Add support to early mount vendor partition so that vendor modules
can be loaded during early init.
Change-Id: I2233f6c3b8eab180631ca0a0472b586b7ec854f1
Signed-off-by: Imran Khan <kimran@codeaurora.org>
Zero address is considered invalid address, hence changing
the secure domain mapping space to start from a
non-zero address
CRs-Fixed: 2053654
Change-Id: I27300845fb3839372bc38be1d322a7ad0cd0472d
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
Add debugfs nodes to indicate source/sink
HDCP support and authentication state accurately
for DRM HDMI driver.
Change-Id: Ia947c97708c99f23b7b4f23bcad75529b155a42c
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Add a utility function to dump HDMI core registers to
command line. This will facilitate debugging issues by
just calling this API wherever necessary.
Change-Id: I4a6727dca34edf4bb1c240adc7b93f50a7530b98
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Workaround to re-enable BMD while getting battery id was required
when driver is probing. This was done to mitigate driver probe
failures because of failed SRAM access. Getting battery_id from
RR_ADC driver itself takes ~200ms and re-enabling BMD takes 200ms
which increases driver probe time.
With the recent changes in reading batt_id channel from RR_ADC
driver, re-enabling BMD is not required anymore when the driver
probes. Hence move getting battery_id and battery profile to
profile_load_work. This way, there won't be any delay when the
driver probes.
CRs-Fixed: 2062261
Change-Id: Ifeb361e0f82fca65f9570fd1f425c7360445d01f
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
The fence object is allocated in the driver during the buffer queuing
time. When user requested the output buffer fence fd, driver should not
hold on to the fence object anymore because there is only a single
reference to the fence. When user destroy the fence by closing the fd,
the reference is cleared. Driver only needs to destroy the fence object
if the user does not request it.
CRs-Fixed: 2059181
Change-Id: Ic83d93fd3c7f404774007065df02b402adbf80af
Signed-off-by: Benjamin Chan <bkchan@codeaurora.org>
When getting power down setting, there is an untrusted pointer
from a user space pointer.Need to copy to the kernel space first.
Change-Id: I0c66e942453581caac5ea00a54a252840a32972f
CRs-Fixed: 2037398 2037403
Signed-off-by: Terence Ho <terenceh@codeaurora.org>
Use mutex lock before using queuing ioctls like
queuing, dequeing buffers to avoid race condition.
Change-Id: I59c70f2abe3daddf38761a2f17671d3d49f43989
CRs-fixed: 2038086
Signed-off-by: Terence Ho <terenceh@codeaurora.org>
UFS PA errors were observed during data read operation hence Hardware
programming guide have recommended setting QSERDES_RX1_UCDR_PI_CONTROLS
register to value 0x81 which will use CDR DIV4 in all bands and fastlock.
Change-Id: I456354b346aca7abfaa4839d538c5054c3e79fbf
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Remove KP and SY countries from regdb.
CRs-Fixed: 2062687
Change-Id: I41575e415844255d48d2f9e6deca20c1dc656d3c
Signed-off-by: Amar Singhal <asinghal@codeaurora.org>
It's possible for POR to fail and endpoint not enter correct reset
state. Add a timeout when waiting for MHI state BHI transition event
during firmware download.
CRs-Fixed: 2062442
Change-Id: I6af16ec1a514d50b9e984e3fb16696d23e605ba0
Signed-off-by: Sujeev Dias <sdias@codeaurora.org>
There is a possibility that load_gpu() can fail to initialize and turn
on the GPU successfully due to system level issues such as a clock or a
regulator not turning on. In such a case setup_pagetable() returns a
NULL pointer which is not handled correctly in msm_open(). This leads to
a crash in INIT_LIST_HEAD() which tries to access a NULL pointer. This
patch properly handles the NULL return and initializes the list head
only if the gpu load was successful, also adds missing NULL checks to
places in the code where the msm_file_private structure was being accessed.
Change-Id: I6eb85227d928a82c3cf2553fd6645affdd986473
Signed-off-by: Sharat Masetty <smasetty@codeaurora.org>
Don't clear io pad power switch setting when 3V support is faked.
This bit has to be cleared for allowing IO pad to switch to 3V when
Vccq is really configured to 3V. But in case of faking 3V where we
configure Vccq to 1.8V only, we shouldn't clear this bit.
Change-Id: Ib1a9a2eb7189e161e22c8baf588453b3f2632eb7
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
The max20010 external buck supplies VDD_GFX for msm8996-auto
boards. Add max20010 regulator device driver to program the
necessary configuration as per the board specific requirements.
CRs-Fixed: 2062515
Change-Id: I68ad73e0c7cff1a87d218dc4677801f9e0206db6
Signed-off-by: Tirupathi Reddy <tirupath@codeaurora.org>
The patch performs Hypervisor mapping for guaranteed memshare
client's alloted memory at boot time and prevent allocation
for clients with size zero.
CRs-Fixed: 2062478
Change-Id: I96cc7b2014693d97f11411a7b62d2c365f21caf1
Signed-off-by: Manoj Prabhu B <bmanoj@codeaurora.org>
The patch updates global variables of mdlog session
for user PD during different scenarios.
CRs-Fixed: 2058648
Change-Id: I503caddf8e9d504649ee8d81115da60267a5f560
Signed-off-by: Manoj Prabhu B <bmanoj@codeaurora.org>
GPU perfcounters gets reset after a soft reset. Currently
GPU busy stats is using previous values after a soft reset.
This might lead to invalid GPU busy calculations. Start GPU
stats from the scratch after a soft reset to get valid GPU
busy calculations.
Change-Id: Ia38c18ad59f438d724ff4710ee2b350853b3810d
Signed-off-by: Abhilash Kumar <krabhi@codeaurora.org>
When getting power down setting, there is an untrusted pointer
from a user space pointer. Need to copy to the kernel space first.
CRs-Fixed: 2037398
Change-Id: I64032a96e62ddfeec85eebe984d8ba52754f6148
Signed-off-by: Haibin Liu <haibinl@codeaurora.org>
Signed-off-by: VijayaKumar T M <vtmuni@codeaurora.org>
Currently the sysfs nodes for receiving minimum
encryption level reside within the SDE HDMI driver.
Move the nodes to the misc hdcp driver so that they
are available for use for targets using DRM SDE driver.
Change-Id: I94daa981536e56930f7b15eb0ca7b895b9bc9c44
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Add the HDCP 2.2 module for DRM HDMI.
Call the HDCP 2.2 APIs from the SDE HDMI driver and
also invoke the HDCP 2.2 ISR routine from the parent
ISR routine of the SDE HDMI driver.
Change-Id: Ib9a427e5712c827fa93b91fc3a416af61ea83745
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
HDCP 2.2 needs additional helper functions for handling
DDC transactions and interface specific calculations.
Add the necessary utility APIs and also move the necessary
ones to the HDMI utils file for better maintenance.
Change-Id: I215f04991c6ff71ec3e1db7674024e85fbc922e8
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
All sysfs nodes have been moved to the misc hdcp driver.
Make the sde hdcp 1x module interact with the misc hdcp
driver to keep the functionality of repeater topology
intact as its used by wi-fi display userspace module.
Change-Id: Iecf404330695369c17e4afeff864eb6d7b92e372
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Add device node for the MSM HDCP driver which manages
communication between HLOS and TZ for HDCP related
operations.
Change-Id: Iecb4c86ae37ddf7cbc84f962895bbe6e1973b989
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
Currently the hdcp lib doesn't have a context and
just acts as a lib for HDCP compliant interfaces like
HDMI and Display Port.
Make this lib a standalone driver capable of supporting
sysfs nodes to make communication with other modules
like TZ easier and less roundabout.
Change-Id: If693a9d4c8561e6d8c94e236a0fc108c8a65c05e
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
