Add support for 384KHz sample rate and S24_3LE
bitwidth for dummy dai.
Change-Id: I7d20311e58f6ec1f444a46908fd4827be32bed17
Signed-off-by: Rohit kumar <rohitkr@codeaurora.org>
The dbg_event function call in dwc3_msm_remove after the
child dwc3 device is removed is resulting in use after
free of dwc_ipc_log_ctx.
Even though the dwc_ipc_log_ctx is set to NULL in
dwc3_remove, the value is reset once the core platform
device is removed because the dwc3 is allocated using
devm_kzalloc. This call ensures that the memory is freed
up once the device driver is de-registered.
Fix this by calling the dbg_event before de-registering
the core driver.
Change-Id: Ib2cc3d233ea0fd4a1523cae23f57d3cb3753a53f
Signed-off-by: Sriharsha Allenki <sallenki@codeaurora.org>
If a gadget provides "OS descriptors" from the userspace then composite
gadget driver designates only one configuration which has OS descriptor
support to be used always with the hosts. With this we are seeing
enumeration issues in multi-config compositions where device does not
allowing Host to identify preferred configuration to choose.
Hence change the logic in config_desc for handling OS descriptor
configuration.
Change-Id: I5daa7631f9f2d634ee6b4b377b8d66283b21a497
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
With multi-config compositions some hosts can bind the configuration
incorrectly and sending os descriptor request on invalid interface.
This could cause accessing NULL pointer and results in panic.
Fix this by bailing out from the OS descriptor setup request handling
if the interface is NULL.
Change-Id: I65f01b876a46b907eb883e48878fc081860b0753
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
Add code to download bdf from device specific path, if the
file is not present then it returns error.
Change-Id: I6358be8a9e09c0caa5f4478d08f1959e0a0f1c44
Signed-off-by: Mohammed Siddiq <msiddiq@codeaurora.org>
Add SOC ID to support socinfo for YupikP platform.
Change-Id: I6f74f83d058c50ae276f7930b63f19418dd6e2ad
Signed-off-by: Naina Mehta <nainmeht@codeaurora.org>
Currently in f_cdev driver, we are not handling the TIOCM_CTS
signal generared by SIO_IOCTL_INBOUND_FLOW IOCTL. Hence we
are not able to send BRK signal to the host and it keeps
sending data leading to data loss.
Fix this by handling TIOCM_CTS in the driver.
Change-Id: If8e4774e3e0772cb7189f3400d7e53e258b7d5a3
Signed-off-by: Ajay Agarwal <ajaya@codeaurora.org>
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
The commit <d30948cecd> (netfilter: Changes to handle segmentation in
SIP ALG) breaks the build by defining the last argument of the function,
sip_tcp_skip_process (s16 *tdiff) incorrectly. Hence, fix the type by
changing it to a variable.
Change-Id: I5a336bd7aa47a649c24d7c34090f138b2e772bdc
Signed-off-by: Raghavendra Rao Ananta <rananta@codeaurora.org>
There is a race possibility in android_disconnect and
configfs_composite_unbind while using cdev leading to a NULL pointer
dereference in composite_disconnect. Combine android_disconnect with
configfs_composite_disconnect and remove the android_disconnect function.
configfs_composite_disconnect already has a gi->spinlock in place to
prevent the race condition.
Change-Id: Idfdebaf69f3aa68d90b55bffd7c2e04410c5a47f
Signed-off-by: Ronak Vijay Raheja <rraheja@codeaurora.org>
By default, the system firmware enables the mpm wakeup functionality
of all the supported GPIOs. However, certain usecases demand that
this capability be disabled on certain gpios. Hence, provide a
generic interface to disable the GPIOs as needed.
Change-Id: Ib36569d422d2336a81d71e451ef47196c956c3a5
Signed-off-by: Mukesh Ojha <mojha@codeaurora.org>
Driver marks func_is_suspended flag to true and queue suspend work in
gsi_suspend. suspend_work may fail the suspend if the LINK is not in U3.
During host hibernation testing gsi_suspend failing as LINK is not in U3
and next time when bus suspend initiated driver checking for
func_is_suspended flag and bailout the suspend with out queuing suspend
work. This causes a mismatch of usage count and usb not entering to
low power mode.
To fix this issue, Check for the state in gsi_suspend. If the state is
!= suspended or != suspend_in_progress then don't bail out the suspend.
Change-Id: Iade9ba330b232261e205cd821ef42053297befbb
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
Propagate following change from 4.14 to 5.4 for SPI SLAVE driver.
'commit 3c7e85615a1d64 ("spi: spi-geni-qcom: Restrict SPI
responder to transfer mode 1")'.
Change-Id: I6170a0e021cc0a09a2bc9ee26dfc504ee85bbf91
Signed-off-by: Ashish Kori <akori@codeaurora.org>
Propagate following change from 4.14 to 5.4 for SPI driver.
'commit 60dc9d0568e31 ("spi: spi-geni-qcom: Suspend SPI slave
after timeout")'.
Change-Id: I08c7c2a861285ce45a28ea1aab1db9a99016a35e
Signed-off-by: Ashish Kori <akori@codeaurora.org>
Propagate following change from 4.14 to 5.4.
'commit f9f1b07419520b ("platform: qcom-geni-se:
Correct macro definition of GENI_IO_MUX_1_EN")'.
Change-Id: I6d5bd2b3e7252ad272d88352a52dc005e4368af1
Signed-off-by: Ashish Kori <akori@codeaurora.org>
Propagate following change from 4.14 to 5.4.
'commit ab708589654b ("platform: qcom-geni-se: Correct macro definition")'.
Change-Id: I7fbc1bfad0788ab90f62b66ecf2e2d0c8ec82485
Signed-off-by: Ashish Kori <akori@codeaurora.org>
On some targets, f_serial switches transport path from
char_bridge to TTY in DUN+SoftAP scenario. As a part of that,
we call usb_cser_disconnect but notify_modem callback is still
marked as usb_cser_notify_modem. Then on getting a SETUP req
from host PC, usb_cser_notify_modem is called which can lead
to NULL pointer dereference because port_usb is NULL.
Fix this by marking notify_modem callback as NULL on
usb_cser_disconnect.
Change-Id: I4470030cde6f38b905f7a88e16ed9f261a8954d6
Signed-off-by: Ajay Agarwal <ajaya@codeaurora.org>
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
When the host PC sends a DTR high during DUN call,
it expects to get a DSR signal from the modem in reply.
Add necessary support for that.
Change-Id: If46c4b75fa6210c2b15cd2880bc93461fa3a4e89
Signed-off-by: Ajay Agarwal <ajaya@codeaurora.org>
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
Cleanup .name for all external clocks as these clocks are
mapped using .fw_name and references to their phandles in
GCC DT node.
Change-Id: I8b4a0600749da29594f5efc9406237945d2c3a1e
Signed-off-by: Jagadeesh Kona <jkona@codeaurora.org>
Configure the Vdd regulator with supplied load
in NPM mode. Configure the Vdd regulator with 0
load in LPM mode.
Adding support for reading load value for Vdd regulator from
device tree. If this property is not present, no explicit vote
required for load.
CRs-Fixed: 2840599
Change-Id: Ic2f12d9f9ce3e6a9ba6b58888cc5c202a74d8cc3
Signed-off-by: himta ram <hram@codeaurora.org>
Add support to deal with different feature set of A635 GPU.
Change-Id: I8add6e83624f833a937364995da9431b02a3f6c4
Signed-off-by: Pankaj Gupta <gpankaj@codeaurora.org>
Enable CONFIG_CMA_DEBUGFS for QGKI. This enables read only nodes on QGKI
build. Enable CONFIG_CMA_ALLOW_WRITE_DEBUGFS for debug config. This
enables writeable cma debugfs node for debug build.
Change-Id: Ia40eea08c42b5c4d29691ea354411216c03e60a7
Signed-off-by: Prakash Gupta <guptap@codeaurora.org>
This reverts commit 8d0b09d4f2.
With CONFIG_DEBUG_FS enable for non-debug build, we can enable read only
CMA debugfs nodes on all builds.
Change-Id: Ia4a0a9431e27ec908feaae54dbe622a1d985958b
Signed-off-by: Prakash Gupta <guptap@codeaurora.org>
Propagate Propagate following change from 4.14 to 5.4 for SPI driver.
'commit 005c3a05b41c3ea ("spi: spi-geni-qcom: Allow MISO<->MOSI
connectivity for slave pins")'.
Change-Id: I66ca5c2f851f8e3d1485580486cd2900a3756341
Signed-off-by: Ashish Kori <akori@codeaurora.org>
By default, the system firmware enables the mpm wakeup functionality
of all the supported GPIOs. However, certain usecases demand that
this capability be disabled on certain gpios. Hence, provide a
generic interface to disable the GPIOs as needed.
Change-Id: I147629e0126b2eb41c645de03958ef480fd5d7d0
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Update pinctrl driver for Yupik, to match latest
configuration.
Change-Id: I067fb20535dd144bf07f1253dbf9b498de8bb123
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
When card max supported clk frequency is less than 50MHz,
driver registers scale up frequency as 0 and scale down
frequency as max card supported clk frequency. So read/write
during high load performace is dropping.
Ideally in this scenario, clk scaling could be disabled by
skipping registration with devfreq, but userspace might see
scaling supported for some card and its not supported for another
which might create confusion. Hence keeping low and high, clk
scaling frequency as same, and scaling would be skipped when
devfreq attempts to scale the clk.
Change-Id: I25977bdff4275c4ebec7309952029f43e0d2455f
Signed-off-by: Ram Prakash Gupta <rampraka@codeaurora.org>
Currently driver performing Abort sequence only when we seeing timeout
for Cancel interrupt. When there is a race condition between rx data
coming into rx fifo and a cancel_s_cmd is given. The rx data will end
up consuming the rx dma buffer mapped for cancel command and due to
this geni secondary sequencer will be in active state even after
cancel command.
This changes is to check for geni status after cancel operation. If geni
is active perform abort sequence.
Change-Id: I1b463284b48eb940de342da6df9233157d48f469
Signed-off-by: Chandana Kishori Chiluveru <cchiluve@codeaurora.org>
Add snapshot for coresight fuse driver from msm-3.18
commit cf126d1be1a("ANDROID: sdcardfs: Add option to
not link obb"). Make necessary change to adopt coresight
framework change.
Change-Id: If0c677c9e7cb33c2c1656fb458649f0b4c069a00
Signed-off-by: Yuanfang Zhang <zhangyuanfang@codeaurora.org>
When FIFO samples are played back to back along with SWR mode, there is
a chance that FIFO samples won't get flushed out if they're not fully
played before FIFO mode is stopped. When this happens, FIFO samples gets
accumulated leading to overflow after which FIFO mode stops working.
Only way to recover from this is to toggle HAPTICS_EN_BIT before
starting FIFO mode playing again. While at it, add logs to print out
FIFO real time fill status which is helpful for debugging.
Change-Id: I07cd9ac34ad5ccfe7011850db554ba727f568482
Signed-off-by: Fenglin Wu <fenglinw@codeaurora.org>
