Peter Ujfalusi [Tue, 5 Apr 2022 17:26:58 +0000 (10:26 -0700)]
ASoC: SOF: ipc3: Implement the set_get_data IPC ops
Add the implementation for the set_get_data callback for handling large
data set and get.
The set_get_data() in IPC3 can be used only for component messages. The
function expects the caller to prepare the message behind the data pointer
for sending/receiving data. The callback only implements the needed code
to be able to split up a message if needed for transfer.
The set_get_data ops is based on the existing
snd_sof_ipc_set_get_comp_data() and sof_set_get_large_ctrl_data() but made
it generic entry point.
Peter Ujfalusi [Tue, 5 Apr 2022 17:26:55 +0000 (10:26 -0700)]
ASoC: SOF: Add high level IPC IO callback definitions to ipc_ops
Add tx_msg(), rx_msg(), set_get_data() and get_reply() ops, which can
be used as a generic API for sending, receiving single messages and
to write and read large data.
Peter Ujfalusi [Tue, 5 Apr 2022 17:26:54 +0000 (10:26 -0700)]
ASoC: SOF: Add helper function to prepare and send an IPC message
The new sof_ipc_send_msg() can be used by IPC dependent code to prepare
the ipc->msg for a new message transmission and then call in to platform
code to send the message.
Higher level code should be handling the completion and reply.
Add support for using the rt5682s codec together with rt1015p on
mt8192-mt6359 machines. All configurations are shared with the rt5682
codec variant, so simply select the SND_SOC_RT5682S config to ensure the
codec is present and set the correct card name. The codec will be linked
to by pointing to it in the headset-codec property in the devicetree.
While at it, also create macros for the names of the different codec
variants supported by this driver, as well as rename occurrences of
rt1015p_rt5682 to rt1015p_rt5682x, since they are shared between rt5682
and rt5682s.
ASoC: mediatek: mt8192: refactor for I2S8/I2S9 DAI links of headset
As part of the refactoring to allow the same machine driver to be used for
the rt1015(p) and rt5682(s) codecs on the MT8192 platform, parse the
rt5682(s) codec from the headset-codec property in the devicetree and wire
it to the I2S8 and I2S9 backends.
ASoC: mediatek: mt8192: refactor for I2S3 DAI link of speaker
As part of the refactoring to allow the same machine driver to be used for
the rt1015(p) and rt5682(s) codecs on the MT8192 platform, parse the
rt1015(p) codecs from the speaker-codecs property in the devicetree and
wire them to the I2S3 backend, instead of hardcoding the links and
selecting through the compatible.
ASoC: dt-bindings: mt8192-mt6359: add new compatible and new properties
1. Adds new compatible string "mt8192_mt6359_rt1015p_rt5682s" for machines
with rt1015p and rt5682s.
2. Adds new property "headset-codec" for getting headset codec.
3. Adds new property "speaker-codecs" for getting speaker codecs.
Huawei Matebook D15 uses two different GPIOs are used to control the output:
- gpio0 controls the speaker output;
- gpio1 controls the headphone output.
Changing both at the same time cause spurious events that are mis-interpreted
as input events, causing troubles on apps. So, a delay is needed before turning
on such gpios.
Also, the headset microphone is connected to MIC1, instead of MIC2 port.
With this patch, plugging a headphone causes a jack event to trigger the speaker
supply, powering down the speaker and powering up the headphone output.
Removing the headphone also triggers the power supply, powering up the speaker
and powering down the headphone.
ASoC: Intel: sof_es8336: Add a quirk for Huawei Matebook D15
Based on experimental tests, Huawei Matebook D15 actually uses
both gpio0 and gpio1: the first one controls the speaker, while
the other one controls the headphone.
Also, the headset is mapped as MIC1, instead of MIC2.
Add mt8195 shutdown callback function implementation to ensure that
shutdown flow is called to shutdown dsp core, disable adsp clock, and
power off dsp sram correctly during reboot flow.
ASoC: codecs: wcd934x: do not switch off SIDO Buck when codec is in use
SIDO(Single-Inductor Dual-Ouput) Buck powers up both analog and digital
circuits along with internal memory, powering off this is the last thing
that codec should do when going to very low power.
Current code was powering off this Buck if there are no users of sysclk,
which is not correct. Powering off this buck will result in no register access.
This code path was never tested until recently after adding pm support
in SoundWire controller. Fix this by removing the buck poweroff when the
codec is active and also the code that is not used.
Without this patch all the read/write transactions will never complete and
results in SLIMBus Errors like:
qcom,slim-ngd qcom,slim-ngd.1: Tx:MT:0x0, MC:0x60, LA:0xcf failed:-110
wcd934x-codec wcd934x-codec.1.auto: ASoC: error at soc_component_read_no_lock
on wcd934x-codec.1.auto for register: [0x00000d05] -110
qcom,slim-ngd-ctrl 171c0000.slim: Error Interrupt received 0x82000000
ASoC: SOF: topology: cleanup dailinks on widget unload
We set the cpu_dai capture_ or playback_widget on widget_ready but
never clear them, which leads to failures when unloading/reloading a
topology in modprobe/rmmod tests
ASoC: rt711/5682: check if bus is active before deferred jack detection
This patch takes a defensive programming and paranoid approach in case
the parent device (SoundWire) is pm_runtime resumed but the rt711
device is not. In that case, during the attachment and initialization,
a jack detection workqueue can be scheduled. Since the pm_runtime
suspend routines will not be invoked, the sequence to cancel all
deferred work is not executed, and the jack detection could happen
after the bus stops operating, leading to a timeout.
This patch applies the same solution to rt5682, based on the
similarities between codec drivers. The race condition with rt5682 was
not detected experimentally though.
Stephen Kitt [Tue, 5 Apr 2022 17:39:52 +0000 (19:39 +0200)]
ASoC: wm8731: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 16:40:31 +0000 (18:40 +0200)]
ASoC: sgtl5000: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 17:11:58 +0000 (19:11 +0200)]
ASoC: rt1016: enable building
The rt1016 driver wasn't listed in Kconfig and Makefile, which AFAICT
means it isn't ever built. This patch adds the appropriate entries to
Kconfig and Makefile.
Stephen Kitt [Tue, 5 Apr 2022 16:36:18 +0000 (18:36 +0200)]
ASoC: pcm3060: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 13:03:26 +0000 (15:03 +0200)]
ASoC: rt*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 16:34:00 +0000 (18:34 +0200)]
ASoC: tscs*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 12:24:11 +0000 (14:24 +0200)]
ASoC: wm*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
wm8731.c is excluded and will be submitted separately.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 16:35:33 +0000 (18:35 +0200)]
ASoC: lm49xxx: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Tue, 5 Apr 2022 16:37:50 +0000 (18:37 +0200)]
ASoC: mt6660: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Renesas is the only user of ak4613 driver on upstream for now.
It has STEREO/TDM512/TDM256/TDM128 mode, but STEREO only is used,
because of Renesas board connection.
I noticed that I can test first 2ch out of TDM256 mode 8ch
Playback even in such a situation.
Ajye Huang [Thu, 24 Mar 2022 08:47:08 +0000 (16:47 +0800)]
ASoC: Intel: sof_rt5682: Add support for max98360a speaker amp on SSP2
Follow Intel's design to replace max98360a amp SSP2 reather than SSP1
by judging DMI_OEM_STRING in sof_rt5682_quirk_table struct.
And reusing max98357's topology since DAI setting could be leveraged.
Ahmad Fatoum [Tue, 5 Apr 2022 15:57:31 +0000 (17:57 +0200)]
ASoC: fsl_sai: fix 1:1 bclk:mclk ratio support
Refactoring in commit a50b7926d015 ("ASoC: fsl_sai: implement 1:1
bclk:mclk ratio support") led to the bypass never happening
as (ratio = 1) was caught in the existing if (ratio & 1) continue;
check. The correct check sequence instead is:
- skip all ratios lower than one and higher than 512
- skip all odd ratios except for 1:1
- skip 1:1 ratio if and only if !support_1_1_ratio
And for all others, calculate the appropriate divider. Adjust the
code to facilitate this.
On a S905W-based system I get the following error:
debugfs: Directory 'c1105400.audio-controller' with parent 'P230-Q200' already present!
Turned out that multiple components having the same name triggers this
error in soc_init_component_debugfs(). The proposed solution allows
other drivers to adopt the same approach with minimal effort.
With the patch the error is gone and that's the debugfs entries.
On a S905W-based system I get the following error:
debugfs: Directory 'c1105400.audio-controller' with parent 'P230-Q200' already present!
Turned out that multiple components having the same name triggers this
error in soc_init_component_debugfs(). With the patch the error is gone
and that's the debugfs entries.
Heiner Kallweit [Wed, 9 Mar 2022 20:21:55 +0000 (21:21 +0100)]
ASoC: soc-core: add debugfs_prefix member to snd_soc_component_driver
Allow the component debugfs_prefix to be set from
snd_soc_component_driver. First use case is avoiding a duplicate
debugfs entry error in case a device has multiple components
which have the same name therefore.
Note that we don't set component->debugfs_prefix if it's set already.
That's needed because partially component->debugfs_prefix is set
before calling snd_soc_component_initialize().
ASoC: dt-bindings: lpass-cpu: Update clocks and power domain names for sc7280 platform
Update required clock-names used for MI2S primary path, VA macro's
codec memory path and HDMI path in sc7280 based platforms.
Update power domain names required for sc7280 platforms.
ASoC: qcom: dt-bindings: Update bindings for clocks in lpass digital codes
Update dt-bindings for clocks as per ADSP bypass solutions, in which macro
dcodec GDSCs are enabled using power domains in lpass digital codec drivers.
AK4613 has STEREO/TDM512/TDM256/TDM128 mode.
Renesas is the only user of ak4613 on upstream for now,
and is using it as STEREO mode, because of board connection.
Thus, current driver is supporting STEREO mode only, and other
modes are not supported.
But I noticed that I can try first 2ch out of TDM256 mode 8ch
Playback even in such a situation.
But because of board connection, I can't test full TDM256 mode,
and/or other TDM mode. Thus I don't want to add new DT propaty for now.
This patch enables TDM256 mode test by "ifdef style",
but it has no effect to current supported STEREO mode.
You can define AK4613_ENABLE_TDM_TEST to try TDM256 mode.
Please don't hesitate to break current code if you can add full TDM256
and/or other TDM mode. You don't need to care compatibility with Renesas.
ASoC: soc-utils: Add kunit test for snd_soc_tdm_params_to_bclk()
Create a new kunit test for soc-utils and use it to test
snd_soc_tdm_params_to_bclk().
The test uses a table of values to avoid the possibility that an
on-the-fly generator contains the same algorithmic error as the
function-under-test and so fails to detect a bug.
There is no need to test every possible combination of values.
Enough test cases are included to give confidence that the function
is producing the correct results.
ASoC: soc-utils: Add helper to calculate BCLK from TDM info
Add a helper function snd_soc_tdm_params_to_bclk() to calculate
the bclk from params info and the tdm sots configuration.
When using a TDM frame of N slots of width W bits:
bclk = sample_rate * N * W
As a convenience to simplify calling code, if the slot count or
slot width are 0 a value will be obtained from the params. This
allows calling code to use this one function to handle cases of
TDM where only one parameter is fixed, or I2S where the slot width
is fixed (for example to set a 32-bit slot for 24-bit samples).
Also as a convenience the slot count can optionally be rounded up
to a multiple. This is mainly useful for I2S systems, since I2S has
two phases of LRCLK the number of slots is always a multiple of 2.
Add a convenience macro for defining a single (mono) TLV control
with a signed value. This can already be done by using
SOC_DOUBLE_R_S_TLV() with the same address for left and right
registers, but a dedicated macro is more readable.
Miaoqian Lin [Mon, 4 Apr 2022 09:07:46 +0000 (09:07 +0000)]
ASoC: rk817: Use devm_clk_get() in rk817_platform_probe
We need to call clk_put() to undo clk_get() in the error path.
Use devm_clk_get() to obtain a reference to the clock, It has
the benefit that clk_put() is no longer required.
This series adds support for Asynchronous Sample Rate Converter (ASRC)
module on Tegra186 and later generations of SoCs. ASRC is a client of
AHUB. The driver and DT support is added to make it work with Tegra
audio graph card. The module can be plugged into audio path using ALSA
mixer controls.
ASRC supports two modes of operation, where it gets the ratio info
from SW and ratio detector module. Presently the support is added for
SW mode.
Following patches contain minor changes, cleaning up code to be easier
to read. Clean up few loops, to be simpler or altogether remove them.
Rename some things to make code easier to understand.
This series contains a bunch of fairly minor changes which overhaul and
modernise the WM8731 driver, there should be no impact on the driver's
functionality.
This series of patches adds support for mt8195 board with mt6359, max98390
and rt5682. To prevent from copy-paste components, mt8195 machine
drivers and dt-bindings are merged in the patch.
At some point the machine driver for the audio subsystem on the
AT91SAM9G20-EK appears to have bitrotted, no longer probing due to
effects of the transition to the common clock framework. The first patch
in this series fixes the initial probe problem, with the rest of the
series being random other fixes and cleanups I noticed while looking at
the driver.
Hui Wang [Thu, 24 Mar 2022 08:18:39 +0000 (16:18 +0800)]
ASoC: cs35l41: Fix a shift-out-of-bounds warning found by UBSAN
We enabled UBSAN in the ubuntu kernel, and the cs35l41 driver triggers
a warning calltrace like below:
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: bitoffset= 8, word_offset=23, bit_sum mod 32=0, otp_map[i].size = 24
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: bitoffset= 0, word_offset=24, bit_sum mod 32=24, otp_map[i].size = 0
================================================================================
UBSAN: shift-out-of-bounds in linux-kernel-src/sound/soc/codecs/cs35l41-lib.c:836:8
shift exponent 64 is too large for 64-bit type 'long unsigned int'
CPU: 10 PID: 595 Comm: systemd-udevd Not tainted 5.15.0-23-generic #23
Hardware name: LENOVO \x02MFG_IN_GO/\x02MFG_IN_GO, BIOS N3GET19W (1.00 ) 03/11/2022
Call Trace:
<TASK>
show_stack+0x52/0x58
dump_stack_lvl+0x4a/0x5f
dump_stack+0x10/0x12
ubsan_epilogue+0x9/0x45
__ubsan_handle_shift_out_of_bounds.cold+0x61/0xef
? regmap_unlock_mutex+0xe/0x10
cs35l41_otp_unpack.cold+0x1c6/0x2b2 [snd_soc_cs35l41_lib]
cs35l41_hda_probe+0x24f/0x33a [snd_hda_scodec_cs35l41]
cs35l41_hda_i2c_probe+0x65/0x90 [snd_hda_scodec_cs35l41_i2c]
When both bitoffset and otp_map[i].size are 0, the line 836 will
result in GENMASK(-1, 0), this triggers the shift-out-of-bounds
calltrace.
Here add a checking, if both bitoffset and otp_map[i].size are 0,
do not run GENMASK() and directly set otp_val to 0, this will not
bring any function change on the driver but could avoid the calltrace.
Hui Wang [Thu, 24 Mar 2022 08:18:38 +0000 (16:18 +0800)]
ASoC: cs35l41: Add one more variable in the debug log
otp_map[].size is a key variable to compute the value of otp_val and
to update the bit_offset, it is helpful to debug if could put it in
the debug log.
Xiaomeng Tong [Sun, 27 Mar 2022 08:10:02 +0000 (16:10 +0800)]
ASoC: rt5682: fix an incorrect NULL check on list iterator
The bug is here:
if (!dai) {
The list iterator value 'dai' will *always* be set and non-NULL
by for_each_component_dais(), so it is incorrect to assume that
the iterator value will be NULL if the list is empty or no element
is found (In fact, it will be a bogus pointer to an invalid struct
object containing the HEAD). Otherwise it will bypass the check
'if (!dai) {' (never call dev_err() and never return -ENODEV;)
and lead to invalid memory access lately when calling
'rt5682_set_bclk1_ratio(dai, factor);'.
To fix the bug, just return rt5682_set_bclk1_ratio(dai, factor);
when found the 'dai', otherwise dev_err() and return -ENODEV;
Xiaomeng Tong [Tue, 29 Mar 2022 01:21:34 +0000 (09:21 +0800)]
ASoC: soc-dapm: fix two incorrect uses of list iterator
These two bug are here:
list_for_each_entry_safe_continue(w, n, list,
power_list);
list_for_each_entry_safe_continue(w, n, list,
power_list);
After the list_for_each_entry_safe_continue() exits, the list iterator
will always be a bogus pointer which point to an invalid struct objdect
containing HEAD member. The funciton poniter 'w->event' will be a
invalid value which can lead to a control-flow hijack if the 'w' can be
controlled.
The original intention was to continue the outer list_for_each_entry_safe()
loop with the same entry if w->event is NULL, but misunderstanding the
meaning of list_for_each_entry_safe_continue().
CATPT_SSP_IFACE_LAST is being used only to calculate CATPT_SSP_COUNT.
Make CATPT_SSP_COUNT part of the enum directly and remove the redundant
constant.
Stephen Kitt [Fri, 25 Mar 2022 17:07:34 +0000 (18:07 +0100)]
ASoC: cs*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Stephen Kitt [Fri, 25 Mar 2022 16:25:15 +0000 (17:25 +0100)]
ASoC: wm*: use i2c_match_id and simple i2c probe
As part of the ongoing i2c transition to the simple probe
("probe_new"), this patch uses i2c_match_id to retrieve the
driver_data for the probed device. The id parameter is thus no longer
necessary and the simple probe can be used instead.
Stephen Kitt [Fri, 25 Mar 2022 16:48:28 +0000 (17:48 +0100)]
ASoC: ada*, ssm*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
Hui Wang [Mon, 28 Mar 2022 12:35:35 +0000 (20:35 +0800)]
ASoC: cs35l41: Fix an out-of-bounds access in otp_packed_element_t
The CS35L41_NUM_OTP_ELEM is 100, but only 99 entries are defined in
the array otp_map_1/2[CS35L41_NUM_OTP_ELEM], this will trigger UBSAN
to report a shift-out-of-bounds warning in the cs35l41_otp_unpack()
since the last entry in the array will result in GENMASK(-1, 0).
UBSAN reports this problem:
UBSAN: shift-out-of-bounds in /home/hwang4/build/jammy/jammy/sound/soc/codecs/cs35l41-lib.c:836:8
shift exponent 64 is too large for 64-bit type 'long unsigned int'
CPU: 10 PID: 595 Comm: systemd-udevd Not tainted 5.15.0-23-generic #23
Hardware name: LENOVO \x02MFG_IN_GO/\x02MFG_IN_GO, BIOS N3GET19W (1.00 ) 03/11/2022
Call Trace:
<TASK>
show_stack+0x52/0x58
dump_stack_lvl+0x4a/0x5f
dump_stack+0x10/0x12
ubsan_epilogue+0x9/0x45
__ubsan_handle_shift_out_of_bounds.cold+0x61/0xef
? regmap_unlock_mutex+0xe/0x10
cs35l41_otp_unpack.cold+0x1c6/0x2b2 [snd_soc_cs35l41_lib]
cs35l41_hda_probe+0x24f/0x33a [snd_hda_scodec_cs35l41]
cs35l41_hda_i2c_probe+0x65/0x90 [snd_hda_scodec_cs35l41_i2c]
? cs35l41_hda_i2c_remove+0x20/0x20 [snd_hda_scodec_cs35l41_i2c]
i2c_device_probe+0x252/0x2b0
Hui Wang [Mon, 28 Mar 2022 12:35:34 +0000 (20:35 +0800)]
ASoC: cs35l41: Add one more variable in the debug log
otp_map[].size is a key variable to compute the value of otp_val and
to update the bit_offset, it is helpful to debug if could put it in
the debug log.
These structs only have their address assigned to the
dapm_{routes,widget} fields in the snd_soc_card struct,
both which are pointers to const data. Make them const to
allow the compiler to put them in read-only memory.
Stephen Kitt [Fri, 25 Mar 2022 17:19:04 +0000 (18:19 +0100)]
ASoC: da7218: use i2c_match_id and simple i2c probe
As part of the ongoing i2c transition to the simple probe
("probe_new"), this patch uses i2c_match_id to retrieve the
driver_data for the probed device. The id parameter is thus no longer
necessary and the simple probe can be used instead.
This patch follows the model set by da7218_of_get_id().
Stephen Kitt [Fri, 25 Mar 2022 17:12:43 +0000 (18:12 +0100)]
ASoC: da[79]*: use simple i2c probe function
The i2c probe functions here don't use the id information provided in
their second argument, so the single-parameter i2c probe function
("probe_new") can be used instead.
This avoids scanning the identifier tables during probes.
It seems that flush_scheduled_work() is called without any real
purpose at sst_context_cleanup() (the driver doesn't put works on the
global queue at all). As the flush_schedule_work() function is going
to be abolished in near future, let's drop it now.
Stephen Kitt [Fri, 25 Mar 2022 16:54:52 +0000 (17:54 +0100)]
ASoC: ad1*, ada*, ssm*: use i2c_match_id and simple i2c probe
As part of the ongoing i2c transition to the simple probe
("probe_new"), this patch uses i2c_match_id to retrieve the
driver_data for the probed device. The id parameter is thus no longer
necessary and the simple probe can be used instead.
ASoC: codecs: Fix error handling in power domain init and exit handlers
Update error handling in power domain init and exit handlers, as existing handling
may cause issues in device remove function.
Use appropriate pm core api for power domain get and sync to avoid redundant code.
Jaroslav Kysela [Tue, 29 Mar 2022 12:00:39 +0000 (14:00 +0200)]
ASoC: SOF: topology: use new sound control LED layer
Use the new sound control LED layer instead the direct ledtrig_audio_set()
call - see 22d8de62f11b ("ALSA: control - add generic LED trigger module
as the new control layer").
ASoC: soc-pcm: use GFP_KERNEL when the code is sleepable
At the kzalloc() call in dpcm_be_connect(), there is no spin lock involved.
It's merely protected by card->pcm_mutex, instead. The spinlock is applied
at the later call with snd_soc_pcm_stream_lock_irq() only for the list
manipulations. (See it's *_irq(), not *_irqsave(); that means the context
being sleepable at that point.) So, we can use GFP_KERNEL safely there.
This patch revert commit d8a9c6e1f676 ("ASoC: soc-pcm: use GFP_ATOMIC for
dpcm structure") which is no longer needed since commit b7898396f4bb
("ASoC: soc-pcm: Fix and cleanup DPCM locking").
Xiaomeng Tong [Sun, 27 Mar 2022 08:13:00 +0000 (16:13 +0800)]
codecs: rt5682s: fix an incorrect NULL check on list iterator
The bug is here:
if (!dai) {
The list iterator value 'dai' will *always* be set and non-NULL
by for_each_component_dais(), so it is incorrect to assume that
the iterator value will be NULL if the list is empty or no element
is found (In fact, it will be a bogus pointer to an invalid struct
object containing the HEAD). Otherwise it will bypass the check
'if (!dai) {' (never call dev_err() and never return -ENODEV;)
and lead to invalid memory access lately when calling
'rt5682s_set_bclk1_ratio(dai, factor);'.
To fix the bug, just return rt5682s_set_bclk1_ratio(dai, factor);
when found the 'dai', otherwise dev_err() and return -ENODEV;
Peter Ujfalusi [Thu, 31 Mar 2022 11:49:57 +0000 (14:49 +0300)]
ASoC: topology: Correct error handling in soc_tplg_dapm_widget_create()
Academic correction of error handling:
In case the allocation of kc or kcontrol_type fails the correct label to
jump is hdr_err since the template.sname has been also allocated at this
point.
projects / linux.git / log