Imported Upstream version 6.10.0.49

Former-commit-id: 1d6753294b2993e1fbf92de9366bb9544db4189b

external/llvm-project/polly/test/Isl/CodeGen/LoopParallelMD/do_not_mutate_debug_info.ll

@@ -0,0 +1,69 @@
+; This test checks that we do not accidently mutate the debug info when
+; inserting loop parallel metadata.
+; RUN: opt %loadPolly < %s  -S -polly -polly-codegen -polly-ast-detect-parallel | FileCheck %s
+; CHECK-NOT: !7 = !{!7}
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+
+@A = common global i32* null, align 8
+
+; Function Attrs: nounwind uwtable
+define void @foo() !dbg !4 {
+entry:
+  tail call void @llvm.dbg.value(metadata i32 0, i64 0, metadata !9, metadata !19), !dbg !20
+  %0 = load i32*, i32** @A, align 8, !dbg !21, !tbaa !23
+  br label %for.body, !dbg !27
+
+for.body:                                         ; preds = %for.body, %entry
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds i32, i32* %0, i64 %indvars.iv, !dbg !21
+  %1 = load i32, i32* %arrayidx, align 4, !dbg !21, !tbaa !30
+  %add = add nsw i32 %1, 1, !dbg !21
+  store i32 %add, i32* %arrayidx, align 4, !dbg !21, !tbaa !30
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !27
+  %exitcond = icmp eq i64 %indvars.iv, 1, !dbg !27
+  br i1 %exitcond, label %for.end, label %for.body, !dbg !27
+
+for.end:                                          ; preds = %for.body
+  ret void, !dbg !32
+}
+
+; Function Attrs: nounwind readnone
+declare void @llvm.dbg.value(metadata, i64, metadata, metadata)
+
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!15, !16}
+!llvm.ident = !{!17}
+
+!0 = distinct !DICompileUnit(language: DW_LANG_C99, producer: "clang version 3.6.0 ", isOptimized: true, emissionKind: 1, file: !1, enums: !2, retainedTypes: !2, globals: !12, imports: !2)
+!1 = !DIFile(filename: "t2.c", directory: "/local/mnt/workspace/build/tip-Release")
+!2 = !{}
+!4 = distinct !DISubprogram(name: "foo", line: 3, isLocal: false, isDefinition: true, isOptimized: true, unit: !0, scopeLine: 3, file: !1, scope: !5, type: !6, variables: !8)
+!5 = !DIFile(filename: "t2.c", directory: "/local/mnt/workspace/build/tip-Release")
+!6 = !DISubroutineType(types: !7)
+!7 = !{null}
+!8 = !{!9}
+!9 = !DILocalVariable(name: "i", line: 4, scope: !10, file: !5, type: !11)
+!10 = distinct !DILexicalBlock(line: 4, column: 3, file: !1, scope: !4)
+!11 = !DIBasicType(tag: DW_TAG_base_type, name: "int", size: 32, align: 32, encoding: DW_ATE_signed)
+!12 = !{!13}
+!13 = !DIGlobalVariableExpression(var: !DIGlobalVariable(name: "A", line: 2, isLocal: false, isDefinition: true, scope: null, file: !5, type: !14), expr: !DIExpression())
+!14 = !DIDerivedType(tag: DW_TAG_pointer_type, size: 64, align: 64, baseType: !11)
+!15 = !{i32 2, !"Dwarf Version", i32 4}
+!16 = !{i32 2, !"Debug Info Version", i32 3}
+!17 = !{!"clang version 3.6.0 "}
+!18 = !{i32 0}
+!19 = !DIExpression()
+!20 = !DILocation(line: 4, column: 12, scope: !10)
+!21 = !DILocation(line: 5, column: 5, scope: !22)
+!22 = distinct !DILexicalBlock(line: 4, column: 3, file: !1, scope: !10)
+!23 = !{!24, !24, i64 0}
+!24 = !{!"any pointer", !25, i64 0}
+!25 = !{!"omnipotent char", !26, i64 0}
+!26 = !{!"Simple C/C++ TBAA"}
+!27 = !DILocation(line: 4, column: 3, scope: !28)
+!28 = !DILexicalBlockFile(discriminator: 2, file: !1, scope: !29)
+!29 = !DILexicalBlockFile(discriminator: 1, file: !1, scope: !22)
+!30 = !{!31, !31, i64 0}
+!31 = !{!"int", !25, i64 0}
+!32 = !DILocation(line: 6, column: 1, scope: !4)

external/llvm-project/polly/test/Isl/CodeGen/LoopParallelMD/loop_nest_param_parallel.ll

@@ -0,0 +1,64 @@
+; RUN: opt %loadPolly -polly-codegen -polly-ast-detect-parallel -S < %s | FileCheck %s
+;
+; Check that we mark multiple parallel loops correctly including the memory instructions.
+;
+; CHECK-DAG:  %polly.loop_cond[[COuter:[0-9]*]] = icmp sle i64 %polly.indvar_next{{[0-9]*}}, 1023
+; CHECK-DAG:  br i1 %polly.loop_cond[[COuter]], label %polly.loop_header{{[0-9]*}}, label %polly.loop_exit{{[0-9]*}}, !llvm.loop ![[IDOuter:[0-9]*]]
+;
+; CHECK-DAG:  %polly.loop_cond[[CInner:[0-9]*]] = icmp sle i64 %polly.indvar_next{{[0-9]*}}, 511
+; CHECK-DAG:  br i1 %polly.loop_cond[[CInner]], label %polly.loop_header{{[0-9]*}}, label %polly.loop_exit{{[0-9]*}}, !llvm.loop ![[IDInner:[0-9]*]]
+;
+; CHECK-DAG: store i32 %{{[a-z_0-9]*}}, i32* %{{[a-z_0-9]*}}, {{[ ._!,a-zA-Z0-9]*}}, !llvm.mem.parallel_loop_access !4
+;
+; CHECK-DAG: ![[IDOuter]] = distinct !{![[IDOuter]]}
+; CHECK-DAG: ![[IDInner]] = distinct !{![[IDInner]]}
+; CHECK-DAG: !4 = !{![[IDOuter]], ![[IDInner]]}
+;
+;    void jd(int *A) {
+;      for (int i = 0; i < 1024; i++)
+;        for (int j = 0; j < 512; j++)
+;          A[i * 512 + j] = i + j;
+;    }
+;
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+
+define void @jd(i32* %A) {
+entry:
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.inc5, %entry
+  %indvars.iv3 = phi i64 [ %indvars.iv.next4, %for.inc5 ], [ 0, %entry ]
+  %exitcond6 = icmp ne i64 %indvars.iv3, 1024
+  br i1 %exitcond6, label %for.body, label %for.end7
+
+for.body:                                         ; preds = %for.cond
+  br label %for.cond1
+
+for.cond1:                                        ; preds = %for.inc, %for.body
+  %indvars.iv = phi i64 [ %indvars.iv.next, %for.inc ], [ 0, %for.body ]
+  %exitcond = icmp ne i64 %indvars.iv, 512
+  br i1 %exitcond, label %for.body3, label %for.end
+
+for.body3:                                        ; preds = %for.cond1
+  %tmp = add nsw i64 %indvars.iv3, %indvars.iv
+  %tmp7 = shl nsw i64 %indvars.iv3, 9
+  %tmp8 = add nsw i64 %tmp7, %indvars.iv
+  %arrayidx = getelementptr inbounds i32, i32* %A, i64 %tmp8
+  %tmp9 = trunc i64 %tmp to i32
+  store i32 %tmp9, i32* %arrayidx, align 4
+  br label %for.inc
+
+for.inc:                                          ; preds = %for.body3
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  br label %for.cond1
+
+for.end:                                          ; preds = %for.cond1
+  br label %for.inc5
+
+for.inc5:                                         ; preds = %for.end
+  %indvars.iv.next4 = add nuw nsw i64 %indvars.iv3, 1
+  br label %for.cond
+
+for.end7:                                         ; preds = %for.cond
+  ret void
+}

external/llvm-project/polly/test/Isl/CodeGen/LoopParallelMD/single_loop_param_parallel.ll

@@ -0,0 +1,86 @@
+; RUN: opt %loadPolly -polly-codegen -S < %s | FileCheck %s -check-prefix=SEQUENTIAL
+; RUN: opt %loadPolly -polly-codegen -polly-ast-detect-parallel -S < %s | FileCheck %s -check-prefix=PARALLEL
+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"
+
+; This is a trivially parallel loop. We just use it to ensure that we actually
+; emit the right information.
+;
+; for (i = 0; i < n; i++)
+;   A[i] = 1;
+;
+@A = common global [1024 x i32] zeroinitializer
+define void @test-one(i64 %n) {
+start:
+  fence seq_cst
+  br label %loop.header
+
+loop.header:
+  %i = phi i64 [ 0, %start ], [ %i.next, %loop.backedge ]
+  %exitcond = icmp ne i64 %i, %n
+  br i1 %exitcond, label %loop.body, label %ret
+
+loop.body:
+  %scevgep = getelementptr [1024 x i32], [1024 x i32]* @A, i64 0, i64 %i
+  store i32 1, i32* %scevgep
+  br label %loop.backedge
+
+loop.backedge:
+  %i.next = add nsw i64 %i, 1
+  br label %loop.header
+
+ret:
+  fence seq_cst
+  ret void
+}
+
+; SEQUENTIAL: @test-one
+; SEQUENTIAL-NOT: !llvm.mem.parallel_loop_access
+; SEQUENTIAL-NOT: !llvm.loop
+
+; PARALLEL: @test-one
+; PARALLEL: store i32 1, i32* %scevgep1, {{[ ._!,a-zA-Z0-9]*}}, !llvm.mem.parallel_loop_access ![[LoopID:[0-9]*]]
+; PARALLEL:  br i1 %polly.loop_cond, label %polly.loop_header, label %polly.loop_exit, !llvm.loop ![[LoopID]]
+
+; This loop has memory dependences that require at least a simple dependence
+; analysis to detect the parallelism.
+;
+; for (i = 0; i < n; i++)
+;   A[2 * i] = A[2 * i + 1];
+;
+define void @test-two(i64 %n) {
+start:
+  fence seq_cst
+  br label %loop.header
+
+loop.header:
+  %i = phi i64 [ 0, %start ], [ %i.next, %loop.backedge ]
+  %exitcond = icmp ne i64 %i, %n
+  br i1 %exitcond, label %loop.body, label %ret
+
+loop.body:
+  %loadoffset1 = mul nsw i64 %i, 2
+  %loadoffset2 = add nsw i64 %loadoffset1, 1
+  %scevgepload = getelementptr [1024 x i32], [1024 x i32]* @A, i64 0, i64 %loadoffset2
+  %val = load i32, i32* %scevgepload
+  %storeoffset = mul i64 %i, 2
+  %scevgepstore = getelementptr [1024 x i32], [1024 x i32]* @A, i64 0, i64 %storeoffset
+  store i32 %val, i32* %scevgepstore
+  br label %loop.backedge
+
+loop.backedge:
+  %i.next = add nsw i64 %i, 1
+  br label %loop.header
+
+ret:
+  fence seq_cst
+  ret void
+}
+
+; SEQUENTIAL: @test-two
+; SEQUENTIAL-NOT: !llvm.mem.parallel_loop_access
+; SEQUENTIAL-NOT: !llvm.loop
+
+; PARALLEL: @test-two
+; PARALLEL: %val_p_scalar_ = load i32, i32* %scevgep, {{[ ._!,a-zA-Z0-9]*}}, !llvm.mem.parallel_loop_access ![[LoopID:[0-9]*]]
+; PARALLEL: store i32 %val_p_scalar_, i32* %scevgep1, {{[ ._!,a-zA-Z0-9]*}}, !llvm.mem.parallel_loop_access ![[LoopID]]
+; PARALLEL:  br i1 %polly.loop_cond, label %polly.loop_header, label %polly.loop_exit, !llvm.loop ![[LoopID]]