crashpad/snapshot/mac/mach_o_image_annotations_reader_test.cc

488 lines
19 KiB
C++
Raw Normal View History

// Copyright 2014 The Crashpad Authors
//
// 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 "snapshot/mac/mach_o_image_annotations_reader.h"
#include <dlfcn.h>
#include <mach/mach.h>
#include <signal.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
#include <unistd.h>
#include <map>
#include <string>
#include <vector>
#include "base/files/file_path.h"
#include "client/annotation.h"
#include "client/annotation_list.h"
#include "client/crashpad_info.h"
#include "client/simple_string_dictionary.h"
#include "gtest/gtest.h"
#include "snapshot/mac/process_reader_mac.h"
#include "test/errors.h"
#include "test/mac/mach_errors.h"
#include "test/mac/mach_multiprocess.h"
#include "test/test_paths.h"
#include "util/file/file_io.h"
#include "util/mac/mac_util.h"
#include "util/mach/exc_server_variants.h"
#include "util/mach/exception_ports.h"
#include "util/mach/mach_extensions.h"
#include "util/mach/mach_message.h"
#include "util/mach/mach_message_server.h"
namespace crashpad {
namespace test {
namespace {
// \return The path to crashpad_snapshot_test_module_crashy_initializer.so
base::FilePath ModuleWithCrashyInitializer() {
return TestPaths::BuildArtifact("snapshot",
"module_crashy_initializer",
TestPaths::FileType::kLoadableModule);
}
//! \return The path to the crashpad_snapshot_test_no_op executable.
base::FilePath NoOpExecutable() {
return TestPaths::BuildArtifact(
"snapshot", "no_op", TestPaths::FileType::kExecutable);
}
class TestMachOImageAnnotationsReader final
: public MachMultiprocess,
public UniversalMachExcServer::Interface {
public:
enum TestType {
// Dont crash, just test the CrashpadInfo interface.
kDontCrash = 0,
// The child process should crash by calling abort(). The parent verifies
// that the system libraries set the expected annotations.
//
// This test verifies that the message field in crashreporter_annotations_t
// can be recovered. Either 10.10.2 Libc-1044.1.2/stdlib/FreeBSD/abort.c
// abort() or 10.10.2 Libc-1044.10.1/sys/_libc_fork_child.c
// _libc_fork_child() calls CRSetCrashLogMessage() to set the message field.
kCrashAbort,
// The child process should crash at module initialization time, when dyld
// will have set an annotation matching the path of the module being
// initialized.
//
// This test exists to verify that the message2 field in
// crashreporter_annotations_t can be recovered. 10.10.2
// dyld-353.2.1/src/ImageLoaderMachO.cpp
// ImageLoaderMachO::doInitialization() calls CRSetCrashLogMessage2() to set
// the message2 field.
kCrashModuleInitialization,
// The child process should crash by setting DYLD_INSERT_LIBRARIES to
// contain a nonexistent library. The parent verifies that dyld sets the
// expected annotations.
kCrashDyld,
};
explicit TestMachOImageAnnotationsReader(TestType test_type)
: MachMultiprocess(),
UniversalMachExcServer::Interface(),
test_type_(test_type) {
mac: Tests that crash intentionally shouldn’t go to ReportCrash Crashpad has many tests that crash intentionally. Some of these are gtest death tests, and others arrange for intentional crashes to test Crashpad’s own crash-catching logic. On macOS, all of the gtest death tests and some of the other intentional crashes were being logged by ReportCrash, the system’s crash reporter. Since these reports corresponded to intentional crashes, they were never useful, and served only to clutter ~/Library/Logs/DiagnosticReports. Since Crashpad is adept at handling exceptions on its own, this introduces the “exception swallowing server”, crashpad_exception_swallower, which is a Mach exception server that implements a no-op exception handler routine for all exceptions received. The exception swallowing server is established as the task handler for EXC_CRASH and EXC_CORPSE_NOTIFY exceptions during gtest death tests invoked by {ASSERT,EXPECT}_DEATH_{CHECK,CRASH}, and for all child processes invoked by the Multiprocess test infrastructure. The exception swallowing server is not in effect at other times, so unexpected crashes in test code can still be handled by ReportCrash or another crash reporter. With this change in place, no new reports are generated in the user-level ~/Library/Logs/DiagnosticReports or the system’s /Library/Logs/DiagnosticReports during a run of Crashpad’s full test suite on macOS. Bug: crashpad:33 Change-Id: I13891853a7e25accc30da21fa7ea8bd7d1f3bd2f Reviewed-on: https://chromium-review.googlesource.com/777859 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Robert Sesek <rsesek@chromium.org>
2017-11-20 13:32:26 -05:00
switch (test_type_) {
case kDontCrash:
// SetExpectedChildTermination(kTerminationNormal, EXIT_SUCCESS) is the
// default.
break;
case kCrashAbort:
SetExpectedChildTermination(kTerminationSignal, SIGABRT);
break;
case kCrashModuleInitialization:
SetExpectedChildTerminationBuiltinTrap();
mac: Tests that crash intentionally shouldn’t go to ReportCrash Crashpad has many tests that crash intentionally. Some of these are gtest death tests, and others arrange for intentional crashes to test Crashpad’s own crash-catching logic. On macOS, all of the gtest death tests and some of the other intentional crashes were being logged by ReportCrash, the system’s crash reporter. Since these reports corresponded to intentional crashes, they were never useful, and served only to clutter ~/Library/Logs/DiagnosticReports. Since Crashpad is adept at handling exceptions on its own, this introduces the “exception swallowing server”, crashpad_exception_swallower, which is a Mach exception server that implements a no-op exception handler routine for all exceptions received. The exception swallowing server is established as the task handler for EXC_CRASH and EXC_CORPSE_NOTIFY exceptions during gtest death tests invoked by {ASSERT,EXPECT}_DEATH_{CHECK,CRASH}, and for all child processes invoked by the Multiprocess test infrastructure. The exception swallowing server is not in effect at other times, so unexpected crashes in test code can still be handled by ReportCrash or another crash reporter. With this change in place, no new reports are generated in the user-level ~/Library/Logs/DiagnosticReports or the system’s /Library/Logs/DiagnosticReports during a run of Crashpad’s full test suite on macOS. Bug: crashpad:33 Change-Id: I13891853a7e25accc30da21fa7ea8bd7d1f3bd2f Reviewed-on: https://chromium-review.googlesource.com/777859 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Robert Sesek <rsesek@chromium.org>
2017-11-20 13:32:26 -05:00
break;
case kCrashDyld:
// Prior to 10.12, dyld fatal errors result in the execution of an
// int3 instruction on x86 and a trap instruction on ARM, both of
// which raise SIGTRAP. 10.9.5 dyld-239.4/src/dyldStartup.s
// _dyld_fatal_error. This changed in 10.12 to use
// abort_with_payload(), which appears as SIGABRT to a waiting parent.
SetExpectedChildTermination(
kTerminationSignal,
MacOSVersionNumber() < 10'12'00 ? SIGTRAP : SIGABRT);
mac: Tests that crash intentionally shouldn’t go to ReportCrash Crashpad has many tests that crash intentionally. Some of these are gtest death tests, and others arrange for intentional crashes to test Crashpad’s own crash-catching logic. On macOS, all of the gtest death tests and some of the other intentional crashes were being logged by ReportCrash, the system’s crash reporter. Since these reports corresponded to intentional crashes, they were never useful, and served only to clutter ~/Library/Logs/DiagnosticReports. Since Crashpad is adept at handling exceptions on its own, this introduces the “exception swallowing server”, crashpad_exception_swallower, which is a Mach exception server that implements a no-op exception handler routine for all exceptions received. The exception swallowing server is established as the task handler for EXC_CRASH and EXC_CORPSE_NOTIFY exceptions during gtest death tests invoked by {ASSERT,EXPECT}_DEATH_{CHECK,CRASH}, and for all child processes invoked by the Multiprocess test infrastructure. The exception swallowing server is not in effect at other times, so unexpected crashes in test code can still be handled by ReportCrash or another crash reporter. With this change in place, no new reports are generated in the user-level ~/Library/Logs/DiagnosticReports or the system’s /Library/Logs/DiagnosticReports during a run of Crashpad’s full test suite on macOS. Bug: crashpad:33 Change-Id: I13891853a7e25accc30da21fa7ea8bd7d1f3bd2f Reviewed-on: https://chromium-review.googlesource.com/777859 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Robert Sesek <rsesek@chromium.org>
2017-11-20 13:32:26 -05:00
break;
}
}
TestMachOImageAnnotationsReader(const TestMachOImageAnnotationsReader&) =
delete;
TestMachOImageAnnotationsReader& operator=(
const TestMachOImageAnnotationsReader&) = delete;
~TestMachOImageAnnotationsReader() {}
// UniversalMachExcServer::Interface:
kern_return_t CatchMachException(exception_behavior_t behavior,
exception_handler_t exception_port,
thread_t thread,
task_t task,
exception_type_t exception,
const mach_exception_data_type_t* code,
mach_msg_type_number_t code_count,
thread_state_flavor_t* flavor,
ConstThreadState old_state,
mach_msg_type_number_t old_state_count,
thread_state_t new_state,
mach_msg_type_number_t* new_state_count,
const mach_msg_trailer_t* trailer,
bool* destroy_complex_request) override {
*destroy_complex_request = true;
if (test_type_ != kCrashDyld) {
// In 10.12.1 and later, the task port will not match ChildTask() in the
// kCrashDyld case, because kCrashDyld uses execl(), which results in a
// new task port being assigned.
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
EXPECT_EQ(task, ChildTask());
}
// The process ID should always compare favorably.
pid_t task_pid;
kern_return_t kr = pid_for_task(task, &task_pid);
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
EXPECT_EQ(kr, KERN_SUCCESS) << MachErrorMessage(kr, "pid_for_task");
EXPECT_EQ(task_pid, ChildPID());
ProcessReaderMac process_reader;
bool rv = process_reader.Initialize(task);
if (!rv) {
ADD_FAILURE();
} else {
const std::vector<ProcessReaderMac::Module>& modules =
process_reader.Modules();
std::vector<std::string> all_annotations_vector;
for (const ProcessReaderMac::Module& module : modules) {
if (module.reader) {
MachOImageAnnotationsReader module_annotations_reader(
&process_reader, module.reader, module.name);
std::vector<std::string> module_annotations_vector =
module_annotations_reader.Vector();
all_annotations_vector.insert(all_annotations_vector.end(),
module_annotations_vector.begin(),
module_annotations_vector.end());
} else {
EXPECT_TRUE(module.reader);
}
}
// Mac OS X 10.6 doesnt have support for CrashReporter annotations
// (CrashReporterClient.h), so dont look for any special annotations in
// that version.
const int macos_version_number = MacOSVersionNumber();
if (macos_version_number > 10'07'00) {
EXPECT_GE(all_annotations_vector.size(), 1u);
std::string expected_annotation;
switch (test_type_) {
case kCrashAbort:
// The child process calls abort(), so the expected annotation
// reflects this, with a string set by 10.7.5
// Libc-763.13/stdlib/abort-fbsd.c abort(). This string is still
// present in 10.9.5 Libc-997.90.3/stdlib/FreeBSD/abort.c abort(),
// but because abort() tests to see if a message is already set and
// something else in Libc will have set a message, this string is
// not the expectation on 10.9 or higher. Instead, after fork(), the
// child process has a message indicating that a fork() without
// exec() occurred. See 10.9.5 Libc-997.90.3/sys/_libc_fork_child.c
// _libc_fork_child().
expected_annotation =
macos_version_number <= 10'08'00
? "abort() called"
: "crashed on child side of fork pre-exec";
break;
case kCrashModuleInitialization:
// This message is set by dyld-353.2.1/src/ImageLoaderMachO.cpp
// ImageLoaderMachO::doInitialization().
// dyld4 no longer sets this, so instead check for fork() without
// exec() like above.
expected_annotation =
macos_version_number < 12'00'00
? ModuleWithCrashyInitializer().value()
: "crashed on child side of fork pre-exec";
break;
case kCrashDyld:
// This is independent of dylds error_string, which is tested
// below. dyld4 no longer sets this.
expected_annotation =
macos_version_number < 12'00'00
? "dyld: launch, loading dependent libraries"
: "";
break;
default:
ADD_FAILURE();
break;
}
bool found = false;
for (const std::string& annotation : all_annotations_vector) {
// Look for the expectation as a leading susbtring, because the actual
// string that dyld uses will have the contents of the
// DYLD_INSERT_LIBRARIES environment variable appended to it on OS X
// 10.10.
if (annotation.substr(0, expected_annotation.length()) ==
expected_annotation) {
found = true;
break;
}
}
EXPECT_TRUE(found) << expected_annotation;
}
// dyld exposes its error_string at least as far back as Mac OS X 10.4.
if (test_type_ == kCrashDyld) {
std::string couldnt_load_annotation =
macos_version_number < 12'00'00
? "could not load inserted library"
// dyld4 no longer writes an annotation for the primary error
// See https://crbug.com/1334418/#c26
: "tried: '/var/empty/NoDirectory/NoLibrary' (no such file)";
bool found = false;
for (const std::string& annotation : all_annotations_vector) {
// Look for the expectation as a substring, because the actual
// string will contain the librarys pathname and a reason, or on
// macOS 12, only the reason.
if (annotation.find(couldnt_load_annotation) != std::string::npos) {
found = true;
break;
}
}
EXPECT_TRUE(found) << couldnt_load_annotation;
}
}
ExcServerCopyState(
behavior, old_state, old_state_count, new_state, new_state_count);
return ExcServerSuccessfulReturnValue(exception, behavior, false);
}
private:
// MachMultiprocess:
void MachMultiprocessParent() override {
ProcessReaderMac process_reader;
ASSERT_TRUE(process_reader.Initialize(ChildTask()));
// Wait for the child process to indicate that its done setting up its
// annotations via the CrashpadInfo interface.
char c;
Make file_io reads more rational and predictable ReadFile() attempted to continue reading after a short read. In most cases, this is fine. However, ReadFile() would keep trying to fill a partially-filled buffer until experiencing a 0-length read(), signaling end-of-file. For certain weird file descriptors like terminal input, EOF is an ephemeral condition, and attempting to read beyond EOF doesn’t actually return 0 (EOF) provided that they remain open, it will block waiting for more input. Consequently, ReadFile() and anything based on ReadFile() had an undocumented and quirky interface, which was that any short read that it returned (not an underlying short read) actually indicated EOF. This facet of ReadFile() was unexpected, so it’s being removed. The new behavior is that ReadFile() will return an underlying short read. The behavior of FileReaderInterface::Read() is updated in accordance with this change. Upon experiencing a short read, the caller can determine the best action. Most callers were already prepared for this behavior. Outside of util/file, only crashpad_database_util properly implemented EOF detection according to previous semantics, and adapting it to new semantics is trivial. Callers who require an exact-length read can use the new ReadFileExactly(), or the newly renamed LoggingReadFileExactly() or CheckedReadFileExactly(). These functions will retry following a short read. The renamed functions were previously called LoggingReadFile() and CheckedReadFile(), but those names implied that they were simply wrapping ReadFile(), which is not the case. They wrapped ReadFile() and further, insisted on a full read. Since ReadFile()’s semantics are now changing but these functions’ are not, they’re now even more distinct from ReadFile(), and must be renamed to avoid confusion. Test: * Change-Id: I06b77e0d6ad8719bd2eb67dab93a8740542dd908 Reviewed-on: https://chromium-review.googlesource.com/456676 Reviewed-by: Robert Sesek <rsesek@chromium.org>
2017-03-16 13:36:38 -04:00
CheckedReadFileExactly(ReadPipeHandle(), &c, sizeof(c));
// Verify the “simple map” and object-based annotations set via the
// CrashpadInfo interface.
const std::vector<ProcessReaderMac::Module>& modules =
process_reader.Modules();
std::map<std::string, std::string> all_annotations_simple_map;
std::vector<AnnotationSnapshot> all_annotations;
for (const ProcessReaderMac::Module& module : modules) {
MachOImageAnnotationsReader module_annotations_reader(
&process_reader, module.reader, module.name);
std::map<std::string, std::string> module_annotations_simple_map =
module_annotations_reader.SimpleMap();
all_annotations_simple_map.insert(module_annotations_simple_map.begin(),
module_annotations_simple_map.end());
std::vector<AnnotationSnapshot> annotations =
module_annotations_reader.AnnotationsList();
all_annotations.insert(
all_annotations.end(), annotations.begin(), annotations.end());
}
EXPECT_GE(all_annotations_simple_map.size(), 5u);
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
EXPECT_EQ(all_annotations_simple_map["#TEST# pad"], "crash");
EXPECT_EQ(all_annotations_simple_map["#TEST# key"], "value");
EXPECT_EQ(all_annotations_simple_map["#TEST# x"], "y");
EXPECT_EQ(all_annotations_simple_map["#TEST# longer"], "shorter");
EXPECT_EQ(all_annotations_simple_map["#TEST# empty_value"], "");
EXPECT_EQ(all_annotations.size(), 3u);
bool saw_same_name_3 = false, saw_same_name_4 = false;
for (const auto& annotation : all_annotations) {
EXPECT_EQ(annotation.type,
static_cast<uint16_t>(Annotation::Type::kString));
std::string value(reinterpret_cast<const char*>(annotation.value.data()),
annotation.value.size());
if (annotation.name == "#TEST# one") {
EXPECT_EQ(value, "moocow");
} else if (annotation.name == "#TEST# same-name") {
if (value == "same-name 3") {
EXPECT_FALSE(saw_same_name_3);
saw_same_name_3 = true;
} else if (value == "same-name 4") {
EXPECT_FALSE(saw_same_name_4);
saw_same_name_4 = true;
} else {
ADD_FAILURE() << "unexpected annotation value " << value;
}
} else {
ADD_FAILURE() << "unexpected annotation " << annotation.name;
}
}
// Tell the child process that its permitted to crash.
CheckedWriteFile(WritePipeHandle(), &c, sizeof(c));
if (test_type_ != kDontCrash) {
// Handle the childs crash. Further validation will be done in
// CatchMachException().
UniversalMachExcServer universal_mach_exc_server(this);
mach_msg_return_t mr =
MachMessageServer::Run(&universal_mach_exc_server,
LocalPort(),
MACH_MSG_OPTION_NONE,
MachMessageServer::kOneShot,
MachMessageServer::kReceiveLargeError,
kMachMessageTimeoutWaitIndefinitely);
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
EXPECT_EQ(mr, MACH_MSG_SUCCESS)
<< MachErrorMessage(mr, "MachMessageServer::Run");
}
}
void MachMultiprocessChild() override {
CrashpadInfo* crashpad_info = CrashpadInfo::GetCrashpadInfo();
// This is “leaked” to crashpad_info.
SimpleStringDictionary* simple_annotations = new SimpleStringDictionary();
simple_annotations->SetKeyValue("#TEST# pad", "break");
simple_annotations->SetKeyValue("#TEST# key", "value");
simple_annotations->SetKeyValue("#TEST# pad", "crash");
simple_annotations->SetKeyValue("#TEST# x", "y");
simple_annotations->SetKeyValue("#TEST# longer", "shorter");
simple_annotations->SetKeyValue("#TEST# empty_value", "");
crashpad_info->set_simple_annotations(simple_annotations);
AnnotationList::Register(); // This is “leaked” to crashpad_info.
static StringAnnotation<32> test_annotation_one{"#TEST# one"};
static StringAnnotation<32> test_annotation_two{"#TEST# two"};
static StringAnnotation<32> test_annotation_three{"#TEST# same-name"};
static StringAnnotation<32> test_annotation_four{"#TEST# same-name"};
test_annotation_one.Set("moocow");
test_annotation_two.Set("this will be cleared");
test_annotation_three.Set("same-name 3");
test_annotation_four.Set("same-name 4");
test_annotation_two.Clear();
// Tell the parent that the environment has been set up.
char c = '\0';
CheckedWriteFile(WritePipeHandle(), &c, sizeof(c));
// Wait for the parent to indicate that its safe to crash.
Make file_io reads more rational and predictable ReadFile() attempted to continue reading after a short read. In most cases, this is fine. However, ReadFile() would keep trying to fill a partially-filled buffer until experiencing a 0-length read(), signaling end-of-file. For certain weird file descriptors like terminal input, EOF is an ephemeral condition, and attempting to read beyond EOF doesn’t actually return 0 (EOF) provided that they remain open, it will block waiting for more input. Consequently, ReadFile() and anything based on ReadFile() had an undocumented and quirky interface, which was that any short read that it returned (not an underlying short read) actually indicated EOF. This facet of ReadFile() was unexpected, so it’s being removed. The new behavior is that ReadFile() will return an underlying short read. The behavior of FileReaderInterface::Read() is updated in accordance with this change. Upon experiencing a short read, the caller can determine the best action. Most callers were already prepared for this behavior. Outside of util/file, only crashpad_database_util properly implemented EOF detection according to previous semantics, and adapting it to new semantics is trivial. Callers who require an exact-length read can use the new ReadFileExactly(), or the newly renamed LoggingReadFileExactly() or CheckedReadFileExactly(). These functions will retry following a short read. The renamed functions were previously called LoggingReadFile() and CheckedReadFile(), but those names implied that they were simply wrapping ReadFile(), which is not the case. They wrapped ReadFile() and further, insisted on a full read. Since ReadFile()’s semantics are now changing but these functions’ are not, they’re now even more distinct from ReadFile(), and must be renamed to avoid confusion. Test: * Change-Id: I06b77e0d6ad8719bd2eb67dab93a8740542dd908 Reviewed-on: https://chromium-review.googlesource.com/456676 Reviewed-by: Robert Sesek <rsesek@chromium.org>
2017-03-16 13:36:38 -04:00
CheckedReadFileExactly(ReadPipeHandle(), &c, sizeof(c));
// Direct an exception message to the exception server running in the
// parent.
ExceptionPorts exception_ports(ExceptionPorts::kTargetTypeTask,
mach_task_self());
ASSERT_TRUE(exception_ports.SetExceptionPort(
EXC_MASK_CRASH, RemotePort(), EXCEPTION_DEFAULT, THREAD_STATE_NONE));
switch (test_type_) {
case kDontCrash: {
break;
}
case kCrashAbort: {
abort();
}
case kCrashModuleInitialization: {
// Load a module that crashes while executing a module initializer.
void* dl_handle = dlopen(ModuleWithCrashyInitializer().value().c_str(),
RTLD_LAZY | RTLD_LOCAL);
// This should have crashed in the dlopen(). If dlopen() failed, the
// ASSERT_NE() will show the message. If it succeeded without crashing,
// the FAIL() will fail the test.
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
ASSERT_NE(dl_handle, nullptr) << dlerror();
FAIL();
}
case kCrashDyld: {
// Set DYLD_INSERT_LIBRARIES to contain a library that does not exist.
// Unable to load it, dyld will abort with a fatal error.
ASSERT_EQ(
setenv(
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
"DYLD_INSERT_LIBRARIES", "/var/empty/NoDirectory/NoLibrary", 1),
0)
<< ErrnoMessage("setenv");
// The actual executable doesnt matter very much, because dyld wont
// ever launch it. It just needs to be an executable that uses dyld as
// its LC_LOAD_DYLINKER (all normal executables do). A custom no-op
// executable is provided because DYLD_INSERT_LIBRARIES does not work
// with system executables on OS X 10.11 due to System Integrity
// Protection.
base::FilePath no_op_executable = NoOpExecutable();
test: Use (actual, [un]expected) in gtest {ASSERT,EXPECT}_{EQ,NE} gtest used to require (expected, actual) ordering for arguments to EXPECT_EQ and ASSERT_EQ, and in failed test assertions would identify each side as “expected” or “actual.” Tests in Crashpad adhered to this traditional ordering. After a gtest change in February 2016, it is now agnostic with respect to the order of these arguments. This change mechanically updates all uses of these macros to (actual, expected) by reversing them. This provides consistency with our use of the logging CHECK_EQ and DCHECK_EQ macros, and makes for better readability by ordinary native speakers. The rough (but working!) conversion tool is https://chromium-review.googlesource.com/c/466727/1/rewrite_expectassert_eq.py, and “git cl format” cleaned up its output. EXPECT_NE and ASSERT_NE never had a preferred ordering. gtest never made a judgment that one side or the other needed to provide an “unexpected” value. Consequently, some code used (unexpected, actual) while other code used (actual, unexpected). For consistency with the new EXPECT_EQ and ASSERT_EQ usage, as well as consistency with CHECK_NE and DCHECK_NE, this change also updates these use sites to (actual, unexpected) where one side can be called “unexpected” as, for example, std::string::npos can be. Unfortunately, this portion was a manual conversion. References: https://github.com/google/googletest/blob/master/googletest/docs/Primer.md#binary-comparison https://github.com/google/googletest/commit/77d6b173380332b1c1bc540532641f410ec82d65 https://github.com/google/googletest/pull/713 Change-Id: I978fef7c94183b8b1ef63f12f5ab4d6693626be3 Reviewed-on: https://chromium-review.googlesource.com/466727 Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-04-04 00:35:21 -04:00
ASSERT_EQ(execl(no_op_executable.value().c_str(),
no_op_executable.BaseName().value().c_str(),
nullptr),
0)
<< ErrnoMessage("execl");
break;
}
default:
break;
}
}
TestType test_type_;
};
TEST(MachOImageAnnotationsReader, DontCrash) {
TestMachOImageAnnotationsReader test_mach_o_image_annotations_reader(
TestMachOImageAnnotationsReader::kDontCrash);
test_mach_o_image_annotations_reader.Run();
}
TEST(MachOImageAnnotationsReader, CrashAbort) {
TestMachOImageAnnotationsReader test_mach_o_image_annotations_reader(
TestMachOImageAnnotationsReader::kCrashAbort);
test_mach_o_image_annotations_reader.Run();
}
TEST(MachOImageAnnotationsReader, CrashModuleInitialization) {
TestMachOImageAnnotationsReader test_mach_o_image_annotations_reader(
TestMachOImageAnnotationsReader::kCrashModuleInitialization);
test_mach_o_image_annotations_reader.Run();
}
TEST(MachOImageAnnotationsReader, CrashDyld) {
TestMachOImageAnnotationsReader test_mach_o_image_annotations_reader(
TestMachOImageAnnotationsReader::kCrashDyld);
test_mach_o_image_annotations_reader.Run();
}
} // namespace
} // namespace test
} // namespace crashpad