diff mbox

[8/8] target-tricore: Add instructions of RR1 opcode format, that have 0xb3 as first opcode

Message ID 1418405504-11175-9-git-send-email-kbastian@mail.uni-paderborn.de
State New
Headers show

Commit Message

Bastian Koppelmann Dec. 12, 2014, 5:31 p.m. UTC
Add instructions of RR1 opcode format, that have 0xb3 as first opcode.
Add helper functions mulh, mulmh and mulrh in four variants, that compute multiplication,
with multiprecision (mulmh) or rounding (mulrh) of 4 halfwords, being either low or high parts
of two 32 bit regs.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
 target-tricore/helper.h    |  13 ++++++
 target-tricore/op_helper.c | 111 +++++++++++++++++++++++++++++++++++++++++++++
 target-tricore/translate.c |  96 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 220 insertions(+)

--
2.1.3

Comments

Richard Henderson Dec. 12, 2014, 8:53 p.m. UTC | #1
On 12/12/2014 09:31 AM, Bastian Koppelmann wrote:
> +#define MUL_H_OP(ARG00, ARG01, ARG10, ARG11)                                \
> +uint64_t helper_mulh_##ARG00 ##_##ARG01 ##_##ARG10 ##_##ARG11(              \
> +         CPUTriCoreState *env, target_ulong arg1, target_ulong arg2,        \
> +         target_ulong n)                                                    \
> +{                                                                           \
> +    uint64_t ret;                                                           \
> +    uint32_t result0, result1;                                              \
> +    int32_t arg00 = sextract32(arg1, ARG00, 16);                            \
> +    int32_t arg01 = sextract32(arg1, ARG01, 16);                            \
> +    int32_t arg10 = sextract32(arg2, ARG10, 16);                            \
> +    int32_t arg11 = sextract32(arg2, ARG11, 16);                            \

Since these functions only differ in the extraction of the arguments,
it'd probably be better to only have 3 helpers instead of 12 and do
this extraction in tcg as part of the call.



r~
diff mbox

Patch

diff --git a/target-tricore/helper.h b/target-tricore/helper.h
index 8d7958d..304c023 100644
--- a/target-tricore/helper.h
+++ b/target-tricore/helper.h
@@ -87,6 +87,19 @@  DEF_HELPER_3(dvinit_b_13, i64, env, i32, i32)
 DEF_HELPER_3(dvinit_b_131, i64, env, i32, i32)
 DEF_HELPER_3(dvinit_h_13, i64, env, i32, i32)
 DEF_HELPER_3(dvinit_h_131, i64, env, i32, i32)
+/* mulh */
+DEF_HELPER_4(mulh_U_L_L_L, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulh_U_L_L_U, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulh_U_L_U_L, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulh_L_U_U_U, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulmh_U_L_L_L, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulmh_U_L_L_U, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulmh_U_L_U_L, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulmh_L_U_U_U, i64, env, i32, i32, i32)
+DEF_HELPER_4(mulrh_U_L_L_L, i32, env, i32, i32, i32)
+DEF_HELPER_4(mulrh_U_L_L_U, i32, env, i32, i32, i32)
+DEF_HELPER_4(mulrh_U_L_U_L, i32, env, i32, i32, i32)
+DEF_HELPER_4(mulrh_L_U_U_U, i32, env, i32, i32, i32)
 /* CSA */
 DEF_HELPER_2(call, void, env, i32)
 DEF_HELPER_1(ret, void, env)
diff --git a/target-tricore/op_helper.c b/target-tricore/op_helper.c
index 0fd38b5..1cf9b8a 100644
--- a/target-tricore/op_helper.c
+++ b/target-tricore/op_helper.c
@@ -1254,6 +1254,117 @@  uint64_t helper_dvinit_h_131(CPUTriCoreState *env, uint32_t r1, uint32_t r2)
     return ret;
 }

+#define MUL_H_OP(ARG00, ARG01, ARG10, ARG11)                                \
+uint64_t helper_mulh_##ARG00 ##_##ARG01 ##_##ARG10 ##_##ARG11(              \
+         CPUTriCoreState *env, target_ulong arg1, target_ulong arg2,        \
+         target_ulong n)                                                    \
+{                                                                           \
+    uint64_t ret;                                                           \
+    uint32_t result0, result1;                                              \
+    int32_t arg00 = sextract32(arg1, ARG00, 16);                            \
+    int32_t arg01 = sextract32(arg1, ARG01, 16);                            \
+    int32_t arg10 = sextract32(arg2, ARG10, 16);                            \
+    int32_t arg11 = sextract32(arg2, ARG11, 16);                            \
+                                                                            \
+                                                                            \
+    int32_t sc1 = ((arg00 & 0xffff) == 0x8000) &&                           \
+                  ((arg10 & 0xffff) == 0x8000) && (n == 1);                 \
+    int32_t sc0 = ((arg01 & 0xffff) == 0x8000) &&                           \
+                  ((arg11 & 0xffff) == 0x8000) && (n == 1);                 \
+    if (sc1) {                                                              \
+        result1 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result1 = (((uint32_t)(arg00 * arg10)) << n);                       \
+    }                                                                       \
+    if (sc0) {                                                              \
+        result0 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result0 = (((uint32_t)(arg01 * arg11)) << n);                       \
+    }                                                                       \
+    /* advanced overflow */                                                 \
+    env->PSW_USB_AV = (result1 ^ result1 * 2u) | (result0 ^ result0 * 2u);  \
+    env->PSW_USB_SAV |= env->PSW_USB_AV;                                    \
+    /* clean v bit */                                                       \
+    env->PSW_USB_V = 0;                                                     \
+    ret = (((uint64_t)result1 << 32)) | result0;                            \
+    return ret;                                                             \
+}                                                                           \
+uint64_t helper_mulmh_##ARG00 ##_##ARG01 ##_##ARG10 ##_##ARG11(             \
+         CPUTriCoreState *env, target_ulong arg1, target_ulong arg2,        \
+         target_ulong n)                                                    \
+{                                                                           \
+    uint64_t ret;                                                           \
+    int64_t result0, result1;                                               \
+    int32_t arg00 = sextract32(arg1, ARG00, 16);                            \
+    int32_t arg01 = sextract32(arg1, ARG01, 16);                            \
+    int32_t arg10 = sextract32(arg2, ARG10, 16);                            \
+    int32_t arg11 = sextract32(arg2, ARG11, 16);                            \
+                                                                            \
+    int32_t sc1 = ((arg00 & 0xffff) == 0x8000) &&                           \
+                  ((arg10 & 0xffff) == 0x8000) && (n == 1);                 \
+    int32_t sc0 = ((arg01 & 0xffff) == 0x8000) &&                           \
+                  ((arg11 & 0xffff) == 0x8000) && (n == 1);                 \
+                                                                            \
+    if (sc1) {                                                              \
+        result1 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result1 = ((arg00 * arg10) << n);                                   \
+    }                                                                       \
+    if (sc0) {                                                              \
+        result0 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result0 = ((arg01 * arg11) << n);                                   \
+    }                                                                       \
+    /* clear advanced overflow */                                           \
+    env->PSW_USB_AV = 0;                                                    \
+    /* clean v bit */                                                       \
+    env->PSW_USB_V = 0;                                                     \
+    ret = (result1 + result0);                                              \
+    ret = ret << 16;                                                        \
+    return ret;                                                             \
+}                                                                           \
+uint32_t helper_mulrh_##ARG00 ##_##ARG01 ##_##ARG10 ##_##ARG11(             \
+         CPUTriCoreState *env, target_ulong arg1, target_ulong arg2,        \
+         target_ulong n)                                                    \
+{                                                                           \
+    uint32_t result0, result1;                                              \
+    int32_t arg00 = sextract32(arg1, ARG00, 16);                            \
+    int32_t arg01 = sextract32(arg1, ARG01, 16);                            \
+    int32_t arg10 = sextract32(arg2, ARG10, 16);                            \
+    int32_t arg11 = sextract32(arg2, ARG11, 16);                            \
+                                                                            \
+    int32_t sc1 = ((arg00 & 0xffff) == 0x8000) &&                           \
+                  ((arg10 & 0xffff) == 0x8000) && (n == 1);                 \
+    int32_t sc0 = ((arg01 & 0xffff) == 0x8000) &&                           \
+                  ((arg11 & 0xffff) == 0x8000) && (n == 1);                 \
+                                                                            \
+    if (sc1) {                                                              \
+        result1 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result1 = ((arg00 * arg10) << n) + 0x8000;                          \
+    }                                                                       \
+    if (sc0) {                                                              \
+        result0 = 0x7fffffff;                                               \
+    } else {                                                                \
+        result0 = ((arg01 * arg11) << n) + 0x8000;                          \
+    }                                                                       \
+    /* advanced overflow */                                                 \
+    env->PSW_USB_AV = (result1 ^ result1 * 2u) | (result0 ^ result0 * 2u);  \
+    env->PSW_USB_SAV |= env->PSW_USB_AV;                                    \
+    /* clean v bit */                                                       \
+    env->PSW_USB_V = 0;                                                     \
+    return (result1 & 0xffff0000) | (result0 >> 16);                        \
+}
+
+#define U 16
+#define L 0
+MUL_H_OP(U, L, L, L)
+MUL_H_OP(U, L, L, U)
+MUL_H_OP(U, L, U, L)
+MUL_H_OP(L, U, U, U)
+#undef U
+#undef L
+
 /* context save area (CSA) related helpers */

 static int cdc_increment(target_ulong *psw)
diff --git a/target-tricore/translate.c b/target-tricore/translate.c
index aacf097..36a85d6 100644
--- a/target-tricore/translate.c
+++ b/target-tricore/translate.c
@@ -4573,6 +4573,98 @@  static void decode_rr_divide(CPUTriCoreState *env, DisasContext *ctx)
     }
 }

+/* RR1 Format */
+static void decode_rr1_mul(CPUTriCoreState *env, DisasContext *ctx)
+{
+    uint32_t op2;
+
+    int r1, r2, r3;
+    TCGv n;
+    TCGv_i64 temp64;
+
+    r1 = MASK_OP_RR1_S1(ctx->opcode);
+    r2 = MASK_OP_RR1_S2(ctx->opcode);
+    r3 = MASK_OP_RR1_D(ctx->opcode);
+    n  = tcg_const_i32(MASK_OP_RR1_N(ctx->opcode));
+    op2 = MASK_OP_RR1_OP2(ctx->opcode);
+
+    switch (op2) {
+    case OPC2_32_RR1_MUL_H_32_LL:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulh_U_L_L_L(temp64, cpu_env, cpu_gpr_d[r1],
+                                cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MUL_H_32_LU:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulh_U_L_L_U(temp64, cpu_env, cpu_gpr_d[r1],
+                                cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MUL_H_32_UL:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulh_U_L_U_L(temp64, cpu_env, cpu_gpr_d[r1],
+                                cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MUL_H_32_UU:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulh_L_U_U_U(temp64, cpu_env, cpu_gpr_d[r1],
+                                cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MULM_H_64_LL:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulmh_U_L_L_L(temp64, cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MULM_H_64_LU:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulmh_U_L_L_U(temp64, cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MULM_H_64_UL:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulmh_U_L_U_L(temp64, cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MULM_H_64_UU:
+        temp64 = tcg_temp_new_i64();
+        gen_helper_mulmh_L_U_U_U(temp64, cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        tcg_gen_extr_i64_i32(cpu_gpr_d[r3], cpu_gpr_d[r3+1], temp64);
+        tcg_temp_free_i64(temp64);
+        break;
+    case OPC2_32_RR1_MULR_H_16_LL:
+        gen_helper_mulrh_U_L_L_L(cpu_gpr_d[r3], cpu_env, cpu_gpr_d[r1],
+                                     cpu_gpr_d[r2], n);
+        break;
+    case OPC2_32_RR1_MULR_H_16_LU:
+        gen_helper_mulrh_U_L_L_U(cpu_gpr_d[r3], cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        break;
+    case OPC2_32_RR1_MULR_H_16_UL:
+        gen_helper_mulrh_U_L_U_L(cpu_gpr_d[r3], cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        break;
+    case OPC2_32_RR1_MULR_H_16_UU:
+        gen_helper_mulrh_L_U_U_U(cpu_gpr_d[r3], cpu_env, cpu_gpr_d[r1],
+                                 cpu_gpr_d[r2], n);
+        break;
+    }
+    tcg_temp_free(n);
+}
+
 static void decode_32Bit_opc(CPUTriCoreState *env, DisasContext *ctx)
 {
     int op1;
@@ -4825,6 +4917,10 @@  static void decode_32Bit_opc(CPUTriCoreState *env, DisasContext *ctx)
     case OPCM_32_RR_DIVIDE:
         decode_rr_divide(env, ctx);
         break;
+/* RR1 Format */
+    case OPCM_32_RR1_MUL:
+        decode_rr1_mul(env, ctx);
+        break;
     }
 }