thanks for Ian. I got it!
in arm_cond_branch and some pattern have code to reduce code by
combine the condition code to next instruction.
2009/10/3 kito <npickito@xxxxxxxxx>:
> thanks for your reply.
> my question is, I don't know why the conditional execution
> instruction is generated without cond_exec in there.
>
> @(insn 82 729 83 ../../../gcc-live/libgcc/../gcc/libgcc2.c:1848 (set
> (reg:QI 18 r18 [320])
> @ (const_int 0 [0x0])) 169 {*arm_movqi_insn}
> (expr_list:REG_EQUAL (const_int 0 [0x0])
> @ (nil)))
> moveq r18, #0 @ 82 *arm_movqi_insn/1 [length = 4] <---
>
> I have try to remove the predicable attribute but it's still generate same code
>
>>>> 2009/10/2 Ian Lance Taylor <iant@xxxxxxxxxx>:
>>>>> kito <npickito@xxxxxxxxx> writes:
>>>>>
>>>>>> I have some question about Conditional Execution Pattern.
>>>>>> In fact, I want to remove Conditional Execution in ARM for experimentation.
>>>>>> I have remove lots of Conditional Execution Pattern from arm.md
>>>>>> But I see a RTX that I can't understand.
>>>>>>
>>>>>> @(insn 74 79 80 ../../../gcc-live/libgcc/../gcc/libgcc2.c:1848 (set
>>>>>> (reg:QI 18 r18 [320])
>>>>>> @ (const_int 1 [0x1])) 169 {*arm_movqi_insn}
>>>>>> (expr_list:REG_EQUAL (const_int 1 [0x1])
>>>>>> @ (nil)))
>>>>>> mov r18, #1 @ 74 *arm_movqi_insn/1 [length = 4]
>>>>>> @(jump_insn 80 74 729 ../../../gcc-live/libgcc/../gcc/libgcc2.c:1848 (set (pc)
>>>>>> @ (if_then_else (ne (reg:CC 40 cc)
>>>>>> @ (const_int 0 [0x0]))
>>>>>> @ (label_ref 83)
>>>>>> @ (pc))) 220 {*arm_cond_branch} (expr_list:REG_DEAD (reg:CC 40 cc)
>>>>>> @ (expr_list:REG_BR_PROB (const_int 5000 [0x1388])
>>>>>> @ (nil)))
>>>>>> -> 83)
>>>>>> @(insn 82 729 83 ../../../gcc-live/libgcc/../gcc/libgcc2.c:1848 (set
>>>>>> (reg:QI 18 r18 [320])
>>>>>> @ (const_int 0 [0x0])) 169 {*arm_movqi_insn}
>>>>>> (expr_list:REG_EQUAL (const_int 0 [0x0])
>>>>>> @ (nil)))
>>>>>> moveq r18, #0 @ 82 *arm_movqi_insn/1 [length = 4]
>>>>>>
>>>>>> I don't know the expr_list:REG_EQUAL is how to generate, I have seem
>>>>>> it in expand pass.
>>>>>
>>>>>
>>>>> I'm not sure what your actual question is. REG_EQUAL is documented in
>>>>> the internals manual and in reg-notes.def.
>>>>>
>>>>> Ian
>>>>>
>>>
>>
>
