x64: port atomic_rmw to ISLE (#4389)

* x64: port `atomic_rmw` to ISLE

This change ports `atomic_rmw` to ISLE for the x64 backend. It does not
change the lowering in any way, though it seems possible that the fixed
regs need not be as fixed and that there are opportunities for single
instruction lowerings. It does rename `inst_common::AtomicRmwOp` to
`MachAtomicRmwOp` to disambiguate with the IR enum with the same name.

* x64: remove remaining hardcoded register constraints for `atomic_rmw`

* x64: use `SyntheticAmode` in `AtomicRmwSeq`

* review: add missing reg collector for amode

* review: collect memory registers in the 'late' phase

cranelift/codegen/src/machinst/reg.rs

@@ -328,6 +328,11 @@ impl<'a, F: Fn(VReg) -> VReg> OperandCollector<'a, F> {
         self.add_operand(Operand::reg_use(reg.into()));
     }
 
+    /// Add a register use, at the end of the instruction (`After` position).
+    pub fn reg_late_use(&mut self, reg: Reg) {
+        self.add_operand(Operand::reg_use_at_end(reg.into()));
+    }
+
     /// Add multiple register uses.
     pub fn reg_uses(&mut self, regs: &[Reg]) {
         for &reg in regs {