seperately.
decouple preview and recording path configuration
by configuring the buffers during preview start
and recording start respectively.
Signed-off-by: Kiran Kumar H N <hurlisal@codeaurora.org>
Signed-off-by: Mingcheng Zhu <mingchen@codeaurora.org>
add necessary changes in camera driver to
migrate to videobuf2 framework.
Signed-off-by: Kiran Kumar H N <hurlisal@codeaurora.org>
Signed-off-by: Mingcheng Zhu <mingchen@codeaurora.org>
Conflicts:
drivers/media/video/Makefile
eMMC card can be put in sleep by sending CMD5 during suspend
so we can turn off the eMMC VCC rail to save power.
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
eMMC card can be put in sleep by sending CMD5 during suspend
so we can turn off the eMMC VCC rail to save power.
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Put MMC to sleep if it supports SLEEP/AWAKE (CMD5) in the mmc suspend
to minimize power consumption.
Signed-off-by: Balaji T K <balajitk@ti.com>
Acked-by: Venkatraman S <svenkatr@ti.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
LSFR and CONFIG registers are configured by secure domain
code and XPU protected.
CRs-Fixed: 284133
Signed-off-by: Mona Hossain <mhossain@codeaurora.org>
Added support for snd pcm client driver. Which
provides interface to write pcm samples to voice path.
Signed-off-by: Rajesha Kini <rkini@codeaurora.org>
This reverts commit 66882a5046143601429632a0b589d9ffeeb3c668.
This change will be submitted later.
Signed-off-by: Kiran Kumar H N <hurlisal@codeaurora.org>
Fix the division of bandwidth between slaves with multiple ports.
Since, interleaving changes involve exposing multiple ports for
a given master, config functions need to be changed for
halting/unhalting the right ports.
Signed-off-by: Gagan Mac <gmac@codeaurora.org>
System Fabric ARB RPM resource was changed from 27 to 29 4-byte words,
update the RPM-id mapping to reflect this.
Signed-off-by: Eugene Seah <eseah@codeaurora.org>
On 8960, all memory banks except the lowest one
(where the kernel lives) are unstable (i.e. may
be powered off under DMM).
This patch temporarily hard codes the bank
size to 512M, a later patch will allow us to
obtain this from shared memory.
Signed-off-by: Larry Bassel <lbassel@codeaurora.org>
While this check is valid in theory, it generates a plethora
of false positives on ALSA drivers and ends up causing more trouble
than it saves.
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
Some clients have specific alignment requirements for
iova addresses. Allow those requirements to be specified
via the upper bits of the flag field. Since all iommu
alignment must be at least 4k, the lower bits can specify
other options while the upper bits specify the alignment.
Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
USB shared memory DUN(dial up networking) driver re-tries multiple times
to open smd channel, due to delay in smd driver initialization and modem
load during boot up. Instead, register a platform driver with smd channel
name. Whenever specific SMD channel is opened by modem, a corresponding
platform device would be added resulting in probe of platform driver being
called. In the probe function, DUN driver opens up SMD channel
Signed-off-by: Hemant Kumar <hemantk@codeaurora.org>
Fix issue where clients can get notifications of failed
writes before they are notified that a reset has occurred.
Signed-off-by: Vamsi Krishna <vskrishn@codeaurora.org>
When a remote processor restarts, the transport corresponding to that
node notifies the MSM_IPC Router of that event. MSM_IPC Router cleans
up the data structures corresponding to that transport and notifies
the local clients/servers of removal of all the concerned remote
servers/clients respectively. When the remote node comes up and
broadcasts the server information, local clients will be notified of
the arrival of the server so that they can continue communication with
the server.
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@codeaurora.org>
The first time the target boots up, complete(charm_boot) is
called. Subsystem restart assumes that complete(charm_boot)
has never been called before the first time. This can result
in the charm code getting into a bad state. Fix this by not
calling complete on bootup.
Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Move msm7x2x_misc_init() to init_machine stage because
acpuclock initialization should not happen before timer
driver initializing the lpj.
CRs-Fixed: 295455
Signed-off-by: Trilok Soni <tsoni@codeaurora.org>
The sampling rate of Fluence topology is either 16K or 8K. Therefore,
we need to reduce the ADM sampling rate to 16K during recording with
Fluence topology
Signed-off-by: Jay Wang <jaywang@codeaurora.org>
Correct the sensor type value for the IMX074 sensor.
Set the mount angle value for this sensor on 8960,
also made sure that the values obtained for angle,type
from probing are recorded in the structure which used
to retrieve this information during device query.
CRs-fixed: 296576
Signed-off-by: Mansoor Aftab <maftab@codeaurora.org>
During power off we disable 5v, power down core, by turning off the
TX channels one by one. When a HPD is triggered the driver has a
software debouncing timer during which interrupts are disabled to
be able to read stable HPD status from the register. When HPD is
detected while the core is being turned off/powered down the driver's
software disables the interrupt and this was making the core to go
into off state. So the fix was to keep the interrupt bit enabled through
the register even during the debounce timeout (this doesn't trigger
unnecessary interrupts due to appropriate clearing and masking of HPD
interrupt bits).
Signed-off-by: Manoj Rao <manojraj@codeaurora.org>
The function arch_add_memory() should only perform logical
memory hotplug, but it was also improperly performing
physical memory hotplug. Even worse, it was adding pages
to the free list before the memory bank they were in
was powered on.
Signed-off-by: Larry Bassel <lbassel@codeaurora.org>
To support sending of calibration in physical shared memory
the number of voice volume steps is only sent in the first
voice volume calibration block. This means that all products
that do not support out-of-band calibration have to pull
the number of voice volume steps from the first buffer and
place it into the rest of the buffers.
Signed-off-by: Ben Romberger <bromberg@codeaurora.org>
Handle VOICE START event when voice state is in voice change
Consider the following case
B calls A (A accept the call), voice state in voice acquire
C calls A (A accept and B in hold) in this scenario following series
Of events issue to voice driver
VOICE_END (this event issues a device change command to modem, in turn
Modem changes the voice state from voice acquire to voice change)
VOICE_START (this event is not changing the voice state from voice change
To voice acquire as a result mismatch in the voice state maintained by
Voice proxy driver)
This patch change the voice state from voice change to voice acquire
When VOICE_START event received and voice state is in voice change
CRs-Fixed: 295375
Signed-off-by: Santosh Mardi <gsantosh@codeaurora.org>
If this function doesn't exist the generic clock layer returns
-ENOSYS which should be sufficient enough to signal that it isn't
supported.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
ANC is a feature which reduces noise on the near-end
by pushing anti-noise onto the audio stream heard
by the user. ANC mixes the anti-noise signal
inside of the WCD9310 codec.
Signed-off-by: Brad Rubin <brubin@codeaurora.org>
Physical Digital MIC 1 on Fluid and MTP is there on Front Bottom Left.
It is connected D_DMIC1_2_data pin on Tabla and drives data on right
channel which is treated as DMIC2 input. Physical Digital MIC 4 on
Fluid is there on Back top. It is connected D_DMIC3_4_data pin on
Tabla and drives data on left channel which is treated as DMIC3 input.
Signed-off-by: Kiran Kandi <kkandi@codeaurora.org>
Bahama's SYS RESET GPIO is put in LOW state in bahama shutdown,
only when the GPIO is available from the I/O expander.
Signed-off-by: Rahul Kashyap <rkashyap@codeaurora.org>
Always reset the modem_reset flag in rpcrouter_sdio_remote_probe function.
Else repeating failed modem restart attempts, cause the probe to free the
already freed resources and hence cause a crash.
CRs-Fixed: 295677
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@codeaurora.org>
Currently, ASM buffers are freed before closing
ASM stream in the closing function. ASM in QDSP6
can still be actively accessing buffers. Existing
sequence can result in QDSP6 crash. Adjust
the closing sequence.
Signed-off-by: Patrick Lai <plai@codeaurora.org>
Wait until the external modem has completely booted up
before monitoring the GPIO lines. Otherwise, there could
be noise on the lines which cause erroneous interrupts.
Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Adds logic to enable analog input 3 and analog
input 4. These inputs use external microphone
bias 3 and 4 respectively on the CDP only.
Adds all possible routes in the codec for each
of the inputs.
Signed-off-by: Bhalchandra Gajare <gajare@codeaurora.org>
Adds logic to enable the analog Input 5 and 6.
These inputs do not need a microphone bias to
be enabled.
Signed-off-by: Bhalchandra Gajare <gajare@codeaurora.org>
Perform the receive_cb(NULL) notification before cleaning
up any pending write skb's. This allows clients to update
their cached reset status based upon the
receive_cb(NULL) in case they need to do special
handling of the cancelled write skb's.
Signed-off-by: Eric Holmberg <eholmber@codeaurora.org>
