efl/src/examples/ecore/efl_net_server_simple_examp...

575 lines
21 KiB
C
Raw Normal View History

#define EFL_BETA_API_SUPPORT
#include <Efl_Net.h>
#include <Ecore_Getopt.h>
#include <fcntl.h>
static Eina_Bool echo = EINA_FALSE;
static double timeout = 10.0;
/* NOTE: client i/o events are only used as debug, you can omit these */
static void
_client_can_read_changed(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: client %s can_read=%d\n",
efl_net_socket_address_remote_get(event->object),
efl_io_reader_can_read_get(event->object));
}
static void
_client_can_write_changed(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: client %s can_write=%d\n",
efl_net_socket_address_remote_get(event->object),
efl_io_writer_can_write_get(event->object));
}
static void
_client_eos(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: client %s eos.\n",
efl_net_socket_address_remote_get(event->object));
}
static void
_client_read_finished(void *data EINA_UNUSED, const Efl_Event *event)
{
/* on _error() we close it, then do not read as it has nothing */
if (efl_io_closer_closed_get(event->object))
return;
if (echo) return;
fprintf(stderr,
"-- BEGIN RECEIVED DATA --\n"
EINA_SLICE_STR_FMT
"-- END RECEIVED DATA--\n",
EINA_SLICE_STR_PRINT(efl_io_buffered_stream_slice_get(event->object)));
}
static void
_client_closed(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: client %s closed.\n",
efl_net_socket_address_remote_get(event->object));
}
/* this is the only event that matters, from here we remove our extra
* reference from the client and let it be deleted.
*/
static void
_client_finished(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: client %s finished sending and receiving, remove extra reference.\n",
efl_net_socket_address_remote_get(event->object));
if (!efl_io_closer_closed_get(event->object))
efl_io_closer_close(event->object);
efl_unref(event->object);
}
/*
* On errors, such as ETIMEDOUT, we want to close the client if not
* happened yet.
*/
static void
_client_error(void *data EINA_UNUSED, const Efl_Event *event)
{
Eina_Error *perr = event->info;
fprintf(stderr, "ERROR: client %s error: %s\n",
efl_net_socket_address_remote_get(event->object),
eina_error_msg_get(*perr));
if (!efl_io_closer_closed_get(event->object))
efl_io_closer_close(event->object);
}
EFL_CALLBACKS_ARRAY_DEFINE(client_cbs,
{ EFL_IO_READER_EVENT_CAN_READ_CHANGED, _client_can_read_changed },
{ EFL_IO_READER_EVENT_EOS, _client_eos },
{ EFL_IO_WRITER_EVENT_CAN_WRITE_CHANGED, _client_can_write_changed },
{ EFL_IO_CLOSER_EVENT_CLOSED, _client_closed },
{ EFL_IO_BUFFERED_STREAM_EVENT_READ_FINISHED, _client_read_finished },
{ EFL_IO_BUFFERED_STREAM_EVENT_FINISHED, _client_finished },
{ EFL_IO_BUFFERED_STREAM_EVENT_ERROR, _client_error });
/* copier events are of interest, you should hook to at least "done"
* and "error"
*/
/* echo copier is about the same socket, you can close it right away */
static void
_echo_copier_done(void *data EINA_UNUSED, const Efl_Event *event)
{
Eo *copier = event->object;
fprintf(stderr, "INFO: echo copier done, close and del %p\n", copier);
efl_del(copier); /* set to close_on_destructor, will auto close copier and client */
}
static void
_echo_copier_error(void *data EINA_UNUSED, const Efl_Event *event)
{
Eo *copier = event->object;
const Eina_Error *perr = event->info;
if (*perr == ETIMEDOUT)
{
Eo *client = efl_io_copier_source_get(copier);
fprintf(stderr, "INFO: client '%s' timed out, delete it.\n",
efl_net_socket_address_remote_get(client));
efl_del(copier);
return;
}
efl_loop_quit(efl_loop_get(event->object), eina_value_int_init(EXIT_FAILURE));
fprintf(stderr, "ERROR: echo copier %p failed %d '%s', close and del.\n",
copier, *perr, eina_error_msg_get(*perr));
efl_del(copier);
}
EFL_CALLBACKS_ARRAY_DEFINE(echo_copier_cbs,
{ EFL_IO_COPIER_EVENT_DONE, _echo_copier_done },
{ EFL_IO_COPIER_EVENT_ERROR, _echo_copier_error});
/* server events are mandatory, afterall you need to define what's
* going to happen after a client socket is connected. This is the
* "client,add" event.
*
* if clients_limit and clients_reject_excess are set, then
* "client,rejected" is dispatched for rejected sockets, they contain
* the string with socket identification.
*/
static void
_server_client_add(void *data EINA_UNUSED, const Efl_Event *event)
{
Efl_Net_Socket *client = event->info;
fprintf(stderr, "INFO: accepted client %s\n",
efl_net_socket_address_remote_get(client));
/* to use a client, you must efl_ref() it. */
efl_ref(client);
/*
* monitor the client if it's done and for debug purposes
* (optional)
*/
efl_event_callback_array_add(client, client_cbs(), NULL);
efl_io_buffered_stream_timeout_inactivity_set(client, timeout);
/*
* Since sockets are reader/writer/closer objects, we can use the
* Efl_Io_Copier utility.
*/
if (echo)
{
/*
* An echo copier is pretty simple, use the socket as both
* source and destination.
*
* This is the same as efl_net_server_example.c
*/
Eo *echo_copier = efl_add(EFL_IO_COPIER_CLASS, efl_parent_get(client),
efl_io_copier_source_set(efl_added, client),
efl_io_copier_destination_set(efl_added, client),
efl_event_callback_array_add(efl_added, echo_copier_cbs(), client),
efl_io_closer_close_on_invalidate_set(efl_added, EINA_TRUE) /* we want to auto-close as we have a single copier */
);
fprintf(stderr, "INFO: using an echo copier=%p for client %s\n",
echo_copier, efl_net_socket_address_remote_get(client));
return;
}
else
{
/*
* Here is where the "simple" kicks in, instead of all the
* complexity listed in efl_net_server_example.c, we just
* "write & forget" the "Hello World!" and wait for all data
* to be received with a simple "finished" event.
*/
Eina_Slice slice = EINA_SLICE_STR_LITERAL("Hello World!");
efl_io_writer_write(client, &slice, NULL);
efl_io_buffered_stream_eos_mark(client); /* say that's it */
}
}
static void
_server_client_rejected(void *data EINA_UNUSED, const Efl_Event *event)
{
const char *client_address = event->info;
fprintf(stderr, "INFO: rejected client %s\n", client_address);
}
static void
_server_error(void *data EINA_UNUSED, const Efl_Event *event)
{
const Eina_Error *perr = event->info;
fprintf(stderr, "ERROR: %d '%s'\n", *perr, eina_error_msg_get(*perr));
efl_loop_quit(efl_loop_get(event->object), eina_value_int_init(EXIT_FAILURE));
}
static void
_server_serving(void *data EINA_UNUSED, const Efl_Event *event)
{
fprintf(stderr, "INFO: serving at %s\n",
efl_net_server_address_get(event->object));
if (efl_class_get(event->object) == EFL_NET_SERVER_TCP_CLASS)
{
fprintf(stderr,
"TCP options:\n"
" - IPv6 only: %u\n",
efl_net_server_ip_ipv6_only_get(event->object));
}
else if (efl_class_get(event->object) == EFL_NET_SERVER_UDP_CLASS)
{
Eina_Iterator *it;
const char *str;
fprintf(stderr,
"UDP options:\n"
" - IPv6 only: %u\n"
" - don't route: %u\n"
" - multicast TTL: %u\n"
" - multicast loopback: %u\n"
" - multicast groups:\n",
efl_net_server_ip_ipv6_only_get(event->object),
efl_net_server_udp_dont_route_get(event->object),
efl_net_server_udp_multicast_time_to_live_get(event->object),
efl_net_server_udp_multicast_loopback_get(event->object));
it = efl_net_server_udp_multicast_groups_get(event->object);
EINA_ITERATOR_FOREACH(it, str)
fprintf(stderr, " * %s\n", str);
eina_iterator_free(it);
}
}
EFL_CALLBACKS_ARRAY_DEFINE(server_cbs,
{ EFL_NET_SERVER_EVENT_CLIENT_ADD, _server_client_add },
{ EFL_NET_SERVER_EVENT_CLIENT_REJECTED, _server_client_rejected },
{ EFL_NET_SERVER_EVENT_SERVER_ERROR, _server_error },
{ EFL_NET_SERVER_EVENT_SERVING, _server_serving });
static const char * protocols[] = {
"tcp",
"udp",
"ssl",
implement efl_net_{socket,dialer,server}_windows This is the local socket for windows, analogous to AF_UNIX. `Efl_Net_Socket_Windows` is the base class doing `ReadFile()` and `WriteFile()` using overlapped I/O, as well as the close procedure (`FlushFileBuffers()`, `DisconnectNamedPipe()` and `CloseHandle()`). These are done on top of an existing HANDLE that is set by `Efl_Net_Dialer_Windows` (from `CreateFile()`) or `Efl_Net_Server_Windows` (from `CreateNamedPipe()`). The overlapped I/O will return immediately, either with operation completed or `ERROR_IO_PENDING`, which means the kernel will execute that asynchronously and will later `SetEvent(overlapped.hEvent)` which is an event we wait on our main loop. That `overlapped` handle must exist during the call lifetime, thus cannot be bound to `pd`, as we may call `CancelIo()` but there is no guarantee the memory won't be touched, in that case we keep the overlapped around, but without an associated object. Windows provides no notification "can read without blocking" or non-blocking calls that returns partial data. The way to go is to use these overlapped I/O, with an initial `ReadFile()` to an internal buffer, once that operation finishes, we callback the user to says there is something to read (`efl_io_reader_can_read_set()`) and wait until `efl_io_reader_read()` is called to consume the available data, then `ReadFile()` is called again to read more data to the same internal buffer. Likewise, there is no "can write without blocking" or non-blocking calls that sends only partial data. The way to go is to get user bytes in `efl_io_writer_write()` and copy them in an internal buffer, then call `WriteFile()` on that and inform the user nothing else can be written until that operation completes (`efl_io_writer_can_write_set()`). This is cumbersome since we say we "sent" stuff when we actually didn't, it's still in our internal buffer (`pd->send.bytes`), but nonetheless the kernel and the other peer may be adding even more buffers, in this case we need to do a best effort to get it delivery. A particular case is troublesome: `write() -> close()`, this may result in `WriteFile()` pending, in this case we wait using `GetOverlappedResult()`, *this is nasty and may block*, but it's the only way I see to cope with such common use case. Other operations, like ongoing `ReadFile()` or `ConnectNamedPipe()` will be canceled using `CancelIo()`. Q: Why no I/O Completion Port (IOCP) was used? Why no CreateThreadpoolIo()? These perform much better! A: These will call back from secondary threads, but in EFL we must report back to the user in order to process incoming data or get more data to send. That is, we serialize everything to the main thread, making it impossible to use the benefits of IOCP and similar such as CreateThreadpoolIo(). Since we'd need to wakeup the main thread anyways, using `OVERLAPPED.hEvent` with `ecore_main_win32_handler_add()` does the job as we expect. Thanks to Vincent Torri (vtorri) for his help getting this code done with an example on how to do the NamedPipe handling on Windows.
2017-03-22 00:29:16 -07:00
#ifdef EFL_NET_SERVER_WINDOWS_CLASS
"windows",
#endif
#ifdef EFL_NET_SERVER_UNIX_CLASS
"unix",
#endif
NULL
};
static const char *ciphers_strs[] = {
"auto",
"tlsv1",
"tlsv1.1",
"tlsv1.2",
NULL
};
static const Ecore_Getopt options = {
"efl_net_server_example", /* program name */
NULL, /* usage line */
"1", /* version */
"(C) 2016 Enlightenment Project", /* copyright */
"BSD 2-Clause", /* license */
/* long description, may be multiline and contain \n */
"Example of Efl_Net_Server objects usage.\n"
"\n"
"This example spawns a server of the given protocol at the given address.",
EINA_FALSE,
{
ECORE_GETOPT_STORE_TRUE('e', "echo",
"Behave as 'echo' server, send back to client all the data receive"),
ECORE_GETOPT_STORE_TRUE(0, "socket-activated",
"Try to use $LISTEN_FDS from systemd, if not do a regular serve()"),
ECORE_GETOPT_STORE_UINT('l', "clients-limit",
"If set will limit number of clients to accept"),
ECORE_GETOPT_STORE_TRUE('r', "clients-reject-excess",
"Immediately reject excess clients (over limit)"),
ECORE_GETOPT_STORE_FALSE(0, "ipv4-on-ipv6",
"IPv4 clients will be automatically converted into IPv6 and handled transparently."),
ECORE_GETOPT_STORE_DOUBLE('t', "inactivity-timeout",
"The timeout in seconds to disconnect a client. The timeout is restarted for each client when there is some activity. It's particularly useful for UDP where there is no disconnection event."),
ECORE_GETOPT_VERSION('V', "version"),
ECORE_GETOPT_COPYRIGHT('C', "copyright"),
ECORE_GETOPT_LICENSE('L', "license"),
ECORE_GETOPT_HELP('h', "help"),
ECORE_GETOPT_CATEGORY("udp", "UDP options"),
ECORE_GETOPT_STORE_TRUE(0, "udp-dont-route",
"If true, datagrams won't be routed using a gateway, being restricted to the local network."),
ECORE_GETOPT_STORE_UINT(0, "udp-multicast-ttl",
"Multicast time to live in number of hops from 0-255. Defaults to 1 (only local network)."),
ECORE_GETOPT_STORE_FALSE(0, "udp-multicast-noloopback",
"Disable multicast loopback."),
ECORE_GETOPT_APPEND('M', "udp-multicast-group", "Join a multicast group in the form 'IP@INTERFACE', with optional '@INTERFACE', where INTERFACE is the IP address of the interface to join the multicast.", ECORE_GETOPT_TYPE_STR),
ECORE_GETOPT_CATEGORY("ssl", "SSL options"),
ECORE_GETOPT_CHOICE('c', "ssl-cipher", "Cipher to use, defaults to 'auto'", ciphers_strs),
ECORE_GETOPT_APPEND(0, "ssl-certificate", "certificate path to use.", ECORE_GETOPT_TYPE_STR),
ECORE_GETOPT_APPEND(0, "ssl-private-key", "private key path to use.", ECORE_GETOPT_TYPE_STR),
ECORE_GETOPT_APPEND(0, "ssl-crl", "certificate revocation list to use.", ECORE_GETOPT_TYPE_STR),
ECORE_GETOPT_APPEND(0, "ssl-ca", "certificate authorities path to use.", ECORE_GETOPT_TYPE_STR),
ECORE_GETOPT_CHOICE_METAVAR(0, NULL, "The server protocol.", "protocol",
protocols),
ECORE_GETOPT_STORE_METAVAR_STR(0, NULL,
"The server address to listen, such as "
"IPv4:PORT, [IPv6]:PORT, Unix socket path...",
"address"),
ECORE_GETOPT_SENTINEL
}
};
static Eo *simple_server = NULL;
EAPI_MAIN void
efl_pause(void *data EINA_UNUSED,
const Efl_Event *ev EINA_UNUSED)
{
}
EAPI_MAIN void
efl_resume(void *data EINA_UNUSED,
const Efl_Event *ev EINA_UNUSED)
{
}
EAPI_MAIN void
efl_terminate(void *data EINA_UNUSED,
const Efl_Event *ev EINA_UNUSED)
{
/* FIXME: For the moment the main loop doesn't get
properly destroyed on shutdown which disallow
relying on parent destroying their children */
if (simple_server)
{
efl_del(simple_server);
simple_server = NULL;
}
fprintf(stderr, "INFO: main loop finished.\n");
}
EAPI_MAIN void
efl_main(void *data EINA_UNUSED,
const Efl_Event *ev)
{
const Efl_Class *cls;
char *protocol = NULL;
char *address = NULL;
Eina_List *udp_mcast_groups = NULL;
char *str;
unsigned int clients_limit = 0;
unsigned udp_mcast_ttl = 1;
Eina_Bool clients_reject_excess = EINA_FALSE;
Eina_Bool ipv6_only = EINA_TRUE;
Eina_Bool udp_dont_route = EINA_FALSE;
Eina_Bool udp_mcast_loopback = EINA_TRUE;
Eina_List *certificates = NULL;
Eina_List *private_keys = NULL;
Eina_List *crls = NULL;
Eina_List *cas = NULL;
char *cipher_choice = NULL;
Eina_Bool socket_activated = EINA_FALSE;
Eina_Bool quit_option = EINA_FALSE;
Ecore_Getopt_Value values[] = {
ECORE_GETOPT_VALUE_BOOL(echo),
ECORE_GETOPT_VALUE_BOOL(socket_activated),
ECORE_GETOPT_VALUE_UINT(clients_limit),
ECORE_GETOPT_VALUE_BOOL(clients_reject_excess),
ECORE_GETOPT_VALUE_BOOL(ipv6_only),
ECORE_GETOPT_VALUE_DOUBLE(timeout),
/* standard block to provide version, copyright, license and help */
ECORE_GETOPT_VALUE_BOOL(quit_option), /* -V/--version quits */
ECORE_GETOPT_VALUE_BOOL(quit_option), /* -C/--copyright quits */
ECORE_GETOPT_VALUE_BOOL(quit_option), /* -L/--license quits */
ECORE_GETOPT_VALUE_BOOL(quit_option), /* -h/--help quits */
ECORE_GETOPT_VALUE_BOOL(quit_option), /* category: udp */
ECORE_GETOPT_VALUE_BOOL(udp_dont_route),
ECORE_GETOPT_VALUE_UINT(udp_mcast_ttl),
ECORE_GETOPT_VALUE_BOOL(udp_mcast_loopback),
ECORE_GETOPT_VALUE_LIST(udp_mcast_groups),
ECORE_GETOPT_VALUE_BOOL(quit_option), /* category: ssl */
ECORE_GETOPT_VALUE_STR(cipher_choice),
ECORE_GETOPT_VALUE_LIST(certificates),
ECORE_GETOPT_VALUE_LIST(private_keys),
ECORE_GETOPT_VALUE_LIST(crls),
ECORE_GETOPT_VALUE_LIST(cas),
/* positional argument */
ECORE_GETOPT_VALUE_STR(protocol),
ECORE_GETOPT_VALUE_STR(address),
ECORE_GETOPT_VALUE_NONE /* sentinel */
};
int args;
Eo *server;
Eina_Error err;
args = ecore_getopt_parse(&options, values, 0, NULL);
if (args < 0)
{
fputs("ERROR: Could not parse command line options.\n", stderr);
goto end;
}
if (quit_option) goto end;
args = ecore_getopt_parse_positional(&options, values, 0, NULL, args);
if (args < 0)
{
fputs("ERROR: Could not parse positional arguments.\n", stderr);
goto end;
}
if (!protocol)
{
fputs("ERROR: missing protocol.\n", stderr);
goto end;
}
if (strcmp(protocol, "tcp") == 0) cls = EFL_NET_SERVER_TCP_CLASS;
else if (strcmp(protocol, "udp") == 0) cls = EFL_NET_SERVER_UDP_CLASS;
else if (strcmp(protocol, "ssl") == 0) cls = EFL_NET_SERVER_SSL_CLASS;
implement efl_net_{socket,dialer,server}_windows This is the local socket for windows, analogous to AF_UNIX. `Efl_Net_Socket_Windows` is the base class doing `ReadFile()` and `WriteFile()` using overlapped I/O, as well as the close procedure (`FlushFileBuffers()`, `DisconnectNamedPipe()` and `CloseHandle()`). These are done on top of an existing HANDLE that is set by `Efl_Net_Dialer_Windows` (from `CreateFile()`) or `Efl_Net_Server_Windows` (from `CreateNamedPipe()`). The overlapped I/O will return immediately, either with operation completed or `ERROR_IO_PENDING`, which means the kernel will execute that asynchronously and will later `SetEvent(overlapped.hEvent)` which is an event we wait on our main loop. That `overlapped` handle must exist during the call lifetime, thus cannot be bound to `pd`, as we may call `CancelIo()` but there is no guarantee the memory won't be touched, in that case we keep the overlapped around, but without an associated object. Windows provides no notification "can read without blocking" or non-blocking calls that returns partial data. The way to go is to use these overlapped I/O, with an initial `ReadFile()` to an internal buffer, once that operation finishes, we callback the user to says there is something to read (`efl_io_reader_can_read_set()`) and wait until `efl_io_reader_read()` is called to consume the available data, then `ReadFile()` is called again to read more data to the same internal buffer. Likewise, there is no "can write without blocking" or non-blocking calls that sends only partial data. The way to go is to get user bytes in `efl_io_writer_write()` and copy them in an internal buffer, then call `WriteFile()` on that and inform the user nothing else can be written until that operation completes (`efl_io_writer_can_write_set()`). This is cumbersome since we say we "sent" stuff when we actually didn't, it's still in our internal buffer (`pd->send.bytes`), but nonetheless the kernel and the other peer may be adding even more buffers, in this case we need to do a best effort to get it delivery. A particular case is troublesome: `write() -> close()`, this may result in `WriteFile()` pending, in this case we wait using `GetOverlappedResult()`, *this is nasty and may block*, but it's the only way I see to cope with such common use case. Other operations, like ongoing `ReadFile()` or `ConnectNamedPipe()` will be canceled using `CancelIo()`. Q: Why no I/O Completion Port (IOCP) was used? Why no CreateThreadpoolIo()? These perform much better! A: These will call back from secondary threads, but in EFL we must report back to the user in order to process incoming data or get more data to send. That is, we serialize everything to the main thread, making it impossible to use the benefits of IOCP and similar such as CreateThreadpoolIo(). Since we'd need to wakeup the main thread anyways, using `OVERLAPPED.hEvent` with `ecore_main_win32_handler_add()` does the job as we expect. Thanks to Vincent Torri (vtorri) for his help getting this code done with an example on how to do the NamedPipe handling on Windows.
2017-03-22 00:29:16 -07:00
#ifdef EFL_NET_SERVER_WINDOWS_CLASS
else if (strcmp(protocol, "windows") == 0) cls = EFL_NET_SERVER_WINDOWS_CLASS;
#endif
#ifdef EFL_NET_SERVER_UNIX_CLASS
else if (strcmp(protocol, "unix") == 0) cls = EFL_NET_SERVER_UNIX_CLASS;
#endif
else
{
fprintf(stderr, "ERROR: unsupported protocol: %s\n", protocol);
goto end;
}
simple_server = efl_add(EFL_NET_SERVER_SIMPLE_CLASS, ev->object, /* it's mandatory to use a main loop provider as the server parent */
efl_net_server_simple_inner_class_set(efl_added, cls), /* alternatively you could create the inner server and set with efl_net_server_simple_inner_server_set() */
efl_net_server_clients_limit_set(efl_added,
clients_limit,
clients_reject_excess), /* optional */
efl_event_callback_array_add(efl_added, server_cbs(), NULL)); /* mandatory to have "client,add" in order to be useful */
if (!simple_server)
{
fprintf(stderr, "ERROR: could not create simple server for class %p (%s)\n",
cls, efl_class_name_get(cls));
goto end;
}
/* get the inner server so we can configure it for each protocol */
server = efl_net_server_simple_inner_server_get(simple_server);
if (cls == EFL_NET_SERVER_TCP_CLASS)
{
efl_net_server_ip_ipv6_only_set(server, ipv6_only);
efl_net_server_fd_reuse_address_set(server, EINA_TRUE); /* optional, but nice for testing */
efl_net_server_fd_reuse_port_set(server, EINA_TRUE); /* optional, but nice for testing... not secure unless you know what you're doing */
if (socket_activated) efl_net_server_fd_socket_activate(server, address);
}
else if (cls == EFL_NET_SERVER_UDP_CLASS)
{
const Eina_List *lst;
efl_net_server_ip_ipv6_only_set(server, ipv6_only);
efl_net_server_udp_dont_route_set(server, udp_dont_route);
efl_net_server_udp_multicast_time_to_live_set(server, udp_mcast_ttl);
efl_net_server_udp_multicast_loopback_set(server, udp_mcast_loopback);
EINA_LIST_FOREACH(udp_mcast_groups, lst, str)
efl_net_server_udp_multicast_join(server, str);
efl_net_server_fd_reuse_address_set(server, EINA_TRUE); /* optional, but nice for testing */
efl_net_server_fd_reuse_port_set(server, EINA_TRUE); /* optional, but nice for testing... not secure unless you know what you're doing */
if (socket_activated) efl_net_server_fd_socket_activate(server, address);
}
else if (cls == EFL_NET_SERVER_SSL_CLASS)
{
Eo *ssl_ctx;
Efl_Net_Ssl_Cipher cipher = EFL_NET_SSL_CIPHER_AUTO;
if (cipher_choice)
{
if (strcmp(cipher_choice, "auto") == 0)
cipher = EFL_NET_SSL_CIPHER_AUTO;
else if (strcmp(cipher_choice, "tlsv1") == 0)
cipher = EFL_NET_SSL_CIPHER_TLSV1;
else if (strcmp(cipher_choice, "tlsv1.1") == 0)
cipher = EFL_NET_SSL_CIPHER_TLSV1_1;
else if (strcmp(cipher_choice, "tlsv1.2") == 0)
cipher = EFL_NET_SSL_CIPHER_TLSV1_2;
}
ssl_ctx = efl_add_ref(EFL_NET_SSL_CONTEXT_CLASS, NULL,
efl_net_ssl_context_certificates_set(efl_added, eina_list_iterator_new(certificates)),
efl_net_ssl_context_private_keys_set(efl_added, eina_list_iterator_new(private_keys)),
efl_net_ssl_context_certificate_revocation_lists_set(efl_added, eina_list_iterator_new(crls)),
efl_net_ssl_context_certificate_authorities_set(efl_added, eina_list_iterator_new(cas)),
efl_net_ssl_context_setup(efl_added, cipher, EINA_FALSE /* a server! */));
efl_net_server_ssl_context_set(server, ssl_ctx);
efl_net_server_fd_reuse_address_set(server, EINA_TRUE); /* optional, but nice for testing */
efl_net_server_fd_reuse_port_set(server, EINA_TRUE); /* optional, but nice for testing... not secure unless you know what you're doing */
if (socket_activated) efl_net_server_fd_socket_activate(server, address);
}
implement efl_net_{socket,dialer,server}_windows This is the local socket for windows, analogous to AF_UNIX. `Efl_Net_Socket_Windows` is the base class doing `ReadFile()` and `WriteFile()` using overlapped I/O, as well as the close procedure (`FlushFileBuffers()`, `DisconnectNamedPipe()` and `CloseHandle()`). These are done on top of an existing HANDLE that is set by `Efl_Net_Dialer_Windows` (from `CreateFile()`) or `Efl_Net_Server_Windows` (from `CreateNamedPipe()`). The overlapped I/O will return immediately, either with operation completed or `ERROR_IO_PENDING`, which means the kernel will execute that asynchronously and will later `SetEvent(overlapped.hEvent)` which is an event we wait on our main loop. That `overlapped` handle must exist during the call lifetime, thus cannot be bound to `pd`, as we may call `CancelIo()` but there is no guarantee the memory won't be touched, in that case we keep the overlapped around, but without an associated object. Windows provides no notification "can read without blocking" or non-blocking calls that returns partial data. The way to go is to use these overlapped I/O, with an initial `ReadFile()` to an internal buffer, once that operation finishes, we callback the user to says there is something to read (`efl_io_reader_can_read_set()`) and wait until `efl_io_reader_read()` is called to consume the available data, then `ReadFile()` is called again to read more data to the same internal buffer. Likewise, there is no "can write without blocking" or non-blocking calls that sends only partial data. The way to go is to get user bytes in `efl_io_writer_write()` and copy them in an internal buffer, then call `WriteFile()` on that and inform the user nothing else can be written until that operation completes (`efl_io_writer_can_write_set()`). This is cumbersome since we say we "sent" stuff when we actually didn't, it's still in our internal buffer (`pd->send.bytes`), but nonetheless the kernel and the other peer may be adding even more buffers, in this case we need to do a best effort to get it delivery. A particular case is troublesome: `write() -> close()`, this may result in `WriteFile()` pending, in this case we wait using `GetOverlappedResult()`, *this is nasty and may block*, but it's the only way I see to cope with such common use case. Other operations, like ongoing `ReadFile()` or `ConnectNamedPipe()` will be canceled using `CancelIo()`. Q: Why no I/O Completion Port (IOCP) was used? Why no CreateThreadpoolIo()? These perform much better! A: These will call back from secondary threads, but in EFL we must report back to the user in order to process incoming data or get more data to send. That is, we serialize everything to the main thread, making it impossible to use the benefits of IOCP and similar such as CreateThreadpoolIo(). Since we'd need to wakeup the main thread anyways, using `OVERLAPPED.hEvent` with `ecore_main_win32_handler_add()` does the job as we expect. Thanks to Vincent Torri (vtorri) for his help getting this code done with an example on how to do the NamedPipe handling on Windows.
2017-03-22 00:29:16 -07:00
#ifdef EFL_NET_SERVER_UNIX_CLASS
else if (cls == EFL_NET_SERVER_UNIX_CLASS)
{
efl_net_server_unix_unlink_before_bind_set(server, EINA_TRUE); /* makes testing easier */
if (socket_activated) efl_net_server_fd_socket_activate(server, address);
}
#endif
/* an explicit call to efl_net_server_serve() after the object is
* constructed allows for more complex setup, such as interacting
* with the object to add more properties that couldn't be done
* during efl_add().
*/
if (!efl_net_server_serving_get(simple_server))
{
if (socket_activated)
fprintf(stderr, "WARNING: --socket-activated, but not able to use $LISTEN_FDS descriptors. Try to start the server...\n");
err = efl_net_server_serve(simple_server, address);
if (err)
{
fprintf(stderr, "ERROR: could not serve(%s): %s\n",
address, eina_error_msg_get(err));
goto end_server;
}
}
return ;
end_server:
efl_del(simple_server);
simple_server = NULL;
end:
EINA_LIST_FREE(udp_mcast_groups, str)
free(str);
efl_loop_quit(ev->object, eina_value_int_init(EXIT_FAILURE));
}
EFL_MAIN_EX();