Align x86-64 stubs by whether the lazy-binding helper exists

ld-prime byte-aligns x86-64 __stubs only when they go through
__stub_helper; with chained fixups, -bind_at_load or weak-lookup
stubs alone they are 2-byte aligned. The classic-vs-chained rule we
had was the common case of that.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

src/passes.rs

@@ -3964,11 +3964,14 @@ pub fn create_output_sections<E: Target>(ctx: &mut Context<E>) {
     if !ctx.stubs.symbols.is_empty() {
         ctx.stubs.hdr.reserved2 = E::STUB_SIZE as u32;
         ctx.stubs.hdr.size = ctx.stubs.symbols.len() as u64 * E::STUB_SIZE;
-        // ld-prime's x86-64 stubs are 2-byte aligned with chained
-        // fixups and byte-aligned with classic dyld info; arm64's are
-        // instruction-aligned.
+        // ld-prime's x86-64 stubs are byte-aligned when they go
+        // through the lazy-binding helper and 2-byte aligned otherwise
+        // (chained fixups, -bind_at_load, weak-lookup stubs); arm64's
+        // are instruction-aligned.
         if E::CPUTYPE == crate::macho::CPU_TYPE_X86_64 {
-            ctx.stubs.hdr.p2align = if ctx.use_chained_fixups() { 1 } else { 0 };
+            let lazy = ctx.lazy_binding()
+                && ctx.stubs.symbols.iter().any(|&id| !ctx.binds_weak_lookup(id));
+            ctx.stubs.hdr.p2align = if lazy { 0 } else { 1 };
         }
         ctx.chunks.push(ChunkId::Stubs);
     }

tests/bind-at-load.sh

@@ -17,3 +17,8 @@ otool -l $t/exe > $t/lc
 not grep -q '__la_symbol_ptr' $t/lc
 dyld_info -fixups $t/exe | grep '__got .* bind .*_printf'
 $t/exe | grep hi
+# Without the lazy-binding helper, x86-64 stubs are 2-byte aligned as
+# with chained fixups.
+if [ $ARCH = x86_64 ]; then
+  grep -A8 'sectname __stubs' $t/lc | grep 'align 2^1'
+fi