An IPA header or processing context can be added once
and later deleted once from user space.
Multiple deletion may cause invalid state of the headers
software cache.
Change-Id: Ic0b8472b7fd8a76233a007d90c832af726184574
CRs-fixed: 1097714
Signed-off-by: Ghanim Fodi <gfodi@codeaurora.org>
USB client needs 19.2MHz frequency support for utmi clock, so that it could
be requested for. Add support for the same.
Change-Id: Icdbf9c5155bdd0ec02d357182d6e020c06a70648
Signed-off-by: Taniya Das <tdas@codeaurora.org>
Enabling deferred resume feature on SDM660. With this feature
enabled, MMC devices won't be resumed as part of the system
resume. It will be resumed only when there's a request to be
processed.
Change-Id: I901a888269b8bc0de006c8174546fd5f5c30baa8
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
It is observed that some PD-capable sink devices have a
longer start-up delay before responding to the Source
Capabilities message with a Request. Since it is PD capable
we'd like to wait until after entering PE_SRC_Ready and VDM
Discovery to start USB host mode. Accommodate these devices by
increasing the number of failed attempts of sending the message
from 5 to 10 before treating it as a non-PD device and starting
host mode.
Change-Id: Ia555699e78729b75dc118cbb929f7c4c4f2d1279
Signed-off-by: Jack Pham <jackp@codeaurora.org>
This reverts commit 85cc38fca2 ("usb: pd: Add support to set the
maximum speed via extcon") which unnecessarily restricted starting
USB host mode to high-speed only. Relax this so that USB host can
start in SuperSpeed when DFP mode is entered.
However, the only exception is if DisplayPort mode is discovered,
we need to drop down to high-speed only due to current hardware
limitations. This is handled during the VDM discovery stage.
Also use this opportunity to refactor all extcon_set_state_() calls
to helper functions that start/stop USB host/peripheral modes.
Change-Id: I2f32f70c57a8a6e2c34c8daf59dea931b445422f
Signed-off-by: Jack Pham <jackp@codeaurora.org>
Configure QRD8998HB analog and digital voltage to correct value.
CRs-Fixed: 1112314
Change-Id: I15d26582a87c9613c2ada010912b53e15e383b18
Signed-off-by: Jin Fu <jinf@codeaurora.org>
Add support to identify fake battery/debug board and suspend the USB
input path when this is detected.
CRs-Fixed: 1107245
Change-Id: Icaa936e4e5b3c9381bc92b6664b32547a820b9c3
Signed-off-by: Ashay Jaiswal <ashayj@codeaurora.org>
In some devices, qcom-i2c-pmic driver fails probe from the initial
poke failure of attached SEM slave.
As software workaround, this issue can be recovered by retrying to
ensure the unresponsive device is powered.
CRs-Fixed: 1083672
Change-Id: Ie828ba236f42b24a3d2acefdf069445cc72503f1
Signed-off-by: Harry Yang <harryy@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
The gcc_gpu_bimc_gfx_src_clk and gcc_gpu_snoc_dvm_gfx_clk need
to left at their default state of ON. Remove controlling them
from the linux clock driver to avoid disabling them during
late_init.
Change-Id: Iefc033998bf87fcc98dfaa1b7321d9cc33dedd5e
Signed-off-by: Taniya Das <tdas@codeaurora.org>
clk_prepare would hold a mutex and as clk_debug_mux_measure_rate has
already acquired a spinlock it would result in a BUG from sleeping context.
Avoid this by moving the prepare_enable before acquiring spinlock.
Change-Id: Ia405c884663ef80e87ae066df09f1c30134faf2e
Signed-off-by: Taniya Das <tdas@codeaurora.org>
The 's' flag is supposed to indicate that a softirq is running. This
can be detected by testing the preempt_count with SOFTIRQ_OFFSET.
The current code tests the preempt_count with SOFTIRQ_MASK, which
would be true even when softirqs are disabled but not serving a
softirq.
Link: http://lkml.kernel.org/r/1481300417-3564-1-git-send-email-pkondeti@codeaurora.org
Change-Id: I084531ce806e0f7d42a38be0a7ad45977c43d158
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Git-commit: c59f29cb144a6a0dfac16ede9dc8eafc02dc56ca
Git-repo: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
Some of the ARMv8.1, ARMv8.2 architectural features(ARM64_HW_AFDBM,
ARM64_PAN and ARM64_UAO) are enabled by default. Do not enable
these features by default, so that target specific defconfig
files can enable or disable based on features supported.
Change-Id: I072216927c3475169156a3d51f20d06675630ab9
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
External sense configuration is not supported. Hence remove it.
Change-Id: If65eb8c4d9720d4c4e9d913c91c8b33f420f29d1
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Prevent NULL pointer dereference while copying the data to memory device
process by adding mutex protection when closing the process.
Change-Id: Idaad73fb8b2a0bb5a4983d5e255bfdf702cecf71
Signed-off-by: Sreelakshmi Gownipalli <sgownipa@codeaurora.org>
The power directory structure has changed in kernel version 4.9. Align
msm-4.4 with this new directory structure.
Change-Id: Iba729f8ef33245ea78cdc05276ba8f5593191509
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
rc could be returned without being initialized. Initialize it.
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Change-Id: I044f74c7853e6853ae41734a95a4d86569c60457
pos is incremented by bytes_read which could exceed the boundary of
kbuf. Fix this by checking the bounds of pos.
resched_ms is used when uninitialized. Initialize it.
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Change-Id: I3a01d59bb5ed53666307a5210a684c7d9fd654fc
param_type is used for array access before the boundary check. Check the
boundary before using param_type.
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>
Change-Id: I3d1fb4ada3182c533e5b937717962dc513837e78
IRQ clean up should be done only after codec devices have been
notified with device down. Without this, there are chances of
race-condition where some of the resources like locks used for
safe handling of the IRQs would be used after those are freed.
CRs-Fixed: 1102353
Change-Id: I276571a7e50fcf227febd86494589b14988e3d41
Signed-off-by: Banajit Goswami <bgoswami@codeaurora.org>
Signed-off-by: Bhalchandra Gajare <gajare@codeaurora.org>
Allow MIC channel selection as the primary channel requested
by the application on Multichannel Echo Cancellation module.
CRs-fixed: 1102004
Change-Id: I94ba0f3171b4c7ebcc9936ca4f192f68de7be034
Signed-off-by: Rahul Sharma <sharah@codeaurora.org>
Signed-off-by: Derek Chen <chenche@codeaurora.org>
Because of recently introduced kernel change, previous change to
unlock mutex is not required as HID device addition is now done in
worker thread instead of main thread. Previous code flow was blocking
as uhid_hid_set_report was called from main thread which is not the
case now. Previous code flow was
uhid_char_write->uhid_dev_create2->uhid_hid_raw_request->
uhid_hid_set_report->__uhid_report_queue_and_wait->
wait_event_interruptible_timeout which is now changed to
uhid_char_write->uhid_dev_create2.
CRs-Fixed: 1112380
Signed-off-by: Hemant Gupta <hemantg@codeaurora.org>
Change-Id: I11c3e5628484739acc208deff84a82f34b1b7beb
11ad device supports 48 bit addresses, reflect that
by setting the dma mask accordingly to be able to
map skbs using 48 bits.
HW has limitation that all vrings should share the same
msb bits out of the 48 bits address. To overcome that,
for vrings mapping use 32 bit addresses only.
Change-Id: I7f710129ba61356f7f29ec69c885f60a31246151
CRs-Fixed: 1109659
Signed-off-by: Hamad Kadmany <hkadmany@codeaurora.org>
IPA needs to be kept disabled in SDA630 APQ platform.
Change-Id: I26831c989edee3f94f6cbafbb9955d669c546442
Acked-by: Mohammed Javid <mjavid@qti.qualcomm.com>
Signed-off-by: Utkarsh Saxena <usaxena@codeaurora.org>
