[Bug target/86314] [7/8/9 Regression] GCC 7.x and 8.x zero out "eax" before using "rax" in "lock bts"

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86314

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |uros at gcc dot gnu.org

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Goes wrong in peephole2:

Splitting with gen_peephole2_28

;; Convert setcc + movzbl to xor + setcc if operands don't overlap.

...
(define_peephole2
  [(parallel [(set (reg FLAGS_REG) (match_operand 0))
              (match_operand 4)])
   (set (match_operand:QI 1 "register_operand")
        (match_operator:QI 2 "ix86_comparison_operator"
          [(reg FLAGS_REG) (const_int 0)]))
   (set (match_operand 3 "any_QIreg_operand")
        (zero_extend (match_dup 1)))]
  "(peep2_reg_dead_p (3, operands[1])
    || operands_match_p (operands[1], operands[3]))
   && ! reg_overlap_mentioned_p (operands[3], operands[0])
   && ! reg_set_p (operands[3], operands[4])
   && peep2_regno_dead_p (0, FLAGS_REG)"
  [(parallel [(set (match_dup 5) (match_dup 0))
              (match_dup 4)])
   (set (strict_low_part (match_dup 6))
        (match_dup 2))]
{
  operands[5] = gen_rtx_REG (GET_MODE (operands[0]), FLAGS_REG);
  operands[6] = gen_lowpart (QImode, operands[3]);
  ix86_expand_clear (operands[3]);
})

looks like the peephole was added by Uros (and refined multiple times already)