KMP Skia Native Render Surface Lifecycle Synchronization

--- description: KMP Skia Native Render Surface Lifecycle Synchronization - Enforces synchronous teardown of Skia hardware render targets during Android SurfaceHolder lifecycle and iOS CAMetalLayer view controller teardown. Asynchronous deallocation permits the native OS compositor to destroy backing native windows while Skia render passes are still in flight, triggering headless graphics driver panics and SIGSEGV memory faults. globs: **/src/androidMain/**/*.kt, **/src/iosMain/**/*.kt, **/src/commonMain/**/render/*.kt alwaysApply: true verified_by: "agentrules.org" license: "CC-BY-4.0" source: "https://agentrules.org/rules/kmp-skia-surface-destruction-synchronization" --- # KMP Skia Native Render Surface Lifecycle Synchronization > **Severity**: CRITICAL | **Category**: mobile > **Anti-Pattern Warning**: Asynchronous or Unsynchronized Native GPU Render-Surface Teardown in Multiplatform Bindings ## System Constraints You are a Principal Mobile Systems Architect enforcing determinism in Kotlin Multiplatform (KMP) Compose/Skia rendering layers. You must enforce that platform render surfaces (ANativeWindow/SurfaceHolder on Android, CAMetalLayer/UIViewController on iOS) synchronously block surface destruction until pending Skia GrDirectContext GPU commands finish and render targets are unregistered. ## Strict Guidelines - Engineers MUST ALWAYS synchronously flush, submit, and close the Skia GrDirectContext/Surface inside the platform destruction callback (e.g., SurfaceHolder.Callback.surfaceDestroyed or UIViewController.viewWillDisappear/deinit) before returning control to the OS window manager. - Engineers MUST NOT rely on Kotlin garbage collection, CoroutineScope cancellation, or dispatching to CoroutineDispatchers.Default to release Skia Surfaces or native ANativeWindow handles. - Platform bridges MUST hold an explicit reentrant lock or mutex between the rendering thread draw invocation and the lifecycle surface destruction thread. - Code MUST unbind the Skia BackendRenderTarget immediately after context flushing to invalidate any pending draw commands prior to native window release. - Native rendering components NEVER allow an ongoing render frame to execute draw calls once the OS has initiated surface destruction signals. ## Bad vs Good Implementations ### Android Surface Destruction Synchronization Guardrail #### Bad Antipattern (AI Common Hallucination) > **Why it fails**: surfaceDestroyed returns immediately without blocking. When the Android framework destroys the native window handle, Skia's background thread submits commands to a dead hardware buffer, crashing with fatal driver panics. ```kotlin class SkiaSurfaceView(context: Context) : SurfaceView(context), SurfaceHolder.Callback { private val scope = CoroutineScope(Dispatchers.Default) private var skiaSurface: org.jetbrains.skia.Surface? = null override fun surfaceCreated(holder: SurfaceHolder) { scope.launch { skiaSurface = initSkiaSurface(holder.surface) } } override fun surfaceChanged(holder: SurfaceHolder, format: Int, width: Int, height: Int) {} override fun surfaceDestroyed(holder: SurfaceHolder) { // BUG: Asynchronous release allows the OS to invalidate ANativeWindow // before Skia completes in-flight GPU commands, causing a driver panic (SIGSEGV). scope.launch { skiaSurface?.close() skiaSurface = null } } } ``` #### Production Standard (Correct) > **Standard**: Acquires renderLock and synchronously calls grContext.flush(syncCpu = true) followed by surface.close(). The OS thread is blocked until the GPU finishes all operations on the native buffer, preventing headless driver panics. ```kotlin class SkiaSurfaceView(context: Context) : SurfaceView(context), SurfaceHolder.Callback { private val renderLock = java.lang.Object() private var grContext: org.jetbrains.skia.DirectContext? = null private var skiaSurface: org.jetbrains.skia.Surface? = null override fun surfaceCreated(holder: SurfaceHolder) { synchronized(renderLock) { grContext = DirectContext.makeGL() skiaSurface = initSkiaSurface(holder.surface, grContext) } } override fun surfaceChanged(holder: SurfaceHolder, format: Int, width: Int, height: Int) {} override fun surfaceDestroyed(holder: SurfaceHolder) { // INVARIANT: Block the OS thread until GPU flush completes and surfaces are released. synchronized(renderLock) { grContext?.flush(syncCpu = true) skiaSurface?.close() skiaSurface = null grContext?.abandon() grContext?.close() grContext = null } } } ```

Architectural Specification & Invariant Rationale

Architectural Invariant Design: KMP Skia Native Render Surface Lifecycle Synchronization

KMP Skia Native Render Surface Lifecycle Synchronization defines an industrial-grade behavioral standard for mobile engineering. Enforces synchronous teardown of Skia hardware render targets during Android SurfaceHolder lifecycle and iOS CAMetalLayer view controller teardown. Asynchronous deallocation permits the native OS compositor to destroy backing native windows while Skia render passes are still in flight, triggering headless graphics driver panics and SIGSEGV memory faults. Modern software architectures require strict adherence to deterministic patterns to guarantee high availability, maintainability, and zero security vulnerabilities.

Framework Paradigm Shift & Drift MechanicsLLMs are trained on millions of open-source repositories containing outdated code, legacy idioms, and unpatched antipatterns. When coding assistants generate implementations for this stack, they default to common statistical tokens: Asynchronous or Unsynchronized Native GPU Render-Surface Teardown in Multiplatform Bindings Without an explicit system instruction file, developers spend hours debugging silent failures.

Failure Modes & Production Blast Radius

Hazard Analysis: What Breaks Without This Rule

The following failure modes have been cataloged from hundreds of hours of production code reviews where AI assistants produced non-compliant implementations:

Failure Mode	AI Symptom (The Anti-Pattern)	Severity	Production Impact & Remedy
AI Hallucination & Deprecation Drift	Asynchronous or Unsynchronized Native GPU Render-Surface Teardown in Multiplatform Bindings	Critical	Violating this industrial standard introduces regression bugs, security risks, and unexpected runtime failures in production. Enforce explicit behavioral system directives across all AI coding assistants in your repository.
Silent Architecture Inconsistency	AI assistants produce mixed paradigms across files, creating fragmented codebase conventions.	Medium	Increases cognitive load during code reviews and leads to maintenance debt across engineering teams. Pin project conventions in .cursor/rules, CLAUDE.md, and AGENTS.md so all team agents follow identical guardrails.

Architectural Directives in Detail

Step-by-Step Production Directives

Each rule constraint below is injected into the AI agent's system prompt to establish deterministic boundaries:

Engineers MUST ALWAYS synchronously flush, submit, and close the Skia GrDirectContext/Surface inside the platform destruction callback (e.g., SurfaceHolder.Callback.surfaceDestroyed or UIViewController.viewWillDisappear/deinit) before returning control to the OS window manager.

Engineers MUST ALWAYS synchronously flush, submit, and close the Skia GrDirectContext/Surface inside the platform destruction callback (e.g., SurfaceHolder.Callback.surfaceDestroyed or UIViewController.viewWillDisappear/deinit) before returning control to the OS window manager.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.