At present, axfer is designed to use several types of backend for
transmission of data frames. This commit is an implementation
example of the backend.
Libffado is a userspace library for transmission of data frames according
to protocols similar to IEC 61883-1/6. This library handles audio and
music units on IEEE 1394 bus.
Unfortunately, this library executes ctor/dtor of instances for some
objects in startup/finish routines of C runtime. As a result, it outputs
some superfluous messages even if the backend is not actually used.
Furthermore, this library brings memory leak internally. Therefore,
it's not practical to build this backend for generic purposes. Although
the backend implementation works fine, this commit is just for technical
preview.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In 2010, ALSA PCM interface got an flag of hardware parameters to suppress
periodical interrupts, according to a request from PulseAudio developer.
In typical PCM operation for usual hardware, PCM drivers configure the
hardware to generate the periodical interrupts to notify that the same
amount of data frames as a period of PCM buffer is actually transferred
via serial sound interface. The flag can suppress this if the driver
support it.
There's some merits of this configuration:
- No interrupt context run for PCM substream. The PCM substream is
handled in any process context only. No need to care of race
conditions between interrupt/process contexts. This is good for
developers of drivers and applications.
- CPU time is not used for handlers on the interrupt context. The CPU
time can be dedicated for the other tasks. This is good in a point
of Time Sharing System.
- Hardware is not configured to generate interrupts. This is good in a
point of reduction of overall power consumption.
Disabling period interrupt is used for 'Timer-based scheduling' to
consume data frames on PCM buffer independently of interrupt context. As
noted, no interrupt context runs for PCM substream, thus any blocking
operation is not released. Furthermore, system calls for multiplexed I/O
is not also released without timeout.
In this scheduling model, applications need to care of available space on
PCM buffer by lapse of time, typically by yielding CPU and wait for
rescheduling. For the yielding, timeout is calculated for preferable
amount of PCM frames to process. This is an additional merit for
applications, like sound servers. when an I/O thread of the server wait
for the timeout, the other threads can process data frames for server
clients. Furthermore, with usage of rewinding/forwarding, applications
can achieve low latency between transmission position and handling
position even if they uses large size of PCM buffers.
But the timeout should be calculated with enough care of hardware
capabilities. To disable period interrupt, used hardware should satisfy
some requirements for data transmission:
1. Even if drivers don't handle interrupts to queue next data transmission,
hardware voluntarily perform the data transmission when needed
(typically by requesting DMA automatically).
2. hardware has a capability to report current position of data
transmission with enough accuracy against the data transmission.
developers refer this as 'granularity'. If hardware can always
reports updated position after the data transmission finishes, the
granularity equals to the size of period of PCM buffer.
3. a fine size of data transmission in one time. This size is decided
depending on configuration of hardware or DMA controller, but for
efficiency it may not be one byte. Thus some amount of data frame is
transferred by one data transmission. Developers refer this as
'burst-ness'.
The timeout should be calculated according to the item 2 and 3, however
in current ALSA PCM interface supplemental information is not delivered
from drivers to applications. Although at present userspace applications
should be written by a speculative way for this point, there's few
problems because there're a few hardware which satisfy the above items.
However, when more drivers supports this feature, the problem may largely
be exposed and bothers application developers.
This commit adds an option to use 'timer-based scheduling' for data
transmission. This commit adds '--sched-model' option, and the scheduling
mode is enabled when 'timer' is assigned to the option by equal sign.
Although there's some TODOs, you can see the scheduling mode in this
simple program, like:
$ axfer transfer --sched-model=timer -P -d 2 -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer --sched-model=timer -C -d 2 -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support of waiter for Linux specific epoll(7) system call.
For portability to the other Unix-like systems such as xBSD, modification
of Makefile.am may be required for conditional build, but this commit
includes no changes for it.
Below lines are examples to use this option:
$ axfer transfer --waiter-type=epoll -M -P -d 2 -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer --waiter-type=epoll -M -C -d 2 -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support of waiter for select(2) system call.
Below lines are examples to use this option:
$ axfer transfer --waiter-type=select -M -P -d 2 -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer --waiter-type=select -M -C -d 2 -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support of waiter for poll(2) system call.
Below lines are examples to use this option:
$ axfer transfer --waiter-type=poll -M -P -d 2 -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer --waiter-type=poll -M -C -d 2 -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
There're several types of system calls for multiplexed I/O. They're used to
receive notifications of I/O events. Typically, userspace applications call
them against file descriptor to yield CPU. When I/O is enabled on any of
the descriptors, a task of the application is rescheduled, then the
application execute I/O calls.
This commit adds a common interface for this type of system calls, named as
'waiter'. This is expected to be used with non-blocking file operation and
operations on mapped page frame.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In alsa-lib PCM API, data frames can be handled in mapped page frame,
instead of calling any system calls.
This commit support for this type of operation. To reduce CPU usage,
this commit uses 'snd_pcm_wait()' to wait for event notification.
Below lines are examples to execute:
$ axfer transfer -M -P -d 2 -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer -M -C -d 2 -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In current aplay, default action is to transfer data frames from/to
devices. This commit adds support for this functionality.
Event loop is included in an added file. In the loop, the number of
handled data frames is manipulated by an appropriate way. As a result, users
can stop data transmission frames by frame.
Unlike aplay, when catching SIGSTP, this application performs to suspend
PCM substream. When catching SIGCONT, it performs to resume the PCM
substream. The aim of this design is to avoid XRUN state of the PCM
substream. If users/developers need to any XRUN-recovery test, it's
better to work for the other ways.
Below lines are examples to execute:
$ axfer transfer -P -D hw:0,3 /dev/urandom -f dat -vvv
$ axfer transfer -C -D hw:1,0 /dev/null -r 48000 -vvv
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In alsa-lib PCM API, snd_pcm_read[i|n]() and snd_pcm_write[i|n]() are used
to transfer data frames from/to hardware. When a handler is not opened with
specific flags, these functions perform blocking operation; i.e. the
function call doesn't return till all of request number of data frames are
actually handled, or call is interrupted by Unix signals, or PCM substeam
corrupts due to hardware reasons.
This commit adds support for this type of data transmission. For cases that
requested data frames are not processed by container interface, this commit
adds internal cache mechanism to handle rest of data frames in next timing.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support fo alsa-lib PCM API as a backend of 'xfer'
module. In a set of alsa-lib PCM API, there're two ways to handle data
frames; by calling ioctl(2) with some specific commands with buffer in
user space, or copying data frames on mapped page frames. To support
both ways, this commit adds an operation structure as abstraction.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In aplay, many command-line options are supported. Some of them have
dependency or conflicts. Furthemore, some of them are just for
runtime configuration of alsa-lib(libasound), and some options can
be used by several xfer backends commonly; e.g. options for file name,
sample format and sampling rate.
This commit adds a parser for the common options below.
* --help (-h)
* Just output 'help' string (not written yet).
* --verbose (-v)
* For verbose output, including information about xfer, mapper and
container.
* --format (-f): string. format literals or one of ['cd'|'cdr'|'dat']
* For sample format supported by ALSA PCM interface. Special format
can be used. For playback, this is auto-detected according to actual
file format.
* --channels (-c)
* For the number of samples included in one data frame. For playback,
this is auto-detected according to actual file format, except for
'raw' format. This option can conflict to above format option.
* --rate (-r)
* For the number of data frames transferred in one second. For playback,
this is auto-detected according to actual file format, except for
'raw' format. This option can conflict to format option above.
* --file-type (-f): string. one of ['wav'|'au'|'voc'|'raw']
* For format of files of given paths. For playback, this is optional
because the format is auto-detected. For capture, this is optional too
because the format is decided according to suffix of given path.
Anyway, this option is used for cases to fail to detect or decide.
* --separate-channels (-I)
* When using several files as source or destination for transmission
of data frame, this option can be used with several file paths.
When '--separate-channels' option is used, users can give several file
paths to source/destination of data transmission, else they can give single
file path for the purpose. When multiple files are handled by this option,
for playback, data frames in first channel is used to construct buffer for
data transmission with multi channel. For capture, data frames in each
channel of buffer are written to each of given path. Furthermore, when a
single path is given for capture, file paths are auto-generated according
to available number of channels. For example, 'name.wav' is given for
2 channels capture, 'name-0.wav' and 'name-1.wav' are generated. In a
case of no suffix, 'name-0' and 'name-1' are generated.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
ALSA has PCM interface to transfer data frames. In userspace, there're
some implementation to utilize this interface to produce application
programming interface; alsa-lib (libasound) and tinyalsa. However, it's
possible to use the interface with raw I/O operations.
This commit adds an common interface to transfer data frames for this
program, named as 'xfer'. This internal interface is designed for users
to select several backend for data transmission. This includes some
functions expected to be called by main program just for data
transmission. In an aspect to maintain PCM substream, suspend feature is
required to handle a pair of SIGTSTP/SIGCONT UNIX signals.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support of mapper for 'multiple' target. This handles
several files via 'container' functions, and constructs data frame buffer
for playback, or splits data frames from data frame buffer for capture.
When playback source files includes data frames with several channels, the
first channel is used to construct buffer. For capture direction, each of
channel of data frame is stored in one file, thus the file includes one
channel of data frame. When handling non-interleaved buffer, a caller
should use an array of buffer for each of channels with non-interleaved
data frames.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In usual use case of aplay, single file is used to playback or capture data
frames.
This commit adds support of single type mapper for this use case. All of
supported file format can include data frame with interleaved alignment,
thus this mapper have a functionality to convert from several types of
data frame alignment to interleaved alignment or vise versa. When
handling non-interleaved buffer, a caller should use an array of buffer
for each of channels with non-interleaved data frames.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In current aplay, several files can be handled as source of data frames for
playback, or destination of captured data frames by an option
'--separate-channels' (-I).
On the other hand, in ALSA PCM kernel/user interface, several types of
buffer are used to communicate between application/hardware;
- mapped page frame for data frames with interleaved alignment
- mapped page frame for data frames with non-interleaved alignment
- buffer in user space for data frames with interleaved alignment
- a list of buffer in user space for data frames with non-interleaved
alignment
This commit adds a common interface, named as 'mapper' to convert frame
alignment between these two sides. This interface includes two types;
'muxer' and 'demuxer'. The 'muxer' is for playback direction, to
construct playback buffer with PCM frames from several files. The 'demuxer'
is for capture direction, to split PCM frames from capture buffer to
each of file. Unlike multimedia containers such as MPEG 2/4 Systems,
the 'muxer' and 'demuxer' are for playback/capture buffer, not for file
contents.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In former commits, container module gets supports of parser/builder for
several types of file format. This commit adds a unit test for them.
This includes positive test cases only. The test cases actually generate
I/O to file systems for many test cases. It takes a long time to finish.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support for raw data without any headers/chunks/blocks.
A parser of container cannot recognize format of sample without
supplemental information.
Additionally, it includes no magic bytes. A parser of container should
process first several bytes as a part of PCM frames, instead of magic
bytes.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support for data of Creative Tech. voice format. In this
data format, values in each of field are represented in little-endian byte
order and available formats of data sample are restricted in little-endian
byte order.
In version 1.10 of this format, sampling rate is represented with
reciprocal number of the rate, thus we cannot calculate original sampling
rate precisely just from its header. For example at 44.1kHz, file header
includes 233 (=256-1,000,000/44,100), but we cannot recover the value just
from the code (43478.2...). For my convenience, this commit adds a
pre-computed table and lookup major rates from the table.
Additionally, this format can includes several blocks with different
sample format. When handling this type of file, we need to start/stop
substream for each of the block, while this brings complicated code.
This type of format is enough ancient and presently quite minor. This
commit takes a compromise and handles a first sample block only.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support for data of Sparc AU format. In this data format,
values in each of field are encoded in big-endian byte order and available
formats of data sample are restricted in big-endian byte order.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds support for data of Microsoft/IBM RIFF/Wave format. In
this data format, values in each of field are encoded in both bit/little
byte order but inner a file the same order is used. Magic bytes in the
beginning of data indicated which byte order is used for the file.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Current aplay supports several types of data format for file; Microsoft/IBM
RIFF/Wave (.wav), Sparc AU (.au) and Creative Tech. voice (.voc). These
formats were designed to handle audio-related data with interleaved frame
alignment.
This commit adds a common interface to handle the file format, named as
'container' module. This includes several functions to build/parse
the format data from any file descriptors. Furthermore, this includes
several helper functions for implementations of each builder/parser.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Original aplay implementation has a feature to output two types of list;
devices and PCMs. The list of devices is a result to query sound card and
pcm component structured maintained in kernel land. The list of PCMs is a
result to parse runtime configuration files in alsa-lib. Entries in the
former list is corresponding to ALSA PCM character device
('/dev/snd/pcm%uC%uD[p|c]'), while entries in the latter list includes
some 'virtual' instances in application runtime.
This commit adds an implementation for the above functionality. This is
executed by taking 'list' sub-command. A 'device' option has the same
effect as '--list-devices' and '-L' of aplay. A 'pcm' option has the same
effect as '--list-pcms' and '-l' of aplay. In both cases, an additional
option is required for stream direction. Below is examples of new command
system for this sub-command.
$ axfer list device -C (= arecord --list-devices)
$ axfer list pcm -P (= aplay -l)
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds a new command, 'axfer' ('ALSA transfer'), to transfer data
frames described in asound.h. This command is intended to replace current
aplay. The most of features and command line parameters come from aplay as
much as possible, while it has more better feature and code to maintain.
This commit adds an entry point for this command. Current option system of
aplay is still available, while this command has a sub-command system like
commands in iproute2.
Currently, two sub-commands are supported; 'list' and 'transfer'. The
'list' sub-command has the same effect as '-l' and '-L' options of aplay.
The 'transfer' sub-command has the same effect as the main feature of
aplay. For the sub-command system, an option for stream direction is
required; '-P' for playback and '-C' for capture. If you create symbolic
links to this binary for aplay/arecord, please execute:
$ ln -s axfer aplay
$ ln -s axfer arecord
Actual code for each sub-command will be implemented in later commits.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
