From 310ef9f40b350957a1a6d6236ce453ff9f73e81d Mon Sep 17 00:00:00 2001
From: Takashi Kokubun <takashikkbn@gmail.com>
Date: Mon, 13 Apr 2020 20:32:59 -0700
Subject: Make vm_call_cfunc_with_frame a fastpath (#3027)

when there's no need to call CALLER_SETUP_ARG and CALLER_REMOVE_EMPTY_KW_SPLAT
(i.e. !rb_splat_or_kwargs_p(ci) && !calling->kw_splat).

Micro benchmark:
```
$ benchmark-driver -v --rbenv 'before;after' benchmark/vm_send_cfunc.yml --repeat-count=4
before: ruby 2.8.0dev (2020-04-13T23:45:05Z master b9d3ceee8f) [x86_64-linux]
after: ruby 2.8.0dev (2020-04-14T00:48:52Z no-splat-fastpath 418d363722) [x86_64-linux]
Calculating -------------------------------------
                         before       after
       vm_send_cfunc    69.585M     88.724M i/s -    100.000M times in 1.437097s 1.127096s

Comparison:
                    vm_send_cfunc
               after:  88723605.2 i/s
              before:  69584737.1 i/s - 1.28x  slower
```

Optcarrot:
```
$ benchmark-driver -v --rbenv 'before;after' benchmark.yml --repeat-count=12 --output=all
before: ruby 2.8.0dev (2020-04-13T23:45:05Z master b9d3ceee8f) [x86_64-linux]
after: ruby 2.8.0dev (2020-04-14T00:48:52Z no-splat-fastpath 418d363722) [x86_64-linux]
Calculating -------------------------------------
                                       before                 after
Optcarrot Lan_Master.nes    50.76119601545175     42.73858236484051 fps
                            50.76388649761503     51.04211379912850
                            50.80930672252514     51.39455790755538
                            50.90236000778749     51.75656936556145
                            51.01744746340430     51.86875277356489
                            51.06495279015112     51.88692482485558
                            51.07785337168974     51.93429603190578
                            51.20163525187862     51.95768145071314
                            51.34671771913112     52.45577266040274
                            51.35918340835583     52.53163888762858
                            51.46641337418146     52.62172484121034
                            51.50835463462257     52.85064021113239
```
---
 vm_insnhelper.c | 9 +++++++++
 1 file changed, 9 insertions(+)

(limited to 'vm_insnhelper.c')

diff --git a/vm_insnhelper.c b/vm_insnhelper.c
index 21b5e93ca3..a4edd9a797 100644
--- a/vm_insnhelper.c
+++ b/vm_insnhelper.c
@@ -1964,6 +1964,13 @@ rb_iseq_only_kwparam_p(const rb_iseq_t *iseq)
            iseq->body->param.flags.has_block == FALSE;
 }
 
+// If true, cc->call needs to include `CALLER_SETUP_ARG` (i.e. can't be skipped in fastpath)
+MJIT_STATIC bool
+rb_splat_or_kwargs_p(const struct rb_callinfo *restrict ci)
+{
+    return IS_ARGS_SPLAT(ci) || IS_ARGS_KW_OR_KW_SPLAT(ci);
+}
+
 
 static inline void
 CALLER_SETUP_ARG(struct rb_control_frame_struct *restrict cfp,
@@ -2508,6 +2515,7 @@ vm_method_cfunc_entry(const rb_callable_method_entry_t *me)
 static VALUE
 vm_call_cfunc_with_frame(rb_execution_context_t *ec, rb_control_frame_t *reg_cfp, struct rb_calling_info *calling, struct rb_call_data *cd)
 {
+    RB_DEBUG_COUNTER_INC(ccf_cfunc_with_frame);
     const struct rb_callinfo *ci = cd->ci;
     const struct rb_callcache *cc = cd->cc;
     VALUE val;
@@ -2555,6 +2563,7 @@ vm_call_cfunc(rb_execution_context_t *ec, rb_control_frame_t *reg_cfp, struct rb
 
     CALLER_SETUP_ARG(reg_cfp, calling, ci);
     CALLER_REMOVE_EMPTY_KW_SPLAT(reg_cfp, calling, ci);
+    CC_SET_FASTPATH(cd->cc, vm_call_cfunc_with_frame, !rb_splat_or_kwargs_p(ci) && !calling->kw_splat);
     return vm_call_cfunc_with_frame(ec, reg_cfp, calling, cd);
 }
 
-- 
cgit v1.2.3