diff options
author | Alan Wu <XrXr@users.noreply.github.com> | 2021-09-24 16:57:53 -0400 |
---|---|---|
committer | Alan Wu <XrXr@users.noreply.github.com> | 2021-10-20 18:19:41 -0400 |
commit | a09adac2d76342fd4a0d97905b3397758689c817 (patch) | |
tree | b36da7570712769a9c129c95df71761ab0d59104 | |
parent | 79015926eb81ac93700146d42c343f568d3ddfbb (diff) |
Add specialization for String#to_s on plain strings
When calling "to_s" on an instance of String, the method simply returns
self. In this situation most of the work comes from setting up the
method call. It turns out that both railsbench and liquid-render do this
a lot.
When generating code for opt_send_without_block, we already generate a
known class guard, so we can detect when the receiver is a String
instance. Since gen_send_cfunc() is also used for gen_invokesuper(), and
gen_invokesuper() doesn't generate a known class guard, a new nullable
parameter for specialized codegen function is added.
Closes GH-245
-rw-r--r-- | bootstraptest/test_yjit.rb | 35 | ||||
-rw-r--r-- | yjit_codegen.c | 38 |
2 files changed, 64 insertions, 9 deletions
diff --git a/bootstraptest/test_yjit.rb b/bootstraptest/test_yjit.rb index 73b87e17fd..91c39323e0 100644 --- a/bootstraptest/test_yjit.rb +++ b/bootstraptest/test_yjit.rb @@ -2038,3 +2038,38 @@ assert_equal '42', %q{ Ractor.new { A.foo }.take } + +assert_equal '["plain", "special", "sub", "plain"]', %q{ + def foo(arg) + arg.to_s + end + + class Sub < String + end + + special = String.new("special") + special.singleton_class + + [ + foo("plain"), + foo(special), + foo(Sub.new("sub")), + foo("plain") + ] +} + +assert_equal '["sub", "sub"]', %q{ + def foo(arg) + arg.to_s + end + + class Sub < String + def to_s + super + end + end + + sub = Sub.new("sub") + + [foo(sub), foo(sub)] +} diff --git a/yjit_codegen.c b/yjit_codegen.c index c05ec53a38..0669bdee91 100644 --- a/yjit_codegen.c +++ b/yjit_codegen.c @@ -2954,13 +2954,15 @@ jit_protected_callee_ancestry_guard(jitstate_t *jit, codeblock_t *cb, const rb_c } // Return true when the codegen function generates code. -typedef bool (*method_codegen_t)(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc); +// known_recv_klass is non-NULL when the caller has used jit_guard_known_klass +// See yjit_reg_method. +typedef bool (*method_codegen_t)(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *known_recv_klass); // Codegen for rb_obj_not(). // Note, caller is responsible for generating all the right guards, including // arity guards. static bool -jit_rb_obj_not(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc) +jit_rb_obj_not(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *known_recv_klass) { const val_type_t recv_opnd = ctx_get_opnd_type(ctx, OPND_STACK(0)); @@ -2988,7 +2990,7 @@ jit_rb_obj_not(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const // Codegen for rb_true() static bool -jit_rb_true(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc) +jit_rb_true(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *known_recv_klass) { ADD_COMMENT(cb, "nil? == true"); ctx_stack_pop(ctx, 1); @@ -2999,7 +3001,7 @@ jit_rb_true(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_ // Codegen for rb_false() static bool -jit_rb_false(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc) +jit_rb_false(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *known_recv_klass) { ADD_COMMENT(cb, "nil? == false"); ctx_stack_pop(ctx, 1); @@ -3011,7 +3013,7 @@ jit_rb_false(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb // Codegen for rb_obj_equal() // object identity comparison static bool -jit_rb_obj_equal(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc) +jit_rb_obj_equal(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *known_recv_klass) { ADD_COMMENT(cb, "equal?"); x86opnd_t obj1 = ctx_stack_pop(ctx, 1); @@ -3028,6 +3030,22 @@ jit_rb_obj_equal(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, cons return true; } +// Codegen for rb_str_to_s() +// When String#to_s is called on a String instance, the method returns self and most +// of the overhead comes from setting up the method call. We observed that this situation +// happens a lot in some workloads. +static bool +jit_rb_str_to_s(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *recv_known_klass) +{ + if (recv_known_klass && *recv_known_klass == rb_cString) { + ADD_COMMENT(cb, "to_s on plain string"); + // The method returns the receiver, which is already on the stack. + // No stack movement. + return true; + } + return false; +} + // Check if we know how to codegen for a particular cfunc method static method_codegen_t lookup_cfunc_codegen(const rb_method_definition_t *def) @@ -3040,7 +3058,7 @@ lookup_cfunc_codegen(const rb_method_definition_t *def) } static codegen_status_t -gen_send_cfunc(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc) +gen_send_cfunc(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const rb_callable_method_entry_t *cme, rb_iseq_t *block, const int32_t argc, VALUE *recv_known_klass) { const rb_method_cfunc_t *cfunc = UNALIGNED_MEMBER_PTR(cme->def, body.cfunc); @@ -3085,7 +3103,7 @@ gen_send_cfunc(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const { method_codegen_t known_cfunc_codegen; if ((known_cfunc_codegen = lookup_cfunc_codegen(cme->def))) { - if (known_cfunc_codegen(jit, ctx, ci, cme, block, argc)) { + if (known_cfunc_codegen(jit, ctx, ci, cme, block, argc, recv_known_klass)) { // cfunc codegen generated code. Terminate the block so // there isn't multiple calls in the same block. jit_jump_to_next_insn(jit, ctx); @@ -3644,7 +3662,7 @@ gen_send_general(jitstate_t *jit, ctx_t *ctx, struct rb_call_data *cd, rb_iseq_t case VM_METHOD_TYPE_ISEQ: return gen_send_iseq(jit, ctx, ci, cme, block, argc); case VM_METHOD_TYPE_CFUNC: - return gen_send_cfunc(jit, ctx, ci, cme, block, argc); + return gen_send_cfunc(jit, ctx, ci, cme, block, argc, &comptime_recv_klass); case VM_METHOD_TYPE_IVAR: if (argc != 0) { // Argument count mismatch. Getters take no arguments. @@ -3842,7 +3860,7 @@ gen_invokesuper(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb) case VM_METHOD_TYPE_ISEQ: return gen_send_iseq(jit, ctx, ci, cme, block, argc); case VM_METHOD_TYPE_CFUNC: - return gen_send_cfunc(jit, ctx, ci, cme, block, argc); + return gen_send_cfunc(jit, ctx, ci, cme, block, argc, NULL); default: break; } @@ -4480,4 +4498,6 @@ yjit_init_codegen(void) yjit_reg_method(rb_cModule, "==", jit_rb_obj_equal); yjit_reg_method(rb_cSymbol, "==", jit_rb_obj_equal); yjit_reg_method(rb_cSymbol, "===", jit_rb_obj_equal); + + yjit_reg_method(rb_cString, "to_s", jit_rb_str_to_s); } |