This change reinitializes the sde_crtc->mixers when CTL
datapath switch occurs during mode set and RM allocation
of CTL hw block is changed. This initialization is required
for CTL_LAYER programming to trigger on the new CTL allocated
from RM.
Issue case:
1. Primary Display is using CTL_0 and it is reserved.
2. Secondary Display is using CTL_1. On suspend, RM adds
CTL_1 into the free list.
3. External Display is powered on, RM allocates CTL_1 hw blk.
4. Secondary Display is powered on, RM allocated CTL_2 hw blk.
5. External Display is suspended/unplugged, RM adds CTL_1 into
the free list.
6. When any mode_set(say fps switch) occurs on secondary, RM
allocates new resources and CTL_1 is allocated.
sde_crtc->num_mixers is non zero, so all the layer programming
happens on CTL_2, but CTL_1_FLUSH bits are programmed causing
hw timeout issue.
Change-Id: I5f1f52b7673740c48b249ab4d36e80b7a1d3db96
Signed-off-by: Jayaprakash Madisetty <quic_jmadiset@quicinc.com>
uidle_db_updates are generated when CTL_x_UIDLE_ACTIVE is set to 1.
It needs to enabled in both uidle enable and disable cases.
CTL_x_UIDLE_ACTIVE is set to 0 only in cases where uidle configuration
is not updated.
Change-Id: If7655e4eae351bac248f0906c473cdfaf93f2b8a
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
This change deprecates idle notify work for video mode
since idle timer will be maintained by userspace.
As part of idle notify work, syscache state is changed from
CACHE_STATE_NORMAL to CACHE_STATE_PRE_CACHE along with
notifying to the userspace. This change removes
CACHE_STATE_PRE_CACHE in the state machine and state is
updated from CACHE_STATE_NORMAL to CACHE_STATE_FRAME_WRITE
whenever the cache property is set.
Change-Id: If3b2c34be954cb625aca76da81fd854c077a8250
Signed-off-by: Yashwanth <yvulapu@codeaurora.org>
This change sends a notification to user mode after msm_drm
driver releases the mmio and irq resources on trusted vm transition
request. This is required as user mode has no other way to know
when the resources where actually released. User mode driver earlier
relied on retire fence signaling but retire fences are send before
releasing the hw.
Change-Id: Ia218cfcbf398b2de1ad9578fb9baedf348b067df
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
Current timeout values in sde driver for vblank, kickoff and
frame complete timeouts are fixed and can be much lower than
a vsync incase of low refresh rate display. Compute timeouts
based on refresh rate for low refresh rate displays.
Change-Id: I9dda41feb15446de7451824e185321de421ad575
Signed-off-by: Prabhanjan Kandula <pkandula@codeaurora.org>
When all instances of driver fd's are closed by user-space client, drm driver
will be closed. When last close of driver is called, custom reset properties
api will be called where driver should cache the properties that it wants to
clear. Current behavior of color processing driver is to clear hardware
configuration instead of caching which can cause crashes if clocks are off.
Change updates the driver to cache the pending disable and update hardware
during display commit.
Change-Id: I9703f860ed0ae3c859d6fc3995b58be13203f259
Signed-off-by: Gopikrishnaiah Anandan <agopik@codeaurora.org>
Delayed work items may touch HW registers. If these work items
run while HW is not owned by this VM it will lead to invalid
access. This happens in video mode as HAL does not disable idle
power-collapse in this mode. It can also happen with ESD status
if lastclose or TUI transition failure occurs.
Although there is a contract with user mode to turn off certain
features, kernel cannot rely on it to always do the right thing.
Prevent potential crashes from certain corner cases by
cancelling all delayed work items when the HW ownership is
transferred.
Change-Id: I08da17f2ce72bf2fddf71924c3e8edd2e2715be8
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
Add changes to modify the phys_enc->in_clone_mode variable
post wb_reset_state since this is a shared variable used
during atomic_check and atomic_commit. In current issue case,
wb_atomic_check has set in_clone_mode to true in commit N,
and in commit N-1 CWB is being disabled and re-sets the
in_clone_mode variable to false causing pp_done timeouts in
primary in commit N.
Change-Id: I8159bbdb5622a351d76bdc4dba75d48df20f4365
Signed-off-by: Jayaprakash Madisetty <jmadiset@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
In PM resume with CWB concurrency usecase, crtc pointer in
conn->state is NULL since drm_mode_config_reset operation is
performed on pm_resume. This change relies on conn_mask in
new_crtc_state for primary connector retrieval and also adds
get_num_lm_from_mode callback to DSI for LM count retrieval
from dsi panel topology. Existing get_mode_info api cannot
retrieve the topology info because mode->priv_info is NULL.
This occurs as WB encoder is added in the drm encoder_list
before primary encoder, introduced as part of commit d28ebf05f4
("disp: msm: sde: populate WB display encoder list before dsi").
Change-Id: I55358fd88ab778bd81475cf3628be13335de1cb5
Signed-off-by: Jayaprakash Madisetty <jmadiset@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
During idle power collapse, dim layer dirty flags are
stored in sde_crtc_state which might be invalid if state
swap occurs. This change adds revalidate mask in sde_crtc
structure to revalidate after coming out of idle power
collapse.
Change-Id: Ie2f34a794896a3f8e729ef7d1f3ae35340123257
Signed-off-by: Yashwanth <yvulapu@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
This change avoids upstream drm issuing drm_atomic_commit in
drm_fb_release which is leading to artifacts on screen or
atomic_check failures due to atomically unstaging each fb
from plane_state and committing remaining planes on hardware.
a) This patch moves the state operations for setting crtc to
connector state to a helper api.
b) This patch clears any dim_layers present in the crtc_state
as part of null commit.
c) This patch removes any framebuffers attached to a drm_file
in msm_preclose whose refcount is not managed by composer kill
inadvertently and issues null flush to hardware in such cases.
d) This patch handles msm_preclose as part of msm_release
operation since legacy feature is not supported
for msm_driver.
Change-Id: Ib2068d74d4b23b73b7c84544858c9f6bb6adfa67
Signed-off-by: Jayaprakash Madisetty <jmadiset@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
In dual display usecases, during pm suspend/resume,
commit is scheduled only on primary crtc thread. If idle
timeout value is very short such as in LP2 mode, it might
result in race condition due to idle pc off work getting
scheduled on its crtc thread. This change adds kickoff in
progress flag to handle such cases as crtc frame pending
count is only updated after rc kickoff.
Change-Id: Iebb331d914b23cc5eeadfeb2a488891e88b3202a
Signed-off-by: Yashwanth <yvulapu@codeaurora.org>
During one of the DP timeout usecases, flush doesn't take
effect due to vid vblank wait failure. As a result, smmu
faults are observed because of fetching the previously
staged planes. This change adds ctl reset in the same
DP atomic commit context to recover and avoid
smmu faults.
Change-Id: I2f9aceca56e27f140607317f7596d6fe0d908af8
Signed-off-by: Yashwanth <yvulapu@codeaurora.org>
When LTM is switching on/off, merge_mode bit value gets toggled between
0x1 (dual pipe merge configuration) and 0x0 (single pipe configuration).
It is illegal to reconfigure LTM to/from dual-pipe merge mode before
both LTM instances have completed their current workloads. This change
adds support to disable merge_mode one frame after histogram is disabled
to make sure both hardware instances are completely idle and avoid
corrupted histogram data collection.
Change-Id: I9a6b5cbfb69e8af7936749e57fe7c8f7c2703b95
Signed-off-by: Ping Li <pingli@codeaurora.org>
Add support to send a data packet of info, written to
predefined buffers, providing information about each submitted frame.
Add required UAPI definitions for frame data buffers and event
notification.
Add support to read ubwc statistics from hw, based on defined rois.
Change-Id: I51f279de98ae4e2a02b0df6943d334764011d5db
Signed-off-by: Nilaan Gunabalachandran <ngunabal@codeaurora.org>
Move the hist irq handling out of callback function, i.e., the hw
interrupt irq_lock context, to avoid dead lock between crtc spin_lock
and irq_lock. This change also extends crtc spin_lock coverage in
_sde_cp_crtc_enable_hist_irq to prevent null pointer dereference on
event node, which can be deleted during crtc event de-registration.
Change-Id: Iadaed54ab93c4c4abe065a8762d2addccb0c65c6
Signed-off-by: Ping Li <pingli@codeaurora.org>
Due to lock sequence inconsistency between sde_crtc->spin_lock and
sde_kms->hw_intr->irq_lock can cause deadlock, to avoid this possible
deadlock this change uses different spin lock for frame events.
Change-Id: I51b1184dfa1069c87653099b95b992b277721daf
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
Calculate line_time once during modeset and allow
each plane to use it instead of calculating for each frame.
It also simplifies the line_time calculation for
command mode display.
Change-Id: I94ce29eec94bfdbee9016fbf93378661ebf79c03
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
HFC feature of demura needs a skip blend plane to be set. If skip blend
plane is not set and HFC feature is requested to be enabled, driver
should skip turning on HFC demura feature. Change adds checks to ensure
that HFC is always enabled with skip blend plane staged.
Change-Id: I923359c7cb143867660b4c1e667f56ed42fa51c9
Encoder is disabling planes and unstaging layers from layer mixer. Some
of the crtc features are dependent on the plane being staged. Change
adds api that encoder can call on crtc to stop the features that are
dependent on source pipes.
Change-Id: I4d875155ceb8f66acfe6ce0096141ca7253bb140
Signed-off-by: Gopikrishnaiah Anandan <agopik@codeaurora.org>
Declare the continuous splash pipeline setup to userspace by filling
in the DRM states for all plane, crtc, encoder, and connector objects
in use. This information will be treated as an 'informative' state,
and will be cleared at the start of the first commit to maintain
the DRM methodology of DRM clients being the only controller of
the pipeline. This ensures any configuration provided by userspace
is accepted and applied, even if it may already align with the setup
done by continuous splash.
This DRM state configuration is done via manual modification of the
DRM object states. Modification via the exposed DRM UAPI functions
is not possible due to no drm_atomic_state object linking the DRM
pipeline objects together.
Change-Id: I67650e05aafbb4e799cf60939f0595bc3786fc6e
Signed-off-by: Christopher Braga <cbraga@codeaurora.org>
As the number of cp features has continued to grow, the file
sde_color_processing has become increasingly harder to parse.
Update function and parameter names to better identify internal
functions and increase readability.
Change-Id: Ib82d6c9a45b36b932ab3a2d573b7cbe13a6c10bc
Signed-off-by: Christopher Braga <cbraga@codeaurora.org>
Recent investigation shows that color processing properties are
incorrectly being cached in validate. This can result in unwanted
color processing properties being applied if a previous commit
failed or was validate only.
Add color processing properties to sde crtc state instead of marking
them dirty in color processing driver. When atomic commit is called
properties from state will be marked as dirty and applied.
Change-Id: If231a1f028e4cbd0b50eb0a947f4d58f94390a0c
Signed-off-by: Gopikrishnaiah Anandan <agopik@codeaurora.org>
Signed-off-by: Christopher Braga <cbraga@codeaurora.org>
This reverts commit 6886d03e4a.
It tries to memset the pstate from UI thread (crtc atomic check)
and crtc commit thread.
Change-Id: Ic9d3e5555d7085832df76025e53488d2b3365739
Signed-off-by: Nilaan Gunabalachandran <ngunabal@codeaurora.org>
DPU has added support for noise injection into the layer stack. Change
adds support for noise layer programming and exposes the hardware block
to the user space modules.
Change-Id: Id176eea54fcdcd5d399457b14133a1ccde07299f
Add new capture/tap point as CRTC property for
D-CWB feature. Update the hardware blocks and
corresponding APIs to configure D-CWB data path.
Add new hardware pingpong blocks that
are dedicated for CWB.
Change-Id: I22576df1768b50f9f47d8527f62913b01ff4d9a7
Signed-off-by: Chandan Uddaraju <chandanu@codeaurora.org>
To avoid kcalloc for the multirect plane states and plane states will now
be stored sde crtc. These states are populated momentarily and accessed in
a single context for a handful of functions then are not used. This will
clean up parameters passed between functions in the commit path as well.
Change-Id: I6a8116a43c140b3f1c0464734032b8db13c1cfb0
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
Move away from the private and private_flags fields from drm_mode,
as it is being deprecated in latest kernel version. Instead, Add
msm_display_mode as a wrapper to be used in downstream to store these
parameters. Also, store msm_mode in connector_state to be accessed
in commit path.
Change-Id: Ia5bdebe75f00aa15fb7db4dc3a0d50c30894a95c
Signed-off-by: Orion Brody <obrody@codeaurora.org>
This change resets the plane and crtc dirty properties on
transition to other VM. This ensures when the VM acquires the
hw again all these properties on the new state are reapplied.
Change-Id: Id95676453eb9ae937b1b22f0e244b47449101cb0
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
There is a race between disable commit swap-state on display
thread and vblank work on event thread which may skip applying
the vblank vote if the encoder_mask has already been cleared.
To avoid the race, use a cached encoder_mask that is gauranteed
to be valid between the vblank_on and vblank_off calls.
Also, vblank queue work is initializing a variable that's never
used and msm_disable_outputs has 2 instances of pointers to old
CRTC state. Remove this dead code.
Change-Id: I5e4a482b8f067e272a2aef5afa08cc0e1ab89434
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
It is not applicable for all DFPS cases to update UIDLE state
according to current frame rate. If DFPS changes frame rate
through vertical front porch values, the SDE clocks and transfer
time will not get changed accordingly, and it always get fixed
at max frame rate configuration of DFPS.
Add this change to check max FPS of DFPS instead of current
frame rate for UIDLE update, if DFPS is enabled with VFP.
Change-Id: I7634bce6a9eb1af212ba19a267735be08b20ae1f
Signed-off-by: Lei Chen <chenlei@codeaurora.org>
Added sysfs entry on crtc to notify a retire frame event
which indicates start of new frame. It is added at same
time of signaling retire fence.
Change-Id: Ie60aae96bd6e6bfb3cfe9db482cb59053f22383f
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
All sde crtc's are virtual when they are created. Resources for the crtc
is allocated when crtc is enabled. All crtc's will not have same
capabilities because some of the dspp blocks have additional hardware
blocks. Change exposes additional dspp capabilities dynamically when
crtc is allocated the dspp hardware block.
Change-Id: I93e76a1335574e4ca30d9419ef6cc6e8149e2c3c
Signed-off-by: Gopikrishnaiah Anandan <agopik@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>
Add a CRTC property to request the VM to acquire/release
HW resources.
Display driver in trusted VM boots up without HW ownership. Set
the default value of the property as RELEASED to handle resource
assignments.
Change-Id: Iea651a2fea902d95d4b954052af4ef016af15a91
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
In some cases like suspend and cwb concurrencies,
the number of mixers in the sde_crtc structure
can become zero. Add support to get the number
of mixers from topology in those cases to
avoid incorrect resource allocation request.
Change-Id: Id9b82e805ff50a107ad06514b4e41c0917abdf33
Signed-off-by: Krishna Manikandan <mkrishn@codeaurora.org>
Increase the maximum number of mixers per crtc to 4 to
support 4LM use case. This change also increases the number
of data path to 4 to support 4LM in continuous splash handoff.
Change-Id: I4655017dcb405fad69513bebb8fd7f848fc5873d
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
When a 4LM topology is used each plane attached to a CRTC
is tagged with a L/R layout value and an offset value
depending on where destination X coordinate lands on the display.
The layout information is used to determine SSPP to LM
pair mapping and local coordinate space.
This change also handles source-split and Z-order
validation checks for planes staged on different mixer
pairs.
Change-Id: I1b20223388e65fc36a8b379ad9df23a277fcd1a5
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
Setup blendstages is done per LM but FETCH_PIPE_ACTIVE is per CTL.
Overloading mixer blendstage setup with fetch pipe logic can lead
to HW programming errors. Refactor the logic for setting
FETCH_PIPE_ACTIVE by adding a new op that allows caller to provide
a bitmask of all pipes required to be active on this CTL. This new
logic includes support for:
- 4LM use-cases, staging pipes for all LMs within a CRTC
- Demura fetch-pipe without need for tracking via active_cfg (removed)
Also, lower the cyclomatic complexity in setup_blendstages by moving
the logic for obtaining the mixer config settings in to a helper
function.
Change-Id: I2907b359ffad5734be5b06f44919b5ddb1ef3f7c
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
