diff mbox series

[RFC,v2,10/15] target/riscv: rvb: rotate (left/right)

Message ID 20201216020150.3157-11-frank.chang@sifive.com
State New
Headers show
Series support subsets of bitmanip extension | expand

Commit Message

Frank Chang Dec. 16, 2020, 2:01 a.m. UTC 
From: Kito Cheng <kito.cheng@sifive.com>

Signed-off-by: Kito Cheng <kito.cheng@sifive.com>
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
 target/riscv/insn32-64.decode           |  3 ++
 target/riscv/insn32.decode              |  3 ++
 target/riscv/insn_trans/trans_rvb.c.inc | 61 +++++++++++++++++++++++++
 target/riscv/translate.c                | 36 +++++++++++++++
 4 files changed, 103 insertions(+)

Comments

Richard Henderson Dec. 16, 2020, 4:39 p.m. UTC | #1 
On 12/15/20 8:01 PM, frank.chang@sifive.com wrote:
> +static bool trans_ror(DisasContext *ctx, arg_ror *a)
>  {
> +    REQUIRE_EXT(ctx, RVB);
> +    return gen_arith(ctx, a, &tcg_gen_rotr_tl);
> +}

Use gen_shift.

> +static bool trans_rori(DisasContext *ctx, arg_rori *a)
> +{
> +    REQUIRE_EXT(ctx, RVB);
> +
> +    if (a->shamt >= TARGET_LONG_BITS) {
> +        return false;
> +    }
> +
> +    TCGv source1 = tcg_temp_new();
> +
> +    gen_get_gpr(source1, a->rs1);
> +    tcg_gen_rotri_tl(source1, source1, a->shamt);
> +    gen_set_gpr(a->rd, source1);

Use gen_shifti.

> +static bool trans_rol(DisasContext *ctx, arg_rol *a)
> +{
> +    REQUIRE_EXT(ctx, RVB);
> +    return gen_arith(ctx, a, &tcg_gen_rotl_tl);
> +}

Use gen_shift.

> +static bool trans_rorw(DisasContext *ctx, arg_rorw *a)
> +{
> +    REQUIRE_EXT(ctx, RVB);
> +    return gen_shiftw(ctx, a, &gen_rorw);
> +}
> +
> +static bool trans_roriw(DisasContext *ctx, arg_roriw *a)
> +{
> +    REQUIRE_EXT(ctx, RVB);
> +
> +    if (a->shamt >= 32) {
> +        return false;
> +    }

Test is impossible due to @sh5.

> +    if (a->shamt == 0) {
> +        TCGv t = tcg_temp_new();
> +        gen_get_gpr(t, a->rs1);
> +        tcg_gen_ext32s_tl(t, t);
> +        gen_set_gpr(a->rd, t);
> +        tcg_temp_free(t);
> +        return true;
> +    }

Why do you need this special case?  The general expansion would appear to work
fine, and surely this needs no special optimization.


r~
diff mbox series

Patch

diff --git a/target/riscv/insn32-64.decode b/target/riscv/insn32-64.decode
index 6d017c70c74..ac0634d754a 100644
--- a/target/riscv/insn32-64.decode
+++ b/target/riscv/insn32-64.decode
@@ -100,9 +100,12 @@  sbinvw     0110100 .......... 001 ..... 0111011 @r
 sbextw     0100100 .......... 101 ..... 0111011 @r
 slow       0010000 .......... 001 ..... 0111011 @r
 srow       0010000 .......... 101 ..... 0111011 @r
+rorw       0110000 .......... 101 ..... 0111011 @r
+rolw       0110000 .......... 001 ..... 0111011 @r
 
 sbsetiw    0010100 .......... 001 ..... 0011011 @sh5
 sbclriw    0100100 .......... 001 ..... 0011011 @sh5
 sbinviw    0110100 .......... 001 ..... 0011011 @sh5
 sloiw      0010000 .......... 001 ..... 0011011 @sh5
 sroiw      0010000 .......... 101 ..... 0011011 @sh5
+roriw      0110000 .......... 101 ..... 0011011 @sh5
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 78ce4b11097..b4677293582 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -617,6 +617,8 @@  sbinv      0110100 .......... 001 ..... 0110011 @r
 sbext      0100100 .......... 101 ..... 0110011 @r
 slo        0010000 .......... 001 ..... 0110011 @r
 sro        0010000 .......... 101 ..... 0110011 @r
+ror        0110000 .......... 101 ..... 0110011 @r
+rol        0110000 .......... 001 ..... 0110011 @r
 
 sbseti     00101. ........... 001 ..... 0010011 @sh
 sbclri     01001. ........... 001 ..... 0010011 @sh
@@ -624,3 +626,4 @@  sbinvi     01101. ........... 001 ..... 0010011 @sh
 sbexti     01001. ........... 101 ..... 0010011 @sh
 sloi       00100. ........... 001 ..... 0010011 @sh
 sroi       00100. ........... 101 ..... 0010011 @sh
+rori       01100. ........... 101 ..... 0010011 @sh
diff --git a/target/riscv/insn_trans/trans_rvb.c.inc b/target/riscv/insn_trans/trans_rvb.c.inc
index 11b5439e703..433cc7f9fc8 100644
--- a/target/riscv/insn_trans/trans_rvb.c.inc
+++ b/target/riscv/insn_trans/trans_rvb.c.inc
@@ -189,7 +189,36 @@  static bool trans_sroi(DisasContext *ctx, arg_sroi *a)
     return gen_shifti(ctx, a, &gen_sro);
 }
 
+static bool trans_ror(DisasContext *ctx, arg_ror *a)
 {
+    REQUIRE_EXT(ctx, RVB);
+    return gen_arith(ctx, a, &tcg_gen_rotr_tl);
+}
+
+static bool trans_rori(DisasContext *ctx, arg_rori *a)
+{
+    REQUIRE_EXT(ctx, RVB);
+
+    if (a->shamt >= TARGET_LONG_BITS) {
+        return false;
+    }
+
+    TCGv source1 = tcg_temp_new();
+
+    gen_get_gpr(source1, a->rs1);
+    tcg_gen_rotri_tl(source1, source1, a->shamt);
+    gen_set_gpr(a->rd, source1);
+
+    tcg_temp_free(source1);
+    return true;
+}
+
+static bool trans_rol(DisasContext *ctx, arg_rol *a)
+{
+    REQUIRE_EXT(ctx, RVB);
+    return gen_arith(ctx, a, &tcg_gen_rotl_tl);
+}
+
 /* RV64-only instructions */
 #ifdef TARGET_RISCV64
 
@@ -289,4 +318,36 @@  static bool trans_sroiw(DisasContext *ctx, arg_sroiw *a)
     return gen_shiftiw(ctx, a, &gen_sro);
 }
 
+static bool trans_rorw(DisasContext *ctx, arg_rorw *a)
+{
+    REQUIRE_EXT(ctx, RVB);
+    return gen_shiftw(ctx, a, &gen_rorw);
+}
+
+static bool trans_roriw(DisasContext *ctx, arg_roriw *a)
+{
+    REQUIRE_EXT(ctx, RVB);
+
+    if (a->shamt >= 32) {
+        return false;
+    }
+
+    if (a->shamt == 0) {
+        TCGv t = tcg_temp_new();
+        gen_get_gpr(t, a->rs1);
+        tcg_gen_ext32s_tl(t, t);
+        gen_set_gpr(a->rd, t);
+        tcg_temp_free(t);
+        return true;
+    }
+
+    return gen_shiftiw(ctx, a, &gen_rorw);
+}
+
+static bool trans_rolw(DisasContext *ctx, arg_rolw *a)
+{
+    REQUIRE_EXT(ctx, RVB);
+    return gen_shiftw(ctx, a, &gen_rolw);
+}
+
 #endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0c00d20ab1b..10b4142a3ab 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -849,6 +849,42 @@  static void gen_packuw(TCGv ret, TCGv arg1, TCGv arg2)
     tcg_temp_free(t);
 }
 
+static void gen_rorw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+    TCGv_i32 t1 = tcg_temp_new_i32();
+    TCGv_i32 t2 = tcg_temp_new_i32();
+
+    /* truncate to 32-bits */
+    tcg_gen_trunc_tl_i32(t1, arg1);
+    tcg_gen_trunc_tl_i32(t2, arg2);
+
+    tcg_gen_rotr_i32(t1, t1, t2);
+
+    /* sign-extend 64-bits */
+    tcg_gen_ext_i32_tl(ret, t1);
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t2);
+}
+
+static void gen_rolw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+    TCGv_i32 t1 = tcg_temp_new_i32();
+    TCGv_i32 t2 = tcg_temp_new_i32();
+
+    /* truncate to 32-bits */
+    tcg_gen_trunc_tl_i32(t1, arg1);
+    tcg_gen_trunc_tl_i32(t2, arg2);
+
+    tcg_gen_rotl_i32(t1, t1, t2);
+
+    /* sign-extend 64-bits */
+    tcg_gen_ext_i32_tl(ret, t1);
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t2);
+}
+
 #endif
 
 static bool gen_arith(DisasContext *ctx, arg_r *a,