Run a block through a frame and step, and compile blocks to AArch64 that call the same steps (021)

- run() keeps a block's operand stack and open loops in a Frame; step()
  runs one instruction (branches and loops too, one dispatch), and
  enter_each, enter_loop, leave_loop and iterate are their own functions.
  The executor is the same machine code, 7% fewer instructions.
- The verifier records each instruction's stack height (from its block's
  start), so compiled code knows each slot's place.
- module-jit.c (AArch64 Linux; elsewhere, or PURE_JIT=0, none): a
  module's handlers', views' and folds' blocks compiled when it loads.
  Branches and loops are branches; every other instruction calls the
  executor's own step, so both run the same meaning. A loop that is not
  opened lands where its iteration would have ended; an empty loop
  iterates in place.
- test-module recompiles a module whose code a test changes after load.
- A search (100,000 nodes, 1,000,000 edges, core only) is 150 ms
  compiled, against 122 ms executed: a call of step for every instruction
  costs more than the executor's switch; the hot instructions are written
  as code next.

1 files changed, 5 insertions, 0 deletions

diff --git a/core/test-module.c b/core/test-module.c
index a509349..b7fea70 100644
--- a/core/test-module.c
+++ b/core/test-module.c
@@ -3,6 +3,7 @@
 #define _DARWIN_C_SOURCE /* mkdtemp, which _POSIX_C_SOURCE hides there */
 #endif
 #include "module.h"
+#include "module-run.h"
 #include "module-save.h"
 #include "module-store.h"
 #include "module-text.h"
@@ -3171,6 +3172,8 @@ static void lane_folds(void) {
         }
         planned += module_a.view_plan[VIEWS];
         memset(module_b.view_plan, 0, sizeof(module_b.view_plan));
+        pure_jit_free(module_b.jit); /* its code changed: compiled again (021) */
+        module_b.jit = pure_jit_make(&module_b);
         PureApp a, b;
         CHECK(pure_app_start(&a, &module_a, (PureData){0}) == PURE_OK && pure_app_start(&b, &module_b, (PureData){0}) == PURE_OK);
         pure_app_work(&a, UINT64_MAX); /* counted, never past it */
@@ -3765,6 +3768,8 @@ static void ranges(void) {
           !module_a.code[118].range && module_a.code[118].plan);
     for (uint32_t i = 0; i < module_b.code_count; ++i) module_b.code[i].range = 0;
     memset(module_b.view_range, 0, sizeof(module_b.view_range));
+    pure_jit_free(module_b.jit); /* its code changed: compiled again (021) */
+    module_b.jit = pure_jit_make(&module_b);
     uint32_t seed = 0x7a45e11u;
     unsigned compared = 0, walked = 0;
     for (unsigned round = 0; round < 60; ++round) {