runtime/wasm: add timers and host event wakeups (based on #2198)

[cpunion]

Depends on #2198.

Problem

The single-worker wasm runtime can park goroutines, but it has no host event source for time. time.Sleep, timers, tickers, timeout selects, and AfterFunc therefore cannot suspend one G while allowing the worker to run another G and later resume the timer owner.

Implementation

• Add a private runtime/internal/wasmevent min-heap with stable timer ordering, Stop/Reset semantics, periodic deadline catch-up, and a lazy dispatch layer. Packages that never use time retain only a small dormant dispatch path.
• Poll due events from the wasm scheduler. When the run queue is empty, wait through the host adapter; callbacks run only after control returns to the scheduler, so there is no synchronous Go re-entry during Asyncify unwind/rewind.
• Use emscripten_sleep for J32/J64 host wakeups and poll(NULL, 0, timeout) for default WASI Preview 1. Keep the existing explicit LLGO_WASI_THREADS=1 timer backend unchanged.
• Use a monotonic clock for runtime deadlines while retaining normal wall-clock/time-zone behavior for time.Now.
• Add standard-library //go:wasmimport lowering so WASI libc/runtime declarations retain their official host module/name.
• Update the Emscripten embind adapter for both legacy and current unified EM_INVOKER contracts.

The compiler, scheduler, timer queue, and host adapters remain separate. No new target or build tag is introduced, and native/embedded source selection is unchanged. Static wasm defer dispatch is supplied by the inherited #2208 dependency and is not part of this PR-specific diff.

Semantics covered

The shared J32, J64, and P1 acceptance fixture covers:

• wall clock and local time zone
• time.Sleep
• active, stopped, reset, and expired Timer behavior
• Ticker
• AfterFunc waking a channel waiter
• timeout select

Validation

Local macOS validation uses GOMAXPROCS=2, GOMEMLIMIT=3GiB, and -p=1 where applicable:

• changed cl/ssa import, locality, and defer tests: pass
• nested runtime timer queue and dispatch tests: pass
• J32/J64/P1 timer fixtures: execute successfully with wasm timers ok
• P1 wasm-tools validate --features all: pass
• Codecov patch coverage: 97.29730% (gate passes)
• The latest main/ssa/wasm: use static defer continuation dispatch #2208 dependency stack passes the full GitHub Ubuntu/macOS matrix and Codecov

Earlier broader validation also covered -O0, -O3, ThinLTO, explicit P1 pthread compatibility, the full nested runtime module, and bounded Ubuntu tests.

Coverage highlights:

• runtime/internal/wasmevent: 93.9%
• lazy dispatch functions: 100%
• wasmImportByDoc: 100%
• SetWasmImport: 100%

Cost

On Apple M4 Max with GOMAXPROCS=2, queue Reset+RunDue is 13.51-14.87 ns/op, 0 B/op, 0 allocs/op.

Against #2198, a no-time J32 scheduler fixture grows from 416,551 to 418,082 bytes with -pclntab=none: +1,531 bytes (+0.37%). With the default pclntab it grows from 568,087 to 571,522 bytes: +3,435 bytes (+0.60%, including metadata). Native output size remains unchanged apart from worktree-path metadata, and the representative bare-metal ELF is byte-identical.

The PR-specific diff above #2198 is 24 files, 1,048 insertions, and 45 deletions.

[codecov]

Codecov Report

❌ Patch coverage is 97.29730% with 5 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
internal/build/wasm_postlink.go	95.23%	2 Missing and 1 partial ⚠️
internal/build/build.go	85.71%	1 Missing and 1 partial ⚠️

📢 Thoughts on this report? Let us know!