When LTM is turning off, user space will turn off LTM histogram and
then wait for LTM hist off event from driver to change the state
from offpending to off. But if device enters IPC during this time,
driver will not notify user space with LTM hist off event and LTM
will stuck in offpending state. This change remove the check
before LTM hist off event notification to avoid LTM state stuck issue.
Change-Id: I073adabf72c6454acb8ab4d333c0547723523781
Signed-off-by: Ping Li <pingli@codeaurora.org>
Connector state in crtc_atomic_check is still the old state, use
connector_mask to update the actual connector in the list to fix
zero connector issue.
Change-Id: I91541063acbb19beced3004a030c7f27abf9ee50
Signed-off-by: Xiaowen Wu <wxiaowen@codeaurora.org>
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
Use MDP scratch registers to pass drm mode index information
between the VMs. Expose the APIs which LA VM can use to set
the current mode for all active displays before handoff.
LE VM relies on this information to set the corresponding
display in correct mode during handoff.
Change-Id: I4569dd58e953e588bca816ac718335d3f3f7e29b
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
The MEM & IRQ lend handlers signals the resources
are lent from the primary VM. Reserve the SDE HW
resources in trusted VM using the cont-splash path
and setup the SDE & DRM software states. Set the
DSI connector status to active for the handoff
display during the start and disable it at the end
of the session.
Change-Id: I75b6df735eb5609cc521808aca0ddfb0b3993e84
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
Install backlight property only for primary VM and avoid
backlight updates when HW is not owned by the VM.
Change-Id: Ief831661e2b9419a0b9d392742bf96300ecc0d48
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Check against the HW ownership before servicing vblank
control requests from clients.
Change-Id: I22544817fde72afa7f668fbc08caa28abe0330ee
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Extends atomic check to validate TUI transition request aginst
SDE and non-SDE client states. Triggers HW assignment after
making sure current frame is completely flushed out of HW
pipeline. Initiate the acquire after TUI end after restoring
the software states.
Change-Id: I6ceccd44c18526ff818723b5631b66af3ff083c1
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
client_pre_release/client_post_acquire hooks in vm_ops allows non-SDE
clients to register for callbacks to be invoked
- before releasing the resources from the current VM so that
the client can updates the respective sw state to indicate
the absence of HW ownership
- before the first frame push after acquiring the HW to update
reset the respective sw state to indicate the presence of
HW ownership and resume operations
Change-Id: I93e5a2b43c79b4e7cc28e6e0129d8091aa0e508d
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Add VM layer files to handle VM specific operations
for primary and secondary VM's. They interact with
Hypervisor RM modules to LEND/ACCEPT/RELEASE/RECLAIM
HW resources. They also handle notifications for each
of the above Hyp RM operations to acquire and update
the SDE software states.
Change-Id: I5982f4bf56550dc464797c62cb356be39925b21c
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Creates a deep copy of a struct which previously was shallow.
Also adds null checks which prevent segmentation faults.
Change-Id: I9155f632736fdb30e31f28f55ebe92954956a82d
Signed-off-by: Orion Brody <obrody@codeaurora.org>
This change adds 2 vsync delay after changing the timing engine
parameters namely the porches to update the refresh rate.
This ensures that panel vsync is updated as per new timing before
modifying it again.
Change-Id: I5866ea2f6f2e68bc8ce7435c4a5dbe27d8ebdd91
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
When hpd disconnect is received, the dp driver sends an uevent to
usermode and waits for the usermode to tear down all external displays
before turning off the dp link. If the wait for uevent completion
times out or exits prematurely due to a race condition with the
connect path, it is possible for the dp link to be turned off
prematurely, with displays still in enabled state. On a subsequent
hpd disconnect, the driver attempts to disable the display causing
this issue since the link has been already turned off.
This change adds a check for display status after uevent handling and
before turning off dp link, to make sure the displays are always
cleaned up before the link is turned off. Any future attempts to
disable from usermode will be ignored.
Change-Id: I404ad90b037fede9c9469a2fd16ca1fbdbd36b7c
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
When dp disconnect is triggered, as part of the teardown process, the
driver sends a notification to audio driver and waits for the audio
driver to complete before proceeding. When back to back disconnects
are triggered currently there is no check if a disconnect is already
in progress. So the driver ends up reinitializing the signal
completion that the previous disconnect thread is waiting on causing
this issue.
This change adds a check at the beginning of disconnect handling and
bail out if a disconnect is already in progress.
Change-Id: I2eaee339408d9dbc120f8efc14d86fca0ebb0758
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
This reverts commit e0b15f4dbc.
Needed display configs are enabled in trusted vm configuration.
We don't need this change anymore.
Change-Id: Id3b591ce5cfebf5f560a0b60816f7fd76253c347
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
In current SDE driver when misr is enabled, for each commit in
encoder kickoff stage misr is configured for both lm and interface
misr blks. This can clear misr data before client could collect misr.
This change avoids misr data clear and configures misr based on
user input.
Change-Id: I85fc19c78afc6d01346219250c82f2ada824eb0d
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
Commit d46cae019e ("disp: msm: sde: trigger pm_qos vote
with irq enable") moves the pm_qos vote with irq
enable/disable state. Such irq enable/disable call may be
triggered from atomic context and lead to scheduling issues
due to mutex_lock usage in pm_qos APIs. This change moves
the vote with encoder idle pc to allow lock usage with sleep.
Change-Id: I2d22566fbfb5399c5d2d2a4efe882a1928cfbbf8
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
This change adds DSI pll support for 10nm architecture.
Change-Id: I3819dd828dbcc168b115bd718c5d656ea9fd12c8
Signed-off-by: Rajeev Nandan <rajeevny@codeaurora.org>
Add required changes to makefile and enable the config
keys for holi target compilation.
Change-Id: I4b48e79c8b26c45f2f54fc1ac02a6e322bda07b9
Signed-off-by: Krishna Manikandan <mkrishn@codeaurora.org>
Reset the QMP PHY into DP-only mode only when the adapter/cable
connected is not multi-function capable. We remove the
peer_usb_comm check since this flag is not being updated when
the cable is connected.
Change-Id: Iafbdf73b3ef827b757b51bfadfa6fa16b840ac27
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Add support to parse ddr type and corresponding
highest bank bit for configuring UBWC parameters.
Change-Id: I54f79a60f55fd6b146783c8f53fce1bf6c8d8d0a
Signed-off-by: Krishna Manikandan <mkrishn@codeaurora.org>
Few v4l2_ioctl_ops are deprecated on kernel upgrade. Made
changes to use the new simulated ops. Addressed compilation
issues on enabling rotator on 5.4 kernel.
Change-Id: I37add012306da433c4d7f8aaada3089b6ecdd819
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Add revision checks required for rotator for holi
target.
Change-Id: Ie35d888d7681c97d681e2b2d6b564e0a7e1f11f2
Signed-off-by: Krishna Manikandan <mkrishn@codeaurora.org>
Add trace logs in display early wakeup function for performance
profiling.
Change-Id: I63392417f03eac60dba1c43cd71fe5032ba59ed3
Signed-off-by: Lei Chen <chenlei@codeaurora.org>
Display clocks and IRQs are disabled during idle state
on command mode for power saving, and will be enabled
when a new frame commits to display driver. But enable
display clocks and IRQs will cause some latency.
So add a new SDE custom IOCTL for user-space to early wake
up display before first frame commits to kernel.
Change-Id: I6ca0188d321c4964f29c46e588b64d06b9634c59
Signed-off-by: Lei Chen <chenlei@codeaurora.org>
Add this new IOCTL for user-space to wake up kernel display as soon
as it gets a new frame update.
Change-Id: Icd45b47e99ca0ea996e46ae8335a29f6636bf406
Signed-off-by: Lei Chen <chenlei@codeaurora.org>
Add support to read register ranges and IRQ lines that needs
isolation during VM switch in dsi ctrl and dsi phy.
Register for VM resource collection event providing the
callback function.
Change-Id: I5eae4699b0a97ffed438627ccea855c401b3fbeb
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Display in trusted VM works in a handover mode i.e, all
display components need to be initialized by the primary VM
before switching to trusted VM on the usecase boundary.
That makes it a hard requirement for the trusted VM not to
re-program any of the display configuration registers before
it could hand the HW back to the primary VM.
Also, most of the linux framework drivers including pintrl, clocks
are not enabled in the trusted VM.
In order to address the above two limitations and to control the
probe/commit sequences, this change adapts the same path as that
of continuous splash to bypass some of the critical functions
which are identified to be HW intrusive. Based on the DT property
read from SDE handle, dsi drivers will choose to bypass
enable/disable HW state updates when executing in the vm environment.
Change-Id: Ica71c924d189ab65fe3be5b0ac870633e3b749e1
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
Avoid sharing and have dedicated dsi HW ctrl ops across versions
even if they are identical. It helps to stub out a specific
version of ops if needed.
Change-Id: Ic14c206b526ebe03dd5f47f62701820e7e542670
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
Trusted VM will be granted access to MDSS HW dynamically on
usecase boundary. As a result, all the attempts to access
HW before the assignment, including the probe time access
will result in Stage 2 faults. This change skips the
PLL clock registration during probe as the clocks will not
be controlled by the VM.
Change-Id: I326f4a775796cd95dcf398449b08f2682e4aca43
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
