glib/gio/tests/gdbus-peer.c

1461 lines
49 KiB
C
Raw Normal View History

/* GLib testing framework examples and tests
*
2010-05-10 14:07:28 +02:00
* Copyright (C) 2008-2010 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General
* Public License along with this library; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place, Suite 330,
* Boston, MA 02111-1307, USA.
*
* Author: David Zeuthen <davidz@redhat.com>
*/
#include <gio/gio.h>
#include <unistd.h>
#include <string.h>
/* for open(2) */
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <string.h>
/* for g_unlink() */
#include <glib/gstdio.h>
#include <gio/gnetworkingprivate.h>
#include <gio/gunixsocketaddress.h>
#include <gio/gunixfdlist.h>
/* used in test_overflow */
#ifdef G_OS_UNIX
#include <gio/gunixconnection.h>
#include <errno.h>
#endif
#include "gdbus-tests.h"
#ifdef G_OS_UNIX
static gboolean is_unix = TRUE;
#else
static gboolean is_unix = FALSE;
#endif
static gchar *test_guid = NULL;
static GMainLoop *service_loop = NULL;
static GDBusServer *server = NULL;
static GMainLoop *loop = NULL;
/* ---------------------------------------------------------------------------------------------------- */
/* Test that peer-to-peer connections work */
/* ---------------------------------------------------------------------------------------------------- */
typedef struct
{
gboolean accept_connection;
gint num_connection_attempts;
GPtrArray *current_connections;
guint num_method_calls;
gboolean signal_received;
} PeerData;
static const gchar *test_interface_introspection_xml =
"<node>"
" <interface name='org.gtk.GDBus.PeerTestInterface'>"
" <method name='HelloPeer'>"
" <arg type='s' name='greeting' direction='in'/>"
" <arg type='s' name='response' direction='out'/>"
" </method>"
" <method name='EmitSignal'/>"
" <method name='EmitSignalWithNameSet'/>"
" <method name='OpenFile'>"
" <arg type='s' name='path' direction='in'/>"
" </method>"
" <signal name='PeerSignal'>"
" <arg type='s' name='a_string'/>"
" </signal>"
" <property type='s' name='PeerProperty' access='read'/>"
" </interface>"
"</node>";
static GDBusInterfaceInfo *test_interface_introspection_data = NULL;
static void
test_interface_method_call (GDBusConnection *connection,
const gchar *sender,
const gchar *object_path,
const gchar *interface_name,
const gchar *method_name,
GVariant *parameters,
GDBusMethodInvocation *invocation,
gpointer user_data)
{
PeerData *data = user_data;
const GDBusMethodInfo *info;
data->num_method_calls++;
g_assert_cmpstr (object_path, ==, "/org/gtk/GDBus/PeerTestObject");
g_assert_cmpstr (interface_name, ==, "org.gtk.GDBus.PeerTestInterface");
info = g_dbus_method_invocation_get_method_info (invocation);
g_assert_cmpstr (info->name, ==, method_name);
if (g_strcmp0 (method_name, "HelloPeer") == 0)
{
const gchar *greeting;
gchar *response;
g_variant_get (parameters, "(&s)", &greeting);
response = g_strdup_printf ("You greeted me with '%s'.",
greeting);
g_dbus_method_invocation_return_value (invocation,
g_variant_new ("(s)", response));
g_free (response);
}
else if (g_strcmp0 (method_name, "EmitSignal") == 0)
{
GError *error;
error = NULL;
g_dbus_connection_emit_signal (connection,
NULL,
"/org/gtk/GDBus/PeerTestObject",
"org.gtk.GDBus.PeerTestInterface",
"PeerSignal",
NULL,
&error);
g_assert_no_error (error);
g_dbus_method_invocation_return_value (invocation, NULL);
}
else if (g_strcmp0 (method_name, "EmitSignalWithNameSet") == 0)
{
GError *error;
gboolean ret;
GDBusMessage *message;
message = g_dbus_message_new_signal ("/org/gtk/GDBus/PeerTestObject",
"org.gtk.GDBus.PeerTestInterface",
"PeerSignalWithNameSet");
g_dbus_message_set_sender (message, ":1.42");
error = NULL;
ret = g_dbus_connection_send_message (connection, message, G_DBUS_SEND_MESSAGE_FLAGS_NONE, NULL, &error);
g_assert_no_error (error);
g_assert (ret);
g_object_unref (message);
g_dbus_method_invocation_return_value (invocation, NULL);
}
else if (g_strcmp0 (method_name, "OpenFile") == 0)
{
#ifdef G_OS_UNIX
const gchar *path;
GDBusMessage *reply;
GError *error;
gint fd;
GUnixFDList *fd_list;
g_variant_get (parameters, "(&s)", &path);
fd_list = g_unix_fd_list_new ();
error = NULL;
fd = open (path, O_RDONLY);
g_unix_fd_list_append (fd_list, fd, &error);
g_assert_no_error (error);
close (fd);
reply = g_dbus_message_new_method_reply (g_dbus_method_invocation_get_message (invocation));
g_dbus_message_set_unix_fd_list (reply, fd_list);
g_object_unref (invocation);
error = NULL;
g_dbus_connection_send_message (connection,
reply,
G_DBUS_SEND_MESSAGE_FLAGS_NONE,
NULL, /* out_serial */
&error);
g_assert_no_error (error);
g_object_unref (reply);
#else
g_dbus_method_invocation_return_dbus_error (invocation,
"org.gtk.GDBus.NotOnUnix",
"Your OS does not support file descriptor passing");
#endif
}
else
{
g_assert_not_reached ();
}
}
static GVariant *
test_interface_get_property (GDBusConnection *connection,
const gchar *sender,
const gchar *object_path,
const gchar *interface_name,
const gchar *property_name,
GError **error,
gpointer user_data)
{
g_assert_cmpstr (object_path, ==, "/org/gtk/GDBus/PeerTestObject");
g_assert_cmpstr (interface_name, ==, "org.gtk.GDBus.PeerTestInterface");
g_assert_cmpstr (property_name, ==, "PeerProperty");
return g_variant_new_string ("ThePropertyValue");
}
static const GDBusInterfaceVTable test_interface_vtable =
{
test_interface_method_call,
test_interface_get_property,
NULL /* set_property */
};
static void
on_proxy_signal_received (GDBusProxy *proxy,
gchar *sender_name,
gchar *signal_name,
GVariant *parameters,
gpointer user_data)
{
PeerData *data = user_data;
data->signal_received = TRUE;
g_assert (sender_name == NULL);
g_assert_cmpstr (signal_name, ==, "PeerSignal");
g_main_loop_quit (loop);
}
static void
on_proxy_signal_received_with_name_set (GDBusProxy *proxy,
gchar *sender_name,
gchar *signal_name,
GVariant *parameters,
gpointer user_data)
{
PeerData *data = user_data;
data->signal_received = TRUE;
g_assert_cmpstr (sender_name, ==, ":1.42");
g_assert_cmpstr (signal_name, ==, "PeerSignalWithNameSet");
g_main_loop_quit (loop);
}
/* ---------------------------------------------------------------------------------------------------- */
static gboolean
on_authorize_authenticated_peer (GDBusAuthObserver *observer,
GIOStream *stream,
GCredentials *credentials,
gpointer user_data)
{
PeerData *data = user_data;
gboolean authorized;
data->num_connection_attempts++;
authorized = TRUE;
if (!data->accept_connection)
{
authorized = FALSE;
g_main_loop_quit (loop);
}
return authorized;
}
/* Runs in thread we created GDBusServer in (since we didn't pass G_DBUS_SERVER_FLAGS_RUN_IN_THREAD) */
static void
on_new_connection (GDBusServer *server,
GDBusConnection *connection,
gpointer user_data)
{
PeerData *data = user_data;
GError *error;
guint reg_id;
//g_print ("Client connected.\n"
// "Negotiated capabilities: unix-fd-passing=%d\n",
// g_dbus_connection_get_capabilities (connection) & G_DBUS_CAPABILITY_FLAGS_UNIX_FD_PASSING);
g_ptr_array_add (data->current_connections, g_object_ref (connection));
/* export object on the newly established connection */
error = NULL;
reg_id = g_dbus_connection_register_object (connection,
"/org/gtk/GDBus/PeerTestObject",
test_interface_introspection_data,
&test_interface_vtable,
data,
NULL, /* GDestroyNotify for data */
&error);
g_assert_no_error (error);
g_assert (reg_id > 0);
g_main_loop_quit (loop);
}
static gpointer
service_thread_func (gpointer user_data)
{
PeerData *data = user_data;
GMainContext *service_context;
GDBusAuthObserver *observer;
GError *error;
service_context = g_main_context_new ();
g_main_context_push_thread_default (service_context);
error = NULL;
observer = g_dbus_auth_observer_new ();
server = g_dbus_server_new_sync (is_unix ? "unix:tmpdir=/tmp/gdbus-test-" : "nonce-tcp:",
G_DBUS_SERVER_FLAGS_NONE,
test_guid,
observer,
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_signal_connect (server,
"new-connection",
G_CALLBACK (on_new_connection),
data);
g_signal_connect (observer,
"authorize-authenticated-peer",
G_CALLBACK (on_authorize_authenticated_peer),
data);
g_object_unref (observer);
g_dbus_server_start (server);
service_loop = g_main_loop_new (service_context, FALSE);
g_main_loop_run (service_loop);
g_main_context_pop_thread_default (service_context);
g_main_loop_unref (service_loop);
g_main_context_unref (service_context);
/* test code specifically unrefs the server - see below */
g_assert (server == NULL);
return NULL;
}
#if 0
static gboolean
on_incoming_connection (GSocketService *service,
GSocketConnection *socket_connection,
GObject *source_object,
gpointer user_data)
{
PeerData *data = user_data;
if (data->accept_connection)
{
GError *error;
guint reg_id;
GDBusConnection *connection;
error = NULL;
connection = g_dbus_connection_new_sync (G_IO_STREAM (socket_connection),
test_guid,
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_SERVER,
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_ptr_array_add (data->current_connections, connection);
/* export object on the newly established connection */
error = NULL;
reg_id = g_dbus_connection_register_object (connection,
"/org/gtk/GDBus/PeerTestObject",
&test_interface_introspection_data,
&test_interface_vtable,
data,
NULL, /* GDestroyNotify for data */
&error);
g_assert_no_error (error);
g_assert (reg_id > 0);
}
else
{
/* don't do anything */
}
data->num_connection_attempts++;
g_main_loop_quit (loop);
/* stops other signal handlers from being invoked */
return TRUE;
}
static gpointer
service_thread_func (gpointer data)
{
GMainContext *service_context;
gchar *socket_path;
GSocketAddress *address;
GError *error;
service_context = g_main_context_new ();
g_main_context_push_thread_default (service_context);
socket_path = g_strdup_printf ("/tmp/gdbus-test-pid-%d", getpid ());
address = g_unix_socket_address_new (socket_path);
service = g_socket_service_new ();
error = NULL;
g_socket_listener_add_address (G_SOCKET_LISTENER (service),
address,
G_SOCKET_TYPE_STREAM,
G_SOCKET_PROTOCOL_DEFAULT,
NULL, /* source_object */
NULL, /* effective_address */
&error);
g_assert_no_error (error);
g_signal_connect (service,
"incoming",
G_CALLBACK (on_incoming_connection),
data);
g_socket_service_start (service);
service_loop = g_main_loop_new (service_context, FALSE);
g_main_loop_run (service_loop);
g_main_context_pop_thread_default (service_context);
g_main_loop_unref (service_loop);
g_main_context_unref (service_context);
g_object_unref (address);
g_free (socket_path);
return NULL;
}
#endif
/* ---------------------------------------------------------------------------------------------------- */
#if 0
static gboolean
check_connection (gpointer user_data)
{
PeerData *data = user_data;
guint n;
for (n = 0; n < data->current_connections->len; n++)
{
GDBusConnection *c;
GIOStream *stream;
c = G_DBUS_CONNECTION (data->current_connections->pdata[n]);
stream = g_dbus_connection_get_stream (c);
g_debug ("In check_connection for %d: connection %p, stream %p", n, c, stream);
g_debug ("closed = %d", g_io_stream_is_closed (stream));
GSocket *socket;
socket = g_socket_connection_get_socket (G_SOCKET_CONNECTION (stream));
g_debug ("socket_closed = %d", g_socket_is_closed (socket));
g_debug ("socket_condition_check = %d", g_socket_condition_check (socket, G_IO_IN|G_IO_OUT|G_IO_ERR|G_IO_HUP));
gchar buf[128];
GError *error;
gssize num_read;
error = NULL;
num_read = g_input_stream_read (g_io_stream_get_input_stream (stream),
buf,
128,
NULL,
&error);
if (num_read < 0)
{
g_debug ("error: %s", error->message);
g_error_free (error);
}
else
{
g_debug ("no error, read %d bytes", (gint) num_read);
}
}
return FALSE;
}
static gboolean
on_do_disconnect_in_idle (gpointer data)
{
GDBusConnection *c = G_DBUS_CONNECTION (data);
g_debug ("GDC %p has ref_count %d", c, G_OBJECT (c)->ref_count);
g_dbus_connection_disconnect (c);
g_object_unref (c);
return FALSE;
}
#endif
#ifdef G_OS_UNIX
static gchar *
read_all_from_fd (gint fd, gsize *out_len, GError **error)
{
GString *str;
gchar buf[64];
gssize num_read;
str = g_string_new (NULL);
do
{
num_read = read (fd, buf, sizeof (buf));
if (num_read == -1)
{
if (errno == EAGAIN || errno == EWOULDBLOCK)
continue;
g_set_error (error,
G_IO_ERROR,
g_io_error_from_errno (errno),
"Failed reading %d bytes into offset %d: %s",
(gint) sizeof (buf),
(gint) str->len,
strerror (errno));
goto error;
}
else if (num_read > 0)
{
g_string_append_len (str, buf, num_read);
}
else if (num_read == 0)
{
break;
}
}
while (TRUE);
if (out_len != NULL)
*out_len = str->len;
return g_string_free (str, FALSE);
error:
if (out_len != NULL)
out_len = 0;
g_string_free (str, TRUE);
return NULL;
}
#endif
static void
test_peer (void)
{
GDBusConnection *c;
GDBusConnection *c2;
GDBusProxy *proxy;
GError *error;
PeerData data;
GVariant *value;
GVariant *result;
const gchar *s;
GThread *service_thread;
gulong signal_handler_id;
memset (&data, '\0', sizeof (PeerData));
data.current_connections = g_ptr_array_new_with_free_func (g_object_unref);
/* first try to connect when there is no server */
error = NULL;
c = g_dbus_connection_new_for_address_sync (is_unix ? "unix:path=/tmp/gdbus-test-does-not-exist-pid" :
/* NOTE: Even if something is listening on port 12345 the connection
* will fail because the nonce file doesn't exist */
"nonce-tcp:host=localhost,port=12345,noncefile=this-does-not-exist-gdbus",
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
_g_assert_error_domain (error, G_IO_ERROR);
g_assert (!g_dbus_error_is_remote_error (error));
g_clear_error (&error);
g_assert (c == NULL);
/* bring up a server - we run the server in a different thread to avoid deadlocks */
error = NULL;
service_thread = g_thread_create (service_thread_func,
&data,
TRUE,
&error);
while (service_loop == NULL)
g_thread_yield ();
g_assert (server != NULL);
/* bring up a connection and accept it */
data.accept_connection = TRUE;
error = NULL;
c = g_dbus_connection_new_for_address_sync (g_dbus_server_get_client_address (server),
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_assert (c != NULL);
while (data.current_connections->len < 1)
g_main_loop_run (loop);
g_assert_cmpint (data.current_connections->len, ==, 1);
g_assert_cmpint (data.num_connection_attempts, ==, 1);
g_assert (g_dbus_connection_get_unique_name (c) == NULL);
g_assert_cmpstr (g_dbus_connection_get_guid (c), ==, test_guid);
/* check that we create a proxy, read properties, receive signals and invoke
* the HelloPeer() method. Since the server runs in another thread it's fine
* to use synchronous blocking API here.
*/
error = NULL;
proxy = g_dbus_proxy_new_sync (c,
G_DBUS_PROXY_FLAGS_NONE,
NULL,
NULL, /* bus_name */
"/org/gtk/GDBus/PeerTestObject",
"org.gtk.GDBus.PeerTestInterface",
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_assert (proxy != NULL);
error = NULL;
value = g_dbus_proxy_get_cached_property (proxy, "PeerProperty");
g_assert_cmpstr (g_variant_get_string (value, NULL), ==, "ThePropertyValue");
/* try invoking a method */
error = NULL;
result = g_dbus_proxy_call_sync (proxy,
"HelloPeer",
g_variant_new ("(s)", "Hey Peer!"),
G_DBUS_CALL_FLAGS_NONE,
-1,
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_variant_get (result, "(&s)", &s);
g_assert_cmpstr (s, ==, "You greeted me with 'Hey Peer!'.");
g_variant_unref (result);
g_assert_cmpint (data.num_method_calls, ==, 1);
/* make the other peer emit a signal - catch it */
signal_handler_id = g_signal_connect (proxy,
"g-signal",
G_CALLBACK (on_proxy_signal_received),
&data);
g_assert (!data.signal_received);
g_dbus_proxy_call (proxy,
"EmitSignal",
NULL, /* no arguments */
G_DBUS_CALL_FLAGS_NONE,
-1,
NULL, /* GCancellable */
NULL, /* GAsyncReadyCallback - we don't care about the result */
NULL); /* user_data */
g_main_loop_run (loop);
g_assert (data.signal_received);
g_assert_cmpint (data.num_method_calls, ==, 2);
g_signal_handler_disconnect (proxy, signal_handler_id);
/* Also ensure that messages with the sender header-field set gets
* delivered to the proxy - note that this doesn't really make sense
* e.g. names are meaning-less in a peer-to-peer case... but we
* support it because it makes sense in certain bridging
* applications - see e.g. #623815.
*/
signal_handler_id = g_signal_connect (proxy,
"g-signal",
G_CALLBACK (on_proxy_signal_received_with_name_set),
&data);
data.signal_received = FALSE;
g_dbus_proxy_call (proxy,
"EmitSignalWithNameSet",
NULL, /* no arguments */
G_DBUS_CALL_FLAGS_NONE,
-1,
NULL, /* GCancellable */
NULL, /* GAsyncReadyCallback - we don't care about the result */
NULL); /* user_data */
g_main_loop_run (loop);
g_assert (data.signal_received);
g_assert_cmpint (data.num_method_calls, ==, 3);
g_signal_handler_disconnect (proxy, signal_handler_id);
/* check for UNIX fd passing */
#ifdef G_OS_UNIX
{
GDBusMessage *method_call_message;
GDBusMessage *method_reply_message;
GUnixFDList *fd_list;
gint fd;
gchar *buf;
gsize len;
gchar *buf2;
gsize len2;
method_call_message = g_dbus_message_new_method_call (NULL, /* name */
"/org/gtk/GDBus/PeerTestObject",
"org.gtk.GDBus.PeerTestInterface",
"OpenFile");
g_dbus_message_set_body (method_call_message, g_variant_new ("(s)", "/etc/hosts"));
error = NULL;
method_reply_message = g_dbus_connection_send_message_with_reply_sync (c,
method_call_message,
G_DBUS_SEND_MESSAGE_FLAGS_NONE,
-1,
NULL, /* out_serial */
NULL, /* cancellable */
&error);
g_assert_no_error (error);
2010-05-06 23:31:51 +02:00
g_assert (g_dbus_message_get_message_type (method_reply_message) == G_DBUS_MESSAGE_TYPE_METHOD_RETURN);
fd_list = g_dbus_message_get_unix_fd_list (method_reply_message);
g_assert (fd_list != NULL);
g_assert_cmpint (g_unix_fd_list_get_length (fd_list), ==, 1);
error = NULL;
fd = g_unix_fd_list_get (fd_list, 0, &error);
g_assert_no_error (error);
g_object_unref (method_call_message);
g_object_unref (method_reply_message);
error = NULL;
buf = read_all_from_fd (fd, &len, &error);
g_assert_no_error (error);
g_assert (buf != NULL);
close (fd);
error = NULL;
g_file_get_contents ("/etc/hosts",
&buf2,
&len2,
&error);
g_assert_no_error (error);
g_assert_cmpint (len, ==, len2);
g_assert (memcmp (buf, buf2, len) == 0);
g_free (buf2);
g_free (buf);
}
#else
error = NULL;
result = g_dbus_proxy_call_sync (proxy,
"OpenFile",
g_variant_new ("(s)", "boo"),
G_DBUS_CALL_FLAGS_NONE,
-1,
NULL, /* GCancellable */
&error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_DBUS_ERROR);
g_assert (result == NULL);
g_error_free (error);
#endif /* G_OS_UNIX */
/* Check that g_socket_get_credentials() work - this really should
* be in a GSocket-specific test suite but no such test suite exists
* right now.
*/
{
GSocket *socket;
GCredentials *credentials;
socket = g_socket_connection_get_socket (G_SOCKET_CONNECTION (g_dbus_connection_get_stream (c)));
g_assert (G_IS_SOCKET (socket));
error = NULL;
credentials = g_socket_get_credentials (socket, &error);
#ifdef __linux__
{
struct ucred *native_creds;
g_assert_no_error (error);
g_assert (G_IS_CREDENTIALS (credentials));
native_creds = g_credentials_get_native (credentials, G_CREDENTIALS_TYPE_LINUX_UCRED);
g_assert (native_creds != NULL);
g_assert (native_creds->uid == getuid ());
g_assert (native_creds->gid == getgid ());
g_assert (native_creds->pid == getpid ());
}
g_object_unref (credentials);
#else
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_NOT_SUPPORTED);
g_assert (credentials == NULL);
#endif
}
/* bring up a connection - don't accept it - this should fail
*/
data.accept_connection = FALSE;
error = NULL;
c2 = g_dbus_connection_new_for_address_sync (g_dbus_server_get_client_address (server),
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
_g_assert_error_domain (error, G_IO_ERROR);
g_assert (c2 == NULL);
#if 0
/* TODO: THIS TEST DOESN'T WORK YET */
/* bring up a connection - accept it.. then disconnect from the client side - check
* that the server side gets the disconnect signal.
*/
error = NULL;
data.accept_connection = TRUE;
c2 = g_dbus_connection_new_for_address_sync (g_dbus_server_get_client_address (server),
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_assert (c2 != NULL);
g_assert (!g_dbus_connection_get_is_disconnected (c2));
while (data.num_connection_attempts < 3)
g_main_loop_run (loop);
g_assert_cmpint (data.current_connections->len, ==, 2);
g_assert_cmpint (data.num_connection_attempts, ==, 3);
g_assert (!g_dbus_connection_get_is_disconnected (G_DBUS_CONNECTION (data.current_connections->pdata[1])));
g_idle_add (on_do_disconnect_in_idle, c2);
g_debug ("==================================================");
g_debug ("==================================================");
g_debug ("==================================================");
g_debug ("waiting for disconnect on connection %p, stream %p",
data.current_connections->pdata[1],
g_dbus_connection_get_stream (data.current_connections->pdata[1]));
g_timeout_add (2000, check_connection, &data);
//_g_assert_signal_received (G_DBUS_CONNECTION (data.current_connections->pdata[1]), "closed");
g_main_loop_run (loop);
g_assert (g_dbus_connection_get_is_disconnected (G_DBUS_CONNECTION (data.current_connections->pdata[1])));
g_ptr_array_set_size (data.current_connections, 1); /* remove disconnected connection object */
#endif
/* unref the server and stop listening for new connections
*
* This won't bring down the established connections - check that c is still connected
* by invoking a method
*/
//g_socket_service_stop (service);
//g_object_unref (service);
g_dbus_server_stop (server);
g_object_unref (server);
server = NULL;
error = NULL;
result = g_dbus_proxy_call_sync (proxy,
"HelloPeer",
g_variant_new ("(s)", "Hey Again Peer!"),
G_DBUS_CALL_FLAGS_NONE,
-1,
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_variant_get (result, "(&s)", &s);
g_assert_cmpstr (s, ==, "You greeted me with 'Hey Again Peer!'.");
g_variant_unref (result);
g_assert_cmpint (data.num_method_calls, ==, 5);
#if 0
/* TODO: THIS TEST DOESN'T WORK YET */
/* now disconnect from the server side - check that the client side gets the signal */
g_assert_cmpint (data.current_connections->len, ==, 1);
g_assert (G_DBUS_CONNECTION (data.current_connections->pdata[0]) != c);
g_dbus_connection_disconnect (G_DBUS_CONNECTION (data.current_connections->pdata[0]));
if (!g_dbus_connection_get_is_disconnected (c))
_g_assert_signal_received (c, "closed");
g_assert (g_dbus_connection_get_is_disconnected (c));
#endif
g_object_unref (c);
g_ptr_array_unref (data.current_connections);
g_object_unref (proxy);
g_main_loop_quit (service_loop);
g_thread_join (service_thread);
}
/* ---------------------------------------------------------------------------------------------------- */
typedef struct
{
GDBusServer *server;
GMainContext *context;
GMainLoop *loop;
GList *connections;
} DmpData;
static void
dmp_data_free (DmpData *data)
{
g_main_loop_unref (data->loop);
g_main_context_unref (data->context);
g_object_unref (data->server);
g_list_foreach (data->connections, (GFunc) g_object_unref, NULL);
g_list_free (data->connections);
g_free (data);
}
static void
dmp_on_method_call (GDBusConnection *connection,
const gchar *sender,
const gchar *object_path,
const gchar *interface_name,
const gchar *method_name,
GVariant *parameters,
GDBusMethodInvocation *invocation,
gpointer user_data)
{
//DmpData *data = user_data;
gint32 first;
gint32 second;
g_variant_get (parameters,
"(ii)",
&first,
&second);
g_dbus_method_invocation_return_value (invocation,
g_variant_new ("(i)", first + second));
}
static const GDBusInterfaceVTable dmp_interface_vtable =
{
dmp_on_method_call,
NULL, /* get_property */
NULL /* set_property */
};
/* Runs in thread we created GDBusServer in (since we didn't pass G_DBUS_SERVER_FLAGS_RUN_IN_THREAD) */
static void
dmp_on_new_connection (GDBusServer *server,
GDBusConnection *connection,
gpointer user_data)
{
DmpData *data = user_data;
GDBusNodeInfo *node;
GError *error;
/* accept the connection */
data->connections = g_list_prepend (data->connections, g_object_ref (connection));
error = NULL;
node = g_dbus_node_info_new_for_xml ("<node>"
" <interface name='org.gtk.GDBus.DmpInterface'>"
" <method name='AddPair'>"
" <arg type='i' name='first' direction='in'/>"
" <arg type='i' name='second' direction='in'/>"
" <arg type='i' name='sum' direction='out'/>"
" </method>"
" </interface>"
"</node>",
&error);
g_assert_no_error (error);
/* sleep 100ms before exporting an object - this is to test that
* G_DBUS_CONNECTION_FLAGS_DELAY_MESSAGE_PROCESSING really works
* (GDBusServer uses this feature).
*/
usleep (100 * 1000);
/* export an object */
error = NULL;
g_dbus_connection_register_object (connection,
"/dmp/test",
node->interfaces[0],
&dmp_interface_vtable,
data,
NULL,
&error);
g_dbus_node_info_unref (node);
}
static gpointer
dmp_thread_func (gpointer user_data)
{
DmpData *data = user_data;
GError *error;
gchar *guid;
data->context = g_main_context_new ();
g_main_context_push_thread_default (data->context);
error = NULL;
guid = g_dbus_generate_guid ();
data->server = g_dbus_server_new_sync ("unix:tmpdir=/tmp/gdbus-test-",
G_DBUS_SERVER_FLAGS_NONE,
guid,
NULL, /* GDBusAuthObserver */
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_signal_connect (data->server,
"new-connection",
G_CALLBACK (dmp_on_new_connection),
data);
g_dbus_server_start (data->server);
data->loop = g_main_loop_new (data->context, FALSE);
g_main_loop_run (data->loop);
g_main_context_pop_thread_default (data->context);
g_free (guid);
return NULL;
}
static void
delayed_message_processing (void)
{
GError *error;
DmpData *data;
GThread *service_thread;
guint n;
data = g_new0 (DmpData, 1);
error = NULL;
service_thread = g_thread_create (dmp_thread_func,
data,
TRUE,
&error);
while (data->server == NULL || !g_dbus_server_is_active (data->server))
g_thread_yield ();
for (n = 0; n < 5; n++)
{
GDBusConnection *c;
GVariant *res;
gint32 val;
error = NULL;
c = g_dbus_connection_new_for_address_sync (g_dbus_server_get_client_address (data->server),
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
error = NULL;
res = g_dbus_connection_call_sync (c,
NULL, /* bus name */
"/dmp/test",
"org.gtk.GDBus.DmpInterface",
"AddPair",
g_variant_new ("(ii)", 2, n),
G_VARIANT_TYPE ("(i)"),
G_DBUS_CALL_FLAGS_NONE,
-1, /* timeout_msec */
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_variant_get (res, "(i)", &val);
g_assert_cmpint (val, ==, 2 + n);
g_variant_unref (res);
g_object_unref (c);
}
g_main_loop_quit (data->loop);
g_thread_join (service_thread);
dmp_data_free (data);
}
/* ---------------------------------------------------------------------------------------------------- */
static gboolean
nonce_tcp_on_authorize_authenticated_peer (GDBusAuthObserver *observer,
GIOStream *stream,
GCredentials *credentials,
gpointer user_data)
{
PeerData *data = user_data;
gboolean authorized;
data->num_connection_attempts++;
authorized = TRUE;
if (!data->accept_connection)
{
authorized = FALSE;
g_main_loop_quit (loop);
}
return authorized;
}
/* Runs in thread we created GDBusServer in (since we didn't pass G_DBUS_SERVER_FLAGS_RUN_IN_THREAD) */
static void
nonce_tcp_on_new_connection (GDBusServer *server,
GDBusConnection *connection,
gpointer user_data)
{
PeerData *data = user_data;
g_ptr_array_add (data->current_connections, g_object_ref (connection));
g_main_loop_quit (loop);
}
static gpointer
nonce_tcp_service_thread_func (gpointer user_data)
{
PeerData *data = user_data;
GMainContext *service_context;
GDBusAuthObserver *observer;
GError *error;
service_context = g_main_context_new ();
g_main_context_push_thread_default (service_context);
error = NULL;
observer = g_dbus_auth_observer_new ();
server = g_dbus_server_new_sync ("nonce-tcp:",
G_DBUS_SERVER_FLAGS_NONE,
test_guid,
observer,
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_signal_connect (server,
"new-connection",
G_CALLBACK (nonce_tcp_on_new_connection),
data);
g_signal_connect (observer,
"authorize-authenticated-peer",
G_CALLBACK (nonce_tcp_on_authorize_authenticated_peer),
data);
g_object_unref (observer);
g_dbus_server_start (server);
service_loop = g_main_loop_new (service_context, FALSE);
g_main_loop_run (service_loop);
g_main_context_pop_thread_default (service_context);
g_main_loop_unref (service_loop);
g_main_context_unref (service_context);
/* test code specifically unrefs the server - see below */
g_assert (server == NULL);
return NULL;
}
static void
test_nonce_tcp (void)
{
PeerData data;
GError *error;
GThread *service_thread;
GDBusConnection *c;
gchar *s;
gchar *nonce_file;
gboolean res;
const gchar *address;
memset (&data, '\0', sizeof (PeerData));
data.current_connections = g_ptr_array_new_with_free_func (g_object_unref);
error = NULL;
server = NULL;
service_loop = NULL;
service_thread = g_thread_create (nonce_tcp_service_thread_func,
&data,
TRUE,
&error);
while (service_loop == NULL)
g_thread_yield ();
g_assert (server != NULL);
/* bring up a connection and accept it */
data.accept_connection = TRUE;
error = NULL;
c = g_dbus_connection_new_for_address_sync (g_dbus_server_get_client_address (server),
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
g_assert_no_error (error);
g_assert (c != NULL);
while (data.current_connections->len < 1)
g_main_loop_run (loop);
g_assert_cmpint (data.current_connections->len, ==, 1);
g_assert_cmpint (data.num_connection_attempts, ==, 1);
g_assert (g_dbus_connection_get_unique_name (c) == NULL);
g_assert_cmpstr (g_dbus_connection_get_guid (c), ==, test_guid);
g_object_unref (c);
/* now, try to subvert the nonce file (this assumes noncefile is the last key/value pair)
*/
address = g_dbus_server_get_client_address (server);
s = strstr (address, "noncefile=");
g_assert (s != NULL);
s += sizeof "noncefile=" - 1;
nonce_file = g_strdup (s);
/* First try invalid data in the nonce file - this will actually
* make the client send this and the server will reject it. The way
* it works is that if the nonce doesn't match, the server will
* simply close the connection. So, from the client point of view,
* we can see a variety of errors.
*/
error = NULL;
res = g_file_set_contents (nonce_file,
"0123456789012345",
-1,
&error);
g_assert_no_error (error);
g_assert (res);
c = g_dbus_connection_new_for_address_sync (address,
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
_g_assert_error_domain (error, G_IO_ERROR);
g_assert (c == NULL);
/* Then try with a nonce-file of incorrect length - this will make
* the client complain - we won't even try connecting to the server
* for this
*/
error = NULL;
res = g_file_set_contents (nonce_file,
"0123456789012345_",
-1,
&error);
g_assert_no_error (error);
g_assert (res);
c = g_dbus_connection_new_for_address_sync (address,
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_INVALID_ARGUMENT);
g_assert (c == NULL);
/* Finally try with no nonce-file at all */
g_assert_cmpint (g_unlink (nonce_file), ==, 0);
error = NULL;
c = g_dbus_connection_new_for_address_sync (address,
G_DBUS_CONNECTION_FLAGS_AUTHENTICATION_CLIENT,
NULL, /* GDBusAuthObserver */
NULL, /* cancellable */
&error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_INVALID_ARGUMENT);
g_assert (c == NULL);
g_free (nonce_file);
g_dbus_server_stop (server);
g_object_unref (server);
server = NULL;
g_main_loop_quit (service_loop);
g_thread_join (service_thread);
}
2010-07-29 08:20:29 +02:00
static void
test_credentials (void)
{
GCredentials *c1, *c2;
GError *error;
gchar *desc;
c1 = g_credentials_new ();
c2 = g_credentials_new ();
error = NULL;
if (g_credentials_set_unix_user (c2, getuid (), &error))
g_assert_no_error (error);
g_clear_error (&error);
g_assert (g_credentials_is_same_user (c1, c2, &error));
g_assert_no_error (error);
desc = g_credentials_to_string (c1);
g_assert (desc != NULL);
g_free (desc);
g_object_unref (c1);
g_object_unref (c2);
}
/* ---------------------------------------------------------------------------------------------------- */
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
#ifdef G_OS_UNIX
/* Chosen to be big enough to overflow the socket buffer */
#define OVERFLOW_NUM_SIGNALS 5000
#define OVERFLOW_TIMEOUT_SEC 10
static GDBusMessageFilterResult
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
overflow_filter_func (GDBusConnection *connection,
GDBusMessage *message,
gboolean incoming,
gpointer user_data)
{
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
volatile gint *counter = user_data;
*counter += 1;
return G_DBUS_MESSAGE_FILTER_RESULT_NO_EFFECT;
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
}
static gboolean
overflow_on_500ms_later_func (gpointer user_data)
{
g_main_loop_quit (loop);
return FALSE; /* don't keep the idle */
}
static void
test_overflow (void)
{
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
gint sv[2];
gint n;
GSocket *socket;
GSocketConnection *socket_connection;
GDBusConnection *producer, *consumer;
GError *error;
GTimer *timer;
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
volatile gint n_messages_received;
volatile gint n_messages_sent;
g_assert_cmpint (socketpair (AF_UNIX, SOCK_STREAM, 0, sv), ==, 0);
error = NULL;
socket = g_socket_new_from_fd (sv[0], &error);
g_assert_no_error (error);
socket_connection = g_socket_connection_factory_create_connection (socket);
g_assert (socket_connection != NULL);
g_object_unref (socket);
producer = g_dbus_connection_new_sync (G_IO_STREAM (socket_connection),
NULL, /* guid */
G_DBUS_CONNECTION_FLAGS_NONE,
NULL, /* GDBusAuthObserver */
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
NULL, /* GCancellable */
&error);
g_dbus_connection_set_exit_on_close (producer, TRUE);
g_assert_no_error (error);
g_object_unref (socket_connection);
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
n_messages_sent = 0;
g_dbus_connection_add_filter (producer, overflow_filter_func, (gpointer) &n_messages_sent, NULL);
/* send enough data that we get an EAGAIN */
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
for (n = 0; n < OVERFLOW_NUM_SIGNALS; n++)
{
error = NULL;
g_dbus_connection_emit_signal (producer,
NULL, /* destination */
"/org/foo/Object",
"org.foo.Interface",
"Member",
g_variant_new ("(s)", "a string"),
&error);
g_assert_no_error (error);
}
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
/* sleep for 0.5 sec (to allow the GDBus IO thread to fill up the
* kernel buffers) and verify that n_messages_sent <
* OVERFLOW_NUM_SIGNALS
*
* This is to verify that not all the submitted messages have been
* sent to the underlying transport.
*/
g_timeout_add (500, overflow_on_500ms_later_func, NULL);
g_main_loop_run (loop);
g_assert_cmpint (n_messages_sent, <, OVERFLOW_NUM_SIGNALS);
/* now suck it all out as a client, and add it up */
socket = g_socket_new_from_fd (sv[1], &error);
g_assert_no_error (error);
socket_connection = g_socket_connection_factory_create_connection (socket);
g_assert (socket_connection != NULL);
g_object_unref (socket);
consumer = g_dbus_connection_new_sync (G_IO_STREAM (socket_connection),
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
NULL, /* guid */
G_DBUS_CONNECTION_FLAGS_DELAY_MESSAGE_PROCESSING,
NULL, /* GDBusAuthObserver */
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
NULL, /* GCancellable */
&error);
g_assert_no_error (error);
g_object_unref (socket_connection);
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
n_messages_received = 0;
g_dbus_connection_add_filter (consumer, overflow_filter_func, (gpointer) &n_messages_received, NULL);
g_dbus_connection_start_message_processing (consumer);
timer = g_timer_new ();
g_timer_start (timer);
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
while (n_messages_received < OVERFLOW_NUM_SIGNALS && g_timer_elapsed (timer, NULL) < OVERFLOW_TIMEOUT_SEC)
g_main_context_iteration (NULL, FALSE);
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
g_assert_cmpint (n_messages_sent, ==, OVERFLOW_NUM_SIGNALS);
g_assert_cmpint (n_messages_received, ==, OVERFLOW_NUM_SIGNALS);
g_timer_destroy (timer);
g_object_unref (consumer);
g_object_unref (producer);
}
#else
static void
test_overflow (void)
{
Bug 626748 – Use async methods for writing and handle EAGAIN If sending a lot of data and/or the other peer is not reading it, then socket buffers can overflow. This is communicated from the kernel by returning EAGAIN. In GIO, it is modelled by g_output_stream_write() and g_socket_send_message() returning G_IO_ERROR_WOULD_BLOCK. It is also problematic that that we're using synchronous IO in the shared GDBus IO thread. It means that one GDBusConnection can lock up others. It turns out that by porting from g_output_stream_write() to g_output_stream_write_async() we fix the EAGAIN issue. For GSocket, we still need to handle things manually (by creating a GSource) as g_socket_send_message() is used. We check the new behavior in Michael's producer/consumer test case (at /gdbus/overflow in gdbus-peer.c) added in the last commit. Also add a test case that sends and receives a 20 MiB message. Also add a new `transport' G_DBUS_DEBUG option so it is easy to inspect partial writes: $ G_DBUS_DEBUG=transport ./gdbus-connection -p /gdbus/connection/large_message [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 0 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 128000 on a GSocketOutputStream ======================================================================== GDBus-debug:Transport: >>>> WROTE 128000 bytes of message with serial 4 and size 20971669 from offset 256000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: >>>> WROTE 43669 bytes of message with serial 4 and size 20971669 from offset 20928000 on a GSocketOutputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 16 bytes of message with serial 3 and size 20971620 to offset 0 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 15984 bytes of message with serial 3 and size 20971620 to offset 16 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 16000 bytes of message with serial 3 and size 20971620 to offset 16000 from a GSocketInputStream [...] ======================================================================== GDBus-debug:Transport: <<<< READ 144000 bytes of message with serial 3 and size 20971620 to offset 20720000 from a GSocketInputStream ======================================================================== GDBus-debug:Transport: <<<< READ 107620 bytes of message with serial 3 and size 20971620 to offset 20864000 from a GSocketInputStream OK https://bugzilla.gnome.org/show_bug.cgi?id=626748 Signed-off-by: David Zeuthen <davidz@redhat.com>
2010-08-16 19:43:35 +02:00
/* TODO: test this with e.g. GWin32InputStream/GWin32OutputStream */
}
#endif
/* ---------------------------------------------------------------------------------------------------- */
int
main (int argc,
char *argv[])
{
gint ret;
GDBusNodeInfo *introspection_data = NULL;
g_type_init ();
g_thread_init (NULL);
g_test_init (&argc, &argv, NULL);
introspection_data = g_dbus_node_info_new_for_xml (test_interface_introspection_xml, NULL);
g_assert (introspection_data != NULL);
test_interface_introspection_data = introspection_data->interfaces[0];
test_guid = g_dbus_generate_guid ();
/* all the tests rely on a shared main loop */
loop = g_main_loop_new (NULL, FALSE);
g_test_add_func ("/gdbus/peer-to-peer", test_peer);
g_test_add_func ("/gdbus/delayed-message-processing", delayed_message_processing);
g_test_add_func ("/gdbus/nonce-tcp", test_nonce_tcp);
2010-07-29 08:20:29 +02:00
g_test_add_func ("/gdbus/credentials", test_credentials);
g_test_add_func ("/gdbus/overflow", test_overflow);
ret = g_test_run();
g_main_loop_unref (loop);
g_free (test_guid);
g_dbus_node_info_unref (introspection_data);
return ret;
}