Export stats primitives from Rust to CRuby (ruby#196)

maximecb · XrXr · web-flow · commit ec62a54933ea · 2022-02-26T12:15:03.000-05:00
* Export stats primitives from Rust to CRuby

* Run yjit.rb during init again and move primitive functions from yjit.h

I removed code to run yjit.rb during init from `inits.c` earlier as an
expedient way to get builds working. It's led to confusion about why
RubyVM::YJIT isn't showing up. Sorry!

As for the functions added to yjit.h, I moved them to yjit_iface.c
because I think we want to keep yjit.h a shortlist of declaration for
code external to YJIT. The primitive functions only need to be visible
before the `yjit.rbinc` include.

I hope you don't mind me pushing to your branch. Feel free to tell me to
stop it.

* preemptive diff minimization

* Remove unnecessary stubs

* Try to fix !YJIT_BUILD builds

too lazy to spin up a separate build config locally so sorry if this
fails CI again 😅

* Fix again

* Stub Init_builtin_yjit() in vm.o rather than in header

Co-authored-by: Alan Wu &lt;XrXr@users.noreply.github.com&gt;
diff --git a/inits.c b/inits.c
@@ -98,6 +98,7 @@ rb_call_builtin_inits(void)
     BUILTIN(array);
     BUILTIN(kernel);
     BUILTIN(timev);
+    BUILTIN(yjit);
     BUILTIN(nilclass);
     BUILTIN(marshal);
     Init_builtin_prelude();
diff --git a/vm.c b/vm.c
@@ -3961,6 +3961,11 @@ Init_vm_objects(void)
     vm->frozen_strings = st_init_table_with_size(&rb_fstring_hash_type, 10000);
 }
 
+/* Stub for builtin function when not building YJIT units*/
+#if !YJIT_BUILD
+void Init_builtin_yjit(void) {}
+#endif
+
 /* top self */
 
 static VALUE
diff --git a/yjit.c b/yjit.c
@@ -379,7 +379,7 @@ rb_get_iseq_body_param_opt_num(rb_iseq_t* iseq) {
     return iseq->body->param.opt_num;
 }
 
-VALUE*
+const VALUE*
 rb_get_iseq_body_param_opt_table(rb_iseq_t* iseq) {
     return iseq->body->param.opt_table;
 }
diff --git a/yjit.h b/yjit.h
@@ -44,15 +44,14 @@
 // Expose these as declarations since we are building YJIT.
 bool rb_yjit_enabled_p(void);
 unsigned rb_yjit_call_threshold(void);
-
 void rb_yjit_invalidate_all_method_lookup_assumptions(void);
 void rb_yjit_method_lookup_change(VALUE klass, ID mid);
 void rb_yjit_cme_invalidate(VALUE cme);
 void rb_yjit_collect_vm_usage_insn(int insn);
 void rb_yjit_collect_binding_alloc(void);
 void rb_yjit_collect_binding_set(void);
 bool rb_yjit_compile_iseq(const rb_iseq_t *iseq, rb_execution_context_t *ec);
-void rb_yjit_init();
+void rb_yjit_init(void);
 void rb_yjit_bop_redefined(VALUE klass, const rb_method_entry_t *me, enum ruby_basic_operators bop);
 void rb_yjit_constant_state_changed(void);
 void rb_yjit_iseq_mark(const struct rb_iseq_constant_body *body);
diff --git a/yjit.rb b/yjit.rb
@@ -134,18 +134,16 @@ def self.disasm_block(cs, block, highlight)
   # Return a hash for statistics generated for the --yjit-stats command line option.
   # Return nil when option is not passed or unavailable.
   def self.runtime_stats
-    # defined in yjit_iface.c
-    Primitive.get_yjit_stats
+    Primitive.rb_yjit_get_stats
   end
 
   # Discard statistics collected for --yjit-stats.
   def self.reset_stats!
-    # defined in yjit_iface.c
-    Primitive.reset_stats_bang
+    Primitive.rb_yjit_reset_stats_bang
   end
 
   def self.stats_enabled?
-    Primitive.yjit_stats_enabled_p
+    Primitive.rb_yjit_stats_enabled_p
   end
 
   def self.enabled?
@@ -157,7 +155,7 @@ def self.simulate_oom!
   end
 
   # Avoid calling a method here to not interfere with compilation tests
-  if Primitive.yjit_stats_enabled_p
+  if Primitive.rb_yjit_stats_enabled_p
     at_exit { _print_stats }
   end
 
diff --git a/yjit/src/stats.rs b/yjit/src/stats.rs
@@ -5,9 +5,6 @@ use crate::cruby::*;
 use crate::options::*;
 use crate::yjit::{yjit_enabled_p};
 
-// TODO
-//extern const int rb_vm_max_insn_name_size;
-
 // YJIT exit counts for each instruction type
 static mut EXIT_OP_COUNT: [u64; VM_INSTRUCTION_SIZE] = [0; VM_INSTRUCTION_SIZE];
 
@@ -139,10 +136,24 @@ make_counters!(
 
 //===========================================================================
 
-/// Primitive called in yjit.rb. Export all YJIT statistics as a Ruby hash.
-/// This should be wrapped on the C side with RB_VM_LOCK_ENTER()
+/// Primitive called in yjit.rb
+/// Check if stats generation is enabled
 #[no_mangle]
-pub extern "C" fn rb_yjit_get_yjit_stats(ec: EcPtr, ruby_self: VALUE) -> VALUE {
+pub extern "C" fn rb_yjit_stats_enabled_p(ec: EcPtr, ruby_self: VALUE) -> VALUE {
+    #[cfg(feature = "stats")]
+    if get_option!(gen_stats) {
+        return Qtrue
+    }
+
+    return Qfalse;
+}
+
+/// Primitive called in yjit.rb.
+/// Export all YJIT statistics as a Ruby hash.
+/// This needs to be wrapped on the C side with RB_VM_LOCK_ENTER()
+#[no_mangle]
+pub extern "C" fn rb_yjit_gen_stats_dict(ec: EcPtr, ruby_self: VALUE) -> VALUE {
+
     // Return Qnil if YJIT isn't enabled
     if !yjit_enabled_p() {
         return Qnil;
@@ -151,8 +162,6 @@ pub extern "C" fn rb_yjit_get_yjit_stats(ec: EcPtr, ruby_self: VALUE) -> VALUE {
     unsafe {
         let hash = rb_hash_new();
 
-        //RB_VM_LOCK_ENTER();
-
         /*
         {
             VALUE key = ID2SYM(rb_intern("inline_code_size"));
@@ -221,8 +230,6 @@ pub extern "C" fn rb_yjit_get_yjit_stats(ec: EcPtr, ruby_self: VALUE) -> VALUE {
             */
         }
 
-        //RB_VM_LOCK_LEAVE();
-
         return hash;
     }
 }
diff --git a/yjit_iface.c b/yjit_iface.c
@@ -546,108 +546,18 @@ comments_for(rb_execution_context_t *ec, VALUE self, VALUE start_address, VALUE
     return comment_array;
 }
 
-static VALUE
-yjit_stats_enabled_p(rb_execution_context_t *ec, VALUE self)
-{
-    // FIXME: moved to Rust
-    return Qfalse;
-}
-
-// Primitive called in yjit.rb. Export all YJIT statistics as a Ruby hash.
-static VALUE
-get_yjit_stats(rb_execution_context_t *ec, VALUE self)
+// Primitive called in yjit.rb.
+// Export all YJIT statistics as a Ruby hash.
+VALUE
+rb_yjit_get_stats(rb_execution_context_t *ec, VALUE self)
 {
-    // TODO: take VM lock and call into Rust code
-    return Qnil;
-
-    /*
-    // Return Qnil if YJIT isn't enabled
-    if (cb == NULL) {
-        return Qnil;
-    }
-
-    VALUE hash = rb_hash_new();
-
+    VALUE stats_hash;
     RB_VM_LOCK_ENTER();
-
-    {
-        VALUE key = ID2SYM(rb_intern("inline_code_size"));
-        VALUE value = LL2NUM((long long)cb->write_pos);
-        rb_hash_aset(hash, key, value);
-
-        key = ID2SYM(rb_intern("outlined_code_size"));
-        value = LL2NUM((long long)ocb->write_pos);
-        rb_hash_aset(hash, key, value);
-    }
-
-#if YJIT_STATS
-    if (rb_yjit_opts.gen_stats) {
-        // Indicate that the complete set of stats is available
-        rb_hash_aset(hash, ID2SYM(rb_intern("all_stats")), Qtrue);
-
-        int64_t *counter_reader = (int64_t *)&yjit_runtime_counters;
-        int64_t *counter_reader_end = &yjit_runtime_counters.last_member;
-
-        // For each counter in yjit_counter_names, add that counter as
-        // a key/value pair.
-
-        // Iterate through comma separated counter name list
-        char *name_reader = yjit_counter_names;
-        char *counter_name_end = yjit_counter_names + sizeof(yjit_counter_names);
-        while (name_reader < counter_name_end && counter_reader < counter_reader_end) {
-            if (*name_reader == ',' || *name_reader == ' ') {
-                name_reader++;
-                continue;
-            }
-
-            // Compute length of counter name
-            int name_len;
-            char *name_end;
-            {
-                name_end = strchr(name_reader, ',');
-                if (name_end == NULL) break;
-                name_len = (int)(name_end - name_reader);
-            }
-
-            // Put counter into hash
-            VALUE key = ID2SYM(rb_intern2(name_reader, name_len));
-            VALUE value = LL2NUM((long long)*counter_reader);
-            rb_hash_aset(hash, key, value);
-
-            counter_reader++;
-            name_reader = name_end;
-        }
-
-        // For each entry in exit_op_count, add a stats entry with key "exit_INSTRUCTION_NAME"
-        // and the value is the count of side exits for that instruction.
-
-        char key_string[rb_vm_max_insn_name_size + 6]; // Leave room for "exit_" and a final NUL
-        for (int i = 0; i < VM_INSTRUCTION_SIZE; i++) {
-            const char *i_name = insn_name(i); // Look up Ruby's NUL-terminated insn name string
-            snprintf(key_string, rb_vm_max_insn_name_size + 6, "%s%s", "exit_", i_name);
-
-            VALUE key = ID2SYM(rb_intern(key_string));
-            VALUE value = LL2NUM((long long)exit_op_count[i]);
-            rb_hash_aset(hash, key, value);
-        }
-    }
-#endif
-
+    VALUE rb_yjit_gen_stats_dict(rb_execution_context_t *ec, VALUE self);
+    stats_hash = rb_yjit_gen_stats_dict(ec, self);
     RB_VM_LOCK_LEAVE();
 
-    return hash;
-    */
-}
-
-// Primitive called in yjit.rb. Zero out all the counters.
-static VALUE
-reset_stats_bang(rb_execution_context_t *ec, VALUE self)
-{
-#if YJIT_STATS
-    memset(&exit_op_count, 0, sizeof(exit_op_count));
-    memset(&yjit_runtime_counters, 0, sizeof(yjit_runtime_counters));
-#endif // if YJIT_STATS
-    return Qnil;
+    return stats_hash;
 }
 
 // Primitive for yjit.rb. For testing running out of executable memory
@@ -662,6 +572,12 @@ simulate_oom_bang(rb_execution_context_t *ec, VALUE self)
     return Qnil;
 }
 
+// Primitives used by yjit.rb
+VALUE rb_yjit_stats_enabled_p(rb_execution_context_t *ec, VALUE self);
+VALUE rb_yjit_get_stats(rb_execution_context_t *ec, VALUE self);
+VALUE rb_yjit_reset_stats_bang(rb_execution_context_t *ec, VALUE self);
+
+// Preprocessed yjit.rb generated during build
 #include "yjit.rbinc"
 
 #if YJIT_STATS
@@ -849,7 +765,7 @@ outgoing_ids(VALUE self)
 
 // Can raise RuntimeError
 void
-rb_yjit_init()
+rb_yjit_init(void)
 {
     if (!PLATFORM_SUPPORTED_P || !JIT_ENABLED) {
         return;

Original file line number	Diff line number	Diff line change
`@@ -379,7 +379,7 @@ rb_get_iseq_body_param_opt_num(rb_iseq_t* iseq) {`
`379`	`379`	`return iseq->body->param.opt_num;`
`380`	`380`	`}`
`381`	`381`
`382`		`-VALUE*`
	`382`	`+const VALUE*`
`383`	`383`	`rb_get_iseq_body_param_opt_table(rb_iseq_t* iseq) {`
`384`	`384`	`return iseq->body->param.opt_table;`
`385`	`385`	`}`