After introducing "samples" to the calculation of wait time, the
driver might timeout at the regmap_field_read_poll_timeout call,
because the wait time could be longer than the 100000 usec limit
due to a large "samples" number.
So this patch sets the timeout limit to 2 times of the wait time
in order to fix this issue.
Fixes: 5c090abf94 ("hwmon: (ina3221) Add averaging mode support")
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Link: https://lore.kernel.org/r/20191022005922.30239-1-nicoleotsuka@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The SDHCI controller on Tegra186 supports 40-bit addressing, which is
usually enough to address all of system memory. However, if the SDHCI
controller is behind an IOMMU, the address space can go beyond. This
happens on Tegra186 and later where the ARM SMMU has an input address
space of 48 bits. If the DMA API is backed by this ARM SMMU, the top-
down IOVA allocator will cause IOV addresses to be returned that the
SDHCI controller cannot access.
Unfortunately, prior to the introduction of the ->set_dma_mask() host
operation, the SDHCI core would set either a 64-bit DMA mask if the
controller claimed to support 64-bit addressing, or a 32-bit DMA mask
otherwise.
Since the full 64 bits cannot be addressed on Tegra, this had to be
worked around in commit 68481a7e1c ("mmc: tegra: Mark 64 bit dma
broken on Tegra186") by setting the SDHCI_QUIRK2_BROKEN_64_BIT_DMA
quirk, which effectively restricts the DMA mask to 32 bits.
One disadvantage of this is that dma_map_*() APIs will now try to use
the swiotlb to bounce DMA to addresses beyond of the controller's DMA
mask. This in turn caused degraded performance and can lead to
situations where the swiotlb buffer is exhausted, which in turn leads
to DMA transfers to fail.
With the recent introduction of the ->set_dma_mask() host operation,
this can now be properly fixed. For each generation of Tegra, the exact
supported DMA mask can be configured. This kills two birds with one
stone: it avoids the use of bounce buffers because system memory never
exceeds the addressable memory range of the SDHCI controllers on these
devices, and at the same time when an IOMMU is involved, it prevents
IOV addresses from being allocated beyond the addressible range of the
controllers.
Since the DMA mask is now properly handled, the 64-bit DMA quirk can be
removed.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
[treding@nvidia.com: provide more background in commit message]
Tested-by: Nicolin Chen <nicoleotsuka@gmail.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Cc: stable@vger.kernel.org # v4.15 +
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
According to the original dma_direct_alloc_pages() code:
{
unsigned int count = PAGE_ALIGN(size) >> PAGE_SHIFT;
if (!dma_release_from_contiguous(dev, page, count))
__free_pages(page, get_order(size));
}
The count parameter for dma_release_from_contiguous() was page
aligned before the right-shifting operation, while the new API
dma_free_contiguous() forgets to have PAGE_ALIGN() at the size.
So this patch simply adds it to prevent any corner case.
Fixes: fdaeec198ada ("dma-contiguous: add dma_{alloc,free}_contiguous() helpers")
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The dma_alloc_contiguous() limits align at CONFIG_CMA_ALIGNMENT for
cma_alloc() however it does not restore it for the fallback routine.
This will result in a size mismatch between the allocation and free
when running into the fallback routines after cma_alloc() fails, if
the align is larger than CONFIG_CMA_ALIGNMENT.
This patch adds a cma_align to take care of cma_alloc() and prevent
the align from being overwritten.
Fixes: fdaeec198ada ("dma-contiguous: add dma_{alloc,free}_contiguous() helpers")
Reported-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tegra186 board under arm64 is using this device, according to
its dts file. So this patch enables its driver with a "=m" as
the other HWMON drivers.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
This patch replaces dma_{alloc,release}_from_contiguous() with
dma_{alloc,free}_contiguous() to simplify those function calls.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Commit fdaeec198ada ("dma-contiguous: add dma_{alloc,free}_contiguous()
helpers") adds a pair of new helper functions so as to abstract code in
the dma-direct (and other places in the future), however it breaks QEMU
boot feature using x86_64 defconfig.
That's because x86_64 defconfig has CONFIG_DMA_CMA=n so those two newly
introduced helper functions are empty in their !CONFIG_DMA_CMA version,
while previously the platform independent dma-direct code had fallback
alloc_pages_node() and __free_pages().
So this patch fixes it by adding alloc_pages_node() and __free_pages()
in the !CONFIG_DMA_CMA version of the two helper functions.
Tested with below QEMU command:
qemu-system-x86_64 -m 512m \
-drive file=images/x86_64/rootfs.ext4,format=raw,if=ide \
-append 'console=ttyS0 root=/dev/sda' -nographic \
-kernel arch/x86_64/boot/bzImage
with the rootfs from the below link:
https://github.com/ClangBuiltLinux/continuous-integration/raw/master/images/x86_64/rootfs.ext4
Fixes: fdaeec198ada ("dma-contiguous: add dma_{alloc,free}_contiguous() helpers")
Reported-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The addresses within a single page are always contiguous, so it's
not so necessary to always allocate one single page from CMA area.
Since the CMA area has a limited predefined size of space, it may
run out of space in heavy use cases, where there might be quite a
lot CMA pages being allocated for single pages.
However, there is also a concern that a device might care where a
page comes from -- it might expect the page from CMA area and act
differently if the page doesn't.
This patch tries to use the fallback alloc_pages path, instead of
one-page size allocations from the global CMA area in case that a
device does not have its own CMA area. This'd save resources from
the CMA global area for more CMA allocations, and also reduce CMA
fragmentations resulted from trivial allocations.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: dann frazier <dann.frazier@canonical.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Both dma_alloc_from_contiguous() and dma_release_from_contiguous() are
very simply implemented, but requiring callers to pass certain
parameters like count and align, and taking a boolean parameter to check
__GFP_NOWARN in the allocation flags. So every function call duplicates
similar work:
unsigned long order = get_order(size);
size_t count = size >> PAGE_SHIFT;
page = dma_alloc_from_contiguous(dev, count, order,
gfp & __GFP_NOWARN);
[...]
dma_release_from_contiguous(dev, page, size >> PAGE_SHIFT);
Additionally, as CMA can be used only in the context which permits
sleeping, most of callers do a gfpflags_allow_blocking() check and a
corresponding fallback allocation of normal pages upon any false result:
if (gfpflags_allow_blocking(flag))
page = dma_alloc_from_contiguous();
if (!page)
page = alloc_pages();
[...]
if (!dma_release_from_contiguous(dev, page, count))
__free_pages(page, get_order(size));
So this patch simplifies those function calls by abstracting these
operations into the two new functions: dma_{alloc,free}_contiguous.
As some callers of dma_{alloc,release}_from_contiguous() might be
complicated, this patch just implements these two new functions to
kernel/dma/direct.c only as an initial step.
Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: dann frazier <dann.frazier@canonical.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Build warning being reported:
sound/soc/fsl/fsl_sai.c: In function 'fsl_sai_remove':
sound/soc/fsl/fsl_sai.c:921:1: warning: no return statement in
function returning non-void [-Wreturn-type]
So this patch just adds a "return 0" to fix it.
Fixes: 812ad463e0 ("ASoC: fsl_sai: Add support for runtime pm")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The CONFIG register has two 3-bit fields for conversion time
settings of Bus-voltage and Shunt-voltage, respectively. The
conversion settings, along with averaging mode, allow users
to optimize available timing requirement.
This patch adds an 'update_interval' sysfs node through the
hwmon_chip_info of hwmon core. It reflects a total hardware
conversion time:
samples * channels * (Bus + Shunt conversion times)
Though INA3221 supports different conversion time setups for
Bus and Shunt voltages, this patch only adds the support of
a unified setting for both conversion times, by dividing the
conversion time into two equal values.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
[groeck: .rst related formatting changes in documentation]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reading back the CONFIG register increases an extra I2C
transaction. This's not necessary and could be replaced
with a local variable caching the register settings.
So this patch replaces two readback regmap_read() calls
with a tmp variable.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The CONFIG register has a 3-bit averaging mode field for users
to setup the number of samples that are collected and averaged
together. This is very useful to filter noise from sensor data.
This patch adds a 'samples' sysfs node using hwmon_chip_samples
of hwmon core, and updates wait time calculation by taking this
samples value into account.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
By default, ina3221, as a hardware monitor, continuously measures
the inputs and generates corresponding data. However, for battery
powered devices, this mode might be power consuming.
The DT binding doc is updated with a new boolean type property to
allow changing the default operating mode from consuming mode to
single-shot mode, which will measure input on demand and then shut
down to save power.
So this patch implements the DT property accordingly.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
By default, ina3221, as a hardware monitor, continuously measures
the inputs and generates corresponding data. However, for battery
powered devices, this mode might be power consuming.
This patch adds a "ti,single-shot" property to allow changing the
default continuous mode to single-shot operating mode.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
There are a few sysfs entries being exposed to user space by the
ina2xx hwmon driver while not getting explicitly documented. So
this patch just adds a description section for them.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
According to the code right before the removed line, data->kind
should be either from DT or from id pointer. So there shouldn't
be an additional overwriting after the if-else statement.
So this patch just removes the overwriting line.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
If all three channels are disabled via in[123]_enable ABI,
the driver could suspend the chip for power saving purpose.
So this patch adds the PM runtime support in order to gain
more power control than system suspend and resume use case.
For PM runtime, there are a few related changes happening:
1) Added a new pm_dev device pointer for all the PM runtime
callbacks. This is because hwmon core registers a child
device for each hwmon driver and passes it back to each
driver. So there might be a mismatch between two device
pointers in the driver if mixing using them.
2) Added a check in ina3221_is_enabled() to make sure that
the chip is resumed.
3) Bypassed the unchanged status in ina3221_write_enable()
in order to keep the PM runtime refcount being matched.
4) Removed the reset routine in the probe() by calling the
resume() via pm_runtime_get_sync() instead, as they're
similar. It's also necessary to do so to match initial
PM refcount with the number of enabled channels.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The data might need some time to get ready after channel enabling,
although the data register is always readable. The CVRF bit is to
indicate that data conversion is finished, so polling the CVRF bit
before data reading could ensure the result being valid.
An alternative way could be to wait for expected time between the
channel enabling and the data reading. And this could avoid extra
I2C communications. However, INA3221 seemly takes longer time than
what's stated in the datasheet. Test results show that sometimes
it couldn't finish data conversion in time.
So this patch plays safe by adding a CVRF polling to make sure the
data register is updated with the new data.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
This change adds a mutex to serialize accesses of sysfs attributes.
This is required when polling CVRF bit of the MASK/ENABLE register
because this bit is cleared on a read of this MASK/ENABLE register
or a write to CONFIG register, which means that this bit might be
accidentally cleared by reading other fields like alert flags.
So this patch adds a mutex lock to protect the write() and read()
callbacks. The read_string() callback won't need the lock since it
just returns the label without touching any hardware register.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
There is nothing critically wrong to read these two attributes
without having a is_enabled() check at this point. But reading
the MASK_ENABLE register would clear the CVRF bit according to
the datasheet. So it'd be safer to fence for disabled channels
in order to add pm runtime feature.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The current register (04h) has a sign bit at MSB. The comments
for this calculation also mention that it's a signed register.
However, the regval is unsigned type so result of calculation
turns out to be an incorrect value when current is negative.
This patch simply fixes this by adding a casting to s16.
Fixes: 5d389b1251 ("hwmon: (ina2xx) Make calibration register value fixed")
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
When using DT configurations, the id pointer might turn out to
be NULL. Then the driver encounters NULL pointer access:
Unable to handle kernel read from unreadable memory at vaddr 00000018
[...]
PC is at ina2xx_probe+0x114/0x200
LR is at ina2xx_probe+0x10c/0x200
[...]
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
The reason is that i2c core returns the id pointer by matching
id_table with client->name, while the client->name is actually
using the name from the first string in the DT compatible list,
not the best one. So i2c core would fail to match the id_table
if the best matched compatible string isn't the first one, and
then would return a NULL id pointer.
This probably should be fixed in i2c core. But it doesn't hurt
to make the driver robust. So this patch fixes it by using the
"chip" that's added to unify both DT and non-DT configurations.
Additionally, since id pointer could be null, so as id->name:
ina2xx 10-0047: power monitor (null) (Rshunt = 1000 uOhm)
ina2xx 10-0048: power monitor (null) (Rshunt = 10000 uOhm)
So this patch also fixes NULL name pointer, using client->name
to play safe and to align with hwmon->name.
Fixes: bd0ddd4d08 ("hwmon: (ina2xx) Add OF device ID table")
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Trace events are useful for people who collect data from the
Ftrace outputs. There're people who analyse the relationship
of cpufreq, thermal and hwmon (power/voltage/current) using
the convenient and timestamped Ftrace outputs, while unlike
cpufreq and thermal subsystems the hwmon does not have trace
events supported yet.
So this patch adds initial trace events for the hwmon core.
To call hwmon_attr_base() for aligned attr index numbers, it
also moves the function upward.
Ftrace outputs:
...: hwmon_attr_show_string: index=2, attr_name=in2_label, val=VDD_5V
...: hwmon_attr_show: index=2, attr_name=in2_input, val=5112
...: hwmon_attr_show: index=2, attr_name=curr2_input, val=440
Note that the _attr_show and _attr_store functions are tied
to the _with_info API. So a hwmon driver requiring the trace
events feature should use _with_info API to register a hwmon
device.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The hwmon core has a newer API which abstracts most of common
things in the core so as to simplify the hwmon device drivers.
This patch implements this _info API to ina3221 hwmon driver.
It also reduces the binary size:
text data bss dec hex filename
5114 1712 0 6826 1aaa drivers/hwmon/ina3221_before.o
4456 440 0 4896 1320 drivers/hwmon/ina3221.o
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The input->shunt_resistor is int type while the value from DT is
unsigned int. Meanwhile, a divide-by-zero error would happen if
the value is 0. So this patch just simply validates the value.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
According to hwmon ABI, in%d_enable is a sysfs interface that
allows user space to enable and disable the input sensor. So
this patch just simply adds the attribute to the list.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
An ina3221 chip has three input ports. Each port is used
to measure the voltage and current of its input source.
The DT binding now has defined bindings for their input
sources, so the driver should read these information and
handle accordingly.
This patch adds a new structure of input source specific
information including input source label, shunt resistor
value and its connection status. It exposes these labels
via in[123]_label sysfs nodes upon available, and also
disables those channels where there are no input source
being connected. Meanwhile, it also adds in[123]_enable
sysfs nodes for users to get control of three channels,
and returns -ENODATA code for any sensor read according
to hwmon ABI.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Texas Instruments INA3221 is a triple-channel shunt and bus
voltage monitor. This patch adds a DT binding doc for it.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Depending on the hardware design, an INA3221 chip might lose
its power during system suspend/resume. So this patch adds
a set of suspend and resume functions to cache the register
values including config register value and limit settings.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
[groeck: Moved call to dev_set_drvdata()]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The three INA3221_CONFIG_MODE macros are not correctly defined here.
The MODE3-1 bits are located at BIT 2-0 according to the datasheet.
So this patch just fixes them by shifting all of them with a correct
offset. However, this isn't a crital bug fix as the driver does not
use any of them at this point.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The MSB (15th bit) of INA3221_CONFIG is a self-clear reset bit.
So this register should be added to the volatile_table of the
regmap_config. Otherwise, we will see this bit is sticky in the
regcache which might accidentally reset the chip when an actual
write happens to the register.
This might not be a severe bug for the current code line since
there's no second place touching the INA3221_CONFIG except the
reset routine in the probe().
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The freq parameter indicates the physical frequency of an actual
input clock or a desired frequency of an output clock for HCKT/R.
It should never be passed 0. This might cause Division-by-zero.
So this patch adds a check to fix it.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
This is a partial revert (in a cleaner way) of commit ebf08ae3bc
("ASoC: fsl_ssi: Keep ssi->i2s_net updated") to fix a regression
at test cases when switching between mono and stereo audio.
The problem is that ssi->i2s_net is initialized in set_dai_fmt()
only, while this set_dai_fmt() is only called during the dai-link
probe(). The original patch assumed set_dai_fmt() would be called
during every playback instance, so it failed at the overriding use
cases.
This patch adds the local variable i2s_net back to let regular use
cases still follow the mode settings from the set_dai_fmt().
Meanwhile, the original commit of keeping ssi->i2s_net updated was
to make set_tdm_slot() clean by checking the ssi->i2s_net directly
instead of reading SCR register. However, the change itself is not
necessary (or even harmful) because the set_tdm_slot() might fail
to check the slot number for Normal-Mode-None-Net settings while
mono audio cases still need 2 slots. So this patch can also fix it.
And it adds an extra line of comments to declare ssi->i2s_net does
not reflect the register value but merely the initial setting from
the set_dai_fmt().
Reported-by: Mika Penttilä <mika.penttila@nextfour.com>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Mika Penttilä <mika.penttila@nextfour.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When the desired ratio is less than 256, the savesub (tolerance)
in the calculation would become 0. This will then fail the loop-
search immediately without reporting any errors.
But if the ratio is smaller enough, there is no need to calculate
the tolerance because PM divisor alone is enough to get the ratio.
So a simple fix could be just to set PM directly instead of going
into the loop-search.
Reported-by: Marek Vasut <marex@denx.de>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
This patch cleans up probe() function by moving all Device Tree
related code into a separate function. It allows the probe() to
be Device Tree independent. This will be very useful for future
integration of imx-ssi driver which has similar functionalities
while exists only because it supports non-DT cases.
This patch also moves symmetric_channels of AC97 from the probe
to the structure snd_soc_dai_driver for simplification.
Additionally, since PowerPC and AC97 use the same pdev pointer
to register a platform device, this patch also unifies related
code.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The _fsl_ssi_set_dai_fmt() is a helper function being called from
fsl_ssi_set_dai_fmt() as an ASoC operation and fsl_ssi_hw_init()
mainly for AC97 format initialization.
This patch cleans the _fsl_ssi_set_dai_fmt() in following ways:
* Removing *dev pointer in the parameters as it's included in the
*ssi pointer of struct fsl_ssi.
* Using regmap_update_bits() instead of regmap_read() with masking
the value manually.
* Moving baudclk check to the switch-case routine to skip the I2S
master check. And moving SxCCR.DC settings after baudclk check.
* Adding format settings for SND_SOC_DAIFMT_AC97 like others.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
AC97 configures most of registers earlier to start a communication
with CODECs in order to successfully initialize CODEC. Currently,
_fsl_ssi_set_dai_fmt() and fsl_ssi_setup_ac97() are called to get
all SSI registers properly set.
Since now the driver has a fsl_ssi_hw_init() to handle all register
initial settings, this patch moves those register settings of AC97
to the fsl_ssi_hw_init() as well.
Meanwhile it applies _fsl_ssi_set_dai_fmt() call to AC97 only since
other formats would be configured via normal set_dai_fmt() directly.
This patch also adds fsl_ssi_hw_clean() to cleanup control bits for
AC97 in the platform remote() function.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The _fsl_ssi_set_dai_fmt() bypasses an undefined format for AC97
mode. However, it's not really necessary if AC97 has its complete
format defined.
So this patch adds a DAIFMT macro of complete format including a
clock direction and polarity.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The trigger() calls fsl_ssi_tx_config() and fsl_ssi_rx_config(),
and both of them jump to fsl_ssi_config(). And fsl_ssi_config()
later calls another fsl_ssi_rxtx_config().
However, the whole routine, especially fsl_ssi_config() function,
is too complicated because of the folowing reasons:
1) It has to handle the concern of the opposite stream.
2) It has to handle cases of offline configurations support.
3) It has to handle enable and disable operations while they're
mostly different.
Since the enable and disable routines have more differences than
TX and RX rountines, this patch simplifies these helper functions
with the following changes:
- Changing to two helper functions of enable and disable instead
of TX and RX.
- Removing fsl_ssi_rxtx_config() by separately integrating it to
two newly introduced enable & disable functions.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The FIFO clear helper function is just one line of code now.
So it could be cleaned up by removing it and calling regmap
directly.
Meanwhile, FIFO clear could be applied to all use cases, not
confined to AC97. So this patch also moves FIFO clear in the
trigger() to fsl_ssi_config() and removes the AC97 check.
Note that SOR register is safe from offline_config HW limit.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The define of fsl_ssi_disable_val is not so clear as it mixes two
steps of calculations together. And those parameter names are also
a bit long to read.
Since it just tries to exclude the shared bits from the regvals of
current stream while the opposite stream is active, it's better to
use something like ssi_excl_shared_bits.
This patch also bisects fsl_ssi_disable_val into two macros of two
corresponding steps and then shortens its parameter names. It also
updates callers in the fsl_ssi_config() accordingly.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
Checking TE and RE bits in SCR register doesn't work for AC97 mode
which enables SSIEN, TE and RE in the fsl_ssi_setup_ac97() that's
called during probe().
So when running into the trigger(), it will always get the result
of both TE and RE being enabled already, even if actually there is
no active stream.
This patch fixes this issue by adding a variable to log the active
streams manually.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch replaces the register read with ssi->i2s_net for
simplification. It also removes masking SSIEN from scr value
since it's handled later by regmap_update_bits() to set this
scr value back.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The hw_params() overwrites i2s_net settings for special cases like
mono-channel support, however, it doesn't update ssi->i2s_net as
set_dai_fmt() does.
This patch removes the local i2s_net variable and directly updates
ssi->i2s_net in the hw_params() so that the driver can simply look
up the ssi->i2s_net instead of reading the register.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
The RX and TX macros were defined implicitly and there was
a potential risk if someone changes their values.
Since they were defined to index the array ssi->regvals[2],
this patch moves these two macros to fsl_ssi.c, closer to
its owner ssi->regvals. And it also puts some comments here
to limit their value within [0, 1].
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Caleb Crome <caleb@crome.org>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch just simply moves tegra_thermctl_set_trip_temp() behind
those function implementations so that it can remove those forward
declarations.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
The name fsl_ssi_rxtx_reg_val is too long to read comfortably.
So this patch shortens it by using an array (fsl_ssi_regvals,
renamed from fsl_ssi_reg_val). To do that, it also introduces
two macros (TX and RX) to replace the wrapper structure. This
will also help further cleanups.
Meanwhile, it unifies all local variable with the name "vals"
to get rid of the name "reg" -- could be confusing with "regs"
in the private struct for regmap.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Acked-by: Timur Tabi <timur@tabi.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch refines the comments by:
1) Removing all out-of-date comments
2) Removing all not-so-useful comments
3) Unifying the styles of all comments
4) Shortening comments to be more conise
5) Adding comments to improve code readablity
6) Moving all register related comments to fsl_ssi.h
7) Adding comments to all register and field defines
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Tested-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Reviewed-by: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
Acked-by: Timur Tabi <timur@tabi.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The machine driver links both imx-ssi (legacy non-DT driver) and
fsl_ssi (up-to-date DT based driver). So It also includes both
imx-ssi.h and fsl_ssi.h header files. This creates a limitation
for two header files -- they can't define anything with identical
names.
Since the eukrea-tlv320 machine driver now does not really need
anything being defined in the fsl_ssi header file, and it's also
going to take some time to clean up two SSI drivers, this patch
takes a quick way to remove the #include fsl_ssi.h line for now.
It can be added back once the header files are done refactoring.
The eukrea-tlv320 driver is still compiled successfully without
any erorr using imx_v6_v7_defconfig, after removing it.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Below is the call trace of tegra210_init_pllu() function:
start_kernel()
-> time_init()
--> of_clk_init()
---> tegra210_clock_init()
----> tegra210_pll_init()
-----> tegra210_init_pllu()
Because the preemption is disabled in the start_kernel before calling
time_init, tegra210_init_pllu is actually in an atomic context while
it includes a readl_relaxed_poll_timeout that might sleep.
So this patch just changes this readl_relaxed_poll_timeout() to its
atomic version.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Acked-By: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Both tegra124-dfll and clk-dfll are using platform_set_drvdata
to set drvdata of the exact same pdev while they use different
pointers for the drvdata. Once the drvdata has been overwritten
by tegra124-dfll, clk-dfll will never get its td pointer as it
expects.
Since tegra124-dfll merely needs its soc pointer in its remove
function, this patch fixes the bug by removing the overwriting
in the tegra124-dfll file and letting the tegra_dfll_unregister
return an soc pointer for it.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Acked-By: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The dev pointer is going through a null check after a dereference.
So this patch removes that useless check since the driver does not
pass a null dev pointer in any case.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
The driver already supports DMA_DEV_TO_DEV in sdma_config(),
DMA_SLAVE_BUSWIDTH_2_BYTES and DMA_SLAVE_BUSWIDTH_1_BYTE in
sdma_prep_slave_sg(). So this patch adds them to the lists.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The snd_soc_component_set_sysclk() and snd_soc_dai_set_tdm_slot()
in the soc-core.c will return -ENOTSUPP if there is no function
implementation for them in the dai and component drivers.
So this patch tries to ignore this errno.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The set_sysclk() now is used to override the output bit clock rate.
But this is not a common way to implement a set_dai_sysclk(). And
this creates a problem when a general machine driver (simple-card
for example) tries to do set_dai_sysclk() by passing an input clock
rate for the baud clock instead of setting the bit clock rate as
fsl_ssi driver expected.
So this patch solves this problem by firstly removing set_sysclk()
since the hw_params() can calculate the bit clock rate. Secondly,
in order not to break those TDM use cases which previously might
have been using set_sysclk() to override the bit clock rate, this
patch changes the driver to calculate the bit clock rate using the
slot number and the slot width from the via set_tdm_slot().
The patch also removes an obsolete comment of the dir parameter.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
According to the datasheets of Realtek ALC5658 (Rev. 0.9)
and ALC5659 (Rev. 0.13), the "IRQ Control 4" register is
located at MX-00BAh while the "IRQ control 5" register is
at MX-00BBh.
There is no "IRQ Control 6" register in the datasheets but
it is supposed to be behind the "IRQ Control 5" register.
So this patch corrects these addresses.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The DMA_SG is still a type of memory copy operation that should
conform the hardware restriction. So this patch just applies the
copy_align to DMA_SG as well.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Since it's possible to have an IRQ without enabling JD3, the
GPIO1 pin then would remain its default GPIO function which
is set as an input direction (seeing from rt5659). Meanwhile,
CPU would also listen this connection via its own GPIO input:
[input] [input]
CPU GPIO <--------> RT5659 GPIO1
The result for the IRQ on the CPU side will be unexpectable.
So this patch enables the IRQ output for GPIO1 pin any way
as long as there's an IRQ assigned from platform data or DT:
[input] [IRQ output]
CPU GPIO <--------< RT5659 GPIO1
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The SND_SOC_DAPM_PRE_PMU case would call startup()/hw_params() that
might access substream->runtime through other functions.
For example:
Unable to handle kernel NULL pointer dereference at virtual address
[....]
PC is at snd_pcm_hw_rule_add+0x24/0x1b0
LR is at snd_pcm_hw_constraint_list+0x20/0x28
[....]
Process arecord (pid: 424, stack limit = 0xffffffc1ecaf0020)
Call trace:
[<ffffffc00086be68>] snd_pcm_hw_rule_add+0x24/0x1b0
[<ffffffc00086c014>] snd_pcm_hw_constraint_list+0x20/0x28
[<ffffffc0008b47a4>] cs53l30_pcm_startup+0x24/0x30
[<ffffffc0008a6260>] snd_soc_dai_link_event+0x290/0x354
[<ffffffc0008a7528>] dapm_seq_check_event.isra.31+0x134/0x2c8
[<ffffffc0008a7768>] dapm_seq_run_coalesced+0x94/0x1c8
[<ffffffc0008a7940>] dapm_seq_run+0xa4/0x404
[<ffffffc0008a8bac>] dapm_power_widgets+0x524/0x984
[<ffffffc0008ab1c4>] snd_soc_dapm_stream_event+0x8c/0xa8
[<ffffffc0008ac7f4>] soc_pcm_prepare+0x10c/0x1ec
[<ffffffc000865b9c>] snd_pcm_do_prepare+0x1c/0x38
[<ffffffc000865600>] snd_pcm_action_single+0x40/0x88
[<ffffffc0008656b8>] snd_pcm_action_nonatomic+0x70/0x90
[<ffffffc000868d28>] snd_pcm_common_ioctl1+0xb6c/0xdd8
[<ffffffc000869508>] snd_pcm_capture_ioctl1+0x200/0x334
[<ffffffc00086a084>] snd_pcm_ioctl_compat+0x648/0x95c
[<ffffffc0001ff4b4>] compat_SyS_ioctl+0xac/0xfc4
[<ffffffc000084cf0>] el0_svc_naked+0x24/0x28
---[ end trace 0dc4f99c2759c35c ]---
So this patch adds a dummy runtime for the original dummy substream
to merely avoid the NULL pointer access.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The TDM mode using PCM format now has two-bit right shift due to the
format configuration in the driver. According to Figure 4-13 in the
CS53L30 datasheet, using ASP_SCLK_INV = 0 and SHIFT_LEFT = 1 should
be the correct combination to create one-bit right shift for the DSP
type A format.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The maximum slot number of CS53L30 is 4 while it should support
the situation that's less than 4 channels based on the rx_mask.
So when the driver validates the last slot location, it should
check the last active slot instead of always the 4th one.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The codec chip has a physical MUTE pin to let users control it via
GPIO. So this patch add a mute control support to the driver.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Acked-by: Paul Handrigan <Paul.Handrigan@cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The driver is using the set_bias_level to control the power on and
off so it should get SND_SOC_BIAS_OFF in order to proceed normal
powering sequences. This patch enables the idle_bias_off option so
the DAPM core will set the bias level to SND_SOC_BIAS_OFF instead
of stopping at SND_SOC_BIAS_STANDBY.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
SND_SOC_DAIFMT_IB_NF = 0x3 (11b) | SND_SOC_DAIFMT_IB_IF = 0x4 (100b)
creates a mask 0x7 (111b) which also includes SND_SOC_DAIFMT_NB_IF
= 0x2 (10b).
So this patch uses the traditional way to check the clock inversion.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
CS53L30 is a Quad-Channel ADC from Cirrus Logic with an I2S/TDM DAI.
So this patch adds a codec driver for CS53L30 that includes 4-channel
24-bit recording and TDM mode supports.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The dev->name of CODEC might not be identical to its codec_dai_name,
so using dev->name to probe the CODEC dai is not a correct approach.
This patch specifies each supporting codec_dai_name instead of using
dev->name any more.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
As commit 6c8ca30eec ("ASoC: fsl_ssi: Don't try to round-up for PM
divisor calculation") mentioned that there's no more need to use a
round up work around to get a better divisor since the clk-divider
driver has been refined a lot.
So this patch applies the same modification to fsl_spdif driver.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
According to i.MX6 Series Reference Manual, the formula to calculate
the sys clock is
sysclk rate = bclk rate * (div2 + 1) * (7 * psr + 1) * (pm + 1) * 2
Commit aafa85e71a ("ASoC: fsl_ssi: Add DAI master mode support for
SSI on i.MX series") added the divisor calculation which relies on
the clk_round_rate(). However, at that time, clk_round_rate() didn't
provide closest clock rates for some cases because it might not use
a correct rounding policy. So using the original formula (pm + 1) for
PM divisor was not able to give us a desired clock rate. And then we
used (pm + 2) to do the trick.
However, the clk-divider driver has been refined a lot since commit
b11d282dbe ("clk: divider: fix rate calculation for fractional rates")
Now using (pm + 2) trick would result an incorrect clock rate.
So this patch fixes the problem by removing the useless trick.
Reported-by: Stephane Cerveau <scerveau@voxtok.com>
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch adds snd_soc_of_parse_audio_routing() to get dapm routes
configurations via Device Tree.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The current helper functions, snd_soc_of_parse_audio_simple_widgets()
and snd_soc_of_parse_audio_routing(), set dapm_widgets and dapm_routes
without caring if they are already set by using build-in widgets and
routes in the card driver. So there could be one of them, build-in one
or Device Tree one, overrided by the other depending on which one was
assigned later.
This patch adds an extra pair of dapm_widgets and dapm_routes for DT
use only so as to prevent unexpected overriding.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
According to i.MX Reference Manual, the bit-clock frequency generated
by SSI must be never greater than 1/5 of the peripheral clock frequency.
This peripheral clock, however, is not baudclk but the IPG clock (i.e.
ssi_private->clk in the fsl_ssi driver).
So this patch just simply fixes the incorrect limitation applied to
the bit clock (baudclk) rate.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
ESAI may need to use fixed slot width to comply with external
CODEC. So this set_tdm_slot() call will allow the ESAI driver
to override its default dynamic slot width policy.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The driver previously used 32-bit fixed slot width as default.
In result, ESAI might use 32-bit length to capture 16-bit width
audio slot from CODEC side when ESAI is running as DAI slave.
So this patch just removes the default slot_width so as to use
dynamic slot width. If there comes a specific situation that
needs a fixed width, the machine driver shall set slot_width
via set_tdm_slot() so as to let the ESAI driver replace the
dynamic width policy with the fixed value.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When cpu-dai is the DAI Master (CBM_CFx), it may need some configurations,
set_sysclk() call for eample, for cpu-dai side in the hw_params(), even if
the set_bias_level() has already taken care of the codec-dai side.
So this patch just simply adds an additional condition.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Kernel dump (WARN_ON) ocurred during system boot-up inside regmap_write():
------------[ cut here ]------------
WARNING: CPU: 0 PID: 47 at kernel/locking/lockdep.c:2744 lockdep_trace_alloc+0xe8/0x108()
DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags))
Modules linked in:
CPU: 0 PID: 47 Comm: kworker/u2:2 Not tainted 3.18.0-rc1-10245-gb75d289-dirty #56
Workqueue: deferwq deferred_probe_work_func
Backtrace:
[<80012294>] (dump_backtrace) from [<80012578>] (show_stack+0x18/0x1c)
r6:8097c73c r5:8097c73c r4:00000000 r3:be33ba80
[<80012560>] (show_stack) from [<806aac48>] (dump_stack+0x8c/0xa4)
[<806aabbc>] (dump_stack) from [<8002a694>] (warn_slowpath_common+0x70/0x94)
r6:80062838 r5:00000009 r4:bd827b30 r3:be33ba80
[<8002a624>] (warn_slowpath_common) from [<8002a6f0>] (warn_slowpath_fmt+0x38/0x40)
r8:00000004 r7:00000001 r6:000080d0 r5:60000193 r4:bd826010
[<8002a6bc>] (warn_slowpath_fmt) from [<80062838>] (lockdep_trace_alloc+0xe8/0x108)
r3:80831590 r2:8082e160
[<80062750>] (lockdep_trace_alloc) from [<800ea5dc>] (kmem_cache_alloc+0x28/0x134)
r5:000080d0 r4:be001f00
[<800ea5b4>] (kmem_cache_alloc) from [<8038d72c>] (regcache_rbtree_write+0x15c/0x648)
r10:00000000 r9:0000001c r8:00000004 r7:00000001 r6:00000000 r5:bd819a00
r4:00000000 r3:811aea88
[<8038d5d0>] (regcache_rbtree_write) from [<8038c4d8>] (regcache_write+0x5c/0x64)
r10:be3f9f88 r9:00000000 r8:00000004 r7:00000001 r6:00000000 r5:00000001
r4:bd819a00
[<8038c47c>] (regcache_write) from [<8038b0dc>] (_regmap_raw_write+0x134/0x5f4)
r6:be3f9f84 r5:00000001 r4:bd819a00 r3:00000001
[<8038afa8>] (_regmap_raw_write) from [<8038b610>] (_regmap_bus_raw_write+0x74/0x94)
r10:00000000 r9:00000001 r8:be3fb080 r7:bd819a00 r6:00000001 r5:00000000
r4:bd819a00
[<8038b59c>] (_regmap_bus_raw_write) from [<8038a8b4>] (_regmap_write+0x60/0x9c)
r6:00000001 r5:00000000 r4:bd819a00 r3:8038b59c
[<8038a854>] (_regmap_write) from [<8038ba24>] (regmap_write+0x48/0x68)
r7:bd81ad80 r6:00000001 r5:00000000 r4:bd819a00
[<8038b9dc>] (regmap_write) from [<80528f30>] (fsl_asrc_dai_probe+0x34/0x104)
r6:bd888628 r5:be3fb080 r4:be3b4410 r3:be3b442c
------------[ dump end ]------------
=============================================================================
2741 /*
2742 * Oi! Can't be having __GFP_FS allocations with IRQs disabled.
2743 */
2744 if (DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags)))
2745 return;
=============================================================================
By looking at 2744 line, we can get that it's because regcache_rbtree_write()
would call kmalloc() with GFP flag if it couldn't find an existing block to
insert nodes while this kmalloc() call is inside a spin_lock_irq_save pair,
i.e. IRQs disabled.
Even though this may be a bug that should be fixed, I still try to send this
patch as a quick fix (work around) since it does no harm to assign default
values of every registers when using regcache.
Signed-off-by: Nicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Nicolin Chen