crashpad/snapshot/win/exception_snapshot_win_test.cc

306 lines
11 KiB
C++
Raw Normal View History

// Copyright 2015 The Crashpad Authors. All rights reserved.
//
// 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/win/exception_snapshot_win.h"
#include <string>
#include "base/files/file_path.h"
#include "base/strings/string16.h"
#include "base/strings/utf_string_conversions.h"
#include "client/crashpad_client.h"
#include "gtest/gtest.h"
#include "snapshot/win/process_snapshot_win.h"
#include "test/errors.h"
win: Dynamically disable WoW64 tests absent explicit 32-bit build output Rather than having the 64-bit build assume that it lives in out\{Debug,Release}_x64 and that it can find 32-bit build output in out\{Debug,Release}, require the location of 32-bit build output to be provided explicitly via the CRASHPAD_TEST_32_BIT_OUTPUT environment variable. If this variable is not set, 64-bit tests that require 32-bit test build output will dynamically disable themselves at runtime. In order for this to work, a new DISABLED_TEST() macro is added to support dynamically disabled tests. gtest does not have its own first-class support for this (https://groups.google.com/d/topic/googletestframework/Nwh3u7YFuN4, https://github.com/google/googletest/issues/490) so this local solution is used instead. For tests via Crashpad’s own build\run_tests.py, which is how Crashpad’s own buildbots and trybots invoke tests, CRASHPAD_TEST_32_BIT_OUTPUT is set to a locaton compatible with the paths expected for the GYP-based build. No test coverage is lost on Crashpad’s own buildbots and trybots. For Crashpad tests in Chromium’s buildbots and trybots, this environment variable will not be set, causing these tests to be dynamically disabled. Bug: crashpad:203, chromium:743139, chromium:777924 Change-Id: I3c0de2bf4f835e13ed5a4adda5760d6fed508126 Reviewed-on: https://chromium-review.googlesource.com/739795 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-10-26 13:48:01 -04:00
#include "test/gtest_disabled.h"
#include "test/test_paths.h"
#include "test/win/child_launcher.h"
#include "util/file/file_io.h"
#include "util/thread/thread.h"
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
#include "util/win/exception_handler_server.h"
#include "util/win/registration_protocol_win.h"
#include "util/win/scoped_handle.h"
#include "util/win/scoped_process_suspend.h"
namespace crashpad {
namespace test {
namespace {
// Runs the ExceptionHandlerServer on a background thread.
class RunServerThread : public Thread {
public:
// Instantiates a thread which will invoke server->Run(delegate);
RunServerThread(ExceptionHandlerServer* server,
ExceptionHandlerServer::Delegate* delegate)
: server_(server), delegate_(delegate) {}
~RunServerThread() override {}
private:
// Thread:
void ThreadMain() override { server_->Run(delegate_); }
ExceptionHandlerServer* server_;
ExceptionHandlerServer::Delegate* delegate_;
DISALLOW_COPY_AND_ASSIGN(RunServerThread);
};
// During destruction, ensures that the server is stopped and the background
// thread joined.
class ScopedStopServerAndJoinThread {
public:
ScopedStopServerAndJoinThread(ExceptionHandlerServer* server, Thread* thread)
: server_(server), thread_(thread) {}
~ScopedStopServerAndJoinThread() {
server_->Stop();
thread_->Join();
}
private:
ExceptionHandlerServer* server_;
Thread* thread_;
DISALLOW_COPY_AND_ASSIGN(ScopedStopServerAndJoinThread);
};
class CrashingDelegate : public ExceptionHandlerServer::Delegate {
public:
CrashingDelegate(HANDLE server_ready, HANDLE completed_test_event)
: server_ready_(server_ready),
completed_test_event_(completed_test_event),
break_near_(0) {}
~CrashingDelegate() {}
void set_break_near(WinVMAddress break_near) { break_near_ = break_near; }
void ExceptionHandlerServerStarted() override { SetEvent(server_ready_); }
unsigned int ExceptionHandlerServerException(
HANDLE process,
WinVMAddress exception_information_address,
WinVMAddress debug_critical_section_address) override {
ScopedProcessSuspend suspend(process);
ProcessSnapshotWin snapshot;
snapshot.Initialize(process,
ProcessSuspensionState::kSuspended,
exception_information_address,
debug_critical_section_address);
// Confirm the exception record was read correctly.
EXPECT_NE(snapshot.Exception()->ThreadID(), 0u);
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(EXCEPTION_BREAKPOINT, snapshot.Exception()->Exception());
// Verify the exception happened at the expected location with a bit of
// slop space to allow for reading the current PC before the exception
// happens. See TestCrashingChild().
constexpr uint64_t kAllowedOffset = 100;
EXPECT_GT(snapshot.Exception()->ExceptionAddress(), break_near_);
EXPECT_LT(snapshot.Exception()->ExceptionAddress(),
break_near_ + kAllowedOffset);
SetEvent(completed_test_event_);
return snapshot.Exception()->Exception();
}
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
private:
HANDLE server_ready_; // weak
HANDLE completed_test_event_; // weak
WinVMAddress break_near_;
DISALLOW_COPY_AND_ASSIGN(CrashingDelegate);
};
void TestCrashingChild(TestPaths::Architecture architecture) {
// Set up the registration server on a background thread.
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
ScopedKernelHANDLE server_ready(CreateEvent(nullptr, false, false, nullptr));
ASSERT_TRUE(server_ready.is_valid()) << ErrorMessage("CreateEvent");
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
ScopedKernelHANDLE completed(CreateEvent(nullptr, false, false, nullptr));
ASSERT_TRUE(completed.is_valid()) << ErrorMessage("CreateEvent");
CrashingDelegate delegate(server_ready.get(), completed.get());
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
ExceptionHandlerServer exception_handler_server(true);
std::wstring pipe_name(L"\\\\.\\pipe\\test_name");
exception_handler_server.SetPipeName(pipe_name);
RunServerThread server_thread(&exception_handler_server, &delegate);
server_thread.Start();
ScopedStopServerAndJoinThread scoped_stop_server_and_join_thread(
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
&exception_handler_server, &server_thread);
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(WaitForSingleObject(server_ready.get(), INFINITE), WAIT_OBJECT_0)
<< ErrorMessage("WaitForSingleObject");
// Spawn a child process, passing it the pipe name to connect to.
base::FilePath child_test_executable =
TestPaths::BuildArtifact(L"snapshot",
L"crashing_child",
TestPaths::FileType::kExecutable,
architecture);
ChildLauncher child(child_test_executable, pipe_name);
ASSERT_NO_FATAL_FAILURE(child.Start());
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
// The child tells us (approximately) where it will crash.
WinVMAddress break_near_address;
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
LoggingReadFileExactly(child.stdout_read_handle(),
&break_near_address,
sizeof(break_near_address));
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
delegate.set_break_near(break_near_address);
win: Crash handler server This replaces the registration server, and adds dispatch to a delegate on crash requests. (As you are already aware) we went around in circles on trying to come up with a slightly-too-fancy threading design. All of them seemed to have problems when it comes to out of order events, and orderly shutdown, so I've gone back to something not-too-fancy. Two named pipe instances (that clients connect to) are created. These are used only for registration (which should take <1ms), so 2 should be sufficient to avoid any waits. When a client registers, we duplicate an event to it, which is used to signal when it wants a dump taken. The server registers threadpool waits on that event, and also on the process handle (which will be signalled when the client process exits). These requests (in particular the taking of the dump) are serviced on the threadpool, which avoids us needing to manage those threads, but still allows parallelism in taking dumps. On process termination, we use an IO Completion Port to post a message back to the main thread to request cleanup. This complexity is necessary so that we can unregister the threadpool waits without being on the threadpool, which we need to do synchronously so that we can be sure that no further callbacks will execute (and expect to have the client data around still). In a followup, I will readd support for DumpWithoutCrashing -- I don't think it will be too difficult now that we have an orderly way to clean up client records in the server. R=cpu@chromium.org, mark@chromium.org, jschuh@chromium.org BUG=crashpad:1,crashpad:45 Review URL: https://codereview.chromium.org/1301853002 .
2015-09-03 11:06:17 -07:00
// Wait for the child to crash and the exception information to be validated.
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(WaitForSingleObject(completed.get(), INFINITE), WAIT_OBJECT_0)
<< ErrorMessage("WaitForSingleObject");
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(child.WaitForExit(), EXCEPTION_BREAKPOINT);
}
#if defined(ADDRESS_SANITIZER)
// https://crbug.com/845011
#define MAYBE_ChildCrash DISABLED_ChildCrash
#else
#define MAYBE_ChildCrash ChildCrash
#endif
TEST(ExceptionSnapshotWinTest, MAYBE_ChildCrash) {
TestCrashingChild(TestPaths::Architecture::kDefault);
}
#if defined(ARCH_CPU_64_BITS)
TEST(ExceptionSnapshotWinTest, ChildCrashWOW64) {
if (!TestPaths::Has32BitBuildArtifacts()) {
win: Dynamically disable WoW64 tests absent explicit 32-bit build output Rather than having the 64-bit build assume that it lives in out\{Debug,Release}_x64 and that it can find 32-bit build output in out\{Debug,Release}, require the location of 32-bit build output to be provided explicitly via the CRASHPAD_TEST_32_BIT_OUTPUT environment variable. If this variable is not set, 64-bit tests that require 32-bit test build output will dynamically disable themselves at runtime. In order for this to work, a new DISABLED_TEST() macro is added to support dynamically disabled tests. gtest does not have its own first-class support for this (https://groups.google.com/d/topic/googletestframework/Nwh3u7YFuN4, https://github.com/google/googletest/issues/490) so this local solution is used instead. For tests via Crashpad’s own build\run_tests.py, which is how Crashpad’s own buildbots and trybots invoke tests, CRASHPAD_TEST_32_BIT_OUTPUT is set to a locaton compatible with the paths expected for the GYP-based build. No test coverage is lost on Crashpad’s own buildbots and trybots. For Crashpad tests in Chromium’s buildbots and trybots, this environment variable will not be set, causing these tests to be dynamically disabled. Bug: crashpad:203, chromium:743139, chromium:777924 Change-Id: I3c0de2bf4f835e13ed5a4adda5760d6fed508126 Reviewed-on: https://chromium-review.googlesource.com/739795 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-10-26 13:48:01 -04:00
DISABLED_TEST();
}
TestCrashingChild(TestPaths::Architecture::k32Bit);
}
#endif // ARCH_CPU_64_BITS
class SimulateDelegate : public ExceptionHandlerServer::Delegate {
public:
SimulateDelegate(HANDLE server_ready, HANDLE completed_test_event)
: server_ready_(server_ready),
completed_test_event_(completed_test_event),
dump_near_(0) {}
~SimulateDelegate() {}
void set_dump_near(WinVMAddress dump_near) { dump_near_ = dump_near; }
void ExceptionHandlerServerStarted() override { SetEvent(server_ready_); }
unsigned int ExceptionHandlerServerException(
HANDLE process,
WinVMAddress exception_information_address,
WinVMAddress debug_critical_section_address) override {
ScopedProcessSuspend suspend(process);
ProcessSnapshotWin snapshot;
snapshot.Initialize(process,
ProcessSuspensionState::kSuspended,
exception_information_address,
debug_critical_section_address);
EXPECT_TRUE(snapshot.Exception());
EXPECT_EQ(snapshot.Exception()->Exception(), 0x517a7edu);
// Verify the dump was captured at the expected location with some slop
// space.
#if defined(ADDRESS_SANITIZER)
// ASan instrumentation inserts more instructions between the expected
// location and what's reported. https://crbug.com/845011.
constexpr uint64_t kAllowedOffset = 500;
#else
constexpr uint64_t kAllowedOffset = 100;
#endif
EXPECT_GT(snapshot.Exception()->Context()->InstructionPointer(),
dump_near_);
EXPECT_LT(snapshot.Exception()->Context()->InstructionPointer(),
dump_near_ + kAllowedOffset);
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(snapshot.Exception()->ExceptionAddress(),
snapshot.Exception()->Context()->InstructionPointer());
SetEvent(completed_test_event_);
return 0;
}
private:
HANDLE server_ready_; // weak
HANDLE completed_test_event_; // weak
WinVMAddress dump_near_;
DISALLOW_COPY_AND_ASSIGN(SimulateDelegate);
};
void TestDumpWithoutCrashingChild(TestPaths::Architecture architecture) {
// Set up the registration server on a background thread.
ScopedKernelHANDLE server_ready(CreateEvent(nullptr, false, false, nullptr));
ASSERT_TRUE(server_ready.is_valid()) << ErrorMessage("CreateEvent");
ScopedKernelHANDLE completed(CreateEvent(nullptr, false, false, nullptr));
ASSERT_TRUE(completed.is_valid()) << ErrorMessage("CreateEvent");
SimulateDelegate delegate(server_ready.get(), completed.get());
ExceptionHandlerServer exception_handler_server(true);
std::wstring pipe_name(L"\\\\.\\pipe\\test_name");
exception_handler_server.SetPipeName(pipe_name);
RunServerThread server_thread(&exception_handler_server, &delegate);
server_thread.Start();
ScopedStopServerAndJoinThread scoped_stop_server_and_join_thread(
&exception_handler_server, &server_thread);
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(WaitForSingleObject(server_ready.get(), INFINITE), WAIT_OBJECT_0)
<< ErrorMessage("WaitForSingleObject");
// Spawn a child process, passing it the pipe name to connect to.
base::FilePath child_test_executable =
TestPaths::BuildArtifact(L"snapshot",
L"dump_without_crashing",
TestPaths::FileType::kExecutable,
architecture);
ChildLauncher child(child_test_executable, pipe_name);
ASSERT_NO_FATAL_FAILURE(child.Start());
// The child tells us (approximately) where it will capture a dump.
WinVMAddress dump_near_address;
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
LoggingReadFileExactly(child.stdout_read_handle(),
&dump_near_address,
sizeof(dump_near_address));
delegate.set_dump_near(dump_near_address);
// Wait for the child to crash and the exception information to be validated.
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(WaitForSingleObject(completed.get(), INFINITE), WAIT_OBJECT_0)
<< ErrorMessage("WaitForSingleObject");
EXPECT_EQ(child.WaitForExit(), 0u);
}
#if defined(ADDRESS_SANITIZER)
// https://crbug.com/845011
#define MAYBE_ChildDumpWithoutCrashing DISABLED_ChildDumpWithoutCrashing
#else
#define MAYBE_ChildDumpWithoutCrashing ChildDumpWithoutCrashing
#endif
TEST(SimulateCrash, MAYBE_ChildDumpWithoutCrashing) {
TestDumpWithoutCrashingChild(TestPaths::Architecture::kDefault);
}
#if defined(ARCH_CPU_64_BITS)
TEST(SimulateCrash, ChildDumpWithoutCrashingWOW64) {
if (!TestPaths::Has32BitBuildArtifacts()) {
win: Dynamically disable WoW64 tests absent explicit 32-bit build output Rather than having the 64-bit build assume that it lives in out\{Debug,Release}_x64 and that it can find 32-bit build output in out\{Debug,Release}, require the location of 32-bit build output to be provided explicitly via the CRASHPAD_TEST_32_BIT_OUTPUT environment variable. If this variable is not set, 64-bit tests that require 32-bit test build output will dynamically disable themselves at runtime. In order for this to work, a new DISABLED_TEST() macro is added to support dynamically disabled tests. gtest does not have its own first-class support for this (https://groups.google.com/d/topic/googletestframework/Nwh3u7YFuN4, https://github.com/google/googletest/issues/490) so this local solution is used instead. For tests via Crashpad’s own build\run_tests.py, which is how Crashpad’s own buildbots and trybots invoke tests, CRASHPAD_TEST_32_BIT_OUTPUT is set to a locaton compatible with the paths expected for the GYP-based build. No test coverage is lost on Crashpad’s own buildbots and trybots. For Crashpad tests in Chromium’s buildbots and trybots, this environment variable will not be set, causing these tests to be dynamically disabled. Bug: crashpad:203, chromium:743139, chromium:777924 Change-Id: I3c0de2bf4f835e13ed5a4adda5760d6fed508126 Reviewed-on: https://chromium-review.googlesource.com/739795 Commit-Queue: Mark Mentovai <mark@chromium.org> Reviewed-by: Scott Graham <scottmg@chromium.org>
2017-10-26 13:48:01 -04:00
DISABLED_TEST();
}
TestDumpWithoutCrashingChild(TestPaths::Architecture::k32Bit);
}
#endif // ARCH_CPU_64_BITS
} // namespace
} // namespace test
} // namespace crashpad