syzbot caught a potential deadlock between the PCM
runtime->buffer_mutex and the mm->mmap_lock. It was brought by the
recent fix to cover the racy read/write and other ioctls, and in that
commit, I overlooked a (hopefully only) corner case that may take the
revert lock, namely, the OSS mmap. The OSS mmap operation
exceptionally allows to re-configure the parameters inside the OSS
mmap syscall, where mm->mmap_mutex is already held. Meanwhile, the
copy_from/to_user calls at read/write operations also take the
mm->mmap_lock internally, hence it may lead to a AB/BA deadlock.
A similar problem was already seen in the past and we fixed it with a
refcount (in commit b248371628). The former fix covered only the
call paths with OSS read/write and OSS ioctls, while we need to cover
the concurrent access via both ALSA and OSS APIs now.
This patch addresses the problem above by replacing the buffer_mutex
lock in the read/write operations with a refcount similar as we've
used for OSS. The new field, runtime->buffer_accessing, keeps the
number of concurrent read/write operations. Unlike the former
buffer_mutex protection, this protects only around the
copy_from/to_user() calls; the other codes are basically protected by
the PCM stream lock. The refcount can be a negative, meaning blocked
by the ioctls. If a negative value is seen, the read/write aborts
with -EBUSY. In the ioctl side, OTOH, they check this refcount, too,
and set to a negative value for blocking unless it's already being
accessed.
Reported-by: syzbot+6e5c88838328e99c7e1c@syzkaller.appspotmail.com
Fixes: dca947d4d2 ("ALSA: pcm: Fix races among concurrent read/write and buffer changes")
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/000000000000381a0d05db622a81@google.com
Link: https://lore.kernel.org/r/20220330120903.4738-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently we have neither proper check nor protection against the
concurrent calls of PCM hw_params and hw_free ioctls, which may result
in a UAF. Since the existing PCM stream lock can't be used for
protecting the whole ioctl operations, we need a new mutex to protect
those racy calls.
This patch introduced a new mutex, runtime->buffer_mutex, and applies
it to both hw_params and hw_free ioctl code paths. Along with it, the
both functions are slightly modified (the mmap_count check is moved
into the state-check block) for code simplicity.
Reported-by: Hu Jiahui <kirin.say@gmail.com>
Cc: <stable@vger.kernel.org>
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Link: https://lore.kernel.org/r/20220322170720.3529-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Quite a few fixes here, including an unusually large set in the core
spurred on by various testing efforts as well as the usual small driver
fixes. There are quite a few fixes for out of bounds writes in both the
core and the various Qualcomm drivers, plus a couple of fixes for
locking in the DPCM code.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEEreZoqmdXGLWf4p/qJNaLcl1Uh9AFAmH5Vi4ACgkQJNaLcl1U
h9Ah6gf/SdlpKMKgAYiXAfmhb/88UgWkXWwHYgK+uHQXy3VpiG3zOso2uHQTbZ0X
+QMZ1LwPbcSsz6Ny/w/kQ4MKeklHou0X6GiQ6PRoDukBPQ6yidivwnVXt35qNBwT
iYmMuLpWsq0CrbUp9cPX+zurX+jUv/CKuwG1JnHxS2+lQ7UjetUOfKzO80X4zUA5
tm12soU9PSUcSyBICg5Jn+Tt9ZPccJ4cDPLfdRVkKjwb+6o4CnuM56Jq9UeHY6KT
SGr8/Jt0I61p8UHPUpqJsj2ItHhy80bNraSaNptQljlIYCPidgFlPKyXxgunm8e5
Z2PUG2mOIR4pk0cj1SxOXwhH9+d3dQ==
=UdmQ
-----END PGP SIGNATURE-----
Merge tag 'asoc-fix-v5.17-rc2' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Fixes for v5.17
Quite a few fixes here, including an unusually large set in the core
spurred on by various testing efforts as well as the usual small driver
fixes. There are quite a few fixes for out of bounds writes in both the
core and the various Qualcomm drivers, plus a couple of fixes for
locking in the DPCM code.
The recent change for DPCM locking caused spurious lockdep warnings.
Actually the warnings are false-positive, as those are triggered due
to the nested stream locks for FE and BE. Since both locks belong to
the same lock class, lockdep sees it as if a deadlock.
For fixing this, we need to take PCM stream locks for BE with the
nested lock primitives. Since currently snd_pcm_stream_lock*() helper
assumes only the top-level single locking, a new helper function
snd_pcm_stream_lock_irqsave_nested() is defined for a single-depth
nested lock, which is now used in the BE DAI trigger that is always
performed inside a FE stream lock.
Fixes: b2ae806630 ("ASoC: soc-pcm: serialize BE triggers")
Reported-and-tested-by: Hans de Goede <hdegoede@redhat.com>
Reported-and-tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Link: https://lore.kernel.org/r/73018f3c-9769-72ea-0325-b3f8e2381e30@redhat.com
Link: https://lore.kernel.org/alsa-devel/9a0abddd-49e9-872d-2f00-a1697340f786@samsung.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20220119155249.26754-2-tiwai@suse.de
Signed-off-by: Mark Brown <broonie@kernel.org>
Regarding to handling [U|S][32|24] PCM formats, many userspace
application developers and driver developers have confusion, since they
require them to understand justification or padding. It easily
loses consistency and soundness to operate with many type of devices. In
this commit, I attempt to solve the situation by adding comment about
relation between [S|U]32 formats and 'msbits' hardware parameter.
The formats are used for 'left-justified' sample format, and the available
bit count in most significant bit is delivered to userspace in msbits
hardware parameter (struct snd_pcm_hw_params.msbits), which is decided by
msbits constraint added by pcm drivers (snd_pcm_hw_constraint_msbits()).
In driver side, the msbits constraint includes two elements; the physical
width of format and the available width of the format in most significant
bit. The former is used to match SAMPLE_BITS of format. (For my
convenience, I ignore wildcard in the usage of the constraint.)
As a result of interaction between ALSA pcm core and ALSA pcm application,
when the format in which SAMPLE_BITS equals to physical width of the
msbits constaint, the msbits parameter is set by referring to the
available width of the constraint. When the msbits parameter is not
changed in the above process, ALSA pcm core set it alternatively with
SAMPLE_BIT of chosen format.
In userspace application side, the msbits is only available after calling
ioctl(2) with SNDRV_PCM_IOCTL_HW_PARAMS request. Even if the hardware
parameter structure includes somewhat value of SAMPLE_BITS interval
parameter as width of format, all of the width is not always available
since msbits can be less than the width.
I note that [S|U]24 formats are used for 'right-justified' 24 bit sample
formats within 32 bit frame. The first byte in most significant bit
should be invalidated. Although the msbits exposed to userspace should be
zero as invalid value, actually it is 32 from physical width of format.
[ corrected typos -- tiwai ]
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20210529033353.21641-1-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
A few drivers want to have rather the exact buffer preallocation at
the driver probe time and keep using it for the whole operations
without allowing dynamic buffer allocation. For satisfying the
demands, this patch extends the managed buffer allocation API
slightly.
Namely, when 0 is passed to max argument of the allocation helper
functions snd_pcm_set_managed_buffer*(), it treats as if the fixed
size allocation of the given size. If the pre-allocation fails in
this mode, the function returns now -ENOMEM. Otherwise, i.e. max
argument is non-zero, the function never returns -ENOMEM but tries to
fall back to the smaller chunks and allows the dynamic allocation
later -- which is still the default behavior until now.
For more intuitive use, also two new helpers are added for handling
the fixed size buffer allocation, too: snd_pcm_set_fixed_buffer() and
snd_pcm_set_fixed_buffer_all().
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20210802072815.13551-4-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch introduces the ops table to each memory allocation type
(SNDRV_DMA_TYPE_XXX) and abstract the handling for the better code
management. Then we get separate the page allocation, release and
other tasks for each type, especially for the SG buffer.
Each buffer type has now callbacks in the struct snd_malloc_ops, and
the common helper functions call those ops accordingly. The former
inline code that is specific to SG-buffer is moved into the local
sgbuf.c, and we can simplify the PCM code without details of memory
handling.
Link: https://lore.kernel.org/r/20210609162551.7842-4-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Current implementation of ALSA PCM core has a kernel API,
snd_pcm_period_elapsed(), for drivers to queue event to awaken processes
from waiting for available frames. The function voluntarily acquires lock
of PCM substream, therefore it is not called in process context for any
PCM operation since the lock is already acquired.
It is convenient for packet-oriented driver, at least for drivers to audio
and music unit in IEEE 1394 bus. The drivers are allowed by Linux
FireWire subsystem to process isochronous packets queued till recent
isochronous cycle in process context in any time.
This commit adds snd_pcm_period_elapsed() variant,
snd_pcm_period_elapsed_without_lock(), for drivers to queue the event in
the process context.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20210610031733.56297-2-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The fix for a long-standing USB-audio bug required one more dependency
variable to be added to the hw constraints. Unfortunately I didn't
realize at debugging that the new addition may result in the overflow
of the dependency array of each snd_pcm_hw_rule (up to three plus a
sentinel), because USB-audio driver adds one more dependency only for
a certain device and bus, hence it works as is for many devices. But
in a bad case, a simple open always results in -EINVAL (with kernel
WARNING if CONFIG_SND_DEBUG is set) no matter what is passed.
Since the dependencies are real and unavoidable (USB-audio restricts
the hw_params per looping over the format/rate/channels combos), the
only good solution seems to raise the bar for one more dependency for
snd_pcm_hw_rule -- so does this patch: now the hw constraint
dependencies can be up to four.
Fixes: 506c203cc3 ("ALSA: usb-audio: Fix hw constraints dependencies")
Reported-by: Jamie Heilman <jamie@audible.transient.net>
Link: https://lore.kernel.org/r/20210123155730.22576-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Kernel-doc markups should use this format:
identifier - description
There is a common comment marked, instead, with kernel-doc
notation.
Some identifiers have different names between their prototypes
and the kernel-doc markup.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/535182d6f55d7a7de293dda9676df68f5f60afc6.1603469755.git.mchehab+huawei@kernel.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This is a very big update for the core since Morimoto-san has been
rather busy continuing his refactorings to clean up a lot of the cruft
that we have accumilated over the years. We've also gained several new
drivers, including initial (but still not complete) parts of the Intel
SoundWire support.
- Lots of refactorings to modernize the code from Morimoto-san.
- Conversion of SND_SOC_ALL_CODECS to use imply from Geert Uytterhoeven.
- Continued refactoring and fixing of the Intel support.
- Soundwire and more advanced clocking support for Realtek RT5682.
- Support for amlogic GX, Meson 8, Meson 8B and T9015 DAC, Broadcom
DSL/PON, Ingenic JZ4760 and JZ4770, Realtek RL6231, and TI TAS2563 and
TLV320ADCX140.
-----BEGIN PGP SIGNATURE-----
iQFHBAABCgAxFiEEreZoqmdXGLWf4p/qJNaLcl1Uh9AFAl6B03sTHGJyb29uaWVA
a2VybmVsLm9yZwAKCRAk1otyXVSH0Jm5B/4pb6Hz1ifM+ikR4qB7BOdm/UGzxWlD
Mng+qN3kh6Tx+x8/xnQRmfOnhFoKnDSTnlxcDr6ujAtz3bH+CKXbo2jjfvmagQrW
j87tYGdh6T38E6mf6543Eth3zJGphZYUY70REXt0HsaQ8bA912nrs5H8LTFiD2i+
lBJd/MmlCYL3gb/ftN9dpTfYYdxLy9HH9j5tR5zpLN3AD3Hew9jHqqX+ioAk+6lA
Nisbx5grP10pWFklYXyJlUFQOqMOYMtPMblLRA9bjqcEtL3quIXPXKkPNtTh+0Kf
5t7Q39RtVXJAk0DcatjMJUGAGQyM9r5g1Nk8WLB1LLv+NfLcO6cUpLAc
=+z/q
-----END PGP SIGNATURE-----
Merge tag 'asoc-v5.7' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Updates for v5.7
This is a very big update for the core since Morimoto-san has been
rather busy continuing his refactorings to clean up a lot of the cruft
that we have accumilated over the years. We've also gained several new
drivers, including initial (but still not complete) parts of the Intel
SoundWire support.
- Lots of refactorings to modernize the code from Morimoto-san.
- Conversion of SND_SOC_ALL_CODECS to use imply from Geert Uytterhoeven.
- Continued refactoring and fixing of the Intel support.
- Soundwire and more advanced clocking support for Realtek RT5682.
- Support for amlogic GX, Meson 8, Meson 8B and T9015 DAC, Broadcom
DSL/PON, Ingenic JZ4760 and JZ4770, Realtek RL6231, and TI TAS2563 and
TLV320ADCX140.
It can be useful to derive min/max rates of a snd_pcm_hardware without
having a snd_pcm_runtime, such as before constructing an ASoC DAI link.
Create a new helper that takes a pointer to a snd_pcm_hardware directly,
and refactor the original function as a wrapper around it, to avoid
needing to update any call sites.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20200305051143.60691-2-samuel@sholland.org
Signed-off-by: Mark Brown <broonie@kernel.org>
ALSA code has SNDRV_PCM_STREAM_PLAYBACK/CAPTURE everywhere.
Having for_each_xxxx macro is useful.
This patch adds for_each_pcm_streams() for it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Acked-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/874kvpbotq.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
snd_pcm_format_t is a strong-typed integer and requires the explicit
cast with __force if converted or compared with a normal integer
value. Since most of use cases do iterate over all formats and test /
set the mask, provide a couple of new helper macros that do the
explicit cast.
Link: https://lore.kernel.org/r/20200206163945.6797-3-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Since we have a bitwise definition of snd_pcm_state_t and use it for
certain struct fields, a few new (and years old) sparse warnings came
up. This patch is an attempt to cover them.
- The state fields in snd_pcm_mmap_status* and co are all defined as
snd_pcm_state_t type now
- The PCM action callbacks take snd_pcm_state_t argument as well;
some actions taking special values got the explicit cast and
comments
- For the PCM action that doesn't need an extra argument receives
ACTION_ARG_IGNORE instead of ambiguous 0
While we're at it, the boolean argument is also properly changed to
bool and true/false, as well as a slight refactoring of PCM pause
helper function to make easier to read.
No functional changes, just shutting up chatty sparse.
Fixes: 46b770f720 ("ALSA: uapi: Fix sparse warning")
Reported-by: kbuild test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20200131152214.11698-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The struct snd_pcm_status will use 'timespec' type variables to record
timestamp, which is not year 2038 safe on 32bits system.
Userspace will use SNDRV_PCM_IOCTL_STATUS and SNDRV_PCM_IOCTL_STATUS_EXT
as commands to issue ioctl() to fill the 'snd_pcm_status' structure in
userspace. The command number is always defined through _IOR/_IOW/IORW,
so when userspace changes the definition of 'struct timespec' to use
64-bit types, the command number also changes.
Thus in the kernel, we now need to define two versions of each such ioctl
and corresponding ioctl commands to handle 32bit time_t and 64bit time_t
in native mode:
struct snd_pcm_status32 {
......
s32 trigger_tstamp_sec;
s32 trigger_tstamp_nsec;
......
s32 audio_tstamp_sec;
s32 audio_tstamp_nsec;
......
};
struct snd_pcm_status64 {
......
s32 trigger_tstamp_sec;
s32 trigger_tstamp_nsec;
......
s32 audio_tstamp_sec;
s32 audio_tstamp_nsec;
......
};
Moreover in compat file, we renamed or introduced new structures to handle
32bit/64bit time_t in compatible mode. The 'struct snd_pcm_status32' and
snd_pcm_status_user32() are used to handle 32bit time_t in compat mode.
'struct compat_snd_pcm_status64' and snd_pcm_status_user_compat64() are used
to handle 64bit time_t.
The implicit padding before timespec is made explicit to avoid incompatible
structure layout between 32-bit and 64-bit x86 due to the different
alignment requirements, and the snd_pcm_status structure is now hidden
from the kernel to avoid relying on the timespec definitio definitionn
Finally we can replace SNDRV_PCM_IOCTL_STATUS and SNDRV_PCM_IOCTL_STATUS_EXT
with new commands and introduce new functions to fill new 'struct snd_pcm_status64'
instead of using unsafe 'struct snd_pcm_status'. Then in future, the new
commands can be matched when userspace changes 'timespec' to 64bit type
to make a size change of 'struct snd_pcm_status'. When glibc changes time_t
to 64-bit, any recompiled program will issue ioctl commands that the kernel
does not understand without this patch.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Since timespec is not year 2038 safe on 32bit system, and we need to
convert all timespec variables to timespec64 type for sound subsystem.
This patch is used to do preparation for following patches, that will
convert all structures defined in uapi/sound/asound.h to use 64-bit
time_t.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
The standard programming model of a PCM sound driver is to process
snd_pcm_period_elapsed() from an interrupt handler. When a running
stream is stopped, PCM core calls the trigger-STOP PCM ops, sets the
stream state to SETUP, and moves on to the next step. This is
performed in an atomic manner -- this could be called from the interrupt
context, after all.
The problem is that, if the stream goes further and reaches to the
CLOSE state immediately, the stream might be still being processed in
snd_pcm_period_elapsed() in the interrupt context, and hits a NULL
dereference. Such a crash happens because of the atomic operation,
and we can't wait until the stream-stop finishes.
For addressing such a problem, this commit adds a new PCM ops,
sync_stop. This gets called at the appropriate places that need a
sync with the stream-stop, i.e. at hw_params, prepare and hw_free.
Some drivers already have a similar mechanism implemented locally, and
we'll refactor the code later.
Link: https://lore.kernel.org/r/20191117085308.23915-7-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch adds the support for the feature to automatically allocate
and free PCM buffers, so called "managed buffer allocation" mode.
It's set up via new PCM helpers, snd_pcm_set_managed_buffer() and
snd_pcm_set_managed_buffer_all(), both of which correspond to the
existing preallocator helpers, snd_pcm_lib_preallocate_pages() and
snd_pcm_lib_preallocate_pages_for_all(). When the new helper is used,
it not only performs the pre-allocation of buffers, but also it
manages to call snd_pcm_lib_malloc_pages() before the PCM hw_params
ops and snd_lib_pcm_free() after the PCM hw_free ops inside PCM core,
respectively. This allows drivers to drop the explicit calls of the
memory allocation / release functions, and it will be a good amount of
code reduction in the end of this patch series.
When the PCM substream is set to the managed buffer allocation mode,
the managed_buffer_alloc flag is set in the substream object. Since
some drivers want to know when a buffer is newly allocated or
re-allocated at hw_params callback (e.g. want to set up the additional
stuff for the given buffer only at allocation time), now PCM core
turns on buffer_changed flag when the buffer has changed.
The standard conversions to use the new API will be straightforward:
- Replace snd_pcm_lib_preallocate*() calls with the corresponding
snd_pcm_set_managed_buffer*(); the arguments should be unchanged
- Drop superfluous snd_pcm_lib_malloc() and snd_pcm_lib_free() calls;
the check of snd_pcm_lib_malloc() returns should be replaced with
the check of runtime->buffer_changed flag.
- If hw_params or hw_free becomes empty, drop them from PCM ops
Link: https://lore.kernel.org/r/20191117085308.23915-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The helper is no longer referred after the recent code refactoring.
Drop the export for saving some bits and future misuse.
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20191108094641.20086-9-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Most of the modern codecs supports 352.8KHz and 384KHz sample rates.
Currenlty HW params fails to set 352.8Kz and 384KHz sample rate
as these are not in known rates list.
Add these new rates to known list to allow them.
This patch also adds defines in pcm.h so that drivers can use it.
Signed-off-by: Vidyakumar Athota <vathota@codeaurora.org>
Signed-off-by: Banajit Goswami <bgoswami@codeaurora.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20190822095653.7200-2-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version this program is distributed in the
hope that it will be useful but without any warranty without even
the implied warranty of merchantability or fitness for a particular
purpose see the gnu general public license for more details you
should have received a copy of the gnu general public license along
with this program if not write to the free software foundation inc
59 temple place suite 330 boston ma 02111 1307 usa
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 1334 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Richard Fontana <rfontana@redhat.com>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190527070033.113240726@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now all callers no longer check the return value from
snd_pcm_lib_preallocate_pages() and co, let's make them to return
void, so that any new code won't fall into the same pitfall.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The proc files are recursively freed by calling with the root
snd_info_entry object, so we don't have to keep each object for
releasing one by one. Move the release of the PCM stream proc root at
the beginning, so that we can remove the redundant code and resource.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
We have currently two global locks, a rwlock and a rwsem, that are
used for managing linking the PCM streams. Due to these global locks,
once when a linked stream is used, the lock granularity suffers a
lot.
This patch attempts to eliminate the former global lock for atomic
ops. The latter rwsem needs remaining because of the loosy way of the
loop calls in snd_pcm_action_nonatomic(), as well as for avoiding the
deadlock at linking. However, these are used far rarely, actually
only by two actions (prepare and reset), where both are no timing
critical ones. So this can be still seen as a good improvement.
The basic strategy to eliminate the rwlock is to assure group->lock at
adding or removing a stream to / from the group. Since we already
takes the group lock whenever taking the all substream locks under the
group, this shouldn't be a big problem. The reference to group
pointer in snd_pcm_substream object is protected by the stream lock
itself.
However, there are still pitfalls: a race window at re-locking and the
lifecycle of group object. The former is a small race window for
dereferencing the substream group object opened while snd_pcm_action()
performs re-locking to avoid ABBA deadlocks. This includes the unlink
of group during that window, too. And the latter is the kfree
performed after all streams are removed from the group while it's
still dereferenced.
For addressing these corner cases, two new tricks are introduced:
- After re-locking, the group assigned to the stream is checked again;
if the group is changed, we retry the whole procedure.
- Introduce a refcount to snd_pcm_group object, so that it's freed
only when it's empty and really no one refers to it.
(Some readers might wonder why not RCU for the latter. RCU in this
case would cost more than refcounting, unfortunately. We take the
group lock sooner or later, hence the performance improvement by RCU
would be negligible. Meanwhile, because we need to deal with
schedulable context depending on the pcm->nonatomic flag, it'll become
dynamic RCU/SRCU switch, and the grace period may become too long.)
Along with these changes, there are a significant amount of code
refactoring. The complex group re-lock & ref code is factored out to
snd_pcm_stream_group_ref() function, for example.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Make a common helper to re-assign the PCM link using list_move() instead
of open code with manual list_del() and list_add_tail(). This assures
the consistency and we can get rid of snd_pcm_group.count field -- its
purpose is only to check whether the list is singular, and we can know
it by list_is_singular() call now.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_pcm_suspend() is no longer called from outside, so let's make it
local static. Also drop a superfluous NULL check there.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Until now we rely on each driver calling snd_pcm_suspend*() explicitly
at its own PM handling. However, this can be done far more easily by
setting the PM ops to each actual snd_pcm device object.
This patch adds the device_type object for PCM stream and assigns to
each PCM stream object. The type contains only the PM ops for system
suspend; we don't need to deal with the resume in general.
The suspend hook simply calls snd_pcm_suspend_all() for the given PCM
streams. This implies that the PM order is correctly put, i.e. PCM is
suspended before the main (or codec) driver, which should be true in
general. If a special ordering is needed, you'd need to adjust the
device PM order manually later.
This patch introduces a new flag, snd_pcm.no_device_suspend, too.
With this flag set, the PCM device object won't invoke
snd_pcm_suspend_all() by itself. This is needed for ASoC who wants to
manage the PM call orders in its serialized way, and the flag is set
in soc_new_pcm() as default.
For the non-ASoC world, we can get rid of the manual snd_pcm_suspend
calls. This will be done in the later patches.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The snd_pcm_lib_read() and snd_pcm_lib_write() inline functions have
the explicit cast from a user pointer to a kernel pointer, but they
lacks of __force prefix.
This fixes sparse warnings like:
./include/sound/pcm.h:1093:47: warning: cast removes address space of expression
Fixes: 6854121372 ("ALSA: pcm: Direct in-kernel read/write support")
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_pcm_lib_mmap_vmalloc() was supposed to be implemented with
somewhat special for vmalloc handling, but in the end, this turned to
just the default handler, i.e. NULL. As the situation has never
changed over decades, let's rip it off.
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently ALSA core blocks userspace for about 10 seconds for PCM R/W IO.
This needs to be configurable for modern hardware like DSPs where no
pointer update in milliseconds can indicate terminal DSP errors.
Add a substream variable to set the wait time in ms. This allows userspace
and drivers to recover more quickly from terminal DSP errors.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This format is similar to existing SNDRV_PCM_FORMAT_{S,U}20_3 that keep
20-bit PCM samples in 3 bytes, however i.MX6 platform SSI FIFO does not
allow 3-byte accesses (including DMA) so a 4-byte (more conventional)
format is needed for it.
Signed-off-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
We have an ioctl to inform the PCM protocol version the running kernel
supports, but there is no way to know which protocol version the
user-space can understand. This lack of information caused headaches
in the past when we tried to extend the ABI. For example, because we
couldn't guarantee the validity of the reserved bytes, we had to
introduce a new ioctl SNDRV_PCM_IOCTL_STATUS_EXT for assigning a few
new fields in the formerly reserved bits. If we could know that it's
a new alsa-lib, we could assume the availability of the new fields,
thus we could have reused the existing SNDRV_PCM_IOCTL_STATUS.
In order to improve the ABI extensibility, this patch adds a new ioctl
for user-space to inform its supporting protocol version to the
kernel. By reporting the supported protocol from user-space, the
kernel can judge which feature should be provided and which not.
With the addition of the new ioctl, the PCM protocol version is bumped
to 2.0.14, too. User-space checks the kernel protocol version via
SNDRV_PCM_INFO_PVERSION, then it sets the supported version back via
SNDRV_PCM_INFO_USER_PVERSION.
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
SNDRV_PCM_IOCTL1_GSTATE was firstly introduced in v0.9.0, however never
be used and the purpose is missing.
This commit removes the long-abandoned command, bye.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Drivers can implement 'struct snd_pcm_ops.ioctl' to handle some requests
from ALSA PCM core. These requests are internal purpose in kernel land.
Usually common set of operations are used for it.
SNDRV_PCM_IOCTL1_INFO is one of the requests. According to code comment,
it has been obsoleted in the old days.
We can see old releases in ftp.alsa-project.org. The command was firstly
introduced in v0.5.0 release as SND_PCM_IOCTL1_INFO, to allow drivers to
fill data of 'struct snd_pcm_channel_info' type. In v0.9.0 release,
this was obsoleted by the other commands for ioctl(2) such as
SNDRV_PCM_IOCTL_CHANNEL_INFO.
This commit removes the long-abandoned command, bye.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Acked-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Takashi 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: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The markup inside the #if 0 comment actually refers to a
kernel-doc markup. As we're getting rid of DocBook update it.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The PCM notifier code is used only by OSS emulation layer, so we can
build it conditionally for reducing the size.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
... so that the driver can avoid ifdef's for the dead PM callbacks.
The compiler should optimize them out in anyway.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai
Takashi Sakamoto
Mauro Carvalho Chehab
Samuel Holland
Kuninori Morimoto
Baolin Wang
Vidyakumar Athota
Thomas Gleixner
Ricardo Biehl Pasquali
Liam Girdwood
Maciej S. Szmigiero
Mauro Carvalho Chehab