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>
32 /**********************************************************************
33 * GStreamer Default File Source
36 * This source uses mmap(2) to efficiently load data from a file.
37 * To do this without seriously polluting the applications' memory
38 * space, it must do so in smaller chunks, say 1-4MB at a time.
39 * Buffers are then subdivided from these mmap'd chunks, to directly
40 * make use of the mmap.
42 * To handle refcounting so that the mmap can be freed at the appropriate
43 * time, a buffer will be created for each mmap'd region, and all new
44 * buffers will be sub-buffers of this top-level buffer. As they are
45 * freed, the refcount goes down on the mmap'd buffer and its free()
46 * function is called, which will call munmap(2) on itself.
48 * If a buffer happens to cross the boundaries of an mmap'd region, we
49 * have to decide whether it's more efficient to copy the data into a
50 * new buffer, or mmap() just that buffer. There will have to be a
51 * breakpoint size to determine which will be done. The mmap() size
52 * has a lot to do with this as well, because you end up in double-
53 * jeopardy: the larger the outgoing buffer, the more data to copy when
54 * it overlaps, *and* the more frequently you'll have buffers that *do*
57 * Seeking is another tricky aspect to do efficiently. The initial
58 * implementation of this source won't make use of these features, however.
59 * The issue is that if an application seeks backwards in a file, *and*
60 * that region of the file is covered by an mmap that hasn't been fully
61 * deallocated, we really should re-use it. But keeping track of these
62 * regions is tricky because we have to lock the structure that holds
63 * them. We need to settle on a locking primitive (GMutex seems to be
64 * a really good option...), then we can do that.
68 GstElementDetails gst_filesrc_details = {
71 "Read from arbitrary point in a file",
73 "Erik Walthinsen <omega@cse.ogi.edu>",
78 #define GST_TYPE_FILESRC \
79 (gst_filesrc_get_type())
80 #define GST_FILESRC(obj) \
81 (G_TYPE_CHECK_INSTANCE_CAST((obj),GST_TYPE_FILESRC,GstFileSrc))
82 #define GST_FILESRC_CLASS(klass) \
83 (G_TYPE_CHECK_CLASS_CAST((klass),GST_TYPE_FILESRC,GstFileSrcClass))
84 #define GST_IS_FILESRC(obj) \
85 (G_TYPE_CHECK_INSTANCE_TYPE((obj),GST_TYPE_FILESRC))
86 #define GST_IS_FILESRC_CLASS(obj) \
87 (G_TYPE_CHECK_CLASS_TYPE((klass),GST_TYPE_FILESRC))
90 GST_FILESRC_OPEN = GST_ELEMENT_FLAG_LAST,
92 GST_FILESRC_FLAG_LAST = GST_ELEMENT_FLAG_LAST + 2,
95 typedef struct _GstFileSrc GstFileSrc;
96 typedef struct _GstFileSrcClass GstFileSrcClass;
102 guint pagesize; // system page size
104 gchar *filename; // filename
105 gint fd; // open file descriptor
106 off_t filelen; // what's the file length?
108 off_t curoffset; // current offset in file
109 off_t block_size; // bytes per read
110 gboolean touch; // whether to touch every page
116 GMutex *map_regions_lock;
119 struct _GstFileSrcClass {
120 GstElementClass parent_class;
124 /* FileSrc signals and args */
141 static void gst_filesrc_class_init (GstFileSrcClass *klass);
142 static void gst_filesrc_init (GstFileSrc *filesrc);
144 static void gst_filesrc_set_property (GObject *object, guint prop_id, const GValue *value, GParamSpec *pspec);
145 static void gst_filesrc_get_property (GObject *object, guint prop_id, GValue *value, GParamSpec *pspec);
147 static GstBuffer * gst_filesrc_get (GstPad *pad);
148 static gboolean gst_filesrc_srcpad_event (GstPad *pad, GstEventType event, gint64 location, guint32 data);
150 static GstElementStateReturn gst_filesrc_change_state (GstElement *element);
153 static GstElementClass *parent_class = NULL;
154 //static guint gst_filesrc_signals[LAST_SIGNAL] = { 0 };
157 gst_filesrc_get_type(void)
159 static GType filesrc_type = 0;
162 static const GTypeInfo filesrc_info = {
163 sizeof(GstFileSrcClass), NULL,
165 (GClassInitFunc)gst_filesrc_class_init,
170 (GInstanceInitFunc)gst_filesrc_init,
172 filesrc_type = g_type_register_static (GST_TYPE_ELEMENT, "GstFileSrc", &filesrc_info, 0);
178 gst_filesrc_class_init (GstFileSrcClass *klass)
180 GObjectClass *gobject_class;
181 GstElementClass *gstelement_class;
183 gobject_class = (GObjectClass*)klass;
184 gstelement_class = (GstElementClass*)klass;
186 parent_class = g_type_class_ref (GST_TYPE_ELEMENT);
188 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_LOCATION,
189 g_param_spec_string("location","File Location","Location of the file to read",
190 NULL,G_PARAM_READWRITE));
191 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_FILESIZE,
192 g_param_spec_ulong("filesize","File Size","Size of the file being read",
193 0,G_MAXULONG,0,G_PARAM_READABLE));
194 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_FD,
195 g_param_spec_int("fd","File-descriptor","File-descriptor for the file being read",
196 0,G_MAXINT,0,G_PARAM_READABLE));
197 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_BLOCKSIZE,
198 g_param_spec_ulong("blocksize","Block Size","Block size to read per buffer",
199 0,G_MAXULONG,4096,G_PARAM_READWRITE));
200 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_OFFSET,
201 g_param_spec_ulong("offset","File Offset","Byte offset of current read pointer",
202 0,G_MAXULONG,0,G_PARAM_READWRITE));
203 g_object_class_install_property(G_OBJECT_CLASS(klass), ARG_MAPSIZE,
204 g_param_spec_ulong("mmapsize","mmap() Block Size","Size in bytes of mmap()d regions",
205 0,G_MAXULONG,4*1048576,G_PARAM_READWRITE));
207 gobject_class->set_property = gst_filesrc_set_property;
208 gobject_class->get_property = gst_filesrc_get_property;
210 gstelement_class->change_state = gst_filesrc_change_state;
214 gst_filesrc_bufcmp (gconstpointer a, gconstpointer b)
216 // GstBuffer *bufa = (GstBuffer *)a, *bufb = (GstBuffer *)b;
218 // sort first by offset, then in reverse by size
219 if (GST_BUFFER_OFFSET(a) < GST_BUFFER_OFFSET(b)) return -1;
220 else if (GST_BUFFER_OFFSET(a) > GST_BUFFER_OFFSET(b)) return 1;
221 else if (GST_BUFFER_SIZE(a) > GST_BUFFER_SIZE(b)) return -1;
222 else if (GST_BUFFER_SIZE(a) < GST_BUFFER_SIZE(b)) return 1;
227 gst_filesrc_init (GstFileSrc *src)
229 src->srcpad = gst_pad_new ("src", GST_PAD_SRC);
230 gst_pad_set_get_function (src->srcpad,gst_filesrc_get);
231 gst_pad_set_event_function (src->srcpad,gst_filesrc_srcpad_event);
232 gst_element_add_pad (GST_ELEMENT (src), src->srcpad);
234 src->pagesize = getpagesize();
236 src->filename = NULL;
241 src->block_size = 4096;
245 src->mapsize = 4 * 1024 * 1024; // default is 4MB
247 src->map_regions = g_tree_new(gst_filesrc_bufcmp);
248 src->map_regions_lock = g_mutex_new();
253 gst_filesrc_set_property (GObject *object, guint prop_id, const GValue *value, GParamSpec *pspec)
257 /* it's not null if we got it, but it might not be ours */
258 g_return_if_fail (GST_IS_FILESRC (object));
260 src = GST_FILESRC (object);
264 /* the element must be stopped in order to do this */
265 g_return_if_fail (GST_STATE (src) < GST_STATE_PLAYING);
267 if (src->filename) g_free (src->filename);
268 /* clear the filename if we get a NULL (is that possible?) */
269 if (g_value_get_string (value) == NULL) {
270 gst_element_set_state (GST_ELEMENT (object), GST_STATE_NULL);
271 src->filename = NULL;
272 /* otherwise set the new filename */
274 src->filename = g_strdup (g_value_get_string (value));
278 src->block_size = g_value_get_ulong (value);
281 src->curoffset = g_value_get_ulong (value);
284 if ((src->mapsize % src->pagesize) == 0)
285 src->mapsize = g_value_get_ulong (value);
287 GST_INFO(0, "invalid mapsize, must a multiple of pagesize, which is %d\n",src->pagesize);
295 gst_filesrc_get_property (GObject *object, guint prop_id, GValue *value, GParamSpec *pspec)
299 /* it's not null if we got it, but it might not be ours */
300 g_return_if_fail (GST_IS_FILESRC (object));
302 src = GST_FILESRC (object);
306 g_value_set_string (value, src->filename);
309 g_value_set_ulong (value, src->filelen);
312 g_value_set_int (value, src->fd);
315 g_value_set_ulong (value, src->block_size);
318 g_value_set_ulong (value, src->curoffset);
321 g_value_set_ulong (value, src->mapsize);
324 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
330 gst_filesrc_free_parent_mmap (GstBuffer *buf)
332 GstFileSrc *src = GST_FILESRC(GST_BUFFER_POOL_PRIVATE(buf));
334 // fprintf(stderr,"freeing mmap()d buffer at %d+%d\n",GST_BUFFER_OFFSET(buf),GST_BUFFER_SIZE(buf));
336 // remove the buffer from the list of available mmap'd regions
337 g_mutex_lock(src->map_regions_lock);
338 g_tree_remove(src->map_regions,buf);
339 // check to see if the tree is empty
340 if (g_tree_nnodes(src->map_regions) == 0) {
341 // we have to free the bufferpool we don't have yet
343 g_mutex_unlock(src->map_regions_lock);
345 // now unmap the memory
346 munmap(GST_BUFFER_DATA(buf),GST_BUFFER_MAXSIZE(buf));
350 gst_filesrc_map_region (GstFileSrc *src, off_t offset, off_t size)
355 // fprintf(stderr,"mapping region %d+%d from file into memory\n",offset,size);
357 // time to allocate a new mapbuf
358 buf = gst_buffer_new();
359 // mmap() the data into this new buffer
360 GST_BUFFER_DATA(buf) = mmap (NULL, size, PROT_READ, MAP_SHARED, src->fd, offset);
361 if (GST_BUFFER_DATA(buf) == NULL) {
362 fprintf(stderr, "ERROR: gstfilesrc couldn't map file!\n");
363 } else if (GST_BUFFER_DATA(buf) == (void *)-1) {
364 perror("gstfilesrc:mmap()");
366 // madvise to tell the kernel what to do with it
367 retval = madvise(GST_BUFFER_DATA(buf),GST_BUFFER_SIZE(buf),MADV_SEQUENTIAL);
368 // fill in the rest of the fields
369 GST_BUFFER_FLAGS(buf) = GST_BUFFER_READONLY | GST_BUFFER_ORIGINAL;
370 GST_BUFFER_SIZE(buf) = size;
371 GST_BUFFER_MAXSIZE(buf) = size;
372 GST_BUFFER_OFFSET(buf) = offset;
373 GST_BUFFER_TIMESTAMP(buf) = -1LL;
374 GST_BUFFER_POOL_PRIVATE(buf) = src;
375 GST_BUFFER_FREE_FUNC(buf) = gst_filesrc_free_parent_mmap;
377 g_mutex_lock(src->map_regions_lock);
378 g_tree_insert(src->map_regions,buf,buf);
379 g_mutex_unlock(src->map_regions_lock);
385 gst_filesrc_map_small_region (GstFileSrc *src, off_t offset, off_t size)
387 int mod, mapbase, mapsize;
390 // printf("attempting to map a small buffer at %d+%d\n",offset,size);
392 // if the offset starts at a non-page boundary, we have to special case
393 if ((mod = offset % src->pagesize)) {
394 mapbase = offset - mod;
395 mapsize = ((size + mod + src->pagesize - 1) / src->pagesize) * src->pagesize;
396 // printf("not on page boundaries, resizing map to %d+%d\n",mapbase,mapsize);
397 map = gst_filesrc_map_region(src, mapbase, mapsize);
398 return gst_buffer_create_sub (map, offset - mapbase, size);
401 return gst_filesrc_map_region(src,offset,size);
409 // This allows us to search for a potential mmap region.
411 gst_filesrc_search_region_match (gpointer a, gpointer b)
413 GstFileSrcRegion *r = (GstFileSrcRegion *)b;
415 // trying to walk b down the tree, current node is a
416 if (r->offset < GST_BUFFER_OFFSET(a)) return -1;
417 else if (r->offset >= (GST_BUFFER_OFFSET(a) + GST_BUFFER_SIZE(a))) return 1;
418 else if ((r->offset + r->size) <= (GST_BUFFER_OFFSET(a) + GST_BUFFER_SIZE(a))) return 0;
425 * @pad: #GstPad to push a buffer from
427 * Push a new buffer from the filesrc at the current offset.
430 gst_filesrc_get (GstPad *pad)
433 GstBuffer *buf = NULL, *map;
434 off_t readend,readsize,mapstart,mapend;
435 gboolean eof = FALSE;
436 GstFileSrcRegion region;
439 g_return_val_if_fail (pad != NULL, NULL);
440 src = GST_FILESRC (gst_pad_get_parent (pad));
441 g_return_val_if_fail (GST_FLAG_IS_SET (src, GST_FILESRC_OPEN), NULL);
443 // calculate end pointers so we don't have to do so repeatedly later
444 readsize = src->block_size;
445 readend = src->curoffset + src->block_size; // note this is the byte *after* the read
446 mapstart = GST_BUFFER_OFFSET(src->mapbuf);
447 mapend = mapstart + GST_BUFFER_SIZE(src->mapbuf); // note this is the byte *after* the map
449 // check to see if we're going to overflow the end of the file
450 if (readend > src->filelen) {
451 readsize = src->filelen - src->curoffset;
452 readend = src->curoffset;
456 // if the start is past the mapstart
457 if (src->curoffset >= mapstart) {
458 // if the end is before the mapend, the buffer is in current mmap region...
459 // ('cause by definition if readend is in the buffer, so's readstart)
460 if (readend <= mapend) {
461 // printf("read buf %d+%d lives in current mapbuf %d+%d, creating subbuffer of mapbuf\n",
462 // src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
463 buf = gst_buffer_create_sub (src->mapbuf, src->curoffset - GST_BUFFER_OFFSET(src->mapbuf),
466 // if the start actually is within the current mmap region, map an overlap buffer
467 } else if (src->curoffset < mapend) {
468 // printf("read buf %d+%d starts in mapbuf %d+%d but ends outside, creating new mmap\n",
469 // src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
470 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
473 // the only other option is that buffer is totally outside, which means we search for it
475 // now we can assume that the start is *before* the current mmap region
476 // if the readend is past mapstart, we have two options
477 } else if (readend >= mapstart) {
478 // either the read buffer overlaps the start of the mmap region
479 // or the read buffer fully contains the current mmap region
480 // either way, it's really not relevant, we just create a new region anyway
481 // printf("read buf %d+%d starts before mapbuf %d+%d, but overlaps it\n",
482 // src->curoffset,readsize,GST_BUFFER_OFFSET(src->mapbuf),GST_BUFFER_SIZE(src->mapbuf));
483 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
486 // then deal with the case where the read buffer is totally outside
488 // first check to see if there's a map that covers the right region already
489 // printf("searching for mapbuf to cover %d+%d\n",src->curoffset,readsize);
490 region.offset = src->curoffset;
491 region.size = readsize;
492 map = g_tree_search(src->map_regions,gst_filesrc_search_region_match,®ion);
494 // if we found an exact match, subbuffer it
496 // printf("found mapbuf at %d+%d, creating subbuffer\n",GST_BUFFER_OFFSET(map),GST_BUFFER_SIZE(map));
497 buf = gst_buffer_create_sub (map, src->curoffset - GST_BUFFER_OFFSET(map), readsize);
499 // otherwise we need to create something out of thin air
501 // if the read buffer crosses a mmap region boundary, create a one-off region
502 if ((src->curoffset / src->mapsize) != (readend / src->mapsize)) {
503 // printf("read buf %d+%d crosses a %d-byte boundary, creating a one-off\n",
504 // src->curoffset,readsize,src->mapsize);
505 buf = gst_filesrc_map_small_region (src, src->curoffset, readsize);
507 // otherwise we will create a new mmap region and set it to the default
509 off_t nextmap = src->curoffset - (src->curoffset % src->mapsize);
510 // printf("read buf %d+%d in new mapbuf at %d+%d, mapping and subbuffering\n",
511 // src->curoffset,readsize,nextmap,src->mapsize);
512 // first, we're done with the old mapbuf
513 gst_buffer_unref(src->mapbuf);
515 src->mapbuf = gst_filesrc_map_region (src, nextmap, src->mapsize);
517 buf = gst_buffer_create_sub (src->mapbuf, src->curoffset - GST_BUFFER_OFFSET(src->mapbuf), readsize);
522 /* if we need to touch the buffer (to bring it into memory), do so */
524 for (i=0;i<GST_BUFFER_SIZE(buf);i+=src->pagesize)
525 *(GST_BUFFER_DATA(buf)+i) = *(GST_BUFFER_DATA(buf)+i);
530 /* we're done, return the buffer */
531 src->curoffset += GST_BUFFER_SIZE(buf);
535 /* open the file and mmap it, necessary to go to READY state */
537 gst_filesrc_open_file (GstFileSrc *src)
539 g_return_val_if_fail (!GST_FLAG_IS_SET (src ,GST_FILESRC_OPEN), FALSE);
541 GST_DEBUG(0, "opening file %s\n",src->filename);
544 src->fd = open (src->filename, O_RDONLY);
547 gst_element_error (GST_ELEMENT (src), g_strconcat("opening file \"", src->filename, "\"", NULL));
550 /* find the file length */
551 src->filelen = lseek (src->fd, 0, SEEK_END);
552 lseek (src->fd, 0, SEEK_SET);
554 // allocate the first mmap'd region
555 src->mapbuf = gst_filesrc_map_region (src, 0, src->mapsize);
559 GST_FLAG_SET (src, GST_FILESRC_OPEN);
564 /* unmap and close the file */
566 gst_filesrc_close_file (GstFileSrc *src)
568 g_return_if_fail (GST_FLAG_IS_SET (src, GST_FILESRC_OPEN));
573 /* zero out a lot of our state */
578 GST_FLAG_UNSET (src, GST_FILESRC_OPEN);
582 static GstElementStateReturn
583 gst_filesrc_change_state (GstElement *element)
585 g_return_val_if_fail (GST_IS_FILESRC (element), GST_STATE_FAILURE);
587 if (GST_STATE_PENDING (element) == GST_STATE_NULL) {
588 if (GST_FLAG_IS_SET (element, GST_FILESRC_OPEN))
589 gst_filesrc_close_file (GST_FILESRC (element));
591 if (!GST_FLAG_IS_SET (element, GST_FILESRC_OPEN)) {
592 if (!gst_filesrc_open_file (GST_FILESRC (element)))
593 return GST_STATE_FAILURE;
597 if (GST_ELEMENT_CLASS (parent_class)->change_state)
598 return GST_ELEMENT_CLASS (parent_class)->change_state (element);
600 return GST_STATE_SUCCESS;
604 gst_filesrc_srcpad_event(GstPad *pad, GstEventType event, gint64 location, guint32 data)
606 GstFileSrc *src = GST_FILESRC(GST_PAD_PARENT(pad));
608 if (event == GST_EVENT_SEEK) {
609 if (data == SEEK_SET) {
610 src->curoffset = (guint64)location;
611 } else if (data == SEEK_CUR) {
612 src->curoffset += (gint64)location;
613 } else if (data == SEEK_END) {
614 src->curoffset = src->filelen - (guint64)location;
616 // push a discontinuous event?