[PATCH v4 11/22] target/xtensa: implement FPU division and square root

[Max Filippov <jcmvbkbc@gmail.com>]

This does not implement all opcodes related to div/sqrt as specified in
the xtensa ISA, partly because the official specification is not
complete and partly because precise implementation is unnecessarily
complex. Instead instructions specific to the div/sqrt sequences are
implemented differently, most of them as nops, but the results of
div/sqrt sequences is preserved.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
---
 target/xtensa/fpu_helper.c |  24 +++++++++
 target/xtensa/helper.h     |   4 ++
 target/xtensa/translate.c  | 104 +++++++++++++++++++++++++++++++++++++
 3 files changed, 132 insertions(+)

diff --git a/target/xtensa/fpu_helper.c b/target/xtensa/fpu_helper.c
index b5faf34ad080..ba3c29d19d91 100644
--- a/target/xtensa/fpu_helper.c
+++ b/target/xtensa/fpu_helper.c
@@ -231,6 +231,30 @@ float32 HELPER(msub_s)(CPUXtensaState *env, float32 a, float32 b, float32 c)
                           &env->fp_status);
 }
 
+float64 HELPER(mkdadj_d)(CPUXtensaState *env, float64 a, float64 b)
+{
+    set_use_first_nan(true, &env->fp_status);
+    return float64_div(b, a, &env->fp_status);
+}
+
+float32 HELPER(mkdadj_s)(CPUXtensaState *env, float32 a, float32 b)
+{
+    set_use_first_nan(env->config->use_first_nan, &env->fp_status);
+    return float32_div(b, a, &env->fp_status);
+}
+
+float64 HELPER(mksadj_d)(CPUXtensaState *env, float64 v)
+{
+    set_use_first_nan(true, &env->fp_status);
+    return float64_sqrt(v, &env->fp_status);
+}
+
+float32 HELPER(mksadj_s)(CPUXtensaState *env, float32 v)
+{
+    set_use_first_nan(env->config->use_first_nan, &env->fp_status);
+    return float32_sqrt(v, &env->fp_status);
+}
+
 uint32_t HELPER(ftoi_d)(CPUXtensaState *env, float64 v,
                         uint32_t rounding_mode, uint32_t scale)
 {
diff --git a/target/xtensa/helper.h b/target/xtensa/helper.h
index 095f754671ce..ae938ceedb80 100644
--- a/target/xtensa/helper.h
+++ b/target/xtensa/helper.h
@@ -83,6 +83,10 @@ DEF_HELPER_4(madd_d, f64, env, f64, f64, f64)
 DEF_HELPER_4(madd_s, f32, env, f32, f32, f32)
 DEF_HELPER_4(msub_d, f64, env, f64, f64, f64)
 DEF_HELPER_4(msub_s, f32, env, f32, f32, f32)
+DEF_HELPER_3(mkdadj_d, f64, env, f64, f64)
+DEF_HELPER_3(mkdadj_s, f32, env, f32, f32)
+DEF_HELPER_2(mksadj_d, f64, env, f64)
+DEF_HELPER_2(mksadj_s, f32, env, f32)
 DEF_HELPER_4(ftoi_d, i32, env, f64, i32, i32)
 DEF_HELPER_4(ftoui_d, i32, env, f64, i32, i32)
 DEF_HELPER_3(itof_d, f64, env, i32, i32)
diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c
index fff29cc25dd1..944a157747cd 100644
--- a/target/xtensa/translate.c
+++ b/target/xtensa/translate.c
@@ -7314,6 +7314,38 @@ static void translate_sub_s(DisasContext *dc, const OpcodeArg arg[],
     }
 }
 
+static void translate_mkdadj_d(DisasContext *dc, const OpcodeArg arg[],
+                               const uint32_t par[])
+{
+    gen_helper_mkdadj_d(arg[0].out, cpu_env, arg[0].in, arg[1].in);
+}
+
+static void translate_mkdadj_s(DisasContext *dc, const OpcodeArg arg[],
+                               const uint32_t par[])
+{
+    OpcodeArg arg32[2];
+
+    get_f32_o1_i2(arg, arg32, 0, 0, 1);
+    gen_helper_mkdadj_s(arg32[0].out, cpu_env, arg32[0].in, arg32[1].in);
+    put_f32_o1_i2(arg, arg32, 0, 0, 1);
+}
+
+static void translate_mksadj_d(DisasContext *dc, const OpcodeArg arg[],
+                               const uint32_t par[])
+{
+    gen_helper_mksadj_d(arg[0].out, cpu_env, arg[1].in);
+}
+
+static void translate_mksadj_s(DisasContext *dc, const OpcodeArg arg[],
+                               const uint32_t par[])
+{
+    OpcodeArg arg32[2];
+
+    get_f32_o1_i1(arg, arg32, 0, 1);
+    gen_helper_mksadj_s(arg32[0].out, cpu_env, arg32[1].in);
+    put_f32_o1_i1(arg, arg32, 0, 1);
+}
+
 static void translate_wur_fpu_fcr(DisasContext *dc, const OpcodeArg arg[],
                                   const uint32_t par[])
 {
@@ -7349,6 +7381,22 @@ static const XtensaOpcodeOps fpu_ops[] = {
         .name = "add.s",
         .translate = translate_add_s,
         .coprocessor = 0x1,
+    }, {
+        .name = "addexp.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "addexp.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "addexpm.d",
+        .translate = translate_mov_s,
+        .coprocessor = 0x1,
+    }, {
+        .name = "addexpm.s",
+        .translate = translate_mov_s,
+        .coprocessor = 0x1,
     }, {
         .name = "ceil.d",
         .translate = translate_ftoi_d,
@@ -7375,6 +7423,22 @@ static const XtensaOpcodeOps fpu_ops[] = {
         .name = "cvts.d",
         .translate = translate_cvts_d,
         .coprocessor = 0x1,
+    }, {
+        .name = "div0.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "div0.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "divn.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "divn.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
     }, {
         .name = "float.d",
         .translate = translate_float_d,
@@ -7475,6 +7539,30 @@ static const XtensaOpcodeOps fpu_ops[] = {
         .name = "madd.s",
         .translate = translate_madd_s,
         .coprocessor = 0x1,
+    }, {
+        .name = "maddn.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "maddn.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "mkdadj.d",
+        .translate = translate_mkdadj_d,
+        .coprocessor = 0x1,
+    }, {
+        .name = "mkdadj.s",
+        .translate = translate_mkdadj_s,
+        .coprocessor = 0x1,
+    }, {
+        .name = "mksadj.d",
+        .translate = translate_mksadj_d,
+        .coprocessor = 0x1,
+    }, {
+        .name = "mksadj.s",
+        .translate = translate_mksadj_s,
+        .coprocessor = 0x1,
     }, {
         .name = "mov.d",
         .translate = translate_mov_d,
@@ -7567,6 +7655,14 @@ static const XtensaOpcodeOps fpu_ops[] = {
         .name = "neg.s",
         .translate = translate_neg_s,
         .coprocessor = 0x1,
+    }, {
+        .name = "nexp01.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "nexp01.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
     }, {
         .name = "oeq.d",
         .translate = translate_compare_d,
@@ -7660,6 +7756,14 @@ static const XtensaOpcodeOps fpu_ops[] = {
         .par = (const uint32_t[]){true, true, true},
         .op_flags = XTENSA_OP_STORE,
         .coprocessor = 0x1,
+    }, {
+        .name = "sqrt0.d",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
+    }, {
+        .name = "sqrt0.s",
+        .translate = translate_nop,
+        .coprocessor = 0x1,
     }, {
         .name = "ssi",
         .translate = translate_ldsti_s,
-- 
2.20.1