summary refs log tree commit diff stats
path: root/net/colo-compare.c
diff options
context:
space:
mode:
Diffstat (limited to 'net/colo-compare.c')
-rw-r--r--net/colo-compare.c277
1 files changed, 207 insertions, 70 deletions
diff --git a/net/colo-compare.c b/net/colo-compare.c
index c07e7c1c09..f15779dedc 100644
--- a/net/colo-compare.c
+++ b/net/colo-compare.c
@@ -32,6 +32,9 @@
 #include "migration/migration.h"
 #include "util.h"
 
+#include "block/aio-wait.h"
+#include "qemu/coroutine.h"
+
 #define TYPE_COLO_COMPARE "colo-compare"
 #define COLO_COMPARE(obj) \
     OBJECT_CHECK(CompareState, (obj), TYPE_COLO_COMPARE)
@@ -51,6 +54,8 @@ static NotifierList colo_compare_notifiers =
 #define REGULAR_PACKET_CHECK_MS 3000
 #define DEFAULT_TIME_OUT_MS 3000
 
+static QemuMutex colo_compare_mutex;
+static bool colo_compare_active;
 static QemuMutex event_mtx;
 static QemuCond event_complete_cond;
 static int event_unhandled_count;
@@ -77,6 +82,23 @@ static int event_unhandled_count;
  *                    |packet  |  |packet  +    |packet  | |packet  +
  *                    +--------+  +--------+    +--------+ +--------+
  */
+
+typedef struct SendCo {
+    Coroutine *co;
+    struct CompareState *s;
+    CharBackend *chr;
+    GQueue send_list;
+    bool notify_remote_frame;
+    bool done;
+    int ret;
+} SendCo;
+
+typedef struct SendEntry {
+    uint32_t size;
+    uint32_t vnet_hdr_len;
+    uint8_t *buf;
+} SendEntry;
+
 typedef struct CompareState {
     Object parent;
 
@@ -91,6 +113,8 @@ typedef struct CompareState {
     SocketReadState pri_rs;
     SocketReadState sec_rs;
     SocketReadState notify_rs;
+    SendCo out_sendco;
+    SendCo notify_sendco;
     bool vnet_hdr;
     uint32_t compare_timeout;
     uint32_t expired_scan_cycle;
@@ -122,12 +146,17 @@ enum {
     SECONDARY_IN,
 };
 
+static const char *colo_mode[] = {
+    [PRIMARY_IN] = "primary",
+    [SECONDARY_IN] = "secondary",
+};
 
 static int compare_chr_send(CompareState *s,
-                            const uint8_t *buf,
+                            uint8_t *buf,
                             uint32_t size,
                             uint32_t vnet_hdr_len,
-                            bool notify_remote_frame);
+                            bool notify_remote_frame,
+                            bool zero_copy);
 
 static bool packet_matches_str(const char *str,
                                const uint8_t *buf,
@@ -145,7 +174,7 @@ static void notify_remote_frame(CompareState *s)
     char msg[] = "DO_CHECKPOINT";
     int ret = 0;
 
-    ret = compare_chr_send(s, (uint8_t *)msg, strlen(msg), 0, true);
+    ret = compare_chr_send(s, (uint8_t *)msg, strlen(msg), 0, true, false);
     if (ret < 0) {
         error_report("Notify Xen COLO-frame failed");
     }
@@ -217,6 +246,7 @@ static int packet_enqueue(CompareState *s, int mode, Connection **con)
     ConnectionKey key;
     Packet *pkt = NULL;
     Connection *conn;
+    int ret;
 
     if (mode == PRIMARY_IN) {
         pkt = packet_new(s->pri_rs.buf,
@@ -245,16 +275,18 @@ static int packet_enqueue(CompareState *s, int mode, Connection **con)
     }
 
     if (mode == PRIMARY_IN) {
-        if (!colo_insert_packet(&conn->primary_list, pkt, &conn->pack)) {
-            error_report("colo compare primary queue size too big,"
-                         "drop packet");
-        }
+        ret = colo_insert_packet(&conn->primary_list, pkt, &conn->pack);
     } else {
-        if (!colo_insert_packet(&conn->secondary_list, pkt, &conn->sack)) {
-            error_report("colo compare secondary queue size too big,"
-                         "drop packet");
-        }
+        ret = colo_insert_packet(&conn->secondary_list, pkt, &conn->sack);
     }
+
+    if (!ret) {
+        trace_colo_compare_drop_packet(colo_mode[mode],
+            "queue size too big, drop packet");
+        packet_destroy(pkt, NULL);
+        pkt = NULL;
+    }
+
     *con = conn;
 
     return 0;
@@ -272,12 +304,13 @@ static void colo_release_primary_pkt(CompareState *s, Packet *pkt)
                            pkt->data,
                            pkt->size,
                            pkt->vnet_hdr_len,
-                           false);
+                           false,
+                           true);
     if (ret < 0) {
         error_report("colo send primary packet failed");
     }
     trace_colo_compare_main("packet same and release packet");
-    packet_destroy(pkt, NULL);
+    packet_destroy_partial(pkt, NULL);
 }
 
 /*
@@ -459,10 +492,12 @@ sec:
         g_queue_push_head(&conn->primary_list, ppkt);
         g_queue_push_head(&conn->secondary_list, spkt);
 
-        qemu_hexdump((char *)ppkt->data, stderr,
-                     "colo-compare ppkt", ppkt->size);
-        qemu_hexdump((char *)spkt->data, stderr,
-                     "colo-compare spkt", spkt->size);
+        if (trace_event_get_state_backends(TRACE_COLO_COMPARE_MISCOMPARE)) {
+            qemu_hexdump((char *)ppkt->data, stderr,
+                        "colo-compare ppkt", ppkt->size);
+            qemu_hexdump((char *)spkt->data, stderr,
+                        "colo-compare spkt", spkt->size);
+        }
 
         colo_compare_inconsistency_notify(s);
     }
@@ -699,65 +734,115 @@ static void colo_compare_connection(void *opaque, void *user_data)
     }
 }
 
-static int compare_chr_send(CompareState *s,
-                            const uint8_t *buf,
-                            uint32_t size,
-                            uint32_t vnet_hdr_len,
-                            bool notify_remote_frame)
+static void coroutine_fn _compare_chr_send(void *opaque)
 {
+    SendCo *sendco = opaque;
+    CompareState *s = sendco->s;
     int ret = 0;
-    uint32_t len = htonl(size);
 
-    if (!size) {
-        return 0;
-    }
+    while (!g_queue_is_empty(&sendco->send_list)) {
+        SendEntry *entry = g_queue_pop_tail(&sendco->send_list);
+        uint32_t len = htonl(entry->size);
 
-    if (notify_remote_frame) {
-        ret = qemu_chr_fe_write_all(&s->chr_notify_dev,
-                                    (uint8_t *)&len,
-                                    sizeof(len));
-    } else {
-        ret = qemu_chr_fe_write_all(&s->chr_out, (uint8_t *)&len, sizeof(len));
-    }
+        ret = qemu_chr_fe_write_all(sendco->chr, (uint8_t *)&len, sizeof(len));
 
-    if (ret != sizeof(len)) {
-        goto err;
-    }
+        if (ret != sizeof(len)) {
+            g_free(entry->buf);
+            g_slice_free(SendEntry, entry);
+            goto err;
+        }
 
-    if (s->vnet_hdr) {
-        /*
-         * We send vnet header len make other module(like filter-redirector)
-         * know how to parse net packet correctly.
-         */
-        len = htonl(vnet_hdr_len);
+        if (!sendco->notify_remote_frame && s->vnet_hdr) {
+            /*
+             * We send vnet header len make other module(like filter-redirector)
+             * know how to parse net packet correctly.
+             */
+            len = htonl(entry->vnet_hdr_len);
 
-        if (!notify_remote_frame) {
-            ret = qemu_chr_fe_write_all(&s->chr_out,
+            ret = qemu_chr_fe_write_all(sendco->chr,
                                         (uint8_t *)&len,
                                         sizeof(len));
+
+            if (ret != sizeof(len)) {
+                g_free(entry->buf);
+                g_slice_free(SendEntry, entry);
+                goto err;
+            }
         }
 
-        if (ret != sizeof(len)) {
+        ret = qemu_chr_fe_write_all(sendco->chr,
+                                    (uint8_t *)entry->buf,
+                                    entry->size);
+
+        if (ret != entry->size) {
+            g_free(entry->buf);
+            g_slice_free(SendEntry, entry);
             goto err;
         }
+
+        g_free(entry->buf);
+        g_slice_free(SendEntry, entry);
+    }
+
+    sendco->ret = 0;
+    goto out;
+
+err:
+    while (!g_queue_is_empty(&sendco->send_list)) {
+        SendEntry *entry = g_queue_pop_tail(&sendco->send_list);
+        g_free(entry->buf);
+        g_slice_free(SendEntry, entry);
     }
+    sendco->ret = ret < 0 ? ret : -EIO;
+out:
+    sendco->co = NULL;
+    sendco->done = true;
+    aio_wait_kick();
+}
+
+static int compare_chr_send(CompareState *s,
+                            uint8_t *buf,
+                            uint32_t size,
+                            uint32_t vnet_hdr_len,
+                            bool notify_remote_frame,
+                            bool zero_copy)
+{
+    SendCo *sendco;
+    SendEntry *entry;
 
     if (notify_remote_frame) {
-        ret = qemu_chr_fe_write_all(&s->chr_notify_dev,
-                                    (uint8_t *)buf,
-                                    size);
+        sendco = &s->notify_sendco;
     } else {
-        ret = qemu_chr_fe_write_all(&s->chr_out, (uint8_t *)buf, size);
+        sendco = &s->out_sendco;
     }
 
-    if (ret != size) {
-        goto err;
+    if (!size) {
+        return 0;
     }
 
-    return 0;
+    entry = g_slice_new(SendEntry);
+    entry->size = size;
+    entry->vnet_hdr_len = vnet_hdr_len;
+    if (zero_copy) {
+        entry->buf = buf;
+    } else {
+        entry->buf = g_malloc(size);
+        memcpy(entry->buf, buf, size);
+    }
+    g_queue_push_head(&sendco->send_list, entry);
+
+    if (sendco->done) {
+        sendco->co = qemu_coroutine_create(_compare_chr_send, sendco);
+        sendco->done = false;
+        qemu_coroutine_enter(sendco->co);
+        if (sendco->done) {
+            /* report early errors */
+            return sendco->ret;
+        }
+    }
 
-err:
-    return ret < 0 ? ret : -EIO;
+    /* assume success */
+    return 0;
 }
 
 static int compare_chr_can_read(void *opaque)
@@ -830,6 +915,12 @@ static void check_old_packet_regular(void *opaque)
 void colo_notify_compares_event(void *opaque, int event, Error **errp)
 {
     CompareState *s;
+    qemu_mutex_lock(&colo_compare_mutex);
+
+    if (!colo_compare_active) {
+        qemu_mutex_unlock(&colo_compare_mutex);
+        return;
+    }
 
     qemu_mutex_lock(&event_mtx);
     QTAILQ_FOREACH(s, &net_compares, next) {
@@ -843,6 +934,7 @@ void colo_notify_compares_event(void *opaque, int event, Error **errp)
     }
 
     qemu_mutex_unlock(&event_mtx);
+    qemu_mutex_unlock(&colo_compare_mutex);
 }
 
 static void colo_compare_timer_init(CompareState *s)
@@ -890,6 +982,7 @@ static void colo_compare_handle_event(void *opaque)
 
 static void colo_compare_iothread(CompareState *s)
 {
+    AioContext *ctx = iothread_get_aio_context(s->iothread);
     object_ref(OBJECT(s->iothread));
     s->worker_context = iothread_get_g_main_context(s->iothread);
 
@@ -906,7 +999,7 @@ static void colo_compare_iothread(CompareState *s)
     }
 
     colo_compare_timer_init(s);
-    s->event_bh = qemu_bh_new(colo_compare_handle_event, s);
+    s->event_bh = aio_bh_new(ctx, colo_compare_handle_event, s);
 }
 
 static char *compare_get_pri_indev(Object *obj, Error **errp)
@@ -1062,6 +1155,7 @@ static void compare_pri_rs_finalize(SocketReadState *pri_rs)
                          pri_rs->buf,
                          pri_rs->packet_len,
                          pri_rs->vnet_hdr_len,
+                         false,
                          false);
     } else {
         /* compare packet in the specified connection */
@@ -1092,7 +1186,7 @@ static void compare_notify_rs_finalize(SocketReadState *notify_rs)
     if (packet_matches_str("COLO_USERSPACE_PROXY_INIT",
                            notify_rs->buf,
                            notify_rs->packet_len)) {
-        ret = compare_chr_send(s, (uint8_t *)msg, strlen(msg), 0, true);
+        ret = compare_chr_send(s, (uint8_t *)msg, strlen(msg), 0, true, false);
         if (ret < 0) {
             error_report("Notify Xen COLO-frame INIT failed");
         }
@@ -1196,12 +1290,21 @@ static void colo_compare_complete(UserCreatable *uc, Error **errp)
                            s->vnet_hdr);
     }
 
-    QTAILQ_INSERT_TAIL(&net_compares, s, next);
+    s->out_sendco.s = s;
+    s->out_sendco.chr = &s->chr_out;
+    s->out_sendco.notify_remote_frame = false;
+    s->out_sendco.done = true;
+    g_queue_init(&s->out_sendco.send_list);
 
-    g_queue_init(&s->conn_list);
+    if (s->notify_dev) {
+        s->notify_sendco.s = s;
+        s->notify_sendco.chr = &s->chr_notify_dev;
+        s->notify_sendco.notify_remote_frame = true;
+        s->notify_sendco.done = true;
+        g_queue_init(&s->notify_sendco.send_list);
+    }
 
-    qemu_mutex_init(&event_mtx);
-    qemu_cond_init(&event_complete_cond);
+    g_queue_init(&s->conn_list);
 
     s->connection_track_table = g_hash_table_new_full(connection_key_hash,
                                                       connection_key_equal,
@@ -1209,6 +1312,16 @@ static void colo_compare_complete(UserCreatable *uc, Error **errp)
                                                       connection_destroy);
 
     colo_compare_iothread(s);
+
+    qemu_mutex_lock(&colo_compare_mutex);
+    if (!colo_compare_active) {
+        qemu_mutex_init(&event_mtx);
+        qemu_cond_init(&event_complete_cond);
+        colo_compare_active = true;
+    }
+    QTAILQ_INSERT_TAIL(&net_compares, s, next);
+    qemu_mutex_unlock(&colo_compare_mutex);
+
     return;
 }
 
@@ -1224,8 +1337,9 @@ static void colo_flush_packets(void *opaque, void *user_data)
                          pkt->data,
                          pkt->size,
                          pkt->vnet_hdr_len,
-                         false);
-        packet_destroy(pkt, NULL);
+                         false,
+                         true);
+        packet_destroy_partial(pkt, NULL);
     }
     while (!g_queue_is_empty(&conn->secondary_list)) {
         pkt = g_queue_pop_head(&conn->secondary_list);
@@ -1276,6 +1390,20 @@ static void colo_compare_finalize(Object *obj)
     CompareState *s = COLO_COMPARE(obj);
     CompareState *tmp = NULL;
 
+    qemu_mutex_lock(&colo_compare_mutex);
+    QTAILQ_FOREACH(tmp, &net_compares, next) {
+        if (tmp == s) {
+            QTAILQ_REMOVE(&net_compares, s, next);
+            break;
+        }
+    }
+    if (QTAILQ_EMPTY(&net_compares)) {
+        colo_compare_active = false;
+        qemu_mutex_destroy(&event_mtx);
+        qemu_cond_destroy(&event_complete_cond);
+    }
+    qemu_mutex_unlock(&colo_compare_mutex);
+
     qemu_chr_fe_deinit(&s->chr_pri_in, false);
     qemu_chr_fe_deinit(&s->chr_sec_in, false);
     qemu_chr_fe_deinit(&s->chr_out, false);
@@ -1289,17 +1417,23 @@ static void colo_compare_finalize(Object *obj)
 
     qemu_bh_delete(s->event_bh);
 
-    QTAILQ_FOREACH(tmp, &net_compares, next) {
-        if (tmp == s) {
-            QTAILQ_REMOVE(&net_compares, s, next);
-            break;
-        }
+    AioContext *ctx = iothread_get_aio_context(s->iothread);
+    aio_context_acquire(ctx);
+    AIO_WAIT_WHILE(ctx, !s->out_sendco.done);
+    if (s->notify_dev) {
+        AIO_WAIT_WHILE(ctx, !s->notify_sendco.done);
     }
+    aio_context_release(ctx);
 
     /* Release all unhandled packets after compare thead exited */
     g_queue_foreach(&s->conn_list, colo_flush_packets, s);
+    AIO_WAIT_WHILE(NULL, !s->out_sendco.done);
 
     g_queue_clear(&s->conn_list);
+    g_queue_clear(&s->out_sendco.send_list);
+    if (s->notify_dev) {
+        g_queue_clear(&s->notify_sendco.send_list);
+    }
 
     if (s->connection_track_table) {
         g_hash_table_destroy(s->connection_track_table);
@@ -1309,15 +1443,18 @@ static void colo_compare_finalize(Object *obj)
         object_unref(OBJECT(s->iothread));
     }
 
-    qemu_mutex_destroy(&event_mtx);
-    qemu_cond_destroy(&event_complete_cond);
-
     g_free(s->pri_indev);
     g_free(s->sec_indev);
     g_free(s->outdev);
     g_free(s->notify_dev);
 }
 
+static void __attribute__((__constructor__)) colo_compare_init_globals(void)
+{
+    colo_compare_active = false;
+    qemu_mutex_init(&colo_compare_mutex);
+}
+
 static const TypeInfo colo_compare_info = {
     .name = TYPE_COLO_COMPARE,
     .parent = TYPE_OBJECT,