| 1 | //===-- sanitizer_bvgraph.h -------------------------------------*- C++ -*-===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | // This file is a part of Sanitizer runtime. |
| 10 | // BVGraph -- a directed graph. |
| 11 | // |
| 12 | //===----------------------------------------------------------------------===// |
| 13 | |
| 14 | #ifndef SANITIZER_BVGRAPH_H |
| 15 | #define SANITIZER_BVGRAPH_H |
| 16 | |
| 17 | #include "sanitizer_common.h" |
| 18 | #include "sanitizer_bitvector.h" |
| 19 | |
| 20 | namespace __sanitizer { |
| 21 | |
| 22 | // Directed graph of fixed size implemented as an array of bit vectors. |
| 23 | // Not thread-safe, all accesses should be protected by an external lock. |
| 24 | template<class BV> |
| 25 | class BVGraph { |
| 26 | public: |
| 27 | enum SizeEnum : uptr { kSize = BV::kSize }; |
| 28 | uptr size() const { return kSize; } |
| 29 | // No CTOR. |
| 30 | void clear() { |
| 31 | for (uptr i = 0; i < size(); i++) |
| 32 | v[i].clear(); |
| 33 | } |
| 34 | |
| 35 | bool empty() const { |
| 36 | for (uptr i = 0; i < size(); i++) |
| 37 | if (!v[i].empty()) |
| 38 | return false; |
| 39 | return true; |
| 40 | } |
| 41 | |
| 42 | // Returns true if a new edge was added. |
| 43 | bool addEdge(uptr from, uptr to) { |
| 44 | check(idx1: from, idx2: to); |
| 45 | return v[from].setBit(to); |
| 46 | } |
| 47 | |
| 48 | // Returns true if at least one new edge was added. |
| 49 | uptr addEdges(const BV &from, uptr to, uptr added_edges[], |
| 50 | uptr max_added_edges) { |
| 51 | uptr res = 0; |
| 52 | t1.copyFrom(from); |
| 53 | while (!t1.empty()) { |
| 54 | uptr node = t1.getAndClearFirstOne(); |
| 55 | if (v[node].setBit(to)) |
| 56 | if (res < max_added_edges) |
| 57 | added_edges[res++] = node; |
| 58 | } |
| 59 | return res; |
| 60 | } |
| 61 | |
| 62 | // *EXPERIMENTAL* |
| 63 | // Returns true if an edge from=>to exist. |
| 64 | // This function does not use any global state except for 'this' itself, |
| 65 | // and thus can be called from different threads w/o locking. |
| 66 | // This would be racy. |
| 67 | // FIXME: investigate how much we can prove about this race being "benign". |
| 68 | bool hasEdge(uptr from, uptr to) { return v[from].getBit(to); } |
| 69 | |
| 70 | // Returns true if the edge from=>to was removed. |
| 71 | bool removeEdge(uptr from, uptr to) { |
| 72 | return v[from].clearBit(to); |
| 73 | } |
| 74 | |
| 75 | // Returns true if at least one edge *=>to was removed. |
| 76 | bool removeEdgesTo(const BV &to) { |
| 77 | bool res = 0; |
| 78 | for (uptr from = 0; from < size(); from++) { |
| 79 | if (v[from].setDifference(to)) |
| 80 | res = true; |
| 81 | } |
| 82 | return res; |
| 83 | } |
| 84 | |
| 85 | // Returns true if at least one edge from=>* was removed. |
| 86 | bool removeEdgesFrom(const BV &from) { |
| 87 | bool res = false; |
| 88 | t1.copyFrom(from); |
| 89 | while (!t1.empty()) { |
| 90 | uptr idx = t1.getAndClearFirstOne(); |
| 91 | if (!v[idx].empty()) { |
| 92 | v[idx].clear(); |
| 93 | res = true; |
| 94 | } |
| 95 | } |
| 96 | return res; |
| 97 | } |
| 98 | |
| 99 | void removeEdgesFrom(uptr from) { |
| 100 | return v[from].clear(); |
| 101 | } |
| 102 | |
| 103 | bool hasEdge(uptr from, uptr to) const { |
| 104 | check(idx1: from, idx2: to); |
| 105 | return v[from].getBit(to); |
| 106 | } |
| 107 | |
| 108 | // Returns true if there is a path from the node 'from' |
| 109 | // to any of the nodes in 'targets'. |
| 110 | bool isReachable(uptr from, const BV &targets) { |
| 111 | BV &to_visit = t1, |
| 112 | &visited = t2; |
| 113 | to_visit.copyFrom(v[from]); |
| 114 | visited.clear(); |
| 115 | visited.setBit(from); |
| 116 | while (!to_visit.empty()) { |
| 117 | uptr idx = to_visit.getAndClearFirstOne(); |
| 118 | if (visited.setBit(idx)) |
| 119 | to_visit.setUnion(v[idx]); |
| 120 | } |
| 121 | return targets.intersectsWith(visited); |
| 122 | } |
| 123 | |
| 124 | // Finds a path from 'from' to one of the nodes in 'target', |
| 125 | // stores up to 'path_size' items of the path into 'path', |
| 126 | // returns the path length, or 0 if there is no path of size 'path_size'. |
| 127 | uptr findPath(uptr from, const BV &targets, uptr *path, uptr path_size) { |
| 128 | if (path_size == 0) |
| 129 | return 0; |
| 130 | path[0] = from; |
| 131 | if (targets.getBit(from)) |
| 132 | return 1; |
| 133 | // The function is recursive, so we don't want to create BV on stack. |
| 134 | // Instead of a getAndClearFirstOne loop we use the slower iterator. |
| 135 | for (typename BV::Iterator it(v[from]); it.hasNext(); ) { |
| 136 | uptr idx = it.next(); |
| 137 | if (uptr res = findPath(from: idx, targets, path: path + 1, path_size: path_size - 1)) |
| 138 | return res + 1; |
| 139 | } |
| 140 | return 0; |
| 141 | } |
| 142 | |
| 143 | // Same as findPath, but finds a shortest path. |
| 144 | uptr findShortestPath(uptr from, const BV &targets, uptr *path, |
| 145 | uptr path_size) { |
| 146 | for (uptr p = 1; p <= path_size; p++) |
| 147 | if (findPath(from, targets, path, path_size: p) == p) |
| 148 | return p; |
| 149 | return 0; |
| 150 | } |
| 151 | |
| 152 | private: |
| 153 | void check(uptr idx1, uptr idx2) const { |
| 154 | CHECK_LT(idx1, size()); |
| 155 | CHECK_LT(idx2, size()); |
| 156 | } |
| 157 | BV v[kSize]; |
| 158 | // Keep temporary vectors here since we can not create large objects on stack. |
| 159 | BV t1, t2; |
| 160 | }; |
| 161 | |
| 162 | } // namespace __sanitizer |
| 163 | |
| 164 | #endif // SANITIZER_BVGRAPH_H |
| 165 | |