FairMQ/fairmq/sdk/Topology.cxx

/********************************************************************************
 *    Copyright (C) 2019 GSI Helmholtzzentrum fuer Schwerionenforschung GmbH    *
 *                                                                              *
 *              This software is distributed under the terms of the             *
 *              GNU Lesser General Public Licence (LGPL) version 3,             *
 *                  copied verbatim in the file "LICENSE"                       *
 ********************************************************************************/

#include "Topology.h"

#include <fairlogger/Logger.h>

#include <boost/algorithm/string/split.hpp>
#include <boost/algorithm/string/trim.hpp>

#include <utility>
#include <thread>
#include <future>
#include <mutex>
#include <condition_variable>

namespace fair {
namespace mq {

auto operator<<(std::ostream& os, AsyncOpResult v) -> std::ostream&
{
    switch (v) {
        case AsyncOpResult::Aborted:
            return os << "Aborted";
        case AsyncOpResult::Timeout:
            return os << "Timeout";
        case AsyncOpResult::Error:
            return os << "Error";
        case AsyncOpResult::Ok:
        default:
            return os << "Ok";
    }
}

namespace sdk {

const std::unordered_map<DeviceTransition, DeviceState, tools::HashEnum<DeviceTransition>> Topology::fkExpectedState =
{
    { Transition::InitDevice,   DeviceState::InitializingDevice },
    { Transition::CompleteInit, DeviceState::Initialized },
    { Transition::Bind,         DeviceState::Bound },
    { Transition::Connect,      DeviceState::DeviceReady },
    { Transition::InitTask,     DeviceState::InitializingTask },
    { Transition::Run,          DeviceState::Running },
    { Transition::Stop,         DeviceState::Ready },
    { Transition::ResetTask,    DeviceState::DeviceReady },
    { Transition::ResetDevice,  DeviceState::Idle },
    { Transition::End,          DeviceState::Exiting }
};

Topology::Topology(DDSTopology topo, DDSSession session)
    : fDDSSession(std::move(session))
    , fDDSTopo(std::move(topo))
    , fTopologyState()
    , fStateChangeOngoing(false)
    , fExecutionThread()
    , fShutdown(false)
{
    fDDSTopo.CreateTopology(fDDSTopo.GetTopoFile());

    std::vector<uint64_t> deviceList = fDDSTopo.GetDeviceList();
    for (const auto& d : deviceList) {
        fTopologyState.emplace(d, DeviceStatus{ false, DeviceState::Ok });
    }
    fDDSSession.SubscribeToCommands([this](const std::string& msg, const std::string& condition, uint64_t senderId) {
        LOG(info) << "Received from " << senderId << ": " << msg;
        std::vector<std::string> parts;
        boost::algorithm::split(parts, msg, boost::algorithm::is_any_of(":,"));
        if (parts[0] == "state-change") {
            boost::trim(parts[2]);
            AddNewStateEntry(senderId, parts[2]);
        } else if (parts[0] == "state-changes-subscription") {
            if (parts[2] != "OK") {
                LOG(error) << "state-changes-subscription failed with return code: " << parts[2];
            }
        } else if (parts[0] == "state-changes-unsubscription") {
            if (parts[2] != "OK") {
                LOG(error) << "state-changes-unsubscription failed with return code: " << parts[2];
            }
        }
    });
    fDDSSession.StartDDSService();
    fDDSSession.SendCommand("subscribe-to-state-changes");

    fExecutionThread = std::thread(&Topology::WaitForState, this);
}

auto Topology::ChangeState(fair::mq::Transition transition, ChangeStateCallback cb, const std::chrono::milliseconds& timeout) -> void
{
    {
        std::lock_guard<std::mutex> guard(fMtx);
        if (fStateChangeOngoing) {
            LOG(error) << "State change already in progress, concurrent requested not yet supported";
            return;
        }
        fStateChangeOngoing = true;
        fChangeStateCallback = cb;
        fStateChangeTimeout = timeout;

        fDDSSession.SendCommand(GetTransitionName(transition));

        fTargetState = fkExpectedState.at(transition);
    }
    fExecutionCV.notify_one();
}

void Topology::WaitForState()
{
    while (!fShutdown) {
        if (fStateChangeOngoing) {
            auto condition = [&] { return fShutdown || std::all_of(fTopologyState.cbegin(),
                                                                   fTopologyState.cend(),
                                                                   [&](TopologyState::value_type i) {
                                                                       return i.second.state == fTargetState;
                                                                   });
            };

            std::unique_lock<std::mutex> lock(fMtx);

            if (fStateChangeTimeout > std::chrono::milliseconds(0)) {
                if (!fCV.wait_for(lock, fStateChangeTimeout, condition)) {
                    LOG(debug) << "timeout";
                    // TODO: catch this from another thread...
                    throw std::runtime_error("timeout");
                }
            } else {
                fCV.wait(lock, condition);
            }

            if (fShutdown) {
                break;
            }

            fStateChangeOngoing = false;
            fChangeStateCallback(ChangeStateResult{AsyncOpResult::Ok, fTopologyState});
        } else {
            std::unique_lock<std::mutex> lock(fExecutionMtx);
            fExecutionCV.wait(lock);
        }
    }
};

void Topology::AddNewStateEntry(uint64_t senderId, const std::string& state)
{
    {
        std::unique_lock<std::mutex> lock(fMtx);
        fTopologyState[senderId] = DeviceStatus{ true, fair::mq::GetState(state) };
    }
    fCV.notify_one();
}

Topology::~Topology()
{
    {
        std::lock_guard<std::mutex> guard(fMtx);
        fShutdown = true;
    }
    fExecutionCV.notify_one();
    fExecutionThread.join();
}

auto operator<<(std::ostream& os, Topology::ChangeStateResult v) -> std::ostream&
{
    return os << v.rc;
}

}   // namespace sdk
}   // namespace mq
}   // namespace fair