x86-64 jit: inline new_k, recycle_leaf and own fast paths, in o3 regions and plain blocks

a call to rt.new_k / rt.recycle_leaf with a constant class, or to rt.own,
does its class-list pop or push (or owner store) inline and goes on after
the call; an empty list, memdebug, a held range, a block outside memory, a
translated word or a type/class mismatch takes the ordinary call. in a
region this replaces the o3_invoke round trip. o3 frees of leaf objects now
use recycle_leaf too (new_k stamps the type inside o3 since b86db60).
ZYNOFASTALLOC=1 turns it off; ZYPERFMAP=1 writes a perf map of translated code.
tests/programs/o3_freelist.zc: reuse, zeroing, an empty list, untyped
neighbours, memdebug; jit vs ZYNOFASTALLOC vs ZYO3INTERP.

1 files changed, 5 insertions, 0 deletions

diff --git a/src/core/stack.h b/src/core/stack.h
index 8c7f7a2..dbc718a 100644
--- a/src/core/stack.h
+++ b/src/core/stack.h
@@ -15,6 +15,11 @@ enum {
 	SC_RETURN_LOW, SC_RETURN_TOP, SC_RETURN_NODE,
 	SC_FREE, SC_STOP, SC_BUMP, SC_MAX, SC_HOLD_LOW, SC_HOLD_HIGH, SC_CUT, SC_GC_ACTIVE, SC_COUNT
 };
+/* optional words after the registered ones, for the native allocator fast paths
+   in compiled o3 regions: a tag (a runtime without them never has it there), the
+   address of rt.dbg, and the entry addresses of rt.new_k, rt.recycle_leaf and rt.own */
+enum { SC_FAST = SC_COUNT, SC_DBG, SC_NEW_K, SC_RECYCLE_LEAF, SC_OWN, SC_FAST_END };
+#define SC_FAST_TAG 0x2B7E1517
 
 /* the including core supplies stack_memory_grow and refreshes m if it moves. */
 static Cell stack_get(int slot) { Cell p = m[stack_control + slot]; return inmem(p) ? m[p] : 0; }