load i64* %3, align 8 + %4 = inttoptr i64 %blockaddress14 to i8* + indirectbr i8* %4, [label %while.body] + +while.body: ; preds = %while.body, %while.body.lr.ph + %index7 = phi i32 [ %index15, %while.body ], [ %index13, %while.body.lr.ph ] + %indvar = phi i64 [ %indvar.next, %while.body ], [ 0, %while.body.lr.ph ] + %type_attrs = getelementptr %struct.ref_s* %tmp9, i64 %indvar, i32 1 + store i16 32, i16* %type_attrs, align 2 + %indvar.next = add i64 %indvar, 1 + %exitcond5 = icmp eq i64 %indvar.next, %tmp4 + %tmp7 = select i1 %exitcond5, i32 1648, i32 0 + %index15 = add i32 %index7, %tmp7 + %tmp8 = select i1 %exitcond5, i64 13, i64 0 + %5 = sext i32 %index15 to i64 + %6 = getelementptr [3891 x i64]* @table, i64 0, i64 %5 + %blockaddress16 = load i64* %6, align 8 + %7 = inttoptr i64 %blockaddress16 to i8* + indirectbr i8* %7, [label %return, label %while.body] + +return: ; preds = %while.body, %if.end, %entry + %retval.0 = phi i32 [ %call, %entry ], [ 0, %if.end ], [ 0, %while.body ] + ret i32 %retval.0 +} diff --git a/test/CodeGen/X86/2012-04-09-TwoAddrPassBug.ll b/test/CodeGen/X86/2012-04-09-TwoAddrPassBug.ll new file mode 100644 index 0000000..2d90165 --- /dev/null +++ b/test/CodeGen/X86/2012-04-09-TwoAddrPassBug.ll @@ -0,0 +1,34 @@ +; RUN: llc -O1 -verify-coalescing < %s +; PR12495 +target datalayout = +"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.7.0" + +define void @func(i8* nocapture) nounwind uwtable ssp align 2 { + br i1 undef, label %4, label %2 + +; <label>:2 ; preds = %1 + %3 = tail call double @foo() nounwind + br label %4 + +; <label>:4 ; preds = %2, %1 + %5 = phi double [ %3, %2 ], [ 0.000000e+00, %1 ] + %6 = fsub double %5, undef + %7 = fcmp olt double %6, 0.000000e+00 + %8 = select i1 %7, double 0.000000e+00, double %6 + %9 = fcmp olt double undef, 0.000000e+00 + %10 = fcmp olt double %8, undef + %11 = or i1 %9, %10 + br i1 %11, label %12, label %14 + +; <label>:12 ; preds = %4 + %13 = tail call double @fmod(double %8, double 0.000000e+00) nounwind + unreachable + +; <label>:14 ; preds = %4 + ret void +} + +declare double @foo() + +declare double @fmod(double, double) diff --git a/test/CodeGen/X86/GC/lit.local.cfg b/test/CodeGen/X86/GC/lit.local.cfg index b05ed3c..a8ad0f1 100644 --- a/test/CodeGen/X86/GC/lit.local.cfg +++ b/test/CodeGen/X86/GC/lit.local.cfg @@ -1,13 +1,6 @@ config.suffixes = ['.ll', '.c', '.cpp'] -def getRoot(config): - if not config.parent: - return config - return getRoot(config.parent) - -root = getRoot(config) - -targets = set(root.targets_to_build.split()) +targets = set(config.root.targets_to_build.split()) if not 'X86' in targets: config.unsupported = True diff --git a/test/CodeGen/X86/SwizzleShuff.ll b/test/CodeGen/X86/SwizzleShuff.ll new file mode 100644 index 0000000..100817a --- /dev/null +++ b/test/CodeGen/X86/SwizzleShuff.ll @@ -0,0 +1,68 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s + +; Check that we perform a scalar XOR on i32. + +; CHECK: pull_bitcast +; CHECK: xorl +; CHECK: ret +define void @pull_bitcast (<4 x i8>* %pA, <4 x i8>* %pB) { + %A = load <4 x i8>* %pA + %B = load <4 x i8>* %pB + %C = xor <4 x i8> %A, %B + store <4 x i8> %C, <4 x i8>* %pA + ret void +} + +; CHECK: multi_use_swizzle +; CHECK: mov +; CHECK-NEXT: shuf +; CHECK-NEXT: shuf +; CHECK-NEXT: shuf +; CHECK-NEXT: xor +; CHECK-NEXT: ret +define <4 x i32> @multi_use_swizzle (<4 x i32>* %pA, <4 x i32>* %pB) { + %A = load <4 x i32>* %pA + %B = load <4 x i32>* %pB + %S = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 1, i32 5, i32 6> + %S1 = shufflevector <4 x i32> %S, <4 x i32> undef, <4 x i32> <i32 1, i32 3, i32 2, i32 2> + %S2 = shufflevector <4 x i32> %S, <4 x i32> undef, <4 x i32> <i32 2, i32 1, i32 0, i32 2> + %R = xor <4 x i32> %S1, %S2 + ret <4 x i32> %R +} + +; CHECK: pull_bitcast2 +; CHECK: xorl +; CHECK: ret +define <4 x i8> @pull_bitcast2 (<4 x i8>* %pA, <4 x i8>* %pB, <4 x i8>* %pC) { + %A = load <4 x i8>* %pA + store <4 x i8> %A, <4 x i8>* %pC + %B = load <4 x i8>* %pB + %C = xor <4 x i8> %A, %B + store <4 x i8> %C, <4 x i8>* %pA + ret <4 x i8> %C +} + + + +; CHECK: reverse_1 +; CHECK-NOT: shuf +; CHECK: ret +define <4 x i32> @reverse_1 (<4 x i32>* %pA, <4 x i32>* %pB) { + %A = load <4 x i32>* %pA + %B = load <4 x i32>* %pB + %S = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 0, i32 3, i32 2> + %S1 = shufflevector <4 x i32> %S, <4 x i32> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 2> + ret <4 x i32> %S1 +} + + +; CHECK: no_reverse_shuff +; CHECK: shuf +; CHECK: ret +define <4 x i32> @no_reverse_shuff (<4 x i32>* %pA, <4 x i32>* %pB) { + %A = load <4 x i32>* %pA + %B = load <4 x i32>* %pB + %S = shufflevector <4 x i32> %A, <4 x i32> %B, <4 x i32> <i32 1, i32 0, i32 3, i32 2> + %S1 = shufflevector <4 x i32> %S, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 3, i32 2> + ret <4 x i32> %S1 +} diff --git a/test/CodeGen/X86/atom-sched.ll b/test/CodeGen/X86/atom-sched.ll index 2301dfc..4dd9a9e 100644 --- a/test/CodeGen/X86/atom-sched.ll +++ b/test/CodeGen/X86/atom-sched.ll @@ -1,5 +1,9 @@ +; XFAIL: * ; RUN: llc <%s -O2 -mcpu=atom -march=x86 -relocation-model=static | FileCheck -check-prefix=atom %s ; RUN: llc <%s -O2 -mcpu=core2 -march=x86 -relocation-model=static | FileCheck %s +; +; FIXME: Atom's scheduler is temporarily disabled. +; XFAIL: * @a = common global i32 0, align 4 @b = common global i32 0, align 4 diff --git a/test/CodeGen/X86/atomic_op.ll b/test/CodeGen/X86/atomic_op.ll index 972dab2..7c5abe2 100644 --- a/test/CodeGen/X86/atomic_op.ll +++ b/test/CodeGen/X86/atomic_op.ll @@ -13,6 +13,7 @@ entry: %xort = alloca i32 ; <i32*> [#uses=2] %old = alloca i32 ; <i32*> [#uses=18] %temp = alloca i32 ; <i32*> [#uses=2] + %temp64 = alloca i64 store i32 %argc, i32* %argc.addr store i8** %argv, i8*** %argv.addr store i32 0, i32* %val1 @@ -106,6 +107,25 @@ entry: ; CHECK: cmpxchgl %17 = cmpxchg i32* %val2, i32 1976, i32 1 monotonic store i32 %17, i32* %old + ; CHECK: movl $1401, %[[R17mask:[a-z]*]] + ; CHECK: movl [[R17atomic:.*]], %eax + ; CHECK: movl %eax, %[[R17newval:[a-z]*]] + ; CHECK: andl %[[R17mask]], %[[R17newval]] + ; CHECK: notl %[[R17newval]] + ; CHECK: lock + ; CHECK: cmpxchgl %[[R17newval]], [[R17atomic]] + ; CHECK: jne + ; CHECK: movl %eax, + %18 = atomicrmw nand i32* %val2, i32 1401 monotonic + store i32 %18, i32* %old + ; CHECK: andl + ; CHECK: andl + ; CHECK: notl + ; CHECK: notl + ; CHECK: lock + ; CHECK: cmpxchg8b + %19 = atomicrmw nand i64* %temp64, i64 17361641481138401520 monotonic + store i64 %19, i64* %temp64 ret void } diff --git a/test/CodeGen/X86/avx-intrinsics-x86.ll b/test/CodeGen/X86/avx-intrinsics-x86.ll index 616601a..b334932 100644 --- a/test/CodeGen/X86/avx-intrinsics-x86.ll +++ b/test/CodeGen/X86/avx-intrinsics-x86.ll @@ -1078,33 +1078,33 @@ define <2 x i64> @test_x86_sse41_pmuldq(<4 x i32> %a0, <4 x i32> %a1) { declare <2 x i64> @llvm.x86.sse41.pmuldq(<4 x i32>, <4 x i32>) nounwind readnone -define i32 @test_x86_sse41_ptestc(<4 x float> %a0, <4 x float> %a1) { +define i32 @test_x86_sse41_ptestc(<2 x i64> %a0, <2 x i64> %a1) { ; CHECK: vptest ; CHECK: sbbl - %res = call i32 @llvm.x86.sse41.ptestc(<4 x float> %a0, <4 x float> %a1) ; <i32> [#uses=1] + %res = call i32 @llvm.x86.sse41.ptestc(<2 x i64> %a0, <2 x i64> %a1) ; <i32> [#uses=1] ret i32 %res } -declare i32 @llvm.x86.sse41.ptestc(<4 x float>, <4 x float>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestc(<2 x i64>, <2 x i64>) nounwind readnone -define i32 @test_x86_sse41_ptestnzc(<4 x float> %a0, <4 x float> %a1) { +define i32 @test_x86_sse41_ptestnzc(<2 x i64> %a0, <2 x i64> %a1) { ; CHECK: vptest ; CHECK: seta ; CHECK: movzbl - %res = call i32 @llvm.x86.sse41.ptestnzc(<4 x float> %a0, <4 x float> %a1) ; <i32> [#uses=1] + %res = call i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %a0, <2 x i64> %a1) ; <i32> [#uses=1] ret i32 %res } -declare i32 @llvm.x86.sse41.ptestnzc(<4 x float>, <4 x float>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestnzc(<2 x i64>, <2 x i64>) nounwind readnone -define i32 @test_x86_sse41_ptestz(<4 x float> %a0, <4 x float> %a1) { +define i32 @test_x86_sse41_ptestz(<2 x i64> %a0, <2 x i64> %a1) { ; CHECK: vptest ; CHECK: sete ; CHECK: movzbl - %res = call i32 @llvm.x86.sse41.ptestz(<4 x float> %a0, <4 x float> %a1) ; <i32> [#uses=1] + %res = call i32 @llvm.x86.sse41.ptestz(<2 x i64> %a0, <2 x i64> %a1) ; <i32> [#uses=1] ret i32 %res } -declare i32 @llvm.x86.sse41.ptestz(<4 x float>, <4 x float>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestz(<2 x i64>, <2 x i64>) nounwind readnone define <2 x double> @test_x86_sse41_round_pd(<2 x double> %a0) { diff --git a/test/CodeGen/X86/avx-shuffle.ll b/test/CodeGen/X86/avx-shuffle.ll index 54f01e9..16c447b 100644 --- a/test/CodeGen/X86/avx-shuffle.ll +++ b/test/CodeGen/X86/avx-shuffle.ll @@ -147,3 +147,58 @@ entry: %shuffle.i = shufflevector <4 x float> %1, <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 4, i32 4, i32 4> ret <8 x float> %shuffle.i } + +; PR12413 +; CHECK: vpshufb +; CHECK: vpshufb +; CHECK: vpshufb +; CHECK: vpshufb +define <32 x i8> @shuf(<32 x i8> %inval1, <32 x i8> %inval2) { +entry: + %0 = shufflevector <32 x i8> %inval1, <32 x i8> %inval2, <32 x i32> <i32 0, +i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 +22, i32 24, i32 26, i32 28, i32 30, i32 32, i32 34, i32 36, i32 38, i32 40, i32 +42, i32 44, i32 46, i32 48, i32 50, i32 52, i32 54, i32 56, i32 58, i32 60, i32 +62> + ret <32 x i8> %0 +} + +; CHECK: blend1 +; CHECK: vblendps +; CHECK: ret +define <4 x i32> @blend1(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline { + %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 7> + ret <4 x i32> %t +} + +; CHECK: blend2 +; CHECK: vblendps +; CHECK: ret +define <4 x i32> @blend2(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline { + %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7> + ret <4 x i32> %t +} + +; CHECK: blend2a +; CHECK: vblendps +; CHECK: ret +define <4 x float> @blend2a(<4 x float> %a, <4 x float> %b) nounwind alwaysinline { + %t = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7> + ret <4 x float> %t +} + +; CHECK: blend3 +; CHECK-NOT: vblendps +; CHECK: ret +define <4 x i32> @blend3(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline { + %t = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 1, i32 5, i32 2, i32 7> + ret <4 x i32> %t +} + +; CHECK: blend4 +; CHECK: vblendpd +; CHECK: ret +define <4 x i64> @blend4(<4 x i64> %a, <4 x i64> %b) nounwind alwaysinline { + %t = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 7> + ret <4 x i64> %t +} diff --git a/test/CodeGen/X86/avx-vbroadcast.ll b/test/CodeGen/X86/avx-vbroadcast.ll index 5bf9f4f..148ae73 100644 --- a/test/CodeGen/X86/avx-vbroadcast.ll +++ b/test/CodeGen/X86/avx-vbroadcast.ll @@ -47,7 +47,7 @@ entry: ;;;; 128-bit versions ; CHECK: vbroadcastss (% -define <4 x float> @E(float* %ptr) nounwind uwtable readnone ssp { +define <4 x float> @e(float* %ptr) nounwind uwtable readnone ssp { entry: %q = load float* %ptr, align 4 %vecinit.i = insertelement <4 x float> undef, float %q, i32 0 @@ -57,6 +57,19 @@ entry: ret <4 x float> %vecinit6.i } + +; CHECK: _e2 +; CHECK-NOT: vbroadcastss +; CHECK: ret +define <4 x float> @_e2(float* %ptr) nounwind uwtable readnone ssp { + %vecinit.i = insertelement <4 x float> undef, float 0xbf80000000000000, i32 0 + %vecinit2.i = insertelement <4 x float> %vecinit.i, float 0xbf80000000000000, i32 1 + %vecinit4.i = insertelement <4 x float> %vecinit2.i, float 0xbf80000000000000, i32 2 + %vecinit6.i = insertelement <4 x float> %vecinit4.i, float 0xbf80000000000000, i32 3 + ret <4 x float> %vecinit6.i +} + + ; CHECK: vbroadcastss (% define <4 x i32> @F(i32* %ptr) nounwind uwtable readnone ssp { entry: @@ -71,7 +84,7 @@ entry: ; Unsupported vbroadcasts ; CHECK: _G -; CHECK-NOT: vbroadcastsd (% +; CHECK-NOT: broadcast (% ; CHECK: ret define <2 x i64> @G(i64* %ptr) nounwind uwtable readnone ssp { entry: @@ -82,7 +95,7 @@ entry: } ; CHECK: _H -; CHECK-NOT: vbroadcastss +; CHECK-NOT: broadcast ; CHECK: ret define <4 x i32> @H(<4 x i32> %a) { %x = shufflevector <4 x i32> %a, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef> @@ -90,7 +103,7 @@ define <4 x i32> @H(<4 x i32> %a) { } ; CHECK: _I -; CHECK-NOT: vbroadcastsd (% +; CHECK-NOT: broadcast (% ; CHECK: ret define <2 x double> @I(double* %ptr) nounwind uwtable readnone ssp { entry: diff --git a/test/CodeGen/X86/avx-vpermil.ll b/test/CodeGen/X86/avx-vpermil.ll index 9707cd9..cb904b9 100644 --- a/test/CodeGen/X86/avx-vpermil.ll +++ b/test/CodeGen/X86/avx-vpermil.ll @@ -45,7 +45,8 @@ entry: ret <8 x float> %shuffle } -; CHECK: vpermilps +; CHECK: palignr +; CHECK: palignr define <8 x float> @funcF(<8 x float> %a) nounwind uwtable readnone ssp { entry: %shuffle = shufflevector <8 x float> %a, <8 x float> zeroinitializer, <8 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9> diff --git a/test/CodeGen/X86/avx2-intrinsics-x86.ll b/test/CodeGen/X86/avx2-intrinsics-x86.ll index 1fb41c0..3f27a02 100644 --- a/test/CodeGen/X86/avx2-intrinsics-x86.ll +++ b/test/CodeGen/X86/avx2-intrinsics-x86.ll @@ -800,22 +800,6 @@ define <8 x float> @test_x86_avx2_permps(<8 x float> %a0, <8 x float> %a1) { declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x float>) nounwind readonly -define <4 x i64> @test_x86_avx2_permq(<4 x i64> %a0) { - ; CHECK: vpermq - %res = call <4 x i64> @llvm.x86.avx2.permq(<4 x i64> %a0, i8 7) ; <<4 x i64>> [#uses=1] - ret <4 x i64> %res -} -declare <4 x i64> @llvm.x86.avx2.permq(<4 x i64>, i8) nounwind readonly - - -define <4 x double> @test_x86_avx2_permpd(<4 x double> %a0) { - ; CHECK: vpermpd - %res = call <4 x double> @llvm.x86.avx2.permpd(<4 x double> %a0, i8 7) ; <<4 x double>> [#uses=1] - ret <4 x double> %res -} -declare <4 x double> @llvm.x86.avx2.permpd(<4 x double>, i8) nounwind readonly - - define <4 x i64> @test_x86_avx2_vperm2i128(<4 x i64> %a0, <4 x i64> %a1) { ; CHECK: vperm2i128 %res = call <4 x i64> @llvm.x86.avx2.vperm2i128(<4 x i64> %a0, <4 x i64> %a1, i8 1) ; <<4 x i64>> [#uses=1] diff --git a/test/CodeGen/X86/avx2-vbroadcast.ll b/test/CodeGen/X86/avx2-vbroadcast.ll index fbabb15..1a78414 100644 --- a/test/CodeGen/X86/avx2-vbroadcast.ll +++ b/test/CodeGen/X86/avx2-vbroadcast.ll @@ -150,3 +150,38 @@ entry: %vecinit2.i = insertelement <2 x double> %vecinit.i, double %q, i32 1 ret <2 x double> %vecinit2.i } + +; CHECK: V111 +; CHECK: vpbroadcastd +; CHECK: ret +define <8 x i32> @V111(<8 x i32> %in) nounwind uwtable readnone ssp { +entry: + %g = add <8 x i32> %in, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> + ret <8 x i32> %g +} + +; CHECK: _e2 +; CHECK: vbroadcastss +; CHECK: ret +define <4 x float> @_e2(float* %ptr) nounwind uwtable readnone ssp { + %vecinit.i = insertelement <4 x float> undef, float 0xbf80000000000000, i32 0 + %vecinit2.i = insertelement <4 x float> %vecinit.i, float 0xbf80000000000000, i32 1 + %vecinit4.i = insertelement <4 x float> %vecinit2.i, float 0xbf80000000000000, i32 2 + %vecinit6.i = insertelement <4 x float> %vecinit4.i, float 0xbf80000000000000, i32 3 + ret <4 x float> %vecinit6.i +} + +; CHECK: _e4 +; CHECK-NOT: broadcast +; CHECK: ret +define <8 x i8> @_e4(i8* %ptr) nounwind uwtable readnone ssp { + %vecinit0.i = insertelement <8 x i8> undef, i8 52, i32 0 + %vecinit1.i = insertelement <8 x i8> %vecinit0.i, i8 52, i32 1 + %vecinit2.i = insertelement <8 x i8> %vecinit1.i, i8 52, i32 2 + %vecinit3.i = insertelement <8 x i8> %vecinit2.i, i8 52, i32 3 + %vecinit4.i = insertelement <8 x i8> %vecinit3.i, i8 52, i32 3 + %vecinit5.i = insertelement <8 x i8> %vecinit4.i, i8 52, i32 3 + %vecinit6.i = insertelement <8 x i8> %vecinit5.i, i8 52, i32 3 + %vecinit7.i = insertelement <8 x i8> %vecinit6.i, i8 52, i32 3 + ret <8 x i8> %vecinit7.i +} diff --git a/test/CodeGen/X86/avx2-vperm.ll b/test/CodeGen/X86/avx2-vperm.ll new file mode 100755 index 0000000..d576d0e --- /dev/null +++ b/test/CodeGen/X86/avx2-vperm.ll @@ -0,0 +1,34 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s + +define <8 x i32> @perm_cl_int_8x32(<8 x i32> %A) nounwind readnone { +entry: +; CHECK: perm_cl_int_8x32 +; CHECK: vpermd + %B = shufflevector <8 x i32> %A, <8 x i32> undef, <8 x i32> <i32 0, i32 7, i32 2, i32 1, i32 2, i32 7, i32 6, i32 0> + ret <8 x i32> %B +} + + +define <8 x float> @perm_cl_fp_8x32(<8 x float> %A) nounwind readnone { +entry: +; CHECK: perm_cl_fp_8x32 +; CHECK: vpermps + %B = shufflevector <8 x float> %A, <8 x float> undef, <8 x i32> <i32 undef, i32 7, i32 2, i32 undef, i32 4, i32 undef, i32 1, i32 6> + ret <8 x float> %B +} + +define <4 x i64> @perm_cl_int_4x64(<4 x i64> %A) nounwind readnone { +entry: +; CHECK: perm_cl_int_4x64 +; CHECK: vpermq + %B = shufflevector <4 x i64> %A, <4 x i64> undef, <4 x i32> <i32 0, i32 3, i32 2, i32 1> + ret <4 x i64> %B +} + +define <4 x double> @perm_cl_fp_4x64(<4 x double> %A) nounwind readnone { +entry: +; CHECK: perm_cl_fp_4x64 +; CHECK: vpermpd + %B = shufflevector <4 x double> %A, <4 x double> undef, <4 x i32> <i32 0, i32 3, i32 2, i32 1> + ret <4 x double> %B +} diff --git a/test/CodeGen/X86/block-placement.ll b/test/CodeGen/X86/block-placement.ll index 167d522..fc7b638 100644 --- a/test/CodeGen/X86/block-placement.ll +++ b/test/CodeGen/X86/block-placement.ll @@ -1,4 +1,4 @@ -; RUN: llc -mtriple=i686-linux -enable-block-placement < %s | FileCheck %s +; RUN: llc -mtriple=i686-linux < %s | FileCheck %s declare void @error(i32 %i, i32 %a, i32 %b) @@ -76,11 +76,11 @@ define i32 @test_loop_cold_blocks(i32 %i, i32* %a) { ; Check that we sink cold loop blocks after the hot loop body. ; CHECK: test_loop_cold_blocks: ; CHECK: %entry +; CHECK: %unlikely1 +; CHECK: %unlikely2 ; CHECK: %body1 ; CHECK: %body2 ; CHECK: %body3 -; CHECK: %unlikely1 -; CHECK: %unlikely2 ; CHECK: %exit entry: @@ -122,14 +122,14 @@ define i32 @test_loop_early_exits(i32 %i, i32* %a) { ; Check that we sink early exit blocks out of loop bodies. ; CHECK: test_loop_early_exits: ; CHECK: %entry +; CHECK: %body1 ; CHECK: %body2 ; CHECK: %body3 ; CHECK: %body4 -; CHECK: %body1 +; CHECK: %exit ; CHECK: %bail1 ; CHECK: %bail2 ; CHECK: %bail3 -; CHECK: %exit entry: br label %body1 @@ -199,6 +199,36 @@ exit: ret i32 %base } +define i32 @test_no_loop_rotate(i32 %i, i32* %a) { +; Check that we don't try to rotate a loop which is already laid out with +; fallthrough opportunities into the top and out of the bottom. +; CHECK: test_no_loop_rotate: +; CHECK: %entry +; CHECK: %body0 +; CHECK: %body1 +; CHECK: %exit + +entry: + br label %body0 + +body0: + %iv = phi i32 [ 0, %entry ], [ %next, %body1 ] + %base = phi i32 [ 0, %entry ], [ %sum, %body1 ] + %arrayidx = getelementptr inbounds i32* %a, i32 %iv + %0 = load i32* %arrayidx + %sum = add nsw i32 %0, %base + %bailcond1 = icmp eq i32 %sum, 42 + br i1 %bailcond1, label %exit, label %body1 + +body1: + %next = add i32 %iv, 1 + %exitcond = icmp eq i32 %next, %i + br i1 %exitcond, label %exit, label %body0 + +exit: + ret i32 %base +} + define void @test_loop_rotate_reversed_blocks() { ; This test case (greatly reduced from an Olden bencmark) ensures that the loop ; rotate implementation doesn't assume that loops are laid out in a particular @@ -348,7 +378,6 @@ define void @unnatural_cfg2() { ; CHECK: %entry ; CHECK: %loop.body1 ; CHECK: %loop.body2 -; CHECK: %loop.header ; CHECK: %loop.body3 ; CHECK: %loop.inner1.begin ; The end block is folded with %loop.body3... @@ -356,6 +385,7 @@ define void @unnatural_cfg2() { ; CHECK: %loop.body4 ; CHECK: %loop.inner2.begin ; The loop.inner2.end block is folded +; CHECK: %loop.header ; CHECK: %bail entry: @@ -928,3 +958,126 @@ entry: exit: ret void } + +define void @benchmark_heapsort(i32 %n, double* nocapture %ra) { +; This test case comes from the heapsort benchmark, and exemplifies several +; important aspects to block placement in the presence of loops: +; 1) Loop rotation needs to *ensure* that the desired exiting edge can be +; a fallthrough. +; 2) The exiting edge from the loop which is rotated to be laid out at the +; bottom of the loop needs to be exiting into the nearest enclosing loop (to +; which there is an exit). Otherwise, we force that enclosing loop into +; strange layouts that are siginificantly less efficient, often times maing +; it discontiguous. +; +; CHECK: @benchmark_heapsort +; CHECK: %entry +; First rotated loop top. +; CHECK: .align +; CHECK: %while.end +; CHECK: %for.cond +; CHECK: %if.then +; CHECK: %if.else +; CHECK: %if.end10 +; Second rotated loop top +; CHECK: .align +; CHECK: %if.then24 +; CHECK: %while.cond.outer +; Third rotated loop top +; CHECK: .align +; CHECK: %while.cond +; CHECK: %while.body +; CHECK: %land.lhs.true +; CHECK: %if.then19 +; CHECK: %if.then19 +; CHECK: %if.then8 +; CHECK: ret + +entry: + %shr = ashr i32 %n, 1 + %add = add nsw i32 %shr, 1 + %arrayidx3 = getelementptr inbounds double* %ra, i64 1 + br label %for.cond + +for.cond: + %ir.0 = phi i32 [ %n, %entry ], [ %ir.1, %while.end ] + %l.0 = phi i32 [ %add, %entry ], [ %l.1, %while.end ] + %cmp = icmp sgt i32 %l.0, 1 + br i1 %cmp, label %if.then, label %if.else + +if.then: + %dec = add nsw i32 %l.0, -1 + %idxprom = sext i32 %dec to i64 + %arrayidx = getelementptr inbounds double* %ra, i64 %idxprom + %0 = load double* %arrayidx, align 8 + br label %if.end10 + +if.else: + %idxprom1 = sext i32 %ir.0 to i64 + %arrayidx2 = getelementptr inbounds double* %ra, i64 %idxprom1 + %1 = load double* %arrayidx2, align 8 + %2 = load double* %arrayidx3, align 8 + store double %2, double* %arrayidx2, align 8 + %dec6 = add nsw i32 %ir.0, -1 + %cmp7 = icmp eq i32 %dec6, 1 + br i1 %cmp7, label %if.then8, label %if.end10 + +if.then8: + store double %1, double* %arrayidx3, align 8 + ret void + +if.end10: + %ir.1 = phi i32 [ %ir.0, %if.then ], [ %dec6, %if.else ] + %l.1 = phi i32 [ %dec, %if.then ], [ %l.0, %if.else ] + %rra.0 = phi double [ %0, %if.then ], [ %1, %if.else ] + %add31 = add nsw i32 %ir.1, 1 + br label %while.cond.outer + +while.cond.outer: + %j.0.ph.in = phi i32 [ %l.1, %if.end10 ], [ %j.1, %if.then24 ] + %j.0.ph = shl i32 %j.0.ph.in, 1 + br label %while.cond + +while.cond: + %j.0 = phi i32 [ %add31, %if.end20 ], [ %j.0.ph, %while.cond.outer ] + %cmp11 = icmp sgt i32 %j.0, %ir.1 + br i1 %cmp11, label %while.end, label %while.body + +while.body: + %cmp12 = icmp slt i32 %j.0, %ir.1 + br i1 %cmp12, label %land.lhs.true, label %if.end20 + +land.lhs.true: + %idxprom13 = sext i32 %j.0 to i64 + %arrayidx14 = getelementptr inbounds double* %ra, i64 %idxprom13 + %3 = load double* %arrayidx14, align 8 + %add15 = add nsw i32 %j.0, 1 + %idxprom16 = sext i32 %add15 to i64 + %arrayidx17 = getelementptr inbounds double* %ra, i64 %idxprom16 + %4 = load double* %arrayidx17, align 8 + %cmp18 = fcmp olt double %3, %4 + br i1 %cmp18, label %if.then19, label %if.end20 + +if.then19: + br label %if.end20 + +if.end20: + %j.1 = phi i32 [ %add15, %if.then19 ], [ %j.0, %land.lhs.true ], [ %j.0, %while.body ] + %idxprom21 = sext i32 %j.1 to i64 + %arrayidx22 = getelementptr inbounds double* %ra, i64 %idxprom21 + %5 = load double* %arrayidx22, align 8 + %cmp23 = fcmp olt double %rra.0, %5 + br i1 %cmp23, label %if.then24, label %while.cond + +if.then24: + %idxprom27 = sext i32 %j.0.ph.in to i64 + %arrayidx28 = getelementptr inbounds double* %ra, i64 %idxprom27 + store double %5, double* %arrayidx28, align 8 + br label %while.cond.outer + +while.end: + %idxprom33 = sext i32 %j.0.ph.in to i64 + %arrayidx34 = getelementptr inbounds double* %ra, i64 %idxprom33 + store double %rra.0, double* %arrayidx34, align 8 + br label %for.cond +} diff --git a/test/CodeGen/X86/br-fold.ll b/test/CodeGen/X86/br-fold.ll index 8af3bd1..2c37194 100644 --- a/test/CodeGen/X86/br-fold.ll +++ b/test/CodeGen/X86/br-fold.ll @@ -1,7 +1,7 @@ ; RUN: llc -march=x86-64 < %s | FileCheck %s ; CHECK: orq -; CHECK-NEXT: jne +; CHECK-NEXT: LBB0_1 @_ZN11xercesc_2_513SchemaSymbols21fgURI_SCHEMAFORSCHEMAE = external constant [33 x i16], align 32 ; <[33 x i16]*> [#uses=1] @_ZN11xercesc_2_56XMLUni16fgNotationStringE = external constant [9 x i16], align 16 ; <[9 x i16]*> [#uses=1] diff --git a/test/CodeGen/X86/call-push.ll b/test/CodeGen/X86/call-push.ll index 8cca10c..e69f8c1 100644 --- a/test/CodeGen/X86/call-push.ll +++ b/test/CodeGen/X86/call-push.ll @@ -7,8 +7,8 @@ define i32 @decode_byte(%struct.decode_t* %decode) nounwind { ; CHECK: decode_byte: ; CHECK: pushl ; CHECK: popl -; CHECK: popl ; CHECK: jmp +; CHECK: popl entry: %tmp2 = getelementptr %struct.decode_t* %decode, i32 0, i32 4 ; <i16*> [#uses=1] %tmp23 = bitcast i16* %tmp2 to i32* ; <i32*> [#uses=1] diff --git a/test/CodeGen/X86/change-compare-stride-0.ll b/test/CodeGen/X86/change-compare-stride-0.ll deleted file mode 100644 index 439f7b0..0000000 --- a/test/CodeGen/X86/change-compare-stride-0.ll +++ /dev/null @@ -1,83 +0,0 @@ -; RUN: llc < %s -march=x86 -enable-lsr-nested | FileCheck %s -; -; Nested LSR is required to optimize this case. -; We do not expect to see this form of IR without -enable-iv-rewrite. - -define void @borf(i8* nocapture %in, i8* nocapture %out) nounwind { -; CHECK: borf: -; CHECK-NOT: inc -; CHECK-NOT: leal 1( -; CHECK-NOT: leal -1( -; CHECK: decl -; CHECK-NEXT: cmpl $-478 -; CHECK: ret - -bb4.thread: - br label %bb2.outer - -bb2.outer: ; preds = %bb4, %bb4.thread - %indvar18 = phi i32 [ 0, %bb4.thread ], [ %indvar.next28, %bb4 ] ; <i32> [#uses=3] - %tmp34 = mul i32 %indvar18, 65535 ; <i32> [#uses=1] - %i.0.reg2mem.0.ph = add i32 %tmp34, 639 ; <i32> [#uses=1] - %0 = and i32 %i.0.reg2mem.0.ph, 65535 ; <i32> [#uses=1] - %1 = mul i32 %0, 480 ; <i32> [#uses=1] - %tmp20 = mul i32 %indvar18, -478 ; <i32> [#uses=1] - br label %bb2 - -bb2: ; preds = %bb2, %bb2.outer - %indvar = phi i32 [ 0, %bb2.outer ], [ %indvar.next, %bb2 ] ; <i32> [#uses=3] - %ctg2 = getelementptr i8* %out, i32 %tmp20 ; <i8*> [#uses=1] - %tmp21 = ptrtoint i8* %ctg2 to i32 ; <i32> [#uses=1] - %tmp23 = sub i32 %tmp21, %indvar ; <i32> [#uses=1] - %out_addr.0.reg2mem.0 = inttoptr i32 %tmp23 to i8* ; <i8*> [#uses=1] - %tmp25 = mul i32 %indvar, 65535 ; <i32> [#uses=1] - %j.0.reg2mem.0 = add i32 %tmp25, 479 ; <i32> [#uses=1] - %2 = and i32 %j.0.reg2mem.0, 65535 ; <i32> [#uses=1] - %3 = add i32 %1, %2 ; <i32> [#uses=9] - %4 = add i32 %3, -481 ; <i32> [#uses=1] - %5 = getelementptr i8* %in, i32 %4 ; <i8*> [#uses=1] - %6 = load i8* %5, align 1 ; <i8> [#uses=1] - %7 = add i32 %3, -480 ; <i32> [#uses=1] - %8 = getelementptr i8* %in, i32 %7 ; <i8*> [#uses=1] - %9 = load i8* %8, align 1 ; <i8> [#uses=1] - %10 = add i32 %3, -479 ; <i32> [#uses=1] - %11 = getelementptr i8* %in, i32 %10 ; <i8*> [#uses=1] - %12 = load i8* %11, align 1 ; <i8> [#uses=1] - %13 = add i32 %3, -1 ; <i32> [#uses=1] - %14 = getelementptr i8* %in, i32 %13 ; <i8*> [#uses=1] - %15 = load i8* %14, align 1 ; <i8> [#uses=1] - %16 = getelementptr i8* %in, i32 %3 ; <i8*> [#uses=1] - %17 = load i8* %16, align 1 ; <i8> [#uses=1] - %18 = add i32 %3, 1 ; <i32> [#uses=1] - %19 = getelementptr i8* %in, i32 %18 ; <i8*> [#uses=1] - %20 = load i8* %19, align 1 ; <i8> [#uses=1] - %21 = add i32 %3, 481 ; <i32> [#uses=1] - %22 = getelementptr i8* %in, i32 %21 ; <i8*> [#uses=1] - %23 = load i8* %22, align 1 ; <i8> [#uses=1] - %24 = add i32 %3, 480 ; <i32> [#uses=1] - %25 = getelementptr i8* %in, i32 %24 ; <i8*> [#uses=1] - %26 = load i8* %25, align 1 ; <i8> [#uses=1] - %27 = add i32 %3, 479 ; <i32> [#uses=1] - %28 = getelementptr i8* %in, i32 %27 ; <i8*> [#uses=1] - %29 = load i8* %28, align 1 ; <i8> [#uses=1] - %30 = add i8 %9, %6 ; <i8> [#uses=1] - %31 = add i8 %30, %12 ; <i8> [#uses=1] - %32 = add i8 %31, %15 ; <i8> [#uses=1] - %33 = add i8 %32, %17 ; <i8> [#uses=1] - %34 = add i8 %33, %20 ; <i8> [#uses=1] - %35 = add i8 %34, %23 ; <i8> [#uses=1] - %36 = add i8 %35, %26 ; <i8> [#uses=1] - %37 = add i8 %36, %29 ; <i8> [#uses=1] - store i8 %37, i8* %out_addr.0.reg2mem.0, align 1 - %indvar.next = add i32 %indvar, 1 ; <i32> [#uses=2] - %exitcond = icmp eq i32 %indvar.next, 478 ; <i1> [#uses=1] - br i1 %exitcond, label %bb4, label %bb2 - -bb4: ; preds = %bb2 - %indvar.next28 = add i32 %indvar18, 1 ; <i32> [#uses=2] - %exitcond29 = icmp eq i32 %indvar.next28, 638 ; <i1> [#uses=1] - br i1 %exitcond29, label %return, label %bb2.outer - -return: ; preds = %bb4 - ret void -} diff --git a/test/CodeGen/X86/dbg-declare.ll b/test/CodeGen/X86/dbg-declare.ll new file mode 100644 index 0000000..5d4cedc --- /dev/null +++ b/test/CodeGen/X86/dbg-declare.ll @@ -0,0 +1,57 @@ +; RUN: llc < %s -O0 -mtriple x86_64-apple-darwin +; <rdar://problem/11134152> + +define i32 @foo(i32* %x) nounwind uwtable ssp { +entry: + %x.addr = alloca i32*, align 8 + %saved_stack = alloca i8* + %cleanup.dest.slot = alloca i32 + store i32* %x, i32** %x.addr, align 8 + call void @llvm.dbg.declare(metadata !{i32** %x.addr}, metadata !14), !dbg !15 + %0 = load i32** %x.addr, align 8, !dbg !16 + %1 = load i32* %0, align 4, !dbg !16 + %2 = zext i32 %1 to i64, !dbg !16 + %3 = call i8* @llvm.stacksave(), !dbg !16 + store i8* %3, i8** %saved_stack, !dbg !16 + %vla = alloca i8, i64 %2, align 16, !dbg !16 + call void @llvm.dbg.declare(metadata !{i8* %vla}, metadata !18), !dbg !23 + store i32 1, i32* %cleanup.dest.slot + %4 = load i8** %saved_stack, !dbg !24 + call void @llvm.stackrestore(i8* %4), !dbg !24 + ret i32 0, !dbg !25 +} + +declare void @llvm.dbg.declare(metadata, metadata) nounwind readnone + +declare i8* @llvm.stacksave() nounwind + +declare void @llvm.stackrestore(i8*) nounwind + +!llvm.dbg.cu = !{!0} + +!0 = metadata !{i32 786449, i32 0, i32 12, metadata !"20020104-2.c", metadata !"/Volumes/Sandbox/llvm", metadata !"clang version 3.1 (trunk 153698)", i1 true, i1 false, metadata !"", i32 0, metadata !1, metadata !1, metadata !3, metadata !1} ; [ DW_TAG_compile_unit ] +!1 = metadata !{metadata !2} +!2 = metadata !{i32 0} +!3 = metadata !{metadata !4} +!4 = metadata !{metadata !5} +!5 = metadata !{i32 786478, i32 0, metadata !6, metadata !"foo", metadata !"foo", metadata !"", metadata !6, i32 6, metadata !7, i1 false, i1 true, i32 0, i32 0, null, i32 256, i1 false, i32 (i32*)* @foo, null, null, metadata !12} ; [ DW_TAG_subprogram ] +!6 = metadata !{i32 786473, metadata !"20020104-2.c", metadata !"/Volumes/Sandbox/llvm", null} ; [ DW_TAG_file_type ] +!7 = metadata !{i32 786453, i32 0, metadata !"", i32 0, i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !8, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!8 = metadata !{metadata !9, metadata !10} +!9 = metadata !{i32 786468, null, metadata !"int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] +!10 = metadata !{i32 786447, null, metadata !"", null, i32 0, i64 64, i64 64, i64 0, i32 0, metadata !11} ; [ DW_TAG_pointer_type ] +!11 = metadata !{i32 786470, null, metadata !"", null, i32 0, i64 0, i64 0, i64 0, i32 0, metadata !9} ; [ DW_TAG_const_type ] +!12 = metadata !{metadata !13} +!13 = metadata !{i32 786468} ; [ DW_TAG_base_type ] +!14 = metadata !{i32 786689, metadata !5, metadata !"x", metadata !6, i32 16777221, metadata !10, i32 0, i32 0} ; [ DW_TAG_arg_variable ] +!15 = metadata !{i32 5, i32 21, metadata !5, null} +!16 = metadata !{i32 7, i32 13, metadata !17, null} +!17 = metadata !{i32 786443, metadata !5, i32 6, i32 1, metadata !6, i32 0} ; [ DW_TAG_lexical_block ] +!18 = metadata !{i32 786688, metadata !17, metadata !"a", metadata !6, i32 7, metadata !19, i32 0, i32 0} ; [ DW_TAG_auto_variable ] +!19 = metadata !{i32 786433, null, metadata !"", null, i32 0, i64 0, i64 8, i32 0, i32 0, metadata !20, metadata !21, i32 0, i32 0} ; [ DW_TAG_array_type ] +!20 = metadata !{i32 786468, null, metadata !"char", null, i32 0, i64 8, i64 8, i64 0, i32 0, i32 6} ; [ DW_TAG_base_type ] +!21 = metadata !{metadata !22} +!22 = metadata !{i32 786465, i64 1, i64 0} ; [ DW_TAG_subrange_type ] +!23 = metadata !{i32 7, i32 8, metadata !17, null} +!24 = metadata !{i32 9, i32 1, metadata !17, null} +!25 = metadata !{i32 8, i32 3, metadata !17, null} diff --git a/test/CodeGen/X86/dec-eflags-lower.ll b/test/CodeGen/X86/dec-eflags-lower.ll deleted file mode 100644 index 190819f..0000000 --- a/test/CodeGen/X86/dec-eflags-lower.ll +++ /dev/null @@ -1,67 +0,0 @@ -; RUN: llc < %s -march=x86-64 | FileCheck %s - -%struct.obj = type { i64 } - -; CHECK: _Z7releaseP3obj -define void @_Z7releaseP3obj(%struct.obj* nocapture %o) nounwind uwtable ssp { -entry: -; CHECK: decq (%{{rdi|rcx}}) -; CHECK-NEXT: je - %refcnt = getelementptr inbounds %struct.obj* %o, i64 0, i32 0 - %0 = load i64* %refcnt, align 8, !tbaa !0 - %dec = add i64 %0, -1 - store i64 %dec, i64* %refcnt, align 8, !tbaa !0 - %tobool = icmp eq i64 %dec, 0 - br i1 %tobool, label %if.end, label %return - -if.end: ; preds = %entry - %1 = bitcast %struct.obj* %o to i8* - tail call void @free(i8* %1) - br label %return - -return: ; preds = %entry, %if.end - ret void -} - -@c = common global i64 0, align 8 -@a = common global i32 0, align 4 -@.str = private unnamed_addr constant [5 x i8] c"%ld\0A\00", align 1 -@b = common global i32 0, align 4 - -; CHECK: test -define i32 @test() nounwind uwtable ssp { -entry: -; CHECK: decq -; CHECK-NOT: decq -%0 = load i64* @c, align 8, !tbaa !0 -%dec.i = add nsw i64 %0, -1 -store i64 %dec.i, i64* @c, align 8, !tbaa !0 -%tobool.i = icmp ne i64 %dec.i, 0 -%lor.ext.i = zext i1 %tobool.i to i32 -store i32 %lor.ext.i, i32* @a, align 4, !tbaa !3 -%call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %dec.i) nounwind -ret i32 0 -} - -; CHECK: test2 -define i32 @test2() nounwind uwtable ssp { -entry: -; CHECK-NOT: decq ({{.*}}) -%0 = load i64* @c, align 8, !tbaa !0 -%dec.i = add nsw i64 %0, -1 -store i64 %dec.i, i64* @c, align 8, !tbaa !0 -%tobool.i = icmp ne i64 %0, 0 -%lor.ext.i = zext i1 %tobool.i to i32 -store i32 %lor.ext.i, i32* @a, align 4, !tbaa !3 -%call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %dec.i) nounwind -ret i32 0 -} - -declare i32 @printf(i8* nocapture, ...) nounwind - -declare void @free(i8* nocapture) nounwind - -!0 = metadata !{metadata !"long", metadata !1} -!1 = metadata !{metadata !"omnipotent char", metadata !2} -!2 = metadata !{metadata !"Simple C/C++ TBAA", null} -!3 = metadata !{metadata !"int", metadata !1} diff --git a/test/CodeGen/X86/fdiv.ll b/test/CodeGen/X86/fdiv.ll new file mode 100644 index 0000000..0749682 --- /dev/null +++ b/test/CodeGen/X86/fdiv.ll @@ -0,0 +1,41 @@ +; RUN: llc < %s -march=x86-64 -enable-unsafe-fp-math | FileCheck %s + +define double @exact(double %x) { +; Exact division by a constant converted to multiplication. +; CHECK: @exact +; CHECK: mulsd + %div = fdiv double %x, 2.0 + ret double %div +} + +define double @inexact(double %x) { +; Inexact division by a constant converted to multiplication. +; CHECK: @inexact +; CHECK: mulsd + %div = fdiv double %x, 0x41DFFFFFFFC00000 + ret double %div +} + +define double @funky(double %x) { +; No conversion to multiplication if too funky. +; CHECK: @funky +; CHECK: divsd + %div = fdiv double %x, 0.0 + ret double %div +} + +define double @denormal1(double %x) { +; Don't generate multiplication by a denormal. +; CHECK: @denormal1 +; CHECK: divsd + %div = fdiv double %x, 0x7FD0000000000001 + ret double %div +} + +define double @denormal2(double %x) { +; Don't generate multiplication by a denormal. +; CHECK: @denormal +; CHECK: divsd + %div = fdiv double %x, 0x7FEFFFFFFFFFFFFF + ret double %div +} diff --git a/test/CodeGen/X86/iv-users-in-other-loops.ll b/test/CodeGen/X86/iv-users-in-other-loops.ll deleted file mode 100644 index e51e61d..0000000 --- a/test/CodeGen/X86/iv-users-in-other-loops.ll +++ /dev/null @@ -1,299 +0,0 @@ -; RUN: llc < %s -mcpu=generic -march=x86-64 -enable-lsr-nested -o %t -; RUN: not grep inc %t -; RUN: grep dec %t | count 2 -; RUN: grep addq %t | count 12 -; RUN: not grep addb %t -; RUN: not grep leal %t -; RUN: not grep movq %t - -; IV users in each of the loops from other loops shouldn't cause LSR -; to insert new induction variables. Previously it would create a -; flood of new induction variables. -; Also, the loop reversal should kick in once. -; -; In this example, performing LSR on the entire loop nest, -; as opposed to only the inner loop can further reduce induction variables, -; and their related instructions and registers. - -target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128" -target triple = "x86_64-unknown-linux-gnu" - -define void @foo(float* %A, i32 %IA, float* %B, i32 %IB, float* nocapture %C, i32 %N) nounwind { -entry: - %0 = xor i32 %IA, 1 ; <i32> [#uses=1] - %1 = xor i32 %IB, 1 ; <i32> [#uses=1] - %2 = or i32 %1, %0 ; <i32> [#uses=1] - %3 = icmp eq i32 %2, 0 ; <i1> [#uses=1] - br i1 %3, label %bb2, label %bb13 - -bb: ; preds = %bb3 - %4 = load float* %A_addr.0, align 4 ; <float> [#uses=1] - %5 = load float* %B_addr.0, align 4 ; <float> [#uses=1] - %6 = fmul float %4, %5 ; <float> [#uses=1] - %7 = fadd float %6, %Sum0.0 ; <float> [#uses=1] - %indvar.next154 = add i64 %B_addr.0.rec, 1 ; <i64> [#uses=1] - br label %bb2 - -bb2: ; preds = %entry, %bb - %B_addr.0.rec = phi i64 [ %indvar.next154, %bb ], [ 0, %entry ] ; <i64> [#uses=14] - %Sum0.0 = phi float [ %7, %bb ], [ 0.000000e+00, %entry ] ; <float> [#uses=5] - %indvar146 = trunc i64 %B_addr.0.rec to i32 ; <i32> [#uses=1] - %N_addr.0 = sub i32 %N, %indvar146 ; <i32> [#uses=6] - %A_addr.0 = getelementptr float* %A, i64 %B_addr.0.rec ; <float*> [#uses=4] - %B_addr.0 = getelementptr float* %B, i64 %B_addr.0.rec ; <float*> [#uses=4] - %8 = icmp sgt i32 %N_addr.0, 0 ; <i1> [#uses=1] - br i1 %8, label %bb3, label %bb4 - -bb3: ; preds = %bb2 - %9 = ptrtoint float* %A_addr.0 to i64 ; <i64> [#uses=1] - %10 = and i64 %9, 15 ; <i64> [#uses=1] - %11 = icmp eq i64 %10, 0 ; <i1> [#uses=1] - br i1 %11, label %bb4, label %bb - -bb4: ; preds = %bb3, %bb2 - %12 = ptrtoint float* %B_addr.0 to i64 ; <i64> [#uses=1] - %13 = and i64 %12, 15 ; <i64> [#uses=1] - %14 = icmp eq i64 %13, 0 ; <i1> [#uses=1] - %15 = icmp sgt i32 %N_addr.0, 15 ; <i1> [#uses=2] - br i1 %14, label %bb6.preheader, label %bb10.preheader - -bb10.preheader: ; preds = %bb4 - br i1 %15, label %bb9, label %bb12.loopexit - -bb6.preheader: ; preds = %bb4 - br i1 %15, label %bb5, label %bb8.loopexit - -bb5: ; preds = %bb5, %bb6.preheader - %indvar143 = phi i64 [ 0, %bb6.preheader ], [ %indvar.next144, %bb5 ] ; <i64> [#uses=3] - %vSum0.072 = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %21, %bb5 ] ; <<4 x float>> [#uses=1] - %vSum1.070 = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %29, %bb5 ] ; <<4 x float>> [#uses=1] - %vSum2.069 = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %37, %bb5 ] ; <<4 x float>> [#uses=1] - %vSum3.067 = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %45, %bb5 ] ; <<4 x float>> [#uses=1] - %indvar145 = trunc i64 %indvar143 to i32 ; <i32> [#uses=1] - %tmp150 = mul i32 %indvar145, -16 ; <i32> [#uses=1] - %N_addr.268 = add i32 %tmp150, %N_addr.0 ; <i32> [#uses=1] - %A_addr.273.rec = shl i64 %indvar143, 4 ; <i64> [#uses=5] - %B_addr.0.sum180 = add i64 %B_addr.0.rec, %A_addr.273.rec ; <i64> [#uses=2] - %B_addr.271 = getelementptr float* %B, i64 %B_addr.0.sum180 ; <float*> [#uses=1] - %A_addr.273 = getelementptr float* %A, i64 %B_addr.0.sum180 ; <float*> [#uses=1] - tail call void asm sideeffect ";# foo", "~{dirflag},~{fpsr},~{flags}"() nounwind - %16 = bitcast float* %A_addr.273 to <4 x float>* ; <<4 x float>*> [#uses=1] - %17 = load <4 x float>* %16, align 16 ; <<4 x float>> [#uses=1] - %18 = bitcast float* %B_addr.271 to <4 x float>* ; <<4 x float>*> [#uses=1] - %19 = load <4 x float>* %18, align 16 ; <<4 x float>> [#uses=1] - %20 = fmul <4 x float> %17, %19 ; <<4 x float>> [#uses=1] - %21 = fadd <4 x float> %20, %vSum0.072 ; <<4 x float>> [#uses=2] - %A_addr.273.sum163 = or i64 %A_addr.273.rec, 4 ; <i64> [#uses=1] - %A_addr.0.sum175 = add i64 %B_addr.0.rec, %A_addr.273.sum163 ; <i64> [#uses=2] - %22 = getelementptr float* %A, i64 %A_addr.0.sum175 ; <float*> [#uses=1] - %23 = bitcast float* %22 to <4 x float>* ; <<4 x float>*> [#uses=1] - %24 = load <4 x float>* %23, align 16 ; <<4 x float>> [#uses=1] - %25 = getelementptr float* %B, i64 %A_addr.0.sum175 ; <float*> [#uses=1] - %26 = bitcast float* %25 to <4 x float>* ; <<4 x float>*> [#uses=1] - %27 = load <4 x float>* %26, align 16 ; <<4 x float>> [#uses=1] - %28 = fmul <4 x float> %24, %27 ; <<4 x float>> [#uses=1] - %29 = fadd <4 x float> %28, %vSum1.070 ; <<4 x float>> [#uses=2] - %A_addr.273.sum161 = or i64 %A_addr.273.rec, 8 ; <i64> [#uses=1] - %A_addr.0.sum174 = add i64 %B_addr.0.rec, %A_addr.273.sum161 ; <i64> [#uses=2] - %30 = getelementptr float* %A, i64 %A_addr.0.sum174 ; <float*> [#uses=1] - %31 = bitcast float* %30 to <4 x float>* ; <<4 x float>*> [#uses=1] - %32 = load <4 x float>* %31, align 16 ; <<4 x float>> [#uses=1] - %33 = getelementptr float* %B, i64 %A_addr.0.sum174 ; <float*> [#uses=1] - %34 = bitcast float* %33 to <4 x float>* ; <<4 x float>*> [#uses=1] - %35 = load <4 x float>* %34, align 16 ; <<4 x float>> [#uses=1] - %36 = fmul <4 x float> %32, %35 ; <<4 x float>> [#uses=1] - %37 = fadd <4 x float> %36, %vSum2.069 ; <<4 x float>> [#uses=2] - %A_addr.273.sum159 = or i64 %A_addr.273.rec, 12 ; <i64> [#uses=1] - %A_addr.0.sum173 = add i64 %B_addr.0.rec, %A_addr.273.sum159 ; <i64> [#uses=2] - %38 = getelementptr float* %A, i64 %A_addr.0.sum173 ; <float*> [#uses=1] - %39 = bitcast float* %38 to <4 x float>* ; <<4 x float>*> [#uses=1] - %40 = load <4 x float>* %39, align 16 ; <<4 x float>> [#uses=1] - %41 = getelementptr float* %B, i64 %A_addr.0.sum173 ; <float*> [#uses=1] - %42 = bitcast float* %41 to <4 x float>* ; <<4 x float>*> [#uses=1] - %43 = load <4 x float>* %42, align 16 ; <<4 x float>> [#uses=1] - %44 = fmul <4 x float> %40, %43 ; <<4 x float>> [#uses=1] - %45 = fadd <4 x float> %44, %vSum3.067 ; <<4 x float>> [#uses=2] - %.rec83 = add i64 %A_addr.273.rec, 16 ; <i64> [#uses=1] - %A_addr.0.sum172 = add i64 %B_addr.0.rec, %.rec83 ; <i64> [#uses=2] - %46 = getelementptr float* %A, i64 %A_addr.0.sum172 ; <float*> [#uses=1] - %47 = getelementptr float* %B, i64 %A_addr.0.sum172 ; <float*> [#uses=1] - %48 = add i32 %N_addr.268, -16 ; <i32> [#uses=2] - %49 = icmp sgt i32 %48, 15 ; <i1> [#uses=1] - %indvar.next144 = add i64 %indvar143, 1 ; <i64> [#uses=1] - br i1 %49, label %bb5, label %bb8.loopexit - -bb7: ; preds = %bb7, %bb8.loopexit - %indvar130 = phi i64 [ 0, %bb8.loopexit ], [ %indvar.next131, %bb7 ] ; <i64> [#uses=3] - %vSum0.260 = phi <4 x float> [ %vSum0.0.lcssa, %bb8.loopexit ], [ %55, %bb7 ] ; <<4 x float>> [#uses=1] - %indvar132 = trunc i64 %indvar130 to i32 ; <i32> [#uses=1] - %tmp133 = mul i32 %indvar132, -4 ; <i32> [#uses=1] - %N_addr.358 = add i32 %tmp133, %N_addr.2.lcssa ; <i32> [#uses=1] - %A_addr.361.rec = shl i64 %indvar130, 2 ; <i64> [#uses=3] - %B_addr.359 = getelementptr float* %B_addr.2.lcssa, i64 %A_addr.361.rec ; <float*> [#uses=1] - %A_addr.361 = getelementptr float* %A_addr.2.lcssa, i64 %A_addr.361.rec ; <float*> [#uses=1] - %50 = bitcast float* %A_addr.361 to <4 x float>* ; <<4 x float>*> [#uses=1] - %51 = load <4 x float>* %50, align 16 ; <<4 x float>> [#uses=1] - %52 = bitcast float* %B_addr.359 to <4 x float>* ; <<4 x float>*> [#uses=1] - %53 = load <4 x float>* %52, align 16 ; <<4 x float>> [#uses=1] - %54 = fmul <4 x float> %51, %53 ; <<4 x float>> [#uses=1] - %55 = fadd <4 x float> %54, %vSum0.260 ; <<4 x float>> [#uses=2] - %.rec85 = add i64 %A_addr.361.rec, 4 ; <i64> [#uses=2] - %56 = getelementptr float* %A_addr.2.lcssa, i64 %.rec85 ; <float*> [#uses=1] - %57 = getelementptr float* %B_addr.2.lcssa, i64 %.rec85 ; <float*> [#uses=1] - %58 = add i32 %N_addr.358, -4 ; <i32> [#uses=2] - %59 = icmp sgt i32 %58, 3 ; <i1> [#uses=1] - %indvar.next131 = add i64 %indvar130, 1 ; <i64> [#uses=1] - br i1 %59, label %bb7, label %bb13 - -bb8.loopexit: ; preds = %bb5, %bb6.preheader - %A_addr.2.lcssa = phi float* [ %A_addr.0, %bb6.preheader ], [ %46, %bb5 ] ; <float*> [#uses=3] - %vSum0.0.lcssa = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %21, %bb5 ] ; <<4 x float>> [#uses=2] - %B_addr.2.lcssa = phi float* [ %B_addr.0, %bb6.preheader ], [ %47, %bb5 ] ; <float*> [#uses=3] - %vSum1.0.lcssa = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %29, %bb5 ] ; <<4 x float>> [#uses=2] - %vSum2.0.lcssa = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %37, %bb5 ] ; <<4 x float>> [#uses=2] - %N_addr.2.lcssa = phi i32 [ %N_addr.0, %bb6.preheader ], [ %48, %bb5 ] ; <i32> [#uses=3] - %vSum3.0.lcssa = phi <4 x float> [ zeroinitializer, %bb6.preheader ], [ %45, %bb5 ] ; <<4 x float>> [#uses=2] - %60 = icmp sgt i32 %N_addr.2.lcssa, 3 ; <i1> [#uses=1] - br i1 %60, label %bb7, label %bb13 - -bb9: ; preds = %bb9, %bb10.preheader - %indvar106 = phi i64 [ 0, %bb10.preheader ], [ %indvar.next107, %bb9 ] ; <i64> [#uses=3] - %vSum0.339 = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %75, %bb9 ] ; <<4 x float>> [#uses=1] - %vSum1.237 = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %80, %bb9 ] ; <<4 x float>> [#uses=1] - %vSum2.236 = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %85, %bb9 ] ; <<4 x float>> [#uses=1] - %vSum3.234 = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %90, %bb9 ] ; <<4 x float>> [#uses=1] - %indvar108 = trunc i64 %indvar106 to i32 ; <i32> [#uses=1] - %tmp113 = mul i32 %indvar108, -16 ; <i32> [#uses=1] - %N_addr.435 = add i32 %tmp113, %N_addr.0 ; <i32> [#uses=1] - %A_addr.440.rec = shl i64 %indvar106, 4 ; <i64> [#uses=5] - %B_addr.0.sum = add i64 %B_addr.0.rec, %A_addr.440.rec ; <i64> [#uses=2] - %B_addr.438 = getelementptr float* %B, i64 %B_addr.0.sum ; <float*> [#uses=1] - %A_addr.440 = getelementptr float* %A, i64 %B_addr.0.sum ; <float*> [#uses=1] - %61 = bitcast float* %B_addr.438 to <4 x float>* ; <i8*> [#uses=1] - %62 = load <4 x float>* %61, align 1 - %B_addr.438.sum169 = or i64 %A_addr.440.rec, 4 ; <i64> [#uses=1] - %B_addr.0.sum187 = add i64 %B_addr.0.rec, %B_addr.438.sum169 ; <i64> [#uses=2] - %63 = getelementptr float* %B, i64 %B_addr.0.sum187 ; <float*> [#uses=1] - %64 = bitcast float* %63 to <4 x float>* ; <i8*> [#uses=1] - %65 = load <4 x float>* %64, align 1 - %B_addr.438.sum168 = or i64 %A_addr.440.rec, 8 ; <i64> [#uses=1] - %B_addr.0.sum186 = add i64 %B_addr.0.rec, %B_addr.438.sum168 ; <i64> [#uses=2] - %66 = getelementptr float* %B, i64 %B_addr.0.sum186 ; <float*> [#uses=1] - %67 = bitcast float* %66 to <4 x float>* ; <i8*> [#uses=1] - %68 = load <4 x float>* %67, align 1 - %B_addr.438.sum167 = or i64 %A_addr.440.rec, 12 ; <i64> [#uses=1] - %B_addr.0.sum185 = add i64 %B_addr.0.rec, %B_addr.438.sum167 ; <i64> [#uses=2] - %69 = getelementptr float* %B, i64 %B_addr.0.sum185 ; <float*> [#uses=1] - %70 = bitcast float* %69 to <4 x float>* ; <i8*> [#uses=1] - %71 = load <4 x float>* %70, align 1 - %72 = bitcast float* %A_addr.440 to <4 x float>* ; <<4 x float>*> [#uses=1] - %73 = load <4 x float>* %72, align 16 ; <<4 x float>> [#uses=1] - %74 = fmul <4 x float> %73, %62 ; <<4 x float>> [#uses=1] - %75 = fadd <4 x float> %74, %vSum0.339 ; <<4 x float>> [#uses=2] - %76 = getelementptr float* %A, i64 %B_addr.0.sum187 ; <float*> [#uses=1] - %77 = bitcast float* %76 to <4 x float>* ; <<4 x float>*> [#uses=1] - %78 = load <4 x float>* %77, align 16 ; <<4 x float>> [#uses=1] - %79 = fmul <4 x float> %78, %65 ; <<4 x float>> [#uses=1] - %80 = fadd <4 x float> %79, %vSum1.237 ; <<4 x float>> [#uses=2] - %81 = getelementptr float* %A, i64 %B_addr.0.sum186 ; <float*> [#uses=1] - %82 = bitcast float* %81 to <4 x float>* ; <<4 x float>*> [#uses=1] - %83 = load <4 x float>* %82, align 16 ; <<4 x float>> [#uses=1] - %84 = fmul <4 x float> %83, %68 ; <<4 x float>> [#uses=1] - %85 = fadd <4 x float> %84, %vSum2.236 ; <<4 x float>> [#uses=2] - %86 = getelementptr float* %A, i64 %B_addr.0.sum185 ; <float*> [#uses=1] - %87 = bitcast float* %86 to <4 x float>* ; <<4 x float>*> [#uses=1] - %88 = load <4 x float>* %87, align 16 ; <<4 x float>> [#uses=1] - %89 = fmul <4 x float> %88, %71 ; <<4 x float>> [#uses=1] - %90 = fadd <4 x float> %89, %vSum3.234 ; <<4 x float>> [#uses=2] - %.rec89 = add i64 %A_addr.440.rec, 16 ; <i64> [#uses=1] - %A_addr.0.sum170 = add i64 %B_addr.0.rec, %.rec89 ; <i64> [#uses=2] - %91 = getelementptr float* %A, i64 %A_addr.0.sum170 ; <float*> [#uses=1] - %92 = getelementptr float* %B, i64 %A_addr.0.sum170 ; <float*> [#uses=1] - %93 = add i32 %N_addr.435, -16 ; <i32> [#uses=2] - %94 = icmp sgt i32 %93, 15 ; <i1> [#uses=1] - %indvar.next107 = add i64 %indvar106, 1 ; <i64> [#uses=1] - br i1 %94, label %bb9, label %bb12.loopexit - -bb11: ; preds = %bb11, %bb12.loopexit - %indvar = phi i64 [ 0, %bb12.loopexit ], [ %indvar.next, %bb11 ] ; <i64> [#uses=3] - %vSum0.428 = phi <4 x float> [ %vSum0.3.lcssa, %bb12.loopexit ], [ %100, %bb11 ] ; <<4 x float>> [#uses=1] - %indvar96 = trunc i64 %indvar to i32 ; <i32> [#uses=1] - %tmp = mul i32 %indvar96, -4 ; <i32> [#uses=1] - %N_addr.526 = add i32 %tmp, %N_addr.4.lcssa ; <i32> [#uses=1] - %A_addr.529.rec = shl i64 %indvar, 2 ; <i64> [#uses=3] - %B_addr.527 = getelementptr float* %B_addr.4.lcssa, i64 %A_addr.529.rec ; <float*> [#uses=1] - %A_addr.529 = getelementptr float* %A_addr.4.lcssa, i64 %A_addr.529.rec ; <float*> [#uses=1] - %95 = bitcast float* %B_addr.527 to <4 x float>* ; <i8*> [#uses=1] - %96 = load <4 x float>* %95, align 1 - %97 = bitcast float* %A_addr.529 to <4 x float>* ; <<4 x float>*> [#uses=1] - %98 = load <4 x float>* %97, align 16 ; <<4 x float>> [#uses=1] - %99 = fmul <4 x float> %98, %96 ; <<4 x float>> [#uses=1] - %100 = fadd <4 x float> %99, %vSum0.428 ; <<4 x float>> [#uses=2] - %.rec91 = add i64 %A_addr.529.rec, 4 ; <i64> [#uses=2] - %101 = getelementptr float* %A_addr.4.lcssa, i64 %.rec91 ; <float*> [#uses=1] - %102 = getelementptr float* %B_addr.4.lcssa, i64 %.rec91 ; <float*> [#uses=1] - %103 = add i32 %N_addr.526, -4 ; <i32> [#uses=2] - %104 = icmp sgt i32 %103, 3 ; <i1> [#uses=1] - %indvar.next = add i64 %indvar, 1 ; <i64> [#uses=1] - br i1 %104, label %bb11, label %bb13 - -bb12.loopexit: ; preds = %bb9, %bb10.preheader - %A_addr.4.lcssa = phi float* [ %A_addr.0, %bb10.preheader ], [ %91, %bb9 ] ; <float*> [#uses=3] - %vSum0.3.lcssa = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %75, %bb9 ] ; <<4 x float>> [#uses=2] - %B_addr.4.lcssa = phi float* [ %B_addr.0, %bb10.preheader ], [ %92, %bb9 ] ; <float*> [#uses=3] - %vSum1.2.lcssa = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %80, %bb9 ] ; <<4 x float>> [#uses=2] - %vSum2.2.lcssa = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %85, %bb9 ] ; <<4 x float>> [#uses=2] - %N_addr.4.lcssa = phi i32 [ %N_addr.0, %bb10.preheader ], [ %93, %bb9 ] ; <i32> [#uses=3] - %vSum3.2.lcssa = phi <4 x float> [ zeroinitializer, %bb10.preheader ], [ %90, %bb9 ] ; <<4 x float>> [#uses=2] - %105 = icmp sgt i32 %N_addr.4.lcssa, 3 ; <i1> [#uses=1] - br i1 %105, label %bb11, label %bb13 - -bb13: ; preds = %bb12.loopexit, %bb11, %bb8.loopexit, %bb7, %entry - %Sum0.1 = phi float [ 0.000000e+00, %entry ], [ %Sum0.0, %bb7 ], [ %Sum0.0, %bb8.loopexit ], [ %Sum0.0, %bb11 ], [ %Sum0.0, %bb12.loopexit ] ; <float> [#uses=1] - %vSum3.1 = phi <4 x float> [ zeroinitializer, %entry ], [ %vSum3.0.lcssa, %bb7 ], [ %vSum3.0.lcssa, %bb8.loopexit ], [ %vSum3.2.lcssa, %bb11 ], [ %vSum3.2.lcssa, %bb12.loopexit ] ; <<4 x float>> [#uses=1] - %N_addr.1 = phi i32 [ %N, %entry ], [ %N_addr.2.lcssa, %bb8.loopexit ], [ %58, %bb7 ], [ %N_addr.4.lcssa, %bb12.loopexit ], [ %103, %bb11 ] ; <i32> [#uses=2] - %vSum2.1 = phi <4 x float> [ zeroinitializer, %entry ], [ %vSum2.0.lcssa, %bb7 ], [ %vSum2.0.lcssa, %bb8.loopexit ], [ %vSum2.2.lcssa, %bb11 ], [ %vSum2.2.lcssa, %bb12.loopexit ] ; <<4 x float>> [#uses=1] - %vSum1.1 = phi <4 x float> [ zeroinitializer, %entry ], [ %vSum1.0.lcssa, %bb7 ], [ %vSum1.0.lcssa, %bb8.loopexit ], [ %vSum1.2.lcssa, %bb11 ], [ %vSum1.2.lcssa, %bb12.loopexit ] ; <<4 x float>> [#uses=1] - %B_addr.1 = phi float* [ %B, %entry ], [ %B_addr.2.lcssa, %bb8.loopexit ], [ %57, %bb7 ], [ %B_addr.4.lcssa, %bb12.loopexit ], [ %102, %bb11 ] ; <float*> [#uses=1] - %vSum0.1 = phi <4 x float> [ zeroinitializer, %entry ], [ %vSum0.0.lcssa, %bb8.loopexit ], [ %55, %bb7 ], [ %vSum0.3.lcssa, %bb12.loopexit ], [ %100, %bb11 ] ; <<4 x float>> [#uses=1] - %A_addr.1 = phi float* [ %A, %entry ], [ %A_addr.2.lcssa, %bb8.loopexit ], [ %56, %bb7 ], [ %A_addr.4.lcssa, %bb12.loopexit ], [ %101, %bb11 ] ; <float*> [#uses=1] - %106 = fadd <4 x float> %vSum0.1, %vSum2.1 ; <<4 x float>> [#uses=1] - %107 = fadd <4 x float> %vSum1.1, %vSum3.1 ; <<4 x float>> [#uses=1] - %108 = fadd <4 x float> %106, %107 ; <<4 x float>> [#uses=4] - %tmp23 = extractelement <4 x float> %108, i32 0 ; <float> [#uses=1] - %tmp21 = extractelement <4 x float> %108, i32 1 ; <float> [#uses=1] - %109 = fadd float %tmp23, %tmp21 ; <float> [#uses=1] - %tmp19 = extractelement <4 x float> %108, i32 2 ; <float> [#uses=1] - %tmp17 = extractelement <4 x float> %108, i32 3 ; <float> [#uses=1] - %110 = fadd float %tmp19, %tmp17 ; <float> [#uses=1] - %111 = fadd float %109, %110 ; <float> [#uses=1] - %Sum0.254 = fadd float %111, %Sum0.1 ; <float> [#uses=2] - %112 = icmp sgt i32 %N_addr.1, 0 ; <i1> [#uses=1] - br i1 %112, label %bb.nph56, label %bb16 - -bb.nph56: ; preds = %bb13 - %tmp. = zext i32 %N_addr.1 to i64 ; <i64> [#uses=1] - br label %bb14 - -bb14: ; preds = %bb14, %bb.nph56 - %indvar117 = phi i64 [ 0, %bb.nph56 ], [ %indvar.next118, %bb14 ] ; <i64> [#uses=3] - %Sum0.255 = phi float [ %Sum0.254, %bb.nph56 ], [ %Sum0.2, %bb14 ] ; <float> [#uses=1] - %tmp.122 = sext i32 %IB to i64 ; <i64> [#uses=1] - %B_addr.652.rec = mul i64 %indvar117, %tmp.122 ; <i64> [#uses=1] - %tmp.124 = sext i32 %IA to i64 ; <i64> [#uses=1] - %A_addr.653.rec = mul i64 %indvar117, %tmp.124 ; <i64> [#uses=1] - %B_addr.652 = getelementptr float* %B_addr.1, i64 %B_addr.652.rec ; <float*> [#uses=1] - %A_addr.653 = getelementptr float* %A_addr.1, i64 %A_addr.653.rec ; <float*> [#uses=1] - %113 = load float* %A_addr.653, align 4 ; <float> [#uses=1] - %114 = load float* %B_addr.652, align 4 ; <float> [#uses=1] - %115 = fmul float %113, %114 ; <float> [#uses=1] - %Sum0.2 = fadd float %115, %Sum0.255 ; <float> [#uses=2] - %indvar.next118 = add i64 %indvar117, 1 ; <i64> [#uses=2] - %exitcond = icmp eq i64 %indvar.next118, %tmp. ; <i1> [#uses=1] - br i1 %exitcond, label %bb16, label %bb14 - -bb16: ; preds = %bb14, %bb13 - %Sum0.2.lcssa = phi float [ %Sum0.254, %bb13 ], [ %Sum0.2, %bb14 ] ; <float> [#uses=1] - store float %Sum0.2.lcssa, float* %C, align 4 - ret void -} diff --git a/test/CodeGen/X86/licm-dominance.ll b/test/CodeGen/X86/licm-dominance.ll index 8a0958d..019f8a3 100644 --- a/test/CodeGen/X86/licm-dominance.ll +++ b/test/CodeGen/X86/licm-dominance.ll @@ -1,7 +1,7 @@ -; RUN: llc -asm-verbose=false < %s | FileCheck %s +; RUN: llc -asm-verbose=true < %s | FileCheck %s ; MachineLICM should check dominance before hoisting instructions. -; CHECK: jne LBB0_3 +; CHECK: ## in Loop: ; CHECK-NEXT: xorb %al, %al ; CHECK-NEXT: testb %al, %al diff --git a/test/CodeGen/X86/lit.local.cfg b/test/CodeGen/X86/lit.local.cfg index b05ed3c..a8ad0f1 100644 --- a/test/CodeGen/X86/lit.local.cfg +++ b/test/CodeGen/X86/lit.local.cfg @@ -1,13 +1,6 @@ config.suffixes = ['.ll', '.c', '.cpp'] -def getRoot(config): - if not config.parent: - return config - return getRoot(config.parent) - -root = getRoot(config) - -targets = set(root.targets_to_build.split()) +targets = set(config.root.targets_to_build.split()) if not 'X86' in targets: config.unsupported = True diff --git a/test/CodeGen/X86/loop-blocks.ll b/test/CodeGen/X86/loop-blocks.ll index faba630..d14102f 100644 --- a/test/CodeGen/X86/loop-blocks.ll +++ b/test/CodeGen/X86/loop-blocks.ll @@ -41,6 +41,7 @@ done: ; CHECK-NEXT: align ; CHECK-NEXT: .LBB1_4: ; CHECK-NEXT: callq bar99 +; CHECK-NEXT: align ; CHECK-NEXT: .LBB1_1: ; CHECK-NEXT: callq body @@ -75,19 +76,21 @@ exit: ; CHECK: yet_more_involved: ; CHECK: jmp .LBB2_1 ; CHECK-NEXT: align -; CHECK-NEXT: .LBB2_4: -; CHECK-NEXT: callq bar99 +; CHECK-NEXT: .LBB2_5: +; CHECK-NEXT: callq block_a_true_func +; CHECK-NEXT: callq block_a_merge_func +; CHECK-NEXT: align +; CHECK-NEXT: .LBB2_1: +; CHECK-NEXT: callq body +; +; LBB2_4 +; CHECK: callq bar99 ; CHECK-NEXT: callq get ; CHECK-NEXT: cmpl $2999, %eax ; CHECK-NEXT: jle .LBB2_5 ; CHECK-NEXT: callq block_a_false_func ; CHECK-NEXT: callq block_a_merge_func ; CHECK-NEXT: jmp .LBB2_1 -; CHECK-NEXT: .LBB2_5: -; CHECK-NEXT: callq block_a_true_func -; CHECK-NEXT: callq block_a_merge_func -; CHECK-NEXT: .LBB2_1: -; CHECK-NEXT: callq body define void @yet_more_involved() nounwind { entry: @@ -136,17 +139,22 @@ exit: ; CHECK-NEXT: align ; CHECK-NEXT: .LBB3_7: ; CHECK-NEXT: callq bar100 -; CHECK-NEXT: jmp .LBB3_1 -; CHECK-NEXT: .LBB3_8: +; CHECK-NEXT: align +; CHECK-NEXT: .LBB3_1: +; CHECK-NEXT: callq loop_header +; CHECK: jl .LBB3_7 +; CHECK: jge .LBB3_3 ; CHECK-NEXT: callq bar101 ; CHECK-NEXT: jmp .LBB3_1 -; CHECK-NEXT: .LBB3_9: +; CHECK-NEXT: .LBB3_3: +; CHECK: jge .LBB3_4 ; CHECK-NEXT: callq bar102 ; CHECK-NEXT: jmp .LBB3_1 -; CHECK-NEXT: .LBB3_5: +; CHECK-NEXT: .LBB3_4: +; CHECK: jl .LBB3_6 ; CHECK-NEXT: callq loop_latch -; CHECK-NEXT: .LBB3_1: -; CHECK-NEXT: callq loop_header +; CHECK-NEXT: jmp .LBB3_1 +; CHECK-NEXT: .LBB3_6: define void @cfg_islands() nounwind { entry: diff --git a/test/CodeGen/X86/loop-strength-reduce3.ll b/test/CodeGen/X86/loop-strength-reduce3.ll deleted file mode 100644 index d6c265f..0000000 --- a/test/CodeGen/X86/loop-strength-reduce3.ll +++ /dev/null @@ -1,37 +0,0 @@ -; RUN: llc < %s -march=x86 -enable-lsr-nested | grep cmp | grep 240 -; RUN: llc < %s -march=x86 -enable-lsr-nested | grep inc | count 1 - -define i32 @foo(i32 %A, i32 %B, i32 %C, i32 %D) nounwind { -entry: - %tmp2955 = icmp sgt i32 %C, 0 ; <i1> [#uses=1] - br i1 %tmp2955, label %bb26.outer.us, label %bb40.split - -bb26.outer.us: ; preds = %bb26.bb32_crit_edge.us, %entry - %i.044.0.ph.us = phi i32 [ 0, %entry ], [ %indvar.next57, %bb26.bb32_crit_edge.us ] ; <i32> [#uses=2] - %k.1.ph.us = phi i32 [ 0, %entry ], [ %k.0.us, %bb26.bb32_crit_edge.us ] ; <i32> [#uses=1] - %tmp3.us = mul i32 %i.044.0.ph.us, 6 ; <i32> [#uses=1] - br label %bb1.us - -bb1.us: ; preds = %bb1.us, %bb26.outer.us - %j.053.us = phi i32 [ 0, %bb26.outer.us ], [ %tmp25.us, %bb1.us ] ; <i32> [#uses=2] - %k.154.us = phi i32 [ %k.1.ph.us, %bb26.outer.us ], [ %k.0.us, %bb1.us ] ; <i32> [#uses=1] - %tmp5.us = add i32 %tmp3.us, %j.053.us ; <i32> [#uses=1] - %tmp7.us = shl i32 %D, %tmp5.us ; <i32> [#uses=2] - %tmp9.us = icmp eq i32 %tmp7.us, %B ; <i1> [#uses=1] - %tmp910.us = zext i1 %tmp9.us to i32 ; <i32> [#uses=1] - %tmp12.us = and i32 %tmp7.us, %A ; <i32> [#uses=1] - %tmp19.us = and i32 %tmp12.us, %tmp910.us ; <i32> [#uses=1] - %k.0.us = add i32 %tmp19.us, %k.154.us ; <i32> [#uses=3] - %tmp25.us = add i32 %j.053.us, 1 ; <i32> [#uses=2] - %tmp29.us = icmp slt i32 %tmp25.us, %C ; <i1> [#uses=1] - br i1 %tmp29.us, label %bb1.us, label %bb26.bb32_crit_edge.us - -bb26.bb32_crit_edge.us: ; preds = %bb1.us - %indvar.next57 = add i32 %i.044.0.ph.us, 1 ; <i32> [#uses=2] - %exitcond = icmp eq i32 %indvar.next57, 40 ; <i1> [#uses=1] - br i1 %exitcond, label %bb40.split, label %bb26.outer.us - -bb40.split: ; preds = %bb26.bb32_crit_edge.us, %entry - %k.1.lcssa.lcssa.us-lcssa = phi i32 [ %k.0.us, %bb26.bb32_crit_edge.us ], [ 0, %entry ] ; <i32> [#uses=1] - ret i32 %k.1.lcssa.lcssa.us-lcssa -} diff --git a/test/CodeGen/X86/lsr-loop-exit-cond.ll b/test/CodeGen/X86/lsr-loop-exit-cond.ll index bdf09df..ebda9f2 100644 --- a/test/CodeGen/X86/lsr-loop-exit-cond.ll +++ b/test/CodeGen/X86/lsr-loop-exit-cond.ll @@ -1,5 +1,6 @@ ; RUN: llc -mtriple=x86_64-darwin < %s | FileCheck %s +; CHECK: t: ; CHECK: decq ; CHECK-NEXT: movl ( ; CHECK-NEXT: jne @@ -136,3 +137,44 @@ bb2: ; preds = %bb store i8 %92, i8* %93, align 1 ret void } + +; Check that DAGCombiner doesn't mess up the IV update when the exiting value +; is equal to the stride. +; It must not fold (cmp (add iv, 1), 1) --> (cmp iv, 0). + +; CHECK: f: +; CHECK: %for.body +; CHECK: incl [[IV:%e..]] +; CHECK: cmpl $1, [[IV]] +; CHECK: jne +; CHECK: ret + +define i32 @f(i32 %i, i32* nocapture %a) nounwind uwtable readonly ssp { +entry: + %cmp4 = icmp eq i32 %i, 1 + br i1 %cmp4, label %for.end, label %for.body.lr.ph + +for.body.lr.ph: ; preds = %entry + %0 = sext i32 %i to i64 + br label %for.body + +for.body: ; preds = %for.body.lr.ph, %for.body + %indvars.iv = phi i64 [ %0, %for.body.lr.ph ], [ %indvars.iv.next, %for.body ] + %bi.06 = phi i32 [ 0, %for.body.lr.ph ], [ %i.addr.0.bi.0, %for.body ] + %b.05 = phi i32 [ 0, %for.body.lr.ph ], [ %.b.0, %for.body ] + %arrayidx = getelementptr inbounds i32* %a, i64 %indvars.iv + %1 = load i32* %arrayidx, align 4 + %cmp1 = icmp ugt i32 %1, %b.05 + %.b.0 = select i1 %cmp1, i32 %1, i32 %b.05 + %2 = trunc i64 %indvars.iv to i32 + %i.addr.0.bi.0 = select i1 %cmp1, i32 %2, i32 %bi.06 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, 1 + br i1 %exitcond, label %for.end, label %for.body + +for.end: ; preds = %for.body, %entry + %bi.0.lcssa = phi i32 [ 0, %entry ], [ %i.addr.0.bi.0, %for.body ] + ret i32 %bi.0.lcssa +} + diff --git a/test/CodeGen/X86/machine-cp.ll b/test/CodeGen/X86/machine-cp.ll index 54fa01c..8e97b99 100644 --- a/test/CodeGen/X86/machine-cp.ll +++ b/test/CodeGen/X86/machine-cp.ll @@ -5,11 +5,11 @@ define i32 @t1(i32 %a, i32 %b) nounwind { entry: ; CHECK: t1: -; CHECK: jne +; CHECK: je [[LABEL:.*BB.*]] %cmp1 = icmp eq i32 %b, 0 br i1 %cmp1, label %while.end, label %while.body -; CHECK: BB +; CHECK: [[LABEL]]: ; CHECK-NOT: mov ; CHECK: ret diff --git a/test/CodeGen/X86/misched-new.ll b/test/CodeGen/X86/misched-new.ll index f3c2af8..8f2f6f7 100644 --- a/test/CodeGen/X86/misched-new.ll +++ b/test/CodeGen/X86/misched-new.ll @@ -1,8 +1,10 @@ ; RUN: llc -march=x86-64 -mcpu=core2 -enable-misched -misched=shuffle -misched-bottomup < %s -; XFAIL: * -; ...should pass. See PR12324: misched bringup +; REQUIRES: asserts ; ; Interesting MachineScheduler cases. +; +; FIXME: There should be an assert in the coalescer that we're not rematting +; "not-quite-dead" copies, but that breaks a lot of tests <rdar://problem/11148682>. declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/postra-licm.ll b/test/CodeGen/X86/postra-licm.ll index 48c48ae..01d6cbe 100644 --- a/test/CodeGen/X86/postra-licm.ll +++ b/test/CodeGen/X86/postra-licm.ll @@ -70,8 +70,8 @@ bb26.preheader: ; preds = %imix_test.exit bb23: ; preds = %imix_test.exit unreachable ; Verify that there are no loads inside the loop. -; X86-32: %bb26.preheader ; X86-32: .align 4 +; X86-32: %bb28 ; X86-32-NOT: (%esp), ; X86-32-NOT: (%ebp), ; X86-32: jmp diff --git a/test/CodeGen/X86/pr12360.ll b/test/CodeGen/X86/pr12360.ll new file mode 100644 index 0000000..f29e50e --- /dev/null +++ b/test/CodeGen/X86/pr12360.ll @@ -0,0 +1,46 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin | FileCheck %s + +define zeroext i1 @f1(i8* %x) { +; CHECK: f1: +; CHECK: movb (%rdi), %al +; CHECK-NEXT: ret + +entry: + %0 = load i8* %x, align 1, !range !0 + %tobool = trunc i8 %0 to i1 + ret i1 %tobool +} + +define zeroext i1 @f2(i8* %x) { +; CHECK: f2: +; CHECK: movb (%rdi), %al +; CHECK-NEXT: ret + +entry: + %0 = load i8* %x, align 1, !range !0 + %tobool = icmp ne i8 %0, 0 + ret i1 %tobool +} + +!0 = metadata !{i8 0, i8 2} + + +; check that we don't build a "trunc" from i1 to i1, which would assert. +define zeroext i1 @f3(i1 %x) { +; CHECK: f3: + +entry: + %tobool = icmp ne i1 %x, 0 + ret i1 %tobool +} + +; check that we don't build a trunc when other bits are needed +define zeroext i1 @f4(i32 %x) { +; CHECK: f4: +; CHECK: and + +entry: + %y = and i32 %x, 32768 + %z = icmp ne i32 %y, 0 + ret i1 %z +} diff --git a/test/CodeGen/X86/pr2659.ll b/test/CodeGen/X86/pr2659.ll index 5dab5c9..8003588 100644 --- a/test/CodeGen/X86/pr2659.ll +++ b/test/CodeGen/X86/pr2659.ll @@ -18,11 +18,12 @@ forcond.preheader: ; preds = %entry ; CHECK-NOT: xorl ; CHECK-NOT: movl ; CHECK-NOT: LBB -; CHECK: jne +; CHECK: je ; There should be no moves required in the for loop body. ; CHECK: %forbody ; CHECK-NOT: mov +; CHECK: jbe ifthen: ; preds = %entry ret i32 0 diff --git a/test/CodeGen/X86/pr3495.ll b/test/CodeGen/X86/pr3495.ll deleted file mode 100644 index 1ce3b49..0000000 --- a/test/CodeGen/X86/pr3495.ll +++ /dev/null @@ -1,79 +0,0 @@ -; RUN: llc < %s -march=x86 -stats -enable-lsr-nested |& grep {Number of spill slots allocated} | grep 1 -; PR3495 -; -; Note: this should not spill at all with either good LSR or good regalloc. - -target triple = "i386-pc-linux-gnu" -@x = external global [8 x i32], align 32 ; <[8 x i32]*> [#uses=1] -@rows = external global [8 x i32], align 32 ; <[8 x i32]*> [#uses=2] -@up = external global [15 x i32], align 32 ; <[15 x i32]*> [#uses=2] -@down = external global [15 x i32], align 32 ; <[15 x i32]*> [#uses=1] - -define i32 @queens(i32 %c) nounwind { -entry: - %tmp91 = add i32 %c, 1 ; <i32> [#uses=3] - %tmp135 = getelementptr [8 x i32]* @x, i32 0, i32 %tmp91 ; <i32*> [#uses=1] - br label %bb - -bb: ; preds = %bb569, %entry - %r25.0.reg2mem.0 = phi i32 [ 0, %entry ], [ %indvar.next715, %bb569 ] ; <i32> [#uses=4] - %tmp27 = getelementptr [8 x i32]* @rows, i32 0, i32 %r25.0.reg2mem.0 ; <i32*> [#uses=1] - %tmp28 = load i32* %tmp27, align 4 ; <i32> [#uses=1] - %tmp29 = icmp eq i32 %tmp28, 0 ; <i1> [#uses=1] - br i1 %tmp29, label %bb569, label %bb31 - -bb31: ; preds = %bb - %tmp35 = sub i32 %r25.0.reg2mem.0, 0 ; <i32> [#uses=1] - %tmp36 = getelementptr [15 x i32]* @up, i32 0, i32 %tmp35 ; <i32*> [#uses=1] - %tmp37 = load i32* %tmp36, align 4 ; <i32> [#uses=1] - %tmp38 = icmp eq i32 %tmp37, 0 ; <i1> [#uses=1] - br i1 %tmp38, label %bb569, label %bb41 - -bb41: ; preds = %bb31 - %tmp54 = sub i32 %r25.0.reg2mem.0, %c ; <i32> [#uses=1] - %tmp55 = add i32 %tmp54, 7 ; <i32> [#uses=1] - %tmp62 = getelementptr [15 x i32]* @up, i32 0, i32 %tmp55 ; <i32*> [#uses=2] - store i32 0, i32* %tmp62, align 4 - br label %bb92 - -bb92: ; preds = %bb545, %bb41 - %r20.0.reg2mem.0 = phi i32 [ 0, %bb41 ], [ %indvar.next711, %bb545 ] ; <i32> [#uses=5] - %tmp94 = getelementptr [8 x i32]* @rows, i32 0, i32 %r20.0.reg2mem.0 ; <i32*> [#uses=1] - %tmp95 = load i32* %tmp94, align 4 ; <i32> [#uses=0] - %tmp112 = add i32 %r20.0.reg2mem.0, %tmp91 ; <i32> [#uses=1] - %tmp113 = getelementptr [15 x i32]* @down, i32 0, i32 %tmp112 ; <i32*> [#uses=2] - %tmp114 = load i32* %tmp113, align 4 ; <i32> [#uses=1] - %tmp115 = icmp eq i32 %tmp114, 0 ; <i1> [#uses=1] - br i1 %tmp115, label %bb545, label %bb118 - -bb118: ; preds = %bb92 - %tmp122 = sub i32 %r20.0.reg2mem.0, %tmp91 ; <i32> [#uses=0] - store i32 0, i32* %tmp113, align 4 - store i32 %r20.0.reg2mem.0, i32* %tmp135, align 4 - br label %bb142 - -bb142: ; preds = %bb142, %bb118 - %k18.0.reg2mem.0 = phi i32 [ 0, %bb118 ], [ %indvar.next709, %bb142 ] ; <i32> [#uses=1] - %indvar.next709 = add i32 %k18.0.reg2mem.0, 1 ; <i32> [#uses=2] - %exitcond710 = icmp eq i32 %indvar.next709, 8 ; <i1> [#uses=1] - br i1 %exitcond710, label %bb155, label %bb142 - -bb155: ; preds = %bb142 - %tmp156 = tail call i32 @putchar(i32 10) nounwind ; <i32> [#uses=0] - br label %bb545 - -bb545: ; preds = %bb155, %bb92 - %indvar.next711 = add i32 %r20.0.reg2mem.0, 1 ; <i32> [#uses=2] - %exitcond712 = icmp eq i32 %indvar.next711, 8 ; <i1> [#uses=1] - br i1 %exitcond712, label %bb553, label %bb92 - -bb553: ; preds = %bb545 - store i32 1, i32* %tmp62, align 4 - br label %bb569 - -bb569: ; preds = %bb553, %bb31, %bb - %indvar.next715 = add i32 %r25.0.reg2mem.0, 1 ; <i32> [#uses=1] - br label %bb -} - -declare i32 @putchar(i32) diff --git a/test/CodeGen/X86/rd-mod-wr-eflags.ll b/test/CodeGen/X86/rd-mod-wr-eflags.ll new file mode 100644 index 0000000..faca3d7 --- /dev/null +++ b/test/CodeGen/X86/rd-mod-wr-eflags.ll @@ -0,0 +1,179 @@ +; RUN: llc < %s -march=x86-64 | FileCheck %s + +%struct.obj = type { i64 } + +; CHECK: _Z7releaseP3obj +define void @_Z7releaseP3obj(%struct.obj* nocapture %o) nounwind uwtable ssp { +entry: +; CHECK: decq (%{{rdi|rcx}}) +; CHECK-NEXT: je + %refcnt = getelementptr inbounds %struct.obj* %o, i64 0, i32 0 + %0 = load i64* %refcnt, align 8, !tbaa !0 + %dec = add i64 %0, -1 + store i64 %dec, i64* %refcnt, align 8, !tbaa !0 + %tobool = icmp eq i64 %dec, 0 + br i1 %tobool, label %if.end, label %return + +if.end: ; preds = %entry + %1 = bitcast %struct.obj* %o to i8* + tail call void @free(i8* %1) + br label %return + +return: ; preds = %entry, %if.end + ret void +} + +@c = common global i64 0, align 8 +@a = common global i32 0, align 4 +@.str = private unnamed_addr constant [5 x i8] c"%ld\0A\00", align 1 +@b = common global i32 0, align 4 + +; CHECK: test +define i32 @test() nounwind uwtable ssp { +entry: +; CHECK: decq +; CHECK-NOT: decq +%0 = load i64* @c, align 8, !tbaa !0 +%dec.i = add nsw i64 %0, -1 +store i64 %dec.i, i64* @c, align 8, !tbaa !0 +%tobool.i = icmp ne i64 %dec.i, 0 +%lor.ext.i = zext i1 %tobool.i to i32 +store i32 %lor.ext.i, i32* @a, align 4, !tbaa !3 +%call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %dec.i) nounwind +ret i32 0 +} + +; CHECK: test2 +define i32 @test2() nounwind uwtable ssp { +entry: +; CHECK-NOT: decq ({{.*}}) +%0 = load i64* @c, align 8, !tbaa !0 +%dec.i = add nsw i64 %0, -1 +store i64 %dec.i, i64* @c, align 8, !tbaa !0 +%tobool.i = icmp ne i64 %0, 0 +%lor.ext.i = zext i1 %tobool.i to i32 +store i32 %lor.ext.i, i32* @a, align 4, !tbaa !3 +%call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %dec.i) nounwind +ret i32 0 +} + +declare i32 @printf(i8* nocapture, ...) nounwind + +declare void @free(i8* nocapture) nounwind + +!0 = metadata !{metadata !"long", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA", null} +!3 = metadata !{metadata !"int", metadata !1} + +%struct.obj2 = type { i64, i32, i16, i8 } + +declare void @other(%struct.obj2* ) nounwind; + +; CHECK: example_dec +define void @example_dec(%struct.obj2* %o) nounwind uwtable ssp { +; 64 bit dec +entry: + %s64 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 0 +; CHECK-NOT: load + %0 = load i64* %s64, align 8 +; CHECK: decq ({{.*}}) + %dec = add i64 %0, -1 + store i64 %dec, i64* %s64, align 8 + %tobool = icmp eq i64 %dec, 0 + br i1 %tobool, label %if.end, label %return + +; 32 bit dec +if.end: + %s32 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 1 +; CHECK-NOT: load + %1 = load i32* %s32, align 4 +; CHECK: decl {{[0-9][0-9]*}}({{.*}}) + %dec1 = add i32 %1, -1 + store i32 %dec1, i32* %s32, align 4 + %tobool2 = icmp eq i32 %dec1, 0 + br i1 %tobool2, label %if.end1, label %return + +; 16 bit dec +if.end1: + %s16 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 2 +; CHECK-NOT: load + %2 = load i16* %s16, align 2 +; CHECK: decw {{[0-9][0-9]*}}({{.*}}) + %dec2 = add i16 %2, -1 + store i16 %dec2, i16* %s16, align 2 + %tobool3 = icmp eq i16 %dec2, 0 + br i1 %tobool3, label %if.end2, label %return + +; 8 bit dec +if.end2: + %s8 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 3 +; CHECK-NOT: load + %3 = load i8* %s8 +; CHECK: decb {{[0-9][0-9]*}}({{.*}}) + %dec3 = add i8 %3, -1 + store i8 %dec3, i8* %s8 + %tobool4 = icmp eq i8 %dec3, 0 + br i1 %tobool4, label %if.end4, label %return + +if.end4: + tail call void @other(%struct.obj2* %o) nounwind + br label %return + +return: ; preds = %if.end4, %if.end, %entry + ret void +} + +; CHECK: example_inc +define void @example_inc(%struct.obj2* %o) nounwind uwtable ssp { +; 64 bit inc +entry: + %s64 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 0 +; CHECK-NOT: load + %0 = load i64* %s64, align 8 +; CHECK: incq ({{.*}}) + %inc = add i64 %0, 1 + store i64 %inc, i64* %s64, align 8 + %tobool = icmp eq i64 %inc, 0 + br i1 %tobool, label %if.end, label %return + +; 32 bit inc +if.end: + %s32 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 1 +; CHECK-NOT: load + %1 = load i32* %s32, align 4 +; CHECK: incl {{[0-9][0-9]*}}({{.*}}) + %inc1 = add i32 %1, 1 + store i32 %inc1, i32* %s32, align 4 + %tobool2 = icmp eq i32 %inc1, 0 + br i1 %tobool2, label %if.end1, label %return + +; 16 bit inc +if.end1: + %s16 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 2 +; CHECK-NOT: load + %2 = load i16* %s16, align 2 +; CHECK: incw {{[0-9][0-9]*}}({{.*}}) + %inc2 = add i16 %2, 1 + store i16 %inc2, i16* %s16, align 2 + %tobool3 = icmp eq i16 %inc2, 0 + br i1 %tobool3, label %if.end2, label %return + +; 8 bit inc +if.end2: + %s8 = getelementptr inbounds %struct.obj2* %o, i64 0, i32 3 +; CHECK-NOT: load + %3 = load i8* %s8 +; CHECK: incb {{[0-9][0-9]*}}({{.*}}) + %inc3 = add i8 %3, 1 + store i8 %inc3, i8* %s8 + %tobool4 = icmp eq i8 %inc3, 0 + br i1 %tobool4, label %if.end4, label %return + +if.end4: + tail call void @other(%struct.obj2* %o) nounwind + br label %return + +return: + ret void +} diff --git a/test/CodeGen/X86/select.ll b/test/CodeGen/X86/select.ll index ce04e07..f465a4f 100644 --- a/test/CodeGen/X86/select.ll +++ b/test/CodeGen/X86/select.ll @@ -75,9 +75,9 @@ define void @test6(i32 %C, <4 x float>* %A, <4 x float>* %B) nounwind { ; Verify that the fmul gets sunk into the one part of the diamond where it is ; needed. ; CHECK: test6: -; CHECK: jne -; CHECK: mulps +; CHECK: je ; CHECK: ret +; CHECK: mulps ; CHECK: ret } diff --git a/test/CodeGen/X86/sibcall.ll b/test/CodeGen/X86/sibcall.ll index a9a5420..2af3559 100644 --- a/test/CodeGen/X86/sibcall.ll +++ b/test/CodeGen/X86/sibcall.ll @@ -147,7 +147,7 @@ define i32 @t11(i32 %x, i32 %y, i32 %z.0, i32 %z.1, i32 %z.2) nounwind ssp { ; 32: t11: ; 32-NOT: subl ${{[0-9]+}}, %esp -; 32: jne +; 32: je ; 32-NOT: movl ; 32-NOT: addl ${{[0-9]+}}, %esp ; 32: jmp {{_?}}foo5 diff --git a/test/CodeGen/X86/sink-hoist.ll b/test/CodeGen/X86/sink-hoist.ll index e13a817..7957eb8 100644 --- a/test/CodeGen/X86/sink-hoist.ll +++ b/test/CodeGen/X86/sink-hoist.ll @@ -7,8 +7,9 @@ ; CHECK: foo: ; CHECK-NEXT: testb $1, %dil -; CHECK-NEXT: je +; CHECK-NEXT: jne ; CHECK-NEXT: divsd +; CHECK-NEXT: movaps ; CHECK-NEXT: ret ; CHECK: divsd @@ -25,10 +26,10 @@ define double @foo(double %x, double %y, i1 %c) nounwind { ; CHECK: split: ; CHECK-NEXT: testb $1, %dil -; CHECK-NEXT: je -; CHECK-NEXT: divsd +; CHECK-NEXT: jne +; CHECK-NEXT: movaps ; CHECK-NEXT: ret -; CHECK: movaps +; CHECK: divsd ; CHECK-NEXT: ret define double @split(double %x, double %y, i1 %c) nounwind { %a = fdiv double %x, 3.2 diff --git a/test/CodeGen/X86/smul-with-overflow.ll b/test/CodeGen/X86/smul-with-overflow.ll index 7c2e247..7ac3840 100644 --- a/test/CodeGen/X86/smul-with-overflow.ll +++ b/test/CodeGen/X86/smul-with-overflow.ll @@ -19,7 +19,7 @@ overflow: ret i1 false ; CHECK: test1: ; CHECK: imull -; CHECK-NEXT: jo +; CHECK-NEXT: jno } define i1 @test2(i32 %v1, i32 %v2) nounwind { diff --git a/test/CodeGen/X86/sse41-blend.ll b/test/CodeGen/X86/sse41-blend.ll index 78604a0..1a1017d 100644 --- a/test/CodeGen/X86/sse41-blend.ll +++ b/test/CodeGen/X86/sse41-blend.ll @@ -80,3 +80,11 @@ define <2 x double> @B(<2 x double> %x, <2 x double> %y) { ret <2 x double> %min } +; CHECK: float_crash +define void @float_crash() nounwind { +entry: + %merge205vector_func.i = select <4 x i1> undef, <4 x double> undef, <4 x double> undef + %extract214vector_func.i = extractelement <4 x double> %merge205vector_func.i, i32 0 + store double %extract214vector_func.i, double addrspace(1)* undef, align 8 + ret void +} diff --git a/test/CodeGen/X86/sse41.ll b/test/CodeGen/X86/sse41.ll index 2ac4cb4..54264b1 100644 --- a/test/CodeGen/X86/sse41.ll +++ b/test/CodeGen/X86/sse41.ll @@ -183,8 +183,8 @@ define <4 x float> @insertps_3(<4 x float> %t1, <4 x float> %t2) nounwind { ; X64: insertps $0, %xmm1, %xmm0 } -define i32 @ptestz_1(<4 x float> %t1, <4 x float> %t2) nounwind { - %tmp1 = call i32 @llvm.x86.sse41.ptestz(<4 x float> %t1, <4 x float> %t2) nounwind readnone +define i32 @ptestz_1(<2 x i64> %t1, <2 x i64> %t2) nounwind { + %tmp1 = call i32 @llvm.x86.sse41.ptestz(<2 x i64> %t1, <2 x i64> %t2) nounwind readnone ret i32 %tmp1 ; X32: _ptestz_1: ; X32: ptest %xmm1, %xmm0 @@ -195,8 +195,8 @@ define i32 @ptestz_1(<4 x float> %t1, <4 x float> %t2) nounwind { ; X64: sete %al } -define i32 @ptestz_2(<4 x float> %t1, <4 x float> %t2) nounwind { - %tmp1 = call i32 @llvm.x86.sse41.ptestc(<4 x float> %t1, <4 x float> %t2) nounwind readnone +define i32 @ptestz_2(<2 x i64> %t1, <2 x i64> %t2) nounwind { + %tmp1 = call i32 @llvm.x86.sse41.ptestc(<2 x i64> %t1, <2 x i64> %t2) nounwind readnone ret i32 %tmp1 ; X32: _ptestz_2: ; X32: ptest %xmm1, %xmm0 @@ -207,8 +207,8 @@ define i32 @ptestz_2(<4 x float> %t1, <4 x float> %t2) nounwind { ; X64: sbbl %eax } -define i32 @ptestz_3(<4 x float> %t1, <4 x float> %t2) nounwind { - %tmp1 = call i32 @llvm.x86.sse41.ptestnzc(<4 x float> %t1, <4 x float> %t2) nounwind readnone +define i32 @ptestz_3(<2 x i64> %t1, <2 x i64> %t2) nounwind { + %tmp1 = call i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %t1, <2 x i64> %t2) nounwind readnone ret i32 %tmp1 ; X32: _ptestz_3: ; X32: ptest %xmm1, %xmm0 @@ -220,9 +220,9 @@ define i32 @ptestz_3(<4 x float> %t1, <4 x float> %t2) nounwind { } -declare i32 @llvm.x86.sse41.ptestz(<4 x float>, <4 x float>) nounwind readnone -declare i32 @llvm.x86.sse41.ptestc(<4 x float>, <4 x float>) nounwind readnone -declare i32 @llvm.x86.sse41.ptestnzc(<4 x float>, <4 x float>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestz(<2 x i64>, <2 x i64>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestc(<2 x i64>, <2 x i64>) nounwind readnone +declare i32 @llvm.x86.sse41.ptestnzc(<2 x i64>, <2 x i64>) nounwind readnone ; This used to compile to insertps $0 + insertps $16. insertps $0 is always ; pointless. diff --git a/test/CodeGen/X86/store_op_load_fold2.ll b/test/CodeGen/X86/store_op_load_fold2.ll index 1168622..8313166 100644 --- a/test/CodeGen/X86/store_op_load_fold2.ll +++ b/test/CodeGen/X86/store_op_load_fold2.ll @@ -1,4 +1,5 @@ -; RUN: llc < %s -mtriple=i686-linux -x86-asm-syntax=intel | FileCheck %s +; RUN: llc < %s -mtriple=i686-linux -x86-asm-syntax=att | FileCheck %s -check-prefix=ATT +; RUN: llc < %s -mtriple=i686-linux -x86-asm-syntax=intel | FileCheck %s -check-prefix=INTEL target datalayout = "e-p:32:32" %struct.Macroblock = type { i32, i32, i32, i32, i32, [8 x i32], %struct.Macroblock*, %struct.Macroblock*, i32, [2 x [4 x [4 x [2 x i32]]]], [16 x i8], [16 x i8], i32, i64, [4 x i32], [4 x i32], i64, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i16, double, i32, i32, i32, i32, i32, i32, i32, i32, i32 } @@ -16,9 +17,14 @@ cond_true2732.preheader: ; preds = %entry store i64 %tmp2676.us.us, i64* %tmp2666 ret i32 0 -; CHECK: and {{E..}}, DWORD PTR [360] -; CHECK: and DWORD PTR [356], {{E..}} -; CHECK: mov DWORD PTR [360], {{E..}} +; INTEL: and {{E..}}, DWORD PTR [360] +; INTEL: and DWORD PTR [356], {{E..}} +; FIXME: mov DWORD PTR [360], {{E..}} +; The above line comes out as 'mov 360, EAX', but when the register is ECX it works? + +; ATT: andl 360, %{{e..}} +; ATT: andl %{{e..}}, 356 +; ATT: movl %{{e..}}, 360 } diff --git a/test/CodeGen/X86/sub-with-overflow.ll b/test/CodeGen/X86/sub-with-overflow.ll index 749b5db..db8313c 100644 --- a/test/CodeGen/X86/sub-with-overflow.ll +++ b/test/CodeGen/X86/sub-with-overflow.ll @@ -20,7 +20,7 @@ overflow: ; CHECK: func1: ; CHECK: subl 20(%esp) -; CHECK-NEXT: jo +; CHECK-NEXT: jno } define i1 @func2(i32 %v1, i32 %v2) nounwind { @@ -40,7 +40,7 @@ carry: ; CHECK: func2: ; CHECK: subl 20(%esp) -; CHECK-NEXT: jb +; CHECK-NEXT: jae } declare i32 @printf(i8*, ...) nounwind diff --git a/test/CodeGen/X86/switch-bt.ll b/test/CodeGen/X86/switch-bt.ll index 8e39342..58a5c03 100644 --- a/test/CodeGen/X86/switch-bt.ll +++ b/test/CodeGen/X86/switch-bt.ll @@ -5,11 +5,11 @@ ; CHECK: movabsq $2305843009482129440, %r ; CHECK-NEXT: btq %rax, %r -; CHECK-NEXT: jb -; CHECK-NEXT: movl $671088640, %e +; CHECK-NEXT: jae +; CHECK: movl $671088640, %e ; CHECK-NEXT: btq %rax, %r -; CHECK-NEXT: jb -; CHECK-NEXT: testq %rax, %r +; CHECK-NEXT: jae +; CHECK: testq %rax, %r ; CHECK-NEXT: j define void @test(i8* %l) nounwind { @@ -60,7 +60,7 @@ define void @test2(i32 %x) nounwind ssp { ; CHECK-NEXT: movl $91 ; CHECK-NOT: movl ; CHECK-NEXT: btl -; CHECK-NEXT: jb +; CHECK-NEXT: jae entry: switch i32 %x, label %if.end [ i32 6, label %if.then @@ -85,7 +85,7 @@ define void @test3(i32 %x) nounwind { ; CHECK: cmpl $5 ; CHECK: ja ; CHECK: cmpl $4 -; CHECK: jne +; CHECK: je switch i32 %x, label %if.end [ i32 0, label %if.then i32 1, label %if.then diff --git a/test/CodeGen/X86/tail-opts.ll b/test/CodeGen/X86/tail-opts.ll index f1b9f20..6e20af5 100644 --- a/test/CodeGen/X86/tail-opts.ll +++ b/test/CodeGen/X86/tail-opts.ll @@ -113,15 +113,16 @@ altret: ; CHECK-NEXT: jbe .LBB2_3 ; CHECK-NEXT: ucomiss %xmm{{[0-2]}}, %xmm{{[0-2]}} ; CHECK-NEXT: ja .LBB2_4 -; CHECK-NEXT: .LBB2_2: -; CHECK-NEXT: movb $1, %al -; CHECK-NEXT: ret +; CHECK-NEXT: jmp .LBB2_2 ; CHECK-NEXT: .LBB2_3: ; CHECK-NEXT: ucomiss %xmm{{[0-2]}}, %xmm{{[0-2]}} ; CHECK-NEXT: jbe .LBB2_2 ; CHECK-NEXT: .LBB2_4: ; CHECK-NEXT: xorb %al, %al ; CHECK-NEXT: ret +; CHECK-NEXT: .LBB2_2: +; CHECK-NEXT: movb $1, %al +; CHECK-NEXT: ret define i1 @dont_merge_oddly(float* %result) nounwind { entry: @@ -336,10 +337,10 @@ return: ; CHECK: two: ; CHECK-NOT: XYZ +; CHECK: ret ; CHECK: movl $0, XYZ(%rip) ; CHECK: movl $1, XYZ(%rip) ; CHECK-NOT: XYZ -; CHECK: ret define void @two() nounwind optsize { entry: diff --git a/test/CodeGen/X86/tls-pie.ll b/test/CodeGen/X86/tls-pie.ll new file mode 100644 index 0000000..e2e58a54 --- /dev/null +++ b/test/CodeGen/X86/tls-pie.ll @@ -0,0 +1,63 @@ +; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu -relocation-model=pic -enable-pie \ +; RUN: | FileCheck -check-prefix=X32 %s +; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu -relocation-model=pic -enable-pie \ +; RUN: | FileCheck -check-prefix=X64 %s + +@i = thread_local global i32 15 +@i2 = external thread_local global i32 + +define i32 @f1() { +; X32: f1: +; X32: movl %gs:i@NTPOFF, %eax +; X32-NEXT: ret +; X64: f1: +; X64: movl %fs:i@TPOFF, %eax +; X64-NEXT: ret + +entry: + %tmp1 = load i32* @i + ret i32 %tmp1 +} + +define i32* @f2() { +; X32: f2: +; X32: movl %gs:0, %eax +; X32-NEXT: leal i@NTPOFF(%eax), %eax +; X32-NEXT: ret +; X64: f2: +; X64: movq %fs:0, %rax +; X64-NEXT: leaq i@TPOFF(%rax), %rax +; X64-NEXT: ret + +entry: + ret i32* @i +} + +define i32 @f3() { +; X32: f3: +; X32: movl i2@INDNTPOFF, %eax +; X32-NEXT: movl %gs:(%eax), %eax +; X32-NEXT: ret +; X64: f3: +; X64: movq i2@GOTTPOFF(%rip), %rax +; X64-NEXT: movl %fs:(%rax), %eax +; X64-NEXT: ret + +entry: + %tmp1 = load i32* @i2 + ret i32 %tmp1 +} + +define i32* @f4() { +; X32: f4: +; X32: movl %gs:0, %eax +; X32-NEXT: addl i2@INDNTPOFF, %eax +; X32-NEXT: ret +; X64: f4: +; X64: movq %fs:0, %rax +; X64-NEXT: addq i2@GOTTPOFF(%rip), %rax +; X64-NEXT: ret + +entry: + ret i32* @i2 +} diff --git a/test/CodeGen/X86/tls.ll b/test/CodeGen/X86/tls.ll new file mode 100644 index 0000000..e8a79bf --- /dev/null +++ b/test/CodeGen/X86/tls.ll @@ -0,0 +1,329 @@ +; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s +; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s +; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s +; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s + +@i1 = thread_local global i32 15 +@i2 = external thread_local global i32 +@i3 = internal thread_local global i32 15 +@i4 = hidden thread_local global i32 15 +@i5 = external hidden thread_local global i32 +@s1 = thread_local global i16 15 +@b1 = thread_local global i8 0 + +define i32 @f1() { +; X32_LINUX: f1: +; X32_LINUX: movl %gs:i1@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f1: +; X64_LINUX: movl %fs:i1@TPOFF, %eax +; X64_LINUX-NEXT: ret +; X32_WIN: f1: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movl _i1@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f1: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movl i1@SECREL(%rax), %eax +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i32* @i1 + ret i32 %tmp1 +} + +define i32* @f2() { +; X32_LINUX: f2: +; X32_LINUX: movl %gs:0, %eax +; X32_LINUX-NEXT: leal i1@NTPOFF(%eax), %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f2: +; X64_LINUX: movq %fs:0, %rax +; X64_LINUX-NEXT: leaq i1@TPOFF(%rax), %rax +; X64_LINUX-NEXT: ret +; X32_WIN: f2: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: leal _i1@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f2: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: leaq i1@SECREL(%rax), %rax +; X64_WIN-NEXT: ret + +entry: + ret i32* @i1 +} + +define i32 @f3() nounwind { +; X32_LINUX: f3: +; X32_LINUX: movl i2@INDNTPOFF, %eax +; X32_LINUX-NEXT: movl %gs:(%eax), %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f3: +; X64_LINUX: movq i2@GOTTPOFF(%rip), %rax +; X64_LINUX-NEXT: movl %fs:(%rax), %eax +; X64_LINUX-NEXT: ret +; X32_WIN: f3: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movl _i2@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f3: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movl i2@SECREL(%rax), %eax +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i32* @i2 + ret i32 %tmp1 +} + +define i32* @f4() { +; X32_LINUX: f4: +; X32_LINUX: movl %gs:0, %eax +; X32_LINUX-NEXT: addl i2@INDNTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f4: +; X64_LINUX: movq %fs:0, %rax +; X64_LINUX-NEXT: addq i2@GOTTPOFF(%rip), %rax +; X64_LINUX-NEXT: ret +; X32_WIN: f4: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: leal _i2@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f4: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: leaq i2@SECREL(%rax), %rax +; X64_WIN-NEXT: ret + +entry: + ret i32* @i2 +} + +define i32 @f5() nounwind { +; X32_LINUX: f5: +; X32_LINUX: movl %gs:i3@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f5: +; X64_LINUX: movl %fs:i3@TPOFF, %eax +; X64_LINUX-NEXT: ret +; X32_WIN: f5: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movl _i3@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f5: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movl i3@SECREL(%rax), %eax +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i32* @i3 + ret i32 %tmp1 +} + +define i32* @f6() { +; X32_LINUX: f6: +; X32_LINUX: movl %gs:0, %eax +; X32_LINUX-NEXT: leal i3@NTPOFF(%eax), %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f6: +; X64_LINUX: movq %fs:0, %rax +; X64_LINUX-NEXT: leaq i3@TPOFF(%rax), %rax +; X64_LINUX-NEXT: ret +; X32_WIN: f6: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: leal _i3@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f6: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: leaq i3@SECREL(%rax), %rax +; X64_WIN-NEXT: ret + +entry: + ret i32* @i3 +} + +define i32 @f7() { +; X32_LINUX: f7: +; X32_LINUX: movl %gs:i4@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f7: +; X64_LINUX: movl %fs:i4@TPOFF, %eax +; X64_LINUX-NEXT: ret + +entry: + %tmp1 = load i32* @i4 + ret i32 %tmp1 +} + +define i32* @f8() { +; X32_LINUX: f8: +; X32_LINUX: movl %gs:0, %eax +; X32_LINUX-NEXT: leal i4@NTPOFF(%eax), %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f8: +; X64_LINUX: movq %fs:0, %rax +; X64_LINUX-NEXT: leaq i4@TPOFF(%rax), %rax +; X64_LINUX-NEXT: ret + +entry: + ret i32* @i4 +} + +define i32 @f9() { +; X32_LINUX: f9: +; X32_LINUX: movl %gs:i5@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f9: +; X64_LINUX: movl %fs:i5@TPOFF, %eax +; X64_LINUX-NEXT: ret + +entry: + %tmp1 = load i32* @i5 + ret i32 %tmp1 +} + +define i32* @f10() { +; X32_LINUX: f10: +; X32_LINUX: movl %gs:0, %eax +; X32_LINUX-NEXT: leal i5@NTPOFF(%eax), %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f10: +; X64_LINUX: movq %fs:0, %rax +; X64_LINUX-NEXT: leaq i5@TPOFF(%rax), %rax +; X64_LINUX-NEXT: ret + +entry: + ret i32* @i5 +} + +define i16 @f11() { +; X32_LINUX: f11: +; X32_LINUX: movzwl %gs:s1@NTPOFF, %eax +; Why is this kill line here, but no where else? +; X32_LINUX-NEXT: # kill +; X32_LINUX-NEXT: ret +; X64_LINUX: f11: +; X64_LINUX: movzwl %fs:s1@TPOFF, %eax +; X64_LINUX-NEXT: # kill +; X64_LINUX-NEXT: ret +; X32_WIN: f11: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movzwl _s1@SECREL(%eax), %eax +; X32_WIN-NEXT: # kill +; X32_WIN-NEXT: ret +; X64_WIN: f11: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movzwl s1@SECREL(%rax), %eax +; X64_WIN-NEXT: # kill +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i16* @s1 + ret i16 %tmp1 +} + +define i32 @f12() { +; X32_LINUX: f12: +; X32_LINUX: movswl %gs:s1@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f12: +; X64_LINUX: movswl %fs:s1@TPOFF, %eax +; X64_LINUX-NEXT: ret +; X32_WIN: f12: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movswl _s1@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f12: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movswl s1@SECREL(%rax), %eax +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i16* @s1 + %tmp2 = sext i16 %tmp1 to i32 + ret i32 %tmp2 +} + +define i8 @f13() { +; X32_LINUX: f13: +; X32_LINUX: movb %gs:b1@NTPOFF, %al +; X32_LINUX-NEXT: ret +; X64_LINUX: f13: +; X64_LINUX: movb %fs:b1@TPOFF, %al +; X64_LINUX-NEXT: ret +; X32_WIN: f13: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movb _b1@SECREL(%eax), %al +; X32_WIN-NEXT: ret +; X64_WIN: f13: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movb b1@SECREL(%rax), %al +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i8* @b1 + ret i8 %tmp1 +} + +define i32 @f14() { +; X32_LINUX: f14: +; X32_LINUX: movsbl %gs:b1@NTPOFF, %eax +; X32_LINUX-NEXT: ret +; X64_LINUX: f14: +; X64_LINUX: movsbl %fs:b1@TPOFF, %eax +; X64_LINUX-NEXT: ret +; X32_WIN: f14: +; X32_WIN: movl __tls_index, %eax +; X32_WIN-NEXT: movl %fs:__tls_array, %ecx +; X32_WIN-NEXT: movl (%ecx,%eax,4), %eax +; X32_WIN-NEXT: movsbl _b1@SECREL(%eax), %eax +; X32_WIN-NEXT: ret +; X64_WIN: f14: +; X64_WIN: movl _tls_index(%rip), %eax +; X64_WIN-NEXT: movq %gs:88, %rcx +; X64_WIN-NEXT: movq (%rcx,%rax,8), %rax +; X64_WIN-NEXT: movsbl b1@SECREL(%rax), %eax +; X64_WIN-NEXT: ret + +entry: + %tmp1 = load i8* @b1 + %tmp2 = sext i8 %tmp1 to i32 + ret i32 %tmp2 +} + diff --git a/test/CodeGen/X86/tls1.ll b/test/CodeGen/X86/tls1.ll deleted file mode 100644 index f39658e..0000000 --- a/test/CodeGen/X86/tls1.ll +++ /dev/null @@ -1,20 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i32 15 - -define i32 @f() nounwind { -entry: - %tmp1 = load i32* @i - ret i32 %tmp1 -} -; X32_LINUX: movl %gs:i@NTPOFF, %eax -; X64_LINUX: movl %fs:i@TPOFF, %eax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: movl _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movabsq $i@SECREL, %rcx - diff --git a/test/CodeGen/X86/tls10.ll b/test/CodeGen/X86/tls10.ll deleted file mode 100644 index fb61596..0000000 --- a/test/CodeGen/X86/tls10.ll +++ /dev/null @@ -1,13 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t -; RUN: grep {movl %gs:0, %eax} %t -; RUN: grep {leal i@NTPOFF(%eax), %eax} %t -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2 -; RUN: grep {movq %fs:0, %rax} %t2 -; RUN: grep {leaq i@TPOFF(%rax), %rax} %t2 - -@i = external hidden thread_local global i32 - -define i32* @f() { -entry: - ret i32* @i -} diff --git a/test/CodeGen/X86/tls11.ll b/test/CodeGen/X86/tls11.ll deleted file mode 100644 index cc14826..0000000 --- a/test/CodeGen/X86/tls11.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i16 15 - -define i16 @f() { -entry: - %tmp1 = load i16* @i - ret i16 %tmp1 -} -; X32_LINUX: movzwl %gs:i@NTPOFF, %eax -; X64_LINUX: movzwl %fs:i@TPOFF, %eax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: movzwl _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movq %gs:88, %rcx -; X64_WIN: movabsq $i@SECREL, %rcx -; X64_WIN: movzwl (%rax,%rcx), %eax diff --git a/test/CodeGen/X86/tls12.ll b/test/CodeGen/X86/tls12.ll deleted file mode 100644 index 3da789e..0000000 --- a/test/CodeGen/X86/tls12.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i8 15 - -define i8 @f() { -entry: - %tmp1 = load i8* @i - ret i8 %tmp1 -} -; X32_LINUX: movb %gs:i@NTPOFF, %al -; X64_LINUX: movb %fs:i@TPOFF, %al -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: movb _i@SECREL(%eax), %al -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movq %gs:88, %rcx -; X64_WIN: movabsq $i@SECREL, %rcx -; X64_WIN: movb (%rax,%rcx), %al diff --git a/test/CodeGen/X86/tls13.ll b/test/CodeGen/X86/tls13.ll deleted file mode 100644 index 0f6a98a..0000000 --- a/test/CodeGen/X86/tls13.ll +++ /dev/null @@ -1,33 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i16 0 -@j = thread_local global i16 0 - -define void @f() nounwind optsize { -entry: - %0 = load i16* @i, align 2 - %1 = sext i16 %0 to i32 - tail call void @g(i32 %1) nounwind - %2 = load i16* @j, align 2 - %3 = zext i16 %2 to i32 - tail call void @h(i32 %3) nounwind - ret void -} - -declare void @g(i32) - -declare void @h(i32) - -; X32_LINUX: movswl %gs:i@NTPOFF, %eax -; X32_LINUX: movzwl %gs:j@NTPOFF, %eax -; X64_LINUX: movswl %fs:i@TPOFF, %edi -; X64_LINUX: movzwl %fs:j@TPOFF, %edi -; X32_WIN: movswl _i@SECREL(%esi), %eax -; X32_WIN: movzwl _j@SECREL(%esi), %eax -; X64_WIN: movabsq $i@SECREL, %rax -; X64_WIN: movswl (%rsi,%rax), %ecx -; X64_WIN: movabsq $j@SECREL, %rax -; X64_WIN: movzwl (%rsi,%rax), %ecx diff --git a/test/CodeGen/X86/tls14.ll b/test/CodeGen/X86/tls14.ll deleted file mode 100644 index 6462571..0000000 --- a/test/CodeGen/X86/tls14.ll +++ /dev/null @@ -1,33 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i8 0 -@j = thread_local global i8 0 - -define void @f() nounwind optsize { -entry: - %0 = load i8* @i, align 2 - %1 = sext i8 %0 to i32 - tail call void @g(i32 %1) nounwind - %2 = load i8* @j, align 2 - %3 = zext i8 %2 to i32 - tail call void @h(i32 %3) nounwind - ret void -} - -declare void @g(i32) - -declare void @h(i32) - -; X32_LINUX: movsbl %gs:i@NTPOFF, %eax -; X32_LINUX: movzbl %gs:j@NTPOFF, %eax -; X64_LINUX: movsbl %fs:i@TPOFF, %edi -; X64_LINUX: movzbl %fs:j@TPOFF, %edi -; X32_WIN: movsbl _i@SECREL(%esi), %eax -; X32_WIN: movzbl _j@SECREL(%esi), %eax -; X64_WIN: movabsq $i@SECREL, %rax -; X64_WIN: movsbl (%rsi,%rax), %ecx -; X64_WIN: movabsq $j@SECREL, %rax -; X64_WIN: movzbl (%rsi,%rax), %ecx diff --git a/test/CodeGen/X86/tls15.ll b/test/CodeGen/X86/tls15.ll deleted file mode 100644 index 7abf070..0000000 --- a/test/CodeGen/X86/tls15.ll +++ /dev/null @@ -1,18 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t -; RUN: grep {movl %gs:0, %eax} %t | count 1 -; RUN: grep {leal i@NTPOFF(%eax), %ecx} %t -; RUN: grep {leal j@NTPOFF(%eax), %eax} %t -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2 -; RUN: grep {movq %fs:0, %rax} %t2 | count 1 -; RUN: grep {leaq i@TPOFF(%rax), %rcx} %t2 -; RUN: grep {leaq j@TPOFF(%rax), %rax} %t2 - -@i = thread_local global i32 0 -@j = thread_local global i32 0 - -define void @f(i32** %a, i32** %b) { -entry: - store i32* @i, i32** %a, align 8 - store i32* @j, i32** %b, align 8 - ret void -} diff --git a/test/CodeGen/X86/tls2.ll b/test/CodeGen/X86/tls2.ll deleted file mode 100644 index e882f53..0000000 --- a/test/CodeGen/X86/tls2.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = thread_local global i32 15 - -define i32* @f() { -entry: - ret i32* @i -} -; X32_LINUX: movl %gs:0, %eax -; X32_LINUX: leal i@NTPOFF(%eax), %eax -; X64_LINUX: movq %fs:0, %rax -; X64_LINUX: leaq i@TPOFF(%rax), %rax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: leal _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movq %gs:88, %rcx -; X64_WIN: addq $i@SECREL, %rax diff --git a/test/CodeGen/X86/tls3.ll b/test/CodeGen/X86/tls3.ll deleted file mode 100644 index ee3f28f..0000000 --- a/test/CodeGen/X86/tls3.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = external thread_local global i32 ; <i32*> [#uses=2] - -define i32 @f() nounwind { -entry: - %tmp1 = load i32* @i ; <i32> [#uses=1] - ret i32 %tmp1 -} -; X32_LINUX: movl i@INDNTPOFF, %eax -; X32_LINUX: movl %gs:(%eax), %eax -; X64_LINUX: movq i@GOTTPOFF(%rip), %rax -; X64_LINUX: movl %fs:(%rax), %eax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: movl _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movabsq $i@SECREL, %rcx diff --git a/test/CodeGen/X86/tls4.ll b/test/CodeGen/X86/tls4.ll deleted file mode 100644 index 2b53ec5..0000000 --- a/test/CodeGen/X86/tls4.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = external thread_local global i32 ; <i32*> [#uses=2] - -define i32* @f() { -entry: - ret i32* @i -} -; X32_LINUX: movl %gs:0, %eax -; X32_LINUX: addl i@INDNTPOFF, %eax -; X64_LINUX: movq %fs:0, %rax -; X64_LINUX: addq i@GOTTPOFF(%rip), %rax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: leal _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movq %gs:88, %rcx -; X64_WIN: addq $i@SECREL, %rax diff --git a/test/CodeGen/X86/tls5.ll b/test/CodeGen/X86/tls5.ll deleted file mode 100644 index 3cc6dab..0000000 --- a/test/CodeGen/X86/tls5.ll +++ /dev/null @@ -1,19 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s - -@i = internal thread_local global i32 15 - -define i32 @f() { -entry: - %tmp1 = load i32* @i - ret i32 %tmp1 -} -; X32_LINUX: movl %gs:i@NTPOFF, %eax -; X64_LINUX: movl %fs:i@TPOFF, %eax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: movl _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movabsq $i@SECREL, %rcx diff --git a/test/CodeGen/X86/tls6.ll b/test/CodeGen/X86/tls6.ll deleted file mode 100644 index c98ad7c..0000000 --- a/test/CodeGen/X86/tls6.ll +++ /dev/null @@ -1,21 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu | FileCheck -check-prefix=X32_LINUX %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu | FileCheck -check-prefix=X64_LINUX %s -; RUN: llc < %s -march=x86 -mtriple=x86-pc-win32 | FileCheck -check-prefix=X32_WIN %s -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-win32 | FileCheck -check-prefix=X64_WIN %s - -@i = internal thread_local global i32 15 - -define i32* @f() { -entry: - ret i32* @i -} -; X32_LINUX: movl %gs:0, %eax -; X32_LINUX: leal i@NTPOFF(%eax), %eax -; X64_LINUX: movq %fs:0, %rax -; X64_LINUX: leaq i@TPOFF(%rax), %rax -; X32_WIN: movl __tls_index, %eax -; X32_WIN: movl %fs:__tls_array, %ecx -; X32_WIN: leal _i@SECREL(%eax), %eax -; X64_WIN: movl _tls_index(%rip), %eax -; X64_WIN: movq %gs:88, %rcx -; X64_WIN: addq $i@SECREL, %rax diff --git a/test/CodeGen/X86/tls7.ll b/test/CodeGen/X86/tls7.ll deleted file mode 100644 index e9116e7..0000000 --- a/test/CodeGen/X86/tls7.ll +++ /dev/null @@ -1,12 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t -; RUN: grep {movl %gs:i@NTPOFF, %eax} %t -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2 -; RUN: grep {movl %fs:i@TPOFF, %eax} %t2 - -@i = hidden thread_local global i32 15 - -define i32 @f() { -entry: - %tmp1 = load i32* @i - ret i32 %tmp1 -} diff --git a/test/CodeGen/X86/tls8.ll b/test/CodeGen/X86/tls8.ll deleted file mode 100644 index 375af94..0000000 --- a/test/CodeGen/X86/tls8.ll +++ /dev/null @@ -1,13 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t -; RUN: grep {movl %gs:0, %eax} %t -; RUN: grep {leal i@NTPOFF(%eax), %eax} %t -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2 -; RUN: grep {movq %fs:0, %rax} %t2 -; RUN: grep {leaq i@TPOFF(%rax), %rax} %t2 - -@i = hidden thread_local global i32 15 - -define i32* @f() { -entry: - ret i32* @i -} diff --git a/test/CodeGen/X86/tls9.ll b/test/CodeGen/X86/tls9.ll deleted file mode 100644 index 7d08df8..0000000 --- a/test/CodeGen/X86/tls9.ll +++ /dev/null @@ -1,12 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t -; RUN: grep {movl %gs:i@NTPOFF, %eax} %t -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2 -; RUN: grep {movl %fs:i@TPOFF, %eax} %t2 - -@i = external hidden thread_local global i32 - -define i32 @f() nounwind { -entry: - %tmp1 = load i32* @i - ret i32 %tmp1 -} diff --git a/test/CodeGen/X86/uint64-to-float.ll b/test/CodeGen/X86/uint64-to-float.ll index e853e77..ca764e7 100644 --- a/test/CodeGen/X86/uint64-to-float.ll +++ b/test/CodeGen/X86/uint64-to-float.ll @@ -7,13 +7,14 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3 target triple = "x86_64-apple-darwin10.0.0" ; CHECK: testq %rdi, %rdi -; CHECK-NEXT: jns LBB0_2 +; CHECK-NEXT: js LBB0_1 +; CHECK: cvtsi2ss +; CHECK-NEXT: ret +; CHECK: LBB0_1 ; CHECK: shrq ; CHECK-NEXT: andq ; CHECK-NEXT: orq ; CHECK-NEXT: cvtsi2ss -; CHECK: LBB0_2 -; CHECK-NEXT: cvtsi2ss define float @test(i64 %a) { entry: %b = uitofp i64 %a to float diff --git a/test/CodeGen/X86/utf16-cfstrings.ll b/test/CodeGen/X86/utf16-cfstrings.ll new file mode 100644 index 0000000..af76a33 --- /dev/null +++ b/test/CodeGen/X86/utf16-cfstrings.ll @@ -0,0 +1,35 @@ +; RUN: llc < %s -mtriple x86_64-apple-macosx10 | FileCheck %s +; <rdar://problem/10655949> + +%0 = type opaque +%struct.NSConstantString = type { i32*, i32, i8*, i64 } + +@__CFConstantStringClassReference = external global [0 x i32] +@.str = internal unnamed_addr constant [5 x i16] [i16 252, i16 98, i16 101, i16 114, i16 0], align 2 +@_unnamed_cfstring_ = private constant %struct.NSConstantString { i32* getelementptr inbounds ([0 x i32]* @__CFConstantStringClassReference, i32 0, i32 0), i32 2000, i8* bitcast ([5 x i16]* @.str to i8*), i64 4 }, section "__DATA,__cfstring" + +; CHECK: .section __TEXT,__ustring +; CHECK-NEXT: .align 1 +; CHECK-NEXT: _.str: +; CHECK-NEXT: .short 252 ## 0xfc +; CHECK-NEXT: .short 98 ## 0x62 +; CHECK-NEXT: .short 101 ## 0x65 +; CHECK-NEXT: .short 114 ## 0x72 +; CHECK-NEXT: .short 0 ## 0x0 + +define i32 @main() uwtable ssp { +entry: + %retval = alloca i32, align 4 + store i32 0, i32* %retval + call void (%0*, ...)* @NSLog(%0* bitcast (%struct.NSConstantString* @_unnamed_cfstring_ to %0*)) + ret i32 0 +} + +declare void @NSLog(%0*, ...) + +!llvm.module.flags = !{!0, !1, !2, !3} + +!0 = metadata !{i32 1, metadata !"Objective-C Version", i32 2} +!1 = metadata !{i32 1, metadata !"Objective-C Image Info Version", i32 0} +!2 = metadata !{i32 1, metadata !"Objective-C Image Info Section", metadata !"__DATA, __objc_imageinfo, regular, no_dead_strip"} +!3 = metadata !{i32 4, metadata !"Objective-C Garbage Collection", i32 0} diff --git a/test/CodeGen/X86/utf8.ll b/test/CodeGen/X86/utf8.ll new file mode 100644 index 0000000..67bc5ae --- /dev/null +++ b/test/CodeGen/X86/utf8.ll @@ -0,0 +1,4 @@ +; RUN: llc < %s -march=x86 | FileCheck %s + +; CHECK: iΔ +@"i\CE\94" = common global i32 0, align 4 diff --git a/test/CodeGen/X86/vec_compare-2.ll b/test/CodeGen/X86/vec_compare-2.ll index 946b126..91777f7 100644 --- a/test/CodeGen/X86/vec_compare-2.ll +++ b/test/CodeGen/X86/vec_compare-2.ll @@ -10,8 +10,10 @@ define void @blackDespeckle_wrapper(i8** %args_list, i64* %gtid, i64 %xend) { entry: ; CHECK: cfi_def_cfa_offset ; CHECK-NOT: set -; CHECK: pcmpgt -; CHECK: blendvps +; CHECK: movzwl +; CHECK: movzwl +; CHECK: pshufd +; CHECK: pshufb %shr.i = ashr <4 x i32> zeroinitializer, <i32 3, i32 3, i32 3, i32 3> ; <<4 x i32>> [#uses=1] %cmp318.i = sext <4 x i1> zeroinitializer to <4 x i32> ; <<4 x i32>> [#uses=1] %sub322.i = sub <4 x i32> %shr.i, zeroinitializer ; <<4 x i32>> [#uses=1] diff --git a/test/CodeGen/X86/vec_shuffle-20.ll b/test/CodeGen/X86/vec_shuffle-20.ll index fc06b95..976cd18 100644 --- a/test/CodeGen/X86/vec_shuffle-20.ll +++ b/test/CodeGen/X86/vec_shuffle-20.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -o /dev/null -march=x86 -mattr=+sse2 -mtriple=i686-apple-darwin9 -stats -info-output-file - | grep asm-printer | grep 3 +; RUN: llc < %s -o /dev/null -march=x86 -mcpu=corei7 -mtriple=i686-apple-darwin9 -stats -info-output-file - | grep asm-printer | grep 3 define <4 x float> @func(<4 x float> %fp0, <4 x float> %fp1) nounwind { entry: diff --git a/test/CodeGen/X86/vec_shuffle-37.ll b/test/CodeGen/X86/vec_shuffle-37.ll index 0608398..430aa04 100644 --- a/test/CodeGen/X86/vec_shuffle-37.ll +++ b/test/CodeGen/X86/vec_shuffle-37.ll @@ -4,10 +4,10 @@ define <4 x i32> @t00(<4 x i32>* %a0) nounwind ssp { entry: -; CHECK: movaps ({{%rdi|%rcx}}), %xmm0 -; CHECK: movaps %xmm0, %xmm1 -; CHECK-NEXT: movss %xmm2, %xmm1 -; CHECK-NEXT: shufps $36, %xmm1, %xmm0 +; CHECK: movaps ({{%rdi|%rcx}}), %[[XMM0:xmm[0-9]+]] +; CHECK: movaps %[[XMM0]], %[[XMM1:xmm[0-9]+]] +; CHECK-NEXT: movss %xmm{{[0-9]+}}, %[[XMM1]] +; CHECK-NEXT: shufps $36, %[[XMM1]], %[[XMM0]] %0 = load <4 x i32>* undef, align 16 %1 = load <4 x i32>* %a0, align 16 %2 = shufflevector <4 x i32> %1, <4 x i32> %0, <4 x i32> <i32 0, i32 1, i32 2, i32 4> diff --git a/test/CodeGen/X86/xor-icmp.ll b/test/CodeGen/X86/xor-icmp.ll index 34875ed..fd1b006 100644 --- a/test/CodeGen/X86/xor-icmp.ll +++ b/test/CodeGen/X86/xor-icmp.ll @@ -9,13 +9,13 @@ entry: ; X32-NOT: andb ; X32-NOT: shrb ; X32: testb $64 -; X32: jne +; X32: je ; X64: t: ; X64-NOT: setne ; X64: xorl ; X64: testb $64 -; X64: jne +; X64: je %0 = and i32 %a, 16384 %1 = icmp ne i32 %0, 0 %2 = and i32 %b, 16384 @@ -43,7 +43,7 @@ define i32 @t2(i32 %x, i32 %y) nounwind ssp { ; X32: cmpl ; X32: sete ; X32-NOT: xor -; X32: jne +; X32: je ; X64: t2: ; X64: testl @@ -51,7 +51,7 @@ define i32 @t2(i32 %x, i32 %y) nounwind ssp { ; X64: testl ; X64: sete ; X64-NOT: xor -; X64: jne +; X64: je entry: %0 = icmp eq i32 %x, 0 ; <i1> [#uses=1] %1 = icmp eq i32 %y, 0 ; <i1> [#uses=1] |