[v3,01/29] tcg-aarch64: Set ext based on TCG_OPF_64BIT

Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 tcg/aarch64/tcg-target.c | 28 +++++++---------------------
 1 file changed, 7 insertions(+), 21 deletions(-)

On 02.09.2013 19:54, Richard Henderson wrote:
> Signed-off-by: Richard Henderson <rth@twiddle.net>
> ---
>  tcg/aarch64/tcg-target.c | 28 +++++++---------------------
>  1 file changed, 7 insertions(+), 21 deletions(-)
> 
> diff --git a/tcg/aarch64/tcg-target.c b/tcg/aarch64/tcg-target.c
> index 55ff700..5b067fe 100644
> --- a/tcg/aarch64/tcg-target.c
> +++ b/tcg/aarch64/tcg-target.c
> @@ -1105,9 +1105,9 @@ static inline void tcg_out_load_pair(TCGContext *s, TCGReg addr,
>  static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>                         const TCGArg *args, const int *const_args)
>  {
> -    /* ext will be set in the switch below, which will fall through to the
> -       common code. It triggers the use of extended regs where appropriate. */
> -    int ext = 0;
> +    /* 99% of the time, we can signal the use of extension registers
> +       by looking to see if the opcode handles 64-bit data.  */
> +    bool ext = (tcg_op_defs[opc].flags & TCG_OPF_64BIT) != 0;
>  
>      switch (opc) {
>      case INDEX_op_exit_tb:
> @@ -1163,7 +1163,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_mov_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_mov_i32:
>          tcg_out_movr(s, ext, args[0], args[1]);
>          break;
> @@ -1176,43 +1175,36 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_add_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_add_i32:
>          tcg_out_arith(s, ARITH_ADD, ext, args[0], args[1], args[2], 0);
>          break;
>  
>      case INDEX_op_sub_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_sub_i32:
>          tcg_out_arith(s, ARITH_SUB, ext, args[0], args[1], args[2], 0);
>          break;
>  
>      case INDEX_op_and_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_and_i32:
>          tcg_out_arith(s, ARITH_AND, ext, args[0], args[1], args[2], 0);
>          break;
>  
>      case INDEX_op_or_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_or_i32:
>          tcg_out_arith(s, ARITH_OR, ext, args[0], args[1], args[2], 0);
>          break;
>  
>      case INDEX_op_xor_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_xor_i32:
>          tcg_out_arith(s, ARITH_XOR, ext, args[0], args[1], args[2], 0);
>          break;
>  
>      case INDEX_op_mul_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_mul_i32:
>          tcg_out_mul(s, ext, args[0], args[1], args[2]);
>          break;
>  
>      case INDEX_op_shl_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_shl_i32:
>          if (const_args[2]) {    /* LSL / UBFM Wd, Wn, (32 - m) */
>              tcg_out_shl(s, ext, args[0], args[1], args[2]);
> @@ -1222,7 +1214,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_shr_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_shr_i32:
>          if (const_args[2]) {    /* LSR / UBFM Wd, Wn, m, 31 */
>              tcg_out_shr(s, ext, args[0], args[1], args[2]);
> @@ -1232,7 +1223,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_sar_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_sar_i32:
>          if (const_args[2]) {    /* ASR / SBFM Wd, Wn, m, 31 */
>              tcg_out_sar(s, ext, args[0], args[1], args[2]);
> @@ -1242,7 +1232,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_rotr_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_rotr_i32:
>          if (const_args[2]) {    /* ROR / EXTR Wd, Wm, Wm, m */
>              tcg_out_rotr(s, ext, args[0], args[1], args[2]);
> @@ -1252,7 +1241,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_rotl_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_rotl_i32:     /* same as rotate right by (32 - m) */
>          if (const_args[2]) {    /* ROR / EXTR Wd, Wm, Wm, 32 - m */
>              tcg_out_rotl(s, ext, args[0], args[1], args[2]);
> @@ -1265,14 +1253,12 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_brcond_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_brcond_i32: /* CMP 0, 1, cond(2), label 3 */
>          tcg_out_cmp(s, ext, args[0], args[1], 0);
>          tcg_out_goto_label_cond(s, args[2], args[3]);
>          break;
>  
>      case INDEX_op_setcond_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_setcond_i32:
>          tcg_out_cmp(s, ext, args[1], args[2], 0);
>          tcg_out_cset(s, 0, args[0], args[3]);

There's not point to change to 'bool' if you pass '0': either we keep as int, and we pass 0,
or we change to bool, and we pass 'false'.
There are instances of this also in successive patches, I point out only this one, but it should be checked in the whole series.

> @@ -1315,9 +1301,11 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          tcg_out_qemu_st(s, args, 3);
>          break;
>  
> -    case INDEX_op_bswap64_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_bswap32_i64:
> +        /* Despite the _i64, this is a 32-bit bswap.  */
> +        ext = 0;
> +        /* FALLTHRU */
> +    case INDEX_op_bswap64_i64:

we waste too much y space here, which gives context and is a scarse resource.
What about 

case INDEX_op_bswap32_i64: /* Despite the _i64, this is a 32-bit bswap.  */
    ext = false; /* FALLTHRU */

>      case INDEX_op_bswap32_i32:
>          tcg_out_rev(s, ext, args[0], args[1]);
>          break;
> @@ -1327,12 +1315,10 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
>          break;
>  
>      case INDEX_op_ext8s_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_ext8s_i32:
>          tcg_out_sxt(s, ext, 0, args[0], args[1]);
>          break;
>      case INDEX_op_ext16s_i64:
> -        ext = 1; /* fall through */
>      case INDEX_op_ext16s_i32:
>          tcg_out_sxt(s, ext, 1, args[0], args[1]);
>          break;
> 

C.

On 12 September 2013 09:25, Claudio Fontana <claudio.fontana@huawei.com> wrote:
> On 02.09.2013 19:54, Richard Henderson wrote:
>>
>> -    case INDEX_op_bswap64_i64:
>> -        ext = 1; /* fall through */
>>      case INDEX_op_bswap32_i64:
>> +        /* Despite the _i64, this is a 32-bit bswap.  */
>> +        ext = 0;
>> +        /* FALLTHRU */
>> +    case INDEX_op_bswap64_i64:
>
> we waste too much y space here, which gives context and is a scarse resource.
> What about
>
> case INDEX_op_bswap32_i64: /* Despite the _i64, this is a 32-bit bswap.  */
>     ext = false; /* FALLTHRU */

Consensus in the rest of the code is for /* fall through */
rather than /* FALLTHRU */ -- there's only 28 of the
latter compared to 169 of the former.

-- PMM

On 12.09.2013 10:58, Peter Maydell wrote:
> On 12 September 2013 09:25, Claudio Fontana <claudio.fontana@huawei.com> wrote:
>> On 02.09.2013 19:54, Richard Henderson wrote:
>>>
>>> -    case INDEX_op_bswap64_i64:
>>> -        ext = 1; /* fall through */
>>>      case INDEX_op_bswap32_i64:
>>> +        /* Despite the _i64, this is a 32-bit bswap.  */
>>> +        ext = 0;
>>> +        /* FALLTHRU */
>>> +    case INDEX_op_bswap64_i64:
>>
>> we waste too much y space here, which gives context and is a scarse resource.
>> What about
>>
>> case INDEX_op_bswap32_i64: /* Despite the _i64, this is a 32-bit bswap.  */
>>     ext = false; /* FALLTHRU */
> 
> Consensus in the rest of the code is for /* fall through */
> rather than /* FALLTHRU */ -- there's only 28 of the
> latter compared to 169 of the former.
> 

I like /* fall through */ better as well.

On 09/12/2013 01:58 AM, Peter Maydell wrote:
> On 12 September 2013 09:25, Claudio Fontana <claudio.fontana@huawei.com> wrote:
>> On 02.09.2013 19:54, Richard Henderson wrote:
>>>
>>> -    case INDEX_op_bswap64_i64:
>>> -        ext = 1; /* fall through */
>>>      case INDEX_op_bswap32_i64:
>>> +        /* Despite the _i64, this is a 32-bit bswap.  */
>>> +        ext = 0;
>>> +        /* FALLTHRU */
>>> +    case INDEX_op_bswap64_i64:
>>
>> we waste too much y space here, which gives context and is a scarse resource.
>> What about
>>
>> case INDEX_op_bswap32_i64: /* Despite the _i64, this is a 32-bit bswap.  */
>>     ext = false; /* FALLTHRU */
> 
> Consensus in the rest of the code is for /* fall through */
> rather than /* FALLTHRU */ -- there's only 28 of the
> latter compared to 169 of the former.

Those 28 may well be all mine too.  The fingers still remember
that one must use FALLTHRU for lint.   ;-)


r~