//! Main evaluator loop and setting up the initial stack frame.
-use std::convert::TryFrom;
use std::ffi::OsStr;
+use std::iter;
use log::info;
use rustc_hir::def_id::DefId;
-use rustc_middle::ty::{self, layout::LayoutCx, TyCtxt};
-use rustc_target::abi::LayoutOf;
+use rustc_middle::ty::{
+ self,
+ layout::{LayoutCx, LayoutOf},
+ TyCtxt,
+};
use rustc_target::spec::abi::Abi;
+use rustc_session::config::EntryFnType;
+
+use std::collections::HashSet;
+
use crate::*;
#[derive(Copy, Clone, Debug, PartialEq)]
Allow,
}
+#[derive(Copy, Clone, PartialEq, Eq)]
+pub enum BacktraceStyle {
+ /// Prints a terser backtrace which ideally only contains relevant information.
+ Short,
+ /// Prints a backtrace with all possible information.
+ Full,
+ /// Prints only the frame that the error occurs in.
+ Off,
+}
+
/// Configuration needed to spawn a Miri instance.
#[derive(Clone)]
pub struct MiriConfig {
pub stacked_borrows: bool,
/// Controls alignment checking.
pub check_alignment: AlignmentCheck,
+ /// Controls integer and float validity (e.g., initialization) checking.
+ pub check_number_validity: bool,
/// Controls function [ABI](Abi) checking.
pub check_abi: bool,
/// Action for an op requiring communication with the host.
pub ignore_leaks: bool,
/// Environment variables that should always be isolated from the host.
pub excluded_env_vars: Vec<String>,
+ /// Environment variables that should always be forwarded from the host.
+ pub forwarded_env_vars: Vec<String>,
/// Command-line arguments passed to the interpreted program.
pub args: Vec<String>,
/// The seed to use when non-determinism or randomness are required (e.g. ptr-to-int cast, `getrandom()`).
pub seed: Option<u64>,
- /// The stacked borrows pointer id to report about
- pub tracked_pointer_tag
: Option<PtrId>,
- /// The stacked borrows call ID to report about
- pub tracked_call_id: Option<CallId>,
- /// The allocation id to report about.
- pub tracked_alloc_id
: Option<AllocId>,
+ /// The stacked borrows pointer ids to report about
+ pub tracked_pointer_tag
s: HashSet<PtrId>,
+ /// The stacked borrows call IDs to report about
+ pub tracked_call_ids: HashSet<CallId>,
+ /// The allocation ids to report about.
+ pub tracked_alloc_id
s: HashSet<AllocId>,
/// Whether to track raw pointers in stacked borrows.
- pub track_raw: bool,
+ pub tag_raw: bool,
/// Determine if data race detection should be enabled
pub data_race_detector: bool,
/// Rate of spurious failures for compare_exchange_weak atomic operations,
/// If `Some`, enable the `measureme` profiler, writing results to a file
/// with the specified prefix.
pub measureme_out: Option<String>,
- /// Panic when unsupported functionality is encountered
+ /// Panic when unsupported functionality is encountered.
pub panic_on_unsupported: bool,
+ /// Which style to use for printing backtraces.
+ pub backtrace_style: BacktraceStyle,
+ /// Which provenance to use for int2ptr casts
+ pub provenance_mode: ProvenanceMode,
+ /// Whether to ignore any output by the program. This is helpful when debugging miri
+ /// as its messages don't get intermingled with the program messages.
+ pub mute_stdout_stderr: bool,
}
impl Default for MiriConfig {
validate: true,
stacked_borrows: true,
check_alignment: AlignmentCheck::Int,
+ check_number_validity: false,
check_abi: true,
isolated_op: IsolatedOp::Reject(RejectOpWith::Abort),
ignore_leaks: false,
excluded_env_vars: vec![],
+ forwarded_env_vars: vec![],
args: vec![],
seed: None,
- tracked_pointer_tag: None,
- tracked_call_id: None,
- tracked_alloc_id: None,
- track_raw: false,
+ tracked_pointer_tags: HashSet::default(),
+ tracked_call_ids: HashSet::default(),
+ tracked_alloc_ids: HashSet::default(),
+ tag_raw: false,
data_race_detector: true,
cmpxchg_weak_failure_rate: 0.8,
measureme_out: None,
panic_on_unsupported: false,
+ backtrace_style: BacktraceStyle::Short,
+ provenance_mode: ProvenanceMode::Legacy,
+ mute_stdout_stderr: false,
}
}
}
/// Returns a freshly created `InterpCx`, along with an `MPlaceTy` representing
-/// the location where the return value of the `start` lang item will be
+/// the location where the return value of the `start` function will be
/// written to.
/// Public because this is also used by `priroda`.
pub fn create_ecx<'mir, 'tcx: 'mir>(
tcx: TyCtxt<'tcx>,
- main_id: DefId,
- config: MiriConfig,
+ entry_id: DefId,
+ entry_type: EntryFnType,
+ config: &MiriConfig,
) -> InterpResult<'tcx, (InterpCx<'mir, 'tcx, Evaluator<'mir, 'tcx>>, MPlaceTy<'tcx, Tag>)> {
let param_env = ty::ParamEnv::reveal_all();
let layout_cx = LayoutCx { tcx, param_env };
tcx,
rustc_span::source_map::DUMMY_SP,
param_env,
- Evaluator::new(&config, layout_cx),
- MemoryExtra::new(&config),
+ Evaluator::new(config, layout_cx),
);
- // Complete initialization.
- EnvVars::init(&mut ecx, config.excluded_env_vars)?;
- MemoryExtra::init_extern_statics(&mut ecx)?;
+
+ // Capture the current interpreter stack state (which should be empty) so that we can emit
+ // allocation-tracking and tag-tracking diagnostics for allocations which are part of the
+ // early runtime setup.
+ let info = ecx.preprocess_diagnostics();
+
+ // Some parts of initialization require a full `InterpCx`.
+ Evaluator::late_init(&mut ecx, config)?;
// Make sure we have MIR. We check MIR for some stable monomorphic function in libcore.
- let sentinel = ecx.resolve_path(&["core", "ascii", "escape_default"]);
- if !tcx.is_mir_available(sentinel.def.def_id()) {
- tcx.sess.fatal("the current sysroot was built without `-Zalways-encode-mir`. Use `cargo miri setup` to prepare a sysroot that is suitable for Miri.");
+ let sentinel = ecx.try_resolve_path(&["core", "ascii", "escape_default"]);
+ if !matches!(sentinel, Some(s) if tcx.is_mir_available(s.def.def_id())) {
+ tcx.sess.fatal(
+ "the current sysroot was built without `-Zalways-encode-mir`, or libcore seems missing. \
+ Use `cargo miri setup` to prepare a sysroot that is suitable for Miri."
+ );
}
- // Setup first stack-frame
- let main_instance = ty::Instance::mono(tcx, main_id);
- let main_mir = ecx.load_mir(main_instance.def, None)?;
- if main_mir.arg_count != 0 {
- bug!("main function must not take any arguments");
- }
+ // Setup first stack frame.
+ let entry_instance = ty::Instance::mono(tcx, entry_id);
+
+ // First argument is constructed later, because it's skipped if the entry function uses #[start].
- let start_id = tcx.lang_items().start_fn().unwrap();
- let main_ret_ty = tcx.fn_sig(main_id).output();
- let main_ret_ty = main_ret_ty.no_bound_vars().unwrap();
- let start_instance = ty::Instance::resolve(
- tcx,
- ty::ParamEnv::reveal_all(),
- start_id,
- tcx.mk_substs(::std::iter::once(ty::subst::GenericArg::from(main_ret_ty))),
- )
- .unwrap()
- .unwrap();
-
- // First argument: pointer to `main()`.
- let main_ptr = ecx.memory.create_fn_alloc(FnVal::Instance(main_instance));
// Second argument (argc): length of `config.args`.
let argc = Scalar::from_machine_usize(u64::try_from(config.args.len()).unwrap(), &ecx);
// Third argument (`argv`): created from `config.args`.
// Store command line as UTF-16 for Windows `GetCommandLineW`.
{
// Construct a command string with all the aguments.
- let mut cmd = String::new();
- for arg in config.args.iter() {
- if !cmd.is_empty() {
- cmd.push(' ');
- }
- cmd.push_str(&*shell_escape::windows::escape(arg.as_str().into()));
- }
- // Don't forget `0` terminator.
- cmd.push(std::char::from_u32(0).unwrap());
+ let cmd_utf16: Vec<u16> = args_to_utf16_command_string(config.args.iter());
- let cmd_utf16: Vec<u16> = cmd.encode_utf16().collect();
let cmd_type = tcx.mk_array(tcx.types.u16, u64::try_from(cmd_utf16.len()).unwrap());
let cmd_place =
ecx.allocate(ecx.layout_of(cmd_type)?, MiriMemoryKind::Machine.into())?;
// Return place (in static memory so that it does not count as leak).
let ret_place = ecx.allocate(ecx.machine.layouts.isize, MiriMemoryKind::Machine.into())?;
// Call start function.
- ecx.call_function(
- start_instance,
- Abi::Rust,
- &[Scalar::from_pointer(main_ptr, &ecx).into(), argc.into(), argv],
- Some(&ret_place.into()),
- StackPopCleanup::None { cleanup: true },
- )?;
+
+ match entry_type {
+ EntryFnType::Main => {
+ let start_id = tcx.lang_items().start_fn().unwrap();
+ let main_ret_ty = tcx.fn_sig(entry_id).output();
+ let main_ret_ty = main_ret_ty.no_bound_vars().unwrap();
+ let start_instance = ty::Instance::resolve(
+ tcx,
+ ty::ParamEnv::reveal_all(),
+ start_id,
+ tcx.mk_substs(::std::iter::once(ty::subst::GenericArg::from(main_ret_ty))),
+ )
+ .unwrap()
+ .unwrap();
+
+ let main_ptr = ecx.create_fn_alloc_ptr(FnVal::Instance(entry_instance));
+
+ ecx.call_function(
+ start_instance,
+ Abi::Rust,
+ &[Scalar::from_pointer(main_ptr, &ecx).into(), argc.into(), argv],
+ &ret_place.into(),
+ StackPopCleanup::Root { cleanup: true },
+ )?;
+ }
+ EntryFnType::Start => {
+ ecx.call_function(
+ entry_instance,
+ Abi::Rust,
+ &[argc.into(), argv],
+ &ret_place.into(),
+ StackPopCleanup::Root { cleanup: true },
+ )?;
+ }
+ }
+
+ // Emit any diagnostics related to the setup process for the runtime, so that when the
+ // interpreter loop starts there are no unprocessed diagnostics.
+ ecx.process_diagnostics(info);
Ok((ecx, ret_place))
}
-/// Evaluates the main function specified by `main_id`.
+/// Evaluates the entry function specified by `entry_id`.
/// Returns `Some(return_code)` if program executed completed.
/// Returns `None` if an evaluation error occured.
-pub fn eval_main<'tcx>(tcx: TyCtxt<'tcx>, main_id: DefId, config: MiriConfig) -> Option<i64> {
+pub fn eval_entry<'tcx>(
+ tcx: TyCtxt<'tcx>,
+ entry_id: DefId,
+ entry_type: EntryFnType,
+ config: MiriConfig,
+) -> Option<i64> {
// Copy setting before we move `config`.
let ignore_leaks = config.ignore_leaks;
- let (mut ecx, ret_place) = match create_ecx(tcx, main_id, config) {
+ let (mut ecx, ret_place) = match create_ecx(tcx, entry_id, entry_type, &config) {
Ok(v) => v,
Err(err) => {
err.print_backtrace();
})();
// Machine cleanup.
- EnvVars::cleanup(&mut ecx).unwrap();
+ // Execution of the program has halted so any memory access we do here
+ // cannot produce a real data race. If we do not do something to disable
+ // data race detection here, some uncommon combination of errors will
+ // cause a data race to be detected:
+ // https://github.com/rust-lang/miri/issues/2020
+ ecx.allow_data_races_mut(|ecx| EnvVars::cleanup(ecx).unwrap());
// Process the result.
match res {
Ok(return_code) => {
if !ignore_leaks {
+ // Check for thread leaks.
+ if !ecx.have_all_terminated() {
+ tcx.sess.err(
+ "the main thread terminated without waiting for all remaining threads",
+ );
+ tcx.sess.note_without_error("pass `-Zmiri-ignore-leaks` to disable this check");
+ return None;
+ }
+ // Check for memory leaks.
info!("Additonal static roots: {:?}", ecx.machine.static_roots);
- let leaks = ecx.memory.leak_report(&ecx.machine.static_roots);
+ let leaks = ecx.leak_report(&ecx.machine.static_roots);
if leaks != 0 {
tcx.sess.err("the evaluated program leaked memory");
+ tcx.sess.note_without_error("pass `-Zmiri-ignore-leaks` to disable this check");
// Ignore the provided return code - let the reported error
// determine the return code.
return None;
Err(e) => report_error(&ecx, e),
}
}
+
+/// Turns an array of arguments into a Windows command line string.
+///
+/// The string will be UTF-16 encoded and NUL terminated.
+///
+/// Panics if the zeroth argument contains the `"` character because doublequotes
+/// in argv[0] cannot be encoded using the standard command line parsing rules.
+///
+/// Further reading:
+/// * [Parsing C++ command-line arguments](https://docs.microsoft.com/en-us/cpp/cpp/main-function-command-line-args?view=msvc-160#parsing-c-command-line-arguments)
+/// * [The C/C++ Parameter Parsing Rules](https://daviddeley.com/autohotkey/parameters/parameters.htm#WINCRULES)
+fn args_to_utf16_command_string<I, T>(mut args: I) -> Vec<u16>
+where
+ I: Iterator<Item = T>,
+ T: AsRef<str>,
+{
+ // Parse argv[0]. Slashes aren't escaped. Literal double quotes are not allowed.
+ let mut cmd = {
+ let arg0 = if let Some(arg0) = args.next() {
+ arg0
+ } else {
+ return vec![0];
+ };
+ let arg0 = arg0.as_ref();
+ if arg0.contains('"') {
+ panic!("argv[0] cannot contain a doublequote (\") character");
+ } else {
+ // Always surround argv[0] with quotes.
+ let mut s = String::new();
+ s.push('"');
+ s.push_str(arg0);
+ s.push('"');
+ s
+ }
+ };
+
+ // Build the other arguments.
+ for arg in args {
+ let arg = arg.as_ref();
+ cmd.push(' ');
+ if arg.is_empty() {
+ cmd.push_str("\"\"");
+ } else if !arg.bytes().any(|c| matches!(c, b'"' | b'\t' | b' ')) {
+ // No quote, tab, or space -- no escaping required.
+ cmd.push_str(arg);
+ } else {
+ // Spaces and tabs are escaped by surrounding them in quotes.
+ // Quotes are themselves escaped by using backslashes when in a
+ // quoted block.
+ // Backslashes only need to be escaped when one or more are directly
+ // followed by a quote. Otherwise they are taken literally.
+
+ cmd.push('"');
+ let mut chars = arg.chars().peekable();
+ loop {
+ let mut nslashes = 0;
+ while let Some(&'\\') = chars.peek() {
+ chars.next();
+ nslashes += 1;
+ }
+
+ match chars.next() {
+ Some('"') => {
+ cmd.extend(iter::repeat('\\').take(nslashes * 2 + 1));
+ cmd.push('"');
+ }
+ Some(c) => {
+ cmd.extend(iter::repeat('\\').take(nslashes));
+ cmd.push(c);
+ }
+ None => {
+ cmd.extend(iter::repeat('\\').take(nslashes * 2));
+ break;
+ }
+ }
+ }
+ cmd.push('"');
+ }
+ }
+
+ if cmd.contains('\0') {
+ panic!("interior null in command line arguments");
+ }
+ cmd.encode_utf16().chain(iter::once(0)).collect()
+}
+
+#[cfg(test)]
+mod tests {
+ use super::*;
+ #[test]
+ #[should_panic(expected = "argv[0] cannot contain a doublequote (\") character")]
+ fn windows_argv0_panic_on_quote() {
+ args_to_utf16_command_string(["\""].iter());
+ }
+ #[test]
+ fn windows_argv0_no_escape() {
+ // Ensure that a trailing backslash in argv[0] is not escaped.
+ let cmd = String::from_utf16_lossy(&args_to_utf16_command_string(
+ [r"C:\Program Files\", "arg1", "arg 2", "arg \" 3"].iter(),
+ ));
+ assert_eq!(cmd.trim_end_matches("\0"), r#""C:\Program Files\" arg1 "arg 2" "arg \" 3""#);
+ }
+}
projects / rust.git / blobdiff