This fixes a crash. sink_input_pop_cb() drains the message queue that receives
memchunks from the combine sink thread to avoid requesting more audio too soon.
The same message queue received also SET_REQUESTED_LATENCY messages, which
generate rewind requests. Rewind requests shouldn't be issued in the pop()
callback, doing so results in an assertion error. Therefore, it was not safe to
drain the message queue in the pop() callback, but usually the queue is empty,
so this bug was not immediately detected.
This patch splits the message queue into two queues: audio_inq and control_inq.
audio_inq receives only messages containing memchunks, and control_inq receives
only the SET_REQUESTED_LATENCY messages. The pop() callback only drains the
audio queue, which avoids the rewind requests in the pop() callback.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=90489
We have crashes related to modules loaded after unload. This added
warning can provide some information about what that module is,
which in turn can help us solve the crashes, hopefully.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=90108
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
When we the underlying sink/source goes away, there is an intermediate
state where the asyncmsgqs that we were using for the sink-input and
source-output go away. This is usually okay if the sink-input and
source-output are moved to another device, but can be problematic if we
don't support moving (which is the case when the filter is autoloaded).
This becomes a problem because of the following chain of events:
* The underlying sink goes away
* Moving the filter sink-input fails (because it is autloaded)
* At this point the sink-input has no underlying sink, and thus
no underlying asyncmsgq
* This also applies to all sink-inputs connected to the echo-cancel
module
* The sink-input is killed, triggering a module unload
* On unlink, module-rescue-streams tries to move sink-inputs to
another sink, starting with a START_MOVE message
* There is no asyncmsgq for the message, so we crash
* We can't just perform a NULL check for the asyncmsgq, since there
are state changes we need to effect during the move
To fix this, we pretend to allow the move to the new sink, and then
unlink ourselves *after* the move is complete. This ensures that we
never find ourselves in a position where we need the underlying
sink/asyncmsgq to be present when it is not.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=90416
In case the sample spec is not known, as can be the case when
pa_stream_new_extended is used, we cannot satisfy the PULSE_LATENCY_MSEC
request.
As a workaround disable being able to use PULSE_LATENCY_MSEC in this case.
Reported-by: Fritsch <fritsch@xbmc.org>
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
...otherwise this code will fail on big-endian architectures.
Cc: Hui Wang <hui.wang@canonical.com>
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Modern versions of MinGW and Visual Studio provide socket errno
defines that make no sense (no API sets them). Make sure we
continue to use the old WSAE ones that are actually returned by
Windows' socket API.
Signed-off-by: Pierre Ossman <ossman@cendio.se>
pulsecore/sink.c: In function 'pa_sink_put':
pulsecore/sink.c:648:53: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
pa_assert(!(s->flags & PA_SINK_DYNAMIC_LATENCY) == (s->thread_info.fixed_latency != 0));
^
pulsecore/source.c: In function 'pa_source_put':
pulsecore/source.c:599:55: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
pa_assert(!(s->flags & PA_SOURCE_DYNAMIC_LATENCY) == (s->thread_info.fixed_latency != 0));
^
rewrite expression to suppress warning:
!(x & MASK) == (y != 0)
<->
!(x & MASK) == !(y == 0)
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
warnings emited by gcc 5.1:
utils/padsp.c: In function 'dsp_trigger':
utils/padsp.c:1902:39: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
while (!pa_operation_get_state(o) != PA_OPERATION_DONE) {
^
utils/padsp.c: In function 'dsp_cork':
utils/padsp.c:1937:39: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
while (!pa_operation_get_state(o) != PA_OPERATION_DONE) {
^
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Fall back to the previous /etc/bash_completion.d dir on failures
(either old bash completion or not installed).
changes over Ville Skyttä's patch:
define PKG_CHECK_VAR macro which became available only in pkg-config 0.28
see https://bugs.freedesktop.org/show_bug.cgi?id=88782 and
https://bugs.freedesktop.org/show_bug.cgi?id=89540
Signed-off-by: Ville Skyttä <ville.skytta@iki.fi>
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
pulsecore/filter/biquad.c: In function 'biquad_lowpass':
pulsecore/filter/biquad.c:52:10: warning: declaration of 'gamma' shadows a global declaration [-Wshadow]
pulsecore/filter/biquad.c: In function 'biquad_highpass':
pulsecore/filter/biquad.c:86:10: warning: declaration of 'gamma' shadows a global declaration [-Wshadow]
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
see https://bugs.freedesktop.org/show_bug.cgi?id=86818
an exit code of 1 makes systemd believe that the service failed;
better return 0 to denote that PA sucessfully stopped on the user's
request
sidenote: systemd's SuccessExitStatus= could be used to turn code 1 into a
code denoting success
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Cc: jan.steffens@gmail.com
On 32bits OS, this test case fails. The reason is when rewinding to
the middle of a block, some of float parameters in the saved_state
are stored in the memory from FPU registers, and those parameters will
be used for next time to process data with lfe. Here if FPU register
is over 32bits, the storing from FPU register to memory will introduce
some variation, and this small variation will introduce small
variation to the rewinding result.
So adding the tolerant variation for comparing the rewind result, make
this test case can work on both 64bits OS and 32bits OS.
Signed-off-by: Hui Wang <hui.wang@canonical.com>
This fixes buffer attr calculation so that we set the source latency to
the requested latency. This makes sense because the intermediate
delay_memblockq is just a mechanism to send data to the client. It
should not actually add to the total latency over what the source
already provides.
With this, the meaning of fragsize and maxlength become more
meaningful/accurate with regards to ADJUST_LATENCY mode -- fragsize
becomes the latency the source is configured for (which is then
approximately the total latency until the buffer reaches the client).
Maxlength, as before, continues to be the maximum amount of data we
might hold for the client before overrunning.
Fix the following warnings.
CC pulsecore/filter/libpulsecore_6.0_la-lfe-filter.lo
pulsecore/filter/lfe-filter.c: In function 'pa_lfe_filter_rewind':
pulsecore/filter/lfe-filter.c:179:9: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'size_t' [-Wformat=]
pa_log_debug("Rewinding LFE filter %lu samples to position %lli. No saved state found", samples, (long long) f->index);
^
pulsecore/filter/lfe-filter.c:183:5: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'size_t' [-Wformat=]
pa_log_debug("Rewinding LFE filter %lu samples to position %lli. Found saved state at position %lli",
^
CC pulsecore/filter/libpulsecore_6.0_la-biquad.lo CC pulsecore/filter/libpulsecore_6.0_la-lfe-filter.lo
pulsecore/filter/lfe-filter.c: In function 'pa_lfe_filter_rewind':
pulsecore/filter/lfe-filter.c:179:9: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'size_t' [-Wformat=]
pa_log_debug("Rewinding LFE filter %lu samples to position %lli. No saved state found", samples, (long long) f->index);
^
pulsecore/filter/lfe-filter.c:183:5: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'size_t' [-Wformat=]
pa_log_debug("Rewinding LFE filter %lu samples to position %lli. Found saved state at position %lli",
^
Installs all the build dependencies, and runs make check and
check-daemon.
V1: Based on Arun Raghavan's travis file. Added trusty repositories to get
newer libs.
V2: Explicitly list all dependencies instead of relying on the Ubuntu
package Build-Dependencies. Send notifications to pulseaudio-discuss
V3: Install libsystemd-daemon-dev, libsystemd-id128-dev,
libsystemd-journal-dev, and libsystemd-login-dev. Send notifications to
pulseaudio-commits. Drop libjson0-dev, libjson-c-dev is the package to
depend on.
Previously pa_parse_volume() clamped the value to fit in the valid
range, but I think it's better to reject values outside the valid
range.
This also changes the percentage parsing to allow non-integer values.
