Oh, v1sub and v2sub look incorrect, need use ^b instead of ^~b.
Thanks.
> From: gang.chen.5...@gmail.com
> To: peter.mayd...@linaro.org; r...@twiddle.net
> CC: qemu-devel@nongnu.org; xili_gchen_5...@hotmail.com;
> gang.chen.5...@gmail.com
> Subject: [PATCH v2] target-tilegx: Implement v*add and v*sub instructions
> Date: Tue, 22 Sep 2015 06:18:38 +0800
>
> From: Chen Gang <gang.chen.5...@gmail.com>
>
> v4* are implemented in normal code, another are implemented in helper
> functions.
>
> Signed-off-by: Chen Gang <gang.chen.5...@gmail.com>
> ---
> target-tilegx/helper.h | 5 +++++
> target-tilegx/simd_helper.c | 23 +++++++++++++++++++++++
> target-tilegx/translate.c | 46
> +++++++++++++++++++++++++++++++++++++++++++--
> 3 files changed, 72 insertions(+), 2 deletions(-)
>
> diff --git a/target-tilegx/helper.h b/target-tilegx/helper.h
> index b253722..6d98f3a 100644
> --- a/target-tilegx/helper.h
> +++ b/target-tilegx/helper.h
> @@ -11,3 +11,8 @@ DEF_HELPER_FLAGS_2(v1shrs, TCG_CALL_NO_RWG_SE, i64, i64,
> i64)
> DEF_HELPER_FLAGS_2(v2shl, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> DEF_HELPER_FLAGS_2(v2shru, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> DEF_HELPER_FLAGS_2(v2shrs, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> +
> +DEF_HELPER_FLAGS_2(v1add, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> +DEF_HELPER_FLAGS_2(v1sub, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> +DEF_HELPER_FLAGS_2(v2add, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> +DEF_HELPER_FLAGS_2(v2sub, TCG_CALL_NO_RWG_SE, i64, i64, i64)
> diff --git a/target-tilegx/simd_helper.c b/target-tilegx/simd_helper.c
> index c03e31a..00265fe 100644
> --- a/target-tilegx/simd_helper.c
> +++ b/target-tilegx/simd_helper.c
> @@ -22,6 +22,29 @@
> #include "qemu-common.h"
> #include "exec/helper-proto.h"
>
> +uint64_t helper_v1add(uint64_t a, uint64_t b)
> +{
> + return ((a & 0x7f7f7f7f7f7f7f7fULL) + (b & 0x7f7f7f7f7f7f7f7fULL))
> + ^ ((a ^ b) & 0x8080808080808080ULL);
> +}
> +
> +uint64_t helper_v1sub(uint64_t a, uint64_t b)
> +{
> + return ((a & 0x7f7f7f7f7f7f7f7fULL) - (b & 0x7f7f7f7f7f7f7f7fULL))
> + ^ ((a ^ ~b) & 0x8080808080808080ULL);
> +}
> +
> +uint64_t helper_v2add(uint64_t a, uint64_t b)
> +{
> + return ((a & 0x7fff7fff7fff7fffULL) + (b & 0x7fff7fff7fff7fffULL))
> + ^ ((a ^ b) & 0x8000800080008000ULL);
> +}
> +
> +uint64_t helper_v2sub(uint64_t a, uint64_t b)
> +{
> + return ((a & 0x7fff7fff7fff7fffULL) - (b & 0x7fff7fff7fff7fffULL))
> + ^ ((a ^ ~b) & 0x8000800080008000ULL);
> +}
>
> uint64_t helper_v1shl(uint64_t a, uint64_t b)
> {
> diff --git a/target-tilegx/translate.c b/target-tilegx/translate.c
> index 9228751..297de5c 100644
> --- a/target-tilegx/translate.c
> +++ b/target-tilegx/translate.c
> @@ -358,6 +358,26 @@ static void gen_v4sh(TCGv d64, TCGv a64, TCGv b64,
> tcg_temp_free_i32(bl);
> }
>
> +static void gen_v4op(TCGv d64, TCGv a64, TCGv b64,
> + void (*generate)(TCGv_i32, TCGv_i32, TCGv_i32))
> +{
> + TCGv_i32 al = tcg_temp_new_i32();
> + TCGv_i32 ah = tcg_temp_new_i32();
> + TCGv_i32 bl = tcg_temp_new_i32();
> + TCGv_i32 bh = tcg_temp_new_i32();
> +
> + tcg_gen_extr_i64_i32(al, ah, a64);
> + tcg_gen_extr_i64_i32(bl, bh, b64);
> + generate(al, al, bl);
> + generate(ah, ah, bh);
> + tcg_gen_concat_i32_i64(d64, al, ah);
> +
> + tcg_temp_free_i32(al);
> + tcg_temp_free_i32(ah);
> + tcg_temp_free_i32(bl);
> + tcg_temp_free_i32(bh);
> +}
> +
> static TileExcp gen_rr_opcode(DisasContext *dc, unsigned opext,
> unsigned dest, unsigned srca)
> {
> @@ -1043,8 +1063,12 @@ static TileExcp gen_rrr_opcode(DisasContext *dc,
> unsigned opext,
> break;
> case OE_RRR(V1ADDUC, 0, X0):
> case OE_RRR(V1ADDUC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V1ADD, 0, X0):
> case OE_RRR(V1ADD, 0, X1):
> + gen_helper_v1add(tdest, tsrca, tsrcb);
> + mnemonic = "v1add";
> + break;
> case OE_RRR(V1ADIFFU, 0, X0):
> case OE_RRR(V1AVGU, 0, X0):
> return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> @@ -1114,12 +1138,20 @@ static TileExcp gen_rrr_opcode(DisasContext *dc,
> unsigned opext,
> break;
> case OE_RRR(V1SUBUC, 0, X0):
> case OE_RRR(V1SUBUC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V1SUB, 0, X0):
> case OE_RRR(V1SUB, 0, X1):
> + gen_helper_v1sub(tdest, tsrca, tsrcb);
> + mnemonic = "v1sub";
> + break;
> case OE_RRR(V2ADDSC, 0, X0):
> case OE_RRR(V2ADDSC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V2ADD, 0, X0):
> case OE_RRR(V2ADD, 0, X1):
> + gen_helper_v2add(tdest, tsrca, tsrcb);
> + mnemonic = "v2add";
> + break;
> case OE_RRR(V2ADIFFS, 0, X0):
> case OE_RRR(V2AVGS, 0, X0):
> case OE_RRR(V2CMPEQ, 0, X0):
> @@ -1181,13 +1213,20 @@ static TileExcp gen_rrr_opcode(DisasContext *dc,
> unsigned opext,
> break;
> case OE_RRR(V2SUBSC, 0, X0):
> case OE_RRR(V2SUBSC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V2SUB, 0, X0):
> case OE_RRR(V2SUB, 0, X1):
> + gen_helper_v2sub(tdest, tsrca, tsrcb);
> + mnemonic = "v2sub";
> + break;
> case OE_RRR(V4ADDSC, 0, X0):
> case OE_RRR(V4ADDSC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V4ADD, 0, X0):
> case OE_RRR(V4ADD, 0, X1):
> - return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> + gen_v4op(tdest, tsrca, tsrcb, tcg_gen_add_i32);
> + mnemonic = "v4add";
> + break;
> case OE_RRR(V4INT_H, 0, X0):
> case OE_RRR(V4INT_H, 0, X1):
> tcg_gen_shri_tl(tdest, tsrcb, 32);
> @@ -1221,9 +1260,12 @@ static TileExcp gen_rrr_opcode(DisasContext *dc,
> unsigned opext,
> break;
> case OE_RRR(V4SUBSC, 0, X0):
> case OE_RRR(V4SUBSC, 0, X1):
> + return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> case OE_RRR(V4SUB, 0, X0):
> case OE_RRR(V4SUB, 0, X1):
> - return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
> + gen_v4op(tdest, tsrca, tsrcb, tcg_gen_sub_i32);
> + mnemonic = "v2sub";
> + break;
> case OE_RRR(XOR, 0, X0):
> case OE_RRR(XOR, 0, X1):
> case OE_RRR(XOR, 5, Y0):
> --
> 1.9.3
>
>
