Added tracepoints obsoleted RULES_DEBUG local macro and relevant codes.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

In a previous commit, tracepoints are added for PCM parameter processing.
As long as I know, this implementation increases size of relocatable
object by 35%. For vendors who are conscious of memory footprint, it
brings apparent disadvantage.

This commit utilizes CONFIG_SND_DEBUG configuration to enable/disable the
tracepoints.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When working for devices which support configurable modes for its data
transmission or which consists of several components, developers are
likely to use rules of parameters of PCM substream. However, there's no
infrastructure to assist their work.

In old days, ALSA PCM core got a local 'RULES_DEBUG' macro to debug
refinement of parameters for PCM substream. Although this is merely a
makeshift. With some modifications, we get the infrastructure.

This commit is for the purpose. Refinement of mask/interval type of
PCM parameters is probed as tracepoint events as 'hw_mask_param' and
'hw_interval_param' on existent 'snd_pcm' subsystem.
Signed-off-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
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>

Although the ack callback is supposed to be called at each appl_ptr or
hw_ptr update, we missed a few opportunities: namely, forward, rewind
and sync_ptr.

Formerly calling ack at rewind may have leaded to unexpected results
due to the forgotten negative appl_ptr update in indirect-PCM helper,
which is the major user of the PCM ack callback.  But now we fixed
this oversights, thus we can call ack callback safely even at rewind
callback -- of course with the proper handling of the error from the
callback.

This patch adds the calls of ack callback in the places mentioned in
the above.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

PCM core code has a few usages of set_fs(), mostly for two codepaths:
- The DELAY ioctl call from pcm_compat.c
- The ioctl wrapper in kernel context for PCM OSS and other

This patch removes the set_fs() usage in these places by a slight code
refactoring.  For the former point, snd_pcm_delay() is changed to
return the  value directly instead of putting the value to the given
address.  Each caller stores the result in an appropriate manner.

For fixing the latter, snd_pcm_lib_kernel_ioctl() is changed to call
the functions directly as well.  For achieving it, now the function
accepts only the limited set of ioctls that have been used, so far.
The primary user of this function is the PCM OSS layer, and the only
other user is USB UAC1 gadget driver.  Both drivers don't need the
full set of ioctls.
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Factor out the common codes in snd_pcm_*_forward() and *_rewind()
functions to simplify the codes.  No functional changes.
Reviewd-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The mostly same codes for checking the current PCM state and calling
hwsync are found in a few places.  This patch simplifies them by
creating a common helper function.

It also fixes a couple of cases where we missed the proper state check
(e.g. PAUSED state wasn't handled in rewind and snd_pcm_hwsync()),
too.
Reviewed-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>

There's a read-only table for each sampling rate, while it doesn't have
const qualifier and can be modified.

This commit add the qualifier. As a result, a symbol for the table
moves from .data section to .rodata.
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>

->fault(), ->page_mkwrite(), and ->pfn_mkwrite() calls do not need to
take a vma and vmf parameter when the vma already resides in vmf.

Remove the vma parameter to simplify things.

[arnd@arndb.de: fix ARM build]
  Link: http://lkml.kernel.org/r/20170125223558.1451224-1-arnd@arndb.de
Link: http://lkml.kernel.org/r/148521301778.19116.10840599906674778980.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Reviewed-by: NRoss Zwisler <ross.zwisler@linux.intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: Darrick J. Wong <darrick.wong@oracle.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jan Kara <jack@suse.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When the stream is in suspended state some applications wait
on "Stream Pipe Error" in response to snd_pcm_avail call to
resume the stream.

In the current implementation snd_pcm_avail() returns zero
when the stream is in suspended state. This causes application
to enter in infinite loop for frames to be available.

"Stream pipe Error" code is getting returned for read/write
call when the stream is in suspended state. Similarly update
snd_pcm_avail to return -ESTRPIPE.
Signed-off-by: NJeeja KP <jeeja.kp@intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

We can't return a negative error code from the poll callback the return
type is unsigned and is checked against the poll specific flags we need
to return POLLERR if we encounter an error.
Signed-off-by: NCharles Keepax <ckeepax@opensource.wolfsonmicro.com>
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

A non-atomic PCM stream may take snd_pcm_link_rwsem rw semaphore twice
in the same code path, e.g. one in snd_pcm_action_nonatomic() and
another in snd_pcm_stream_lock().  Usually this is OK, but when a
write lock is issued between these two read locks, the problem
happens: the write lock is blocked due to the first reade lock, and
the second read lock is also blocked by the write lock.  This
eventually deadlocks.

The reason is the way rwsem manages waiters; it's queued like FIFO, so
even if the writer itself doesn't take the lock yet, it blocks all the
waiters (including reads) queued after it.

As a workaround, in this patch, we replace the standard down_write()
with an spinning loop.  This is far from optimal, but it's good
enough, as the spinning time is supposed to be relatively short for
normal PCM operations, and the code paths requiring the write lock
aren't called so often.
Reported-by: NVinod Koul <vinod.koul@intel.com>
Tested-by: NRamesh Babu <ramesh.babu@intel.com>
Cc: <stable@vger.kernel.org> # v3.18+
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The action_ops structures are never modified, so declare them as const.

Done with the help of Coccinelle.
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Reviewed-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
Tested-by: NTakashi Sakamoto <o-takashi@sakamocchi.jp>
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>

snd_pcm_release_substream() always calls hw_free op when the stream
was opened.  This is superfluous in most cases because it's been
already released via explicit hw_free ioctl.  Although this double
call is usually OK as this callback should be written to be called
multiple times, it's better to avoid superfluous calls.
Reported-by: NVinod Koul <vinod.koul@intel.com>
Tested-by: NJeeja Kp <jeeja.kp@intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

As far as I can see, having an invalid ->tstamp_mode is harmless, but
adding a check silences a static checker warning.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

With the nonatomic PCM ops, the system may spew lockdep warnings like:

 =============================================
 [ INFO: possible recursive locking detected ]
 4.2.0-rc1-jeejaval3 #12 Not tainted
 ---------------------------------------------
 aplay/4029 is trying to acquire lock:
  (snd_pcm_link_rwsem){.+.+.+}, at: [<ffffffff816fd473>] snd_pcm_stream_lock+0x43/0x60

 but task is already holding lock:
  (snd_pcm_link_rwsem){.+.+.+}, at: [<ffffffff816fcf29>] snd_pcm_action_nonatomic+0x29/0x80

 other info that might help us debug this:
  Possible unsafe locking scenario:

        CPU0
        ----
   lock(snd_pcm_link_rwsem);
   lock(snd_pcm_link_rwsem);

Although this is false-positive as the rwsem is taken always as
read-only for these code paths, it's certainly annoying to see this at
any occasion.  A simple fix is to use down_read_nested() in
snd_pcm_stream_lock() that can be called inside another lock.
Reported-by: NVinod Koul <vinod.koul@intel.com>
Reported-by: NJeeja Kp <jeeja.kp@intel.com>
Tested-by: NJeeja Kp <jeeja.kp@intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

readv() and writev() should _not_ ignore all but the first ->iov_len,
among other things.  Really weird abuse of those syscalls - it
expects a vector element per channel, with identical lengths (it
actually assumes them to be identical - no checking is done).
readv() and writev() are really bad match for that.  Unfortunately,
userland API is userland API and we can't do anything about them.

Converted to ->read_iter/->write_iter.  Please, _please_ don't do
anything of that kind when designing new interfaces.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

struct kiocb now is a generic I/O container, so move it to fs.h.
Also do a #include diet for aio.h while we're at it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

When a PCM draining is performed to an empty stream that has been
already in PREPARED state, the current code just ignores and leaves as
it is, although the drain is supposed to set all such streams to SETUP
state.  This patch covers that overlooked case.

Cc: <stable@vger.kernel.org>
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>

Let userspace select audio timestamp config when the
STATUS_EXT ioctl is used, ignore and zero all
other fields
No change for the existing STATUS ioctl, parameters
are treated as read-only.
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Don't use generic snapshot of trigger_tstamp if low-level driver or
hardware can get a more precise value for better audio/system time
synchronization.
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

timestamp in RUNNING mode is already taken in update_hw_ptr routine,
getting a new timestamp introduces offset between hw_ptr, audio_tstamp
and system time

Add else condition to read timestamp as fallback and only when
enabled
Signed-off-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Nowadays it's recommended.  Replace all in a shot.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add SNDRV_PCM_TRIGGER_DRAIN trigger for pcm drain.

Some audio devices require notification of drain events
in order to properly drain and shutdown an audio stream.
Signed-off-by: NLibin Yang <libin.yang@intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Fix a copy & paste error:
 Warning(sound/core/pcm_native.c:1112): Excess function parameter 'state' description in 'snd_pcm_stop_xrun'

The state argument was dropped from snd_pcm_stop_xrun().
Reported-by: Nkbuild test robot <fengguang.wu@intel.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add a new helper function snd_pcm_stop_xrun() to the standard sequnce
lock/snd_pcm_stop(XRUN)/unlock by a single call, and replace the
existing open codes with this helper.

The function checks the PCM running state to prevent setting the wrong
state, too, for more safety.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Some state changes (e.g. snd_pcm_stop()) sets the runtime state after
calling snd_timer_notify().  This is basically racy, since the
notification may wakes up the user even before the state change.
Although the possibility is low, we should set the state before the
notifications.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Just a small code refactoring to reduce more lines.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The function snd_pcm_action_lock_irq() can be much simplified by
simply wrapping snd_pcm_action() with the stream lock.  This was
rather the original idea, but later it was open coded for
optimization.  However, looking at the optimization part closely, one
notices that the probability of the optimized path is quite low; in
normal situations, the linked stream action happens only for the
triggered substream, thus the operation becomes identical.  So the
code simplification has a clear win, especially because we have now
doubly codes for both atomic and non-atomic locks.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add proper kerneldoc comments to the exported functions.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

On x86, using dma_mmap_coherent() for the pages allocated via
dma_alloc_coherent() results in a warning like:
  aplay:32536 map pfn RAM range req uncached-minus for [mem 0x21d500000-0x21d51ffff], got write-back

Until the issue is addressed in the core side, take back to the old
good way in PCM code only for x86.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Some architectures like PARISC is known not to support mmap properly
with the DMA buffer, where dma_mmap_coherent() returns -EINVAL
unconditionally.  From the API POV, we should rather drop the mmap
support there and expose it before the user-space tries to call mmap.

The patch contains again ugly ifdef's, unfortunately, as there is no
global flag indicating this.  Once when such macro is defined, we can
get rid of this instead.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Since we have consistently dma_mmap_coherent() for all architectures,
the current ifdef and arch-specific codes in pcm core can be cleaned
up gracefully.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

As PCM core handles the multiple linked streams in parallel, lockdep
gets confused (partly because of weak annotations) and spews the
false-positive warnings.  This hasn't been a problem for long time but
the latest PCM lock path update seems to have woken up a sleeping
dog.

Here is an attempt to paper over this issue: pass the lock subclass
just calculated from the depth in snd_pcm_action_group().  Also, a
(possibly) wrong lock subclass set in snd_pcm_action_lock_mutex() is
dropped, too.
Reported-and-tested-by: NArthur Marsh <arthur.marsh@internode.on.net>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

This avoids following kernel crash when try to playback on arm64

[  107.497203] [<ffffffc00046b310>] snd_pcm_mmap_data_fault+0x90/0xd4
[  107.503405] [<ffffffc0001541ac>] __do_fault+0xb0/0x498
[  107.508565] [<ffffffc0001576a0>] handle_mm_fault+0x224/0x7b0
[  107.514246] [<ffffffc000092640>] do_page_fault+0x11c/0x310
[  107.519738] [<ffffffc000081100>] do_mem_abort+0x38/0x98

Tested: backported to 3.14 and tried to playback on arm64 machine
Signed-off-by: NAnatol Pomozov <anatol.pomozov@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>