snd_hctl_wait() currently makes a lot of assumptions that fall outside
the defined API. snd_pcm_wait(), however, follows the API nicely so the
new version is based on that.
Signed-off-by: Pierre Ossman <ossman@cendio.se>
Some io plug-ins might want to adjust the reported delay value and not
strictly follow the current buffer usage (that's why we have two calls
after all).
Allow them to specify a delay() callback and use the previous behaviour
if they don't.
Signed-off-by: Pierre Ossman <ossman@cendio.se>
When this new parameter is specified, the behavior of opening a file PCM
for input is changed to the following: reads from the device, instead of
being passed down to the slave PCM, will read the raw data from the
specified file. No file writes will take place in this case. In the
absence of this parameter, previous behavior is unchanged.
The input file name and descriptor are kept in the snd_pcm_file_t struct
in a manner analogous to the already existing output file parameter.
TODO:
(1) Only interleaved reads (snd_pcm_file_readi) have been implemented
for now.
(2) File read()'s that return a number of bytes that's not a multiple of
the frame size will result in data loss (choppiness). The rbuf,
rbuf_size_bytes, and rbuf_used_bytes members of the snd_pcm_file_t
struct will be used to address this problem in the future.
(3) Mind whether the PCM was opened in blocking mode. If so, we'll have
to loop until the buffer has been filled with read()'s.
Signed-off-by: Juan Carlos Castro y Castro <jcastro@vialink.com.br>
Add four new information flags SNDRV_SEQ_PORT_TYPE_HARDWARE, _SOFTWARE,
_SYNTHESIZER, _PORT for sequencer ports. This makes it easier for apps
like Rosegarden to make policy decisions based on the port type.
Add the internal start_pending state in the rate plugin for supporting
delayed start in the case no data is committed to slave PCM.
This fixes the problem of CPU hog with artsd.
Fixed the segfault in the error path of snd_pcm_hook_add_conf()
(calling invalid install_func pointer).
Also, added the missing snd_dlclose() to release the handle properly.
The fix to enable finer sync of dmix in delay callback seems
rather harmful for some apps like xmms on SMP.
The problem of aoss was solved by max_periods=0, so better to
revert this change.
Sync pointers in delay callback in dmix. This solves some problems
of aoss with apps that don't check poll_revents.
The rate plugin is fixed to call delay of slave appropriately, too.
Introduce "max_periods" option to specify the max number of periods
per buffer to each plugin.
- When max_periods = -1, the fixed buffer size as the slave size is
used (old behavior).
- When max_periods = 0 (or 1), the number of periods is variable
between 2 and the slave buffer size.
- When max_periods greater than 2 is given, it specifies the max
periods of that pcm explicitly.
When no option is given in the PCM defintion, the value
"defaults.pcm.dmix_max_periods" is referred as default.
The default value is 0, as defined in alsa.conf.
You can override this in ~/.asoundrc or /etc/asound.conf as you like.
Created a new rate converter plugin SDK.
A rate converter can be replaced as an extra plugin now.
The default rate converter is a built-in linear converter.
You can find a sample external converter in alsa-plugins package.
Fixed the noisy output of dmix with two (or less) periods.
The dmix tends to give noise or XRUN when running with two periods
because of its implementation nature. To avoid this, the start
position is aligned to the period size, so that the updates are synced
with interrupts of slave PCM.
Fix the PCM auto-start conditions
- The capture stream should be started the PCM anyway before reads
- Fixed the endless poll() when the written size doesn't fit to the
buffer size due to lack of auto-start
Here's a patch which adds support for S24_3LE and byte-swapped S16 and
S32 to softvol. I've tested S24_3LE and byte-swapped S16 on powerpc
with snd-usb-audio. All other cases are untested so far.
(Config at http://blog.blackdown.de/static/alsa/USB-Audio.conf)
Signed-off-by: Juergen Kreileder <jk@blackdown.de>
snd_timer_tread struct is a bad design for 32/64bit compatibility,
and reading this struct on 32bit program returns zero. This results
in tight poll looping (bug#1938, #1945).
For avoiding this bug, now more bigger buffer is read to cover the
64bit tread struct, too. Also this optimizes the read without
checking -EAGAIN in the case both user-space and kernel have the same
tread size.
