Merge subtree update for toolchain nightly-2025-08-26

.github/workflows/flux.yml

@@ -9,7 +9,7 @@ on:
 
 env:
   FIXPOINT_VERSION: "556104ba5508891c357b0bdf819ce706e93d9349"
-  FLUX_VERSION: "a17246965a8752e3d3d4e3559865311048bb61f7"
+  FLUX_VERSION: "b0cec81c42bc6e210f675b46dd5b4b16774b0d0e"
 
 jobs:
   check-flux-on-core:

library/Cargo.toml

@@ -59,4 +59,3 @@ rustflags = ["-Cpanic=abort"]
 rustc-std-workspace-core = { path = 'rustc-std-workspace-core' }
 rustc-std-workspace-alloc = { path = 'rustc-std-workspace-alloc' }
 rustc-std-workspace-std = { path = 'rustc-std-workspace-std' }
-compiler_builtins = { path = "compiler-builtins/compiler-builtins" }

library/alloc/src/alloc.rs

@@ -17,6 +17,7 @@ unsafe extern "Rust" {
     #[rustc_allocator]
     #[rustc_nounwind]
     #[rustc_std_internal_symbol]
+    #[rustc_allocator_zeroed_variant = "__rust_alloc_zeroed"]
     fn __rust_alloc(size: usize, align: usize) -> *mut u8;
     #[rustc_deallocator]
     #[rustc_nounwind]

library/alloc/src/collections/btree/map.rs

@@ -40,30 +40,15 @@ pub(super) const MIN_LEN: usize = node::MIN_LEN_AFTER_SPLIT;
 
 /// An ordered map based on a [B-Tree].
 ///
-/// B-Trees represent a fundamental compromise between cache-efficiency and actually minimizing
-/// the amount of work performed in a search. In theory, a binary search tree (BST) is the optimal
-/// choice for a sorted map, as a perfectly balanced BST performs the theoretical minimum amount of
-/// comparisons necessary to find an element (log<sub>2</sub>n). However, in practice the way this
-/// is done is *very* inefficient for modern computer architectures. In particular, every element
-/// is stored in its own individually heap-allocated node. This means that every single insertion
-/// triggers a heap-allocation, and every single comparison should be a cache-miss. Since these
-/// are both notably expensive things to do in practice, we are forced to, at the very least,
-/// reconsider the BST strategy.
-///
-/// A B-Tree instead makes each node contain B-1 to 2B-1 elements in a contiguous array. By doing
-/// this, we reduce the number of allocations by a factor of B, and improve cache efficiency in
-/// searches. However, this does mean that searches will have to do *more* comparisons on average.
-/// The precise number of comparisons depends on the node search strategy used. For optimal cache
-/// efficiency, one could search the nodes linearly. For optimal comparisons, one could search
-/// the node using binary search. As a compromise, one could also perform a linear search
-/// that initially only checks every i<sup>th</sup> element for some choice of i.
+/// Given a key type with a [total order], an ordered map stores its entries in key order.
+/// That means that keys must be of a type that implements the [`Ord`] trait,
+/// such that two keys can always be compared to determine their [`Ordering`].
+/// Examples of keys with a total order are strings with lexicographical order,
+/// and numbers with their natural order.
 ///
-/// Currently, our implementation simply performs naive linear search. This provides excellent
-/// performance on *small* nodes of elements which are cheap to compare. However in the future we
-/// would like to further explore choosing the optimal search strategy based on the choice of B,
-/// and possibly other factors. Using linear search, searching for a random element is expected
-/// to take B * log(n) comparisons, which is generally worse than a BST. In practice,
-/// however, performance is excellent.
+/// Iterators obtained from functions such as [`BTreeMap::iter`], [`BTreeMap::into_iter`], [`BTreeMap::values`], or
+/// [`BTreeMap::keys`] produce their items in key order, and take worst-case logarithmic and
+/// amortized constant time per item returned.
 ///
 /// It is a logic error for a key to be modified in such a way that the key's ordering relative to
 /// any other key, as determined by the [`Ord`] trait, changes while it is in the map. This is
@@ -72,14 +57,6 @@ pub(super) const MIN_LEN: usize = node::MIN_LEN_AFTER_SPLIT;
 /// `BTreeMap` that observed the logic error and not result in undefined behavior. This could
 /// include panics, incorrect results, aborts, memory leaks, and non-termination.
 ///
-/// Iterators obtained from functions such as [`BTreeMap::iter`], [`BTreeMap::into_iter`], [`BTreeMap::values`], or
-/// [`BTreeMap::keys`] produce their items in order by key, and take worst-case logarithmic and
-/// amortized constant time per item returned.
-///
-/// [B-Tree]: https://en.wikipedia.org/wiki/B-tree
-/// [`Cell`]: core::cell::Cell
-/// [`RefCell`]: core::cell::RefCell
-///
 /// # Examples
 ///
 /// ```
@@ -169,6 +146,43 @@ pub(super) const MIN_LEN: usize = node::MIN_LEN_AFTER_SPLIT;
 /// // modify an entry before an insert with in-place mutation
 /// player_stats.entry("mana").and_modify(|mana| *mana += 200).or_insert(100);
 /// ```
+///
+/// # Background
+///
+/// A B-tree is (like) a [binary search tree], but adapted to the natural granularity that modern
+/// machines like to consume data at. This means that each node contains an entire array of elements,
+/// instead of just a single element.
+///
+/// B-Trees represent a fundamental compromise between cache-efficiency and actually minimizing
+/// the amount of work performed in a search. In theory, a binary search tree (BST) is the optimal
+/// choice for a sorted map, as a perfectly balanced BST performs the theoretical minimum number of
+/// comparisons necessary to find an element (log<sub>2</sub>n). However, in practice the way this
+/// is done is *very* inefficient for modern computer architectures. In particular, every element
+/// is stored in its own individually heap-allocated node. This means that every single insertion
+/// triggers a heap-allocation, and every comparison is a potential cache-miss due to the indirection.
+/// Since both heap-allocations and cache-misses are notably expensive in practice, we are forced to,
+/// at the very least, reconsider the BST strategy.
+///
+/// A B-Tree instead makes each node contain B-1 to 2B-1 elements in a contiguous array. By doing
+/// this, we reduce the number of allocations by a factor of B, and improve cache efficiency in
+/// searches. However, this does mean that searches will have to do *more* comparisons on average.
+/// The precise number of comparisons depends on the node search strategy used. For optimal cache
+/// efficiency, one could search the nodes linearly. For optimal comparisons, one could search
+/// the node using binary search. As a compromise, one could also perform a linear search
+/// that initially only checks every i<sup>th</sup> element for some choice of i.
+///
+/// Currently, our implementation simply performs naive linear search. This provides excellent
+/// performance on *small* nodes of elements which are cheap to compare. However in the future we
+/// would like to further explore choosing the optimal search strategy based on the choice of B,
+/// and possibly other factors. Using linear search, searching for a random element is expected
+/// to take B * log(n) comparisons, which is generally worse than a BST. In practice,
+/// however, performance is excellent.
+///
+/// [B-Tree]: https://en.wikipedia.org/wiki/B-tree
+/// [binary search tree]: https://en.wikipedia.org/wiki/Binary_search_tree
+/// [total order]: https://en.wikipedia.org/wiki/Total_order
+/// [`Cell`]: core::cell::Cell
+/// [`RefCell`]: core::cell::RefCell
 #[stable(feature = "rust1", since = "1.0.0")]
 #[cfg_attr(not(test), rustc_diagnostic_item = "BTreeMap")]
 #[rustc_insignificant_dtor]

library/core/src/array/mod.rs

@@ -621,11 +621,11 @@ impl<T, const N: usize> [T; N] {
     /// assert_eq!(strings.len(), 3);
     /// ```
     #[stable(feature = "array_methods", since = "1.77.0")]
-    #[rustc_const_unstable(feature = "const_array_each_ref", issue = "133289")]
+    #[rustc_const_stable(feature = "const_array_each_ref", since = "CURRENT_RUSTC_VERSION")]
     pub const fn each_ref(&self) -> [&T; N] {
         let mut buf = [null::<T>(); N];
 
-        // FIXME(const-hack): We would like to simply use iterators for this (as in the original implementation), but this is not allowed in constant expressions.
+        // FIXME(const_trait_impl): We would like to simply use iterators for this (as in the original implementation), but this is not allowed in constant expressions.
         let mut i = 0;
         while i < N {
             buf[i] = &raw const self[i];
@@ -652,11 +652,11 @@ impl<T, const N: usize> [T; N] {
     /// assert_eq!(floats, [0.0, 2.7, -1.0]);
     /// ```
     #[stable(feature = "array_methods", since = "1.77.0")]
-    #[rustc_const_unstable(feature = "const_array_each_ref", issue = "133289")]
+    #[rustc_const_stable(feature = "const_array_each_ref", since = "CURRENT_RUSTC_VERSION")]
     pub const fn each_mut(&mut self) -> [&mut T; N] {
         let mut buf = [null_mut::<T>(); N];
 
-        // FIXME(const-hack): We would like to simply use iterators for this (as in the original implementation), but this is not allowed in constant expressions.
+        // FIXME(const_trait_impl): We would like to simply use iterators for this (as in the original implementation), but this is not allowed in constant expressions.
         let mut i = 0;
         while i < N {
             buf[i] = &raw mut self[i];

library/core/src/char/methods.rs

@@ -1874,28 +1874,33 @@ pub const unsafe fn encode_utf8_raw_unchecked(code: u32, dst: *mut u8) {
     // SAFETY: The caller must guarantee that the buffer pointed to by `dst`
     // is at least `len` bytes long.
     unsafe {
-        match len {
-            1 => {
-                *dst = code as u8;
-            }
-            2 => {
-                *dst = (code >> 6 & 0x1F) as u8 | TAG_TWO_B;
-                *dst.add(1) = (code & 0x3F) as u8 | TAG_CONT;
-            }
-            3 => {
-                *dst = (code >> 12 & 0x0F) as u8 | TAG_THREE_B;
-                *dst.add(1) = (code >> 6 & 0x3F) as u8 | TAG_CONT;
-                *dst.add(2) = (code & 0x3F) as u8 | TAG_CONT;
-            }
-            4 => {
-                *dst = (code >> 18 & 0x07) as u8 | TAG_FOUR_B;
-                *dst.add(1) = (code >> 12 & 0x3F) as u8 | TAG_CONT;
-                *dst.add(2) = (code >> 6 & 0x3F) as u8 | TAG_CONT;
-                *dst.add(3) = (code & 0x3F) as u8 | TAG_CONT;
-            }
-            // SAFETY: `char` always takes between 1 and 4 bytes to encode in UTF-8.
-            _ => crate::hint::unreachable_unchecked(),
+        if len == 1 {
+            *dst = code as u8;
+            return;
+        }
+
+        let last1 = (code >> 0 & 0x3F) as u8 | TAG_CONT;
+        let last2 = (code >> 6 & 0x3F) as u8 | TAG_CONT;
+        let last3 = (code >> 12 & 0x3F) as u8 | TAG_CONT;
+        let last4 = (code >> 18 & 0x3F) as u8 | TAG_FOUR_B;
+
+        if len == 2 {
+            *dst = last2 | TAG_TWO_B;
+            *dst.add(1) = last1;
+            return;
         }
+
+        if len == 3 {
+            *dst = last3 | TAG_THREE_B;
+            *dst.add(1) = last2;
+            *dst.add(2) = last1;
+            return;
+        }
+
+        *dst = last4;
+        *dst.add(1) = last3;
+        *dst.add(2) = last2;
+        *dst.add(3) = last1;
     }
 }
 

library/core/src/cmp.rs

@@ -1554,6 +1554,9 @@ pub fn min<T: Ord>(v1: T, v2: T) -> T {
 ///
 /// Returns the first argument if the comparison determines them to be equal.
 ///
+/// The parameter order is preserved when calling the `compare` function, i.e. `v1` is
+/// always passed as the first argument and `v2` as the second.
+///
 /// # Examples
 ///
 /// ```
@@ -1574,7 +1577,7 @@ pub fn min<T: Ord>(v1: T, v2: T) -> T {
 #[must_use]
 #[stable(feature = "cmp_min_max_by", since = "1.53.0")]
 pub fn min_by<T, F: FnOnce(&T, &T) -> Ordering>(v1: T, v2: T, compare: F) -> T {
-    if compare(&v2, &v1).is_lt() { v2 } else { v1 }
+    if compare(&v1, &v2).is_le() { v1 } else { v2 }
 }
 
 /// Returns the element that gives the minimum value from the specified function.
@@ -1646,6 +1649,9 @@ pub fn max<T: Ord>(v1: T, v2: T) -> T {
 ///
 /// Returns the second argument if the comparison determines them to be equal.
 ///
+/// The parameter order is preserved when calling the `compare` function, i.e. `v1` is
+/// always passed as the first argument and `v2` as the second.
+///
 /// # Examples
 ///
 /// ```
@@ -1666,7 +1672,7 @@ pub fn max<T: Ord>(v1: T, v2: T) -> T {
 #[must_use]
 #[stable(feature = "cmp_min_max_by", since = "1.53.0")]
 pub fn max_by<T, F: FnOnce(&T, &T) -> Ordering>(v1: T, v2: T, compare: F) -> T {
-    if compare(&v2, &v1).is_lt() { v1 } else { v2 }
+    if compare(&v1, &v2).is_gt() { v1 } else { v2 }
 }
 
 /// Returns the element that gives the maximum value from the specified function.
@@ -1745,6 +1751,9 @@ where
 ///
 /// Returns `[v1, v2]` if the comparison determines them to be equal.
 ///
+/// The parameter order is preserved when calling the `compare` function, i.e. `v1` is
+/// always passed as the first argument and `v2` as the second.
+///
 /// # Examples
 ///
 /// ```
@@ -1769,7 +1778,7 @@ pub fn minmax_by<T, F>(v1: T, v2: T, compare: F) -> [T; 2]
 where
     F: FnOnce(&T, &T) -> Ordering,
 {
-    if compare(&v2, &v1).is_lt() { [v2, v1] } else { [v1, v2] }
+    if compare(&v1, &v2).is_le() { [v1, v2] } else { [v2, v1] }
 }
 
 /// Returns minimum and maximum values with respect to the specified key function.

library/core/src/intrinsics/mod.rs

@@ -267,53 +267,72 @@ pub unsafe fn atomic_fence<const ORD: AtomicOrdering>();
 pub unsafe fn atomic_singlethreadfence<const ORD: AtomicOrdering>();
 
 /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
-/// if supported; otherwise, it is a no-op.
+/// for the given address if supported; otherwise, it is a no-op.
 /// Prefetches have no effect on the behavior of the program but can change its performance
 /// characteristics.
 ///
-/// The `locality` argument must be a constant integer and is a temporal locality specifier
-/// ranging from (0) - no locality, to (3) - extremely local keep in cache.
+/// The `LOCALITY` argument is a temporal locality specifier ranging from (0) - no locality,
+/// to (3) - extremely local keep in cache.
 ///
 /// This intrinsic does not have a stable counterpart.
 #[rustc_intrinsic]
 #[rustc_nounwind]
-pub unsafe fn prefetch_read_data<T>(data: *const T, locality: i32);
+#[miri::intrinsic_fallback_is_spec]
+pub const fn prefetch_read_data<T, const LOCALITY: i32>(data: *const T) {
+    // This operation is a no-op, unless it is overridden by the backend.
+    let _ = data;
+}
+
 /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
-/// if supported; otherwise, it is a no-op.
+/// for the given address if supported; otherwise, it is a no-op.
 /// Prefetches have no effect on the behavior of the program but can change its performance
 /// characteristics.
 ///
-/// The `locality` argument must be a constant integer and is a temporal locality specifier
-/// ranging from (0) - no locality, to (3) - extremely local keep in cache.
+/// The `LOCALITY` argument is a temporal locality specifier ranging from (0) - no locality,
+/// to (3) - extremely local keep in cache.
 ///
 /// This intrinsic does not have a stable counterpart.
 #[rustc_intrinsic]
 #[rustc_nounwind]
-pub unsafe fn prefetch_write_data<T>(data: *const T, locality: i32);
+#[miri::intrinsic_fallback_is_spec]
+pub const fn prefetch_write_data<T, const LOCALITY: i32>(data: *const T) {
+    // This operation is a no-op, unless it is overridden by the backend.
+    let _ = data;
+}
+
 /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
-/// if supported; otherwise, it is a no-op.
+/// for the given address if supported; otherwise, it is a no-op.
 /// Prefetches have no effect on the behavior of the program but can change its performance
 /// characteristics.
 ///
-/// The `locality` argument must be a constant integer and is a temporal locality specifier
-/// ranging from (0) - no locality, to (3) - extremely local keep in cache.
+/// The `LOCALITY` argument is a temporal locality specifier ranging from (0) - no locality,
+/// to (3) - extremely local keep in cache.
 ///
 /// This intrinsic does not have a stable counterpart.
 #[rustc_intrinsic]
 #[rustc_nounwind]
-pub unsafe fn prefetch_read_instruction<T>(data: *const T, locality: i32);
+#[miri::intrinsic_fallback_is_spec]
+pub const fn prefetch_read_instruction<T, const LOCALITY: i32>(data: *const T) {
+    // This operation is a no-op, unless it is overridden by the backend.
+    let _ = data;
+}
+
 /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
-/// if supported; otherwise, it is a no-op.
+/// for the given address if supported; otherwise, it is a no-op.
 /// Prefetches have no effect on the behavior of the program but can change its performance
 /// characteristics.
 ///
-/// The `locality` argument must be a constant integer and is a temporal locality specifier
-/// ranging from (0) - no locality, to (3) - extremely local keep in cache.
+/// The `LOCALITY` argument is a temporal locality specifier ranging from (0) - no locality,
+/// to (3) - extremely local keep in cache.
 ///
 /// This intrinsic does not have a stable counterpart.
 #[rustc_intrinsic]
 #[rustc_nounwind]
-pub unsafe fn prefetch_write_instruction<T>(data: *const T, locality: i32);
+#[miri::intrinsic_fallback_is_spec]
+pub const fn prefetch_write_instruction<T, const LOCALITY: i32>(data: *const T) {
+    // This operation is a no-op, unless it is overridden by the backend.
+    let _ = data;
+}
 
 /// Executes a breakpoint trap, for inspection by a debugger.
 ///