The C based gRPC (C++, Python, Ruby, Objective-C, PHP, C#) https://grpc.io/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

2331 lines
81 KiB

/*
*
* Copyright 2015 gRPC 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 <mutex>
#include <thread>
#include "absl/memory/memory.h"
#include "absl/strings/ascii.h"
#include "absl/strings/match.h"
#include "absl/strings/str_format.h"
#include <grpc/grpc.h>
#include <grpc/support/alloc.h>
#include <grpc/support/log.h>
#include <grpc/support/time.h>
#include <grpcpp/channel.h>
#include <grpcpp/client_context.h>
#include <grpcpp/create_channel.h>
#include <grpcpp/resource_quota.h>
#include <grpcpp/security/auth_metadata_processor.h>
#include <grpcpp/security/credentials.h>
#include <grpcpp/security/server_credentials.h>
#include <grpcpp/server.h>
#include <grpcpp/server_builder.h>
#include <grpcpp/server_context.h>
#include <grpcpp/support/string_ref.h>
#include <grpcpp/test/channel_test_peer.h>
#include "src/core/ext/filters/client_channel/backup_poller.h"
#include "src/core/lib/gprpp/env.h"
#include "src/core/lib/iomgr/iomgr.h"
#include "src/core/lib/security/credentials/credentials.h"
#include "src/proto/grpc/testing/duplicate/echo_duplicate.grpc.pb.h"
#include "src/proto/grpc/testing/echo.grpc.pb.h"
#include "test/core/util/port.h"
#include "test/core/util/test_config.h"
#include "test/cpp/end2end/interceptors_util.h"
#include "test/cpp/end2end/test_service_impl.h"
#include "test/cpp/util/string_ref_helper.h"
#include "test/cpp/util/test_credentials_provider.h"
#ifdef GRPC_POSIX_SOCKET_EV
#include "src/core/lib/iomgr/ev_posix.h"
#endif // GRPC_POSIX_SOCKET_EV
#include <gtest/gtest.h>
using std::chrono::system_clock;
namespace grpc {
namespace testing {
namespace {
bool CheckIsLocalhost(const std::string& addr) {
const std::string kIpv6("ipv6:%5B::1%5D:");
const std::string kIpv4MappedIpv6("ipv6:%5B::ffff:127.0.0.1%5D:");
const std::string kIpv4("ipv4:127.0.0.1:");
return addr.substr(0, kIpv4.size()) == kIpv4 ||
addr.substr(0, kIpv4MappedIpv6.size()) == kIpv4MappedIpv6 ||
addr.substr(0, kIpv6.size()) == kIpv6;
}
const int kClientChannelBackupPollIntervalMs = 200;
const char kTestCredsPluginErrorMsg[] = "Could not find plugin metadata.";
const char kFakeToken[] = "fake_token";
const char kFakeSelector[] = "fake_selector";
const char kExpectedFakeCredsDebugString[] =
"SecureCallCredentials{GoogleIAMCredentials{Token:present,"
"AuthoritySelector:fake_selector}}";
const char kWrongToken[] = "wrong_token";
const char kWrongSelector[] = "wrong_selector";
const char kExpectedWrongCredsDebugString[] =
"SecureCallCredentials{GoogleIAMCredentials{Token:present,"
"AuthoritySelector:wrong_selector}}";
const char kFakeToken1[] = "fake_token1";
const char kFakeSelector1[] = "fake_selector1";
const char kExpectedFakeCreds1DebugString[] =
"SecureCallCredentials{GoogleIAMCredentials{Token:present,"
"AuthoritySelector:fake_selector1}}";
const char kFakeToken2[] = "fake_token2";
const char kFakeSelector2[] = "fake_selector2";
const char kExpectedFakeCreds2DebugString[] =
"SecureCallCredentials{GoogleIAMCredentials{Token:present,"
"AuthoritySelector:fake_selector2}}";
const char kExpectedAuthMetadataPluginKeyFailureCredsDebugString[] =
"SecureCallCredentials{TestMetadataCredentials{key:TestPluginMetadata,"
"value:Does not matter, will fail the key is invalid.}}";
const char kExpectedAuthMetadataPluginValueFailureCredsDebugString[] =
"SecureCallCredentials{TestMetadataCredentials{key:test-plugin-metadata,"
"value:With illegal \n value.}}";
const char kExpectedAuthMetadataPluginWithDeadlineCredsDebugString[] =
"SecureCallCredentials{TestMetadataCredentials{key:meta_key,value:Does not "
"matter}}";
const char kExpectedNonBlockingAuthMetadataPluginFailureCredsDebugString[] =
"SecureCallCredentials{TestMetadataCredentials{key:test-plugin-metadata,"
"value:Does not matter, will fail anyway (see 3rd param)}}";
const char
kExpectedNonBlockingAuthMetadataPluginAndProcessorSuccessCredsDebugString
[] = "SecureCallCredentials{TestMetadataCredentials{key:test-plugin-"
"metadata,value:Dr Jekyll}}";
const char
kExpectedNonBlockingAuthMetadataPluginAndProcessorFailureCredsDebugString
[] = "SecureCallCredentials{TestMetadataCredentials{key:test-plugin-"
"metadata,value:Mr Hyde}}";
const char kExpectedBlockingAuthMetadataPluginFailureCredsDebugString[] =
"SecureCallCredentials{TestMetadataCredentials{key:test-plugin-metadata,"
"value:Does not matter, will fail anyway (see 3rd param)}}";
const char kExpectedCompositeCallCredsDebugString[] =
"SecureCallCredentials{CompositeCallCredentials{TestMetadataCredentials{"
"key:call-creds-key1,value:call-creds-val1},TestMetadataCredentials{key:"
"call-creds-key2,value:call-creds-val2}}}";
class TestMetadataCredentialsPlugin : public MetadataCredentialsPlugin {
public:
static const char kGoodMetadataKey[];
static const char kBadMetadataKey[];
TestMetadataCredentialsPlugin(const grpc::string_ref& metadata_key,
const grpc::string_ref& metadata_value,
bool is_blocking, bool is_successful,
int delay_ms)
: metadata_key_(metadata_key.data(), metadata_key.length()),
metadata_value_(metadata_value.data(), metadata_value.length()),
is_blocking_(is_blocking),
is_successful_(is_successful),
delay_ms_(delay_ms) {}
bool IsBlocking() const override { return is_blocking_; }
Status GetMetadata(
grpc::string_ref service_url, grpc::string_ref method_name,
const grpc::AuthContext& channel_auth_context,
std::multimap<std::string, std::string>* metadata) override {
if (delay_ms_ != 0) {
gpr_sleep_until(
gpr_time_add(gpr_now(GPR_CLOCK_REALTIME),
gpr_time_from_millis(delay_ms_, GPR_TIMESPAN)));
}
EXPECT_GT(service_url.length(), 0UL);
EXPECT_GT(method_name.length(), 0UL);
EXPECT_TRUE(channel_auth_context.IsPeerAuthenticated());
EXPECT_TRUE(metadata != nullptr);
if (is_successful_) {
metadata->insert(std::make_pair(metadata_key_, metadata_value_));
return Status::OK;
} else {
return Status(StatusCode::NOT_FOUND, kTestCredsPluginErrorMsg);
}
}
std::string DebugString() override {
return absl::StrFormat("TestMetadataCredentials{key:%s,value:%s}",
metadata_key_.c_str(), metadata_value_.c_str());
}
private:
std::string metadata_key_;
std::string metadata_value_;
bool is_blocking_;
bool is_successful_;
int delay_ms_;
};
const char TestMetadataCredentialsPlugin::kBadMetadataKey[] =
"TestPluginMetadata";
const char TestMetadataCredentialsPlugin::kGoodMetadataKey[] =
"test-plugin-metadata";
class TestAuthMetadataProcessor : public AuthMetadataProcessor {
public:
static const char kGoodGuy[];
explicit TestAuthMetadataProcessor(bool is_blocking)
: is_blocking_(is_blocking) {}
std::shared_ptr<CallCredentials> GetCompatibleClientCreds() {
return grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kGoodMetadataKey, kGoodGuy,
is_blocking_, true, 0)));
}
std::shared_ptr<CallCredentials> GetIncompatibleClientCreds() {
return grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kGoodMetadataKey, "Mr Hyde",
is_blocking_, true, 0)));
}
// Interface implementation
bool IsBlocking() const override { return is_blocking_; }
Status Process(const InputMetadata& auth_metadata, AuthContext* context,
OutputMetadata* consumed_auth_metadata,
OutputMetadata* response_metadata) override {
EXPECT_TRUE(consumed_auth_metadata != nullptr);
EXPECT_TRUE(context != nullptr);
EXPECT_TRUE(response_metadata != nullptr);
auto auth_md =
auth_metadata.find(TestMetadataCredentialsPlugin::kGoodMetadataKey);
EXPECT_NE(auth_md, auth_metadata.end());
string_ref auth_md_value = auth_md->second;
if (auth_md_value == kGoodGuy) {
context->AddProperty(kIdentityPropName, kGoodGuy);
context->SetPeerIdentityPropertyName(kIdentityPropName);
consumed_auth_metadata->insert(std::make_pair(
string(auth_md->first.data(), auth_md->first.length()),
string(auth_md->second.data(), auth_md->second.length())));
return Status::OK;
} else {
return Status(StatusCode::UNAUTHENTICATED,
string("Invalid principal: ") +
string(auth_md_value.data(), auth_md_value.length()));
}
}
private:
static const char kIdentityPropName[];
bool is_blocking_;
};
const char TestAuthMetadataProcessor::kGoodGuy[] = "Dr Jekyll";
const char TestAuthMetadataProcessor::kIdentityPropName[] = "novel identity";
class Proxy : public grpc::testing::EchoTestService::Service {
public:
explicit Proxy(const std::shared_ptr<Channel>& channel)
: stub_(grpc::testing::EchoTestService::NewStub(channel)) {}
Status Echo(ServerContext* server_context, const EchoRequest* request,
EchoResponse* response) override {
std::unique_ptr<ClientContext> client_context =
ClientContext::FromServerContext(*server_context);
return stub_->Echo(client_context.get(), *request, response);
}
private:
std::unique_ptr<grpc::testing::EchoTestService::Stub> stub_;
};
class TestServiceImplDupPkg
: public grpc::testing::duplicate::EchoTestService::Service {
public:
Status Echo(ServerContext* /*context*/, const EchoRequest* /*request*/,
EchoResponse* response) override {
response->set_message("no package");
return Status::OK;
}
};
class TestScenario {
public:
TestScenario(bool use_interceptors, bool use_proxy, bool inproc,
const std::string& credentials_type, bool callback_server)
: use_interceptors_(use_interceptors),
use_proxy_(use_proxy),
inproc_(inproc),
credentials_type_(credentials_type),
callback_server_(callback_server) {}
bool use_interceptors() const { return use_interceptors_; }
bool use_proxy() const { return use_proxy_; }
bool inproc() const { return inproc_; }
const std::string& credentials_type() const { return credentials_type_; }
bool callback_server() const { return callback_server_; }
std::string AsString() const;
static std::string Name(const ::testing::TestParamInfo<TestScenario>& info) {
return info.param.AsString();
}
private:
bool use_interceptors_;
bool use_proxy_;
bool inproc_;
const std::string credentials_type_;
bool callback_server_;
};
std::string TestScenario::AsString() const {
std::string retval = use_interceptors_ ? "Interceptor" : "";
if (use_proxy_) retval += "Proxy";
if (inproc_) retval += "Inproc";
if (callback_server_) retval += "CallbackServer";
if (credentials_type_ == kInsecureCredentialsType) {
retval += "Insecure";
} else {
std::string creds_type = absl::AsciiStrToLower(credentials_type_);
if (!creds_type.empty()) creds_type[0] = absl::ascii_toupper(creds_type[0]);
retval += creds_type;
}
return retval;
}
class End2endTest : public ::testing::TestWithParam<TestScenario> {
protected:
static void SetUpTestSuite() { grpc_init(); }
static void TearDownTestSuite() { grpc_shutdown(); }
End2endTest()
: is_server_started_(false),
kMaxMessageSize_(8192),
special_service_("special"),
first_picked_port_(0) {}
void TearDown() override {
if (is_server_started_) {
server_->Shutdown();
if (proxy_server_) proxy_server_->Shutdown();
}
if (first_picked_port_ > 0) {
grpc_recycle_unused_port(first_picked_port_);
}
}
void StartServer(const std::shared_ptr<AuthMetadataProcessor>& processor) {
int port = grpc_pick_unused_port_or_die();
first_picked_port_ = port;
server_address_ << "localhost:" << port;
// Setup server
BuildAndStartServer(processor);
}
void RestartServer(const std::shared_ptr<AuthMetadataProcessor>& processor) {
if (is_server_started_) {
server_->Shutdown();
BuildAndStartServer(processor);
}
}
void BuildAndStartServer(
const std::shared_ptr<AuthMetadataProcessor>& processor) {
ServerBuilder builder;
ConfigureServerBuilder(&builder);
auto server_creds = GetCredentialsProvider()->GetServerCredentials(
GetParam().credentials_type());
if (GetParam().credentials_type() != kInsecureCredentialsType) {
server_creds->SetAuthMetadataProcessor(processor);
}
if (GetParam().use_interceptors()) {
std::vector<
std::unique_ptr<experimental::ServerInterceptorFactoryInterface>>
creators;
// Add 20 phony server interceptors
creators.reserve(20);
for (auto i = 0; i < 20; i++) {
creators.push_back(std::make_unique<PhonyInterceptorFactory>());
}
builder.experimental().SetInterceptorCreators(std::move(creators));
}
builder.AddListeningPort(server_address_.str(), server_creds);
if (!GetParam().callback_server()) {
builder.RegisterService(&service_);
} else {
builder.RegisterService(&callback_service_);
}
builder.RegisterService("foo.test.youtube.com", &special_service_);
builder.RegisterService(&dup_pkg_service_);
builder.SetSyncServerOption(ServerBuilder::SyncServerOption::NUM_CQS, 4);
builder.SetSyncServerOption(
ServerBuilder::SyncServerOption::CQ_TIMEOUT_MSEC, 10);
server_ = builder.BuildAndStart();
is_server_started_ = true;
}
virtual void ConfigureServerBuilder(ServerBuilder* builder) {
builder->SetMaxMessageSize(
kMaxMessageSize_); // For testing max message size.
}
void ResetChannel(
std::vector<
std::unique_ptr<experimental::ClientInterceptorFactoryInterface>>
interceptor_creators = {}) {
if (!is_server_started_) {
StartServer(std::shared_ptr<AuthMetadataProcessor>());
}
EXPECT_TRUE(is_server_started_);
ChannelArguments args;
auto channel_creds = GetCredentialsProvider()->GetChannelCredentials(
GetParam().credentials_type(), &args);
if (!user_agent_prefix_.empty()) {
args.SetUserAgentPrefix(user_agent_prefix_);
}
args.SetString(GRPC_ARG_SECONDARY_USER_AGENT_STRING, "end2end_test");
if (!GetParam().inproc()) {
if (!GetParam().use_interceptors()) {
channel_ = grpc::CreateCustomChannel(server_address_.str(),
channel_creds, args);
} else {
channel_ = CreateCustomChannelWithInterceptors(
server_address_.str(), channel_creds, args,
interceptor_creators.empty() ? CreatePhonyClientInterceptors()
: std::move(interceptor_creators));
}
} else {
if (!GetParam().use_interceptors()) {
channel_ = server_->InProcessChannel(args);
} else {
channel_ = server_->experimental().InProcessChannelWithInterceptors(
args, interceptor_creators.empty()
? CreatePhonyClientInterceptors()
: std::move(interceptor_creators));
}
}
}
void ResetStub(
std::vector<
std::unique_ptr<experimental::ClientInterceptorFactoryInterface>>
interceptor_creators = {}) {
ResetChannel(std::move(interceptor_creators));
if (GetParam().use_proxy()) {
proxy_service_ = std::make_unique<Proxy>(channel_);
int port = grpc_pick_unused_port_or_die();
std::ostringstream proxyaddr;
proxyaddr << "localhost:" << port;
ServerBuilder builder;
builder.AddListeningPort(proxyaddr.str(), InsecureServerCredentials());
builder.RegisterService(proxy_service_.get());
builder.SetSyncServerOption(ServerBuilder::SyncServerOption::NUM_CQS, 4);
builder.SetSyncServerOption(
ServerBuilder::SyncServerOption::CQ_TIMEOUT_MSEC, 10);
proxy_server_ = builder.BuildAndStart();
channel_ =
grpc::CreateChannel(proxyaddr.str(), InsecureChannelCredentials());
}
stub_ = grpc::testing::EchoTestService::NewStub(channel_);
PhonyInterceptor::Reset();
}
bool is_server_started_;
std::shared_ptr<Channel> channel_;
std::unique_ptr<grpc::testing::EchoTestService::Stub> stub_;
std::unique_ptr<Server> server_;
std::unique_ptr<Server> proxy_server_;
std::unique_ptr<Proxy> proxy_service_;
std::ostringstream server_address_;
const int kMaxMessageSize_;
TestServiceImpl service_;
CallbackTestServiceImpl callback_service_;
TestServiceImpl special_service_;
TestServiceImplDupPkg dup_pkg_service_;
std::string user_agent_prefix_;
int first_picked_port_;
};
void SendRpc(grpc::testing::EchoTestService::Stub* stub, int num_rpcs,
bool with_binary_metadata) {
EchoRequest request;
EchoResponse response;
request.set_message("Hello hello hello hello");
for (int i = 0; i < num_rpcs; ++i) {
ClientContext context;
if (with_binary_metadata) {
char bytes[8] = {'\0', '\1', '\2', '\3',
'\4', '\5', '\6', static_cast<char>(i)};
context.AddMetadata("custom-bin", std::string(bytes, 8));
}
context.set_compression_algorithm(GRPC_COMPRESS_GZIP);
Status s = stub->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
}
}
// This class is for testing scenarios where RPCs are cancelled on the server
// by calling ServerContext::TryCancel()
class End2endServerTryCancelTest : public End2endTest {
protected:
// Helper for testing client-streaming RPCs which are cancelled on the server.
// Depending on the value of server_try_cancel parameter, this will test one
// of the following three scenarios:
// CANCEL_BEFORE_PROCESSING: Rpc is cancelled by the server before reading
// any messages from the client
//
// CANCEL_DURING_PROCESSING: Rpc is cancelled by the server while reading
// messages from the client
//
// CANCEL_AFTER PROCESSING: Rpc is cancelled by server after reading all
// the messages from the client
//
// NOTE: Do not call this function with server_try_cancel == DO_NOT_CANCEL.
void TestRequestStreamServerCancel(
ServerTryCancelRequestPhase server_try_cancel, int num_msgs_to_send) {
RestartServer(std::shared_ptr<AuthMetadataProcessor>());
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
// Send server_try_cancel value in the client metadata
context.AddMetadata(kServerTryCancelRequest,
std::to_string(server_try_cancel));
auto stream = stub_->RequestStream(&context, &response);
int num_msgs_sent = 0;
while (num_msgs_sent < num_msgs_to_send) {
request.set_message("hello");
if (!stream->Write(request)) {
break;
}
num_msgs_sent++;
}
gpr_log(GPR_INFO, "Sent %d messages", num_msgs_sent);
stream->WritesDone();
Status s = stream->Finish();
// At this point, we know for sure that RPC was cancelled by the server
// since we passed server_try_cancel value in the metadata. Depending on the
// value of server_try_cancel, the RPC might have been cancelled by the
// server at different stages. The following validates our expectations of
// number of messages sent in various cancellation scenarios:
switch (server_try_cancel) {
case CANCEL_BEFORE_PROCESSING:
case CANCEL_DURING_PROCESSING:
// If the RPC is cancelled by server before / during messages from the
// client, it means that the client most likely did not get a chance to
// send all the messages it wanted to send. i.e num_msgs_sent <=
// num_msgs_to_send
EXPECT_LE(num_msgs_sent, num_msgs_to_send);
break;
case CANCEL_AFTER_PROCESSING:
// If the RPC was cancelled after all messages were read by the server,
// the client did get a chance to send all its messages
EXPECT_EQ(num_msgs_sent, num_msgs_to_send);
break;
default:
gpr_log(GPR_ERROR, "Invalid server_try_cancel value: %d",
server_try_cancel);
EXPECT_TRUE(server_try_cancel > DO_NOT_CANCEL &&
server_try_cancel <= CANCEL_AFTER_PROCESSING);
break;
}
EXPECT_FALSE(s.ok());
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
// Make sure that the server interceptors were notified
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
// Helper for testing server-streaming RPCs which are cancelled on the server.
// Depending on the value of server_try_cancel parameter, this will test one
// of the following three scenarios:
// CANCEL_BEFORE_PROCESSING: Rpc is cancelled by the server before writing
// any messages to the client
//
// CANCEL_DURING_PROCESSING: Rpc is cancelled by the server while writing
// messages to the client
//
// CANCEL_AFTER PROCESSING: Rpc is cancelled by server after writing all
// the messages to the client
//
// NOTE: Do not call this function with server_try_cancel == DO_NOT_CANCEL.
void TestResponseStreamServerCancel(
ServerTryCancelRequestPhase server_try_cancel) {
RestartServer(std::shared_ptr<AuthMetadataProcessor>());
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
// Send server_try_cancel in the client metadata
context.AddMetadata(kServerTryCancelRequest,
std::to_string(server_try_cancel));
request.set_message("hello");
auto stream = stub_->ResponseStream(&context, request);
int num_msgs_read = 0;
while (num_msgs_read < kServerDefaultResponseStreamsToSend) {
if (!stream->Read(&response)) {
break;
}
EXPECT_EQ(response.message(),
request.message() + std::to_string(num_msgs_read));
num_msgs_read++;
}
gpr_log(GPR_INFO, "Read %d messages", num_msgs_read);
Status s = stream->Finish();
// Depending on the value of server_try_cancel, the RPC might have been
// cancelled by the server at different stages. The following validates our
// expectations of number of messages read in various cancellation
// scenarios:
switch (server_try_cancel) {
case CANCEL_BEFORE_PROCESSING:
// Server cancelled before sending any messages. Which means the client
// wouldn't have read any
EXPECT_EQ(num_msgs_read, 0);
break;
case CANCEL_DURING_PROCESSING:
// Server cancelled while writing messages. Client must have read less
// than or equal to the expected number of messages
EXPECT_LE(num_msgs_read, kServerDefaultResponseStreamsToSend);
break;
case CANCEL_AFTER_PROCESSING:
// Even though the Server cancelled after writing all messages, the RPC
// may be cancelled before the Client got a chance to read all the
// messages.
EXPECT_LE(num_msgs_read, kServerDefaultResponseStreamsToSend);
break;
default: {
gpr_log(GPR_ERROR, "Invalid server_try_cancel value: %d",
server_try_cancel);
EXPECT_TRUE(server_try_cancel > DO_NOT_CANCEL &&
server_try_cancel <= CANCEL_AFTER_PROCESSING);
break;
}
}
EXPECT_FALSE(s.ok());
// Make sure that the server interceptors were notified
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
// Helper for testing bidirectional-streaming RPCs which are cancelled on the
// server. Depending on the value of server_try_cancel parameter, this will
// test one of the following three scenarios:
// CANCEL_BEFORE_PROCESSING: Rpc is cancelled by the server before reading/
// writing any messages from/to the client
//
// CANCEL_DURING_PROCESSING: Rpc is cancelled by the server while reading/
// writing messages from/to the client
//
// CANCEL_AFTER PROCESSING: Rpc is cancelled by server after reading/writing
// all the messages from/to the client
//
// NOTE: Do not call this function with server_try_cancel == DO_NOT_CANCEL.
void TestBidiStreamServerCancel(ServerTryCancelRequestPhase server_try_cancel,
int num_messages) {
RestartServer(std::shared_ptr<AuthMetadataProcessor>());
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
// Send server_try_cancel in the client metadata
context.AddMetadata(kServerTryCancelRequest,
std::to_string(server_try_cancel));
auto stream = stub_->BidiStream(&context);
int num_msgs_read = 0;
int num_msgs_sent = 0;
while (num_msgs_sent < num_messages) {
request.set_message("hello " + std::to_string(num_msgs_sent));
if (!stream->Write(request)) {
break;
}
num_msgs_sent++;
if (!stream->Read(&response)) {
break;
}
num_msgs_read++;
EXPECT_EQ(response.message(), request.message());
}
gpr_log(GPR_INFO, "Sent %d messages", num_msgs_sent);
gpr_log(GPR_INFO, "Read %d messages", num_msgs_read);
stream->WritesDone();
Status s = stream->Finish();
// Depending on the value of server_try_cancel, the RPC might have been
// cancelled by the server at different stages. The following validates our
// expectations of number of messages read in various cancellation
// scenarios:
switch (server_try_cancel) {
case CANCEL_BEFORE_PROCESSING:
EXPECT_EQ(num_msgs_read, 0);
break;
case CANCEL_DURING_PROCESSING:
EXPECT_LE(num_msgs_sent, num_messages);
EXPECT_LE(num_msgs_read, num_msgs_sent);
break;
case CANCEL_AFTER_PROCESSING:
EXPECT_EQ(num_msgs_sent, num_messages);
// The Server cancelled after reading the last message and after writing
// the message to the client. However, the RPC cancellation might have
// taken effect before the client actually read the response.
EXPECT_LE(num_msgs_read, num_msgs_sent);
break;
default:
gpr_log(GPR_ERROR, "Invalid server_try_cancel value: %d",
server_try_cancel);
EXPECT_TRUE(server_try_cancel > DO_NOT_CANCEL &&
server_try_cancel <= CANCEL_AFTER_PROCESSING);
break;
}
EXPECT_FALSE(s.ok());
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
// Make sure that the server interceptors were notified
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
};
TEST_P(End2endServerTryCancelTest, RequestEchoServerCancel) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.AddMetadata(kServerTryCancelRequest,
std::to_string(CANCEL_BEFORE_PROCESSING));
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
}
// Server to cancel before doing reading the request
TEST_P(End2endServerTryCancelTest, RequestStreamServerCancelBeforeReads) {
TestRequestStreamServerCancel(CANCEL_BEFORE_PROCESSING, 1);
}
// Server to cancel while reading a request from the stream in parallel
TEST_P(End2endServerTryCancelTest, RequestStreamServerCancelDuringRead) {
TestRequestStreamServerCancel(CANCEL_DURING_PROCESSING, 10);
}
// Server to cancel after reading all the requests but before returning to the
// client
TEST_P(End2endServerTryCancelTest, RequestStreamServerCancelAfterReads) {
TestRequestStreamServerCancel(CANCEL_AFTER_PROCESSING, 4);
}
// Server to cancel before sending any response messages
TEST_P(End2endServerTryCancelTest, ResponseStreamServerCancelBefore) {
TestResponseStreamServerCancel(CANCEL_BEFORE_PROCESSING);
}
// Server to cancel while writing a response to the stream in parallel
TEST_P(End2endServerTryCancelTest, ResponseStreamServerCancelDuring) {
TestResponseStreamServerCancel(CANCEL_DURING_PROCESSING);
}
// Server to cancel after writing all the respones to the stream but before
// returning to the client
TEST_P(End2endServerTryCancelTest, ResponseStreamServerCancelAfter) {
TestResponseStreamServerCancel(CANCEL_AFTER_PROCESSING);
}
// Server to cancel before reading/writing any requests/responses on the stream
TEST_P(End2endServerTryCancelTest, BidiStreamServerCancelBefore) {
TestBidiStreamServerCancel(CANCEL_BEFORE_PROCESSING, 2);
}
// Server to cancel while reading/writing requests/responses on the stream in
// parallel
TEST_P(End2endServerTryCancelTest, BidiStreamServerCancelDuring) {
TestBidiStreamServerCancel(CANCEL_DURING_PROCESSING, 10);
}
// Server to cancel after reading/writing all requests/responses on the stream
// but before returning to the client
TEST_P(End2endServerTryCancelTest, BidiStreamServerCancelAfter) {
TestBidiStreamServerCancel(CANCEL_AFTER_PROCESSING, 5);
}
TEST_P(End2endTest, SimpleRpcWithCustomUserAgentPrefix) {
// User-Agent is an HTTP header for HTTP transports only
if (GetParam().inproc()) {
return;
}
user_agent_prefix_ = "custom_prefix";
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello hello hello hello");
request.mutable_param()->set_echo_metadata(true);
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
const auto& trailing_metadata = context.GetServerTrailingMetadata();
auto iter = trailing_metadata.find("user-agent");
EXPECT_TRUE(iter != trailing_metadata.end());
std::string expected_prefix = user_agent_prefix_ + " grpc-c++/";
EXPECT_TRUE(iter->second.starts_with(expected_prefix)) << iter->second;
}
TEST_P(End2endTest, MultipleRpcsWithVariedBinaryMetadataValue) {
ResetStub();
std::vector<std::thread> threads;
threads.reserve(10);
for (int i = 0; i < 10; ++i) {
threads.emplace_back(SendRpc, stub_.get(), 10, true);
}
for (int i = 0; i < 10; ++i) {
threads[i].join();
}
}
TEST_P(End2endTest, MultipleRpcs) {
ResetStub();
std::vector<std::thread> threads;
threads.reserve(10);
for (int i = 0; i < 10; ++i) {
threads.emplace_back(SendRpc, stub_.get(), 10, false);
}
for (int i = 0; i < 10; ++i) {
threads[i].join();
}
}
TEST_P(End2endTest, ManyStubs) {
ResetStub();
ChannelTestPeer peer(channel_.get());
int registered_calls_pre = peer.registered_calls();
int registration_attempts_pre = peer.registration_attempts();
for (int i = 0; i < 1000; ++i) {
grpc::testing::EchoTestService::NewStub(channel_);
}
EXPECT_EQ(peer.registered_calls(), registered_calls_pre);
EXPECT_GT(peer.registration_attempts(), registration_attempts_pre);
}
TEST_P(End2endTest, EmptyBinaryMetadata) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello hello hello hello");
ClientContext context;
context.AddMetadata("custom-bin", "");
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, AuthoritySeenOnServerSide) {
ResetStub();
EchoRequest request;
request.mutable_param()->set_echo_host_from_authority_header(true);
EchoResponse response;
request.set_message("Live long and prosper.");
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
if (GetParam().credentials_type() == kTlsCredentialsType) {
// SSL creds overrides the authority.
EXPECT_EQ("foo.test.google.fr", response.param().host());
} else if (GetParam().inproc()) {
EXPECT_EQ("inproc", response.param().host());
} else {
EXPECT_EQ(server_address_.str(), response.param().host());
}
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, ReconnectChannel) {
if (GetParam().inproc()) {
return;
}
int poller_slowdown_factor = 1;
// It needs 2 pollset_works to reconnect the channel with polling engine
// "poll"
#ifdef GRPC_POSIX_SOCKET_EV
grpc_core::UniquePtr<char> poller = GPR_GLOBAL_CONFIG_GET(grpc_poll_strategy);
if (0 == strcmp(poller.get(), "poll")) {
poller_slowdown_factor = 2;
}
#endif // GRPC_POSIX_SOCKET_EV
ResetStub();
SendRpc(stub_.get(), 1, false);
RestartServer(std::shared_ptr<AuthMetadataProcessor>());
// It needs more than GRPC_CLIENT_CHANNEL_BACKUP_POLL_INTERVAL_MS time to
// reconnect the channel. Make it a factor of 5x
gpr_sleep_until(
gpr_time_add(gpr_now(GPR_CLOCK_REALTIME),
gpr_time_from_millis(kClientChannelBackupPollIntervalMs * 5 *
poller_slowdown_factor *
grpc_test_slowdown_factor(),
GPR_TIMESPAN)));
SendRpc(stub_.get(), 1, false);
}
TEST_P(End2endTest, RequestStreamOneRequest) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
auto stream = stub_->RequestStream(&context, &response);
request.set_message("hello");
EXPECT_TRUE(stream->Write(request));
stream->WritesDone();
Status s = stream->Finish();
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
EXPECT_TRUE(context.debug_error_string().empty());
}
TEST_P(End2endTest, RequestStreamOneRequestWithCoalescingApi) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_initial_metadata_corked(true);
auto stream = stub_->RequestStream(&context, &response);
request.set_message("hello");
stream->WriteLast(request, WriteOptions());
Status s = stream->Finish();
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, RequestStreamTwoRequests) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
auto stream = stub_->RequestStream(&context, &response);
request.set_message("hello");
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Write(request));
stream->WritesDone();
Status s = stream->Finish();
EXPECT_EQ(response.message(), "hellohello");
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, RequestStreamTwoRequestsWithWriteThrough) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
auto stream = stub_->RequestStream(&context, &response);
request.set_message("hello");
EXPECT_TRUE(stream->Write(request, WriteOptions().set_write_through()));
EXPECT_TRUE(stream->Write(request, WriteOptions().set_write_through()));
stream->WritesDone();
Status s = stream->Finish();
EXPECT_EQ(response.message(), "hellohello");
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, RequestStreamTwoRequestsWithCoalescingApi) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_initial_metadata_corked(true);
auto stream = stub_->RequestStream(&context, &response);
request.set_message("hello");
EXPECT_TRUE(stream->Write(request));
stream->WriteLast(request, WriteOptions());
Status s = stream->Finish();
EXPECT_EQ(response.message(), "hellohello");
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, ResponseStream) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
auto stream = stub_->ResponseStream(&context, request);
for (int i = 0; i < kServerDefaultResponseStreamsToSend; ++i) {
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message() + std::to_string(i));
}
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, ResponseStreamWithCoalescingApi) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
context.AddMetadata(kServerUseCoalescingApi, "1");
auto stream = stub_->ResponseStream(&context, request);
for (int i = 0; i < kServerDefaultResponseStreamsToSend; ++i) {
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message() + std::to_string(i));
}
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
// This was added to prevent regression from issue:
// https://github.com/grpc/grpc/issues/11546
TEST_P(End2endTest, ResponseStreamWithEverythingCoalesced) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
context.AddMetadata(kServerUseCoalescingApi, "1");
// We will only send one message, forcing everything (init metadata, message,
// trailing) to be coalesced together.
context.AddMetadata(kServerResponseStreamsToSend, "1");
auto stream = stub_->ResponseStream(&context, request);
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message() + "0");
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, BidiStream) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
std::string msg("hello");
auto stream = stub_->BidiStream(&context);
for (int i = 0; i < kServerDefaultResponseStreamsToSend; ++i) {
request.set_message(msg + std::to_string(i));
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
}
stream->WritesDone();
EXPECT_FALSE(stream->Read(&response));
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, BidiStreamWithCoalescingApi) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.AddMetadata(kServerFinishAfterNReads, "3");
context.set_initial_metadata_corked(true);
std::string msg("hello");
auto stream = stub_->BidiStream(&context);
request.set_message(msg + "0");
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
request.set_message(msg + "1");
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
request.set_message(msg + "2");
stream->WriteLast(request, WriteOptions());
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
EXPECT_FALSE(stream->Read(&response));
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
// This was added to prevent regression from issue:
// https://github.com/grpc/grpc/issues/11546
TEST_P(End2endTest, BidiStreamWithEverythingCoalesced) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.AddMetadata(kServerFinishAfterNReads, "1");
context.set_initial_metadata_corked(true);
std::string msg("hello");
auto stream = stub_->BidiStream(&context);
request.set_message(msg + "0");
stream->WriteLast(request, WriteOptions());
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
EXPECT_FALSE(stream->Read(&response));
EXPECT_FALSE(stream->Read(&response));
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
// Talk to the two services with the same name but different package names.
// The two stubs are created on the same channel.
TEST_P(End2endTest, DiffPackageServices) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
std::unique_ptr<grpc::testing::duplicate::EchoTestService::Stub> dup_pkg_stub(
grpc::testing::duplicate::EchoTestService::NewStub(channel_));
ClientContext context2;
s = dup_pkg_stub->Echo(&context2, request, &response);
EXPECT_EQ("no package", response.message());
EXPECT_TRUE(s.ok());
}
template <class ServiceType>
void CancelRpc(ClientContext* context, int delay_us, ServiceType* service) {
gpr_sleep_until(gpr_time_add(gpr_now(GPR_CLOCK_REALTIME),
gpr_time_from_micros(delay_us, GPR_TIMESPAN)));
while (!service->signal_client()) {
}
context->TryCancel();
}
TEST_P(End2endTest, CancelRpcBeforeStart) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
context.TryCancel();
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ("", response.message());
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
TEST_P(End2endTest, CancelRpcAfterStart) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
request.mutable_param()->set_server_notify_client_when_started(true);
request.mutable_param()->set_skip_cancelled_check(true);
Status s;
std::thread echo_thread([this, &s, &context, &request, &response] {
s = stub_->Echo(&context, request, &response);
EXPECT_EQ(StatusCode::CANCELLED, s.error_code());
});
if (!GetParam().callback_server()) {
service_.ClientWaitUntilRpcStarted();
} else {
callback_service_.ClientWaitUntilRpcStarted();
}
context.TryCancel();
if (!GetParam().callback_server()) {
service_.SignalServerToContinue();
} else {
callback_service_.SignalServerToContinue();
}
echo_thread.join();
EXPECT_EQ("", response.message());
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
// Client cancels request stream after sending two messages
TEST_P(End2endTest, ClientCancelsRequestStream) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
auto stream = stub_->RequestStream(&context, &response);
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Write(request));
context.TryCancel();
Status s = stream->Finish();
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
EXPECT_EQ(response.message(), "");
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
// Client cancels server stream after sending some messages
TEST_P(End2endTest, ClientCancelsResponseStream) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("hello");
auto stream = stub_->ResponseStream(&context, request);
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message() + "0");
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message() + "1");
context.TryCancel();
// The cancellation races with responses, so there might be zero or
// one responses pending, read till failure
if (stream->Read(&response)) {
EXPECT_EQ(response.message(), request.message() + "2");
// Since we have cancelled, we expect the next attempt to read to fail
EXPECT_FALSE(stream->Read(&response));
}
Status s = stream->Finish();
// The final status could be either of CANCELLED or OK depending on
// who won the race.
EXPECT_GE(grpc::StatusCode::CANCELLED, s.error_code());
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
// Client cancels bidi stream after sending some messages
TEST_P(End2endTest, ClientCancelsBidi) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
std::string msg("hello");
// Send server_try_cancel value in the client metadata
context.AddMetadata(kClientTryCancelRequest, std::to_string(1));
auto stream = stub_->BidiStream(&context);
request.set_message(msg + "0");
EXPECT_TRUE(stream->Write(request));
EXPECT_TRUE(stream->Read(&response));
EXPECT_EQ(response.message(), request.message());
request.set_message(msg + "1");
EXPECT_TRUE(stream->Write(request));
context.TryCancel();
// The cancellation races with responses, so there might be zero or
// one responses pending, read till failure
if (stream->Read(&response)) {
EXPECT_EQ(response.message(), request.message());
// Since we have cancelled, we expect the next attempt to read to fail
EXPECT_FALSE(stream->Read(&response));
}
Status s = stream->Finish();
EXPECT_EQ(grpc::StatusCode::CANCELLED, s.error_code());
if (GetParam().use_interceptors()) {
EXPECT_EQ(20, PhonyInterceptor::GetNumTimesCancel());
}
}
TEST_P(End2endTest, RpcMaxMessageSize) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message(string(kMaxMessageSize_ * 2, 'a'));
request.mutable_param()->set_server_die(true);
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
}
void ReaderThreadFunc(ClientReaderWriter<EchoRequest, EchoResponse>* stream,
gpr_event* ev) {
EchoResponse resp;
gpr_event_set(ev, reinterpret_cast<void*>(1));
while (stream->Read(&resp)) {
gpr_log(GPR_INFO, "Read message");
}
}
// Run a Read and a WritesDone simultaneously.
TEST_P(End2endTest, SimultaneousReadWritesDone) {
ResetStub();
ClientContext context;
gpr_event ev;
gpr_event_init(&ev);
auto stream = stub_->BidiStream(&context);
std::thread reader_thread(ReaderThreadFunc, stream.get(), &ev);
gpr_event_wait(&ev, gpr_inf_future(GPR_CLOCK_REALTIME));
stream->WritesDone();
reader_thread.join();
Status s = stream->Finish();
EXPECT_TRUE(s.ok());
}
TEST_P(End2endTest, ChannelState) {
if (GetParam().inproc()) {
return;
}
ResetStub();
// Start IDLE
EXPECT_EQ(GRPC_CHANNEL_IDLE, channel_->GetState(false));
// Did not ask to connect, no state change.
CompletionQueue cq;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::milliseconds(10);
channel_->NotifyOnStateChange(GRPC_CHANNEL_IDLE, deadline, &cq, nullptr);
void* tag;
bool ok = true;
cq.Next(&tag, &ok);
EXPECT_FALSE(ok);
EXPECT_EQ(GRPC_CHANNEL_IDLE, channel_->GetState(true));
EXPECT_TRUE(channel_->WaitForStateChange(GRPC_CHANNEL_IDLE,
gpr_inf_future(GPR_CLOCK_REALTIME)));
auto state = channel_->GetState(false);
EXPECT_TRUE(state == GRPC_CHANNEL_CONNECTING || state == GRPC_CHANNEL_READY);
}
// Takes 10s.
TEST_P(End2endTest, ChannelStateTimeout) {
if ((GetParam().credentials_type() != kInsecureCredentialsType) ||
GetParam().inproc()) {
return;
}
int port = grpc_pick_unused_port_or_die();
std::ostringstream server_address;
server_address << "localhost:" << port;
// Channel to non-existing server
auto channel =
grpc::CreateChannel(server_address.str(), InsecureChannelCredentials());
// Start IDLE
EXPECT_EQ(GRPC_CHANNEL_IDLE, channel->GetState(true));
auto state = GRPC_CHANNEL_IDLE;
for (int i = 0; i < 10; i++) {
channel->WaitForStateChange(
state, std::chrono::system_clock::now() + std::chrono::seconds(1));
state = channel->GetState(false);
}
}
TEST_P(End2endTest, ChannelStateOnLameChannel) {
if ((GetParam().credentials_type() != kInsecureCredentialsType) ||
GetParam().inproc()) {
return;
}
// Channel using invalid target URI. This creates a lame channel.
auto channel = grpc::CreateChannel("dns:///", InsecureChannelCredentials());
// Channel should immediately report TRANSIENT_FAILURE.
EXPECT_EQ(GRPC_CHANNEL_TRANSIENT_FAILURE, channel->GetState(true));
// And state will never change.
auto state = GRPC_CHANNEL_TRANSIENT_FAILURE;
for (int i = 0; i < 10; ++i) {
channel->WaitForStateChange(
state, std::chrono::system_clock::now() + std::chrono::seconds(1));
state = channel->GetState(false);
}
}
// Talking to a non-existing service.
TEST_P(End2endTest, NonExistingService) {
ResetChannel();
std::unique_ptr<grpc::testing::UnimplementedEchoService::Stub> stub;
stub = grpc::testing::UnimplementedEchoService::NewStub(channel_);
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
Status s = stub->Unimplemented(&context, request, &response);
EXPECT_EQ(StatusCode::UNIMPLEMENTED, s.error_code());
EXPECT_EQ("", s.error_message());
}
// Ask the server to send back a serialized proto in trailer.
// This is an example of setting error details.
TEST_P(End2endTest, BinaryTrailerTest) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
request.mutable_param()->set_echo_metadata(true);
DebugInfo* info = request.mutable_param()->mutable_debug_info();
info->add_stack_entries("stack_entry_1");
info->add_stack_entries("stack_entry_2");
info->add_stack_entries("stack_entry_3");
info->set_detail("detailed debug info");
std::string expected_string = info->SerializeAsString();
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
auto trailers = context.GetServerTrailingMetadata();
EXPECT_EQ(1u, trailers.count(kDebugInfoTrailerKey));
auto iter = trailers.find(kDebugInfoTrailerKey);
EXPECT_EQ(expected_string, iter->second);
// Parse the returned trailer into a DebugInfo proto.
DebugInfo returned_info;
EXPECT_TRUE(returned_info.ParseFromString(ToString(iter->second)));
}
TEST_P(End2endTest, ExpectErrorTest) {
ResetStub();
std::vector<ErrorStatus> expected_status;
expected_status.emplace_back();
expected_status.back().set_code(13); // INTERNAL
// No Error message or details
expected_status.emplace_back();
expected_status.back().set_code(13); // INTERNAL
expected_status.back().set_error_message("text error message");
expected_status.back().set_binary_error_details("text error details");
expected_status.emplace_back();
expected_status.back().set_code(13); // INTERNAL
expected_status.back().set_error_message("text error message");
expected_status.back().set_binary_error_details(
"\x0\x1\x2\x3\x4\x5\x6\x8\x9\xA\xB");
for (auto iter = expected_status.begin(); iter != expected_status.end();
++iter) {
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("Hello");
auto* error = request.mutable_param()->mutable_expected_error();
error->set_code(iter->code());
error->set_error_message(iter->error_message());
error->set_binary_error_details(iter->binary_error_details());
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(iter->code(), s.error_code());
EXPECT_EQ(iter->error_message(), s.error_message());
EXPECT_EQ(iter->binary_error_details(), s.error_details());
EXPECT_TRUE(absl::StrContains(context.debug_error_string(), "created"));
#ifndef NDEBUG
// grpc_core::StatusIntProperty::kFileLine is for debug only
EXPECT_TRUE(absl::StrContains(context.debug_error_string(), "file"));
EXPECT_TRUE(absl::StrContains(context.debug_error_string(), "line"));
#endif
EXPECT_TRUE(absl::StrContains(context.debug_error_string(), "status"));
EXPECT_TRUE(absl::StrContains(context.debug_error_string(), "13"));
}
}
//////////////////////////////////////////////////////////////////////////
// Test with and without a proxy.
class ProxyEnd2endTest : public End2endTest {
protected:
};
TEST_P(ProxyEnd2endTest, SimpleRpc) {
ResetStub();
SendRpc(stub_.get(), 1, false);
}
TEST_P(ProxyEnd2endTest, SimpleRpcWithEmptyMessages) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_TRUE(s.ok());
}
TEST_P(ProxyEnd2endTest, MultipleRpcs) {
ResetStub();
std::vector<std::thread> threads;
threads.reserve(10);
for (int i = 0; i < 10; ++i) {
threads.emplace_back(SendRpc, stub_.get(), 10, false);
}
for (int i = 0; i < 10; ++i) {
threads[i].join();
}
}
// Set a 10us deadline and make sure proper error is returned.
TEST_P(ProxyEnd2endTest, RpcDeadlineExpires) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
request.mutable_param()->set_skip_cancelled_check(true);
// Let server sleep for 4 secs first to guarantee expiry.
// 4 secs might seem a bit extreme but the timer manager would have been just
// initialized (when ResetStub() was called) and there are some warmup costs
// i.e the timer thread many not have even started. There might also be other
// delays in the timer manager thread (in acquiring locks, timer data
// structure manipulations, starting backup timer threads) that add to the
// delays. 4 secs might be still not enough in some cases but this
// significantly reduces the test flakes
request.mutable_param()->set_server_sleep_us(4 * 1000 * 1000);
ClientContext context;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::milliseconds(1);
context.set_deadline(deadline);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(StatusCode::DEADLINE_EXCEEDED, s.error_code());
}
// Set a long but finite deadline.
TEST_P(ProxyEnd2endTest, RpcLongDeadline) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::hours(1);
context.set_deadline(deadline);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
}
// Ask server to echo back the deadline it sees.
TEST_P(ProxyEnd2endTest, EchoDeadline) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
request.mutable_param()->set_echo_deadline(true);
ClientContext context;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::seconds(100);
context.set_deadline(deadline);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
gpr_timespec sent_deadline;
Timepoint2Timespec(deadline, &sent_deadline);
// We want to allow some reasonable error given:
// - request_deadline() only has 1sec resolution so the best we can do is +-1
// - if sent_deadline.tv_nsec is very close to the next second's boundary we
// can end up being off by 2 in one direction.
EXPECT_LE(response.param().request_deadline() - sent_deadline.tv_sec, 2);
EXPECT_GE(response.param().request_deadline() - sent_deadline.tv_sec, -1);
}
// Ask server to echo back the deadline it sees. The rpc has no deadline.
TEST_P(ProxyEnd2endTest, EchoDeadlineForNoDeadlineRpc) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
request.mutable_param()->set_echo_deadline(true);
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
EXPECT_EQ(response.param().request_deadline(),
gpr_inf_future(GPR_CLOCK_REALTIME).tv_sec);
}
TEST_P(ProxyEnd2endTest, UnimplementedRpc) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
Status s = stub_->Unimplemented(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), grpc::StatusCode::UNIMPLEMENTED);
EXPECT_EQ(s.error_message(), "");
EXPECT_EQ(response.message(), "");
}
// Client cancels rpc after 10ms
TEST_P(ProxyEnd2endTest, ClientCancelsRpc) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
const int kCancelDelayUs = 10 * 1000;
request.mutable_param()->set_client_cancel_after_us(kCancelDelayUs);
ClientContext context;
std::thread cancel_thread;
if (!GetParam().callback_server()) {
cancel_thread = std::thread(
[&context, this](int delay) { CancelRpc(&context, delay, &service_); },
kCancelDelayUs);
// Note: the unusual pattern above (and below) is caused by a conflict
// between two sets of compiler expectations. clang allows const to be
// captured without mention, so there is no need to capture kCancelDelayUs
// (and indeed clang-tidy complains if you do so). OTOH, a Windows compiler
// in our tests requires an explicit capture even for const. We square this
// circle by passing the const value in as an argument to the lambda.
} else {
cancel_thread = std::thread(
[&context, this](int delay) {
CancelRpc(&context, delay, &callback_service_);
},
kCancelDelayUs);
}
Status s = stub_->Echo(&context, request, &response);
cancel_thread.join();
EXPECT_EQ(StatusCode::CANCELLED, s.error_code());
EXPECT_EQ(s.error_message(), "CANCELLED");
}
// Server cancels rpc after 1ms
TEST_P(ProxyEnd2endTest, ServerCancelsRpc) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
request.mutable_param()->set_server_cancel_after_us(1000);
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(StatusCode::CANCELLED, s.error_code());
EXPECT_TRUE(s.error_message().empty());
}
// Make the response larger than the flow control window.
TEST_P(ProxyEnd2endTest, HugeResponse) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("huge response");
const size_t kResponseSize = 1024 * (1024 + 10);
request.mutable_param()->set_response_message_length(kResponseSize);
ClientContext context;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::seconds(20);
context.set_deadline(deadline);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(kResponseSize, response.message().size());
EXPECT_TRUE(s.ok());
}
TEST_P(ProxyEnd2endTest, Peer) {
// Peer is not meaningful for inproc
if (GetParam().inproc()) {
return;
}
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("hello");
request.mutable_param()->set_echo_peer(true);
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
EXPECT_TRUE(CheckIsLocalhost(response.param().peer()));
EXPECT_TRUE(CheckIsLocalhost(context.peer()));
}
//////////////////////////////////////////////////////////////////////////
class SecureEnd2endTest : public End2endTest {
protected:
SecureEnd2endTest() {
GPR_ASSERT(!GetParam().use_proxy());
GPR_ASSERT(GetParam().credentials_type() != kInsecureCredentialsType);
}
};
TEST_P(SecureEnd2endTest, SimpleRpcWithHost) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
context.set_authority("foo.test.youtube.com");
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(response.has_param());
EXPECT_EQ("special", response.param().host());
EXPECT_TRUE(s.ok());
}
bool MetadataContains(
const std::multimap<grpc::string_ref, grpc::string_ref>& metadata,
const std::string& key, const std::string& value) {
int count = 0;
for (std::multimap<grpc::string_ref, grpc::string_ref>::const_iterator iter =
metadata.begin();
iter != metadata.end(); ++iter) {
if (ToString(iter->first) == key && ToString(iter->second) == value) {
count++;
}
}
return count == 1;
}
TEST_P(SecureEnd2endTest, BlockingAuthMetadataPluginAndProcessorSuccess) {
auto* processor = new TestAuthMetadataProcessor(true);
StartServer(std::shared_ptr<AuthMetadataProcessor>(processor));
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(processor->GetCompatibleClientCreds());
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
request.mutable_param()->set_expected_client_identity(
TestAuthMetadataProcessor::kGoodGuy);
request.mutable_param()->set_expected_transport_security_type(
GetParam().credentials_type());
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
// Metadata should have been consumed by the processor.
EXPECT_FALSE(MetadataContains(
context.GetServerTrailingMetadata(), GRPC_AUTHORIZATION_METADATA_KEY,
std::string("Bearer ") + TestAuthMetadataProcessor::kGoodGuy));
}
TEST_P(SecureEnd2endTest, BlockingAuthMetadataPluginAndProcessorFailure) {
auto* processor = new TestAuthMetadataProcessor(true);
StartServer(std::shared_ptr<AuthMetadataProcessor>(processor));
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(processor->GetIncompatibleClientCreds());
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAUTHENTICATED);
}
TEST_P(SecureEnd2endTest, SetPerCallCredentials) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
std::shared_ptr<CallCredentials> creds =
GoogleIAMCredentials(kFakeToken, kFakeSelector);
context.set_credentials(creds);
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORIZATION_TOKEN_METADATA_KEY,
kFakeToken));
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORITY_SELECTOR_METADATA_KEY,
kFakeSelector));
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedFakeCredsDebugString);
}
class CredentialsInterceptor : public experimental::Interceptor {
public:
explicit CredentialsInterceptor(experimental::ClientRpcInfo* info)
: info_(info) {}
void Intercept(experimental::InterceptorBatchMethods* methods) override {
if (methods->QueryInterceptionHookPoint(
experimental::InterceptionHookPoints::PRE_SEND_INITIAL_METADATA)) {
std::shared_ptr<CallCredentials> creds =
GoogleIAMCredentials(kFakeToken, kFakeSelector);
info_->client_context()->set_credentials(creds);
}
methods->Proceed();
}
private:
experimental::ClientRpcInfo* info_ = nullptr;
};
class CredentialsInterceptorFactory
: public experimental::ClientInterceptorFactoryInterface {
CredentialsInterceptor* CreateClientInterceptor(
experimental::ClientRpcInfo* info) override {
return new CredentialsInterceptor(info);
}
};
TEST_P(SecureEnd2endTest, CallCredentialsInterception) {
if (!GetParam().use_interceptors()) {
return;
}
std::vector<std::unique_ptr<experimental::ClientInterceptorFactoryInterface>>
interceptor_creators;
interceptor_creators.push_back(
std::make_unique<CredentialsInterceptorFactory>());
ResetStub(std::move(interceptor_creators));
EchoRequest request;
EchoResponse response;
ClientContext context;
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORIZATION_TOKEN_METADATA_KEY,
kFakeToken));
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORITY_SELECTOR_METADATA_KEY,
kFakeSelector));
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedFakeCredsDebugString);
}
TEST_P(SecureEnd2endTest, CallCredentialsInterceptionWithSetCredentials) {
if (!GetParam().use_interceptors()) {
return;
}
std::vector<std::unique_ptr<experimental::ClientInterceptorFactoryInterface>>
interceptor_creators;
interceptor_creators.push_back(
std::make_unique<CredentialsInterceptorFactory>());
ResetStub(std::move(interceptor_creators));
EchoRequest request;
EchoResponse response;
ClientContext context;
std::shared_ptr<CallCredentials> creds1 =
GoogleIAMCredentials(kWrongToken, kWrongSelector);
context.set_credentials(creds1);
EXPECT_EQ(context.credentials(), creds1);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedWrongCredsDebugString);
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORIZATION_TOKEN_METADATA_KEY,
kFakeToken));
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORITY_SELECTOR_METADATA_KEY,
kFakeSelector));
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedFakeCredsDebugString);
}
TEST_P(SecureEnd2endTest, OverridePerCallCredentials) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
std::shared_ptr<CallCredentials> creds1 =
GoogleIAMCredentials(kFakeToken1, kFakeSelector1);
context.set_credentials(creds1);
EXPECT_EQ(context.credentials(), creds1);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedFakeCreds1DebugString);
std::shared_ptr<CallCredentials> creds2 =
GoogleIAMCredentials(kFakeToken2, kFakeSelector2);
context.set_credentials(creds2);
EXPECT_EQ(context.credentials(), creds2);
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
Status s = stub_->Echo(&context, request, &response);
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORIZATION_TOKEN_METADATA_KEY,
kFakeToken2));
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORITY_SELECTOR_METADATA_KEY,
kFakeSelector2));
EXPECT_FALSE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORIZATION_TOKEN_METADATA_KEY,
kFakeToken1));
EXPECT_FALSE(MetadataContains(context.GetServerTrailingMetadata(),
GRPC_IAM_AUTHORITY_SELECTOR_METADATA_KEY,
kFakeSelector1));
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedFakeCreds2DebugString);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
}
TEST_P(SecureEnd2endTest, AuthMetadataPluginKeyFailure) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kBadMetadataKey,
"Does not matter, will fail the key is invalid.", false, true,
0))));
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAVAILABLE);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedAuthMetadataPluginKeyFailureCredsDebugString);
}
TEST_P(SecureEnd2endTest, AuthMetadataPluginValueFailure) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kGoodMetadataKey,
"With illegal \n value.", false, true, 0))));
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAVAILABLE);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedAuthMetadataPluginValueFailureCredsDebugString);
}
TEST_P(SecureEnd2endTest, AuthMetadataPluginWithDeadline) {
ResetStub();
EchoRequest request;
request.mutable_param()->set_skip_cancelled_check(true);
EchoResponse response;
ClientContext context;
const int delay = 100;
std::chrono::system_clock::time_point deadline =
std::chrono::system_clock::now() + std::chrono::milliseconds(delay);
context.set_deadline(deadline);
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin("meta_key", "Does not matter", true,
true, delay))));
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
if (!s.ok()) {
EXPECT_TRUE(s.error_code() == StatusCode::DEADLINE_EXCEEDED ||
s.error_code() == StatusCode::UNAVAILABLE);
}
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedAuthMetadataPluginWithDeadlineCredsDebugString);
}
TEST_P(SecureEnd2endTest, AuthMetadataPluginWithCancel) {
ResetStub();
EchoRequest request;
request.mutable_param()->set_skip_cancelled_check(true);
EchoResponse response;
ClientContext context;
const int delay = 100;
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin("meta_key", "Does not matter", true,
true, delay))));
request.set_message("Hello");
std::thread cancel_thread([&] {
gpr_sleep_until(gpr_time_add(gpr_now(GPR_CLOCK_REALTIME),
gpr_time_from_millis(delay, GPR_TIMESPAN)));
context.TryCancel();
});
Status s = stub_->Echo(&context, request, &response);
if (!s.ok()) {
EXPECT_TRUE(s.error_code() == StatusCode::CANCELLED ||
s.error_code() == StatusCode::UNAVAILABLE);
}
cancel_thread.join();
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedAuthMetadataPluginWithDeadlineCredsDebugString);
}
TEST_P(SecureEnd2endTest, NonBlockingAuthMetadataPluginFailure) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kGoodMetadataKey,
"Does not matter, will fail anyway (see 3rd param)", false, false,
0))));
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAVAILABLE);
EXPECT_EQ(s.error_message(),
std::string("Getting metadata from plugin failed with error: ") +
kTestCredsPluginErrorMsg);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedNonBlockingAuthMetadataPluginFailureCredsDebugString);
}
TEST_P(SecureEnd2endTest, NonBlockingAuthMetadataPluginAndProcessorSuccess) {
auto* processor = new TestAuthMetadataProcessor(false);
StartServer(std::shared_ptr<AuthMetadataProcessor>(processor));
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(processor->GetCompatibleClientCreds());
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
request.mutable_param()->set_expected_client_identity(
TestAuthMetadataProcessor::kGoodGuy);
request.mutable_param()->set_expected_transport_security_type(
GetParam().credentials_type());
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(request.message(), response.message());
EXPECT_TRUE(s.ok());
// Metadata should have been consumed by the processor.
EXPECT_FALSE(MetadataContains(
context.GetServerTrailingMetadata(), GRPC_AUTHORIZATION_METADATA_KEY,
std::string("Bearer ") + TestAuthMetadataProcessor::kGoodGuy));
EXPECT_EQ(
context.credentials()->DebugString(),
kExpectedNonBlockingAuthMetadataPluginAndProcessorSuccessCredsDebugString);
}
TEST_P(SecureEnd2endTest, NonBlockingAuthMetadataPluginAndProcessorFailure) {
auto* processor = new TestAuthMetadataProcessor(false);
StartServer(std::shared_ptr<AuthMetadataProcessor>(processor));
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(processor->GetIncompatibleClientCreds());
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAUTHENTICATED);
EXPECT_EQ(
context.credentials()->DebugString(),
kExpectedNonBlockingAuthMetadataPluginAndProcessorFailureCredsDebugString);
}
TEST_P(SecureEnd2endTest, BlockingAuthMetadataPluginFailure) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
context.set_credentials(grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(
TestMetadataCredentialsPlugin::kGoodMetadataKey,
"Does not matter, will fail anyway (see 3rd param)", true, false,
0))));
request.set_message("Hello");
Status s = stub_->Echo(&context, request, &response);
EXPECT_FALSE(s.ok());
EXPECT_EQ(s.error_code(), StatusCode::UNAVAILABLE);
EXPECT_EQ(s.error_message(),
std::string("Getting metadata from plugin failed with error: ") +
kTestCredsPluginErrorMsg);
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedBlockingAuthMetadataPluginFailureCredsDebugString);
}
TEST_P(SecureEnd2endTest, CompositeCallCreds) {
ResetStub();
EchoRequest request;
EchoResponse response;
ClientContext context;
const char kMetadataKey1[] = "call-creds-key1";
const char kMetadataKey2[] = "call-creds-key2";
const char kMetadataVal1[] = "call-creds-val1";
const char kMetadataVal2[] = "call-creds-val2";
context.set_credentials(grpc::CompositeCallCredentials(
grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(kMetadataKey1, kMetadataVal1,
true, true, 0))),
grpc::MetadataCredentialsFromPlugin(
std::unique_ptr<MetadataCredentialsPlugin>(
new TestMetadataCredentialsPlugin(kMetadataKey2, kMetadataVal2,
true, true, 0)))));
request.set_message("Hello");
request.mutable_param()->set_echo_metadata(true);
Status s = stub_->Echo(&context, request, &response);
EXPECT_TRUE(s.ok());
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
kMetadataKey1, kMetadataVal1));
EXPECT_TRUE(MetadataContains(context.GetServerTrailingMetadata(),
kMetadataKey2, kMetadataVal2));
EXPECT_EQ(context.credentials()->DebugString(),
kExpectedCompositeCallCredsDebugString);
}
TEST_P(SecureEnd2endTest, ClientAuthContext) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
request.mutable_param()->set_check_auth_context(
GetParam().credentials_type() == kTlsCredentialsType);
request.mutable_param()->set_expected_transport_security_type(
GetParam().credentials_type());
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
std::shared_ptr<const AuthContext> auth_ctx = context.auth_context();
std::vector<grpc::string_ref> tst =
auth_ctx->FindPropertyValues("transport_security_type");
ASSERT_EQ(1u, tst.size());
EXPECT_EQ(GetParam().credentials_type(), ToString(tst[0]));
if (GetParam().credentials_type() == kTlsCredentialsType) {
EXPECT_EQ("x509_subject_alternative_name",
auth_ctx->GetPeerIdentityPropertyName());
EXPECT_EQ(4u, auth_ctx->GetPeerIdentity().size());
EXPECT_EQ("*.test.google.fr", ToString(auth_ctx->GetPeerIdentity()[0]));
EXPECT_EQ("waterzooi.test.google.be",
ToString(auth_ctx->GetPeerIdentity()[1]));
EXPECT_EQ("*.test.youtube.com", ToString(auth_ctx->GetPeerIdentity()[2]));
EXPECT_EQ("192.168.1.3", ToString(auth_ctx->GetPeerIdentity()[3]));
}
}
class ResourceQuotaEnd2endTest : public End2endTest {
public:
ResourceQuotaEnd2endTest()
: server_resource_quota_("server_resource_quota") {}
void ConfigureServerBuilder(ServerBuilder* builder) override {
builder->SetResourceQuota(server_resource_quota_);
}
private:
ResourceQuota server_resource_quota_;
};
TEST_P(ResourceQuotaEnd2endTest, SimpleRequest) {
ResetStub();
EchoRequest request;
EchoResponse response;
request.set_message("Hello");
ClientContext context;
Status s = stub_->Echo(&context, request, &response);
EXPECT_EQ(response.message(), request.message());
EXPECT_TRUE(s.ok());
}
// TODO(vjpai): refactor arguments into a struct if it makes sense
std::vector<TestScenario> CreateTestScenarios(bool use_proxy,
bool test_insecure,
bool test_secure,
bool test_inproc,
bool test_callback_server) {
std::vector<TestScenario> scenarios;
std::vector<std::string> credentials_types;
GPR_GLOBAL_CONFIG_SET(grpc_client_channel_backup_poll_interval_ms,
kClientChannelBackupPollIntervalMs);
#if TARGET_OS_IPHONE
// Workaround Apple CFStream bug
grpc_core::SetEnv("grpc_cfstream", "0");
#endif
if (test_secure) {
credentials_types =
GetCredentialsProvider()->GetSecureCredentialsTypeList();
}
auto insec_ok = [] {
// Only allow insecure credentials type when it is registered with the
// provider. User may create providers that do not have insecure.
return GetCredentialsProvider()->GetChannelCredentials(
kInsecureCredentialsType, nullptr) != nullptr;
};
if (test_insecure && insec_ok()) {
credentials_types.push_back(kInsecureCredentialsType);
}
// Test callback with inproc or if the event-engine allows it
GPR_ASSERT(!credentials_types.empty());
for (const auto& cred : credentials_types) {
scenarios.emplace_back(false, false, false, cred, false);
scenarios.emplace_back(true, false, false, cred, false);
if (test_callback_server) {
// Note that these scenarios will be dynamically disabled if the event
// engine doesn't run in the background
scenarios.emplace_back(false, false, false, cred, true);
scenarios.emplace_back(true, false, false, cred, true);
}
if (use_proxy) {
scenarios.emplace_back(false, true, false, cred, false);
scenarios.emplace_back(true, true, false, cred, false);
}
}
if (test_inproc && insec_ok()) {
scenarios.emplace_back(false, false, true, kInsecureCredentialsType, false);
scenarios.emplace_back(true, false, true, kInsecureCredentialsType, false);
if (test_callback_server) {
scenarios.emplace_back(false, false, true, kInsecureCredentialsType,
true);
scenarios.emplace_back(true, false, true, kInsecureCredentialsType, true);
}
}
return scenarios;
}
INSTANTIATE_TEST_SUITE_P(
End2end, End2endTest,
::testing::ValuesIn(CreateTestScenarios(false, true, true, true, true)),
&TestScenario::Name);
INSTANTIATE_TEST_SUITE_P(
End2endServerTryCancel, End2endServerTryCancelTest,
::testing::ValuesIn(CreateTestScenarios(false, true, true, true, true)),
&TestScenario::Name);
INSTANTIATE_TEST_SUITE_P(
ProxyEnd2end, ProxyEnd2endTest,
::testing::ValuesIn(CreateTestScenarios(true, true, true, true, true)),
&TestScenario::Name);
INSTANTIATE_TEST_SUITE_P(
SecureEnd2end, SecureEnd2endTest,
::testing::ValuesIn(CreateTestScenarios(false, false, true, false, true)),
&TestScenario::Name);
INSTANTIATE_TEST_SUITE_P(
ResourceQuotaEnd2end, ResourceQuotaEnd2endTest,
::testing::ValuesIn(CreateTestScenarios(false, true, true, true, true)),
&TestScenario::Name);
} // namespace
} // namespace testing
} // namespace grpc
int main(int argc, char** argv) {
grpc::testing::TestEnvironment env(&argc, argv);
::testing::InitGoogleTest(&argc, argv);
int ret = RUN_ALL_TESTS();
return ret;
}