If video state set to DEINIT before processing all frame done
packets in the list may create video failures as explained below,
the client communication to video hardware will fail because of
DEINIT state and client will close the session upon failure which
will happen in parallel to response thread processing the response
packets in the list. It may happen that client already free'd the
buffer references and response thread might access the same buffer
reference and results in use-after-free memory fault. So In case
of sys error from video hardware, set video state to DEINIT after
processing all packets in the list to avoid use-after-free failure

Change-Id: I688c3ec3feb2b5621d75c4da93ee9870aa0e6dfe
Signed-off-by: Darshan Kumsi Srinivasa <darssr@codeaurora.org>

No need to process response messages from video hardware
after device went into invalid state. Processing responses
may result in use-after-free memory fault because client
might free all the resources after error.

Change-Id: I3bfb26e5aa52aba33b7b62cda7820dcbc5fe033f
Signed-off-by: Darshan Kumsi Srinivasa <darssr@codeaurora.org>

Subsys notifier chain for modem taking too long, MODEM SSR is
blocked, mutex held by wlan which is also in waiting state,
for event completion in wait_for_event_completion() routine.
To Avoid the above senario platform driver needs to send FW_DOWN
uevent during defered smp2p error handling as part of FW down event
wlan host completes the wait for completion events and release
the acquired lock.

CRs-Fixed: 2312089
Change-Id: I98b5c944328ab445f6dc112433bcd21c1f4e0773
Signed-off-by: Anurag Chouhan <achouhan@codeaurora.org>

On SDM670, GPU Fmax needed is 610MHz, so add support for
the same.

Change-Id: I75b4a3bb2ecb90ad931de67c076c54cef5c9872a
Signed-off-by: Deepak Kumar <dkumar@codeaurora.org>

Add initial device tree files to support QM215 QRD.
Delete all references of PM8937 and include PM8916.

Change-Id: I04a106b3c08c01baa2246090131ba5dee4189e9a
Signed-off-by: Teng Fei Fan <tengfei@codeaurora.org>

New soc id is assigned for sdm712 target. Add support
for the same.

Change-Id: If89ead70045a5009d9ac29062840326edbb68d82
Signed-off-by: Trishansh Bhardwaj <tbhardwa@codeaurora.org>

This change enables the squashfs with LZ4 compression
for msm8937go project.

Change-Id: I91678eee8586d5af6dce72d923776d87008b5f3c
Signed-off-by: Ankit Jain <jankit@codeaurora.org>

Enable device tree and overlay support for all sdmdolphin
supported platforms.

Change-Id: I668fc844298fc74df9ca48b90af5ac214d3ec513
Signed-off-by: Kaushal Kumar <kaushalk@codeaurora.org>

Add support for getting soc-id and dummy handle
on sdmdolphin.

Change-Id: I06f52d5093fb7f1646725fee70c7d18cd00a135e
Signed-off-by: Kaushal Kumar <kaushalk@codeaurora.org>

Protect ipa default routing table from
addition, deletion and modification once after
default rule added by ipa-driver.

Change-Id: I045d9c29fed23edf796d826e440b81124e1f666a
Signed-off-by: Mohammed Javid <mjavid@codeaurora.org>

Currently it is possible for an ION client to allocate non-HLOS memory
(ie memory which isn't assigned to the HLOS vmid), map this memory, and
then attempt to access this memory from the CPU.
Attempting to access non-HLOS memory from the CPU will cause a
stage-2 fault.

Fix ION so that non-HLOS memory cannot be mapped by the CPU.

Change-Id: Ifb51de2eabc076cddc744c13f01ef97b4a7c6874
Signed-off-by: Liam Mark <lmark@codeaurora.org>

By default all the interrupt sources for TYPE_C_OR_RID_DETECTION_CHANGE
are enabled and are getting fired frequently on any event. Disable all
the interrupt sources not handled by charger driver.

Change-Id: I613e06472a6f83b66b702b7420afd890792515dc
Signed-off-by: Umang Agrawal <uagrawal@codeaurora.org>

Add configs required for all common platforms sharing the
same msm8937go*_defconfig

Change-Id: I29e44e3b2e20303511ecbb77a4e4ae6ef4e28ecd
Signed-off-by: Balamurugan Thanikachalam <bthanika@codeaurora.org>

Sanitize debugfs inputs to only allow access to mdp memory block
specified in dtsi file. This change will allow only one single block
to be read at the time and will avoid accessing memory outside of valid
decode space which can trigger AHB error bus response.

Change-Id: Icede9a8939a66faa59d674c18183fb0ebcf67908
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>

Remove volume up and down configuration from base apq8053-lite-dragon
based devices and add into apq8053-lite-som based device for supporting
in dev console. This way volume up/down functionality gets enabled for
developers while apq8053-lite-dragon based device variants can expose
the gpio for other purposes using pio hal.

SPMI RESIN interrupt line is responsible for triggering the fastboot
command and doubles as a VOLUME_DOWN pin. For apq8053-lite-dragon based
device variants, in certain use-cases, these keyevents need to be
suppressed and instead, is directly read the line state (high/low)
from user-space. In order to support this, set the `resin-gpiobase`
entry to expose this interrupt line as a virtual GPIO.

Change-Id: I97720d1c0ce376a3fa620d1fe5d54d37c51f0988
Signed-off-by: Mayank Chopra <mak.chopra@codeaurora.org>

By current definition, CONFIG_VMALLOC_SAVING uses the VA space of
carveout areas as vmalloc area, but reduces the size of default
vmalloc space, thereby keeping the total vmalloc size same as
without CONFIG_VMALLOC_SAVING, and in turn increasing the lowmem.
This means, changes in carveout sizes will have the side effect
of altering the vmalloc size, which will then need "vmalloc"
kernel param changes for every change in carveout size. And these
changes in values passed to vmalloc kernel param, can have a side
effect on targets that shares the kernel params but does not save
any VA space using CONFIG_VMALLOC_SAVING (for e.g. when some or
all of the carveouts are in highmem area.).

It can be argued that this patch changes the meaning of vmalloc
kernel param, and is bad. For e.g. when vmalloc=300M is passed,
the user may expect the vmalloc to be 300M, but the actual total
vmalloc will be 300M plus VA recovered from carveouts. This won't
seem to be an issue if "vmalloc" param is understood as a method
to increase the "default" vmalloc area size.

Change-Id: I30bc22f1c685a7de92ca331bf32ca152798fc13b
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>

Enable linear charger, vm-bms, vibrator and  smb1360
drivers for qm215.

Change-Id: Ic4cd26261acc221fd99667e6ddf6b1fdc3b21b05
Signed-off-by: Kavya Nunna <knunna@codeaurora.org>

Modify the function names of batterydata interface to avoid conflict with
the qg battery interface as it uses the same names.

Change-Id: I4ea1b91c664d920d20e0a588901138e27bbe5bb8
Signed-off-by: Kavya Nunna <knunna@codeaurora.org>

If driver fails to register extcon notifier from probe, it defers
same to work_queue which is invoked when extcon provider becomes
available (psy_nb is used for same). If power_supply notifier call
gets invoked back to back before extcon_register_work could
unregister psy_changed notifier, it results in driver registering
extcon notifier two times without unregister. This causes notifier_nb
list corruption where notifer_nb->next points to itself and any
extcon event later results in inifinite notifer calls causing
watchdog bite.

Change-Id: Ic7963971db198817e5643b6f3968cd8a4bc2e43c
Signed-off-by: Manu Gautam <mgautam@codeaurora.org>

During driver loading larger memory pools get allocated to smaller
memory requirement which result in unavailability of memory for
larger modules.
Updated prealloc memory table by adding memory blocks.

CRs-Fixed: 2322354
Change-Id: I642b7f2b0af5a9661ef7a3515b5340c0bced7c77
Signed-off-by: Sandeep Singh <sandsing@codeaurora.org>