kernel/linux.git/sound/soc, branch v6.12.81

ASoC: Intel: boards: fix unmet dependency on PINCTRL

2026-04-11T12:24:33+00:00

[ Upstream commit e920c36f2073d533bdf19ba6ab690432c8173b63 ] This reverts commit c073f0757663 ("ASoC: Intel: sof_sdw: select PINCTRL_CS42L43 and SPI_CS42L43") Currently, SND_SOC_INTEL_SOUNDWIRE_SOF_MACH selects PINCTRL_CS42L43 without also selecting or depending on PINCTRL, despite PINCTRL_CS42L43 depending on PINCTRL. See the following Kbuild warning: WARNING: unmet direct dependencies detected for PINCTRL_CS42L43 Depends on [n]: PINCTRL [=n] && MFD_CS42L43 [=m] Selected by [m]: - SND_SOC_INTEL_SOUNDWIRE_SOF_MACH [=m] && SOUND [=y] && SND [=m] && SND_SOC [=m] && SND_SOC_INTEL_MACH [=y] && (SND_SOC_SOF_INTEL_COMMON [=m] || !SND_SOC_SOF_INTEL_COMMON [=m]) && SND_SOC_SOF_INTEL_SOUNDWIRE [=m] && I2C [=y] && SPI_MASTER [=y] && ACPI [=y] && (MFD_INTEL_LPSS [=n] || COMPILE_TEST [=y]) && (SND_SOC_INTEL_USER_FRIENDLY_LONG_NAMES [=n] || COMPILE_TEST [=y]) && SOUNDWIRE [=m] In response to v1 of this patch [1], Arnd pointed out that there is no compile-time dependency sof_sdw and the PINCTRL_CS42L43 driver. After testing, I can confirm that the kernel compiled with SND_SOC_INTEL_SOUNDWIRE_SOF_MACH enabled and PINCTRL_CS42L43 disabled. This unmet dependency was detected by kconfirm, a static analysis tool for Kconfig. Link: https://lore.kernel.org/all/b8aecc71-1fed-4f52-9f6c-263fbe56d493@app.fastmail.com/ [1] Fixes: c073f0757663 ("ASoC: Intel: sof_sdw: select PINCTRL_CS42L43 and SPI_CS42L43") Signed-off-by: Julian Braha Acked-by: Arnd Bergmann Link: https://patch.msgid.link/20260325001522.1727678-1-julianbraha@gmail.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: ep93xx: Fix unchecked clk_prepare_enable() and add rollback on failure

2026-04-11T12:24:32+00:00

[ Upstream commit 622363757b2286dd2c2984b0d80255cbb35a0495 ] ep93xx_i2s_enable() calls clk_prepare_enable() on three clocks in sequence (mclk, sclk, lrclk) without checking the return value of any of them. If an intermediate enable fails, the clocks that were already enabled are never rolled back, leaking them until the next disable cycle — which may never come if the stream never started cleanly. Change ep93xx_i2s_enable() from void to int. Add error checking after each clk_prepare_enable() call and unwind already-enabled clocks on failure. Propagate the error through ep93xx_i2s_startup() and ep93xx_i2s_resume(), both of which already return int. Signed-off-by: Jihed Chaibi Fixes: f4ff6b56bc8a ("ASoC: cirrus: i2s: Prepare clock before using it") Link: https://patch.msgid.link/20260324210909.45494-1-jihed.chaibi.dev@gmail.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: ak4458: Convert to RUNTIME_PM_OPS() & co

2026-04-02T11:09:52+00:00

commit 9f9c8e9064ea8ceb13540a283f08550c097bb673 upstream. Use the newer RUNTIME_PM_OPS() and SYSTEM_SLEEP_PM_OPS() macros instead of SET_RUNTIME_PM_OPS() and SET_SYSTEM_SLEEP_PM_OPS() together with pm_ptr(), which allows us dropping ugly __maybe_unused attributes and CONFIG_PM ifdefs. This optimizes slightly when CONFIG_PM is disabled, too. Signed-off-by: Takashi Iwai Reviewed-by: Charles Keepax Link: https://patch.msgid.link/20250317095603.20073-4-tiwai@suse.de Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman

ASoC: SOF: ipc4-topology: Allow bytes controls without initial payload

2026-04-02T11:09:42+00:00

commit d40a198e2b7821197c5c77b89d0130cc90f400f5 upstream. It is unexpected, but allowed to have no initial payload for a bytes control and the code is prepared to handle this case, but the size check missed this corner case. Update the check for minimal size to allow the initial size to be 0. Cc: stable@vger.kernel.org Fixes: a653820700b8 ("ASoC: SOF: ipc4-topology: Correct the allocation size for bytes controls") Signed-off-by: Peter Ujfalusi Reviewed-by: Bard Liao Reviewed-by: Liam Girdwood Reviewed-by: Seppo Ingalsuo Reviewed-by: Kai Vehmanen Link: https://patch.msgid.link/20260326075618.1603-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman

ASoC: adau1372: Fix clock leak on PLL lock failure

2026-04-02T11:09:41+00:00

[ Upstream commit bfe6a264effcb6fe99ad7ceaf9e8c7439fc9555b ] adau1372_enable_pll() was a void function that logged a dev_err() on PLL lock timeout but did not propagate the error. As a result, adau1372_set_power() would continue with adau1372->enabled set to true despite the PLL being unlocked, and the mclk left enabled with no corresponding disable on the error path. Convert adau1372_enable_pll() to return int, using -ETIMEDOUT on lock timeout and propagating regmap errors directly. In adau1372_set_power(), check the return value and unwind in reverse order: restore regcache to cache-only mode, reassert GPIO power-down, and disable the clock before returning the error. Signed-off-by: Jihed Chaibi Fixes: 6cd4c6459e47 ("ASoC: Add ADAU1372 audio CODEC support") Reviewed-by: Nuno Sá Link: https://patch.msgid.link/20260325210704.76847-3-jihed.chaibi.dev@gmail.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: adau1372: Fix unchecked clk_prepare_enable() return value

2026-04-02T11:09:40+00:00

[ Upstream commit 326fe8104a4020d30080d37ac8b6b43893cdebca ] adau1372_set_power() calls clk_prepare_enable() but discards the return value. If the clock enable fails, the driver proceeds to access registers on unpowered hardware, potentially causing silent corruption. Make adau1372_set_power() return int and propagate the error from clk_prepare_enable(). Update adau1372_set_bias_level() to return the error directly for the STANDBY and OFF cases. Signed-off-by: Jihed Chaibi Fixes: 6cd4c6459e47 ("ASoC: Add ADAU1372 audio CODEC support") Reviewed-by: Nuno Sá Link: https://patch.msgid.link/20260325210704.76847-2-jihed.chaibi.dev@gmail.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: Intel: catpt: Fix the device initialization

2026-04-02T11:09:39+00:00

[ Upstream commit 5a184f1cb43a8e035251c635f5c47da5dc3e3049 ] The DMA mask shall be coerced before any buffer allocations for the device are done. At the same time explain why DMA mask of 31 bits is used in the first place. Cc: Andy Shevchenko Fixes: 7a10b66a5df9 ("ASoC: Intel: catpt: Device driver lifecycle") Signed-off-by: Cezary Rojewski Reviewed-by: Andy Shevchenko Link: https://patch.msgid.link/20260320101217.1243688-1-cezary.rojewski@intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: fsl_easrc: Fix event generation in fsl_easrc_iec958_put_bits()

2026-04-02T11:09:28+00:00

[ Upstream commit 54a86cf48eaa6d1ab5130d756b718775e81e1748 ] ALSA controls should return 1 if the value in the control changed but the control put operation fsl_easrc_iec958_put_bits() unconditionally returns 0, causing ALSA to not generate any change events. This is detected by mixer-test with large numbers of messages in the form: No event generated for Context 3 IEC958 CS5 Context 3 IEC958 CS5.0 orig 5224 read 5225, is_volatile 0 Add a suitable check. Signed-off-by: Mark Brown Link: https://patch.msgid.link/20260205-asoc-fsl-easrc-fix-events-v1-1-39d4c766918b@kernel.org Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: fsl_easrc: Fix event generation in fsl_easrc_iec958_set_reg()

2026-04-02T11:09:28+00:00

[ Upstream commit 31ddc62c1cd92e51b9db61d7954b85ae2ec224da ] ALSA controls should return 1 if the value in the control changed but the control put operation fsl_easrc_set_reg() only returns 0 or a negative error code, causing ALSA to not generate any change events. Add a suitable check by using regmap_update_bits_check() with the underlying regmap, this is more clearly and simply correct than trying to verify that one of the generic ops is exactly equivalent to this one. Signed-off-by: Mark Brown Link: https://patch.msgid.link/20260205-asoc-fsl-easrc-fix-events-v1-2-39d4c766918b@kernel.org Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

driver core: platform: use generic driver_override infrastructure

2026-04-02T11:09:25+00:00

[ Upstream commit 2b38efc05bf7a8568ec74bfffea0f5cfa62bc01d ] When a driver is probed through __driver_attach(), the bus' match() callback is called without the device lock held, thus accessing the driver_override field without a lock, which can cause a UAF. Fix this by using the driver-core driver_override infrastructure taking care of proper locking internally. Note that calling match() from __driver_attach() without the device lock held is intentional. [1] Link: https://lore.kernel.org/driver-core/DGRGTIRHA62X.3RY09D9SOK77P@kernel.org/ [1] Reported-by: Gui-Dong Han Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220789 Fixes: 3d713e0e382e ("driver core: platform: add device binding path 'driver_override'") Reviewed-by: Greg Kroah-Hartman Link: https://patch.msgid.link/20260303115720.48783-5-dakr@kernel.org Signed-off-by: Danilo Krummrich Signed-off-by: Sasha Levin