runtime: run syscall/js finalizers on wasm without a manual GC (#5545)

* runtime: run syscall/js finalizers on wasm without a manual GC

* runtime: address review feedback on finalizer idle GC

* runtime: clear a finished task's args pointer so its arguments are collectable

* runtime: skip the finalizer scan with a per-block registration bit

* runtime: guard the finalizer registration bitmap with gcLock

* testdata: cover finalizer invariants on every scheduler

* main_test: limit the finalizer scheduler variants to linux and darwin

* testdata: wait for the finalizer queue to drain before asserting

* testdata: make the finalizer counters atomic and wait for a known drain count

* runtime: add finalizer bookkeeping asserts under runtime_asserts

* runtime: address finalizer GC review feedback

* testdata: strengthen blocked stack finalizer test

* runtime: fix finalizer cleanup edge cases

* runtime: decouple wasm export scheduling from finalizers

* runtime: avoid redundant wakeups for re-entrant wasm exports

* runtime: simplify finalizer comments

compiler/goroutine.go

@@ -309,10 +309,10 @@ func (c *compilerContext) createGoroutineStartWrapper(fnType llvm.Type, fn llvm.
 	}
 	defer b.Dispose()
 
-	var deadlock llvm.Value
-	var deadlockType llvm.Type
+	var exitGoroutine llvm.Value
+	var exitGoroutineType llvm.Type
 	if c.Scheduler == "asyncify" {
-		deadlockType, deadlock = c.getFunction(c.program.ImportedPackage("runtime").Members["deadlock"].(*ssa.Function))
+		exitGoroutineType, exitGoroutine = c.getFunction(c.program.ImportedPackage("runtime").Members["exitGoroutine"].(*ssa.Function))
 	}
 
 	if !fn.IsAFunction().IsNil() {
@@ -377,7 +377,7 @@ func (c *compilerContext) createGoroutineStartWrapper(fnType llvm.Type, fn llvm.
 			b.CreateCall(fnType, fn, params, "")
 
 			if c.Scheduler == "asyncify" {
-				b.CreateCall(deadlockType, deadlock, []llvm.Value{
+				b.CreateCall(exitGoroutineType, exitGoroutine, []llvm.Value{
 					llvm.Undef(c.dataPtrType),
 				}, "")
 			}
@@ -528,14 +528,13 @@ func (c *compilerContext) createGoroutineStartWrapper(fnType llvm.Type, fn llvm.
 		b.CreateCall(fnType, fnPtr, params, "")
 
 		if c.Scheduler == "asyncify" {
-			b.CreateCall(deadlockType, deadlock, []llvm.Value{
+			b.CreateCall(exitGoroutineType, exitGoroutine, []llvm.Value{
 				llvm.Undef(c.dataPtrType),
 			}, "")
 		}
 	}
 
 	if c.Scheduler == "asyncify" {
-		// The goroutine was terminated via deadlock.
 		b.CreateUnreachable()
 	} else {
 		// Finish the function. Every basic block must end in a terminator, and

compiler/testdata/goroutine-wasm-asyncify.ll

@@ -22,14 +22,14 @@ entry:
 
 declare void @main.regularFunction(i32, ptr) #0
 
-declare void @runtime.deadlock(ptr) #0
+declare void @runtime.exitGoroutine(ptr) #0
 
 ; Function Attrs: nounwind
 define linkonce_odr void @"main.regularFunction$gowrapper"(ptr %0) unnamed_addr #2 {
 entry:
   %unpack.int = ptrtoint ptr %0 to i32
   call void @main.regularFunction(i32 %unpack.int, ptr undef) #11
-  call void @runtime.deadlock(ptr undef) #11
+  call void @runtime.exitGoroutine(ptr undef) #11
   unreachable
 }
 
@@ -53,7 +53,7 @@ define linkonce_odr void @"main.inlineFunctionGoroutine$1$gowrapper"(ptr %0) unn
 entry:
   %unpack.int = ptrtoint ptr %0 to i32
   call void @"main.inlineFunctionGoroutine$1"(i32 %unpack.int, ptr undef)
-  call void @runtime.deadlock(ptr undef) #11
+  call void @runtime.exitGoroutine(ptr undef) #11
   unreachable
 }
 
@@ -96,7 +96,7 @@ entry:
   %2 = getelementptr inbounds nuw i8, ptr %0, i32 4
   %3 = load ptr, ptr %2, align 4
   call void @"main.closureFunctionGoroutine$1"(i32 %1, ptr %3)
-  call void @runtime.deadlock(ptr undef) #11
+  call void @runtime.exitGoroutine(ptr undef) #11
   unreachable
 }
 
@@ -130,7 +130,7 @@ entry:
   %4 = getelementptr inbounds nuw i8, ptr %0, i32 8
   %5 = load ptr, ptr %4, align 4
   call void %5(i32 %1, ptr %3) #11
-  call void @runtime.deadlock(ptr undef) #11
+  call void @runtime.exitGoroutine(ptr undef) #11
   unreachable
 }
 
@@ -193,7 +193,7 @@ entry:
   %6 = getelementptr inbounds nuw i8, ptr %0, i32 12
   %7 = load ptr, ptr %6, align 4
   call void @"interface:{Print:func:{basic:string}{}}.Print$invoke"(ptr %1, ptr %3, i32 %5, ptr %7, ptr undef) #11
-  call void @runtime.deadlock(ptr undef) #11
+  call void @runtime.exitGoroutine(ptr undef) #11
   unreachable
 }
 

compiler/testdata/large.ll

@@ -203,13 +203,13 @@ entry:
   ret void
 }
 
-declare void @runtime.deadlock(ptr) #0
+declare void @runtime.exitGoroutine(ptr) #0
 
 ; Function Attrs: nounwind
 define linkonce_odr void @"main.readLargeValue$gowrapper"(ptr %0) unnamed_addr #6 {
 entry:
   %1 = call i8 @main.readLargeValue(ptr %0, ptr undef)
-  call void @runtime.deadlock(ptr undef) #9
+  call void @runtime.exitGoroutine(ptr undef) #9
   unreachable
 }