SurfaceFlinger's (Layer's) shadow copy of the BufferQueue queue was
getting out of sync for a few reasons. This change fixes these by
doing the following:
- Adds a check to re-synchronize the shadow copy every time we
successfully acquire a buffer by first dropping stale buffers before
removing the current buffer.
- Avoids trying to perform updates for buffers which have been rejected
(for incorrect dimensions) by SurfaceFlinger.
- Adds IGraphicBufferConsumer::setShadowQueueSize, which allows the
consumer to notify the BufferQueue that it is maintaining a shadow
copy of the queue and prevents it from dropping so many buffers
during acquireBuffer that it ends up returning a buffer for which the
consumer has not yet received an onFrameAvailable call.
Bug: 20096136
Change-Id: I78d0738428005fc19b3be85cc8f1db498043612f
Adds a NATIVE_WINDOW_BUFFER_AGE query, which returns the age of the
contents of the most recently dequeued buffer as the number of frames
that have elapsed since it was last queued.
Change-Id: Ib6fd62945cb62d1e60133a65beee510363218a23
Changes Surface::detachNextBuffer to return an sp<GraphicBuffer>
instead of an ANativeWindowBuffer* to ensure that reference counting
works correctly.
Bug: 20092217
Change-Id: I3979ea6121aaf14845f0554477b778770413581e
Adds a new method IGBP::allowAllocation, which controls whether
dequeueBuffer is permitted to allocate a new buffer. If allocation is
disallowed, dequeueBuffer will block or return an error as it
normally would (as controlled by *ControlledByApp).
If there are free buffers, but they are not of the correct dimensions,
format, or usage, they may be freed if a more suitable buffer is not
found first.
Bug: 19801715
Change-Id: I0d604958b78b2fd775c2547690301423f9a52165
BufferQueue used to choose free buffers by scanning through its array
of slots and picking one based on timestamp. This changes that
mechanism to use a pair of free lists: one with buffers attached and
one without. This makes it easier to choose either type of free slot
depending on the needs of the current operation.
Fixes an issue with the first version of this change, found in bugs
20482952, 20443314, and 20464549.
Bug: 13175420
Change-Id: I9b6e83cfe8f9b4329a976025cb8e291d51fb6d4a
On normal Layers, we defer applying resize transactions until a buffer
of the correct size arrives. This breaks with sideband streams, because
buffers are not seen by SurfaceFlinger. This change applies
transactions immediately for Layers which have a sideband stream
attached to avoid that problem.
Bug: 20428254
Change-Id: I379d8aaef460d467cdf8432764b4a504928fb65a
(cherry picked from commit c300b8b506536e1bf94ffd8a3b75d46e776923e0)
When the surface damage code went in, it incorrectly assumed that if
an application was doing CPU rendering, it would be using lock and
unlockAndPost instead of dequeue and queue, so it repurposed the dirty
region too aggressively. This change keeps it from clobbering the
dirty region if a CPU producer is attached.
Bug: 20431815
Change-Id: Id4dfd71378311ea822f0289f6de2d20a7bd84014
These tests hadn't been run in a while and had accumulated some rot.
This fixes most of the compile warnings and some tests which fail
incorrectly, but 2 failures remain on N6.
Change-Id: I7fee78cd058a32f4d7df40ec9f899ff767f68517
BufferItem and GraphicBuffer were not parceling correctly, which had
not been noticed because the libgui tests (specifically, one that
tests placing a BufferQueue in a separate process from the IGBP/C)
had not been run recently. This change fixes the errors found in
those classes.
Change-Id: Ie224361a534a79115a3481d83ff97f21d154d4f5
BufferQueue used to choose free buffers by scanning through its array
of slots and picking one based on timestamp. This changes that
mechanism to use a pair of free lists: one with buffers attached and
one without. This makes it easier to choose either type of free slot
depending on the needs of the current operation.
Bug: 13175420
Change-Id: Ic8398e7511bd11a60a1c82e3ad2ee271c9822be1
This modifies EventThread such that its phase offsets for both
Choreographer and SurfaceFlinger may be modified at runtime. It also
plumbs this functionality up to the SurfaceFlinger debug interface so
that it is possible to rapidly test different offsets without
restarting the framework.
Change-Id: I426873f8553f931250dfebc9a8a4a78e9f1f4309
Passes the surface damage from the incoming SurfaceFlingerConsumer
BufferQueue down to the hardware composer HAL interface, if the
HWC version number is 1.5 or greater.
Bug: 11239309
Change-Id: Ic4305210593874a8d6deba3319055b2b8c57e926
This change adds support for passing surface damage all of the way
down from the EGL interface through the consumer side of the
BufferQueue. Depends on system/core change
Ie645e6a52b37b5c1b3be19481e8348570d1aa62c
Bug: 11239309
Change-Id: I4457ea826e9ade4ec187f973851d855b7b93a31b
Fence was flattening incorrectly, causing issues with any fields
flattened after it, and Region was relying on being the last object
flattened. This change addresses both.
Change-Id: If49de061de6e2f11f4d846b4d8c18627dfb7e109
This adds EGL wrapper functions for the following EGL extensions:
EGL_EXT_buffer_age
EGL_KHR_partial_update
EGL_KHR_swap_buffers_with_damage
Change-Id: I407acda1e0310f7f01a5efe9c915721a941138a4
To avoid nan or infinity when orientation value is used for
calculation without being initialized, check mOrientedRanges.haveOrientation
value before using mOrientedRanges.orientation.min or .max value.
Change-Id: I68ed9ab36819c5faa6422e9f061e1275aeed11e3
Signed-off-by: Baik Han <baik.han@lge.com>
New "mvuserdata" command will move all private app data from one
volume UUID to another. It leverages the existing "cp" toybox
command to do the heavy lifting for all known users, preserving
details like timestamps and permissions. It invokes restorecon()
to correctly label the new location when the copy is finished.
Changes installd to no longer drop capabilities, so we run as root
again. This also allows us to exec "cp" with CAP_DAC_OVERRIDE and
CAP_FOWNER still in effect.
Bug: 19993667
Change-Id: I1f407a7c4a1af97ca5afc27b04eb16b4936cbdef
There is memory corruption in below code
const Rect* prev = &dst[prevIndex];
dst.add(Rect(prev->right, top, right, bottom));
prev points to a memory of vector dst, when dst resize in add()
call, the memory that prev points to will be copy to the new
allocated vector memory and the old memory will become undefined
Avoid pointer in this case, use a local copy instead
Change-Id: I4d95ceedd00c8fb615ac153082ade1b1ce0d0fa8
We're now parsing and passing through volume UUIDs sent across the
command socket. The "!" argument value is treated as null, which
means internal storage.
Bug: 19993667
Change-Id: I17729a769ce687a2e94e85991a6338c77ded0b66
Teach free_cache() and restorecon_data() about building per-volume
paths. Also clean up restorecon_data() by using std::string when
building paths.
Clearer names for path building utility methods, and tests to verify.
Bug: 19993667
Change-Id: Iacfbcdaa5b901cc2490bc8eba366dfdeb44f1d93
