An atomicity violation occurs when the validity of the variables
da7219->clk_src and da7219->mclk_rate is being assessed. Since the entire
assessment is not protected by a lock, the da7219 variable might still be
in flux during the assessment, rendering this check invalid.
To fix this issue, we recommend adding a lock before the block
if ((da7219->clk_src == clk_id) && (da7219->mclk_rate == freq)) so that
the legitimacy check for da7219->clk_src and da7219->mclk_rate is
protected by the lock, ensuring the validity of the check.
This possible bug is found by an experimental static analysis tool
developed by our team. This tool analyzes the locking APIs
to extract function pairs that can be concurrently executed, and then
analyzes the instructions in the paired functions to identify possible
concurrency bugs including data races and atomicity violations.
Fixes: 6d817c0e9f ("ASoC: codecs: Add da7219 codec driver")
Cc: stable@vger.kernel.org
Signed-off-by: Qiu-ji Chen <chenqiuji666@gmail.com>
Link: https://patch.msgid.link/20240930101216.23723-1-chenqiuji666@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Some Qualcomm X1E laptops have only two speakers. Regardless whether
this sound card driver is suitable for them (we could re-use one for
some older SoC), we should set reasonable channel map depending on the
number of channels, not always 4-speaker setup.
This change is necessary for bringing audio support on Lenovo Thinkpad
T14s with Qualcomm X1E78100 and only two speakers.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://patch.msgid.link/20241023124152.130706-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
i.MX6SL and i.MX6SX SPDIF blocks are compatible with i.MX35.
Document 'fsl,imx35-spdif' as a fallback compatible for these two
chip variants.
This fixes the following dt-schema warnings:
compatible: ['fsl,imx6sl-spdif', 'fsl,imx35-spdif'] is too long
compatible: ['fsl,imx6sx-spdif', 'fsl,imx35-spdif'] is too long
Signed-off-by: Fabio Estevam <festevam@denx.de>
Link: https://patch.msgid.link/20241028020749.36972-1-festevam@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Prabhakar <prabhakar.csengg@gmail.com>:
This patch series aims to rename "sh" to "renesas", along with this
the references to this path have been updated.
Merge series from James Calligeros <jcalligeros99@gmail.com>:
This patch set adds a driver for the Cirrus Logic CS42L84 codec. This chip
is (so far) found only on Apple Silicon Macs. In keeping with proud Apple
tradition, the CS42L84 is essentially just a CS42L42 with a different
regmap and no publicly available datasheet. It may also be missing its
parent's S/PDIF capabilities as none of Apple's devices support S/PDIF out,
however this cannot be positively confirmed.
This driver has lived in the downstream Asahi tree for quite a while now,
and gained some refinements along the way. I have squashed most of these
into the initial driver commit as they were small changes like tweaking
msleep()s or filling out TLVs, but left seperate a larger change to
tip/ring sense IRQ handling as it differs significantly from what is found
in the CS42L42 driver.
Merge series from Ryan Walklin <ryan@testtoast.com>:
The Allwinner H616 has a playback-only audio codec, with a single stereo
or differential-mono line output. This patch series adds support for
the H616 (and H313/H618/H700/T507) SoC. Based on the Allwinner kernel
SDK driver, and tested on the H700.
Some jacks integrated on devices with this codec, such as certain
Apple Silicon Macs, have quite trigger-happy tip sense switches that
cause a tip sense IRQ before the plug is fully seated. If users are
unfortunate with their timing, this can lead to headsets being detected
as mic-less headphones among other issues with the codec's device
detection routines.
Introduce some rudimentary ring sense interrupt handling so that we
can re-trigger the codec's detection routines when we are certain
that the plug is fully seated.
This seems to differ from what other Cirrus drivers do, but is
necessary for devices to be reliably detected properly here.
Signed-off-by: James Calligeros <jcalligeros99@gmail.com>
Reviewed-by: Neal Gompa <neal@gompa.dev>
Link: https://patch.msgid.link/20241020-cs42l84-v2-3-37ba2b6721d9@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The H616 has an audio codec compatible with the sun4i-a10 driver.
The codec is relatively cut down compared to some of the other Allwinner
SoCs and only has a single line-out route (relying on a separate digital
microphone IP block for input). HDMI and SPDIF audio are handled
separately by an audio hub IP block, which is not currently implemented
in mainline kernels. This requires some additional flexibility in the
DMA bindings.
Add compatible string and routing for the H616 audio codec, and update
the required DMA descriptions.
Signed-off-by: Ryan Walklin <ryan@testtoast.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://patch.msgid.link/20241023075917.186835-3-ryan@testtoast.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The H616 SoC codec is playback-only with a single line-out route, and
has some register differences from prior codecs.
Add the required compatible string, registers, quirks, DAPM widgets,
codec controls and routes, based on existing devices and the H616
datasheet.
Signed-off-by: Ryan Walklin <ryan@testtoast.com>
Link: https://patch.msgid.link/20241023075917.186835-6-ryan@testtoast.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>:
Now, we don't need to use dpcm_playback/capture flags for DPCM since v6.11,
and thus we can use playback/capture_only flags instead too for DPCM,
which is same as non-DPCM.
Let's remove dpcm_playback/capture flags, and use playback/capture_only flags
instead.
[Before] [After]
dpcm_playback = 1; /* no setting is needed */
dpcm_capture = 1;
dpcm_playback = 1; playback_only = 1;
dpcm_capture = 0;
dpcm_playback = 0; capture_only = 1;
dpcm_capture = 1;
And then, because no-one uses dpcm_xxx flags, we don't need to have the code
for both DPCM and non-DPCM. These can be handled by same code.
Merge series from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>:
Add new debug only parameters to override the default IPC and boot
timeout, these are only available if CONFIG_SND_SOC_SOF_DEBUG is
enabled, only for developers.
The second patch moves the firmware, topology name/path and IPC
version overrides from acpi/pci/of platform code to core and markes
the platform parameters deprecated, to be removed in few kernel
cycles.
There is really no point of duplicating these overrides as they
are not platform dependent.
DPCM and non-DPCM validation check are very similar. The big difference
is that DPCM doesn't check Codec validation. This is historical reason.
It should be checked, but it breaks existing driver/behavior.
Anyway, if we uses dummy DAI as Codec when DPCM case, there is no
difference between DPCM and non-DPCM. Let's merge these.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://patch.msgid.link/8734kq9vgq.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
