ZJIT: Support invokebuiltin opcodes (#13632)

* `invokebuiltin` * `invokebuiltin_delegate` * `invokebuiltin_delegate_leave` These instructions all call out to a C function, passing EC, self, and some number of arguments. `invokebuiltin` gets the arguments from the stack, whereas the `_delegate` instructions use a subset of the locals. `opt_invokebuiltin_delegate_leave` has a fast path for `leave`, but I'm not sure we need to do anything special for that here (FWIW YJIT appears to treat the two delegate instructions the same).
author: Daniel Colson <[email protected]> 2025-06-18 18:58:34 -0400
committer: GitHub <[email protected]> 2025-06-18 15:58:34 -0700
commit: 3290d3d7f004fdf19311c3f4e40b01e45f14c23c (patch)
tree: 13413ac6c8983b65d80bf0fb9ee967cbef24e38e
parent: 521b2fcba4e96898bfd237c79f17f19530b7a030 (diff)
4 files changed, 126 insertions, 7 deletions
diff --git a/test/ruby/test_zjit.rb b/test/ruby/test_zjit.rb
index 2b171b02b1..e10e9a8742 100644
--- a/test/ruby/test_zjit.rb
+++ b/test/ruby/test_zjit.rb
@@ -76,6 +76,21 @@ class TestZJIT < Test::Unit::TestCase
     }
   end
 
+  def test_invokebuiltin
+    assert_compiles '["."]', %q{
+      def test = Dir.glob(".")
+      test
+    }
+  end
+
+  def test_invokebuiltin_delegate
+    assert_compiles '[[], true]', %q{
+      def test = [].clone(freeze: true)
+      r = test
+      [r, r.frozen?]
+    }
+  end
+
   def test_opt_plus_const
     assert_compiles '3', %q{
       def test = 1 + 2
diff --git a/zjit/src/codegen.rs b/zjit/src/codegen.rs
index b1869f71c0..90c3ce640e 100644
--- a/zjit/src/codegen.rs
+++ b/zjit/src/codegen.rs
@@ -260,6 +260,7 @@ fn gen_insn(cb: &mut CodeBlock, jit: &mut JITState, asm: &mut Assembler, functio
         Insn::IfFalse { val, target } => return gen_if_false(jit, asm, opnd!(val), target),
         Insn::SendWithoutBlock { call_info, cd, state, self_val, args, .. } => gen_send_without_block(jit, asm, call_info, *cd, &function.frame_state(*state), self_val, args)?,
         Insn::SendWithoutBlockDirect { cme, iseq, self_val, args, state, .. } => gen_send_without_block_direct(cb, jit, asm, *cme, *iseq, opnd!(self_val), args, &function.frame_state(*state))?,
+        Insn::InvokeBuiltin { bf, args, state } => gen_invokebuiltin(jit, asm, &function.frame_state(*state), bf, args)?,
         Insn::Return { val } => return Some(gen_return(asm, opnd!(val))?),
         Insn::FixnumAdd { left, right, state } => gen_fixnum_add(jit, asm, opnd!(left), opnd!(right), &function.frame_state(*state))?,
         Insn::FixnumSub { left, right, state } => gen_fixnum_sub(jit, asm, opnd!(left), opnd!(right), &function.frame_state(*state))?,
@@ -311,6 +312,26 @@ fn gen_get_constant_path(asm: &mut Assembler, ic: *const iseq_inline_constant_ca
     val
 }
 
+fn gen_invokebuiltin(jit: &mut JITState, asm: &mut Assembler, state: &FrameState, bf: &rb_builtin_function, args: &Vec<InsnId>) -> Option<lir::Opnd> {
+    // Ensure we have enough room fit ec, self, and arguments
+    // TODO remove this check when we have stack args (we can use Time.new to test it)
+    if bf.argc + 2 > (C_ARG_OPNDS.len() as i32) {
+        return None;
+    }
+
+    gen_save_pc(asm, state);
+
+    let mut cargs = vec![EC];
+    for &arg in args.iter() {
+        let opnd = jit.get_opnd(arg)?;
+        cargs.push(opnd);
+    }
+
+    let val = asm.ccall(bf.func_ptr as *const u8, cargs);
+
+    Some(val)
+}
+
 /// Lowering for [`Insn::CCall`]. This is a low-level raw call that doesn't know
 /// anything about the callee, so handling for e.g. GC safety is dealt with elsewhere.
 fn gen_ccall(jit: &mut JITState, asm: &mut Assembler, cfun: *const u8, args: &[InsnId]) -> Option<lir::Opnd> {
diff --git a/zjit/src/cruby.rs b/zjit/src/cruby.rs
index e0334ed44d..3a1c45ffd3 100644
--- a/zjit/src/cruby.rs
+++ b/zjit/src/cruby.rs
@@ -1038,8 +1038,8 @@ pub mod test_utils {
     }
 
     /// Get the ISeq of a specified method
-    pub fn get_method_iseq(name: &str) -> *const rb_iseq_t {
-        let wrapped_iseq = eval(&format!("RubyVM::InstructionSequence.of(method(:{}))", name));
+    pub fn get_method_iseq(recv: &str, name: &str) -> *const rb_iseq_t {
+        let wrapped_iseq = eval(&format!("RubyVM::InstructionSequence.of({}.method(:{}))", recv, name));
         unsafe { rb_iseqw_to_iseq(wrapped_iseq) }
     }
 
diff --git a/zjit/src/hir.rs b/zjit/src/hir.rs
index fbca1f4418..276e14a639 100644
--- a/zjit/src/hir.rs
+++ b/zjit/src/hir.rs
@@ -9,7 +9,7 @@ use crate::{
 use std::{
     cell::RefCell,
     collections::{HashMap, HashSet, VecDeque},
-    ffi::{c_int, c_void},
+    ffi::{c_int, c_void, CStr},
     mem::{align_of, size_of},
     ptr,
     slice::Iter
@@ -477,6 +477,9 @@ pub enum Insn {
         state: InsnId,
     },
 
+    // Invoke a builtin function
+    InvokeBuiltin { bf: rb_builtin_function, args: Vec<InsnId>, state: InsnId },
+
     /// Control flow instructions
     Return { val: InsnId },
 
@@ -636,6 +639,13 @@ impl<'a> std::fmt::Display for InsnPrinter<'a> {
                 }
                 Ok(())
             }
+            Insn::InvokeBuiltin { bf, args, .. } => {
+                write!(f, "InvokeBuiltin {}", unsafe { CStr::from_ptr(bf.name) }.to_str().unwrap())?;
+                for arg in args {
+                    write!(f, ", {arg}")?;
+                }
+                Ok(())
+            }
             Insn::Return { val } => { write!(f, "Return {val}") }
             Insn::FixnumAdd  { left, right, .. } => { write!(f, "FixnumAdd {left}, {right}") },
             Insn::FixnumSub  { left, right, .. } => { write!(f, "FixnumSub {left}, {right}") },
@@ -1027,6 +1037,7 @@ impl Function {
                 args: args.iter().map(|arg| find!(*arg)).collect(),
                 state: *state,
             },
+            InvokeBuiltin { bf, args, state } => InvokeBuiltin { bf: *bf, args: find_vec!(*args), state: *state },
             ArraySet { array, idx, val } => ArraySet { array: find!(*array), idx: *idx, val: find!(*val) },
             ArrayDup { val , state } => ArrayDup { val: find!(*val), state: *state },
             &HashDup { val , state } => HashDup { val: find!(val), state },
@@ -1123,6 +1134,7 @@ impl Function {
             Insn::SendWithoutBlock { .. } => types::BasicObject,
             Insn::SendWithoutBlockDirect { .. } => types::BasicObject,
             Insn::Send { .. } => types::BasicObject,
+            Insn::InvokeBuiltin { .. } => types::BasicObject,
             Insn::Defined { .. } => types::BasicObject,
             Insn::DefinedIvar { .. } => types::BasicObject,
             Insn::GetConstantPath { .. } => types::BasicObject,
@@ -1727,6 +1739,10 @@ impl Function {
                     worklist.extend(args);
                     worklist.push_back(state);
                 }
+                Insn::InvokeBuiltin { args, state, .. } => {
+                    worklist.extend(args);
+                    worklist.push_back(state)
+                }
                 Insn::CCall { args, .. } => worklist.extend(args),
                 Insn::GetIvar { self_val, state, .. } | Insn::DefinedIvar { self_val, state, .. } => {
                     worklist.push_back(self_val);
@@ -2614,6 +2630,35 @@ pub fn iseq_to_hir(iseq: *const rb_iseq_t) -> Result<Function, ParseError> {
                     let insn_id = fun.push_insn(block, Insn::NewRange { low, high, flag, state: exit_id });
                     state.stack_push(insn_id);
                 }
+                YARVINSN_invokebuiltin => {
+                    let bf: rb_builtin_function = unsafe { *get_arg(pc, 0).as_ptr() };
+
+                    let mut args = vec![];
+                    for _ in 0..bf.argc {
+                        args.push(state.stack_pop()?);
+                    }
+                    args.push(self_param);
+                    args.reverse();
+
+                    let exit_id = fun.push_insn(block, Insn::Snapshot { state: exit_state });
+                    let insn_id = fun.push_insn(block, Insn::InvokeBuiltin { bf, args, state: exit_id });
+                    state.stack_push(insn_id);
+                }
+                YARVINSN_opt_invokebuiltin_delegate |
+                YARVINSN_opt_invokebuiltin_delegate_leave => {
+                    let bf: rb_builtin_function = unsafe { *get_arg(pc, 0).as_ptr() };
+                    let index = get_arg(pc, 1).as_usize();
+                    let argc = bf.argc as usize;
+
+                    let mut args = vec![self_param];
+                    for &local in state.locals().skip(index).take(argc) {
+                        args.push(local);
+                    }
+
+                    let exit_id = fun.push_insn(block, Insn::Snapshot { state: exit_state });
+                    let insn_id = fun.push_insn(block, Insn::InvokeBuiltin { bf, args, state: exit_id });
+                    state.stack_push(insn_id);
+                }
                 _ => {
                     // Unknown opcode; side-exit into the interpreter
                     let exit_id = fun.push_insn(block, Insn::Snapshot { state: exit_state });
@@ -2907,7 +2952,7 @@ mod tests {
 
     #[track_caller]
     fn assert_method_hir(method: &str, hir: Expect) {
-        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq(method));
+        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq("self", method));
         unsafe { crate::cruby::rb_zjit_profile_disable(iseq) };
         let function = iseq_to_hir(iseq).unwrap();
         assert_function_hir(function, hir);
@@ -2934,7 +2979,7 @@ mod tests {
 
     #[track_caller]
     fn assert_method_hir_with_opcodes(method: &str, opcodes: &[u32], hir: Expect) {
-        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq(method));
+        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq("self", method));
         for &opcode in opcodes {
             assert!(iseq_contains_opcode(iseq, opcode), "iseq {method} does not contain {}", insn_name(opcode as usize));
         }
@@ -2956,7 +3001,7 @@ mod tests {
 
     #[track_caller]
     fn assert_compile_fails(method: &str, reason: ParseError) {
-        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq(method));
+        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq("self", method));
         unsafe { crate::cruby::rb_zjit_profile_disable(iseq) };
         let result = iseq_to_hir(iseq);
         assert!(result.is_err(), "Expected an error but succesfully compiled to HIR: {}", FunctionPrinter::without_snapshot(&result.unwrap()));
@@ -4180,6 +4225,44 @@ mod tests {
               Return v10
         "#]]);
     }
+
+    #[test]
+    fn test_invokebuiltin_delegate_with_args() {
+        assert_method_hir_with_opcode("Float", YARVINSN_opt_invokebuiltin_delegate_leave, expect![[r#"
+            fn Float:
+            bb0(v0:BasicObject, v1:BasicObject, v2:BasicObject, v3:BasicObject):
+              v6:BasicObject = InvokeBuiltin rb_f_float, v0, v1, v2
+              Jump bb1(v0, v1, v2, v3, v6)
+            bb1(v8:BasicObject, v9:BasicObject, v10:BasicObject, v11:BasicObject, v12:BasicObject):
+              Return v12
+        "#]]);
+    }
+
+    #[test]
+    fn test_invokebuiltin_delegate_without_args() {
+        assert_method_hir_with_opcode("class", YARVINSN_opt_invokebuiltin_delegate_leave, expect![[r#"
+            fn class:
+            bb0(v0:BasicObject):
+              v3:BasicObject = InvokeBuiltin _bi20, v0
+              Jump bb1(v0, v3)
+            bb1(v5:BasicObject, v6:BasicObject):
+              Return v6
+        "#]]);
+    }
+
+    #[test]
+    fn test_invokebuiltin_with_args() {
+        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq("GC", "start"));
+        assert!(iseq_contains_opcode(iseq, YARVINSN_invokebuiltin), "iseq GC.start does not contain invokebuiltin");
+        let function = iseq_to_hir(iseq).unwrap();
+        assert_function_hir(function, expect![[r#"
+            fn start:
+            bb0(v0:BasicObject, v1:BasicObject, v2:BasicObject, v3:BasicObject, v4:BasicObject):
+              v6:FalseClassExact = Const Value(false)
+              v8:BasicObject = InvokeBuiltin gc_start_internal, v0, v1, v2, v3, v6
+              Return v8
+        "#]]);
+    }
 }
 
 #[cfg(test)]
@@ -4190,7 +4273,7 @@ mod opt_tests {
 
     #[track_caller]
     fn assert_optimized_method_hir(method: &str, hir: Expect) {
-        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq(method));
+        let iseq = crate::cruby::with_rubyvm(|| get_method_iseq("self", method));
         unsafe { crate::cruby::rb_zjit_profile_disable(iseq) };
         let mut function = iseq_to_hir(iseq).unwrap();
         function.optimize();
author	Daniel Colson <[email protected]>	2025-06-18 18:58:34 -0400
committer	GitHub <[email protected]>	2025-06-18 15:58:34 -0700
commit	3290d3d7f004fdf19311c3f4e40b01e45f14c23c (patch)
tree	13413ac6c8983b65d80bf0fb9ee967cbef24e38e
parent	521b2fcba4e96898bfd237c79f17f19530b7a030 (diff)