After commit 0edb555a65 ("platform: Make platform_driver::remove()
return void") .remove() is (again) the right callback to implement for
platform drivers.
Convert all drivers below sound/soc to use .remove(), with the eventual
goal to drop struct platform_driver::remove_new(). As .remove() and
.remove_new() have the same prototypes, conversion is done by just
changing the structure member name in the driver initializer.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Link: https://patch.msgid.link/20240909151230.909818-2-u.kleine-koenig@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
With ARCH=x86, make allmodconfig && make W=1 C=1 reports:
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/renoir/snd-acp3x-rn.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/yc/snd-soc-acp6x-mach.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/acp/snd-acp-i2s.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/acp/snd-acp-pdm.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/acp/snd-acp-legacy-common.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/acp/snd-acp-pci.o
WARNING: modpost: missing MODULE_DESCRIPTION() in sound/soc/amd/ps/snd-soc-ps-mach.o
Add the missing invocations of the MODULE_DESCRIPTION() macro.
Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Link: https://lore.kernel.org/r/20240612-md-sound-soc-amd-v1-1-ad1de0409c11@quicinc.com
Signed-off-by: Mark Brown <broonie@kernel.org>
*-objs suffix is reserved rather for (user-space) host programs while
usually *-y suffix is used for kernel drivers (although *-objs works
for that purpose for now).
Let's correct the old usages of *-objs in Makefiles.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20240507155540.24815-5-tiwai@suse.de
Signed-off-by: Mark Brown <broonie@kernel.org>
To align with AMD SoundWire manager driver license, update license as
GPL-2.0-only for Pink Sardine ACP PCI driver and corresponding child
drivers.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://msgid.link/r/20240222102656.631144-1-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add machine select logic for SoundWire interface and create a machine
device node based on ACP PDM/SoundWire configuration.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://msgid.link/r/20240214104014.1144668-5-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Consider the below scenario, When ACP and SoundWire managers are in
D3 state and SoundWire manager power off mode is selected and acp and
SoundWire manager instances are in runtime suspended state.
In this case, for the ACP PME wake event, the ACP PCI driver should resume
SoundWire manager devices based on wake enable status set.
Add code for handling ACP PME wake event for runtime suspend scenario
when SoundWire power off mode is selected.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://msgid.link/r/20240214104014.1144668-4-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The earlier acp_reset flag is set to true in two instances as mentioned
below.
1. When active SoundWire manager instances power mode is set to
Power off mode when SoundWire configuration is selected.
2. For other acp configurations
As code being refactored and common function being used for scanning
SoundWire controller, acp_reset flag update logic is dropped.
Instead of it, check the SoundWire manager instance enable state, based on
it update sdw_en_stat flag which will be used to apply ACP init/de-init
sequence during suspend/resume callbacks based on flag set value when
SoundWire configuration is selected.
For other acp configurations, acp init/de-init will be called by default.
Refactor existing pm ops logic for SoundWire configuration and use
sdw_en_stat flag for invoking acp init/de-init sequence.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://msgid.link/r/20240214104014.1144668-3-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Refactor ACP child platform device creation code based on acp config.
Use common SoundWire manager functions for device probe and exit
sequences.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://msgid.link/r/20240214104014.1144668-2-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC is now unified asoc_xxx() into snd_soc_xxx().
This patch convert asoc_xxx() to snd_soc_xxx().
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/878r9cs25b.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The compile warning with -Wformat-truncation at
sdw_amd_scan_controller() is false-positive; the max loop size is
AMD_SDW_MAX_MANAGERS (= 2), hence it fits with the given size.
For suppressing the warning, replace snprintf() with scnprintf().
As stated in the above, truncation doesn't matter.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20230915082207.26200-1-tiwai@suse.de
Signed-off-by: Mark Brown <broonie@kernel.org>
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20230707072830.3395789-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
On Pink Sardine machines that don't have SdW controllers in use, the property
`mipi-sdw-manager-list` won't exist. There is no point in showing an error
to a user when this situation is encountered.
Furthermore if the machine doesn't have a DMIC connected to the ACP, there
may be no platform devices created either.
Downgrade the associated message to debug.
Fixes: d1351c30ac ("ASoC: amd: ps: create platform devices based on acp config")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20230708025208.54272-1-mario.limonciello@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
acp_get_byte_count() function should return zero bytes instead of
-EINVAL for invalid SoundWire manager instance.
Fixes: f722917350 ("ASoC: amd: ps: add SoundWire dma driver dma ops")
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230626105356.2580125-5-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
For AUDIO0 RX stream, AUDIO0_RX position registers should be used.
DMA error is reported due to referring wrong position register set for
AUDIO0 RX stream.
Correct the position register set for AUDIO0 RX stream.
Fixes: f722917350 ("ASoC: amd: ps: add SoundWire dma driver dma ops")
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230626105356.2580125-3-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Correct the DMA irq mask macro to program DMA irq bits correctly for
SDW0 instance rx streams.
Fixes: 298d4f7b17 ("ASoC: amd: ps: add support for SoundWire DMA interrupts")
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230626105356.2580125-2-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add comments for DMA stream id and IRQ bit mapping in
ACP_EXTERNAL_CNTL & ACP_EXTERNAL_CNTL1 registers for
SDW0 and SDW1 manager instances.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230626105356.2580125-1-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
pm_runtime_allow() is not needed for ACP child platform devices.
Replace pm_runtime_allow() with pm_runtime_mark_last_busy()
& pm_runtime_set_active() in pm_runtime enable sequence for
ACP child platform drivers.
Signed-off-by: Arun Gopal <arungopal.kondaveeti@amd.com>
Link: https://lore.kernel.org/r/Message-Id: <20230623214150.4058721-1-arungopal.kondaveeti@amd.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
AMD SoundWire manager supports different power modes.
acp_reset flag will be set to false only when SoundWire manager power
mode is selected as ClockStop Mode. For rest of the combinations
(ACP PDM + SDW), acp_reset flag will be set to true.
When acp_reset flag is set, acp de-init and acp init sequence should
be invoked during suspend/resume callbacks.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230612095903.2113464-10-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Move to request_threaded_irq and use thread for handling
SoundWire DMA interrupts.
Whenever audio data equal to the SoundWire FIFO watermark level
are produced/consumed, interrupt is generated.
Acknowledge the interrupt and wake up the irq thread.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230612095903.2113464-6-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
SoundWire DMA platform driver binds to the platform device created by
ACP PCI device. SoundWire DMA driver registers ALSA DMA component
with ASoC framework.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230612095903.2113464-4-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Handle SoundWire manager related interrupts in ACP PCI driver
interrupt handler and schedule SoundWire manager work queue for
further processing.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230612095903.2113464-3-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Based on ACP pin configuration and scanning child devices
under ACP pci device ACPI scope, platform device configuration
(pdev_config) and platform device count(pdev_count) will be
calculated.
Using pdev_config and pdev_count values, ACP PCI driver will
create platform devices for Pink Sardine platform.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230612095903.2113464-2-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Sending the mutex address(acp_lock) as platform
data during ACP PDM platform driver register sequence,
its creating copy of the platform data.
Referencing this platform data in ACP PDM driver results
incorrect reference to the common lock usage.
Instead of directly passing the lock address as platform
data, retrieve it from parent driver data structure
and use the same lock reference in ACP PDM driver.
Fixes: 45aa83cb93 ("ASoC: amd: ps: use acp_lock to protect common registers in pdm driver")
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230525113000.1290758-1-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of a busy waiting while loop using udelay in
acp63_power_on and acp63_reset functions use readl_poll_timeout
function to check the condition.
Signed-off-by: Syed Saba Kareem <Syed.SabaKareem@amd.com
Reviewed-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com
Link: https://lore.kernel.org/r/20230426122219.3745586-2-Syed.SabaKareem@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org
Instead of acp63_readl() and acp63_writel() wrappers
readl and writel functions can be used directly.
Remove acp63_readl() and acp63_writel() wrappers.
Signed-off-by: Syed Saba Kareem <Syed.SabaKareem@amd.com
Reviewed-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com
Link: https://lore.kernel.org/r/20230426122219.3745586-1-Syed.SabaKareem@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org
The most recent changes to ASoC, such as new module parameters, date to the
year 2023. Update copyright statement accordingly.
Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com
Link: https://lore.kernel.org/r/20230420180212.3101178-1-carlos.bilbao@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org
We have SOF and generic ACP support enabled for Rembrandt and
pheonix platforms on some machines. Since we have same PCI id
used for probing, add check for machine configuration flag to
avoid conflict with newer pci drivers. Such machine flag has
been initialized via dmi match on few Chrome machines. If no
flag is specified probe and register older platform device.
Signed-off-by: Syed Saba Kareem <Syed.SabaKareem@amd.com>
Reviewed-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230412091638.1158901-1-Syed.SabaKareem@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Takashi Iwai <tiwai@suse.de>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20230315150745.67084-19-u.kleine-koenig@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
In case of regressions for any users that the new pdm_gain value is
too high and for additional debugging, introduce a module parameter
that would let them configure it.
This parameter should be removed in the future:
* If it's determined that the parameter is not needed, just hardcode
the correct value as before
* If users do end up using it to debug and report different values
we should introduce a config knob that can have policy set by ucm.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Link: https://lore.kernel.org/r/20230131184653.10216-7-mario.limonciello@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
No issues have been reported yet for DMIC audio level on ps platforms,
but as problems were found both on YC (Rembrandt) and Renoir based
designs it's very likely they happen on ps too.
Increase the PDM gain to solve this problem.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Link: https://lore.kernel.org/r/20230131184653.10216-6-mario.limonciello@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Clang warns:
sound/soc/amd/ps/pci-ps.c:218:2: error: variable 'ret' is used uninitialized whenever switch default is taken [-Werror,-Wsometimes-uninitialized]
default:
^~~~~~~
sound/soc/amd/ps/pci-ps.c:239:9: note: uninitialized use occurs here
return ret;
^~~
sound/soc/amd/ps/pci-ps.c:190:9: note: initialize the variable 'ret' to silence this warning
int ret;
^
= 0
1 error generated.
ret is used uninitialized if 'goto de_init' is taken. As this is not an
error nor should the ACP be deinitialized, just directly return 0 in
this case statement, which resolves the warning.
Fixes: 1d325cdaf7 ("ASoC: amd: ps: refactor platform device creation logic")
Link: https://github.com/ClangBuiltLinux/linux/issues/1779
Suggested-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Syed Saba Kareem <syed.sabakareem@amd.com>
Link: https://lore.kernel.org/r/20230105-wsometimes-uninitialized-pci-ps-c-v2-1-c50321676325@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Sparse warning:
error: symbol 'acp63_fill_platform_dev_info' was not declared. Should
it be static?
Also reduce line lines below 100 characters.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20230104145708.25051-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Retrieve acp_lock mutex as platform data and use it for protecting
ACP common registers access in acp pdm driver.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20230104055435.321327-2-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of using hard coded index value for platform device index,
retrieve the device index based on platform devices created.
In PDM config case, ACP PCI driver retrieves dev index from
pdm_dev_index variable.
This will avoid overhead when multiple endpoint combinations
support is added later.
platform device index will vary based on audio endpoint
configuration.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20221221172855.1618766-3-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Refactor platform device creation implementation.
Based on platform dev count and pdev mask create platform devices.
Use common API to fill platform dev information.
Use pdev count variable in remove callback for unregistering
platform devices.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20221221172855.1618766-2-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Implement API to retrieve acp device config and calculate platform
device count and dev mask for platform device node creation.
Currently for DMIC configuration, mask and dev count are calculated.
Same api will be used to extend support for different ACP device
configurations.
Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Link: https://lore.kernel.org/r/20221221172855.1618766-1-Vijendar.Mukunda@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
