2 * Copyright (C) 1999,2000 Erik Walthinsen <omega@cse.ogi.edu>
3 * 2000 Wim Taymans <wtay@chello.be>
7 * This library is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Library General Public
9 * License as published by the Free Software Foundation; either
10 * version 2 of the License, or (at your option) any later version.
12 * This library is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Library General Public License for more details.
17 * You should have received a copy of the GNU Library General Public
18 * License along with this library; if not, write to the
19 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 * Boston, MA 02111-1307, USA.
25 #include <sys/types.h>
33 /**********************************************************************
34 * GStreamer Default File Source
37 * This source uses mmap(2) to efficiently load data from a file.
38 * To do this without seriously polluting the applications' memory
39 * space, it must do so in smaller chunks, say 1-4MB at a time.
40 * Buffers are then subdivided from these mmap'd chunks, to directly
41 * make use of the mmap.
43 * To handle refcounting so that the mmap can be freed at the appropriate
44 * time, a buffer will be created for each mmap'd region, and all new
45 * buffers will be sub-buffers of this top-level buffer. As they are
46 * freed, the refcount goes down on the mmap'd buffer and its free()
47 * function is called, which will call munmap(2) on itself.
49 * If a buffer happens to cross the boundaries of an mmap'd region, we
50 * have to decide whether it's more efficient to copy the data into a
51 * new buffer, or mmap() just that buffer. There will have to be a
52 * breakpoint size to determine which will be done. The mmap() size
53 * has a lot to do with this as well, because you end up in double-
54 * jeopardy: the larger the outgoing buffer, the more data to copy when
55 * it overlaps, *and* the more frequently you'll have buffers that *do*
58 * Seeking is another tricky aspect to do efficiently. The initial
59 * implementation of this source won't make use of these features, however.
60 * The issue is that if an application seeks backwards in a file, *and*
61 * that region of the file is covered by an mmap that hasn't been fully
62 * deallocated, we really should re-use it. But keeping track of these
63 * regions is tricky because we have to lock the structure that holds
64 * them. We need to settle on a locking primitive (GMutex seems to be
65 * a really good option...), then we can do that.
69 GstElementDetails gst_filesrc_details = {
72 "Read from arbitrary point in a file",
74 "Erik Walthinsen <omega@cse.ogi.edu>",
78 //#define fs_print(format,args...) g_print(format, ## args)
79 #define fs_print(format,args...)
82 #define GST_TYPE_FILESRC \
83 (gst_filesrc_get_type())
84 #define GST_FILESRC(obj) \
85 (G_TYPE_CHECK_INSTANCE_CAST((obj),GST_TYPE_FILESRC,GstFileSrc))
86 #define GST_FILESRC_CLASS(klass) \
87 (G_TYPE_CHECK_CLASS_CAST((klass),GST_TYPE_FILESRC,GstFileSrcClass))
88 #define GST_IS_FILESRC(obj) \
89 (G_TYPE_CHECK_INSTANCE_TYPE((obj),GST_TYPE_FILESRC))
90 #define GST_IS_FILESRC_CLASS(obj) \
91 (G_TYPE_CHECK_CLASS_TYPE((klass),GST_TYPE_FILESRC))
94 GST_FILESRC_OPEN = GST_ELEMENT_FLAG_LAST,
96 GST_FILESRC_FLAG_LAST = GST_ELEMENT_FLAG_LAST + 2,
99 typedef struct _GstFileSrc GstFileSrc;
100 typedef struct _GstFileSrcClass GstFileSrcClass;
106 guint pagesize; // system page size
108 gchar *filename; // filename
109 gint fd; // open file descriptor
110 off_t filelen; // what's the file length?
112 off_t curoffset; // current offset in file
113 off_t block_size; // bytes per read
114 gboolean touch; // whether to touch every page
120 GMutex *map_regions_lock;
122 gboolean seek_happened;
125 struct _GstFileSrcClass {
126 GstElementClass parent_class;
130 /* FileSrc signals and args */
148 static void gst_filesrc_class_init (GstFileSrcClass *klass);
149 static void gst_filesrc_init (GstFileSrc *filesrc);
151 static void gst_filesrc_set_property (GObject *object, guint prop_id, const GValue *value, GParamSpec *pspec);
152 static void gst_filesrc_get_property (GObject *object, guint prop_id, GValue *value, GParamSpec *pspec);
154 static GstBuffer * gst_filesrc_get (GstPad *pad);
155 static gboolean gst_filesrc_srcpad_event (GstPad *pad, GstEventType event, gint64 location, guint32 data);
157 static GstElementStateReturn gst_filesrc_change_state (GstElement *element);
160 static GstElementClass *parent_class = NULL;
161 //static guint gst_filesrc_signals[LAST_SIGNAL] = { 0 };
164 gst_filesrc_get_type(void)
166 static GType filesrc_type = 0;
169 static const GTypeInfo filesrc_info = {
170 sizeof(GstFileSrcClass), NULL,
172 (GClassInitFunc)gst_filesrc_class_init,
177 (GInstanceInitFunc)gst_filesrc_init,
179 filesrc_type = g_type_register_static (GST_TYPE_ELEMENT, "GstFileSrc", &filesrc_info, 0);
185 gst_filesrc_class_init (GstFileSrcClass *klass)
187 GObjectClass *gobject_class;
188 GstElementClass *gstelement_class;
190 gobject_class = (GObjectClass*)klass;
191 gstelement_class = (GstElementClass*)klass;
193 parent_class = g_type_class_ref (GST_TYPE_ELEMENT);
195 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_LOCATION,
196 g_param_spec_string("location","File Location","Location of the file to read",
197 NULL,G_PARAM_READWRITE));
198 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_FILESIZE,
199 g_param_spec_ulong("filesize","File Size","Size of the file being read",
200 0,G_MAXULONG,0,G_PARAM_READABLE));
201 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_FD,
202 g_param_spec_int("fd","File-descriptor","File-descriptor for the file being read",
203 0,G_MAXINT,0,G_PARAM_READABLE));
204 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_BLOCKSIZE,
205 g_param_spec_ulong("blocksize","Block Size","Block size to read per buffer",
206 0,G_MAXULONG,4096,G_PARAM_READWRITE));
207 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_OFFSET,
208 g_param_spec_ulong("offset","File Offset","Byte offset of current read pointer",
209 0,G_MAXULONG,0,G_PARAM_READWRITE));
210 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_MAPSIZE,
211 g_param_spec_ulong("mmapsize","mmap() Block Size","Size in bytes of mmap()d regions",
212 0,G_MAXULONG,4*1048576,G_PARAM_READWRITE));
213 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_TOUCH,
214 g_param_spec_boolean("touch","Touch read data","Touch data to force disk read before push()",
215 TRUE,G_PARAM_READWRITE));
217 gobject_class->set_property = gst_filesrc_set_property;
218 gobject_class->get_property = gst_filesrc_get_property;
220 gstelement_class->change_state = gst_filesrc_change_state;
224 gst_filesrc_bufcmp (gconstpointer a, gconstpointer b)
226 // GstBuffer *bufa = (GstBuffer *)a, *bufb = (GstBuffer *)b;
228 // sort first by offset, then in reverse by size
229 if (GST_BUFFER_OFFSET(a) < GST_BUFFER_OFFSET(b)) return -1;
230 else if (GST_BUFFER_OFFSET(a) > GST_BUFFER_OFFSET(b)) return 1;
231 else if (GST_BUFFER_SIZE(a) > GST_BUFFER_SIZE(b)) return -1;
232 else if (GST_BUFFER_SIZE(a) < GST_BUFFER_SIZE(b)) return 1;
237 gst_filesrc_init (GstFileSrc *src)
239 src->srcpad = gst_pad_new ("src", GST_PAD_SRC);
240 gst_pad_set_get_function (src->srcpad,gst_filesrc_get);
241 gst_pad_set_event_function (src->srcpad,gst_filesrc_srcpad_event);
242 gst_element_add_pad (GST_ELEMENT (src), src->srcpad);
244 src->pagesize = getpagesize();
246 src->filename = NULL;
251 src->block_size = 4096;
255 src->mapsize = 4 * 1024 * 1024; // default is 4MB
257 src->map_regions = g_tree_new(gst_filesrc_bufcmp);
258 src->map_regions_lock = g_mutex_new();
260 src->seek_happened = FALSE;
265 gst_filesrc_set_property (GObject *object, guint prop_id, const GValue *value, GParamSpec *pspec)
269 /* it's not null if we got it, but it might not be ours */
270 g_return_if_fail (GST_IS_FILESRC (object));
272 src = GST_FILESRC (object);
276 /* the element must be stopped in order to do this */
277 g_return_if_fail (GST_STATE (src) < GST_STATE_PLAYING);
279 if (src->filename) g_free (src->filename);
280 /* clear the filename if we get a NULL (is that possible?) */
281 if (g_value_get_string (value) == NULL) {
282 gst_element_set_state (GST_ELEMENT (object), GST_STATE_NULL);
283 src->filename = NULL;
284 /* otherwise set the new filename */
286 src->filename = g_strdup (g_value_get_string (value));
290 src->block_size = g_value_get_ulong (value);
293 src->curoffset = g_value_get_ulong (value);
296 if ((src->mapsize % src->pagesize) == 0)
297 src->mapsize = g_value_get_ulong (value);
299 GST_INFO(0, "invalid mapsize, must a multiple of pagesize, which is %d\n",src->pagesize);
302 src->touch = g_value_get_boolean (value);
310 gst_filesrc_get_property (GObject *object, guint prop_id, GValue *value, GParamSpec *pspec)
314 /* it's not null if we got it, but it might not be ours */
315 g_return_if_fail (GST_IS_FILESRC (object));
317 src = GST_FILESRC (object);
321 g_value_set_string (value, src->filename);
324 g_value_set_ulong (value, src->filelen);
327 g_value_set_int (value, src->fd);
330 g_value_set_ulong (value, src->block_size);
333 g_value_set_ulong (value, src->curoffset);
336 g_value_set_ulong (value, src->mapsize);
339 g_value_set_boolean (value, src->touch);
342 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
348 gst_filesrc_free_parent_mmap (GstBuffer *buf)
350 GstFileSrc *src = GST_FILESRC(GST_BUFFER_POOL_PRIVATE(buf));
352 fs_print ("freeing mmap()d buffer at %d+%d\n",GST_BUFFER_OFFSET(buf),GST_BUFFER_SIZE(buf));
354 // remove the buffer from the list of available mmap'd regions
355 g_mutex_lock(src->map_regions_lock);
356 g_tree_remove(src->map_regions,buf);
357 // check to see if the tree is empty
358 if (g_tree_nnodes(src->map_regions) == 0) {
359 // we have to free the bufferpool we don't have yet
361 g_mutex_unlock(src->map_regions_lock);
363 // now unmap the memory
364 munmap(GST_BUFFER_DATA(buf),GST_BUFFER_MAXSIZE(buf));
368 gst_filesrc_map_region (GstFileSrc *src, off_t offset, size_t size)
373 g_return_val_if_fail (offset >= 0, NULL);
375 fs_print ("mapping region %08lx+%08lx from file into memory\n",offset,size);
377 // time to allocate a new mapbuf
378 buf = gst_buffer_new();
379 // mmap() the data into this new buffer
380 GST_BUFFER_DATA(buf) = mmap (NULL, size, PROT_READ, MAP_SHARED, src->fd, offset);
381 if (GST_BUFFER_DATA(buf) == NULL) {
382 fprintf (stderr, "ERROR: gstfilesrc couldn't map file!\n");
383 } else if (GST_BUFFER_DATA(buf) == MAP_FAILED) {
384 g_error ("gstfilesrc mmap(0x%x, %d, 0x%llx) : %s",
385 size, src->fd, offset, sys_errlist[errno]);
387 // madvise to tell the kernel what to do with it
388 retval = madvise(GST_BUFFER_DATA(buf),GST_BUFFER_SIZE(buf),MADV_SEQUENTIAL);
389 // fill in the rest of the fields
390 GST_BUFFER_FLAGS(buf) = GST_BUFFER_READONLY | GST_BUFFER_ORIGINAL;
391 GST_BUFFER_SIZE(buf) = size;
392 GST_BUFFER_MAXSIZE(buf) = size;
393 GST_BUFFER_OFFSET(buf) = offset;
394 GST_BUFFER_TIMESTAMP(buf) = -1LL;
395 GST_BUFFER_POOL_PRIVATE(buf) = src;
396 GST_BUFFER_FREE_FUNC(buf) = gst_filesrc_free_parent_mmap;
398 g_mutex_lock(src->map_regions_lock);
399 g_tree_insert(src->map_regions,buf,buf);
400 g_mutex_unlock(src->map_regions_lock);
406 gst_filesrc_map_small_region (GstFileSrc *src, off_t offset, size_t size)
412 // printf("attempting to map a small buffer at %d+%d\n",offset,size);
414 // if the offset starts at a non-page boundary, we have to special case
415 if ((mod = offset % src->pagesize)) {
418 mapbase = offset - mod;
419 mapsize = ((size + mod + src->pagesize - 1) / src->pagesize) * src->pagesize;
420 // printf("not on page boundaries, resizing map to %d+%d\n",mapbase,mapsize);
421 map = gst_filesrc_map_region(src, mapbase, mapsize);
422 ret = gst_buffer_create_sub (map, offset - mapbase, size);
424 gst_buffer_unref (map);
429 return gst_filesrc_map_region(src,offset,size);
437 // This allows us to search for a potential mmap region.
439 gst_filesrc_search_region_match (gpointer a, gpointer b)
441 GstFileSrcRegion *r = (GstFileSrcRegion *)b;
443 // trying to walk b down the tree, current node is a
444 if (r->offset < GST_BUFFER_OFFSET(a)) return -1;
445 else if (r->offset >= (GST_BUFFER_OFFSET(a) + GST_BUFFER_SIZE(a))) return 1;
446 else if ((r->offset + r->size) <= (GST_BUFFER_OFFSET(a) + GST_BUFFER_SIZE(a))) return 0;
453 * @pad: #GstPad to push a buffer from
455 * Push a new buffer from the filesrc at the current offset.
458 gst_filesrc_get (GstPad *pad)
461 GstBuffer *buf = NULL, *map;
463 off_t readend,mapstart,mapend;
464 GstFileSrcRegion region;
467 g_return_val_if_fail (pad != NULL, NULL);
468 src = GST_FILESRC (gst_pad_get_parent (pad));
469 g_return_val_if_fail (GST_FLAG_IS_SET (src, GST_FILESRC_OPEN), NULL);
472 if (src->seek_happened) {
473 src->seek_happened = FALSE;
474 return gst_event_new(GST_EVENT_DISCONTINUOUS);
478 if (src->curoffset == src->filelen) {
479 gst_element_set_state(src,GST_STATE_PAUSED);
480 return gst_event_new(GST_EVENT_EOS);
483 // calculate end pointers so we don't have to do so repeatedly later
484 readsize = src->block_size;
485 readend = src->curoffset + src->block_size; // note this is the byte *after* the read
486 mapstart = GST_BUFFER_OFFSET(src->mapbuf);
487 mapend = mapstart + GST_BUFFER_SIZE(src->mapbuf); // note this is the byte *after* the map
489 // check to see if we're going to overflow the end of the file
490 if (readend > src->filelen) {
491 readsize = src->filelen - src->curoffset;
492 readend = src->curoffset;
495 // if the start is past the mapstart
496 if (src->curoffset >= mapstart) {
497 // if the end is before the mapend, the buffer is in current mmap region...
498 // ('cause by definition if readend is in the buffer, so's readstart)
499 if (readend <= mapend) {
500 fs_print ("read buf %d+%d lives in current mapbuf %d+%d, creating subbuffer of mapbuf\n",
501 src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
502 buf = gst_buffer_create_sub (src->mapbuf, src->curoffset - GST_BUFFER_OFFSET(src->mapbuf),
505 // if the start actually is within the current mmap region, map an overlap buffer
506 } else if (src->curoffset < mapend) {
507 fs_print ("read buf %d+%d starts in mapbuf %d+%d but ends outside, creating new mmap\n",
508 src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
509 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
512 // the only other option is that buffer is totally outside, which means we search for it
514 // now we can assume that the start is *before* the current mmap region
515 // if the readend is past mapstart, we have two options
516 } else if (readend >= mapstart) {
517 // either the read buffer overlaps the start of the mmap region
518 // or the read buffer fully contains the current mmap region
519 // either way, it's really not relevant, we just create a new region anyway
520 fs_print ("read buf %d+%d starts before mapbuf %d+%d, but overlaps it\n",
521 src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
522 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
525 // then deal with the case where the read buffer is totally outside
527 // first check to see if there's a map that covers the right region already
528 fs_print ("searching for mapbuf to cover %d+%d\n",src->curoffset,readsize);
529 region.offset = src->curoffset;
530 region.size = readsize;
531 map = g_tree_search (src->map_regions,
532 (GCompareFunc) gst_filesrc_search_region_match,
535 // if we found an exact match, subbuffer it
537 fs_print ("found mapbuf at %d+%d, creating subbuffer\n",GST_BUFFER_OFFSET(map),GST_BUFFER_SIZE(map));
538 buf = gst_buffer_create_sub (map, src->curoffset - GST_BUFFER_OFFSET(map), readsize);
540 // otherwise we need to create something out of thin air
542 // if the read buffer crosses a mmap region boundary, create a one-off region
543 if ((src->curoffset / src->mapsize) != (readend / src->mapsize)) {
544 fs_print ("read buf %d+%d crosses a %d-byte boundary, creating a one-off\n",
545 src->curoffset,readsize,src->mapsize);
546 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
548 // otherwise we will create a new mmap region and set it to the default
550 off_t nextmap = src->curoffset - (src->curoffset % src->mapsize);
551 fs_print ("read buf %d+%d in new mapbuf at %d+%d, mapping and subbuffering\n",
552 src->curoffset,readsize,nextmap,src->mapsize);
553 // first, we're done with the old mapbuf
554 gst_buffer_unref(src->mapbuf);
556 src->mapbuf = gst_filesrc_map_region (src, nextmap, src->mapsize);
558 buf = gst_buffer_create_sub (src->mapbuf, src->curoffset - GST_BUFFER_OFFSET(src->mapbuf), readsize);
563 /* if we need to touch the buffer (to bring it into memory), do so */
565 for (i=0;i<GST_BUFFER_SIZE(buf);i+=src->pagesize)
566 *(GST_BUFFER_DATA(buf)+i) = *(GST_BUFFER_DATA(buf)+i);
569 /* we're done, return the buffer */
570 src->curoffset += GST_BUFFER_SIZE(buf);
574 /* open the file and mmap it, necessary to go to READY state */
576 gst_filesrc_open_file (GstFileSrc *src)
578 g_return_val_if_fail (!GST_FLAG_IS_SET (src ,GST_FILESRC_OPEN), FALSE);
580 GST_DEBUG(0, "opening file %s\n",src->filename);
583 src->fd = open (src->filename, O_RDONLY);
586 gst_element_error (GST_ELEMENT (src), g_strconcat("opening file \"", src->filename, "\"", NULL));
589 /* find the file length */
590 src->filelen = lseek (src->fd, 0, SEEK_END);
591 lseek (src->fd, 0, SEEK_SET);
593 // allocate the first mmap'd region
594 src->mapbuf = gst_filesrc_map_region (src, 0, src->mapsize);
598 GST_FLAG_SET (src, GST_FILESRC_OPEN);
603 /* unmap and close the file */
605 gst_filesrc_close_file (GstFileSrc *src)
607 g_return_if_fail (GST_FLAG_IS_SET (src, GST_FILESRC_OPEN));
613 /* zero out a lot of our state */
618 GST_FLAG_UNSET (src, GST_FILESRC_OPEN);
622 static GstElementStateReturn
623 gst_filesrc_change_state (GstElement *element)
625 g_return_val_if_fail (GST_IS_FILESRC (element), GST_STATE_FAILURE);
627 if (GST_STATE_PENDING (element) == GST_STATE_NULL) {
628 if (GST_FLAG_IS_SET (element, GST_FILESRC_OPEN))
629 gst_filesrc_close_file (GST_FILESRC (element));
631 if (!GST_FLAG_IS_SET (element, GST_FILESRC_OPEN)) {
632 if (!gst_filesrc_open_file (GST_FILESRC (element)))
633 return GST_STATE_FAILURE;
637 if (GST_ELEMENT_CLASS (parent_class)->change_state)
638 return GST_ELEMENT_CLASS (parent_class)->change_state (element);
640 return GST_STATE_SUCCESS;
644 gst_filesrc_srcpad_event(GstPad *pad, GstEventType event, gint64 location, guint32 data)
646 GstFileSrc *src = GST_FILESRC(GST_PAD_PARENT(pad));
648 if (event == GST_EVENT_SEEK) {
649 if (data == SEEK_SET) {
650 src->curoffset = (guint64)location;
651 } else if (data == SEEK_CUR) {
652 src->curoffset += (gint64)location;
653 } else if (data == SEEK_END) {
654 src->curoffset = src->filelen - (guint64)location;
656 src->seek_happened = TRUE;
657 // push a discontinuous event?