diff options
| author | Ralf Jung <post@ralfj.de> | 2023-07-10 10:16:05 +0200 |
|---|---|---|
| committer | Ralf Jung <post@ralfj.de> | 2023-07-10 10:16:05 +0200 |
| commit | a3fdf75d126a02da4c27fd9e1cc16bdc93a5f133 (patch) | |
| tree | 5803fda00e2ed5bfbff938e19453324144075fa4 /src/tools | |
| parent | d4034a2035d67637e044876c1e5801ef68d29889 (diff) | |
| parent | 743333f3dd90721461c09387ec73d09c080d5f5f (diff) | |
| download | rust-a3fdf75d126a02da4c27fd9e1cc16bdc93a5f133.tar.gz rust-a3fdf75d126a02da4c27fd9e1cc16bdc93a5f133.zip | |
Merge from rustc
Diffstat (limited to 'src/tools')
29 files changed, 1522 insertions, 140 deletions
diff --git a/src/tools/build_helper/Cargo.toml b/src/tools/build_helper/Cargo.toml index 99f6fea2ecf..66894e1abc4 100644 --- a/src/tools/build_helper/Cargo.toml +++ b/src/tools/build_helper/Cargo.toml @@ -6,3 +6,5 @@ edition = "2021" # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html [dependencies] +serde = "1" +serde_derive = "1" diff --git a/src/tools/build_helper/src/lib.rs b/src/tools/build_helper/src/lib.rs index 3fa970373b3..575f3677155 100644 --- a/src/tools/build_helper/src/lib.rs +++ b/src/tools/build_helper/src/lib.rs @@ -1,3 +1,4 @@ pub mod ci; pub mod git; +pub mod metrics; pub mod util; diff --git a/src/tools/build_helper/src/metrics.rs b/src/tools/build_helper/src/metrics.rs new file mode 100644 index 00000000000..2d0c66a8f33 --- /dev/null +++ b/src/tools/build_helper/src/metrics.rs @@ -0,0 +1,92 @@ +use serde_derive::{Deserialize, Serialize}; + +#[derive(Serialize, Deserialize)] +#[serde(rename_all = "snake_case")] +pub struct JsonRoot { + #[serde(default)] // For version 0 the field was not present. + pub format_version: usize, + pub system_stats: JsonInvocationSystemStats, + pub invocations: Vec<JsonInvocation>, +} + +#[derive(Serialize, Deserialize)] +#[serde(rename_all = "snake_case")] +pub struct JsonInvocation { + // Unix timestamp in seconds + // + // This is necessary to easily correlate this invocation with logs or other data. + pub start_time: u64, + pub duration_including_children_sec: f64, + pub children: Vec<JsonNode>, +} + +#[derive(Serialize, Deserialize)] +#[serde(tag = "kind", rename_all = "snake_case")] +pub enum JsonNode { + RustbuildStep { + #[serde(rename = "type")] + type_: String, + debug_repr: String, + + duration_excluding_children_sec: f64, + system_stats: JsonStepSystemStats, + + children: Vec<JsonNode>, + }, + TestSuite(TestSuite), +} + +#[derive(Serialize, Deserialize)] +pub struct TestSuite { + pub metadata: TestSuiteMetadata, + pub tests: Vec<Test>, +} + +#[derive(Serialize, Deserialize)] +#[serde(tag = "kind", rename_all = "snake_case")] +pub enum TestSuiteMetadata { + CargoPackage { + crates: Vec<String>, + target: String, + host: String, + stage: u32, + }, + Compiletest { + suite: String, + mode: String, + compare_mode: Option<String>, + target: String, + host: String, + stage: u32, + }, +} + +#[derive(Serialize, Deserialize)] +pub struct Test { + pub name: String, + #[serde(flatten)] + pub outcome: TestOutcome, +} + +#[derive(Serialize, Deserialize)] +#[serde(tag = "outcome", rename_all = "snake_case")] +pub enum TestOutcome { + Passed, + Failed, + Ignored { ignore_reason: Option<String> }, +} + +#[derive(Serialize, Deserialize)] +#[serde(rename_all = "snake_case")] +pub struct JsonInvocationSystemStats { + pub cpu_threads_count: usize, + pub cpu_model: String, + + pub memory_total_bytes: u64, +} + +#[derive(Serialize, Deserialize)] +#[serde(rename_all = "snake_case")] +pub struct JsonStepSystemStats { + pub cpu_utilization_percent: f64, +} diff --git a/src/tools/cargo b/src/tools/cargo -Subproject 5b377cece0e0dd0af686cf53ce4637d5d85c2a1 +Subproject 45782b6b8afd1da042d45c2daeec9c0744f72cc diff --git a/src/tools/clippy/clippy_lints/src/derivable_impls.rs b/src/tools/clippy/clippy_lints/src/derivable_impls.rs index ec0ca50cfec..020ffe7f8fa 100644 --- a/src/tools/clippy/clippy_lints/src/derivable_impls.rs +++ b/src/tools/clippy/clippy_lints/src/derivable_impls.rs @@ -9,7 +9,7 @@ use rustc_hir::{ Body, Expr, ExprKind, GenericArg, Impl, ImplItemKind, Item, ItemKind, Node, PathSegment, QPath, TyKind, }; use rustc_lint::{LateContext, LateLintPass}; -use rustc_middle::ty::adjustment::{Adjust, PointerCast}; +use rustc_middle::ty::adjustment::{Adjust, PointerCoercion}; use rustc_middle::ty::{self, Adt, AdtDef, SubstsRef, Ty, TypeckResults}; use rustc_session::{declare_tool_lint, impl_lint_pass}; use rustc_span::sym; @@ -116,7 +116,7 @@ fn check_struct<'tcx>( let is_default_without_adjusts = |expr| { is_default_equivalent(cx, expr) && typeck_results.expr_adjustments(expr).iter().all(|adj| { - !matches!(adj.kind, Adjust::Pointer(PointerCast::Unsize) + !matches!(adj.kind, Adjust::Pointer(PointerCoercion::Unsize) if contains_trait_object(adj.target)) }) }; diff --git a/src/tools/clippy/clippy_lints/src/pass_by_ref_or_value.rs b/src/tools/clippy/clippy_lints/src/pass_by_ref_or_value.rs index 0d78c3048ba..eab725de17f 100644 --- a/src/tools/clippy/clippy_lints/src/pass_by_ref_or_value.rs +++ b/src/tools/clippy/clippy_lints/src/pass_by_ref_or_value.rs @@ -14,7 +14,7 @@ use rustc_hir as hir; use rustc_hir::intravisit::FnKind; use rustc_hir::{BindingAnnotation, Body, FnDecl, Impl, ItemKind, MutTy, Mutability, Node, PatKind}; use rustc_lint::{LateContext, LateLintPass}; -use rustc_middle::ty::adjustment::{Adjust, PointerCast}; +use rustc_middle::ty::adjustment::{Adjust, PointerCoercion}; use rustc_middle::ty::layout::LayoutOf; use rustc_middle::ty::{self, RegionKind}; use rustc_session::{declare_tool_lint, impl_lint_pass}; @@ -195,7 +195,7 @@ impl<'tcx> PassByRefOrValue { .adjustments() .items() .flat_map(|(_, a)| a) - .any(|a| matches!(a.kind, Adjust::Pointer(PointerCast::UnsafeFnPointer))) + .any(|a| matches!(a.kind, Adjust::Pointer(PointerCoercion::UnsafeFnPointer))) { continue; } diff --git a/src/tools/clippy/clippy_utils/src/qualify_min_const_fn.rs b/src/tools/clippy/clippy_utils/src/qualify_min_const_fn.rs index c9938caefb0..fbf4ab2722e 100644 --- a/src/tools/clippy/clippy_utils/src/qualify_min_const_fn.rs +++ b/src/tools/clippy/clippy_utils/src/qualify_min_const_fn.rs @@ -16,7 +16,7 @@ use rustc_middle::mir::{ }; use rustc_middle::traits::{ImplSource, ObligationCause}; use rustc_middle::ty::subst::GenericArgKind; -use rustc_middle::ty::{self, adjustment::PointerCast, Ty, TyCtxt}; +use rustc_middle::ty::{self, adjustment::PointerCoercion, Ty, TyCtxt}; use rustc_middle::ty::{BoundConstness, TraitRef}; use rustc_semver::RustcVersion; use rustc_span::symbol::sym; @@ -119,18 +119,18 @@ fn check_rvalue<'tcx>( | CastKind::FloatToFloat | CastKind::FnPtrToPtr | CastKind::PtrToPtr - | CastKind::Pointer(PointerCast::MutToConstPointer | PointerCast::ArrayToPointer), + | CastKind::PointerCoercion(PointerCoercion::MutToConstPointer | PointerCoercion::ArrayToPointer), operand, _, ) => check_operand(tcx, operand, span, body), Rvalue::Cast( - CastKind::Pointer( - PointerCast::UnsafeFnPointer | PointerCast::ClosureFnPointer(_) | PointerCast::ReifyFnPointer, + CastKind::PointerCoercion( + PointerCoercion::UnsafeFnPointer | PointerCoercion::ClosureFnPointer(_) | PointerCoercion::ReifyFnPointer, ), _, _, ) => Err((span, "function pointer casts are not allowed in const fn".into())), - Rvalue::Cast(CastKind::Pointer(PointerCast::Unsize), op, cast_ty) => { + Rvalue::Cast(CastKind::PointerCoercion(PointerCoercion::Unsize), op, cast_ty) => { let pointee_ty = if let Some(deref_ty) = cast_ty.builtin_deref(true) { deref_ty.ty } else { diff --git a/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.rs b/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.rs deleted file mode 100644 index 67115f7c5a7..00000000000 --- a/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.rs +++ /dev/null @@ -1,9 +0,0 @@ -//@run-rustfix -//@compile-flags: -C incremental=target/debug/test/incr - -// see https://github.com/rust-lang/rust-clippy/issues/10969 - -fn main() { - let s = "Hello, world!"; - println!("{}", s.to_string()); -} diff --git a/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.stderr b/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.stderr deleted file mode 100644 index a992c542914..00000000000 --- a/src/tools/clippy/tests/ui/to_string_in_format_args_incremental.stderr +++ /dev/null @@ -1,10 +0,0 @@ -error: `to_string` applied to a type that implements `Display` in `println!` args - --> $DIR/to_string_in_format_args_incremental.rs:8:21 - | -LL | println!("{}", s.to_string()); - | ^^^^^^^^^^^^ help: remove this - | - = note: `-D clippy::to-string-in-format-args` implied by `-D warnings` - -error: aborting due to previous error - diff --git a/src/tools/compiletest/src/header.rs b/src/tools/compiletest/src/header.rs index c835962ad12..ad10c3e07ce 100644 --- a/src/tools/compiletest/src/header.rs +++ b/src/tools/compiletest/src/header.rs @@ -588,21 +588,22 @@ impl TestProps { } } +/// Extract a `(Option<line_config>, directive)` directive from a line if comment is present. pub fn line_directive<'line>( comment: &str, ln: &'line str, ) -> Option<(Option<&'line str>, &'line str)> { + let ln = ln.trim_start(); if ln.starts_with(comment) { let ln = ln[comment.len()..].trim_start(); if ln.starts_with('[') { // A comment like `//[foo]` is specific to revision `foo` - if let Some(close_brace) = ln.find(']') { - let lncfg = &ln[1..close_brace]; + let Some(close_brace) = ln.find(']') else { + panic!("malformed condition directive: expected `{}[foo]`, found `{}`", comment, ln); + }; - Some((Some(lncfg), ln[(close_brace + 1)..].trim_start())) - } else { - panic!("malformed condition directive: expected `{}[foo]`, found `{}`", comment, ln) - } + let lncfg = &ln[1..close_brace]; + Some((Some(lncfg), ln[(close_brace + 1)..].trim_start())) } else { Some((None, ln)) } diff --git a/src/tools/compiletest/src/runtest.rs b/src/tools/compiletest/src/runtest.rs index 672779325ae..ac19fe078f0 100644 --- a/src/tools/compiletest/src/runtest.rs +++ b/src/tools/compiletest/src/runtest.rs @@ -41,7 +41,7 @@ use crate::extract_gdb_version; use crate::is_android_gdb_target; mod debugger; -use debugger::{check_debugger_output, DebuggerCommands}; +use debugger::DebuggerCommands; #[cfg(test)] mod tests; @@ -997,16 +997,13 @@ impl<'test> TestCx<'test> { }; // Parse debugger commands etc from test files - let DebuggerCommands { commands, check_lines, breakpoint_lines, .. } = - match DebuggerCommands::parse_from( - &self.testpaths.file, - self.config, - prefixes, - self.revision, - ) { - Ok(cmds) => cmds, - Err(e) => self.fatal(&e), - }; + let dbg_cmds = DebuggerCommands::parse_from( + &self.testpaths.file, + self.config, + prefixes, + self.revision, + ) + .unwrap_or_else(|e| self.fatal(&e)); // https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/debugger-commands let mut script_str = String::with_capacity(2048); @@ -1023,12 +1020,12 @@ impl<'test> TestCx<'test> { // Set breakpoints on every line that contains the string "#break" let source_file_name = self.testpaths.file.file_name().unwrap().to_string_lossy(); - for line in &breakpoint_lines { + for line in &dbg_cmds.breakpoint_lines { script_str.push_str(&format!("bp `{}:{}`\n", source_file_name, line)); } // Append the other `cdb-command:`s - for line in &commands { + for line in &dbg_cmds.commands { script_str.push_str(line); script_str.push_str("\n"); } @@ -1058,7 +1055,7 @@ impl<'test> TestCx<'test> { self.fatal_proc_rec("Error while running CDB", &debugger_run_result); } - if let Err(e) = check_debugger_output(&debugger_run_result, &check_lines) { + if let Err(e) = dbg_cmds.check_output(&debugger_run_result) { self.fatal_proc_rec(&e, &debugger_run_result); } } @@ -1088,17 +1085,14 @@ impl<'test> TestCx<'test> { PREFIXES }; - let DebuggerCommands { commands, check_lines, breakpoint_lines } = - match DebuggerCommands::parse_from( - &self.testpaths.file, - self.config, - prefixes, - self.revision, - ) { - Ok(cmds) => cmds, - Err(e) => self.fatal(&e), - }; - let mut cmds = commands.join("\n"); + let dbg_cmds = DebuggerCommands::parse_from( + &self.testpaths.file, + self.config, + prefixes, + self.revision, + ) + .unwrap_or_else(|e| self.fatal(&e)); + let mut cmds = dbg_cmds.commands.join("\n"); // compile test file (it should have 'compile-flags:-g' in the header) let should_run = self.run_if_enabled(); @@ -1132,13 +1126,14 @@ impl<'test> TestCx<'test> { ./{}/stage2/lib/rustlib/{}/lib/\n", self.config.host, self.config.target )); - for line in &breakpoint_lines { + for line in &dbg_cmds.breakpoint_lines { script_str.push_str( - &format!( + format!( "break {:?}:{}\n", self.testpaths.file.file_name().unwrap().to_string_lossy(), *line - )[..], + ) + .as_str(), ); } script_str.push_str(&cmds); @@ -1279,7 +1274,7 @@ impl<'test> TestCx<'test> { } // Add line breakpoints - for line in &breakpoint_lines { + for line in &dbg_cmds.breakpoint_lines { script_str.push_str(&format!( "break '{}':{}\n", self.testpaths.file.file_name().unwrap().to_string_lossy(), @@ -1315,7 +1310,7 @@ impl<'test> TestCx<'test> { self.fatal_proc_rec("gdb failed to execute", &debugger_run_result); } - if let Err(e) = check_debugger_output(&debugger_run_result, &check_lines) { + if let Err(e) = dbg_cmds.check_output(&debugger_run_result) { self.fatal_proc_rec(&e, &debugger_run_result); } } @@ -1372,16 +1367,13 @@ impl<'test> TestCx<'test> { }; // Parse debugger commands etc from test files - let DebuggerCommands { commands, check_lines, breakpoint_lines, .. } = - match DebuggerCommands::parse_from( - &self.testpaths.file, - self.config, - prefixes, - self.revision, - ) { - Ok(cmds) => cmds, - Err(e) => self.fatal(&e), - }; + let dbg_cmds = DebuggerCommands::parse_from( + &self.testpaths.file, + self.config, + prefixes, + self.revision, + ) + .unwrap_or_else(|e| self.fatal(&e)); // Write debugger script: // We don't want to hang when calling `quit` while the process is still running @@ -1430,7 +1422,7 @@ impl<'test> TestCx<'test> { // Set breakpoints on every line that contains the string "#break" let source_file_name = self.testpaths.file.file_name().unwrap().to_string_lossy(); - for line in &breakpoint_lines { + for line in &dbg_cmds.breakpoint_lines { script_str.push_str(&format!( "breakpoint set --file '{}' --line {}\n", source_file_name, line @@ -1438,7 +1430,7 @@ impl<'test> TestCx<'test> { } // Append the other commands - for line in &commands { + for line in &dbg_cmds.commands { script_str.push_str(line); script_str.push_str("\n"); } @@ -1458,7 +1450,7 @@ impl<'test> TestCx<'test> { self.fatal_proc_rec("Error while running LLDB", &debugger_run_result); } - if let Err(e) = check_debugger_output(&debugger_run_result, &check_lines) { + if let Err(e) = dbg_cmds.check_output(&debugger_run_result) { self.fatal_proc_rec(&e, &debugger_run_result); } } diff --git a/src/tools/compiletest/src/runtest/debugger.rs b/src/tools/compiletest/src/runtest/debugger.rs index 379ff0bab40..eebe5f3580b 100644 --- a/src/tools/compiletest/src/runtest/debugger.rs +++ b/src/tools/compiletest/src/runtest/debugger.rs @@ -2,18 +2,25 @@ use crate::common::Config; use crate::header::line_directive; use crate::runtest::ProcRes; +use std::fmt::Write; use std::fs::File; use std::io::{BufRead, BufReader}; -use std::path::Path; +use std::path::{Path, PathBuf}; +/// Representation of information to invoke a debugger and check its output pub(super) struct DebuggerCommands { + /// Commands for the debuuger pub commands: Vec<String>, - pub check_lines: Vec<String>, + /// Lines to insert breakpoints at pub breakpoint_lines: Vec<usize>, + /// Contains the source line number to check and the line itself + check_lines: Vec<(usize, String)>, + /// Source file name + file: PathBuf, } impl DebuggerCommands { - pub(super) fn parse_from( + pub fn parse_from( file: &Path, config: &Config, debugger_prefixes: &[&str], @@ -21,7 +28,7 @@ impl DebuggerCommands { ) -> Result<Self, String> { let directives = debugger_prefixes .iter() - .map(|prefix| (format!("{}-command", prefix), format!("{}-check", prefix))) + .map(|prefix| (format!("{prefix}-command"), format!("{prefix}-check"))) .collect::<Vec<_>>(); let mut breakpoint_lines = vec![]; @@ -29,63 +36,88 @@ impl DebuggerCommands { let mut check_lines = vec![]; let mut counter = 0; let reader = BufReader::new(File::open(file).unwrap()); - for line in reader.lines() { + for (line_no, line) in reader.lines().enumerate() { counter += 1; - match line { - Ok(line) => { - let (lnrev, line) = line_directive("//", &line).unwrap_or((None, &line)); - - // Skip any revision specific directive that doesn't match the current - // revision being tested - if lnrev.is_some() && lnrev != rev { - continue; - } - - if line.contains("#break") { - breakpoint_lines.push(counter); - } - - for &(ref command_directive, ref check_directive) in &directives { - config - .parse_name_value_directive(&line, command_directive) - .map(|cmd| commands.push(cmd)); - - config - .parse_name_value_directive(&line, check_directive) - .map(|cmd| check_lines.push(cmd)); - } - } - Err(e) => return Err(format!("Error while parsing debugger commands: {}", e)), + let line = line.map_err(|e| format!("Error while parsing debugger commands: {}", e))?; + let (lnrev, line) = line_directive("//", &line).unwrap_or((None, &line)); + + // Skip any revision specific directive that doesn't match the current + // revision being tested + if lnrev.is_some() && lnrev != rev { + continue; + } + + if line.contains("#break") { + breakpoint_lines.push(counter); + } + + for &(ref command_directive, ref check_directive) in &directives { + config + .parse_name_value_directive(&line, command_directive) + .map(|cmd| commands.push(cmd)); + + config + .parse_name_value_directive(&line, check_directive) + .map(|cmd| check_lines.push((line_no, cmd))); } } - Ok(Self { commands, check_lines, breakpoint_lines }) + Ok(Self { commands, breakpoint_lines, check_lines, file: file.to_owned() }) } -} -pub(super) fn check_debugger_output( - debugger_run_result: &ProcRes, - check_lines: &[String], -) -> Result<(), String> { - let num_check_lines = check_lines.len(); - - let mut check_line_index = 0; - for line in debugger_run_result.stdout.lines() { - if check_line_index >= num_check_lines { - break; + /// Given debugger output and lines to check, ensure that every line is + /// contained in the debugger output. The check lines need to be found in + /// order, but there can be extra lines between. + pub fn check_output(&self, debugger_run_result: &ProcRes) -> Result<(), String> { + // (src_lineno, ck_line) that we did find + let mut found = vec![]; + // (src_lineno, ck_line) that we couldn't find + let mut missing = vec![]; + // We can find our any current match anywhere after our last match + let mut last_idx = 0; + let dbg_lines: Vec<&str> = debugger_run_result.stdout.lines().collect(); + + for (src_lineno, ck_line) in &self.check_lines { + if let Some(offset) = dbg_lines + .iter() + .skip(last_idx) + .position(|out_line| check_single_line(out_line, &ck_line)) + { + last_idx += offset; + found.push((src_lineno, dbg_lines[last_idx])); + } else { + missing.push((src_lineno, ck_line)); + } } - if check_single_line(line, &(check_lines[check_line_index])[..]) { - check_line_index += 1; + if missing.is_empty() { + Ok(()) + } else { + let fname = self.file.file_name().unwrap().to_string_lossy(); + let mut msg = format!( + "check directive(s) from `{}` not found in debugger output. errors:", + self.file.display() + ); + + for (src_lineno, err_line) in missing { + write!(msg, "\n ({fname}:{num}) `{err_line}`", num = src_lineno + 1).unwrap(); + } + + if !found.is_empty() { + let init = "\nthe following subset of check directive(s) was found successfully:"; + msg.push_str(init); + for (src_lineno, found_line) in found { + write!(msg, "\n ({fname}:{num}) `{found_line}`", num = src_lineno + 1) + .unwrap(); + } + } + + Err(msg) } } - if check_line_index != num_check_lines && num_check_lines > 0 { - Err(format!("line not found in debugger output: {}", check_lines[check_line_index])) - } else { - Ok(()) - } } +/// Check that the pattern in `check_line` applies to `line`. Returns `true` if they do match. fn check_single_line(line: &str, check_line: &str) -> bool { // Allow check lines to leave parts unspecified (e.g., uninitialized // bits in the wrong case of an enum) with the notation "[...]". @@ -101,21 +133,19 @@ fn check_single_line(line: &str, check_line: &str) -> bool { } let (mut rest, first_fragment) = if can_start_anywhere { - match line.find(check_fragments[0]) { - Some(pos) => (&line[pos + check_fragments[0].len()..], 1), - None => return false, - } + let Some(pos) = line.find(check_fragments[0]) else { + return false; + }; + (&line[pos + check_fragments[0].len()..], 1) } else { (line, 0) }; for current_fragment in &check_fragments[first_fragment..] { - match rest.find(current_fragment) { - Some(pos) => { - rest = &rest[pos + current_fragment.len()..]; - } - None => return false, - } + let Some(pos) = rest.find(current_fragment) else { + return false; + }; + rest = &rest[pos + current_fragment.len()..]; } if !can_end_anywhere && !rest.is_empty() { false } else { true } diff --git a/src/tools/miri/tests/fail/overlapping_assignment.rs b/src/tools/miri/tests/fail/overlapping_assignment.rs new file mode 100644 index 00000000000..84994c179f9 --- /dev/null +++ b/src/tools/miri/tests/fail/overlapping_assignment.rs @@ -0,0 +1,23 @@ +#![feature(core_intrinsics)] +#![feature(custom_mir)] + +use std::intrinsics::mir::*; + +// It's not that easy to fool the MIR validity check +// which wants to prevent overlapping assignments... +// So we use two separate pointer arguments, and then arrange for them to alias. +#[custom_mir(dialect = "runtime", phase = "optimized")] +pub fn self_copy(ptr1: *mut [i32; 4], ptr2: *mut [i32; 4]) { + mir! { + { + *ptr1 = *ptr2; //~ERROR: overlapping ranges + Return() + } + } +} + +pub fn main() { + let mut x = [0; 4]; + let ptr = std::ptr::addr_of_mut!(x); + self_copy(ptr, ptr); +} diff --git a/src/tools/miri/tests/fail/overlapping_assignment.stderr b/src/tools/miri/tests/fail/overlapping_assignment.stderr new file mode 100644 index 00000000000..42a000dfcc6 --- /dev/null +++ b/src/tools/miri/tests/fail/overlapping_assignment.stderr @@ -0,0 +1,20 @@ +error: Undefined Behavior: `copy_nonoverlapping` called on overlapping ranges + --> $DIR/overlapping_assignment.rs:LL:CC + | +LL | *ptr1 = *ptr2; + | ^^^^^^^^^^^^^ `copy_nonoverlapping` called on overlapping ranges + | + = help: this indicates a bug in the program: it performed an invalid operation, and caused Undefined Behavior + = help: see https://doc.rust-lang.org/nightly/reference/behavior-considered-undefined.html for further information + = note: BACKTRACE: + = note: inside `self_copy` at $DIR/overlapping_assignment.rs:LL:CC +note: inside `main` + --> $DIR/overlapping_assignment.rs:LL:CC + | +LL | self_copy(ptr, ptr); + | ^^^^^^^^^^^^^^^^^^^ + +note: some details are omitted, run with `MIRIFLAGS=-Zmiri-backtrace=full` for a verbose backtrace + +error: aborting due to previous error + diff --git a/src/tools/opt-dist/Cargo.toml b/src/tools/opt-dist/Cargo.toml new file mode 100644 index 00000000000..5a1794d3336 --- /dev/null +++ b/src/tools/opt-dist/Cargo.toml @@ -0,0 +1,22 @@ +[package] +name = "opt-dist" +version = "0.1.0" +edition = "2021" + +[dependencies] +build_helper = { path = "../build_helper" } +env_logger = "0.10" +log = "0.4" +anyhow = { version = "1", features = ["backtrace"] } +humantime = "2" +humansize = "2" +sysinfo = { version = "0.29", default-features = false } +fs_extra = "1" +camino = "1" +reqwest = { version = "0.11", features = ["blocking"] } +zip = { version = "0.6", default-features = false, features = ["deflate"] } +tar = "0.4" +xz = "0.1" +serde = { version = "1", features = ["derive"] } +serde_json = "1" +glob = "0.3" diff --git a/src/tools/opt-dist/README.md b/src/tools/opt-dist/README.md new file mode 100644 index 00000000000..05a75870da1 --- /dev/null +++ b/src/tools/opt-dist/README.md @@ -0,0 +1,7 @@ +# Optimized build pipeline +This binary implements a heavily optimized build pipeline for `rustc` and `LLVM` artifacts that are used for both for +benchmarking using the perf. bot and for final distribution to users. + +It uses LTO, PGO and BOLT to optimize the compiler and LLVM as much as possible. +This logic is not part of bootstrap, because it needs to invoke bootstrap multiple times, force-rebuild various +artifacts repeatedly and sometimes go around bootstrap's cache mechanism. diff --git a/src/tools/opt-dist/src/environment/linux.rs b/src/tools/opt-dist/src/environment/linux.rs new file mode 100644 index 00000000000..d4c55c46f7c --- /dev/null +++ b/src/tools/opt-dist/src/environment/linux.rs @@ -0,0 +1,54 @@ +use crate::environment::Environment; +use crate::exec::cmd; +use crate::utils::io::copy_directory; +use camino::{Utf8Path, Utf8PathBuf}; + +pub(super) struct LinuxEnvironment; + +impl Environment for LinuxEnvironment { + fn python_binary(&self) -> &'static str { + "python3" + } + + fn checkout_path(&self) -> Utf8PathBuf { + Utf8PathBuf::from("/checkout") + } + + fn downloaded_llvm_dir(&self) -> Utf8PathBuf { + Utf8PathBuf::from("/rustroot") + } + + fn opt_artifacts(&self) -> Utf8PathBuf { + Utf8PathBuf::from("/tmp/tmp-multistage/opt-artifacts") + } + + fn build_root(&self) -> Utf8PathBuf { + self.checkout_path().join("obj") + } + + fn prepare_rustc_perf(&self) -> anyhow::Result<()> { + // /tmp/rustc-perf comes from the x64 dist Dockerfile + copy_directory(Utf8Path::new("/tmp/rustc-perf"), &self.rustc_perf_dir())?; + cmd(&[self.cargo_stage_0().as_str(), "build", "-p", "collector"]) + .workdir(&self.rustc_perf_dir()) + .env("RUSTC", &self.rustc_stage_0().into_string()) + .env("RUSTC_BOOTSTRAP", "1") + .run()?; + Ok(()) + } + + fn supports_bolt(&self) -> bool { + true + } + + fn executable_extension(&self) -> &'static str { + "" + } + + fn skipped_tests(&self) -> &'static [&'static str] { + &[ + // Fails because of linker errors, as of June 2023. + "tests/ui/process/nofile-limit.rs", + ] + } +} diff --git a/src/tools/opt-dist/src/environment/mod.rs b/src/tools/opt-dist/src/environment/mod.rs new file mode 100644 index 00000000000..f66b9ab41ea --- /dev/null +++ b/src/tools/opt-dist/src/environment/mod.rs @@ -0,0 +1,75 @@ +use camino::Utf8PathBuf; + +#[cfg(target_family = "unix")] +mod linux; +#[cfg(target_family = "windows")] +mod windows; + +pub trait Environment { + fn host_triple(&self) -> String { + std::env::var("PGO_HOST").expect("PGO_HOST environment variable missing") + } + + fn python_binary(&self) -> &'static str; + + /// The rustc checkout, where the compiler source is located. + fn checkout_path(&self) -> Utf8PathBuf; + + /// Path to the downloaded host LLVM. + fn downloaded_llvm_dir(&self) -> Utf8PathBuf; + + /// Directory where the optimization artifacts (PGO/BOLT profiles, etc.) + /// will be stored. + fn opt_artifacts(&self) -> Utf8PathBuf; + + /// The main directory where the build occurs. + fn build_root(&self) -> Utf8PathBuf; + + fn build_artifacts(&self) -> Utf8PathBuf { + self.build_root().join("build").join(self.host_triple()) + } + + fn cargo_stage_0(&self) -> Utf8PathBuf { + self.build_artifacts() + .join("stage0") + .join("bin") + .join(format!("cargo{}", self.executable_extension())) + } + + fn rustc_stage_0(&self) -> Utf8PathBuf { + self.build_artifacts() + .join("stage0") + .join("bin") + .join(format!("rustc{}", self.executable_extension())) + } + + fn rustc_stage_2(&self) -> Utf8PathBuf { + self.build_artifacts() + .join("stage2") + .join("bin") + .join(format!("rustc{}", self.executable_extension())) + } + + /// Path to the built rustc-perf benchmark suite. + fn rustc_perf_dir(&self) -> Utf8PathBuf { + self.opt_artifacts().join("rustc-perf") + } + + /// Download and/or compile rustc-perf. + fn prepare_rustc_perf(&self) -> anyhow::Result<()>; + + fn supports_bolt(&self) -> bool; + + /// What is the extension of binary executables in this environment? + fn executable_extension(&self) -> &'static str; + + /// List of test paths that should be skipped when testing the optimized artifacts. + fn skipped_tests(&self) -> &'static [&'static str]; +} + +pub fn create_environment() -> Box<dyn Environment> { + #[cfg(target_family = "unix")] + return Box::new(linux::LinuxEnvironment); + #[cfg(target_family = "windows")] + return Box::new(windows::WindowsEnvironment::new()); +} diff --git a/src/tools/opt-dist/src/environment/windows.rs b/src/tools/opt-dist/src/environment/windows.rs new file mode 100644 index 00000000000..36f4c0f29cc --- /dev/null +++ b/src/tools/opt-dist/src/environment/windows.rs @@ -0,0 +1,78 @@ +use crate::environment::Environment; +use crate::exec::cmd; +use crate::utils::io::move_directory; +use camino::Utf8PathBuf; +use std::io::Cursor; +use zip::ZipArchive; + +pub(super) struct WindowsEnvironment { + checkout_dir: Utf8PathBuf, +} + +impl WindowsEnvironment { + pub fn new() -> Self { + Self { checkout_dir: std::env::current_dir().unwrap().try_into().unwrap() } + } +} + +impl Environment for WindowsEnvironment { + fn python_binary(&self) -> &'static str { + "python" + } + + fn checkout_path(&self) -> Utf8PathBuf { + self.checkout_dir.clone() + } + + fn downloaded_llvm_dir(&self) -> Utf8PathBuf { + self.checkout_path().join("citools").join("clang-rust") + } + + fn opt_artifacts(&self) -> Utf8PathBuf { + self.checkout_path().join("opt-artifacts") + } + + fn build_root(&self) -> Utf8PathBuf { + self.checkout_path() + } + + fn prepare_rustc_perf(&self) -> anyhow::Result<()> { + // FIXME: add some mechanism for synchronization of this commit SHA with + // Linux (which builds rustc-perf in a Dockerfile) + // rustc-perf version from 2023-05-30 + const PERF_COMMIT: &str = "8b2ac3042e1ff2c0074455a0a3618adef97156b1"; + + let url = format!("https://github.com/rust-lang/rustc-perf/archive/{PERF_COMMIT}.zip"); + let response = reqwest::blocking::get(url)?.error_for_status()?.bytes()?.to_vec(); + + let mut archive = ZipArchive::new(Cursor::new(response))?; + archive.extract(self.rustc_perf_dir())?; + move_directory( + &self.rustc_perf_dir().join(format!("rustc-perf-{PERF_COMMIT}")), + &self.rustc_perf_dir(), + )?; + + cmd(&[self.cargo_stage_0().as_str(), "build", "-p", "collector"]) + .workdir(&self.rustc_perf_dir()) + .env("RUSTC", &self.rustc_stage_0().into_string()) + .env("RUSTC_BOOTSTRAP", "1") + .run()?; + + Ok(()) + } + + fn supports_bolt(&self) -> bool { + false + } + + fn executable_extension(&self) -> &'static str { + ".exe" + } + + fn skipped_tests(&self) -> &'static [&'static str] { + &[ + // Fails as of June 2023. + "tests\\codegen\\vec-shrink-panik.rs", + ] + } +} diff --git a/src/tools/opt-dist/src/exec.rs b/src/tools/opt-dist/src/exec.rs new file mode 100644 index 00000000000..d05ddbd4c0e --- /dev/null +++ b/src/tools/opt-dist/src/exec.rs @@ -0,0 +1,169 @@ +use crate::environment::Environment; +use crate::metrics::{load_metrics, record_metrics}; +use crate::timer::TimerSection; +use crate::training::{LlvmBoltProfile, LlvmPGOProfile, RustcPGOProfile}; +use camino::{Utf8Path, Utf8PathBuf}; +use std::collections::BTreeMap; +use std::fs::File; +use std::process::{Command, Stdio}; + +#[derive(Default)] +pub struct CmdBuilder { + args: Vec<String>, + env: BTreeMap<String, String>, + workdir: Option<Utf8PathBuf>, + output: Option<Utf8PathBuf>, +} + +impl CmdBuilder { + pub fn arg(mut self, arg: &str) -> Self { + self.args.push(arg.to_string()); + self + } + + pub fn env(mut self, name: &str, value: &str) -> Self { + self.env.insert(name.to_string(), value.to_string()); + self + } + + pub fn workdir(mut self, path: &Utf8Path) -> Self { + self.workdir = Some(path.to_path_buf()); + self + } + + pub fn redirect_output(mut self, path: Utf8PathBuf) -> Self { + self.output = Some(path); + self + } + + pub fn run(self) -> anyhow::Result<()> { + let mut cmd_str = String::new(); + cmd_str.push_str( + &self + .env + .iter() + .map(|(key, value)| format!("{key}={value}")) + .collect::<Vec<_>>() + .join(" "), + ); + if !self.env.is_empty() { + cmd_str.push(' '); + } + cmd_str.push_str(&self.args.join(" ")); + if let Some(ref path) = self.output { + cmd_str.push_str(&format!(" > {path:?}")); + } + cmd_str.push_str(&format!( + " [at {}]", + self.workdir + .clone() + .unwrap_or_else(|| std::env::current_dir().unwrap().try_into().unwrap()) + )); + log::info!("Executing `{cmd_str}`"); + + let mut cmd = Command::new(&self.args[0]); + cmd.stdin(Stdio::null()); + cmd.args(self.args.iter().skip(1)); + for (key, value) in &self.env { + cmd.env(key, value); + } + if let Some(ref output) = self.output { + cmd.stdout(File::create(output.clone().into_std_path_buf())?); + } + if let Some(ref workdir) = self.workdir { + cmd.current_dir(workdir.clone().into_std_path_buf()); + } + let exit_status = cmd.spawn()?.wait()?; + if !exit_status.success() { + Err(anyhow::anyhow!( + "Command {cmd_str} has failed with exit code {:?}", + exit_status.code(), + )) + } else { + Ok(()) + } + } +} + +pub fn cmd(args: &[&str]) -> CmdBuilder { + assert!(!args.is_empty()); + CmdBuilder { args: args.iter().map(|s| s.to_string()).collect(), ..Default::default() } +} + +pub struct Bootstrap { + cmd: CmdBuilder, + metrics_path: Utf8PathBuf, +} + +impl Bootstrap { + pub fn build(env: &dyn Environment) -> Self { + let metrics_path = env.build_root().join("build").join("metrics.json"); + let cmd = cmd(&[ + env.python_binary(), + env.checkout_path().join("x.py").as_str(), + "build", + "--target", + &env.host_triple(), + "--host", + &env.host_triple(), + "--stage", + "2", + "library/std", + ]) + .env("RUST_BACKTRACE", "full"); + Self { cmd, metrics_path } + } + + pub fn dist(env: &dyn Environment, dist_args: &[String]) -> Self { + let metrics_path = env.build_root().join("build").join("metrics.json"); + let cmd = cmd(&dist_args.iter().map(|arg| arg.as_str()).collect::<Vec<_>>()) + .env("RUST_BACKTRACE", "full"); + Self { cmd, metrics_path } + } + + pub fn llvm_pgo_instrument(mut self, profile_dir: &Utf8Path) -> Self { + self.cmd = self + .cmd + .arg("--llvm-profile-generate") + .env("LLVM_PROFILE_DIR", profile_dir.join("prof-%p").as_str()); + self + } + + pub fn llvm_pgo_optimize(mut self, profile: &LlvmPGOProfile) -> Self { + self.cmd = self.cmd.arg("--llvm-profile-use").arg(profile.0.as_str()); + self + } + + pub fn rustc_pgo_instrument(mut self, profile_dir: &Utf8Path) -> Self { + self.cmd = self.cmd.arg("--rust-profile-generate").arg(profile_dir.as_str()); + self + } + + pub fn rustc_pgo_optimize(mut self, profile: &RustcPGOProfile) -> Self { + self.cmd = self.cmd.arg("--rust-profile-use").arg(profile.0.as_str()); + self + } + + pub fn llvm_bolt_instrument(mut self) -> Self { + self.cmd = self.cmd.arg("--llvm-bolt-profile-generate"); + self + } + + pub fn llvm_bolt_optimize(mut self, profile: &LlvmBoltProfile) -> Self { + self.cmd = self.cmd.arg("--llvm-bolt-profile-use").arg(profile.0.as_str()); + self + } + + /// Do not rebuild rustc, and use a previously built rustc sysroot instead. + pub fn avoid_rustc_rebuild(mut self) -> Self { + self.cmd = self.cmd.arg("--keep-stage").arg("0").arg("--keep-stage").arg("1"); + self + } + + pub fn run(self, timer: &mut TimerSection) -> anyhow::Result<()> { + self.cmd.run()?; + let metrics = load_metrics(&self.metrics_path)?; + record_metrics(&metrics, timer); + Ok(()) + } +} diff --git a/src/tools/opt-dist/src/main.rs b/src/tools/opt-dist/src/main.rs new file mode 100644 index 00000000000..39fa7e1fb2a --- /dev/null +++ b/src/tools/opt-dist/src/main.rs @@ -0,0 +1,175 @@ +use anyhow::Context; +use log::LevelFilter; + +use crate::environment::{create_environment, Environment}; +use crate::exec::Bootstrap; +use crate::tests::run_tests; +use crate::timer::Timer; +use crate::training::{gather_llvm_bolt_profiles, gather_llvm_profiles, gather_rustc_profiles}; +use crate::utils::io::reset_directory; +use crate::utils::{clear_llvm_files, format_env_variables, print_free_disk_space}; + +mod environment; +mod exec; +mod metrics; +mod tests; +mod timer; +mod training; +mod utils; + +fn is_try_build() -> bool { + std::env::var("DIST_TRY_BUILD").unwrap_or_else(|_| "0".to_string()) != "0" +} + +fn execute_pipeline( + env: &dyn Environment, + timer: &mut Timer, + dist_args: Vec<String>, +) -> anyhow::Result<()> { + reset_directory(&env.opt_artifacts())?; + env.prepare_rustc_perf()?; + + // Stage 1: Build PGO instrumented rustc + // We use a normal build of LLVM, because gathering PGO profiles for LLVM and `rustc` at the + // same time can cause issues, because the host and in-tree LLVM versions can diverge. + let rustc_pgo_profile = timer.section("Stage 1 (Rustc PGO)", |stage| { + let rustc_profile_dir_root = env.opt_artifacts().join("rustc-pgo"); + + stage.section("Build PGO instrumented rustc and LLVM", |section| { + Bootstrap::build(env).rustc_pgo_instrument(&rustc_profile_dir_root).run(section) + })?; + + let profile = stage + .section("Gather profiles", |_| gather_rustc_profiles(env, &rustc_profile_dir_root))?; + print_free_disk_space()?; + + stage.section("Build PGO optimized rustc", |section| { + Bootstrap::build(env).rustc_pgo_optimize(&profile).run(section) + })?; + + Ok(profile) + })?; + + // Stage 2: Gather LLVM PGO profiles + // Here we build a PGO instrumented LLVM, reusing the previously PGO optimized rustc. + // Then we use the instrumented LLVM to gather LLVM PGO profiles. + let llvm_pgo_profile = timer.section("Stage 2 (LLVM PGO)", |stage| { + // Remove the previous, uninstrumented build of LLVM. + clear_llvm_files(env)?; + + let llvm_profile_dir_root = env.opt_artifacts().join("llvm-pgo"); + + stage.section("Build PGO instrumented LLVM", |section| { + Bootstrap::build(env) + .llvm_pgo_instrument(&llvm_profile_dir_root) + .avoid_rustc_rebuild() + .run(section) + })?; + + let profile = stage + .section("Gather profiles", |_| gather_llvm_profiles(env, &llvm_profile_dir_root))?; + + print_free_disk_space()?; + + // Proactively delete the instrumented artifacts, to avoid using them by accident in + // follow-up stages. + clear_llvm_files(env)?; + + Ok(profile) + })?; + + let llvm_bolt_profile = if env.supports_bolt() { + // Stage 3: Build BOLT instrumented LLVM + // We build a PGO optimized LLVM in this step, then instrument it with BOLT and gather BOLT profiles. + // Note that we don't remove LLVM artifacts after this step, so that they are reused in the final dist build. + // BOLT instrumentation is performed "on-the-fly" when the LLVM library is copied to the sysroot of rustc, + // therefore the LLVM artifacts on disk are not "tainted" with BOLT instrumentation and they can be reused. + timer.section("Stage 3 (LLVM BOLT)", |stage| { + stage.section("Build BOLT instrumented LLVM", |stage| { + Bootstrap::build(env) + .llvm_bolt_instrument() + .llvm_pgo_optimize(&llvm_pgo_profile) + .avoid_rustc_rebuild() + .run(stage) + })?; + + let profile = stage.section("Gather profiles", |_| gather_llvm_bolt_profiles(env))?; + print_free_disk_space()?; + + // LLVM is not being cleared here, we want to reuse the previous PGO-optimized build + + Ok(Some(profile)) + })? + } else { + None + }; + + let mut dist = Bootstrap::dist(env, &dist_args) + .llvm_pgo_optimize(&llvm_pgo_profile) + .rustc_pgo_optimize(&rustc_pgo_profile) + .avoid_rustc_rebuild(); + + if let Some(llvm_bolt_profile) = llvm_bolt_profile { + dist = dist.llvm_bolt_optimize(&llvm_bolt_profile); + } + + // Final stage: Assemble the dist artifacts + // The previous PGO optimized rustc build and PGO optimized LLVM builds should be reused. + timer.section("Stage 4 (final build)", |stage| dist.run(stage))?; + + // After dist has finished, run a subset of the test suite on the optimized artifacts to discover + // possible regressions. + // The tests are not executed for try builds, which can be in various broken states, so we don't + // want to gatekeep them with tests. + if !is_try_build() { + timer.section("Run tests", |_| run_tests(env))?; + } + + Ok(()) +} + +fn main() -> anyhow::Result<()> { + // Make sure that we get backtraces for easier debugging in CI + std::env::set_var("RUST_BACKTRACE", "1"); + + env_logger::builder() + .filter_level(LevelFilter::Info) + .format_timestamp_millis() + .parse_default_env() + .init(); + + let mut build_args: Vec<String> = std::env::args().skip(1).collect(); + log::info!("Running optimized build pipeline with args `{}`", build_args.join(" ")); + log::info!("Environment values\n{}", format_env_variables()); + + if let Ok(config) = std::fs::read_to_string("config.toml") { + log::info!("Contents of `config.toml`:\n{config}"); + } + + // Skip components that are not needed for try builds to speed them up + if is_try_build() { + log::info!("Skipping building of unimportant components for a try build"); + for target in [ + "rust-docs", + "rustc-docs", + "rust-docs-json", + "rust-analyzer", + "rustc-src", + "clippy", + "miri", + "rustfmt", + ] { + build_args.extend(["--exclude".to_string(), target.to_string()]); + } + } + + let mut timer = Timer::new(); + let env = create_environment(); + + let result = execute_pipeline(env.as_ref(), &mut timer, build_args); + log::info!("Timer results\n{}", timer.format_stats()); + + print_free_disk_space()?; + + result.context("Optimized build pipeline has failed") +} diff --git a/src/tools/opt-dist/src/metrics.rs b/src/tools/opt-dist/src/metrics.rs new file mode 100644 index 00000000000..cabe07eda32 --- /dev/null +++ b/src/tools/opt-dist/src/metrics.rs @@ -0,0 +1,106 @@ +use crate::timer::TimerSection; +use build_helper::metrics::{JsonNode, JsonRoot}; +use camino::Utf8Path; +use std::time::Duration; + +#[derive(Clone, Debug)] +pub struct BuildStep { + r#type: String, + children: Vec<BuildStep>, + duration: Duration, +} + +impl BuildStep { + pub fn find_all_by_type(&self, r#type: &str) -> Vec<&BuildStep> { + let mut result = Vec::new(); + self.find_by_type(r#type, &mut result); + result + } + fn find_by_type<'a>(&'a self, r#type: &str, result: &mut Vec<&'a BuildStep>) { + if self.r#type == r#type { + result.push(self); + } + for child in &self.children { + child.find_by_type(r#type, result); + } + } +} + +/// Loads the metrics of the most recent bootstrap execution from a metrics.json file. +pub fn load_metrics(path: &Utf8Path) -> anyhow::Result<BuildStep> { + let content = std::fs::read(path.as_std_path())?; + let mut metrics = serde_json::from_slice::<JsonRoot>(&content)?; + let invocation = metrics + .invocations + .pop() + .ok_or_else(|| anyhow::anyhow!("No bootstrap invocation found in metrics file"))?; + + fn parse(node: JsonNode) -> Option<BuildStep> { + match node { + JsonNode::RustbuildStep { + type_: kind, + children, + duration_excluding_children_sec, + .. + } => { + let children: Vec<_> = children.into_iter().filter_map(parse).collect(); + let children_duration = children.iter().map(|c| c.duration).sum::<Duration>(); + Some(BuildStep { + r#type: kind.to_string(), + children, + duration: children_duration + + Duration::from_secs_f64(duration_excluding_children_sec), + }) + } + JsonNode::TestSuite(_) => None, + } + } + + let duration = Duration::from_secs_f64(invocation.duration_including_children_sec); + let children: Vec<_> = invocation.children.into_iter().filter_map(parse).collect(); + Ok(BuildStep { r#type: "root".to_string(), children, duration }) +} + +/// Logs the individual metrics in a table and add Rustc and LLVM durations to the passed +/// timer. +pub fn record_metrics(metrics: &BuildStep, timer: &mut TimerSection) { + let llvm_steps = metrics.find_all_by_type("bootstrap::llvm::Llvm"); + let llvm_duration: Duration = llvm_steps.into_iter().map(|s| s.duration).sum(); + + let rustc_steps = metrics.find_all_by_type("bootstrap::compile::Rustc"); + let rustc_duration: Duration = rustc_steps.into_iter().map(|s| s.duration).sum(); + + // The LLVM step is part of the Rustc step + let rustc_duration = rustc_duration.saturating_sub(llvm_duration); + + if !llvm_duration.is_zero() { + timer.add_duration("LLVM", llvm_duration); + } + if !rustc_duration.is_zero() { + timer.add_duration("Rustc", rustc_duration); + } + + log_metrics(metrics); +} + +fn log_metrics(metrics: &BuildStep) { + use std::fmt::Write; + + let mut substeps: Vec<(u32, &BuildStep)> = Vec::new(); + + fn visit<'a>(step: &'a BuildStep, level: u32, substeps: &mut Vec<(u32, &'a BuildStep)>) { + substeps.push((level, step)); + for child in &step.children { + visit(child, level + 1, substeps); + } + } + + visit(metrics, 0, &mut substeps); + + let mut output = String::new(); + for (level, step) in substeps { + let label = format!("{}{}", ".".repeat(level as usize), step.r#type); + writeln!(output, "{label:<65}{:>8.2}s", step.duration.as_secs_f64()).unwrap(); + } + log::info!("Build step durations\n{output}"); +} diff --git a/src/tools/opt-dist/src/tests.rs b/src/tools/opt-dist/src/tests.rs new file mode 100644 index 00000000000..6b35b13e586 --- /dev/null +++ b/src/tools/opt-dist/src/tests.rs @@ -0,0 +1,101 @@ +use crate::environment::Environment; +use crate::exec::cmd; +use crate::utils::io::{copy_directory, unpack_archive}; +use anyhow::Context; +use camino::Utf8PathBuf; + +/// Run tests on optimized dist artifacts. +pub fn run_tests(env: &dyn Environment) -> anyhow::Result<()> { + // After `dist` is executed, we extract its archived components into a sysroot directory, + // and then use that extracted rustc as a stage0 compiler. + // Then we run a subset of tests using that compiler, to have a basic smoke test which checks + // whether the optimization pipeline hasn't broken something. + let build_dir = env.build_root().join("build"); + let dist_dir = build_dir.join("dist"); + let unpacked_dist_dir = build_dir.join("unpacked-dist"); + std::fs::create_dir_all(&unpacked_dist_dir)?; + + let extract_dist_dir = |name: &str| -> anyhow::Result<Utf8PathBuf> { + unpack_archive(&dist_dir.join(format!("{name}.tar.xz")), &unpacked_dist_dir)?; + let extracted_path = unpacked_dist_dir.join(name); + assert!(extracted_path.is_dir()); + Ok(extracted_path) + }; + let host_triple = env.host_triple(); + + // Extract rustc, libstd, cargo and src archives to create the optimized sysroot + let rustc_dir = extract_dist_dir(&format!("rustc-nightly-{host_triple}"))?.join("rustc"); + let libstd_dir = extract_dist_dir(&format!("rust-std-nightly-{host_triple}"))? + .join(format!("rust-std-{host_triple}")); + let cargo_dir = extract_dist_dir(&format!("cargo-nightly-{host_triple}"))?.join("cargo"); + let extracted_src_dir = extract_dist_dir("rust-src-nightly")?.join("rust-src"); + + // We need to manually copy libstd to the extracted rustc sysroot + copy_directory( + &libstd_dir.join("lib").join("rustlib").join(&host_triple).join("lib"), + &rustc_dir.join("lib").join("rustlib").join(&host_triple).join("lib"), + )?; + + // Extract sources - they aren't in the `rustc-nightly-{host}` tarball, so we need to manually copy libstd + // sources to the extracted sysroot. We need sources available so that `-Zsimulate-remapped-rust-src-base` + // works correctly. + copy_directory( + &extracted_src_dir.join("lib").join("rustlib").join("src"), + &rustc_dir.join("lib").join("rustlib").join("src"), + )?; + + let rustc_path = rustc_dir.join("bin").join(format!("rustc{}", env.executable_extension())); + assert!(rustc_path.is_file()); + let cargo_path = cargo_dir.join("bin").join(format!("cargo{}", env.executable_extension())); + assert!(cargo_path.is_file()); + + // Specify path to a LLVM config so that LLVM is not rebuilt. + // It doesn't really matter which LLVM config we choose, because no sysroot will be compiled. + let llvm_config = env + .build_artifacts() + .join("llvm") + .join("bin") + .join(format!("llvm-config{}", env.executable_extension())); + assert!(llvm_config.is_file()); + + let config_content = format!( + r#"profile = "user" +changelog-seen = 2 + +[build] +rustc = "{rustc}" +cargo = "{cargo}" + +[target.{host_triple}] +llvm-config = "{llvm_config}" +"#, + rustc = rustc_path.to_string().replace('\\', "/"), + cargo = cargo_path.to_string().replace('\\', "/"), + llvm_config = llvm_config.to_string().replace('\\', "/") + ); + log::info!("Using following `config.toml` for running tests:\n{config_content}"); + + // Simulate a stage 0 compiler with the extracted optimized dist artifacts. + std::fs::write("config.toml", config_content)?; + + let x_py = env.checkout_path().join("x.py"); + let mut args = vec![ + env.python_binary(), + x_py.as_str(), + "test", + "--stage", + "0", + "tests/assembly", + "tests/codegen", + "tests/codegen-units", + "tests/incremental", + "tests/mir-opt", + "tests/pretty", + "tests/run-pass-valgrind", + "tests/ui", + ]; + for test_path in env.skipped_tests() { + args.extend(["--exclude", test_path]); + } + cmd(&args).env("COMPILETEST_FORCE_STAGE0", "1").run().context("Cannot execute tests") +} diff --git a/src/tools/opt-dist/src/timer.rs b/src/tools/opt-dist/src/timer.rs new file mode 100644 index 00000000000..2b29ba8d59f --- /dev/null +++ b/src/tools/opt-dist/src/timer.rs @@ -0,0 +1,167 @@ +use std::ops::{Deref, DerefMut}; +use std::time::{Duration, SystemTime}; + +pub struct Timer { + root: TimerSection, +} + +impl Timer { + pub fn new() -> Self { + Timer { root: TimerSection::new(None) } + } + + pub fn format_stats(&self) -> String { + use std::fmt::Write; + + let mut items = Vec::new(); + for (name, child) in &self.root.children { + match child { + SectionEntry::SubSection(section) => { + section.collect_levels(0, name, &mut items); + } + SectionEntry::Duration(duration) => items.push((0, name, *duration)), + } + } + + let rows: Vec<(String, Duration)> = items + .into_iter() + .map(|(level, name, duration)| (format!("{}{name}:", " ".repeat(level)), duration)) + .collect(); + + let total_duration = self.total_duration(); + let total_duration_label = "Total duration:".to_string(); + + const SPACE_AFTER_LABEL: usize = 2; + let max_label_length = 16.max(rows.iter().map(|(label, _)| label.len()).max().unwrap_or(0)) + + SPACE_AFTER_LABEL; + + let table_width = max_label_length + 23; + let divider = "-".repeat(table_width); + + let mut output = String::new(); + writeln!(output, "{divider}").unwrap(); + for (label, duration) in rows { + let pct = (duration.as_millis() as f64 / total_duration.as_millis() as f64) * 100.0; + let duration_fmt = format!("{:>12.2}s ({pct:>5.2}%)", duration.as_secs_f64()); + writeln!(output, "{label:<0$} {duration_fmt}", max_label_length).unwrap(); + } + output.push('\n'); + + let total_duration = Duration::new(total_duration.as_secs(), 0); + let total_duration = format!( + "{:>1$}", + humantime::format_duration(total_duration).to_string(), + table_width - total_duration_label.len() + ); + writeln!(output, "{total_duration_label}{total_duration}").unwrap(); + + writeln!(output, "{divider}").unwrap(); + output + } +} + +impl Deref for Timer { + type Target = TimerSection; + + fn deref(&self) -> &Self::Target { + &self.root + } +} + +impl DerefMut for Timer { + fn deref_mut(&mut self) -> &mut Self::Target { + &mut self.root + } +} + +pub struct TimerSection { + name: Option<String>, + children: Vec<(String, SectionEntry)>, + duration_excluding_children: Duration, +} + +impl TimerSection { + pub fn new(name: Option<String>) -> Self { + TimerSection { + name, + children: Default::default(), + duration_excluding_children: Duration::ZERO, + } + } + + pub fn section<F: FnOnce(&mut TimerSection) -> anyhow::Result<R>, R>( + &mut self, + name: &str, + func: F, + ) -> anyhow::Result<R> { + let full_name = match &self.name { + Some(current_name) => { + format!("{current_name} > {name}") + } + None => name.to_string(), + }; + log::info!("Section `{full_name}` starts"); + let mut child = TimerSection { + name: Some(full_name.clone()), + children: Default::default(), + duration_excluding_children: Duration::ZERO, + }; + + let start = SystemTime::now(); + let result = func(&mut child); + let duration = start.elapsed().unwrap(); + + let msg = match result { + Ok(_) => "OK", + Err(_) => "FAIL", + }; + + child.duration_excluding_children = duration.saturating_sub(child.total_duration()); + + log::info!("Section `{full_name}` ended: {msg} ({:.2}s)`", duration.as_secs_f64()); + self.children.push((name.to_string(), SectionEntry::SubSection(child))); + result + } + + pub fn add_duration(&mut self, name: &str, duration: Duration) { + self.children.push((name.to_string(), SectionEntry::Duration(duration))); + } + + fn total_duration(&self) -> Duration { + self.duration_excluding_children + + self.children.iter().map(|(_, child)| child.total_duration()).sum::<Duration>() + } + + fn collect_levels<'a>( + &'a self, + level: usize, + name: &'a str, + items: &mut Vec<(usize, &'a str, Duration)>, + ) { + items.push((level, name, self.total_duration())); + for (name, child) in &self.children { + match &child { + SectionEntry::Duration(duration) => { + items.push((level + 1, name, *duration)); + } + SectionEntry::SubSection(section) => { + section.collect_levels(level + 1, name, items); + } + } + } + } +} + +enum SectionEntry { + Duration(Duration), + SubSection(TimerSection), +} + +impl SectionEntry { + fn total_duration(&self) -> Duration { + match self { + SectionEntry::Duration(duration) => *duration, + SectionEntry::SubSection(timer) => timer.total_duration(), + } + } +} diff --git a/src/tools/opt-dist/src/training.rs b/src/tools/opt-dist/src/training.rs new file mode 100644 index 00000000000..10f4a603695 --- /dev/null +++ b/src/tools/opt-dist/src/training.rs @@ -0,0 +1,202 @@ +use crate::environment::Environment; +use crate::exec::{cmd, CmdBuilder}; +use crate::utils::io::{count_files, delete_directory}; +use anyhow::Context; +use camino::{Utf8Path, Utf8PathBuf}; +use humansize::BINARY; + +const LLVM_PGO_CRATES: &[&str] = &[ + "syn-1.0.89", + "cargo-0.60.0", + "serde-1.0.136", + "ripgrep-13.0.0", + "regex-1.5.5", + "clap-3.1.6", + "hyper-0.14.18", +]; + +const RUSTC_PGO_CRATES: &[&str] = &[ + "externs", + "ctfe-stress-5", + "cargo-0.60.0", + "token-stream-stress", + "match-stress", + "tuple-stress", + "diesel-1.4.8", + "bitmaps-3.1.0", +]; + +const LLVM_BOLT_CRATES: &[&str] = LLVM_PGO_CRATES; + +fn init_compiler_benchmarks( + env: &dyn Environment, + profiles: &[&str], + scenarios: &[&str], + crates: &[&str], +) -> CmdBuilder { + // Run rustc-perf benchmarks + // Benchmark using profile_local with eprintln, which essentially just means + // don't actually benchmark -- just make sure we run rustc a bunch of times. + cmd(&[ + env.cargo_stage_0().as_str(), + "run", + "-p", + "collector", + "--bin", + "collector", + "--", + "profile_local", + "eprintln", + env.rustc_stage_2().as_str(), + "--id", + "Test", + "--cargo", + env.cargo_stage_0().as_str(), + "--profiles", + profiles.join(",").as_str(), + "--scenarios", + scenarios.join(",").as_str(), + "--include", + crates.join(",").as_str(), + ]) + .env("RUST_LOG", "collector=debug") + .env("RUSTC", env.rustc_stage_0().as_str()) + .env("RUSTC_BOOTSTRAP", "1") + .workdir(&env.rustc_perf_dir()) +} + +fn merge_llvm_profiles( + env: &dyn Environment, + merged_path: &Utf8Path, + profile_dir: &Utf8Path, +) -> anyhow::Result<()> { + cmd(&[ + env.downloaded_llvm_dir().join("bin/llvm-profdata").as_str(), + "merge", + "-o", + merged_path.as_str(), + profile_dir.as_str(), + ]) + .run() + .context("Cannot merge LLVM profiles")?; + Ok(()) +} + +fn log_profile_stats( + name: &str, + merged_profile: &Utf8Path, + profile_root: &Utf8Path, +) -> anyhow::Result<()> { + log::info!("{name} PGO statistics"); + log::info!( + "{merged_profile}: {}", + humansize::format_size(std::fs::metadata(merged_profile.as_std_path())?.len(), BINARY) + ); + log::info!( + "{profile_root}: {}", + humansize::format_size(fs_extra::dir::get_size(profile_root.as_std_path())?, BINARY) + ); + log::info!("Profile file count: {}", count_files(profile_root)?); + Ok(()) +} + +pub struct LlvmPGOProfile(pub Utf8PathBuf); + +pub fn gather_llvm_profiles( + env: &dyn Environment, + profile_root: &Utf8Path, +) -> anyhow::Result<LlvmPGOProfile> { + log::info!("Running benchmarks with PGO instrumented LLVM"); + + init_compiler_benchmarks(env, &["Debug", "Opt"], &["Full"], LLVM_PGO_CRATES) + .run() + .context("Cannot gather LLVM PGO profiles")?; + + let merged_profile = env.opt_artifacts().join("llvm-pgo.profdata"); + log::info!("Merging LLVM PGO profiles to {merged_profile}"); + + merge_llvm_profiles(env, &merged_profile, profile_root)?; + log_profile_stats("LLVM", &merged_profile, profile_root)?; + + // We don't need the individual .profraw files now that they have been merged + // into a final .profdata + delete_directory(profile_root)?; + + Ok(LlvmPGOProfile(merged_profile)) +} + +pub struct RustcPGOProfile(pub Utf8PathBuf); + +pub fn gather_rustc_profiles( + env: &dyn Environment, + profile_root: &Utf8Path, +) -> anyhow::Result<RustcPGOProfile> { + log::info!("Running benchmarks with PGO instrumented rustc"); + + // The profile data is written into a single filepath that is being repeatedly merged when each + // rustc invocation ends. Empirically, this can result in some profiling data being lost. That's + // why we override the profile path to include the PID. This will produce many more profiling + // files, but the resulting profile will produce a slightly faster rustc binary. + let profile_template = profile_root.join("default_%m_%p.profraw"); + + // Here we're profiling the `rustc` frontend, so we also include `Check`. + // The benchmark set includes various stress tests that put the frontend under pressure. + init_compiler_benchmarks(env, &["Check", "Debug", "Opt"], &["All"], RUSTC_PGO_CRATES) + .env("LLVM_PROFILE_FILE", profile_template.as_str()) + .run() + .context("Cannot gather rustc PGO profiles")?; + + let merged_profile = env.opt_artifacts().join("rustc-pgo.profdata"); + log::info!("Merging Rustc PGO profiles to {merged_profile}"); + + merge_llvm_profiles(env, &merged_profile, profile_root)?; + log_profile_stats("Rustc", &merged_profile, profile_root)?; + + // We don't need the individual .profraw files now that they have been merged + // into a final .profdata + delete_directory(profile_root)?; + + Ok(RustcPGOProfile(merged_profile)) +} + +pub struct LlvmBoltProfile(pub Utf8PathBuf); + +pub fn gather_llvm_bolt_profiles(env: &dyn Environment) -> anyhow::Result<LlvmBoltProfile> { + log::info!("Running benchmarks with BOLT instrumented LLVM"); + + init_compiler_benchmarks(env, &["Check", "Debug", "Opt"], &["Full"], LLVM_BOLT_CRATES) + .run() + .context("Cannot gather LLVM BOLT profiles")?; + + let merged_profile = env.opt_artifacts().join("bolt.profdata"); + let profile_root = Utf8PathBuf::from("/tmp/prof.fdata"); + log::info!("Merging LLVM BOLT profiles to {merged_profile}"); + + let profiles: Vec<_> = + glob::glob(&format!("{profile_root}*"))?.into_iter().collect::<Result<Vec<_>, _>>()?; + + let mut merge_args = vec!["merge-fdata"]; + merge_args.extend(profiles.iter().map(|p| p.to_str().unwrap())); + + cmd(&merge_args) + .redirect_output(merged_profile.clone()) + .run() + .context("Cannot merge BOLT profiles")?; + + log::info!("LLVM BOLT statistics"); + log::info!( + "{merged_profile}: {}", + humansize::format_size(std::fs::metadata(merged_profile.as_std_path())?.len(), BINARY) + ); + + let size = profiles + .iter() + .map(|p| std::fs::metadata(p).map(|metadata| metadata.len())) + .collect::<Result<Vec<_>, _>>()? + .into_iter() + .sum::<u64>(); + log::info!("{profile_root}: {}", humansize::format_size(size, BINARY)); + log::info!("Profile file count: {}", profiles.len()); + + Ok(LlvmBoltProfile(merged_profile)) +} diff --git a/src/tools/opt-dist/src/utils/io.rs b/src/tools/opt-dist/src/utils/io.rs new file mode 100644 index 00000000000..43546e5fcfa --- /dev/null +++ b/src/tools/opt-dist/src/utils/io.rs @@ -0,0 +1,48 @@ +use anyhow::Context; +use camino::Utf8Path; +use fs_extra::dir::CopyOptions; +use std::fs::File; + +/// Delete and re-create the directory. +pub fn reset_directory(path: &Utf8Path) -> anyhow::Result<()> { + log::info!("Resetting directory {path}"); + let _ = std::fs::remove_dir(path); + std::fs::create_dir_all(path)?; + Ok(()) +} + +pub fn copy_directory(src: &Utf8Path, dst: &Utf8Path) -> anyhow::Result<()> { + log::info!("Copying directory {src} to {dst}"); + fs_extra::dir::copy(src, dst, &CopyOptions::default().copy_inside(true))?; + Ok(()) +} + +#[allow(unused)] +pub fn move_directory(src: &Utf8Path, dst: &Utf8Path) -> anyhow::Result<()> { + log::info!("Moving directory {src} to {dst}"); + fs_extra::dir::move_dir(src, dst, &CopyOptions::default().content_only(true))?; + Ok(()) +} + +/// Counts all children of a directory (non-recursively). +pub fn count_files(dir: &Utf8Path) -> anyhow::Result<u64> { + Ok(std::fs::read_dir(dir)?.count() as u64) +} + +pub fn delete_directory(path: &Utf8Path) -> anyhow::Result<()> { + log::info!("Deleting directory `{path}`"); + std::fs::remove_dir_all(path.as_std_path()) + .context(format!("Cannot remove directory {path}"))?; + Ok(()) +} + +pub fn unpack_archive(path: &Utf8Path, dest_dir: &Utf8Path) -> anyhow::Result<()> { + log::info!("Unpacking directory `{path}` into `{dest_dir}`"); + + assert!(path.as_str().ends_with(".tar.xz")); + let file = File::open(path.as_std_path())?; + let file = xz::read::XzDecoder::new(file); + let mut archive = tar::Archive::new(file); + archive.unpack(dest_dir.as_std_path())?; + Ok(()) +} diff --git a/src/tools/opt-dist/src/utils/mod.rs b/src/tools/opt-dist/src/utils/mod.rs new file mode 100644 index 00000000000..9305d4989aa --- /dev/null +++ b/src/tools/opt-dist/src/utils/mod.rs @@ -0,0 +1,36 @@ +pub mod io; + +use crate::environment::Environment; +use crate::utils::io::delete_directory; +use humansize::BINARY; +use sysinfo::{DiskExt, RefreshKind, System, SystemExt}; + +pub fn format_env_variables() -> String { + let vars = std::env::vars().map(|(key, value)| format!("{key}={value}")).collect::<Vec<_>>(); + vars.join("\n") +} + +pub fn print_free_disk_space() -> anyhow::Result<()> { + let sys = System::new_with_specifics(RefreshKind::default().with_disks_list().with_disks()); + let available_space: u64 = sys.disks().iter().map(|d| d.available_space()).sum(); + let total_space: u64 = sys.disks().iter().map(|d| d.total_space()).sum(); + let used_space = total_space - available_space; + + log::info!( + "Free disk space: {} out of total {} ({:.2}% used)", + humansize::format_size(available_space, BINARY), + humansize::format_size(total_space, BINARY), + (used_space as f64 / total_space as f64) * 100.0 + ); + Ok(()) +} + +pub fn clear_llvm_files(env: &dyn Environment) -> anyhow::Result<()> { + // Bootstrap currently doesn't support rebuilding LLVM when PGO options + // change (or any other llvm-related options); so just clear out the relevant + // directories ourselves. + log::info!("Clearing LLVM build files"); + delete_directory(&env.build_artifacts().join("llvm"))?; + delete_directory(&env.build_artifacts().join("lld"))?; + Ok(()) +} diff --git a/src/tools/tidy/src/deps.rs b/src/tools/tidy/src/deps.rs index ecc84c1618c..2c237c418b1 100644 --- a/src/tools/tidy/src/deps.rs +++ b/src/tools/tidy/src/deps.rs @@ -40,10 +40,12 @@ const EXCEPTIONS: &[(&str, &str)] = &[ ("ar_archive_writer", "Apache-2.0 WITH LLVM-exception"), // rustc ("colored", "MPL-2.0"), // rustfmt ("dissimilar", "Apache-2.0"), // rustdoc, rustc_lexer (few tests) via expect-test, (dev deps) + ("encoding_rs", "(Apache-2.0 OR MIT) AND BSD-3-Clause"), // opt-dist ("fluent-langneg", "Apache-2.0"), // rustc (fluent translations) ("fortanix-sgx-abi", "MPL-2.0"), // libstd but only for `sgx` target. FIXME: this dependency violates the documentation comment above. ("instant", "BSD-3-Clause"), // rustc_driver/tracing-subscriber/parking_lot ("mdbook", "MPL-2.0"), // mdbook + ("openssl", "Apache-2.0"), // opt-dist ("ryu", "Apache-2.0 OR BSL-1.0"), // cargo/... (because of serde) ("self_cell", "Apache-2.0"), // rustc (fluent translations) ("snap", "BSD-3-Clause"), // rustc @@ -142,6 +144,8 @@ const PERMITTED_RUSTC_DEPENDENCIES: &[&str] = &[ "elsa", "ena", "equivalent", + "errno", + "errno-dragonfly", "expect-test", "fallible-iterator", // dependency of `thorin` "fastrand", diff --git a/src/tools/tidy/src/pal.rs b/src/tools/tidy/src/pal.rs index 6fd41e83362..62c070221dc 100644 --- a/src/tools/tidy/src/pal.rs +++ b/src/tools/tidy/src/pal.rs @@ -53,13 +53,9 @@ const EXCEPTION_PATHS: &[&str] = &[ // FIXME: platform-specific code should be moved to `sys` "library/std/src/io/copy.rs", "library/std/src/io/stdio.rs", - "library/std/src/f32.rs", - "library/std/src/f64.rs", "library/std/src/path.rs", "library/std/src/sys_common", // Should only contain abstractions over platforms "library/std/src/net/test.rs", // Utility helpers for tests - "library/std/src/personality.rs", - "library/std/src/personality/", ]; pub fn check(path: &Path, bad: &mut bool) { |
