From patchwork Mon Jan 30 16:00:44 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Joel Cunningham X-Patchwork-Id: 2373 Delivered-To: ffmpegpatchwork@gmail.com Received: by 10.103.89.21 with SMTP id n21csp1519643vsb; Mon, 30 Jan 2017 08:01:21 -0800 (PST) X-Received: by 10.28.20.144 with SMTP id 138mr9481236wmu.41.1485792081644; Mon, 30 Jan 2017 08:01:21 -0800 (PST) Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id j133si13873237wma.124.2017.01.30.08.01.21; Mon, 30 Jan 2017 08:01:21 -0800 (PST) Received-SPF: pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) client-ip=79.124.17.100; Authentication-Results: mx.google.com; dkim=neutral (body hash did not verify) header.i=@me.com; spf=pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) smtp.mailfrom=ffmpeg-devel-bounces@ffmpeg.org; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=me.com Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 94CEF68A5AD; Mon, 30 Jan 2017 18:01:16 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from st11p00im-asmtp004.me.com (st11p00im-asmtp004.me.com [17.172.80.98]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id B142468A559 for ; Mon, 30 Jan 2017 18:01:09 +0200 (EET) Received: from process-dkim-sign-daemon.st11p00im-asmtp004.me.com by st11p00im-asmtp004.me.com (Oracle Communications Messaging Server 7.0.5.38.0 64bit (built Feb 26 2016)) id <0OKL00300OH0HU00@st11p00im-asmtp004.me.com> for ffmpeg-devel@ffmpeg.org; Mon, 30 Jan 2017 16:00:57 +0000 (GMT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=me.com; s=4d515a; t=1485792057; bh=BEbAtDaNzsLnttuDSPvL22LIog6Je8vPr5r5aNXYrAM=; h=From:To:Subject:Date:Message-id:MIME-version; b=arfyZXf5iNnbVrz3Nt4BfL/k4NQvElGDOoGEXNfQ+oyTNSGgq27OpdhDJ7Up732cr m1LpglTJ/GunGVXWfdqURghtU6U2ry5OPt2aAjG+oc2piGozqs+6z9WKKeJfeM6BTR FsXoNjAWRIk6e/MNWQD8RGv3La5tJWnk3yyftXksaPllvA+fEIYRne0wfLANyPyOJf 4+olB1yjikKCLddSHFeTDMdWcrBJIqZG4bxLzLDW63CAaOAse71Ae7PP9YlDJaX6XD JGRB9EXWQQescyLO1LBIwZD9w+LO1whIToHhYAI5xv4irKrL6ig3woYCZUQ5fmwTg4 r1TP1pE0VkmFA== Received: from icloud.com ([127.0.0.1]) by st11p00im-asmtp004.me.com (Oracle Communications Messaging Server 7.0.5.38.0 64bit (built Feb 26 2016)) with ESMTPSA id <0OKL00HV8OHJ1Y30@st11p00im-asmtp004.me.com> for ffmpeg-devel@ffmpeg.org; Mon, 30 Jan 2017 16:00:56 +0000 (GMT) X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10432:,, definitions=2017-01-30_10:,, signatures=0 X-Proofpoint-Spam-Details: rule=notspam policy=default score=0 spamscore=0 clxscore=1034 suspectscore=4 malwarescore=0 phishscore=0 adultscore=0 bulkscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.0.1-1603290000 definitions=main-1701300153 From: Joel Cunningham To: ffmpeg-devel@ffmpeg.org Date: Mon, 30 Jan 2017 10:00:44 -0600 Message-id: <20170130160044.27794-1-joel.cunningham@me.com> X-Mailer: git-send-email 2.10.0 In-reply-to: <20170127005218.GU4698@nb4> References: <20170127005218.GU4698@nb4> MIME-version: 1.0 Subject: [FFmpeg-devel] [PATCH] HTTP: improve performance by reducing forward seeks X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Cc: Joel Cunningham Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" 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 --- libavformat/avio.c | 7 +++++++ libavformat/avio.h | 6 ++++++ libavformat/aviobuf.c | 19 ++++++++++++++++++- libavformat/http.c | 8 ++++++++ libavformat/tcp.c | 21 +++++++++++++++++++++ libavformat/url.h | 8 ++++++++ 6 files changed, 68 insertions(+), 1 deletion(-) diff --git a/libavformat/avio.c b/libavformat/avio.c index 3606eb0..62233a6 100644 --- a/libavformat/avio.c +++ b/libavformat/avio.c @@ -645,6 +645,13 @@ int ffurl_get_multi_file_handle(URLContext *h, int **handles, int *numhandles) return h->prot->url_get_multi_file_handle(h, handles, numhandles); } +int ffurl_get_short_seek(URLContext *h) +{ + if (!h->prot->url_get_short_seek) + return AVERROR(ENOSYS); + return h->prot->url_get_short_seek(h); +} + int ffurl_shutdown(URLContext *h, int flags) { if (!h->prot->url_shutdown) diff --git a/libavformat/avio.h b/libavformat/avio.h index e2cb4af..8040094 100644 --- a/libavformat/avio.h +++ b/libavformat/avio.h @@ -313,6 +313,12 @@ typedef struct AVIOContext { */ enum AVIODataMarkerType current_type; int64_t last_time; + + /** + * A callback that is used instead of short_seek_threshold. + * This is current internal only, do not use from outside. + */ + int (*short_seek_get)(void *opaque); } AVIOContext; /** diff --git a/libavformat/aviobuf.c b/libavformat/aviobuf.c index bf7e5f8..4ade4d0 100644 --- a/libavformat/aviobuf.c +++ b/libavformat/aviobuf.c @@ -119,6 +119,7 @@ int ffio_init_context(AVIOContext *s, s->ignore_boundary_point = 0; s->current_type = AVIO_DATA_MARKER_UNKNOWN; s->last_time = AV_NOPTS_VALUE; + s->short_seek_get = NULL; return 0; } @@ -233,6 +234,7 @@ int64_t avio_seek(AVIOContext *s, int64_t offset, int whence) int64_t pos; int force = whence & AVSEEK_FORCE; int buffer_size; + int short_seek; whence &= ~AVSEEK_FORCE; if(!s) @@ -254,13 +256,21 @@ int64_t avio_seek(AVIOContext *s, int64_t offset, int whence) if (offset < 0) return AVERROR(EINVAL); + if (s->short_seek_get) { + short_seek = s->short_seek_get(s->opaque); + /* fallback to default short seek */ + if (short_seek <= 0) + short_seek = s->short_seek_threshold; + } else + short_seek = s->short_seek_threshold; + offset1 = offset - pos; // "offset1" is the relative offset from the beginning of s->buffer if (!s->must_flush && (!s->direct || !s->seek) && offset1 >= 0 && offset1 <= buffer_size - s->write_flag) { /* can do the seek inside the buffer */ s->buf_ptr = s->buffer + offset1; } else if ((!s->seekable || - offset1 <= buffer_size + s->short_seek_threshold) && + offset1 <= buffer_size + short_seek) && !s->write_flag && offset1 >= 0 && (!s->direct || !s->seek) && (whence != SEEK_END || force)) { @@ -858,6 +868,12 @@ static int64_t io_seek(void *opaque, int64_t offset, int whence) return ffurl_seek(internal->h, offset, whence); } +static int io_short_seek(void *opaque) +{ + AVIOInternal *internal = opaque; + return ffurl_get_short_seek(internal->h); +} + static int io_read_pause(void *opaque, int pause) { AVIOInternal *internal = opaque; @@ -919,6 +935,7 @@ int ffio_fdopen(AVIOContext **s, URLContext *h) (*s)->read_pause = io_read_pause; (*s)->read_seek = io_read_seek; } + (*s)->short_seek_get = io_short_seek; (*s)->av_class = &ff_avio_class; return 0; fail: diff --git a/libavformat/http.c b/libavformat/http.c index 944a6cf..b354c87 100644 --- a/libavformat/http.c +++ b/libavformat/http.c @@ -1524,6 +1524,12 @@ static int http_get_file_handle(URLContext *h) return ffurl_get_file_handle(s->hd); } +static int http_get_short_seek(URLContext *h) +{ + HTTPContext *s = h->priv_data; + return ffurl_get_short_seek(s->hd); +} + #define HTTP_CLASS(flavor) \ static const AVClass flavor ## _context_class = { \ .class_name = # flavor, \ @@ -1545,6 +1551,7 @@ const URLProtocol ff_http_protocol = { .url_seek = http_seek, .url_close = http_close, .url_get_file_handle = http_get_file_handle, + .url_get_short_seek = http_get_short_seek, .url_shutdown = http_shutdown, .priv_data_size = sizeof(HTTPContext), .priv_data_class = &http_context_class, @@ -1564,6 +1571,7 @@ const URLProtocol ff_https_protocol = { .url_seek = http_seek, .url_close = http_close, .url_get_file_handle = http_get_file_handle, + .url_get_short_seek = http_get_short_seek, .url_shutdown = http_shutdown, .priv_data_size = sizeof(HTTPContext), .priv_data_class = &https_context_class, diff --git a/libavformat/tcp.c b/libavformat/tcp.c index 5f00ba7..3055e48 100644 --- a/libavformat/tcp.c +++ b/libavformat/tcp.c @@ -266,6 +266,26 @@ static int tcp_get_file_handle(URLContext *h) return s->fd; } +static int tcp_get_window_size(URLContext *h) +{ + TCPContext *s = h->priv_data; + int avail; + int avail_len = sizeof(avail); + +#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, @@ -274,6 +294,7 @@ const URLProtocol ff_tcp_protocol = { .url_write = tcp_write, .url_close = tcp_close, .url_get_file_handle = tcp_get_file_handle, + .url_get_short_seek = tcp_get_window_size, .url_shutdown = tcp_shutdown, .priv_data_size = sizeof(TCPContext), .flags = URL_PROTOCOL_FLAG_NETWORK, diff --git a/libavformat/url.h b/libavformat/url.h index 5c50245..910f1e0 100644 --- a/libavformat/url.h +++ b/libavformat/url.h @@ -84,6 +84,7 @@ typedef struct URLProtocol { int (*url_get_file_handle)(URLContext *h); int (*url_get_multi_file_handle)(URLContext *h, int **handles, int *numhandles); + int (*url_get_short_seek)(URLContext *h); int (*url_shutdown)(URLContext *h, int flags); int priv_data_size; const AVClass *priv_data_class; @@ -249,6 +250,13 @@ int ffurl_get_file_handle(URLContext *h); int ffurl_get_multi_file_handle(URLContext *h, int **handles, int *numhandles); /** + * Return the current short seek threshold value for this URL. + * + * @return threshold (>0) on success or <=0 on error. + */ +int ffurl_get_short_seek(URLContext *h); + +/** * Signal the URLContext that we are done reading or writing the stream. * * @param h pointer to the resource