Line data Source code
1 : //===-- llvm/CodeGen/DIEHash.cpp - Dwarf Hashing Framework ----------------===//
2 : //
3 : // The LLVM Compiler Infrastructure
4 : //
5 : // This file is distributed under the University of Illinois Open Source
6 : // License. See LICENSE.TXT for details.
7 : //
8 : //===----------------------------------------------------------------------===//
9 : //
10 : // This file contains support for DWARF4 hashing of DIEs.
11 : //
12 : //===----------------------------------------------------------------------===//
13 :
14 : #include "DIEHash.h"
15 : #include "ByteStreamer.h"
16 : #include "DwarfDebug.h"
17 : #include "llvm/ADT/ArrayRef.h"
18 : #include "llvm/ADT/StringRef.h"
19 : #include "llvm/BinaryFormat/Dwarf.h"
20 : #include "llvm/CodeGen/AsmPrinter.h"
21 : #include "llvm/CodeGen/DIE.h"
22 : #include "llvm/Support/Debug.h"
23 : #include "llvm/Support/Endian.h"
24 : #include "llvm/Support/MD5.h"
25 : #include "llvm/Support/raw_ostream.h"
26 :
27 : using namespace llvm;
28 :
29 : #define DEBUG_TYPE "dwarfdebug"
30 :
31 : /// Grabs the string in whichever attribute is passed in and returns
32 : /// a reference to it.
33 192 : static StringRef getDIEStringAttr(const DIE &Die, uint16_t Attr) {
34 : // Iterate through all the attributes until we find the one we're
35 : // looking for, if we can't find it return an empty string.
36 482 : for (const auto &V : Die.values())
37 414 : if (V.getAttribute() == Attr)
38 : return V.getDIEString().getString();
39 :
40 68 : return StringRef("");
41 : }
42 :
43 : /// Adds the string in \p Str to the hash. This also hashes
44 : /// a trailing NULL with the string.
45 285 : void DIEHash::addString(StringRef Str) {
46 : LLVM_DEBUG(dbgs() << "Adding string " << Str << " to hash.\n");
47 285 : Hash.update(Str);
48 285 : Hash.update(makeArrayRef((uint8_t)'\0'));
49 285 : }
50 :
51 : // FIXME: The LEB128 routines are copied and only slightly modified out of
52 : // LEB128.h.
53 :
54 : /// Adds the unsigned in \p Value to the hash encoded as a ULEB128.
55 2035 : void DIEHash::addULEB128(uint64_t Value) {
56 : LLVM_DEBUG(dbgs() << "Adding ULEB128 " << Value << " to hash.\n");
57 : do {
58 2035 : uint8_t Byte = Value & 0x7f;
59 2035 : Value >>= 7;
60 2035 : if (Value != 0)
61 0 : Byte |= 0x80; // Mark this byte to show that more bytes will follow.
62 4070 : Hash.update(Byte);
63 2035 : } while (Value != 0);
64 2035 : }
65 :
66 94 : void DIEHash::addSLEB128(int64_t Value) {
67 : LLVM_DEBUG(dbgs() << "Adding ULEB128 " << Value << " to hash.\n");
68 : bool More;
69 94 : do {
70 94 : uint8_t Byte = Value & 0x7f;
71 94 : Value >>= 7;
72 94 : More = !((((Value == 0) && ((Byte & 0x40) == 0)) ||
73 1 : ((Value == -1) && ((Byte & 0x40) != 0))));
74 : if (More)
75 0 : Byte |= 0x80; // Mark this byte to show that more bytes will follow.
76 188 : Hash.update(Byte);
77 : } while (More);
78 94 : }
79 :
80 : /// Including \p Parent adds the context of Parent to the hash..
81 19 : void DIEHash::addParentContext(const DIE &Parent) {
82 :
83 : LLVM_DEBUG(dbgs() << "Adding parent context to hash...\n");
84 :
85 : // [7.27.2] For each surrounding type or namespace beginning with the
86 : // outermost such construct...
87 : SmallVector<const DIE *, 1> Parents;
88 19 : const DIE *Cur = &Parent;
89 23 : while (Cur->getParent()) {
90 4 : Parents.push_back(Cur);
91 4 : Cur = Cur->getParent();
92 : }
93 : assert(Cur->getTag() == dwarf::DW_TAG_compile_unit ||
94 : Cur->getTag() == dwarf::DW_TAG_type_unit);
95 :
96 : // Reverse iterate over our list to go from the outermost construct to the
97 : // innermost.
98 : for (SmallVectorImpl<const DIE *>::reverse_iterator I = Parents.rbegin(),
99 : E = Parents.rend();
100 23 : I != E; ++I) {
101 4 : const DIE &Die = **I;
102 :
103 : // ... Append the letter "C" to the sequence...
104 4 : addULEB128('C');
105 :
106 : // ... Followed by the DWARF tag of the construct...
107 4 : addULEB128(Die.getTag());
108 :
109 : // ... Then the name, taken from the DW_AT_name attribute.
110 4 : StringRef Name = getDIEStringAttr(Die, dwarf::DW_AT_name);
111 : LLVM_DEBUG(dbgs() << "... adding context: " << Name << "\n");
112 4 : if (!Name.empty())
113 1 : addString(Name);
114 : }
115 19 : }
116 :
117 : // Collect all of the attributes for a particular DIE in single structure.
118 264 : void DIEHash::collectAttributes(const DIE &Die, DIEAttrs &Attrs) {
119 :
120 1137 : for (const auto &V : Die.values()) {
121 : LLVM_DEBUG(dbgs() << "Attribute: "
122 : << dwarf::AttributeString(V.getAttribute())
123 : << " added.\n");
124 873 : switch (V.getAttribute()) {
125 : #define HANDLE_DIE_HASH_ATTR(NAME) \
126 : case dwarf::NAME: \
127 : Attrs.NAME = V; \
128 : break;
129 : #include "DIEHashAttributes.def"
130 : default:
131 : break;
132 : }
133 : }
134 264 : }
135 :
136 26 : void DIEHash::hashShallowTypeReference(dwarf::Attribute Attribute,
137 : const DIE &Entry, StringRef Name) {
138 : // append the letter 'N'
139 26 : addULEB128('N');
140 :
141 : // the DWARF attribute code (DW_AT_type or DW_AT_friend),
142 26 : addULEB128(Attribute);
143 :
144 : // the context of the tag,
145 26 : if (const DIE *Parent = Entry.getParent())
146 18 : addParentContext(*Parent);
147 :
148 : // the letter 'E',
149 26 : addULEB128('E');
150 :
151 : // and the name of the type.
152 26 : addString(Name);
153 :
154 : // Currently DW_TAG_friends are not used by Clang, but if they do become so,
155 : // here's the relevant spec text to implement:
156 : //
157 : // For DW_TAG_friend, if the referenced entry is the DW_TAG_subprogram,
158 : // the context is omitted and the name to be used is the ABI-specific name
159 : // of the subprogram (e.g., the mangled linker name).
160 26 : }
161 :
162 29 : void DIEHash::hashRepeatedTypeReference(dwarf::Attribute Attribute,
163 : unsigned DieNumber) {
164 : // a) If T is in the list of [previously hashed types], use the letter
165 : // 'R' as the marker
166 29 : addULEB128('R');
167 :
168 29 : addULEB128(Attribute);
169 :
170 : // and use the unsigned LEB128 encoding of [the index of T in the
171 : // list] as the attribute value;
172 29 : addULEB128(DieNumber);
173 29 : }
174 :
175 122 : void DIEHash::hashDIEEntry(dwarf::Attribute Attribute, dwarf::Tag Tag,
176 : const DIE &Entry) {
177 : assert(Tag != dwarf::DW_TAG_friend && "No current LLVM clients emit friend "
178 : "tags. Add support here when there's "
179 : "a use case");
180 : // Step 5
181 : // If the tag in Step 3 is one of [the below tags]
182 244 : if ((Tag == dwarf::DW_TAG_pointer_type ||
183 122 : Tag == dwarf::DW_TAG_reference_type ||
184 122 : Tag == dwarf::DW_TAG_rvalue_reference_type ||
185 43 : Tag == dwarf::DW_TAG_ptr_to_member_type) &&
186 : // and the referenced type (via the [below attributes])
187 : // FIXME: This seems overly restrictive, and causes hash mismatches
188 : // there's a decl/def difference in the containing type of a
189 : // ptr_to_member_type, but it's what DWARF says, for some reason.
190 : Attribute == dwarf::DW_AT_type) {
191 : // ... has a DW_AT_name attribute,
192 38 : StringRef Name = getDIEStringAttr(Entry, dwarf::DW_AT_name);
193 38 : if (!Name.empty()) {
194 26 : hashShallowTypeReference(Attribute, Entry, Name);
195 26 : return;
196 : }
197 : }
198 :
199 96 : unsigned &DieNumber = Numbering[&Entry];
200 96 : if (DieNumber) {
201 29 : hashRepeatedTypeReference(Attribute, DieNumber);
202 29 : return;
203 : }
204 :
205 : // otherwise, b) use the letter 'T' as the marker, ...
206 67 : addULEB128('T');
207 :
208 67 : addULEB128(Attribute);
209 :
210 : // ... process the type T recursively by performing Steps 2 through 7, and
211 : // use the result as the attribute value.
212 67 : DieNumber = Numbering.size();
213 67 : computeHash(Entry);
214 : }
215 :
216 : // Hash all of the values in a block like set of values. This assumes that
217 : // all of the data is going to be added as integers.
218 43 : void DIEHash::hashBlockData(const DIE::const_value_range &Values) {
219 137 : for (const auto &V : Values)
220 188 : Hash.update((uint64_t)V.getDIEInteger().getValue());
221 43 : }
222 :
223 : // Hash the contents of a loclistptr class.
224 3 : void DIEHash::hashLocList(const DIELocList &LocList) {
225 : HashingByteStreamer Streamer(*this);
226 3 : DwarfDebug &DD = *AP->getDwarfDebug();
227 : const DebugLocStream &Locs = DD.getDebugLocs();
228 10 : for (const auto &Entry : Locs.getEntries(Locs.getList(LocList.getValue())))
229 4 : DD.emitDebugLocEntry(Streamer, Entry);
230 3 : }
231 :
232 : // Hash an individual attribute \param Attr based on the type of attribute and
233 : // the form.
234 438 : void DIEHash::hashAttribute(const DIEValue &Value, dwarf::Tag Tag) {
235 438 : dwarf::Attribute Attribute = Value.getAttribute();
236 :
237 : // Other attribute values use the letter 'A' as the marker, and the value
238 : // consists of the form code (encoded as an unsigned LEB128 value) followed by
239 : // the encoding of the value according to the form code. To ensure
240 : // reproducibility of the signature, the set of forms used in the signature
241 : // computation is limited to the following: DW_FORM_sdata, DW_FORM_flag,
242 : // DW_FORM_string, and DW_FORM_block.
243 :
244 438 : switch (Value.getType()) {
245 : case DIEValue::isNone:
246 : llvm_unreachable("Expected valid DIEValue");
247 :
248 : // 7.27 Step 3
249 : // ... An attribute that refers to another type entry T is processed as
250 : // follows:
251 : case DIEValue::isEntry:
252 122 : hashDIEEntry(Attribute, Tag, Value.getDIEEntry().getEntry());
253 122 : break;
254 109 : case DIEValue::isInteger: {
255 109 : addULEB128('A');
256 109 : addULEB128(Attribute);
257 109 : switch (Value.getForm()) {
258 94 : case dwarf::DW_FORM_data1:
259 : case dwarf::DW_FORM_data2:
260 : case dwarf::DW_FORM_data4:
261 : case dwarf::DW_FORM_data8:
262 : case dwarf::DW_FORM_udata:
263 : case dwarf::DW_FORM_sdata:
264 94 : addULEB128(dwarf::DW_FORM_sdata);
265 94 : addSLEB128((int64_t)Value.getDIEInteger().getValue());
266 94 : break;
267 : // DW_FORM_flag_present is just flag with a value of one. We still give it a
268 : // value so just use the value.
269 15 : case dwarf::DW_FORM_flag_present:
270 : case dwarf::DW_FORM_flag:
271 15 : addULEB128(dwarf::DW_FORM_flag);
272 15 : addULEB128((int64_t)Value.getDIEInteger().getValue());
273 15 : break;
274 0 : default:
275 0 : llvm_unreachable("Unknown integer form!");
276 : }
277 : break;
278 : }
279 161 : case DIEValue::isString:
280 161 : addULEB128('A');
281 161 : addULEB128(Attribute);
282 161 : addULEB128(dwarf::DW_FORM_string);
283 161 : addString(Value.getDIEString().getString());
284 161 : break;
285 0 : case DIEValue::isInlineString:
286 0 : addULEB128('A');
287 0 : addULEB128(Attribute);
288 0 : addULEB128(dwarf::DW_FORM_string);
289 0 : addString(Value.getDIEInlineString().getString());
290 0 : break;
291 46 : case DIEValue::isBlock:
292 : case DIEValue::isLoc:
293 : case DIEValue::isLocList:
294 46 : addULEB128('A');
295 46 : addULEB128(Attribute);
296 46 : addULEB128(dwarf::DW_FORM_block);
297 46 : if (Value.getType() == DIEValue::isBlock) {
298 1 : addULEB128(Value.getDIEBlock().ComputeSize(AP));
299 1 : hashBlockData(Value.getDIEBlock().values());
300 45 : } else if (Value.getType() == DIEValue::isLoc) {
301 42 : addULEB128(Value.getDIELoc().ComputeSize(AP));
302 42 : hashBlockData(Value.getDIELoc().values());
303 : } else {
304 : // We could add the block length, but that would take
305 : // a bit of work and not add a lot of uniqueness
306 : // to the hash in some way we could test.
307 3 : hashLocList(Value.getDIELocList());
308 : }
309 : break;
310 : // FIXME: It's uncertain whether or not we should handle this at the moment.
311 : case DIEValue::isExpr:
312 : case DIEValue::isLabel:
313 : case DIEValue::isDelta:
314 : llvm_unreachable("Add support for additional value types.");
315 : }
316 438 : }
317 :
318 : // Go through the attributes from \param Attrs in the order specified in 7.27.4
319 : // and hash them.
320 264 : void DIEHash::hashAttributes(const DIEAttrs &Attrs, dwarf::Tag Tag) {
321 : #define HANDLE_DIE_HASH_ATTR(NAME) \
322 : { \
323 : if (Attrs.NAME) \
324 : hashAttribute(Attrs.NAME, Tag); \
325 : }
326 : #include "DIEHashAttributes.def"
327 : // FIXME: Add the extended attributes.
328 264 : }
329 :
330 : // Add all of the attributes for \param Die to the hash.
331 264 : void DIEHash::addAttributes(const DIE &Die) {
332 264 : DIEAttrs Attrs = {};
333 264 : collectAttributes(Die, Attrs);
334 264 : hashAttributes(Attrs, Die.getTag());
335 264 : }
336 :
337 97 : void DIEHash::hashNestedType(const DIE &Die, StringRef Name) {
338 : // 7.27 Step 7
339 : // ... append the letter 'S',
340 97 : addULEB128('S');
341 :
342 : // the tag of C,
343 97 : addULEB128(Die.getTag());
344 :
345 : // and the name.
346 97 : addString(Name);
347 97 : }
348 :
349 : // Compute the hash of a DIE. This is based on the type signature computation
350 : // given in section 7.27 of the DWARF4 standard. It is the md5 hash of a
351 : // flattened description of the DIE.
352 264 : void DIEHash::computeHash(const DIE &Die) {
353 : // Append the letter 'D', followed by the DWARF tag of the DIE.
354 264 : addULEB128('D');
355 264 : addULEB128(Die.getTag());
356 :
357 : // Add each of the attributes of the DIE.
358 264 : addAttributes(Die);
359 :
360 : // Then hash each of the children of the DIE.
361 494 : for (auto &C : Die.children()) {
362 : // 7.27 Step 7
363 : // If C is a nested type entry or a member function entry, ...
364 230 : if (isType(C.getTag()) || C.getTag() == dwarf::DW_TAG_subprogram) {
365 150 : StringRef Name = getDIEStringAttr(C, dwarf::DW_AT_name);
366 : // ... and has a DW_AT_name attribute
367 150 : if (!Name.empty()) {
368 97 : hashNestedType(C, Name);
369 97 : continue;
370 : }
371 : }
372 133 : computeHash(C);
373 : }
374 :
375 : // Following the last (or if there are no children), append a zero byte.
376 264 : Hash.update(makeArrayRef((uint8_t)'\0'));
377 264 : }
378 :
379 : /// This is based on the type signature computation given in section 7.27 of the
380 : /// DWARF4 standard. It is an md5 hash of the flattened description of the DIE
381 : /// with the inclusion of the full CU and all top level CU entities.
382 : // TODO: Initialize the type chain at 0 instead of 1 for CU signatures.
383 43 : uint64_t DIEHash::computeCUSignature(StringRef DWOName, const DIE &Die) {
384 43 : Numbering.clear();
385 43 : Numbering[&Die] = 1;
386 :
387 43 : if (!DWOName.empty())
388 11 : Hash.update(DWOName);
389 : // Hash the DIE.
390 43 : computeHash(Die);
391 :
392 : // Now return the result.
393 : MD5::MD5Result Result;
394 43 : Hash.final(Result);
395 :
396 : // ... take the least significant 8 bytes and return those. Our MD5
397 : // implementation always returns its results in little endian, so we actually
398 : // need the "high" word.
399 43 : return Result.high();
400 : }
401 :
402 : /// This is based on the type signature computation given in section 7.27 of the
403 : /// DWARF4 standard. It is an md5 hash of the flattened description of the DIE
404 : /// with the inclusion of additional forms not specifically called out in the
405 : /// standard.
406 21 : uint64_t DIEHash::computeTypeSignature(const DIE &Die) {
407 21 : Numbering.clear();
408 21 : Numbering[&Die] = 1;
409 :
410 21 : if (const DIE *Parent = Die.getParent())
411 1 : addParentContext(*Parent);
412 :
413 : // Hash the DIE.
414 21 : computeHash(Die);
415 :
416 : // Now return the result.
417 : MD5::MD5Result Result;
418 21 : Hash.final(Result);
419 :
420 : // ... take the least significant 8 bytes and return those. Our MD5
421 : // implementation always returns its results in little endian, so we actually
422 : // need the "high" word.
423 21 : return Result.high();
424 : }
|