Re: [Qemu-devel] [RFC] TCG new op: setcond

[Avi Kivity <avi@redhat.com>]

Laurent Desnogues wrote:
> this patch implements a new TCG op, setcond, that sets a temp
> to 1 if the condition is true, else to 0.  The benefit is the potential
> removal of brcond instructions, and helpers size reduction which
> can lead to using TCG instead of helpers.
>
> This patch is only posted here to get comments before I dig
> further into that and propose a proper update.
>
> One of the ARM helpers (sub with flag settings) has been
> converted to TCG and uses setcond to compute carry.
>
> setcond has been implemented only for x86_64 TCG back-end.
>
>   
> +// TODO should apply to setcond_i64 but not tested
> +static void tcg_out_setcond(TCGContext *s, int cond,
> +                            TCGArg ret, TCGArg arg1, TCGArg arg2, int rexw)
> +{
> +    /* clear ret since setcc only sets the lower 8 bits */
> +    tcg_out_modrm(s, 0x01 | (ARITH_XOR << 3) | rexw, ret, ret);
> +    /* cmp */
> +    tcg_out_modrm(s, 0x01 | (ARITH_CMP << 3) | rexw, arg2, arg1);
> +    /* setcc */
> +    // TODO this should use tcg_out_modrm
> +    //      however currently tcg_out_modrm outputs an extra byte for [abcd]l
> +    //tcg_out_modrm(s, (0x90 + tcg_cond_to_jcc[cond]) | P_EXT | P_REXB, ret, 0);
> +    if (ret > 3)
> +        tcg_out8(s, 0x40);
> +    else if (ret > 7)
> +        tcg_out8(s, 0x41);
>   

if ret == 8 this fails.  You need to reverse the order of the tests.

> +    tcg_out8(s, 0x0f);
> +    tcg_out8(s, 0x90 + tcg_cond_to_jcc[cond]);
> +    tcg_out8(s, 0xc0 + (ret & 7));
> +}
> +
>  
>   

-- 
error compiling committee.c: too many arguments to function