We utilized a `FenceCycle` to keep track of if the buffer was mutable or not and introduced another cycle to track GPU-side requirements only on fulfillment of which could the buffer be utilized on the host but due to the recent change in the behavior this system ended up being unoptimal.
This commit replaces the cycle with a boolean tracking if there are any usages of the resource on the GPU within the current context that may prevent it from being mutated on the CPU. The fence of the context is simply attached to the buffer based off this which was allowed as the new behavior of buffer fences matches all the requirements for this.
An atomic transactional loop was performed on the backing `std::shared_ptr` inside `BufferView`/`TextureView`'s `lock`/`LockWithTag`/`try_lock` functions, these locks utilized `std::atomic_load` for atomically loading the value from the `shared_ptr` recursively till it was the same value pre/post-locking.
This commit abstracts the locking functionality of `TextureView`/`BufferDelegate` into `LockableSharedPtr` to avoid code duplication and removes the usage of `std::atomic_load` in either case as it is not necessary due to the implicit memory barrier provided by locking a mutex.
`PresentationEngine` and `GraphicBufferProducer` methods that utilized textures for the surface utilized the `Texture` type rather than the `TextureView` type, this was never correct but at the time of authoring this code `TextureView` was not finalized and in a major flux which is why it was not utilized and `Texture` was utilized instead. Now that is is far more stable, it has been replaced with `TextureView`.
We want to block on the host thread during presentation while the host surface isn't present to implicitly pause the game, this can end up being fairly costly as it involves locking the `PresentationEngine` mutex which can lead to a lot of contention with the presentation thread. This fixes the issue by polling if there is a surface and only if there isn't then doing the wait as it isn't mandatory to wait always, we'll eventually run into the guest thread stalling.
Multiple threads concurrently accessing the `TextureManager`/`BufferManager` (Referred to as "resource managers") has a potential deadlock with a resource being locked while acquiring the resource manager lock while the thread owning it tries to acquire a lock on the resource resulting in a deadlock.
This has been fixed with locking of resource manager now being externally handled which ensures it can be locked prior to locking any resources, `CommandExecutor` provides accessors for retrieving the resource manager which automatically handles locking aside doing so on attachment of resources.
GPU resources have been designed with locking by fences in mind, fences were treated as implicit locks on a GPU, design paradigms such as `GraphicsContext` simply unlocking the texture mutex after attaching it which would set the fence cycle were considered fine prior but are unoptimal as it enforces that a `FenceCycle` effectively ensures exclusivity. This conflates the function of a mutex which is mutual exclusion and that of the fence which is to track GPU-side completion and led to tying if it was acceptable to use a GPU resource to GPU completion rather than simply if it was not currently being used by the CPU which is the function of the mutex.
This rework fixes this with the groundwork that has been laid with previous commits, as `Context` semantics are utilized to move back to using mutexes for locking of resources and tracking the usage on the GPU in a cleaner way rather than arbitrary fence comparisons. This also leads to cleaning up a lot of methods that involved usage of fences that no longer require it and therefore can be entirely removed, further cleaning up the codebase. It also opens the door for future improvements such as the removal of `hostImmutableCycle` and replacing them with better solutions, the implementation of which is broken at the moment regardless.
While moving to `Context`-based locking the question of multiple GPU workloads being in-flight while using overlapping resources came up which brought a fundamental limitation of `FenceCycle` to light which was that only one resource could be concurrently attached to a cycle and it could not adequately represent multi-cycle dependencies. `FenceCycle` chaining was designed to fix this inadequacy and allows for several different GPU workloads to be in-flight concurrently while utilizing the same resources as long as they can ensure GPU-GPU synchronization.
If we want to allow submitting multiple pieces of work to the GPU at once while still requiring CPU synchronization, we'll need to track all past fence cycles associated with a resource alongside the current one. To solve this the concept of chaining fences has been introduced, fences from past usages can be chained to the latest fence which'll then recursively forward operations to chained fences.
This change also ends up mandating a move away from `FenceCycleDependency` as it would prevent fences from concurrently locking the same resources which is required for chaining to work as two fences being chained fundamentally means they're locking the same resources. The `AtomicForwardList` is therefore used as the new container.
Resources on the GPU can be fairly convoluted and involve overlaps which can lead to the same GPU resources being utilized with different views, we previously utilized fences to lock resources to prevent concurrent access but this was overly harsh as it would block usage of resources till GPU completion of the commands associated with a resource.
Fences have now been replaced with locks but locks run into the issue of being per-view and therefore to add a common object for tracking usage the concept of "tags" was introduced to track a single context so locks can be skipped if they're from the same context. This is important to prevent a deadlock when locking a resource which has been already locked from the current context with a different view.
We currently present all frames synchronously on the thread that calls into SurfaceFlinger functions, this is unoptimal as it doesn't match guest behavior which can lead to delaying the guest from working on the next frame. This commit queuing up frames to non-blocking and handles all waiting then presenting the frame on a dedicated thread.
We utilize `pthread_setname_np` to set the thread names but didn't check for any errors which resulted in the `Skyline-Choreographer` and `ChannelCmdFifo` not having proper names as they exceeded the 16 character limit on thread names for the pthread function. This has now been fixed by changing the names and introducing error checking to invocations of this function.
We waited on the host GPU after `Execute` but this isn't optimal as it causes a major stall on the CPU which can lead to several adverse effects such as downclocking by the governor and losing the opportunity to work in parallel with the GPU.
This has now been fixed by splitting `Execute`'s functionality into two functions: `Submit` and `SubmitWithFlush` which both execute all nodes and submit the resulting command buffer to the GPU but flushing will wait on the GPU to complete while the non-flush variant will not wait and work ahead of the GPU.
We need move-assignment semantics to viably utilize these objects as class members, they cannot be replaced without move-assign (or copy-assign but that is undesirable here). This commit fixes that by introducing a move assignment operator to them while making the `slot` a pointer which has the necessary nullability semantics.
A `Setting` delegate class has been introduced, holding the raw value of the setting and adding support for registering callbacks to that setting. Callbacks will then be called when the value of that setting changes.
As a result of this, raw setting values have been made accessible through pointer dereference semantics.
Mali GPU drivers utilize the `ppoll()` syscall inside `waitForFences` which isn't correctly restarted after a signal, which we can receive at any time on a guest thread. This commit fixes that by recursively calling the function on failure till it succeeds or returns an unexpected error.
Co-authored-by: PixelyIon <pixelyion@protonmail.com>
Co-authored-by: Billy Laws <blaws05@gmail.com>
Certain non-indexed quad draws would mistakenly take the indexed quad path because of the assumption that they would not have a bound index buffer. This resulted in a crash for most games using quads due to a faulty exception `Indexed quad conversion is not supported`, when in fact they were not using indexed quads.
Co-authored-by: PixelyIon <pixelyion@protonmail.com>
Co-authored-by: Billy Laws <blaws05@gmail.com>
This commit implements several key optimisations in megabuffering that are all inherently interlinked.
- Megabuffering is moved from per-buffer to per-view copies, this makes megabuffering possible for small views into larger underlying buffers which is often the case with even the simplest of games,
- Megabuffering is no longer the default option, it is only enabled for buffer views that have had inline GPU writes applied to them in the past as that is the only case where they are beneficial. In any other case the cost of copying, even with a 128KiB limit can be significant.
- With both of these changes, there is now possibility for overlapping views where one uses megabuffering and one does not. In order to allow GPU inline writes to work consistently in such cases a system of 'host immutability' has been implemented, when a buffer is marked as host immutable for a given cycle, all writes to the buffer from that point to the point the cycle is signalled will be performed on the GPU, ensuring that the backing contents are correctly sequenced
Has the same guarantees of pointer stabilty while also being significantly faster in cases where a buffer has thousands of views. This is the case in RE4 and this change leads to an almost 1000% performance improvement in that game.
We currently have a global `MegaBuffer` instance that is shared across all channels, this is very problematic as `MegaBuffer` fundamentally works like a state machine with allocations (especially resetting/freeing) and is thread-specific. Therefore, we now have a pool of several `MegaBuffer`s which is allocated from by the `CommandExecutor` and kept channel specific as a result which also limits its usage to a single thread, this allows for individually resetting or freeing any allocations.
There was a lot of redundant code in the `CommandScheduler` when the same functionality could be achieved with much shorter and cleaner code which this commit fixes. This includes no changes to the user-facing API and does not require any changes on the user side as a result.
Some games remap rendertargets or map them late which would lead to weird graphical bugs or crashes. Drop the caching since VMM lookup is fairly cheap anyway.
The `VkBufferImageCopy` offset calculations were wrong inside `CopyIntoStagingBuffer` as it multiplied the mip level's linear size by `levelCount` rather than `layerCount`. This led to substantial UB in games which called this function as it led to an overflow and resulted in writing to other areas of the buffer which caused major issues such as vertex/index buffer corruption and corresponding graphical glitches alongside likely being the cause of some crashes.
BC7 CPU decoding had the red and blue channels swapped around as it outputted a BGRA image after decoding while we expected an RGBA image to be produced. This should fix the colors of certain textures in titles such as Cuphead or Sonic Forces.
We can't render to a 3D texture through a 3D view, we instead have to create a 2D array view into it and render to that. The texture manager previously didn't support having a different view type/layer count between a guest texture view and the underlying storage texture that is required to support this so that was also implemented by reading the view layer count from the dimensions depth instead if the underlying texture is 3D (and the view type is 2D array). Additionally move away from our own view type enum to Vulkan, inline with other guest texture member types.
Sampler anisotropy was made a required feature in an earlier commit due to its widespread availability but this was determined to be incorrect as certain Mali GPUs that can otherwise run 2D games in Skyline do not have this feature, while they are still not officially supported as this was the only roadblock to support them, it has now been made an optional feature.
Certain GPU vendors such as ARM's Mali do not have support for BCn textures whatsoever while other vendors such as AMD only have partial support (BC1-BC3). Most titles on the guest utilize BC textures and to address this on host GPUs without support for BCn, we need to decompress the texture on the CPU. This commit implements a CPU BCn texture decoder based off Swiftshader's BC decoder, it also adds the necessary infrastructure to have different formats for the `GuestTexture` and `Texture` objects.
The iterations of the inner loop for sector deswizzling was miscalculated as `SectorWidth * SectorHeight` while the result was correct at `32`, it should be determined by the amount of sector lines within a GOB i.e.: `(GobWidth / SectorWidth) * GobHeight`.
Support for mipmapped textures was not implemented which is fairly crucial to proper rendering of games as the only level that would load is the first level (highest resolution), that might result in a lot more memory bandwidth being utilized. Mipmapping also has associated benefits regarding aliasing as it has a minor anti-aliasing effect on distant textures.
This commit entirely implements mipmapping support but it does not extend to full support for views into specific mipmap levels due to the texture manager implemention being incomplete.
Maxwell DMA requires swizzled copies to/from textures and earlier it had to construct an arbitrary `GuestTexture` to do so but with the introduction of the cleaner API, this has become redundant which this commit cleans up and replaces with direct calls to the API with all the necessary values.
The API for texture swizzling is now more concrete and abstracted out from `GuestTexture`, this allows for neater usage in certain areas such as MaxwellDMA while having a `GuestTexture` wrapper as well allowing for neater usage in those cases.
The code itself has also been cleaned up slightly with all usage of `u32`s being upgraded to `size_t` as this is simply more efficient due to the compiler not needing to emulate wraparound behavior for integer types smaller than the processor word size.
The Fermi 2D engine implements both image blit and resolve operations, supporting subpixel sampling with both linear and point filtering.
Resolve operations are performed by sampling from the center of each pixel in order to resolve the final image from the MSAA samples
MSAA images are stored in memory like regular images but each pixels dimensions are scaled: e.g for 2x2 MSAA
```
112233
112233
445566
445566
```
These would be sampled with both duDx and duDy as 2 (integer part), resolving to the following:
```
123
456
```
Blit operations are performed by sampling from the corner of each pixel, scaling the image as one would expect.
This implementation isn't fully complete as Vulkan blit doesn't support some combinations which Fermi does, most notably between colour and depth stencil. These will be implemented properly at a later date, likely after the texture manager rework.
Out of Bounds Blit, used by some OpenGL games is also missing since supporting it requires texture aliasing, this will also be supported after the texture manager rework.
Co-authored-by: Billy Laws <blaws05@gmail.com>
Certain writes during swizzling went out of bounds due to incorrect `blockExtentY` calculation, the previous commit to fix this ended up breaking it further. This commit returns to the original commit's calculations with the proper addendum of a check for exact alignment with a GOB which is the case that was broken earlier.
The `GuestTexture::GetLayerStride` function was not always being utilized to retrieve the layer stride inside `Texture`, it would instead directly access the `guestTexture::layerStride` member. This is problematic as it may not be initialized and return `0` which would lead to a broken image copy.
