Auto merge of rust-lang#147022 - Zalathar:no-args, r=wesleywiser

Remove current code for embedding command-line args in PDB

The compiler currently has code that will obtain a list of quoted command-line arguments, and pass it through to TargetMachine creation, so that the command-line args can be embedded in PDB output.

This PR removes that code, due to subtle concerns that might not have been apparent when it was originally added.

---

Those concerns include:
- The entire command-line quoting process is repeated every time a target-machine-factory is created. In incremental builds this typically occurs 500+ times, instead of happening only once. The repeated quoting constitutes a large chunk of instructions executed in the `large-workspace` benchmark.
  - See rust-lang#146804 (comment) for an example of the perf consequences of skipping all that work.
  - This overhead occurs even when building for targets or configurations that don't emit PDB output.
- Command-line arguments are obtained in a way that completely bypasses the query system, which is a problem for the integrity of incremental compilation.
  - Fixing this alone is likely to inhibit incremental rebuilds for most or all CGUs, even in builds that don't emit PDB output.
- Command-line arguments and the executable path are obtained in a way that completely bypasses the compiler's path-remapping system, which is a reproducibility hazard.
  - rust-lang#128842

---

Relevant PRs:
- rust-lang#113492
- rust-lang#130446
- rust-lang#131805
- rust-lang#146700
- rust-lang#146973

Zulip thread:
- https://rust-lang.zulipchat.com/#narrow/channel/131828-t-compiler/topic/Some.20PDB.20info.20bypasses.20the.20query.20system.20and.20path.20remapping/with/541432211

---

According to rust-lang#96475, one of the big motivations for embedding the command-line arguments was to enable tools like Live++. [It appears that Live++ doesn't actually support Rust yet](https://rust-lang.zulipchat.com/#narrow/channel/131828-t-compiler/topic/embeded.20compiler.20args.20and.20--remap-path-prefix/near/523800010), so it's possible that there aren't any existing workflows for this removal to break.

In the future, there could be a case for reintroducing some or all of this functionality, guarded behind an opt-in flag so that it doesn't cause problems for other users. But as it stands, the current implementation puts a disproportionate burden on other users and on compiler maintainers.

Author: bors

compiler/rustc_codegen_llvm/src/back/command_line_args.rs

@@ -1,5 +1,4 @@
 pub(crate) mod archive;
-mod command_line_args;
 pub(crate) mod lto;
 pub(crate) mod owned_target_machine;
 mod profiling;

compiler/rustc_codegen_llvm/src/back/command_line_args/tests.rs

@@ -38,8 +38,6 @@ impl OwnedTargetMachine {
         output_obj_file: &CStr,
         debug_info_compression: &CStr,
         use_emulated_tls: bool,
-        argv0: &str,
-        command_line_args: &str,
         use_wasm_eh: bool,
     ) -> Result<Self, LlvmError<'static>> {
         // SAFETY: llvm::LLVMRustCreateTargetMachine copies pointed to data
@@ -66,10 +64,6 @@ impl OwnedTargetMachine {
                 output_obj_file.as_ptr(),
                 debug_info_compression.as_ptr(),
                 use_emulated_tls,
-                argv0.as_ptr(),
-                argv0.len(),
-                command_line_args.as_ptr(),
-                command_line_args.len(),
                 use_wasm_eh,
             )
         };

compiler/rustc_codegen_llvm/src/back/mod.rs

@@ -31,7 +31,6 @@ use rustc_span::{BytePos, InnerSpan, Pos, SpanData, SyntaxContext, sym};
 use rustc_target::spec::{CodeModel, FloatAbi, RelocModel, SanitizerSet, SplitDebuginfo, TlsModel};
 use tracing::{debug, trace};
 
-use crate::back::command_line_args::quote_command_line_args;
 use crate::back::lto::ThinBuffer;
 use crate::back::owned_target_machine::OwnedTargetMachine;
 use crate::back::profiling::{
@@ -253,19 +252,6 @@ pub(crate) fn target_machine_factory(
 
     let use_emulated_tls = matches!(sess.tls_model(), TlsModel::Emulated);
 
-    // Command-line information to be included in the target machine.
-    // This seems to only be used for embedding in PDB debuginfo files.
-    // FIXME(Zalathar): Maybe skip this for non-PDB targets?
-    let argv0 = std::env::current_exe()
-        .unwrap_or_default()
-        .into_os_string()
-        .into_string()
-        .unwrap_or_default();
-    let command_line_args = quote_command_line_args(&sess.expanded_args);
-    // Self-profile counter for the number of bytes produced by command-line quoting.
-    // Values are summed, so the summary result is cumulative across all TM factories.
-    sess.prof.artifact_size("quoted_command_line_args", "-", command_line_args.len() as u64);
-
     let debuginfo_compression = sess.opts.debuginfo_compression.to_string();
     match sess.opts.debuginfo_compression {
         rustc_session::config::DebugInfoCompression::Zlib => {
@@ -326,8 +312,6 @@ pub(crate) fn target_machine_factory(
             &output_obj_file,
             &debuginfo_compression,
             use_emulated_tls,
-            &argv0,
-            &command_line_args,
             use_wasm_eh,
         )
     })

compiler/rustc_codegen_llvm/src/back/owned_target_machine.rs

@@ -2330,10 +2330,6 @@ unsafe extern "C" {
         OutputObjFile: *const c_char,
         DebugInfoCompression: *const c_char,
         UseEmulatedTls: bool,
-        Argv0: *const c_uchar, // See "PTR_LEN_STR".
-        Argv0Len: size_t,
-        CommandLineArgs: *const c_uchar, // See "PTR_LEN_STR".
-        CommandLineArgsLen: size_t,
         UseWasmEH: bool,
     ) -> *mut TargetMachine;
 

compiler/rustc_codegen_llvm/src/back/write.rs

@@ -346,12 +346,6 @@ pub struct CodegenContext<B: WriteBackendMethods> {
     pub split_dwarf_kind: rustc_session::config::SplitDwarfKind,
     pub pointer_size: Size,
 
-    /// All commandline args used to invoke the compiler, with @file args fully expanded.
-    /// This will only be used within debug info, e.g. in the pdb file on windows
-    /// This is mainly useful for other tools that reads that debuginfo to figure out
-    /// how to call the compiler with the same arguments.
-    pub expanded_args: Vec<String>,
-
     /// Emitter to use for diagnostics produced during codegen.
     pub diag_emitter: SharedEmitter,
     /// LLVM optimizations for which we want to print remarks.
@@ -1153,7 +1147,6 @@ fn start_executing_work<B: ExtraBackendMethods>(
         remark: sess.opts.cg.remark.clone(),
         remark_dir,
         incr_comp_session_dir: sess.incr_comp_session_dir_opt().map(|r| r.clone()),
-        expanded_args: tcx.sess.expanded_args.clone(),
         diag_emitter: shared_emitter.clone(),
         output_filenames: Arc::clone(tcx.output_filenames(())),
         module_config: regular_config,

compiler/rustc_codegen_llvm/src/llvm/ffi.rs

@@ -269,7 +269,6 @@ pub fn run_compiler(at_args: &[String], callbacks: &mut (dyn Callbacks + Send))
         make_codegen_backend: None,
         registry: diagnostics_registry(),
         using_internal_features: &USING_INTERNAL_FEATURES,
-        expanded_args: args,
     };
 
     callbacks.config(&mut config);

compiler/rustc_codegen_ssa/src/back/write.rs

@@ -376,12 +376,6 @@ pub struct Config {
     /// enabled. Makes it so that "please report a bug" is hidden, as ICEs with
     /// internal features are wontfix, and they are usually the cause of the ICEs.
     pub using_internal_features: &'static std::sync::atomic::AtomicBool,
-
-    /// All commandline args used to invoke the compiler, with @file args fully expanded.
-    /// This will only be used within debug info, e.g. in the pdb file on windows
-    /// This is mainly useful for other tools that reads that debuginfo to figure out
-    /// how to call the compiler with the same arguments.
-    pub expanded_args: Vec<String>,
 }
 
 /// Initialize jobserver before getting `jobserver::client` and `build_session`.
@@ -480,7 +474,6 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
                 util::rustc_version_str().unwrap_or("unknown"),
                 config.ice_file,
                 config.using_internal_features,
-                config.expanded_args,
             );
 
             codegen_backend.init(&sess);

compiler/rustc_driver_impl/src/lib.rs

@@ -78,7 +78,6 @@ where
             "",
             None,
             &USING_INTERNAL_FEATURES,
-            Default::default(),
         );
         let cfg = parse_cfg(sess.dcx(), matches.opt_strs("cfg"));
         let cfg = build_configuration(&sess, cfg);