Currently, whenever the switch LED device is triggered, IRES is
updated based on the LEDs belonging to that switch LED device.
However, this can overwrite IRES configuration if flash and torch
LED devices have different IRES since they share the same id.
Fix this by checking the active LED status which will be updated
based on the brightness level set so that IRES will be updated
for the correct LED (flash/torch) device.
CRs-Fixed: 2173127
Change-Id: Ic3b5db2f56758ccd68fd80139aeb22f31723130e
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Currently usb stack is teared down if no activity is detected for
10 seconds when floating charger is connected. USB 3.1 compliance
equipment usually reported as floating charger. Introduce sysfs
parameter to indicate if usb compliance in progress and do not
tear down the stack based on this parameter.
Change-Id: I26013c281827f35eac7f21ed68b3880e541d82b9
Signed-off-by: Vamsi Krishna Samavedam <vskrishn@codeaurora.org>
Signed-off-by: Vijayavardhan Vennapusa <vvreddy@codeaurora.org>
Most often a source will send fewer than the maximum number
of PDOs (7). Since the rx_msg buffer is now allocated up to
the size of the actual data, honor the rx_msg->data_len when
copying to pd->received_pdos rather than always 28 bytes.
This fixes out-of-bounds read access as reported by KASAN.
Change-Id: I5f98f7ccba027c1ab436ccf6fc822e2a319bafa1
Signed-off-by: Jack Pham <jackp@codeaurora.org>
When in source mode if a hard reset is received just before or
while usbpd_sm() is in the middle of trying to send an outgoing
message, the signal handling would get delayed due waiting for
the PD PHY to complete retrying and failing both the message as
well as the subsequent Soft_Reset message. Instead, treat the
incoming hard reset with priority try to bail out of further
attempts to send a message so that we can immeidately re-queue and
process the hard reset on the next usbpd_sm() invocation.
In case the TX attempt still manages to win the race, this will
cause delay that affects our tPSHardReset (25-35ms) required time
resulting in VBUS getting turned off too late. Handle this by
keeping track of when the hard reset signal arrived so that
we can schedule SRC_TRANSITION_TO_DEFAULT within this window.
Also promote the error message when a hard reset is received
from KERN_DEBUG to KERN_ERR, as this is abnormal enough
to deserve printing to the kernel log with higher priority.
Change-Id: Ie503f7b776022067ec3a4788d5229ec508b9c55f
Signed-off-by: Jack Pham <jackp@codeaurora.org>
commit 07040df2790e ("usb: pd: policy_engine: Handle spec revision
properly") changed the spec revision handling to be 2.0 mode unless
a programmable power supply is seen. Extend this via a module
parameter to allow PD 3.0 in sink-only mode regardless of PPS when
a 3.0 source is connected.
Since the revision of the source is first encountered when
receiving the Source Capabilities message, this information is
lost by the time pd_eval_src_caps() is called. So bring back the
handling in phy_msg_received() removed in commit 07040df2790e ("usb:
pd: policy_engine: Handle spec revision properly").
Change-Id: I0485224aeadcffbfaecd0c7942c6dbcd9ddd3813
Signed-off-by: Hemant Kumar <hemantk@codeaurora.org>
Issue:
When total_steps is updated, after that, copy_from_user
fails with an error, then, i2c_reg_tbl is not allocated.
In this case, when calling msm_actuator_parse_i2c_params,
it lead to out-of-bound memory write.
Fix:
1) Assign total_steps to zero when error from copying.
2) Add NULL pointer check for i2c tbl.
3) Fixing the issue where the function can return
with an error code leaving "a_ctrl->i2c_reg_tbl"
and "a_ctrl->total_steps" out of sync.
Change-Id: Ib46deceb7bd8efff1cb606b894396e7016271dd3
Signed-off-by: Rahul Sharma <sharah@codeaurora.org>
Memory cgroups is an unused feature on sdm660. But
enabling it in kernel means that user space can
configure it. Avoid such user configurations until
the feature is decided to be deployed.
Change-Id: Ie77a4de51a59179ff3cb543e3cc955ce48412a59
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
The ath10k wlan firmware supports GTK rekeying while asleep.
Enable GTK rekeying failure wakeup source in wow suspend.
This adds support to wakeup the device from wow suspend
state if GTK rekeying process failed in the wlan firmware.
CRs-Fixed: 2169340
Change-Id: I321a6139d74b0fcf476e961ba83879fae9c39f04
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
Variable map may pointing to the same buffer on race conditions
in functions fastrpc_internal_mmap and fastrpc_internal_munmap,
use mutex to avoid race conditions on same buffer.
Change-Id: I96ed884c44a36f574677ba3ba189dfbf2ce3751d
Acked-by: Krishnaiah Tadakamalla <ktadakam@qti.qualcomm.com>
Signed-off-by: Tharun Kumar Merugu <mtharu@codeaurora.org>
Currently, recharge SOC is adjusted based on the SOC at which
charge termination occurs. However, it gets reset back to the
original threshold only if input is removed. This does not work
well if charging resumes again once the battery is out of JEITA
as the recharge SOC threshold stays configured at a lower value.
Fix it by adjusting recharge SOC to the original threshold once
charging begins.
CRs-Fixed: 2163058
Change-Id: I07c714d32ba0005d3b0ff61142928eda58f049d0
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
KTM lmh dcvs frequency mitigation uses given online/first cpu max/min
mitigation request. There can be cases like emergency frequency
mitigation where mitigation request is for a particular cpu only and
online/first cpu may be one of the other cpus from the same cluster.
In this case lmh dcvs takes online/first cpu max request which can be
in no mitigation state. It leads to unmitigated state even though one
of the cpus is triggered.
If device supports cluster mitigation, use min/max request of that
cluster instead of min/max request of given online/first cpu.
It ensures lmh dcvs mitigation if one of the cpu mitigation is
triggered and other unmitigated cpu of same cluster is given as
online/first cpu.
Change-Id: Ibbb913eb67a7f84d4c3658d0edae495990ca9010
Signed-off-by: Manaf Meethalavalappu Pallikunhi <manafm@codeaurora.org>
To support reset function in virtual clock driver, we register
a reset controller in virtual clock front driver. It is a virtual
reset controller which talks with host backend via virtual clock
message.
Change-Id: I13d48180534110260d66e42f3a02b63c2afcaeca
Signed-off-by: Zhiqiang Tu <ztu@codeaurora.org>
Fix warning that the variable is used before initialized, which
might be treated as error in strict case.
Change-Id: I95a34fa31e87fde2055af745f46f65a0d0402efc
Signed-off-by: Ferry Zhou <tianguiz@codeaurora.org>
AIS camera voting on VFE and ISPIF clocks cause XO shutdown failure.
So, add new AIS manager driver, to control AIS camera clock votes.
Also add a new method in early camera driver to turn on/off clocks
when it's needed.
Change-Id: I43090b51cb29ca9de62dfa191f77b7aa9dae8613
Signed-off-by: Rahul Sharma <sharah@codeaurora.org>
Issue:
i2c_reg_tbl may be null under error condition when set param.
then, other actuator function still may use the i2c_reg_tbl as null.
Fix:
1) the assignment total_steps follow on kmalloc buffer.
2) Add NULL pointer check for i2c tbl.
CRs-Fixed: 2152401
Change-Id: Ieec3d88e6dae0177787da0906f53d59ac4f5a624
Signed-off-by: Haibin Liu <haibinl@codeaurora.org>
