ASoC: SOF: Intel: start splitting top-level from common parts
The existing code relies on the 'HDA_COMMON' module and namespace. We
need to start splitting top-level parts from the low-level ones,
otherwise we will not be able to reuse the low-level parts DMA support
for SoundWire/BPT.
This patch adds the initial split between the sof-pci-intel-xxx
modules and the common parts, in a follow-up patch we will further
split the HDA_COMMON parts
Since the PCI modules are not all independent, i.e. the CNL parts are
also used in JSL and TGL, additional Kconfig and namespace modules
were added.
ASoC: SOF: Intel: regroup all SoundWire/Intel functions in hda.c
To avoid circular dependencies between SOF/Intel and SoundWire/Intel,
we need to split the top-level hda.c from the rest of the code. This
patch first regroups all SoundWire related code in hda.c.
Export this helper so that we can report the DPIB position if the BPT
DMA do not complete - this is very useful to see if the DMA started or
gets stuck somehow with invalid bandwidth configurations.
ASoC: Use inline function for type safety in snd_soc_substream_to_rtd()
A common pattern in sound drivers is getting 'struct snd_soc_pcm_runtime'
from 'struct snd_pcm_substream' opaque pointer private_data field with
snd_soc_substream_to_rtd(). However 'private_data' appears in several
other structures as well, including 'struct snd_compr_stream'. The
field might not hold the same type for every structure, although seems
the case at least for 'struct snd_compr_stream', so code can easily make
a mistake by using macro for wrong structure passed as argument.
Switch from macro to inline function, so such mistake will be build-time
detectable.
Mark Brown [Tue, 30 Apr 2024 15:00:17 +0000 (00:00 +0900)]
ASoC: doc: dapm: various improvements
Merge series from Luca Ceresoli <luca.ceresoli@bootlin.com>:
This series applies various improvements to the DAPM documentation: a
rewrite of a few sections for clarity, style improvements and typo fixes.
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
---
Changes in v2:
- avoid wrapping in patch 3 as suggested by Alex
- Link to v1: https://lore.kernel.org/r/20240416-dapm-docs-v1-0-a818d2819bf6@bootlin.com
---
Luca Ceresoli (12):
ASoC: doc: dapm: fix typos
ASoC: doc: dapm: fix struct name
ASoC: doc: dapm: minor rewording
ASoC: doc: dapm: remove dash after colon
ASoC: doc: dapm: clarify it's an internal API
ASoC: doc: dapm: replace "map" with "graph"
ASoC: doc: dapm: extend initial descrption
ASoC: doc: dapm: describe how widgets and routes are registered
ASoC: doc: dapm: fix and improve section "Registering DAPM controls"
ASoC: doc: dapm: improve section "Codec/DSP Widget Interconnections"
ASoC: doc: dapm: update section "DAPM Widget Events"
ASoC: doc: dapm: update event types
Mark Brown [Tue, 30 Apr 2024 14:35:00 +0000 (23:35 +0900)]
ASoC: Intel: avs: PCM code cleanup
Merge series from Cezary Rojewski <cezary.rojewski@intel.com>:
A set of changes that aims to improve readability of cohesiveness of the
pcm code for the avs-driver.
Start off with a change that synchronizes DAI open/close - DAIs are
started up in ascending order yet their shutdown does not follow the
scheme - it is done in the ascending order too, rather than desceding
one. This patch is a dependency for the next one in line.
To align the HDAudio DAI startup/shutdown with the non-HDAudio
equivalents, relocate the code from component to DAI. The reason above
is a dependency stems from codec driver requirements - HDAudio code
found in sound/pci/hda/ expects substream->runtime->private_data to
point to a valid stream (HOST) pointer.
With the hard part done, the follow up changes update the existing code
to reduce it is complexity - removal of duplicates, renaming of
ambiguous functions and adding new fields to DAI-data object so that the
number of local variables and casts is reduced.
Mark Brown [Tue, 30 Apr 2024 14:34:51 +0000 (23:34 +0900)]
ASoC: Constify local snd_sof_dsp_ops
Merge series from Krzysztof Kozlowski <krzk@kernel.org>:
The core code does not modify the 'struct snd_sof_dsp_ops' passed via
pointer in various places, so this can be made pointer to const in few
places. This in turn allows few drivers to have the local (usually
static) 'struct snd_sof_dsp_ops' as const which increased code safety,
as it is now part of rodata.
Not all drivers can be made safer that way. Intel and AMD rely on
customizing that 'struct snd_sof_dsp_ops' before passing to SOF, so they
won't benefit. They don't lose anything., either.
Mark Brown [Tue, 30 Apr 2024 14:34:44 +0000 (23:34 +0900)]
ASoc: SOF: misc updates for 6.10
Merge series from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:
Fixes when fw_lib_prefix is not set, updated error messages, improved
dmesg logs to SoundWire configurations not supported by ACPI
tables/topology and support for IEC61937 passthrough.
Mark Brown [Tue, 30 Apr 2024 14:34:37 +0000 (23:34 +0900)]
ASoC: Intel: updates for 6.10 - part5
Merge series from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:
This patchset corrects a couple of mistakes corrected, improves
snd_soc_card allocation. The new functionality is mostly for
SoundWire platforms, with new SKUs for Dell and Acer, and support for
the Cirrus Logic bridge/sidecar amplifier topology.
ASoC: sunxi: DMIC: Add controls for adjusting the mic gains
The AllWinner H6 and later SoCs that sport a DMIC block contain a set of registers to control
the gain (left + right) of each of the four supported channels.
Add ASoC controls for changing each of the stereo channel gains using alsamixer and alike
ASoC: dt-bindings: tegra30-i2s: convert to dt schema
Convert NVIDIA Tegra30 I2S binding to DT schema and
add "clock-names" property used by multiple tegra i2s blocks
in arch/arm64/boot/dts/nvidia/tegra132.dtsi. This is not a
required property by the binding.
'struct snd_sof_dsp_ops' is not modified by core code, so it can be made
const for increased code safety.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Tested-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20240426-n-const-ops-var-v2-14-e553fe67ae82@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org>
'struct snd_sof_dsp_ops' is not modified by core code, so it can be made
const for increased code safety.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Tested-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20240426-n-const-ops-var-v2-13-e553fe67ae82@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: SOF: Constify stored pointer to snd_sof_dsp_ops
The core code does not modify the 'struct snd_sof_dsp_ops', stored in
'struct sof_dev_desc'. Make the pointer to 'struct snd_sof_dsp_ops'
const to annotate this, make code a bit safer and allow individual
drivers to also define this structure as const.
If a PCM is set to use ChainDMA then add it to the card->components string
after a marker of iec61937-pcm:, for example on current HDA platforms where
HDMI is set to use ChainDMA:
iec61937-pcm:5,4,3 (the order of the PCM ids can differ)
UCM is expected to parse and use this property to allow applications to
use bytestream passthrough in a standard way.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426153902.39560-6-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: SOF: Intel: hda: list SoundWire peripherals on mismatch
Most of the SoundWire support issues come from bad ACPI information,
or configuration reported by ACPI that are not supported by the SOF
driver/topology. The users see a "No SoundWire machine driver found"
message without any details, and the fallback to HDaudio w/ HDMI is
used.
We can reduce our support load with a clear dev_info() log that will
give us a clear hint on the mismatch and why a machine driver/topology
were not found.
Example log on a MTL device:
[ 13.158599] sof-audio-pci-intel-mtl 0000:00:1f.3: No SoundWire machine driver found for the ACPI-reported configuration:
[ 13.158603] sof-audio-pci-intel-mtl 0000:00:1f.3: link 0 mfg_id 0x025d part_id 0x0713 version 0x3
[ 13.158606] sof-audio-pci-intel-mtl 0000:00:1f.3: link 1 mfg_id 0x025d part_id 0x1316 version 0x3
[ 13.158608] sof-audio-pci-intel-mtl 0000:00:1f.3: link 2 mfg_id 0x025d part_id 0x1316 version 0x3
In parallel, we will also provide an update to `alsa-info` to log all
SoundWire peripherals found in ACPI tables.
ASoC: SOF: topology: remove incorrect widget id in error message
In sof_widget_ready() function, the shift field of
struct snd_soc_tplg_dapm_widget is incorrectly used to print
widget id in dev_err(scomp->dev, "error: failed to add widget id %d ..",
this patch removes the useless tw->shift from the error output.
Peter Ujfalusi [Fri, 26 Apr 2024 15:38:58 +0000 (10:38 -0500)]
ASoC: SOF: debug: Handle cases when fw_lib_prefix is not set, NULL
The firmware libraries are not supported by IPC3, the fw_lib_path is not
a valid parameter and it is always NULL.
Do not create the debugfs file for IPC3 at all as it is not applicable.
With IPC4 some vendors/platforms might not support loadable libraries and
the fw_lib_prefix is left to NULL to indicate this.
Handle such case with allocating "Not supported" string.
Reviewed-by: Marc Herbert <marc.herbert@intel.com> Fixes: 17f4041244e6 ("ASoC: SOF: debug: show firmware/topology prefix/names") Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426153902.39560-2-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Maciej Strozek [Fri, 26 Apr 2024 15:21:23 +0000 (10:21 -0500)]
ASoC: intel: sof_sdw: Add support for cs42l43-cs35l56 sidecar amps
The cs42l43 has both a SPI master and an I2S interface, these can
be used to populate 2 cs35l56 amplifiers as sidecar devices along
side the cs42l43. Giving a system that looks like:
Add a quirk to specify this feature is present and use it to add
codec to codec DAI link to connect the amplifiers into the sound
card, add appropriate widgets, and setup clocking on the
amplifiers.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com> Signed-off-by: Maciej Strozek <mstrozek@opensource.cirrus.com> Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426152123.36284-13-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Charles Keepax [Fri, 26 Apr 2024 15:21:22 +0000 (10:21 -0500)]
ASoC: Intel: sof_sdw: Add callbacks to register sidecar devices
Add support for systems that have additional non-SoundWire devices
(sidecars) connected to one of the SoundWire devices in the
system. This is done through the addition of two callbacks, one used
at endpoint parsing time that will return the number of devices and
DAI links to be added, and another called later as the DAI links are
created that will populate those devices into the appropriate arrays.
Charles Keepax [Fri, 26 Apr 2024 15:21:21 +0000 (10:21 -0500)]
ASoC: Intel: sof_sdw: Delay update of the codec_conf array
Move the population of the codec_conf array from endpoint parsing
time to link creation time. This is slightly cleaner as the
population is done whilst the DAI links are also being populated,
putting all population together. However, primarily this facilitates
allowing additional non-SoundWire devices to be easily added into
the array in future feature additions.
ASoC: Intel: sof_sdw: add a space before cfg-amp in components
UCM parse amp with Regex " cfg-amp:([0-9]+)". The "ASoC: Intel: sof_sdw:
remove FOUR_SPEAKER quirks" patch removed "cfg-spk:%d " from components
which removed the necessary space as well and cause UCM can't parse the
amp number properly.
UCM can load a board-specific file based on the card long_name. Remove
the constant "Intel Soundwire SOF" long_name so that the ASoC core can
set the long_name based on DMI information.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com> Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426152123.36284-8-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
The static card_sof_sdw struct is modified during runtime and in case the
module is not removed, but the card is, then the next time the card is
created the card_sof_sdw will contain information from the previous card
which might lead to hard to debug issues, side effects.
Move the snd_soc_card into mc_private and use that to make sure that the
card is initialized correctly.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426152123.36284-7-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
The static hda_soc_card might be modified during runtime which might cause
issues on next time when the card is created.
For example if the dmic_num was set with module parameter then removed for
the next module loading then the card's components will still going to
point to the previous boot's cfg-dmics:X string.
There might be other places where devm allocated memory have been freed but
the hda_soc_card still pointing to the now unallocated memory (the memory
is freed when the platform device is removed).
Fix this issue by moving the snd_soc_card into skl_hda_private and use it
for the card registration to ensure that it is correctly initialized every
time.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426152123.36284-2-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: Intel: avs: Store pointer to link_stream in dma_data
While the HDAudio codec driver expectations must be met - store valid
pointer to HDAudio LINK stream in substream->runtime->private_data - the
code is more readable and easier to maintain if dma_data stores pointers
to both HOST and LINK stream.
DAI BE operations can refer to the LINK stream with data->link_stream,
similarly to how DAI FE operations access the HOST stream with
data->host_stream.
ASoC: Intel: avs: Remove redundancy around DAI prepare
Drop unused arguments in the avs_dai_prepare() function. With the
function updated, it matches its template in snd_soc_dai_ops and can be
referenced throughout the pcm.c file without need of any wrappers.
ASoC: Intel: avs: Remove redundancy around DAI startup
Half of the arguments in avs_dai_startup() are unused and can be
dropped. With the function updated, it matches its template in
snd_soc_dai_ops and can be referenced throughout the pcm.c file without
need of any wrappers.
ASoC: Intel: avs: Remove redundancy around DAI shutdown
Move avs_dai_nonhda_be_shutdown() to avs_dai_shutdown() as the function
is common for all transfer types, not just non-HDAudio ones. Use it
to simplify avs_dai_fe_shutdown().
While at it, fix explicit kfree(data) and use the destructor instead.
ASoC: Intel: avs: Relocate HDA BE DAI specific operations
DAI's startup()/shutdown() shall deal with allocation and freeing of
resources needed to facilitate streaming over it. Currently for HDAudio
BE DAIs some of that task is done in component->open()/close(). Relocate
the relevant pieces to address that.
ASoC: pcm: Reverse iterate DAIs when shutting them down
During startup snd_soc_dai_startup() is launched in ascending order and
the exact same thing is done during shutdown procedure. Reverse the
order in the latter so that it is symmetric to the former.
The example in this section is not in the kernel sources anymore. Replace
it with an up to date code fragment. Reword the initial paragraph. Remove
"Please" which is not standard practice in documentation.
Improve wording in a few places, cleanup ReST colon syntax, remove space
before colon, and remove the "codec" parameter in the parentheses of
snd_soc_dapm_new_widgets() (there should be no parameters in the docs, and
that function takes a card, not a codec).
ASoC: doc: dapm: fix and improve section "Registering DAPM controls"
Extend the first paragraph to mention the {,num_}dapm_routes fields just
like the widget conterparts. Mention the route fields also in the code
example. Fix "at build time", this really means "at probe time".
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com> Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> Link: https://lore.kernel.org/r/20240426-dapm-docs-v2-9-87b07547eb5b@bootlin.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: doc: dapm: describe how widgets and routes are registered
The small paragraph describing how to register widgets is incomplete (does
not mention routes) and mentions snd_soc_dapm_new_control() which is not
really used. Moreover it appears as a part of the "Virtual Widgets"
subsection.
Replace it with a detailed and current description of how widgets and
routes are registered, mentioning both static declaration and runtime
registration. Also make this a section on its own.
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com> Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> Link: https://lore.kernel.org/r/20240426-dapm-docs-v2-8-87b07547eb5b@bootlin.com Signed-off-by: Mark Brown <broonie@kernel.org>
Extend the initial description providing clearer definitions of "widget"
and "route", and to stop using the word "component" to mean "widget". Give
more details and clarify wording and add a picture representing a real DAPM
graph.
Group all the introductory paragraphs before the "DAPM power domains", and
split the latter to a specific section.
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com> Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> Link: https://lore.kernel.org/r/20240426-dapm-docs-v2-7-87b07547eb5b@bootlin.com Signed-off-by: Mark Brown <broonie@kernel.org>
Brent Lu [Fri, 26 Apr 2024 15:25:29 +0000 (10:25 -0500)]
ASoC: Intel: sof-rt5682: support bclk as PLL source on rt5682s
For rt5682s codec, we could use bclk as PLL source when the frequency
is 3.072MHz but no 2.4MHz. Update the code to select correct pll_id
and clk_id for 3.072MHz bclk.
Brent Lu [Fri, 26 Apr 2024 15:25:28 +0000 (10:25 -0500)]
ASoC: Intel: sof-rt5682: get bclk frequency from topology
A different bclk frequency 3.072MHz was introduced to tgl platform and
is used in mtl topologies. Use SOF API to get frequency from topology
instead of hardcoding.
Brent Lu [Fri, 26 Apr 2024 15:25:26 +0000 (10:25 -0500)]
ASoC: Intel: sof-rt5682: add driver_data to sof_rt5682 board
Move default BYT/CHT quirk to driver_data of sof_rt5682 board. This
fixes a problem that DMI quirk of Minnowboard board got overwritten in
probe function since it's a BYT board.
Fixes: c68e07970eca ("ASoC: intel: sof_rt5682: Add quirk for number of HDMI DAI's") Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com> Signed-off-by: Brent Lu <brent.lu@intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20240426152529.38345-21-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Brent Lu [Fri, 26 Apr 2024 15:25:21 +0000 (10:25 -0500)]
ASoC: Intel: sof_rt5682: support MAX98357A on glk boards
For glk boards, MAX98357A speaker amplifier is supported by machine
driver glk_rt5682_mx98357a with sound card name glkrt5682max. Use same
name for backward compatibility with existing devices on market.
Brent Lu [Fri, 26 Apr 2024 15:25:20 +0000 (10:25 -0500)]
ASoC: Intel: sof_rt5682: add glk_rt5682_def for glk boards
Add the board config glk_rt5682_def to rt5682 machine driver for all
glk boards using default SSP port allocation (headphone codec on SSP2,
speaker amplifiers on SSP1).
Brent Lu [Fri, 26 Apr 2024 15:25:18 +0000 (10:25 -0500)]
ASoC: Intel: sof_da7219: support MAX98357A on glk boards
For glk boards, MAX98357A speaker amplifier is supported by machine
driver bxt_da7219_max98357a with sound card name glkda7219max. Use
same name for backward compatibility with existing devices on market.
Brent Lu [Fri, 26 Apr 2024 15:25:17 +0000 (10:25 -0500)]
ASoC: Intel: sof_da7219: add glk_da7219_def for glk boards
Add the board config glk_da7219_def to da7219 machine driver for all
glk boards using default SSP port allocation (headphone codec on SSP2,
speaker amplifiers on SSP1).
Brent Lu [Fri, 26 Apr 2024 15:25:15 +0000 (10:25 -0500)]
ASoC: Intel: sof_da7219: support MAX98357A on cml boards
For cml boards, MAX98357A speaker amplifier is supported by machine
driver bxt_da7219_max98357a with sound card name cmlda7219max. Use
same name for backward compatibility with existing devices on market.
Brent Lu [Fri, 26 Apr 2024 15:25:11 +0000 (10:25 -0500)]
ASoC: Intel: sof_da7219: support MAX98390 on cml boards
For cml boards, MAX98390 speaker amplifier is supported by machine
driver bxt_da7219_max98357a with sound card name cml_max98390_da7219.
Use same name for backward compatibility with existing devices on
market.
Brent Lu [Fri, 26 Apr 2024 15:25:10 +0000 (10:25 -0500)]
ASoC: Intel: sof_da7219: add cml_da7219_def for cml boards
Add the board config cml_da7219_def to da7219 machine driver for all
cml boards using default SSP port allocation (headphone codec on SSP0,
speaker amplifiers on SSP1).
Brent Lu [Fri, 26 Apr 2024 15:25:07 +0000 (10:25 -0500)]
ASoC: Intel: nau8825/rt5682: move speaker widget to common modules
Two machine drivers sof_rt5682 and sof_nau8825 always register two
speaker widgets 'Left Spk' and 'Right Spk' regardless the actual
number of speakers. Move the widget registration to speaker common
modules to avoid useless speaker widgets for 1 or 4 speaker boards.
Merge tag '9p-for-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs
Pull 9p fix from Eric Van Hensbergen:
"This contains a single mitigation to help deal with an apparent race
condition between client and server having to deal with inode number
collisions"
* tag '9p-for-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
fs/9p: mitigate inode collisions
Merge tag 'acpi-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull ACPI fixes from Rafael Wysocki:
"These fix three recent regressions, one introduced while enabling a
new platform firmware feature for power management, and two introduced
by a recent CPPC library update.
Specifics:
- Allow two overlapping Low-Power S0 Idle _DSM function sets to be
used at the same time (Rafael Wysocki)
- Fix bit offset computation in MASK_VAL() macro used for applying a
bitmask to a new CPPC register value (Jarred White)
- Fix access width field usage for PCC registers in CPPC (Vanshidhar
Konda)"
* tag 'acpi-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI: PM: s2idle: Evaluate all Low-Power S0 Idle _DSM functions
ACPI: CPPC: Fix access width used for PCC registers
ACPI: CPPC: Fix bit_offset shift in MASK_VAL() macro
Merge tag 'net-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from netfilter, wireless and bluetooth.
Nothing major, regression fixes are mostly in drivers, two more of
those are flowing towards us thru various trees. I wish some of the
changes went into -rc5, we'll try to keep an eye on frequency of PRs
from sub-trees.
Also disproportional number of fixes for bugs added in v6.4, strange
coincidence.
Current release - regressions:
- igc: fix LED-related deadlock on driver unbind
- wifi: mac80211: small fixes to recent clean up of the connection
process
- Revert "wifi: iwlwifi: bump FW API to 90 for BZ/SC devices", kernel
doesn't have all the code to deal with that version, yet
- Bluetooth:
- set power_ctrl_enabled on NULL returned by gpiod_get_optional()
- qca: fix invalid device address check, again
- Bluetooth:
- lots of fixes for the command submission rework from v6.4
- qca: fix NULL-deref on non-serdev suspend
Misc:
- tools: ynl: don't ignore errors in NLMSG_DONE messages"
* tag 'net-6.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (88 commits)
af_unix: Suppress false-positive lockdep splat for spin_lock() in __unix_gc().
net: b44: set pause params only when interface is up
tls: fix lockless read of strp->msg_ready in ->poll
dpll: fix dpll_pin_on_pin_register() for multiple parent pins
net: ravb: Fix registered interrupt names
octeontx2-af: fix the double free in rvu_npc_freemem()
net: ethernet: ti: am65-cpts: Fix PTPv1 message type on TX packets
ice: fix LAG and VF lock dependency in ice_reset_vf()
iavf: Fix TC config comparison with existing adapter TC config
i40e: Report MFS in decimal base instead of hex
i40e: Do not use WQ_MEM_RECLAIM flag for workqueue
net: ti: icssg-prueth: Fix signedness bug in prueth_init_rx_chns()
net/mlx5e: Advertise mlx5 ethernet driver updates sk_buff md_dst for MACsec
macsec: Detect if Rx skb is macsec-related for offloading devices that update md_dst
ethernet: Add helper for assigning packet type when dest address does not match device address
macsec: Enable devices to advertise whether they update sk_buff md_dst during offloads
net: phy: dp83869: Fix MII mode failure
netfilter: nf_tables: honor table dormant flag from netdev release event path
eth: bnxt: fix counting packets discarded due to OOM and netpoll
igc: Fix LED-related deadlock on driver unbind
...
Merge tag 'nfsd-6.9-5' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux
Pull nfsd fixes from Chuck Lever:
- Revert some backchannel fixes that went into v6.9-rc
* tag 'nfsd-6.9-5' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
Revert "NFSD: Convert the callback workqueue to use delayed_work"
Revert "NFSD: Reschedule CB operations when backchannel rpc_clnt is shut down"
Merge tag 'for-linus-2024042501' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Pull HID fixes from Benjamin Tissoires:
- A couple of i2c-hid fixes (Kenny Levinsen & Nam Cao)
- A config issue with mcp-2221 when CONFIG_IIO is not enabled
(Abdelrahman Morsy)
- A dev_err fix in intel-ish-hid (Zhang Lixu)
- A couple of mouse fixes for both nintendo and Logitech-dj (Nuno
Pereira and Yaraslau Furman)
- I'm changing my main kernel email address as it's way simpler for me
than the Red Hat one (Benjamin Tissoires)
* tag 'for-linus-2024042501' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
HID: mcp-2221: cancel delayed_work only when CONFIG_IIO is enabled
HID: logitech-dj: allow mice to use all types of reports
HID: i2c-hid: Revert to await reset ACK before reading report descriptor
HID: nintendo: Fix N64 controller being identified as mouse
MAINTAINERS: update Benjamin's email address
HID: intel-ish-hid: ipc: Fix dev_err usage with uninitialized dev->devc
HID: i2c-hid: remove I2C_HID_READ_PENDING flag to prevent lock-up
Jakub Kicinski [Thu, 25 Apr 2024 15:46:53 +0000 (08:46 -0700)]
Merge tag 'nf-24-04-25' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Pablo Neira Ayuso says:
====================
Netfilter/IPVS fixes for net
The following patchset contains two Netfilter/IPVS fixes for net:
Patch #1 fixes SCTP checksumming for IPVS with gso packets,
from Ismael Luceno.
Patch #2 honor dormant flag from netdev event path to fix a possible
double hook unregistration.
* tag 'nf-24-04-25' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
netfilter: nf_tables: honor table dormant flag from netdev release event path
ipvs: Fix checksumming on GSO of SCTP packets
====================
Peter Münster [Wed, 24 Apr 2024 13:51:52 +0000 (15:51 +0200)]
net: b44: set pause params only when interface is up
b44_free_rings() accesses b44::rx_buffers (and ::tx_buffers)
unconditionally, but b44::rx_buffers is only valid when the
device is up (they get allocated in b44_open(), and deallocated
again in b44_close()), any other time these are just a NULL pointers.
So if you try to change the pause params while the network interface
is disabled/administratively down, everything explodes (which likely
netifd tries to do).
Link: https://github.com/openwrt/openwrt/issues/13789 Fixes: 1da177e4c3f4 (Linux-2.6.12-rc2) Cc: stable@vger.kernel.org Reported-by: Peter Münster <pm@a16n.net> Suggested-by: Jonas Gorski <jonas.gorski@gmail.com> Signed-off-by: Vaclav Svoboda <svoboda@neng.cz> Tested-by: Peter Münster <pm@a16n.net> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Peter Münster <pm@a16n.net> Reviewed-by: Michael Chan <michael.chan@broadcom.com> Link: https://lore.kernel.org/r/87y192oolj.fsf@a16n.net Signed-off-by: Jakub Kicinski <kuba@kernel.org>
tls: fix lockless read of strp->msg_ready in ->poll
tls_sk_poll is called without locking the socket, and needs to read
strp->msg_ready (via tls_strp_msg_ready). Convert msg_ready to a bool
and use READ_ONCE/WRITE_ONCE where needed. The remaining reads are
only performed when the socket is locked.
dpll: fix dpll_pin_on_pin_register() for multiple parent pins
In scenario where pin is registered with multiple parent pins via
dpll_pin_on_pin_register(..), all belonging to the same dpll device.
A second call to dpll_pin_on_pin_unregister(..) would cause a call trace,
as it tries to use already released registration resources (due to fix
introduced in b446631f355e). In this scenario pin was registered twice,
so resources are not yet expected to be release until each registered
pin/pin pair is unregistered.
Currently, the following crash/call trace is produced when ice driver is
removed on the system with installed E810T NIC which includes dpll device:
Fix by adding a parent pointer as a cookie when creating a registration,
also when searching for it. For the regular pins pass NULL, this allows to
create separated registration for each parent the pin is registered with.
projects / thirdparty / linux.git / log
