WebKit's websrc depends on the main-thread for download completion
rendezvous. This exposed a number of deadlocks in adaptivedemux due to
it holding the MANIFEST_LOCK during network requests, and also needing
to hold it to change_state and resolve queries, which frequently occur
during these download windows.
Make demux->running MT-safe so that it can be accessed without using the
MANIFEST_LOCK. In case a source is downloading and requires a MT-thread
notification for completion of the fragment download, a state change
during this download window will deadlock unless we cancel the downloads
and ensure they are not restarted before we finish the state-change.
Also make demux->priv->have_manifest MT-safe. A duration query happening
in the window described above can deadlock for the same reason. Other
src queries (like SEEKING) that happen in this window also could
deadlock, but I haven't hit this scenario.
Increase granularity of API_LOCK'ing in change_state as well. We need to
cancel downloads before trying to take this lock, since sink events
(EOS) will hold it before starting a fragment download.
For each lib we build export its own API in headers when we're
building it, otherwise import the API from the headers.
This fixes linker warnings on Windows when building with MSVC.
The problem was that we had defined all GST_*_API decorators
unconditionally to GST_EXPORT. This was intentional and only
supposed to be temporary, but caused linker warnings because
we tell the linker that we want to export all symbols even
those from externall DLLs, and when the linker notices that
they were in external DLLS and not present locally it warns.
What we need to do when building each library is: export
the library's own symbols and import all other symbols. To
this end we define e.g. BUILDING_GST_FOO and then we define
the GST_FOO_API decorator either to export or to import
symbols depending on whether BUILDING_GST_FOO is set or not.
That way external users of each library API automatically
get the import.
While we're at it, add new GST_API_EXPORT in config.h and use
that for GST_*_API decorators instead of GST_EXPORT.
The right export define depends on the toolchain and whether
we're using -fvisibility=hidden or not, so it's better to set it
to the right thing directly than hard-coding a compiler whitelist
in the public header.
We put the export define into config.h instead of passing it via the
command line to the compiler because it might contain spaces and brackets
and in the autotools scenario we'd have to pass that through multiple
layers of plumbing and Makefile/shell escaping and we're just not going
to be *that* lucky.
The export define is only used if we're compiling our lib, not by external
users of the lib headers, so it's not a problem to put it into config.h
Also, this means all .c files of libs need to include config.h
to get the export marker defined, so fix up a few that didn't
include config.h.
This commit depends on a common submodule commit that makes gst-glib-gen.mak
add an #include "config.h" to generated enum/marshal .c files for the
autotools build.
https://bugzilla.gnome.org/show_bug.cgi?id=797185
If connection-speed property is in use, this value should be used as the
current download rate since subclasses might read it to figure out
which playlist variant they will use.
https://bugzilla.gnome.org/show_bug.cgi?id=784592
Regardless of LIVE or VOD, "a manifest has next period but
currently EOSed" state is meaning that it's time to advance period.
Previous behavior of adpativedemux, however, was able to period
advancing only for VOD case, since the adaptivedemux tried to
update and wait new manifest without respecting existence of the next period.
https://bugzilla.gnome.org/show_bug.cgi?id=781183
We need different export decorators for the different libs.
For now no actual change though, just rename before the release,
and add prelude headers to define the new decorator to GST_EXPORT.
When switching bitrates we set the old streams as cancelled, but it
could also be confused with a cancel due to other reasons (as an error)
and it would lead the element to stop the pipeline mistankely. This
would happen when the stream being replaced was waiting for a manifest
update on live. Ss make it sure that we are stopping for switching
bitrates to avoid erroring out.
https://bugzilla.gnome.org/show_bug.cgi?id=789457
Scenario:
A manifest starts out in live mode but then the recording is finalized
and a subsequent update changes the state to a non-live manifest when
the server has finished recording/transcoding/whatever with the full
list of fragments.
Without this patch, the manifest update task is never stopped on the
live->non-live transition and will busy loop, burning through one CPU
core.
https://bugzilla.gnome.org/show_bug.cgi?id=786275
This commit ensures that the idle probe which GstAdaptiveDemuxStream
adds to the upstream source pad is removed after use. Previously a new
probe was added to the pad whenever a fragment was downloaded, meaning
the number of pad probe callbacks being executed increased continually.
https://bugzilla.gnome.org/show_bug.cgi?id=785957
There can be twice as many stream tasks running as there are output
pads for playback of variant HLS playlists. Half of them are the
current pads, and the other half are the pads that are about to be
switched to due to a bitrate change.
The old code only stopped the current streams which could result
in a deadlock on stopping the pipeline. The changes force stopping
and joining of any prepared streams too.
https://bugzilla.gnome.org/show_bug.cgi?id=785987
This is a workaround for a regression introduced by
f4190a49c0
( adaptivedemux: Check live seeking range more often )
The goal of the previous commit was to be able to cope with non-1.0
rates on live streams which have a "seeking window" (i.e. the server
keeps around quite a bit of the live stream so you can seek back into
it).
Without that commit, two different kind of issues would happen:
* When doing reverse playback, you would never check whether you
are outside of the seekable region. And would then continuously
try to download fragments that are no longer present.
* When doing fast forward, you would end up requesting fragments
which are not present yet.
In order to determine whether one was *really* outside of the seekable
window, we check whether the current stream position is still
within the seekable region.
The *problem* though with that commit is that it assumes that subclasses
will return continuously updated seeking ranges (i.e. dependent on the
current time), which is *NOT* the case.
For example:
* dashdemux does use the current UTC to determine the seekable region
* hlsdemux uses the values from the last updated manifest
Therefore if one downloads fragments faster than realtime, for HLS
we would end up at the end of the last manifest seekable range, and
the previous commit would consider the stream as being ended... which
is not the case.
In the long run, we need to figure out a way to cope with non-1.0
rates on live streams for all types of stream (including HLS).
https://bugzilla.gnome.org/show_bug.cgi?id=783075
This is a race that was exposed by the {hls|dash}.scrub_forward_seeking
validate test.
The "race" is that a subclass might want to change format, causing
a new stream to be created (but not exposed/switched yet) and put on the
prepared_streams list. That stream will have values (including pending
segment) from the pre-seek state.
Before the stream is exposed/switched, a new seek comes in and the stream
values get updated ... but the ones that will be changed don't get updated
causing them to push out wrong segments once they are exposed.
https://bugzilla.gnome.org/show_bug.cgi?id=773159
Adaptive demuxers are special demuxers that runs their own sources
internally. In this patch we flag the demuxer as being a source in order
to receive the downstream events. We then handle the EOS event by
resetting the internal state and pushing EOS on all pads. This handling
is done asynchronously to avoid blocking user thread.
https://bugzilla.gnome.org/show_bug.cgi?id=723868
Previous commit let demux call gst_uri_downloader_cancel() on _demux_reset().
Note that, _demux_reset() called during PAUSED_TO_READY and READY_TO_PAUSED.
And, it will set "cancelled" on uridownloader which blocks the use of
uridownloader. The issue is that, subclass can use the uridownloader not only
live streaming for manifest update, but also for fetching another manifests
such as variant and rendition m3u8 of hls streaming. So to unblock it,
demux should clear "cancelled" before processing initial manifest.
https://bugzilla.gnome.org/show_bug.cgi?id=783401
before broadcasting preroll.
The deadlock was as follows:
-> The subclass pushes a buffer on a newly-created stream in T1
-> We take the preroll lock in T1, to handle_preroll
-> The demuxer is stopped in T2, we take the MANIFEST_LOCK
-> T1 starts blocking because it received a reconfigure event
and needs to take the MANIFEST_LOCK
-> T2 deadlocks because it now wants the preroll_lock.
https://bugzilla.gnome.org/show_bug.cgi?id=783255
Make sure the manifest update loop is stopped before proceeding with the
resetting of the manifest data. Otherwise, the updates loop will try to
use it and it leads to a segfault
https://bugzilla.gnome.org/show_bug.cgi?id=783028
As we release the MANIFEST_LOCK in stop_tasks,
demux->priv->old_streams can be set, we need to free these
otherwise we may end up trying to dispose elements in the
READY state.
https://bugzilla.gnome.org/show_bug.cgi?id=783256
When an accurate seek is requested on a live stream, only requests the
exact value for the "starting position" (i.e. start in forward playback
and stop in reverse playback).
https://bugzilla.gnome.org/show_bug.cgi?id=782698
The live seeking range was only checked when doing actual seeks. This was
assuming that the rate would always be 1.0 (i.e. the playback would
advance in realtime, and therefore fragments would always be available
since the seeking window moves at the same rate).
With non-1.0 rates, this no longer becomes valid, and therefore we need
to check whether we are still within the live seeking range when advancing.
https://bugzilla.gnome.org/show_bug.cgi?id=783075
What we want is to retry downloading the fragment on 4xx/5xx errors
however returning EOS will cause waiting for a manifest update for live
(which may be a really long time) or stop everything for non-live.
Change that to only return EOS/ERROR once we've reached the error limit.
https://bugzilla.gnome.org/show_bug.cgi?id=776609
This commit fixes the following assumptions with live seeking:
1) start was always valid and of type GST_SEEK_TYPE_SET
2) direction was always forward
3) stop should be offsetted when handling non-accurate seeks before
the range start position.
In order to handle more live seeking use-cases (including reverse playback),
only do non-accurate start/stop value clamping for GST_SEEK_TYPE_SET values.
Also add a bit more debugging lines for issues
https://bugzilla.gnome.org/show_bug.cgi?id=782330
When dealing with live streams, we can't rely on GstSegment calculation
since it uses the segment duration to calculate the absolute values.
But since we are dealing with live *and* we know the ranges, we can
compute the absolute seeking values using the range stop (i.e. "now")
as the END position.
Allows seeking back to "live" by using start_type:GST_SEEK_TYPE_END
and start:0
https://bugzilla.gnome.org/show_bug.cgi?id=782228
The allowed live seek ranges returned by subclasses are "inclusive", that is
to say that the "range_stop" value they return is the highest acceptable position
one can seek to (i.e. "now").
Allow seeking to exactly that value
Rationale is to allow the manifest update task to continue running while
seeks are occurring. Otherwise, if the user reliably performs a seek
before the manifest is updated, then as the manifest task is reset on
seeks (and thus the time to wait between manifest updates), the manifest
would never be updated.
This fix makes the manifest update task free-running and continously
update even during seeks.
Some actions (Qos, reconfigure, ...) might take place before we finish pushing out flush_start.
One problem would be that:
1) The QOS handling in adaptivedemux takes the MANIFEST LOCK
That QOS event comes from basesink with its PREROLL_LOCK taken
2) FLUSH_START is sent from adaptivedemux with the MANIFEST_LOCK taken and the basesink flushing handler needs to take the PREROLL_LOCK
=> deadlock
https://bugzilla.gnome.org/show_bug.cgi?id=781320
At the moment that demux is waiting manifest update, the target sequence
of fragment was advanced already. So, checking stream_has_next_fragment()
means looking for the next fragment of target fragment.
This might cause unexpected buffering if each fragment has
large duration and manifest is listing only limited number of fragments.
https://bugzilla.gnome.org/show_bug.cgi?id=780494
When there are new pads pending for a bitrate switch, don't allow
EOS through from the old streams. It will be sent when the new pads are
ready, just before the old streams are removed.
This fixes racy bitrate switching with hlsdemux in urisourcebin
where old pads EOS before new pads appear and the entire pipeline can
EOS if those propagate fast enough
