DIV_ROUND_CLOSEST() does not work on 64-bit variables when building for
a 32-bit target:
ld.lld: error: undefined symbol: __udivdi3
>>> referenced by loongson_i2s.c
>>> sound/soc/loongson/loongson_i2s.o:(loongson_i2s_hw_params) in archive vmlinux.a
Use DIV_ROUND_CLOSEST_ULL() instead.
Fixes: d24028606e ("ASoC: loongson: Add Loongson ASoC Sound Card Support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230616090156.2347850-4-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The new driver fails to build when PCI is disabled:
WARNING: unmet direct dependencies detected for SND_SOC_LOONGSON_I2S_PCI
Depends on [n]: SOUND [=y] && !UML && SND [=y] && SND_SOC [=y] && (LOONGARCH || COMPILE_TEST [=y]) && PCI [=n]
Selected by [y]:
- SND_SOC_LOONGSON_CARD [=y] && SOUND [=y] && !UML && SND [=y] && SND_SOC [=y] && (LOONGARCH || COMPILE_TEST [=y])
sound/soc/loongson/loongson_i2s_pci.c:167:1: error: type specifier missing, defaults to 'int'; ISO C99 and later do not support implicit int [-Wimplicit-int]
module_pci_driver(loongson_i2s_driver);
Add the appropriate Kconfig dependency.
Fixes: d24028606e ("ASoC: loongson: Add Loongson ASoC Sound Card Support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230616090156.2347850-3-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Build testing without CONFIG_PM_SLEEP causes a warning:
sound/soc/loongson/loongson_i2s.c:246:12: error: unused function 'i2s_suspend' [-Werror,-Wunused-function]
sound/soc/loongson/loongson_i2s.c:255:12: error: unused function 'i2s_resume' [-Werror,-Wunused-function]
Use the modern SYSTEM_SLEEP_PM_OPS() instead of the old one to avoid this.
Fixes: d24028606e ("ASoC: loongson: Add Loongson ASoC Sound Card Support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230616090156.2347850-2-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The PM functions are never referenced when CONFIG_PM_SLEEP is
disabled:
sound/soc/codecs/max98388.c:854:12: error: unused function 'max98388_suspend' [-Werror,-Wunused-function]
static int max98388_suspend(struct device *dev)
^
sound/soc/codecs/max98388.c:864:12: error: unused function 'max98388_resume' [-Werror,-Wunused-function]
static int max98388_resume(struct device *dev)
Fix this by using the modern SYSTEM_SLEEP_PM_OPS() macro in place of
the deprecated SET_SYSTEM_SLEEP_PM_OPS() version, and use pm_sleep_ptr()
to hide the entire structure as well.
On a related note, the of_match_ptr() and ACPI_PTR() macros have the same
problem and would cause the device id table to be unused when the driver
is built-in and the respective subsystems are disabled. This does not
cause warnings unless -Wunused-const-variable is passed to the compiler,
but it's better to just not use the macros at all here.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230616090156.2347850-1-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Mark Brown <broonie@kernel.org>:
The RealTek SoundWire CODECs don't support bulk register operations so
they gain no advantage from using the rbtree cache, convert them to use
the maple tree cache instead - it is more modern.
Merge series from Bard Liao <yung-chuan.liao@linux.intel.com>:
Currently, ASoC supports dailinks with the following mappings:
1 cpu DAI to N codec DAIs
N cpu DAIs to N codec DAIs
But the mapping between N cpu DAIs and M codec DAIs is not supported.
The reason is that we didn't have a mechanism to map cpu and codec DAIs
This series suggests a new snd_soc_dai_link_codec_ch_map struct in
struct snd_soc_dai_link{} which provides codec DAI to cpu DAI mapping
information used to implement N cpu DAIs to M codec DAIs support.
And add the codec_ch_maps to SOF SoundWire machine driver.
Merge series from Mark Brown <broonie@kernel.org>:
A lot of the Cirrus Logic devices only support single register
read/write operations so they get no benefit from using the rbtree cache
over the more modern maple tree cache, convert them to use maple tree.
Current ASoC is assuming that trigger starting order is
Link -> Component -> DAI as default, and its reverse order for stopping.
But some Driver / Card want to reorder it for some reasons.
We have such flags, but is unbalance like below.
struct snd_soc_component_driver :: start_dma_last
struct snd_soc_dai_link :: stop_dma_first
We want to have more flexible, and more generic method.
This patch adds new snd_soc_trigger_order for start/stop at
component / DAI-link.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87r0qmfnzx.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently, ASoC supports dailinks with the following mappings:
1 cpu DAI to N codec DAIs
N cpu DAIs to N codec DAIs
But the mapping between N cpu DAIs and M codec DAIs is not supported.
The reason is that we didn't have a mechanism to map cpu and codec DAIs
This patch suggests a new snd_soc_dai_link_codec_ch_map struct in
struct snd_soc_dai_link{} which provides codec DAI to cpu DAI mapping
information used to implement N cpu DAIs to M codec DAIs
support.
When a dailink contains two or more cpu DAIs, we should set channel
number of cpus based on its channel mask. The new struct also provides
channel mask information for each codec and we can construct the cpu
channel mask by combining all codec channel masks which map to the cpu.
The N:M mapping is however restricted to the N <= M case due to physical
restrictions on a time-multiplexed bus such as I2S/TDM, AC97, SoundWire
and HDaudio.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20230607031242.1032060-2-yung-chuan.liao@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Commit 3ed2b549b3 ("ALSA: pcm: fix wait_time calculations") corrected
the PCM wait_time calculations and in doing so reduced the calculated
wait_time. This exposed an issue with the Tegra Master Volume Control
(MVC) device where the reduced wait_time caused the MVC to fail. For now
fix this by setting the default wait_time for Tegra to be 500ms.
Fixes: 3ed2b549b3 ("ALSA: pcm: fix wait_time calculations")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20230613093453.13927-1-jonathanh@nvidia.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Brent Lu <brent.lu@intel.com>:
This patch series remove redundant dapm routes declared in multiple machine
drivers. These routes will be created by snd_soc_dapm_connect_dai_link_widgets()
during soundcard initialization. Following is the kernel log from a KBL
chromebook nocturne.
dmic:
[ 13.773455] avs_dmic avs_dmic: connected DAI link dmic-codec:Capture -> dmic-platform:DMIC Rx
[ 13.773460] avs_dmic avs_dmic: connected DAI link dmic-codec:Capture -> dmic-platform:DMIC WoV Rx
max98373:
[ 14.079536] avs_max98373 avs_max98373.1: connected DAI link avs_max98373.1-platform:ssp0 Tx -> i2c-MX98373:00:Right HiFi Playback
[ 14.079545] avs_max98373 avs_max98373.1: connected DAI link i2c-MX98373:00:Right HiFi Capture -> avs_max98373.1-platform:ssp0 Rx
[ 14.079550] avs_max98373 avs_max98373.1: connected DAI link avs_max98373.1-platform:ssp0 Tx -> i2c-MX98373:01:Left HiFi Playback
[ 14.079554] avs_max98373 avs_max98373.1: connected DAI link i2c-MX98373:01:Left HiFi Capture -> avs_max98373.1-platform:ssp0 Rx
hdaudio:
[ 14.094818] avs_hdaudio avs_hdaudio.2: connected DAI link hdaudioB0D2-platform:hdaudioB0D2-cpu0 Tx -> hdaudioB0D2:HDMI 0 Playback
[ 14.094824] avs_hdaudio avs_hdaudio.2: connected DAI link hdaudioB0D2-platform:hdaudioB0D2-cpu1 Tx -> hdaudioB0D2:HDMI 1 Playback
[ 14.094828] avs_hdaudio avs_hdaudio.2: connected DAI link hdaudioB0D2-platform:hdaudioB0D2-cpu2 Tx -> hdaudioB0D2:HDMI 2 Playback
