X-Git-Url: http://xvm.mit.edu/gitweb/invirt/third/libt4.git/blobdiff_plain/5fd8cc8409d0efadc07dfe8d6774ad9ff477663d..f0dcb6b97d6d40f67698d1f71ac26970f1776f82:/rpc/rpc.cc diff --git a/rpc/rpc.cc b/rpc/rpc.cc index e18506a..7937785 100644 --- a/rpc/rpc.cc +++ b/rpc/rpc.cc @@ -1,646 +1,447 @@ -/* - The rpcc class handles client-side RPC. Each rpcc is bound to a - single RPC server. The jobs of rpcc include maintaining a connection to - server, sending RPC requests and waiting for responses, retransmissions, - at-most-once delivery etc. - - The rpcs class handles the server side of RPC. Each rpcs handles multiple - connections from different rpcc objects. The jobs of rpcs include accepting - connections, dispatching requests to registered RPC handlers, at-most-once - delivery etc. - - Both rpcc and rpcs use the connection class as an abstraction for the - underlying communication channel. To send an RPC request/reply, one calls - connection::send() which blocks until data is sent or the connection has failed - (thus the caller can free the buffer when send() returns). When a - request/reply is received, connection makes a callback into the corresponding - rpcc or rpcs (see rpcc::got_pdu() and rpcs::got_pdu()). - - Thread organization: - rpcc uses application threads to send RPC requests and blocks to receive the - reply or error. All connections use a single PollMgr object to perform async - socket IO. PollMgr creates a single thread to examine the readiness of socket - file descriptors and informs the corresponding connection whenever a socket is - ready to be read or written. (We use asynchronous socket IO to reduce the - number of threads needed to manage these connections; without async IO, at - least one thread is needed per connection to read data without blocking other - activities.) Each rpcs object creates one thread for listening on the server - port and a pool of threads for executing RPC requests. The - thread pool allows us to control the number of threads spawned at the server - (spawning one thread per request will hurt when the server faces thousands of - requests). - - In order to delete a connection object, we must maintain a reference count. - For rpcc, - multiple client threads might be invoking the rpcc::call() functions and thus - holding multiple references to the underlying connection object. For rpcs, - multiple dispatch threads might be holding references to the same connection - object. A connection object is deleted only when the underlying connection is - dead and the reference count reaches zero. - - The previous version of the RPC library uses pthread_cancel* routines - to implement the deletion of rpcc and rpcs objects. The idea is to cancel - all active threads that might be holding a reference to an object before - deleting that object. However, pthread_cancel is not robust and there are - always bugs where outstanding references to deleted objects persist. - This version of the RPC library does not do pthread_cancel, but explicitly - joins exited threads to make sure no outstanding references exist before - deleting objects. - - To delete a rpcc object safely, the users of the library must ensure that - there are no outstanding calls on the rpcc object. - - To delete a rpcs object safely, we do the following in sequence: 1. stop - accepting new incoming connections. 2. close existing active connections. - 3. delete the dispatch thread pool which involves waiting for current active - RPC handlers to finish. It is interesting how a thread pool can be deleted - without using thread cancellation. The trick is to inject x "poison pills" for - a thread pool of x threads. Upon getting a poison pill instead of a normal - task, a worker thread will exit (and thread pool destructor waits to join all - x exited worker threads). - */ +// +// The rpcc class handles client-side RPC. Each rpcc is bound to a single RPC +// server. The jobs of rpcc include maintaining a connection to server, sending +// RPC requests and waiting for responses, retransmissions, at-most-once delivery +// etc. +// +// The rpcs class handles the server side of RPC. Each rpcs handles multiple +// connections from different rpcc objects. The jobs of rpcs include accepting +// connections, dispatching requests to registered RPC handlers, at-most-once +// delivery etc. +// +// Both rpcc and rpcs use the connection class as an abstraction for the +// underlying communication channel. To send an RPC request/reply, one calls +// connection::send() which blocks until data is sent or the connection has +// failed (thus the caller can free the buffer when send() returns). When a +// request/reply is received, connection makes a callback into the corresponding +// rpcc or rpcs (see rpcc::got_pdu() and rpcs::got_pdu()). +// +// Thread organization: +// rpcc uses application threads to send RPC requests and blocks to receive the +// reply or error. All connections use a single PollMgr object to perform async +// socket IO. PollMgr creates a single thread to examine the readiness of socket +// file descriptors and informs the corresponding connection whenever a socket is +// ready to be read or written. (We use asynchronous socket IO to reduce the +// number of threads needed to manage these connections; without async IO, at +// least one thread is needed per connection to read data without blocking other +// activities.) Each rpcs object creates one thread for listening on the server +// port and a pool of threads for executing RPC requests. The thread pool allows +// us to control the number of threads spawned at the server (spawning one thread +// per request will hurt when the server faces thousands of requests). +// +// In order to delete a connection object, we must maintain a reference count. +// For rpcc, multiple client threads might be invoking the rpcc::call() functions +// and thus holding multiple references to the underlying connection object. For +// rpcs, multiple dispatch threads might be holding references to the same +// connection object. A connection object is deleted only when the underlying +// connection is dead and the reference count reaches zero. +// +// This version of the RPC library explicitly joins exited threads to make sure +// no outstanding references exist before deleting objects. +// +// To delete a rpcc object safely, the users of the library must ensure that +// there are no outstanding calls on the rpcc object. +// +// To delete a rpcs object safely, we do the following in sequence: 1. stop +// accepting new incoming connections. 2. close existing active connections. 3. +// delete the dispatch thread pool which involves waiting for current active RPC +// handlers to finish. It is interesting how a thread pool can be deleted +// without using thread cancellation. The trick is to inject x "poison pills" for +// a thread pool of x threads. Upon getting a poison pill instead of a normal +// task, a worker thread will exit (and thread pool destructor waits to join all +// x exited worker threads). +// #include "rpc.h" -#include "method_thread.h" -#include "slock.h" -#include #include #include -#include #include +#include +#include -#include "jsl_log.h" -#include "gettime.h" -#include "lang/verify.h" - -const rpcc::TO rpcc::to_max = { 120000 }; -const rpcc::TO rpcc::to_min = { 1000 }; - -rpcc::caller::caller(unsigned int xxid, unmarshall *xun) -: xid(xxid), un(xun), done(false) -{ - VERIFY(pthread_mutex_init(&m,0) == 0); - VERIFY(pthread_cond_init(&c, 0) == 0); +inline void set_rand_seed() { + auto now = time_point_cast(steady_clock::now()); + srandom((uint32_t)now.time_since_epoch().count()^(uint32_t)getpid()); } -rpcc::caller::~caller() -{ - VERIFY(pthread_mutex_destroy(&m) == 0); - VERIFY(pthread_cond_destroy(&c) == 0); -} +static sockaddr_in make_sockaddr(const string & hostandport); -inline -void set_rand_seed() +rpcc::rpcc(const string & d) : dst_(make_sockaddr(d)) { - struct timespec ts; - clock_gettime(CLOCK_REALTIME, &ts); - srandom((int)ts.tv_nsec^((int)getpid())); -} + set_rand_seed(); + clt_nonce_ = (nonce_t)random(); -rpcc::rpcc(sockaddr_in d, bool retrans) : - dst_(d), srv_nonce_(0), bind_done_(false), xid_(1), lossytest_(0), - retrans_(retrans), reachable_(true), chan_(NULL), destroy_wait_ (false), xid_rep_done_(-1) -{ - VERIFY(pthread_mutex_init(&m_, 0) == 0); - VERIFY(pthread_mutex_init(&chan_m_, 0) == 0); - VERIFY(pthread_cond_init(&destroy_wait_c_, 0) == 0); - - if(retrans){ - set_rand_seed(); - clt_nonce_ = random(); - } else { - // special client nonce 0 means this client does not - // require at-most-once logic from the server - // because it uses tcp and never retries a failed connection - clt_nonce_ = 0; - } - - char *loss_env = getenv("RPC_LOSSY"); - if(loss_env != NULL){ - lossytest_ = atoi(loss_env); - } - - // xid starts with 1 and latest received reply starts with 0 - xid_rep_window_.push_back(0); - - jsl_log(JSL_DBG_2, "rpcc::rpcc cltn_nonce is %d lossy %d\n", - clt_nonce_, lossytest_); + char *loss_env = getenv("RPC_LOSSY"); + if (loss_env) + lossytest_ = atoi(loss_env); + + IF_LEVEL(2) LOG("cltn_nonce is " << clt_nonce_ << " lossy " << lossytest_); } // IMPORTANT: destruction should happen only when no external threads // are blocked inside rpcc or will use rpcc in the future -rpcc::~rpcc() -{ - jsl_log(JSL_DBG_2, "rpcc::~rpcc delete nonce %d channo=%d\n", - clt_nonce_, chan_?chan_->channo():-1); - if(chan_){ - chan_->closeconn(); - chan_->decref(); - } - VERIFY(calls_.size() == 0); - VERIFY(pthread_mutex_destroy(&m_) == 0); - VERIFY(pthread_mutex_destroy(&chan_m_) == 0); -} - -int -rpcc::bind(TO to) -{ - int r; - int ret = call(rpc_const::bind, 0, r, to); - if(ret == 0){ - ScopedLock ml(&m_); - bind_done_ = true; - srv_nonce_ = r; - } else { - jsl_log(JSL_DBG_2, "rpcc::bind %s failed %d\n", - inet_ntoa(dst_.sin_addr), ret); - } - return ret; +rpcc::~rpcc() { + cancel(); + IF_LEVEL(2) LOG("delete nonce " << clt_nonce_ << " chan " << (chan_?(int)chan_->fd:-1)); + chan_.reset(); + VERIFY(calls_.size() == 0); +} + +int rpcc::bind(milliseconds to) { + nonce_t r; + rpc_protocol::status ret = call_timeout(rpc_protocol::bind, to, r); + if (ret == 0) { + lock ml(m_); + bind_done_ = true; + srv_nonce_ = r; + } else { + IF_LEVEL(2) LOG("bind " << inet_ntoa(dst_.sin_addr) << " failed " << ret); + } + return ret; }; // Cancel all outstanding calls -void -rpcc::cancel(void) -{ - ScopedLock ml(&m_); - printf("rpcc::cancel: force callers to fail\n"); - std::map::iterator iter; - for(iter = calls_.begin(); iter != calls_.end(); iter++){ - caller *ca = iter->second; +void rpcc::cancel(void) { + lock ml(m_); + if (calls_.size()) { + LOG("force callers to fail"); + for (auto & p : calls_) { + caller *ca = p.second; + + IF_LEVEL(2) LOG("force caller to fail"); + + lock cl(ca->m); + ca->done = true; + ca->intret = rpc_protocol::cancel_failure; + ca->c.notify_one(); + } - jsl_log(JSL_DBG_2, "rpcc::cancel: force caller to fail\n"); + destroy_wait_ = true; + while (calls_.size () > 0) + destroy_wait_c_.wait(ml); + + LOG("done"); + } +} + +int rpcc::call1(proc_id_t proc, milliseconds to, string & rep, marshall & req) { + + caller ca(0, &rep); + xid_t xid_rep; { - ScopedLock cl(&ca->m); - ca->done = true; - ca->intret = rpc_const::cancel_failure; - VERIFY(pthread_cond_signal(&ca->c) == 0); + lock ml(m_); + + if ((proc != rpc_protocol::bind.id && !bind_done_) || (proc == rpc_protocol::bind.id && bind_done_)) { + IF_LEVEL(1) LOG("rpcc has not been bound to dst or binding twice"); + return rpc_protocol::bind_failure; + } + + if (destroy_wait_) + return rpc_protocol::cancel_failure; + + ca.xid = xid_++; + calls_[ca.xid] = &ca; + + req.pack_header(rpc_protocol::request_header{ + ca.xid, proc, clt_nonce_, srv_nonce_, xid_rep_window_.front() + }); + xid_rep = xid_rep_window_.front(); } - } - while (calls_.size () > 0){ - destroy_wait_ = true; - VERIFY(pthread_cond_wait(&destroy_wait_c_,&m_) == 0); - } - printf("rpcc::cancel: done\n"); -} + milliseconds curr_to = rpc::to_min; + auto finaldeadline = steady_clock::now() + to; + + bool transmit = true; + shared_ptr ch; + + while (1) { + if (transmit) { + get_latest_connection(ch); + if (ch) { + if (reachable_) { + request forgot; + { + lock ml(m_); + if (dup_req_.isvalid() && xid_rep_done_ > dup_req_.xid) { + forgot = dup_req_; + dup_req_.clear(); + } + } + if (forgot.isvalid()) + ch->send(forgot.buf); + ch->send(req); + } + else IF_LEVEL(1) LOG("not reachable"); + IF_LEVEL(2) LOG(clt_nonce_ << " just sent req proc " << hex << proc << + " xid " << dec << ca.xid << " clt_nonce " << clt_nonce_); + } + transmit = false; // only send once on a given channel + } -int -rpcc::call1(unsigned int proc, marshall &req, unmarshall &rep, - TO to) -{ + auto nextdeadline = min(steady_clock::now() + curr_to, finaldeadline); + curr_to *= 2; - caller ca(0, &rep); - int xid_rep; - { - ScopedLock ml(&m_); - - if((proc != rpc_const::bind && !bind_done_) || - (proc == rpc_const::bind && bind_done_)){ - jsl_log(JSL_DBG_1, "rpcc::call1 rpcc has not been bound to dst or binding twice\n"); - return rpc_const::bind_failure; - } - - if(destroy_wait_){ - return rpc_const::cancel_failure; - } - - ca.xid = xid_++; - calls_[ca.xid] = &ca; - - req_header h(ca.xid, proc, clt_nonce_, srv_nonce_, - xid_rep_window_.front()); - req.pack_req_header(h); - xid_rep = xid_rep_window_.front(); - } - - TO curr_to; - struct timespec now, nextdeadline, finaldeadline; - - clock_gettime(CLOCK_REALTIME, &now); - add_timespec(now, to.to, &finaldeadline); - curr_to.to = to_min.to; - - bool transmit = true; - connection *ch = NULL; - - while (1){ - if(transmit){ - get_refconn(&ch); - if(ch){ - if(reachable_) { - request forgot; - { - ScopedLock ml(&m_); - if (dup_req_.isvalid() && xid_rep_done_ > dup_req_.xid) { - forgot = dup_req_; - dup_req_.clear(); - } - } - if (forgot.isvalid()) - ch->send((char *)forgot.buf.c_str(), forgot.buf.size()); - ch->send(req.cstr(), req.size()); - } - else jsl_log(JSL_DBG_1, "not reachable\n"); - jsl_log(JSL_DBG_2, - "rpcc::call1 %u just sent req proc %x xid %u clt_nonce %d\n", - clt_nonce_, proc, ca.xid, clt_nonce_); - } - transmit = false; // only send once on a given channel - } - - if(!finaldeadline.tv_sec) - break; - - clock_gettime(CLOCK_REALTIME, &now); - add_timespec(now, curr_to.to, &nextdeadline); - if(cmp_timespec(nextdeadline,finaldeadline) > 0){ - nextdeadline = finaldeadline; - finaldeadline.tv_sec = 0; - } - - { - ScopedLock cal(&ca.m); - while (!ca.done){ - jsl_log(JSL_DBG_2, "rpcc:call1: wait\n"); - if(pthread_cond_timedwait(&ca.c, &ca.m, - &nextdeadline) == ETIMEDOUT){ - jsl_log(JSL_DBG_2, "rpcc::call1: timeout\n"); - break; - } - } - if(ca.done){ - jsl_log(JSL_DBG_2, "rpcc::call1: reply received\n"); - break; - } - } - - if(retrans_ && (!ch || ch->isdead())){ - // since connection is dead, retransmit - // on the new connection - transmit = true; - } - curr_to.to <<= 1; - } - - { - // no locking of ca.m since only this thread changes ca.xid - ScopedLock ml(&m_); - calls_.erase(ca.xid); - // may need to update the xid again here, in case the - // packet times out before it's even sent by the channel. - // I don't think there's any harm in maybe doing it twice - update_xid_rep(ca.xid); - - if(destroy_wait_){ - VERIFY(pthread_cond_signal(&destroy_wait_c_) == 0); - } - } - - if (ca.done && lossytest_) { - ScopedLock ml(&m_); - if (!dup_req_.isvalid()) { - dup_req_.buf.assign(req.cstr(), req.size()); - dup_req_.xid = ca.xid; + lock cal(ca.m); + while (!ca.done) { + IF_LEVEL(2) LOG("wait"); + if (ca.c.wait_until(cal, nextdeadline) == cv_status::timeout) { + IF_LEVEL(2) LOG("timeout"); + break; } - if (xid_rep > xid_rep_done_) - xid_rep_done_ = xid_rep; + } + if (ca.done) { + IF_LEVEL(2) LOG("reply received"); + break; + } } - ScopedLock cal(&ca.m); + if (nextdeadline >= finaldeadline) + break; + + // retransmit on new connection if connection is dead + if (!ch || ch->isdead()) + transmit = true; + } + + { + // no locking of ca.m since only this thread changes ca.xid + lock ml(m_); + calls_.erase(ca.xid); + // may need to update the xid again here, in case the + // packet times out before it's even sent by the channel. + // I don't think there's any harm in maybe doing it twice + update_xid_rep(ca.xid, ml); + + if (destroy_wait_) + destroy_wait_c_.notify_one(); + } + + if (ca.done && lossytest_) + { + lock ml(m_); + if (!dup_req_.isvalid()) { + dup_req_.buf = req; + dup_req_.xid = ca.xid; + } + if (xid_rep > xid_rep_done_) + xid_rep_done_ = xid_rep; + } - jsl_log(JSL_DBG_2, - "rpcc::call1 %u call done for req proc %x xid %u %s:%d done? %d ret %d \n", - clt_nonce_, proc, ca.xid, inet_ntoa(dst_.sin_addr), - ntohs(dst_.sin_port), ca.done, ca.intret); + lock cal(ca.m); - if(ch) - ch->decref(); + IF_LEVEL(2) LOG(clt_nonce_ << " call done for req proc " << hex << proc << + " xid " << dec << ca.xid << " " << inet_ntoa(dst_.sin_addr) << ":" << + ntoh(dst_.sin_port) << " done? " << ca.done << " ret " << ca.intret); - // destruction of req automatically frees its buffer - return (ca.done? ca.intret : rpc_const::timeout_failure); + // destruction of req automatically frees its buffer + return (ca.done? ca.intret : rpc_protocol::timeout_failure); } -void -rpcc::get_refconn(connection **ch) -{ - ScopedLock ml(&chan_m_); - if(!chan_ || chan_->isdead()){ - if(chan_) - chan_->decref(); - chan_ = connect_to_dst(dst_, this, lossytest_); - } - if(ch && chan_){ - if(*ch){ - (*ch)->decref(); - } - *ch = chan_; - (*ch)->incref(); - } +void rpcc::get_latest_connection(shared_ptr & ch) { + lock ml(chan_m_); + if (!chan_ || chan_->isdead()) + chan_ = connection::to_dst(dst_, this, lossytest_); + + if (chan_) + ch = chan_; } -// PollMgr's thread is being used to -// make this upcall from connection object to rpcc. +// PollMgr's thread is being used to +// make this upcall from connection object to rpcc. // this funtion must not block. // -// this function keeps no reference for connection *c +// this function keeps no reference for connection *c bool -rpcc::got_pdu(connection *c, char *b, int sz) +rpcc::got_pdu(const shared_ptr &, const string & b) { - unmarshall rep(b, sz); - reply_header h; - rep.unpack_reply_header(&h); - - if(!rep.ok()){ - jsl_log(JSL_DBG_1, "rpcc:got_pdu unmarshall header failed!!!\n"); - return true; - } - - ScopedLock ml(&m_); - - update_xid_rep(h.xid); - - if(calls_.find(h.xid) == calls_.end()){ - jsl_log(JSL_DBG_2, "rpcc::got_pdu xid %d no pending request\n", h.xid); - return true; - } - caller *ca = calls_[h.xid]; - - ScopedLock cl(&ca->m); - if(!ca->done){ - ca->un->take_in(rep); - ca->intret = h.ret; - if(ca->intret < 0){ - jsl_log(JSL_DBG_2, "rpcc::got_pdu: RPC reply error for xid %d intret %d\n", - h.xid, ca->intret); - } - ca->done = 1; - } - VERIFY(pthread_cond_broadcast(&ca->c) == 0); - return true; -} + unmarshall rep(b, true); + rpc_protocol::reply_header h; + rep.unpack_header(h); -// assumes thread holds mutex m -void -rpcc::update_xid_rep(unsigned int xid) -{ - std::list::iterator it; + if (!rep.ok()) { + IF_LEVEL(1) LOG("unmarshall header failed!!!"); + return true; + } - if(xid <= xid_rep_window_.front()){ - return; - } + lock ml(m_); - for (it = xid_rep_window_.begin(); it != xid_rep_window_.end(); it++){ - if(*it > xid){ - xid_rep_window_.insert(it, xid); - goto compress; - } - } - xid_rep_window_.push_back(xid); + update_xid_rep(h.xid, ml); -compress: - it = xid_rep_window_.begin(); - for (it++; it != xid_rep_window_.end(); it++){ - while (xid_rep_window_.front() + 1 == *it) - xid_rep_window_.pop_front(); - } + if (calls_.find(h.xid) == calls_.end()) { + IF_LEVEL(2) LOG("xid " << h.xid << " no pending request"); + return true; + } + caller *ca = calls_[h.xid]; + + lock cl(ca->m); + if (!ca->done) { + *ca->rep = b; + ca->intret = h.ret; + if (ca->intret < 0) { + IF_LEVEL(2) LOG("RPC reply error for xid " << h.xid << " intret " << ca->intret); + } + ca->done = 1; + } + ca->c.notify_all(); + return true; } +void rpcc::update_xid_rep(xid_t xid, lock & m_lock) { + VERIFY(m_lock); + if (xid <= xid_rep_window_.front()) + return; -rpcs::rpcs(unsigned int p1, int count) - : port_(p1), counting_(count), curr_counts_(count), lossytest_(0), reachable_ (true) -{ - VERIFY(pthread_mutex_init(&procs_m_, 0) == 0); - VERIFY(pthread_mutex_init(&count_m_, 0) == 0); - VERIFY(pthread_mutex_init(&reply_window_m_, 0) == 0); - VERIFY(pthread_mutex_init(&conss_m_, 0) == 0); + for (auto it = xid_rep_window_.begin(); it != xid_rep_window_.end(); it++) { + if (*it > xid) { + xid_rep_window_.insert(it, xid); + goto compress; + } + } + xid_rep_window_.push_back(xid); - set_rand_seed(); - nonce_ = random(); - jsl_log(JSL_DBG_2, "rpcs::rpcs created with nonce %d\n", nonce_); +compress: + auto it = xid_rep_window_.begin(); + for (it++; it != xid_rep_window_.end(); it++) { + while (xid_rep_window_.front() + 1 == *it) + xid_rep_window_.pop_front(); + } +} - char *loss_env = getenv("RPC_LOSSY"); - if(loss_env != NULL){ - lossytest_ = atoi(loss_env); - } +rpcs::rpcs(in_port_t p1) : port_(p1) +{ + set_rand_seed(); + nonce_ = (nonce_t)random(); + IF_LEVEL(2) LOG("created with nonce " << nonce_); - reg(rpc_const::bind, this, &rpcs::rpcbind); - dispatchpool_ = new ThrPool(6,false); + reg(rpc_protocol::bind, &rpcs::rpcbind, this); +} - listener_ = new tcpsconn(this, port_, lossytest_); +void rpcs::start() { + char *loss_env = getenv("RPC_LOSSY"); + listener_.reset(new connection_listener(this, port_, loss_env ? atoi(loss_env) : 0)); } -rpcs::~rpcs() -{ - // must delete listener before dispatchpool - delete listener_; - delete dispatchpool_; - free_reply_window(); +rpcs::~rpcs() { + // must delete listener before dispatchpool + listener_ = nullptr; + dispatchpool_ = nullptr; } -bool -rpcs::got_pdu(connection *c, char *b, int sz) -{ - if(!reachable_){ - jsl_log(JSL_DBG_1, "rpcss::got_pdu: not reachable\n"); - return true; - } +bool rpcs::got_pdu(const shared_ptr & c, const string & b) { + if (!reachable_) { + IF_LEVEL(1) LOG("not reachable"); + return true; + } - djob_t *j = new djob_t(c, b, sz); - c->incref(); - bool succ = dispatchpool_->addObjJob(this, &rpcs::dispatch, j); - if(!succ || !reachable_){ - c->decref(); - delete j; - } - return succ; + return dispatchpool_->addJob(std::bind(&rpcs::dispatch, this, c, b)); } -void -rpcs::reg1(unsigned int proc, handler *h) -{ - ScopedLock pl(&procs_m_); - VERIFY(procs_.count(proc) == 0); - procs_[proc] = h; - VERIFY(procs_.count(proc) >= 1); -} +void rpcs::dispatch(shared_ptr c, const string & buf) { + unmarshall req(buf, true); -void -rpcs::updatestat(unsigned int proc) -{ - ScopedLock cl(&count_m_); - counts_[proc]++; - curr_counts_--; - if(curr_counts_ == 0){ - std::map::iterator i; - printf("RPC STATS: "); - for (i = counts_.begin(); i != counts_.end(); i++){ - printf("%x:%d ", i->first, i->second); - } - printf("\n"); - - ScopedLock rwl(&reply_window_m_); - std::map >::iterator clt; - - unsigned int totalrep = 0, maxrep = 0; - for (clt = reply_window_.begin(); clt != reply_window_.end(); clt++){ - totalrep += clt->second.size(); - if(clt->second.size() > maxrep) - maxrep = clt->second.size(); - } - jsl_log(JSL_DBG_1, "REPLY WINDOW: clients %d total reply %d max per client %d\n", - (int) reply_window_.size()-1, totalrep, maxrep); - curr_counts_ = counting_; - } -} + rpc_protocol::request_header h; + req.unpack_header(h); + proc_id_t proc = h.proc; -void -rpcs::dispatch(djob_t *j) -{ - connection *c = j->conn; - unmarshall req(j->buf, j->sz); - delete j; - - req_header h; - req.unpack_req_header(&h); - int proc = h.proc; - - if(!req.ok()){ - jsl_log(JSL_DBG_1, "rpcs:dispatch unmarshall header failed!!!\n"); - c->decref(); - return; - } - - jsl_log(JSL_DBG_2, - "rpcs::dispatch: rpc %u (proc %x, last_rep %u) from clt %u for srv instance %u \n", - h.xid, proc, h.xid_rep, h.clt_nonce, h.srv_nonce); - - marshall rep; - reply_header rh(h.xid,0); - - // is client sending to an old instance of server? - if(h.srv_nonce != 0 && h.srv_nonce != nonce_){ - jsl_log(JSL_DBG_2, - "rpcs::dispatch: rpc for an old server instance %u (current %u) proc %x\n", - h.srv_nonce, nonce_, h.proc); - rh.ret = rpc_const::oldsrv_failure; - rep.pack_reply_header(rh); - c->send(rep.cstr(),rep.size()); - return; - } - - handler *f; - // is RPC proc a registered procedure? - { - ScopedLock pl(&procs_m_); - if(procs_.count(proc) < 1){ - fprintf(stderr, "rpcs::dispatch: unknown proc %x.\n", - proc); - c->decref(); - VERIFY(0); - return; - } - - f = procs_[proc]; - } - - rpcs::rpcstate_t stat; - char *b1; - int sz1; - - if(h.clt_nonce){ - // have i seen this client before? - { - ScopedLock rwl(&reply_window_m_); - // if we don't know about this clt_nonce, create a cleanup object - if(reply_window_.find(h.clt_nonce) == reply_window_.end()){ - VERIFY (reply_window_[h.clt_nonce].size() == 0); // create - reply_window_[h.clt_nonce].push_back(reply_t(-1)); // store starting reply xid - jsl_log(JSL_DBG_2, - "rpcs::dispatch: new client %u xid %d chan %d, total clients %d\n", - h.clt_nonce, h.xid, c->channo(), (int)reply_window_.size()-1); - } - } - - // save the latest good connection to the client - { - ScopedLock rwl(&conss_m_); - if(conns_.find(h.clt_nonce) == conns_.end()){ - c->incref(); - conns_[h.clt_nonce] = c; - } else if(conns_[h.clt_nonce]->compare(c) < 0){ - conns_[h.clt_nonce]->decref(); - c->incref(); - conns_[h.clt_nonce] = c; - } - } - - stat = checkduplicate_and_update(h.clt_nonce, h.xid, - h.xid_rep, &b1, &sz1); - } else { - // this client does not require at most once logic - stat = NEW; - } - - switch (stat){ - case NEW: // new request - if(counting_){ - updatestat(proc); - } - - rh.ret = f->fn(req, rep); - if (rh.ret == rpc_const::unmarshal_args_failure) { - fprintf(stderr, "rpcs::dispatch: failed to" - " unmarshall the arguments. You are" - " probably calling RPC 0x%x with wrong" - " types of arguments.\n", proc); - VERIFY(0); - } - VERIFY(rh.ret >= 0); - - rep.pack_reply_header(rh); - rep.take_buf(&b1,&sz1); - - jsl_log(JSL_DBG_2, - "rpcs::dispatch: sending and saving reply of size %d for rpc %u, proc %x ret %d, clt %u\n", - sz1, h.xid, proc, rh.ret, h.clt_nonce); - - if(h.clt_nonce > 0){ - // only record replies for clients that require at-most-once logic - add_reply(h.clt_nonce, h.xid, b1, sz1); - } - - // get the latest connection to the client - { - ScopedLock rwl(&conss_m_); - if(c->isdead() && c != conns_[h.clt_nonce]){ - c->decref(); - c = conns_[h.clt_nonce]; - c->incref(); - } - } - - c->send(b1, sz1); - if(h.clt_nonce == 0){ - // reply is not added to at-most-once window, free it - free(b1); - } - break; - case INPROGRESS: // server is working on this request - break; - case DONE: // duplicate and we still have the response - c->send(b1, sz1); - break; - case FORGOTTEN: // very old request and we don't have the response anymore - jsl_log(JSL_DBG_2, "rpcs::dispatch: very old request %u from %u\n", - h.xid, h.clt_nonce); - rh.ret = rpc_const::atmostonce_failure; - rep.pack_reply_header(rh); - c->send(rep.cstr(),rep.size()); - break; - } - c->decref(); + if (!req.ok()) { + IF_LEVEL(1) LOG("unmarshall header failed"); + return; + } + + IF_LEVEL(2) LOG("rpc " << h.xid << " (proc " << hex << proc << ", last_rep " << + dec << h.xid_rep << ") from clt " << h.clt_nonce << " for srv instance " << h.srv_nonce); + + marshall rep; + rpc_protocol::reply_header rh{h.xid,0}; + + // is client sending to an old instance of server? + if (h.srv_nonce != 0 && h.srv_nonce != nonce_) { + IF_LEVEL(2) LOG("rpc for an old server instance " << h.srv_nonce << + " (current " << nonce_ << ") proc " << hex << h.proc); + rh.ret = rpc_protocol::oldsrv_failure; + rep.pack_header(rh); + c->send(rep); + return; + } + + handler *f; + // is RPC proc a registered procedure? + { + lock pl(procs_m_); + if (procs_.count(proc) < 1) { + LOG("unknown proc 0x" << hex << proc << " with h.srv_nonce=" << h.srv_nonce << ", my srv_nonce=" << nonce_); + VERIFY(0); + return; + } + + f = procs_[proc]; + } + + // have i seen this client before? + { + lock rwl(reply_window_m_); + // if we don't know about this clt_nonce, create a cleanup object + if (reply_window_.find(h.clt_nonce) == reply_window_.end()) { + VERIFY (reply_window_[h.clt_nonce].size() == 0); // create + reply_window_[h.clt_nonce].push_back(reply_t(-1)); // store starting reply xid + IF_LEVEL(2) LOG("new client " << h.clt_nonce << " xid " << h.xid << + " chan " << c->fd << ", total clients " << (reply_window_.size()-1)); + } + } + + // save the latest good connection to the client + { + lock rwl(conns_m_); + if (conns_.find(h.clt_nonce) == conns_.end()) + conns_[h.clt_nonce] = c; + else if (conns_[h.clt_nonce]->create_time < c->create_time) + conns_[h.clt_nonce] = c; + } + + string b1; + + switch (check_duplicate_and_update(h.clt_nonce, h.xid, h.xid_rep, b1)) { + case NEW: // new request + rh.ret = (*f)(forward(req), rep); + if (rh.ret == rpc_protocol::unmarshall_args_failure) { + LOG("failed to unmarshall the arguments. You are " << + "probably calling RPC 0x" << hex << proc << " with the wrong " << + "types of arguments."); + VERIFY(0); + } + VERIFY(rh.ret >= 0); + + rep.pack_header(rh); + b1 = rep; + + IF_LEVEL(2) LOG("sending and saving reply of size " << b1.size() << " for rpc " << + h.xid << ", proc " << hex << proc << " ret " << dec << rh.ret << ", clt " << h.clt_nonce); + + add_reply(h.clt_nonce, h.xid, b1); + + // get the latest connection to the client + { + lock rwl(conns_m_); + if (c->isdead()) + c = conns_[h.clt_nonce]; + } + + c->send(rep); + break; + case INPROGRESS: // server is working on this request + break; + case DONE: // duplicate and we still have the response + c->send(b1); + break; + case FORGOTTEN: // very old request and we don't have the response anymore + IF_LEVEL(2) LOG("very old request " << h.xid << " from " << h.clt_nonce); + rh.ret = rpc_protocol::atmostonce_failure; + rep.pack_header(rh); + c->send(rep); + break; + } } // rpcs::dispatch calls this when an RPC request arrives. @@ -655,35 +456,32 @@ rpcs::dispatch(djob_t *j) // returns one of: // NEW: never seen this xid before. // INPROGRESS: seen this xid, and still processing it. -// DONE: seen this xid, previous reply returned in *b and *sz. +// DONE: seen this xid, previous reply returned in b. // FORGOTTEN: might have seen this xid, but deleted previous reply. -rpcs::rpcstate_t -rpcs::checkduplicate_and_update(unsigned int clt_nonce, unsigned int xid, - unsigned int xid_rep, char **b, int *sz) +rpcs::rpcstate_t +rpcs::check_duplicate_and_update(nonce_t clt_nonce, xid_t xid, + xid_t xid_rep, string & b) { - ScopedLock rwl(&reply_window_m_); + lock rwl(reply_window_m_); - std::list &l = reply_window_[clt_nonce]; + list & l = reply_window_[clt_nonce]; VERIFY(l.size() > 0); VERIFY(xid >= xid_rep); - unsigned int past_xid_rep = l.begin()->xid; + xid_t past_xid_rep = l.begin()->xid; - std::list::iterator start = l.begin(), it; - it = ++start; + list::iterator start = l.begin(), it = ++start; - if (past_xid_rep < xid_rep || past_xid_rep == (unsigned int)-1) { + if (past_xid_rep < xid_rep || past_xid_rep == -1) { // scan for deletion candidates - for (; it != l.end() && it->xid < xid_rep; it++) { - if (it->cb_present) - free(it->buf); - } + while (it != l.end() && it->xid < xid_rep) + it++; l.erase(start, it); l.begin()->xid = xid_rep; } - if (xid < past_xid_rep && past_xid_rep != (unsigned int)-1) + if (xid < past_xid_rep && past_xid_rep != -1) return FORGOTTEN; // skip non-deletion candidates @@ -694,12 +492,10 @@ rpcs::checkduplicate_and_update(unsigned int clt_nonce, unsigned int xid, if (it != l.end() && it->xid == xid) { if (it->cb_present) { // return information about the remembered reply - *b = it->buf; - *sz = it->sz; + b = it->buf; return DONE; - } else { - return INPROGRESS; } + return INPROGRESS; } else { // remember that a new request has arrived l.insert(it, reply_t(xid)); @@ -708,379 +504,56 @@ rpcs::checkduplicate_and_update(unsigned int clt_nonce, unsigned int xid, } // rpcs::dispatch calls add_reply when it is sending a reply to an RPC, -// and passes the return value in b and sz. -// add_reply() should remember b and sz. -// free_reply_window() and checkduplicate_and_update is responsible for -// calling free(b). -void -rpcs::add_reply(unsigned int clt_nonce, unsigned int xid, - char *b, int sz) -{ - ScopedLock rwl(&reply_window_m_); +// and passes the return value in b. +// add_reply() should remember b. +void rpcs::add_reply(nonce_t clt_nonce, xid_t xid, const string & b) { + lock rwl(reply_window_m_); // remember the RPC reply value - std::list &l = reply_window_[clt_nonce]; - std::list::iterator it = l.begin(); + list & l = reply_window_[clt_nonce]; + list::iterator it = l.begin(); // skip to our place in the list for (it++; it != l.end() && it->xid < xid; it++); // there should already be an entry, so whine if there isn't if (it == l.end() || it->xid != xid) { - fprintf(stderr, "Could not find reply struct in add_reply"); - l.insert(it, reply_t(xid, b, sz)); + LOG("Could not find reply struct in add_reply"); + l.insert(it, reply_t(xid, b)); } else { - *it = reply_t(xid, b, sz); + *it = reply_t(xid, b); } } -void -rpcs::free_reply_window(void) -{ - std::map >::iterator clt; - std::list::iterator it; - - ScopedLock rwl(&reply_window_m_); - for (clt = reply_window_.begin(); clt != reply_window_.end(); clt++){ - for (it = clt->second.begin(); it != clt->second.end(); it++){ - if (it->cb_present) - free(it->buf); - } - clt->second.clear(); - } - reply_window_.clear(); -} - -// rpc handler -int -rpcs::rpcbind(int a, int &r) -{ - jsl_log(JSL_DBG_2, "rpcs::rpcbind called return nonce %u\n", nonce_); - r = nonce_; - return 0; -} - -void -marshall::rawbyte(unsigned char x) -{ - if(_ind >= _capa){ - _capa *= 2; - VERIFY (_buf != NULL); - _buf = (char *)realloc(_buf, _capa); - VERIFY(_buf); - } - _buf[_ind++] = x; -} - -void -marshall::rawbytes(const char *p, int n) -{ - if((_ind+n) > _capa){ - _capa = _capa > n? 2*_capa:(_capa+n); - VERIFY (_buf != NULL); - _buf = (char *)realloc(_buf, _capa); - VERIFY(_buf); - } - memcpy(_buf+_ind, p, n); - _ind += n; -} - -marshall & -operator<<(marshall &m, bool x) -{ - m.rawbyte(x); - return m; -} - -marshall & -operator<<(marshall &m, unsigned char x) -{ - m.rawbyte(x); - return m; -} - -marshall & -operator<<(marshall &m, char x) -{ - m << (unsigned char) x; - return m; -} - - -marshall & -operator<<(marshall &m, unsigned short x) -{ - m.rawbyte((x >> 8) & 0xff); - m.rawbyte(x & 0xff); - return m; -} - -marshall & -operator<<(marshall &m, short x) -{ - m << (unsigned short) x; - return m; -} - -marshall & -operator<<(marshall &m, unsigned int x) -{ - // network order is big-endian - m.rawbyte((x >> 24) & 0xff); - m.rawbyte((x >> 16) & 0xff); - m.rawbyte((x >> 8) & 0xff); - m.rawbyte(x & 0xff); - return m; -} - -marshall & -operator<<(marshall &m, int x) -{ - m << (unsigned int) x; - return m; -} - -marshall & -operator<<(marshall &m, const std::string &s) -{ - m << (unsigned int) s.size(); - m.rawbytes(s.data(), s.size()); - return m; -} - -marshall & -operator<<(marshall &m, unsigned long long x) -{ - m << (unsigned int) (x >> 32); - m << (unsigned int) x; - return m; -} - -void -marshall::pack(int x) -{ - rawbyte((x >> 24) & 0xff); - rawbyte((x >> 16) & 0xff); - rawbyte((x >> 8) & 0xff); - rawbyte(x & 0xff); -} - -void -unmarshall::unpack(int *x) -{ - (*x) = (rawbyte() & 0xff) << 24; - (*x) |= (rawbyte() & 0xff) << 16; - (*x) |= (rawbyte() & 0xff) << 8; - (*x) |= rawbyte() & 0xff; -} - -// take the contents from another unmarshall object -void -unmarshall::take_in(unmarshall &another) -{ - if(_buf) - free(_buf); - another.take_buf(&_buf, &_sz); - _ind = RPC_HEADER_SZ; - _ok = _sz >= RPC_HEADER_SZ?true:false; -} - -bool -unmarshall::okdone() -{ - if(ok() && _ind == _sz){ - return true; - } else { - return false; - } +rpc_protocol::status rpcs::rpcbind(nonce_t & r) { + IF_LEVEL(2) LOG("called return nonce " << nonce_); + r = nonce_; + return 0; } -unsigned int -unmarshall::rawbyte() -{ - char c = 0; - if(_ind >= _sz) - _ok = false; - else - c = _buf[_ind++]; - return c; -} - -unmarshall & -operator>>(unmarshall &u, bool &x) -{ - x = (bool) u.rawbyte() ; - return u; -} - -unmarshall & -operator>>(unmarshall &u, unsigned char &x) -{ - x = (unsigned char) u.rawbyte() ; - return u; -} - -unmarshall & -operator>>(unmarshall &u, char &x) -{ - x = (char) u.rawbyte(); - return u; -} - - -unmarshall & -operator>>(unmarshall &u, unsigned short &x) -{ - x = (u.rawbyte() & 0xff) << 8; - x |= u.rawbyte() & 0xff; - return u; -} - -unmarshall & -operator>>(unmarshall &u, short &x) -{ - x = (u.rawbyte() & 0xff) << 8; - x |= u.rawbyte() & 0xff; - return u; -} - -unmarshall & -operator>>(unmarshall &u, unsigned int &x) -{ - x = (u.rawbyte() & 0xff) << 24; - x |= (u.rawbyte() & 0xff) << 16; - x |= (u.rawbyte() & 0xff) << 8; - x |= u.rawbyte() & 0xff; - return u; -} - -unmarshall & -operator>>(unmarshall &u, int &x) -{ - x = (u.rawbyte() & 0xff) << 24; - x |= (u.rawbyte() & 0xff) << 16; - x |= (u.rawbyte() & 0xff) << 8; - x |= u.rawbyte() & 0xff; - return u; -} - -unmarshall & -operator>>(unmarshall &u, unsigned long long &x) -{ - unsigned int h, l; - u >> h; - u >> l; - x = l | ((unsigned long long) h << 32); - return u; -} - -unmarshall & -operator>>(unmarshall &u, std::string &s) -{ - unsigned sz; - u >> sz; - if(u.ok()) - u.rawbytes(s, sz); - return u; -} - -void -unmarshall::rawbytes(std::string &ss, unsigned int n) -{ - if((_ind+n) > (unsigned)_sz){ - _ok = false; - } else { - std::string tmps = std::string(_buf+_ind, n); - swap(ss, tmps); - VERIFY(ss.size() == n); - _ind += n; - } -} - -bool operator<(const sockaddr_in &a, const sockaddr_in &b){ - return ((a.sin_addr.s_addr < b.sin_addr.s_addr) || - ((a.sin_addr.s_addr == b.sin_addr.s_addr) && - ((a.sin_port < b.sin_port)))); -} - -/*---------------auxilary function--------------*/ -void -make_sockaddr(const char *hostandport, struct sockaddr_in *dst){ - - char host[200]; - const char *localhost = "127.0.0.1"; - const char *port = index(hostandport, ':'); - if(port == NULL){ - memcpy(host, localhost, strlen(localhost)+1); - port = hostandport; - } else { - memcpy(host, hostandport, port-hostandport); - host[port-hostandport] = '\0'; - port++; - } - - make_sockaddr(host, port, dst); - -} +static sockaddr_in make_sockaddr(const string & hostandport) { + string host = "127.0.0.1"; + string port = hostandport; + auto colon = hostandport.find(':'); + if (colon != string::npos) { + host = hostandport.substr(0, colon); + port = hostandport.substr(colon+1); + } -void -make_sockaddr(const char *host, const char *port, struct sockaddr_in *dst){ - - in_addr_t a; - - bzero(dst, sizeof(*dst)); - dst->sin_family = AF_INET; - - a = inet_addr(host); - if(a != INADDR_NONE){ - dst->sin_addr.s_addr = a; - } else { - struct hostent *hp = gethostbyname(host); - if(hp == 0 || hp->h_length != 4){ - fprintf(stderr, "cannot find host name %s\n", host); - exit(1); - } - dst->sin_addr.s_addr = ((struct in_addr *)(hp->h_addr))->s_addr; - } - dst->sin_port = htons(atoi(port)); -} + sockaddr_in dst{}; // zero initialize + dst.sin_family = AF_INET; -int -cmp_timespec(const struct timespec &a, const struct timespec &b) -{ - if(a.tv_sec > b.tv_sec) - return 1; - else if(a.tv_sec < b.tv_sec) - return -1; - else { - if(a.tv_nsec > b.tv_nsec) - return 1; - else if(a.tv_nsec < b.tv_nsec) - return -1; - else - return 0; - } -} + struct in_addr a{inet_addr(host.c_str())}; -void -add_timespec(const struct timespec &a, int b, struct timespec *result) -{ - // convert to millisec, add timeout, convert back - result->tv_sec = a.tv_sec + b/1000; - result->tv_nsec = a.tv_nsec + (b % 1000) * 1000000; - VERIFY(result->tv_nsec >= 0); - while (result->tv_nsec > 1000000000){ - result->tv_sec++; - result->tv_nsec-=1000000000; - } -} + if (a.s_addr != INADDR_NONE) + dst.sin_addr.s_addr = a.s_addr; + else { + struct hostent *hp = gethostbyname(host.c_str()); -int -diff_timespec(const struct timespec &end, const struct timespec &start) -{ - int diff = (end.tv_sec > start.tv_sec)?(end.tv_sec-start.tv_sec)*1000:0; - VERIFY(diff || end.tv_sec == start.tv_sec); - if(end.tv_nsec > start.tv_nsec){ - diff += (end.tv_nsec-start.tv_nsec)/1000000; - } else { - diff -= (start.tv_nsec-end.tv_nsec)/1000000; - } - return diff; + if (!hp || hp->h_length != 4 || hp->h_addrtype != AF_INET) { + LOG_NONMEMBER("cannot find host name " << host); + exit(1); + } + memcpy(&a, hp->h_addr_list[0], sizeof(in_addr_t)); + dst.sin_addr.s_addr = a.s_addr; + } + dst.sin_port = hton((in_port_t)stoi(port)); + return dst; }