TL;DR: Offscreen surface is created on the render thread and device to host
transfer performed there before task completion on the UI thread.
While attempting to snapshot layer trees, the engine was attempting to use the
IO thread context. The reasoning was that this would be safe to do because any
textures uploaded to the GPU as a result of async texture upload would have
originated from this context and hence the handles would be valid in either
context. As it turns out, while the handles are valid, Skia does not support
this use-case because cross-context images transfer ownership of the image from
one context to another. So, when we made the hop from the UI thread to the IO
thread (for snapshotting), if either the UI or GPU threads released the last
reference to the texture backed image, the image would be invalid. This led to
such images being absent from the layer tree snapshot.
Simply referencing the images as they are being used on the IO thread is not
sufficient because accessing images on one context after their ownership has
already been transferred to another is not safe behavior (from Skia's
perspective, the handles are still valid in the sharegroup).
To work around these issues, it was decided that an offscreen render target
would be created on the render thread. The color attachment of this render
target could then be transferred as a cross context image to the IO thread for
the device to host tranfer.
Again, this is currently not quite possible because the only way to create
cross context images is from encoded data. Till Skia exposes the functionality
to create cross-context images from textures in one context, we do a device to
host transfer on the GPU thread. The side effect of this is that this is now
part of the frame workload (image compression, which dominate the wall time,
is still done of the IO thread).
A minor side effect of this patch is that the GPU latch needs to be waited on
before the UI thread tasks can be completed before shell initialization.
Also adds //flutter/benchmarking which, similar to //flutter/testing, allows for the creation of a benchmarking executable. This is also the target that contains benchmarking utilities.
- Switch core snapshot to Dart 2 and remove support for loading platform.dill.
- Remove support for loading script snapshots.
- Remove support for loading source.
- Remove settings and fix names to reflect the above.
- Remove support for loading the service isolate from source.
When these delegate methods were initially added, it was expected that a single
shell would be able to own mutliple platform views, engines and animators. This
plan was abandoned in favor of creating multiple shells with their own platform
views, engines, etc.. The arguments were meant to ease the disambiguate the
instances of the variaous objects managed by the shell. This is no longer
necessary.
This change adds a delay before Animator::BeginFrame calls its
delegate's OnAnimatorNotifyIdle. This is because under certain
workloads, such as our parent view resizing us, which is communicated
via viewport change events, we won't have a frame scheduled yet in the
animator, despite the fact that we will go on to schedule a frame once
the viewport event arrives.
In Fuchsia's resizing performance test, on our reference high end x86-64
hardware, the previous logic was resulting in a ~45ms garbage collection
right after the first frame of an animation.
* Flush all embedded Android view on hot restart.
Adds an OnEngineRestarted method to PlatformView, this is currently only
implemented for Android where we need to use it for embedded views.
* review comments followup
* rename to OnPreEngineRestart, call before Clone
Don't emit a "VSYNC" event when running on Fuchsia, as traces on Fuchsia
are typically recorded across the whole system, causing the events to
collide with each other.
