Adding the x86 part of behavioural testing for std::arch intrinsics

ci/run.sh

@@ -99,6 +99,9 @@ fi
 # Test targets compiled with extra features.
 case ${TARGET} in
     x86_64-unknown-linux-gnu)
+        TEST_CPPFLAGS="-fuse-ld=lld -I/usr/include/x86_64-linux-gnu/"
+        TEST_CXX_COMPILER="clang++-19"
+        TEST_RUNNER="${CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_RUNNER}"
         export STDARCH_DISABLE_ASSERT_INSTR=1
 
         export RUSTFLAGS="${RUSTFLAGS} -C target-feature=+avx"
@@ -189,6 +192,14 @@ case "${TARGET}" in
             --linker "${CARGO_TARGET_AARCH64_BE_UNKNOWN_LINUX_GNU_LINKER}" \
             --cxx-toolchain-dir "${AARCH64_BE_TOOLCHAIN}"
         ;;
+    x86_64-unknown-linux-gnu*)
+        CPPFLAGS="${TEST_CPPFLAGS}" RUSTFLAGS="${HOST_RUSTFLAGS}" RUST_LOG=warn \
+            cargo run "${INTRINSIC_TEST}" "${PROFILE}"  \
+            --bin intrinsic-test -- intrinsics_data/x86-intel.xml \
+            --runner "${TEST_RUNNER}" \
+            --cppcompiler "${TEST_CXX_COMPILER}" \
+            --target "${TARGET}"
+        ;;
      *)
         ;;
 esac

crates/intrinsic-test/Cargo.toml

@@ -22,3 +22,5 @@ pretty_env_logger = "0.5.0"
 rayon = "1.5.0"
 diff = "0.1.12"
 itertools = "0.14.0"
+quick-xml = { version = "0.37.5", features = ["serialize", "overlapped-lists"] }
+serde-xml-rs = "0.8.0"

crates/intrinsic-test/src/arm/argument.rs

@@ -0,0 +1,12 @@
+use crate::arm::intrinsic::ArmIntrinsicType;
+use crate::common::argument::Argument;
+
+impl Argument<ArmIntrinsicType> {
+    pub fn type_and_name_from_c(arg: &str) -> (&str, &str) {
+        let split_index = arg
+            .rfind([' ', '*'])
+            .expect("Couldn't split type and argname");
+
+        (arg[..split_index + 1].trim_end(), &arg[split_index + 1..])
+    }
+}

crates/intrinsic-test/src/arm/json_parser.rs

@@ -79,20 +79,26 @@ fn json_to_intrinsic(
 ) -> Result<Intrinsic<ArmIntrinsicType>, Box<dyn std::error::Error>> {
     let name = intr.name.replace(['[', ']'], "");
 
-    let results = ArmIntrinsicType::from_c(&intr.return_type.value, target)?;
-
+    let mut results = ArmIntrinsicType::from_c(&intr.return_type.value)?;
+    results.set_metadata("target".to_string(), target.to_string());
     let args = intr
         .arguments
         .into_iter()
         .enumerate()
         .map(|(i, arg)| {
-            let arg_name = Argument::<ArmIntrinsicType>::type_and_name_from_c(&arg).1;
-            let metadata = intr.args_prep.as_mut();
-            let metadata = metadata.and_then(|a| a.remove(arg_name));
-            let arg_prep: Option<ArgPrep> = metadata.and_then(|a| a.try_into().ok());
+            let (type_name, arg_name) = Argument::<ArmIntrinsicType>::type_and_name_from_c(&arg);
+            let ty = ArmIntrinsicType::from_c(type_name)
+                .unwrap_or_else(|_| panic!("Failed to parse argument '{arg}'"));
+
+            let arg_prep = intr.args_prep.as_mut();
+            let arg_prep = arg_prep.and_then(|a| a.remove(arg_name));
+            let arg_prep: Option<ArgPrep> = arg_prep.and_then(|a| a.try_into().ok());
             let constraint: Option<Constraint> = arg_prep.and_then(|a| a.try_into().ok());
 
-            let mut arg = Argument::<ArmIntrinsicType>::from_c(i, &arg, target, constraint);
+            let mut arg =
+                Argument::<ArmIntrinsicType>::new(i, arg_name.to_string(), ty, constraint);
+            arg.ty
+                .set_metadata("target".to_string(), target.to_string());
 
             // The JSON doesn't list immediates as const
             let IntrinsicType {

crates/intrinsic-test/src/arm/mod.rs

@@ -1,10 +1,11 @@
+mod argument;
 mod compile;
 mod config;
 mod intrinsic;
 mod json_parser;
 mod types;
 
-use std::fs::File;
+use std::fs;
 
 use rayon::prelude::*;
 
@@ -68,28 +69,35 @@ impl SupportedArchitectureTest for ArmArchitectureTest {
 
         let (chunk_size, chunk_count) = chunk_info(self.intrinsics.len());
 
-        let cpp_compiler = compile::build_cpp_compilation(&self.cli_options).unwrap();
+        let cpp_compiler = compile::build_cpp_compilation(&self.cli_options);
+
+        match fs::exists("c_programs") {
+            Ok(false) => fs::create_dir("c_programs").unwrap(),
+            Ok(true) => {}
+            _ => return false,
+        }
 
         let notice = &build_notices("// ");
         self.intrinsics
             .par_chunks(chunk_size)
             .enumerate()
             .map(|(i, chunk)| {
                 let c_filename = format!("c_programs/mod_{i}.cpp");
-                let mut file = File::create(&c_filename).unwrap();
+                let mut file = fs::File::create(&c_filename).unwrap();
                 write_mod_cpp(&mut file, notice, c_target, platform_headers, chunk).unwrap();
 
                 // compile this cpp file into a .o file
-                let output = cpp_compiler
-                    .compile_object_file(&format!("mod_{i}.cpp"), &format!("mod_{i}.o"))?;
-                assert!(output.status.success(), "{output:?}");
-
+                if let Some(compiler) = cpp_compiler.as_ref() {
+                    let output = compiler
+                        .compile_object_file(&format!("mod_{i}.cpp"), &format!("mod_{i}.o"))?;
+                    assert!(output.status.success(), "{output:?}");
+                }
                 Ok(())
             })
             .collect::<Result<(), std::io::Error>>()
             .unwrap();
 
-        let mut file = File::create("c_programs/main.cpp").unwrap();
+        let mut file = fs::File::create("c_programs/main.cpp").unwrap();
         write_main_cpp(
             &mut file,
             c_target,
@@ -99,20 +107,22 @@ impl SupportedArchitectureTest for ArmArchitectureTest {
         .unwrap();
 
         // compile this cpp file into a .o file
-        info!("compiling main.cpp");
-        let output = cpp_compiler
-            .compile_object_file("main.cpp", "intrinsic-test-programs.o")
-            .unwrap();
-        assert!(output.status.success(), "{output:?}");
-
-        let object_files = (0..chunk_count)
-            .map(|i| format!("mod_{i}.o"))
-            .chain(["intrinsic-test-programs.o".to_owned()]);
-
-        let output = cpp_compiler
-            .link_executable(object_files, "intrinsic-test-programs")
-            .unwrap();
-        assert!(output.status.success(), "{output:?}");
+        if let Some(compiler) = cpp_compiler.as_ref() {
+            info!("compiling main.cpp");
+            let output = compiler
+                .compile_object_file("main.cpp", "intrinsic-test-programs.o")
+                .unwrap();
+            assert!(output.status.success(), "{output:?}");
+
+            let object_files = (0..chunk_count)
+                .map(|i| format!("mod_{i}.o"))
+                .chain(["intrinsic-test-programs.o".to_owned()]);
+
+            let output = compiler
+                .link_executable(object_files, "intrinsic-test-programs")
+                .unwrap();
+            assert!(output.status.success(), "{output:?}");
+        }
 
         true
     }

crates/intrinsic-test/src/arm/types.rs

@@ -1,25 +1,21 @@
+use std::collections::HashMap;
+
 use super::intrinsic::ArmIntrinsicType;
 use crate::common::cli::Language;
 use crate::common::intrinsic_helpers::{IntrinsicType, IntrinsicTypeDefinition, Sign, TypeKind};
 
 impl IntrinsicTypeDefinition for ArmIntrinsicType {
     /// Gets a string containing the typename for this type in C format.
+    /// This assumes that the metadata hashmap contains this value at the
+    /// "type" key
     fn c_type(&self) -> String {
-        let prefix = self.0.kind.c_prefix();
-        let const_prefix = if self.0.constant { "const " } else { "" };
-
-        if let (Some(bit_len), simd_len, vec_len) =
-            (self.0.bit_len, self.0.simd_len, self.0.vec_len)
-        {
-            match (simd_len, vec_len) {
-                (None, None) => format!("{const_prefix}{prefix}{bit_len}_t"),
-                (Some(simd), None) => format!("{prefix}{bit_len}x{simd}_t"),
-                (Some(simd), Some(vec)) => format!("{prefix}{bit_len}x{simd}x{vec}_t"),
-                (None, Some(_)) => todo!("{:#?}", self), // Likely an invalid case
-            }
-        } else {
-            todo!("{:#?}", self)
-        }
+        self.metadata
+            .get("type")
+            .expect("Failed to extract the C typename in Aarch!")
+            .replace("*", "")
+            .replace("  ", "")
+            .trim()
+            .to_string()
     }
 
     fn c_single_vector_type(&self) -> String {
@@ -59,7 +55,7 @@ impl IntrinsicTypeDefinition for ArmIntrinsicType {
             bit_len: Some(bl),
             simd_len,
             vec_len,
-            target,
+            metadata,
             ..
         } = &self.0
         {
@@ -69,7 +65,11 @@ impl IntrinsicTypeDefinition for ArmIntrinsicType {
                 ""
             };
 
-            let choose_workaround = language == Language::C && target.contains("v7");
+            let choose_workaround = language == Language::C
+                && metadata
+                    .get("target")
+                    .filter(|value| value.contains("v7"))
+                    .is_some();
             format!(
                 "vld{len}{quad}_{type}{size}",
                 type = match k {
@@ -123,6 +123,8 @@ impl IntrinsicTypeDefinition for ArmIntrinsicType {
 
     fn from_c(s: &str, target: &str) -> Result<Self, String> {
         const CONST_STR: &str = "const";
+        let mut metadata: HashMap<String, String> = HashMap::new();
+        metadata.insert("type".to_string(), s.to_string());
         if let Some(s) = s.strip_suffix('*') {
             let (s, constant) = match s.trim().strip_suffix(CONST_STR) {
                 Some(stripped) => (stripped, true),

crates/intrinsic-test/src/common/argument.rs

@@ -20,6 +20,15 @@ impl<T> Argument<T>
 where
     T: IntrinsicTypeDefinition,
 {
+    pub fn new(pos: usize, name: String, ty: T, constraint: Option<Constraint>) -> Self {
+        Argument {
+            pos,
+            name,
+            ty,
+            constraint,
+        }
+    }
+
     pub fn to_c_type(&self) -> String {
         self.ty.c_type()
     }
@@ -36,14 +45,6 @@ where
         self.constraint.is_some()
     }
 
-    pub fn type_and_name_from_c(arg: &str) -> (&str, &str) {
-        let split_index = arg
-            .rfind([' ', '*'])
-            .expect("Couldn't split type and argname");
-
-        (arg[..split_index + 1].trim_end(), &arg[split_index + 1..])
-    }
-
     /// The binding keyword (e.g. "const" or "let") for the array of possible test inputs.
     fn rust_vals_array_binding(&self) -> impl std::fmt::Display {
         if self.ty.is_rust_vals_array_const() {

crates/intrinsic-test/src/common/constraint.rs

@@ -5,13 +5,15 @@ use std::ops::Range;
 pub enum Constraint {
     Equal(i64),
     Range(Range<i64>),
+    Set(Vec<i64>),
 }
 
 impl Constraint {
-    pub fn to_range(&self) -> Range<i64> {
+    pub fn to_vector(&self) -> Vec<i64> {
         match self {
-            Constraint::Equal(eq) => *eq..*eq + 1,
-            Constraint::Range(range) => range.clone(),
+            Constraint::Equal(eq) => vec![*eq],
+            Constraint::Range(range) => range.clone().collect::<Vec<i64>>(),
+            Constraint::Set(values) => values.clone(),
         }
     }
 }