add -Zmiri-strict-provenance

Author: RalfJung

README.md

@@ -294,6 +294,10 @@ environment variable:
   entropy.  The default seed is 0.  **NOTE**: This entropy is not good enough
   for cryptographic use!  Do not generate secret keys in Miri or perform other
   kinds of cryptographic operations that rely on proper random numbers.
+* `-Zmiri-strict-provenance` enables [strict
+  provenance](https://github.com/rust-lang/rust/issues/95228) checking in Miri. This means that
+  casting an integer to a pointer yields a result with 'invalid' provenance, i.e., with provenance
+  that cannot be used for any memory access. Also implies `-Zmiri-tag-raw-pointers`.
 * `-Zmiri-symbolic-alignment-check` makes the alignment check more strict.  By
   default, alignment is checked by casting the pointer to an integer, and making
   sure that is a multiple of the alignment.  This can lead to cases where a

src/bin/miri.rs

@@ -363,6 +363,10 @@ fn main() {
                 "-Zmiri-tag-raw-pointers" => {
                     miri_config.tag_raw = true;
                 }
+                "-Zmiri-strict-provenance" => {
+                    miri_config.strict_provenance = true;
+                    miri_config.tag_raw = true;
+                }
                 "-Zmiri-track-raw-pointers" => {
                     eprintln!(
                         "WARNING: -Zmiri-track-raw-pointers has been renamed to -Zmiri-tag-raw-pointers, the old name is deprecated."

src/eval.rs

@@ -108,9 +108,13 @@ pub struct MiriConfig {
     /// 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,
+    /// Whether to enforce "strict provenance" rules. Enabling this means int2ptr casts return
+    /// pointers with an invalid provenance, i.e., not valid for any memory access.
+    pub strict_provenance: bool,
 }
 
 impl Default for MiriConfig {
@@ -136,6 +140,7 @@ impl Default for MiriConfig {
             measureme_out: None,
             panic_on_unsupported: false,
             backtrace_style: BacktraceStyle::Short,
+            strict_provenance: false,
         }
     }
 }

src/intptrcast.rs

@@ -15,23 +15,27 @@ pub type MemoryExtra = RefCell<GlobalState>;
 pub struct GlobalState {
     /// This is used as a map between the address of each allocation and its `AllocId`.
     /// It is always sorted
-    pub int_to_ptr_map: Vec<(u64, AllocId)>,
+    int_to_ptr_map: Vec<(u64, AllocId)>,
     /// The base address for each allocation.  We cannot put that into
     /// `AllocExtra` because function pointers also have a base address, and
     /// they do not have an `AllocExtra`.
     /// This is the inverse of `int_to_ptr_map`.
-    pub base_addr: FxHashMap<AllocId, u64>,
+    base_addr: FxHashMap<AllocId, u64>,
     /// This is used as a memory address when a new pointer is casted to an integer. It
     /// is always larger than any address that was previously made part of a block.
-    pub next_base_addr: u64,
+    next_base_addr: u64,
+    /// Whether to enforce "strict provenance" rules. Enabling this means int2ptr casts return
+    /// pointers with an invalid provenance, i.e., not valid for any memory access.
+    strict_provenance: bool,
 }
 
-impl Default for GlobalState {
-    fn default() -> Self {
+impl GlobalState {
+    pub fn new(config: &MiriConfig) -> Self {
         GlobalState {
             int_to_ptr_map: Vec::default(),
             base_addr: FxHashMap::default(),
             next_base_addr: STACK_ADDR,
+            strict_provenance: config.strict_provenance,
         }
     }
 }
@@ -43,8 +47,12 @@ impl<'mir, 'tcx> GlobalState {
     ) -> Pointer<Option<Tag>> {
         trace!("Casting 0x{:x} to a pointer", addr);
         let global_state = memory.extra.intptrcast.borrow();
-        let pos = global_state.int_to_ptr_map.binary_search_by_key(&addr, |(addr, _)| *addr);
 
+        if global_state.strict_provenance {
+            return Pointer::new(None, Size::from_bytes(addr));
+        }
+
+        let pos = global_state.int_to_ptr_map.binary_search_by_key(&addr, |(addr, _)| *addr);
         let alloc_id = match pos {
             Ok(pos) => Some(global_state.int_to_ptr_map[pos].1),
             Err(0) => None,

src/machine.rs

@@ -205,7 +205,7 @@ impl MemoryExtra {
         MemoryExtra {
             stacked_borrows,
             data_race,
-            intptrcast: Default::default(),
+            intptrcast: RefCell::new(intptrcast::GlobalState::new(config)),
             extern_statics: FxHashMap::default(),
             rng: RefCell::new(rng),
             tracked_alloc_id: config.tracked_alloc_id,

tests/compile-fail/box-cell-alias.rs

@@ -1,4 +1,4 @@
-// compile-flags: -Zmiri-tag-raw-pointers
+// compile-flags: -Zmiri-strict-provenance
 
 // Taken from <https://github.com/rust-lang/unsafe-code-guidelines/issues/194#issuecomment-520934222>.
 

tests/compile-fail/stacked_borrows/zst_slice.rs

@@ -1,4 +1,4 @@
-// compile-flags: -Zmiri-tag-raw-pointers
+// compile-flags: -Zmiri-strict-provenance
 // error-pattern: does not exist in the borrow stack
 
 fn main() {

tests/compile-fail/strict-provenance-offset.rs

@@ -0,0 +1,9 @@
+// compile-flags: -Zmiri-strict-provenance
+// error-pattern: not a valid pointer
+
+fn main() {
+    let x = 22;
+    let ptr = &x as *const _ as *const u8;
+    let roundtrip = ptr as usize as *const u8;
+    let _ = unsafe { roundtrip.offset(1) };
+}

tests/run-pass/btreemap.rs

@@ -1,4 +1,4 @@
-// compile-flags: -Zmiri-tag-raw-pointers
+// compile-flags: -Zmiri-strict-provenance -Zmiri-check-number-validity
 #![feature(btree_drain_filter)]
 use std::collections::{BTreeMap, BTreeSet};
 use std::mem;

tests/run-pass/concurrency/channels.rs

@@ -1,5 +1,4 @@
 // ignore-windows: Concurrency on Windows is not supported yet.
-// compile-flags: -Zmiri-disable-isolation
 
 use std::sync::mpsc::{channel, sync_channel};
 use std::thread;