ffmpeg/libavformat/tcp.c

379 lines
12 KiB
C
Raw Normal View History

/*
* TCP protocol
* Copyright (c) 2002 Fabrice Bellard
*
* This file is part of FFmpeg.
*
* FFmpeg 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 2.1 of the License, or (at your option) any later version.
*
* FFmpeg 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 FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "avformat.h"
#include "libavutil/avassert.h"
#include "libavutil/mem.h"
#include "libavutil/parseutils.h"
#include "libavutil/opt.h"
#include "libavutil/time.h"
2014-10-09 22:05:27 +02:00
#include "network.h"
#include "os_support.h"
#include "url.h"
#if HAVE_POLL_H
#include <poll.h>
#endif
typedef struct TCPContext {
const AVClass *class;
int fd;
int listen;
char *local_port;
char *local_addr;
int open_timeout;
int rw_timeout;
int listen_timeout;
int recv_buffer_size;
int send_buffer_size;
int tcp_nodelay;
#if !HAVE_WINSOCK2_H
int tcp_mss;
#endif /* !HAVE_WINSOCK2_H */
} TCPContext;
#define OFFSET(x) offsetof(TCPContext, x)
#define D AV_OPT_FLAG_DECODING_PARAM
#define E AV_OPT_FLAG_ENCODING_PARAM
static const AVOption options[] = {
{ "listen", "Listen for incoming connections", OFFSET(listen), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 2, .flags = D|E },
{ "local_port", "Local port", OFFSET(local_port), AV_OPT_TYPE_STRING, { .str = NULL }, 0, 0, .flags = D|E },
{ "local_addr", "Local address", OFFSET(local_addr), AV_OPT_TYPE_STRING, { .str = NULL }, 0, 0, .flags = D|E },
{ "timeout", "set timeout (in microseconds) of socket I/O operations", OFFSET(rw_timeout), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, .flags = D|E },
{ "listen_timeout", "Connection awaiting timeout (in milliseconds)", OFFSET(listen_timeout), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, .flags = D|E },
{ "send_buffer_size", "Socket send buffer size (in bytes)", OFFSET(send_buffer_size), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, .flags = D|E },
{ "recv_buffer_size", "Socket receive buffer size (in bytes)", OFFSET(recv_buffer_size), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, .flags = D|E },
{ "tcp_nodelay", "Use TCP_NODELAY to disable nagle's algorithm", OFFSET(tcp_nodelay), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, .flags = D|E },
#if !HAVE_WINSOCK2_H
{ "tcp_mss", "Maximum segment size for outgoing TCP packets", OFFSET(tcp_mss), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, .flags = D|E },
#endif /* !HAVE_WINSOCK2_H */
2014-10-09 22:05:27 +02:00
{ NULL }
};
2014-10-09 22:05:27 +02:00
static const AVClass tcp_class = {
.class_name = "tcp",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
static int customize_fd(void *ctx, int fd, int family)
{
TCPContext *s = ctx;
if (s->local_addr || s->local_port) {
struct addrinfo hints = { 0 }, *ai, *cur_ai;
int ret;
hints.ai_family = family;
hints.ai_socktype = SOCK_STREAM;
ret = getaddrinfo(s->local_addr, s->local_port, &hints, &ai);
if (ret) {
av_log(ctx, AV_LOG_ERROR,
"Failed to getaddrinfo local addr: %s port: %s err: %s\n",
s->local_addr, s->local_port, gai_strerror(ret));
return ret;
}
cur_ai = ai;
while (cur_ai) {
ret = bind(fd, (struct sockaddr *)cur_ai->ai_addr, (int)cur_ai->ai_addrlen);
if (ret)
cur_ai = cur_ai->ai_next;
else
break;
}
freeaddrinfo(ai);
if (ret) {
ff_log_net_error(ctx, AV_LOG_ERROR, "bind local failed");
return ret;
}
}
/* Set the socket's send or receive buffer sizes, if specified.
If unspecified or setting fails, system default is used. */
if (s->recv_buffer_size > 0) {
if (setsockopt (fd, SOL_SOCKET, SO_RCVBUF, &s->recv_buffer_size, sizeof (s->recv_buffer_size))) {
ff_log_net_error(ctx, AV_LOG_WARNING, "setsockopt(SO_RCVBUF)");
}
}
if (s->send_buffer_size > 0) {
if (setsockopt (fd, SOL_SOCKET, SO_SNDBUF, &s->send_buffer_size, sizeof (s->send_buffer_size))) {
ff_log_net_error(ctx, AV_LOG_WARNING, "setsockopt(SO_SNDBUF)");
}
}
if (s->tcp_nodelay > 0) {
if (setsockopt (fd, IPPROTO_TCP, TCP_NODELAY, &s->tcp_nodelay, sizeof (s->tcp_nodelay))) {
ff_log_net_error(ctx, AV_LOG_WARNING, "setsockopt(TCP_NODELAY)");
}
}
#if !HAVE_WINSOCK2_H
if (s->tcp_mss > 0) {
if (setsockopt (fd, IPPROTO_TCP, TCP_MAXSEG, &s->tcp_mss, sizeof (s->tcp_mss))) {
ff_log_net_error(ctx, AV_LOG_WARNING, "setsockopt(TCP_MAXSEG)");
}
}
#endif /* !HAVE_WINSOCK2_H */
return 0;
}
/* return non zero if error */
static int tcp_open(URLContext *h, const char *uri, int flags)
{
struct addrinfo hints = { 0 }, *ai, *cur_ai;
int port, fd = -1;
TCPContext *s = h->priv_data;
const char *p;
char buf[256];
int ret;
char hostname[1024],proto[1024],path[1024];
char portstr[10];
s->open_timeout = 5000000;
av_url_split(proto, sizeof(proto), NULL, 0, hostname, sizeof(hostname),
&port, path, sizeof(path), uri);
if (strcmp(proto, "tcp"))
return AVERROR(EINVAL);
if (port <= 0 || port >= 65536) {
av_log(h, AV_LOG_ERROR, "Port missing in uri\n");
return AVERROR(EINVAL);
}
p = strchr(uri, '?');
if (p) {
if (av_find_info_tag(buf, sizeof(buf), "listen", p)) {
char *endptr = NULL;
s->listen = strtol(buf, &endptr, 10);
/* assume if no digits were found it is a request to enable it */
if (buf == endptr)
s->listen = 1;
}
if (av_find_info_tag(buf, sizeof(buf), "local_port", p)) {
av_freep(&s->local_port);
s->local_port = av_strdup(buf);
if (!s->local_port)
return AVERROR(ENOMEM);
}
if (av_find_info_tag(buf, sizeof(buf), "local_addr", p)) {
av_freep(&s->local_addr);
s->local_addr = av_strdup(buf);
if (!s->local_addr)
return AVERROR(ENOMEM);
}
if (av_find_info_tag(buf, sizeof(buf), "timeout", p)) {
s->rw_timeout = strtol(buf, NULL, 10);
}
if (av_find_info_tag(buf, sizeof(buf), "listen_timeout", p)) {
s->listen_timeout = strtol(buf, NULL, 10);
}
if (av_find_info_tag(buf, sizeof(buf), "tcp_nodelay", p)) {
s->tcp_nodelay = strtol(buf, NULL, 10);
}
}
if (s->rw_timeout >= 0) {
s->open_timeout =
h->rw_timeout = s->rw_timeout;
}
hints.ai_family = AF_UNSPEC;
hints.ai_socktype = SOCK_STREAM;
snprintf(portstr, sizeof(portstr), "%d", port);
if (s->listen)
hints.ai_flags |= AI_PASSIVE;
if (!hostname[0])
ret = getaddrinfo(NULL, portstr, &hints, &ai);
else
ret = getaddrinfo(hostname, portstr, &hints, &ai);
if (ret) {
av_log(h, AV_LOG_ERROR,
"Failed to resolve hostname %s: %s\n",
hostname, gai_strerror(ret));
return AVERROR(EIO);
}
cur_ai = ai;
#if HAVE_STRUCT_SOCKADDR_IN6
// workaround for IOS9 getaddrinfo in IPv6 only network use hardcode IPv4 address can not resolve port number.
if (cur_ai->ai_family == AF_INET6){
struct sockaddr_in6 * sockaddr_v6 = (struct sockaddr_in6 *)cur_ai->ai_addr;
if (!sockaddr_v6->sin6_port){
sockaddr_v6->sin6_port = htons(port);
}
}
#endif
if (s->listen > 0) {
while (cur_ai && fd < 0) {
fd = ff_socket(cur_ai->ai_family,
cur_ai->ai_socktype,
cur_ai->ai_protocol, h);
if (fd < 0) {
ret = ff_neterrno();
cur_ai = cur_ai->ai_next;
}
}
if (fd < 0)
goto fail1;
customize_fd(s, fd, cur_ai->ai_family);
}
tcp: set socket buffer sizes before listen/connect/accept From e24d95c0e06a878d401ee34fd6742fcaddeeb95f Mon Sep 17 00:00:00 2001 From: Joel Cunningham <joel.cunningham@me.com> Date: Mon, 9 Jan 2017 13:37:51 -0600 Subject: [PATCH] tcp: set socket buffer sizes before listen/connect/accept Attempting to set SO_RCVBUF and SO_SNDBUF on TCP sockets after connection establishment is incorrect and some stacks ignore the set call on the socket at this point. This has been observed on MacOS/iOS. Windows 7 has some peculiar behavior where setting SO_RCVBUF after applies only if the buffer is increasing from the default while decreases are ignored. This is possibly how the incorrect usage has gone unnoticed Unix Network Programming Vol. 1: The Sockets Networking API (3rd edition, seciton 7.5): "When setting the size of the TCP socket receive buffer, the ordering of the function calls is important. This is because of TCP's window scale option, which is exchanged with the peer on SYN segments when the connection is established. For a client, this means the SO_RCVBUF socket option must be set before calling connect. For a server, this means the socket option must be set for the listening socket before calling listen. Setting this option for the connected socket will have no effect whatsoever on the possible window scale option because accept does not return with the connected socket until TCP's three-way handshake is complete. This is why the option must be set on the listening socket. (The sizes of the socket buffers are always inherited from the listening socket by the newly created connected socket)" Signed-off-by: Joel Cunningham <joel.cunningham@me.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-01-09 21:54:47 +01:00
if (s->listen == 2) {
// multi-client
if ((ret = ff_listen(fd, cur_ai->ai_addr, cur_ai->ai_addrlen, h)) < 0)
goto fail1;
} else if (s->listen == 1) {
// single client
if ((ret = ff_listen_bind(fd, cur_ai->ai_addr, cur_ai->ai_addrlen,
s->listen_timeout, h)) < 0)
goto fail1;
// Socket descriptor already closed here. Safe to overwrite to client one.
fd = ret;
} else {
ret = ff_connect_parallel(ai, s->open_timeout / 1000, 3, h, &fd, customize_fd, s);
if (ret < 0)
goto fail1;
}
h->is_streamed = 1;
s->fd = fd;
freeaddrinfo(ai);
return 0;
fail1:
if (fd >= 0)
closesocket(fd);
freeaddrinfo(ai);
return ret;
}
static int tcp_accept(URLContext *s, URLContext **c)
{
TCPContext *sc = s->priv_data;
TCPContext *cc;
int ret;
av_assert0(sc->listen);
if ((ret = ffurl_alloc(c, s->filename, s->flags, &s->interrupt_callback)) < 0)
return ret;
cc = (*c)->priv_data;
ret = ff_accept(sc->fd, sc->listen_timeout, s);
if (ret < 0) {
ffurl_closep(c);
return ret;
}
cc->fd = ret;
return 0;
}
static int tcp_read(URLContext *h, uint8_t *buf, int size)
{
TCPContext *s = h->priv_data;
int ret;
if (!(h->flags & AVIO_FLAG_NONBLOCK)) {
ret = ff_network_wait_fd_timeout(s->fd, 0, h->rw_timeout, &h->interrupt_callback);
if (ret)
return ret;
}
ret = recv(s->fd, buf, size, 0);
if (ret == 0)
return AVERROR_EOF;
return ret < 0 ? ff_neterrno() : ret;
}
static int tcp_write(URLContext *h, const uint8_t *buf, int size)
{
TCPContext *s = h->priv_data;
int ret;
if (!(h->flags & AVIO_FLAG_NONBLOCK)) {
ret = ff_network_wait_fd_timeout(s->fd, 1, h->rw_timeout, &h->interrupt_callback);
if (ret)
return ret;
}
ret = send(s->fd, buf, size, MSG_NOSIGNAL);
return ret < 0 ? ff_neterrno() : ret;
}
static int tcp_shutdown(URLContext *h, int flags)
{
TCPContext *s = h->priv_data;
int how;
if (flags & AVIO_FLAG_WRITE && flags & AVIO_FLAG_READ) {
how = SHUT_RDWR;
} else if (flags & AVIO_FLAG_WRITE) {
how = SHUT_WR;
} else {
how = SHUT_RD;
}
return shutdown(s->fd, how);
}
static int tcp_close(URLContext *h)
{
TCPContext *s = h->priv_data;
closesocket(s->fd);
return 0;
}
static int tcp_get_file_handle(URLContext *h)
{
TCPContext *s = h->priv_data;
return s->fd;
}
HTTP: improve performance by reducing forward seeks This commit optimizes HTTP performance by reducing forward seeks, instead favoring a read-ahead and discard on the current connection (referred to as a short seek) for seeks that are within a TCP window's worth of data. This improves performance because with TCP flow control, a window's worth of data will be in the local socket buffer already or in-flight from the sender once congestion control on the sender is fully utilizing the window. Note: this approach doesn't attempt to differentiate from a newly opened connection which may not be fully utilizing the window due to congestion control vs one that is. The receiver can't get at this information, so we assume worst case; that full window is in use (we did advertise it after all) and that data could be in-flight The previous behavior of closing the connection, then opening a new with a new HTTP range value results in a massive amounts of discarded and re-sent data when large TCP windows are used. This has been observed on MacOS/iOS which starts with an initial window of 256KB and grows up to 1MB depending on the bandwidth-product delay. When seeking within a window's worth of data and we close the connection, then open a new one within the same window's worth of data, we discard from the current offset till the end of the window. Then on the new connection the server ends up re-sending the previous data from new offset till the end of old window. Example (assumes full window utilization): TCP window size: 64KB Position: 32KB Forward seek position: 40KB * (Next window) 32KB |--------------| 96KB |---------------| 160KB * 40KB |---------------| 104KB Re-sent amount: 96KB - 40KB = 56KB For a real world test example, I have MP4 file of ~25MB, which ffplay only reads ~16MB and performs 177 seeks. With current ffmpeg, this results in 177 HTTP GETs and ~73MB worth of TCP data communication. With this patch, ffmpeg issues 4 HTTP GETs and 3 seeks for a total of ~22MB of TCP data communication. To support this feature, the short seek logic in avio_seek() has been extended to call a function to get the short seek threshold value. This callback has been plumbed to the URLProtocol structure, which now has infrastructure in HTTP and TCP to get the underlying receiver window size via SO_RCVBUF. If the underlying URL and protocol don't support returning a short seek threshold, the default s->short_seek_threshold is used This feature has been tested on Windows 7 and MacOS/iOS. Windows support is slightly complicated by the fact that when TCP window auto-tuning is enabled, SO_RCVBUF doesn't report the real window size, but it does if SO_RCVBUF was manually set (disabling auto-tuning). So we can only use this optimization on Windows in the later case Signed-off-by: Joel Cunningham <joel.cunningham@me.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-01-30 17:00:44 +01:00
static int tcp_get_window_size(URLContext *h)
{
TCPContext *s = h->priv_data;
int avail;
socklen_t avail_len = sizeof(avail);
HTTP: improve performance by reducing forward seeks This commit optimizes HTTP performance by reducing forward seeks, instead favoring a read-ahead and discard on the current connection (referred to as a short seek) for seeks that are within a TCP window's worth of data. This improves performance because with TCP flow control, a window's worth of data will be in the local socket buffer already or in-flight from the sender once congestion control on the sender is fully utilizing the window. Note: this approach doesn't attempt to differentiate from a newly opened connection which may not be fully utilizing the window due to congestion control vs one that is. The receiver can't get at this information, so we assume worst case; that full window is in use (we did advertise it after all) and that data could be in-flight The previous behavior of closing the connection, then opening a new with a new HTTP range value results in a massive amounts of discarded and re-sent data when large TCP windows are used. This has been observed on MacOS/iOS which starts with an initial window of 256KB and grows up to 1MB depending on the bandwidth-product delay. When seeking within a window's worth of data and we close the connection, then open a new one within the same window's worth of data, we discard from the current offset till the end of the window. Then on the new connection the server ends up re-sending the previous data from new offset till the end of old window. Example (assumes full window utilization): TCP window size: 64KB Position: 32KB Forward seek position: 40KB * (Next window) 32KB |--------------| 96KB |---------------| 160KB * 40KB |---------------| 104KB Re-sent amount: 96KB - 40KB = 56KB For a real world test example, I have MP4 file of ~25MB, which ffplay only reads ~16MB and performs 177 seeks. With current ffmpeg, this results in 177 HTTP GETs and ~73MB worth of TCP data communication. With this patch, ffmpeg issues 4 HTTP GETs and 3 seeks for a total of ~22MB of TCP data communication. To support this feature, the short seek logic in avio_seek() has been extended to call a function to get the short seek threshold value. This callback has been plumbed to the URLProtocol structure, which now has infrastructure in HTTP and TCP to get the underlying receiver window size via SO_RCVBUF. If the underlying URL and protocol don't support returning a short seek threshold, the default s->short_seek_threshold is used This feature has been tested on Windows 7 and MacOS/iOS. Windows support is slightly complicated by the fact that when TCP window auto-tuning is enabled, SO_RCVBUF doesn't report the real window size, but it does if SO_RCVBUF was manually set (disabling auto-tuning). So we can only use this optimization on Windows in the later case Signed-off-by: Joel Cunningham <joel.cunningham@me.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-01-30 17:00:44 +01:00
#if HAVE_WINSOCK2_H
/* SO_RCVBUF with winsock only reports the actual TCP window size when
auto-tuning has been disabled via setting SO_RCVBUF */
if (s->recv_buffer_size < 0) {
return AVERROR(ENOSYS);
}
#endif
if (getsockopt(s->fd, SOL_SOCKET, SO_RCVBUF, &avail, &avail_len)) {
return ff_neterrno();
}
return avail;
}
const URLProtocol ff_tcp_protocol = {
.name = "tcp",
.url_open = tcp_open,
.url_accept = tcp_accept,
.url_read = tcp_read,
.url_write = tcp_write,
.url_close = tcp_close,
.url_get_file_handle = tcp_get_file_handle,
HTTP: improve performance by reducing forward seeks This commit optimizes HTTP performance by reducing forward seeks, instead favoring a read-ahead and discard on the current connection (referred to as a short seek) for seeks that are within a TCP window's worth of data. This improves performance because with TCP flow control, a window's worth of data will be in the local socket buffer already or in-flight from the sender once congestion control on the sender is fully utilizing the window. Note: this approach doesn't attempt to differentiate from a newly opened connection which may not be fully utilizing the window due to congestion control vs one that is. The receiver can't get at this information, so we assume worst case; that full window is in use (we did advertise it after all) and that data could be in-flight The previous behavior of closing the connection, then opening a new with a new HTTP range value results in a massive amounts of discarded and re-sent data when large TCP windows are used. This has been observed on MacOS/iOS which starts with an initial window of 256KB and grows up to 1MB depending on the bandwidth-product delay. When seeking within a window's worth of data and we close the connection, then open a new one within the same window's worth of data, we discard from the current offset till the end of the window. Then on the new connection the server ends up re-sending the previous data from new offset till the end of old window. Example (assumes full window utilization): TCP window size: 64KB Position: 32KB Forward seek position: 40KB * (Next window) 32KB |--------------| 96KB |---------------| 160KB * 40KB |---------------| 104KB Re-sent amount: 96KB - 40KB = 56KB For a real world test example, I have MP4 file of ~25MB, which ffplay only reads ~16MB and performs 177 seeks. With current ffmpeg, this results in 177 HTTP GETs and ~73MB worth of TCP data communication. With this patch, ffmpeg issues 4 HTTP GETs and 3 seeks for a total of ~22MB of TCP data communication. To support this feature, the short seek logic in avio_seek() has been extended to call a function to get the short seek threshold value. This callback has been plumbed to the URLProtocol structure, which now has infrastructure in HTTP and TCP to get the underlying receiver window size via SO_RCVBUF. If the underlying URL and protocol don't support returning a short seek threshold, the default s->short_seek_threshold is used This feature has been tested on Windows 7 and MacOS/iOS. Windows support is slightly complicated by the fact that when TCP window auto-tuning is enabled, SO_RCVBUF doesn't report the real window size, but it does if SO_RCVBUF was manually set (disabling auto-tuning). So we can only use this optimization on Windows in the later case Signed-off-by: Joel Cunningham <joel.cunningham@me.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2017-01-30 17:00:44 +01:00
.url_get_short_seek = tcp_get_window_size,
.url_shutdown = tcp_shutdown,
.priv_data_size = sizeof(TCPContext),
.flags = URL_PROTOCOL_FLAG_NETWORK,
2014-10-09 22:05:27 +02:00
.priv_data_class = &tcp_class,
};