Introduces a mechanism to allow backdrop filters to 1) share backdrop inputs and 2) fuse filter applications for faster blurs.
This is a proposed solution to https://github.com/flutter/flutter/issues/131568
Implemented:
* Developer can specify a "backdrop id" which indicates that a backdrop layer should share the input texture and potentially cached filter for a layer.
* Removes second save layer for each backdrop filter
* Removes save layer trace event for backdrop filter
* Can fuse backdrop filters if there is more than one identical filter
TBD:
* Adjust heruristic to avoid applying bdf filter to entire screen
Suggestions: applying a bdf should be a distinct operation from a save layer in the DL builder/dispatcher. The saveLayer implmenentation in the impeller dispatcher is super convoluted because it needs to handle both.
### Video
Video starts with normal bdf then I hot reload to specify that the bdfs share inputs/filters. This is running on a pixel 8 pro
Change to the macrobenchmark app is just:
```dart
Widget build(BuildContext context) {
Widget addBlur(Widget child, bool shouldBlur) {
if (shouldBlur) {
return ClipRect(
child: BackdropFilter(
filter: ImageFilter.blur(sigmaX: 5, sigmaY: 5),
backdropId: 1, // Added ID
child: child,
),
);
} else {
return child;
}
}
```
https://github.com/user-attachments/assets/22707f97-5825-43f1-91b4-1a02a43437f5
Requires framework changes in https://github.com/jonahwilliams/flutter/pull/new/backdrop_id
Avoid Sprintf'ing labels for textures/cmd buffers/render passes. For the render target textures, defer combined label construction until we know the label will be used.
Many of our labeling APIs accepted a std::string. When provided with a string literal (const char*), this would force the allocation of a new std::string object and a copy of the original char*. std::string_view is a view of either a std::string, or a const char* (or other), which allows an API that doesn not need to store the string to accept both without a copy. Of course, we can't store a std::string_view, as that doesn't provide memory ownership - so many of the GLES implementations will still copy the string_view into a string locally.
Also updates several of the debug labelling functions to no-op in release mode and removes some expensive interpolations that aren't super necessary.
Fixes [#146251](https://github.com/flutter/flutter/issues/146251).
On Windows, a "pointer add" event to a new view is sometimes sent before the "pointer remove" event from the old view. When `PointerDataPacketConverter` handles the add event, it asserts that the device's pointer state has been cleared. Since the pointer state is cleared only when the remove event is processed, the assertion fails if the events arrive out of order.
The solution proposed here is to synthesize a remove event if an add event is received while the pointer state hasn't been cleared (i.e., if the pointer is added without being removed from the previous view).
To avoid duplicate remove events, the view ID is now tracked in `PointerState`. If the original "remove" arrives after the synthesized one, the state's view ID will differ from the pointer data's view ID, allowing it to be safely ignored.
When synthesizing the remove event, it's possible that the old view has already been destroyed, meaning no remove event will be received later. For example, this occurs when a window is destroyed while the pointer is inside it. However, the framework expects an "add" event to always follow a "remove" event, and "remove" events with invalid views are ignored. To meet the framework's expectations, the view ID of the "add" event will be used for the "remove" event if the old view has been destroyed.
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
I used this fixture data early on to bootstrap testing for Flutter GPU before shader bundles were finished. But today we have the build system compile shader bundles for us, so this is no longer used.
b/371512414
This issue had no reproduction but the stacktrace was provided. The theory of this fix is that the `ImageResult` type is capturing something that doesn't handle being copied and deallocated twice so instead of copying it, we std::move it into a shared_ptr so that it will only be deallocated once.
Also, I just avoid using a `std::move` when overflowing. The idea here is maybe sometimes deleting the std::move'd item isn't kosher. We know from the stacktrace that it has something to do with overflowing because that would be the only case where something is being deleted.
I've added 2 integration tests that exercises the code where the crash appears to be happening. This is good coverage to have but neither of them reproduced the issue, even with MallocScribble enabled. That's the best we can do without a reproduction.
```
Thread 7 (id: 0x0000a103)crashed
0x000000010a48ae8c(Flutter -function.h)std::_fl::allocator<impeller::ContextMTL::PendingTasks>::destroy[abi:v15000](impeller::ContextMTL::PendingTasks*)
0x000000010a48b9fc(Flutter -allocator_traits.h:309)impeller::ContextMTL::StoreTaskForGPU(std::_fl::function<void ()> const&, std::_fl::function<void ()> const&)
0x000000010a48b9fc(Flutter -allocator_traits.h:309)impeller::ContextMTL::StoreTaskForGPU(std::_fl::function<void ()> const&, std::_fl::function<void ()> const&)
0x000000010a4d49e4(Flutter -image_decoder_impeller.cc:422)std::_fl::__function::__func<flutter::ImageDecoderImpeller::UploadTextureToPrivate(std::_fl::function<void (sk_sp<flutter::DlImage>, std::_fl::basic_string<char, std::_fl::char_traits<char>, std::_fl::allocator<char>>)>, std::_fl::shared_ptr<impeller::Context> const&, std::_fl::shared_ptr<impeller::DeviceBuffer> const&, SkImageInfo const&, std::_fl::shared_ptr<SkBitmap> const&, std::_fl::optional<SkImageInfo> const&, std::_fl::shared_ptr<fml::SyncSwitch> const&)::$_1, std::_fl::allocator<flutter::ImageDecoderImpeller::UploadTextureToPrivate(std::_fl::function<void (sk_sp<flutter::DlImage>, std::_fl::basic_string<char, std::_fl::char_traits<char>, std::_fl::allocator<char>>)>, std::_fl::shared_ptr<impeller::Context> const&, std::_fl::shared_ptr<impeller::DeviceBuffer> const&, SkImageInfo const&, std::_fl::shared_ptr<SkBitmap> const&, std::_fl::optional<SkImageInfo> const&, std::_fl::shared_ptr<fml::SyncSwitch> const&)::$_1>, void ()>::operator()()
0x000000010a06c234(Flutter -function.h:512)fml::SyncSwitch::Execute(fml::SyncSwitch::Handlers const&) const
0x000000010a4d42f8(Flutter -image_decoder_impeller.cc:408)flutter::ImageDecoderImpeller::UploadTextureToPrivate(std::_fl::function<void (sk_sp<flutter::DlImage>, std::_fl::basic_string<char, std::_fl::char_traits<char>, std::_fl::allocator<char>>)>, std::_fl::shared_ptr<impeller::Context> const&, std::_fl::shared_ptr<impeller::DeviceBuffer> const&, SkImageInfo const&, std::_fl::shared_ptr<SkBitmap> const&, std::_fl::optional<SkImageInfo> const&, std::_fl::shared_ptr<fml::SyncSwitch> const&)
0x000000010a4d3838(Flutter -image_decoder_impeller.cc:538)std::_fl::__function::__func<flutter::ImageDecoderImpeller::Decode(fml::RefPtr<flutter::ImageDescriptor>, unsigned int, unsigned int, std::_fl::function<void (sk_sp<flutter::DlImage>, std::_fl::basic_string<char, std::_fl::char_traits<char>, std::_fl::allocator<char>>)> const&)::$_1, std::_fl::allocator<flutter::ImageDecoderImpeller::Decode(fml::RefPtr<flutter::ImageDescriptor>, unsigned int, unsigned int, std::_fl::function<void (sk_sp<flutter::DlImage>, std::_fl::basic_string<char, std::_fl::char_traits<char>, std::_fl::allocator<char>>)> const&)::$_1>, void ()>::operator()()
0x000000010a06dd1c(Flutter -function.h:512)fml::ConcurrentMessageLoopDarwin::ExecuteTask(std::_fl::function<void ()> const&)
0x000000010a06737c(Flutter -concurrent_message_loop.cc:101)void* std::_fl::__thread_proxy[abi:v15000]<std::_fl::tuple<std::_fl::unique_ptr<std::_fl::__thread_struct, std::_fl::default_delete<std::_fl::__thread_struct>>, fml::ConcurrentMessageLoop::ConcurrentMessageLoop(unsigned long)::$_0>>(void*)
0x000000021d4ff378(libsystem_pthread.dylib + 0x00006378)_pthread_start
```
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
This enhances the overlay optimization by delaying combining pictures to get tighter bounds for the pictures that make up the scene, enabling more sophisticated optimization since we can determine if they intersect with platform views on a per-picture basis.
Fixes https://github.com/flutter/flutter/issues/149863
Fixes https://github.com/flutter/flutter/issues/155833
On a Macbook in Chrome in an example app with an infinite scrolling grid of platform views, this brings the ratio of dropped frames from 93% to 55% (roughly 4 fps to 30 fps).
This is a reland of https://github.com/flutter/engine/pull/54878 with a fix for scenes with pictures and shader masks that are eventually entirely clipped out. It also fixes a performance issue caused by making too many Canvases just to record the size of the picture elements in the scene.
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
This is an attempt to reland the overlay optimization for skwasm and fixing the golden diffs from the framework tests.
Original PR description:
This PR refactors the scene builder's logic in order to more aggressively merge flutter content and platform view content together. This essentially covers the case discussed in this flutter issue: https://github.com/flutter/flutter/issues/149863
This optimization ensures that each picture or platform view is applied to the lowest possible slice in the scene, which avoids the proliferation of redundant slices and overlays in the scene.
This enhances the overlay optimization by delaying combining pictures to get tighter bounds for the pictures that make up the scene, enabling more sophisticated optimization since we can determine if they intersect with platform views on a per-picture basis.
Fixes https://github.com/flutter/flutter/issues/149863
On a Macbook in Chrome in an example app with an infinite scrolling grid of platform views, this brings the ratio of dropped frames from 93% to 55% (roughly 4 fps to 30 fps).
This is a reland of https://github.com/flutter/engine/pull/54878 with a fix for scenes with pictures and shader masks that are eventually entirely clipped out.
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
Reverts: flutter/engine#55402
Initiated by: chingjun
Reason for reverting: caused internal tests to fail.
See b/369740500 for more details.
Original PR Author: harryterkelsen
Reviewed By: {yjbanov}
This change reverts the following previous change:
This enhances the overlay optimization by delaying combining pictures to get tighter bounds for the pictures that make up the scene, enabling more sophisticated optimization since we can determine if they intersect with platform views on a per-picture basis.
Fixes https://github.com/flutter/flutter/issues/149863
On a Macbook in Chrome in an example app with an infinite scrolling grid of platform views, this brings the ratio of dropped frames from 93% to 55% (roughly 4 fps to 30 fps).
This is a reland of https://github.com/flutter/engine/pull/54878 with a fix for scenes with pictures that are eventually entirely clipped out.
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
This enhances the overlay optimization by delaying combining pictures to get tighter bounds for the pictures that make up the scene, enabling more sophisticated optimization since we can determine if they intersect with platform views on a per-picture basis.
Fixes https://github.com/flutter/flutter/issues/149863
On a Macbook in Chrome in an example app with an infinite scrolling grid of platform views, this brings the ratio of dropped frames from 93% to 55% (roughly 4 fps to 30 fps).
This is a reland of https://github.com/flutter/engine/pull/54878 with a fix for scenes with pictures that are eventually entirely clipped out.
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
In the Vulkan backend, the CommandPoolRecyclerVK stores command pools that can be reused while rendering a frame. For each frame, SurfaceContextVK::AcquireNextSurface calls CommandPoolRecyclerVK::Dispose to clear the cached pools.
However, the CommandPoolRecyclerVK's cache of pools is placed in thread-local storage. So command pools used on the IO thread will not be cleaned up if frame rendering calls Dispose on the raster thread.
This PR adds a Context API that the IO thread can call after using a command pool in order to ensure that cached resources are disposed.
Fixes https://github.com/flutter/flutter/issues/155519
This enhances the overlay optimization by delaying combining pictures to get tighter bounds for the pictures that make up the scene, enabling more sophisticated optimization since we can determine if they intersect with platform views on a per-picture basis.
Fixes https://github.com/flutter/flutter/issues/149863
On a Macbook in Chrome in an example app with an infinite scrolling grid of platform views, this brings the ratio of dropped frames from 93% to 55% (roughly 4 fps to 30 fps).
[C++, Objective-C, Java style guides]: https://github.com/flutter/engine/blob/main/CONTRIBUTING.md#style
In certain situations, semantics elements get assigned `pointer-events: none` when they aren't supposed to.
One such situation is when a text field has a [decoration error text](https://api.flutter.dev/flutter/material/InputDecoration/errorText.html). The semantics node become a container, and we always set `pointer-events: none` on container nodes.
This PR introduces an `acceptsPointerEvents` getter on `SemanticRole` and `SemanticBehavior` to control when `pointer-events` should be `all` or `none`.
Fixes https://github.com/flutter/flutter/issues/141975
This is placeholder stuff that I added before the rest of the API to prove that we can call the engine symbols.
Today this is totally redundant as Flutter GPU has a bunch of automated tests which exercise every FFI call.
Part of https://github.com/flutter/flutter/issues/150953.
Provide a way to get the required minimum uniform byte alignment when referencing uniform blocks in a device buffer. Allow the user to explicitly flush DeviceBuffers (necessary for devices without shared memory).
Reverts: flutter/engine#54949
Initiated by: eyebrowsoffire
Reason for reverting: Incorrect golden diffs on engine roll, see https://github.com/flutter/flutter/pull/155181
Original PR Author: eyebrowsoffire
Reviewed By: {harryterkelsen}
This change reverts the following previous change:
This PR refactors the scene builder's logic in order to more aggressively merge flutter content and platform view content together. This essentially covers the case discussed in this flutter issue: https://github.com/flutter/flutter/issues/149863
This optimization ensures that each picture or platform view is applied to the lowest possible slice in the scene, which avoids the proliferation of redundant slices and overlays in the scene.
ui.Canvas and ui.SceneBuilder now use the DlPath object directly from the ui.Path object. This results in increased sharing of the wrapper objects which then increases the sharing of both the converted Impeller paths and Skia's volatile flag.
The VolatilePathTracker mechanism is deleted and rather than count the number of frames that a path is stable for, instead we count the number of times it is used for rendering. If a path is used 100 times in a single frame, it will become non-volatile and start being cached almost immediately. The cached Impeller paths are now also tracked for all instances of the same path, rather than for each call site that originated from a DisplayList dispatch.
