cpubits: add 64-bit override for ARMv7 (#1419)

ARMv7 is one of the main architectures for which we've received requests
for 64-bit overrides in the past (see discussion on #826).

Though natively 32-bit, ARMv7 supports certain "doubleword" instructions
which model 64-bit values as a pair of 32-bit registers, e.g.
`ADDS`/`ADC` and `SUBS`/`SBC` for 2x32-bit addition/subtraction, as well
as `UMULL`/`SMULL` for widening multiplication with 2x32-bit outputs.

Many ARMv7 CPUs internally fetch 64-bits of instruction at once and can
move 64-bits of data via `LDRD`/`STRD` in one cycle on optimized paths.

Some high-performance ARMv7 CPUs internally combine the barrel shifter +
ALU to speed multi-word shifts.

If we use 64-bit implementations when targeting ARMv7, codegen is able
to leverage these optimizations.

Author: tarcieri

.github/workflows/cpubits.yml

@@ -52,6 +52,23 @@ jobs:
       - run: ${{ matrix.deps }}
       - run: cargo test --target ${{ matrix.target }}
 
+  # Test on foreign architectures using `cross test`
+  test-cross:
+    strategy:
+      matrix:
+        include:
+          - target: armv7-unknown-linux-gnueabi # ARM32
+          - target: powerpc-unknown-linux-gnu   # PPC32
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+      - uses: dtolnay/rust-toolchain@master
+        with:
+          toolchain: stable
+          targets: ${{ matrix.target }}
+      - run: cargo install cross
+      - run: cross test --target ${{ matrix.target }}
+
   # Test WASM using `wasm32-wasip1` target on `wasmtime`
   test-wasm:
     strategy:

cpubits/src/lib.rs

@@ -4,6 +4,7 @@
     html_logo_url = "https://raw.githubusercontent.com/RustCrypto/media/6ee8e381/logo.svg",
     html_favicon_url = "https://raw.githubusercontent.com/RustCrypto/media/6ee8e381/logo.svg"
 )]
+#![allow(clippy::doc_markdown)]
 
 //! # Supported bit sizes
 //!
@@ -93,7 +94,8 @@
 /// certain targets from 32-bit to 64-bit ones.
 ///
 /// This 64-bit promotion occurs if `any` of the following `cfg`s are true:
-/// - `target_family = "wasm"`
+/// - ARMv7: `all(target_arch = "arm", not(target_feature = "thumb-mode"))`
+/// - WASM: `target_arch = "wasm32"`
 #[macro_export]
 macro_rules! cpubits {
     // Only run the given block if we have selected a 16-bit word size, i.e. the code will be
@@ -173,9 +175,13 @@ macro_rules! cpubits {
         64 => { $( $tokens64:tt )* }
     ) => {
         $crate::cpubits! {
+            // `cfg` selector for 64-bit target overrides
+            // Implicitly `cfg(any(...))`
             #[cfg(enable_64bit(
-                // `cfg` selector for 64-bit targets (implicitly `any`)
-                target_arch = "wasm32"
+                // ARMv7
+                all(target_arch = "arm", not(target_feature = "thumb-mode")),
+                // WASM
+                target_arch = "wasm32",
             ))]
             16 => { $( $tokens32 )* }
             32 => { $( $tokens32 )* }
@@ -294,7 +300,12 @@ mod tests {
     /// Return the expected number of bits for the target.
     fn expected_bits() -> u32 {
         // Duplicated 64-bit override predicates need to go here
-        if cfg!(target_arch = "wasm32") {
+        if cfg!(any(
+            // ARMv7
+            all(target_arch = "arm", not(target_feature = "thumb-mode")),
+            // WASM
+            target_arch = "wasm32"
+        )) {
             64
         } else {
             detect_pointer_width()
@@ -306,6 +317,13 @@ mod tests {
         assert_eq!(detected_bits(), expected_bits());
     }
 
+    /// Explicit test for ARMv7 so we can see the predicate is working
+    #[cfg(all(target_arch = "arm", not(target_feature = "thumb-mode")))]
+    #[test]
+    fn cpubits_on_armv7_is_64bit() {
+        assert_eq!(detected_bits(), 64);
+    }
+
     /// Explicit test for WASM so we can see the predicate is working
     #[cfg(target_arch = "wasm32")]
     #[test]