Update the voltage swing and pre-emphasis settings based on the
latest hardware programming guide.
Change-Id: If90db2833aba2bd0613276eff22f850bf34859e5
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
sde_connector_init will return an ERR_PTR if connector
creation failed, so need to use IS_ERR_OR_NULL to check
the return value.
Change-Id: I4fee5a624261898bbd079c54705e6eaebc71bac6
Signed-off-by: Yuan Zhao <yzhao@codeaurora.org>
Current dp_mst_bridge has variables that are accessed by both
check phase and commit phase, which causes racing issues. This
change will add private state to dp_mst_bridge to separate check
phase and commit phase.
Furthermore, this change is a partial rollback of commit 2446602565ec
("drm/msm/dp: add private state to dp_mst_bridge") where active_enc_cnt
is removed. In this change we retain the encoder availability check in
mode_valid.
Change-Id: I8ac05cf5f1755375e4e9f34e42dbaea1d23bac64
Signed-off-by: Xiaowen Wu <wxiaowen@codeaurora.org>
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
The single "port" refcount has been split in 2 on 5.4. MST sim
layer is only getting the topology refcount but never initializes
or obtains references for the memory allocated for this port.
Add the new refcount logic required on 5.4 to MST sim layer.
Change-Id: I6e25c048fa26352c4fb718996514a1ca91432408
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
Log the status of the call to kobject_uevent_env so we know
whether the uevent was sent successfully to the userspace.
Change-Id: I4fd5f9b51d9771dab399e322d678411f824c8aba
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Return ERR_PTR with -ENODEV when the USBPD module is not
enabled. This allows the DP driver to handle the failure
correctly and prevents any potential NULL pointer
dereferencing in the clean up path.
Change-Id: I4fbc559276bebe770a27935243a59c10c1bb5491
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
When CONFIG_DEBUG_FS is disabled skip creating root directory.
Change-Id: I97bed925392b781b73687164ca55e6cb09f951fc
Signed-off-by: Rajat Gupta <rajatgu@codeaurora.org>
Currently DP driver determines if a mode is DSC capable
based on a DTSI entry and the required number of DSC
to support it. This approach does not scale when there
is an overlap in DSC requirement between DSI displays
and external DP display, thus causing one of the display to
report modes that cannot be supported.
This change compares the resources reserved for DP driver
calculated at initialization time and the currently available
ones to determine the correct number of resources that DP driver
can use. It also adds DSC and topology filtering logic and moves
DSC hardware specific from DP driver to SDE driver.
Change-Id: I8e601de33422b7c6d786826f7bfe152c4af8a6b5
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
Primary and secondary dsi displays are built-in displays
and they are supported during all concurrency usecases
without resource allocation failure. DP mode filter
logic should provide supported mode information based
on free mdp resources after dsi resource assignment.
Change-Id: I3a9637a91ea1ffcc31997e25caff7f13605283ac
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
Set the aborted flag during module initialization so the default
behavior is to error out on aux transactions. The flag will be
reset when there is a hot plug connect and a sink device is
actually present.
Change-Id: I9e31b9293e817ed951ba5464d403e7178615418f
Signed-off-by: Yuan Zhao <yzhao@codeaurora.org>
The DRM framework has added support to check for and calculate the
correct checksum for test cases in which the EDID is corrupted e.g.
test 4.2.2.6. Update the implementation to use the correct checksum
that is available as a field in the DP connector structure.
Change-Id: Ia9f65926e2c0fb3444aa87a4ac97850fbb4d70f3
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Register for USB cable notifications so we can manually trigger
a disconnect event if the USB cable is disconnected during a
simulation.
Change-Id: I08a9313bcde3e74f2a76804e934423e0fc2d6ab6
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Add event logs to capture HDCP state transitions and errors.
Change-Id: Id82049b93eb2e8bf84faddebaa0959f8cda1c91f
Signed-off-by: poyewu <poyewu@codeaurora.org>
Some sinks need more time to stabilize after synchronization and
before it can handle an HDCP authentication request.
This change adds the delay for better interoperability.
Change-Id: I4fffb9024f782dd68e3a883208cb8e9d0c5350a1
Signed-off-by: poyewu <poyewu@codeaurora.org>
If a DP MST display is connected after a DP SST unplug, the
DISCONNECT_NOTIFIED state is not cleared and CONNECT_NOTIFIED
state is not set. These states are not relevant for DP MST and
hence skipped during the send_notification. But the
DISCONNECT_NOTIFIED state that was not cleared from the previous
SST unplug will prevent the wait for completion when DP MST is
disconnected. This will trigger the host_unready to get executed
before the dp_display_disable happens for the DP MST displays.
As the result, the dp_display_disable exits early causing all the
stream clocks and the panels objects to remain active even after
unplug.
As CONNECT_NOTIFIED and DISCONNECT_NOTIFIED state are not relevant
for DP MST, clearing them if DP MST hpd detected in dp_display.
Change-Id: Iaf0e762633eb276e96cb860fda480adb04718eb9
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Sometimes the link clock is not getting locked when dp
display is resumed from PM suspend. This is causing NOC
error when registers in link clock domain are accessed
without proper checks.
The set_colorspace, config_hdr and update_pps connector
ops are accessing the registers in link clock domain
without checking if the necessary clocks are enabled.
These changes will ensure the DP stream clocks are not
enabled when link clock enable fails and prevents the
connector ops from accessing the registers whose clocks
are not enabled.
Change-Id: If89d2552ee0ce96493ee8c1666d7677221705d9a
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
While running DP simulation script for stability and the rare
occasions when link clock fails to enable, the ctrl power_on
flag is wrongly set indicating the link is powered on. This
flag is used in the ctrl sub-module to check that the link
clock is on before accessing the registers powered by the
link clock. When this flag is wrongly set, the link registers
are accessed while processing the cable disconnect to cause a
NOC error.
The change will ensure the power_on flag is set when the link
clock is on so that the access to registers in the link clock
domain can be correctly filtered by the power_on flag.
Change-Id: I40af054a5738172f5ea86079a9258518f8fdd44e
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Exit stream enable on cable disconnect considering that the
controller and display modules would have transitioned to aborted
state. It is possible that the disconnect might have prevented
the link clock from being enabled, so we have to skip any register
programming in the link clock domain. We remove the unprepare
call from the DRM wrapper error handling since that path can lead
to unclocked register access when the link clock is not enabled.
Change-Id: I92595c0824193df63c2746bc8cd644f59c33604a
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Deregister the altmode notifier in the client teardown path since
the client will no longer be available to handle altmode
registration.
Change-Id: I06c6c61d4eb0b0fc9e9c1df1da4a0c86e4556de1
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Enable compilation of DP MST feature. Also ported drm mst implementation to
use updated API definitions. Changes to the API include; removal of hotplug
callback; separate refcounting for topology and memory allocation in mst
port; pass the entire port instead of slot count in
atomic_release_vcpi_slots API.
Change-Id: I951c2978339229ea7e124c2addc114ea5d3cbb3d
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
This change updates resource topology mapping tables and includes
logic to compare compression types for dsc vs vdc.
Change-Id: I1735edeb07aec8ed0065f84ac0824c58158412f3
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
Update the altmode client registration by replacing the notifier
mechanism with a callback.
Change-Id: I550e8c1b58b5ccfd10e46f240ee54a8acdfe9bec
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
This change sets the regulator voltage to zero for all
those regulators which support voltage scaling before
disabling the regulator. Failing to do so causes the regulator
vote to set to the last min value specified in the previous
set voltage call.
Additionally modified the config regulator api and renamed to
get regulator because of the above change.
Change-Id: Ie4c4842db0c08a4d98926ed79503cf84e4d5762a
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
Make msm_drm into single module and all child driver
registers and unregisters are handled from parent's
register and unregister respectively.
Change-Id: I017513d1de3b6b25dd5543d7fa7741c0bac1740d
Signed-off-by: Shashank Babu Chinta Venkata <sbchin@codeaurora.org>
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
Signed-off-by: Chandan Uddaraju <chandanu@codeaurora.org>
This change addresses out of range and null checks in
sde and dp driver.
Change-Id: I142196d7394f0bf0abab1bfa89abfd784a5521c8
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
If DP is running in mst mode, DP host should not power down during
unprepare call as aux need keep alive to handle further hot-plug
sideband message. Unless it is a PM suspend scenario where it is
required to deinit the host.
Change-Id: I208894e1cbc75e727044c196194b0a4e3936921e
Signed-off-by: Xiaowen Wu <wxiaowen@codeaurora.org>
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
During suspend, there might be a chance that lib auth work is queued
but not yet started. So during pm cycle it might start execution
before host_init which can lead to noc error while accessing dp_aux
registers. To prevent that set abort flag to abort lib auth work and
set interrupts before host_deinit.
Change-Id: Ie2c3ac9b0846644d3c2b37f410c341659b030c16
Signed-off-by: Rajat Gupta <rajatgu@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Ensure host_init is executed when the DP display driver is resuming
from a video session teardown. The host_deinit is executed when DP
video session teardown is initiated by the userspace while DP cable
is connected. If the DP display is disconnected from dongle while
the video session is in teardown, the host_init is not executed
when the session is restored. All subsequent display connections
will fail because the core clocks are not on and aux abort is not
reset.
This change will ensure host_init is executed whenever the
host_deinit is executed during DP video session teardown by userspace.
Change-Id: If1ae3eb731a96bd929cd023733c098b513436a1c
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
The hpd notification is skipped in connect_work if there is an
attention_work pending. So, the notification has to be sent in
attention_work.
Some dongles issue multiple attention events with LINK_STATUS_UPDATED
while the connect_work is in progress. So, once the link training
is successfully completed the LINK_STATUS_UPDATED attention event is
considered handled and the attention_work scheduled to execute after
the connect_work will return early without handling the
hpd_notification cases.
This change will ensure the hpd_notification is sent from
attention_work in such cases.
Change-Id: I5ef6781354b80842d2a6f44360cde25ee6d2b0b7
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Change to update the voltage swing and pre-emphasis settings for
lito to support dp phy cts.
Change-Id: I97f0b9882291d7d52e10ef4bef6ea6a23780e12a
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
The DP display driver issues an audio disconnect notification
after the video disconnect notification. Sometimes audio driver
is waiting for the video commit to complete before turning off
the audio engine. In such cases, there is a brief screen freeze
observed on the primary. In other cases, when the off call from
audio driver gets delayed, a momentary screen freeze is seen.
The audio notification happens as part of the display commit call.
The order of notification is video and then audio while
processing display connect and the same order is followed for
disconnect also, which is causing this issue. These changes will
modify the order of notification to send audio disconnect first
and then video while processing DP cable disconnect.
Change-Id: I8bece39c164620b319d971e5a2597d9dc187566e
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
When panel object for a dp connector is destroyed, the instance
of dp_panel in the connector object is made NULL to prevent
further panel object dereferences.
Change-Id: I63409ab8c470f14cf9c032e373cb0c0872577504
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
When DP_LINK_STATUS_UPDATED irq_hpd occurs, the mst audio instances
are turned off, irq_hpd handled and audio instances are turned on
again. While turning on, an audio instance different from the ones
turned off is being sent, causing ops mismatches eventually leading
to a NOC error. These changes will correct the pointer to the audio
or the panel instances to match the ones that are turned off.
Change-Id: Idd9c013fd64cd6db06bbec34f5083029e4ab522c
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
The min_enc_lvl update is asynchronous to the hdcp2x state machine and
can come anytime. In an instance where authentication with the sink is
continuously failing, the enforce engine blocks the hdcp22 secure content
playback and resets the min enc lvl. This is causing the sde_hdcp_2x_main
to perform a send message operation out of sync with the current
authentication attempt. This is causing the state machine to go into bad
state by changing the next message id, sometimes causing a pagefault error
in the sde_hdcp_2x_main while cp_irq is in progress. This change will
proceed with the min_enc_lvl change only when the mainlink is hdcp2p2
authenticated.
Change-Id: I977ac3a083b9593f8f7e760d4e477d4af424d5c9
Signed-off-by: Sankeerth Billakanti <sbillaka@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Modify the macros to use default kernel logging function if DRM log
levels are not explicitly set. This helps in being able to selectively
enable function or file level logs.
Change-Id: I4f894a2ecf314a3a5fc0bb22a75a0e10dae31e07
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Delay the hotplug connect notification to check for any test
requests issued by the sink. This will avoid unwanted connection
event callbacks.
Change-Id: Id7a04dbbb5ad5b86c3791ac2db5de2c92e98fc46
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Upon completing the link setup, check to see if any IRQ HPD
notification is pending prior to sending the connect notification
to the user mode. If an IRQ HPD is pending, address the IRQ first
prior to sending the notification. This will ensure that the link
setup is stable first before the connection setup is started. This
will also help in adhering to the CTS timing requirements for
responding to IRQ HPD notifications. The current patchset limits
this logic only for SST connections.
Change-Id: I3f5e8e9290f915b98ec5e9e6cf9ff14aa9929d96
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
Ensure that session lock is held while handling link maintenance
to ensure that connect/disconnect handling and link maintenance are
serialized. This will avoid un-intended interleaving of session
setup/teardown with link maintenance which can result in link training
failures.
Change-Id: I96bb5b01a80433e26d675a019b8827de9fecb75b
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
It is possible that by the time DP display is initialized by the
user mode in response to a hotplug connect event, the DP display has
already been disconnected. In such cases, the link aborted flag is
set by the driver prior to sending the HPD disconnect event. Current
implementation will abort the connection callback when this aborted
flag is set. As a result, the DP stream clocks would not be enabled
and the DP interface will not be able to generate any vsyncs. This can
cause a lot of instability. Allow the connection callbacks from the
usermode to complete successfully to avoid these issues. This will also
help in subsequent disconnect and connect notifications to be handled
gracefully.
Change-Id: I9a0d44492a3590e1b6026bc545ee81f7414925cd
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
DP host initialization and de-initialization need to be done only
in response to cable connect/disconnect events. De-initializing host
(which includes turning off all regulators), for every HPD low events
is unnecessary and can cause unintended issues with DP CTS tests.
Change-Id: I95ac71853dc14a46e4175cf0bee4c3eae315fc1e
Signed-off-by: Aravind Venkateswaran <aravindh@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
The HDCP 2.3 specification added new timing
requirements for message read duration. Ensure
that these requirements are enforced in the DP
display layer.
Delay for a specification-allowed amount of time
before reading messages to ensure that sinks have
enough time to process messages and make their
replies available to read.
Change-Id: I40b210823a7cbaca6efc12abc4c8e8b98a10e071
Signed-off-by: Fuad Hossain <fhossain@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
