During LDAR dump, LRME, FD registers are dumped. Due to race
conditions or concurrency, there might be a situation
where the hw is in OFF state and still the registers are
read. This will cause aborts on core.
This commit protects the dumping logic with the hw_mutex for
the two drivers.
Also it fixes one crash scenario while dumping the FD registers.

Change-Id: I9b58b43a005a717b059a1a59d64b1810969004aa
Signed-off-by: Gaurav Jindal <gjindal@codeaurora.org>

Change-Id: I39070ea2f33842dfc61af9d0241862730a5750c4

Force a disconnect if the simulation client disables simulation
before disconnecting from the simulated sink. This ensures that
the driver will not erroneously attempt AUX transactions in
subsequent interactions after simulation is disabled.

Change-Id: Ibc581deafe46753c514bccc70ba5c953c8d49bd8
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>

Remove session lock for power state update to improve functionality
for fast hotplug use cases. This allows the controller and aux modules
to be updated asynchronously based on hotplug state. For example,
in disconnected state all aux transactions will be aborted.
Updates to the power state can be done without the session lock
since this path is only exercised during compliance testing.

Change-Id: Ibe175624ef56d9f7b953b9b54928682f1cac906a
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>

Reject composition if any seemless transition such as VRR, dynclk
is requested during power on/off commits.

Change-Id: I731bfc06b3bd1e7ae920c12cbc68f95f5cc01687
Signed-off-by: Kalyan Thota <kalyant@codeaurora.org>

* commit 'a4ad2182':
  msm: camera: isp: Disable the PPI hw during PPI overflow case
  msm: camera: isp: Move CSID error propagation workq create code
  msm: camera:icp: Fix deadlock during ICP pagefault handling
  msm: camera: isp: Reset sof counter before enable the CSID path
  msm: camera: isp: Disable tasklet once
  msm: camera: isp: Mask unused rdi interrupts
  msm: camera: isp: CSID error propagation
  msm: camera: isp: Skip reapply of DMI config on bubble recovery
  ARM: dts: msm: Add svsl1 level for CSID clock in atoll

Change-Id: I1fb0374d7363d9c459af780e31b9380b13696d51
Signed-off-by: Sridhar Gujje <sgujje@codeaurora.org>

The original IOMMU mapping space starts from 0x80000000 with
size 0x60000000 bytes. But actually there is no such limitation
for NPU IOMMU which can start from 0 with size 4GB.

Change-Id: Ic988ca1ae242438d0c5eaede2a3eaab55f8e508e
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>

This fix configures the maximum HOLB drop timeout
value.

Change-Id: I04ea5aac20757a4cc31f251dc9a4d6dc8fdfbebe
Signed-off-by: Praveen Kurapati <pkurapat@codeaurora.org>

In mdss_dsi_cmd_write, a failure in copying the cmds to
'string_buf' can cause an early return. In this case,
the 'pcmds->string_buf' won't be pointing to a valid
buffer. This can lead to use-after-free and memory leak.
To avoid this, assign the newly allocated buffer to
'pcmds->string_buf' after returning from krealloc call.

Change-Id: I286f12c86078d1989cb09453c8a395a4ad94b324
Signed-off-by: Nirmal Abraham <nabrah@codeaurora.org>
Signed-off-by: Vishnuvardhan Prodduturi <vproddut@codeaurora.org>

Don't unregister listener when data->released is true and only update
listener  id in private data when listener is registered successfully,
to avoid unnecessary unregister operation.

Change-Id: I10c3b353ae93a0366d6ea00416cbbe35a6d4ec36
Signed-off-by: Zhen Kong <zkong@codeaurora.org>

To prevent kthread unloading app/unregistering listener and freeing
private data when qseecom_release thread was still accessing it,
make change to let kthread unload and unregister after qseecom_release
was  complete.

Change-Id: I48a019aa7fd2707b2b95c29f633fa9d519600df6
Signed-off-by: Zhen Kong <zkong@codeaurora.org>

To avoid race condition on app_access_lock between client thread
loading app and waiting on a listener response and another thread
unloading  app during device reboot, make change to add unloading
app requests into a waiting list and start a separate kthread
to unload app when lock is released.

Change-Id: I24391bc34bb9a076b269067cfd885d62b0b53e3e
Signed-off-by: Zhen Kong <zkong@codeaurora.org>

Enable hid-generic to handle hid devices
which are not attached to any of the drivers.

Change-Id: Iceca82d1407d34bbfcfc4a3d0da772d024515944
Signed-off-by: rbandi <rohitbandi@codeaurora.org>

There is a potential race between evict() and __mark_inode_dirty()
observed with EXT4 FS. The ext4/jbd2 context is marking the inode as
dirty while at the same time evict() is happening on another core.
Due to this race, __mark_inode_dirty() is adding the inode to
bdi_writeback IO dirty list, after evict() path has removed this
inode from the writeback list. This results NULL pointer dereference
issue when accessing inode->i_wb from writeback thread as below -

locked_inode_to_wb_and_lock_list+0x2c/0x2a0
inode_to_wb_and_lock_list+0x24/0x30
writeback_sb_inodes+0x2cc/0x444
__writeback_inodes_wb+0x78/0xbc
wb_writeback+0x1c8/0x3d0
wb_workfn+0x1a8/0x4d4
process_one_work+0x1c0/0x3d4
worker_thread+0x224/0x344
kthread+0x120/0x130
ret_from_fork+0x10/0x18

Fix this by checking for the inode->i_state I_WILL_FREE or I_FREEING
before and after adding to the writeback list. If it is observed to be
in freeing path before adding, then simply return. If it is after adding
to the writeback list, then remove this inode from the writeback list
in __mark_inode_dirty() context itself.

Change-Id: I1e5f15d81dae01f0dd7a108b1861adc83105ed7b
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>

To improve read calibration data latency use a minimum
but enough timeout value. To improve consistency in
controlling IMU stream use a different value.

Change-Id: Iff196d055f0202e642de218020605f7797e3ae4a
Signed-off-by: Rohit Bandi <rohitbandi@codeaurora.org>

To improve read calibration data latency use a
minimum but enough timeout value.

Change-Id: I629ae61ca5005d0f3d0a48e9780ad0940a1ca860
Signed-off-by: Rohit Bandi <rohitbandi@codeaurora.org>

Add support for 128 byte format to comply
with optimized viewer requirements.

Change-Id: Icfa7ca1cadf3f18533912f403af50532eea47c60
Signed-off-by: Rohit Bandi <rohitbandi@codeaurora.org>

Refactoring code for hid-qvr to improve
readability.

Change-Id: I00f679b6e1cf6f59d576243757cba1a88c9286b6
Signed-off-by: Rohit Bandi <rohitbandi@codeaurora.org>

Adding new vendor id and product id for
hid-qvr support.

Change-Id: I777568f1a68b15345695e09d2ad89e9230508886
Signed-off-by: Rohit Bandi <rohitbandi@codeaurora.org>

Disable the PPI interrupt to stop IRQ storm during overflow errors.

Change-Id: I4fac7c521ef6774933d836ad326ed2f4091a6322
Signed-off-by: Chandan Kumar Jha <cjha@codeaurora.org>

Register node in debugfs needs to free buffer in release function
which is allocated while reading registers.

Change-Id: I39a3bd94e6659ba5201c0984248558bd2a3e887c
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>

* quic/dev/msm-4.14.c2-display:
  msm/sde/rotator: disable rotator clock during pm suspend

Change-Id: I96492e839be320b70632c25a43cd8c8a09bba3fe
Signed-off-by: ckosuru <kosuru@codeaurora.org>

The page->mapping will be set to NULL when it is truncated.
The same metadata page if it is currently under processing by
ext4/jbd2, then it is possible for pfk_bio_get_inode() to see
inconsistent page->mapping, for the requests submitted by jbd2
on this page. To avoid this inconsistency, use cached mapping at
all places in pfk_bio_get_inode(). It doesn't matter whether the
pfk driver sees the mapping or sees null mapping for metadata
requests because, pfk_allow_merge_bio() will always return true
for 2 journal metadata requests and  for one journal and one
non-journal request, it always returns false (independent of
the metadata page mapping state).

Change-Id: Ia452f972346e1ba3b25560ffa140d023dbb3409d
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>

In ioctls like kgsl_ioctl_submit_commands(), if both syncobj
type and cmd/marker/sparseobj type are submitted, the syncobj
is queued first followed by the other obj type. After syncobj
is successfully queued, in case of failure in get_timestamp
while queuing the other obj, both the command objs are
destroyed. As sync obj is already queued, accessing this
later would cause a crash.

Compare the user generated timestamp with the drawctxt
timestamp and return early in case of error. This avoids
unnecessary queuing of drawobjs.

Change-Id: I336c95c42ab1075d7653cba02772f92c918c884c
Signed-off-by: Archana Sriram <apsrir@codeaurora.org>
Signed-off-by: Harshitha Sai Neelati <hsaine@codeaurora.org>