This change initializes a variable as null to account for cases
when it is not set. This change also ensures proper pointer
checks before attempting to access function pointer.
Change-Id: I2f06a0877293668e80bee9d9b82d412476dc5184
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
Remove idle time from the Qsync calculation as this was added by
mistake to allow writes to complete just before the next TE but
this causes tearing. Also, to ensure tearing doesn't recur,
round down the threshold lines to the nearest multiple of 4, and
remove a couple of extra lines to compensate for potential
latencies.
Change-Id: I8a53a989e26cbd7f0e2b94caa8df8f5bee3ad26c
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
SDE RSCC solver state and autorefresh enable concurrency
is not supported. This change moves the rscc solver state
to disable to avoid concurrency. It also resets the autorefresh
software structure state when encoder is disabled. This allows
autorefresh reconfiguration with next encoder enable.
Change-Id: Idb8c722c823d9f46d3cd03e1b046da69c8d88fc4
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
In dual DSI cases sde_encoder_helper_phys_disable needs to be
called only for master encoder.
Change-Id: If2d327a4800914fe365c6877bf7854275a4434ae
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Currently, SDE_DBG_DUMP takes any number of hw block names along with
few defined strings as arguments. This set of arguments is used to
determine which HW block registers needs to be dumped. Move to a
blk bitmask to avoid passing a large set of arguments. The bitmask is
split based on the clks required to access the HW block for ease of use.
The lower 0-23 bits are used for HW blocks which can be accessed by just
enabling the MDP clks. DP is kept separate as it needs DP specific
clks to be enabled. Add a debugfs node through which the mask can be
modified, which can be useful while using the debugfs dump option to
force a panic.
As part of the change, remove in-log/in-mem enable mask debugfs node
for every debugbus and use a single node to control the logging
mechanism for all the HW blocks debugbus.
Change-Id: Ibb6354b3e3265c9911104bb0f964616eb8a898c9
Signed-off-by: Veera Sundaram Sankaran <veeras@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>
Avoid irq enable/disable during modeset and trigger
frame as posted start frame. This saves mode_set time
and also avoids unbalanced vblank_irq in this usecase.
Change-Id: I06958da5e52bc2aca0ddc60d2783615f80a839a4
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
Signed-off-by: Tatenda Chipeperekwa <tatendac@codeaurora.org>
From MDSS 8.x INTF line/frame counters can be reset through
a register. Reset these counters during timing engine enable /
tear-check enable to keep track of meaningful counters, which
would be useful during debugging. Additionally, reset the
counters during cont-splash modeset to track the number of
auto-refresh frames while disabling it during the first frame.
Change-Id: I66b45f5b29793df1fb4635972b1c614ad8c3b5b3
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
Watchdog timer is moving from TOP to INTF. This change adds
support for movement and ensures backwards compatibility.
Vsync select only needs to specify whether or not to use
Timer 0 associated with the interface. It does not need to
select between Timer 0-4.
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
Change-Id: I9d89a8cb1ea607e9fc0bdbffa0a6a9acceff7f13
Fix invalid condition in TE configuration missed during
porting of previous change.
Change-Id: Ie6b1926e6fdd1e0ec3b520593b1e5cbfbf9e4202
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
Use of drm_display_mode vrefresh is being deprecated in
upstream DRM framework. Downstream driver need to use
drm_mode_vrefresh API from now on.
This change removes dependency on drm_display_mode vrefresh
and replaces it with drm_mode_vrefresh API in SDE, DSI and
DP driver. In addition, it also modifies drm_display_mode clock
to align with upstream approach where an uncompressed mode clock
is required to match drm_mode_vrefresh API.
Change-Id: Ie972a2e140adfd81c4e68df8e7bc69feaaca22e1
Signed-off-by: Nilaan Gunabalachandran <ngunabal@codeaurora.org>
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
The SDE RM determines the number of displays enabled in cont-splash by
checking the INTF_ACTIVE bits in the ctl. The trusted-vm relies on the
same cont-splash path during transitions and the stale ctl INTF_ACTIVE
on dual display usecases leads to wrong status check. Currently, only
video/wb encoders use the phys_encoder disable helper during encoder
disable to clear out few ctl/intf registers. Extend it for cmd encoder
as well to remove the stale entries.
Change-Id: I7f21ef46178ce3943ee85abdf29f7cdef93d7c52
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
On frame timeouts, the driver should only be responsible for
reporting the error event to user space clients. Handling of
the error events should be left to the user space clients.
This change removes explicit mechanism in SDE driver to track
error events for a pre-defined threshold levels and RESET hints
sent to the user space on handling those error events during
frame timeouts.
Change-Id: I7bdf2495fae6430384b4031a7edf043b6efe88c5
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
The irq register and unregister calls in ppdone timeout
path can result in race condition which can lead to
scenarios where irqs will be enabled with no
callback registered. Add vblank mutex lock to avoid
such scenarios.
Change-Id: I02b07b773d6b9a28c3af431eb3d8314292769f12
Signed-off-by: Krishna Manikandan <mkrishn@codeaurora.org>
In current implementation qsync min fps is single value.
It is same for all the list of supported dfps list.
Added support for new dt entry dsi-supported-qsync-min-fps-list
corresponding to the fps supported in the dfps list
dsi-supported-dfps-list.
Change-Id: Ifd5309c2f51865a3c0d9fadb65cbcd291b6ef42b
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Add changes to extend the current evtlogs for ease
of debug purposes.
Change-Id: Ibe136afbf25b8b8decfc06ef3868e3f628df4dcb
Signed-off-by: Jayaprakash <jmadiset@codeaurora.org>
Currently driver reads CTL status register before checking
pending kickoff counter. This can lead to a register access
violation when there is a race condition between the ESD and
commit thread.
This change checks pending kickoff counter before reading
CTL status register.
Change-Id: I5828b580c16d075df19eb349ee88d8b7da47941e
Signed-off-by: Amine Najahi <anajahi@codeaurora.org>
Remove the output parameter "extra_frame_trigger_time" from this
function as callers are not using it, so it is not needed.
Change-Id: I98fcfe974c64da9ce62456a693330b898707e6b4
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
Adjust the QSYNC line count calculation to compensate for the
idle time, when no transfers are actively taking place.
Change-Id: If91eab25321eea6e6880f07605c5a9c1b7b7ee05
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
Vblank refcount can reach out of sync with below case
1. event_thread triggers the vblank_enable
2. commit_thread triggers the modeset
2.a modeset resets the vblank refcount with mode_set
3. event_thread triggers the vblank_disable
Event 2.a resets the vblank refcount and vblank disable
request after 2.a is going to fail. This can be fixed
by avoiding concurrency between mode_set call and vblank
request.
Change-Id: Ibb810ec90e81d63feee443f1c37dd736d5cfac0d
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
VM switches during TUI usecase are expected to be seamless i.e without
display reset. In SDE language, this translates to respective display
drivers not tearing down the HW pipeline while releasing the HW.
In Primary VM, this taken care by keeping the DRM pipeline alive when
TUI is active.
In Trusted VM, since the client creates and destroys the display per
session, checks are needed to bypass the physical encoder disable(s).
Change-Id: Iac42f02806962405c9364b1ffed85778229977e9
Signed-off-by: Jeykumar Sankaran <jsanka@codeaurora.org>
This change addresses out of range and null checks in
sde driver.
Change-Id: I4ee82760ce3ee7053c336e49ec9eaae8b4c31b1e
Signed-off-by: Narendra Muppalla <NarendraM@codeaurora.org>
With posted start enabled SW no longer waits for ppdone events
unless more than one frame is in queue. HW recovery logic relied
on these waits to know when we recover from a timeout. This
change moves the HW recovery SUCCESS event signalling outside of
the wait to ensure this event is sent to user-space regardless
of the status of posted start.
Change-Id: I8896e8126284b415513499723ccf0155ee8bc6a7
Signed-off-by: Steve Cohen <cohens@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>
From Lahaina onwards, widebus is enabled for compressed DSI stream.
This change adjusts interface timing parameters to account for widebus.
Change-Id: Ie6b739ed2cdb515064e3a94404b3e0fe07755d7e
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
Update autorefresh disable sequence-2 by avoiding
tearcheck enable configuration. Updated sequence
will trigger the frame by resetting the write
line count with tearcheck start position.
Change-Id: I984251c0cb23475f20cd5ea62122a167324d6670
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
Autorefresh might be still in enabled state after
write_ptr reached to display height. This patch
checks the autorefresh status bit and keeps polling
the bit till autorefresh sequence is in enabled
state. It times out after 1 second if autorefresh
is still in enabled state.
Change-Id: I5d4f4cb35e5cc8c680c1878f52cee385f709d764
Signed-off-by: Dhaval Patel <pdhaval@codeaurora.org>
In Dual display concurrencies there are certain cases where
irq_idx and vblank_refcount state mismatch can occur.
To avoid it, during setup of irq_hw_idx, reset the vblank_refcount
and unregister read ptr irq if not yet done by then along with
maintaining global mutex lock for vblank_refcount.
Also, if register IRQ fails, correct vblank_refcount so
that IRQ registration can be tried again.
Change-Id: I06bcbf71c6a43bd161ff61093d9f6063a292d6bc
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Signed-off-by: Samantha Tran <samtran@codeaurora.org>
wr_ptr_wait_success is true only if wr_ptr interrupt arrives, so
initialize it after display kickoff.
Change-Id: I5790d9dac25352898ece160f6b258b50ca2edefa
Signed-off-by: Ray Zhang <rayz@codeaurora.org>
There can be few cases of ESD where CTL_START is cleared but
wr_ptr interrupt does not come. Signaling retire fence in these
cases to avoid freeze and dangling pending_retire_fence_cnt.
Change-Id: I167f69dce5cbe43b4771e5056d8a73bd7587e76e
Signed-off-by: Raviteja Tamatam <travitej@codeaurora.org>
Reset MDP ctl path and DSI ctl on autorefresh
disable failure. This will enable the hardware
to recover from the hang.
Change-Id: Ia9acc8573c22e0713179ef4f6ef604caacabfadb
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
Using individual flush functions for each active hw blk
is not scable-able for future use. Clean up the ops to merge
all flush functions into one and manage HW block id
with same API.
Change-Id: I62afbc51fa7d345b3a1f5721e5e09661a4215f7a
Signed-off-by: Nilaan Gunabalachandran <ngunabal@codeaurora.org>
This change enables the interface hw block to accept
64-bit compressed pixels. This configuration is enabled based
on hw capability. For current hw, this is required any time
the compressed pixels flow through the interface block,
whereas in the previous version of DPU hw this configuration
is only required for topology using 4 way dsc merge.
Change-Id: I7bf79d035dba5084c5057022a7fa1117479e8d52
Signed-off-by: Abhijit Kulkarni <kabhijit@codeaurora.org>
Set the INTF TE capability bit only on interfaces which have a non-zero
value in the device tree node qcom,sde-intf-tear-irq-off instead of
enabling it for all interfaces based only on the HW version. The HW
doesn't support TE programming for non-TE enabled interfaces, so this
patch only populates the TE ops for those which support it.
Change-Id: I00518e846dc44e1e0808a049625dc14099656e11
Signed-off-by: Steve Cohen <cohens@codeaurora.org>
Due to interrupt delays, sometimes the pp-done interrupt
for an in-between frame is lost, as with posted-start
frames are queued to the hardware before the completion
of the previous frame. Handle the lost pp-done interrupt
in the case where frame-n pp-done interrupt is missed
and frame-n+1 pp-done interrupt is processed before
frame-n+1 wr-ptr interrupt.
Change-Id: I36ec7ac494b2720fc005dab75047d2f4a5a2a699
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
Update the wr_ptr_success state to avoid waiting for
two frames pp_done_tx_irq. It ensures frame done callback
is triggered, if IRQ's are disabled for long time by
some external entity.
Change-Id: I9ee72bd65dd7251bf8db0b6843fad02b675935bc
Signed-off-by: Jayaprakash <jmadiset@codeaurora.org>
With posted-start, ctl scheduler status and pending-kickoff
count are checked to address the missing pp-done interrupt
cases, which can occur if the IRQs are disabled for a long
time by some entity. Currently this check is done after
the wr-ptr-irq. At this point the scheduler status may
or may not be idle, based on the frame-transfer. Move this
handling after the timeout, so the scheduler is guaranteed
to be in idle state for working use-case.
Change-Id: I3fa9ecce8139ff667c1882e286571169c543c797
Signed-off-by: Veera Sundaram Sankaran <veeras@codeaurora.org>
By adding spin locks around the handle frame done, a race
condition, between the wait for idle of one frame, and the pp
done of the next frame, is mitigated.
Change-Id: I78650fc9688b22db6d314b9bed023e0ddad20dbf
Signed-off-by: Nilaan Gunabalachandran <ngunabal@codeaurora.org>
