Without this, new developers, unfamiliar with pipewire/wireplumber
architecture, can easily be confused about why their debug messages
are not showing up.
Add a new overflow-safe function to check if region p2 of size s2 fits
completely in p1 of size s1 and, if it does, return the amount of bytes
in p1 that come after the end of p2. Use this to bounds check the pod
iterators while ensuring that the pointer is bounds checked before being
dereferenced.
The spa_pod*_next() functions can still create an out-of-bounds pointer,
but this will not be dereferenced. Fixing this requires either
additional complexity in these functions or forbidding POD structs,
objects, and sequences that have a length that is not a multiple of 8
bytes.
Fixes: 92ac9a355f ("spa: add spa_ptrinside")
Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
Primark True Wireless earbud doesn't support sbc-xq. Having it
enabled causes bluez to enter into a loop enabling/disabling
the device dozens of times per minute, making it unusable.
When the session manager sends an error to the client, it typically
also destroys the node after the error, which causes the stream to go
to STATE_UNCONNECTED via proxy_removed(). In that case, make sure
we exit the loop early, otherwise it will take 30 seconds to unblock
gst_element_set_state()
This is a revised version of the fix that was commited via !1763
and then reverted, as it was problematic. Now the code ensures
that it breaks out only if the state was previously CONNECTING
or higher.
GStreamer uses a time stamp for the decoding time, but PipeWire uses an
offset to the presentation time. Thus, the pipewiresink must not use the
DTS as dts_offset, but has to calculate the offset.
If the buffer's DTS is invalid, assume that dts is pts.
Add a new overflow safe function to check if region p2 of size s2 fits
completely in p1 of size s1. Use this to bounds check the pod iterators.
Fixes#3727
That is indeed 0 for nearly any device. However the NTP value in the session identification part plays a crucial role for distinguishing between streams in some implementations, e.g. Dante.
Dante Controller does not recognize next stream having the same NTP value. Work around that by adding current number of sessions to the time and the magic value.
Co-authored-by: Dewi Seignard <dewiweb@gmail.com>
This should be done to match packet size requirements (e.g. 1 ms) while allowing user's software to run at higher buffer size to not stutter.
This will require scheduling multiple rtp_audio_flush_packets calls per one rtp_audio_process_capture call
It's better for PHC identification in multi-NIC systems. PHC numbering might be dependent on the probe order yet no bugs was observed. Still recommend this for more comfortable configuration.
Also added some guidance on what must be changed
The monitor sources also list the port of the sink and so the active
port needs to be collected as well so it doesn't fall back to the first
port (which might not be available).
Keep a list of ports for the node. When the node goes away, clear the
port links to the node. Handle the case where the port no longer has a
node.
This avoids a crash when, for example, the node permission is removed
and the port points to the now freed node_data.
Fixes#3708
Check for errors when loading the geometry instead of silently failing.
The points need to be given in the user locale and so might fail to
parse when given in JSON format.
Format the geometry nicely when loading the module.
Some AirPlay devices will announce their IPv4 addresses
over IPv6 mDNS if both are available, so the determined
IP version was not reliable.
The prop is not used by module-raop-sink, so its
removal should be safe.
As part of the setup for IRQ based scheduling, a period event
was installed. Not only is a timer based polling unecessary for
IRQ scheduling, depending on the state of the system, the timer
could fire far enough from the IRQ, causing alsa wakeup events
with no data in the ring buffer. Pipewire would identify these
events as an "early wakeup", adding an extra quantum of time
to the next_time estimate, skewing the clock and causing issues
with apps that depend on precise timing.
Update the started and ready state after we suspend/pause the node so
that we don't complain if scheduling happens between setting the fields
and actually stopping the follower.
Also only complain when the scheduling happens when the node is not
ready. It is possible that the node is scheduled before we manage to set
the started field.
Move the driver and warned bits after the int field in the struct so
that they are placed in separate memory.
Otherwise, a write from the data thread might race with a write from the
main thread and leave the bits in the wrong state.
This reverts commit 49cdb468c2.
We should not do this, the nsec field should be relatable to the clock
monotonic time. If we use the estimated time, without actually using it
as a timer, we might end up with a wakeup time in the future compared to
the MONOTONIC clock time.
Instead, you can use the estimated current time simply by subtracting
the rate corrected duration from the next_nsec. This is really only
useful for some selected use cases (like in the JACK library).
This fixes some issues where in pro-audio mode, a client would try to
compare the current MONOTONIC time to nsec and find that it is in the
past.
This commit was done in an attempt to fix#3657 but it turned out the
real problem was something else.
We should not use nsec from the clock as the current_usec because
current_usec is supposed to be an idealized time when the wakeup would
have happened and nsec is when it actually happened.
Instead use next_nsec and subtract the rate corrected period from it to
get the idealized current_usec.
We can still use nsec to calculate the elapsed time since the wakeup and
be sure that it is always in the past.
