Initial work of multithreaded key-db. Note: Fails tests

src/ae.cpp

@@ -30,7 +30,9 @@
  * POSSIBILITY OF SUCH DAMAGE.
  */
 
+#include <mutex>
 #include <stdio.h>
+#include <fcntl.h>
 #include <sys/time.h>
 #include <sys/types.h>
 #include <unistd.h>
@@ -41,11 +43,17 @@
 #include <errno.h>
 
 #include "ae.h"
+#include "fastlock.h"
 extern "C" {
 #include "zmalloc.h"
 #include "config.h"
 }
 
+fastlock g_lock;
+thread_local aeEventLoop *g_eventLoopThisThread = NULL;
+
+#define AE_ASSERT(x) if (!(x)) do { fprintf(stderr, "AE_ASSER FAILURE\n"); *((volatile int*)0) = 1; } while(0)
+
 /* Include the best multiplexing layer supported by this system.
  * The following should be ordered by performances, descending. */
 #ifdef HAVE_EVPORT
@@ -62,6 +70,59 @@ extern "C" {
     #endif
 #endif
 
+enum class AE_ASYNC_OP
+{
+    PostFunction,
+    DeleteFileEvent,
+};
+typedef struct aeCommand
+{
+    AE_ASYNC_OP op;
+    int fd; 
+    int mask;
+    aePostFunctionProc *proc;
+    void *clientData;
+} aeCommand;
+
+void aeProcessCmd(aeEventLoop *eventLoop, int fd, void *, int )
+{
+    aeCommand cmd;
+    for (;;)
+    {
+        auto cb = read(fd, &cmd, sizeof(aeCommand));
+        if (cb != sizeof(cmd))
+        {
+            if (errno == EAGAIN)
+                break;
+            fprintf(stderr, "Failed to read pipe.\n");
+        }
+        switch (cmd.op)
+        {
+        case AE_ASYNC_OP::DeleteFileEvent:
+            aeDeleteFileEvent(eventLoop, cmd.fd, cmd.mask);
+            break;
+
+        case AE_ASYNC_OP::PostFunction:
+            {
+            std::unique_lock<decltype(g_lock)> ulock(g_lock);
+            ((aePostFunctionProc*)cmd.proc)(cmd.clientData);
+            break;
+            }
+        }
+    }
+}
+
+int aePostFunction(aeEventLoop *eventLoop, aePostFunctionProc *proc, void *arg)
+{
+    aeCommand cmd;
+    cmd.op = AE_ASYNC_OP::PostFunction;
+    cmd.proc = proc;
+    cmd.clientData = arg;
+    auto size = write(eventLoop->fdCmdWrite, &cmd, sizeof(cmd));
+    AE_ASSERT(size == sizeof(cmd));
+    return AE_OK;
+}
+
 aeEventLoop *aeCreateEventLoop(int setsize) {
     aeEventLoop *eventLoop;
     int i;
@@ -83,6 +144,16 @@ aeEventLoop *aeCreateEventLoop(int setsize) {
      * vector with it. */
     for (i = 0; i < setsize; i++)
         eventLoop->events[i].mask = AE_NONE;
+
+    fastlock_init(&eventLoop->flock);
+    int rgfd[2];
+    if (pipe(rgfd) < 0)
+        goto err;
+    eventLoop->fdCmdRead = rgfd[0];
+    eventLoop->fdCmdWrite = rgfd[1];
+    fcntl(eventLoop->fdCmdRead, F_SETFL, O_NONBLOCK);
+    aeCreateFileEvent(eventLoop, eventLoop->fdCmdRead, AE_READABLE|AE_THREADSAFE, aeProcessCmd, NULL);
+
     return eventLoop;
 
 err:
@@ -107,6 +178,7 @@ int aeGetSetSize(aeEventLoop *eventLoop) {
  *
  * Otherwise AE_OK is returned and the operation is successful. */
 int aeResizeSetSize(aeEventLoop *eventLoop, int setsize) {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     int i;
 
     if (setsize == eventLoop->setsize) return AE_OK;
@@ -129,19 +201,25 @@ extern "C" void aeDeleteEventLoop(aeEventLoop *eventLoop) {
     zfree(eventLoop->events);
     zfree(eventLoop->fired);
     zfree(eventLoop);
+    fastlock_free(&eventLoop->flock);
+    close(eventLoop->fdCmdRead);
+    close(eventLoop->fdCmdWrite);
 }
 
 extern "C" void aeStop(aeEventLoop *eventLoop) {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     eventLoop->stop = 1;
 }
 
 extern "C" int aeCreateFileEvent(aeEventLoop *eventLoop, int fd, int mask,
         aeFileProc *proc, void *clientData)
 {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     if (fd >= eventLoop->setsize) {
         errno = ERANGE;
         return AE_ERR;
     }
+
     aeFileEvent *fe = &eventLoop->events[fd];
 
     if (aeApiAddEvent(eventLoop, fd, mask) == -1)
@@ -155,8 +233,22 @@ extern "C" int aeCreateFileEvent(aeEventLoop *eventLoop, int fd, int mask,
     return AE_OK;
 }
 
+void aeDeleteFileEventAsync(aeEventLoop *eventLoop, int fd, int mask)
+{
+    if (eventLoop == g_eventLoopThisThread)
+        return aeDeleteFileEvent(eventLoop, fd, mask);
+    aeCommand cmd;
+    cmd.op = AE_ASYNC_OP::DeleteFileEvent;
+    cmd.fd = fd;
+    cmd.mask = mask;
+    auto cb = write(eventLoop->fdCmdWrite, &cmd, sizeof(cmd));
+    if (cb != sizeof(cmd))
+        fprintf(stderr, "Failed to write to pipe.\n");
+}
+
 extern "C" void aeDeleteFileEvent(aeEventLoop *eventLoop, int fd, int mask)
 {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     if (fd >= eventLoop->setsize) return;
     aeFileEvent *fe = &eventLoop->events[fd];
     if (fe->mask == AE_NONE) return;
@@ -178,6 +270,7 @@ extern "C" void aeDeleteFileEvent(aeEventLoop *eventLoop, int fd, int mask)
 }
 
 extern "C" int aeGetFileEvents(aeEventLoop *eventLoop, int fd) {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     if (fd >= eventLoop->setsize) return 0;
     aeFileEvent *fe = &eventLoop->events[fd];
 
@@ -211,6 +304,7 @@ extern "C" long long aeCreateTimeEvent(aeEventLoop *eventLoop, long long millise
         aeTimeProc *proc, void *clientData,
         aeEventFinalizerProc *finalizerProc)
 {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     long long id = eventLoop->timeEventNextId++;
     aeTimeEvent *te;
 
@@ -231,6 +325,7 @@ extern "C" long long aeCreateTimeEvent(aeEventLoop *eventLoop, long long millise
 
 extern "C" int aeDeleteTimeEvent(aeEventLoop *eventLoop, long long id)
 {
+    AE_ASSERT(g_eventLoopThisThread == NULL || g_eventLoopThisThread == eventLoop);
     aeTimeEvent *te = eventLoop->timeEventHead;
     while(te) {
         if (te->id == id) {
@@ -270,6 +365,7 @@ static aeTimeEvent *aeSearchNearestTimer(aeEventLoop *eventLoop)
 
 /* Process time events */
 static int processTimeEvents(aeEventLoop *eventLoop) {
+    std::unique_lock<decltype(g_lock)> ulock(g_lock);
     int processed = 0;
     aeTimeEvent *te;
     long long maxId;
@@ -343,6 +439,62 @@ static int processTimeEvents(aeEventLoop *eventLoop) {
     return processed;
 }
 
+extern "C" void ProcessEventCore(aeEventLoop *eventLoop, aeFileEvent *fe, int mask, int fd)
+{
+#define LOCK_IF_NECESSARY(fe) \
+    std::unique_lock<decltype(g_lock)> ulock(g_lock, std::defer_lock); \
+    if (!(fe->mask & AE_THREADSAFE)) \
+        ulock.lock()
+
+    int fired = 0; /* Number of events fired for current fd. */
+
+    /* Normally we execute the readable event first, and the writable
+    * event laster. This is useful as sometimes we may be able
+    * to serve the reply of a query immediately after processing the
+    * query.
+    *
+    * However if AE_BARRIER is set in the mask, our application is
+    * asking us to do the reverse: never fire the writable event
+    * after the readable. In such a case, we invert the calls.
+    * This is useful when, for instance, we want to do things
+    * in the beforeSleep() hook, like fsynching a file to disk,
+    * before replying to a client. */
+    int invert = fe->mask & AE_BARRIER;
+
+    /* Note the "fe->mask & mask & ..." code: maybe an already
+        * processed event removed an element that fired and we still
+        * didn't processed, so we check if the event is still valid.
+        *
+        * Fire the readable event if the call sequence is not
+        * inverted. */
+    if (!invert && fe->mask & mask & AE_READABLE) {
+        LOCK_IF_NECESSARY(fe);
+        fe->rfileProc(eventLoop,fd,fe->clientData,mask);
+        fired++;
+    }
+
+    /* Fire the writable event. */
+    if (fe->mask & mask & AE_WRITABLE) {
+        if (!fired || fe->wfileProc != fe->rfileProc) {
+            LOCK_IF_NECESSARY(fe);
+            fe->wfileProc(eventLoop,fd,fe->clientData,mask);
+            fired++;
+        }
+    }
+
+    /* If we have to invert the call, fire the readable event now
+        * after the writable one. */
+    if (invert && fe->mask & mask & AE_READABLE) {
+        if (!fired || fe->wfileProc != fe->rfileProc) {
+            LOCK_IF_NECESSARY(fe);
+            fe->rfileProc(eventLoop,fd,fe->clientData,mask);
+            fired++;
+        }
+    }
+
+#undef LOCK_IF_NECESSARY
+}
+
 /* Process every pending time event, then every pending file event
  * (that may be registered by time event callbacks just processed).
  * Without special flags the function sleeps until some file event
@@ -413,55 +565,19 @@ int aeProcessEvents(aeEventLoop *eventLoop, int flags)
         numevents = aeApiPoll(eventLoop, tvp);
 
         /* After sleep callback. */
-        if (eventLoop->aftersleep != NULL && flags & AE_CALL_AFTER_SLEEP)
+        if (eventLoop->aftersleep != NULL && flags & AE_CALL_AFTER_SLEEP) {
+            std::unique_lock<decltype(g_lock)> ulock(g_lock, std::defer_lock);
+            if (!(eventLoop->beforesleepFlags & AE_THREADSAFE))
+                ulock.lock();
             eventLoop->aftersleep(eventLoop);
+        }
 
         for (j = 0; j < numevents; j++) {
             aeFileEvent *fe = &eventLoop->events[eventLoop->fired[j].fd];
             int mask = eventLoop->fired[j].mask;
             int fd = eventLoop->fired[j].fd;
-            int fired = 0; /* Number of events fired for current fd. */
 
-            /* Normally we execute the readable event first, and the writable
+            ProcessEventCore(eventLoop, fe, mask, fd);
-             * event laster. This is useful as sometimes we may be able
-             * to serve the reply of a query immediately after processing the
-             * query.
-             *
-             * However if AE_BARRIER is set in the mask, our application is
-             * asking us to do the reverse: never fire the writable event
-             * after the readable. In such a case, we invert the calls.
-             * This is useful when, for instance, we want to do things
-             * in the beforeSleep() hook, like fsynching a file to disk,
-             * before replying to a client. */
-            int invert = fe->mask & AE_BARRIER;
-
-            /* Note the "fe->mask & mask & ..." code: maybe an already
-             * processed event removed an element that fired and we still
-             * didn't processed, so we check if the event is still valid.
-             *
-             * Fire the readable event if the call sequence is not
-             * inverted. */
-            if (!invert && fe->mask & mask & AE_READABLE) {
-                fe->rfileProc(eventLoop,fd,fe->clientData,mask);
-                fired++;
-            }
-
-            /* Fire the writable event. */
-            if (fe->mask & mask & AE_WRITABLE) {
-                if (!fired || fe->wfileProc != fe->rfileProc) {
-                    fe->wfileProc(eventLoop,fd,fe->clientData,mask);
-                    fired++;
-                }
-            }
-
-            /* If we have to invert the call, fire the readable event now
-             * after the writable one. */
-            if (invert && fe->mask & mask & AE_READABLE) {
-                if (!fired || fe->wfileProc != fe->rfileProc) {
-                    fe->rfileProc(eventLoop,fd,fe->clientData,mask);
-                    fired++;
-                }
-            }
 
             processed++;
         }
@@ -497,9 +613,14 @@ int aeWait(int fd, int mask, long long milliseconds) {
 
 void aeMain(aeEventLoop *eventLoop) {
     eventLoop->stop = 0;
+    g_eventLoopThisThread = eventLoop;
     while (!eventLoop->stop) {
-        if (eventLoop->beforesleep != NULL)
+        if (eventLoop->beforesleep != NULL) {
+            std::unique_lock<decltype(g_lock)> ulock(g_lock, std::defer_lock);
+            if (!(eventLoop->beforesleepFlags & AE_THREADSAFE))
+                ulock.lock();
             eventLoop->beforesleep(eventLoop);
+        }
         aeProcessEvents(eventLoop, AE_ALL_EVENTS|AE_CALL_AFTER_SLEEP);
     }
 }
@@ -508,10 +629,22 @@ const char *aeGetApiName(void) {
     return aeApiName();
 }
 
-void aeSetBeforeSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *beforesleep) {
+void aeSetBeforeSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *beforesleep, int flags) {
     eventLoop->beforesleep = beforesleep;
+    eventLoop->beforesleepFlags = flags;
 }
 
-void aeSetAfterSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *aftersleep) {
+void aeSetAfterSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *aftersleep, int flags) {
     eventLoop->aftersleep = aftersleep;
+    eventLoop->aftersleepFlags = flags;
 }
+
+void aeAcquireLock()
+{
+    g_lock.lock();
+}
+
+void aeReleaseLock()
+{
+    g_lock.unlock();
+}

src/ae.h

@@ -34,6 +34,7 @@
 #define __AE_H__
 
 #include <time.h>
+#include "fastlock.h"
 
 #ifdef __cplusplus
 extern "C" {
@@ -71,6 +72,7 @@ typedef void aeFileProc(struct aeEventLoop *eventLoop, int fd, void *clientData,
 typedef int aeTimeProc(struct aeEventLoop *eventLoop, long long id, void *clientData);
 typedef void aeEventFinalizerProc(struct aeEventLoop *eventLoop, void *clientData);
 typedef void aeBeforeSleepProc(struct aeEventLoop *eventLoop);
+typedef void aePostFunctionProc(void *pvArgs);
 
 /* File event structure */
 typedef struct aeFileEvent {
@@ -110,16 +112,23 @@ typedef struct aeEventLoop {
     int stop;
     void *apidata; /* This is used for polling API specific data */
     aeBeforeSleepProc *beforesleep;
+    int beforesleepFlags;
     aeBeforeSleepProc *aftersleep;
+    int aftersleepFlags;
+    struct fastlock flock;
+    int fdCmdWrite;
+    int fdCmdRead;
 } aeEventLoop;
 
 /* Prototypes */
 aeEventLoop *aeCreateEventLoop(int setsize);
+int aePostFunction(aeEventLoop *eventLoop, aePostFunctionProc *proc, void *arg);
 void aeDeleteEventLoop(aeEventLoop *eventLoop);
 void aeStop(aeEventLoop *eventLoop);
 int aeCreateFileEvent(aeEventLoop *eventLoop, int fd, int mask,
         aeFileProc *proc, void *clientData);
 void aeDeleteFileEvent(aeEventLoop *eventLoop, int fd, int mask);
+void aeDeleteFileEventAsync(aeEventLoop *eventLoop, int fd, int mask);
 int aeGetFileEvents(aeEventLoop *eventLoop, int fd);
 long long aeCreateTimeEvent(aeEventLoop *eventLoop, long long milliseconds,
         aeTimeProc *proc, void *clientData,
@@ -129,11 +138,14 @@ int aeProcessEvents(aeEventLoop *eventLoop, int flags);
 int aeWait(int fd, int mask, long long milliseconds);
 void aeMain(aeEventLoop *eventLoop);
 const char *aeGetApiName(void);
-void aeSetBeforeSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *beforesleep);
+void aeSetBeforeSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *beforesleep, int flags);
-void aeSetAfterSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *aftersleep);
+void aeSetAfterSleepProc(aeEventLoop *eventLoop, aeBeforeSleepProc *aftersleep, int flags);
 int aeGetSetSize(aeEventLoop *eventLoop);
 int aeResizeSetSize(aeEventLoop *eventLoop, int setsize);
 
+void aeAcquireLock();
+void aeReleaseLock();
+
 #ifdef __cplusplus
 }
 #endif

src/ae_epoll.cpp

@@ -83,7 +83,11 @@ static int aeApiAddEvent(aeEventLoop *eventLoop, int fd, int mask) {
     if (mask & AE_READABLE) ee.events |= EPOLLIN;
     if (mask & AE_WRITABLE) ee.events |= EPOLLOUT;
     ee.data.fd = fd;
-    if (epoll_ctl(state->epfd,op,fd,&ee) == -1) return -1;
+    if (epoll_ctl(state->epfd,op,fd,&ee) == -1)
+    {
+        perror("epoll_ctl failed");
+        return -1;
+    }
     return 0;
 }
 

src/aof.c

@@ -105,7 +105,7 @@ void aofChildWriteDiffData(aeEventLoop *el, int fd, void *privdata, int mask) {
         ln = listFirst(server.aof_rewrite_buf_blocks);
         block = ln ? ln->value : NULL;
         if (server.aof_stop_sending_diff || !block) {
-            aeDeleteFileEvent(server.el,server.aof_pipe_write_data_to_child,
+            aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.aof_pipe_write_data_to_child,
                               AE_WRITABLE);
             return;
         }
@@ -162,8 +162,8 @@ void aofRewriteBufferAppend(unsigned char *s, unsigned long len) {
 
     /* Install a file event to send data to the rewrite child if there is
      * not one already. */
-    if (aeGetFileEvents(server.el,server.aof_pipe_write_data_to_child) == 0) {
+    if (aeGetFileEvents(server.rgel[IDX_EVENT_LOOP_MAIN],server.aof_pipe_write_data_to_child) == 0) {
-        aeCreateFileEvent(server.el, server.aof_pipe_write_data_to_child,
+        aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], server.aof_pipe_write_data_to_child,
             AE_WRITABLE, aofChildWriteDiffData, NULL);
     }
 }
@@ -738,7 +738,7 @@ int loadAppendOnlyFile(char *filename) {
         /* Serve the clients from time to time */
         if (!(loops++ % 1000)) {
             loadingProgress(ftello(fp));
-            processEventsWhileBlocked();
+            processEventsWhileBlocked(IDX_EVENT_LOOP_MAIN);
         }
 
         if (fgets(buf,sizeof(buf),fp) == NULL) {
@@ -1470,7 +1470,7 @@ void aofChildPipeReadable(aeEventLoop *el, int fd, void *privdata, int mask) {
     }
     /* Remove the handler since this can be called only one time during a
      * rewrite. */
-    aeDeleteFileEvent(server.el,server.aof_pipe_read_ack_from_child,AE_READABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.aof_pipe_read_ack_from_child,AE_READABLE);
 }
 
 /* Create the pipes used for parent - child process IPC during rewrite.
@@ -1488,7 +1488,7 @@ int aofCreatePipes(void) {
     /* Parent -> children data is non blocking. */
     if (anetNonBlock(NULL,fds[0]) != ANET_OK) goto error;
     if (anetNonBlock(NULL,fds[1]) != ANET_OK) goto error;
-    if (aeCreateFileEvent(server.el, fds[2], AE_READABLE, aofChildPipeReadable, NULL) == AE_ERR) goto error;
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], fds[2], AE_READABLE, aofChildPipeReadable, NULL) == AE_ERR) goto error;
 
     server.aof_pipe_write_data_to_child = fds[1];
     server.aof_pipe_read_data_from_parent = fds[0];
@@ -1507,8 +1507,8 @@ error:
 }
 
 void aofClosePipes(void) {
-    aeDeleteFileEvent(server.el,server.aof_pipe_read_ack_from_child,AE_READABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.aof_pipe_read_ack_from_child,AE_READABLE);
-    aeDeleteFileEvent(server.el,server.aof_pipe_write_data_to_child,AE_WRITABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.aof_pipe_write_data_to_child,AE_WRITABLE);
     close(server.aof_pipe_write_data_to_child);
     close(server.aof_pipe_read_data_from_parent);
     close(server.aof_pipe_write_ack_to_parent);

src/blocked.c

@@ -109,15 +109,15 @@ void blockClient(client *c, int btype) {
 /* This function is called in the beforeSleep() function of the event loop
  * in order to process the pending input buffer of clients that were
  * unblocked after a blocking operation. */
-void processUnblockedClients(void) {
+void processUnblockedClients(int iel) {
     listNode *ln;
     client *c;
 
-    while (listLength(server.unblocked_clients)) {
+    while (listLength(server.rgunblocked_clients[iel])) {
-        ln = listFirst(server.unblocked_clients);
+        ln = listFirst(server.rgunblocked_clients[iel]);
         serverAssert(ln != NULL);
         c = ln->value;
-        listDelNode(server.unblocked_clients,ln);
+        listDelNode(server.rgunblocked_clients[iel],ln);
         c->flags &= ~CLIENT_UNBLOCKED;
 
         /* Process remaining data in the input buffer, unless the client
@@ -153,7 +153,7 @@ void queueClientForReprocessing(client *c) {
      * blocking operation, don't add back it into the list multiple times. */
     if (!(c->flags & CLIENT_UNBLOCKED)) {
         c->flags |= CLIENT_UNBLOCKED;
-        listAddNodeTail(server.unblocked_clients,c);
+        listAddNodeTail(server.rgunblocked_clients[c->iel],c);
     }
 }
 

src/cluster.c

@@ -493,7 +493,7 @@ void clusterInit(void) {
         int j;
 
         for (j = 0; j < server.cfd_count; j++) {
-            if (aeCreateFileEvent(server.el, server.cfd[j], AE_READABLE,
+            if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], server.cfd[j], AE_READABLE,
                 clusterAcceptHandler, NULL) == AE_ERR)
                     serverPanic("Unrecoverable error creating Redis Cluster "
                                 "file event.");
@@ -601,7 +601,7 @@ clusterLink *createClusterLink(clusterNode *node) {
  * with this link will have the 'link' field set to NULL. */
 void freeClusterLink(clusterLink *link) {
     if (link->fd != -1) {
-        aeDeleteFileEvent(server.el, link->fd, AE_READABLE|AE_WRITABLE);
+        aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], link->fd, AE_READABLE|AE_WRITABLE);
     }
     sdsfree(link->sndbuf);
     sdsfree(link->rcvbuf);
@@ -645,7 +645,7 @@ void clusterAcceptHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
          * node identity. */
         link = createClusterLink(NULL);
         link->fd = cfd;
-        aeCreateFileEvent(server.el,cfd,AE_READABLE,clusterReadHandler,link);
+        aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],cfd,AE_READABLE,clusterReadHandler,link);
     }
 }
 
@@ -2132,7 +2132,7 @@ void clusterWriteHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
     }
     sdsrange(link->sndbuf,nwritten,-1);
     if (sdslen(link->sndbuf) == 0)
-        aeDeleteFileEvent(server.el, link->fd, AE_WRITABLE);
+        aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], link->fd, AE_WRITABLE);
 }
 
 /* Read data. Try to read the first field of the header first to check the
@@ -2208,7 +2208,7 @@ void clusterReadHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
  * from event handlers that will do stuff with the same link later. */
 void clusterSendMessage(clusterLink *link, unsigned char *msg, size_t msglen) {
     if (sdslen(link->sndbuf) == 0 && msglen != 0)
-        aeCreateFileEvent(server.el,link->fd,AE_WRITABLE|AE_BARRIER,
+        aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],link->fd,AE_WRITABLE|AE_BARRIER,
                     clusterWriteHandler,link);
 
     link->sndbuf = sdscatlen(link->sndbuf, msg, msglen);
@@ -3402,7 +3402,7 @@ void clusterCron(void) {
             link = createClusterLink(node);
             link->fd = fd;
             node->link = link;
-            aeCreateFileEvent(server.el,link->fd,AE_READABLE,
+            aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],link->fd,AE_READABLE,
                     clusterReadHandler,link);
             /* Queue a PING in the new connection ASAP: this is crucial
              * to avoid false positives in failure detection.

src/config.c

@@ -968,15 +968,18 @@ void configSetCommand(client *c) {
                 server.maxclients = orig_value;
                 return;
             }
-            if ((unsigned int) aeGetSetSize(server.el) <
+            if ((unsigned int) aeGetSetSize(server.rgel[IDX_EVENT_LOOP_MAIN]) <
                 server.maxclients + CONFIG_FDSET_INCR)
             {
-                if (aeResizeSetSize(server.el,
+                for (int iel = 0; iel < MAX_EVENT_LOOPS; ++iel)
-                    server.maxclients + CONFIG_FDSET_INCR) == AE_ERR)
                 {
-                    addReplyError(c,"The event loop API used by Redis is not able to handle the specified number of clients");
+                    if (aeResizeSetSize(server.rgel[iel],
-                    server.maxclients = orig_value;
+                        server.maxclients + CONFIG_FDSET_INCR) == AE_ERR)
-                    return;
+                    {
+                        addReplyError(c,"The event loop API used by Redis is not able to handle the specified number of clients");
+                        server.maxclients = orig_value;
+                        return;
+                    }
                 }
             }
         }

src/fastlock.h

@@ -21,6 +21,11 @@ struct fastlock
     int m_lock;
 
 #ifdef __cplusplus
+    fastlock()
+    {
+        fastlock_init(this);
+    }
+
     void lock()
     {
         fastlock_lock(this);

src/module.c

@@ -2696,7 +2696,7 @@ RedisModuleCallReply *RM_Call(RedisModuleCtx *ctx, const char *cmdname, const ch
 
     /* Create the client and dispatch the command. */
     va_start(ap, fmt);
-    c = createClient(-1);
+    c = createClient(-1, IDX_EVENT_LOOP_MAIN);
     c->puser = NULL; /* Root user. */
     argv = moduleCreateArgvFromUserFormat(cmdname,fmt,&argc,&flags,ap);
     replicate = flags & REDISMODULE_ARGV_REPLICATE;
@@ -3546,7 +3546,7 @@ RedisModuleBlockedClient *RM_BlockClient(RedisModuleCtx *ctx, RedisModuleCmdFunc
     bc->disconnect_callback = NULL; /* Set by RM_SetDisconnectCallback() */
     bc->free_privdata = free_privdata;
     bc->privdata = NULL;
-    bc->reply_client = createClient(-1);
+    bc->reply_client = createClient(-1, IDX_EVENT_LOOP_MAIN);
     bc->reply_client->flags |= CLIENT_MODULE;
     bc->dbid = c->db->id;
     c->bpop.timeout = timeout_ms ? (mstime()+timeout_ms) : 0;
@@ -3692,7 +3692,7 @@ void moduleHandleBlockedClients(void) {
                 !(c->flags & CLIENT_PENDING_WRITE))
             {
                 c->flags |= CLIENT_PENDING_WRITE;
-                listAddNodeHead(server.clients_pending_write,c);
+                listAddNodeHead(server.rgclients_pending_write[IDX_EVENT_LOOP_MAIN],c);
             }
         }
 
@@ -3794,7 +3794,7 @@ RedisModuleCtx *RM_GetThreadSafeContext(RedisModuleBlockedClient *bc) {
      * access it safely from another thread, so we create a fake client here
      * in order to keep things like the currently selected database and similar
      * things. */
-    ctx->client = createClient(-1);
+    ctx->client = createClient(-1, IDX_EVENT_LOOP_MAIN);
     if (bc) selectDb(ctx->client,bc->dbid);
     return ctx;
 }
@@ -4300,7 +4300,7 @@ RedisModuleTimerID RM_CreateTimer(RedisModuleCtx *ctx, mstime_t period, RedisMod
         if (memcmp(ri.key,&key,sizeof(key)) == 0) {
             /* This is the first key, we need to re-install the timer according
              * to the just added event. */
-            aeDeleteTimeEvent(server.el,aeTimer);
+            aeDeleteTimeEvent(server.rgel[IDX_EVENT_LOOP_MAIN],aeTimer);
             aeTimer = -1;
         }
         raxStop(&ri);
@@ -4309,7 +4309,7 @@ RedisModuleTimerID RM_CreateTimer(RedisModuleCtx *ctx, mstime_t period, RedisMod
     /* If we have no main timer (the old one was invalidated, or this is the
      * first module timer we have), install one. */
     if (aeTimer == -1)
-        aeTimer = aeCreateTimeEvent(server.el,period,moduleTimerHandler,NULL,NULL);
+        aeTimer = aeCreateTimeEvent(server.rgel[IDX_EVENT_LOOP_MAIN],period,moduleTimerHandler,NULL,NULL);
 
     return key;
 }
@@ -4659,7 +4659,7 @@ void moduleInitModulesSystem(void) {
 
     /* Set up the keyspace notification susbscriber list and static client */
     moduleKeyspaceSubscribers = listCreate();
-    moduleFreeContextReusedClient = createClient(-1);
+    moduleFreeContextReusedClient = createClient(-1, IDX_EVENT_LOOP_MAIN);
     moduleFreeContextReusedClient->flags |= CLIENT_MODULE;
     moduleFreeContextReusedClient->puser = NULL; /* root user. */
 

src/networking.c

@@ -82,7 +82,7 @@ void linkClient(client *c) {
     raxInsert(server.clients_index,(unsigned char*)&id,sizeof(id),c,NULL);
 }
 
-client *createClient(int fd) {
+client *createClient(int fd, int iel) {
     client *c = zmalloc(sizeof(client), MALLOC_LOCAL);
 
     /* passing -1 as fd it is possible to create a non connected client.
@@ -94,7 +94,7 @@ client *createClient(int fd) {
         anetEnableTcpNoDelay(NULL,fd);
         if (server.tcpkeepalive)
             anetKeepAlive(NULL,fd,server.tcpkeepalive);
-        if (aeCreateFileEvent(server.el,fd,AE_READABLE,
+        if (aeCreateFileEvent(server.rgel[iel],fd,AE_READABLE|AE_THREADSAFE,
             readQueryFromClient, c) == AE_ERR)
         {
             close(fd);
@@ -106,6 +106,7 @@ client *createClient(int fd) {
     selectDb(c,0);
     uint64_t client_id;
     atomicGetIncr(server.next_client_id,client_id,1);
+    c->iel = iel;
     c->id = client_id;
     c->resp = 2;
     c->fd = fd;
@@ -186,7 +187,7 @@ void clientInstallWriteHandler(client *c) {
          * a system call. We'll only really install the write handler if
          * we'll not be able to write the whole reply at once. */
         c->flags |= CLIENT_PENDING_WRITE;
-        listAddNodeHead(server.clients_pending_write,c);
+        listAddNodeHead(server.rgclients_pending_write[c->iel],c);
     }
 }
 
@@ -779,9 +780,9 @@ int clientHasPendingReplies(client *c) {
 }
 
 #define MAX_ACCEPTS_PER_CALL 1000
-static void acceptCommonHandler(int fd, int flags, char *ip) {
+static void acceptCommonHandler(int fd, int flags, char *ip, int iel) {
     client *c;
-    if ((c = createClient(fd)) == NULL) {
+    if ((c = createClient(fd, iel)) == NULL) {
         serverLog(LL_WARNING,
             "Error registering fd event for the new client: %s (fd=%d)",
             strerror(errno),fd);
@@ -849,6 +850,32 @@ static void acceptCommonHandler(int fd, int flags, char *ip) {
     c->flags |= flags;
 }
 
+struct AcceptCommonHandlerAsyncArgs
+{
+    int fd;
+    int flags;
+    char cip[NET_IP_STR_LEN];
+    int fUseCip;
+    int iel;
+};
+static void AcceptCommonHandlerAsync(void *args)
+{
+    struct AcceptCommonHandlerAsyncArgs *aargs = args;
+    acceptCommonHandler(aargs->fd, aargs->flags, aargs->cip, aargs->iel);
+    zfree(args);
+}
+static void EnqueueAcceptCommonHandler(int fd, int flags, char *ip, int iel)
+{
+    struct AcceptCommonHandlerAsyncArgs *args = zmalloc(sizeof(struct AcceptCommonHandlerAsyncArgs), MALLOC_LOCAL);
+    args->fd = fd;
+    args->flags = flags;
+    if (ip != NULL)
+        memcpy(args->cip, ip, NET_IP_STR_LEN);
+    args->fUseCip = (ip != NULL);
+    args->iel = iel;
+    aePostFunction(server.rgel[iel], AcceptCommonHandlerAsync, args);
+}
+
 void acceptTcpHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
     int cport, cfd, max = MAX_ACCEPTS_PER_CALL;
     char cip[NET_IP_STR_LEN];
@@ -865,7 +892,24 @@ void acceptTcpHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
             return;
         }
         serverLog(LL_VERBOSE,"Accepted %s:%d", cip, cport);
-        acceptCommonHandler(cfd,0,cip);
+        int ielCur = 0;
+        for (; ielCur < MAX_EVENT_LOOPS; ++ielCur)
+        {
+            if (el == server.rgel[ielCur])
+                break;
+        }
+        serverAssert(ielCur < MAX_EVENT_LOOPS);
+        int iel = rand() % MAX_EVENT_LOOPS;
+        if (iel == ielCur)
+        {
+            aeAcquireLock();
+            acceptCommonHandler(cfd,0,cip, iel);
+            aeReleaseLock();
+        }
+        else
+        {
+            EnqueueAcceptCommonHandler(cfd, 0, cip, iel);
+        }
     }
 }
 
@@ -884,7 +928,7 @@ void acceptUnixHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
             return;
         }
         serverLog(LL_VERBOSE,"Accepted connection to %s", server.unixsocket);
-        acceptCommonHandler(cfd,CLIENT_UNIX_SOCKET,NULL);
+        EnqueueAcceptCommonHandler(cfd,CLIENT_UNIX_SOCKET,NULL, rand() % MAX_EVENT_LOOPS);
     }
 }
 
@@ -928,26 +972,26 @@ void unlinkClient(client *c) {
         }
 
         /* Unregister async I/O handlers and close the socket. */
-        aeDeleteFileEvent(server.el,c->fd,AE_READABLE);
+        aeDeleteFileEventAsync(server.rgel[c->iel],c->fd,AE_READABLE);
-        aeDeleteFileEvent(server.el,c->fd,AE_WRITABLE);
+        aeDeleteFileEventAsync(server.rgel[c->iel],c->fd,AE_WRITABLE);
         close(c->fd);
         c->fd = -1;
     }
 
     /* Remove from the list of pending writes if needed. */
     if (c->flags & CLIENT_PENDING_WRITE) {
-        ln = listSearchKey(server.clients_pending_write,c);
+        ln = listSearchKey(server.rgclients_pending_write[c->iel],c);
         serverAssert(ln != NULL);
-        listDelNode(server.clients_pending_write,ln);
+        listDelNode(server.rgclients_pending_write[c->iel],ln);
         c->flags &= ~CLIENT_PENDING_WRITE;
     }
 
     /* When client was just unblocked because of a blocking operation,
      * remove it from the list of unblocked clients. */
     if (c->flags & CLIENT_UNBLOCKED) {
-        ln = listSearchKey(server.unblocked_clients,c);
+        ln = listSearchKey(server.rgunblocked_clients[c->iel],c);
         serverAssert(ln != NULL);
-        listDelNode(server.unblocked_clients,ln);
+        listDelNode(server.rgunblocked_clients[c->iel],ln);
         c->flags &= ~CLIENT_UNBLOCKED;
     }
 }
@@ -1092,6 +1136,7 @@ int writeToClient(int fd, client *c, int handler_installed) {
     while(clientHasPendingReplies(c)) {
         if (c->bufpos > 0) {
             nwritten = write(fd,c->buf+c->sentlen,c->bufpos-c->sentlen);
+
             if (nwritten <= 0) break;
             c->sentlen += nwritten;
             totwritten += nwritten;
@@ -1113,6 +1158,7 @@ int writeToClient(int fd, client *c, int handler_installed) {
             }
 
             nwritten = write(fd, o->buf + c->sentlen, objlen - c->sentlen);
+
             if (nwritten <= 0) break;
             c->sentlen += nwritten;
             totwritten += nwritten;
@@ -1145,7 +1191,8 @@ int writeToClient(int fd, client *c, int handler_installed) {
              zmalloc_used_memory() < server.maxmemory) &&
             !(c->flags & CLIENT_SLAVE)) break;
     }
-    server.stat_net_output_bytes += totwritten;
+    
+    __atomic_fetch_add(&server.stat_net_output_bytes, totwritten, __ATOMIC_RELAXED);
     if (nwritten == -1) {
         if (errno == EAGAIN) {
             nwritten = 0;
@@ -1165,7 +1212,7 @@ int writeToClient(int fd, client *c, int handler_installed) {
     }
     if (!clientHasPendingReplies(c)) {
         c->sentlen = 0;
-        if (handler_installed) aeDeleteFileEvent(server.el,c->fd,AE_WRITABLE);
+        if (handler_installed) aeDeleteFileEvent(server.rgel[c->iel],c->fd,AE_WRITABLE);
 
         /* Close connection after entire reply has been sent. */
         if (c->flags & CLIENT_CLOSE_AFTER_REPLY) {
@@ -1187,16 +1234,17 @@ void sendReplyToClient(aeEventLoop *el, int fd, void *privdata, int mask) {
  * we can just write the replies to the client output buffer without any
  * need to use a syscall in order to install the writable event handler,
  * get it called, and so forth. */
-int handleClientsWithPendingWrites(void) {
+int handleClientsWithPendingWrites(int iel) {
     listIter li;
     listNode *ln;
-    int processed = listLength(server.clients_pending_write);
+    list *pending_writes = server.rgclients_pending_write[iel];
+    int processed = listLength(pending_writes);
 
-    listRewind(server.clients_pending_write,&li);
+    listRewind(pending_writes,&li);
     while((ln = listNext(&li))) {
         client *c = listNodeValue(ln);
         c->flags &= ~CLIENT_PENDING_WRITE;
-        listDelNode(server.clients_pending_write,ln);
+        listDelNode(pending_writes,ln);
 
         /* If a client is protected, don't do anything,
          * that may trigger write error or recreate handler. */
@@ -1219,7 +1267,7 @@ int handleClientsWithPendingWrites(void) {
             {
                 ae_flags |= AE_BARRIER;
             }
-            if (aeCreateFileEvent(server.el, c->fd, ae_flags,
+            if (aeCreateFileEvent(server.rgel[c->iel], c->fd, ae_flags,
                 sendReplyToClient, c) == AE_ERR)
             {
                     freeClientAsync(c);
@@ -1268,15 +1316,15 @@ void resetClient(client *c) {
  *    path, it is not really released, but only marked for later release. */
 void protectClient(client *c) {
     c->flags |= CLIENT_PROTECTED;
-    aeDeleteFileEvent(server.el,c->fd,AE_READABLE);
+    aeDeleteFileEvent(server.rgel[c->iel],c->fd,AE_READABLE);
-    aeDeleteFileEvent(server.el,c->fd,AE_WRITABLE);
+    aeDeleteFileEvent(server.rgel[c->iel],c->fd,AE_WRITABLE);
 }
 
 /* This will undo the client protection done by protectClient() */
 void unprotectClient(client *c) {
     if (c->flags & CLIENT_PROTECTED) {
         c->flags &= ~CLIENT_PROTECTED;
-        aeCreateFileEvent(server.el,c->fd,AE_READABLE,readQueryFromClient,c);
+        aeCreateFileEvent(server.rgel[c->iel],c->fd,AE_READABLE|AE_THREADSAFE,readQueryFromClient,c);
         if (clientHasPendingReplies(c)) clientInstallWriteHandler(c);
     }
 }
@@ -1630,6 +1678,14 @@ void readQueryFromClient(aeEventLoop *el, int fd, void *privdata, int mask) {
     UNUSED(el);
     UNUSED(mask);
 
+    int iel = 0;
+    for (; iel < MAX_EVENT_LOOPS; ++iel)
+    {
+        if (server.rgel[iel] == el)
+            break;
+    }
+    serverAssert(iel == c->iel);
+
     readlen = PROTO_IOBUF_LEN;
     /* If this is a multi bulk request, and we are processing a bulk reply
      * that is large enough, try to maximize the probability that the query
@@ -1650,18 +1706,24 @@ void readQueryFromClient(aeEventLoop *el, int fd, void *privdata, int mask) {
     qblen = sdslen(c->querybuf);
     if (c->querybuf_peak < qblen) c->querybuf_peak = qblen;
     c->querybuf = sdsMakeRoomFor(c->querybuf, readlen);
+    
     nread = read(fd, c->querybuf+qblen, readlen);
+    
     if (nread == -1) {
         if (errno == EAGAIN) {
             return;
         } else {
             serverLog(LL_VERBOSE, "Reading from client: %s",strerror(errno));
+            aeAcquireLock();
             freeClient(c);
+            aeReleaseLock();
             return;
         }
     } else if (nread == 0) {
         serverLog(LL_VERBOSE, "Client closed connection");
+        aeAcquireLock();
         freeClient(c);
+        aeReleaseLock();
         return;
     } else if (c->flags & CLIENT_MASTER) {
         /* Append the query buffer to the pending (not applied) buffer
@@ -1682,7 +1744,9 @@ void readQueryFromClient(aeEventLoop *el, int fd, void *privdata, int mask) {
         serverLog(LL_WARNING,"Closing client that reached max query buffer length: %s (qbuf initial bytes: %s)", ci, bytes);
         sdsfree(ci);
         sdsfree(bytes);
+        aeAcquireLock();
         freeClient(c);
+        aeReleaseLock();
         return;
     }
 
@@ -1692,7 +1756,9 @@ void readQueryFromClient(aeEventLoop *el, int fd, void *privdata, int mask) {
      * was actually applied to the master state: this quantity, and its
      * corresponding part of the replication stream, will be propagated to
      * the sub-slaves and to the replication backlog. */
+    aeAcquireLock();
     processInputBufferAndReplicate(c);
+    aeReleaseLock();
 }
 
 void getClientsMaxBuffers(unsigned long *longest_output_list,
@@ -1775,7 +1841,7 @@ sds catClientInfoString(sds s, client *client) {
     if (p == flags) *p++ = 'N';
     *p++ = '\0';
 
-    emask = client->fd == -1 ? 0 : aeGetFileEvents(server.el,client->fd);
+    emask = client->fd == -1 ? 0 : aeGetFileEvents(server.rgel[client->iel],client->fd);
     p = events;
     if (emask & AE_READABLE) *p++ = 'r';
     if (emask & AE_WRITABLE) *p++ = 'w';
@@ -2323,7 +2389,7 @@ void flushSlavesOutputBuffers(void) {
          * of put_online_on_ack is to postpone the moment it is installed.
          * This is what we want since slaves in this state should not receive
          * writes before the first ACK. */
-        events = aeGetFileEvents(server.el,slave->fd);
+        events = aeGetFileEvents(server.rgel[IDX_EVENT_LOOP_MAIN],slave->fd);
         if (events & AE_WRITABLE &&
             slave->replstate == SLAVE_STATE_ONLINE &&
             clientHasPendingReplies(slave))
@@ -2395,13 +2461,13 @@ int clientsArePaused(void) {
  * write, close sequence needed to serve a client.
  *
  * The function returns the total number of events processed. */
-int processEventsWhileBlocked(void) {
+int processEventsWhileBlocked(int iel) {
     int iterations = 4; /* See the function top-comment. */
     int count = 0;
     while (iterations--) {
         int events = 0;
-        events += aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
+        events += aeProcessEvents(server.rgel[iel], AE_FILE_EVENTS|AE_DONT_WAIT);
-        events += handleClientsWithPendingWrites();
+        events += handleClientsWithPendingWrites(iel);
         if (!events) break;
         count += events;
     }

src/rdb.c

@@ -1862,7 +1862,7 @@ void rdbLoadProgressCallback(rio *r, const void *buf, size_t len) {
         if (server.masterhost && server.repl_state == REPL_STATE_TRANSFER)
             replicationSendNewlineToMaster();
         loadingProgress(r->processed_bytes);
-        processEventsWhileBlocked();
+        processEventsWhileBlocked(IDX_EVENT_LOOP_MAIN);
     }
 }
 

src/replication.c

@@ -856,7 +856,7 @@ void putSlaveOnline(client *slave) {
     slave->replstate = SLAVE_STATE_ONLINE;
     slave->repl_put_online_on_ack = 0;
     slave->repl_ack_time = server.unixtime; /* Prevent false timeout. */
-    if (aeCreateFileEvent(server.el, slave->fd, AE_WRITABLE,
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], slave->fd, AE_WRITABLE,
         sendReplyToClient, slave) == AE_ERR) {
         serverLog(LL_WARNING,"Unable to register writable event for replica bulk transfer: %s", strerror(errno));
         freeClient(slave);
@@ -918,7 +918,7 @@ void sendBulkToSlave(aeEventLoop *el, int fd, void *privdata, int mask) {
     if (slave->repldboff == slave->repldbsize) {
         close(slave->repldbfd);
         slave->repldbfd = -1;
-        aeDeleteFileEvent(server.el,slave->fd,AE_WRITABLE);
+        aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],slave->fd,AE_WRITABLE);
         putSlaveOnline(slave);
     }
 }
@@ -989,8 +989,8 @@ void updateSlavesWaitingBgsave(int bgsaveerr, int type) {
                 slave->replpreamble = sdscatprintf(sdsempty(),"$%lld\r\n",
                     (unsigned long long) slave->repldbsize);
 
-                aeDeleteFileEvent(server.el,slave->fd,AE_WRITABLE);
+                aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],slave->fd,AE_WRITABLE);
-                if (aeCreateFileEvent(server.el, slave->fd, AE_WRITABLE, sendBulkToSlave, slave) == AE_ERR) {
+                if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], slave->fd, AE_WRITABLE, sendBulkToSlave, slave) == AE_ERR) {
                     freeClient(slave);
                     continue;
                 }
@@ -1075,7 +1075,7 @@ void replicationEmptyDbCallback(void *privdata) {
  * performed, this function materializes the master client we store
  * at server.master, starting from the specified file descriptor. */
 void replicationCreateMasterClient(int fd, int dbid) {
-    server.master = createClient(fd);
+    server.master = createClient(fd, IDX_EVENT_LOOP_MAIN);
     server.master->flags |= CLIENT_MASTER;
     server.master->authenticated = 1;
     server.master->reploff = server.master_initial_offset;
@@ -1276,7 +1276,7 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
          * handler, otherwise it will get called recursively since
          * rdbLoad() will call the event loop to process events from time to
          * time for non blocking loading. */
-        aeDeleteFileEvent(server.el,server.repl_transfer_s,AE_READABLE);
+        aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.repl_transfer_s,AE_READABLE);
         serverLog(LL_NOTICE, "MASTER <-> REPLICA sync: Loading DB in memory");
         rdbSaveInfo rsi = RDB_SAVE_INFO_INIT;
         if (rdbLoad(server.rdb_filename,&rsi) != C_OK) {
@@ -1464,7 +1464,7 @@ int slaveTryPartialResynchronization(int fd, int read_reply) {
         if (reply != NULL) {
             serverLog(LL_WARNING,"Unable to send PSYNC to master: %s",reply);
             sdsfree(reply);
-            aeDeleteFileEvent(server.el,fd,AE_READABLE);
+            aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_READABLE);
             return PSYNC_WRITE_ERROR;
         }
         return PSYNC_WAIT_REPLY;
@@ -1479,7 +1479,7 @@ int slaveTryPartialResynchronization(int fd, int read_reply) {
         return PSYNC_WAIT_REPLY;
     }
 
-    aeDeleteFileEvent(server.el,fd,AE_READABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_READABLE);
 
     if (!strncmp(reply,"+FULLRESYNC",11)) {
         char *replid = NULL, *offset = NULL;
@@ -1626,7 +1626,7 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
         serverLog(LL_NOTICE,"Non blocking connect for SYNC fired the event.");
         /* Delete the writable event so that the readable event remains
          * registered and we can wait for the PONG reply. */
-        aeDeleteFileEvent(server.el,fd,AE_WRITABLE);
+        aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_WRITABLE);
         server.repl_state = REPL_STATE_RECEIVE_PONG;
         /* Send the PING, don't check for errors at all, we have the timeout
          * that will take care about this. */
@@ -1841,7 +1841,7 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
     }
 
     /* Setup the non blocking download of the bulk file. */
-    if (aeCreateFileEvent(server.el,fd, AE_READABLE,readSyncBulkPayload,NULL)
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd, AE_READABLE,readSyncBulkPayload,NULL)
             == AE_ERR)
     {
         serverLog(LL_WARNING,
@@ -1860,7 +1860,7 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
     return;
 
 error:
-    aeDeleteFileEvent(server.el,fd,AE_READABLE|AE_WRITABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_READABLE|AE_WRITABLE);
     if (dfd != -1) close(dfd);
     close(fd);
     server.repl_transfer_s = -1;
@@ -1884,7 +1884,7 @@ int connectWithMaster(void) {
         return C_ERR;
     }
 
-    if (aeCreateFileEvent(server.el,fd,AE_READABLE|AE_WRITABLE,syncWithMaster,NULL) ==
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_READABLE|AE_WRITABLE,syncWithMaster,NULL) ==
             AE_ERR)
     {
         close(fd);
@@ -1905,7 +1905,7 @@ int connectWithMaster(void) {
 void undoConnectWithMaster(void) {
     int fd = server.repl_transfer_s;
 
-    aeDeleteFileEvent(server.el,fd,AE_READABLE|AE_WRITABLE);
+    aeDeleteFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],fd,AE_READABLE|AE_WRITABLE);
     close(fd);
     server.repl_transfer_s = -1;
 }
@@ -2236,7 +2236,7 @@ void replicationResurrectCachedMaster(int newfd) {
 
     /* Re-add to the list of clients. */
     linkClient(server.master);
-    if (aeCreateFileEvent(server.el, newfd, AE_READABLE,
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], newfd, AE_READABLE,
                           readQueryFromClient, server.master)) {
         serverLog(LL_WARNING,"Error resurrecting the cached master, impossible to add the readable handler: %s", strerror(errno));
         freeClientAsync(server.master); /* Close ASAP. */
@@ -2245,7 +2245,7 @@ void replicationResurrectCachedMaster(int newfd) {
     /* We may also need to install the write handler as well if there is
      * pending data in the write buffers. */
     if (clientHasPendingReplies(server.master)) {
-        if (aeCreateFileEvent(server.el, newfd, AE_WRITABLE,
+        if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], newfd, AE_WRITABLE,
                           sendReplyToClient, server.master)) {
             serverLog(LL_WARNING,"Error resurrecting the cached master, impossible to add the writable handler: %s", strerror(errno));
             freeClientAsync(server.master); /* Close ASAP. */

src/scripting.c

@@ -1104,7 +1104,7 @@ void scriptingInit(int setup) {
      * Note: there is no need to create it again when this function is called
      * by scriptingReset(). */
     if (server.lua_client == NULL) {
-        server.lua_client = createClient(-1);
+        server.lua_client = createClient(-1, IDX_EVENT_LOOP_MAIN);
         server.lua_client->flags |= CLIENT_LUA;
     }
 
@@ -1278,7 +1278,7 @@ void luaMaskCountHook(lua_State *lua, lua_Debug *ar) {
          * here when the EVAL command will return. */
         protectClient(server.lua_caller);
     }
-    if (server.lua_timedout) processEventsWhileBlocked();
+    if (server.lua_timedout) processEventsWhileBlocked(IDX_EVENT_LOOP_MAIN);
     if (server.lua_kill) {
         serverLog(LL_WARNING,"Lua script killed by user with SCRIPT KILL.");
         lua_pushstring(lua,"Script killed by user with SCRIPT KILL...");

src/sentinel.c

@@ -2013,7 +2013,7 @@ void sentinelReconnectInstance(sentinelRedisInstance *ri) {
             link->pending_commands = 0;
             link->cc_conn_time = mstime();
             link->cc->data = link;
-            redisAeAttach(server.el,link->cc);
+            redisAeAttach(server.rgel[IDX_EVENT_LOOP_MAIN],link->cc);
             redisAsyncSetConnectCallback(link->cc,
                     sentinelLinkEstablishedCallback);
             redisAsyncSetDisconnectCallback(link->cc,
@@ -2037,7 +2037,7 @@ void sentinelReconnectInstance(sentinelRedisInstance *ri) {
 
             link->pc_conn_time = mstime();
             link->pc->data = link;
-            redisAeAttach(server.el,link->pc);
+            redisAeAttach(server.rgel[IDX_EVENT_LOOP_MAIN],link->pc);
             redisAsyncSetConnectCallback(link->pc,
                     sentinelLinkEstablishedCallback);
             redisAsyncSetDisconnectCallback(link->pc,

src/server.c

@@ -1619,6 +1619,16 @@ void getExpansiveClientsInfo(size_t *in_usage, size_t *out_usage) {
     *out_usage = o;
 }
 
+
+void AsyncClientCron(void *pv)
+{
+    client *c = (client*)pv;
+    mstime_t now = mstime();
+    if (clientsCronHandleTimeout(c,now)) return;
+    if (clientsCronResizeQueryBuffer(c)) return;
+    if (clientsCronTrackExpansiveClients(c)) return;
+}
+
 /* This function is called by serverCron() and is used in order to perform
  * operations on clients that are important to perform constantly. For instance
  * we use this function in order to disconnect clients after a timeout, including
@@ -1661,12 +1671,20 @@ void clientsCron(void) {
         listRotate(server.clients);
         head = listFirst(server.clients);
         c = listNodeValue(head);
-        /* The following functions do different service checks on the client.
+        if (c->iel == IDX_EVENT_LOOP_MAIN)
-         * The protocol is that they return non-zero if the client was
+        {
-         * terminated. */
+            /* The following functions do different service checks on the client.
-        if (clientsCronHandleTimeout(c,now)) continue;
+            * The protocol is that they return non-zero if the client was
-        if (clientsCronResizeQueryBuffer(c)) continue;
+            * terminated. */
-        if (clientsCronTrackExpansiveClients(c)) continue;
+            if (clientsCronHandleTimeout(c,now)) continue;
+            if (clientsCronResizeQueryBuffer(c)) continue;
+            if (clientsCronTrackExpansiveClients(c)) continue;
+        }
+        else if (IDX_EVENT_LOOP_MAIN > 1)
+        {
+            aePostFunction(server.rgel[c->iel], AsyncClientCron, c);
+        }
+        
     }
 }
 
@@ -2070,14 +2088,20 @@ void beforeSleep(struct aeEventLoop *eventLoop) {
     moduleHandleBlockedClients();
 
     /* Try to process pending commands for clients that were just unblocked. */
-    if (listLength(server.unblocked_clients))
+    if (listLength(server.rgunblocked_clients[IDX_EVENT_LOOP_MAIN]))
-        processUnblockedClients();
+    {
+        aeReleaseLock();
+        processUnblockedClients(IDX_EVENT_LOOP_MAIN);
+        aeAcquireLock();
+    }
 
     /* Write the AOF buffer on disk */
     flushAppendOnlyFile(0);
 
     /* Handle writes with pending output buffers. */
-    handleClientsWithPendingWrites();
+    aeReleaseLock();
+    handleClientsWithPendingWrites(IDX_EVENT_LOOP_MAIN);
+    aeAcquireLock();
 
     /* Before we are going to sleep, let the threads access the dataset by
      * releasing the GIL. Redis main thread will not touch anything at this
@@ -2085,6 +2109,24 @@ void beforeSleep(struct aeEventLoop *eventLoop) {
     if (moduleCount()) moduleReleaseGIL();
 }
 
+void beforeSleepLite(struct aeEventLoop *eventLoop)
+{
+    int iel = 0;
+    for (; iel < MAX_EVENT_LOOPS; ++iel)
+    {
+        if (server.rgel[iel] == eventLoop)
+            break;
+    }
+    serverAssert(iel < MAX_EVENT_LOOPS);
+    
+    /* Try to process pending commands for clients that were just unblocked. */
+    if (listLength(server.rgunblocked_clients[iel]))
+        processUnblockedClients(iel);
+    
+    /* Handle writes with pending output buffers. */
+    handleClientsWithPendingWrites(iel);
+}
+
 /* This function is called immadiately after the event loop multiplexing
  * API returned, and the control is going to soon return to Redis by invoking
  * the different events callbacks. */
@@ -2702,6 +2744,8 @@ void initServer(void) {
     signal(SIGPIPE, SIG_IGN);
     setupSignalHandlers();
 
+    fastlock_init(&server.flock);
+
     if (server.syslog_enabled) {
         openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
             server.syslog_facility);
@@ -2715,9 +2759,12 @@ void initServer(void) {
     server.clients_to_close = listCreate();
     server.slaves = listCreate();
     server.monitors = listCreate();
-    server.clients_pending_write = listCreate();
+    for (int iel = 0; iel < MAX_EVENT_LOOPS; ++iel)
+    {
+        server.rgclients_pending_write[iel] = listCreate();
+        server.rgunblocked_clients[iel] = listCreate();
+    }
     server.slaveseldb = -1; /* Force to emit the first SELECT command. */
-    server.unblocked_clients = listCreate();
     server.ready_keys = listCreate();
     server.clients_waiting_acks = listCreate();
     server.get_ack_from_slaves = 0;
@@ -2726,12 +2773,15 @@ void initServer(void) {
 
     createSharedObjects();
     adjustOpenFilesLimit();
-    server.el = aeCreateEventLoop(server.maxclients+CONFIG_FDSET_INCR);
+    for (int i = 0; i < MAX_EVENT_LOOPS; ++i)
-    if (server.el == NULL) {
+    {
-        serverLog(LL_WARNING,
+        server.rgel[i] = aeCreateEventLoop(server.maxclients+CONFIG_FDSET_INCR);
-            "Failed creating the event loop. Error message: '%s'",
+        if (server.rgel[i] == NULL) {
-            strerror(errno));
+            serverLog(LL_WARNING,
-        exit(1);
+                "Failed creating the event loop. Error message: '%s'",
+                strerror(errno));
+            exit(1);
+        }
     }
     server.db = zmalloc(sizeof(redisDb)*server.dbnum, MALLOC_LOCAL);
 
@@ -2808,7 +2858,7 @@ void initServer(void) {
     /* Create the timer callback, this is our way to process many background
      * operations incrementally, like clients timeout, eviction of unaccessed
      * expired keys and so forth. */
-    if (aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL) == AE_ERR) {
+    if (aeCreateTimeEvent(server.rgel[IDX_EVENT_LOOP_MAIN], 1, serverCron, NULL, NULL) == AE_ERR) {
         serverPanic("Can't create event loop timers.");
         exit(1);
     }
@@ -2816,20 +2866,23 @@ void initServer(void) {
     /* Create an event handler for accepting new connections in TCP and Unix
      * domain sockets. */
     for (j = 0; j < server.ipfd_count; j++) {
-        if (aeCreateFileEvent(server.el, server.ipfd[j], AE_READABLE,
+        for (int iel = 0; iel < MAX_EVENT_LOOPS; ++iel)
-            acceptTcpHandler,NULL) == AE_ERR)
+        {
-            {
+            if (aeCreateFileEvent(server.rgel[iel], server.ipfd[j], AE_READABLE|AE_THREADSAFE,
-                serverPanic(
+                acceptTcpHandler,NULL) == AE_ERR)
-                    "Unrecoverable error creating server.ipfd file event.");
+                {
-            }
+                    serverPanic(
+                        "Unrecoverable error creating server.ipfd file event.");
+                }
+        }
     }
-    if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
+    if (server.sofd > 0 && aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN],server.sofd,AE_READABLE,
         acceptUnixHandler,NULL) == AE_ERR) serverPanic("Unrecoverable error creating server.sofd file event.");
 
 
     /* Register a readable event for the pipe used to awake the event loop
      * when a blocked client in a module needs attention. */
-    if (aeCreateFileEvent(server.el, server.module_blocked_pipe[0], AE_READABLE,
+    if (aeCreateFileEvent(server.rgel[IDX_EVENT_LOOP_MAIN], server.module_blocked_pipe[0], AE_READABLE,
         moduleBlockedClientPipeReadable,NULL) == AE_ERR) {
             serverPanic(
                 "Error registering the readable event for the module "
@@ -4738,6 +4791,17 @@ int redisIsSupervised(int mode) {
     return 0;
 }
 
+void *workerThreadMain(void *parg)
+{
+    int iel = (int)((int64_t)parg);
+
+    int isMainThread = (iel == IDX_EVENT_LOOP_MAIN);
+    aeSetBeforeSleepProc(server.rgel[iel], isMainThread ? beforeSleep : beforeSleepLite, isMainThread ? 0 : AE_THREADSAFE);
+    aeSetAfterSleepProc(server.rgel[iel], isMainThread ? afterSleep : NULL, 0);
+    aeMain(server.rgel[iel]);
+    aeDeleteEventLoop(server.rgel[iel]);
+    return NULL;
+}
 
 int main(int argc, char **argv) {
     struct timeval tv;
@@ -4940,10 +5004,13 @@ int main(int argc, char **argv) {
         serverLog(LL_WARNING,"WARNING: You specified a maxmemory value that is less than 1MB (current value is %llu bytes). Are you sure this is what you really want?", server.maxmemory);
     }
 
-    aeSetBeforeSleepProc(server.el,beforeSleep);
+    pthread_t rgthread[MAX_EVENT_LOOPS];
-    aeSetAfterSleepProc(server.el,afterSleep);
+    for (int iel = 0; iel < MAX_EVENT_LOOPS; ++iel)
-    aeMain(server.el);
+    {
-    aeDeleteEventLoop(server.el);
+        pthread_create(rgthread + iel, NULL, workerThreadMain, (void*)((int64_t)iel));
+    }
+    void *pretT;
+    pthread_join(rgthread[IDX_EVENT_LOOP_MAIN], &pretT);
     return 0;
 }
 

src/server.h

@@ -51,6 +51,7 @@
 
 typedef long long mstime_t; /* millisecond time type. */
 
+#include "fastlock.h"
 #include "ae.h"      /* Event driven programming library */
 #include "sds.h"     /* Dynamic safe strings */
 #include "dict.h"    /* Hash tables */
@@ -845,6 +846,8 @@ typedef struct client {
     /* Response buffer */
     int bufpos;
     char buf[PROTO_REPLY_CHUNK_BYTES];
+
+    int iel; /* the event loop index we're registered with */
 } client;
 
 struct saveparam {
@@ -1005,6 +1008,9 @@ struct clusterState;
 #define CHILD_INFO_TYPE_RDB 0
 #define CHILD_INFO_TYPE_AOF 1
 
+#define MAX_EVENT_LOOPS 2
+#define IDX_EVENT_LOOP_MAIN 0
+
 struct redisServer {
     /* General */
     pid_t pid;                  /* Main process pid. */
@@ -1019,7 +1025,8 @@ struct redisServer {
     redisDb *db;
     dict *commands;             /* Command table */
     dict *orig_commands;        /* Command table before command renaming. */
-    aeEventLoop *el;
+    int cel;
+    aeEventLoop *rgel[MAX_EVENT_LOOPS];
     unsigned int lruclock;      /* Clock for LRU eviction */
     int shutdown_asap;          /* SHUTDOWN needed ASAP */
     int activerehashing;        /* Incremental rehash in serverCron() */
@@ -1051,7 +1058,7 @@ struct redisServer {
     int cfd_count;              /* Used slots in cfd[] */
     list *clients;              /* List of active clients */
     list *clients_to_close;     /* Clients to close asynchronously */
-    list *clients_pending_write; /* There is to write or install handler. */
+    list *rgclients_pending_write[MAX_EVENT_LOOPS]; /* There is to write or install handler. */
     list *slaves, *monitors;    /* List of slaves and MONITORs */
     client *current_client; /* Current client, only used on crash report */
     rax *clients_index;         /* Active clients dictionary by client ID. */
@@ -1272,7 +1279,7 @@ struct redisServer {
     /* Blocked clients */
     unsigned int blocked_clients;   /* # of clients executing a blocking cmd.*/
     unsigned int blocked_clients_by_type[BLOCKED_NUM];
-    list *unblocked_clients; /* list of clients to unblock before next loop */
+    list *rgunblocked_clients[MAX_EVENT_LOOPS]; /* list of clients to unblock before next loop */
     list *ready_keys;        /* List of readyList structures for BLPOP & co */
     /* Sort parameters - qsort_r() is only available under BSD so we
      * have to take this state global, in order to pass it to sortCompare() */
@@ -1362,6 +1369,8 @@ struct redisServer {
     pthread_mutex_t lruclock_mutex;
     pthread_mutex_t next_client_id_mutex;
     pthread_mutex_t unixtime_mutex;
+
+    struct fastlock flock;
 };
 
 typedef struct pubsubPattern {
@@ -1504,7 +1513,7 @@ size_t redisPopcount(void *s, long count);
 void redisSetProcTitle(char *title);
 
 /* networking.c -- Networking and Client related operations */
-client *createClient(int fd);
+client *createClient(int fd, int iel);
 void closeTimedoutClients(void);
 void freeClient(client *c);
 void freeClientAsync(client *c);
@@ -1571,8 +1580,8 @@ void disconnectSlaves(void);
 int listenToPort(int port, int *fds, int *count);
 void pauseClients(mstime_t duration);
 int clientsArePaused(void);
-int processEventsWhileBlocked(void);
+int processEventsWhileBlocked(int iel);
-int handleClientsWithPendingWrites(void);
+int handleClientsWithPendingWrites(int iel);
 int clientHasPendingReplies(client *c);
 void unlinkClient(client *c);
 int writeToClient(int fd, client *c, int handler_installed);
@@ -2014,7 +2023,7 @@ int ldbPendingChildren(void);
 sds luaCreateFunction(client *c, lua_State *lua, robj *body);
 
 /* Blocked clients */
-void processUnblockedClients(void);
+void processUnblockedClients(int iel);
 void blockClient(client *c, int btype);
 void unblockClient(client *c);
 void queueClientForReprocessing(client *c);