Re: [Qemu-ppc] [PATCH 6/6] target-ppc: add vclzlsbb/vctzlsbb instructions

[David Gibson]

On Wed, Sep 28, 2016 at 11:15:18AM +0530, Rajalakshmi Srinivasaraghavan wrote:
> The following vector instructions are added from ISA 3.0.
> 
> vclzlsbb - Vector Count Leading Zero Least-Significant Bits Byte
> vctzlsbb - Vector Count Trailing Zero Least-Significant Bits Byte
> 
> Signed-off-by: Rajalakshmi Srinivasaraghavan
> <address@hidden>

This will do for now, but I think you could do better than the loop.

Something like:
        tmp = vector & 0x0101010101010101;
        count = (clz(tmp) + 1) >> 3;

I think would do it (obviously more care would be needed with edge cases).

> ---
>  target-ppc/helper.h                 |    2 ++
>  target-ppc/int_helper.c             |   28 ++++++++++++++++++++++++++++
>  target-ppc/translate/vmx-impl.inc.c |   14 ++++++++++++++
>  target-ppc/translate/vmx-ops.inc.c  |    2 ++
>  4 files changed, 46 insertions(+), 0 deletions(-)
> 
> diff --git a/target-ppc/helper.h b/target-ppc/helper.h
> index 9c3095f..30c4429 100644
> --- a/target-ppc/helper.h
> +++ b/target-ppc/helper.h
> @@ -343,6 +343,8 @@ DEF_HELPER_2(vpopcntb, void, avr, avr)
>  DEF_HELPER_2(vpopcnth, void, avr, avr)
>  DEF_HELPER_2(vpopcntw, void, avr, avr)
>  DEF_HELPER_2(vpopcntd, void, avr, avr)
> +DEF_HELPER_1(vclzlsbb, tl, avr)
> +DEF_HELPER_1(vctzlsbb, tl, avr)
>  DEF_HELPER_3(vbpermd, void, avr, avr, avr)
>  DEF_HELPER_3(vbpermq, void, avr, avr, avr)
>  DEF_HELPER_2(vgbbd, void, avr, avr)
> diff --git a/target-ppc/int_helper.c b/target-ppc/int_helper.c
> index f132f7b..759ead9 100644
> --- a/target-ppc/int_helper.c
> +++ b/target-ppc/int_helper.c
> @@ -881,6 +881,34 @@ VCT(uxs, cvtsduw, u32)
>  VCT(sxs, cvtsdsw, s32)
>  #undef VCT
>  
> +target_ulong helper_vclzlsbb(ppc_avr_t *r)
> +{
> +    target_ulong count = 0;
> +    int i;
> +    VECTOR_FOR_INORDER_I(i, u8) {
> +        if (r->u8[i] & 0x01) {
> +            break;
> +        }
> +        count++;
> +    }
> +    return count;
> +}
> +target_ulong helper_vctzlsbb(ppc_avr_t *r)
> +{
> +    target_ulong count = 0;
> +    int i;
> +#if defined(HOST_WORDS_BIGENDIAN)
> +    for (i = ARRAY_SIZE(r->u8) - 1; i >= 0; i--) {
> +#else
> +    for (i = 0; i < ARRAY_SIZE(r->u8); i++) {
> +#endif
> +        if (r->u8[i] & 0x01) {
> +            break;
> +        }
> +        count++;
> +    }
> +    return count;
> +}
>  void helper_vmhaddshs(CPUPPCState *env, ppc_avr_t *r, ppc_avr_t *a,
>                        ppc_avr_t *b, ppc_avr_t *c)
>  {
> diff --git a/target-ppc/translate/vmx-impl.inc.c 
> b/target-ppc/translate/vmx-impl.inc.c
> index 0bc7188..1649b34 100644
> --- a/target-ppc/translate/vmx-impl.inc.c
> +++ b/target-ppc/translate/vmx-impl.inc.c
> @@ -691,6 +691,18 @@ static void glue(gen_, name)(DisasContext *ctx)          
>                \
>          tcg_temp_free_ptr(rd);                                          \
>      }
>  
> +#define GEN_VXFORM_NOA_3(name, opc2, opc3, opc4)                        \
> +static void glue(gen_, name)(DisasContext *ctx)                         \
> +    {                                                                   \
> +        TCGv_ptr rb;                                                    \
> +        if (unlikely(!ctx->altivec_enabled)) {                          \
> +            gen_exception(ctx, POWERPC_EXCP_VPU);                       \
> +            return;                                                     \
> +        }                                                               \
> +        rb = gen_avr_ptr(rB(ctx->opcode));                              \
> +        gen_helper_##name(cpu_gpr[rD(ctx->opcode)], rb);                \
> +        tcg_temp_free_ptr(rb);                                          \
> +    }
>  GEN_VXFORM_NOA(vupkhsb, 7, 8);
>  GEN_VXFORM_NOA(vupkhsh, 7, 9);
>  GEN_VXFORM_NOA(vupkhsw, 7, 25);
> @@ -905,6 +917,8 @@ GEN_VXFORM_NOA_2(vctzb, 1, 24, 28)
>  GEN_VXFORM_NOA_2(vctzh, 1, 24, 29)
>  GEN_VXFORM_NOA_2(vctzw, 1, 24, 30)
>  GEN_VXFORM_NOA_2(vctzd, 1, 24, 31)
> +GEN_VXFORM_NOA_3(vclzlsbb, 1, 24, 0)
> +GEN_VXFORM_NOA_3(vctzlsbb, 1, 24, 1)
>  GEN_VXFORM_NOA(vpopcntb, 1, 28)
>  GEN_VXFORM_NOA(vpopcnth, 1, 29)
>  GEN_VXFORM_NOA(vpopcntw, 1, 30)
> diff --git a/target-ppc/translate/vmx-ops.inc.c 
> b/target-ppc/translate/vmx-ops.inc.c
> index 009e9b1..20d243f 100644
> --- a/target-ppc/translate/vmx-ops.inc.c
> +++ b/target-ppc/translate/vmx-ops.inc.c
> @@ -223,6 +223,8 @@ GEN_VXFORM_300_EO(vctzb, 0x01, 0x18, 0x1C),
>  GEN_VXFORM_300_EO(vctzh, 0x01, 0x18, 0x1D),
>  GEN_VXFORM_300_EO(vctzw, 0x01, 0x18, 0x1E),
>  GEN_VXFORM_300_EO(vctzd, 0x01, 0x18, 0x1F),
> +GEN_VXFORM_300_EO(vclzlsbb, 0x01, 0x18, 0x0),
> +GEN_VXFORM_300_EO(vctzlsbb, 0x01, 0x18, 0x1),
>  GEN_VXFORM_300(vpermr, 0x1D, 0xFF),
>  
>  #define GEN_VXFORM_NOA(name, opc2, opc3)                                \

-- 
David Gibson                    | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au  | minimalist, thank you.  NOT _the_ _other_
                                | _way_ _around_!
http://www.ozlabs.org/~dgibson

signature.asc
Description: PGP signature