This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Question on operand_equal_p on different type conversion expressions
- From: "Bin.Cheng" <amker dot cheng at gmail dot com>
- To: Richard Biener <richard dot guenther at gmail dot com>
- Cc: Andrew Pinski <pinskia at gmail dot com>, "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Tue, 21 May 2013 17:48:31 +0800
- Subject: Re: Question on operand_equal_p on different type conversion expressions
- References: <CAHFci29t23t0hDJ3y8OA_o2jJrPDCoa=3g9FwKQjnpS34EPvYg at mail dot gmail dot com> <CA+=Sn1m+M3Z3=5XybvAkVM=1CqgQ9t5_Ly2tSEDe06=Cr52yCA at mail dot gmail dot com> <CAHFci29DkmuFp_Syo7yoaR-4gD+5xXx3t9yStjz=kAHACb9EZg at mail dot gmail dot com> <CAFiYyc363G_XZb19+A8Zh5POQmE+NAKKZdPcieHk=4gtxPAzGQ at mail dot gmail dot com>
On Tue, May 21, 2013 at 4:50 PM, Richard Biener
<richard.guenther@gmail.com> wrote:
> On Tue, May 21, 2013 at 8:38 AM, Bin.Cheng <amker.cheng@gmail.com> wrote:
>> On Tue, May 21, 2013 at 1:55 PM, Andrew Pinski <pinskia@gmail.com> wrote:
>>> On Mon, May 20, 2013 at 10:50 PM, Bin.Cheng <amker.cheng@gmail.com> wrote:
>>
>>>
>>>
>>> NOP_EXPR here is a misnamed tree really. It could also be a
>>> CONVERT_EXPR and still have the same issue as the types are not the
>>> same.
>>>
>>>
>>>> The problem is operand_equal_q simply return false because arg0/arg1
>>>> have different tree code.
>>>
>>> No it returns false because the types are two different. One is
>>> signed and the other is unsigned.
>>>
>>>>
>>>> Should operand_equal_q take two kinds of conversion expression into
>>>> consideration, or arg0/arg1 are not equal? Thanks.
>>>
>>> Yes why would it not? Look at the resulting types again.
>>
>> Thanks very much. The dumped tree codes are different (my mistake).
>> But the problem still exists in operand_equal_q.
>> For now with below tree nodes,
>> arg0:
>> <convert_expr 0xb72ddb04
>> type <integer_type 0xb74602a0 short int sizes-gimplified public HI
>> size <integer_cst 0xb744e7c4 constant 16>
>> unit size <integer_cst 0xb744e7e0 constant 2>
>> align 16 symtab 0 alias set 4 canonical type 0xb74602a0
>> precision 16 min <integer_cst 0xb744e770 -32768> max <integer_cst
>> 0xb744e78c 32767> context <translation_unit_decl 0xb760dd80 D.6120>
>> pointer_to_this <pointer_type 0xb7241600>>
>>
>> arg 0 <ssa_name 0xb72882f8
>> type <integer_type 0xb7460420 long int sizes-gimplified public SI
>> size <integer_cst 0xb744e55c constant 32>
>> unit size <integer_cst 0xb744e578 constant 4>
>> align 32 symtab 0 alias set 5 canonical type 0xb7460420
>> precision 32 min <integer_cst 0xb744e888 -2147483648> max <integer_cst
>> 0xb744e8a4 2147483647> context <translation_unit_decl 0xb760dd80
>> D.6120>
>> pointer_to_this <pointer_type 0xb74677e0>>
>> visiteddef_stmt _23 = *_22;
>>
>> version 23>>
>>
>> arg1:
>> <nop_expr 0xb72e1b54
>> type <integer_type 0xb74602a0 short int sizes-gimplified public HI
>> size <integer_cst 0xb744e7c4 constant 16>
>> unit size <integer_cst 0xb744e7e0 constant 2>
>> align 16 symtab 0 alias set 4 canonical type 0xb74602a0
>> precision 16 min <integer_cst 0xb744e770 -32768> max <integer_cst
>> 0xb744e78c 32767> context <translation_unit_decl 0xb760dd80 D.6120>
>> pointer_to_this <pointer_type 0xb7241600>>
>>
>> arg 0 <ssa_name 0xb72882f8
>> type <integer_type 0xb7460420 long int sizes-gimplified public SI
>> size <integer_cst 0xb744e55c constant 32>
>> unit size <integer_cst 0xb744e578 constant 4>
>> align 32 symtab 0 alias set 5 canonical type 0xb7460420
>> precision 32 min <integer_cst 0xb744e888 -2147483648> max <integer_cst
>> 0xb744e8a4 2147483647> context <translation_unit_decl 0xb760dd80
>> D.6120>
>> pointer_to_this <pointer_type 0xb74677e0>>
>> visiteddef_stmt _23 = *_22;
>>
>> version 23>>
>>
>> The operand_equal_p still returns false because below piece of code in it:
>>
>> #if 1
>> if (TREE_CODE (arg0) != TREE_CODE (arg1)
>> /* This is needed for conversions and for COMPONENT_REF.
>> Might as well play it safe and always test this. */
>> || TREE_CODE (TREE_TYPE (arg0)) == ERROR_MARK
>> || TREE_CODE (TREE_TYPE (arg1)) == ERROR_MARK
>> || TYPE_MODE (TREE_TYPE (arg0)) != TYPE_MODE (TREE_TYPE (arg1)))
>> #else
>> if ((TREE_CODE (arg0) != TREE_CODE (arg1)
>> && !(CONVERT_EXPR_P (arg0) && CONVERT_EXPR_P (arg1)))
>> /* This is needed for conversions and for COMPONENT_REF.
>> Might as well play it safe and always test this. */
>> || TREE_CODE (TREE_TYPE (arg0)) == ERROR_MARK
>> || TREE_CODE (TREE_TYPE (arg1)) == ERROR_MARK
>> || TYPE_MODE (TREE_TYPE (arg0)) != TYPE_MODE (TREE_TYPE (arg1)))
>> #endif
>> return 0;
>>
>>
>> The code in else part should be used instead, right?
>
> As this code is used by frontends who may actually have different behaviors
> for NOP_EXPR vs. CONVERT_EXPR (which is the only reason the two
> tree codes still exist!) it isn't 100% obvious. Though if it passes testing
> then it's ok IMHO, but I'd like you to refactor the above to
>
> if (TREE_CODE (arg0) != TREE_CODE (arg1)
> /* NOP_EXPR and CONVERT_EXPR are considered equal. */
> && !(CONVERT_EXPR_P (arg0) && CONVERT_EXPR_P (arg1)))
> return 0;
>
> /* This is needed for conversions and for COMPONENT_REF.
> Might as well play it safe and always test this. */
> if (TREE_CODE (TREE_TYPE (arg0)) == ERROR_MARK
> || TREE_CODE (TREE_TYPE (arg1)) == ERROR_MARK
> || TYPE_MODE (TREE_TYPE (arg0)) != TYPE_MODE (TREE_TYPE (arg1)))
> return 0;
>
> ok if that passes testing.
>
Will be done. Thanks.
--
Best Regards.