2011-02-07 12:59:50 +02:00
|
|
|
/*
|
2013-08-16 00:12:51 +03:00
|
|
|
* Copyright (c) 2007 The FFmpeg Project
|
2011-02-07 12:59:50 +02:00
|
|
|
*
|
2013-08-16 00:12:51 +03:00
|
|
|
* This file is part of FFmpeg.
|
2011-02-07 12:59:50 +02:00
|
|
|
*
|
2013-08-16 00:12:51 +03:00
|
|
|
* FFmpeg is free software; you can redistribute it and/or
|
2011-02-07 12:59:50 +02:00
|
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
|
|
|
* version 2.1 of the License, or (at your option) any later version.
|
|
|
|
*
|
2013-08-16 00:12:51 +03:00
|
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
2011-02-07 12:59:50 +02:00
|
|
|
* 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
|
2013-08-16 00:12:51 +03:00
|
|
|
* License along with FFmpeg; if not, write to the Free Software
|
2011-02-07 12:59:50 +02:00
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
|
|
*/
|
|
|
|
|
2013-08-08 22:26:42 +03:00
|
|
|
#include <fcntl.h>
|
2011-02-07 12:59:50 +02:00
|
|
|
#include "network.h"
|
2015-05-27 12:57:51 +02:00
|
|
|
#include "tls.h"
|
2013-05-31 04:05:13 +03:00
|
|
|
#include "url.h"
|
network: Add RFC 8305 style "Happy Eyeballs"/"Fast Fallback" helper function
For cases with dual stack (IPv4 + IPv6) connectivity, but where one
stack potentially is less reliable, strive to trying to connect over
both protocols in parallel, using whichever address connected first.
In cases with a hostname resolving to multiple IPv4 and IPv6
addresses, the current connection mechanism would try all addresses
in the order returned by getaddrinfo (with all IPv6 addresses ordered
before the IPv4 addresses normally). If connection attempts to the
IPv6 addresses return quickly with an error, this was no problem, but
if they were unsuccessful leading up to timeouts, the connection process
would have to wait for timeouts on all IPv6 target addresses before
attempting any IPv4 address.
Similar to what RFC 8305 suggests, reorder the list of addresses to
try connecting to, interleaving address families. After starting one
connection attempt, start another one in parallel after a small delay
(200 ms as suggested by the RFC).
For cases with unreliable IPv6 but reliable IPv4, this should make
connection attempts work as reliably as with plain IPv4, with only an
extra 200 ms of connection delay.
Signed-off-by: Martin Storsjö <martin@martin.st>
2018-08-10 09:25:59 +02:00
|
|
|
#include "libavutil/avassert.h"
|
2012-08-15 23:27:52 +03:00
|
|
|
#include "libavutil/mem.h"
|
2012-10-09 15:26:02 +03:00
|
|
|
#include "libavutil/time.h"
|
2011-02-05 00:25:07 +02:00
|
|
|
|
2015-01-22 16:50:48 +02:00
|
|
|
int ff_tls_init(void)
|
2011-02-05 00:25:07 +02:00
|
|
|
{
|
2017-05-23 10:15:28 +02:00
|
|
|
#if CONFIG_TLS_PROTOCOL
|
|
|
|
#if CONFIG_OPENSSL
|
2015-05-28 01:39:55 +02:00
|
|
|
int ret;
|
2015-05-27 12:57:51 +02:00
|
|
|
if ((ret = ff_openssl_init()) < 0)
|
|
|
|
return ret;
|
2011-02-06 01:06:49 +02:00
|
|
|
#endif
|
2017-05-23 10:15:28 +02:00
|
|
|
#if CONFIG_GNUTLS
|
2015-05-27 12:57:51 +02:00
|
|
|
ff_gnutls_init();
|
2011-02-05 00:25:07 +02:00
|
|
|
#endif
|
|
|
|
#endif
|
2015-01-22 16:50:48 +02:00
|
|
|
return 0;
|
2011-02-05 00:25:07 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void ff_tls_deinit(void)
|
|
|
|
{
|
2017-05-23 10:15:28 +02:00
|
|
|
#if CONFIG_TLS_PROTOCOL
|
|
|
|
#if CONFIG_OPENSSL
|
2015-05-27 12:57:51 +02:00
|
|
|
ff_openssl_deinit();
|
2011-02-06 01:06:49 +02:00
|
|
|
#endif
|
2017-05-23 10:15:28 +02:00
|
|
|
#if CONFIG_GNUTLS
|
2015-05-27 12:57:51 +02:00
|
|
|
ff_gnutls_deinit();
|
2011-02-05 00:25:07 +02:00
|
|
|
#endif
|
2017-05-23 10:15:28 +02:00
|
|
|
#endif
|
2011-02-05 00:25:07 +02:00
|
|
|
}
|
2011-02-07 12:59:50 +02:00
|
|
|
|
|
|
|
int ff_network_init(void)
|
|
|
|
{
|
2011-12-21 05:19:10 +03:00
|
|
|
#if HAVE_WINSOCK2_H
|
|
|
|
WSADATA wsaData;
|
|
|
|
|
2011-02-07 12:59:50 +02:00
|
|
|
if (WSAStartup(MAKEWORD(1,1), &wsaData))
|
|
|
|
return 0;
|
|
|
|
#endif
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
int ff_network_wait_fd(int fd, int write)
|
|
|
|
{
|
|
|
|
int ev = write ? POLLOUT : POLLIN;
|
|
|
|
struct pollfd p = { .fd = fd, .events = ev, .revents = 0 };
|
|
|
|
int ret;
|
2016-01-27 05:00:17 +02:00
|
|
|
ret = poll(&p, 1, POLLING_TIME);
|
2011-02-07 12:59:50 +02:00
|
|
|
return ret < 0 ? ff_neterrno() : p.revents & (ev | POLLERR | POLLHUP) ? 0 : AVERROR(EAGAIN);
|
|
|
|
}
|
|
|
|
|
2012-10-09 15:26:02 +03:00
|
|
|
int ff_network_wait_fd_timeout(int fd, int write, int64_t timeout, AVIOInterruptCB *int_cb)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
int64_t wait_start = 0;
|
|
|
|
|
|
|
|
while (1) {
|
2013-07-12 10:35:51 +03:00
|
|
|
if (ff_check_interrupt(int_cb))
|
|
|
|
return AVERROR_EXIT;
|
2012-10-09 15:26:02 +03:00
|
|
|
ret = ff_network_wait_fd(fd, write);
|
|
|
|
if (ret != AVERROR(EAGAIN))
|
|
|
|
return ret;
|
2013-07-06 02:31:37 +03:00
|
|
|
if (timeout > 0) {
|
2012-10-09 15:26:02 +03:00
|
|
|
if (!wait_start)
|
2014-05-07 00:16:51 +03:00
|
|
|
wait_start = av_gettime_relative();
|
|
|
|
else if (av_gettime_relative() - wait_start > timeout)
|
2012-10-09 15:26:02 +03:00
|
|
|
return AVERROR(ETIMEDOUT);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-01-02 18:05:03 +02:00
|
|
|
int ff_network_sleep_interruptible(int64_t timeout, AVIOInterruptCB *int_cb)
|
|
|
|
{
|
|
|
|
int64_t wait_start = av_gettime_relative();
|
|
|
|
|
|
|
|
while (1) {
|
|
|
|
int64_t time_left;
|
|
|
|
|
|
|
|
if (ff_check_interrupt(int_cb))
|
|
|
|
return AVERROR_EXIT;
|
|
|
|
|
|
|
|
time_left = timeout - (av_gettime_relative() - wait_start);
|
|
|
|
if (time_left <= 0)
|
|
|
|
return AVERROR(ETIMEDOUT);
|
|
|
|
|
|
|
|
av_usleep(FFMIN(time_left, POLLING_TIME * 1000));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-02-07 12:59:50 +02:00
|
|
|
void ff_network_close(void)
|
|
|
|
{
|
|
|
|
#if HAVE_WINSOCK2_H
|
|
|
|
WSACleanup();
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
#if HAVE_WINSOCK2_H
|
|
|
|
int ff_neterrno(void)
|
|
|
|
{
|
|
|
|
int err = WSAGetLastError();
|
|
|
|
switch (err) {
|
|
|
|
case WSAEWOULDBLOCK:
|
|
|
|
return AVERROR(EAGAIN);
|
|
|
|
case WSAEINTR:
|
|
|
|
return AVERROR(EINTR);
|
2012-06-23 15:00:17 +03:00
|
|
|
case WSAEPROTONOSUPPORT:
|
|
|
|
return AVERROR(EPROTONOSUPPORT);
|
|
|
|
case WSAETIMEDOUT:
|
|
|
|
return AVERROR(ETIMEDOUT);
|
|
|
|
case WSAECONNREFUSED:
|
|
|
|
return AVERROR(ECONNREFUSED);
|
|
|
|
case WSAEINPROGRESS:
|
|
|
|
return AVERROR(EINPROGRESS);
|
2011-02-07 12:59:50 +02:00
|
|
|
}
|
|
|
|
return -err;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
int ff_is_multicast_address(struct sockaddr *addr)
|
|
|
|
{
|
|
|
|
if (addr->sa_family == AF_INET) {
|
|
|
|
return IN_MULTICAST(ntohl(((struct sockaddr_in *)addr)->sin_addr.s_addr));
|
|
|
|
}
|
|
|
|
#if HAVE_STRUCT_SOCKADDR_IN6
|
|
|
|
if (addr->sa_family == AF_INET6) {
|
|
|
|
return IN6_IS_ADDR_MULTICAST(&((struct sockaddr_in6 *)addr)->sin6_addr);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-05-30 02:08:51 +03:00
|
|
|
|
2013-06-01 20:38:57 +03:00
|
|
|
static int ff_poll_interrupt(struct pollfd *p, nfds_t nfds, int timeout,
|
|
|
|
AVIOInterruptCB *cb)
|
|
|
|
{
|
|
|
|
int runs = timeout / POLLING_TIME;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
do {
|
|
|
|
if (ff_check_interrupt(cb))
|
|
|
|
return AVERROR_EXIT;
|
|
|
|
ret = poll(p, nfds, POLLING_TIME);
|
2018-08-09 20:30:48 +02:00
|
|
|
if (ret != 0) {
|
|
|
|
if (ret < 0)
|
|
|
|
ret = ff_neterrno();
|
|
|
|
if (ret == AVERROR(EINTR))
|
|
|
|
continue;
|
2013-06-01 20:38:57 +03:00
|
|
|
break;
|
2018-08-09 20:30:48 +02:00
|
|
|
}
|
2013-06-05 12:56:56 +03:00
|
|
|
} while (timeout <= 0 || runs-- > 0);
|
2013-06-01 20:38:57 +03:00
|
|
|
|
|
|
|
if (!ret)
|
|
|
|
return AVERROR(ETIMEDOUT);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2022-01-10 18:34:04 +02:00
|
|
|
int ff_socket(int af, int type, int proto, void *logctx)
|
2013-08-08 22:26:42 +03:00
|
|
|
{
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
#ifdef SOCK_CLOEXEC
|
|
|
|
fd = socket(af, type | SOCK_CLOEXEC, proto);
|
|
|
|
if (fd == -1 && errno == EINVAL)
|
|
|
|
#endif
|
|
|
|
{
|
|
|
|
fd = socket(af, type, proto);
|
|
|
|
#if HAVE_FCNTL
|
2013-10-20 17:34:32 +03:00
|
|
|
if (fd != -1) {
|
|
|
|
if (fcntl(fd, F_SETFD, FD_CLOEXEC) == -1)
|
2022-01-10 18:34:04 +02:00
|
|
|
av_log(logctx, AV_LOG_DEBUG, "Failed to set close on exec\n");
|
2013-10-20 17:34:32 +03:00
|
|
|
}
|
2013-08-08 22:26:42 +03:00
|
|
|
#endif
|
|
|
|
}
|
2015-05-30 20:50:43 +02:00
|
|
|
#ifdef SO_NOSIGPIPE
|
2018-08-05 14:19:28 +02:00
|
|
|
if (fd != -1) {
|
|
|
|
if (setsockopt(fd, SOL_SOCKET, SO_NOSIGPIPE, &(int){1}, sizeof(int))) {
|
2022-01-10 18:34:04 +02:00
|
|
|
av_log(logctx, AV_LOG_WARNING, "setsockopt(SO_NOSIGPIPE) failed\n");
|
2018-08-05 14:19:28 +02:00
|
|
|
}
|
|
|
|
}
|
2015-05-30 20:50:43 +02:00
|
|
|
#endif
|
2013-08-08 22:26:42 +03:00
|
|
|
return fd;
|
|
|
|
}
|
|
|
|
|
2015-07-03 02:22:25 +02:00
|
|
|
int ff_listen(int fd, const struct sockaddr *addr,
|
2022-01-10 18:34:04 +02:00
|
|
|
socklen_t addrlen, void *logctx)
|
2013-05-30 02:08:51 +03:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
int reuse = 1;
|
2013-06-03 17:11:29 +03:00
|
|
|
if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &reuse, sizeof(reuse))) {
|
2022-01-10 18:34:04 +02:00
|
|
|
av_log(logctx, AV_LOG_WARNING, "setsockopt(SO_REUSEADDR) failed\n");
|
2013-06-03 17:11:29 +03:00
|
|
|
}
|
2013-05-30 02:08:51 +03:00
|
|
|
ret = bind(fd, addr, addrlen);
|
|
|
|
if (ret)
|
|
|
|
return ff_neterrno();
|
|
|
|
|
|
|
|
ret = listen(fd, 1);
|
|
|
|
if (ret)
|
|
|
|
return ff_neterrno();
|
2015-07-03 02:22:25 +02:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
int ff_accept(int fd, int timeout, URLContext *h)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct pollfd lp = { fd, POLLIN, 0 };
|
2013-05-30 02:08:51 +03:00
|
|
|
|
2013-06-01 20:38:57 +03:00
|
|
|
ret = ff_poll_interrupt(&lp, 1, timeout, &h->interrupt_callback);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
2013-05-30 02:08:51 +03:00
|
|
|
|
|
|
|
ret = accept(fd, NULL, NULL);
|
|
|
|
if (ret < 0)
|
|
|
|
return ff_neterrno();
|
2013-10-21 18:23:35 +03:00
|
|
|
if (ff_socket_nonblock(ret, 1) < 0)
|
2019-09-30 08:41:06 +02:00
|
|
|
av_log(h, AV_LOG_DEBUG, "ff_socket_nonblock failed\n");
|
2013-10-21 18:23:35 +03:00
|
|
|
|
2013-05-30 02:08:51 +03:00
|
|
|
return ret;
|
|
|
|
}
|
2013-05-31 04:05:13 +03:00
|
|
|
|
2015-07-03 02:22:25 +02:00
|
|
|
int ff_listen_bind(int fd, const struct sockaddr *addr,
|
|
|
|
socklen_t addrlen, int timeout, URLContext *h)
|
|
|
|
{
|
|
|
|
int ret;
|
2022-01-10 18:34:04 +02:00
|
|
|
if ((ret = ff_listen(fd, addr, addrlen, h)) < 0)
|
2015-07-03 02:22:25 +02:00
|
|
|
return ret;
|
|
|
|
if ((ret = ff_accept(fd, timeout, h)) < 0)
|
|
|
|
return ret;
|
|
|
|
closesocket(fd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-05-31 04:05:13 +03:00
|
|
|
int ff_listen_connect(int fd, const struct sockaddr *addr,
|
2013-08-05 19:44:20 +03:00
|
|
|
socklen_t addrlen, int timeout, URLContext *h,
|
|
|
|
int will_try_next)
|
2013-05-31 04:05:13 +03:00
|
|
|
{
|
|
|
|
struct pollfd p = {fd, POLLOUT, 0};
|
|
|
|
int ret;
|
|
|
|
socklen_t optlen;
|
|
|
|
|
2013-10-21 18:23:35 +03:00
|
|
|
if (ff_socket_nonblock(fd, 1) < 0)
|
2019-09-30 08:41:06 +02:00
|
|
|
av_log(h, AV_LOG_DEBUG, "ff_socket_nonblock failed\n");
|
2013-05-31 04:05:13 +03:00
|
|
|
|
|
|
|
while ((ret = connect(fd, addr, addrlen))) {
|
|
|
|
ret = ff_neterrno();
|
|
|
|
switch (ret) {
|
|
|
|
case AVERROR(EINTR):
|
|
|
|
if (ff_check_interrupt(&h->interrupt_callback))
|
|
|
|
return AVERROR_EXIT;
|
|
|
|
continue;
|
|
|
|
case AVERROR(EINPROGRESS):
|
|
|
|
case AVERROR(EAGAIN):
|
2013-06-01 20:38:57 +03:00
|
|
|
ret = ff_poll_interrupt(&p, 1, timeout, &h->interrupt_callback);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
2013-05-31 04:05:13 +03:00
|
|
|
optlen = sizeof(ret);
|
|
|
|
if (getsockopt (fd, SOL_SOCKET, SO_ERROR, &ret, &optlen))
|
|
|
|
ret = AVUNERROR(ff_neterrno());
|
|
|
|
if (ret != 0) {
|
|
|
|
char errbuf[100];
|
|
|
|
ret = AVERROR(ret);
|
|
|
|
av_strerror(ret, errbuf, sizeof(errbuf));
|
2013-08-05 19:44:20 +03:00
|
|
|
if (will_try_next)
|
|
|
|
av_log(h, AV_LOG_WARNING,
|
|
|
|
"Connection to %s failed (%s), trying next address\n",
|
|
|
|
h->filename, errbuf);
|
|
|
|
else
|
|
|
|
av_log(h, AV_LOG_ERROR, "Connection to %s failed: %s\n",
|
|
|
|
h->filename, errbuf);
|
2013-05-31 04:05:13 +03:00
|
|
|
}
|
|
|
|
default:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
2013-06-15 12:41:36 +03:00
|
|
|
|
network: Add RFC 8305 style "Happy Eyeballs"/"Fast Fallback" helper function
For cases with dual stack (IPv4 + IPv6) connectivity, but where one
stack potentially is less reliable, strive to trying to connect over
both protocols in parallel, using whichever address connected first.
In cases with a hostname resolving to multiple IPv4 and IPv6
addresses, the current connection mechanism would try all addresses
in the order returned by getaddrinfo (with all IPv6 addresses ordered
before the IPv4 addresses normally). If connection attempts to the
IPv6 addresses return quickly with an error, this was no problem, but
if they were unsuccessful leading up to timeouts, the connection process
would have to wait for timeouts on all IPv6 target addresses before
attempting any IPv4 address.
Similar to what RFC 8305 suggests, reorder the list of addresses to
try connecting to, interleaving address families. After starting one
connection attempt, start another one in parallel after a small delay
(200 ms as suggested by the RFC).
For cases with unreliable IPv6 but reliable IPv4, this should make
connection attempts work as reliably as with plain IPv4, with only an
extra 200 ms of connection delay.
Signed-off-by: Martin Storsjö <martin@martin.st>
2018-08-10 09:25:59 +02:00
|
|
|
static void interleave_addrinfo(struct addrinfo *base)
|
|
|
|
{
|
|
|
|
struct addrinfo **next = &base->ai_next;
|
|
|
|
while (*next) {
|
|
|
|
struct addrinfo *cur = *next;
|
|
|
|
// Iterate forward until we find an entry of a different family.
|
|
|
|
if (cur->ai_family == base->ai_family) {
|
|
|
|
next = &cur->ai_next;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (cur == base->ai_next) {
|
|
|
|
// If the first one following base is of a different family, just
|
|
|
|
// move base forward one step and continue.
|
|
|
|
base = cur;
|
|
|
|
next = &base->ai_next;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
// Unchain cur from the rest of the list from its current spot.
|
|
|
|
*next = cur->ai_next;
|
|
|
|
// Hook in cur directly after base.
|
|
|
|
cur->ai_next = base->ai_next;
|
|
|
|
base->ai_next = cur;
|
|
|
|
// Restart with a new base. We know that before moving the cur element,
|
|
|
|
// everything between the previous base and cur had the same family,
|
|
|
|
// different from cur->ai_family. Therefore, we can keep next pointing
|
|
|
|
// where it was, and continue from there with base at the one after
|
|
|
|
// cur.
|
|
|
|
base = cur->ai_next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_address_list(void *ctx, const struct addrinfo *addr,
|
|
|
|
const char *title)
|
|
|
|
{
|
|
|
|
char hostbuf[100], portbuf[20];
|
|
|
|
av_log(ctx, AV_LOG_DEBUG, "%s:\n", title);
|
|
|
|
while (addr) {
|
|
|
|
getnameinfo(addr->ai_addr, addr->ai_addrlen,
|
|
|
|
hostbuf, sizeof(hostbuf), portbuf, sizeof(portbuf),
|
|
|
|
NI_NUMERICHOST | NI_NUMERICSERV);
|
|
|
|
av_log(ctx, AV_LOG_DEBUG, "Address %s port %s\n", hostbuf, portbuf);
|
|
|
|
addr = addr->ai_next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
struct ConnectionAttempt {
|
|
|
|
int fd;
|
|
|
|
int64_t deadline_us;
|
|
|
|
struct addrinfo *addr;
|
|
|
|
};
|
|
|
|
|
|
|
|
// Returns < 0 on error, 0 on successfully started connection attempt,
|
|
|
|
// > 0 for a connection that succeeded already.
|
|
|
|
static int start_connect_attempt(struct ConnectionAttempt *attempt,
|
|
|
|
struct addrinfo **ptr, int timeout_ms,
|
|
|
|
URLContext *h,
|
|
|
|
void (*customize_fd)(void *, int), void *customize_ctx)
|
|
|
|
{
|
|
|
|
struct addrinfo *ai = *ptr;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
*ptr = ai->ai_next;
|
|
|
|
|
2022-01-10 18:34:04 +02:00
|
|
|
attempt->fd = ff_socket(ai->ai_family, ai->ai_socktype, ai->ai_protocol, h);
|
network: Add RFC 8305 style "Happy Eyeballs"/"Fast Fallback" helper function
For cases with dual stack (IPv4 + IPv6) connectivity, but where one
stack potentially is less reliable, strive to trying to connect over
both protocols in parallel, using whichever address connected first.
In cases with a hostname resolving to multiple IPv4 and IPv6
addresses, the current connection mechanism would try all addresses
in the order returned by getaddrinfo (with all IPv6 addresses ordered
before the IPv4 addresses normally). If connection attempts to the
IPv6 addresses return quickly with an error, this was no problem, but
if they were unsuccessful leading up to timeouts, the connection process
would have to wait for timeouts on all IPv6 target addresses before
attempting any IPv4 address.
Similar to what RFC 8305 suggests, reorder the list of addresses to
try connecting to, interleaving address families. After starting one
connection attempt, start another one in parallel after a small delay
(200 ms as suggested by the RFC).
For cases with unreliable IPv6 but reliable IPv4, this should make
connection attempts work as reliably as with plain IPv4, with only an
extra 200 ms of connection delay.
Signed-off-by: Martin Storsjö <martin@martin.st>
2018-08-10 09:25:59 +02:00
|
|
|
if (attempt->fd < 0)
|
|
|
|
return ff_neterrno();
|
|
|
|
attempt->deadline_us = av_gettime_relative() + timeout_ms * 1000;
|
|
|
|
attempt->addr = ai;
|
|
|
|
|
|
|
|
ff_socket_nonblock(attempt->fd, 1);
|
|
|
|
|
|
|
|
if (customize_fd)
|
|
|
|
customize_fd(customize_ctx, attempt->fd);
|
|
|
|
|
|
|
|
while ((ret = connect(attempt->fd, ai->ai_addr, ai->ai_addrlen))) {
|
|
|
|
ret = ff_neterrno();
|
|
|
|
switch (ret) {
|
|
|
|
case AVERROR(EINTR):
|
|
|
|
if (ff_check_interrupt(&h->interrupt_callback)) {
|
|
|
|
closesocket(attempt->fd);
|
|
|
|
attempt->fd = -1;
|
|
|
|
return AVERROR_EXIT;
|
|
|
|
}
|
|
|
|
continue;
|
|
|
|
case AVERROR(EINPROGRESS):
|
|
|
|
case AVERROR(EAGAIN):
|
|
|
|
return 0;
|
|
|
|
default:
|
|
|
|
closesocket(attempt->fd);
|
|
|
|
attempt->fd = -1;
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Try a new connection to another address after 200 ms, as suggested in
|
|
|
|
// RFC 8305 (or sooner if an earlier attempt fails).
|
|
|
|
#define NEXT_ATTEMPT_DELAY_MS 200
|
|
|
|
|
|
|
|
int ff_connect_parallel(struct addrinfo *addrs, int timeout_ms_per_address,
|
|
|
|
int parallel, URLContext *h, int *fd,
|
|
|
|
void (*customize_fd)(void *, int), void *customize_ctx)
|
|
|
|
{
|
|
|
|
struct ConnectionAttempt attempts[3];
|
|
|
|
struct pollfd pfd[3];
|
|
|
|
int nb_attempts = 0, i, j;
|
|
|
|
int64_t next_attempt_us = av_gettime_relative(), next_deadline_us;
|
|
|
|
int last_err = AVERROR(EIO);
|
|
|
|
socklen_t optlen;
|
|
|
|
char errbuf[100], hostbuf[100], portbuf[20];
|
|
|
|
|
|
|
|
if (parallel > FF_ARRAY_ELEMS(attempts))
|
|
|
|
parallel = FF_ARRAY_ELEMS(attempts);
|
|
|
|
|
|
|
|
print_address_list(h, addrs, "Original list of addresses");
|
|
|
|
// This mutates the list, but the head of the list is still the same
|
|
|
|
// element, so the caller, who owns the list, doesn't need to get
|
|
|
|
// an updated pointer.
|
|
|
|
interleave_addrinfo(addrs);
|
|
|
|
print_address_list(h, addrs, "Interleaved list of addresses");
|
|
|
|
|
|
|
|
while (nb_attempts > 0 || addrs) {
|
|
|
|
// Start a new connection attempt, if possible.
|
|
|
|
if (nb_attempts < parallel && addrs) {
|
|
|
|
getnameinfo(addrs->ai_addr, addrs->ai_addrlen,
|
|
|
|
hostbuf, sizeof(hostbuf), portbuf, sizeof(portbuf),
|
|
|
|
NI_NUMERICHOST | NI_NUMERICSERV);
|
|
|
|
av_log(h, AV_LOG_VERBOSE, "Starting connection attempt to %s port %s\n",
|
|
|
|
hostbuf, portbuf);
|
|
|
|
last_err = start_connect_attempt(&attempts[nb_attempts], &addrs,
|
|
|
|
timeout_ms_per_address, h,
|
|
|
|
customize_fd, customize_ctx);
|
|
|
|
if (last_err < 0) {
|
|
|
|
av_strerror(last_err, errbuf, sizeof(errbuf));
|
|
|
|
av_log(h, AV_LOG_VERBOSE, "Connected attempt failed: %s\n",
|
|
|
|
errbuf);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (last_err > 0) {
|
|
|
|
for (i = 0; i < nb_attempts; i++)
|
|
|
|
closesocket(attempts[i].fd);
|
|
|
|
*fd = attempts[nb_attempts].fd;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
pfd[nb_attempts].fd = attempts[nb_attempts].fd;
|
|
|
|
pfd[nb_attempts].events = POLLOUT;
|
|
|
|
next_attempt_us = av_gettime_relative() + NEXT_ATTEMPT_DELAY_MS * 1000;
|
|
|
|
nb_attempts++;
|
|
|
|
}
|
|
|
|
|
|
|
|
av_assert0(nb_attempts > 0);
|
|
|
|
// The connection attempts are sorted from oldest to newest, so the
|
|
|
|
// first one will have the earliest deadline.
|
|
|
|
next_deadline_us = attempts[0].deadline_us;
|
|
|
|
// If we can start another attempt in parallel, wait until that time.
|
|
|
|
if (nb_attempts < parallel && addrs)
|
|
|
|
next_deadline_us = FFMIN(next_deadline_us, next_attempt_us);
|
|
|
|
last_err = ff_poll_interrupt(pfd, nb_attempts,
|
|
|
|
(next_deadline_us - av_gettime_relative())/1000,
|
|
|
|
&h->interrupt_callback);
|
|
|
|
if (last_err < 0 && last_err != AVERROR(ETIMEDOUT))
|
|
|
|
break;
|
|
|
|
|
|
|
|
// Check the status from the poll output.
|
|
|
|
for (i = 0; i < nb_attempts; i++) {
|
|
|
|
last_err = 0;
|
|
|
|
if (pfd[i].revents) {
|
|
|
|
// Some sort of action for this socket, check its status (either
|
|
|
|
// a successful connection or an error).
|
|
|
|
optlen = sizeof(last_err);
|
|
|
|
if (getsockopt(attempts[i].fd, SOL_SOCKET, SO_ERROR, &last_err, &optlen))
|
|
|
|
last_err = ff_neterrno();
|
|
|
|
else if (last_err != 0)
|
|
|
|
last_err = AVERROR(last_err);
|
|
|
|
if (last_err == 0) {
|
|
|
|
// Everything is ok, we seem to have a successful
|
|
|
|
// connection. Close other sockets and return this one.
|
|
|
|
for (j = 0; j < nb_attempts; j++)
|
|
|
|
if (j != i)
|
|
|
|
closesocket(attempts[j].fd);
|
|
|
|
*fd = attempts[i].fd;
|
|
|
|
getnameinfo(attempts[i].addr->ai_addr, attempts[i].addr->ai_addrlen,
|
|
|
|
hostbuf, sizeof(hostbuf), portbuf, sizeof(portbuf),
|
|
|
|
NI_NUMERICHOST | NI_NUMERICSERV);
|
|
|
|
av_log(h, AV_LOG_VERBOSE, "Successfully connected to %s port %s\n",
|
|
|
|
hostbuf, portbuf);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (attempts[i].deadline_us < av_gettime_relative() && !last_err)
|
|
|
|
last_err = AVERROR(ETIMEDOUT);
|
|
|
|
if (!last_err)
|
|
|
|
continue;
|
|
|
|
// Error (or timeout) for this socket; close the socket and remove
|
|
|
|
// it from the attempts/pfd arrays, to let a new attempt start
|
|
|
|
// directly.
|
|
|
|
getnameinfo(attempts[i].addr->ai_addr, attempts[i].addr->ai_addrlen,
|
|
|
|
hostbuf, sizeof(hostbuf), portbuf, sizeof(portbuf),
|
|
|
|
NI_NUMERICHOST | NI_NUMERICSERV);
|
|
|
|
av_strerror(last_err, errbuf, sizeof(errbuf));
|
|
|
|
av_log(h, AV_LOG_VERBOSE, "Connection attempt to %s port %s "
|
|
|
|
"failed: %s\n", hostbuf, portbuf, errbuf);
|
|
|
|
closesocket(attempts[i].fd);
|
|
|
|
memmove(&attempts[i], &attempts[i + 1],
|
|
|
|
(nb_attempts - i - 1) * sizeof(*attempts));
|
|
|
|
memmove(&pfd[i], &pfd[i + 1],
|
|
|
|
(nb_attempts - i - 1) * sizeof(*pfd));
|
|
|
|
i--;
|
|
|
|
nb_attempts--;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (i = 0; i < nb_attempts; i++)
|
|
|
|
closesocket(attempts[i].fd);
|
|
|
|
if (last_err >= 0)
|
|
|
|
last_err = AVERROR(ECONNREFUSED);
|
|
|
|
if (last_err != AVERROR_EXIT) {
|
|
|
|
av_strerror(last_err, errbuf, sizeof(errbuf));
|
|
|
|
av_log(h, AV_LOG_ERROR, "Connection to %s failed: %s\n",
|
|
|
|
h->filename, errbuf);
|
|
|
|
}
|
|
|
|
return last_err;
|
|
|
|
}
|
|
|
|
|
2013-06-15 12:41:36 +03:00
|
|
|
static int match_host_pattern(const char *pattern, const char *hostname)
|
|
|
|
{
|
|
|
|
int len_p, len_h;
|
|
|
|
if (!strcmp(pattern, "*"))
|
|
|
|
return 1;
|
|
|
|
// Skip a possible *. at the start of the pattern
|
|
|
|
if (pattern[0] == '*')
|
|
|
|
pattern++;
|
|
|
|
if (pattern[0] == '.')
|
|
|
|
pattern++;
|
|
|
|
len_p = strlen(pattern);
|
|
|
|
len_h = strlen(hostname);
|
|
|
|
if (len_p > len_h)
|
|
|
|
return 0;
|
|
|
|
// Simply check if the end of hostname is equal to 'pattern'
|
|
|
|
if (!strcmp(pattern, &hostname[len_h - len_p])) {
|
|
|
|
if (len_h == len_p)
|
|
|
|
return 1; // Exact match
|
|
|
|
if (hostname[len_h - len_p - 1] == '.')
|
|
|
|
return 1; // The matched substring is a domain and not just a substring of a domain
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int ff_http_match_no_proxy(const char *no_proxy, const char *hostname)
|
|
|
|
{
|
|
|
|
char *buf, *start;
|
|
|
|
int ret = 0;
|
|
|
|
if (!no_proxy)
|
|
|
|
return 0;
|
|
|
|
if (!hostname)
|
|
|
|
return 0;
|
|
|
|
buf = av_strdup(no_proxy);
|
|
|
|
if (!buf)
|
|
|
|
return 0;
|
|
|
|
start = buf;
|
|
|
|
while (start) {
|
|
|
|
char *sep, *next = NULL;
|
|
|
|
start += strspn(start, " ,");
|
|
|
|
sep = start + strcspn(start, " ,");
|
|
|
|
if (*sep) {
|
|
|
|
next = sep + 1;
|
|
|
|
*sep = '\0';
|
|
|
|
}
|
|
|
|
if (match_host_pattern(start, hostname)) {
|
|
|
|
ret = 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
start = next;
|
|
|
|
}
|
|
|
|
av_free(buf);
|
|
|
|
return ret;
|
|
|
|
}
|
2018-08-04 11:48:15 +02:00
|
|
|
|
|
|
|
void ff_log_net_error(void *ctx, int level, const char* prefix)
|
|
|
|
{
|
|
|
|
char errbuf[100];
|
|
|
|
av_strerror(ff_neterrno(), errbuf, sizeof(errbuf));
|
|
|
|
av_log(ctx, level, "%s: %s\n", prefix, errbuf);
|
|
|
|
}
|