Cargo.lock

@@ -117,9 +117,9 @@ checksum = "1fd0f2584146f6f2ef48085050886acf353beff7305ebd1ae69500e27c67f64b"
 
 [[package]]
 name = "cc"
-version = "1.1.15"
+version = "1.1.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "57b6a275aa2903740dc87da01c62040406b8812552e97129a63ea8850a17c6e6"
+checksum = "72db2f7947ecee9b03b510377e8bb9077afa27176fdbff55c51027e976fdcc48"
 dependencies = [
  "shlex",
 ]
@@ -161,7 +161,7 @@ dependencies = [
  "heck 0.5.0",
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -310,9 +310,9 @@ dependencies = [
 
 [[package]]
 name = "fastrand"
-version = "2.1.1"
+version = "2.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e8c02a5121d4ea3eb16a80748c74f5549a5665e4c21333c6098f283870fbdea6"
+checksum = "9fc0510504f03c51ada170672ac806f1f105a88aa97a5281117e1ddc3368e51a"
 
 [[package]]
 name = "fixedbitset"
@@ -424,7 +424,7 @@ checksum = "4fa4d8d74483041a882adaa9a29f633253a66dde85055f0495c121620ac484b2"
 dependencies = [
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -510,9 +510,9 @@ checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe"
 
 [[package]]
 name = "libc"
-version = "0.2.158"
+version = "0.2.157"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d8adc4bb1803a324070e64a98ae98f38934d91957a99cfb3a43dcbc01bc56439"
+checksum = "374af5f94e54fa97cf75e945cce8a6b201e88a1a07e688b47dfd2a59c66dbd86"
 
 [[package]]
 name = "libloading"
@@ -752,7 +752,7 @@ dependencies = [
  "phf_shared 0.11.2",
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -856,7 +856,7 @@ dependencies = [
  "proc-macro2",
  "pyo3-macros-backend",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -869,14 +869,14 @@ dependencies = [
  "proc-macro2",
  "pyo3-build-config",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
 name = "quote"
-version = "1.0.37"
+version = "1.0.36"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b5b9d34b8991d19d98081b46eacdd8eb58c6f2b201139f7c5f643cc155a633af"
+checksum = "0fa76aaf39101c457836aec0ce2316dbdc3ab723cdda1c6bd4e6ad4208acaca7"
 dependencies = [
  "proc-macro2",
 ]
@@ -942,9 +942,9 @@ dependencies = [
 
 [[package]]
 name = "redox_users"
-version = "0.4.6"
+version = "0.4.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ba009ff324d1fc1b900bd1fdb31564febe58a8ccc8a6fdbb93b543d33b13ca43"
+checksum = "bd283d9651eeda4b2a83a43c1c91b266c40fd76ecd39a50a8c630ae69dc72891"
 dependencies = [
  "getrandom",
  "libredox",
@@ -989,9 +989,9 @@ dependencies = [
 
 [[package]]
 name = "rustix"
-version = "0.38.35"
+version = "0.38.34"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a85d50532239da68e9addb745ba38ff4612a242c1c7ceea689c4bc7c2f43c36f"
+checksum = "70dc5ec042f7a43c4a73241207cecc9873a06d45debb38b329f8541d85c2730f"
 dependencies = [
  "bitflags",
  "errno",
@@ -1035,29 +1035,29 @@ checksum = "61697e0a1c7e512e84a621326239844a24d8207b4669b41bc18b32ea5cbf988b"
 
 [[package]]
 name = "serde"
-version = "1.0.209"
+version = "1.0.208"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "99fce0ffe7310761ca6bf9faf5115afbc19688edd00171d81b1bb1b116c63e09"
+checksum = "cff085d2cb684faa248efb494c39b68e522822ac0de72ccf08109abde717cfb2"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.209"
+version = "1.0.208"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a5831b979fd7b5439637af1752d535ff49f4860c0f341d1baeb6faf0f4242170"
+checksum = "24008e81ff7613ed8e5ba0cfaf24e2c2f1e5b8a0495711e44fcd4882fca62bcf"
 dependencies = [
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
 name = "serde_json"
-version = "1.0.127"
+version = "1.0.125"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8043c06d9f82bd7271361ed64f415fe5e12a77fdb52e573e7f06a516dea329ad"
+checksum = "83c8e735a073ccf5be70aa8066aa984eaf2fa000db6c8d0100ae605b366d31ed"
 dependencies = [
  "itoa",
  "memchr",
@@ -1147,7 +1147,7 @@ dependencies = [
  "proc-macro2",
  "quote",
  "rustversion",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -1163,9 +1163,9 @@ dependencies = [
 
 [[package]]
 name = "syn"
-version = "2.0.76"
+version = "2.0.75"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "578e081a14e0cefc3279b0472138c513f37b41a08d5a3cca9b6e4e8ceb6cd525"
+checksum = "f6af063034fc1935ede7be0122941bafa9bacb949334d090b77ca98b5817c7d9"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -1232,7 +1232,7 @@ checksum = "a4558b58466b9ad7ca0f102865eccc95938dca1a74a856f2b57b6629050da261"
 dependencies = [
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]
 
 [[package]]
@@ -1310,9 +1310,9 @@ checksum = "0336d538f7abc86d282a4189614dfaa90810dfc2c6f6427eaf88e16311dd225d"
 
 [[package]]
 name = "unicode-xid"
-version = "0.2.5"
+version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "229730647fbc343e3a80e463c1db7f78f3855d3f3739bee0dda773c9a037c90a"
+checksum = "f962df74c8c05a667b5ee8bcf162993134c104e96440b663c8daa176dc772d8c"
 
 [[package]]
 name = "unicode_names2"
@@ -1510,5 +1510,5 @@ checksum = "fa4f8080344d4671fb4e831a13ad1e68092748387dfc4f55e356242fae12ce3e"
 dependencies = [
  "proc-macro2",
  "quote",
- "syn 2.0.76",
+ "syn 2.0.75",
 ]

nac3artiq/src/codegen.rs

@@ -2,7 +2,7 @@ use nac3core::{
     codegen::{
         classes::{
             ArrayLikeIndexer, ArrayLikeValue, ArraySliceValue, ListValue, NDArrayType,
-            NDArrayValue, ProxyType, ProxyValue, RangeValue, UntypedArrayLikeAccessor,
+            NDArrayValue, RangeValue, UntypedArrayLikeAccessor,
         },
         expr::{destructure_range, gen_call},
         irrt::call_ndarray_calc_size,
@@ -22,7 +22,7 @@ use inkwell::{
     module::Linkage,
     types::{BasicType, IntType},
     values::{BasicValueEnum, PointerValue, StructValue},
-    AddressSpace, IntPredicate, OptimizationLevel,
+    AddressSpace, IntPredicate,
 };
 
 use pyo3::{
@@ -32,7 +32,6 @@ use pyo3::{
 
 use crate::{symbol_resolver::InnerResolver, timeline::TimeFns};
 
-use inkwell::values::IntValue;
 use itertools::Itertools;
 use std::{
     collections::{hash_map::DefaultHasher, HashMap},
@@ -487,10 +486,13 @@ fn format_rpc_arg<'ctx>(
             let buffer = ctx.builder.build_array_alloca(llvm_i8, buffer_size, "rpc.arg").unwrap();
             let buffer = ArraySliceValue::from_ptr_val(buffer, buffer_size, Some("rpc.arg"));
 
+            let ppdata = generator.gen_var_alloc(ctx, llvm_arg_ty.element_type(), None).unwrap();
+            ctx.builder.build_store(ppdata, llvm_arg.data().base_ptr(ctx, generator)).unwrap();
+
             call_memcpy_generic(
                 ctx,
                 buffer.base_ptr(ctx, generator),
-                llvm_arg.ptr_to_data(ctx),
+                ppdata,
                 llvm_pdata_sizeof,
                 llvm_i1.const_zero(),
             );
@@ -526,298 +528,6 @@ fn format_rpc_arg<'ctx>(
     arg_slot
 }
 
-/// Formats an RPC return value to conform to the expected format required by NAC3.
-fn format_rpc_ret<'ctx>(
-    generator: &mut dyn CodeGenerator,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    ret_ty: Type,
-) -> Option<BasicValueEnum<'ctx>> {
-    // -- receive value:
-    // T result = {
-    //   void *ret_ptr = alloca(sizeof(T));
-    //   void *ptr = ret_ptr;
-    //   loop: int size = rpc_recv(ptr);
-    //   // Non-zero: Provide `size` bytes of extra storage for variable-length data.
-    //   if(size) { ptr = alloca(size); goto loop; }
-    //   else *(T*)ret_ptr
-    // }
-
-    let llvm_i8 = ctx.ctx.i8_type();
-    let llvm_i32 = ctx.ctx.i32_type();
-    let llvm_i8_8 = ctx.ctx.struct_type(&[llvm_i8.array_type(8).into()], false);
-    let llvm_pi8 = llvm_i8.ptr_type(AddressSpace::default());
-
-    let rpc_recv = ctx.module.get_function("rpc_recv").unwrap_or_else(|| {
-        ctx.module.add_function("rpc_recv", llvm_i32.fn_type(&[llvm_pi8.into()], false), None)
-    });
-
-    if ctx.unifier.unioned(ret_ty, ctx.primitives.none) {
-        ctx.build_call_or_invoke(rpc_recv, &[llvm_pi8.const_null().into()], "rpc_recv");
-        return None;
-    }
-
-    let prehead_bb = ctx.builder.get_insert_block().unwrap();
-    let current_function = prehead_bb.get_parent().unwrap();
-    let head_bb = ctx.ctx.append_basic_block(current_function, "rpc.head");
-    let alloc_bb = ctx.ctx.append_basic_block(current_function, "rpc.continue");
-    let tail_bb = ctx.ctx.append_basic_block(current_function, "rpc.tail");
-
-    let llvm_ret_ty = ctx.get_llvm_abi_type(generator, ret_ty);
-
-    let result = match &*ctx.unifier.get_ty_immutable(ret_ty) {
-        TypeEnum::TObj { obj_id, .. } if *obj_id == PrimDef::NDArray.id() => {
-            let llvm_i1 = ctx.ctx.bool_type();
-            let llvm_usize = generator.get_size_type(ctx.ctx);
-
-            // Round `val` up to its modulo `power_of_two`
-            let round_up = |ctx: &mut CodeGenContext<'ctx, '_>,
-                            val: IntValue<'ctx>,
-                            power_of_two: IntValue<'ctx>| {
-                debug_assert_eq!(
-                    val.get_type().get_bit_width(),
-                    power_of_two.get_type().get_bit_width()
-                );
-
-                let llvm_val_t = val.get_type();
-
-                let max_rem = ctx
-                    .builder
-                    .build_int_sub(power_of_two, llvm_val_t.const_int(1, false), "")
-                    .unwrap();
-                ctx.builder
-                    .build_and(
-                        ctx.builder.build_int_add(val, max_rem, "").unwrap(),
-                        ctx.builder.build_not(max_rem, "").unwrap(),
-                        "",
-                    )
-                    .unwrap()
-            };
-
-            // Setup types
-            let (elem_ty, ndims) = unpack_ndarray_var_tys(&mut ctx.unifier, ret_ty);
-            let llvm_elem_ty = ctx.get_llvm_type(generator, elem_ty);
-            let llvm_ret_ty = NDArrayType::new(generator, ctx.ctx, llvm_elem_ty);
-
-            // Allocate the resulting ndarray
-            // A condition after format_rpc_ret ensures this will not be popped this off.
-            let ndarray = llvm_ret_ty.new_value(generator, ctx, Some("rpc.result"));
-
-            // Setup ndims
-            let ndims =
-                if let TypeEnum::TLiteral { values, .. } = &*ctx.unifier.get_ty_immutable(ndims) {
-                    assert_eq!(values.len(), 1);
-
-                    u64::try_from(values[0].clone()).unwrap()
-                } else {
-                    unreachable!();
-                };
-            // Set `ndarray.ndims`
-            ndarray.store_ndims(ctx, generator, llvm_usize.const_int(ndims, false));
-            // Allocate `ndarray.shape` [size_t; ndims]
-            ndarray.create_dim_sizes(ctx, llvm_usize, ndarray.load_ndims(ctx));
-
-            /*
-                ndarray now:
-                  - .ndims: initialized
-                  - .shape: allocated but uninitialized .shape
-                  - .data: uninitialized
-            */
-
-            let llvm_usize_sizeof = ctx
-                .builder
-                .build_int_truncate_or_bit_cast(llvm_usize.size_of(), llvm_usize, "")
-                .unwrap();
-            let llvm_pdata_sizeof = ctx
-                .builder
-                .build_int_truncate_or_bit_cast(
-                    llvm_ret_ty.element_type().size_of().unwrap(),
-                    llvm_usize,
-                    "",
-                )
-                .unwrap();
-            let llvm_elem_sizeof = ctx
-                .builder
-                .build_int_truncate_or_bit_cast(llvm_elem_ty.size_of().unwrap(), llvm_usize, "")
-                .unwrap();
-
-            // Allocates a buffer for the initial RPC'ed object, which is guaranteed to be
-            // (4 + 4 * ndims) bytes with 8-byte alignment
-            let sizeof_dims =
-                ctx.builder.build_int_mul(ndarray.load_ndims(ctx), llvm_usize_sizeof, "").unwrap();
-            let unaligned_buffer_size =
-                ctx.builder.build_int_add(sizeof_dims, llvm_pdata_sizeof, "").unwrap();
-            let buffer_size = round_up(ctx, unaligned_buffer_size, llvm_usize.const_int(8, false));
-
-            let stackptr = call_stacksave(ctx, None);
-            // Just to be absolutely sure, alloca in [i8 x 8] slices to force 8-byte alignment
-            let buffer = ctx
-                .builder
-                .build_array_alloca(
-                    llvm_i8_8,
-                    ctx.builder
-                        .build_int_unsigned_div(buffer_size, llvm_usize.const_int(8, false), "")
-                        .unwrap(),
-                    "rpc.buffer",
-                )
-                .unwrap();
-            let buffer = ctx
-                .builder
-                .build_bitcast(buffer, llvm_pi8, "")
-                .map(BasicValueEnum::into_pointer_value)
-                .unwrap();
-            let buffer = ArraySliceValue::from_ptr_val(buffer, buffer_size, None);
-
-            // The first call to `rpc_recv` reads the top-level ndarray object: [pdata, shape]
-            //
-            // The returned value is the number of bytes for `ndarray.data`.
-            let ndarray_nbytes = ctx
-                .build_call_or_invoke(
-                    rpc_recv,
-                    &[buffer.base_ptr(ctx, generator).into()], // Reads [usize; ndims]. NOTE: We are allocated [size_t; ndims].
-                    "rpc.size.next",
-                )
-                .map(BasicValueEnum::into_int_value)
-                .unwrap();
-
-            // debug_assert(ndarray_nbytes > 0)
-            if ctx.registry.llvm_options.opt_level == OptimizationLevel::None {
-                ctx.make_assert(
-                    generator,
-                    ctx.builder
-                        .build_int_compare(
-                            IntPredicate::UGT,
-                            ndarray_nbytes,
-                            ndarray_nbytes.get_type().const_zero(),
-                            "",
-                        )
-                        .unwrap(),
-                    "0:AssertionError",
-                    "Unexpected RPC termination for ndarray - Expected data buffer next",
-                    [None, None, None],
-                    ctx.current_loc,
-                );
-            }
-
-            // Copy shape from the buffer to `ndarray.shape`.
-            let pbuffer_dims =
-                unsafe { buffer.ptr_offset_unchecked(ctx, generator, &llvm_pdata_sizeof, None) };
-
-            call_memcpy_generic(
-                ctx,
-                ndarray.dim_sizes().base_ptr(ctx, generator),
-                pbuffer_dims,
-                sizeof_dims,
-                llvm_i1.const_zero(),
-            );
-            // Restore stack from before allocation of buffer
-            call_stackrestore(ctx, stackptr);
-
-            // Allocate `ndarray.data`.
-            // `ndarray.shape` must be initialized beforehand in this implementation
-            //   (for ndarray.create_data() to know how many elements to allocate)
-            let num_elements =
-                call_ndarray_calc_size(generator, ctx, &ndarray.dim_sizes(), (None, None));
-
-            // debug_assert(nelems * sizeof(T) >= ndarray_nbytes)
-            if ctx.registry.llvm_options.opt_level == OptimizationLevel::None {
-                let sizeof_data =
-                    ctx.builder.build_int_mul(num_elements, llvm_elem_sizeof, "").unwrap();
-
-                ctx.make_assert(
-                    generator,
-                    ctx.builder.build_int_compare(IntPredicate::UGE,
-                        sizeof_data,
-                        ndarray_nbytes,
-                        "",
-                    ).unwrap(),
-                    "0:AssertionError",
-                    "Unexpected allocation size request for ndarray data - Expected up to {0} bytes, got {1} bytes",
-                    [Some(sizeof_data), Some(ndarray_nbytes), None],
-                    ctx.current_loc,
-                );
-            }
-
-            ndarray.create_data(ctx, llvm_elem_ty, num_elements);
-
-            let ndarray_data = ndarray.data().base_ptr(ctx, generator);
-            let ndarray_data_i8 =
-                ctx.builder.build_pointer_cast(ndarray_data, llvm_pi8, "").unwrap();
-
-            // NOTE: Currently on `prehead_bb`
-            ctx.builder.build_unconditional_branch(head_bb).unwrap();
-
-            // Inserting into `head_bb`. Do `rpc_recv` for `data` recursively.
-            ctx.builder.position_at_end(head_bb);
-
-            let phi = ctx.builder.build_phi(llvm_pi8, "rpc.ptr").unwrap();
-            phi.add_incoming(&[(&ndarray_data_i8, prehead_bb)]);
-
-            let alloc_size = ctx
-                .build_call_or_invoke(rpc_recv, &[phi.as_basic_value()], "rpc.size.next")
-                .map(BasicValueEnum::into_int_value)
-                .unwrap();
-
-            let is_done = ctx
-                .builder
-                .build_int_compare(IntPredicate::EQ, llvm_i32.const_zero(), alloc_size, "rpc.done")
-                .unwrap();
-            ctx.builder.build_conditional_branch(is_done, tail_bb, alloc_bb).unwrap();
-
-            ctx.builder.position_at_end(alloc_bb);
-            // Align the allocation to sizeof(T)
-            let alloc_size = round_up(ctx, alloc_size, llvm_elem_sizeof);
-            let alloc_ptr = ctx
-                .builder
-                .build_array_alloca(
-                    llvm_elem_ty,
-                    ctx.builder.build_int_unsigned_div(alloc_size, llvm_elem_sizeof, "").unwrap(),
-                    "rpc.alloc",
-                )
-                .unwrap();
-            let alloc_ptr =
-                ctx.builder.build_pointer_cast(alloc_ptr, llvm_pi8, "rpc.alloc.ptr").unwrap();
-            phi.add_incoming(&[(&alloc_ptr, alloc_bb)]);
-            ctx.builder.build_unconditional_branch(head_bb).unwrap();
-
-            ctx.builder.position_at_end(tail_bb);
-            ndarray.as_base_value().into()
-        }
-
-        _ => {
-            let slot = ctx.builder.build_alloca(llvm_ret_ty, "rpc.ret.slot").unwrap();
-            let slotgen = ctx.builder.build_bitcast(slot, llvm_pi8, "rpc.ret.ptr").unwrap();
-            ctx.builder.build_unconditional_branch(head_bb).unwrap();
-            ctx.builder.position_at_end(head_bb);
-
-            let phi = ctx.builder.build_phi(llvm_pi8, "rpc.ptr").unwrap();
-            phi.add_incoming(&[(&slotgen, prehead_bb)]);
-            let alloc_size = ctx
-                .build_call_or_invoke(rpc_recv, &[phi.as_basic_value()], "rpc.size.next")
-                .unwrap()
-                .into_int_value();
-            let is_done = ctx
-                .builder
-                .build_int_compare(IntPredicate::EQ, llvm_i32.const_zero(), alloc_size, "rpc.done")
-                .unwrap();
-
-            ctx.builder.build_conditional_branch(is_done, tail_bb, alloc_bb).unwrap();
-            ctx.builder.position_at_end(alloc_bb);
-
-            let alloc_ptr =
-                ctx.builder.build_array_alloca(llvm_pi8, alloc_size, "rpc.alloc").unwrap();
-            let alloc_ptr =
-                ctx.builder.build_bitcast(alloc_ptr, llvm_pi8, "rpc.alloc.ptr").unwrap();
-            phi.add_incoming(&[(&alloc_ptr, alloc_bb)]);
-            ctx.builder.build_unconditional_branch(head_bb).unwrap();
-
-            ctx.builder.position_at_end(tail_bb);
-            ctx.builder.build_load(slot, "rpc.result").unwrap()
-        }
-    };
-
-    Some(result)
-}
-
 fn rpc_codegen_callback_fn<'ctx>(
     ctx: &mut CodeGenContext<'ctx, '_>,
     obj: Option<(Type, ValueEnum<'ctx>)>,
@@ -953,14 +663,63 @@ fn rpc_codegen_callback_fn<'ctx>(
     // reclaim stack space used by arguments
     call_stackrestore(ctx, stackptr);
 
-    let result = format_rpc_ret(generator, ctx, fun.0.ret);
+    // -- receive value:
+    // T result = {
+    //   void *ret_ptr = alloca(sizeof(T));
+    //   void *ptr = ret_ptr;
+    //   loop: int size = rpc_recv(ptr);
+    //   // Non-zero: Provide `size` bytes of extra storage for variable-length data.
+    //   if(size) { ptr = alloca(size); goto loop; }
+    //   else *(T*)ret_ptr
+    // }
+    let rpc_recv = ctx.module.get_function("rpc_recv").unwrap_or_else(|| {
+        ctx.module.add_function("rpc_recv", int32.fn_type(&[ptr_type.into()], false), None)
+    });
 
-    if !result.is_some_and(|res| res.get_type().is_pointer_type()) {
+    if ctx.unifier.unioned(fun.0.ret, ctx.primitives.none) {
-        // An RPC returning an NDArray would not touch here.
+        ctx.build_call_or_invoke(rpc_recv, &[ptr_type.const_null().into()], "rpc_recv");
-        call_stackrestore(ctx, stackptr);
+        return Ok(None);
     }
 
-    Ok(result)
+    let prehead_bb = ctx.builder.get_insert_block().unwrap();
+    let current_function = prehead_bb.get_parent().unwrap();
+    let head_bb = ctx.ctx.append_basic_block(current_function, "rpc.head");
+    let alloc_bb = ctx.ctx.append_basic_block(current_function, "rpc.continue");
+    let tail_bb = ctx.ctx.append_basic_block(current_function, "rpc.tail");
+
+    let ret_ty = ctx.get_llvm_abi_type(generator, fun.0.ret);
+    let need_load = !ret_ty.is_pointer_type();
+    let slot = ctx.builder.build_alloca(ret_ty, "rpc.ret.slot").unwrap();
+    let slotgen = ctx.builder.build_bitcast(slot, ptr_type, "rpc.ret.ptr").unwrap();
+    ctx.builder.build_unconditional_branch(head_bb).unwrap();
+    ctx.builder.position_at_end(head_bb);
+
+    let phi = ctx.builder.build_phi(ptr_type, "rpc.ptr").unwrap();
+    phi.add_incoming(&[(&slotgen, prehead_bb)]);
+    let alloc_size = ctx
+        .build_call_or_invoke(rpc_recv, &[phi.as_basic_value()], "rpc.size.next")
+        .unwrap()
+        .into_int_value();
+    let is_done = ctx
+        .builder
+        .build_int_compare(inkwell::IntPredicate::EQ, int32.const_zero(), alloc_size, "rpc.done")
+        .unwrap();
+
+    ctx.builder.build_conditional_branch(is_done, tail_bb, alloc_bb).unwrap();
+    ctx.builder.position_at_end(alloc_bb);
+
+    let alloc_ptr = ctx.builder.build_array_alloca(ptr_type, alloc_size, "rpc.alloc").unwrap();
+    let alloc_ptr = ctx.builder.build_bitcast(alloc_ptr, ptr_type, "rpc.alloc.ptr").unwrap();
+    phi.add_incoming(&[(&alloc_ptr, alloc_bb)]);
+    ctx.builder.build_unconditional_branch(head_bb).unwrap();
+
+    ctx.builder.position_at_end(tail_bb);
+
+    let result = ctx.builder.build_load(slot, "rpc.result").unwrap();
+    if need_load {
+        call_stackrestore(ctx, stackptr);
+    }
+    Ok(Some(result))
 }
 
 pub fn attributes_writeback(

nac3core/src/codegen/classes.rs

@@ -1404,7 +1404,7 @@ impl<'ctx> NDArrayValue<'ctx> {
 
     /// Returns the double-indirection pointer to the `data` array, as if by calling `getelementptr`
     /// on the field.
-    pub fn ptr_to_data(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+    fn ptr_to_data(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
         let llvm_i32 = ctx.ctx.i32_type();
         let var_name = self.name.map(|v| format!("{v}.data.addr")).unwrap_or_default();