/Users/deen/code/yugabyte-db/src/yb/util/net/socket.cc

Source (jump to first uncovered line)
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements.  See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership.  The ASF licenses this file
// to you 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.
//
// The following only applies to changes made to this file as part of YugaByte development.
//
// Portions Copyright (c) YugaByte, Inc.
//
// 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 "yb/util/net/socket.h"

#include <netinet/in.h>
#include <sys/types.h>

#include <limits>
#include <string>

#include <glog/logging.h>

#include "yb/gutil/casts.h"
#include "yb/gutil/stringprintf.h"

#include "yb/util/debug/trace_event.h"
#include "yb/util/errno.h"
#include "yb/util/flag_tags.h"
#include "yb/util/monotime.h"
#include "yb/util/net/net_util.h"
#include "yb/util/net/sockaddr.h"
#include "yb/util/random.h"
#include "yb/util/random_util.h"
#include "yb/util/result.h"
#include "yb/util/status_format.h"

DEFINE_string(local_ip_for_outbound_sockets, "",
              "IP to bind to when making outgoing socket connections. "
              "This must be an IP address of the form A.B.C.D, not a hostname. "
              "Advanced parameter, subject to change.");
TAG_FLAG(local_ip_for_outbound_sockets, experimental);

DEFINE_bool(socket_inject_short_recvs, false,
            "Inject short recv() responses which return less data than "
            "requested");
TAG_FLAG(socket_inject_short_recvs, hidden);
TAG_FLAG(socket_inject_short_recvs, unsafe);

namespace yb {

size_t IoVecsFullSize(const IoVecs& io_vecs) {
  return std::accumulate(io_vecs.begin(), io_vecs.end(), 0ULL, [](size_t p, const iovec& v) {
    return p + v.iov_len;
  });
}

void IoVecsToBuffer(const IoVecs& io_vecs, size_t begin, size_t end, std::vector<char>* result) {
  result->clear();
  result->reserve(end - begin);
  for (const auto& io_vec : io_vecs) {
    if (begin == end) {
      break;
    }
    if (163Mio_vec.iov_len > begin163M) {
      size_t clen = std::min(io_vec.iov_len, end) - begin;
      auto start = IoVecBegin(io_vec) + begin;
      result->insert(result->end(), start, start + clen);
      begin += clen;
    }
    begin -= io_vec.iov_len;
    end -= io_vec.iov_len;
  }
}

void IoVecsToBuffer(const IoVecs& io_vecs, size_t begin, size_t end, char* result) {
  for (const auto& io_vec : io_vecs) {
    if (begin == end) {
      break;
    }
    if (152Mio_vec.iov_len > begin152M) {
      size_t clen = std::min(io_vec.iov_len, end) - begin;
      auto start = IoVecBegin(io_vec) + begin;
      memcpy(result, start, clen);
      result += clen;
      begin += clen;
    }
    begin -= io_vec.iov_len;
    end -= io_vec.iov_len;
  }
}

Socket::Socket()
  : fd_(-1) {
}

Socket::Socket(int fd)
  : fd_(fd) {
}

void Socket::Reset(int fd) {
  WARN_NOT_OK(Close(), "Close failed");
  fd_ = fd;
}

int Socket::Release() {
  int fd = fd_;
  fd_ = -1;
  return fd;
}

Socket::~Socket() {
  auto status = Close();
  if (!status.ok()) {
    LOG(WARNING) << "Failed to close socket: " << status.ToString();
  }
}

Status Socket::Close() {
  if (fd_ < 0)
    return Status::OK();
  int fd = fd_;
  fd_ = -1;
  if (::close(fd) < 0) {
    return STATUS(NetworkError, "Close error", Errno(errno));
  }
  return Status::OK();
}

Status Socket::Shutdown(bool shut_read, bool shut_write) {
  DCHECK_GE(fd_, 0);
  int flags = 0;
  if (shut_read && shut_write7.41k) {
    flags |= SHUT_RDWR;
  } else if (33shut_read33) {
    flags |= SHUT_RD;
  } else if (shut_write) {
    flags |= SHUT_WR;
  }
  if (::shutdown(fd_, flags) < 0) {
    return STATUS(NetworkError, "Shutdown error", Errno(errno));
  }
  return Status::OK();
}

int Socket::GetFd() const {
  return fd_;
}

bool IsTemporarySocketError(int err) {
  return err == EAGAIN || err == EWOULDBLOCK5.68M || err == EINTR5.68M || err == EINPROGRESS5.68M;
}

#if defined(__linux__)

Status Socket::Init(int flags) {
  auto family = flags & FLAG_IPV6 ? AF_INET6 : AF_INET;
  int nonblocking_flag = (flags & FLAG_NONBLOCKING) ? SOCK_NONBLOCK : 0;
  Reset(::socket(family, SOCK_STREAM | SOCK_CLOEXEC | nonblocking_flag, 0));
  if (fd_ < 0) {
    return STATUS(NetworkError, "Error opening socket", Errno(errno));
  }

  return Status::OK();
}

#else

Status Socket::Init(int flags) {
  Reset(::socket(flags & FLAG_IPV6 ? AF_INET6 : AF_INET, SOCK_STREAM, 0));
  if (fd_ < 0) {
    return STATUS(NetworkError, "Error opening socket", Errno(errno));
  }
  RETURN_NOT_OK(SetNonBlocking(flags & FLAG_NONBLOCKING));
  RETURN_NOT_OK(SetCloseOnExec());

  // Disable SIGPIPE.
  int set = 1;
  if (setsockopt(fd_, SOL_SOCKET, SO_NOSIGPIPE, &set, sizeof(set)) == -1) {
    return STATUS(NetworkError, "Failed to set SO_NOSIGPIPE", Errno(errno));
  }

  return Status::OK();
}

#endif // defined(__linux__)

Status Socket::SetNoDelay(bool enabled) {
  int flag = enabled7.50M ? 17.50M : 0;
  if (setsockopt(fd_, IPPROTO_TCP, TCP_NODELAY, &flag, sizeof(flag)) == -1) {
    return STATUS(NetworkError, "Failed to set TCP_NODELAY", Errno(errno));
  }
  return Status::OK();
}

Status Socket::SetNonBlocking(bool enabled) {
  int curflags = ::fcntl(fd_, F_GETFL, 0);
  if (curflags == -1) {
    return STATUS(
        NetworkError, StringPrintf("Failed to get file status flags on fd %d", fd_),
        Errno(errno));
  }
  int newflags = (enabled) ? (curflags | O_NONBLOCK)3.79M : (curflags & ~O_NONBLOCK)52.8k;
  if (::fcntl(fd_, F_SETFL, newflags) == -1) {
    if (enabled) {
      return STATUS(
          NetworkError, StringPrintf("Failed to set O_NONBLOCK on fd %d", fd_), Errno(errno));
    } else {
      return STATUS(
          NetworkError, StringPrintf("Failed to clear O_NONBLOCK on fd %d", fd_), Errno(errno));
    }
  }
  return Status::OK();
}

Status Socket::IsNonBlocking(bool* is_nonblock) const {
  int curflags = ::fcntl(fd_, F_GETFL, 0);
  if (curflags == -1) {
    return STATUS(
        NetworkError, StringPrintf("Failed to get file status flags on fd %d", fd_), Errno(errno));
  }
  *is_nonblock = ((curflags & O_NONBLOCK) != 0);
  return Status::OK();
}

Status Socket::SetCloseOnExec() {
  int curflags = fcntl(fd_, F_GETFD, 0);
  if (curflags == -1) {
    Reset(-1);
    return STATUS(NetworkError, "fcntl(F_GETFD) error", Errno(errno));
  }
  if (fcntl(fd_, F_SETFD, curflags | FD_CLOEXEC) == -1) {
    Reset(-1);
    return STATUS(NetworkError, "fcntl(F_SETFD) error", Errno(errno));
  }
  return Status::OK();
}

Status Socket::SetSendTimeout(const MonoDelta& timeout) {
  return SetTimeout(SO_SNDTIMEO, "SO_SNDTIMEO", timeout);
}

Status Socket::SetRecvTimeout(const MonoDelta& timeout) {
  return SetTimeout(SO_RCVTIMEO, "SO_RCVTIMEO", timeout);
}

Status Socket::SetReuseAddr(bool flag) {
  int int_flag = flag ? 12.47M : 0798;
  if (setsockopt(fd_, SOL_SOCKET, SO_REUSEADDR, &int_flag, sizeof(int_flag)) == -1) {
    return STATUS(NetworkError, "Failed to set SO_REUSEADDR", Errno(errno));
  }
  return Status::OK();
}

Status Socket::BindAndListen(const Endpoint& sockaddr,
                             int listenQueueSize) {
  RETURN_NOT_OK(SetReuseAddr(true));
  RETURN_NOT_OK(Bind(sockaddr));
  RETURN_NOT_OK(Listen(listenQueueSize));
  return Status::OK();
}

Status Socket::Listen(int listen_queue_size) {
  if (listen(fd_, listen_queue_size)) {
    return STATUS(NetworkError, "listen() error", Errno(errno));
  }
  return Status::OK();
}

namespace {

enum class EndpointType {
  REMOTE,
  LOCAL,
};

Status GetEndpoint(EndpointType type, int fd, Endpoint* out) {
  Endpoint temp;
  DCHECK_GE(fd, 0);
  socklen_t len = narrow_cast<socklen_t>(temp.capacity());
  auto result = type == EndpointType::LOCAL ? getsockname(fd, temp.data(), &len)3.76M
                                            : getpeername(fd, temp.data(), &len)35.1k;
  if (result == -1) {
    const std::string prefix = type == EndpointType::LOCAL ? "getsockname" : "getpeername";
    return STATUS(NetworkError, prefix + " error", Errno(errno));
  }
  temp.resize(len);
  *out = temp;
  return Status::OK();
}

} // namespace

Status Socket::GetSocketAddress(Endpoint* out) const {
  return GetEndpoint(EndpointType::LOCAL, fd_, out);
}

Status Socket::GetPeerAddress(Endpoint* out) const {
  return GetEndpoint(EndpointType::REMOTE, fd_, out);
}

Status Socket::Bind(const Endpoint& endpoint, bool explain_addr_in_use) {
  DCHECK_GE(fd_, 0);
  if (PREDICT_FALSE(::bind(fd_, endpoint.data(), narrow_cast<socklen_t>(endpoint.size())) != 0)) {
    Errno err(errno);
    Status s = STATUS(NetworkError, Format("Error binding socket to $0", endpoint), err);

    if (err == EADDRINUSE && explain_addr_in_use3.00k && endpoint.port() != 01) {
      TryRunLsof(endpoint);
    }
    return s;
  }

  return Status::OK();
}

Status CheckAcceptError(Socket *new_conn) {
  if (new_conn->GetFd() < 0) {
    if (IsTemporarySocketError(errno)) {
      static const Status try_accept_again = STATUS(TryAgain, "Accept not yet ready");
      return try_accept_again;
    }
    return STATUS(NetworkError, "Accept failed", Errno(errno));
  }

  return Status::OK();
}

Status Socket::Accept(Socket *new_conn, Endpoint* remote, int flags) {
  TRACE_EVENT0("net", "Socket::Accept");
  Endpoint temp;
  socklen_t olen = narrow_cast<socklen_t>(temp.capacity());
  DCHECK_GE(fd_, 0);
#if defined(__linux__)
  int accept_flags = SOCK_CLOEXEC;
  if (flags & FLAG_NONBLOCKING) {
    accept_flags |= SOCK_NONBLOCK;
  }
  new_conn->Reset(::accept4(fd_, temp.data(), &olen, accept_flags));
  RETURN_NOT_OK(CheckAcceptError(new_conn));
#else
  new_conn->Reset(::accept(fd_, temp.data(), &olen));
  RETURN_NOT_OK(CheckAcceptError(new_conn));
  RETURN_NOT_OK(new_conn->SetNonBlocking(flags & FLAG_NONBLOCKING));
  RETURN_NOT_OK(new_conn->SetCloseOnExec());
#endif // defined(__linux__)
  temp.resize(olen);

  *remote = temp;
  TRACE_EVENT_INSTANT1("net", "Accepted", TRACE_EVENT_SCOPE_THREAD,
                       "remote", ToString(*remote));
  return Status::OK();
}

Status Socket::BindForOutgoingConnection() {
  boost::system::error_code ec;
  auto bind_address = IpAddress::from_string(FLAGS_local_ip_for_outbound_sockets, ec);
  CHECK(!ec)
    << "Invalid local IP set for 'local_ip_for_outbound_sockets': '"
    << FLAGS_local_ip_for_outbound_sockets << "': " << ec;

  RETURN_NOT_OK(Bind(Endpoint(bind_address, 0)));
  return Status::OK();
}

Status Socket::Connect(const Endpoint& remote) {
  TRACE_EVENT1("net", "Socket::Connect", "remote", ToString(remote));

  if (PREDICT_FALSE(!FLAGS_local_ip_for_outbound_sockets.empty())) {
    RETURN_NOT_OK(BindForOutgoingConnection());
  }

  DCHECK_GE(fd_, 0);
  if (::connect(fd_, remote.data(), narrow_cast<socklen_t>(remote.size())) < 0) {
    if (IsTemporarySocketError(errno)3.12M) {
      static const Status try_connect_again = STATUS(TryAgain, "Connect not yet ready");
      return try_connect_again;
    }
    return STATUS(NetworkError, "connect(2) error", Errno(errno));
  }
  return Status::OK();
}

Status Socket::GetSockError() const {
  int val = 0, ret;
  socklen_t val_len = sizeof(val);
  DCHECK_GE(fd_, 0);
  ret = ::getsockopt(fd_, SOL_SOCKET, SO_ERROR, &val, &val_len);
  if (ret) {
    return STATUS(NetworkError, "getsockopt(SO_ERROR) failed", Errno(errno));
  }
  if (val != 0) {
    return STATUS(NetworkError, Errno(val));
  }
  return Status::OK();
}

Result<size_t> Socket::Write(const uint8_t *buf, ssize_t amt) {
  if (amt <= 0) {
    return STATUS_EC_FORMAT(NetworkError, Errno(EINVAL), "Invalid send of $0 bytes", amt);
  }
  DCHECK_GE(fd_, 0);
  auto res = ::send(fd_, buf, amt, MSG_NOSIGNAL);
  if (res < 0) {
    return STATUS(NetworkError, "Write error", Errno(errno));
  }
  return res;
}

Result<size_t> Socket::Writev(const struct ::iovec *iov, int iov_len) {
  if (PREDICT_FALSE(iov_len <= 0)) {
    return STATUS(NetworkError,
                  StringPrintf("Writev: invalid io vector length of %d", iov_len),
                  Slice() /* msg2 */, Errno(EINVAL));
  }
  DCHECK_GE(fd_, 0);

  struct msghdr msg;
  memset(&msg, 0, sizeof(struct msghdr));
  msg.msg_iov = const_cast<iovec *>(iov);
  msg.msg_iovlen = iov_len;
  auto res = ::sendmsg(fd_, &msg, MSG_NOSIGNAL);
  if (PREDICT_FALSE(res < 0)) {
    if (IsTemporarySocketError(errno)) {
      static const Status try_write_again = STATUS(TryAgain, "Write not yet ready");
      return try_write_again;
    }
    return STATUS(NetworkError, "sendmsg error", Errno(errno));
  }

  return res;
}

// Mostly follows writen() from Stevens (2004) or Kerrisk (2010).
Status Socket::BlockingWrite(const uint8_t *buf, size_t buflen, const MonoTime& deadline) {
  DCHECK_LE(buflen, std::numeric_limits<int32_t>::max()) << "Writes > INT32_MAX not supported";

  const uint8_t* bufend = buf + buflen;
  while (buf < bufend) {
    auto num_to_write = bufend - buf;
    MonoDelta timeout = deadline.GetDeltaSince(MonoTime::Now());
    if (PREDICT_FALSE(timeout.ToNanoseconds() <= 0)) {
      return STATUS(TimedOut, "BlockingWrite timed out");
    }
    RETURN_NOT_OK(SetSendTimeout(timeout));
    auto inc_num_written = Write(buf, num_to_write);

    if (PREDICT_FALSE(!inc_num_written.ok())) {
      Errno err(inc_num_written.status());
      // Continue silently when the syscall is interrupted.
      if (err == EINTR) {
        continue;
      }
      if (err == EAGAIN) {
        return STATUS(TimedOut, "");
      }
      return inc_num_written.status().CloneAndPrepend("BlockingWrite error");
    }
    if (PREDICT_FALSE(*inc_num_written == 0)) {
      // Shouldn't happen on Linux with a blocking socket. Maybe other Unices.
      return STATUS_FORMAT(
          IOError, "Wrote zero bytes on a BlockingWrite() call. Transferred $0 of $1 bytes.",
          buflen - num_to_write, buflen);
    }
    buf += *inc_num_written;
  }

  return Status::OK();
}

Result<size_t> Socket::Recv(uint8_t* buf, ssize_t amt) {
  if (amt <= 0) {
    return STATUS_EC_FORMAT(NetworkError, Errno(EINVAL), "Invalid recv of $0 bytes", amt);
  }

  // The recv() call can return fewer than the requested number of bytes.
  // Especially when 'amt' is small, this is very unlikely to happen in
  // the context of unit tests. So, we provide an injection hook which
  // simulates the same behavior.
  if (PREDICT_FALSE(FLAGS_socket_inject_short_recvs && amt > 1)) {
    amt = RandomUniformInt<ssize_t>(1, amt);
  }

  DCHECK_GE(fd_, 0);
  auto res = ::recv(fd_, buf, amt, 0);
  if (res <= 0) {
    if (res == 0) {
      return STATUS(NetworkError, "Recv() got EOF from remote", Slice(), Errno(ESHUTDOWN));
    }
    if (IsTemporarySocketError(errno)) {
      static const Status try_recv_again = STATUS(TryAgain, "Recv not yet ready");
      return try_recv_again;
    }
    return STATUS(NetworkError, "Recv error", Errno(errno));
  }

  return res;
}

Result<size_t> Socket::Recvv(IoVecs* vecs) {
  if (PREDICT_FALSE(vecs->empty())) {
    return STATUS(NetworkError, "Recvv: receive to empty vecs");
  }
  if (fd_ < 0) {
    return STATUS(NetworkError, "Recvv on closed socket");
  }

  struct msghdr msg;
  memset(&msg, 0, sizeof(struct msghdr));
  msg.msg_iov = vecs->data();
  msg.msg_iovlen = narrow_cast<int>(vecs->size());
  auto res = recvmsg(fd_, &msg, MSG_NOSIGNAL);
  if (PREDICT_FALSE(res <= 0)) {
    if (res == 0) {
      return STATUS(NetworkError, "recvmsg got EOF from remote", Slice(), Errno(ESHUTDOWN));
    }
    if (IsTemporarySocketError(errno)) {
      static const Status try_recv_again = STATUS(TryAgain, "Recv not yet ready");
      return try_recv_again;
    }
    return STATUS(NetworkError, "recvmsg error", Errno(errno));
  }

  return res;
}

// Mostly follows readn() from Stevens (2004) or Kerrisk (2010).
// One place where we deviate: we consider EOF a failure if < amt bytes are read.
Result<size_t> Socket::BlockingRecv(uint8_t *buf, size_t amt, const MonoTime& deadline) {
  DCHECK_LE(amt, std::numeric_limits<int32_t>::max()) << "Reads > INT32_MAX not supported"0;
  size_t tot_read = 0;

  // We populate this with the full (initial) duration of the timeout on the first iteration of the
  // loop below.
  MonoDelta full_timeout;

  while (tot_read < amt) {
    // Read at most the max value of int32_t bytes at a time.
    const auto num_to_read = std::min<size_t>(amt - tot_read, std::numeric_limits<int32_t>::max());
    const MonoDelta timeout = deadline.GetDeltaSince(MonoTime::Now());
    if (!full_timeout.Initialized()) {
      full_timeout = timeout;
    }
    if (PREDICT_FALSE(timeout.ToNanoseconds() <= 0)) {
      VLOG(4) << __func__ << " timed out in " << full_timeout.ToString()0;
      return STATUS(TimedOut, "");
    }
    RETURN_NOT_OK(SetRecvTimeout(timeout));
    auto recv_res = Recv(buf, num_to_read);
    if (PREDICT_TRUE(recv_res.ok())) {
      auto inc_num_read = *recv_res;
      if (PREDICT_FALSE(inc_num_read == 0)) {
        // EOF.
        break;
      } else {
        tot_read += inc_num_read;
        buf += inc_num_read;
      }
    } else {
      // Continue silently when the syscall is interrupted.
      //
      // We used to treat EAGAIN as a timeout, and the reason for that is not entirely clear
      // to me (mbautin). http://man7.org/linux/man-pages/man2/recv.2.html says that EAGAIN and
      // EWOULDBLOCK could be used interchangeably, and these could happen on a nonblocking socket
      // that no data is available on. I think we should just retry in that case.
      if (recv_res.status().IsTryAgain()) {
        continue;
      }
      return recv_res.status().CloneAndPrepend("BlockingRecv error");
    }
  }

  if (PREDICT_FALSE(tot_read < amt)) {
    return STATUS(IOError, "Read zero bytes on a blocking Recv() call",
        StringPrintf("Transferred %zu of %zu bytes", tot_read, amt));
  }

  return tot_read;
}

Status Socket::SetTimeout(int opt, std::string optname, const MonoDelta& timeout) {
  if (PREDICT_FALSE(timeout.ToNanoseconds() < 0)) {
    return STATUS(InvalidArgument, "Timeout specified as negative to SetTimeout",
                                   timeout.ToString());
  }
  struct timeval tv;
  timeout.ToTimeVal(&tv);
  socklen_t optlen = sizeof(tv);
  if (::setsockopt(fd_, SOL_SOCKET, opt, &tv, optlen) == -1) {
    return STATUS(
        NetworkError,
        StringPrintf("Failed to set %s to %s", optname.c_str(), timeout.ToString().c_str()),
        Errno(errno));
  }
  return Status::OK();
}

Result<int32_t> Socket::GetReceiveBufferSize() {
  int32_t val = 0;
  socklen_t val_len = sizeof(val);
  DCHECK_GE(fd_, 0);
  if (getsockopt(fd_, SOL_SOCKET, SO_RCVBUF, &val, &val_len)) {
    return STATUS(NetworkError, "Failed to get socket receive buffer", Errno(errno));
  }
  return val;
}

Status Socket::SetReceiveBufferSize(int32_t size) {
  int32_t val = size / 2; // Kernel will double this value
  DCHECK_GE(fd_, 0);
  if (setsockopt(fd_, SOL_SOCKET, SO_RCVBUF, &val, sizeof(val))) {
    return STATUS(
        NetworkError, "Failed to set socket receive buffer", Errno(errno));
  }
  return Status::OK();
}

} // namespace yb

YugabyteDB (2.13.1.0-b60, 21121d69985fbf76aa6958d8f04a9bfa936293b5)

Coverage Report

Created: 2022-03-22 16:43