[v2,1/3] docs: sound: add 'pcmtest' driver documentation

Add documentation for the new Virtual PCM Test Driver. It covers all
possible usage cases: errors and delay injections, random and
pattern-based data generation, playback and ioctl redefinition
functionalities testing.

We have a lot of different virtual media drivers, which can be used for
testing of the userspace applications and media subsystem middle layer.
However, all of them are aimed at testing the video functionality and
simulating the video devices. For audio devices we have only snd-dummy
module, which is good in simulating the correct behavior of an ALSA device.
I decided to write a tool, which would help to test the userspace ALSA
programs (and the PCM middle layer as well) under unusual circumstances
to figure out how they would behave. So I came up with this Virtual PCM
Test Driver.

This new Virtual PCM Test Driver has several features which can be useful
during the userspace ALSA applications testing/fuzzing, or testing/fuzzing
of the PCM middle layer. Not all of them can be implemented using the
existing virtual drivers (like dummy or loopback). Here is what can this
driver do:

- Simulate both capture and playback processes
- Check the playback stream for containing the looped pattern
- Generate random or pattern-based capture data
- Inject delays into the playback and capturing processes
- Inject errors during the PCM callbacks

Also, this driver can check the playback stream for containing the
predefined pattern, which is used in the corresponding selftest to check
the PCM middle layer data transferring functionality. Additionally, this
driver redefines the default RESET ioctl, and the selftest covers this PCM
API functionality as well.

Signed-off-by: Ivan Orlov <ivan.orlov0322@gmail.com>
---
V1 -> V2:

- Rename the driver from from 'valsa' to 'pcmtest'.
- Implement support for interleaved and non-interleaved access modes
- Add support for 8 substreams and 4 channels
- Extend supported formats
- Extend and rewrite in C the selftest for the driver

 Documentation/sound/cards/index.rst   |   1 +
 Documentation/sound/cards/pcmtest.rst | 119 ++++++++++++++++++++++++++
 2 files changed, 120 insertions(+)
 create mode 100644 Documentation/sound/cards/pcmtest.rst

On 5/22/23 17:01, Bagas Sanjaya wrote:
> On Sun, May 21, 2023 at 10:32:16PM +0400, Ivan Orlov wrote:
>> diff --git a/Documentation/sound/cards/pcmtest.rst b/Documentation/sound/cards/pcmtest.rst
>> new file mode 100644
>> index 000000000000..ea8070eaa44e
>> --- /dev/null
>> +++ b/Documentation/sound/cards/pcmtest.rst
>> @@ -0,0 +1,119 @@
>> +.. SPDX-License-Identifier: GPL-2.0
>> +
>> +The Virtual PCM Test Driver
>> +===========================
>> +
>> +The Virtual PCM Test Driver emulates a generic PCM device, and can be used for
>> +testing/fuzzing of the userspace ALSA applications, as well as for testing/fuzzing of
>> +the PCM middle layer. Additionally, it can be used for simulating hard to reproduce
>> +problems with PCM devices.
>> +
>> +What can this driver do?
>> +~~~~~~~~~~~~~~~~~~~~~~~~
>> +
>> +At this moment the driver can do the following things:
>> +	* Simulate both capture and playback processes
>> +	* Generate random or pattern-based capturing data
>> +	* Inject delays into the playback and capturing processes
>> +	* Inject errors during the PCM callbacks
>> +
>> +It supports up to 8 substreams and 4 channels. Also it supports both interleaved and
>> +non-interleaved access modes.
>> +
>> +Also, this driver can check the playback stream for containing the predefined pattern,
>> +which is used in the corresponding selftest (alsa/test-pcmtest-driver.c). To check the
>> +PCM middle layer data transferring functionality. Additionally, this driver redefines
> "... corresponding selftest ... to check the ..."
>> +the default RESET ioctl, and the selftest covers this PCM API functionality as well.
>> +
>> +Configuration
>> +-------------
>> +
>> +The driver has several parameters besides the common ALSA module parameters:
>> +
>> +	* fill_mode (bool) - Buffer fill mode (see below)
>> +	* inject_delay (int)
>> +	* inject_hwpars_err (bool)
>> +	* inject_prepare_err (bool)
>> +	* inject_trigger_err (bool)
>> +
>> +
>> +Capture Data Generation
>> +-----------------------
>> +
>> +The driver has two modes of data generation: the first (0 in the fill_mode parameter)
>> +means random data generation, the second (1 in the fill_mode) - pattern-based
>> +data generation. Let's look at the second mode.
>> +
>> +First of all, you may want to specify the pattern for data generation. You can do it
>> +by writing the pattern to the debugfs file (/sys/kernel/debug/pcmtest/fill_pattern).
>> +Like that:
> "For example::"
>> +
>> +.. code-block:: bash
>> +
>> +	echo -n mycoolpattern > /sys/kernel/debug/pcmtest/fill_pattern
>> +
>> +After this, every capture action performed on the 'pcmtest' device will return
>> +'mycoolpatternmycoolpatternmycoolpatternmy...' for every channel buffer.
>> +
>> +In case of interleaved access, the capture buffer will contain the repeated pattern
>> +for every channel. Otherwise, every channel buffer will contain the repeated pattern.
>> +
>> +The pattern itself can be up to 4096 bytes long.
>> +
>> +Delay injection
>> +---------------
>> +
>> +The driver has 'inject_delay' parameter, which has very self-descriptive name and
>> +can be used for time delay/speedup simulations. The parameter has integer type, and
>> +it means the delay added between module's internal timer ticks.
>> +
>> +If the 'inject_delay' value is positive, the buffer will be filled slower, if it is
>> +negative - faster. You can try it yourself by starting a recording in any
>> +audio recording application (like Audacity) and selecting the 'pcmtest' device as a
>> +source.
>> +
>> +This parameter can be also used for generating a huge amount of sound data in a very
>> +short period of time (with the negative 'inject_delay' value).
>> +
>> +Errors injection
>> +----------------
>> +
>> +This module can be used for injecting errors into the PCM communication process. This
>> +action can help you to figure out how the userspace ALSA program behaves under unusual
>> +circumstances.
>> +
>> +For example, you can make all 'hw_params' PCM callback calls return EBUSY error by
>> +writing '1' to the 'inject_hwpars_err' module parameter:
>> +
>> +.. code-block:: bash
>> +
>> +	echo 1 > /sys/module/snd_pcmtest/parameters/inject_hwpars_err
>> +
>> +Errors can be injected into the following PCM callbacks:
>> +
>> +	* hw_params (EBUSY)
>> +	* prepare (EINVAL)
>> +	* trigger (EINVAL)
>> +
>> +
>> +Playback test
>> +-------------
>> +
>> +This driver can be also used for the playback functionality testing - every time you
>> +write the playback data to the 'pcmtest' PCM device and close it, the driver checks the
>> +buffer of each channel for containing the looped pattern (which is specified in the
>> +fill_pattern debugfs file). If the playback buffer content represents the looped pattern,
>> +'pc_test' debugfs entry is set into '1'. Otherwise, the driver sets it to '0'.
>> +
>> +ioctl redefinition test
>> +-----------------------
>> +
>> +The driver redefines the 'reset' ioctl, which is default for all PCM devices. To test
>> +this functionality, we can trigger the reset ioctl and check the 'ioctl_test' debugfs
>> +entry:
>> +
>> +.. code-block:: bash
>> +
>> +	cat /sys/kernel/debug/pcmtest/ioctl_test
>> +
>> +If the ioctl is triggered successfully, this file will contain '1', and '0' otherwise.
> 
> The rest is LGTM, thanks!
> 

Hi Bagas!

Thank you for the review! I'll fix this in the next version of the patch 
set after getting response from the sound maintainers - probably it 
isn't the only part which needs to be fixed :)

Kind regards,
Ivan Orlov.