aarch64: Avoid a copy in the extract_vector lowering (#6252)
* Avoid a move in the aarch64 lowering of extract_vector * Update tests
This commit is contained in:
@@ -97,7 +97,7 @@
|
||||
|
||||
;;; Rules for `extract_vector` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
|
||||
(rule (lower (extract_vector x 0))
|
||||
(value_reg (fpu_move_128 (put_in_reg x))))
|
||||
(value_reg (put_in_reg x)))
|
||||
|
||||
;;;; Rules for `swiden_high` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
|
||||
|
||||
|
||||
Reference in New Issue
Block a user