public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: "H.J. Lu" <hjl.tools@gmail.com>
To: Jan Beulich <jbeulich@suse.com>
Cc: Binutils <binutils@sourceware.org>
Subject: Re: [PATCH] x86: simplify expressions in update_imm()
Date: Mon, 31 Oct 2022 11:58:01 -0700	[thread overview]
Message-ID: <CAMe9rOp9K7iGmJYUgUT_rwmQpZ_+HjTHoiORZd3eGtZ5=XbsTQ@mail.gmail.com> (raw)
In-Reply-To: <c8d25b29-37fe-9623-a799-e67e1f991743@suse.com>

On Mon, Oct 31, 2022 at 10:01 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> Comparing the sum of the relevant .imm<N> fields against a constant imo
> makes more obvious what is actually meant. It allows dropping of two
> static variables, with a 3rd drop requiring two more minor adjustments
> elsewhere, utilizing that "i" is zeroed first thing in md_assemble().
> This also increases the chances of the compiler doing the calculations
> all in registers.
>
> --- a/gas/config/tc-i386.c
> +++ b/gas/config/tc-i386.c
> @@ -1908,12 +1908,9 @@ operand_type_xor (i386_operand_type x, i
>  static const i386_operand_type disp16_32 = OPERAND_TYPE_DISP16_32;
>  static const i386_operand_type anydisp = OPERAND_TYPE_ANYDISP;
>  static const i386_operand_type regxmm = OPERAND_TYPE_REGXMM;
> -static const i386_operand_type imm8 = OPERAND_TYPE_IMM8;
> -static const i386_operand_type imm8s = OPERAND_TYPE_IMM8S;
>  static const i386_operand_type imm16 = OPERAND_TYPE_IMM16;
>  static const i386_operand_type imm32 = OPERAND_TYPE_IMM32;
>  static const i386_operand_type imm32s = OPERAND_TYPE_IMM32S;
> -static const i386_operand_type imm64 = OPERAND_TYPE_IMM64;
>  static const i386_operand_type imm16_32 = OPERAND_TYPE_IMM16_32;
>  static const i386_operand_type imm16_32s = OPERAND_TYPE_IMM16_32S;
>  static const i386_operand_type imm16_32_32s = OPERAND_TYPE_IMM16_32_32S;
> @@ -3976,7 +3973,7 @@ process_immext (void)
>
>    exp = &im_expressions[i.imm_operands++];
>    i.op[i.operands].imms = exp;
> -  i.types[i.operands] = imm8;
> +  i.types[i.operands].bitfield.imm8 = 1;
>    i.operands++;
>    exp->X_op = O_constant;
>    exp->X_add_number = i.tm.extension_opcode;
> @@ -7652,18 +7649,12 @@ static int
>  update_imm (unsigned int j)
>  {
>    i386_operand_type overlap = i.types[j];
> -  if ((overlap.bitfield.imm8
> -       || overlap.bitfield.imm8s
> -       || overlap.bitfield.imm16
> -       || overlap.bitfield.imm32
> -       || overlap.bitfield.imm32s
> -       || overlap.bitfield.imm64)
> -      && !operand_type_equal (&overlap, &imm8)
> -      && !operand_type_equal (&overlap, &imm8s)
> -      && !operand_type_equal (&overlap, &imm16)
> -      && !operand_type_equal (&overlap, &imm32)
> -      && !operand_type_equal (&overlap, &imm32s)
> -      && !operand_type_equal (&overlap, &imm64))
> +  if (overlap.bitfield.imm8
> +      + overlap.bitfield.imm8s
> +      + overlap.bitfield.imm16
> +      + overlap.bitfield.imm32
> +      + overlap.bitfield.imm32s
> +      + overlap.bitfield.imm64 > 1)
>      {
>        if (i.suffix)
>         {
> @@ -7700,12 +7691,12 @@ update_imm (unsigned int j)
>        else if (i.prefix[DATA_PREFIX])
>         overlap = operand_type_and (overlap,
>                                     flag_code != CODE_16BIT ? imm16 : imm32);
> -      if (!operand_type_equal (&overlap, &imm8)
> -         && !operand_type_equal (&overlap, &imm8s)
> -         && !operand_type_equal (&overlap, &imm16)
> -         && !operand_type_equal (&overlap, &imm32)
> -         && !operand_type_equal (&overlap, &imm32s)
> -         && !operand_type_equal (&overlap, &imm64))
> +      if (overlap.bitfield.imm8
> +         + overlap.bitfield.imm8s
> +         + overlap.bitfield.imm16
> +         + overlap.bitfield.imm32
> +         + overlap.bitfield.imm32s
> +         + overlap.bitfield.imm64 != 1)
>         {
>           as_bad (_("no instruction mnemonic suffix given; "
>                     "can't determine immediate size"));
> @@ -8064,7 +8055,7 @@ build_modrm_byte (void)
>              immediate operand to encode the first operand.  */
>           exp = &im_expressions[i.imm_operands++];
>           i.op[i.operands].imms = exp;
> -         i.types[i.operands] = imm8;
> +         i.types[i.operands].bitfield.imm8 = 1;
>           i.operands++;
>
>           gas_assert (i.tm.operand_types[reg_slot].bitfield.class == RegSIMD);

OK.

Thanks.

-- 
H.J.

      reply	other threads:[~2022-10-31 18:58 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-31 17:00 Jan Beulich
2022-10-31 18:58 ` H.J. Lu [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAMe9rOp9K7iGmJYUgUT_rwmQpZ_+HjTHoiORZd3eGtZ5=XbsTQ@mail.gmail.com' \
    --to=hjl.tools@gmail.com \
    --cc=binutils@sourceware.org \
    --cc=jbeulich@suse.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).