Remove LoadSplat opcode, in preparation for pattern-matching Load+Splat.

This was added as an incremental step to improve AArch64 code quality in
PR #2278. At the time, we did not have a way to pattern-match the load +
splat opcode sequence that the relevant Wasm opcodes lowered to.
However, now with PR #2366, we can merge effectful instructions such as
loads into other ops, and so we can do this pattern matching directly.
The pattern-matching update will come in a subsequent commit.

cranelift/codegen/meta/src/isa/x86/legalize.rs

@@ -396,7 +396,6 @@ fn define_simd(
     let insertlane = insts.by_name("insertlane");
     let ishl = insts.by_name("ishl");
     let ishl_imm = insts.by_name("ishl_imm");
-    let load_splat = insts.by_name("load_splat");
     let raw_bitcast = insts.by_name("raw_bitcast");
     let scalar_to_vector = insts.by_name("scalar_to_vector");
     let splat = insts.by_name("splat");
@@ -821,7 +820,6 @@ fn define_simd(
     narrow.custom_legalize(fcvt_to_sint_sat, "expand_fcvt_to_sint_sat_vector");
     narrow.custom_legalize(fmin, "expand_minmax_vector");
     narrow.custom_legalize(fmax, "expand_minmax_vector");
-    narrow.custom_legalize(load_splat, "expand_load_splat");
 
     narrow_avx.custom_legalize(imul, "convert_i64x2_imul");
     narrow_avx.custom_legalize(fcvt_from_uint, "expand_fcvt_from_uint_vector");

cranelift/codegen/meta/src/shared/instructions.rs

@@ -4491,24 +4491,5 @@ pub(crate) fn define(
         .other_side_effects(true),
     );
 
-    let Offset = &Operand::new("Offset", &imm.offset32).with_doc("Byte offset from base address");
-    let a = &Operand::new("a", TxN);
-
-    ig.push(
-        Inst::new(
-            "load_splat",
-            r#"
-        Load an element from memory at ``p + Offset`` and return a vector
-        whose lanes are all set to that element.
-
-        This is equivalent to ``load`` followed by ``splat``.
-        "#,
-            &formats.load,
-        )
-        .operands_in(vec![MemFlags, p, Offset])
-        .operands_out(vec![a])
-        .can_load(true),
-    );
-
     ig.build()
 }

cranelift/codegen/src/isa/aarch64/lower_inst.rs

@@ -1219,6 +1219,7 @@ pub(crate) fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
             }
         }
 
+        /*
         Opcode::LoadSplat => {
             let off = ctx.data(insn).load_store_offset().unwrap();
             let ty = ty.unwrap();
@@ -1234,6 +1235,7 @@ pub(crate) fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 size,
             });
         }
+        */
 
         Opcode::Store
         | Opcode::Istore8

cranelift/codegen/src/isa/x86/enc_tables.rs

@@ -1892,31 +1892,3 @@ fn expand_tls_value(
         unreachable!();
     }
 }
-
-fn expand_load_splat(
-    inst: ir::Inst,
-    func: &mut ir::Function,
-    _cfg: &mut ControlFlowGraph,
-    _isa: &dyn TargetIsa,
-) {
-    let mut pos = FuncCursor::new(func).at_inst(inst);
-
-    pos.use_srcloc(inst);
-
-    let (ptr, offset, flags) = match pos.func.dfg[inst] {
-        ir::InstructionData::Load {
-            opcode: ir::Opcode::LoadSplat,
-            arg,
-            offset,
-            flags,
-        } => (arg, offset, flags),
-        _ => panic!(
-            "Expected load_splat: {}",
-            pos.func.dfg.display_inst(inst, None)
-        ),
-    };
-    let ty = pos.func.dfg.ctrl_typevar(inst);
-    let load = pos.ins().load(ty.lane_type(), flags, ptr, offset);
-
-    pos.func.dfg.replace(inst).splat(ty, load);
-}

cranelift/interpreter/src/step.rs

@@ -545,7 +545,6 @@ where
         Opcode::Shuffle => unimplemented!("Shuffle"),
         Opcode::Swizzle => unimplemented!("Swizzle"),
         Opcode::Splat => unimplemented!("Splat"),
-        Opcode::LoadSplat => unimplemented!("LoadSplat"),
         Opcode::Insertlane => unimplemented!("Insertlane"),
         Opcode::Extractlane => unimplemented!("Extractlane"),
         Opcode::VhighBits => unimplemented!("VhighBits"),

cranelift/wasm/src/code_translator.rs

@@ -1414,17 +1414,16 @@ pub fn translate_operator<FE: FuncEnvironment + ?Sized>(
         | Operator::V128Load16Splat { memarg }
         | Operator::V128Load32Splat { memarg }
         | Operator::V128Load64Splat { memarg } => {
-            let opcode = ir::Opcode::LoadSplat;
-            let result_ty = type_of(op);
-            let (flags, base, offset) = prepare_load(
+            translate_load(
                 memarg,
-                mem_op_size(opcode, result_ty.lane_type()),
+                ir::Opcode::Load,
+                type_of(op).lane_type(),
                 builder,
                 state,
                 environ,
             )?;
-            let (load, dfg) = builder.ins().Load(opcode, result_ty, flags, offset, base);
-            state.push1(dfg.first_result(load))
+            let splatted = builder.ins().splat(type_of(op), state.pop1());
+            state.push1(splatted)
         }
         Operator::V128Load32Zero { memarg } | Operator::V128Load64Zero { memarg } => {
             translate_load(
@@ -2103,7 +2102,7 @@ fn mem_op_size(opcode: ir::Opcode, ty: Type) -> u32 {
         ir::Opcode::Istore8 | ir::Opcode::Sload8 | ir::Opcode::Uload8 => 1,
         ir::Opcode::Istore16 | ir::Opcode::Sload16 | ir::Opcode::Uload16 => 2,
         ir::Opcode::Istore32 | ir::Opcode::Sload32 | ir::Opcode::Uload32 => 4,
-        ir::Opcode::Store | ir::Opcode::Load | ir::Opcode::LoadSplat => ty.bytes(),
+        ir::Opcode::Store | ir::Opcode::Load => ty.bytes(),
         _ => panic!("unknown size of mem op for {:?}", opcode),
     }
 }