/* * Copyright (C) 2016 The Android Open Source Project * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include "ti_heap.h" #include #include #include "android-base/logging.h" #include "android-base/thread_annotations.h" #include "arch/context.h" #include "art_field-inl.h" #include "art_jvmti.h" #include "base/logging.h" #include "base/macros.h" #include "base/mutex.h" #include "base/utils.h" #include "class_linker.h" #include "deopt_manager.h" #include "dex/primitive.h" #include "events-inl.h" #include "gc/collector_type.h" #include "gc/gc_cause.h" #include "gc/heap-visit-objects-inl.h" #include "gc/heap-inl.h" #include "gc/scoped_gc_critical_section.h" #include "gc_root-inl.h" #include "handle.h" #include "handle_scope.h" #include "java_frame_root_info.h" #include "jni/jni_env_ext.h" #include "jni/jni_id_manager.h" #include "jni/jni_internal.h" #include "jvmti_weak_table-inl.h" #include "mirror/array-inl.h" #include "mirror/array.h" #include "mirror/class.h" #include "mirror/object-inl.h" #include "mirror/object-refvisitor-inl.h" #include "mirror/object_array-inl.h" #include "mirror/object_array-alloc-inl.h" #include "mirror/object_reference.h" #include "obj_ptr-inl.h" #include "object_callbacks.h" #include "object_tagging.h" #include "offsets.h" #include "read_barrier.h" #include "runtime.h" #include "scoped_thread_state_change-inl.h" #include "stack.h" #include "thread-inl.h" #include "thread_list.h" #include "ti_logging.h" #include "ti_stack.h" #include "ti_thread.h" #include "well_known_classes.h" namespace openjdkjvmti { EventHandler* HeapExtensions::gEventHandler = nullptr; namespace { struct IndexCache { // The number of interface fields implemented by the class. This is a prefix to all assigned // field indices. size_t interface_fields; // It would be nice to also cache the following, but it is complicated to wire up into the // generic visit: // The number of fields in interfaces and superclasses. This is the first index assigned to // fields of the class. // size_t superclass_fields; }; using IndexCachingTable = JvmtiWeakTable; static IndexCachingTable gIndexCachingTable; // Report the contents of a string, if a callback is set. jint ReportString(art::ObjPtr obj, jvmtiEnv* env, ObjectTagTable* tag_table, const jvmtiHeapCallbacks* cb, const void* user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (UNLIKELY(cb->string_primitive_value_callback != nullptr) && obj->IsString()) { art::ObjPtr str = obj->AsString(); int32_t string_length = str->GetLength(); JvmtiUniquePtr data; if (string_length > 0) { jvmtiError alloc_error; data = AllocJvmtiUniquePtr(env, string_length, &alloc_error); if (data == nullptr) { // TODO: Not really sure what to do here. Should we abort the iteration and go all the way // back? For now just warn. LOG(WARNING) << "Unable to allocate buffer for string reporting! Silently dropping value." << " >" << str->ToModifiedUtf8() << "<"; return 0; } if (str->IsCompressed()) { uint8_t* compressed_data = str->GetValueCompressed(); for (int32_t i = 0; i != string_length; ++i) { data[i] = compressed_data[i]; } } else { // Can copy directly. memcpy(data.get(), str->GetValue(), string_length * sizeof(uint16_t)); } } const jlong class_tag = tag_table->GetTagOrZero(obj->GetClass()); jlong string_tag = tag_table->GetTagOrZero(obj.Ptr()); const jlong saved_string_tag = string_tag; jint result = cb->string_primitive_value_callback(class_tag, obj->SizeOf(), &string_tag, data.get(), string_length, const_cast(user_data)); if (string_tag != saved_string_tag) { tag_table->Set(obj.Ptr(), string_tag); } return result; } return 0; } // Report the contents of a primitive array, if a callback is set. jint ReportPrimitiveArray(art::ObjPtr obj, jvmtiEnv* env, ObjectTagTable* tag_table, const jvmtiHeapCallbacks* cb, const void* user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (UNLIKELY(cb->array_primitive_value_callback != nullptr) && obj->IsArrayInstance() && !obj->IsObjectArray()) { art::ObjPtr array = obj->AsArray(); int32_t array_length = array->GetLength(); size_t component_size = array->GetClass()->GetComponentSize(); art::Primitive::Type art_prim_type = array->GetClass()->GetComponentType()->GetPrimitiveType(); jvmtiPrimitiveType prim_type = static_cast(art::Primitive::Descriptor(art_prim_type)[0]); DCHECK(prim_type == JVMTI_PRIMITIVE_TYPE_BOOLEAN || prim_type == JVMTI_PRIMITIVE_TYPE_BYTE || prim_type == JVMTI_PRIMITIVE_TYPE_CHAR || prim_type == JVMTI_PRIMITIVE_TYPE_SHORT || prim_type == JVMTI_PRIMITIVE_TYPE_INT || prim_type == JVMTI_PRIMITIVE_TYPE_LONG || prim_type == JVMTI_PRIMITIVE_TYPE_FLOAT || prim_type == JVMTI_PRIMITIVE_TYPE_DOUBLE); const jlong class_tag = tag_table->GetTagOrZero(obj->GetClass()); jlong array_tag = tag_table->GetTagOrZero(obj.Ptr()); const jlong saved_array_tag = array_tag; jint result; if (array_length == 0) { result = cb->array_primitive_value_callback(class_tag, obj->SizeOf(), &array_tag, 0, prim_type, nullptr, const_cast(user_data)); } else { jvmtiError alloc_error; JvmtiUniquePtr data = AllocJvmtiUniquePtr(env, array_length * component_size, &alloc_error); if (data == nullptr) { // TODO: Not really sure what to do here. Should we abort the iteration and go all the way // back? For now just warn. LOG(WARNING) << "Unable to allocate buffer for array reporting! Silently dropping value."; return 0; } memcpy(data.get(), array->GetRawData(component_size, 0), array_length * component_size); result = cb->array_primitive_value_callback(class_tag, obj->SizeOf(), &array_tag, array_length, prim_type, data.get(), const_cast(user_data)); } if (array_tag != saved_array_tag) { tag_table->Set(obj.Ptr(), array_tag); } return result; } return 0; } template bool VisitorFalse(art::ObjPtr obj ATTRIBUTE_UNUSED, art::ObjPtr klass ATTRIBUTE_UNUSED, art::ArtField& field ATTRIBUTE_UNUSED, size_t field_index ATTRIBUTE_UNUSED, UserData* user_data ATTRIBUTE_UNUSED) { return false; } template class FieldVisitor { public: // Report the contents of a primitive fields of the given object, if a callback is set. template static bool ReportFields(art::ObjPtr obj, UserData* user_data, StaticPrimitiveVisitor& static_prim_visitor, StaticReferenceVisitor& static_ref_visitor, InstancePrimitiveVisitor& instance_prim_visitor, InstanceReferenceVisitor& instance_ref_visitor) REQUIRES_SHARED(art::Locks::mutator_lock_) { FieldVisitor fv(user_data); if (obj->IsClass()) { // When visiting a class, we only visit the static fields of the given class. No field of // superclasses is visited. art::ObjPtr klass = obj->AsClass(); // Only report fields on resolved classes. We need valid field data. if (!klass->IsResolved()) { return false; } return fv.ReportFieldsImpl(nullptr, obj->AsClass(), obj->AsClass()->IsInterface(), static_prim_visitor, static_ref_visitor, instance_prim_visitor, instance_ref_visitor); } else { // See comment above. Just double-checking here, but an instance *should* mean the class was // resolved. DCHECK(obj->GetClass()->IsResolved() || obj->GetClass()->IsErroneousResolved()); return fv.ReportFieldsImpl(obj, obj->GetClass(), false, static_prim_visitor, static_ref_visitor, instance_prim_visitor, instance_ref_visitor); } } private: explicit FieldVisitor(UserData* user_data) : user_data_(user_data) {} // Report the contents of fields of the given object. If obj is null, report the static fields, // otherwise the instance fields. template bool ReportFieldsImpl(art::ObjPtr obj, art::ObjPtr klass, bool skip_java_lang_object, StaticPrimitiveVisitor& static_prim_visitor, StaticReferenceVisitor& static_ref_visitor, InstancePrimitiveVisitor& instance_prim_visitor, InstanceReferenceVisitor& instance_ref_visitor) REQUIRES_SHARED(art::Locks::mutator_lock_) { // Compute the offset of field indices. size_t interface_field_count = CountInterfaceFields(klass); size_t tmp; bool aborted = ReportFieldsRecursive(obj, klass, interface_field_count, skip_java_lang_object, static_prim_visitor, static_ref_visitor, instance_prim_visitor, instance_ref_visitor, &tmp); return aborted; } // Visit primitive fields in an object (instance). Return true if the visit was aborted. template bool ReportFieldsRecursive(art::ObjPtr obj, art::ObjPtr klass, size_t interface_fields, bool skip_java_lang_object, StaticPrimitiveVisitor& static_prim_visitor, StaticReferenceVisitor& static_ref_visitor, InstancePrimitiveVisitor& instance_prim_visitor, InstanceReferenceVisitor& instance_ref_visitor, size_t* field_index_out) REQUIRES_SHARED(art::Locks::mutator_lock_) { DCHECK(klass != nullptr); size_t field_index; if (klass->GetSuperClass() == nullptr) { // j.l.Object. Start with the fields from interfaces. field_index = interface_fields; if (skip_java_lang_object) { *field_index_out = field_index; return false; } } else { // Report superclass fields. if (kCallVisitorOnRecursion) { if (ReportFieldsRecursive(obj, klass->GetSuperClass(), interface_fields, skip_java_lang_object, static_prim_visitor, static_ref_visitor, instance_prim_visitor, instance_ref_visitor, &field_index)) { return true; } } else { // Still call, but with empty visitor. This is required for correct counting. ReportFieldsRecursive(obj, klass->GetSuperClass(), interface_fields, skip_java_lang_object, VisitorFalse, VisitorFalse, VisitorFalse, VisitorFalse, &field_index); } } // Now visit fields for the current klass. for (auto& static_field : klass->GetSFields()) { if (static_field.IsPrimitiveType()) { if (static_prim_visitor(obj, klass, static_field, field_index, user_data_)) { return true; } } else { if (static_ref_visitor(obj, klass, static_field, field_index, user_data_)) { return true; } } field_index++; } for (auto& instance_field : klass->GetIFields()) { if (instance_field.IsPrimitiveType()) { if (instance_prim_visitor(obj, klass, instance_field, field_index, user_data_)) { return true; } } else { if (instance_ref_visitor(obj, klass, instance_field, field_index, user_data_)) { return true; } } field_index++; } *field_index_out = field_index; return false; } // Implements a visit of the implemented interfaces of a given class. template struct RecursiveInterfaceVisit { static void VisitStatic(art::Thread* self, art::ObjPtr klass, T& visitor) REQUIRES_SHARED(art::Locks::mutator_lock_) { RecursiveInterfaceVisit rv; rv.Visit(self, klass, visitor); } void Visit(art::Thread* self, art::ObjPtr klass, T& visitor) REQUIRES_SHARED(art::Locks::mutator_lock_) { // First visit the parent, to get the order right. // (We do this in preparation for actual visiting of interface fields.) if (klass->GetSuperClass() != nullptr) { Visit(self, klass->GetSuperClass(), visitor); } for (uint32_t i = 0; i != klass->NumDirectInterfaces(); ++i) { art::ObjPtr inf_klass = art::mirror::Class::GetDirectInterface(self, klass, i); DCHECK(inf_klass != nullptr); VisitInterface(self, inf_klass, visitor); } } void VisitInterface(art::Thread* self, art::ObjPtr inf_klass, T& visitor) REQUIRES_SHARED(art::Locks::mutator_lock_) { auto it = visited_interfaces.find(inf_klass.Ptr()); if (it != visited_interfaces.end()) { return; } visited_interfaces.insert(inf_klass.Ptr()); // Let the visitor know about this one. Note that this order is acceptable, as the ordering // of these fields never matters for known visitors. visitor(inf_klass); // Now visit the superinterfaces. for (uint32_t i = 0; i != inf_klass->NumDirectInterfaces(); ++i) { art::ObjPtr super_inf_klass = art::mirror::Class::GetDirectInterface(self, inf_klass, i); DCHECK(super_inf_klass != nullptr); VisitInterface(self, super_inf_klass, visitor); } } std::unordered_set visited_interfaces; }; // Counting interface fields. Note that we cannot use the interface table, as that only contains // "non-marker" interfaces (= interfaces with methods). static size_t CountInterfaceFields(art::ObjPtr klass) REQUIRES_SHARED(art::Locks::mutator_lock_) { // Do we have a cached value? IndexCache tmp; if (gIndexCachingTable.GetTag(klass.Ptr(), &tmp)) { return tmp.interface_fields; } size_t count = 0; auto visitor = [&count](art::ObjPtr inf_klass) REQUIRES_SHARED(art::Locks::mutator_lock_) { DCHECK(inf_klass->IsInterface()); DCHECK_EQ(0u, inf_klass->NumInstanceFields()); count += inf_klass->NumStaticFields(); }; RecursiveInterfaceVisit::VisitStatic(art::Thread::Current(), klass, visitor); // Store this into the cache. tmp.interface_fields = count; gIndexCachingTable.Set(klass.Ptr(), tmp); return count; } UserData* user_data_; }; // Debug helper. Prints the structure of an object. template struct DumpVisitor { static bool Callback(art::ObjPtr obj ATTRIBUTE_UNUSED, art::ObjPtr klass ATTRIBUTE_UNUSED, art::ArtField& field, size_t field_index, void* user_data ATTRIBUTE_UNUSED) REQUIRES_SHARED(art::Locks::mutator_lock_) { LOG(ERROR) << (kStatic ? "static " : "instance ") << (kRef ? "ref " : "primitive ") << field.PrettyField() << " @ " << field_index; return false; } }; ATTRIBUTE_UNUSED void DumpObjectFields(art::ObjPtr obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (obj->IsClass()) { FieldVisitor:: ReportFields(obj, nullptr, DumpVisitor::Callback, DumpVisitor::Callback, DumpVisitor::Callback, DumpVisitor::Callback); } else { FieldVisitor::ReportFields(obj, nullptr, DumpVisitor::Callback, DumpVisitor::Callback, DumpVisitor::Callback, DumpVisitor::Callback); } } class ReportPrimitiveField { public: static bool Report(art::ObjPtr obj, ObjectTagTable* tag_table, const jvmtiHeapCallbacks* cb, const void* user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (UNLIKELY(cb->primitive_field_callback != nullptr)) { jlong class_tag = tag_table->GetTagOrZero(obj->GetClass()); ReportPrimitiveField rpf(tag_table, class_tag, cb, user_data); if (obj->IsClass()) { return FieldVisitor::ReportFields( obj, &rpf, ReportPrimitiveFieldCallback, VisitorFalse, VisitorFalse, VisitorFalse); } else { return FieldVisitor::ReportFields( obj, &rpf, VisitorFalse, VisitorFalse, ReportPrimitiveFieldCallback, VisitorFalse); } } return false; } private: ReportPrimitiveField(ObjectTagTable* tag_table, jlong class_tag, const jvmtiHeapCallbacks* cb, const void* user_data) : tag_table_(tag_table), class_tag_(class_tag), cb_(cb), user_data_(user_data) {} template static bool ReportPrimitiveFieldCallback(art::ObjPtr obj, art::ObjPtr klass, art::ArtField& field, size_t field_index, ReportPrimitiveField* user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { art::Primitive::Type art_prim_type = field.GetTypeAsPrimitiveType(); jvmtiPrimitiveType prim_type = static_cast(art::Primitive::Descriptor(art_prim_type)[0]); DCHECK(prim_type == JVMTI_PRIMITIVE_TYPE_BOOLEAN || prim_type == JVMTI_PRIMITIVE_TYPE_BYTE || prim_type == JVMTI_PRIMITIVE_TYPE_CHAR || prim_type == JVMTI_PRIMITIVE_TYPE_SHORT || prim_type == JVMTI_PRIMITIVE_TYPE_INT || prim_type == JVMTI_PRIMITIVE_TYPE_LONG || prim_type == JVMTI_PRIMITIVE_TYPE_FLOAT || prim_type == JVMTI_PRIMITIVE_TYPE_DOUBLE); jvmtiHeapReferenceInfo info; info.field.index = field_index; jvalue value; memset(&value, 0, sizeof(jvalue)); art::ObjPtr src = kReportStatic ? klass : obj; switch (art_prim_type) { case art::Primitive::Type::kPrimBoolean: value.z = field.GetBoolean(src) == 0 ? JNI_FALSE : JNI_TRUE; break; case art::Primitive::Type::kPrimByte: value.b = field.GetByte(src); break; case art::Primitive::Type::kPrimChar: value.c = field.GetChar(src); break; case art::Primitive::Type::kPrimShort: value.s = field.GetShort(src); break; case art::Primitive::Type::kPrimInt: value.i = field.GetInt(src); break; case art::Primitive::Type::kPrimLong: value.j = field.GetLong(src); break; case art::Primitive::Type::kPrimFloat: value.f = field.GetFloat(src); break; case art::Primitive::Type::kPrimDouble: value.d = field.GetDouble(src); break; case art::Primitive::Type::kPrimVoid: case art::Primitive::Type::kPrimNot: { LOG(FATAL) << "Should not reach here"; UNREACHABLE(); } } jlong obj_tag = user_data->tag_table_->GetTagOrZero(src.Ptr()); const jlong saved_obj_tag = obj_tag; jint ret = user_data->cb_->primitive_field_callback(kReportStatic ? JVMTI_HEAP_REFERENCE_STATIC_FIELD : JVMTI_HEAP_REFERENCE_FIELD, &info, user_data->class_tag_, &obj_tag, value, prim_type, const_cast(user_data->user_data_)); if (saved_obj_tag != obj_tag) { user_data->tag_table_->Set(src.Ptr(), obj_tag); } if ((ret & JVMTI_VISIT_ABORT) != 0) { return true; } return false; } ObjectTagTable* tag_table_; jlong class_tag_; const jvmtiHeapCallbacks* cb_; const void* user_data_; }; struct HeapFilter { explicit HeapFilter(jint heap_filter) : filter_out_tagged((heap_filter & JVMTI_HEAP_FILTER_TAGGED) != 0), filter_out_untagged((heap_filter & JVMTI_HEAP_FILTER_UNTAGGED) != 0), filter_out_class_tagged((heap_filter & JVMTI_HEAP_FILTER_CLASS_TAGGED) != 0), filter_out_class_untagged((heap_filter & JVMTI_HEAP_FILTER_CLASS_UNTAGGED) != 0), any_filter(filter_out_tagged || filter_out_untagged || filter_out_class_tagged || filter_out_class_untagged) { } bool ShouldReportByHeapFilter(jlong tag, jlong class_tag) const { if (!any_filter) { return true; } if ((tag == 0 && filter_out_untagged) || (tag != 0 && filter_out_tagged)) { return false; } if ((class_tag == 0 && filter_out_class_untagged) || (class_tag != 0 && filter_out_class_tagged)) { return false; } return true; } const bool filter_out_tagged; const bool filter_out_untagged; const bool filter_out_class_tagged; const bool filter_out_class_untagged; const bool any_filter; }; } // namespace void HeapUtil::Register() { art::Runtime::Current()->AddSystemWeakHolder(&gIndexCachingTable); } void HeapUtil::Unregister() { art::Runtime::Current()->RemoveSystemWeakHolder(&gIndexCachingTable); } jvmtiError HeapUtil::IterateOverInstancesOfClass(jvmtiEnv* env, jclass klass, jvmtiHeapObjectFilter filter, jvmtiHeapObjectCallback cb, const void* user_data) { if (cb == nullptr || klass == nullptr) { return ERR(NULL_POINTER); } art::Thread* self = art::Thread::Current(); art::ScopedObjectAccess soa(self); // Now we know we have the shared lock. art::StackHandleScope<1> hs(self); art::ObjPtr klass_ptr(soa.Decode(klass)); if (!klass_ptr->IsClass()) { return ERR(INVALID_CLASS); } art::Handle filter_klass(hs.NewHandle(klass_ptr->AsClass())); ObjectTagTable* tag_table = ArtJvmTiEnv::AsArtJvmTiEnv(env)->object_tag_table.get(); bool stop_reports = false; auto visitor = [&](art::mirror::Object* obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { // Early return, as we can't really stop visiting. if (stop_reports) { return; } art::ScopedAssertNoThreadSuspension no_suspension("IterateOverInstancesOfClass"); art::ObjPtr klass = obj->GetClass(); if (filter_klass != nullptr && !filter_klass->IsAssignableFrom(klass)) { return; } jlong tag = 0; tag_table->GetTag(obj, &tag); if ((filter != JVMTI_HEAP_OBJECT_EITHER) && ((tag == 0 && filter == JVMTI_HEAP_OBJECT_TAGGED) || (tag != 0 && filter == JVMTI_HEAP_OBJECT_UNTAGGED))) { return; } jlong class_tag = 0; tag_table->GetTag(klass.Ptr(), &class_tag); jlong saved_tag = tag; jint ret = cb(class_tag, obj->SizeOf(), &tag, const_cast(user_data)); stop_reports = (ret == JVMTI_ITERATION_ABORT); if (tag != saved_tag) { tag_table->Set(obj, tag); } }; art::Runtime::Current()->GetHeap()->VisitObjects(visitor); return OK; } template static jvmtiError DoIterateThroughHeap(T fn, jvmtiEnv* env, ObjectTagTable* tag_table, jint heap_filter_int, jclass klass, const jvmtiHeapCallbacks* callbacks, const void* user_data) { if (callbacks == nullptr) { return ERR(NULL_POINTER); } art::Thread* self = art::Thread::Current(); art::ScopedObjectAccess soa(self); // Now we know we have the shared lock. bool stop_reports = false; const HeapFilter heap_filter(heap_filter_int); art::StackHandleScope<1> hs(self); art::Handle filter_klass(hs.NewHandle(soa.Decode(klass))); auto visitor = [&](art::mirror::Object* obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { // Early return, as we can't really stop visiting. if (stop_reports) { return; } art::ScopedAssertNoThreadSuspension no_suspension("IterateThroughHeapCallback"); jlong tag = 0; tag_table->GetTag(obj, &tag); jlong class_tag = 0; art::ObjPtr klass = obj->GetClass(); tag_table->GetTag(klass.Ptr(), &class_tag); // For simplicity, even if we find a tag = 0, assume 0 = not tagged. if (!heap_filter.ShouldReportByHeapFilter(tag, class_tag)) { return; } if (filter_klass != nullptr) { if (filter_klass.Get() != klass) { return; } } jlong size = obj->SizeOf(); jint length = -1; if (obj->IsArrayInstance()) { length = obj->AsArray()->GetLength(); } jlong saved_tag = tag; jint ret = fn(obj, callbacks, class_tag, size, &tag, length, const_cast(user_data)); if (tag != saved_tag) { tag_table->Set(obj, tag); } stop_reports = (ret & JVMTI_VISIT_ABORT) != 0; if (!stop_reports) { jint string_ret = ReportString(obj, env, tag_table, callbacks, user_data); stop_reports = (string_ret & JVMTI_VISIT_ABORT) != 0; } if (!stop_reports) { jint array_ret = ReportPrimitiveArray(obj, env, tag_table, callbacks, user_data); stop_reports = (array_ret & JVMTI_VISIT_ABORT) != 0; } if (!stop_reports) { stop_reports = ReportPrimitiveField::Report(obj, tag_table, callbacks, user_data); } }; art::Runtime::Current()->GetHeap()->VisitObjects(visitor); return ERR(NONE); } jvmtiError HeapUtil::IterateThroughHeap(jvmtiEnv* env, jint heap_filter, jclass klass, const jvmtiHeapCallbacks* callbacks, const void* user_data) { auto JvmtiIterateHeap = [](art::mirror::Object* obj ATTRIBUTE_UNUSED, const jvmtiHeapCallbacks* cb_callbacks, jlong class_tag, jlong size, jlong* tag, jint length, void* cb_user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { return cb_callbacks->heap_iteration_callback(class_tag, size, tag, length, cb_user_data); }; return DoIterateThroughHeap(JvmtiIterateHeap, env, ArtJvmTiEnv::AsArtJvmTiEnv(env)->object_tag_table.get(), heap_filter, klass, callbacks, user_data); } class FollowReferencesHelper final { public: FollowReferencesHelper(HeapUtil* h, jvmtiEnv* jvmti_env, art::ObjPtr initial_object, const jvmtiHeapCallbacks* callbacks, art::ObjPtr class_filter, jint heap_filter, const void* user_data) : env(jvmti_env), tag_table_(h->GetTags()), initial_object_(initial_object), callbacks_(callbacks), class_filter_(class_filter), heap_filter_(heap_filter), user_data_(user_data), start_(0), stop_reports_(false) { } void Init() REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { if (initial_object_.IsNull()) { CollectAndReportRootsVisitor carrv(this, tag_table_, &worklist_, &visited_); // We need precise info (e.g., vregs). constexpr art::VisitRootFlags kRootFlags = static_cast( art::VisitRootFlags::kVisitRootFlagAllRoots | art::VisitRootFlags::kVisitRootFlagPrecise); art::Runtime::Current()->VisitRoots(&carrv, kRootFlags); art::Runtime::Current()->VisitImageRoots(&carrv); stop_reports_ = carrv.IsStopReports(); if (stop_reports_) { worklist_.clear(); } } else { visited_.insert(initial_object_.Ptr()); worklist_.push_back(initial_object_.Ptr()); } } void Work() REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { // Currently implemented as a BFS. To lower overhead, we don't erase elements immediately // from the head of the work list, instead postponing until there's a gap that's "large." // // Alternatively, we can implement a DFS and use the work list as a stack. while (start_ < worklist_.size()) { art::mirror::Object* cur_obj = worklist_[start_]; start_++; if (start_ >= kMaxStart) { worklist_.erase(worklist_.begin(), worklist_.begin() + start_); start_ = 0; } VisitObject(cur_obj); if (stop_reports_) { break; } } } private: class CollectAndReportRootsVisitor final : public art::RootVisitor { public: CollectAndReportRootsVisitor(FollowReferencesHelper* helper, ObjectTagTable* tag_table, std::vector* worklist, std::unordered_set* visited) : helper_(helper), tag_table_(tag_table), worklist_(worklist), visited_(visited), stop_reports_(false) {} void VisitRoots(art::mirror::Object*** roots, size_t count, const art::RootInfo& info) override REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*helper_->tag_table_->GetAllowDisallowLock()) { for (size_t i = 0; i != count; ++i) { AddRoot(*roots[i], info); } } void VisitRoots(art::mirror::CompressedReference** roots, size_t count, const art::RootInfo& info) override REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*helper_->tag_table_->GetAllowDisallowLock()) { for (size_t i = 0; i != count; ++i) { AddRoot(roots[i]->AsMirrorPtr(), info); } } bool IsStopReports() { return stop_reports_; } private: void AddRoot(art::mirror::Object* root_obj, const art::RootInfo& info) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { if (stop_reports_) { return; } bool add_to_worklist = ReportRoot(root_obj, info); // We use visited_ to mark roots already so we do not need another set. if (visited_->find(root_obj) == visited_->end()) { if (add_to_worklist) { visited_->insert(root_obj); worklist_->push_back(root_obj); } } } // Remove NO_THREAD_SAFETY_ANALYSIS once ASSERT_CAPABILITY works correctly. art::Thread* FindThread(const art::RootInfo& info) NO_THREAD_SAFETY_ANALYSIS { art::Locks::thread_list_lock_->AssertExclusiveHeld(art::Thread::Current()); return art::Runtime::Current()->GetThreadList()->FindThreadByThreadId(info.GetThreadId()); } jvmtiHeapReferenceKind GetReferenceKind(const art::RootInfo& info, jvmtiHeapReferenceInfo* ref_info) REQUIRES_SHARED(art::Locks::mutator_lock_) { // TODO: Fill in ref_info. memset(ref_info, 0, sizeof(jvmtiHeapReferenceInfo)); switch (info.GetType()) { case art::RootType::kRootJNIGlobal: return JVMTI_HEAP_REFERENCE_JNI_GLOBAL; case art::RootType::kRootJNILocal: { uint32_t thread_id = info.GetThreadId(); ref_info->jni_local.thread_id = thread_id; art::Thread* thread = FindThread(info); if (thread != nullptr) { art::mirror::Object* thread_obj; if (thread->IsStillStarting()) { thread_obj = nullptr; } else { thread_obj = thread->GetPeerFromOtherThread(); } if (thread_obj != nullptr) { ref_info->jni_local.thread_tag = tag_table_->GetTagOrZero(thread_obj); } } // TODO: We don't have this info. if (thread != nullptr) { ref_info->jni_local.depth = 0; art::ArtMethod* method = thread->GetCurrentMethod(nullptr, /* check_suspended= */ true, /* abort_on_error= */ false); if (method != nullptr) { ref_info->jni_local.method = art::jni::EncodeArtMethod(method); } } return JVMTI_HEAP_REFERENCE_JNI_LOCAL; } case art::RootType::kRootJavaFrame: { uint32_t thread_id = info.GetThreadId(); ref_info->stack_local.thread_id = thread_id; art::Thread* thread = FindThread(info); if (thread != nullptr) { art::mirror::Object* thread_obj; if (thread->IsStillStarting()) { thread_obj = nullptr; } else { thread_obj = thread->GetPeerFromOtherThread(); } if (thread_obj != nullptr) { ref_info->stack_local.thread_tag = tag_table_->GetTagOrZero(thread_obj); } } auto& java_info = static_cast(info); size_t vreg = java_info.GetVReg(); ref_info->stack_local.slot = static_cast( vreg <= art::JavaFrameRootInfo::kMaxVReg ? vreg : -1); const art::StackVisitor* visitor = java_info.GetVisitor(); ref_info->stack_local.location = static_cast(visitor->GetDexPc(/* abort_on_failure= */ false)); ref_info->stack_local.depth = static_cast(visitor->GetFrameDepth()); art::ArtMethod* method = visitor->GetMethod(); if (method != nullptr) { ref_info->stack_local.method = art::jni::EncodeArtMethod(method); } return JVMTI_HEAP_REFERENCE_STACK_LOCAL; } case art::RootType::kRootNativeStack: case art::RootType::kRootThreadBlock: case art::RootType::kRootThreadObject: return JVMTI_HEAP_REFERENCE_THREAD; case art::RootType::kRootStickyClass: case art::RootType::kRootInternedString: // Note: this isn't a root in the RI. return JVMTI_HEAP_REFERENCE_SYSTEM_CLASS; case art::RootType::kRootMonitorUsed: case art::RootType::kRootJNIMonitor: return JVMTI_HEAP_REFERENCE_MONITOR; case art::RootType::kRootFinalizing: case art::RootType::kRootDebugger: case art::RootType::kRootReferenceCleanup: case art::RootType::kRootVMInternal: case art::RootType::kRootUnknown: return JVMTI_HEAP_REFERENCE_OTHER; } LOG(FATAL) << "Unreachable"; UNREACHABLE(); } bool ReportRoot(art::mirror::Object* root_obj, const art::RootInfo& info) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { jvmtiHeapReferenceInfo ref_info; jvmtiHeapReferenceKind kind = GetReferenceKind(info, &ref_info); jint result = helper_->ReportReference(kind, &ref_info, nullptr, root_obj); if ((result & JVMTI_VISIT_ABORT) != 0) { stop_reports_ = true; } return (result & JVMTI_VISIT_OBJECTS) != 0; } private: FollowReferencesHelper* helper_; ObjectTagTable* tag_table_; std::vector* worklist_; std::unordered_set* visited_; bool stop_reports_; }; void VisitObject(art::mirror::Object* obj) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { if (obj->IsClass()) { VisitClass(obj->AsClass().Ptr()); return; } if (obj->IsArrayInstance()) { VisitArray(obj); return; } // All instance fields. auto report_instance_field = [&](art::ObjPtr src, art::ObjPtr obj_klass ATTRIBUTE_UNUSED, art::ArtField& field, size_t field_index, void* user_data ATTRIBUTE_UNUSED) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { art::ObjPtr field_value = field.GetObject(src); if (field_value != nullptr) { jvmtiHeapReferenceInfo reference_info; memset(&reference_info, 0, sizeof(reference_info)); reference_info.field.index = field_index; jvmtiHeapReferenceKind kind = field.GetOffset().Int32Value() == art::mirror::Object::ClassOffset().Int32Value() ? JVMTI_HEAP_REFERENCE_CLASS : JVMTI_HEAP_REFERENCE_FIELD; const jvmtiHeapReferenceInfo* reference_info_ptr = kind == JVMTI_HEAP_REFERENCE_CLASS ? nullptr : &reference_info; return !ReportReferenceMaybeEnqueue(kind, reference_info_ptr, src.Ptr(), field_value.Ptr()); } return false; }; stop_reports_ = FieldVisitor::ReportFields(obj, nullptr, VisitorFalse, VisitorFalse, VisitorFalse, report_instance_field); if (stop_reports_) { return; } jint string_ret = ReportString(obj, env, tag_table_, callbacks_, user_data_); stop_reports_ = (string_ret & JVMTI_VISIT_ABORT) != 0; if (stop_reports_) { return; } stop_reports_ = ReportPrimitiveField::Report(obj, tag_table_, callbacks_, user_data_); } void VisitArray(art::mirror::Object* array) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { stop_reports_ = !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_CLASS, nullptr, array, array->GetClass()); if (stop_reports_) { return; } if (array->IsObjectArray()) { art::ObjPtr> obj_array = array->AsObjectArray(); for (auto elem_pair : art::ZipCount(obj_array->Iterate())) { if (elem_pair.first != nullptr) { jvmtiHeapReferenceInfo reference_info; reference_info.array.index = elem_pair.second; stop_reports_ = !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_ARRAY_ELEMENT, &reference_info, array, elem_pair.first.Ptr()); if (stop_reports_) { break; } } } } else { if (!stop_reports_) { jint array_ret = ReportPrimitiveArray(array, env, tag_table_, callbacks_, user_data_); stop_reports_ = (array_ret & JVMTI_VISIT_ABORT) != 0; } } } void VisitClass(art::mirror::Class* klass) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { // TODO: Are erroneous classes reported? Are non-prepared ones? For now, just use resolved ones. if (!klass->IsResolved()) { return; } // Superclass. stop_reports_ = !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_SUPERCLASS, nullptr, klass, klass->GetSuperClass().Ptr()); if (stop_reports_) { return; } // Directly implemented or extended interfaces. art::Thread* self = art::Thread::Current(); art::StackHandleScope<1> hs(self); art::Handle h_klass(hs.NewHandle(klass)); for (size_t i = 0; i < h_klass->NumDirectInterfaces(); ++i) { art::ObjPtr inf_klass = art::mirror::Class::ResolveDirectInterface(self, h_klass, i); if (inf_klass == nullptr) { // TODO: With a resolved class this should not happen... self->ClearException(); break; } stop_reports_ = !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_INTERFACE, nullptr, klass, inf_klass.Ptr()); if (stop_reports_) { return; } } // Classloader. // TODO: What about the boot classpath loader? We'll skip for now, but do we have to find the // fake BootClassLoader? if (klass->GetClassLoader() != nullptr) { stop_reports_ = !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_CLASS_LOADER, nullptr, klass, klass->GetClassLoader().Ptr()); if (stop_reports_) { return; } } DCHECK_EQ(h_klass.Get(), klass); // Declared static fields. auto report_static_field = [&](art::ObjPtr obj ATTRIBUTE_UNUSED, art::ObjPtr obj_klass, art::ArtField& field, size_t field_index, void* user_data ATTRIBUTE_UNUSED) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { art::ObjPtr field_value = field.GetObject(obj_klass); if (field_value != nullptr) { jvmtiHeapReferenceInfo reference_info; memset(&reference_info, 0, sizeof(reference_info)); reference_info.field.index = static_cast(field_index); return !ReportReferenceMaybeEnqueue(JVMTI_HEAP_REFERENCE_STATIC_FIELD, &reference_info, obj_klass.Ptr(), field_value.Ptr()); } return false; }; stop_reports_ = FieldVisitor::ReportFields(klass, nullptr, VisitorFalse, report_static_field, VisitorFalse, VisitorFalse); if (stop_reports_) { return; } stop_reports_ = ReportPrimitiveField::Report(klass, tag_table_, callbacks_, user_data_); } void MaybeEnqueue(art::mirror::Object* obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (visited_.find(obj) == visited_.end()) { worklist_.push_back(obj); visited_.insert(obj); } } bool ReportReferenceMaybeEnqueue(jvmtiHeapReferenceKind kind, const jvmtiHeapReferenceInfo* reference_info, art::mirror::Object* referree, art::mirror::Object* referrer) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { jint result = ReportReference(kind, reference_info, referree, referrer); if ((result & JVMTI_VISIT_ABORT) == 0) { if ((result & JVMTI_VISIT_OBJECTS) != 0) { MaybeEnqueue(referrer); } return true; } else { return false; } } jint ReportReference(jvmtiHeapReferenceKind kind, const jvmtiHeapReferenceInfo* reference_info, art::mirror::Object* referrer, art::mirror::Object* referree) REQUIRES_SHARED(art::Locks::mutator_lock_) REQUIRES(!*tag_table_->GetAllowDisallowLock()) { if (referree == nullptr || stop_reports_) { return 0; } if (UNLIKELY(class_filter_ != nullptr) && class_filter_ != referree->GetClass()) { return JVMTI_VISIT_OBJECTS; } const jlong class_tag = tag_table_->GetTagOrZero(referree->GetClass()); jlong tag = tag_table_->GetTagOrZero(referree); if (!heap_filter_.ShouldReportByHeapFilter(tag, class_tag)) { return JVMTI_VISIT_OBJECTS; } const jlong referrer_class_tag = referrer == nullptr ? 0 : tag_table_->GetTagOrZero(referrer->GetClass()); const jlong size = static_cast(referree->SizeOf()); jlong saved_tag = tag; jlong referrer_tag = 0; jlong saved_referrer_tag = 0; jlong* referrer_tag_ptr; if (referrer == nullptr) { referrer_tag_ptr = nullptr; } else { if (referrer == referree) { referrer_tag_ptr = &tag; } else { referrer_tag = saved_referrer_tag = tag_table_->GetTagOrZero(referrer); referrer_tag_ptr = &referrer_tag; } } jint length = -1; if (referree->IsArrayInstance()) { length = referree->AsArray()->GetLength(); } jint result = callbacks_->heap_reference_callback(kind, reference_info, class_tag, referrer_class_tag, size, &tag, referrer_tag_ptr, length, const_cast(user_data_)); if (tag != saved_tag) { tag_table_->Set(referree, tag); } if (referrer_tag != saved_referrer_tag) { tag_table_->Set(referrer, referrer_tag); } return result; } jvmtiEnv* env; ObjectTagTable* tag_table_; art::ObjPtr initial_object_; const jvmtiHeapCallbacks* callbacks_; art::ObjPtr class_filter_; const HeapFilter heap_filter_; const void* user_data_; std::vector worklist_; size_t start_; static constexpr size_t kMaxStart = 1000000U; std::unordered_set visited_; bool stop_reports_; friend class CollectAndReportRootsVisitor; }; jvmtiError HeapUtil::FollowReferences(jvmtiEnv* env, jint heap_filter, jclass klass, jobject initial_object, const jvmtiHeapCallbacks* callbacks, const void* user_data) { if (callbacks == nullptr) { return ERR(NULL_POINTER); } art::Thread* self = art::Thread::Current(); art::gc::Heap* heap = art::Runtime::Current()->GetHeap(); if (heap->IsGcConcurrentAndMoving()) { // Need to take a heap dump while GC isn't running. See the // comment in Heap::VisitObjects(). heap->IncrementDisableMovingGC(self); } { art::ScopedObjectAccess soa(self); // Now we know we have the shared lock. art::jni::ScopedEnableSuspendAllJniIdQueries sjni; // make sure we can get JNI ids. art::ScopedThreadSuspension sts(self, art::kWaitingForVisitObjects); art::ScopedSuspendAll ssa("FollowReferences"); art::ObjPtr class_filter = klass == nullptr ? nullptr : art::ObjPtr::DownCast(self->DecodeJObject(klass)); FollowReferencesHelper frh(this, env, self->DecodeJObject(initial_object), callbacks, class_filter, heap_filter, user_data); frh.Init(); frh.Work(); } if (heap->IsGcConcurrentAndMoving()) { heap->DecrementDisableMovingGC(self); } return ERR(NONE); } jvmtiError HeapUtil::GetLoadedClasses(jvmtiEnv* env, jint* class_count_ptr, jclass** classes_ptr) { if (class_count_ptr == nullptr || classes_ptr == nullptr) { return ERR(NULL_POINTER); } class ReportClassVisitor : public art::ClassVisitor { public: explicit ReportClassVisitor(art::Thread* self) : self_(self) {} bool operator()(art::ObjPtr klass) override REQUIRES_SHARED(art::Locks::mutator_lock_) { if (klass->IsLoaded() || klass->IsErroneous()) { classes_.push_back(self_->GetJniEnv()->AddLocalReference(klass)); } return true; } art::Thread* self_; std::vector classes_; }; art::Thread* self = art::Thread::Current(); ReportClassVisitor rcv(self); { art::ScopedObjectAccess soa(self); art::Runtime::Current()->GetClassLinker()->VisitClasses(&rcv); } size_t size = rcv.classes_.size(); jclass* classes = nullptr; jvmtiError alloc_ret = env->Allocate(static_cast(size * sizeof(jclass)), reinterpret_cast(&classes)); if (alloc_ret != ERR(NONE)) { return alloc_ret; } for (size_t i = 0; i < size; ++i) { classes[i] = rcv.classes_[i]; } *classes_ptr = classes; *class_count_ptr = static_cast(size); return ERR(NONE); } jvmtiError HeapUtil::ForceGarbageCollection(jvmtiEnv* env ATTRIBUTE_UNUSED) { art::Runtime::Current()->GetHeap()->CollectGarbage(/* clear_soft_references= */ false); return ERR(NONE); } static constexpr jint kHeapIdDefault = 0; static constexpr jint kHeapIdImage = 1; static constexpr jint kHeapIdZygote = 2; static constexpr jint kHeapIdApp = 3; static jint GetHeapId(art::ObjPtr obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { if (obj == nullptr) { return -1; } art::gc::Heap* const heap = art::Runtime::Current()->GetHeap(); const art::gc::space::ContinuousSpace* const space = heap->FindContinuousSpaceFromObject(obj, true); jint heap_type = kHeapIdApp; if (space != nullptr) { if (space->IsZygoteSpace()) { heap_type = kHeapIdZygote; } else if (space->IsImageSpace() && heap->ObjectIsInBootImageSpace(obj)) { // Only count objects in the boot image as HPROF_HEAP_IMAGE, this leaves app image objects // as HPROF_HEAP_APP. b/35762934 heap_type = kHeapIdImage; } } else { const auto* los = heap->GetLargeObjectsSpace(); if (los->Contains(obj.Ptr()) && los->IsZygoteLargeObject(art::Thread::Current(), obj.Ptr())) { heap_type = kHeapIdZygote; } } return heap_type; }; jvmtiError HeapExtensions::GetObjectHeapId(jvmtiEnv* env, jlong tag, jint* heap_id, ...) { if (heap_id == nullptr) { return ERR(NULL_POINTER); } art::Thread* self = art::Thread::Current(); auto work = [&]() REQUIRES_SHARED(art::Locks::mutator_lock_) { ObjectTagTable* tag_table = ArtJvmTiEnv::AsArtJvmTiEnv(env)->object_tag_table.get(); art::ObjPtr obj = tag_table->Find(tag); jint heap_type = GetHeapId(obj); if (heap_type == -1) { return ERR(NOT_FOUND); } *heap_id = heap_type; return ERR(NONE); }; if (!art::Locks::mutator_lock_->IsSharedHeld(self)) { if (!self->IsThreadSuspensionAllowable()) { return ERR(INTERNAL); } art::ScopedObjectAccess soa(self); return work(); } else { // We cannot use SOA in this case. We might be holding the lock, but may not be in the // runnable state (e.g., during GC). art::Locks::mutator_lock_->AssertSharedHeld(self); // TODO: Investigate why ASSERT_SHARED_CAPABILITY doesn't work. auto annotalysis_workaround = [&]() NO_THREAD_SAFETY_ANALYSIS { return work(); }; return annotalysis_workaround(); } } static jvmtiError CopyStringAndReturn(jvmtiEnv* env, const char* in, char** out) { jvmtiError error; JvmtiUniquePtr param_name = CopyString(env, in, &error); if (param_name == nullptr) { return error; } *out = param_name.release(); return ERR(NONE); } static constexpr const char* kHeapIdDefaultName = "default"; static constexpr const char* kHeapIdImageName = "image"; static constexpr const char* kHeapIdZygoteName = "zygote"; static constexpr const char* kHeapIdAppName = "app"; jvmtiError HeapExtensions::GetHeapName(jvmtiEnv* env, jint heap_id, char** heap_name, ...) { switch (heap_id) { case kHeapIdDefault: return CopyStringAndReturn(env, kHeapIdDefaultName, heap_name); case kHeapIdImage: return CopyStringAndReturn(env, kHeapIdImageName, heap_name); case kHeapIdZygote: return CopyStringAndReturn(env, kHeapIdZygoteName, heap_name); case kHeapIdApp: return CopyStringAndReturn(env, kHeapIdAppName, heap_name); default: return ERR(ILLEGAL_ARGUMENT); } } jvmtiError HeapExtensions::IterateThroughHeapExt(jvmtiEnv* env, jint heap_filter, jclass klass, const jvmtiHeapCallbacks* callbacks, const void* user_data) { if (ArtJvmTiEnv::AsArtJvmTiEnv(env)->capabilities.can_tag_objects != 1) { \ return ERR(MUST_POSSESS_CAPABILITY); \ } // ART extension API: Also pass the heap id. auto ArtIterateHeap = [](art::mirror::Object* obj, const jvmtiHeapCallbacks* cb_callbacks, jlong class_tag, jlong size, jlong* tag, jint length, void* cb_user_data) REQUIRES_SHARED(art::Locks::mutator_lock_) { jint heap_id = GetHeapId(obj); using ArtExtensionAPI = jint (*)(jlong, jlong, jlong*, jint length, void*, jint); return reinterpret_cast(cb_callbacks->heap_iteration_callback)( class_tag, size, tag, length, cb_user_data, heap_id); }; return DoIterateThroughHeap(ArtIterateHeap, env, ArtJvmTiEnv::AsArtJvmTiEnv(env)->object_tag_table.get(), heap_filter, klass, callbacks, user_data); } namespace { using ObjectPtr = art::ObjPtr; using ObjectMap = std::unordered_map; static void ReplaceObjectReferences(const ObjectMap& map) REQUIRES(art::Locks::mutator_lock_, art::Roles::uninterruptible_) { art::Runtime::Current()->GetHeap()->VisitObjectsPaused( [&](art::mirror::Object* ref) REQUIRES_SHARED(art::Locks::mutator_lock_) { // Rewrite all references in the object if needed. class ResizeReferenceVisitor { public: using CompressedObj = art::mirror::CompressedReference; explicit ResizeReferenceVisitor(const ObjectMap& map, ObjectPtr ref) : map_(map), ref_(ref) {} // Ignore class roots. void VisitRootIfNonNull(CompressedObj* root) const REQUIRES_SHARED(art::Locks::mutator_lock_) { if (root != nullptr) { VisitRoot(root); } } void VisitRoot(CompressedObj* root) const REQUIRES_SHARED(art::Locks::mutator_lock_) { auto it = map_.find(root->AsMirrorPtr()); if (it != map_.end()) { root->Assign(it->second); art::WriteBarrier::ForEveryFieldWrite(ref_); } } void operator()(art::ObjPtr obj, art::MemberOffset off, bool is_static) const REQUIRES_SHARED(art::Locks::mutator_lock_) { auto it = map_.find(obj->GetFieldObject(off)); if (it != map_.end()) { UNUSED(is_static); if (UNLIKELY(!is_static && off == art::mirror::Object::ClassOffset())) { // We don't want to update the declaring class of any objects. They will be replaced // in the heap and we need the declaring class to know its size. return; } else if (UNLIKELY(!is_static && off == art::mirror::Class::SuperClassOffset() && obj->IsClass())) { // We don't want to be messing with the class hierarcy either. return; } VLOG(plugin) << "Updating field at offset " << off.Uint32Value() << " of type " << obj->GetClass()->PrettyClass(); obj->SetFieldObject(off, it->second); art::WriteBarrier::ForEveryFieldWrite(obj); } } // java.lang.ref.Reference visitor. void operator()(art::ObjPtr klass ATTRIBUTE_UNUSED, art::ObjPtr ref) const REQUIRES_SHARED(art::Locks::mutator_lock_) { operator()(ref, art::mirror::Reference::ReferentOffset(), /* is_static */ false); } private: const ObjectMap& map_; ObjectPtr ref_; }; ResizeReferenceVisitor rrv(map, ref); if (ref->IsClass()) { // Class object native roots are the ArtField and ArtMethod 'declaring_class_' fields // which we don't want to be messing with as it would break ref-visitor assumptions about // what a class looks like. We want to keep the default behavior in other cases (such as // dex-cache) though. Unfortunately there is no way to tell from the visitor where exactly // the root came from. // TODO It might be nice to have the visitors told where the reference came from. ref->VisitReferences(rrv, rrv); } else { ref->VisitReferences(rrv, rrv); } }); } static void ReplaceStrongRoots(art::Thread* self, const ObjectMap& map) REQUIRES(art::Locks::mutator_lock_, art::Roles::uninterruptible_) { // replace root references expcept java frames. struct ResizeRootVisitor : public art::RootVisitor { public: explicit ResizeRootVisitor(const ObjectMap& map) : map_(map) {} // TODO It's somewhat annoying to have to have this function implemented twice. It might be // good/useful to implement operator= for CompressedReference to allow us to use a template to // implement both of these. void VisitRoots(art::mirror::Object*** roots, size_t count, const art::RootInfo& info) override REQUIRES_SHARED(art::Locks::mutator_lock_) { art::mirror::Object*** end = roots + count; for (art::mirror::Object** obj = *roots; roots != end; obj = *(++roots)) { auto it = map_.find(*obj); if (it != map_.end()) { // Java frames might have the JIT doing optimizations (for example loop-unrolling or // eliding bounds checks) so we need deopt them once we're done here. if (info.GetType() == art::RootType::kRootJavaFrame) { const art::JavaFrameRootInfo& jfri = art::down_cast(info); if (jfri.GetVReg() == art::JavaFrameRootInfo::kMethodDeclaringClass) { info.Describe(VLOG_STREAM(plugin) << "Not changing declaring-class during stack" << " walk. Found obsolete java frame id "); continue; } else { info.Describe(VLOG_STREAM(plugin) << "Found java frame id "); threads_with_roots_.insert(info.GetThreadId()); } } *obj = it->second.Ptr(); } } } void VisitRoots(art::mirror::CompressedReference** roots, size_t count, const art::RootInfo& info) override REQUIRES_SHARED(art::Locks::mutator_lock_) { art::mirror::CompressedReference** end = roots + count; for (art::mirror::CompressedReference* obj = *roots; roots != end; obj = *(++roots)) { auto it = map_.find(obj->AsMirrorPtr()); if (it != map_.end()) { // Java frames might have the JIT doing optimizations (for example loop-unrolling or // eliding bounds checks) so we need deopt them once we're done here. if (info.GetType() == art::RootType::kRootJavaFrame) { const art::JavaFrameRootInfo& jfri = art::down_cast(info); if (jfri.GetVReg() == art::JavaFrameRootInfo::kMethodDeclaringClass) { info.Describe(VLOG_STREAM(plugin) << "Not changing declaring-class during stack" << " walk. Found obsolete java frame id "); continue; } else { info.Describe(VLOG_STREAM(plugin) << "Found java frame id "); threads_with_roots_.insert(info.GetThreadId()); } } obj->Assign(it->second); } } } const std::unordered_set& GetThreadsWithJavaFrameRoots() const { return threads_with_roots_; } private: const ObjectMap& map_; std::unordered_set threads_with_roots_; }; ResizeRootVisitor rrv(map); art::Runtime::Current()->VisitRoots(&rrv, art::VisitRootFlags::kVisitRootFlagAllRoots); // Handle java Frames. Annoyingly the JIT can embed information about the length of the array into // the compiled code. By changing the length of the array we potentially invalidate these // assumptions and so could cause (eg) OOB array access or other issues. if (!rrv.GetThreadsWithJavaFrameRoots().empty()) { art::MutexLock mu(self, *art::Locks::thread_list_lock_); art::ThreadList* thread_list = art::Runtime::Current()->GetThreadList(); art::instrumentation::Instrumentation* instr = art::Runtime::Current()->GetInstrumentation(); for (uint32_t id : rrv.GetThreadsWithJavaFrameRoots()) { art::Thread* t = thread_list->FindThreadByThreadId(id); CHECK(t != nullptr) << "id " << id << " does not refer to a valid thread." << " Where did the roots come from?"; VLOG(plugin) << "Instrumenting thread stack of thread " << *t; // TODO Use deopt manager. We need a version that doesn't acquire all the locks we // already have. // TODO We technically only need to do this if the frames are not already being interpreted. // The cost for doing an extra stack walk is unlikely to be worth it though. instr->InstrumentThreadStack(t); } } } static void ReplaceWeakRoots(art::Thread* self, EventHandler* event_handler, const ObjectMap& map) REQUIRES(art::Locks::mutator_lock_, art::Roles::uninterruptible_) { // Handle tags. We want to do this seprately from other weak-refs (handled below) because we need // to send additional events and handle cases where the agent might have tagged the new // replacement object during the VMObjectAlloc. We do this by removing all tags associated with // both the obsolete and the new arrays. Then we send the ObsoleteObjectCreated event and cache // the new tag values. We next update all the other weak-references (the tags have been removed) // and finally update the tag table with the new values. Doing things in this way (1) keeps all // code relating to updating weak-references together and (2) ensures we don't end up in strange // situations where the order of weak-ref visiting affects the final tagging state. Since we have // the mutator_lock_ and gc-paused throughout this whole process no threads should be able to see // the interval where the objects are not tagged. struct NewTagValue { public: ObjectPtr obsolete_obj_; jlong obsolete_tag_; ObjectPtr new_obj_; jlong new_tag_; }; // Map from the environment to the list of pairs that were changed. std::unordered_map> changed_tags; event_handler->ForEachEnv(self, [&](ArtJvmTiEnv* env) { // Cannot have REQUIRES(art::Locks::mutator_lock_) since ForEachEnv doesn't require it. art::Locks::mutator_lock_->AssertExclusiveHeld(self); env->object_tag_table->Lock(); // Get the tags and clear them (so we don't need to special-case the normal weak-ref visitor) for (auto it : map) { jlong new_tag = 0; jlong obsolete_tag = 0; bool had_obsolete_tag = env->object_tag_table->RemoveLocked(it.first, &obsolete_tag); bool had_new_tag = env->object_tag_table->RemoveLocked(it.second, &new_tag); // Dispatch event. if (had_obsolete_tag || had_new_tag) { event_handler->DispatchEventOnEnv( env, self, &obsolete_tag, &new_tag); changed_tags.try_emplace(env).first->second.push_back( { it.first, obsolete_tag, it.second, new_tag }); } } // After weak-ref update we need to go back and re-add obsoletes. We wait to avoid having to // deal with the visit-weaks overwriting the initial new_obj_ptr tag and generally making things // difficult. env->object_tag_table->Unlock(); }); // Handle weak-refs. struct ReplaceWeaksVisitor : public art::IsMarkedVisitor { public: ReplaceWeaksVisitor(const ObjectMap& map) : map_(map) {} art::mirror::Object* IsMarked(art::mirror::Object* obj) REQUIRES_SHARED(art::Locks::mutator_lock_) { auto it = map_.find(obj); if (it != map_.end()) { return it->second.Ptr(); } else { return obj; } } private: const ObjectMap& map_; }; ReplaceWeaksVisitor rwv(map); art::Runtime::Current()->SweepSystemWeaks(&rwv); // Re-add the object tags. At this point all weak-references to the old_obj_ptr are gone. event_handler->ForEachEnv(self, [&](ArtJvmTiEnv* env) { // Cannot have REQUIRES(art::Locks::mutator_lock_) since ForEachEnv doesn't require it. art::Locks::mutator_lock_->AssertExclusiveHeld(self); env->object_tag_table->Lock(); auto it = changed_tags.find(env); if (it != changed_tags.end()) { for (const NewTagValue& v : it->second) { env->object_tag_table->SetLocked(v.obsolete_obj_, v.obsolete_tag_); env->object_tag_table->SetLocked(v.new_obj_, v.new_tag_); } } env->object_tag_table->Unlock(); }); } } // namespace void HeapExtensions::ReplaceReference(art::Thread* self, art::ObjPtr old_obj_ptr, art::ObjPtr new_obj_ptr) { ObjectMap map { { old_obj_ptr, new_obj_ptr } }; ReplaceReferences(self, map); } void HeapExtensions::ReplaceReferences(art::Thread* self, const ObjectMap& map) { ReplaceObjectReferences(map); ReplaceStrongRoots(self, map); ReplaceWeakRoots(self, HeapExtensions::gEventHandler, map); } jvmtiError HeapExtensions::ChangeArraySize(jvmtiEnv* env, jobject arr, jsize new_size) { if (ArtJvmTiEnv::AsArtJvmTiEnv(env)->capabilities.can_tag_objects != 1) { return ERR(MUST_POSSESS_CAPABILITY); } art::Thread* self = art::Thread::Current(); ScopedNoUserCodeSuspension snucs(self); art::ScopedObjectAccess soa(self); if (arr == nullptr) { JVMTI_LOG(INFO, env) << "Cannot resize a null object"; return ERR(NULL_POINTER); } art::ObjPtr klass(soa.Decode(arr)->GetClass()); if (!klass->IsArrayClass()) { JVMTI_LOG(INFO, env) << klass->PrettyClass() << " is not an array class!"; return ERR(ILLEGAL_ARGUMENT); } if (new_size < 0) { JVMTI_LOG(INFO, env) << "Cannot resize an array to a negative size"; return ERR(ILLEGAL_ARGUMENT); } // Allocate the new copy. art::StackHandleScope<2> hs(self); art::Handle old_arr(hs.NewHandle(soa.Decode(arr))); art::MutableHandle new_arr(hs.NewHandle(nullptr)); if (klass->IsObjectArrayClass()) { new_arr.Assign( art::mirror::ObjectArray::Alloc(self, old_arr->GetClass(), new_size)); } else { // NB This also copies the old array but since we aren't suspended we need to do this again to // catch any concurrent modifications. new_arr.Assign(art::mirror::Array::CopyOf(old_arr, self, new_size)); } if (new_arr.IsNull()) { self->AssertPendingOOMException(); JVMTI_LOG(INFO, env) << "Unable to allocate " << old_arr->GetClass()->PrettyClass() << " (length: " << new_size << ") due to OOME. Error was: " << self->GetException()->Dump(); self->ClearException(); return ERR(OUT_OF_MEMORY); } else { self->AssertNoPendingException(); } // Suspend everything. art::ScopedThreadSuspension sts(self, art::ThreadState::kSuspended); art::gc::ScopedGCCriticalSection sgccs( self, art::gc::GcCause::kGcCauseDebugger, art::gc::CollectorType::kCollectorTypeDebugger); art::ScopedSuspendAll ssa("Resize array!"); // Replace internals. new_arr->SetLockWord(old_arr->GetLockWord(false), false); old_arr->SetLockWord(art::LockWord::Default(), false); // Copy the contents now when everything is suspended. int32_t size = std::min(old_arr->GetLength(), new_size); switch (old_arr->GetClass()->GetComponentType()->GetPrimitiveType()) { case art::Primitive::kPrimBoolean: new_arr->AsBooleanArray()->Memcpy(0, old_arr->AsBooleanArray(), 0, size); break; case art::Primitive::kPrimByte: new_arr->AsByteArray()->Memcpy(0, old_arr->AsByteArray(), 0, size); break; case art::Primitive::kPrimChar: new_arr->AsCharArray()->Memcpy(0, old_arr->AsCharArray(), 0, size); break; case art::Primitive::kPrimShort: new_arr->AsShortArray()->Memcpy(0, old_arr->AsShortArray(), 0, size); break; case art::Primitive::kPrimInt: new_arr->AsIntArray()->Memcpy(0, old_arr->AsIntArray(), 0, size); break; case art::Primitive::kPrimLong: new_arr->AsLongArray()->Memcpy(0, old_arr->AsLongArray(), 0, size); break; case art::Primitive::kPrimFloat: new_arr->AsFloatArray()->Memcpy(0, old_arr->AsFloatArray(), 0, size); break; case art::Primitive::kPrimDouble: new_arr->AsDoubleArray()->Memcpy(0, old_arr->AsDoubleArray(), 0, size); break; case art::Primitive::kPrimNot: for (int32_t i = 0; i < size; i++) { new_arr->AsObjectArray()->Set( i, old_arr->AsObjectArray()->Get(i)); } break; case art::Primitive::kPrimVoid: LOG(FATAL) << "void-array is not a legal type!"; UNREACHABLE(); } // Actually replace all the pointers. ReplaceReference(self, old_arr.Get(), new_arr.Get()); return OK; } void HeapExtensions::Register(EventHandler* eh) { gEventHandler = eh; } } // namespace openjdkjvmti