When MDP cannot support a layer, we send the whole list of layers to
GPU. If such layers are not updating they can be cached on FB while
allowing the rest to take the MDP route.
Example # 1: Volume bar on video in 90 rotated mode. Since volume bar
is not h/w accelerated, there is no prerotation and we invoke the GPU
for redraws each round. With this change we won't because of caching.
Example # 2: If prerotation is applied to layers on primary, the
secondary layers would show a transform and the whole list is redrawn
with GPU. Now we could make use of caching.
This change also imposes a 1 layer limit on secondary, but now allows
full and partial MDP comp, thus benefiting secondary-only content.
Change-Id: I7eb43ba817f1081f317fefa99fa0c507344060e8
The ab (average bandwidth) is just the average and there could be
underruns when ib (instantaneous bandwidth) is attempted to be used.
Add additional factor to use ib instead of ab. This means we need to
assume that all the layer contents could be requested together, rather
than one after another.
Also use the trimmed layer dimensions for calculations.
Change-Id: I771f964e1c80b5db564faa01cdfb75d75b54f97a
In video handling this patch makes sure:
1) Sufficient VG pipes are available, not just any pipes
2) If VG pipes are insufficient in multi-video scenarios
preference is given to secure videos
Change-Id: I170592463a1c28348108a1b12d60908cf3063d7d
When checking for pipes needed for each mixer, the pipes completely
unused are counted twice.
Add a per display check on top of per mixer check to make sure this
is taken care of.
For example: Each mixer needs 4 pipes, and total completely unused
pipes are 6. Each mixer will get 6 as available pipes, which is ok
at a mixer level, but at a display level 8 pipes are needed and 6
available. Need to account for that
Change-Id: I9811255aab96c7fe47331f8aa125fef2a4a2f704
- needsScaling does not check for scaling required in left & right
mixers when the crop values are sanitized. So it leads to issue
if DMA pipes are picked-up as high resolution layer does not need
scaling but left or right mixer configuration needs scaling after
the split.
- needsScalingWithSplit takes care of this with checking for the
sanitized cropped values.
Change-Id: I1f5b1952a59a84f5ccbaef123251b3c8cd46e4dd
Layer handle values will always be stored irrespective of the
composition type. Layer caching will choose a layer for FB
composition if its handle has not changed.
Change-Id: Ibc41a8cd6232c0c5c04efee01b38513a2431f4e1
Writeback for assertive display feature is needed only for
split displays and only if ro.qcom.ad is set to 2 by OEM.
This patch removes the unnecessary code and adds the relevant
conditional checks for split displays and property value
Change-Id: Ic36f28441d1a6e93132316b33fe1cf98a6e1efe4
A display could be split if its dimensions exceed 2048 or in case
of primary if we make it so via device tree property.
Add an API to find if a display is split. Use this API instead of
going through a set of conditions in various use cases.
Rename the HighRes, LowRes objects to a more appropriate Split and
NonSplit respectively
Change-Id: Id847c2bf36ea9250e1cdbc0259fdec64bd124537
Add support for checking whether the current frame's required
bandwidth can be handled by MDP.
The max limit is configurable via a property and the value is
derived from simulation runs and profiling
Conflicts:
libhwcomposer/hwc_mdpcomp.cpp
Change-Id: I46b46c346973fcb6487a8a656268171470ce69b9
Make all pipes in overlay available, if programming a frame (list)
fails. The failure could be due to commit failure, in which case the
overlay itself will make pipes available.
The failure can also be owing to lack of pipes, since we don't allow
switching pipes across mixers any more. This is particularly a problem
for split-panels. This fix addresses this latter case.
Change-Id: Ieac7ae7d03e5b9a93c7245fa160c8933593c2e09
Do not allow video if it requires pre-rotation and DMA pipe
is configured in LINE mode. Fall back to GPU in this usecase.
This is not applicable for wfd on 8x26 since Driver supports
both multiplexing of DMA pipe in LINE and BLOCK mode as same
mixer control block is used for rotation and writeback mode.
Change-Id: I5b99b214619134b1bd67f3c441b14f76051e3d3b
Avoid mdpcomp dump for frames having layers more than
MAX_NUM_APP_LAYERS, since our HAL doesn't support it.
Change-Id: I96365b4cbdeef2c00524eeb53877ee568ba0c172
Remove the unnecessary blank lock, mdp comp lock, secure lock.
Rename the ext lock to a more appropriate draw lock.
The mdp comp lock is at an incorrect place and causes unwanted
objects to show up in dumpsys, since configDone hasnt cleaned
them up yet.
dump(), blank(), draw() should all acquire a common lock.
draw() includes prepare() and set().
Change-Id: I595547dd5a393a8af6cd8c9297d50793b715e658
An earlier patch disabled plane alpha completely for video.
Instead, do this only when we are in video only mode.
Change-Id: I2edc8c5ff5c90ec205b63175e5f3a4e6ff7eae69
Command mode panels have an internal framebuffer with which they
refresh, hence the idle invalidator isn't needed for them.
Change-Id: I6acf0cc1f8c520cbcf706f109a42e47304c534d3
- cleans up external library
- add separate library for virtual display
- process virtual updates in its separate path
in hwc.
- Acquire blank mutex lock for one complete drawing
cycle
Change-Id: Ib984c578464a131ecdb27ee48960f58d68b7a5a7
Add assertive display support to hwc. This feature takes effect when
node /sys/class/graphics/fb*/ad is present. This signifies that that LM0
is being used with fb* for writeback.
When a video playback begins we write "1" to this node to indicate to
post processing that a writeback will happen. Likewise a "0" is written
to this node when playback stops.
The original contents are worked upon and the modified output is fed
via writeback to either rotator or mdp as appropriate.
The feature doesnt trigger when either:
1) Buffer size exceeds 2048
2) External display is connected
3) Multiple yuv streams are present
When this mode is active, MDP comp is applied only to yuv layer.
Change-Id: If5520f9dc849de3189c9f9ed4e9072c8f8f760e1
If the left split is not published by the driver then the default should
be even split. This was missing from the configuration in utils. This is
already taken care of in framebuffer configuration and mdp composition.
Change-Id: I3db44481fd791809146752ddff56420dcbe89fb4
For split displays, earlier we allowed pipes to switch mixers in
subsequent rounds. This change prevents that and makes sure there
is one composition round where a pipe being transferred to another
mixer of the same display is UNSET
Change-Id: I3c679cc4256363eeb70c5cf8bcaf5047b8a064c2
If framebuffer needs to be configured, configure it before all the
other layers in mdp composition. This helps in cases where we are
out of SMPs and framebuffer configuration fails owing to that.
Framebuffer being the fallback path, should always get highest
priority when reserving SMP blocks
Change-Id: Ie7a6903d1b9fb98b308689c81522571449bf2e8e
Added the ability to disable idleInvalidator when the
debug.mdpcomp.idletime property is set to "-1". The hwc test app
displays image on the screen with surface flinger turned off.
This causes timeout, so this change is to allow the testapp to
disable the timeout functionality by setting property.
Change-Id: I1adf28eea33ea73f3695644bec95056d16bd85f8
If debug property for MDP composition is enabled,
we call dump function from composition thread.
Make sure we are not holding the MDP composition
lock before calling dump function as this also tries
to acquire the same lock and causes a dead lock.
Change-Id: I05e5fd58be096f1ab683d796cc1beb4972c2168f
Reset the current frame data before returning from MDP
composition prepare to avoid accessing previous frame data
in MDP composition draw phase. Avoid the number of app layers
going beyond the layer count limit in MDP composition.
Change-Id: I2a3035ad5cb1731060b9359ccdbe9fe0c07e843c
CRs-fixed: 510683
Dumpsys runs in different thread not in composition thread.
Accessing mdp composition frame info in dumpsys with out
synchronizing could cause crash since frame info can get reset
in composition thread.
CRs-fixed: 503268
Change-Id: If3f7d8cc9fce8b14622e666fa52583ff0cfdedd2
Add support for unevenly split primary displays.
The driver provides info about the split via msm_fb_split sysfs node
For external we assume even split. If driver doesn't specify any split
for primary, we default to even split.
Change-Id: I4d541f41de2d7a5d2b62653fa33cab079a6d5d30
If an incoming layer has HFLIP tranform it needs two lines of
SMP buffer(or latency buffer). For panels with width greater than
1k this is costly. Fallback to GPU if RGB layer transform has HFLIP.
Change-Id: I9de63bf088d6afa0924ece7ac75e68f0bab7a6b5
When WFD is connected to 8x26
1) On Video transition, have 1 padding round if external connected.
Required to shift pipes across mixers.
2) Request DMA pipe always for FB on WFD.
Necessary for rotation + writeback
3) Disable non-worm-hole calcs on external (because of 2)
4) Disable action-safe calcs on external (because of 2)
Change-Id: I63b50b2477db443f9ea1d3fa610b59295c9359b7
In video-only cases, don't use yuv count for fb Z order, instead use
the mdp count which indicates how many video layers are actually
being sent to overlays.
Change-Id: I53ab4316fbb17f5d054066ef2dd85d89183d6a8c
This change adds adb property support to disable mixed mode in
MDP Comp. Will be helpful in debugging issues to narrow down the
root cause.
Change-Id: I49b737bb8755dc3a35db2e1ea8ccb2043135d2b0
Move the condition for checking maximum number of app layers
ourside the isFrameDoable function to avoid heap corruption while
updating hnd array in cacheAll function.
Change-Id: Id4458fcadc7775a2d78b1849de6782857c6ac17f
Fix MDP comp array index access in case of High Res pipe allocation.
This is already correct in case of Low Res allocation routine.
Change-Id: I7eaa0711ef8998ea03edba18b74db80459daf4c7
- Check for MAX_NUM_APP_LAYERS before updating yuv indices array.
- Fall back to GPU composition when number of app layers exceeds
MAX_NUM_APP_LAYERS to avoid heap corruption.
Change-Id: Ieb91b705a0a5f50ce2f8829d1f1ee048d44b7d2e
In scaling check, consider out of bounds layers, apply trim
and then decide if scaling is necessary or not.
Change-Id: Iad7f662f42096b3f7a4a1c4cfe71048d275f4b5f
These include constrained targets, with external connected which could
lead to a scarcity of pipes. Furthermore owing to a scarcity of writeback
interfaces, even if pipes are available, we may not be able to still
use rotator.
Change-Id: Ie2fb4c1726f32addcace89c268a126abe0cddbe0
