gh actions: upgrade to 1.2.11-1

debian/changelog

@@ -1,3 +1,27 @@
+libserverframe (1.2.11-1) unstable; urgency=medium
+
+  * upgrade to 1.2.11-1
+
+ -- YuQing <384681@qq.com>  Sun, 23 Nov 2025 10:48:22 +0000
+
+libserverframe (1.2.11-1) unstable; urgency=medium
+
+  * upgrade to 1.2.11-1
+
+ -- YuQing <384681@qq.com>  Sun, 23 Nov 2025 10:00:56 +0000
+
+libserverframe (1.2.11-1) unstable; urgency=medium
+
+  * upgrade to 1.2.11-1
+
+ -- YuQing <384681@qq.com>  Sun, 23 Nov 2025 09:06:43 +0000
+
+libserverframe (1.2.8-1) unstable; urgency=medium
+
+  * upgrade to 1.2.8-1
+
+ -- YuQing <384681@qq.com>  Sat, 16 Aug 2025 16:32:03 +0000
+
 libserverframe (1.2.7-1) unstable; urgency=medium
 
   * upgrade to 1.2.7-1

debian/substvars

@@ -1 +1 @@
-libfastcommon:Version=1.0.77
+libfastcommon:Version=1.0.83

libserverframe.spec

@@ -2,7 +2,7 @@
 %define CommitVersion %(echo $COMMIT_VERSION)
 
 Name: libserverframe
-Version: 1.2.8
+Version: 1.2.11
 Release: 1%{?dist}
 Summary: network framework library
 License: AGPL v3.0
@@ -12,9 +12,9 @@ Source: http://github.com/happyfish100/libserverframe/%{name}-%{version}.tar.gz
 
 BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)
 
-BuildRequires: libfastcommon-devel >= 1.0.78
+BuildRequires: libfastcommon-devel >= 1.0.83
 Requires: %__cp %__mv %__chmod %__grep %__mkdir %__install %__id
-Requires: libfastcommon >= 1.0.78
+Requires: libfastcommon >= 1.0.83
 
 %description
 common framework library 

make.sh

@@ -9,10 +9,13 @@ DEBUG_FLAG=0
 
 if [ -f /usr/include/fastcommon/_os_define.h ]; then
   OS_BITS=$(grep -F OS_BITS /usr/include/fastcommon/_os_define.h | awk '{print $NF;}')
+  USE_URING=$(grep -F IOEVENT_USE_URING /usr/include/fastcommon/_os_define.h | awk '{print $NF;}')
 elif [ -f /usr/local/include/fastcommon/_os_define.h ]; then
   OS_BITS=$(grep -F OS_BITS /usr/local/include/fastcommon/_os_define.h | awk '{print $NF;}')
+  USE_URING=$(grep -F IOEVENT_USE_URING /usr/local/include/fastcommon/_os_define.h | awk '{print $NF;}')
 else
   OS_BITS=64
+  USE_URING=''
 fi
 
 uname=$(uname)
@@ -49,6 +52,9 @@ LIBS=''
 uname=$(uname)
 if [ "$uname" = "Linux" ]; then
   CFLAGS="$CFLAGS"
+  if [ -n "$USE_URING" ]; then
+    LIBS="$LIBS -luring"
+  fi
 elif [ "$uname" = "FreeBSD" ] || [ "$uname" = "Darwin" ]; then
   CFLAGS="$CFLAGS"
   if [ "$uname" = "Darwin" ]; then

src/idempotency/client/client_channel.c

@@ -174,7 +174,6 @@ static struct fast_task_info *alloc_channel_task(IdempotencyClientChannel
         *channel, const uint32_t hash_code, const FCCommunicationType comm_type,
         const char *server_ip, const uint16_t port, int *err_no)
 {
-    int len;
     struct fast_task_info *task;
     SFAddressFamilyHandler *fh;
     SFNetworkHandler *handler;
@@ -195,12 +194,7 @@ static struct fast_task_info *alloc_channel_task(IdempotencyClientChannel
         return NULL;
     }
 
-    len = strlen(server_ip);
+    fc_safe_strcpy(task->server_ip, server_ip);
-    if (len >= sizeof(task->server_ip)) {
-        len = sizeof(task->server_ip) - 1;
-    }
-    memcpy(task->server_ip, server_ip, len);
-    *(task->server_ip + len) = '\0';
     task->port = port;
     task->arg = channel;
     task->thread_data = g_sf_context.thread_data +
@@ -209,7 +203,8 @@ static struct fast_task_info *alloc_channel_task(IdempotencyClientChannel
     channel->last_connect_time = g_current_time;
     if ((*err_no=sf_nio_notify(task, SF_NIO_STAGE_CONNECT)) != 0) {
         channel->in_ioevent = 0;   //rollback
-        sf_release_task(task);
+        __sync_sub_and_fetch(&task->reffer_count, 1);
+        free_queue_push(task);
         return NULL;
     }
     return task;
@@ -221,6 +216,14 @@ int idempotency_client_channel_check_reconnect(
     int result;
     char formatted_ip[FORMATTED_IP_SIZE];
 
+#if IOEVENT_USE_URING
+    struct fast_task_info *task;
+    task = channel->task;
+    if (SF_CTX->use_io_uring && FC_ATOMIC_GET(task->reffer_count) > 1) {
+        return 0;
+    }
+#endif
+
     if (!__sync_bool_compare_and_swap(&channel->in_ioevent, 0, 1)) {
         return 0;
     }
@@ -237,6 +240,9 @@ int idempotency_client_channel_check_reconnect(
                 formatted_ip, channel->task->port);
     }
 
+    if (channel->task->event.fd >= 0) {
+        channel->task->handler->close_connection(channel->task);
+    }
     __sync_bool_compare_and_swap(&channel->task->canceled, 1, 0);
     if ((result=sf_nio_notify(channel->task, SF_NIO_STAGE_CONNECT)) == 0) {
         channel->last_connect_time = g_current_time;
@@ -348,8 +354,8 @@ int idempotency_client_channel_push(struct idempotency_client_channel *channel,
     receipt->req_id = req_id;
     fc_queue_push_ex(&channel->queue, receipt, &notify);
     if (notify) {
-        if (__sync_add_and_fetch(&channel->in_ioevent, 0)) {
+        if (FC_ATOMIC_GET(channel->in_ioevent)) {
-            if (__sync_add_and_fetch(&channel->established, 0)) {
+            if (FC_ATOMIC_GET(channel->established)) {
                 sf_nio_notify(channel->task, SF_NIO_STAGE_CONTINUE);
             }
         } else {

src/idempotency/client/client_channel.h

@@ -76,13 +76,13 @@ static inline void idempotency_client_channel_set_id_key(
 static inline int idempotency_client_channel_check_wait_ex(
         struct idempotency_client_channel *channel, const int timeout)
 {
-    if (__sync_add_and_fetch(&channel->established, 0)) {
+    if (FC_ATOMIC_GET(channel->established)) {
         return 0;
     }
 
     idempotency_client_channel_check_reconnect(channel);
     lcp_timedwait_sec(&channel->lcp, timeout);
-    if (__sync_add_and_fetch(&channel->established, 0)) {
+    if (FC_ATOMIC_GET(channel->established)) {
         return 0;
     } else {
         /*

src/idempotency/client/receipt_handler.c

@@ -49,6 +49,10 @@ static IdempotencyReceiptGlobalVars receipt_global_vars;
 
 static int receipt_init_task(struct fast_task_info *task, void *arg)
 {
+#if IOEVENT_USE_URING
+    FC_URING_IS_CLIENT(task) = true;
+#endif
+
     if (RDMA_INIT_CONNECTION != NULL) {
         return RDMA_INIT_CONNECTION(task, arg);
     } else {
@@ -92,7 +96,6 @@ static void receipt_task_finish_cleanup(struct fast_task_info *task)
 
     if (task->event.fd >= 0) {
         sf_task_detach_thread(task);
-        task->handler->close_connection(task);
     }
 
     sf_nio_reset_task_length(task);
@@ -282,7 +285,7 @@ static int deal_setup_channel_response(struct fast_task_info *task)
     }
 
     channel = (IdempotencyClientChannel *)task->arg;
-    if (__sync_add_and_fetch(&channel->established, 0)) {
+    if (FC_ATOMIC_GET(channel->established)) {
         format_ip_address(task->server_ip, formatted_ip);
         logWarning("file: "__FILE__", line: %d, "
                 "response from server %s:%u, unexpected cmd: "
@@ -520,6 +523,7 @@ static int do_init(FCAddressPtrArray *address_array)
 {
     const int task_arg_size = 0;
     const bool double_buffers = false;
+    const bool need_shrink_task_buffer = false;
     const bool explicit_post_recv = false;
     int result;
     int bytes;
@@ -552,8 +556,8 @@ static int do_init(FCAddressPtrArray *address_array)
             NULL, sf_proto_set_body_length, NULL, NULL, receipt_deal_task,
             receipt_task_finish_cleanup, receipt_recv_timeout_callback,
             1000, sizeof(SFCommonProtoHeader), TASK_PADDING_SIZE,
-            task_arg_size, double_buffers, explicit_post_recv,
+            task_arg_size, double_buffers, need_shrink_task_buffer,
-            receipt_init_task, pd, NULL);
+            explicit_post_recv, receipt_init_task, pd, NULL);
 }
 
 int receipt_handler_init(FCAddressPtrArray *address_array)
@@ -565,7 +569,6 @@ int receipt_handler_init(FCAddressPtrArray *address_array)
     }
 
     sf_enable_thread_notify(true);
-    sf_set_remove_from_ready_list(false);
     fc_sleep_ms(100);
 
     return 0;

src/sf_define.h

@@ -96,6 +96,8 @@
 #define SF_BINLOG_SOURCE_USER        'U'  //by user call
 #define SF_BINLOG_SOURCE_REPLAY      'R'  //by binlog replay
 
+#define SF_LOG_SCHEDULE_ENTRIES_COUNT 3
+
 #ifdef __cplusplus
 extern "C" {
 #endif

src/sf_global.c

@@ -47,11 +47,12 @@ SFGlobalVariables g_sf_global_vars = {
     {0, 0}, NULL, {NULL, 0}
 };
 
-SFContext g_sf_context = {{'\0'}, NULL, 0, false, sf_address_family_auto,
+SFContext g_sf_context = {{'\0'}, NULL, 0, false, false, false,
-    {{AF_UNSPEC, {{true, fc_comm_type_sock}, {false, fc_comm_type_rdma}}},
+    sf_address_family_auto, {{AF_UNSPEC, {{true, fc_comm_type_sock},
-        {AF_UNSPEC, {{true, fc_comm_type_sock}, {false, fc_comm_type_rdma}}}},
+        {false, fc_comm_type_rdma}}},
+    {AF_UNSPEC, {{true, fc_comm_type_sock}, {false, fc_comm_type_rdma}}}},
     {DEFAULT_MAX_CONNECTONS, SF_DEF_MAX_PACKAGE_SIZE, SF_DEF_MIN_BUFF_SIZE,
-    SF_DEF_MAX_BUFF_SIZE}, 1, DEFAULT_WORK_THREADS, 0, true, true, true,
+    SF_DEF_MAX_BUFF_SIZE}, 1, DEFAULT_WORK_THREADS, 0, true, true,
     {false, 0, 0}, {sf_task_finish_clean_up}
 };
 
@@ -669,6 +670,10 @@ int sf_load_context_from_config_ex(SFContext *sf_context,
     int inner_port;
     int outer_port;
     int port;
+#if IOEVENT_USE_URING
+    bool global_use_send_zc;
+    bool use_send_zc;
+#endif
     int i;
     int result;
 
@@ -705,6 +710,17 @@ int sf_load_context_from_config_ex(SFContext *sf_context,
         outer_port = config->default_outer_port;
     }
 
+#if IOEVENT_USE_URING
+    global_use_send_zc = iniGetBoolValue(NULL, "use_send_zc",
+            config->ini_ctx.context, true);
+    if (config->ini_ctx.section_name == NULL) {
+        use_send_zc = global_use_send_zc;
+    } else {
+        use_send_zc = iniGetBoolValue(config->ini_ctx.section_name,
+                "use_send_zc", config->ini_ctx.context, global_use_send_zc);
+    }
+#endif
+
     for (i=0; i<SF_ADDRESS_FAMILY_COUNT; i++) {
         fh = sf_context->handlers + i;
         fh->ctx = sf_context;
@@ -747,9 +763,15 @@ int sf_load_context_from_config_ex(SFContext *sf_context,
         rdma_handler->inner.enabled = sock_handler->inner.enabled;
         rdma_handler->outer.port = sock_handler->outer.port;
         rdma_handler->outer.enabled = sock_handler->outer.enabled;
-
     }
 
+#if IOEVENT_USE_URING
+    sf_context->use_io_uring = (config->comm_type == fc_comm_type_sock);
+    sf_context->use_send_zc = sf_context->use_io_uring ? use_send_zc : false;
+#else
+    sf_context->use_io_uring = false;
+#endif
+
     sf_context->accept_threads = iniGetIntValue(
             config->ini_ctx.section_name,
             "accept_threads", config->ini_ctx.context, 1);
@@ -934,6 +956,12 @@ void sf_context_config_to_string(const SFContext *sf_context,
             ", address_family=%s, accept_threads=%d, work_threads=%d",
             get_address_family_caption(sf_context->address_family),
             sf_context->accept_threads, sf_context->work_threads);
+
+#if IOEVENT_USE_URING
+    len += snprintf(output + len, size - len, ", use_io_uring=%d, "
+            "use_send_zc=%d", sf_context->use_io_uring,
+            sf_context->use_send_zc);
+#endif
 }
 
 void sf_log_config_to_string_ex(SFLogConfig *log_cfg, const char *caption,
@@ -1004,18 +1032,27 @@ void sf_global_config_to_string_ex(const char *max_pkg_size_item_nm,
 
     len = snprintf(output, size,
             "base_path=%s, max_connections=%d, connect_timeout=%d, "
-            "network_timeout=%d, thread_stack_size=%d KB, "
+            "network_timeout=%d, thread_stack_size=%d KB, %s, ",
-            "%s, tcp_quick_ack=%d, log_level=%s, "
+            SF_G_BASE_PATH_STR,
-            "run_by_group=%s, run_by_user=%s, ", SF_G_BASE_PATH_STR,
             g_sf_global_vars.net_buffer_cfg.max_connections,
             g_sf_global_vars.net_buffer_cfg.connect_timeout,
             g_sf_global_vars.net_buffer_cfg.network_timeout,
-            g_sf_global_vars.thread_stack_size / 1024,
+            g_sf_global_vars.thread_stack_size / 1024, pkg_buff);
-            pkg_buff, g_sf_global_vars.tcp_quick_ack,
+
+#if IOEVENT_USE_URING
+    len += snprintf(output + len, size - len, "use_io_uring=%d, "
+            "use_send_zc=%d, ", g_sf_context.use_io_uring,
+            g_sf_context.use_send_zc);
+#endif
+
+    len += snprintf(output + len, size - len,
+            "tcp_quick_ack=%d, "
+            "log_level=%s, "
+            "run_by_group=%s, run_by_user=%s, ",
+            g_sf_global_vars.tcp_quick_ack,
             log_get_level_caption(),
             g_sf_global_vars.run_by.group,
-            g_sf_global_vars.run_by.user
+            g_sf_global_vars.run_by.user);
-            );
 
     sf_log_config_to_string(&g_sf_global_vars.error_log,
             "error-log", output + len, size - len);

src/sf_nio.c

@@ -41,6 +41,9 @@
 #include "sf_service.h"
 #include "sf_nio.h"
 
+static int sf_client_sock_write(int sock, const int event, void *arg);
+static int sf_client_sock_read(int sock, const int event, void *arg);
+
 void sf_set_parameters_ex(SFContext *sf_context, const int header_size,
         sf_set_body_length_callback set_body_length_func,
         sf_alloc_recv_buffer_callback alloc_recv_buffer_func,
@@ -59,19 +62,43 @@ void sf_set_parameters_ex(SFContext *sf_context, const int header_size,
     sf_context->callbacks.release_buffer = release_buffer_callback;
 }
 
+#if IOEVENT_USE_URING
+#define CLEAR_OP_TYPE_AND_RELEASE_TASK(task) \
+    FC_URING_OP_TYPE(task) = IORING_OP_NOP;  \
+    sf_release_task(task)
+
+static int sf_uring_cancel_done(int sock, const int event, void *arg)
+{
+    struct fast_task_info *task;
+
+    task = (struct fast_task_info *)arg;
+    if (event != IOEVENT_TIMEOUT) {
+        CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+    }
+    return 0;
+}
+#endif
+
 void sf_task_detach_thread(struct fast_task_info *task)
 {
-    ioevent_detach(&task->thread_data->ev_puller, task->event.fd);
+#if IOEVENT_USE_URING
+    if (SF_CTX->use_io_uring) {
+        if (FC_URING_OP_TYPE(task) != IORING_OP_NOP) {
+            task->event.callback = (IOEventCallback)sf_uring_cancel_done;
+            uring_prep_cancel(task);
+        }
+    } else {
+#endif
+        ioevent_detach(&task->thread_data->ev_puller, task->event.fd);
+#if IOEVENT_USE_URING
+    }
+#endif
 
     if (task->event.timer.expires > 0) {
         fast_timer_remove(&task->thread_data->timer,
                 &task->event.timer);
         task->event.timer.expires = 0;
     }
-
-    if (SF_CTX->remove_from_ready_list) {
-        ioevent_remove(&task->thread_data->ev_puller, task);
-    }
 }
 
 void sf_task_switch_thread(struct fast_task_info *task,
@@ -92,28 +119,38 @@ static inline void release_iovec_buffer(struct fast_task_info *task)
     }
 }
 
+void sf_socket_close_connection(struct fast_task_info *task)
+{
+#if IOEVENT_USE_URING
+    if (SF_CTX->use_io_uring) {
+        if (uring_prep_close_fd(task) != 0) {
+            close(task->event.fd);
+        }
+    } else {
+#endif
+        close(task->event.fd);
+#if IOEVENT_USE_URING
+    }
+#endif
+    task->event.fd = -1;
+}
+
 void sf_task_finish_clean_up(struct fast_task_info *task)
 {
-    /*
-    assert(task->event.fd >= 0);
-    if (task->event.fd < 0) {
-        logWarning("file: "__FILE__", line: %d, "
-                "task: %p already cleaned",
-                __LINE__, task);
-        return;
-    }
-    */
-
-    if (task->finish_callback != NULL) {
-        task->finish_callback(task);
-        task->finish_callback = NULL;
-    }
-
     release_iovec_buffer(task);
     sf_task_detach_thread(task);
-    task->handler->close_connection(task);
 
-    __sync_fetch_and_sub(&g_sf_global_vars.connection_stat.current_count, 1);
+#if IOEVENT_USE_URING
+    if (!SF_CTX->use_io_uring) {
+#endif
+        task->handler->close_connection(task);
+        __sync_fetch_and_sub(&g_sf_global_vars.
+                connection_stat.current_count, 1);
+
+#if IOEVENT_USE_URING
+    }
+#endif
+
     sf_release_task(task);
 }
 
@@ -139,46 +176,115 @@ static inline int set_write_event(struct fast_task_info *task)
     return 0;
 }
 
+#if IOEVENT_USE_URING
+static inline int prepare_first_recv(struct fast_task_info *task)
+{
+    if (SF_CTX->callbacks.alloc_recv_buffer != NULL) {
+        return uring_prep_recv_data(task, task->recv.ptr->data,
+                SF_CTX->header_size);
+    } else {
+        return uring_prep_first_recv(task);
+    }
+}
+
+static inline int prepare_next_recv(struct fast_task_info *task)
+{
+    int recv_bytes;
+
+    if (task->recv.ptr->length == 0) { //recv header
+        recv_bytes = SF_CTX->header_size - task->recv.ptr->offset;
+        return uring_prep_recv_data(task, task->recv.ptr->data +
+                task->recv.ptr->offset, recv_bytes);
+    } else {
+        recv_bytes = task->recv.ptr->length - task->recv.ptr->offset;
+        if (task->recv_body == NULL) {
+            return uring_prep_recv_data(task, task->recv.ptr->data +
+                    task->recv.ptr->offset, recv_bytes);
+        } else {
+            return uring_prep_recv_data(task, task->recv_body +
+                    (task->recv.ptr->offset - SF_CTX->
+                     header_size), recv_bytes);
+        }
+    }
+}
+#endif
+
 static inline int set_read_event(struct fast_task_info *task)
 {
     int result;
 
-    if (task->event.callback == (IOEventCallback)sf_client_sock_read) {
+#if IOEVENT_USE_URING
-        return 0;
+    if (SF_CTX->use_io_uring) {
-    }
+        if (FC_URING_OP_TYPE(task) != IORING_OP_NOP) {
+            if (FC_URING_OP_TYPE(task) == IORING_OP_RECV) {
+                logWarning("file: "__FILE__", line: %d, "
+                        "trigger recv again!", __LINE__);
+                return 0;
+            } else {
+                logWarning("file: "__FILE__", line: %d, "
+                        "another operation in progress, op_type: %d!",
+                        __LINE__, FC_URING_OP_TYPE(task));
+                return EBUSY;
+            }
+        }
 
-    task->event.callback = (IOEventCallback)sf_client_sock_read;
+        if (task->event.callback != (IOEventCallback)sf_client_sock_read) {
-    if (ioevent_modify(&task->thread_data->ev_puller,
+            task->event.callback = (IOEventCallback)sf_client_sock_read;
-                task->event.fd, IOEVENT_READ, task) != 0)
+        }
-    {
+        if ((result=prepare_first_recv(task)) != 0) {
-        result = errno != 0 ? errno : ENOENT;
+            ioevent_add_to_deleted_list(task);
-        ioevent_add_to_deleted_list(task);
+            return result;
+        }
+    } else {
+#endif
 
-        logError("file: "__FILE__", line: %d, "
+        if (task->event.callback == (IOEventCallback)sf_client_sock_read) {
-                "ioevent_modify fail, "
+            return 0;
-                "errno: %d, error info: %s",
+        }
-                __LINE__, result, strerror(result));
+        task->event.callback = (IOEventCallback)sf_client_sock_read;
-        return result;
+        if (ioevent_modify(&task->thread_data->ev_puller,
+                    task->event.fd, IOEVENT_READ, task) != 0)
+        {
+            result = errno != 0 ? errno : ENOENT;
+            ioevent_add_to_deleted_list(task);
+
+            logError("file: "__FILE__", line: %d, "
+                    "ioevent_modify fail, "
+                    "errno: %d, error info: %s",
+                    __LINE__, result, strerror(result));
+            return result;
+        }
+
+#if IOEVENT_USE_URING
     }
+#endif
 
     return 0;
 }
 
 int sf_set_read_event(struct fast_task_info *task)
 {
+#if IOEVENT_USE_URING
+    if (SF_CTX->use_io_uring) {
+        return 0;
+    }
+#endif
+
+    /* reset recv offset and length */
     task->recv.ptr->offset = 0;
     task->recv.ptr->length = 0;
+
     task->nio_stages.current = SF_NIO_STAGE_RECV;
     return set_read_event(task);
 }
 
-static inline int sf_ioevent_add(struct fast_task_info *task,
+static inline int sf_ioevent_add(struct fast_task_info *task)
-        IOEventCallback callback, const int timeout)
 {
     int result;
 
     result = ioevent_set(task, task->thread_data, task->event.fd,
-            IOEVENT_READ, callback, timeout);
+            IOEVENT_READ, (IOEventCallback)sf_client_sock_read,
+            SF_CTX->net_buffer_cfg.network_timeout);
     return result > 0 ? -1 * result : result;
 }
 
@@ -196,8 +302,7 @@ static inline void inc_connection_current_count()
 static inline int sf_nio_init(struct fast_task_info *task)
 {
     inc_connection_current_count();
-    return sf_ioevent_add(task, (IOEventCallback)sf_client_sock_read,
+    return sf_ioevent_add(task);
-            SF_CTX->net_buffer_cfg.network_timeout);
 }
 
 int sf_socket_async_connect_check(struct fast_task_info *task)
@@ -212,7 +317,7 @@ int sf_socket_async_connect_check(struct fast_task_info *task)
     return result;
 }
 
-static int sf_client_connect_done(int sock, short event, void *arg)
+static int sf_client_connect_done(int sock, const int event, void *arg)
 {
     int result;
     struct fast_task_info *task;
@@ -220,16 +325,31 @@ static int sf_client_connect_done(int sock, short event, void *arg)
 
     task = (struct fast_task_info *)arg;
     if (task->canceled) {
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring && event != IOEVENT_TIMEOUT) {
+            CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+        }
+#endif
         return ENOTCONN;
     }
 
-    if (event & IOEVENT_TIMEOUT) {
+    if (event == IOEVENT_TIMEOUT) {
         result = ETIMEDOUT;
     } else {
-        result = task->handler->async_connect_check(task);
+#if IOEVENT_USE_URING
-        if (result == EINPROGRESS) {
+        if (SF_CTX->use_io_uring) {
-            return 0;
+            CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+            result = (task->event.res < 0 ? -1 * task->event.res :
+                    task->event.res);
+        } else {
+#endif
+            result = task->handler->async_connect_check(task);
+            if (result == EINPROGRESS) {
+                return 0;
+            }
+#if IOEVENT_USE_URING
         }
+#endif
     }
 
     if (SF_CTX->callbacks.connect_done != NULL) {
@@ -260,14 +380,26 @@ static int sf_client_connect_done(int sock, short event, void *arg)
 int sf_socket_async_connect_server(struct fast_task_info *task)
 {
     int result;
-    if ((task->event.fd=socketCreateEx2(AF_UNSPEC, task->server_ip,
-                    O_NONBLOCK, NULL, &result)) < 0)
-    {
-        return result > 0 ? -1 * result : result;
-    }
 
-    return asyncconnectserverbyip(task->event.fd,
+#if IOEVENT_USE_URING
-            task->server_ip, task->port);
+    if (SF_CTX->use_io_uring) {
+        if ((result=uring_prep_connect(task)) != 0) {
+            return result;
+        }
+        return EINPROGRESS;
+    } else {
+#endif
+        if ((task->event.fd=socketCreateEx2(AF_UNSPEC, task->server_ip,
+                        O_NONBLOCK, NULL, &result)) < 0)
+        {
+            return result > 0 ? -1 * result : result;
+        }
+
+        return asyncconnectserverbyip(task->event.fd,
+                task->server_ip, task->port);
+#if IOEVENT_USE_URING
+    }
+#endif
 }
 
 static int sf_async_connect_server(struct fast_task_info *task)
@@ -287,10 +419,7 @@ static int sf_async_connect_server(struct fast_task_info *task)
         }
 
         if (result == 0) {
-            if ((result=sf_ioevent_add(task, (IOEventCallback)
+            if ((result=sf_ioevent_add(task)) != 0) {
-                            sf_client_sock_read, SF_CTX->
-                            net_buffer_cfg.network_timeout)) != 0)
-            {
                 return result;
             }
 
@@ -331,10 +460,12 @@ static int sf_nio_deal_task(struct fast_task_info *task, const int stage)
         case SF_NIO_STAGE_RECV:
             task->nio_stages.current = SF_NIO_STAGE_RECV;
             if ((result=set_read_event(task)) == 0) {
-                if (sf_client_sock_read(task->event.fd,
+                if (!SF_CTX->use_io_uring) {
-                            IOEVENT_READ, task) < 0)
+                    if (sf_client_sock_read(task->event.fd,
-                {
+                                IOEVENT_READ, task) < 0)
-                    result = errno != 0 ? errno : EIO;
+                    {
+                        result = errno != 0 ? errno : EIO;
+                    }
                 }
             }
             break;
@@ -345,10 +476,7 @@ static int sf_nio_deal_task(struct fast_task_info *task, const int stage)
             result = SF_CTX->callbacks.deal_task(task, SF_NIO_STAGE_CONTINUE);
             break;
         case SF_NIO_STAGE_FORWARDED:  //forward by other thread
-            if ((result=sf_ioevent_add(task, (IOEventCallback)
+            if ((result=sf_ioevent_add(task)) == 0) {
-                            sf_client_sock_read,
-                            SF_CTX->net_buffer_cfg.network_timeout)) == 0)
-            {
                 result = SF_CTX->callbacks.deal_task(task, SF_NIO_STAGE_SEND);
             }
             break;
@@ -416,6 +544,7 @@ int sf_nio_notify(struct fast_task_info *task, const int stage)
         }
     }
 
+    sf_hold_task(task);  //since 1.2.11
     PTHREAD_MUTEX_LOCK(&task->thread_data->waiting_queue.lock);
     task->notify_next = NULL;
     if (task->thread_data->waiting_queue.tail == NULL) {
@@ -435,7 +564,7 @@ int sf_nio_notify(struct fast_task_info *task, const int stage)
         {
             result = errno != 0 ? errno : EIO;
             logError("file: "__FILE__", line: %d, "
-                    "write eventfd %d fail, errno: %d, error info: %s",
+                    "write to fd %d fail, errno: %d, error info: %s",
                     __LINE__, FC_NOTIFY_WRITE_FD(task->thread_data),
                     result, STRERROR(result));
             return result;
@@ -463,26 +592,32 @@ static inline void deal_notified_task(struct fast_task_info *task,
     }
 }
 
-void sf_recv_notify_read(int sock, short event, void *arg)
+void sf_recv_notify_read(int fd, const int event, void *arg)
 {
     int64_t n;
     int stage;
-    struct nio_thread_data *thread_data;
+    struct ioevent_notify_entry *notify_entry;
     struct fast_task_info *task;
     struct fast_task_info *current;
 
-    thread_data = ((struct ioevent_notify_entry *)arg)->thread_data;
+    notify_entry = (struct ioevent_notify_entry *)arg;
-    if (read(sock, &n, sizeof(n)) < 0) {
+    if (read(fd, &n, sizeof(n)) < 0) {
+#if IOEVENT_USE_URING
+        if (errno == EAGAIN) {
+            return;
+        }
+#endif
+
         logWarning("file: "__FILE__", line: %d, "
-                "read from eventfd %d fail, errno: %d, error info: %s",
+                "read from fd %d fail, errno: %d, error info: %s",
-                __LINE__, sock, errno, STRERROR(errno));
+                __LINE__, fd, errno, STRERROR(errno));
     }
 
-    PTHREAD_MUTEX_LOCK(&thread_data->waiting_queue.lock);
+    PTHREAD_MUTEX_LOCK(&notify_entry->thread_data->waiting_queue.lock);
-    current = thread_data->waiting_queue.head;
+    current = notify_entry->thread_data->waiting_queue.head;
-    thread_data->waiting_queue.head = NULL;
+    notify_entry->thread_data->waiting_queue.head = NULL;
-    thread_data->waiting_queue.tail = NULL;
+    notify_entry->thread_data->waiting_queue.tail = NULL;
-    PTHREAD_MUTEX_UNLOCK(&thread_data->waiting_queue.lock);
+    PTHREAD_MUTEX_UNLOCK(&notify_entry->thread_data->waiting_queue.lock);
 
     while (current != NULL) {
         task = current;
@@ -499,6 +634,7 @@ void sf_recv_notify_read(int sock, short event, void *arg)
             __sync_bool_compare_and_swap(&task->nio_stages.notify,
                     stage, SF_NIO_STAGE_NONE);
         }
+        sf_release_task(task);  //since 1.2.11
     }
 }
 
@@ -508,16 +644,31 @@ int sf_send_add_event(struct fast_task_info *task)
     if (task->send.ptr->length > 0) {
         /* direct send */
         task->nio_stages.current = SF_NIO_STAGE_SEND;
-        if (sf_client_sock_write(task->event.fd, IOEVENT_WRITE, task) < 0) {
+#if IOEVENT_USE_URING
-            return errno != 0 ? errno : EIO;
+        if (SF_CTX->use_io_uring) {
+            if (task->event.callback != (IOEventCallback)sf_client_sock_write) {
+                task->event.callback = (IOEventCallback)sf_client_sock_write;
+            }
+            if (SF_CTX->use_send_zc) {
+                return uring_prep_first_send_zc(task);
+            } else {
+                return uring_prep_first_send(task);
+            }
+        } else {
+#endif
+            if (sf_client_sock_write(task->event.fd, IOEVENT_WRITE, task) < 0) {
+                return errno != 0 ? errno : EIO;
+            }
+#if IOEVENT_USE_URING
         }
+#endif
     }
 
     return 0;
 }
 
 static inline int check_task(struct fast_task_info *task,
-        const short event, const int expect_stage)
+        const int event, const int expect_stage)
 {
     if (task->canceled) {
         return ENOTCONN;
@@ -537,6 +688,18 @@ static inline int check_task(struct fast_task_info *task,
         return 0;
     }
 
+#if IOEVENT_USE_URING
+    if (SF_CTX->use_io_uring) {
+        logWarning("file: "__FILE__", line: %d, "
+                "client ip: %s, event: %d, expect stage: %d, "
+                "but current stage: %d, close connection",
+                __LINE__, task->client_ip, event,
+                expect_stage, task->nio_stages.current);
+        ioevent_add_to_deleted_list(task);
+        return -1;
+    }
+#endif
+
     if (task->handler->comm_type == fc_comm_type_sock) {
         if (tcp_socket_connected(task->event.fd)) {
             return EAGAIN;
@@ -553,28 +716,69 @@ static inline int check_task(struct fast_task_info *task,
     }
 }
 
+#if IOEVENT_USE_URING
+static inline int prepare_next_send(struct fast_task_info *task)
+{
+    if (SF_CTX->use_send_zc) {
+        return uring_prep_next_send_zc(task);
+    } else {
+        return uring_prep_next_send(task);
+    }
+}
+#endif
+
 ssize_t sf_socket_send_data(struct fast_task_info *task,
         SFCommAction *action, bool *send_done)
 {
     int bytes;
+    int result;
 
-    if (task->iovec_array.iovs != NULL) {
+#if IOEVENT_USE_URING
-        bytes = writev(task->event.fd, task->iovec_array.iovs,
+    if (SF_CTX->use_io_uring) {
-                FC_MIN(task->iovec_array.count, IOV_MAX));
+        bytes = task->event.res;
     } else {
-        bytes = write(task->event.fd, task->send.ptr->data +
+#endif
-                task->send.ptr->offset, task->send.ptr->length -
+        if (task->iovec_array.iovs != NULL) {
-                task->send.ptr->offset);
+            bytes = writev(task->event.fd, task->iovec_array.iovs,
+                    FC_MIN(task->iovec_array.count, IOV_MAX));
+        } else {
+            bytes = write(task->event.fd, task->send.ptr->data +
+                    task->send.ptr->offset, task->send.ptr->length -
+                    task->send.ptr->offset);
+        }
+#if IOEVENT_USE_URING
     }
+#endif
+
     if (bytes < 0) {
-        if (errno == EAGAIN || errno == EWOULDBLOCK)
+#if IOEVENT_USE_URING
-        {
+        if (SF_CTX->use_io_uring) {
-            if (set_write_event(task) != 0) {
+            result = -bytes;
-                return -1;
+        } else {
+#endif
+            result = errno;
+#if IOEVENT_USE_URING
+        }
+#endif
+        if (result == EAGAIN || result == EWOULDBLOCK) {
+#if IOEVENT_USE_URING
+            if (SF_CTX->use_io_uring) {
+                if (prepare_next_send(task) != 0) {
+                    return -1;
+                }
+            } else {
+#endif
+                if (set_write_event(task) != 0) {
+                    return -1;
+                }
+#if IOEVENT_USE_URING
             }
+#endif
+
             *action = sf_comm_action_break;
             return 0;
-        } else if (errno == EINTR) {  //should retry
+        } else if (result == EINTR && !SF_CTX->use_io_uring) {
+            /* should try again */
             logDebug("file: "__FILE__", line: %d, "
                     "client ip: %s, ignore interupt signal",
                     __LINE__, task->client_ip);
@@ -585,30 +789,40 @@ ssize_t sf_socket_send_data(struct fast_task_info *task,
                     "client ip: %s, send fail, task offset: %d, length: %d, "
                     "errno: %d, error info: %s", __LINE__, task->client_ip,
                     task->send.ptr->offset, task->send.ptr->length,
-                    errno, strerror(errno));
+                    result, strerror(result));
             return -1;
         }
     } else if (bytes == 0) {
         logWarning("file: "__FILE__", line: %d, "
-                "client ip: %s, sock: %d, task length: %d, offset: %d, "
+                "client ip: %s, task length: %d, offset: %d, "
                 "send failed, connection disconnected", __LINE__,
-                task->client_ip, task->event.fd, task->send.ptr->length,
+                task->client_ip, task->send.ptr->length,
                 task->send.ptr->offset);
         return -1;
     }
 
     task->send.ptr->offset += bytes;
     if (task->send.ptr->offset >= task->send.ptr->length) {
-        if (task->send.ptr != task->recv.ptr) {  //double buffers
+#if IOEVENT_USE_URING
-            task->send.ptr->offset = 0;
+        if (SF_CTX->use_io_uring && FC_URING_IS_SEND_ZC(task) &&
-            task->send.ptr->length = 0;
+                task->thread_data->ev_puller.send_zc_done_notify)
-        }
+        {
-        *action = sf_comm_action_finish;
+            *action = sf_comm_action_break;
-        *send_done = true;
+            *send_done = false;
-    } else {
+        } else {
-        *action = sf_comm_action_continue;
+#endif
-        *send_done = false;
+            if (task->send.ptr != task->recv.ptr) {  //double buffers
+                task->send.ptr->offset = 0;
+                task->send.ptr->length = 0;
+            }
 
+            *action = sf_comm_action_finish;
+            *send_done = true;
+#if IOEVENT_USE_URING
+        }
+#endif
+
+    } else {
         /* set next writev iovec array */
         if (task->iovec_array.iovs != NULL) {
             struct iovec *iov;
@@ -637,6 +851,25 @@ ssize_t sf_socket_send_data(struct fast_task_info *task,
             task->iovec_array.iovs = iov;
             task->iovec_array.count = end - iov;
         }
+
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring) {
+            if (!(FC_URING_IS_SEND_ZC(task) && task->thread_data->
+                        ev_puller.send_zc_done_notify))
+            {
+                if (prepare_next_send(task) != 0) {
+                    return -1;
+                }
+            }
+            *action = sf_comm_action_break;
+        } else {
+#endif
+            *action = sf_comm_action_continue;
+#if IOEVENT_USE_URING
+        }
+#endif
+
+        *send_done = false;
     }
 
     return bytes;
@@ -646,30 +879,56 @@ ssize_t sf_socket_recv_data(struct fast_task_info *task,
         const bool call_post_recv, SFCommAction *action)
 {
     int bytes;
+    int result;
     int recv_bytes;
     bool new_alloc;
 
-    if (task->recv.ptr->length == 0) { //recv header
+#if IOEVENT_USE_URING
-        recv_bytes = SF_CTX->header_size - task->recv.ptr->offset;
+    if (SF_CTX->use_io_uring) {
-        bytes = read(task->event.fd, task->recv.ptr->data +
+        bytes = task->event.res;
-                task->recv.ptr->offset, recv_bytes);
     } else {
-        recv_bytes = task->recv.ptr->length - task->recv.ptr->offset;
+#endif
-        if (task->recv_body == NULL) {
+        if (task->recv.ptr->length == 0) { //recv header
+            recv_bytes = SF_CTX->header_size - task->recv.ptr->offset;
             bytes = read(task->event.fd, task->recv.ptr->data +
                     task->recv.ptr->offset, recv_bytes);
         } else {
-            bytes = read(task->event.fd, task->recv_body +
+            recv_bytes = task->recv.ptr->length - task->recv.ptr->offset;
-                    (task->recv.ptr->offset - SF_CTX->
+            if (task->recv_body == NULL) {
-                     header_size), recv_bytes);
+                bytes = read(task->event.fd, task->recv.ptr->data +
+                        task->recv.ptr->offset, recv_bytes);
+            } else {
+                bytes = read(task->event.fd, task->recv_body +
+                        (task->recv.ptr->offset - SF_CTX->
+                         header_size), recv_bytes);
+            }
         }
+#if IOEVENT_USE_URING
     }
+#endif
 
     if (bytes < 0) {
-        if (errno == EAGAIN || errno == EWOULDBLOCK) {
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring) {
+            result = -bytes;
+        } else {
+#endif
+            result = errno;
+#if IOEVENT_USE_URING
+        }
+#endif
+        if (result == EAGAIN || result == EWOULDBLOCK) {
+#if IOEVENT_USE_URING
+            if (SF_CTX->use_io_uring) {
+                if (prepare_next_recv(task) != 0) {
+                    return -1;
+                }
+            }
+#endif
             *action = sf_comm_action_break;
             return 0;
-        } else if (errno == EINTR) {  //should retry
+        } else if (result == EINTR && !SF_CTX->use_io_uring) {
+            /* should try again */
             logDebug("file: "__FILE__", line: %d, "
                     "client ip: %s, ignore interupt signal",
                     __LINE__, task->client_ip);
@@ -680,7 +939,7 @@ ssize_t sf_socket_recv_data(struct fast_task_info *task,
                     "client ip: %s, recv fail, "
                     "errno: %d, error info: %s",
                     __LINE__, task->client_ip,
-                    errno, strerror(errno));
+                    result, strerror(result));
             return -1;
         }
     } else if (bytes == 0) {
@@ -712,7 +971,18 @@ ssize_t sf_socket_recv_data(struct fast_task_info *task,
     task->recv.ptr->offset += bytes;
     if (task->recv.ptr->length == 0) { //pkg header
         if (task->recv.ptr->offset < SF_CTX->header_size) {
-            *action = sf_comm_action_continue;
+#if IOEVENT_USE_URING
+            if (SF_CTX->use_io_uring) {
+                if (prepare_next_recv(task) != 0) {
+                    return -1;
+                }
+                *action = sf_comm_action_break;
+            } else {
+#endif
+                *action = sf_comm_action_continue;
+#if IOEVENT_USE_URING
+            }
+#endif
             return bytes;
         }
 
@@ -765,7 +1035,18 @@ ssize_t sf_socket_recv_data(struct fast_task_info *task,
     if (task->recv.ptr->offset >= task->recv.ptr->length) { //recv done
         *action = sf_comm_action_finish;
     } else {
-        *action = sf_comm_action_continue;
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring) {
+            if (prepare_next_recv(task) != 0) {
+                return -1;
+            }
+            *action = sf_comm_action_break;
+        } else {
+#endif
+            *action = sf_comm_action_continue;
+#if IOEVENT_USE_URING
+        }
+#endif
     }
 
     return bytes;
@@ -841,8 +1122,7 @@ static int calc_iops_and_remove_polling(struct fast_task_info *task)
                     ioevent_set_timeout(&task->thread_data->ev_puller,
                             task->thread_data->timeout_ms);
                 }
-                result = sf_ioevent_add(task, (IOEventCallback)
+                result = sf_ioevent_add(task);
-                        sf_client_sock_read, SF_CTX->net_buffer_cfg.network_timeout);
 
                 format_ip_address(task->client_ip, formatted_ip);
                 logInfo("file: "__FILE__", line: %d, client: %s:%u, "
@@ -904,7 +1184,7 @@ int sf_rdma_busy_polling_callback(struct nio_thread_data *thread_data)
     return 0;
 }
 
-int sf_client_sock_read(int sock, short event, void *arg)
+static int sf_client_sock_read(int sock, const int event, void *arg)
 {
     int result;
     int bytes;
@@ -914,10 +1194,15 @@ int sf_client_sock_read(int sock, short event, void *arg)
 
     task = (struct fast_task_info *)arg;
     if ((result=check_task(task, event, SF_NIO_STAGE_RECV)) != 0) {
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring && event != IOEVENT_TIMEOUT) {
+            CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+        }
+#endif
         return result >= 0 ? 0 : -1;
     }
 
-    if (event & IOEVENT_TIMEOUT) {
+    if (event == IOEVENT_TIMEOUT) {
         if (task->recv.ptr->offset == 0 && task->req_count > 0) {
             if (SF_CTX->callbacks.task_timeout != NULL) {
                 if (SF_CTX->callbacks.task_timeout(task) != 0) {
@@ -929,7 +1214,8 @@ int sf_client_sock_read(int sock, short event, void *arg)
             task->event.timer.expires = g_current_time +
                 SF_CTX->net_buffer_cfg.network_timeout;
             fast_timer_add(&task->thread_data->timer,
-                &task->event.timer);
+                    &task->event.timer);
+            return 0;
         } else {
             if (task->recv.ptr->length > 0) {
                 logWarning("file: "__FILE__", line: %d, "
@@ -946,10 +1232,14 @@ int sf_client_sock_read(int sock, short event, void *arg)
             ioevent_add_to_deleted_list(task);
             return -1;
         }
-
-        return 0;
     }
 
+#if IOEVENT_USE_URING
+    if (SF_CTX->use_io_uring) {
+        CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+    }
+#endif
+
     total_read = 0;
     action = sf_comm_action_continue;
     while (1) {
@@ -996,23 +1286,72 @@ int sf_client_sock_read(int sock, short event, void *arg)
     return total_read;
 }
 
-int sf_client_sock_write(int sock, short event, void *arg)
+static int sock_write_done(struct fast_task_info *task,
+        const int length, const bool send_done)
+{
+    int next_stage;
+
+    release_iovec_buffer(task);
+    task->recv.ptr->offset = 0;
+    task->recv.ptr->length = 0;
+
+    if (SF_CTX->callbacks.send_done == NULL || !send_done) {
+        task->nio_stages.current = SF_NIO_STAGE_RECV;
+    } else {
+        if (SF_CTX->callbacks.send_done(task,
+                    length, &next_stage) != 0)
+        {
+            ioevent_add_to_deleted_list(task);
+            return -1;
+        }
+
+        if (task->nio_stages.current != next_stage) {
+            task->nio_stages.current = next_stage;
+        }
+    }
+
+#if IOEVENT_USE_URING
+    if (!SF_CTX->use_io_uring || task->nio_stages.
+            current == SF_NIO_STAGE_RECV)
+    {
+#endif
+
+        if (set_read_event(task) != 0) {
+            return -1;
+        }
+
+#if IOEVENT_USE_URING
+    }
+#endif
+
+    return 0;
+}
+
+static int sf_client_sock_write(int sock, const int event, void *arg)
 {
     int result;
     int bytes;
     int total_write;
     int length;
-    int next_stage;
     SFCommAction action;
     bool send_done;
     struct fast_task_info *task;
 
     task = (struct fast_task_info *)arg;
     if ((result=check_task(task, event, SF_NIO_STAGE_SEND)) != 0) {
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring && event != IOEVENT_TIMEOUT) {
+            if (event == IOEVENT_NOTIFY || !(FC_URING_IS_SEND_ZC(task) &&
+                        task->thread_data->ev_puller.send_zc_done_notify))
+            {
+                CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+            }
+        }
+#endif
         return result >= 0 ? 0 : -1;
     }
 
-    if (event & IOEVENT_TIMEOUT) {
+    if (event == IOEVENT_TIMEOUT) {
         logError("file: "__FILE__", line: %d, "
             "client ip: %s, send timeout. total length: %d, offset: %d, "
             "remain: %d", __LINE__, task->client_ip, task->send.ptr->length,
@@ -1023,6 +1362,42 @@ int sf_client_sock_write(int sock, short event, void *arg)
         return -1;
     }
 
+#if IOEVENT_USE_URING
+    if (event == IOEVENT_NOTIFY) {
+        if (!FC_URING_IS_SEND_ZC(task)) {
+            logWarning("file: "__FILE__", line: %d, "
+                    "unexpected io_uring notify!", __LINE__);
+            return -1;
+        }
+
+        FC_URING_OP_TYPE(task) = IORING_OP_NOP;
+        if (!task->canceled) {
+            if (task->send.ptr->offset >= task->send.ptr->length) {
+                length = task->send.ptr->length;
+                if (task->send.ptr != task->recv.ptr) {  //double buffers
+                    task->send.ptr->offset = 0;
+                    task->send.ptr->length = 0;
+                }
+
+                result = sock_write_done(task, length, true);
+            } else {
+                result = prepare_next_send(task);
+            }
+        }
+
+        sf_release_task(task);
+        return result == 0 ? 0 : -1;
+    }
+
+    if (SF_CTX->use_io_uring) {
+        if (!(FC_URING_IS_SEND_ZC(task) && task->thread_data->
+                    ev_puller.send_zc_done_notify))
+        {
+            CLEAR_OP_TYPE_AND_RELEASE_TASK(task);
+        }
+    }
+#endif
+
     total_write = 0;
     length = task->send.ptr->length;
     action = sf_comm_action_continue;
@@ -1038,27 +1413,9 @@ int sf_client_sock_write(int sock, short event, void *arg)
 
         total_write += bytes;
         if (action == sf_comm_action_finish) {
-            release_iovec_buffer(task);
+            if (sock_write_done(task, length, send_done) != 0) {
-            task->recv.ptr->offset = 0;
-            task->recv.ptr->length = 0;
-            if (set_read_event(task) != 0) {
                 return -1;
             }
-
-            if (SF_CTX->callbacks.send_done == NULL || !send_done) {
-                task->nio_stages.current = SF_NIO_STAGE_RECV;
-            } else {
-                if (SF_CTX->callbacks.send_done(task,
-                            length, &next_stage) != 0)
-                {
-                    return -1;
-                }
-
-                if (task->nio_stages.current != next_stage) {
-                    task->nio_stages.current = next_stage;
-                }
-            }
-
             break;
         } else if (action == sf_comm_action_break) {
             break;
@@ -1067,3 +1424,8 @@ int sf_client_sock_write(int sock, short event, void *arg)
 
     return total_write;
 }
+
+bool sf_client_sock_in_read_stage(struct fast_task_info *task)
+{
+    return (task->event.callback == (IOEventCallback)sf_client_sock_read);
+}

src/sf_nio.h

@@ -68,15 +68,6 @@ static inline void sf_set_connect_done_callback_ex(SFContext *sf_context,
     sf_set_connect_done_callback_ex(&g_sf_context, done_callback)
 
 
-static inline void sf_set_remove_from_ready_list_ex(
-        SFContext *sf_context, const bool enabled)
-{
-    sf_context->remove_from_ready_list = enabled;
-}
-
-#define sf_set_remove_from_ready_list(enabled) \
-    sf_set_remove_from_ready_list_ex(&g_sf_context, enabled);
-
 static inline TaskCleanUpCallback sf_get_task_cleanup_callback_ex(
         SFContext *sf_context)
 {
@@ -99,10 +90,9 @@ static inline void sf_nio_reset_task_length(struct fast_task_info *task)
     }
 }
 
-void sf_recv_notify_read(int sock, short event, void *arg);
+void sf_socket_close_connection(struct fast_task_info *task);
+void sf_recv_notify_read(int sock, const int event, void *arg);
 int sf_send_add_event(struct fast_task_info *task);
-int sf_client_sock_write(int sock, short event, void *arg);
-int sf_client_sock_read(int sock, short event, void *arg);
 
 void sf_task_finish_clean_up(struct fast_task_info *task);
 
@@ -158,11 +148,6 @@ static inline int sf_nio_forward_request(struct fast_task_info *task,
     return sf_nio_notify(task, SF_NIO_STAGE_FORWARDED);
 }
 
-static inline bool sf_client_sock_in_read_stage(struct fast_task_info *task)
-{
-    return (task->event.callback == (IOEventCallback)sf_client_sock_read);
-}
-
 static inline void sf_nio_add_to_deleted_list(struct nio_thread_data
         *thread_data, struct fast_task_info *task)
 {
@@ -173,6 +158,8 @@ static inline void sf_nio_add_to_deleted_list(struct nio_thread_data
     }
 }
 
+bool sf_client_sock_in_read_stage(struct fast_task_info *task);
+
 #ifdef __cplusplus
 }
 #endif

src/sf_service.c

@@ -33,9 +33,8 @@
 #include "fastcommon/sched_thread.h"
 #include "fastcommon/ioevent_loop.h"
 #include "fastcommon/fc_memory.h"
-#include "sf_nio.h"
+#include "sf_proto.h"
 #include "sf_util.h"
-#include "sf_global.h"
 #include "sf_service.h"
 
 #if defined(OS_LINUX)
@@ -61,12 +60,14 @@ struct worker_thread_context {
 static void *worker_thread_entrance(void *arg);
 
 static int sf_init_free_queue(SFContext *sf_context, const char *name,
-        const bool double_buffers, const int task_padding_size,
+        const bool double_buffers, const bool need_shrink_task_buffer,
-        const int task_arg_size, TaskInitCallback init_callback,
+        const int task_padding_size, const int task_arg_size,
-        void *init_arg)
+        TaskInitCallback init_callback, void *init_arg)
 {
     int result;
+    int buffer_size;
     int m;
+    int max_m;
     int alloc_conn_once;
 
     if ((result=set_rand_seed()) != 0) {
@@ -75,17 +76,26 @@ static int sf_init_free_queue(SFContext *sf_context, const char *name,
         return result;
     }
 
-    m = sf_context->net_buffer_cfg.min_buff_size / (64 * 1024);
+    if (strcmp(name, "cluster") == 0 || strcmp(name, "replica") == 0) {
+        buffer_size = FC_MAX(4 * 1024 * 1024, sf_context->
+                net_buffer_cfg.max_buff_size);
+        max_m = 64;
+    } else {
+        buffer_size = sf_context->net_buffer_cfg.min_buff_size;
+        max_m = 16;
+    }
+    m = buffer_size / (64 * 1024);
     if (m == 0) {
         m = 1;
-    } else if (m > 16) {
+    } else if (m > max_m) {
-        m = 16;
+        m = max_m;
     }
     alloc_conn_once = 256 / m;
+
     return free_queue_init_ex2(&sf_context->free_queue, name, double_buffers,
-            sf_context->net_buffer_cfg.max_connections, alloc_conn_once,
+            need_shrink_task_buffer, sf_context->net_buffer_cfg.max_connections,
-            sf_context->net_buffer_cfg.min_buff_size, sf_context->
+            alloc_conn_once, sf_context->net_buffer_cfg.min_buff_size,
-            net_buffer_cfg.max_buff_size, task_padding_size,
+            sf_context->net_buffer_cfg.max_buff_size, task_padding_size,
             task_arg_size, init_callback, init_arg);
 }
 
@@ -101,12 +111,14 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
         sf_recv_timeout_callback timeout_callback, const int net_timeout_ms,
         const int proto_header_size, const int task_padding_size,
         const int task_arg_size, const bool double_buffers,
-        const bool explicit_post_recv, TaskInitCallback init_callback,
+        const bool need_shrink_task_buffer, const bool explicit_post_recv,
-        void *init_arg, sf_release_buffer_callback release_buffer_callback)
+        TaskInitCallback init_callback, void *init_arg,
+        sf_release_buffer_callback release_buffer_callback)
 {
     int result;
     int bytes;
     int extra_events;
+    int max_entries;
     int i;
     struct worker_thread_context *thread_contexts;
     struct worker_thread_context *thread_ctx;
@@ -132,8 +144,8 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
     }
 
     if ((result=sf_init_free_queue(sf_context, name, double_buffers,
-                    task_padding_size, task_arg_size, init_callback,
+                    need_shrink_task_buffer, task_padding_size,
-                    init_arg)) != 0)
+                    task_arg_size, init_callback, init_arg)) != 0)
     {
         return result;
     }
@@ -161,7 +173,11 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
 
     if (SF_G_EPOLL_EDGE_TRIGGER) {
 #ifdef OS_LINUX
+#if IOEVENT_USE_EPOLL
         extra_events = EPOLLET;
+#else
+        extra_events = 0;
+#endif
 #elif defined(OS_FREEBSD)
         extra_events = EV_CLEAR;
 #else
@@ -171,6 +187,40 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
         extra_events = 0;
     }
 
+    max_entries = (sf_context->net_buffer_cfg.max_connections +
+            sf_context->work_threads - 1) / sf_context->work_threads;
+    if (strcmp(sf_context->name, "cluster") == 0 ||
+            strcmp(sf_context->name, "replica") == 0)
+    {
+        if (max_entries < 1024) {
+            max_entries += 8;
+        } else {
+            max_entries = 1024;
+        }
+    } else {
+        if (max_entries < 4 * 1024) {
+            max_entries = max_entries * 2;
+        } else if (max_entries < 8 * 1024) {
+            max_entries = (max_entries * 3) / 2;
+        } else if (max_entries < 16 * 1024) {
+            max_entries = (max_entries * 5) / 4;
+        } else if (max_entries < 32 * 1024) {
+            max_entries = (max_entries * 6) / 5;
+        } else if (max_entries < 64 * 1024) {
+            max_entries = (max_entries * 11) / 10;
+        } else if (max_entries < 128 * 1024) {
+            max_entries = (max_entries * 21) / 20;
+        }
+
+#if IOEVENT_USE_URING
+        if (sf_context->use_io_uring) {
+            if (max_entries > 32 * 1024) {
+                max_entries = 32 * 1024;
+            }
+        }
+#endif
+    }
+
     g_current_time = time(NULL);
     sf_context->thread_count = 0;
     data_end = sf_context->thread_data + sf_context->work_threads;
@@ -195,18 +245,43 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
             thread_data->arg = NULL;
         }
 
-        if (ioevent_init(&thread_data->ev_puller, 2 + sf_context->
+        if ((result=ioevent_init(&thread_data->ev_puller, sf_context->
-                    net_buffer_cfg.max_connections, net_timeout_ms,
+                        name, sf_context->use_io_uring, max_entries,
-                    extra_events) != 0)
+                        net_timeout_ms, extra_events)) != 0)
         {
-            result  = errno != 0 ? errno : ENOMEM;
+            char prompt[256];
+#if IOEVENT_USE_URING
+            if (sf_context->use_io_uring) {
+                if (result == EPERM) {
+                    strcpy(prompt, " make sure kernel."
+                            "io_uring_disabled set to 0");
+                } else if (result == EINVAL) {
+                    sprintf(prompt, " maybe max_connections: %d is too large"
+                            " or [%s]'s work_threads: %d is too small",
+                            sf_context->net_buffer_cfg.max_connections,
+                            sf_context->name, sf_context->work_threads);
+                } else {
+                    *prompt = '\0';
+                }
+            } else {
+#endif
+                *prompt = '\0';
+#if IOEVENT_USE_URING
+            }
+#endif
+
             logError("file: "__FILE__", line: %d, "
-                "ioevent_init fail, "
+                "ioevent_init fail, errno: %d, error info: %s.%s"
-                "errno: %d, error info: %s",
+                , __LINE__, result, strerror(result), prompt);
-                __LINE__, result, strerror(result));
             return result;
         }
 
+#if IOEVENT_USE_URING
+        if (sf_context->use_io_uring && send_done_callback != NULL) {
+            ioevent_set_send_zc_done_notify(&thread_data->ev_puller, true);
+        }
+#endif
+
         result = fast_timer_init(&thread_data->timer, 2 * sf_context->
                 net_buffer_cfg.network_timeout, g_current_time);
         if (result != 0) {
@@ -487,7 +562,9 @@ struct fast_task_info *sf_socket_accept_connection(SFListener *listener)
     }
     FC_SET_CLOEXEC(incomesock);
 
-    if ((task=sf_alloc_init_task(listener->handler, incomesock)) == NULL) {
+    if ((task=sf_alloc_init_server_task(listener->handler,
+                    incomesock)) == NULL)
+    {
         close(incomesock);
         return NULL;
     }
@@ -498,12 +575,6 @@ struct fast_task_info *sf_socket_accept_connection(SFListener *listener)
     return task;
 }
 
-void sf_socket_close_connection(struct fast_task_info *task)
-{
-    close(task->event.fd);
-    task->event.fd = -1;
-}
-
 void sf_socket_close_ex(SFContext *sf_context)
 {
     int i;
@@ -782,15 +853,13 @@ int sf_setup_signal_handler()
     return 0;
 }
 
-#define LOG_SCHEDULE_ENTRIES_COUNT 3
-
 int sf_startup_schedule(pthread_t *schedule_tid)
 {
     ScheduleArray scheduleArray;
-    ScheduleEntry scheduleEntries[LOG_SCHEDULE_ENTRIES_COUNT];
+    ScheduleEntry scheduleEntries[SF_LOG_SCHEDULE_ENTRIES_COUNT];
 
     scheduleArray.entries = scheduleEntries;
-    sf_setup_schedule(&g_log_context, &g_sf_global_vars.error_log,
+    sf_logger_setup_schedule(&g_log_context, &g_sf_global_vars.error_log,
             &scheduleArray);
     return sched_start(&scheduleArray, schedule_tid,
             g_sf_global_vars.thread_stack_size, (bool * volatile)
@@ -801,7 +870,7 @@ int sf_add_slow_log_schedule(SFSlowLogContext *slowlog_ctx)
 {
     int result;
     ScheduleArray scheduleArray;
-    ScheduleEntry scheduleEntries[LOG_SCHEDULE_ENTRIES_COUNT];
+    ScheduleEntry scheduleEntries[SF_LOG_SCHEDULE_ENTRIES_COUNT];
 
     if (!slowlog_ctx->cfg.enabled) {
         return 0;
@@ -814,8 +883,8 @@ int sf_add_slow_log_schedule(SFSlowLogContext *slowlog_ctx)
     }
 
     scheduleArray.entries = scheduleEntries;
-    sf_setup_schedule(&slowlog_ctx->ctx, &slowlog_ctx->cfg.log_cfg,
+    sf_logger_setup_schedule(&slowlog_ctx->ctx, &slowlog_ctx->
-            &scheduleArray);
+            cfg.log_cfg, &scheduleArray);
     return sched_add_entries(&scheduleArray);
 }
 

src/sf_service.h

@@ -25,6 +25,9 @@
 #include "fastcommon/ioevent.h"
 #include "fastcommon/fast_task_queue.h"
 #include "sf_types.h"
+#include "sf_proto.h"
+#include "sf_global.h"
+#include "sf_nio.h"
 
 typedef void* (*sf_alloc_thread_extra_data_callback)(const int thread_index);
 typedef void (*sf_sig_quit_handler)(int sig);
@@ -45,8 +48,9 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
         sf_recv_timeout_callback timeout_callback, const int net_timeout_ms,
         const int proto_header_size, const int task_padding_size,
         const int task_arg_size, const bool double_buffers,
-        const bool explicit_post_recv, TaskInitCallback init_callback,
+        const bool need_shrink_task_buffer, const bool explicit_post_recv,
-        void *init_arg, sf_release_buffer_callback release_buffer_callback);
+        TaskInitCallback init_callback, void *init_arg,
+        sf_release_buffer_callback release_buffer_callback);
 
 #define sf_service_init_ex(sf_context, name, alloc_thread_extra_data_callback,\
         thread_loop_callback, accept_done_callback, set_body_length_func,   \
@@ -56,7 +60,7 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
         thread_loop_callback, accept_done_callback, set_body_length_func,     \
         NULL, send_done_callback, deal_func, task_cleanup_func, \
         timeout_callback, net_timeout_ms, proto_header_size, \
-        0, task_arg_size, false, false, NULL, NULL, NULL)
+        0, task_arg_size, false, true, false, NULL, NULL, NULL)
 
 #define sf_service_init(name, alloc_thread_extra_data_callback, \
         thread_loop_callback, accept_done_callback, set_body_length_func,   \
@@ -65,8 +69,8 @@ int sf_service_init_ex2(SFContext *sf_context, const char *name,
     sf_service_init_ex2(&g_sf_context, name, alloc_thread_extra_data_callback, \
         thread_loop_callback, accept_done_callback, set_body_length_func, NULL,\
         send_done_callback, deal_func, task_cleanup_func, timeout_callback, \
-        net_timeout_ms, proto_header_size, 0, task_arg_size, false, false,  \
+        net_timeout_ms, proto_header_size, 0, task_arg_size, false, true,   \
-        NULL, NULL, NULL)
+        false, NULL, NULL, NULL)
 
 int sf_service_destroy_ex(SFContext *sf_context);
 
@@ -108,8 +112,6 @@ int sf_socket_create_server(SFListener *listener,
 void sf_socket_close_server(SFListener *listener);
 struct fast_task_info *sf_socket_accept_connection(SFListener *listener);
 
-void sf_socket_close_connection(struct fast_task_info *task);
-
 int sf_socket_server_ex(SFContext *sf_context);
 #define sf_socket_server() sf_socket_server_ex(&g_sf_context)
 
@@ -163,6 +165,11 @@ static inline struct fast_task_info *sf_alloc_init_task_ex(
         return NULL;
     }
 
+    if (task->shrinked) {
+        task->shrinked = false;
+        sf_proto_init_task_magic(task);
+    }
+
     __sync_add_and_fetch(&task->reffer_count, reffer_count);
     __sync_bool_compare_and_swap(&task->canceled, 1, 0);
     task->handler = handler;
@@ -170,12 +177,42 @@ static inline struct fast_task_info *sf_alloc_init_task_ex(
     return task;
 }
 
-#define sf_hold_task_ex(task, inc_count) __sync_add_and_fetch( \
+#define sf_hold_task_ex(task, inc_count)  fc_hold_task_ex(task, inc_count)
-        &task->reffer_count, inc_count)
+#define sf_hold_task(task)  fc_hold_task(task)
-#define sf_hold_task(task)  sf_hold_task_ex(task, 1)
 
 #define sf_alloc_init_task(handler, fd) sf_alloc_init_task_ex(handler, fd, 1)
 
+static inline struct fast_task_info *sf_alloc_init_server_task(
+        SFNetworkHandler *handler, const int fd)
+{
+    const int reffer_count = 1;
+    struct fast_task_info *task;
+
+    if ((task=sf_alloc_init_task_ex(handler, fd, reffer_count)) != NULL) {
+#if IOEVENT_USE_URING
+        FC_URING_IS_CLIENT(task) = false;
+#endif
+    }
+
+    return task;
+}
+
+static inline struct fast_task_info *sf_alloc_init_client_task(
+        SFNetworkHandler *handler)
+{
+    const int fd = -1;
+    const int reffer_count = 1;
+    struct fast_task_info *task;
+
+    if ((task=sf_alloc_init_task_ex(handler, fd, reffer_count)) != NULL) {
+#if IOEVENT_USE_URING
+        FC_URING_IS_CLIENT(task) = true;
+#endif
+    }
+
+    return task;
+}
+
 static inline void sf_release_task(struct fast_task_info *task)
 {
     if (__sync_sub_and_fetch(&task->reffer_count, 1) == 0) {
@@ -187,6 +224,15 @@ static inline void sf_release_task(struct fast_task_info *task)
                 "used: %d, freed: %d", __LINE__, task,
                 alloc_count, alloc_count - free_count, free_count);
                 */
+
+#if IOEVENT_USE_URING
+        if (SF_CTX->use_io_uring) {
+            task->handler->close_connection(task);
+            __sync_fetch_and_sub(&g_sf_global_vars.
+                    connection_stat.current_count, 1);
+        }
+#endif
+
         free_queue_push(task);
     }
 }
@@ -262,6 +308,19 @@ static inline SFNetworkHandler *sf_get_rdma_network_handler3(
     return sf_get_rdma_network_handler(sf_context3);
 }
 
+static inline bool sf_get_double_buffers_flag(FCServerGroupInfo *server_group)
+{
+    if (server_group->comm_type == fc_comm_type_sock) {
+#if IOEVENT_USE_URING
+        return true;
+#else
+        return false;
+#endif
+    } else {  //RDMA
+        return true;
+    }
+}
+
 #ifdef __cplusplus
 }
 #endif

src/sf_types.h

@@ -115,6 +115,7 @@ typedef struct sf_listener {
 
 struct sf_context;
 struct sf_address_family_handler;
+
 typedef struct sf_network_handler {
     bool enabled;
     bool explicit_post_recv;
@@ -179,8 +180,9 @@ typedef struct sf_context {
     struct nio_thread_data *thread_data;
     volatile int thread_count;
 
-    //int rdma_port_offset;
+    bool is_client;    //since v1.2.5
-    bool is_client;   //since v1.2.5
+    bool use_io_uring; //since v1.2.9
+    bool use_send_zc;  //since v1.2.9
     SFAddressFamily address_family;
     SFAddressFamilyHandler handlers[SF_ADDRESS_FAMILY_COUNT];
 
@@ -190,7 +192,6 @@ typedef struct sf_context {
     int work_threads;
 
     int header_size;
-    bool remove_from_ready_list;
     bool realloc_task_buffer;
     bool connect_need_log;  //for client connect
     FCSmartPollingConfig smart_polling;

src/sf_util.c

@@ -281,7 +281,7 @@ int sf_logger_init(LogContext *pContext, const char *filename_prefix)
     return 0;
 }
 
-ScheduleEntry *sf_logger_set_schedule_entry(struct log_context *pContext,
+ScheduleEntry *sf_logger_set_schedule_entries(struct log_context *pContext,
         SFLogConfig *log_cfg, ScheduleEntry *pScheduleEntry)
 {
     INIT_SCHEDULE_ENTRY(*pScheduleEntry, sched_generate_next_id(),

src/sf_util.h

@@ -96,14 +96,14 @@ void sf_parse_cmd_option_bool(int argc, char *argv[],
 
 int sf_logger_init(LogContext *pContext, const char *filename_prefix);
 
-ScheduleEntry *sf_logger_set_schedule_entry(struct log_context *pContext,
+ScheduleEntry *sf_logger_set_schedule_entries(struct log_context *pContext,
         SFLogConfig *log_cfg, ScheduleEntry *pScheduleEntry);
 
-static inline void sf_setup_schedule(struct log_context *pContext,
+static inline void sf_logger_setup_schedule(struct log_context *pContext,
         SFLogConfig *log_cfg, ScheduleArray *scheduleArray)
 {
     ScheduleEntry *scheduleEntry;
-    scheduleEntry = sf_logger_set_schedule_entry(pContext,
+    scheduleEntry = sf_logger_set_schedule_entries(pContext,
             log_cfg, scheduleArray->entries);
     scheduleArray->count = scheduleEntry - scheduleArray->entries;
 }