As of v4.12, snd_pcm_hw_params_choose() is just called in a process
context of ioctl(2) with SNDRV_PCM_IOCTL_HW_PARAMS. The function locates
in a different file, which has no tracepoints.

This commit moves the function to a file with the tracepoints for later
commit.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Now all materials are ready, let's allow the direct in-kernel
read/write, i.e. a kernel-space buffer is passed for read or write,
instead of the normal user-space buffer.  This feature is used by OSS
layer and UAC1 driver, for example.

The __snd_pcm_lib_xfer() takes in_kernel argument that indicates the
in-kernel buffer copy.  When this flag is set, another transfer code
is used.  It's either via copy_kernel PCM ops or the normal memcpy(),
depending on the driver setup.

As external API, snd_pcm_kernel_read(), *_write() and other variants
are provided.

That's all.  This support is really simple because of the code
refactoring until now.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Use the existing silence helper codes for simplification.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Both __snd_pcm_lib_read() and __snd_pcm_write() functions have almost
the same code to loop over samples.  For simplification, this patch
unifies both as the single helper, __snd_pcm_lib_xfer().

Other than that, there should be no functional change by this patch.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

This patch proceeds more abstraction of PCM read/write loop codes.

For both interleaved and non-interleaved transfers, the same copy or
silence transfer code (which is defined as pcm_transfer_f) is used
now.  This became possible since we switched to byte size to copy_*
and fill_silence ops argument instead of frames.

And, for both read and write, we can use the same copy function (which
is defined as pcm_copy_f), just depending on whether interleaved or
non-interleaved mode.

The transfer function is determined at the beginning of the loop,
depending on whether the driver gives the specific copy ops or it's
the standard read/write.

Another bonus by this change is that we now guarantee the silencing
behavior when NULL buffer is passed to write helpers.  It'll simplify
some codes later.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Make snd_pcm_lib_read() and *_write() static inline functions that
call the common helper functions directly.  This reduces a slight
amount of codes, and at the same time, it's a preparation for the
further cleanups / fixes.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Just shuffle the codes, without any change otherwise.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Now that all users of old copy and silence ops have been converted to
the new PCM ops, the old stuff can be retired and go away.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

For supporting the explicit in-kernel copy of PCM buffer data, and
also for further code refactoring, three new PCM ops, copy_user,
copy_kernel and fill_silence, are introduced.  The old copy and
silence ops will be deprecated and removed later once when all callers
are converted.

The copy_kernel ops is the new one, and it's supposed to transfer the
PCM data from the given kernel buffer to the hardware ring-buffer (or
vice-versa depending on the stream direction), while the copy_user ops
is equivalent with the former copy ops, to transfer the data from the
user-space buffer.

The major difference of the new copy_* and fill_silence ops from the
previous ops is that the new ops take bytes instead of frames for size
and position arguments.  It has two merits: first, it allows the
callback implementation often simpler (just call directly memcpy() &
co), and second, it may unify the implementations of both interleaved
and non-interleaved cases, as we'll see in the later patch.

As of this stage, copy_kernel ops isn't referred yet, but only
copy_user is used.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Acked-by: NMark Brown <broonie@kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Several files are used to construct PCM core module, a.k.a snd-pcm.
Although available APIs are described in 'include/sound/pcm.h', some of
them are not exported as symbols in kernel space. Such APIs are just for
module local usage.

This commit adds module local header file and move some function prototypes
into it so that scopes of them are controlled properly and developers
get no confusion from unavailable symbols.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

In a function snd_pcm_hw_params_choose(), target parameters are arranged
into a table. Though each entry of this table is read-only, they don't
have const qualifier.

This commit adds the qualifier.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

ALSA pcm core has hw_param_interval_c() to pick up parameter with const
qualifier for safe programming.

This commit applies it to the cases.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h>

Fix up affected files that include this signal functionality via sched.h.
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>

When snd_pcm_add_chmap_ctls() is called to the PCM stream to which a
chmap has been already assigned, it returns as an error due to the
conflicting snd_ctl_add() result.  However, this also clears the
already assigned chmap_kctl field via pcm_chmap_ctl_private_free(),
and becomes inconsistent in the later operation.

This patch adds the check of the conflicting chmap kctl before
actually trying to allocate / assign.  The check failure is treated as
a kernel warning, as the double call of snd_pcm_add_chmap_ctls() is
basically a driver bug and having the stack trace would help
developers to figure out the bad code path.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Currently kill_fasync() is called outside the stream lock in
snd_pcm_period_elapsed().  This is potentially racy, since the stream
may get released even during the irq handler is running.  Although
snd_pcm_release_substream() calls snd_pcm_drop(), this doesn't
guarantee that the irq handler finishes, thus the kill_fasync() call
outside the stream spin lock may be invoked after the substream is
detached, as recently reported by KASAN.

As a quick workaround, move kill_fasync() call inside the stream
lock.  The fasync is rarely used interface, so this shouldn't have a
big impact from the performance POV.

Ideally, we should implement some sync mechanism for the proper finish
of stream and irq handler.  But this oneliner should suffice for most
cases, so far.
Reported-by: NBaozeng Ding <sploving1@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The commit [d507941b: ALSA: pcm: Correct PCM BUG error message]
made the warning prefix back to "BUG:" due to its previous wrong
prefix.  But a kernel message containing "BUG:" seems taken as an Oops
message wrongly by some brain-dead daemons, and it annoys users in the
end.  Instead of teaching daemons, change the string again to a more
reasonable one.

Fixes: 507941beb1e ('ALSA: pcm: Correct PCM BUG error message')
Cc: <stable@vger.kernel.org> # v3.19+
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The ALSA core does not modify the constraints provided by a driver. Most
constraint helper functions already take a const pointer to the constraint
description, the exception at the moment being the ratden and ratnum
constraints. Make those const as well, this allows a driver to declare them
as const.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

While there is nothing wrong with the transfer_ack_begin and
transfer_ack_end callbacks per-se, the last documented user was part of the
alsa-driver 0.5.12a package, which was released 14 years ago and even
predates the upstream integration of the ALSA core and has subsequently
been superseded by newer alsa-driver releases.

This seems to indicate that there is no need for having these callbacks and
they are just cruft that can be removed.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

PCM timer is not always used. For embedded device, we need an interface
to disable it when it is not needed, to shrink the kernel size and
memory footprint, here add CONFIG_SND_PCM_TIMER for it.

When both CONFIG_SND_PCM_TIMER and CONFIG_SND_TIMER is unselected,
about 25KB saving bonus we can get.

Please be noted that when disabled, those stubs who using pcm timer
(e.g. dmix, dsnoop & co) may work incorrectlly.
Suggested-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NJie Yang <yang.jie@intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Currently in snd_pcm_update_hw_ptr0 during interrupt,
we consider there were double acknowledged interrupts when:
1. HW reported pointer is smaller than expected, and
2. Time from last update time (hdelta) is over half a buffer time.

However, when HW reported pointer is only a few bytes smaller than
expected, and when hdelta is just a little larger than half a buffer time
(e.g. ping-pong buffer), it wrongly treats this IRQ as double acknowledged.

The condition #2 uses jiffies, but jiffies is not high resolution
since it is integer. We should consider jiffies inaccuracy.
Signed-off-by: NKoro Chen <koro.chen@mediatek.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Introduce more generic .get_time_info to retrieve
system timestamp and audio timestamp in single routine.
Backwards compatibility is preserved with same functionality
as with .wall_clock method (to be removed in following commits
to avoid breaking git bisect)
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add helper functions to allow drivers to specify several disjoint
ranges for a variable. In particular, there is a codec (PCM512x) that
has a hole in its supported range of rates, due to PLL and divider
restrictions.

This is like snd_pcm_hw_constraint_list(), but for ranges instead of
points.
Signed-off-by: NPeter Rosin <peda@axentia.se>
Reviewed-by: NLars-Peter Clausen <lars@metafoo.de>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently the msbits constraints requires to specify a specific sample
format width for which the constraint should be applied. But often the
number of most significant bits is not sample format specific, but rather a
absolute limit. E.g. the PCM interface might accept 32-bit and 24-bit
samples, but the DAC has a 16-bit resolution and throws away the LSBs. In
this case for both 32-bit and 24-bit format msbits should be set to 16. This
patch extends snd_pcm_hw_constraint_msbits() so that a wildcard constraint
can be setup that is applied for all formats with a sample width larger than
the specified msbits. Choosing the wildcard constraint is done by setting
the sample width parameter of the function to 0.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

If the sound card is made up of discrete components, each with their own
driver (e.g. like in the ASoC case), we might end up with multiple msbits
constraint rules installed. Currently this will result in msbits being set
to whatever the last rule set it to.

This patch updates the behavior of the rule to choose the minimum (other
than zero) of all the installed rules.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

ALSA PCM core has a mechanism tracking the PCM hwptr updates for
analyzing XRUNs.  But its log is limited (up to 10) and its log output
is a kernel message, which is hard to handle.

In this patch, the hwptr logging is moved to the tracing
infrastructure instead of its own.  Not only the hwptr updates but
also XRUN and hwptr errors are recorded on the trace log, so that user
can see such events at the exact timing.

The new "snd_pcm" entry will appear in the tracing events:
  # ls -F /sys/kernel/debug/tracing/events/snd_pcm
  enable  filter  hw_ptr_error/  hwptr/  xrun/

The hwptr is for the regular hwptr update events.  An event trace
looks like:

  aplay-26187 [004] d..3  4012.834761: hwptr: pcmC0D0p/sub0: POS: pos=488, old=0, base=0, period=1024, buf=16384

"POS" shows the hwptr update by the explicit position update call and
"IRQ" means the hwptr update by the interrupt,
i.e. snd_pcm_period_elapsed() call.  The "pos" is the passed
ring-buffer offset by the caller, "old" is the previous hwptr, "base"
is the hwptr base position, "period" and "buf" are period- and
buffer-size of the target PCM substream.
(Note that the hwptr position displayed here isn't the ring-buffer
 offset.  It increments up to the PCM position boundary.)

The XRUN event appears similarly, but without "pos" field.
The hwptr error events appear with the PCM identifier and its reason
string, such as "Lost interrupt?".

The XRUN and hwptr error reports on kernel message are still left, can
be turned on/off via xrun_debug proc like before.  But the bit 3, 4, 5
and 6 bits of xrun_debug proc are dropped by this patch.  Also, along
with the change, the message strings have been reformatted to be a bit
more consistent.

Last but not least, the hwptr reporting is enabled only when
CONFIG_SND_PCM_XRUN_DEBUG is set.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

While converting to dev_*(), the message showing the invalid PCM
position was wrongly tagged as if an XRUN although it's actually a
BUG.  This patch corrects the message again.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The calculated frame size was wrong because snd_pcm_format_physical_width()
actually returns the number of bits, not bytes.

Use snd_pcm_format_size() instead, which not only returns bytes, but also
simplifies the calculation.

Fixes: 8bea869c ("ALSA: PCM midlevel: improve fifo_size handling")
Signed-off-by: NClemens Ladisch <clemens@ladisch.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Changing an interval boundary to a multiple of the step size makes that
boundary exact.
Signed-off-by: NClemens Ladisch <clemens@ladisch.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The min parameter was not used by any caller.  And if it were used,
underflows in the calculations could lead to incorrect results.
Signed-off-by: NClemens Ladisch <clemens@ladisch.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

BugLink: http://bugs.launchpad.net/bugs/1305480

The kerneloops-daemon scans dmesg for common crash signatures, among
which is 'BUG:'. The message emitted by the PCM library is really a
warning, so the most expedient thing to do seems to be to change the
string.
Signed-off-by: NTim Gardner <tim.gardner@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Allow modules to use it, fixing a build failure when the newly added
ADAU1977 driver is built as a module.
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NMark Brown <broonie@linaro.org>
Acked-by: NTakashi Iwai <tiwai@suse.de>

Use dev_err() & co as much as possible.  If not available (no device
assigned at the calling point), use pr_xxx() helpers instead.

For simplicity, introduce new helpers for pcm stream, pcm_err(), etc.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When the process is sleeping at the SNDRV_PCM_STATE_PAUSED
state from the wait_for_avail function, the sleep process will be woken by
timeout(10 seconds). Even if the sleep process wake up by timeout, by this
patch, the process will continue with sleep and wait for the other state.
Signed-off-by: NJongHo Kim <furmuwon@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The use of snd_printd_ratelimit() supresses superfluous output from
printk_ratelimit() when CONFIG_SND_DEBUG is not defined. For example,

[   43.753692] snd_pcm_update_hw_ptr0: 26 callbacks suppressed
[   48.822131] snd_pcm_update_hw_ptr0: 25 callbacks suppressed
[   53.894953] snd_pcm_update_hw_ptr0: 25 callbacks suppressed
[   58.997761] snd_pcm_update_hw_ptr0: 25 callbacks suppressed
[   64.100952] snd_pcm_update_hw_ptr0: 25 callbacks suppressed

fills the log even when no debug output is actually produced.

Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Yacine Belkadi <yacine.belkadi.1@gmail.com>
Signed-off-by: NTim Gardner <tim.gardner@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The ops field of the snd_pcm_substream struct is never modified inside the ALSA
core. Making it const allows drivers to declare their snd_pcm_ops struct as
const.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

script/kernel-doc reports the following type of warnings (when run in verbose
mode):

Warning(sound/core/init.c:152): No description found for return value of
'snd_card_create'

To fix that:
- add missing descriptions of function return values
- use "Return:" sections to describe those return values

Along the way:
- complete some descriptions
- fix some typos
Signed-off-by: NYacine Belkadi <yacine.belkadi.1@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

ALSA did not provide any direct means to infer the audio time for A/V
sync and system/audio time correlations (eg. PulseAudio).
Applications had to track the number of samples read/written and
add/subtract the number of samples queued in the ring buffer.  This
accounting led to small errors, typically several samples, due to the
two-step process.  Computing the audio time in the kernel is more
direct, as all the information is available in the same routines.

Also add new .audio_wallclock routine to enable fine-grain synchronization
between monotonic system time and audio hardware time.
Using the wallclock, if supported in hardware, allows for a
much better sub-microsecond precision and a common drift tracking for
all devices sharing the same wall clock (master clock).
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Keep track of boundary crossing when hw_ptr
exceeds boundary limit and wraps-around. This
will help keep track of total number
of frames played/received at the kernel level
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

A few files have been slipped from the previous commit to add MONO
channel type.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>