Merge pull request #177 from rust-secure-code/fix-cargo-c

Fix cargo-c

Author: Shnatsel

cargo-auditable/Cargo.toml

@@ -18,7 +18,7 @@ auditable-serde = {version = "0.8.0", path = "../auditable-serde"}
 miniz_oxide = {version = "0.8.0"}
 serde_json = "1.0.57"
 cargo_metadata = "0.18"
-pico-args = "0.5"
+pico-args = { version = "0.5", features = ["eq-separator"] }
 serde = "1.0.147"
 wasm-gen = "0.1.4"
 

cargo-auditable/src/rustc_arguments.rs

@@ -12,10 +12,12 @@ use std::{ffi::OsString, path::PathBuf};
 // https://github.com/rust-lang/rust/blob/26ecd44160f54395b3bd5558cc5352f49cb0a0ba/compiler/rustc_session/src/config.rs
 
 /// Includes only the rustc arguments we care about
+#[derive(Debug)]
 pub struct RustcArgs {
     pub crate_name: String,
     pub crate_types: Vec<String>,
     pub cfg: Vec<String>,
+    pub emit: Vec<String>,
     pub out_dir: PathBuf,
     pub target: Option<String>,
     pub print: Vec<String>,
@@ -35,18 +37,96 @@ impl RustcArgs {
     }
 }
 
+impl RustcArgs {
+    // Split into its own function for unit testing
+    fn from_vec(raw_args: Vec<OsString>) -> Result<RustcArgs, pico_args::Error> {
+        let mut parser = pico_args::Arguments::from_vec(raw_args);
+
+        // --emit requires slightly more complex parsing
+        let raw_emit_args: Vec<String> = parser.values_from_str("--emit")?;
+        let mut emit: Vec<String> = Vec::new();
+        for raw_arg in raw_emit_args {
+            for item in raw_arg.split(',') {
+                emit.push(item.to_owned());
+            }
+        }
+
+        Ok(RustcArgs {
+            crate_name: parser.value_from_str("--crate-name")?,
+            crate_types: parser.values_from_str("--crate-type")?,
+            cfg: parser.values_from_str("--cfg")?,
+            emit,
+            out_dir: parser
+                .value_from_os_str::<&str, PathBuf, pico_args::Error>("--out-dir", |s| {
+                    Ok(PathBuf::from(s))
+                })?,
+            target: parser.opt_value_from_str("--target")?,
+            print: parser.values_from_str("--print")?,
+        })
+    }
+}
+
 pub fn parse_args() -> Result<RustcArgs, pico_args::Error> {
     let raw_args: Vec<OsString> = std::env::args_os().skip(2).collect();
-    let mut parser = pico_args::Arguments::from_vec(raw_args);
-
-    Ok(RustcArgs {
-        crate_name: parser.value_from_str("--crate-name")?,
-        crate_types: parser.values_from_str("--crate-type")?,
-        cfg: parser.values_from_str("--cfg")?,
-        out_dir: parser.value_from_os_str::<&str, PathBuf, pico_args::Error>("--out-dir", |s| {
-            Ok(PathBuf::from(s))
-        })?,
-        target: parser.opt_value_from_str("--target")?,
-        print: parser.values_from_str("--print")?,
-    })
+    RustcArgs::from_vec(raw_args)
+}
+
+pub fn should_embed_audit_data(args: &RustcArgs) -> bool {
+    // Only inject audit data into crate types 'bin' and 'cdylib',
+    // it doesn't make sense for static libs and weird other types.
+    if !(args.crate_types.contains(&"bin".to_owned())
+        || args.crate_types.contains(&"cdylib".to_owned()))
+    {
+        return false;
+    }
+
+    // when --emit is specified explicitly, only inject audit data for --emit=link
+    // because it doesn't make sense for all other types such as llvm-ir, asm, etc.
+    if !args.emit.is_empty() && !args.emit.contains(&"link".to_owned()) {
+        return false;
+    }
+
+    // --print disables compilation UNLESS --emit is also specified
+    if !args.print.is_empty() && args.emit.is_empty() {
+        return false;
+    }
+
+    true
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn cargo_c_compatibility() {
+        let raw_rustc_args = vec!["--crate-name", "rustls", "--edition=2021", "src/lib.rs", "--error-format=json", "--json=diagnostic-rendered-ansi,artifacts,future-incompat", "--crate-type", "staticlib", "--crate-type", "cdylib", "--emit=dep-info,link", "-C", "embed-bitcode=no", "-C", "debuginfo=2", "-C", "link-arg=-Wl,-soname,librustls.so.0.14.0", "-Cmetadata=rustls-ffi", "--cfg", "cargo_c", "--print", "native-static-libs", "--cfg", "feature=\"aws-lc-rs\"", "--cfg", "feature=\"capi\"", "--cfg", "feature=\"default\"", "--check-cfg", "cfg(docsrs)", "--check-cfg", "cfg(feature, values(\"aws-lc-rs\", \"capi\", \"cert_compression\", \"default\", \"no_log_capture\", \"read_buf\", \"ring\"))", "-C", "metadata=b6a43041f637feb8", "--out-dir", "/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps", "--target", "x86_64-unknown-linux-gnu", "-C", "linker=clang", "-C", "incremental=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/incremental", "-L", "dependency=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps", "-L", "dependency=/home/user/Code/rustls-ffi/target/debug/deps", "--extern", "libc=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/liblibc-4fc7c9f82dda33ee.rlib", "--extern", "log=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/liblog-6f7c8f4d1d5ec422.rlib", "--extern", "rustls=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/librustls-a93cda0ba0380929.rlib", "--extern", "pki_types=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/librustls_pki_types-27749859644f0979.rlib", "--extern", "rustls_platform_verifier=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/librustls_platform_verifier-bceca5cf09f3d7ba.rlib", "--extern", "webpki=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/deps/libwebpki-bc4a16dd84e0b062.rlib", "-C", "link-arg=-fuse-ld=/home/user/mold-2.32.0-x86_64-linux/bin/mold", "-L", "native=/home/user/Code/rustls-ffi/target/x86_64-unknown-linux-gnu/debug/build/aws-lc-sys-d52f8990d9ede41d/out"];
+        let raw_rustc_args: Vec<OsString> = raw_rustc_args.into_iter().map(|s| s.into()).collect();
+        let args = RustcArgs::from_vec(raw_rustc_args).unwrap();
+        assert!(should_embed_audit_data(&args));
+    }
+
+    #[test]
+    fn multiple_emit_values() {
+        let raw_rustc_args = vec![
+            "--emit=dep-info,link",
+            "--emit",
+            "llvm-bc",
+            // end of interesting args, start of boilerplate
+            "--crate-name",
+            "foobar",
+            "--out-dir",
+            "/foo/bar",
+        ];
+        let raw_rustc_args: Vec<OsString> = raw_rustc_args.into_iter().map(|s| s.into()).collect();
+        let mut args = RustcArgs::from_vec(raw_rustc_args).unwrap();
+
+        let expected = vec!["dep-info", "link", "llvm-bc"];
+        let mut expected: Vec<String> = expected.into_iter().map(|s| s.into()).collect();
+
+        args.emit.sort();
+        expected.sort();
+
+        assert_eq!(args.emit, expected)
+    }
 }

cargo-auditable/src/rustc_wrapper.rs

@@ -7,7 +7,8 @@ use std::{
 use crate::{
     binary_file, collect_audit_data,
     platform_detection::{is_apple, is_msvc, is_wasm},
-    rustc_arguments, target_info,
+    rustc_arguments::{self, should_embed_audit_data},
+    target_info,
 };
 
 use std::io::BufRead;
@@ -20,12 +21,7 @@ pub fn main(rustc_path: &OsStr) {
     if env::var_os("CARGO_PRIMARY_PACKAGE").is_some() {
         let arg_parsing_result = rustc_arguments::parse_args();
         if let Ok(args) = rustc_arguments::parse_args() {
-            // Only inject audit data into crate types 'bin' and 'cdylib',
-            // and only if --print is not specified (which disables compilation)
-            if args.print.is_empty()
-                && (args.crate_types.contains(&"bin".to_owned())
-                    || args.crate_types.contains(&"cdylib".to_owned()))
-            {
+            if should_embed_audit_data(&args) {
                 // Get the audit data to embed
                 let target_triple = args
                     .target