The current spinlock lockup detection code can sometimes
produce false positives on some platform, because of the
unfairness of the locking algorithm itself, so we disable
DEBUG_SPINLOCK.

Change-Id: Ic909d2aac3f669c5324bfe32ece66d3d9d1c6db9
Signed-off-by: Tengfei Fan <tengfeif@codeaurora.org>

Merge "msm: adsprpc: restrict user apps from sending kernel RPC messages" into kernel.lnx.4.9.r12-rel

Merge "msm: camera: reqmgr: Hold device lock while retrieving session info" into kernel.lnx.4.9.r12-rel

There is off-by-one read issue in function ipa3_is_vlan_mode.
Add fix to overcome this issue.

Change-Id: If087a762166f09d8786352de680db6d46b70b794
Signed-off-by: Praveen Kurapati <pkurapat@codeaurora.org>

There is a possibility for destroy session and link to occur
concurrently, if this does we end up using a freed pointer while
linking. Avoid this by increasing the scope of the device lock.

Change-Id: I6ffb2d0633455a838cca9194e4717d82a9647183
Signed-off-by: Karthik Anantha Ram <kartanan@codeaurora.org>

Compute buffer size only if num_out_res is less than MAX resources.

Change-Id: I99740682ccc84cbd3aaa90afe1e42ab46160cc2c
Signed-off-by: Trishansh Bhardwaj <tbhardwa@codeaurora.org>
Signed-off-by: Raja Mallik <rmallik@codeaurora.org>

Verify that user applications are not using the kernel RPC message
handle to restrict them from directly attaching to guest OS on the
remote subsystem

Change-Id: Icfa114a12f2bebbe815eb9930027fded51f717fd
Acked-by: Thyagarajan Venkatanarayanan <venkatan@qti.qualcomm.com>
Signed-off-by: Tharun Kumar Merugu <mtharu@codeaurora.org>
Signed-off-by: Mohammed Nayeem Ur Rahman <mohara@codeaurora.org>

Add thermal-sensor-cells property to adc_tm on PM8937
to enable thermal registration for adc_tm channel.

Change-Id: Ia2fec41274960cafc42806fa899dd3419129f88f
Signed-off-by: Jishnu Prakash <jprakash@codeaurora.org>

Change-Id: I9b62885540917d69db24c8150fc812f8264daaf0

This change updates the wait time before turning off the reference clock.
The wait time of UFS version 3.0 and onward devices can be read out from
their device attributes. Otherwise, fix the wait time to 100 micro seconds.

Change-Id: Ie5085e62ef3efa2bfaa3f632593e4ade52ffd645
Signed-off-by: Can Guo <cang@codeaurora.org>
Signed-off-by: Sayali Lokhande <sayalil@codeaurora.org>

There can be many ice instances present in dtsi file but
not all of them will be initialized by storage driver.
Check if crypto instance is initialized before setting
it up for data encryption/decryption usage.

Change-Id: I7c9227007474052513b277dec5963a973781c524
Signed-off-by: Neeraj Soni <neersoni@codeaurora.org>

This change adds a device quirk for some UFS devices which need long delay
after device reference clock is turned on but before initiation of the
state transition to STALL from a LS-MODE or from the HIBERN8 state.

Change-Id: I587dba737633034cca93b8ba839ab367dbc326e1
Signed-off-by: Can Guo <cang@codeaurora.org>
Signed-off-by: Sayali Lokhande <sayalil@codeaurora.org>

Sometimes TZ side images won't have support for SCM7 call in this cases
we need to fallback to SCM6 call.

Change-Id: I276547cfddfc1c12784f43f790440f8c5ed53330
Signed-off-by: Tony Lijo Jose <tjose@codeaurora.org>

Panic ctrl is enabled on dma start and cleared
on dma stop, if supported. Doing the same again
in ctrl on/off is redundant.

Change-Id: I60dab92e54d01973fc11a3dc42933d1ab6f84a4e
Signed-off-by: Animesh Kishore <animeshk@codeaurora.org>

Fix compilation errors in smmu, flash, ife and cdm.

Change-Id: If9c80f74c5b0b2c69ea8506d7f0b256f91c80bc0
Signed-off-by: Trishansh Bhardwaj <tbhardwa@codeaurora.org>

Rename sdmnobelium to sdm712.

Change-Id: I9fdfbf3ebde32c7f6aa97968668f03e5ba80fdfa
Signed-off-by: Kaushal Kumar <kaushalk@codeaurora.org>

Even when ION_FLAG_POOL_FORCE_ALLOC is set for secure alloc,
alloc_from_pool_preferred tries to get the pages from pool
first. In case if it enters split_page_from_secure_pool,
free_buffer_page can end up calling __free_pages on the page
from pool which is hyp assigned, resulting in issues for e.g.
a fault if page poisoning is enabled.

Change-Id: I68759dc17551a5705693506a5c137977d429fe36
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>

pppolac driver incorrectly enqueues the packet into the sock queue
without pulling UDP headers. The application will receive data along
with UDP header when L2TP control packets are received.

The issue was introduced after moving UDP header removal functionality
from process rcvmesg context to BH context.

Instead of pppolac driver directly queuing L2TP control packets into
socket queue, return packet to udp_queue_rcv_skb, which will deliver the
packet to the application after pulling the UDP header.

Fixes: e6afc8ac ("udp: remove headers from UDP packets before queueing")

Change-Id: Icfa0fd8da43ea9c14fa7c718746a6529651ac202
Acked-by: Sharath Chandra Vurukala <sharathv@qti.qualcomm.com>
Signed-off-by: Tejaswi Tanikella <tejaswit@codeaurora.org>

The GCC wrapper writes any error message from GCC to stdout
along with the messages from the wrapper itself. This is okay
for most case, but when GCC is used with -print-xxx flags,
the stdout output is supposed to be taken as input to some
other build command, so putting error messages in there is
pretty bad. Fix this by writing error messages to stderr.

Change-Id: I4656033f11ba5212fdcc884cc588f8b9d2c23419
Signed-off-by: Shadab Naseem <snaseem@codeaurora.org>

Change-Id: Ia05dffce6b2cb28ef83b3775f0bfe48644b5fff9

Add the dynamic wakeup source property to IPC router LPASS xprt.

CRs-Fixed: 2057391
Change-Id: If5dae17b105b91f401522a6b27356a632abc125c
Signed-off-by: Arun Kumar Neelakantam <aneela@codeaurora.org>

Writeback timeout is seen during writeback disable. This is
because IRQ registration is now moved out of prepare kickoff
and is only enabled by resource control during kickoff. So,
refcount is added for wb irq registration balancing. Also,
add wb ctrl for null commit during writeback disable.

Change-Id: I5b1b2abbf275824158a9732b89c1f35deaf347b7
Signed-off-by: Shubhashree Dhar <dhar@codeaurora.org>

Panel may go bad during esd test, in such cases do not wait
for PP done when switching from ON to LP1/LP2, which will cause
fifo error's and black screen.

Change-Id: I2a416e2a5e332f0be00215734342405292410c41
Signed-off-by: Raghavendra Ambadas <rambad@codeaurora.org>

scm_call2() API can block indefinitely if another client is using
this API. This is due to a mutex in this API to serialize calls to
the TZ. This blocks the GPU wake up which in turn can result in
kgsl fence timeouts. Since CPZ register programing is handled at
the hypervisor, we can safely avoid this serialization by using
scm_call2_atomic() API which doesn't block.

Change-Id: I48ba3e1a682e1027463a1c6b067e6cfcb4a0e8bc
Signed-off-by: Akhil P Oommen <akhilpo@codeaurora.org>