diff options
| author | Boxy <rust@boxyuwu.dev> | 2025-01-29 16:46:09 +0000 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-01-29 16:46:09 +0000 |
| commit | 0b4890851287f7552e25beb43ab67243a87a1ff4 (patch) | |
| tree | 6b28cb50611b2bbcca661aaef967dd4a6eb7f26f /compiler/rustc_codegen_ssa/src | |
| parent | bec4359db9bf5fd3c6e9d4a88d26dcf091b3c82a (diff) | |
| parent | 815c5d4eee36e836c7b75aa9288a58c4e8e7830b (diff) | |
| download | rust-0b4890851287f7552e25beb43ab67243a87a1ff4.tar.gz rust-0b4890851287f7552e25beb43ab67243a87a1ff4.zip | |
Rustc pull
Diffstat (limited to 'compiler/rustc_codegen_ssa/src')
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/assert_module_sources.rs | 8 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/back/command.rs | 17 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/back/link.rs | 77 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/back/lto.rs | 4 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/back/symbol_export.rs | 2 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/base.rs | 24 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/codegen_attrs.rs | 10 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/errors.rs | 62 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/lib.rs | 22 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/mir/debuginfo.rs | 6 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/mir/naked_asm.rs | 173 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/mir/operand.rs | 59 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/mir/rvalue.rs | 73 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/traits/builder.rs | 21 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/traits/consts.rs | 1 | ||||
| -rw-r--r-- | compiler/rustc_codegen_ssa/src/traits/write.rs | 2 |
16 files changed, 444 insertions, 117 deletions
diff --git a/compiler/rustc_codegen_ssa/src/assert_module_sources.rs b/compiler/rustc_codegen_ssa/src/assert_module_sources.rs index ab65319e3d3..27331ce4ca6 100644 --- a/compiler/rustc_codegen_ssa/src/assert_module_sources.rs +++ b/compiler/rustc_codegen_ssa/src/assert_module_sources.rs @@ -46,8 +46,12 @@ pub fn assert_module_sources(tcx: TyCtxt<'_>, set_reuse: &dyn Fn(&mut CguReuseTr return; } - let available_cgus = - tcx.collect_and_partition_mono_items(()).1.iter().map(|cgu| cgu.name()).collect(); + let available_cgus = tcx + .collect_and_partition_mono_items(()) + .codegen_units + .iter() + .map(|cgu| cgu.name()) + .collect(); let mut ams = AssertModuleSource { tcx, diff --git a/compiler/rustc_codegen_ssa/src/back/command.rs b/compiler/rustc_codegen_ssa/src/back/command.rs index b3c5b86ccf4..63023fdba20 100644 --- a/compiler/rustc_codegen_ssa/src/back/command.rs +++ b/compiler/rustc_codegen_ssa/src/back/command.rs @@ -13,6 +13,7 @@ pub(crate) struct Command { args: Vec<OsString>, env: Vec<(OsString, OsString)>, env_remove: Vec<OsString>, + env_clear: bool, } #[derive(Clone)] @@ -36,7 +37,13 @@ impl Command { } fn _new(program: Program) -> Command { - Command { program, args: Vec::new(), env: Vec::new(), env_remove: Vec::new() } + Command { + program, + args: Vec::new(), + env: Vec::new(), + env_remove: Vec::new(), + env_clear: false, + } } pub(crate) fn arg<P: AsRef<OsStr>>(&mut self, arg: P) -> &mut Command { @@ -79,6 +86,11 @@ impl Command { self } + pub(crate) fn env_clear(&mut self) -> &mut Command { + self.env_clear = true; + self + } + fn _env_remove(&mut self, key: &OsStr) { self.env_remove.push(key.to_owned()); } @@ -106,6 +118,9 @@ impl Command { for k in &self.env_remove { ret.env_remove(k); } + if self.env_clear { + ret.env_clear(); + } ret } diff --git a/compiler/rustc_codegen_ssa/src/back/link.rs b/compiler/rustc_codegen_ssa/src/back/link.rs index df35b5e8426..105a4cb81f0 100644 --- a/compiler/rustc_codegen_ssa/src/back/link.rs +++ b/compiler/rustc_codegen_ssa/src/back/link.rs @@ -15,12 +15,14 @@ use rustc_ast::CRATE_NODE_ID; use rustc_data_structures::fx::{FxIndexMap, FxIndexSet}; use rustc_data_structures::memmap::Mmap; use rustc_data_structures::temp_dir::MaybeTempDir; -use rustc_errors::DiagCtxtHandle; +use rustc_errors::{DiagCtxtHandle, LintDiagnostic}; use rustc_fs_util::{fix_windows_verbatim_for_gcc, try_canonicalize}; use rustc_hir::def_id::{CrateNum, LOCAL_CRATE}; +use rustc_macros::LintDiagnostic; use rustc_metadata::fs::{METADATA_FILENAME, copy_to_stdout, emit_wrapper_file}; use rustc_metadata::{find_native_static_library, walk_native_lib_search_dirs}; use rustc_middle::bug; +use rustc_middle::lint::lint_level; use rustc_middle::middle::debugger_visualizer::DebuggerVisualizerFile; use rustc_middle::middle::dependency_format::Linkage; use rustc_middle::middle::exported_symbols::SymbolExportKind; @@ -29,6 +31,7 @@ use rustc_session::config::{ OutputType, PrintKind, SplitDwarfKind, Strip, }; use rustc_session::cstore::DllImport; +use rustc_session::lint::builtin::LINKER_MESSAGES; use rustc_session::output::{check_file_is_writeable, invalid_output_for_target, out_filename}; use rustc_session::search_paths::PathKind; use rustc_session::utils::NativeLibKind; @@ -749,6 +752,14 @@ fn link_dwarf_object(sess: &Session, cg_results: &CodegenResults, executable_out } } +#[derive(LintDiagnostic)] +#[diag(codegen_ssa_linker_output)] +/// Translating this is kind of useless. We don't pass translation flags to the linker, so we'd just +/// end up with inconsistent languages within the same diagnostic. +struct LinkerOutput { + inner: String, +} + /// Create a dynamic library or executable. /// /// This will invoke the system linker/cc to create the resulting file. This links to all upstream @@ -981,6 +992,11 @@ fn link_natively( match prog { Ok(prog) => { + let is_msvc_link_exe = sess.target.is_like_msvc + && flavor == LinkerFlavor::Msvc(Lld::No) + // Match exactly "link.exe" + && linker_path.to_str() == Some("link.exe"); + if !prog.status.success() { let mut output = prog.stderr.clone(); output.extend_from_slice(&prog.stdout); @@ -991,22 +1007,16 @@ fn link_natively( command: cmd, escaped_output, verbose: sess.opts.verbose, + sysroot_dir: sess.sysroot.clone(), }; sess.dcx().emit_err(err); // If MSVC's `link.exe` was expected but the return code // is not a Microsoft LNK error then suggest a way to fix or // install the Visual Studio build tools. if let Some(code) = prog.status.code() { - if sess.target.is_like_msvc - && flavor == LinkerFlavor::Msvc(Lld::No) - // Respect the command line override - && sess.opts.cg.linker.is_none() - // Match exactly "link.exe" - && linker_path.to_str() == Some("link.exe") - // All Microsoft `link.exe` linking error codes are - // four digit numbers in the range 1000 to 9999 inclusive - && (code < 1000 || code > 9999) - { + // All Microsoft `link.exe` linking ror codes are + // four digit numbers in the range 1000 to 9999 inclusive + if is_msvc_link_exe && (code < 1000 || code > 9999) { let is_vs_installed = windows_registry::find_vs_version().is_ok(); let has_linker = windows_registry::find_tool(&sess.target.arch, "link.exe").is_some(); @@ -1028,8 +1038,49 @@ fn link_natively( sess.dcx().abort_if_errors(); } - info!("linker stderr:\n{}", escape_string(&prog.stderr)); - info!("linker stdout:\n{}", escape_string(&prog.stdout)); + + let stderr = escape_string(&prog.stderr); + let mut stdout = escape_string(&prog.stdout); + info!("linker stderr:\n{}", &stderr); + info!("linker stdout:\n{}", &stdout); + + // Hide some progress messages from link.exe that we don't care about. + // See https://github.com/chromium/chromium/blob/bfa41e41145ffc85f041384280caf2949bb7bd72/build/toolchain/win/tool_wrapper.py#L144-L146 + if is_msvc_link_exe { + if let Ok(str) = str::from_utf8(&prog.stdout) { + let mut output = String::with_capacity(str.len()); + for line in stdout.lines() { + if line.starts_with(" Creating library") + || line.starts_with("Generating code") + || line.starts_with("Finished generating code") + { + continue; + } + output += line; + output += "\r\n" + } + stdout = escape_string(output.trim().as_bytes()) + } + } + + let (level, src) = codegen_results.crate_info.lint_levels.linker_messages; + let lint = |msg| { + lint_level(sess, LINKER_MESSAGES, level, src, None, |diag| { + LinkerOutput { inner: msg }.decorate_lint(diag) + }) + }; + + if !prog.stderr.is_empty() { + // We already print `warning:` at the start of the diagnostic. Remove it from the linker output if present. + let stderr = stderr + .strip_prefix("warning: ") + .unwrap_or(&stderr) + .replace(": warning: ", ": "); + lint(format!("linker stderr: {stderr}")); + } + if !stdout.is_empty() { + lint(format!("linker stdout: {}", stdout)) + } } Err(e) => { let linker_not_found = e.kind() == io::ErrorKind::NotFound; diff --git a/compiler/rustc_codegen_ssa/src/back/lto.rs b/compiler/rustc_codegen_ssa/src/back/lto.rs index efccf7687a1..9fd984b6419 100644 --- a/compiler/rustc_codegen_ssa/src/back/lto.rs +++ b/compiler/rustc_codegen_ssa/src/back/lto.rs @@ -4,7 +4,6 @@ use std::sync::Arc; use rustc_ast::expand::autodiff_attrs::AutoDiffItem; use rustc_data_structures::memmap::Mmap; use rustc_errors::FatalError; -use rustc_middle::ty::TyCtxt; use super::write::CodegenContext; use crate::ModuleCodegen; @@ -89,13 +88,12 @@ impl<B: WriteBackendMethods> LtoModuleCodegen<B> { pub unsafe fn autodiff( self, cgcx: &CodegenContext<B>, - tcx: TyCtxt<'_>, diff_fncs: Vec<AutoDiffItem>, config: &ModuleConfig, ) -> Result<LtoModuleCodegen<B>, FatalError> { match &self { LtoModuleCodegen::Fat(module) => { - B::autodiff(cgcx, tcx, &module, diff_fncs, config)?; + B::autodiff(cgcx, &module, diff_fncs, config)?; } _ => panic!("autodiff called with non-fat LTO module"), } diff --git a/compiler/rustc_codegen_ssa/src/back/symbol_export.rs b/compiler/rustc_codegen_ssa/src/back/symbol_export.rs index 60ab2919352..f8f7bb2dbc6 100644 --- a/compiler/rustc_codegen_ssa/src/back/symbol_export.rs +++ b/compiler/rustc_codegen_ssa/src/back/symbol_export.rs @@ -293,7 +293,7 @@ fn exported_symbols_provider_local( // external linkage is enough for monomorphization to be linked to. let need_visibility = tcx.sess.target.dynamic_linking && !tcx.sess.target.only_cdylib; - let (_, cgus) = tcx.collect_and_partition_mono_items(()); + let cgus = tcx.collect_and_partition_mono_items(()).codegen_units; // The symbols created in this loop are sorted below it #[allow(rustc::potential_query_instability)] diff --git a/compiler/rustc_codegen_ssa/src/base.rs b/compiler/rustc_codegen_ssa/src/base.rs index 544578b29f1..e438bd70c51 100644 --- a/compiler/rustc_codegen_ssa/src/base.rs +++ b/compiler/rustc_codegen_ssa/src/base.rs @@ -44,7 +44,8 @@ use crate::mir::operand::OperandValue; use crate::mir::place::PlaceRef; use crate::traits::*; use crate::{ - CachedModuleCodegen, CompiledModule, CrateInfo, ModuleCodegen, ModuleKind, errors, meth, mir, + CachedModuleCodegen, CodegenLintLevels, CompiledModule, CrateInfo, ModuleCodegen, ModuleKind, + errors, meth, mir, }; pub(crate) fn bin_op_to_icmp_predicate(op: BinOp, signed: bool) -> IntPredicate { @@ -490,8 +491,8 @@ pub fn maybe_create_entry_wrapper<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>( let ptr_ty = cx.type_ptr(); let (arg_argc, arg_argv) = get_argc_argv(&mut bx); - let (start_fn, start_ty, args, instance) = if let EntryFnType::Main { sigpipe } = entry_type - { + let EntryFnType::Main { sigpipe } = entry_type; + let (start_fn, start_ty, args, instance) = { let start_def_id = cx.tcx().require_lang_item(LangItem::Start, None); let start_instance = ty::Instance::expect_resolve( cx.tcx(), @@ -512,10 +513,6 @@ pub fn maybe_create_entry_wrapper<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>( vec![rust_main, arg_argc, arg_argv, arg_sigpipe], Some(start_instance), ) - } else { - debug!("using user-defined start fn"); - let start_ty = cx.type_func(&[isize_ty, ptr_ty], isize_ty); - (rust_main, start_ty, vec![arg_argc, arg_argv], None) }; let result = bx.call(start_ty, None, None, start_fn, &args, None, instance); @@ -530,7 +527,8 @@ pub fn maybe_create_entry_wrapper<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>( } } -/// Obtain the `argc` and `argv` values to pass to the rust start function. +/// Obtain the `argc` and `argv` values to pass to the rust start function +/// (i.e., the "start" lang item). fn get_argc_argv<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>(bx: &mut Bx) -> (Bx::Value, Bx::Value) { if bx.cx().sess().target.os.contains("uefi") { // Params for UEFI @@ -621,7 +619,7 @@ pub fn codegen_crate<B: ExtraBackendMethods>( // Run the monomorphization collector and partition the collected items into // codegen units. - let codegen_units = tcx.collect_and_partition_mono_items(()).1; + let codegen_units = tcx.collect_and_partition_mono_items(()).codegen_units; // Force all codegen_unit queries so they are already either red or green // when compile_codegen_unit accesses them. We are not able to re-execute @@ -927,6 +925,7 @@ impl CrateInfo { dependency_formats: Lrc::clone(tcx.dependency_formats(())), windows_subsystem, natvis_debugger_visualizers: Default::default(), + lint_levels: CodegenLintLevels::from_tcx(tcx), }; info.native_libraries.reserve(n_crates); @@ -1052,14 +1051,11 @@ pub(crate) fn provide(providers: &mut Providers) { config::OptLevel::SizeMin => config::OptLevel::Default, }; - let (defids, _) = tcx.collect_and_partition_mono_items(cratenum); + let defids = tcx.collect_and_partition_mono_items(cratenum).all_mono_items; let any_for_speed = defids.items().any(|id| { let CodegenFnAttrs { optimize, .. } = tcx.codegen_fn_attrs(*id); - match optimize { - attr::OptimizeAttr::None | attr::OptimizeAttr::Size => false, - attr::OptimizeAttr::Speed => true, - } + matches!(optimize, attr::OptimizeAttr::Speed) }); if any_for_speed { diff --git a/compiler/rustc_codegen_ssa/src/codegen_attrs.rs b/compiler/rustc_codegen_ssa/src/codegen_attrs.rs index 1daa17fbaf3..a0bc2d4ea48 100644 --- a/compiler/rustc_codegen_ssa/src/codegen_attrs.rs +++ b/compiler/rustc_codegen_ssa/src/codegen_attrs.rs @@ -575,7 +575,7 @@ fn codegen_fn_attrs(tcx: TyCtxt<'_>, did: LocalDefId) -> CodegenFnAttrs { codegen_fn_attrs.inline = InlineAttr::Never; } - codegen_fn_attrs.optimize = attrs.iter().fold(OptimizeAttr::None, |ia, attr| { + codegen_fn_attrs.optimize = attrs.iter().fold(OptimizeAttr::Default, |ia, attr| { if !attr.has_name(sym::optimize) { return ia; } @@ -587,17 +587,19 @@ fn codegen_fn_attrs(tcx: TyCtxt<'_>, did: LocalDefId) -> CodegenFnAttrs { inline_span = Some(attr.span); if items.len() != 1 { err(attr.span, "expected one argument"); - OptimizeAttr::None + OptimizeAttr::Default } else if list_contains_name(items, sym::size) { OptimizeAttr::Size } else if list_contains_name(items, sym::speed) { OptimizeAttr::Speed + } else if list_contains_name(items, sym::none) { + OptimizeAttr::DoNotOptimize } else { err(items[0].span(), "invalid argument"); - OptimizeAttr::None + OptimizeAttr::Default } } else { - OptimizeAttr::None + OptimizeAttr::Default } }); diff --git a/compiler/rustc_codegen_ssa/src/errors.rs b/compiler/rustc_codegen_ssa/src/errors.rs index c7213bbc801..5e684632fb2 100644 --- a/compiler/rustc_codegen_ssa/src/errors.rs +++ b/compiler/rustc_codegen_ssa/src/errors.rs @@ -351,6 +351,7 @@ pub(crate) struct LinkingFailed<'a> { pub command: Command, pub escaped_output: String, pub verbose: bool, + pub sysroot_dir: PathBuf, } impl<G: EmissionGuarantee> Diagnostic<'_, G> for LinkingFailed<'_> { @@ -364,6 +365,8 @@ impl<G: EmissionGuarantee> Diagnostic<'_, G> for LinkingFailed<'_> { if self.verbose { diag.note(format!("{:?}", self.command)); } else { + self.command.env_clear(); + enum ArgGroup { Regular(OsString), Objects(usize), @@ -398,26 +401,55 @@ impl<G: EmissionGuarantee> Diagnostic<'_, G> for LinkingFailed<'_> { args.push(ArgGroup::Regular(arg)); } } - self.command.args(args.into_iter().map(|arg_group| match arg_group { - ArgGroup::Regular(arg) => arg, - ArgGroup::Objects(n) => OsString::from(format!("<{n} object files omitted>")), - ArgGroup::Rlibs(dir, rlibs) => { - let mut arg = dir.into_os_string(); - arg.push("/{"); - let mut first = true; - for rlib in rlibs { - if !first { - arg.push(","); + let crate_hash = regex::bytes::Regex::new(r"-[0-9a-f]+\.rlib$").unwrap(); + self.command.args(args.into_iter().map(|arg_group| { + match arg_group { + // SAFETY: we are only matching on ASCII, not any surrogate pairs, so any replacements we do will still be valid. + ArgGroup::Regular(arg) => unsafe { + use bstr::ByteSlice; + OsString::from_encoded_bytes_unchecked( + arg.as_encoded_bytes().replace( + self.sysroot_dir.as_os_str().as_encoded_bytes(), + b"<sysroot>", + ), + ) + }, + ArgGroup::Objects(n) => OsString::from(format!("<{n} object files omitted>")), + ArgGroup::Rlibs(mut dir, rlibs) => { + let is_sysroot_dir = match dir.strip_prefix(&self.sysroot_dir) { + Ok(short) => { + dir = Path::new("<sysroot>").join(short); + true + } + Err(_) => false, + }; + let mut arg = dir.into_os_string(); + arg.push("/{"); + let mut first = true; + for mut rlib in rlibs { + if !first { + arg.push(","); + } + first = false; + if is_sysroot_dir { + // SAFETY: Regex works one byte at a type, and our regex will not match surrogate pairs (because it only matches ascii). + rlib = unsafe { + OsString::from_encoded_bytes_unchecked( + crate_hash + .replace(rlib.as_encoded_bytes(), b"-*") + .into_owned(), + ) + }; + } + arg.push(rlib); } - first = false; - arg.push(rlib); + arg.push("}.rlib"); + arg } - arg.push("}"); - arg } })); - diag.note(format!("{:?}", self.command)); + diag.note(format!("{:?}", self.command).trim_start_matches("env -i").to_owned()); diag.note("some arguments are omitted. use `--verbose` to show all linker arguments"); } diff --git a/compiler/rustc_codegen_ssa/src/lib.rs b/compiler/rustc_codegen_ssa/src/lib.rs index 65c6067c740..40299ae2630 100644 --- a/compiler/rustc_codegen_ssa/src/lib.rs +++ b/compiler/rustc_codegen_ssa/src/lib.rs @@ -29,18 +29,23 @@ use rustc_ast as ast; use rustc_data_structures::fx::{FxHashSet, FxIndexMap}; use rustc_data_structures::sync::Lrc; use rustc_data_structures::unord::UnordMap; +use rustc_hir::CRATE_HIR_ID; use rustc_hir::def_id::CrateNum; use rustc_macros::{Decodable, Encodable, HashStable}; use rustc_middle::dep_graph::WorkProduct; +use rustc_middle::lint::LintLevelSource; use rustc_middle::middle::debugger_visualizer::DebuggerVisualizerFile; use rustc_middle::middle::dependency_format::Dependencies; use rustc_middle::middle::exported_symbols::SymbolExportKind; +use rustc_middle::ty::TyCtxt; use rustc_middle::util::Providers; use rustc_serialize::opaque::{FileEncoder, MemDecoder}; use rustc_serialize::{Decodable, Decoder, Encodable, Encoder}; use rustc_session::Session; use rustc_session::config::{CrateType, OutputFilenames, OutputType, RUST_CGU_EXT}; use rustc_session::cstore::{self, CrateSource}; +use rustc_session::lint::Level; +use rustc_session::lint::builtin::LINKER_MESSAGES; use rustc_session::utils::NativeLibKind; use rustc_span::Symbol; @@ -200,6 +205,7 @@ pub struct CrateInfo { pub dependency_formats: Lrc<Dependencies>, pub windows_subsystem: Option<String>, pub natvis_debugger_visualizers: BTreeSet<DebuggerVisualizerFile>, + pub lint_levels: CodegenLintLevels, } #[derive(Encodable, Decodable)] @@ -302,3 +308,19 @@ impl CodegenResults { Ok((codegen_results, outputs)) } } + +/// A list of lint levels used in codegen. +/// +/// When using `-Z link-only`, we don't have access to the tcx and must work +/// solely from the `.rlink` file. `Lint`s are defined too early to be encodeable. +/// Instead, encode exactly the information we need. +#[derive(Copy, Clone, Debug, Encodable, Decodable)] +pub struct CodegenLintLevels { + linker_messages: (Level, LintLevelSource), +} + +impl CodegenLintLevels { + pub fn from_tcx(tcx: TyCtxt<'_>) -> Self { + Self { linker_messages: tcx.lint_level_at_node(LINKER_MESSAGES, CRATE_HIR_ID) } + } +} diff --git a/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs b/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs index 843a996d2bf..5924c8991ad 100644 --- a/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs +++ b/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs @@ -19,9 +19,7 @@ use crate::traits::*; pub struct FunctionDebugContext<'tcx, S, L> { /// Maps from source code to the corresponding debug info scope. - /// May be None if the backend is not capable of representing the scope for - /// some reason. - pub scopes: IndexVec<mir::SourceScope, Option<DebugScope<S, L>>>, + pub scopes: IndexVec<mir::SourceScope, DebugScope<S, L>>, /// Maps from an inlined function to its debug info declaration. pub inlined_function_scopes: FxHashMap<Instance<'tcx>, S>, @@ -232,7 +230,7 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { &self, source_info: mir::SourceInfo, ) -> Option<(Bx::DIScope, Option<Bx::DILocation>, Span)> { - let scope = &self.debug_context.as_ref()?.scopes[source_info.scope]?; + let scope = &self.debug_context.as_ref()?.scopes[source_info.scope]; let span = hygiene::walk_chain_collapsed(source_info.span, self.mir.span); Some((scope.adjust_dbg_scope_for_span(self.cx, span), scope.inlined_at, span)) } diff --git a/compiler/rustc_codegen_ssa/src/mir/naked_asm.rs b/compiler/rustc_codegen_ssa/src/mir/naked_asm.rs index 8df270abc81..dc406809874 100644 --- a/compiler/rustc_codegen_ssa/src/mir/naked_asm.rs +++ b/compiler/rustc_codegen_ssa/src/mir/naked_asm.rs @@ -1,10 +1,14 @@ +use rustc_abi::{BackendRepr, Float, Integer, Primitive, RegKind}; use rustc_attr_parsing::InstructionSetAttr; +use rustc_hir::def_id::DefId; use rustc_middle::mir::mono::{Linkage, MonoItem, MonoItemData, Visibility}; use rustc_middle::mir::{Body, InlineAsmOperand}; -use rustc_middle::ty::layout::{HasTyCtxt, HasTypingEnv, LayoutOf}; -use rustc_middle::ty::{Instance, TyCtxt}; -use rustc_middle::{bug, ty}; +use rustc_middle::ty::layout::{FnAbiOf, HasTyCtxt, HasTypingEnv, LayoutOf}; +use rustc_middle::ty::{Instance, Ty, TyCtxt}; +use rustc_middle::{bug, span_bug, ty}; use rustc_span::sym; +use rustc_target::callconv::{ArgAbi, FnAbi, PassMode}; +use rustc_target::spec::WasmCAbi; use crate::common; use crate::traits::{AsmCodegenMethods, BuilderMethods, GlobalAsmOperandRef, MiscCodegenMethods}; @@ -32,7 +36,8 @@ pub(crate) fn codegen_naked_asm<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>( let item_data = cx.codegen_unit().items().get(&MonoItem::Fn(instance)).unwrap(); let name = cx.mangled_name(instance); - let (begin, end) = prefix_and_suffix(cx.tcx(), instance, &name, item_data); + let fn_abi = cx.fn_abi_of_instance(instance, ty::List::empty()); + let (begin, end) = prefix_and_suffix(cx.tcx(), instance, &name, item_data, fn_abi); let mut template_vec = Vec::new(); template_vec.push(rustc_ast::ast::InlineAsmTemplatePiece::String(begin.into())); @@ -103,6 +108,7 @@ enum AsmBinaryFormat { Elf, Macho, Coff, + Wasm, } impl AsmBinaryFormat { @@ -111,6 +117,8 @@ impl AsmBinaryFormat { Self::Coff } else if target.is_like_osx { Self::Macho + } else if target.is_like_wasm { + Self::Wasm } else { Self::Elf } @@ -122,6 +130,7 @@ fn prefix_and_suffix<'tcx>( instance: Instance<'tcx>, asm_name: &str, item_data: &MonoItemData, + fn_abi: &FnAbi<'tcx, Ty<'tcx>>, ) -> (String, String) { use std::fmt::Write; @@ -169,7 +178,7 @@ fn prefix_and_suffix<'tcx>( } Linkage::LinkOnceAny | Linkage::LinkOnceODR | Linkage::WeakAny | Linkage::WeakODR => { match asm_binary_format { - AsmBinaryFormat::Elf | AsmBinaryFormat::Coff => { + AsmBinaryFormat::Elf | AsmBinaryFormat::Coff | AsmBinaryFormat::Wasm => { writeln!(w, ".weak {asm_name}")?; } AsmBinaryFormat::Macho => { @@ -264,7 +273,161 @@ fn prefix_and_suffix<'tcx>( writeln!(end, "{}", arch_suffix).unwrap(); } } + AsmBinaryFormat::Wasm => { + let section = link_section.unwrap_or(format!(".text.{asm_name}")); + + writeln!(begin, ".section {section},\"\",@").unwrap(); + // wasm functions cannot be aligned, so skip + write_linkage(&mut begin).unwrap(); + if let Visibility::Hidden = item_data.visibility { + writeln!(begin, ".hidden {asm_name}").unwrap(); + } + writeln!(begin, ".type {asm_name}, @function").unwrap(); + if !arch_prefix.is_empty() { + writeln!(begin, "{}", arch_prefix).unwrap(); + } + writeln!(begin, "{asm_name}:").unwrap(); + writeln!( + begin, + ".functype {asm_name} {}", + wasm_functype(tcx, fn_abi, instance.def_id()) + ) + .unwrap(); + + writeln!(end).unwrap(); + // .size is ignored for function symbols, so we can skip it + writeln!(end, "end_function").unwrap(); + } } (begin, end) } + +/// The webassembly type signature for the given function. +/// +/// Used by the `.functype` directive on wasm targets. +fn wasm_functype<'tcx>(tcx: TyCtxt<'tcx>, fn_abi: &FnAbi<'tcx, Ty<'tcx>>, def_id: DefId) -> String { + let mut signature = String::with_capacity(64); + + let ptr_type = match tcx.data_layout.pointer_size.bits() { + 32 => "i32", + 64 => "i64", + other => bug!("wasm pointer size cannot be {other} bits"), + }; + + // FIXME: remove this once the wasm32-unknown-unknown ABI is fixed + // please also add `wasm32-unknown-unknown` back in `tests/assembly/wasm32-naked-fn.rs` + // basically the commit introducing this comment should be reverted + if let PassMode::Pair { .. } = fn_abi.ret.mode { + let _ = WasmCAbi::Legacy; + span_bug!( + tcx.def_span(def_id), + "cannot return a pair (the wasm32-unknown-unknown ABI is broken, see https://github.com/rust-lang/rust/issues/115666" + ); + } + + let hidden_return = matches!(fn_abi.ret.mode, PassMode::Indirect { .. }); + + signature.push('('); + + if hidden_return { + signature.push_str(ptr_type); + if !fn_abi.args.is_empty() { + signature.push_str(", "); + } + } + + let mut it = fn_abi.args.iter().peekable(); + while let Some(arg_abi) = it.next() { + wasm_type(tcx, &mut signature, arg_abi, ptr_type, def_id); + if it.peek().is_some() { + signature.push_str(", "); + } + } + + signature.push_str(") -> ("); + + if !hidden_return { + wasm_type(tcx, &mut signature, &fn_abi.ret, ptr_type, def_id); + } + + signature.push(')'); + + signature +} + +fn wasm_type<'tcx>( + tcx: TyCtxt<'tcx>, + signature: &mut String, + arg_abi: &ArgAbi<'_, Ty<'tcx>>, + ptr_type: &'static str, + def_id: DefId, +) { + match arg_abi.mode { + PassMode::Ignore => { /* do nothing */ } + PassMode::Direct(_) => { + let direct_type = match arg_abi.layout.backend_repr { + BackendRepr::Scalar(scalar) => wasm_primitive(scalar.primitive(), ptr_type), + BackendRepr::Vector { .. } => "v128", + BackendRepr::Memory { .. } => { + // FIXME: remove this branch once the wasm32-unknown-unknown ABI is fixed + let _ = WasmCAbi::Legacy; + span_bug!( + tcx.def_span(def_id), + "cannot use memory args (the wasm32-unknown-unknown ABI is broken, see https://github.com/rust-lang/rust/issues/115666" + ); + } + other => unreachable!("unexpected BackendRepr: {:?}", other), + }; + + signature.push_str(direct_type); + } + PassMode::Pair(_, _) => match arg_abi.layout.backend_repr { + BackendRepr::ScalarPair(a, b) => { + signature.push_str(wasm_primitive(a.primitive(), ptr_type)); + signature.push_str(", "); + signature.push_str(wasm_primitive(b.primitive(), ptr_type)); + } + other => unreachable!("{other:?}"), + }, + PassMode::Cast { pad_i32, ref cast } => { + // For wasm, Cast is used for single-field primitive wrappers like `struct Wrapper(i64);` + assert!(!pad_i32, "not currently used by wasm calling convention"); + assert!(cast.prefix[0].is_none(), "no prefix"); + assert_eq!(cast.rest.total, arg_abi.layout.size, "single item"); + + let wrapped_wasm_type = match cast.rest.unit.kind { + RegKind::Integer => match cast.rest.unit.size.bytes() { + ..=4 => "i32", + ..=8 => "i64", + _ => ptr_type, + }, + RegKind::Float => match cast.rest.unit.size.bytes() { + ..=4 => "f32", + ..=8 => "f64", + _ => ptr_type, + }, + RegKind::Vector => "v128", + }; + + signature.push_str(wrapped_wasm_type); + } + PassMode::Indirect { .. } => signature.push_str(ptr_type), + } +} + +fn wasm_primitive(primitive: Primitive, ptr_type: &'static str) -> &'static str { + match primitive { + Primitive::Int(integer, _) => match integer { + Integer::I8 | Integer::I16 | Integer::I32 => "i32", + Integer::I64 => "i64", + Integer::I128 => "i64, i64", + }, + Primitive::Float(float) => match float { + Float::F16 | Float::F32 => "f32", + Float::F64 => "f64", + Float::F128 => "i64, i64", + }, + Primitive::Pointer(_) => ptr_type, + } +} diff --git a/compiler/rustc_codegen_ssa/src/mir/operand.rs b/compiler/rustc_codegen_ssa/src/mir/operand.rs index 19101ec2d1b..9ca7d4f8f00 100644 --- a/compiler/rustc_codegen_ssa/src/mir/operand.rs +++ b/compiler/rustc_codegen_ssa/src/mir/operand.rs @@ -204,14 +204,30 @@ impl<'a, 'tcx, V: CodegenObject> OperandRef<'tcx, V> { let alloc_align = alloc.inner().align; assert!(alloc_align >= layout.align.abi); + // Returns `None` when the value is partially undefined or any byte of it has provenance. + // Otherwise returns the value or (if the entire value is undef) returns an undef. let read_scalar = |start, size, s: abi::Scalar, ty| { + let range = alloc_range(start, size); match alloc.0.read_scalar( bx, - alloc_range(start, size), + range, /*read_provenance*/ matches!(s.primitive(), abi::Primitive::Pointer(_)), ) { - Ok(val) => bx.scalar_to_backend(val, s, ty), - Err(_) => bx.const_poison(ty), + Ok(val) => Some(bx.scalar_to_backend(val, s, ty)), + Err(_) => { + // We may have failed due to partial provenance or unexpected provenance, + // continue down the normal code path if so. + if alloc.0.provenance().range_empty(range, &bx.tcx()) + // Since `read_scalar` failed, but there were no relocations involved, the + // bytes must be partially or fully uninitialized. Thus we can now unwrap the + // information about the range of uninit bytes and check if it's the full range. + && alloc.0.init_mask().is_range_initialized(range).unwrap_err() == range + { + Some(bx.const_undef(ty)) + } else { + None + } + } } }; @@ -222,16 +238,14 @@ impl<'a, 'tcx, V: CodegenObject> OperandRef<'tcx, V> { // check that walks over the type of `mplace` to make sure it is truly correct to treat this // like a `Scalar` (or `ScalarPair`). match layout.backend_repr { - BackendRepr::Scalar(s @ abi::Scalar::Initialized { .. }) => { + BackendRepr::Scalar(s) => { let size = s.size(bx); assert_eq!(size, layout.size, "abi::Scalar size does not match layout size"); - let val = read_scalar(offset, size, s, bx.immediate_backend_type(layout)); - OperandRef { val: OperandValue::Immediate(val), layout } + if let Some(val) = read_scalar(offset, size, s, bx.immediate_backend_type(layout)) { + return OperandRef { val: OperandValue::Immediate(val), layout }; + } } - BackendRepr::ScalarPair( - a @ abi::Scalar::Initialized { .. }, - b @ abi::Scalar::Initialized { .. }, - ) => { + BackendRepr::ScalarPair(a, b) => { let (a_size, b_size) = (a.size(bx), b.size(bx)); let b_offset = (offset + a_size).align_to(b.align(bx).abi); assert!(b_offset.bytes() > 0); @@ -247,20 +261,21 @@ impl<'a, 'tcx, V: CodegenObject> OperandRef<'tcx, V> { b, bx.scalar_pair_element_backend_type(layout, 1, true), ); - OperandRef { val: OperandValue::Pair(a_val, b_val), layout } - } - _ if layout.is_zst() => OperandRef::zero_sized(layout), - _ => { - // Neither a scalar nor scalar pair. Load from a place - // FIXME: should we cache `const_data_from_alloc` to avoid repeating this for the - // same `ConstAllocation`? - let init = bx.const_data_from_alloc(alloc); - let base_addr = bx.static_addr_of(init, alloc_align, None); - - let llval = bx.const_ptr_byte_offset(base_addr, offset); - bx.load_operand(PlaceRef::new_sized(llval, layout)) + if let (Some(a_val), Some(b_val)) = (a_val, b_val) { + return OperandRef { val: OperandValue::Pair(a_val, b_val), layout }; + } } + _ if layout.is_zst() => return OperandRef::zero_sized(layout), + _ => {} } + // Neither a scalar nor scalar pair. Load from a place + // FIXME: should we cache `const_data_from_alloc` to avoid repeating this for the + // same `ConstAllocation`? + let init = bx.const_data_from_alloc(alloc); + let base_addr = bx.static_addr_of(init, alloc_align, None); + + let llval = bx.const_ptr_byte_offset(base_addr, offset); + bx.load_operand(PlaceRef::new_sized(llval, layout)) } /// Asserts that this operand refers to a scalar and returns diff --git a/compiler/rustc_codegen_ssa/src/mir/rvalue.rs b/compiler/rustc_codegen_ssa/src/mir/rvalue.rs index eb4ef599b82..e775d219c7b 100644 --- a/compiler/rustc_codegen_ssa/src/mir/rvalue.rs +++ b/compiler/rustc_codegen_ssa/src/mir/rvalue.rs @@ -8,7 +8,7 @@ use rustc_middle::ty::{self, Instance, Ty, TyCtxt}; use rustc_middle::{bug, mir, span_bug}; use rustc_session::config::OptLevel; use rustc_span::{DUMMY_SP, Span}; -use tracing::{debug, instrument}; +use tracing::{debug, instrument, trace}; use super::operand::{OperandRef, OperandValue}; use super::place::PlaceRef; @@ -93,6 +93,8 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { return; } + // If `v` is an integer constant whose value is just a single byte repeated N times, + // emit a `memset` filling the entire `dest` with that byte. let try_init_all_same = |bx: &mut Bx, v| { let start = dest.val.llval; let size = bx.const_usize(dest.layout.size.bytes()); @@ -117,13 +119,33 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { false }; + trace!(?cg_elem.val); match cg_elem.val { OperandValue::Immediate(v) => { if try_init_all_same(bx, v) { return; } } - _ => (), + OperandValue::Pair(a, b) => { + let a_is_undef = bx.cx().is_undef(a); + match (a_is_undef, bx.cx().is_undef(b)) { + // Can happen for uninit unions + (true, true) => { + // FIXME: can we produce better output here? + } + (false, true) | (true, false) => { + let val = if a_is_undef { b } else { a }; + if try_init_all_same(bx, val) { + return; + } + } + (false, false) => { + // FIXME: if both are the same value, use try_init_all_same + } + } + } + OperandValue::ZeroSized => unreachable!("checked above"), + OperandValue::Ref(..) => {} } let count = self @@ -365,10 +387,15 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { use abi::Primitive::*; imm = bx.from_immediate(imm); - // When scalars are passed by value, there's no metadata recording their - // valid ranges. For example, `char`s are passed as just `i32`, with no - // way for LLVM to know that they're 0x10FFFF at most. Thus we assume - // the range of the input value too, not just the output range. + // If we have a scalar, we must already know its range. Either + // + // 1) It's a parameter with `range` parameter metadata, + // 2) It's something we `load`ed with `!range` metadata, or + // 3) After a transmute we `assume`d the range (see below). + // + // That said, last time we tried removing this, it didn't actually help + // the rustc-perf results, so might as well keep doing it + // <https://github.com/rust-lang/rust/pull/135610#issuecomment-2599275182> self.assume_scalar_range(bx, imm, from_scalar, from_backend_ty); imm = match (from_scalar.primitive(), to_scalar.primitive()) { @@ -389,7 +416,14 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { bx.bitcast(int_imm, to_backend_ty) } }; + + // This `assume` remains important for cases like (a conceptual) + // transmute::<u32, NonZeroU32>(x) == 0 + // since it's never passed to something with parameter metadata (especially + // after MIR inlining) so the only way to tell the backend about the + // constraint that the `transmute` introduced is to `assume` it. self.assume_scalar_range(bx, imm, to_scalar, to_backend_ty); + imm = bx.to_immediate_scalar(imm, to_scalar); imm } @@ -411,31 +445,8 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> { return; } - let abi::WrappingRange { start, end } = scalar.valid_range(self.cx); - - if start <= end { - if start > 0 { - let low = bx.const_uint_big(backend_ty, start); - let cmp = bx.icmp(IntPredicate::IntUGE, imm, low); - bx.assume(cmp); - } - - let type_max = scalar.size(self.cx).unsigned_int_max(); - if end < type_max { - let high = bx.const_uint_big(backend_ty, end); - let cmp = bx.icmp(IntPredicate::IntULE, imm, high); - bx.assume(cmp); - } - } else { - let low = bx.const_uint_big(backend_ty, start); - let cmp_low = bx.icmp(IntPredicate::IntUGE, imm, low); - - let high = bx.const_uint_big(backend_ty, end); - let cmp_high = bx.icmp(IntPredicate::IntULE, imm, high); - - let or = bx.or(cmp_low, cmp_high); - bx.assume(or); - } + let range = scalar.valid_range(self.cx); + bx.assume_integer_range(imm, backend_ty, range); } pub(crate) fn codegen_rvalue_unsized( diff --git a/compiler/rustc_codegen_ssa/src/traits/builder.rs b/compiler/rustc_codegen_ssa/src/traits/builder.rs index 3ee13b19f66..bbf87a59942 100644 --- a/compiler/rustc_codegen_ssa/src/traits/builder.rs +++ b/compiler/rustc_codegen_ssa/src/traits/builder.rs @@ -217,6 +217,27 @@ pub trait BuilderMethods<'a, 'tcx>: dest: PlaceRef<'tcx, Self::Value>, ); + /// Emits an `assume` that the integer value `imm` of type `ty` is contained in `range`. + /// + /// This *always* emits the assumption, so you probably want to check the + /// optimization level and `Scalar::is_always_valid` before calling it. + fn assume_integer_range(&mut self, imm: Self::Value, ty: Self::Type, range: WrappingRange) { + let WrappingRange { start, end } = range; + + // Perhaps one day we'll be able to use assume operand bundles for this, + // but for now this encoding with a single icmp+assume is best per + // <https://github.com/llvm/llvm-project/issues/123278#issuecomment-2597440158> + let shifted = if start == 0 { + imm + } else { + let low = self.const_uint_big(ty, start); + self.sub(imm, low) + }; + let width = self.const_uint_big(ty, u128::wrapping_sub(end, start)); + let cmp = self.icmp(IntPredicate::IntULE, shifted, width); + self.assume(cmp); + } + fn range_metadata(&mut self, load: Self::Value, range: WrappingRange); fn nonnull_metadata(&mut self, load: Self::Value); diff --git a/compiler/rustc_codegen_ssa/src/traits/consts.rs b/compiler/rustc_codegen_ssa/src/traits/consts.rs index 9af463a691a..d0de7ff0b5f 100644 --- a/compiler/rustc_codegen_ssa/src/traits/consts.rs +++ b/compiler/rustc_codegen_ssa/src/traits/consts.rs @@ -9,6 +9,7 @@ pub trait ConstCodegenMethods<'tcx>: BackendTypes { /// Generate an uninitialized value (matching uninitialized memory in MIR). /// Whether memory is initialized or not is tracked byte-for-byte. fn const_undef(&self, t: Self::Type) -> Self::Value; + fn is_undef(&self, v: Self::Value) -> bool; /// Generate a fake value. Poison always affects the entire value, even if just a single byte is /// poison. This can only be used in codepaths that are already UB, i.e., UB-free Rust code /// (including code that e.g. copies uninit memory with `MaybeUninit`) can never encounter a diff --git a/compiler/rustc_codegen_ssa/src/traits/write.rs b/compiler/rustc_codegen_ssa/src/traits/write.rs index 51e2255efe1..97fe614aa10 100644 --- a/compiler/rustc_codegen_ssa/src/traits/write.rs +++ b/compiler/rustc_codegen_ssa/src/traits/write.rs @@ -1,7 +1,6 @@ use rustc_ast::expand::autodiff_attrs::AutoDiffItem; use rustc_errors::{DiagCtxtHandle, FatalError}; use rustc_middle::dep_graph::WorkProduct; -use rustc_middle::ty::TyCtxt; use crate::back::lto::{LtoModuleCodegen, SerializedModule, ThinModule}; use crate::back::write::{CodegenContext, FatLtoInput, ModuleConfig}; @@ -65,7 +64,6 @@ pub trait WriteBackendMethods: 'static + Sized + Clone { fn serialize_module(module: ModuleCodegen<Self::Module>) -> (String, Self::ModuleBuffer); fn autodiff( cgcx: &CodegenContext<Self>, - tcx: TyCtxt<'_>, module: &ModuleCodegen<Self::Module>, diff_fncs: Vec<AutoDiffItem>, config: &ModuleConfig, |
