Otherwise the change isn't atomic.
Signed-off-by: Christian König <christian.koenig@amd.com>
CC: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Otherwise it is possible that we will have page table corruption
if we change a BOs address multiple times.
Signed-off-by: Christian König <christian.koenig@amd.com>
CC: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If we unmap BOs before releasing them them the intervall tree locks
up because we try to remove an entry not inside the tree.
Based on a patch from Michel Dänzer.
Signed-off-by: Christian König <christian.koenig@amd.com>
CC: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Create a cpumask that limits the affinity of all unbound workqueues.
This cpumask is controlled through a file at the root of the workqueue
sysfs directory.
It works on a lower-level than the per WQ_SYSFS workqueues cpumask files
such that the effective cpumask applied for a given unbound workqueue is
the intersection of /sys/devices/virtual/workqueue/$WORKQUEUE/cpumask and
the new /sys/devices/virtual/workqueue/cpumask file.
This patch implements the basic infrastructure and the read interface.
wq_unbound_cpumask is initially set to cpu_possible_mask.
Cc: Christoph Lameter <cl@linux.com>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Mike Galbraith <bitbucket@online.de>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Current apply_workqueue_attrs() includes pwqs-allocation and pwqs-installation,
so when we batch multiple apply_workqueue_attrs()s as a transaction, we can't
ensure the transaction must succeed or fail as a complete unit.
To solve this, we split apply_workqueue_attrs() into three stages.
The first stage does the preparation: allocation memory, pwqs.
The second stage does the attrs-installaion and pwqs-installation.
The third stage frees the allocated memory and (old or unused) pwqs.
As the result, batching multiple apply_workqueue_attrs()s can
succeed or fail as a complete unit:
1) batch do all the first stage for all the workqueues
2) only commit all when all the above succeed.
This patch is a preparation for the next patch ("Allow modifying low level
unbound workqueue cpumask") which will do a multiple apply_workqueue_attrs().
The patch doesn't have functionality changed except two minor adjustment:
1) free_unbound_pwq() for the error path is removed, we use the
heavier version put_pwq_unlocked() instead since the error path
is rare. this adjustment simplifies the code.
2) the memory-allocation is also moved into wq_pool_mutex.
this is needed to avoid to do the further splitting.
tj: minor updates to comments.
Suggested-by: Tejun Heo <tj@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Mike Galbraith <bitbucket@online.de>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
The polling mode of the driver is designed for transfers that run
less than 30us - it will only execute under those circumstances.
So it should run comfortably without getting interrupted by the
scheduler.
But there are situations where the raspberry pi is so overloaded
that it can take up to 80 jiffies until the polling thread gets
rescheduled - this has been observed especially under heavy
IO situations.
In such a situation we now fall back to the interrupt handler and
log the situation at debug level.
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The way that the timeout code is written in the polling function
the timeout does also trigger when interrupted or rescheduled while
in the polling loop.
This patch changes the timeout from effectively 20ms (=2 jiffies) to
1 second and removes the time that the transfer really takes out of
the computation, as - per design - this is <30us and the jiffie resolution
is 10ms so that does not make any difference what so ever.
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The internal chip select CS0 wasn't initialized properly to work with
CS HIGH chips.
Signed-off-by: Alban Bedel <albeu@free.fr>
Signed-off-by: Mark Brown <broonie@kernel.org>
Clocks should be prepared and unprepared, fix this by using
clk_prepare_enable() and clk_disable_unprepare() instead of
clk_enable() and clk_disable().
Signed-off-by: Alban Bedel <albeu@free.fr>
Signed-off-by: Mark Brown <broonie@kernel.org>
Set the OF node of the spi controller and use the generic GPIO based
chip select instead of the custom controller data. As the controller
data isn't used by any board just drop it.
Signed-off-by: Alban Bedel <albeu@free.fr>
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix a typo in the TX DMA interrupt name for AUART4.
This patch makes AUART4 operational again.
Signed-off-by: Marek Vasut <marex@denx.de>
Fixes: f30fb03d4d ("ARM: dts: add generic DMA device tree binding for mxs-dma")
Cc: stable@vger.kernel.org
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Added entry in sst driver to support rt5650 codec
for intel Braswell platform.
Signed-off-by: Fang, Yang A <yang.a.fang@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
we have defined SNDRV_CTL_ELEM_ID_NAME_MAXLEN as size of name array so use
this define instead of numeric value
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
kbuild robot reports following warning:
"sound/soc/intel/haswell/sst-haswell-ipc.c:2204:1-6:
WARNING: invalid free of devm_ allocated data"
As julia explains to me, the memory allocated with devm_kalloc
is freed automatically on failure of a probe function. So this
kfree should be removed otherwise the double free will be got in
error handler path.
Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently LO_VOL_* stays at it's default (0x4 each) but this should be
calculated after setting VAG_VAL and LO_VAGCNTRL.
LO_VOL_* = 40 * log10(VAG_VAL / LO_VAGCNTRL) + 15
To avoid the log10 operation a table for all valid register values is
precalculated which contains the corresponding value
(VAG_VAL * 100 / LO_VAGCNTRL).
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
This is a preparation for calculating lo_vol which needs both vag
and lo_vag.
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The property '#pwm-cells' is currently missing. It is not possible to
use pwm4 without this property.
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Fixes: 5658a68fb5 ("ARM i.MX25: Add devicetree")
Cc: <stable@vger.kernel.org>
Reviewed-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Remove including <linux/version.h> that don't need it.
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: Brian Austin <brian.austin@cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
According to the schematics the both LHPOUT and RHPOUT are connected to the
external connector. RHPOUT is missing from the DAPM routes, but otherwise
they seem to be complete.
This patch adds the missing route and then sets the fully_routed flag for
the card. This allows to remove all the manual calls to
snd_soc_dapm_nc_pin().
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The DSP control information contains various hints about the usage of
the control use these when handling the control.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Using uXX for little endian data, was triggering some warnings through
sparse:
sound/soc/codecs/wm_adsp.c:716:26: sparse: cast to restricted __le16
sound/soc/codecs/wm_adsp.c:736:23: sparse: cast to restricted __le16
sound/soc/codecs/wm_adsp.c:739:23: sparse: cast to restricted __le32
Correct this by changing the casts to use __leXX instead of uXX.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
There are very few version 0 firmwares in the wild and at some point in
the future it would be nice to remove support for them from the driver,
as they require several work arounds to be present to create controls
properly.
This patch adds a depreciated warning if someone is using this file
format.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Version 2 of the firmware file format includes length fields for the
various strings associated with control creation, to reduce file size.
However this does increase the parsing complexity slightly. This patch
adds support for the revision of the file format.
This patch also adds a new naming scheme for controls created from rev 2
firmware files. This version of the file format is commonly used to
add multiple controls per algorithm per memory region and the old
control naming scheme would cause multiple controls to have the same
name in this case.. Note that the naming scheme for older firmware
versions is left intact to ensure backwards compatibility.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Revision one of the file format includes new algorithm and coefficient
blocks which provide additional information about the controls exported
by the firmware. This patch updates the processing to handle this
version of the file format. Note that whilst this version of the format
adds support for specifying a name for the control through the firmware
file this has not been used and to keep compatibility with existing
deployments no changes to the firmware control naming are made by this
patch.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
This is slightly logically better and avoids some unnecessary forward
declarations in the following refactoring.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The private field in wm_coeff_ctl is currently unused and given the
controls are entirely handled within the ADSP code it is not clear what
it would be used for in the future. Remove the field for now it can be
readded if it is ever required.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Tidy up the code a little by factoring out the creation of the algorithm
regions.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Physically reading and writing controls to/from the DSP are handled by
two low level functions (wm_coeff_{write|read}_control, these currently
take in a snd_kcontrol pointer but immediately pull out a wm_coeff_ctl
pointer from the private data. These functions don't handle the kcontrols
at all they just shuttle data to and from the chip and all the call
sites have a wm_coeff_ctl pointer available. This patch just passes the
wm_coeff_ctl pointer straight into these functions.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Now we only allocate 44 bytes for the control name keep it on the stack
to avoid a lot of pointless memory allocation.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
ALSA only supports control names up to 44 bytes, so there is no point
allocating a whole page of memory to hold the control name, just limit
the control name to 44 bytes.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The algorithm region information in the firmware doesn't contain a
length field, explicitly pass this to the create_control function rather
than bundling into wm_adsp_alg_region.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
We have wm_adsp_region, wm_adsp_alg_region, and wmfw_region, the
variables for which are all frequently called region, this can get quite
confusing when reviewing the code especially given some functions are
quite long. Consistently use mem for wm_adsp_regions, alg_region for
wm_adsp_alg_region and region for wmfw_region.
Additionally, we use a mix of adsp and dsp for pointers to the wm_adsp
structure standardise this on dsp.
Finally, we use algs to refer to the number of algorithms quite
frequently, change this to the more descriptive n_algs.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The vast majority of the wm_adsp_setup_algs function is case statements
for ADSP1 or ADSP2, this patch splits this out into two separate
functions wm_adsp1_setup_algs and wm_adsp2_setup_algs. The small amount
of shared code between them is factored out into an extra helper
function. This makes the code a lot cleaner.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The adau1701 has two power domains, DVDD and AVDD.
Enable them both as long as the codec is in use.
Signed-off-by: Pascal Huerst <pascal.huerst@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
The fixed-regulator bindings require a separate property enable-active-high,
the standard gpio phandle property polarity setting is ignored.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Fixes: 4fe69a934b ("ARM: dts: Add Phytec pfla02 with i.MX6 DualLite/Solo")
Cc: stable@vger.kernel.org
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Otherwise the driver may try and send audio which may confuse the
monitor.
v2: set pin to NULL if no audio
v3: avoid crash with analog encoders
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Don't enable the audio and avi infoframes and audio stream
until all the state is set up.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
It's mostly duplicated with evergreen_dp_enable. This
is a prerequisite for fix implemented in another patch.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Set the line first, then enable the stream. May fix
pink line problems on some displays.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
The number of relocs is passed in by userspace and can be large. It has
been observed to cause kcalloc failures in the wild.
Cc: stable@vger.kernel.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This reverts commits 0a4e6be9ca
and 80f7fdb1c7.
The task migration notifier was originally introduced in order to support
the pvclock vsyscall with non-synchronized TSC, but KVM only supports it
with synchronized TSC. Hence, on KVM the race condition is only needed
due to a bad implementation on the host side, and even then it's so rare
that it's mostly theoretical.
As far as KVM is concerned it's possible to fix the host, avoiding the
additional complexity in the vDSO and the (re)introduction of the task
migration notifier.
Xen, on the other hand, hasn't yet implemented vsyscall support at
all, so we do not care about its plans for non-synchronized TSC.
Reported-by: Peter Zijlstra <peterz@infradead.org>
Suggested-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The kvmclock spec says that the host will increment a version field to
an odd number, then update stuff, then increment it to an even number.
The host is buggy and doesn't do this, and the result is observable
when one vcpu reads another vcpu's kvmclock data.
There's no good way for a guest kernel to keep its vdso from reading
a different vcpu's kvmclock data, but we don't need to care about
changing VCPUs as long as we read a consistent data from kvmclock.
(VCPU can change outside of this loop too, so it doesn't matter if we
return a value not fit for this VCPU.)
Based on a patch by Radim Krčmář.
Reviewed-by: Radim Krčmář <rkrcmar@redhat.com>
Acked-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pin direction configuration was incorrectly overwritten
by output and function values in set_mux(). Fix this.
Signed-off-by: Ivan T. Ivanov <ivan.ivanov@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
