mscp: maintain mscp_thread structs in list

Instead of m->threads array, struct mscp_thread instanes are
maintained in m->thread_list. This enables stable counter access
via mscp_get_stats().

include/mscp.h

@@ -184,7 +184,7 @@ int mscp_scan_join(struct mscp *m);
  *
  * @param m	mscp instance.
  *
- * @return 	0 on success, < 0 if an error occured.
+ * @return 	number of threads on success, < 0 if an error occured.
  *              mscp_get_error() can be used to retrieve error message.
  *
  * @see		mscp_join()

src/atomic.h

@@ -20,6 +20,8 @@ static inline refcnt refcnt_dec(refcnt *cnt)
 }
 
 
+/* mutex */
+
 typedef pthread_mutex_t lock;
 
 static inline void lock_init(lock *l)
@@ -44,12 +46,58 @@ static inline void lock_release_via_cleanup(void *l)
 	lock_release(l);
 }
 
-#define LOCK_ACQUIRE_THREAD(l)				\
+#define LOCK_ACQUIRE(l)						\
-		lock_acquire(l);			\
+	lock_acquire(l);					\
-		pthread_cleanup_push(lock_release_via_cleanup, l)
+	pthread_cleanup_push(lock_release_via_cleanup, l)
 
-
+#define LOCK_RELEASE()				\
-#define LOCK_RELEASE_THREAD() \
 	pthread_cleanup_pop(1)
 
+
+
+/* read/write lock */
+typedef pthread_rwlock_t rwlock;
+
+static inline void rwlock_init(rwlock *rw)
+{
+	pthread_rwlock_init(rw, NULL);
+}
+
+static inline void rwlock_read_acquire(rwlock *rw)
+{
+	int ret = pthread_rwlock_rdlock(rw);
+	assert(ret == 0);
+}
+
+static inline void rwlock_write_acquire(rwlock *rw)
+{
+	int ret = pthread_rwlock_wrlock(rw);
+	assert(ret == 0);
+}
+
+static inline void rwlock_release(rwlock *rw)
+{
+	int ret = pthread_rwlock_unlock(rw);
+	assert(ret == 0);
+}
+
+static inline void rwlock_release_via_cleanup(void *rw)
+{
+	rwlock_release(rw);
+}
+
+#define RWLOCK_READ_ACQUIRE(rw)						\
+	rwlock_read_acquire(rw);					\
+	pthread_cleanup_push(rwlock_release_via_cleanup, rw)
+
+#define RWLOCK_WRITE_ACQUIRE(rw)					\
+	rwlock_write_acquire(rw);					\
+	pthread_cleanup_push(rwlock_release_via_cleanup, rw)
+
+
+#define RWLOCK_RELEASE()			\
+	pthread_cleanup_pop(1)
+
+
+
 #endif /* _ATOMIC_H_ */

src/main.c

@@ -380,11 +380,11 @@ int main(int argc, char **argv)
 
 	ret = mscp_start(m);
 	if (ret < 0)
-		fprintf(stderr, "%s\n", mscp_get_error());
+		fprintf(stderr, "mscp_start: %s\n", mscp_get_error());
 
 	ret = mscp_join(m);
 	if (ret != 0)
-		fprintf(stderr, "%s\n", mscp_get_error());
+		fprintf(stderr, "mscp_join: %s\n", mscp_get_error());
 
 	pthread_cancel(tid_stat);
 	pthread_join(tid_stat, NULL);

src/mscp.c

@@ -40,11 +40,15 @@ struct mscp {
 	int			ret_scan;	/* return code from scan thread */
 
 	size_t			total_bytes;	/* total bytes to be transferred */
-	struct mscp_thread	*threads;
+
+	struct list_head	thread_list;
+	rwlock			thread_rwlock;
 };
 
 
 struct mscp_thread {
+	struct list_head	list;	/* mscp->thread_list */
+
 	struct mscp	*m;
 	int		id;
 	sftp_session	sftp;
@@ -56,7 +60,7 @@ struct mscp_thread {
 };
 
 struct src {
-	struct list_head list;
+	struct list_head list;	/* mscp->src_list */
 	char *path;
 };
 
@@ -211,7 +215,7 @@ struct mscp *mscp_init(const char *remote_host, int direction,
 	int n;
 
 	if (!remote_host) {
-		mscp_set_error("empty remote host\n");
+		mscp_set_error("empty remote host");
 		return NULL;
 	}
 
@@ -238,6 +242,9 @@ struct mscp *mscp_init(const char *remote_host, int direction,
 	INIT_LIST_HEAD(&m->path_list);
 	chunk_pool_init(&m->cp);
 
+	INIT_LIST_HEAD(&m->thread_list);
+	rwlock_init(&m->thread_rwlock);
+
 	if ((m->sem = sem_create(o->max_startups)) == NULL) {
 		mscp_set_error("sem_create: %s", strerrno());
 		goto free_out;
@@ -339,11 +346,14 @@ static int get_page_mask(void)
 
 static void mscp_stop_copy_thread(struct mscp *m)
 {
-	int n;
+	struct mscp_thread *t;
-        for (n = 0; n < m->opts->nr_threads; n++) {
+
-                if (m->threads[n].tid && !m->threads[n].finished)
+	RWLOCK_READ_ACQUIRE(&m->thread_rwlock);
-                        pthread_cancel(m->threads[n].tid);
+	list_for_each_entry(t, &m->thread_list, list) {
-        }
+		if (!t->finished)
+			pthread_cancel(t->tid);
+	}
+	RWLOCK_RELEASE();
 }
 
 static void mscp_stop_scan_thread(struct mscp *m)
@@ -448,10 +458,10 @@ int mscp_scan(struct mscp *m)
 		return -1;
 	}
 
-	/* need scan finished or over nr_threads chunks to determine
+	/* We wait for there are over nr_threads chunks to determine
-	 * actual number of threads (and connections). If the number
+	 * actual number of threads (and connections), or scan
-	 * of chunks are smaller than nr_threads, we adjust nr_threads
+	 * finished. If the number of chunks are smaller than
-	 * to the number of chunks.
+	 * nr_threads, we adjust nr_threads to the number of chunks.
 	 */
 	while (!chunk_pool_is_filled(&m->cp) &&
 	       chunk_pool_size(&m->cp) < m->opts->nr_threads)
@@ -474,9 +484,40 @@ int mscp_scan_join(struct mscp *m)
 
 static void *mscp_copy_thread(void *arg);
 
+static struct mscp_thread *mscp_copy_thread_spawn(struct mscp *m, int id)
+{
+	struct mscp_thread *t;
+	int ret;
+
+	t = malloc(sizeof(*t));
+	if (!t){
+		mscp_set_error("malloc: %s,", strerrno());
+		return NULL;
+	}
+
+	memset(t, 0, sizeof(*t));
+	t->m = m;
+	t->id = id;
+	if (m->cores == NULL)
+		t->cpu = -1;	/* not pinned to cpu */
+	else
+		t->cpu = m->cores[id % m->nr_cores];
+
+	ret = pthread_create(&t->tid, NULL, mscp_copy_thread, t);
+	if (ret < 0) {
+		mscp_set_error("pthread_create error: %d", ret);
+		free(t);
+		return NULL;
+	}
+
+	return t;
+}
+
+
 int mscp_start(struct mscp *m)
 {
-	int n, ret;
+	struct mscp_thread *t;
+	int n, ret = 0;
 
 	if ((n = chunk_pool_size(&m->cp)) < m->opts->nr_threads) {
 		mpr_notice(m->msg_fp, "we have only %d chunk(s). "
@@ -484,63 +525,46 @@ int mscp_start(struct mscp *m)
 		m->opts->nr_threads = n;
 	}
 
-	/* scan thread instances */
-	m->threads = calloc(m->opts->nr_threads, sizeof(struct mscp_thread));
-	memset(m->threads, 0, m->opts->nr_threads * sizeof(struct mscp_thread));
 	for (n = 0; n < m->opts->nr_threads; n++) {
-		struct mscp_thread *t = &m->threads[n];
+		t = mscp_copy_thread_spawn(m, n);
-		t->m = m;
+		if (!t) {
-		t->id = n;
+			mpr_err(m->msg_fp, "failed to spawn copy thread\n");
-		if (!m->cores)
+			break;
-			t->cpu = -1;
+		}
-		else
+		RWLOCK_WRITE_ACQUIRE(&m->thread_rwlock);
-			t->cpu = m->cores[n % m->nr_cores];
+		list_add_tail(&t->list, &m->thread_list);
-
+		RWLOCK_RELEASE();
-                ret = pthread_create(&t->tid, NULL, mscp_copy_thread, t);
-                if (ret < 0) {
-                        mscp_set_error("pthread_create error: %d", ret);
-                        mscp_stop(m);
-			return -1;
-                }
         }
 
-	return 0;
+	return n;
 }
 
 int mscp_join(struct mscp *m)
 {
+	struct mscp_thread *t;
 	int n, ret = 0;
 
 	/* waiting for scan thread joins... */
 	ret = mscp_scan_join(m);
 
         /* waiting for copy threads join... */
-        for (n = 0; n < m->opts->nr_threads; n++) {
+	RWLOCK_READ_ACQUIRE(&m->thread_rwlock);
-                if (m->threads[n].tid) {
+	list_for_each_entry(t, &m->thread_list, list) {
-                        pthread_join(m->threads[n].tid, NULL);
+		pthread_join(t->tid, NULL);
-                        if (m->threads[n].ret < 0)
+		if (t->ret < 0)
-                                ret = m->threads[n].ret;
+			ret = t->ret;
-                }
+		if (t->sftp) {
-        }
+			ssh_sftp_close(t->sftp);
+			t->sftp = NULL;
+		}
+	}
+	RWLOCK_RELEASE();
 
         if (m->first) {
                 ssh_sftp_close(m->first);
 		m->first = NULL;
 	}
 
-        if (m->threads) {
-                for (n = 0; n < m->opts->nr_threads; n++) {
-                        struct mscp_thread *t = &m->threads[n];
-			if (t->ret != 0)
-				ret = ret;
-
-                        if (t->sftp) {
-                                ssh_sftp_close(t->sftp);
-				t->sftp = NULL;
-			}
-                }
-        }
-
 	return ret;
 }
 
@@ -567,7 +591,7 @@ void *mscp_copy_thread(void *arg)
         }
 
 	if (sem_wait(m->sem) < 0) {
-		mscp_set_error("sem_wait: %s\n", strerrno());
+		mscp_set_error("sem_wait: %s", strerrno());
 		mpr_err(m->msg_fp, "%s", mscp_get_error());
 		goto err_out;
 	}
@@ -577,7 +601,7 @@ void *mscp_copy_thread(void *arg)
 	t->sftp = ssh_init_sftp_session(m->remote, m->ssh_opts);
 
 	if (sem_post(m->sem) < 0) {
-		mscp_set_error("sem_post: %s\n", strerrno());
+		mscp_set_error("sem_post: %s", strerrno());
 		mpr_err(m->msg_fp, "%s", mscp_get_error());
 		goto err_out;
 	}
@@ -629,6 +653,7 @@ void *mscp_copy_thread(void *arg)
         return NULL;
 
 err_out:
+	t->finished = true;
 	t->ret = -1;
 	return NULL;
 }
@@ -658,6 +683,13 @@ static void free_chunk(struct list_head *list)
 	free(c);
 }
 
+static void free_thread(struct list_head *list)
+{
+	struct mscp_thread *t;
+	t = list_entry(list, typeof(*t), list);
+	free(t);
+}
+
 void mscp_cleanup(struct mscp *m)
 {
         if (m->first) {
@@ -674,10 +706,9 @@ void mscp_cleanup(struct mscp *m)
 	chunk_pool_release(&m->cp);
 	chunk_pool_init(&m->cp);
 
-	if (m->threads) {
+	RWLOCK_WRITE_ACQUIRE(&m->thread_rwlock);
-		free(m->threads);
+	list_free_f(&m->thread_list, free_thread);
-		m->threads = NULL;
+	RWLOCK_RELEASE();
-	}
 }
 
 void mscp_free(struct mscp *m)
@@ -694,16 +725,19 @@ void mscp_free(struct mscp *m)
 
 void mscp_get_stats(struct mscp *m, struct mscp_stats *s)
 {
+	struct mscp_thread *t;
 	bool finished = true;
-	int n;
 
 	s->total = m->total_bytes;
-	for (s->done = 0, n = 0; n < m->opts->nr_threads; n++) {
+	s->done = 0;
-		s->done += m->threads[n].done;
 
-		if (!m->threads[n].done)
+	RWLOCK_READ_ACQUIRE(&m->thread_rwlock);
+	list_for_each_entry(t, &m->thread_list, list) {
+		s->done += t->done;
+		if (!t->finished)
 			finished = false;
 	}
+	RWLOCK_RELEASE();
 
 	s->finished = finished;
 }

src/path.c

@@ -27,10 +27,10 @@ void chunk_pool_init(struct chunk_pool *cp)
 
 static void chunk_pool_add(struct chunk_pool *cp, struct chunk *c)
 {
-	LOCK_ACQUIRE_THREAD(&cp->lock);
+	LOCK_ACQUIRE(&cp->lock);
 	list_add_tail(&c->list, &cp->list);
 	cp->count += 1;
-	LOCK_RELEASE_THREAD();
+	LOCK_RELEASE();
 }
 
 void chunk_pool_set_filled(struct chunk_pool *cp)
@@ -54,7 +54,7 @@ struct chunk *chunk_pool_pop(struct chunk_pool *cp)
 	struct list_head *first;
 	struct chunk *c = NULL;
 
-	LOCK_ACQUIRE_THREAD(&cp->lock);
+	LOCK_ACQUIRE(&cp->lock);
 	first = cp->list.next;
 	if (list_empty(&cp->list)) {
 		if (!chunk_pool_is_filled(cp))
@@ -65,7 +65,7 @@ struct chunk *chunk_pool_pop(struct chunk_pool *cp)
 		c = list_entry(first, struct chunk, list);
 		list_del(first);
 	}
-	LOCK_RELEASE_THREAD();
+	LOCK_RELEASE();
 
 	/* return CHUNK_POP_WAIT would be very rare case, because it
 	 * means copying over SSH is faster than traversing
@@ -363,7 +363,7 @@ static int prepare_dst_path(FILE *msg_fp, struct path *p, sftp_session dst_sftp)
 {
 	int ret = 0;
 
-	LOCK_ACQUIRE_THREAD(&p->lock);
+	LOCK_ACQUIRE(&p->lock);
 	if (p->state == FILE_STATE_INIT) {
 		if (touch_dst_path(p, dst_sftp) < 0) {
 			ret = -1;
@@ -374,7 +374,7 @@ static int prepare_dst_path(FILE *msg_fp, struct path *p, sftp_session dst_sftp)
 	}
 
 out:
-	LOCK_RELEASE_THREAD();
+	LOCK_RELEASE();
 	return ret;
 }
 

src/path.h

@@ -201,11 +201,16 @@ static int mscp_stat(const char *path, mstat *s, sftp_session sftp)
 
         if (sftp) {
                 s->r = sftp_stat(sftp, path);
-                if (!s->r)
+                if (!s->r) {
+			mscp_set_error("sftp_stat: %s %s",
+				       sftp_get_ssh_error(sftp), path);
                         return -1;
+		}
         } else {
-                if (stat(path, &s->l) < 0)
+                if (stat(path, &s->l) < 0) {
+			mscp_set_error("stat: %s %s", strerrno(), path);
                         return -1;
+		}
         }
 
         return 0;

test/test_python.py

@@ -104,6 +104,12 @@ def test_login_failed():
     with pytest.raises(RuntimeError) as e:
         m.connect()
 
+def test_get_stat_before_copy_start():
+    m = mscp.mscp("localhost", mscp.LOCAL2REMOTE)
+    m.connect()
+    (total, done, finished) = m.stats()
+    assert total == 0 and done == 0
+
 
 param_invalid_kwargs = [
     { "nr_threads": -1 },