kernel.git/sound/soc/fsl, branch master

ASoC: fsl_easrc: Fix event generation in fsl_easrc_iec958_set_reg()

2026-03-01T23:48:02Z

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

ASoC: fsl_easrc: Fix event generation in fsl_easrc_iec958_put_bits()

2026-03-01T23:48:01Z

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

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51Z

This was done entirely with mindless brute force, using git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28Z

This is the result of running the Coccinelle script from scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to avoid scalar types (which need careful case-by-case checking), and instead replace kmalloc-family calls that allocate struct or union object instances: Single allocations: kmalloc(sizeof(TYPE), ...) are replaced with: kmalloc_obj(TYPE, ...) Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...) are replaced with: kmalloc_objs(TYPE, COUNT, ...) Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...) are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...) (where TYPE may also be *VAR) The resulting allocations no longer return "void *", instead returning "TYPE *". Signed-off-by: Kees Cook

ASoC: fsl_xcvr: Revert fix missing lock in fsl_xcvr_mode_put()

2026-02-10T19:24:04Z

This reverts commit f51424872760 ("ASoC: fsl_xcvr: fix missing lock in fsl_xcvr_mode_put()"). The original patch attempted to acquire the card->controls_rwsem lock in fsl_xcvr_mode_put(). However, this function is called from the upper ALSA core function snd_ctl_elem_write(), which already holds the write lock on controls_rwsem for the whole put operation. So there is no need to simply hold the lock for fsl_xcvr_activate_ctl() again. Acquiring the read lock while holding the write lock in the same thread results in a deadlock and a hung task, as reported by Alexander Stein. Fixes: f51424872760 ("ASoC: fsl_xcvr: fix missing lock in fsl_xcvr_mode_put()") Reported-by: Alexander Stein Closes: https://lore.kernel.org/linux-sound/5056506.GXAFRqVoOG@steina-w/ Signed-off-by: Ziyi Guo Link: https://patch.msgid.link/20260210185714.556385-1-n7l8m4@u.northwestern.edu Signed-off-by: Mark Brown

ASoC: Merge up release

2026-02-10T19:23:02Z

We need to fix some commits that went into the final release.

ASoC: fsl_asrc_dma: allocate memory from dma device

2026-02-06T17:22:37Z

The dma-coherent property is used to indicate a device is capable of coherent DMA operations. On i.MX952, one of EDMA devices support such feature, in order to support the EDMA device, the memory needs to be allocated from the DMA device. Make this driver to support both non dma-coherent and dma-coherent dma engine. Remove dma coerce_mask_and coherent() because DMA provider already set it according to its capability. Signed-off-by: Shengjiu Wang Reviewed-by: Frank Li Link: https://patch.msgid.link/20260206014805.3897764-5-shengjiu.wang@nxp.com Signed-off-by: Mark Brown

ASoC: fsl_asrc: Add support for i.MX952 platform

2026-02-06T17:22:36Z

Add a compatible string, clock mapping table and enable the option 'start_before_dma' to support ASRC on the i.MX952 platform. The clock mapping table is to map the clock sources on i.MX952 to the clock ids in the driver, the clock ids are for all the clock sources on all supported platforms. Signed-off-by: Shengjiu Wang Reviewed-by: Daniel Baluta Reviewed-by: Frank Li Link: https://patch.msgid.link/20260206014805.3897764-4-shengjiu.wang@nxp.com Signed-off-by: Mark Brown

ASoC: fsl_asrc_m2m: Add option to start ASRC before DMA device for M2M

2026-02-06T17:22:35Z

There is a limitation on i.MX952 that dma request is not cleared at the end of conversion with dma slave mode. Which causes sample is dropped from the input fifo on the second time if dma is triggered before the client device and EDMA may copy wrong data from output fifo as the output fifo is not ready in the beginning. The solution is to trigger asrc before dma on i.MX952, and add delay to wait output data is generated then start the EDMA for output, otherwise the m2m function has noise issues. So add an option to start ASRC first for M2M before ASRC is enabled on i.MX952. Signed-off-by: Shengjiu Wang Reviewed-by: Daniel Baluta Reviewed-by: Frank Li Link: https://patch.msgid.link/20260206014805.3897764-3-shengjiu.wang@nxp.com Signed-off-by: Mark Brown

ASoC: fsl: imx-rpmsg: use snd_soc_find_dai_with_mutex() in probe

2026-02-05T17:29:00Z

imx_rpmsg_probe() calls snd_soc_find_dai() without holding client_mutex. However, snd_soc_find_dai() has lockdep_assert_held(&client_mutex) indicating callers must hold this lock, as the function iterates over the global component list. All other callers of snd_soc_find_dai() either hold client_mutex via the snd_soc_bind_card() path or use the snd_soc_find_dai_with_mutex() wrapper. Use snd_soc_find_dai_with_mutex() instead to fix the missing lock protection. Signed-off-by: Ziyi Guo Reviewed-by: Frank Li Link: https://patch.msgid.link/20260205052429.4046903-1-n7l8m4@u.northwestern.edu Signed-off-by: Mark Brown