qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Laurent Desnogues" <laurent.desnogues@gmail.com>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] another SPARC issue
Date: Tue, 5 Aug 2008 22:36:43 +0200	[thread overview]
Message-ID: <761ea48b0808051336o120c3e9j286d9523a693f98f@mail.gmail.com> (raw)
In-Reply-To: <20080805161615.S57426@stanley.csl.cornell.edu>

On Tue, Aug 5, 2008 at 10:20 PM, Vince Weaver <vince@csl.cornell.edu> wrote:
>
> The instruction is
>   ldda  [ %o1 + 0x40 ] %asi, %f16
>
> but under Qemu for some reason 0x80 is added to %o1 (instead of 0x40).
>
> Indeed if you look at the generated TCG (see below), the 0x40 is being added
> in twice.  I was digging around in translate.c but can't seem to see why
> this is happning.
>
> Any help would be appreciated...

Since I am not afraid to be ridiculous again, here is my take:

- in disas_sparc_insn, line 4199:
            } else if (IS_IMM) {     /* immediate */
                rs2 = GET_FIELDs(insn, 19, 31);
                tcg_gen_addi_tl(cpu_addr, cpu_src1, (int)rs2);

- line 4307:
                case 0x13:      /* load double word alternate */
                    save_state(dc, cpu_cond);
                    gen_ldda_asi(cpu_val, cpu_addr, insn, rd);

static inline void gen_ldda_asi(TCGv hi, TCGv addr, int insn, int rd)
{
    TCGv r_asi, r_rd;

    r_asi = gen_get_asi(insn, addr);
    r_rd = tcg_const_i32(rd);
    tcg_gen_helper_0_3(helper_ldda_asi, addr, r_asi, r_rd);
    tcg_temp_free(r_rd);
    tcg_temp_free(r_asi);
}

static inline TCGv gen_get_asi(int insn, TCGv r_addr)
{
    int asi, offset;
    TCGv r_asi;

    if (IS_IMM) {
        r_asi = tcg_temp_new(TCG_TYPE_I32);
        offset = GET_FIELD(insn, 25, 31);
        tcg_gen_addi_tl(r_addr, r_addr, offset);
        tcg_gen_ld_i32(r_asi, cpu_env, offsetof(CPUSPARCState, asi));
    } else {
        asi = GET_FIELD(insn, 19, 26);
        r_asi = tcg_const_i32(asi);
    }
    return r_asi;
}

I am not sure I followed the right path, but it seems that indeed
for asi ld/st the constant gets added twice:  once in the disas
function and once in gen_get_asi.


Laurent

  reply	other threads:[~2008-08-05 20:36 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-29  3:32 [Qemu-devel] x86 tcg problem Vince Weaver
2008-07-29  8:46 ` Laurent Desnogues
2008-07-29 13:43   ` Vince Weaver
2008-07-29 18:24   ` malc
2008-07-29 18:33     ` Laurent Desnogues
2008-07-29 17:18 ` Blue Swirl
2008-08-05 20:20   ` [Qemu-devel] another SPARC issue Vince Weaver
2008-08-05 20:36     ` Laurent Desnogues [this message]
2008-08-06  2:33       ` Vince Weaver
2008-08-06  8:17         ` Laurent Desnogues
2008-08-06 15:29         ` Blue Swirl
2008-08-06 19:31           ` [Qemu-devel] one more " Vince Weaver
2008-08-06 19:45             ` Julian Seward
2008-08-06 19:55             ` Blue Swirl
2008-07-29 17:51 ` [Qemu-devel] x86 tcg problem Blue Swirl

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=761ea48b0808051336o120c3e9j286d9523a693f98f@mail.gmail.com \
    --to=laurent.desnogues@gmail.com \
    --cc=qemu-devel@nongnu.org \
    /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 NNTP newsgroup(s).