| /* |
| * Copyright (C) 2022 The Android Open Source Project |
| * |
| * Licensed under the Apache License, Version 2.0 (the "License"); |
| * you may not use this file except in compliance with the License. |
| * You may obtain a copy of the License at |
| * |
| * http://www.apache.org/licenses/LICENSE-2.0 |
| * |
| * Unless required by applicable law or agreed to in writing, software |
| * distributed under the License is distributed on an "AS IS" BASIS, |
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| * See the License for the specific language governing permissions and |
| * limitations under the License. |
| */ |
| |
| #define LOG_TAG "RpcTransportTipcAndroid" |
| |
| #include <binder/RpcSession.h> |
| #include <binder/RpcTransportTipcAndroid.h> |
| #include <log/log.h> |
| #include <poll.h> |
| #include <trusty/tipc.h> |
| |
| #include "FdTrigger.h" |
| #include "RpcState.h" |
| #include "RpcTransportUtils.h" |
| |
| using namespace android::binder::impl; |
| using android::binder::borrowed_fd; |
| using android::binder::unique_fd; |
| |
| namespace android { |
| |
| // RpcTransport for writing Trusty IPC clients in Android. |
| class RpcTransportTipcAndroid : public RpcTransport { |
| public: |
| explicit RpcTransportTipcAndroid(android::RpcTransportFd socket) : mSocket(std::move(socket)) {} |
| |
| status_t pollRead() override { |
| if (mReadBufferPos < mReadBufferSize) { |
| // We have more data in the read buffer |
| return OK; |
| } |
| |
| // Trusty IPC device is not a socket, so MSG_PEEK is not available |
| pollfd pfd{.fd = mSocket.fd.get(), .events = static_cast<int16_t>(POLLIN), .revents = 0}; |
| ssize_t ret = TEMP_FAILURE_RETRY(::poll(&pfd, 1, 0)); |
| if (ret < 0) { |
| int savedErrno = errno; |
| if (savedErrno == EAGAIN || savedErrno == EWOULDBLOCK) { |
| return WOULD_BLOCK; |
| } |
| |
| LOG_RPC_DETAIL("RpcTransport poll(): %s", strerror(savedErrno)); |
| return adjustStatus(-savedErrno); |
| } |
| |
| if (pfd.revents & POLLNVAL) { |
| return BAD_VALUE; |
| } |
| if (pfd.revents & POLLERR) { |
| return DEAD_OBJECT; |
| } |
| if (pfd.revents & POLLIN) { |
| // Copied from FdTrigger.cpp: Even though POLLHUP may also be set, |
| // treat it as a success condition to ensure data is drained. |
| return OK; |
| } |
| if (pfd.revents & POLLHUP) { |
| return DEAD_OBJECT; |
| } |
| |
| return WOULD_BLOCK; |
| } |
| |
| status_t interruptableWriteFully( |
| FdTrigger* fdTrigger, iovec* iovs, int niovs, |
| const std::optional<SmallFunction<status_t()>>& altPoll, |
| const std::vector<std::variant<unique_fd, borrowed_fd>>* ancillaryFds) override { |
| auto writeFn = [&](iovec* iovs, size_t niovs) -> ssize_t { |
| // TODO: send ancillaryFds. For now, we just abort if anyone tries |
| // to send any. |
| LOG_ALWAYS_FATAL_IF(ancillaryFds != nullptr && !ancillaryFds->empty(), |
| "File descriptors are not supported on Trusty yet"); |
| return TEMP_FAILURE_RETRY(tipc_send(mSocket.fd.get(), iovs, niovs, nullptr, 0)); |
| }; |
| |
| status_t status = interruptableReadOrWrite(mSocket, fdTrigger, iovs, niovs, writeFn, |
| "tipc_send", POLLOUT, altPoll); |
| return adjustStatus(status); |
| } |
| |
| status_t interruptableReadFully( |
| FdTrigger* fdTrigger, iovec* iovs, int niovs, |
| const std::optional<SmallFunction<status_t()>>& altPoll, |
| std::vector<std::variant<unique_fd, borrowed_fd>>* /*ancillaryFds*/) override { |
| auto readFn = [&](iovec* iovs, size_t niovs) -> ssize_t { |
| // Fill the read buffer at most once per readFn call, then try to |
| // return as much of it as possible. If the input iovecs are spread |
| // across multiple messages that require multiple fillReadBuffer |
| // calls, we expect the caller to advance the iovecs past the first |
| // read and call readFn as many times as needed to get all the data |
| status_t ret = fillReadBuffer(); |
| if (ret != OK) { |
| // We need to emulate a Linux read call, which sets errno on |
| // error and returns -1 |
| errno = -ret; |
| return -1; |
| } |
| |
| ssize_t processSize = 0; |
| for (size_t i = 0; i < niovs && mReadBufferPos < mReadBufferSize; i++) { |
| auto& iov = iovs[i]; |
| size_t numBytes = std::min(iov.iov_len, mReadBufferSize - mReadBufferPos); |
| memcpy(iov.iov_base, mReadBuffer.get() + mReadBufferPos, numBytes); |
| mReadBufferPos += numBytes; |
| processSize += numBytes; |
| } |
| |
| return processSize; |
| }; |
| |
| status_t status = interruptableReadOrWrite(mSocket, fdTrigger, iovs, niovs, readFn, "read", |
| POLLIN, altPoll); |
| return adjustStatus(status); |
| } |
| |
| bool isWaiting() override { return mSocket.isInPollingState(); } |
| |
| private: |
| status_t adjustStatus(status_t status) { |
| if (status == -ENOTCONN) { |
| // TIPC returns ENOTCONN on disconnect, but that's basically |
| // the same as DEAD_OBJECT and the latter is the common libbinder |
| // error code for dead connections |
| return DEAD_OBJECT; |
| } |
| |
| return status; |
| } |
| |
| status_t fillReadBuffer() { |
| if (mReadBufferPos < mReadBufferSize) { |
| return OK; |
| } |
| |
| if (!mReadBuffer) { |
| // Guarantee at least kDefaultBufferSize bytes |
| mReadBufferCapacity = std::max(mReadBufferCapacity, kDefaultBufferSize); |
| mReadBuffer.reset(new (std::nothrow) uint8_t[mReadBufferCapacity]); |
| if (!mReadBuffer) { |
| return NO_MEMORY; |
| } |
| } |
| |
| // Reset the size and position in case we have to exit with an error. |
| // After we read a message into the buffer, we update the size |
| // with the actual value. |
| mReadBufferPos = 0; |
| mReadBufferSize = 0; |
| |
| while (true) { |
| ssize_t processSize = TEMP_FAILURE_RETRY( |
| read(mSocket.fd.get(), mReadBuffer.get(), mReadBufferCapacity)); |
| if (processSize == 0) { |
| return DEAD_OBJECT; |
| } else if (processSize < 0) { |
| int savedErrno = errno; |
| if (savedErrno == EMSGSIZE) { |
| // Buffer was too small, double it and retry |
| if (__builtin_mul_overflow(mReadBufferCapacity, 2, &mReadBufferCapacity)) { |
| return NO_MEMORY; |
| } |
| mReadBuffer.reset(new (std::nothrow) uint8_t[mReadBufferCapacity]); |
| if (!mReadBuffer) { |
| return NO_MEMORY; |
| } |
| continue; |
| } else { |
| LOG_RPC_DETAIL("RpcTransport fillBuffer(): %s", strerror(savedErrno)); |
| return adjustStatus(-savedErrno); |
| } |
| } else { |
| mReadBufferSize = static_cast<size_t>(processSize); |
| return OK; |
| } |
| } |
| } |
| |
| RpcTransportFd mSocket; |
| |
| // For now, we copy all the input data into a temporary buffer because |
| // we might get multiple interruptableReadFully calls per message, but |
| // the tipc device only allows one read call. We read every message into |
| // this temporary buffer, then return pieces of it from our method. |
| // |
| // The special transaction GET_MAX_THREADS takes 40 bytes, so the default |
| // size should start pretty high. |
| static constexpr size_t kDefaultBufferSize = 64; |
| std::unique_ptr<uint8_t[]> mReadBuffer; |
| size_t mReadBufferPos = 0; |
| size_t mReadBufferSize = 0; |
| size_t mReadBufferCapacity = 0; |
| }; |
| |
| // RpcTransportCtx for Trusty. |
| class RpcTransportCtxTipcAndroid : public RpcTransportCtx { |
| public: |
| std::unique_ptr<RpcTransport> newTransport(android::RpcTransportFd fd, |
| FdTrigger*) const override { |
| return std::make_unique<RpcTransportTipcAndroid>(std::move(fd)); |
| } |
| std::vector<uint8_t> getCertificate(RpcCertificateFormat) const override { return {}; } |
| }; |
| |
| std::unique_ptr<RpcTransportCtx> RpcTransportCtxFactoryTipcAndroid::newServerCtx() const { |
| return std::make_unique<RpcTransportCtxTipcAndroid>(); |
| } |
| |
| std::unique_ptr<RpcTransportCtx> RpcTransportCtxFactoryTipcAndroid::newClientCtx() const { |
| return std::make_unique<RpcTransportCtxTipcAndroid>(); |
| } |
| |
| const char* RpcTransportCtxFactoryTipcAndroid::toCString() const { |
| return "trusty"; |
| } |
| |
| std::unique_ptr<RpcTransportCtxFactory> RpcTransportCtxFactoryTipcAndroid::make() { |
| return std::unique_ptr<RpcTransportCtxFactoryTipcAndroid>( |
| new RpcTransportCtxFactoryTipcAndroid()); |
| } |
| |
| } // namespace android |