Auto merge of rust-lang#120080 - cuviper:128-align-packed, r=<try>

[WIP] pack u128 in the compiler to mitigate new alignment

This is based on rust-lang#116672, adding a new `#[repr(packed(8))]` wrapper on `u128` to avoid changing any of the compiler's size assertions. This is needed in two places:

* `SwitchTargets`, otherwise its `SmallVec<[u128; 1]>` gets padded up to 32 bytes.
* `LitKind::Int`, so that entire `enum` can stay 24 bytes.
  * This change definitely has far-reaching effects though, since it's public.

r? ghost

Author: bors

compiler/rustc_ast/src/ast.rs

@@ -27,6 +27,7 @@ pub use UnsafeSource::*;
 use crate::ptr::P;
 use crate::token::{self, CommentKind, Delimiter};
 use crate::tokenstream::{DelimSpan, LazyAttrTokenStream, TokenStream};
+use rustc_data_structures::packed::Pu128;
 use rustc_data_structures::stable_hasher::{HashStable, StableHasher};
 use rustc_data_structures::stack::ensure_sufficient_stack;
 use rustc_data_structures::sync::Lrc;
@@ -1829,7 +1830,7 @@ pub enum LitKind {
     /// A character literal (`'a'`).
     Char(char),
     /// An integer literal (`1`).
-    Int(u128, LitIntType),
+    Int(Pu128, LitIntType),
     /// A float literal (`1.0`, `1f64` or `1E10f64`). The pre-suffix part is
     /// stored as a symbol rather than `f64` so that `LitKind` can impl `Eq`
     /// and `Hash`.

compiler/rustc_ast/src/util/literal.rs

@@ -374,7 +374,7 @@ fn integer_lit(symbol: Symbol, suffix: Option<Symbol>) -> Result<LitKind, LitErr
     };
 
     let s = &s[if base != 10 { 2 } else { 0 }..];
-    u128::from_str_radix(s, base).map(|i| LitKind::Int(i, ty)).map_err(|_| {
+    u128::from_str_radix(s, base).map(|i| LitKind::Int(i.into(), ty)).map_err(|_| {
         // Small bases are lexed as if they were base 10, e.g, the string
         // might be `0b10201`. This will cause the conversion above to fail,
         // but these kinds of errors are already reported by the lexer.

compiler/rustc_ast_lowering/src/expr.rs

@@ -1908,15 +1908,21 @@ impl<'hir> LoweringContext<'_, 'hir> {
     pub(super) fn expr_usize(&mut self, sp: Span, value: usize) -> hir::Expr<'hir> {
         let lit = self.arena.alloc(hir::Lit {
             span: sp,
-            node: ast::LitKind::Int(value as u128, ast::LitIntType::Unsigned(ast::UintTy::Usize)),
+            node: ast::LitKind::Int(
+                (value as u128).into(),
+                ast::LitIntType::Unsigned(ast::UintTy::Usize),
+            ),
         });
         self.expr(sp, hir::ExprKind::Lit(lit))
     }
 
     pub(super) fn expr_u32(&mut self, sp: Span, value: u32) -> hir::Expr<'hir> {
         let lit = self.arena.alloc(hir::Lit {
             span: sp,
-            node: ast::LitKind::Int(value.into(), ast::LitIntType::Unsigned(ast::UintTy::U32)),
+            node: ast::LitKind::Int(
+                u128::from(value).into(),
+                ast::LitIntType::Unsigned(ast::UintTy::U32),
+            ),
         });
         self.expr(sp, hir::ExprKind::Lit(lit))
     }

compiler/rustc_attr/src/builtin.rs

@@ -1185,9 +1185,9 @@ fn allow_unstable<'a>(
 
 pub fn parse_alignment(node: &ast::LitKind) -> Result<u32, &'static str> {
     if let ast::LitKind::Int(literal, ast::LitIntType::Unsuffixed) = node {
-        if literal.is_power_of_two() {
+        if literal.get().is_power_of_two() {
             // rustc_middle::ty::layout::Align restricts align to <= 2^29
-            if *literal <= 1 << 29 { Ok(*literal as u32) } else { Err("larger than 2^29") }
+            if *literal <= 1 << 29 { Ok(literal.get() as u32) } else { Err("larger than 2^29") }
         } else {
             Err("not a power of two")
         }

compiler/rustc_attr/src/session_diagnostics.rs

@@ -296,7 +296,7 @@ impl<'a> IncorrectReprFormatGenericCause<'a> {
     pub fn from_lit_kind(span: Span, kind: &ast::LitKind, name: &'a str) -> Option<Self> {
         match kind {
             ast::LitKind::Int(int, ast::LitIntType::Unsuffixed) => {
-                Some(Self::Int { span, name, int: *int })
+                Some(Self::Int { span, name, int: int.get() })
             }
             ast::LitKind::Str(symbol, _) => Some(Self::Symbol { span, name, symbol: *symbol }),
             _ => None,

compiler/rustc_builtin_macros/src/concat_bytes.rs

@@ -54,7 +54,7 @@ fn invalid_type_err(
             val,
             ast::LitIntType::Unsuffixed | ast::LitIntType::Unsigned(ast::UintTy::U8),
         )) => {
-            assert!(val > u8::MAX.into()); // must be an error
+            assert!(val.get() > u8::MAX.into()); // must be an error
             dcx.emit_err(ConcatBytesOob { span });
         }
         Ok(ast::LitKind::Int(_, _)) => {
@@ -86,7 +86,7 @@ fn handle_array_element(
             Ok(ast::LitKind::Int(
                 val,
                 ast::LitIntType::Unsuffixed | ast::LitIntType::Unsigned(ast::UintTy::U8),
-            )) if val <= u8::MAX.into() => Some(val as u8),
+            )) if val.get() <= u8::MAX.into() => Some(val.get() as u8),
 
             Ok(ast::LitKind::Byte(val)) => Some(val),
             Ok(ast::LitKind::ByteStr(..)) => {
@@ -148,7 +148,7 @@ pub fn expand_concat_bytes(
                     if let Some(elem) =
                         handle_array_element(cx, &mut has_errors, &mut missing_literals, expr)
                     {
-                        for _ in 0..count_val {
+                        for _ in 0..count_val.get() {
                             accumulator.push(elem);
                         }
                     }

compiler/rustc_codegen_llvm/src/context.rs

@@ -145,6 +145,14 @@ pub unsafe fn create_module<'ll>(
                 .replace("-Fi64", "");
         }
     }
+    if llvm_version < (18, 0, 0) {
+        if sess.target.arch == "x86" || sess.target.arch == "x86_64" {
+            // LLVM 18 adjusts i128 to be 128-bit aligned on x86 variants.
+            // Earlier LLVMs leave this as default alignment, so remove it.
+            // See https://reviews.llvm.org/D86310
+            target_data_layout = target_data_layout.replace("-i128:128", "");
+        }
+    }
 
     // Ensure the data-layout values hardcoded remain the defaults.
     if sess.target.is_builtin {

compiler/rustc_codegen_ssa/src/codegen_attrs.rs

@@ -658,7 +658,7 @@ fn check_link_ordinal(tcx: TyCtxt<'_>, attr: &ast::Attribute) -> Option<u16> {
         // if the resulting EXE runs, as I haven't yet built the necessary DLL -- see earlier comment
         // about LINK.EXE failing.)
         if *ordinal <= u16::MAX as u128 {
-            Some(*ordinal as u16)
+            Some(ordinal.get() as u16)
         } else {
             let msg = format!("ordinal value in `link_ordinal` is too large: `{}`", &ordinal);
             tcx.dcx()

compiler/rustc_data_structures/src/lib.rs

@@ -93,6 +93,7 @@ pub mod aligned;
 pub mod frozen;
 mod hashes;
 pub mod owned_slice;
+pub mod packed;
 pub mod sso;
 pub mod steal;
 pub mod tagged_ptr;

compiler/rustc_data_structures/src/packed.rs

@@ -0,0 +1,71 @@
+use crate::stable_hasher::{HashStable, StableHasher};
+use rustc_serialize::{Decodable, Decoder, Encodable, Encoder};
+use std::cmp::Ordering;
+use std::fmt;
+
+#[repr(packed(8))]
+#[derive(Copy, Clone, Debug, Hash, PartialEq, Eq, PartialOrd, Ord)]
+pub struct Pu128(pub u128);
+
+impl Pu128 {
+    #[inline]
+    pub fn get(self) -> u128 {
+        self.0
+    }
+}
+
+impl From<u128> for Pu128 {
+    #[inline]
+    fn from(value: u128) -> Self {
+        Self(value)
+    }
+}
+
+impl PartialEq<u128> for Pu128 {
+    #[inline]
+    fn eq(&self, other: &u128) -> bool {
+        ({ self.0 }) == *other
+    }
+}
+
+impl PartialOrd<u128> for Pu128 {
+    #[inline]
+    fn partial_cmp(&self, other: &u128) -> Option<Ordering> {
+        { self.0 }.partial_cmp(other)
+    }
+}
+
+impl fmt::Display for Pu128 {
+    #[inline]
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        { self.0 }.fmt(f)
+    }
+}
+
+impl fmt::UpperHex for Pu128 {
+    #[inline]
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        { self.0 }.fmt(f)
+    }
+}
+
+impl<CTX> HashStable<CTX> for Pu128 {
+    #[inline]
+    fn hash_stable(&self, ctx: &mut CTX, hasher: &mut StableHasher) {
+        { self.0 }.hash_stable(ctx, hasher)
+    }
+}
+
+impl<S: Encoder> Encodable<S> for Pu128 {
+    #[inline]
+    fn encode(&self, s: &mut S) {
+        { self.0 }.encode(s);
+    }
+}
+
+impl<D: Decoder> Decodable<D> for Pu128 {
+    #[inline]
+    fn decode(d: &mut D) -> Self {
+        Self(u128::decode(d))
+    }
+}

compiler/rustc_expand/src/mbe/metavar_expr.rs

@@ -124,7 +124,7 @@ fn parse_depth<'sess>(
     };
     if let Ok(lit_kind) = LitKind::from_token_lit(*lit)
         && let LitKind::Int(n_u128, LitIntType::Unsuffixed) = lit_kind
-        && let Ok(n_usize) = usize::try_from(n_u128)
+        && let Ok(n_usize) = usize::try_from(n_u128.get())
     {
         Ok(n_usize)
     } else {

compiler/rustc_hir_typeck/src/expr.rs

@@ -2982,10 +2982,11 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
                         // fixed expression:
                         if let ExprKind::Lit(lit) = idx.kind
                             && let ast::LitKind::Int(i, ast::LitIntType::Unsuffixed) = lit.node
-                            && i < types
-                                .len()
-                                .try_into()
-                                .expect("expected tuple index to be < usize length")
+                            && i.get()
+                                < types
+                                    .len()
+                                    .try_into()
+                                    .expect("expected tuple index to be < usize length")
                         {
                             err.span_suggestion(
                                 brackets_span,

compiler/rustc_hir_typeck/src/fn_ctxt/suggestions.rs

@@ -11,6 +11,7 @@ use crate::ty::TypeAndMut;
 use core::cmp::min;
 use core::iter;
 use rustc_ast::util::parser::{ExprPrecedence, PREC_POSTFIX};
+use rustc_data_structures::packed::Pu128;
 use rustc_errors::{Applicability, Diagnostic, MultiSpan};
 use rustc_hir as hir;
 use rustc_hir::def::Res;
@@ -1409,8 +1410,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
                 }
                 let (_, suffix) = snippet.split_at(snippet.len() - 3);
                 let value = match suffix {
-                    "f32" => (lit - 0xf32) / (16 * 16 * 16),
-                    "f64" => (lit - 0xf64) / (16 * 16 * 16),
+                    "f32" => (lit.get() - 0xf32) / (16 * 16 * 16),
+                    "f64" => (lit.get() - 0xf64) / (16 * 16 * 16),
                     _ => return false,
                 };
                 err.span_suggestions(
@@ -1440,7 +1441,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
         };
 
         // Provided expression needs to be a literal `0`.
-        let ExprKind::Lit(Spanned { node: rustc_ast::LitKind::Int(0, _), span }) = expr.kind else {
+        let ExprKind::Lit(Spanned { node: rustc_ast::LitKind::Int(Pu128(0), _), span }) = expr.kind
+        else {
             return false;
         };
 

compiler/rustc_hir_typeck/src/op.rs

@@ -4,6 +4,7 @@ use super::method::MethodCallee;
 use super::{has_expected_num_generic_args, FnCtxt};
 use crate::Expectation;
 use rustc_ast as ast;
+use rustc_data_structures::packed::Pu128;
 use rustc_errors::{struct_span_code_err, Applicability, Diagnostic, DiagnosticBuilder};
 use rustc_hir as hir;
 use rustc_infer::infer::type_variable::{TypeVariableOrigin, TypeVariableOriginKind};
@@ -834,7 +835,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
                                     hir::Expr {
                                         kind:
                                             hir::ExprKind::Lit(Spanned {
-                                                node: ast::LitKind::Int(1, _),
+                                                node: ast::LitKind::Int(Pu128(1), _),
                                                 ..
                                             }),
                                         ..

compiler/rustc_lint/src/invalid_from_utf8.rs

@@ -111,7 +111,7 @@ impl<'tcx> LateLintPass<'tcx> for InvalidFromUtf8 {
                         .map(|e| match &e.kind {
                             ExprKind::Lit(Spanned { node: lit, .. }) => match lit {
                                 LitKind::Byte(b) => Some(*b),
-                                LitKind::Int(b, _) => Some(*b as u8),
+                                LitKind::Int(b, _) => Some(b.get() as u8),
                                 _ => None,
                             },
                             _ => None,

compiler/rustc_lint/src/types.rs

@@ -492,7 +492,7 @@ fn lint_uint_literal<'tcx>(
     let lit_val: u128 = match lit.node {
         // _v is u8, within range by definition
         ast::LitKind::Byte(_v) => return,
-        ast::LitKind::Int(v, _) => v,
+        ast::LitKind::Int(v, _) => v.get(),
         _ => bug!(),
     };
     if lit_val < min || lit_val > max {
@@ -555,7 +555,7 @@ fn lint_literal<'tcx>(
         ty::Int(t) => {
             match lit.node {
                 ast::LitKind::Int(v, ast::LitIntType::Signed(_) | ast::LitIntType::Unsuffixed) => {
-                    lint_int_literal(cx, type_limits, e, lit, t, v)
+                    lint_int_literal(cx, type_limits, e, lit, t, v.get())
                 }
                 _ => bug!(),
             };
@@ -842,7 +842,7 @@ impl<'tcx> LateLintPass<'tcx> for TypeLimits {
                             ast::LitKind::Int(
                                 v,
                                 ast::LitIntType::Signed(_) | ast::LitIntType::Unsuffixed,
-                            ) => v as i128,
+                            ) => v.get() as i128,
                             _ => return true,
                         },
                         _ => bug!(),
@@ -853,7 +853,7 @@ impl<'tcx> LateLintPass<'tcx> for TypeLimits {
                     let (min, max): (u128, u128) = uint_ty_range(uint_ty);
                     let lit_val: u128 = match lit.kind {
                         hir::ExprKind::Lit(li) => match li.node {
-                            ast::LitKind::Int(v, _) => v,
+                            ast::LitKind::Int(v, _) => v.get(),
                             _ => return true,
                         },
                         _ => bug!(),

compiler/rustc_middle/src/mir/syntax.rs

@@ -13,6 +13,7 @@ use crate::ty::{self, List, Ty};
 use crate::ty::{Region, UserTypeAnnotationIndex};
 
 use rustc_ast::{InlineAsmOptions, InlineAsmTemplatePiece};
+use rustc_data_structures::packed::Pu128;
 use rustc_hir::def_id::DefId;
 use rustc_hir::{self, CoroutineKind};
 use rustc_index::IndexVec;
@@ -829,7 +830,7 @@ impl TerminatorKind<'_> {
 pub struct SwitchTargets {
     /// Possible values. The locations to branch to in each case
     /// are found in the corresponding indices from the `targets` vector.
-    pub(super) values: SmallVec<[u128; 1]>,
+    pub(super) values: SmallVec<[Pu128; 1]>,
 
     /// Possible branch sites. The last element of this vector is used
     /// for the otherwise branch, so targets.len() == values.len() + 1

compiler/rustc_middle/src/mir/terminator.rs

@@ -3,6 +3,7 @@ use rustc_hir::LangItem;
 use smallvec::SmallVec;
 
 use super::TerminatorKind;
+use rustc_data_structures::packed::Pu128;
 use rustc_macros::HashStable;
 use std::slice;
 
@@ -14,15 +15,16 @@ impl SwitchTargets {
     /// The iterator may be empty, in which case the `SwitchInt` instruction is equivalent to
     /// `goto otherwise;`.
     pub fn new(targets: impl Iterator<Item = (u128, BasicBlock)>, otherwise: BasicBlock) -> Self {
-        let (values, mut targets): (SmallVec<_>, SmallVec<_>) = targets.unzip();
+        let (values, mut targets): (SmallVec<_>, SmallVec<_>) =
+            targets.map(|(v, t)| (Pu128(v), t)).unzip();
         targets.push(otherwise);
         Self { values, targets }
     }
 
     /// Builds a switch targets definition that jumps to `then` if the tested value equals `value`,
     /// and to `else_` if not.
     pub fn static_if(value: u128, then: BasicBlock, else_: BasicBlock) -> Self {
-        Self { values: smallvec![value], targets: smallvec![then, else_] }
+        Self { values: smallvec![Pu128(value)], targets: smallvec![then, else_] }
     }
 
     /// Inverse of `SwitchTargets::static_if`.
@@ -31,7 +33,7 @@ impl SwitchTargets {
         if let &[value] = &self.values[..]
             && let &[then, else_] = &self.targets[..]
         {
-            Some((value, then, else_))
+            Some((value.get(), then, else_))
         } else {
             None
         }
@@ -75,15 +77,15 @@ impl SwitchTargets {
 }
 
 pub struct SwitchTargetsIter<'a> {
-    inner: iter::Zip<slice::Iter<'a, u128>, slice::Iter<'a, BasicBlock>>,
+    inner: iter::Zip<slice::Iter<'a, Pu128>, slice::Iter<'a, BasicBlock>>,
 }
 
 impl<'a> Iterator for SwitchTargetsIter<'a> {
     type Item = (u128, BasicBlock);
 
     #[inline]
     fn next(&mut self) -> Option<Self::Item> {
-        self.inner.next().map(|(val, bb)| (*val, *bb))
+        self.inner.next().map(|(val, bb)| (val.get(), *bb))
     }
 
     #[inline]

compiler/rustc_middle/src/ty/context.rs

@@ -745,7 +745,7 @@ impl<'tcx> TyCtxt<'tcx> {
                 ],
             ) = attr.meta_item_list().as_deref()
             {
-                Bound::Included(a)
+                Bound::Included(a.get())
             } else {
                 self.dcx().span_delayed_bug(
                     attr.span,