13933486559

Committed 18 Mar 2025 09:06PM UTC coverage: 83.197% (-0.4%) from 83.637%

Build # 13933486559

Build Type

push

github

Committed by

web-flow

Commit Message

Refactoring (#35)

Run Details

1027 of 1482 branches covered (69.3%)

Branch coverage included in aggregate %.

86 of 87 new or added lines in 10 files covered. (98.85%)

17 existing lines in 7 files now uncovered.

1924 of 2065 relevant lines covered (93.17%)

3950.95 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

83.26

/lib/type/DITypeExtractor.cpp


//  llvm-dimeta library
//  Copyright (c) 2022-2025 llvm-dimeta authors
//  Distributed under the BSD 3-Clause license.
//  (See accompanying file LICENSE)
//  SPDX-License-Identifier: BSD-3-Clause
//

#include "DIFinder.h"
#include "DIRootType.h"
#include "DIUtil.h"
#include "DataflowAnalysis.h"
#include "DefUseAnalysis.h"
#include "GEP.h"
#include "TBAA.h"
#include "Util.h"
#include "ValuePath.h"
#include "support/Logger.h"

#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/ilist_iterator.h"
#include "llvm/BinaryFormat/Dwarf.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/IR/Argument.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/InstIterator.h"
#include "llvm/IR/InstrTypes.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Metadata.h"
#include "llvm/IR/Operator.h"
#include "llvm/IR/Value.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"

#include <cassert>
#include <iterator>
#include <type_traits>
#include <unordered_map>

namespace dimeta::type {

namespace reset {

using GepToDIMemberMap = std::unordered_map<const llvm::GEPOperator*, llvm::DIDerivedType*>;

namespace detail {

template <typename T, typename InstTy>
std::optional<const T*> get_operand_to(const InstTy* memory_instruction) {
  static_assert(std::is_same_v<InstTy, llvm::StoreInst> || std::is_same_v<InstTy, llvm::LoadInst>,
                "Expected load or store instruction");
  const auto* memory_target = memory_instruction->getPointerOperand();
  if (llvm::isa<T>(memory_target)) {
    return llvm::dyn_cast<T>(memory_target);
  }

  if (auto bcast = llvm::dyn_cast<llvm::BitCastInst>(memory_target)) {
    if (llvm::isa<T>(bcast->getOperand(0))) {
      return llvm::dyn_cast<T>(bcast->getOperand(0));
    }
  }

  return {};
}

}  // namespace detail

template <typename T>
bool store_to(const llvm::StoreInst* store) {
  return detail::get_operand_to<T>(store).has_value();
}

template <typename T>
bool load_to(const llvm::LoadInst* load) {
  return detail::get_operand_to<T>(load).has_value();
}

std::optional<llvm::DIType*> reset_load_related_basic(const dataflow::ValuePath&, llvm::DIType* type_to_reset,
                                                      const llvm::LoadInst* load) {
  auto* type = type_to_reset;

  if (load_to<llvm::GlobalVariable>(load) || load_to<llvm::AllocaInst>(load)) {
    // if (auto* maybe_ptr_to_type = llvm::dyn_cast<llvm::DIDerivedType>(type)) {
    //   if ((maybe_ptr_to_type->getTag() == llvm::dwarf::DW_TAG_pointer_type ||
    //        maybe_ptr_to_type->getTag() == llvm::dwarf::DW_TAG_reference_type)) {
    //     LOG_DEBUG("Load of pointer-like " << log::ditype_str(maybe_ptr_to_type))
    //     if (load->getPointerOperandType()->isOpaquePointerTy()) {
    //       return maybe_ptr_to_type->getBaseType();
    //     }
    //   }
    // }
    LOG_DEBUG("Do not reset DIType based on load to global,alloca")
    return type;
  }

  if (auto* maybe_ptr_to_type = llvm::dyn_cast<llvm::DIDerivedType>(type)) {
    if (di::util::is_pointer(*maybe_ptr_to_type)) {
      LOG_DEBUG("Load of pointer-like " << log::ditype_str(maybe_ptr_to_type))
    }

    auto* base_type = maybe_ptr_to_type->getBaseType();

    if (auto* composite = llvm::dyn_cast<llvm::DICompositeType>(base_type)) {
      LOG_DEBUG("Have ptr to composite " << log::ditype_str(composite))
      auto type_tbaa = tbaa::resolve_tbaa(base_type, *load);
      if (type_tbaa) {
        return type_tbaa;
      }
    }
    return base_type;
  }

  return type;
}

std::optional<llvm::DIType*> reset_store_related_basic(const dataflow::ValuePath&, llvm::DIType* type_to_reset,
                                                       const llvm::StoreInst* store_inst) {
  auto* type = type_to_reset;

  if (store_to<llvm::GlobalVariable>(store_inst) || store_to<llvm::AllocaInst>(store_inst)) {
    // Relevant in "heap_lulesh_mock_char.cpp"
    LOG_DEBUG("Store to alloca/global, return " << log::ditype_str(type))
    return type;
  }

  if (!llvm::isa<llvm::DIDerivedType>(type)) {
    LOG_DEBUG("Store resolved, return " << log::ditype_str(type))
    return type;
  }

  auto* derived_type = llvm::cast<llvm::DIDerivedType>(type);

  if (di::util::is_non_static_member(*derived_type)) {
    auto* member_base               = derived_type->getBaseType();
    const bool is_array_type_member = member_base->getTag() == llvm::dwarf::DW_TAG_array_type;
    // Need to look at base type for array-type member of struct. Tests w.r.t. gep:
    // 1. array_composite.c
    // 2. array_composite_offset_zero.c
    // 3. array_composite_sub_offset_zero.c
    // 4. global_nested.c
    if (is_array_type_member) {
      return llvm::cast<llvm::DICompositeType>(member_base)->getBaseType();
    }
    return member_base;
  }

  if (di::util::is_pointer(*derived_type)) {
    if (auto* may_be_ptr_to_ptr = llvm::dyn_cast<llvm::DIDerivedType>(derived_type->getBaseType())) {
      // Pointer to pointer by default remove one level for RHS assignment type w.r.t. store:
      const auto is_ptr_to_ptr = di::util::is_pointer(*may_be_ptr_to_ptr);
      if (is_ptr_to_ptr) {
        LOG_DEBUG("Store to ptr-ptr, return " << log::ditype_str(may_be_ptr_to_ptr))
        return may_be_ptr_to_ptr;
      }
    }

    if (auto* ptr_to_composite = llvm::dyn_cast<llvm::DICompositeType>(derived_type->getBaseType())) {
      if (store_to<llvm::LoadInst>(store_inst)) {
        // Triggers for "heap_lhs_obj_opt.c" (llvm 14/15)
        auto composite_members = ptr_to_composite->getElements();
        assert(!composite_members.empty() && "Store to composite assumed to be store to first member!");
        // auto store_di_target = llvm::dyn_cast<llvm::DIDerivedType>(composite_members[0])->getBaseType();

        for (auto* member : composite_members) {
          if (di::util::is_non_static_member(*member)) {
            auto* store_di_target = llvm::dyn_cast<llvm::DIDerivedType>(member);
            LOG_DEBUG("Store to a 'load of a composite type', assume first member as target "
                      << log::ditype_str(store_di_target))
            return store_di_target->getBaseType();
          }
        }
      }
    }
  }

  LOG_DEBUG("Store resolved, return " << log::ditype_str(type))
  return type;
}

template <typename Iter, typename Iter2>
std::optional<llvm::DIType*> reset_ditype(llvm::DIType* type_to_reset, const dataflow::ValuePath& path,
                                          const Iter& path_iter, const Iter2&) {
  std::optional<llvm::DIType*> type = type_to_reset;
  if (!type) {
    LOG_DEBUG("No type to reset!")
    return {};
  }

  const auto& next_value = path_iter;
  LOG_DEBUG("Type to reset: " << log::ditype_str(*type));
  LOG_DEBUG(">> based on IR: " << **next_value);

  if (llvm::isa<llvm::GEPOperator>(*next_value)) {
    LOG_DEBUG("Reset based on GEP")
    auto* gep = llvm::cast<llvm::GEPOperator>(*next_value);
    // LOG_DEBUG("Path iter gep for extraction is currently " << *gep);
    // TODO: Maybe we could somehow get more info on the underlying type from the dataflow path
    //       if this returns an empty result due to forward decls?
    const auto gep_result = gep::extract_gep_dereferenced_type(type.value(), *gep);
    if (gep_result.member) {
      LOG_DEBUG("Using gep member type result")
      return gep_result.member;
    }
    return gep_result.type;
  }

  if (const auto* load = llvm::dyn_cast<llvm::LoadInst>(*next_value)) {
    // Re-set the DIType from the gep, if presence of:
    // - a load after a gep is likely the first element of the composite type
    // - a load also resolves to the basetype w.r.t. an array composite
    LOG_DEBUG("Reset based on load")
    return reset::reset_load_related_basic(path, type.value(), load);
  }

  if (const auto* store_inst = llvm::dyn_cast<llvm::StoreInst>(*next_value)) {
    // - a store with a ditype(array) is likely the first element of the array
    LOG_DEBUG("Reset based on store")
    return reset::reset_store_related_basic(path, type.value(), store_inst);
  }

  LOG_DEBUG(">> skipping");

  return type;
}

}  // namespace reset

std::optional<llvm::DIType*> find_type(const dataflow::CallValuePath& call_path) {
  auto type = root::find_type_root(call_path);

  if (!type) {
    LOG_DEBUG("find_type_root failed to find a type for path " << call_path.path)
    return {};
  }

  const auto path_end = call_path.path.path_to_value.rend();
  for (auto path_iter = call_path.path.path_to_value.rbegin(); path_iter != path_end; ++path_iter) {
    if (!type) {
      break;
    }
    LOG_DEBUG("Extracted type w.r.t. gep: " << log::ditype_str(*type));
    type = reset::reset_ditype(type.value(), call_path.path, path_iter, path_end).value_or(type.value());
    LOG_DEBUG("reset_ditype result " << log::ditype_str(type.value_or(nullptr)) << "\n")
  }

  if (type) {
    // If last node is a store inst, try to extract type via TBAA
    const auto* const start_node = llvm::dyn_cast_or_null<llvm::StoreInst>(*call_path.path.start_value());
    if (start_node) {
      auto type_tbaa = tbaa::resolve_tbaa(type.value(), *llvm::dyn_cast<llvm::Instruction>(start_node));
      if (type_tbaa) {
        type = type_tbaa.value();
      }
    }
  }

  return type;
}
}  // namespace dimeta::type

1
2	// llvm-dimeta library
3	// Copyright (c) 2022-2025 llvm-dimeta authors
4	// Distributed under the BSD 3-Clause license.
5	// (See accompanying file LICENSE)
6	// SPDX-License-Identifier: BSD-3-Clause
7	//
8
9	#include "DIFinder.h"
10	#include "DIRootType.h"
11	#include "DIUtil.h"
12	#include "DataflowAnalysis.h"
13	#include "DefUseAnalysis.h"
14	#include "GEP.h"
15	#include "TBAA.h"
16	#include "Util.h"
17	#include "ValuePath.h"
18	#include "support/Logger.h"
19
20	#include "llvm/ADT/ArrayRef.h"
21	#include "llvm/ADT/STLExtras.h"
22	#include "llvm/ADT/SmallVector.h"
23	#include "llvm/ADT/ilist_iterator.h"
24	#include "llvm/BinaryFormat/Dwarf.h"
25	#include "llvm/Config/llvm-config.h"
26	#include "llvm/IR/Argument.h"
27	#include "llvm/IR/Constants.h"
28	#include "llvm/IR/DebugInfoMetadata.h"
29	#include "llvm/IR/Function.h"
30	#include "llvm/IR/GlobalVariable.h"
31	#include "llvm/IR/InstIterator.h"
32	#include "llvm/IR/InstrTypes.h"
33	#include "llvm/IR/Instructions.h"
34	#include "llvm/IR/IntrinsicInst.h"
35	#include "llvm/IR/Metadata.h"
36	#include "llvm/IR/Operator.h"
37	#include "llvm/IR/Value.h"
38	#include "llvm/Support/Casting.h"
39	#include "llvm/Support/Debug.h"
40	#include "llvm/Support/ErrorHandling.h"
41	#include "llvm/Support/raw_ostream.h"
42
43	#include <cassert>
44	#include <iterator>
45	#include <type_traits>
46	#include <unordered_map>
47
48	namespace dimeta::type {
49
50	namespace reset {
51
52	using GepToDIMemberMap = std::unordered_map<const llvm::GEPOperator, llvm::DIDerivedType>;
53
54	namespace detail {
55
56	template <typename T, typename InstTy>
57	std::optional<const T> get_operand_to(const InstTy memory_instruction) {	3,374✔
58	static_assert(std::is_same_v<InstTy, llvm::StoreInst> \|\| std::is_same_v<InstTy, llvm::LoadInst>,
59	"Expected load or store instruction");
60	const auto* memory_target = memory_instruction->getPointerOperand();	3,374✔
61	if (llvm::isa<T>(memory_target)) {	3,374!
62	return llvm::dyn_cast<T>(memory_target);	578✔
63	}
64
65	if (auto bcast = llvm::dyn_cast<llvm::BitCastInst>(memory_target)) {	3,064!
66	if (llvm::isa<T>(bcast->getOperand(0))) {	268!
67	return llvm::dyn_cast<T>(bcast->getOperand(0));	×
68	}
69	}	268✔
70
71	return {};	2,796✔
72	}	3,374✔
73
74	} // namespace detail
75
76	template <typename T>
77	bool store_to(const llvm::StoreInst* store) {	2,092✔
78	return detail::get_operand_to<T>(store).has_value();	2,092✔
79	}
80
81	template <typename T>
82	bool load_to(const llvm::LoadInst* load) {	1,282✔
83	return detail::get_operand_to<T>(load).has_value();	1,282✔
84	}
85
86	std::optional<llvm::DIType> reset_load_related_basic(const dataflow::ValuePath&, llvm::DIType type_to_reset,	686✔
87	const llvm::LoadInst* load) {
88	auto* type = type_to_reset;	686✔
89
90	if (load_to<llvm::GlobalVariable>(load) \|\| load_to<llvm::AllocaInst>(load)) {	686✔
91	// if (auto* maybe_ptr_to_type = llvm::dyn_cast<llvm::DIDerivedType>(type)) {
92	// if ((maybe_ptr_to_type->getTag() == llvm::dwarf::DW_TAG_pointer_type \|\|
93	// maybe_ptr_to_type->getTag() == llvm::dwarf::DW_TAG_reference_type)) {
94	// LOG_DEBUG("Load of pointer-like " << log::ditype_str(maybe_ptr_to_type))
95	// if (load->getPointerOperandType()->isOpaquePointerTy()) {
96	// return maybe_ptr_to_type->getBaseType();
97	// }
98	// }
99	// }
100	LOG_DEBUG("Do not reset DIType based on load to global,alloca")
101	return type;	440✔
102	}
103
104	if (auto* maybe_ptr_to_type = llvm::dyn_cast<llvm::DIDerivedType>(type)) {	492!
105	if (di::util::is_pointer(*maybe_ptr_to_type)) {	246✔
106	LOG_DEBUG("Load of pointer-like " << log::ditype_str(maybe_ptr_to_type))
107	}	192✔
108
109	auto* base_type = maybe_ptr_to_type->getBaseType();	246✔
110
111	if (auto* composite = llvm::dyn_cast<llvm::DICompositeType>(base_type)) {	318✔
112	LOG_DEBUG("Have ptr to composite " << log::ditype_str(composite))
113	auto type_tbaa = tbaa::resolve_tbaa(base_type, *load);	72✔
114	if (type_tbaa) {	72!
115	return type_tbaa;	72✔
116	}
UNCOV 117	}	×
118	return base_type;	174✔
119	}	246✔
120
121	return type;	×
122	}	686✔
123
124	std::optional<llvm::DIType> reset_store_related_basic(const dataflow::ValuePath&, llvm::DIType type_to_reset,	1,014✔
125	const llvm::StoreInst* store_inst) {
126	auto* type = type_to_reset;	1,014✔
127
128	if (store_to<llvm::GlobalVariable>(store_inst) \|\| store_to<llvm::AllocaInst>(store_inst)) {	1,014!
129	// Relevant in "heap_lulesh_mock_char.cpp"
130	LOG_DEBUG("Store to alloca/global, return " << log::ditype_str(type))
131	return type;	124✔
132	}
133
134	if (!llvm::isa<llvm::DIDerivedType>(type)) {	890✔
135	LOG_DEBUG("Store resolved, return " << log::ditype_str(type))
136	return type;	56✔
137	}
138
139	auto* derived_type = llvm::cast<llvm::DIDerivedType>(type);	834✔
140
141	if (di::util::is_non_static_member(*derived_type)) {	834✔
142	auto* member_base = derived_type->getBaseType();	420✔
143	const bool is_array_type_member = member_base->getTag() == llvm::dwarf::DW_TAG_array_type;	420✔
144	// Need to look at base type for array-type member of struct. Tests w.r.t. gep:
145	// 1. array_composite.c
146	// 2. array_composite_offset_zero.c
147	// 3. array_composite_sub_offset_zero.c
148	// 4. global_nested.c
149	if (is_array_type_member) {	420✔
150	return llvm::cast<llvm::DICompositeType>(member_base)->getBaseType();	98✔
151	}
152	return member_base;	322✔
153	}	420✔
154
155	if (di::util::is_pointer(*derived_type)) {	414✔
156	if (auto* may_be_ptr_to_ptr = llvm::dyn_cast<llvm::DIDerivedType>(derived_type->getBaseType())) {	696✔
157	// Pointer to pointer by default remove one level for RHS assignment type w.r.t. store:
158	const auto is_ptr_to_ptr = di::util::is_pointer(*may_be_ptr_to_ptr);	302✔
159	if (is_ptr_to_ptr) {	302✔
160	LOG_DEBUG("Store to ptr-ptr, return " << log::ditype_str(may_be_ptr_to_ptr))
161	return may_be_ptr_to_ptr;	282✔
162	}
163	}	302✔
164
165	if (auto* ptr_to_composite = llvm::dyn_cast<llvm::DICompositeType>(derived_type->getBaseType())) {	176✔
166	if (store_to<llvm::LoadInst>(store_inst)) {	64✔
167	// Triggers for "heap_lhs_obj_opt.c" (llvm 14/15)
168	auto composite_members = ptr_to_composite->getElements();	14✔
169	assert(!composite_members.empty() && "Store to composite assumed to be store to first member!");	28!
170	// auto store_di_target = llvm::dyn_cast<llvm::DIDerivedType>(composite_members[0])->getBaseType();
171
172	for (auto* member : composite_members) {	36!
173	if (di::util::is_non_static_member(*member)) {	22✔
174	auto* store_di_target = llvm::dyn_cast<llvm::DIDerivedType>(member);	14✔
175	LOG_DEBUG("Store to a 'load of a composite type', assume first member as target "
176	<< log::ditype_str(store_di_target))
177	return store_di_target->getBaseType();	14✔
178	}	14✔
179	}	22✔
180	}	14!
181	}	50✔
182	}	98✔
183
184	LOG_DEBUG("Store resolved, return " << log::ditype_str(type))
185	return type;	118✔
186	}	1,014✔
187
188	template <typename Iter, typename Iter2>
189	std::optional<llvm::DIType> reset_ditype(llvm::DIType type_to_reset, const dataflow::ValuePath& path,	4,236✔
190	const Iter& path_iter, const Iter2&) {
191	std::optional<llvm::DIType*> type = type_to_reset;	4,236✔
192	if (!type) {	4,236!
193	LOG_DEBUG("No type to reset!")
194	return {};	×
195	}
196
197	const auto& next_value = path_iter;	4,236✔
198	LOG_DEBUG("Type to reset: " << log::ditype_str(*type));
199	LOG_DEBUG(">> based on IR: " << **next_value);
200
201	if (llvm::isa<llvm::GEPOperator>(*next_value)) {	4,236✔
202	LOG_DEBUG("Reset based on GEP")
203	auto* gep = llvm::cast<llvm::GEPOperator>(*next_value);	1,018✔
204	// LOG_DEBUG("Path iter gep for extraction is currently " << *gep);
205	// TODO: Maybe we could somehow get more info on the underlying type from the dataflow path
206	// if this returns an empty result due to forward decls?
207	const auto gep_result = gep::extract_gep_dereferenced_type(type.value(), *gep);	1,018✔
208	if (gep_result.member) {	1,018✔
209	LOG_DEBUG("Using gep member type result")
210	return gep_result.member;	660✔
211	}
212	return gep_result.type;	358✔
213	}	1,018✔
214
215	if (const auto* load = llvm::dyn_cast<llvm::LoadInst>(*next_value)) {	3,904✔
216	// Re-set the DIType from the gep, if presence of:
217	// - a load after a gep is likely the first element of the composite type
218	// - a load also resolves to the basetype w.r.t. an array composite
219	LOG_DEBUG("Reset based on load")
220	return reset::reset_load_related_basic(path, type.value(), load);	686✔
221	}
222
223	if (const auto* store_inst = llvm::dyn_cast<llvm::StoreInst>(*next_value)) {	3,546✔
224	// - a store with a ditype(array) is likely the first element of the array
225	LOG_DEBUG("Reset based on store")
226	return reset::reset_store_related_basic(path, type.value(), store_inst);	1,014✔
227	}
228
229	LOG_DEBUG(">> skipping");
230
231	return type;	1,518✔
232	}	4,236✔
233
234	} // namespace reset
235
236	std::optional<llvm::DIType*> find_type(const dataflow::CallValuePath& call_path) {	1,174✔
237	auto type = root::find_type_root(call_path);	1,174✔
238
239	if (!type) {	1,174✔
240	LOG_DEBUG("find_type_root failed to find a type for path " << call_path.path)
241	return {};	10✔
242	}
243
244	const auto path_end = call_path.path.path_to_value.rend();	1,164✔
245	for (auto path_iter = call_path.path.path_to_value.rbegin(); path_iter != path_end; ++path_iter) {	5,400✔
246	if (!type) {	4,236!
247	break;	×
248	}
249	LOG_DEBUG("Extracted type w.r.t. gep: " << log::ditype_str(*type));
250	type = reset::reset_ditype(type.value(), call_path.path, path_iter, path_end).value_or(type.value());	4,236✔
251	LOG_DEBUG("reset_ditype result " << log::ditype_str(type.value_or(nullptr)) << "\n")
252	}	4,236✔
253
254	if (type) {	1,164!
255	// If last node is a store inst, try to extract type via TBAA
256	const auto* const start_node = llvm::dyn_cast_or_null<llvm::StoreInst>(*call_path.path.start_value());	1,164✔
257	if (start_node) {	1,164✔
258	auto type_tbaa = tbaa::resolve_tbaa(type.value(), *llvm::dyn_cast<llvm::Instruction>(start_node));	1,014✔
259	if (type_tbaa) {	1,014✔
260	type = type_tbaa.value();	994✔
261	}	994✔
262	}	1,014✔
263	}	1,164✔
264
265	return type;	1,164✔
266	}	1,174✔
267	} // namespace dimeta::type

ahueck / llvm-dimeta / 13933486559

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous