aboutsummaryrefslogtreecommitdiffhomepage
path: root/interp
diff options
context:
space:
mode:
authorAyke van Laethem <[email protected]>2022-12-15 21:24:17 +0100
committerAyke <[email protected]>2023-01-17 19:32:18 +0100
commit655075e5e0e09ae3d9a3247d04343c020d87162e (patch)
tree25f8f2677484aa38e9be0fe0cb6c9ed2385f84d5 /interp
parentf9d0ff3becbe6835d32a146895d797d53a09fc15 (diff)
downloadtinygo-655075e5e0e09ae3d9a3247d04343c020d87162e.tar.gz
tinygo-655075e5e0e09ae3d9a3247d04343c020d87162e.zip
runtime: implement precise GC
This implements the block-based GC as a partially precise GC. This means that for most heap allocations it is known which words contain a pointer and which don't. This should in theory make the GC faster (because it can skip non-pointer object) and have fewer false positives in a GC cycle. It does however use a bit more RAM to store the layout of each object. Right now this GC seems to be slower than the conservative GC, but should be less likely to run out of memory as a result of false positives.
Diffstat (limited to 'interp')
-rw-r--r--interp/memory.go2
1 files changed, 2 insertions, 0 deletions
diff --git a/interp/memory.go b/interp/memory.go
index 88b7783e4..1f9ed99f3 100644
--- a/interp/memory.go
+++ b/interp/memory.go
@@ -1236,6 +1236,8 @@ func (r *runner) getValue(llvmValue llvm.Value) value {
// readObjectLayout reads the object layout as it is stored by the compiler. It
// returns the size in the number of words and the bitmap.
+//
+// For details on this format, see src/runtime/gc_precise.go.
func (r *runner) readObjectLayout(layoutValue value) (uint64, *big.Int) {
pointerSize := layoutValue.len(r)
if checks && uint64(pointerSize) != r.targetData.TypeAllocSize(r.i8ptrType) {