aarch64: Migrate ishl/ushr/sshr to ISLE (#3608)

* aarch64: Migrate ishl/ushr/sshr to ISLE

This commit migrates the `ishl`, `ushr`, and `sshr` instructions to
ISLE. These involve special cases for almost all types of integers
(including vectors) and helper functions for the i128 lowerings since
the i128 lowerings look to be used for other instructions as well. This
doesn't delete the i128 lowerings in the Rust code just yet because
they're still used by Rust lowerings, but they should be deletable in
due time once those lowerings are translated to ISLE.

* Use more descriptive names for i128 lowerings

* Use a with_flags-lookalike for csel

* Use existing `with_flags_*`

* Coment backwards order

* Update generated code

cranelift/codegen/src/prelude.isle

@@ -148,6 +148,10 @@
 
 ;;;; Helper Clif Extractors ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 
+;; An extractor that only matches types that can fit in 16 bits.
+(decl fits_in_16 (Type) Type)
+(extern extractor fits_in_16 fits_in_16)
+
 ;; An extractor that only matches types that can fit in 32 bits.
 (decl fits_in_32 (Type) Type)
 (extern extractor fits_in_32 fits_in_32)
@@ -287,12 +291,15 @@
 ;; `ValueRegs` containing the first consumer's result and then the second
 ;; consumer's result.
 (decl with_flags_2 (ProducesFlags ConsumesFlags ConsumesFlags) ValueRegs)
-(rule (with_flags_2 (ProducesFlags.ProducesFlags producer_inst producer_result)
+(rule (with_flags_2 (ProducesFlags.ProducesFlags producer_inst _producer_result)
                     (ConsumesFlags.ConsumesFlags consumer_inst_1 consumer_result_1)
                     (ConsumesFlags.ConsumesFlags consumer_inst_2 consumer_result_2))
       (let ((_x Unit (emit producer_inst))
-            (_y Unit (emit consumer_inst_1))
-            (_z Unit (emit consumer_inst_2)))
+            ;; Note that the order of emission here is swapped, as this seems
+            ;; to generate better register allocation for now with fewer
+            ;; `mov` instructions.
+            (_y Unit (emit consumer_inst_2))
+            (_z Unit (emit consumer_inst_1)))
         (value_regs consumer_result_1 consumer_result_2)))
 
 ;;;; Helpers for Working with TrapCode ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;