DIEHash.cpp revision 47f66d5a756d4eaeba9082ff4c82023a213daf45 
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#define DEBUG_TYPE "dwarfdebug"
15
16#include "DIE.h"
17#include "DIEHash.h"
18#include "DwarfCompileUnit.h"
19#include "llvm/ADT/ArrayRef.h"
20#include "llvm/ADT/StringRef.h"
21#include "llvm/Support/Debug.h"
22#include "llvm/Support/Dwarf.h"
23#include "llvm/Support/Endian.h"
24#include "llvm/Support/MD5.h"
25#include "llvm/Support/raw_ostream.h"
26
27using namespace llvm;
28
29/// \brief Grabs the string in whichever attribute is passed in and returns
30/// a reference to it.
31static StringRef getDIEStringAttr(DIE *Die, uint16_t Attr) {
32  const SmallVectorImpl<DIEValue *> &Values = Die->getValues();
33  const DIEAbbrev &Abbrevs = Die->getAbbrev();
34
35  // Iterate through all the attributes until we find the one we're
36  // looking for, if we can't find it return an empty string.
37  for (size_t i = 0; i < Values.size(); ++i) {
38    if (Abbrevs.getData()[i].getAttribute() == Attr) {
39      DIEValue *V = Values[i];
40      assert(isa<DIEString>(V) && "String requested. Not a string.");
41      DIEString *S = cast<DIEString>(V);
42      return S->getString();
43    }
44  }
45  return StringRef("");
46}
47
48/// \brief Adds the string in \p Str to the hash. This also hashes
49/// a trailing NULL with the string.
50void DIEHash::addString(StringRef Str) {
51  DEBUG(dbgs() << "Adding string " << Str << " to hash.\n");
52  Hash.update(Str);
53  Hash.update(makeArrayRef((uint8_t)'\0'));
54}
55
56// FIXME: The LEB128 routines are copied and only slightly modified out of
57// LEB128.h.
58
59/// \brief Adds the unsigned in \p Value to the hash encoded as a ULEB128.
60void DIEHash::addULEB128(uint64_t Value) {
61  DEBUG(dbgs() << "Adding ULEB128 " << Value << " to hash.\n");
62  do {
63    uint8_t Byte = Value & 0x7f;
64    Value >>= 7;
65    if (Value != 0)
66      Byte |= 0x80; // Mark this byte to show that more bytes will follow.
67    Hash.update(Byte);
68  } while (Value != 0);
69}
70
71void DIEHash::addSLEB128(int64_t Value) {
72  DEBUG(dbgs() << "Adding ULEB128 " << Value << " to hash.\n");
73  bool More;
74  do {
75    uint8_t Byte = Value & 0x7f;
76    Value >>= 7;
77    More = !((((Value == 0 ) && ((Byte & 0x40) == 0)) ||
78              ((Value == -1) && ((Byte & 0x40) != 0))));
79    if (More)
80      Byte |= 0x80; // Mark this byte to show that more bytes will follow.
81    Hash.update(Byte);
82  } while (More);
83}
84
85/// \brief Including \p Parent adds the context of Parent to the hash..
86void DIEHash::addParentContext(DIE *Parent) {
87
88  DEBUG(dbgs() << "Adding parent context to hash...\n");
89
90  // [7.27.2] For each surrounding type or namespace beginning with the
91  // outermost such construct...
92  SmallVector<DIE *, 1> Parents;
93  while (Parent->getTag() != dwarf::DW_TAG_compile_unit) {
94    Parents.push_back(Parent);
95    Parent = Parent->getParent();
96  }
97
98  // Reverse iterate over our list to go from the outermost construct to the
99  // innermost.
100  for (SmallVectorImpl<DIE *>::reverse_iterator I = Parents.rbegin(),
101                                                E = Parents.rend();
102       I != E; ++I) {
103    DIE *Die = *I;
104
105    // ... Append the letter "C" to the sequence...
106    addULEB128('C');
107
108    // ... Followed by the DWARF tag of the construct...
109    addULEB128(Die->getTag());
110
111    // ... Then the name, taken from the DW_AT_name attribute.
112    StringRef Name = getDIEStringAttr(Die, dwarf::DW_AT_name);
113    DEBUG(dbgs() << "... adding context: " << Name << "\n");
114    if (!Name.empty())
115      addString(Name);
116  }
117}
118
119// Collect all of the attributes for a particular DIE in single structure.
120void DIEHash::collectAttributes(DIE *Die, DIEAttrs &Attrs) {
121  const SmallVectorImpl<DIEValue *> &Values = Die->getValues();
122  const DIEAbbrev &Abbrevs = Die->getAbbrev();
123
124#define COLLECT_ATTR(NAME)                                                     \
125  Attrs.NAME.Val = Values[i];                                                  \
126  Attrs.NAME.Desc = &Abbrevs.getData()[i];
127
128  for (size_t i = 0, e = Values.size(); i != e; ++i) {
129    DEBUG(dbgs() << "Attribute: "
130                 << dwarf::AttributeString(Abbrevs.getData()[i].getAttribute())
131                 << " added.\n");
132    switch (Abbrevs.getData()[i].getAttribute()) {
133    case dwarf::DW_AT_name:
134      COLLECT_ATTR(DW_AT_name);
135      break;
136    case dwarf::DW_AT_accessibility:
137      COLLECT_ATTR(DW_AT_accessibility)
138      break;
139    case dwarf::DW_AT_address_class:
140      COLLECT_ATTR(DW_AT_address_class)
141      break;
142    case dwarf::DW_AT_allocated:
143      COLLECT_ATTR(DW_AT_allocated)
144      break;
145    case dwarf::DW_AT_artificial:
146      COLLECT_ATTR(DW_AT_artificial)
147      break;
148    case dwarf::DW_AT_associated:
149      COLLECT_ATTR(DW_AT_associated)
150      break;
151    case dwarf::DW_AT_binary_scale:
152      COLLECT_ATTR(DW_AT_binary_scale)
153      break;
154    case dwarf::DW_AT_bit_offset:
155      COLLECT_ATTR(DW_AT_bit_offset)
156      break;
157    case dwarf::DW_AT_bit_size:
158      COLLECT_ATTR(DW_AT_bit_size)
159      break;
160    case dwarf::DW_AT_bit_stride:
161      COLLECT_ATTR(DW_AT_bit_stride)
162      break;
163    case dwarf::DW_AT_byte_size:
164      COLLECT_ATTR(DW_AT_byte_size)
165      break;
166    case dwarf::DW_AT_byte_stride:
167      COLLECT_ATTR(DW_AT_byte_stride)
168      break;
169    case dwarf::DW_AT_const_expr:
170      COLLECT_ATTR(DW_AT_const_expr)
171      break;
172    case dwarf::DW_AT_const_value:
173      COLLECT_ATTR(DW_AT_const_value)
174      break;
175    case dwarf::DW_AT_containing_type:
176      COLLECT_ATTR(DW_AT_containing_type)
177      break;
178    case dwarf::DW_AT_count:
179      COLLECT_ATTR(DW_AT_count)
180      break;
181    case dwarf::DW_AT_data_bit_offset:
182      COLLECT_ATTR(DW_AT_data_bit_offset)
183      break;
184    case dwarf::DW_AT_data_location:
185      COLLECT_ATTR(DW_AT_data_location)
186      break;
187    case dwarf::DW_AT_data_member_location:
188      COLLECT_ATTR(DW_AT_data_member_location)
189      break;
190    case dwarf::DW_AT_decimal_scale:
191      COLLECT_ATTR(DW_AT_decimal_scale)
192      break;
193    case dwarf::DW_AT_decimal_sign:
194      COLLECT_ATTR(DW_AT_decimal_sign)
195      break;
196    case dwarf::DW_AT_default_value:
197      COLLECT_ATTR(DW_AT_default_value)
198      break;
199    case dwarf::DW_AT_digit_count:
200      COLLECT_ATTR(DW_AT_digit_count)
201      break;
202    case dwarf::DW_AT_discr:
203      COLLECT_ATTR(DW_AT_discr)
204      break;
205    case dwarf::DW_AT_discr_list:
206      COLLECT_ATTR(DW_AT_discr_list)
207      break;
208    case dwarf::DW_AT_discr_value:
209      COLLECT_ATTR(DW_AT_discr_value)
210      break;
211    case dwarf::DW_AT_encoding:
212      COLLECT_ATTR(DW_AT_encoding)
213      break;
214    case dwarf::DW_AT_enum_class:
215      COLLECT_ATTR(DW_AT_enum_class)
216      break;
217    case dwarf::DW_AT_endianity:
218      COLLECT_ATTR(DW_AT_endianity)
219      break;
220    case dwarf::DW_AT_explicit:
221      COLLECT_ATTR(DW_AT_explicit)
222      break;
223    case dwarf::DW_AT_is_optional:
224      COLLECT_ATTR(DW_AT_is_optional)
225      break;
226    case dwarf::DW_AT_location:
227      COLLECT_ATTR(DW_AT_location)
228      break;
229    case dwarf::DW_AT_lower_bound:
230      COLLECT_ATTR(DW_AT_lower_bound)
231      break;
232    case dwarf::DW_AT_mutable:
233      COLLECT_ATTR(DW_AT_mutable)
234      break;
235    case dwarf::DW_AT_ordering:
236      COLLECT_ATTR(DW_AT_ordering)
237      break;
238    case dwarf::DW_AT_picture_string:
239      COLLECT_ATTR(DW_AT_picture_string)
240      break;
241    case dwarf::DW_AT_prototyped:
242      COLLECT_ATTR(DW_AT_prototyped)
243      break;
244    case dwarf::DW_AT_small:
245      COLLECT_ATTR(DW_AT_small)
246      break;
247    case dwarf::DW_AT_segment:
248      COLLECT_ATTR(DW_AT_segment)
249      break;
250    case dwarf::DW_AT_string_length:
251      COLLECT_ATTR(DW_AT_string_length)
252      break;
253    case dwarf::DW_AT_threads_scaled:
254      COLLECT_ATTR(DW_AT_threads_scaled)
255      break;
256    case dwarf::DW_AT_upper_bound:
257      COLLECT_ATTR(DW_AT_upper_bound)
258      break;
259    case dwarf::DW_AT_use_location:
260      COLLECT_ATTR(DW_AT_use_location)
261      break;
262    case dwarf::DW_AT_use_UTF8:
263      COLLECT_ATTR(DW_AT_use_UTF8)
264      break;
265    case dwarf::DW_AT_variable_parameter:
266      COLLECT_ATTR(DW_AT_variable_parameter)
267      break;
268    case dwarf::DW_AT_virtuality:
269      COLLECT_ATTR(DW_AT_virtuality)
270      break;
271    case dwarf::DW_AT_visibility:
272      COLLECT_ATTR(DW_AT_visibility)
273      break;
274    case dwarf::DW_AT_vtable_elem_location:
275      COLLECT_ATTR(DW_AT_vtable_elem_location)
276      break;
277    case dwarf::DW_AT_type:
278      COLLECT_ATTR(DW_AT_type)
279      break;
280    default:
281      break;
282    }
283  }
284}
285
286// Hash an individual attribute \param Attr based on the type of attribute and
287// the form.
288void DIEHash::hashAttribute(AttrEntry Attr) {
289  const DIEValue *Value = Attr.Val;
290  const DIEAbbrevData *Desc = Attr.Desc;
291
292  // 7.27s3
293  // ... An attribute that refers to another type entry T is processed as
294  // follows:
295  // a) If T is in the list of [previously hashed types], use the letter 'R' as
296  // the marker and use the unsigned LEB128 encoding of [the index of T in the
297  // list] as the attribute value; otherwise,
298
299  // [TODO: implement clause (a)]
300
301  if (const DIEEntry *EntryAttr = dyn_cast<DIEEntry>(Value)) {
302    DIE *Entry = EntryAttr->getEntry();
303
304    // b) use the letter 'T' as a the marker, ...
305    addULEB128('T');
306
307    addULEB128(Desc->getAttribute());
308
309    // ... process the type T recursively by performing Steps 2 through 7, and
310    // use the result as the attribute value.
311    computeHash(Entry);
312    return;
313  }
314
315  // Other attribute values use the letter 'A' as the marker, ...
316  addULEB128('A');
317
318  addULEB128(Desc->getAttribute());
319
320  // ... and the value consists of the form code (encoded as an unsigned LEB128
321  // value) followed by the encoding of the value according to the form code. To
322  // ensure reproducibility of the signature, the set of forms used in the
323  // signature computation is limited to the following: DW_FORM_sdata,
324  // DW_FORM_flag, DW_FORM_string, and DW_FORM_block.
325  switch (Desc->getForm()) {
326  case dwarf::DW_FORM_string:
327    llvm_unreachable(
328        "Add support for DW_FORM_string if we ever start emitting them again");
329  case dwarf::DW_FORM_strp:
330    addULEB128(dwarf::DW_FORM_string);
331    addString(cast<DIEString>(Value)->getString());
332    break;
333  case dwarf::DW_FORM_data1:
334  case dwarf::DW_FORM_data2:
335  case dwarf::DW_FORM_data4:
336  case dwarf::DW_FORM_data8:
337  case dwarf::DW_FORM_udata:
338    addULEB128(dwarf::DW_FORM_sdata);
339    addSLEB128((int64_t)cast<DIEInteger>(Value)->getValue());
340    break;
341  // TODO: Add support for additional forms.
342  }
343}
344
345// Go through the attributes from \param Attrs in the order specified in 7.27.4
346// and hash them.
347void DIEHash::hashAttributes(const DIEAttrs &Attrs) {
348#define ADD_ATTR(ATTR)                                                         \
349  {                                                                            \
350    if (ATTR.Val != 0)                                                         \
351      hashAttribute(ATTR);                                                     \
352  }
353
354  ADD_ATTR(Attrs.DW_AT_name);
355  ADD_ATTR(Attrs.DW_AT_accessibility);
356  ADD_ATTR(Attrs.DW_AT_address_class);
357  ADD_ATTR(Attrs.DW_AT_allocated);
358  ADD_ATTR(Attrs.DW_AT_artificial);
359  ADD_ATTR(Attrs.DW_AT_associated);
360  ADD_ATTR(Attrs.DW_AT_binary_scale);
361  ADD_ATTR(Attrs.DW_AT_bit_offset);
362  ADD_ATTR(Attrs.DW_AT_bit_size);
363  ADD_ATTR(Attrs.DW_AT_bit_stride);
364  ADD_ATTR(Attrs.DW_AT_byte_size);
365  ADD_ATTR(Attrs.DW_AT_byte_stride);
366  ADD_ATTR(Attrs.DW_AT_const_expr);
367  ADD_ATTR(Attrs.DW_AT_const_value);
368  ADD_ATTR(Attrs.DW_AT_containing_type);
369  ADD_ATTR(Attrs.DW_AT_count);
370  ADD_ATTR(Attrs.DW_AT_data_bit_offset);
371  ADD_ATTR(Attrs.DW_AT_data_location);
372  ADD_ATTR(Attrs.DW_AT_data_member_location);
373  ADD_ATTR(Attrs.DW_AT_decimal_scale);
374  ADD_ATTR(Attrs.DW_AT_decimal_sign);
375  ADD_ATTR(Attrs.DW_AT_default_value);
376  ADD_ATTR(Attrs.DW_AT_digit_count);
377  ADD_ATTR(Attrs.DW_AT_discr);
378  ADD_ATTR(Attrs.DW_AT_discr_list);
379  ADD_ATTR(Attrs.DW_AT_discr_value);
380  ADD_ATTR(Attrs.DW_AT_encoding);
381  ADD_ATTR(Attrs.DW_AT_enum_class);
382  ADD_ATTR(Attrs.DW_AT_endianity);
383  ADD_ATTR(Attrs.DW_AT_explicit);
384  ADD_ATTR(Attrs.DW_AT_is_optional);
385  ADD_ATTR(Attrs.DW_AT_location);
386  ADD_ATTR(Attrs.DW_AT_lower_bound);
387  ADD_ATTR(Attrs.DW_AT_mutable);
388  ADD_ATTR(Attrs.DW_AT_ordering);
389  ADD_ATTR(Attrs.DW_AT_picture_string);
390  ADD_ATTR(Attrs.DW_AT_prototyped);
391  ADD_ATTR(Attrs.DW_AT_small);
392  ADD_ATTR(Attrs.DW_AT_segment);
393  ADD_ATTR(Attrs.DW_AT_string_length);
394  ADD_ATTR(Attrs.DW_AT_threads_scaled);
395  ADD_ATTR(Attrs.DW_AT_upper_bound);
396  ADD_ATTR(Attrs.DW_AT_use_location);
397  ADD_ATTR(Attrs.DW_AT_use_UTF8);
398  ADD_ATTR(Attrs.DW_AT_variable_parameter);
399  ADD_ATTR(Attrs.DW_AT_virtuality);
400  ADD_ATTR(Attrs.DW_AT_visibility);
401  ADD_ATTR(Attrs.DW_AT_vtable_elem_location);
402  ADD_ATTR(Attrs.DW_AT_type);
403
404  // FIXME: Add the extended attributes.
405}
406
407// Add all of the attributes for \param Die to the hash.
408void DIEHash::addAttributes(DIE *Die) {
409  DIEAttrs Attrs = {};
410  collectAttributes(Die, Attrs);
411  hashAttributes(Attrs);
412}
413
414// Compute the hash of a DIE. This is based on the type signature computation
415// given in section 7.27 of the DWARF4 standard. It is the md5 hash of a
416// flattened description of the DIE.
417void DIEHash::computeHash(DIE *Die) {
418
419  // Append the letter 'D', followed by the DWARF tag of the DIE.
420  addULEB128('D');
421  addULEB128(Die->getTag());
422
423  // Add each of the attributes of the DIE.
424  addAttributes(Die);
425
426  // Then hash each of the children of the DIE.
427  for (std::vector<DIE *>::const_iterator I = Die->getChildren().begin(),
428                                          E = Die->getChildren().end();
429       I != E; ++I)
430    computeHash(*I);
431
432  // Following the last (or if there are no children), append a zero byte.
433  Hash.update(makeArrayRef((uint8_t)'\0'));
434}
435
436/// This is based on the type signature computation given in section 7.27 of the
437/// DWARF4 standard. It is the md5 hash of a flattened description of the DIE
438/// with the exception that we are hashing only the context and the name of the
439/// type.
440uint64_t DIEHash::computeDIEODRSignature(DIE *Die) {
441
442  // Add the contexts to the hash. We won't be computing the ODR hash for
443  // function local types so it's safe to use the generic context hashing
444  // algorithm here.
445  // FIXME: If we figure out how to account for linkage in some way we could
446  // actually do this with a slight modification to the parent hash algorithm.
447  DIE *Parent = Die->getParent();
448  if (Parent)
449    addParentContext(Parent);
450
451  // Add the current DIE information.
452
453  // Add the DWARF tag of the DIE.
454  addULEB128(Die->getTag());
455
456  // Add the name of the type to the hash.
457  addString(getDIEStringAttr(Die, dwarf::DW_AT_name));
458
459  // Now get the result.
460  MD5::MD5Result Result;
461  Hash.final(Result);
462
463  // ... take the least significant 8 bytes and return those. Our MD5
464  // implementation always returns its results in little endian, swap bytes
465  // appropriately.
466  return *reinterpret_cast<support::ulittle64_t *>(Result + 8);
467}
468
469/// This is based on the type signature computation given in section 7.27 of the
470/// DWARF4 standard. It is an md5 hash of the flattened description of the DIE
471/// with the inclusion of the full CU and all top level CU entities.
472// TODO: Initialize the type chain at 0 instead of 1 for CU signatures.
473uint64_t DIEHash::computeCUSignature(DIE *Die) {
474
475  // Hash the DIE.
476  computeHash(Die);
477
478  // Now return the result.
479  MD5::MD5Result Result;
480  Hash.final(Result);
481
482  // ... take the least significant 8 bytes and return those. Our MD5
483  // implementation always returns its results in little endian, swap bytes
484  // appropriately.
485  return *reinterpret_cast<support::ulittle64_t *>(Result + 8);
486}
487
488/// This is based on the type signature computation given in section 7.27 of the
489/// DWARF4 standard. It is an md5 hash of the flattened description of the DIE
490/// with the inclusion of additional forms not specifically called out in the
491/// standard.
492uint64_t DIEHash::computeTypeSignature(DIE *Die) {
493
494  if (DIE *Parent = Die->getParent())
495    addParentContext(Parent);
496
497  // Hash the DIE.
498  computeHash(Die);
499
500  // Now return the result.
501  MD5::MD5Result Result;
502  Hash.final(Result);
503
504  // ... take the least significant 8 bytes and return those. Our MD5
505  // implementation always returns its results in little endian, swap bytes
506  // appropriately.
507  return *reinterpret_cast<support::ulittle64_t *>(Result + 8);
508}
509