platform_system_core/libunwindstack/tests/ElfTest.cpp

585 lines
17 KiB
C++
Raw Normal View History

/*
* 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 <elf.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
#include <gmock/gmock.h>
#include <gtest/gtest.h>
#include <unwindstack/Elf.h>
#include <unwindstack/MapInfo.h>
#include <unwindstack/RegsArm.h>
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
#include "ElfFake.h"
#include "ElfTestUtils.h"
#include "LogFake.h"
#include "MemoryFake.h"
#if !defined(PT_ARM_EXIDX)
#define PT_ARM_EXIDX 0x70000001
#endif
namespace unwindstack {
class ElfTest : public ::testing::Test {
protected:
void SetUp() override {
memory_ = new MemoryFake;
}
void InitElf32(uint32_t machine_type) {
Elf32_Ehdr ehdr;
TestInitEhdr<Elf32_Ehdr>(&ehdr, ELFCLASS32, machine_type);
ehdr.e_phoff = 0x100;
ehdr.e_ehsize = sizeof(ehdr);
ehdr.e_phentsize = sizeof(Elf32_Phdr);
ehdr.e_phnum = 1;
ehdr.e_shentsize = sizeof(Elf32_Shdr);
if (machine_type == EM_ARM) {
ehdr.e_flags = 0x5000200;
ehdr.e_phnum = 2;
}
memory_->SetMemory(0, &ehdr, sizeof(ehdr));
Elf32_Phdr phdr;
memset(&phdr, 0, sizeof(phdr));
phdr.p_type = PT_LOAD;
phdr.p_filesz = 0x10000;
phdr.p_memsz = 0x10000;
phdr.p_flags = PF_R | PF_X;
phdr.p_align = 0x1000;
memory_->SetMemory(0x100, &phdr, sizeof(phdr));
if (machine_type == EM_ARM) {
memset(&phdr, 0, sizeof(phdr));
phdr.p_type = PT_ARM_EXIDX;
phdr.p_offset = 0x30000;
phdr.p_vaddr = 0x30000;
phdr.p_paddr = 0x30000;
phdr.p_filesz = 16;
phdr.p_memsz = 16;
phdr.p_flags = PF_R;
phdr.p_align = 0x4;
memory_->SetMemory(0x100 + sizeof(phdr), &phdr, sizeof(phdr));
}
}
void InitElf64(uint32_t machine_type) {
Elf64_Ehdr ehdr;
TestInitEhdr<Elf64_Ehdr>(&ehdr, ELFCLASS64, machine_type);
ehdr.e_phoff = 0x100;
ehdr.e_flags = 0x5000200;
ehdr.e_ehsize = sizeof(ehdr);
ehdr.e_phentsize = sizeof(Elf64_Phdr);
ehdr.e_phnum = 1;
ehdr.e_shentsize = sizeof(Elf64_Shdr);
memory_->SetMemory(0, &ehdr, sizeof(ehdr));
Elf64_Phdr phdr;
memset(&phdr, 0, sizeof(phdr));
phdr.p_type = PT_LOAD;
phdr.p_filesz = 0x10000;
phdr.p_memsz = 0x10000;
phdr.p_flags = PF_R | PF_X;
phdr.p_align = 0x1000;
memory_->SetMemory(0x100, &phdr, sizeof(phdr));
}
MemoryFake* memory_;
};
TEST_F(ElfTest, invalid_memory) {
Elf elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_FALSE(elf.Init(false));
ASSERT_FALSE(elf.valid());
}
TEST_F(ElfTest, elf_invalid) {
Elf elf(memory_);
InitElf32(EM_386);
// Corrupt the ELF signature.
memory_->SetData32(0, 0x7f000000);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_FALSE(elf.Init(false));
ASSERT_FALSE(elf.valid());
ASSERT_TRUE(elf.interface() == nullptr);
std::string name;
ASSERT_FALSE(elf.GetSoname(&name));
uint64_t func_offset;
ASSERT_FALSE(elf.GetFunctionName(0, &name, &func_offset));
bool finished;
ASSERT_FALSE(elf.Step(0, 0, 0, nullptr, nullptr, &finished));
}
TEST_F(ElfTest, elf32_invalid_machine) {
Elf elf(memory_);
InitElf32(EM_PPC);
ResetLogs();
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_FALSE(elf.Init(false));
ASSERT_EQ("", GetFakeLogBuf());
ASSERT_EQ("4 unwind 32 bit elf that is neither arm nor x86 nor mips: e_machine = 20\n\n",
GetFakeLogPrint());
}
TEST_F(ElfTest, elf64_invalid_machine) {
Elf elf(memory_);
InitElf64(EM_PPC64);
ResetLogs();
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_FALSE(elf.Init(false));
ASSERT_EQ("", GetFakeLogBuf());
ASSERT_EQ("4 unwind 64 bit elf that is neither aarch64 nor x86_64 nor mips64: e_machine = 21\n\n",
GetFakeLogPrint());
}
TEST_F(ElfTest, elf_arm) {
Elf elf(memory_);
InitElf32(EM_ARM);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_ARM), elf.machine_type());
ASSERT_EQ(ELFCLASS32, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, elf_mips) {
Elf elf(memory_);
InitElf32(EM_MIPS);
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_MIPS), elf.machine_type());
ASSERT_EQ(ELFCLASS32, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, elf_x86) {
Elf elf(memory_);
InitElf32(EM_386);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_386), elf.machine_type());
ASSERT_EQ(ELFCLASS32, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, elf_arm64) {
Elf elf(memory_);
InitElf64(EM_AARCH64);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_AARCH64), elf.machine_type());
ASSERT_EQ(ELFCLASS64, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, elf_x86_64) {
Elf elf(memory_);
InitElf64(EM_X86_64);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_X86_64), elf.machine_type());
ASSERT_EQ(ELFCLASS64, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, elf_mips64) {
Elf elf(memory_);
InitElf64(EM_MIPS);
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.valid());
ASSERT_EQ(static_cast<uint32_t>(EM_MIPS), elf.machine_type());
ASSERT_EQ(ELFCLASS64, elf.class_type());
ASSERT_TRUE(elf.interface() != nullptr);
}
TEST_F(ElfTest, gnu_debugdata_init_fail32) {
TestInitGnuDebugdata<Elf32_Ehdr, Elf32_Shdr>(ELFCLASS32, EM_ARM, false,
[&](uint64_t offset, const void* ptr, size_t size) {
memory_->SetMemory(offset, ptr, size);
});
Elf elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.interface() != nullptr);
ASSERT_TRUE(elf.gnu_debugdata_interface() == nullptr);
EXPECT_EQ(0x1acU, elf.interface()->gnu_debugdata_offset());
EXPECT_EQ(0x100U, elf.interface()->gnu_debugdata_size());
}
TEST_F(ElfTest, gnu_debugdata_init_fail64) {
TestInitGnuDebugdata<Elf64_Ehdr, Elf64_Shdr>(ELFCLASS64, EM_AARCH64, false,
[&](uint64_t offset, const void* ptr, size_t size) {
memory_->SetMemory(offset, ptr, size);
});
Elf elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(false));
ASSERT_TRUE(elf.interface() != nullptr);
ASSERT_TRUE(elf.gnu_debugdata_interface() == nullptr);
EXPECT_EQ(0x200U, elf.interface()->gnu_debugdata_offset());
EXPECT_EQ(0x100U, elf.interface()->gnu_debugdata_size());
}
TEST_F(ElfTest, gnu_debugdata_init32) {
TestInitGnuDebugdata<Elf32_Ehdr, Elf32_Shdr>(ELFCLASS32, EM_ARM, true,
[&](uint64_t offset, const void* ptr, size_t size) {
memory_->SetMemory(offset, ptr, size);
});
Elf elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(true));
ASSERT_TRUE(elf.interface() != nullptr);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.gnu_debugdata_interface() != nullptr);
EXPECT_EQ(0x1acU, elf.interface()->gnu_debugdata_offset());
EXPECT_EQ(0x8cU, elf.interface()->gnu_debugdata_size());
}
TEST_F(ElfTest, gnu_debugdata_init64) {
TestInitGnuDebugdata<Elf64_Ehdr, Elf64_Shdr>(ELFCLASS64, EM_AARCH64, true,
[&](uint64_t offset, const void* ptr, size_t size) {
memory_->SetMemory(offset, ptr, size);
});
Elf elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.Init(true));
ASSERT_TRUE(elf.interface() != nullptr);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ASSERT_TRUE(elf.gnu_debugdata_interface() != nullptr);
EXPECT_EQ(0x200U, elf.interface()->gnu_debugdata_offset());
EXPECT_EQ(0x90U, elf.interface()->gnu_debugdata_size());
}
TEST_F(ElfTest, rel_pc) {
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ElfFake elf(memory_);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
ElfInterfaceFake* interface = new ElfInterfaceFake(memory_);
elf.FakeSetInterface(interface);
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
MapInfo map_info(0x1000, 0x2000);
ASSERT_EQ(0x101U, elf.GetRelPc(0x1101, &map_info));
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
elf.FakeSetLoadBias(0x3000);
ASSERT_EQ(0x3101U, elf.GetRelPc(0x1101, &map_info));
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
elf.FakeSetValid(false);
elf.FakeSetLoadBias(0);
ASSERT_EQ(0x101U, elf.GetRelPc(0x1101, &map_info));
}
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
TEST_F(ElfTest, step_in_signal_map) {
ElfFake elf(memory_);
RegsArm regs;
regs[13] = 0x50000;
regs[15] = 0x8000;
regs.SetFromRaw();
ElfInterfaceFake* interface = new ElfInterfaceFake(memory_);
elf.FakeSetInterface(interface);
memory_->SetData32(0x3000, 0xdf0027ad);
MemoryFake process_memory;
process_memory.SetData32(0x50000, 0);
for (size_t i = 0; i < 16; i++) {
process_memory.SetData32(0x500a0 + i * sizeof(uint32_t), i);
}
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
bool finished;
ASSERT_TRUE(elf.Step(0x1000, 0x1000, 0x2000, &regs, &process_memory, &finished));
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
EXPECT_FALSE(finished);
EXPECT_EQ(15U, regs.pc());
EXPECT_EQ(13U, regs.sp());
}
class ElfInterfaceMock : public ElfInterface {
public:
ElfInterfaceMock(Memory* memory) : ElfInterface(memory) {}
virtual ~ElfInterfaceMock() = default;
bool Init(uint64_t*) override { return false; }
void InitHeaders() override {}
bool GetSoname(std::string*) override { return false; }
bool GetFunctionName(uint64_t, uint64_t, std::string*, uint64_t*) override { return false; }
MOCK_METHOD5(Step, bool(uint64_t, uint64_t, Regs*, Memory*, bool*));
MOCK_METHOD2(GetGlobalVariable, bool(const std::string&, uint64_t*));
MOCK_METHOD1(IsValidPc, bool(uint64_t));
void MockSetDynamicOffset(uint64_t offset) { dynamic_offset_ = offset; }
void MockSetDynamicVaddr(uint64_t vaddr) { dynamic_vaddr_ = vaddr; }
void MockSetDynamicSize(uint64_t size) { dynamic_size_ = size; }
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
};
TEST_F(ElfTest, step_in_interface) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
RegsArm regs;
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
MemoryFake process_memory;
bool finished;
EXPECT_CALL(*interface, Step(0x1000, 0, &regs, &process_memory, &finished))
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
.WillOnce(::testing::Return(true));
ASSERT_TRUE(elf.Step(0x1004, 0x1000, 0x2000, &regs, &process_memory, &finished));
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
}
TEST_F(ElfTest, step_in_interface_non_zero_load_bias) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0x4000);
RegsArm regs;
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
MemoryFake process_memory;
bool finished;
EXPECT_CALL(*interface, Step(0x7300, 0x4000, &regs, &process_memory, &finished))
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
.WillOnce(::testing::Return(true));
ASSERT_TRUE(elf.Step(0x7304, 0x7300, 0x2000, &regs, &process_memory, &finished));
Multiple bugfixes, small restructuring. - Move the load bias stored out of ElfInterface into Elf. For the compressed sections, the load bias was not the same as the data from the uncompressed section. - Move the initialization of the compressed section into Init. It was too easy to forget to call the init of the compressed section. - Do not automatically add in load bias to the pc before calling ElfInterface code. Do all of the pc manipulations in the Elf object. - Change the interface GetFunctionName code to pass in the load_bias instead of modifying the pc inside the code. - Modify the Step function to pass in the elf offset, not add it to the pc. It is necessary to have two different relative values when executing the Step: a pc that is relative to the beginning of the elf for the reading data the actual instructions when trying to determine if this is in a signal frame, and a pc that is relative to the map for finding the appropriate unwind information. - Add a feature to Unwinder so that an unwind can be stopped if it ends up in map that has a specified suffix. This is so that the ART unwinding code doesn't require skipping the compressed section. Instead, stop at if trying to unwind through a known suffix code that means the code is in java code. This is important because the compressed section data is not only used by the jave compiled code, so that will continue to work. - Fix tests for restructuring, add new tests for new functionality. Test: Ran art test 137-cfi using new unwinder as default. Test: Ran new unit tests. Change-Id: I42e658c64c5e14f698ba34944a3043afac967884
2017-10-20 01:08:58 +02:00
}
TEST_F(ElfTest, get_global_invalid_elf) {
ElfFake elf(memory_);
elf.FakeSetValid(false);
std::string global("something");
uint64_t offset;
ASSERT_FALSE(elf.GetGlobalVariable(global, &offset));
}
TEST_F(ElfTest, get_global_valid_not_in_interface) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset)).WillOnce(::testing::Return(false));
ASSERT_FALSE(elf.GetGlobalVariable(global, &offset));
}
TEST_F(ElfTest, get_global_valid_below_load_bias) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0x1000);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x300), ::testing::Return(true)));
ASSERT_FALSE(elf.GetGlobalVariable(global, &offset));
}
TEST_F(ElfTest, get_global_valid_dynamic_zero) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
ElfInterfaceMock* gnu_interface = new ElfInterfaceMock(memory_);
elf.FakeSetGnuDebugdataInterface(gnu_interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset)).WillOnce(::testing::Return(false));
EXPECT_CALL(*gnu_interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x500), ::testing::Return(true)));
ASSERT_TRUE(elf.GetGlobalVariable(global, &offset));
EXPECT_EQ(0x500U, offset);
}
TEST_F(ElfTest, get_global_valid_in_gnu_debugdata_dynamic_zero) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x300), ::testing::Return(true)));
ASSERT_TRUE(elf.GetGlobalVariable(global, &offset));
EXPECT_EQ(0x300U, offset);
}
TEST_F(ElfTest, get_global_valid_dynamic_zero_non_zero_load_bias) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0x100);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x300), ::testing::Return(true)));
ASSERT_TRUE(elf.GetGlobalVariable(global, &offset));
EXPECT_EQ(0x200U, offset);
}
TEST_F(ElfTest, get_global_valid_dynamic_adjust_negative) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
interface->MockSetDynamicOffset(0x400);
interface->MockSetDynamicVaddr(0x800);
interface->MockSetDynamicSize(0x100);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x850), ::testing::Return(true)));
ASSERT_TRUE(elf.GetGlobalVariable(global, &offset));
EXPECT_EQ(0x450U, offset);
}
TEST_F(ElfTest, get_global_valid_dynamic_adjust_positive) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
interface->MockSetDynamicOffset(0x1000);
interface->MockSetDynamicVaddr(0x800);
interface->MockSetDynamicSize(0x100);
elf.FakeSetInterface(interface);
uint64_t offset;
std::string global("something");
EXPECT_CALL(*interface, GetGlobalVariable(global, &offset))
.WillOnce(::testing::DoAll(::testing::SetArgPointee<1>(0x850), ::testing::Return(true)));
ASSERT_TRUE(elf.GetGlobalVariable(global, &offset));
EXPECT_EQ(0x1050U, offset);
}
TEST_F(ElfTest, is_valid_pc_elf_invalid) {
ElfFake elf(memory_);
elf.FakeSetValid(false);
elf.FakeSetLoadBias(0);
EXPECT_FALSE(elf.IsValidPc(0x100));
EXPECT_FALSE(elf.IsValidPc(0x200));
}
TEST_F(ElfTest, is_valid_pc_interface) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
EXPECT_CALL(*interface, IsValidPc(0x1500)).WillOnce(::testing::Return(true));
EXPECT_TRUE(elf.IsValidPc(0x1500));
}
TEST_F(ElfTest, is_valid_pc_non_zero_load_bias) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0x1000);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
EXPECT_CALL(*interface, IsValidPc(0x500)).WillOnce(::testing::Return(true));
EXPECT_FALSE(elf.IsValidPc(0x100));
EXPECT_FALSE(elf.IsValidPc(0x200));
EXPECT_TRUE(elf.IsValidPc(0x1500));
}
TEST_F(ElfTest, is_valid_pc_from_gnu_debugdata) {
ElfFake elf(memory_);
elf.FakeSetValid(true);
elf.FakeSetLoadBias(0);
ElfInterfaceMock* interface = new ElfInterfaceMock(memory_);
elf.FakeSetInterface(interface);
ElfInterfaceMock* gnu_interface = new ElfInterfaceMock(memory_);
elf.FakeSetGnuDebugdataInterface(gnu_interface);
EXPECT_CALL(*interface, IsValidPc(0x1500)).WillOnce(::testing::Return(false));
EXPECT_CALL(*gnu_interface, IsValidPc(0x1500)).WillOnce(::testing::Return(true));
EXPECT_TRUE(elf.IsValidPc(0x1500));
}
} // namespace unwindstack