Distorted audio appears occasionally, affecting either playback or
capture and requiring the affected substream to be closed by all
applications and re-opened.
The best way I have found to reproduce the bug is to use dmix in
combination with Chromium, which opens the audio device multiple times
in threads. Anecdotally, the problems appear to have increased with
faster CPUs. I ruled out 32-bit counter wrapping; it often happens
much earlier.
Since applying this patch I have not had problems, where previously
they would occur several times a day.
The patch targets the following issues:
* Check for progress using the counter from the hardware, not after it
has been truncated to the buffer.
This is a clean way to address a possible bug where if a whole
ringbuffer advances between interrupts, it goes unnoticed.
* Move last_period state from chip to pipe
This more logically belongs as part of pipe, and code is reasier to
read if it is "counter position last time a period elapsed".
Now the code has no references to period count. A period is just
when the regular counter crosses a threshold. This increases
readability and reduces scope for bugs.
* Treat period notification and buffer advance independently:
This helps to clarify what is the responsibility of the interrupt
handler, and what is pcm_pointer().
Removing shared state between these operations means race conditions
are fixed without introducing locks. Synchronisation is only around
the read of pipe->dma_counter. There may be cache line contention
around "struct audiopipe" but I did not have cause to profile this.
Pay attention to be robust where dma_counter wrapping is not a
multiple of period_size or buffer_size.
This is a revised patch based on feedback from Takashi and Giuliano.
Signed-off-by: Mark Hills <mark@xwax.org>
Link: https://lore.kernel.org/r/20200708101848.3457-5-mark@xwax.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
These are valid conditions in normal circumstances, so do not "warn" but
make them for debugging.
Signed-off-by: Mark Hills <mark@xwax.org>
Link: https://lore.kernel.org/r/20200708101848.3457-4-mark@xwax.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Use of atomics does not make these statements robust:
atomic_inc(&chip->opencount);
if (atomic_read(&chip->opencount) > 1 && chip->rate_set)
chip->can_set_rate=0;
and
if (atomic_read(&chip->opencount)) {
if (chip->opencount) {
changed = -EAGAIN;
} else {
changed = set_digital_mode(chip, dmode);
It would be necessary to atomically increment or decrement the value
and use the returned result. And yet we still need to prevent other
threads making use of "can_set_rate" while we set it.
However in all but one case the atomic is misleading as they are already
running with "mode_mutex" held.
Decisions are made on mode setting are often intrinsically connected
to "opencount" because some operations are not permitted unless
there is sole ownership.
So instead simplify this, and use "mode_mutex" as a lock for all reference
counting and mode setting.
Signed-off-by: Mark Hills <mark@xwax.org>
Link: https://lore.kernel.org/r/20200708101848.3457-2-mark@xwax.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This check is always false, as it's not the responsibilty of the
device-specific code to make this check. It is already checked
in snd_echo_digital_mode_put.
I do not have a Mona interface to test this change.
This patch is in preparation for follow-up patch to modify the
behavior of "opencount".
Signed-off-by: Mark Hills <mark@xwax.org>
Link: https://lore.kernel.org/r/20200708101848.3457-1-mark@xwax.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning. Variables are declared in a header file included from
multiple C files, replace by #defines as suggested by Takashi
sound/usb/line6/driver.h:70:18: warning: ‘SYSEX_EXTRA_SIZE’ defined
but not used [-Wunused-const-variable=]
70 | static const int SYSEX_EXTRA_SIZE = sizeof(line6_midi_id) + 4;
| ^~~~~~~~~~~~~~~~
sound/usb/line6/driver.h:69:18: warning: ‘SYSEX_DATA_OFS’ defined but
not used [-Wunused-const-variable=]
69 | static const int SYSEX_DATA_OFS = sizeof(line6_midi_id) + 3;
| ^~~~~~~~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200707184924.96291-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The commit 3ad796cbc3 ("ALSA: pcm: Use SG-buffer only when direct
DMA is available") introduced a check of the DMA type and this caused
a build error on m68k (and possibly some others) due to the lack of
dma_is_direct() definition. Since the check is needed only for
CONFIG_SND_DMA_SGBUF enablement (i.e. solely x86), use #ifdef instead
of IS_ENABLED() for avoiding such a build error.
Fixes: 3ad796cbc3 ("ALSA: pcm: Use SG-buffer only when direct DMA is available")
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20200707111225.26826-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
External HDMI receivers have analog circuitry that needs to be powered-on
when exiting standby, and a mechanism to detect PCM v. IEC61937 data.
These two steps take time and up to 2-3 seconds of audio may be muted
when starting playback.
Intel hardware (Haswell and beyond) can keep the link active
with a 'silent stream', so that the receiver does not go through those
two steps when valid audio is transmitted. This mechanism relies
on an setting the channel_id as 0xf, sending info packet and preventing
the codec from going to D3, which will increase the platform
static power consumption. The info packet assumes a basic 2ch stereo,
and the silent stream is enabled when connecting a monitor.
In case of format changes the detection of PCM v. IEC61937 needs to
be re-run. In this case there is no way to avoid the 2-3s mute.
The silent stream is enabled with a Kconfig option, as well as a kernel
parameter should there be a need to override the build time default.
This approach is used based on the power_save capability as an example,
but in the future, it may be used with a kcontrol,
depending on UCM support for HDaudio legacy.
Signed-off-by: Harsha Priya <harshapriya.n@intel.com>
Signed-off-by: Emmanuel Jillela <emmanuel.jillela@intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/1594068797-14011-1-git-send-email-harshapriya.n@intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning.
sound/drivers/vx/vx_core.c: In function ‘snd_vx_threaded_irq_handler’:
sound/drivers/vx/vx_core.c:515:3: warning: suggest braces around empty
body in an ‘if’ statement [-Wempty-body]
515 | ; /* so far, nothing to do yet */
| ^
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-23-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warnings. Mark variables as __always_unused.
sound/pci/via82xx.c: In function ‘snd_via82xx_codec_wait’:
sound/pci/via82xx.c:547:6: warning: variable ‘err’ set but not used
[-Wunused-but-set-variable]
547 | int err;
| ^~~
sound/pci/via82xx_modem.c: In function ‘snd_via82xx_codec_wait’:
sound/pci/via82xx_modem.c:401:6: warning: variable ‘err’ set but not
used [-Wunused-but-set-variable]
401 | int err;
| ^~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-21-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning. Mark the 'audiostatus' variable as __always_unused.
sound/pci/es1938.c: In function ‘snd_es1938_interrupt’:
sound/pci/es1938.c:1622:24: warning: variable ‘audiostatus’ set but
not used [-Wunused-but-set-variable]
1622 | unsigned char status, audiostatus;
| ^~~~~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-19-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning. Mark the 'req' variable as __always_unused.
sound/xen/xen_snd_front.c: In function ‘xen_snd_front_stream_close’:
sound/xen/xen_snd_front.c:117:21: warning: variable ‘req’ set but not
used [-Wunused-but-set-variable]
117 | struct xensnd_req *req;
| ^~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-18-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warnings. Two variables are only used for debug logs, mark
with __maybe_unused:
sound/pci/korg1212/korg1212.c: In function ‘snd_korg1212_create’:
sound/pci/korg1212/korg1212.c:2152:36: warning: variable ‘iomem2_size’
set but not used [-Wunused-but-set-variable]
2152 | unsigned ioport_size, iomem_size, iomem2_size;
| ^~~~~~~~~~~
sound/pci/korg1212/korg1212.c:2152:11: warning: variable ‘ioport_size’
set but not used [-Wunused-but-set-variable]
2152 | unsigned ioport_size, iomem_size, iomem2_size;
| ^~~~~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-15-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning. The loopsize variable is only used in compiled-out
code, so mark with __maybe_unused.
sound/pci/emu10k1/emu10k1_patch.c: In function
‘snd_emu10k1_sample_new’:
sound/pci/emu10k1/emu10k1_patch.c:30:22: warning: variable ‘loopsize’
set but not used [-Wunused-but-set-variable]
30 | int truesize, size, loopsize, blocksize;
| ^~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-14-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warnings. Mark variables used for reads as __always_unused.
sound/pci/emu10k1/emu10k1_main.c: In function
‘snd_emu10k1_cardbus_init’:
sound/pci/emu10k1/emu10k1_main.c:626:15: warning: variable ‘value’ set
but not used [-Wunused-but-set-variable]
626 | unsigned int value;
| ^~~~~
sound/pci/emu10k1/emu10k1_main.c: In function
‘snd_emu1010_load_firmware_entry’:
sound/pci/emu10k1/emu10k1_main.c:656:15: warning: variable
‘write_post’ set but not used [-Wunused-but-set-variable]
656 | unsigned int write_post;
| ^~~~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-12-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning by nothing variable as always unused.
sound/pci/aw2/aw2-saa7146.c: In function ‘snd_aw2_saa7146_interrupt’:
sound/pci/aw2/aw2-saa7146.c:333:15: warning: variable ‘iicsta’ set but
not used [-Wunused-but-set-variable]
333 | unsigned int iicsta;
| ^~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-8-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning. One variable is only used in a conditionally-compiled
block, mark as __maybe_unused
sound/pci/echoaudio/echoaudio.c: In function ‘snd_echo_probe’:
sound/pci/echoaudio/echoaudio.c:1958:6: warning: variable ‘i’ set but
not used [-Wunused-but-set-variable]
1958 | int i, err;
| ^
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warnings, mark variables as __always_unused
sound/pci/asihpi/asihpi.c: In function ‘snd_asihpi_tuner_band_get’:
sound/pci/asihpi/asihpi.c:1907:6: warning: variable ‘num_bands’ set
but not used [-Wunused-but-set-variable]
1907 | u32 num_bands;
| ^~~~~~~~~
sound/pci/asihpi/asihpi.c: In function ‘snd_asihpi_tuner_band_put’:
sound/pci/asihpi/asihpi.c:1934:6: warning: variable ‘num_bands’ set
but not used [-Wunused-but-set-variable]
1934 | u32 num_bands;
| ^~~~~~~~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warnings by removing 2 unnecessary initializations and
removing a variable that's not used.
sound/pci/asihpi/asihpi.c: In function ‘snd_asihpi_tuner_band_get’:
sound/pci/asihpi/asihpi.c:1907:6: warning: variable ‘num_bands’ set
but not used [-Wunused-but-set-variable]
1907 | u32 num_bands = 0;
| ^~~~~~~~~
sound/pci/asihpi/asihpi.c: In function ‘snd_asihpi_tuner_band_put’:
sound/pci/asihpi/asihpi.c:1934:6: warning: variable ‘num_bands’ set
but not used [-Wunused-but-set-variable]
1934 | u32 num_bands = 0;
| ^~~~~~~~~
sound/pci/asihpi/asihpi.c: In function ‘snd_asihpi_mux_info’:
sound/pci/asihpi/asihpi.c:2164:6: warning: variable ‘err’ set but not
used [-Wunused-but-set-variable]
2164 | int err;
| ^~~
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix W=1 warning by noting variable as __always_unused.
sound/isa/gus/gus_uart.c: In function ‘snd_gf1_interrupt_midi_in’:
sound/isa/gus/gus_uart.c:16:22: warning: variable ‘data’ set but not
used [-Wunused-but-set-variable]
16 | unsigned char stat, data, byte;
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200702193604.169059-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Also documented the galpess transitions. Please note that these are not
really stream states, but show how the stream steps in gapless mode
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Link: https://lore.kernel.org/r/20200629134737.105993-3-vkoul@kernel.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
So we had some discussions of the stream states, so I thought it is a
good idea to document the state transitions, so add it documentation
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Link: https://lore.kernel.org/r/20200629134737.105993-2-vkoul@kernel.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The module argument passed to snd_card_new() must be a valid non-NULL
pointer when the module support is enabled. Since ASoC driver passes
the argument from each snd_soc_card definition, one may forget to set
the owner field and lead to a NULL module easily.
For catching such an overlook, add a WARN_ON() in snd_card_new().
Also, put the card->module assignment in the ifdef block for a very
minor optimization.
Link: https://lore.kernel.org/r/20200624160300.21703-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This is a patch set inspired by the recent patch Kai-Heng posted about
the HD-audio mic-mute LED control. Currently HD-audio driver deals
with the mute and mic-mute LED in several different ways: primarily
with the direct callback of vmaster hook and capture sync hook, while
another with the LED class device binding. The latter has been used
for binding with the platform device LEDs like Thinkpad, Dell,
Huawei. And, yet, recently we added our own LED classdev for the
mic-mute LED on some HP systems although they are controlled directly
with the callback; it's exposed, however, for the DMIC that is
governed by a different ASoC driver.
This patch set is an attempt to sort out and make them consistent:
namely,
* All LEDs are now controlled via LED class device
* The generic driver provides helper functions to easily build up the
LED class dev and the relevant mixer controls
* Conversion of the existing framework and clean ups
The patches are lightly tested in my side with a couple of machines
and also through hda-emu tests. Some devices receive new kcontrols
for the mute LED behavior (that have been missing so far), but
anything else look good though my tests.
Link: https://lore.kernel.org/r/20200618110842.27238-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In the end we already enabled the sync-write mode for most of HD-audio
controllers including Intel, and it's no big merit to keep the async
write mode for the rest. Let's make it as default and drop the
superfluous AZX_DCAPS_SYNC_WRITE bit flag.
Also, avoid to set the allow_bus_reset flag, which is a quite unstable
and hackish behavior that was needed only for some early platforms
(decades ago). The straight fallback to the single cmd mode is more
robust.
Link: https://lore.kernel.org/r/20200618144051.7415-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Set LED_CORE_SUSPENDRESUME to LED cdev flags, so that the LED core
would store and restore the LED status at suspend/resume.
In theory, the codec driver should be responsible for all LED bits,
but this might be safer and cover the overlooked cases.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-14-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Both mute and mic-mute LED callbacks do almost similar tasks with just
different bits. Factor out the common code and use them from the
callbacks for simplification.
This ended up with covering the forgotten stuff, too; e.g. VREF LED
handling required the temporary power up/down that was missing for the
mute LED, or some forgotten polarity checks are added.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-13-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Convert the mute LED handling in Sigmatel/IDT codec to the new vmaster
mute helper. A point to be cautiously handled is that the value
passed to the callback is inverted; the vmaster passes "enabled"
(0 = mute), while LED classdev passes "brightness" (1 = mute).
A positive side-effect by this change is that the driver gets also the
enum controls for the mute behavior like other drivers already had.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-12-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Convert the mute LED handling in Conexant codec to the new vmaster
mute helper. A point to be cautiously handled is that the value
passed to the callback is inverted; the vmaster passes "enabled"
(0 = mute), while LED classdev passes "brightness" (1 = mute).
Also the assignment of the default vmaster hook is moved at a later
point after the mute hook is set up. This assures no nested hook.
Finally, since we enable the mute-LED kcontrols always in the helper
side, the extra vmaster_mute_enum flag set up is dropped.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-11-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Convert the mute LED handling in Realtek codec to the new vmaster mute
helper. A point to be cautiously handled is that the value passed to
the callback is inverted; the vmaster passes "enabled" (0 = mute),
while LED classdev passes "brightness" (1 = mute).
The code in Thinkpad helper is also converted. In that case, just
call the new function and remove the open-code.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-10-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Like mic-mute LED handling, add a new helper to deal with the master
mute LED with LED classdev. Unlike the mic-mute case, the playback
master mute is hooked on vmaster, and we suppose no nested hooks
allowed there.
The classdev creation code is factored out to a common function that
is called from both mute and mic-mute LED helpers.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-9-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Now all users of the old snd_hda_gen_add_micmute_led() have been
converted to the new LED-classdev variant, and we can make it local,
and remove the unused hda_gen_spec.micmute_led.update callback field.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-8-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch converts the remaining user of
snd_hda_gen_add_micmute_led() in IDT/Sigmatel codec driver into the
new snd_hda_gen_add_micmute_led_cdev().
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-6-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch converts the remaining user of
snd_hda_gen_add_micmute_led() in Conexant codec driver into the new
snd_hda_gen_add_micmute_led_cdev().
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-5-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch converts the remaining user of
snd_hda_gen_add_micmute_led() in Realtek codec driver into the new
snd_hda_gen_add_micmute_led_cdev().
The Thinkpad helper code is updated accordingly, too.
Also, the usage of snd_hda_gen_fixup_micmute_led() is replaced with
either the local alc_fixup_micmute_led() or the explicit call of
snd_hda_gen_add_micmute_led_cdev() with NULL callback.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-4-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
A new helper, snd_hda_gen_add_micmute_led_cdev(), is introduced here
for creating a LED classdev and setting up the hook to the capture
control for controlling the mic-mute LED to follow the capture switch
change. This will replace the existing users of
snd_hda_gen_add_micmute_led() in later patches.
Also, introduce a new kconfig CONFIG_SND_HDA_GENERIC_LEDS, to indicate
the usage of mute / mic-mute LED helpers. It's selected by the codec
drivers (Realtek, Conexant and Sigmatel), while it selects the
necessary LED class dependencies.
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Link: https://lore.kernel.org/r/20200618110842.27238-3-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Mark Hills