1 //===- ExportTrie.cpp -----------------------------------------------------===//
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 is a partial implementation of the Mach-O export trie format. It's
10 // essentially a symbol table encoded as a compressed prefix trie, meaning that
11 // the common prefixes of each symbol name are shared for a more compact
12 // representation. The prefixes are stored on the edges of the trie, and one
13 // edge can represent multiple characters. For example, given two exported
14 // symbols _bar and _baz, we will have a trie like this (terminal nodes are
15 // marked with an asterisk):
16 //
17 //              +-+-+
18 //              |   | // root node
19 //              +-+-+
20 //                |
21 //                | _ba
22 //                |
23 //              +-+-+
24 //              |   |
25 //              +-+-+
26 //           r /     \ z
27 //            /       \
28 //        +-+-+       +-+-+
29 //        | * |       | * |
30 //        +-+-+       +-+-+
31 //
32 // More documentation of the format can be found in
33 // llvm/tools/obj2yaml/macho2yaml.cpp.
34 //
35 //===----------------------------------------------------------------------===//
36 
37 #include "ExportTrie.h"
38 #include "Symbols.h"
39 
40 #include "lld/Common/ErrorHandler.h"
41 #include "lld/Common/Memory.h"
42 #include "llvm/ADT/Optional.h"
43 #include "llvm/BinaryFormat/MachO.h"
44 #include "llvm/Support/LEB128.h"
45 
46 using namespace llvm;
47 using namespace llvm::MachO;
48 using namespace lld;
49 using namespace lld::macho;
50 
51 namespace {
52 
53 struct Edge {
54   Edge(StringRef s, TrieNode *node) : substring(s), child(node) {}
55 
56   StringRef substring;
57   struct TrieNode *child;
58 };
59 
60 struct ExportInfo {
61   uint64_t address;
62   uint8_t flags = 0;
63   explicit ExportInfo(const Symbol &sym) : address(sym.getVA()) {
64     if (sym.isWeakDef())
65       flags |= EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION;
66     if (sym.isTlv())
67       flags |= EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL;
68     // TODO: Add proper support for re-exports & stub-and-resolver flags.
69   }
70 };
71 
72 } // namespace
73 
74 struct macho::TrieNode {
75   std::vector<Edge> edges;
76   Optional<ExportInfo> info;
77   // Estimated offset from the start of the serialized trie to the current node.
78   // This will converge to the true offset when updateOffset() is run to a
79   // fixpoint.
80   size_t offset = 0;
81 
82   // Returns whether the new estimated offset differs from the old one.
83   bool updateOffset(size_t &nextOffset);
84   void writeTo(uint8_t *buf) const;
85 };
86 
87 bool TrieNode::updateOffset(size_t &nextOffset) {
88   // Size of the whole node (including the terminalSize and the outgoing edges.)
89   // In contrast, terminalSize only records the size of the other data in the
90   // node.
91   size_t nodeSize;
92   if (info) {
93     uint32_t terminalSize =
94         getULEB128Size(info->flags) + getULEB128Size(info->address);
95     // Overall node size so far is the uleb128 size of the length of the symbol
96     // info + the symbol info itself.
97     nodeSize = terminalSize + getULEB128Size(terminalSize);
98   } else {
99     nodeSize = 1; // Size of terminalSize (which has a value of 0)
100   }
101   // Compute size of all child edges.
102   ++nodeSize; // Byte for number of children.
103   for (Edge &edge : edges) {
104     nodeSize += edge.substring.size() + 1             // String length.
105                 + getULEB128Size(edge.child->offset); // Offset len.
106   }
107   // On input, 'nextOffset' is the new preferred location for this node.
108   bool result = (offset != nextOffset);
109   // Store new location in node object for use by parents.
110   offset = nextOffset;
111   nextOffset += nodeSize;
112   return result;
113 }
114 
115 void TrieNode::writeTo(uint8_t *buf) const {
116   buf += offset;
117   if (info) {
118     // TrieNodes with Symbol info: size, flags address
119     uint32_t terminalSize =
120         getULEB128Size(info->flags) + getULEB128Size(info->address);
121     buf += encodeULEB128(terminalSize, buf);
122     buf += encodeULEB128(info->flags, buf);
123     buf += encodeULEB128(info->address, buf);
124   } else {
125     // TrieNode with no Symbol info.
126     *buf++ = 0; // terminalSize
127   }
128   // Add number of children. TODO: Handle case where we have more than 256.
129   assert(edges.size() < 256);
130   *buf++ = edges.size();
131   // Append each child edge substring and node offset.
132   for (const Edge &edge : edges) {
133     memcpy(buf, edge.substring.data(), edge.substring.size());
134     buf += edge.substring.size();
135     *buf++ = '\0';
136     buf += encodeULEB128(edge.child->offset, buf);
137   }
138 }
139 
140 TrieNode *TrieBuilder::makeNode() {
141   auto *node = make<TrieNode>();
142   nodes.emplace_back(node);
143   return node;
144 }
145 
146 static int charAt(const Symbol *sym, size_t pos) {
147   StringRef str = sym->getName();
148   if (pos >= str.size())
149     return -1;
150   return str[pos];
151 }
152 
153 // Build the trie by performing a three-way radix quicksort: We start by sorting
154 // the strings by their first characters, then sort the strings with the same
155 // first characters by their second characters, and so on recursively. Each
156 // time the prefixes diverge, we add a node to the trie.
157 //
158 // node:    The most recently created node along this path in the trie (i.e.
159 //          the furthest from the root.)
160 // lastPos: The prefix length of the most recently created node, i.e. the number
161 //          of characters along its path from the root.
162 // pos:     The string index we are currently sorting on. Note that each symbol
163 //          S contained in vec has the same prefix S[0...pos).
164 void TrieBuilder::sortAndBuild(MutableArrayRef<const Symbol *> vec,
165                                TrieNode *node, size_t lastPos, size_t pos) {
166 tailcall:
167   if (vec.empty())
168     return;
169 
170   // Partition items so that items in [0, i) are less than the pivot,
171   // [i, j) are the same as the pivot, and [j, vec.size()) are greater than
172   // the pivot.
173   const Symbol *pivotSymbol = vec[vec.size() / 2];
174   int pivot = charAt(pivotSymbol, pos);
175   size_t i = 0;
176   size_t j = vec.size();
177   for (size_t k = 0; k < j;) {
178     int c = charAt(vec[k], pos);
179     if (c < pivot)
180       std::swap(vec[i++], vec[k++]);
181     else if (c > pivot)
182       std::swap(vec[--j], vec[k]);
183     else
184       k++;
185   }
186 
187   bool isTerminal = pivot == -1;
188   bool prefixesDiverge = i != 0 || j != vec.size();
189   if (lastPos != pos && (isTerminal || prefixesDiverge)) {
190     TrieNode *newNode = makeNode();
191     node->edges.emplace_back(pivotSymbol->getName().slice(lastPos, pos),
192                              newNode);
193     node = newNode;
194     lastPos = pos;
195   }
196 
197   sortAndBuild(vec.slice(0, i), node, lastPos, pos);
198   sortAndBuild(vec.slice(j), node, lastPos, pos);
199 
200   if (isTerminal) {
201     assert(j - i == 1); // no duplicate symbols
202     node->info = ExportInfo(*pivotSymbol);
203   } else {
204     // This is the tail-call-optimized version of the following:
205     // sortAndBuild(vec.slice(i, j - i), node, lastPos, pos + 1);
206     vec = vec.slice(i, j - i);
207     ++pos;
208     goto tailcall;
209   }
210 }
211 
212 size_t TrieBuilder::build() {
213   if (exported.empty())
214     return 0;
215 
216   TrieNode *root = makeNode();
217   sortAndBuild(exported, root, 0, 0);
218 
219   // Assign each node in the vector an offset in the trie stream, iterating
220   // until all uleb128 sizes have stabilized.
221   size_t offset;
222   bool more;
223   do {
224     offset = 0;
225     more = false;
226     for (TrieNode *node : nodes)
227       more |= node->updateOffset(offset);
228   } while (more);
229 
230   return offset;
231 }
232 
233 void TrieBuilder::writeTo(uint8_t *buf) const {
234   for (TrieNode *node : nodes)
235     node->writeTo(buf);
236 }
237 
238 namespace {
239 
240 // Parse a serialized trie and invoke a callback for each entry.
241 class TrieParser {
242 public:
243   TrieParser(const uint8_t *buf, size_t size, const TrieEntryCallback &callback)
244       : start(buf), end(start + size), callback(callback) {}
245 
246   void parse(const uint8_t *buf, const Twine &cumulativeString);
247 
248   void parse() { parse(start, ""); }
249 
250   const uint8_t *start;
251   const uint8_t *end;
252   const TrieEntryCallback &callback;
253 };
254 
255 } // namespace
256 
257 void TrieParser::parse(const uint8_t *buf, const Twine &cumulativeString) {
258   if (buf >= end)
259     fatal("Node offset points outside export section");
260 
261   unsigned ulebSize;
262   uint64_t terminalSize = decodeULEB128(buf, &ulebSize);
263   buf += ulebSize;
264   uint64_t flags = 0;
265   size_t offset;
266   if (terminalSize != 0) {
267     flags = decodeULEB128(buf, &ulebSize);
268     callback(cumulativeString, flags);
269   }
270   buf += terminalSize;
271   uint8_t numEdges = *buf++;
272   for (uint8_t i = 0; i < numEdges; ++i) {
273     const char *cbuf = reinterpret_cast<const char *>(buf);
274     StringRef substring = StringRef(cbuf, strnlen(cbuf, end - buf));
275     buf += substring.size() + 1;
276     offset = decodeULEB128(buf, &ulebSize);
277     buf += ulebSize;
278     parse(start + offset, cumulativeString + substring);
279   }
280 }
281 
282 void macho::parseTrie(const uint8_t *buf, size_t size,
283                       const TrieEntryCallback &callback) {
284   if (size == 0)
285     return;
286 
287   TrieParser(buf, size, callback).parse();
288 }
289