In the shutdown hook, driver forces adapter to 5V but
there is a possibility that the source change IRQ triggers
and moves back the adapter to 9V.
To avoid this and any race between the irq's and shutdown path,
disable all irqs as the first step in shutdown path.
CRs-Fixed: 2032575
Change-Id: I0d8f59a73b6631e9bf6f5d64dd215c239e7772fe
Signed-off-by: Ashay Jaiswal <ashayj@codeaurora.org>
Module parameter allows to load specific FW used
for FTM testing.
Change-Id: Ic6b4e1596edc0068a6cbd656ec7dfb82746c3f8a
Signed-off-by: Hamad Kadmany <hkadmany@codeaurora.org>
global_state file output is prepared by reading the members of CPU
and cluster data structures. These members are updated under state_lock,
so acquire the same lock while preparing the output.
Change-Id: I8cb3ca6f9100a24f0bc9b293d364ad6aa8aabefb
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
Fix HDCP 1.x state transitions by ensuring that transitions to
authenticating state can be made from both an uninitialized and
a failure state, and making sure that authentication thread is
synchronously canceled before transitioning to a disabled
state.
CRs-Fixed: 2006096
Change-Id: Ic097e16451166f589acbd7c23f87bd9b4c1a65f3
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Current implementation assumes that the first extension
block of the EDID data always has the CEA extension tag.
This is incorrect and can lead to incorrect parsing of
resolution data. Fix this by removing the hardcoding of
the extension tag for the first extension block.
CRs-Fixed: 2027108
Change-Id: I5ef4cdb186591e291b2217db7ccec6f942d79ca4
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Trigger the reading of DPCD if there was a change in the AUX
configuration caused by a failure while reading the EDID.
This is required to ensure the integrity and validity
of the sink capabilities read that will subsequently be used
to establish the mainlink.
CRs-Fixed: 2006096
Change-Id: If3a51b5efd9124fd20dc9860e5f2cdb4e466a2c6
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Ignore the PHY AUX stop error if we receive this error
during an AUX transaction. We can safely ignore this
error and proceed to read the buffered data from the
hardware since this error happens after the data has
been received by the AUX controller hardware.
CRs-Fixed: 2006096
Change-Id: Idb3f0e59b572be565e7572db500699471665f287
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Ensure that all previous interrupt status are cleared before
triggering a new aux transaction, and do not set the
NO_SEND_STOP request bit for read transactions to align with
the hardware programming recommendations.
CRs-Fixed: 2006096
Change-Id: Icc17ac7b09d70ab5b330189be38fa32e3bff850e
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Each block of EDID data is usually 128 bytes long. Reading
each block in a single burst can sometimes result in incorrect
or corrupted data returned from the sink. Add support to read
EDID in shorter burst sizes and set the default burst size
to 16 bytes.
CRs-Fixed: 2006096
Change-Id: I681f2d2eb01de11eadc1857dd5089cbe35befad2
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Update the EDID read sequence to first write the offset to
request the appropriate EDID block prior to reading the EDID
block data. In addition, write the correct segment address
when reading more than two extension blocks.
CRs-Fixed: 2006096
Change-Id: Ic4b2bd4d4cf9da5e247c5735b4e768b9e2b87b27
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Retry AUX read/write transactions that have failed either due
to the AUX controller hardware indicating an error via the ISR,
or due to a software based timeout while waiting for transaction
completion. The transaction retry strategy is as follows: first
repeat the transaction using the same PHY AUX settings, and then
retry the transaction using updated PHY AUX settings if repeating
the transaction has failed.
CRs-Fixed: 2006096
Change-Id: Id9c3c7ae1ab320540545b9c178d947a3cd023079
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Make the parsing of the AUX controller configurations more
extensible by adding support for multiple settings per configuration,
as well as the parsing of the register offsets. This enables the
extension to support different targets, and combinations of AUX
controller settings depending on the use case.
CRs-Fixed: 2006096
Change-Id: I4e6b623a4d9fafcfcc89477dfa57880eb798c350
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Increase the AUX controller hardware timeout limits to the
maximum value supported. This increases the AUX controller
hardware's robustness in handling transactions in the face
of differing timing from one DisplayPort sink to another.
CRs-Fixed: 2006096
Change-Id: Ia0da13720526a96f3a88a849043a3ffbb1185cf7
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Add interrupt handling for PHY AUX errors that might
happen during AUX communication. Read the interrupt
status and clear it before initiating the next AUX
transaction. This will help identify any issues that
result in AUX transaction failures.
CRs-Fixed: 2006096
Change-Id: I0733bfb163c8c3c108002fbe4309e36dc105ccb8
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
A branch device usually notifies a change in downstream connections
using the HPD IRQ pulse. Handle this by checking for a change in
downstream sink count and appropriately handling EDID reads. It is
also possible that the branch device may not have any local EDID.
In such cases, when the downstream sink count is zero, do not read
EDID.
CRs-Fixed: 1112711
Change-Id: I230560c995d7c3b395e37aef5483e5468e1d1dec
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Implement a failsafe mechanism that will prevent the DisplayPort
driver from entering an irrecoverable state in the face of
delayed power off/on events from the display framework. For example,
the driver must indefinitely postpone the handling of other events
from the sink until the power off event is complete, ensuring that
there is no interleaving of events that could lead to a bad state,
or un-clocked register access.
CRs-Fixed: 2006096
Change-Id: I1bfa887ba5ee94cbf44c87aa0d60766cafc854b4
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
EDID utility always lists the default resolution as part of the list
of available modes. If the client does not specify a default resolution
it can result in incorrect behavior. Ensure that the Display Port
driver always initialized with a default resolution.
CRs-Fixed: 2030915
Change-Id: Ib6c75d655e6fcbd5c792ca1aa9da6b08ba2f416e
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Currently, RATE_CFG registers are updated when the auto resonance
is enabled and no errors are seen with auto resonance or the play
rate is within allowed limits. However, it is getting updated
multiple times with the same play rate multiple times. This adds
up some delay in a critical path when the haptics duration is
small. While at it, add APIs to do multiple read/writes to the
peripheral registers like AUTO_RES_LO/HI and RATE_CFG1/2.
Change-Id: I1b5625cc1f9aa3f07552a6b4ef23f2d34241b6a1
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Currently, variable "val" is used for caching the last value
written to EN_CTL1 register and also it is used to read STATUS
register. This causes an incorrect value being written to
EN_CTL1 register when the module is disabled. Fix it.
CRs-Fixed: 2032850
Change-Id: Ib3cbe1b41962e5dff38d9c4d343d8f3f43fc0009
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Add support for sensor and Region of Interest (ROI) events.
Create new ioctl to receive sensor data from userspace.
Create binary sysfs entries to expose data from kernel to
userspace.
Change-Id: Ifa76f08eb293c3da0618c22c07e65976319ea926
Signed-off-by: Alex Sarraf <asarraf@codeaurora.org>
Update the open-loop and closed-loop VDD_APC virtual
corner margin adjustments to match the latest characterization
guidelines.
CRs-Fixed: 2006091
Change-Id: Ib7a0657302b4ae76107ba53b21f9bd4b39f104df
Signed-off-by: Osvaldo Banuelos <osvaldob@codeaurora.org>
Add debug events for CPU_STARTING and CPU_DYING events when a CPU is
hotplugged. The events will help correlate with the TZ events using the
ARM timers counts for debug.
Change-Id: Ieb6449575b4a95974b787ceea430fc5362c77c29
Signed-off-by: Mahesh Sivasubramanian <msivasub@codeaurora.org>
If any data traffic is ON, when SSR is triggered there is a crash
since we allowed HTT packets to go through.
Stopping these packets to go through fixes this crash.
Also changing the QMI client ID for the qmi client.
CRs-Fixed: 2033262
Change-Id: I1236ef901fd8943f228497a79f11feaec1c16d8e
Signed-off-by: Rakesh Pillai <pillair@codeaurora.org>
Configure APC0/1 CPR instances to operate in closed-loop operation
to maintain an optimal voltage set point on the APC0/1 voltage rails.
Also, add static open-loop and closed-loop voltage margins as per
the voltage characterization results.
CRs-Fixed: 2032723
Change-Id: I1a01d78252dc71ca7def251e99581cad3714cdd1
Signed-off-by: Tirupathi Reddy <tirupath@codeaurora.org>
Set resetting flag early when stopping AP to avoid
disconnect events as a result of disconnect command
sent during AP stop procedure.
Change-Id: I59647d5775e68705b361b1aaacbc6a6088e9c813
Signed-off-by: Hamad Kadmany <hkadmany@codeaurora.org>
Existing code that ignores connection events during
reset flow will never take effect since it locks the
same mutex taken by the reset flow.
In addition, in case of unsolicited disconnect events ignore
those as well since device is about to get reset.
Change-Id: I132040822b8c1b3d1d92965d578ed5bee0b3d1cf
Signed-off-by: Hamad Kadmany <hkadmany@codeaurora.org>
The minimum supported voltage on the SMPS5A is 1.224V, update
it accordingly.
Change-Id: Ib973febdf5b719836f5fb54e6cd6e797e7101e3e
Signed-off-by: Anirudh Ghayal <aghayal@codeaurora.org>
To get the 2MB aligned IOVA section mappings, set
ARM_DMA_IOMMU_ALIGNMENT to 9.
Change-Id: I3254c39e4864c32d6f25a4d416b1b7aec6756ecd
Signed-off-by: Charan Teja Reddy <charante@codeaurora.org>
CONFIG_CPU_BOOST config option is only enabled for
msmcortex-perf_defconfig. Enable it for msmcortex_defconfig too.
Change-Id: I31d42aa4ac3a78e6bc3c52aa0ba7cebd8c5a7893
Signed-off-by: Rohit Gupta <rohgup@codeaurora.org>
Currently both sde_crtc_atomic_begin() and
sde_crtc_atomic_flush() add the CRTC state event
to the cached sde_crtc->event.
This has a potential NULL ptr issue in the
case of vblank event firing in between sde_crtc_atomic_begin() and
sde_crtc_atomic_flush() because the upstream DRM vblank API
send_vblank_event() doesn't consider the case when the VBLANK
interrupt could have already freed any pending vblank events.
Remove the caching from sde_crtc_atomic_begin() to avoid this
condition.
Also make sure that a page_flip event was indeed submitted before
signaling the complete_flip() by setting a PENDING_FLIP flag right
after HW flush.
Change-Id: Ib201d2851e57bf22ec1f00814fc2e4dd2f35bfa1
Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
