Rollup merge of #122405 - celinval:smir-new-const, r=oli-obk

Add methods to create StableMIR constant

I've been experimenting with transforming the StableMIR to instrument the code with potential UB checks.

The modified body will only be used by our analysis tool, however, constants in StableMIR must be backed by rustc constants. Thus, I'm adding a few functions to build constants, such as building string and other primitives.

One question I have is whether we should create a global allocation instead for strings.

r? ``````@oli-obk``````

compiler/rustc_smir/src/rustc_smir/context.rs

@@ -23,7 +23,8 @@ use stable_mir::mir::Body;
 use stable_mir::target::{MachineInfo, MachineSize};
 use stable_mir::ty::{
     AdtDef, AdtKind, Allocation, ClosureDef, ClosureKind, Const, FieldDef, FnDef, ForeignDef,
-    ForeignItemKind, GenericArgs, LineInfo, PolyFnSig, RigidTy, Span, Ty, TyKind, VariantDef,
+    ForeignItemKind, GenericArgs, LineInfo, PolyFnSig, RigidTy, Span, Ty, TyKind, UintTy,
+    VariantDef,
 };
 use stable_mir::{Crate, CrateDef, CrateItem, CrateNum, DefId, Error, Filename, ItemKind, Symbol};
 use std::cell::RefCell;
@@ -341,15 +342,56 @@ impl<'tcx> Context for TablesWrapper<'tcx> {
             .ok_or_else(|| Error::new(format!("Const `{cnst:?}` cannot be encoded as u64")))
     }
 
-    fn usize_to_const(&self, val: u64) -> Result<Const, Error> {
+    fn try_new_const_zst(&self, ty: Ty) -> Result<Const, Error> {
         let mut tables = self.0.borrow_mut();
-        let ty = tables.tcx.types.usize;
+        let tcx = tables.tcx;
+        let ty_internal = ty.internal(&mut *tables, tcx);
+        let size = tables
+            .tcx
+            .layout_of(ParamEnv::empty().and(ty_internal))
+            .map_err(|err| {
+                Error::new(format!(
+                    "Cannot create a zero-sized constant for type `{ty_internal}`: {err}"
+                ))
+            })?
+            .size;
+        if size.bytes() != 0 {
+            return Err(Error::new(format!(
+                "Cannot create a zero-sized constant for type `{ty_internal}`: \
+                 Type `{ty_internal}` has {} bytes",
+                size.bytes()
+            )));
+        }
+
+        Ok(ty::Const::zero_sized(tables.tcx, ty_internal).stable(&mut *tables))
+    }
+
+    fn new_const_str(&self, value: &str) -> Const {
+        let mut tables = self.0.borrow_mut();
+        let tcx = tables.tcx;
+        let ty = ty::Ty::new_static_str(tcx);
+        let bytes = value.as_bytes();
+        let val_tree = ty::ValTree::from_raw_bytes(tcx, bytes);
+
+        ty::Const::new_value(tcx, val_tree, ty).stable(&mut *tables)
+    }
+
+    fn new_const_bool(&self, value: bool) -> Const {
+        let mut tables = self.0.borrow_mut();
+        ty::Const::from_bool(tables.tcx, value).stable(&mut *tables)
+    }
+
+    fn try_new_const_uint(&self, value: u128, uint_ty: UintTy) -> Result<Const, Error> {
+        let mut tables = self.0.borrow_mut();
+        let tcx = tables.tcx;
+        let ty = ty::Ty::new_uint(tcx, uint_ty.internal(&mut *tables, tcx));
         let size = tables.tcx.layout_of(ParamEnv::empty().and(ty)).unwrap().size;
 
-        let scalar = ScalarInt::try_from_uint(val, size).ok_or_else(|| {
-            Error::new(format!("Value overflow: cannot convert `{val}` to usize."))
+        // We don't use Const::from_bits since it doesn't have any error checking.
+        let scalar = ScalarInt::try_from_uint(value, size).ok_or_else(|| {
+            Error::new(format!("Value overflow: cannot convert `{value}` to `{ty}`."))
         })?;
-        Ok(rustc_middle::ty::Const::new_value(tables.tcx, ValTree::from_scalar_int(scalar), ty)
+        Ok(ty::Const::new_value(tables.tcx, ValTree::from_scalar_int(scalar), ty)
             .stable(&mut *tables))
     }
 
@@ -556,7 +598,9 @@ impl<'tcx> Context for TablesWrapper<'tcx> {
         global_alloc: &GlobalAlloc,
     ) -> Option<stable_mir::mir::alloc::AllocId> {
         let mut tables = self.0.borrow_mut();
-        let GlobalAlloc::VTable(ty, trait_ref) = global_alloc else { return None };
+        let GlobalAlloc::VTable(ty, trait_ref) = global_alloc else {
+            return None;
+        };
         let tcx = tables.tcx;
         let alloc_id = tables.tcx.vtable_allocation((
             ty.internal(&mut *tables, tcx),

compiler/stable_mir/src/compiler_interface.rs

@@ -14,7 +14,7 @@ use crate::ty::{
     AdtDef, AdtKind, Allocation, ClosureDef, ClosureKind, Const, FieldDef, FnDef, ForeignDef,
     ForeignItemKind, ForeignModule, ForeignModuleDef, GenericArgs, GenericPredicates, Generics,
     ImplDef, ImplTrait, LineInfo, PolyFnSig, RigidTy, Span, TraitDecl, TraitDef, Ty, TyKind,
-    VariantDef,
+    UintTy, VariantDef,
 };
 use crate::{
     mir, Crate, CrateItem, CrateItems, CrateNum, DefId, Error, Filename, ImplTraitDecls, ItemKind,
@@ -101,8 +101,17 @@ pub trait Context {
     /// Evaluate constant as a target usize.
     fn eval_target_usize(&self, cnst: &Const) -> Result<u64, Error>;
 
-    /// Create a target usize constant for the given value.
-    fn usize_to_const(&self, val: u64) -> Result<Const, Error>;
+    /// Create a new zero-sized constant.
+    fn try_new_const_zst(&self, ty: Ty) -> Result<Const, Error>;
+
+    /// Create a new constant that represents the given string value.
+    fn new_const_str(&self, value: &str) -> Const;
+
+    /// Create a new constant that represents the given boolean value.
+    fn new_const_bool(&self, value: bool) -> Const;
+
+    /// Create a new constant that represents the given value.
+    fn try_new_const_uint(&self, value: u128, uint_ty: UintTy) -> Result<Const, Error>;
 
     /// Create a new type from the given kind.
     fn new_rigid_ty(&self, kind: RigidTy) -> Ty;
@@ -200,7 +209,7 @@ pub trait Context {
 
 // A thread local variable that stores a pointer to the tables mapping between TyCtxt
 // datastructures and stable MIR datastructures
-scoped_thread_local! (static TLV: Cell<*const ()>);
+scoped_thread_local!(static TLV: Cell<*const ()>);
 
 pub fn run<F, T>(context: &dyn Context, f: F) -> Result<T, Error>
 where

compiler/stable_mir/src/ty.rs

@@ -128,13 +128,38 @@ impl Const {
 
     /// Creates an interned usize constant.
     fn try_from_target_usize(val: u64) -> Result<Self, Error> {
-        with(|cx| cx.usize_to_const(val))
+        with(|cx| cx.try_new_const_uint(val.into(), UintTy::Usize))
     }
 
     /// Try to evaluate to a target `usize`.
     pub fn eval_target_usize(&self) -> Result<u64, Error> {
         with(|cx| cx.eval_target_usize(self))
     }
+
+    /// Create a constant that represents a new zero-sized constant of type T.
+    /// Fails if the type is not a ZST or if it doesn't have a known size.
+    pub fn try_new_zero_sized(ty: Ty) -> Result<Const, Error> {
+        with(|cx| cx.try_new_const_zst(ty))
+    }
+
+    /// Build a new constant that represents the given string.
+    ///
+    /// Note that there is no guarantee today about duplication of the same constant.
+    /// I.e.: Calling this function multiple times with the same argument may or may not return
+    /// the same allocation.
+    pub fn from_str(value: &str) -> Const {
+        with(|cx| cx.new_const_str(value))
+    }
+
+    /// Build a new constant that represents the given boolean value.
+    pub fn from_bool(value: bool) -> Const {
+        with(|cx| cx.new_const_bool(value))
+    }
+
+    /// Build a new constant that represents the given unsigned integer.
+    pub fn try_from_uint(value: u128, uint_ty: UintTy) -> Result<Const, Error> {
+        with(|cx| cx.try_new_const_uint(value, uint_ty))
+    }
 }
 
 #[derive(Clone, Copy, Debug, PartialEq, Eq)]

tests/ui-fulldeps/stable-mir/check_transform.rs

@@ -0,0 +1,147 @@
+//@ run-pass
+//! Test a few methods to transform StableMIR.
+
+//@ ignore-stage1
+//@ ignore-cross-compile
+//@ ignore-remote
+//@ ignore-windows-gnu mingw has troubles with linking https://github.com/rust-lang/rust/pull/116837
+
+#![feature(rustc_private)]
+#![feature(assert_matches)]
+#![feature(control_flow_enum)]
+#![feature(ascii_char, ascii_char_variants)]
+
+extern crate rustc_hir;
+#[macro_use]
+extern crate rustc_smir;
+extern crate rustc_driver;
+extern crate rustc_interface;
+extern crate stable_mir;
+
+use rustc_smir::rustc_internal;
+use stable_mir::mir::alloc::GlobalAlloc;
+use stable_mir::mir::mono::Instance;
+use stable_mir::mir::{Body, Constant, Operand, Rvalue, StatementKind, TerminatorKind};
+use stable_mir::ty::{Const, ConstantKind};
+use stable_mir::{CrateDef, CrateItems, ItemKind};
+use std::convert::TryFrom;
+use std::io::Write;
+use std::ops::ControlFlow;
+
+const CRATE_NAME: &str = "input";
+
+/// This function uses the Stable MIR APIs to transform the MIR.
+fn test_transform() -> ControlFlow<()> {
+    // Find items in the local crate.
+    let items = stable_mir::all_local_items();
+
+    // Test fn_abi
+    let target_fn = *get_item(&items, (ItemKind::Fn, "dummy")).unwrap();
+    let instance = Instance::try_from(target_fn).unwrap();
+    let body = instance.body().unwrap();
+    check_msg(&body, "oops");
+
+    let new_msg = "new panic message";
+    let new_body = change_panic_msg(body, new_msg);
+    check_msg(&new_body, new_msg);
+
+    ControlFlow::Continue(())
+}
+
+/// Check that the body panic message matches the given message.
+fn check_msg(body: &Body, expected: &str) {
+    let msg = body
+        .blocks
+        .iter()
+        .find_map(|bb| match &bb.terminator.kind {
+            TerminatorKind::Call { args, .. } => {
+                assert_eq!(args.len(), 1, "Expected panic message, but found {args:?}");
+                let msg_const = match &args[0] {
+                    Operand::Constant(msg_const) => msg_const,
+                    Operand::Copy(place) | Operand::Move(place) => {
+                        assert!(place.projection.is_empty());
+                        bb.statements
+                            .iter()
+                            .find_map(|stmt| match &stmt.kind {
+                                StatementKind::Assign(
+                                    destination,
+                                    Rvalue::Use(Operand::Constant(msg_const)),
+                                ) if destination == place => Some(msg_const),
+                                _ => None,
+                            })
+                            .unwrap()
+                    }
+                };
+                let ConstantKind::Allocated(alloc) = msg_const.literal.kind() else {
+                    unreachable!()
+                };
+                assert_eq!(alloc.provenance.ptrs.len(), 1);
+
+                let alloc_prov_id = alloc.provenance.ptrs[0].1 .0;
+                let GlobalAlloc::Memory(val) = GlobalAlloc::from(alloc_prov_id) else {
+                    unreachable!()
+                };
+                let bytes = val.raw_bytes().unwrap();
+                Some(std::str::from_utf8(&bytes).unwrap().to_string())
+            }
+            _ => None,
+        })
+        .expect("Failed to find panic message");
+    assert_eq!(&msg, expected);
+}
+
+/// Modify body to use a different panic message.
+fn change_panic_msg(mut body: Body, new_msg: &str) -> Body {
+    for bb in &mut body.blocks {
+        match &mut bb.terminator.kind {
+            TerminatorKind::Call { args, .. } => {
+                let new_const = Const::from_str(new_msg);
+                args[0] = Operand::Constant(Constant {
+                    literal: new_const,
+                    span: bb.terminator.span,
+                    user_ty: None,
+                });
+            }
+            _ => {}
+        }
+    }
+    body
+}
+
+fn get_item<'a>(
+    items: &'a CrateItems,
+    item: (ItemKind, &str),
+) -> Option<&'a stable_mir::CrateItem> {
+    items.iter().find(|crate_item| (item.0 == crate_item.kind()) && crate_item.name() == item.1)
+}
+
+/// This test will generate and analyze a dummy crate using the stable mir.
+/// For that, it will first write the dummy crate into a file.
+/// Then it will create a `StableMir` using custom arguments and then
+/// it will run the compiler.
+fn main() {
+    let path = "transform_input.rs";
+    generate_input(&path).unwrap();
+    let args = vec![
+        "rustc".to_string(),
+        "--crate-type=lib".to_string(),
+        "--crate-name".to_string(),
+        CRATE_NAME.to_string(),
+        path.to_string(),
+    ];
+    run!(args, test_transform).unwrap();
+}
+
+fn generate_input(path: &str) -> std::io::Result<()> {
+    let mut file = std::fs::File::create(path)?;
+    write!(
+        file,
+        r#"
+        #![feature(panic_internals)]
+        pub fn dummy() {{
+            core::panicking::panic_str("oops");
+        }}
+        "#
+    )?;
+    Ok(())
+}