[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-ppc] [PATCH 02/14] target-ppc: Add xxinsertw instruction
From: |
David Gibson |
Subject: |
Re: [Qemu-ppc] [PATCH 02/14] target-ppc: Add xxinsertw instruction |
Date: |
Fri, 6 Jan 2017 08:57:04 +1100 |
User-agent: |
Mutt/1.7.1 (2016-10-04) |
On Thu, Jan 05, 2017 at 04:56:07PM +0530, Nikunj A Dadhania wrote:
> xxinsertw: VSX Vector Insert Word
I think this still has problems with out of bounds values.
>
> Signed-off-by: Nikunj A Dadhania <address@hidden>
> ---
> target-ppc/helper.h | 1 +
> target-ppc/int_helper.c | 25 +++++++++++++++++++++++++
> target-ppc/translate/vsx-impl.inc.c | 5 +++--
> target-ppc/translate/vsx-ops.inc.c | 1 +
> 4 files changed, 30 insertions(+), 2 deletions(-)
>
> diff --git a/target-ppc/helper.h b/target-ppc/helper.h
> index 8b30420..6c5b194 100644
> --- a/target-ppc/helper.h
> +++ b/target-ppc/helper.h
> @@ -541,6 +541,7 @@ DEF_HELPER_2(xvrspiz, void, env, i32)
> DEF_HELPER_2(xxperm, void, env, i32)
> DEF_HELPER_2(xxpermr, void, env, i32)
> DEF_HELPER_4(xxextractuw, void, env, tl, tl, i32)
> +DEF_HELPER_4(xxinsertw, void, env, tl, tl, i32)
>
> DEF_HELPER_2(efscfsi, i32, env, i32)
> DEF_HELPER_2(efscfui, i32, env, i32)
> diff --git a/target-ppc/int_helper.c b/target-ppc/int_helper.c
> index 63ba0e3..24e5964 100644
> --- a/target-ppc/int_helper.c
> +++ b/target-ppc/int_helper.c
> @@ -2059,6 +2059,31 @@ void helper_xxextractuw(CPUPPCState *env, target_ulong
> xtn,
> putVSR(xtn, &xt, env);
> }
>
> +void helper_xxinsertw(CPUPPCState *env, target_ulong xtn,
> + target_ulong xbn, uint32_t index)
> +{
> + ppc_vsr_t xt, xb;
> + size_t es = sizeof(uint32_t);
> + int ins_index, i = 0;
> +
> + getVSR(xbn, &xb, env);
> + getVSR(xtn, &xt, env);
> +
> +#if defined(HOST_WORDS_BIGENDIAN)
> + ins_index = index;
> + for (i = 0; i < es && ins_index < 16; i++, ins_index++) {
> + xt.u8[ins_index] = xb.u8[8 - es + i];
Unlike extract, you don't mod the ins_index here. Which IIUC, means
for UIMM > 12 you will overwrite data beyond xt, which sounds like a
very bad thing.
> + }
> +#else
> + ins_index = 15 - index;
> + for (i = es - 1; i >= 0 && ins_index >= 0; i--, ins_index--) {
> + xt.u8[ins_index] = xb.u8[8 + i];
> + }
> +#endif
> +
> + putVSR(xtn, &xt, env);
> +}
> +
> #define VEXT_SIGNED(name, element, mask, cast, recast) \
> void helper_##name(ppc_avr_t *r, ppc_avr_t *b) \
> { \
> diff --git a/target-ppc/translate/vsx-impl.inc.c
> b/target-ppc/translate/vsx-impl.inc.c
> index 7977f24..c9ba0f5 100644
> --- a/target-ppc/translate/vsx-impl.inc.c
> +++ b/target-ppc/translate/vsx-impl.inc.c
> @@ -1180,7 +1180,7 @@ static void gen_xxsldwi(DisasContext *ctx)
> tcg_temp_free_i64(xtl);
> }
>
> -#define VSX_EXTRACT(name) \
> +#define VSX_EXTRACT_INSERT(name) \
Plus for UIMM > 15, this macro will set the target to 0. That seems
right for extract, but doesn't really seem right for insert (although
*maybe* it's what the hardware does).
> static void gen_##name(DisasContext *ctx) \
> { \
> TCGv xt, xb; \
> @@ -1208,7 +1208,8 @@ static void gen_##name(DisasContext *ctx)
> \
> tcg_temp_free_i32(t0); \
> }
>
> -VSX_EXTRACT(xxextractuw)
> +VSX_EXTRACT_INSERT(xxextractuw)
> +VSX_EXTRACT_INSERT(xxinsertw)
>
> #undef GEN_XX2FORM
> #undef GEN_XX3FORM
> diff --git a/target-ppc/translate/vsx-ops.inc.c
> b/target-ppc/translate/vsx-ops.inc.c
> index 473d925..096d358 100644
> --- a/target-ppc/translate/vsx-ops.inc.c
> +++ b/target-ppc/translate/vsx-ops.inc.c
> @@ -285,6 +285,7 @@ GEN_XX2FORM(xxspltw, 0x08, 0x0A, PPC2_VSX),
> GEN_XX1FORM(xxspltib, 0x08, 0x0B, PPC2_ISA300),
> GEN_XX3FORM_DM(xxsldwi, 0x08, 0x00),
> GEN_XX2FORM_EXT(xxextractuw, 0x0A, 0x0A, PPC2_ISA300),
> +GEN_XX2FORM_EXT(xxinsertw, 0x0A, 0x0B, PPC2_ISA300),
>
> #define GEN_XXSEL_ROW(opc3) \
> GEN_HANDLER2_E(xxsel, "xxsel", 0x3C, 0x18, opc3, 0, PPC_NONE, PPC2_VSX), \
--
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
- [Qemu-ppc] [PATCH 00/14] POWER9 TCG enablements - part10, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 02/14] target-ppc: Add xxinsertw instruction, Nikunj A Dadhania, 2017/01/05
- Re: [Qemu-ppc] [PATCH 02/14] target-ppc: Add xxinsertw instruction,
David Gibson <=
- [Qemu-ppc] [PATCH 06/14] target-ppc: Add xsaddqp instructions, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 01/14] target-ppc: Add xxextractuw instruction, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 09/14] target-ppc: Add xscvdpqp instruction, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 08/14] target-ppc: Use correct precision for FPRF setting, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 04/14] target-ppc: Replace isden by float64_is_zero_or_denormal, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 07/14] target-ppc: Add xscvdphp, xscvhpdp, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 10/14] target-ppc: Add xscvqpdp instruction, Nikunj A Dadhania, 2017/01/05
- [Qemu-ppc] [PATCH 03/14] target-ppc: Use float64 arg in helper_compute_fprf(), Nikunj A Dadhania, 2017/01/05