Skip to content

L
linux
Switch branch/tag
  1. 29 Apr, 2024 1 commit
    • Jerome Brunet's avatar
      ASoC: meson: axg-fifo: use threaded irq to check periods · b11d2666
      Jerome Brunet authored 
      With the AXG audio subsystem, there is a possible random channel shift on
TDM capture, when the slot number per lane is more than 2, and there is
more than one lane used.

The problem has been there since the introduction of the axg audio support
but such scenario is pretty uncommon. This is why there is no loud
complains about the problem.

Solving the problem require to make the links non-atomic and use the
trigger() callback to start FEs and BEs in the appropriate order.

This was tried in the past and reverted because it caused the block irq to
sleep while atomic. However, instead of reverting, the solution is to call
snd_pcm_period_elapsed() in a non atomic context.

Use the bottom half of a threaded IRQ to do so.

Fixes: 6dc4fa17 ("ASoC: meson: add axg fifo base driver")
Signed-off-by: default avatarJerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20240426152946.3078805-2-jbrunet@baylibre.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      b11d2666
  3. 23 Apr, 2024 2 commits
  5. 20 Apr, 2024 1 commit
  7. 19 Apr, 2024 2 commits
    • Joao Paulo Goncalves's avatar
      ASoC: ti: davinci-mcasp: Fix race condition during probe · d18ca863
      Joao Paulo Goncalves authored 
      When using davinci-mcasp as CPU DAI with simple-card, there are some
conditions that cause simple-card to finish registering a sound card before
davinci-mcasp finishes registering all sound components. This creates a
non-working sound card from userspace with no problem indication apart
from not being able to play/record audio on a PCM stream. The issue
arises during simultaneous probe execution of both drivers. Specifically,
the simple-card driver, awaiting a CPU DAI, proceeds as soon as
davinci-mcasp registers its DAI. However, this process can lead to the
client mutex lock (client_mutex in soc-core.c) being held or davinci-mcasp
being preempted before PCM DMA registration on davinci-mcasp finishes.
This situation occurs when the probes of both drivers run concurrently.
Below is the code path for this condition. To solve the issue, defer
davinci-mcasp CPU DAI registration to the last step in the audio part of
it. This way, simple-card CPU DAI parsing will be deferred until all
audio components are registered.

Fail Code Path:

simple-card.c: probe starts
simple-card.c: simple_dai_link_of: simple_parse_node(..,cpu,..) returns EPROBE_DEFER, no CPU DAI yet
davinci-mcasp.c: probe starts
davinci-mcasp.c: devm_snd_soc_register_component() register CPU DAI
simple-card.c: probes again, finish CPU DAI parsing and call devm_snd_soc_register_card()
simple-card.c: finish probe
davinci-mcasp.c: *dma_pcm_platform_register() register PCM  DMA
davinci-mcasp.c: probe finish

Cc: stable@vger.kernel.org
Fixes: 9fbd58cf ("ASoC: davinci-mcasp: Choose PCM driver based on configured DMA controller")
Signed-off-by: default avatarJoao Paulo Goncalves <joao.goncalves@toradex.com>
Acked-by: default avatarPeter Ujfalusi <peter.ujfalusi@gmail.com>
Reviewed-by: default avatarJai Luthra <j-luthra@ti.com>
Link: https://lore.kernel.org/r/20240417184138.1104774-1-jpaulo.silvagoncalves@gmail.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      d18ca863
    • Amadeusz Sławiński's avatar
      ASoC: Intel: avs: Set name of control as in topology · 4cbb5050
      Amadeusz Sławiński authored 
      When creating controls attached to widgets, there are a lot of rules if
they get their name prefixed with widget name or not. Due to that
controls ended up with weirdly looking names like "ssp0_fe DSP Volume",
while topology set it to "DSP Volume".

Fix this by setting no_wname_in_kcontrol_name to true in avs topology
widgets which disables unwanted behaviour.

Fixes: be2b81b5 ("ASoC: Intel: avs: Parse control tuples")
Signed-off-by: default avatarAmadeusz Sławiński <amadeuszx.slawinski@linux.intel.com>
Link: https://lore.kernel.org/r/20240418142621.2487478-1-amadeuszx.slawinski@linux.intel.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      4cbb5050
  9. 17 Apr, 2024 1 commit
  11. 15 Apr, 2024 1 commit
  13. 11 Apr, 2024 1 commit
  15. 09 Apr, 2024 15 commits
  17. 08 Apr, 2024 7 commits
    • Richard Fitzgerald's avatar
      ASoC: cs35l56: Prevent overwriting firmware ASP config · dfd2ffb3
      Richard Fitzgerald authored 
      Only populate the ASP1 config registers in the regmap cache if the
ASP DAI is used. This prevents regcache_sync() from overwriting
these registers with their defaults when the firmware owns
control of these registers.

On a SoundWire system the ASP could be owned by the firmware to
share reference audio with the firmware on other cs35l56. Or it
can be used as a normal codec-codec interface owned by the driver.
The driver must not overwrite the registers if the firmware has
control of them.

The original implementation for this in commit 07f7d6e7
("ASoC: cs35l56: Fix for initializing ASP1 mixer registers") was
to still provide defaults for these registers, assuming that if
they were never reconfigured from defaults then regcache_sync()
would not write them out because they are not dirty. Unfortunately
regcache_sync() is not that smart. If the chip has not reset (so
the driver has not called regcache_mark_dirty()) a regcache_sync()
could write out registers that are not dirty.

To avoid accidental overwriting of the ASP registers, they are
removed from the table of defaults and instead are populated with
defaults only if one of the ASP DAI configuration functions is
called. So if the DAI has never been configured, the firmware is
assumed to have ownership of these registers, and the regmap cache
will not contain any entries for them.
Signed-off-by: default avatarRichard Fitzgerald <rf@opensource.cirrus.com>
Fixes: 07f7d6e7 ("ASoC: cs35l56: Fix for initializing ASP1 mixer registers")
Link: https://msgid.link/r/20240408101803.43183-5-rf@opensource.cirrus.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      dfd2ffb3
    • Richard Fitzgerald's avatar
      ASoC: cs35l56: Fix unintended bus access while resetting amp · d4884fd4
      Richard Fitzgerald authored 
      Use the new regmap_read_bypassed() so that the regmap can be left
in cache-only mode while it is booting, but the driver can still
read boot-status and chip-id information during this time.

This fixes race conditions where some writes could be issued to the
silicon while it is still rebooting, before the driver has determined
that the boot is complete.

This is typically prevented by putting regmap into cache-only until the
hardware is ready. But this assumes that the driver does not need to
access device registers to determine when it is "ready". For cs35l56
this involves polling a register and the original implementation relied
on having special handlers to block racing callbacks until dsp_work()
is complete. However, some cases were missed, most notably the ASP DAI
functions.

The regmap_read_bypassed() function allows the fix for this to be
simplified to putting regmap into cache-only during the reset. The
initial boot stages (poll HALO_STATE and read the chip ID) are all done
bypassed. Only when the amp is seen to be booted is the cache-only
revoked.

Changes are:
- cs35l56_system_reset() now leaves the regmap in cache-only status.

- cs35l56_wait_for_firmware_boot() polls using regmap_read_bypassed().

- cs35l56_init() revokes cache-only either via cs35l56_hw_init() or
  when firmware has rebooted after a soft reset.

- cs35l56_hw_init() exits cache-only after it has determined that the
  amp has booted.

- cs35l56_sdw_init() doesn't disable cache-only, since this must be
  deferred to cs35l56_init().

- cs35l56_runtime_resume_common() waits for firmware boot before exiting
  cache-only.

These changes cover three situations where the registers are not
accessible:

1) SoundWire first-time enumeration. The regmap is kept in cache-only
   until the chip is fully booted. The original code had to exit
   cache-only to read chip status in cs35l56_init() and cs35l56_hw_init()
   but this is now deferred to after the firmware has rebooted.

   In this case cs35l56_sdw_probe() leaves regmap in cache-only
   (unchanged behaviour) and cs35l56_hw_init() exits cache-only after the
   firmware is booted and the chip identified.

2) Soft reset during first-time initialization. cs35l56_init() calls
   cs35l56_system_reset(), which puts regmap into cache-only.
   On I2C/SPI cs35l56_init() then flows through to call
   cs35l56_wait_for_firmware_boot() and exit cache-only. On SoundWire
   the re-enumeration will enter cs35l56_init() again, which then drops
   down to call cs35l56_wait_for_firmware_boot() and exit cache-only.

3) Soft reset after firmware download. dsp_work() calls
   cs35l56_system_reset(), which puts regmap into cache-only. After this
   the flow is the same as (2).
Signed-off-by: default avatarRichard Fitzgerald <rf@opensource.cirrus.com>
Fixes: 8a731fd3 ("ASoC: cs35l56: Move utility functions to shared file")
Link: https://msgid.link/r/20240408101803.43183-4-rf@opensource.cirrus.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      d4884fd4
    • Richard Fitzgerald's avatar
      ALSA: hda: cs35l56: Exit cache-only after cs35l56_wait_for_firmware_boot() · 73580ec6
      Richard Fitzgerald authored 
      Adds calls to disable regmap cache-only after a successful return from
cs35l56_wait_for_firmware_boot().

This is to prepare for a change in the shared ASoC module that will
leave regmap in cache-only mode after cs35l56_system_reset(). This is
to prevent register accesses going to the hardware while it is
rebooting.
Signed-off-by: default avatarRichard Fitzgerald <rf@opensource.cirrus.com>
Link: https://msgid.link/r/20240408101803.43183-3-rf@opensource.cirrus.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      73580ec6
    • Richard Fitzgerald's avatar
      regmap: Add regmap_read_bypassed() · 70ee853e
      Richard Fitzgerald authored 
      Add a regmap_read_bypassed() to allow reads from the hardware registers
while the regmap is in cache-only mode.

A typical use for this is to keep the cache in cache-only mode until
the hardware has reached a valid state, but one or more status registers
must be polled to determine when this state is reached.

For example, firmware download on the cs35l56 can take several seconds if
there are multiple amps sharing limited bus bandwidth. This is too long
to block in probe() so it is done as a background task. The device must
be soft-reset to reboot the firmware and during this time the registers are
not accessible, so the cache should be in cache-only. But the driver must
poll a register to detect when reboot has completed.
Signed-off-by: default avatarRichard Fitzgerald <rf@opensource.cirrus.com>
Fixes: 8a731fd3 ("ASoC: cs35l56: Move utility functions to shared file")
Link: https://msgid.link/r/20240408101803.43183-2-rf@opensource.cirrus.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      70ee853e
    • Hans de Goede's avatar
      ASoC: Intel: bytcr_rt5640: Apply Asus T100TA quirk to Asus T100TAM too · e50729d7
      Hans de Goede authored 
      The Asus T100TA quirk has been using an exact match on a product-name of
"T100TA" but there are also T100TAM variants with a slightly higher
clocked CPU and a metal backside which need the same quirk.

Sort the existing T100TA (stereo speakers) below the more specific
T100TAF (mono speaker) quirk and switch from exact matching to
substring matching so that the T100TA quirk will also match on
the T100TAM models.
Signed-off-by: default avatarHans de Goede <hdegoede@redhat.com>
Link: https://msgid.link/r/20240407191559.21596-1-hdegoede@redhat.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      e50729d7
    • Sameer Pujar's avatar
      ASoC: tegra: Fix DSPK 16-bit playback · 2e93a29b
      Sameer Pujar authored 
      DSPK configuration is wrong for 16-bit playback and this happens because
the client config is always fixed at 24-bit in hw_params(). Fix this by
updating the client config to 16-bit for the respective playback.

Fixes: 327ef647 ("ASoC: tegra: Add Tegra186 based DSPK driver")
Cc: stable@vger.kernel.org
Signed-off-by: default avatarSameer Pujar <spujar@nvidia.com>
Acked-by: default avatarThierry Reding <treding@nvidia.com>
Link: https://msgid.link/r/20240405104306.551036-1-spujar@nvidia.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      2e93a29b
    • Cezary Rojewski's avatar
      ASoC: Intel: avs: Fix debug window description · 7a1625c1
      Cezary Rojewski authored 
      Recent changes addressed PAGE_SIZE ambiguity in 2/3 locations for struct
avs_icl_memwnd2. The unaddressed one causes build errors when
PAGE_SIZE != SZ_4K.
Reported-by: default avatarkernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202404070100.i3t3Jf7d-lkp@intel.com/
Fixes: 275b583d ("ASoC: Intel: avs: ICL-based platforms support")
Signed-off-by: default avatarCezary Rojewski <cezary.rojewski@intel.com>
Link: https://msgid.link/r/20240408081840.1319431-1-cezary.rojewski@intel.comSigned-off-by: default avatarMark Brown <broonie@kernel.org>
      7a1625c1
  19. 07 Apr, 2024 4 commits
  21. 06 Apr, 2024 5 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7