1 //===- InputChunks.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 #include "InputChunks.h"
10 #include "Config.h"
11 #include "OutputSegment.h"
12 #include "WriterUtils.h"
13 #include "lld/Common/ErrorHandler.h"
14 #include "lld/Common/LLVM.h"
15 #include "llvm/Support/LEB128.h"
16 
17 #define DEBUG_TYPE "lld"
18 
19 using namespace llvm;
20 using namespace llvm::wasm;
21 using namespace llvm::support::endian;
22 
23 namespace lld {
24 StringRef relocTypeToString(uint8_t relocType) {
25   switch (relocType) {
26 #define WASM_RELOC(NAME, REL)                                                  \
27   case REL:                                                                    \
28     return #NAME;
29 #include "llvm/BinaryFormat/WasmRelocs.def"
30 #undef WASM_RELOC
31   }
32   llvm_unreachable("unknown reloc type");
33 }
34 
35 bool relocIs64(uint8_t relocType) {
36   switch (relocType) {
37   case R_WASM_MEMORY_ADDR_LEB64:
38   case R_WASM_MEMORY_ADDR_SLEB64:
39   case R_WASM_MEMORY_ADDR_REL_SLEB64:
40   case R_WASM_MEMORY_ADDR_I64:
41     return true;
42   default:
43     return false;
44   }
45 }
46 
47 std::string toString(const wasm::InputChunk *c) {
48   return (toString(c->file) + ":(" + c->getName() + ")").str();
49 }
50 
51 namespace wasm {
52 StringRef InputChunk::getComdatName() const {
53   uint32_t index = getComdat();
54   if (index == UINT32_MAX)
55     return StringRef();
56   return file->getWasmObj()->linkingData().Comdats[index];
57 }
58 
59 void InputChunk::verifyRelocTargets() const {
60   for (const WasmRelocation &rel : relocations) {
61     uint64_t existingValue;
62     unsigned bytesRead = 0;
63     unsigned paddedLEBWidth = 5;
64     auto offset = rel.Offset - getInputSectionOffset();
65     const uint8_t *loc = data().data() + offset;
66     switch (rel.Type) {
67     case R_WASM_TYPE_INDEX_LEB:
68     case R_WASM_FUNCTION_INDEX_LEB:
69     case R_WASM_GLOBAL_INDEX_LEB:
70     case R_WASM_EVENT_INDEX_LEB:
71     case R_WASM_MEMORY_ADDR_LEB:
72       existingValue = decodeULEB128(loc, &bytesRead);
73       break;
74     case R_WASM_MEMORY_ADDR_LEB64:
75       existingValue = decodeULEB128(loc, &bytesRead);
76       paddedLEBWidth = 10;
77       break;
78     case R_WASM_TABLE_INDEX_SLEB:
79     case R_WASM_TABLE_INDEX_REL_SLEB:
80     case R_WASM_MEMORY_ADDR_SLEB:
81     case R_WASM_MEMORY_ADDR_REL_SLEB:
82     case R_WASM_MEMORY_ADDR_TLS_SLEB:
83       existingValue = static_cast<uint64_t>(decodeSLEB128(loc, &bytesRead));
84       break;
85     case R_WASM_TABLE_INDEX_SLEB64:
86     case R_WASM_MEMORY_ADDR_SLEB64:
87     case R_WASM_MEMORY_ADDR_REL_SLEB64:
88       existingValue = static_cast<uint64_t>(decodeSLEB128(loc, &bytesRead));
89       paddedLEBWidth = 10;
90       break;
91     case R_WASM_TABLE_INDEX_I32:
92     case R_WASM_MEMORY_ADDR_I32:
93     case R_WASM_FUNCTION_OFFSET_I32:
94     case R_WASM_SECTION_OFFSET_I32:
95     case R_WASM_GLOBAL_INDEX_I32:
96       existingValue = read32le(loc);
97       break;
98     case R_WASM_TABLE_INDEX_I64:
99     case R_WASM_MEMORY_ADDR_I64:
100       existingValue = read64le(loc);
101       break;
102     default:
103       llvm_unreachable("unknown relocation type");
104     }
105 
106     if (bytesRead && bytesRead != paddedLEBWidth)
107       warn("expected LEB at relocation site be 5/10-byte padded");
108 
109     if (rel.Type != R_WASM_GLOBAL_INDEX_LEB &&
110         rel.Type != R_WASM_GLOBAL_INDEX_I32) {
111       auto expectedValue = file->calcExpectedValue(rel);
112       if (expectedValue != existingValue)
113         warn(toString(this) + ": unexpected existing value for " +
114              relocTypeToString(rel.Type) + ": existing=" +
115              Twine(existingValue) + " expected=" + Twine(expectedValue));
116     }
117   }
118 }
119 
120 // Copy this input chunk to an mmap'ed output file and apply relocations.
121 void InputChunk::writeTo(uint8_t *buf) const {
122   // Copy contents
123   memcpy(buf + outputOffset, data().data(), data().size());
124 
125   // Apply relocations
126   if (relocations.empty())
127     return;
128 
129 #ifndef NDEBUG
130   verifyRelocTargets();
131 #endif
132 
133   LLVM_DEBUG(dbgs() << "applying relocations: " << toString(this)
134                     << " count=" << relocations.size() << "\n");
135   int32_t off = outputOffset - getInputSectionOffset();
136 
137   for (const WasmRelocation &rel : relocations) {
138     uint8_t *loc = buf + rel.Offset + off;
139     auto value = file->calcNewValue(rel);
140     LLVM_DEBUG(dbgs() << "apply reloc: type=" << relocTypeToString(rel.Type));
141     if (rel.Type != R_WASM_TYPE_INDEX_LEB)
142       LLVM_DEBUG(dbgs() << " sym=" << file->getSymbols()[rel.Index]->getName());
143     LLVM_DEBUG(dbgs() << " addend=" << rel.Addend << " index=" << rel.Index
144                       << " value=" << value << " offset=" << rel.Offset
145                       << "\n");
146 
147     switch (rel.Type) {
148     case R_WASM_TYPE_INDEX_LEB:
149     case R_WASM_FUNCTION_INDEX_LEB:
150     case R_WASM_GLOBAL_INDEX_LEB:
151     case R_WASM_EVENT_INDEX_LEB:
152     case R_WASM_MEMORY_ADDR_LEB:
153       encodeULEB128(value, loc, 5);
154       break;
155     case R_WASM_MEMORY_ADDR_LEB64:
156       encodeULEB128(value, loc, 10);
157       break;
158     case R_WASM_TABLE_INDEX_SLEB:
159     case R_WASM_TABLE_INDEX_REL_SLEB:
160     case R_WASM_MEMORY_ADDR_SLEB:
161     case R_WASM_MEMORY_ADDR_REL_SLEB:
162     case R_WASM_MEMORY_ADDR_TLS_SLEB:
163       encodeSLEB128(static_cast<int32_t>(value), loc, 5);
164       break;
165     case R_WASM_TABLE_INDEX_SLEB64:
166     case R_WASM_MEMORY_ADDR_SLEB64:
167     case R_WASM_MEMORY_ADDR_REL_SLEB64:
168       encodeSLEB128(static_cast<int64_t>(value), loc, 10);
169       break;
170     case R_WASM_TABLE_INDEX_I32:
171     case R_WASM_MEMORY_ADDR_I32:
172     case R_WASM_FUNCTION_OFFSET_I32:
173     case R_WASM_SECTION_OFFSET_I32:
174     case R_WASM_GLOBAL_INDEX_I32:
175       write32le(loc, value);
176       break;
177     case R_WASM_TABLE_INDEX_I64:
178     case R_WASM_MEMORY_ADDR_I64:
179       write64le(loc, value);
180       break;
181     default:
182       llvm_unreachable("unknown relocation type");
183     }
184   }
185 }
186 
187 // Copy relocation entries to a given output stream.
188 // This function is used only when a user passes "-r". For a regular link,
189 // we consume relocations instead of copying them to an output file.
190 void InputChunk::writeRelocations(raw_ostream &os) const {
191   if (relocations.empty())
192     return;
193 
194   int32_t off = outputOffset - getInputSectionOffset();
195   LLVM_DEBUG(dbgs() << "writeRelocations: " << file->getName()
196                     << " offset=" << Twine(off) << "\n");
197 
198   for (const WasmRelocation &rel : relocations) {
199     writeUleb128(os, rel.Type, "reloc type");
200     writeUleb128(os, rel.Offset + off, "reloc offset");
201     writeUleb128(os, file->calcNewIndex(rel), "reloc index");
202 
203     if (relocTypeHasAddend(rel.Type))
204       writeSleb128(os, file->calcNewAddend(rel), "reloc addend");
205   }
206 }
207 
208 void InputFunction::setFunctionIndex(uint32_t index) {
209   LLVM_DEBUG(dbgs() << "InputFunction::setFunctionIndex: " << getName()
210                     << " -> " << index << "\n");
211   assert(!hasFunctionIndex());
212   functionIndex = index;
213 }
214 
215 void InputFunction::setTableIndex(uint32_t index) {
216   LLVM_DEBUG(dbgs() << "InputFunction::setTableIndex: " << getName() << " -> "
217                     << index << "\n");
218   assert(!hasTableIndex());
219   tableIndex = index;
220 }
221 
222 // Write a relocation value without padding and return the number of bytes
223 // witten.
224 static unsigned writeCompressedReloc(uint8_t *buf, const WasmRelocation &rel,
225                                      uint64_t value) {
226   switch (rel.Type) {
227   case R_WASM_TYPE_INDEX_LEB:
228   case R_WASM_FUNCTION_INDEX_LEB:
229   case R_WASM_GLOBAL_INDEX_LEB:
230   case R_WASM_EVENT_INDEX_LEB:
231   case R_WASM_MEMORY_ADDR_LEB:
232   case R_WASM_MEMORY_ADDR_LEB64:
233     return encodeULEB128(value, buf);
234   case R_WASM_TABLE_INDEX_SLEB:
235   case R_WASM_TABLE_INDEX_SLEB64:
236   case R_WASM_MEMORY_ADDR_SLEB:
237   case R_WASM_MEMORY_ADDR_SLEB64:
238     return encodeSLEB128(static_cast<int64_t>(value), buf);
239   default:
240     llvm_unreachable("unexpected relocation type");
241   }
242 }
243 
244 static unsigned getRelocWidthPadded(const WasmRelocation &rel) {
245   switch (rel.Type) {
246   case R_WASM_TYPE_INDEX_LEB:
247   case R_WASM_FUNCTION_INDEX_LEB:
248   case R_WASM_GLOBAL_INDEX_LEB:
249   case R_WASM_EVENT_INDEX_LEB:
250   case R_WASM_MEMORY_ADDR_LEB:
251   case R_WASM_TABLE_INDEX_SLEB:
252   case R_WASM_MEMORY_ADDR_SLEB:
253     return 5;
254   case R_WASM_TABLE_INDEX_SLEB64:
255   case R_WASM_MEMORY_ADDR_LEB64:
256   case R_WASM_MEMORY_ADDR_SLEB64:
257     return 10;
258   default:
259     llvm_unreachable("unexpected relocation type");
260   }
261 }
262 
263 static unsigned getRelocWidth(const WasmRelocation &rel, uint64_t value) {
264   uint8_t buf[10];
265   return writeCompressedReloc(buf, rel, value);
266 }
267 
268 // Relocations of type LEB and SLEB in the code section are padded to 5 bytes
269 // so that a fast linker can blindly overwrite them without needing to worry
270 // about the number of bytes needed to encode the values.
271 // However, for optimal output the code section can be compressed to remove
272 // the padding then outputting non-relocatable files.
273 // In this case we need to perform a size calculation based on the value at each
274 // relocation.  At best we end up saving 4 bytes for each relocation entry.
275 //
276 // This function only computes the final output size.  It must be called
277 // before getSize() is used to calculate of layout of the code section.
278 void InputFunction::calculateSize() {
279   if (!file || !config->compressRelocations)
280     return;
281 
282   LLVM_DEBUG(dbgs() << "calculateSize: " << getName() << "\n");
283 
284   const uint8_t *secStart = file->codeSection->Content.data();
285   const uint8_t *funcStart = secStart + getInputSectionOffset();
286   uint32_t functionSizeLength;
287   decodeULEB128(funcStart, &functionSizeLength);
288 
289   uint32_t start = getInputSectionOffset();
290   uint32_t end = start + function->Size;
291 
292   uint32_t lastRelocEnd = start + functionSizeLength;
293   for (const WasmRelocation &rel : relocations) {
294     LLVM_DEBUG(dbgs() << "  region: " << (rel.Offset - lastRelocEnd) << "\n");
295     compressedFuncSize += rel.Offset - lastRelocEnd;
296     compressedFuncSize += getRelocWidth(rel, file->calcNewValue(rel));
297     lastRelocEnd = rel.Offset + getRelocWidthPadded(rel);
298   }
299   LLVM_DEBUG(dbgs() << "  final region: " << (end - lastRelocEnd) << "\n");
300   compressedFuncSize += end - lastRelocEnd;
301 
302   // Now we know how long the resulting function is we can add the encoding
303   // of its length
304   uint8_t buf[5];
305   compressedSize = compressedFuncSize + encodeULEB128(compressedFuncSize, buf);
306 
307   LLVM_DEBUG(dbgs() << "  calculateSize orig: " << function->Size << "\n");
308   LLVM_DEBUG(dbgs() << "  calculateSize  new: " << compressedSize << "\n");
309 }
310 
311 // Override the default writeTo method so that we can (optionally) write the
312 // compressed version of the function.
313 void InputFunction::writeTo(uint8_t *buf) const {
314   if (!file || !config->compressRelocations)
315     return InputChunk::writeTo(buf);
316 
317   buf += outputOffset;
318   uint8_t *orig = buf;
319   (void)orig;
320 
321   const uint8_t *secStart = file->codeSection->Content.data();
322   const uint8_t *funcStart = secStart + getInputSectionOffset();
323   const uint8_t *end = funcStart + function->Size;
324   uint32_t count;
325   decodeULEB128(funcStart, &count);
326   funcStart += count;
327 
328   LLVM_DEBUG(dbgs() << "write func: " << getName() << "\n");
329   buf += encodeULEB128(compressedFuncSize, buf);
330   const uint8_t *lastRelocEnd = funcStart;
331   for (const WasmRelocation &rel : relocations) {
332     unsigned chunkSize = (secStart + rel.Offset) - lastRelocEnd;
333     LLVM_DEBUG(dbgs() << "  write chunk: " << chunkSize << "\n");
334     memcpy(buf, lastRelocEnd, chunkSize);
335     buf += chunkSize;
336     buf += writeCompressedReloc(buf, rel, file->calcNewValue(rel));
337     lastRelocEnd = secStart + rel.Offset + getRelocWidthPadded(rel);
338   }
339 
340   unsigned chunkSize = end - lastRelocEnd;
341   LLVM_DEBUG(dbgs() << "  write final chunk: " << chunkSize << "\n");
342   memcpy(buf, lastRelocEnd, chunkSize);
343   LLVM_DEBUG(dbgs() << "  total: " << (buf + chunkSize - orig) << "\n");
344 }
345 
346 // Generate code to apply relocations to the data section at runtime.
347 // This is only called when generating shared libaries (PIC) where address are
348 // not known at static link time.
349 void InputSegment::generateRelocationCode(raw_ostream &os) const {
350   LLVM_DEBUG(dbgs() << "generating runtime relocations: " << getName()
351                     << " count=" << relocations.size() << "\n");
352 
353   unsigned opcode_ptr_const = config->is64.getValueOr(false)
354                                   ? WASM_OPCODE_I64_CONST
355                                   : WASM_OPCODE_I32_CONST;
356   unsigned opcode_ptr_add = config->is64.getValueOr(false)
357                                 ? WASM_OPCODE_I64_ADD
358                                 : WASM_OPCODE_I32_ADD;
359 
360   // TODO(sbc): Encode the relocations in the data section and write a loop
361   // here to apply them.
362   uint64_t segmentVA = outputSeg->startVA + outputSegmentOffset;
363   for (const WasmRelocation &rel : relocations) {
364     uint64_t offset = rel.Offset - getInputSectionOffset();
365     uint64_t outputOffset = segmentVA + offset;
366 
367     LLVM_DEBUG(dbgs() << "gen reloc: type=" << relocTypeToString(rel.Type)
368                       << " addend=" << rel.Addend << " index=" << rel.Index
369                       << " output offset=" << outputOffset << "\n");
370 
371     // Get __memory_base
372     writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET");
373     writeUleb128(os, WasmSym::memoryBase->getGlobalIndex(), "memory_base");
374 
375     // Add the offset of the relocation
376     writeU8(os, opcode_ptr_const, "CONST");
377     writeSleb128(os, outputOffset, "offset");
378     writeU8(os, opcode_ptr_add, "ADD");
379 
380     bool is64 = relocIs64(rel.Type);
381     unsigned opcode_reloc_const =
382         is64 ? WASM_OPCODE_I64_CONST : WASM_OPCODE_I32_CONST;
383     unsigned opcode_reloc_add =
384         is64 ? WASM_OPCODE_I64_ADD : WASM_OPCODE_I32_ADD;
385     unsigned opcode_reloc_store =
386         is64 ? WASM_OPCODE_I64_STORE : WASM_OPCODE_I32_STORE;
387 
388     Symbol *sym = file->getSymbol(rel);
389     // Now figure out what we want to store
390     if (sym->hasGOTIndex()) {
391       writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET");
392       writeUleb128(os, sym->getGOTIndex(), "global index");
393       if (rel.Addend) {
394         writeU8(os, opcode_reloc_const, "CONST");
395         writeSleb128(os, rel.Addend, "addend");
396         writeU8(os, opcode_reloc_add, "ADD");
397       }
398     } else {
399       const GlobalSymbol* baseSymbol = WasmSym::memoryBase;
400       if (rel.Type == R_WASM_TABLE_INDEX_I32 ||
401           rel.Type == R_WASM_TABLE_INDEX_I64)
402         baseSymbol = WasmSym::tableBase;
403       writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET");
404       writeUleb128(os, baseSymbol->getGlobalIndex(), "base");
405       writeU8(os, opcode_reloc_const, "CONST");
406       writeSleb128(os, file->calcNewValue(rel), "offset");
407       writeU8(os, opcode_reloc_add, "ADD");
408     }
409 
410     // Store that value at the virtual address
411     writeU8(os, opcode_reloc_store, "I32_STORE");
412     writeUleb128(os, 2, "align");
413     writeUleb128(os, 0, "offset");
414   }
415 }
416 
417 } // namespace wasm
418 } // namespace lld
419