* commit 'c9cd2387b6938a6fbefc731d2177902266f2a130':
Fix a race that could cause GL commands to be executed from the wrong thread.
RefBase subclasses can now decide how they want to be destroyed.
Fix a race in SurfaceFlinger that could cause layers to be leaked forever.
Fix a race-condtion in SurfaceFlinger that could lead to a crash.
The transaction flags were atomically read-and-cleared to determine if
a transaction was needed, in the later case, mStateLock was taken to
keep the current state still during the transaction. This left a small
window open, where a layer could be removed after the transaction flags
were checked but before the transaction was started holding the lock.
In that situation eTraversalNeeded would be set but only seen during the
next transaction cycle; however, because we're handling this transaction
(because of another flag) it will be commited, "loosing" the information
about the layer being removed -- so when the next transaction cycle due
to eTraversalNeeded starts, it won't notice that layers have been removed
and won't populated the ditchedLayers array.
Bug: 4483049
Change-Id: Ibb5989312f871339928ee1aa3f9567770d72969b
Client::mLayers could be accessed from different threads.
On one side from Client::attachLayer() which is currently
called from a binder thread; on the other side from
Client::detachLayer() which is always called from the main
thread.
This could lead to a corruption of Client::mLayers.
We fix this issue by adding an internal lock to Client.
Bug: 4483046
Change-Id: I5262bf1124d9a65ec6f8ffd8e367356fc33a7536
Without that lock, there is a chance of race condition
where while composing a specific index, requestBuf with
the same index can be executed and touch the
same data that is being used in initEglImage.
(e.g. dirty flag in texture)
SW renderer of Stagefright reaches loadTexture() to draw image.
The first time loadTexture() is called, it just initializes OGL
texture, then returns. Thus, the first time call doesn't draw.
This patch fixes to move on to draw stage after the initialization.
Change-Id: I3ec1ad68fb8d376a4ad7aefded1c18a002d175c4
6d0f6cb Revert "Fix [3513017] in lockscreen but showing empty launcher (live wallpaper) only"
6154412 Revert "partially fix [3306150] HTML5 video with H/W acceleration blackout (DO NOT MERGE)"
37c2a37 fix [3408713] Dialog window invisible sometimes
It looks like there is a surface leak, it's unclear where it is.
Without those reverts, this would cause a leak of the associated buffers
which is far more problematic. this change might hide the surface leak.
Bug: 4078032
Change-Id: Iedcda3ffcdd2f69d41047b5c3134c1e867ff90d7
37c2a37 fix [3408713] Dialog window invisible sometimes
d35c666 fix [3385504] Surface flinger hang when adding dim surface
1723b04 fix [3389263] OMX.Nvidia.h264.decode fails to shutdown
1b0114f fix a surface leak in SurfaceFlinger
Bug: 3513017
Change-Id: Ia13ed8c9cdcb1f484e177cdcaff687e7c88a10c3
When some Surfaces are overlapping and one of them changed to
transparent or opaque, the back Surfaces should appear or
disappear. This patch calls SurfaceFlinger::computeVisibleRegions()
to re-calculate region of each Surface to implement the behavior.
Change-Id: Iffb1caf1b4ce28dff252e114fe5b9b07d9c84a6f
Generally we never want to lock a buffer for write access if it is at
the "head" on the surfaceflinger side. The only exception (1) is when
the buffer is not currently in use AND there is at least one queued
buffer -- in which case, SurfaceFlinger will never use said buffer
anymore, because on the next composition around, it will be able to
retire the first queued buffer.
The logic above relies on SurfaceFlinger always retiring
and locking a buffer before composition -- unfortunately this
didn't happen during a screenshot.
This could leave us in a situation where a buffer is locked by the
application for write, and used by SurfaceFlinger for texturing,
causing a hang.
Here, we fix this issue by never assuming the exception (1), it was
intended as an optimization allowing ANativeWindow::lockBuffer() to
return sooner and was justified when most of SF composition was
done in software. The actual buffer locking is now ensured by
gralloc. We could have handled screenshots in a similar way to
a regular composition, but it could have caused glitches on screen,
essentially, taking a screenshot could cause to skip a frame.
now that we removed the notion of a "inUse" buffer in surfaceflinger
a lot of code can be simplified / removed.
noteworthy, the whole concept of "unlockClient" wrt. "compositionComplete"
is also gone.
We used to guarantee that a layer in SurfaceFlinger would never be
destroyed before all references (to its ISurface) on the client
side would be released. At some point, this guarantee got
relaxed to allow to free gralloc resources sooner. This last
change was incorrect, because:
- in implementations with reference-counting the gralloc resources
wouldn't be released anyways, until all the mapping were gone
- in implementations without ref counting, the client side
would most likely crash or do something bad
- it also caused the SharedBufferStack slot to be reallocated
to another surface, which could be problematic if the client
continued to use the surface after the window manager destroyed it.
So, we essentially reinstate the guarantee that layers won't be
destroyed until after all references to their ISurface are
released.
NOTE: This doesn't entirely fix 3306150 because there is another
problem there where the Browser continues to use a surface after it
has been destroyed.
also improve SurfaceFlinger 'dumpsys' log
list the purgatory, which shows windows that have been closed,
but for which the client still has references.
we were not clearing the screen entirely, which caused garbage when
the screen wasn't entirely covered by windows.
Change-Id: Ia7aa13c36a8a314e0e8427d419b16b9aa2165ddf
we make sure to call compositionComplete after everytime we do
composition with the GPU (even for the screenshot case), which is
where the buffer locks are released.
Change-Id: I450430d1e4d1ee9ce1023970642378c42cdcfa4c
* commit 'a2977c383d363e1e88a5b36230b1fa4c312807d2':
[3171580] SurfaceFlinger Bypass mode. (DO NOT MERGE)
[3171580] Add transform field to native buffers. (DO NOT MERGE)
* commit '05813b0eb92cb1bc79607ee402f14ca1e4b43f6d':
[3253328, 3171580] Treat GONE and INVISIBLE views the same when calculating transparent regions
[3171580] Fix two typos related to fixed-size buffers
This is a poor's man precursor to the h/w composer HAL.
Basically we detect when a window is full screen and in
that case we bypass surfaceflinger's composition step, which
yields to much improved performance.
Change-Id: Ie03796ae81a1c951949b771c9323044b980cb347
mFixedSize was never set, this bug was introduced during some "cleanup", in
practice this could cause some issues when a fixed-size buffer was used and
the window was resized.
Layer::drawForSreenShot() had a typo that had no effect.
mFixedSize was used to determine if filtering was needed, which was a bit too
conservative and created a dependency between filtering and "fixed size" states
which should exist.
Now we enable filtering based on the size of the buffer vs. the size of the layer.
Change-Id: I32044e91b0c944c1b137efdceb3f01dfaa78119d
some of these failures are not fatal and even expected in some cases
so they should not emit a dump in the log in those cases.
Change-Id: Idcfa252e3bfa9d74e27fe4ad8f8623aa01aa9c5e
Two bugs were counter acting each other.
- rotation matrices are on the left-hand side of multiplies
- the transform of the overlay is applied before that of the layer
Change-Id: Ia79bd368e9b719235c89ecf244ea263f01ce906a
Merge commit '457bed2bc6561dd67429dde238453fee8602fa9b' into gingerbread-plus-aosp
* commit '457bed2bc6561dd67429dde238453fee8602fa9b':
fix [3123221] Video sticks playing back upside down following orientation switch
Merge commit 'ce4d36ad729f83253d4c5ec9906148f45cc00f8e' into gingerbread-plus-aosp
* commit 'ce4d36ad729f83253d4c5ec9906148f45cc00f8e':
fix [3118445] Transform * Transform does not work as expected
the overlay wasn't reconfigured when the screen-orientation changed. It was
only done when a parameter of the surface itself changed.
Change-Id: I0ca0925bf58ded4c91ab89d12cb1fe4d1477c96c
The problem wasn't in the multiply operator, but rather in the code
that built the transform from the HAL bitmask.
We now use the multiply operator to build the Transform from the bitmask,
which guarantees, it'll always be correct.
Also added a simple test for Transform.
Change-Id: I09bf3b0e51d92f59d83ea91c4cc94fc2aa0bf227
Merge commit '583fefc8dcadecc7fc933513d3569dd9c16c100c' into gingerbread-plus-aosp
* commit '583fefc8dcadecc7fc933513d3569dd9c16c100c':
Use the context_priority extension when present.
Merge commit '043581382e43acbe29a8a62420bc765f49a9dd90' into gingerbread-plus-aosp
* commit '043581382e43acbe29a8a62420bc765f49a9dd90':
[3095807] screen takes a long time to turn on
Merge commit 'd4e03f37423bee383d17f7292753a5f67e497a28' into gingerbread-plus-aosp
* commit 'd4e03f37423bee383d17f7292753a5f67e497a28':
addresses parts of 3096779 and 3097475