[Qemu-devel] [PATCH 10/10] TCP based live migration

[Anthony Liguori <aliguori@us.ibm.com>]

This patch introduces a tcp protocol for live migration.  It can be used as
follows:

qemu-system-x86_64 -hda ~/images/linux-test.img -monitor stdio
 <vm runs for a while>
(qemu) migrate tcp:localhost:1025

On the same system:

qemu-system-x86_64 -hda ~/images/linux-test.img -incoming tcp:localhost:1025

The monitor can be interacted with while waiting for an incoming live migration.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

diff --git a/Makefile.target b/Makefile.target
index 6bf5229..0fa585c 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -473,7 +473,7 @@ endif #CONFIG_DARWIN_USER
 ifndef CONFIG_USER_ONLY
 
 OBJS=vl.o osdep.o monitor.o pci.o loader.o isa_mmio.o machine.o net-checksum.o
-OBJS+=migration.o
+OBJS+=migration.o migration-tcp.o
 ifdef CONFIG_WIN32
 OBJS+=block-raw-win32.o
 else
diff --git a/migration-tcp.c b/migration-tcp.c
new file mode 100644
index 0000000..6b46bd4
--- /dev/null
+++ b/migration-tcp.c
@@ -0,0 +1,243 @@
+#include "qemu-common.h"
+#include "qemu_socket.h"
+#include "migration.h"
+#include "qemu-char.h"
+#include "sysemu.h"
+#include "console.h"
+
+/* FIXME resume monitor on error */
+
+
+typedef struct FdMigrationState
+{
+    QEMUFile *file;
+    int64_t bandwidth_limit;
+    int fd;
+} FdMigrationState;
+
+static void fd_put_notify(void *opaque)
+{
+    FdMigrationState *s = opaque;
+
+    qemu_set_fd_handler2(s->fd, NULL, NULL, NULL, NULL);
+    qemu_file_put_notify(s->file);
+}
+
+static ssize_t fd_put_buffer(void *opaque, const void *data, size_t size)
+{
+    FdMigrationState *s = opaque;
+    ssize_t ret;
+
+    do {
+        ret = write(s->fd, data, size);
+    } while (ret == -1 && errno == EINTR);
+
+    if (ret == -1)
+        ret = -errno;
+
+    if (ret == -EAGAIN)
+        qemu_set_fd_handler2(s->fd, NULL, NULL, fd_put_notify, s);
+
+    return ret;
+}
+
+static int fd_close(void *opaque)
+{
+    FdMigrationState *s = opaque;
+    return close(s->fd);
+}
+
+static void fd_wait_for_unfreeze(void *opaque)
+{
+    FdMigrationState *s = opaque;
+    int ret;
+
+    do {
+        fd_set wfds;
+
+        FD_ZERO(&wfds);
+        FD_SET(s->fd, &wfds);
+
+        ret = select(s->fd + 1, NULL, &wfds, NULL, NULL);
+    } while (ret == -1 && errno == EINTR);
+}
+
+static void fd_put_ready(void *opaque)
+{
+    FdMigrationState *s = opaque;
+
+    if (qemu_savevm_state_iterate(s->file) == 1) {
+        vm_stop(0);
+        qemu_savevm_state_complete(s->file);
+        qemu_fclose(s->file);
+        qemu_free(s);
+        monitor_resume();
+    }
+}
+
+static void tcp_connect_migrate(FdMigrationState *s)
+{
+    int ret;
+
+    s->file = qemu_fopen_ops_buffered(s,
+                                      s->bandwidth_limit,
+                                      fd_put_buffer,
+                                      fd_put_ready,
+                                      fd_wait_for_unfreeze,
+                                      fd_close);
+
+    ret = qemu_savevm_state_begin(s->file);
+    if (ret < 0) {
+        fprintf(stderr, "savevm failed %d\n", ret);
+        return;
+    }
+
+    monitor_suspend();
+
+    fd_put_ready(s);
+}
+
+static void tcp_wait_for_connect(void *opaque)
+{
+    FdMigrationState *s = opaque;
+    int val, ret;
+    int valsize = sizeof(val);
+
+    do {
+        ret = getsockopt(s->fd, SOL_SOCKET, SO_ERROR, &val, &valsize);
+    } while (ret == -1 && errno == EINTR);
+
+    if (ret < 0) {
+        fprintf(stderr, "Could not query connect success\n");
+        return;
+    }
+
+    qemu_set_fd_handler2(s->fd, NULL, NULL, NULL, NULL);
+
+    if (val == 0)
+        tcp_connect_migrate(s);
+    else {
+        fprintf(stderr, "failed to connect to host\n");
+        close(s->fd);
+        qemu_free(s);
+    }
+}
+
+int tcp_start_outgoing_migration(const char *host_port, int64_t bandwidth_limit)
+{
+    struct sockaddr_in addr;
+    FdMigrationState *s;
+    int ret;
+
+    if (parse_host_port(&addr, host_port) < 0) {
+        fprintf(stderr, "invalid host/port combination: %s\n", host_port);
+        return -EINVAL;
+    }
+
+    s = qemu_mallocz(sizeof(*s));
+    if (s == NULL)
+        return -ENOMEM;
+
+    s->bandwidth_limit = bandwidth_limit;
+    s->fd = socket(PF_INET, SOCK_STREAM, 0);
+    if (s->fd == -1) {
+        qemu_free(s);
+        return -errno;
+    }
+
+    fcntl(s->fd, F_SETFL, O_NONBLOCK);
+
+    do {
+        ret = connect(s->fd, (struct sockaddr *)&addr, sizeof(addr));
+        if (ret == -1)
+            ret = -errno;
+
+        if (ret == -EINPROGRESS)
+            qemu_set_fd_handler2(s->fd, NULL, NULL, tcp_wait_for_connect, s);
+    } while (ret == -EINTR);
+
+    if (ret < 0 && ret != -EINPROGRESS) {
+        fprintf(stderr, "failed to connect to host\n");
+        close(s->fd);
+        qemu_free(s);
+    } else if (ret >= 0)
+        tcp_connect_migrate(s);
+
+    return 0;
+}
+
+static void tcp_accept_incoming_migration(void *opaque)
+{
+    struct sockaddr_in addr;
+    socklen_t addrlen = sizeof(addr);
+    int s = (unsigned long)opaque;
+    QEMUFile *f;
+    int c, ret;
+
+    do {
+        c = accept(s, (struct sockaddr *)&addr, &addrlen);
+    } while (c == -1 && errno == EINTR);
+
+    if (c == -1) {
+        fprintf(stderr, "could not accept migration connection\n");
+        return;
+    }
+
+    f = qemu_fopen_fd(c);
+    if (f == NULL) {
+        fprintf(stderr, "could not qemu_fopen socket\n");
+        goto out;
+    }
+
+    vm_stop(0); /* just in case */
+    ret = qemu_loadvm_state(f);
+    if (ret < 0) {
+        fprintf(stderr, "load of migration failed\n");
+        goto out_fopen;
+    }
+
+    /* we've successfully migrated, close the server socket */
+    qemu_set_fd_handler2(s, NULL, NULL, NULL, NULL);
+    close(s);
+
+    vm_start();
+
+out_fopen:
+    qemu_fclose(f);
+out:
+    close(c);
+}
+
+int tcp_start_incoming_migration(const char *host_port)
+{
+    struct sockaddr_in addr;
+    int val;
+    int s;
+
+    if (parse_host_port(&addr, host_port) < 0) {
+        fprintf(stderr, "invalid host/port combination: %s\n", host_port);
+        return -EINVAL;
+    }
+
+    s = socket(PF_INET, SOCK_STREAM, 0);
+    if (s == -1)
+        return -errno;
+
+    val = 1;
+    setsockopt(s, SOL_SOCKET, SO_REUSEADDR, (const char *)&val, sizeof(val));
+
+    if (bind(s, (struct sockaddr *)&addr, sizeof(addr)) == -1)
+        goto err;
+
+    if (listen(s, 1) == -1)
+        goto err;
+
+    qemu_set_fd_handler2(s, NULL, tcp_accept_incoming_migration, NULL,
+                         (void *)(unsigned long)s);
+
+    return 0;
+
+err:
+    close(s);
+    return -errno;
+}
diff --git a/migration.c b/migration.c
index 8a8b4a5..6023e02 100644
--- a/migration.c
+++ b/migration.c
@@ -200,10 +200,20 @@ QEMUFile *qemu_fopen_ops_buffered(void *opaque,
 
 void qemu_start_incoming_migration(const char *uri)
 {
-    fprintf(stderr, "unknown migration protocol: %s\n", uri);
+    const char *p;
+
+    if (strstart(uri, "tcp:", &p))
+        tcp_start_incoming_migration(p);
+    else
+        fprintf(stderr, "unknown migration protocol: %s\n", uri);
 }
 
 void do_migrate(const char *uri)
 {
-    fprintf(stderr, "unknown migration protocol: %s\n", uri);
+    const char *p;
+
+    if (strstart(uri, "tcp:", &p))
+        tcp_start_outgoing_migration(p, 20 << 20);
+    else
+        fprintf(stderr, "unknown migration protocol: %s\n", uri);
 }
diff --git a/migration.h b/migration.h
index 2119a59..314cc87 100644
--- a/migration.h
+++ b/migration.h
@@ -18,4 +18,8 @@ void qemu_start_incoming_migration(const char *uri);
 
 void do_migrate(const char *uri);
 
+int tcp_start_incoming_migration(const char *host_port);
+
+int tcp_start_outgoing_migration(const char *host_port, int64_t bandwidth_limit);
+
 #endif