Commit graph

1431 commits

Author SHA1 Message Date
Mike Ruprecht
d397ea972d funnel: Only emit EOS event if all sinkpads have received one
If multiple sources are plugged into the funnel and one of the
sources emits an EOS, that event is propogated through the funnel
even though other sources connected to the funnel may still be
pushing data. This patch waits to send an EOS event until the
funnel has received an EOS event on each sinkpad.
2012-05-31 12:58:00 -04:00
Mike Ruprecht
03ce735b07 funnel: Fix buffer leak 2012-05-31 12:58:00 -04:00
Andre Moreira Magalhaes (andrunko)
8f267451f4 inputselector: Don't try to sync on the segment if it has no TIME format
...and wait until it is actually configured and has a format before
trying to sync.
2012-05-31 17:45:29 +02:00
Sebastian Dröge
e8e9c45c21 inputselector: No need to broadcast the signal in flush-stop
Everything stopped at this point already.
2012-05-31 17:03:54 +02:00
Bastien Nocera
cdb3087746 queue2: Fix property name in the docs
temp-template, not temp-tmpl

https://bugzilla.gnome.org/show_bug.cgi?id=677170
2012-05-31 12:56:47 +02:00
Andre Moreira Magalhaes (andrunko)
cd798a0a52 inputselector: Properly sync when changing streams
This adds properties to use the clock time for deciding when
to drop buffers for inactive pads and a property to buffer all
not rendered buffers for the active pad to allow pad switching
without losing any buffers at all.
2012-05-31 11:49:29 +02:00
Idar Tollefsen
7a70c91ac3 configure: explicitly check for sys/mman.h header
And use header-specific guards.

https://bugzilla.gnome.org/show_bug.cgi?id=667292
2012-05-15 19:05:42 +01:00
Tim-Philipp Müller
f85885474c filesrc: remove unused or questionable madvise() calls in deprecated mmap code paths
use-mmap functionality has been removed in 0.11 and doesn't really
have any performance advantages in most cases anyway. Deprecate
"sequential" property to hint sequential access in mmap mode and
make it non-functional. Chances are no one was using this ever
anyway, or inappropriately.

Remove madvise() calls. We would need to do extensive configure
checks to handle these properly on different systems, and it
doesn't seem worth adding that for code that's already removed
in 0.11 or questionable anyway (like madvise DONTNEED right
before munmap).

https://bugzilla.gnome.org/show_bug.cgi?id=667292
2012-05-15 18:43:59 +01:00
Wim Taymans
bdda711d03 queue2: signal delete before waiting
When we don't have the requested data in the ringbuffer and we move our read
pointer to the requested position, signal the delete cond to inform the writer
that we changed the current fill level. If we don't, the writer might stay
blocked and we might wait forever.
2012-04-11 12:45:46 +02:00
Wim Taymans
ccbeb19881 queue2: update current read position before waiting
When we don't have enough bytes in the ringbuffer to satisfy the current
request, first update the current read position before waiting. If we don't do
that, the ringbuffer might appear full and the writer will never write more
bytes to wake us up.
2012-04-11 12:45:41 +02:00
Wim Taymans
463a8d9abf queue2: add range only on sinkpad
Only add the range when we receive a segment event on the sinkpad. The add_range
method will modify the write position, which only makes sense to do on the
sinkpad.
2012-04-11 12:45:37 +02:00
Wim Taymans
e3ca88ea7b queue2: fix debug message
We're not writing to the offset of the buffer
2012-04-11 12:45:31 +02:00
Wim Taymans
30c425c29f queue2: add_range already updates the level 2012-04-11 12:45:24 +02:00
Wim Taymans
2e28804a6e queue2: clear seeking flag in all cases
Also clear the seeking flag when downstream is in pull mode.
2012-04-10 16:06:41 +02:00
Wim Taymans
b7a8051d95 queue2: set seeking flag with the queue lock 2012-04-10 12:44:38 +02:00
Wim Taymans
5755e24379 queue2: Keep track of the seeking state
Set the seeking flag right before we send a seek event upstream and discard all
data untill we see a flush-stop again. We need to do this because we activate
the range that we seek to immediately after sending the seek event and it is
possible that we receive data in our chain function from before the seek
which would then be added to the wrong range resulting in data corruption.
2012-04-10 12:44:35 +02:00
Wim Taymans
49a4b801ac queue2: make range on newsegment for ringbuffer
When using the ringbuffer, handle the newsegment event like we handle it when
using the temp-file mode: create a new range for the new byte segment. The new
segment should normally already be created when we do a seek.
2012-04-10 12:44:30 +02:00
Sebastian Dröge
829eafad0e multiqueue: Don't use buffer after pushing it downstream 2012-04-09 16:40:47 +02:00
Tim-Philipp Müller
d05d29d0c9 filesrc: set default block size from local define
Doesn't actually change the default value, just makes use of the
define there is. Superficial testing with fakesink and jpegdec did
not reveal improved performance for bigger block sizes, so leave
default as it is.
2012-04-07 15:21:29 +01:00
Wim Taymans
4638f15de2 queue2: don't update the current reading_pos in flush
A flush from the upstream element should not make buffering go to 0, the next
pull request might be inside a range that we have and then we don't need to
buffer at all. If the next pull is outside anything we have, buffering will
happen as usual anyway.
2012-04-06 16:46:58 +02:00
Wim Taymans
d8a61930af queue2: check the pad mode on the right pad 2012-04-06 12:32:49 +02:00
Wim Taymans
f06458aabc queue2: forward flush events correctly
We want to forward the flush events received on the sinkpad whenever the srcpad
is activated in pushmode, which can also happen when using the RINGBUFFER or
DOWNLOAD mode and downstream failed to activate us in pull mode.
2012-04-06 12:24:22 +02:00
Wim Taymans
67ccad0621 queue2: stop waiting for more data after EOS
When we have EOS, read the remaining bytes in the buffer and make sure we don't
wait for more data. Also clip the output buffer to the amount of remaining
bytes.
2012-04-05 10:17:46 +02:00
Wim Taymans
9a1185673e queue2: check for filled buffer correctly
When using the ringbuffer mode, the buffer is filled when we reached the
max_level.bytes mark or the total size of the ringbuffer, whichever is smaller.
2012-04-05 10:16:52 +02:00
Wim Taymans
0bb0c68566 queue2: avoid waiting for a filled buffer
Use a threshold variable to hold the maximum distance from the current position
for with we will wait instead of doing a seek.
When using the ringbuffer and the requested offset is not available, avoid
waiting until the complete ringbuffer is filled but instead do a seek when the
requested data is further than the threshold.
Avoid doing the seek twice in the ringbuffer case.
Use the same threshold for ringbuffer and download buffering.
2012-04-05 10:16:46 +02:00
Sebastian Dröge
9441e711df multiqueue: Wake up all not-linked streams when a stream switches from linked to not-linked
We reset all the waiting streams, let them push another buffer to
see if they're now active again. This allows faster switching
between streams and prevents deadlocks if downstream does any
waiting too.

Also improve locking a bit, srcresult must be protected by the
multiqueue lock too because it's used/set from random threads.
2012-03-29 14:49:53 +02:00
Sebastian Dröge
17e691421f multiqueue: Recompute high-time too when flushing, not only high-id 2012-03-29 14:47:57 +02:00
Sebastian Dröge
d03e3d7772 inputselector: Only wait until the active pad's running time is reached if the active pad already saw data
Otherwise we might block forever because upstream (e.g. multiqueue) is waiting
for the previously active stream to return forever (which is waiting here
in inputselector) before pushing something on the newly selected stream.
2012-03-29 14:47:57 +02:00
Sebastian Dröge
f000bbe924 tee: Fix 'use of logical '&&' with constant operand' compiler warning
This is actually a real bug.
2012-03-06 12:27:33 +01:00
Tim-Philipp Müller
4594e4311d fdsink: implement GstBaseSink::query instead of messing with the pad 2012-02-25 15:18:00 +00:00
Tim-Philipp Müller
965c610028 fdsink: implement SEEKING query
We may or may not support seeking. stdout to a
terminal doesn't support seeking, for example, but
... ! fdsink > file.foo just might.
2012-02-25 15:08:55 +00:00
Tim-Philipp Müller
82a9544385 filesink: implement SEEKING query
We may or may not do seeking, depends on the
output file/device really, it doesn't have to
be a file after all.
2012-02-25 15:07:43 +00:00
Tim-Philipp Müller
c5b335a0e2 fakesink: answer SEEKING query
We don't do seeking, in case anyone wants to know.
2012-02-25 15:07:05 +00:00
Idar Tollefsen
b9b3795c3c plugins: explicitly cast initialization values to the correct type
https://bugzilla.gnome.org/show_bug.cgi?id=667288
2012-01-04 23:59:57 +00:00
Thiago Santos
71fe1ad070 output-selector: Do not false warn about unlinked pad
When output-selector didn't receive a newsegment event it would
warn about pad being unlinked when switching pads. Making the logs
wrong and misleading.
2011-12-29 16:59:29 -03:00
Tim-Philipp Müller
8a932dbca6 filesrc: return any remaining data on EOS before returning FLOW_UNEXPECTED 2011-12-25 12:42:11 +00:00
Tim-Philipp Müller
cf5de908fb filesrc: minor cosmetic changes
Rename woffset variable, maintain separate bytes_read in addition
to length variable.
2011-12-25 12:29:46 +00:00
Vincent Penquerc'h
431fc714c9 filesrc: do not mistake short reads for EOS
While local filesystems will usually not cause short reads,
this may happen on seekable files on some remote filesystems.
Instead, loop till we get the requested amount of data, or
an actual EOS (ie, 0 bytes).

https://bugzilla.gnome.org/show_bug.cgi?id=665921
2011-12-12 13:35:06 +00:00
Sebastian Dröge
47de97df84 inputselector: Don't send a NEWSEGMENT event if a buffer arrived before the segment was configured 2011-12-07 11:01:49 +01:00
Sebastian Dröge
7952a54299 outputselector: Don't send last segment/buffer when no segment was configured yet 2011-12-07 11:01:31 +01:00
Sebastian Dröge
c8909ab059 fakesink: Make event/buffer verbose output consistent with identity 2011-12-06 14:55:57 +01:00
Sebastian Dröge
b7b84671ba identity: Print buffer flags in the verbose output 2011-12-06 14:55:57 +01:00
Sebastian Dröge
675e54f1f4 fakesrc: Make event/buffer verbose output consistent with identity and print buffer flags 2011-12-06 14:55:57 +01:00
Tim-Philipp Müller
2666450864 Suppress deprecation warnings in selected files, for g_static_rec_mutex_* mostly
GStaticRecMutex is part of our API/ABI, not much we can do here in 0.10.
2011-12-04 14:38:26 +00:00
Tim-Philipp Müller
5889260d5a Work around deprecated thread API in glib master
Add private replacements for deprecated functions such as
g_mutex_new(), g_mutex_free(), g_cond_new() etc., mostly
to avoid the deprecation warnings. We can't change most of
these in 0.10 because they're part of our API and ABI.
2011-12-04 13:35:38 +00:00
Mark Nauwelaerts
bfcb5a43a6 identity: unlock clock wait when appropriate
... notably FLUSH and state change to READY.
2011-12-02 13:41:54 +01:00
Tim-Philipp Müller
3b28af5171 queue2: fix refactoring of draining-on-eos, munge flow return to FLOW_OK 2011-11-28 01:10:20 +00:00
Tim-Philipp Müller
2a3e7c4feb queue2: add bufferlist support
We want to maintain buffer lists if possible.
2011-11-28 00:16:40 +00:00
Tim-Philipp Müller
d663e18402 queue2: split out draining of queue on FLOW_UNEXPECTED into separate function 2011-11-28 00:16:40 +00:00
Tim-Philipp Müller
c9df12754d queue2: pass item type enum to _enqueue instead of simple isbuffer boolean
Avoids some unnecessary GST_IS_EVENT()
2011-11-28 00:16:40 +00:00