compile: translate iseq in-place

running "ruby -rpp -e 'pp GC.stat'", a reduction in
malloc usage is shown:

before:
	:malloc_increase=>118784,
	:oldmalloc_increase=>1178736,

after:
	:malloc_increase=>99832,
	:oldmalloc_increase=>1031976,

For "ruby -e exit", valgrind reports over 300K reduction in
overall allocations (and unnecessary memory copies).

before:
  total heap usage: 49,622 allocs, 20,492 frees, 8,697,493 bytes allocated
after:
  total heap usage: 48,935 allocs, 19,805 frees, 8,373,773 bytes allocated

(numbers from x86-64)

v2 changes based on ko1 recommendations [ruby-core:64883]:
- squashed in-place direct thread translation to avoid alloc+copy
- renamed rb_iseq_untranslate_threaded_code to rb_iseq_original_iseq,
  cache new iseq->iseq_original field.

* compile.c (rb_iseq_translate_threaded_code): modify in-place w/o copy
  (rb_vm_addr2insn): new function for debug
  (rb_iseq_original_iseq): ditto
  (iseq_set_sequence): assign iseq_encoded directly
  [Feature #10185]

* vm_core (rb_iseq_t): move original ->iseq to bottom

* iseq.c (iseq_free, iseq_free): adjust for new layout
  (rb_iseq_disasm): use original iseq for dump
  (iseq_data_to_ary): ditto
  (rb_iseq_line_trace_each): ditto
  (rb_iseq_build_for_ruby2cext): use iseq_encoded directly

* vm_dump.c (rb_vmdebug_debug_print_pre): use original iseq

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@47508 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

1 files changed, 41 insertions, 6 deletions

diff --git a/compile.c b/compile.c
index 434d385aa8..5b778e0ccf 100644
--- a/compile.c
+++ b/compile.c
@@ -569,21 +569,56 @@ rb_iseq_translate_threaded_code(rb_iseq_t *iseq)
     const void * const *table = rb_vm_get_insns_address_table();
     unsigned int i;
 
-    iseq->iseq_encoded = ALLOC_N(VALUE, iseq->iseq_size);
-    MEMCPY(iseq->iseq_encoded, iseq->iseq, VALUE, iseq->iseq_size);
-
     for (i = 0; i < iseq->iseq_size; /* */ ) {
 	int insn = (int)iseq->iseq_encoded[i];
 	int len = insn_len(insn);
 	iseq->iseq_encoded[i] = (VALUE)table[insn];
 	i += len;
     }
-#else
-    iseq->iseq_encoded = iseq->iseq;
 #endif
     return COMPILE_OK;
 }
 
+#if OPT_DIRECT_THREADED_CODE || OPT_CALL_THREADED_CODE
+static int
+rb_vm_addr2insn(const void *addr) /* cold path */
+{
+    int insn;
+    const void * const *table = rb_vm_get_insns_address_table();
+
+    for (insn = 0; insn < VM_INSTRUCTION_SIZE; insn++) {
+	if (table[insn] == addr)
+	    return insn;
+    }
+    rb_bug("rb_vm_addr2insn: invalid insn address: %p", addr);
+}
+#endif
+
+VALUE *
+rb_iseq_original_iseq(rb_iseq_t *iseq) /* cold path */
+{
+    if (iseq->iseq) return iseq->iseq;
+
+    iseq->iseq = ALLOC_N(VALUE, iseq->iseq_size);
+
+    MEMCPY(iseq->iseq, iseq->iseq_encoded, VALUE, iseq->iseq_size);
+
+#if OPT_DIRECT_THREADED_CODE || OPT_CALL_THREADED_CODE
+    {
+	unsigned int i;
+
+	for (i = 0; i < iseq->iseq_size; /* */ ) {
+	    const void *addr = (const void *)iseq->iseq[i];
+	    int insn = (VALUE)rb_vm_addr2insn(addr);
+
+	    iseq->iseq[i] = insn;
+	    i += insn_len(insn);
+	}
+    }
+#endif
+    return iseq->iseq;
+}
+
 /*********************************************/
 /* definition of data structure for compiler */
 /*********************************************/
@@ -1645,7 +1680,7 @@ iseq_set_sequence(rb_iseq_t *iseq, LINK_ANCHOR *anchor)
     }
 #endif
 
-    iseq->iseq = (void *)generated_iseq;
+    iseq->iseq_encoded = (void *)generated_iseq;
     iseq->iseq_size = pos;
     iseq->stack_max = stack_max;