Merge changes from topic 'spec-2089' 6.99.4 guppy/6.99.4 guppy_6.99.4
authorJan-Simon Moeller <jsmoeller@linuxfoundation.org>
Fri, 18 Jan 2019 18:28:56 +0000 (18:28 +0000)
committerGerrit Code Review <gerrit@automotivelinux.org>
Fri, 18 Jan 2019 18:28:56 +0000 (18:28 +0000)
* changes:
  hello3: Fix warning unused variable
  afb-hsrv: Avoid call to getnameinfo
  jobs: Remove conditionnal REMOVE_SYSTEMD_EVENT
  jobs: Separate internal threads from others
  jobs: Refactor event loop handling
  afb-proto-ws: Serialize incoming message

bindings/samples/hello3.c
src/afb-fdev.c
src/afb-hsrv.c
src/afb-proto-ws.c
src/afb-proto-ws.h
src/afb-stub-ws.c
src/jobs.c
src/jobs.h
src/sig-monitor.c
src/sig-monitor.h

index 131e0e4..1d7c195 100644 (file)
@@ -627,7 +627,6 @@ static void in_after (afb_req_t request)
        char *te;
        double td;
        struct timespec t;
-       void (*calling)(afb_req_t);
 
        /* get the type */
        ty = afb_req_value(request, "type") ?: "call";
index 8e3358d..386e1e4 100644 (file)
@@ -17,8 +17,6 @@
 
 #include "fdev.h"
 
-#if !defined(REMOVE_SYSTEMD_EVENT)
-
 #include "afb-systemd.h"
 #include "fdev-systemd.h"
 
@@ -26,16 +24,3 @@ struct fdev *afb_fdev_create(int fd)
 {
        return fdev_systemd_create(afb_systemd_get_event_loop(), fd);
 }
-
-#else
-
-#include "jobs.h"
-#include "fdev-epoll.h"
-
-struct fdev *afb_fdev_create(int fd)
-{
-       return fdev_epoll_add(jobs_get_fdev_epoll(), fd);
-}
-
-#endif
-
index 2a08a2e..0d0bd06 100644 (file)
@@ -540,13 +540,19 @@ static int hsrv_itf_connect(struct hsrv_itf *itf)
        }
        fdev_set_events(itf->fdev, EPOLLIN);
        fdev_set_callback(itf->fdev, hsrv_itf_callback, itf);
+       memset(&addr, 0, sizeof addr);
        lenaddr = (socklen_t)sizeof addr;
        getsockname(fdev_fd(itf->fdev), &addr, &lenaddr);
-       rgni = getnameinfo(&addr, lenaddr, hbuf, sizeof hbuf, sbuf, sizeof sbuf, NI_NUMERICSERV);
-       if (rgni != 0) {
-               ERROR("getnameinfo returned %d: %s", rgni, gai_strerror(rgni));
-               hbuf[0] = sbuf[0] = '?';
-               hbuf[1] = sbuf[1] = 0;
+       if (addr.sa_family == AF_INET && !((struct sockaddr_in*)&addr)->sin_addr.s_addr) {
+               strncpy(hbuf, "*", NI_MAXHOST);
+               sprintf(sbuf, "%d", (int)ntohs(((struct sockaddr_in*)&addr)->sin_port));
+       } else {
+               rgni = getnameinfo(&addr, lenaddr, hbuf, sizeof hbuf, sbuf, sizeof sbuf, NI_NUMERICSERV);
+               if (rgni != 0) {
+                       ERROR("getnameinfo returned %d: %s", rgni, gai_strerror(rgni));
+                       hbuf[0] = sbuf[0] = '?';
+                       hbuf[1] = sbuf[1] = 0;
+               }
        }
        NOTICE("Listening interface %s:%s", hbuf, sbuf);
        return 1;
@@ -583,4 +589,4 @@ int afb_hsrv_add_interface_tcp(struct afb_hsrv *hsrv, const char *itf, uint16_t
                return 0;
        }
        return afb_hsrv_add_interface(hsrv, buffer);
-}
\ No newline at end of file
+}
index f74869c..f407086 100644 (file)
@@ -35,7 +35,6 @@
 #include "afb-ws.h"
 #include "afb-msg-json.h"
 #include "afb-proto-ws.h"
-#include "jobs.h"
 #include "fdev.h"
 #include "verbose.h"
 
@@ -161,7 +160,7 @@ struct afb_proto_ws
        void (*on_hangup)(void *closure);
 
        /* queuing facility for processing messages */
-       int (*queuing)(void (*process)(int s, void *c), void *closure);
+       int (*queuing)(struct afb_proto_ws *proto, void (*process)(int s, void *c), void *closure);
 };
 
 /******************* streaming objects **********************************/
@@ -348,7 +347,7 @@ static void queue_message_processing(struct afb_proto_ws *protows, char *data, s
                        binary->rb.head = data;
                        binary->rb.end = data + size;
                        if (!protows->queuing
-                        || protows->queuing(processing, binary) < 0)
+                        || protows->queuing(protows, processing, binary) < 0)
                                processing(0, binary);
                        return;
                }
@@ -1068,7 +1067,7 @@ void afb_proto_ws_on_hangup(struct afb_proto_ws *protows, void (*on_hangup)(void
        protows->on_hangup = on_hangup;
 }
 
-void afb_proto_ws_set_queuing(struct afb_proto_ws *protows, int (*queuing)(void (*)(int,void*), void*))
+void afb_proto_ws_set_queuing(struct afb_proto_ws *protows, int (*queuing)(struct afb_proto_ws*, void (*)(int,void*), void*))
 {
        protows->queuing = queuing;
 }
index b5f84e9..d797cb6 100644 (file)
@@ -61,7 +61,7 @@ extern int afb_proto_ws_is_server(struct afb_proto_ws *protows);
 extern void afb_proto_ws_hangup(struct afb_proto_ws *protows);
 
 extern void afb_proto_ws_on_hangup(struct afb_proto_ws *protows, void (*on_hangup)(void *closure));
-extern void afb_proto_ws_set_queuing(struct afb_proto_ws *protows, int (*queuing)(void (*)(int,void*), void*));
+extern void afb_proto_ws_set_queuing(struct afb_proto_ws *protows, int (*queuing)(struct afb_proto_ws *, void (*)(int,void*), void*));
 
 
 extern int afb_proto_ws_client_call(struct afb_proto_ws *protows, const char *verb, struct json_object *args, const char *sessionid, void *request, const char *user_creds);
index 5dfacb7..7ec2d63 100644 (file)
@@ -646,9 +646,9 @@ static void on_hangup(void *closure)
        }
 }
 
-static int enqueue_processing(void (*callback)(int signum, void* arg), void *arg)
+static int enqueue_processing(struct afb_proto_ws *proto, void (*callback)(int signum, void* arg), void *arg)
 {
-       return jobs_queue(NULL, 0, callback, arg);
+       return jobs_queue(proto, 0, callback, arg);
 }
 
 /*****************************************************/
index f2c9d52..4a88c30 100644 (file)
 #include "sig-monitor.h"
 #include "verbose.h"
 
-#if defined(REMOVE_SYSTEMD_EVENT)
-#include "fdev-epoll.h"
-#endif
-
 #define EVENT_TIMEOUT_TOP      ((uint64_t)-1)
 #define EVENT_TIMEOUT_CHILD    ((uint64_t)10000)
 
@@ -75,20 +71,20 @@ struct evloop
        unsigned state;        /**< encoded state */
        int efd;               /**< event notification */
        struct sd_event *sdev; /**< the systemd event loop */
-       pthread_cond_t  cond;  /**< condition */
        struct fdev *fdev;     /**< handling of events */
        struct thread *holder; /**< holder of the evloop */
 };
 
 #define EVLOOP_STATE_WAIT           1U
 #define EVLOOP_STATE_RUN            2U
-#define EVLOOP_STATE_LOCK           4U
 
 /** Description of threads */
 struct thread
 {
        struct thread *next;   /**< next thread of the list */
        struct thread *upper;  /**< upper same thread */
+       struct thread *nholder;/**< next holder for evloop */
+       pthread_cond_t *cwhold;/**< condition wait for holding */
        struct job *job;       /**< currently processed job */
        pthread_t tid;         /**< the thread id */
        volatile unsigned stop: 1;      /**< stop requested */
@@ -96,7 +92,7 @@ struct thread
 };
 
 /**
- * Description of synchonous callback
+ * Description of synchronous callback
  */
 struct sync
 {
@@ -123,19 +119,13 @@ static int remains = 0; /** allowed count of waiting jobs */
 /* list of threads */
 static struct thread *threads;
 static _Thread_local struct thread *current_thread;
-static _Thread_local struct evloop *current_evloop;
 
 /* queue of pending jobs */
 static struct job *first_job;
 static struct job *free_jobs;
 
 /* event loop */
-static struct evloop evloop[1];
-
-#if defined(REMOVE_SYSTEMD_EVENT)
-static struct fdev_epoll *fdevepoll;
-static int waitevt;
-#endif
+static struct evloop evloop;
 
 /**
  * Create a new job with the given parameters
@@ -163,6 +153,7 @@ static struct job *job_create(
                job = malloc(sizeof *job);
                pthread_mutex_lock(&mutex);
                if (!job) {
+                       ERROR("out of memory");
                        errno = ENOMEM;
                        goto end;
                }
@@ -204,6 +195,7 @@ static void job_add(struct job *job)
 
        /* queue the jobs */
        *pjob = job;
+       remains--;
 }
 
 /**
@@ -215,6 +207,8 @@ static inline struct job *job_get()
        struct job *job = first_job;
        while (job && job->blocked)
                job = job->next;
+       if (job)
+               remains++;
        return job;
 }
 
@@ -269,23 +263,6 @@ static void job_cancel(int signum, void *arg)
        job->callback(SIGABRT, job->arg);
 }
 
-#if defined(REMOVE_SYSTEMD_EVENT)
-/**
- * Gets a fdev_epoll item.
- * @return a fdev_epoll or NULL in case of error
- */
-static struct fdev_epoll *get_fdevepoll()
-{
-       struct fdev_epoll *result;
-
-       result = fdevepoll;
-       if (!result)
-               result = fdevepoll = fdev_epoll_create();
-
-       return result;
-}
-#endif
-
 /**
  * Monitored normal callback for events.
  * This function is called by the monitor
@@ -300,13 +277,9 @@ static void evloop_run(int signum, void *arg)
 {
        int rc;
        struct sd_event *se;
-       struct evloop *el = arg;
 
        if (!signum) {
-               current_evloop = el;
-               __atomic_store_n(&el->state, EVLOOP_STATE_LOCK|EVLOOP_STATE_RUN|EVLOOP_STATE_WAIT, __ATOMIC_RELAXED);
-               __atomic_store_n(&el->holder, current_thread, __ATOMIC_RELAXED);
-               se = el->sdev;
+               se = evloop.sdev;
                rc = sd_event_prepare(se);
                if (rc < 0) {
                        errno = -rc;
@@ -320,8 +293,7 @@ static void evloop_run(int signum, void *arg)
                                        ERROR("sd_event_wait returned an error (state: %d): %m", sd_event_get_state(se));
                                }
                        }
-                       __atomic_and_fetch(&el->state, ~(EVLOOP_STATE_WAIT), __ATOMIC_RELAXED);
-
+                       evloop.state = EVLOOP_STATE_RUN;
                        if (rc > 0) {
                                rc = sd_event_dispatch(se);
                                if (rc < 0) {
@@ -331,69 +303,151 @@ static void evloop_run(int signum, void *arg)
                        }
                }
        }
-       __atomic_and_fetch(&el->state, ~(EVLOOP_STATE_WAIT|EVLOOP_STATE_RUN), __ATOMIC_RELAXED);
 }
 
+/**
+ * Internal callback for evloop management.
+ * The effect of this function is hidden: it exits
+ * the waiting poll if any.
+ */
+static void evloop_on_efd_event()
+{
+       uint64_t x;
+       read(evloop.efd, &x, sizeof x);
+}
 
-#if defined(REMOVE_SYSTEMD_EVENT)
 /**
- * Monitored normal loop for waiting events.
- * @param signum 0 on normal flow or the number
- *               of the signal that interrupted the normal
- *               flow
- * @param arg     the events to run
+ * wakeup the event loop if needed by sending
+ * an event.
  */
-static void monitored_wait_and_dispatch(int signum, void *arg)
+static void evloop_wakeup()
 {
-       struct fdev_epoll *fdev_epoll = arg;
-       if (!signum) {
-               fdev_epoll_wait_and_dispatch(fdev_epoll, -1);
+       uint64_t x;
+
+       if (evloop.state & EVLOOP_STATE_WAIT) {
+               x = 1;
+               write(evloop.efd, &x, sizeof x);
        }
 }
-#endif
 
 /**
- * Main processing loop of threads processing jobs.
- * The loop must be called with the mutex locked
- * and it returns with the mutex locked.
- * @param me the description of the thread to use
- * TODO: how are timeout handled when reentering?
+ * Release the currently held event loop
  */
-static void thread_run(volatile struct thread *me)
+static void evloop_release()
 {
-       struct thread **prv;
-       struct job *job;
-#if !defined(REMOVE_SYSTEMD_EVENT)
-       struct evloop *el;
-#endif
+       struct thread *nh, *ct = current_thread;
+
+       if (evloop.holder == ct) {
+               nh = ct->nholder;
+               evloop.holder = nh;
+               if (nh)
+                       pthread_cond_signal(nh->cwhold);
+       }
+}
+
+/**
+ * get the eventloop for the current thread
+ */
+static int evloop_get()
+{
+       struct thread *ct = current_thread;
+
+       if (evloop.holder)
+               return evloop.holder == ct;
+
+       ct->nholder = NULL;
+       evloop.holder = ct;
+       return 1;
+}
+
+/**
+ * acquire the eventloop for the current thread
+ */
+static void evloop_acquire()
+{
+       struct thread **pwait, *ct;
+       pthread_cond_t cond;
+
+       /* try to get the evloop */
+       if (!evloop_get()) {
+               /* failed, init waiting state */
+               ct = current_thread;
+               ct->nholder = NULL;
+               ct->cwhold = &cond;
+               pthread_cond_init(&cond, NULL);
+
+               /* queue current thread in holder list */
+               pwait = &evloop.holder;
+               while (*pwait)
+                       pwait = &(*pwait)->nholder;
+               *pwait = ct;
+
+               /* wake up the evloop */
+               evloop_wakeup();
+
+               /* wait to acquire the evloop */
+               pthread_cond_wait(&cond, &mutex);
+               pthread_cond_destroy(&cond);
+       }
+}
 
+/**
+ * Enter the thread
+ * @param me the description of the thread to enter
+ */
+static void thread_enter(volatile struct thread *me)
+{
        /* initialize description of itself and link it in the list */
        me->tid = pthread_self();
        me->stop = 0;
        me->waits = 0;
        me->upper = current_thread;
-       if (!current_thread) {
-               started++;
-               sig_monitor_init_timeouts();
-       }
        me->next = threads;
        threads = (struct thread*)me;
        current_thread = (struct thread*)me;
+}
+
+/**
+ * leave the thread
+ * @param me the description of the thread to leave
+ */
+static void thread_leave()
+{
+       struct thread **prv, *me;
+
+       /* unlink the current thread and cleanup */
+       me = current_thread;
+       prv = &threads;
+       while (*prv != me)
+               prv = &(*prv)->next;
+       *prv = me->next;
+
+       current_thread = me->upper;
+}
+
+/**
+ * Main processing loop of internal threads with processing jobs.
+ * The loop must be called with the mutex locked
+ * and it returns with the mutex locked.
+ * @param me the description of the thread to use
+ * TODO: how are timeout handled when reentering?
+ */
+static void thread_run_internal(volatile struct thread *me)
+{
+       struct job *job;
+
+       /* enter thread */
+       thread_enter(me);
 
        /* loop until stopped */
        while (!me->stop) {
-               /* release the event loop */
-               if (current_evloop) {
-                       __atomic_and_fetch(&current_evloop->state, ~EVLOOP_STATE_LOCK, __ATOMIC_RELAXED);
-                       __atomic_store_n(&current_evloop->holder, NULL, __ATOMIC_RELAXED);
-                       current_evloop = NULL;
-               }
+               /* release the current event loop */
+               evloop_release();
 
                /* get a job */
                job = job_get();
                if (job) {
                        /* prepare running the job */
-                       remains++; /* increases count of job that can wait */
                        job->blocked = 1; /* mark job as blocked */
                        me->job = job; /* record the job (only for terminate) */
 
@@ -404,31 +458,21 @@ static void thread_run(volatile struct thread *me)
 
                        /* release the run job */
                        job_release(job);
-#if !defined(REMOVE_SYSTEMD_EVENT)
-               } else {
-                       /* no job, check events */
-                       el = &evloop[0];
-                       if (el->sdev && !__atomic_load_n(&el->state, __ATOMIC_RELAXED)) {
-                               /* run the events */
-                               __atomic_store_n(&el->state, EVLOOP_STATE_LOCK|EVLOOP_STATE_RUN|EVLOOP_STATE_WAIT, __ATOMIC_RELAXED);
-                               __atomic_store_n(&el->holder, me, __ATOMIC_RELAXED);
-                               current_evloop = el;
-                               pthread_mutex_unlock(&mutex);
-                               sig_monitor(0, evloop_run, el);
-                               pthread_mutex_lock(&mutex);
-                       } else {
-                               /* no job and not events */
-                               running--;
-                               if (!running)
-                                       ERROR("Entering job deep sleep! Check your bindings.");
-                               me->waits = 1;
-                               pthread_cond_wait(&cond, &mutex);
-                               me->waits = 0;
-                               running++;
+               /* no job, check event loop wait */
+               } else if (evloop_get()) {
+                       if (evloop.state != 0) {
+                               /* busy ? */
+                               CRITICAL("Can't enter dispatch while in dispatch!");
+                               abort();
                        }
-#else
-               } else if (waitevt) {
-                       /* no job and not events */
+                       /* run the events */
+                       evloop.state = EVLOOP_STATE_RUN|EVLOOP_STATE_WAIT;
+                       pthread_mutex_unlock(&mutex);
+                       sig_monitor(0, evloop_run, NULL);
+                       pthread_mutex_lock(&mutex);
+                       evloop.state = 0;
+               } else {
+                       /* no job and no event loop */
                        running--;
                        if (!running)
                                ERROR("Entering job deep sleep! Check your bindings.");
@@ -436,49 +480,57 @@ static void thread_run(volatile struct thread *me)
                        pthread_cond_wait(&cond, &mutex);
                        me->waits = 0;
                        running++;
-               } else {
-                       /* wait for events */
-                       waitevt = 1;
-                       pthread_mutex_unlock(&mutex);
-                       sig_monitor(0, monitored_wait_and_dispatch, get_fdevepoll());
-                       pthread_mutex_lock(&mutex);
-                       waitevt = 0;
-#endif
                }
        }
+       /* cleanup */
+       evloop_release();
+       thread_leave();
+}
 
-       /* release the event loop */
-       if (current_evloop) {
-               __atomic_and_fetch(&current_evloop->state, ~EVLOOP_STATE_LOCK, __ATOMIC_RELAXED);
-               __atomic_store_n(&el->holder, NULL, __ATOMIC_RELAXED);
-               current_evloop = NULL;
-       }
+/**
+ * Main processing loop of external threads.
+ * The loop must be called with the mutex locked
+ * and it returns with the mutex locked.
+ * @param me the description of the thread to use
+ */
+static void thread_run_external(volatile struct thread *me)
+{
+       /* enter thread */
+       thread_enter(me);
 
-       /* unlink the current thread and cleanup */
-       prv = &threads;
-       while (*prv != me)
-               prv = &(*prv)->next;
-       *prv = me->next;
-       current_thread = me->upper;
-       if (!current_thread) {
-               sig_monitor_clean_timeouts();
-               started--;
-       }
+       /* loop until stopped */
+       me->waits = 1;
+       while (!me->stop)
+               pthread_cond_wait(&cond, &mutex);
+       me->waits = 0;
+       thread_leave();
 }
 
 /**
- * Entry point for created threads.
- * @param data not used
- * @return NULL
+ * Root for created threads.
  */
-static void *thread_main(void *data)
+static void thread_main()
 {
        struct thread me;
 
-       pthread_mutex_lock(&mutex);
        running++;
-       thread_run(&me);
+       started++;
+       sig_monitor_init_timeouts();
+       thread_run_internal(&me);
+       sig_monitor_clean_timeouts();
+       started--;
        running--;
+}
+
+/**
+ * Entry point for created threads.
+ * @param data not used
+ * @return NULL
+ */
+static void *thread_starter(void *data)
+{
+       pthread_mutex_lock(&mutex);
+       thread_main();
        pthread_mutex_unlock(&mutex);
        return NULL;
 }
@@ -492,7 +544,7 @@ static int start_one_thread()
        pthread_t tid;
        int rc;
 
-       rc = pthread_create(&tid, NULL, thread_main, NULL);
+       rc = pthread_create(&tid, NULL, thread_starter, NULL);
        if (rc != 0) {
                /* errno = rc; */
                WARNING("not able to start thread: %m");
@@ -524,7 +576,6 @@ int jobs_queue(
                void (*callback)(int, void*),
                void *arg)
 {
-       const char *info;
        struct job *job;
        int rc;
 
@@ -532,16 +583,13 @@ int jobs_queue(
 
        /* allocates the job */
        job = job_create(group, timeout, callback, arg);
-       if (!job) {
-               errno = ENOMEM;
-               info = "out of memory";
+       if (!job)
                goto error;
-       }
 
        /* check availability */
-       if (remains == 0) {
+       if (remains <= 0) {
+               ERROR("can't process job with threads: too many jobs");
                errno = EBUSY;
-               info = "too many jobs";
                goto error2;
        }
 
@@ -550,13 +598,12 @@ int jobs_queue(
                /* all threads are busy and a new can be started */
                rc = start_one_thread();
                if (rc < 0 && started == 0) {
-                       info = "can't start first thread";
+                       ERROR("can't start initial thread: %m");
                        goto error2;
                }
        }
 
        /* queues the job */
-       remains--;
        job_add(job);
 
        /* signal an existing job */
@@ -568,7 +615,6 @@ error2:
        job->next = free_jobs;
        free_jobs = job;
 error:
-       ERROR("can't process job with threads: %s, %m", info);
        pthread_mutex_unlock(&mutex);
        return -1;
 }
@@ -614,8 +660,6 @@ static int do_sync(
        /* allocates the job */
        job = job_create(group, timeout, sync_cb, sync);
        if (!job) {
-               ERROR("out of memory");
-               errno = ENOMEM;
                pthread_mutex_unlock(&mutex);
                return -1;
        }
@@ -624,7 +668,10 @@ static int do_sync(
        job_add(job);
 
        /* run until stopped */
-       thread_run(&sync->thread);
+       if (current_thread)
+               thread_run_internal(&sync->thread);
+       else
+               thread_run_external(&sync->thread);
        pthread_mutex_unlock(&mutex);
        return 0;
 }
@@ -658,41 +705,6 @@ int jobs_enter(
        return do_sync(group, timeout, enter_cb, &sync);
 }
 
-/**
- * Internal callback for evloop management.
- * The effect of this function is hidden: it exits
- * the waiting poll if any. Then it wakes up a thread
- * awaiting the evloop using signal.
- */
-static int on_evloop_efd(sd_event_source *s, int fd, uint32_t revents, void *userdata)
-{
-       uint64_t x;
-       struct evloop *evloop = userdata;
-       read(evloop->efd, &x, sizeof x);
-       pthread_mutex_lock(&mutex);
-       pthread_cond_broadcast(&evloop->cond);
-       pthread_mutex_unlock(&mutex);
-       return 1;
-}
-
-/**
- * unlock the event loop if needed by sending
- * an event.
- * @param el the event loop to unlock
- * @param wait wait the unlocked state of the event loop
- */
-static void unlock_evloop(struct evloop *el, int wait)
-{
-       /* wait for a modifiable event loop */
-       while (__atomic_load_n(&el->state, __ATOMIC_RELAXED) & EVLOOP_STATE_WAIT) {
-               uint64_t x = 1;
-               write(el->efd, &x, sizeof x);
-               if (!wait)
-                       break;
-               pthread_cond_wait(&el->cond, &mutex);
-       }
-}
-
 /**
  * Unlocks the execution flow designed by 'jobloop'.
  * @param jobloop indication of the flow to unlock
@@ -701,7 +713,6 @@ static void unlock_evloop(struct evloop *el, int wait)
 int jobs_leave(struct jobloop *jobloop)
 {
        struct thread *t;
-       int i;
 
        pthread_mutex_lock(&mutex);
        t = threads;
@@ -713,15 +724,8 @@ int jobs_leave(struct jobloop *jobloop)
                t->stop = 1;
                if (t->waits)
                        pthread_cond_broadcast(&cond);
-               else {
-                       i = (int)(sizeof evloop / sizeof *evloop);
-                       while(i) {
-                               if (evloop[--i].holder == t) {
-                                       unlock_evloop(&evloop[i], 0);
-                                       break;
-                               }
-                       }
-               }
+               else
+                       evloop_wakeup();
        }
        pthread_mutex_unlock(&mutex);
        return -!t;
@@ -755,13 +759,16 @@ int jobs_call(
        return do_sync(group, timeout, call_cb, &sync);
 }
 
-/* temporary hack */
-#if !defined(REMOVE_SYSTEMD_EVENT)
-__attribute__((unused))
-#endif
-static void evloop_callback(void *arg, uint32_t event, struct fdev *fdev)
+/**
+ * Internal callback for evloop management.
+ * The effect of this function is hidden: it exits
+ * the waiting poll if any. Then it wakes up a thread
+ * awaiting the evloop using signal.
+ */
+static int on_evloop_efd(sd_event_source *s, int fd, uint32_t revents, void *userdata)
 {
-       sig_monitor(0, evloop_run, arg);
+       evloop_on_efd_event();
+       return 1;
 }
 
 /**
@@ -770,74 +777,41 @@ static void evloop_callback(void *arg, uint32_t event, struct fdev *fdev)
  */
 static struct sd_event *get_sd_event_locked()
 {
-       struct evloop *el;
        int rc;
 
        /* creates the evloop on need */
-       el = &evloop[0];
-       if (!el->sdev) {
+       if (!evloop.sdev) {
                /* start the creation */
-               el->state = 0;
+               evloop.state = 0;
                /* creates the eventfd for waking up polls */
-               el->efd = eventfd(0, EFD_CLOEXEC);
-               if (el->efd < 0) {
+               evloop.efd = eventfd(0, EFD_CLOEXEC|EFD_SEMAPHORE);
+               if (evloop.efd < 0) {
                        ERROR("can't make eventfd for events");
                        goto error1;
                }
                /* create the systemd event loop */
-               rc = sd_event_new(&el->sdev);
+               rc = sd_event_new(&evloop.sdev);
                if (rc < 0) {
                        ERROR("can't make new event loop");
                        goto error2;
                }
                /* put the eventfd in the event loop */
-               rc = sd_event_add_io(el->sdev, NULL, el->efd, EPOLLIN, on_evloop_efd, el);
+               rc = sd_event_add_io(evloop.sdev, NULL, evloop.efd, EPOLLIN, on_evloop_efd, NULL);
                if (rc < 0) {
                        ERROR("can't register eventfd");
-#if !defined(REMOVE_SYSTEMD_EVENT)
-                       sd_event_unref(el->sdev);
-                       el->sdev = NULL;
-error2:
-                       close(el->efd);
-error1:
-                       return NULL;
-               }
-#else
-                       goto error3;
-               }
-               /* handle the event loop */
-               el->fdev = fdev_epoll_add(get_fdevepoll(), sd_event_get_fd(el->sdev));
-               if (!el->fdev) {
-                       ERROR("can't create fdev");
-error3:
-                       sd_event_unref(el->sdev);
+                       sd_event_unref(evloop.sdev);
+                       evloop.sdev = NULL;
 error2:
-                       close(el->efd);
+                       close(evloop.efd);
 error1:
-                       memset(el, 0, sizeof *el);
                        return NULL;
                }
-               fdev_set_autoclose(el->fdev, 0);
-               fdev_set_events(el->fdev, EPOLLIN);
-               fdev_set_callback(el->fdev, evloop_callback, el);
-#endif
        }
 
-       /* attach the event loop to the current thread */
-       if (current_evloop != el) {
-               if (current_evloop) {
-                       __atomic_and_fetch(&current_evloop->state, ~EVLOOP_STATE_LOCK, __ATOMIC_RELAXED);
-                       __atomic_store_n(&current_evloop->holder, NULL, __ATOMIC_RELAXED);
-               }
-               current_evloop = el;
-               __atomic_or_fetch(&el->state, EVLOOP_STATE_LOCK, __ATOMIC_RELAXED);
-               __atomic_store_n(&el->holder, current_thread, __ATOMIC_RELAXED);
-       }
-
-       /* wait for a modifiable event loop */
-       unlock_evloop(el, 1);
+       /* acquire the event loop */
+       evloop_acquire();
 
-       return el->sdev;
+       return evloop.sdev;
 }
 
 /**
@@ -847,30 +821,38 @@ error1:
 struct sd_event *jobs_get_sd_event()
 {
        struct sd_event *result;
+       struct thread lt;
+
+       /* ensure an existing thread environment */
+       if (!current_thread) {
+               memset(&lt, 0, sizeof lt);
+               current_thread = &lt;
+       }
 
+       /* process */
        pthread_mutex_lock(&mutex);
        result = get_sd_event_locked();
        pthread_mutex_unlock(&mutex);
 
-       return result;
-}
-
-#if defined(REMOVE_SYSTEMD_EVENT)
-/**
- * Gets the fdev_epoll item.
- * @return a fdev_epoll or NULL in case of error
- */
-struct fdev_epoll *jobs_get_fdev_epoll()
-{
-       struct fdev_epoll *result;
-
-       pthread_mutex_lock(&mutex);
-       result = get_fdevepoll();
-       pthread_mutex_unlock(&mutex);
+       /* release the faked thread environment if needed */
+       if (current_thread == &lt) {
+               /*
+                * Releasing it is needed because there is no way to guess
+                * when it has to be released really. But here is where it is
+                * hazardous: if the caller modifies the eventloop when it
+                * is waiting, there is no way to make the change effective.
+                * A workaround to achieve that goal is for the caller to
+                * require the event loop a second time after having modified it.
+                */
+               NOTICE("Requiring sd_event loop out of binder callbacks is hazardous!");
+               if (verbose_wants(Log_Level_Info))
+                       sig_monitor_dumpstack();
+               evloop_release();
+               current_thread = NULL;
+       }
 
        return result;
 }
-#endif
 
 /**
  * Enter the jobs processing loop.
@@ -883,7 +865,6 @@ struct fdev_epoll *jobs_get_fdev_epoll()
 int jobs_start(int allowed_count, int start_count, int waiter_count, void (*start)(int signum, void* arg), void *arg)
 {
        int rc, launched;
-       struct thread me;
        struct job *job;
 
        assert(allowed_count >= 1);
@@ -913,9 +894,9 @@ int jobs_start(int allowed_count, int start_count, int waiter_count, void (*star
                sd_event_set_watchdog(get_sd_event_locked(), 1);
 #endif
 
-       /* start at least one thread */
-       launched = 0;
-       while ((launched + 1) < start_count) {
+       /* start at least one thread: the current one */
+       launched = 1;
+       while (launched < start_count) {
                if (start_one_thread() != 0) {
                        ERROR("Not all threads can be started");
                        goto error;
@@ -925,18 +906,12 @@ int jobs_start(int allowed_count, int start_count, int waiter_count, void (*star
 
        /* queue the start job */
        job = job_create(NULL, 0, start, arg);
-       if (!job) {
-               ERROR("out of memory");
-               errno = ENOMEM;
+       if (!job)
                goto error;
-       }
        job_add(job);
-       remains--;
 
        /* run until end */
-       running++;
-       thread_run(&me);
-       running--;
+       thread_main();
        rc = 0;
 error:
        pthread_mutex_unlock(&mutex);
index 58adc1c..1a91968 100644 (file)
@@ -48,10 +48,5 @@ extern int jobs_start(
                void (*start)(int signum, void* arg),
                void *arg);
 
-#if !defined(REMOVE_SYSTEMD_EVENT)
 struct sd_event;
 extern struct sd_event *jobs_get_sd_event();
-#else
-struct fdev_epoll;
-extern struct fdev_epoll *jobs_get_fdev_epoll();
-#endif
index 15fe260..9e13fa1 100644 (file)
@@ -299,3 +299,8 @@ void sig_monitor(int timeout, void (*function)(int sig, void*), void *arg)
        else
                function(0, arg);
 }
+
+void sig_monitor_dumpstack()
+{
+       return dumpstack(1, 0);
+}
index 5fdac16..03c8ec4 100644 (file)
@@ -25,3 +25,5 @@ extern int sig_monitor_enable();
 
 extern void sig_monitor(int timeout, void (*function)(int sig, void*), void *arg);
 
+extern void sig_monitor_dumpstack();
+